From 7e9354be5f0e2210f209ecf35cb141afa55f1eb6 Mon Sep 17 00:00:00 2001
From: roomrys <roomrys@users.noreply.github.com>
Date: Thu, 19 Dec 2024 17:52:41 +0000
Subject: [PATCH] deploy: 7991f14dcf31877f0363f1f2a4392586a04d6eee

---
 .buildinfo                                    |     2 +-
 CODE_OF_CONDUCT.html                          |    36 +-
 CONTRIBUTING.html                             |    44 +-
 ...642a87dad38e9d08625d98d4cf0ffd2b414281.png |   Bin 0 -> 45294 bytes
 ...b9545d9a293bed86bd144682549fe750d7a34d.png |   Bin 0 -> 52093 bytes
 ...a631a9de6f98afaf55b8f4f59ee8351db6feeb.png |   Bin 0 -> 44455 bytes
 ...06868f157496f5fefdd2f4e8ad741179cb6e69.png |   Bin 0 -> 70587 bytes
 ...4303852fcba727e88b0345cabc9999244acf84.png |   Bin 0 -> 55940 bytes
 ...e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png |   Bin 0 -> 24913 bytes
 ...97b24f30f433e554ae3059a3d226b457135a00.png |   Bin 0 -> 34680 bytes
 ...d73bdc855130d58eaa29f55050031f80d0a148.png |   Bin 0 -> 75959 bytes
 ...60ddf3eba8874e355ff047933d9a606da9318e.png |   Bin 0 -> 14548 bytes
 ...00b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png |   Bin 0 -> 169439 bytes
 ...1a4ec1072e73e79320eae51395ff6e3a9fe551.png |   Bin 0 -> 96489 bytes
 ...a678d09aa630125b1a04b4b4429de3f8b4f1fc.png |   Bin 0 -> 75337 bytes
 ...88b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png |   Bin 0 -> 41697 bytes
 _images/bonsai-connection.jpg                 |   Bin 0 -> 113805 bytes
 _images/bonsai-filecapture.jpg                |   Bin 0 -> 68990 bytes
 _images/bonsai-predictcentroids.jpg           |   Bin 0 -> 76547 bytes
 _images/bonsai-predictposeidentities.jpg      |   Bin 0 -> 85845 bytes
 _images/bonsai-predictposes.jpg               |   Bin 0 -> 79334 bytes
 _images/bonsai-workflow.jpg                   |   Bin 0 -> 33481 bytes
 ...d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png |   Bin 0 -> 48824 bytes
 ...7c1961ff1a43b6b62847e515dadb3047025028.png |   Bin 0 -> 29473 bytes
 ...0470e9c921794a84a0af190ce345f8321a14e9.png |   Bin 0 -> 27526 bytes
 ...75704e720b95ba9694ecd512a2cc28d501fbab.png |   Bin 0 -> 68307 bytes
 _sources/api.rst                              |    13 +-
 _sources/guides/bonsai.md                     |    75 +
 _sources/guides/cli.md                        |    27 +-
 _sources/guides/gui.md                        |     2 +-
 _sources/guides/index.md                      |     5 +
 _sources/guides/proofreading.md               |     2 +
 _sources/installation.md                      |   424 +-
 ..._and_inference_on_an_example_dataset.ipynb |     4 +-
 .../_sphinx_javascript_frameworks_compat.js   |   134 +
 _static/basic.css                             |    56 +-
 _static/bonsai-connection.jpg                 |   Bin 0 -> 113805 bytes
 _static/bonsai-filecapture.jpg                |   Bin 0 -> 68990 bytes
 _static/bonsai-predictcentroids.jpg           |   Bin 0 -> 76547 bytes
 _static/bonsai-predictposeidentities.jpg      |   Bin 0 -> 85845 bytes
 _static/bonsai-predictposes.jpg               |   Bin 0 -> 79334 bytes
 _static/bonsai-workflow.jpg                   |   Bin 0 -> 33481 bytes
 _static/css/tabs.css                          |    91 +
 _static/doctools.js                           |   418 +-
 _static/documentation_options.js              |     4 +-
 _static/jquery-3.6.0.js                       | 10881 ++++++++++++++++
 _static/jquery.js                             |     4 +-
 _static/language_data.js                      |   100 +-
 ...69c37c29e427902b24a333a5f9fcb2f0b3ac41.css |  2342 ++++
 _static/searchtools.js                        |   821 +-
 _static/sphinx_highlight.js                   |   144 +
 _static/tabs.css                              |    89 +
 _static/tabs.js                               |   145 +
 api.html                                      |   115 +-
 api/sleap.info.align.html                     |    80 +-
 api/sleap.info.feature_suggestions.html       |   202 +-
 api/sleap.info.labels.html                    |    24 +-
 api/sleap.info.metrics.html                   |    94 +-
 api/sleap.info.summary.html                   |   125 +-
 api/sleap.info.trackcleaner.html              |    52 +-
 api/sleap.info.write_tracking_h5.html         |   102 +-
 api/sleap.instance.html                       |   532 +-
 api/sleap.io.convert.html                     |    54 +-
 api/sleap.io.dataset.html                     |   609 +-
 api/sleap.io.format.adaptor.html              |   100 +-
 api/sleap.io.format.alphatracker.html         |   138 +-
 api/sleap.io.format.coco.html                 |    84 +-
 api/sleap.io.format.csv.html                  |    86 +-
 api/sleap.io.format.deeplabcut.html           |   124 +-
 api/sleap.io.format.deepposekit.html          |    84 +-
 api/sleap.io.format.dispatch.html             |    24 +-
 api/sleap.io.format.filehandle.html           |    76 +-
 api/sleap.io.format.genericjson.html          |    88 +-
 api/sleap.io.format.hdf5.html                 |    88 +-
 api/sleap.io.format.labels_json.html          |   100 +-
 api/sleap.io.format.leap_matlab.html          |    84 +-
 api/sleap.io.format.main.html                 |    60 +-
 api/sleap.io.format.ndx_pose.html             |    98 +-
 api/sleap.io.format.nix.html                  |    88 +-
 api/sleap.io.format.sleap_analysis.html       |    90 +-
 api/sleap.io.format.text.html                 |    88 +-
 api/sleap.io.legacy.html                      |    64 +-
 api/sleap.io.pathutils.html                   |    68 +-
 api/sleap.io.video.html                       |   424 +-
 api/sleap.io.videowriter.html                 |   132 +-
 api/sleap.io.visuals.html                     |    96 +-
 api/sleap.message.html                        |    88 +-
 api/sleap.nn.architectures.common.html        |    66 +-
 ...leap.nn.architectures.encoder_decoder.html |   306 +-
 api/sleap.nn.architectures.hourglass.html     |   194 +-
 api/sleap.nn.architectures.hrnet.html         |   106 +-
 api/sleap.nn.architectures.leap.html          |   100 +-
 ....nn.architectures.pretrained_encoders.html |    92 +-
 api/sleap.nn.architectures.resnet.html        |   244 +-
 api/sleap.nn.architectures.unet.html          |   140 +-
 api/sleap.nn.architectures.upsampling.html    |   124 +-
 api/sleap.nn.callbacks.html                   |   156 +-
 api/sleap.nn.config.data.html                 |   158 +-
 api/sleap.nn.config.model.html                |   564 +-
 api/sleap.nn.config.optimization.html         |   312 +-
 api/sleap.nn.config.outputs.html              |   204 +-
 api/sleap.nn.config.training_job.html         |   132 +-
 api/sleap.nn.config.utils.html                |    56 +-
 api/sleap.nn.data.augmentation.html           |   198 +-
 api/sleap.nn.data.confidence_maps.html        |   200 +-
 api/sleap.nn.data.dataset_ops.html            |   232 +-
 api/sleap.nn.data.edge_maps.html              |   124 +-
 api/sleap.nn.data.general.html                |   124 +-
 api/sleap.nn.data.grouping.html               |    56 +-
 api/sleap.nn.data.identity.html               |   120 +-
 api/sleap.nn.data.inference.html              |    80 +-
 api/sleap.nn.data.instance_centroids.html     |   110 +-
 api/sleap.nn.data.instance_cropping.html      |   150 +-
 api/sleap.nn.data.normalization.html          |   176 +-
 api/sleap.nn.data.offset_regression.html      |    64 +-
 api/sleap.nn.data.pipelines.html              |   424 +-
 api/sleap.nn.data.providers.html              |   156 +-
 api/sleap.nn.data.resizing.html               |   206 +-
 api/sleap.nn.data.training.html               |   100 +-
 api/sleap.nn.data.utils.html                  |   100 +-
 api/sleap.nn.evals.html                       |   160 +-
 api/sleap.nn.heads.html                       |   382 +-
 api/sleap.nn.identity.html                    |    72 +-
 api/sleap.nn.inference.html                   |  1106 +-
 api/sleap.nn.losses.html                      |   106 +-
 api/sleap.nn.model.html                       |    86 +-
 api/sleap.nn.paf_grouping.html                |   319 +-
 api/sleap.nn.peak_finding.html                |   163 +-
 api/sleap.nn.system.html                      |   122 +-
 api/sleap.nn.tracker.components.html          |   171 +-
 api/sleap.nn.tracker.kalman.html              |    64 +-
 api/sleap.nn.tracking.html                    |   265 +-
 api/sleap.nn.training.html                    |   326 +-
 api/sleap.nn.utils.html                       |    88 +-
 api/sleap.nn.viz.html                         |    88 +-
 api/sleap.skeleton.html                       |   599 +-
 api/sleap.util.html                           |   168 +-
 datasets.html                                 |    72 +-
 docs/_static/bonsai-connection.jpg            |   Bin 0 -> 113805 bytes
 docs/_static/bonsai-filecapture.jpg           |   Bin 0 -> 68990 bytes
 docs/_static/bonsai-predictcentroids.jpg      |   Bin 0 -> 76547 bytes
 docs/_static/bonsai-predictposeidentities.jpg |   Bin 0 -> 85845 bytes
 docs/_static/bonsai-predictposes.jpg          |   Bin 0 -> 79334 bytes
 docs/_static/bonsai-workflow.jpg              |   Bin 0 -> 33481 bytes
 docs/_static/css/tabs.css                     |    91 +
 genindex.html                                 |   145 +-
 guides/bonsai.html                            |   512 +
 guides/choosing-models.html                   |    30 +-
 guides/cli.html                               |    81 +-
 guides/colab.html                             |    28 +-
 guides/custom-training.html                   |    24 +-
 guides/gui.html                               |    52 +-
 guides/index.html                             |    36 +-
 guides/merging.html                           |    24 +-
 guides/proofreading.html                      |    42 +-
 guides/remote.html                            |    38 +-
 guides/skeletons.html                         |    28 +-
 guides/training.html                          |    24 +-
 guides/troubleshooting-workflows.html         |    24 +-
 help.html                                     |    58 +-
 index.html                                    |    48 +-
 installation.html                             |   370 +-
 notebooks/Analysis_examples.html              |    68 +-
 notebooks/Data_structures.html                |    52 +-
 .../Interactive_and_realtime_inference.html   |    56 +-
 .../Interactive_and_resumable_training.html   |    68 +-
 notebooks/Model_evaluation.html               |    48 +-
 notebooks/Post_inference_tracking.html        |    48 +-
 ...g_and_inference_on_an_example_dataset.html |    66 +-
 ...ning_and_inference_using_Google_Drive.html |    58 +-
 notebooks/analysis_example/README.html        |    22 +-
 notebooks/index.html                          |    50 +-
 objects.inv                                   |   Bin 20518 -> 20650 bytes
 overview.html                                 |    24 +-
 py-modindex.html                              |    25 +-
 search.html                                   |    20 +-
 searchindex.js                                |     2 +-
 tutorials/analysis.html                       |    28 +-
 tutorials/assisted-labeling.html              |    26 +-
 tutorials/initial-labeling.html               |    32 +-
 tutorials/initial-training.html               |    30 +-
 tutorials/new-project.html                    |    32 +-
 tutorials/proofreading.html                   |    26 +-
 tutorials/tutorial.html                       |    24 +-
 184 files changed, 26732 insertions(+), 5883 deletions(-)
 create mode 100644 _images/1b0b6e7baba606deccb7311077642a87dad38e9d08625d98d4cf0ffd2b414281.png
 create mode 100644 _images/35c10296694a5eb3635e87b56ab9545d9a293bed86bd144682549fe750d7a34d.png
 create mode 100644 _images/3f78e9f7b93df6d2f3452d162ca631a9de6f98afaf55b8f4f59ee8351db6feeb.png
 create mode 100644 _images/42e0a303da85179624b2247d2d06868f157496f5fefdd2f4e8ad741179cb6e69.png
 create mode 100644 _images/4c35c39e8dcf293ab2362ab3534303852fcba727e88b0345cabc9999244acf84.png
 create mode 100644 _images/507b37f5c35704f694d1e81f64e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png
 create mode 100644 _images/6a87f28526a92715315c4766f197b24f30f433e554ae3059a3d226b457135a00.png
 create mode 100644 _images/6dcf42bdd95e0ded30ed1fb43ad73bdc855130d58eaa29f55050031f80d0a148.png
 create mode 100644 _images/7d06f5baca09a6bf3099c8ae1060ddf3eba8874e355ff047933d9a606da9318e.png
 create mode 100644 _images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png
 create mode 100644 _images/b21d966bde1fb74635e96843f91a4ec1072e73e79320eae51395ff6e3a9fe551.png
 create mode 100644 _images/b44ac03a6b1270ea102001c424a678d09aa630125b1a04b4b4429de3f8b4f1fc.png
 create mode 100644 _images/b63ef4ea81a463798c2aa2540688b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png
 create mode 100644 _images/bonsai-connection.jpg
 create mode 100644 _images/bonsai-filecapture.jpg
 create mode 100644 _images/bonsai-predictcentroids.jpg
 create mode 100644 _images/bonsai-predictposeidentities.jpg
 create mode 100644 _images/bonsai-predictposes.jpg
 create mode 100644 _images/bonsai-workflow.jpg
 create mode 100644 _images/ca67ed7d78a0a607f47eea0f22d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png
 create mode 100644 _images/d2b61f464140b9d880c782c1347c1961ff1a43b6b62847e515dadb3047025028.png
 create mode 100644 _images/d8517178d972bc0703903cfab70470e9c921794a84a0af190ce345f8321a14e9.png
 create mode 100644 _images/dc4bfaee117747c1eeba0b62c375704e720b95ba9694ecd512a2cc28d501fbab.png
 create mode 100644 _sources/guides/bonsai.md
 create mode 100644 _static/_sphinx_javascript_frameworks_compat.js
 create mode 100644 _static/bonsai-connection.jpg
 create mode 100644 _static/bonsai-filecapture.jpg
 create mode 100644 _static/bonsai-predictcentroids.jpg
 create mode 100644 _static/bonsai-predictposeidentities.jpg
 create mode 100644 _static/bonsai-predictposes.jpg
 create mode 100644 _static/bonsai-workflow.jpg
 create mode 100644 _static/css/tabs.css
 create mode 100644 _static/jquery-3.6.0.js
 create mode 100644 _static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css
 create mode 100644 _static/sphinx_highlight.js
 create mode 100644 _static/tabs.css
 create mode 100644 _static/tabs.js
 create mode 100644 docs/_static/bonsai-connection.jpg
 create mode 100644 docs/_static/bonsai-filecapture.jpg
 create mode 100644 docs/_static/bonsai-predictcentroids.jpg
 create mode 100644 docs/_static/bonsai-predictposeidentities.jpg
 create mode 100644 docs/_static/bonsai-predictposes.jpg
 create mode 100644 docs/_static/bonsai-workflow.jpg
 create mode 100644 docs/_static/css/tabs.css
 create mode 100644 guides/bonsai.html

diff --git a/.buildinfo b/.buildinfo
index 82e156059..344c2f8f3 100644
--- a/.buildinfo
+++ b/.buildinfo
@@ -1,4 +1,4 @@
 # Sphinx build info version 1
 # This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
-config: 9d36406655356348384633483bebff27
+config: 061a3cab3e5ef79e32ad32ceda8f5410
 tags: 645f666f9bcd5a90fca523b33c5a78b7
diff --git a/CODE_OF_CONDUCT.html b/CODE_OF_CONDUCT.html
index 0a812695b..0b9108aa1 100644
--- a/CODE_OF_CONDUCT.html
+++ b/CODE_OF_CONDUCT.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Contributor Covenant Code of Conduct &#8212; SLEAP (v1.3.3)</title>
+    <title>Contributor Covenant Code of Conduct &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'CODE_OF_CONDUCT';</script>
     <link rel="canonical" href="/develop/CODE_OF_CONDUCT.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
     <link rel="index" title="Index" href="genindex.html" />
     <link rel="search" title="Search" href="search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -333,13 +331,13 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="contributor-covenant-code-of-conduct">
-<h1>Contributor Covenant Code of Conduct<a class="headerlink" href="#contributor-covenant-code-of-conduct" title="Permalink to this headline">#</a></h1>
+<h1>Contributor Covenant Code of Conduct<a class="headerlink" href="#contributor-covenant-code-of-conduct" title="Permalink to this heading">#</a></h1>
 <section id="our-pledge">
-<h2>Our Pledge<a class="headerlink" href="#our-pledge" title="Permalink to this headline">#</a></h2>
+<h2>Our Pledge<a class="headerlink" href="#our-pledge" title="Permalink to this heading">#</a></h2>
 <p>In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to make participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, gender identity and expression, level of experience, nationality, personal appearance, race, religion, or sexual identity and orientation.</p>
 </section>
 <section id="our-standards">
-<h2>Our Standards<a class="headerlink" href="#our-standards" title="Permalink to this headline">#</a></h2>
+<h2>Our Standards<a class="headerlink" href="#our-standards" title="Permalink to this heading">#</a></h2>
 <p>Examples of behavior that contributes to creating a positive environment include:</p>
 <ul class="simple">
 <li><p>Using welcoming and inclusive language</p></li>
@@ -358,21 +356,21 @@ <h2>Our Standards<a class="headerlink" href="#our-standards" title="Permalink to
 </ul>
 </section>
 <section id="our-responsibilities">
-<h2>Our Responsibilities<a class="headerlink" href="#our-responsibilities" title="Permalink to this headline">#</a></h2>
+<h2>Our Responsibilities<a class="headerlink" href="#our-responsibilities" title="Permalink to this heading">#</a></h2>
 <p>Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.</p>
 <p>Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.</p>
 </section>
 <section id="scope">
-<h2>Scope<a class="headerlink" href="#scope" title="Permalink to this headline">#</a></h2>
+<h2>Scope<a class="headerlink" href="#scope" title="Permalink to this heading">#</a></h2>
 <p>This Code of Conduct applies both within project spaces and in public spaces when an individual is representing the project or its community. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.</p>
 </section>
 <section id="enforcement">
-<h2>Enforcement<a class="headerlink" href="#enforcement" title="Permalink to this headline">#</a></h2>
+<h2>Enforcement<a class="headerlink" href="#enforcement" title="Permalink to this heading">#</a></h2>
 <p>Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by contacting the project team at <a class="reference external" href="mailto:talmo&#37;&#52;&#48;salk&#46;edu">talmo<span>&#64;</span>salk<span>&#46;</span>edu</a>. All complaints will be reviewed and investigated and will result in a response that is deemed necessary and appropriate to the circumstances. The project team is obligated to maintain confidentiality with regard to the reporter of an incident. Further details of specific enforcement policies may be posted separately.</p>
 <p>Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project’s leadership.</p>
 </section>
 <section id="attribution">
-<h2>Attribution<a class="headerlink" href="#attribution" title="Permalink to this headline">#</a></h2>
+<h2>Attribution<a class="headerlink" href="#attribution" title="Permalink to this heading">#</a></h2>
 <p>This Code of Conduct is adapted from the <a class="reference external" href="https://contributor-covenant.org">Contributor Covenant</a>, version 1.4, available at <a class="reference external" href="https://contributor-covenant.org/version/1/4/">https://contributor-covenant.org/version/1/4</a></p>
 </section>
 </section>
@@ -436,7 +434,7 @@ <h2>Attribution<a class="headerlink" href="#attribution" title="Permalink to thi
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/CONTRIBUTING.html b/CONTRIBUTING.html
index 3e6261c52..fa15f0b81 100644
--- a/CONTRIBUTING.html
+++ b/CONTRIBUTING.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Contributing to SLEAP &#8212; SLEAP (v1.3.3)</title>
+    <title>Contributing to SLEAP &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'CONTRIBUTING';</script>
     <link rel="canonical" href="/develop/CONTRIBUTING.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
     <link rel="index" title="Index" href="genindex.html" />
     <link rel="search" title="Search" href="search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -339,11 +337,11 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="contributing-to-sleap">
-<h1>Contributing to SLEAP<a class="headerlink" href="#contributing-to-sleap" title="Permalink to this headline">#</a></h1>
+<h1>Contributing to SLEAP<a class="headerlink" href="#contributing-to-sleap" title="Permalink to this heading">#</a></h1>
 <p>As our community grows it is important to adhere to a set of contribution guidelines. These guidelines may change as needed. Please feel free to propose changes to source code in a pull request!</p>
 <section id="table-of-contents">
-<h2>Table Of Contents<a class="headerlink" href="#table-of-contents" title="Permalink to this headline">#</a></h2>
-<ol class="simple">
+<h2>Table Of Contents<a class="headerlink" href="#table-of-contents" title="Permalink to this heading">#</a></h2>
+<ol class="arabic simple">
 <li><p><a class="reference internal" href="#code-of-conduct"><span class="std std-doc">Code of Conduct</span></a></p></li>
 <li><p><a class="reference internal" href="#contributing"><span class="std std-doc">Contributing</span></a></p>
 <ul class="simple">
@@ -356,21 +354,21 @@ <h2>Table Of Contents<a class="headerlink" href="#table-of-contents" title="Perm
 </ol>
 </section>
 <section id="code-of-conduct">
-<h2>Code of Conduct<a class="headerlink" href="#code-of-conduct" title="Permalink to this headline">#</a></h2>
+<h2>Code of Conduct<a class="headerlink" href="#code-of-conduct" title="Permalink to this heading">#</a></h2>
 <p>Everyone contributing to SLEAP is governed by the <a class="reference internal" href="CODE_OF_CONDUCT.html"><span class="doc std std-doc">SLEAP Code of Conduct</span></a>. As such, all are expected to abide by this code. Please report any unacceptable behavior to <code class="docutils literal notranslate"><span class="pre">talmo&#64;salk.edu</span></code>.</p>
 </section>
 <section id="contributing">
-<h2>Contributing<a class="headerlink" href="#contributing" title="Permalink to this headline">#</a></h2>
+<h2>Contributing<a class="headerlink" href="#contributing" title="Permalink to this heading">#</a></h2>
 <p>Github has made it easy to separate issues from discussions. Generally speaking, if something warrants a pull request (e.g. bug fixes, TODOs) it should be submitted as an issue. Conversely, general questions about usage and improvement ideas should go into discussions. These both act as staging areas before adding the items to the SLEAP Roadmap. Once added to the Roadmap, items can be given priority, status, assignee(s), etc.</p>
 <section id="issues">
-<h3>Issues<a class="headerlink" href="#issues" title="Permalink to this headline">#</a></h3>
+<h3>Issues<a class="headerlink" href="#issues" title="Permalink to this heading">#</a></h3>
 <ul class="simple">
 <li><p>Check <a class="reference external" href="https://github.com/talmolab/sleap/issues">open/closed issues</a>, <a class="reference external" href="https://github.com/talmolab/sleap/discussions/categories/ideas">ideas</a>, and <a class="reference external" href="https://github.com/talmolab/sleap/discussions/categories/help">the help page</a> to make sure issue doesn’t already exist / has been solved.</p></li>
 <li><p>Create new issue using the <a class="reference external" href="https://github.com/talmolab/sleap/blob/arlo/contributing_guide/.github/ISSUE_TEMPLATE/bug_report.md">issue template</a>.</p></li>
 </ul>
 </section>
 <section id="discussions">
-<h3>Discussions<a class="headerlink" href="#discussions" title="Permalink to this headline">#</a></h3>
+<h3>Discussions<a class="headerlink" href="#discussions" title="Permalink to this heading">#</a></h3>
 <ul class="simple">
 <li><p>This is a place to go to ask questions and propose new ideas.</p></li>
 <li><p>3 categories: Help, Ideas, General</p>
@@ -383,8 +381,8 @@ <h3>Discussions<a class="headerlink" href="#discussions" title="Permalink to thi
 </ul>
 </section>
 <section id="pull-requests">
-<h3>Pull Requests<a class="headerlink" href="#pull-requests" title="Permalink to this headline">#</a></h3>
-<ol class="simple">
+<h3>Pull Requests<a class="headerlink" href="#pull-requests" title="Permalink to this heading">#</a></h3>
+<ol class="arabic simple">
 <li><p>Install source code <a class="reference external" href="https://sleap.ai/installation.html#conda-from-source"><code class="docutils literal notranslate"><span class="pre">develop</span></code> branch</a> and follow instructions to create conda env, etc.</p></li>
 <li><p>Create a fork from the <code class="docutils literal notranslate"><span class="pre">develop</span></code> branch.</p>
 <ul class="simple">
@@ -424,14 +422,14 @@ <h3>Pull Requests<a class="headerlink" href="#pull-requests" title="Permalink to
 </section>
 </section>
 <section id="style-guides">
-<h2>Style Guides<a class="headerlink" href="#style-guides" title="Permalink to this headline">#</a></h2>
+<h2>Style Guides<a class="headerlink" href="#style-guides" title="Permalink to this heading">#</a></h2>
 <ul class="simple">
 <li><p><strong>Lint</strong> - <a class="reference external" href="https://black.readthedocs.io/en/stable/">Black</a> version 21.6b0 (see <a class="reference external" href="https://github.com/talmolab/sleap/blob/develop/dev_requirements.txt">dev_requirements</a> for any changes).</p></li>
 <li><p><strong>Code</strong> - Generally follow <a class="reference external" href="https://peps.python.org/pep-0008/">PEP8</a>. Type hinting is encouraged.</p></li>
 <li><p><strong>Documentation</strong> - Use <span class="xref myst">Google-style comments and docstrings</span> to document code.</p></li>
 </ul>
 <section id="thank-you-for-contributing-to-sleap">
-<h3>Thank you for contributing to SLEAP!<a class="headerlink" href="#thank-you-for-contributing-to-sleap" title="Permalink to this headline">#</a></h3>
+<h3>Thank you for contributing to SLEAP!<a class="headerlink" href="#thank-you-for-contributing-to-sleap" title="Permalink to this heading">#</a></h3>
 <p>:heart:</p>
 </section>
 </section>
@@ -502,7 +500,7 @@ <h3>Thank you for contributing to SLEAP!<a class="headerlink" href="#thank-you-f
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/_images/1b0b6e7baba606deccb7311077642a87dad38e9d08625d98d4cf0ffd2b414281.png b/_images/1b0b6e7baba606deccb7311077642a87dad38e9d08625d98d4cf0ffd2b414281.png
new file mode 100644
index 0000000000000000000000000000000000000000..6d63e3f04297a1c1cb7d3cdab1165e9351c50eb5
GIT binary patch
literal 45294
zcmX_nWmH>j({%+e#idZBxE6{-v7*6(OMw6_5L}8o+{N9cxVt++ON+ZxC{}`da0&1p
zo^P%1M;2KjM{=&Y_RQY1Crnit^cssC>)EqsujS>W)t^0kj`Q@wKm$H;ue~V*{s=kC
ze0J8bH*<FT>Ii)X{_6bA#@^Y+()gV#)X~Y(-j0WZkAsK(orSaWH>Z!BoVNe}3mo>2
z=A5H!uqogqnBU}dot`~=Mfmi3{wyuy&9i4U`ts5en(i5g_xh18y78+ay+Wx>;|!du
zl^h1Me9NMvBnuR36l0+;rvotp7<m)FzNtgG{=N~g`dWPG1Jdp%8ulMwzBTigbNMKL
zxG{%crb$WW^w5M5=<_U#E^Eyi7+e^fWqEL)WgLcV_-)!h^31K=ZRpGf|Kt2m+U#h!
zwI`ey5%alrj*1UF-6QM?I^LnQ{RDrs7g(e0$agz`^g35|Xb0Xz^B1{=!=FBD;Ofwz
z0KV3dAGMpg?*6|cl#Yq|pANP#Sv+dJ*%(X+4Sf?Oi;u6S53@~UoZENtP@dN=pOuXw
zi;|t>1@nTZSrvJyR`*~Ro<7yS9bFwAl=~O0v)TqaH=gF^=1i`xuImQ}6^RSHU?s+c
zH!=OS5Dt?@y<T?_yQGvxPHWEnw^6cO<Wa`f)@;DpmzI_qZi(;<57#%X8d?3!H@zNt
zro8q7+<))0gRW*S5$+UlgrIi0mLiUG<s6CFz9Vp~-XWe*d;6lh8~*Up%1U&4dioUS
z_RGIQ1rVle(lWP3y(HIbFT=;!(~v3PvV&rqj0rT^Og14KU9y?s`TXDS$uJ2}&(d=;
zGc!X&KXas%ev*|=jm4iNc*n|sxp8sfV*PB^^)=G@V^n#0d1$DY<e{X|TNE9^_M)rV
zrv3eWuPedY1<vEE*xJR##j?5M!&yES?rIY+FRz-f1B?7rF~js4qRQ=a^Ib|`<w?9D
zR+}xgyI%bs)sGCg4;(i;hd%m4L{dd9t$Tk9-e-o><^4^&aT1f~FBCIV_&T$f^Uc=Q
zHYsd7`8$TZ!mr9oG^4p%v#)K7nV4Q?hlS2oJO0Olw8MdP*6Bk<#3c6Z*r%nIX@3?M
zS?}jXS!QT7rA&~S7r8~TA#)Ojak!1!E*;fk_Vuc%L#~=kREp77_dajeKZkoOm)RSH
z*N?a1y_TdfNP^B;5$;W$8jPP(+ON*dX{%9?A-AMg7Z<H^iDsRAe3t8i8y*M<x{|?S
zO%X>SY_FTtj$``aUWdadx#h|$`jy(XKqv?c3p+SDX#}Qv=T6~ax|GCJwbyYCK8)R}
z($S!~aav!@SZGA7%=2%nO)s^!Q-DDK)MxN*uX5(HZ8%`w7Z(X<zo7I^lv?Po6-h5x
zaU{K}E6mjHn5TbM6<$C_B?Xn}1Gp=9uL^VR?d*<p`M``4UVb3$KX-7KIU;`*O@jnr
zpW4{luM*acfCsOIQ<e0pVww~DKL<4Uj^1yR&n7C^j@(wP6<wH~tD@3wzV9vf5*t&m
zYM<Zst)+_Y8vcz(p99+%_Yg{zTzAqvc%L7oqXqdZmUu2kW>`M_7+una`atM_&Y+uJ
z(M1I6d9S{(y!^gfZI&lh>1Y;gQ9?bN--U`OrEswawnEw#KVCT|J#XFv#L%$s;^N|+
zZyn`cF&ky>2h@+f)L;x2MbQ)(n4Rfud78euQ+IIt^u?J=JeymzgDvf(j6-1-Dr?)u
z53~q{pV@Sl>~$U+3Ubg>A+n!?rUY$mZ5<a$zHTe1&~Vy-7Ef0>T?%AXuq$TGP&sCz
z3O(+B1X{NC_Wh>ZED5?VZ_an8xj*Ua>vI85gpl33{1pXp=pJ>HEGlu$^3<&qO0g_^
z%CF56al%%P{7?I{AaCXSE6hn5XMx!{r7ixr4zPSxFq=Lo$2UqGI18d1ow~<c%$~(u
zM~PKYHYB*3@i974evI=J<T7LdF*54k-QE3CxU^s94o!{WMcML%*+525BV`Z27|xM#
zxrk+_Xc0(wvtY5HG%8{-rT>Ar&a7o-Wvyrsu|`xyU$NBJP+1gylWqubOIHfD@dtiL
zv$i_#E+}6uq8PY-Wm0ivWu<mx2(%2BFNlSRFMV`<W23;uk3YQ^*u8u;i@adVkha*p
z-!}Xaz!uiz4=P$oW+t_IiH|;M!7soqG&|KfR3tnd)M9l-m&g9K+|bRk<eP|CzhLG=
zG4g6q|F{fGI0cNN9m_S9y}hdIk^Wg*w5zS{?e_Mz%+Buddy<IHmD*g{<0df3ij<%;
zQ&UruvAKD%5(Ri2xZa<Ig|Mp?FHU6@6@CBXZhG(xHvdP+;a=A|Mg2m>=XNFe`lZ9e
zLq}UTH-W7)kH~&_OM81{L`1|NAe!!th14qm%&V%dHUWaDsk!-;%B=R59-c(S{C@m#
zo_S#nEbX@z*Zb{Or1X+e(}rouCdrJdYKpKXU39l+haGByZc)}I3e&Ht`tsQ_1|Bkk
zr_HHXsmJVJ6*T`AzjD)zl^^t)UCJW0eclNj9UZYXnl2$i9w<2AZ1vM-LTJLY;FScJ
z$1Bq=H4}j7RBW4oH!dC?WM?-bj0y4*v;;3IKMbgIJ#b|z!`IjIW-oy|*Y?xwv*2a5
z{aig?`Pu(AHW%H{L_8=*Sy_4g^t28t>P=fa0wAaH==sX%DEQ#upl5bgOYz!Lt9aTH
z*c~MFt_}`ri6zGD5xum)MKZFp`+@IyBoEKwt0m4`HHHIN4IniTzY0nr%pf=uZQfuB
z<I8mBAev(<2w8RB$iQn&qY`@Vc`3L+s?t>Ai=oUtJX|YG6Ps<|O4hRx@PAmN(Ek*6
zPfluPV>rkQ1pjnb-C1G3l*Jhq!d$u))(D=k;0*_U%ADzXftaZ1`^5|ibzQ{r^0F~;
zAca}IZYt9MU_>q{$TsaSJr_s1ArR)K++n-Ui$i)8bsu|H{cj!~fDig(sQZBvk#ZUm
zC2k&DJ94HB1FLHTV<8XNbkv1>V4>=j0L~_%As`@t+Ne}K{e;}#^;^FKX_Wad@H6L5
zSAaWKlLZ!!88*^ldaCvIt~qz{Ze5^X3yUkSgMxFVG#lS-C_zp-TuyG-U}{8=$Ksgf
zrKd7;+E9dD>{Y~Bz&H-qUV%iFLN;1iw*+fk5$wXJ(4e`w+2ms;$$rUp2fH#kX+h+!
z=@{^kzn1V;GK!=uNCV&vCHR(SSy$uTN=ZPD8Lx`&(NBFFm-<U96OL#Jay&oUpRazU
z9^O)a1o^I8Js(u99!X@%g3;vCVc0)AJKJF@z@o)M#;U5!nlwC17Zs3LY5H!Rqh-|9
z+=7*^i?jEL3nwM2%2QA_UI%8c?u?c7s=IQ2pCnDo=nbLMl`>qyr0-7v1hdsC<1K6X
ztI_)LgM{Its}8mzN%DrWy08lXvEFBA6Y9?Cj{^}xDhKZR`xREDXf13U7D}*27q^RZ
z(~iUS-YJR*{i`2jfE!hUfO{Y@KeW37*p<MJb@=v8b%=1u2zUp?^Ym_mNVaalh<`%A
z`U|SFSQ(<n_D3sW9qR!m-1E~#07)Er>xeOQiVTi0lbBLwd4`DMjLtfv`r!oWIdvC4
z!jTeT_F4JP4i1x8v1Oyfo7}*jRrCSYwfA!@brtUqvdYIzeseZn&lrx+vSO}$d(spJ
zs#TW|el<#*CkWw4xk5|7A;vkpsM2SPn+OO5lV?wQaK+meZWWz4s#`hEy65_;>zgY`
zXa}qQxKB_84iXd<jRN9m2@?k1=%~<7Da>IwC8pPLJURmMUca|k%<td7XY_=*lBJ;C
z1$mP@t}aku&E1qX@xQkOjfU2~kNz4QyDV3eJ;rIkbPW5*p!=<$_q1?nFk-<*TYE>u
zeeaYbORZMe@ryVGtfV6i9ir?i)8iS!8XZOO(`rsLkxwq_D^toabY%O?i?E6D+A}!Q
zp|EFd)wtH8@VUl9q^#RRQ3|J{>=_lhcKClV8uD*g<iw?PgfLmwl;`@}+1pv}nwbid
zf720sitia?#E~bzuUD=*-Beia{><#1|174xO-V^X++OT^!3R5Wa<MmO6Vd*PH6TMR
zXi$=;`4C~Bqyczzb#sf5IBinp-;p#gcjDV%hiwT<@~MC7)c->)YGP=>D=4U-zRmp@
zLC)o+lKGWZ9lYSoC&rcy1(FGAM5(nECmFl@8;531KJw))yB%G|g8>zD5y{Q5?K5tR
z@T_t>4#MgF#B4Tcpq?QFfk0ZOYWYCkJ??t+^+xoPYIWj^vGvi5u^tuh+uPdq4ySO!
zXUlbZn~POXF`iw>3_2Vn5TON(fSppfhTU<A$%rLL9bS0a$V&H?Kla{{L#u4eb_D+|
zW5BxcR-q^?0an7>%G$WqgkO0_Z7X8or49~(y$A=Cu{#4^X2Bu<SxhU<`A{<0gz&+U
zXNt@G&}JJcDXGi{K7>}WzegIcy7B2FQ4KM+_7}~EDm2w&BLf3sad)7`2J%aVcKO_#
zAGii_%|8oa6AC}W0}G*08dGY`mFrY{GNoAnyVr)F%{#7!og~5$tg0`fD*lIbDBzmm
zpZQ*^{YRoM=7N<gvG}zZ+2-zz`uh4rbLrQ&ei)^dRTi4nZ43B^&q~<hpAsIBtl$8?
zZYrymhPfRnsOO%Loet`7-ZZlboW_n1e6l#OZiridBT06C<%NGJ=Xp66EB_mX4($87
zL>ry=1W1k->_H)Np9pmy%8ls(JYqOQ+}?gn9^MrFAxIvj@Fn6`Sr|Y76qJKb87Vw3
z5j9UJ$sQEnjxG6RwV|-HRloNQJ~0IoZnj$2o8Nhk-`uP3Dy&0aI5gKrm>Zg<(5y>O
zg%c+gtUF0I20o6=X&D9f^uZlOe3612oSg831NNe6OU2+qC9B~XUt6(kkaPpVg?9!e
zm)e4vT`{&N_4*&dzR3%}xo>`kN^3sJ@P{2dS`@d@%15RI`GezQP!Hv4Y>=tlcXm3B
z8NQl2Y;_8Cs=5q-7%}AZ?KJw%?YEio<)(4?LXo>|tLCN4lCbeHsAMbQdOo9~bRy>O
zdtr#$TFf4lT^60EmRib~VuPvj1dmEB%=*ljf0tJl+sb-32_q0IlnbSmCI0uU6*0B%
z^;b4|WecU_h-?7SAmy-oATmm8zxZ9a1%E*@<=IBI?MSqlgkX478n28ef{H#{Nwa}o
z;ZKfzX@|kW;BdHCZ0cN%Wg0hPSA3|<(9kd$TQsSZL8I|*Z?Eh#9f+_@k&6HyRY>DU
zI<vai{6+L)wb-D70t+{fqe$nl$;ndm@bEB!y9R}vg~AXhHL6U}mT+8}0=Mep%<asT
z+iUPQg0j{61$H3ol^a-f5iIZ`Q2()j;@ONU^U~Y(D%Z=aJHs~}$$;f2hPX6Ko~|Dm
zFz%~cj7>+PL;5N`P|B_l@cz<+#JZ2ijHqYFiyJFFk~o#s+UCKAyg#z$bD?4(RuuF}
zga75Q@#*S-vE!<CeVX~nzmTej#OglYE~U_5No~9+xExw`6f?6$fc?~eq-8)nwLPdR
zw{Z-qb!4yWA6kg3BwSAWn366#L_E)##cp@<y^e>M<Jk%>xwl+=eOKbySmb+KS}c!m
z&bATjy%Fo$uV9|qJaMn?vIpP${aH=(FCjCd-J7%^>hJHra#vmy?2CJST67Fry6UD~
z?5?$`spq!p%&2(i?v`NDuj&R$(?L^p<P7la{s1JlqoYGTIuPpT*QNA~Cc3inuahwE
zLZCWHgp6f6b*QA<Hc&==+SIFCVmt~ttBu)cwoY$e(eagO(xcyC=(M2AxUL+mZ)?+M
z^zV}>2QO~to;IecCPb?mh?HM6mXD{uDZ22a<TPv#1qzdf6(-uX2cp5G%YUKSogZm8
z<u^P4lmq_Dy!#?Nq*Q4T$mUvu6lUT<nn}E54#i()aS%~w4g8>z&!NUvNBEX>y+9>q
zT>H6aa&nTfT1)<A3OK&P<RLO4`Sy*)v{=-ng=XMPfU6{)Qj|6&5dDS9EL4~+mqiw9
zLjN?I)`SIuC@Clsy`u!|R|+hr_)wRj@*%9ai2vlB>v$oprqSWYWzJA0U!E3)D|XtU
zGNPB|M8?WQoWif(%!M9XLadHkBpiPiwgkJ&&P#Zu4G{Ky3MC2My1cx6;sFJT!c$u5
zd}zF~8>FwJmRLn~_U(fwGCN&+b3b<r$V78kon*|H(zl$&kbENXC)XmB@MT6f70V_q
z%xOJ8=<Ss$1u1}R^Nfza8638Dc1F)v8}(max{CPR@<vBT>v>S(2;HHjj$=FE-$*g!
zy3HEuaX0E$y?p<{KoYZ%$o~$}x}_Vr(lsJ_wjIo@eNjB#VRGkfjA=OqUaRYLC%|Vg
zEWi{=Ra)PjE=JsJXOqHIz2z|?Is7Low4eE$-dcXs{gjS2pAy`b_}(K9CaC;&+rc|i
zQ;I0P;rG`Recq}Zi#0$H!AI+cVzn2S?}Oeph!_IW4CVt+nfzH@<s_80)ocsWyu7-C
z1F`Dh;^Ja|su8-o$}K<ewL*?f+5m>dnhwH&04c&y-lGf$M!Gn{u!0eOmbTXBAD%y1
za($=w9<Lgx%!l>N=T_dc(Rj`=ZR@AlKo$IZ!Uk^LNh+$QEp~5#0?GT~X1nr14u&!4
zv8;Kp5!BZnoYtd}LY9<78N4~vy?>M!MHJ$8ZA*kHtV4L(sf<^2^Uf6Wqa;S!J57$M
z+g8ThEtR6l1vD(!w&Vv3&u`!GfK==7R8WOlo#M-Q7(tgc84cK4HO96N(pFqZC730J
zC4Cltc*%>DalWeFv|v?av3HYc$*8&&3Qlczfl*4u>QrLmr$hx2&>p#}I%!3vb3zjv
zY`_e{RI6qz*2gi}$wZE>V<4kd7PeI5ZyU0w+b*^P@z=89CKi3ms_UPGO@b{tLN9m6
z2)^8;p5GA_sLvJY^T2rU-HJVTeXP!b(9SY>_DzH;WvxVPV}k^2`46~fsBX-mxebF9
z5f0r&v$ru;s{IQhKY+;af;H+Sw5k{T!3KI_#Nqhsr&pvAGD1iYQx#WLj>fF%#f&+c
z>|pVT!88lxz<ktTklD!vnMFs^7~tm=DGk(;3?&tJywBEauf3g}HL;VcI7uRY(PjQ(
z6dLf)Y=lYeMQF?n5PQhtDhREb{*+#EFgWuv%ue}N@F)1l>BZ2g{O7_$*jAMYiUe4f
z5VIal|0QdAX={BTQ{~LHDG)6&`T6+--=OI(Lk&$@^(rfVq&CU$*c2qgi;03L3l%gL
ziuZ*4VBxpwo1xWD)c;_i-u>njnvt2=$lgfQo4};9zO_|U5U7Fa)F@woZ9w~}abB&r
zs7|h72Z77#$%;RYGVAW7B`+;4v#{}8OxkiOvh)(+yOrkg7KtJvN;j{{orR#4qB9a+
zR}Hw8i`ycHy;Un`BE(XrQqCD+)u0*p$i1@>t~!x1r)Qp|@UDcpofGR_R#p~?`?$nJ
znCZqOAXc?MKHOEF$wPE!(9}Q=(7D8;nAjHr5<)6GUmfU+?;uI!QFbb{!(&cbFo40o
zXLz@AX0N9P>w!!i`TylRH815o_nVLBT;>{&h>f6KKV>sf<NYG-zJKu%D4&3|-j6^y
zu-Qv_h}pdX2(iRZpg^syuO|e!>PowZhACu7wdADZYK09+s(?Hl<~;8GoT1}$<}~B{
zEo)Wx<T;*`^;H|%%-LO=F@<tNnzw8h!@<l6u+cUMEdE19-#+ue#2z}e#{Dm=0|v8v
z{bNxKkPz>NH|yBjs3<5J5|giKHh*vG>XGPJ95}fy->vYG5PZH+BQfwvvggQ7X9^#L
z=6+K27%{7Vs-yvcZx01lHQurM5MDmJ94u}$Rmh+tY;9*XI(wqUcI;)QN&(&z%T|kw
z=##cB;G%7vD#Mt>HlgDdjAu#O*IyPbG}`bWkL+$)4{f=(Kmu$QDEi!wJFcD(-?%n3
zh0};+lSWJ}wQTI{?5qspbC|yLr#7@npO)sI;}WCvztep(O*hRv>a68lE&N}Kztodt
zu7bYJ?dPi__|-mHOfKbliHph&nbR(Yyb*$V$(0<i@p^;q&4umk>=clpQ16Rbok}O8
z8;X}L88(_`8&EJ~$uq4g6;mc<dXO<I53K@s_E_=e2B}^`oz6QA+s@WjD^!6&4Q090
z>G9DuD}&(B>%VH;9$Trc>es*?a&UAcF!H^5;w(Q#%^03XZMy&0710^DTze2H@ad|4
z_BnS^xi{y0W&My2tZHB*ofTcGllCv+<k)56J50WvH2=`Lq6Suu0s<mhjN=$K+|Kbw
zp4H}8n61C?X>;P&%%3fN1W6AYz&$9G(zVH0btKHiXBD?yOK=<Vs+3AiLHU2bRP^lw
zDcp+8dSK?x$J0~q%j4jnJW!giznHo`6H!BaSFFIF)Z5w?BQ<-)W0(Ij+XJKFdZ2~C
zV?OB9f;qIAgCwHPPtGEnqxLOEN_nJb3{Hh*mQv_KC<pc6m3&JwrKC59)I01I57`nl
zXAvsW^W67i_=hp<PO$(h9p0kWWy<Q!C4QxYk*XL+gMO^Yc%WQSjys6&GGH`^5-WLT
zt*`H99vO=6sdEn<kt*HsG}KQ2E-h)=Slst^J1!CV@ZbvwP;~{Fy7H$1ToiuVi0n}2
zP0xu8=HcO-5yh3Bx7j0*hpjVSh%=ODp(^f5F+70nGf=~C@?b1jL8zzjY>^DjRl&Cf
zUt=f#shUNFv~X4t8|;_ipt5?LlphiDEBq{F7x|Ahk8k$B`g=a!e}qBp1Gl}~!6yMH
zd~fNJ%f-r$D?-}U+L8j+;F~vZJyVL4;`$9c6B{ckRve(Et^=c!c25F|`~9uUjkY}L
zuUW772<$N4(dA~(W<a3x^AiWEdlA0}5n4OAa3V1oQ|r&xz7D#EhYf@2rA+_6FJ-DV
zZ&x^Yb>Hq6!NB5zLP9KXG4f^DvH|o86(n%$YP0otW=LW)`N-G3lDK3(GBhh{7Bxhm
z@QJ8q8iuZ;qw|!yK04@0R$Jd9*LHU;f%ZlZQ?@rjL~kCp2(G(B;hG5E9zn${E-~BV
zvE<tg(c_y{b?vlIN`~>DUa8d_g$gT{_As(wVz#|&aZQ*S;`yP_d-8C1-0#+TmZ^QB
zc08L?cSKD^mAqVPAf43)tsIQ*a~I3*KRU8|MW&HDIcuD%#84bJV2lCTOXl9fQkqMN
z<PIWp*dIdhkRgS=e0Pvb*CT3rOqk8D!g+-9?gzUnCKs&avwI6w4l<K-)E!E^yZK?H
z&gLjQyNl9GKFHv>YxD7}B$MMHF(L?|l1*CC%xIXDsic_dGZ~-7>~S^3PXEh}vi;lJ
zR-2i4TD(<;5q0i!IPrYk`r4Lcv*#0uv&)$!nFP4zP|wU+;@fTX$yd#7`^%ATkdAnN
zM~^eNp8D#}TUKCg_RD>b#<-c}hAskd#hBR~p7!jkAfa&MI-1R2%K}ZKNx1z<XxlqG
zC5<l%q{^*A72E$^kYAvI7)~fqNR$%)4fzMq?)RN9i_suj+m%3+9M3W`!fSX`q@Xoa
zZiGAgJk?)#$s19Axyf+bfB!Yw%DuYc1hHVS_TPX;c-suHp|UUE)>5XGVE+hjxqqi|
zorWE5-=aUw+YCA}fIu3w3EQbt|J(O7Qoe8#z4AYuLcO_Y$NuH+ox)sl@P$Ql_WepG
zpj-lCfC&(6hMnHJ+)|r{4jLI*q*jEhJ$@;q$Kp_^mzOX=g~CAiYK|s($|+2^6qv3E
z*_pQ#qsdq@89g}fh2-$v%t=Zjl_rGb>zYgxF(ViA<%IH)LuwPxSluLqupo@n%Aj9f
z;E$H)05azZHk8Nh{32%8c}=^hn&f3Muv*m_pY5}wBuS%KDrRjHiuUHitkWVMvl#bt
zg~LRd=|({2^tZaU*)o~$TX<QHW#L3N2bJ}{_)_HSo|zEDAtSvMf&T3RB%XG{dVGB?
z2ngB3Unk-zMSovfXyjyn?$2lAaaP#)n5k?6NZ854&~}eA#wZ>xssY#t@uyZppP5p@
zH*)2eW@a?Zf(va1dhgRI>|CrSlE<fK@d{=G!&(S2vL<N?9ejNaC)9clG1U}Mi@Ax3
zh)Ttv7V4#QD3F9$FS7R4Xi4d#<XC|XI6#(qjVEtw2!dqhon=E**gN4EQQB;@dKSvI
z@9G@NkL6@U-u*K1gSFB*r@)NK#Fu{TiW#uy$-n?9z2%_$-yhQ(9)MpqmZ#ah@EGlM
zoN^S8eQS9dR@K$XJs!w4CK{!}uaXEGLR7)z3%ZTU$5FcYyXlU2oAnKkRdqUaMGcPc
z_|W))ns}cuF@N$tVMM1~ONdg*rBb_GP)v+t?(@8b2Xjk!`T%#tj<tEhlt%vob!Yko
zY}g>IMN}{teQ;vvo7c1HU9y)sZ2#EM!ng6LBAPu72R@}g{(!O+?jEJuztquZ__Hp)
z%z*Ic2+yMNJIB5O@#b$M)0@_b`whEbcPTJ2b*$Dg{yVC7h=hS8teAeh@Zzkj@mfbH
z>q?!<ZE(!n95J^-4TlfSv~0}NE72b)E?Wu_w<1A*a$D)rDxz-npSJ^z2Y2MP1HKKH
z?hzKDG31BK@nk-zvQQ5*AZxSiyM78bYE5RA<H7hfD@jo1e8zYlK+>e8g|}#^4vvnJ
zT?yA6<C$yZ-K03OP!c`3(w>PuH~3KcQ|N0)CA-lTgo8VGzGp*zjW3aW@_|u7nrJ>!
zTTqIuBKx@ptKhb^tqqS59%mqii!E=fOB^Q$haN_%$WDG_Ohub4?yYNu&Vj#J1Gc?$
zPtEIZx*fhTYR92X?-WXiKd5(jD#zK0s`w>8*S_Bp=GfrCutP0egV4k;iRRQ*fL|=T
zF;rgN|M`eI@y7_OEHJT6bkJT`S{bFvRC=jW5cXoMZ-y8<*Jn=Ufdq=_z-u8L!Rtv8
z0Q8aSa4mUTm(L1(u^5z#r+>YL24HO}FrIfh>NxMjUCYh~aIZV|9NvPRNa<+y_N30q
zY;VT|>U0?BaHL6=x7OzI<SUoR>xod3;w{qJP~P6us8axN&-@G;z)Jl2aBdA~NBZcg
z7n{pS?@P^6d!I;1O9MHm{ilL*4ZtOE{0EUk;D^i3Hmy_hq|sd?V~r0g$E*+QHNL#$
zNv47{yuCp!sft+kwV#PLcjOv^O?9*T!rQz;$oa@(*dr1NXAH7J+cXeCO&hMU+u&&;
znQqS@XAdwI=x~13Tl;jN)&aUT)a#LaZv-DN!jf%3xM^kFl56j)>UcClVV!Q2-5;`Y
zpyFc1hkdt`!pPuIG?l@xtQu9XySy1_+F&qI@>mXxGaR50_P%OcwXJOUV6c4mOYV`8
zdBVp>Wk=lz#{JVhhB5rMVpCp)0A5$HM~p~hR~9+WcOPBN$N;pcO3WjuY5Q04ftRP>
z4=;$h?%xM7qe{Nafx6?XyiACr;f8V$5_UM3!4e@PJEqyo@5e1lDr}5P6c@n#Zx^*!
zwRTi>=rvGoE1%P$4AX9U$Gpt$RTESrD!n`|lU}lPfBi0hZ{!*M$H~CKvFMF%iIJZ3
z=jy_OyWX%gk{}`4g*Ovewc-sW5A8!xVP|m9nrqP?l={N|-tb<8m0a=8v3Hq$Ij<NF
zWW$E(Ez&cJP4cauRlRE=_~jZ)y<WS0=H4ZQX48E8>kV1I!I1QYoTa%P`F8l3X(+*o
zkj#~y=A8nhr+QM0#>FKza)VDfnkEt-qN(dBWXhFZ6b8sT#!x;7K$!@SUwh6h@}SJ|
z3RTMDsafc<r(T32s&GUoOlW8Tw?cUtUS-qLnmYn&k~q=p<Vz_c<1Jc`e_Ll*tytDv
z!B9zGS*p#-i-sjwrqqgl#9A}XA<X;c7M!B$J@pv*3UOY~-t6Dk3&-9Fp(8ff7Hs)y
zEM6*iE=C;K_Z_PTx&(z{<c(Lo9$bF+dWhn_XdpEX`~$o3TYO7RT+7yY^>jHlRv5}5
zz~x+ILxz&H5RGoSFEIxIj<0L^^(HF;Z}MSG?zKT}^#)-=HtzxtbM+TW5BXP08=t->
zPvNUTaxTPz(%NSN;V5Uf^<u&o%T7{~&;ovE7>{&uwEw2rKuu)7+u%3wbi<gb);CDq
zJNhEW(t+Kh51%S8r&~@4DW%*$XnwzLr>&!qv=0&t);^u68~hZq?pdVha1DDKDK}jP
z^hxtR?<8>B<$`Sj({4u_o!-l<cdN=fsKcgC9?pDPb&e7o^G=r2P5{=ep3s;9Dj)w5
za-HK?_$A5u-Ac=arY>Z0%YE_}wC}s$zHUs+x7o<kpATLiruPYrU<V2X1d}*BJW`rh
z)m2sDl3yt-05J|_mIXJyA7DAOPVwSqTV<5It~NZI0dg^aQLGYSfl`y+J!s)3gbe=(
zg{Dt^+0$0wJPlub$16?d4EtD!;;!ri<Wz|=$fIN#q!O!oZf-pGmjXkXg<mi97g>g<
z|FX6^8yi!;Ni4X^{<S7@BG$TUs9!b`>OqEV$R#T3>nElq&K%?XSHIqT%ODU`ju~kV
z)l9%_`h7JPD?lai%@ppP_*XKhUsRZ!2c!)#o~u^~&c9P2iw@QP8>BAM#}rU-Y2AK7
zrE{8Lj#Qn)qs0|KaCj1fbY5b-s_&ayxq!U3no`=GNep2o<flyzZR7LNrFmUdWdLZv
z@^sQayXj65o}~jsp|d`^-$N}b^;D<5+7>{E-zRajM7PVu(qrF!Slul6a@W49-k{Fk
z5Sqr0GfsDe<1@#nyQm^|Q{eq$s5E(6L!*dHk|A42r~`-9L!}{pzxygoc6NT(nLLI{
zl=bak!A6OyR?KZRf7xrzcFmQgr1GUG8=Mj_l>pqyD5UsAd$^YG`=Qkjv|ud2sPxHd
zC_sC>(vd`!*mT-4EfBKpaKes7DLB$SUF)3oQ=L(-N-MQ|_V5T~1}=L|AadJ4RyR~Z
zbMsHc;lX3I-V13Yl2}&|GY*yPv&Z;%Gb@mXdqnfqI{=-eJPXj%tBtxkNpt97Y?AZ?
z`+EtDZ`DA3q|`5~SOr&?4>hSpa(sM!shF6UP-pji`n*b>op+6iXtjyrtJvZZe1+M~
z?8thpwqf5C4Q*5sqcc5BxY0U(AILGkjFfUeZXpUw{Eot(ci(MMHLP()Na8pvvFZ3s
z=x1c<w@=n{5W*BiclusK=PgjMrZ=tgy$ULQ;m7Q8eMAPf7QXyIA>3)N&DD8`R$;*^
zD7<vrQ{{$iLD0+74LtZ!g~C!$uaZSOsxXTzroWO$vd{)9XiC(uUT{sq0}5mHuGsCY
zF+2@r{-F>c5|6B18p&zlRwMg(=tp>~mu%8|GBWNet7_4zP3T-{mdT^}YoI1@VHpiD
zwG4g#LQ4YJMV!lf1fL~ErR@JYY#;uz33QHca3WSlhfY7%ajgkAn686*mJzOj{Z^!p
zTV#2fIqI!8<;mV;Z`XUqRhj-~B1Z1`P{L|Eu@|7w^3m4u7-<=4Htk-{TKP$=I9X}T
zigB}gT56@b=^<<tEsabYfqq@&sP=U0ZAtASAX9~w_}Ujotaxe_;&MCvth>6YsG2lD
zT#1q&W9OZ!9`(6<yWvH!7iA?qmPyPNF{7$$9gmJ4lBEKsh3J@=<D0niI}fy?JvCvg
zxf*^%$7(7c%_(AEcLD6=W}3%NM*K!6y63H=31-`?%EU(OP`~xbpe%)uc^~W>2JRHH
z*2)bXuviIlP+<JS?2lJuJa`9(wrQ88j!Vt-hM@wGQgt$CPG=DYoyb49<d8m92v;oW
z*vs5El;;_RSm%@aEI2q*MdXSv`wR+O$4)b8h#JWE@TM8C{bTSC-<66{6+EWy4vKO8
z-248?J_gjtbG~sua9g&`+s!TBQ1of(%Iqijiospvaw0D5u|(v(#N7}o?p3a2-f$f3
zbw&-V@Ng^xfph9W*5Ug8%8KEWIJdaNk<h0xr0NJbH*}2&qcN^A)d-Tfb<PUD0Ypo_
z4DOPxgX{_(SqO>TIkl-H2i?l@fUI9{-*+m{f3Y%3MLIa=AW`FDUsS$G=y-#D_Wp+a
zSrX^I(tATf=j@%-(-H0eEGjcVRlKxE$<Hoqrk8I1DrK7qU$+|-=jBGo_A0Edy!6B=
zHJF@=W!Bevw1zxUp6nJ{vnB?Q9&SKIqW~HJzkd)hn@1`hCQy7!k~4a-L6)`TFycUh
zHL0w>DkDlBHKeS+4DXX<9l=g@s8~=j;&$~IOx#oT+y~n&^<L$m(SIl};%#M*c$oVF
zC@GG=>JvfIMCUZw^_FRKOca&|NiHn8s^W#8Fc=!k@@_@HY$m7QZ>+8~k5tn1%-?MU
zonm_HlqUvV)44@XWOASzkmz~pg11<{^~xuxbgNhW6ucemi*JX|PhMv%DaS~CJDt5=
z;RR=8Wc2*|_pbx0!Ae89SK#YJtlU036c=0BR*3J`R<t}C=f8G+c5h?c*2dP`C-zZP
zQ^<KzpAAPQRv`6&7L#)Fc}lO<hGZfp#l5D4Pah@<8>N4_64nIJW4N)Q`nzj=B<5a)
zw=^_^*6_+!=ojK6!$6%QtID1J5828Nj@EIJS8+oQSv;9(F3dlNpt>x0eF32&p=d5f
z^P<Nx?J52r1qCfw&TX*&*cl|eFw-vYcN4J>?q`)#yJg}E{c%272AClPg@xH_Iimvl
zKKX26h4pc1{797?G%ajuJ$0D~Fd&v>RtY1%xcSn$ogNg-95Z4+x|SO#G0%5%^AD=a
z@vFNCHXN<lqi?S4p4|uk8>6E5xq+h8OZQnr1XJqCofp|+Y|^}X!*4sxhFgVKaw%i`
z(;G?Z6QEe$ZV`K#&E=&fqHOWM3Eeq4L(r}r?HhVkZ&`=udh2>MDu+&{G*hCCE~S|w
znm@*wsU!qI!}aU)KvyOXwB78#`+zEXG{VR^qf@Mubv`EWU6AU$@iz4GXRG!&yV>ik
zkCUJSo0i$Q$^fqoy?N<KMY%x?f5m4OwtVtCZ1p2molb@SsWwJ%U&|-Tik);yIf-!&
z7x{i8eETvXZ2}@270NBI$ynKbUAd}G;Tc^b<;Np97Wh%}n;rQ^0eoYt*}_BmQ`C$s
zt0*~T(zq7=MaY1bG^!-(W?g-Av(9D7j#!F&<G73**ks*g$YJ}3eID~|R{nG6MGObi
z#TZQuqSUDpzQy;>{9H&6?nv^?At`aO1-c`D5XTHM+qbFhGtq~p)wu6$k>IcN%UD<z
z(upYS4DAkDEE-gM+3UE@-`^;mMM0xkmt6AJZPK=~*c7t}3>+&AZp^}e-c*Dw_iCX1
zCzM-Xo9Mms5&l*)QInlLBDEQ4U8FcMfDYeHC3s0S{6-W*xN!0k_L=GUHhMECrWj9K
z;?{G1m4I#3%x;1S-fPaGg;=F>Crm`kHg8KS)C4g)Yrex>o*j3D*}NaN+RmSl>p5I`
zb+btDxmWTW`t5avHrnkN$%WXiKyK}Wj%Y@XItu<FgFse!E`2@_YmQ>US(~zam;yR+
z%oUyM1@e+^a?&Wp*WrF>|Ao(*5ObjKF~M0D%KTR@<dr(j2f#zqlgVtnI*E1-=Sn>W
zROrk2O-XjtFaPl6y)f7J_HB)j606BfC$B~C7xZ5KTdHDtEt;+5gH?()soQyXVz*H6
zAJMh?EuW6D>2u!iL>T$n#ox?q$2;eRdca>j0NEYtQ5bm6ez%|SuD(dU3?OSCJ$y{G
z&1x}^wZxMWt1M{=lBXGHPv6C>>uZYtEp|8!`{$jkeg(B(sb+Cv`Pl6W_2stv{r<!%
zod{LuS1A1-v@9()-(!FsSA0baknuSSP$&*w6yVxY##N^<ko%oIeN01c%lt7{apEG<
zk<UU)=!diJ1`}ak8Mxu+$nt~_rq5QmY20C2tY??y;4<{&irf0P`9);anT--Aj8Xj2
ze8K<s$afmZyzlEgs#SFObA2P%19C4=(g9mPH~dDGEYf0I9OL;3&l_ib+9MqMG$lqE
zObS3IbYt)W8Xl;JdyzQzB}xYKZ>|w&S?RP?X7D2{Ewq~P6P(n)Z`)WKrj!i=>cjFT
zSC6V%3I`q-<t<>vpCo^Z?Fi}Cgj&cnHY)87w~E02&Yv#F$MFXR9f;>L4m<V!bVJ^U
z{!Qrjc?^Znistirv4T&NbsIcGIz7w`21#|SB*D5}Jcg}xrLg9qT{8buO)B^&CD^>v
zI%(8zAW?ar4l*G&Q;eFDZ(+ZRy{TDMbE}GA!pYx@=U~>G?Rbye(!aLrX2+E*ans7d
z=mI`kGL+M4eo<C0MS5KRJ=$fqw;wF)Es$H>!uT52?Bsfu`#2uSA+M=GOTA<iukk!_
zC3GGHZhnA&a&*>Uvr1r6=>V#(#b^f=qyMe|oE7pOA&8Zf<-Ub!WFSqtP2<He<A&>P
zwO*}V9<NlK@hY17hW({<;?JZvu(#6sd)0Wa1zHim*UFgg(oOzFpn7U^-~S7t?mOEa
zA9itr3Q_6pjak68h|w_Gvgo8dd<d<}pKEdSE-dJ>xP`iE`sjKwSDdF~9>2lR6gs<U
zr!ZD}9%{jcr9R5%p}*&U=xZby_g;J|o*PZ#5mMvT(J4H@lqWh@rOJ@F(BzQ(zTo!{
zQ1*-SCZ3{+M&pT)FAE$js+q>Jos-W(kroP7kjaj(Y2pckn&Ok;i>WOBkxbH56zDWR
zIlGW`h3q938B7M}Q3qPP=_69ZIgU>n&_d(lW}S0VHm2~Z`9T1au6(RB2iiU@Yo00$
zO{b$_Maq3*K80NVa~)=R9h7D;f*k3lnUBlO<@Vi2^G1?fVRBoS;kFbU&LQ>NOIJwh
zM$XslaINR8mFJ~I!3b3QNFvTZ!Ey2LU$$)8FGdi<up#&H{y+Fteg5}fYTQMr`hk`L
zkV~=o?E;8ETZArJv6jO$F4s=_a5d!Nam;$PcX^l*9~fDj60<(~57B`+{?Qua_v84`
znke@IDsP?-nyI#7%1<w0YiSfNlp^&BYwf||E96TwwhqH2Hik8_T+b}52oZR)qkaBa
zYxN=iRm|rS1RoE_1NNOhysKHBP@PQ^IeCZONocTpumB*{Tv_@OVe9(^<n*evSk<Sc
z*tmid8DIl#ijoC!kQ2dAg6&^e2eabn@Wa)5{7%Km`j35hqi^VqVdTAjUah2$0guDb
z3?p`7rZQ^_zhpUoC91l1rGMWx5n#2^i<4I9Xd+574UZq_{?ztB1S>WnLo)5Xecfnu
zD{ow@b`N)cLR9^ga31AB^v+6o_oYWhgzC!Orw0#Xa$+TZv&gr*-{>kE-0uJ%w9q~H
zxK4%Q)hv9oMrDi#n0{b=VBw8bV=~)9W0K4OTQDz9fMRBDRv$;zkSn14KYzpikjt<)
zrC&E^vBUjts2~1vrH|?jd$R_dw3(^>VZ?gzwXIn{VDwWSTVXEVac*K`$UM+A8M?Bm
z&#vNT2n-yY-sJ^k488FS@E1<~@$+1sPM!@RrXx2sACa2U@<=5iSnV*LQm|@|a39fi
z^UCa23!qJt>`k{xx+t9_wDt1~s+9s3@OoVy)PsSvo$8q5%<zB-g(L_0;^8Trprcm6
zx4S-=z=QyXP(HJhAcU!Hd%fEX9$cM>A6JZS1bt5es2k$#+9|P{pjfif8$a*gLXtOl
zTC*w(oWVKG44iaQssNTtAJuO!^l{9{?U#v)7|Qky=N=a0J8tqzlg(@AT^~HzRO^-G
zQ{Qe3)#BP>6sq;rU1DdxEKBH1^PnngYud;dEbEuNU5sN%Ab0r{QndRJuW9>5%VQMi
z_Dm}DK7rQV-_>Z9l<IY37YeVUMLX^NYkj|c`(A($J=q~pllMbHwq%!V$@d6p8|S-T
zP4Nx3`e#Vp-n#%p4_N~%g}k)uStkwzcfvr!2(dGn54gTNNb#r?C?3;jPt0#$^h*(Q
zF?}wg;8`z2CnkDd=e~wjq|zCDYTIGh+&Q05K@eTJq3MA2C>BA4*AG+F-8%*^9j%=B
zQqway=;T^kh)-n|i~$tB4`{jk&s*i^a4CVDuP$~FT?}KCos4WUCc=N*8h+GAqIaG7
zB0IXSuGZarx802G?Ky#Ru3|hOAmGUw&}*#L)ML&eYqk(fOV63#S8u=QLW6lbr}})k
zHrE!yk%B<poQ*Ylt5fG4@Q5|IT3tz5?x5sYzk9_=&sBgucmsGsq0?sLqJe<N?g{H_
zo15nyRW@%+AbDSCWR5^;y0_We`(_H?<W_|P@*{fJk0g|u^m4Oi@|90JXkLD#ih-}K
zz4JY}UyXdcS$%kgSiOHR^j;8&ScN>4yI(FmifY3xkvif?qu#EMeguyv)DOR}^TluT
zpCozed$p=C9EGpk@g-(l%o|eZ5Fg`@tzCCuI-J>m%rO^k(EGV8>h<Hfy^)E<iIxCy
zjdFNyA`RF5=z;E#|8Vost_`VC*uY-SorZPF(WqJd(<7Olo}NpAhmE+px(=nFhcPY}
zrYaX#eyJKdAts|utXi`uUjOs=Z`-!tCBE>*1SWm1&*_Fda^?8m5P5X`V23R1I!{@@
z>$>T<KIytP>c8u{_UmC~MBKnvBQDa%;YJi3>yOt)cD<zIM#-6#B1N5wmLCqAftyjs
z;2=tk^&<>731!<M%(0fq-xuZV+7fo_#s0IEmi@3VX~ox@`{?2P-S@}zKMpM;$WJDg
z?WD8ts}c(eN?tG_Ffjs{o)<p$Yo01iowZk6ujnB1UfJTPQ%IFED*4)Clm-C3I>aDT
zcF*d4o1ehL)+#a`f0Wf1-i2K4Y0i3-NKEF#FD^g4c6+Qp?kny>YK2+O#9LlBuhWQd
zPW`Rw;@L0nE_`<*z;(@jw>@l`erhowop9l8@v(_5U~$#pxV(}KI#2y8B1E-l<_`yK
z9)eZKxBY_$&X<6n8_&`C+QFfo;d5iDS%rwkj0r#0!F)ZgVKk+2YI|;KuU83PD2<Uf
zLI!!?^>_t;atylggzqog%Mjwz-AB@1HrcKFA?9iS2NKB)yLYvek%9A<vMnzPX2+t3
z2@$a-Za!9as?>ppd7W4MED7Or#7z5+GLXB4hDuhAaU7Q<DoGAP8v!^*bn>;gv=>>!
z9KSdkP5E@S5R`klmyXJ(+q2<pA?fq#pSX|2{df^|EVF2-3Go+J_qR;qH}c3czlZTB
z3ow`84QGNs-NR|Z@bxu32JW?bXDH%zd9Z@{hn9<jRsp-9JZ5U<2;DRLkD)A#2X!jQ
zFn9{^un3u!$Vw^q5ezmCyRP0odb@_=MerY**m9edT<$YQG4<OLZqN@U$NG$`ntibv
z3dD*keYxTaYA3JP6JeCv;oE~mU%dW+Z+{%JXJBhF#C(<jn6#18ezz*{^K;%TQBTe+
zhGU=Fi$g&{!BC*3({p#+MGaIT>>8aOXGN%!B9<j{oS3CKi~{ytqQdn(2LZp=nm2PN
z{&@Oa^>SY2(ybdhcf{0fs$wSC`zF+QWXN&(H@gS~cV@F7s`hbn7kw#)(1|Ty8#vPp
zD{-`{8U+{Vk|e<%vkQy8goSS=T@vO3u*-<c@^^xjusD1?g5L^%po&qn>jy+pWWL{d
z%(}Sf%~wd&)}(Y|wwm<FIwRnHWKS6u6cJ%BX9WbeLR@^e)n}W0my>>gmLTGBTCmAv
zHm+t;k6WMw!=^bZICo6q8)A^S(zUid`Dpc>oAg5-znQEVl16BpN>cYubn63=qgx!E
zY_76>m#OAzOl4BOt4xF1%VFBb;G8%bkrC4jTZDN?qL6z+I$9x#DKMxQz~3+EkdP=B
znX1HysI>QCqZ0R}s5a~f9~XZFTm%74kHxExK&4Xs@IkEome<fw1<olT9{xDClij;}
z!0mfHp55!m)fU$ESpOF+1vDV(e|U)cL{h}`Bd!r}BFagu$uv_RM61=2aunbed$Je;
zBQpv_2Fdq8L%o0#G=tmVu$QK(tJ|rPXPNe6d>i%FF&_eq4lZgEuIDy<W?u(<M-k&<
zcQz-&K0UtkjM+Otcifa`9@3!{cFEGjVJaEKe{Nt8VhjjKRGXf3q}qVL4dYf%v!44I
z*-GBg*hoz6bFLOCoB!nMnw$Wf^;u#*<bK=7$jtwliGb<{Fg4)&9bG)KJ^bFKmYrUr
zjg6RPez@MGe<*$&5BRgV5(dm1Xp?grFuo2f-T_FXmxb8PM#A;^w8}9h|G`Vno_{qT
z_Nn0v1(D`}+=f>-l<>%!WqyNdNQED$f_9AVk#l#DUbWj%^+vVQXa&@&m!21gaECaO
zx2xC(*fGbnZ-uM1Uzkv}l>{pp$ktzJQVZ;Oct{a}Tvgz5h96yMakz`ox)(zAh*}ll
zW-dS%&;9a17vMi6imdyKyv@72;*U4u@U1PzCtL5+Oln3ab-5Xlem<iter?-wkCCq{
zr5{{B1%cbuE=mk3oi39alDwbRX{J32)baX6K?Y6W`AeUI@b~WvE551=YX(fqz>sT8
z1hAsaM84lEhe1cJb5{ozyjC!U^;Ejj5CY<9M!85^SW8RE?e>C8HB*8HE&FQ4c9&>M
zJ|m}Y<^yED;L3F_QTp}djC%v-YgaAEPW{mRBRgkab@%?oVVI(k&xIBni{h1R{@=<K
z@!QFmb#w7MUF4koy|(wU*J{L-PFMJ3%V@!ZE8*kZ9sc7Dr`^!cl)5y=tcP?zI!0=5
zaU5+B*X=0?v02?23goTiUUwiD2|XWDL98q+urTF!dEam!R=07tD?e)gtzbZUj@__A
ziy2m{-Z}MMDebg(NSk<Si!W)%`s9LnQG(5nXKp(N?+mq@?4hClE}cAh@2N~3epbOG
zq@HGS(`8Ew>K+t*L)i6Z7IXPesGKB%r}~)wt93XC{i0|4B{=?Ltfk%S8DN&{$x!>`
zig~=IeuQ67i{DQR`g#)i-E&5p)cRblY+4aL9{wxfd;V}i&2|6xQH@<3DFLwqubX*M
ztlpJB`XaBCALNk-<G$qCzPFY@TL{32NG6r69>8$}j4;gAZG}+#z5(2MfCFQ}v2aE?
zJ8<DsJV>eud{@8jLG4oLamC`~1}<U0)h=|e9FL-vzLZ}9=`P?VbyXa0d`}F~-fUn|
zFK2Dd21{N{CV?60_}~<iIGD~X3u#kdvZ=yFJPnAow|bOXtG1~W`1*hurl+x-1sjt2
z`S~YHZ+|)I^fh%yGjZ4b%15)m&FGI$<rm`d!I30?a=AZFHeb!gZ@aqZ`u-SHe^K3a
z=rP@+5>5MXYlL)L_q#m4KXgCa0z=rV;o=YW(d+dQnLj9fsQeBux<cW<=X=3Vdlh-z
zFV2C%5D?{qgs;t>Qd_;$cltiiR&`b1WL3Wg%mHSKBqe2nBiY%w$~m3vgIj9(PF-7R
z9X165tx94Xs9;h5&)x`uK$iX%rd@PZQ}Q(=SB}-#FO*FQ$Ygd;0+VbyUB7h+Be7*~
zyfR;xeKzF^_1*G1?jWW1LpmS_vyt*ogM3HEB!2hz_16(9cIzrJ#fXq`QU_T3y;W+C
zk95!2G4hV<zWFgAbn@QfoGFG8u}vKVr@l8P(<yJeZ3Z6K_QU3*P*n=)6Uu~yggAJ5
zk^}b2N_uW(RvQ1b7X?XBKLN~)y$<rUlT)P7mc^3^jl0O&uuDa~^*Pqe?l(iDEJpD<
z+|N*pf?(CUPV=ECTdJ~1D-PUczC6J`Gv=46>nC;I2NadAEYywTlTu)z*+iT(C)%Tp
z&F=2Ws+g;}hV^-w#=60gr?{*)A2;Ss?E&;gKvBL9*w;mFcTLC^Ky_O;(vW+z#rWm>
zp@)_%ZX&JsH@%ZC$n}!*x}`2MmAsqE)tD=>$KURVb(saBiY4^-SB?a8IrXmuD@~~-
zcl@&(F%xOflo^#VeEbs2>G}Q~+U7>2z8`i2pXy}?pRXRbL9gUnF|&Ejr+7y@{zUgr
z^!!%h3Y|B=@*7whN*uy7O^W(k#5<rhNUly;RIQ)t{QAVgxeXsgc!5NU03ICqz^Ohy
z(!{NlNA9!~@rXlQKfL`(SEXhUA{{;d8c0)*QP<Pk+la3F0HfY&<>OZm#R<nfcdMTG
z=<@;&mpcM<P1OAdv*uj4=h;cO9m_t1rqg2cOYgE1#Yh&Ur|KVIFWp3~eRk3Z8eOj_
z=g#;W!Wy-`GL}5sywvND#F(r&hJgvtZdV;wX;0l)O~3sblhyO4!5_|Q)#ekMWjiDC
zQ>x&Wt}bf8*y@rtnS<u?X=X%0%<GaJaI;fTQZ}-?NqepRNNfP+$l;Zhe;&{L-gjQ<
zJY-(a9WZu^iiLMwtUjhAfKle_t_Oi)1odON#<2&+_edR`TB;~}gUnf%$W?Fh`Ny?K
zJf{BsN?a}ZAKdPPU9%iUM^1TPpc8*qbuzQ4|Ahxp{}<hoUH-d7c-2I|wPJ1pQ#dfg
zEmh`9geaGZs&)<QUHNY{^woTb%!~kfxX45P8P|X{HOSXb#zlIGGJF#xYkDTVIk%Le
zRJyxeDK@a7bcRw)4rWZa?#A`SdpyX#uSd^~<rwI-MsDQ&A5Uk&6$RUM;YUSjX=xCo
z1`wpXhwc!h6&O;wV?a6tX^;>Yx{+=Wr9nERrAxZ|yJx*?eLuj(ICIz8XYcFiYe>qT
zOUGQVse&U5oOKbuE|#K=_YZnTm(wnnar{X9&gR<j9{N1+lXr#kRZNJM^zUeW$*6m-
zjZxpeqBZd*+htsey1IC<p1<uAYjp3_?}WQd<x$wcr?_}JTI!vGHs6|$b6?=6(5@H-
zwbey;1V=)4oE)v;g^-AP=_1{4Q0|;j+y9(_QR_Jlj;8eZTRYxm3et}ObZ2a;d>0m%
zjS;0ELvU|Xu(oC9AKts4D`^`=c@c%$ui?8Afo?~O@7?vH^)IP1rM-pGt@Gw`g%Fc8
zp8tH3;x|Av3;7ZQ;VE#?lD1z@R;;K#oT{!;-v;{Oirp^btomK*w-bqx`LQoXQui#b
zf4<!7Bd#<u>^VMW6`gx1L#wl}!C$plirU-r^HQx;@Nb0S8jIx3A$*k8aqwP<4#+1a
z_PGiAY*g>#d*&JsoA<{ZOd6E4^&3S0BW1q<e;@Sr#(iDY_XN)MFM69l2D3_8>CV)m
za4z+3T~r_SU*~wnJhi`B3#)OAG~HUWPh|k#!FbiJp3V&X+tDA%&V_qo-$w?+b^y3g
zItnsK$PxqOk(02!K#J!`*6&ode!zzP!OG|27tZSMG`^#Xd)L2*K1YHwe#=WtijDO|
zMm<w~6=T0YSPL=6vwZr;Z7nIozDv_N;f0O}G7K2Pw(D2V;>;QG`X}cYmzq}!$)4R`
zs2J!iWS4sUiKnK2SL;V=ke>VszlEjdx&p6Z$K}_7OZ5*H81(~B-g_@R^&+2Ta!$bh
zuu83v;x?Edrz0X7gQq_9Yvlg<B%+n!ns6CJb>!D{%e2>b|01uS_{>{6qjq0Pv8z5+
zq+dJp#@UYzY0<l__9Mx;@D*Cwc7IjJ0M0Rz)_d+RzNev%wFcUJ*Xl)<KPMQz5)8NW
zGJMtdX=l%T2NvsAP(83nH8eDY0eK8OKVNscxw^hhDC-}~$OQ2v(ZUu3HpP%KUvQ&=
zudR;ZYeIcoVUzYgFM_69cG&euv2#u4K_ktBArDH&a-n?IzxTJQejm^9x+m91IzQyS
z_14dufni5cGor4dixV{K^!o~)s{Es--VZSBI`gwBl7bEg9r0-wT7FfpsimN3q-28h
z!)|8!QGsg$8RK2L8ggIxw?<d@Q2k&y*#4kndbhRuuJx9i&F|c}>-d-x)anY)K9seS
zu1$+Disodci0PB3MP~Z$q<C}^I*Cz)8LL=$rYdRF%`!VUJCDyAm0y4~-<6!klsVa_
zS+ngGO=4Ji!u0a-9JjJWsh;)pWxqAZS9W>$cz-{45qzEL+|t&@79N1nZv;G|gx_)P
z$GOX*Wu>7-rmc3=rE3CWhG~!B`<%6{Ni$98*cudYC{uHvQ0?`+q=RI96^_iPw7RZ!
za+|uu*U}CxvU=9Ppd<pRgJjLlp`oFbv)yT=N`$=g4eqvZ6leM7V)8y{`r`$d2kM+?
z8ukeeTi?1)%f}n;r}a-Vh-X&<4>S<aK;$3XWC^^xmvD&bYyVx+Bgv(PWd(vBsF<#w
z_1)|Q%#B!ndP_Tu0((n@G#yY>0Jp%ZF}+D8Wjsg}6p6qjnl0?fX4SbMz<_j7`b^<H
z1qs8+7b3%L4M&bt+i+vLm&TV53{;&jb@Dc39pB`SZ@1=6BVMc`*rI2Db-ah_J?uVd
zGng8SW_=tio^%DWp-Pd!V=<g577g;E6+R??Zc_I6`YzYC%Q4j_X)jU~C1J}7;rp@A
z?-9pj(FbkM7+HEYnJ$4c(e-;YE{nC|(j5*@t89M9s%M#3ceiwN=c@5KO;<+7wA=AR
zma}N)(m>ERC5ClpbUbx4Iqy{3Y&r%zkoYz0K1v_ngz+fcj`p-bWZ^S^NGIpD??omS
z`YaDUDV{^Pej5|q>#a(c;}xN`P~DxUqv(oRIbKhaQTUceYsWkifuQIb!Aai1_V@R@
z!M)pECNJw(zw3-er&pGWDaYT>X^oqhB>ymdYn?4Mj2RnD6xW|*BT}}QEWTYWbZjzL
zwdG6Q7?9AMc%-`b8+Xyer=BnLk@YkD?J6}-=DJ@^mImw39X_75d%WUixtG^9V~T|D
zIXij|8#=I$KYEnu>CoRutR+QduSkWcu<FOMZaM~>9{R1u*470<Rb*`g>PPS#OBfVl
z#?aO2l2}_DnFU6G#J|*H*W6?aAd2y5n{1x#Vj%O#=Jwl>+>3!kjwf*#9{ncP*(W!R
zJAsz~ViIx4+}5%B1ae9Ba)Rf}#T@R{n_SlN`3C<nfhryo!N;qTc<wE3(-y*QB3eu<
z&%<$Nie5vhB|Wt?X#3%rW4-h9kyHVka;+TCFP^3ZPs`?vmWh4)mR@#KEStnuo#a#A
z&8i~9#w9baD{^&Kg*LIK_ttUXCwiXy)9J1!&u~5MQ0J1*7HWMqK64e8+*?*zpxZ$;
zYFtj+@63+#8}GDBAG@=9HyR}6=`M`B+a$b0l0QHp#za@1B`1JcUTvGk95dMExV9pE
z*6>MFXDIkGk3Fw%tB|7cdkaHjyzr>>$JDuZ8l8st$!1pG-pwe<c^+oqY3Lfu5`=i>
zt4TkW))RXrDK+#W@4uY~J;wQkzL*|dWtPub2i;XhIN8O*-wMPZ5?vkQIM&?-zjZq`
zvfjr0^~&PiZEnVe-|g;H7_Y9G7|+9jak8=g)C(L>phV2mqVI(UL*L)W+TBq__rSK3
zl2V@)+G}_(`TS$boXK?(#d7_oTR<2>m_^ul$w1HCRV;{d68w*j#R=ZY_;l?|6z@B(
za?(M^y`|EYou^on+N4w=I4wU~LJ1$0Kpu1G52V-nT##uUzkkm_0$0MFU&Ti}dy{QT
zM0iEaJofVU<_SsEEF-s7%CXSWmCEgg{k{2Z2lZ`-$Z59kjlxegTH~$hb6s-P5vT0+
zc*XpM@2f|9kcFDaCiaWZTjsoJME)5{432JYQ+n39aqIZ3j7^#Q`EBm8&Hh6rsW!=p
zT)CnP7xvv65`s?g{@IT=Ir)hgiP$!=Ro$Ieznw+h4sPvgli`U(NEv@JW=jTXoG}vv
z+aG83C<9N!ha2-xovH-W(J5YH6ePz&Xt5&$%p#%+vC0aJ0T=`U%5`Lmu=_JH`wF0q
zphc9{%_Hh{S=^_C#dtor6Gz{UbiziQfvFxOxIb6rzCZE7)0+-sa-{w8&45SjQg_(8
zC^jPIS@!VGiM<Z*Q1!n%8cQA)Mbmx*YEp`CmgUrEiv95%d0|X*v`o<&b;2HAylJ%~
zm_3zKDhH+^b~l$b5ZB`DrOT+LFpz}EEB}S-gjT}I6iIlA6>xQ8IHKwQ8=RCTN2K+C
zIdg&x`UDhaN2`JD6;v{M26*VB)~?=`-xER}f5cOd!Ne)*LlLP?_ce@AR`f3?yzyA%
zS?)x|fOX!*_e*D-^{p$yVw(aG9cryB)C2npCMGvF0+Qg<I$iCpyxJXG(#SU8$g$a7
zWzb#bIZr&1u&a2yud%ae-XsSS`dtGaOViTd!d8K09|Y6l^JOpY#Dc6b!nF&BX=Ax9
zg*$Z=B=x_pR~<&mC05uE9j8CZQ}=$T$^Nmvbmc<@03<IHBCQ?LUVQE6J~3>bhbE;!
zSZNA@WBgb-A<PmY1`$<`1QZwAc}Wt`M*nx9zCwILT&NYixW;V#_!2xPDbjOu5yvls
zYWswIv<t;6v1>D7B?|^|1+<gZE3IbocNdj9OMpXi0MAYB^%`%Snj8*QD_HS3+j(o1
zRdMG>KsUg^vO(~kKsv{NyxD*%xAy&r%9H&9E{;b%%Q-g<x19kyuP_e#_V!cmr=e?`
zG?(f9uSNw&cVg{VK6|`=%CcOb|LV_gdRf>~KV@v(=(0SN=yY$UoAqmNW$t7Jq(jt+
z3j6G2d$tk_@%$l<COK<HyN_KxL|bj*#r~vv3MfquMfG{ctBsQfj?{LrzbIVJFI+MZ
zhLIRJ>)y`%vqz_A^v_Ud0}sbf&M&1;Cq1Lrtpf_NHr0*AZo(A%XVwM`@R?r?M=e7w
zyw`mXJ;~6gDma5Azs}JF7hR-?pM&b<E07XPjPHt1qtryD$Wk=sp&M04hhFn9<&EYY
zF8kA1phv;wA)FSJwtL_Wu*%w46jumsfi+wE6rG9ex=%Z&F=t1KMgPDE)$SIjsr?>L
z9hcUjbiYTSe<5&m=TR3XlpNwquO3+MuV{u(Tee)SDv7S;tCE86;@c#&O7P?9($a9i
zG?Y`)SPw(X_i=opC5NjrTh{UOTybtkrdvF91B%+%z2op+-^-;m4A<#3T4VWtEL<Og
zIeYoe6OT7ZYb1@7uw%v?j7r$-0#=)cnj^V~Tx{DRE@Nr`b@zTJ%l5wicz&NBRW5Nc
zTTL3sot26;A1Onk%D{D(47Fj+e&|n3(VM%~BQ`ep>Y{wW#V`M8wVG5gXXVf-m?RA4
zGP2nv-RJiBTfYu@u1ltv-=f~*<hS`h%?*d5>k*>4KP%%A;SKs@o82`)BsA-bzvZt}
z+Q&aQ%}Eq$`k<Ar_t>2;zPn<I(LMO8Sw?C!T(e^3Fs~^zrdeQSNq50;Uw-<8jDJt>
zv?I<t`^@Qe4|7ShZ%Nx@KcGB-tRUdkDKa<DCxZ77GAqc>oB=1`pz$SC5O#c8yrhMd
znv3DUIl`z&;^>}{hiYjBMO&_&mO8_QL-qvqna$^OH{2%6r<s~{q&SAeWa~8&GUN)|
ziN;4M9DhLx5p;s=0}GN(sb0w;A(eHJ?_{tY%{$L=@`e~E5y&jUa#_i8ZiVSj{#T(B
zMP8(%N!fqJd$UJC6sxRs#G129AI1^G?5DM^;LwCpJLOx~>b5knt4L_li0`t`rB14M
zJ;{?FUh;d;FvjWqen18ccgpmM#^PdmsGgXpn3mCHde7t^ov7MK^ovVNOyR^@RAG9Y
z{-foH&(TqYHEJpB(L$2WagWTr=`u+x_h|-J%9@(&#sXL#4YEu+mqR?+^#jiZt9Dtd
z&6PR*oo&QG00`U;#zrym^6&9R1=FRU^or1)T-_So1(bok)cKzQQwUbr&As7TVZaS3
z=i+yzXXgz8*$*=z`Bwc@gj28SU->7DD9UXwBIz|C#b(1rq2oM%c61x2*Uly3j`{n-
zTCSd#H!d2<%ld?qspAJv58s(!FdWxtl~Y`ygA}pucRQ9g=4h;(;(KJd^IGt}9l;|<
zkqy-g{O6UKshgAfv<zD~`0bSOerSNcN$TGw(N~a%G>JDHndstG{K6S(-fyjM+(S+c
z4=+GIGw+mlxq2rXKbzCS$A8n@GN4zZ>2aKpEFy6%v1EgKJxMZ88vG`@`4JTk8n{m0
z-V_?sYf&W{i_~qac6rUXqWLQzti;mafe>B)Y$@~Xv&CNxEAMt2xF%&=)i}5Em8EB<
zO|~F)HaG1v)*kbL98y?BFqT@W1SORQ!$5Jz>Ff!(dpo#6ZL`kgbs?4rE3Q)*aY;?Q
zR@p^xaxasJIoNR2_+^JO>lg<YC$2VI2gh+E(0>y>iIaSVd@&Dk9f`9s#xq3cK)S+e
zPNtqz{0kgVU&_s1G0{lc-}Mtw4BlFov*L8txm%ffoLEFzS*ouP#0g{~pJ|aks=&WP
z7ESLyat##G#Lw~^`4rNUx=^7CE$Hyq+L-TEUBKgthCq4ocyA&^6jKS)6M!+uO2Kcb
zOVQL#@YU?f<pt`ic?}G3m0X_TW1$nfM^zc}5D$l^1d(bnZd2iyA<0^M=jeqNq!wy(
zbcJMmt&EwTfO7^a9leW%LR@Wvp<P`B`Ps?o>3*Qj?GIPp{Oh?<SU0uu%Ey#s05;&1
z?i6n_o4y%*64__7QE*wsW~g7MEJv4pM~dREP7!QgPF7)IX$U2CZhNjyg|rw{@~SPO
z>pV9s4|b2ZKq6{D-l>b()wP?ajF$%4CkN}!e{jA;Qp|<BN#shugccbY4*wNuNIF(^
zIKhxnlI3Kh5E!_T#C>9A(;2=p6>`M|l?GQlAfV}^J30x}Od1W3dd0Uh*VpZcQ*nuf
zYG=le_q@Bapl<SzU@3J9f7WN8tSTy{m4DBHTkcNhyv9O=!DSlkRy|3dxC!6<qd}_9
z;1g)cpqbDZ+C)z(!t_XKP2m@h=d(vQrA#!-;O+D5DTaVLnizy`zN5i_Z2<4ny#32=
zo3o&UU0n>6<?`!6I2C-)4Q)S6(bdBa-Lnz!-g=AADQ>#o=a*(bXEtAaej~jO;gb%7
zHVZ#p4t@bb=Qad1)pph4`|-L2Z--%b{b3Q}wpy7D$E&O5NAaa2uZd_U?HfJys)(1?
zUZ@s~gOc*Kz=Yig(&?~f6saSbkhq!F4E<Nx%khsbl=>27u(!p6cr_^4w!Sboy;Oac
zc1^}+yL6WEpjz@Ws<aOwz73KNKr3pXpkt9pUZ;wZfsw|=3?;jeT9(!pDXLGsWL(-^
zjs104&N*RDm97X~8E~8ZvtjzuQW-_2JyS8}{`JBjev{KNx}h@<=b{6T1T?vCp|_Nx
zLf@ykt<CP6lS&r@VRkzy_xb9xAg?~YJX9t(8NPl~o|GG@iC^Aq1n^^St&|p`mG`ie
zV~2i}Mw>CSqF$b|ZJ71<uYWpyLV;ZcSyF}IjZXdba<j21DRmi%yLu$il!OS>o4dJO
z6W-B~WEOJ{t+ITubh+>^&bLx{^G?wn32yru)?5idZV|}55B*yP_of}Vn7w8|S<}~H
z@q65;u-re~mDiU$cId0B>7Dm&vjd4QNGa~v*Tj=0;mb7g28pBar5ovb1vw;9&-FNU
z_8pf>j{3c89Zk1@w*{JQDZfDudhxDa8KV5o8spPpiw(XP>FL9k<W@B{nFIPh@R%XF
zi^!E6iqCMDpT9|G)~WwJje3VhkaX6!UB%Z^*YM3ZXh)Jmu_}i4wYm6NGK);Hanf8F
z%iK7(`;Crp67+ui088CTSf0uwU&F--)|Z$@R$(v$tr4q7gd>wQ#Z|Qk<@KWzynh5w
zay6_qekjont_W#;N}x$*-26S)z@QRHg!)AI3mXM(WAN$F<X_j{d?z+MV?O|DB=MI!
zQ3<YOSXkBLDzJV<kw8CuxphP(01&SKdk)F!GQk5Ko1x)hedB#EU`}Hx>a)`PDkn|i
zh^h;(I=(7Z5*au;Aqlwrx%svt39WWDezoU)!)Bx0B}7Q0mvk&E;7bGhWt)Y<nI^-x
z=%B-tw1lJdXR$?BWKRqW#$vM9f-D5=97t_OU2%?^<4$>}uMKwRrsc`$KLM(mlPF5X
zXU@VwZ?l@EX)fK?l&UciVthRecx=Wr%-@Xk^6s4~V^U`SM@3rVq!J=>d*d==9x97z
zijE0fS>lIR*_V^EGc3{odMhBS9ZFO&?T(5c1$wI2r&dU}xK}INjaxz-o4ZPcZc4UR
zoiIs?VxAUL%@&VTNBD|%iw!M0M^QPA>CZ`egihb%1)bLiTu%)pv>{sO`g?9>`Gk*K
zE8b}+FJJjwx{k?|M+lNRif#Ukl~I18z+bh=Nxyb>iiaAf7)%8mdFV#eCpT8=Z`sPy
z-xsmSV0_<69(hhv8~nmD?7P|z%gVDaoKX7Cj6FyLrDK65klw)utkG6=iQubN=gM8?
zjBPZWWi4?0@Gn@)*~VG`{l2t2rm)Do#q7Hluc5_J6wK1+>e~xuG!aH4#{&#^hF#sR
zx16~h4XAz>>eSZMJC9{QhVg&3IEX5f5M6ZMRa5*idPnxS2=P4eH&Y#JAnT`~v?<BG
zk>ysP>0BJYDCAGw*cW~dp_c9xQH+`5At$6IJ(m~Wc9OrI=RcOYQ=7^!P00Y5Ra?l<
zFeZP|*8~_vM_Ag4We1<aa&ZY2)s0EQ%x}eqy@vv#U5T?b^f^<;){<e<%~&y0W@`Rw
zUr@O|L)^(!$u!HA03XNWbW5FrnwnVG%#T$3jJJ}FZ%T#qnj)Jp_jEzp8Bigu2i{rZ
z%6~hO!BKxq3Uh@WIA-@3B>lMgajc0ZNepB#*tW+G!Ek%b;Ryu>+ar9^H-Bf7r<b3w
z?EY&x&3c#5f5n#^NqTttx4MT<OD5?uFTB%f;#oGwkq%R^u*HAc2giZBqEk{&m#)MG
z5O{@<k2-=NwCSE~D&K{NM?cc?j+J)hc_I|5UR=Bj^yYPxH%kS-@{X~7duO((ko7Nx
zJ&Fh<8%pk>oh5fQV5yE2B`FfTE>wJErNNRSZ%isACdQr$2WQuhAAk}ln!uIzdtX8k
zLewRAT3X5`R*<zWiLS%&Nq4T1b@R^AcM2D6$-Qgu%sb8ikj~Qh8&7jAwiluE5cI2u
z88_!_^rMjDm`2l@HYz%wH2GiXI}S^UGQ{{B2j#mvm_)#w9rq%W==xZgkb%w1yZZbL
zF~W8%?oNz-9hZh0-$W?>;`_=0LMzR|=u5pKhRT!5_V%TKbtmalqw|hDptb`w7sex@
zrdjC^j?(*3LIr5&c!8>ZxkQvs3jXMWYKCw)Xk2IKUVgb8c(76EG$fov$ctnSe|C+p
zdYjtr!PHgAs&=HBQj8Jsr3X~Kz-|M;0Z91C<;<|JHeoLiQU&P3jaxKz=@8_@uDzxE
zxHYqz{S#pdfdx@q#9>euHD`DWyFtDiv8*jIj=|`0BMp52)sTMEl;MOY6{;yNo_kMG
zrQ|^*?<Y^Q`Q$<aF0YsDY1~qBdTCQb0|+&*tJ(3Dy>qw9;hC3ilZH6}{{=WHM!WA{
z*1i=B=!ci%>=GGYddhaO^s+!O^<a(Gl62R26<VRkIH#0U%xsArmmwzse$Df)M=^y=
zt9<K2yIai0PE7`EtGBl;bg;fKYlu<#_mS8k$HV};)O=w^>C&b~p?9m2Iit1K3F>=B
z$&uIIwJjtCDxJ#?(l%E$g!>_KX*>43rTMRcitTgQfr_M+O<iTd2h5G?IWKx+K$;Ps
zaQyFKI+~qHPVo8|qBNmRF5lv<OkG7+Srp5R5?X40IQ7G2&&A@&CgtSeUF08zGYM}c
z?&U6matt+3Jg0%Y4~sv+qXhI95}cVkooTUrD+lpX60+yzpS|HqNGT&ZgRG}I79)6T
zyLuwz9q*-}ITu?_qYLr)AFkUjXQq?eywk}I`bv*-F#8J5xG7)HFD)Isj*JhEk0c15
z<yyIS+W^%N(ysv8PfdCLTuW@G#ZroR9|TqJj|+K{71B2B(ylDIn+nWwJwVBOP3ZXW
z-Sn7#RaYjc@H%t7vOIOtZBgz|?Z@JK0gU%rnVaF((tY)C#Tm|T#=>Krs~YTO)dMQ#
zmJ~2tL;i=VJiF_?mhN|wAtuljiE(avN`r2hGxd!I5u-h=zlI8d3bHL9gz{SG{oq@5
zyc_dHb3FD?Ut`66(z9;<CX9Ug93sbErVrWlrZXjz%t+TkacCnIt@#wJ^MZDt1rvJ=
z(2`$|@1%0R3U*2O+sW<ndbz*YIz@+o4b!I*ZkVLi(q}++KMWhB(M}>zRvg2-l6kY?
z0r(qFq<@ySalTO)FmOiQGLcU)J2+Lw(z_u$xD+F@%@0ejUwqzAko=c)_q`{lpg~l8
z6>!}?KvAHzQcSCq*e=jq1y9Qq9x{-TDSdN5!FA97={9XZiB$~qook0|{mU8yEN$z!
zlLp@F>%or+!HGnzO-;y~dFmNsqEP1WKrzNJ1Q2jr-ET86t!!~A5s=>^nlDpKwrQC?
zKon#qMsY+>T7!J8eb)d#Djxq=L!uDv|FCfC1Hq&C^kvER$(s=Y3^6ye<tRRJBOdBh
zg>R~1j!n6uY-FghNd`qec__5|2^ECFRi$TN0!m$adI@N;YLhuvp3OqiynWmY2eCZu
zE<Lv#lL=y;+BRw9Rms4Tc+v==tUL&O^nj8m1N=zK13N8&CK(SoM8Y9gQKZ$DH85d~
zPR2?*ZlxZD|J)5%;}MPHSHEqN6JV;zl~e#gy0W}KNWXU+&z!PL-Vtzb*nV51!~=X-
z+@<^TddzEr;rZY~#DC0K%oNm&QHiL_uB{tWRxO-Hpi|i8)EEh<iE(b_KFMG&RjF&(
zjz)|Fn$MN!frl^~^&6eKD$xTbzIQS#v>5n9-IbU{=EA2(o9Vk;;_!Y`hmAs<I2#{P
zu3T0t&pXu*$DCY2v9-$a>8Y6x>W^O|@`H&Ia#?8oVWfDD_B=ls6Hh|eD9S$kcAD^L
zg{Y5c6<BaY;M8=l1m3J*VGWQC2wG2y?@gN_`AmAYc$FV9T)Kt6kY26Ae=|UNpRp;<
zE5t%L`$Of9Y1Z<NCUV;WE2%4%?iDxZTCzpVn4HG8tF3M?VbrVpRCUs9B{}lFw?{nd
z<$eFAeKcTp*vS}Y{}{d-aG=uLUfn-d_UYyz+|sL<UJtSq0qwBemt=X_e9D^7zTp}j
zJsR%7W%r>K8qGeMqaZ!~N1HV47`HOAWY9IT`ReOrkpYp&fH)P-Zg`keLUw-={=>>=
zcD?xch(c8kFY$_QO$fxDuW2oWAhs~c)#@yK4lNgIW)MbKq@){4&e$zq__1a}nV2W-
zA8T4C?6{O2s>_pciXa=!`V|3x!xV|}^7wG3yoz?mLe^Y%)LwFjpXT+zpo%WO*Lw<R
zpu$(hOf6V1r}Nto7F1>yf{jSj;`)*UxUgfFQChkd3ZA<o>e*+e7>oee(RYjW1Ev_j
z5Z7IVeSdrB@=INl3Txl(Z2W|WV&7S#Vbr;CE$QOydY?pWvLMiBYh^?{*1l<jQyRv&
zG8*tPdUlNO+!z1b!=8*;VO(z^-=lv>j~?(R%{R;1R;XfmJeiP_sSSP{B7Uem60!q|
zmoK37Sc$hLJp#bZkEqj?pTH>s)NE6bkpPBsJQBDD!fO8<vZebCBEMH2o^3n5=x3jU
zLwf*enCr@<Nzz(xbd6P04A<x|V_{RMm=-B7E)-^dv5*O*M$kG$Iz9p27dPT0IDYBr
zf~gd#{0Ek;fnDLO*@gdP#a!sOrV3x%v3}Fpkxeozu3Gpmm*NmUhvUXDxz^pow_0gj
z(#n_Mg%`>W3!h)#I+?Bx0jY)`R~NWn_a7C2ZFHQ+iv9;6-&?s)!F65bB+duDFPS>X
zdzm0ZiT65L(xEl~jdftorpTn=Drsg#BE4_pA@EG$56;ca*+^`Hhl+McE?UiPZ^iiM
zXgvYk;TmVXerIm7E`ZA1eNT7h-{uQ)JCC^TT@jea;g<d>Y2_rs?+1b;7B?HId<yFD
z=@u0Qs$XlZDUT;Os_9Ba{hJ%hU%&9BS_VF0g@Z~{vtrHs{|aZ|xJyYvZ}<zVK4E>r
zUs3pO`PAU0$h)tvG2N7}Bj<w=pGv-zA9DsJOK99_-vr0GU<eRB&RpL~IL72ZvMTDj
zK@yU+6NpzgZ7k=T31z3m7*&Z6Nn~Gq945p*P^YS;rG>G)$|Yig;K5j&U*!@qVaLkd
z7#Avqhk(4RP5rT2b^wB$vj(ee?qVnl<qbx4IcA*WRiSnfj)C@KjN57KGYwG)PSQ~x
zbb(ks=B$nIb3u|b?B{s#OcLUbRyCw;F4)LNH{~KIA(x4preQacAm#{NPstZ5Kb@Ce
ziz+n`cf7D2XaUcS)4I!u-D%l%M!mvX|I+7Is(Qdd^O=Q;n@-8P<Ve`Y8C-Ep6|X7Q
z%96L&Cz#?THcxO<D719Ok0w@oqV<Cvue8FoPW!lu4J}lpbQlH6Gc9$N7;|i}_|;3e
zl5^^m<hAIH%#4rY;*RR7U-M8o#=v<H-30Z_L4S5o4T&Movs-~xB-eei*?>*olMUbU
zEZXYAfcaFu7f!NL!T?R;NNr@3^yAd<-FdNe)XBcI+;@??)%bJmmT2XxN<kNF{6r&8
z@UpMkijqT(^v&VU_L`QAOR$Zdo-!}Udo=pq>aTVYNf$&VnWBG7%ZKPD2N*4em4~;Z
z2aze9et1fB6{U&ZP)eJ;QZ*=Q?4(6`3b(=|HCBJ9Li3EGo)LGVmz@>~>!}N$f9#cz
z*{>S<+F7ZA?ll#^f|g(HPjPpe>T@cWf3J>$1n787;AHh5e}2cdV`}kR!HbH0P}6;L
zr8caoKH_@pJwsK@>g++FG53c(yPVBsLYAEI)3&|UWY|qlh59FJC}K-br@WS?nHrD2
za${$Q88A6bxtoNR4eYv1Dw!0d0}z3h<>lpiLeTKy31sX7{O=v%aKq<UZ%LIHG}5Zw
znw;oSwBIBh2w^b{pXmow>6P`@+lI5|zCm`o^egoiGN_9u^JYE*)1y+O&QuXY_=-Hp
zyl04bzxW^cPDA5xQe~eb4soP36n2y8|M%<IUDn8n0p`(5O%zhRWsY9?^*$Wu(N^ba
z9m_9Fk+uwa#5jcw{JzCKjLico)SmAr6T9a~4x;!P<^Su0E^ao%DN8wWr&3beHyl)5
zy*=^85h&Q4i0}R5bg@qn;+Io?hVe0#t-1lRm~j`fd!sxZI^AT}1oF0H9l}tNp!3ub
zjnaLKUEPcv8Y+;V2YMV5gbKKl$V@UvRDL5ihINq<>LWb)E0U{*ZGk9ARIDnqEI-ey
zcoF@ad@E$b*UTS)5eXL|_TeRRPKGi$%>?Uf_hs7djJpHD1PT`hv^2>|GCp4>@jI*9
zCOo>@Lz&~F42XDAcmLL^2&|PPCQCOolO2#H3}(&!J+Q%tyk}09k>@9a%Bt7Ql5!rC
zpg(Odp?==l7Ygq87dUxX{4)!o!=(!5@_DDEOX^ofAq-(7!HYUyLbR2Ue(hsoL?tE<
zqPl+0Q(-4A^WV2r&g|t^nSWl$q7Amj%nfkDi?q9CJ5*KdmQbOmwDmC-NtUJaXLBgw
z4cU~EQjzyrT(a;gua^<u4!@vzu~2E0{2~`cF+aA4JjGy3eq`0?ag4u=U}fVgbR2&%
zObDC(9u(5E?=5B!YyN$nCYPl*>_#Y(t8H>Is$q^p$z9^18jd#_;pmZT>`4yfWIma;
z%azY#q9PS7FNkm1v|`R!q@vkSXPQYLm-oFU!XbKODB3i+PkWUw+xqL<&lO_Lr!8d}
zmBKX@GtU1)mo6?snJ~w)5}fbdB>vs#$1IT#bvK#pH0_@iQdl`ca5+9U8obzJ|2-rp
z_TkOy6Aei(?RA;?y{B1c+3&BaOn6_4X2E+|R~3~#niyijsP$<BKHw&_Ea$V7@v1*k
zK<s!zK2Xt0JyqdeTh+I5Pp+akrq~@&i|UT;7#X#k(94*1R`?Vzft<7w66nNC+?Ct}
z%Q6UmgA<O{b+ZyGCN!LTiu9}(6cnt}wU7<-s*~Y8F(l$>9p3C;{M<mmTi**m`|cy@
ziJh)b+elI6c5`zB`>}SS$`FtA(=%I3F9Sgswp%^c5Aa@KAj8j7(ED+6sL&WIDn(PY
z!L7tp8Jk=YQQ`F~7&zV%5M`n7U-%hWutkZ^8`uj{amY*nnEiu9)+3RDDS48_z=&1m
z%+Di{SRvk(BWNRoD6vmz9=8a$!3btborbhis5bfG)rUZjf842Q4MuXoV9qIeGKC>+
zw6o1-GrZ>#c5A5AKDZ$DaI`a7qJ5XbdfpdrdANz!K*yV_5&FCK8OPVY>r+0|sbt<e
zkvn$iOGN}s3LB|p4@ftHVIp^TciJ+UHLcq3gG61f8)H5b%0d$~JO3mh)*g3@UM)!x
z)oYH==T8Hq`}W@|YvA&Sxo-dcn}v1+?p&m!4>`F5;Lt$^<M!PC!406lx}*JWS1*sE
z{V0HO7~uC!L0Sy(kWE^Rk_moCfBAykm$i?EZuF(1DK^<MT<^rzD4E~Vs&+Yn^cM@Z
zLB9kWtE!4-TFN`7Raz<NDnz<p;AexZj=63AJ0G^v>3Vq&Yp$CmdOj&|T{})-fLL4<
zcoz)`KQ=P8lhT&`Z8E~Q2A9r)KMmf8LcIkPo^&J<{FIhFV*T64`-b>OV$SSDvdXuk
zGS&B-U6t~bJLHK>>QoE$;8eZ*8;BD@D<CTR1tb-aR&8KyPR^|i6MFYLuVmdmD%K;<
z%ZIq8KHh&xbp41II#ZN~vK8b+8XMMpfG;~w8T=jOB*6MALt#rC4vf*N(0Eo8$zJ0I
z&>Esi`&CJZK_$^|po39-=Uhw!E9mG@R~Z8zeW%(EyZX$_Pge5Z)-Rz|bRjc3xDIT=
zkyO({ndxgm75CQdCAYhiQyd4sLZ%jmqNS7t|EYTZoz3GhS@^+L(__T`!LC+unEvh+
z7}0_Z!t~r1+56r0_)#F!0l+hx=tI0G!!UdN%NOm4>y_k8pTSB!g3P{f(9*~3dyT9G
zpcD_k>YX%E2)NuBSk#z^ivm`KsZHrKR!<r#`4=}~!UKeIznLx?UEvX?YW<Blk28C9
zW&eCCm*LkcfjpN*DUG&0x&6;D9y}n>>SFg}aKp2htMM8)X-{IgjBF)&mZt7Z7k9mv
z?yQ+53{Cwm`6{lQcy`^pjGU$5CJj~d2^R1}Z#346_nw`$$4kImy#kQbWriafEzp`c
z;=3<l*_MBdX-!lgy%!=u)ey3ImUsYz@1=6-dKQ0Pig4R;ecOV6Czd&kl*JPd)Cm5}
z{!ovx`KD%OB%ylYZ|K;-Y3jwF7;(W;Eh=1@&hpFv@4e^Dm=1O0$n*+{q4pm}58fB$
zN{#iMj8O)!GY@yfcK<$T{a_xx`o(Mh{p)nw^Uh_{i7tX8$vxoV`vNeOJBuhA8f`2R
zIB~m%2*ORAtBQ+lUN?x84ns-@+0S2BK8yX4`^EJ&)w|}}MS`7K@AqbCL>u*+l-{-9
z?b-L}7{lM<=eT5|bqttg1p6GanK!^({KuEH5=8lnOP+BwU&f&GB&OZJfS$V%l)0i~
zCx);nKona`a2&QpB%CXTN42r#)s8sP{^W@_T_!>hY-Z_`@so^=#{{)d7gaj_+a=H9
z-4qUogx~J1wro20O1$nNOwVl--<w!a+a?GUcdOYny*RVDnj`0iRw6Czw<p8nua_uT
z(dnOj!gJ?{z_J>#J-UITDcFu|FU+o1h38kqGH^4x=f7Mq+#z?saGVl5l24l=QFH6s
zHBCh?!Q^F3EWdit^=>ti{mRp&zN$DK0iEVbz$V4D_luI{(c(vWdN#^{!IuN}*3^U4
ze3wLd1jaJW59RS(wa*Fblpa?mEvlBWLCM?oW<GU0lJvih&(fHB=JZY9l|}2ushnRh
zJlQGX4Z2;V#=nX)thgF325<jFgJd8)AiWaD>jU0Ng^J&+b}6RW(8>i#EVg#z_0Lm@
z7K1o*J9&%P{~7Cuq&(g!PHWd~%UKB>nO?q^{`l2tzm1f9d@@$g!sdWOT>gV|f9d0D
z$yT4b99{$$<_>h)48&<4Td)=kCp!0gsO$IdTi5TH{Ed(JwOm;a32UjZCkBy8<e?rn
zI63(&sb~Eobv@Qki#6N~nf_gBTK7e5_Y0TS60E89N*>u4y5Cxo`tH;wzTyqsG!aTs
zpHWfvzKqfTNfWg<7L!6h`ryQ8HkyL5H@w@{|1Y4(r~6?^E3I0kJ}yn8>|4fX4R29`
ze_?gvO+@aPVpL0ZC4NYunv${$37YwBubZYj4b1hK&$IA{M<j@;n+WZBQ}`srzc4i|
znl6dR7TCvm{rC2^Z&4FooP&91e!8nc0!$*0`kqHnsN4ui^$rxIb(~PNz*9+lRV<A~
zBFQJ7S#7;{FtJ#wIcy3|D!O;TFKUk*_81Ity`kHc3dMKJi`V41$z@=lvCE*y!_q$w
zhMgW|*=J9dyF-F#R@Iy_c~i3V&Nr^u;Lc{;O*gj|uhs43umXms3{L719Vt2GOq`#^
zceMjebSdNe=#b1I))GVt>sQ^tXu~Ov>UWFU_xt2@q1#4w#sYBL!rJ<fHj1g{xxewT
z3f8qvB7Ye5-Oev4{fo<#P)g``ZD~7s0@b9*;T*xD>yl>T!9552W}G|gp29J@9ko{@
z$(QS#+PnHw1^*8IHPR?p9WvCnDTFFj%{K4%IKKz(<X!UbU+hAD7`1O4F8EEWN&Y@|
zaFxA>?|n;^EQ!5n)NbE*^se93yj)s*ejFr=Ibt%28<mWQcaboRi#X+C^6(hN4PI3f
z)5M1p$lhvaCv!&Ww%0RihW)e24}%X)ES^=J&Ba2OeEvXnFGQM0NBvVqnrlfp5v`G7
zmg_kDzpNGt#T#0ic}cIYR}~`SUya|kargfT(Ra0%&Fm^|jc>Fr|8J(?<5^!dLF)l6
zP9Y~`aBjPwulh%t<PBDror37LuwtFeBK!V_{{90=L7o}s9LWp_Z$j6JLAXW7iZLU$
z-j_2QpcqxDVNN}Nx$Io~V6*3-6Yh;>9jH%WuA`)4XuHy^w7$1vN{s#Xdpgt(?qee4
zWJ(JE=?juK0AWwxp&AxFJVltxlE(J*?e43K?$e{*(QL6a2!h7;(5srmJf-~b4{rgM
zfL!Ik+tUkyjKQNLrz<1d{qw#JwO?XP-OK)+VK>Get8~SP_|#9kl<gaIyNF-@mwQ8G
zj=RUW#T}w5BWx1*u!w;V@V$k9er@p`b{ShBrF13@y#z+`)vnLP7cm*_`e|8`!c^}V
zADXnWK*cnWbS9b{(B3NC{i`89b=X{?KE!G3{)p5ssodHYzqx9>4r3p?Z$zeSkc>T)
z38Uh`1`IExJBzqYIr5knJe+!}^;i@o@e*p2+MIJsyYH9Sth#-8km}H&W(NO@F*I^*
zGd%fhD)IT#^!78~*Lk7@4j9J3!R-;!-@mw5Z6bxjmJL@aVMzq>pryGl#%cehJU$cc
zPmPwznbP*S+i)sc+|`Jr(iu6Km8o!yId+*yJaXzyLi=6T2AdT%f;oSMlSh2=PBtxS
ze9MuD@?i0*dN;^~XJ;(NB%)c0=N&)U!qe#{hQfVFw7*57c|XGKdaNSQ`f&^*7P0Pe
zD+i1FGfdY0YIZB{*<|(2N-Z^4UsLG$@1=EL^2D5%oEJ-K8P(S_&zi9D>gBfA*63W}
zHt{pC(3T3w5ZObY>JWh+`v95tf)v3i?K8Shu@bueK@O=}cBC@`OaI2D4RO6KNo#rj
zVet+;U8z-iRLl-GBS*W<iswxWRk@(cB&gpY@5qEI_*Qk!o2S?-lFzQ(T<R2|IxsZJ
zlY7|uPJJ^6-rgk1f!`GI9n)56BC&DbEYyyljkNL6g}3j*D8WtI<Zf@>XD_L0yAOu(
z(og{f)#2$Wgw538?m$9t4oRJ9FP&C%m9jjPSg^I&USwg}#`ZfGW;42=<0I0a`ez)0
zwI5M~=ZVbng{wrv!Gaj)-Ifi7OwRwuhk+wOY@OD>)nBu|6MV6>UNZ3hi2VDa#gZUL
zXX^M~5lx)v&TQ7^nlr+jKb@*3e|7Mjl}|aBEd-}_$}o~@OkJfAi|k@FMq&;OVUyi<
z?Ksi9;Z31~!(y!^mT9IUT-CR&e>&2F$Ehr)pcwoy82Yv}Coy<`u(y}U-UHj)jSr^j
z%~d3AqCGpn_4ZENVB4rOsu=iUB069rW$@qriDS-;PDl<Nwa&x)cf@CXW(6fb*V4j-
zFcWQ)oUmt})}x@?K!Xo?r#yS_+QNmPjz_{F6=@+^HRTt#aO0^dk<bpqh}wZWbxJlI
z14z-4eyG%>sGCwrp_@6|)0=g4PrA%V<?;LSFMc13&V8XbrW@DB{w{_Gig!t@MarTf
zy0h*3LSs%wfIXYF7D_8P`gc$rJAf}NKD{y~Qaz~5uqUpPRXvk_X~Np#fIlZxWoq`g
zbjuc19`><VIHV@0dm6Dm;=tcOKOy1`a&43E`Stz<tfUM~^6<`7r{;qAcvc?ki_SWa
zjYDMOP;=+@grbm*&vEOkgQy@jX&{1=mnk|&QA$vlj;s2@<;{1lDA>m_f4#KUf8|bG
z5)`zDtntRh<ea_tO37(l31Q5fA4(Iq+b&`R8!sf2W*MzEL@&Q$DSy^!)CF(LCl%M0
z_&R;JlzNug9}qdfiCXpems_KaN^*c}qdE<lbKctlL(-K1feOwY&abx)Nz*B#FS}@J
zXl#oQqz*IOXA$KjUA@kkpGRb8>UnN@gXv7Qo}F0R612F4AzSy440~4Nt?kCuq(>*~
zL~RVb?a{o(fuo}pIiLq9{x#{Pk>#xVO#;{7Am3)UVT-73(}};8bbAwGQPqnQfFh%h
zA?+-2z~$G;XNlo?Qt~NnP2DWvtaaBfuA;6baix%Fa3S*JkKUIUuXBcHQ1*-Ggv$L#
zI}_e}CY2BvGIWfUd(rM&^b4xW;m0Yrm1(e)Ox@E=^Qjl|1idQ()2aR~qJF1rI-$gw
z=>OiJeoQdwV`35inP6(cw28!S`Hby>SipKf+w!!-iJ??_{@eGd^$eG?OW_1wN<UiV
zx?#hT$Y+o*K*fA+PLoa`ukQ()$ORJY1uJ`*l8#KeKI$DDXiwRPZC%YJtkbt2aVko%
zd?ytH_~`=exBS_Si9L;ElD$_Z3iv^~lBnI!9%_4pUbtE9;@0*n{!cSu9XZVgHO$mt
zmtY`vP^Jcl%v{K2iGZA*94Jv`2a{WdwoT&#lv1OMpgBcwP`UpxW3o`wam!TNh_qd-
z*xuY6JO3qcu@sOwT1ve-&YJJ)>S{d)fNc|OP7Qdist|v5&#EIRNvp)VpTO>ol+{8{
z4{S?S&2?wk76wXfGxZrBt(r}_>QLQ&cYo#|PF1t=vSiQO&N@Qs?o7V#+}elhf<cpP
zE%+pC7{Df{+(Nfq^zeUBxN@B5UDPi1<9SB4yvCuY9vJLBrNLI;B+I_#f<IGDeJhE0
z0;G$GXKIo8=1DP?nze2&9Mi#{9TUJnS;~RYC8W>>%<;#v)fb^Z@7Ty@r`d`dF>5V!
z?RBn#4tS(XaJ2j$ZkD3KH$N}W&<Df#%=ASun_KigUfC(}cX~XTm>Ef{u{HZm&jboB
zbM0GWjL5L5&;k)mPcX*;wz98A9RO6Km>d>e!-CIaI>wAENJSqk$fU`vE)X=Zg0sa&
zUG#4+7RHv#g>W}Se-$)%_46I`V~lK^A0VPYT2%|<P7>?n*}4|Ik|M(0J<F<1-d;*C
z($yCkXq!~a?mk@~NQUtQvZ0s~y2uuuE_{766KPut;d~HNd>HEf)04qgw-2CLNdsk@
z9uBp*#iECM<<CzzwzRQH#}Qgm-{3Tl#3T`5y0R3WGkU`_Yk)WQiM#|P0-pcu=s62z
z!zQMeOPx!5GN<2<ADGSYZ&w?Z?`(o**}h5=$2Q!2E(HJSmnFX|!NP88SqyxWqll>t
z+d<xQ&lGMffLmxv4%d}}81~Dpa?p*Wx28qKFA-Z^O<{xit}LV#Gs#=E%Kp<|v66m2
zx;szoP#&S3<5+8BiA!Cfmi!j+i)%8M60k5@yv}wq-1=+k7tgw@mxo4_`*uN62vCq#
zK&O4(8yE<H90USoy3c_3iv>A73HU+ZVgCFB0!v_*fPpWX`{S@>#)q#e3$plcB%LxW
zd2`!&+#!M4%$bniE0|(Jg}xAxl!Cz%t`H5ErhE&^)RJ#nv~Y!&@u)a1cFd<K-fL?T
zZ_BSO(*@3%*)ap0Pttb^I2o>F3;FVBPGpX=NWPIDgYK(Sg_|6w)rkAss&+24bcqkX
zempGA=E2v!3)qI5JxD6kiz^^eac-~MZ8b4`Aq*4^E{Mjr3MetS+LL?<=1>B`1qe3J
z1$8>CdXN4~)A(&7-|Oz?7LonAp|)lvC#iap7=m#$M^r?JN{|OEKtS};#wWtjaze?e
zYBqTZXQ2buWrKf#ewlA`7e;TP44w$^&)sOV;o`iblH=Skjvwl5H!I~YqfYV_ucYK`
z9alxj7M}S4!+}?+giI^y)kK`z{Eon@)AaV;H3C1TBT79CUwMNktQQLCx;2Likd3Ar
zz&!EPj#tqBs;xiGgeyvOeXV+&uu+i)dL`(C=)RZ;VEOF=o<YEI=mwX|(_a?Vy6Vip
zG6Edhn6y@@Zd0YA;f+QWpQC|3pznxbo?cwBzaPWvzhCPug`IE>S#<#81ztP`A3VYa
z@wY&m1aPy`!Qh~o*!f<P?WRk5Dc-c-5m<^=ebV<$^V;e!6g4S<;?`<09plB<*qF(%
zl6%<8!!x7ykZ|CNchKCu37!5_L_M})jWfsa0hlpPQ>I{G_{Qj{I*Vu2lw5yeWg;SQ
zU)zZ(lS)#f)CC5n0)w?B45lZ5aS-i3G26uLm(nG+;MS~F5D(1eM>CuTu?ysQwC6_<
z#fbi656iBr2dT@fBMnMj@x>@Q!_lyF3(Oj(mjdyuTi=(#h3Aw>zER~Q9Y@YEiS0;w
z-Rs#ea5BPg+W9Hrgf)JcWwiBQ?+<0}f}5Iu?_Rv?pB;Ehbp21EX_-yAn|Q67tEgwV
z&E=7MN2*Eza#}a|%Q!pkug#Z*NfgR^Uu9>#k-7H#vI|a(DcLj7paZc_xrNBE8&e|3
zwDy;nLg3r#0$aw>$7kMJpLlHLf$H=2jy<v1_2t?l2MBibNg>t7w?g68%r3F5B=w9G
z*6#F}P_>Om4L{xME7(JzJ6aT}O86wapAw~OG~s!0D+K&|Goc!^W(CbCx&6RTQT>o0
z4vOw#TXpJ^kH6b#+sITZ6(Jiy-+PZM-S+_L-H*0YV2#&)YlopuFdF1%oG831@>K5C
zK2~VAvVy^?1DGMK7Ir<+6|HCN^Em7}Mn6O^=%>$k>UB?>JOd#Uy`ZZUohN<VSt1KW
zxyd2EFBb(exuJahRXzU?*WyRXK#d^?#n3HQ6fmPF_m;`dw>~k%O=xcxB^;Coiw+sG
z1B(gx8weGF^H(tv$;$!(<W-Q+#B)r$ss|^8kLp$GAlig8RuWzn;E;fC&@ykm8iy37
zXN<wD$UCRy>9N3^wti^FrtENigyq1ppJSs@sL)d|S8RbCF7Y4r*(OaAw-8F0UGmAi
zr~Yc}b@4+VIWroWcz4q~)KN5b%o{~Mo!qxAVC4YO`Snx9S+m=>?3;$vOMRc&Kw_nd
zJxSzO`}D3Wa*W~slJ$w*W3$bfiMlW0(}4s$rl~AA9J`<y1(5z9e)@<;gH#idKe5#M
z&qB5hEjfn?nPbTjBl3H*U|z~USU;_Q7Vj(f@2(85ufq(0j}z@0fX>hZnB`cXBKOy9
z)~8LdVB{nf_-Xsu)cZbX7=kg)U_)9aqQ6f&jk;Slv&d{z&6kOj_8r$QabQi8aAQ`d
zQOfUC6eWD3@k5iv@ykl5Gqd4S<4j>se6as6uD9$OSeDq5r--bQk4bi@utn=4^d`tz
zKTJcBqkSr6sBn<RO6a_;&d}Ga&9moEzg|?kSrx?;Rp*Q{fp*#oxXq9p38V_Y@C{GB
zv^b1~lZ^HCa+NApDopN`v+U)naQ;g1QU2$us%?|P3Nm)X;TSX`dUeAW4LJj{OAp9t
z;{E)I+_3zR7HhbcVNLEzT0ANYoIr|6r)~X7g8*TRq`4O<?oAw55q?St*(vZYZ6{@k
zKN`Ey8+zDxZzo5_lRZ4%jgT8=DL**?qX^?pYI&2JEDyoCjvgbz9z`D_G`Qa{e(FTK
zKP0ckCVkHKG}R$?O46<uG+{p<$Glbd)ccuugV@D)F!qlP*fBlET`XeLd4c19TI-cC
z_w%N|1B?)UE#3e25e<j3sGe80B=e)EQarJwsv1~*fe~aZ#tU1{7ZYDn2Y$+z2vgY(
z<=%9@#YRHQwHTbKaz6$cXeD`wf|1&@UN0=Xf#=re_UQk^b`fI?vxjZ2Q70qb_EN18
zfiL)IhcD?)NV#*&<MTxY1gy~Yv>9n;0CJ^qG0&3j@2yJVGi`<789yvD`7n`QdaYcE
zgr8X~Iq2ONeQBX;+TGJM|4G`4&Rj|hGcY=isf|qX8ahj}r2aN9qgNAEtnXc-#7LIT
zNc(Z$w4xqU_u}(0Tu8&|U%dwtpVU4&UNmls0crUZr1qOBn1}$(al*(J1c+dQmIt60
zT~1c`P71Pczsk%Z4aYH5{GQhPCkxa%;@BqfS5$j$9f#;iuHrK(dQ%f;sZX=B1%!k?
zgB=9S$KW?20Y4Re7-T6!A?AfRhH{??MosZ^2o;1*La&kd-!Cab)YqPMgvv%@L3MR@
zs;0zR59{oHx(t9dt)tX}k23v-0O6_f*>|_od*Aa&>vq<Oc|q}xDm$z>z%xZiE^mG9
zxa+T|I?Cv$YiQu=*&C7_>i-^gN-9i1X7Ty+=O)}uQPSYnf7|Z01t+JqfdWlw>4b-@
zr8w?wZ+Am2EXT5gj{CVzQOw#CYtC_zm)YSBeiYxd)j;2Ae)c0>Z=rYw!!^jE?~R^-
z3HMJK12b&Fl%^cpixt$wvXipUrx_Q?u*Sy~(u3i+r}M~3mY~1mdw-Hi%q_h>Xz^-a
z`^8!vTAX6x+fdY2Qml!k1oHmRm{wC6WDlGZBR$6JWrvh}N?OnH-1VOT9na<#re>x^
zXUxe=${zf#hR)a3l_=Bi-qR{N!L}@u^sLdD$LK^aWW;)6x7h5>7`cMEY)!Jg0Qx)g
zq-lgVxW6y7UvP%4JLnvMcrp=og&>OR*&uif%*+&4p_Qa;Gm#XvNzX&-L8lkM|Byu$
zL4fZ+HIx;FpEsMmQl_Is1TQhPK3ycY-@@>el+m)EMMbkqf#+KsUcUzv2se|YM<AP<
z1nW{d<jQQ_=s6LUSNu9}F(E5>TNIKG$<gv?3u1O@0xcaN@B&LJb^B!(dxYc*YH(45
zGd{n87f`UMM7#)*O_uGzn932}WH_Jet3Nnh-#rX{%o^{T?YnG7jEJsfYyE%MB;s%;
zenus0UOYly)(&~`F6j`&+<(YmcY*_RAd^@VygDSP4AR_5n3IDfI*~M1SK0&y$eoHY
z?`fMxqh<DWYq>uiPM=sbhcp!V*a6RCZPkImni#x`;hl+E&`3(E3J-gsp?Oo=pM@Q`
zLTQ+W=;Uw{GfeUp%>=K(>HZ3c1AAuG4x0QkIrvycB53}_{i36PMS4;>S^PXvW2&3X
zi&6o()?HjCB!3xc@7cH*a$5`jUq@FN*5ud56$K=u5$W!dkdlz@E=L-4gLH~CN|)fD
z7%;jV5+jBnITR#D=SV?7kk0q;@`;Q1vgbVKocotr%Rp_bsSNl~_UL5CEwBbN?|exi
zbv;pjzw4pjZ7UVC&2pkBnT;oS#RNfa%l-yBc1eO3fv+nYPny|F->)q-E8A0hH#t2K
zTDeDQYf^&JE10?vU{k&}baWbR)^i@+j<Js=NffcV&2+h~z5%W6w^gdsN*kT~IgbNr
zo3*rUR99H;b!{i6TQb}4vlsUM`012$Y>(}I-jM_&!=Eqfe?W>>hm_Ot)W-ZOQcc3g
z44i66dabaf=j!iSQ7VL_nybm9Y1pctPF2$kt`c+FD4m#&z<$<?u(IsxIA$#kxawd1
z*N>E}bj&d^;T%gu`{FZiv+s@LURxVxLMGgLSrZ8|3OBT#EcGtcaeSRw+h@Jn!#9<c
znf*EN7vx(CXZsX)lECB_HYKwJO=(z?NCUEKyq(~K-&#?xt*I3UP2?hX9b${qAp)jX
z!C^15Mw2hm{{G(H=?ia>b~u$EaIjSYF0}j28S+Nr_Z;VHE16i0*&(~3BO)VhFs)xY
zmV0@^#@=ff9IK$8&|GB-hezJ!sv7x);n>{y=)VH$uwaZmi4nfKkG@2|xz1x9UwH>r
zbKsQ#*GmnP(o^m=XuFsc&vT#sIii~MSWfSt$0DibT&t>nz2%Sie5gHSC95WH^5A>H
z$3krHM<OSb)-QxvR`yR9G6Q(928d7oJP1D{L)tvWOl&$TLjN&;o})>c(V3yq@Vj=?
zh>*L>jj6!e^At8rv$xgz9)Oj>p;Q7=|68%om!CV^m+Zx&gXXSxTdCG&|NSA#to1rf
z<Y1xHbbS)-bPR6rx3WY~%F8AoAfVB>0}@Hi6tue6A8AP!vB-gtPI60Ew4;N@M~oM5
zMq;oZw;himHx4&}sSU>3x^8MdItS|&UHpj%!V?oxXnmlUHe#zIrxj10nhFQk4E}Vr
zh+;2xK+KlHhz63^O|xDxpBeZ~{!2|FMA%Ai70PpOxJ=my#}Uh2)_H#%^sz$lCh{rp
zK~cz`sq&s~o_4fGjT>xAbEZ;X-GH5oy^tTO_wB=7_L2vcGGL6l^`s$)sK=bsir1Y6
zXoTJ0J#&ZP&g1bsJjl;X-$?J@a>P`2%xw0-y!I~-4-fN{)8R1mN+wn+tGJbESj+t;
z0_1kiM#N!;a-q-&{B8HuexW%hEZED-pjX{iQ(a{FM@sV6YhO{&yx?jp7o6uZx)Gyi
z?{WB@Sg!3PW8VeSTB{RH64D&C>K@k1Um;?-ClyTb{?$@qAJswAq&(*OAzHWT8-L~&
z%-|qYc){SGCYr;o3q?XVz)uv=Z95U$I>1!?O}3l?uVseyB>6(*Yl`Nu6CBU;R3c!#
z3`3XBi5#<iSk5M53KJ`tZnimz8@+@asP!xC{x>6<Kl}!CBCqtkFz?Q-64|WzGO^xC
zqwzL<6-P=$AvD+4n|ldtO+hD0xi$EPjYqS($p(tSdJ>&<UR$zJ{DmQ?RC6}_$2TR)
z=mMU9BA6lZq*&o(ve(+LkBrGbll<3dOvIHzV2a-oxv<P~p<?BQhH5%W=)NGXHM|?d
zr>+qZeDW@@*U-L_Z2rmastT)SPifZyx-O^9p{?IO2|DsbIx<+xqH}wXuX*634fFtT
zw-vA_M|>HH{(KNG#m@ue0m8tZ+F*9XkiPRkbi00WyKuoc;|JESLK0>he0FDvMRuoH
zhduv4g)=rFr;qDeO4rpSn{d_h+hJ8f#poE??!1BFPI`~$g8zj>fJj~{C+P`s!;x@n
zm7Pgr`dwm)(Gpp_{9c7PZQyjZeNLeTQ&@`o>jm-VS9A$fwM}(*5xnSTe<{xM$9-~@
zx<Zd@n8vSAu10~|3cBrd0RqMdjMdlSlw8avs+2?dX67QHHlmTLfi=3}geW?nvRU*4
zA@Rc8Ppg>%m{92m6B*u)bu6Vh$oNs9WipN_2TS48e$o-L(39$Nz12NUiLb@|QcpdF
z!-xD=C(yVb-mit!rcAD3@rK*yASk#<S)fvuybKE~ca<gv$l;r)(+Uf*A?iR%$*=*&
zL`Oz+x=ycd2LH6TcZq$mKa8XU^w`3?HCSPSNFmsNToLiv@%?KS(J8pf#Pa5fls~{u
zER|3FwR(Zq*L(R17lDU<|AT*RB00~IUR}q{M+Ved>>X`weCGW?-(?>2>zAvfoSb_I
z%D{))QT68So)+l3hZ?=r&J_5nWUH^JH8du>V;$5xpI?HI*?JDiv}jCd4(o+wE;>>M
z*x^wU?-7KnXK&qq+Q4j1`i8c=UvhnVf$VDiU{i6PA{8~E?&TtPp?S@dkn2fMNCGQW
zoePoIrEb+8n*&gjD^#;kN+e5QTl1#}`g~qgdpg6a$jY>b*N-6ODhw#TK}YV$fdVa!
z%|onus%d`?WBueZe>giL)`9-b28sE-GWg*<#rYRPt>5lw_DBWI=l;4#Kz^Wl#t_a}
zOjd_GT*vzrDKi7$*3tuiTv_m}WZ?TZk*Loa(JMh<LgK6LzqjcAhS_)slqIvU05b;F
zT5itBHY<{|$p~KIEKRS?#bs_eZpC#3Irw88YS@nSQar4oq@rF!JE3RZu6Wj)<v+~>
z(`p9w*4|MRftUi__yON16Ce-HGORi>BskPpw;_4@Ir(W+&v_Ki<EjH=+RWjjfEF#}
zbiyxd{BCSDRQbBe7sBIk{7CRxro;Dt?IGiODakBX9PKd=L1XR(<Qa@lG&69*aOTPq
zMLPeU#x}$`r{qsQ)I~xtnk7Zc@xli|rX_x28`|<{tI~cOo*><o$|gDZ=;7r_GBQD>
zDZhpZ|HS?<H&<Xw`_9`YHViVq?^_86Hc+qG_P<aFySAo%5c9=?k?PHF1BaymmB^v&
z<c^ww)evu(%2avF>h}t19U?R12J7MGQCr5`Cc2oJQ=63xnWBPHR)+i~!t?U^EkV-c
znIU7GLMa_Scj9foH_tJj#?>;LdeflfABHuCK5_F#ZD~%Ex~@=oiZ2IPu2e3`B+%Mw
zwtWV;USMG0a9#X{nItHw)haK0ICGv4#ZS<ll+B2JD0}xe){urpL`yGc@X@<`Ufd=v
zBM!bIJ6lRR`UI6PFZfVP&Os;O^7Dnbvjg0cK*IBf{nO)QOQB;RnK~VssVw^USoF?;
z<q<-?A?|8qQC81Vt*G||$@OfTDh?O*94+qiyGwbFe&~C~BLN{^8EffBd%-Y`V1s9r
zxCcVM_cQqk#y0B~BmdM`FuWJ-f2OUqjuR(dVfuSjfx+8$)u6x3BlY}1Wcwfh_Qj*8
z>>8L5DAz_yMaPe0TltI+GOpZnN+{cYI*?$1Nw<2ePNMv5wxRGbyX%gg!JbCzMJti{
z0EwyId48uZv2?j?;e*_@C0j5xnGnS=A&an~Kpj3|R$?P(%o%{{p#Sp4Ivc(o`Sw~5
z5zE$%av6_XC)F3W_CYnyfJ`YAw-K3BNS(TL=4cAp@&<{qfMFRsaiWey^lpoP{$bOp
zlno^;sec2gH$ZgDZNq0GVpV+yC7c-B)pskb6P6?%L;_rd+kn|d+q;Jb;C=cvQo<hX
z$g2kvK<*C7@X889SoDG|rXt^<ET#1WF|;#=@?Va*2A*8TJ#kUnN6cf%veXjgP8#3U
zfz(PC!4+toX4&<v{|t-~UosPOME|mMDsNDq39;Q#n;aVSyq+w0#+geHiKKlC&Rz@A
zo!VbIua5n|{B;`$dmB`19%ORp$pcLSon$~g3sUWFZ>rmqwI7sW?}xl7`Uk7zh=%7k
z$4{5ixu-sk?*UQSi`vQR+I{}4K^9Y&dn?4$G<K~4`6}I6`_yU)&C0m9=CuJEvfCMA
z`v;*w(h0VRIFuP5CbD)pU2lK#2^1dJU&8U-)CZLE=Xjn)90tbx*m%bCfjTcZI%J9G
zLDXnU(n*XY2T^s;y<W*FMttSorr3)hQ47zHDoye~W;pK%D{V-}$HJl7J5VxqnqT##
zP5rJQ<qs(L@B_yf2oa*^Oq35}P{+F3$+edwA`6UGCqlGNju(04VT?es2q=Gk22DF7
zqK@(|XH2f%eM(8n3W})L^hT+3l7qpC-~ocw^nd-~Lc-8&f8(M3Tk&nboxQZQG@cF|
z1drAt2VqI2f-+R);^ja3j@`65>;Zn&5($-}<AbsyrcSr>0w;3Th3mU4=%Pk{dLC9o
zqXiSwf>lkNzPLIHRc4JcsNhCMrUv&XOdPKb_p3!KrUefLwG^&TDtUV46EnR+c2|iy
zpVWH%qDM(q!P3KkTpQ33e(O5Zb<|6V#dRv3(Gu5{Gw40nC_pWhpIP+0Hjk5ve7A~n
zz!g~G(or0FaUAsEp@!r}iu;zDnB-<{qRPoV)61O+(++2R)$W(sD1A)I<Enyrpn3$Q
zfxeCSaLs~j^TJ~uk_x7>ypg~*Sw8*`G@d8biLnmYBU4SXuNMaxkB?4IPxHvUuH%Fk
z)Tv2mTJx*h2(T40K6rrDg#|r8#2FcN1bgEYT}`!^-CVgh9Z+|OL36hGQw?T-s8KbE
zp<zfgWsduwp(}<bb{)T7iSpe7Fpq-|c7l_Yr4aVgpPvBVnGh_)KL7&<IFr8jf8w8c
z9x^bArb|)cc3kP1%ciYY;(2E2wZ5}Cxs=4?zH;EXBk*=a<4JuQxgm7`SvT6rm@KCm
zHzNO^Ip3qGWq=ie^nEJu(0@bgELVq*s-GZ^hSF7gMN$kl0(Q56rwQPIJpg%Hxs5Xi
zjCboZIgsMqj6iUQHr)R}ojfxJI*zD7HnF(J;WDr8$HL3;OHZn68bO9a;3ZO&Rhfe#
z0axmf!vPF%kgbvKOw#z+Os(7AWQ+9@gt<>END%~!s?eMA!5S4V=zc{p%UL4$3|8^$
zCP=D=$ErG<ZrjVizH~4!!3-*dg`3{LpTka>Z#jKCUe`AKE~l!fPgG)mUgoup^<;2{
zGAePzg&R-)HD2U(R&pI9P&%77gU?Lzely2^Ar8+cvx6s$6Hy+1Rho|Kn^7F&eZZY^
z`u97rxr07x2@L70WNP7DD<;9?!nH`;!j`W*J+Ntk7xOT+vp*#o5qrnnM}YcqUufg@
zVvAy;(%#6HU|f}blJ~v(SXjxR0-tC**=3z^B!OAKByKl`q%RZBD*AJej~EKZj*7^K
zcAU*T1)X+VV$Z7mVvQ^wA6^5t%B)qF)N>S7cn~13+RJHi=8+VR!c2tie*e{OeSMI%
zMt|_!wF)wz_0E~ANK7p6nPHI5(Qn1EbL%~NfDGt{osW&~0!{j9z1&4TAOY3LKgJ?~
zXMP2!D#1CmyJKNYKey6_B+m&lzml{n(rn^Bd)z#S`c$@<vl=|HqDy-6@bSY(Z*@Ki
zJTKPyzKT;ev536(`$>&%uv#qnCekY$DP?NB5O#BAQaqpWPHLb`WO7@iDS9(ph&(kp
zLQLH;bra(E<^^YjAK@M0Cl%XsXXRc~#EpaGX+||T*1IW8QW-c`mtCj#Wm8Qtt~(Iq
z82fXhL>|MDz)T^)5LpKtP-mqNxT{)Lm9EQrDm*dF7;L20&vqir`1UmpQzlbMNf7!b
zMg!t%9~3jOk}0F87dVuGL##^k`Z4@^pPu8@Rv4PqcHI`>BqkTrA0(VSgbHQiLQ)aW
zNi+QW=HRgd<6pDUq*$rGra`*8y5@QYo$Vl}lIV{ypW>RQFsWI^0(kv;yrmKVeL8{v
zo~a#asZVRuC_@ljA&uv{eypb^8uyiOGU}yKCSydYv%&Sp^GZwa?756W60fTALkn(?
zs`+Gpx%sUF5{^Vus<Up?92C5_6rBbo&08FA6xRkHe#BIJh0$e~S-9ocyqiwS-+Wse
z&83$`IuNx-nbuk(p_72KOrH1fn+7%EL86zcJx&K*L)kK!gtmgLrm}*oP84fLj)^+5
z+{8MVSW_t-eaG<}yjPeR5OjSy(S)Q-Vw6j~jViuf4b4Ec3VwM?5VkCM3rPPzmD$<Q
z!nIEF%t9UH?7bI&F)IU}3uD@GZoDx?+Tc<3X^LIBIHL%(dmxM60xk#j$oQhT4#WpS
zxc}Ch8<(PfQbZQ@&0J~uf_+zQ*nWU>pv4A)%0fH|%!IR3>%ntZ(@|<~`^SD-PD`gt
z=nALn78M?QL*+&b9^>*u#ZlPSgue0{6Ha?ZN_k@`maOF#3vaw0@(Y|8o*I;&L;ZlL
z{1#~jJdy>|AHmT54;bQjaB7*QI6xjBy!$(m5;>WV_avV@A21@)Am_hA0zNP$s;wTL
znQBpO!f&c)p9(sZ0f@;`_@7M!C|)y|pp-B|DDkD@@`s7MnY<Xqz3BP;XODUBJV=(h
zHF+7d*n;z16}X<(+8lR%F#s$CN81`Cx4LtM%cm+*tSU}-S78xg`d}=B>}DS9?{fyo
zom)^Mvk;v|TtCPw^h+_x1)6%ReZVzvOTL@dT2Y3Nz!QgFzNChXCNAvj*rk<f-jkXu
zH7>@CS>09DAoXK(8?6T@g6!4dn8`Dslm{?KG3AsQs7)ziyeeA@ZDy>M$2jRLUA(06
z`~IT#hQAVIThfEh`?Yp<i~Y<SL$&)+1GtF0W~9wP(SKXn01A^F7r|2|z>fc_8)4sA
zkuZ~^4O`YgU#wpZmHBq;wez_aRMH&`zPV58DTuQS8OXkDA*k^kxpRI64K7ci<+!eH
zUIMZT8N}~G`2aer0@0#KT`cJ3b?ouozFnQNPRKo(7nz-=H~DL1j{*xONDY<DC1Mk~
z*vRf@%1^0pFKHO;sY&9Y>;aX0I_DpJvCl@&rTFR)S>CRoXtCv-GrOjEl)G{XhMe0>
zH$-m+DK~>mE09Ftf2tXfvhTX_`Q}l|ytYqm%97{;d^-{ExGOLFQ+qRTk}G|Ln*~H#
zRGNB*m7sqr@@hhTp``+o0qs^)qQ&<W4@y)bUej50(I&W&n*yZP{4T4H3Hk8T>Q-DA
ziTMJkf^t-6f`XK#`sZc}zvNnVDwVaeb~1c<Jv)OTyLvc^)xS2u3ZT!}6BBCpHsYUd
zlb=0b&>P^!mLw=tXDVm=l&?UA!Zu&CxfbB(PDwpgjp8l-9pW{axwGJX;Agqbg2PhI
z7iWcL%+g#WR%T}^=2m1$AG|(i{%O!J#n}_k3LwtCj+@Xk395+YkX`lYCzsa%>XfH-
z<7;*JI}*P2HQ2<A+=ONQ>}lr4pcCndE>6)pN8BqVQJW>OK6%S|Sx`Wt2r8jj3vDC;
zcL-LfZT*B_$HnbEdVktJKcjYmcjPt{q6dB(LJFoy3^>FZ=4UL0c?g=^Sl0VIUOy|m
z?Hayk<nVnkO?ZWq2~-)IZA5Qsp@H_~n8ne32%JIr^2r~%YBL5nCtr98>y3by0Zu9R
zwpGnC?PYAx7rd0xbP*I|2S@GmmiPHItk?-pIVrw_73h*X!u-9Y1NP3r!TAya_8qVy
z(+y8z_>^0_%Ag4#1gfuzuHkrOspuYzq<MZS#905)oz-VE@mByv{fd+=8>b}JV=@vT
z(5)S1%dGQ&BnXTyFTVmeoQ4`k&{apM4$%(j(oKt%-tRnW^y*RX8biw+3l;82oj(4>
zLrS@&rkcz8#i|x?>2jef;-Re9r%5->e-$;>ijrjGZ*leFa^B{xjh2cV>5nJC(i7EW
zz*TeSf}(43XQ)i;DEz{=O~8#Koy-M%$kN#Yp&*};DBSLyA?<8>umor@khgJ5NrsQK
zYQ2#FU`1}fvQn3Gb}Jet`r?uP1j)FZny14OV)sXQ*1E5Ja<q4IU~n)UNQtf{A`|mf
z{^9JA=b$2IWCAz^Z1}dNf<AZbQ6PPO$T38nZR~y!$n)29RGG^yUem1eCT$ap+_Fez
zWt+I;B?2j73-Pdo*zZEc?rh~5{GwzsY=seu%A!h&kBp|zt*?>i)v`7d<MTacjQ%^@
zO#As?2vh@uRixjDd5wu;hbzrdE6Z>i^?4q9$sDV`=~{fsFn-OMit6uszp@m(x%y^o
zzgv|t#3(on@5|TfR3*CTdU=ba<HXU$<KANklPO)cosXQpPfIlWclfswUNIgnzgme`
zjwxq!l!T_4Q>8Aj?pchHAs>>W%2+7z?v-=OoLI`i%f}+jg0TBC#8LT+M$hV#c+RHW
z6h5=035fPPJ8_P6A}+FdZD=%m&}_N}WYY49u@N<FeT9O~U!$F3m>%AEuJXP)TKXqp
zB<UA|e|^FBc95s!R`1cr7)E#dneht$*btd4sg#+1N)#O%|Ne<q*AKknlZN%0v5aw@
zPLCcciMtK2yJVX$iW;RXhR+QR;n|e?wo4zwsuo|>WoDdJaVQ`61n$KUH2&*ck#bv`
z+OwuAzj^A-YPCJ+N4j&#)RMg$U&nWoLr+?=WsqGFL5<LX%q0c?n<FeGgpSo6%p><7
zU*h%uV6pkQ!A8kvNRk>V8D{?%U2U|(;0QcVE#6LO0NL-Gsus<<Bg+JG$r0cr60J^5
z0@<&(jVM?h8P8Rs%CFBN*}{UkXHw$+8BJ1?mv!ox=1_Y|n%Ku}UcNXRC)KBb-G^Lv
zkdSl6ypIB@ni-;xJpBrY<-`Xf+%I~QjTD(fL)0J*2hJ$--5QfsmcqZ9ZgHnIk7F}u
zdlXBs)f^qf`TRl%93OPrJFpaj^|jfR@oK`<`seFxZ<v3(N<>pP!iQPO$F^`I-Tg1v
zMqt0v@jIzgI;TET*SwTIizYWE@6`SB`j4W%=LE3yPYdWc{}f?7n#N|7koP01lFLH8
ztSBN5tmv>fu4MiUE}txIPHBxl3{PuoYX{|fq!;qbHw@eistA=TcLD{SKU=gjUjzmJ
z2)XPKIDj8!$e}*HHp54Lur@gS!J0~^9+#-%_xps75xbrnOxZ~)ydvKelcPJndqqW4
zl>RRKc|5AZ(2m|S`yQO?-$zaGyNK^!Dw}h88-W0^R}iuai(jL=dMyNqQMc8^a8Ptk
z69<M$Ib~Wq6AYgS8_8`?Lm5pgC`8>AOFlN!Tan~Uew_iHS&ag3Kfk8G=D2itKsUhu
zPx!*5(^!Xt({1(K2f-cUU@zKY>6yCL0VWh+kV<7@KddsQUc2z1ec!8i>5x<MH&&o!
z)j&fb+`i;rrS!GkMJq)mt^YvYu4Md1D`DcK!(T5KyLpkRl~9>y{WzL;>N8@XvU%nX
zGUT3>YanV({l`HDLHlrSfj4uRjRk^q$acVOHEE8M2>%TaHKE_yGOBTS3(lzV!k{i~
zKo|72cBP3MxH;8s{7^$|B44w);D-2I9iO5U=RQCk#<^)3Z>%dd6-X@PQ8wXa8RTQO
zi?(?&XHU}7$<4Ey-HguX<^=MZ_XjLG1)7?SO~yLUlH{(N>zUnz8f-0TJ03$}1S1V|
zGs<8n>`4vyd?-hrda4|dw?yGtD(Wze5MII>`tdD?q{a9zWvSK~d<T-9jd}TNa@+na
z-uZ`Jy@ZrhJcwuUGDEb+vNS~oCSGw4PvahsQKsBC*!?*>I?I)-H4RQMU}ip%b)z7y
zzL(q_Sa$u(a-4yF|Hbfc7t>H(6ZnMa&OLan#+0w>dV|FMjoe}J5=78+Rgfr+%%wDu
zftVhV4<KMPuz>sKR-lG;Q3kX85O|j8ed^a}RPHBw=W9*go4OZ#@q6Xr2fio`!)!IK
z#>U;DZ^^V|q$dZ)n~OoTt^Qm6@9`v`#^!3@zp5?THNr_l8zIxNLR{TzO+FVlvu`5X
zs!qIL)99XnYU%Ukomx`TJAF?5&#Yo<gxcS6J+JxC=6QT_xaUB^5HDK7AQRA7X41e&
zKq;G8&nw=T)+>M1esg^x1e)qf<*pn(i^h>IJ;H`G3Jcz?K<bfQWZ5mhD%HeMOtW)W
z!P0)=-d@m^R6(gp$IP7Q2@nk8iRS7IPJU_LfeatmI15?Y8(V-F{me}xfsA|DV?|2~
zCZ&Yf;}qJ_f&Hp5RS}Pkex<9GMIJ?z(w_51sjiWL<bfpg*->20;-orH(1U<?CT+2?
z6L+OS5p~tP<*T~Sol3{rxc7dC$aygpCV(^zr_XZJ)(K-^09w47UrboMxmrxhzq#D_
z{{`cl*anRyd(W;!Q`t9R)^z>S$^+1E+W2V}O)A80NEED-pJvmgkMlhm|C85ggVZ-a
z5qWxWH@VsilVf2CJs-<2Ay0nB#O|WRNP1Ro3W4zGl9vz4cuq!43Ky@}8SmqB_k(qF
zkW!q}@N=n^0#{d!l{lSeM>GDb+t*PeQ7w)?m>!ZqT6Uu}M^wGNze=}EIx}LUdv!yk
z?7!TYRCe8}AMliEYjj6Phjs7ON|Ib+{?#A*TaT8J{8mVB`P8viw(OMg03SCm$vrx4
z8+o-e3*=a%U5ZmqQrh#h)QLLXH#L*n|DT?N>ZbL(9lSr6)I^xmz1HDKyRG*FR~m1Q
z9Z~JHmOF}HrC#x#4Zh<`+v-;3PMV2{5Mjsc7{<$9s5C`S!hP-G-#?Mn-bj^y*faVM
zeoeaL>(tdcI&on{D3E3$%|yfAg61*t{X}GY2xMa*jd?6*_4axQ2G^22g)DiKXEFUY
z0i^_<!G|NNuq6S4fsO$_QfL)!Fq{Z+klf=n_=!7CR{Z(V!khawJ`mP<`(Ixgt2TEA
zXJ%v8r+iIkibgiA3iKV`X_VIf;2VCApF_@8|IUaE5GvtLR4J3QzZ5BkNqR<%1xO!c
zl(LG`=w+%aYG!d_bmG+HCv`9fp6Hv~K@#leL4)6~iPmC;mkt~5t2*noZHGP8x+Y%+
z#^SKNHPZHGGHRaS>}>->mP9k*s4>x|0liBr`+VW8bGf(dmy%(VStZ8N7KI>ne4hDo
z*I>yi#(L##1I!N9Nn$glyhvd%;J>tz%Fk*qAe>@cP1GW|D>ZJCcPiT#2KbA1cFVWu
zjF@!etV`SRY*URF7UYldiC*~}&sN)6&sS-hT*t^=zEY>%3leNql^fE>@zZQE_VhY|
zCc)wG*b|Pc{f4m2p5Pps;nrkhMz0lr@E}coNd%clSMxIy!36r}RTU|XwL9&FzY7`N
zL?5Quswhb%Ii(L+MtA*w?{9vwwYjY*h*=WQYeLp!qnS+d_y>-(-6p<M#({jY+uAVm
z8HxH+>i=qFG!HJUasN9OD^I+owL!rc*a?BHUQp_S4VKwsAl=c;{10%=0O0~bkUaaJ
zXlyk~+~1R_A5!Bqr4qwx#1O$U-h{2cUa!X!ru?oo%zIHfXfh`1_=mgjqKdY$Uf0um
z!4+7PPT_a`c(b&%@LQebRQ_xIq*0Qeo7)36L5Tb&>tsuaLbZBDJFSzmE0Y*|L-_1(
zpR!uHB40<Z363apa!X)ji3jfHCSiw*naqaogNe-|_@sWctJ2C=whnKyS}Q&qtQXZ+
z6Nk?cLuquK4nZKb*WZ8Q(;U~@rffwnyLBNUnse<U?p(oM2Hz2Qs8|2-CJP++{04wG
zoDO}mt9QbVaZ`9gX?|fP_QfW#_wrVWd@IXq^WQK0GsLbR;Dgpzv{0a(2`UQCfwtLA
z<4*)EI=N-){w;CDp^Y=UX>-Y3j3Pqn6cSF(g?|L-|7}K@yKAs;q@clsSU`<f8mLq{
zCZaD&5JRX(_&iK+#mQbu`wd~LIQ~4mdH;)6ofF~mzXsX<7T&^K{(%5V2vhtn7PGWv
z9f*#ABSWvG_`?7~Y~Lo3^kqD~h2@Wy9cc1n@A!U^g7BTjo1JFJsHz3jp!p*Pr*}?m
zYkM_bSJ_SS&jRhqi<zb5<dXJ+H=eE4KT^CHMn3{Oom|(k?Q&!dBH8l9D4xbgpf40O
znXb8aY3U5P6jr?bQ&#hjlbo<)c$;XkeXM~f$m}pVya)yY&D?XJ?D$M9RQG7LU<4FV
z8xnub9^C+qc@IqQ!spYMMEt)8Y#yO6#z>GLm=LaCg_9;OU*k?o3qnYND#==BZG%%V
zrcLwpsEKJ%z1LU##yH-!_y|a8`j;P{Dy?{k3rUcDesb5RE6pIZ3jh*5fUl@@rS(Q3
zbk}A3clFd~QnF!Bo71@;IZgh+oW5+v>_1fg{9qRlS&Xl@3^EsDsuJ2vSxpla`BcON
zT7V63MB{?$JqTI=Rb~?{lcu`ScXD$|5T!4#t~m}H-C{_Rw$^;FtyQm_zH@M&tO|*^
zIJ>_2W*9?%As)r8yL()3333G5?LCzUUnH(6p;u?StuyS3x|Z0y8TF5vUzbB4X6#Lj
za=J2NmY$v{PuDPq_<F&X%UHH&lLLhp_Eod<rQa&rZs{@XYJ1vzI(ntHNH6^Fy!LZW
z33vFiz#0fY%DRf-N%^sLBEv{}^09jGzOC}g>1{3Jb?@RSZTVmd_OB2I21drdEbGwe
z*<PI<dFX3)$h{V|h)nLtoMZ%6e--zv0qx_3N?7zo<N`<^f=nVqiS~2r<!6*gTJvi0
z6qRP>jypsK(-Ou=u-A9~-gG#^pkFXex|g8|!Z7pJMb*6FE{hz2s2pk~1S(x{Cu)G?
zw}WVWmZ0_j^AUn1jUZw0mY9}kS_Ds|=pVg@CMRdxinVdU8%)o1AM`e=?c3&4m!Yfo
zk|Yeu<>#RI&!;1ULmi%I%$_(q_Gj0TOA4VOSlQ@GS@-vk=(y1{|Fn6CGlqwS2&-#3
zVrn*9)h^RV*CKaPWlSG`H1%NoAY2lV_i<~?pXcS(@6!B|qlfen%vU?}Ce~w#;{|P|
z&iy1?L5tW65egLNga^dbms<BK7YNZFPyE>mfNP$kA}s_1!LN$If5Y{MGLWS9Q&wTx
z+|?O~?>XdVu(F6awnVK#>#m{Qp08W{sJ#BbyDvLRv_vYaS*coWB!iM-<4}T{c#}YR
zJ5OXq&#-z%Rh8759uf)ybG(IVyisQy8n4M|%v9F#9k4dtc|>CRLEPo|vVvmqD5dD&
z<Dc8oO;DBWJ7U@w(s6t<d%+dTBhEf5Un9nWTy1~9kFl>bKQsqlIN~#svv;sYe*L3#
zG^VW#vZ{Sv-Vu^yxC61G-AX(}?}Svq0_|h3Kz}r#dS}e`Z!^kTc}Y)9@g}!t1s|TT
zo-{TKjk<?u*CeSH1L&*f`PA=^Q(OXyXLbcMLD_|6ju&Rs*%PL;Q?iiqDr-nkal*V`
zLYe)nZH>J4z)^~Mu52oHruD5$6~tnIKBZgy@wRdWI00Sn0vO-~8NOW>FQD6f_0`YS
z@UL`4l_lwkT=!0yJ6CzGCH{YJwn{5~xI3!&pR)B{WDBFtOB?0Ft}hpk1e&;4Dx3Q1
z(~o3q24v3G+)NUrC}YAfLz7{U6u(Hl4;~**gf;adlFZF_hVkGF&__vY%5!pIBN?om
z;@FnMK&X$&8y_vFj}J5-r`nId{Ouez+3k&?TC1t8UHLaGmjr~amFSA+b6RTXyqRJE
zf7sw|2PCbWE<TOoleHmV?1O#K4bedV<39ftEtfSvSXSAFY^)!unM<K@hcFD2jQEB7
zY_*t7a&3o{CMy$q82gV>@;}Vpd7!vwms$Klf|-6OsGVq%n=d3b8Y@p`Z_S}L>Ks~h
z-#9w!^U=VpnwM#TF5YQOiqoF@^N`uL`&`dbsx-?r1EetheD)KZI(Nc)GrXys^lk!t
ztFjDNe13ZQ=`WtT#eamK&%X-@QJl>ol@p{URrWkR3C~z5eCa~<Q0i^^;M;MX$l_hB
zIw1;t?Lgs3as!1OtmHGd>~F{VcKWR(X?$D2Jpy`^`T)wri&F>G`9tDAXO6_uAvC$t
zji5Re;IYM(_FX`K(0F7kOQdi118>BijQ>f0m2m5f>RD>xHjQrHX{Jop5p(^jBFB0%
zUaLB7OP*);q7uT^w#DMFxuBHOThW&JNC}Woad_zMp%HA4I(rtVy0=Mhd+1wQ(N^Sl
zn0S4Yy1<<XzF8RQJ=F31|Ifrq{Nh*W+l&0Z3bh;3fj@6AO8A|N@xT2~SFIby@|O*q
UmlIb~cfgmrvW`-X!pryn0~poYegFUf

literal 0
HcmV?d00001

diff --git a/_images/35c10296694a5eb3635e87b56ab9545d9a293bed86bd144682549fe750d7a34d.png b/_images/35c10296694a5eb3635e87b56ab9545d9a293bed86bd144682549fe750d7a34d.png
new file mode 100644
index 0000000000000000000000000000000000000000..e547d8996d1b688ea46432650d8a1162aa9eb1ca
GIT binary patch
literal 52093
zcmZ^~1yojT)Gdl47^EObgLHSNba!`mcY{hxNq2vgba!`mcXxN+hwnRQ{O6uKuESS+
z*}Qu{v7R~SnrjgxBP9&?3F8w41O%L@h=3df#G54W%J={czEk6h&H%o=v*#04_yGQR
zd@u+E??2jzsM<q7z#zT8-i$AKJ%VqtItZ#b$XgpZIP2OOLRjiL*jQLQSeWYJIT_m7
zn_62jP}5V>Q{tI8IM}e$(EQJJYHK@VnpoVBAP5LN2vGrE1()Q*C1;JPtLN6U8T7*@
zmS6a+Ht9daj*}CZB`GXbYaFSHteCBy_LY=s>SHVy48;d1)y?Vzqt_s8@!b2h-oqPv
zw6yU2iR%5r_UqhrIe=x@xjFpp$;F)qPuh*}w}J8yybr(d;v>p)Aras4B2s(e&w=+J
zj7TCw!37<s|KGQx2a+l5G0rQVHMt$TQ|IUBA7EezzJ1Hc<^Sxfl<&?Ck#X_~2FBdE
zB`pnCy~e8P`<O$CukSmn<~yV3`vaQCrzdxIEZ$vP@Bl8Oo8zUN5(F&P*70odyR#nL
zv(r<E@7Zz*SI_Jz@Z#d)70X!!ZcR8aFu^#XV36;|LWwWY5D*Y1a7jo=AT%^JB^O@*
z>Dj+>OZ9$s&CJRQ3JTKQ>W@K#^*4IZ_PkOgmWcoE;=-z|VGD)HVrAjfQd&yee7g>V
z<8`Ym6pk~L70&&s^`N9>_;9WQolfsv6tU#`UT#`WPENPOGQ$XG$7H^IesM9grTgPq
zPXe1gF-B-k*F*onz{YlxHT3HF01XKxWk^@>SIE;@{czg_%bK&TIBj~zeFc9wOn;>U
z!;w@jO0CBD-CdK{$7KozvwGaxR2-NwVraRnK@m$P3V`9s=Spoo-K<z&oH$?YPYDPM
zBje&CFjiNiQfr6^2)z6J`SV~BOSoo(Q`H3x;fEy=R9`JTYHBHXIJg)X7#Q`#BznW&
zSm@|v-luKv?Tn0a9>I+I`Tb!dW{Ne<*|)Rfv?8;zvbI*d9wdi*d;hGspK!5JQc8q{
zhB9FzV_=XA3kwfud);Lxr=(CB8yWSNQx2RoUW8%c;?nGFZ}*p#m7NbJFl-6qdQLi|
zDiy?(G@gsO?u}=IpF1Dnd7(tWW>?_0+Y~7{>C&O?^?ts~#R7N5Wp@yvrlyXiASX9a
zjom__Q0fIUIrqF-BEjWyQ{Z+wjJKdg+Z;+_8EARFFGx;JrA}kD<?~t(BH7H#Bdu#}
z)QB}yjU^P-sx*n<1WW2V5KCnzB^B1_{o-}kaMBzFuG-t%Z;g)1uyJ#1F=K?Ddp}>`
zx<1~XotNqNfQK43Y0(G^3ieSb7e_NNFl-GDO1NHZ57~KmcqqU0y<1BTG$6SAgiVwL
z?$$6@O(Y)MI}k?`2kv{YzrUHmAJA_;U1;hw_&R19t;X3)c1A`8FAxeGA%@^S=c5_?
zhFN1K^G@fRqOP5RpU(?Qn#qL)1bWjvuT>jn0>i`O`1$#_78e(d7L*HBzJ=p(sTCI$
zS>AYLjgblq3H9?lT}Dw3lw)CHQUv(>_Xi;0T3y}d7$Af9G@A7asZ8dv$!TfU^V+Ja
zWUoJ9YioP7tc*A)$XW53`jgLl1gq=$iHWGjMox(Gii)v=D*-`4CJth9^50Vh$}9`a
z9=E5cGs}&x3~AoaY;$vSCX5(_qI@MY@^W&&;^MwSKtjW=jtS##%^9Z|u6G74x*Zg0
zdeMOIwhj*$&s8rjYA6+{b+!4tC!?U?k2TysI3Ob>g}@^sf&wQHg!uGv6|VQAZDCQ-
z9XJp+wzhPZ&USWs&yRPMs^z|56Ogd6hbm1cc2`=w<5mLjl3~sUg=$HYpAa=OWbm9b
zbm~GwB`anE;IY@+-s2L|(?>cly9Qhx%v>B+jJypU`8oUr7dKoiJ2NjY(BI$RzS&uw
z4r0-MgzF0`>iOk_bcNHA8neahkEA5^U^I4m({VxoNot;`7mpVs6<$wn)D4I9*C#6@
zi*@#Q*9+DSkKFIpG{B~7T;5+D1Y~Dt@A0}s-QK!D!(+`J*8%jXt*tHfxU&-qLF2Wt
zp%oMqqzVHth7DeZLy1ga`IL$^=8hecgz7{iCv_iDP*9RNT_YQg>w`Zc6C)xa{fdb}
zF*G#H(QohQ$p8BnpN*|<$$Th<BQ8$cD}f}{<ri44vaV+kvQ{lmPTcpq1jP-<?9;`X
z3%VY^MzIB05^csR%F1-l*1HPTsy-<zD?`I!bmmAV-GVcH3ue5)#lp%Oe|~N^U8LSK
zTc$56A+ZTASvg4v2|<H6-wYy2)tStbZM>d0r&Oy9Jlp7D@VviNDmgef2$st?=n2Et
zu<b%oQBzxAT^-#olai9Et*<Y)T4H|-37KD5xEZ6Y)iysr?=T2%9uyK{aCg4Nc|EJ|
z>*pt{q4D~Vi=7c`rzST0?R2FA*UboFS~S?fG@i7>LklpCfoL-M)3dXT+S>TeP62;d
z^znoiJziF8fcJ5}TLW=f3goJFc0^#^-a$j3_fwS6S}oQ(jQ|AubhBuWP}1_u?&#<U
zqWEk#EBySl0|7GB-WJDg_ZvX+)?ZLC5P^Y#IskeeA68z7IXMADrJ>F^Iy>)V1mNfa
z@GR7*i|p+cX?{E(P|}wnk7)#P$@%hl0RR+1TU-0}PFjE$Cg?u-{X*cr{V}8MonoBo
z>fhfl#>vH{T5F?QFp~^$-mEI|b+i+E@!X#5;NWmxj`Byx#cdoO4sP~Esg~&p=01Bt
zt$C1b7CU&r0P1XKLmaUHRQj$0xlagiDlJFEe{VT<-n>;Sij9u`!pX@ws6{FZF4nT5
zg$jv@iG4soux5@d?_(oL0@vjuc){m+StpdWJp~O641$k#k}VSZTfDEPlwI82)4UB{
zpCecT&O(xP^ISB+u+wl~PX58*!h$;GCn7-Lj1$uvD0JVy|DVtQzlMYC#|p)}JMp)m
zfQ*`&d#B3h9-o7Q1Mspe`AAa={wx1`Dmgkb<w~WrkBlH-Vq$*zBq$v)X)YQm;N`^)
zj*N?&Tk6GKCZ;5!FLyz0EgBCGkA<bBVCLQrx%|$7f#AP?{}!jL44r}TXfR&7c@_*f
zz!X)KmbU89I_zn@uFCF&Kyc2*@Xm=CAtsn{+o3@DXRs(DyAJHJG}=-Q4vc^(K){t*
zuQaE^7s&sw%NFyMLI3<iSZ%e`{}b~4@Pivz*4#9&#8+|+#po(G9LOsyJoFb)M|$0X
zHT;OAByk-byzA@hCA+2=3gyVi$oB3~%(-eyOb}y*DrK4K#h%RnZWHxeY3@71ljaBN
zj~_n@2ng7`Jl!N+qlXF+OC|S$&m^T>bYohVx}2>mmE=yD&6Vmxg6P51($ZpOV=L3^
z`c!SbBE`$gThl{B?D**yUUEhu-p}#JTNpbU=N8gM4|QtB>BO&EE^n-@tzYp7puxxQ
zDk^BBD$_6im{vMR^HqZ>9A8Z&zVkeuLWXl)s}#%(^!I0GWc<|a@K30QL!*TNTM>rC
z^}D3x<gT#_Rbb$Cj!3lL&mQnttly1{h#?lLEjx5*X<tu!cu7$aRBKyX#+=jp_wOf~
zj|tV7LGT#cT!bDvm}jw^ouBLVN0Sw5HpBv6E(NE*xNaMRl8Q=HL}V2p5<T|AS^#ch
z>ciyZ<m&#u*}Ss6{DkF*@9=;3i3`NEZj6A#Wd0Ks)p5MkpjxW)1|lLN0*lqA^^ydj
zczQZsVPT<&s3=`ZT5774rze*WG~E96?0-8c5Q5194@5^SmAa6E!k4G}EAtB#b93|J
z=6f^1H-dmh6crU40X#?~GJfRYX|W2Ie2o-eabLa+owi?4SZsEklYdP|kfk}>R@;0M
z7>wcNb0pgO`+vVY-?SJ7NdkI&4Jb;@`pUV3jh^rtR$@Fn2o@F=fTYuFSAefvkLz|^
zz^k6ByE-ryM14CEU)I<uP@(n{k}SHqx?>Y-j0CU8p`AzvK!0avr;fS#0JifYxKN2#
zKlUn{-vpyl>5XLya$yT4{ktLmbL{``WB30Cg@3QE;b%<jsWf}Ajg5_gm_tNG?QZsb
zaAPl&`utx2mzhowgP@(Qccgd^Hu>bh^Q`OZE7pD{5TbS5pR@`rwftA9YY^Jm+kb?G
zJ*>Vm0x(v0G0ai7lj>f-<#zNL<sBI5a47*pm}NrQ*G(<?U;6Q1IIHD*xbHTj9s#2O
zD8k3b4-5*TJ7T~Hl}MoX^L{$e&N0k*v+S|~hlGqQIUtf|IbEm<QK;SW%(*PM^WS<2
zuKIn|T0L$!VW#rE9_BcSzs^@6x2j!(1~gzkZ^Vdk2SP{%AgZ+Q!Y`B8{32Ob?sQhF
zhS@FZ?qcWS`QaogAwd+-?EeIADm)w_)xSIaDo?<k(p6c#5*^cwhEH}j(SL_C{+W=F
zP~-IWe;onajZl_XrmzQJ%iV9RQeD<RLdegkOaBK0NOV?<tXSCCs?`?gva+(52Qwv0
z8*tE&LRp`E+I~<DFfuYi=$M)Y@n!5z&jSAaHxe>rN_E=ahGDU_1I}URd~Ui~*1tVn
z<Gh-b|MKMvEzt1(2)I&SUS2gl#Q&bGuR+=3am?<XKzpM0ynFX<b$h!T=pJ^*riv0%
z>P$m|55H`JMokz!fOuSVn2=OaQMq(r9RgAWT==L`VFX7<M>jh^Z`XtE3@%UwS~0%H
zdn?EoR4Pqmr31!@#{fuXNJkmg<?`2>B#*4bh=3z~sfvn<Iyp0Q<AC-Ru1PCgZK(>L
z?q7_p0D;gqGLl<dELCa4jwg!ris*fP`Qkgl#{YgWEG!HHOeTHiz@0U5d1QDP-r4zv
zI)&n23P=PA1KHk(ZGwLT{r^L1|3`GM;mZ^ityOUq+-=>aabf?)a|l8UCH~N)AuCBY
zV(A?fS#Ni}g{>nz&E0(~tYfU}6Oe)#;nm^JF2K9{2TWl!7-c5*{IHs3RMhPl%ah=J
zu$(Bxkd9@)&;*^e))^%?d@2`OHlBk7K!DwB*vN>l&P_hkv#l?ZT2g<4c^LSJ-=?$d
zvuPy`dsnyg{cQi@ft7P@Q~7I@8%Rw=xBMXz=Rew-olj#mh@0yx7HNL{zcNyF^J2{4
zzhWDjcjL%f3GPn4hLRh)Xl!o*xxrF|_D5T|_O9*(kYK>t92u*9)hd{YR7&;ZrT?Q4
z&nKt*NwA8DRc#J()$*k6!^t<5Jhz#*YJ4k1VC639PmW$9u~ls$k|{^l&!}kLoX+|_
zpl@MHxoQ%P9Jxe2?q?e9ShCjV&mBOm+1EMsTFC!G@=YZKpN*Hb^znbXd*7Eo*h4C=
z!E*U2xd`*X2Af8bg!=_?v{p()(#M~jpQ}fl0c~|Y{PNr)xeyBtllHeHiSy}ru`}8(
zs)U*<F}Yu);YREq&4iI~;TN4!;SW7x;}Lfa6h{rwp&oJN5hEp@*qAT}IIb_hMg?9M
zl^A`CO~I_drnEe_vN*Y^PLyyHw;K;#qQ`>0&gbkY)3(BPJ6As`DWyc|uVjysX~COX
z&&jv=PySglq@+a!cn$OjKS*05I^PK)A*1z$c9gvRZaf*K@JB+)kU9v8%s_ej^AQV@
z_*e|_<y!z(d@QaO1HnUD3#zkt19HI|PR1wJSC0%)NUc@;Mhw(CUw|~@5Tb%eZ)}b<
zY~kv-^^&C;u2E2>W!Cwk4=Ge;GpnfSiTO(h^}l=Rt;>0FoqY}XE%#=07gPBM))7ZY
z0rXI(T(*$;sH}pzM0R!JJT%-Jsb*GHj#K@8Tr8Ra6Q`yuB2$Zz#evOL+ngQhV&-8=
z^eRM{nA<TFYU*_kCB*tZk+g3im*{W5^F4o~C%3byveh-Sffd_Ysetu&fw!<dHhw?7
zcCad#r|wNi82yPTVT-Nt>k6yNDsBYJ+QBYd>zcf9(17H+pDBeJY!IsN@~hUN^$}_L
zEN~yXJ6T4S8z!%y-Y(fuvc2ENAh|%Ks%I`lsg12RLXf1@%9ov^RkGDoM<L4Qm<$4b
zXao6~3@OY;7BPw)D&u_$>CkEn*U84t?oj3PM$(Zyd9(y3ZR(8i3+~u6LH>;S8=Eqx
z^*&K?+04!Sq7rbrX71K{t+_ro<2fXPGiUR&uUtY&5QhhMp`Bqtt7okLCkY8v*K&XH
z)JBSBvI?`+M<-Nf2+|m_w<%99Uh^$zJ3POUB+<>oANhQEg?)IPzpa3Sp13Q7><c4e
zJbLE!h~c!;35j@&-hCd=Vav*O(ZZ#R^G#t=F;S~wuHI%PUVS#TF3$PJyKNm8=iTa3
z9nGrMfmtdG-UrRs(AR))o$|9_tq-a!m-uZa7SH1Vk9;q<VR$mOhyD!b@2|8oSUb<k
z;{G(BI@QSfRAv>Vj#wWIW};(cTPONRbIlL7K==5m&T#W8<Qm&aYKB$AA$oG1(Fh?K
z_J=IJT_dB9T8jVdqJV-J3d4Je&Y7_!JYzbY@4KF_l7wMC3V9^So#zdPa2yxIRtEJH
z`NRDwbR#yW3hN^|eSgIr-ZKR~g}vm1Cl+U&<s{q^Y7L$P3&7bnSb#{1wrPlQNs|^2
znYcKe9n&b;d~+&263}1L<?>`G6!X3e^&I{aW6b3Kv>6-MS47=8Hy`nkWlhxD7bl;H
z#Cfzh^cePT?eUchmR1!kKPI5mDNHJtF_`1-q0HrY@^kSgj9SAg&YS(MoQj@SPG^1M
z`SJr+XfBYK9=}hBP?flWK78MQ-st9o!+A$ncaUhCeqG~>%OR!Zo81!(mv1Y70Vcoh
zda9fh_C_6cff}XOwvOS5eH>5wbED6r+oSaP-{4nBk!@ZJIsZt~O9YXfxpn9R=9P#P
zs+Rv6p>^LL`(e(t-nD6f1@pbINvHMu>fAiB-|W^D3r;EI_jq;hZMZ}!Gf@n4c`sed
zewrgKXWtvd<FH0_XAU}F)mp~a38p9%6@4qiLl12ZE3X#=2Uj^*pHdNZ5Q7zJYmXPA
z8kITZD!g@}ZNgy?*L2~IaIeAXglK@iC1v(^l^+wcnR(%4EE*BPC%?DC6z=}+a8~w6
zd3cvXuTVguU~cYWc2MwT`b$my%5=Mb0L{bujz;JzckH5!pOU5=HWiL*DQqKAYLVUj
zDc1UbU@5Vf5zMPl-0h=!^ZR#PRH0uZtz-kBI)gk;`_y#Z)YGrgg@1>As&*I<J*`MP
z?~E1%)t7SwMq(5owbnm6z%ey`%V<gk#guz{v-uvRoWAaUW!v>EW8%<YtPmQX*{2*D
z2}Mmp;TX{z@}hwkhF^gnH>1z9zE@OI?y~79a2INFH5ZO|N&Gqs3Io+`k+YBIe__a`
zqirh*;aa0jGzwGVIZQC3e7YU$In14%I2Y*k0do`@QQ>9F1;vb)PuWL{5mk#}T1d|b
zLZa<4M0-nn{iwO|>^65-ni!c&156WiiMc#gL-C)*n=BYfG>xS(LQ}*U^~Po6xACdt
zDSNIt>QwBhZu^XH9fHpD=moAS^C(sP%aB_l7u&~2n@yJIh6k=<Z2DlRwTpU2Qw799
zAj;PwmGX;>9NT9AXV6=^yARvD%-3*#9M@qywR(&zqF#Dzom~j}7{a+jP2n@vnpB3H
zr9kRpbBLW6{ny?d*;0l~Q34i&UqXivm&mK*ev*fDEZW|BQg2d^NIE0#^GZZJQFdE8
zK?Qz5`TZv`CX`1fJqnb0@(C|xOyfhUm*Z>lA_`;UX>JQDi%|@WTjIN1)C)u6_u(73
zS=RgKG^Zs|-uO-UNy5|(3zYl`Q(`F2!pkAKDVQt7FA}PnP|kwOzAJk%Tu>P-h<!q}
zguUewAb?MfUw$D;swVX`F&~goPQ@yi7Tm3`A{?4!LroZr1)A5=JE&C<rHm-b@qTK-
zdWk~+h8%Q_JGXV<Adn$rbM$AfcA@@UcAbUuvBzHGW0<bjZ8EB222=(K2qg`NznEro
zsptb4nF@a^s}xu!WhP{TWn)jvZ&own8&2&njkK3}r?Vf&ReTOB7rZ@cI}Y6E>MqU)
zNfNoDPkx`@i!IN+ustO|)&i{CaV#26L}T>-yzHZzcK<$U!nvG^0oB-C7=+9RTNphs
zSjnvAcWZn=3-3<#?~{vAvW5WI7@!9rb2ua^K!g6l3BGk=IR&u>53A@9P`$3)C>o_=
zY$4kt0hrr_12%`adNHmY6i4}b@83a4o%tUhZ#mp5&Sul1Q;L<D?N<q{&~2Aoc`p7C
z2Gy){3j#LK8H`k#$)Nj6VlHfj>j<grE2Zmhgk^Yz0${EDZc_>Xs;fhL6X0X5QkQpc
z0HC;)0X3sFe?<*YZh{%Jib?*yhgQE;0ML(GvV3mIPD{fIK|XC)<<ala{6+e6c0GHf
z-r|1aJY!M{;>CcN{>YQ?M=@1TD6g>PMdYeL+M72^NDn~&tb((29I6)5k|kZEaiQTl
z=U%Eat1)r#<oaF}<>sUG^PP5T{H?S{7X$X%+t)>kL4t{NCAcQQD#CU#z29(|t;>`l
zEmcxwGIuPLeoC>9^0vCd01|6Kc&U|iMb{(gk%0$$VGuF=XJNbzu9#nB&9!Iu@!=DQ
z>W-xd7X2~FqHv%!3c1c<HXyU&u{ev>i|4R2pM=i`qXS%AXBs%~W+)yH%H7e{*-CRe
zknTT3Somrg5A<nzL|?+S4Zq=JUgJ)-OrGbgZV_MDx_a;GYg0-a6CTwJiQfHAfR@4l
zNnPp~pPA(6<!=jA_@o{tA7jUb#b_XT3f8&1CWccu2&@ks?X~z8f4S(+XB95qF^`<6
zbakfF`PN?<+l|g%V0`%Y0UM;8yG(?t>kFG{b2bjhSsv}GKm|;rP8Z=&4A7ezCV&q*
zCC!LaNxf1G5Li03Gdf=43*A6Jyl?Fo4mI33CM33TsOh2^pg-yI76r0vhvsF*6@ixa
ztF*gwYRjXx$DdP_hYR>}CVhm4@*7+4%c9KwN*!hZdtAkB>H5mo++G}Ip^f`|IdJi<
zk+Lvq5?+gG&VQsMn>z-UeLkdRczC+-S3F5eYjq=z{8TVCFO^f*E2`$RW-MLXCDsBn
z6U+Xc<gYyxy=rUZ`n`E?6Mp>y!FHhc0DX17WT9jw<*?|`<+j`#Z-5?z^DXO{`tie-
zXyf5x0<ISkvQPJ=Lts0R5FuuUaPnWOZ+DZnQqe|#RBYlgQv7vxAaKy!IpUy%s^*^@
z8W~f-LF_-E2r|otp9P6WTZ6rJ>(VrekAZ=@5amklvk*WRu;D6<>uME>qQifCw}O3m
zX!}W$*^Ix8A0B=A1Qq4eL)S~-3nFnYE~#_jft#QVFF~e82e8Jh`aE4f36?mWff_3U
zSe~!Zvh<Na(95f93tCTZ=om`Z&fppOOX!sfi!_X|Qp1E8Z0xNKYGXaNRX(jm;jHS&
zlD`5|$T=1~YR<X-{z>#?3`1)+E<#9CN*m6$J7V#axvTz3539<Yt@b5V(=nT*vCD!x
zt@WtGwy%^SNP%vTI+?Gxi4(6E)aPwoUk2jQ=|vr?K57Do(1!PmdH?NGCv&yYafVij
zV1~@;`eo1V3DRu7_Ha361vudl_?`h#GB&f7Ct%ip<YN$8_kBz9tn!FDnF5KjtO?tV
z^X9zoydx`P+Y|Lt6Cd%yrs0j6;_C9F$x-1OeglTdn=&26g!?l2<*cU<`_aP#2542j
zcz17*MACNi@Dbu<OaR;%{41qB0&YE*loW^r1+oLWSpwhhBri^B=Fu~X<!qiLit`w-
zkv~B~mgD&>#VyiHm_mKC5^Vy(SkHcp;fXu}9nykF6_TR*Fq>Kxci^wVTdN#DcjBBu
z`sEW=;o%#bDo5f)bHwaKyK3VjL+#@?+&s5jX}FK6&(uFz7g#d{Y~4^~-!~p{p<yEa
zjB64`Xo1OK0cVL8Q#4Yg>INAeHJhHr0;2+2s(Y#WNxFUOD&|$@wv0RLM-zecZFKiQ
z;z83g`R;BO73d(BxjK6>`A@C1CA=zsDdUmZV@~0f1`etHn7TZoZ{d6v-`hNmQc(AD
zxkomnP|P5tbX@W8$f(E<Ha9|Ik#h10dJ>gWf;)q_m6?+1P!8}m0LL`vI{e!#5h9>W
zS>oJqkh{~ufYGI#qKNvDEwbGU8Tm>ao+kS~b~KjV5CA>f$NMTTEI5yrL@7nKw=_@A
zcW>PTFEM{o4sbnd2v|>EbT5Yk=H{d=eu#yuRk>w#vAbGL(BN`03idQ~M6nnMF$Bqs
z_2JWO<<fLJGzSdkwy);01(RHcHK5sDC%3FrKotbkG(lj+$E-5h6xG|04_)@*;2!NB
zrGII@b+K7Q0*m^ZGawr#cqCq8b6=;zURfOG?C%Lco=&a787OU@mz>==%&HxjFIs5U
zdhKErV{;}U8=`}wn+;OOPnIa&hKIW1_tX&<_jmwQt+GMPJC2yar?K}_5tQuomky|N
z6@&N*&-Xq(2r5D&goGey3^<h;)SOjOiBemv)%YH@lgss65)+w|rn2*w0YIl1RY(%#
zEtbsOj=d_NxqLYJ&^ABo&2O7*wa!O}c5V|o9r&tc+<-OSWrl;sra5*C(9hmm{OT#5
z<k<GNiixw{l?cSZD?5aC6t@qLqf*3V=TF&XQfAGu0jt9(yoI6vfc%yq;NpMx0=PP_
z?7U-ps4(fQ;=&#l+}Y=4U{!GOy;*NDNLNS1akApVj<oq1s3*c#11Nar+M~5>+yMv6
zUAO^?wC10aOt2QYBs3e%yXv=lyGB-by67;|%RLWg<`D5rHuN<ijm64E%w(cSfG&7R
z&{dd^+o>^a@?{DpIa#ef_R^<etISNSefSEz8onA4ijhr#Ey2PX!V)?LNj=!q?EN>w
zUWSz?wCzppzzkRok6Ms?s4+UsP|9snv+Rn|odDVrzqRk?ZKF?E5zjqGK9T7nn^^|w
zh&N86P*v9sE<QC&es0HSYQ)z)`kXYuy5LcHQpy~jKQVK0zZ?z}k?vmn!W)}vJCiX!
z5t<QZPtEM2w!XhRBw{(Jxn;+_B@%}${vhdbXHO1ZP9DvF$9i|(GK+9Glk;_SAFNrV
zjuF7noueIn^;<4WOa^}9ss{^0VIwRIO>Af3rH~>+eL(~9Dx#O9C+{ItDs_7su7-+@
zF$|2}66(%|0Fw92)R?Cqmz{7Z(VI$z_3IS!6wOjo&JmUUwWOR3ae(r7V<&7I?1Xr;
z)_B+(qCb>z6#|1cbo7~}(lwb-{4t^$o9DgtA*BUrF;$nGWsq3LYmz7)gTW(g?e7Yw
zr+n?D)aCiBqKW{+lZ(K^U!&D|GMRF~BmNc+$RVv=L4tK^sB0ys5E!f>8j|PC^^hqQ
zKk+EAA8|w|gcn+maj+MkHq!L1ES=Aa_yzO7Ort`U9^M!OUw}*zBuf>1#P&UTP?O9i
zvnj~oaZT?Lx*mOaah56+?3E9P!;G!1@L>Xmvb<@unOF5T%_da*8q)6-8Z=AWyQBdZ
ziqrSo-)bKQfEc{!HN0`C{pVa-m<3WNt9pAn@IW>_y)CZ;4?TZUY0YRuFVVTo;F1pj
z3jSb|cevCl<y7>A?RbrE-4%#iD&u~bosp^Tgr8PzkG6Ik75usB$NDj7_efyt?TP!4
zyptEh_%b%nf^M8=tD?0v1%McbxCyU|F?oqN!m&0sUC1gzH}MERIeftqjlJEOUFeML
zGXesw7QmI(@)fmv`9{}juz39KN2KAogEHJQE_)h+MrdcVt^uC$>ZIndOUt@lRqM-z
zpdu_X?l#NM7lK!M447NUQhzFVw6X4g0(&kMUYHgc49Mx#W$`~b-g<rY@w^Pw^zlUe
z@eb_c-B4k_zJ?zunV6~o5(5Hw40f`Wh<4-`clW2#0{8I7Vm1=e0rq|Jq+d9{*nb70
zaHpmuv$425mmg^`{pMv8hHHPY{Pa~+Nk2h;`D6r_yqF((R%Tpy&C22MTUGB+>J=fc
zJ^fSs;6Q{fQvCM1hFXSC<z%xZQ-osz4Dxzn2ET*d(R@mNy4#6gNVHU^H+%D>3CZF3
zJUkJcDhOPrvpf2fGUlzX<rEjS9A>h)^&k9>kvG?+C0b1Yn`KLBuYiNS1dP=Tehf{`
zikT>%`C@mcob`FRu$UZ3M7z(hm{L?y8{pj30zU!&i8{<!JT4I+U=|Z&DVM#qbO62s
z109B>32!9P-%6(FDMdi_BVd^}On1704$X*z%~MQysGdt`TDh!h$}XyFKSFi9cUzvj
zK~KFtUSvP7PleAN`D6>z-m@Gv8Od0-J@l{qN?82+R%_|$r|Dbb@99yv;nIz`Eb4Jm
z+QP<@7s~fm69+k0hwcR<k<+qj+HLmoxNT|*A9rPehM-cL_%~?VHVsr!U*?{nr_Cs=
z!fW3kaIgT9_+WFn=U@(<u2o`qcQPWUwZi(><cB#Aeq%r)#o<LY%?ZRp#kC#C-+b$D
zPWEi>IYH_Jz}Sr)HX|v>fFS;GgePz~Z|7?6#aXH-o|ru=m5Rjq95~e&uy7C|r#OQ?
z8?l%3(}0Ch3N~6YTg34ILr;J;{vnnj5WAHM&%gVMMBg{;6G6ITL0Q-kZSvb(+!bU2
zn@V33tDlVb3AnUArgN23UU)smnS`>(hBawMi|rA?Mt+XDt*PA%95ZR()2=rc;dTrG
zN&U#(W9B^4^WA~`7gMo1Dm`wq>_lMLx;tjRnT>3Yg7gzs;as2Bs=_zXHjWO-i&?in
z%?9%JC*ZPb<6}d$m$O0G1c%|B7m(dUq$XZM02j4P=C3OF&@p_4I?817`m1Vfcc%q5
z`?UjWyJ<P4yo!2Amg5LbJc$DhU{+S3cmd$uvbosz?D|m|4X&InHF?CJIu`6~D}Ft1
z;leE^pa|2|@&W2^>-*a0?Evkd;Lr)b4=v(AX#f?_#b?;oG**+Y5Qyuu{A!ILNzMt%
z^Gok?jwu~&cyU)J{7RJPWPe^Z8leORSfS({)DNn?2H;j3ua&w&fF#!7qA%s2*nf1h
z%Qx>clxayW?UXFEM1Mrl0M&IIR5NaRAn#4m#F?~3=bSh#|Jf}7c<eHyU-aaTh(rjJ
zxGl7o-!w};Ck(tLP^Z+WqTomcx0fiH7^wKsZT8l>%hNxp2xbO0wGSmDZo|Y-+RPTC
zE0G=~1HvU<qsuBQN6y5T!A2Uby8t*?2bvpwV9(G3^=Z=Lh>}4Mz&&XSZx0an#_8sW
zH$qfa)1FG8Trz{~y=HXqtETC_@NelqcKfVdRoRVSUx=XlTF-686N@y^lcIja^p`7p
z<4<8=xNsaZXjmlI>QDZ9eMa!~=yxy2&%hwPaUlJL9%wKjJcVA#h=u{`dZ+^oe_6b3
zPv3j7J(EDtNO{9ryCYQpbFzNWfqA=ABCVUDU=Q-HA*4`Q`f?VUpt0tTQYCfG)0=Ts
zbNJ%iylUUvT*{8(xVVwMU|>`L);(?-dGIYp?#l^}B>V_`LDx8C`AqaJj*vIG$o(-H
zar)-Ek||o^z9TaCY!56PB-yvDOmJ187@+l|w}Chx&u?gtqmJGJ1K60L^)Ca3D3Kmw
zv4M6(eD|r7Hkq58ndJ?{PDTH&x5pI)1ndol!F-NVe!^$^pn~@!77->8PKHN4-R+4H
z+)WjPl8*dTueoJ<RGR4@pV5cT9qf0lD`U9<1fK_aD=qF3MPpqoE`o14?n_x%B1l_4
z4?#nVM0U;uc)x8GTB<vzpMtA;4H&T(U~!-t1wDB5mJjaVmW=N1-`@O_-9#^xVJkSh
zGEvPyV1Qp-t>?R%{Td+x){K7?6NHrvaezX5)p8X--w0N9J;szJwc&gVIYl4OH@6y^
z_0ghtFs0x1D+K`$LQIu$JyA*BkniM=%AXKWw;nq+cVqw3CLzG0hMUPrQ1-a8xkZH&
z$-<T1<kWj8HN+D^7zFTyev%;?f@I|wt{U?TD2$not3kWNZuXCwa$$#!`s+!etB*U=
zjZ)Rlpjw~|r-?9oipUc!v1P>7%qz-H$w2<dQrr>qBSayVms4J*P-luHJR5!<c~k2>
zSVS7LuHM8FY4EBMe-Mf<;(ChLXf2SU-;u6iGG`wl4oKX*&BQE1h%k|0re+Q+I<^)Q
z>aP%+u5>&E5_whkKrDJC@EE2%=`|e_7MHREf#fn@cjds!hY2)7Z(;9NYM!7Ahl|re
zRWOsGh1AN&R*-=Z66-Jm?h<$i@0@8XP(L98r97~vk(dD-6!Rvd^U25Uc{3~yE=8O2
zJwRWs7mu-A)!Ur{S*4P64g=~L322NtfbQn;`j9|k?Jt&(KMsql0vq|}hdJ~}W@htm
zVCz@<Lwj-FzenhK^jTZ>X?RNITM$^`!yS*dm^Z6qEK;759RMY*yX{ly+PPX<HW!PU
zZkJ|YwMKQGl*;7S*((gRXY!qt>0Ja9q7dpwhiocrJ3m~7Mt(NYAiHj+t2ym*v_1v-
zhD^!K`D2(Ou$t@teMb2n|NfQQ>Iv%L9AVvAKt&dL?}41xH()t-5tMXf#=#MV4T!Iw
z=4VmDrUls36mwF22a*v&>l&xAeCNqZn$>@zZ+o!H7Wfni&C?j{R&6j#!eDv+$>D|k
z(Uh%O@L*qnp*&YvRcU&G0kXi?<T1aEuJhe>F<<PJ8x>IJ!hb4SRI+m79;jWj_6}Z>
z8eu77h;T_fXNqn{(XSv>@}8xb^&bf&w1FW^?Z5*d=}%;Rw;dk55>=FBlh(jUPYpnL
zn+61FXXBIitNzr5yQWqLVUJ>f#+<{N%VxNQk)D;7WCjrOLt&1K5-)HUk2F5ooHKFh
z?@>G6vqKzGK1-2jc?OiZbUKF8Y+5y?jJ~1!o*r&OyL)>-3e9=AxO*~Q()FAcZkdsj
z5ev0u8D}41n6Sil2787zdtl_cqp*De6qy8E*tf|hZ|MMn$_`-P1D*BidM{t&Kuqo(
zELhk&D;6C)l(U(0Cj_KWKx8{^&DNZ^iPIvCmG2P%*N&`^%BU|7N+m^h3th(>c&NG0
z!?#{zzO_(QuAzg!u|ApqM#n`n+%_2r1>!pS5*rj778&^lzn>gUilN+u)@9EL==hBN
z_=~A55n~YK>G~d|xU4Bm)`etkC{wWx3s1gbL9vp!W0Qhw_DWw>K|vR!Oy4@IaBl>b
z7mu0r?2%>36hl4+n%vaFSOQTW7W>axcyCf?rutVhZDDJvRIT(A57EfZcE=mYRlOr$
zLY5Pn3-%O?Pi9h%15dugpQ)0#GfjO}Cr)>Wdj`hG(61Kh+;l4(E!E@|<bRbk8P_;F
zCnJOwhtVKBe-AWAlz-5k2I4~V%OTSDpd8qtz2cG!TcfH#kond;-6FDhzd*F4&H59H
zgLHOch3HROm=$r&-_^#GnZ(3%hjAnx{F~VkG&uH(#SJ}zNJ-o^r+akq37ZLGwiqbV
zFTbUxm<w1;%F1>%30#`Rmqn_IV>yIZX4n2$X54&Y*Cf0;JQ+F8WZA#FY*mZ=6#2$w
zv%Ddn#pG^(G~B38`@U@?pPEC;c9=S>yD3ac-8A+=L!wjK829rwGX!uVoQ!)|7E;yh
z-Xf^aNX0CdOw81<3Fnd+MDLW6dA?_WmiVW|TnB_gazKtjlw)e2kh#g~(~~yy!ga5q
z^F>=@9K`MZDKzjoR;_k^r};Qnu(2+g&UTWV!l!=v&aX*^9?*Qa8Y2D*kiK&Hgv7Lw
zkb=T_TwE(gKrLWVvGtj(tfkZoWmjX691NrEF|vSsg7pj&i|VNBmq{FiF(yfJA$DJ3
zFIUL{z=8@UhU&R$;35HYVXHduZd3W|ncATlgo6<FH0ri$w?=k#_tMFUoG7(c$^&{s
zbdbSBfWRGG4=FmEZ0wuw7E*6t4SUDUE9_Kt4RtY9O~Dc}omwh_*vfNC)X>`Mu2H46
z^)H);-pf=;IfnGZU(&n-b8DK-)ZSDAd^<{TC(Ebk>1lQk_D<~^-#*zYLNPMq)szqb
z@S`DBB0F&jsVb}O7w)5++kav`uo(On3i;v6Lw6w2U#*$D@h4!Z%+YNk!dappz+Hvy
zE|xyFuqOMt<6KDZ`2@@q2+$W$R@5=Fkl=V{7j8+r;7N^&zjXag>2LR#kz3XqP(K4Y
zv-{@2WjKUT^tEnobJA9umWntN)y(~f4ANF{fsCPGrIhAW<ZN`12c(b9$QBe3(7n(K
z^@(UXe&eNnW0|TBN`JZJX=<9-AO|pWbQCG~9nRZ$AP9~5h*NiHB^7I2wziw2R5GUU
z&qQ*F@uZf#`tNUd_;Z)vBwOMQ3D+;+zY#c_RX>5sI$1%<YHZ-aAhuq>ct8DIVp#xN
zhWtl|>w`TTU#Y)SrmRYKk)Zl3CqCX=ILK6#_AZqFBFTTtJNlN21+L{cV$G@%q_Tq$
zR}~zF*5Z{^K-P5TqY&Z|Po!VRcA3r5Q&Bf0t!Jz6v5R}r--73c(mX%K19OopqSwKQ
zVMK-U(w9H)o&ckPe$ZSdV8?POu&Zr6tXB%2koK!+<SpqLX0jy~FqI9%yYU^Qqzt)n
z2|c>{$|(&DD5K4w>!IwAR_6ft1%~^OT%(4LELy3yVEq6|PwLz?J9+jeP}@}agi^&W
zSwTKD1D`Nd`$KkCbym;Z)aNY}$v}mR)kD&3UXdhjA*g!4q?{3kTX|(n-C~4nCqZUS
z#6s$rZv5th9r)UpLt~m4fE1S`0OaC8H`Shvl~`4^|Lre+Ap0jb20j9c7Bv#2e!Zl~
zU+On)33gUEXhp|dx8w1Y0)3tC4`NNaQo!5`urPk?P<#7AkZ6Z}Ny96mj54!?gPsfY
z$>xGrass_D8x}{sB55|0#whC8+f0Gd#a<|(&3B!nX<=1|Vo7v5_+Z!~)a=uq*0+>!
zS17G}N<*qq`49N7Y?LF`+aoq%3j}@POCJ0P$QE(43=GU{2HBRxV~Gf};IsS7)w!n>
z%?j?A?~Gp^6OHdDYV)gldVfMlQ9e~j{g^_(@<^<VZJ~E!0P5*|&+^#uJReX#2WKs>
zO)>Eg5b0%@g%ks!3+i)B6vg>d=0QWE21dPv$7oeQLVIdXCC^84wlarAxi#^{<YRoA
z<OkMkJVOC?^d>;D!bN6u$eKh_o+)=*LDxg4Ok70NC%%=!x&6{qjYOgD3>~iNFXriW
zfj*6;Cha5J11cAt=|8B&$VJwcp3)e`-AP}t+>=y~ZU0vLz8u~C8SK|6#^H<t`5TAE
z@wetx#9YSgMGJrYzjaf7y}htEIIn7PIURivlhFrM2Y_e!{`&^fPRgDXx3X1p#1}&=
zl%m850bxDdOa;<1R<7lj!*$;xHkEiWu+1Rn**jHHjqT;SM&5hRv$4B86F~&2#B9Yl
z8DnR2(2`HdJ-$^N9{bO0f5An2&c?y+($>MHA}}0A5Ijc28<yKh(ZLpkZ^|#Nct$Jy
z8T>s&BUeaf*ya0BYkE7CqNZ%JQO0vtZcj3AZBWw_?df(k;sw+zrQk#O=5k&a1ZT~Y
zq!^>&lGh1?`ejv2f<m&G&~Zo5tEv%ly${e{3bKj@Yrbl<(OxT6*qd*im%(aWLzd54
z7qIsE)_F$VYyL#@_Y6>Lq-W}!7Mj3{KBU&{z`}rKkCevV00q6*s*G$ZKR8xE5k;Hz
z=NFb_D&_KeyH)k3$^kRH<;BfUTl%r9@Mo0kL;ucMrTs?KKPSWH+yisZ5-VK;4J)9Y
zloDsm*=Dk$TZH4h{ee+5mNbv8E2tcn_?@7$S@TR?KWM@dvRvB#^F;QfkA^uq8`DVS
zsa&(N&|dF4quZT9lUG7^c>SP3`6}lpDvbgy-$OCk7kpxs^Y|r5P;7aIcWt32N!N4T
zv}B>K%8UQU#LaSJbt2p4MPxz9O7YDLhPQOEql?=(OCQe}5I@*BxpIHw5z&Sk+a@F#
zW*ahrG5z&@^K0i9Dme_L0xIkkS})gF@B-&+xtm+;5xPolC~M>7+xaEKp9OrYRZt;)
zbsiTn=w2G@i40YZKoRHexwk~%*0+LlTdGifkDg047EFFQTTOW&dKGOOR1Z8c(*(Gi
zG;(rN+~f33q@5JGe@Zo5a%b(IAc<&LA(4nN)ITkYWhb_6#M(QY9OPJj5?s;bRRUX?
zW9omk$UCvXRcwmWL7H~T91bMiA;AGD#ce}vZgwceQ^3+-*^@OB$27z0A~f3f4K<F}
z#*k=fmvDpz?x^X}*6KkOp$B+<sogDLy+il?(4tRg!Jripf{ny!8vCv*NeA4RSMwl(
zn;*1p7#j5o+^0};ehd}2#HATZ|C<3V8vDnv?#?@+uYf#LL%r{|l@y?BwNbYnIA}!)
zWxy3BXdU~SBD-=cJT9MIv~AKS>hA(;us6G4Tk+K3AokCS4Y1w@txXqZBT#yBXP!&M
zy%B(}^zghT0bOaJILi9`+0niLR{}U_F6$$4Z*0o#bbzv3w~GP>g2u8pC#aJyJaHe8
ztHy$^Jpy9d@*#bpJKu||H8|i%l-c#D1F0bjFkvDJm#`5tbz_#Du_VW<d4!q<I6;@P
zK<8LEdz)%*luF((c!{P+Dr-xr@^Th)f4&VE-f(&3m}S3DFyp)PAIxm74*Z?zGpS&p
zbAsd{Md#&p>K;<wd6Z?ftLXFg+Mq|iWX5hP9Ub=k!`|sFW_!uzP36(yaL^9$T%@K@
zK|+h9L{0(p%4=0v$TGAbltC;TDVr|{3F`MmRyU|ere0e~AS8gq5Da<l@V+lMEJ`cY
z$FKkG0K+rrlnF0*KV(tmz7N#TtD`N+v86dLr=;8l?0<dtX)TCiT=sxF{+-G@>uANZ
zn|FimS=PGhMwCis;>8$@S@PoQ4vL^YkKyTA(dot9GvQ~5LYg<Jl(fBsMIp&kVj#be
z&YM*NYRd^fc$FjE4|BV*Jzbh#Ay`59X#;dCfNlZ3rG5|8zPz2&upL?q3~$-GJ0X2t
z!*IWB$SWqnyvb?OXfuZ^ly?XXL13lOu6C+Pr^g~0K^4=E|28Hk2E`pxHFnO!js)}q
z=IGXA+g3X1w3XITqA|w-AIcEyE5H|Pr$OYX=ehkW05`Ao9=IOnDiL~!lM(mNGZRcF
z!+vrZ|Fnqf<l?NR9Tn9|i&6}1o<GOgUZp2s$I|xa_X}q$oQ;EO=+5B?>B_NnS7i=a
z9~65PS7dqqz)8-Uvi;O5`WMh=vzhj?RdUG%d=jVpQwuj&=9jr^=qqR{h-))*p?sq5
zxyw|m)l|I5&{4jW;#A*X9hSpIeeC7a(I&szzpk&&--?eD;Q0@3mXM09gd~la^nXG=
zgHMG`cDKxHViQ%#r%d=d?d81U=Ty!a3<&2KYpJT<0vJ%+_1`MSb`K&9(f#;-hR$Pn
z=f;lM15_LWRa7;L8aY+^Zk}JRJ~_^0V~n!o(a|RTx4<4qKl|BKmj4je{nB3^5)vYo
z!I!U)DVHZw(hFL;SwY*be(rIUYy`h|T&t5h+V&sW2p_%-yx#1ihriN1{B-@HRpX$X
z=VCVQ!`ZSS`D}gg0{#!V!+6M3QvWB0W4V2~KTMsVm_;#OY9f|lxErY`=biXL^$wP1
zKhU^Yzs<yLs+}yZbgeYBkN{JO^6IN|DVG$zZC&!Fv}*PaSFpcA7ZtRn&%S#=XEY)M
zUsy42h#~gvHN@^^WXbtC`8hQH-t#Sh%?FM!u|wH?m0I@k<>^t5>XoFwn?F672_?vp
z>lJu;S7bt)GoWMUp{ugFHXi*esLXNQ?baocxfI2!@(+shDG;8e1R@D_v#Rx6NE+e^
zMKkkqd^5B%cCb$F`ty3*M%o&tPR@uWsU8_%?4j#!FVgZ6E>2pswjv})E<l|e^x2*L
zGYEW~5o>Kfx%~BfQH0f_(OM5)UPYYu7Z?Fb>X(^BQ#6b~8O7%?@#EZHnL=o#+?}AN
zEnaVSZfva1?IXmGrDs*sBSXCWRIZ%0K3NR!mk%%<jZWrnE~99hFTPm%4E#h;24b*Z
z{YNUz{@WzuBvMjTYxHV7j2!h&HhOp#!@dleLf#j#$-l!-wPs92o7gKNKGkmx&_%}y
z&>hhnhGyK(psbW2uZ^9u(PjTmk;b5sDx&W0V;;EYQ(BajU34*=(kF9u$U0;`6oFP%
z(CI9zq^&#cs6QX6nmVy&APJ!5tzfXLadX&mj+%uw?sBVq6t2Y$JJ3-rTm`1IjHiQ>
z0TMyo=_SI7F@i;u@;bEil`u=Oai74<fsrQ>wV0^DG|A|`q%#Wo;8NxwNowXL|KWUs
zkv7y7Z^}V)-`W1y*WpcR2=&o`*xQHC^e5#WI-;$6@W#A-DJja2h-H3u&tbn(E(Z7}
z105le2Oo<1V6}X{A`3PF%>uHF;+*s|KT6$E)6oMn@RyEpQbse?U(VFF-d+AD=pN3f
zY#pEfi0O_Sa_<IvuNF;C5lt>0O%7ww7f6XdW=a7d=no-wq^$Jap25aCQ}TW}vH_Ns
zk6#SIbTv*M)sc+qY$hWW?DXvuC{W3gDkhI&k^zl@orOv}6fzZ~F0Z96>R0_&e!!0c
z<&!+ie1a8@7t?EN0f-I9`+KKcn9eQY2-%1Zo}NV^9%FC0VCmDr>h5nNPM@IYOJ%bR
z6DOTX9Df_w2eB(ei-9r;@H28B7Gpr0VU8;t1ZZcW9WA>1PzJh>0KaxLy{~c8^y3q4
zO*aSHkRaRDmMSt(&M-*37yJt2DV-OjvAm)kEJ;nNeb^ExPo`;=lan$^Q9n^xQuuH|
zCt8|9s?zH*n$V$v7%=ODIA3`&Di6_87MI$;*{zHjl2#%uE3Fi^Gz|})`S8eJvKw1|
zjq`>M%>g5sD*$f*Ns%Z&`YX=&6&Zd{%g@j$(5lRKMh)L%LliKtJk*QYqzpy>X+n<i
z3sRY@)wkRUn;&mG%z|O)w}$}U(OCdiw%nANU?`hz+=kh~;NPaSQ_u*^$-VZhq#g&|
z*9%Hc$g7(Jx2@e*8N-h<xe<oRlon`(^msLqllLH=sCl8YetvaooYH%kQ%pJuRa?|!
z9AA^~k)fD^wX=2?u~rn-kNHvl>m9(>%Gnjh%enUXgwoxIVJfFS5wSB{y|?)8OEx>@
z^`DOSH~G)($7Hm5hsUlI;+)ETL-<5xZuu;*pJYkD<WU`hU60QWc}l<~{_=@vICYKH
z-H9{tu+GvteU0$KQqh3Kd&wfLDutx2!`h1OeRjd65SP0h66m&3GSK0@jWJNzbYf*q
z4^qp>6jNW?Z?=7tHLtOk^}l-oOlfyroL|0L3@vtq{MpuQ5bEnC30jc@jpRTIGI2pB
zVzc`#&Ds7Fvtv_~!URdB_~TPwh&aY2_=SFrf5Kvt2|<g|W1@U-xo=bz0g!sIb_m~?
zSu*CzVA^*gwIcZWo->RV)G@CULH9K8G30!5k@e_|tftf3AFFG@YFyp@-6%5Y`Gajk
zxf}iC#_X!GqQF25WL8t++PbN<1o<!$*apxDv8niVO~<qc)K@vGE`L=QM~^SEhCox5
zTYQfU>5Y7&=?Qg}2&$vxjlYQXO}Bje+o7}1FPTtsULe0)>IQ)2K~-vU-s;Ir)B`Cz
zZ~@t0OU2fCz^)Mj3aKe4US4&9t@?xtEPI(fEMNgx=4(y!F0Z)82wXW#<sDd?%QigB
zyOZM^S&AuaczBc7-%OYu<dVRp&6w(XM*<bP)zN7Ux&b(ruTzKAo<OTgJTUxf-A*Nd
zB+XKh@8<EobFR4Pfz$p$Y>jYLC8&U!n|$F>A?_!_{59kDk^j+p#7u_twA8Un&C*b<
z4e5K4#q)@owaMyt-A|}uO!pIpBTZ^#2y*#F#0}(Wm;sto+f7$+p=l640W|GX4Od8~
zVlhxw%~@o9p<R9Yh*HNXX;{txwzBz5W62wL2(mKE)Q$G`exh8*R)Z!{{62b62!>@u
zC1Ap+M02PD=L*&h&MgJn((*G=T_`dhW}0oD4CxeDQ_7;0Qyc66<JEVC7W1j!-RVZ3
zdE9rfmz7+v*B-De`^qg`t$V_vX6M}BN0z>ghVD1{NuBQTQ;w7hvp&6Y^expW?|{`@
zg@uU72hm=q<j$ZJ*0PNz!?|(;%ajv@(Aly^WId#%nJz<B<8_f8*OeRB)RmE=)s%7R
zyVF&*O?oI8dZSDQOup=9GtIYSXB%|H(phW0YKJLsm6KK(b#4T18!DIGoB|10esbYr
zm;uD(P;?S_H?B`m2p{-JM<EugZlPiKO{THu+7}*pa4vsrJxoyRDtpqM#Cr9o1iO!P
zLh-H!Oi*tvjZ3@QEx;@`pT7JFwp(`;s-8l%J)UK#wd?O6U0GGpiz3#QIUrkkdMl}j
zNt+`fjQMmG`!gIGzik_{d$^v9v--}|PLHAPCM0Q${7F<ie5M0oxeg_PbR+1(TR4P0
zF@X8mC)3}v(OZX!L5}32qK;=hHr(5%%M|WFrnbuH>-QFIIqmN%^o-}nK7)_JWvXWR
z)ShvD!6!JEFDt3tNarDQyV3K!xc2yGE?k`2mqNQY)}cJX_C2Y*jl;bwHBG(qza^8S
z=upNSepGj7a>t6k3n7~j9pHG0J}$ex8CITEX!0SNv#$da;&%V>g`6t^`4PdmW<bpR
z$VttNd17w5!c-9vWpktnJ#<-{d|Ba+sLBxc+V$zY1V`xG2SSK~N<1d(ESoE*PpFSb
zk_szZ+iHIPdt}ISat*;oNkIad8YP~?=ne0R_=b?x({`GxyrFd@ETAFI=CdS=ge+m{
zW&LM%1QFl0oaPZ;jS6sRT)vHpTjcl~8`J{-x{RIA+w?d~kBs_y^ltAkPH(I^UE7{^
zqtzzuP)?#f_+&e{hle!>e>Z0TBSiA;lb(su2D?z`?3V4B)n2Yd;E17w#R>{1uO&&^
z%-jz5D$SQb=CJHKH)KPORK3H&gL*|Ozn=9<xiO_7!zk)q=N4qih9;jOCci^V%IxlO
z#K<%r66cS~l1*r$=VJ=moyYU|axrES?poiY$w1WcXG!xGPiDyM{-JQ}jgjp|sV^nC
zw8pTCi;<ep4IyhPg3!=n@w@7x98m53JGU06+>s$86cJewvPmZ%;eq49&0=LtQXOVC
zjoS39{MHc54m*ueS=UMJRP;HON;Wzwo$ho-b*H>6KcvR|chC^Kk4g5lcIoQHn8FoN
z<OqM?ws4Siny}Gxm*;jBHI_H(f?2SO&i_NxHAhw2z3rNsY)s~4+n8)m)|+ixQ%$yA
zlWp6!ZQK5yertVewNC$p^K9Jvx-abWzWZi<OwVBM0XY(wQY#TA+iiKdwJr0x^|u+q
zR~!EI$@zm5M3~Zv$L-^Nt^I(gfDFTSJe+!piq#RkWFk-BL8u(85GUrpTllWiTl+q(
zUSN-qX`uJFo4l2YWkb(sTV9X;)J#cd$vidot{j9KK{eIkR{8EC^DH{{fMA{JBGdVk
zuDNG3{=jY-qU|zqr5waDb@aLg?=RRM)&ncvFG2m2MXj5&`=%p%jU^S`v4d3Nheljc
zL1I=azwjdP4(@Fl9m?nU*aAg!s)ZSy)`*C68An$wPjc3zyKwcs!f}BLOuX7^R1vr7
z+a<1&X&eWIqC4J(cbCW2CEiGR)&+TFa8QAR%kvoOx_I?f*VTh9;PQjtxqnst(drXr
zqcHNbv-=lB^T)@^rEGqxhx@1M!yVs>UyxbFM>9t4z$1Q}(r86b%FDR3!w4Ej%*Rn|
z@)(xkdD@SAGhWYoI~mgy4?k(1qk;NnpwEnnnkJA)6fxqfW@R=?q(lGV%UF$tMzV)o
z$xKX?6HnRVB^R(sS-d(qG)C4#k-NUpoyE@EpVJ@-GJ)f9;438!0DU|4lz5g7v8`wl
z%00cA_xFVgwhg-HKN2CkDalJdg^kiV0`bL1H3(;;$aVc043~)lq;-gOiFt{oLwLf?
zC)dAXNf0o2#hjF&Z8Xm7o=M<xyLjr;zvcV8rr>GS%PWj3rUW;Z;k7k9@=i$M4<fk0
z&c>13Z|tOXBzoY%bF{uO^L&;)m{xIGmh9E`>ekSh5XHp?sT^!z5XJEmE=Ld0-c}mv
z&I)8wdZM?41Xq+kQN`z>weq%+X0fB0`m1Rm%Fnp_0;3fhzuu|`%oIW=*X@ss=Qlm;
zmm^ivJ#+ldu1*p!8i$!eg=L-5O*&WKL+AF5pDzEDO^@tXmvtJ`E@e9?NFJb{k`zb@
z8Xih!E941LEJcy|h6lwU7ZnVO2j`3i=8OvROW~3IPEE8WDj3to1d8?&8*m2>PKXzi
z*H7QNGb6UfL}N)Z*Isz6vw}~J;k6b{>3uT9J!W-0QB@wAc??VFg~=AP<&US~nXs@G
zx%fcIGzNEW-HJG`@F8&?ZxiSJ5JR%|CXCGW@%<@;et{DaE)xF@M;4ae*L==(Mzqhu
zP8;kdMHP!mSzYuSJkJ_$!Nn`86BBzco|YgHA(fx2LM)vXjW;)prTHO7PJpWg)^#4H
z!btrU*uOfIk-J}cOcCs?NRy7&n;XzSEy}Sx?0n&ABphOx#|VOF12NdG84el$*P|7&
zzMEa-O(5mqZUGL_$<K%l=0evI;~X5A8tFc=T0&vQ*#}gA8d>Vab{9xF?YqILfQ4ip
zJPQwWu~1#3om_4REe*Z|4zK$PF`{6Zgg2MT)w2@%feYtoJ@NhYHPW|_vQ<uM7e-(K
zL;OCra#3^jJgp@$jTn=;Y;M3GTLht*NCo><Tb?kJ*sH~N-6KG~hOK10b<c5hR7l_V
zTlX*Fh&X}IboK;vwn~w$YdY^5ZWMX!q*P(P-^ogb<;I(82hHYCsR}t&O8RD&6T8Xe
z3vx&btb}-$nCA=b&a5$N5?5GLJQikLq22hJg?N~(v6Kwa?ydu5`HN=>k|l~&nO0Qg
zs)<DDlx^=>DdQ?Pg*ATj`rSdQAk1kPV^Q8PvvIKT_Tt3R`YbxhPp<FH6umR<AX$qc
zidKY9XKo}$Ss@xOzp2Dx+m_1V-@OM^#`(;}O`HN%sT(=CribP!j9|b82vrRJleFi<
zz(^nDm9pMRwSH_ld2h7?$FG-d4VhbWu;m=%uqqFD4M-<W=hM5#`6Uyq>%Z41<%6v9
z-Ic+zlC}q^tcWJ-%a``6hy`AdS7DGn(<7<fNLILEgd^4Uswin_8z1Z(`m<ONK05VU
z5hM~E?G*y|VsJksVv9<bd(v7r^;_3vAMr1$2P`ktJcX9>V+|CB6V{k6vGJczbhY@7
zdv=|DHqmwzO=ev8U}?%Bt5T`|Msm08VgGitu1^tjjF5DIpFDTR^F;_I{U-;(+j?`0
zdh@j3bJ~32(+DQ!5F;D>`I7RCF{|~Cem?#R;`6h(cY&wKV<=MQFvj6>w_(zI^C#j;
zhOFu)Glu{Y2mPh=1fB;(r_L)Bqzw96-2<a{Os1I9K12<f&OxV1BkS39Ayv$O%%M=&
zcD!}hmv8RJ<$6*Kue%wsa<1i1ULZ5Q6tJa+?KrFf2`YnRLa(Ipm&e@QDNGNz2$mkP
zK{u1b$qY@)5uz6ujgGGAgS1PoYtDR25A=LR^u1IPW3oeff}V8K3L(N4`pL;0`1?py
z{WS*6t~dUWvaL1BM`mVc@Ov^m*4!BdLUYPi;MsZY_DhE@9_wq>FNnMwt<Srgwv(}0
zjI{Z_`<$@I-;wj!g8cvR+?7z^^Bv+}XuU(4)55c8Y67zZdZ#9-WQK~#e^^K;ai;#M
zc{flE_VM~X?h|FZek>gyBbPA9lDLNm>KkIZ6E|YnCw0FJq+8hol1tx_sSOoGWCL<K
zov{Bn1#waMswMOG-JPxmZsZQxCyHnVBX!xadK0}!+InR*E1E8vk{40zWI#ifdKeuS
zUOgLVlEKx__3YT1-g94pF_xEJJ&WyM+Q{^mpUG57agODF+y~n16ei<-9C%>xxXM>Q
zggIA^9BF_@Cip<p6l2}!s)=tdP9FGqDn6#cwCCBkQE(+?yYOp2Pb0^;zbT)$AtGda
z>)3{WEV8A&&m3;3CP}gR^M2|f92x`Ida&xhkiW1CW}nyhP7D>3C!;zogCDD>&bO6#
z^mglGm-nSZ@1@K>6D5zA1>+kN{M;=~w97uSjnquFA+{)h3EafF9imIlSclWFb*;BR
z@`3%;$k6{ex%)pSr<|D5UiZOA5v6_ix)&DxxPrUpsP&b<wWX|vWo(%KwOFqV->e=_
zf1G=y$(<nDPMr(<{Oig8X-St{cVj`x!;t_T>Vq9uIBRQ2ThM}+W7Q3$1&g@vpzQ-A
zpYC$pKK_j6z3F$$B_~aH?!^e;jjK0@flLWMzHN{gKS*DM<)REPc48LH?(USo6M#l<
z4sPwOJ|Aw~wYV1Up@y$u(q-M1drv;Un<ZU6B_1BdlPowTr$7$9D#Y-ihk&Ok^py5r
zUb{VusvKmEE*k{X4SvWT7XpA2U(5S7dM>t{4#*j2e8`*a9G#Sb4MPe(5h^R+urW2K
z#;134TS1txIX7n1EWng-pQuN3)+W`p&KvDk6~{8oBA2Y5R(YH)ts|5*sGMEU(eOd%
zQyk`2hpBU3YYG?NZ_{Du{RKxoFL3vXmcXyY6?$`>Q;_&zdi-3M@Eq_orTtg3N6Ydu
z=k4H}R_b!&_Dr^d_hanA+0vRD!pRX)VXXOhHrO#r@_ri2y=ezcAl|@=zbB{1WQeAT
zBzt!;F6p5yVD)r6@p<vC_=*;Wcr(KjxfSQ3wl(LE6uHh+sY^LGEVIe1>y=QOJ!@Ev
z@r-R1wiAB|L?JP--C=+^j~ZQIq&E=lY%%mwSl1Z@`Wr2_Ykz^(U;er=|4BI!#RXHi
zV<oYrMe^pWNgH9JxD!xI@U?K6150$x#wTL*v|!%2=b*vBVR;$#sv7#m7o@G@5-c$?
zB}%lSpnGMEXcqZ|Cc@W;lUh0Jk0{yT;(X}FA<=xap!1CI1c8rp)JG81Px7X>_~d4&
zcE-k}ms*Oyf9>I6oZzg*xDhlRmj2s~^K4Uj<NKt35C83javX$aAEEBZRsQBWu>WQ@
zt2F00ce}FhLvRzDua!euyQQe*?4Wku0aSWZi>1J3GkVAdJ^=tt_mRjP+hvGaFdIau
z)8GmRoTT5Gef?efbMIgR>CU>z_cnaUWkn3ilrY;>Y{5r&E@Wu7JdoB^ztK~&Ne*MR
zXWvgewFr%howhb+NH4(9z*D`?iITVTe4I`IckCYY$I}po_TXTrs(YZ)^6Y$3ZvU4Q
zXOxZGk2qdW*~6BUO;TR21l6$3-S(Mzaq*)BuA7~KDh>;T-JOAGi?IQ#;sVddup#}3
zJ?Z!EXz3MVxTw49O&ID|Vi;nbEOL?s><UOJ00^z#bCn|6{O2}7T#;cXQt3IEcld5D
z{FpQ~d6Th`9I%lrr1*dWAV<0WjijNm&Od*`pVu#91XfwWVeR>7G*@k^sx{xiNGvot
zGY;(MT1X6Qc2(`3WxkLZXs-;%t0&(E)pIg0{3{klQh+y9%`NxoqrS7_RYnMPkR$HX
zXexTik-Ugv<txnZ<t7qWy+xck4Qvk%q8E~ofbxBkqn#&UJW$Fb1q-K$QH{QO5fA`k
zMyS%g;vmv=!b`;3daP4?<Anz~C}^KM_du6A75Lo(eq*fVA{nL?iSTmig8~*AkBYH2
zkkeCGyz=?uT;_Es^Q}FaV&9_wvKGkK)1}(7sme5C3Lss%t+U{(2hUpP=Iw(phI|(u
zJbWys^5&60MM-JeV*A3+<9?5SL&u?qm&C<;QDTp6)fny+mh`Px!U%_epE_nqisS!0
z45(S)OZW@aiC<L;v$N$2)(k8fX+j~=X1<W!)o*UIN<e2dCJwzx&Ww~xaXG!Tn2{|#
zZc^W0RzZ!x5mW-^C-@5_@sMV(q!{CIXb_i8`GI3H!DgcmNA(R=07)LABn0658_vK|
z!SWkmlv$<+ct|#nl||j{p+6~pd`HeA$yp$cX4E6S-!C<@pdq$edv^7_TFY!)mT!+U
z;)HGXe6y!1VEUp$Sbyjtu4wI6I<w4U?AST`Z&lKuiF~(LMw3*Aql!Q-C-Ukuw)3a>
zS-Pbd%KqUu5L*8)f}0?p3nGh!=k(xx+=yl<eg4!-;zrMUEr3W*km7Q`CQacK#r@oa
z3iWHfyAclKPe8Www@`nRmP_F%GVs4S1bL)s4sq4K1V8>yp5k7OyTq1r#e>8MG&0%3
zwR~1Kp&2tka<X!=%}D+^?myNQpl|oIKZ4co9s8~2tVA*8yQy0U@Hf^;f$4l>Ys8+s
zz}naN_F@K17>DKfoYO+J)%&H}qH*EE?(W$inu&!H%8jenDLcuCm^CwgQPc$S`H#lG
zQfVT<w<6A`Odm)AVWhq(LI^@OGpgmd3e4)%20^?4(JV<#1LS*I9Go+{2lgaJZY$eb
zssKs3cuBF0{1|^e$M4*$RJPo-06cx<{Kbv;HQ&Azp-}$5tK4;n`V6LP)u!moie$G_
z13pNFA$k%IS_|<u|3-}lkDZ7xi~!iCX^1r=IOwYjY2NfNsxx|~T+Y_;k~%!iOUTsC
z!s?mqqyE;2<(n3xGb4r!PI?^aFAYbg!wB9uIRvln$wS@IxA2tq2L-d;6ox|+7G<vx
zpg19&<sY@%{HaU=<UY%a<&eQc5xSA01RU34%EEcQe)8^l^;z*t9{*x%9Ox9-5SjF|
zM#G~-3IcWybfjou`_=2FR(W%MWD!oCM!cljfY|XUU~LC9n;|k!c)JQS;|d@pGd%wI
zsDBXXC|U}X&DkEIq*jk<bnG%%#rwgPo}tCy%}ISbdG679pf8+)A|xnq#5`)YREiF)
z1@H{PuyRDaf(|H2$03pLC@8Hb?-|1uBkw;y(Y%P=jnG&dRJ`jSV(?48Rj0=6v(L2%
zE7~V2xf~g+$x2b^)->u7xGNYc>r*ioXoXus{ge|xdYZl)ecBl^mefE9)Q$x?GNtWb
ztHX%ki@2_}fpR+&=4b{M-50$TU!UjzuoPjfo5ZwSDMp;LYO%aY)ppCkf@#uLxvOqJ
zPip-MrQd~kR74UEw9*dJSf2AA?8B+9%T?bZ$LISCO7MOsmp<o~XG&0gZ>(tYk>`}I
zoE(kQ%o3s-fz(pIYO<vG4I!NJt+5u2<DF`nkhqW2{I7!8>_r`1n*qB&Cw?=95ZTDn
zV<W)@W0-AI2F*z7c<<p_@7}*_px_Dp_-5=e`1g{t=<U?bDQLriYej)jf#E%PC}_Hl
zER=Y;8q2bDU8c6aQ9ft#v><24p{w&ZDl|)8u!O{J3^eiQ#xo54g*UA4CVgjyg|QBD
zsz|9*2pB}_2S>x{#_!I!$i$p5p9bgp>s=S0;o4Rm1EGDD3F(ZXXKYfd{Ve#Cr+{*F
zG<Li_;FdPjnZQ%ntVukLi_?=bv0_G1f0;kPvHjPvxkuIIzJEF;!p9#wFY;htDr+l(
z`eWE8>BHa9>l@HuWW4fyqM97iZKovo?7S+`b6P**vP{`?NAZF9uP?tZUm^6#8G9o}
zP4<GF5)_^$hc;#>=Mv@k#xsrvdqqI~B=TgG6wzeW85Z+0K{cBm=D9ED$9M6#%r^Iv
zu0XhdSCfVe{)eUm;0mv-*YXVdG*?i1og0h_It~m0B-8TsP$#4p>vM$9|Ev^ux;}XM
zDsd>~9SPG1qSj^~KRx3asjXmMQxR+O3v8zb9kZ5kr3CcxJGEb6FSvJ&>m+G?R_^5Z
zc%C?q*9~n&srD=7A^c?;QzA&*K=p+A@KPAf))-!2S(Y_~+-*Q2D^!5vAWAFeC-Z0Q
zY=z^UA?g|IwVSqHV<-L*C#dVwS!=Hbbrd*22`bhG7A<K37?)VpG(;*xMrz!+<e@V%
z|Ng(D9Q;Twsh_&PEe{q?@kIo8CJ#480gh3aQZ;wJFz9y~Z-Cmwh3ymn?R0>eecZ6J
z>-vX7sJ*hd!G+^b;DCirl5O)1KSJB!uQHw*eEQa%%Rj$NIsDAyx`3CS!kg9j;4`*z
zBa?G}fvhqBo;2K+nOBRLwZ}^*m@wiic%P^HjuSN**1I6M99bCciGO3PO>kVy2|e^p
zir?DAq|(X*TZ+T91cZ>6i}ZJiH&3<G+?IVZ(4XT_-A|^vQc5gKG0tD^OjVcw&}jaY
zLn8JIK3l)<xx6_H`h?75KE2<pdeHa`C$vPjMJ#gLGEgupXD62aRYb5vBbVrqzZgDr
zifA%?O_Tih>%VwH3?q_L)ZPQi7eVwVL2lk{u~dP0rthT(S`qz5`}`go2<?|QEkPue
zJ~WxW*p*t_pTrWeh{L;GM1G^Y*wbJp=JtG&Dc_P7$AxtHhGDw>wQd-#7d~^Q5hqW{
zdxoa4m_Uy#x!3Z<{x#(JiFLfV96es96GF~zo&CxW={;PJA{;>mkGZoujAJsU>`%kb
z&C8E(`t749k+eFY%f>GRFdab?#y^ZFc6Q!w#plWI6=|#LRh2pp`@5cBZ6HnS8mgKL
zq5*2%;+LH&C4;=8*;@xk?4A(WW%-adm#urylG2}S{)k!bD#;p(ziTXy!F^7(C`FAd
z5l1iG9gqrlkYxoW@O!XymVET|=WZ^-hd;pmlIXnrm1`>tnw|+OlT(Y>W%YxLQ=QP$
z)V*B6aHtx$u?eR{IlPOULMT`VbP&n`aJ6&CF<qLQoX(;rpQ=B&W6SNne;7$xB)uKS
z<5-7FIAZ%dRlHydDT@uqVOB@uv;pj6W{)?5l)^}|BE{}I`{XUGMNlnRmAW}_w*A}B
zAb=_Hy7oRc-}4WL9e__~E~D3ts4Wd_np<{kQ2y_<2myXVmYtY6&4%Gx<7Y16jqET$
zsT1HR%3Op|<VWh#G-b9fydnD*I=A&1P<!*9ord`#r|QCvi@k5pOG~9llPi23?gjbM
zhN36`KqqGSN9}*?I3>F^JpIh%uWDF)NP6rT?SDS3<}umMesv>;Lli(_=#%UE-3XTT
zXd0+oS`MF~7&drFB*E^A$_JF>ivbiiWh)c0b^O>o6t}&$*O#?TRJ97LQu@9_r&yK)
z>8?)cgDZ}ZNiAuId)2x*OCmli1m{K{&a=0+e426fsYNre!>dF0o!LY9rm_0Q;`o$e
zB?wD`Xc7XM1YMLg9(aQh76j{0NIDOkpeMgD7+DeGZsT~y`1p1OoZhc<c~>)Yznea5
zFO%8&gmln{hI2sLo8w$jXYKt*eIf)xLn1Su_f;nII<z+>DTz;i#H36-V{j?ckDAJ7
z!q2r20({QP4!-+bh|yR^*lCr@K{^TVT{<fsRAU`6Z4+jji_d6)8J!4*<Ua)Zv{(#w
zW8uZnDSklXGAmvVFVBN$pVsYJGqez)Rnn_T3*LKrF~5HOQjb6YeS~91Ct!RYy}`e{
zyfh<fgk52M-&p4ze@6dyX-#V)X;n9;AdioQS`j3OE<|GW166SzD|}zHD%76g87;0o
z!%Yu6GK!Q!yp-9H__>YfXZL@4<Hw>TOP;+cNsM3$NYA4Og>JU*s9ywlPm81T;hWv4
zpcfMGRJtGTA2wp11RY(9nbHM<luEXvZ5B-)<3d2KkXma#3PMJQr<i9nJfzDTRgCI)
zL~3pf!;J7#v!X#o9jpds!ROL5P`A<}9$;)zQn*sghEe4-#l}%3<Wq-sACg*KxXRCL
z=z4c=38B8t#==5)Rfwe2x_x$;*T9+Os2^vs@%BGG;d;g*W=#gm-6$BC{RfcMBqV6R
zLS{os@=!fQkkx}&ei)6D3E+q+ts_{{QO6Pm<K24*%FJ}?JVJrE1DH6)ANa2+aeU|z
zS~4v}ffV@&QLCYZLrA@9g#=-fhdA-e+2qj%vD7S0>3B`J=ce~&N4q_vE9oa+1dt$R
z;zHS`(k=#U<>7SY$GC0-&nRA)N?;#vuZ|c}avWd8d3|f~rR5Glk~RJNk3>Qn+NMjJ
z`N?QJT*AP7L>T7uab~|Es%TRX34G>eqR>8TbC|_uz$pj0h!9FzhHh)y;o_jVXHnh9
z5Ekg@w#w2)v`|Gm_<Dw-hXc_6evp<#D6a~UJ{=xp3OHMC8k-Hq;-yWLEN>n*YgcZ=
z>={UEA}B0qTs#gS8~yeF6;73UDXk??(Te9Wi_gZjR<uiArT?}Z9yJLMzcrRj`2Su2
zs*x6|>lVko$xmpVtR#R6UhDj*%sF*}c<EkHvDCK;?4)8|OztO`?0}6L4He8Mu=Uk<
zT8WOHLSV=Sy8_3S--ppQ9VRAn*fz5WKz=66Bp#DZ8*yUp9PM=J=sA7m+X5iw>$;EN
zn_9!-1;kq>F1Puk?f4jVP&*v@Sv!>TT~fVDRjoB*m!0JqDzXt(-E{jpYF<2-b*2xz
zb!8g+qfg(bB>(PciIud>$l(AdKcS?PJDE`wdyTalmoI&!r5+$Po+H4HgCbAlRMD7#
znix;^2KNE1mcHnTp*7rY<=(9u#mk#DOD;fW>B@L&OcH}<Ee4jiV;h8SE-R6ZB*(L4
zvR6EjJ0*&)>99jMv9P&de4Xh50p94LHw{!I5<RDOa3zwK9$U;Ac+9r)PT$*7<FO5j
zKv#@xaITt^62z-G133VyUDF(9c~*_-zp7#A@L$SIq_UM^_73Qi=p7RTlg$=tix2Jr
z%sA-ymBFtIA8u-g$B3=<M3&jdk%9$Zqm+`&5vhIus;1Ntqg(Frhd`z?{4Dtn`7lb>
zr6uZB8y`dYydGd<zNPE16X3tK>B=Q_Ac-I&1L~ASLNnVLpYrP_TzjNU>WoF}to6X5
z6AF6Vx&?3tOsel@)&pKeraN>#ZSb+9`j>JRy?cBATbtgBvk%z;UY+r{&&)vG!4-@E
z_bvVUy4j10=dq34=StZu{4dk@mnE%K3%0oA=Az(()nmbbcW!7v1H|!qG=WcJpkG^b
zQB=6R@JYeU!Q|ks-3N}3uzx_cfV;7@ska?NqG8bpyrya+RCu16<jA-UoaJQE!(|-6
z0PX8h?(1R>3=?a!FTrF}y2s=bmeqqZ{@aZNfI+-9tgLz;s6%Y)%%zZo0U00eR}fLc
zvYFDA1Rp|5B;4!TGUM@_8ury61%4OK6`{&jc4SCK_Gc5^0ILUda6kV(o`V88#5X#@
z#>xJm5vw=nbzRq-BN_aaCq(@9yUHocqeQnwI4YCbX?Ku%I>z7iNkGS~)0hEhnE@?>
zN)h5!SR#PlJZ))_UvKfGbdK(3LJXsicpPnOnt&af`xfB_2vGZ+VIZz8R|x{_i4cp8
zn&yr6*~XBQE1qk$+XOl1KCM@eYn^Wm+ix~^xiwF$@z_5l3?LG{FDLr??7pN~Elwj+
zm81w9{<b6>zNTQGOV^a1mmTq}38a{BnNnh0GJxC<0ra3i^Ga3YOCCo7rz>a*EO-2@
zU>FCdN-Op6o6zkD4^H2!oqo^CcR(17z~$-(tEwMgFA0TW0;}=)^#OxYe9S=Q1T7t(
zTp7=)Y>YoimCnmv9$#KN<9=zS^sihRs=HjCoM>@mw;qnV&nkPu5Yq~UzQTbeAC=}H
z0sC79Wa4^|3a8`z9)}gp;UVq*JD5e&e``$k;a=oF$@>mx46VBN?Y<c0px6SNv(M<X
zgAdcGMH<${n~HgX_?0dh<_bbUTb|qV-&KJplR%*}fgD~o$(2VK@4+V2C_n7EHpxjW
z1?Cl~V70~WkXfph!1=<!Y4(j-sgQx$>E>@ahOCo6X%8fKDnL=D9qI;1$8RHW<RcjO
zZDJ_J!OH)&XH!1^Ps>xz=xAW%K4-DqX*DDg2SY>g+>#oS(4eO=_&w#e$#Q7V+<D=A
zJA`(Uk&iBz8)NhDyqSUGjO*P9g>j(_mSeXB;D`WMz*l630<ivd%g+^4WLc)->ESi5
zPcCz=lVu)ofe8ck6dyW>`b>W7C+-BVGCweB*;J4x6Sf-PZtSNdBx&sfz9<Jbt)@2X
zbUu2IbnA_yGh$J)-R|0p{{RZ1ARl+L#-MT#G}cDA=FDDUz|GCX{-eq0&3^lyr}buX
z&0}Cvjm?;7{@m%!<8btC^VP-f@g&&Q;!_NlaWP!2=<YWON3EH|GDA_Gke4?1=>sug
zbaoidL$i}4ZS8Eqjq-gYzLc4O=02fQFkgY==`<B{)M^mn2oP64vZO8?RRmKgezRH)
zGaNLwU5w4j1>pQ57N^e|(+`tLtg8HM?uQ9?$7!r~duD$M*mrojFKnnFDj<jtvfz8M
zWIIJ!|K@yL&QtXP?fuwqu-3j;pz~9N1KYC8X5aI;@7B#=6B49N=M9v7;#?iQ?&T_}
zt^U(s1&`OurDmV0BQ0Gfoy~daag9Krmoo6=X`Bm<={qokg3TF#>gl|2nO|j1Z*%eb
zm;J#@Je@5ug%Y^jnJpzGWLNCpPnQxx2|j>1Tec?R%YHPn(|F#Sa{pmCK_ZgHk$aoP
zXpEm`0kb?`6~%cx3*ysJK%alS&~nbA?2V;Zi-`aXhiEZl&9HI*h27OGCqHf#r#oSr
z4*3W}*5a)CJtxqQVZo$)OCb#U#}*`I?#T2=|4E8i=93P&z(0`@W@D(RD3=nh20)q!
z(AxF@@<zT%p~JQD$%Bt%@<!gtw=pc<hXj(4M}#sW$>QC$Y7@e52SI}1tsx<joG++q
z`}+vVBb2gJtbioROcpVg%tA-0CeYg3Q}N=3ws5&WI#OOnx9KZcp)WZDUO;vjhngm0
zEWJXB;|VChk6s_-*yXf2k2`hMS}Y%yj^c0S>kv!gaAW#)_N*Z-90+xZv->rO7SHTM
z@&7#a=cdal*RWyO)(Iy3iwNnQ$ja+At2J?iM%`D6w!>;y%4Ni_pFQ3G1M9M*jC*uT
z|J$-L;LX$w(P*$Y5V$G)g@*w}3_H-cuN#6;0Auid)O+~-l;8{Z^b<^UgaJd4FYSkE
z8k(NP7zaoO3KnKt^X->*(|>9~i>xb6+(#6&Yk>*Xuq3O<aw0X^rz^-ehMjl%A^%L!
zL}ev-iNhg_d*1iqJ|bc;Hw`#vx@ywD13Tt{o#0<pMRL!JzBBj>jiJAI3sFM)lJuv~
zfHVLD$%QpOPX{b*E|E<^q_vGR>D}`NJ-TyInx!Z1Yb$mdb$JFRUfOAfGr7+!MuoAY
zY7TFvyX)0(GlWa;>)0<Lf8k`veza)<R8Jl2S#GsHUZ`)bpCzW0DPYfFOxV8VT3&Z^
zTz1uzK|ValU_ASTglchk_@Eiu$L0Zpip(teLV?0`BeoG$;9TP216|r-HhT~Ai1wp{
zouF?pQ`OjDsHg+m&}9UL$8pa^mbA=D<T3aeJsk&Np4^}5Q*MB&4VoOog>{9KGF`64
zLV8%XYu~Ad)zx??oEuM>oXHe&Oc~dt@<K<eZS}iOSwXMj;BRwkhLgt=vHUk07-Fnx
z%5X;GAb)ukY4;(RoisreX)YVBYdJcf_n;|{&*nXPH7&N4F)@nOR~T61x8$e2;hghj
zP#a7s%*wPJd(j6TfM^JkndeqZ<%$Ukon$N+`)_RT64d)L<<H$Si?S6VnH@31-DD8J
zP??_d=4@K^O#9>+zF%W;1h)RHTnwR?m$qy9l0lFfCmpoJhC&pszie|$3XMsaPAmvf
zaby;-A`!1bDELkp5g-7Nv;zTz*rs-CV79_n)PUy)oDY4NRTNk5xI^e%JfKMvsLTJz
z0{0~3Di%@*p-4`4#OL}M>3ciJC?rq{=$O$D%&Eqdfw8D9?MS{CYXJEGiT{zYW&jb7
z7k3>2^=nf54QjFNlwSMNo*Vd0<b4PIxY205Te%#CBINIO!p%Th-w$gyNvwfMtQ2fm
z%>06r_d#MzOI&F_30M(eU7UX=JOLBXmmerwelVJOpX!F?+Hlb5b|%xs?8uUgo$a)#
zRF*}hVsNb~?Wt~i@~va|YZ>HKbtKaz;l0~=yUIzXcKx&d4hq6IESBHQX2@wv<Fj<d
z{zW;{Hv2LueFsP$!n*NYr+Z8pF3WM#TnLp#JOnY;T~S@9WJ0LE({SJ`2RDF3BKgdi
zYNbSoIeQ)c^|{90;wahC%f1my{zvhP@Lfrf0x3Rr{@nE<dgu0G)>|YkYaqbZy3seM
z68aPVCx6ViO1mw~YB({JIe05#WBachMgd9aTed(8yXXj-hE}9Ai3Il5<>76MpdXpn
zS|EuLfi_(l6911co6bbqEC0?4mTDEDgUZq^+@HN$V%%8=L+hmo!Vw8P4_N}SjFqp)
z&K<QntkldTC1;BRcoM%+MNnY@lELKR{W!Y4aW8`Fu@U}jg=MY{%4cw2KwJPhv!((4
z&*5ISqiju5>Q}~p6|$2lexUARqaUb0EQM3}r?JsqanGUrf3+=$w>6l6Fs*XG)e<<c
zqFc=a<~$>z{e~$!s^(DP@)&S!+>jPmKwcB9M0FexP)8!LQOt|OLXGP_ILYw)Z*~=i
zN4#x9{FLEtx~H1K0?5JKvYk@1RZEu&auoQ8t;$FcA8o0(Z3_%cG(nW})lc63C~s%K
zds$k3YI?%WY})cpEiZq2F7>(ITR2Qf;4&I2g^MNGzUT7Tw**`flE)3?W)&Zyb#17-
z#$2gi8uFj<`M}xTy|Zempx?9niUlP5w*wuB%bQvQ`Ka{&8SJxZe4~O8IA`&FnaInS
zG*=MqpbNT7@1Mi>><msHx;{8EeMw7GLk9=VV}k(v3+iS{LYIxptHmsfQ>Ioq%f`9=
z#<{?2-MwukF><*eDiog>n0ziBUtOI`E*~>2lgahXhKpH4LoCu!rVC`EZfjmFrCo4L
zDi*?Y2?d!bx#G<9P)h0s;%joB#F9N7iMG8yUk#b993G(CzC)w|1W@vc3&<9H9)1jN
zLRV@pj;(vT7VG|w9<9l0IlMR@;z?1a(H7uH_>KK@VizsC)9h;n?49a}+c~B&4vAnr
zX1X!+e?B#W=+&=y^L=vdXu@2(X^|Mw7F-Xtpss*mX$jgbrH1hJxBUuFO_s+$Q!Asx
z8COB2)h7FtOA9g9q=wHJ2d4S_-X>2~B&I0hClM7Ln(vzcWh8LE)N7%a2;#!rGMn50
z&$+CnG&J(&FpV=fqtTkMcgE~KhCl+LGk{GKyJt{Vn9&|9Y+`SdW)?g+yb1H3h>EyJ
z-@VcV>q{rKDKi#J68H@sNyX%HR_OG3YwIwb*;%59$)QN?Tf_3ST{|A<zNGe$`_}&`
zS>%rBq3uN+>jf20k$A<yv0wS>x`_XRfirKgk!NtTfp_v;qWI|BkFLQ`One)XVSWvK
zS+`U66o4zE$cw}^Fs`)vGW$6{B*Q=JHYyrb7@4eZL{Ca1fr8cx2C7sDl*;ERIDQrj
z<jLp3caW>*DVx9XqhsgBbA6%hFN<qb=iixfpZpGe_!QI^RSW_;ou~lZX7PBfe(SVA
zqe06{Yrl`w;us72$^7OwX6Jdd8@r@noWnN|1vHOzo6iF=q0Xo*ub(Z+nLdjEBjM(~
zh%+Xc(Zf1hiTUtYm3L2w?Pac}AfL4uq2=W66&diqeciiVc^0bT<<(IAqcGvzsdbp;
z-%aPtWKk_kO??B!Sw|L-NzSVeOsOt@!ZnVCp@X(WZcz@$^YL#^?S6_#rvp?o0lWwS
zJiLW~A@XO_MorU(`WZ{+>FuL`qgI0#4YT{sQ~TGQ7mmxO4JqGNaluL=eLl}MqoIi<
z1%6qni#g7J$Y;YGJn1aI^0V02^Q)EA9VN7MS$xVRNONultesv6l0l8hju{!O>#Ve_
zFq(TzfK;YE^8!+G0#In$s8nlz$3Bf*<W`vD)Xo$z&>KL|vT9ao;@CQ;*mAD`R*U$j
zI?kdx-t$t#l*p^6g~K6MW4qBe3lH#;oGqu->IT)=bT8wj^IU0UO0P)U8?)NiTl|4a
z1-P1qJbZwa1iECgG1enxg|Z<{NdTCBf|un5c&axioFzL>D2552!ZP%|maYPyyHw4b
z)&=E7653OU19*$<g;7xk5MDUGc>42fYRvdG=QDiB`$V#j)<Y|_+;rFh`od~Gc5ONw
z$?_PwMm&eb>GX9-NdCdoQ_W8)q7k9)!=+z)pW0!y_25aXqyHGcdn)MmWOcfSLMCo*
z2pu_i55NI8YFnl2U;x~4t~_<o?F)bKC7#-N_#d7UsZGFCDqnPmi2@@&%*jpiOaNrg
zG6Ag#A|VrKe;9IW99zqwNeEi_);RWI@aV)gTLKE104s>()w8E%`E@0C7DYt0EgDJW
z?lA0y7^PCKs>uBlQ&boZDwBZH5E_sBg;(F{4#n~lB-)5|ExiCw4kwPlosV<?gcY(Z
zuD!y6&lSO}0TZ(mlL-7P!}7SN%}rx*qMNA=rQXR-UPM->`XdEEI;Z6~?%cFgw^*uW
zumMFP!*4Nz4o5{5L#2@Qt|g}yy0kb<HmQ7ZhM@HgEE8a4Dm-t=04U7+?e}BK(vM)D
zfv;u}Fm9>UK02Y0hS<&egl{Z0-v83`VoW%%U*kB0{K>WmD@h$VD*`5OfqG<0*{@H<
zy?+x0_JwBtn7FhL0Ioy2EJq^u`6;72F~O-tvcGaS)quO;Yj9a$Se+Ze8W$tx=S6Dx
zkShyK)C&p!5-F9V3QQ?JIwm}0(gHeqKR|d%8+p*e1pcUOEmgIVvYi6F!7sGlP{DvF
zkXLpD(MwW!Uv#G6{V#$l#w13&-}pYg4VzQSaHz!KK^<VX3D&4MQP}z>zqThg)#=(-
znO>p&>5?0-0fRD7Z9XxdlWr*CfaQ@UNm;fj7E6~YqV9l6X)Pu{NbgxCG0ck-c1(!w
zp+oHRQC`HNyV-SPh+2(_`5m~q3J^bOhErGb>~0L@0r9PJk#EGIRi}{94%er3%WsNz
z{r0(cN!!^|2SYz9AHolH)t1v88&A#zAW($J{(366qc#f$khopDMI8{#<SKDYX-sr4
z+~{TKB7x7ruwq9=z<Kc~DUrvSiT56yoNP1CEf^gH&(iN!qjw(kOMUdv|9B$-?%bBx
z+_@lSM<&0WoH{I!F#x*Z8F(Oczt(eL?gO#{9GnDfZxJq%eEB8qiM4}t_%RrANK&gr
zdm}}B=Lm-O=mrH^5sL<edHIHYi?3u~Gv^EBDwcQ=U{G#lS9pQ`L5JEqid*&555j_(
zef%96@F_sIQ-w;wM6zY`YCpjc62;*u*P(-NgVX0d6MTD^ctLfJfC2+fA4bw<-|=~F
zWTEj4?VNK}x#nc>@|aF554)vB0INDW&1L_76?;=>G~4Sl&W)vXyDH-bDYC5G&zvMU
zXtjBZr`~9a8D!(f<NAp8M>h=Ps|Z!hL)$Mxqx#ZhTxc{z0Q>mjqgU6U(P@*<yRO7k
ze<0uLb|_6y=<J55m=xu;nZrAJs@GOysLIT;^V6RSsBktQ89`!j!gTw1{z)_%?W|RB
z-P?WPY<rm+>{hs^^uu`kfS%gBH}<SKXCJwlVmFhN9ki<0ojkuHGqk~<0+IAu4jje^
z?fUrE6p$5~i=X~zllOGdjHQcMsNyKI`OH>Uu6DMn6ziEyeCvEIb#X;kB~2sfHDlKk
z=g{!1VfxepI?lSK6~GrS_saVbZ1b7^Y3Y^t?w@Ej{M}QzK0rp9`{}%8u;>JE>aO&_
zir6sm8liM*K1g+NKNIV;?a}X%hkIr(D()stNem?~UV8fzMkhD_eK_#?eeSbc$9NnD
zo)#w#h;u?XACZHAY3=)B(7o6-Fx(9ZyJt*+ZQPOM!0j7=TT#KP_%KqWw8`uu>MueZ
zb7~?}O!4q{kTx5KnQguo24p$gD$t?yDwGs0p~B-S=yusZA1RqTs_GR-aB4`n%uI-q
zC*Te=LISb><nvV?Q*P7i%>Idxf5Eln^!P~;)R%=H4zAjs`?fU=?cXC5!=&S{`xXv4
zY{zCctY9k;y(U-(CEU6fbVBIJN|PcrITQyo#ROkGrSAiq&I*Hbe~~Pu3+n*4^8xCR
zcl8zdKe1qTAM-LTx#fw%rN(0-=koU8mz(*)d;HWgTFVN`z1TN&Uc=u#*SlCr!8-gt
zreJ(op8`KaL-8#h$-fE2=geW+WfmCa<$Z6bf&tbqS?eZ<^vyjS8$^9)vVq6>1yL5a
z+He9uS3+am3SEZf4k;_CC~hwI`GjuhYZ7?<C*a>xF7efynWEcQyWquZIlo!tt<h0S
z&To}0gA4pBxI1)`O)}mGXzLI(HZ*2tCrQ%*B!tFZe?Skn*fBEARRqlH_={#PDaTO1
z#Cfy%74Id|llGGkIM=4TJ`p@2o;qppABXbyAqDe8x>>;~pP|TuSiDdo{5n#zTZE#I
z0huK=eG>2~?C$ek?qK>uXZM~oHbapgaWOl=e1~@Pba&<D%Sr;2Q8E>)U!+PkmSf*0
zMw16{6TX4E0-Nh;mh;he`l&W0RjbPi_Oo)Z+1{Y{u`3)+?=LL~;B83}Ts2|rE(VmV
zbP|9O>Nh6tooX~C1n~#)|K+)Yi^7btQ`0!D)qdTyCL|-qJ9+7z0hk4W{Tsk_P_|bs
z9QF~j3)sh$7BsdOJ?s&^mJN%J;M4&(3BVS)^;Dwg*Z@z=0$B#mrGfacR1i<y>@Xcc
zgH!7Z&^rev1fN0k!JTm9{;xfUPD(CM%)$zOWZ}^Tgivb?hp%6xZdCdsb>cqF-okhO
z?(^HJ<2t5MgPrgvxZn2Z1c~LO+-yHdayfrGp&W*5qj0H?LZkHHmT>v*-4X_+IAweh
z4HCplZEVp|?POF~fb7!$TU6_WrfwQ*F42oce_t(WuwMx0(+`jhGMT-D^nG8Uh&jgI
z<SUH?+d3sKfc#a#I{7yAomRSdY|Ug>wEVC0oG1^Hf}fu40HbYSw;YXCV^+MrQkpAC
z7guXO@%Rh~{SqOPTPZz@Peqp6orjji>yk<NL1)98rw;$J&=R@;@GZod2o2q=#e1*s
zJa@TY`aq^QNHdr{qqL^Ba$3tleSd^<ouDKeo|#6WwBuZd*wG`6L_7}=s5}Cc&ix=3
zgV4u7hR1;8_Y=#$$mFb1F%cCRPCwyd(Cbc|(5YE<N|W10l*C*={7LddB1&p$Co!)R
z1W-;!qpRlvUKbU&?eCf9xHh2t{}SbZkZhWjrgLM9&%Y)$^HlEw5bxNO6Un*L?wa$>
zdYOJW45wF<l;EqE6Pl<RM4@^6D}(LI*~pC%t~n8({JW#7waICAHu>y)jp=g*Kcc_`
zN9y@9vk??{53DqJwS;HQ01Wxco_P=wu>J<(bO3t>$$R?%yOfEaV2T9IMjp|vffvR^
z`hw|$_)E;<%S(a9e*2>HA}iE=Qg_`sBbSoNRGj2t;=`h9C7d7ewXLB1y8w9^hY8YS
z_GGI+m)$OFOVBSn)%TUdbzKY-A27^-s2P6f;I!7~OnL$7#20l{CT1fhO$%qc%@uy7
zHAi0+O&t8JB*1b4=v{aP0bZGT{6JhlZESk7r=Gb2$T@gNd3~AG%16h5mfTS9piGB2
z8<3mmqb&DRU2JyicOE^IE8r8K>Y9ZI<uKovE}>O6Fs;)0V1x4Dl(Xgf`Glbp>)LK#
ztT1ula7r5V99Ew<_G}PeFiljc8zQmw;|d?392+M(Bf0sX@#YtP@Nw4!YNUm5|5&$O
z8<B`2<;ZhbUIm;QvO}!IcXGLO2q_eK-7EeANZ0V$Xc_2n6XUm)+rLKjKH82c@zp+E
zE4?mcBt^=@5FH#h#rf#_J#-E-D)JWTf$Yo8nh5Vh>7}K<V9!$332rNn<Wp}DxknpD
zB`d-)pFcnvIIeL3PjAazP5}{}fr2Dgr14_<lc|<dV#vro*717I@s5`YE|&G!bojc0
z)g`iNtf{*8d%2*Hv!%@~5NKJ`#$CCm0nCxjX(rshK+PW&DV2v14SP)+Z91>r{IbdK
z!tMUyP-IaU27-Ud*e~=x5w<(gAxwZVq3F_5FrE?;j!l4!Bz%7u7Juu4lz;!zSbN?;
zGpsfa$VmZj5-@3QShCNaT-zV+;52i5^028`6_(W#*G0+QOfN0tYP_6IKtl2%e#cIb
z9@Y-?89_b#j09Y5LM>4j;MB)wej_a(NW#d&r+h^34tX(O<$ACGu&VQ!FcY-QQAA<l
zuJ!}-rQIG}dyp?HidTjdOtsr74<I%t=uiy`-=XU1suv&7gs4>ap5_s6!qtKBPg(hr
zu!Pid)kQVGu!5}TiY1qZq-+to)1MECd2$6CGeI(?9X!J-(9V&ryLbtjN$oxWyn#1U
z2>t9TRl%U`=(gBeE|<Zj^iEL8>Qm>0O$Ma=u-i#e6@#s3ZJR&gP}(}|)>-6ppQ!(G
z%S*)Y4?C0iCc@_=NKQR`I)eR)vmbbpReXz*aY=@y#4MiD7ynnog?|`L<jxuyxa@v_
zKB9|ohNy>uGO%A1%(ApWK1aUrvN*EU65AYSKe%wyvMgE-Br|0cEnU4R#P0pny;I1J
zSF*0qGx}&gLc`XfaF2`hg}JyU;j_Xgoc@urp3X9p8)sHmPdHTX==2-0l*`uo#qnY?
zqg&aX);`65)B73|QUhWlsVU<J0;KVZKZV#2wOAGEIW69Ol#y}eGX9;UNN5!={M9A5
z#nfPs<k<Mh+_H^^*<Mn$k+UjR#HLCjZyx*n%!L*_unP~Wui6j;xHpAMbD#F6UO-=w
zKospK5cX*hF1`Iuz&?9_>v9~O0{1t1<d;f>O8%$wawl-E>u?1I42Y!lruUiEeJkLD
z+=oP+B$ve*We%*osdeKAa&ZSKzH)EZ!uhPksz<h&WK0BkF~E_glZ7F2f_t_giXC;Q
zyMl}0u#3;7Aw@aF83#Ghg8JpdUF1bkG=}i$$S=~<J#wiW9NyNR+_qDfNb^Kz5)f4-
z)~zG{e=h*xrCU{fI*bFW+kmCc`YV5nmiMCqh&uajJ3vAUR@$V$1`ut!_)xHl!a;v$
ziLdNlP*}h00Co_a69is)zOGc($}pKCp)E>G7*mImEg#V1bboP?<KZ7(llw66yT+Sc
zA2NtPpLyzj`-s3Ip>snd7;pVEmKY^MXjHZb6htDzZD);Z_m-<?9@E@dp$-<ED6ph<
zcpzDGFtGkFk{7YQ;c~BB5(5#+wfugK=W+1uA#Cue%@g%@#^g_%NF+ND6rdZtlKlNM
zJpDg}z7iHBtFhj--_qvRo2f?I4D1+r@>t?HE}-~X>~nvf7ilM_E{^{5alqpK=C|#)
z-$*1bul)d;aX$YqoHwfT@*N~;vA^dRv^UU45+nddrBAWT4}TzOWSAQc4gJwk71}`c
zQFlxUVy#<eF7C}80DKHF6It26y!<~EfQlA0=KZ)RV&c2kdw#^Rp8);^=xg=Y6g65_
zp02hI|DD9CF;;x;)7G<P+SOWTTOKM6fiW@SCt1mNAGFU*G8EmR4UvQ{Y=n{E$w&{7
z$q;$HHBv0UVd_jy!!$qLxsgPvFI>7g1<^PL!2yLZnd{lZ(`UuyOXb#Qg&2!O06e~9
zGwE0D@Vz3%Y{LI6WQmKXKmJqNT>d+;9@x9GG%8-s4=f!T+Wfu(X(1%x2bPPxl6JPb
z_rP3skF(Fx`lvge?h6;tFW{YwOR?GMxfA$Sc~^_ynox(UKYv#G)wyol1tm`YXNw}!
zD*&ga4=0+=RuuLlWm))1qsjfGe@-3S&Fr#OT4Dn6p7AaaNF*rYp(+k3pQ6016J*b5
zec8R}Mp7kaa}WW=Hj0RdpjlALLkb$0?&P$GZ`lzOgq*K%fN;KYbjL?PXk8dsP!a*y
z<$z2uz~leX$}tkzP#lpKd+Q+gA5!Ah6)c@&a6yuC2>_mWYFQ<6aaWj372fmLpIym#
z#bAN_!i|}}x@74VuG1?w?SPj?V4>mCx#KHTAdCkY=gVvZaXFzp`Q5vgr~4JL5M?ok
z%M^Acvyy)eo~#HKU=qi-AyS$C6hnXKF~yG)r;<kZ0OwBE&e>i?KmYy!;)R};Cmmpb
z`*wBEfe6&l+fDj6uc7K&*Pp5<G1T{JtEJ*&T632j-MDg(ew)RTBqq!-NZ(@s7%WPv
zbkk@>4_Cs-L}sRVRA1~(PA4X|6l)6bbig&Xs~(%8v`P#U)4mhY5*K#_G0DwH1|yF4
z1zLeL{{<OIr%W`dFQTULpl1-}ZN`{;J)5UeEh_me@u$KiR>1gom8h!%Qk1_|yx#A@
zuSnF=s#@saVSe2{62NK+z2M+US<bKk{l7A)O$M7cVP3@C5y=Tx+LY+4)DdLq$yOEH
zc1>eUVq<PFu8*-aHgpTeOrDbK#g3S(V}E5xBIVSaHQ~6C{_|Zar@_g-v}Ecp0ezjQ
zU#NzFSwZ>S3D>&R5U9F<zW<fF0yH5{+bai<51R#LBI-Y=euPk__r)Bo8j!WMpND2w
zSdn9ZHx`i77AIR`fv$QSFuJDc(zSKyw%7bCV>Ye0fwf=2b&W~|yChOQNom`|@9YdG
znbxZJYZ1zty0EKVpXs+u_?9y2Z#Ooh^UZfQah?C)YXS-7zsUn<=cn<xmhWWsg5y?z
zk?VpLan{&IOmOFOnS1i8sl)=l)5JZF<)oc`G}fy5F{qldAV?qT4+1pCOjC5UvxGQk
zY&6ohoIeOT;0*tz+BMo$+Q-d%ESi#xo4Rv84h{tblBy@}dv6?0_r`looMG|YX0$(O
zY`*lq6b4L|qMK)S+_`wIyyGJhA!Csk(&iTLQ0prh=dmN;<0N6ap$P6QZK(ZGAy5l(
zVL2xuI?WY`x_(`oOY(;C50AbY6BumC(2l=+VW$gD8MABD564=<duO<vDY!Npmi(A$
zr3bs(nP?1Xj1$@oV3Q(84U~);SBzS;5@H0`<0c5@&QBP}>Z{v>USqu$b!I2n3e5P#
zx(n6LhyF=rhD{iq*!kyZ56g93<)Q55yNAkqY`X<PPR<##V4giCqJ$XP(9^|P7K8Cb
z<zLwpw;qRcOUfSe3aAa$)7&U~HU{n%Ti1+t?-(r|jLQpMXF&%BQAiaR#i+DdAwh8(
z6;(uUD{$m%RMQHbfb|)ME!WOTq5e$O<Kr^o_ASx5fZr$FR_F27pBUCA?8~SZ*s)W;
zx=8K<sk62aVexmimme8lH1Rk-?WZe)9}Z-rljGH&+UZIjJDHC-wy)%>>S|&*<^fN>
zP||GA7kyd;y<1SN3XeGwDw*zItVhk%9QrSPRQ@#g%I)mQ%Nn?q;bqE@7?u3?q=cWD
ziYx5TP&jg0VR<V)(om)=qKX5SyTOWrfn;RS9)vhMnbP{b%#)?|{STm2Ns0DdsgQ>1
za3fV}-birckwC<OYYdOIe{;CCrol7nu3CwC^1`PI)9E#@ASwS38u>c#JyQwoouh0#
z{?Z#-IA(2_|LD;Zj3|`Y*aL&tnp^YK<?))+<2rOOW>|LozMhXJB{@E&ZowLp!A=nE
zUC17rLp8eZ0B*14tzRu&B{89lkdh_R?RfSTuMj>&3~BGqQ5hAVJ#lFxchY9ooSBid
z`jh_A7TR#-ZxM?giLvg=WpYc!0N*1b?D$%p84YDqdax?rz^Mqq)#@`pp8b#67>al<
zG%B!yfec|%8SYB=fOKeFSs&}Vx>(c{Kf)ov7B3dTO~aBHX|H%&Ny?<B-+72C60=V2
zg8;6$>Oi)!UTM((%k`b_qxoLBVGi=dG2WgO!(`GxQlF?(_Y-kr<8k=!;ym2Ib}C0@
zE4eA52g!H>wONHmAdu8x{bM;(i$@vm3_%R_%|%*E(Akj7=JJ(8E&0L9km<YLbRHx7
zC03pPDt+jC!vHU`dy6>;N7M1-JJ;fe@xAWyN)?VTP;tQ{#6YcaeDD-rpc2uvx^o_G
z3>u#Cpai00xL&Yv7erI3L>sgl`l>@NSobVz8i4=B6xbrS!h=i8q7b44!J{iSJF#`H
zpmf1amjmi*ylP=tyxCv>oOBcX9r3zn|Npvs%crWlH(nSl5T(0IT0pu%x*MfIN*bg)
zq`Ol}x<Q&vO30?WySuwNi~Bdv%=7B}0mpH8f#Y|_TI-5WTr1LM?s`w{OATtq613qL
z{mWWTPF(oPzv+7|7m=%fh?;$DW@QUl#kvL_WLp;`huA(;a0y4hWws*r;?tIUM~vmg
zwOn7&N#o4N-Y^#e$G0S5{h6o?S)#W&OBZ{$u(=e@lYLLe3fGI?Nv?_XdS0NeB}Dh!
zPO$5}QahT9`H|9m2P-yP7X}Iz-><>RW+>4|rYz9pte%flpT8<mUq?V(B-iSH;77(3
zEx$w<aCnIj4mHZncHyXjds)Mlyk>nAo6C-IS(oq!VnsEz!MJ(r)7KN+`}g;x^4QGZ
zuGs^adA^fIDz<kAb3x+v2m9p|K|^bwGZEf?m`a6{(6@UNi<#VYe~$Z+P962W-92Tv
zfFN#Hp`ZR`IZ}kcXSP=&0&fg%y)zs8Y#G;ErPB?RR@Lr31vDeon)j{v|562hB}wdo
zINaA%&v0L_;-nG437ZFX9PD#j&)W@s$T~+&viP~G;i1N?Z3*||yn6WxQaF7^I~jp6
zdRk6JTjBM@)H>AhGhAMw>WOf?K?31BUl^0CNe@gQs%W&+rH4=sLn_@_OHY9Lf$7#3
z-6KX*>nHQeKkYRgqTAt=VQF{pW(y;T2p&3cr?KN9Yce7Twizq)Tqkcd+-6-I=k~%P
z21$jcu0oD}TuFgL>`nw{n|rw(3*R0zm1S~1lB9XLYHOc08Gy`v*$O9M;56oGMbi`@
zFZFfwp!#_5QNBAw#z1DDnvx=HBYSnDUpghg2B&bC*78-5CqeC3`-k)LL{1lxJ@P{4
zsTWbHZWLebC@z2)DL3@5Ss6*`vMU+-8#y+sGM(_HEiwiV*c04$0TUQMtoul5``PZl
zoemVNONGq!s%zX+z|BdaqPkeDbQt<&%#J&2I(dBhNFDMJSl>(p$33vTlxFwV)W|L%
zJBAKo(2nMwl`+)&b#34HB|EJRf@enHZ6!c*1HWb-Ue%+MQaXV9+q2!LLY*T};g-}I
z(@suaYbCgm)`hpSt3e*_9G?Cfa-21{oBz1uYjXSjbVK~kz@Es~KVjCFSfV&ZS*Dud
zMe$s~*qTToQX7u4_rPorfT^7Krv?4<vQ4q+znBjPd_c*tY5U^`W>OrR?)3h(jEohZ
z<n6JZdJLaE6l3UtcxTA8#`Xt|D>r1FxDcpdHAl?N$m$QS!fMvTePw^pB1q2$yALah
zAmQ)yG%r7$_CEvsbi~VD&M*)pr-6lm^>BMCm08H#&`a}eTE~9rBC&=rfjau(--CT%
zpKU!s9JU!hbR7P^M?%Fsoc`^5uc1!E=}_B#J*^$#2`Q6S3|nN)my|->fGTe#LMf@m
ziD~)^`x@jHCkm!Yk-e8InT~EGao^#k{~bB3+BhJaWgD7J%K;B5d~o*%r&EuP*}7?k
zs&Kc22VRrQTd@(i>I?fHch`F^jweTu+~Z%$N`G%G@p=WIYn4%7obhc;&))2p>|MDH
z6~FI2+e%CS2Oy#H&Bv&C;P5k@YAh*UL_2rm4(mcGSUnuNXPiWl4K9qC3gus>8apDW
zaKQ~THN;w!+DKQWSvq3|SyHI%n(|F^;|@%u*TYAYKV)x-@+Z%hOOD_+=kOOBDW8-k
zrz|#V_>RuHKPhDjliR`M1eoO~nRt*G)h&Fd<ocK?k5HaJ)skni_09y*=0hYxHK}#+
zoQy}0xS!Z&jJorlO{vMO1sE!`<+0m{XIkT709QqAj<nR%_$r(7JZ@t2T->dLT?TP3
zNH&e9pxA$D>&?iLHRTp0q`Hzz8s*hb|75%F(80~pW@0^?OL{}zROph%-G!oFY3(B8
zMo*Atpy);n|7GKJ!ye7tjXv?!>^S+W&Eww9IRY%qW+~{!>qO;w9ir-Z0ity&diYAx
zJ3Gm+_an{V;$1g5A}8d8h|y)cG<vr}u49I;cCRfR@NLekm2@k^Zet?hLy2TJujc|j
zvT8?oyIq&7J|w4wyWPPf{-s|lBNZ~{Z5C2)O>S64h7ZS^fYw6Sb~B3)VA4wOeDxN0
zAC6Y9Tq1aIS1rWQQmWtuQR>}B!#SJez+`p2Me<9k>HudzNF7|HJf|Z28jrCDJ9l?j
zgk&Fw2~D})rgj5p4A<dTf=S9=M_ytrZIRvfn}yfNz}k=zSHipL!T?wpB@<q2EDz2P
zETiP8uW&VlKN89|`lybhQDuAzhlR+<r4019IJ9Qxx*{UP!}|An5AFR1$7(jp`CN|%
z!`d^Zwjo@EQs#rf{8a><-}(8Zs*4pgrQ2wURd9^1Mk|Ral!s~_W#8Yf0Yx-AVb=jC
zY)8Cw(hR9Z+-PJ^<-o(Y?9$g+3#(!F5<QVWW4K^`jB`&p^aHywu4LxPJlo2f5RDL*
zs_l;DHCkDmwM{;u!?-&|bMk_ZuHg^(k0KsDR9exEAoF4mTUhd$4)ljHW%vC00p@oE
zgMXIjsr4Ei`y#9Za2w_P<1hbQ`y<+fwE~qhH7zMyKI%Hn+e!4b5|u;6v`<F$THXb(
zwXGel{0$I5^T`PLRqh)Znf}<iF84@mOZ9}jwkaTGKWcBW;NT}3`F($0H}>m~4x6{(
zb9;I7S}(n)u;I~{nDpL0wd^v^v)oWJ^rrJ<9hq}pHzCar%DqR$6(paJ=t?7)y=BEx
zmi(mHKDpox#SD@nh(sKMcvl<SVO5gYVNv8seZBfBrW@-;voiY$tnC{ET8Q!ybo>uJ
zPRozkS@k%#*Nu2rC(w&0L$gt5vKlc2eQ_oRhVr3_)nF97<c{qff!rzw;=BOMxI^cN
zoPBAC<9V({(L&qkDkH>vHQYDdKx^IvVNL98By;0qcT%tp+xek?!|(^NSM(1`xG%p2
zqqiTO4J*~)l{W^$$72<fKvB}=%pM)6C*H)Fz1+@56%{x@)ExGwj8=E<^G@0DRh(N0
zD&|O05;2r)+W;&6?jQU^BjMZm2yfgHtujH3hjA<~3Gp?@@doM`TSh8wQ*Ll%O4-a!
zOV!_f`j#;Jzf%ff-Fgwzfu^Qe1?)MG?*=Aom+hF7eLxBN9BX#B_x{+&>Ct*-h=Vw#
zP7_<WTe4Q%i^%(UQN>%kG6Q??BfwXTSk}Gkg_!&kP1p7aYn+FeQ%|6bEKWyWLz-~W
z6~t!$%k;cs0wU@8rRBP<!?&|9t2v_+o>5>2!*mAb>tnvQrCOTH_o$~18OQVr>X?<W
zA3fK58$PzSU*$tgtoxVkCNV22Lw%}87E_lLm@uV&NtqA0a^uxKZ{7Q3WYiRHxMVHK
z&H9!W3*Q<I0qp3`&5sm(yva~3{{Ev)0q?%O)x_fT-vu|@)k_6AG=^c#xNc5-h;F(Y
zcWdP<8KJ}*6U7<pWx*V_zf0rtYnymi&tmCn48t1z$M5%ksHHI3#{(#wX@)!CeDQmp
z%<5U?1T?LTK9&wO(A0yRiB+UE^759Rq#K9hNfC{lWye+{qLRSu<I@Y}HOJLaY?OG`
zKU5lC&lPh0qv_%Oy0PHaX8*AYC=VZknj@pLs*D&H@`O9{2D0<3eszUx_(3DcN%vy-
zU(m(UN%psVaDM9UA2@@Rv>3FlBpbV4^A{kO!fpe(aHYdnm*7j_bNOF#Qnyx_OucZz
zsD;Er3yA^hfO`FLp%4wEN{H2#jdd^aJ3*8nsZ^oIez;(G={LWfr=+J#Wud%|VE0_#
z^8@GXzmw)Xt>C4j=mw|sSE;p*zi3;s?4hmOphqKJ6T*->K*crPT%MksEeBgj-2U|3
zfL&!TGUn3uvX($!YL)Xa{aPCaYfHyl&fz<z;4jTfrJ5(pbG<^)Qi6U+)I5ls?P=C?
z?8-ptdo2IucU#JZ&>wI7VdY=POsBr=#LK$r@Ci&x<MzKoa_`k;=Z_q)JYu0_rA^$I
zyaE~>>V-!}0=`lh8Ts@evCPdvdPZb03Z9yJA`@fN3(lJm3!P#vJu}<o>Nm90zY|Cq
zv5Je?M{Pxdj+4rB5pulUv=Ub2)mEuNAc?cMqZ;^93>ur9O<i&KO`+u!jg$@|(=6&y
zm5S8iZh{zb#QDfV0Rkc_aErSiwB-@oH8-J=a#2D{;x{L&jNSJG8#S3+l_M{FVKrJ(
z)a2?3%MY|RwAp#{0?S3Bl&WZ`91&4L(R=g`wIsm<Ts|6^t1X3fwopF`3^w9*Y&v3r
zqKvS%B(cGDVv-PkDX05)(&D%~D?LrN6S{8|@u~T;M#d##QKNhR&74hJ)s5r^=X`(S
z0k4-YuRU1z99pOiL)I!QMglrt4qI|_sXJ}qYajf3&I_XlPeNtmN}u+&)`#5xMEaiT
zVgF>;gwv@D_t+!;cVw5X;7>5cA3k)n29Crx(-Ye~O03hZP)~oI+;_G4$4H}MYbtAw
zj*5^g+3H(0v`U7oTvz%P8~>cJmhWn{JX<fpVhMw~1-!HCgV)H*rwun7MN|>Cp_Bd}
zKAH7Cfc1zwat;3Fdt|Uz@?CH9qa*8~k?6nMSoT)ws;!qK@~MLEUH(a$As}^@dzkq|
zbQT~Q*c^g!2Xny2WF=9q2g<E)GWnV7=me9Ss3~_hL8Xo9U6n&NJdHtMw*BGy)6P^g
zr;xb_W%s8Uo=<F69eam@)-vP{=WN2Br2ruW+ea!p*Z(emu`4of2^E7$*jj9LaZ?|!
zL4N5;xpfG=iF4OVz;9`*-H7A0CV9@fwucno^dLKE+TeP^d)I?STS-{R!okfsy~T}B
zbNMhTrJ;u7ad*KoN!}0_WMCv%zrF9H3Bw=2@1QOqpl#3@6AYXZ6`B-XmwiB$s`=TQ
zHDyw!HFLz@hY1M<Fm6-TQ!Ht707+~t)8ThR1UB%5m`kewNfL`oRyj?4)RQ&IUiC1k
zu~zarElMgS^BhR=&#u$YN4cqzOT?#`p0PFY86Fo@{QT#b2KwTJ8^&0fY+0&iH(BoV
z$fef?e<n9o^y}CwygG=vsyAp#Hl=ovUYkMgJ&^%U1i~G4oUr_u05Mn9DIA5^`)u~G
zrT#mnr<ps{jKtEsqN)hE@AI<G<^HW&nUtl=Z{l`8X%tQU__k(%VWv-g>M}i2f0cuS
zLxhFzR&JWlz#6Tn>?Ly0c4b<m^>fZ>r<wbH6Hg+4)Yf6~UlZg?7&sA>P~7%I%yYTp
zuEX9HGTz`6LwL7SC~Yk6JsE4@1pdwwn3Sf1;|~%a(OLbEn)KURefSY3fGY)PRqtH7
zTtCcuEn`3MwjAp*lPN{wWh<IFuY^Ed%=YN^E#`Sv!(rgskFJ!wwhcZK<~F~nBXhQh
zJ8<Xc{NA6cOX!qccZLb#q#i~y=8bxEA#dvgy=Tj+JnyG9D=|^y*U(PmkDoJ{hsV?<
zMfPB0Sl>L~m4oLW@b>L%avoyHi(D;S41D<USA<y2lfP=5wZ1ZUEsk#$U1#k>+!bpB
zz~yKC#EkbdQ9_uClFFuq>f&}Aw2~G9-g6_t5F^ITgarWefH+P?)^eQm@UvkSY1ES2
zZ*<Z}6r(%iX4vN8`p}v_vty)Z==;K4`7ptz?<(W;Ono94;;Bfql^z70@SV_#SUQ4z
z1!mu96n%Xe6>~+H6`OBR$F9ewx@@EIgQsyAM}K>#sBKnJPC|q|I#@YK?pMIrWB|m4
zBd_bT^Q}8wKQAiS%E(E{%*k<oL>?PXUVGME9=Vqv<1zH46K~@)s)B@>0@|l-^GO@o
z9^)p)qU*i9t(%z=H=3{^;t>d#gtdRzasUskDl}V4D!&wVr4tlu_g+_aEup@b2JL7W
zU47V=Q&Z1n7g-)<xWMLKw;=)<I_0QR&70OCkHzXmLL^7&mqv|`;X?-EAnZREsgPd)
z%4Ah7&&)Mrv-J;{j#3NZ!BaGeTpYs<noM`pnZF@!XVVJ{?Zf|}6*<=h4Q$bO+>PTu
z_}3aveS6}PcFim2t|NYw5nJ0Jt2Ti7Ks2USLEwEGBaiZr+?BNE1$<6KOF}_TVOt<Q
zp-~c^({GNv7{13rJ<Ou?a?nJkKR1yE?}#^(56aJG)2|A*3s%y5IDk>W*I;#Y6O<K@
z%fE0<jI=j>U0XMI6p<(fI1nl0W~ER|??nzKnkt1Np~A%{_T0G)3iEdDHiqAJz-<L^
zS8xsO+gPCS?9qp!Bp?J;PnIC(?MS%`*A@B&j1K*nkp%%U?3XPuo=5+0=rnye5P%am
zvPJFSBy%Q@H<^E$vM`c`Y9<||LGQ1~LdZ0Coxt1stFvmo^<+OhA(hAy)y2=>LNe`)
z{Ni6OfZ+P+8+WW2+qMp#mM}jXwtrY<!6v(H?#9HsEPdjG;g3vf5SD%$-cP~yq=7*z
zY@LFHo3(vVg*>Ly>faztjEWX5`F+K*JM#Kaf&d>J6mol7dgQ^Tz({0|e{OdIxn+&P
zaK?+mTHPE1pGG8=qX7|Gxkay2!3mVp#`9Xv^qFQILH(p!`0NFw#?JF<V4QHgXt`p;
z11>lzVWhl-z>O1bP`+NCqVvsTU^4sm=3LxloSJNvYFGvGLfDy`6jdF{yXpQJLlmH;
ztO;H}rwqse7QsZra);;z*(X%%$E@E!;Iupd36@&L6^$7?6H$`nqbsWq$+>A3d(yQ1
zG>?lOR1>2HBbF$=-$U8zU~I2I0BP!Id&5@4ytQjmJ9O-O`@tu;25s7E9D`Fu(+)5s
z7JNY%QkKWHTA8ANLWO4r&~Z4#MLy$<YD-7bzBkPJD^1%&p7qy`Qno!)>btN%1x?-t
zHF1mX^Us@y6y4C>_?*H}Eax<mzWn&O7Ygy(wUm*%r?fyrRqc*@%ly^SJ=%e7?o)wq
zeeREK(-L33aX+PzxhhW*Bq!?<&^3*w4WVav!h(+<7AiwdM;87>bR7l?>PS;MM)(&r
zE>qvE7Q-jBD1s4TGr<DqRc^DrO7>Y-R7Fc;EnTt!h^a}TdgS&M3B2w2tqC}7(McX}
z5#n2hu1ywJD9wE5P@5aQaN*|qAeQ$ITB&SDNe{-et}-A1A_~Q8q*6$m9)sCHVW7|7
z?uda8#S=9dFCeA*;I3(6Xv;Q2!vZ%ChanohS@eZfwd^krPr^A9N4n|)n>gBzGLJki
zHpee0_JdEWrE9c+wR^--Sc~7@8RI+88*+W9t?uP}BF2Tzwgh3I!5k+aW)b;j;s8Gz
z#d(_dkTkoB{6hD#=nITY;-1Uc$Uo2&^baN3^q+NO(ymJ_xGbp*n3`wTUJKvzbAX@$
z15rK{UWy#>5I?gryKF^o$`Cjs9d9TQyrLp=`1xb2qO;48BNjKkJ(mlcz;2LInkBNt
z^+n3mKPN<uumGG<goNL0okVBTKa>6(zuiv|8=nozepMx)VLmmc1@Q^G4$j%oby`&K
z5BmV>7-;7<xr=?9Z&TBrb^?}{WG*CuSibS4FZVHpOuu@6$t|Dz<U;q+SH3}EJ1+v6
zG$IBvviB926yG*)_~r*0Qft{km3$u$J7Q`wEhq2H8*u!{<4qK-Wb}7~vlez6;s-Tq
z(=YqGw%>KIpVjPXcMR#=msX%g6n`5|9ReanZmH)x5DTM%K8`}&og_%fbG0Pv*~1)D
zT!w8H<miGkZG0mtH1SMJm)e)4C&vyhY`qp=hsmb<@qgU#^V)wc7|=46f23BQw!r$O
zgybW4MzzmguAw)&$nE){or*1mK$1Jll}tmhucgtw`-VZAnMXWjSRd6=@5D;<xIM?&
z@~ODGVnw=?-NBzw3gz~9Uk(KRKP>>7X-zWc4+7bFVE@RB=IYGp|6qLD40CQjrPQzn
z`G&Cj4acDC+YWVyJ}uwov>j|YR1n}|1v5`b;TyVOEBne2P4^SAY)y6M>JCHw)iFeQ
z4~e)nuSV1!DZoe0nWu}`<#hRoCZu1uXGLHi*0+T;J9IAYbVn*xD<J%3O-`dphhZqa
zFRWrp4_;niqd27Nlu__JHr37l11q^V;Cnu!yn4n_elZ18aWr>ZPu|BjQZ05jo$ufn
zAk^F5diYu2(`C%0HdXI|ro?D)rA*S7LufDZ9~btYJXe<TLzfIH#>W90Vy|fv7N#?O
zAwq}e(A+vDE3;by8yP&L_E~k#6R(%w;GAn7%rAFP)~$?-$onk#UF80i(rF8e$z&V=
z*$w9eY<OM{w4xXq%GcYe`G;g@RjRpeOYakj|M9L{av@6ApAV%DNy4jDP(6KXC5z$q
z>GKwUUM?{Y>bl`KD~9&@rqJo?n*z71><@D9B`rSq7>eedbY|%uk0&JFOkHwN9~ErY
zo=jyT1{-gl6UR(dMvm?x?xRnX6wH5?Q+Z=Inn!fkDgIlj0EaMbV0L!^d|pBfpVZD_
z=V$r01A*!YGsLI=q)xf{NB}U2K0y)|yZf+Rv~ME`rR^rnnyluc^e>Fji4y$*unT&L
zSD$5eEriG`nf#2&hz&CatW0wtW|&}xImm(V;r)W<LHLfM?mG5Z`V|t%g+)uZC2yoz
z4r-AY27gU$*(7Gif>NES2BYHj%Q+1*RFrz!BeK!3{1yjW%>^+~N9cXZL(iyKj%$5{
zgnVE}^MIf5J1&L@tfxKnseP^V`8DWDl;naO-NpJI{MDoJBX&;w9v=XPhotOx+ph}0
zDIegBanyt}FMfFkmty9$jQLRC#K1mov5{~xg_a#jFfR2dQH+)>BiJ>a0$2nD5I2J*
zlexatlYiunOxUf7J9JM6t#bBcI=*T#Xo>(@<wt#s@`}sUz>0*d?8mTy$cVh#1J<r|
z;!kQctax#yK_e=Hyx!p6&kWtvR*Xlg7L%&hqLkf9X;|<pE_{xlSRVdb_qLq+)E`I4
z#_8SJnDGzoHvKHQlMu3DycIf)B2|O19x9K1S*4IOc(UN|PDMc7SP?m~siPL&o8RH{
z{<&8y8)^*;t=vLI)y0|an>isXx7T!W?W4Q*Z1WYi8gpARZbe%3yx{Q?#t^M#-``BO
zEI%v$;`L^Ew<#2&*N^#-mraC~(lG?j%(Y^(1^^qI<vl1a+eIMt1gSV-KC(2ad3zdw
z*Smg}6APDyQhM}<3om};SDtA+tZSP?$Kv<wKVgGg$_F@vXGZ9s0^ZTC`A_<9V(|Qb
zg$1b~)evi!9$nqdC^UwjoZR5BWpcV|fC~}Oh>ST%ZTy?e&GM$T+7{4T``Sw$RST%l
zM~<=&hm~Z)=Yj#;Ahqa}(_e)uhK3&uo1DP=Q^f$SeBUz?pxWPLCn&8l{`mP!YsKi>
zs>@Ii+fSY@{*;H9&QCxc4#yXdgTQ1I`E&_D39_YKSd=bxhzBNA)p&BqmD6#Gp}aDL
zcXnHQD985r+_(_-H4}KMR!8$!cS=4K=Eom<cD^BDWlm`_707ICP69<^oRwF?dQ&lg
zKwQ77cR1BxuT`cTbGDi$r>Q&7DXtMre*d^>+hnD=JY<kvd_eNX08g=JEGUSM;3WL0
zX&Vf6(gcv>p)y=ejUIAZgw>{OlAOFMB4(qmU=)o+nls1Kd}A1zGQ|_i2G(Y-Jn_g4
z!MoO_&_X?YaLXA8;z50CpM^F>jVD?It6qb1{)U48LxAoeBqL4s&e?M4j<?~ECXY(q
z(8l6TvL#$HR&-}&Ong{)AE1^B4AQOj+i@J(4yg7ueNb{kNdWre=f?zK4uu7E@gS*4
zFvLbwz*9s-F=t>>FrWOGyfM#ASs@Ck%44n7<=Y^#B;t|4Wci4A_Q(+6Z_y_1mh#H3
zg)OveI*ql0r0{LJp1P;$5mgRdFkuAqK4Zs7%qHP0nugUAraALDf;2a~S2sNNMfT`3
zh7=z{e0o%n44NI*m8LEW8X(W~)q4<n2iVr3uih(2&(eW=Me%d+V|yn26<hWw;&xHc
zy5{?1=a#ZSs|m7eEG`YYd~fO-Ae~l%Tp^YaZkPzKO7`gr8LoCWrb)j->0ImAvotaJ
z*5>oNM5p;zPL6V`3zf|+m=pSuPP<1uutg}~uTK(d<w1DUUd)PEmoCHOMU;i*eHJmS
zo|?cuP#QRS-lG5-D?-|i;V<K@jmrW+?k6o=Lx%b->wgl<-!dq(yctyR{kte?x6%8p
zjoRmZ4Z9DJ2MpusgKf<WXqvm6w;4$8I#hrT*fE)vlsCo%0s_$Z`Yj!09xlUy{eSyQ
z;sv1iMta7A!!m~Fn9jz73I@QHLy<)5v-i@FvSsG`tPK1|H)0sCG&f?xe@B@ryCdsb
z{{$RcWCNRHGY0E{8G22w=N`#`O;u6#6je@B^RESq%_)t?gI%<NsV~TIILx&EA?9b;
zEL4_w*84Z!jCi3zDyH3uF4Oy`r}vi}E+&)65jP!__&<jS^r*n-0^#EE-(Sq?Z7*Mj
zSx@&UixQvI5k9RrN~2?A>j)fky~2H~j*W^cKvWDttE-AVyc7m8AwZ_A>&Q+~^PYcT
zq8P0Cwooi!b6CEiS46jiswLY7PN)23WS^ANoFG6H%(Pigl0!kb1l$J`o0WLxIBG06
z-5*=UU2r`H=yX7BDKBjqq?1cNktAVb{<bs#g}o*LE`tTTRUCqChDnOy@xb_=mN2sa
z=`BPtIp4w)=sgf~Ge&0`uyea4Ca-zmc>>2)hCY0#h~~7?Kab|++`HSiq}iNUII3uy
zee-uP&!ePd{%=+ZgzBIMnX%&zofRMr{r&H*wPlC|?sl1UXtYE7M%w&uRDVxE-zTZ(
zj6h#4M4&S>W;N^4$!XFtyZjv<u`=yE*w1OULU)uVm*ff^@uy6WBeu&gO#{r&x4Lm(
zl9!4i*fHKrwJ`Br66Oa}oF%gG%oKWxE;(V3NKzBsNjVX(JTp@Iqi`lU_A_Va+8edm
zVJcN<TO8|e8lXZslhD-hLZGyGzxI|}i2xa}B9&>30K*I#(vr}c<t@DEq!9rDQICbe
ziq)%mo1>}X(q=L#<wqzP<AG~p*s%(LdF2K}kw|y(QPGBCL%oKMRQbBRb}uU`PheQs
zX!CMq$(z6R{DvYhb1~68V570CmOhJ#q-yzOLAS-+<!L?4osda(>!HR|Y*M5aQQCOe
zO$7rCk-iC!2)XI~?V4^22}7wxj;%}5!adBh_;skOwriN8kCSw)Fr>B=x>ssR8T`I(
zlpanJ^kiaa#gLdHk0;oXp#qw#%y)$IrW*R|(Vb$eehS@~T1iRrwI;@C%n1*r6!*3h
z=l~tG-h2%5A^m`u`o6tdkGJAFvV=zSA2iqey8FAHa?cxvTE|cPlyI`vYVTpI1KOG!
zkkAubypQXJ8s?Bj6)w5pS{~5>lM4cHa1g%xkRXQWeu8U1y=Xn5HXyAm7G5^$a%?MC
zkQgYQa>g}hGUak(#CJJ&O?cK5q^TUaz0-6eRFoVGLXzEBlV!ZZt%B;U6*Yy{+w{lh
zTRe5|RT!8}2MmmUR^9UidIIKm)`X`|*1^%DU`f!$N<)RHT(AFyHC`9F(t+dxhQfl_
zOstjCnzZ+h0l#AEHg9lqT0qS<In_U#Vbv$~ow{;^Pi|nsnb?A70tb;{o|tH@64Ca*
z7+znHO2-?5J|kn~vYU_-cW1m!3JeJ<7}pb_kONuc6Wih6N$tRzs%J8<ZTQ#Rh(7}>
zEb!U^%?+?Yu^#+2;y~oq6?T3z7-gtKy%bh>Mgk+9x8R{Qh>(v4vTU}C(;rS&fUdxY
zGAz#=loPXE{t$u|wohuyPb_lS@bD;_wm-8aN`w0A2m>@Jz9PEM=SmMrvn2%&{8uMf
zAk$eJ8~xJMd}P4SdCk?=H+9l))Ji<gXkPFh`lt--A)k)W7`nAPLEH8$+T-6fr4q|D
z3=sRNRtJfsv-R%KXRlVRc9{6KSM+xA$ZSp^XAOvisAz4Nzl&oZ5nrcHxCtQhUtPEB
zrbY{`Im`jPHow`E#mX4l)c|kjH3wP2`+Li?)|o}MapU_Z*=B#A>R4R?@({AxGCU1)
zj^PPMlw8WNd73MOPkHI*Qwf~xG1C&Mp8`3i@cRU)nDt&rIZ&%^Xm|F<NzPx*l<(Ng
z)m!eooCkzpXys<A8oS0(`oIq|Ca_psrpn35%)zn_JU4%?-|^Shq83Mz$>EC?2uT(Q
zi3Wk6TEU9B6f(J@`3hH;H>j@fWHzW(-Wb|a@JFHxmVG{Vt8kbCpWqfgT&u8StAf0n
z%`b1TZPQ^brtr^A6X$ELFKMMsttX4wn|Q=%9pDaRK`;kY*=(53AuOWjyX&b)wG?6&
zwbIf;=g(Rt%1-R_omyAs<2zmz)w0!ABl8`5Q23E`vr&(eZd+Kjha`-Sj>&y7UQmQK
z^d^j22639I>+)%{RZn<^4~%-x>4}vMqXGVFdYo%36?t(c3{=xpVKr3<@Z*lu5#3=S
zlj5ft<{6t)#ZN*G#(d8;m#SkcbnP86CHbmIIeQ@SmB2^tDkaLQLMy72UB}^-?X20^
zkbe&Lv23|Af}jokQ5p?j8vVl(2UUXuwK?)eQ~)ayGnDK_sJuT#Q`BXd20{0Z4d=F_
z^B)(>puAk+U%FxZXxWq)@mj%iTmKA?a(dOF0kf7(4R@_!^3=au8iU28-@$+pV?N+g
zl|;o6@(iK<^M+3~L^-AUT$+uVGiyI$3lPawup@*r1wa2iu&aW2PH~*67AXMUxvu(?
zB%xX|SBWe+13TbpzScztR1PUnJK(glfML+$BKMuBlJikG>i^r|PHpx){pa=*C}`=Z
z?s7D$r=X)!T2v%F=XZ17SnHQe>n&X?r!a&U{VmSi#K8SP9{XB`6d7Omg|j<RLn>B(
z1h~RsA&F4C*=xa`yP;sG$k^!&R%^+Yfiuk8Td<T$#gas&lvEOXa65@WXoB=7lr>=4
z5=<X>Ds^FY-uJtHahP`6>=`a0s+iN1{R^l9fUr{pQ2vQBsp4YJv`_Q5w5_^V7y_w|
zJYz9bVbfa%c5{%2?5EeC>0(!o?-c3Ac|CWs+saD$TNAxX?|3M@1yR&caEEHDJln0O
z1f**j{9c=(*09IapWl$PGs1|=(>VktjE<>`s@e1Jyq1oN{G9Z4b)6Kn8*kA8J(}3;
zdwDwm=v9fw{=f*lSi<DN#wme5FJ$-L(bWx$f`u-yQqQ_jhW;dnex36`uJl(wSANIQ
z1y72jl|~XcBV`43?7c^b(TaNgy7Tua@mce=s6a-1*(%U>z<#1AjM_WJjzvr0{FawM
z0>3U|6onq&#usb7dp3W>=pH2C^+4dYbgyJxVEF?3dq2^s@htIoY4pf)FK%Z6?hg?Z
z24y5XWc1XmY3%4jC3pCDZrsVZ(cPu7R#`n8>xFFc3Z^@s8R36V#>Tnb{MkCO1dMB5
zZ}~GKrd391bk6qEFPw!MQ7g*}EI$rvjJUR>_?R3hTz8S^DFAjzx{~ZMYH#EdhTo;t
zQ&8~nv3dm;MrZZ{38dRVFes=hXW7=iC{sHvUG5!=APSgr1u*a!p|!}n>jrSc=dhu8
z&dGBBQO2;l(8QZZ>&3{M?O|05>*?oHZe+K>P}bWRgi-_RD;21dE1W-$RIf{U>rtPY
zR)0V^?xb=u-ZHd(3AgM1oic1?amVR(I2*9ZWDsc#QU$iO)89gvjb0>%HAZB0@iqO~
z@`!83ex*!?*0wZZoa;vByueO-)glN=(CD|Z68-A#VwlxZ?<2rzC--?~oR3feMgc<V
zyMqf&(AzwVm`J*BwJNZ?w$8>WB9E^I7!~K{#IE~JYuwda5&$JGv-$0giyT*+M^;Vi
zD@-1R?DEfKr^s%uF9?{@U0Q=cbW_ptosRLylaYEmFuHI3xtNcTp`xNqOYMG7R<mMX
zNBsM6SiJ4Wwp1e>HR}?g8>VW62FCr9DhO?B`01Goi;}fUZHWF;Ri#zI#O}B6U{*Jn
z%UHIPKCR`GYjB#@u`5Eh*m$kzK+pYw*>k<!A1(h)Pmr>H=ZWk7zV=-=Y9xI3x|AU^
zg^vts6QJ2oV6n;roo07UZFpdlu%f%eXY7K4XgG@~%4Zh;A@e`{Gu?AzMg#o1lx+OE
zsBw6Zyp|V`%Kj>`Jg>I|+5^{lc5tF;xr+)vO)a0mh4EbYeEtO*`w`QeydlBT@b$rX
zFW46z{#FETM#rgln$p*MU0cz2&Fh62&cZPTJ4_=x+ncK;kZoBw956&`bK6{eT%fG6
z*{{gM8cd4o{@8R-awKS9Muux6m1>J2MF=EloqCpH7#`OiI-qLK8uDE_YCF0Dq9T+B
zBEu|+9Iby3l92tHB<H2^?yq*_nsv<`h;I-fE0TVG`Q)nLuDNj%<*Wo6YRmZ^90+#n
zP`FH)NG9uV1za%Af!ghfF1C26I{^recuiWpDglOqQgo9S?z??8qH)J&z;m^RFfp7R
zo{5R+P)2m=$UbJYvy#~Lem1hZOH=81K&Q_uB;izYM+!!N-PVy)TcnS>C$v}JaaLvZ
zK$g|1J}8)wL;rut`rvoPy);le0qVzdR91mPj_5i4Vzg(8p<+3>e`VwPD`QS4aI;zm
z(=@kcSlg769I+6@&k<-~63iU^(sDz_i49fa=b@KP*`%0$7@@H9K8}56k8#d%ss(6P
zf#6COv(ILB76416<OYHn9&xL+wvUx35PoeH(x}oY6VsKG3;8j#qxbxa<kS$Gej_a=
zGg6Su)5|oZkedVzR}IO~#PX=Zrz{Pgro7sZ?9lO&rMv1=a(ADq2tb!TNDUB1m%~$m
zlQ}l8?1}HIdxwrKZyr5+$UFDQk6qHn6G}H0RyT5Fo^TZds6|?_%NyY$8{xpu^0@%0
z<v+8s-!+ae0Ve>Ub3cHHB^E1~b^szcNNIuNALp{}<Y!}PR0eS>&y2gE_*kNfDV#Bs
zbTLqZvo;ir9lu2+%wG^0cy#*XFqJte8msP%-<5RFZ?M~#qNWIe^%594PjqbV=nX0N
z#q}1Tu8(ucX5jBlDb24!^9KE>n9T6Oq!_@(O<!}MpbD%sH7RgTMS-e2j9dwOFazr`
zl_NBvvh4di7=edVr+L;(v$B3o5xrICMS2~%LddV}D3ezZw0L%cc=o0tm(|g&wTtN;
zE4sWYOM$O;qMws^+41NS48mwY!|%O)wQ>d7W{+R4t|0{sLjyI?KrurQ%Ar5Utb|b3
zgFG42O5J|S0|ov6xC3gTc9ubEUl~m&J^+D;iF|N4uK{|rtO={Ueii=#1L&d6@Z{LD
z`bn`j9zdEg`0@Wr;w10&C(t>0<-V#|fN(x75;TsN9LMYrR<p^9PRaqEC=oMbO3-)@
z_Vee=@$l|>4;`wD+zfEU-i8jw`fU;jN})t?uBq#^M1WcbyXk2UT&9V<m+>9QZ;5{0
zcmQ?pTngAuOl%e%<JB)Sph(^%-6MgyDAAvykEbiB-_W4g&&sOFP}rwP`M$^-#gQaF
z(|*6Q)8Um(4$Ax_ygEhQ)3;Jz;AhZ}!2M3AyN6GeMP_S?Yi9<w*CW+D@B&%oVM|*Z
zTe|nZG~AqyD~KsJvTE3nZfDW4A2$59j(~2W=1SIL;lfb9DRj6ywy}<a$(BGG|G6lk
z0~U$vwo2XirWT$uNS-5+Lq|}m+vB^3YZ0&Cgf%2T0D}%D^@LI10;QJ^3GwAb)zVhi
zsG^+kPJTVXcYwn`w=_eX4oFaxZK^Sq6HFkNwgD5P6vUKhi$3wt17ghPW?n0y%DckG
zr45fXO-j0+a=+k8$H{~SB!?ej;Tyyou)d=if%ri~^(ey5l+2J9;5s!a^oS|Tj#+g{
z6GQel2D5**j?Or@O`gFKS`fQTZ=~TP@!iOXgF9)&dc|Tnse*>_T&`}7UEG$afFPV{
zTdj2)m_rn3<GPKQ4TYBF8M;k(OxJcw%ZIJ@mU|ULeM(;igvyk)<HX8Kh)~kA?U@v6
zgcs`V(O}|dOi3Dz;HwR6;E>s=Zz)$~=Eijna{@UUb=Xk;Co5L}lN%^M?($g!Y!dsX
zhQ7YIa>WrL;1Z*v928sEgvkO_t9Pzkw#|5ga<k(G6Pm9dh=4p2h<b?a(GoeGp3h2&
zAZ2&G%lrgJ=lU^FNuGpy9%i`W$D%<;hZp)^_7QIYZLPjo?_XR+FjaM{wzY@nPGAKj
zOBu<JhGwVEV&cXb3Re;kAz}T=f7BQnAlYVvh)_{z7vRy7xoUHvP_bpSKS?=yn)q!D
zTD4Lv0Pwlxo={Jgk9~N#+^oZa(7nGpk6e<sL7#07;m4Gk6E=(DyD{+E*P!TWyo4SF
z81!7K&#iPx_9y;~|3At08n=o%j2Mw%u1kzY)%K}+4h)5lpm%6g*R6`jjK}x(_L@sW
zw;~}YGn%$z*;`dVm>Ox`YgK5%6Eyt>6!@T$^B3f+biCp(4;>{OT6J`c1sY_%5f{Hu
z&u5@f`hlUPY+-Qz+!X4eqj*jh!r$xE?6-Xd9MgAH6bGE_mPlPm?$rUDe#-Nd8mr^v
zpq5mZ4C9Xx$L8(!y7cVq?JZYdXuRRW@Rds@F#!DycxvZ{LseaFQtrR!YI~rAj+mT-
zHuh#`OcF`@bNT4Xg^TGQ*0mECMLLzA0ua1#A%Bg`uC+6TsRa=aC-jVjGNIz(Z&8Jc
zv$EY2E#J@~T&837WLH9~Li#3uSm7>}nylve<&T<PJlGoggpG{Mh;!+?J@CsovRLv_
zmK+i_%u0W9Vna`1&mc;9Z7~OwuS^w?XD3J1oNj(YRfHx~*GbDwsYOTJ9N=H6!B?r(
zQ3+a^xQf#F!7P9c|1hWlShG6$`t)LP-Tsh}i+z)M@sN{i&!%w%OywAa<QSEpg@bRF
z7vB>S%hzrQtr5-TSYoJbht^Yud%vh0xP_M0zQoDD_lvLLQPXYqpZJ2`^Ikr;U5jPg
z1&AB&f6Q89bVjtVo&9O8iA!{9m2NCu+irtPM0QSR?463yms#}*!xi>KCJdj09Sqv%
z4z`1Pep9YtzZA+C>r?mM$@TZ@<d0T@#yxSpKYwOcOj@1?De`493D5%yCQ-ghoC+V?
zYu~V2J8k&ky3qaVr-{dhirWHyET)kj7H~|w>l_j5)^*wtUnica&zL2|zp~IJO7|dM
znf>!^&sY)TW%#rf-!mIZq<w{KFGY_OiH;h59fi{(utV8iqh~_!9j`YbI>reTP=fG*
zQ}*o~=8bh=E7S(?;iz;JdjV3(?nOqP-+wqm)HN1x80(u503MUdo(ORNC&}+j%a^2P
z2Df^E%ZelN4ha;FB7x1J)yF?MI*2vTKj|iv=JuHimXAi`;PtI{w*UUb7Uv19o1#{`
z{rt^ZNOyxYaT=bS7wpeF*mkv^bO>Ux724s|frjQw5*+T=DG40xQBjbf0>7STA#PE#
zwZf;nM85(==)OXcVgbbU{GSIAx+p40Y}xtAlMUhU@SJWMihJ274VKU9j70aaDFZ6f
zJ+gKKgf|2IckD!}zz*iJd~u<)ThZAK_#5yVV!4P^?3P3q6vT=RYwGK^h(NsnG8Fzf
zry>ThE9C+DK-=;z(vygK@wARVve02<TL{gBT3bwbhZ*Rrc+0F)VY+m$r@d=L0h*rL
zbM;26DxN%ri;5I9Y<`_IBxF_uKo|xvRv19$FuQg=k-^)>?M8jp4uuA6U#;3)Ej!we
zYSkd*ms<X{>&odDI=_peQ7Nd227m>q%}Be~)p&Y=7Jhl#G;>Y`%_8ip+p$NCEra;X
zP_Gn<AL?Ea4e5y`$VyS1uYt%y>9h<X1o899CY5j91_C{QxxwGH3q2ac@Ya`!*}g!^
z7t3?AC>YKtoNfCyyccot%7uVcwKT8PioEua8;EzO%9kZ1Qa}FpJlPz~kC~;8<c{VB
ztwOQS8ig<L8p2wCcyfhusl2+Lh_1|g;x&LExdTV@^UNT2;YZha0Pjdg^rGcI^PeB&
zNCDGmQv43%_pg+zZ*^}{<G$dL1E}2=f~T#9rua5C(k%HC8cvw$vw^Z=-;S4|?{^ww
zvlTu_hrn_oQJO>{hx6Zq$g-$c&$y2pEb9pffz-8a3&d$4mqE*DS&o~n*%o?1GbN{a
z_gi0C@f;7pO(1iG{krW?X5$OC!}#w3Fo_Fcg*J2mBqs4DA-UxR{w-cBe4R){R~~mF
zN=hTFx9W)`ss(a=sqsaBb%59j{|6`!x)!yOz55m;&D&tqxM@v_5uFEc+sA@sQ|9uK
zBv$XAQN}7CfcTjpo&dN!x<`H}lGU5~w!(?yTOIr8f@hGG^4rhl&_tLZz~|9gez^UE
z0o-{A@kGxzoudb66$%$e@|AYQhUkdH0)a6=Fzv-T>96bo23H15hawmr;K`h>zkp~8
z)D1F-nqX^p$8(R*{s6YeXDo{nI$upkQsym2EL^vPQhVq+3Y~Wrb6|?+<de`Rx9NJW
zE#yX@;$UD9VkJccRe)61+xvBY&!5I4WW{sxTaadTnym0QH~Er&IWF5c!v0q9M@h7S
z9g!SUFb&H$lq3HsWJJPiMX+!{0d)7stD}sci4#Jm46-kmy<nqgSf4otKn*T{bPwFD
z?9wlplhFW#Ti%$@QAkSiXGyH^Z3jyFPvG06nB(bAxh3zJx#1ND-x~{2#uRSTf=(-L
zXklDu>OUZqev8HLSi28|t!voLL7v2(8Q>-s(hyJ3JTfQ0@bCQ$WSMOPR*D!E_Hm;l
z)#zaIKh)^oS|tfQ^K-Qim0KcK`Z?Tf5mU)1YAqw2%ak5Jqvb<FZm>T9ws>aYsx%;Y
z32r_@lpo<d)TEZ}P(-yH(MxOagJe==Knn@1lOl>D=Wg33v!Wh5%qXI(%vdP#oc>m0
zq`Sx-VPD?YqXO~()SI8W-j}+hbV7nb=WfgkNEDOeIY7Prl*#KfHP1CW!Z$f*0t3nc
z&JBw1xFQroW=v-P7lM@HQ8Xy8U|W|0x6i=#P^x^uJ<sfc_jz9qjToB@5yJG5A`DR?
z^x<bkRm4Cwto(bkPWuXsvDT6_5kWH3?%a9g6nmwJ0m3dpQw9(dp6m61E6n+i6j5HG
z*)a$k1a)jVjA~RNMfUlC@8#%9uZ}&7tuvZ`Q;hW%2PP{v4-ocaV_$sd^t!&ukbhvl
zi)$4RW_kCVmYhXe;Q&}Yr-#H(8C&eC%KM@)8`wi_PMJ8efFTVq@4(DQ7UBsi*Jqkc
z>X&Q#LFN!8==*9M`G{0GfOc`)WIjU5V86-<onY@u(QR9p8_($#5y}g)5|u!bSeg;M
zRf+EUPt>e|BY66`lY?j;oL((@XvXva^j+J0Vo2iGYIsn;=Xd%WvviaTWC~!XqKXm|
zCBB9T^-V)sxy+O`GRI?hcYDXq83qZ!N{*_WP~84e0lm=JNIjvZ(}JRoB_MJ@ni#R<
z7SZZ@0qof1s%E+S%)xrk;F&t`zbX{HR2>IUMW+$V>r0b+EAV<pI9z#M2f$(i%0Y3A
zX3+Swf0N;9juq(O+}}@STR~b!pzlgz0y{{|T884I83qU;lY>Q{M+$R)#x)fO%MrPr
zvAqp`{I5VIXZCFO9ulP*S{|;=ce^4N=^qdcOFp9mg8RahW!wm;5$xZsifn~l3A*~9
zFY$??<MqDjgTEI2U{U>P{laWG(2001&*jusQ$K?vVh8TyAX{U+e(574XK<4UR7M6z
zvroch<D`kj&ku+S>L2497?bL%*Q8#&{{S@eq{gFHKuEskpt)nV0QXts-Wg;H;X)+7
z0hV~q#OJq|x^}E1Yq4tqRuX)-8PK1AvN+X#qWsU^ye5D4Q6Bj2K61sjfRSg|pXxA9
zo(>iktTl~cp(jffLnr2`oF~#voC9twlYZ4<wJ+J*0>0_5ZRn4&2LEMn{E31+e~kR2
z<280r@!L#OG&D4=PoIC)FB>)4>ysE+<`t>u)k^=$GhH1t#WJp4<E%x-cF@jcwD`$;
zhUNwCe&Dwu5)WzNy0lBDdb&KqkIY3D$UOWhexvoqEn<54ukHPT)Ft2ib%PSa_P&5;
zfq9~?vYUlUZMg#ne5gHQpQ8}ZVFm85>soH*CZjhj*H^FUJ1+e(r85+43?x~e?rl8y
zswQJ~b%+1$B;rZ;&5M|PLsMlpp$QagEvw!Bov?J1&5-&CITSeIdreLgFU@6%sao?z
zYvMb9wvx@4(QrC%344hZlZ@Wn)XA@d1z(E`s_wR18X`MOztl9iH!uy%IJWQN-h^cJ
zFj;Q0giuDXx&?7A2O(;NOueN|=g*Xw3i|$Zflfd4Aj#Jexwm>QJ-gV;HhAVtpiZQT
zf8oz=z?bp0&{^C6;<cpG_eX1~OMxtf9MuQ`GOs^<-H$A8nb(h%SjJw(LRN}DYvWvx
zo&r{P0`ZFF;O3NMiAXhID+Hxa;dmzExr-)>l{LJa@@(c$r!R`r?$%GHi{jojINroD
zzsvtnm-oZ|bBB1wSo~S}_H270;m`~beko(2_1@`zU(r`~qr2I-eK(Pl^fg({AlsGs
z@|xr!ljGPW731CM7Of$~JK_q#zFnTvYP3N9Vxhwjt-7L!PBjnQ1WhljTE>#|6{>i=
z`z$H{lKadg3#y#VV975;&P6S9(*Xz&(dNeM28MJOFz^TYAEs<XVSPo(znR3E>&qj@
zkMMZ#(0IJtn06<1vhU?_(p=g_fs`XxKct8L3ZJ3kSAXm~r5;~&n__FT+Nyc!WD#z`
zBRC_c&{k_z14Sx!TYjgZnfw(}U3U05)MpO2-cV*cQ>W6ctyo%Hs^Op+Z>w{2@j5NT
zHM6D3E5pZ<hU@XN>x+1el9Jv-rCE(7#7Z`X0=mSYOF|1JzEMz)wVP4jSV~dmeTY(e
z8TlaJO8%`J1^-*5GjHwcPApG^dP1&=t=7o;b;R17ri(WB6<uln1H|Sd0inXtwrWiS
zYi>=vDa7icOaOY~wEc6){pqE3gZ(|KBnN-&jrE+sOXAo1FIj)CS&wa>92F=rn6P9h
zk~)dfot5w2kH4$@Siag!1-_v^$My>kBeU7XS%t#-qFk_KX&hde5!Y04c<$%vJc*7U
zy2hh9vvS{gDQHeb2}P$zUVnK}hl2k**h2TZ>wLFNPc+s*L(@gM;{5H+5qA}RDUzgB
zY1SUym`1YK7kqPU*3-f*Bc4p2!P<|~FV}{4=gkF_98VlBBKZvNFJP|PWz+sbMMzQ=
zhPv0T;0Ndn+n}tM5MmRPRZZH>1pAfn3^uH(rnCdU@nr72iH$K=FU!HsKR=QEr(y0?
zjLn3|k@?F@2iOK4@A4a*H$~~2F7Qi@274W>EsZ@)&dpC>XZ;emlm0@tK^cKl8df*r
z?NP$VUNo|{^ks8&=?g<?alOr#sd70}J_i-;Lb#2S+v;T9uMUmWJu&N^!^$tjz>;LV
zb0Qc$uGsn@S*EG{{jppl{VsH|FV!*?J?z}XL|;i~ZmC2s>mMrqGEBwC$V2?cO8uQ!
zaP|r{XxI){V1ETq!@xx`4!1S4U90K&z`*?F)PD){DPkH1CXAdN24*b?@&Acm4K5yV
zs1$#ZRHBTXIctM~xy(<It)B^+JK$!YvG^Lp%Efi#_5K~q4-}&+^^2>kC7XF$?Wu(U
z^mr+Xf&PB#Xo>E<i^CZhSeQH28jUaPDHGS%^)G(@C5$}c2H(;?7Wyj~m@ZU&7;l@-
nH!$A_Dqvx7#4Y|Wzj}flN>n79tqDkg0e>V#<wVMazWMwgWqAuR

literal 0
HcmV?d00001

diff --git a/_images/3f78e9f7b93df6d2f3452d162ca631a9de6f98afaf55b8f4f59ee8351db6feeb.png b/_images/3f78e9f7b93df6d2f3452d162ca631a9de6f98afaf55b8f4f59ee8351db6feeb.png
new file mode 100644
index 0000000000000000000000000000000000000000..2a09d3420fcd30f654e43052e2817e849da40c70
GIT binary patch
literal 44455
zcmX_n1yCFB7j0X96)5gbafjj*cWH5VC=gtOOP~b`!6mpBr)Y79LLm@bOY!0!N^pPM
z|Gf8@$*_>jW_Q1P?|1Gw=SFI%DPm!eV?25C1WQ>-PV32&r`Z2quU-KE@vXfs27bNs
zlsE9ycD3>J`3MF*QT^!Y=H%+>WN-P#8wB>ScXi?865!(Fcw^`3>E<EA&F%dEKfvV*
zw&fmWZ=C{8g6^hd=<(#qYl45Tr%y7oNuNBaB2<=>*740gx;Kn*UQS&_k8Py4^qI`w
zc0BfRUE#_}*~8b<voX?OeGOW}w<lWvf}ezk<5>e0GV}eqBE9~c)(p4f&7_jj8`=t7
z-KJsE$fWPn-e#paOI)sxIoqN;eUV-BwIn!ek1?w^1L=-cR69%jWn&34wU-Nyqc$Tv
z16T1p1KAb!i%qdOXP+e|UM_rMdTgWk`|t4PQS#=i@V^5ois_gAC;xjpU4kX<f1iY(
zNxKJr1r7yEaH2NQAYXABc$_cquQm}dDAX>KiH^>Jz1evYK02z};7p-9oqRiY;B}r6
zlUcIY8O)oJ4DJjP@9yk`)U1@)TiV)|dbln&)pmALMfcOGsj0osI5-#a_V%_62oUAs
z;#%L{W@3UVmx`TFFRskdoPZo1*?%2)Rq9o!<<S*1KswD7r}^DIJhUZjw*nsdrabmT
zeSaQ|k<89DIjb<i(p8v%H>VTPOe-4uK};~(3uoSpB?|rf?YRggFK1WR0f)x@v#MXI
zwB^fL2<gS9Ep-S3_dD!kSfC1vF_)}ED1o1+mzQo;g}%Er0YkfOr`f=-U%xsLjjUMW
zd^u|Lbn7S-%53qhlF7P{s=leYd00G~v|^TRY<zt9KC{8{!AX#P<W#vAT!h(iUq0U5
zt_8C&H8)??=BEsl*tn<W(<7&^2_--frTDIUIk~!)Mp<zrx(ZktL#E<sS+Sbh)$OlZ
zd;@M5N&^DLm|zI;(B1J)9LvW#KjCoK4%^)av;Ll_2)5)O_agk?lrRc^31STX`lV<4
z=Pn3BthHV=d3vZl+FQqA>n)hUM@u~Jup2UNO^_eIdZP-O#<2Bj{oCwJVQg$%{CHC}
zPF`DCxpbJT@_A6iHOfeSzK{C!P6I-gxv;RnhDtn%S8^rT5|Xnbnjw8&7r90wM_EKp
zIoH{sB3pzVPxkxJOoE=YY_++mDQev<S(-8bHcZRzb4^W6^c%Y=#La8H+eRK_%n}Y(
z!+%3x-a{;JvzxOF(waKGn46oMtweI*z*4XTvXRy(mW2|u;;h9qn#f|PBQ;gOI^I_{
zf-tdE8!+YH+qd*;EEZc+N&mW;Dz$Lj+?^c$7;r~TPTQsh;Wu)(PMHl$$Erk+XmHDW
z8+YPr%KE(u8?h_{#DYPU-lawo`*U&Gs;<Vy*=x><>FctmCld8IT6RFp0c#iVycND!
z(i*!raem_MzdGY;c8a>cB%e*ttKzo8o&N0)*ZS2R)9m@JOoy)_E0*;jiFf3N`Fd!n
z*C=4#W~b^w+a~f;nVoeJ>9D`8em|m-jp8_llr187p3kc4ThNYFVm>+H3kq?iZ;n8)
zSJ%{tqP5i5!{BNrJno-5MzzQthR4R#OiWBXnOYuAO-zb-;$jpfUg(6&S;~mZDG?Im
zq}rU%TJYTN&pAXjs$X|TTawBO_!ZLGO7YR>9afuAqR-qg6t8y_KN~dJk_h`+Yv~k?
z@s<vwV0nH%@;Z=uU|^tlNn*k7T~kT^Ex((&p|wNarx|48Y)y*%hauq*K8<l#ASZx}
zI6OS`w78otRUa-|Tv&+8$;k;ZxLMeD{usdcpB8zC{t@}>YrX=R<i9m@2Q6J)T+YxI
zf<S$K2!TIGtEPr1{7o!$w<U}1XH{t$pN<U;=#F?iFoyDT9VF0f<6*Bf4{lk@0&eQ;
zG|ja&+$5-(UC3f}Kb9?-iF`}vVx`B-%v?C3(D$N(?ZzZDAF}GW<sZ>$60tP@Bg%bv
zi<gU=`&gP2Ysph2Haa@`+3Xnw-m84_1JQTy_WD$>Dm7GIsCG(X_JRHQ3J%(8V|c{8
z8uA|;iQan+i6G>sJNL7KRu>lb`+PfMvHRy=(DEB`$4&$1%F4=W>FB7ua2Lr@SZvB6
z!n^n=kvA|qt6R>$l5+D$JXTInA$cH1GD%gUG%}COx`cX$iElD)*u=(3AkPFIN!Kbh
zp-53xZjJFX`1lr?XX<ASi#_<;*u<GOEG#O@E~5?~-#Kk*ZT+_tJ3Bj<9K9D8-nJ-}
z(rI3Ieq|sY{$+t84TKQ5f5zV_J^pQIE9O|9Z~{fL#l=N;2!!hNT1>A(S8>mp;Zkvn
zJ8js2D~%^ZWqo~}ZftDK(#J<AcJGzFDidsK?&#>KSNm5Y?JHIdH7-_Z^mSlSZF?Qx
z&252RT`TRyFy{S+Y(`0ee^`JoV$_A%$jQ*pO)nMX6ceK(vM1JS+er?y*qcVg*E?l~
zlh9U9cP9&dPq#-|x?CtJy<S`&LnKT}@EpB<4GpK~E05Wp`MA4lGKHYt{kM^_TO2vI
zSuyXVM;IG+eRtPEKZBylD%oQZ?$MluvEt?7#tCkp`k9!UlO)Trr=N5?yB1#J^Q(TJ
zbs(DF;|CtY#_sMe^-NsLsD5ip3yabra1IvDRzV<VN~Z0zb8@_r|4XYnH@(Xq-*NZz
zGhC9S<C{%+Ze1_;U1DhBz@R)|0P*Fj!KiA8)XoGFa_du&im=wYS_@C|knNzoLHHJ|
zW|pv`)P)kpHEYq&Tksqg1+j5(M343hC!^>c18v@YW1N+eF>_FCeqf-PIFx`eCw-mz
zy|NMqgTZ(w$P*3@4hr0Te0*NZ3S94;`r%~*S6(q&jy4UV@~j!^Y-?j9exG0*Y`E0x
zr^HgBj2u5cJnY*Th%dCy^u0!bqz!}e76O3_52?gf<>uyoJyr0`tV-|8OW+(VZ&^Qz
z2S6aE6H#HOz_QX6EHT>!g27Yn`oM#kma@Rgx8Z*w?}QQ~s^^{?-qq+L6W*$OcMN*K
zG<`1Qnel9cVf<MkzmgfTAHos2fO~>X<~3EOn?j3~tm<T>ETF|Yi`NYEmB>?+BmC%5
z_FHI=vh#1-kQ=+S^$+wyLEg7RO@ok+*K!si%lK@qq)yC_uauXGCZ3G#oC13(RdS)N
zd3O-i^j<^5+V*xyDE183iLP5l2*>k$^a9&3BQYO2XjrNW=t^#FKM|>UFGG#zLvVO=
z(?+5S+UySSZ)M?=>d0NLX_|n*3Scon4E)P4Cue6Oe)cVj$x+6sqBQUt1`S%ikdV-q
zfn@f^;l(^wL)6MnQcx+!!D9(eoMr`^l|Xws@<UqY8_M56&I(IP#t#n<tNq|y#}3)6
zA38a=y_q4GN@RMS&y$iMi-7LF62I-}>|91HYGiPPqK#afW;msM3M}|;YHHdKpcG0<
z%H)ez{Yy(ref|CNBR2I~FZ{+t9J%l}2&hjdgd>9+Dg(`&i+SE0m;Dvc)54GLUpEWj
z`)x*l61sK4nsq3&gL9mv!o)aMbW`?UKgo&m9O+w<A-hFEpiV_sR+!6xwccxK{Y&^%
zuX(i>_2;FI88LfiX{xBX*t-eu5VL+w3bEr?#~VdaEyt^ha)(DpL|7HOohHvs2?$fA
zlzxauCyrPfH#og#FY96Q;2Z5Qs}Ut$W=_4?>)0hQF{Hj2@MF$jSYGCIk7{;ZCI0*u
zu_lv~NKcLqWnZi021<v#d!nL1tJ@Z<S$$Ccw|+}^_0Me~)LPU@Bp0)sJT<mk7l>vP
z3pTG<su~_MsUj~RmlLWjDfvyxKMcknX|5JQ>n{8xTlN(nfJpYleBvjQ@dia&5Q`GN
z31gxY8T=fKO#*E6Rz5z>+FXw?7L#mBa;f_u064v{Z#98?iX5?E^6EM6nnmodZL>Rk
zT;T={d@Wz6#aqS#$%!G`CIq6ON%*;jXw{WUe`csLzQkvw$sX*%((WG)>&Mp@S064q
zN8LANOE@|?Ots}6fy7wp@bgd?oO!H$C6py*m~nWSfLY(^d2(~ZLN?)r_)Ianyu2KO
z&nLHESAfb{!rk{f8Pj(L0Ob&0F8&+F!L~+NN4(%d1zmyqw$V!B%bLi4o62)|N0$>M
zYo=L@XwuYj#*4=|>gRnBx;FqZ!C-tm%^{220RYgw4<+!gHAohzuCCT%mt(K&vGNRT
zH0ob$b`?#{8!Ql10|3>@ylq~8P0)D-+iDAW9RP3Z=PT*Ib~H6M5}8BJbvyje%L|?Q
zKlE3$)P4w(pubzMNGL`o_9l;;Yg<D$dG%h?-MF6ZIBqH>gYYy7K40Cpjc+CKDpsib
zM2tHI?kb$bzl!9Fi+Eyyy){2j(R8M=z_Mk6)(<}kUbS_7h2|IJYQC(dqf>BtdP<ib
z0xQ?<>6mx=bz-<R?XYh5MfyV%3{h0s5A5CrVuTFQ(4;z~vYHmJk?wCGjG$|Whm}94
z7r>Xe%Ro|8=-2q?<?9;P%rTCARTj@s8ETw$m^GyRXpz;C=%q!|FZq1!)ww|HjKYi8
zG*dpRl_jmF5|W~#AC5ir>(xJ6kFy0a^~0|c()R2^z_%v@Gy}l4*8#Z|1fCy~bN(P`
zJo01kJk1XCt)a0vVrbJfQ%Lrj7iXn{_bec95pPoUaA>HK6(cDuBlQDX`Rr4yhPw~T
zABN=^zDADRVH+wLs`Xp2Y_^$Kya{W;{Hs<v-2znL3PL$qh}gx`v7n84y1UWqatUpN
zZyd|Lb&$K`qd8j_jKY+dmaI3}u*he^7Z6lQz%KDj?&<0cCLS||ekfsfS4RiO3n+#>
zxEC!$r)1&~DBm0V`)q3J>Q;hTUe*TS!5H(Cg+=x-3cMUS@ngy5cRoim3h<ZpecxjQ
z<J`Y%LcUz0OFnles4$}E5i6)@efREPR}&m8Nry~`NI>&=ZB16}hc4d;@drW&D1wG@
z)92IE(*+u0S1dh$V`gZBVowIH?+p66%<$MmMW)zNs^yoe%rpl#NeEm5HuJbBOi7QA
z)BjLjXXLMvbJMrI&#urKqle>$VSlM>$?Qev{EywQh$25g@_G{%SNUG=cSKcHRr;t+
zu}b6OqJL0OF8S+=x$7a^r>3;CTgEgFXJ7dgJ9{g88hdt*DK{mA8OyI7?_i&_cl9fD
zmCJ_S)W0ww$m79M7|eL%<>fUyA3V|I+*;!Mxmx>&tBZ>WL3l_uYh>-ja^A7F@Hzsh
zq>9%(KR$deF3=VA^v}Fu&t+XxnFCPv>)tlt$M#;hsT@Y}$`L|-V<Jl+m+4TCe=+E&
z>R_6NmvpP}F3CjmsX7y^z`F0$c04y_Qv|LQ=tLInk-S-#%Lto==N(myYR4`D%R+#V
zHtg&s2f+jjd5C{X+ERHbfR0YQ-QH%*;kl>D{*c6a@-i~W`4=%K#T1*O=kt|N+Z)Hs
z#jQb1L0dggj0|I}G>eO$wIO70hL*jm-StL$C6cL>oxI>byi3>bc@BAV(kS76(zI}m
zPW#NHhx@COjjQ!&Ze$8`Q9gS4mRs!U))pu`H`hDW1SHm=`$_uG&UoHHGzqtevc-DV
z%`lP3Y=6*|@89@KKegCPc8@I`o9QJRuFFX{4>lG`_S<%-{E~Wr*Lo+JY#C-U7~<xp
zG4$&P0Pfb!b~V-k<PCtlV#@r;(9lr)A5-rUKznzHL=yMYURt-Bq<Zrtlw5o}{2vY$
z|24f6Ja@&LdR!UtS@pzLA}{T0E#ap?`|3az?G$u1t)cezc1AR<-pWt4mU5o=XO#4J
zxdz+Gh%s4T#^MLm*6^9(xoV_Ea>bsf(ZsiEcQ3hD@)b%lMcn3HF<W3VQBhGwYB(pB
zsz=t{&MT$x1dMT_u8HIt)XX{+QlN}8ZiIJ1$=wdz3Ay!#4V3p?J*_#BWYs2Sz`{p!
z&YV%5CZuH~1iJmAWLxsPv$KY!()2KnZx`-ve=;!T^c9f_$r&hte}&yOLAd2Ol()-1
zFY<=I_)jhN`ImM!J`(M_)Hl=TxXSi+Bg@Wfq-CZK-Q273TUbQD6I4LNM($V$tjR=P
z{ck@=(LPQjVps!sI4(Z^wWK{}PB3PjN5!*6T0<)v_o{9yT&FxCTHsn)^!^^u$;|Rx
zYF#m0$#G}4WENSLhe;z}nf{0MDLG`JC}4biTokxPQ;R_PA?2XXq1LXhxVakhe<e%|
za?@ZPULMm`+#694J91piJtmT&rkY;aV0V(#apH?@%KfXxAM!2Xo&bNwBLja~uJ#vx
zi%#5rK8oNQw^2TQCSm3u39A0+B0HcotnW>io}A`AcGTA~>#N;1LXiy&V?Vl}wq?+?
z{lX<G)CgKbR+|T;!JT^GmO;=BR%Z$|`C)fWbV#r{^HqL{H>lObd+Nux^S#mozneT&
z4~iPQ#isur0run{z(N0G_g)DElpnJWUw8G|?~-HeihP@B%seRu3jP=C>jf>&;H@G;
zl4yBNO7LX*R{0VRoB8!2p5|A<{BvvGNC3|J7UmR4TsWw2_wHrVCrHR^$&!;L`$%Xq
z)uF<$p$eXl=jUE`>m=Q~V$4N5UVb$7LD%U4=7ICSU~MMVf#+d^wo2|!gp}tSV=G_E
zh-k_a9EdX?vW9FV!fUAQ@ZEwxRif)t2)}g_U8ZOII`Yi%aRAs+H;xvYbI=PShjyEN
z8K>k@KvrM%?mhK?3=C^#Udeq-VPstT1z8#$@^t)F<Gm+AD@@{o8eC+4u&GuER~s0z
zhU;^iTfO-*_^C9RPJfcWsg0|++&$5$O0NQp{6gr}N98cQklqMm`NqVa++!46f4vdK
zTNsNwpz*_6+tSvQ1&;Bwkq73*zx%!<5tq$z+KL^+SAMCL+<%4W*>s%%Yna)YsCqm8
zKPss-FNoZAwTpAuA77+bx6udc8g?I6Or~DU(y*qSYJRRD$Gm0sp(5lrrti(^@Kzn=
zL%j)?=#U(NDobco>TrESP?yYC-n!}Gx~>aEB+|!vr%cwl;B{f`HtB9=$k2ZKK*+BD
zpvBwmLJ!jhe|payuf1SrWWvpo*^Vj=K4Mpzb;s&{xC>0zM(~IBJwl-7AwB))ldqN3
z^nZrl{rP5rDf7+LUynd3bi^bWu4A3^Ta6a$xk*17t7w%H&{ZuHOvD6B^kg@4naC;0
zVBwjDz6qoK@ZKj3X7-QNSuc#2h^pzj`L_kW#_6kXa^66|m}`xy)`3h;miC#dDp!+)
ztyU{<L$rd0q`T$#$Zlxky>Kr7C^13b+D#U#^gFH$&k~;-M`a&vuwj9+oau)Uyg{=;
zAhtP^?3j4UpXBFkaVpX=nxvR+AdsIF7rTejo3gjO4Kgg%Pjx#R!pmiQS;Ot0aS}%Z
zeL6>P;Z0qa-48!5)|ZORNr{U8Ko^2JFFCS!hxhp(=L-)pJ_GhNn@%$l#th9=S(*QS
z)J0NemPG#BPG;F@noC+g{M@lV=h^UxVf5>Eh~kIocBc`9j@!r;7P)&-53*=h_qi4f
zG&zg9bP*1KTRGrieiRt#ZMUR&@{=ZJW&o^>1&X5UN(U>iTA@(h_y&-{TU3KGk>k1I
z?GfJ%UZ#AE&1_O60Pwu7fG*8+1a4&BujJ>01~`#=uuBb=F#fW6qHn42IlSQL{sPKt
z=GR}vpTv9`22GN^&V1!>wi5v~M*R;DLDKYcc3KiqT@~1u^7RtYs}ih^fvDPqALa3N
zyN4Aa=?iK3Tehrx7!83J^G?L8*JD2GTf=GYo}M~!tU43v7@tIW_x2NCX7sp8m*5~X
zU=P{DY-xkbzLUFd1cm#jQY^6P^iY{*!@2@1>K5nE(!QwAPhHg9xU@xaOx)#`y7;^7
zGlj*)W8)<sS|}~<XZ6kd>utx82ns7LJk4c((vwgBeWykB?EI&{^DbYIcCNPr$(8)N
z;Dz*uBJx0V;Wf#R--f27rtw>!(VvZ?UZ#^svuB`79DmE_6Q}<?#?fU=eEE~l@^O9u
zfRzaQiOqnAJ3nTdjX(|F9B&$e5RbjaaNbT=OlsST=X7ce4~aeischW^G~;R-Fm_2J
zKW!<aqbYH6G_DFTxE?3BTJ18umhEH;iwBegDJ`(iY;PW|(W?-akcb1;zTdJI3h*yL
z3+tihs^>z}>C=xl*%*^iLBAb2*vG>al{S2v5lBBwC#fB?)l=&quE8I=>!16MB+ARg
z)HLN&Wp4hqPqExb)WhFq4yD}=H}eL#sD2s7e-)hv=WVCtPcDqddB=JV)sBckZ&K`M
zWVjDQnWAr0*x!RPRWrB1^N75JsiB)}T5YwsbAKv;>Un-aLqn4$&YDcbwV|zS2d)q|
zO49FU6q2zNVJ>lA1T*{uxSM&m(og&;iQ~){J<Jlfm50B!z=lgi`S&rlm{DDGf68#)
zA9-MTdw_j+W+!S{!eT}m3imA}!kMa<uvKpW!U6z(@`<}<FCzcbl)Tw31X!qL^x=-@
z=uw3!DtH(6XB>Yrq<vD=`-{5fd`m8rg$16Twfjo>s6>8P|Ex8(*R<=y^|8XqJo9|3
zhvWOsed!@URu2HNSvhwAG~mZ9La5nEg_7^O-AqYCx^C%+pf00Lt2g>`gE4T8V6C!p
zeS!?>z+CCf*pIOt=3108G^!r$_UgkuOAq6oA+!y6JgEQP(hREDU7OB9+3e(b*)!zl
zZWjpMBe*(Dqg;VRNcllIY?DfLdP}zFBSc?75tIDs`~9K1V16e~SaD+y3bFBMpUC?F
z+Omma_V_vDg8DFyXLM>tqj-}3UiOmBU_cv5MfP2#3dyp4c5{C4dwzZ|J9ZtSw_4P3
zm1~9q6pK1EBlmJ3fw;{&pwM=b0L{4Tax^$|i%-}u^K+h3rNFP6g;Q0doiUUdRdWZw
z7Z;-o3k&07{{Cf1t>T4qY{aQO8RqAB9;2tHF`8FOOcl+M&3kYFQ*b`<5Go=h9EpL3
z-~%nsIzVUDn^g2n3vd!(I0tS}tg^U%&`z50??hJiR2)08Lp&jRIYBs841VLJVJ~ld
zyMy)S*JDjM&>INpHA8}?`)1i<LrNFk4$R`t(m%D~kO|HUL@5WJX00B!L~$`Uo=A#`
z#Z**O;Qy#V5cuthjU-oXZftHc5J$`V9mi47#Lx0n(_e-U<yeii(syZU4nD6&_$yRv
z%n(ySzB&ei!dH*qm&Xf=c+7Cp)0)UFO<N-K>Qodl<{c0kioEc0vFmv)N-bPGp8?|c
zWUs_`gH$mc0&h^6l{AnuO)^r`9i+0V4AK^Hvk3uFzZb+706&+{uRc*qHF2KvVr*4L
z*fA<ujis=}QZbkAXR4L<uiI13tm@a8r9|!##NN^Mt~-3m+f1buYI|-%w#F;sSLMyV
zE1ESrWPDFou%Y#~RPpm~gf)I%(2c#cJWy*7vdIou-oJXk@#obT>juF<%9dS*(7M{M
z>w>pwFXL7dP^9Nb13L6liSA6Jbmh0$DLOb1bCORZA?-K^&!ikjKd0cR*-RhNvK0{h
zu~lVR-f5>3f0M<!B(a(clkaX_l<E4uEFqZj48?hmXF8Wekn&3Aa$52Ny=+JBy%spy
z4`A}Je={CC+9CY<bn{hOMJ%nv$IO!!SVAZGRi$A=*tsf9mzJV<%^%YLwf=|4#{>H2
z5M9jQK-cF!`dg9ecAyH^p`D)oSTbc5z?FtP`C3HaWcR@PoAb^T^1Rnpn6nno(r+ud
zb`1vt6Li!reqF~nV2$Sv4{7Jr&vT>>gI+S)`Pj-OPk&oz;?#6zs@XF`|4Nvf9A76@
zEKDkQruz&P@?HyVvB{@emlHZyDi#<RPFNF?s}%>{@#Ib;^2)RP4?K{5&LXdDO@hZ*
zetaC}&NwRsUmBp*55Z8KyRKw<?YE6I&8x!;Hhg(-@<j6Tj*^P1AAwNz%c3!rZB+G;
zaS6L&0Wl;S$s!%5t-<Z?ZUFlJC*YkQ%;gAHbo0l!I9Y$<8_!qhj0zp*YebB_XW7Ms
zg0RLolxb8@%7u?E{l48{d0BJ2>IX|2Jn^434L<<_A%Iu`#@p~xPZa(6>`qf>K>YJ6
zJ5^W(mtjc^bH)_$F(Opuvw|F6<<Fl#)o@{ATDM=|V%;*`?TfPg^bN(Up0S3-VOJ}P
zbVRHU7b#*k2_hcvc;$z9PQYF{cA7nI$ey*|EtXMG%M-G<%Ye>>4$?6(8J)7N3Y*yu
zxnek5X=*km1A3Ba=>%DJ{r#Un$ph}(%Qt=bfS(P%HuO1Xu*@EL0d3JeL;TrlO&JZU
z$S3ZG|0n?yrdz$KK=?xMq#BO^6*Dd_4%z9V`OuaeyHJ^IWoUy}&w%(KKG4Z|7|&rb
zD)Wnw**-Y9+k>C33KNewg?)`*Un8(zsw}2s_NpCh0e>4>g<R0ce;0*|6M48K%=0~)
zj6o4o2W-?7v)k1h+JTyt=DK#&7g(fE`+S0T652}67Co5qB0bVxt{+q4)`8@fbEQt*
zKO&wYkJhIckTcz1J(|^n1=tO>v8}<XJb&M{X=KQ!DWt7W;C)iroL^jI4;`CEw9ZK(
zcui<0^7Kf3w<Ao7x{CRY$Y(WWO$46r-B-oR%JpRlxniiFW%+)cQx#P8;I^TE3c}Us
zJC(_3U*k57gNDqIW_$9(!`?Ne1Zppe>8(Na`XE$eLbC^ATVWe0&dG9mcb^Ec>jV$!
zR@oJj^J%OwWMs?YYx(E-^sl~1K?w(Nj}K&<)k1mVDxcM~=*=e_^o)=FRYhpVFbQ!+
zFVN}BOSeN^&XzBy-0W3A4s}W!+Vj2|W11viK7Y@8CO337xl{0$T7N2&a=%asL%SnE
zrucxCk!&(nv<zESr*V$6stcBu`!TDP^v;A&L@m=KE_V>Ett95@uHVTP3RTzAaC+Sl
z(`tgjgfj&7mHC>~M))SB79aKXK$^_`9ayfXSdVQwKQiyc5oR02$-1LP+2Dlp60x@R
z8&$7%K>4ayQ4PD55Kq%Ar1Fz}-?AhAxnY@gKtTI{T`e*{(QG8|&!(YJaJc|;EzK3(
zS~)G&an?W%uM&;ss&_CTY@=Q>d<DpG&}tkKCNna1Qp{5928cZ(dow0qj5vsM{*+O~
zVj}8BG1}O?`QhSL^PZy}G!i_B4^{v1)P$Unu~O$7p<>Aj*y8KdH+oLZ8RNleUao>P
zOHZzaPXl5lv)D){Lsu$f%5|3_H0}&#lzQiCK{nk4UX^pr-w@av|Gq0DxB(lf_D^}S
zs8D0uL1lzFH5#RSDu;fs21J>z^D0-b)Vh0<7eO{f!G~d*##fZr<qz(sxi_47h6M7#
z3r`Pg&=bBpCSj;4E>US5;u>GhZGOJDTWr59_+=>`sMePr%Z?S~z16ezzPiu7I9%y7
zS7+CO`YM^{_NkWAdyVAYBVz||LeW$jvRREK+;f+TIf|`YjAc1@>+1N&($zG3)oz7@
z9)A|Xkkqp^5bZAPpsn%?N1eo}S?X{XYJ!A~78{pYkU6IiyUjV{3j@w8cdxviw+WIC
z^)-6JlJ;wSkH%0njqFw~i8!|S4Sf&ML1s0}q3HHJFgt1A=#n|>`}5cyc~_d_IWHk%
zCz@rWwr)8|tvAU;=aD;!S}uY8vF@D<Ci{h${WpyZ#_MzsNhU<4m4)-RE8_^BaGtua
z2j|PTGIU6ee?7AF^BZb+inl4RJ-oe%56+!~{uLEzf~XF@q4803fG1|aZ2jS3>HA}L
zQT^iy5jm)&?OE|`a-;Qjp>J8Fnmj^9qhCrmakt(PC&Bl_Cfpk%M_w-S`P*u_FI&|i
z`$Kx?Wp`wJ2&P)!ssO*6jAJ42OjbV`<+tyva)Q(xB}wk`1wq#sIU0GxxVL-YS;dU}
z(!-cy&5S~v{CspI@$bCReV;iSC`0nX{X&xuJ1J(k{of)|!fp%6Kr$7ya&)ezlWK};
z{+IH)H6aCrzsb3;DNNDta}9kB6(xP>jaL}|(>MOFxR|g%+6i5Myq0R6^FA2GUyn3g
zy@i4msKJOCW)ryaN@){f^bb#XGFI-cntVn2^Je%k;~ArYxoUw}c^oDv;Fj+vf9mfz
zo>R7UD!hT;?;Vm^u~Ls*hMB0Mp#eKCZ`omb@;yS*@@>L5rV<kQB@Y3ARBH&_$~Y|v
zSDD>P<mP|HxXO|FGqHc@Zge(QlOuTcbyk0|K}0H=9a?3A_Coqwf)24yLR)k)S)3ET
z^;3@mjANSNk0$IJ(eTU@b{lVe!=YE#27pMEa%fci?StPS6{tfV#UjV9(_&T=UGV3C
zahDtqqPaaI;d1cXy4opZn_f_Z631T7T=s?ulPJts(KRAyICW_=SV&bHn|IP7BuqY}
z;E%qXr`cp#ll`zj8m%8G#+M|7`HCW^z0E_nu>$JNsq#0LrI*!-Oay3I63PI1^AhBy
ztc?K@%pzLZH)A#KMgBu}uZNaCJAIp9ta#X$#G#-&rbl+eieH?Z#|+qnHR$*(HuV!t
z3$%H3AU}L}VzyEa%&n7Rg!74*F!?5b8lL#=Or@5^5(`c2nIvzW9juf;YF#Lr_G$j^
z(u1x{PuD1h?}XltES%87)hS_mdZ0mZ)isML<Dfx>Qd}aLBBy(k*YHw7R7bal``r0G
zWsMVlIfOa}#7oLar`^v;lFcA;W*AZxc?!t<sU+H<XX6+4mG3057w5et3RfM&2X03&
z4cm<5zU8iqXN_9Ay4E-`Op9;Q4ycM*wyI8hYQ%2pDGK<CeF>ZrKwKnfF@eBEl!-Ab
z_V1dQt{}aY%7<?okKZ->XS}uUk(W_uW5>7Xx`nw~!e(cXqtJGcU_W1&8GlB@v3#nc
z^Motmjmp~hN5VihTax5V)VF*1akbmGvsM)){TTKaQ91mo)6z6uIu)p5l^mgkWL*lM
zRZPXu9|!B4`n4<s0=G(Xe_Sek2U}8O(1z!B>g(%;MMU<qYx7riBs`oL97RrbkeIiM
z8(YJzZ3<`ah5R`{P;cl8D}1fGC;b*(yk5x@t5%qvY|=Zu_*g8*nClr@NkBjV)y!I#
z2y?q*VBT*^w_xX|x0{2Jzm$Lu7-&CAOH035ZrY;a&q37*Vf=KKttKm>oz)gIHrMN(
zsyfs}72bUgA+1@1U5;Mx0fX*<A#XUEV&^9Mt?b#KE$fYZgqoqoNC+%-ry1o|=;Vqq
z3_HX~@qN^#A$mG6v4feS2<K+L=Mw5X*6}Fy!HxqzYt1`?y{pI1(17NfDd;t)4UBhv
z9?U>KD@FfNtfGi!68n+j)v~>FT@T|dii*zVxhc0Rxz-B^lvA6xtDMk1tx73O`srLS
z+Y@uOV!D(spoP|Uh3efpZ&7nz>r`>rNvN|I{pRJj(YE*scD%{?{x$;+3=G)pj{9es
zR$j}^dgQDGF+xLbN!AJi&O0erP@V2uSuvB7JHGXHIJ4+a?Agx0J<GWi0e_88tzxit
z+&83X@f6HZd2jZG_e`hMgpAr^<TAQUC1;l(U}1K5%l?D69G2eRVX9zZ5T!1N5zZ7@
zPK8kpJEEX>&WlYP>eEEDtBiLsw!1;-qE)5va3tn9$y%7MvFRREaWf5U`I`h4uMmDW
zWxIo}uR4R>*dP1f<^EgTLtc(|Nq26mkyi{7;NmC@!#d#mM2~Z;$llO7n<Kr=W+vLP
z{i|D}AZKc&*|sn%j`MzY2QTY$%_q<|ESn<RTm=DYgMeCKtHpvp*0UDO>EYg*G{muv
z^mJ2Hk+65TOgJ-zWhArWu%V+OUk!%#`tL;MUMlYmsfp1A+RkSZ#%=F%wLJY3mO|%|
z7P57nv}wRTY<qEWF|=pQ&tO)K`FaroXt$OGfk&Yeq$wZ5b^0Z%aYuJf0sG0pS05a=
zwmWOyz6%_@(<H&-OX7{evKK_>6dq3;V1pkGvFdEQ58@L<3n=QLEm@dG{eR~1@q8BR
z{ZwIL1m(`hYTw4N^G^i9yu#hQ@@10hzvx=0G=lw9ds$zX3Hpyj>P<2?tNW_HQsrWM
zp(euSbpf;Lk$~pim|FcPVo#-gux8gcgCmhHe&__jm%eD4sc^#{uiRz2=Nm5)pt9U(
zFqyOh-AYC0ldoQ*P@5M9eJQkBqZ-aK#~tPC%>I|l2XA?Ab5zC&36`IAN64_t<!57F
z#~9sf@TOI}e@p1icf@o;_j`L)%sV4t&NY%OhoxF;uA>t3<*fD7Lhb$*UyVFW(YClQ
zHtHNVvh|$dra&e!kDGCcUv&VKUsrW*C^l1X5;+9{i;81QGwcSIqD0=kiv;xaRAokz
z@Ik4QSEkdOJNwY9s-}IzHd%0dr1!V%d2V=Gwuwzguv&~dsDV|Bz7w{NTmKuOLOYY|
z1$Uo1Pkf!&<<a~8_j1I}Ddz5t7RgJ4jHHq64Ti6cpvAz94<R4etAcROVMXZg<$ERk
zq<GiWGT!!l_$<td(aPzRHU!q$6_xj@!T;LAap2lo*v_yrZOQ8ayVuF$cX18~aLQ$G
z=|b<R1vyF26)`l#RNI~{>n9Yke6#5L31~QazX7r~`lF^>MnF+nGK*El=ihxA-O=FP
zLLZE{a?Oky8clQbzhKfO_p(KONyKs5j{~9E8fep-zoVne#vK9%4iOxg7PVv98u$1s
zRaJ|8th!tTwq@DjCY%?elL*|Kq`!_y>^@<=ijIO=qaucdTEbl&oW+A6MaI@i=jayM
zXZ~t#xh!E(KRvLaK{H&j#V@}Q{r!bc2U0@n+w_UK$iCBh1H+JV-pcRoLQ9Q=WZwC^
z`}8^ECfUVD>GgQBnouXpp>1>=!Dvd&BT&AjBbpJ}jmPt@0z-j)4QqN4+_fT^Cn;E_
z(_snc43()~npF6hc~I)z;hS7(1$}I_jD9!LkpqFCQ}t9;&|B`$pOu;D6R-Sb=tLt{
zWnPa2=i}>A6enBf7bh4BhQhKv8;~YQ&^9P`7ZB|5RC(uQ*c}bu>INn!ox0<iUjlhH
z|5oP}8*Fb|3_Ahj|06RGQJCqpyMTKRigO7QF57!`J!63<8Sv6^D@trXPZbv*PsTeB
zr|w1M`!7(bZa<1Zflpo<KkQWf)CGOB*!1+DG|1F}Wymw{4q-z!-woU;FG699Y5TkQ
z&?S8E)Q8Dk^5f!z?WfAKy4wFCkxG$gRY(wuO=rzYd6WJ!gx$@d9d1gnwEX(MTKn{z
zSV}F3PQ3HUd-(7c^ayrX!9k6uk(GtFoECBWJFYmE@Fn|cr0NmkK!Wy$DO|lMlgPWd
zFTWE>Fb$|acx?6_ojqA@23x%T)OesB`p{y4l->*+)cxoG1I!d);hF`39JLQ3{I~CH
zzVe3A>dRhZVP=Ep4aL@4$wfSO_V>$MEV4ai47Eh$%ssbS*86Jp^GGiXqKmFD^vfwZ
z{YQQ(W2jeuVwXHo-9$LJ{JslfRaZ3TnwCp!^pg(l4Gmbjih7LkI^#?cskzxE@msI<
z2u~Q82n{VXoJqsd7^ssj7vD(`D$I0EYki9G1RR<Lt)Od*4`r90+g$&IVpPTVe=)GM
zZ)eMSYd}#(zWOQ+Fjs9{UA6R)72#$eSisT6$3|6u>S+89YOz;f`BLyiQdc}>X^j(Y
zbU!1ni_%q{f&Tvf&aa-t8kMyDMy3pgkj{PBJ>Aem5K954{<9G-rtSv+pS1w>$IeBt
zOn;Ft*zADXKn|1o6f59h#Mz|acb!PwugVo-c|p0PSH&^Hc(v78w;cFhCU9X^l_f{s
zrg9W+M+sM6uHN2oJx^_Y{SYHNjd|j}Bz9HuhmyJ1nwOQi6hNyJ7(TCHFEfQ5YSf8O
zeuhRn9g}XX&k2*+_b5k6V<s|*E!L#aR#pC$U&hnKO-6|TvIf9^pen8Z93FB2pmCn2
z$eYb~NDL?TGmUFPzj`>Xa4bdq_?fL>eyk}`@3kr$_4M00D7D6!atee=8Sk$J6~@z#
z0@|Y>U7ftCSu%BKi|Es+OUJME#XTd%3t@_xZ@S)VYhxpk?$5Z1qt)Jv2T$`Exobg!
zkZk(gXW4KSx4hx#usI#iyUrKmx5-W~nyJM7zPi3@1&=2YGA^o0eCYYdd(Cw}ocjXo
zex$udE?q%=w9yOY9(z-QkeQ*?1WZk??3#}8aUQ&J>JSM*BVE9oE>+$Ib+p0r>m0|~
z@G)0AL)F<jf&vm|8f!_Fg!;`jiXwmBzYusj!(h>RNOFKIRB^ZIm^z++kDmu3i~D|t
z$=uvdLj5zUg+K`|O{$;xWxG7q7Rv+QrOqb=X~`+eyOF2<HjXjn3)A0=i)BYX;UF#{
zt<1Iv!=a!-Y(F~rLVpI2u3?eek=iO+bHtf#9^@Z6%raMhLU?|Zy6mgVDEgDPT^;N6
zAghhTMY=OCe`k6BqA@Xu>y(_XfTIyms=%9B>RsgKf!~GJq+@S1K447|gr8EgY3KiE
zIw#PJ<ha$ow8sY*|7ecY{#Gsle-=1W=YH9Ai_hlA)5o`icRSVOlToIHA*be4vDrfe
zwN^aZ*nbH_oM9#^gj4e|7|N=fY_$WHQb5`8N`7v7uDUx19`ai)qJAA2j;t`e=CLao
z2YdL%&e0WIm1XnG$%h^Kr>@%D!zW7qYI<qeWmw@)Vx@u^zx~qd^)S+9^tNwJ>6qcm
zeOk`IQjpn&K4ojSM0CRVk?nxo{5F{&<q2B*op@G<+z=W05Mj37l_!mUg8?(pJQkA=
zy;*pZM|uEQ;kLG>t?ZN62LlTvIL~phQd2ToSsLqDSp@i{KcD<4|M`RCm3;8@m&Lr}
z1KX<Lc}R5_Joy+_tCz;XZiN@N8fPXvZ#roBGJe?OcQyK+vnL5}R`SYA{iy0Yhva2h
zkRnJq`p6mIGQn}}RfcoXq{c+XZHW$zxk0rsK9Q_Vy4Xc(m6~60@n87`R+YjDMI0H*
zuD;8zGMlFkmsjVs{e_NUZg|Ti+t%l|eeR#B7EALbth;T8;vrj|b-RIxx-^(!wF+XQ
z*gXsvZ*R)=mtLy6^u@)8NKhFw-wWBl*%<tVnbFeJ^(wsq_pB>Kld5lLjx+KT>t=oT
zGIK#4(4GU8cQ;ozow6@-hZ6RlS?T5Z6GePwx-58G_weY*Pw3>i9nG(ii=CsY)}6b2
zdhn)MSc0_?FaFuvCKW}CIuU@=1^lRrJr73nc$cR2n0U3WHyz57!ig-Uh9eA)wt{$k
zIvK4zV-c7sqLv6^5*<Zv4yn;|2EVV<>Y6*QT<X-6)JZ#L{lMshPf<p)B4?F=RS41w
z?UC1hkjRY|x{Lk&wLFuOm7aoNaXw<bV}?o8Fd7rxT86+~az$1;b9^TUq`z{UH_+5P
z&ek^NR*uMS4lS*=O-3+;b_s-SjPvHd^?=lO;UJ0<!jo<TD~rS&#Qq3hvhK#*m{63i
zuq^v`at)-<vZ2Pi<T9ko%wK|F1%kT|l#NcQKT-{gF~^eNM#B+F#y~q&4`xvvo&ETD
zHDA?^q6SJA;BApG;IUB}z3z2{R6jqalFi8alUIM?9}}S<<NO98eg@$%%Nzb+-p<X%
z^`2}YbG$99g!MzmLeP(!S*~A;%<XK(_cT_R0Yn2#8YXTTwmQO23vy5D$1aw4<{WZ-
zh8LM9WRJ=X2hE$O+@_MuIbCIggIqB51+xEXGk|VdsNNMc7<P${I6jVA4>y6S7$P$H
z78}Rkd;_$@L~5s;<!*GqVtoY{>caci%E{Lh>$0J9p*gxHFECqI9uK*k&r01(C?H}n
z;#{6J`x54ekl}3|M)=@CYsW7b8-4w5D&)^2!aj9NdaB;#Zw{)>@e0$p2X7U`Z(d;L
zKbg%)<fc$qGDk-YU3s-PX0Q)j*zEjlI5O2^PP1KXI&$(nyJ-Ia-wC&~wPgnUEuyc&
zcuOE0A{9DbEZUj8Bc2}c_kr>Xpbpsjpvb!-{<3AbgiyExEI-}>^N)p%QJNPj<B)oM
zhVti``9R-bGPL@D9M%q>m_UwHu8hdsdNe-F-TXaZNMn!9-)Y%xc6vT0w9x9v&@dny
zrR!pz2RwFvbzgl^vj5KIL=%TjFfH0xWGt+<7vO%BkQt%fV-ZbiIwgru8J4(xmDhcZ
z$}DtklU`k@+U;W-iqs&gza}A@!{Js#C(1}?%8-*bW`esI-sfj&(uR3=p7l@{@@LE+
zfzc-m-e#$pQoDMv?%3X*eQCDSu$bnTO|C48EYcgZejXdcaGiR(<50m9h+}|~@2EPe
z91Hg#P3rX!7)J7PXgvIHTg2V5s-Ixf5>GO#OzIje%f(Cq;#_|JYsaP2b)VwW>6`B5
zgpn<VUXkO%PVdfntGXIr0rQ{Wz!h!gJL(rx1EhpnHS1SicK1F+P}-{BB#Uq861J(b
zE`I=GqYSOvrdIt#WATcr(%Te?@7g7Vjz#}5Y)L~qoYaX*|6Vsr>}yj)Hh4C%#lsv}
z=9>w9@2jPFElVnuko^W9^72j8=_gLx;2_W0X_7}PGFh&pP%BCtDt@Ct^%in^C&*<>
zaXA*02}$}TF-3KA4b7<Dg7Eugz8K@okfg>v-trTpQHNhF(dysKA-86~sZVS$hy2a?
ze~m6Rj4ZUAWSW?j{ITU7MAd!u*p9P?cy3C`OI=qM54WpG$Pot+y`*5vPb?WxvSnm*
zZi@GX%gQ~(sWeF^7Zc2d^#_v_F5y%fX2skvg)XG9W>iA5Uh&;)t7GD$EHL;rHD5H=
zA7s$v9=(i&j$z0;KeJw4jd+G&(h;m()wRFAs>Ocsu2K0AFiQf|G4go|Yen^h<IySl
z(ftsig5HbaairHZWi+wkm*2&mVw{UWFJ=|ldfV7`?RO|gGact-iIb9bZ{|qyWh5SM
zb#JxMR$ou>PvG=>-LXGYQVAGZ1GtmP>3-#*v*kO>`TWe>*T9zTGc9j6>8X_wo=K&3
z$YcT}u{*fy9emdoo=P#fJS^v^=6`$PyklBlL@q)qyD2Xyp+i@>LE8|yDCQ5#`Z*or
z8N#==tdRjhhM8;MXE2>d4(|j;MnV2AM|&3kHt!v&#IA_|odh(wtjw8DIy6RYY;1sc
zqoG_Rrs<WjgmoN$b6VdV8gJ8>o0t#=IWLqNb>b**bhorXNdUwEdu`yD797{oZ3M6-
z`w~r9jVQ-!K}g+snrdhieQ1zCIcti9on%u`(?_MBHslha!wr;x$xgF0^3S)OhNi{~
zkI32ELu15!H*%Cvr>`nR7-&sPT2;3*#&;&0jyi?DgNDlEOM|K|I_|EFR^AJy3V5dD
zK~5rCJY-A-LIp}jG7lE{M>|io{mjD$>%iGJ@CqDV(QzSpyW>!W9ru_MvWTTGJS^7Q
z<&0UN;6SF+Bj0U1y<CZdVDIYx7T5KPT;06Ha>DY)1Lj4{)R{kAGc|Pi!wa6^`|UCt
zUazEAR`VeL$XW#I#RZqNZxL`iI#zW++hg7ECf~&jVCm|0#Yc0C{k>cu-OE@gkau3X
zg-5_WV2txVB2P!w*w>ER?S9mPSdi;TMOAw@jb+7i=@G(M?-*1{cpTILgRWg+<=QX{
z0-5XdN1FQP(267h6DzNGuG76|IJ%RQ%!8FA>yBkFMQbH>UTINB8Sed91ALQ7oqMhx
z1lJ|Kv|GF{#GNfi9Ib9~LnT@eRWC|~4METM9J^o74(-a~;Vp+c-ajm~@vj=@@0?fO
zd3Ofd1LKDhfHD&&slFxu>RmXFQ!r<aAAx7ueIgE-pD94npMssc6<7JbY$0=AS6SOk
zTNl=D5(lU}4kY)3#UkV*3_J>roe%vWN^r+;I4BGP<9HB>^LHZ`8;r=Xm37m849=0z
ze?cSu#yB)zf~D;dFjJiXzJxfpgKoXIbaWDcNdpzt)!Esp^#EuF|1?K_Ax<#OvwTJZ
zzl22RBYJMwuEXG~iG{hOFkS|Dy&<p*4)+G;LED`S4RMN95J{yjx<ZZBUClqS63udW
z8fe{M*MGZwr}ewrRkMJ~wkx{Fep*lc%r859i4fz{&xR2J#L)VGZ9vQd56}1J%$|~(
zI`!PI2k5{;0mD^Hsd^sVr~7tU7_b+pOI$N}m(&|TZWi5n9)-7Sf{E+%9;SWo_bb5J
zonejy2O<n_WhVr*6Ug!m;$FutsGp^r(>lJKluJF(mVE+|5m(5qduVkpLyTy2^Ug;?
z18Q>)rK}VSD!X0h1z`tWO=stYu7~nC(w_>o$2un)KTY4tkP93b)b}=&=MJlXS(lmT
zSIy4O?gQ*R++A1y<^c$RhV^c*><0L=K&-u!FE%gmwwv5GTD>sGN%8bZ_wX)7GAjOI
z<xwmba$TX;e&@MY3#rj$dUa>M@li0*Gyd)js&uOA0RA4C|K`~riTuk*6<s~^b>r1d
z56@m(2J_rJnXVgB+s~)acG%eU!t<tDZY$%*bEL(u3|jf2AoH%|>4}y&>^K`VQ-bUg
z`>5hysO^&El%w9`P$OB2nD9AL)c$P6)!pg%2F~MOoDE0wTO&%bAo9n8>)5OLYqda1
z#S>TYcd^Z`X8nNaew{mWe8cJ5uK{L!(P%u~vhYYpyn46%`+6VruF>ecy9UwofNH-<
zXEdDz4P9*l&i;MLdN01M;Q+tUNX64wNJ;$71j+G<unv}2Ozmzp<r3BiTy(>chHkdA
z%k+~SqbQA#N57!JuGPRt9qO0NR-DJn_5h}M7&5<RKVc3=F(320cLY5iOW1~sSZ{gC
zWCUvUs2VN<!!}pDh4JgFH}~Dtk1)w#!kvRff3v=gtA{MXE^$AqZvVQNdITy+w%L2#
z;c`0Q^S46}EQOoK%;=O8KJAz|;f1CEwUQv8Fll*B$I`A_<>i(5%3}#GD?U@cA;Pqm
z_CA199+)NZ_zSMeF)KN(34af7(!Kn(vPSNH>5ra9@{pUdx4Wu;pM5Fod{`*m-;*6t
zi|Mc^_)n)lK^yq7K>Ipj`S|6C*41JYCqm1JYz<q(;Z|Mpp8w%0{!#d0H~xX;;itJ%
z|8?6fI8&GZ<mdSS9t?^Y^{Qkx#=}}itEe#Oyi>NV=)BBwNeS90Lhv#>Mit?}MckiV
z`C(@NXXqkr!*a7w6q&*}eEoaoh9|X7IWeX1kJsoI#E=fnZ)Q{5N8u98@=MmTu2vo6
z_3C1(zx6M3JI6Q-RLGO)r{FwswXkmIMIe<_hWFbA<(3lnacArS%LK4<0kf(o;=fSi
zM>od+=G5-%xs9yuTcP!4$;H^3#u^}W;H%K!yF;djvCH6;nI(9T2U+m_zWT-{$!fgh
z?N$7G;57iI!5&m6`^@eS8^JoUAh$OuUr619(UR5mX#UiSdGxHg=+2HV!`i9?ek;(4
zG%7n2J{BiOjLEoH9NAjAEU4GW9&OlGcbumj4D}E^zplK>FZQF3RZbhb?K6mEC#AEy
zXWdR_^}1**i4oSnvOM!Sc}7?8SUi1udwWT)G40#v2*KKv!d~9V16F{dyv)8o8}Ld3
z?{0k=Xb?W_=N)@Eee5)lytb*#I!bpDs+G*?>o1RmRT$yc$F45=5qF<lsBe@?-rIM)
z3#6#A$s5>O-)ISZ*bu%Wc^pr_t7ro4yqUnCwcvW5ykqvuH~-iwS%f(hJLR8mfj~b3
zADFbnCZD$70VWC;E`I6jhaUmX8eq~A+7j@(9WHmiHya)(R2g;!OtG~Ab1waf`5mfE
zudc^Wm3Bpxc$73<TwQ(4u&X4rc#`rwzuCXJyXV#-Tvz8D9DCR2_k)Os)|ItjLGyLB
z0bmfX=pb*b>MB9xtbo2dZYC^DhX40=J+<}a%a`teng2x)u&e+NyJ6uf^zau3yQ;cB
zy-FJuT9*twlV9R*zBH*`Xfat%Ju#*kU0=N)UA|g<fC;(#FDvvpLqw0*N+j=YPdaV8
zqX5HW3s9V=H}aMqB*l+9db;2Q&Ex{bYan5cYSyU*@tfEqnAE9P>`?bt)V~>+i1jIq
z57$cha&ppMfZzCy?9uOKU=SFX15(x)GO=jOm3=>bTRNU`Qf^v-Xh;Uwr$_)@)wQ|_
zX30-4b@d+s81Dbt$Mr?_%{t$Fs9;V@OQVu#ayBOXZssZ_#{pDaTvV==cpt&Nnw8$L
z({h*sYh28dez`o<5x1P_@4R^1IP(Z_;(y*zC;R_5&gESj0*oTKySw8tE4pq-ULhqT
zgHQi(Y#Tdg*F4(a39SbcKUN()YE~wS-5v)yl0V)*_D=tsoPAuM%k6vImb~Zfx;*5#
z;4<rTHM<L5R~+w-0eVSb#QWc%7BJ@gyyVu!sA-WXH;isEUEvC;R}D{-0EbQUtKxxI
z__9W$`jcfc)JkhjxQ2@o8>dcgW$;NzNOHtEz<x}y_O#dU%ysmdD$Nh(46L_S`Jk)%
zCCw3KpT-9>0VC<yf#lg;PmlCFq|_x!0`aMd7kUg#_wi(nPBr;1I{qJ`&N?c}?~V2r
zC`gA=(lw-X57OP;D&5_sfG~6p-CcrogZL$fZs`s|K)T@`e)q0*|8}{A*O~V`XFq#?
zwomTsbrIELn!OW(xt+6DhJ-oOHr!>TpOPS7s<6^bFIOrw%O1^x_P%?1k3a45-GTdh
z+VZWLXQ;w_z2ohc>8-&2$ohp@QPo)`>GGm(_hFrIgunT*()!_d5Feu4npf~?_QHNx
zA-g?!JeOHJZrO)HLZi)mWGzJa_S3BocQ|mkVmIJk3oKtLw)$MLX9{^wJBoO~Vq56{
zx|`W%5ODNSZgAb3Jfr!R-`}py#Lq_@-@iYQ^rCLcY_Vt?02{zp<x%Fm(=Z|x_S26^
z()^zAdOiMo<%rk@J%t!vz70fETx(_zi%5^@&BZXkN@-I<DAP4?NoRqV_oKxz8xvv%
z@6cvCX#1htC+4&v9q4Q}!7C9dmBtVWbN{`urKx)8MvyDN;;ZWL3r4bJ2fWT;J0$3J
zO7Hh@;V)dls<hs5ekyfDynnH#rPqNiynO9;yVWfmx}18aGq`^5&A;tP_+WOEcj>)<
z*gc#dX@97+ww1IT{;;@w*K+jeQU)O;dD;d#0Ruxrx23JcdRsyNd%^RIi}jNc;pik*
zqb0S}D0UxL%5|x8@1Af9r`6m&Dw>}tgD@eUAN>kB1yez}H~DaeW?`S6yZ!dirEOFo
z-EG3|<>jNF{ve?-RtY_J&q09*`(xu^R%~r#$<vw=2Vx;%%T938kZxGsw5!sS;zpuF
z1{5sRs7R|;_pq({)^qZmNGi-8&UBRR!HeO3*XXY7sNvy!IZEoL<;du2`ynv-zU<@h
z_R-zt^U0IrEpk#N#m_}exDnwRxc>L7{#g4{4_x)EM!oSPg-2v}1OC_@m))%VuS9UQ
z=6~~{fPC&EHgZn7yv%!gZnM!p&s0AKXV)_&G8(u0i=lpPub>rL9IkJBJzQmfybphy
zcTKo*(uyOcr{iz!LvC7OFyxx&GnF~y(d~Ujti{u%*8hU8F`!>nuwn)F{x?Rqq@b(8
zh)mVY^{}<r@ustl8WFr#ge(pEk6@fHp9?1Y39)hsvi%WRibY3&2;y<NqjK9xb}Mo_
z)N#eLuD5)x>2s%aU&LlR63Aphwe~jNzH8a?e&DFuljLFO>;cl=cx5_ByVuxgxt7&7
z7y59Sb<0#y;YXq*{2+KgK}P1L{G2W-(S@)#8=su0KXG<4seHMh4S%UvURb|rEAb!m
zPHnB#TVoY@t`3kgYfx*nO#Ow6tjR@aXJ_{qkO#?gkkxwgjryV0_HfCmzNTgZ1?_rQ
zjP|E7J{q0bd77U6w)dMP98B=y(;>t;R@R&1IaXjK11VY>J8?X^C9JDIOmEtXnk-Xu
zi#>6gTNp{jqZ<$1&?p-Io)8cXuK)X$wsoEl_6H(4+jWDKyt%6!;0(%y`VxWj;ArN4
z&M3U&tl?ghgp91${^1<NM)u5)F3Ij*Ry~CK$MAf%exr4>9C@{Adduo}#glinx*WN@
zAQUO=%@z7EA8ohx_f`-J5`7?yukaW;yEXWwn$ek~wFk6j;PH0!@*=`+m~J$ivTDtl
z|8Fn<ZzfFei5$}3L<0-*wc$08d%=kbZ^YjKfwAO~ZHXLdN1PP;6^QYN%{=e?Y!J^|
z2iZ?;*@k-^vdcyPk;}eh`SX^{?~yV?z8f2pQARnhE6^O(sNj@IIdC|Wp|zzt%hIE#
zX&hDeCT+ti*gfLy2lpf2v7Hz1x#XW|TGQv0E?NIkKbd{^OF|snViwtj<}MHKBP+Z8
z?e3M<r*;|6A2$5SZaQxL#l^dtQ^Y4K{*&`zyHNyY`|czoKf;H<N)Mj5$4O<4tr6e^
zhyv5+u_6vLoNbiO8XysTThhH9l$m}t02kV2<nm?YWyr%nGQe+;aFj0DZBo$Z4>=ta
z#WT9ee7AnxmkL=aLCMjM^0C()0#~cyINSLpX0nQzhF*-vPep#`Ik4&^u=FO%>fD{0
zltozap>a)$zGYqiWMrVuA0{^HtKsdDm6?~eWuXx2dhpC~Nm|B{BfDW{Ov=b;sbL%j
zyMOVcH=3*$91oA-c3_aj*PS6!__UHOXB=-O3eKNFbAcnWhlY#w%7b@O&dxe`U0&?r
z|M{<XkdAcR9QNAZNj+rzZicdq7<38V_e%K-_?^BLjz08%xB-WSY9bOBFOD{Z-O?Wt
z^~Z$>oW>GWkFQKX&Iv2<JKNPLV%0AlG^}608_K3ze7$`!+T~=3b0$MyutS-P+m-)W
z`L4FYMGZCx=1iSLvnP6+h-RG5VyG4gk^OmK){Ulq@WRK$#g}KU8Sw_d-CVwDIr1~I
z(R()OvvkBem9Hg&q)$s|tQd5W6R!C7tR+)s-Y1f#KRN$y@5}i3;Td4<AUW;jaC$|o
zqGXu6xpD!6J&g0%tt}_Fh4rIr8m4}~>^{81T;FeaXnPbf{4ebvDvhrm?u~j5_WVix
zy&7Q;|FO1X8@1p1yC$0Pk)z^S1u#BW#=5<|6$Aq9XW)YxDF4q3NJ|Klaq86;g<#HB
z#@DDghuBuQ<Fi5A2hXfXHtHvtd=jI@E($GE70JMW7S3LHJf`i{T>GC{R87OT<M-2d
z#By8-aZRet(Z=M4Ucsz~LwI2u%K1+?=0zvBZOsgfulK9Z+l#dvA1~hAW$0I7X=v%f
z^HTo%>#jCj$Q<*of)Dq?VGkQ4X^;D?R=+y{ErLdqJzO%iU$h;9)Q%}QK`(d9DiKfU
zdq0E)#3);c%uJCfJfbK9{BfQjh0=&S>0F;bbD+e+k9jHnRMo*sjjY2u5#9klg17ru
zK89Fd5L$lEkS?yTn_oPE(?V0^!v=CU-fyv5jK4?Ecskc_;&kZ3XV3Ex9s3<xCa~51
zm$V}NJ-UH!Y{1HD_g|`L+ys)V`!|UtezF)LJuWu>!PiX?5U~PId)L}2+LvDigI*lo
z3WociQ?A+ISGiyC$2LaoUkW^2njiV!(k(Ocz3n*D>%b2Q2?3h~T<}@j95O?-QpJJ~
zNrkfx{JCIQ%g_;~q!BWDO)`^!?5~w#mFKg^HvNmap@j!Wm-5Bcq39?5IE}#oX$768
zR8CPUC;!H`qMNUe#%wo(5FVVg*fy}Y?bkeJaih#sWTt2y3Hc2ZfhjJy0HQ!ec|aIc
zZyT<v<<(B*i<s|FBocI1$KE3f^3I}0ZZW;<c<AQ8KdONVUow5J!gciI2=zbLz<nrt
z*r0Q)X(#>*e#`&8KR_&uX6DQ0dk?zR=_gh8nP1{*2t(;o9k?Dih)uMJ?n#nd3Kql<
z0!(C?%A{EEli)hglh=*611~rkv~#cwCL~$wvKUpBG6h3{Dsnkh>F0XrelB%8xQ3Ia
z_~vYe?uXyHKf^TImBl8k=bIz*acJ55L3eCj&)Z8!K=SDVP@@+sfA|fJjNF%|wVcbY
zuN8N_?c$wei1#&)I_GPD4I<)76ZvV64j;1SFV_z_kL>-udG7yCQG=u2jfh8@WFqr%
zVn144baJ$e`4Jm0XAA#QW{~dUqC!nlEARWGN$D!CH$uVXGWLU~P}(u3tMtFAmnbC*
z=nNmW^V%65n2^>G-w-;uv|`p3$vy#SXJwLN>~lqfXM$sZl_eI=J9b7ANors~L;^wY
z$GZYk5}lH`q>QPvOl$ouQ+vGo;TmGn!a=W%$4M<A_A04v^EX>p5c|bD^BNtdH{rk!
z$M98xg>0`Kc4O~2^91rWEKn-^Loy_s()xR-dbrR(=ViboG4|>)$gzntF^%7j;;l_6
z)p&!#VT@mzG=i!tCGPmCG3@7%(JrovZZLl_({jk#sJj#iQyNUI)%1NlvJ#?G3z;Td
z9CbJ3k{5gn<V*FD+84hc^KCP_*0o|B3eT&mQF7gzmEh%ybgibI`;LnxgdFZpKAd?i
zhFV*EOt#4rN%D3Ap>-4=HaRXz#0PsYY+hY?(iwPzjI@lKrzauscz<vgM*8a;_tJ22
zdtXgD{Ri49mx9`GuxWx(e5ARf$KZ#)z^_W(DlfYEQ1XMSIVcE-cbt2z67CWqw*J8t
zWgd_M3MgN60ez?{A=NQdjc<G<^zFR>eso1sc+wgCc%@goU5Tg;nul&`9Qw{XE54}b
z{ry&_I0Q4(j;7FT)^n~~dO#{kyzF&4bu~Rnxvs!Xc3x%)lBjjSh&V+S{cYang)i+K
z+g_0$L?ZRV5nn`#!>i?wXsP0n4C>&z8Rl6-+c2By<;LIw-kAoYX>*Sl>^6qX;H|KQ
z3x58#Dij}WEaf1F#2X5mFk%1m!oF#4d!xDvf_|WZt$sKV^!H@eHF9t>d+N4MSFgJI
zj>o=8`T6rm(bw_^Ia3s(*p5g6)Ue9X*g9_uhhmG6gMx0#Uq{F6OUi7czU!!JT?Ws7
zgGJ3t*PRrQ6<#A=ph5g(fpafgu^L*DMeM$tu&);iPuc093Tkispjn9)GYUb6%_&`c
zwZ#p~ju@C~0ax42<MSWZ%)Wc3Yo%J%2{itu;dqn~4B@}L0<lp(Ua;x|4HA7>9Mi<e
z4OwcnH~G4P*_9l4XiIdTlD5o<PG@Izs^YNSOFB>sj|Wk79nOq!yJ}cWg*0fNtEI)z
z&VuUz{X;uuor!GVuc8I4O`KPde${J}1OcfW{H)U!#D+=iXY9~uQR1M<_C2*T58;9r
z*B-yT3Dgrr`b>i3)koAI@zKDF->soRnBXYX>xUxLw|M4f+BrQ)Fg=>mlZ4`F_)?|m
z!cuLM&5F-WJ_oAPC#^%R<gw~(M{y}2urnt_g)#(&V{=O^ZrK>TEM)<Fm(4L-?6E3_
zfJ#*oeYlE)3)(M&fnJJZnQ^UP<%DaR@RH=0IZd?fctnAPx(Avm<<Vhkohs)F@pj;Y
z%lt(2o$pg&vXfLahX6}?13TU^4r)=PpF&jPxeF1J)Y%%?EPwXg5xDarfra=im;vSx
z><u{LkJw>U>GC(Qlhzwy6IBRb*Yzr{nyi;cKB>OREc#}mt}QFFExAkNj0s}ti0c+w
zT&16soaP6?w6E0WWsE(GHB#Byi~J`3sDd2koj9IuEFK+x#$~t1OJnF-zY9OJz*pbr
zN%t+`xtA~LianT(jMQH@*F+u^o*q%>djE6?Zl6q&e|i<nO(ddKYDeAY-t4b#IQa3+
z2%*?mO@&7*nG3T<1D-VcqBWBRL5+8?g$t?FkhxnAu9GSHpFrJPa1gE?t@j?Sb%$fL
z4)RMN%gQ{)ufg{jSWJ@8<iBTe{G=Ig$3y=Mtd_mDIi^OH_9M#Vctu?Q{7kd17U^Cw
zuPfh%Oh*X9c1UH**a5v_4cKM?TWb67XnmW|7mq4!Jeq8Rpdc07@#9b83wjw8h=iM8
zH1fK(3@JpXTaDfi*LaLtJ*P~bGJL0kq)}YP#!i7(t=n#W7T9fkrT^}yOe8NHXujQH
z`S(lJ4{dgBZ?=pEMq^NW8e0)DJFBs8=9n9``*>DH4AzYa)u!G?zOS&g2hZExU0`~A
zz8}vDf(fA(g-(@0YYtMMja{Uw;-4X*Lx40z-x3lZ9Ir($Cb(YmG4n;yL-m^IS5fLP
zd=N=o>v#lB0YMz_sLajHzeR^VSK({^HJNNtJ=TWs_~Uzivtd;-)H;Hn<Uwva#;L$2
zD5}gZ%P#6Oo3yY7zO@0_rE)Kc_;e=@tdBH<`|(3?uacSfFjRsQd&VWh{`=Ny2AP7j
zhpulHM3#|I&r7L4t9}w8A-;e*4iXHr>y#o%UsNq1D|qbF-7A(^(--Xmw+CuiTuuO;
z5OdW9h%-!CPJan5@h#1NB#F^b(v;qbq-r-;SZ}0UgjB?z^q_yKux01dh6ZME;qA4m
z6T8g44g6<=nY8xJp<65RC65BT;V)itxMtTKFY2Pqa9fpN$(d{XHdzpFYYx}ukuv4v
zmizQ!mO*iG=lnT~y;vY^K~+nOUXY9<Rz(QUkbzs)ORa*V0+;-JE5Wj}yl8XZf8ztE
zH`G{}Wt329ST`93a?TS49HBh6M;TE>OaTYE{VYP$av@_vmFk3$-EVY!ysexto=7};
zhXuEcnaiXP{)Xn}=mKD<ei0`E6q~GkMC+J+eE$aN8s%}!XB^;8WFj*<wx-|FX}M1d
z;Qjr?3W933f6r!Gil$o?VfD_XL&aR&DYV8i)#>jt2m}gx9_M#)a5mOX4=#?9O$2MO
zP7E(@F9f|+BrT^#S0JyCwpDA=`Ae~;NP*)U`Zk4qyls~w$)jIb%6T>{m{I7k@}8Zl
z`_K*^;v_*o!Wm#C0B61tX=k_`R&Z3GjzWH?BK>sYL({m1-c`m<zL`n7@S$uTeh96`
zRy@1c-*-g0(C|Mq*DH8Pij`E87+tWIh6o9liP9`#e?uSHs}0#n%T(G&!*TsJxItE@
zH-g8@bP9fM01WH`m!6WWfK}fTY<OhLyWu1v%o=u$iXNq%pS>aXJOk4>IhyIed654j
z^eF7k7o9#svuw1Efgp&g_nid<hv7&vvqrg^2PL76;pS`u@%vL;^>Ey@r)txzP*b0O
zZl<<=#i|}V&UyG!02r#zvVMBX`HctD@|GZc!JH<&PyH%cmuBVTD=RDZ9Yd5CtM7TN
zO5Tano%)B$|30Yl>Brj|HR+G@tUjhZcxEB0?d4vssSG%9mCUV8O@Gp_?}1w3B))QV
z<hUBXMh?WlZm!wI#eEam`u+WZ5YA9Q9Sy}Cohc%Yt;_Q1^GI#_;WIsg@lK4BlD|>9
zq9Vc3tsNQu+CY@lf|`aY;TPUBce<PFy{F3^XmfoUtHVl)VNmT%bX8l?E0AyknIJHi
znm)%#8cwE}<5Zx#KD0GyfBuQf*BK?7Of~-KE3TQ%kEAdbXEk+h=H!O2BgOuLI$<)N
zs>-URL7M<q35*}%4%ReuYb-Ct=%ZeeI;W_enHrap5KNWGq(GvoB&pYn7UbU{^7n?-
z47ADvSoFIVTdcLY$4Cq!3oZExo8qV_)S@Z_iq_x_(LLhoYD6ha7)LDLg_6V+82dcR
zG=`8v*W`T%^%*NHXEgDVIr{G`t3n{@jYgmnoDi|Qs+ug-M|Gg-Y2rvtfa9f9WhPub
z6$HBpd4YLVxUooZF%&xvV(Xp|PWL&Mq=8=oLS+{{41GR+eug-|)D629ryZA5$tC`q
zL6xqe5AMdtY~o$=JB&ZcaqD;}+;ox3ZceI}mKap~>zXlBOhCg!P5W9>9Dq!^#>P};
zwSLvlXln9q9hgP26&u4Vw$M+U*@+5x1z~U$_6{{|>*s$Xhl;nFC;RAdqHn_+Pn0|l
zf)ZY8eWW~C(l5MuC5*(uHAcYcqSg5jJU86lvSQ*}4wsqi6iepP5o*yuXd2UXnq)`M
zHZ=EYp%A>p^@p#o4osuDNG`lf`7qmm3AsN#OY^}iK~MKc$<4(9Mn#h?RnlT!%G|#*
zV+U5bj7Sq)Az&G$wq}2SE=<VGFe(2PV<;tryau|y$yv<Ht4t_<oUORkEZBA7wZox`
zs%w=2v>QLarBLmYQ-)2#zJ_SnTF-kK*%2vmzNg`dBK|%im?hi<DPSEf=ds&I!#IK8
zpvMGxCh9;zd-Tn(ZdZLJ{(VNLQa67WU3Y{_mU8f1pWAeT*!i7-G_LwBwZB#T1UjjT
zq-MR6$5gk|ZHyU&I#Q`?5%C81ZCl&zJrwpN$Bih<Vlr)k#P7oL#%9*qV#V6QyWI86
zP3Dp=UWTt*{wpKW*`y-vi~ef*QJxG%V`!H9dA5Zr#IV}1a$x!dtcRE8Kk>VLbX|_8
zQ$h%Qyq#q_J^<gls2-(aD*(9y>aGf&tY|Zktfit!;oBR{LuNJAk(nu_r~!(=Pz9h$
zYZFSsVEwt=V|3eXgol)iPY2Q4@9o!b)~ygzz>cEO%1`g&^8DNufr4V-MNDA+FEEN1
zgE}V$cAz)EP>?jf!}01+HoakYM473=y-oc&WIV|D+foX!!e>JS4&~JCY!}$4Z?7rj
zSQ293+wqZ`kl%lLaU@-8By}*p%X9aoVc1P@ZhkbH77Jf*b=i#7`32j_4uIVVXql2X
zK2>mKl*}Q(*1~ql%Dx7iPL~m(7zqu%YPG>@lTL^jMT&lm)IC}ftW7^E$4!e7B8R#U
z#I#nVn2J!ms(TR4zCelGHDo@-)p-<~@>V51%2x*~%sW&>fLE{tFEDN4w}YK7Wz(Ar
zb$#CoSBL-SL3m_FzC^x>nxiPl85Dw-Jl!J2Rd^vPhQi|BUCIK%xhhRn_Yrw-MEJA5
zRG(px!tdd-SGc{`|9tAsUdw`O+!@?$UUVe?rb*a(s;CvG;QW@!<k{pv600#F21Ti>
z%S+>aU?JBk&`zR#cM)zPhzJPYP&tptBtAj&92UgKPFlN*CVp~8vJTxPGqA$MF%PSf
z^bLdC)sE<BVTgrY8n)Rgy}r1(km6!r!&qu`9nL(HZvd<@kK+v)AdED-aG&gN(9Q76
zki`yy@8UYmWCY<qKb;o}sI-Dd)Xt=sZyl5FwqJHwh>nI~=3pzT)<@p1`})3-F_r$p
z8QZ@OyScH=*FSAZ$g{sO<A)*{6$jvsoap-${94D@Al&?-cO#=y$TmR7sx7sgzbd0I
z`?82h!o~fSvv=$6EC9W>`X#)A$LqNcWW?{qQ&g3IS9B`#)hIeIf4E=K`p_(iV4$r%
z>R-K3g<akN9Hh_LlvYMFfwZxAFfGY9DLVIwI7QlqwASRGRE(wjr|Nj%alVgvqZJ>>
zM^-n{>p6s<t7hg%0v!-l$+nUeyUW8Gm?FnHHoheMt==k%kPJ}d`JR|Zq1v?Z^!vH?
zj3R33^s>Ap$NGwsUYdFW9nCOyN%aDZzUAAW(&EJpl^$ctU#V>Wpyw2xEFVuHWJPQi
z=BM+!HNSM8B2kHvq<m>yLE~C_nK@IE)3oo+pON_7o4OY9q!sM@JW0WpPo9)E1oE*o
zVgN>8#)Yl&><bvu(Vc4x?U+-Vc6`Z~^ZS0AZ+B`_q9yJXFfT`_@uI3MMOE8`BB!V@
zp4Ne56#$nmqA0YP-q3uCKowgK)h6iMq|S*QcSnRWK3$^wmSP*@oYz~;g+cip0l{BX
z`F@$>EDN_}Oz2gcfJvuU61^TlElf*NAw!Vmz1JSbid1~!jIJPk5Fd#C?~tjivNPoi
z7MI61=ZEV`?p+7c#s8e>i*+SF%~RtA^!SYIIyL=eDjQ3cT+qtL8=fP_X_o7p1Yw>j
z^dsFrXM%0|2=OvzuPDCvsc1y@3?I(+H#YJ~VEJrbVA0@nd(?Y0k#)~Gl(6xe-4g+4
zjD3D1OEcs+D!w6<Qn1BhwOybe0R-f@OIlKky>H{)8Vv_CR_Vi3Dn{y2F+VosEP^Ci
zIaqu*PUyrwa^+8JX-X0wMiq1)1{QJwF;jd^{2#kDYM)GC_E!_y+yvC;dUz~4y67x*
z&AY~g*zS(cn=$sz44h(w4~xs{H#WoUaIHlU*?jA7s*mj=`{@*TuhwU=A4KhaQsA^x
z0In<`=uj*3mOK6_SSO*yQMD3@4w%Rcd?ys0HpjdisQdUn169M^M%=|nwHr=~91dn$
zH(I`l-~UpG=8BIbW~`rjCDm0~xKVYlPB+`L&v2&cA!BgY;JH2|M~#xfF-K^sU<X%m
zKHiB;m9=fOk#oOV(XJ}1nNm3qkS|xcc3C~+(b<h(mCmeQ6Ckpiq0SFcLHg#3J7*ZG
zlwduZfg);n?2uY-d)VeLG1GB-JTM%zwmfR3mSumy-brq@pZV952WtZ!@x|kVjyfqa
z2>T1H|2j^#BzAmTyJ5Fj_=k4okM5S}R9{m6Sp;$EoN5{40F>y=5Z*3YD|WoGVe*eJ
z1n%eWhIdDsABB4y7pfbuF~73mf4uS;gF%bfL}#ON9OsXr=ux}h<S4Qhvz0H27#;zG
zMeNN@X_C42u~)tPz+yIHS%qA~N7*Tz3fZja9D=!}HyKmhA{3>0+uV}JAXnY5k#=@{
zKp?z0@Lx&4QeBekXyawQAEg>Zt63~%e24g<U97ro+4mwcr&9wS4ZWrZyopxN6WZ_P
zz@H&3%=*ssKlZml1KAZ<R+^^e1-7Y$`LR;s{ZW%Cgg_DngQx&2ZZn|5b$?b-v{S`8
z1;_vs0fy8HnYxY+BklpV<N_^HuUm6@GL*0H9Qn=^IJ5L?v#|}_hcTNs!@kkJ`<@)M
zyhzY&?c@zsPE3I5TzQBgWY%&0Y~99N!7;|(ia3<z%^nC-6*7fCHDHl5?@N^l4$}}{
z$WTqBfWn8Wvg0Uo!*~3vYNJoE&W7@;Uj1i@<!C~b7P<1oRNdLURZxr4?~q`VVk5hA
zrex7-8@%HljZ@&trs4un^}w1SCz<b~G*y=TWqcI$0X&+rS)+-R$?LhXuaTwpb0nPR
z#AD55sn+bMQZVS_3R=Jk-oN4IAAMPT>T>Yr4H?{z8M-rwx;W4I2-C%*?&W}RhKAg_
zjy!QYrt#*{O*c^=OzF%`$F)^qLt*=;k13_9uS>H#N@gS}kw<lJMbc<*Rel@Y#+s*)
zwWmP>#@kF`|IB;ZV8@14g1tWPrm&wk%iof_`^o9utD1>d4<r(eetslJ>KQ0exQ@8K
zEa2ycDD58!8*GRD2HwmopCuh=p#+}Os0o$+#St!RzK=%XD4CWTC$}-3C>a~8=>8F#
z^K%K#O{ZSRscqnSSc?LTZuZrt(p(OiZyx}xx^+^nN5V9`p&hTHDz<exVze@R!FMlD
z#gpWKif3qXSS3?VrHnIh#EvGo{D>e<GAZVDR9PECuTgpa*Z7|&4rF$X<4)fz<=^Ly
z)T7m<RiwAL9De5WQH^Pg)Vmu$y2sgIdMkQ6W^(plXx(12-eq)D6lfBtu2EgNqcvqz
zL4%%L$1Iq@%A$LhxDC=qwoTW{vee66W-J8T43@OIdus1Q4phZ?Q0&!ax(-hqO}u>!
z1uCzLc{Xo>GLj*?AO-JU<9p-#E{7{yW5;#{8CMz+aAx&^#Ok%vbzI(V{x{mw1c!eL
zh7n)9zs$_5TJtdH>Gcx{Tm33&XXnnYS9RQR*HQic%CRGcHdOz$+0ZWGQL?CXE#<w3
zX6b-reU>P~;Na;Yg{)I;Cjs)_zXFS?_#+KR#wK0CgP1B-rLDLCiU&`Fu0hpiKHZ^E
z*e;_4$5pYywLF`N>9JwIXe768)Y(p5fZ?3iU-(VFKiQkQXzFw#2byz-f2YI~Av6{d
zTJ-OwKW(<*=pI<dzW%@?6>rkF=}`Pv&Kx6$qX8y(_JY#ytuS@M%*@P1lrpuxW?K{<
zQbW^XdRTye`i5~R3rzX5f*mi#ZG`sHe=>vZ$E4>v^aGJa3#>yZ{UG*_D-m-ju6C_Z
z#m6ALEYS)v4eP?T`+iFi*siigFFf#Bs;ey&WK=?l6y!&8xBNO9awJj148cwkTypzE
zv!7V>?}5PEPm6*aBOvY9wkkrCazh)1u5AG_9>}05s0?od5Fy0{)2t0%sy4FU+`Gt+
zU-^n(tQe8nu2d19>1aM~v;ZJSmLD1$O`C($acWc{)3ra)CVYRnu3b5EbiO}7Ea1?z
z_LIG4D6Ec`kdBWo@LIyqp{~OIa|FeiN_>V~g$varG_!wqC-tEGpujP^+`-bwcV8@(
zD_8Vno5V9hg*HF(ytMeE=!|ZN75cv$1IF_l?1!O1Bhnu(St(G4B#b6wl@SIq5g8S{
zSM8nprsmTLxV@eq8u~o*&UbXTPMRq{5!!mE4feF+#v&-^)(A=3nff)J9VINgBKJO$
z>OolD(amidCF(fAnE{zhfau+EbbPH@MBnR&k!mZSfq;T*A1^Omw0`i|njS7MdBF|L
zZ3bj+8wa;fa$oI?Uxw~rtC+|m%c9a-myJ}RedA-FboR~s(lR0T%5!~C1wR#~9kZE-
zfie1rTSVdd#&GV=>ESGff~2L_fm1?^oU$qQtmKG54&yHZl$U+@sV5=J{V#1_t)Doh
z5Tp-yfGz}4lbbsjzb^5S2uwav>vj%s+`)!63EdCagfei)T1lh~XI~6m#~uyc^sz6z
zYV!VhZZbUuPwkP|f75KzQj5CwDm*MjFqU!XM4#H(3$XCw3I=P8eKIF|gl;PCFOIx%
ziF~){J2)I+fty6C&kGt#&{2NUCS2JHervUD=^Z&CLAA|D%8hmSfr;2PCgR6@pIAq(
zyIS~OWs7kdxqK^CxgU$mlo~pj3C<R68OXpNm$}f_SYBn!L2ZDCpkMoFxI7+UrOGXX
zC08hHs)!-^8ZVlQ++51EI~+GIn61|WAb2mMEJk&^B|Nq#&{#sew@tyO458vp7*424
zuw(=m`7d_rLYWBeRPi8RGE9`_+YUsJe2)_zvO#%XtE{%jfQ8TP1ce!q`XC;JOvWdj
zE{%HcnH^X4yva6Q+pps_6e-4cN3DEACHTuVY+BRx-vTzP+?n!@hA_VA(6>*%qLj5;
z5w^CUg{%V(eEJU7sS%ku0r@S{oC=}NKkQM%2C_l0XUTgY!=r14+yS3KjnYR>2)C=1
zmEA_CT$(dOae%O%pUw@IWNeuShVpS3L6UzCN}`TQN7kDp*T(FtLR$3Z+(S@+@Tfk0
z0qqzGFABht6j{gdf4527RWFJ<j6Xu4XCbibgeHjHVskM!YBT9YOYj_fv2c%-!{DUf
zlti)oW+(nLim?An#5+3uGXdzjkG^ZLjE3L_R?fL(zTX){5hFE~&8xJ(*ZZE{ZhlE(
z+8-&e(D|pMei(PwsrP#e4~cO_%vdC1fj{o0((GHJ5ReAAJe^PiVN+hiHm@lW?Fi$>
z^7bSRIG{($iE-`qDUQ>f^nW<cR$msU+BkO`r34K&PF;Fki-dU1<^cOdEeLV@0p^7j
z0nCD_f$)4@u<J&pRN=3i0~RX&XL%xj)e?s*nWEORZZ4;!Y`rg3nF|)JHB9StU6=e@
z08|`UD~|w~_;IHWDAJF=R7=NA3s66ag)ciX|K|$>K%W*1a8DcXEA0M^4NNIOLwa<3
zob0|hw4e6NZDb2BIrkmpNU)?-LSo;XV~?WXTSesRYZ<Ct>6)r4zQX5=M1p&gJ<||U
zWU;rMcf-K_{9rBXR%|#Xd8jeA4MM_z61hJ_wZgT<ZL&F`JL#hxGR2P=&S7(B8uuN|
zrvDA%4`n~~(RJP58vgQaa8Qq_El$JwR&?%7RvCgwd}dNlNr!fK3DkjZkhRpr>jz>F
zFroJV8-<|XZDZ?|!U!Zp0+Dd2nM3pT#i4c4L1f<uwy)7fb$qNMy?tt7>x1UNxK&Y1
z3f6Cc(@O^{n*oi=m=x8goJ!zT`x)Gpej`Sy3yF1FE>2Hfs2UPM9M1#6_fX~tIE*DL
zW-O9#-X56~JUN!+_`Pv&i9A*;2-Y-T60vRJa=a`O@fl7m9phtfm(>_QIqm7LD@Bt1
z)so~|x)7p5xHT}mVzBT7B^`hW`p6CFacX+>BnIT)pDI0Yf`Nwg5e^%+3S>*r4)pqZ
zoEo;%oZ#ZlVIo5_&6C1YKxD7}(lo+$)Grfw1Ppg2cqC=x8KoGsW(u?)OEDg?fuEEw
z&WDH(1J%qdYM|z>r~QbSH3GQPLCmZ{a5<ID2cx2rZ|`#X)?kef6>r@G7d~5yUgkiC
zO{O|KXQ+)8gSI~uGbxcRO=|pEr|^tjoVXyaEKvwpwQ}F0{6pAamU-HH)8zC)Vna}?
zt@(jM5OF<ko^G{ayWI%-AO7e0?hb#O5eOF?3)lNMYfs5y=Zxj)l-@uUKc?H!C$}9L
zZO`no(z?=;n}LiSqLn)WZV{UqG_CHqh&8eb&RVUnkltotF@}o4My}L1tYtIbF&hGI
zGGBEh-F>?0{rBZ!W{sd7b;DHIp{DAiKnmusYMPG<6Taw3^IOxi*c1cnviLk@kt)TQ
zu|>Lk#vZ>O*8T(DBVg*g&5X;xkt-l~p=gIiV=t-=8%SO)o<v+Hd_p(1kp;Pc_*U-1
z^0Ab>^m1Eo_j^${`(>_&MU`OS%(=Us0oI!=Vb)NU=8#vN15-aCOQQ`o1?KdEDL=1V
zS)<>$sdbJ7rH1wS`Y-+Go#Czs4fPaS8%9NNt&fF91pa8ZaSgx4QEJRfkfy6ioYkJf
zGn+QA6kx2Wgc4!UHBnsuUFNvpAE+KSp$^ya5Wb!pu{Pd}uDl>A^O0oj^`Z-X#m-DL
z7O8$hc(-+7771R=p;-^N+ZWTFRJnOaXRV2T1wvP}+?8oqbFtFX%jFM(QKKVXKE+5M
za}yXNDAdQK@YbL?Q`L1VygVOc*;gxaDs+GF-Td7jn%k9SR3oX7*0D1y*S1q|oWio0
z4k>5$V<rw%_><#A%P%PS*N2R%g*bx`aq`US9UWiZZG-&pTSFAqi{44vC|DV<-Ca#5
z^%Ujo;7DJlI}c2q7^Ztu!dWb?FLIOrodW%slCrc68ZakoF$avrz;dzp+ClAz|H{N=
zW;_yhA%zeY!`L8(w=t5lR#rxG9Vdshz<l|z6fs5CE52CKn@l(@zqL(pQWW+^I}`a2
zCZIiFs*%Y0sZwIrmmN%C$#S%PI@8&D05NpU^*z&pt8?$OQAlxNmlbR{YalavhnBE}
zd6?LD8Jy|B&IP&)B6;kW623q4g?-t)3R4K?x57=mXxfAKFDUbIiRXOav+R3P%_0?O
z7f}tcl7$weVI=p}2}qXzc7DR_N?pZ;U0jHWe$Q>hsm2nkrt#MCbM(oS4c$Jw=<mhI
zZlPTYQ78`X`<-dCNfp?y9!2uDVa;xpE`I*M6jC8CJn;2?{O4VKPEtP~2m02X4!PV_
z37Yy_RmYw+Y5FNZt#+7Nl4Pqa56JlSsW7v5Q~vTrRc+kmgl(_}O124br@`trGg38(
zJ_jgiqq^;L7BPkRmm?itN=h2TpTtJ=KC@!^O+Y!9`#J8|In$Zl1mPG_7ykpmW8Dou
z#bjXS9A-j?<MTXHy$aE8V2I-BrMa`Y-?pC#s++ml&=8O{OUIP-_|&a~v#dJtCN{QL
zOMA{|zN3<`|Gv#Ii*9)zU!?xreUbu>4@IF{{lGon_^)SnluU##-ZlHTwa?4QBz=_U
zOQWwCMDTa8YYqB3-HlbFtp35u8LU3>T<)MY<k;?vZztt<<k*NOERcW0ts*!K;z-G)
z%PneY&#5c6nm(2EGu0$>P=@vQ<o7IJ><`=AqV}bd3HMY_6CFbSK;PH;aD6IQ$#T1B
zzMk6|If1ftk9CY>mQYkjY(zUZjoevJRXR0#*S%r<$K6y$OKN^#k!6$>8A6psaZJvk
zgh8wqY;SK4%xV$z#<Hj5Ul*3N{CoOMSqEs=d<}ycO}2e|W+<qx+bxW{R})Z<Y4a7Q
z4L|1}*9<}xN)-}9jGXrdtW&a5nB#YAgtMmLb)Tdy9glZ9lr)91blep`)Wug|46J3o
zyX%S6w(&-HGrurZSCfZw&)ym1hh_`ka8F{4#he`yH|Wk+YvjXPp2xs<c1%NAiQZm3
z3*Plz=45cLJ@hh?q&j>XIdyyAdSeO38L)!-(X6>0GkTsAL-G0A+lcJOB7!*oM*rzf
zs#0Tn=ypfb6hDw5wu5&wnO+TEpUpO4dL))G@*bK`EZ9*4+r-Y~EKe}uUw18)ERGol
z#))1|V&nRzN?S!jr);MazHO)OO;6Sf-?8LsNeYaSQI}8Tf!4TjqU|Nto#w8`ac*y>
zrO)1Py31$i6AdRk)glj0EZ`9l5knM?WLTvZIzMq3IzG8iEdy0cgrEqZb$jav8krbF
zB^04AuO%2wD#T##b;!snCWr0m_f0=Yp1Ub*RnVA_km38;*hShY;R>w9#YLE=vat9X
zL84|aNwUsnVR!mE#VYcjcS2K&22)2Yae6hzByy$g{BUQp4^3d^cX)sObkG6?-?!iq
zQ%%d_gSk3|(NLw&^{y5?cF0%r-2P=LtUPA2J@Z__J=!;qhBmY*g4vV$=7Hins(VQ;
zWFj5K9&?rYIJ!Vu7fEoXRpsC`t`~;|Ct6wyl7?sZQYMLX$1>|xf2w<B?`EKKlf!S_
z_4J?Xk#5LUA(z$@mIQ?HicWz}Gyk`E+kv^g4%=LMb_8B$a_8{Dk)soJoKRL1a+>Op
zmG!f!kC)m@SNCrZ8dUgDDV-l8=_t*_?`tye?(Uo{6wz(X6BJINE}0Z}Q0QdhXAVvd
zZV}ecT~6g3eXmVk7*&WMmrW(-82DGvaMZ3`oW%xFtC84fbuC^Mw_BFcqWMmSh%cG0
zId*@P7pV5u*D;fNW>U<0Bdq>5Yo@Dw-*?P6t|%3!Y-(hi_fs{edo$bGbM(zoYN}Mr
z`ST@&oS98X0!!A}&8=?agE)G{tKMk=NvKpodd+_+EA$8C^xA1Az5&WD4N9>QQ>~P`
z%}gtGhLwl!%Jpmfs7K1r>UW!Tjkt4i6qM5i>-xLvO=OGDnfg_dm{{heMbz2Vvwr<n
zhfFQJ<v}B?tl{Rz49;%B3sDprm|OK^=jtBZzPw-jdjUlv-_<Mc%bdlViiA2R%nj64
zF}))GzP#WR9!wjB>0yh5CgLXbhH#u<_eR<KEZ6#75wGN|M^+C*th*z2wbxnO$P`g&
zgROn^n?bbTmFk#CGVP)I9S2K)b4%`$vqkukcIU|TN?3elj6(tK>7YjU-=SbtCUHJr
zTfF9x4VYl@&}%{r7UUh$wrdwO^rI>*M~`ELA@(ckq@R;S8`FsX)g^5rV;bxIKh^^=
z!A9RTtF)OU8%~VqOm7U*cUU55-2L@Q2_`F!d_b2`e}6wc&kv=Z9@XG4gm0DCJ_QR_
zL)n9Y?{r{ajM(?3@s(gXJjaQ|RoPNly)#UbA}UpG1@=Y7M6##k`y;48$@h${0!IZC
z+WwlV`s*wn(oRad#YGF+xY$BVPIa&!@os~jqVKsz$>8tPG0-A1UbTH<e7*bXlZE~Q
zmX#ogX7FVvAHJxdpf$tTgYK*K2PBYESHJ}lCudSc7hinB94r1WtVYBER{W46Cpw<P
z<GZ(&=0a|5uj#&3txPe!5Y9H&W;?G)IVUy;?J?-1cc!~Ow!BT95h%oT5%D-EjP@V;
zZ6b>Z^x~rG_P!(@n8-5BE1qFC6BqA%0!g~9d-mV51wDy^vLo*==cHwWRSGS$2+@+o
ziS4DGzF$@Bg4YSosA0qO=>P-5{4eF(KXC%G1d1{?@k|?zp1K^3k|`d}FVG{afkEB*
znWIZF&iXr>UM#CMXWvb8L|&y|TiO<=e>3D$A7mPd-}As+B)M~kz7D`tV5ye~>+B6F
z=8Z@3-NIo5`vO%ta_iI<vB(t3;r6{nC|Cac_PAhI-=uyPtg!+2A_GrKa_&?N`4X92
z^8jsiSf=6DHvwKbLvHbTydk_e>sN&^a_;r6O2?V^GWGg}(YAmH<?d7Chg%lSQnzLB
zAKmu$<4jM2a$7JruZqE+%Vbl*BGYYi#W^&#8IO(P$n{FMe;+?F2okEAahKUGJ65nS
zo_W6#>T3IxMR`!&b;JIhN|GT8+)**i9ARXRJiiT7{AL_)>z+pu5|_9R3J^Fl;?32F
zMnX+=ZuD6&DCa91>g2FO<``+VTw58>boBf?9mz_a5X##7vo@*&2CFxcR;C^Ym;+O;
z{kVIjofZsC^-7?dqnG=Lj{(^~r2GKezpjIt18VEtGRK=Xba19Mf>LGD`i>pd7_!dl
ze}eaHTP70NzA)C#5HC6#{;$`c|4N+8Lr3iu7b5dQ0Wifba+l{yzr0ZUH@J84vw~gL
z)!#N!i1@d{>K4JN29>Zfopq9)vtmL7ucX~wn+-!W>_FwKj_{PCnVnZaT3>u)anSE}
z`6iR&?Gfi+$0GxvX=f&2ygOO3h~{a;U3jC!e>PeZoEbie@}D}C#Kex<J$`>2yk{I+
zF#t;5CYMJwbt<pd><>nanH9iHO>3|WJsrk#qs^nR-ofo;{vgR>VH5az5KR5RrR^K<
zSNvV4rPq+`T48?(??^et*ki#RlMVs?*yo~U8PqfAoQIkqf)Ac78ZeL%<;p)b3sd&@
zi-~H`3Xa{+mJ`op=lW{7YgP8Fib1OieFRZZc6gw?OhRBS-$UMuw%?>mSX%kk@v%RJ
zYC}9I8Wi-MG>daDB~315V2KX|Xda59X*$NWIPHZ8?vP@XE^SFts|Cgzt2{SijzyYT
zf^#C*u{UXnb(m}FTuk?6=>S;|Xbs-1V@^ofj@e&<ao|KX4_2Z2$pE@)ZCkoubbPE&
z2#Lzchcq#xxlMcrf1&$MSuw1IUtD`vk8JJ4kwA`X^3Q?H_e^&ujY&sjzZXNN>}!Di
zwEM(lG**<o@V~lYQwLSs(*l|Xn`X)733%#g3~l0p3XTGLg+2Q!#9{Wv=-L9N#C7j+
zqh7S8m4ft!%<2ZNo#;2iG6Soq!geINBNDb2?0J!Tv)gyMiEIi`#i#I101*t}zN3HH
zCYWkhndrY*?VZ(q2&GCo8?cNRjNmde9st6<Pg8aBC}+b!xX-erQDSlCP<D(Gh*f5^
zh_|)_iUXK-9%;giAAt8|Gn=4&z!@kXjmEp-AO8Yi%vY6+y`THDXwuW?yC&36jqjf;
zaT4#|5IA-9zRM9O5>>%jdAt`F4&~E|I6#Pw)QQ_J-*D8&7t_fdd?vQOBCOk-rEm%y
zyGZ|RY^TgQEGzs^=0L=O;j<wXrnDdLL?LG|X8QA}>$&<!;0M6ataukE5Ar(2SAs#=
zb-~sdUgh=klhMxVAfxf;5|k{>QdZ*Tf#E(<b1v|w>m_i1uVoGJEP;NsUwQ7<;6hye
z@Zl({un9dmvMz#psUuW_U`^@VK(s;(E>!qa2xsjpnH+f;BT^|!_Xu!}f_Napd9B1_
zA9XzPOHaZ~4SpO2X=B9bUEkVABJ&*7#Y30_e})@c2^HA5c{>ZT&xEoH<gt4hO@^F;
zM3HzK_yN4r&a60h&t5U4XT4)KN#qVZR)iki6l+57#M@=`HPsH_F@g6WhG7g0jH{z7
zmJ_Ln_#e_ci!Tea_QF0EKjbqpJYelQ(**qlF{ma07kKwLzM&YTe$a;|yx)@+VGaa6
z2z*VlKA`Nad&&R=oqV0$3%G|_0(FF~eaf1|2_bHvS$Vdx_%iX5{9Ca#ST5fWo7D`*
zl_iwo4THR$3e!fOhg|n*yR^>`4i^>WBcSH#9}1#{jYOLA9}c4GA*QLyT>E?*&iq#^
z96B(>hv&3?n}-dXUt4Wj%1b`BB70Y9cY)%c!*@3$gR?M!KH+mtI(S`488_a1F*ZI$
z&K@v&re|gK0z@d-mi=B{jy3|VLm34K(5)d;_*r=ZUYPKWyBxdDqXP50mzg-by6U<_
z7W=1nIv)%#JP1XNeI&zv#(pytclGr(P!2~!S2|8gk2_6|FbI|nE_SOVzMu%?M!vE;
zMeWN$><F8-Z_++n-hO>i2M-)6;MILnsiqlpMO^(v(4!A`YU8??g(4ZqJdyB_-M{Mt
zg?92=M&5q|p831F;)C_zV+qSb6>OeO5RCB84n-Dh%7G9Swh5|BB~TvSnn2+yvh`gb
zyuTU9V3!!DGdCW6W$^k;;h2JI)$ti2q10MRS}uFdZ7f)7l^hy``t;4Mwo)pEMEw5f
zppN@xpn_C8w9!AkJ<4n5jteI4M6t@>`m*;Ls`eA<fwK9K<ddF(b`Xinhj1q8L^l#_
zi9fHZ1+Gg~(5p{$bi+d$Sq%&41{=a<KC$!4*2=ih<!q~#C<nbX?AGfadwk9yVTJ`%
zE?-{_;+j3*5wl_m$#1f8$K=qy4}FoqNttUHw-H(V4G|(O7c4D;&|MC>RNK2K3@(b6
zY$NSl&BpwpghTq}SGE$e6bh9N>%hod&T?xy5VS#5mn}n4;3*Khb(z5Z-4FzAy7y)(
z1)wFGlC^Ub;gGMf5;3C2qyskG;g8H0Z|3x1S}A6)ks|`RVe{q9lB3nq*twBS9bdC_
zsm@O{!fbTnj<KNj_Oa>Ys#-WLHvqO{W4mp#6S%_l8gb)^+S~%czA&2mZk?&IF9~C5
zS>9%8O&9}qSs)$ce(XIiQi7a9EISk@7p3w7@_dO6*^@6o?d@ai_z?rz<GJ+0fKU+)
zdmSa}RrPhr(WcoB6B5ma&*_+y#LA!b*H&IHGD;=}og!gzukxKGFfcI2^Lfs!%PFz5
zf)xYF7{X`ND>T2F)#!kV(STy$Cq%)qMoZY}`odw*0L-_Sd2pH~Q2qK?2_xYdt}PWT
z8iJSCqxEd~hAZ~UsZy8lo*xB~vJp@%yUCc>x#`ir0N$EUmAYq1x*`7tBh-*KvIIR<
zqdvO)ry{pmNr8II@S1P*=g|X!345{bf>-l-?vK4C&|T(EIYClRr<9q{Y5G~?yIu3E
z<B?evU6Lks(20}xm-_Uvqyj966Y!j_3!?A<F#B7tI2l}y8lJ;i?+WAW^TS)GRSz7r
zdFhOnoVUXAg-*<J?freuGW}$;o3pbs>q)Xt?%<Ch8@2DzrQy?UJv`A6u<T)Pnq@(h
zDo+8IzKE(-USd)mri36qj%gu%QiH2Pn3UWtI29gkosXGikeo@F80S(Ht#euI&Bbn>
zA`(a%Lu*5aM5`QXP>2?b2tz08Z9Tu%{%3<>!>lMmkve3Cjg53LeNc3sRQK_xn%u`|
zGGP{$ZyYKpAAAhsWo(lQKpPr;amaPltaB}`*bhILPej7JXq;ZjEs{{4A9g6Vp&DDu
z+EkVkpGyaVJT-T0^e}=3{VeO3k+aTaYX@fHCKh3fJexErb_VYYQ^f=3tFBttW=QG!
zqZR#XJFE9(djw#hle%q>P!^?{N6;Dkm+>@C$BJ#w(n;#RI!11`$8nwkdufQHVy!p=
ziD0)Jd02}tVQ_6O(Hm18Q|n`65G|6Jw5(Q74XyR_29&0oU;Rfclbr5+&zSVKQ+HDH
zo;J%zAQg9rC;sKY520$#S=pj(rCM(5I3TkAYcizM&@viFXJm_uSrD|b>)4c<Qg>bH
z1W_t~G|GawANbTg3S@<rYapa2KZ5E?*xy%)nRBDCVU6Y3=+wusVv>_M$>n#Bk~mfJ
zSDt412lH%Del8y<egf-)N0@<!hxU?KGbrt@UQcci2`6%VI|4GKG+YK})<5c0|EqNg
zIlxpQ!l6E9KG7qS4d?>rJuuaS>}E$qVUu2IB4eRR5p3rLm(XUq?X<tyA4JFn&7^J|
ziF@{L2XRTnb%5tEYi`)!p?OAxbH4%>nNWMKcZtQKgsSgCOO0XW`qjyH6<BqE9|d^V
zAH!8((Q0n$X!BszF|$U7$?QHk_7ytDu|H#Oi}E}rawpMngt3@<Yfnc9sSdJs|N1NZ
zz{kO1)YG#nNIO3Bjk2aVWczrP%Yi}bz22Wcw(~vzzvNRO-ZJ925ATPVqQS^Vx^E-B
zlC7)B#Y#$9&`6N+>qSz)=2qL`O5tj+j?z`2;E&}}%MdkYnukjnk#XEl-d?jIzFq7I
zqCwYthT(w+Swm)3;IWOCBXCpOK&7?QFN5}7VU?`Ju1dh1+ZKNCED~xr$`5{Zt?Uu}
zj3a-IWsm?GB}zi5Ka3YxDZ0Zmk`eRlANy%qq8Ox>u*>Od0$c!WS~e&^IoG#E%uKkc
z`n(1v-Pa06>7H{2dokHC)Pw^9gIp*g@rd%e;oF`&R0+q(r7}oebT3q7P$Bl&raeFU
zFmr<@DkudaYyoedM0}1;vOMyd`<U9GlF0)$5z}|~2@t!)5s$>{k>k3-4)EX~+auRC
ztJop<g!mNo{Tt_-bC1ILCZaFYUw?6SG7?M1uPGv9F|YmPA*LLI0Rxp;3$xmrBVQ!f
zi7@_m!6C-{BErE5fp|ws`Ucs7Fyg*78G-Wi)B*(sNg7vi-{btR4&S>$c#%>Vks$4O
zHgR|RIXd-sLU4*303f0<dx}edhp5g%*-sj4EsS(&>@l4}4FRrLxL|jqXd=g$6k+sb
zT2xGV#?F?hsW{Yme3vIW-r<j-w<gH^%l$}smGV`EQolar;EkugSdQ_Zl@;=(cHj9H
zab}ur2$t_;tx(<t*a(jA0PO}xtdH}|g}==mt?fS=b{WyKWofiQ6=f?_$<fg|+N7QZ
zBfI&Q8hm;t!T!)#|1F2h4Qjk0yudl}cGE<7SYQVqnFWnUW#u8G+v<KS&2fQHbDVtH
zdBMKTgG9YoZb8oo^o9ux!=Hi2^Rate)<pWNvgcLTmQ_-8__KO-_NX71i;FlAUaj91
z;xLSazgz1P`2)>Y!B60fhGirSl_8^fQqgOm)oGF8P;%zR%gs%b+5NQs8XT3w_868H
zt{zJ0B3Q$Fh7HZP&Pk!TQ+qMoN!_?EZy=W8&&3Lgo@(ASiWfO?<?O8e*smq$!i#}m
zDr#pndKrpHZ)1BZZf@^u*~py59;V}w3V*%@EowOXNo!js6wf@5)+EG1-vkI~pCNN*
zuy4?Aitow$v3I9Q2AUpuCh;xhIZg2pkyQLJt!x<7EaL1@4<eyTf<p^Ly5VDBvh!9+
zyj;IPgOrXPMxAw@cjhXGvs(-Ydy)j<#zf?;plbes>L@5+*@@$bgJ&FQ{-PiME@|%<
z2NAWj{+y{d-_hh-MOf-T&mVEmC83DYi8~6T9cZJ(^r8MD7O?UB1}#2(%wRBeIRf_u
zAO8%hvg#L~LVGTQ+3lWh<1~5<)G(1CDQB{O8XEnMn9=^%e7WQ2OB<4@))mT<(2t_j
z4dKzMA}?5%3&wlM_Ac};@Tgs2C6yHKzj7cfRWQlzhzR!hr%cUpjUPj)AiX8}59+SK
z&hs^JAJJV`e)EOs!}RJK0|5@MqghP|COQkYQrA1#)!BEox20L$d<lZx+RsCM{(T)h
zSO#EafJNbbT*Cq{$Ri>E8@tl5O?HubdQ~}7^Z&H;m2pkD-`igU5JWnpr5UAkgLHSd
z3P^*1bSd2(lA}um=@Jkn2NF`!NcZT$2J*l6dmi4f&)5sL-Eq!!t~dZr_KzxS?1zl)
zJ{)JOD0X#_G}QJgy(_IQwJv?QRFz78q+t!si%3(0qGXrPL{5@K$dGw#Y4^O2hO&Sa
zMU(pN_O1iFA&&Bq(&XY6nGNQcc8hM=Grr~4LiaxYUej`IpX4VCU~mH#WA&qtpH@>>
z(v^Diyx#NE4=P49P%l3}919>&ComByixgvGqLu>mf1_8%e#XSpUjk`!U!dz-Kfn1Q
z&kAEDy7&0}s8zl|6E=EX{u(*H9SvHD-4IIPr0==&`MOTqkr)`Tn;uj;g^%b#Vdplk
zO>-D+3R$R^`8JOddHMC4iQ}UUWEk2{7*EwzPQV-;EpS*XNM?{JrZ4@WhttvjGNQsY
z#8%voE#yy{UuZEj@X~m$b#k^9ms!gWa!gA&8?3=Jv)#i9b;>>nOC~h;8CidPZMo_5
zlDDYiBCAH4?2%2UK={k4AFYF(_w0`hB3q@awgzGpM=ht}KNN>E+eC4vDkXdDY6$F}
zC$fa+7o(KPW4l2=<YsTCz$@b6>polVs?&b+RlGZL9lA7p6>~*4ArbNh2=Uazkc?7)
z1i6Djo_75M)0>*MHv&~{EXJ0?u8;;~turIVMLRpS*=;$df$SpAsfpl(z#lm-LhlhP
zisci^w2SX41)jZmM0VL_xFtu)&#pSnT-M#WCSs4Qg4xtG^dh(kMuoOC&oR2|MF;V>
zp)f1|KT=BrR;N$-XGE%Hp1KFt%LdG`;?%HwI_pcvTBtcmDV>$T9c=g)*RLk3$mi67
zJnT`Yf4tNUIzDLdKi_XYCq<!Jon}RLzqn?0iRVy%_*+mP!;X;1?RA~?VWVE&J#jn(
z&cflw08FXCA7TM40s3uO%w8~<<AN#FJgB^$ddZK-0cIN?H7ebw;%{im-}dxBr|hGl
zH@Val?2LCvWZ%&GU>RIQqgCzCr(Z*#!fGbWC!4Q<X2}1Z=FxAawPKmV$Ddy^vOdIr
zII2~EP`}wG`kE-!E%pSGPN~^x0lSCl<39!a_p1LNcI7*a7~VmYZ~`~hfi7JL44Jx=
zGqjg<pCV60MTyicxKcwdR`qb`+efulnyt5aBb;nsI=fWI+pc5oX<o|TKOi$nmA2nA
zo0-DFNw27d3FA9X*%~`Y%LaED7_xlbqsnBLpwPU7OvzY^KlzpEt6$FY(T8pG@Uy!f
zS^i1Z$w;!wlT+Joud)J(Bb|#FUVY81AG_pu7XV6ikUqx|V}c{Y!M+v^TQ~B!OSKl@
zgP9Dyu90`XiM&#wt<VH@s0N%ke%=6@{1c)`l%*;irY*}BHg*Gi)b=uS3LE%TSzB6$
z{dM2t7|ErbkzKK?ru&0{ZP)ak*IN!v?3G)MUj+=Ai|)dh;f7GsOP*tY-l_6uFEqU}
z8IFY$4fKUYBqZfOl1q`~&Zs;IZQp<<20kez8b!GxH8BJ*^DYRvzG<Wkr%?_A8DtBS
z5~h7zMVPspqoaJUxzTrzgeh9uun|GwwJ~HWB#|=sy?G?|wA~bokby<0*n4x(#H;~}
ztOIzG!pt+>)j%yD{90eUu*XeN&!-9v$dO6<N!{Fk4Fjuk2t5+Lwd^pim$K^*+xs8+
z3*O_kP#vuQ&&By0Qc1j=D07^>hLA9i-@a@&HnOoDW~sfXxc*t6?X>^FN}!hiqc4@Y
zpeF$yAV8SlfBhAbn`sIQn)o%T!*5PHQ`qv&O(5|z>ruiL><KBCA-4W=#A2KqTt8EX
z6v-<a30rvTW-Up+^bxEWjXIV%U1Jh5bzIBn@xXs#04e1~mh6eIAYk=V7Pr^Pz^zai
z4ZZs+Q;E*`DkWDa7Ms2KT<yU*CmpnIrU_U^+|L7Pg+bV&V~49}dwj#=CBKb!5uC5~
zl3JS`k}e=QS0H)8j%4XmT&7i8#7ZWGAYuBw_}E!ho0pkMr-W`T65mmW7p*CSxEf!i
z+7qKc)I6KN^jhcxP406jCh>P(=D%H;j%uC;hzVG$Liy$zFFA#hv-;okE&ZkXEq28j
z6AlOt(w-noLYb?!y|Qs<!TrZ0gSRJCc=eo398V&5j$|9E=dZ4|)r@+dW5kKLjF^s}
zH5Mc+24F<j5gIn9(KNHJ_ZVfyZZ<lba>5Dd`jcAZd9b-F-=YHSa@iDdan7<`^~VEV
zi>tuP+<sPxeCM+=^J@6FpAU)$_BZ63sBxuUOt}6KWuNpIN%YTP4N8F&6H(n4X46`o
zw89KS-f3^s|8w+5fc6t6J3y%MKdY$Np$)-GRo9LAm6hjR4VQa6a_?zcSi3D0sjDor
zeta>GQ5(PJ46el}RURciSp`wKbrvgKjPmfIO$A{HNjoR?7(i=%MsSrobyCPY>x|M)
zCYPUG4yUtB(2>RtKipK0`1#;BzV_2mDGc7MH*{{9y%i5)Hu`-CvT)LG|E7G#k{hDB
z70lCsLljOwJnsS#6RK8KX<il{jdvO~F*s)2nR(TM2CGRUR7GqJnhy=h&O`Y)_Kt9p
z6HTQG|Fie+!cX0dI&q>4*K@;DPl)Q^d-|}CSzJYAX2Z?0t6AnpS88Nd3he{cN;A3o
zSD-})=4BHi6^2>0IzbLUfw!D54{>U9t7GcFmXINb6*L9(-vp&U-a}MMArCy&)Yj78
z*D$4SqOhm_AO%8;Cof+$^chhE;)kMb*j~GUIJNZ`<&*Z(?{sQFnH^s1*FlRfvlW2$
z_d<)MX!fnHS0k8nTxCJ(V6r_|&=}|avHEwd*@20JLVbjoXLFR)n+lvwwvPKpNfzz6
zK)4!ad=%HBs1j{U#zydYMZ%@$->gz<X*1Dg;Anx=@Pw_T6i8nJj40xPrW9V_i1xjH
z1Vns%ISv>ADRr-vvel5L+?jyfNd`)iJ?rt5a;J_o8?K=5|E`2R%p!d{AixA$SXdY>
zHU{oz@;p3{B<FXtz1qyRXWq-QQ@x))A%nok`2Ra+pj}cF|6RG{)b&gt<crPH4{#Hr
z78SSU-WmP6GyV{D|L=v|BEHOTIud(~4>LKCT|5zevUOJO9d<)9dBoAd!TaBJ=HoQX
zQ<y@!5ecFXp{Q3SkDKRf7zz70AR>8b3Tb3+9hApF{n6OV4&o5dWHd?1qk@Mt2Oa?d
zW<>qM8M%+%bCrGz47l!m`%;)2?qAR51>iZ5%xAD9?w?U7k{pLm%BoE<pzzW5)NJ=Z
zLhn;xHi0g&(RDhojzSk`=+B~elJ`9_Q#v$C_Wu6vHlg<kR?UYf-j?vOs|&TZtKQ1%
zI_qbd)r%(BULHi1>Lj>5DxC<K-j@|DAsrOUf<a>}56vZ98%%eBA)=w7Vcw^Em6{)z
zzKN%si)+)wX&+nM4jrxva4noV85*8*V+lOm$4Nuj{gBi%j{A7(yg*ip21x+ufcB09
zl#|2WdHxd{+>Ln~1$FUt-{*_p%N%)qfE|5G=6%C}XseCer^BgT@IZd{%|$~#2Lf%x
ziHM1_E$9Bk-N<RmQ{uuzj*l6xodBLkW5HboAq6q=1U5>4ntr<jt>$Xy(F8qLz!Lpr
zvyHRFTRtxWh?9UQh`?eij0-MOdsHG!K|Y?{b>q||-J{<3929L<qIvyq>UNxV)?hjd
z&6`HSd0J8=krfcr%O1iQDY1+Ql1#G*@a>?T!M4?i<2eShx9wXA8Qk?v?f^OhkfH<>
zcd2Zn!Z}RP5f|XpAU?YffJOixEJf+nx22xQeE*>0gd6XNv5~Yn=0!B~ui={$(K+Nv
z5v9a%L2~_Nz?`n+vj>$(^_(y}&-8I7X8Q0V6n`aH6ljhR5O!~iCe%3vZj7}PRbl^p
z+VlEr%27X>%nYvT1iF96ks#C@gl_et!zutwNz-jYtjzQ5a2II8V=w&7Luv!M9~nwb
zLDkbBcdl;;8|od#_}&8TH#)17fuRez_gSaK+U&DE;lvQx<*llTuSu<I>L8l7`u3#_
z)num_fPz&j$9$BMIi9NiR9j8}C;tez9Qf6BE#^cc6l@Yi7e~wpg8MOg;lJu-8QaAO
zvJRD@2_x>Mr3G}P1ct^@1=cpDo^35J5tRWyR41(6QDimfIpe*fSgSaC<Y3_DIL*@#
zxSv8TP08A1*xf<HnCLpY%{u_#gz2fa0stj1os)*Y1o2UBk;b6<Wz~E#1yLp+lmH&v
z&$bLwscS8}xncSu_%W|5EZwQj<8<~nw^UrhY2R;E;U=Y3x~7mYjFMCjyhJV{DBa@p
z6PFtn%fIGgh8Bk{JG~H-zRPu-Y&^(%KD<gH#<e2^qJiMs)Nev0`CeA}H=1VLlWLYl
z`1m?>!>pmemTfJw`^E-{oPM8aK_}sdy5F1-pH54tC35)Iyexlc2A!t<t_7r|t6j}t
zN`HU<K8JeozM%O_LG2yDMg*Gy-sz+4eA!Y~S3TlP<8B%8>gs7wl$i@K9z|tO3BDyc
zlUjh)9mf|S;T?65?zo(<P=`s`4VZ5_<Ow|k?WIJ7AO0zDRMQkC<{h5a`pP4V+eAui
z>-CmrlSL_Yc8!BwcVrdkF;-F?HoMr{$a)%!#OGLwa<<_z#6yhB?F!5pcyenQ*X+ud
zBS|)8!v(y%qVHJC@Sv54i+Xo7o^`ZMLz$+K>i16~@8|2uEsiu-sv328^R_KDFp+<#
zSJumfv=#aV6t&}82b84z?eC&5(bKJGu&ht#-7rOb%Lj}yJ(S2^ng{QMK;htQnK=hy
zlik|w31$?VxYGkxc0Ty0YsjDvIH!TX!j*^ff<(AVNDc2UyMQ`+N5_CfIYmW9W*Wmy
z;QfXUuTyJPBixaCVA<%EC0`LAee)XIf2EUsII(BanuO@&TWnOT(6vOLB!hRY9+cyi
zV!m<GBC6mn&Ng13n)~x8D;H}A=kH_X6Ggd!KQCO?sh(Dr7)!BMTAST_JDM)pXuQ8Z
zE|F`?K8uKfVD>L<QpAf^e-fh>iDpi+5SfUk{8P@~>QwIc6ynt<!_um6Q9ElxC6T2-
zm*Q{4e`WF{DGAJUMu$Bm?{3a~Zmxj+@U3Xjc`;od80n%9g8w<@fzu%14z&U=fT~gK
zd0;+X`aK<FNP*tH;LTbFU2k@tz>-p6Q${DZbbzLO0|(Nqv?DNSiFQ`{o>SHVD@ng^
zqfeCJ2ub0<wn|FTji<A+!!6nFOp#?UNHCWZvQ!dJ=iQ#;x>0+;Dhpc?_O%IH4(FX)
zH!IaX=?kUL2gwd8ZZ%B|?XEUwi+B>6bYHjpE7iPyEXWWs8f9QEAQISGOY8fnTL$+p
z4jpm3fwUQIlSFVy-hCyX^{xoe#eIMsJtQ%7^>7-lSD+VH!x=Ds4K@ne*aR$T-M+sU
zdv{;}v$(jJ*0>e5ka_gzHO~tKV0hbQFRT_AV1Sc$7~9cl0XX?Q%V_S5545w?oXMa6
z^tLe?zJ&Ou8I_G9ccWUIrl9iUvRYU^>_vwgOTYcn$d<q|w^?^z#b(=)y@^gCt4h>!
zG~Ov*vnC3K)MSEEylc<>hgTYhnQ28qx5{w~+&`%S>+8VjTm;<p;2+2PR(k)L!$~fN
zXF@8_V7|c-sIm<L^6mQHKP0b>%aOOIr@TSPNvPLeFnEi=<&a!eG;7gC`tD5X?kq&_
z4~t=3yFx{6Js-2lehbMW_+D>-T>0qHebVP3boP;-7|}$bfU`dpwYa!AXN7Fcvtjbq
zuCAjEqVd~_87KDVfn}dl=~XYiIiEln`WO4t<b$qR_|3OF1M=9+wImg-kB%Ojd;H5k
z>4WIfa_e#@5^7(*hV%vD(lKI^U<JN0A$A^}keh^qJ(0JF&C?fJ5-^%otSsJ0z3cM0
zu$2)<p|3$Yncyofj>r-gdv`VdGTrm%Vz_p=Cz2TPV<6C=Eqa4+fun0jbBsVj>zzec
z00pQC*k@aL*J}2GvThFX1G|`JXznQ<28B>5&ku{7sExd?;XkdvwH8$arre$cjQXn1
z6?6V}o|p-bIy%w=xlG~}$-_DkoHw-Xo4{+W6d%4(6?IcE17si~tRG$|s-0lYykW}P
zq;Mo<IAUQTO<0cXNo^PSZjwUk_M%DOWdAHAd!DJK$<zNT+Q1J0x>+w|?N@c!6S|bR
zzi(h@$9h@0r_^)fR6fkQhEP68$ZVNQFM@5J$vr)Xde%a<8BTHg)K*c}&p14EL1;x|
zf0}8b1cMm+y`gNM0-?Fdol~QpD6#NrVDsvVCXJ+ON=Ac)fX)<bg$3TOZF6>2z+<rd
zGQ3%!B@>ERRm3V=$pVE*yB2NS0#C+B?=g_oCXNYTpQaxH<sfH<W)Kp5P|6>d!BuRh
z%&P4eS>Bu(yl=V^15(|k?ykpT$qb5SL9Dc>X#tpM5rNMHVeubx(Z#bhx6|qS>=gZO
zfHaAsedOBr0OIwwT|v%v|C>c?Nh{_UxwhaAuo(cmhJ-vk+OwV3ZM*z4xHX)Ywt2V5
zt!t$ZDII*t8ET^&-<$LBLL5#$puDZ)_*of>tRr~ibES1GhPM(Cw@9QY0A*txzcla%
zxY&$cnireB@R2Z}Pm~lLXMINBjsJ%1zYRVReGJMIfTb{8*b4Y$-R0ZsS>j}H^xdP+
zwZ;*hhza|9eEii~V-F5~<=#{`=K;|r2>9w@@`l!T?MhmvY+_0)LX3Ui(+Feg=~42X
z$YYHIs|=lGG_p&za-|Re_g~qsZgm1?WwTwEO=x(0=-t#c(R(Z~Hc+|q)EGk36j)2H
z6eD;e2(~)dBTceW?sby=uxYh-X=!j&)_Jb%u+W@a{Q0GOuk(lNwKWE?I)E=>2!Bab
z*47xGS{q5th>ipYg>kKGgGMyBW5Cc>1mmYT@&wujkwbnJHzljn#s6mg&SR(SkE%AA
z(1%zDLFKjlT-8Id;nUo#e9;>0b|6sE(uVH79csThu$2PV!xk*%Bv|Qhd-J}(pwhq5
zX`2hvrs!fC2=eI)<DdnCDcB;_H0%CQbS;<+&W0y@rqjRgr^cTyl~k#GA=R_=Lv79m
zr@8F?N_uZ_KX<O?Ug~_!nW|4BK`ycrI88KxNw{w~i%kUYkW_b~cyX*qFaC2~o2l{O
zGX4R!gyoTm+Q37{34ZUcZjwk&(H5MJH3DkJR3lP2&7QTF6ui`N9Wj|bIcqH<Cv2VH
zq2`psDpZG$`>131B}Va3Uzbkd4sG+qwGk#S;Uh^+o3Sq^GIk@`aZyCaBqsuiI_+}S
z?8B1{n_YEu|8$s}h4v?(&OgjMFO;^>6)WY>F7wZSyr~isC00f1>X^N?yl}vrNol(|
zJ?MO{-(KC_<TXnmwoay}m*xRM0iK6t@(Ucd<0*Q1pTL%4lwG|OTF28zZrM8)blIR`
z{U_7O*tU#GIIA;|@H*?6WE~heeN|i2MtLe#>(uIVk3rJ!PoMq!K|f9Z(f#3fNgMGo
zf=~1CjTrj%;<GF^=S~Z6i!?Rz(G?^3aC`ff_}KU>^?S4P`{MI*Jh+7EpE9}rd0<Pe
z;ay+bZ|pr6YN@1N@sbW3k~}l#7qBAX`Bs>`dD=Y~A6>D<f|%GbeEF8G0H5#$m!atw
z?Bt7Y91m?N35Sx7t7rFhW<d)K6kcQ)1%QJR8mcn}YWYTIDp0S<PXaIXOmK1F<HIrf
zDqc)Wal55dR8#Ye-@ke`Bv3+NyYV9@nIy&_i{#a$(eaHQyI!Fha-&OSnqS@6Ii*Nz
zTyHpbJ_gz(p-Za)*L`hmGn*zPBoBMXj!Dm?TwHXrP4+0EDGgI|uDlPp4(B}ypG&Ez
zwGeIBTA0I!m2}9sl=$3<eW?=W6RM*ox&jd7)!bOn%<ESMn18v>Uc9$dI0B633}yAt
z4+0l~9W6ccPd^d*^W0Z#EeW<EZ1>!S245Eko%zv}3Mbp*Nzi<OMD(=GKwOIT2{07X
z8H;3t=9>>o@>Sb-1lqmW8XteXB64e)S}ErjbTeKK|3>>IL93r+^38D`7V;goS$$6c
z9^_2|e&%bD3=Mvh1Y8GRnPo94dmn~&Y^WIoX4kkzaQ)0SG~ctzAVSl{vt(?Wl#ll_
z^XDoh;j<S$5g(rwJ-!1PXS92@X{{}1q|(O>W=3^AdRpW6?al%fBXI_bfP#Z%(BPK*
z?5yCev4%Nt=*G}i8e%Ko+}b+&z-T-dwL|S+Fa!qZLvkdE)HdyJhAO-Qadme6YXmLS
zJ*ea|-IMAHIConB<7gCOzgCAxEbpCAFI!;nPOcb5ZOzgaIeGp%bGtFd9B}0K6A^4U
zs`=0CS%rM3n#o$B4U44eO4jVTajMe5;%*+v91IwQHL&errN<RNE5b_T5*`X<&|HLx
z{(iS9O~MnQiCmiaD0r1@Ez&m61g26Ae<H4k32FL1qn<vTIvWw&H5H+nY$<e<MOgQ{
z09eSzHlceZll~36mBsoe<!E;-uF#>428-qmg>b8!i{cqW$?Ze$2Yd>Ih?=SZu$Am(
z|GX4@S(`Dbvp+L4)2Vz)@R|^?MAG9;n%rA^VT=bpA-cVJbgJ~V5bP)KdT)iwQmIMb
zedTR&i7NSS=GIzt(V8sD(c=hU{$D=x)mtGyzw`VRWDUQ+ZVaWlS3hQEt{bpXc+P$L
zlQ+N}%-8|_BrpA8Q{FIa)Uj;k@mA_UyCz4n=$!$=|9gH-cw*<UHd*Ajk=sE9u3kik
zObzg$LT5L!<12Q6l6`C+C~7y4P5X%c99Q=cu4fGBK?js_nbH;|Yxn;jK|&Q*_CZl<
ze}Jm+m!cYub$yw1X)ld~`50}wy1Th8_VhgC;Je!5qxVu)pp5d?Isz5q)%>!>7eRZW
zFST2brguaz$ghKb({G9>VDi3UFr@Q6_~1U(ykw{>YC?Q^`ZLE`kN@PocqmJuwZbv9
zCTG0d(xXwPxM!#nR2UW&c#Fu1#2^yDXv=%)7@?Pk<<Nn9%a)J0{nK(T%TO}AtCpRF
z`)N8U61D46cVUU*Q?e*K(w}dual8-ge%!ATSniDtXszKF7T5m=$jrv!1i|Mr)Xj<;
z+Q;-T7QEVN$=KGLa++Uih#4l<Ziq0qoVo|1Srh~L-W(QHIo+<YJia$Qe_O+^U|*r0
zu)Mf*imi!=xwVZ8d4B4WT~<IQ0Uy4op5>BQqFE+)vup-@o11^|yPzB3y6kqd!72rT
z-~+r0VL>c&zTr|tFve2RI;(#TTI3^ve7WI`6CB=9#I9!wB^!>w@cEqpNW-gZXwVg0
z^6j{}1_jCru<K`JSW}cjYMXSW_&l1O7Gh70+L3>`7Hh!b)BEO0&#p*!#%8Fn)8+}5
zj9Gn=Y()Z&(fzU#rrXg{1!G_D6Fq#4%()UZ7V6mV4gr4TFzf8uod|+T0vlmoQU*fT
zR^n{#kfYMN6zvqT$;c7??q%n10G(O6*%DgElDd*cw?z6$YdQ0OU23UPo1N*7s-`k<
z%|G;QRNwQPDIQs-SJ*Iy60?3}Vb;dWN}2HMOh8i>5RDwDKHX~1cd?prU}K|sbD*_c
zROLWizU6+o(!1HDcHQne;@_)C&#6<ERB;Y*0zjlqy=tpZJ_o5mziSb}E%^v}H*dIL
z1Z~y3b0<1CcWs`iohqN>%zoX$Ue=DS`>foAP%|3*IKLIc|9NPXrFr=HJVl3h0TpMN
z3TuwxTrR~vLxPG30;iMb=0K8^YQRRFm!89tSQ^*%>%eerQ}ha|+S!Y%BhrR#XOK@R
zBJE=^9b+#NDxum(Pel1H{HLuMq4p>9PrRk+7s<pQ5qw>J%&T?xMSFA9?JJfgOXb#k
zhIWT6)*`7a3~33M^f>~_vMXt(SJ$iHph4_vxybh@bxnASx!My+Y8zsL0ut_sXnI*?
z>2~O*??I@FHSu2$j8APCPO|ZTkpZ5b`T<uaA_D)W!WuMRIUZUE_v&hrl3(kz&E{a2
zwERrBE)y1K8)zS3_$ldE@%vB7fzt@%X)AO9m#Kw<=oc>QH8a1qY@HL`0&mvhVbD>|
zXYN|bpukf#$_va4%0{pKKDgYY8;hKWvYrCJoWEXp2W@++%=;W(&wFPaUMNB2HyoEN
zFtJ=l9uBfta}k!;N-<fYwpaacdXhCSvu2SsXWrKr`K`Su>|MU_X6$K#y{CQ^MIa{E
z1?Nr>!Ur-_?|9pWwOU)ED~ci{BaobwW(l6%UV}Pkcc2p8rf9dRPX1|)3<Z<ilT_%9
z^}8WEyOgOt#L5b(hOM&>4RF0bIT@hKOodS-4$d1l{1*3xCF}G$_ZQ*8Bmlg+xs;Sr
z$n0U&vx2-5^D3=LL|Rv}Smq>bK7EiT{ZAcs?wG?ao2fghsd-M!lkh2BakezkFPPM6
zFI1xmGpd!Te8n!X-_8LK$C6#bpds-~XomK1jT0R8EPWug;JiksTk&x1u2AxHerD$I
z-tT~)*IEnp?n^o#Ay+U!yJt2-tf|gXE!&r_4cC`K?%i@ELYcyE@AEpy(gMooh2cAI
z&u#~{&>m+UvdO3JL!x8&0u@_#c*&Hd%n+8TW7^v-nv?cNW8ee^C5t=5?4RbTpmli2
z@}ka{(qFigGj>Wo=!Fb69O(0md0mv}Ep><S&QAu9CjX&>-O}W$c~p{I%I4RuG4?Wp
z$Kt|BHdwUsZ_gcN&$~#&2W>o4{{{TJ-tL@{eADU*VotS~ee;*LBfi^n!Ug9ySf0!n
zR(4=!CFzJ=LRh>_eZF`N4ie}cT2%yD*3NGJS{k7pDQHJz%~I=}S*VYqD;HoW{yVN8
zAji%N8H@+txY;L7sLreVrTw)U7q@Q*!wl!*Hz;%m9A8*$Ue_NRJ<!`XyfpQy-J_&L
z)j3BGCiZQ>v&4b{ShNE2D^*@~*9H(7n%Z&lU!&?@d2;DG3G@}YD?+xLw`*Jj)~Pld
z6gTF^aZg{g?p#x3S&rCVi`CkWye_}UA;$7Q*kU(<0j3P_$blaOp7=G`HcVN#GWu@Z
zR%>&0t=?t>3c|WUZ{9)$iGL~77N(`}1GRDJMLYOw*k|*QPS)y*M-EiG%yf#rPpy{I
zwk|~9)HKab@US9TZr%F++W(^r&F1m$qwn%?L3&sQoX7*h$^T-pF7{`qo8v_{7g^5;
ztjm+?QuB?o?SHD(X+_z))gz5C)5Fb5)y-=zke5qE`TB<Y9eZQ*gDXk`L3o9%K{du^
zozd~0mKOiZq6hxfM5uS0+Miu3@<0@TSVU4IHr&zERYi|z(e`Y=0&0*W{+FehzUaV%
zPUnj=|A^?pAI*&I+`%|#zLc04m3S=U8*5rDQiWzU5hkXr=5tX8nju(qVdDJvuS)`7
z#vH@amf_xn?YMQ<4>y|V>8X(B99(LtGkktF^x>loS_6-Ky~fI5*R#st`_%mrnbv~H
zT7Tj7`|RJyhbil^{n(wY-l;Vri`i{xFm3mlST}L-9sKUrpTjx9r=hDW5G4qzYAn=z
zY{%0nrgbEX6x^e4{Uh_OfrrE(2|P;i*ioiu#la=U)U<OV2)F&T-C>rG*7?n^h@-Kz
zayWTW?ZLfq9#89wHlhH!M}O%{yms;=auRYj`263dp1eWI$Qf$z8N*SW<QZ1OBdgW;
zXlm%qDwfpF!2;~aCNPT{_p$*#ko}w{8R~EpPvtQ3$B|kK>G39`#)hEU{>Xww!dm`f
z0U&P<q|3}$(jmL!lUxe%je*@65xLXkZr%<8x}}gLfqUosfrkqurMMGG9G{ZsM-U4M
z@)CgxT6b7n$K7kJLU+;ihGt&_1vYyEIo%ts4dfvy&CbIaKAng^LY96@de(EUvbQvg
z22kuIM=gU-6J-XR{ubRfREc&xxfFr8?(UBdo%ciA3AF_mr&;n$@0rw>*~60G&*jsy
z<3@G~F!XwQ>$0|?ZUJe?8|H(Tup*|1Ah?J}l*v8*^&XB}cUh==cQh!}0p%@bBzN4f
znRup5_-<L*S4ncwXg@a#J!=vbt?Dx9*;a&7ZSHUsYnSKRsuba$c7HA!WTVo%t!?9D
zT3KHcBf0Lb{UShrnu$L>4<BoGM(7h7HwF{$TYk=rS_gErE`SuGaqjs#Zb~amA;;mM
zZ{1@y)$!4cMy`&%SEMg$BSWB&8<UuqZ|w(+EQK_d%S{q=1j{9XXABs3B><S*9W;HJ
zjCkMjVnXdO<eXo<fNrA9COWxM@Dkx~G6paB3+7g*t3{xd>4!0euARme)=HoEIy%ga
z{RbZy%J`MJTBT${vDaB}dMY(rJ^yI*I%}N+0x36$QU}=NW`!7Y)NXn_XKG+taNZ$a
zy8QbVHb>L4gN83xImZfPHtR3DY;%t5;WzLAt*mbuL&jd6uC)JC=;PKq2Og5ZZ-x3v
z5nvV~LS^jt!BydlcJ10Yvio5;EHu7~xoD@`dl0^fh_feLI#X>x2}wV5NY%WrNPLs)
z%!bw^NK}^UcIY1!OZFM>M@pG?BszZs)>~-3JP_>xrUDDVMa720vwn=1Tq-NK9z5c9
zwEU76_B(6!t(Ez@qnERiZ@e^_v52g~*SF6c#%y4AgBO3ds{_p>CLP7o^+!_o^}REg
z5BQ(14_a{k9NGv`;8L)WecKjtoY1kV^~w==yEFZQl?!Zq<PD;$b*v5F+Y)nYz|OpY
zIC{iK9R0=Kh|`@4n@>eBy_k!(&SD@__HC!O#wp5pxOuqV(A6nk3>GdeMQ2-{obeY3
zzG9wLx&lb0rSTiBWQlA-4WukmcQUC)8gbmyy4{|uYEpMEvgrHkN_^IXKtv*6R`48o
zPxG>e|2L&g{$ESj=uX|c85B?0Ctz8I4r~H79cU#qa9@EL9MEDY7IwxCOIICA&4E9F
zI|CTrpyfyGoL;muNp$vczt@H?bbhdBatOO{_NQrn&TIpzquw>n@!Nq)srA>~I6oKo
zn}dQ!<1M*<y-IEVDt>DId;rW<sGH{aTXG1<W90!&yrk#0G5<19`5FuPl5R(e?>y)-
zCb$6q{Nq19c{-8-^s{zW$O2)>aqy68607gzBX~IfNs=uAyhw7q|6JNF0R4Kwvb*P3
XA~Y0y552Yj13$|08gex<uR{J0<9<lK

literal 0
HcmV?d00001

diff --git a/_images/42e0a303da85179624b2247d2d06868f157496f5fefdd2f4e8ad741179cb6e69.png b/_images/42e0a303da85179624b2247d2d06868f157496f5fefdd2f4e8ad741179cb6e69.png
new file mode 100644
index 0000000000000000000000000000000000000000..d060eb486a93afbe45c19ebb3736f9d27a9b7f80
GIT binary patch
literal 70587
zcmc$GWmHsA`!6MegkT_zv`P+*q@+lxC|xpyN`rJNNJxitD5xmiodQFHNQXl+NDVO5
zkar*7SO0gd`{{nU_pCK*fEmu&XYXh4^Zer3p$|2b$u7}e!ok5IQ@OA37zc--3kL`H
zh=>sUg&~NJAG{Gb%c(pe0*5z|c?kHO*x|mOGY$^P73hWg^Ox@l_@me>Mcr3g_Li^Q
zo;g|I*gkvZU}OKv=7rf!R|_ZS7xs37yaK!e+&7=UdgXAJkMDmz=e2jT;!9u(55>W`
ziKC)$?}>Zb`Y*S7hNlhZ=VSVUWe3v2pRPvd#ouT1r^b!DdGpq|KSj}7k~{AXDM*Y9
zeHGiCN?x;EGP}0NeCy_iX2o~!FP`z<Ql|*@zh;j^#>2rTF7ASsDjPjIg)ODU$vwK+
z=^^8hcCkyk%t{-ZFOa!zCP#!z`QIakxYuqIFXZoGWA`97^uND!2IjCO|NTJC{r}>F
zPMNvXnG=qS^bXC-yojOApQ?NoIq9yG)9pAJ$s)cT63CVu>Ai29l6D;oZi_oTe17^x
zJS;3M1M1t$y#C;loz4@S(%ZzoM?;x*TZdy!XD-M?gs8Obm|l$&yUEed#KT``-;2$S
z$D2JlGN&)TXrv#N8_6sAU{Yt;B9*JYqdmv!#+`b2usexOlAG0Qg>{igB+{zR9bBZ8
z=9UTvc;6v#o$+zgkU8CTjFL(Zy^5CVpm!!kQ2TBLFj2_OI%gi!Go8#{tl9cb0z*jJ
zj>S$>8Q5%picFREITAFy0{YLR@W&L7qJn}#hHNaGTBfwmv}S|x%xv{VChur*=jE*q
znKQ@fg{o;M5<ZVERrkZWU{TLAI=kxEGUwO<@4Y;erO$jmx9c`4>2lJXE>-<<r4ya;
z#&GmJZ6w=X%_~}VDe_`~K1I8CNF1)2oZsE7wk~Q&5N-IazFv>%=hiLuzp5!2GdDN4
zdVur-ck+sgTerE#cwZn=VpHts_*k?Wxi@Aq@GO`CQCPRxjr2nIamrv-MhlJVx^{ea
zvzeB@W=BrWZf8eHr<@<0pUud48~bOT?dH_ak=dJ~(%mdGefH<dhcsAwN*z1W{LZi=
zFcK>6*>QMOfv@QPMndLc8)x-4f)4vz)2|y`v=e%v{m#8iTdNGn$;h_b`Fvv<4v?vd
zjk}-eJy$*&x{N$sC_GL_dk5<@9gRS@(VHogc~Fm%5=!msD2M`sZ{z!)P?J=zoUo+Z
zpS0%VZ9JTJj52vOsJLD+W^B{2H%jM&WI8`oHr{%3qxhHW)YGWG45?I}s_9q*`-VNo
zqMDZ@cX;c!<TRxAp7@?DFbN6?wGnfsJv=?j@Y<X^Mf=Irp&KO<O?<^o&UO-OmQ$<`
z(dTFA)g(;6L=@B6RxrKa$+n^A@6WyI5{c}p3E_495u(V>i_%-TP#`+BU)R|xu=fcI
zX7mh!M@c16d{^qJbl*0!tKU{SJ3DC7i7GIxiG_w*{<73AlQdd*%&-+SpgBu1ta_)L
z3){vi-E!fEA%9`7RhlDVPKM{5U?yKN9b*snne(#)Q7nyh=FxD@xX-~akJB}i0hckO
z4(J8Owv;TAgxR4N*5QC|IB=SLk+WTvc^ZCx7<E3&%7i5nMY^-%=oq<NpE*Vz91iD5
zw?EY}YQOBeg})DmKf+iC3Ri<&S5*+dF`N#VKG7)El-?v*fXVsM7@gF<t-Tjr2a}&9
zborId%`Z(&nR$h^@3b_fT;fnB6)#7WW=2JACxx^N^!KVQtgEJ=IJW4Hzb$)qry$CE
zk+oc(XLawcRk{n6#74Q+d$B3s;|=U?j!a*$^e!_Lmz79Q>;AiI+>w$y_o?`dNt?Z9
zj*T~~7sv}6k5*Uj&YbO6c|a?K1adoawWP><iEF)T#&_CzK&pO`op0iU4ap5&^U3lT
zh8|1r>3wzwaiAa6Qbi)(lH53JMV%rK8arNJWb6d@FO5_A%hdohwSBB{Us!tU_2p;_
z8mNP@1Lr52NN^>cfm8{%O{J;goSa}bmFGgVQOEn;srHS@pH&l`TFFFz3=QEgo;JMt
z{kUHOEeTC8)+ovEyT6AfywD9Qv4I7ZlP|MfC_rDL7}>i(-zsD@LsLzsap(Oi`q4Gp
zo$=Fynas^iyTa=EOXM;q4$!-&r{}nR(`iObX<rg7`$6=3zst<Uo_jCB$+!}D3=k(+
zOox=!fYbq7U~{QOPm<7!D}Lw4GotQG!}T5*1bY|MT?xEK-B?s+(><EAqp>DCAxiUt
zrD%EbIk21-#hOF04b~z4(Mi9<`kn-y1u}d0RzjDT75b8gvjL=JWP>j&#=&YL;=U=f
z0EXqJy4L=LRi<N9&WmmmlS4TiV-OF{n>DD?8Q<e=a6hqF3<{GnRv)yvxtUYJ4K1L%
zTCr;2ju(lAXd~_r?<mZRE%({`3f?gBhC$H`z1w>Zr6PRd=8wK;X7-bbP8YW<<mnXo
ze!m>48Ec*$<%Q0C;*rUXmO}fWz4XeT4-z(|ir7waWV)?<4Bi3rG!x8II)BjDiA8>=
zYv;Qjq8HIp;4pORVy8PPTn|nbwA#rtIenLER~rhw1P3JJK(M|K<7-UPE3=fEQltWd
zlOvn?pEX~|tH%KOUFu-E_3aJgPOvh8Uw0BZq<oR}9+JC1<QJqJ{3c2*B#&0I>ra;W
z&U3m;TgfFHdU&eBz1Ejg?Ge<i>l0<aH#aSpdy}VKr)(D@_0D@Px*|QSpn(HD_npkQ
zwzgKp<8-3TO3-c7khkf?iMMWDV7WiNhg@QPOEcX~%jKmItxZ{<h)vV!{_vVhVbj?W
z=ZI1?XHA-OlvgkKww#}zOo6s_K&f5Di1vtMg?@z3{zULXluQO!QGKjsnfT^Y4`>9Y
zKmB<^cMjRC6;xS*1z}m$<md`mqD@n6D?O$cZtxm*PS-eN$2+7E8&xy?U~T7&lsR(;
z4Hw-9Q_3%w8|^XR1wA0eet(LqMmHWGZrJ<v+-N#fB$qy%y;I^m{7Bgmgn`q9MdCJi
z5Wg=QkghwrdGs?_=H!_#c-}hfba6T#{BkSv#ya^PPt7=eC+3u#6BT6ytNqEkmCOQ4
z=G<ei#_GJ+W9Fdo;3-q|c{BdytuT|5H(+7i+H`A5)h|%e(Yb;4J*d+wH77=8`bl4w
z+`>7ScD_7wy6GmWhqzHcd*K8-;Ae^5S<2Td3ldTn6|+1$!Ezf`M?=eRUQuHT^w5<+
z*T+FCco=Ujkv|jx;D->e&6;2kj!9gF)o(t2{OA~Z4|;Ul{rvBv8`s`v&7cN6Wjxjl
zW{v_o^sTY|ZjG%_4?rtaIBXr<vG=lIW(ze&^F6#jO+3wGIoW%GZf5ok3Gauxx>T^X
zRCt4jOi`o<N@AufMc7&lyBPPPLjTWyf^1?jUQ6R!WU`E(G%C%RTs47vZmab&XJ9k9
z7561B6Ku6T`vtL;4*|K(4cig+^api*i}lBwb$p!ZudYDLxY$CtV3MSV05qwXj(!B^
z!l8Bai(2yc8dOG47iXrLy*(e8D90Q<?qImK6Pr%nDuz+PBqe*HFK`)ezjN%MhKpf(
z1)&axPRpfJX#Vq)N^i1Y&QWiLN*sqHnEjZMBH#C}b$udJHm%U8f#q(i+LjFjT52Y7
z>XDr-lrv(3(?K5K4LWN+#k%mLcAo9#Jd`AMUj4!Kp9N4!US2+5x8zM(|J|_G%QA+<
z0?@O@kMc<E&ks3+HyXDtQ1R%uf$8wD;ezI$RrB~ik0j^{l*wDWu05y^mStV(ZpbdM
z*oJ03cfO$veGeRqCK=9=PxG`(z(@~07ZN#ah&3O$=Z5s)RHk=qTjmXdS_6(wOh2?d
z4%#<p$F7S%Kid6`(`&yWdr_F%6vjqvMw-z~k<szItiNiTpjI8PS@t>Rp*2}l#BQ9Y
zvRgrmYr5bBETh+)MQMv06Oiw$UYkYhY{RD|uaG0UtN0q1o;aba<>!KF=Xa;nooFV@
z0=l`uQGEgQN`XtMi>*MMZ8{DqHXe6ozT}mE<-j=0Z!@Q9m0xYcC1TB#L<Ahkv@{tX
zqX$t+&#~dI_rE0*Sd&sFj4PP8JPRvoC`zTaCUv<w#7~*trpAH5TXT?{B|Kc~e4tH!
z0Q-h)kN3Q6aBZw8`DDbMCga{%)=WgShcfGf`)c7N8DH!LV=Hh&@;_T*KE`-<)b8{5
zOR&II8t|JhXcFPVNvTxb8Z=r*w+MpDs90an1u#=Mv`{98^ncQD`)T@kT1DqcPRA&w
zHYv%@>~68{P4*ug(Ui~U^nNS5vl;bFj#5~s1`IC*R`w)Kt~}F*3#3;fitRDaoq6(7
z@pp&su7tc)%i9wW-~9B<?8rM#cmgMi{9v$O0Ea`gHc1=(nRB%h-Vw1?)$(3wp0>`_
zzEz|$jyAzp;7IO)4XM^Q*qe(Vf<x{m;L_*Ghc%UJZ_A!vT74wvRuDk5r=s@m?n=4v
z&spkSp2z5w2)f-QTC%Hyh_Xpx8dshxbTHx7Pj)PZTlw2Oqm}`ZlI1C#O1q<XS3>Wu
zyrUhqRe0k4DXTS(P_rk8jpK_QtCIHv6A`tc3gSidEp*80t8v<fM*%^hDryRSx}*;q
zQT5m93FU)2Ot)Gvx*~1S_K!qiMvv0D-Ob=C@2VdJw5HT^lU*IMFq<S`>>>Q^#+q+S
z_Oo|VjOo750~`1k*Ij=0-<ElkRIgG(gxs~w1M)R4KCg)vGFRYW5lASkZi%l>O(`KT
z{ZZkb|7#$`dWdptC1jP2W4=UNCfzZ8gLcRxX2Vi|lIC^r3rsPc`AxcKKZJ=1=AG0C
zIvhr`-hcjtj8TBaln7-7dd5>sg$!E>8R<*eMOfUj6_}qStw>EW3&)E`Bcp5;%#=+o
zOuf1tNB8T*phL(~$!J(jK|+kNj}Yrs7ZsD=XIdA^HTl!*TRTkfE(fftguU#y>6GgS
zE_d$NE*#5)Z4SXS^XK|`)M}|hKB=w=bHwJnQG-q=E@O8OtNMSdMbsskTbIAF68ls=
z8ye2u@{^y^C9X=Z$-ZoqEY3sOhSjo7N=-x<@AX-9>1YK1`c%G6LH5#POIJ0ssYxM@
zRN6u@%O6!V`?g2Fh$>iIxf3)}DQ;|Ek^696AcZ-bkIwR`GsCx@6}8p+ug-*V*&4O#
zy$+q!JGOYH5zHj7zu1?RHblN}UE664#q50M&E62ed)*nn|C*#QrR)KrR)rS!Kz8<<
z_oV>X7f)k1*ukqd;VX341!>z9xf9`|wDLAA-bAD|9IcJ*AhyAL&_r@J>}k|45SCO7
ztHln~5Cj>8#pFkjWDHAu5%v6CENpfawrCJ;&T}*=uM5WrP|Gh%^_lsizq$ON9NS|4
z;dI+J^BMneR#HmDH)Z06_WJ8IVI<hvl3Lv<ectvP8w&w^tMC%UY(vfleZ?)a2id{4
zbVJD&5)*GiLIz!n;KjH>YNi)>MvtiSO7Ce`UA*rTB%Dy$^SrKnl@BEq8u7fXqJGm(
zXw;}vM<<|#THc)dR|TO<>MSwo7j>ndR579Bg>D?j!&R13+-9*D?jk)#ex6PBu^(uq
z5X7$4ZA?bxK<*J8X~eB(j|<{xRQMXcC>4Dh)Z24!wAv;}As7v$%JY%75|U3HmY9@=
zw^dNvR5-Le3Nx#<W)R9E2>7@mXt<@THgZ|dq1-Ttsq8G_935i4%f_Il?)h$fhxD3t
z4+zhVvMk<6?F76hV`+wKS)8jMrbf5hr7zqhc`alv<kKxH@;$^F#KtN^zeL*QLo+EM
z>U>zBtYgr__hp7m-6fA1(c6c!DmH3Y+Sx=*j_U^oCT^{kD^9;hYssay<R=e|aAZ*}
z9Vf=*OW7Gmg<pv4P}9zTQ4&$C=p9Ird)8aeS^|sdI}l*hTn;<D^qS&Bnm~7LM7m%%
z<3af7h-pavw^c&~p6se2OmyO<O+tx5r3t;BrHjTh&!4KT3SxM??m<4OxMQDwPb#XR
zFp7FX{9n-B)%R@_lpnLgu#?(RSMFxtB}+AhnYWbnlcCl=sM2T{;IN70E`LZ-q29J7
z9HJZ*3C39|9Qz$9u@Xs34Zkn1L?sy-djB1^acVU8Y?GgoIDs}H!%Re18Da7Z=iu^r
zx=Y_PT9xl%wU+!`n-;a-$|B>HOK~?D+zI4vhR%Dho6|JxMZV0O+_hEu=_)KDF;B~&
z(#Y5BfBE&Wl}FQ2c&`VGb%>2Zwl2!36BU!9zVeQ4SAmR_@P_0@^fTD7t(@U{YOERh
zZr||l+m`#xlCAFgV)`vHl$~#KBH%wNo*!L{;^fFRb73oAv5(?^vM%2(C2MYT-77-O
zMC+vpM+r>ok#tpskCKT<TgvFo*UeNFyp|)JL;SFUS>5!#p)p2FHfwmqN5R6l@sr8d
zvQrk=v?uEW-7o`$t$jm4Vf8)bc)#Bd#O>d2r$$A2^YnYe>>|Ex8l=3?@9PVV_+8-k
z^ZMBDJCh<L<iV_52#Li5s${qJv)uPl*4Dr8zf@?i*WMh<?>a87&p2STwiT8c%S!71
zs$sbPEj!Do&er$*kzj(hL%s_uX*Gw>Z;I26XOr#p26x5+I>U{*xHjK*yEIAI6)XOH
z8u6BL4B??kORwj6ZXW)uSnBZxD%&ib&T%7m!+Q%`FT8w0Hx(7{{Ciy7Wj(D9wxq9*
zNE9Yy+NoLE&Nwi9GCD=8YJ;idS$o>Q^r&Pr-YchCSFrn#5B`-tY$@saarm%AN_%yv
zjRH?Ms=~=0Eq{~5gNd^9fCUMQ7x|`0thM?p|K+}5!Jxd`@I8r7;AgF<E1t20c?S46
zq?P(l(_cFFe$waEvX)YdygSkDDi9X<BKetWVvL#YRmEwV9B$k5^!EiW<wl6!NR?d$
z23=BCX$#xC_lGJvHm-L|8JMx=Gn|U~+S3nNu{}jv_&kV}y7=C{Wc`B@rwcd5TS=9K
zYpD^Yl6yERMDBRPJ1UaIxb^Mg4vh7J3d~;uXn!#ys5P(XD=@PX6Q>XVCZ;rEZrXN2
zJR(WRFipO8^U{yk?8LVp-6uv;)VH{+Ub(XRQ3+=fH*2=pV2g^5B{upY>ldYxuG=E~
zWG_Ft|1=Gedq;-5AXRzaTS3>1MXpm+2L}eDOrV9in9;VqiZOjHLgN@hjEuIxWeV6%
zxYgFy#-^A=C;7;M+P>vtcxbUm6FKsVx5riZx5lCG7rh4)uwMlI#4sPn1MivRA!+b}
zaIo9F*SZpAm+{B_t7PM=)UVCw`hZiX1orc%u!X7LY=^($B&I<qKRq?W*P|kmMo)>g
zDYQh5%Dge{t(Fc6bB?#bCz6)oAon;Pr<kKBu+AO-dO?*=l}Gl?=hk83_O|gCX0$D_
z$>I7$)b^O20=G^5`!{v7W*uH<sR^$tP)j;?7BjQni7dE^3jI1;EX@~3yy$3wapquV
zY?wn_eZxw^j7C!o@pi|FddA@JQ(3=5#XV%b(+yfj!rb(%QZcsoLKGS0mulB*W^!9k
z!<U)Oa!VIc>NuL$%)+Y_Qv)!sdoQS6%iX`dW-aIIO%sfy%NHhyi>eS?WW7mI%-j>@
z=wIQ{V8h6CZJ{*o?k{G|Lm_{$wsDGxO63ywD?8&nsIU-kE!uEgIBYyuZ3K?c9vdiq
z_)B&vhimdKo`hu+OvKol8SV8Z4=jon%lM;#-fGF-^kec7cB^XSuNgBryYeGwNQ3Ua
z5Ekk7bVszg!_K@Fgc-re6GaO76GtZPEVF;nP*Au@{Q6egq3rcUvK_W3f)9HQ3SVAa
z$X=z^p}%@j^V&)EcxdM_=WTm#`Q{7q*QjrBAyZ%r%*-S~+MTEYB9iu++X#B9`mTlX
z>!(+U7VbV<PkVajTr7ocYWK@4t#BG2WATGo9_-g!h_4Zy=Ug56%p5b%-#h9d{7F1~
zy4^*>`-eX2&g{VOkT)=Ix^&33PWr{)rMA}lBT*Z8*(mnF!L#gJq8{GLB87-FVsg6D
z!L{1?dVj9eu7?Ge`!CQxftRojuXGPQsnE;5AA9eE#?9Ny$pVoP!cV#5PnW!r=`E3?
zIpKTx)h&XGex?(q>fH-$_UOWgR*yrH=q0TKKHvMg%%#O9u4JXSLlW`LZ^9U%JmQVK
z7onlG@6c%F$YHuQEO%?!BY67}(jQ3$7vb}JvL8_*+c-e<@oPZW`Fhw|ovaJK(b)_g
z{-c^r<&kNr%4rJ;OFiSNofL$!f{0062S*oChBtP9;dthT&qkR01sj-s&_kq6@G8e-
ziW6CP6qh{WddK)JzOc(!&6)usk*E>xU$>|FuhbVU!{v-GnzRvG%uvg3F`KBW(U5Rt
zHieza)l`MzB}L101jDwkE5)hHJ{A2?`#a=aZB15TOy5GZY3|)%mcU!{BZ=_oMcWm5
z>M>u!To*oO#1;OQr<mVI(8ICxS5GM|zDPG}$W(c(*h#vQTF$XQSL&TuNqa`?q4yb0
zy1~06349Tm#1>QUW`doX&d5!lZo0#}Fkx;N?(}w<t;*sN<VcNoly3hq!xtsN#WjC7
zk2K|LC8(X6yyw4PZuFF%P+HwouoNPH(;*>9Tg5BOpY*U?JbQsF8ZFL3hw6%%EoNTq
zS~as-vbI!+2qf3a;aXgfrZ!o>Y;;y{cA2j8z)$4R+r*@ItT}1V>x2<Ug4PQeUy_bi
z>Bdf6_gOT889ar>&Ay>=vqPk@X2QQ$gL!P;9~JqODb+0M=?Pp5X8+XdD}~5&1FifE
z7Ws+JWf4ka83L9zFeFW%^7VxYoej#)h3+YD&6U#?U!Ej0dr8ZPo8fu`@1*Q&eD%30
zvrF_iSpxIN2NXsolP$6rzNY1MpDNsnGl%c%N_S6<1`U=(4`?w*aa|mx<+q&n&477z
zm2Mcq&LgDG5pm{E){Mk+BGa8kzZk0K9p2d|O7=TJN!@tA7X>E1#WaoJ-NfLRNX9bh
zq`||E8=Bnci|=4CSHpPNGAw(6lPL{*d8t<Rx(s8kqh7lg<<L@NOWi!v`SkNtFZl}O
zd#@b3mui=SzDZ|Av`mDM_a>cQ89B0rjp%RuI&6@uu_}ALZIH^sA+k59ee{Pn{x3+6
zXS@36Cjjn=HB}&N|K~{F`hW9*7a_z68+vcIr`Ayej^fX;!**L`pH#Q@)$L4duN5^7
z&BV0g--g9--cu>fiE$=JOI>7mMFFvhey4kyhiPYLz7SjGD8r?l?@slA>+i`vyF1^Q
z#j?F)!KtmQn*d~K^<u25d!My2rg}geP&vA#=3QlvB>$d&f`Gj%LrKTPM;PL;lVNbe
z2&Dj-v+XF!{c$)TsBPm0#CJMb_3GWXt2?ywbz3Nv>9^HxK#csK<C16HlGbb*Yhdx<
z439};8lcCDpAWKVh%a5U>`PGulpr|w-a<xMRaGLyh%NY5PCNBjWgY%)Zz(Z&++TQf
z?OFIUK$LMCRN73VAvKHNW5*ogvOK-Ku!kKo{g57Xell=w^X2Ee|8ONCopGP54D13A
z7~!Odq}3Abymt0P!$0)W?Ci~<98D|0@Sz~xL@P(5omxJ@0>@lWiN1s0sgwC+kyB<4
z5bmZ^wpG6jNB;Ik4Ce=(*oe7?i_Q8Hy!Qwl79vD<fG8LQk(g^mO&QQTM6zR-67h=x
z|LAmndNA~|B4k1Gpc+t#azv0+G++nGYk!W#+mB*xNUbr;klqB*SYt+R3<cUG{y_eT
zi1OXdu0I+nBK!dKiuYVPPN#cgWq?rM|K(-^#!h@L@TzW!*#*6XZHRk(r&BNYFJ5C%
z)x>dZKxUf~+bAD$UyM=4xj*zO0M6*M`v~Ie&yJhUQ^6^BmiYXV`Fwt>0`9z|oyvgU
zYu~03+h;G}cjoPT*veE2__tEKUw7?lmZ&&1GkQZgGwz$1Fex**T=QBlxoAC_KNl|8
z{{!_;UX|=nPa0rbTUnU}JlZqBKidGVmrdn6=<ovVuy0y9cn*M*E9>E{+v?Kvz2bk2
zZpws5iU&aiC-hEdUmpW8-b|+}%In3g<kLVPqIVsC1_gEfrb`jL<;{h4!2eAH0S1z>
z^b0D*^B~^-&!PD<P$jnEu$}J<h$e~6>Hr{7Knmh!<FOq?RZq>Fxsr>`;%qh?7)OfD
z6V)GhAJok(0b29|An+{#cZhQ`bAHMMNeha6o|FeOeljx069Y+r<y46+w(O$?{G{A9
z#rky%zNy}d>;Dalm>je3Y+d;kKpDn80ivN<`@8FqF6g~sZL+XCq=^E3^g%l>hK(4K
zL^Ds<t&%0&csm%qxGyrgo4Q1LV3tdUG$ghs#~Kratw*;S4;rGNL4nj)!BI$!_;Z};
zUFA?A+yc5t2HG1X0rBkK+hHaIvu#&37a>_SPrD!sP`5z7c^9WCoeWJU98yx!t+>pS
zKp=Do+9M@_f>X4$k{z}4GqLfNv-ZD0%uu?z9B_;a2^xT@{~l(%%QCqkf&(~uBiD(y
zQ}#{iK($NE$Y7Mlc4_{h&F3!rCE;I|j0T);6i?+io<EJC<*j>=1`QoMl7gMTegTug
zf4l%6W6g1Z6nWkcMA>dgom?$x54(W@bo*=D%88%>-~BRf<N73MiLn8aF(gth80-B#
z8OwmoRbG`!hdH^UqoZnY!MqORy<D9VFc;2FP%?tf%WBBoFQVIfMSkf(_K5@%i7n!P
zR)Ml--`Tssgh`h25&>GYAc)#;j7>lR_ZX`~zJ*x|CV))uKZi&#@m|a6XV5n1k*7Ly
z4MmJoc6A#f@KBLL?sVcWM|GSOILbH4GF&^in)z>shVCrD-0`H{^?5B6>!QXLcE@RX
z(C*xHZ5K5Y3u3OL$OzROc?{B;EJ(j~75yg0Y2z@n^@?s>rsyk-zg=O+!o3kam`D!$
z!N`n>oXc-FEl_FYc%?{o<?6EtX0phq&q7=~pC5@xFQUi@wxW}WthhhcTNbw=wD;bX
zu<3F~r<~MmJgj`+tG3E;`kP-~@2WpfEU)K0Id@QkCQZM-kkNxxLE1+=e2H%5KH3#N
zHY{25wJ{~6byUoOt9O|R!eP|F8KAG%eSeh0nf!)iOi8)X19Mhd$B0f{0W@`Y%CL>v
zFAl{9W_ErGfj6OF-`3k%?0ecYu)w!luD%nY>zV`v;EG@Z<(Q?BR}uxDdg7HS&|<)2
zO>)4EM0#aY4%C+!Vk*iF0Xrf$Wks-q8d(UOD{og{OK?>F*=(8=|Ekob^a7byaF$Ku
zp$c(0L5J*f!F7G=9bE<ugK}#c-tL~&t?F6F!}@_R!*cd0Q9$*^WeOfTx)dqGeEL4t
z*LM*aIRYfEkyEfhT<-mVYUeCWUBL|F$iJ>1Zf*<MDcaWP7qJSkQ^x>u1Nh_VuEnnb
zcPUFs6g=#&Ji|#%EVzYwIBMQj{^BS;K@&y|II>@l@C}}P$xr0!Hj-X`GXVd5xXsM}
zgYnyZp50Z$t*1=b=&DCts(HF3fm!F6BM~G0cAVxAYn|<i{F>N~SmwL7WC71QtEfAb
zs2yz|XU2>tb?URgiKR`|rA^`;@4*J*mqe3Z+P$c`c0-a?FSW8~0pS%TT$yV^^qK<i
zo7+<??vX+3Nnd^zA5B)?^D5lLj!z_w&rj=ljHuI7drHMfUr4JCiu@43)hUnQd;*V!
zEtIF3DE$<P5*p)=7ot(~3{_XS`H*Y%{roq|+(tn~n}qEVQX1Phif19k)N00dnFk}Q
zqd0CCP139CU1<^t<);?6Cs0b%PU?h4g6OM>Pxmw%T7#P%M$HuSqFiHPw^uulU-lY4
z?7Pg2ZhF|Jce}JQ_xPyW^^#(Jr0v1Z$?FK2gKg@a@&2(4{z7y&wJQP+Q&7N)ag8OG
ztjRihYxj1IQX9h174k|wpP%-YmwJ<v5CE#FGNi?F@{8a%c*JL_cHOuGohb{z_UZGo
z^BnP&Vs-5+9%r#c9?vs-#zvl`Q3n}@eau(kQthZH!(hkBIqZjqEzA|CMU7Hpq6O|N
z@i|gzQQ$gup3~!RCv97VRq~-;=&V0tvCgZdffu(=!}dKbj^9#jwM0ki4NUU3QfrL}
z%TP^wV&31Lmq6E@crVdY5qZoc5BeFnv<#6)R&IYf%yP{>4k^EL<Edo0ZIb|Rw{Mjq
z_vAbiSNZ`M_>R#5v#p@jbBp@wS(SinOEVUD-cIc-rsRQiQ349aaqeT5<>>QpREpw$
z7%c7WpeteSep^(fmUcdw2h)<2@d$rlJ@fhsk|j{qCi^96j$+d>MXn1s3GWQobz7at
z3-)+CiA5Kd>gj1z%~)e^6;9%)3d=0p5|%{pw<l4%@;pwYoC=Ao1~h8*E%?;ZLcvJ`
z#)CfIh5NV-P&q6D=wd}0v($mN7IlexE=UxL&kPocv2GX1qG~rs!jeSvMhumwtgy)2
z)h{E95wl^gx6Y+Sd|Q(a?{tJ?GDJKWXl45rH{z;_m?9doG}8~B8d$YFVrA4plx7AQ
zaTVo-=I&6#l@^!}U!JF&Pdvf+GifdFcv$#V1+a2u*uwx-T-j5MxRShGLXJJ;9a}9q
z|6y`Aas2=f-ZRd-94SE9(->~6Fl7~f=*oz%xSqJwt41I)BWmStSYfx^6cye=D37DB
zXr%Q0(Bo>iSMnhCW(lQj+>Fn77Lx$#zL65EMy{uM%U~kevrg1or?V)v#+=(z=Hpc~
z-TEUZ)KkxUr*8n-N@#IorDL5&jnX8h<Z1eb=Z3y9mq7H<3A%;gG5yE!t2iURcbbB7
zip1L-hRuwG)L0+nmOM7@p$%|uoP=8-DR02P(Z{veREs38>3V1i?AMZ;Tri-bZy`wG
z$nxs-G14y2<C-t1AG>vcx0E!nCVb>>lkK<`1;ZXa+I%`Rn_Fox){WMjzm+>#WQl=|
zJKerbjxww`FIe`3m$V<D<ZBwSFVE*2I&N5-c_3dR_4t;%8@`eh=9<)0COGdM4S4W!
z!Ap}XFg~8ww8G}S)djG=dEu%3Vjo#^9|^)7OGaCsbJxw%Y;T;L7phNL9nO}tPcPT!
z#0;-Q(!H9z`*ED^axKR<w@=b08}L-7^Me@a)klD(7EA;5_Ye*0AlhexSzqtLl*Q9&
z@~TvY(O4>sI_-{Ldytb(iSe(F@Fk~49t>mScoCZj3p1bQDnnnn`qbxz{pvk7!dUgn
z?Lf8I4UmgL`Pl!CR_I}KNW0tr$pe5hP5M6;66gVI{{27Zia!rq4$1$|A7tP%6UY($
z?=5Eos0(L68IjayMDRD|_qY_#=0z)p3dstgz(Y{iADjTRVF^-eIMC{&j;8_60;aSM
zsY;ObS(R4*yKUf{$sk<g>;xZT5u%MP$yOP}4^Qpj{o*M41a7?@8>)X8X*8Q{0$`+8
z{knVK*CwrtmOjO4%tM|fD9>&>+bf!Qk&~VzYR`!D%Kx>i+{ECq7z5HJ_h~r5cC#U@
zk?u;XH~(Ik?s4q;?T{=Yf8ZUU@Gq!B(pvLimird)nGSJeK=O77vZM%DkktM-{;cCw
zbDV~ThTZ*r>9Pwq{${fspY5xr!ObF%9iuArc_19H1Er?GGZG%<vkp&t^-~3BcXt=U
zanRj>^!y_5?L3HO%QtSQ^*lR5{e#dba`rBf@6=iic<VsFX9t!i2|r*ezyhOV$94zP
z0A%FzJ00_LJj0@#H!8;PJg#u7T(KO;=mUlnJpg#oz()mn%U+*cw_pzJ)8_m;EEDC#
z>40sVO`DqCuG_2)gX~RijhOT_mr*^O-Zb$zC<BJvcw(sM75Z(Cy?lOHRl<uGOGaZY
zXsVE&Twzo^^K~<)ybx*z2A3NUP;%(E-Dy4kn-rU4Qi?NGw0%GLV#VQb2DfQPp;UDw
zCgZ>X;3@4W(NuNFiO1Wp`}k~Q<}9MB>2$-;q2&T(--4hE$h;vZn)9?{2gv;Ym^a;K
zd?tWPC9^O0pB+_8=$VO%Wx?m7qA18(0AN4_lx@F$jf(?WxrSzD1`-VHh-K@35zc-|
zwVMY%toe&Fhqw>~7oYXV2RC;D>?ZiLSz)|p0Bu?U3y<vC17Y^g|75hlyLFu5rc%1p
zlL#4~@UVI9rQB01ju4^`yTP|yIc1j(uKj0RT_<1QK(;gx587QN`UQl|*L4i<)%}8O
zL6Rq1t<w&zWb5D`xQ%MB`DaJo{ev<9!~!w&_`g?MG-+!(nc|TrT8Mqr1b}2U@RLA>
zKPZB8FVEUn=mTqz98tbORU`!QpuPhB21*=5=RRS2=U!O-UV*&@u%p<32?lroo^{}6
zvH{2xr+!(uXb!U64rVFt9vnEYmvlf@s-_3;*j0A8$Bh5oE&2IQ5t|@_pP#>ZbtDfL
zxH_fAEmLhF2asxlg*PzYF3$#BUWPD}z27k%f?AFUvOEF9CX+vRkNZfTaXaL7155#4
zGq9(6LasE3NLkefUNbt#Py|IvH3Y3!zh>tW{4*1utQ{q4#{TkLE2zh$qZdIx;hO>r
zA>;-<0bpqsfWVuPa1F6Jg3Wr2MkAQkm`>IxeqiDRwjnbJVdJDZ2T~*flmd(_1hXFk
z$jyLj3j&lZfe(%B5#cnwsiNqKqObp@H$o)gXG?MBAh3IYcWXITm3I!<H@P5#9KdZL
z8tte3&QH_a=R+Y=Ei}SVw?qC-$WscO%b@^<lYa1|TmI!X(+9|o`$Vt7Xf^`7jzcGN
z@Jd<xzfMI#7lPa<kcX-1{1mNMY>ErbqF@GBYC)_1PsT+$9Sy)+X0<x}Z5ng|Fy;Wa
znFlbui7h9=#rB4^07C>Tv+OgQ`~LRnV;c~uIXO92&QI0<U4#96N2L@%@pfo^0Un-U
z&_ZH=&>-OH(u9Z$u#)1gtgf0t7T%@q_t2<8ezTLUV3|eXA`dc%jgV(LqL?`vtx6KG
z=w8l`{M%DDgwV<L3lP%B>++hKmqn)?@WIzV4nE(CO{^zCO99@+$YmY)5OwQZc|?D`
zx{4#QUW{)KykO$ch=YkHqI3OkQ_Tg#MBQ*RcrmB}RP9>f0&r}@!AL`%Uz5{Sos(tJ
zCSWyo$HxH+6BjUp(?djzCGSSu0<rr|JeHj*f5Myi0(gaoxSd-R9~kI)-2p-nVZlT;
z?o-5+>?)H1Z;OwTId{wizcQ^_^My`+G5puE;mjgb!2{=ZtgDak0OG*sj~CYaMc?e9
zNw2tD6{I&s|EX@&R6jcrS$#E(nIw{VjD=3Ce_SCji!n4bB)r{}s3mR==WB%cmy<v-
zO~z(MhZZ=^>AD!4G~h+t*(iUb#KS6LhRPrqA)GX(!X1w%b9s#AVHqsrVTCFc{cC>)
zO!BLs`5I;lIp2KYpmC*F94-Rr=XK$|r~&n?q{wPQS#xe-x~nv*Jvrgy@hPD=Asz9*
zZpeH*JRf(zwPqR7azS&-ikj3mGe%s;$IE(|p|$X-QI!ncGcoHySDKiLPrR&q5C{%=
zG2}|~Z6N8%nps}t?P1&R#n!Osh^>`QeeR%^uK~UWz{<12@*S_App=4$fRvC)vi~^u
zR9`yVmj1=oQ*6&ZcN2N5X>KFe@%VM&1A_b#o57Xu7c~8>lU|P!rK_ZGetHcH`dF?q
zWdS5?8-*)Nw$rXOJc$Bnk(KgN8br*|N|BwqcWvbQfNu#N{@HFPMq4AC*PTAY1?aE5
z@CVtcG5P8|@4ih?XJwhZJMzG*{Hod4#wPOJw9}9g{ke_G(T;Q@8a6+fZ+kfwTW<*A
z-v?Z(q4@=f2j(_|9E8P(vou{hFtKqvfr7Fs+J&f3@}%ahyTZXM4_0;IFBFJFBXXVM
zQRRy;8`#K<R%Xm~pZA2p?X$#U=n}-kFf(r=0Mn@j?IfeIgSxvd@Q6NC)+hx};?KBp
z%r=#rGzaYIu>wa?5f7tg0H^TlgbJZI%y$~^w+Gd4s8dQTRd;$bo_`b*m7V$wEHxga
z<AE9m9klX+%p#A3n-aRm%N1+%nX|9cN3ckDQ|~J_;gDMKn9<G2OBb0V=jsRGUzcB8
zbEc5BdSVq;9oD9MG))Aw>C2Z35lP;ppI;zkB_Ho;1{XZ*sU#!UL8O$72&-DEO?Yo_
zjW5W-j9)wYohiSiXnxoy@vX-}q$Sv27<*WVkoL&bhi@%5g@uXB@Hv0&Pdt$SV3e<O
zts`+ED7s_9VKR3)YFI*_w{JkvIK!#1SEopJnLadP1-LXg(qAdMs5z?9xJfUr-TTGy
zbh>K(q)auHm)*JWPA0~K6krjycaG9cu_PJKX1=^&=qHRdEmpL#Ae`sF+Z44CK(XT+
zk$*w7QYNgCgqf+;F=AY)+$RW=(JSE%3+fiGWn97-QK2^Ci*&Nr8sR6C+LZ=+Ux-xW
zHHF0ztLDaQZrw&i@*Hy{y;R@#=XyzX9o|P}ZA+!p^D=_H<jd-UEcMkYkK}G~+@1Z-
zHCiIelW{&&;R$`azQD3ape)Zj<xCTV!3d$p`J#{UFtLPPhM`}^?Xf!V)6kA#<|i(s
z@a38vU8a+2Uz4WHo{nx*V04uaRwC>pBl}aoN<)1Hcmv<QQ{yV1R{6lVPo{7}|EGRg
z>rb^B=3%E%ZrLRd4Fws?T;uVwo!@bd7CEiIN9^0$umWJ*5ZCf#?S+*gr$9n~wZ@bc
z=*4+6CW-ke4~C9jP75zZ%U9<Q6J~<>3pGj@(yb<ec#fUv*nf?|p0T4v4GMzcYkGy&
zB@wx&(}cDHmc~2_KPCrDEW<MQ#u!k_$D&-LSf~CngblyulM>U>yqpT4IfCRzuYw9?
z9l%KtH)DDYb*saurYby~p-ShVecU!UTtB2a+#a>#%l7>TKcy}GMs$>LFjLjVI~}?n
z4duFJJ+2|Pw-ArcByLESgzm_s4tjp(#osBpLiDNSDT;gQo21oC299g+;-n}uy1|g!
zlbo|Xs2k3?B#o2j@f%O}LoYw;)g3tSrIxo*A33vAGZ1|#419Rm)qZyl)GbJB%rv$F
zSEWDa2VgR8mC~JtuXj$G8TAb$dx#c>ml*N&cRO7#iVN*o^U?%>X$m_lT5T*oGhxO<
z%$zrRn7@*mps5n(LYz8tpti9>Gsd)(F>no|E8Y(x(S=6(Va4ljlZ=McO(tC^xWuM(
zUy)nyk(%RxcDwC}^7*v!%EUFf=C}c^(1_6NkT#{K4c+3EmJ2WWh8TpuCRf*H?lVUx
zeXYIBu(NT3?6Mk})ZTA0vOn!y=+iznf-F?Px%g4yB~G#8bbC+ii`}>J)`W?$Jn7_#
zu4`YCY0piZ#e0IamoT=J=tgQmp3I-6G(3|1{`k{LwTV);9lIt}h80b>$7-DhbRw~A
zwCW!3_f%p#{0hdKPPn=r0Fqt?A0L0(eIWuALO|Yckeflt^zXGb{GHb&1JcJorl#tW
zRsV@h1MCB-I<XBaihT1w<S#Fk_V6x%H0Y1d*x@_zpOS}Z#s5JG)hP8?QV7VH7B%df
zjppkuG*0}@C>0jBC4u!nx+&`)5x)4%9qG;SFY)S>dGJs805;qIM?zhSIBE-NziIY2
z#pu+cu;0X!WD2Cf;YUy?0@J@B!e*!D^z;;@t5IGyKoijPWB!LN;viHRuKvczYoh|H
z^?>Ss=zS080Cd4PW_kf!RcziBYgHTa7uW5Gf&X|1^f~JecEo6bL5HT_k=7UWR0I(1
zT0w0E76=z6kgx$U^gG{gG&;l;-IDk}q`C01zfDkZ0>YY#MXdE{l@=bPr*P|)1_D2Q
z29W6Z0i{j_Z1$jdqHn<O%o(aOu@O)iotjd$w?~&fy6_K%LGmCQoD}H=;UHgZk4Y)u
z=gWY$Q4GBUHhu>PQ*TD{bRd>K&AEQ}GoQurMimNxnij99ilfNCKsyF?$HHzj&O&eR
z-U)nfZ_n{$yTiGQJ+T;6fkCR5Gypu9<C*gepfeN$DS**qkuoSaI5aDyzh46qcRD+X
zpv)9k#MAg+>rQqb2q66ME$C6|KWAoSHf^Juj(?$b8@<JE7`=M5HxAWG0NiE^08VR<
z$(a)*CIWnH<bxR3kl0Yj9-Mv0RcHz{r>J9|#J?j^id7rei3M~b5<h!JgGO#vS>}im
zC^<JFj0XXpTpd4<1~$z@nhYc%LB(eiAQjgGY8}8_kj?}lR3Q5R`Q{|67OJq2BNtz!
zAiulv!K&%a-wW-0QU@2JfrEIGn}zUmLqo$KU<A(Pz3H5}0D-|?!GTZN*#SVZ5Oi66
z0_m!Mih97{{m#kk2j95V49xo=AHGz#UP=NLV?{~rQxAPnXN`nIs)_A#%0IpIs)sSa
z>-%J}5HUjM?QjN=TL<(4<V9Dl%i7BC-*a=pOv{+eb4jQkBhkopuH_=rA~d(R=i9mU
z%R`{z1gKmElnUZ(0r0imfc60w91p-J7xO>epNfKnDBH<$b>Gy_f2WS*fuwz``MHdj
z8!eBH9?I+FSkmk3U6*C~WtC1GT$VNp-g^qqXa>n&2~yj<NO`MjRQ#dgpmKctr!9Wp
zhTWhAUwhB^SASkQg59iL_BV3PF$#}l<$Bq1V<}=~oe;Usxi8VDk_YXDxz2bm<--VK
zXM$a*#Oqu)JeN`5&L`wQKmG_h_Ig?hHNDr1;#KGdLG>B;s6JHT0*bkMp~@PlB6D(T
z3U_5?#T4k!5csGE5)!@3@S}WC71wlJ*8~U9R1zQ^o5}Lfs~Tdj<;lef{6Q5$g|s~G
zUn4<`Xq3-<hPlk?Z%qhTxX-=4p^qy7Cmq%Q%M(Et>ra=!O#%`j;BDZhrv6}T7aFX<
z0s@t^?X8Uoiii+4e+*{qesh7OdT-PK_fJXOpW+5E?9f<6`JKR}4{8QrR{d!W32uMq
z1GaVJ>b=s!LS9f?+70SfqyXYH*F!)>c$sIvGk1WF>;RBZiPO>pP_Wb%e_IPO9ejQq
znv>~U>oPh9Q8u9Xp_hcO9!J*AO%&2e!9>R0DeHk|pvg(Qi35VhilSp1*;V{^bnCw$
z%kK%GrSuYTA;t(WMU$ST<`e{y$LnR@pa2YYv{r-y1tXWnjJ<e*>79t6LccmUv?Yir
zwG^RHs9v=Io&m@SsAD-Zy|4K5lVSnr>lFnI!<bCKzv)uzF+|uE4kNhNjEJDC!7isJ
z3tLl7RocQqwBt;JdMkxTZJ;uWDwd6zRK8#8z)-{<*$b9Xv8`r&2+&3WJHc8K)C2%j
zvy6+X@EjnQ0GR$>>{0)&a^x;g*yRgmuEUJa0d6##3YV$yqXdhSA(_*wP+KSJJp`kq
zk8bfb9<l&M7H?;Ew9<AeWFd*aH&fFAtZJAO%mDsFZYpt5v*xk-El0P^Qhlv@29#hz
zIGOm(Y*f1X-x2c&aBr3RC7^Pp5f}mH(-x{1BCVxtcRoX_A5=)Z189v@s4N$pQ|s2X
zYOq-Tsfz;4mYp;RGHB)0xci&!&COXbcsYRSk^6|#{PVLa7OJ8Ig`4^v75ii~qSyX=
z!5En=(I?Tp;>tkehFA<hzcH?j7Qn%h^`~4&yaU!j-`f423q1Pa2b;gbBO?W@`svf0
zq0l@4a~<M|{+K5~OBd3d`{%$4$MR>z$<agsPGkXmhfkT&gAE5F-oWdJasM9LvQu-<
zI%wlN@s!+7Fqs(IK;oDfU4fvO^dGhuE{pG#kYt=0pP>!2rN?h}FJ>eqyeNiW=qM{X
zS6xE(`*!)&6#vby3eWMbBtHI_VuEMQ^SY$`x*6i<-6hZWX4G26z{}HF&x*lP;>V$S
z7aWmSsU0GOJOwjSi%T+FX1C1`wNvP^Uivaf$|NP)?<Xyn_PE*iXlTd-n1yg%b_8yZ
zhcxQ~EiX1kJXjQ4l2SwiCkp?(c_kbsmgyHMoG=WaWJo}9MH7ZGs>3|!<zvoL+>d1X
zdq1wfzcrck;n79hpdZw|fk)3r{rI2dwTH_zRFnwwC;V8nlA`(^;w0rKV0|sHO@vK^
znI>`=wb3l{(>}uXbC1`|OFjwZi^P*W7ZW)I#d-Nxu8uwt(JF@xwUV20Gj?w2-rNd`
z-Y%zu>tFuvfQ*~#q^+=+5R5AD7T&Ql-tjD>$F~kumws~nO-M<(#xlbQH|~enVW6Hc
zhldaqC!ORx(k>iC)*l2)clI>IAbKUzwslEqGKf_DTjo^gHhR4@rBn@U={_W|dnDHi
z6=9TWeX?&L5I-%m>(zC`Iz)fQp>-~iyb5MNdgFShu%FuHabi`|Ez(dMS=vPx?SY@$
zwUzhs$d)}ir#vFrUnl+b0{r^icCT|+j<4Yvalo^S@}!*E*IGm5D1^H`Poi+X4N&*2
zvhiBSSFvIU7sx_;h~CsA9%ojpiZstgaJppl*3_TE9eht_7WNVksy+-NNP{Z*zesUE
zE`F5FV_9%SWGfr{`BTKMdH*A4;@gpdQo<cAAH+F?q?`0KV#qMHwvoajoomAu1t&sN
z@a(F}${HmGk&gHi4CzN2`{RDmj?{WaJoxRm7arWhMN8GKu|D~EXOFum@dtq}yvu<=
zchW}9eCJvl$0w;<KlGkw``MV+`x+t}Qx@362QCoG%NpAy)A>nZB87VNT_}alRu;T7
z4wetkA0$~e+3a4QS&8b)Q5H3^wzIG7>6M68Y6`LWu`1{_n%N@KUHWZ8IlROpSr|#D
zMt#oARL2o{y>Cn~vno+Pp&YR;Ms%CdaqTpe%mJxI3ZHzFwEsc2MJb}v_Jw_T52NO<
zs4CR5cl6r@j7}N1AWDBBV-GuS9r5uR)vcuH4<nT+9s{V!Wpw1W$24`^K%Zqj)5LPp
zDyF2ELZQIv8(mpPy>yAMJlzkfHZin{_vk89Upe*79c<DKI@Ok-f_nTw%lgeUIvKk%
zH3784p?9RszWl;R7($WS%-#{PvE3MY?ux9M98fi}Zp=swvyaDJSi5)}E3D>hXSb_5
zv(nUeHqooPS|-&^cXOLYIPr>T;Z6QN@j&lWnB!(QDt=oUC&F<?eqPH4o>oa1%RR$K
zE*}mX{`m=}twB8i3uhO`qkH)7(kp~*fsk^<-DI(4Wf8Wvs?{8g$<?VoA>k=a%t5cj
zmZ`0@FP&PR-(B>rK+2C~A#aK+G*We$`W6lm)RXiE9Lr65U%vHh9d4gVJR^M>ne)0N
zK1&XXO!+Osk=bD*ZS2~?e)QdS(iS)>_`X;KQ8O?#8LW{8+wP_=8xGs(AR@=~Fz0<|
z^IHV6=5Iz1>BsBUN|osB7@f@w(sT=30>?OCHZp3g(DhZ`bv#||LKTNv$9j`e%UGIo
zZ#;|9Pgq>jl6<taQ)JiY`kU6-_Bp;8-7>t%13MztH*jg)odK)VAUb7|hooJ@aE$2&
zb0S~UoU=^OXS0&h1SR>mE$&@)pVGCeNgOB;=^T?3VSRC}pApW2m5dV^C5h0UShe0`
zV}H3QjKR`R^#&)4C2W@T?-n>Rv&+~m|I{{^$Tr;iv7&Qt;TF~lvwf==>xL`(F@rgi
zP9U*jQu}uI7*YGpmp(VT*Q~usNyFI%JUrMU(4>86`((EQtnYmFsNt9<ca?VC^ZKHv
zgq8!s#rQuhGhgIaHDJUt6MdY+9~h1554{z6UQm=@QXph0Cih0W0e4qnV^`>CyEw13
zPWn6Mz3<%xk@=f#X8O*rmKi>heVqt&YPkE6FREy1M{-;Wk!dNN`r^b#-#U@lI<?*>
zS!(&CXsIP**3D1j)ODPNWhooq)t<9npc{Nfa!vjIV7STc8*m_##vO&*84&E0m^+oE
zCcl-=V4Uj8bmd*_Z!r<f{?W7hAz(u<Iku~5NXICB*h<ZQja`J<C3EMM1t*Dwr9f=N
zAk(Pc;*_qTC)a_XY>-uFPH&{=M0ernfmr}^DXx^G1iQga5!^bX_a2p<@j`IzXYn8P
z)naKb#eDunwes#kfp>V#iPtmd@LO*$JWhnAjDD?)fBASO+c)*J{yxDqKPx@--q4p)
z?p9$5AGoaN)D%+5z<>_q2J=v1^Pd7(XipB1$<+X<>vm$#@JnG}j%)`Rw>~-*WbXaq
zjNV)KaCoa{@t~aa&ihLsRUv{((E*GZH15jzmuoq0><M<I9Gub8&^!Z~4^&GG($GMF
znYN%zfrPi(A6V@!`-2>?aqBH#G1yW9D)BzY%A5y-@>55!od-UUBotD<OnacxAs^~1
zH#f|FRZ}|9QCy>%IR0KKM6G-*z#Kw1(B=%OqV^D~0m4u`q%(kxFVJ&<70(7o1sZZ@
z0#|4Y@H4XZOJdZ{fJ-wCs%A#fFF7TX(TO+=D&C+ZXkkobxF!3~^VKw9OcpBz@GG#o
z&CEer15{@WZ8Cxdh5O$@*?sfON#6{lRBVAwKF;xXdITgSBql1XZ-8tbs7%0&dJp8N
zGC+I*dm8v12K$XKaGd;2<5-H?!1?=9gu@`4C8R+W*tWK|=Id9yYr>{B%|Y@jRH&`j
z;3>?R;m)F$Py+8DkCfWiofm5woN?@6=m4*(s?vYL>#=_WEnDhcmpi~sx&&&OXM^bM
znyFh$Iv9JJn{i|gTQ0kUy5|m{$aX`ix^|IqpIg(h5fW5pJI#l3ZuXgn{kL;Fr7eqQ
z;?_Y8^q>3-+O-0LtbFxbcu>$lsBZ>Hj=I$j%uw(^GTwI*K4|+0FNv<dH&_}qv39St
zz@Taevj0KbUBtvt{@G%I&!ivhuyU`)a2eVT#>dCUOCtH-e_fse{{?NGQq{Ou0P;6+
zo9XHp8El-)oZm^*c?+oS-go`^PY97I0YL!H0Lgcfpc0!NDAU)%+8ff@{wA;!1M7P?
zX_^6)fFv2D#@|BbkoBB(qz1Gt&Yz<b`Q_$6LTCNp+W)w)Pcz<_|7+chR`j~C0FcZ8
zN-C?0WB&WW`33us9D_{~M!yqB-r5xiWE|yY42y$e@vW^b#wpkJ@ojz1KSkM4+kG!z
z=sF{*Kfs}e>rX*r3i%!nn$CUtfj*?;oo`@Q^A?of9|>z_|7~Es)1}ykOHDjtl`n=k
zfDLhFV`B-Z*rJfwWc)OG9oV*)fYoybJ{8FzVLSpnQ2#j=FK6GJ$ooMWSzywXq0HnH
zKR5Txt`Tb~4Q!=5pe8>Ka^XhUdk3heioooV^`po22VMlD9ILCVS$TPDE-FobTX5o9
z>v(P-V?m>r09b&7i~*34gRM!J=Dy`*%mbpv!reUyfJ3dKF+C6%&Zzz+NZ!VIk#dH_
z`Pa4eu=T`yrOOs^P*U&UK!JlY<l?6yz`aLNzHdEQ?Z^t*U8}*y0X4fnRe{-Dy7hKW
z+IQgdY=$r?gimS!uY*g}N&zR;l-wrBJU>CvFfjNul-~L~7DeBEX@CH`1-1ilUV5+d
zm^Tr(c=QSZZnVs|9Y7Q;EXHXrL#i3Tsko*alNC7huYTNPQ^8w2xaPH-Y&8xVf#bJt
z<%a+p`sh}?ybi%ZX!907Y@%Nm*#F>u%9kRyf6mp#>*IWAmE|-zkV+wQc+>`@-T}4}
zn2sT+^6om<Auul>!19HF+y)SA9@3OiFVM~gR#_o+aUj)iApzc_jg&rm4)JNwQUz^t
z1g(n}iAIQ>rJjSw8h=@{#rNZU0p_F2oIfHWBDR5&>;%jJ0&oOr?#P|<efx6+AcG!B
z@5Dsm{1_YS0K4HdgUJteNBEeR7X~?!FZ-;>mseFm7rZMWp;u+c0Jds@2pmAOVIa86
zPtS)GQZE^;ls5PqLrAgxo{&Dk`9XsMAx5x4k#q}~!??|W<I(u|(6lR-y};)N*nI%#
zrjXkNx7p-;&jeHw9Bftj9akCpVwb$<NK9toOvlt~V?|c(?$rJ-hQ3e+F~3Nu+;pZ2
z`d_?#cR1GV|2GN|5-Pi7gjDt_D@As;tZ>;QS!GkAl##tjR<bu`6)r0wx~z=ICM&5F
z&-?WKHh$0Z&-48AbR7469Novw^|{W^x!$k$YortWS!9k^x6S*pVq2vSsEUO`oM>I3
z{YJyDTL=(DLt5m!dAeiTc~PSsrb5wXJ+OjdukH0Tn}#Mo7CG}Ii@Dr<>jdB(LId#1
z!^Uer`q2gzY$yMrJunPN(y@8X(;1+K@}46V%sDz*+9&!2&fxY?V$pdNm8}2;H$m}?
z=MFeQH+asdWog(jE#9M6H38@W?M$M34isrnan?*eJS_^xMGesI_pp=MZsK*~eZN;v
zTT&~0;^NB6ENqqbDTj561`>c#6A%!9pd4U4XcP^=M798IXKAc12~H)vVqokiyC6=B
zTP@h|>~ZC5TxJO#vym_e+fSgv7C8c?ZGV_WJbXX^Cl*t6e{Xz%wM5C3!q`~8O>PIL
zDE5wNeuSCby!m*eTPh8a+}RWcODzs=dZx5TopM=yMO=1k8NopF05Q$K%gQ%;S?N}5
zR?>z%gK%`V>NM)yF@q%-VniZ8B>C+Ga?SXv@kdPX=}?6l5wOtFp$+Vun_WT!l&yO$
zcQ_c3bp@ixoH(Fw6`lf=07MP2AFpTfKL-=T(5dA}!<GQz-Nfh`x5H+l!r1Ef<-(uW
zb-Wt}Zy#V7Fs0(qi49T&moL5YW$>OPfAog)^YbHj7g%<92o19<tO){~h|ymocYnyZ
z|MlM!`K!anG1bQ7eMm+C#}6<TOo2p&tViC!L5zka<@{}DfJpXz*W`=?l2qx|BGI2C
zd;7kR_?4qUEC5D{-yU^41`T{ubMyUe=VEF>Jz=n2_r)gOktRqYz?I?wuYW5(T`HNp
ze~X8jNux)2>l8;ap;hQB54F*({n*HFWQ?NgJ{nPJ8nl&Rhj6lHCT~8aQJXwambNBX
z$aKE|H3C-Cjl?ak@@;KrEvpNKzlPQm6ILD~J02|U$>Zhk;(n0LO5oyke3??8M0{-=
z)EN_4xfy!0r1q!al@$pr?FHr)iQvw*+6)?+VyqcM>sF{YF}Bi|e+An)N_&x-a)dHG
zhvsNoat5Vb|GRshu_hrbyhk0^%dN(sZcVtId3N8EysJ+WaZY1>M`8|ySQHzxY<^#S
z>qRy7Z0bQRi(rub#TWy651UUCPIN?l9aGIeDyRuwyq$HMs%`zq_sNFbp;1qNoHgy-
znFdrJb5hbPjh^S0ELv>)<Tj&XEXDPBLV>)$)~-LJ$z*kmda(8<DaV_tj~hSiqFqeL
zSU4P;Lbuf3$VzA2;OH-H5ghi&P;r;H-glCr{Hv3sl<^Ex1y50U&y39>hHZ;b*I~I)
zkrBf$2CDfE%z4)&q3{ZLyazelDG9|)+lbKeGX{5$FZzrfCASP0o>?te4_B5pu|K{1
z_B_91o8iuirGWr|Jx|_$4&loT6;6FYhDl6YUUj!Uo1U6gAY3MIB+@YoaEQ`pcB{9R
zV;|{#a%<A~zLyZ#t$lJB$B!U4Xmy{btM2ky_pG7`crX~GM~0no$iO{gm}OFbeEhV{
zmy9+30J64_bX&6EwFy9VB1oR8x`$hJ-<qU%J2LY4ZYbyQGofsuR+;b|I8`%z6TbtJ
zV7}hV+F1U>Lcmd-QscnM;eeBu86rI7&6{+j3M@#1l@ts~rR`tWN}1V+9Nzz3)Vgg^
z#No@vMOt;TlNYJW*AIQKo*i;YcEnlDtiSM^|IlmSP;`&aV?6$wq)7U*aFxgUx(M~K
zuQ`k=)ljka4nP{=i<RnU3wzU&c2$(inr4^BXEx|Vx{u^`>pM7_9lt4gJja_-{gB4O
zHOW&#+j1LQmOmEHvL9qjC*d4^@LXw7v@&Wr#&_nH3P%4!lZk%16qkTr-lwTrmk^GM
zwY3Iy;-j*T=>b(0-IievIiJp)e^$cIU=!M3seYg0z=jQ@_1mebU-=0|Gu79+-{A&h
zszdZ+;jO5jn{pe@&zPw`DvUkSw=_;Y>_KWyq=9+%G1pVRpjR?!bHb;-el^s|W-__&
zjg%nv%9KKz-@4>G>r<9LCAe;c3-(wayi;L*_V{d~i_AuKS7~wRT3H3wdTHeppbmA@
z%Zu#u6-OB_nf7pZ_gj3Azaje?yG@?`zB_668fK$hwaul>FmW*u>ie!W?9~HLvEpp(
zE@8ZFQdOxnIwK|6${D2<3DydAZtOd4&cnnkhp(0&W~w}Jl)DgXtv`Yx=WTG{juKzb
zJ8s-2ckQ}wSi}$7tY^{&l+^IfnX$r|G&I&_d<nNh%rPNs7||6v727wQn|x<JzaluI
zI~(|eGN6JqrkPq#iSb_aYl6TN1W#h8?WFvBGqJU)*J=wy1DL7RH|ZEkWm1dCi}S>l
z`#Jp_H#`K}HJ*?0#rE@Ido=B><VwO*pM*Zl6Nn65vv)fdK|V(|e8oqui+$-C8E-*d
zTw(PktA4rBtdRcJb*Cr2XZ<B`L@}Az(!wg<({m@9i@UUG+ijbLPuP4Bup<K~p@%4h
zL#TGEz{A$;{KJ69`iYc-*bn3GK9lzEm(J(j{{Dhc?Jg_tlwFL!cb8SD&87ureeL;-
z0991?v|Cz4b$N*@a~cM=tOUo<!(H8*ULhpselbfgmL*Qz+o%_E6dYKlCd5Ym$k*;J
zSbK#H_4dH7>aCZ&`R;weC3NK!U17<Y-7WWF`o)7^mCL}qU}Gm%e4KkLc-{N^nnzIy
zk!|77qGZ;>LVxN06`#C1%dtC3PNBZZ-$E`@@|7*)DGFyVP!aW!PM0g5QJD)KnI{?H
z8h()2J-ceDZ$@2-i|le@amDX^G0WE(TTF|~+4s|Vj;>?V_OxTy7ZJawB;!&^8+tAU
z$A(S3^X>XI%RVGx^n2X>dg9aW&fI=7<5^Jhw!PLa|4<n=ob^U|>&#A<n`K#;Lx#^i
z+&e(+&fn-Ba6}N^JL>Y1im~p>-qG4D2Y#igT$(zIvg6ai3%7!Pak9?SVy>lpErufj
z(6(z##E+}<g>WJRatXKl%hqaMCWJK(<F-CniV1Ja+WCwaKKgm8GkIn>E48k+pyApj
zVf%c6eU@<?d;*`@8+j+NgW{Esj<p8iuZv~&hH=Y;_7=;%k}Ds%xOiTvE32QuVf(wi
zf4XO&b6v;Tpdp<h6%(H8c{_KTEiTNbO?|IUd0QNM__^o`;K{tCv|YAeUh?B;vyXVW
zv(TAw(ph-9`5y;fgRJ=LlSEB=I#$n&Uro(BEH*}l*F{a8%_dGMPw+2Q!@IJ!M6pk~
zfawq!Q7ktXrts}ee#J?C0?hn@hj<}1;sFZm*QFj8b*oHmBrR@Q`fZL)DHHwtTv@C}
z+>RR0HA|(Pt79M|tHlZX3#yPb`Mfyr`)G%Icn<5OE|WoUX1iWVe%1d^x8C398}`@l
z;^1Ua^#3g9{`_xWy-5G42-|$sWPW$d4iBwxUHO9wY)cUC`Ej9-HE+oIzpfcS)wOyx
z{8KO%Mcy?p!k=Zt!_tQEd5H=Ae!g|vK(F)o+XQwSdQ<40i>h?^Mq)r_Ob~P7xE2dI
zg|TY-?_1?hNz%w)ci?6>VCX^)y_a(U6W9;&BX>@hQMUmnFt6Lm!7qiH4_eh*gO!-p
zFe%H(KeYrq4cqp<Um)n4LDXJwM@3$6(!O#-sHYFt-EP#zNp>kH>oQ(|KHEq5KYS&k
z{>UBArK8famTR|pJB$gKw8|;sF|S^|!c~ff0U<P7BNy-ITRgf(WS0hYLVar`)A?IS
ze%TPK@Q3U@E-O>2J5fd03Jt$YAngR~WDmW8CbBt@EPuB8{Uk<gNzH8KkKn8NLoIyt
z{rk(hx&z10EdLp_w7$MG*YC_|o*pQ=-i-hbgw9;wnZCXb3|e}BRN{i>5MAuokQ-Z+
zsyK<=ncYRdU*{0HyfUnhtn$xQli|@_b867x##)>k`CK<EkoKMI`1dw>cPM)Xp|b`U
z%>^)`1AGUSBVcZ)gIeD8Kz7|2-`)0`&Aa}K8+GuIRU{&;7SASF|9MVq$#tU}^L4GX
zO#rRM!}|oev?$OxsQ1e(DnxhvN)*2Fp$)`{0)=&7^`Z0oVc2SO;eE3Adixu2Vnq~c
zpe{tQkzi>-PKZ>;ssJ#d?miecQtU1_{Hm&pbo^5Q{Q8<F0PRTB5km}*=uV3h4Y|Go
zqF=vFg}aP-%sL^HL6{|L(89+nhPyu=^#4ypvf)plz8RFqrc>BZykyUX+@;})M?k3v
z;UzPO(?LTvv$MUT`)7%LBDZf4k9s543bvYhmG(o-BaScTIiQkI=!-_8+9RiRRx65p
zLeZ~4DrH3Ehf8!KSQNtZ+ax*<n%@6dH;WqI@R}^4TtM_&#Ht4^^JgcHVfgFyTelK_
z&-z-YC0(n?Ye`nO?d?MVHC)?kXcp+yz%`uYu0JwNO@?+YvJad^+z8;n)Gj+XX8gAV
zR?NCc8-zy1)y-`U(122xSyenV;=n9gAOA5#`uO)`sC8P_-NUe)=d<G0`!&_pg4Mk5
z0vt!+4>f$s#q`EVb|w97=Kt_!qWpHZ{D>!tc7GJXl=u2J((uawk(KH}5WKGIx2c;C
z5DY9uJg9S2Xg}+on*IGbOVnR;Pb9c}wirP?<6A3Z<e#h}M-e3kc`?EWB3duJzu)ix
zP$C9AuLF-5=`Z$B7BGUQM?i(k%$rwYtw?50XO)cKBksE)`KPd6pD9%%Y)_2S6qbO7
zvp4ck1hIz&H27^RI@INvS^#lC{Br0AlLL0UZvsiv5yBdynB|s_7zJkxDvO<_<R<rg
zBE13F`vmlq31E?b0!+~C^M@M5V((#(`90wsb+)x0=Yx728urC*a(Z>fRv^hHqo8mG
zS$i*#dEh^|x<<pqF7w*DA#6X0*9d5b(_2v1F5uSC0t42G1+?}Lf)}*?VDX)E(Dz0P
z3&cLbAeFQG#GN89<3Af(MHzfUFoH#E3prYI#&sg@`KwnlZ>4s<Z@{dKWHW`*9hNVW
zlLL3;&2WX?{K`rWARsW=N^dO><JKOwoJ8)>`c}^_t6$31GHVTEN_OtxLLRT)bF#;W
zm1_PduLn9LkFb^DoYz{8_4V~Nfg`ajV=F2q0o^DuYavBY+uak^DH68;Vnu5FGN0{R
zcnJEw33kUn$GQ>S9+6_K(tL!S7<YR&*nSV5pEnF18Bq53m;&~LjR-#Kl|SAEuj$*I
zyi^Aej$mek%mc3&2or@hGxSJWkEfVL^@L&dH<r8P__{8|Un&TBT07UgK{Er$T4I2e
zB60$<1qJ-)6YL1YC&*-xE~s>?LH$OfRMY+b60*eTH(uFPy&uU(?6Io4S>4UiuOB~D
zo8#xh50atgrHbW@K53j!p{}yXI3{p^OHLpm@meZ5Kdy$Qqfl<>eY`0Fh4KYi+-klG
z?HZQ=CtgsEoBXTuxTCVE&Th}onzjz&`2?zRv>Nd|_<=|Jm84!Y@VVL$?~9a1y)pg7
z%i+PnyrL6?B3OS0*2YrN)F8s=qPJdhQ*>AUfHp1O9N9w}<+iKBMAGls%dZ6p#U4<4
zZ>4o5K~XtQ<WhFy4TDeq)lqNG^RL9KFJUx#RKw3)ZGN}1sO5dJE9V}^lT)@<`X6K$
z1a7gjW=dvA@Hk8NE40hm7i>}|_`_XCzO#Gv?LLa>q2aSXo>EXyc!>=j^XS<UNY)<G
zA8m0;*${nQwIMs+jV(#h!|>9+P>M>RzvUdi{M4TR^Ckm+Jo7D$H4?WpzAl=zM#f`&
z=be@)fET_<_WfS-`CC~Q*i!MVVrlk#Grn^9%LGYr8{Q)njX{-%nG5VqtA!PHwVu+@
z#M9jBv3eA8&uHo4Y5V6(CtSi5HCnvXtOn)euV#*7wG2~2r<ct|J(Zj9@?aBA3WyJX
zyX-)`kHh3v&#Om;v2l4-muMF_PL?>h$~n-yKHxvZ6eU{D8)HlzHSx{Ejp3ZueP|bD
z%a6KJ@qeNY@(xXD=Otv|JH7NY>A?M*@qzYQR^C{sRPK~C%Un@Q9x!n1k*f-x=Vdwf
zg1$RT7G#{~7t&jsm%a+!%4Ur&<IcDi3gf0=N}NtIC&+i1`szuOxFYpUF7~J#ujW%8
zNw+F+v&Zb4m%<cW?_DMf;&Ku(l8LxQt-e8rk<9ADg7@ObU=`_>X@@R)rQ08X)u5;W
zMV3O=*`u+Ryb{({;)HUcnA?NxA=URI^09A!oJhKVy^+2X-@!=v>@BBDPDCm9mcD3L
z_HPdvdK~r-5YKj&9t6^X)Yv9NcAybYSskOg8`~}Z^PX=Hr0D$S0`TN@@xqr`$|rw!
z_-Ml6t*e48`-W(fT8+(l3NU=-fqWSXq5YOH09-tHJ%?#>x1%G>$uMM0rV-AGRcZYe
zmxP<A+UeDfeRa<-*^kvA!$5a2$S&i{(ervZCrTHtlB9cGg>P78)uO6H8ue<oXfd4o
z565J7MDC-=6yYgpyu~X}VeT)$J35dMA%?XN<|lzGzisSQGK?av?X`<%REml6`Dk7f
zFsVndo*#dGr!qkDWU6Jf(GcgpA<hVrxne?b<xBXToOJF7LRvg!ulKxF9#Ksq!E0On
zLdgEYA>yNXG^%)dTG<TDg3i2~-xp~uey(z0C^5#v-dvz;go96H!QxF0-hRD@945k-
zhn0yjPUpH?qQ0qrx2X)Py+U(OIqYRT-PSwI!KGUmGOO=;*LYb*E3X((@==BL9&K@L
zVXxg1z0g#lZuodNz11sxgE`<7*}jEqcZFiZs;@{WC49hjRXXpu&h!xRETsnX1zB~T
z<JCCMT8@vS)reCk!)U}sh{;dB!>x^EoTLbEDSl!{M0{_#-v?WUIY%|*KX~s;AQLRm
zNn4=dggv=({pwSJ2OekLp#G4pAWbI+Hpv3Nt<6QMoG(YjJ(RrhG#wZy+{Kwb<#v4a
zRt_Qka#VaSbeIpz8yUtI8N!!Q7yimKFk+ZCCiCGdVD~iYUu59DdS<HD>g7tS38x(2
z{aamT3fqj{+xFa&rf=9?_g{V}Y$=x>9$_Dn*x5=PkRGWsBq4lvE6hx*(W8|b)4j2g
z|F*E#G`n((x5|N=Xl@gq=*lV$Z**;XkX2(Q?V)9k*YrwWdgkW#{?2uY-GNRF4yMX{
z587)pW~gAM0I?vy#NGR}gbJ3HTOr0XR^uV*e%w1L;fvP|PC=br(~HlOYhYQ;eJ(Nn
z@Dv>@Ym>Ta-WR#$0t^FLuo~U7IIdG3iQ-RB);SR~tG$v_CLf;XJIp-y75_5XFODY9
z_(=ky5KL5_Ky(;jj<XUH>gNfm`5wo56qY9STYT%Es>`8m2E?lJM0T=HE}M4pR`gSX
z7G4h0i`<*s7#rA$3GC*%EBmM``YbWS4)_QAedLn5y1zPB68bl1T@q$<qpNdZDktP~
z5Y9fw**g$_{<$Q3<IG}a{=-(yPO6}VjQ%p3StcS`6P{p(-MVX*UA|O=Ea~L&S_@B4
z+2)AYTZuAY>@1F9Y?xLgoG7V?SprWy{jBk$RZib!EF|orr2I&HX<pP1p6kRg7lO3P
z?zPHS_Q2>kmIT;1)G2eh_EA>}#J)H1TP0VKW-s86*Dl6V60($#x7T7aQ{$9aLI$-N
z-VTIbyO?OPU-)>dlKLGO<3iosY!V!8@|Ip7PXO606GKpCiK(^0;<v7^Bv-rn8Na)0
z*_3B#M0fMOJ&#@zhY$e<p)!l9*f$rC<7hcXsD$`mG*P9PqHUD+Ieh43ZARYJla(wK
zYC;8?D%6IUGkV<`*;QK_Z=e6<$m3tIg=ZfOf|<<9NtN`NMWiEa9Z&tS-x*SS<F})x
z!NLqBNu)959gTVkjru%9Bi{<II)*GR%S1$mI=P!sQwhz47*q2-IvK!W5tpZ4j?eD6
zA&DVX=63D9ekRJCE1le0<S;uy2ODg}BONd2&z4WMZ(rf!=B+5WuW_}}j8^&dQsK-p
zhZ`O1CLd9rh4`efqWYbSl!s|}rZW{QdL0wpsEJu%5|xnCsmhTsczmW%YUVKEWeA`-
zNeyPFL?)3Ac3pu{58{^zhzhJ|<a_tCi3S;7e2~@DY>vAVP|16vBI(Wf1xX9Rf{OLA
zpP%2`R|WDujpcy#$4PWJ;db<0g1)CW#8t+ss<N7XdJkys$MRo2SN6p2Y+bhW-6^h@
zNmlaLpCsa74qDz;%|6fFYl#VZrydlgo_cPptm)GxGp&+@Qmz$=Q-)L+;AA!465@&P
zpIJaDWR|ot&ggX}D54zeXTNKaXCx;h`PTK=@O;R~T=1~LazU~-oy%q|BZEVQ32TQ<
zOmDedjd54&Gc#fdC&~IONeti@nYE94h{y)y0L!a)q<Q_oGnPNAzY9oW!QJ}fd3csB
zJ~S#12#1xgMAZ7kI)onkKnVxK+v?D^{pMrps?W6+v=?BvE)dXQU@i7Y6<E=qYQ%R)
zfm7SW%fUaac+hscf`_FXU+-C_(2ESJeVhEjo$g9D{L3<Tb~4|5UBJ_PByovjf-q<`
z)Ym?g-Gy#SjMj=*k2<7^?!w!b5;JFKN0~RwtOIc0?3L?h9DJ?XlQBF6QNmb$JUjb5
z3E}Y3pZRgA${2dL9`AzhUM}9-MjZ8yW+Q$!*R6MMgtBi?tZazc{F-pFC%t7tk8wJq
zV*SWWJVThVzQR$cb7c5gE}yMkzmejMk>tQ|Jb=-Jgz%W<O;^ORZ+m6Lrz&x`@4&vf
z@2TpHd>z?$@ox5j3)h#e6w?{*Td^@=1am8}YuTx0UYIVe5e%6SEjCh|x%-CPpzW>0
zLvR%5@N+FqJe$DtH1bWRR13+NAro4mZO5)}9!sr?<`6i&#339jql@9m+xThiH{~VG
zeEy0Nj<ezC)=ct<5f4E|Lz(Z__7m4=gOYuJ%p*r8WLFPtf)fo|>sSUy`{k{7Ii`O|
z<U2NLWRW0Wz#*HUk9DZxC~$x_G*!^HIqpxT&|s1lLL(Gl3VmvyAz|r11nI*=0fm3r
zsQ-KO|NjeG^glkNZcl3gzFzhJqDOZDJxWIb4nP|CwWm`au2?)$2YuCes620krvcLP
zGeVR5cP&AUP$fsAF|QE5h7Gq_Mx1`7<M`YDT+Jcq=Rud2F^WS4pcNm8WVpHcUbi8(
z{ZECXtAGHSEy6au{ECqOww%Fs%o&*D>mXuAl<L>3TgTS^50q#3)jXs`?EQ(*m5Vc|
zCjjik&}Z{~!-g?t%nLv(sQi0mh@BvIWzQmN^yTe+k1d8)F5H6P0%%YVPgwyJ;1})v
zyY6pu%sDZ%@-cbele)ST)WE#{bMhFQLTRgHmlEGf!>+HmxH#A(pwb+{3#eN4{pET0
z#2`caAG}HTicD!-TWS6g#xmh^sV5GgRDti1f-v@irD2Piwt>dQ+WG+^663kLy7G#O
zwtb&X>xQPJf}h{+VhZ_haK+<@t6FKCb?WAH-7GL{fXeU~703wbxb3%f`@*j;ec+z;
z%VLY`C1_AkKlnP(hMsA015bIM-5C%r&;e~_cfM=)_lz`OQGdL|4oYtp<Q4{L7pM$J
zCIP714CDfMoK>i_L6t)_JUqNQHkHh(qKRTB(3?bkeqb87|CrAagIt5b7jCCvVr%~o
zgyvwR#kC;5Y&WjyOJF5Hj35QE>1a*l>i}^&l-(CL$dF<lv3ih@1bsY^-T6TS(B!2?
zxgS{VP&{5_Y%Rcaf$<qc0>z*hp3OI@zhBSKV@qXQYz81;MzJ?{B+|~(?Z?OI2L%Pu
zGBe+BIL^Q@^>xG%nUPHt?JN}Sv5~UkjQ8{*c!=6DRO!HgfQNOjScFiZzU)uXu@GQa
zv}ckv|9%-2v$P9Q`Nl;XD5M8zq!CYOzl^^Nh20tSRS+23Ewy_AM$ZhG3CV!NEq$no
zI&mOgz)E#d6eUujTqvNRWat+lt2sn=12VtL9*hTqvEKLZTL6ea+zkIUhVQ$T0`LFe
z%%L`EA)mm{0oYgafL?)b8`1ic_khe)PsbYQL6R1j`W-+1t&tdUN5(;mjcmVS6a$w>
zs-m*Fz(BPHxLj+1b0J?lJVY|^-PuUf<c|3UIP6;((q^fjO;SBWLp360e~?x`f1i_A
zVlC+2ikbX$Zmp`FK7)`U<iaNHdus@MCzueNfs&M<`UePgX|N;!(8%3$GL)-`JO{B(
ztw1CID-MvIoD9Pt6FMXzz*{}$w{{FV{jrGMVhREh_Z5Y{^39sx8^`)gnc99J^wH&M
zaTl0q13(KP9#6Q`jwmc<7zHE(1H9$itly3m&;)KHSG`Dh8~1mAHSCJuA&LMDItB8a
z+mlSB{TfOufhmk07d-hCA}NEuA9>zY))m5c&-7W2?(bc^Hn?v7vjTi$8Oo)SWt1-k
zIWd-i6|shi?qy&gdKhBO0ha{@rMLjBjGQi<AYSMuh}Tw;u^^Bp2#Vbnb&S2&F3)tQ
zb;D%qK<H|!&(~`xF$leO#5DjQO`WDfvW$-Bk0el}L8L3eyi1@t5x*3eQTVGcE7nk&
zmG{KH<9p0HuWm%DK#DVGM4f<~DkOfx1KE}!tWQyxI>@yT02;(iL!1cv(H9kN?K6Op
ziZ14t#RK|+R5$>U0t;`{3*_?u{j=@~nL_Qxp5f;}8V~dkFzu+FVp^F%SRDvSkWvmI
z<fed*BNf-}x<#X(kjVB4lBinrI}zP)rYi}FK#;%zL6Y5wJ_T&bRz#&jJfz*9X}eJX
z2$gJJ>QKyj-E!>u8x$8M3Fs!$hahi@&!J|Jzx1EYd*IdSddpQ<ALKHW#U`%^PDlly
zRF|VDv<%W(zMV7jovWJ#!2T1?z;PNn3CILcP{nUiq8@E!JDb;cV?aW*fq+ruu|am|
z&;}^8oB&Z@4~L8`B*Mwj<Jmh=U>?olA@4$@=RhGchLBH&C>_X{#46PzC;%{Oi}|l(
zeZ9S}Kt_UgFn6%g+kC#k{(Ar&&%T-l`2G9op*2KH{JJ#6=)L+1EeQynN#fGXUTr$Q
z^B6EeWN^~3aH!$)i^Xras`{2|Rw*q2=<;Gs3R=Vl#6OqY^x+}l8Zw%=7V88j9?-~I
z%pxEaCJ7}cA<IBD-}NZ(N$NC2fwCW|Bwvw$eH{sf97ZbFPWi2$lJ?aZ#wTX<uZZP=
z=2g|(lXgV2D5nMdF1|hA_@WPc<BZ+)ZThVI`)&8WeD^oLSQ)kYR3Tx-%rcRn;VWxM
zwKx+;wh4i_%9RHnhs^y!+zj#niuuI$56{AFngLN@=oT2_J--m(O5{LhFi8KJ=J-1q
z@Z@#J=*8>F<y&al)3%f)i3i_?du?HH=+n@6rgCbAgR~bhYtjHFy~37{#P63+ps^z>
zA&IvfNuVz8lX)QUsmS!%kLb_E_xOUkVvY^+zj$OuE5N#tJ&Faq3p;UV$3FTlNy3eh
zIkqQe7miGQQz=avCzF()<j;=rOrW{NFT@!!=^@Ny*%Nv=!GG*YXs-IHI@9wRg{pq~
z9D?`>x7j2u*cjm1Gbk^GWMY=`DGn3jcZb!tS?gm?aFzMGp0Zm!@4-iB$y;!>n<|w@
zsyk!;?1H&gnRvqO{kYOH@n+2!rVHwR5d}lMkYsUS1Ul%Wm5q<{O8P?0z%_wfjc_-u
z?);&+JWgt<odR*Cw6V{LA3iKm5|FU5vH4RTmU+C1PaLH3fb1n{*uk?Gq?l>&j}tSQ
z>+L^s^{^@J(K|9U2gphTjd-6(>qSZ9yj-g4>rx69w&cgl@0}oG=lS8AIxDH>{9-#9
z<b4ZT4My6h(zB1?C(;KUknL^nNz=JzZ2V}&y+86nb0#fM;UfQ~jzycmtdIISx<u9_
z^t10YDI8Q|@2)=^yjRiAyU#SCar0($_W1$T`v#5L6#Hxfg+88WQV*>?R-85ds{B%4
z<lPJ-Mv<n+p*@OGeP=xki5}2#uwykivTZTtICG=4W_sgCKdv!;rL>Kos}c~K%bc*3
z;krmh+G5RgwW;TL+`L*IO)`(i^dwiPHtD02<$RaVb;dsly-^fXasyx22A4=*#Ln(L
zO=*2dO-xi#lQqlyB2grNi9)*=9a-!VQz;|gg9XFHCunm8gfw(G*tvy?2<b1sr0ywj
z$?SL+6^%Fd<)Qt@iYeRsZP}T6hJtnNygmwq*S+?4wBU&?Ym{+im1EaC^c77B4`vEy
zrC$1Si5)l0joswr3_m(ewXe>GE{i6kte`qfHXhsm?!a{Qw_!}_+uqNgf|)M+36ha6
z2+N#SQC7sR!SgH>Dm7$0vwc9bf7mKy{BDs1v)=ZF&c$ENnXcam*NvsIn>U&6J8M~K
z6$X+Fshd}*jW+X~+48@uBPB2Ab&tbT^trmq$&d<KI|Y@88l>`e`3~h7gi?=Y1q&6}
zDQwnU%B#in%hevT7hPDrKomv#UbOkPp;;7V(!E&au$^bhW1UoUWK10xjbtjWF1tec
zwDki=cytfjO;QD`aJ9NIX4!Op;`(tdT0?76T<qLF|M!%qFF%p!sO<g9nPu%(DyrD$
zxS(}~PWZ5FHdeGMQ1w)Jj^G_<PR_Xe_g{t|m{PxsqS)}XI1{YW{(^3%Iwhp)MutSG
z#dEex!WP~=GI`XiTAOahg@r7u1d@}kyaJo;8ScjKuSI`Zvv)W`*NV~b6%;r~%_)A(
zEJ=8zzK4Sorc>u@NgTFme^|J7=<(Y@Pr20-Cl;`g`zu77-&1%<J)s&{Gdk4m#Gqy_
z9MrgZMOo379IJn!HA{-Stn^)YNo?b$Q#pqsHGi0VZ};R&W<?oWa(x@a<q2^L(r3+X
zwj5@I{Ui>`X5<xuuhkP9*V-1zE;C&fI6PghI?C}x@Xpt^^Z2Q=8fuu~QbNW8x)g`3
zuI?%B`vXmq#`#-481CwhS1yb%?g}rpIogPc$m(o|er@k6el9MJ+YDBo^0v<6`i6g5
zO35m_ZSG0TS7rP>>C>07YjfcpWW)4_i9)g36m+BCUfmEa;7jg*p!_;<;e<~Gl_`DH
zrnj>(5$24nX`mvNhT2Jj0gQKL6qoUNjH|R1NmHg?))o9?mx(Utxy`Bb)nzL>UU&mD
zl(Asw>0np?MG?YtN@nE0KHi*4<TzX*O3y`IDK<G)y3@AZxH<pks7*Ib+e`fV#ga<>
zlY)H1Za41|C|a~Jbn5Gte873syhv>0ptZ;B;>_}2nQ&F(UlWnFCI7X_@3U!2Fpx86
zS)v%Be%539o%{!D$NcQ@__$nk9ir=Emsqz6_0$)wY-;Y0EnLkA@x)IVI_cn>I><(L
z?}(=z`!K%;$BZ1G*m)S`V22L)rX>mrD#LHwY{`C^CfrTSY9U#TQL-ohrI7G3Q&_{&
zkcTph1TPFHsy}u9Z0fDtp?H3h%Q5NNR&AX3ABN^;vba%W^3mH6qj)dMCKT$4abUf!
zZ!=s*Na`#7s(bSMS@lGd{H)k}T78dWHt}_|fAHx|miOKeSqOi9+fcuj@k|0^*p29c
zkg4>mY=f4@C!;1fRm`c(k|?vCru4swos4=M(m&fzzhI+mPIWS>CxVM#7~tu+oLD(p
zo?wOc7m|fHzT9~&`0BWrC9d-cjn)(;=8SrxV0Qu?7lSEPl)jR!av4wI%R{1yB1I}@
z^jmduUQv-wKBa|fwq&_N?qQGem|nACP38}f!sN~|yfFRPM4^Os>B>l?SxJ8s)-71&
z-7^LS(bBGvpF)Ld@MIbb9N825X+H4KzZWIOU9IAal}oBB^vja_d?@u+yX|quOs3p~
z2cA@#Di>c1W_NY>EmuX}CQ!7Aan{d??l0C8Pd3a5SL#qSV7*F#(e_u5r6SoUJo)`5
zc@aes$4wvcy9D*OoGO%_F!+#U(1<>)|G|3{){_}|lR~J5p#^85#>3}Ylf~C|H+&nP
z86Kn?kUc$Wf84vZ#U%IC;rgkFE&k98Ifo@SS%{iQ>lY)PF7OYtmShic2<|iz1Z1m{
zeiNlBSIR#@XhxXlkix8)A@}+*2_3&1<FrX}s9=uv#VFXnb;Z67v#W<XPl$)Q*xP6L
z(I(Mg7>X<m!&Muu$a(ZTym^w_7pF59sz@Vn?V)jOZIY=dIUTPXk#fp&l9I(?`K)KB
z4o5gdPvu9k-1IhVW#K2O5M@(kXvMIl<qC;D@0A;jC{1y?;H(<-G^s_CLRKMWOh3c%
zcos>9^BI84hXV){xfT?ZZONyrgWf%S5lp9QcaL?mG&V@jzRcTW*e3h5FiDeoO!~79
zX~XQqpiEkVf#bOw*Gi3vx}DlR==6UvE+-H~D9n*7g|wIRy|QD^^HH=(NiXk^*0&hE
z*WAHxBI~IfW2%`mVcA~IEK$Mo!>%bhgT6%JDkrS139EKbzZ<8=^k$m`gCa_63hRRH
zAFjCbzLVm*tgREheV{KsWbtrf{GIC8CWh?FX0WP=p7O_tvI#4c9iO+&8Ty=WcO6$z
zwfH%p{fbx#PPp?&(n_#&8BaR>HXn{MD^$3=SCW0P0(11`j_*###`iJrH{_DL9FB}+
zUuK3}{O0Ro<wi!XJi2)L^4P()uzm}9J1We9KKHs>xC-q2;0;z$SZe$JT0`ZcqEl#M
z@U!g06@pLC*?a4Tyt`*hrJ-=~xx!TH00YIEb6uQvFJ*<`$wl*LS%;ex3Wlt%Dt;6T
zt=F{IU@T}CWya~hF7Nd%dbxG&NxaRM0T<(-O6Am$uhveb_c|uI?(-z#`md@}*cqub
zVDu7no@KLKJ11-3C+CG}ituIsRHvNbC^to4WJf$;G{C6>?|eE_&h8OXSIa6XbrQ|`
zWZd^V1BM*Pz|kL3F~EmD7Ew(u0LMMq2}|b<5xV1pMnbAuwCmuD2ilPfNboIk{UsCy
z)jK~Hhrk@|*}x*eASoz2ln;aDwZri|;bY=IK!Arl<ezd*f=wX?Fdk$Q4)zz3$Bf<Y
zpo=wN9XFvmi0qse=Cp4lI)qPo@c_%c|NLv$TNqf_t$Yc}Q^&r8E+ZMd2Bz@rA=Mbl
z*sn}XT3nYY_ai0}RDaHZhxCCA0!b>s6ru^rg10Ci5(>5&^ps$rAO?w;$Qz?(?&%}o
zuF$2Vq~I}vw><LrgF14;gBihNNX>$97;ss`>%vRm%N8?J{_{i^AGisNo0Ws!zYMA)
zokGKQsN(w#+v=;Tw$9vt_G{1Q5S6b`UL$=M%Fskq7_ftbhXGQ%(e2R~)Sri>O=^G0
zBqjDo?uyM$7<Vg<sE^3N9SAP7kazhJ)FYD+HcbLbj;Wx87`%kP1LJ;>fl25G_^vb}
zL(39?0ZYN|q70nh2#{R~j8D>S?+K*40}J+I@H*gww{iEx6FnZ`>sz(hpYOr7{NuJH
zc<JzjO`-T_061<V%lV1!4fmHoH)@lYygTcWX``5!y&|ge_k~QR^~tGKA<ca|@KWR<
zH4~~-yBleMD*-~ebzbtOygV_oCjs6C2Hf}l%VPoW2n1IlbVQLW)qRo#8-83(IEioq
zSYF7a59R>gV$r$EuzUfI|Gs+o%(385ib(trdZCBv8FGBQywV_KROABtl#Sc$Cg4zb
z0dOOfGYcgwGh%+AnTPN|L6w@1K@5YyvPfKfgkr5BNOsq#mFAzX5i0EY!XhynmFi1i
z9spd!vpm=98U2wu#1Q!Yel7$kWf>tCp#BP2e1V84d)Wa0wa`ITCUQGm^ug+H>mHBe
z{`=U?*x|VoL*Qol3c1ibgDPdh!`NjTWrBhi;=0Hw6nTn#^HGj6QZa(;0_1mZkdFog
z*?I!8W0gExjXP`1IWg<+JE!z}1j8{3K9zgV0y?He33lhaf=V;cZ}{n+tQ`0CLh>>I
zBoasNo@k_HJ-B-C_d>e(K;=fp0Gk28R0d#iNi@RLnr>XQYZQ#oN|^AYCff`FPJajJ
zqCGbQQy{&R?*$P2jhSqNz~Y0U*6s?+VhG9K-?`N(7tVGboGHJ7E2t7j6h4xYptpjo
zNq;t=K=jY|+UPNVfA8e4(%%0_{-7k^`8B?q;#v1r;Z({cnEOZ|2yx(Z*~(M*Vcs!Z
zu$KP2?7WK|TpyF|>0nU$>bbfvR5*Qb7EwGKH=<>;eO3S1iGxkWVhdoZ_xnwF7uQY2
ztOcAvu^JX2?%~fLO8w0RpgQ4o5NS!#S_e9lXG7G<-#>PMMflxF7~A=sZ&M<k@<;=^
z`?=Y>r@*Va(Llnjk2V082uzw(piv=r5@WAr*n%8ic5CDUdr04L1bjfp-@E=Z@Ksxt
zGNrH~3hp`-On-j4!jaa!w@^{IeFPvf0Oi!#XnvP0#hHb?a$=0u;LJ+IFA;gCUk2n(
z5PI$9*8+v-J469@!p^c&x!thvJP#$Q2wNQe`_#?X*WiOdGgI|_1t?gLkq-_EHE-Ow
zfCstKkQ~BcWMq^i9J9ijrR9=Rf}CF=a3J*S*IOujS#mR14L?*mQRAU#7zzj3`&ig7
zJtpoDqxkHwDZ%j#O9ZEa=jW>D2qLQjxk$uk=I<lyU!}mYZ3FAF1w^OcE&a*hvk~YC
zq3&3f$m&CPfs}+Ckbtju3TcaBDa%?}okDyzz?bySeRG3mLn)v?_{fL}L46|o>=n84
zD;Sm<va$k6s5$bt21ZAJjagj%`<m^&ALnc2f&i}S)z|uZN!yCymx<2~5n#yhhU{Ze
z$OCe^LN=k^U(3>_5&9xv`DOw%v(`jy`4w7u+MvgzXha%YaPe(NoGSzqCT2T|MgFtx
z*Y3he1ePK%*HCIbc#7QftW&Kso{f-a{s>7Q!h7Mt%9CEFRQfD$+K(n;U|)sYZv=k=
zWQqr!3VY#~W1bjLEnQe^rY^6j*!x8&!$p}xHj)h;LtTe0{#hvjRv7{2h$)67zOY9j
zLo-B`L_Y$!)=k_sD~H!?^|u*YW$Z#LVj-^B33<d6@9u0PfD*w@NM8ja(<ZQ9v9_~g
zRiMHHW^k?+H_8eDM<xiuNOJ-lb*<xcQ4bFHRdPo0@38(9a(sDnm!YDv67eBRVJ;eb
zOdr8RW<Hq;QNq_I$LtVY5Oonak3K>HJqVx#bJrBGn(7@M?f^ezZS<m%3ftKCyTzWW
z4-Ae(>etB8zKRsu!sFg9q=5u@NX#fl8}zMyS912>K~I4luRy?K{nO0r^#tV^BekCF
z@*76QZkk-<bI=U@1k+56nDHha>}iDPX+ZmEiky7l8UlwJ`S=z(={adg_dq)^8=B59
zdpn8FRybp=U|TS9Oj}-FM*Ptg?XQ35ar3x!Mvei&2iHE8rHTRLuK8G4qaqOJv5~Df
zGQI@nHnYHvCZfMn;)YYnVPs^?)Z!MZ8n6{Y+xH#>iiB=UFegx30}l<(p4TzD5Wr1E
zc;HNuq>s21XxR`TIMA>oWJ-JEF<3cn{Q7qS*ks5@8bO9nGQ>qAu`@cUq*-a*0RGm@
z)k2=PBG*oi2Xj6|nLFqfhM40{jKml+eY?-lAww={=!OEbCIPm2gv-OZPhPlKDB|V{
zi*6DrV_~vq85fm7KZD<C@)BaV&V1=)QvmC!31ESLe+116jdd-hKfCIKL)Gw7i~#3G
z7ej=~cR3(QWCCEPNY5+DE5kp&1>IPqY7qa|Adesrklw*10`V~rDZ}Bu6pj43kUshN
z&g%Vgm|<YutC6dPnkWj9PbxSL5WEva-pk0fX)k$Z33Gf3R+>#1t&35WhcYH)fAinB
zC7B6_`Sf$`Jm_wX5sl1tCAGp|#@=75qqa#D&JobJyMujU1LjJz09g~#XLR1E+tiT@
zXYK+Hvg+Bm4CpU{W!wg=;-jrjdt}Skfd|op_v#~&r1yAESy>c1tMM59ey%|mvf)Wy
zo?U>ebpgBd-`mO>O9h;4pTQgXhoR#a;8=<rTlbU7H?V@uNie{iiHQjemBSILt&w|g
zWgAhF(Ix=0?zrE6*vOz784T~A0L0t^xga6OD)?2inrT01;Ha|rGS>@)TB=sQ)LH!h
zajx!~pfP|UYDY4EIFy8tQ92&FY9X#@JviYxl<u!rPs|n#D;ujsdq3)@!tRIqBIxH}
zZ|j5`YTN7v(x@VPVHCy#rp+NpppS|HSXE5^1K0gh#>R}%Hh{dpSi>*8M#wzaCB*ZZ
z5`al@PC-Ed-HCl;>+|vd%;hG#j0}lW$u|j*^Az&4{`O&3vUOH+OAI=C<2Q$Ga*zd|
zV~3!y{DXpnkL$pi0uQN14dJ?l_HV;mLl&S3`8N3oNs<R`^9lwbD@PA}wDzo>V5djX
zJ`ni?&U<^UAGFIDLyptTTrotTC%|(?!tsH0;y|f~K>JyauC%{*D|bJKFIV{ekzE{g
z2%%ez7kJS$imgNc<=78nMJ{lk^nk}h-*OOS$D+`Os75zT2s9J@(j$vM2=x+*m`Bfp
zUK6ko&@Big;vrok<H-VVMFWt%0~%P=MgRgo()W7QRp7u~1Ge~EF3)G+EFE_~Bm8Hb
z9asy;zySc}+NTSv?}GJ0e0_aAId1v|T=vI$?-W#^Zs{a+Q2dZ(y^Vtdy8@Nrz^DoS
zPjsjwc`J-^*q+Q4%(pdio8e~M)$p^k0f{ou(Mr7^AH=<i8zjll-W3A&{}j@N>`5A!
zJvmW73GJp}D(@(nvL0%~O+$)|=27evvSnZ4R9mTx$tLY!mkkQhripo^9h*$p+Bjo;
zXW$(v3s<nvrx;n|nvjhV&KB+U8`6G;*~vH-+D@7ni&`h}(mPcWl>gOQSF-+{^@$dZ
zQWBy^9If$e?EMsO1@TKX09m~VJ^(q@`b`^h5oL+uTt&_M@Z3900rzXS%esLAbt}2j
zRcMv%*n}-Y2JQ;*KX5uxz@Lnx4dE{T?jc8bhOAY{GlMECQ|@+l=wP;aw33dO^|EHa
z<mG8f{mv64`#7Gqv`236TW7gjSPT7<NtjL`!0Mlj;vIHOzihP{ZE3;(=HPUU`>^Ta
zPxlMctDgOUC=o1ww7oA|+y3}Y(xe3&M^>j@`ZJ#$ErZ}9qo7IdK9{M%F1i*zqrB?;
zoFw_a3L4j-e)=EpxPw(}wn*(ix=m!%<dw(MeJ3ej^nXm_zovSpCp+Xl{>H^PX5F<H
zs=3^<49rAIqpz!#&jTb<bO>;~!W$vac|`FsDRRa7rvpo3D>kV{1BJ0(4)EM!%q9wM
z<ceI_nPCt34s-(;i;6Uvs=FD25eg69-5V9JvV5p%9=Y}{AZ<00Yh!gn$=)bgx=QoJ
zRSKb`NKU1r_xapi=~Rl|QoalVs9T7Hjx<ltu0#)~qom9eX6Bt8KygamoWY}L<X)uB
z1+L1b(@(CF#fzt_I`!TYZc$GZKBtzH)N00+^f*1j*}S;?rN@FW3*+}il^}(J^f~i+
z*GG7IJVgWBH=}3XdmfTdZPY7@p~$&%-%Um~C(bl0m;W(t&ei!#rZyw}a*umMige8v
zjw7GxeN*oqd)&)he{25cL07MWeuujOC$W@6(L(LK5h{Re<no!<k(p8z={CG?+c`t2
zAE5IP0rg$Q7M6lGmNXIy3ECg=Bzh_HUgM`qo93H}7OIox;fdbK9<j5~$_~t`V3vF7
zp{k70;kahJ#rQd^Ia2@7@zctt$3_*sbQ`1I4*;yJ)RgzE1u&EVIljUknYUe(iSf4d
z8exm@&Krz5DuRNC)M-+9>Tl77oX#(ZrJp_}L~F-8-{$k`wMpHFH+edVG1I;{En<%f
z-iNA1P7Xh&_yORIFIP8~lOxYz-HID{UKiqM^U7@g)&eJwtW(f{QTWq4H$)}~mnZDv
zL{**<$eI#j&S2@iTQQ>fy++o=CYJ@J!@1L=U+XLR8Cp8u5=>DIQ&5c%=VDHL`gLC6
z^p}<<=zZiEoN#rvra8>zvD}xY6F<x!oBuU8o8{n?wrdTcqD`B$M*775FPCIR6$Qo|
ztrUA&9_uz~QmBnkDKVvy>~GeivU|qCQl#z0Wm!r_EF@lcl5Tn3^Frfr@ui&W&Ds>H
zIOPsks^NQg%rkd}<$hJ*BI5yVR5K`SVKJqmuJ$ea=`hA0WJ_OKsc&YO9&l7jUWjDA
zfTB0WG3PmhTj`tySL>!6U0<S5%qvSD{o5(g8f@whv(|Ps^s?yhi?SXnr8`nSVRQe;
zrmy#Mq?@;XGLOTJRKQ|SofZ1bFl>=O9+n|!XB#@D6)C=KoGQIdGs<I^K?R_ou^Nux
z@uv5>aiesd@+)%d+s?ImMharj!7xD2k{IJ5yuiZ8mAA03e48*f8dq$aX^0onRVBbD
zLK<ksTUeYeeC4`NJkv;}Nv)<y^g3^&bIR0__wv?e3ksKp^h;CKyt1Wj$t&2N(%)y}
zU|%5waPLJh9<2paDK0CUXINT0G}?*haz<Nd$m4S_uB{vU=P>aPXV{)3nj)1|542^}
z2wND8aQ>y^-O8$n5idJKQ^#|y6<~4r9s!%Aqk*@$ol*(}jNXJvSPt`_Im5E)OssSv
z9OhUi?yBmdYc-+g6xCk4g`!A0Zs6j5-H9VSxqM&lTU5Sr>wC^$&J`YvwW!e0<5|{;
zQkN_#{S?!rx5zi#$!m?#nddKx9heY1t8epE@YQWYn$T`~dW%7`Tn?6!q=`9yX)kl8
z29K4EU;am>?q>)#Ig$I@((`B9nOAh*)Mw>a$dEK4VLYBiV@ic7vCaxsQ8d%QR$q4R
zXws!W_$5jF^rVclxa@u1C&KSm)?$<-LV0nMw)Z(5`nl}R<f_c)#g45HXG$k?bnV_(
z$gsbZYx6mS{{AJKH-i_6jbu(=KGk_sALCI_r0%$ZiKh!;&oSJg*(`n`38P(RAc84>
zt6oOZJ)l!kTN@iP#V0wgaJxs<deU;p*k&6r(W6Sy99mC3%b1V-^iLj{leI8B6#BZC
zUiUm1?lfoCgTDP3hq9tx>%6`h$q>FLOd6Og+8N38Jg^d-B%(Dkv6Q0B9_OUf6Q}$1
zmE<e0ygGO@M%J<Jl6-pO{v7GN^|`9U0j$3QYZ-XO`#MYvO+7r$d~V|#$l6E0HBfnC
zoJRoqJNXsO9HxdnGQI6)C1h+zJQvwt3u12@=2wU+QdfuW=yK(L<R+H?Rik1Mx8+F{
zahpIc^aF#UO=x=cu;~5aCYj!JZ@k!H{w875@^!jO64DkurpJ~q9w^_<Ns%)!JINEP
z-a$-AH_iJ=+oS6XKgWQaIaSbWZl^@SjpTTm0vdfA$fkGm0dzipN@T%RBZo$pD^Ppg
zw@388anDj`|GNkM^pS!NA5Q(mx*d#uv|>)TUuS@LuI-5Rdwjki3MSq_9=$96Yq^#W
z-(Og?%oUf+$nR8ofn6xSsJ^X5)?XgL|FEB4!_L}?pmKG1eqCSh^H!O=i5Q`jQZ%;9
zOL6Sr0mHOc#ZE=1Vb+WH<=?h@9=hE!XMfPh_q5pZ$J22%0dnbAKP8hjse*9-CH{1#
zsv7?#GY!M#o)`!A_Dr&%&`R9<w^x4M=pT~acJD|($DEzqPZwrO^Gl8)PaW2p@=3XA
zeL9Bns9xq06^4QLPHRkY=9#JeeFt>Um%pfyqob?w+qx$k{hC|sT+=h<gBX4XbEdSI
z7~iK6b`p%Z2Pzp$<>t=i-^BFmfCbrMlLGs)VXI6AbTWSkScs=EicKFDmE4T@T03WM
z(@RJCT)kUK_~{L5{*J5xrq}N&>ZkY_h2IXn@NItaViPMq&P+p_uWQ{tJy^pVh_A`o
z@{S>nmY&Nc4EM%1<&9`n+auF+)8%i!DHCf`mpsMYcQRb}m3Fi^Airp7kb5-~TW^u{
z(BhLpKUPywXnQnQL1Dbd*uL27<wvJ++mZ39&|jPdf?>G$2;bGD?Q8n;mM`siyS!9M
zzYGKobYdocT(WtcIcd#7loGmOO&+0gKt-9Ie68Y9k&%81BXnyNr=IjYUmF{w_nc%g
zJ}2_S0Q#urRJZICF3O439z8V4*~M7&NZtKsJH>|U^d{El<nzNEe1%hj?`-JjyF^1*
z2H*zq1UI*Bm&F;VP4oP_kt?*YZW4&QO8aLERnRCjzVPP=z^cFK4XOzhR_1zNEh6%k
zk!~4g^+jsUum!m3nvra>ofHne^;S=IV!xwTC?F=c0Hr{%L<wLaK(yPuE!DmTxECKH
zC8A<M&)(bA+2|OC40+5z<$wY?@w`#FhislfpbT1A2D>xxCxlrd5#^L3XV{PHo#(8Q
z94he<5eIyB!Kelarj_E(Ixr77F(SkGCneaS7Y?hl9i?a6{uXI0J7V^(kC(tLZVJdu
z6B!T-H0~yCcgFPpy{$nX(Etc|9OuJ5_Lr~&Js>-K)U3c`#uK1K7Ke%t$!R}1*Ua#Z
z(SGpSkj493vo=fRKo#wg+6s@p`*|OC)`7eJP`#7_c#kq<!ShRW@&T(mSX5R2T<=8c
zJv2ba3~%B?+k^#5fV0Ra9}M{p1JIsPRz~d)T(4M&2ZJ^S<Av@iAn0O1qL~Z^_{s_V
zgiq*B2w>YZxdQ)`Jk18fbI*E_5EPV#yM>Id0GoX9-snQM+Q^#^G>(Y3rJna5k=L9U
z%TT~Dk~Kmg4nZ+A4)FHyOVJzoA0M}yivk5!Z1?BJ9zzR?`ES4sd<_s=^((%AYh;$4
z$WUg5N)Rvt3+pVBzXJ9aHtVrix0ABsZIy9o7`T|a_3K9vl??E50Z>yGDiMXM+u`(~
zP7Q>z84v-)7DKfSa<~F;J__nO!L++WqoXQ7(JcNb{tw)I@uv_26B{^n(LtMHp9zRa
z7wf*N9$Q-hbOZdS&5@56a4P|p1o{`M-qfUi36@!P%W<_pkLK04wW)T?GHHz92p|IB
zl=SUUYs^IapE11Wo`a@1;!rOVr$JgZ@|QwBRw%9*5#~TdjC|Tosc7D){zwFMSK7=E
zGx+tEKvP@s6#sWRRKav`AVrl})G4(Y#DN4SJvL<dgb1$4t@^!f402b5f6VK;_3ZB4
zmh@El3)ekp_~b(#>2~{e*?S!WM2swUo!3A^s4Ds09EM8t?g{tK?khL{Mb}2g2EZ*3
zXu4{Eg23XdIdgy@gip`TB9!lM8Cud_w<1#tR#g1AKYt*vv@00D-~Z3Y`pF%Ca3#8;
zvIh6hxKOTCZ27R!{lpXqa2z`Ncf7u>ewW_gz7z59V@x*9?&BL>KMtI0DKxi%Zu)II
zh320Jh%jdxdfAL{cSHk6SQ9F&k?`AZ(*^v(HFC?xDgnW9d~Qzq=f@sYoscOFEkAqU
zYn*jK?!Y2a5g-#ggz0_R-0U41`d+zR_g@@+Z28kaKY3gFt~tZ}GJ(0HcXz-c$u@+d
zuJ~@9M@d@1EZwb46GJgmAP^oHsRZ$I!w<1PAA4Z!bA(X4@H^mV4>~sd8oR~)X&1U2
ze2_-~HsP&+SzrMX_^G=!^6uU1dnDAj?ZwZZeMO?q{+$auq$H7b<jAER4|-g(z`}*8
z<lec3kC>5$XNen;p(9KXF+j{4Awog5K*wu^>YraQb{KvEsPEVm{IOpYvS{!M7a?T)
ze(m%e0H=Iu-s=izK7eAs23BF4IX^;!k-j?9=EvAS0QALLRw_0x>H?v{5xwUQU`+t6
zF{ufFitr~;rWp`PKLN*fc)|sEt0>I@k#$i<GdP7tKwtBISQrJu{z(I0dnF$FJ3u9C
zJGsX`SK)-^9~K}LH4-T!o+`9Kx)8zm?g<kV$p%T@sIQC65uqPI4h|qrjQ3%)g>Wns
z%LVvF4k-M=@$BoEm!q-IrZ$9Sult&-|DBrl_xutg%olM-1+aoSlJ#QJL^mNR9eLp)
zIxuQ^SiXsW1)&U(GI4GXMM5cN!A}oh$ik-ZdSE9)6k3E4f<ghO4<vnL3&Hj;<cGg?
zh=DEdy)CjdL)_KIw0&m!c8skoDw@#fkwotEhVHu{DnPP<DKJMh1yB$58h--y^;Bov
zxmh?Y@X$__31;aQow4Z^u;ZtGEbmjrK-k#?yk#5a-O$#k&9=YuLvnB@J=*{h|8DlF
zh&aKP9NA^*0ecV_+L?gGXb7DEglfTNfg<nFYzI%(zIX2~!rlm5H*?^{J4pHt_OBva
zH4wtyGd8xzLkC8$GJH<}1|`zKiV<KyXo{D?z5~SZpLYSVs)wuxw~Vg8&(TcWC)IzY
z6!pTC&efHRJ+%X&L%L2La#<h<RX0MsaYW2VnUZ^UvjD|cLvsT4X5&SjPD1k?+2k)S
zi^O%V!7f<_4tRKBVl%YJ-5+6#(6ta<somHy^Z#el@?QNVyT5%>l*k0+E1#y0?~0<F
zM?65?vaWKj!iFm}?gikp6HIts%Qv*C9y4^v?E#qbkoH24C(r1#AEgnZXl^`MS0)f4
z2it*AfH~T;Q73)~PIde%k^2E-{&&zbCAC0Ha@{upwbsC6q_Aof`fAW?MUo0XaLIV4
zlNZ7u5{3?O)GGlN`jZl!K}5-RLAn>P-zY;C0RTOtG`@W|q3==xqX9VjKrx)HC|@uQ
zrRxD4UJC7-Cnet+_IB097uLVh#bf@#QpqK0>>>-2u<>jpuxZ1+$g^%B4Os5TixgLA
z)Cn^H8HR#uhZUr@qqfmrBp|r5SfXw*fRlUk6}sHeu_>zwM{5l2?;h*sC}R@&f`ss)
z97JhW%E}wf!!P{f^M(H2q1M7b%JThe*FV_!Wxh`gxCyHlifSA{69cm#dTpxJ1y&iv
zp7ZD%7}P*l!5<sl3GMqeh+oIYkfCM`B2@#&1VQ2GqY*R<{-=omZ_e-m_aFi023Ym3
z`0Z}Hr?$ckWpm8R_@8C%e0(m#97|h{TU4>|h82knL{Yp&Jlw(5BsOUdN&VzShp-Ig
zO|Tllzlma9VB~>u&<zjl0Y-uzh>8G01=QknhZlqBAwXV1)}V~AQ&xXYQ)>mM2=L0S
zZ-U40LK70XO~&(K#x`6|{9OXVb%(Sb+mDK0U%acWZx_OcY>cf7hJeIRC*diGEPBwa
zf~yBePfB62T`kZdxr>Lw>#st50<=fOKHj5-hhze03*zdP`fVhifc!oN^bagy!?Oby
zc3wPTY=n!*cj!B`!7XLgXSYs@?w{!(k+Jl`<Izfn2vN@1+<V>Y(-81Nkf0m{<|ByY
zzuM*KyO&Q2`!Tv4qQ+2R?Yd$Z<WGDZ_w&O;M+y21+6|%ErNr2ZG&JavrRKAs3C-xm
z<N||xy8-3(gCLU`f(El#4gVwG)+wQj&xK8s-}3iX=l77Dc1XE!_6U{dJMMbtHm<my
z>@PXmCzE(-b@eov%(W^56<%>Z6yMjFXL>GIC}drKaOil`(cni1^IgfbdE6{`^*9*9
z!?f1Qk8R{+9uG8g6TFg?l+!}OndPPvxbMrreW!}1czXuIQzu&BS6XkZ)3{2@Y%9Ax
z+^^H8_Qoz-+1R6;D{a5hJyV4X(p<Ukma9r)`s9qoE+kQuiqk(mN6#_k4wd1Jq9cK?
zbty&`h3^K;IR54W2rCI{ll%3voXZZlfZL+Fv#9=v)!8_xjZK*=#58W{lAJ-DOw77_
zdC-8Joz2$4CEAcy`LYh2o|bQ@(0iGX3Uk9l*Iz2Px$&RrYe_7B?b1}Fxmw!H@k%}M
zjLT^BJMqp%A3Zy|eBHUXRehhQJt}I0e9l`%>MK1oy{s`=lRhw1fe{~B&<VQYe)fOS
z_LgB)g<HEYh++W>A}&fPm$Y;#2+}1b0!nuyAV`Pkk}fIfE)i*^1nH2L7En?IL`fCV
zGamPg`@O#JIzP{kz1_NXGUt41j(gm(Y(ZC`s>p~vlJ&`TBlJ?EWNBg~e=hDBU-z8E
zt&yLOF<2!N!j-cz>i5h;0OuwT?r$l29-DuW^frs`8dtr`p(k|$8(mrLH+5)<)o!d*
zxVuw4_l^~9vsLVtd-_R`|JK9(qM|Kckx7MT@g_-cF&jcR<=sVD)C9TD)C<s%XSoob
zQf|c(J>fF^yw%qH%jyHaNV#;EV@l8uHQTeHMv9t-kqJvrd^|GgW!yQGc11ea;>D;-
z87OlFIo0kS64;zL9wYg{tmENLySL8G-N19C9L@S<hR<B1gY{UeuuO{%VxA22&u+S_
z&dANep&#y_uawZ{7OptZGiwvS7VvF8rddJ5gzkCivq|Ytv!Lf?JhW}v-RAn#9Ivvq
zaq<{G`WKJ9w?-b1m3!MLV2!z+b5Hc}Djdny_D~9o{YDq0{B-{QpqP^n1JQnNZ+Yy(
z)~(0=3R$tnO2ukghuZTJaYUE+M{W|zKgSnmYlUKZ;{Davt}J;j0M!BH-c`ZVYeQgY
zr*5t+Bxgzn&&?G?srAXw!=7gD9?4PZG-CIfn&gkKWCpsUoq24jLv!@p-#!1hY)O1M
z=9X%RB%Vl_wC<X!3PYD#3J?#4&KJ7^5D##+W1Wl+5!=nVyYrQ=#&K3TrP_e(Disrl
z<m8F8^rqY?ea0L7V7M;x`j_H5CBPzEvZbsZR8nYNVZxfpv@>}vl;zBHx(Y<JMwbWg
zyvyWZ-yl}W*^g05@qZQiLJt_XEX9Kc)b?N6{~}p!-xAak8hdL3Bw~3CO@1n+bmTH&
zWL9V>Fjb_IR?rP;!DbAx4HlH&Tk*?u2`nwjzJVWa@k9((nipS#@R{S5<-x&aSM46_
zkQLjdMnqU#;szu!$1D}&_J|{RDtqLERXpcQIem9#Lo<=L&xZM#u+R?Zane)T6e*Z^
zIBfN_A8Sc&GQ~iWc+v4%A4wf<{S~l%Ll$044UXz12{x!aYZD#Q4#-MbZ1@w$d3~fC
zURmx-mGnuQ96bx2sF+VdH~=^D3{2}raWbV0=S?-e{ro+iwq!Gtw^5z&f%gJ*`^gGr
z)94xh0AuCbiz<>Y0Lv%<6rRex%r(GEc;XG;N(^8XKb{tQQQPUKzN0F%8v7n)KVQG*
z;}0<ER12`2Ve(lMhYyDR#EcfS(HHP%W2IS!izW-KRvEu~i3Bhl5i|^WYk4~+BEsC6
z=U#3wy`4_Znd>&%b}MzN2E^hjc7f{J7w0_kqxLCnI@~D_jb2*$4jqEThX6pnnxGd5
ztC_PrD!VK*)0VhsA=-YCA6GY%Ut3#o!>p)x&`$K~NrJADF_KMVZ{>0he9KMhB@5<4
zE}G#Eq4R{@W4fMmy3FynUVf9z%HkVJ7P#%We5noZd6(H3mR(*E^Z|4QB?9cw5g6qh
zW-=ABtr%Be3HFIAwq?^<Gbw>-j+>)e$PX_CQas>5Z%%%#P_!lc`MDV;*7_^ZkZ_!w
zgYJu)Lb(7eG8#jjf}^rF!9Env`GVx-l@t=%EOUUdJ?BJg8$}uR7~2|YT#yi0qkOl&
zD40<D&Skam&#xvVBChsa<+j?qdWCWF?1>ONdtYZ!XPKY1K=$^aJfZKG#aAj{(BhKG
zXZ7XUl379mwe(}HMr}ovt5ODI9lU^&Tgyv{1|^!EPXA;<nYO}|MSZI3w%!FpA$#5C
z=Noe!Hfc4=+PBs(a<fQZR<tG#P<xb7qA;>us$J+9SQ_^(+f+9#GtI8foC96}bCL1$
zR_FIo&QZnj2K-K9BeC*$(u`qoGM;Hvl5N`^Li$-ZbR-A*cqSa^mrTFLNDfQEJ5q3r
z<qBhw-QNCOmqmNRr}3_KmgpTr-Kv1(UR6O-(T}ISD~%=|T(x<2v1%>h?UvFx{6sP8
z4O^FsY;5~0Ikj-xNp_c>QD2MRm6C*g7>`RErj1U!Y)0Z*zze+%dw9m0x?3{K>=&t{
z)JU>!ia&aos63{A?rV|pnw+njG+$%4<TDs=7@^)CF0mnrklxdGnkTZARQRMW76Ygj
zwl^X7dS)zjMeyDJK=mydv)gd%Q!PYN3yYPv@9W`Z^4aTOurM|7Um5zm^tB@77EnRl
zdOJiJvP4hZ$a;I@?j6d!02za6s?QiWtKE6t{YqU(FG!5kti;C?;g!aki5!pVjw~q8
z9+5Qg#`t;Ae6$R^&OiCtWmIs^qvEIV#Qq`5HDFMu{ye?S)MVKC+jr;7)f<JMd;k1d
z`Jy>IHGHLKgzie9r7fh!Jqnb;52FgDnj<sWGFpr>Q^E%6D78slO6ru7%nab|H03Q$
zF!TwQArG}NB@LZ@aOe>6RKJbpsO!s_3({Ir0}1OG@l7AG^ZS0Uy;r84?Nb(}`ykZG
zE`H?Z8O>bT)3WzsJeFBanF+;qzF5>*Uy{hu3b4~Y>?o(jEvY4@Q6PwCL~fKsVRc#4
zUW+~6noZ5!b@kLnLhXxqyO@ZiQ;J(_#a9-;{H6=eQb;p@YMxVc$Gp0S(st%(S_)p7
zk-t);g+*}$b2v9KL3>$&i$=CB{D$(_3Df=S3qM6_C->h7ys5M;Okt`{muqs8)8-s@
zJwy<otIcFp&8Mv?gp*ELF6HRxKWqA_WA|3C&Iy9x3>A!Y?GWV`#|SQM9L-Q$Ic?F@
zTK<JcaAkUZ4*p!N*RmX2gRy-#yR@EbNjo!yHSCF9I!Cd5d7*=RGuGzn*o#wDv1&By
zl@Y#XI<6c?N6Xmu@3IgbyC0wS#^-K{9pP8@#}C(TZSakSABq*=2xTyy6`d%0vmx}o
z`px>s*kvl)UBddB3vQKfJOkf2H%!;R|KQF(RJbJ>u>Z(pd||0&YF8#;X8gmpmC{IL
znW$|uLnz}f=Jc0Beit6SiqjL`HPllWU$FReUA1saC3PVt4gamv9lcPmt3&i>=9had
zxQL%Dno>1$F-zlXJd+_KE3YnjQhXhETTuE*JoZ9dV9+uLaS7X~N|@H`nUFb5pVQZ@
z$fyFD#8vCIJ67SGB1Az7<Rxx&MkPhxd9DQ4-y_-LsrF&wzZD&r&})2Q>Jd?#)vbxk
z&0L=@FT$Y1Z+`xYnNwT3$da>%$lHvv;a)x4VstXm#DpD47mqB{AS_H6#O2GYtMh;n
zFQFt4B<Ms6aInuq;flzV07*8hss!R7oC1+<5slQwh7F23Qz({Ds$Ly{u84R1!fS!e
zGugSM|I_oB1I4M6xKtn_KcXNz^u5d)I}5Bbl%9^byifoUAnH4+q7Yz&?Egx&O@|9L
z-wh9Ij+}&C#BSi?A-qM`sq-QvIf%j!DkKXF%M=WsKTt$>T+1UcX3}wS;YYG|fPuXB
zCT9o@-{q@$ebz`TR0^QsF+GNOfxz~mg>QXC(E&g>4gG4=j4K7727Qm^i&WIqR`&MA
zG*^h4ySVwR|LZle=od&8cHu9Nkjv})*RHCpaFO7e8WF%<zjQ(9MD9oTzeO!#x_nw)
zWjf~C_r_HZSbIv&&b(^mnlfRuayDTHj)oon%YMhApXPQ#q&a{yvj7bp?vi)`Y?^Y2
z747mt+!BC{fcVPm{qr7JT`JkwtmsGNshJ$S(+zW4uBdFpY=nGGP;8HPRh{zr8h|3?
zkvV?kjl3TyR{<`pJ1)Bt`tJw-8OA@uWMgHOk&)pK{sqzg{`hz#V&4smYd#l><{DAq
zB*^!-E4Qe)|9QIp(K>qbp%oaxBa;9nEZ7hLjbFXs-TA$zxj{ylJrVbR_eA0%UN_~`
zj~rRSH^QJ?G~DD!>pyFN3yXeK^5&BZ2~*ETDgWzNAoV+nB%f}a4v^1-3++W`svZ3A
z4HXrMvX}iIs09u7eGmL4giBcDb||QHudD$=JS<{;@RuDJAqJ!sxk{qC95_qo@tU~a
z_py*2h_xG9r)Q44)q<)(D17E?4yyWeIol<U(zYlIPPH7rl>AjNVsd6i1?(*T6HDTI
z<Hdo2L!CRg$r7uP8zo|KMmKX+T>&QQ@5hS?un3rhXd8DBKm%3l^&4=T5En?wsj2aZ
z@YY{D*YN$w$#b0V1)m<=d{n)9<f;RPMi`(_Yah^wAkLgj7*#6@zW{)+M)vCpGR*)2
zwo>ZVr!YIe0n;uHB6&U$M*y*L$r;FBYo&JW;j~^`Sr@!lco`3#-!q55^EVFaR>V<5
znV7e4-$uNvzh)B1^%b-fN5l(m9{K>OB$>InxSu~g#31^n3`n<lu3z5&sA*Tp8w#Zn
zc8W;LkbL}m?!RMl_C)lWG`Vz(Bayhbx9IEj&G9~;IMiMWrM(h|_*#f828}9Gyiu3s
zwY3~@jD$ENJU>7(N{yNcQLPL=T)I^Iy~B@xZ$dD4Ij$C5B*~@6s_NTke*x~=#gcay
z8K|IWiW2utbM|iJoEf_haU95v1(`vDsYEWIf4RBX>FwhC-%2CALz#nj+*AL}w4MRE
zvMHsZma1;<W@M@hSl=dCN*qG0xo@z6pw*6SBYTDBh*EujyCZX5U{&J*v(+sX*;f47
zxcsLHZnqV>LU$<bc4Qn{|3xtFucSFKjZaStjE|3ZLYD*a+<r736>CJu)e|`;A>b(@
zL<RuxTFCnKX#%#Y3+iSd^Adow1<<toz3clY9z+(K#rNw&04VL^Rso+}0LS!qlMe>d
z$<VF+lmnIS9rj>`Ca%L{HYL`;@I>MtXjXhp-K~2J&>0E^y-6{Z209UCWo4x4kYad)
zsM#n#6xv#=<$a4NyW~7Rk6&CI1q{Qtnx%iwLePALQ!cJ=2_cNH787m!8VK{?y_K<N
zG8;oH5qAB8vKwPa8w5K|FKALQ{MO{W7QKIikM6v*UjU@`s3G8Ea91~T4RJ<tA}NdD
z#IOL!ZN`fi4%g;y9Sq9!iS@wcIB@hu4h@K1O!rg?DWe94h9p4&1N}-VH7U;cIheI9
z2%waFOmZ>h2JgL3s;?n7SVKdj8B9dLZa`rn0d~i%*RNAS`4EVpFnBEF?D^Mg0z8b3
zbSUuPRiOQU+Cb-@v(3xKE5=OYKet!@jtSUI;2{^(A6}`@yTe~lGlh4Qrqkdf3KYJN
zU>hWgnn=h$7AU<Lm6fq=Z88ub;x8JWq>!Tx9xm|kM?NrK_;Cz!W<0$$%uXNv^FjB&
zhf&w0h}`0y7~UDI(uV|-@Q<LFIF$bUdD!RA)^HN)jqf)6UTHw>Gk9*aXHz@WML{nT
zvA*#T{o7~znL#Olns{ikAnA{mmexf%2t>UKNK{w->jI~*JqZUBWn*Wv>R7ZL7{DR~
z8!SK8OQ#%Z@F1f|2$Uwssis;t{A2TDUY}oXa0y)<jkjDxe_5me25ulr66D5!!kZzC
z#+LxOR-iHcg`BTp*(@RpT||X0c>a6=_S+*}A%9N&KG+pIuC(Ke{$dF?A!?>OcE}n4
zAA9@hxpmYQM6VDGRg3{KMrQ}O|8U*^qKn1~1;+qhh?a>gtqVDFutvE+;0!IdfI*#Z
zJ3BknzD58k(kL+pV+8GKRL|232pQWsFm@bFs`<5bzwn>1CpFG@K0%=5U7EbI=PzHz
zqn=kNhWi2H_#>+HojcvM_uu2k<mn3V^Or(L1JrfdPi~+s5D@ivmz5Eu=lU8I;q?$X
z|A6(eX0I^JJ%D7PpXs$*JJae&gW{*W)7%esuW<{Kjwn-f9xR?`e2G4A!UU%~q|fQK
z0Nr03g2_+{9LiP=l6r6o0B=0hIE6ROZQODU(fLsi2VR0O>=%E5Q=njf@B8yh1?c^G
zsUEC=wU3#Img?2I{?$DIf<8Dh(gw$gG;HcSKluL{ka#b_w?wO>5(XD*7;BXX-F@H}
zb1NeYlc>c-bC|aHNaci8PK5Pk^8S_Mypu3l&|y6I8iGSA5t<vwTLgt9p%^>#i3cql
zBm_i#Y?wJWSv7_K7^;#Owgi%TL8vW?l6zHLyMDQiZMDLf!MKnpB*rKzT-jRkYw4I-
z!dBbG>&9fbFdJ?1xkx*o%Y9q6Z+MihPSRV|@>N@rl-U&U)6~q8NhaU^CB%fqo;eyr
zl+;FNo@ZuwMuRrVG%OIGF&96e(_gYpVbdugBQM5mQv6)%8W+!a$E*3%*j(J;nfbAp
zm3cGYOKLP4D>?bjg(XB!ly1&{xqfs!-Opy?-2{v6(ClT2)V1pK!Tu3u2{&tx<Af@=
zKAOZFB{HVOYPvp<$-~7ga}5?|jFc>gd+Rf)OyBrG&0U+x@U%p{GAX?wA#3Cr0~X8>
z>iQc}w(IXWawKK1k+qTLaHYS9$;yklXy)mbD;f5rC7-9`=*3JShF}x@Pcyx|A1Q2y
z7_hl{kr%^FXu5M>3W+NghQ};$H`y3k>k-hZ`7ec}2<5gGb3L{g`&GRzok(V7{w|`K
zpFJ+ND*9x|t6b%-jf6x@7n3@ZNkAq(1=qzJvkBNMbB7#hhGP3p%h0E;&5o0_Hd+K~
zbsGg4F_n&yF1hoZ2nv)RKOd$~{3ySr+vjqz(GWFJzed-f*4u&e_vdbt_)b)wAL4RU
zIyqU9@AxE}=2Uw{!UqA3N77ol^S!h--#_#VSlc~jAgx!6BJuUZvbBDcA$C!Fz*!#L
z@ap{dJ^$z9oS8z#-Hn|3pJ(25nI?{yLV4Le)^ggKC|-%CpEmyMT%-q^d{c$h<?X>U
zXXT!foTmskQVWB6vl-Xq`inXHQrg8_U7=Lomqt&qVL?<^yXk}HoP?dove_eY3#pof
zGV8}AWnae&g^%~YOkr_hv~so17tkA#s(<(NJU_0P-j$Ky;ry&$m}Bbg=Bzh&kB22B
zB;Yr63>G&;ZievX3~tp9*S)GOzNK?qrk0qAIEkEM$}#9%$dz%;)pjl-m;_b{Esjbo
zCxT`Z9xe=;YO3*^f$<Y#9?+tpTW|xMl)QaDlYx|5g62@a!P|i+ePx|rvPYB$K8^Ix
z=ah1dli%m5H98?i`lDlMyd)_y@Hd?-_m=!^PMrpE^7dcF&ctdoYI9ht8}E2B%rQ8l
ze!Z?igRX&;K0`y*1#G@68D9<myD@n(T#xuoV~c_`b@E7#g>Sa;#;OxpYr(S{HDRsk
zE#3Pm7#x0MaCk$>Aui(bvK~I7Od)J=))eC-4w`Or&)1I$UnF|bd^p1^;y1yS{)U;<
z$S{V?)<wgXZM)=IzY6_VmPbBjiDG^x6NlC09E|Y%yi_xvdH7*#6LoJg1Tq%Bb-H4}
zV4NRi`0#1=j8Z9Mmqs8#%`c)UnYib<AvR<0bdF~VnV)mlZN`Q&+OKH`nT*fx8!u3e
z<v#NFx6={I?aBCZTIP-wj-OWM{!K%R$curJ9!)g0vw^SaWbboJlDBW2Jqi`>bQ9iX
zciM{F#Jj{v4?h}GVnx;bxx(Q$afKox&fZKk=dKYQ%L+`u5D@gael2pkW|%~QCqW*~
zve0=nN3?j~kXrWU@83}U^C&%FF7itAoV&eHDPh-bY^uq~e~S#N+wG%nuV1Dt&Dw=3
z2=cNHv?NIJiW-r$mvpp>RhmCwnLdpZvPu<r`4=cq?!9b`AylK`K27m$F7n~Bx&xL{
zXSk#H`n5;tQr?+VefQ)NNC=Fge6Gn>-E8FMU>|DX>GJlZ-#XrOx|@sgiG9Yn@)LnG
z^CoGuIr=prIuDq$d!NnHy*Pf2#Oj{<t8nR<z~y7*rp<>^TE!I9#3vL<TS<>kwze2r
zyuQ%QXQdFgvY`4hF3OyLSf-jL+xST`sRYpx0y0PUJ2*d?W|PoSC5uS<1#+m^xrhRE
z3skFnh3VGgnMKZ<6yC{KyvRLBnBsE{uI+J>pmtB7LH+kp_c%#QnRo7*N(#htW3S?l
zyQdYAmiGpPlFqhLgxw&fz^0gApuDCs=w#*c*-?XHGJt5U)xPEgWtFMxKtj<Qf#(Wh
zCdSsK!ps%<g33Qb1L-rQ2TCv{g>J7aLJ}~e6$SU7S(Q9vX~myhwRb)H^LSI_>sh@0
zR+=~~o}$8x_pht1c=S_ubOp!NQ<Yc@s>K3k)*^3mNS@hVaI#DwpY<anOf_;=%?wIN
zQ1+%Jzn%sU#ZrEJ(hP>L$|$NTS{oB{tW#Qyzjnowg``tv!Sh_7)29gMn5KN5XCGsm
zn3=q$64fbMN!7lO{v?+nO(Ce~%N>O2lrL~5X4a46hu&(}?vq_bsoIrtD{(P)QW;W~
zb`#seN326RY#2k?jK{@ZxEiUm{5Qz_!mnR$?Z0H^`L1^Oov<bhlJ6G}|GW27iXERM
z;3P&as|FtmmZ&D)Zr-4tG%>avYI3<cs2_Yg&ucyTlp^`{hUh%4U3^vhCA*<ame}0h
zf+p@Fr|GiVyU#baD~O&1I?dQoF9;NNZCKv3mP<PSV{EC_>yi;oRMnHvLi;0zk7r~(
zi<68$X&h5{s3te_#q~8nJ6gpTm|w`+Qej!zo6Bk1-c`ly)#UpH1gW&#eR00M!Y$)v
z3hN~a?yY4OQ#f>t^hJAI`Ik)}h~x!ZcvOA3El`~vXs7*7ghj#ifqyzIb$G69rFg2L
zi)LePxYk}~m9C2Xc5}8PJdPco#*Q=_<$a9g5v6(dKAz9r3R}}dB=}@CG{lx<PkXdH
zC}u8-WwCo<ny<i#3?|gOE5lNbQUdJ#V*SdeANk3v&?`C^`O%#yP`RbCbSuhLJtosH
z)i}<U#lU`Z4l`LMH^%uWy``j(@N){QjF1ZxQIe9bv5ApI_4LHThK9*SN@LBa5N|SE
zYXUf~e-yrwej-6GI>Y{fzYvp|_G)PGmXy2Pbphtc;~TcT9{DeyF-wU_kk3BG&oz3T
z>myzrQL6N0V7TX|&he{R?**%I%AceU+skRuOJbzqMkuZ*r!1?P)i80yZ%^I%YWtY(
z#Q4*dtTy?<)f`9S5UOXoTCR&f?{YjgR&A=J`NiDGN2DJ2*k)qu`$Cx+CH8pt8v*t*
z34hLOlH6bA$2eiha&`5cc5eu@N=T?pYdzr;gfk_UtQUa?gQ=pDVBu}fHc42MuX2;O
z<#|qL0lk@fY{f*b<n-lR6KNYi^Z>h@zD#;XgXovSg|!3$_)uVpGQXL5-0wz1#EzGs
zRo*XMfu<Wn*1300p6GiOD#ec#o7=pTrfDaaY*7D0+?-$G@}lk(Y@Z1UT;Z$>x%>li
z^c$TI-6u{?CB#g|CM3X=Zf(4@5Hp$iaCw%B6yU#Anon7rYOVdBJ${A=D;SB_d1g*_
ze8yU*kXx~JRv8Ok-d_9g0jCD1TUuHc=0_}_h2jyV=Jb+U4i6V<(S$s$->Do621E91
zc7NV1g-aM!SSbGa?A+@3E3vN*k{gfDH4J<H+P-Q;erwS1>8<J_qsaGDdG6-b5tEv$
zmqHn>lM(Wmy&NqdO4FQ;c^soo-*fB4$$-cNU7C{N^@WA^H-0J#JJoiV+Kh)jx6kM;
zW(bu0*?jfQ7onyP<?Njsyb2?iVMcQViB!eabRLmTtiNrp6~@Sv%iC36{FR56px&)Y
zl3aG{oU?efu5uCejV!%p&sB4hZ_3d=N_I3KOvvzAh~S*yX2;zcE*Bkb&aaWdCCKNB
zC|^zc^f5>GsBnYa=|>A2<neo@Fx4yc)T?9%@5$?sOl{SF*`9te#Buaoqacw<#DFdm
z#{;?60_y9(3PO<gW<$}3tOt?o@S%|$Yb-;o!Z+Y|t~tJi)Lz6`!GnUd4P|5iBbDt4
zTeyjO0l6AEDQ@A=P8X?7t`(2nM8*Djx__pL-WCnvOw>d->}Fo4pqv-@qIZG<Y!F}s
z<h+4A1`*i>K?D%k3Bc>gS3f^L8?UfB1Jw|CEhHnGN<{G*8M%c-su1Ie_*HJe9-)M?
z966}&2w9kIJ)rpbuj-xJvB|asCK_^RK+3>ub(UU{6?VwTXoDhw65I;#fZsKn9zZgS
zB87luNgg77>-lYI5Wbt$mmo1Cz`p2q5bF%Mqs{qhS-;<Cs>K~}0Ys$uJoXC?N>K)e
zJ(vG03folHWpjpzlIJ19Eug3jSHd<U^*D-BLcf+6HEUN`lA^(Ys+vzwP%E4kXS9H!
z;{S&XI+!-dL?B`cBo<@Dyxhztf$jn~OA8E=`(~4f?Qt~%WlEz&l6$VMWQP}yZ0nAi
zZOnJV_0CAEnY|k`{AZOaHSaL~LV{<A4`q$Q6W4t={vqlJc61<Vgs>0waR;cDkZ;BO
zraQ_Yo^1^Q+2GX5xBrC<|JL6V&AbN~&=8CxI4Is3-+6P&ClV@31bg8PF7p}eO^m(*
z9VXkJWeWE}@#y|+b(c;@QSl@(F)>O^LOCUO!Fa`aSW|0_n|O$eD3iI6z3m0*e-YnX
z!gLJEXqGPZrkL_P?W#I*k|I(eXJ@RBPnoEBY_#*=OuPST|EFuW1hQ%XYMD>Zj!^hn
z@_+xY?<+vcpW4NnxV+G}Q2Jli&gV!1|35fvK&JY^$o>b1)yClVt{ngOXZHcg<S+-o
zrPbK+THZlbE;Y@5q;uLBwGm+EBFdVB^&dbts_ds~+WR1Y`S+WGe-M;b+3x*A(=3_H
zO?=8DFARWKi*(B$LBo3~O0MqX67pdMhn6Pq7=?p(H%<+gXUMfG19;G>nq~&7=fCc-
z$f*-Cj=nmh;1-Z;PyEGj-$2OGfUl|3ftB$5S-z<a<-p8ieo$tgnAtVBe+}f<kZX*p
z?oQKaaBC<ji}a;w_uN+!Nim=Ai1PeJL@ToOyzk)w=uyWBcW`)!PY3kZNRbWydd}vz
z7zVEvXoA9E0fY|@w_=TG!FFB$jWepVPycT#;qsAAeOa3JSW;6wkXYxHmcsM8&-3q5
z;qb+wK!>S{hulDr^d057A~7o$5!>EB2;W~CET;qKUh#+yauF#FnL<Jl_}1Xv5c+r+
z^-s#iS8c>5r>1(K{+EE2fugq1jeNTrL3ix?Si9IQO0|FM>F9a5nKxWS?A;7RlAu_=
z=7_LY2z*EBJa{NOb_0lpo3mibvsnHgupE(Ty=IUTsqZU5T-<>7UhwSM<fzWU@I*B<
zUt5^uIGU&a#qQoHkayn#$EpY1v?&E_F(}yj=hia$i6e;J3!BL+AbmXrB3iDMI&17*
zy?9ba85CbTG%=x|tW1oD7?`cbq@ymoE)@R;aQ$mIN%Z#a?voyKN--~?zmU8mjh0Mk
z{cT{TqIvh%Sr`=jz?}0E>r3Rf87na&K*3`uofc6U5pf!p-m7B8Qtf#_X9xgp!OKrZ
zxFSkX+7p@l?<DtVSAI$36)+FG=6B>E1TjVgzg!%>p%nS@#T?9dP+tnkl))GT4|e9D
z^*zKX#p~NdY|Kw9D@a^G@Bw;WnShFe6d@dvnIn+UczGzV9JvnxD%4(0IJDVy4)0)!
zb@(sQp8ZS22l_jq;ivu#r6d7lJpxJxR{v+PzkEZ;9l9DY^ALcK{6Wgb*(GsA8Ch9b
zJYDSJCIBdRn3tpd1tOHAZ{$4%#W|8hpjdRI!>$HCC9*5Q048NK$oKsa-oYRW->uIA
z&Md7+40IC^M-n)9Z2+>lBdx+v+V2)*L6f11TtN)Mq6#sh2@FwMFz6?Kf$z#3N`^~a
zDg}xvj+j~o`f{Kt0J9c=+FO7S+K4=-P^3GMmtn%kI?Jj_(X@Mn<owU$cTDzA`UGl_
zOCOfNYlNxe-sd^p3}jD5JOpDw>k+8HD3S@0hS7nBlo5!{iTKm-H(G$FD2B)R`nxe&
z5zyOWMP3@{S7e|fmS!<ptcMp0?L3HfKGV)EaT=nSbFEm#57xzoz;@LIdGkbUuGQdw
zR??f*@p876NHF){G3z1+o^lu12qVto_V#x2y+L`UWT7bJHjkp@Dk>_tz<cTRVdVFK
zLc@`H09gDYr95DWwh%cDQzUkigtOOS(qT{WAe((dW!*3ID-mJ3{_=J%FQG&u;?aZS
zt`LytI~JVa5JCqr0_MOu0x2hu!UZW?0J@}^nwkPmdA3t1zs<;9FzP^lygaU}+NkjY
zEKW8nWG5w`H-wH=vCYq0|F)CKQT%gEHVzIJ-@kr*XcTCsBc*nelA}rx<;Sm#zF0!m
zTtK`>lgPAO2RWc{L$?;Il+Xx!E%MP20?(Uj@CtMP{_*i&TLbXEK|bRE9YW{e3B4o^
zZ-GwF)AI_#Lg9)QV0na$ppZew-iG?W2`}a(exwu0Twa60JWQO8uE5eaAI9K4SA$ar
z^pwzH2JcQ7NUr~hC9{D0`u@5P{Ea!`>*XqP@7VZp6~KQ)N`(dasx18vYC6MKgLdk_
zp6>82cjO)$Vc;V0YT+9~2wYBRfpI#yG%9&;g6$A9hR?bAUNTnAP^F8K!&{KvnG;$>
z_F{m41M8eaJ5Jifgb`uZ;M)A>udxFHe^C?B!FTUYVPPSjHZ&8^Zlk5^cm(|lU8}{y
zXT`=a15rNP<VtljR|{VcficwrWXB_R?U|D%R#@qjkCDO340IM!*>g@GzAg)OYy7aw
zUL)slB_=k1Aku6kd(%+a{^Vr~*V)@pW;DJ}dGS+<pJkFUlo97i{PrGBJxt!)v>(%U
z?>6C;w~<Cs(w2=wGO2b0c=F9gbgWSk-oJ#h%g5T|-rmsd&LEE)we~bq;~_rN&F9FT
z)V$Kh)oPEIeub0!^d8jmbR?xjX9xlk5}3mQ5J<@p1Nfs;k@wvVjbKTo=Ef$b3wxn2
zttoLK__i4xz4?th>1hpHwr%A!PCikrf)e!1+?h2|@|{K)KND@C1gVw;3=?sPg<|cP
zP`^ZO%%xBUnGrs2h5_-c*v!*Mw;mtSry29N5Kh3XzTG8IU2qeMQ}*S4Y!(wi*}iZ!
zX48SI*qoW)HQH9&P^<Jpdj;j^;i%)A^UY-2H~6!XMeumkf*u<2zE?7$1i(ilK)LtW
za%eD7ix{oMH5C^JtW_=1_aj7IGgjoXu}(=PdU#j6jnsRN;l_rwF}A08D$dm|V?MQ(
z)1HtB>U3c&s3NoNav8isLuXYxmMhJ@^)2hPU(6Z+n9n=0>bZ&&TX^{bG-TQdekKVO
zvhKIc1P9>nnt75xAUir*HE29)c<9Olb^;^(W-iR<cj{(B?=U2IdTRJ-fiY^~<Fewh
zMnj||B{^d~xM^3VW~N}qtCOvjwz!<nlxzI>CH3#Rcdnm;0){-k;b=D>%OiXva$_oN
zl~t6})j|7U>njqc9|M9Jpr@?#R$QBj!8}_-cH(Xc_g(JgU`5-l+Rnr7PDWA5225+Y
z%rasCbxkNom0jI13bMZPM#7$vKewAn{T`KA-;w&Obt}TU-MN^bht&e@ZUfk7GmgQ;
zt_93o5--{NIhp&&ERmcbts>sov@q)S6=_b^fjNuuxV0t$juR5c2?)CR*#1xuT<B(B
zs_&9lonxZ>GOS_e0dSd~502P-@|$g*Qx6dl;M}+x7bc}7|NNGb>*uZzdhEze&U_Yt
z2ggq9X_38qB@e*)Ir4(CGogwkM%kK@T_u(g8@8-2-|p>p$Ce#eUBh<%kS1}U>6Xfi
z+I`4!dp|PSs%$!}6GQq`=tFh-V_R*y3$Z{XQa?fPU~0ybrX3!~ERpZxP`2%<(F*ye
z=eJ#hD?XBvMRD0O7;iZGU68#;ND07wD5HX7qveaB<F^uT+NyIg{G9cpRSSPylWOL)
zCil6BiP)TMl8Z9XTOht>nHh19xR^A%TM9oCl^0?H3C@4=z&Y|uQqE>CA9ub-vG+9#
z@7jsZm<@moy1ofjn3GEuvBIph)rAIuQnO)#rZSebv}b*x;7bn^P5c7`hTts0UpICh
zWos1gJ?HPKz*3bGhxq4r4AoygBSXqqNbP=q*m7{p^w@VrqoXRjbxN^M3>z=s{9I^-
zskkLLu$s-y^@*5pEP7HpH{qz8hI^$_h)}q^@Oc;gvsG54QuN+AG*6AgW0KM_-$UhJ
z_J4X8$wTWDrNTl280x28z})<j4T|1Hy?OnPPbqn7+pXj`Nu03F{c#oH^X?5T63g_V
z+lj&*h31y4k(Y9*)2M7^WGRx0{Cd<<pG#a&adlwA=3dm4t-qsu`Z3XAY1hR))+M~@
zlHbB{VHXJBVGHwJp50XnQ6~qSt!9%_W|+_Uu9MJFYJMdmrJu@fZ#LZN8iJa>ysjHp
zHlf7Uzrp)aBE`uZU81WylnWK%9ZVySvh{nWZF~Nn8;LBgqNzrM=Z-RQS-ml0JhI^^
z7i&~5N@bXrMjy&pqd-}+@l>JVtQ38velXDAhHpKf3;)EtT~_ddV2I1AxG)FMJ3b;<
z0Km8MSFgSnG<$4_>znQ4D9x~U;|qI@Ypv&<BuKltthTJ>?yv-x_U&9(xwtQZ0r0r_
z8I|0myn>^6iXKG3MRKtRJIZmBjD6fKE$=rdPO6giWdZD%4vP)VRB*3QdbY7&_fvn6
zFSP45&M3y<dqr*Cso&?%XnlE1R;BnniNUBk)@$j3aZZ}z1ZSL%{hWH$V+-Mjj#4Vz
ziWq5V&NOpDI}1=lBCjT)CY~-kqvt7^6AJwo79EL`vg(ZR*!vnnO(v38!)F7pcl?PD
z)m~X=^>y&URir@Y3G?ILS>9K<JtJ?6yF?8z#QGr{-k%LqX}gPmi6s<Q#Am*F#yKOo
z*6SM^%bX{?;qGpD$%mY$tdP`C?3(O-u>g~CSnNXKrOGG!T;VoiH0)=xgg6bJrE3?k
z1?)B{ck(sYdFVFhUVb~E_NkRe2|yspsTzfunv3Nj)`Aj%7pg6rC%)$kH;OtLVtaY3
zU!%h~i9&udM_5g9(5dtA!cHf@U18B-EFm<AOgv1p$cA&f@0F7fjLG=}c$qiTH&Gmx
zfRk6$q;;`B+ng+dov3t6SPfPz;;_vtACpzg45uR*PhFwc=rT#7$mxIa{tD4fnq=W;
zOp0yp#MZIEG4)Tz&c2rb5^lV_TDgy54+!{kp)LaRYa08GD635Y`hOIJqDJ8{%Y)W=
z@<vfS>_kK-96Q2ai3_muoR0l;(rLbl!H7o9?oxhfG~bx!d?wkHMBKSOS3?Y8Yx-8A
z2(R|J=y2IV&X?B!>D>&8{oPbft2<<VPm}jdy}7#fR;ySac~yC7Ljb_IzTC2sWw?wW
z5t^ag)3IYZpYqsdR4S90?sPpV(NhnX<^IZo6FWYcb2<j2Gu86l^dt%PGR6HeGsANd
z7#=t|KkAF(!NjZyp=@H#Bq#i?rYir(uz?`*ZrAQoYh!z*E&Op#VK`P$Id*w~ny3Uu
z^HKUkyW0=d9$H-CyEq8oa7&A6z@fG3FI}5$hAGb+pYVC`Gbue*rzyEu8)E#4x2z^e
z<|W~o09$S5QRU_qURw4p@8J);L1qbXtUgswO*VM%9iBh<0#3jQwi!u_fJ5+yA}`)g
z@0@dv-J1ZYa!px-M|qHSX;Q%Tl+SodobjjlN&LL`6qK0%uuqR=mgu(MzjkM>$8eAG
zj8^eC!#x`+^Bi9{vudC1(w(X`aRE06y@eQ)i~ROwgFC0bANd(9j8`}%Rqb7-Tex}{
zR<b$`bdiS55BcsoyJW>4qC59vm567Rb!X`53Zqr+cQrY3+3flaF=hS4r=4oa2BE^1
zH&6X2YOde!(bdp3@QDlZlDU*v&2DMzX6$!aMoa+vIHA%u)Omi+&!iwv;K2n#)Cho9
z2g~7wy>P(09iQC1avab|+_+WS?O_^nk#|z6XIr;!6HXQd63=VLJ6O;$pKc?yy7b5_
zAtn=Vw4(FzMHnLmBR>mapn+CDK5r}efZ0@Anbux13ceP(o#s16_1C?><-8t*Ptj^H
zo`tX3Bk9ms{~agC1Y|UDb|ddDW3_n<oFf<H#fM^hko%$;aG#2_OCLaikG+u`XqAZ2
z6whQvB1s?U#2?}U1sW>*p!CTc&Avpn+Rw4#ckEw4GkEOw<=y!L(H;iBH{yZT6%BkD
zve7A8oy!ye2LK`Cuhz?Ym=#Av*@SAQ^Pa2RVDzRZ+=k#tguQ$PBnC3&uObC3l7T|`
z3HE?j!Rr&TA`l<s;xV8Dl7mjLMSp11!zSflWpORS`#gOj`1S#)vb{2jGD#Z#>@{rq
zpYrlUs0I>30=g&@ig^QeV1!7s^%T_(KoD8L0CMzhbhI4q4G$S@ehYK(3V>+o2OwK^
z28!KX;E8?Bx$J#>QY<}A?;o~B&@Xl>y1KX<>DIu6@%yrhZ}5`47qDs2SG(67^$cBn
zH<ExlffN*<3+f}p!~h4<;}15juH$foz#MHxu9i@_5HunQ8!)_dDlFB^=zlW@vktKw
zU=ULJeLhcZzGs~B4n)N~glYZF)B(4^zXo+s&_hbIHR4^MiWiWB{NiGBAi|KJcqA(G
zqz7CnRe%Hq+_|c<9xkFLE+(Sxli#1_7A*f`&29*;#UC4l8VfjAHy4zF0>JC_{VqaH
z5Em<wUJ}_Tyn;+`e8i0@UMX{0=<4Do5#3S+_1ztJ>P5QlvJ!Ti88exKs%Cm;M=|m(
zpE4?vAt4>TAXbOiWXRz`DnOuSp&0eFKfjjnP|=E<P(N(KZ_l^8U#YLsJh%#p?w%dz
zEI2x<0A&tr4Q$OOgD$0ZK$VZsXylFya#5EBhk&eJeJie42PHMbhHvSN{{9#UYb7NK
zY(|QV0Bi<0y71Yv9|I9SEdU5e0N|4)8$sA#WBd<%IaO(prSiW)PvqK=n3%Yqz<K~x
z{IDpWk%`8C3$a@;azFSd=Y&RX{ueB?*jbU|ndEH$l*v3{QhVrM-gnt$+S4K4^-o4<
zhNuM(hIQI!T6my<;GxW$gIQd+&kR&kN*akBaKb}MO#q#Cor4eF#(dAji<7$zwDr@9
zM+jf;Bf&Q)agBiBL<2k~-%U4&I09Yp`d_-HgR=3#51LnoJTv9LJXCJwP2xO_Vn~?+
zlhXco?^@8}7d*?f*LvCK4NSk8uK%)}XCQ7)5b!#pW<!AekouO7kB`f4g75RnN>FD0
z#x4LvKqvg`6$lLk@T5%7$yp$d@c#eNM)zgS<qsZ0UjnYmB~-ZqaLOVO-i)SdT#z)E
z?ie15?J8Mo&~HLCAsAmC;E+*KQGEv;aR(Ipo46&pyrJx&WnW9NgR9MlS$+I12<Ii@
zfW_;3z*8xGMCtW#<&Vfz_j29WLrCh0XM`cc(Ss9VVj`98<l<Z2zIg;T0OSE%p7FJ%
zg9}Pc2ST79csK%&+9yrKg%!F3h;=H(fZzD_W3zmYdhC%Mj_z5|>$W-407B4-Itpl&
zz@$X<SGgSZy^ELs&8bDpnv1@~p*nGu?LSRG-VNM71fC*@2w8~$Ox=Z`Dfq{2VEC>f
zECUHxNP3RA@s*X82;u^Qm?Llv!OK&2wf+6UT<Ms}q@!H0ij6ZP#nB`77hql=5x9uz
ze-!p5gj^q?*8s-sSK+yvNv?zfJPDlm?sl;qZb0ku3kzp~6qy55LD(7DyE6+5gOPPE
zb1;`O@x;M4C#<@!0l2NO@#-)NZ3o`>1umD*H_;8(xU3|=jtfs!YSNCND#QQ=_{4vs
zqZG^+;}Aa|8NwI=!4HAu5hgyr<}NH{hHy??azD6!)Svt79dtV2F9A0R#6c|?M`EBH
zI^dU0PwS)p)NAmt1$PTwNNWQIB_K}Fx*;SHo<Du#>NhD6IFrEk0-*3;G!=YZ|1t@Y
zS#|0Xw?%KVX!sR3&M>BfpC(*VZ^9txNJCI!)=lv@`yCYCcnbwF0e^sEjqp*<>UlCm
zf(Hx<4>SiYyWX%(6ZkE;Bi${C(GjN?|1WtDT5}Ng^(%+$t3r-C@+$xJYw6k&wC%j7
zDp>x(drpR)vQiGH2?nShK1SpoGg2Ibt0^d|6+oU`<KngPByAp9!T?%>tQt@w%vPWq
zc3oIf`Jump=MrFi0Bwa0hY1qpvWKkx=NJQ1UeDw|;L={0EPF74VOJrQC<8+EARK`J
zU4bzJ0FQHfV!m*Q<H2zX(so(|?;@N&8N4=;r>W`w&+WfRDrBp*Ze9gm-VK4IRxY9{
z+wli;LP(GS7&g8iYCq_DEa!Npl(8Jla0dg2NjjbXm*8paCee+~e?6T0;V;*k_H(U_
z6R`k8BDZEZNovbTbO5M@#cu?FpY3$r(}65m4FiKdK#p5lg3h>uRqNmANN9qA6A%O_
zQd%Ux0qW!{_&(C9bK{0xe+lTz<}d``UmFOoe>jPE@IvQP=KO$<&VhGUn$(e|!93XH
z^kIa&qM<Sg7Y0thEP%YA7<rq>9|HRiSc{(e3W5qJpwl|R%N_tw7tlY}c0`UvQWw+?
zFli4rgLws?U-i*ASh0!sM3kMKofmjd{QID)+oa)=D1$u~ck03o3=U90a_QIsW6w<#
z^$B5`_OE}KBV-iLA$ftM1&eEYl+a!^2AC2W0ywmgQAY<EA|QKG6a*;$1p5sJII4R=
z2pr2PL8PKeGFbWYC9)DKg)1VY86B_CBdWf9`OyJbx@sar)Bf%EjTS^w2dZ~{7nSjg
z3h0L+KV=V3Po%#<kYehRBQ#Ss5w{n-m8oZTVd|ls92r!@_epdo)64fxC&96P5dtQ7
zz(E5818DpHRTQwIx#8my+$9o5lN|sq_X2)2f{2qrYXVc%?t-qrsNGaetg!RqT_wj8
znoK45Yyx+g|KOR1My@$VP3OzWDTRGB#@lxc3VW1dOZ|x<@H!OpPpoCW{&{(S{hOAi
z%kOprn>Fv=AJf&WB!&hAhov6%zvC<5>a2alD-bK}wk#0F>uyz1q*R{7sBC5(|M=0w
zO)=GPJpN&n)*8V%pDf9QG6aq>?5^+Y?&z=Vw!ijjFp$VG=sG95ZgJ}Thw9aG&a`cv
zm7<+#3mGH6EN{68$D+V1<5atYwjrw*!VQy8T5aHZMuyCcgQt1A?%-Ngxxe1jp3@m^
zUVIflEV@$A!Y?CO@of<&iZP?cA6sO#4fh;VT~1$QkU4vYQpYl5fSfteuP`FWv1NL|
zSn-{gQd-2P@a~(sVTBRQ<A?l9AvRv1B`1TDN^#)u;kH4`6MRXLWO297kQt6G+FU(z
zf8*Zz$)c6%>&~VA1H^VjY<y}wN2qM+8-7ezy^2sCzEQW!_vm6Uw&mk;$g?dev*cZf
zKF5$B56vMP{gl6>9%M}(5NS>G@tYLqk4HoOgyat*M~oSzsRz$#E*HcG=_bd|P)S-6
zT&Gq&af0Zw057|(aYV_Zk5y7^^v%bt8c!Pse!WL2dqg%ZLTAO}dqbCb)y<mm_oMR8
z8UC;HyAl>Z)>S*IC}6YOHQlQm2$DRmP)_$Nu`qL$^*5E21j*b<=jANV83`R?<2&F2
z^wdmqyqi5{!~VQg26ryWikKR1R<O9F>OQZs5s2FyjFasS8D_mu-R5q#3o6Z(ijQz~
z!)c@+Zx%RH-W%&)=fI(IzbKEK!jd%xGb0onWyIT4B^U8*aDi3c_8|w+OJaC~tl`9t
zr}T=i2e5FlS~9iF`YRX5b5z==u1tlt58bxRuI!h4!ZBcYGDfMa{ldrmSAm5QmL*3{
zVAB^(s@~Nshp|a<V!7hfRnrG3D9;NFS)SxeIz)_ZH0BZy&1=83E6l#f?CYiYUQza4
z2tPI=KANN7R#7{Bk=9?Hq3F#+zQo95?VFW1(53QsRKpmJ*O_HCuTz)E4so@`;5Kuc
z?ouhqy@+rtt_tfJ=480AWG2woPbSb$A)v;EUwZ#utG8`|t40W8;z?pYLh0*gzFv2y
zwtIxd3Vi9DH^el1iHBNLt}pZr2#93q*JSZz9JeG2#)COkeEWj8TI`r`utXp~F1JKV
zQZ?K1iNKdIHn;WCHsi32+FQATg{nc;V_U1YmqdH7-~~(I+bH=g+FdN<&dhjJpCrR8
zVw>TYqwu!v_Bm=++Rks*XOr)f2ONqDp}b{n{GyGtF2{&{Z%+oRDcbC%6c98%d-AFV
z(f8-evz;58_oTh0ICH6F;lId615>)YhtzFNZ8zm)&fcaprw+V((_2X{`p$I|D?WN@
zcJ9>YR$^@JnkNM29hwAW#>gln<6nHc{$BaMuDN7RK&tiAfiOEI%dsZ{cNAtjG?yJU
zSziOCH1`dK<+0$y+PI@(d<lMdilJFUnqw5ZrIjN+xvy_nTTR<LOa#D`c%CaaF1z)e
zisG2E_`HdSZmi1!2U<of?5af3$vR<-){bQD{>qs`-8woK4!IGk?BlW<e);*lryT!=
zcUv(G7FA*sS;p~=Us`<hN2qg6x>U>LzYu;3?Y<{fZ_9qaU+dZIHw=-d0CCpy$d-C;
z+}IU{3q!QA%OQpGzO{i3R#Xy-{o<B~YUX(K)GolfuwsolR&<doZtRMO%~|6mOE&%i
z48=FG&b_H&9=tH%JQ|-Nmhk^*HO-nBnHaHr&rCq`3M^^&WcT%;`et_(`kwXD`#*G{
z11Yp85NcFpXH3-T<}8kHxkY&L9#x+RtL@I&k()2eG6p3K5*d^4zZ(>D{_f>fMNXBU
z!=2jdIbtwA{yl>jbIUq6r{C{Qzs^cd`Jf^TbBV*G_O!<Pqym0P$PO>C)UoO+%D2m8
z`_gQ<;T1<)zWAi7xH|Cls}qrZ+;hQ<7<)w`!6Uc5{d!ic@;|Upx62q9!M}b>;CO-m
z+?gK<FazO%Dups(eR2p4(g+xaWqcm2`FK3EXYva5q`ap`*+p9uZBwa&M+1)p678an
zwU6XK{*td`!SS@Yd{mYb>l&VWJ*Z0kOjhu7qLYc1rLFgD#gi8U^|fG==s%}b@QNz-
zRos%|y7&1P?w&bwUG6H$#b?SKW4D!eevVyg?`U(dBruLIdV+WR1aZaLM2wzg&Mc$H
z+zA+l();9gas#-SIVw>#vgQos#qxgaN=^}R`4S#Gp=|Ue4{I|R8FupHcH#rLP9a5;
zt04|cdZ`*4UQr_#?<|>epUK-a;y$+NVbR|CTlDL8D$}JSBgR9n+-)WIH!a)9CpLry
zyWMHkF4sFgvWq>>szw~q*B!MjJMw#ciH{<f@1e9pmAHv&_C?F|gvG4jh_Tx%xtEL8
z!wcQ*Sq{@@QQd~k(5gcEr*5oW<#~Z!my6ySw&}vQhT0iHjoJ+7G&Z>(-FrSk#m57C
zS|b@ok^uh*5k78Y>(|2@47b{8i&k~o<D3LxGIc&Tp`@<j9)XWfE%)52*lYKG8BFI6
z3uRKthT2+o?p5T)&9^ArWGv~t4IAmk4{8W7WIlztGJ*+~kYiz<uV!=NhRt}SB#P9C
zmr!z8cB`6|LNb6uoeQ5>AbWt^l84?zIq0KUBf*rAM_ATp+WM9lZzW>cv=;+swLZC=
zWw@YG_NdOlbMC~XwP;Kh)45UIGBFN9BmEH=+uZ9c>kUTB9{IBj`uX-IHaXOOqExuj
z`&LY0x#DRrN?(t>sj@y*ulA@>JMm=aRkaccF@c}66@7C;VH)v$kFu!lSdma$k|^N8
zO=-r52G|l^S^pIB>__tH2|9x_v0%_}?*`jgBxxAO5075sKrf~7;lTQ4sweYA>-p7z
z)A`p`o<}c*EnW}0K7n1xS6HQOOOP<Z-VwwJq!*}mnTRG`BQRk}jHI|@e6wGUR52~W
zo&HNM<7YGH%8+pDbyqu5?U2p$HcDc*Y#d|e#jo}}BD@_E7Rp+sUP2!g0Y}HpSt|Yu
z8sUW)A{dEj=BJNk1uE1l=+A1I+{wB0r+2=Ny290yML_SuDP3QZI;ypW5pEsYvdV#8
zujwxx38tcFC;S?*`}!pkud&^)J<Zpmy|p?<q{KH;I&p66F#SaxAEn+4(+p!Q;&%UH
zbwZKvF)`nfoNQmcu()*ps7DkcnfkEn^zGWW8joc;V)|ShjIU4X!?#6BuUhj7<rrK4
zKWn#TEeUE^y~TMPDk!Ka)vS&BX1QbzhgBTq2)%z~CzN<DGmr{LY}B}MJ=ScYvS@Ac
z`n>Uu{rR525=uXPEi-R-HoeN%*#<*D=t7o@o|n1z-|o{l2|?cZ#=pLQo2>k1?}!~o
zHkbebonjIk-<^g$f?qgbGcGpp2tyf)<hKZe*lYOX2_-F*H(Og<{qK|cZb1DuavKV{
zw=?@wGu&2#OxBfD4L^rpKpjd5;&fA>S|PUG4~urk&`N;|+5iSE;;8M%w!`QSN*hV@
z`2Cs%q)3G5wJspe;DL?U4`-nAxUIVbfA07Lx`z8>DAEC`no)|ww*nRKU`Uw{M{@6;
zBtZyAF)gZm52j@SVi4(qyw3h+v{;C3QX;~Wz1I2o-`B_`PJa^n$Gvm0k57@WLLc~b
z-((Q|5ZEv>r3w09GzwH%W06N2itKm;=xP&~Lv_N(dM79hEgc;CA`nFISFj0R`6#i$
z8SJ#2!EbF~U_hc;_b;a)n620l;;KRasG)&0a2+Jwo9eqF2la8(&+=~9VDvbkPXB>`
z7et^zY2BXFo97Yyo9gkI1<B_TI0%7RhZ>=ZLL3of&$nrUT)L1J4$6<ieyHX^1gBrW
zcN;7~;)&u`>VR*t00;mB(ftrM`F{wNi($N7^w|(;^pjJ4ep>yHD_q*%`z2I1p)?Fc
z%78)zrCJ{{genKIQy049&q#NMk;#VBbi%|iN02h`R`44^7G4d6)*px-ITJV|9B**r
zU&Xqi;JVEJ7}6C^0GZ?zpo1SFyq+B;xL3~xvmwDH0$`x3{!73B(>~zXpxhn}&iJD#
z$~zFIm4xgql@hq<ecxgvY1&&HH`v5RZtMN`M_jRs|H4!cz{R4ngNMd%-lS>Iws6{R
zREYwoX$%Upr!Wms=r}$y+XLnp)M!H2+WyD0&C=2mb8nUI*s%wOqniJPRJmN2u}>@G
zkk$@iz$lHbfz`d~_@%F?B!qNzBvCE}#XKI^EjFQqA5LJ+Nwj~PUhgsvIbb_czvpbg
zIs&!D7DSwAB}LGJ585w4uK}t~{eZHJfLry$d0fYWz%0q;U@<Vf!0G)y6KT>GY6MFM
z1-sR)=&ffK6a;}L782kckf~okFzI07`F&yg2gTO5M6O;>0T>Hadd0=Xze<st!N|x+
z26Om9h54>p&j(BSH{}Tm_KK;wX-_?cf8$mAJ~COJ0@pd6>}=mZH}PMRa|Tzo;QpX!
z`$45{dOuz7A5vT3Il<%^i-?U+Zioay`vGs#K;fr4NW0ykFelsx{q4Su&dmd0*FB29
z+VKyyT83&mkg+c+#v!1yZuiReYij#f{@{T3lb<u?;C}09?4rSZd{<}*MyP)dG{Q(g
z|2KmiP?ky2Kk7gvYKJ*^R1_^_2%@jhdj9s{O>S)5t6nhZz(}4xEVT4}`-N{0(DNdk
z!yyIa&h!QaG(Fs4bMa59oi77u4lFB9{X|xe^&`XoWSIE(u;XlBr%ggT|I~ufrcH4*
z4%DaIfS2MSH9i>KwF8mu=~Z_gByWbf<>oW;bk`@ZkNw|5*Hf{Gt`D^hSOt?yL+DrF
zp>+Eq{rW_B_7c#nBSIVG4H=;#AHhwCL<f;-N`OyFoHl$8jTqQafyXSV%Xsh($4Tv?
z>xus37LB}dsRiw`Axsb12_hsNAQ{wRK`LV;`~>vyZZ9OS5y4KQ^V&-o53NQ7UP6P4
z4>(=@r$^-uM%Fdk{_@MMX`s&;3B5Xwq@zgLo9qnEZNg4WC>9dL&*@+>2@Rw(IC^9q
z3FJ>i`udA@rCz1<10V&wbg!XZcMxVNFfkFtiqL+5CMDH04<5q(h)b=!d1$9#2U6Lh
z6${G(;Uz!dR6_8G92kZ)<A`ZMq5bPepD$9VqTnwu2aZLqh$s)|$2SC}plmNNkwr;R
z<M8}K(8dx5gVc>tQ)-8Uxlra=FsWPywi+PX7C~qyfIYpy@fQX%C-RoAhA2y9mlDd`
zn+|-=@bK_o*nPwu1~+XpFg5uL^*~e;j6P{Qcf79CaXb_PhyK&=+jiU4m*?QQ#`oV%
z6^HGf_k(E5Cvb&eBnv>3297A?c{v1DqX266g0lw!pp`PPj6rvvqs@K@t%TRV7B0j7
z_hT^rbqll!gaLN@@h!p^88Z`lhdXm#GqL+{+4P?#{Y%c5UMAi)lv#;Lk%%n`Ek!wS
zJ3z((;13LT1gO!6IEKJ*ZAJ$fB7MT4vgsKNTW$z6w+1FB6+w&dv+5tOL1j|$&n09-
zMD4ysG(POgXt+0lsON+diU1=2t1AblIu~3T51B5Y4i$U@Bs)Y3FKw8(ty@`JHwgnD
z8m}f<{eR}xyh|b6D}klGJp%ubZUG^js5_-zaJUgpR^(g>4MG&_h}81{R;fL|wW-MI
z|BId5?w{F7-S6QA5*3(LAY#vjO%WzGoYu%W65;UQq3a|CUcn%(eg7V5IH30!&6#ol
zYPu6V>*npmLwrpU5s_EGOLT)C3Rqfy0Nwa3?DH1jE%RdqsIl-u&OrFlF-5k68T2Bj
zW{QbG2<*MbtL&%&Q3I-}ProyGbRy^v>EL~T|DtEjIM};ZPq@^}6N0bIU6^Vl;(vi#
zap;nvqjtlfg2tNSu3z&RW3K5B{)G|W+`6O=x6OXjHaTV|zDm=Eu@=j1Zzm7U>1&Vi
zmKv6=cuGiwW8M#aaFnRUjYgGM)o0*HwRF-$rBzz*VwrRJBbY+5roH)cd}A8;mgU1g
zt;!YGNL@s1#Y0a@5kg-jPZ`F0xXoDZH%W;hP28&6`6XI>ZelE5WB>e<(Q~DTF*N5R
zE;(^S9h6CYyxEWN%%a(e&>TxiYk*Y*tS+r@2y9MkoH+3i8$@72<l1;SmXP3!#JMYv
zH&dZ1;ffQwE0lJ1vF=iZ{A5e7T99?ZgA8u=L7EuL<5b7Oa@u8YI>v-$4L^(-e0}7r
z_`|+yeR+Sv>zntoQ$>rfw@Ss#u||cIww=y#Apq>nTqE6_R+@To?dfoiF!O+|V&|Th
zQcvraw5EvOQ-M3#{(iSK(xdgC>b=om{DB`*=-7hs7)l8CN*<kwtk4;}h97mDCFz)h
z*F>Lun#?rm+_&jVeMT%#{eo3|h?+kZv{Y7bJo!zcEN=PW>PXp6u$R&>Wr^x0cj@|{
zjIbhh5|#1Muf226FQrSp8N+`L<(jz$U88EN+BT`?xd9A0%lLKQy}g9~!}4!)7FwC?
zXbhPrA~T7D>a!M&8P3fOjRoo|spy*DQOy`=k+RwnS<i3bH(r84<(4ykX?Q3Y%haY;
zhJdSC8>l<NB%EHfG-}sad*7qzzBh(F(e?o^BqOqFyzyI7Yv7v?-h+q6g*sB7)dY^!
zUd=6|k@R>!)++R*?iw{CHmtN!Tj^)AV!QUhLqNNfvDvdGB6h+K_m}hrWv#}oIhUJ%
zdzpU1ra$ou?;o`BnilXcz3IiVzS)nR8?F7FN~|1xGEvzdqjG_iKdq|%v7b<<@3e;E
zUb&%UFg6=Nv1pp|qXEi)U_gnkQw1<6SBVE45B%1GWhx72{&FEiXV10v#H82YjQoD|
zPS5PST;H+{fke8nN#@S=t(TjP25JJBBKkh9|5MzRhf}@od*V=g6QzNOM2l4BF`=Se
zrbx<6SxSZsi7eAnc15O*V<@sTn2JP6WNOEfRLVSLk!furu@cL$-0#;p_PzI<=iGCj
z=l*ef9{;phYrVhU`+MK_`}uwbD@vth(`NmfMIj7Y*k$)^P6M?A3~ka*4u|8>5(kC(
z%R=Q>OQ&qhZ|;kz$)wPmyf<W%Bn0N2SUJzP#OZg*A03d8R~4J@bR$yUN?6Tvwh;z+
z`8GZ9)*FlWlReye;Jo?c1GH;B#oDg2H`hHppm%Q7cFJ3+V3(MD>bkf{1!`YV{0rNz
zWUJIV+DE{9yw1R&&k8pUCa8vAe!V~Cq?!RELqg-|Nj17wtNNP(!hFJDLb$8QY2zkQ
z+I*1h*@^s`;)Oz*^;K(;)6vGGM3L^MPjET+QiIYrm(SEl2KP^z+nLebZJjjkjY<G=
zb)~R=pL_hL<d){+uZ}Zhw;tD*^`+4lZx^>$lT@Y@s8Wm+!}(^%c<;pFrp<Lfs=fOr
z&4cv*O3gDSYCPU#8*frH^Za53KQrY)Y>t0wfo@R%mnBRXwLph0VYK@Ui?a8tnmXQ2
zxvmeiVsT87x+|rx-JW3vY>$*6@+3b0g_7TNzJWPMaGN)<m3VXEz7Mmj?vyl_cxJ}Z
z==I%gPgMARqsJN8Hhf7Pn(NS9<>vi<pjj+WCabO?W+Putpr284V$ZGeN~;&Ysz&sv
z8?>KxS!lI2I(1&#qv06ugjUZn4Xam~ZVD4x8!Z~>_jeBy)zsC_kZJ-=PVG6(&m-f|
zdZ^~I;e2cVw?zd?o69y6#Z6Nrxvev#PtU`ztO?hjGEHKHO8$uErbJ!m5->GtH`8D6
ze=t(kkm+-<^)Nt=v9URs1BK$!AC5m993LsaC--3Zd`i0zWl1x&AX<TxdO)$Q$ny1^
zNV`}VJdv9%cBX`0;f)9~GYNZH`qP~?IZQ!kt+s<|;s?rxM^s`&#XcGuB|jYQHF+>H
zJ2lcJPi)p8c+RsgZYfodEeLw_J?6&4j~l0ld=6^M-PgE3c5h+A`j&2d>Br&o{B^Rq
z?V-L_Z6!1*%I0C8_CBWaF7it*l8y712d*N=BHTzG4a4zyo-)VwMEDRH#l#=C_slOy
zpO17;FC<tQZn`4>E^N?y3m1FlWBGG=XNqjCiq~B>>Hjn@DOI~6m&8qQ$u4?$Y+66c
zDLDSE+5sA$)%+ZLsmx9v#iXBAjVg6{|7hGzPQJ40ePF=7pX?H~K|I4O)A%79JK!q9
zJNH%ou|q=Rm{@9Oz&iI^4({m=4jvVXUTK!iRWU^!7D2sV&-RsR+Q`+=jvZ<LCaNnU
zRT!VYu_)EO)}?YJ+?Dw}^2vnQ$iCn5&Q}XkVlz9Wh}k=jRH}JA@+-``B=B2a@;!Ac
z;ffC>(PpNwZT!qhuu9nBmZxqP*&2HBsP{-(((sK3vAO<--|csg_dWhDB`A^DT;7p7
z{5}2o75%EZEkx-K{&0DYHla4jN%JFgj)8<R)`iiK1osSikxwEo8n|w*?3CrtMyIHY
zEBowr4wr;KbycJ3s`V>p)B9ilpI;a|aT3xwq4#IH`A<;+Yt?7d&3)0}N<I4*9mxIk
zhj<Gxw&p<r<|{Flyu>e2)q$}JXLBM~1R>ky4x~DfGf^lCM%1Rc=Cc?hj_yuS<^K+6
zKJeNZ23ZlDpZGa7gUCS00gw1FQ5&eE9!Zz2T(tX7^?z8%Y>LE#DP|6!Q5+l6Wk8d{
z00VgT@<A@4YP1Ifp;m?A8<~~stC?k7bbGL(_p{JxbfkMN5$gn@CkxFwivnyQXP|1-
zAxn%6&ipS1x~l+Ni3L~mwE^zXxyFRpf&+V2Ziv(2r}`j^SKvGl>Jx+FtTAXu;`jM`
zf#p#;v<a&lJbYpXxT1T9JQ_L`L5<(s)&|W^&3*UsmRS+c#B<_SM&<+~TW=rqWD)pr
znA(XEzQ7`5N9P(;j$!k8X>o>*ycYPXqG(-(S+rnLg^`n}iUlu0*PI6{Bi>$cs;-@_
zt@M9cg8{@LeqH2ukg@rHvFSD7=fp%`&)_dMpt510nGh6igLRw!e%_DkiBzrks}_8P
z85vg2w`<{OJp%EkP-pZ5ih^0CSQ|tC|2lvuPds||rujf26b5PMkXo^Jf2P@;($bp+
z!Q*b_^F0$;nVCCuo~>MoG5gW47Yrw~yhQ}Ag*~hUc{1(CvyvgTPO*PijDZ~&2Hs#X
zh7d(#$PXnNbd--tZ(!>HvLPyZ{770WsE`xAwdbZ2H6<v_9c0C9*y~4?6NcbO-(W+P
z9cJ01nHD=*vqS2H7@lHh$34jn81N3dUCXkIJCPa$1shuJmk(^6*cLM|9`a5xX28zU
zaHTk68n(bE!F<NMgh8M@V%ljyYPmoo(Y=9(<pV*R&5${P7K%V~x_3{aN`DD#?cP8@
zMFZ3bg79det3ApA-)kU+WH~DVJK9@AzQ%-fu2dU(etf<@1S`K+3;(7JLsHN`mthEZ
zLjI}4TvM{;b1!ksT1Q?MXd1gR8b7{cu3R~LhtXjpMIsM^TcB1N1i;O)za|31CV=yW
zMGbhkF^1nQ8(l)k!5Gj_L?DwEF&d~`*w7;rfP_P!aD;-5S@>i=<qSn)W#Sgd!5D=0
z1D66*B#@PY8R2NI1){VGD{vZvx8~>rqRbEgk!RpBW^&zx8u8jpI{4_N`7nd2_CZSu
zD;J*>oSO>O&#lx{jETZ99VBvr5;4nMhP}SP<|vR=fB=KiI}k5ndN9JUJCZ+;vr<&4
zD^ER5ebhBO+j92>z&vq~fb&Wx8eq_C8sLx%e;n4i3qF48F|drC9qT9r)}el~$x6Z6
za=C180#6?>K(e69hQNmy<ct12BL#U{V40X$<uTq>tgNB}v-R5Jrz;nuXje-QU{_#H
z*9ZtPiZlVa4olya++gef`6}{h!D*9EpIH=mJ2BT${@4BIjFzetzW3*t?od?OXR5?6
zY#6<Tw%48QvRX&AKx%kVZd2qvcagXVFGAK?*Y9R`x;aDdbo&QH6jDvMZ&Ni@{a_gV
zAa`BrYN8ewU)#FW@LRd3qv~TLd<7>aRgY5CIJtu6bcj56CTrAPYxD@)38VrA(J*n<
z^w*)UKI^|9WKbRDTcWZZGeYliUw`@Gu-;#j_VDl?V50^sgetD8kz9)Q4Sm%u=)8Mv
za<VX`C3=D9_q;vgkN3~L%A^TKsVpXg(<uj^=qm1x`?R*-L>dS~zwb@cAGI@{Zy~OZ
zk^q-2jk8hNSF+=ibViz4b5{kUrqu-mgkzs5sD9(T{4#6$lIpyOq0{#A=ryZW%UkNQ
zSg-XgiCzQ}%{tszuh#rB+x2V%Blq3G{%;5ojiD)Qx*E|t@){bP!^S#P9#G9Bwu|y#
zI>C`5oE<yO#<spxU+;i9#3FL&Yw?^oUd!XZZ`YGz{q8f~8-)^Pj{VvgJ02yj$}d1)
zOsoA;JyS;zQaVBOqMq$BP{}`Trjygpc>bRDrOBd>8ai3;oast6Q%P6fb9~9t$nke|
z-BZS*l(>i-Arrmzo2dc2MLnynb);h?5^B6yW$fNxn^{js=qHU5?f0$CNpZDK5)29;
zA0mDuM0%R^RnIENG&dS|@qN=5-jbj-oG|0R`jw=8#YEMHo>Qyc_u3mNo;^ru(NJA?
z;<$voY;>-4fuSH_;APgp?_7QP_Z%Nw6nNaae@!fRSyX1X{{wSx)e$k_*};}`l!Zxx
za40`#A7j#V2bZg!UX38p%QDZ~Utr*yo=c^Vzaz(q<C%?43TE#H_6Rd*RCD8KgRU7B
zfpwXrcQ)6R>QjoS0}7Iqes6yf+1xM_;*Rl#ynGpI_JDZ(G}Zs^)A1<<=Y&dz%cpBD
zl#YFMl9S1c--=Y~iB?r-$f>(~vldep#}^DK$8J|NGFgR`sO0-Jt}vxg{i)KIr=m<Q
zybqmBd}ozDxM5~DR|UIgkwdg<eabU^2X4Z<Tvh5#N&Bd2{d6UL+T%{QV6v!co0N;j
zdATOj`l-7njEnB^toj%?<0bl*gz`q`>7zugBLh`cd=I}nP7*Y{dsnTy#`cFw`cEfZ
zvs6zlBm~9iGq^~CbfR`{V;}R(DY7+jjxHImvD2tw#K=f)<6dJPYV-N{;$M;q$|n~;
zj_4n@y~PpK#6^gHwfp)B<LDxVUu!y)uN};|mVGOxd3WQyvK)Q#YxOKk)$oFms?rbg
z$se5u_8RlWyktFZpjsvEVthI^$B0uSxV+kT{Vd}~?GTa5cgyCnzOdyS@ze`nHiE-m
ze~M_G(zykvMdFb+ea*$S)j>DE6PtQFZ@v6H$?@i8wcnblO#LQ%7tJk{XLrOs>vk9L
z<wd?57LNCO)1_whVw!$AFQ2>Z-F>%xN9I0uurtY~QagNqR<LZE=I=9_vNlw-PhhE#
zDu~U!omT0}vq5`^?oH`8$B0iBl7}Q=_>bLth*3KTqhh`6oU*x|&+Hr8B4|{SNB!^{
zmhFMP-9PAeojgi79{m!bK&U=FVtm-LQ_@?_Ydk4G?hVz<xZSu@y}nC1gM2%=Rb$(E
z8>MY^R(yo<c4euv=auAjh6_*45v5$_FK+c9Zfj`g7Y>iQ<H{3EU7z(OOWW}cA?u*o
z^>9IxB2ga#72as~$X?#p-6{?o&y~r%f;J?(d=57+RcfC%b;##>@#27v)vg^3qTg7M
z9kFSl|4D$e&;BiR7g^pYp_WZFNs>G>VeOU)hQEli+I@ww^JH5tj?uXM&V8Nbs-}t_
z3}0=IxdbPV)wX7|Q~P6PI)m84=00}27k~C1u}vFtcjhT(CuvYLh2(p?S-wIpqVoC&
z=IE*Hq|w@e4V2*=<;ve5FHNVsBuQuJl&CM=Z7l9C(48uk-IHKKpG~$Y;0fxjUaC+x
zsj{~u##<UkigsMIO;zsY(6;%>xi>6Izv^en$;eBBCS4{jA#6utg)N^o@YV-BB##|b
zq+U5y9eCtTd|t^MeX|=^tU$v1%&;nncIi}^p8eMql)J)ZlGYMOh}8i*7t{NUcw8#C
zDATihxmV3G0zyji-1YUBDjOn&suox0Yl6$meWhm#*N`>1bg1DDhNPn@FK;J(rKMAT
zY!S}tpY!1`=ojmj;I3van_LQ{XAMBqmmNqkalYlydXjOTLsdEib{TuidY-%Clgs^|
zKcst+11ZtMY8R^6xj%gP@VJNJODY}u`UtF>AvTOZv;uN`UKk!cL_4>EZpQcDf2MwE
zqtRILjjgQ(HY+=7_*nSADv|8HfDvfFx8>?@TX!WI;IZSc2AOwcZ@hN<#aniI|3CZ~
zReAi`3Gy{b{__Kl^fVLxEY(`H&d0A1qyOoJ@v~_@3(z3e1lb5Dp30Hv30t|Y-R`*C
zKOoxhI*>qPfx$RTey_8%u)qY+^c(6%pkVrjp1EL{>iuE9VMr12b#Q`Rj$cPDd^ha)
zt*!h*Sg8d4p(}>oHjZ!px)VA;t~JwOh3i9Oq|eBMtv2TTLEGYG`_rS_<$f4=YXe&v
zTwb)Ls|u`(Wym4UZSp9x2bXL`rL%9*#`QE%(+fxydR%1?Adg8z>VIT}se|1|DcHMo
zLh~e}T{^qrOV9h7(hFs;4ZjuHFrm#l;oO)$Zc$)A4pB{u=$?U1y*YIdnV#FOIRujM
zDdg~8nzgmCNY(+jp~Xv;jI9o`$&27kKFSBAYBDI)hYFZrJ<$i&FQA}zm?$~lJq)fU
zv2(G(W5gE^X18m8hXl>8C9vhlu7du=B3LzWYt6ujp(ouB!`NR1>4pxu<_l0YE<)k8
z05s%FR$C9*%f7Fe2>}a+5a?XUF<BPYufSk{3YVcr&i45S3z?1hhdOY%mY_tbIP1Hf
zROksh8dkj&<$J4Zt7q5KdJ6QAS<sjEXO&+lD_>V`mKIo$COc&Y4Ka^lXbtzHPcAlt
zpn4%Mwv&f2u?>3ULZf#h7Fsd6GOrj3*Z0k3CX-6aucG5C1<XkHPc6Nrp<6IGxI022
z#3QU41uodjf*bI^;PoEKOvk<iHecb;ANYc9AN|}O&fs|L#j`4$JVCNh1Tj%X=RV%G
zvo^`~l-o?#97t@FN766mMnd@u{S?K(AMW5g{|@x~-(64Y4xYIUKHEshK6s}ZiaQMJ
zWcEyiB<sMqJUI(C&^2{6>_!uznv;NbTOLR^IM%9fBV&OIWfW_+PG{}udOemR#+><g
zchit(j`p<JH%A7$6If<gc2fe@s`sH`$eL;}4l>L%`wTSYw-gF=pF!cUg`=oDK&3wb
zm;qrImLg-ti(a7L%5r!rh-|=9{H<8;O_ivATff6-1Oaxg@0M&d>}cWkB6Ojli*e&u
zb?w+j+IxC<CD)wGfbQ;sMqE!1C>ma2Odm8nI-y$Y1<P{mq?H0g0uIVq#l^lKzI1nW
zP0hY%g-p6Bb2(l5;aUSJnKOQ+RWLlrAR8{0dG^|2Z0<{eGjwMx-K-#50B={s-Eh9(
zKXf&VuR&3ZEx53BG4+#re#KA(dqJ_r5L=oQV-nwtO#SlYo;o*t!{ob4a0S5vd4G|Q
zjwx6iv(jFF^?=C;;*+&R9_VX--sQmj33`@;;189VnR(_##FHmalwv)j*-2V(amw(&
z=-i^ZPM5sEeFE&*_d~m9>x4hIbDB37hy$cD)RWsOh;x9LCJrsR$EucC%mMroQ)b?4
z-;8XRJCy45A`U#YkUfD<;zA*y34lScVqp#PTy030T51#wQ{bA=f=he`NagDEDLRBg
z)-QaY05bG-vs~qKJlLfL^}yEj?XpFh{&3qs>lKzxo>b8DJKSbnK2!ZZ|MUhr+(LLb
zOCix<7HaIeZm7`@H15o95Uy>gWyVNnM@j-rlDh0B4zdwl%;EzwcPXqkptB3c0L|q^
z;|#1r?wv({2?fZ7aaAtpa-76rMF`Bql?|LsdKbT5%02lIHF_9O8wBzY&J5TzfaVcw
z>)Uz)Nd{0yVWt@kypt3?K0Uvam()_K9+PV>)8eYaCX@J6qQPJ2pB}Dn$M&fe6zns-
zWSwH`XzZwBqY`l*tJ|5Za@#Z=pqe8LVE*s~m@1c4E$)F<Z{M#1M_>0Y8Mvk3ha#$h
ze`%dhMP&{a=-k@hE{o?Dv}B|6HB@5^K)}4Ne{mbSNRsf+RJS;~ndCxs?yfDe@p?ar
zfovFjS6QizWE7;l00~E@<&Z$<PReYJl>b;^9oXzZB5O1TzG4a+xL|W&h#meeK97oE
zx{SR%)@L!6IGGI08=P5Hv$b-d!Ggx7H&AjEAlTu~gZR~uAkh!>lQF;8UjKn;+zbm0
zY(W;MLr4D(0Jq@HeZcm}2d3Oa{9HtJLmdfP@MujBi?h_~Rmdyh)|q`7G{ENP!OKAL
zL2*HFw9XQ<aSX&hsPGUWFHF9;AF`bdAS!Vl-5a(28Fib183m8!Weey0yJ@nzBZ>V%
zpq#)d3NYYGZEH4jv2?{HC;WfhWB=vm`;RXixxJiQfNTPUXP%z>Xu27-D^4Yj<uf5U
zKa1mM7vwK7DNCOrbWsO=y<ta0QpUAw*Y3zB0SNWlLqcS96{1nE-6mkN)>=2fBU|x`
z?pC}7nxmVO4e$+T?kFu|w6s2L#?3~wso@Q9M~N$(fzWIZ{p-=$2)3F@KpVfo+=+$e
zF0|3-$4kWUxCM?`jlG(%Q)|P;WWeuuK~^^0I4^v|Z$C<^Zlh;7NE{R|3?7j!|8@}Z
zo`ql#EI0y#2Kf}`G8+-+f|Q18!0uK_3j3na1TZ*9U|b-^gi{D5crUW<z<0j?=x<Q+
zAnzk!bwG0-l1nYT0953tU#!Qy3t0~=xHLjUo&<uiBiu1~gfJjQq|TPSIK7;3jc)JY
zvBdeCJH$C_-GbECW@KgIaPb|1e{=z<Vd(#c{##%2OD;@Q%;1~_WC>YO0#Zi@iDL7p
zibahJl1nhmIYL0H2)IR<Yl~h%m<S7|gqFbEfc1~c|12HuN6XJW&mCYOFaWq=u+Fjl
z=pgdp4_jXGLU+}oFfMCwLIMR2SPCK~0&Su|)hdV9VYEON68SzT0IuW@%2GohAz(bh
zK|muxF?IZ?#tw2Gpg47+fjptG(tiMHjan7ccrBbcy1KiWb>xM^$W23`8XFo(Aq4}I
zF0kXS9bBf0X6{bOAGMYz#%K$$I{*$6{oVnlUC$wH7@zHqa*`4)5J)(dB_2H)?Z&tF
zsw~aCj+0n(B^1OxgRzi!5(PUo5PaCdS4gVksmt*Ymor(``NKNfzLkKfJ+j=85Q6A8
ztWe`fC2%a9eb8ou9}d_wQlreWf%Q@nI3_@u<peCgFWyl0vnOlV69|lWki2{HA#Z@4
zphs0y&?J#6Sfm-KPDi2`s9mT1CK^=hk;;O;{ge2-1IRX+SP*m-+<>~#_dO{oX=x$=
zPvT=de>)eiW&<wA5KILDSSk1rv1WtSqp^`|-c7`-njJ|pXe8y@l9L={2O%;(!*5Rm
zTpEOOJ(%}{@Vsn=gHDTs9T%W>amn|Lmw6@Z@rIH}y+hx%RN8j}9EYheH*Qi!rrXtF
zsm7=Wbb+HHAqnpVTs7RmafYGUf7zEuL?G`tK@jRa*fU`36yRis4Dm(>2Mzoji;R`e
z5oW^JdY6Smct!^K)X)N>(HR~y@^qnL<}wP(Y%8g(1sRkTiYmj`Nq{ynAf(M2V9Is}
zvKFX^chdq<Fm=Mht`>MXtkFiat|aZCI;-7SVOFinJ?@wD^Pdb*#g@O_uH^;M*W|yh
gjQ_V6CYIQ1FN;OJKjL?s4gTpLI;!{R;HfMB0rvRKDF6Tf

literal 0
HcmV?d00001

diff --git a/_images/4c35c39e8dcf293ab2362ab3534303852fcba727e88b0345cabc9999244acf84.png b/_images/4c35c39e8dcf293ab2362ab3534303852fcba727e88b0345cabc9999244acf84.png
new file mode 100644
index 0000000000000000000000000000000000000000..641d4e200db41b6735abe123c5a8aa8a0b679dc1
GIT binary patch
literal 55940
zcmeFZWmuJK)HS;3PU%i*NdZARq(nqoLXeh}ZlpxIB&0zE1f-R21q7u9q(KBE1*GGf
zYro&Q&cE~H{C>IId$Vz|*7Mx=J?9*Aj4_{ATAKF=aA|N62n2zuilR0Gfo6z6pghCI
zgrBhTDzw0V(B0%!b+F;fAKNk#{vXFh#n26bz#~QegEBf7cmY2YcULlSf8=cK?q&Ym
z3gKw(?&9F=?qFxZ<Z1QX&Cc0L_?FNuAzmh1cXt;Fe*XXW?{7Ijx8YA^jCqAXFd<YG
z<#fEWw&uLN$yR@2?)iPRtW{;g&<-9JWG2PV4;@9NisNKalE0@V7gyrWwV$v3_LYd7
zY?0b)7SvI4bj1V)Ql@Y=^rQsSPv6fk6TAP=Fa>Awu3ni?QZ|nrO=t9P%}egKu;MVI
zNB;NKca!YsR`7q{uG@7bC@CVpEEVt#|NBN!wCDf(M~4PDI5?g9W&u{aLJZHuoMwaD
z+hsI!r5$&*Bl|SH%X#Nle9N~cD;aNUX4!t4Txjy($qBkTJ=Z+o-8Exx4N!ACuQaU1
ztT3vV-~L%&?XuL>toc+ZLm=?NTg34v+11sRxX(VD^w~NqKAm`#j(uI)=v1#Rdy{*?
z_21*;_T3Y;*W63P`3glkv5wi&0W<+;hn=0FSWBCm&vu_juH$@p%AOd&XWT%{Jz-`R
zk3~p_&B(~;e|5U%S^j5<orB})+2KaIh#jef`#QqLX02KClk<Y~#o^ckgL->%Zv7AV
z`uh3~GaoBb$^s+=EgzPiY=2tdjUUR9ieBn|`E=%^<KE#RI)d}{SJHrh&`tyv)=ScW
z*k!6%<&p{e<{z?(@nQ@-KIQ(z@!g)jMvaY))6M9aDk>_$yaU)cI4{!DCZ&p-11}|7
zT3bi*<U+N}ccwdjhv;zMm6GDSHa0OqK`<~dq?#lDWf!Z=`c5-D>gN47c}md~>Cyob
z8aI<CPVTt<&a)aw#m2|)8Q^b^$(HmoYR!R7(5p0!eORcRZ{8h2CF#kFcrxGOb21ck
z6`GgFROd8Dlg#lzSyMC4s*UT>J0><Z^y%s8RP&YI_<OdtocsIxQ7x6UVorGva-|W$
zp4&e=7Zw;V&;Fqy1~!gvKl_3F1&_@i;hM$W)lYkeE6rp?>_*WEXhjtl+5%jcx-iOC
zf3L2Z|E#mKool*{iiSo)N*Yq6o;L4xdF$~<{QGZ2I!hlFBgn2S=qM;678jqOV__A#
zt$i2wJzNj?NK7I{=w0rA5&yb3p7Gx9d@BhVnPn-Dv@`>PDXc?m!9NtPh*>~@czfnk
z@7LEnyUi^@wb^m>k`Yx^0^?=+M0}4w_MH4(NlZ*M&v=k(<$t=GW;@?<CnPlV@6t;~
zBgfkFk<EYU;;t+-BDTTr?q@DX^S8m4AaI9{!XhFlGHfnR{!$4%>5xjBhFjA}7eG1q
z{Y?S>87X<lI+!U&RKxppq(I4bqyXd1?Pn1_2fq!!eR5mVJX{|^LoUMDyZe8?OYSm>
zivF;FGmyr=5-T<@`{W6esqY5+)^u&LO0qyuP*C~fPlQju#K^h02pHAdhgY^=HL1jU
zy*UUt9Jwd%x{UtOaRw7!rta3{cUx(DI6qXkwxy>>5uf2M)xSRz3Y$MFs;$4hVcG1l
z6|q?A3QrAae^jhFRb}2SHMvCJes#vv<hdQy(IG41JWu!NorYXspmcE8w%gin#r;gt
zz2oEFSX!~2CaK>;x$$9P=qKmrtZZy=`$RY%<S5F@BAn)$tZw_CA8p%CRpqbVzUJcM
z!s&iVbPw+B;U_1Px@VJwJbL9-F+>pmEj|Z3jqy_JP5lD{-Q%T?Mry3o(J?Uc)zfZG
zeR9s75P_Y*#>1=KqHXrtZCiL>PcNtQHSqe<YiVtb13pOC(ilQXsZcq#XVNT)&*JIR
z&mT;hdk%(V_(dImT4wx)?<nEWkif4~2nq^X^e3@*Ei9PP$pp>{#pdSb+K#?`Wl(D!
z8yOvaqq@3U9-Enh-;`G~TY{aGl(gsXO0SZwZ3%=yxs`ObCWoodD!HKli3@WWvCP#;
zVacQ7IBX)K_Dw&*42|WE7ie8WLl3fKf=u7OeH$4QgAXrT`4W4hv$N9zR+oi^MP5-^
z88^9~?Pz<315P4)L1E$Ni8B4qVK_Gw;gw3hzD>mP*oLV}5DZExDv?v|LG6+8@svRq
ze_u{DxE2!2N5X<5?;UxCgVwbtkZ?Z38RTSPWeq2mhx9Sh;QCZB;LM$bmX-&(9Bl#T
zJrFSr9?2w$^m_)UC-gk=^pZF2CqM8(Qs9P*Qq<8&nX0o(j*N;TmB+63+%`gPQ?+?F
zCh|SnuCM$^sHkk~P0o(Cd2VW??^S*pL#|7!@4ucu6J<&c4&~xOS1s!^pDNeKisKWk
z46&JeAhwjWv=XM8Jjx<py&@S%<+}mLn+yd7<=8)7o@~bq)?^v74%x`ahz0B*3n!;a
zxpAX5{A$JHPjHHYA$z~I%E`*2d)OWrQ2lK3gWoO4ux$l$p_mpw$_-g=-c)v&Z!tVM
zJ$(gfJ32$yCMrFhs@h{y54k*nr>m3${*OO8kYEupkh!|LvfRCEj8843=)OKIxBkx;
zcINc-bUB>Tlvp7Gp9H>^A-oD~hH?r_TYWA5E`QeBtcK{cfL&p^bxX6{q*)I>%CKUy
z-8pw_%VDJ1%lU677Gq>e3Rz-O63d})>qOf>FGwqtcoe+EaEG}jP3|Nn&7S0I-?ML|
z@taYgVG)wSUF*#R%{I6$zt%S;xfjns{`c=+i^&fr1(0m@&Ie_nibe(pBLp3Pija_#
zvm<X~q{-u%pzonATr!ja3rK3nQx^8}Wi))Yxll$FF9$2D0<tm{IL>@zVP{u*TU2Cd
zPb2Uo*s5uGSPM#{e3_5mzd!6ZALRHC(Y(Z`X2Ha#V!CD4#)~}TuWoSke9xBTTZx|O
zWw4x^y@PqUJ)@r`;i1;k+ZzqL7AfKJ$Ka!#oZJiVzn>{TKXz_qhf~u7E3ODv9=%1&
z{(?DL&BWD=E&`w0;<m%yVy8t<B#C@v7n)tRgh$WM`ezmIs6`#p*C#7;R~sSdsAK~D
ztLwLBK0SbtznA{adZ4PLgcEW#+wZ{)R!FuLSBty9;qQ=l%ErcKK)@*Gl!Z(uec!&d
z?p|2e8j9GDt87hGE5Lz)!x#>4+5iy1q`2M9MJDi~flAuHVSTE)P(?$dE@o(OkSs7T
zaQS#4sC{>nbYvyASM6DRcXf?^rRj3E*L>sJiC%@#E4U#ef=p4zH&9@!?_Qs$oj`3s
z9(wDc9LL+df`Si`hYgZ49&?!ObWPC-B>(bv)YASZXyN5>Y77Z7CA~h_PP~7;dYZtj
z&{OiTkauXZ(v1I&sG~iOwAa`tXS0z;H=8nZ{`QL+@734(<q@rCf6A62N42j#7ZMT*
zgWC`fIIDv&hYTBa_HQ#9&f0*;!-o%HB`Q_=&bDfY%2Ow*pWYL+`A%0xbIlNf+{!Pp
zv^KYgLc_|R_Qv7J$jFe~xbc)aH#PO<pC1*^3<LfBTOetF3OkgfxMgf$DRnuwK2oUk
z$!V@UVtb+84Dxq0RBUYKouG@8W(+)vNJvkQ3C?Z$&W>iBm0(49gRZ@PY_|sLkQBJk
z&KpvCJ(`t1V%q!-3rhLo#S6U|6KrNA()zp8yo<{mhD;_UX6<c#gM)Q3?ax=$5(5g^
z_TKRNZlR;28%Q1R@7HAgj*E^C9kE4by^42svkii&e2=YfBbQMrwH1(5xluhXgl*rz
zKq&8k08MN|z_}=39TWs$h@PqHr+@o7a`SY{^o||Rw6dU-QZ!D9Ff*e-@O3o0ttF(U
zVlt!S;^Mj-{JxLq0Q|7{=TB6V$L74(G$56ig@vp?D$R(zyu1cn0H-0VZzvX#a;^2Z
z;kT+c>YNuCZrhK0|2mdajAvxvQNK2zW?+Z|Of=_V?BsMSDJdzc`~cw86R4(~oSeaW
zW%V@2&L77<M^n1)%-(_W{6U2ecK+V#*Vl<zSup@p%3KL(M4&=o?VK{~!r}xbp0-_`
zJ^S_hch~3780O~Y-q)AMiLYN{AYtRrIvl9!zSWaa6CX@w{`j2kuRd(^_bD^a$;nYs
zSFgPr9~s$EpcG9^LSjBaxBvTFCd3AFK|z65p)x@V*P|~&AE3T-Li($&++AG_A0Hn-
z|8c<yFIe2!iG!s5IKs!<+d!(_eS<WNczwRLd7|ZJ@<!XFnX-~nzG1DkUiTXjJ5*+L
z<Xe_%v(vUl%R@RtR%@5T^@m6h2=6Ti!M!u1z8Z{zwik4L93=ky7dHG_2ozqj>7m_E
z@lt(FjoyD*Vw70{rC8mLh3%eruZz-A<2-8@NPvKl)H_#GHSa6v{Vv1royvwIANkVs
zyNt}>_gtBHBIy&vZp#7TmpJ<98mocji_-%tF(<kh<-&=&V$JLj?NUW;ZPG(uDG{1)
z-R+6Z;|PSpOTzHxW=TkMF}UR1vK4xTUc2*3_MDmkc>Z=1$t=RLU!QHPooIl9SP*o5
zCGNRJg#>@62Wu-U@iNK=22-r9L@d4+P$!?{q&{7hnM~vz@IK%E^!MkpO7e`4{-^u#
zb^Lc@AYJ4eH$I1Rh(#mj6tlPIGMFJ;Jau|C@5ZRrurOaym@GN@n~aR?$#9+=a&N2$
z)6H=#VZmyCQnIP~<~2F~^(xewZB-9#H@k7lw9~SK40eYFtN@v!&bc}I{QUf>79Zg^
zV$N|;Q&!;OTXyfxHF+pNC3S@@$hzx8>fC-Q`ty_Xwd^_e-jhw`R4dqdOxPDLq3e}l
zA418tJPtKj?<F{Wjkc$80P&U@*K(S7TcC~Nn4OywwCKS`VlCf)e^k`exP|869J#I!
zGefPT@bU41KH)4N&2LS=x!tVXho7378URAqjB|S|glYk-z_6`|Ui%aVCMF9H4<04I
zY0GYnX}~|-!K^zsZrs#pc;;0GJM4Y%O|07DGp_sQ4`goT{wU4jD*?+(`Iv!cbG$SZ
zGS-|&R&+F;zrQ~O6`(4#wzf7ef~0F}N#t8|bK^pk^X(5!PgCEzb!+eF=*x7iO~r3X
zx8H+sb-l246*;+5eiZZb^DeHgC|Xn^&<VWwmMQiW@TM)4+E;byk#TX#zP=L5F;p)A
zvll>Y^nLwWgR5L`KOyh#UZ=YT$z70Ww!ilhnlC6vW#(QId8(?a)%fb_>I%=F*XZ7d
z+iwWGY_c){h^G6J-s8_W3St(}O`mAFcF|X<XYX~(ewZ8!t0qOwIWy=&o9BA6XYMfD
zKq8S%2;0u>$-vLgk6Z-EF<o_bV^sk9A^45HEI2dzC3rjn#ARt?qfj|Xneooza})x1
zazf@f9AJ;(HV=C^b94@;Q3VC`QQaXJcvUt-3_V}ISZs_IiF@xcMG(^aF8Gw`SH>M4
zx&tPnhQ@lTEkF{9+{)}MEGz^74kB9-WCogTbgQ-+px9byzs?z`FsNoB!^H@^{HG6?
zVeXh{a$<rUT1Mp21MKK}v}G7WcZW1CE-uv$*{&$Z(wIYK7k61i0CLz{%MG%C)<!9H
zvh)$hV74SBbcP>e;-jOx0ZX}_9on(oyorQWCs5!X!hN5AU}W+2-7~QJrq}xw0MAli
zB5S&+BMo3zs92!@`~cNqBKDvGdICK?l#v@B%-Z7sf;@pcrxLR4Gf6|H+4J4@cu3b4
zUt*|%evtL|zZ*w)Cj{rlP1)ix6a*##O&DP6r_g}Fj!Cp#9zD*2rl@V8m6e4B6_4_k
zqJsmk=g#aa<clCY^~T(H=jq`9An$tKdnFdjXs7!<tPC9;U6u7W8bCoS0FvNB-%MoW
z0AWG?{rh)hV+@z=dUfHWktI$7D|CGplTca7-__ks!pIm8I{{c+TtWhgR}HEyFpw|a
z7-f&4m;CX*)Dl+hVY8=W47D&O<t<}bMMZSDm!}>#EA1!Vum7l^mA*Q8>afs84;2vl
zh#m;IC=trbW&pY?`>O-U4jwuO*I#`k8rc%$m2GG5*^Q@+pb64y_4Po;A2QK|ZTp=<
zckp+!vYiDQV1R1f&<kkg%VU&TcOL(T^+cZiz~?in>)2cB=IafHp3D5(o7<_3>c}=n
zTvGDsV7egk<DsAmn0KL<JylgTca!s7j-e7n#SKRtD}HcKSsC-aPH6~)WFesC<LpX+
zpi~m>oRCUhUR?N#IL%T)gxA+qic#TH-FePt^wHr}j+7rN0ze+Jk<xf0f&&Pm1THIa
zHI?6tf%RTo52SnKkp+mi{$Z{THN5@eTdoWv<S!+_(hsXGl;BW~{yA)9Q$0D`tQ2&f
zzpIfUgprYvvA4h9WOMKuKyQf-H}WbmLs94@++~f8>6GK>Ucsg`G&bf#4-1uK7OleU
zzf@%RjiUD}rI|mnN$G~7TIITe<Y4YzUmT=9yYG9v6B`nO%4^a@mg~Qd16fMIew-5k
z!%EJ8|M>QJxuGx;szg)p$pbQH4wDVXCEo*JXw=qsyt}}-v9XaZ_!JY$6tYpjzB+DC
z2LKb0MnLAd8vxh@nGx*kk>r!y_0?s?yTAMU@{lH7Ab8vX-cDIb;}Q`O4dzJEn)(05
z2s~NBf=@)I*ABoEd}FrtM!dj^Aba%VojJbq0S>sxzWWBdR3f&-A52@LCQbbapwTgC
z^}P)!G-_k4`0<-mUmqWHxP#OW7caUl-JsV=zw09ebQT)mAtXnD8QN)onj3ILjf@@S
zog>+~+FIcpznweklv`ep;OO!i)}SA(4fO&9B%-9G1o9*GWX(UGK?-*65DRx)9$O<<
znjSD_h+Y|Vg2b@5PUODOVnjU_OJXyJHr$}jmgL=o+&JiB8k(94pdI$!toVeusCV|?
z@8<weB8gmVs8NtvCciu@=2T5_FEoM9gT<HI10oig&7g1IGKqQ(+(uYf*b{hDuh#1u
zQZ0aBmVm{9BzWF#s;8oY-Q=|!2MK37ucr6LkR`I+8&Obw@SqofcnAP`%0{0mDOhea
zC?b3R{(Xk(A?SDHpz~e_hrjKlpjF*(dG*&XAY~|EZ@S*9CU2R1zY)qDjU+!A)F7Cb
z09`^KRq%ZE>(MPNIEe-gF3d^E$;h8WK6-L>ar)Vl>%nJQQ3om{xPk7Sd?P=3C-xi0
ztN$2OhsahuQsi}!Ob;uFQI2o8Hvl#RGH=&U%pUG|KjWMAY+C>=#7fcV>UMHmZla>%
zP{gt{Bcp^RK(mduIW<y}kue*<2)5%TiLC?PTYrU<BG;#%R}S^QBs2u%n2K?Ba&mG0
zH8OI&`gKZ57}R%zxJ&n-ix@T*Us1B-)01K5P0a%=;5kpu&aRsOzzbh>ymKas`0v9~
z{Q{(GR`8G;z`{an(f2>!)G$Zz{?B*+`=kF|iT^zl|E<9PhQj~vP<Y*n$AnzN0*=hi
zC$hO!UikxD!9lS!RKlwP-uinrUWT6C!XPmbF#}Yjm-JKw{i^CoCn`4cmI)-##JSsc
zR8gVlaZG2%WP}bfVVbc`%r}L8%v((pC{RGwhfb#p+H@3zu+8@#kblfYXts^X;THp}
z-r2|{v6YbFHgJlRymRkOA112uS~2xYv!+C9T15GvodBH_R#wJ?SS-#Bl((=b$S#9R
zt*Mkc7n3iX=rAnNn(0JaWy7)!CvCs>@0u@~#z1~3D|8#E2wM>voie>|ZJ(iQpEm3`
z7+SLas)XKfBFo2>mrBN2rj<W=-nwP-#zX)LlvIe=yf+#mZ*tRIG;`aS01DPh=ff35
zeF77XEEn6`oY-nFRnJb4*bcFuySeej-ykL3GQdH;OrbD2KNy{zO(h}0@87h8_Ry`_
zSCI^zs{?6#@zZ}Xpu{i@2!UFdgh0^|=2?7P5z;zPgO}3}RYpa*kOZ2C65~c8gly6B
zt}aM<HLA+f_NINI)C|_t6Je-E{qwrhSnbrHw7h-%$*I#p^zj=UAxn~xR<;F|FTXKp
z^My0>_%_2jpJH`A+_c@TiB#*Ep5q{9%)}N(Urr%45-(YrtA_eigBK{wlsWa$5o+QT
zqTVRhb>&MwF#1QgUX}FdlpBzI`0yc>0h1h86ql)>eci*iJo8%0zf7m+eg8ndFcpG1
z!Tg8$!5_!;qq$=5eBDQWX{wfP^Fycd92&#@)>P)oA0Enj?cip;_nm33)9y`Bir*S7
zQlF}SMtQU|7tN#pp<`x-w!gn$-2X(NtE)@EYsc6xO+|-AhnpNgc(}~<#SMfDbi?i&
zqZn3SQ>xvp6c5Fz#(f29oZfZcAC*l)5ReTQR#9pPv@x_cwS*@xVLi+bNz3M8uc_en
zc+aZ(T_H*k%eR)~DU<o?^(AA-bIi_0XBodV{r4vaJW@5P$v4eALogbiui_%|fC)}@
zg@p(v?lLVRdEEeq(|{oe)Et90|3>;U=7L_;op0_ZoOQ!$YPNa7@+y5on_5h?3H-DO
z#Q70jjoWKQcGfZDVAMiF`*2cR)fbi8HHqzhX{`g?9{7fb!?#SD!updrR)A8~>!##0
zvh>+nUylq42^qZ=TQY9k<lg`1P9{F7)J(yU@#lo#dGAWopayjsPIinP<K3W{^jPlz
ziJET`medDND1>4#kJQv26z^O%bv!J35g$tHPd}T8un~9Pe0UM_fI5&T@8v#pe5@iO
zWY2z-SK~{<3Ip!9yKMq`RaRD(TJzlk?UmoZBVY*(b8IXvr>bdKYKqN{7(ORVUPot6
zQ#WKM;<O4VP@vxI>w8I}@6OeCZPp=#Bh026qFG$)HLb{Qv0N^==jr*f?Cp!<QJfrZ
z{c@A@+W&jEvcUTZ0Ab#8c{JnV?v5m?_r8fO7<@TBIYDkZQsCm&eIEv#=TxIbK@uee
zg|!bI1A|kqT4bN;kIJ7RS5hyR>^(^f2<%jDNTgIOzp6d`<WkLq@!^Bj$d7+_8o240
z@u-7?wY3*7aw)e=Bp$xnm}uwu%1yV(#I?v<>Uvwccb@&VL<BE)DsW!VyH*EY`u5et
zfB<=8XJ-dwBSH}xf_EAjA<$7X-?>A*w6uhci))eg<50uJ!z0|N0u%9TZEdEJhK1X4
z?YnVLB$0H4DOG-2Y@f(4c?357FwSb$*RF4!)QuTFT4m$j)w6EF2~2V*tsj>iCjTEV
zK!0RHg6I)nP6@ilpK~D{dvcK^R2o&3`r)bOx7xkVT{`76pPX66#3&I}z|n#*f`ZU$
zd~Sto<ab+-8IYY1ND(oNG7Lyqd=z-p^v9nLn>isP<G)U-xmlrXq9891fq{WREZ+e_
z&VIbyl<0p-<<ELNW!B0{Cf^@}S)z=-$<=#5c~nE*CJ0=NnXiu*-#7|tCwe7_!x@aj
zts1_U#ildbz^cEVpQ__s7)0rpcCFJ~b%0(F|FWA6XFXQgSkjMNtiY**&GJ`Q%zoQy
z{)D)EZn1)%wgiwDWQaj~oSXGwbp63ozcPn22we0e@^WlEL>w7nu5VEjybk?2@4Q>B
zztQ-W>Uj0>$@3GTKGb2{&{lPBkBd~L&rL4*RjU6fv>!|)%Uq!&5Tc`}ueKWlsY?I_
z_D(@hj{;PHXYty(TpXHN9a(qW5>iso3rDb%`@;qYHP-)~VZAXc(l$iZx3}jG?mCxg
zz1SaEpD3dQjm;!!RpxIPak`KduID!Qmf%MHq-m|2lA_MfqwRBY?r>66RC$C+Q+v#<
zP~7gZ#aE#Mn-!(R2q()TIZnLz=A9%Od#u|jYDEFjK_M9gH`y5RUJ?g82q|duB&H?3
zYN>Ph_121YY2QQfngk`(FoHUmPzYZ@A+}5Sm0EgYQUYobo7{`8pad&K@LMbmi5rJC
zgdglVIN30S@EaS_X|PsNZs-Nok(E}f_G$Kq1$P|e$0lT&1uH=Py!hflMUQT;R~98i
zV{^E^fH8-4*Uwx+n2UZ1>nbeh0D-^pD`KzDy^djiZr<v@!~P>X+6_3Cgq)l>U@0Dd
zu@j*I1_6y#IYt(=K&unTKI*OCi@!k^`WNU{RD`9WGA!=s=qU16>gq&hTYPSi^XMT}
z@O<skJUz-l+jEH7$w^9}--b8;xd0XCE?+RVkvF!XrKR=6#~&z=x9{E|8NCwSG72Ci
z!{PVy9Hwhr+}uK;Qo~j<tA<}%swSa_%cHzj3rC)4YT@4*TfZ!UK0EoH=`3J7%;XOg
z@7-jjL;O(*3Hcwi`DD1i@Pq?rt0~pnxmxa7YL%V2c-=xXX(P0Wv2)Q)eZgyq%_-pf
zHE7%TbfrHfnMlEiP~+5heXJ(m-BVF>oX@t!IDc3o&)DTY+odNp<?oB&L%eF@nqUMI
z4_i8Uuqwqc7Xl4Eg)~n`EdfOqlT9V{@jyCro(?xN4vMNuswoq@ih$fL<9Eaf*J7=_
zg93dwRS=qv*91w%;>NOzY0h4SvgzieH+6nAtbF8RT)P#~JuMd<>8igs^R9a@mG1BI
zvx>!Y|Jx1csrNE)SM%JxecMC$*F~ut`>C}gC7$dJ&aV&OBFLdP1w2)>voY6YDGXQ&
z9bsv0z5J^`d40ST{2@pL0z!QU@MMC(d9?ifLYum&(`O_nhZ1<U!GTm6ze_`3nyztk
zK&Ey&ro5c@0{9q}`}f0P<=Ir_r)#VpelWh{?duDYJ7fK`$uB@!<WJaN>)w~g-kNKs
z1Zw-GT9G_Zvz}X1FQ%ppEG;b!vgw--P}m2A0M=CdpFRiv5DU8Ay}i9);P!#GM(V*p
zTy2?ER`yPVpt88UY+-nP=g!c)Qmf)n(?}>5>+?%Nglr5KUfO|^;g{0egeUv)<lg*M
zj4dhORc=$rGOG&U-e@9R%-EF-o5r8ERsQ*agZPdw^&kdjI@#B*pqk)^6(Q5M5yS5v
ze)J@_RGbesJk}o*#zFjQ-x=UiC1ndowT~rbe(EQ$tA`hEgcCX++>@?|{t+kQL9Cie
zs}JpmRpb2US|~z!Iy{MqS#c$Mww6{ltn9bx2U08L^o$tl?D>AS4b$z$2hu)M=9hEw
zr5OCum8m{(?JW}}DWh_=d-~ZeQAJCF=}23|pZ#75y4Lr}8&Z0);rG7bvVXGJE@9%j
zp;n9kYrWcc%1cXy)O7nVM~KbPM=!ggDCX!W4~U9^m1s(S3LHa(R=qtE#v!GuBEz3g
zxnyJ*LG}zEu?_2d65MYouc=9hBw1mTSp@~F`|fhT#)Us>MWjGWX=-YE?=-F<%LCYK
z(rjCk*bGBN5Z>P2i%?Mg^L&8siDFkzWdbZ-{2*7^+WOsUj3zKqBOe_f|6gf=UAtMT
zb*~Gr!E>7isdyp6CwVk;?oNN#C3$F<)Sdj0oKOI*Pe;7V&?>;S{f82sA$#Y-G!_d#
z&W?wFwbi885!a9JVZ?nuU9tNgwS{O&%H~=H->ZyMF|8*`V59mLJZc+3P0BWHoNh`+
z8rF%fLTXt^?enTs4@KSN2hMlh*2VP|%w7nB=WZc|pCq%mUx(_;k~iocwb)4QnC>V=
zG^Tnf)Jxpv9a^|tcVz2fXW<tks^2_8K}gSwQHoXeunM!IaZNgRV!t?9N8g(9pi(uw
z)2B~Gym(FbJh?Vs2*ItibUBKS8<Z!aAg|Zu=sy`q7e-t%|Bj(4P>cq_%cq><{!_?k
zUFd)*U%){P_2l0g+juT}&YR;CCaRhe0|4ZRU?55J0d{(2zXR@Ss?|>nDBlRLd7omD
zfm$BJ)Wq!USSU5f@C3rGclmR4Vopxr<OUiVnjipISZZ(<pumpdhA+R8TCa1M9{s~$
zCy)K>*RLnQ?-?|?bA5D}h8|?Rgaj#%(A>6{EgZ=M3mm6e+r^FZHk2p7`Vx@@Z5S~S
zVBw~fy^y4lo+ZCDWyPYmcXfs{5ZUNt=^iDMLM7`Wfo&?-@!^%V9KUEW-}V&e`F4yQ
zci%u@hoYQj{^%rQBs$m@!8XJ(v<VyAtYr20dbbbr#xH*yc`^hh`>5E?OMy2`2-#F7
zxsawTo)jySMdN(BEPLJHPTnVdV{%I+$`ygzuBV%GE~U#DYf-COh39xKE*`f=j&&OC
z*~`{6KYTBlV9V=s?_P_o@^un=Q{~;oqL|%=>XFLu;@<_c(0&4aHu@t?>9EE<x~YF=
z{L9V5e^r`Ou|s*g*-%b#QEHnh<SN@E1Cw!s4*ly*_>CnfUpNHZa<{&rqsC5#cyoeL
z@Ww`I{otR#fVH>&k@DcGfh`@q*5>IOD(P)0Z{2<FjMr{N&L8Mr77lnapIdV~<;J2Q
zepwBqFSt<6|CB_#mE%-o6yK_{ta+Y`<-YQT0BL#xCrVglrL*dUQhcJ_*gJDjbwQvw
zlPVU>{N{Z;=LII0TNf7xLnY8xZIBy8fk7h%n8+3AQbF`Z!Vt^prfS(Ty$Yml4)}S6
z3O(c>Kw-#@eEhLIo0~rX&=^woKsrZ|Ix@t1J<aa&a!5h~F_N>DU8+C@t;(9(3^Mb;
zO1l}+O;u7*I`T$riRtNMK}-0$ldTEN=Lz%HuU}!{+mg}kRc1I~B%GYAwHg{=WL+`F
zXHs3SdWL&<bCO%={>6lM%&r7Q<8-3{4z*^ERn^(2*TWK*&F|*p(aZ&8gZ9}5?Mf3?
zp60uFD0LFiGT~!l3o{J{8&(GY(nUF&nh93JP{;{3TtyFv!&yCR51Dx!C%f?nRmq3Q
z``D4FGKcvn@bR&T)w0IL)lBYGlg%otS+sFNMs)5yJ_T!kq0vI8e0i7qvrX>w)~zH$
z==j8(!NCkd+U%^xe{>YxpR<rIJISk{LLX)rVf>5QWB9#@i-(K6RV;@HH#u3t6N(=T
zs&w>p0<P1Fb+!w$6lII~*vxwanb)7IyA4x+CwoYU-AfDKeeKJ~p5L${HDvj)y3U9X
zB<&dG6{*U4ZSL3UNOf=fXE+GSNOKO@DS21b9Qi>LmGs=|7%9Bxy84w|@!ma!BJF&T
zGQF3=DHnhL#uO71YiMrn{`QUFHMedT^h?gWHH;a=sG#G03A#F9^!B-IHfw{=LPbR-
zCMFKfS9r-HBt!~8sBdWK9%x=iy9@D*K^G);c6I`p-@qU7apJ=3P0{=*{rc|p=09J+
zMJy5$G~E#d-GF$InqfmTrQ_8@8o)$gWgswa^QQ)V>EikQ!x)$2<8bB+i7<3=pZUv^
zN1rQkFJ^vx?%(cZLa?#VeMz`tt@a&A8s$4eS$~d~@FpE~_q}w&0l~0$$_odf?({bp
zSk7L3rxJbBJ_*932n0r^sI1qPc;2K!DMpI4bjP)k4fgArs$S`Ow7PAh?kaP!6sHe+
zCW4e|yZgb!31qlBSBL$!mqI=l?hcpMSY%14auY_&MqhK|>*y*n@l3C+->x>QI`^FA
zj3u};V_Mfz3n6ZxYY+z`SiUQ>m{H|xb3$J#64l=Ja_)}$^Y%dQSmhhmEvM9Haoh2e
zm|iamlUG<LE>-X)L~sO{*fn492!=21G7oLv9YJPbFW(y)_wOr(;}Ur$T5D(wKBcZ9
zWVjm%LQNjTDTqoTyaNeV)L>`|&E)NN5M^a$MMKJI^R0gQ07Q$_$iWz7-Vu!Aefe(^
zS~>IGPueWY%gad745=zWJl9V@XIGV%4fZ&Q0(D9wQ<MOB77)-&Oj{+O4M)lx<0TJQ
z7}U?yO-&gtE^uf1IM~>b1{W%j41k~5BK-(aiXh9GU@x#ySk1E+xZ1-*;4}?PE*Y23
zniS8}*$~YR5+)$1%)i8*WEL4(SC49sj#Rpusnp2UNYj4j#%d`li41;`Pu$DTbp4gU
zyLBFQd1_Z+?RWRXXJKEchjsX{W*Z$N)d;L6>|Hhvr7a6daMy&ibV|^v@U?VH@jo3$
z7__h7_+fl;{`#?&sUapQTZv`(x74!Hk`d=~GDLV96V?7dgMZVw7cH)gxG|ViS0nWi
zs<*Z_hWk57CV0*Y27=`CU*PVn&LWD?duxmw9=8z~mKtiQg?p$aR5=|P`OA3G|NQqS
z_}He!x%O;VT@+1T*@p-=%WHnG)@>6N`sp9p{Jr<$=;T0P-@P*j=nsJag@#1}iXeBr
z6eo0KL^2n|U?@aGfW+NC<?}U-KOO`}7G7R_upI{!R5JQ%fj)x>1_mhQ<e4r=JZ1s^
zCKQH_dl`e6BqeFVrdLM;YDpY;M3zCIHfRrQjeYMgki?b`%{o$*L6&9;KBGPBA;lqt
zn<oQ8R$E2D>i4|In;9|fgw(7Ejd0W*8cIuQ+`R*_0kzt}^QObG-KnF6a9f1FapJy~
z@@gcVx$Fl+<=Oc5LSuWHC#i`aukmM7Fgq+G9%6~xTG<+X#1u-(d#{J_9(R7YWtcQ+
zC3yjV-ziKrFRM^^JPDQhT?nZU0dzF?DEa13FQmpD7qIUgT}4c-dyD&}aJBRr`H1-o
z{527uD=OU>@*{jIJJqbEu2%Az+q=|cseGJl9!(=3JuG;d_vz42sddphNp9T=1D4&T
zUA79ml%Y?-ezU47QK(c_FYsh8;}++cOMd*?Xs6_8$Iz+<><|WGCQ>Rs#y13KDi&zh
zXb6KAZ&$li{qps7TmO@#m*Q@}W2~mYn1<AA!R6l7+;Y903#J`t3ixbp=CFG?#lb${
zFf+891_RXp3!TG)*Bg+3nK(GGfN<j~dXGDqsaN*dCO04gEPhX6%f0_i=&Nu4d(|h+
z-8zT!Zd_xXkWOhQBAatk6#M=D56XYnHGdI*#C3a0XALYfafhpQ)$f{3E@+aYoy{0F
z)L)9vW}YmKemZw`TvHXt#lclY@hWjO5$%{6q0X_I(+aYnQta`Vop;5bkokkrJtCt;
zOdVR|!PAlNZ<s<qP8${ELz~DEeveweA%iY2iA;<0g_#$l(q`IY3Yitw1U4K#36Xoz
zghYY#6X(0JFVS(53aTEo-_3u?hU+IP^3~dw{We?a-kS=+P>$tuH1>soqO8}L2OFU!
z8}=bKKKIyWQ_V#(T%LTt*;2o;u^XEPP@Y|t6v@LpkhvrSKiQ{0S&MPw<2r~AaAK7^
z={XRAz>SoxLFjY=l)=Wqfh;I+pvj@5!iJfG-dlI?-m;&(+Dt0~yT3um!OHUY%O-BU
z@<^ZqpMsBe7?ehzv8Ijta&mH|5#`_<N7`{u7DI^;2=GG}9sN|alu@g|{BHVvxR|3m
z!G|_uhA98yu}Q_de)WNfmtQdWZr4A2SB{tZ;qj9+qj2IrX<vpkjt`1Q|GI{@-0RVF
zBv`+y$wuKuioENrxAxhR__*!t_EElq>k^Og$l=i+51CsKS5c3IKgqGPak9Obe)_xM
z-vA%;fRN|ao^aruJ=&~+hQ&D|%slN?wa@&+U%1b_nR#yCs38{U#t(IfAjGsyAsp_%
zX_;}bl<)Zhk&xwa%#h%Cf3yh^r>VttXnS76ulFGe4e|C%amv@t8CnGjsRYT^;?>pF
zIS+rN5d$0%NM}BHb?YBLDur1Bq&SW=mO$x5HZnS;+DL;QSdTuc-+z#llG0A-4~|R(
zu1tD&cXze#KYO5rFz1@RdI70zYv*=qc?YFria8ri?0u#*Lo$vaiK8IAedo{TIoXkK
z4gmktYK|cW2K!h_SR{iC_Z=85SgKYn6(Of)9B)F2a}<vhvL^pZHQwO!SHk!)<@TS}
zx7f^2)HD^odNn81S%{_hvg%lq584y3g`*c0((oyJD0Do|KQVAS{$7Dy!e*DqtCN7>
zTns)5;oX`GMC9V<5+?V@7(YV6?+UxU;Mh^~`e1LAhq>r1^h2Q$8V;$3wXkKltdShI
zMh*1`8V()zTQ7A2L`Ml41^RDlYza5%*eB76xSZ{&la*&;ecoL|A^S-DRegZ__2efp
zB+Uu9MXw?xIy#!&;W8Qe^!dPZNB^^p;=cZVu8BA-?MIJXx26Qahl2`UywJ2X%C_?z
z$(F-W^&-9QG4RH8?~5}rAz)~u_eaI!iVv<xqbo6Zf&e~UZ!%yrrpCv|PaMFI20oPl
z0$N!#1bC=0fxa&ad{={He~jw6f#gO9R5g%c!s6SM0nP)v1y6DDOmm@Ud7k{fm{Q_(
zIDbUE8Agg(nenZE71^X21(YsaMm~PjUOh{q$E8k87`JCPfgq3YjCfQLHEi>{vtF;?
zEcf8;qYVugVKiLC9bA;ck+8qJnp|`oTubH8!yLGSAEpafE@F=pV0V095KT1-xSuK3
zbLzH~IwVuX)`{OS(MjF>;4=8deJqy#1dMy(xaO@p3tRK4Y{fze$)x#71OnImmP^zd
zOkX|c3Wp!@+`UVOTm&%WEdBO8rtKpK3n83Q7cZ}uwQl|MRC{sYbwHYCz{K(y1~&FC
zj~5C7<0DlJv>&Xawj!y$0P4W#_>%p}9f_-n?M-mK3IMsQy&I*Q#HRDXczWOeXlpt(
z7y|)WR?gHkN7sOt0EkUl4Gny#qcLFHY4%`$pjf2twMVRIdW%f0qtwnsyj}aScJ1P<
zyH7lTz=(5kIWHpFt>3ge#qTysZ^+BV_%ml_pN$_~PUGK%o+K?TZe!-=S2902<@6T8
z3x4XOk6DiO>tQwB?6;dKVwsD!sfjk@j{`7qu{n2U$4E#&C@PzeMKP*XwB@umcgShW
zad9m&l`iIK7l!!Vp_4wp!F{F|{4R#I<@N`chjt^4a$e=<P41|9SBr${D;DNqG`@Yq
zauM$b3+JA0sdgtQeSU9PdlR@KaN{=mY1Z0La3d)Y7<FNvHxGl+Be2PYP?~z9fIcCY
zcFUy7@7T$>)t4MRYQFEBPM$prBGN`WJ;2%O0>{?4*^>Yq{K^&<EXyk^C<vHyshd@4
z5CL<gckK{(=N0m+px10_Li(&Aj~Y}xkxQgiN=t!hq7KN%;8Y3je8M?uB1IyVE%NP+
z{BtdZ=Q?sTRfJ6BmrEOqHk*q~Y*3GGhh8o+k{Vn5Grry+q6-Y*u5q?zzSJWew3=fs
z)paTPgO;Y(SyJuE?caM8A^$BxgJ@vF5KSx`$4XbZq5mc)8;)90O;s%pBJR6Uhg{99
zyFSmICnMuUMS5nkc`_+#>~v(@*w`TISU6>eJFj*TS3GwQHtb7ahS{jD2dR0_qUYHI
zXGkvR`G{{)-}{iBNF;nH&|UZ=bdGRcwYOb}I~BY<xzYhu-hVB?Sh?~+%qaHNE4jw!
ztHV$?NGsUEtckSjelTtf0don|5a|Y+8_K@kUh)Hw8b`q3*F0}vSgM2MYr&Hl@q@5c
zjnAM80Yg4s6^{GA`qzi^(SWE30m)|hC1YR*jFCKqITO+fE(i>usCvPOhk^hfDEpbO
z;-R<J<a>-y*O`%7Loy1TzjvAM*e{M3rcH<>@#Ybe)+G$P8uN2H`sXq+qaXVCW|_9(
zB}RM77l|p9X?5s|cAr=1FU<+}ysJDeAGIZO@fs(e_ckI;$m{sH@T2q7PqPKNq#KX+
zd{=S=uy2qhK)qLqdJyt}JolUzTCl;+Z_i$FZayZwIgMW=kFtwPG)ZTc#D;&lcYiSV
z{I(@C6BoJO{WYwZ^_ToIrr60Fm6@m<#_ca~6y7Z!tmdYw2>Y3f$bzFMafACJm9rCH
zn2rrYWM}A$4I8pRL6g91REG`RkQad=4A`;%RQv-LV`P|vlZp`JQlL{)z4QE1rH*Ia
zya)unz=5?~R@v4CK1~F|x$X3KK$?wE#(y@`wKZ0Db_}r9!t!=YW#-PuzkG>hRA=it
z=?Ydx+l4kM7}0o{X6lOr{iGlm!9t;!$&5aeLJk|jph+Q22_nti(1r}xSoOPasw>7b
z8zNA{zwO5;Nw#S%eP6u__WO4pG#zN5+)9$$B1(vp=(?w4yO`iCYp5!}qSt9F{+@RB
z%~WOCNpY+i`=XI4mMW?34r5r=gc%vmP0%=uv$OJfhUE}iG%w7z=RZ?QuH_qT+v|<t
zGUp9^zc<^+$?M~W33|bcghYq$z2pHxe?;+ng6`qV8=rX8*-qlG-j&Mp_|Y}}Zyv>J
z>+P|i=jgdtD~5U+O7@7AtPSs*CbBY$%1~MJ+An6FjGb#we?Pg%B3-{f<$$Rn#Ny18
zfQ2F}OCfu};OKB|6c3&zUnWh>@NF9pi8+(rO?Qd1ch(;*L^B1g2dNxpPxNP&kZv!a
z>O_P1y3ov?-gaG9)G5(IL7IEP9ySakJ3x}kf++NT8EM->c8joKkgze08(cc=si7D7
z+`QWs4MSo;>LEt}pydFB5z*d$-Hy8r%?4P1_x}EsLpti9xdF#xjbb1+G$T?#tb<h$
zY5s(P#Y)91ipcf9Z*j}<!u1FqCJ<RoyjS{?sS$L$S6k99DkexL=)xCwH!L7n#DowU
z2ZlE_fK!#H!|Lymy4bCtXSPzHa8uqUqalAH(p4JcJ}E52kh5dZ@KpLsctF@zi~}b-
z4rVq==9**h<2L+&4}|phT3>Y6iZM&xLEBKZ6W~FF^D;4YqL9|%-mY`$XdI&)K99{0
zCN8-<a6|e}jf>tLSv+=by|Um3$3Z?rwFSqLJpbn3=C2Eu@XY&qP$fAR$ue_QJ-cL>
zjneE1aWwns5P%Z=t0ep&-A;wRnQZyDBO>^H3~Pu&=FdY927S=7Wg#bEsU{&0w)>y)
z(!P2DiR|z1$Iu_&abHi+?LZBNmL2KMKW-5ey7VI&bwFD1;hvLk-mj5x8>@e&zwR!v
zHQzcTG&nGTr=H3yr=vrL<Zs}8pTfdIzddD=whhKTkOKG3yq=7L#S{E+%iw$u@3-s%
zF6C-s|EeEGCjlSY&V1y9h6sJJGnAV|iheLLjalQSY--~1oHi$qZpeG1`vcy|#@-qU
zFRte}e~9W0KE}xIN5slwh`qedo=ZSL>R>()w|w>n5oF6_$AlZ>^2dGmW_taf@t%wA
z1A&X8>n>XXc`nrk{}Wzr6`FK><??k&0{e&F+o@q)!R)HI*yhS{0-a2Q31q<=K{iSR
zH8_~&A&4hJndT|Mj|D&NdH!?4rkQ16)bQEOUx@Zn;vY1F9#^6TgBd*~!DT~7e9&4r
zMQujJ=5{d~tKTSgheO2E*fT9fvZJQ5s-=aFl{k8{VYI&&K}X*u1)`pBb`K3Y^uJ4%
zm*3?JMn&b;-VO=DXkp4TRw90@(wI8*3}v7d+qgsdQ3vXjtehgc97YPueY^b01k{cG
zo6VBKyT`pXC~pDfho}2o*wmhX_xt{a_dMs$+J&YAFMrD(>O#OV{n)$gnZAs~#LlbB
z^Gl(NdT{=N!MA9A4W>E-Jhu!OyqR3~3vW^u!?@nXar^b^YML1gSP0%d{gn)M)fC8L
zi7;d>48~wlr`bN}*Vlj6^FV893G+DMrOkx8Tm_f_GU<1q)sS9!U;};uSZ9-z%mnf+
z_i%G^H|iqUDg$Z`h>^6S&hrj~nbOff$0omdqXw;Lw4&&s+~m!ht9Sz*-xA97uw>ad
z5)XETg(I&F8ZhJE<*T9IM^zQwnfmu+;X@X8AM-@U=eF_{ZxM3z^(I+D(cOI7*d`mv
zBo}6NnxBlDw;Dg%XTLwBAHVU^q4O43-RAS?<7PQ;qIr4L=+7a1<8t9YU)wJ1I$fRi
zA2bwfPjQDQF7<a8K3kVx)N<&#r}&7O8<qM61~m?La>Z8mKjAy2X~nnSsl%fM{^z*>
z?;e+#-)ErS6VL;RO~k}zsW<x8vr4=rqgqm_+tgBg<oAv;s&CdfI6PRck2&QM;SvOc
zJ&2=i=3vBMvt-Kg{lf(VYGCaySrR?i--r1+aRf|G0-0$lFdNaoy5b@Z6QH|yh5+nf
zrwu3pM(6j<W>f9?{5e~gNrIa@3MTxncCWvuG?W`uhrrmz7-TOz$44w%o|543w*-qU
zFqSkDDL*;L&A`SEE*(6VxSp0*ulg4l?}&%>^^f(hPmtqgJ^9N}ArxQq_g|!4Uofgp
zdnt2`g>k~a!FE&TDU0s>sXq5LHzu)g#<t8MzE5vf53f43o29iAFvxg*&_?ez2e`QN
zJNZ<x)}U9{b0=PF6SixZ3_1&^X<tU19p>CgN+@xgNNArrn`S?5sr(}q)<+@F?rclO
z*7>VBXXW$#N9uefEUKiTrVM?(V=@XPq!J-!dX<HK?O6DYh9%j9injBa@#pbY1uNe^
zDC_acuxbu!q^iD3WP1f;SIJ39+J{-hA^$cf$AH3bKU8&unZ|WjP|OGx7J9saGrUFp
zguHjU-7OgApBAIk_l!O{L2HhJbm^}?HyJUtd8DO<@D9D7TNe<DX<gX(O4Uap$%?oB
zStve>7@ibQa?A<`3APCOd!=_n{!jhhhGYrphCVH&Ayv;}O-k4!7Vg+*EnuQ#H&%Xy
z6Or|c)Z~VKGih~w%kx#{hfD5()kIH;0>1`hROV1Vd5vCX*-7vVhXC(MVjdGuyaKIf
znlC@&Mvk%2T^~#DeS;d?JC%3Oe_QQ(h}s)T*fW&)B;@LA9s190Zqn2Q8Ft{;Xs9Ux
zQ(anGswA_vwpM6I1C&C<j08NG0W9mC*0snXV^|lP27+X_YNLvd&ls8)v6XLkt9u#H
z@&F{l0ErMx2>{g{G<|RvymnCxyKu_B(o4k-<sKd$7CfJxoi*-05)%~UlEk)Y)4%_1
zMb7JYqMHz&7%I7(_ncZpD(=;Sj2VlzYo5yde2zd_D9j4der)eQjcjpv%fzEyMb7tc
z-D>w>GUq61zc=cB&pl-(Vz$yoY!exibu2gE_I}z5tm>UH@}l`qT5gX@nz_bj*VlPZ
z-yQlHW8Y>+GkHBMMJh<~0s{w+w2319^z|gFb%K>PH_Y9D?XPRw6D=V-n;^?C@^;AO
z`7yHoz{wss+!*8Mq;<BqE1huw_fcjM^gZhi42O>vKVU>e5YT>s4$crf<S@Okm|>IC
zSRMdgogtV~K=Jm&j_oRLqNbxWCUZHJroD;GLH%Mk<sS~m(<>d=+LC&+X@xkXwKaq_
z%rPiS?bEg2GFzh;_dBCH?ws7!ZVz-+a7Z_%;*y}a!-hTWq%?{b)!y+$I>3pBW%B-8
zi(=wfDI-xKRGeM(7cF(XZ9aCMr93hbcgGmxwMN1sh(jWYpPp?hFziGOTcHd2#6^yw
z<oA@W*hOraQ3!<cCTk}qCx><~aenoNr(Y249gR_ah6g<$hgvj?lR;kk6Gi@8LoII3
ziG!K>1*BIq@^fh8MI5H8R1{gn#nVYxmG?jmaqFe`s)BL50bB&Mx8EK;kivMfzGt7y
z8vs)Q4_#I(eT?DGCq-F=2hL@6Nk*#6ul8cP-t!ElE1WhwTrq6@BZe!o+b$;fSMcIZ
zT~w;05mill|L4Te?oMm0GJ5u_ef4GjH12<8=C3V76zJKe3QDb`nab^Xb>B>&^p}}(
z1)IrS<?Ea~T{)q>-WpN-DgUZ2e6+wPndeBigyAj^B?(y}>t~}XY-h8A7{yj=f=ch}
znh<Ksj1=6of$%w}dr54BbmAQ_0}K!C$O||>ikP0NvBEPnG;BGZ_XGdJU{%0<HDzc%
zPoRZY)x2f+tuawNTNCDDPES_itEV06Y?q|O$15JEC9>Wtc+U-O5Vwcf1kfY>eSJvz
z{jI0t)14z2@^}N1a4vMAwL#Z`ra_k;9huv=As&JvyYB9)j-o~{T%@ct6S8WGoliRY
ztzzj1zn1yHKKEK(%^Z_jociC9N|t6ZvtOn$4VO4s@9?WBKYS&-X`FmtRQv0`h`B!b
zytT#>J6El_KY0q3Gt`IQp3RQ83wN;;;&m>INn4#0wWoP6Yz@+t*JxBZa12rm<^}z0
zw=lz#ri%}>rDRGEnv(l7f?u8#qnlK^P;f=QYo_<#IU%8-PzJ37VHR|G8^#Th1C%8W
zU$OnW9ok{+1PJyXP}J~eX5M=5v>uxxpo&vyHOpgbW{JyLTi-;Er@*R8act_qGz!ch
zbwMRpn>+r+v+4)FNrGepGYB7OECD%$!%!0WYi62LLjh!-=zA!rLaL(2Q%p+ay+(uL
zA!#~<_VmgwGWPJo0XHmD!%JFqy4XAK;OY1Wf3$mXY}9aYiW22n_Z%)TkBJx;jU*~g
z@bFozf=C&dwNP}|71e}}Th2Je2=)Wb?oCGB)4eT%-gK)^L&9*mwSthgCZX!9&`><B
z>Rgz_XEPoAvr<dT%sA2;J*w;xOl%=qRm?X`8ZTxeMUcVcdW3SDgX1L-T?zx0!7MML
zf%#2=u`gjTDJdL_HHtvvP^4X?1`a3oJkPPQu{)4U;n@~2{Rvsa^ZO<g)!*Q<LE@$H
zX_n%#SKq9K-6M3tVgp(dJH*#(9(~l|_qsg)lQbWr&y5jq^+gz0v-BFvtrs5{c*Xlp
z3D63@E<}l}!ni?CrWwhOjq{4gCeS{eAXz=H6x-%jiPPa=%JL*_&k=d13og-gSM0rY
zPN&!sH9;i{Zd%HV=gKUt&qc(fp8LHz)>K!&myG^~*#G?Vam2~9*3ui#EioKEd3I(b
zzKjWr`D8rIU@6w4{6EDawz5=+pTClG(~=D`!x}u%qbM-IS#4U3V}Q*CBvi^*Z_uMa
z$nGd}KuT@7lq#oc==}GrgzntQ014crG1&%u5t{CG#~sOuiE=<SX54XmmoKWJq4AY)
zC$|l2BDC)-lhOBkG8Md3q*gHrh>r2;YqSoh>im4amp|2T3zcQ8;5MUK4Yv)Pb|Q4D
zxqtk~Mza>H@#Qt_y8mG>foIm>E{!YSkU!;5`IZ;aae33Nj#aPaZI#I9^Lr&iZq4ns
zS(t?rd$(Pq<51=}mPKO<sH^Ba(~egA_ccBt1DPM9YK2kYK@Lha6UvWjt?xroE72~B
zy2H514aRViJ*H=XE|uaJU=Z=ZbfLkO6$quYA`yVY<|X3M#32A<$p{!eXyv0a#P{+S
z;v?@3Q?2{mlUk;Jp?*Vd26S!-x%@A$iNzdr(teejV9R!%2i*1efzHAbt9x?0QT(i$
zK98bc=j0AkfZ>~We^<QPUDQ0DYk%W-;4!w>Mq3@8d)oZ-eDgcYQnVUD*=zftX3IPD
zwi|Uqva(7^Y;V`awRjEQ>+~*F6*8G8TPg=ge;JNSdPXnYC)DV0@w$H>Ir!?-@>jWi
z+xfX`x!$JL(Ay?D7@1!mEy7#eK7Tg)>*D5L0McXds24iId!@(vWJM~~!`J{Xp<r^8
zjv#v~kW8(>9}zK)jM&GKS&vlo@cJYul<xIr(kJkJVW>~_|0^}I^-$-deqIQB*ip8V
zsC!VdI%~OJm4Tcd69T2Xi>UqgOl44>Y)M1-7o5bQ*s-CMvCXEqVf4|BYf=eO{)sC=
zH37lr*siGpba8n+NBWX&j7#4!Uec>4|4`H~i;M|lSuEt>NHMZ#<9&*0c(X_uyAvi8
z4L&*1!`$$GbnK^vz)YBUeIxfnw`=7uS^psxD7nJ>UnoiIkrSV<UcDNx@bgJtwlWkI
zu{Ca)k~|zM4upZ-Cx0d@hrne=P-|{sq5YBg&U9XBgNYMcKacz($`;{@ua=Ou@X4Xs
zSn2z)8#kw2`t(@1(+$gIS$cKC_i$Cn%Y4!JWZW+La74ny6xdYP&^_g`P3;ccM}{b(
zLs(nfGsYSg1z5->A12Y4U4CspR+3{#V)L8h9W1^sU#l|WLu_3;a$roK>N^ycKSAD@
zd=i_xi48oV1|GGs2ZQd^PK9}S!JxUh4M`mjm+A9&AB%CpX$$>qfs@_}PXlp*K|R!Y
z?UCAty02|=d?J7*G6i45z)u+-o`I@U-S_W>V15%k-~9w)vxwkt)`cLhz~erKh>>u`
zis*4~bB8(WqfZIioGU12Emv$c4zmw8&kYxrWxw3&SM2kYkRI-@YQ3)(k&}%n&xHG2
zPXE%Wa+39!BuYMl=cT(BAqBJG{!&n4Zst&R_Dk@tqRXQC#_|XJADX^0tje@&8xTc6
zx(1Y#?vf7KbcjeJDIg`%DP1BVDP1B6NQX*yi-3R#NJ@8ierrF+_s*Z0ImSWvz3*#X
zYn|&1`^-;R2y-?8JzD16!vs3Z9Lv0a=e`2tTg?5UCiOXF&$P@ii#*ORbOJw!cbE3a
z>9rwy)R3}70ZB>Z6-opw@S68PWOOmi2R8xc@S7dNl=SHCg;onT&xOr|eGt=D!sOVZ
z&$u3MgQ1|PNXy0+0oyktZBHAhv0xMS5!BmnPtO4@KyEzl;6CseIH0AaB_<`MFm3PX
z$lZ*sI#f3Am;NBF#)gnasXnUhQ@cO9p(!hch^1%Myt{IAVuXB`=PQ1{eoy=y^>df2
zLii1vL}ilQ@|)<}ig96;!Gw9}6|bh<7@KzmCbeH5V|}E#8;+T?bZblL+evGQm()lF
z&5LT?Qla?I+t2Z8Q$E(knz22+?OA10DJz9i=ifH9&o|bO{N$a$;?p=B&o8#%Z=$V{
zD{4*okp9@R5)76uTf;zvkPGhtu=G(4+-_#UWk>-$U|!^)fO_$^TCVTfLm{F6495mp
zeE8vzO0u6?^XZyzx{+)77yma~wlKif!OFtiGxhs($-j#<;%K*Ze1(seRPt8WWKBYQ
z8=L|)jm~A8akuMpO{D+LB{s5qTwoil3JrzX>DD=l6i&8x$akR-Vi9h=NiaCRGSyvw
zbaGszS3J|5&+@ZZzDu;|Vt2Vw$2R|e?=S-+3riS~mB_9(xfs5I#mBAxB>ehv7tuxH
zmo74Hxxf^=>IicbvTZ<t&2@coBe)<~f-Y1QfoMw*_M|W#95=42At0D(diS`9H#Irg
zz-Re-_MVF#f^6;&JzWsGQ_P-Tl$mG)xi98BR(WZ|G6wnxO8TX*yCFBxLmPNImeq;M
zncZQ^?pbbV2<f0b53!n%C=48V^R?fZ)%N3IeQO935xO44;MlowMcMQ9XS*IK<bjCr
zOgty_?7M2{H3xl9JLTu>w9-C!UZ-N7zS@a=cFKEOF2OzJ16B1TWx0yBo15DZ_biZi
zc$^j$;NnOYb?1a`(tfCJI0}2$L(CgGA21bA&mTy5T7L4+<&83>w{xTS>6@m?%1ZE&
z68xD7<Mg-p_<@hIu6pb}Q>L^T?iniBN-|iTjK6+6`9s)snNd}yn8zoN@w0SPNl0QP
zp7JO<y(G5r1EZ>puNr#eB-m$EWbz6tbzSd&RT@5+ObU*Sa(iCFRr$Qv-kV9YIZVlh
zc`s3oSTXO{f<T8cha=G{#VW;{ecDBboz)L_UZ?G@tQ4WXd&eT_SB;#LOt1g`_N!I0
z@BA14SD;dFJlCb5kfNMs1ANbD`1s<%1|=0*-do)8fl1kQ9R_IDhkKgxK7Sr+_VWW&
z=3a}>$+Tthh{H^AbfZ@z1AbF>FLELZ`nkYVu|{~J8BNQoi%6vZJNkQ7V9vMw^^)a^
zIy1jHy=`mP7t+)Z_p-lO$$v*hJL;j<HNDCy-!vN<*-=m6-gf=nGW3yfA@Rv)?y{H}
z4zKvKI!vuiU3#!ToH+CB+)heJc!Jmx2=0+u)SD*$WJVB(<vM?FZ*2IG^lPJj>+<8$
zp`KR<<BGUtGXXpI`Cd+<O4}yf{;5%{7>18Xe^|0Lm|fJ=?&3+I>s~^DbQ+jJx;G6$
zDfxtg-JqwEKhg^jb9g43HTev94#>h1a02`Zz%IuOB)-Ts9=9JC@h04((QrT=Ctp_^
zI&>u=!IJ-;Frla9dfN*;CJGGl^pxqoZ#>{Wuggm1Vqle<pL@TlS&^rQcl-ejn?mb#
ze~%Sa!ezR*=IwP6R$kB1P@zD|)z#cLPCfgJ|9(#sud;^Jbq$BGsf|VyQEpFq#RR7c
zW^afTs`YAM3zu3rr;A-54OG27Ew<CYHQ?RTkC1G*?|)l*hE`BG$LdME-I8T#8nWrR
z8{O?Z@DDaT@xJ$&SJDO--R{RX39yJmgKqxv<=J!;ejxY{CNkfp0J;;>>TNL<{2Yj1
zHUk)U^9zRJLY*?ZcV!aZt@+cw%bb-@(HSKqW;8qac~~s(YB9+u*gj1<b;u7ljV1p}
zcaNjf<BN^*J4!qXD*nj(Wp?W7^om`77IE3ND9lQxhs(Z4?Fa{ItO^kbZShvh3U37k
zqr@4I=hjv_tlOp6rN!qDdmc6gkWH7lMu_fZ{7o;6fGfw@FSD~uUPnDYfQMXId^QXp
z?Vp?UyK4K1XSBqM`^N;&i?NREU%hx&QtAJ{Pn^YnG{InQ8A0-2SXng~Ms{;S6?`fq
z6FeKlZaB6E12GB)Zo07qx8r<=Gx6njyXhZnWerZf;+8*?pB<he!b9Yc`ZF+k7{tum
zuG`>*MXvU5eGBLrd>vLq9Pfr9rk~hyGZXd@T4A_T*eQ<_yU?Q(vNh?xIK5`<%%K|h
ze{+a8HQ98MRyP-R)lEKjHM^)pQ5&ZkB@~|>9uB@{#b_w<2dd8syd2i3#H=nWJTWj%
zs2k@p<LLkW(WHNzNWU-6x&BA6)R^2q{eesr(Z53)&jty}!)>)d=?6l=!J{SdK9{)_
z0%b3L()8uZoSv(ITTdTceK@CFBg8}t92H0dJL6+PmQ3Jy#3m>Az-PF$SODV?&h?Zx
z^I=Q_UX4f)7knxjj{fM&4qgEbDCy@%H_&FR?&U;F*nF%uGT<ZwTbPILQ;yx9Wusp-
zLO*=?pq%(BIXU_AYS&z}g7z6LQ&Jh8dAv@+3mnY$;TqRHvZT1;Lv~_%bNaaB#H{Sn
zF+#w+vOVz!Y=5`=oqjH~ocP)4fi<14EbWSzXH7piFDBuu`MiWfylTfQe&!MUR$pf*
zWX`Xw*=NT0>B;@~g>Ts6(y0V%OVO~j76u)TLo;Pc)iGXeOvbzt*p(n#Z{h46N7uc*
zslT3e|Kk`Pvar#8x1YuJup~LF)RPFn%Gpc67I~l};5oIpPtI$87wn}>fbi#gA%4M|
z4MMa<@K&f@tpi5k0l?72SDPAHV)tp$LD|ppTo7o4NCTzv?sD+v1FqECRPBLh*%>i;
zzmpVE{C*IAptQ1}DZ#E)=br{gWxlo`cIIuwSHrY>)nk=r%Edjpv0+6Pth+`y#YdC=
zF^`WvQvD6gOt6JUCqBg^`0*|C`UKRSXrYxOrAMq4R-@hmrN+=^VdO6(h-R`olu66x
zU~l+9)FZu@tFf=-N!{*rHCwrEv5+_?I+|YM7wyp+qn9X0J*ZgCS6GvVjhJuDnTPe}
zU%nv7R`6Ic*Q%RoHl-aSL_QRDSY_U$6!`#_LylvH?j&EXty7oDyqb7hn+IrlF&_Yn
zUF>+=?18XThyBYSFphlO`}X;)Qy+({^9U?9yL#hioPuD%;|OM1BValMYiBvwkio(o
z74VV$h(RDjJ!EH({Qi278g1P{NKI>WlL2ak`~$;p@5Qml=&;Ls2xB?YXf*>Iitt}Y
znodpxhXk9%o7={zR0)hBYvW!|_A_S3Z+oire`1TJ-fnXRo6IhYTls5Px%h24778{x
z=ov?l_A3IHwYu~ZCgqb=5B>zw{%w;M!=O`-Wp8!$;TToC2e{uoeKJIMd~m7w1#&(}
zZp%(day_)aRDOrHB3w3FyK<*EJaW{6wloZCMKZDD?wk}{(H>yphlzv3QKzv}M$_l6
zaP;!4N-1trWq)L3q^WjdCWunwVKqJs>&3u8rQ_R2;7Lj<u1y@$)`pnz+7SoGi7)!q
zoY9ikLLewu%1g83oie6|LBzm)+I8if(dq8h4fxf@%x<PZf1__sEgrV4J`_Nap-(VY
z!&;{=eX{!Qt7iLFt@C~b?F(r{KTafCaqk)P&?qevrPDCM?b=RbHLT^POH@xi%f<1c
z-z9H)k6=DEuis};GhFn-V*Wk-5gQA;q46-h*d(4kSU+gZ+lxu{C8LvZbv@1z`Qz|N
zw0g>3Ds(?`RH6suZ`u2pD?7rD!ro3cY4fa@rw}U}WuCNe6XovS*1KkXc<@pHm_tv5
zH;dlwoXkbjyqfO_gFQFz>%U(KjYqy(Yrt5EVo!73LG&~R)<v*+tpUv^&dDAEm_~uP
zlp+SZ-(a-|V>QxU<qT@emB?~SN5|sdk2HK^Mf*$v9IW@mAKi{!5}&eR8o5`U&x|#^
z7A6hD(c2w+b>5CIBW=bA`KO|hg!=6<KCXr<E1qm>D3q(btG?^st<*MMQ_e0h@k&D?
z;k*27rbfBXkl`>P6|k7+S+rs0wZd^o{dbIuPa&jGJ=t1bl!f3}(~rJrD}AhlR;njF
zcu`&N;_!m_fq1X0bUnp(_qVqxy}T@>#xW!m6j*a7POSiSG;hP7!&^%GP2FCVE~uoW
zz}6@%qQ1c?74*^-Fp2=hlB^tsje!AN-ZX*P@76l@>UTk_Jhczf*%6#woF04g^6=Ew
z-IbNQ#orKWN{L-a^cxNTzuBy_mII%=xVUY0-<$?xFkl0bYu+=OM&m;MY!2~7ce30k
z+E(&ip0z?;ux#8}_W=W$!w*L)x;{^TdDd@q!&EW_=CmTr!FWw?lA@c_TK)e;pE0zc
z?LH8SeC0)F^ER}Bt66mCh%tJ4pr7kV?$2#fL(8$R2V<9%Jb-kWp1q1$B5K-A3Qx7a
zyEl|WmPq>u^B&OZ{P(2(0F(!#Rkn6XUs0cg>pC%PY5&183fy8O-k!U$z^*TVy-)1$
zmo%B-2tZlL%7C;_>!L#52|h#oI9h00ZK8J-aOIqptj-AEH0=pu;k{OoAtel#x7<^9
z?osBNh;bt_Bbdmi5hrotb+qSZG{HxdpKoL}Mx%P^u5~saU)J1)$B=0LX^|%Hnm;J7
z+Q9OiUs9VVHx7#!nt9l-?URg?I{H5P!^ZdJ_!!!#PT%7-mOk44g0U)msZ1)dOy`HO
zls2r0tw#$<;4}h<g2T7sg}NBjIFz?{u!V(%Wj_&s>W=;B2o@w_K)D3Z&2RvFDVB8{
zp(#Q-9R32zjA`h7$;;h#ScMEXdbvCQNec_3?)lkcI-NSJIQ1>Mp+T%!!7Ay)A_s=5
zWjT$cIL2pFu~c*E(bo$3_L}61D*aFeH_<AOnfQuM(%w#TvrzRCVBJ+K?)lesnW-~i
zJWx5I{44meJibnu(wEYw1{En>7|rh~LlipY7B=pyRlMDsXY*sQJQBdgJ)^ev64N~w
zQ(O&N!O)v)44^XN!_`Z?yYHcU^Xt`S)?VZ6Refr*)SpK`K%(RcMOqI`en19MQ`iJd
zL}T)-vb9-b_jL|ZY_M;W8sP`4=`FJRjGYV7kxN;A*LsIFU)N)S7JRrq8nL)&4hF+d
zfroYPpa`7$#sfyyazQBHS-Le*#RLo&MgWN(-S|FHxyL9dnv_Pk{c+{$se^qz)z=a)
zGLE9}D}()Ac-I`Ul8TmcO1hicGBWHTDWnFq+F~hPc!edthLm-Y+hI7EqA^|_r%hb(
z^RW+2@1A!bjBmer@M?1V?dDnE@FSX*hb~iye5V}c*<Ye*Sw>lXI#)NL>-+e5y4J47
z_wV>!SJL{Dbk_+kf2tQ-<xFVr6EJpqT^}t+Fk8;rdN)ur!HoL4F^FB`be`fRP(K$}
zvaD^ONZDHted^TOW8mV#i|AC(MO*2!T3V(kDsUEnmKZF@oPpobJi%SdFA+fWFsK+<
z=_oT3x)O@8qyl#-Cuj^M&K8NDR}1*|tbE?>9lsz(k7#EiB%-F?5zGQe@S(onY?QoK
zEK0p;J0pv3NZVIO>fLLyc(fh0fcttVg!~OBa$+n@EX<l}(`entw`9&+%{MZu_b2aX
zy+A!gb3Q#AeOy3=wW2EbRODS_w@I;}OWnM<VDW6A7TGvM$h%YKH}5$;g)jejO42T5
zNNZL5Xi%N*#Onsxb{bmT?LNs1w2)3jXX*8v@(%kJb&3_N^PbVN+t*&v{BYJwGdiWW
z?hZ`IAaXx|-tPd(Vr0guvK(x(VITnGJ>b%07wG(O|8*Ej(>C$DP%8pfrWeyuy&<p|
z(JlG&DZ5oCJkQ!2-Y{62sn<HKy`?GDt9lC-9X!Bw#2zTv7i`*mt!Qa-_5;00yYcH`
zX;_%z8URHtoO&@ayH>1jy@7PS4`w}+e=l??=aI?3aJr-2!rDSL-=dx)!kKh?G$*E!
zt&>Y&D;u|1fw@j5nN{~<J>R#&%*pZax7jKaZcyI**~>kWos<3*<(}R_rl%eXW7t-+
z7r1S#AJ|d!u7>4x4XEhp1A{hQ(oQv7(AfPyoT?3%$FbWkGH9~8`8ed>lHDIjDB4qy
zzY+By;Rz~!g<mwP1Y{Jruff6o1FiJia<J~T7)U2~Fb>9lAmreG2WLRx9awbIDiVSP
zW?e^u&Hzw~KpbWUp1Z!^%$!~j%l^B**jWK07_Us1Pg^i%F+3IELe#UzYyBS=fb8}r
z!%NVmAO$CdR~)D3BT9eGmg;HexfqzBom3i0d>{5LM5B}H$yJQUC+W7Z50nL-c5mIN
zDAcr`S6AUH`qMR#2Z#4>n7)rdtH<fLVJgVxshshX)i!Tql_y}OQD~V->FXI!TO`wG
zB$JCOdTUV2h)}>c#1`N9(e%*x!7&AEAYxiNTdlU*a)r<^DJ{J<%UPSI{t@DK<kS^v
z>$4X+i^|qtoRrP9C&h~^yvpTCsQ;7uY8`1B|I8MPpd>sgpQ|sK$^X>{5P(#IICZ<>
zKoz*`w;sE>&1nTAxyRR_qjSE~z#5JKbL-ilVsNw_f!E9V+(QMe56C(!%*-ahg4Jml
z13nKd>F(URbB}q{6-v+M-)|9A;%`29I8+QFw{O72LoU_QRV^zvu1Z?0P1TLQ7&YG!
z{9UO0NLlDb+<Tg88U^igZ1SX$kYBen_Wn7N?(1C)Db(o{OcRhsj=pH8APUp2=JZTW
z$Q|{%(hk^P{Hur_GX5<1F(Fz+|K{~w`nXP|-CvGYwQg9=`iOzmfTYH;)*yKTgxbei
z2OCZaBZ>I&npHx>dcPr&GUDvrncwlHZSNcNwo(afznR=}<Ctinp1pQk!*)6-f9_gL
zigTo?{n)$VIkFCi??hdJAv(C;%ul<|A`1}eQR|uuFYwjCG^`DH#)F<5jEr47Ql!Wv
z2%u{V=S%j}JNzy)&Vv1W1vD9KGeza)<-bkIz{^f-zvm%yIHKNdt8_fHkAagj%6--+
z2`2YSaI`layEvctsb>#Aj;@c_teGm5i+r2%CmR7dQSW`{*%nhAm2yX$6@v6HdUsn0
zDP)p72T2;;oK~I<Pm8R{598b2Oj*TpVBv@uK5fWT*BiqZ?wq~k;h_>2#WvtPOyBIo
zEzse-3Pbb<$?G`Ua8mGpG5)kcD0va`-u@?-NRLi|(O%`IF-!kWV@hJiBz|z9EZOC8
z>*r4PbbX?V^$^`7v~>NO8QyJ+7}xhsZYe{6O}9YdF<vJuVw1Yt>b27tnrAHrSM5rI
zcxPuP2}|i2QI=2}sE00Xx)<kxL2Td{Q*SjyN3Y0wG&cw72Qe_w!T|#_=p~XfhtE@P
zldpqALK2v(Y%D5OR4zl!wU|W)u*wyApT4J5h`*nP-r;*(5r27jF6M$6jK24p;#q}W
z#-}lC5^N@t@z)8b2Ytqu%b^F+LCpSi=)Kru=lxN=3|2SX<lSwKT6WHQmS2gO;>a)d
z(a}Yyva@rNPX9h3BJ<E%3@BKym=ZmyH7(o2sHAgV(U^9id59g>aH}`Tz#ivW`J~pS
z8;PRp?Vi4(AaZlVcBEsWKbRu#L*x8$gW+b^kOi2FrU1`eE>Pdh{n57<-#G%~O2^An
zM`*gXbtD1wmr+o-3%_c=h%!C`+=iTFLCOA{0x;Kn*hz|A9gZ+G*}>|HUANpHPFs?Q
z&;dpzdwfhx%l34AgyF`0caslR;051Hj(FJqu80`hgHNvxUxo5aAk>`->)qAMLBGet
z)EH{F@75=f1`TFlYi9eHMLkEij_YdESw8KLzB4EpNv}L!j88=tes_+%wWF{1sV~h#
zL&n!@7QZ8hlbWW7Kk*TM%o+%M)FhET{wf%se&7bvva%XaC63{G@4%z2s59|z)3c60
zMTi<(S{VghXKV*Ux+)u-s)ve3%_T36QQnYy{G{2S$b*lL)XMr(`<?~}&ffd^s}*`{
zXtcXas%L(H=OE{J!z}cj2x)1>Cr4$e8s_HoAn|wsO6%o-BUS<m3MC7R<(ElK9PmS;
zyW6q38vNUzRbn1{9dBks2Q$9qX5=f(T_vG}05(=w2N%p_kKtzkZ_N><ffrWyDn^p=
zvC>{<@`$?ueGrIok;TBrF;eC>f|swByo0Ei{DrVR)DzEC;uz%EsdRs(7{^R+Y-Nsl
z^RkL<cJ}Rk33AzFX(rcU@8LYZ1wp97jO~<C7TFnFp17))dUdgUXP&C*w{N4|MYo%7
zkJCEf@&ruy^^R|!{mu1%iu+Xa_pftq{m|O(jpx8hXm5W<h|g|5MNUS&M>Mfv^kJNt
zeTbk(O1opWrg-LFWvL@H;Buk3lqg6kET6^CIly<imiC{9YmK=g9f8HD@Sp*1ED-q%
z^`^JKf9d<A-9ZGk<-u+{9&m!d_h`4CBmQ7*yy78H>lfi6vmSkep7&N+!#)<EqMyN9
zl%cMR%tVi!wCQqR8EBsxm$ImIKp+wKI^G2vegmhB_J7F85pctd%%#^j>g(vB=HGI@
zZ}=IX-5-C>$;+c~RC3L*OjOe6I_KZ9!2rTy>V`Me{ecWl<^K4v;N3Fs?ahX%$OqSa
zmmI^gwA~@`;YUh>c;&wfC=_VEP>JbXrx{v1z1Tc@RXJ74qtm;z*2LvtrP2HRelWg^
z{YaV`!#hO(lkDZ?<H<=fujxdV*>cZe1b-67giUtmI}9tyrkjK&9O=7`<CsJA4+>3|
zrn$073_KRY8trgc8->xnaA*DOrH;E=8$&(jcR-fzC|K=1SOsmJG^mO_K_9N%n1qi8
zTR3Qg=MJZdcJF8dngaG1p(~kQQQ&EI@RM&)4P-bG+e@NC+J%~zD#2`@<#5XGZ36-4
z(ynjO8_>5dX*}@`cz%Ot^(>e`ZOj4Bl>bpLp024I^r4cryd!uB^^76Q_xYtkD%3<W
z!?6hvL7|B8!F!3`DL>@+uTOLM=}FsN2h0Qm9oi8i@8*t+Maz}F0gVOE{k-nl`g8HG
zMQ);7uclgm$QJ!~ep>bWU2Uf7UdDgZt32r`o>ql#OBG~BC2lUt%3DJ6Hp~lD)qg*5
z^SVsa!)`8N$ga?SAZ52`dlT_2C0%*Mx_x}Hmqv$;UW08hNRdHfUfh*82*ABh#V=$g
z+iJhP)|WVUT9+Mr#ff)dPt}@-iT!JeHLf3lXco1Sxb)4L`9U)BqKn(s4V#p9cTr?-
z(fG3A<C17B0~wa;*IXjhL*DXaDQ$8Zphf`e9YBSVPb*6LA}2%g)W{eNn?J>h04;$g
zpA)=*7JzD#57hw{rN|v9FEC6Y^PmlyFwWXp5R0Oof^JFa=B-i~ypFI4gsX-`Q9<GH
z2tO{hgu(nP?NWnYaG(Y{>Z`CH=tyD{Xe2!P@955H;ed}cR!B%lXXz2nmVjwjGL*`Y
zFAmb;cN(sVaTu_*v!8t|B3`B1SK1}N-J{fN^7ogbtx;2enPRU|g5Z;-T)s5ar;*DS
zos#OSj?el@!_XH$(b5HZTvqvIRgV>?IIsOGWfR?xivO}4TUMu_GG)0dDnTYLh{qy$
z%PR(Zj4SU{_G5)y@LVkVH}=@y-(>)7TsAE`^I6aq;cvpAMGyakZ;icFwIk86dj9hk
zQ}UP_y=S-TeEgr9?Jt~DhdV|+q@5q#ot9)Cm4tDy|A(pO8X$TzpC$FG6N_0p=4F8I
z{EpXmJ`A=G1mKapHd)OzC~?7#Lc*%ay!Q>J!GC~z=E9nI%KM;Sw?rGDtQ27Cf&UE3
zN08)!3tI?O8Q@}9MQlIv+&2fd0CLF=S|=o6sB_$<qySraq-8W1ATVri@bE+dAqFib
zE)D^{$l->5^BtHFmGBsWKvXza1mk-{*#I-<M8FOBKn3(^bLm^p(U|cd7yu*9(AAq8
z@ZIXGv&R<p(0@E4ig!|+_I2O;)j&O7Zbp4zZtoY0-P%fdQSsMSv&gv7s)Vh~zj)DQ
zm@vA+Z_Diu`&oimK9S*Ag(5F2Z{aToY?2)mI!y8~O^z)52omyDo+WG6KV@$+sc!_&
zU5hIN<;NWA#nIW^{#G}yHMq)|hJ|-fVR^XFSMJd>nu=fTd{Z<jAR$-wXbQgf+f|07
z{`$K9f{=Z3C?qOarhS==x+l$*|GhnxFOiN7DK8D`UdjDm=Mz+K&+O@sZFq#9E`hqJ
z(wpr`R>MW$M)7DN&BW`EJ>$;}$V|9yg_4^m&>~ww;QA_p#RLQ>JdT4sbU_QC@6Y-D
z6h`<V5Jv^1Pr%Xx%bz`{?ZcmBu_fTbcUJli6s<1kFIQ)qC4t@8Y;`j_0fu2stB!!T
zebFh)+_OJKgV>c|5Lc%oDcO)(<Uh_YEDbAIR3WEa^Y@Jai{zA-$J|`5NG7GfUvjEN
zf}m$G<JLe~1K%vASl-!hN{B$6t2N+6Cf%#_&dPsW7?_&;LC3Vto!+h67sWR2HXHAt
zD|v`kg~OoP$*--o?cA*=^9T8}T?rX@Xj6zBe4N{1CcSNJ5nb8JXe>Q#)738W-ChN*
zx|;?kE$=l;iJ#7*-q0`#EQ?(*X*?yg#+Ie|6{7m)1+FK;@?6LM;;L~{-yr&Yh0$F0
z;X=v_nZZkZ<!?5=)vxA1uf&U=dj3GPh|Vml17h*<3+wo%`I~*m4JrX|_ZJDv9a+^C
zi(lN!MGVJLDXlL45b=BF#<%XY2(d~!@O4*H%An|dStB#d?~ozlTWa=j+on4WunTbJ
zZb4QhaD0e+(f*bPNbj(8?}Xxna=IA*v}UL2!a?W(q-CUn4)8^VE~FRbK)C_O>~}B{
zWsVWS?mk_8;%4K#P;0O`euWD;ApsYm2Yv$BM>H?v&HD)gSSSbkL~==AG4KnoGtXO#
ze=6}0f@zRWmw*^|?x=mv%NvG8D@d~1pt_Osn978qT@^h+2CnI{gFjjTSc4N16T7s!
z0=jQXD8!1hsx+ze%$xi45kJ((&7N_7UC|4S@BMmVbS;6J9|%y&X|qvDUTJ?)t>P2J
zNh&=HylJ?UwE0~+5Oh-$z44sCso=HYSnbWSQSkG?U%^ot&a}$BREBsx-hJb}_AzHu
zRUY48ADzd?<G|zjSX1#z|4&o(fBYGj51GzhT&_$VmHDMtb1o_HMpyf!D@8HKrIY*5
zE#2JroDrc#H*OM1d*ZB(9~Oy$sc$PpX}l?$v%Q738^uYj?;>;?7rRjZ<tWB1(L|xd
zlN}q^8Idhbk(7t>s2{1upDsq+kXh;6YWwE*iAOjTY+JzP{VCXy3;SLQfsGXt7gscx
zpdU2fTtx#`!~OCHF~H9q&@<{)nq$JDngdh&2GdFS;y3{S1b@p|NK<?MP2?tM;i>ma
zo9o2)g2L3Y&yt0m8SSdmV5z?o2pueV%g4fj%Y<XLf3}ikthtF)XS-}qrNCVUhv^=y
z)wCGV*mrSzG9i>~W2|fiR7LfxB3Wn!$i)Ks+jpvh+z9L?ktd9DaiD_fBvZxw*Uw9P
zw}|*jj;Q|dTw)%-&2(~oR#O)3!N68U;cNkCRkL@NPPe#UFA6pdn|MULa7u2y^E1~3
z?{D}4H2nPI#`Z@K=p#aYF<s=K>pv@|F<k754)ZI}&F8(j+%@iLyWeng-a35nhu&_g
zcNaGZ|8XRtQvWZ24G&SIjsDa9+O@F9?bOjHrdGrz*?&ewi@8q_{HL&|-T3+<6|Wp+
zeRd>#b=j5q1Ns@lZ0)8=$o_p|#*<~guAHTMf)GwIQFo);Pj3aK^s<^kBMm`u?MN5z
z!or_feinIwU}#a+ZsX*{`<~*FDKN+s@|#=G6pz+t>dqlABN1v4#_#W8I!nj^G$>~o
zC#l^F$XJX8;2f{OI%OHU&o-NHnbmd>!37rV1iKd9Pyh`yeRUFdapYB1n|94FVBi@-
z619PXfwr6k6#Op2b(#WGcEh-**-F}9{ah_K-ewfk-Ud>nhF5&S8z0I$K6_Z<0fDG|
zvuE#eFEyAe5pmtJwjQ^O!`}}!rCTSzpZ`VZ9#9&zoIRJb#yr><_<)dt`(i2MEoEu9
z>9b#2IQ=D`zC0hA9Y3i`+Fg59^?0m;_8}Xg&-EUDJENHM0M*INblPUd7tM;0c!t`F
z7fke}Xjh~S2wjzlQ>uzP69qF@76PxQ0_YU!w}pGojRhS?I5;ByEDoF#kDWiEY2e$=
zZFbCRgv&q*GJGT;C#VP9QrG@$sc;lY@}WvR`i6YrT-P--eb23%C+hx#+{I6Tze2Ik
z1h4;*CZ9b%29-z%hMI49(E#*D&}sVvhm#MCP2e4M9YUERGeJd)a|#OP(}K%nLGz9J
zzfi$W(|=#`GeuU8V8<H-d#vSw41NBUfsT%j8=2YJM6;@-Ao`)N*5Y>5=P|RdqcXTd
z3bQME;(oGNlE%jV+nl;do88KMzm=!&FmQv8Ri)-`$FNhPbC;Tr=cB|2cerMi)-%n5
z<w75_%^ZwRIZ{_F`s_Sl){DbTG?!EH%lzY>MIh7L*K4_4?>t`ZN*(*hM6AuLed_Ng
ziZWlrW8cuVZ86?G8tk6m*lUvuEZ?bhm1Ek|UL-8ct7&2Lvi>PhYrBKXZqZ$-JW((R
zl>wID8PKSZisFwd@bV4*Jhy@-yJa);J-s6GFu!|uH<4Y>Nz0pg2k?X#$ObAb^cRdj
zXt{;A3;!z*;4-k#^fB`PydeUBxbUsP^H3MKOJ54K1Uf6D_^5AL#)JQ6rzxyRa8U<O
z;Tp7DZyrvY`~IZoj_=h9{eDz>mZxCq=^V3d<?RT14W1gbM>cnwsFlLXWJ11F#S{yu
zX}MVXw)imbSlN*nj`R-xWVlkmvX_uZESoj<kJ??=R3h9dP4iUscRg(3*?zr`Ld2xP
zFt#8gyHX$;|N5*uMeWWlcK_cR`Qt0SbLC8-L^hmm^;_vb9k&PdpBa))*<n;<`35$+
zn&>}8gY9O5l)2tAG}Fla2UxO5e*E4YMX3yQmRmLlwhwSIAV~!o!>Q9Bx~~Fkc*t>N
zzPZ}*o2a&@hM9*N#^ojOip&F+09k_tn1_H8G7hNBJV1zs)DQLLm#0;gl_>u`vLQx_
zR(FoTCJnk5lRLVMjHq%Tc*k@}3V*)3{HyYbZ?&g#8sBHNO7^`a!KzEU14PfWAzl6I
zJ%Kw*i!?v?2v8Z=1<>?j_dGOb)`j9YB(HJN_|c<wP4bE}<Xo1<ol8Am{ql3HZM>=v
zr+4ILOVNz&pW(l`nLfYOztnnk;2X94*)eyHey#LDlm5|}@YXFlF6!y6#%F9(eml+2
z_<C4#kZ})C<e^j}?lw<GL`Ubt_@$C5mI7DE#q`W)$gAN2=ZNJ&$(x2k-w*Os-n&v%
zZ&qmm1nzxR)+h%IHV80juu=q3>4gWxB%q!`(|QIPIG<-ck(F&glDC4`Oou&B*Cra6
zf$4XDU#?IRKYzJzKE69+I*j-Ebadu)&7MY)ux_%m{m=Sk^*^)AI-4Nl*GBhV-xvjR
zszqnf+$`v!3R5S0tcjb^t83nUWyn(OmhpJh)dgKZ9)BTGmG1jGxu^S+k>ugC+v;xA
z47wErj2JJf|NbX7#}hZ=_KA6HAm-n&gH<Q#Z%QUB0;Bl4R!5h21@iBouUNPpQ5^O5
zf6Ezr9m%kLsT9*pWD%n!7Zx48GGUqZ5)c%2qlWRFV|vAT_}07rIsD3z#{)W|d-G(&
ze>w8ej3Oeb(BXWt{;PG6(=ZGa08Oozpy~sEwI~SWGH{rN!-rJ+fzSvboZxU0wmah*
z3%x&nn1T!xDCAY3U5B3Ez10;Ka7pQSK&P3PWUaQ2Eb%2VsFZnQ!HFH(w4vkWm|*I;
zpW^F+`$d*to5JMZNmE(Q63OZ4c8J5JGd|*b(4CM7(z!EdeK<*F=pnbO*hCWy8e_yt
zVjgCy7$9CoD15TG`onKMzx9yo=fFx)>yCr%*b7!ADH&EN1L|iP{gXQTe?-kn_v>rC
z)djczwKQY_+oy+6_39(@o13CR&ka@1)7$wLv<KBPE=jN#u9e$+c^9V02l!r~Bwhw0
zDx-ivBIKo2K<I@3_!S!%@ZO`~e+LgKE;iN?D$JnEx{C$A>6<$D@i|zHCkxse4IRM2
z%K@AO2xKW>+-NXLsFBk%Fkk`ZRghW94F%q==HJ_h#9MXKb{cil9gD0l9LKGW)wzbv
zGZ1<Xwu8_3qt|uPq_e0M!j0ZZul^?Cv}<dN^}|p#{1tk&d@QuR6Hu;8P#9%v@%xt2
zx8PLZ(uAvmd#7a=!LVxXzw&V*$3=AVP>s`u6F=b^Z@xHxlZm#i68pKTTi9V|$}BJR
zr0z-$56|jNTyUuSO-Lnk?p^MOHMr05*{%DCcH>c5j;!|=6w1Aj<79cmEu0#ng2p#L
zXQTdo!|jK#K-znN>jAM;g^y`HR+=&E^%yA)X$yhkUCjjpJTlTo?D}-U_nRjz(#rwK
z;1&38d)=-{s;H!7y8M$5BXrjH;yt`$q?oaqCC?Ke4{vu=$9M=b_h52IMud$zjd256
zBm8`@E1LQ#B)_28bAJO^c>mx4eotgH6f(XRvY++}nH_*g>alk>!<L_zK7AWuRPV^$
z-sG84uzuZF;OAt@l4I`?EN4xaIP(5_edwV=l*tDa-=})1{PI~Bw-tCZeK5*3%gDJW
zmtv>V#Sn;~wW%fy>x+R8d04bCRASmc!bg*6c!{n{LXXO$ExMKImz+k`e6cQUkJ2k?
z_ws6FC%ab5yeLpyBpvIBwd&#|qhab`wnoejD_HI*m|`maMgGTMLB+c7zOnjUIB7$L
zYFCJpf%=SWS^@Q#j}qNB0e=XH@fj~`Lqnb^gPIt)diIBvxsY~+;1wVZ_uDOPdGLcU
zvlX}x^W6)G3m&)q?mx~E4SqybpwkAkZB72)B6M_tkkE-#MS(jopdxdiQ-MV<c4#}2
zo#DCmIq8S@O+qxhmJUFNLG}Uk(phO~eE>prL-TygLskkl|E*xQgH#AWiM-T09Ph~W
z*0||HuY6WPqv0>-Acw2*H~ORLXRAVEozQwFoDK@x%U?5?X)CecZ1_Sh;_pTzDfm@N
zD%Bv=7Ffr(SEIHPz3x??>3sb_jQ(+{$h<)~Wi`ej(C++#wZ@Pm7@xnul_+2J^TM!J
z;oA;;8%O-8AohOD#|00AH7ywhIs^4JDwwel7_VXyE-a(-mN}BBTntIaZ~i$QBJkWk
z4pa79AKOsRY@gG0yHPQ!Bw@yF*0?f-x>3%*BaYBdEHHUibCRZ`7I-~lDOhDt9?1u^
zc|sy0Plz1`a{(M+dM?y$lp+&@eZSp*@ZcvPb>w0m+z`Bz4{1$Pz871z({-a|Za%Qf
z3j(@FrQNg$?1kg~$JxscHz(q|sl1ArThMmOL6>q71$841HWr{Ai|kuMMs&gwFo?&j
z>+&^B5aEXY?%n};){rQb0KM@Z%qtF%-R2Al#FgM)?zS`YL`ew?=Jtb)DG?%~cx5x^
zaqC}sEVr$KpWEuj#Dvf+$cTh$gKT4DCw75JDzZo4p2qjF&etoi)`eEmd>USLM;|_-
z6Z?oEK8-8WvCxG9(TeH~|6ea<j3o4SR1S(7_kYAk--9sI5w<~GD@6B~IyD03xJY_M
zU1ID}#HTMS?_QdOEHmTYKdS$lHx`(5LyAiJc$pd_VB`jmPTHe+QaRU7GO~Ch*lbTC
zx+SN`yM)nMg^fm%@xwa{*V4k-fEKaampG6nYzA*0=nVtFDlP^lG(M1Iz~5HJn(*}!
zJt5*0@Xq^S7eEI>EWkGQ*Ne*HGBUOV>m`uaG02qx3GuDgq8j3Vf3aO34Y-F!nUOeD
zodE!2C&0TE_(re>dIqs|5Hr>bqD?6n6v@Tk@JmT`%~Cd{LM|XOSPW^T_K6vZj)2e!
z8A4qNITjKZn|w&j7sLdhFA=4R4n+!^UROU+`5`_y;Fk~`&A!-pbx>O?&$QVph0cPN
zs=w3`UgACGdXH(0>Ac@pA*)F@O%THiUAMR=!R(nJXXiJE%2``91%dkngHdu=ki-1a
zwm<D`)X39(h-|iE{lXZ<-htNN`zdWmT^doy+NR2M3$^^%aG4&)c`*NC>)hXt32E}L
zuUSu2Ddh}Cp6iVI<uD5+Jn^U!p$Sl<zt*jezl;}V_74s=t)}Zn*Pziv|M>8<KmMnk
zQ|b2&)Yf~I*K*L!J3u#g05dGezPXSa+DwmmOh^mAv=KbMnW*2{WiH$fSeQ+jn6`aG
zn;k;<CR8m~=h<WU3qO`o>zBX3(z6&&N~6yT=;*%0pP88(m92);4pY?-ZzlS`r$7QV
zg|QLAZFT&W_IQlSltyQZEbe;C<r(L|K%}P(FpBfQMZg>U;{f{(1+aYK=ez_Zc<}?7
zfPK2V87z{~!FUO_>LQSz77u;YFkG1!h0Bz%c89Nq5Fa0@<#giyTTB~l0`7E>qZbOx
z(LLx|kRoPe`YC8z*#IsMYc=_Z2*HAj=HY_p(hqTJ9cI_Ml>mChuc?1SU#yS~y}GbO
zPoCyhAdEUCK6Xw0dF5kSW&ZkIoTwn`h~>QaOY&`L3!2&#sYEH&ENw1LvgcE;gHz9B
zg@fC}_*%o4(KoKmQY5aMwUpGc$etRR9~THiv;eyzb9i8yt;urI!kO=w(>8tj!+smH
zL*<}%jsJyE22e83p2<dT+X?&JP2nl=Y56}cfX%D!!FM9w0UlKW8@?8CcHN(P2s5uG
zzGe(3v8*JPulA_lv%B+k_BvWIaC&TT$#!mH@hoL9rw<%VyWmN(=znL=@;zsPY0wmk
z6ohb@ula&P7wLBoSnf-B2;P@_-QFP9i@Yxw1&x<wuY|amn1XWOHE-=8TT)*qoIGKl
zbAF(BwT_J?A@QXDPP>grC9e*N#NQl#gkU(>R)&jt?u(gs_;JGj01FC8#Sn)1{zr1=
zU&zx1tfWzrIK>)#Y+%9GYQn7s?2!lI&+|vb!~H5Qzg|1CW3(;DzBy(roNgVoSNUW$
znD>El@{#f}3k%lUIl3PEd(w@LxSwU-{CjW_CHwAlQlY7qM{sZvAsil=%J`UBXouc>
zd;CT{^gE@g!C=LBNCf|AK<E&qo<hxq*X}~57BRafibAu{*e6Z4Kn6S~Weawz#WI7)
z<pr<BT>|sWi~O<!pM1stG;BMt$P{^F<~#15lKp(rycN%LrGq#-TPJsUJ#!zY{1vBK
z<#&b2DnxVqQF<JG7*b#cu`NWDluw|`07tr>L?7^>%0vPg@CxS?vI1^rdb;~h&#9Y*
zn_Df5Gg%QV1dv*5C`y*_g=m%;Wx!}5$=|B3l>mE-ST19UKc$8j28_BucrXEc1nKGy
zPOvFJjsRLEKV+v~{$kAx|HKWe4GD<ATiHJPwU)``%Mj#gJ?fz3;)N;r*zL>Ir46>c
zF?kr9@P>t459%4WwkPI<C<+RJABHAbTioNnZ>@#OERCMBO_(EycsY}TN7fNFhnaCt
z5hMJG{kn21toFQ3V?s1OHXrqP=NQroo_*t$=h_@lDCd8YquBUe0qs-Bd#e`@3!d}b
zDLq>`Zaz0YQ8irZHsW)n*RSPzCBsIcs(1@CGSlL@@1KnTbnK6aQ77-K_9Ty`qt^bx
zjI2TX!R*zlq)c=174R46hL@~2LsSs?xS%!DlL)~Zj&Fc@q1>z+7kZd>(DgwkuLR&{
zgTJlGA$%1W$idDl*Ob+nRS3BD;O5K&d=*A^ZrDA5(Od@%7C@c6Za)V5UF1^++si`j
z5+$ftxH1{vkbMfUaC?A-iA<Y>Ed}x=hT`(lGYvgMHR(vjH#0bDN2$V&E8C+cyMR-D
zV0vz~kJ_oOU5M~Fqj&$m$VbMvn#D}`1J@CaF8Gg@8P^+&o?c0?kiTa>+ouZ-?xQ6V
z_hCWoOrPi$f6-p{5y>K{(vBgHZyGR&qgROw^yBn9zqi&}Nr}yq8c9A#P9c!$Y{?Zj
zr_%Z5i-A)YpY2w#pEqqmjhEwtzfSmRqg`<WM5hw_V#*oS<0jE`e_nP+(R@_1YdcoS
zx0>F^qs&o1!rGJsT2PW7oero^ywQcFpH^PqhJ6YM2XDca_9bL@BgwfSXRw{9p0Kq5
zvWYhoee^qbkQg$oyppVOKFk75`SS7-#GuM@W(L!<)$Th+;8ocJLKb8SG?EIBMBuFW
zuh*QBlx83gAe+B-P>xjb{4F-Cp2Wn%TUqQqWb1`j_Kpc~7LZ*G1P2jN6mn6G!Mhqq
z|IAJuukbBihZR4~1JzBzPl}go{FWk750Cy%vu`b!F{@)bN4Upc9^`iu>E7)Quy=T1
z6tU6uC7)`x*-KyZiJU89b4)U>#&<^}`pmv!Cp2r-=;n2ke!A}q-G|~*3Fcu5Oxl&(
zKW*ypwvJnzmj|y??RC6EALDPR{Co49Up32bTeg;8nL^&}3T(nKJa#(Wldrcm>$H6%
zI54Yg7>dUoV;R$@h*^4OEM{g!iOQzl9BkYi52}{9`lVG@3&ID>0|*{${=&et6U?7m
zl(`yZU^cGuzCn8Az(X<z?>Qu%1f3cWQfm3tYEN?k=#p8$>KJbDbcS>M6XJk*S>{*L
zrtM%EAdHM-MBZe)FWQ3Tk9CL;NY7zVP^myTBak=-zc*RI=+z30dI(ZSLvG(!2D7?H
zM)-1uxE}}I!6|@5R6tk;YahFc7$#zQf5AH9;+$8>52L^FrlLjBD$oBzgGQl~eH<QL
zX*(N7Z)W%FHZA{e{gq#z1_JO+l&>DO5-AMF8{(bpT(TA#F3kLXX@!V!+v4;s3_rw^
zu(y1qrIjLTC6G3@qhKtGUf|CtRWI>+_0i=PL?2_))*SY|?D7#IA;*-}22vD4y06M}
zaQ%T$xFl2X-0~4K?zXtfXPYG!qe3gWtfA9!boF@k8fqf>riNO~S5$8wb%vwYDBc0@
zEFtu5r*qkH`u;m~R~L2mk47y$#ey@<ulo3{htcir?ZM;>nKKPj4bmPJob-A7PD+@u
zfa8Rug@7?oG+5jHs<M0v8LbQXHjD<~qYg5{0pyGcxWiJP<fZmETyXzDrsTpHx&S>p
za*l*IbO^dhNuQ`un<a=3LxR&tt{I$*D)70mWNalw;FFUBv?rDL68Y9chY=PLVFrCj
z1H>ojvXk0P)kpzL1iH^f_@cox2mCRdXc~8(gMAD#XB(K|egAoNhLZ@p>WXA=-e<|Q
zVD-J=krbo|Kl*DpJaWILd4qUTKRWw*Oxoe#<VWR|a$KZ(`GdC9U<YBfKxMNFH#DKG
z<kUl%Udnx$h2`~Y*8kp5{de`qeYv!oJW1(am=CYu>5BizvdOmpTKc%_y`puGIm__g
z7Ow@2U-SLTejY~gzy5SgMaUAZxz}lC(_)4>E%hyWfIxe0J4By-i!bB1Irb1Kb{H;O
zZ)%XxXy6mqK)vN}=kJA%Mp)8QvD4HPwMNQqaM6glrbx`>2D$b&lAC$x4&<WUzS_)2
ze$rA}o(hm?KcwF<E&$*RS)@hXw~=o65YW>XLn{rjr)uB%)i~5XqJa3{4ET@8mhXY2
zZ#ozbX?%Pfz(nN?R7oTb4Fi!X6tE81f*~!o!JGt4!zw)X<}oPl-Ech_5dvccaD6ps
z_NA<v_z9#RxK{O}3@Qi_CRG!VangUP1aY_kJwe+0cJ?5MQgC1_`=5%Mk@5Q93i1Yq
zIm5%#GuxU1Wa7)fIz;xKfG3zA(XP4FS-!Isc;>o6KY*Jo{o}|BpCr0{yT$ihq}o?V
zCajd}<=QoaUYBkjCrT-se5*CaXxN*h4k9MKIQE5SpYP#@zFiMorN&)T{Nq>av`^^O
zJl&a)HAv*Mg`1l9LZ&M?xOGl$gtOmjwp!d1>(c1w>c*NfX4W_K2)@S3ir`@OUJnJk
z+<!F2O=6qxJWjpTwQH`XeeF1=6^Bc70hQ&<51%&eW>C(xVGXBuh<_~ADa||gfj=@X
zko8rr?ubzN2ENf9clcN7f@E_d-0T-*oKr7}JonL`Rk<<PUikg{*?2*JHvlt*^U=n^
zac;nbTl$U6o46ycL`gqiV14%kWg;13j3NEl1dbV)<(Z!~Jz*pQ`~aCpjvR^rZQ%u8
zfV}UjAF;8qF{@7LPr&w&dEzYEn`b-p*Y59Oe-FH1=bagGq@*5s*}|c^aiO(yap6Z&
z?qRzLmMZ(p{e5u^%E*8qFxi_RtsYrhgyeRx&<KIvh=_zFr0;0gmluWs_zkeZG6PER
zblsJNPLpx7FWhLu)%M1hS69G1E+|w2Ls7oGJI_nBp-vzJqwO1yNI`Dww^+xYp?c!Y
zd+wp`IT4C7HR+z%XqZpS!kZc5T(woga+-KIuWNNck1CYV04pgJuPyd1{KKT6M(k80
zcjLZ5+}_?B4$lv4fo}s@HRHDPosO^0KBNtu7>hS5-)%iJTS*z;BXY2YdH5c)0!=UR
zkBlU1f&=c_6ru7hlu4bP#|5ljV>ca;7+zi=KqKCI{$#}BeR6}zgG?j61M(kKp41y>
zX&v(E3`d(d#Y4S|enmVhU1f+GO>~o_m*{Iaak2FGc#~rs6xTL>*djW^-bxqbesMEM
zRf-fR=L#Yx3Ai(^DO#wu$L+N@TIztd-RXFCRhv7l`tiduD&s@0%UOAJfr1!@?lZBW
z_+md{@Zp6XNa~CF=Va~|XztrnNFTA9+Luq;8^GB`4QeKwSY?oId8}l(pRVRaXRhAb
zhCALGkR51m1-wqyvAuIXeL4jq0Fvhc2vRnD8_-fL;Pq-D5NL4LR-ohp``dz;0R(Y;
zz+U|XJBA5$DQ<8B+1+(O?q5T&$w2muF-)`%)Td!EoM>olgox!Irx(?Jp~1m(3k&VA
zd(Z<1(_0Hs|M4JQOE1@UenYWG<`2UK2wF@Y08$Q&ul`Brk9_XaFB5wm<riQZGjaM?
zh;TsO$?L>=>froY#m-0x(=}UMz2Ne9YLvP0S~QM*G_>0szNBw|)&0jhD%a>m2A1c!
z&ChB%Yy8+auPtw>(mMJq895311x#`2^bebPXmi{%9{9u(v8Tj0&y_ZOAL~~5M8_`%
zYK=O%H+*MYE9?y3-a@E%o_;Xi(WXTwq<NT4i+>lPL~JoVv!usMr0}F7&TWk`F+QpG
zk)tqzq<|AAADbUlr|XhBkH#;(m(&24ERLO4=3G;!H`iYZEgo@}H=5k&Ab>@P%dZkn
z+xRgQ|1?efzj{+N5*r<vIQmTiC<8=Q_2k`a5mTFdcOH=t=LDb#hWx--F~K_QKQ!~9
zT2oBKRg7C5E2D;`Hu7#H=d<AUKHasQ^aHjiwA0Av2j<$*@I=+Ws54-ZajF2Q<q5WN
z*MEc<ls<fP-<g5^2YEziT#373CW$~APSNNOSZ*L0B`ELj3t9l#&u?|w1)}Kh;8FdA
z`B05G1PpKp2ni9;l^{(;pjGPM_jvTg3|=v0sRzxB)6KOHppG)YSM5pSq6SDY7ibg^
zsnPfi|7JR(e{e7wUI<>3mMp<DF%THRq}gIxu?xGD5%9Immtcgme0lnsGST$fXWBu@
z{2pQ?rt-xbasLMfXsSE1_4h;F7H~_6FdRKUv^Z{}({caT{AGk214G;?lJ231^Fnvd
zc-i!=#~M|oHuF_B2X(LStsc=<9Ge%OlE}uPTepQssTT|tQe{YPB1{6NbUCz*?0+Xr
z3hvD1)ErEf&<23?wt@RufdV(qWTS$@m$rt>^$0ye`@glX47NllbP!kwq8`>7O>%nS
zEBB}ByR3%EzwG7XyqE7ZC4|$Z-4lu{saZL?Y41e5{%1?Chj7AVFT{yAeJss=wh^XT
zc;R(qP*s1wrk;S$Ta4xY30=CyR6p->>Jkh0KT7e|H}~GeeeT-@*sBskLjZ?WhD>uS
zsM6n$+#qQL*_J{fu5?*bhyAxMz?4WDA?y^894$K8@XIC2w_s0?<^bP8I5A5cW%i0t
z{j@5$OT&iZtIg<oK-O&xj4Jrm7*rZs3GikN=XCE(uLwdf3J0;0!3ZUyvkg3tQUvX(
zfK}QCumG$$Du7|0&zc|sW~lH-x~TsC9=2xyf<lN&72uOfNt{<57o%$~gVlLOZ>!vY
zdDER_EN$)^8W_0y7Y6-R`!Yx!L2E7`y+#&uK-_kHWop2Li!vv6f!$j{NTw>PP$Ga|
zqJ}>zvHtkWch_r|bW`cS&O)|E=NAR?$3L}u6}$aJR8Ny{8`QhpBBOAO(o}!2Mm9gp
z&s0gb@SFc0D%Q@xf5E0|r*=apKd)VIPc(>l(x%F>$--myjt@RNPIjR$;dDs$WOZ%b
zx9p9?$hMxM7EJm*ExAyW^lnj8DghO<*QzC7PT1+>4M}G=QBboWmFxj+JvpY0E|1y0
zp!~CJv$0~G1W2}}>*dX?jAebqHdJL#mfA498tr&Dz;)8dT-JoQes}GRwDwKY+c;Sb
z+AJOT9=%$i*ZcvHJ-9P+!H~ryCMKqTG7Ab)a2CYL^~0wI`Va6^OdyLC2_i$zn!G>-
zn^qw)3IvE*z-%=^=&VWsl4W3)p0_U`ziYT!{gRt#{)dBk173;(CwvDlOiz0f(p3Rh
zpn*h72oXnK0Py`)`h?FU^Ozwe2*5ELB4UTLrNGF==bC?LZk>R7SF1hNbZM@JWa_W`
z+dB(ha~eYKiQC4<gF&?iH0hl`9<H$X)3PS&2Q5GJN0)7LV|sfQbR(UeFzI2x#Dv%Q
zIMq#(z#KE5QAom#;e8@mi}>Kd*x2g)n(|MIYihbdqwdev1&EKa*GrbW^nX)2Mi;$^
zcYlOveLFZ9$WU1m4>(OmSIv;XD{PtWf6uJU?m2a(s9v%uQ7_)NpiLC_E=>$<ya@^f
za1{N`xnF=-lH)7;%n<_9`>J;*_4RnorTuHB;bPrudzNLgA1@@GLk#S{cltg4Qzq?A
zwbuX66%R~(Edc&S9uDAY0h2*PyKEqZHy2tF<VkAuI^l(IJ%9$FjB!93?2jyoe5h|o
z^d4?6CqQ>>BLyyi=>l+tjFlBLd<}H=GYv>vGnn-%!4m-Nwh;haQYn-J?_M6*j?6|4
zLdM>Ex54Ub8Sd1HRFhf*PFNu;R@+T8>z30X%}zla3hWl1Nupx-Rp6jcfh0$!@V=~c
zofXO0&gvC!Nh5pvU0UGsZxiD2?`^(fZ_oeSqYQuSUnw*4<(+fze8B2#+*AH=4dGW*
zaw>5$A3nu}cdca&^6ni)SJ3NS?_SbnP|WL8Fc;DjcwLjq%W?GWCs&Px?nHfhk|F4k
zzvN{g;Y0pJ*;w*uvOXfov}<HK$l}quljM>ZQx2ZP{`%8<uPM5QGURKV&avW9WaIWe
zOj{Whq29+UU2eO2s7K}XyhG~KEP0^o1r_mw9>&|!*J7vN?G(OH9x(4`wLVQS5y$QO
zv6Ne)AN0UAQ;aBdPI7L`c4Fs1p*jBZ#VzRR0ndf7CUaPKA@`<0GUY+uTx{?5X2Z$2
zRAJoy0=ghTF+T&ifrM`WB$9BsKrRj7i>Y*6khA$+5Zgawfy92G8w7j~<s&fFk*igh
zqAgV7FO6|SgOFey>~eq*=>v_eY&IeU8Nwzd6`k3hr5B8k904I|#%cK{xslh_Bglk}
z<IrNiNg7!*hpQ8o;mL4O5t5Q7Zt<+T)cKXN?tJfJE0gqk>l?B>jPK9Z-&$SverdE*
zn&r=FZ@#D`XIZ7J)tlEWh10!c_E8JJhn#cg=YQ@U+<Q&f)zFSjJwCIzqWGC7W8(K4
zcC?yrHxDXS><M)9)V_D7lYr5}`@$xr%M780_f*LYx8mq?6e$P^X&JEiJpR%sX|jj4
zo?dj%q~(|L){p#0$kAMV!mYfv@u85vpf!1Ep_1S3pHp^cXHMSpcho$0{yEUyrw?v@
zf2kq=@7Yd>Z6nGTxoLD+spErNvZ&~BWlh)X)~kw%+u2Z&z*m+TXmjRJfxvIv`TEim
zy2KA=w;4cUaR!7WAkpB!V%q{tV(~IWR+pA!flY$+$Apb9jAa_ItMqeG+y>9u;U*t1
zn#R8~r+W))Q)h$rv$M^9S^pp0L>{tCF;65%8#zW)v6-X57u5w`oWfqm?9jpd%#zFk
zvz15sW~Wc!z5;*2B~UEDe-8y&uEk`kG_oyVo@kQrG-mD(c_it0cyJ`-$3a(dyk*_M
zo5Mx&ceL-12eIJ89beH@Ei=-Ct2pWy{|+xll^yZ0lNG;rm&WfGd?GL&rrpaGHl%0J
z*By6353i3MTgJ&9kC;fJvB-0Con74CvQkk$KbvAS_<!xaWmlG4*e*<mbci5GH%Kd;
z(wz!|bSjHhx{>Z~6cGfJR=PnzQcywZ5<$9=+UND`_xT6=<I5OptTk|fa9{VF^PER!
z-emGuIxK_<K>wNHT_z*f;-5y}T-ZI|EGsK4$aF`Bsi7$-$NqaEfvHB`!$$us1#eL|
zYW)Y<(ea&`+uyVT$ILrE=&M{{5p;ZzJJztcCASbc74>{uA4Nh*^RrzKn{vKsCg-JK
zAY1;;e{A@(OjL1-$TkBQ0u?+FFlM|0)<l#-tJ<;`(Uri7Pz;+t916>?r-!oyBQDMk
zYJs$ZAZ1PXnk-4*`z&f#2TjT1H1V+*cng4S5Vba~x`-f~KRALBJ)Lr8)h$}u5r{fT
z{_v<LSCb8iw1H*@at_@vm>!WP0U8w#0|Stfg@6Sz5e!7Ypml=!C4I{kmcUws*5Zsl
z-FW`R(U*e<1wRd!^u{n~aOZQkR49T9**$2dW9~lt^u~yWT{-txK<YcSByvu&<m$Yf
z0=D6Y3D*AgrS=04!!Tatgx>VG)vhzeoHN{&ir)KqZSQ6AS10~ro^VM|x6xISlmSC?
zK%tB)?v+X7>A$`+a-N4cgT2ZlC%KB*4|++d5?()4A>C#*&%mf*1};Zof32@x+4L8_
zU!LxFI;jV&f})za^bTDZb(tdCm?IN*{r3(8O(rXrzZO!}MEIunq0r|NI+kZFNRS5K
ze$Lg(+VVA$!u!qv*FRrt+r%*91cN?72u%jy8e!0Hkh3QmxUT2(zi7@Au=#>ijsZm#
z-t2pvi$SY6A@mEvVWBDoXk#;4qd9R9$$-!dB-(o16nJ1kM;bzm#zz@zkG5xX;7=e_
z0mw^$u@Uc#^-QaphK3Cs?1)#Cj7?KQyNw2^gM=ds)7g2)F-{!91wd$m#3)pnww^*f
zsrx3+68vZ&2*v<W06DvM5NwIMZGLAoQL6>2(rePRmq-6jqQrcXuAfQMhIRb0DEt{#
z&oTMQDZa<eZO7I2MNZ<Le_HXqKa$T;Xr+ioWE%bdNJzIxhLY#pDK9Hz&}K<g$5L%K
zNn21=7%&jwUCq(ATC2G`)9#uuDZLT<b@;HodFUZePUyt5lfC7B1;dx&1aVwXNp<3x
zH@g&^%*-AhdlB3l3E$jOISei@%5Br+T;Gu;q$8@B+A{p=Ldh&Q{r$~l>gKZ{N9n49
zmGz0Sr6pPZpn5tf^llgDZ}i5;*VK2V((4THTxq9pHQD5F#ySkk8S*cs<=rOBajxzJ
zK5v^$;hvrU1V`sw;I%)T1)TsgxYZGTNA&P;ErD9ULg)E5Wa3R?8vx}41UjQbD2yh&
zOze8a_*Yj~$WQ>+6htiy0ESkcSfJ_EeQ`vk1ohKMbU94H98a2r0PC{`bdQ|dkQl@g
zAZ`VV^h|po9gNj*&4ODJE0pI*OG^VvF_0t73%bkE;}BAS1>eAcY|`Wr1ztt#8J$5V
zD4j#{&*63q8|yk;+u&<QlAfDid;Al<-uqKI*`5IC-p^-N4cCJjDEQIp(=S6gezDsX
zKaO(}O{KQ0ED2<imAR669MP)T9nX7dVa$M`l<s6VQKD9jM{qRB>BmW*gYQL-zw$ZF
z7^h>RFXdmQaG}Sr8+oZV&2<Z^(LY+N`l%{oP`>`F3Y^0diiimt^lWT6!g-MqO+e7h
zT!nKBO?I#=KUG7e!{Vz;BRO*l$wy^=T^{+QWY!LA=1H~JCekjn8EUI(J4$5QZ7%7(
z&T7@kw$-nj^wIDzETlamdvur*^>3o~U&*fg`}b_6{U^47AdUt)eW|B$*!K`-{X}aQ
z_wKE<7Y&}@VoW3mB?rcB%68hX^C5E;7@RLa7r8rX#K8Fr#1KRn$HAtDNB|JaEL1UG
z9g51~{HQb=MSyasv<0_5Br`1i{i~*{ONs*8pVt6a01r9_cUYzLn83$!I02`N*Z><n
zfOc>ghnE0wz+qrFs$quU6licr05KyF$posY?7;;EbkNcPWdoq8NA9i&$oYv)FVWO8
z5Rha<?fmhX5;(-+&IIvS`S2JaE-Q13A7uLbu=Gi3X*J<5K_Dr>(FT%2?vCPJ=3aXp
zPX3+$JI#|4CE-Q;eO=B~a6Dj7`^Mxz+f`_LqxZI$)i1xW4Sk)!%SbljZ~sjv3Wq=M
z6dpSXe*PGRe{!4e;4P7h!(+wdRhe#9{mqLjejO|zn7CE+{ok^-wo|6E2tIH!YwEpK
z8P~PyqVwK_#8S%yy+gQOI{iE<((F2Qzed~S(;~yO%n-N5==vo$9Gd%71NN`RaxEId
zLK<C6^$ohq<pUC#i5b;$ROk%rUJfs(ClxysHOXeo@#^}mJ*}GVwC>R_#t+c5qqkY+
zCOUuJYG@ay@Fj)L^_sRJ7Y;Em!aFh5erb*|%ie`-MosGVpz;HJ-s=8;zvKZJaucEe
z;utQa2;>Os4v8Rw8C&PQA^*iC-RyPu(~pSu0`#OI-HY^rR~G;UzXbFmB{dZpxd8?S
zKeY5jiISqCBEpiwUHG7589|&t+YAs?4fF@m```HmLv^XKuw8M%MJ2>_y8}0c{bhxS
zMkEmc2y;G4!%vP>P)~jlqf^6L0lS3TVi#t%u=5BY!$<p@Su+2-Z@T@Yaiozbh@lQt
zL|HYfRElx%cV|xbo!oH&v-FQqBPCToH^l;Wf?$;BRNgO%-gHSNV>@XMDLZT~4m%E(
z&WdEzRKq_nOH?$KKQFlzCY#=PlH73zlPWPOG130m)c7&sK&}y<9L~l~zskE#y#K2O
zNRmqy_O|v+?9fWr{&pDlPI(VS!pdoxi)O{McDd9Zr#G_Hhu~s<J>1rPp}_}FG%$c%
z%X5Z{d^EM;Dt?9f>*o43rc*$EuBZ#)Ylk9+1}5>+%@oDICKSwXLK=0LTDq|q?3Nwr
z|3<%@5`K^N9oNZr=#I}T1=GF{M%4Jdhf+yes#-N3S=*ovgE;U+uv18DCo-_*N&2RP
z91nT1P?%LLL4ouQUjC@h#-LGyYI&4#vH}>Ob-}~bm(RHF)~Kwh$;!{4R8+(a1-e5a
zn0O-aK$!~Gwzh_WA%M1<OEcD(pDggdz)-gj25@wEG=IQ*I+#%+EW)i@gfp1ITjH|#
zJ||$0uYLCQ1)O(~a|$MdWJJI441ZMce;3u(@5O`nfA~;|Dg7m>e0*f%Kc~3x#z#2^
zeAz13Y1hR3HeP~xgJ2$o4(-Mo>u~^0+TTa7RcF7n%xs*Pst@o-(M)?0HxI|N{OlUW
z6XtKCYHZr!j{hdKVkWMQi=pW3$=82)>BncY@zBd!ePr5G`P79{Sz!Cpr}nD5bVgH8
zrTUlQtpfSPfgJPR7?Wgb0a69YyX-uI`O%8CH&}}@?S(Qhu*g)ENoJ_O^)^rEhzh(x
zAD<8932clSjHxFhyBWG^%c!DP?Qok=h)3nV7)G)7cjMmWj}N`t4+xVv{@Oj7I^OK^
z{>MCGrrACFmp<4!RXaMp$9FR@ac0Pfw)<eaa^z$*0GDc35Xwq1LOVM^t>7+}iVy#?
z1O5fLlRTgn4?z75)R*`T3T|Z?85Eea8%%iS{7;_1cKa`58O#)Mu$k6CdO3J0lfTy0
z!=20T_@`Z9(<e5VJOH#ffb0Uqiw<)OU1I?pih6bSIueqS0OcZNG%(CdJ3H4y<ZF=O
z0~cb&cp4R8-C@}RrxynE5+A6&Vz*$|E_ge^bv+4h7lh&iR+Qo1y)@{30S6#=^sUi0
zt_(4l@aO&TRzEBFkeQr6xQ%8feBIKR+4R@FOZp;2vfR_+{A2%18Di>Nqj?R6o7pjg
zYECsPf1SsOF-}i)KRI%kz9UgJ^>mi|a_f!brRt1m0A?uh45v<N!7&ZX#?bRHWjd~V
zy&;?!DRZ~u0+Z+&EArQz+-TYqJ<BLKl_R9??1}lr8dq~*i0X>?u+ZWvi{8&Fry#f)
z`9h2s=Y3hI$T%OR^$?cVy#h=+X<Uo?<=Gvk@lQ?<j`uq4{%9A`<lo5myAYVP3shRE
z5EL=tq)iTcl7#DDE@d~ly`)|Cl3a-Am-R;5SMq}3&>9KWd#qT0Po7o#w^Qh2Cu1b5
z$R%6Hel+cQw%cb!8_JWHZ@o^N^VJ@lTw%a*FopG3mn&Tlg21Lw9v99lSr|UfPj=W~
zP6dHg4a_X`t<HKd%MTPtOJm+3N2(7)JA;VPG$yWg>%kGLoWzJ!00x1aq6Tug*rK(-
zYWS!EB!|!K$6ld=?n3nO+r3Kg<y3ng*gkx3*oh24Fe`!RomohT96~YS(r(!5a}Ncs
z>%*Tg{(_7Ev3$e+N(@IP9IrS81aARnVZW{&ktvy2uSt2tW9Enxt2{Lp*+iofOp56)
z#_gl!oiV33$HKz@(uvoj*S417(uYm;!2&5~HgUrMmI(!ReY{!k^L-xKnkbxCF2OA#
z+boHrOLAzOzmMKTyDS%S#hC?2Zi$UmmGU?}X;Q7bM@tuq-KMDJG^1hpZ_1^`T9}zR
zHaRSW#3(sH<CIxufg?*fX<XUENJalc%hlJM!_8}o$GW6n0$bZN%^Y*LD(9VZJ*}$d
z8)8;BpI?|(jI0Z@EvYK^l4;LMa(ruIU5#kh^qb7b@HP^$eBRB>TonGCzx~`ay@#2G
zP^kRoTB)A0scF$rW;6wLXi;Ows~~Qf7u~)~9{6u-+vHofJrNv`5zuy|N)YbX6D3*&
zxrA#(8$Ujrcib+9k>6!)<c`x^%K*UZk<rnJt*QKb^u)S+`f4mM{qj7tj0Oe<B2Nlt
zPq<558lR8CBc7X^gXZ}~n0?>@BJ^lF%`@Akz-uYMsFr}fM!e||oB->-AFP|}U*BLW
zzDa_tXxK)TKyL@b(m=!Xqfd@T@Loz<=Xc2d4BrC@w1=j2grTF1Q$(a8;2n*voJ%qV
zJr@%`EBaT~n?GMKC1Ws62lP*tJdzN+#ngJ)a(0~YU0z9n%$p_I-}t!J_pempvQ@SC
zlb}?7f-&v^dK6sU9aeL<A~r+Pg;))_9~i6m-WB2M-n`Xp(-~)W@&>oFvYTG^b7s4A
zvZf*gVq1rH8Pj<%unenqT4H)P&hDfUr#a>_?6Etr@wn9f<rB(5FL}!7Gba{{+Gc+r
zyCx!Qqo7?$`>h;(sBN{|NEzv>2kLJ3iFCF9DF6Mub16NmaZ<9Vt+_i%xfs#-(IK0*
z7#VmL0*jZXl%H?hOu9r%tZmGT0Y=(G>reehtzVU^G6l~8nkxoGFiUD1ZhJVOdjXT0
z9T)&1ql5_2>4I6=_Fr3#?*HY(>|o`=($RUR%>Yw{CM*Ud*rtZts@f1z$LZm$q=%IX
zckIOU^v_VOq@t#_3_(Fi>_^t^NuWZJz^<R|fznu^!7v^DPyQI7DQM{Eg3ry6jj?Wg
ze7wOaj;($vSj}D}#ZmOmI2~myR!9s6nF(`dHt(eMpSn&HG2bjesjI98O&NLTyEcET
zl<_U5bM77<6n>vF%{8pK^T2~t*W^ap=jRM8OEg%X!V+k$iMcOSXQO^sX#ae%%!3=D
z*J!6Ar`Yv`-y$8Kk_A1Qd2w+o*y#S!*3NKDr7-KGIg`|KgWy+ajqd;6dcQm6rLg*0
zgKhQw%kDEObd?5T&9pDmpGn#eM3$y3a>>}-#6?l9UYlQp6W9*2xuYt57D+Es;-&L4
z894Zz^6AB090DN%;#ZWrUbT3S{b+tG@>+6#xE1(C<$qSqWM#?80&-e_d0=qS-c1w6
za^OO!P(TY);(M|!N<BtaHN6Eokz_FLj6+ux9Aa`wj0yq)JYnVtkV24$j)ry$iC?AA
z^$m$47N1{$xcEz5CWIFv0nSKUFX-(+R|YX4K?A9^aeU7Kg4NM15Q|W7cOeQ!&2r~R
zl2@23Hh*nbj}y(Tl3J#ntb0>+yftu5dbYwb<~EVk?k~D?RSL`v6sw@}M-_U_8(cW%
zqgOjdqV6oP=7@@kI%f}l=5bzApTDK<;WdqUzD!|_EA;Gk{!#fNblw=grfD>FR{zzx
zTr`r4oo-ET1T<9+kD~k2;}eremEJH^6*Pjl`_A>=8<|2eFY7*t^QH5=Jj!vNNa3Pd
zk??jcr-^<INAj=s{I{#?&jSnK<jC&_)<J;t1S0M>oo}X$Ng9K)JqxT)#-AJ+AUO!|
zZa~8)us^|N)&k5q2p<nIGl-f4QPM%iPy<w+5GwsMXaW_Kwd!j^1`L?Y>_8BMWZ^+3
zh1A9FgBuh>XZHw@cju$0-`*>>2;S5R&0>ytssCNGTrto4gSX!8`ybRU)35Wtl=Md5
zVh~%G`+TxY5kJsrf0ezgiT~7%^GEHZ>s&nRRr&Y+cp+w$&OgPL+%)P}jRn?@1ubjr
zI-w;@Z=mvt(DS7v4_Qu%fr*N6wb*}@arro?6axe2Rs{hX54-;CO(oD`t8}%RKH2mQ
z?3yX!2w)1AKALJ8DuD|Ln=Vv>!u+8H0Xc|L1kn?MbPI6jKUrD66~5YlHX_X(u>T>#
zVgTJ>RF{G1I|M7%$QJT8KL@iV6le-V-^&+R)_-Q35t%t+djck0uRXvYF7z-CEx|P~
zVB?gOloY9OMg$7*S0PHgR9>^I(mvqFnl?Q+jww3>R<&P-M|2WAljHn5>1a9Fw1vNU
zNpVzW&@nq5=IeqIJrPZ+OF1x1kdz#@vW$&WRU8sge-)7Z(&WT4(Z6MkhG!}LqQ8~*
z2YoJ27&DW}?dU=aK2A#Vn?Hx;E0WNI4`YH)epf7S$EGbLx(Yh5vqtSa_~N$F^EUR!
zFqRR+7Yb~docOy1BATzuL^ShywfQJ5YJ&$cvC!wxu@;0kQXK!YfA1@SEI~_<x>-R4
z3Wzge;KmDw6LRl{TU<k9qvWoQKo)S4RNv>BRk|GSFK=zdfqDbLsF5H0^Ve7~99qDa
z!uy4kntBkBbP(?J&d(PBZV8z2b3{2Dai0(Z2MX39Q!u+2aL(vhMn>pfT!lmGTg8(c
z2wS&#TsseAte|-ZilvnmxGS{&Oq7fY>UF`1v`#yD^oQ(l>q4rJFt>x=T1F0~`PNnB
ze_x%@@_24E7~|r+FefDJXjKiU8?@us{}8i6`0GOq`|7jT<agu#e&b0^<~1+-d3&)g
zxHF9Ta_@pY429va)G}e$_}SD-+TW<0Pbfbp(3@AF3k0oZL++-4oQ88NcO%taK<RB+
z=l1AQdj|6BLq@f=CO^flcq{gr@a72IQN{25(lu;Q=t-x7_9js$@!d%mUdz-64pt&z
z)2DO0QC6C{VmYvgWpr@xuW6<~<z;fF_azEzv9>Tay)T0IYsu2uS|1RHs&KGV0`~_S
zUGKVwW&L#Z^j5;XhhW2ma0gmIxVoT2QA<k;W^w*3bvS$=i=gVKqnEGm^4eNNd%JW%
zKmg@Yxd;{H*NJ&Vf(WY?@Z;FN*4Eb2C(ZYv%f!!`j-z>Y1b_`HQ8#878W+IXRty-5
zjL-NdKt3V&djy%#S?22Mgupu61$S%?{ddHGO?%mo!vpZRW@Z&8i<ul+1jxznmpAhy
z)6UXeDvmI!h_ctX#MvD-uCM2sBYdu;lDQYfp5w5X{fqEETH=>?BX6L2w|2a^I~wP+
zN}L=)=XBIe|6WBz(+Ta3c=3K*HQhC;iDv9gQ@K`u!bLZ8&U40%%W|CWFZqiElJ|Em
zr|f2H;(A7l8Tn*_mlix2Heb|{V>bvf8+~S~!~Y!O?aFI};{Lkj^wYaghzWhtJh<*U
zpICKHzT_lfs+_6a|9#!&M*Q|b`t@|OX(=V{8)m&$@4HY|f#TfGR{p2tGaIq=kvmNU
zVd!kW=}gyf7_-Xh;a1i_Wh_tmh3|F!y1@5A-u9vOi4r<5b)4Yt#eU@m@TQcN6$|r{
zqvZ@Mf@~S2f*1RPywJRn0D5t|_XlUqAb`Qe$1j4-2MLP+Mrv_+IVCINWN1Pt4r&qk
z>_%ub^CeW^o`$shAu0S?`4WLYrN21U3qlSxApP-B=K*3yKLDFg&(2Um4fp3;DH)t#
zpifkf@`kF@!|hq>oME7;TDm~|xc5VCXajJ2W#v7&Ayes`VZDd1S&NpoNHH4pTXaeH
z@PrmidlvRKxY(H&XPQa$4uliQZ;g=o^8I*5aBTfeEg}goRk7KZBuu7=g=sw5byh4?
zSq;_qc18cht0ri7amtrBT$0suH4zI^ZpsOWP~$hpFO5z2po5O#mEV|};%?iC5t?e`
z>DDG|fr_{sG1~WX&p(AcqVHOWk{m{P!PZvU8hMC+!bV<(H(im}-xk2El@tD6l%y-F
z+Do=BRY;vBhIFci19#q$<&~OR%Tvh<#`zJcYlcNRW)xIq5qzgqnjiTgJD(Oay4Oy9
zW*!BFhp4amq)~|lPU-dL@%a;e%JY>~-C?xF8Hur*en|P?hV=J_rp>po&bXOWbU%A^
zwj+Kzqz?-M)j$T(^Yo)7tfR*t0P%wo>%GInp5fs*n51D_7ifc1w9)_kJ`#xzL~ak<
z#y&j02O75+s9gs&+!}HMyxOZQcAyb|5et`R_oEF17~CKw@_`OHAjE>VNgtKzM?s>v
z4$ufgw(bwxOtdz^i-%7ql?&fdyTQc*NT42AhgbE~ca#(qzJdS_c~ihdiHtsg8GklE
zy*zYTA7=qvvl~t<I&Zd3SI1v<xC%OmS_^)R$w^v-tVwpBLz&C`dd*ueSbx@<QMleq
zNOetd&lt8lpPd^cg5o-EH)Zw@l?L?4k47*PZBiF%>{U)k;)U=rzbv*;q9%tdWl%pF
zmrP{l^L}MSI<e!$(6Nr8p}_Hwz=)se94q)I+0nR8D%#ts4_Tp+jReuCcNG+RW)$qr
zW6+)zh751SXFp|mbn;`}^+qt)*O1Q9gDjDPXQib@BL!-|Mz+>l8{E#`%*A1;d~S9Z
zN0YvrUj4x{RGMOETt6*k{wmBaNcgwOT)&f|Pz2d<(ZbIy7JT0D{e$e*_fI~dW=vp`
zrdAlGFG7k>=@La*^yO*k{9v!I4>6wuez9)We1vYs$3ted6LnWl&l14;wGK_-1`A<a
zQ~%c_mMh_%40Y~6;74U-+`9L)D5KzcqY1KN!7pVlkcEK6w6wIyZt5Wx6&owBLk<)z
zT(xb%A(lC83)WUS&^&pA#8vR_SFhX9<^&I4z+s#jic)*vu^<_<=mTHP5K|hGJOaHu
zfmeR`Sm>}wnc$KKvdZ@lx%m0{bEN}ml?%qRsK6XD2#Q$5U<pN@{i*!ODC*5jHNmTY
zdq1rsap#2Sm#R6#P@GQ)H`Blk)#1s&YoiJ@dLd@8T{)YM>8}K#MW#+264aUHU_RsA
z7_NP{=Q%<5$VPt2KX;$Cs<~rTJ)zTDksWKy<W<@FGuxLk7rz{DE$Amu&CL7Y&nvDr
z&b;HhYvvqPU^XHY<Ec2yYOx?AZ;58l>D91!pefBP>>0Z{cxhubN4tgPV!dPWbnvWy
zbyZ{V<d^GFzXG|lV13xa#rOwDlTTTL*)J*WI0LYVm{GAu^L@)2UJ}cn;6I3qQ<6jL
zm1#TTO7cH6qD6G6Z+)HMvI*v`ES<-+G6n{Ov=aSrB*EN?j}^*b)%?iR`z5fDdr5lV
zU}_hE7H(+zLHw(rs?P(U5^3{9;&i~21acQ+paPKG0whudVsnI?=W>xk^VQX(`aWVg
zVhW1SFo_@(5Td(*6jI>buuy7$#tz$dJKnbvx(x#{10NqSpNS>Wk|1V;CMJ-GyFLA0
zmbN||gFKOWc>xJOjT-9<|KBf{3a(H3{6qNC%OcWmX9Se=I6R1Hju{h5w2@IWyAF*7
z48}EYD&n<=K~YePNWymab6!-U81y%GQY}x=u7z{#V(-*$QQF^*64y#;*rC~pHyx4M
zA{5@q8!GD#3?dH6)VhAKkuJ$@$k)$9o)W&f82{+{s^OmfY5jIp>#r)N@q2hI68o-C
zJ}#GfllpJo$X3IQ2%BA(#~DM3`Q7m{+(1oBai_0W()A*JSUmFF@xI~vb)w8zjN#l@
zY=+n7$ctaW5iJL@K*6$QNgjAlW(zqa>&!_2iwVCu;u%Co8bA~i0-8!u6RZGO1b1ap
zc6Kz-C2$)n1e!iU@?JNziGuqT1z}qd<On2cVD((i^PKDK>%#^(1GdNPVprG<flS?%
zS5O3M6u}uFFUJbR1hCcqJJ}fpa_0JnzHXjo$%GB<gR5jBKVHXDk<lHc<I#xRa+HnU
zf+W0j#R0OAeO2YpFP!3x58esB`1*QFRQQZq*P05K{ljS$|Cw$&O?P)g;dau+<qH=v
z5>iSud`i`@g^6~$VormrW0CrRMg@L%bDt#K-nc^9#wyn0C6bkxoJzKjbACw$SG9z^
zbe@~{1*J_yHBE^EO|n04<f&a92W+xwrPMgyJf2uX#U1*6h@qCAbm_PE_xk?t+df3&
zfdW8ft<ye06b_Jndj=SR1MJpd7AXSA1_K-0Zan}*lE7*qlQg(WkrE-mbYdeS`pDu=
zj72YjvO*Mpi2W2^SL6aApjvW$|5G7*D!0+TPWNT$yyXue0w5ZZL|ie?ZFLX^=G#mB
z&#T2Vu=%ode<$wG>eJzT`&`S=X}9K7>5TK2Rw(gemaZxI>Kl^zM865U&4;oRa?>Xs
zRpmE^2OeZo9$+uVHbr#K$0|+yIwhdnpJLf+H@u%^)LO!bD~md#P4;0XL8W-yIL<+$
zO&sOB&*jza%Qf8PcxnN(yWgY7#(9&o%fesHH78Vik?iwiy*X+3Ap~8C<_QJ@Jb~%}
zv1K3zMd%bs<$A=UcKp`iHps4Q;i-clAqi~LQqM^Up|!KD3_+|ApdV5MU_we;Z!JYA
zL0IRd4?o<7o787(0h^6exNVmBzWC<+HS%9loV{)n3oZ|~Ld$`h<KSqlpEteLeT}}&
zq(;ml8qej1u@@L9D$NNxjt@khT9SspV%upps(#EeL~9%J-wml`)Qs7v&S7PjBi*c<
z_fg!;dBW5brgD|Hw@=&W>8z9)_^1k^x)_>DN<80BtmkV(t1=>oNo0;f5OZLcxqprV
z4-fM=9^6pu;Mr#Bu8ys$n!Ue(;ZXbIQv+BENi5F>fhX*T_qGGn4d<J$dx*`p_xyWP
z?x5d{b#{|Zy+-cvs5!Rm$F1*gZ8rkKy0c7!x)<d%G#&`t&gbkzBhC(Q*XVlEkE5}`
zH_&o5jH6m#mZqXpGT6G{j;2V&H2?cM#?D+&slIEGnXkoqJ`t?x^Xubn39)E_`Rmi>
zvgQdDF>z}P0{RtJw3-;48eX0Tl)}P7?hyrY5)8WqHl`308yXqR9vgXW#o#2_uuhVj
z+T{B8qYBYF87$cN!!}zdcM{BhyOAVSGjJD7isnu{d2;zp?`=t$bZJRj!UYx;6&1Hx
zTP{S#!>jR#%2bYp7>5shQ84#FQ?*yWW}A(TO<LvkzsdG|ov-*LBuQqZBsV+1)nE5~
zpZ&Ge!%RxOEg&58xh-te-`3qCKl*%3r_JZorUlG>bZ+QJ1#~x)UzpKI<CWOx#w#@5
zTqs`ee)nDAaCKsMijuMEMOQv?rTOzD-G0)uI3zK$J>U{jHHv}SWoUi#(dXzXQtq*m
zq@*eSdN&J^%%VV8Vb`kL%lA*<7LPI^L?w95ubV=KE2!SZ*Kg7m+#Vkgocu<iz40<{
zw6&Xv+OsCmm++I~5w4mh%BzM~nM7vSY@+j9qRaPB?C2IQPkxYgFb4IK_z5*OUOI1=
zn`9rVrz_(~>~^5TrEG6a`Z^0JG$gJPXtMpB=~=Msb>RM&mXU#!q(?_bBNmvwy*;~c
zzs;`pZz0}n6c7aN=A5%LGP-ekY6b?i&s}rp3lLqIC;nG{-%VtVICL$R^|kFsvwNvC
zdivq@L&NcL^KPdyJbJ7n3lTLE>fH-9{mf6VNAc?B4{z>O^E>+N;Hln-t`3pCSY#zp
zsTSJ#mAtuAJ7AwCTiG^SI{WH))X8pT%c`H}JZS|Kb?VgOGZ7E`MOH`)6In>nKuY(w
zqB$M_%WQB*2?AUOuPHK)!iX#VU+I2G2pX{Ey#N*ufj3k0-vtyIvaqmV1_7ItR8Wv$
zLD>S66(m_?#>V2`cbtN(@cTe&*g*_J$;?AHH#eIc$nZTQ@V(2(xIZ{v(HJr?KeF8X
zaoWpuS3g8H#(&;MuxZ)l*!M=m->A+RX}_`20O3DGRp)sGe*5Q*?+<=g6A__uDR<cF
zcBgglJ9GHXDplW5zdAc1iW2Ouw$fK$iFsMsqvSncG57P5@U1LSLBKxdL4k3Hc+M?7
z$+!PhrroEu<RcVnKDn?32Ni73isj9QGTplOk~@hwIxfxzeC&vJ1F7RfIvJ5?1gb+|
zJBzETi0h{{R8(*wcgqjJ3=j@3-<tmT(-FCI!<s-qbYRdBNgEIVOwd@EAoVX2inCFW
zG%vt38kgh6174Vx=QuS*!9N^_;=J(KZ)A8nXv(|czf`(`V<?N?{i;|n{%69d#hKd9
zQk{4Mk3^Ir2S@7GAQSfo))?veo&-~2iXaE`dd}Yq0^EFjk#!L|OV8;*$fWjKyj!`{
zEHrSdAB*$1$<=&Fdhe7bk-)#+`HA;`qYkts+r7jpr4DXm$NwhDARp~12?+d=5M$DU
zKlrcx`uws6yi^l_eSvQXq6eh|n+;ZUPu5Ec+@K>mckiUDyBlg=S>dvK2)WbXuNo_z
zs61E8wlM;0Ry@FM)xZhD#D!c)Nt2(3xvh_jUJw28ipu@fIeXsNeECElIv9l1Rk?ZJ
zrYDRCVNdY!^SRt+rMSuP#VJ#SK>U7pPhCeSoiOSRzqh9!P;8`#?oRObt=3NzK5@_f
z*c|e%f&KAOOY-R4lfL9z8CB;Ijx7&^bUj50bMov3W0@IUdK)PBOR(*|FAWU*4Oh^f
zHyyNYTSw($UVDzbM@7einE$SVC-so%1vlp$AV^@5>Wbp-LL1bmeNMQ=7{nYQ&k$~8
zBUTBLpTUfl3Cf1RPh<dwG3eJ?{1?wE^I}^{m#NIkQhS9L6kqs!_&6PVodFaS+TLs(
zL0lD1Bj|eJ2hF~UdM1Wl$?dH&KFbA_7n9Xmc4Eiteq92d4x;D14=+$qsIpb$W%W~>
zFVG{dpA$Q~Qn@Qhs2$AWX^ASW*!^v@{t=s@Qkv9Nmz5Im!XvW3lpXm+z~Ic*a?N25
z+NT^4HULnhqM`8?exah0k_eCKa0@<&>w+Iy-$-FsKb}_d3xq2`*1QGiIoXbk<>gl<
zZ=|MLHE}9x8JbGB73zir(<_KJZqDax9H&1qxuchG67R@Va-lR6!27aI_t)op9nZ8y
zS9aKvDsW7V&pBCMvM~QpW#%llY)T@!&wZKj@78woLf}K`(W&B4c3zH%SG|jQ=kKzT
ze8!u&=DZ7Zt|e&;=Cy+8Nk|se^Y%_1>K9M`!N!E26E*Cl=Y6b}DVN?~X>CLSQ9ySA
zh%T(FOC2BA{W3VnWtNeiuKVzzqJskm1kIu#vpwX@LgG@V<jIVlcgBVZ0@=c7(Y?Hv
zTeJHauqz}?2L-U)Lbwx@{5QFCQVVlxizzRtx)4_i!&ZG{q=`s~DP4H?2vI_t{I+p<
z)Rq_Kwxw`yPVMfmO)u_H#)vyw^vX%YjF>T+i_cGg<SAuYjIr^=G3Sug+T5AJ)nqJW
za`@IjlIOfBuNx<i-%Vd_YuW$r870=eXKb~eN`$N~r{}pTat|F`IYu>dHBzZG`S9bt
zW6su^>2{?@B^13?s94SvdM4k63ku|#)AxV%Zxng5n0klHLUX>4gPW^_`41D%y{OI!
z#j%E?RffOqCL-iDD=qq_ri3!!icdR!11}L!B8U|OLFu4NRqHW$4RZVYmEq3HlMw^5
zLkL`IJQxuNmH8J3F1-A^pi|S=1RUa~@QllYRSDyCn`#gN2nY&#pKsN3O^$x~QV75!
zA_cIvv$LYa0>ZGfhC4bY29=$YbE&$IwtMXW)x3V13g{Vl&x^0R;E|Bv#*Io(_x~qM
zi2BO(gV>O_|7{K9ds4&oa-$h${X}nUm(TYo-<n;=>|)(>r5nN?EZj{|_fza0RrBpg
zn69!qIIF(Xu9zwFv5Nh1xYGDl`ZN3S4Y#M(-`@T<%G^ECfBN>vLf@TpzPq6YZ_GaX
zRt1$S&J1~)`E?&K*(C-t7Z?tr_zEe=3cu)9)RE<%{LxH{u3W7|vW3%4f+A>dLEDbK
zR?qE+u}54qw)6W<<z#Kcnpl(Pc7MxT_mQ7h181@fY9r#XJ+mV!E)*n89hf8OzteTx
zcm<;olYiefG-Sf-Rtvu5w$mlNfHHy!@NyJu<{^X$aJMj2kiaxFt`hQK9e{+sMHld6
zAu_Rk08U-S@{mIia$}p^T9IiV`s9#+_WwyQfj=<&9su$h2xGH@gp>DX?fH|SNq;o#
z3EK<|5SI~T%Ptmz>c?~vJ^jXXi`Q|qiuKseopo8Mt1E6Uy|<V>or8rd6CVt=(?8cH
zy76iI_aD1?%b)69WqszNcfL`NWxd0?$2mQB{@tnV(jL!Oz>|!OF3w<y^%oV^8H>vT
z`{y|m#_l`%<s=%8e}!&HU)>0yFr=6Z^=6*^n%B}^$+UjVlX^!?@k+^<K`Q)RCi`ZR
zMQX%P+|)F(ESd*CZypzUIx41U4wtWp6c>HBSSuUuD<|)=8@-5Y3=9ahya);&7O1L<
z6xvF~Ai#f9=(8B$j5_eiJ{q5h1d5}Af^p4_D6yvQqVp0v5WTUmCVh*_yW?M#y{>0w
zKGc9+ZXNjc>AR*^xt`hg|N8=JTdw(%l9B^Z2O&*#P|Jt7+NRq4A+rY?9@Xetx!-^F
z)2<>kU_nK8?`|S#b?~L4AbcNk3`4pQVBtE3hLIMOhUvZV4RJBh*bd)-%Y4SAo4~}i
z_a;&U1C&_q#4_;>1qB5t3&^dL0O68dfqyzoF@OXbK^rv^lL*CGNSZYimk{G%Of`G*
zW@Tkb097H=`=l8zIJ(-}9Z=>B@1GBj;Hk|5tKLrTR!%Op*jfSA9bVC#aoxvcR`TX3
zgq!k|FDh`#6+fJcJg$Exx0kKX&eKi1*yG__<%iluYPEMOgUrLXItT>|6+IS*?W)L7
zl^f4Ym;2}ZPA3@-aYq-CJeoyj52l7C8Zk+o3ub5!l~f`1kB&e-A~ac-smP~uk%#qH
zkyI}%$ivyW*loX#Ind0t<V4d_pAy74vv(FUnxEk;#qo(Q@f5zQ!$IdGNZI(re*00S
ziglwx2VI9P)l@w8-2#inzAh_F_By}R#{sG#PVr}h?A7N}{^ss!+x21TWq4mOk@yoZ
z(bVo=Lg+yQY&%hLalOlX3n%ADAs48YMQGx`Xxl)Z1O}Ry1TD<KUXVbB5V9cY#6UxZ
zZ3e#LEsPIvhLKTI8#l(m$qXyOMJx}VD7dFIfX3xH#eV3yIpx}Q4q+|S-F(VgT5M8M
zZEoHWWr?uHNSGr$cx!NzK-ZC;iHSNyIKpX(T){zqtxuam?bHyAn3$o_X?$Wr>(rT*
zk@2<X<&osn>D8^CsV>|;#RiLI<*_))_MqYGQ_!@>@C|m^8c~v1c>Nc^xLmlKk!m&d
z-%Tb#G>UW!V(ycr`E#ZKGF77AR`tOd(KiB7iXFP0sQS!8@7i0S-EUU0Y}aI-^Sc+`
z9bc=dn|3GRI(5;b>h~X0!gKfar&{m1>q+8OL@Ay)?R5HU5%k&CFLH3Mmep;iVqZ=H
zNmryjoE0Q4U9kTuiZ0YGN7nFt%(D#j4ug6zUddy|XLCksyIwDQEX_n)&rKhtHFF!t
zAR8J`c+1dkY7O)!ly~Y&6&M&ADw~+l5J?^qgZ`^`?ZEZ$j}C$;0L`xTlnt7HK`6QK
zq8Uurh(;FF3gFXzUg24)dLi`a5s15x3;aMb3lN~&MGJq9e6oZbNP!700Y(FY3=v2i
zQB=cv3zi82C-4|t-pSfZ3U4tou05P|+WPs)mdo~)C1EK}x2DhwzYbKim6ySze1C+G
zDdzlbM0+u5bvLDGAM@inYw|oCLu*z!o|jr($X4IQ#6gpt=%89%*`NDwVAeOR?pWhp
zis`NHEhd2<3i6KR#nUl@{8%}&il>;A@2~2QsU$hLc>8)@4VlJ8F1x!+9?nocK3vys
zoJ*LlHTr5NkdRHlM-byu&gl1xvAbHd8^`^;YFfi&k<iC(r1ST?QzOZLjY^akQ^ckL
zdjxG~zs2V+2X&X0thByTHyUzS4|pRIInOz7>>vm~LVn9H77pNODyL~WP5}xN>4Zqk
z&Q=BdgEK?;*^gL&HS(I{3+w{2@z22S$_Bd-bbzOTBu*gU!VASuiVlhXKrTk$%a{qg
zzJ!Ed+U6%Em?TLDUkN--BKo#$27e82&eu>Id+(m=$tw10%S7B6Tw5C?y02rX9oBba
z_cEX6s`;dw`E4=Z_IKQG8s%zVp8wZBpR2I#DN25AC@(iS`0i_wlulHJA~8;bf6o--
zt`X{fH+|<LVCDAooQaK_ZyTGP2QlTmu(-p~gKKy{bmK_Z-sLhatg)-qQ%m?p`@m-s
zGL`i?iLxhI52>nN6^2;&=}IiKN}1T4vm0o&7q7HD1Wvr|d|O;qP0h7=Htia}yJOMC
zWvBH7fw%zbBkBMWIM9G<Urhz!*}>g{#Hs)qYvMk9&+l|G0!c>lYBhxwpDAcZ2rrVm
zfhWX-1q1{Cbv@Pw)sDAdXy%Rcir*Y<>)gkTs-S(~_d$NYmdJg!yS~Q#kKXC!&F=3K
zo%GEoo2ysP=;*ih2VJH%FHTl0OagY^dbV=k<zSk@?{@eV+0kima%8>prLupzAcCzj
z=+-|i#*v!E8qWOM8e`%D)0_%t303{!r)gWvrV|)J4YmYj_F5}`C5bF+t>8P;OLsGC
z@!~IlBrce=u<85~k*ufu{CKDefe9-1_PvedyLSVSC_l^I8z7GyfYvJ@v5BBX&(`MS
z=V_<rIHVy9Kqb7IcM*~OZ@Fh~cE0mo);jafkexi}YzbUAz~0r>nHtzRNXa`x2eC|n
z^w{;H_#eUjPk2{n%Yr29(H%~iQ}MS<&G^=wJshUdPTC%r2N%$rcN0w!@I7N4+vzUh
zEwy5P{W_85Rvlc&+@Js!_2vUy*gj=7H3?T&R}nBBhT>hQdVKiP#K6D-@GKD;!mF#3
z_6g9BLbtH-+0RDKIJle|EFa^ZjBm~T#j)gXSk}I{E)S_7qsHJ)A2O8B6pWoaIVtVR
z&o_pP$M&d|*YU~dGK)-9bLz*x!aCg~aZxJd=CxV)ruWizGnEUIm&mJDqyc2O6$dYy
zLE9(p7>JTc1rdRO@3CVz5uGL+yx*T!+yL3?cH<gtI?odyxsG2D2x?>GOG45pLKQA+
zIctq#dSOm@?f2I|{`XB{l11hNI<p0?O7``u9}~`rms_-oLq0`M#r@(j&}Ewro5JbZ
zzASCM<cj<hL1u&3XN8G_<!>@sE7;0P$SifQv8<W%mdH!)Ms}+3PLO+e!5N>$Z-oUK
z5yS(>%yQYxIIgAF1EQ@2$jrVi?Bq8-CO25a)6j0-IO_y5l0FbV1JE>#Ti_2w?F~al
z#*EBwb^6<ZmWHe6262yS7NQ<izoVqBe{t5=^G*5H^}Fe_iJ~l;DfggDqx0P>L)E%P
z<{=>mR|COc$4***D?S)WT1d@>D&cy*oN(PBUZVakcJz6aOsYQPr1&Cj=V3rp$-CHZ
zToWpMd8t_bC|)1vd&C<2w=?m!n|RmQ;8(FnE{BNG0#t1RwnLQRgm=CGQZrR+M+6r|
z8}P`oy{7{X6saT-w8aO&E|x}=v~p4n1RTgG;E?RItu^>Mmzxif?NC!-0gDw*lf=~2
z9`F}806GLVv#zZd1^Ykm!(;LApc>12uOWWw_I)k&?857=bYRvImX_o-+LcU@y)Ooj
zU@ZGY)GVLNW7g)j)AZap@l(I09C@Emc|-HZyH`gEDQ*T>|CJE_j%dF~$-FMp*a#Yd
zIW3{Qp5*YXTJ%oC?Qisjrr&4o_#6dOZT@rNvOvWXG_2n2ueiK$k}*pUsLJ;DKD+68
zB;}=nswSt%pvLnQ>|5Z0&Vt`}4`d`rAO(_PebxwxB(EcA8Nt&5=h{9r__@Oae-t1E
zw`MRx_JAUuUn2#^4<ySB;<3Kmt)G#DH3K5pUOs$KMN^5m(1W*U%}@I7U?HIW);2a@
z0V9IO2W9XL;o{+yn<$~G$;+d3b>+oo3Wj6fjluSn=%~0Ot!<sQymwwj%}PNLz!f=K
z<`vI{Go?-zVvc!<k)SpIS4chNEp5ExX}$*0(C?}A*-S#};AA(#p=;ukIg?l!;t5-G
zCgnKe%GYxZsN~K+Uy7gejbPo2da9eS`zy-M27VsPf!z9^S92Ltd^ncozZ}Dw8UNf1
zK)LI*)vqynvHjlBSRP&(TyTH(IMvtspTB@f+O^SS21=iy$uW3pDsU8UKSKH<7zTnW
z5QJ1TZEzxB4iW<hS4VGe@0Hb61mOYXw-4OTdp@n;$XIe~L|8gl0F3wU-7J|<@$wP@
zfC$W3(XhtzKAx4peuw<;kPMBuv|t|l``!^$sXRQeXPj0uNJv>To-){D?xD|_2L+K7
zJ$&<6Ld=_-V?gEN*&wLbpQ65H2?z<7&6<66^1y$J-(9qLwkIuH{_CV~75CK9DK&8&
zA4(1?MhK;1B2&zyZu&>M6HnjW=1cSJg{qYMYZiq$Pr?N`2q&&~NSI^`=$r#cMpnGg
z-a55lvdI59I|_FE5~@`ja3*&9rzHyeJ?_su-pOKjZ-*btL|@>;YGOt<BjBzkD%L#2
z8lgK~%kUvszf+>0BAzAKyOvx;D4|>yKwZRQ48G)~jEsoB<<~7Oxj^h^gG3*>LprtW
zDZ=GY*~iCcwP*ka2c!pqO<J0+eLlg)2zp3Ua&wWy4y{a+Gkxem7y`wVQh}5|q`Guv
z-JyhKR`hXwJp5(OoMms<iQ&Iv>3%7XAlk|2zuOkkRR&KzDOHBwDK$&o8JN!)%s-kR
zKm5lP^+6}@u({bX!NbyBOk}K-qv9WjW0Nc6L%tx)Z#3V~okvJirG|T4rmFfHgQA}~
zjNPJNs6Bo92meBrjB6;K>jTB3kvqfjY!r$0_sIs=N`BnQ6KQZG3l<=Xqh0S?{64))
zd$)M2MJZBK{OGyc#4inuDiOy<S4&c(l5Di}T)`VgjCTEe_>h?<3zmCL!+em#PH0K{
zkN-nBa)@fO1x0My;CAbYDIh*ZP{AJ1er;E8mCXw4?JWe+w@(!SIgCD5q(YXQVnCoG
zFF&GgawLK!<-vwd*K}}e63_Fc9#<>R$ML@pd192K%l)2|m|>$-nD9phN;aoHMi)zQ
zyG%n#^*iAG@?xvRXXgdYxixjKam}gg@crnvd8H0J5@zMGzi&>bX?}d7j&m?vsPqcj
z=r8+#(Han9Dvdp7r1iA|-B310t1?e!7B%m*(zSjuQmQe;rG+cHW=S_Xgtm~0F{|xQ
z>Ac^2`G5~<?&%Mc-2cG2ijZDVuq2F#=0RJ30F~GXwTx85fV~~j3xd3u3C3?oFU$mm
zCQ^X~xhv}MStKQgW4e9o(=;MJ)GJoEfKRTui~{_lNiko;@PXo@O~4#g{@#Bwy6y3q
zn%KCK(}`kNnvX9I95!a=b9%}JaVjZ7EwU-oU*5#g$KcKhW@5<xthddw+Vn%dWRk)E
zV$0m%B8VWpX&~#G!8qVVqSpvpbE9t;ItKo8Cv_0(AT!@9n3sr9lI3|Gex!HjubBe`
zM?<E593Uc(t}vg#bAVkD?!x-z54TeP0&0Z(a)><>KqVxAHvtMPRcwBvs;a8m!z~bc
z#1RP6ZtIM<@$gS>psh!X{FVD$m+t+!OMJ|n+J?<eYL(Vzyb2!<wh#4Xs}CB8E+PqV
zvazN(^S3s3Po$H__67WGoHj``@w?5}((kDMImhPaR?ms!*tv+RvvpVtSX(J{2e>yN
zDoQ{2+qZ8M6-ZVvGJ!z=79~V)ASGBJ?OX+iAu9(*6kw#<Ft{S@7;=3iV^v>B2)tMx
zB{U`{ArUQj+1vD|%92zza3Kth^7>K2@Tgv9W5lcJ?%#{N+CD`=<{f=(m&L^2tNDE=
z9pAi|{Ou9=oc#JbXI_t(%F4EQ$o!7oVcZ)p*Wtl${~4~)Y-gE7lh??PxI_9G2+1bk
z4GpPVpV<oh(0~9mhwKTDjc>Tkv!>n=&p>Em@Dtdor1_N$Gzxu*TAKMK0hC|7c^#ig
z+}TgXsDm#5ej=+gpY1a?oZHph-wGgXj)@fLK6Uu>W6~_(O(Kh39-yk29ensmWg1wq
z*txit!TJVlMJUK4kgTsSM^NkK1v6GzRSC@9Ynu9@w>=C+;)nWfEks6zZSXt#S%})q
z2}IwO1`uABTQ!6>HyqtgA4BrovvR{}kOeNAD`;d?R4M=m!uKSE9dQX9{BXaAuw{AJ
z#{!SD?!xa7V*vNo`=j$_XHx#!M8%%88fRbMq*A}22`3n|b(v)HbRX^Zd}N`ustaSG
z8>*U`8q-}3qTT4o$O~qPAH7y)t!P2bLvc()!COKPh!x~C7~pV)!$Wm3Y0+R7vr*o`
z1gbvZLm7v|I_iJ^D=bki{r~?I@(TRFeI+i{+X3=Ifz#flUmQeDYZ(P2N78xcs~PLJ
zIT+m>L}rinVjny-HKj%IoP9p&{0E+bi83TlgB19~hEwi|M;tjxV_ojXKYPP;n>>{V
zP`?CYbxLMt<ghKC5w?Q-$^{t?M*ra88=&66J@e?f=ZJ5UVsxa*ZB0*qiRDB{DnLF;
tin{<^Kt33X{=d(W%E4Ll|LjxObZPnJU*wa<rBUERML|=(Le4zo{{X~r?Fs+@

literal 0
HcmV?d00001

diff --git a/_images/507b37f5c35704f694d1e81f64e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png b/_images/507b37f5c35704f694d1e81f64e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png
new file mode 100644
index 0000000000000000000000000000000000000000..bd8a50e6ce25f4b0f2a89b08163206e34f9149d6
GIT binary patch
literal 24913
zcmce81z1(<*6tEEvIT*yQj!KL0umx3NLYY?N~eT?fOJcTTLA?WMUd`pk!}<LMLMNH
zx;qzf$G7n8v(Gtu-*fMC|L6LgfW=&Me)Eel-tmrij2F`P#7Kx}h%guoiTLeXvKS2h
zG6r*?^4Ej#$)n>E=ivYNtVNWq<;?V~ZM7_QF?Y4B&5g~hjSaL<Jl3_eGB7jcX6Iq&
zW;^l7+S=TTpM%5Xj|<q%EcH3+Sqod>CWp*#t5{($N3GF+5A^FX+F>vc%EWJ7m$wU>
z?Q?Z%T;1JU#5(%a>*W~ha0!c2e@$AQe)Q$Vt6R5jPnr9j|6!dH+dxAUPA*cDw*BR+
zuUCw3E1&x-nNt)t+7CnNj4#*fU9Q%fpJy!VnKTKVEnRn<U!Pue8OR9y#=hiKsxjX=
zyt+1#+c~_my_lz*u9ymA!eHDCHVUrde!@Ip^s>kObjL%R0r$h(=paJ$L&Zbpu><G_
z&g1`2mmJoto9{2Zzus%*`}@<U?NQh9ZRUDamd6`8dNkza<0e}ZQ}2dePSGqc>&`KV
zrP6egsj8|n^`SwJcI%ew>^q?LwfZG3&w~n#h?p3wkdVfcCr>1#q&l;7YHHVx^BDSf
zq^ZPm=~Pd6(3&<igkPm#WXwD*_G+S@cOVBFR#;frU1XE0n)4`nYjyhSvL(7NhIxE@
zbD7^^B}rRHr@g>pc#8fZPf}~5)cwj&pVHR*9Wz&dRC!a(WgB&}r!pBiIhEYSzLgR(
z$W><Rm@nV!pIgq*EbnJ|!RWFv<oW0SQQOLP+1WDe%<z~E*32~w_>i5=V>NbXb7LbZ
zD(c$q?(Wj!=0KfWqwt=Z-zBFF^H3d6GR6!gO=pGq()EZU>q+fS`};YMPKv9RyOeh4
znM@BVv}J5>JI!|IN^)^=5gk5kZF{AH;!P$ettF-2jj~V1UsQr#urBdjxN!IBK@yf!
zvDC`<Roi3rp%mogaXvnWNoX$JCM08!zjn=y!D-=oeJHP#i;D}t>uw3t*MWh7vF`^-
zXm0zJ?X2A|ZoJQ*Q)beWH`SK>_NkIg@zT<ge?UO|EBdP}%*<c<hA>LE_MdlXlPLS7
zzSh}{Iwi*h2im+Ksk?U-EG)7PlF}u?q8P5tbV=O1*YP_6v43c2YI~|usxx-YFK)Ic
zKh<WsL$~JL+IRhj57idw&fN=Z`ug^qNq0+23zPcR${7xh#^AsaYXSm-%dU8$qW|`|
zFP$(G#lBgvkBu&T{_+0Hm)rFG*0+aie3i7dwN<bii^Bo_{&7zU$x=kTDW{sEpAr!h
z=b8`Qu$k{y*K}SDID0233Klt4_uD(&DtnQHyP*lN{get!cDFaZo<BF=vcYZm1tYOn
zbj@+%Z*y$t`we}6<27z{nrzxUa$59zjxQabthl&%cb1N4cdk*QukfyNr3Xpt>ytN`
z)M?%=EPdlJ8mkLdXwVPh)cW0nl-{sakzeWL*|VbbS07vdpy1QIe*Jn^?@Ko~jIAfW
z+L50#Il@}!5#}CTa7uH1e!y&<%j3dWWo2cq*?^F{yL(rTr-z4xpQWWGlXy74O-p~N
zV}PGuOjsD5h=>R*qNZHJ9T{We8H&auG&HJf4uQmn4_~U^4u@y7ekOYTFIz?2WX$L;
zWyf-|1J(LOe`#^t(sDmHH}@IwlP6E=noxR+pa+<_EYlJY5*mGbcY)>nd6`HdXEiEL
z+_7QT{;uuv{k@i7+!@Jt-*Y%LN|a$m6&fQ1Sq0{L3$46-eU-BH>T5}w_f5z#>c>IO
z*48=IKF4F@B?D6m3Y4?5ve4!T;W4^oGbQit=_y$xzvAiX$;!eaVQ+7*T4bF}O3xpM
zCYFTuvh}h8?!kjjFePUUnqx#@oti%Ae#<fK<Igqjl3gCFZyovi_S;(FkOys!ao5E&
zXU?EG5p-B-N>NB{%QaHUuOdPBxMp(EnrnV~I`!GJU(PTw$rm-YN(rYYM!Iyn%ixxM
zP{Sjfn+*3j7EvzT4{l$n9fAZ6+0SrGy|~{R5aKL4`-z5y{j#jgkdr5J5*Rx?FMBe7
zzu?4=FM1S|<xh$8zqSz#mHn`y)oVh&g02`mK&?0b<Y<0EAjMKVk++0@0Z-rEBe*Aq
zz0mH*kJpRLN~K3vK42$bW=K+3zFFLj`}5chuK6cFv`>$Y8)2oU$ERZpzTf5*`2}5S
zewMOg`I^ko^)1f1%Zk~!*(MX=nvah8^fc$et(@XYbQPuz&!Y<;{qW!-IXF1@7j(Ev
z0QbY$oBw!4>AF8IIU4|PlGw@EBstL8+&WFhe1^NR?<8(8YR@tgxTF2+1W4}nW_*e{
zD&Df_|5oDOjtFCQU$yO5bPG4CDE<Ct`)}$e;wP^xAM-hVa*i0eS<E1e;H-a$b9KgJ
z^sAeGbP#X!V=a<dL)_r)D81~Dd3nF|k2ISU4p;CS*?--20s?fkIk=!>UPNRqo)-DZ
zn)1&-(ca#PxuLvGivC#gYid`hyu96O4TiSs_w(qKGBxPQ$#0#+tziG}5&xt9t$Lwq
zd#<n8rcrR6`N|b#Ft5_%4dFblyW1aTVe5gLc~kUwHl?EC8l2q8jN-YLKuwp)>LlD8
zCJEZ~(N_c*_ZG<DA3Q7#2jj=^Fa<Sr(#nc8*oQRZ$ixT-39Ipj_9U6uOIn|P866vg
z6DSU6rn}6!7>r!69d14_S0c5%H%m$cUNQ)(?qJtD!4{1U%enk^9nA32W5;X;oYyDA
zZTrs~G4fj{A-@{HswBI=&H}HrJi9Vgb0xaEx<dG@5}cfz<fWyf{QQoFI(vC}S<m(I
z{ln<g*3^iB&&oJ_9L+N(kkgo}+<sXP%=7uvr*DEiS9<Uu>i6G&C#RxH{PgLDxL<jB
zxoOel+0vc09@Rq2_*Zm%F?-wN!VcgiZQ2!-40;PJOv`ZF0Tb5vzH&xRPEJ8hE$s<D
zq2bK;tPnxRJioVZb02)fgBLekooe@sqhM8>*u~Cw!i3bmGHFeCnP)bTaoME%-@S+U
z=H_PYw{KQU3Vce}uRp1ebbIrkPEuHSU`~fkF|}QlIO_`A#f!3@r1UMwYXk);cL=#`
z+rbJ>4t{(J{yQ8z_iCG*)Olv++qS(Ho*zDZSZR^uNWT8`;MVqXV>@^UiAFAy?n`$g
zg-Sn}0Gv^}b?aGcCNY{QH_fY-BhSy?4T&ptTu+fpyw}cO{a$c6ez{M~hgz#u$a3U1
zJjtNu6}@3!QSR2tWNd_>qv6Nj33Rb5ldWRN28Z!kt#p>}=C)Yf3F5qd(~~sUKvGh&
zHR}3P@?*zh!6)-CdWB^IrYL?qD|_dT4;~&~FFmQb$WKr5)9U~5a_TF0E$^$L&I=+V
zBkL@4#<HmO-b215$E4@-Y+tbw?6%ylMD^mwQ{S^<VRBunte$yzyesV=l^CB0y$-m+
zK%3x;QV<CrG51MLt;-{9b#a%Mc{n+n=5!7hV>H(1`p$?SKYsl3Y;bU}U)<~0zv=4h
zOZu^cYna=Ji8)0<Mb&BPk7fl^5CNN3tI|W<?;V?pc59j4qIP@weI-H1wYJ<&4OVf7
z^*&pZwXRafe3$J-F#vnTXO(eZRQm#ZB{nveN3V|hf_jlW07S2sFBL2;KOpFrESJb4
zAfPrhG_*Bq=9)VCEs&LiL&~<_KJgW!a0US}<rnY=Z_i0RUf0FlZP5DfeM$ED7U8Xl
z)1^zbtY|c=U(ZD*X_PulHOG2t*bNbYNA7q2h%=Ycl84pnuv=R!Cr@5Kbm&mP-0=K-
z=EjCyRZUGX=SAFudTZWW?L+8mRsS<s`v2Qqri!j&3Zx|TBEXun@?d?s-Ff8IjN`z&
z^wLwJmy04IU)Ep&zxec6e(L+ATYlCjrhP8QKb)*A=wY5f)~;d@{9gQD`n?oy?S9EW
z`8_(gSIR&AUiEdZ`S_oHuW|lg{9Y0}_&xBxfA)LFVuIGS;UiVfQ{8=9xS5~%%;|{x
z;V+_^zl}!%ndpU_lqtRATO!36>_68x<JK9FhX%%MxcVgx=L5Hx*jN-@fOU^gOq~9n
zrGwk>rOORhM_*w_-b&?Ij!N{|kB4WF;w+|{s{VIvA=r$uV!n!dyF0m$XH_>h{26ck
z_S<h!GO?loMy}Y|e1y~w;2$hFC661jTCG9l3^6gW&F;3nercFux@!E)jB$M!pPZ18
z5Fl}BfS&<1^ujJ`_w$UK!CH-ukB{d-TnYa;y=bVGzkdB1a#7a{yahS97C?bKkD9Il
zmXZfEaaJlKZDeGGhA{8%?f54NUq{&c3bL{>gba33)#E5A##DTnq0!TXfGsYLsXqLw
z62vWv=H~B<Y^D?7O(s)PGyK{O74F@8<+3$-E?qh61~Q3&AXi~prkceV12QwsavK9e
zLQ;PG&?cd~BCS*Ns(q#_J0L794bCVtFRxOnQf9_N0eGqum2ADKzGB;e@hcWTp52QO
zFlc@E?%i4O+QvpD4GoRlVH?lif5!_IK5+W<zD0WG<lgl<oAa2`h>?p+SznEEnqq2h
z(baqah*a!qO*hE|<G}7D-_JKy>fP%xD{lw*xw2lqrw%@*b*R!am0tPb!#IRH7l&(D
zxwx{@oVPa?B_t$TK0YOE&rrV_szb^kkhC`2V^r&ZKIE#cX~1~UWs{HusYs2Xyv^lt
z)>Efqx)<@#L`fg}XY`<91HqJP&G&Zhy_xLBDToynceYkjb6vq|wwF5E)(3IjggMF{
zmtR_64`g~t&Gjk4dF0C*^klH{wG_*&N*VrfhGbch{|F~wA`-v}o!<{Ba#;10Secod
zGSrKaU-&O!#bb(PU!kp}zH_hR99E~goQ>`53gDd*-@oS!U2AHR7MHzu?_Qw}+H&x5
zs^j`xJHObYM~}Lu)ibPLl^y@(mtTU;D!Tog>?I1ln+>1t7;<ak<Kxc?RU4#s?RC2o
zQ_)38Nvpnw2=9B2zMh^$(F*-?`C7Y@=J^X3Ub@etDZ(%`Mr-f(*$%Y(pO-W4hFQ#j
zBgJFds|=yH*~_y_k}dDGDw)J3-?3#GXqMPzO?9NFf{X0VLAa$QUTXt)7rTqa9$lJq
zaBBr?YwNCo^6e!_aL4KZh?Ky$A{>E^O}2hx9ya&CxQ`O!0dFdfsnYfSR^ScNAhZ-?
zKZJWqrBzap6wnQ@iR9PT6;c(G4O$aa^G(Cipa3;FFMK5AVfo($Cnjq>CHC0}Q1n_g
z@>%lXZX5Lb%_`s^#^LlTt*yVL<(&esDBD*QsBsx0Az~`_w8iOvx)-J1OtXQq55u%;
zxG!_^c#hTOoA#-O@|wM4a$c_Im5`K_D|gwc&C}P{?|h^ngu6_d&(mRy(RoEiaPjjo
z332geh|b$e?5+Bo*ZXvP^34V`W;sf5@Eg-UEX^fs@#7Z2sL9B^?ZhJ6`3(ICdr^RY
zesO?c&a<*gLZEFc5rw<+;89O6dvS-6H)1PaPW$>Bk&%*i@D4Z)Pkh4}O}B4XF6r;n
zNd*y=g3Fpti|+5()l&e+oIiIiu{UK2Ld^l!z0%l395BZ0MZ#=A0Er;i*TzOEge}3k
zWdFb&DNKNQQcxtf{wves*~=#1-(P=8!(*v<XMZF_bRpcXYcnc>3l;d}OSsPpUf-;8
z%G4-bsU#D=!iw`(7-<qxie+L-wkG{Z=L97Udy&L@;eWaZ%NWou(PE6w@~*qfkrYRd
z#=uzupxd?0`3ocNrsI~64wQ^Xx@vIiH?;9|s>RZp7CiZc3*0bNFg{Iy{#)zwnXA*C
z-)d^2@44W<C4HAvVGwMPG?@SMBe<6}Kk1pWP&{9DKhkx#Fk5?n0@??qxguY`j`H^>
zzYBnGzy+HQb`PM<y#Va~6%E5pVViI)CH=~<b5{XBDijS=c~jBz;i6m228E|mcD^!O
zgV%n9ApaC*G{>Mdj2vgaF-h$2qH#-I@o<J^OusQA9g!(geu06BUD^6Q&UCopGB;*u
zyR@e$7&gbaBen#51uMI;5}18pj*I}lqwssGJ(Zqxe^$Tq>F)~fZTcdUfhN_P@`5Z_
zxb{LT{ao0dO6oQNV2S~5u0XJD06ZWI3rmYlr+Ol9zuiqTfrV=(>bMPq(WuJM?$^$m
zctXg~gij_A1)%T|+z9b3kLP-mHC;AE{H#~ol~Q2~IgADCUos!O2Gm2kBHsQ>-EX(!
z!A?v(0@Qew`yVTwruQ?7|5J#Id(Hn{6b#$Mw$COIy%AtzKGq1!$MgN^j&p@#A3uIH
zY>IlqZ#$O;2e)U&YUmR|eF%>vkWjB9BO^b3`qWnL>WTunj{SXGp>DV>xa?;=(R31a
ztOW3uMPT4o2cMEny}CO0x=PzZ!km`Zw9V}RJ~=&oI-o~_g9ljz1^W_ibFi?;!j@Sb
z4b+@^BgP;T$gb}H_HArt*_K42RHP!%9^prEu?9wTNSbSJ*jIRJI<vgPi-L6p!X=|B
zFABr%oMh2gbjjeuuK~PIUo;vY9W_{)Xr3A<FYm}uXJw4^B0N$Dgdn)ZcHmv3gMxx+
zB5^Z@$u_rZ*nY{|pVX=8st)LWYg~9Q8HguVh(^IND?!Y=3S3v`qXI0jEdJr)>A~Fk
zC|rai0}*RhS7Kb;t<3V>qPfy_jjGR|lYsewJ!iAxhns)SONkfU#)pSvAt0H8IJNoB
zX)$YvS<jz2BZAH!>_FC4h=dqjw=Hrt&ERc`h=^oGMLoLn%@F!R5hy$RJ&05N;^4&S
z>glCO2C~!enub9fL=NDJUA^cLIM#-uOStD9u8P)L3&j5L=E%{mB1%^;R)o_pivo9t
zl}DRfTZWS@@kQeijw@iG4S=jtC{r{vO!4ye=7DYYlK!e9_z*=1Hz?VZqk4PQ(lts|
zlarJE*hzB1@F~j5`Y{<mBr?^R8CkZwWn=?)Yq1KuVsYb~R74Jn+ukyPnTMz(Gr`PV
zlAN6U`}QH~<Nv|^EiW&F8`aKLO@9@331)&-F%6wHn~^U^+&w%D0T(!culZIV%m^<k
zv@wXk152P#X8QNgCD&^F9-;^U=E$k4h5$QpQk)+`Q(%@;0T3C$HVKH(0e&Dv$l1ZY
zXRI+Y^ZE1Vh9&!FvaJ`E<&r@Q6N(cdC{5|$B3-Sso$Xc--d=bLZ<~n%_n}YjQ{YSD
zz1!g_lHf!$I*hPbtfZglzsI_8c9<sfdQ4eaN6wr-b?Qd>X1&=8Y;cW5x?^YWG&M!v
z#4*8_9yOgnTLu9{uXc4<`5d6(;S=T)Co2AmhW^qC9>2%g2aG=8^wqIYGsB6dX!u@=
z<sOW13OK%$4kO?gjRwk`ffCL(AR{3OU>Y46NdmgtPu?`LWd}HGtN)A0myTuZ?c2Ar
z-Q@tq47)z)=B~{Gn+m*yd>%$4xvuVxxa{rQ&%vYf!|q}^eLA*x5yE4lLx-|h_g!;c
zMRfZN(97unm=uegVG_6Iisubix`9o8%XC5I!{-F&#UE8E_wzFx)@G7DNNA^Q%dz2Q
z0CDKAm`52$!YfaI<0x-4bH(=1gwx`-`Dv>p{oTUxh_}GQfc2IhT65jQ4&ctN>-I3?
z<MPM<m5a@>+(8Z!j=E7xoVXwR1q~`Xx-|GGZ;T3OY2B{#wQ=vw<`1NZ`7#8U&O>Be
zwmB9`Q@$G`;(ZAS5ofT{Ap-UvU`Au@m&ee-(8b;o5itO3jZzC5PKzSQYC<{#ky;sr
zV<9w=`{R$bgz?t}bHiIN1uiTuHUm3vw7peURwi-#w)Bk~pD5PhCZTAWhxh5s2WMeN
z<{Ne112dXtjXf9X6a#;~V!6*TnIHG`+62rq1?I7Xac@HmXr*_#<G?Keir<33<kI6A
zWicHbqlzj2(Zq=re|vTkTyv{YW@!T0_`nMifx*Gafbs-42A`s+N)B(|y#`*9ue;Ai
zdEtKE9SH6^2pOHCz%z;ap&ZD`n;uDE>>8dQ!L7=+7i`RAU=fmmnQX8-kNZ%X01({%
zG3x|?nXj4-m>&^2A(sqd_-&^Av4t&4o%^02%O`IjytC2@aYqIs7k|cEmWpb)TSXlO
zm-Q`>qZ5)4$oOAn-A7*V2#>DFtG4%?w^l3^+4ircft@>FFih&av)cJDP~nq1F|rnD
zWdKKeXKU{xFD1YwaB3V;qGfS$k%)vu0VpBh{ES^_cnn9<lyO6Jqf^6}eR+NI!@uDV
z2(@P8I0x%xFD?V^t)ASa-^QL64WT^+1;fbPdo-kq_b{RZV(|~h49tH#rl1VC?MM1B
zyg9+2&6(PkmOv0)BmXIy58No&!IzA}8h{Nu6_nKD;p~e0x$JK0-wnAW?zlE%apksM
z1`y7=?oP=-P@gfqYb$pPef@zGY!YxtNw8zVIW^zq?TgFFxdhCVq#v4Yj<Ss#;89dR
zJbWSdyZg9!y^xL)E<IZDQG0hXF;Wq%3Yr@gHcGI=-Ny6aK>Ho%1wHYe=^IgXNE3dj
zrJ94pK+#ED(vp)9=H09(@2-V~1)TU~a|P`4=fCPcd<fgpt^hdv&CTwF=*Uc8n1?=c
z+#aEhyeYU0kPhI*_4pns0DdS}i7roCIDzIN6-VQ{(ZgI7L&S};BVaO_^V7kv&X-{`
zOv^U)#O@v7_>F=gj^K1mA3YBA1$o@8oB>alA|<qW3r5pjY@5k+{_55H5H@sx+f7DB
zq|tRZ`=7hIjoY<T-F6brcgHa)foQ<=a(<T#+n80-6=8}d=Uw-TP0M#4C;qa3!q_d6
zf&$}U_e(?8tpgHj{t)29aS;&}O@V;L@cV0uudqW{-ueaC6menPXyE#RJd;rV$1_Pl
z6#{*p)Vl^aSmpfr^D7N-9@NhMJ*T8#`KSTGdO)=a68WxcUHaf115YQy5-;3_{LD9^
zO<8x5GZMWaa|zEe7`-3TTxPkdnFTu#X8{ZQ>gScP3uhqdFxWZ6pyvvE3Rs_lnR!A|
zx~4pnp4!<!$XNo%0m-G7Xbg;9K&JS;2mfn@GIKMk>X4flx0EjRIu1DZkVy=z(tcB=
z-3%rRvb=ScSeT{uWH~WlXzm>%hrRvwbCsx}oW#3&X-8pA#;PjOA}$t;o62|fIx^bJ
zcfzmQzH2vZhfo(e+0}{WN<BvJDinCwfREME(z-`XSzXCe0(qP}klqmTDSkN+3v-xN
z95fGPEzBYFq%4QCv2mJeuHooR!2`UsPsA8Eoj>J(f?nSw9u!<@x|yr-<?pia{Qz!0
zoZKtlo$SoiOs^(T9KxA6TZ;+Du`|i<-%B$(Ej&RO7^P0*KIau=E3?5_5r3)S_;<9}
zWh{?bzZwd>4T#CeGMo8ioYBzW-yA}#1ADtlr4Fktz`Q{6T_;?~SrxE#ZGC+^1T;Xc
z9|IWNF_F8xv~&sZ+3Ng&rcq}`fUnf((_g{p#DV>5U%4oK^N(AC;YNuQ0LC4V|F>Bf
zybcQip}ZpyMxB~_TS}W)>@>>uH5>g$;zK&4bFUMSUO;4|W~=0N5?Ha?#?H&`%I;KX
zyoM++E#KQQLs1#HEM38I#Z(1|@jB(C1mmyH72E`7IiuOgtiMDBQoFTZzDQ%&W>HQj
zD(VFA)G&WddINAcSWcY4k#MaEcM*jM%(xgkNxO6r(Y)JqWqY*2O_BnkuqqrqQb9mR
zqX!{v*9U?Hu(S`el&v90@L@K;*V21;B~iG2BKN;1PJqEccnl76I6_dsJeF?!@aDH$
zhyv49${I2J`Nc7Q`=z+M_W>9(ofj6?gzPJiSDyl0ZM>UCBCI!V^!1gFgs@7~m)Dq=
zrSXN`o;zk!x0bIN=(+Yp*u|EcnUg7vunyjVi~!=_s88yrIy4YS6rR#Fx>(zbU@VWI
z7Gzj1X@C9|60W!XPMtc1G!LcQOW*wB&dVkElLY1~{XKXE-mRF(mmaWgzr*x}+3#65
z!5Ue`{|E8*X|6y}T=dB$Gc$UM0A(hXGb$s~I_>@qR5o|rier(FHS`!)qptreFbh1~
zeMFF`aP2Krk{Q9A*M9vfBQE};-+r7PkaLk}=Is1Dhd2$F_OrX#FK<pGfdNE#@~U;h
zn<H->rqlCt0OYVzF8>XJ{W+t5AlQmc+i*y*O3R2(ej)H?N>2J5vq#R`nxsJ=<|1U5
z*WI90(FKwYB(botZCDB;cKkyW0A3!&Mysoxn$yUlfnp}Yc})d^A48CbAOnZOy+W&T
zI68l&9$*p7gZLcB=BO{yMMp>LX4)3E@NMY<8!gS6#UP}hY*4JyPfIOjGRi7_{~x1J
zf>C&fB-J~mf6cGp%fV<sp3CF#@fh$>l78FU+c}WN0Vq8I)JO;Td-Y*&1cjL(Ls*$0
z$jvS1M}G9E>|_>fgcM6g`ScH&LBpxzuqY4o1jaud0~pnr@RvCuc@2$-hDc$v>^{TX
zsVXW>f2P&ho=%>R4$Nf^DDNBbd|b^A4;aA4Ag;=}W6Erlkb3EHR{Jjymz9rC<+P|z
z)AJ*z+wx5AgAKk1n6ed!72qGQ0HshrxBSCZqTGSq!l+0#yAkE%!Ei6NgNzHq$rE8U
zbO`v9#Jiz->xT?g>m*wM_L3ht5)GOXgT5jg2nOOoeWSG2w+h%$-j_iTDHU{eb-zMz
z3>&8np=8tE0-Eq+d4ICg7)+SgxLQ!SdX*D650fgHlbV{EaA^F~-?<icq~D+3Ss$=9
z*k`;7Lbb;Cp`)9Z%s+GYFg!@_`>0!uJqfw}wN58q8hv1X%FDhj<&U(zr1v?7FZ{80
z&#NHe0FeW_B<Z9_SF9qyB@WgRNdREX@ZzDTaPV*0N|noXd&9zcZ)Y7+59S?`pYY@U
z&?`vuczSOFp^Q{G(h#FVGLILbf4?~7(VLr_k%%ca(d|VpvC*S_L=|NSyO{Q}aTDY%
zkfDcH{9>5<&IE#u7Z3^1z?{_$ozT*-$a=H1oO_r1r`J4x3JsJ~7~uSIZoP@M9<xqF
zSky2$ruDAfH!T!F@Na~ivv40Sg&tI6%pko>Ygn;5#O8tvTSzS*@n9hvYX~xx08=tz
z;?@_(&e8D^2wVftO6l|bu`M-v#2^>lRIpi7)JF=clB?DeuQSTF3;|c6KVjH`A{|Io
z;<%t$rUsUMG!IsFGCjY)1(4L_mwi}Yp$iC2zacCcg4fAvS~Gdz2G}UQg?3itc1Af|
zn}PZ|m)f2J5)}QaeFQb~0Ee%g7ptg>wx&|SH*mlOd{$$RAaE+0E1F)}9FI%|+W8(4
z1#TF(N!i`b;e<f2&jNrG0w@bsDG5Y$fHo~eop4tTUL>Vjj?XBXV(GIRe%au#PdZB<
z0%+P8DXa;CtaNLvx@}+l=g+tP{DS2Vza9lykwTFmu+U~@n+Z73gu&#)mTLXn?$|b;
zYZ6l6;0uVpm;R`9b+opSprAV9ysRDJi2FxG%<WG_#VA~I>54`1Wv9qwd=>qiA?zKH
zj@-LWRtaQjrc!1Eq|{mZTwr%?Zd|d(O@?_Pz;_llmMjBSHWry5iu|lvhe;K>!R%-E
zyI@ViL9Ni62Pnh{BJLsxI_ddvNToo{@bM<ixWuemDql{B^Nkz7BUK+Lf9^&Ir~q%_
zha)gDOV>hBLKvo!@67($2mt+w&52Z}p?-);Xk!l4LL^o`nGl+zRe4ZwYa-U+$BT2|
zeVN3;<)`1z@7UWJ*qeNJ-_&BC1s`OAZQAYcr_0^e-`FVzznJ<EZ7_P4>+j~5@1%pM
zEeFED4^?xgPM?Oek^kWbE-<TOViC9vs+k0MS&4d~SQrEa6>TzXU0ctCw=HjQafm~v
zZwibkknnRpfw(9!-y75~zaMFyBgId8E$6iftQQK&`1trfsHTtRft~3-{*z0$e2>dl
zfMNo?P6m;;oj8PLKjlAH-T<75^%Y71(c*P9BMvJQ(Wx0lwJH|-ft&;wWB~5!jY8P8
zv-iR#()JUJ*C?Fm_zEC8Hf&3J^JmUl7*rD|+egKrVXY{D+j>z%q$JtpYu!dRi$6rv
zU3XS#xb=LH{t|_?J!T7oA9;joKtscW$O72*ENpD@5YvNIyaJJ2a&H+(HjB{Lmqx_X
zRkB4P7peKOabH*=!zW({d4nC@y`4nio#ao>Y?P8O9p`vWdy^oA%FNHNih|U`hvgj{
ziXefeIJ65=%aVo&!TYgRxEFjt2?hufrxuEIqa**$hq5;Q`W-?f5YViG#$y@|US@6q
z?w!ne!SsCpuB=}r)9@dJC*<xjbCO(Pvsi<Q6Y(Osl*e#i6CG^Avx^5WmI)_h0_MF1
zhBM<6*!E9>CP)g9pcPP;teRRFuu~>$>e+fP5z&Ky(mg(2%oUVpKc+M<Q{I_p@MB6)
z;%l9i?QM9#!U+g$C0hUtpt(da+7Du*v#S7Vg!i_y`K>3SKz*d5Q|*(5wfXOm5$Nto
zDc7yH33q~Hfc~Cb7^-X+ssTG`4J#3v2X;nW7SgLrOFlab>eA8=md`zS*zcc-kYd}?
zyiHlbPZhx2w-ztY3?=OQYrJBfQ%eb9?wdcgIG{(VZ*0uRQp0Q_hq*NJl?I}(xsv63
zp)sYfy&&tG-w)X-8j5AeMDaVWrB(YfW&#r<S(L-F3tX8$;4g3iNx*13OvVSwLuLZV
zsg5}-Hnjq2Fb$dJJpda)x2n5tR8$E+_R`}?oAkfV{6@Al;gmI7k+@uSCZu%tzM!?q
zi_n-GckLmEY4uYsBOAKu00CjKFGQ)X;K&gBn+_RAlvpeT64aK1_4DlLU*W+oot;V$
z0X7fpuJIBa>>lXh0reIMJ%3xmg?(i%qIMu)Q;z+v9{-=J9hgP4IA_-6ZC<01P6JS!
z6cq}cXnW2(%2wCNwz;MAxa{4`m4ubI1YU;^HGRJy!ma&PGWd-CkdSDR=AZGQfq?;q
z%Yb9@;NT#U_x9AU`;s4WN=m&}1{xX`aR9f?%Z<XBMj3^okYQn?)zIt?b%E7I%7-II
zjs&oWQ0;3-S;J$^4jwpgAmIhYf``-WZ~e;2=e|)b&xZ0WZ5XZJ>VKWRKquPJ3BW>d
zV|`vh+1lWl?9|unpMt=IBy-qw-TEV8?pW$o7n#sa)wc}>o|F}kLRtjZDk&~51A&my
zN;gbuAQNyS`?+Gj8#itsnkPO!J{kNOQb$3iRlVb_Fia;Tl)4v|`w!Sf>~kpn8aqzT
z<W#hi3C{<O+01|a&S7EX(Dl^*Pvnn}7gl8T3h<T113UMSH$$wbX%A!_RyyvRwLdB=
zVFF>q9gfl!BuR<^)-y2oPoh?eBS&QBDxU#tU;){=l<-cDY2hfP$1KMDA3)<@<S|Tw
zQVgf|>5D+YhjjZm%XIz%#)lmwH0raM5~_9YDYUo1>e;m3<xPeYhh!yy5Z-ZUW`II=
z>M5C!G$?3S!2Wbj{I{ev(maj6f;r7%LSUl})M0c5{Fo#(x(n1tX6HDii8I>nSCsQf
z7w)^d?WcCYW|%%d==R2<Jc$9&0MO#o0NWKHdAtIV;a33LC?NxcGm6`0dkf`|QC^(p
zP|djwOV$cn^?1;18-eCGgx4%WuRdh-_IF??s7nd1)pHUYx@H46j-XkuixhU%h;-RX
z25urA!g?cMvr)VWG$$?~h_PB*SY1^BXMObO(G=%9a6jE#ReB*_42B{)ACq)E;lzO?
z_BpC{xnb=xAWvyP$f)7EV*|+*k(PLoNes1^gdPL$^@^`Rqvf7I^Gz>F5Qeb&k0GeO
z^tJ~w<?(E{x&@CTX0Gy-`nB1pe6X6-PY@@9oyp!BX534PzUi3)XsE&KklV5Mx2dEZ
zCoTYSvs2<so&v5C(FrK)2n#jUm7NIha>muwH5A;AHyVuR@4q|{rkunpd73Z$1(c}A
zVx{TX#s&a!ec0Ot#Rp5Tr#6E*8C-$W5HIsGHm!S3PiQ8RJ(7(i^>E@GynYZfm`?3J
zHK^1IAWnYM3VbF&!wOQTctUCJc!H&K=g-+}r>%(BiX-7IoHtuf#fAWehhY65`rNR2
zA4K>-E#BK2*aJ2CkjrgACQPQ|r+m!2mUsK2>65yrhcR}(@U6RQrD8i8Jqa&T*Qu9o
z>M5ld=OpvKykYYw$T=BeAVWx&@q^X|<$PhvCPA3PZ$9XjQM#rITnzv4CCtp{Q|fLz
zb@C^;!w+Ig@LN-Ahf}rbXM)Skhkidu#*o%yT6`Bov0*5O1dO<bgBBwrl0wx55@Le9
zf`uma`AIN#gb4}(HO%sa-cS)?w*&Jpt*sq{)P5@I9Un}fG65~K5Ge3GpZ<cWP%tJQ
zku2}@Q9ZGI0w2@=@lI^DTCz3y)Va52oq&Zj*0GI{*8|y*BDhYd5lMtJaeVI@IDSLm
zv0sHj3=_z5;q|zgP}ClsP&9k`jl;zurJt>7Fy_2UP9n^$Dbh#46CFEuZ=z^ABh}gw
zz@jxs2@oK5`zGQ}I5~QqK1cVP1u!W<?o_iu0)K8arEsQ>OiTngnt>P*Vw8<(0ZQ6T
zsVC3mCm^NvOC9^bolZnyk)-beX*)#RqmUW03W|OYv(&a|BJ=wlB%zX|Ov7RFOik=B
z7&`%_!*cM-!D@Vg;RXr?VX*nE8U+(!ZKfdHL3$S5EagkOwd62H3HYsb-evvj-QbH=
zDp`0K0dmbkd<^I5u@9Bu5P(d#rzRrG7sNwtpbSPhEF4awdXaUrH%D0#(6#T=5*aXb
zB8}<eMJ=2nW^?W*#*IAINtk{2I1CA`x`yt7R0ySvP_CMVf2-wbJFq8kZexIZGiXba
z$vt`kb9<%NfxpZ$QNXfH&u+kNX=5*D&A&l+{OQqxr|QImr2yf&As^l~0Qu!~2(0-8
znK1=nsgyxa9OW<5Wu2`gmxBnwVwGXl(IoV2Mt>}9Q?W`b;(y-mu|3u{ywZ7QDi{T9
za8he`O><Sl8GggaKa05EIF_d6714&h)zIma-Oua6bTfsYtj_P8UP)xoXemMO*9`44
z*S*~kE*+1{fDd7RNcw3wuPT&c=ZXUubh4CJK$gZw&V+ei;iGTW!6_mY;wT08;_l~_
z+LHd2M@fMp=)6?jV_~jVN8*CF>Q&p2Vb1`Wlu6eI5g_^A0`uD(D~2Rq624jo1TYO1
zc-#?<$2ALes{0dbUoZ0v`>fZ$6(YY;uMj|}S9>hGPpWREA8+oV5*H)ns6pfwu+s;$
z0K(xw-At}j04c5f;9$E}UVkF%$#J9V<MW~CoSC16q_U9J`JdA(jtrVaFBYAd(DOmp
zJ0>P3jDi@W_J-TRbf~MY11nDfb4TYq5Vjiqo0Us_dWryNkEf?zQDo38e61>@eTIXI
zmNpfLcl|0njN45K$H+fL%5_=K`hKb|pFf=Uyb*8Z!uv0=n`yV1(t@bM=o9XxUns0P
z9H-9@nBvJ-M@QI0#ELQnpx0`KRRW=coGD_mLFp%U4<B<aH#m9yQ1yt<$**;rJmo1%
z6~;F^>pvf#zu-Obkro1uze{da+!8RkT^P$q#5LZ@EgwMWWT)&%=WMoljikRKTH9rF
zX_V&T!{`03d(Qe*_=9`PIfOcGq)WeuUL#Ys7Y$uDN^o=!)^@j8d_q(51kMvsV{M62
zjPUM*d&3{8+pkAdfK-*W@wPWt$fSl=g0hyeDE1mj5(()AH`^L|+X-0j$B;PNEoT*^
z#*hDUjG933!RT=VE>Z42);e-lpO*wgLnJ^LXcm}9Z>`Oy!irA9o@xh0+H6_@CbRk}
zS4t)A^0T4*>RV*WRX!IZyYH+MaaR%P=sdY$hYzL)$?QNW0YM9t&IE#p1U8dSwlC}~
zNPW^19Ze#Ya`-Y1V;wTF|Fm#__+razOMm_S_gXwYiq8|32lGdE{NgQ2Z?FUyo-qTm
zO#$jyezkyeqMK1Z_FdH?WI<WMleS*L+5{L|2yl_)G@M*?6fydcNSOqSAg!epjfw;;
zhN}%w);A8l$lmV!9)$B5Aa_ba<y4ULig29cbAR@%&B2NDpp+ww>$|g$Y=NRD>(b}j
zjQ(8P9p5JoaI<so+~bl<bIZTjn(dbe56m?AD~*df*f7?faRfa&Gt_y27v?ekE{$EE
z@4RnT8jliOpa%&@vUWQHa|(UI5!%VIhyH{*r%TV}rWd)VzUhD4-Ei!*vvl>s{irdY
zrK|D$b>ykM7hC496-2xd8WAJ$+5~Xf<LqbJSJdTv2iQbWUxxJE9jp<srAv#+c0)U4
zLK{(PJs0wOW56UxAzBJTo>?K@x|-u;o4sMR%X#Z@{O+mwrRPMaNc1-@V5Afjo*4V%
zb@#Xs#t58t{^CnRzRcTeelvIc>-}^^$PcQQ*fq~pAnnk<Xm($hm%GM`>}2K|woCXy
zQJm4bz1F=3yD#Ho@)2<N_Va3`vw~c*W>+fB4%o${5IsCi(bT_o^|#UElQ&)3qJlK@
zj8#lN%QXBmZ2arGK3%M8x*}4Au&}cu1<NiJU!>j(=SSE@QrjKNH+(WORL;JCEkZug
zeD(ACh=ia}!wiY7XyZ}z<dFR*M{HdEZH=qNb35re*B*#x+ePU{;}Q~3^%g!MX&cC%
zc^ub_&zw7F1SJb7YXfPxZ)^1p4T`AtYquX1hA3(Rexl$9-@ULa(U2nP&H+p>?zgnO
z91s{d{+ZRX45#2NRo~Ai;8jMrRw{mUq391&CBPmRqB8X69yrE5GCe{2DaS(dT?5tL
zd~~vlI^)Z-lW8*!f)}J*>KZzeMh61bm!G6@cNcI-vN9BD%vR#H3rIWYudJ$m7!D%c
zaZ)r1xr73yRQo9BJy!)AhSL{sHsHm7QCjAI9eiI4>sup*WI46786zf*1k9QIjS>-G
zC^kbPBcTe9GPT_>-CdcE^7!#&NbVaTN${OJY(|X2E+u|`e%6p@$_43xpEc;fAQ^oZ
z<VfAw`o2Qj3pekEU6FwZsPG5he-s(^6MS5_NEB!9IGvRA`P5gIuctH_QxmMql8j2M
z=|#+(q~MSB$AxP7^=0H-Q@7OQw5`%3)*{%1Ct~}`{UgiTW~Yuq@G4FqMvWD~NUO)M
zV~5At3>81_^edzk3DZpOP8yWs%Wx|QvPSEn3LFTlk@5~Ki(Kw2E@Q3%5a$74!O6ZL
z&#Tb^MW2_es;do=Jr#7YcCT^S91Sd*&Feut7}AOcM8tlG*aWGI0Z67$oro2C!aI;L
zJ%W?#_2PwAyUV#vuB+v0WN|W%B<fWv^DoYgrk)%9`Z~W%0fNHO?o1Mrrf%Wy<#x%%
zjl*{${IeIFMpB*G880xYPpPN9Y`zg+&fOwK$$XJ1s(0QeMcYC*t^K3bfs!|@OE#c_
zLtNi1%NPl^-5Meu(10PWopJZgn>Wvh0|OSo;0SpbBpQdJqfrX)&8DWN51qwoxZ<R`
ztRz`n5erWd?bF|_b&AbxJ++t{T%-g}Y6_m0;(x}Q@&1eHK-wF*y>K;~i0;8deA9<T
zRqoO%iIyog4hY}NKmY2?q?5&q-QBx+ls@MDJAx46k%1lgz&D?Ku91arl`6Dq+Vn6{
zOVObx0lRw&?T?j)3!@&J`|@~p*L7B0!wQpg!>ohc+oU@OmC6``phCq_PcMP#JQLH1
zR}B><<rLKY<kFm0t7$bgk9aJ|nUP{yub{bJ-B}wP`bM$!VAw|m;z9-6#iA6-^@od6
zjn#sbJPNDpC52!AQYf7Q6oY=_yZOmYz;8%52wK&TXR=}*2lNmAD<7NTf5(db!;nFz
z4%7Kyd&KN+mFC?AvT1Q^J%_(gaqI4H_!(qMo(Uv8gIf!#_IBp&_R?1~<X>a-<hyEn
z8Z_R7whzeA_v*;VLCLf4t5+RCTsk?{ld_<$V7+`<5sDF1fzC<*5yHZPIqW6dsn$f_
zx$-?1n~A3Dz;kYaTnmZvkZF_>SbSE@fq(BoT;@ZIgHpu|U#U-%Y{(Z2=&%0DM=U5g
zCUPB5YB;6&hSN(baIEuLqIQp0e);=q9#k26MtzH@2oyOJkaW&RQ8r{=QQ;_fm;&U_
z)FERrSBkmfp4c}iFzetWd&}{MC)2B?9JQxd>DLvl{;w!gX0?AmFPcPZR$N0f+SPoF
zx+rM<kI<E+w;$y(&YUT{!cq|O$4%r4-i;d+9$b$;@#LIk)uWS2&W|uw%XyAzG7<EG
zPoll;uPv3(h9#=|%6L5@>Rb}I&^ahX&UCGwtLj&U#h;fQs1i8bxW6Gp&&Qls`?d^&
za+G<Exi-6)uzx4G<jzNQiM3z@OANjGukN36wIlqEFosmbvIoow_#cEp1iGNzU}tV6
zdSo`uU7A_#K$M>HYa+MQrUSzQ<jZb>mSaytV?{0zM`PR^Ef_trIX~gIUnlgj0@18S
z?QJqswr6-sKdv8R_Qk&6JU-N4{VQ2TE)*hh=*AK9Qr>uyx)Ak3R+QPv=Na~W;qlRa
zy|`C+6|}ep19LK6_#syrdiPPK<JFXOcWIOr_*cxI7>QqP6Hj>8&bQ^0e&u>Fs-UGU
zYit|*Jgea>Y(DBt-JtL982W~vNSZ6+sQ|IyvfJC>`23pNtS^cTPm;&3s8u-IH9fmE
zuO*`Vy0+KH(oE->XVxO!h&_z*+{M<rEsw}dGn9XICxqnpp|AGtzA8c`>Fkto7dc{k
zdwkkt+?MP}m(Wxk3!XhYX+mZyI$7w3rJlXs+Uj*Z+gx-=CxOgtXcKPCttxiF_#b!v
zan`_i|H{ehZ?&78!(=vsW-IU8T|V8{Q<5HAapxa*wRn5!G8$#6&r*~)*Mv|5v(v@`
zrKpa-Bdw43_=!fS#JzHaS6qi%wvpf(*M!C^N+BmZ;sUBx|Ct)A*W=KSA?$f*++wrU
zf!HYOOT0oQ1ihXH5<KuXD550{^9f~gn~I`NJ081Iy*PAXQ0R^v9(#0<XD4SZeiSc|
z#jV@*geyBk{9B!q_#thH`167#3{_D(J{`0*p80N$oOcfuJkw^-`1pv-t$$=`eb=s5
z{7mBRCr7g9+D#fO)dyY&CH&&QWMPYOV^y0x@LTS;wbK0CKT3h0K!0SvD+j5%3p#L-
z*{w!_5<D`uxk^W}3YCeMsX=cNE^@__WK>)_GkDa?-j09))91O)aP7Ppu^=T0TjOoH
z;GmX+O4=c^vxE42R$4)VC3uxty*2Ov33xz>j6rlh<c;WoN+TK2Hw1wg=)@GTJQ9w>
ze+pbQG?{4(SdPhcW2~&IowqtOdH%V<<YGIowNC_lDm}RJ4(ngGjwl#c(n#q*#gm{8
zw>znSK|{<e?8520x)){HFf-$6*h{h2YH(&Wo)TVLe*sCCZ%_F_N<0i0=jrdi7repm
z_qdVynA(8l?Hla_4}|sm5;w<=bi&9Xk!lk_VNLGvimOHbY(ma~xzUNH$LQ}19X6<N
z?^=3nQ!0JDT=15NJ9EjCHV&50CO1`7S}{pMtYf}cb;^pf=UScYruh$}dhXm;xsryy
z`}6mHG-Y$Rfx3Gx4V&|(%>Hw$(s&|j(!nkhextFKu>@mBrBu&o+X0VrI-&1RW>%_o
zQV5!Ma58(*T=7=1{7TF4^BUf6JbzK@_+b*_MJ_1NIQ%7eV3lZIFG1~xW9<Pu8n}zT
zy5CaC&##|AUk}E8eUMJ!0`r&`L5<s#xx$-zJpWPq${S8-r1&vOPqKr~2l`yTLtZW4
zZhbhxsN%L7pKI^0H+x=wzD>N?PFcDBH!VN{nZgAf6+yu|U_Bmgi#)jD_xI6IWd{|x
zNgyRI;<Lmg@tJL0I-a4Gpr#$foD!STd{&R<CEmQW>O<Fyn~#C!toM{)|3MN<ak!7W
z++tXigqrgj<epJ00ssIlt*tgQ-?_wFATGTG^(}oCHS{>%59B${Ta$?)P*;Tt44`VX
zc#{YdrB^c4`mn$QUYeQpMCJDs8WZh=W@e0GV4|-gls`v8xdg=gslY3!f~MZdsg~^2
zBM{m{9}q%B1f<QP_z$^2Xhi^egjB$V8HZ^xVZX;5Y|zeATUsIH;O4H)KPAsY|8$ZD
zQ{oOKv5=8OqC5z+{C$0WArmw*KW`@2#%s`W3%J+wwlg4QLOm-C0181K?~XUz>jwKr
z8PO0=jrK>!xL=AMz=%;lZunE%kN?OjW!f1lD!yY<P@cfYg!u-<9%$5yv!76?f4b^n
z{{eI2c!9nU503(*wP}FQs(wYE3PmAEDghNU(x#>vikTYfLl&r(4k=@iAOZO1lD|av
z`ZJ!*e^loQcKUX0(&TPChI8y{`u0^jtEetEtr{{?PK-?$zJit)ZI8<szSO0^->Df+
z>{}{n6Oaz+8puoB>9ml~rJt)2q}2N(XH-e~r<{;xp2H4k2y6g)gYQL}eJI*iFC1oo
zih=mJxCv1FD)tpAm6brw$Pm<16sA3p5LO{8Z-(gaTi;1cyJ(q1#3-S&b=mjFQJtGV
zj@E1Jr85-e#reh7Jw*dkEW6CMC?kT;kIxa&{O-bi!_lw>(q0{{v$?^UP~n$_rG^3F
z%A@|RF^E{9wo5}8AJPV`d?MzU0Nu7alnefK0iX2(iy<ml2hue@D2STimtP95(hYij
zQEFTj?audWcxTj@*(zP?U}NQc?cNPkZg*9)GS*mZ@V4wukx&Rfb9dWm^Fbjru8e=^
z;9*Mv)<O^;G=!IDLFwZc2i-zK*YXEk?l%f7`=Fd!`SvKsUsw073ebO|WnF!^sa@ur
zEy}QHHcP^1<gZXWRM1`92^c!Z^u`H$yu@V>M>YX5=qK_ZYKPH>CNBtLjDJ_a-&W~?
zlBzi2?IAK60lW9RYx(7c0*ZE23=RkGr_rf%W%siB<Vf)OyfhQRV05@<4A=fr{(fE@
z3Kl{=pi>SK7y+x9wt$^Dpnu+ZqHn=D0m+3u=DB&USX^bbtxV0bw6@nPQld5pu!Lev
z=l=s#x&xJ<?SkvQr-8CUM1WC8T9*AwC?QFLq7u_0Ji8cH!IhT#2h2o$!!H2!1FE3!
z?I1z&`S`x7gy{k#Mj&;jnr9pg=OPj96o`P40-g!9!!9jA<eU-r6<kv;96omRDD&8m
zyub=x5-HpVimKb5q%B#1JZ8e<5p*cAIsR|RGFx8?l#Enj1ZLGYIozf=TT6fEA@gZu
zo8|zetoNZYH~OE1T{#I^DTW=0Ap#=?>d@d0O{lhi7wtH9AQmPm-t(7=fS_58k0FHB
zx(TaP4_tik4%E48GrT_r-%5GvND%h%&NN3w??#A3g%h<NS&%$bej`TKCDR0nKBz4Z
zVHlD=*iiZu?BdnE%uN2J&NITk;@QRu9i#$?$v-Jj6W|4r0WdqXILQYrKPB_CR8FE_
zI2j$r4jlW@YK2)~NGr!EJs|(}fzV1F{@Kropa81~{?oS#Hy2lhF_>5_xCGjuxE%_f
zy_%-wO>h{p_Jhn`_7!m(E4?KdIc_KS5+H|@L<;J8H}5RBlkf|j2(&U`ug3F$pUugk
zze#~I)T(bgnV6NcJ@^$qbmRnIr{?p=%`oQpl$@aYcRLL)3xBWB1ys7SvW4$3iRE{2
z*f{_wqwp$Ih5tk_5)Ta?h0us5te|v;Rn<2!Ox|*e1F5GHLJvHS!BsiBNiyNoQ)fWX
z2et4ZMed^_e;~WBxllW`pZsGH!!=x~I0MU`lS+SW6k!5brdznNN=p8g%Jg+#MFJgD
zIM46*++sfhf^M>5Q@Qe_-`sg$19Eu3K@ZB;pTh<P*~(HBY4?ZG0x0MOmquhCb#{Ny
z|6~2T_J{eWE?@~6#rT{dP~I)+2NIWFI`MX3bx@Jg>_E9DNEzlx2@6jBEDrwJIKo1t
z>|IA>{OThc={n7^iRM_pIA9YBd4Gcmz_=;#K_f0GBv(v#zJQu@d|M2X0H2d*+LuCk
zK{W{)YR!?3d%?3%Wl*|ixdu2@o4;&>wF***KXvsoq`%p?tqWygsKZFn)x*Wm^g<Ex
zCziI<*K+^8G19n^jpK-97G}SA;;-A|P(lU-v|GS|7F;EdlK9tV8`^D7x{zr^|L%DF
zA9bBMhWc<t*pD4|_w@V*O)7Z%tlvWU3RFf6j-o;`!QG8v02OahQ5#4MK_-d1lZ>vR
zo*p2-OhlD%goM){2^rfUCENEK&w?-l5mnkUz|(v)!2%Ju8al8oin`Q5?v@8s=1_-L
zd{V-$7^u-{2!Cau$d=~PqbSI88bK4R%UGBdueLUM5=J34sQ6TZo?>m$HwjUlI!Kvv
zyL`jkQ&A}!G`)c8OgZ>AWI{okHA}RfK4)bF!dVTMO<h#_3AwzfmUwRj98x=F@)b4d
zbALendk5sBc<h%RBI3-fd?(Lky?DR?l8;;UW-h~X3xn~yTb<>GkX=EEK<I}Sla|I-
zx>!vM#nz#K(^M2sfRg48II?1-1P8!~l*^eVOI`?pfoqq9RGrmwKm{$j4CqMUey$ao
zAS!tds$KvJ%0Mc&9cqGBYFRUx0rKp@^Ll@~P}2VQJ?5b82|0**6hOj(9$Lfbss?g(
zRGJE@>lL8Tra;-_cdOx(AgV?|a|QmOXm@KGdK;(>xvUA~3qkigBv(Pj(i13CkGvve
z)M~Y}9&IcP<yxIo46239x^BS3aBls3Dm{>@NPF|<O$TTjt6skO*kdwK*5?2XAwu&&
zO0GKHnQ7)c_=K>w`^S$TX@EgG_IB6HbxlA;^+DppS>PhrGzL2n$zlXfIUU*ubpQ$6
zWn4U+QJe;u(d-Bz=c2S(&@RH|(61pegOox8QBWPw31Aw~rco(~dgHO6+VO(APSrCi
zpp!Og6#_ASD$_I+C8sH}@)55yv$5qc2)pzeg4A2nbthNB&jy5fqid;iR?{%UX;8P-
zfh3w~>2YzNyJN-3j6fus4j41nlKyZuZ!qkE6c`trPc@ZL$bBH#iF8t6A548n`)5F+
zaiKG{AKEVfv=)cd0EZD4_$+H6kx;*dv~;ORAtV%Ag`%foHjOCI^Y-9`?(o@Y-YBTM
zVAI2j`GbhQwi{M39S)2oC407EI}0?aL5<p$78hjzeSs*!cny}B2{iD2akzE}AZ$RA
zUD!^i^@lEl;9m)DmfD+H)R+!}@i)*Z2^us#`*0aBLqWk6wA-#eo-v;z*mi|JS5W!D
zpzb)Edzn{290<}k)&@yOACQBJ5PB{dg47kMqi>?_cO<;M(xw9d{ARndYuZnK`UVLE
z8OTSsG43u9f;`h^G!F^WfzD?S%!aYZ0m%KTQc(BLM8JO_{WGv?@1R+NgddtuC=R*@
zZ{7}?|3r{_LOdZ3s-6+H!}Ss49UuXMn#yZ1c(#1a$+hllTxJ(*d>L^q+;Ri+Kv`bI
z&lDvKfJ+`~Z_$1d4_s1`jpuWkYC8`Z<^FiCPsIrT?Cs3Et^(GTFDOk38;u}V68~pM
zKvd!js+J}c6g_G_wF5gf5E=<8BLhaIVRs!itt@D^e-;NI;v9C#TR2VMXY5S{UA7B=
z>je>J?d-XBjhEGUc4w`iLsfgGX1Q81lx%c8uK1Od?n)Hwvpc9hg27?pr1;*q#uUqf
zm|3@MlnZg#-5f_H7q4HxCR$Hs(JvhTq<}mYB>JJ8tju_I3z~dPL;V}<f)=0S++UQa
z@1-h&Ja#tU2x{7pDtC3Uh7nMq3aY6_5r!qDhhC#qYO&!ZFZ;KPApd(m{=xcrIFwzc
z-o&f@T3d^1w^0+X4>k;Od!y(04xirmrnwt4IQRfwFssEmrB-E$B11OB2&Cxm4H)?>
z;~=4sqf<kQdbJ$gB+{{iO^yWV^a6I$V2A^)C-IX|8RBvsH`F5hvaU=WBG{Ndm(onF
zXLb>eG%M}WqEWmF(Bm1Thu&Lktr+xy))*ycpe9oL?c4a9FIm0~!SfwPhmJz+vIF#t
zK>h&McnXqU{7xJ1L3^kGa=J=SvS~;aD2%bKs})EXWT=+~DL;CY!~~5SJe}O5k<Jdn
zYgA;1YU05ww8LimqJkR_3`7pX%rr%#!kxJ8)Mn<fy9uB{g{qSWT}09GWiXW=X5ETG
zkY-rnb^sNl>@NDcGPAJM>S0k!1;|W0!@x#g_2n8-Z4w85g>oA3DX5bcVC%pgC(z=c
zZ-FQ*9jO4pmP4(mX<=9?s8CQ_1UTKTK(vboTCR6WjB9y3dj<~8A98VUuCmk|Wn|tk
zB^ML~g}Fy}E%$;eQUd;=V~jQ|t~(PAb`0*(3SuQxL;NlmX(qrmfl#Y24uD7j7r$LL
zY7&C_k%7h}N3Wiy&+^*^iaFkgUw%VkmiL=9RUn+b0}VIZpqg2og1tlyJTbV#B1>Vw
z9kNh5fno$Oj;M$eG2x17%F)PY)9*E0oo5ynjtq1Eb{HgSemNk-4qbz>q{6o$e|K(^
z=w=}pwp=hRuqm1noWUF*^97v&VD>d^BJV>98#9y*Tu>``pH^>}3q2e}ipD~WtRWEX
zkZ_`$9@RPUzz%h*1gwRWB)NcwP~9k~Bv-ohBPT&;Vry%+APecb4kUE|MXkS}^=>!p
zV<vIr0O6P-Fm*1%?gpqcL!coCNWe0n%s)hH%{&#h!4xV20Ttc1Zj@byM6QC2j0`Hv
zhv`Z{T~3kXMq+qgQDQT0kVB$wUN9m#s9{5$-(W`XhVv_c-}VC&aRux-szjciO3mzm
zfIcHNvrKhwqegfo({)GEb)%9IsrbZCfezBFWYNRY$|?Ygb-~J!FbF8a9`%a@?JAn+
z$AbCr&{i~Hq!I$=LIzxDCphvn=r)83>vOG+C<Zkkuj{(i8kq*#!`zNc%>jqum%OS_
z=yS~0lK@F4i~{5*8#6%(+UWp%l`8F}M-NOueYXl-E8gqr>K2~29CQWkc{;2$ODC9}
zNSK0E#Z>TZJWNcVx4!H6sz6a&?{6ndx`3BR1CNrO2X%&1)6+(PY13^Wa2#EOo<`)T
zTj?s4kADC-*oOX$)HLa^QF16LD7t3E8a~lOz^O86H4Umn`8-Wneb6mF3g<C;mkUf$
z2SfuZ;58qrXK9P1Yv2p+S4#&S{@IW0|FNa-e}n|M-oa1pq0Jxa++gb|(}0~X&jf1&
zSTQgcQbE(Gj24=Mj6lx^g#0-wRITBlqmOZw7yil9bAX!{4nq4aRV)Z>;@_Q44u->q
z+N*&-lz;HxfeN-IUXn+@fey*~q0~kWP7`fxnjm5{L|@Xx3xh2|2D-QsxK@<AwXucx
z!mbEf2!7uN2epUTi8$yBSCg=V^!d<@81=XUKU942Gb^f9K{__ju1$p)WkBad5GI#x
z{2&JJSogBIy$r0cx>Y?7>UncXr}{9oZ^d9xFU`4yjbYzuGgoYA-I9^rV8s(8a~Im%
zwhCQCdl$;=pzL;@!j}V5@?gt=M^*~e>PUgDX$U(Xz*l67U<;^mx<7^4V=#xumiU&T
zx7It5fUstk#FX#Mni)ddLj=;V&X>r+F${p3QDTmgD8K*^G?b?;l=YVHu5v*A=uL3#
zh(JYcXh74ba2yDrXd&3X&`Tpx?GEY}2HqH)Dg4JjEX?@WJYw^pQV`s!CRE@SffbE#
z-L*&ka-m@f%9}!cfQ+fb6V^;VXaEFtkn`=3-A9*$sc)@#f^W02pcT5~0u1gH{52Eo
zXEDnFH<aAbqOHdH0?yYIxYm#>7BK+z;*lI3op*qxsPPx-Fa(-q)%Opcpu$E2Xq=&2
zQwHtS944YY+w#pccUCi9+kuKieDJ{b7%$}48!4!Iv0$nY^@lpZq+MWQx;0<2{7MhX
zZlI@4J&67N+Ol-SP-&MR;$bPk-9YOaR4@l?4*G*cq(}Yy{0Lf0>#@4y&~D)t%oi#+
zL=9sDHJ723J_ow*A~XV0h89>XB&?_Bv-~{fJc-KN3?M}Ddi5&ft1vJS4LVSV!oH4<
zH!GHB8zd7dFxM(;>!5#y1V7EuquTZL^%smHp@)Wo-|E(MW<DI_fLveb_obStA!r24
z)D9@RF^o?KXHTHxOBSCLc+y<3UZdR{`2fcOd6NlmOPk&YTbT=efi9cGFCHWyP>~9|
zqLmGG;NSP-10ed0l$C>Q>Vmk~%g5k1_#QNq%-CS?Pl2Kv9(MK&RV*GxIuubKZsLdO
z`Q>4$S+H2}a!ld~Ekmef)anAW3$pCne(20V%M)l#b^~e(th%JYwpoKqR4Dv?m!o|9
zlz7opO6rW4Eg&V--ohXihSpWG38w`vlmM__gc|cYuA2sGW&=Kg?VmU@@(>CJ<?ZjH
z_CP3PD((lG#Sp)82JM({<!k9VU;bAuXZO-%7>4n;1CJt$B8V_4=wM<WAkDI{%+1Io
zi{&gc{GuS8J8_MXvk`$vRun@TL`B1CX5ibF%LZA33_39S7;|JwAChZU&?-%4Wxsnx
zCmr_>Y!C1AKKFB9_j_GGGnZHM$BtkvE*4MGnXqxSM$>gL+BXt7P|2oe)>qzd3W%Ez
z|79M5rTwE1?x+Dq1&Q~SbKUej24<h)8?e$ynV{Y*b);Mg*Byy(QBJNB98lF_M1}We
zfA7q?j0_s|)Z+yyldAo0W&=G2Ok&MOBE2K=#O$ogRPL>5PIt^yi|J-h?o?a8PG`Jc
zyh7fzSS<E-+7{H3*)N0X`h;pS#gP`jIIrrkoR|j7xW!^g5<`2IDN026>H~u$<aasw
zA#<rm?SVYE(<$IQY57xwKGM4W+RzR{I_}*IZFr=0sy)Z9=7224*Xg<1^trGDLk_J~
zBdhh-1Y3P4va+(oh=Rm@dIL-7-X%i}uLcGQO)_8J4)E^cqg&Gkh&1=!#8Lke!2SyA
zY*SiM?fHb-@<9R*qG<=!AEt2s1gr0s>~c8)-R0T!UCeu4^on=P>ETif|NaYRdCIHT
z$g4Kar%(VW%OToH;b#egUAP<OzOhbdPwmfO(0`ZDJazQ<V#?HcNAUwBkT+I|&G1P@
zilJx(_7hqE(jEEfpRD*ocW%XvcNFe4stZJ#m4<4w2sos3wi@7I7U=b#{KphUcTJw{
z{Ifv1>4ZgbZ{WU=J}z>|a!#^NxB$Kx)3*T<U4UthYPz;>=5nrVE@?hWV4+ynF;*ry
zL4=R+FK8WOttHV*!Utp0|L*qVtsyFLJ57}E%o>SCY7alW;EM=aLO39FcRr>_1B=#_
z+FhMghKeXKd6q6vJNY8Ogx^)pbt7?I$iG_C@DNxVB~Rvq-BINCKJ3im+(E9|TE-2c
zYh>*d&*x)qMOk_J52`;@{deB<pWyxf-1HN3^IT;mvGr4FNwQL5L9xCkzv|R4V2Tqv

literal 0
HcmV?d00001

diff --git a/_images/6a87f28526a92715315c4766f197b24f30f433e554ae3059a3d226b457135a00.png b/_images/6a87f28526a92715315c4766f197b24f30f433e554ae3059a3d226b457135a00.png
new file mode 100644
index 0000000000000000000000000000000000000000..1731012bb97220af66a56cc669f9955a3895917b
GIT binary patch
literal 34680
zcma(32Q=6J8$OOJ?Uba93Jp{ek&LV|GD8|9*)vK;A}cAg$jHd9h=!T$l}$)0WUo**
z+2eOTdVjv>ch3L(&+qqhPN%}_^?E)ZkNa`o*L_{r{q#~%lG(Op&lU;_ifwYTQmPab
zl&%yMYhpKTz;`;sQXBEtI%^5JOPlcJxXHj1|KEID_Np}n#nzqV&zi68&P({FppCS;
z%|#0%o4dMJh7>n-ZEl-e*qEE@?YA|wvNpB2b&`vhi<fi1iH*(e)7;$u_s_X3tc<yX
zSsr^)Q0%9WlRA6JE@HU-u3hcy>gt5UjPHfbJKrW;2rs2`c)O5r{(_B)H^=gq8wzsr
zX0#tJ@0K3flkqYm^GfX%+Dkim0-jci2n{T)j9VAfdIaAHpsnh)y`H^iY~GILN#KC}
z@xAFS=I5mTeNC>rX0~w$4f*mPetA&k-#1yQFVGx0NWKC*5;pDn_tPrLOY6Kn$yeV2
zj<wSNy?DoNSNxJo?A3ImbsINc6L$CXl#Nfzw51^LcBQ~^zOa4l&!3vn_T2QNPw=xD
zRr0>}bt}t@$2(W28<nZ{@<0Cjr>VEUfBpT)$m_u!;>V9~laZ<5%=hs0tkk}6SA~YA
z$~rGEk5hYovY#?Wr}#ne&GlR9KI}6)!4S%4x--M9Dc*T`Vcmue0Z~y=x~s;~PUl{p
zdy=4;%{%flL%%WpJmaWZfup^IgapH>8@n=YeAoLN%0q!aZXa*9s=n7awMXFA=Z!3m
zXCx#Z^ym4BEqEH$KBr+}v1sA99ah)KymekrPcJx+A8WCgNY-M0Z>QtbwvLXDcN!Vq
zvz;!{hBd*d&dc{|R)!j1r{DW~qx{L<ium*4gUzXjUds6W`VuS2m~_&jL*&M{gyMz2
ze;9f6)=?-aDLF1pR5-3K&pJ+49o{E;FGwzs{c0Vfur24UTetF7R~9MEe}3>fAI{IC
zUryE5)|S>iBrU}%kj4{UYh>nu9ZL;9B;)%f(?WY@xJA<cpfvBW*^hUO{ngL(iyy2H
zzUj7M+o#6HKwMK`M8xC5LUH{H-?+PD9l`n?j#KA%?%cU)^X4_98S4tBzbat$R&EAI
z3g6|m>XUVz@25AvIp(CamyWJnfjhsnhP&`{nqdutn4{qG!rz*A74$UzBC`4<23F3M
zdG9{C)Y94-?&-OSB1tp*L9tFzXQy}x&&tY*<I4O1<HyVIHU7GqjWwsbu_*-K?yugx
zla`hzOp$?V!>x`yx$aMo)*QJQPo-b}SohPz&Fj{!OQ6_A<L`35{VaKZcQ-dlck$Nd
z=5f|Y6HQG`-lh3(i5IC@#P*4ci~p((ugJOcC-j8w+S?Wp-O)_Kwp7Bx!Y^LFlskWZ
zxnh<Mn|1PEgeoT(R0Rm0K5aGJoJ!T|$0S^CT`xMDIS?gi`C#q(Ehli>Z!W(rPp=nK
zu(6q@taz`L|GYi-?o_Qn=hM*8xk|1$gt^~~(HQ)tTv=B3C`8*itR>y}Vv5dLj!Owe
zGb62=w{0_==qYn?7$xswZqE5uEjj*HdrkyCqQ54@Sdd*rB%-M345CA(_W7wW{+oN8
z7N%o$OWZ%0x9}cgIWHw;#q;lCS3gzyA4<*0U^{W*X;zk?e9)2N*U`@M`S<>=%;X`k
zCd>C)oIp?n1O$xF&Ke_>AMZJ}B`Y(ytJR|LJc56)BVU-})QxXC%_+Jhk}f4_hP{yX
zF`A$9@{}EY*+NzpUz0nYZ2o^Z+fOwGcogO3w;0z&6xY;j#lo6Z{L2J1Raefv^bZ%X
zP|db7jEIb6X^#HmU3Aj8?n~}n3$u#<l1b_t9E<LOf!U%NO|^CR?CsagGH1#9AH1S;
z2y2MLl#gv1YR}D#<6qFuPTG{$x&8xwyzAJN&()^?E>G%RZ?b!ZsUE|>5A_$3l8*V8
zeq@<n{QpT)eK}sbRxkfFC0ox7Uv*wxzMG_#FQ0Byi)8)fwCBH%ob*fFwB+%%v2lLr
zb#%hzw{#yruI3&1+gBC%?s{&pw)6aYV`Jmr-x4Kpek^Tm*-xE%#%oyp^!f8@WT<TG
zf$i*pUPyHQ5fSFu<!ahahGEO!ii_8{xVUKD8)B;tJ{}(C$MNrzvQ?tE);>9RD1=+P
zT<65??oaFU^79k5@=qrvB~kBlnEG`d=}5iMSu}*}`r6U=a?Lq+%x-*tBOT6vQ_x|C
z6G`i(X=D7QMAb)#oZg<EPdJ**1x~^g{QUf(wMHb7;0Ho><6VcW>A1BbEPKnl|NK$#
z+R26;r#x`rz|u^sg`K@UkHd_)QGL{t=x7$}fx0-miJq`U2`{pWvPru7{Nc4fw)=<S
zicLE5gkoc3U2>ITstjHzM~e-&q#rJ^rKhL=nQ5UsJKCP8mduQ8AgRi{CGF9N4}1z+
zBk-=dcgNT{IjLSYc;@Ge4h;=a=ZA!Z&<j{}tcy{KJdKiY@s%u%B0H^|oZPGUcm(XH
z#EY-9>?RH(6}-Z4>=Sp6z#6N3G{3r!iRq+fqdC7mZ&*rt`XL;aQ)WL>Zg)!{!sb^N
z$Btfp^R3X~z`xM#OPo%4engqhm4}DN=Jz+R@$uX4+nMB%`{Wc94jnz}#o|2Y`sIrX
z_S(|Q>fotUyRTilrjc=@Y`nW95r4azF+Mqmv{_R9?ARIx!Q1<BeXNHMUyM1k=FO$V
z2jSs+5#Vkcse8u8{L<5{gGCm@P8yR_h5d7!X)*qmtj)4<qX%bp$iIl}Q{-5?%F4mv
zjTCOVFm3v&q=XT*2I;x9yqvwnGBq{Takl-gW{%Bf85x<JyJJe(R{d#HJW<W5`W58k
zaT{}Ub0-|<ZX=z|oKbl3-`gtTZN<}D6qB{`1@BECdB$;Buju|7+;Ek*T)GlTF<0In
zWMzGLOf5+z<HqGHNt#}-U%$?H`QHxHIb5O%HU6<#Guw*IVRp3Njap2nxBN-=M~Zbv
z)Km6b<V|eAw+*X<jz)+&1S5l`JIvm?)t03@V)rkI^`!0fY*k)1ROZ=FC0uy9fxB=i
zEu+_9uY-fb!t_wd$$F6z<SG04{t)D<IK-D?grL9U{Nz<(Tz;Q~QF<6wa3>ATSt~2v
zqeqWQ$;enQ%~@hMk6zE=o0ypR_EybLugtsQgK6Vo4O=-WsVK#$BWu;(sh_p86D%$+
zwx2v7f3x)?(!IyU*Gj23zH|I|cU98Bfc5Ozvq79!Cu4ipD@O<xyKmc@l9^f2(Ge9C
zv<sp3Xv^+{JUsqUq7JG#Hm1i!06PqJ<1TQ9Q5DKe`1JJk-QK?+q147^3JMGN_Vr0=
zYVOOt)!wr-KNTTjACQuAVrHyUSwrI|WfQk{fyjjm7xtaD*>oSZc#R6;8Fw$4Yddm8
zEvb!C|1XNNb^ij7D>j+$#y8I`yE#&fYWL))-cCN^JP>)8Vtjh~11GNqe>l&cC#4k?
z)_^fo19o<HPoF;3nNLqoKdJ5XclVRM0(BqTkQK8LvMm{A^vJHifE6_J>>AQuHS{K^
zryiu|((I})l;6dv{xtabHPx5rw(a3J_w?}CSXWmU+|yvd>mMBK{^ZFvOH0e8@e*ps
zzu(m&gzs*@zmrXV$BrFmT0h=2ZB990YilcY?wl2(l;iRnDP#=Z<vX%UN=7}UUby$2
z0s;a*KbV%7CFchZ%>sIH**oin-tBY>*?sIvihjj*6g!;}JE8Kbz{5iJldOv~Bk6gJ
ztv`Q~MU9MPLNpcE*HhzNem5oiAt#ucnvw#Dgwf$JCW^RHPc!h^a+b#bwIbpdnTqpz
zPSD_>$-aI25|AGO;l3qZ4?igFtz%(faW(LV37^s2%Jfhp$vU_Zl^0T*H9nfl0a+#>
zC1D4ZgNDeDI}0Ll-T_@gJeYgCyU(t!EL+;xyaJTN!O}rqAIi|Pu<&zpqnw$UX>Muh
z?&+Bf5xqa!kzWG{R}YjlKbkw{>ElyUU+?$o6>D9jaK*9YJPBlElP@u@VrS_5Un~8W
z)VJ;Bf90~eLg3pE-{7~JxwcQZuIB(I$eNfOsTY}goN>s@vr@k~(}IO^<91fW;$(l#
z)K9bIh5{!C5fPDhOtlBu+1*KsLZl5358uXxnl!xl{QQ)ULI{`W%pLDtM<sx(0Fo%0
zo11m?^(AlI;E?fSW)l@<x-b8XW8%j(i-uUqjbFchW#BW}fy_-ovLNG0Ba=ine{mfD
zuAdoZC=?3H_dT3eQ=Dd3m(&JChL>Ed>a^Rr`T2vRT~-)T(3h5%CuU~K+uI|N2zy6I
zZwZZl4m$}jx+}1%^!^%3B;uOSAzZC}LkDjoPndV)g-LpBzkoAfCR1~#E?Qg}=V0(l
zET#3Y&#AWi2sTE^GnyW3*e7E56p$kg`5T4Ja72Ao-O4Jh-%V@p-o1~{9YU#i_4>8`
z2?oy9e1}=C!^%;8NI;Allb?-Fe*E}xZeihTwzaXw2a|~}KR(UP**$ph0B|)Hsq8TW
z&jTEc-rv6$GLJt#7&hd5EHB^bmr^gvQ*m<PioBvCm0GfvAp(Gvl~rA1ajbK8WtKf~
zb70lOVx5#FN6{xpK{$)RpEdkQ#{iXCNM4i`nK!<B$Hc_EPV==x;xs$#o^3m_=g*J#
zy#RrZ9IHK18*m2t`a`shhQ7uDae1r8$zK%?@bnasva#V4y*J&3>!R8x`1F9JyGX{^
z6D}?;o53#!YQ}nphU8RKcHnI2;*?P_@|LD0W54GX7fETiC|o`ldHRk?>&N3r@Sn=c
z*lN<{M(Js3OM;GGR8Q4=6vwu=2rz@ft@!@ix5w`M{+7BVes{HGI}X`%KGRP?n?A2!
zA8tyzUeeyqggS?#*phSSSWO6by~-wgd;1INM&}V_DH$1Mb#>dZ>oq507hA{fX3mU{
z$K3n-JD6^AakilI=gPp?>Wxk(m%E>}M;yM1uKsprlMgB~;j{Yt`!YEv-`}dWAf`RG
zGxgwfTzaSOkFZB(DMAJg6~4=bT+oYF;T4LY^7P%Xi~T=?i?s6ZP0SBOTO&p|8Kt;p
z$zcJ4#menQwMtYzCoN?aN|0l^u`*f`^>m|c&d$z1e5b!<J1z1r&W;WIcz>-e=Z+6b
zi4|Z|Kwuzt<#~VBr$<|9_w7?`&$;tH^h(r~s;a8tj{I=^R+?Vf7S)RvO&eaU#{~}{
zH<Xr^s-Vro#g?M&OjM2Af@&W;k#jpaZV_jH;6h;sVAyjs6SPQ^ct<4AhUO-ZIiN&x
zyf$y%tQf}o2su|H$M=_w{YPOFJ-vg+uYEKC-2dSl^yj#CK}5q#nW?J7(S3<`7;sC^
zLqiz={iRX)5X}Hn9>5PheSM`(+AinOmcD1=Y09?j-5_wQtqWJUZvA>1T3S+5LU|3h
zknYx2q^kH<XTkYPmv&7J)DNI%815|WOiKp*q+&W9bo$P3k%l|mnps6-g{yV=qjcsB
z1yC33!GjNgAbR@yrGYSbr!H7!R94EKwjGv5gx*eGGyy^y`tg1*;;s5ZZ{jLvJ)LVo
zLBYGL>E~7{j41&%{Fp`OuwiIz>MypI$x8K^+Wh%(9EG<8doE<vS4nxCB2NBU57}rG
z?<?}5cggkut<vn=DS_J7^6_Rkdr(yga+BcgE(+B7SFc`;|4A;K#7lJe!?E_=DCxn`
zYDmhYyT%DK{g&YG>{6?X*EHwe<sPVyCU;D7I<6k4HB88wojY&*LB=7m{kUSFf&Maj
z?GLsQA_OL$d$NZI$f^7>qyC5>$c0b1c>AHa5LsE-qNb)0tjE_pdutr?6XCUe_0iEt
zjc>!%_^$zeBboI#DvQS<Vf^qt)Y4VzMY}ZD>(^G`6po4o7$;=evo0|)k<s@p3h?~Y
z0AL2Qwzl@3Q#T&sETR%P_Viz7T?|uXmynd)Ohwh_C%&SCgAjZt)7o3)+PanRlnOuW
zM>3i=ZQI7i$H%8|*b5u81?ci?oPu$SsMSHY;-WZ(5WCUrf$6Q0p0}~s*jxfg2w(@U
z-xCrkiA=6#*S(%gyWq3;?qdn6ao!n+<mgBMk*;~5F0v?p^CRGAR6N%G`|oeqwnyc^
z&dk!<x-mc_`!leQDIh-1v(D77+AH${f}*04frpg?077C9%OYj6IF7q|pQ*ZsH1ZM$
zcy4}v0%_0SMywUusSc;v>q*yhjfMh8dV5PeX$}pdH;!*Vd{|yt*|@LL-@NtX6VyzQ
z6r9m?3yX^<?~ZClL`1xcP?btky?5_4PUk*h+sDXA))PI_SKezpD%R0KC;|c2|D@@6
zSpvNOVAkYYS}L_LGg5;OIHH-AlA$lhE~}uRUlV*BZ9*HbYt7r_f@-iSm;3t9&JL81
z&&<SMczJGmxFryf-q2KS9DDZS#fv$-+trWejA}`m_FW99H@L1#>*QzVe!op-5O)><
zv{e4a)MU<2OGkGOCm9f>E!)}yYc$n<w{zn2$@(uZq_?)Ux68@P_Y}EOqRz7n|I9e1
z=65ITlv!|`13+84;l07;X7!KeEfQ$UNtQOu_3Yc;-QCTp@xc>Oi)tn#FE8)D{Qb>k
zeefDc_ScTaipb+YW2x6%%e++v#EK}tn4ofi^qjz--`*bb@bU9I%*pxKtSQNIdho(%
zTec)N2QWQ@Gg&3`*7b5<rngt`PNnqq8L*#!UP>*#q;s?!$sHJF&H3kiu@;>Lt2c&1
z@nfVrDOJ^-{{H@irebkfYDnwEX()Evco7^J)z((+Lg&cLj(ZByF$UCt$6LAcik=-)
zQ_r<MzMJEc-LN&<bi3uhP4mbGEX_@+`Z73RB`7v#Ep+tsGPvsm?LskN*URyJ^`^+u
z6)CzUb`xdvVaP+2g(p5o3WvnR?0<gBth}ep=cE(QRhlX-#63FMG}A_Q=eh2UNYh9U
zvFJ*{`g9>@fx4&+;f}sXAtGHDAyg6}WSw=pdu>BQLtDN>2vTR{L~q609iPwQV78<g
zZY9m<fdj6c&Wlf`r)_GSSy)(b$_bf-4~f6?=g0Fo8BVGVl*p?G#l)gYJgCzxI>ivk
z?r89sPTS~!3__DEI`bD?K{XB}GCdG!aawzEu`7N}HBprf?VtqM8kU7I&+>Kh9~YNr
zJBQk`Pol}_SsW{Ld8aH6Hp>*e8;S<W%s8JK3*YFuwU)S7x!izSwr@A;EO5%cmAZgU
z0Vq}mbZtgQCD{}QCXX;=q;0Yt`8ffA+@I5QmNi+>c9;Q3=5$lCcEyh$!Pw~W&efF=
z0Kcok&kYV@vv3GnvMdiF5!U{d+!M6{?bYXK@n{qokf2JS)CkwW%*@Pp^XD$4D?qoG
zjKcja_`rQ)j-jNl22BYVhJ5|<tlNgZnh<IrA+#R&4TdM@&_$t5AXyLqR((q$(%qN0
zYTSUO50Lf1JiI>Ry4EkR(ECiPx2VGm`5D^pYIL(K_DcYe^B{Jw=GpPQ*T}dxyeK6t
z-H@Uyi6Acli)dB*Y7<&RWHuCxPY4fYp$=n&7HO*~L>7nBgHVfcdamU>^_H!p1*k(h
z=V72C$l>V4_%0wOe9jOsCwhe+zDKT>8t^u}xqOU(QLvMyZCN~oTe$r8%8hOir9a;3
zetjj^mS-Ob7|L_=Cm(=rNbI%9%ECK6rJKc8m#qKMo#y7U=g+f&CqQ#cuu^Jj@sA&L
zpf%9bko(2jf3WC`8USwgx<>^{{4MEwTk;(Qe0_bXcNt>??C#wQKm*odPIx0<CShKe
zl_eL)=tccs?8;G!fTJgm=KsawcDSrMf&D=W6~O*Ey_+JTjv_@^5%J||S_Bel@s!2p
zA7UBAr}qK(wU}SIOXqJM1peeP<4J(6Tka#M8o$CV3SvQQUXf>SISTp#UCdMowYWOF
z$lFVaJGattR$8ZPUn_}~^dJ=JYvt&8z|Op>I$=`Nn;b(qL8QH7M}t!4LyvO9RKpJh
zfD8rP^%_OM&cQ+7%*+Q{zPvbl7`L-JCx)#Pb6U7PH?C_t*0FqPb?u$SnVZ196*!uv
zO-W2!>A4(V2)sfj{%Fxz2Wl<*-d~R13O{2I7vNt5LPIN&Y=scvC|f1v<?etLCyhQc
zpf5`6N|fp=Ks&&D=l3P&g`vYYe!Qz1Y<M{_Ir#|(65Zb*Wtd2%3E$4Wduxzz!f;Or
zQ*wm(Ouv@?Y_NI{in}4-;Vw#{JyI7VuYsqA4qFqn40{RO7BW1!49D@}jX!*+nn!2X
z6pE)00I#ZN-9A8jSc!CXtM@+L$3fjd#Zc6d?_hcNZUvGivM~e!WkAOxs;@R7XOX@d
z@OuL_^Jio|zT(HSbcCJ*oM6)~a6}&9Q=8U&`gu1B3^tzh5y&g3cFMFBBC)7){5O9}
zA+Muc;5;&5gBwg(PoI8H8m61A{GbpFMiLHQm5`3<V+4(_hou6B97qh^*N?wRi}CQ`
zIxUywI|LB`eVeqTv60>Wu=px2E)^_}^U{P2REyiTw((e!4<-$tCTq&NOFY6v?0L~^
zl{Pf^llVf7Fj0Ic%bsmBv=dmSw4}rhO|+wv(|aa4s|1j0fJ!nF5^Ir44bhL@IIyx2
z{TW(fJrw<5!^`I8=A>x{(TifnOifJ<<nbz6vg$j^Z4?b|9z^EQw?tZi$&a=Xg?~X$
zH=wP<dw{Y4AJjWIC=0v|co%<}ZdZZx@@b2Ld3G7!eZ>a67vdEe>9{mQ-!|#&s^UVE
z#NX*O%Y?|Rz}HVpR*zi3AFL`7a&TuPhRkH{acCAehM}tYU^}};M$V({0<}`!(sK3q
zn}7k4H|$>r)v8V*AC_R>NUcLxoq<c+y=#{gz&hyTfv#e=ch@p~@dhbqCEo73Pw{9g
zJ$Qsugj4+wsh0qV$Qu~Ex;*LGM+e@=yg7x`i#HTO6+I&(wZdb0lISSV*KDuwWlFbw
zwa*NH^O^m4jCQCY<}4+M&oYw2(~YTb{`^o}RwlhZrcyG2O~vXNVh=)u`^%SyP@sxH
zrJ!F!O>r+QEX?TD<qw|+jGiBS>1PK%(f-f-%y&#f+#~I|A)vd!Z~UWbucouF-?A$P
z=sf7ig-<^xq@wl!>ahz7hN3J^E=|>66ZRF6Mz>}1{mpJr+Sqx*YMIGYUm*FAT``p6
zer~Stnl)=klZ#+f7Yg(In<$*nC$yK}TnDXOsA4=-YvIa*8K@Z3xJ$F<ZCU;VSfI<|
zH*eMn+rj?&G=MXpCZdn2ZT0o1mX<IO=wFeipD^ySf2i%U{EY6n#uuE`8jYxftEI>u
zXsWjYfEXcWZhT)@CFuhU1Yvn__3`NZ@7<JIfd<eghh+Od@km2s4=PltIX|H;-yV;O
z3JyL8P87($3)`st-R7H@mlyiK+r8!6I-D18e){z32nap!rKL4BHCP&}cTm?zF;t2W
zJPX?YJ(En7DgdWG!irNXFBDkys<7|(>p3=F=>H~x4aa-^#Cv*r^hRV43i$#OMD-01
zKL!bhbQ~}7XbS?w5)xOzT+i0Gh8oB#fC2`HH|5J~Hg)~_bq<8v_|%j>nx4nZqR$C*
zA$=z*ArSR@rVFeUUOqnh2sUKs^Xp@lj^TK`1o!oLuYfOT5VML@Z-7g&g~3AvN|d;7
zBYgLfBeIAeT1cz*66gcLnHqwkyuzsCf-5nqi;x2VLgJVH*eQN8J3E_4uXNLor0WVR
zOY=l+Bmfnc18Q;#=aYts>1NpTy`xVa<pFvj@f_mi4FJ+TCgv(TLk<kOq3My<rk@x3
zjt+En6`>qu;oM$~S4_2yFsH|v$BxhU_;6Q7i;H7#46vAp{Leodq@px_Z|ATtM^{dA
z0<ab-`8LDNJguXmCSUf?jJDUHdkTGPw2?2o7J%D_{4XUXg=!I495fZFtHO6j4NQ)I
z$jmeblQR9Mi97d@kAY8XrbT0xrM@>kcO6y}duZhaVm-}q-iq+J=+fUa9lk~m>FUaC
zAwnmZTU+U07cSSAIG*slvSUk_)VmkJq2%C$BRFX{u7gIxuJ?&R+ERh8t}cM08v=ab
z&DtHK$BSCi3>AL=cwd78Aq2HXOG^v%uqW7O&Fj~p8<hjI0rG<NsX|Z?8W5~|FTR7d
z+MukgY$on48?OPT1CU=2o%IjjgV~iN9f79@6=DGX^@neu>1f-EuvzQ>@d9L*pn3tF
zfz9md?k>hbVGyx9@meYJ1x^NG4v{r9gx|;}`6nKC1siujP*AX>bT24t5(^Y4SiWAg
zhpkqY?6>SXf+9%CYZHi8`~W0I#EU;rBO-^zX_^L2uFEjhjm<$aa!N|SGMcpX0bj{(
z0J|Br<bzx?ZO=J{V?{XTE7vGbyvEz;LUuxwuE*WIx$<7VDu69fC3fxh?b}iEdwP2%
z5vZqH!$s}@a*}imz7ZdagVo1NpaD)k*p1c1RZ(M9OHR$KrvL!}vfJDK994~#SYH<<
z>I34|BL6QP5w%N7N~%qzqqtf+b`A~<Q~<ZLG&|3ZUPtkff>Hv4c{5=UaOowGrBP}r
zb}%qh0vk7fFr`J|AqS5z+9a*pDpr-CX9(rd-vDAw7l#A~*svuZ@Ir=gFTZLpu9-WG
z7dfJ{8AhIc`gF&^Lx--W>XlK}+~d5O)|bd4dDOlhcrOb@6s;xM%`bZ`@`^$F0xIS7
zOU}G*z`q$q@A2V%@`&2LZXFR75l~)Fp2UH^pxVQCL_s`ZcU4W=d*MO67q~jZ%BKgK
zfBsa>zjr6ccBC5b0FGrFQXf&0fUzqkYbJBY3KVd*J-xl}nF!<{G&UMldHed-1rPYY
zz7{L-h%mD_+WJd#6InSqMhl{(&jH9t06HWT(Da{y4M_929*S8Flfb}0;A~blHiFi-
z?-PtUs1n7o!OfLeBJlBzA8%eWtMJ#X1Rw&pm1@;5PbmE;`(KP^G7B;&>e`MI>uK2)
zV{k`=)rYRGJ_Ex+yZ4aPq_1K$(=;yo+xu%FAOLx9d^>`R*;0f}umgZg%~&qH21YOp
z_>toF{G=gzL%=O>LzggK!`qgY-H>Le^P2+L6d+aFyC>u58P6P&<NE$unT2qZXcc*v
ze}cw|L&QTp9>=8_F8sAV1POBs9j7c>B70<;iP_o1e0<ZcKhF9`m{0aq1xGue?;&?}
zqpQdj;7Rn3XdHsn)zwuo^n@#zm=v@PRO@3r4Wbo~uSM5hiv%NNGq?kp`#8X7-emPL
zR(5tPNTeGVtR3#%(=ejSTkTvGEiWnQ<!;&!K*0ifp0NPcBRI^DH9HidDbl+>#7E-&
z2(zk17n`P2ycVZk5>&}vfm^T88--bo<o$Z4rj@Gr=1@Q*=%ruv(e?D&j%N|ie0TB<
z5^);8CS4aMZ>uX56)_AZfzRa2{jM$@D7N)T25GP8eLY2@6i2peBxwrO%#vuw9TQ>)
zj92&QsFI?hqI)mC0CG2`EC~QhmAuG7S`Kzvt!Zq$n7*Z&Ha#`98(60YY?}d&I8-b3
z9QGC3Op8t<6w8P1?mL;8m0%GdbsKN~E#bnUtpNkuMp`og(ma3*ua<A~>Ei>b)Qj3;
z`1kjBWUEJLrf3)#5;i|!t@%6D=mGjN3Mhl9pYp++a$wJ(*{0ck*ogg@Soo7{4c>Px
zCFM|SX0r1+sdt<T^gfyB&&ta;pj$yBE(o^CJ?Xk}eN-F@Fhvl@<t>K~A4XTD-@<}=
zL}UjB|Ifd=y3X9beL_r3jNoU8{r#!+r%-3wvIDVp$h-T%SP{z(u*oTCWzd1Rw+0S4
zLoTuzZ3{!2OSm~WAgYQCcu6CP0$dNX5rJxf?C16HlIQE>)1;B=On-hjigQyhw)hzC
zjs{TMkI%i#cxeds(|#;}mPolo;RgZ3K_P8*AAyw#9tm<71$5xYEaFiQ)^8>99$2kM
z4JHPt(_|$;pAw*QZ)SLNTln!5^p2mSM8gp;x6!$YEl*#BkLhu8vV~FkuBsSV2Z*PE
zSX*EcC>(ZPGDh2iJmZP-MoEAzzeUGq6eP3`aUET}a;WssL^!nz+J0V;;)IV!0@W#0
z%;_{rm=W%Yu<L4C_G~3tcm3ZPn3>fe?UL4xh?0e?%iI9%x=1xYe2pyOcBo;!pPWDQ
zoL!E#>MZZy_8TaGgx|wiYDm0z7-yTX1EZs(S-H7W5LD9cj%h6|Eiv$$GoS%5a&KwG
zmdK#vFzqZ5#epuXs`7$wz{p6---AOl>jdgsH~M&&TnY>Dcj#QpAljj#On~9D!a+X#
z`y7WK+-LTLbHH1@XOdobzcyT;5%|-0pU@{@L!z-lY2a8X<v<rhv=KBv5N=e^c0g=`
zBKq-0(zkDy2vh)s5yGAwI`>!yne<96>p@1t-+B{5u5>;jF8VkZB!~b~#cYZeh+hir
zNFv-Nh-)#p(>_X6=NTet!DD3w{o9-F_>B?40GUW3Yj>Q<G#JN7JA(XXk39IENrBZ3
zu<6VDYXa_6jCx^V3*f!_(O?S-3BgU{K|FFsf&kZ&?o<cB_Jf+)0UZReGHp8Q;AD)<
zcd3Oz)jnnLs0!%Z&=%S0bb?(FbeuZ@pl<+*go7eo4!5+1Pa#3Q7CH*4=K%IGSU_N;
zv+HB}WVh|%i$Uh*R8P5&O)ul!YFi&AN_a0+4MK@JE{<dp>jo0yvf<m<7V}@xA7}|l
zhZ;2j@D~siR0@*+YW}^9Eoc1qakSF;(A)1p?rA`Vc;My+g)#<GCooD1%1FWDs4%#t
zPmm$fmI5!$i-Q^_px3-T=NYVh1Sjx-co8BkI@qeu5kj<}`pjf@&R-+&0hdPb8B)kN
znxl)kwl~QhBen;Byl<r(SQ-#tjYXSRT2^K`-gOqPBlu?aqwo3s{VgO<Utsro@l}z_
zZ!Q%#G|;#epFtMJY3RjmBIx(theE~!^WOQVtN|rxf+|;KL3)>Z(!2xxSu%sZYZCl0
zggm`C1N){57MIU(29=?;%SZ|`*o(IV2}USMkl#)91tO==f(TmwqJo?u;Xd((FaTi2
zaZY!DR7D4hb5wO=^e4DsWOzU)I2Dc~E5JNsg{|8dypWTu5-Wj<{UyUp?TJ5q)dA+q
zw25pgeJ;(dt%P^)upeN;J+1+f$^yOegH>AITND(sm4wX8@$fdtK$8!40H-@=%NhGO
zkQv?K!4XNxw;B2h_U&}ZM}mP-q47RY5MJWjU<k)tt!s@=qMhkO2?U6SVt@Pg?V|7B
zpMe$A3OwW`VhDri(j#bl7v5Yt3LhA;0D%>o1ZVnoQ;>N&wt5Ramm;#|J0@vXg=eT;
zxXG0f9=Mu<j;Z-t^;a8S<e;gNUuZrjWhH_n4t=N=Bp47T)a5El*ax7a^v8v0ed^GP
z4)KPW?ZsVw8hZ|OOHa^OoP;S347u5eV73HWLz0`|JcSHQ#Cdgf2G|AoZncFXb%;-k
zK;d?!+?ZuQ#g00T``OMcQu+Q`<_QKW3ZPaZKO>_433fO~aC?$h-oH-)D{9)*h1FDW
zBD#EghC0!gzl2-I;54jzla~l54kcIz#66L(AmyoK-PQ%U1tlG9*M=;CTm37ulP+Lb
z!5a|QC-QizSrZq+f>XM6bqf{ME0{9i2n;~$*^T`V45UQ~BAE5dmoHQmuv)05>Pe$k
zQDi|#0{UB9gv<qClm>1aZ5`2qVS9T40yJ#laoFFx7QA7^6h^e=rX~T}0Gz!(P8fLm
z5_v3pq#PX`L&aUhkQ_73xi7p}2Q?6ugB>XAz>g+CoIu@Id^PGw8;H&m4lYPC<W9hi
zwwRY2@Lv7VmzI%{;n?0eTR58B{kKrtrNfvvOweh8555^#N&>*1U9IER$`i!GkZ=K=
zK~O!KrHS!zvOzdWQAD{lX*dW{V7h2|Bal404)`B%Eue9pz|~~vf*nQ9=*J7tfQFDG
z10E2to<YcpH92o`3m7nx!~ELYzZ&pbg53qhk%bZ^zPf~i(bf&#Q8WMEGrUmc?Um!?
z<EN)zp+LiIFgcuFA0Y1HjK&_;s4mH5ClCgNQHJm0uu?=9vLvwUK&llMPvqtYsfV5-
zcbu31s)02itr($h(Gd|x1Aln7<2)|{8uV7<D>+&uJC@Up*RS|Id18pp7)gi}^4pVr
zatP)$(T#24G+%0Jd{Cq=!UG37e5N`5dN~*iFs!JP$F(80E~o{@5nc;m4~cahxcBvh
z*>hT)0i8^)5*Z7hh_zCIm^3^xa#B&Bx(lWQv@=IFJ{$q2Lwp0)LWjfq2uV&1gNZ=p
zl=h)3L(YRl2F{zaYzDO+yPLjG8xaO`bo|8qgNW&Sm(IU};nIpJz2!4}sMVk#kD>@G
z1ar>UOx(PB01_-(mYwwUxS8g7EGV%CAZvF2`bEHzQF@E2XB)TO=zu_&VyGf}&4)kK
ze}37mMtC^2ca9AWg&nHsybdi2XD`ieLJuMYRxmA2aO&^hvWkk0z<9Vlcu)`U@X!Fw
z6Z0~WU%^np!P}<&ak>NIoew?t&fUAqKqC>KW`!S1MSXq!<j<A)8hCXfu3Y_b?%1`D
zoM^yWmfpbqc*V=lk7&#;LR3AbeOL`Nbj&$YKHg_&s?w2hD!{DE;d8+`UJpKh6TDJ6
zSafwP9$--qdU@<M{0GEVjij>Y^&Ge9yVr<~M36K|+J(N!$=tAiWcMsXouFZ4^uj4d
z3ZWi$!CJB%lwxgyg_8ycqaXA)I!^zvX#22QBLoX0Fc6L|QKW;bRNts2*WtrVzrNZG
zQ+ReyZTZjXJ>H&`fjJ;Qaa6N^OJ0TJ16dVv+uf)^$8EMd{540i2cq=bvMhl(NE$=f
zxW9e-r!XhJ%0O>!yc;CP_WQ`;jp&`QlGDxlBj}h|9T=$GHRo4`fs^muTcC}68;rI1
z3lJg!5*7867zI~<YXJMlLuDjP3A5<EaTL5A5dJ~BS@rLIjFr9h^qL06>Tst>AX&@5
z09S*pncS#d`iRDTa1eMU1x7{HU9-q9gL+wnwxAX$I{0O94Dk)y??Yx{t+xYAK7won
z$khewz%SHdKc^(X43rWya6x$GQ%Jx4)yli!2W6dY48>=igw6|Xiw@+$ul(5#$`~L6
z@Mfdj^faA7d&B|>&`N+T^8O_JScb;t_(*@nI7Ij4F4%aLqQ%aGmM8QHT$Sizh?^gd
zw=*yulAH=7FOg<o8exHXK}1xPSoz2DrhGG4&SDAsw8_+jy)QN!Z3ecr0UcV0UGKK%
z0v1Y7_MAeavJdoGFrnLO0PCK3czEol^}(WP{ZIEV1u}&vT7UwdriXO#PAh*wyZ*Lw
zG+E0=Y*s0bWhk8E*pE1H(gaf?m4K8%A##adb~}cw@*@*hO?v5@vwvV&y`F39^iF>(
zY8g_PJxqbb0Rxfg;lqcdxm8?tBF5@IduTw$zdnaCendj6lP4BzR_-ZKl_(yt+ebj(
z2Qz#I?KtmW=Ksxqu+|U$%=n|9`T42>TD&lLfDq)+$;29ylM&_rzO0@S(hbm0mp3d_
zz(OltTl+Yf8NruODUd@Ql+J1ZRFD`iuwmccT@3`<HU263ZRh_cP$&^K4?#<vH2cAc
zH7Cpx!h;!NaSwho%$6nN|IDROfWzq~DJtbXDg|H*+r_b$2xIc~|K=b6|1jo)OSbpK
zPW*p)yXI@c_y3uAWo?0iPl0j%{BiNlSVtsu0<j;d$F961fo(*rA~%7w;}FdRqJeQR
zkZ1qDZwM0&1a4skO4Dwli`sQmm9=rp85M>9tnhsY$nFjYXCtBiBFyX6j2qA52Cv~0
z!6(;$TkNY0J_Dcwk5M9!2^k4Mdj$WdhH&tIR~P&nPH$n`;r(1(zDNkf5&=>XTP&2e
zX$f$;CC~mO8F%?lc<~VEu%isDG8i%Ww}A(*#>4MRrVG$oAecCap9k&gN%C^x!OTwQ
znwYb0M_~7*z#t46(}5RsEe=|ZCAv{C&c=cQr%#{0mj2&dkwHArG%v6=OvuF`JN$>5
z>Mf;p0Ked3z^AmN)ncFpPzN;TJe718{sk#C>6G68UhW014^*G0pC2_^yD;bFeQN7K
z6wzm?%`J!FtQmGXArJM6ug)vG--N4RdbC~GO!9v(rU4NIlRmgeg<wtbirubgPmA*D
zip3RVWmAT8>l+)VT=g#e3+X=7k8r|*ARYvy`0tmYwF5?2{rb!Z!@%0MV-%}%z0@36
z-kJ2<{BNaRqcEaevnxf+ojIg#RAni>{|wAl>K{yK043v)S{xUKUW2^wM%xda|9{)l
zQt09m+<8cb&K1<>LAV_tAz=VP5nTnrZ67}hSxImGFCt%`^W|K7^X5%rJ;&ilv;A+j
z!@fF9F<!GN#(+OOdakYQul&Dv_FVFU2?Fa1YGL{9r=y3#Ekh{j8XjJf8W=#-{P&)p
zRo<!dfbA<oQb6C6X&a%~J3Oq2Y8f@C`EOVSo@KL^pFg2{J;}_>%y`+y0G9vXJ^%lV
zj;YHbi=vhJ;fuZ%C+V^G8Gmyl-Y^2+$PgDH+Q3gh9W|?fYa9nwg0{rdAQerxs;VlH
zXhEbAEobECDJ{ndX;6|fva*}tBqi=VAhc`3N#xf0tdN~n;y{<DTlSs<hRO!S$K}VP
z8(D(-MKB<&Ab|k^dO`(o?kGlyAXFA~VPHnphPRw_^MEg@Amz{j7?d6UcQtK_($H9B
z<>hhh4>r@X6CFquvKjjMYNSpf<aQLnw0+k;m|Q?-DFY@N+EoW-)1K2dM{qI55FEf4
zKLoQ(%fKLu--XGbYI>MZJ^=S9Hbg$dMUp9olgnacjRH<$;LKhT;HMiX03Goo<O6lO
zy(3^1(fL*Y+dwY_|ET~lhh78R?RswRWEnbZSUiaLnYe;sUJYDBxj_YH7O-d&Q&Wpg
zE&6(TM9bMo%?u&;5Dp>{E5NXjfjXq7%76zKNu*mb?{*C^1~9Q5Za$0t31=qa^c`;~
z@q}}N%1&ICm}CRia07AU4#l`8n{WqU{xBjDO&lv9bF;Ew+IWDm5DGXpR*wV%Gdau*
z>q3_K94_F8BmWiCFNDB?;jy;BX%TT&VZaNm6ZrhY2Zj~w8DMk6yQ?SB^}xgT(8B|Q
zGBa5!qy`j**VwJR<(XD^tkTg#h|P6TiY<*xqFlvDka<Q{Y86Oh@={Xkz?QM1kwMQz
zNYFLJS&dK@yz_e(1qeGJP>s4{fVO}g<Q5>GJ^&SGboCr?Cjmmp)?rWzfZGg_Nc^0f
zL4pp!8AfDdfJq9Z0A3J@&-rh9=P$1=Hy45(ju3Nt4xCNKLJF2=uHhDVm&3^7vJT*?
zvIAPoH;nGS3tI&WEb!kyhuQj(Fykm)u&H?qzu$r*yXLzdNzE9%F;cfGCOuBu3_gvu
ztKJT$ivemjk(Tk#FW~L>@R{647^9`7AY~JEN5{-e0UZM<E;dNOpqx|@p}5+A7OqQ)
zUnPBn0U^Q3laV+X9i9Eq@^ry-QQxHD3sc-Dc$)%u@4PZ^47>-?n-$%}{{i^Fe|rP#
z0Ma1v`~w0W0PPcZFj6Q+UA&OTcp#)<HJ<k;$C9HdQd#b}n%SXAju5IE?VdgISpIiR
zGR)*V_EW#e5R(tx@gpiR6xi4L4|E9$2}j6q6|@3xOcQKyD^5#>GvzUpFu$7eL-=jV
zA@n6c@dn!VbLxOI2$#1e<*@}M!ivU5TJSIM+PL>dMn;1BuYm6XT@&UcW<r`irl-F|
zXYlaRqpKQ+L917z;!ODQl5$Y_QuGbGszQuGE)IH%VGSJxD8c&<`U6*zwIYv}wIJDe
zd3)=PcNIgSA!7x&G{~rAW-HsdEgZ{$Loin{;566mAu`)$jVT~ve8B-jcex)J30w`K
zo_u_^q9Y;10!WkY!NE;uz@_5`peR88s)O;!{#Tgs5i~9v+|~ebqGv^##!rc0K!jb$
z4q&GF(IZy(sSyY&wIZK%bF(IrCES2eWJQj%u{}a|g=NMEgPA%fKr<1=jI<eWFcZHp
z#9gD$p=U8bst!^XEn*f9?BJkhU|^&8b4@lJ7qA|ru!HTP6N_BBhSa48!VRq)F{R>+
zKZ`vx3oj-xDZU3<=w0X~TuI;1WI}L}=CyTnM^x_PXsKej3Lb+%2t^>%4ba~Q--J0D
zTnm}gA$V2S>KUT=1ggJ>mKg+iz|JNA**we#K!5KBU^WB-0<+!UDeh^Ih}j8Zfdn0J
zT;bMI7`6abERss;wd`CWa%mihBU{<X2FEb&iX;bQM9t1+?!(W)YlDF&#{RyM%+7Mi
zcD!L#CAfM*Mb<3b`OjfMTr{?$ZB<Z98N?j33?~Cuv6NgSw-DnEMnL7j`lnA$ux@2|
zQcmR(g||g!BMmWhY6=zTE<_z8B$ms{6bmUPPZCI87t)A2zLr=gvGBz}F%ah;W$5=G
zMP4$5AWZZI7`;V5m^f&H`pR2b`16m6gt|cv6k5g#AMOHKh?ASQY|#VNL|o0v;>$q@
zXTu*aWw$>?@PLFS_B8a7{c{t&!BvK&USb8t$H$9cZNrf#l?%59D&eD*6=}V2*I+F`
zVGzd^W(n@YWdt_ilxt`uZDqQ27aGaI#@Ey^t?M9pgM(BT)`%ih28jdAY9JeK<?Uy=
z1)cx|B3ivopp{cNe`<l1K#DE{aX@@~&|1Fjm1c$-LRfY5>I1M{qpoa$530%_jr074
z*Gf#mT-UwB!^1Or<+<*_GfcJu){=iem&Vu#G3bFxA}ZKR8S1rb*M2)B-KUJu!iMPM
z;)vl-d}YoJqI_4!iTk*BIN55DQ&KZCU$?F5kJa@$zM0+VTUaO5d58aG`H6Dg4-96u
z2aX2i#J)0*mA&3gmsvLA&Hp_)!6vz%eV11nzoBXG!Gk<fS;2I*Ivd<tj0>igqo(GU
zCnD#ytS<)fbn12}``L%iM~%%|cNi}^)-5`S4OsvFehAJ}>{#H_r$wilG!In-?HEwe
z4J<-|jxpVHxBU)17Qh%KAL`xy=I-l1M3|{v78;`~KYum=Vkk5_6EYPSq&Ie%mD>IP
zcmX={i8Tex*V@|{FypnKG+-2wqBw!d6Ls(R(HxthPr#-9QfGtS08DAw_wUBk1Z6V#
zQ9tlYAjKY(bQh&@^78UF!$T?-?T7KJm4&84-rHSg5?@O_y3<!l1KFbvegSnf5}huq
zqQHs7{s`U}@8GV>K)B7vkEs+JOZ%;K(_1PlD@iv~-aA2|li5mUMFICQ$}h3xy(MxT
zB03Ogr4fX+grwxv9Gjz{K8v?cr2YM>z)h4ilmZzr?%cogCa0!`z!tzGU<A)}p?zU&
zf*m1zP|AxC@|9?h-I#o{sKD6pVmLpT^Qr_C35F9o6yO%o!i=i@iN<viFoChxT7ZAT
z^Tc{Ezi?Sy383=cQqnsy?&q@P=Cahw;@Sj$oe={b5Y$<qehyq#I(v2v4gujKfvb>`
z<iI&LUBAmT+u^{C)<-{PZ{iM#fHM?CJOd{&)bMg6#wgD{J(MQ`=J5%PUx9#<Jm70|
zJPvUP)}&>P6)eup7OvJ3bM$a(#VfOfxe}6O+^JdWaAZp<sYQxVwJc9Z4lLUzH`U-l
z1V!l9;Jsi+;5dxshu|m?Dcpx4dq=SVG*YpZIeoInct_;7BhMOgoy1t6J{iG8+cz-q
z(sU1&mjFNE(JVKV^__Hd?D=P<9<j;k_@h`(7H?!p!Nu<MV^2r`t}8~O;GuHEAV5*f
zo}8rKsreIB<jDwn3>yHg2XpGcsiAjt_OeoJG=S@xIFXR^Q!(QRCx+XHOA<>>@SPHS
z0~vTn-g>8=DxtKYbnHct9yXEq^YHr<m|2Hp(Il=JRX?DQ_fZVze*%S!IkWXVQG8EF
z3PvgcD#)Yz%=-Q2t-R&Yhnu3H{WV~woGca=iQ6P5OB*hQ5+owx<$!W-go@y5>aP0*
zOv|vXFn{>A+;-pgTdzT$Ki(%a`J(OPK`HwKxO1YZL3$*<PK@=b9#GogBQS#NCPL(+
zM~|9EY(L&=uLgH;udW`01B@XxWf<w9Og@F;42aq@Vb=`DIXEQD;6*|be~ufis`5E<
zmjEXs`T;yuU^90N-!5hA>s(zCK)=@GD?HkTDI%`#t(&o=*r1aT7UNS=o<hpH4Ma!*
z8OGDs_g8%Qtsa<p;{k07|5U%+(SP>J6*>gmp<u&QfQqTn8T(j}Q?=kb=0R1X3weQr
zI_D^sr^Sr}ytL$qBMvi3KcuMFU{QDzBC<H;cG|fUY7L3vE5I}Hzrhzk_v#3D6563O
z-I5K+s70VPoq4V|-64Y)VAog}8kCMcGQ6rD{HWGQAFK`$;b3|JEsOUzBwir!Qzss%
z!5xE&EK2^sOe3(}<K^R!OP<yj<t<xpW)ejM<CJ3A9+x6AH!|DZC*1+&fP9>(qg9Dz
ze7wat+8-Vm9DRjvog`p~zCOaH5qBhnb?({S-c~o^C?EjXak@d~1IUHu?x1mq#l$+|
z%OdVuYxEL>>Qu}kq41lm7v?)4|Af`m_y{Mb0oGaZIUh0XrQvOTzMhKZG%zwsyVZW0
zg19W;LRe!50h1B4AQYrqL4SKB&aTUKqX~TcTFLYwHGe&nnH3vZ#l=|=P9@z5JJc?f
z#f_nkiY!kzLL`qUTpqCynXEiiyj?EH8a2Ajd2xieCE+FvnqPXR3cYR;gA;)OsAvY~
zu`QG@51qpVG;$45*0PgTf`o}MK_(T5l>hwHvF7fGXQt1I(TE^CNc2IfWS|(LG<mLq
zIsdg3$z44UWOv&AQ<h<l6C}n+tbyUa?ngCsb-(I_$NC7l(Uk!pK$KSC9cj|DoHX{O
zdU&9eF-Vn6Ph#^YT~=pZh-ik)t>eK~eE^vUP+$&whS=grzhyq4s;*uMTEDr_MI4Bn
zkdYWB2m(F}3TJpkVb|v`GuIcWOrBu?F0pv~Km;N%42%TES`R?PT(ck0lJGGAp0!3L
zC#Y<$SMZ&`mDRlg(`ZLTW2C+?A!B^ePe$9%7LQP+WA^jFMu6M+4A*fY_%KEpdtEYC
z09;;->OvS)@{xr0SigQfk;uUHgZ1A%eB#lOySyh)4t;-1kJ~M|l{wXBU2h5Qe!Vc~
zh3&9dhk_03#SuomgGxkHe39udlEvFc-UX?WrW{AT5i??Roa+1CG(D}lvD27JcZb-_
zqx<Qhj>lAPXnYlPD>BixI$!V7oNawrESkEJkLP^gyLa!39AKjfJAH@axW<P~dZjN8
zs>)CNZPHH1oXZM+yJu>iBH`le?cgt5wS4F2iJe1k=UgFHjNl=)>vwj;jm?Ux$Iij=
zrmb^d9{douV1z&*mtgXKZ88h|g(#`S@iX+|XsJbIeZ4W9p-X?XSBW|Noo03bI=YRx
z+YN-nMj4Y&EOPpS{@!4$BVRG{^fL%NMGn_8Zb&2JWG5?&qeo_T80G;O9mJ|(X<xvq
zi(aVPHn-u<^--o>hQATNgb*f7KT;+yjvq&KeFaXHApB2w7yvI)^$UQ0WHt9=Nm&P*
zyS=vP!EBf}*W*J0&-nwSmNa5UM;CNv@k(3y&Gwwj0Y&M=zD(v3&^J<KX>#GhD=;!!
zeehdnZ5ro>?fWW8ddC!@<_K-iTGBov42g3X%68<`!e{?OGMLqfL$}DAr6?2}5>kN{
z0272Xw=%<ptbY-yS5@_EvNJozoRK?WXnmr~fT8lmfb7F2Xxj;#27Y0xNt0K!&em*m
znr#>P`|YAtoQkFmre(>6qu)3IW44Zm^#u?vk70p$guc1QDtm6RbF~iI!m;a8Wzx_5
z{fqB3T_*;4m|gaOd!GN!eJoCRP{_d0Z~~bs0KWi8+zjynVp%MTuHlZ6v2>&|X2<cf
zQFc9>&-pPk4c~s4)t3*ZfY4Ytn_m%m+D@Gip~#5IJS;5~#IJ|GH4ce_xGovo&q6~_
zfb-rqm+5t*oVvRD)#L6(olnewBhiYdLPo=Ju*_`-T@dj6xjZluu>ztcJlk}K@4RHw
z9h7cj`v9R6GJHn2EJ&4L9t>^#Z)y`es|V8$ag<@5*KeSq?a{LX)?^U4wGTq)2c%Mj
z!@s9&&>T9Kf=z~1U<j-fq^q*-+qY|uXKH>t)Bw-ub9MkK(gVZX;f|k1rqc(qx)w^_
zy2S<Rx{U8U(BZ|=4Od7*`p<clX%X^vheXf=(0;dahS+hT^B{m+EaUos07T~d0?E7w
zw2LnwMiI+|#=7>=e7-hq6%ck3!<*Wl$EE#|8F|f{PXMP;&B-VhjC^b*7SEidDJ<Vf
zouc~+VpWST5odG^($kn}%k3L=qYrV<%M&4uE-E)#Vax0>`RmUCdO!&i5Q7a^ua((4
zLRw6auoob`?>~2h%Mj~9yn<M~`%}#QngK<fLl6(UNixPDBDb0VpRs|#7g&Z#g#}XB
zMK;`=j7;KPEGwAqo`AAe^oN{Pka5GUnSL;alIx_UrKOz9H3E_N{PSl;R4%uQUR~tj
zU37FCl4Nn1o+TG7$SY-WrvY7C1E!MPnPu6F2a4!q7iJ1GU^?*=BvMI`8|rvY(WBF6
z=?(SaizDF<gXRa#nRt$4NV~Z0g4GoVLaG7;ZJEoBkip0nCjNji*k)h&Or^Q3I+J_^
zooRk5dQ}mitA26rw6W4Fu9Mi_H@FcFHnx815FY0xyWRBkuf;IOrTTdIlgLc7{u>md
z8z3r*w@Z80)B!rZfy(}kR#_yLvqlwM(*@KIn1(etcz7o6hTMhjeHnv}r_hY_Pt9{l
zl)h-jLt8E*+kHpx@qkG_TL3N>@*&K2U2C4fP6@<i*oV&GclEL4O9E}T-FVNRM0D;I
zi&l1m-1QrF5mg9=DZjtOx_SHfTqK`|@^S;<l-{O*gUbava>Sd9uYFJ3Pxg(i3};)t
zQ0PsX<(&a;k3>@DHK^P{05`f+K%i|JA{Z*`22>)mL@3ZlOe{B9W<#6XMd~?ND`Pa~
zOPGKoZ9F)FdFty`Z}IpP<kx;e>tT~|(xuOt<ZU8GxKzFM)M9h{fdjVN3ifwEFA+rJ
zjFVp0Z}lu|@^~o=;T(`w-GIm7XWRX;^=}I@016#($8{B2%z}x=Df2?A8c%jv_928X
z387|vzpDE#J@hNfV=j$&^aKZR?Qg>x|9f=El;pW(q+LNDdIFE3iYX&p8U}9@VCnLb
zk=S%V-mlgO{`Zb90x3b|5pEuXHGdxDSZ&9gBRTfaeyqM{ELjoG6X)bh%-Q1Y{Q+oe
zT~-%6p+P%fK9+G6{wp4Ah_07HGYGIjBnf;xs4WHh3hOU?Pe_*n888apAqF9`(73J=
z&@y$X%5F?AjsQ&veZmb5HK+DUAJrb+EaE{b3nUm|YPukc%qZst(_eLwHBb@B<%iX4
zZ?e>la-5U{x~w2mnhhM2-baAj$x}@rJpWB&@h<o<?<j_$YvzK5-^UP;@IkU6b||4s
zk>mpKlx=mqt}2alG=EkLwwfA1XJ_k&h7+x%|2p@W;VhZ^g?Izs*k)VV0rI2=T&HSJ
zwhaX!WO9>&uI>Rqa{v!(kY&7HLoWLBorONyl;puvneDcFq2O?2&dL_f{|YDh1A(>@
zyRS=A$lqq8sK>Qi*^daLgQNpDb_s9fEJ!;-^<l64*Qt0ad?SzF!Ts3-9HSN427Axx
zKa1Xb&o}Z>77q0Hr{c&WJrei^^6L_kNa^?Rp7KBDtGVsB?oAJ#m6ffO)~ZgM?e<_%
zRZ-b5%DO@kURw?uBIb9Qk<$nc1=1Bc(S9rQA|BVE$wja9IuK+bdFV{ipTQhoT?SW{
zM!rNmXiuQsh87e-|E)_mEiIukd!X);@#S(q7Vgbm>vO=(!#JaZ03jF>{9N}7@pO%&
zFnAs#zC=bw#{X$9x|yvFeB1mgO0ct))zzN{rn;x*$M5C4@^c6X%;pW?*n2b%<#{Sh
zX+Mu&@9aD>hSo=HzK<4xc^ug(dvP?E5zkY?gfmAVLwyrIz~t+zD(RXveezvG0Oem{
zVkc93=+@M>&_qRpm;4HFos|7+51@|qg@v_S4^)?fugS(W6KWrgTIq=GKa(mNCUikq
zO<f#tblM>cc@R!3!L@MbZe+MS@^d#95MlCwshW+Oo0~jE1}<;6))8V~*tKidQ;>AF
zw*167jwzhbcrU}Ng?9Uw!${YQ^hPYBs5}7;mfVtL+zm<nHTqvN-VPcdvrS7b`x_bq
z@`S#kXbL`yj=XQxSLFSntIWOl<#jj;PexqN*3oO5tDt!bZlgFvMYB0SnqqfdkTR()
z=HO&W8QQm9#!IT^R%XcDW=!loes?{J2n8t3sdx+>a`pNm<hFVs;2LQMdt~FJ{22|H
z>h?AtUXf)qobmL8GQtDmU}uT-VOCbmQnIXDQNlwV(9szpDP<wNW=Cr=w@JC%)9`8@
z8h?VYu}_2{oT?Xh>DVS`df1hMT6B5`^cVxY#xPBvTm#@g&7M8gu;+O_ew;TrwM#-%
z6NUWOcQx(x`|<FE@#8muWXN-1Kq!EZV)iI;Z5*QFi@``ER>mFOvGsqpuiLrpCeN?D
zsHq9_UR^;d#*=ZVz^ms$(FVGTrwW<tA;Wh-AjbnOYRb5e#-(s>jjs1d`r`=9MI4zp
zxBnjg7Z@lBP9Ab3tQZ<Qi)$*O-)A8<vVkJWV=1^ojx`^+ZcDlk%&!6A6BRdZvXVSM
zf*`yf$=b^3pEhmXst-Dk*uThwfDqnR{&VYUU_>F38yNK$X#FwyGOp$@67;MmDF!tg
zGwUnV&o0FKd8+5wBqH@UCM(U(JO2LIdg~%8@I@gZq2C@X&MH_FRS>&d@{G|hppGg1
z`!0TZ1G?NBP!<ECeYZ_T4)(ab$8$c)%E}OFSDtBv6|!@4E4SbK^Pbf28<@PIcXF2S
zOw+v@{QUHtH#eG+B5~A1v1|VF3JSS{Q`hmYH`;{-q=sTl0~>T~D|y_Jx1Zmtq@<Bg
zS}K&1JKry>C4w+{12YadFs}Rc0~e7^Zon|h!NsMNVbUPso9|-|98Uf#igud*QXP`}
zcj%Cp@dQrt=4j9_;$UtUiSNO8<91E-cLQ@@M5($2Apq=f<dz$pI58JVJzD?FGASZK
zF|Y_e5)Bv}WDSjsHfV*5yA=G!$?O_~>YE7OcjHRk=g*hW=cvNXP|ICigZ}}#usH+q
zX#AFoY5H>+^4dk%5V<mV>)Ay`uWw`IebpxVQe~5+slxg51b`6aa03t<-R=W`us3E&
z+2^)l*0(v<)0$XYx=N=A)&P>DvX$)-07;2#a3vAt{mgOAEEV$S2};B%iw;e)3+S07
zMCZEP(wJ&0fo(5qYDWA-u4h?Tf~o2q);<As=@O9BcQimjsw>>AQ<IZ@_P&LB7~6k?
z-k$6=+vKw}jD^1-m<o<-?yLx2C#5bjP6@zIL5Ng5jEOw1?mkKh8Dd5j=vv12e*snX
z_4UOBku_#>XYlp(+z-Z)lxx8IoYFi}QSpun4m8fiM(LKomps<L6p;bBs}s(HdLzsO
zXqCy#!sRD86u}E34|nXoh$#m0d<&E*U0BV?cnq1+Yt~&pfFM7J8VpW$qwp@v?UWm+
z_Cs63NdF0pdlR;nXd2OCPJPU2R(KYl2LS~T)-R&F(R-fsM-yB*VOSs#%g@&HLNpU3
z51jybufcDw<ep2=#^4ux4=<3z2xX*+P;@mdb947fCpIXq5Fen1$uE-B6M*JHEiwFM
z-Wz}a?#N%1j;yzDrzO<FB#Pn6?`cweA=pHNu<H7I4-E2QVc~ExR1Ee5pvMZkT8=aV
zW0a(q!#ZlF)1}8V+dd)tT#&Z}WST@KBtcr_fN#kFu)`%L<uz|)Xz2g!+2x)LWCqZX
zDsN`jt;N_|G>T0l^lViYUq_@utQPh4CVS9Rl+?R-OL%?_0xkoHFCnP{<~Cf0S&HPs
zb;tySm`1;71m&^WoVVz2oF008JYDJ`+zr+^hL!-o*8R`ac<)V{2pu>si7aUVvH)Re
z_50gV(j364E)A0-a@{~u?kL<PYsNbl8EI$^P&Te_K+&n_T$vF_Dp<HdxV(kgF~6Gr
zgF-@KK(Y-eBxGO{5N*Ov<i2=SxS-|1cvCzcGCC-0{9s)tfCNdppx1OT`#|u==U?qm
z0QPO)2!laZjFo&p4Gl^hd3tH#>HImPuO~gOJbC=s0RKm-eGX4d10HP)TXbbr)zOKT
z9#$?+PR<jdXQ6C+%^DaR!>KgmGmnQD2es5ig1o+dx4n!Ao2dRsw8x%V5$k4@!>AA_
z4s<_ja3x2cSv}g(<5E^$PS`a(hOsVsb;+1KzXR69l#<gHNK}NlLxz@HYv)CtmWMQ~
zM>r5%5n(NHdqid-VipQqBD$88l00+lVqs}zcCev;SXdPZd=)*DRXhzBb-fy;lw0&^
zbu0`RT@!YpC>Y4x#G45d!^XP65nExc<k@%8k$k2BtoP~Bn63d6jWjNaW6ucV2<8J^
zhsxZthH~mK=YQnLkr+H<0KTQg2T{kIyNcG3XIP@9YVYxx2(Np)d)u~c%O&+ILe}eY
z5qnekNgy6zq3#b=G8PHLl{Fb$DRDqTIi*-xcS1-gTa#4=7{>r<4J^{+gTSIspCp*=
z`-lRh1dSAa1UjU>xLpuFBCf-4gJRZ$3e?N(GQSlw4Ts_1YSxVmAp@0I7IcjjIvxjC
zC`wM)4aRzqKNPEj87NOZc(x9CJ}#I!QKC-Za%gt%zH@e}y1sq@!|<Rgf`<g=`is#r
zRiT@sRCD(>eSQwhWe?rtpTr;H11?p8z8*>hA0Y<uTcmvg<S#ox6{JdrKnODr3obZD
z^&amWbTHn>k~Zn}g?Qkh6)G61Vs@IofIhd(y1)e*L>}NG0Vta26WQVojg75AIqwu@
zs{e)d<6P4n;AG1KvC$!_%N%!5l<<^^hdBG7?MeW0g^;#lif>_t9u1li9x_762?8Ae
zyc!B!R?je^$dNg}4>wzb!E!<sVnq>`#ORH}$#3FN&d4kjqR;+!;_H>$^mL7U<X^;N
zj6F64N7+0wqw0HQ(}B7q;2j87t1(Eq(_@{{=#@g|MX3v$KoZU4sRZ`jkLdfH(`riJ
zXLeNMk31gygYJP3X~hT^0Mk?od@)$OS)Y3HP-$%V)9`Q~)h%dK&YjypG;2IP1io2x
z0!4VC6bQuU@jOr@$p95NE;YT+t~Cwhc!9Jfw~Kj7oN1=GUGBP-me!k3Qn`^n4u1|c
z$6gX7nF*v!H-s#<(XjKFBSuAA7lR5!?46*aC!ze2RH>S*C5pxZ@CuI*pt;_Z%ISlX
zasuZ4^g1(45fg(DvNkL)-+c<q;}#-{jHB$`0R-0NSnxxckzuxfNZ9|5i%Stj<Ox)_
z1W<YA%dFQT+3mAa(!pZeqsyyXTq*FGh$uP<R?iC+=%iuwE=22Jk81LWBl3JMLjPA!
zubGzEWEoG~#f3}#+=L*9Iz}dnU?QpR7uCSfl`a%!tYM;5$Ue(?5Rry3^AZLW02iZq
zRW~!+&jG~~c?#k;9_%_HP-B8bLi$Q~_pf-c6nQiN8p^$)&SksRzQCUi4M!_k;7eQ$
zv`MhWcvjv4_<M>D`H84<6*#UOM+NCcv4(o?0aTO*;SFe$IIpn1T!DeI2c+bHWQYNZ
zj0WQwjf~^eDc}CD=Ds@~>;LQj3Pmyt4I(QeBa%=d+hxz}k*@45WhG_HC@!+HvNIxk
zeyD^}QTCppQkR4V_jytGXWaMweSGiV<No94FW38euh)5<*Lj`ue4cYIxG^N`biiOy
zwLPLaAeazw=tmu;0ds2w%cKEAfoRVx-&e>uhdM%#TEKI!jUfX(1QAiF;Nglt(3vZ!
zU=mF)?A6asxyyv6djkXqiBJk*eSk`}0(gr&;+u4O=si%K9}*r~4S!H6np{K~K@Z!Q
zt>9s6zKzX`!Im!IS@!^K4`3q-yTQ~$Yc#stp0FHSYe$`)o*op?!etj=X9IgY=^|D-
z^a0Av&U|V_>GkylEHsBu2gL>6h~mKU^v9v&*7t-{*XwMP5L*Ra5v1HJLkAAj6|2w0
zUCp+H8Bz#QKMqtw9*@J{_3TkQB?L-PvtKpT4ajgeH#efTtiHt=h&5)w*<^>mPmO4?
zCnhQbzzg*Sl8a*8{(&3;`~N^bM~b*RA^}j?UGkDU8=;E`5(fu#95T!B&gDTgGc5p2
zqxM;bK#~AvwBrYRSR;fDP`$ExUMe>Z>d1-fM$GIY<NFp7tB6WqAak3XoE+LgA49Ex
zP-j@UBR76MlprW2l?HqV7$4bt@p$-Ug6-c30LuzRt5z0AYP+HGafgE_BXA=y%`~65
zBEb)^1{2;?$X7R_4xUjzRH9mxMj`hG>_2L91|c&9hOp(pb>JXDvX!soSjp(#%(zK#
zsxE<QBM$`t)>6Oko*2a5pfos00)si5fycswG#gDACiljqroKS1Jv2M`l>vj8um!7k
z9J}ku&}>1fYnRszj~pnqO94j91^kk-i|u>ob65kps}SuEtRh6y>5yAj4>IE1wScnn
z5N3wJRCp8yDHd3+jgNnjQL#uq2igF&PzThV6|k@8&!0Dcfrhr}ixP+!1Lu-Ebtm2A
z8K6jEv3y?l<=k6X{<C<|dp!K&wQG_g)GU`^5kSzmaY6UX2_S&A3=EWDk*7bH>=DkU
zgPI>QCaYjeV6RkQCq~2)gbsHCHOM{}q7lVuqqiw6t__FGfTdGoKHyNn)Y?Ow9MQLk
z120SXQ4LMzFhm&_VP`-smry<=gq~VZ&kt4#Qmw7w3qXwmU|+wGbQBef0*wt7Lnz)x
z@H{lKf~;z}zLmWVNBKH0K*}?!_XQ={Wu~E|ln!1l^v<{qDvHu4y&SGrPiN=mmS0}f
z-{OP?KS4nET^#r->2a)utM--IO_;{}fab$?q~FC7NHbTYdv6~%wp%0;>No%c^nlu~
zF7M?v2)}`7xk7X*2%Iaxrs?vVgJ=>05r^NhM-UP13Q<gGbzwHS+p#=@m5!Z|4@bPT
z1gSx2!+}7_`h!9r^y5G+N}%r~2-5;KTV(pdIf{W;p&P!eiB>j=hh*uNw2d#h79p&G
zG@Q@y%~J1xg3pBHebgKpNU{5J3AaG&QL+aD>436;2rvVca$F<xWDOKU4a!QrmM@`2
z1LPhul|Z~~yd`~+oYM|b5V9nIa6_QeyaZI$PBIDt+sHs54LH_ZBj*G&Au@y#*#pR%
z0OZ_7j0sh?BD)WpWjU1QLc)rm9R+L(JSd&X(9m!zO(%m4UN?b%p30kc1n48gNT9ia
z%o6~zJAmoP18ff{A3*BJST__wbuI>Kz$I+YKzT3#qoE-|0E22ZF@S0_9GFo@w*=U=
zgHf4C<qy=159~jYw5sZ@;VHJQZ(yBp?B?Sn3`=dF`8Zpe+7knpY+o0}5%BAOtpKIH
z+}oGX(1-rzu!$#e>)o8)(ndx`G3>P4XNQ)ZPuW91r?9zLra}bmw0EH1AUBx8(7V7r
zNM6E{gWS{2CyPni*qyBRB68BxuS389<it(O17P5PKR?bkDFqrm!?K)mtGOoh6K)su
z>n1-W4jxeWbsjH`^6%rBM7KFWK3i}4|Mg)p4Zx02NSUoa3PujP3mxI-FM88K2QuoC
zzatGpNN4~@0g?M@my_D2GGMEPwDOu;St!jFrJ;jfkT(PBgE|wzUXEjq02K_E{dFzG
zv;=UEA3|M~xhha6FQ0XF<v>*kjya)q;>gjXyG#>i4aivK25#$vydUFAz%Bxm^bqm}
z%~g#dg9SNUP;FWNsb!Ce|IZbKvH}K!G7vT|)u0SJ;02BW;5?aRUu+35EOHb$N8xut
zo`aTc^|*$X1*mQnLMQ3=kLk!hf+Ap0Rt6*iLT8o7FmSY;0%^<H0Yy^x0R^0%nK6A<
z0E(owLm6q@WLfT_1WIXxxF8pLKYscKzS|n)wV(<x1ouG^i8XlGkRWsejMFmkz-3^N
z%H#9nP(v})d9@zUaFpBzDQO>@Hm?DAv$GjA5U~cZ_*By*Q^EIz1d6<Jv-LtCK*5fZ
z{s@~#n*nsr?KAV?2S*EWTaUl*=QolA+YM94Q}mW2kS6m+3{p3_Ae2qv=I)M~6cU&4
z!^5CiFRXLqNJGOG2@q2z1a_##6d6XqS<B;B5t|P*jG51elfZ?@gGYdX;du{~h>L;V
z9ML(p0dBB&y-+(f@|g0}@TUj@MQb`_^<q$U3_Jif`yKvQdZEJdf*}J&>c=Ny$QprX
zRO$BJMyhabOBLv4z|*as2xcTQm1!^&M7b15o?!2uEUt6|BL|ga%xi`aA^_L4Y}k1k
zD%k;e--0^&AhZT&hZ5!MVNk#aqTTOXQ4Hxos0tVM5=<%VxIk3N0ktcbX98+VAIsmr
z11&*ZHs~JuEgsaB9QLAnAlM#|69_^cy78l0&nMRJJ$-1{_)()l<U#;8(*v6U1TYX3
zc$Bv81uGeL=NPg=ytht}hVn4e<Uo*zf~E%#_yDhfnc+kgwp~WO;AN>PC@72<ut)<<
z0!->8(A>Myft7<GZmOQt*;R}Q6X29)y#Shn((NJM00bSnB2<mlt-%vWjwq*t*vRi+
zy?z&DN<a|t$9#L}p$#IP0~qHL#A|@xAZi+x;FVrl`3NcZ7bp8Z0V{SJx)T7LascoN
z*c(6-iCusKv@m3Vy#FKMOhD2oVtgQiera?GI`P&Pqpd*w-Wg>8A1sUj$O4*40F@lW
zUdPVI_aLq>3e6#81|pea0fuSs>|~daNCN%c0el#urmI1-QCH%%p0tg}0P`7CIUR<{
zmarFxNDGo6i>oKe!4-H4c0dg<8N>hpxU(xAB1cph2C0`D0L}G4kT&Ma?FWA6RWL}X
z0~})Z;l<bF-<y>rK$siXJqUJJi`Ev)L;wB=vaYU9D8dI%5d588r|FQnI$pE2MZ~!e
z3=E*=PEer$sm0@f*upa2%awJL9L|~<20{>dxiHKrSRK@!Vjy|}*Y3l*zs+_`=SQps
zM2n~uE8M*G_>IFLimh+sczQ%o`((fxpU$}byzmW|Kmw&I$q?X?oTh-=qtnO=U<Zb_
z9EIlf*9JJz&v{@}sq$qWsI<CpJIF_CW2N3*QZ%DGxJ&md5BbSRN(Cz8a_S7tH`I|?
z4Q7c4bf$cg!>HxCqBCt1_`cjFBVSFGP8`o;445s8tr5Ce5~XK#lRw&dnaYYcVm6~9
z^y#@>W>oZ-Vm`iL>t?{#vxhGvMPLsYTqliCRTb7_ee%<FBq@@H!o=pCP4(;D<st=l
z!|mb|?tU_9;q_8Fd*}}?Pp-mc>f6o?+HbXX*E!MH@5!QaB0TJOzT`HTpcuzbQuXmj
z&-1xB!Ts>hAVYZ2q>bgSHiG(;ke@-vro{!*p4|^myB-i##c7W(XJ$<Z`2~ubh<Pxo
zx2o3{k{7KgEimvF@CxFGcgpDsG`l3NkyDgu)b&q=jgNgjQ+e^HT*eIXc=#(j-G@Q0
z!wD~<2W-V0J?*NqY%NQ$56WDnk69R-k#iAFxDFkDTvw<)!VtXsesaxyZ}ok4?%D$k
z>M69?(`m_1EBPEm{eRF3tzS%aSq>}N8@4Xx))6bKcZH<ywEgR6_AZ03#P0j#7dwdX
ze3@vO=4ySiDjK^Ge@?9=OtDAF&A{O+(}dC}Pi9KSV~NZ>+#W8bn<lF2g)-!L<-m{4
zxpeuXeCpmmUoMkZ)r3_BX%CUj*>%Fj$_I{0ijcoP9&xSlGOuUi!N3JiRa{1vYY;P*
zd)KaUZTJx*emBZgXNR2*f?`;GOXh)~b?%)HGj5BxoX-uOWuuN@@l8zmf}5)Fy&Er&
zQxtRkO6Fj5@t7c2{lHbca`l9aEPK!MtSI)RV=KpbkEQA)J+Z`VBngs+IAxW6Y$1wT
zFH`Q43V2yYR$iC4kSBM4LScuOJ}ZWKX6n(Ykl39AbZg{7BE+UWn&eEPjKMipynEu<
z1Re=rKCG8lT*J*BemeU|s;Y(nCOD_(NG5~qa~lbd7tOs7`2!VOG)R^FzAMgpvGwa;
zy~StF>?KLPvr$GBH&LP5a@c7h=!t72t^ECNS8mE&qJ$f~In<2dX%3!Lhs=Dri$9C!
zS<lAaj7(K$G=8pg|Ejw76XtW5jSs*ZWPIX|I8zwI-70@MrT>*uRt%jv!@<Tg<ots1
z+SBW20tZYVk>|@3&pgEiC=#F6O#ef`O{qiqX`yd51u1inO5@8Ws>>p(I4zrRn(`Gl
zf`g0aZEW%~{hze)HV=3&FD6Ch*ai-ozM%*^`jkPT`eU<nYf!+R!=#5S-}Tcoz4m>-
znmVhwvAgVySD-}7<C@H{3Pqayf%Z0z7YVIA&Qt@VzP#cxH5~_KLoda=@0_$y=fyLD
zo6_=$TafGQ05=`zb%0@LaIo5m<M^P~QY{!4r4I>hQmV+Yv@~+{n7QS2&9!jm7^l9U
zOE3Aq&`fyr9IdDFBTs!(k{rb@lOZm0ONq3@%8jr6kDL}IQQLtZCRf8{Iu+wjQJ3H@
z^iet_or`0q*cUssZaD3A)}TWs<JHMj&ANz=uHznRTn$I|1ZBEkB+_`2cJ#}~UW2jC
zXQv`4B;AvyEicHn%1!7o3>3tel*cd~W4Uhk6)W<Z-FE(A9^-9!8ys~~u_l&%^?vM6
z!`tuOc(u8Y^by=z5@hZE;2|^w>Fx=;AIwg!*7NMs%#_Yi3<HDsC{cX}1(Tx0A)zj;
zi05OPlJ^&F!-;uAWbegZ^jdzl>iuk$ezH<Ll)4aWrAV<meqX}Q_}Q`j(ep<YdB#Ie
zML3jS_e~>>o_S)QXb>--VS8a>>=H(@r$zTDd5~K9MA^{iBuNpqR$Teis}E~@n_o1O
zBa*s0-W+e?ZidNFY19kVsk&|N{fsbc$-5`llB=5`T&HSEPtv!eC1*bH9hJ!VBf;o5
zH+rtBvEWlO0{fXTba&0vecX7ytiKWo?+uH}3CLLNzs#6}yOBd{dBy6TgjpxW<Ak{t
zi+8)6l8NgsjURUiJhb859_=HXKT)8HV-J*3<;#9`-J~#veP`~o6^3;C4WSj)&d1h<
z?~)X%Th*#ZB8{)m*hiDSDHKnd^ua4)Cv1i5Yo@=QkrS8ApDR$ny--fVw*~wVA@~L9
zx(4ZVnk60|iBdGIK6z8={hLY$U155v>`;Blg1}i?!?V#pC`cED3rhRX$Qb=$I~dqE
z5hZg<s5Yd+bN{hDR2gwCx(o+{Y980JKeQKB_i-09jk%dkrzguxi=%D@l@(ro#{Hb~
zO|Ds|V%|v5kjq+C;Y4O7dw54uERiY!A7*?aFEG-^;jvdn#@Xb8^V1eytXw0?-U{6L
zNwNDxBV%jj8IIPhxnx?&&OIWy^d~%V)i|0<aCaH_0OlK3`h{V9o&z<z>XpfE2W~ps
zMu#$AugCrDHWvz2IVrAXUT?T(HJuzJ*Dn()GTbk6ldt@G?5IOo!}?WKK`c*wyfXb2
zlJm8Nd#wU3GQ}L}F&479nu!<J2%un-LO)j|)>ef|=38HRv@kibeaHx8bdtGzTKz>Z
zXkbM>)8~6fPYM#pv31ZZ#PQ#fnVIsiq%xanEnXjzRqJ_mGLoiM`R)YUAQ`J_`S(S2
ziUO<Ldu?NT>C%pI>`yt;8mp2qe>%(WRq5h}nX_4^T%QPx#pZpn`mNdNKtjwy=jSEj
zMw9M6WiCTZx}L#Wt;|8x5v})hE;4*ud@hqlt&?aYKVz2^{PNK!9`p1_Frc}XkYVT>
zAJ#U<|5{*&d}l6c$8r8W&ab^JN<Ye!FM3j#w%R9FQ=F!aqG1ZAH0^P<5tX6jx*ePh
zlM$2YdgAcA{#Vahrc3di6D>U{+|T4>TOHjspU|E=t0<!PF!rW=67P{ALd=IwuA%2`
zkA7UX38cLd7L+mHfAXeDPwQ@tOX)>|U)J*uACh&q<}*EW=!J1wvzPm^(88O84XHn!
zQ!f@w%xF$1AF8>K9}w%nN42Kz-koq$n>>1kv&~rgUf({0O9`zCs(L<0PJJ^^Fx_h~
zyi-KsRG!nj1ETU;bpZjXN;8jjY4YiqMVJa`aJB~I4jt;%ZhP}ZWbk3<b>xg0Tm=}t
z^N6Ecg9$t-tezrO#8po0+TyOfECv$XF1(~pFO(a*U*UJeM-@{rb({;G<xpLZm1blJ
zT6=b$<c{=y(cz>3#>29<f$94)MA-0|;n`Y|GA)$*(%fh=Q!M7WJEd4;2jD&#Vy{K^
zB$s6!BTg=x()6U5kuUdv4XOCO0k?=so5&v`^f@!pEwAwYS|XI4-N)7ZeRJ|lRW)tW
z=lk)g$7eXOJO$j5+^TGsSwxeA&!2L!np1jRs}&!NV{jzK9Sd&@w!M*Ud!R#sOuYq9
zxu1dV32xr#_4mVfxhZhODvIS44)UFh`6;?jqA2}$(&4VUPDE3xv<aUpr;u^kn@^#d
zJ)yvEG7wftF`N{w)1e$H9HzMo&mT)<%1*LpMEHlRv4kkDHJd5WMpSjCHG>X2FNP;J
zH4Jd1B<Za)77x`_=QTb`yY?w_iU*7FOH#<w?ho$`DS7u;fMiqh0>&wuHk<D~wk@0Q
zQRWB7Bzb&ZHpoCmgbSP4A4mD3JL2mey+>mM59w3iFLp?m(*--p%F34i7!r~b-+f+U
zr-4{pScZeVt9WU6ZynCgfWhM3fsSC)0j=v=x}O=tn?0A;>E4lN+j3`@io6mjrQ@C%
zV}9alFYmfva6jfye_SN@3`dt^o4z;`f9Bo=U9$K~TqM>Oe*68qbjMH7$4_iJd=?hR
z2de}apL>L}kag7_aLt;~3CDIQ)8i&}JzK)#8C9s?UlIMGoUVN|sH4TugJ}pKTyEnp
zU^d6SDRg6EOt1wXOS5z5TCDWx@$|gLhd<hG<2J6;(93sD=vZ?EY*=aE%p}p=sg51*
zyCXegK340WJcK`-OW+D{0*MFcfJ%iBRn3}@PF}W6G_n|TooH^MVDjl8ES;jlwS^Vg
zj2+Q@IUi4-rxB5$AuBt72)mw8Q!VLX*=DADg~30En^bOMY)n-rgO4)AjmA+t->KiF
zpWh=SYN9Obm1cjKraBW{y74XpeGC7JIRDXF;U6<;T4V53f4`9IEL(4_!*eBVg+mU3
z^1&B^Btk14s^h4aZ*geqFnDGkj|;lAK6JRQY+pf&ZY5uVttY7Wi%){Gm6zsVB^A)A
z^N~kHwKq#z%x`!zQkv3d7EU@o$l%-jw#=e)h<sDH27mKY=C{`wXE+zg(ljjXHoU6C
z7uc$Z_PAgMIv$77_lNvk@p$bmXtbK6O*f`%nM7{>+Pf!waCZ-%MCD0;)d0tm&8jZl
zyZv#b4S7zlBL)Lctdb^|tye`03w_!wqUX=@Po~ZLzN7ScbLATm%Czj`BAZLx{(cvY
zHu~c-c1SGT^6L>fr?a-IQp6pVxz847YyHM`<G!(Mt7fnNB2SS`s=?YS=ZmKIDX*V0
z^qnIQlJ$Cx|2$S%`tE?+RC94X=U~C)lpfDME>Rf~QrZrM``Th?o$JsGz7}QlN{S?#
zp@{pSNyTfdenCV&&+S`PAtCJ|T|&v?Q~#m<Oz$C0DBXF*&tB@t*VE+N@u`GG^Ia`%
zsIlb*UURKm8{bZ7pVK4Hb2?QTrx1UMbS53c#hk5`&R3sAo!0G|Vb$m1hL_0N;E~uI
zZ)%hKT<J2dK@@pn;oaodZspfPHuC{r`=SE0bergwWDD|AHRFSeY|bWgzn3?!?tk^L
zOYBOyD!=9!!_p}mlCC-j>&KTAb4&C5%EM>r^Y-J(-+X`aJtb#8-S%BlK*ms4o-@4@
zMT?83ArBu}%CR^bs~x7xX@&HC^f^1R$3~leIO4Oe{^VV58Zzx6`ktik(S=JSSujbP
zahCOH^{1e3`Z}hdV99XO2uxYqx3o&7o*#}{=@gw!AN`9)c=yz1NZYQd`u`}%JKP%{
zwK=)0v!6;v#NTgRQ^HdOhFixgcT7^B^;Knc+)YiUyGdcPU>&@kiieZG5hSO~dY4x0
zYmmbrws}g;-}J-Nb2t^QF3Z+~2gCI__a19`!ZXucEM5Mkxu#L$#HO+T{?vU&*6&#`
zPxNL>Mwq#xLjfPfvx!mOK%uOhJST!AwnAUsBX>wp0F4gsWP`(Od5<O;9kvf7ZoyDn
z|8Jw7=QxR5w<Bi$|2;N(juX8GT+VHbo_PbSKuZ7DkHy<B68Fzr)9U_;;xio(c1OlL
zPsNMgNvo1$H^MdH|4?$dEkFZ(%9H*Vhe`0a{*8-C@MpNT;NI;gZi{LQpAJpG6JWU|
zkdFAjz}5c)ij5AF@gCg9Ab!97FM#>?$8B`uzYa!DQ~8SK*4WDa#cXX4|G!M)zj62m
zR(6jk+182l;r|#48E+8zU%!J<e@x#x6{qs=%9(-!@R@&qlH@eo-M7*9e<9e(&7ryk
zj+#imJ?X1El<AASADG(YjoA1NMg|+*?7+q;0b@oawY1~9QdyyWFXdhOE}^W0iS4&Y
zm_ELAZ9U62t}Dq#gc&!GIpV@wvYP}W7EeQl!ML!Z&qL>SV4^rlh;qvg)BQT-7*LmO
zdx-*c41b-Q(t$mKX?_?ejlo>sLknLFP?2L?_*iMg@E_8*&fS5*2*{C<sN+s^Z=E=<
zTX~ZR!=!=^ank%x_x<NPU@#`3f4-SB1@W(k^pX6(-YccUo@je~+xPzS^S?FTab5ji
zm*<x4CEI!%3}##=Q2NjDVKCJ+tTexmjvm<c4>|uscE1F*eU%vgRPc_!&BoIIjbQ)t
zp??nG_xL?EG1skj2Q=q9kna53qyJZm@Xz~h&m;!Z97G1ngW1mpGqyFEDIJP(m~BOv
zVQ0hqQnkOy^0!iXsckFSxUSj0ZCQ)q@zPsw`|Fl}nX^CN_HXz4znPcq+5G)p>+q}T
z+amLoBl)G>zt!URApezA7dCmaFqmCA%5vv)|Gw1zTX`|IZp-Vpid;eS)Z02h`pMt%
z-W*4J9Zp?Ja6UIa#Pa7T%faP`PU|M^$#w?|gx|z07}Y*fJvr+>La5}8K532ZOf&6t
z7SyS2V{JX2{i6MfhLxT2hEb0117oL?93yugaaJ87q)h8u#V5r&wQIWvA0osFI2XLU
zGab+QGV#R4>M^3#iu)>Q@!0ctPE&3CQo)S#$lZ>+ZC)L9hX}fO50!!nzLVC))B3W^
zJ#*qJIUiOvPF~9EFwq(G<g<!2?C!4R^h^7auUqBRzF5$`VmNf)=v$?v&a1hV#rgPK
zCudbg2pLABD4$TUGwxc_34Z>xs>;TZp(}NgCvkZ~?rQVj?=Kq9CK2L{CHqnmiujv0
zMm1$;KUnwh+m58<=M39DYg<_Md3xre@}2tUI+D1vl^!a!lJH&bFyY}h-+~VJr@qCU
z{+>Fcm%7Z~&g9Sf7Dz7DiB}bz!L@h4nEwVs$j_;N6LTtY8HRuBVEM)v&xdP8vBN_@
zYl3|Xru|M?KTF}kv*NwJFB|ads4cFv-x0v)4r@bGd*N(CM*JE5goHyd%(oR^{JICd
ze})`djx)YzfA(oFzjFK=PgU7SL+yj*UqgEN6Ya6t-Ia@Lx>dgF3Asj4v1Fo?gAG1Z
z=I~*3h2ONuQK&EO;nPPylQ<GjjFb$?7Jkv2w>#Ta{POE6hhqctch{yjcWP$8dlgL`
zNwn(-shz2s6QW*{;<8Hnu@dKBxia_lYvZPg3Ug)S>9=vYwKH`dHQ8<|ZYqRVYF<j6
zxkiK(A)SgJdhIMHr&ajZwe6b*U%KY8x?eWz7QR@sT9d2GT;Mh|Xy?5A$a6^$^g`V4
zDb@*JuT0QxkC~9H?A@q|dU(-Cp_r38-$yO4h~drG3W4@+v#OVso19Wt^zw`Nzka?}
zXw>HQ?U6@ff%ZJ%XuEq(g@%LUz)FeVT=pH~+3)>@PDPBekzRSYMaG)Gn(XAI6%VN<
ztoB>&B0n8!^8(z1&l}aIm6Gud8ytS-`Ng<AR)0^t|8t#2M&p9`FM_6nj$KRhgnTD^
z^QBpi&W0vWQ<b;gl106rhXw5$?9JcSd8ioSiuN1Y&r9X-Y_5D3>YR#^s=GWqChR;_
zg?nD6ACLr@7^4U6pIeO(8jT4rMQ1zDWE-j2<J&}MWk(1G{8be8hl{;F5O<ncKh=FZ
z7;nf@{&rCE?O?-U`@$11PfoiVF&E;Vt*J#`wZUIoBo;h5t)mx=jW_x}sq<p&%Y<|B
zImcz01xW#8-hsiQvB4q@?TW973r4&peb-77s~$|*rRMg!552TrqSjGMJb`P!Q)85~
z@Nuqs^IpT(64LVCuZ?;IcRcU{i(}sf=4&gW?wid$=ve%`()rEbZjt4?>w7=v5?8BP
zzQsiU(+Ud}`V~dJADXOL+<i<(N?u$W_3>x4d+vUDRjBl(-r`lYV7__#b{!+hw<!hj
zOWoa<b?ykYip<m$`T6_5&j_wx`u64I^uyW;(F#xJlHM1Fl|CVj8U+o#U-B0uuN4-#
zdv3n<nkf{Ysi_?<>#|=IDqdN(;%u^3eJb7==WA=Vk;lqXK*;Bft{7b%w!?PguvKpd
z8XV1+PJG)KTF)tokM}H*|L)hCyx`th`Zcn0cy7#L#nyS~T44fpMQ>dG(_uS<Hn&J)
zmL*|a<4Zx)v9kSBy}fPyLl>8-tn_N_3=4X_T$`M2%-=qFzFKQOl=kM1gJa{^{)RUj
z_ifSyZN1DFo4)!4eP^m`2)vNW(kt0C>N72=K|N$rQMGhuCfTR3p+s${Z`G^TdGV&M
z-nSJ7kHg|VxnocL$C{$bL-fa<t0?!1zbkB0DstRhnk{1gGU9!Pdg$|#U!7j%w@2n7
z0wbP%5_$<Cy7x{v_g!%;?68-*n$H?->(wPuyy{)mm)du&w71GBIkvWOB#mTgNWbUA
zwBPx4$Ge28=;-z_sT@D=_PRqKoBSL4O3qk&HIUl!FAgW1Ia6;I;Z!*0Bl<0^prTx$
zZ8I7w5IG^4tY;tu+yj>a)JPpV<4iUA{*)3RX{@iGh87MlP&wc=fAW2>IkC-pbpf46
zfFF{Z1~PS%@N=^o`AyADbCH&?S0Y#JuYSq)PWCs~<F8xXpI^jm?ve|;a8&2Uy;`^p
OMp<56u2{z6#(x0u$oy;o

literal 0
HcmV?d00001

diff --git a/_images/6dcf42bdd95e0ded30ed1fb43ad73bdc855130d58eaa29f55050031f80d0a148.png b/_images/6dcf42bdd95e0ded30ed1fb43ad73bdc855130d58eaa29f55050031f80d0a148.png
new file mode 100644
index 0000000000000000000000000000000000000000..41fefbeb018eb34aed1c01e2596a3d4f130e2b82
GIT binary patch
literal 75959
zcmb@ucQjmG8#gRX2@*Y{geXxGA)^x|(M1cAAiB|__dW!PI!F+`1VM^6y6C}(9z?Iv
zduMdtK6&ok&$Hh3t@r!qJ8Ov;=A2!wYwvyi%7i>okt4oJa}@^%hgd;gS{(=Haw`tb
zg(G}i@XF0V+WX+!WhY66r}*IV#y1HDKNHx?YdhiK5RyS(7kbBiPQZ&o&N5og8g^#R
zZbpu#I5tMk_SSaJ)-R23xtcmUy|lCC<>KMu;k;$x>})T>&HdlsbJ;nXbH_5ELU3?y
z;V4Kydg`9MI_^=;@(g)?J}8{?<=ck%@kOEsKX`5th0XZm^W3@orspy#iv0*Dih=d5
zbN~hiQ*g^ClQRN0oaF=4o#59MZ+^^#T@AiPbkSdpirOsR$ZS2OQgKrMOP6q+#Fr{_
zu1le~bj}m145#&@dU`I)$=%w^yyN|>7#Qi_YdCN?PY>tsPw(=S6?p#sFg(wpz{>Q`
zH7a_$Rr;UHBq@g_A^4xmdMAd3{GYeHvq_CIf!=UVn0E9lN2|!Bb$|=Av)=5izmV*7
zZNq!vLE0da-edNi_dIXfVd-eDZe>(l_3Z7FSvCm|{m6UZt=uZ$J+Hvce}oMSFC;e#
z)^3I0OFajF(3*DAnkLHKVj+-}?0a@7y8BZ~^klQKBTXXh{A}agM}Plb_v*Km{LB}5
zuIKtiCIoh#eM#;cUX!Cp+q09yG0&Aer6d6xD+)Ex51;P;H2aYFn)rbsSLJLJM>p}c
zYvOjt*oo+D4z^_bi`WS-?)Ert-wAMEYYd9I&VhD;uGqZ$Oe{tg+(Dmc*F_O@{N3a&
zWev0LRIxd=&ri#S)<f7*4GU`ar&GPw6ShzE+{T+Hyf!+&Jo}pJa9&eWbJ8bqO3s$z
zOh!n>>E1QB>`lD&)!_7w9G%1EFsgggDBdo^)1y6&vSW;3Z3Dc8uJQXDLXW+#2A$0}
z43yL;#s`x+Sr<~==<#kklLXN@;zR$>du|mrgIv*>oUWLOK-8V?&-`GHoc&J7ZQ|&-
zt=;Z69m2L9d43S7BkH@`Bj>c$eB-?9I2`WSOsTm!GFTh?hVX{tFEu%bx%ZkX!`bXt
zWAEu;)Z}h9gGURm7B(LIkQ47kB6kXO&JNSgh3`JkH5@K71=Gg;`!=@+!ClRKZ%lzs
zzCm58Q<rd1bJz{1DL*_%@KDD=JvSGF?ww>p?&`Oibw~Z2x)pCVYnGoSBlm~9J-NVi
zZ%?QHiKQPqCJp{qs<+=4%#K(Uq->uI*6#D`AGT>0o$ilHY_wFLpB*1=Mf!G_OB`$G
zS4>~^*$OkT?~%FTVo*01t8*}gC-K9f<z|cT*@CZQm+->o9Jg<RX?v_o`C${Ac!RRw
zc$3Txy(Z1tZAwosugat4uQ_xJM=J#jCp+B>dK;cMCeQZED;G1uD!}uEH>!71ORW{-
zxaO)RPuF8!k0P<-Et!$xZCfo2=MOfp?FMH&ar&MdIVr9*lPB|n9@whMPB!ta5RSdG
z>*t%-b&-4h$sY3w-p8Yp^BN!*idk<-?0r^6gPS`of6j4S$t#Dxczn-J?v+{b(Nxde
z-1zYP>sG<}QNfsfBQc%N!SCu7y^ZO2a`YFb6b+8ZCJ(#Z4(iUa=Z<~Kf?EmQ-hpwo
zyIKFoAlli3aBu{p@5)f%d%BY5GFo1{mj}{r(#Y8O;wv;7Zu@PCb>e8Cy!Mu9dl4!U
zxr$h=Tu2>P6*yg#lUS}jtL#Qb4A!i>B>NmKtroR&(K&U%p`M(K)}7MkJr8SbRmiWL
zr|A+uwvG^4W_KO65j+FQmzv@>;nC&*1<vG2=cN6~R!gyA!^LaDvm{+&SY6xdEm>JP
zxiqx390QWC(=0m4uD0@cj9752{sJCdJy+H0HW*c>v|~Cu-{Vma5Xo&_4>meRC-wy<
z_RGp04(H-7Z63@-IyDjt+D~{aI6(KGiFn`$-Y&Y*d7h)36zCK&F)^{ub|uA0#~Qn4
z=bID~LcmpfV7xzSC*eBslI?7o`^;Y^jB4wBf%lbH&%QoK?{qzIT1*cr8h0H<gYc5r
zkPx{Atw2QM(Mlk}X9b6R&rfr_MOG5X*l)_wd$63Io{qWCgeN=qDnQ?fh=}Z)Nad<;
znU1;)>a`Z-9%GY*owB{hw|@sQ%*P`l61G#Nu8YN(6}OVBiattRD0y-3n&@Ud$0(ii
z&qrg9?e}m{C{&?k{~hstbmSI@?&>OK6e)YO`te5HIh83b9o;OH$P#BW5<*9dnXNm?
zb|A}7_6*MVHmWB<{|0ik8ySP>ikgLyzX=UK=ojo1r=1eLf>Nw5d)~(jJLxO2GFUJ<
zJ(TzS>oGHqokzQl`(V{XbJT4`_vX9$4R;L=A4dpH2Xol4{8@V+M}>f;_XM}!@_mZy
z$da&45(|OvQI2n0>7ahQ0$WNG^gU_v#4Kv|jkasnOOZJ{gpy$Et~UWA&YUL9V81BN
zX6vZhV<&a5riH;*)I8OLm6@4&%&A*UNkzrgmvYy)cruUib-iuXn!w!GFE>+IhN0!L
zQ*g3lt$ea8()R=*e$*4`xRe#QU0Q$)KRY{{bbBWQW~2N=flerBoM`viYWEo`?fhgx
z*Qt~Lr2G8jy6?WZZ&`m^(eZkfY9(4U#()9zms6|o)lIjtcYMm_2Qr<VzTSplxS-vl
zh29GVdiOG{h4(c0QOj5Ty*q({fd?QdnxTj!m7{C;`ST|j;j=TJ`OYLPYN4d@+Jm2c
zeT}up>xA`r<r7EAD;&xi)z~`MB(QdFH<7FUoSdwa(O&3E86D3rD2TkR7`vH%9T^N|
zq3HGx<}r_jl<rDSFvdkLi_w4f$nIj$;Z$&ARHV;tX7veJuvF@_4mTUGM=uDk4CFC)
zJJEYCGU9*~D?jho+*6S_b?y?HAri+{cB4&N$)I0C`BpWt$5q71+|HvSx+M)ZGBUHS
z<W`VT-$6{aAR@&u9SOWu$HFVQ#Rorw#G9t3ZWSQctQ4YNgH()RIGMtOcBH;ydI<%x
z2(f)FD7&BMYBiX~=?Tp3O`6t0YupVzXEp4Ef!mk^-k`1%@j|K>7tVvV5^_858QR0m
zAPJ@7;o162*G*<H;hYAw$r>ucbJW;M&^P;TG3@4Sw5odE^8^8#5vtMZC1tzmFJJSg
zOZrr1R*SpMz=CDw?jHYOy&?eQVR&YQNNBh3*$!xQPz`n>$+EN!EMfPOopi=#l%ZYM
z`#<}($Ax73?*!iG$CC@M$yMA{M$^EA=DsNXa_qJ+$Y57{Xy!U(6f{{$7sTKrbVJwv
zLfY}zLIqgS;>D37(Dn-=)66ZY?zpNhv)^*-3|q(W3K1b37_BOvu;GKP(rD|Oob0YU
zTd8vdi_c6l0_;wl+C{U1u|`Y1S-Ku`?>pk*0nnbNp?|Em@71fordBXmd9vLOjYkK_
zfCI2%j&h_8Rc(GJJlxHZ=*)`KZyPDKN`NCU1*Z*fu(jBlWUuwg>fgk6L(?s|B6r?j
zH8`2R0gfVy2laU3^?}5+P-DSbAQ5!a752oZcdhMi!7O+tC0y2Z+(i>iG7tDZIH{%j
zC$JI@yeCT}!5~Rjfc@>yGw+p;Wh9k_K{<WlC)=MBT&i5@i*SCc0ei6f+k<9vJc15+
z+jx?K>o>r0#)BR?%gpc_I!Iv0-4?_)Jh}(CBESuTMdgmZWkLUeF6*6Lm-ZX!vY8E+
zu8Y>M4i`Ig*{n_f=Xc;?4LcLXw&NmHT)Oh5FUF@`Znx^crgxvZpdx-ao2N(g_q-?}
zqG^ES$7Ar`r0Tw_u+PNjwQr4sYsnjLbev%Nb@zAh!)Tm(mQ?=c%U@}jtu0gxQAR4=
zB*lG<`^wq8+^oNp3@N%Vr#@^Vr`~t*5vYyQ?tgSGw~MWz=+3W$)<C(TbYE89s0(J5
z^tkdD<nNonmM;rnsJ@26IdRQ7rOHD2W7Q_7(5*S2=7`nfesVf~QNSdpKFY2X_q35U
z<|K%f<#fNpfI#SHQaW1Zu{wF>I6U0gU%lyN4_js}itxux-O?&7i=SFR+;tdmsQ;<5
ziU^X2i_)-F+Uy$L;WG6x6(LV=u=pxVdNHK5BToqtWKC0*ARd^J^^z&F?;}0o9gJFY
zPZKYdlO>}$51)h9a=J{XAU?czSGY?Hvm-K!?VQSbf2hQqu43X6|B=e+MS4Arl(BAk
zmsth<#dos16!4!`WJgNh`N%i}{pC~YM}`b5K9L%msEE`nMU%dIH)O(80xP6+6%bPj
zyJk0~@IqAA+N48udvQ^N=(*=%bb5^kEwQW5=q;mDcGVL>@B0O6O|O~@zrZ_g;JTT`
zxpg;JTNC`Qt80+b6ckP)O|`T&aTSnfp3kLYFe^x3(~jx<{j764@Qo@9fgrp?RgM7<
z0qgJ_$`VNu&X4o|A!1-2v0`+MT1@Ls&;3zasskz<dMQc6!d@l^zb#Kpfzta}c*oWi
zN~T6uBNjMn^ap#Yh@bxtmMh_l)H}q<4o}0Gc_UimdFN?yh$`vTt;_Q@`GY(hjbEQ;
zZ#4Nk;|Cd~q?uHFy6|)14$d#{XyQDA(8UC`CG*6U%x*J^%ZUxV>nmDBVqBS)Wa=}U
z(z0P9ssgh17@jJgCY5169?s511(YclB@YoZg=oz6;(UIL^u&kMa{SBQn@^)rmQhCG
zjMr1m-jOr2bg3(=8^0!D7RNp6yK}s2E5}Raz2A9ivLT-HK!pN(E)j8FMY~x1?72Xo
z1BOSN?<JKmqipCCoT*&qriUm}1QD(ji?OnYVq0xVTgNg^)l)Ng`_ua{OzZFz?JudW
z&?Avt^ocL>*s)izd&2poWn-qPukOe1VNv>=2JXZag3GL-xT^gEy=i-IU&!61*v)zN
zg^T(66>^V&?Cu2pC3DXaReC}**3%B_@;S`U*@jj&YMa>8j6@S@roCzz(yPh$j<tg?
z1YBTmsheQT&N&rAd=6q=p`AgMFT;q$JyU(m{kjgZEYtzo+=QNH=pb5SeWm!dcCkop
z_ogzf-z^-$EHJYweY1B_#?&pmbV;Z;tc^0^TNx_NSshA8FEJOHR+E(xRMgC~xqH!~
zJckqsids(+S}zWE)2RpgX9?DmD@<uRL_KSU5%w~=W(DavNp;frtUI!@FESg^M5J24
zhywjc6pWh?S}Br^4<DH@rz%`c-Htx_QX66RgIdY!^AJTOo3zCNRc4+7&%&o~M1%R!
zo;^(0Q!B-!h$_Xu2p9>a8CLL5<ZXyj5PcL=Qaw^YWzjnh@CiI=r-f;hDe^0{T0geD
z$tau4v2e|bQ%nF+zxRaynmMO9t|ep6JZ-L3xLSA8&xz~s-tAgzIVM~w<C{qBD`B_E
zz+4iikt0}&KIf4~zZI6dt?;=bhI$F7FN*$1Nj6R4lcB>Jyz<st!=>ilpe!8BmgsVe
zdxmfah$XcW`vJ7xhBFSAA(t1O8uPFZ?%e{0g_YVgog`KDso}3YK`%c(EzHenej(9X
zK(j(^qdXmuGZ$uSuo^Q;9mF44n=5Icp0%t#pDnd&c)ZkNeBb((xt>KKxV?pbrrK=B
zN=lcz{z8El*7O~1rf*>@_ki+34rP0dQB1%Z9~D2LC;P$sylr`!Lj`z;J-n;hIr!4J
zvnx9%KVhRlVJL3+V%44&|FxiWYO%~UcZPPh(Y0hWQXoaoMj}8cQJQ{}_Hb^NW?nO=
z*i9swo(@4F8f!Nja^#tew$@B}WENc{7M>nVinlvVr23J{Q5~@)$ibZc=q|-m>oyY$
zp5L@<E;7RON(Z=~m4`Tz9;GRB7#r54nIOfx5?hHuiqXftN4#tOH1yhx^V1<Kxkrj9
zl~ViBuy~S6GYcw+^=lICr(-Har-J&5{lIU;2ky^rwOUMiSg==!)5;r5tlL8s7ZvUv
zEGj5}-_1??qWtoWe~)og2kY*)tDiZJWyAzt9{F!otE>}x6U{cvpt2HUOjBd`-z9bp
z-PpRkoZcTH7seag!D_;q6`YzwLkq{n?bDw(>x-Z9bbzrMH`uFXQW9F$K5D-aZ(GNC
ztV1uVoQ6Q9FRPnn1?&zvDhpA+Kpo?PR)=@K@QOfd5q(x%bf8wKN@d-~jrJh7B6`nq
zuYG~+jzBX^3iXc3^9xt=@A#x^j-RMAxM^6GO!7EjnTDm@EdvitY@C<_m7UB<MHvIv
zbp>liR1MUPL<^jHr)pOxc%-L}wFBn%p3Mnq@L5m@i(mF8S{1sVv{t;dYH|FMRO`m_
ziMJzv<K9;1JYi#Q>LYI=&QD#l^(L%HtpRm=O(#m)7yPd437xgAL}k|roI>TPn4<?a
z+xa%1z|57!+rG`Si%YvMr*xt|#)&8K&?H)OE-M{~e0G5mAv{t-p5J7+RP*vpcsNb5
zC+?D}SFp6DN5EEf*Tazj-q`2tu(qQit|W~0Xrs#0MZ&=BkX+tW(if=LxVVi<I;L+g
zUl*mMRSKj^a9Fj$8~$nv3%$Zm7ik*kl5H4O?@R`^LQ}(df8vs%MOr@NhF5v_EqIUQ
z+u;87_7?Z2*}#NeRJ}{_E0k3^&23>@g~<6VsifuNt4gF)hZ5dmF*Q=4(J&q*>_If=
zd8_2xAk%5tckj`5ea1|-`rt58lVze~k<8i#3pGRuzk-;Z%R*q62s>sdJgte9BK-sR
zic!!+LWeS*158#bpfpd-QYJGw$iF2>K9;59%e6u)Q-O^iUc!q*BItzTZ9I+zN@B~}
z)-Ol+U~(tv6}|fIIC*8m1#g;UuNZMrXSzQRqERgqh)Q_fB$SAjdHkL0J@xKXJnBkg
zdc$kGpP|B3ioK~1q`Yj%jfZ~4!GB4Gnf-vn?g*ILiv0M{DUSwYwh~HUj4~16XlWmS
zG4By1-^2RB=PJxO!LnRaCfOTN+QOQef$t|4R5nxLjOS``MK)!wMIq&w=BqY8yh9PU
zyoIK*!fx{^lKH`I7fUfg-jl+jvQ*0`wC1@h1)PUE(YnJ5wXnB-SmIq78&1mf?w9aY
zh96QYmKtQ5`I_GN<5=hEoMg=hGun?V$$R%Jp1%G$p3S>KLzJLsahRbKiQA{g7U(*r
zt<lwxVv|~%zrJd>vzN_Jg&-1lN5RLD))Qta{GR5>(nIgw5LE;bWgO4bCB$=qA0NK<
z+VLi?t$gh7B(}CtG7Mj$&msTn=U8KXA9iUJ+ihZY>{%yc*dgei0q?LV7Rb6Lu#uL6
z6hXAF#JGxoFSE0o6qZ#@d^faB7e-=1anGAbV?KVO<on4oQ8rtRw^>N~_8<e-=*iBF
z?FQOQ5_HIT5@zJfH`QVo!O^+p5^UeGot&kH_wFtW^@IIRGgpy%(NT?vujpobFS@bT
zy`9A8T`hR@Yx8=kDhfq6(z`S{ZJa;CvuJodl_hbMygux?>na5|Od*e7Flj}_|5SYA
z7mwo?yG4PnFON1+=pdk1Dtw9%*2+ddTW1!EihmM9EtbNP0DoFYA{I@m%NZ?tO_4@}
zE|!IA+)=3B3qN}-t(f2?CpSAer#OwZbCV%{t@rf?^$_YyX#d8kV5Ru&_)ZIZI7j;p
z#&8ZKS~8R|>`H?a-<-H|;blAg3zEk|+Dc2A-yfEIl>Q>^UGmg^Hzw<L*iRJc4Y6xs
z;q?A*U?B}8;Y$Om!&Dq3@(6wXWZA;_bz`w~(X@vwO&7^8QFA;a^HdHE3mlUUPUVbU
zlP5BPXFX0}X*a(0Qex2R&aPY#-j^5~L;p;6CPqJco_XnR*Hg;xCXb72l!NINwY!~+
z+gt<0hJ<-HPbmjlg0rlHiq=IZ<gA@{)s1A?M;-ZCSw5<Up<^C**sr;NX+r61qzk_I
z(sQ4JxOfEd+>i10{d=-x7yIMU@2)Bn@%m>zzBLrX6dPb~D=Mw3(u5#D*EjoLmU}E2
z$Q(IkoPe|3^!%2(^rZ%M3ev#{L=a<$ut2QMjB)R}j%i1rVRqAJMrXt?b&S5`W9lVR
zTAlIu!XC8{YU)^+7)PxBou~JIq36*4x7|0zrx8s-!Gfme^)cvWq%bwjtrE6O5=QCZ
zuD2Kp2IA2@)2H*@YZ2tl(E+Bz%DwkAJaV_0zHH9DrnRx==3~OVN-ut8!JTU@Oz%A1
zZERsAZ|oQLZi*N~gNt~xCMa<>>kLMq`6{L4aDkvzz|<}ESORl6#`ZbGoS%*J5J}aH
zO5pAA^arW0wFaJj%x(NGm$Al)kMK%Ba|qZt3(>L0Af(i!bN!f_%;lrr`+8l1!$7o2
z$@(f$hh``5Gw({v8xJo$KvoT!nDnT}`*X@hX{Ab}i>lrr*X0~7?US(Vd0R4rrxDTL
zAK#;jF)Z1Yfra>SQjH(1WqvQQkk3fRlcK*T#+uH}x?6TUVu<><#xg9;$+nj)-$!v^
zQPn~;@v2yB2A@<3HNKA@W13NOi?iIWn_#WBhr0X$<-~AWn1a_Bk;X!%iYVfhqv*X$
zg}KaDZHKJ485`7_9+DxjLP;BJ7}w*VE`KUiCQ2EBk4AmY2-aa~MTH-@YTqw=P!Nl7
zd`uZ;#b)6sIUE_DDe=Pb!~H9*F}1o~GNAiyoY82#QiWcFbeVu!EsmC7^1^Y6D9e(k
zuUO`j_Q^-)Y_QF?l{VH`(@O)!O4<{t@jX^_tTfB8-hI*j1UcoxCPP>#<@{S73$Ycc
zH?>+_{io>f9fnK%vEgV*1J2`=mj$#(#&6Os97t*4PZ&u?W;h>aAurHnZeYCF=#|(b
zdlcW)F3PiGZ%PRVx(?jWv#45$wHTSP;^anYa5lYaeA`T+`jMKMBLanDp*p}<P_pkq
zNNNj~r3f);DS7FE92UUUwSH7M&ac2X=Wds`vad$I-8p}8G$Q=<N4>fftrD~~mV|MH
zC;P$H6N)vofUNF=iHy5gY8zqAig?3O<TYz;+_uWe%Xo~=*ahaf1DjE&H!CQcwcB1q
zbEbM#-cR=Rq&#c|)b|ZE_Ipf<4-5PCHC)=r0ytXc_hlTs>}?`7Ow*Q1G=zf!*By{-
zK{V~~5w!t_;mM^gWKaCQ&;-4C;W#0;@oE`@Mg#ZLPlk4F%Q2zEo~R}>uWNSQa@L6t
z2pOtw{OBobA78%#MsHV(we=3XWYER=7<GMa?B`1?eOrY>mDPyyj6C0*4}x>>23Vpb
zVjU3h1{Xw4+|>ZMktgb@F-urrEleJ9iB{JuoTh5#$!)WbaKllwC400<%F!OFKmVLo
zPlQ_V_+aMUWA!mkpLUO|?+^L-f`6fMs2Ia*mExOF^cwqUPR|$cx9X{VyB6{EZN6`H
zZxHj(;P>>&wLL81^HV@1px#PHS{1anuLNE;Ct}6z-5(ji1cyAtANR=OjiZ%v(<4JP
zRVX`&q?5aGXe+fhS-c)P3nCjllbh-)l~?V6p(s6>XW{A%(M=bt4&$q|bG5*Gekm7L
z6zMkDd^>ZDcdQw%thHzfzm8r_S`$g5zGFS8Lqc34?|`l^D^L@0(mXEthOjla`H8yn
z(%MK;m`h7nSS7(RT)SCKdQ_Yth~UELO({C(sLSY8+6c!}%K+{NjM<}g0(r&WUY%4`
zgM{Ppk#mwg@s8ESWvF{rZ8O8L(bF8%jGvZ~ClSfL_xnS-5eaKiB`dx=4nL0_>68&K
zB96xob46Ky>Roc3mu=4;LMJtcUbl{ASs_wDi&6DV)Y3H~IotKFIVG{Rwrzc7Z*yPF
zeXr+Lv(%VfIg+@&(K+L`<37Tp@Z7e~L)Dxi<utK#f-tGTt8L~@XR7K4ukqNX*PPj}
zEgX{cf?8ZhjRcy%M9#t66|eV&H*lK0O4~l`T)BnGNi(<RUT8yJtmu024T&;rx7y^G
zaki){;B78hXy(hd9z=X9E4Y`4q<#+*X!a{^BEe5VGuv8j<X|UlkyS-?9=!aK#&A9+
z7IeMgCqS^Rm5Yz&<?XJ-0zB%^Rn(~w1tIZ&t}>7RfYT7h^g}@P{}-MTddmOxAw@Up
z0UGI)U$Z~*U)YM&BO->y>%y4VM$O5L#2NFRC(`N$fT;g-rCz$-nkqNpC5cbxwTgx~
z0s!EiVz8+2aJ1uG8%HF--cINr-~CS)nV4<@zO$Zze1Q%J&<lSEFu-BXe9DYCD&44D
z{_<6;%zAhu)BW$CZ*ziY(<1?`IRhwEBB%L|^cQ)K0A?-pIrTK_Ox(-e5T1+C(E8@Q
z6te*F1FdXnCz3Zb|2A&~;C7S#*i_gIZ*7@z^B=AcBF<VQ&K%~u)4GR-V3WraSf4d<
zJUqN%qoAAFq7ViyNn!B!4ykXVOzN&I&|zQ@PTY81LUgXw_Z;hMezd!k?AXRZbnV(U
z7QB$`J{{tHx?(W-IE(Tx;J5JDpxhgm@iGbyLwxJzPg2gWS{JF^e8YVo4sQS`qVT~~
zz+qV30U-dfL*RiK>^SV9{wxaaeMu8d0Qh`H@3jhd9X5^2a}oJ>KxhG=o&@*@!@QKJ
zTbB-cB7K@5I^lTo+y`PACeF{WP8&7YZHDtL2AtKp^D|MvrSK&wM}6!N%hy8t!<Yrf
zo%=ShfBOfhh76Wrv)p@MXXGS~NZVuY%?#++Wer<jv<Gz4X1VWidG&s2!4*SmYgDP#
zpd*-GxTt6c;CxhkeI+17UAx$fSVnDg2obqe6j!74Z>YYzQE9`fQ09H;2oY1M-r5Ua
z9{tq<Y%Abwi%uFR^?%iL0;sG5V7bt)YcKvVL!e7BFY+)YFZ;kP#SVVuh&N9+hgEL3
z-4zA|IQ7k&b6^;aq2UL-xb;}YI}KH__fpq{U%z`t3W3DqaeuRDR>n4|l1*Dx^?Lc3
zBOxH1jjPtmn3XkHSy=(C-jOU!{t9%@_Ya4BiQxbTCsO=~-8|U=7ec0mXHdL<h8M6Z
z4S;>T?K+8cqxar?1ThbPhI%tlaC1AZE*WCI02dPpXwx{(_Of{n2!V#KA5JBmEUo)v
zPO7+gc$I*3TiV!Ih*ISGogPGQJc-240pjj~Vl3xxh{qK=oQ)=6+#K%qiG^rEqU+ar
z`S>o^bNe1D()%13xlVYP(<67)Y7aZ@irlx1d-+UFOkPv-noT9o05W2%&HUdSs|yF?
zZkA`-8g=^wK=G!XiTrfFCtKdr<Y`mT(*e;52$WvFxfzCYA&A`ojpyC~N#-%WqCYs^
zk|7f|WS)j}2hrU9;|G3DnvVzrAWRd)y!M}q^8ORtq=%CuW>F?M5XBj1;3Eu1miBz%
z{Dc9gOXA#n0*p-|fW(iEkB7lnRRC51$}Suj1H%F&00_ImZPt?^qwCxwldE5o2$8!w
zJ5T?{<g7<Y=U4@U;r{CIYWbwkYV}SRAQN$lIRDHPprvL2Z)9$0`0Bd%)Fp^OKIu0&
zr7Z9oUvZy(Cl@bZLxTfW3N`<K;~=s_;62MF_ja9$@5}vP51YAtdXTn&*?^Q7Mu6*p
zm<dk3Z+{3PNK~Qgd%C9~uKlkb#Mis5l4ba={?=6i+p!;DzJDRt`3pRV8yg!>R?Q_4
z>tmJGU|!~<(@t2-Q(UNa7P==x|GN%M2H9{a3;>1&By<Ddj$42HkjEZc<yQu})$XLg
z5M{<of4gpCHI!{THi>p#QisTEN&N8zi1f(Q$}k_m7ry>W(~;p_!Z1!ieh{NhR(JL+
zoFw=Fb|3S^y&LnK*qrmO?)8cNs3MNvapd^$i-zGwNGFtm>CA3IOvq0bbi|7ullr+?
zfuF6EvbWfJnSOqc4)Ra5Q<v_R?lt_m2ln(3m)EeW7K=9kNs4HAc=5FCBJRbx6>(hF
z9actM*~FRL{DJtCr?Ajid=?lHZV)59EU9{|VSi`7x*=JcJXin0LjeS%ppE<6CLwR4
zru}n~7TEQBsgCz(1kx2gkc!btraa#rKO&Uo41_V?yXs9u(ZH+C_jyh^x%A@z8vSgh
zY*T0e#u~;-o5Y_e&W4Tv2r}xiL`}Ybqk5!p_*)HWRo+<>*7Rkb`3j%x;j7vu?o8oY
zhWFhslBb6h<AwobFFL*1#NxM}VOd_3@#o(TLq)DABUhK;(=e3SgkZq#N9={d<LR$M
zGs$>b3^bxWG)7Vge>Pg=>NjJ4vpsGCW<-hPd-VNQEvKl%{v{kCD{ALwxS3i0IU(sr
zMi~}l!4xunUq@_-X$?CD4?S!sckcAX$6Z6`(0&wTRhqV!=A2Sm#TcQUI0cq+Z9kok
z*JLq)Pa9WC5Prl=2Oli9RoIkk7+sY9fw*=33iIJ*E9!ggwwwvdoA4*fygu+V7OSFg
zAw;VbKox}$!xmN)4SiRyh+kfo3Cr%LR>4%k5%;waKdiswYGg`$?p?W~MYQ+Rn8&U+
z?iojhxhWj;AuozXQH(;A6a(9FVMP|B3=n7xNyuwggYjRg7sW3d=NObBCSU=PS7bI>
zFg*ye39{_ylr1+6OwdTyU=&RYdHCKqHV761SHC<t-etFcju`tTad8f9<q^O(x+VUV
z=f{VY^udop8f7m1S$y6q{7S;2t+sehiR16A8a-*!r*34b8iyOb$ck*7Ukzw1EaLyJ
zu@^pmHY{8h!4zxyHU9NO6Xkf5kDtK4{5eS_#pX=EkA4ybXBmJ`e9^dU5E0bSx9Q<F
z`iqsKfqeQ`G(LQ4G&0ey==(1%E3)1uo`;fVoO!U#HJDOIwy$QWG^lIBg!A@2X45}B
z=Yw^!eIAMd>2WQ9EsFdHVi9Bfflfybm1D~M%!-;Zhm??}w)v_ofdY%FmjjGut>_rm
zz73OVd}t9#jYF&a!A#VnDqMpzK|$W;yGArEb$(<|^z%gp9<r4-_uzM)JsERhJP{)<
zbjBtitrR9r*q5^CUtDGan5}2u)ex6o$grGgS)xEu3O=NM*`ulyK|48fPDV+;oax@8
zv<o3<AtIS76+ych7{X2T+S(95Op>U?yDQMsuegksZ~*3goQ#5N7_*Q8Bv96(2cMG6
zYKfoVU6<&M5r@(xgX?1NGHwQhIA|f-eMp%<JIwaXSRZ;?m$6?}v}LkNq@|YkEL<zI
zOL*udIOJMG(O|b6=FlfJC51!SnSq*Y%}}5hMZ654PwGKLIi~e!lyZc)jPK)pl3rZ1
zQoGqk@Rc|dv(tV=^Ws_e6M-OMna6!FGNuFvB`W3AVCIg+CBKgaqGnaqsme}A?(rXQ
z{Uo}Qr}acSl0TM2s@&!((*^vsZ=*w>-Y@hMUQ=W)fwi)Ivndj1$Xs&TUR-ja^wgEL
zq|3yRvLkSH(oMaGRSEbcRd{D-8xXlhy@Kto_g(-H6683e^wo>xwU2mX;KSQjN~@F0
z7xq;Ux0V>;5%M3GoM^;z-p!(pM3j?wCVcGj`adLoEb)GHmwQox7c3f#0u~f`w#P2h
z?{SnJ6Zhz>IHlfCIQsA?2o^p^6(DRU`OH3^YH48XtOs4#CLSRB5*@`IW6<XC%S>?R
z0Vja(E`;s*pqfN<y*i&da40-b89bu70PAYe9;I8e)wrx`7QG+;{47Ooa;<9ou;^wc
zqwMw*k`Yrvc<fF*d>lTQSS|Ma3fT&1g(Gr~aJj7c;qw6`shDz_2KF9L()~qGn99XV
z4d_D3A%@1_mES?1=khZ!?rJ&Ovl^J!DiV?K=>T&nq`z5WbmeP2M0idSN<s8GUz+jW
zEN=F*qi+a3a}^Nc+6$N}#9q8z%+^4ytbG}3$7N5=>^a^dngbA0c+Zar0(E9`)9Apc
z`8l^6RBfQ7F)>#DEDpO>cb-L)7=NU91Tg=2d0V_)T%MC{v`po-wd%UyTAkXh=P4bx
z$G6;{SzX#=$q{scC6SFH(O}g^*62P+pb5x-Y(}=d*owHtd*4=cB|2s-ZYD$U3CU-*
zsSqVCyq_sy$>@(lq5C7>Ea|KTx?GMb!?VRE>=JcS#k<@vwel=Z95Ws%-|Tv3c;-!{
zl-z1GFhD;?E<fs`(^5Qudla5-#$;P;Q-$_#C`;RC6m9hUz^U1O19ihC(B^S73!fyl
z(>;UzVijSX+Eu(%`Fg{^qx*gLco*`FFwv(d3UU)Jqg?7j%=GB#Hw`zOD|2Igp{c*;
zg~TssYaU(Dii?E5h_MP$ncRr!>(g(<P`ONISYU^{D$67v{T2&xdaBp9x=fP#kk`|n
zqg@xy(Nj_oU|er3uUpB*-V*<cAD5h%T!uOgF3Y+tTYf6Lqp;p(nfv9r#gjrNhD^$<
zUliIf8MrZn0hDpu(l%lViX$tI-O>|=U@Q|9J_S(`PmcW3kC_RLOs7?(3Mi(!(Z2UF
zEVfCckOY~`lT)TcZDoC|7Lk-9v}kQ55Ek8~JGkVQQYwQT;VI*ODiAhG(!N4Af^5`Q
zTn}^970j-hU(XHsW@V3Pv$3vpU9P4V&|z~Mkm{;f-wK`jJV$J)5b#BLDP$>T+j%~s
zNyOfHVcs!ZE&qMGt|n?!q<??Npk#D1e&3_Tabjtg96{8-{|)VwyRO@B`?YAjz08Kf
zqnxuV)=87pn_3m-8HKu_nPp7pdbX>)@biuqL*gzgC1%G{Saqu5MepLRS9S&uT-x@W
zJ=j~YFD9lmOG~YPSpM=d;~jtAI%;(axCYiD+~Q(kooZjFQIn|EpHkQ7z}NUlGs)jd
z=|+gn`hP27Ce{Bh$t_e4{a;Jj|37rS|E~|J!7)ZP!OC6Xe@|?)!jH>bFPy=s5oNBV
z|GrrMfn|87-Ye260>+GiN#c9Y-8C=*0#Vbg>E^oq8Uz9XDr993TW~h#HgWh{PS3uC
z%3V>hM*;!@7)U`=chX*m<80u?s;TEf32L)rB-OvQZK|w75vXnHt=6ar3O0Ptk9F-E
zu2k>k+0FgZP@RLsJkTj#tHgqo@p3M5U#Bhwl&j#B>)hsKHIVxo=#+dObv^pq;<Aq{
z!*&hAc=bTMccC802qQ~lA^l*S?_r!bP~`r(;eCw#55^#zlab5R^*~w`5D>s^-SBVY
z2YU;2!~<Z=jITNV6uZ^ikM+-w2Av>5E|C935^`1Gyz=wA5b<DLDF55?9jKsMufghr
zQpVlA;`l6QL-lWK^hLoy;9}y}+3LXwdBYwQOAe5@s*&OBEw=9bbPVXkAXzOWJ%waZ
zcx^0NiP@{@7JBFl!YUw1=gETa8QT+|3{Ff%f0x4Fd#8>Iq_Bj+r)wss4}lox?SqZ#
zXg459)pcm5#OX+R5C`dZfgrckedDD-5{trja7e!iO2)zgb&4x)lSsjqn{ISzr>o||
z`$LFBpceiP#L6h_nYP^o_un#IXHdXPwWy1D?*W#2bWGbkMJBUnYJ9^-FelmZG4>D$
zi6qW<B%Dt71|ebR_I`O?3Mlm2tAPT^2ndnPy}gs6W)*jdgz+hR07YvnD7d>I<>;{}
z|GyR3J)7b*mg0dThZ$K&Bn*|=>s1@#n3k3c?m&y%1<;M1`OZrp`YkWmgGO&o1zg|W
z0P68S!3e5D01NmWG%{S=gtUjuk)lzRP74?3*!p8`{nfa8Y#nDRM&$wAfZO{w;lsn1
z=nVE$hbpL8gzBZa_PAp3@Cl&qasWE(OOQbHzL3z6a)hZiZhg4mg#6cjV^ACcsaoi$
zU0uqhYeKKU+}wBtmF&<k1I-~0vPS)%shPdr#Befr4}sE%QWKo?<~2Cmj)O3gVGzgP
z1gGo-lAhk7G)AE!BDS-*ks(+xT*=<stvHalG}!?|4T<Ddfo2nu@{R#PARSP}0TDd$
zD@bRzQc#x)#ySGXf+qmvvwhg@Tbo3o*td)I-v2vo+ujEO=YYxz+}d&o3<|m+H2YRc
ztI+U*#6iP#0L@%SE=pBzh1<<SGHM`>Zh@$7XbeC-juV<O1(k%5fDPpN%|C@Pq(}vV
z@m2a2ppX6SPslYj9mZ{TxILHbx%_$j^1rr%lix=Hs!`(rK!BvthTsoyAi*7^wge^R
z>vpax6pO+i)i<O+W0N>kNcG-Uuq>qkbt{|OxXTTk2`^(56#BV(RaCD4Y!VBUfbnT(
zyE#puo=f62zXs5cqtmoce{c!#PqpFMRI~{ss|ProF1Q6Wc2M6~+Ii@GPaGhq0V?G6
z!4&L3$=lPTW?Oga3ed3@@#9sXOdiAec}g=6YT>38eXNKvY8Q)jf^@9ef;9?Qs=onr
z$c)8HWAOqaZFkVT14z>j<m8U)wa2y)jxyhsqL`(=Wb5|x5gy!WUI}^~1^<%<b8YXT
zIX&5tIOhf0bW>nzsGg=QSY&OE5Z#u?#l<Zf_zcN+q3g#+D}#5GHCQ2L!5~A11miU1
z{)G>9b!iYd15hSj0Rh~H1_lg-H@HmJMoJ;w<`qIJd*A_pyZ~?>p6?0fsjWE@r!UPq
z;?ex7a(`EgGJlB!7VTKm>ZASwGXOb2VGMw#Fa38wAPX{HwyoJ4STBRHrN%(wF@P=N
ztO9`IWGU^80}^5abZ7WQ4kWbxbLnUol@e}xYgfC&%O^bVMv89Z^O|)Sf=$62i0bF&
zlk6ZR^##N8Q|v-+ao6;#OV@_lvXHj>pF525tco(wsUaDUjOk9xB!l0{HyPVwrct+*
zZ&JoxoC=wRPoWw$&K3b{L!)R+*Utiu-plD%sDR9_Icv9qh~W}BS_TsEv?HwdKaHS5
zl?kNpu%;Wgku0lz`FSbCfzLY(oB%_r{osMWgve&@vphvH^AdDvfdd6A^=D=F3^ul(
zt)n7D?}FLLf|b7RUbr6+6IX2(zq%R6V}5$yPM7MyW77JL03#d-jZz<yiu%dl^;~wh
zJ}$o@$+=?oT@a{V24Y;|6JGz!?~`wW0nJfAOP1Q~E{o|nIOHXxcPmhOxbwY4!r++r
zoE?Z^kok{|={$CEfWYSM%5AcCSDskQV>((O36W;fuA)aCCo_`<Wl<7hB%^<6@yu+(
zjl&y^Ke2ES-r*fWKbwk>$fFfpQp?%wb*C@?X`(OEH2bhv6#JWI>z%<9UwhPh+hEk|
zCXqX=&7%*4=g>|}!nmr3r88fYtM;N`a<5!<9+{CP%5)Wkmj3d3!MV89nDm;!3QbD)
z7IXi0zRj_yVEn^*y5j0|GR82|Djjmcu-8SM$xGGcoOloK+1(#W{>at}8iQ1zWVj@y
z%^mPkdjXs|DX)f3acuqyqk`l$4M7jr3BmD-g!!2!1}IEf28_aQH$5b6XV}LNEV`C^
z=@v%blI}AaMwG6{CXV}D)z<LQB<B}_tb|!`2l1E!9Yi6<aO65WHVN;#%-7>&-q?`p
zF6U??MANR)>2i51QTW@wnKtoCc)@Q}GfX2kMS0KmzL3tqr|YI2%~}`_{?<k1yB`G@
zF<O|99FvIqK({&bJEXs3S7Vgaa4CaQ_bmEH$Hz!y_C3R1!Gt(Kf(a%)<{MLY5Y?t-
zEWO5XX#i&S3|E_6pirlqE~2FPWxRao?=zuCq+-R<KC*qS>j%FP!>S@K>!V1k9j8lj
zq;319Ygn)*&2-pJBR)_&TEf}brZN}>{rvVRUiygwtrcIUdGBR)#3q_94M<T&UW>Y?
zCJxniOgswg)xXNr3A2as%=`c(fo_$r2a^kV>sE>JVuT06Cn^V(nJ5&fG_tNb=XLeN
zt*E5!O1j{42`@K#8%@2*^Yg<)@BIn$3Pe=W1)jHl594z_UM9Um)?@ckl09Gj-IcH3
zcE6VDS*9#zST4Dpe&igysEe|c*|o~DApNa#k^Qv@87JktbPfV3NlD2lFS~X7bz!dq
zZy(=Pr|SglZ=Msg7U12r8uwZkKK9077n~+mi`~5FJZ=%vw}%b(=6;JG9`9rK-{|lQ
z5p6aWaVWq~eV+O|paoDt1bd$(4W8zVJ{#R(yq_xYa!$L_tML9W8U&M<!`KlCySK_J
zQ=*5gwHQ}fN^P=kc^fwHd6h)plegu^!_<l|W?>o;qd#I#S4JJiwB_=|Y?)W^8GruJ
z;VXL1=0n>Cqpx(-yC36Lrh-)AHEC+-CA!iDyVFqCLMnu?Zzvh%HWWb+C6{`kS|5JU
z7LsJfv-;f6bHqv~MUH540qV=AXrURAQv<YvcHP?z0iO#R5q1os`8Q8rc*xEZx|Y(+
z>Rl023;&i!LOuHOF8jvz+UN{Yf5P~PN7ea&Db{-QyDD4u{kOY3js4V&BoCwpeA<U%
zjst%!Sl`7Y(A?fFZSgiA7tGd;ll<^oBT*)l&3WF*BCc-9DRH2WFEv+;&Ft6lr>d*Z
z$fE<3w=$aTt&-O(9(Ue`5q<6+?zbel<5yeJMWM;Y69J#hSmMQmnwg(o`VqiAXickj
zxm$kZgSJVEM{<1dw~&{C$zPNgLo6v=Cbi40%Ep|YDtEkAZRgMU(HiD&_O6$GWW{vz
zqm0Ttq32x}&*`FE@yoI=h<llsT}jblBZ%61no;evRa<n6TK?&k*ddYZU*AGT_7Zga
zlF=+&cUme53=LKXTMwQ-87Qzc{=$W!Q6v@hP<wmqpI`B~8FTq!9@AY!l*0iDY@3)y
zCM+dSjYd1{?6Iaf9aTe|!>yRDdTK>mCG!)0rLfSjrHHS)7#G&in0L94@cmvq9q{Ej
zyVx|~vY*{Gyzlv9AnU#PuI9JJh%Y7Hvc)w5RG7Y9>de_)_x3G3n@rb-4L|w_R@{`$
zIqjyU3_lFC$D$j$M09Hf-9_KIn24J*anE2d-tk5{_ZB{PfBF``7OpqfW=2$7qcYSw
zL=w7R+ZX4=9NLQPNyvPIHA(+q$;25*b%nf^&I|co<X##n{w%J#hRSj*T<hX9a<yWD
zknoW$E_0??7L3b$G;eI^J?vIUc>JxkrVHkr={4BNniKjkep6aiO#emIuQwTpocF&c
zbW`6=T=RJPq*Q?|_9J5s^;H_y3|k^9d&VccR6*&(6LKwDq5Fwy{BslIK@abz?|2{_
zeHT&9R}_ufCdJadZQUKt;nZ~^b$CBQlixiL2n|DrEajr-UR?1hlJL)wQAv57Tz@(J
zW`RX_15bo%S$Oe&^XYZ&+VgeO1aCc3(S+`X?dh|cn#d4c;r(T`E9RWV*5L^XmmaL#
zV!s+y+qo;s0N|ef+d?||DOGhXv*z!9xvRIGx`};>St;&j_cs@{BE#oX3QMh!9e*YQ
zZ+z0lt)#6FV1=E1@hgU!10ND{nz%AX=rudv@{+cdeI2dvzUfDhMmFr|8+DZYJk@3Y
z^E($UqVVcx>!j{mX09_H(F@z8fi_vF+wR%TVlX0dnYNC{zSPlH@T3<3C=+7Wtne=H
z5dpZ8*cWGvpA9xT$vfbpqX9|j=tPj?^>u6Ef%e)AayPfh6&j+DyAq?y6R*0<ur|@B
z^p!l~DrYv^NY)m`X{M_N`By5rKglSVOGHr#zx@4KTH&yIildslWoMQ?sYu22WTqKD
z$z@)dGj3yBZlHKvnSXOe%6+T2W!A^cEgJbOWV*`<u_E8vD!P`UPi<a}<r@`#mE9dO
zx)S@cJFck1`G^G#fB8<$0WPKxo~Zc4yTL-B;fYgjGA-A4Jk?dEqbN^vAC*~-ylT4c
zXw$pZY<qKO=BaDnIfZjIA`7kj)2mw_X2tzlU|BnPIH7Nl@)CQKdq3T(fIE7}vgqDd
zHGd#sU07|Sl~(gInBx_QT>t3xc<*-Jfny!kJWAPfCuy={agkl(bV>DP=?mycWeP}P
zP-z3ok|-tx3CxJSqoY-EivMzD1#HJ12_}HASpfE&`La>lnH7U`5r~t6%!!a6^latE
zzkHsLO`1Z~1&DWq>JB=<ka~>`|E*r6XWy@CDDbd%v;RH(l6u_RPq6x9GW&B(J#qUV
zkMXhC|9Fh+{s-__#r}`S`1Bu-F?4z?n*1Muz`oc22n2loS0J$Qe>}veF*KvztFAnZ
z@jqu!CZ>W%CFO*L-1vY+Ml}dyi|VKmK5JHTva&F0>fsr5o=*AMJR!&5zpje~4ETjA
zwuUoc#d{A~mxhbYck<n8!V-ARYp1R308f=)d$c_1O8PGpK*_GKPeUo=SLzzsG?cGb
zb<pKj+gM+Z1Gr~$;4(G`ut%;=c|;g@ZG6Brk;TDOHP_#7ZWLMee-Wb}{pNCaA<g&9
zt_HHq{<+lEqgJ0KDmm8duEgL#hi1q~0~v^yzZRIycO*crqAB3+yL!IOef}K}lxqe_
zEGuhkS>@$eH?=GO7R>wrRd|nryKhwP4d`gsc=AJq1#tI5wJcO90PN#k$^$nZP>FzY
z4xGe|07(fke2rd_rQhWI|I8vNH&)*B2y(w$@K}&MsiLK)pC?Z{xC~ipAuYmCr4!H~
z%(7`$x!wb3`zc~O_e(lu*6BSKQ^}e71x`?j4CsI0jk@47_b>8eveB1dvg(1?#(pQs
z1YD5?kIsO#3i6mi)on`4K%PzuWJd&EMxo!Y$rsnw=0L5#u8UX&TxbpuMI5h{LsWXr
zv>w;rX)xIsw(DjgWSpwO)31JDitl%R?0e1)5!}GyMhYy_=R?BBD}YU(wz2b~RTbN#
zh|_f>u=CkT8U}Ol^?6<fF!zB;b_51gCtzK3otXI$aBceLhC3cGl+AC{`u{UOOxs`9
z%FUyw5%)19l7@TB{k!}7MZgi8>bcBuwr_VH3*3>(;6V|9*9YG4c!89kyGehw05&ub
ze{2AdB9BJeiTd}N`mH^348#z?q>-aC2ecv605LLbiMR>mB_uMWz+B&@pa1!vxd=Yd
z-2conT<|Tx1mBP3dplQholzsMfb~#4<Ol?zx45)q1W*uN&!sHL%1Y<*OU-(uBn0v*
zLa+n`vFO-VzJXc?e2h4Sh?Z-$2hH3?*26p#+p*>LBQJB%K*InEPM@`L&c8!DyJrrT
zg7C=5zW)Ax1(mf_I<J>S%YJzDQxL#{9@vIF%TOH-!8nRsm9O*jxp2N!RcWJ*uRzcS
zclG8)$aW9W@4lylzMa7O+zRY?t-uHONB;pN0$1zn>mM$N{TGcnJVmSOMz4bUp8%Rf
za1TkgM`oyo7wurICJ!)z6~N>>TxLUy1C_ZTbcu+G9YEPlCTR$&X*%~QKhdowawR1t
z9D86FL<7_UQWIR<{H%)nV?zX%zt<2WeFEIk)7$Ob<`dN(UU}>PPO?O~DoZOclqU!~
zvHpWPfHZ*wJdtOMkrf-Y#}#X3qe9p*r{O}QOHg48IQCyFE+7m(cWD~{?Ca`v(M=qv
z?>I;#5-P%zY`+n(cXR1f*fVZ;4ps?)$(4^X*?6k;xBm9Y^s$RJVZ0ogfinf{!6x92
zDGwAy7To6(%ntzY)Ur2d&@FM&C}F==I_L-t&W`=hG|^5ADv-SsOxgi3J5CDE0|==T
z8d?y^7lJc_C60?FX41}<(tNO#6n{bJ$%M&K0RG@E&;b|mADIGB5kRqHJydw%Rn<mq
zl2C;3;)j4AMJV=ANYdc3KWq;56l@>npoE7$qag(WoT}-NpOa7d0YL=>c!7@y*my4j
znR%h>#|CDFM*-x-z8&Nf#xc<zP|#Y-XTN&ZeSRpB42<QX$E(HAM?F|i)}{dD;*zPO
z6=(yCf+WRzESnIt!W5Yj;sDPwq<LXW^Ns?t9M_2Ba69h+5X^)NV|G=E|4y0676>}i
zSW^h{Yw+p@@~E-O(OE??&{RnL@Ed4-Ou;N$4;Q_8ob`R&ZL%{+e3ub&-$LeENKr*?
zp6uro0Sw`<KWpTue}0Ms#8w>82VMvPr|*N7(tVc%!T%0n#^AH44lh-@76{>jFgQ^F
zl7hV51XlZN8)9s!?o3d~gTyuk_F+C6VgM)wVN(ID^|rib%?9|zFP$8;NX!FtdLF{d
zAP?w`x`QitboMw<-awXcVStg~K(HM6P>G@T(zCsJbw*eC%iC;wz)2|Tha<!p;@~Xv
z3<XD-D~C=wVALB(X~V8~<9~ElI*zB3YR3H1d($L6n|^Cg80T|$!hH^8y6mG5w@5Tn
zZ7twJoCC0WL&;n4YgR;+SiUe6ll!;KI|Hy6GP3s`f2w@_MOhxjED*s-0T^d|dK`1k
zafkQxI_dOuq=Cs#VsEnJ3vuS`ak755H{)d%p1EGgmaH;#04qLGjeoz}i%-oTci)z|
zm#KwD%DT;U^RjHK1591?N-8Hs)4=Z$HVscnkCX=&6S&*b0$PF|x~XSP8%yFVDSF1!
zxM{#i^;lVfA5CyLHT=HXN3Aw&ULuU}sgiWRw5b<vWW$snet-=&l<f&WTu%T<lu2-c
zQMac?y8Sra9ou+3F7|y9qnkVOC?RtgIpF$)b-{E!_*T9%t|uJgz?t7U{r2ophLkzE
z>oxRx9It!H^kIPeH6R$6DTGn&cg;Q=?5zI&?rl+HGYZyCp<}An&FLe2z<KI$BVW%o
z$Vh*zCFYD~bO)aLvijju!6tfLi^y`4&4>C*WyMP;_ZoR^IguC=JbC|gOCkX+9`YXq
z09@jktKdyh(=F2`6$Pio)Q<xmqSz~YpT7U9pbbl3;&~yC_@2YJO_j&g81#xaut}zA
z0G`#WY<)8%Wb?ARW<=VLimqgywXIpHTTjXWX4@o07nq?paP}!~DL?EdeK|T@q#$V9
zYC8Iy^$L(a<fc2h>6_o{vpS{IOlx%VNiI(-Tr=G=8XtMtQNZV6G5hih7b}|d{bMD!
z<Y*v(DG6b7UXC9M?zllzCQGfD#G)pwd50@oVm)L1&QlwuPR!GAL{quERM(>|ifpW#
z({|K?{`0^T)d4i0yuGy*v6Cd$dzsZ(A2*ZA^}s>Y&+Od=5*@kD+Xp1Us7xxNzUuUH
zeCxcYL+i99lGGbyv_0%Egb)ObhT#q6qvla?Tq{m-cT~tHX?T}bFB99nZP${^lM#Y-
zp>W(uakEmf0^%5S_`}x?)QZMT8LmXm@%y^hw1}S3k|uE-Wr$?TL@)L0Ukg6I5N}KN
zrJ+|!+J1Ax<FV$`+a$32a8{V*+998jXo$m8Yo!k;G$m)g*#6DHFX*TKjF>0HxP&Ha
zzkl)g$+F9Dv1ncNb8TYfiD;4qTAf6dw`IpHMfp1gRJT(O`)_K(VWCej?X3bzVaYF&
zpE$r}Hk$;P@}+@nMwxB1fn0_F?P5~<;hBl_0KlJ&gRO@>J=iUdh^(~>hP3>{*{iwR
zmovIVdJ-gH7w3jPOT+cl-hT;e^;YW|X9Ds6oACbI(m?@p&TC=3b6!Ct_hDp+B{(TO
z5cQBo)Q@Mvvu%j-m9yjccIwN1`iQ*l+r(Gcl6+L$-3YzLmR>Ii6gl9Z-J>=eE(vH2
zuq$W`<aGarY8;^{<4Ji4iZsIt;VwTtjfB))^<n~Y@!Q104_+{XcteLRMpiAlKEolS
zRpW(UF+EH{14$0Dfw4cKqs&Fc+kg@)$H0tQHC$}-olh+cA-{{q15?ky$t?4>`6?!`
zobT@v_%a_u(N@Mj;u|JnAi&;oyUf4&Z6KZK$Jvxpo|5yOhZCu8<7)3es|jAp2Pz1P
zhT_Kmhq<>5tE!FKc2NmYBm{I3iXc*=u;>y+N;;)PYLOBO2-097jdTepDIzG1N(wAQ
zkW^YiN~K$RkK6ZEpS^$WAK#D9<KQ`bR9I`xdCz;!agFObN66#NCm-Ia+P##+(jM2!
zm$6{7diHdJ-)yYkyKKN<wt4-dw1$yKdZpYiN{$5)+;=Q#)%CDI7H`MFUQG4CV(>__
z$W41Ii5@%6R`=th_p{A#FR`;zv8%7~X6((McP~)B&U$&DqKY~kF0105&7e@MhW;d6
zLFGcTZ?O7dl*-xmTB)ETeC_VAh-&TsED)Qh^0C0~G?{<zTh-*C5xQ19)a(|9D;B<I
z&ZnKc3Q`IgUw2>>Goh35T-%}f!2ZT1=5}Mwo`9xJBt3Q0<8F0#CP!oIO7^UrM-Mw0
zk4T%C&MSLdPhL!GdV4)d!Gq;z{$*{Km&2<PUln2mn&`4eA~K%Zj&ZAAF+SCHZrQ1M
z`zOIbk%q)w_X}NgpV-A5odlAfMwEQ;t}Qm;IsDF2vjoe;ku~O2US@#5d#9Uxa-?P9
zOi@gP*}X&+xGr07NB22iEV4ggOLm|6+!l8d$Jpa_P<SrnzkX@8FZ-cSm_teL%ndL0
zj~L~*2`AT=TIyZN{Po>UTIf3o-Z6@p#O7&+p18_DZQ8Ub@wI=buV=<Ctkd`h3tw24
zTF%=Cq;Z<gq;(<uuRO(q&vq@=j8R_$Nv7rJyp^ZeWn)7tkHO}wN!6^z3uNVD-fU^_
zIKr=&h1Hk0((}F%&VfW_u<|J^RbWARyOHhX8io%8Sl-%B3l_!+3EMfLVcRSTpTYcK
z`|znI_i2vC+dAXpRUAE;YX((YMjz;conzm};mYXGd`>^)lB6yeX35dt_-KFl&G|M%
z&z+Ed-=vTM{|oB|cuXftoqpqI0}8>ocIuwq7{OjKq56A6PJ<Qg{S8<8O$<7v8)VF9
z+l*xw#${_SRmR=F(It8&v1zm2<UB38&z8ud_{9(=hBQgXbIyC|O^0J_>~0)6`YymS
z*7c|Dz#i;s99N|83i{_b3hQf^wpj~FnV+b1tlWNCc&#EJ<TQnBUn?FKcCQ|9?2FKx
z{*1wNv))*%2WwXTs^Z1#!pd_w(tT^h(sPFs_4^%o1~#7TS9ey(KEbwB98y+WRPZLf
zC~K<f`^&3vpK-Z{A8lIYI_JbqJTS!ci}HJ5^M&nOVncz)yW>@NLa^qN?CTbGfiF*=
z{U~}}P{FDq<|-#n&-}z{`Q>wp*0gl#KM3KscBkbGLY*q~d&0(Mr=NHZ`}nHy-ffB7
zx?;txu{B_U`$1Szt~r<R<n;Q4Ra39nQpNX~bFNIzr5A{sy<ax;9luEZ%T*$Uarq(7
zt!|e|>2XGUQtMz&BdytSm+sTUM!C2#G0%yn^~CKhb5XDneY{|?eMfNfdPxg)d|HvE
zvJcz7D~2*RGh46ZN}W^0T=7Uav0flL&|s}oYV(bSZ7G_&IYEEkQn6E|0YCoCr0%=U
zlaiJfDyG+P@pBnL{IO@Q7yT%sJukUH^i69o!;mM%^IrB`j+0?$<88C^OkS}K*GYKi
zj%K|*?@CJ@LaS1$LKa0UdR?KUNYbT&_U$(=7u{zX(YAN_EuJXZ)$4U|HwLTl;$J_#
z$yW4D(sHsQt<tAoTZ6{U*x><t+c&Yf^0DV%ZXkm$>i#l1)>EP9Q!h|i#-YQC14p<{
zZUaDoP-{=@RT36QJmfi7z2`$9xp0?a0aZZIM@8fo$=${9W$<<1e)U=F&x5^D_1b2F
zc~s9X{fQG6@PLbo4^)GrN5N}1D`Swiw*nf?2cDLd#huMBzX#94MNo(#Kfr5`)f)iv
zIY{6-+b^3k^&32e%*;?!{FVYTqE!3G_SpXCi5vbq<M4{T30V!mXUww*N)I0|Z%t>W
z(gtG0TpL<CHuq-gBs?gKw_Sh-Z&7f}Z;Vp-eh?qtP~9C@+kJqTHE#2C3=FuhZ(Y1r
z(wQ1SHF-7u_vdemc?v45fv05=di-&SLWR5-NH`c{+8t#oAN6PW)!QhEhobHy@_O%|
z=;8D9@aQeTk!WQpGplIn<QpnLcM|oPp}MHvNUH7|W7+=m6wmSsZ=@3W35p7YQDO96
z=}2yrlatc}0fT9)w^0Pn69Rc5BR=BS$ZEWHnbuJi`@tHSQBku1&-^X$RZ;fhZyR#U
zuQHJ}392}LtJd*r!<)y12X7s~W6-xCHV~9dZHPvNszcN#PeA|`q%R<t7u3VEpUnPB
zVv?4xjH70t89s${1fYM>05NUk;c+7*10b7*7of|`!EAt*56XymE!<}^hr+`1-G@J)
zuQ0rCTA@%SNoWyHVtj7?+aTyR!$G`R0Fe#UiN@pc<p7&W5OeuCbUy3PVV|j6^N57X
z#RApz=zzZjDBIp@_3m--toSy;!d!Fh__N&1FBbGfnIl0V(KdhdM|LV|v_hC<ZkFY5
z7um-WmI(ngfNMkO_}Bhfe9SLeth?^}^OI!Og#Q03ch$`ATxY)l+-q>)pVw*6zo+5r
zxpg9>V<BI?K7?0n-)c>P)gItco)V=l)7`_j7yrE6IMULFDK8V$V7lCAU}dAG@QCN|
z)=4Ci<w|xV+S?ah;~fT`^gl<Y!b35#lDL{nv=p?EEY;+X(VSW$J1boqV2GLOiDpHV
z9{rLVUrvVGe|@_uctZNmi|ty0LUx2L3a2KqNJw#Un-|{EZ}$&UHAGiGYl%W}zky(V
zP)vihNMp){`N4&ww(tIkZJIuR#kMcEM|X{%aU2RtbL$&pCPXBs#gz4mN2p{4`N$u-
z=qYaBNWuLBs{^s}ptXL&Pom8j`<v18dvUy?*s_SUTe6J`zp#b@w>^Q|ym%mXim~-g
z>oQv&oNi=sF!Y=)@!IT4okU1A#NYx44bwoW{ecnGLI>@L;8YMx2o?olympsUvpcM}
z^IvfMS$wZ*eG*vt@lbeD+=8qP7K{RZCC8>1)rWGHz~;h)!WeJn-cGx1{>a|vwops5
z50#$!vHK<j8&Z5WZMt6>Q`CqLEt0#=^wvVU0Hv5vmF@=W^QX1hfu`l{JJ+xNIV$e7
zYt6Qi?(_OpUZNm`4v#G|+^mRr_KdlZdknx@=AhGf&Dt7|p_;eH*!KcrAt7N7V02Yg
zl1ma(p!Tspr~&m*ztoPd$nw*-@9E)_r<pbm!cUt*87a8@dX)re|A-p8xN7A{{c9<g
zgh6!NSs!jglts{~t^mX;*m#@eckZLjbLc?L=NzpKL}F@hr$jKb%F0SF1X)ZAKA)DG
z8w?^tCosVfVyo1C;QQR%<Z|ks#DN0`WIBRBB~6KT_qTff-4v!9eG&@#ifyRT>2Y6b
zV&BNE-aL!0WCEQ$1KCW?QNY{HOROV72SOge)fy%k0dE&C++Kc*ryH<!Hq+`F)*>|m
ztrFPo2;v28k7>{#*;UL+gXVuoPUtKyiIu2X+E=Xec+P-J$`W*iOYUra0E2pu#8Oba
zHee3}GnC#`ZX_&UwJ8_yya0@|(7EwrQJZ8E+Hl;LCr2x%z@0M*>f*ldjJ`|v_B3c*
z@}2)?KwcSdfp;;}$__*lZ*(6-%MJv#4m%-H5M+DH9Zw@s&bB8X-OUKL7n%032&y?Y
zHkOR51Q3`RXt`*IVpR}9xi182L>51EKJ3fB;t_rl{`~F|p5V{YT-w5=Dn|mSbE%fX
z=e4sOSlqY`iUUSRO&tLFm8G0;IWY?)6ckKrO%~TLb2vc&XTQd?Em6=`tnnH^L6L(>
zAG9?B{n!GC<K%-SZ+As)FS?nh`bo@LT|{IYcEA2t)bw?L*IQj*=La&?@0=Rg5FvX-
z2NL(ryzfQfA-B`;gTV0x7cB&^7@a;aOTc@E3|!~eub<$C>JHb3cw??kF0vBMJb8Ak
z5xeS<3=IV*kn5bl)#3~g%h}BSziQatXUPPxuq}Zx=n0s?5RWaA)wK1AXiK7CFucR1
zFRy%PrwpfN3c6IqL&0(r2fN-$KlD7H#8bIKJB+e7h7giNO&L~&rE9OE{%nHVYco*_
ze15D7NUB2z2VGC#p~S&?s93L@x43;fy0%sdZg>PpvV1d{XbWt>(*)0ZrfY(ak03l4
z`W4}Q4Z0-@1GnME2b%*P%ziDkBup*n;Q(3>btriUR|K9j`^o<3JI^F`e~lThd0hAV
zlOQzTd#%Ral*VE9<;7KWKLm$XEZp9nZoe2uv8UE$BM;%0p5T0w@9!hx1LUmtsS~Q1
zdTsFE{XC^S&e^TUB}9Cxv5So8VF69FRh>Q=`Tc!~vWh>QL6xGVhAw69hwp*5^>~+*
z`MWHH2KcV`QJU(6;q35r#2+JnO5FupaL!m<BLga*h6X6cYpDD8^vUoTZj1e$B?FR2
zh>mDc67gc3PNl!JDpTpz48b<8-Wq&?9bAfe5U9ac<;5VZYt=6Op7E{QDTZ`?bI4`z
z%HTU}&ngR&N)jI0uX2dd;u_I@Vkj<^14^CKuql=ryNs7jSkxTxa}ffe1)%g$fWVa@
z*+`-$wc%-9_Re-)ihrDNKQ{K#_s=|sQz-Wrl(34tRgoF;sbY_|56z}?V$7OZehlHH
zpwKto+X^Q=diIt7SYE@3`w9=qeiG{K?GLPN_cWpP9Mp{G3vO7yuF}@!+-piUpMQR`
z(KJ9_UWhMSGBnETc;lAiw~Uc}3`K0i_%*u+^)if~#Sa8CQ+~JymI24qMQvMg@x;?B
zY%XkVl`jd0Ps&eR#IZG+K0UgB)#Ssi&-ZGHIs85oCKB>3f1`Vna_Lx*(>VQW{Ns^;
zkUF|%w{DigFoS5QSo%n6<Hjn#S`k&vFzt_U5wYd0V2}5sm?#ekN&k6tNxisH4e~w-
zJOO#rjQr|VmQHp8-m2#3j!DOb1{P$@t~=xJJsLcoyeZOka$YvfK0Nz_*r9pH?#S|h
z;txLpZCeyK>DJEezdLI6Vx@{zQbH|;uA8?^vIdWL^{@-Q{=-dFUb+S&{iz|Ol<i8e
z=}*UpLBToY@t-G>`IJkqf5HE9O%4%U=|9w}x9jb7y7BfWi-o@PSrr*%MLdbJJ`+(A
zQFG(PDGhpLho5&&98jxI8SNHn?JXR(dq81v?GdH*Lq@W4g0M?eAFEv1-qUv_Nb^{P
zxcBK-Wqlov+Ucab_-fH^q+aC3Easqkohd=tfI?4HLGJR&SvNh(-iT0Fe7kywi3MLA
zE^)!+%&q3{IGJuonR!G1B}XncGj1Af6*_751-n%h{~%JiWm3)5C^^xq)E&0eg(HEX
znTm8;cDyu~EWg)pq@A$3BfN2qN<wz)xYfBVUeDrW{!*>ArqaW+8*#HjJV%d5Y~0g{
zaOoc;QA;^)!R-vjXFlV{TWS2Q&t}Anu5Gz=KWuf}&m8T1lW$HnpDL@;^}$+i3U^uU
z?IGOKYI4<$M!%*_m0iOTn}<&X+!ATp=avg!sfat!9nxyeEZ_04C9BJG^sy0gzQp%1
z(_N=~>|zhrZNPFjM=XZKzq{zEEn8q{z(9Hfy;gpv8~$>?&Dh{^JA1N%60D4x=2&^k
zM$2nK{FsddUqM2mQv8YaqGn+mnsds$VCxwVv3T08^hsugGU7r39sm5J_!P;k6;<A9
zY>XU7{EeeOhaS^m`1et)^geLV%{-Wr7%1kAyZ4OlP$^A1cr+ZS6RQlbbv=&Stf01f
zS->j9f(^qA^a=k`&wrH%9t0=0%R2ejtG=(~@W=LwI8A@VEtyrirdmZgbK8eY+j0j6
z))Y3naox-Q;Vh>xut?8>%XzaDiJ@hP8_V1&=PxZ{CuIgNi(Cf_w&Ub7M*-E<YjZK0
zsak;_cT`8ct*&D$Fhwl7VcAUu^!mDZ$^!+C$rlXk$^t^BTk@KOR-bK)I#Lz5unKWC
z+H<V$uY^@3|CP<Wsxp@Fd*}@U<zG%AvfWez0f|bGow+dipcXU(yRxd8mgcp#Rs*j3
z;vh1n{ehw1on2VQFe>cdU)^>wiu!G&=zqFU1(onu$(?+;>Rjy5_1H{H>MWV-tXeGo
zomd0ObSkIK=LM%uX4+Z?sU2~?W4Utmg&nzc1#dxc$L;;LWZimB#`xq*PtWE%U9vy)
zZi$%v%kjcKy%a~fVW#1lh>`<zLPUE<%P@9fSizr@!^FZP_pQKEQ|cAX&9zQ}t7=de
zlN)Uwb~ff!pU(Vw^rXGrWaz3-O32xlcZl||1)-m(X84W>g*s%HZQU}!w@)9SW%>|e
z(Gr8h9!ZI9qa&r8525(R%kL-b@B2bPuZ!M7%0;YEPOI;jX6Y6wZ1G?ODb9Gb(WHGW
z$>9(+CRZ89%VcHnS%pf1L9UF?UXh}RYsXIY5hh&IPU4}1Z4R+?$TJ6)LlQXmqMd}=
zc7~oGVK0*3d?rgFrT#AD|NV@TRghIhC53&-)Q@pPuJuf!Be?#vljhrkBHZVrozE5K
z7g^3oc86e-cU(p<B(w#-Opho79WOxZtxD$*b=<Oq(1g)>zE}c*%dL+c-B=`Ue>BC;
zidgh-y(}cd*+lu<V0-DOLsqTijX7|^!SMrvmtN(%Oae;dmtydfdO3lhV;4=uZSU-|
zFI%OwZ2BtbG42pCuS>XCy$=s^TpH=?qqn$`ne?eoH3U0e1@0~L+WU72;kpn|ejpfI
zZ0+)`WnA)k2q}?&NN@S^ffpBUN^dz!{d7)S8$LT5Pg}D7N_HW6{kFwtPweES?9ijf
zM@a!ac45<^@C3{Bg6#K~kFs|3o9UDmEJWGXk3^*&XI~$ic;fv7tCG!~(w^_!XXR6d
zk)Xrf8CFvZ%dXf-uwCDg9UAiewYJORI<d>YP88vsJl5-)N;9MtXa=Cceto5bJacyn
zkzB2osoD^OEm)Y4u%^|2lii=^aMy{)te|(z3Rg5x)m?g#<oGPMddrYN%rmupsO!O+
zm=@tMXY=?kpPXWji@w^*6^#;SFB1$H-5IctE9r@L9A+t?+KD-jFJ#M;m=5VV5lVH)
znD?D6eOFq(>W<gK1fy`QWvphP8Syoo#6=??rB~uPn!%AvMFvBxjuszn1d>PkCS9|>
zOgYv~9w^VaE+;(2JDL1Cc*|Fy_1*NNC3*hvv5ktB0g11JtuGZX?qmg5ydA*Qbezy_
zy8MP+#amTp9Qt1`f)g|Y9pBCF+yAU6*1>Le-1wc5UAnPFsw$K*|J&%M+z)1qpAy86
zr$jZ1)r9Z<c?7FCv~}CxnvbsZ*691V3&b4%3cDIfuhIW470;G4rT_5)z~h##(qDoA
zRp0;D%krmAHUB?sm|0_>xqRxu>;HOWB9QB4$ddrU8`hj`{=D*dNzj_>zBQum*f)4Y
zWqzq&<PSc5WETJ|zKRBo$kdt>>E9>n7%1K^3czEdV2@!PR1WYDP}>{T{m6Ac2$ugn
zi+gqu<&TJn`0ag}zFG8F7+E^17`a2S1CW7nJj2a#<aTKYqP^rU*GF{Rum*_1%^+li
ztyejh_9+3?K?TKOs+(E<)?rG>hNEm0?Yll5cJ5kf2zWX8BqRpixfPCXZf-&tpX65w
zDW4xoVgMq1cCi;+CZc=W%f3;ve;<E!A<2}xkQ<3(5VAA*`8h{9f$w2i#l*Saq8qfZ
zC-!!#_ZksO8blLhP1$Y6t*@<t&uZ)1*>`_d!gkN?JI%VLOZhQKSwIx|Hz11ru8|Sp
zfB<pR5#~TuWR$Q9xcIm5bXF#L6Z5-s^Pq|Ow2QWy7a}6^?_Wm7(X0b-Bxz(93mc^m
zdd)efjsd!tLQQgmwM}x18$=8A7@lGC5fy+xpQno~x9}s&&u`eq?;j6Ffu%pFg_;zr
zw7(h<l&-5T&6NA`7)0MdrxY+3h;Je?C3Yw6iPuoZ>U7U?SL$xG?W@XID2GdzpMX5J
zfFe8)pD6^z%Hhq#u|n=Tj5H{EOb`(IK0Q4K+yztUvv1cjZPyy*l>Ip@6@@!LmHBXF
z$QBWI-IWe<VVB%BcACa}A&&DNVk(kDU@LM40>I)*m)Zf8;MxBgj5USa2S!kP98SL_
z3{W6=;?k`@(_3&Kfh%M*K5yKZNZsj6bp#Ly*WYVyNvza3Sz)r$3gRWO2ED=?Zvv<u
z$P_?YN5&^qZJ{6#`WXm-fT;WuAo&1#0T#=L<tG3?c|y4#y;W32frQtRBpg2A@Gdk{
zDQ?b@tsjUc#;&f3e?ObgjII?(4?1_u%zlfv7R|>95<jsSZ+6DrYW*z*q3S!_Yy9*!
zvO3+m^%&)IhK7ug^oNy{RHczcegN<iErDE6WIrH80B9Nm$*sP&-H+=!|L_j0w$+M}
z_F(E!^WE%{0FK6Q9t47U5FitiW)33s-%0UA)+?xNccC<=@}uzmRfj<3UJ$Vv2ar6s
zu95p0B+j<N!Kei>j=*`U^-R!OZq)Lhb+<;kr48<ey~4)dM2@*Z(rBy5uVBwY+BG_a
zXhT45L)&idFv(48P?b|!r-6Z?|HkPIfGorbAjl+yqI{qb)Oog{QWB<q9B3Ea9e4Tg
z=T*knn<ZmdleWbsG7wb+EIojHs*DAV2o4_d3^52hjHAgJkYlkMs-W{j#QaI%%3FAO
zRXD#C2CV^6QDB)yv64c-jW_jIJ$v~|rxNl_AcKw`2m<KOpT5%mQ-^O?^=C(#L!J+*
zvolYqVzW3=qy<prFuNduQ!DFn7H6p3ecH1+ufbXay97y15bfE8ArPPvTNh)w_20);
zZ(9HW`{1Qm3*_uO`>G-mK<gJLkc|jLEg;eF-7(VrpOTu2YU5KMydKCfEh!h0P;&wC
zg20aip(}TY^R=PqSgZy57y>E@fCWdhKPV)Bp~T_sJ9mnJjsYZ=+SWyg;UFrzxYhFp
zN99T*D<H|E`3-}<dpn3N;+a$WS4&AMR6Eo3vczk9j42y(L5RM=st^y!DEJI!NOpse
zV}odqsOJoVwZq|@*qBNs83{U@Khmkh<K3_xFMj7dx<)~OkW`4=?w&?N1b+Xj_J(yq
zc1q#sd2nPQqaN}k6<RbSO$h00z^t)=)eUFx_dbV2j^EvjpQ~a|4=?=Lt!Hj3_mc2m
zc29AJtllk%(>a5f0?hxOTM&*au<2quU%8?wx$=y{u?5>WQt8{*uxm^Kb5}%>TW1rW
z<!pGHKxz39|I)o!LAx;SfjoA>6sGKJeF6pdJ+w`q3(5&9>ggplG0E}A$6YhJb$3rd
z#s4y3IRjsZY#w7s5|58qS;T+5t9^fBc!%<0{{|D|MGNMx>&Hp@#E$h+@$pHJ#2-q3
zX2P2xf_sf2_Ku<}m6I}glsYne=Yq-%%b|I?Cm0LWHiBi_t{)||^pZGdv3hGssXN*Y
zbiuW+-xNgMy}`pDDH#@$wKyq&C+elS=3^$;rx{XcV@4vLl+HSfOBGn9Brp6rPMC6D
z&h7T?(-AK9c&tVS%@BWHzqFb*j{7E`;N0Xdb7E)jkh+Bj38zj|LLZ#!B!-N$Z6aL{
z2fl3GA(NNZolJK$u-Kw>BcTlO1;W-m?vY;0yF}eO=hQk_YdbScu3=RhWIn632t=t<
zLYid1^A){F`kzC~QCt#gxiiy{?-k*~Q<>87(!LED(;qhRwiPE~-KmJNX=#&iT#(QR
z{3V4klgGAFpKdF7?a9`tW4Rc|j9c8=I(Qn03=a3#!)c4XRPSc%XH?{grG~c*UmnhG
zy(__fZO8lkRvhPGltbDCf2rUjfM=iiG5NBer4<*FKX*)RKK?yz6GwQ6eFl5};@~e;
z)r7YX>aO#g$E0L8_$kC6Wh4++Ru$BXEhFo)DM&u_Wkl4ipEAav71yRR^@bzO!qmv%
zmWe<uzie_|!5Q+AyLqieU&o8@S&3xO5N@hmK0PjxL=m;|<bw!#$kvtDcARnw-sCm0
zlc%7l4Vz1CV_wvzXSVWwdq2YK-g#Y99}D_VkBcraSS;6{3ax19<O~*RP_c3~b?<gB
zjVf#qex4N1eDN-StV2{g8Pw|q&BFTP#VP5V<Yrd|k}jF#p6flS>7h~|O<B-PBcT;i
zkdCo>X}vU_f9`g}RZu9-rHhL`_2dG>$M32@Iu8qRi=)TKURpfuSJKrsmEyY;u7X)n
zao8C+9k}=1cb@8OtH_-1n6!xq(^TK$EKP0HaUz=t(-8ND2fi0_I<F+P{GfZvp6_Vp
za{iJ3A%bb;VELk1s5hX3y3#!|{RQW?oW5r~wYtI6b8FjOc%gffCGnX_^b=Id#M&J{
zXB=y4C8lTqx>TiO^2@Kf&b9{_X;v4lozITR249>p_^_YLPFR$}O59!Ygqe{Ay_#@Y
z5zg!Z#CafuWA6DGf+CWHZ=<;QZFfG~U8a)YDR#wvW=#KP*2$86@FbH2^hq2FHG|Kx
z3Vr)kMZ_JpDRQcwF}n6Czg-a#!{zi4XW9F`dXsNXZmK9Vk{uayFT|Mc)C&oPIatY6
zh9`XSAx?8(^TK!Ej=x(TcV<zW%>Vc{;pr+_dZ#A^jNd(eNe@fEqR_rP^6*1p-?44p
zVSH1Ea@YrM@_1hV(1<Cz#Lte5TG|Vjy57J4tVKbTj*ob7iQr_eUP=*N9WxGB72M%<
zk0@rJ8MpWjAp`w21#i{Uwxgu=-5u0D3xt;s)(fTmG_z#a`d&&4o+j_NQTBPJQ2vK=
z&BSA#oz3CP77OLnNIuxyCYNGnZ5EPKa&rIO-Hc$Jh*8p&1{V**P;MGMvVkrtOYKd}
zt?RcF;u+h*(rQg4@kO`+8%s_C$%^g#FR$dz<%7tOl7W}H&L2;;I>NisoAHuA)`(i;
z!KDU0*N0Cc9)Ek?QYuKk62?5%CxyTN>60DiYY^7oGM29>H<d0%@@2uGL<t^z&L%Cj
zkN6fFu5I*Jf12<lTVkSa<3BfN>Lc-xD693nlh77%xSE}5bML7OG4yqx91j`KXgJ_N
z4S*BamFrEws3Gn*royu`v~X`#LOxif0TUi(pLQ?>bhX}i2^zx_C-6*+l@I$nw0c+T
zH0JufPTeUtwHy$WA18G#y0@2FDNa$8`<!}jFaL#e>sL>;?4s<VhzG0>PWivaJU*dO
z6KLIHE~AYZ3n6FO_BS>6tSXn_Ckslu9}w`1_CT5$4Vjnzy~mh)AIwf2E#&#CFn_1o
zDtDMk*hUJ&_Wa<7shI54smYaD=NGLPvQ<~Jb0r0>1>T(~7oK}Y(_`JP5GIh<@(Mp*
zGD<}8R$JY-`&w$`5xTNh#_F$`1<i=Z=&N7d?lL~BPU=N&pHOdJmQcXT%b_D*xk5z5
z#%5rnebWA8x?)S|^AD2-6w*=&oIl9r61h1oIQso+@Yb2}`A!5gPZdpYY2J-z<*v%T
zIpT&%$#qFr&nSMRky!h>>bs_u8KH`plr_%d8{fwB_iLIIM~o6*D)0^2GG{U}W<Ht^
zHPZ&3Tw07!lqE#(g1lXIwgT*wt901<%RD@!SweG<lyE$yCn|rtJL6nz`c%{WcIm^6
zMGkqvLp~4Jv`=%+ez~Yib2s#$z0Cb6wsh*-@nj4)Jo_}}ES?(3oK>NbFUK>3yU-zW
zv5zWbEioW3lV&S=j6RI-!PeRrtxr+?J$t$MeH+hPQa+wK7s#m3t`OYB+axf;A!~i^
z-5B4s9Z9Z;a376p&(&A79|h!{nQ$v+bE@yw4D@2llH>hoPA4CEMPCRnd-A*IvGkwb
z*0$%ugap_NSwoeuFEtOn5b3AHQ>AO2TlwWqU{Jw5ucT_3Ptn*q*s$mA!zjGLut6av
zy4HP%DsPuIsruT8eCgfBF{zZZ^Onls!JR3}EPr$ZLn$+Qw)9A76B+4P=eWf$^&9j>
zH)4qy7Ly*--Q+0jm^Za`<nf`2ylrS^n=mZ#Fi=eXF!{jqrHbL^uAOU+#mv3yGnyP?
zg`ZZg9^$sY_(bc*c~OQX&z>>%nT9f|L$XN<{5sk3m^bBHA3tZGQ|o0G>hb8)I+1aV
zSh=ON>zid0S9pfgLbP1^_b|noC6Zo|@}vi2nxr?(D?J7VQ*M}Rr_rg2!g5(Mt@JgE
zw(L?ro_(iftiC)#LF&Uu@Ek5~(`wMxyHJF%-aJ}*zifxx?Ex;)i9%lGE-HV*Ey=F0
zVV3U#zo+2~OcNDML;3u(sr}hG3Pry+a(t3<ID776`WN~edWX2o<P9yQT`0cB?7JtD
z)}a*TtVQTpe<|+@=krKcYC^{yclARrG6o54*rhn1cP6%A1;5r>M_F(y5MTY^e~WUs
zJ9Xl|X<+mkW<tp6vCY%dLK*gzDerFAA1+mHA|4;(Qaj>s#>NoV6mOVOfpJFkd;aq&
zj+lfSW?pHueu+ou4mGck$R6!)Fl*YfP?S2-CQ^KcAwGh>p>c=!*vVlP?Ep8|>*B|!
zjl39^E>{Jo8B1|K%smm(o$;0Mk&tqR#u>~r>XmvyS9zlS3=!I^xt6#$!J>VOjIRuA
z!c1yv=>2<tu18UvZMhPfrqBMoKOosOrCp8&BtL5V;%S<FSXu_nQs$~q44G2MEobv+
zr6%H;=d1jfXzRs7#%PR!fhZYXu=H2uOddH^16GtrHaJO^Q&#i^R<c7;$6ed*L$?9X
zyXGCnA%TQF9_x(LTHpE<Pgg6-NEr6DW?dFOY>-7|N@_2o!yccV^eg#XJ5x-b(d@*B
z8CJ_a0Z*6utmidZ1{Bv0doG4Z3%<O#mtH<}vr94R_Jtd!m$dk#ZCM^=GX`nIRNc56
zY208Ps*}hY#;b@cu+A*#Hp?kfBh|WCO;mmP<Kn@!>lDY7F!JBM*Sma78BNz$gV&7N
zTdRb8efvU$;=LQc?(K1e@%EFIy^^B&Mt3}urXci@udDAxZ{Jky*>}3@J#;l@3SuYW
z1HLaGe?oJ@VXUf(?Dj+Vx48#aw3poQrk5=VZM5c_*MGTt<Zg@(A4!a)3zCYVJAzC2
zS<<Nxlor%?GEPr-@Z+VD0F6XV^%D*~d)hqtT9wA>aBY`jqv*x0U7VF~CH-2H*J71i
z^(9!?+j5fJnB}u=9>|fCd>6b^H#O2QWwSS}%Vqpgg|Y$rDTkX{^EBqi_;+;$7BGN6
zd{(Ylq=41867A-3(r`G9un)M#Fz>#qCl}wW-82guyVb34wTMImXrpU#(D-d?H$mUu
z@t5~aJV65G<%7%*e*fp!^i~pC<SY}jD~w^n<ykC^H}TL39`k1oTNc&_r^TOtxyb%2
zDTHR0!e1p(;D^Nk0qJAXzkeAJ_jtqq?@y%sX0LY7{Qu7X^)*E~3y?5;(u4oL{e#R#
z2S_7#i1J`qCYo0@7JJR%8N|5!eArUkil8Kq^K!X!r@HkF?Y|GjGg;9=<aYLTcUtr&
zQ#Df~*1-1d+qW1|C)TfDzs?N4i7G8E-E$d65Kt5|+KX)X&o{q{AZ#M9$tC(gP}~i8
z5oA|G1N_jiF&K>{4h8PRv5HO3T?BdokjrYMJ^WBK9INeo*`M>l{6K`4@9Pr5__BM|
z?&_<(@HhCEB{adZ?|CzdilR1_r<TE`(b>~Oa*0NEZ>)-eo*pCaI-^u2CMISS=ke$4
zQ(g|xWs@q@rE0_yBq;l%${XdTP8~E_O4-rE!jwFp!anZ&X(w1r1@6n1r8kFLAwu#9
zI~mGR^`EzL=_u>gX$_hqB#p)SsV8w{Fx1LZugu}=t;;^P=N2C2#7j#{BXli+d67QI
zKS=P~=!~n9@%7y^en<cBuznMK{KVYXeD67r?K$K@f#Mq(xSqi5vEKs{lZ0<tetxdO
z)2!tR6)UezCo4i$DrO!{mk*b-{WKA0tg58cyqCCb{O=Q1II><@lLr$B=$L1%t*w!W
zfm;12Ds?(c<er^SGvv=0AQd9d74zOeiLvp4=?Ab#NbBkzN2xl|+2XF=3U_|QJL&`b
zTojv1<bST9`lQ~0^H+|^QYq$)=0x+L_qC8b@Aarl?P79M6nha&9|XYgdoVA>gR1bq
zFp%H45xCU!Y_WKgkWrB8YL_cUz2VvlMCukWR<{Ln+5cXA%)5lXMWRhe9t8M@&PZmi
z5<RZn0BEf;lBrBXMmgHZjQ3l{$d+8?gz0cJ2un-Qoqzps4WwymA(YlzVtZV@O@9qc
zM_`ro@<ZXl8YZWh21o9Fe1ETnfU<z<rU>(@d5|dRV|k2Q!C`^LQ=;{ajv!JQ=Dxm5
z09y>`Nn=1s#lc`!3=VVAXM4d1;f+D#k_La0T8V)LSinL~h{po0W*M36K<@2mZ?7R`
ze^LWsrMnkdmwYoLZv5xmfBJEgFLKQNr2GjNn$XZtWb?Rm!MX%w2qbSFJ7-Xj<YyGC
zy0AW2w=~gAC+xtOsTl}4xA(lPVYtX9Rt=g(i18ExshI#d3qcr>7fYT*i6gK=ASV`k
zTytf_)nor&)z`l|@RUA{Y`r7dw&~lk2c*>KG@+=OI@5Jm!0KSl>JOr6oSi=eK7ING
zQqVGZnD*ny(7a0Ir$ZW%fh|PT7A3;FyLAxBhCp(Alwujti68QK-35=NY2F#Ky3w&i
z|2fPl@Unfo61YrF8xg%Pot^LR5gkT;ucoPGx0FUOb=lmwfwHEP$ZUlWK9E8~Hmw6A
z;1)97{<Y|}(QG`4#Bk&ogOh=)51;~9%u&L9VOpur-uf0NXdY^GG+$Nx&x^h;GgK|)
z2;#+8m=B=f)t&0C>mMf|d)i_OVKO9)i8@a{ZZg+{$RG+NqadU|g9FOcA?@7~3^)Wx
z^G!)P309!@Fv1WIzMo!5ZX$&d?mx&^Cm<*|`gU#syH3*lZvod|8@^`A+P6z(wspW9
zh6L|7ffE8G=ZhmK1P!lIW6+G!lSl{!3^NkGfMP^A(!$@y8l#9fJ;-|mx;S#wpeQM9
zW|)*hchFZ~gNTO@JJAfsjI{3YSop==BhO3z?M=UI-MWaIc10TFG=so<gDkm7D21uV
zH(#Tr1%x4l1wzX!9*aXlC2AMIi_6%4f^*Uf*+ww-rNgEQ?w@FqlWu-U9z|Doc8W((
z8w^@zgt3ivXEkFL&zyhzodqWFnIlmyOmq@+j9YI(?g{bVGjMQG_g0NFv9MjQAJ!kT
zZTnZL43Yl~c|?J_G@dH3({9iV*5A{J_KX}?^b8E7bqlo4UI>VV1Vj-0oj{=D_1*Pa
z+K@ndB(i`ZAsJZ^o6I9&ChsD&FDU(B6XEq(v7W0~Rz>=Cp>+pKtHiQ0Y(|%@tfrbj
z+UK1~`m_{wr{X^PFwjt{n)LY3J-zIhGYl{^bUby_5Nv;7V8Znm+aM(rMq~C547mRq
zeMI*^^HaS(k?c(hS2rQdFj%(|gm0cy<Sg#4ca*3Y4hthEfPei4G+@X;!h16fVNOu~
z-Lim<3|LIfNRbBb?gyCYh_<M11NhYoK0ah)1*#m;&-6}%Yu7pqsAZ|T`$sj*x9&S5
z`Se?T;Hx*PzN8yoDM;^Rv|{fFkWUTk`;M2e-S3*~$3j)v%}(r}cIbY|(PyCv5ht3j
zL^0=_aE*VgQo8?8zxF-Ln5XazOxeXiW3x_wg)`3K8tzp7<J%XuMPzL5YY?YmC{)Bx
zk_r^;f5>vvTka=oJ|B(!rQjuX#bqyd3Vq>*c<r<4o#!%C?$yYxJRnjJ(qXR?dG%0^
zUF8h<_vOQ%(;h31<7h*i7_j}B!yH`njP~vQ?UNH%-YC31d;fu>j1pfX4ZA$ad4(~J
zb}BiW{z^8(vDeuWSD2rx2Ol7BRyE1$QtqaY)~rAD-g^~uU;drT^ZoluB{i`xMmX*h
zsjDBdwQIRk(8cuh#(}I%G0d|_iCH;`F`1d-t84WZ#C9=XO;6?tz@xl9-xJMc>jsPV
zn(p#|bTS5&n#W_jLPF!WIG)EpaQ#kxSlj-COcbtWldRVdo+&RW2P;H%os^batT-Ir
zCnco)whu<`1W3RAcz`kbL_@hjBDPg^`-DZQz3t>)&LQy^m1IN~@RuRYAtpXY=#N)Y
zC-!&5NIoG^B4aBg(@PyPfC{7d4864@U&wPd&gdvRXTy}Nnz?=gAI3<RsM~A?O<Zvj
z%m=#JGa2!X<b8Tr&d;~~+n>Ftw>Ha-|15BYg_iGg>U+6F=Gk^CnT$Y=2J&OF;pzI>
z&17^+3JnrW<efKji>Cz)I1On+`zK7k$_Pu|Hzm!b3FBiN7r=JdJndlZ)0;jsTF#hw
zvW`acHV4Io+f$Vtdg+c6$Dqy;@US7^Q;p0}ScvA*)TVN0c+JkOemd=wdBAIJQU%UO
zUWT~Yn!_246q&SlqfCS)&#;h(^D|x#7Udx0OeCx(Ute0O(IMVBYRae)>J{YkFgody
ztqSF%&)@v2ZtP1Jq;{GOxgyj)lF8|R)s@DBQ&XC*&O~oAUS2UoWIlS-_5CKTQTc}O
z_KV_}%i5VVpBL)<c*E2cD}M&P+ZqrJZV)8TGdiHcCaa*o_QKC+S}X7XSAm+EOMqA2
z<xEFHs|xCs=bWp_to(NGX5DqRu1n|)1_}B<aBbp@%kyDc*2IbQ+SfWpV#c1vWc7XG
z6~5p<_qrxuk*TVbF@e6>?2A;A%dfoi%UienrVfe!IDh1I<zTaV-UFsbdfWwhJh5iQ
zwO`1O`IlY2M_(0sFc8YJeuc-bQNJ9oh`#(XKyWQp^HmJKgIGPJxirEtGUPUco9+{f
zLZ&YB(|iT*Ii#)0q=hbVUL7W(G;@rR+h;Su+iCms%$F+-n=|Lv`A-W5(Nh`NkSm3Z
z2k<s&R9Q&a=nU+O?9Z$-SKzAi3kwWo)=9JqZJQt4pU=0_Pfy!part?Ub%QH)nufs0
zYbjkizSx{+rtKV;qF%>NQTsOz`kkMM(jX?0&7hf84Xf0M$TtbNuuI9=nEu7~)EL|I
z2a<;xB0{yG@^Zry5wT~8Cp_m9Q7ocKdqeie3n0q)t!yXBMqx5O>4ch^x{OW3;0)p9
z8++@_KH5pn#?@ts3&Q^TIUU*u`>xMr3nrx3*}gAhj(~!>xxBx9-28)ec(D_RR>!$K
zdlQG1T9q_{_qS9{73P(FT$Q}n=@zEs%-8%~`K)~oF`rQoMv3xKIlk76G>y_v@q3r<
z+1KotG#>8AGD*7R8ZG~$RM=Ik%Q|!h(U0zD(LH+kMMkkh{>sqi_#jQbBj!bnV<=sf
zFy5ImSzIW}`UU0e$&9yxb%zFmHF-bUr(I=VESL#Unjt-GcXm$ptSskqs)mlULrKSi
z(~n$;kx&UFOCOYr?0>59;E?OQrvpRYkGx6FU$lwT`dkreRvrR@UpY5Z?s_(#v%9t5
zRqjk!ID<%zJCA)6mnWBvkT~`XZBb6)qv9x&#MX|R0`C}=jy%tax7P^7D1~I2jcpbk
zmHpJ6`tIsw!`WLhT{M<rD&>k%m%7Xap9wuZ@~|8)5|#NyEc&W)Y95y6%f&s_S23>1
z^-rHuF4~m@UoY9W#*uo-_E2|Yw$IDj=a-|R-e2{%m%7;SIw<!<u_h+7Ly`XiHgdL>
zpx_=`vU#Gz(=X<?Dp;jW6P`q+*&WzSJL-9FKk00zwWdfvb0OASYM*g~Ks*gcui@Z*
zriUR-#eAFtQc8tBVOzPfX58`uYG)=Ow!=vz_+Esi5bN12{W*;igclD{a#dH}Pxe_&
z8H)vE56*blx`X&|w|xJlR+Ujvav#Q{eVBAI(ni^5n5E{Kmn&L@f(C=!xIwwj{&aTk
z*Y5_~iU+N;&J666mCV-+yFvCjNZ<?~V<h7zcc0kB2<dQ*!31Nqeud6afycZj+sNl?
z1j$Q8SS=(#zsn0`#KuZh%ZR?vAhi{6v++n{dH3zmwe<b(cBG2CIexZ9cAFmv>&9ei
zC{p#sx3*k?@A0&i{w2gZmGND<#JNJ)U)sANJuu+iR#Zsv^{~#o8CL;{9WU{BCwq$f
z{Bkrta17KqtJb_73a#(gB)=iRUR0dT;23(pn?7AoF3F=>!{&y@rH={=+U<2WPo&<J
zNetmD<~rV{@Qp!9^u;rELdPpN)t^ps@UbbVikhoLw%_$quf_UX8*|0>8oU>I&l<Vn
zadP@ZZ<=ki-6hHVPP5M2iQIIy7^e5N8V|cUEQ`|n7>|Fa2p-=#q|k^nPYYyAX3b<Q
zeRO|@+`o5{g7RqBNQRAaMou5W#Y2uI&yN0zzDK_!6;b<I=97QDo*Y}^hj`Z0S}G#_
z?Y7+4PRM?-76_tA{8airmZNRW#IG<YA>Ua!n9)$+ns&@})07NXI=*(x>;2y-iR_$d
zz2EM>%@XtssgpY{lnV-BB)<av?R)o*JU<$xNqz`(%G@*B>P|_j98cYYmV<cK#6^49
z4#_soyd=t_k<uQvEY4q#x||w4jg`D`k1Yxw4N4niUkI_~oQVAR%-UOs_U#WYijlYH
zlO3qf6+F98tX-U@+bA(HpUM7weU-jK;PlZW7+gR>-QKPqjr^+Bqaz9;=e52bV@qO8
z|4Mw>zE}T;#cQ4S;g@eb3&)WiF$%I|uN&2fR9HTj)pWXl)Y(^WKK%h#)-e)}pR*p_
zO4tkH#P`{ldM?@UeU!@e=u0^_XRiAueDQ}6aXMc$spqd`u4T__PiaU(To^<@*q+Wc
zG&)GqdxFf-&WVg)<+OZ)5b^C6XCI2TXxGA=isLx45|O7J3j9Okm~73AXr-(7ua@q*
z*IxyxHs`2j=mF5Uo+PE!s!;a^IBDao3c_b6jD&R83Ui7{8Znj=c(f<nNgyY`;1H0*
z|02Y#J2UpiXu{8vF9?gp^lu26OS=mX*e6gbVe|P$;(lE>WoNKUc|>Atf-6V@QLm%#
z7(RO-K7yg{gEl|R?8KRI`_NDh#B+q@$Vfs{z)U@pesfjr^JVkX-@NBZO6F6JJQ~O_
zz^I4*oL^7(;|L2VkZ}y@AhtbMP$DD5*54DS!|^M<WaH&douIk>4qc8q&kw8aby|vb
zjdVBTzO~g)C7aTNroO&A4Ly^=#@T~}QZ$_@cIHlIgw!gr8mx3VM>ACR_b9y1=~d}m
z?7cdJyYnXX>x2)ntt<$Vln+Jl%mLCz@fNoW6QZ3sbK;MuF&jOfv(PPoS&B=KnpeN^
z^oWS`n-t&mu9mLoOnbPsWo=+{RP{ofr+K=^@X;$G=jqs+n42^Vg>xyav}51vR<~`Q
zWtzZe3*_@Lj)s{$8<08^Wu}{%8Xa}2(A79q{YdJt?om9|8yOHj;}hphvK&l2#?y-N
zx<LfN=>mBlaY1=YN;mzGyG1}ipu*0WL}J?(d^5mxJ%AXvaR{OW^IwuswhnPq0gy3w
z>CVHFSYYJfLmBiTWW==D{>Z=j4uDMxA^6Cq`UJ*{qud3`x`i-`MnZi^u=)c^#G`&N
zgq~=4{dayxisVE+pq9{FqhFBPXq%E;7lp_8axe^C$?mFabX$D1*%(alPm>8KOFcjy
zNhEcNC4yiC7_wv6=RJ-SP_wIMfbJ%ELYq^TFTou20|@|r7&<kbSy^30Ma^Rn?h+A?
z1JO>5cYf;r&S3#(WEAC!P!JLQA;Qt*<>e7Tp)86Ju|Q`TeS03cw2)?jW7@)igtH$*
z{<y(WLc&b>KksUXT*T~iUtiP5n6rJAp2A?~`F@LKD<vgGmoaUv8gW-HLaY>}i+=q4
z`3)#EIM9+n!GlnSG$kRUa|;%EJ*L2za)tmhXM6;063b*#lmw5ju$IHr`Etm6z_ckk
zWE0xk8Q%K@*`eK~kBWZ(c@Gu$4usKr|I|wThH%;<;3-2yu=op}Lcb|gke-Bc%i-qD
zDdXJ<W0W)kgB>}PeTL_Sx4nB@3cC5MS{WtUaA48G;7}P2ITH%cwV8dpyzC|E>Gu(A
zr#xSYMpmYy+1dNwXFs^8NbdG_0;w|)iIsr?MnLLqUsyxPTEuvoJWy6#ya4o@IQxOp
zMMzPpb00%w8a?nCp(+&FWfVD)69ua1E18-EUVwR?1jNv@%uFJ;i44a`tN(t=V<e2M
z2dr&v)ARB|fzO?on5eGk2oAYjeIgW9Le=HQ>MP$*?}m3AAgS7f%!g;<3@_8j4*G1(
zA{1KPMgzI%I3;{NzBnEKPtdts(YsHrEF>)e;gyho43U_Du(f|7b#EgT*}k@;4)fyh
zWNgauHIPmV9}__sZixOefgJH`A40yMfjUfkzZ#8~_`Cb}9Dk_&tKgZ9a$)6ubsGys
zgN`P^r_<o_b#xY`S^<*OlxpY)C2y8W5*~zzRyZ6@6HdSFUP;5-Y$z`Wa)(dfibCe+
zv#Lcj|K9aCHOCsMLL@h$H-ijCJo;yxpyIZJv}7AsJxnsfgKtY(ORH<4)SOR&u`$x`
zzZ&Npxd^*H2$oUoYCe6gK-a&4z6)yVc;TBv&85{MnE$AW_0to6>FPLqh$#Ssc?POH
zR)3CWZ2FQn@w5F?Y=$^UBqC=}LJpD70%i95L8fX6vVhL8IV>Smov*7DYh(OBrNu3s
z@z!HI(2CH2GPS)GHSOH1l8b?Au=+(n@YOH2?wHocmql{~8I`vC{pb6f#Yaxj)gV0i
za-f>my-)^Wvpl38W_11N$+f$w0FqIOfxX-3<o|w;=ck0ls3z7Tt2P*ACVKRgClSan
zB<KnUBb9f0Fw4J>Lf)Pm#ww031ULENjn+H6Gw&<<W{nL;VsyTSjA8Nwf#S?K;gkSz
zZ$yrdRAw`FNd;lSOLFhG^zPf?bolQI2+&=BAypW0TAsYRwA%?N@exhtzb{X8t%DjF
z1;XR%xe103NJ+-v)D@pa(~}2O#&;*eg`;sk4(Z<YNI3%|pgagMxpRBaZx{sz;O4YG
z(chOjxLED><y0SpG8;yFj5j{&{x+^2uUa57>>)aJaQ`msAIWQLYdvMrH&11LzLipd
zXi?ynYabkJp^|5;MNZDAC$DHJQW5S07~H6m@#azFv-~aGT=bE^$B(XW7oeL3a&q$Y
zbwivEysJigAODF!2S)_QS+^iXUylrb<Rl*m40|RJc@7SB<FsbE@O`s_4<Nd?Zf`6j
zbFBz4o^0&xCy-tV{x%*`pRa(Y1|uGm9~_oI@D|<_smY3dZ?ILkBP<}m3;9*(t=zu7
zy~@S$pVK9AMTaW}EUcb=Ymh+Yiuq_<tbUW&k@3J|#Z2I&`P3o}`1+l%4R8n>k#>Z_
zai9tFGoC8y);u4F?0W>r?+jsbpE^*m(ImIPwC}X?k^eqhfbh^hk|yvEewezr=ML6V
z=@NrXgS<k}z0X%~I|0k9x2GV$_rmY4#@ZAkKmHbE(ZOPOY;ErAVZZsU49%RlaRvqq
z#sm`jAbKvM*g7;v;$&E)F#hP~Q+N8~zfV=4)CRAI;*dx>g~4|k<QG=d(HAlzv#BK3
zgW9$zc5LxgBL*JCq5lEaVu}kZlmzI$22SV^NYsxbTsHz$!S6G2asqkD_Hoc0iU#Eo
z6vum~MzWF|zNItmaDc*^mXq`2$jg|2m&~u>iW}c{kK5YWjc-oBvI0>Ppmd~u;I!oR
z+5Xj&cgxovtgGMPXo$Ws41=yAAM%s}(hOA0OE}zVG=CuBqOT+}BqE*~noEG*7Nm8}
z-ajBkiQ1>2lG+I!u0e1#_1rHD67xs$|I=%96NFuV;3oXzy1)tQz<T3~ORutb+-uj=
zGQA2A<#FC?N*qm5P)?CZfObIv9BhQuTlvw25~V3qixAC>fqo4vszwxIMR8p?Bj~;g
zm*ASxS|2Qa5EQq2UZnizgHxYaxUJ-euy`OZJHlN~fjzlC+;@8nb{Ai8)_R5j{ekv5
z34U$hVzLw1CcPvtR#a-{yoN_0G~I9+(dyuF=OIw>$^78R>_<zX^UIe8uuStkDYJtI
zq~9N3onrxPy&(p9YlVe+wvUE+{+<4zC**FdzZ1g9mJc#1nk;zNgZC%k#xk_IVES5T
zPu}BK59ha#y&n<&!EqXbt#=IRPJ~0@h~_qUt~)?)umN^;k@<?{-XbfCu(1mBpWor?
z`wfrST%FF*$yH7ki*X#o7oc7JN1Q396PgQZK1)%1j*6L@iaI6#?mBK!u<`c6a0~C>
zK2LDHliw5vJAJ%>%~3S6F^TQU<o~o|s<$V2P|#KsxDV|HMgM8ZKrgiwz=pOkyy8_y
zZ|?_~<i`=Cc-zH=%AeW5_W#wAAs3(FMbbN5Yy1*sEsjq{ad>DxVgK%V0cFn*q&W1!
zaRuSssMH(A<iQWxkv+U>Afq85k=$W$`15j>N~&b?yScZApNF4+2lsdk){GE&oPXR-
zoSd`-q6V6^2S9l=Tmm8Fz;;ZTCkWAUw-k9>+q2Lf5FtQ3xZ$0-J2p3O`s#}ocON}P
z{r6Vv7A!V$N?%^S@s>~iap*lZ`crRIQ>{(WosChflPZjjJqFi94M^rHlBtc$5<0Jp
zRJ63H(LEB)B|L8Ig$2Urgst%wJpqsf$;TSQr(pa5A9_icDY^ys<CCZ5LQG(lojQ2y
z`teInA^-NI?NHH?AF(j0a1!M%w$^6RYz6|fzo3aDQ_#&{nbxT}9-MX%*P>L4hBqMK
zY^LFPc>G{I#$(h1et7ge2(4V?n+BymOzbCzL-{Ci?D^{-%c^(mv(!?_Yr;;5BSa<K
zVNr}3tv3{xhR0;A>x>>at1rV-#w|ZGw}vm+hkq}l`fEM=<UP^2Oh*V*BTNPw)_}fz
zps(hP%)?+luS&6CJjq;o@lq+Q5tO`dpzv8Zfcj_XUA!zo36$SN$RwdqlwEL|P$WQI
z3C$cD_zdWgA3d?dVn?a%0_Z4>c9_aDIufGkjnH{|31AprGEo$G?%Diz+yaEV_l2__
z(1Z9nw5^`$aqoZp$<jFPvc9cx<ggNF1F$L)&EdBV9>9sf@A&-`3G-7S+<wpxdLK?u
z51}`MhE1SO96W!bG9o!S8Bsh@Cg3GBV2&^d@xrDkxt!2e!P;Yxqx}u8UdEIcZ-6$U
z+tTo#r>9+4coupIF@Uj9;X**iK%BuSHg0-%Yu;Cdq_lepHt7J==|vrPJYMcSZ@uG_
zxHC1z(0J=HbUp&60|X5*g$WeU;73f;R*^9>bn`&dR^Y0FvW~&%t}_hv4LNxs5p8X4
zK5IoNs?Xmh5e31U_P##R3bBdc=IBqB?v3I@wdg{Hz`wCp@@}Wj7uQjF>^X<n4mSDV
zCCnqG_gsg#EI%kbR@0_iQVLNNE~@NT@?@(s9DZD!C^?3GdM-&jw3x40uWqxqv5VYZ
zEHoGoS?bC!%wT0QFOtvgy4(%S?(;#Vdd|^IT{n-uyd~{0V<7GMHbTG8kw*Ip@uQ9(
z$F@bxL0P$wDb?w4uOLAO)T4J%nODy|rz+~0PpA13U>Cyl5R<216P_T{^wK~k6+7c(
z5|_tIsjJSxXc4eZq{`oq)1KiKdtFQy?ogI6yea-Yi1_=fW9o_;hrP`pa9tgkoFf#i
z8RAGEpTVDzH<@7V=VBW8>~%}L%wd~ZEG5{D=1e_Y$-?+(?j2Hb@Yj-SnJjpJ3}g)w
z&r}O%2g0JrbQ_J6hcH1O1ri136+=w5XI@KPZ5r@<cND9o&~c0bbQ*<RN53__E~@&B
zP|LX7J_cQQ!(Q5^H~|?&QL=`s9@`QV&+%8-8y{ThqT8g2mQL?cex;jMt#fKB6O&+P
zCiP_Wf$z<lo%X$bBMbSVjO#=5i_;-Wg5-S`n{*2IoJ8v-3+O|hE!8-0bM&=(XC5bj
z<Dx-7>`z#~BI1vImWd5Z_$F8=w6ZV$i^$2M`Q4cU!ILMvQuvdsioQIs3sVwEB79V@
z@KGbEGAvD$_|EIQ*6I;DRHP>~KQ9buD<2flQ)#_Ho=YSkW)nnDZx(MY$NSVQ()#4V
z^@aKoOM+Irq~?g$gRY`S1`DQypYWO;BVy1j-bV*jOr4Y+#qyx|0{6ATbOmb?l27W|
zlL^~GX(E<prO7!t<a{43YVr!76$ei(i8sAfAa4q@dfD1O@`kKlP*JF+jQN}xQOzao
ztaE(?ej4<;Huwlu&Dhkw3xhvo%!}m@>=TVL*OVq12)<xm!CJyg0E)c=eH=zf^ts02
z1E9A&(_t=Mz3Q|cp5hYy?5nwmT(~8^zpXmMJ0PVh)%c-jeC){v|0M^_FnNr;`kscw
z*wH>vj*PlX^=wJg9>;2jr5QgIDCP*av^{wDMNpHx1@Sfu4usLH)CM2osrINx>(3Yl
zL|cTTzq$^65lNKVPZl(l&WvX_$mtDWPxpD@STkuRd^j=7NoVlA7MF>XB}<-uI2MGw
zU=Mcz#&1t$#K{v=PN_Mw1ZxHn3x_sn2JZ(A6=Zijl?EeKE~aEYHT_6vpg@?=FO3gj
z20Rl_{Ts;lIwl&Xm`kyu!m@Z3KJ~L{(cj!n+*zKAMr7QOmitL_GO^-F--*e<HE$uo
z+0jI@141U(;-~uFlN0BjF&r~lQ)%!`t4Mwl#ga!&WA}ka<w#>kw0%OA(_KqprAGrn
zQqS&|s$==Vx)_NOmcL_Ykv@E14l6HAeZVGHA>F3T|B7dEgTIDpjccRDqg;`f#|mrK
z6E5GAj4~<FGQI9{ys|_B-kG8&Yq8S(1J}Ro`<NwH6bIf^Os0O{78x`#Bxc)ZEc+N~
z>4I{jztK;Vl=A01yYHku*7VJ_xD%6^)Bkw2<5WQjcpBx-Br0S|c-)hyX6RS#SE0ly
zBoXF!YM)LGh&NzVKjQ6?>pm&y%GWg>FP})kp7=xItizgcN8$(UM_Wr??vAmkA6#o=
zH%KlXdj3$dq&>S#Vx`blJmB)v-A2J+(C^zh!faD&pK-|x=}6?AUAe|IEGIBc$Q=7b
zP(V+%UyK#E@N$Cmx~aE4MZ@!;>=P^uQHzma?40o@{UtPBdHS8TSDW4yH@oXSAFekA
zn1AnnwJ|OtH#Yi|<NIUDt1m9DN6Vdkc$NO$p8wZ7#vb?vesUk0Z`RV3@?`b2^wCP1
zHik4d6C<1?7s&&^VdNR){vX!<GAhgM{Q^cokdg)!B%~yi1_|j9r4bBTI+X4Xkxm8a
z5Gf0!L|UX91SJKeL<B@YTHw8w=P&y=#`*M~aX#$rP#*5*j<v3JuQ}&6r&n}B<3-E7
zJu3G#vzrT)7Kc;N7#EaNY5mJnh{b*FVM>RH_#OqBw@vina~09oQnAs&DzC3!x~iko
zsYF2<e}kJlHQmXo94E-6EMjzddtl-0_g7~e-eymA@b_9X<O!ApXI6Gre^6EFb>n4q
zTl`EW;TY|6^MWn8NM~_HaY}nDOLW_g8O<z3VR;(i!X2X6_&M)GRwk2yD>9-*WTw{c
zgHf(&BRW5sZOP?@@-48KXf|vL)70KXCpcN%;Lc$VdUcw{rn|`Gymt?&sX$@Wdg>Z&
zliK^N?59z0H-#)*UIq)XDc7i#a1l^t*kzx=HB))OH7`N=hd_x#d{$%qm3Za0dz0;^
z-_`fBTKvjf1i-z?8B?F^CKBHg;f@`u;FY)>;r^YK6*uCoVS|9o`2*ueb@PmzXRaQM
z6$WGCx3%a--bjs#&uucZ9>h1f&*7!I%(=O?-^e}mxpUJpa`usHd^xMLP0y;iy?NNR
zT0w!qoQ`Qe?$5211>v{F-M$mx3db(&{`tALxzda+PZ$hZ+S7TI!Med)F3dfw(K3$)
z>FHB@I*TvAPioRD;O(%Usd}1J_m+`~X3&84>#sx(kIw!Tk{d)=lo1RBz>JMPSzhpC
zljC>bXt``hGx8y!^+sjqugPX?C8d(f-5AAOnDytK2HHgh7qQcVFf*yYDEPk>`NQeK
zS4hb`s46e??e!NIVgkC-K=Zo$o<nCm>Z?P4c4B<yZ4e;-+;3^;%qEb2hBfYKQrL&e
z))*7}y6ExYEbMkq4wEVV30q^o1g1CBp0~`k%z3;!gMR0q^RTpYtMi{@|HH1TDs%Rh
z_<WqR=I@u{ySgc|rLwYO^4jE`m{_W6A*}Icyf0z!P`yqBv{~UkIpQCc<I`T;-Egdv
zx)F2Fit<V2WJPQ1ObTa4c0L#3qHu^GkK%{Dhe9TI3bO)TC}#Ntuf;jiJ|4S}L(6Ft
zK-NJpbWvy=r$VR8gFLs82}hw7CntXS16xpbY+QoV^-H+8Ru$*eJ%6V{9Btj}&+P@I
zx4#z=YW#k+S#Lr4_1cH?djbInp@&Ct_Pa^JU>oqOlLcc|(VVf>lCii3He5f=4F_3>
z-Sl|b_g<?jhAwaFrxvo_zdTqgl=;GKx{Wc~E>-e<L*_aIhB0m3a{UBP6!RZ;GiOH<
zU2EM8LgG{t{R<=864!<;FGs7r>(}dXR`5DEcn<eQG||JpRr59tF}-Keg4Nb;XS{E)
z1L3#mMR{=XGR@;be1&Aep3XcoW}(ksAp1Xaxz9MkwTfq1A=@LCx`Ugpl<%TOgzve8
zHbuY3bVVk)gCA@^TkcLa#8c&2D+@X(IFX-9$Il(4sBELGe7EWjB4Ja?uWway<k&>s
z{^-x7{os{hqH0W`*X0e&)>^rM{<RfrS(+QsCuC{XiYlnBD1&YUpH(u5Qmt6mcVxR`
z`OR^_+uM;?>|tP<UsoZ1PgMJK<`>RMJ+GLSaI$m}3t=pY4BEjBTFbXR^UVw$PYgjU
z?!uC)ot)AArE;?EP^K~?<FlmOh@n@PCq~o|J7rN(8?KXKgp>Acf4-VRGGX0~l!XYf
zlz5sx&Vt~AM#;S(rfUVQj#ghw1}Htsf1SvQ=Mqd4EM&cEWU8OHYw$KwX%5c*MU{2#
z$3L&jAFTw_o80g4<j{<110_t}bEV!~D<Nz8WgXzHw48H*$THP+6J!(1@n9J1IyG+x
zrx7&-L}S#ia~^gMY>^7|LOvPug?(7Qef&AboS>SY$n#5$*Aa{WifO_&mvY#n+W!)M
zn>lE$^PVfn)__jM{-AEtIY+R=_aLyUFn;9rSmgcm=?aIVRSIKTSZA$;WvH%+bQ&cv
zHEKrOC)_oM%WZF^JRs1uekLNz5quj!75GeJ<>tOH@r#F&=I^UlwKwbE<Pg!NSd39b
zo4FiCpHldz1$dCj&^LMSpeoUg90?wtlZ}NPR(7VSU->eq^K=(ic!StzYxAhL&A!O#
z{!tkQkh|r1=bk_GJIK6sZ|cDE$NSY~F<gx!2{(T|32m;&4&r_im5&NvSmd*g^W84F
z)IGSg&>_n47+=mk!KcN|=slpy1_1pcM|WgoUjFXg(dZ>n2u7HigN{pM5Z28(Re2fH
zhKj`O0t|8$5E<AdocxOPfe4*LArJ_JAJRPsT3^@{BO|ok{r5-Se|wujL5Bg-uLrC9
zQRgm4fF`j7JpZO=&zUA)alNR~=UCxS!>eC#AR~0V8^>GQasW&^l-?^0H3P$bdcE(;
z)Bm5yB8m0wqcZn}8@E5XA0B1ha$T~kTP0Ez-X9AXAkE>!CfEOPvZ=C--PA=eUS&o&
zJbuo~g7rsVr5O_r1=i^0C>qj_d`1r*=s93wH`ZU58*~w_oBCNZw7#FSLgUy+ee%=r
zY`WyNmrFx#(2VQ1dsw>!y~+$3YR-qd3j!d41I_baS_-W3cA;rr)zOg;<t2WE4yG94
z5N~@wynufl<Uag$t8v{Gwm5`Tzx)lejj0|s87o9leE?jq4V*9~#>U2wx7R4N*@6|k
z5vb6RcRs-OQq#zc5-ClA3_@M%dwX6=N=g`&GZB}pfUW`Bfw9823=pj!KYmmK9!V`^
zLnST5upzFlo_B0!|8BEhD+gsa0BZ3IOJ@;WzR;%jP5s_QToU>@*!+D1c#GxHbza9A
zdHxhaBpgH`B)}}-e*HbT?=_8rCRFn6X}#GgXOLqfBCK|Q+<D)*(I?f1^s^uke(~<L
z7UpphWNrMmf#MKAHz*4N<eXFj7K9bH=ePDDC}Gj}O?ho?I?Ctx1MsFrdmQ7#rfHkI
z{NG>cZun+}Q?QW};*5eg`t94dwElnA{w%8do4}@&@SGs~cwAL!>Eq5$rLTp)t^YdB
z<tMxLi93EQEm2E&31Z+9(~2y+A6Zcx&CHk}kN|$*2`B<WCw_nb_<vh&2LPLmsc?AS
z|KY!N>y|u|uX-I~(>&mV?A+XbZJQTq|Ha=t^8W;fhuc(7;xiBDBn~bN2;V^c-sotE
z6GC#q%Azh{xtd{5oVmi|2FG`V|9roWALrF%*{q7O?DHc1@`oVYGC&59sQM6+A>$!#
zGYevB$zt0O)B@UB(dEnat5`=Ko5wx`?P8)Y8FkFfFXSjh&wvA~$?9YsAvOO)G>-uF
zASVu;+PX~8_V1?51-adn;|~C}I7PGq2PbE@+NsbckIc~3{U3bCW`O6bbLbyK^2d!k
zceWE3^R4NF3;(sYKT>~Y$`rotD>Z&!<@lJ_#eO_7`?dB%ARqVC@#y|dOTjqW7T(DB
zf4}-DHq+Fka=uN`+gl82p26R4;N#@K;YjA0byIN*b4{FNP!Q!uB5%RX%}tcsxw>2Z
z^QRr4Z6s+!pFaIIw$1<d9koQkW~*BMcw1^JgDfd31Axea>D&cCr&RRxXf9m1uv+~E
zVQ;8CKWI|_-!<&wva@(+O>!}K&x#|e_B~+t9)PlHgKK4yq1R$f{Ws|AUjW6qn2j#r
zQ{MGz|1MQh<cH#xB%O#u0AL>*2BETt^F01YB27pu5@$V_UjpdNw!WR8pJWgY(Us+b
z($e*yFE0PS*pq~yV={QbbElp}`s&q_wzjq~P-biwu&iusSk6VCZ~(P!F-%)zc?j^a
z4AUbKCz|7!pd->?Gy54&`j}Y^-zGMqCJY8rJYLJ*^mm2COXnl1+*Wi^-YDE%<YAGm
z5r)WXS`i0~Vb|C%YX5FoyOccRXj*c<B18Z1%R+$l5uWoNx;ikB=_3+c1AkdR?+R#C
z?bg_ReUIs2lthevfTX)HcTl#mle2T|cfaFf?r(?NeOcZ2tI&LXS6TVg2INga>6k-K
z1BjoVc1Tu?XXU&1i5-<U#E~5CEnxy6*;NXZJ)(pXl9B{`_gp%=x~#3Or{VXI)Sr+>
zXxVD{*}re8CCN%wDm|Pu$y_iGm8=(Kz(P)?pp!@Ds>q&VY)lt9mJ-ul!EOL>a(C+J
z;4UcXGZ&0?&S1>X&;NCL1zVA)wY9WU+Z;SRmH>7V92--_IG!Jm<9u|z*kM{0pgG(>
z5~KpE6kVkb21$q?a9^FYnLR+fBrr~h_`Y}Vo(Pz6R6#d{HFSCaf{KCeRQcPtH}2kz
zwN6F$4k(ouNRMKt`Of9NsY=OB=6}^gw8Hmwh@P8^3zC#0&lb{fd4#gl(a{MC3v1oD
z5xC&lJnINRTNZMTxOnMOM_(Tdr^kr9`0_;`h@mbpST5_LmGlw0diAO}u%77vGNL=_
zxo(D>pdZD@%fV;>>G;L>2P^i}&^IwoJiYuz=-)X!G97Gt`Pc>)GL+Ra8c_sQIYpic
zg*OG(zbt~O5DwRiynm3lDS|6O_wEVI`SYTp{M_7!ppe)ro>$1Z0mO;4oE#PgN;-0N
zb^WWg2hZMlFsJos3plP5F}~*`<2d6xFnFEjuKzX61y5KEjXQU)+1YW!S#<U8gSO%U
z=sq-V-1vZ0B9QKldYs1Jk^FV>L53lJQ;Yx+hM^>*r>7^xvfp^NzM;?m&>7*)0ka|=
znFYsRImi5i`SfEDPDBEV9|=CoP5I*$MAF8_jBFV$<j|kRMjtES@KqWe%-2@}g#H+N
zUcm?8j%@_Stw`|!6DR<{_5|dU`!_xt3Ndo}6mb1<6A9;$wqHX-qXAaWfckGo^<a-5
zeewjR$_BG5`>{<DktDf9$6hGkl^8i08BB9?^Uwo7V2{z#4Q(9v?%x-6`A+)k&YN%0
zf?zg~tB8D55&jD^2DO$T;t(Yx@2t(1b)kG3xW|O_;>l=*0i6xBu{4E};O;#rK8$#K
zD6LX=yaGznOOns4!-BzU(M<)#A<}ky)EZ_sdfYc}-6{dA9VEMfa{?aW<X(t=Vqzj{
zXc6z51#e*{9-dg_W(*v8DG)^e^0(o&q;%=^>({x})v3^&fo_7$_%~=P5?vQNP6$6=
zZCd@fdqZ4hP~3)nw%T<G1+^ng9Ga&rVA|Efiw#;V9YngphlHd7XeCjHuNbiAg3VG0
zh=z*6+8Gg*s6hc*i@JbSq5v%V|0sCu6aG@Ppu713u28k`g7fL~XT(~;v~g^XJ@)7F
z*@<}(_xgR`9e-`}RMyM|6m+FBh|Fc4zJ48r^gVyQSpd}b5bymp-4xz73$I`p@H{$t
zdd$H`NZ)h*2{MF2e$G$>_@D<2$t|F1B&=;#KIh}|-oBh@z{{5xQCl2(`+M-=>mw<u
zf5s_J%~XCzl`T2nqo+U7tc;e0qA08jlmujvEjxb}r23Gn6oj5B!#koBbn&3XHnbE~
z?#x7@3@Qvnr9-jFuc$y-^}zjCf`icW!oqnVd_V9;*<_U+vi@n)^e+hs!9lj(Z|f))
zi)||Q>XzNBt-6}o-)>NowllBPEGxeoB(=|BPf+gca*+M0VNtcqLO4QxUEq>2Kg*l6
zv*ay*I_-^$4chUZxlj?;Oi;Zdv7Oesz|3CpEP;yg9$rZ#_I(N>pK<z@j;cbN#f#s<
zHB!Zy8+7oKYUC7!e*Y|Jh~XCye5lrBoHp2Ts&Krrd<NUZFNAJUj#kBmm-SslxEsEl
zc-on!c~WaP5xR0qFDfd|+(Fg;WE-}7!FZ{dPKCV8K^GFJ88J-in<Rx>mFa4zGVgRG
z_8xq>c_>=%WP-I-iJ2=bxPDZ4zF}rW=w)NeYg@3$VoBPPkCLg%qa3e_B246+=*Vt#
z6t&V5w`}S7@a#m|4eFpXT}JUKMsn3DamE6o+G?%aBR}2NO~%!11{*ge2;@Ygx~Uu3
zU&V=V*9g#+o12W}Uv*?iT64oMTGd+J{58BqM-?K&dAfilRo%5b;lw4SG57CIGqq>#
zvo7ZD-dw!iI-aLZRlaX68*-KMRBV0hoT0EnC=J_GM4Tl~3Geq`j|KCue0P0pCZ6=e
zies4ia=N=5-`4IFHbdEWV(vpntO>)2_pYl=f+|8oPvVcxCs91lQ216AI(js5tJ}(&
z!ECg${@VR#Y0n(Eo0v|k5G4q3Ee@D#+nl22R9>~b_WDuzk_LH)sr2(O{0~m~mBwS7
zS+0J<d^|mtg_;U^*~5bEn?5+Zie3j8d2Wx_G5H=;ac_=C{^Gk{?VI&(;0yMf&AAdG
zn~&F1X`7yPsFz6ihLuXw7p*?pGH~i)2u-IaH%xtg`q`Jx3pUsz@z%`4gj)BmrkKrr
zyjLkcRF=^<M>rq(g10cSV=&P0wwoRE&s=zFicbu}g4zymisc)VcCm!v$VL(0F*%W8
z7-7iMk{{#s(pj00C<yZdQ)eOj$q^gZu|=nun>eSeRNVP5Yc`1pw<o)#ax*vhjnA2v
zC9N#p)*YM>+;sc8D_Orf`~GmUe!_gX;zVWUwGraS_e1ehG5PD2?HIaSK7RbOzuj59
z@JG+>aOC|TfwsbXmR7mxd(Dg{DcIpE=Z3c(B(^2q2pkN{bd~?|&C+PdXGy=N#-K7%
zuINQ=qV~X1^wzTSTgNoM2LeZK8^&Eb9};Lpq$7lQIubpY=d({}8isJu?Tc2~5MuN!
za_dj!FlkG_OKVs&;LlWk6W_d9SHC3xG`Y@@2O~7ykeFrsdNuWx#}Xd(p|&``*6?+N
zrlm>GgSwJ&<%;|Y=?yXYG1uo2lv?EN7upC1ezuCn-78@JEO)1Ln`_@b<xIY@#uGi+
z&K?c&Zc|MUhPQqX;|Co%2Ny0&bsEwTrxNqzrCE11Q=Q|U^(YHd>U;KSbM$_b*ZZ_5
zwt9HD4+tX?dAYS?f&(M_?yIp$Pl*PokGYqw`VVwz&|S=Lz<Xpl!s1}<<n77zQP#Ec
z;fKHzn)S+!?(JWvXE!*?lGm(*lC6Ke%&<s4Yo^~hEzQu-6-bHUs4wobj#td;8gQ{D
zPDwzCh;eAwpt+7vQZmi^T$rV1!c~Ij#7PbepVb;wO-*=NEg8;w=RuV*%JaA|f!**v
zIf&zzxMs+Q4YyP-HL;)*yBIb~^v!DqviYiG&Q_|fv-^@)v-QdCQUi?nSTtKZZd1$A
z#)qwLX!aZX=0AZ+GW)QRY-fj(igC;0%E(AbQk=OKH5Ze~Nj|>z9d3>{Z=|f8YvZ5O
znMhZLy_l9+b}_a2F`@4HJYe?0X+bZ?t-P_3&#rOL>6kJVZ_slw%vPzizxw7<mQH3J
zn;{cqd7!qHH_lPvT=nUVQF)qK7WK^h)2{QTde*8grc^i`Dyf}cm>yUJI6ey0&kE=L
zsm1lMfRL%`Cq~~J>h{7R8LuxaErI$3g}idzwBk=kl}m*OKO`0%V5>xyFU7eDgc^|8
zTnd=2nhSk4H)4Du<SLw(<|W+m;_qvkf|k|RYHkDZYUJf7STY1JdzyRH_nC{DQYQwU
z7@kcD>zn$*y&lbADOwdJ;DsN`zr56A>dfBIl;Yd`1>5uX!)MY7G8Wu<BjxeggMFkz
zoP-PkfkEmy<4vt2k|FY4*4Ha;SS@H+EX>Yo(X|)!7cXwnbw)8bPR(EDQCL{Y+70ob
z$v2L(zqYm(uFH5&?lj?r2a$cm@00G#?!)xbjh~>zb{3Iv$DU1RkP{$|4KXlOq5Ep|
zEU43gaYF4u*y*|4>qMRk0djs3IBczBF%y&_PTxCN6h*0JG1JzsV`4TjP3K#yx55zd
z-S+pJI=Jh|5<969;TZh6o8Wf*YT9eY)|i{8R#NC^QYd#nu;Y4<S3R2<%w6-~7?n1t
zj9Bw~XxFr0{wCw>m^-N;mI~Loo(n1B3fNxS3-cSrqe{)iFNAU^ch1ZX;*j;_wDId$
zv*}FKgjU8g^~DqyyyEbd)V7$~&NWZg&JQ=HBJAWa*7v{&R)~u;F6~(~v;4!rAuK0I
zIuq$>y0}i48D(2=c``ay{gxem-UlD0*R&O_G%pCwI`e*(QGRx6c;lB32}YYUelpBZ
z%`+Wr=kWMMZYqV+v^BEa8=Fg(F)>WVdf^7+BwrqPC7kNyG%Z|I)A}W9@{4($QRGe2
zPY30E*GaAWjIZaW%LMwQexX07MVElRu{X3-Ls+KlQ9FO@P}_k?%U=Y;v=M)2DYIbV
zqbldv?IvbJW-)K#nOTayTn7oBeS_g60~}HU>=Kpui0h1WFX-2UZoF)|Fqg_WP%r14
zpO%os`15xLR_KfY<wGZ^!O0;q>a2>ZOA9dz-lOKBCuE$KHMTOXVXp9tk(`RhS|o;f
zI8y$i+-X<PGd)i2@w#d%R~i<kN2k3dxJQ;ReBGm;b$ZGE&WSf5$c~3Ag?k`VuAw-r
zhNF7Wk&qCBT<u)n_VZ3IMjp}+%}*r;gO<E0M{W?eit$F*d){~Iq@v)=`GlwWUGG!;
z%+q1zVCuXUWry}~GQkrgAy?ld!4Sbu^$o_)!Y7%TE012;UoIiKLKsVmQ==I7+u)hP
zln=G|l{!iZLxR3-j#4JdWJXRxAy$%s6&5@_q1h&^5~BDZJCZch{cuao(Ivy9gSaH$
zDI1$X6SGO_>I@O0I@^Sh4UST0_OcCj{{0Asq2IL%&h~6r&XLq)FP=8wO=HnZ)@X*}
z>B-=nBw!bM@Ox%pF>fTx&Qk9hR(62ufTF39Q3`9(@b)22q4X9#myym$e#_hgW5SZX
z-o84!C|83w=6!XCs~f2Pl!v8i;7_S9tiPbS#6Zck+pJq(QuOd`IL?e+M2Pk!SO|tU
z+DWcbr-%iFyK$VZSEY@$7-;ql?Oz%mkAA_bLReHBYO?F=-?!rI%^@h){D*MK`-f=4
z_tp)G^#^3|pk?!6gf_0!;6~{(ayGHt3mzGr2n+2Cup-NI0~h)(5eegxD>G#mtIiMd
zhSzXhd~lMm+^K|5=W|(P`GN-ZMKk}}Cw-JLxvkyPaal8uf_Mb#J$O&vPRPy5zC58?
zblSqE@&WA$K~G1w3wrG3&hXGjlkC{c#?r}Yo0aV@dB+#}uv+4gaq@f2(FkR!Q^~LG
z#FS(iOyIug<;u%`7u~$nMnnAGUdT{tYX784WYm%dZ3nXyRSUlg3>wRqNtrGu_BWkt
zk2aP)xHDIc&sl8kZGSWO*H{ZPrlO+S*%@8VJ}Nu-=K5ae$oHPAAvKYNtxvrRtyVM>
zC!sE35*CZ%9x<Ka<BU4oMU_qRQk!RPQ&XRdR>_pvzN@vAs8F;ybM$d6*BI&unY&3=
zS%gKc4kRp^crRcw1sZLM4Q%QYG>ZPtjgR7a$=DYW_gi3Dy?h{LIgjP>HSzdomp&_t
zkapMK4KXI=<)ZDoK#NEIcE9&pfc6!c8zM<3#FVNG8cdHg)qbTOooE=Xe0<knjZiA)
zD7rGfaYN!te4TcFO5N?>cg6N%=hLQMH9BxZeYsj5$M*<xE=30CoUy*o*a4xc9S_l>
z+nAOFVc40t_Ei*L3n=DyuC}A(=PXMk?Zour?Fc^wo5@B%G7u($402Ka2}-R&RsgVD
zL|SBI@wq%vlT=-O1qq<QAyf;tFE5IUzCqBiF`%OeqxIe2x}l}j1me4crlC=Pqs^O2
z{}3rrR`O>6&aXk*NjehsS_yVuJ!{YmO$jK%1;k?*Mj#AFcp{Q6E)EvNE9GfJ^mA}<
z@Ti|L!q#5AdSwb(Ncv#qh)_>Z*V{CH_gu{!`Deb_&Nj$*LoqQetqa6X$^7O7?<T!2
zVO+j^8CeW;5eISd@D#uf2Z6`UCVVK)M@wCz1>#z$d5tg-@C{H<LPCPZ?c0e46-(8l
z=FjgPBMOtfe2>mWQ}I0j`4+#!*M{YCr2TDdlm$LQ%Dg}|nKw{_y&1{l!K+jmpeex4
zNC|PEU7Zki;U1Yuh3CW)y`qopaZ47H@j6rn%>Y*i9H#_CkWYDj15SMgG*Df=!ah3=
zfJ$6rdSnEC8pstD@bz~FnJ<qw+;az4EFR~=`9A^e9zX(#8E(Cz3&jvA0#RNkr2!U1
zQB)w{b9JNNq&~)Bj$WCvvb=i!{Ap!nGW_rfn8aeh7qvmii_-|NN)zudah~z0yQfY6
z?JU>*T~#iqKyUg?f3{q+Ny&0KfF`8T3M6TLf~tf~A_Hk`|2}$NtF3vEYd-^I{s^A3
zPx-lDFU$Au`S14AYCSOvl)h`R{Y}#JKQL^r$wejD?BJ8-@vV+w$-D9InCsyMam!er
zj+%RfkB-fLX~!ua`v&ZJ7l8I$)%VjeVk!&&+<-i1z-_n(Y*8>A47S<rA05MbNev$-
z17ZoLw%2@*49;I9jckF<H5zfq{{S$@1Bs;Zi3wGA_if`r7+r^b<Hzy#t68$8+h0BD
z^Q}#&6x{I40qtsOX_3C`F(g%&mlq63=KA@gwe#H>Vo88h0BA{)w$S4TGDbc~hG8xf
zIKG|=%Ext~hn<M0as+Uv48R9~_mV{%nSi{RFrFTUF_0`|dk*z6Gz4J&={x#E0%*qP
z8t$Y2MT#j9dW`%2e~$Ypepv$WJOSG=L4*+iK=A>DaewoF;7?EJ0KfnU6a&rD+FJ35
zhzJxBjrw*lLW$|T+U6G)76h#NpEmgc04Vi`?s&v-0u<b|A?`DU`}FzqFcsD&;rcbA
zzdC*BvvB-y@+^Yb4gqAqg5kcnyqsHAl>%(Q2@R+NV9<i5A+nx$ar4LXmDuA)Ax8Ah
zCu6}A)T<1nBCkdLd8y_ADl&(tZPb&6@o9M2RObwUAr@}+FuvhoRPKIDL)kM11_lX<
ziGPFC!N}pO=Ooz_+rJ^^=?}pEl|ey#>^=Ub{Crkq5^=aU{QiW)q~}73@4g$-xS}Qn
z)OQ1N(UBT=9xg5}46qd^CkRG-)~mY*7tz^*8eFGhY<VIxg5%$lyEe>L`gnOIQ)m>F
z#{eREf;JSH=<EP2uP8!m0bm;9ASu*D0=<7$g`Nj!62Zlwko}wa49eAC<5P!;P}rQo
zL|IE)JK5kHDisC>6JkZjVSX`0=l9q1|3c5u-PJH2atrFxQ@Eu_%?Zy?Q%|o8q2spI
zQh;JBDJvtk22gLL<A&I;45&fSP}B&f9xmF_I0)Q?#Ii}pjDJ_*WV168Y!)v60FK9R
zKPmoobvnU1RXY)=0u<t2=e=#Q|IijftU>bo2*IQ96jAQo%DX8wz+Dk+X#^z;2AXB4
zQFW0r)ET6h2R(4Xt5-|x%Z2}*un(!^nI<l=D_69EVnLlAVe(9{z=T2{QEFP+i|5Z9
zHpFMvM5MsO1AXGjlP7Zv3tBol9dzzrWDE`Iy|;gw0~Ii8PJCW6{n;}n<j*U4u&9p4
zb*^IkU*;18?+?v&Fwl5dDtt#@XmIScPSVi!>=<r_gevlY{{@^cmj1s7oRI_?Xde6h
zT%|Tg$+^0JY|<l=u4`9Vy3F`7P$MC<9y%i-Fb2Sck6;}FII?u(aOUiLBUnLi-n@zG
zGkCJdySobmijVcR@f6Qo{r*o=;K4D_nnuQ~&_6VkNYnu<Zby&s)Sn9Vm&^YA$264B
zZyy%fQiB>dN2nGcO!pXZ8Lz-(=j5D&JUbKUjsqYchnTiKjEwpDn>T%1lmdi72=GHD
zIEbdh!m}6G|EC4OCH@68`#SVa(geVa1;ZQ_6gr94MU<n3<~acY0kVqeAO7?H=-P3e
zm5f<jXPesQBgATRWM^jwh;zAk%6_so1;W2Rj*ThV+HyfnF@d!jwS1|7{S$a-qe`a`
zeXa#{DWHP%H4qp4m3-W&0T9kBDvAL@(i!a8`M_5k8vS;DO-tVm*7k218^0BIyl7sV
z;>*|rN~;07@lY3Fg{@_QtDBn>bUJC=x`p)hsF0pKd6Lthn({l1la8)#C$tZMhY<w2
zM?%@l^jTLFvis7Gm?6MEt5)mxp<S=5s0`?iA3yGSd3hmU?_%3DV5{faS8M-`W?H^B
zvkNH;p2_*1+VeyIe+QXEH!1EzgAwX;ftRUlxEqMhff~*1NBHlq&$aXYIUIPM)4uJi
zKP<uddR;sg`YB9PS&>uD{`RUR&<F^m?oJntM{Z{*8UPg_;L(soq7hmo>FS}+$Z7t`
zol=M}I|LC$8j!Q-rkf|x#kPej<b)Q!l8W09kL^hl{y&VF?8=GI`@lKD`<;As<S&V!
z+@ud}1)q8ghPbn-(zt1d^3q1h?wc8vC-<uG#5>^%ba&>omUV15U|mtkidNw1vE!*y
zh*KztY?T+25oj1mzxcA7mfz#Gx{xLXC+kWLWvVe})6)?=^5)$3cVRolQ!XR3t?dFq
z@+DzZXZF7s>P{F6xeig_V~6KanCasV@^+A1?Kq`Cb++(XYd7Z89`!N-->`3HY=vuf
z3Y5eQQlu@L60Ik@o^L#(^y)X2ty?VZSeo12TDd7t0AS$BZ@l?`BExUolwO@3uQDB$
zPGq~wd@s12PmZNluxRz;eDk8~V|gq_z)fCcTnnUT1vK)}J!8vl8Q2fOD|(_DZ|&2^
zFOA4B9C_rNtK$>p&lW9A?@R7xv(t6ZhkU!kE2yvV39o-8Z-m;OY0;usHL^tX7Onkh
zt$+?0=BCIOLCxvd@j0JwlCeg}gj)_+cnQWch6Dx6yjd-l8dR)5%{uGtQsoHP)JqM%
zT?@S7ZMYSLg0kB6GMA{m96cl}(u^Mw_^EKW)$PqT^=2P!zgX2;*83ys9A{gCb9i?R
zzq{hfi$nd0Emxeb@k^D$WohNBycaZk#^kXkQbY#Vukk0c%Lx=WZH=m>XDr?x8{lJ=
z$NjiuS<H)f?hIqovnJ)z3qt|6x;G!5dVx3RoUVh>KswAcCs#7d&px6oV@%FwyRK3C
zFxrpgjrmv#`Rh4(wq^{rroe!n!vOVe2aexdsYIc+YHa(OwuFSI$@ovq(a8MBP)mJG
zC)lhTKEfDKza3t$tYgS4tA0hXp%egl*ky7y<qj(jJX-!xaY8#pk%ct((rL@<hN4xs
zlflYrzS&x(T<Z+Cky#~e-*z~a?{FR^cb#GCQ!*pl?fVb8(&9Hl;;#iz(vJ(++WdIk
z23x-(InVL{aWUEUa@dnr^E>K#iSZFp^lf|8D8Dh+#Q>=6i%~M5<Z{a<KRH!+t>)`O
zd&XT`^y_zCM$dZxu2$UoEjU|0$@6ieZrO3ohop(OwCggjgYW9*N;hxr(+9QE4+o^V
zgI<dhL}3t1hSpLNsR{4iPBsX)@G~cx74WZFDYn)2|MG6<u9y<D_1GUue`P^UYECnU
zrvbE_6*U=x1vOS`E|1Sf@v`g7b}RNj{xS2Q5nn%{^OtnHMf{*lHnmCRqsZtFt2M&p
z%4~Qi_g;pcdLX*9)69=QHZ!4}A8bwKQLnDZ5?g$_S%?b;iwEmf_lq&cRF#=&68ZL$
z-zrWwtmzZTibnOEZ|;|7ggxNqi`xK^P^IfB+$s*!XKt5aQ|Rn7GBToO4LWghj#`Lb
zlFRUcTS1cCq8;(?@<%j;@oW0}eL3Q~4V6}>Z&@U`NvXb~VVlmQwmHk!PZ<AOVjZ8f
zk-^S`a{r8?PsM&6twS~Sy|erYcpN^@sD7WL=bJJom5C4!io-pl@uL#^*C+*Hzv<4h
zu)y5jnUnR3Ws45&>J-hn!|prxZG;ULMG}o4y9|tO<YpODQHD4>h$qDpT;qpQ((rEN
zYI<RZg`3M|6Dl0*cW;Pfj`%JULR#i|squ^1caPpX2a(Ap_B^b;hoj?3$Pk9*GFv>j
zVsf|iNz6?WNlAl_$z}e@oqFmJnP%c?FZx-6S8$%DG_0XtMD)B#`ag}!h>)@@=8}xZ
zI51rA?BEr7pH3~>L_Hna{<W;@`oLlV{<$+x6q1>yn1+S%cE9#-`;!l+;3vA<=ZJlF
z{}S@;*4G_w*XfIX+MiU{!ZB-qQu^x~w5ZqB39^p8=%Q(45C~PWi4c$XO>9yVZcUVD
z-kji~ytWRsz?MAWI00<@HPu*aVlfAQ>@S-jg2lv+=(7d!F18hO53o~vMRm;r91m^>
zj6FD)I+!7TH%<Km(OC+S!<z$AHIH!5CVzP}s%BnHI=s#P_Swx8-)*TCzf?P(=*U>x
zWJ3O&D7T<oLMuM~vFn3i!M8&nY=K9T1)!SPEUw}eYoHVQ3F0X51OzpEo2B`lWCjO+
zU}f&jx+3K4b4dE*wTXU{W$bX*w`Q4v%%IsCI1$=DsIE@Syg@nUQTU0P_kK@!7rKuD
zO+{FI>ba0}`(@58Pkh546J3Wt<>Z%1V;0DHxj38Vnr9xV2q>{T6A|l3zB&*R+FULW
z&dB9WXoA%sB)O_`N6-_0X#3M(=5hyj6AauYQOc9f_-R7>vxT_^p^TpMR8&wliArue
z1(6LD1<h8C-skFf_w_h19M}{;ksln}SNul_ib$K{0PSkcnZ$?d7$d{ZVF24xcIJ)z
z)O$1Qy{t_8?84;LegdF(fbuxkVd>(6(NySTL#TTosHerX9CP_&n&M;AP{k^0@hPcg
zXKQL!Ny*GrQ)<=_<FEUEZu?w=joZG@+`>w)W6Cw={r0jR<LBFI)1@}lP*h|fIUAek
zayDbp&}%?Ze2N*6oKl+zxhf#zS|{C9nv9Ip_?~=_1?Zl7Dvv!bDDcDC6Q@0PP}#0f
zi7IneDDJgh(8x~{qwc03XEkm3TCPkLdw`j?f|*n>NW@Yn2uH3SaO@tAeH$p#*%Rl)
zk8SvV2)7U;?~#vM548*#rnph;`33t91pv_jZ<-6<`+0;B!$Z}HaS8WS5Ik|*0Xmml
zZ(hQnsHJ|u|KG~Zon&qvz*pm<Z(DXD-p%Y9D{f)#kXZhHFPFa{EX6a2Y6*wK!FP-^
z$Ze20P@Tqlt42aFcO3s5rb1{Zhk#tOPpCl(@ngf($I{#@c0BmPs@JMRhdhLhYgTIt
zf&C57xb~OsO4Z6R^GS-cj;FyB;MK;H9&lzaZzSYpqFo5V+v&IcpvHVPGOHPDpk8^*
zy?uwCda6Z`@PR@Aejrn?+RmH~b@c?pc#yAe*0wHjk6)_H!!Q%_wwxz~8Ga$YYHYg}
z$(Xl^S#h{Qd|=Ee8={CfSC?O!z`0(MW{U6hyS=yRAEMIK%->VLUnCtDDB3coL)Lks
z>Si(-C*fm{a&JX9)->xFci=<Pbp>NO{D`k|vk+Rn^*MLln1+>aYWi|PJlDiusMq_M
z``~-JgQ-Z*Y`m6i@pUXR{i#`v>NL4JgHLlw(;nOfbI~(*>jZP&8%9hGDle4^)sNB}
zyZ<0n4JQVy4&&hop%7z@3@s&F4~z>6q-V1L%S}`f_V(7LO4d^8$Ln9pf4*v$%~TNA
z=|FbN4O4L{yD6N)nSv6>iM^O9A=!8xFRnGGL%Dz1g@Tv!T>OgQwwoqB*Hc!_R)-oq
znxF}_vwXrI;*9vPy74G781}hh-|zAr4(tgt1u|rG7;+A>%wECf65FTy;o-XwVV#_C
z?UYrxyQ4`5d%li&lv2|v#!xjWx%hGAv_ZLd@lX4Iitzxxe<y!+l|LZtG|P(U-1rTJ
z;7jeoL7&77Y6mJJBdi#0JE%g8<GOQ!qzIm-1$=n6iV{1A<%?j6%ZWDUj*_OiQ8M5}
z;cw&R{~xmO_XI^XhtCw-qDAQ&{F651DU?+o@MJDQ+xbH*85P1AwTL+Lc!z<Vom>Mp
zZ(F{*QfI<dQlFk?Slg4hd9*py)+VT;($s!vKu(O!M;P>UbX(%QGFuq-10WqZGa4Eg
ztTqoa&^$iB=Pe1<eoZB!(Si8nhx+$90-16n1e!Jrnr2rQfe9<98JTd6b6{xiuO(Q2
zjdx>^^;_Qe?%z)AqgMwy;vDJLWqqganm;<h`i<~eXVlD|q<0+^jp_dP>s#<cCbuwo
z)+Yz`#f2CZr=RjB0MdN6&9!WJ{OReZY#A(+-}kvs=?Mh~(q73h^h##|Y6O;RL*56m
z>Z)qgK;EhGb?tmxVJ@-M4;e3hvKRTyxicG!Fec$+=AyAtaa@lC(QmW8-n6Ktue*L*
zlkNPfDox8>kN(v6xgAYu=Np?)vto8VW72GI&>)Lhb>#hw+2APpSwEp<C-zSv>~$Z@
z1mV<e>i#Oa6G5WmkLA<EOMI#wHU~|hf<ZbqWGI2`KtOegG>=Hm1=cBo=PsO!2AhD?
zro>V3vjtcdGnv=;1ain6DK_?(3iQnaL#hSO$~W&n*uT~mP>si@V5NO~tbdY^sm;0a
zrBo{YipMVoB&mE;yG{TantVXVP%<~N6cvZ&aO8c1`dtBAa9GH2FLhrv1ZY$H?Y+}d
zu)BXzSeOT5A29w52ILeunS#3cnyM<%6f~@p=vn$R6|w&Lhh+?`K4-lOVy;=#+>cO*
zk5&nh+|^-_zlCgmKxu{900J5+eltAOCp%c6&h|v91jLI3r%sg}G}iA=tZLr5gOp|{
zvKYC;0#rQugRjqfdqDRernGA&_VS5YRaF(Ib}@1lZ~&y?FSa#Yzjc}OmHi)kPft<M
z<o<!+WtKCMd?40Bjjd=C9f0{)HwZR0*&tynEiGLxH5l&Umb>*IxM~-3&r1-jc#+d<
zLm*b!AQEB0>Fpwm^9&*5Mr_$^Qq<gy8m$4U5{bc*=A|+N@1uk#Po_N$EAG9PV)zHu
zk$Li5yUYSl1~8$4p`li=G!CrqzYmS+oO%_UVpE_yLf*mv{vZcuXrq4l>eZB+=gV{d
z4gfmi8|n#a5}P-XTWN)@p4%9>vorxdJPIDB7`wZ>UmTi<P&VGE=l|fflI+(c{@-EB
zGNP=H^Fgie_`RReq(s|oEM3$&47Q&=mZ`Ab(k}KKpHXa8&Ynw`9}1&~cK9v5{Qdej
z8plYnq)VIp1s#U|q|bF#eaR>I|2;HO@|f_r*I9c}$H>9Qr#F@U?W85w5Kau6=&>}0
ze`D#0A=vCZZ&xV>nEYJxRQ~tTxdYAXTbSO^Hx1;&EC69XcQ@SK-3R;o9Y)kauNzwX
z<lhn26#{eWFwdU_5oju_-aESs^syTN?;uYFq>SEleGlU;S<3&AZ0wj5@f0KIkYvcW
z*1UUHU7UUa>@&!uz^|Zk@5d|ZD=yz_>VN)!VZNbNRmb0mbIxv8(jybM0i<H>X8fxn
zcUj}7LFQ!&vt6EvEEB*N8)pBhZIETdvqQX3aaLz%Cy5{1@%7Phuy~&b8UloiDEJ!*
zaqWQ-#MlB$oDGn@p9?^7;ZjGteWMD8A;z_)8pqP@Ye}-D?(l9@@qiGK6hNGS(vGLr
z+^~rk18{Xv*Wn;osT3w0vh4=nr^8dju4Nm?&!prk<0#EeeHI<PH5E?2H3S_hgO(Zt
z)N)tQe3g)pu=lhDpdCr%fd%@4+=)P{E|9H(*XD>i5K2#*|GO7X%QvoX;e;3)>6|HA
zS_eBKWDJb-!{B5<)PUmQK~xIvSKE4qG=E`#pC!F1rudFufcQqBvjhSL=3VUXqJNN2
zyVO^TU6-!I>;(He6Pu71=mg~o!LKl_Q63q160xSJ@B>i;_`L_1XQe>7%(s55R)CKa
z1d>OT8K`MspbA(f21<`t%+tgU`oLTl`R|n1Np`eD%|LYe`}c2NP=<>4H3_QvO!eK4
zs?)KwWJB~Dg6>2?><_=oLXHmx<b8ms18OD?Hbh6wK_rsx02e8OsPBVO+|7$Eh<$>Y
zgi*T0e(~ZA*r7-b)NuX#x}&GD6AoR6Y}bEIz0X3HB%lc;+3|MvZQ=#w+Y);65wV7-
zCg|M9F!Y>14V@Trp~NQ{C<)d8<4yrXaxls#ZUEL!-q~4T{GI&*Bu<_}N?f=X7H|Mi
zuzoFe-L&tcIgfEXv)VFaZ(Xl|Rs$rOm4@yC3Rab;B_)rFii&I;-U1$qq6m=#KZ2RT
z_GIU$`+H?D>H--xLYZH6`%D!8AO@_8q3`CcD`H}xPPUwDe}t$O=;a89ZjE(tR7SCm
zTDNX(Cffe+$2e|$p{<d;-Nbe2(j{v<I}5OdLZ;S8_K8q=s3(aC7Bu7U05}jmJ;az{
zAgT+FwJ`kt@bb}?{gDDYB}$wpq~Z-gV(8VWSle;LgCfcYIjDe28emXpFHv!u$OJH`
zKGWnMXRuE8`w@@hx;ROjBL)&vf(CgGs8I<Fkf!FtG6zCWOmJ272ha1$@6RH4z%8ls
z)*XWuNJnV-{xJt}H>Vmv-P{AF{;j~!HG;uP)qJ8Jpr1HU#X-=Efn@IR*I&Np+Z^G4
zgMz1z(NZ6I-?>XbF*XaTB4=v|!T<*$uL=?!SJ%tP)Erc&K(h#V{4zph(9-KNs>}B%
zQSf3FP%Qw^KYjP^EX*@-yQ6}x7GzZcC^d>0K%G1YPFG=#H?a42K9v5-?2q?Pfh@Te
z)A6uCD!?np{ao*N5n}YF0hH#fbJjlXnnyZlzG1yhK?GM@%sh0ed`q#+oG$I+o=wnF
z%dV@-05fg#q!LgD(|gX5A!`VT7-$6!BU#+j$&E(Heo_S}2Vhi3hig6ekIPujQu3(|
za22N(abPG~GSN8$%KcX@lf7L$@4LaS44J%8Ly86JID)eB2RI=MAe#|lpDh=<j>`0?
zd6FX_qw9Y*`Zn@-e8WHHT-yLf3i@HHpsay}A9qi-8fFHprcH1}#91TY6}6+q&SjJS
z-^%ldc`litBhzCBA33lCPyVEkYKf!`8i1<coe0>mIVepQ0H!{aWkgbT-~pJg#jwP8
zpkU1nG|Q0E8myMAtvQaQyxr+7?l_+JIMTaiS#|%-W8F)#r*tv0wfPyBlUo=Phd(en
znxuUtinVODCtqIk@TMR%WMyG_6K~{@9>u5DDu>sOAD&e50kbVYFaX2&Ax0pJ?mfjq
zL;I=Cbgf$(Z7cQHad<qZv>Siz|N1_?=&ZK@RQUM8i<iCOY?<w_>Wk6rO+hFp3jRRC
zbY!=KS|~#E<T$R<wxP&fXm`odv#SaQ88z+NpJP{*hosj+)+hIKGQy~PEaLUKam$5A
zw<kA(&i@Hm8Gk1rA`-?Ia(L*6?vdZ>wqNYKx)oc*(a&y;+b{SZ_8lpfXk(*H)^8{^
z5*!I<;ZGwNMn2G_Fpp-Z99TMDG91^99NN2866FNMJ!&d>ZC_~QKQz<w4NuEDC{dLo
zSX6;V>-RV1KT4h)br)G%O2l6MYM$e+-PT@pm@KQ<){Z%~FY@_JyY~o%$Imj{QNb<R
zl(J^l<u7kGiKAsXu72YT=KYr3X!8x*T<v3BcWFfFMnc9yrhcpY{XWw=W_u=LK~`;z
z?+XUgnnMMu>#VapK>_Cl-8O!HAsDLQ7y3p^B7|Ed`0imusd!4v`be{yl5tFP#oby5
zlVLuWMXDS@--}GL@n5Z`sY-8|oG$K8*v*+*WUc%o^)<xZY4ynU?0Gk_5e@xbxeiG;
zl@-&A*1QG%s*GVeceI-o>#t8<((M!|qLnLuzMH^(dacPUxwNB1=sO!mW{H;gxVZF0
zML)Ym+XRM9)|aMbAK|VIEULV<@(oYR4Tjmmi-AIBy^lu-7bFTfjnzs*os!u!LSDR<
z5m0`f$5lOhqV7g`;PbQy+c;6>yV@p++)o42`EUAk3iKS@EMryPz>@FmU#C7jZp2i`
zr=K-fS>*Y;m*`ioMU1eNBVA`L<;LOcroS|Xxa^aj&l!IleVEP~It2VGr?B>2#jlu@
ziloWG=@Z-rbCUovy@j>#jAohkV2L(o!`nt`iG246r-*3&ya_MpNY4}I{&+T#J%x5d
z<x#WP&$JOWF55NbI%N@w^sdyK&b2T3(&yW)lHRmbT{O0`t~hZ~u{(x=b&H<lQ~9%D
zul(|v65i~2IxR`_fp_gM0#cRnASe*;N%lALmqk~bE(qGlJ|~ipc@&~dbua1$@y6@k
zz9vug&uoNmTRUo7*n4W9Cq0fC$+|Uii>}q{nuP6itz33{Fd4=MLEXrrsrRcKLQ3!N
zr<7GJTn%5-vZVJenNY=F>1s5QT3ATpouYo(xxwk0X|;UP$Z{Sd8{aA)KO6tLO@!?9
z8apNY)f~T!V4}O%mFz`;(vVTEj`8iH$h$m6bK3VSe`NVllsR0+D3a`$E$~Y{nj<mk
zr^=&NE$}dQk}_}%{G@Zv`jr+vB|(nsLiJG9mc*stR6hlU`%W?sZ=_CE%BySJWC^D*
zS;|V<C2r8D>-%f#z1@Cr)DV)b_}cw-caQnC`&0`O8?%O*_KO3eKWRo4{Oh9>9AzCD
zaZ3sM)K?awC6pHjJu!x!lM_@96K}{KNY@ZGX%Vjz-oD1GOR<QtME5p#v=?i&Ux^|$
zz$u}C@|(kCe05a#ZK5UGA+gtGJoinXXwEeyckB999f{n=miWHt)BbDO@z%HRCzLx%
zL<Ysr<kJFMFgfs>nlZJOfc+I|0QR*(y)CnN!pm*nqm|4iHT(Grnm0uFPJN_z!;N|W
zMoM|2LZrSD9^!R<1+~!b%NN9X=P9->P>73(KKz&(C`<DGrnbDy$?n@^zLURBtcYtX
z#bEc*>pryF!?R-Vd#*n6C<bnIvt@#dirPt)YS!^r-W};zv)O}1f9OPh(vh{bGQ;on
z-rF!=XJyg|skXhLoW{vPOu87ronb1HX2MihQp2wz+MA<MZqr%1uJ84L>CyxGR1w`t
z%{BbkTW%%gA`ObuHi^4oG8t=4Hc>q}eI?bFjAakV+kWsn<Cip?5*(R4XR$QG9bp^;
zXInySU6tUw=@>OI7Zm1{+~$SbApa`lV`{hs(Gq7iKHS-qfKz*FUsId=4Qw1=acwwU
zX`J`szJPmo%Ug%ar6h(tr<3d%Ucimde!V%8#(Ml;ikQonQZ_Q*wp?9SWZNO};<MMe
ze0{CaJU4cA?t?e(j!4boDP`L8zsa8|Ixf4bw(1A9vAR8b@+`CBy6w<nkmc9|%>?Ft
zoK=Zv&g@jd?|trM_kuJDas(%20^v<{l0S2Z%KjF0`!;Xp&E1U5&Gd<Td<8X~-(L99
z6H;_y)7~dqvdPd8+p-kRj66ZFspVt2%u%~dnM2D)c~<k&O8vXs6a3<gsbo`gqu($?
z?wMkb8PzchF)Bx5b)V{-yD3fFgE?#@y^x}=VjybhaJkhcW94(5U@7*hL=We5`RFST
zIg!ZII_<Oh9i|<aUaEnnswq%xAv^TEL>1GJSf_DjW^sG%^U!pHe(4`{x3Y>~ImSJ*
zIJqhj-ZN%<r~YAnhB{QW2QeIJ$<GM*teiT3sqi-@*cS8tn7xzIvmw)pIW;3TJN`Xp
zY{hU=-X%p`uF!F7xUHmU!th+7qHW0S75@fR+mJNwAb~#q{t%}GHbKmj6ocjX&dzo6
z7mZj7UB<H4Za<k-+9(m>pQ7KeOQT{rEB|Tm%q{~Hu`Iu3%9EvMv~%G%YgSWAi4^p-
z=2V&$TY|eaMagMv$`rRMT1WKgT0GU=^^Ih*vo@UDG=8TpSpAw<vt`dqj$g`8eCtud
zufox(rJR;g?zzf0R=&X$h36`LSO1yKdQ`sF@~db8SRO>%wcZuF-F;1{Ypt1dLp5#1
z^4aAc!EClW@(^m*JxG(7+<9&}^W3*jI?A^?ivDQ<F6o$c+SsmG3Z~3V%wJ9)RMGM8
zz^)ckufG3y|GaL4zm~U$laKJcNUUqji(Jfr%%4pb>xvw~3XUq)LS>Eom&a$M?>`uI
z6i#^;#j5pbfb|r?K~(ugxb|pHOa^{TC|R}Eu~;vy<8{Y!pX7OnXA>vivXTl7QKh=C
zb|cD(lc(EO%H7<fEK_y)^ubWsr^_9<)Ug`h^t>Khrs0b8Co}Yiv;>d3q|jcV8~nu)
zRLDo3<S1(_X>id*kH346!uv{}+@-6GQH!x%bFT6~xU_VF!f!`=1{<2t@jtozJYO|c
zr*Ul}(nCum>qG>;dCG5UlFPU$Doa#(_zBr!oc-qO<@yRTp>wywZbt-HH*{FbD}S(?
z_4pKBB1~OityTR%+_{mP9BL&W*2QZin72J)iM^>G`8ei!?S+C9oS9fUdCn7w%GCC^
zU$7D>-N=-t>HcjNmcsYDyNp+9wzh57Iv1Z_{hq+k7SD992Y1=c>sW0E(G!H9q0}j-
z>*#m#HRpN=jgq+ug-~1H6v-D%%Y1r$Mu512bySb;!l#wB5yBiIU5?Hhne~+_YN5F>
z7jIgd;m@<HPs&3z3f#la#uq8GKdtwQXMJk@{_fi)91Z0=)AW9E@AqU|kMzbnALvWP
zM*aLQ{(iK|&a07$Yq<XBm*TYDYH2C5%jUbrS2~9}CSEV=N+<R5S=|3*8#(z@KB?nI
ziE70yoVCfpq)W54rjC9>b*$=!Cu@_{)=g#^*b>fYYLcC6sk~oYC)BK6revZOvvSVJ
z?NhKU-{afYGcG{P!yL}_I@3G%AJ&Fz({8Q5FK{!do6TM4i!!@KEXd~Ip(GRY$Rc>M
z@}X?@lbPhotj{t$d=Y02g4oV^>ttlLVNvvzW^ZGwy*uegRld5f_jw`Ar9m{<bad%Q
zGZjvmB0ZgNrf9I;)^MxXfjeLKBa&ym&Pw_U-)stMS%h*(TeVkV$#v|fPdeS0Jn|6!
z$@4m$x{ZRbmk(`ffQ?ramDS)b_tONX4DQGk`-?8EeWB483jEJDsBGg4n0HzYa!fg|
zuT(hk7YYo#cJSF|s&<K~>k#3&O=6PJYBjNFNuKsG>5`+3G4YFc<@%it_aA$WJtJ)`
zFHI2LwHum0%l~w2-uM2M&RGgE>(O$7R;`x@HoqAc`qD2D%gRn<MhObsR<qf^`VgvD
zHqd5w3fgYjFW>D=bvP3lHX_$J<nquU+PJ!D%RNSSDEHBYghu@Zv9+5648*b~6q*E1
zQEy*H30>^Iw278&wp2UGgS;}lNeAN^se}E#jYS#-4_>kM!Os%uHzh;pqOqa`FJ=fa
zc?a_ybPE{>Y}+K{Q_d%+tzH<w+z7JUb0v<n%qm=0@jV~jBU0l|t=`;+0E|+H>Bgjt
z?XdR~P7A#mUm(j$MMVW8jzw{yOrJY+I}>a)@Jv-KSBudm07O2><8Y$};s$L-iUX0Q
zz{ig_pv63JC>0?Np$1L5U7%_rY<PL=)y_zpWGifL4@!TF4AXz1GBrH?KZ$L4j&pQt
z^NFBq`K!r0b(VV$Hi(Q7V;`q)ze1Z2|D%Jy=GGstoi_*2Bi*s}uK>9v@~nHW<Ct}U
z<GAfRl<kLlNs*l+_*h(FWn;4zO0xl>5{+GDAnZiq&8dTB17R<_dwOtVcI(Q2ewn0i
z&R-LkEr)$5%A=Ax5Cl_)uUc7ql*lxq>Sy(m$JVGN(j$UA?`yBvDz*`9rxExLQqrf+
zF`(}JI6mH$6Dvv30QTE52;-pl+c8E$9Ay4Ixt{p$7dQ|nFnHXut@$5^MZ@42k2uZ<
zCWQ3>nAQMyONF)#d2O44w<eNsdDZnI-82aDh#LU6&;jHex;GiS#+HKi_x6yl2EvQM
z2N&xr0uPa<2uzYSEt&lm^esj6{%4OS)q6wd{?5?5`X@-R2q|KcAdGf~>vC=)5-tW{
z9-S>n6uS`UL`w^Pu8xiPZ2dMuBx$N{pr~A{->!?zQ%-V=;m9i<GO_j65-xnpS@53|
zE2-4y8{rIG04bTqh~{swVy;R+!#*;qM9%;ifMUP~k;N)zF?4=EW1j(I^~LmABxM7v
zCgYuPFE8W>iaH51|K~XCH%<41A!R5P#g>(n@EL+@K(F2fo~k)=r0r^LYXijExpH$z
z4-<K~XYGGk*a35NpApa6Z62}z$(bZQx_!--YX&JnfBoVGFl&{z1~3?CbjO9^z=AA*
zK?uL>BP6i2tRe8>@_z<0jE6@e_ulx?Qc*p9^ClLF)xi`wR5n{0j59&pE{-hn!FC%)
zB3667$*}_~>CD29=L%V%D<emu`%j;4E{q`kXO+HW#0&EOKK4fVnk_4J1Hc}Y%cWh0
zZeP54lsW##u$0XCrX^$yh$$?ji!E1TxBqu*njigsIkLR;wtOME)=}%&1b^mi-+i_#
zn->|&zW(=FzqRXe%TpcQgr!LP4E_aE1-gaw^OXPNKAQFU#zf`O*^jZ_FB-Z_q;}(_
zjA2wEWj8{Kr8a$@C<QC!IgOMY>~A0Kv)n%ZLfJnmkWDG@YjU>k+WiBHBrQo$L?X@y
z5dSr)!?vTHlm}+-CfsO{8}EBdr~kRQ_mCm#hoi)ip!w<EWkk?G)@K_iqQ9@Kw1^Jx
z`dC?oK}C`LolUZK|NYi|)G{9!7?|bp^HouG1b{t%icL|;pj-85>)S~D?{b?N;fyG{
zISc>5aBl&L=?8@2fcRM_5z_&(ElII?%%_o)tl#Dx|0kfU8-(vgIf9eOSiSi^81{br
z{CU&4j^w{_06y)l2U3BW5hyVah1h}7`Ud1sqT@NBItcL58f0eTV%+B%(tY{)A`A)!
zJdg45$=0cZi77`&wsQsr`1yG;vj3jGR5;v#huicE`U?w4N(?9onA|FHCt(=@z)DFF
z+>5i{N`)AvM&3CDgdjH@lpop_L-*(N-GkG?veVDMME*C+cShTtlnn`@s7;B-{a=sT
z3`|Y+A#-tLG>(j>GbOxtW~B~Ik>>*7j-N%_c6LNi=pORt25=hLazw?%I022}1R(eM
zmYV-5l2Xyw9acQA1Y3;?=fSNUl_>wi1^>UOO`*N-B~<kP>MQ|7e17}YYlE6Zg`5v4
zv0$Ufy)F`}<PFX<f5|tfq`>4gQOT8P&%u$JMMH=9t<{f*)Xe7}uBTbVTv|tBtHA{W
z#^NsAooQ@z*62yT_mcpi32G5?rNzO)*#XHdUb$t=!EV-(Xx2_M+fK&RE*)A6CU;te
z>D_v_A*&P-S=Z(!|7VmOxxeXP(8<%x)-B`*-F`1;QHt=bdd&bE6O#wXy#%Q;!7ZWx
z-diidbK+YgrpT-W8G<0%5&WW;qN1Ze4i3uM+w=TCt$k%wRb99)7+?S*p>zmRf;4PG
zLZzex3{pt}5s;KdMd_3-Q5xx#M(IXEKvGHRQX0AM^qkMnGseAl+<V4xjQvA}wO7nF
z=X{^{3EG`zneE}411&UIA&3^52V;7(@nsS1jQ`$12RrAE>lT$;SSVW{B*^83_+Fv-
z(II(O<6q(e0*#7-qxHCAtJsLK`sowl@2CvZm;&hBVNGCW-n;L;kB*C*0ss~J>wC9_
z_sG-(bqqwXbFaSrus43_hGJF0hkb-LR^abs<I3_XrbI#1q(xrNSYWti)DgpvhmX%K
zB=qGZgaed^A8ykgG9d4N)Qv&CRocD5CU}G>qU0zpf$ak~#}%~veSR2@$fiC%Cm|^K
zcXt_rANz;ua@=9S*ysE`aj#V}JvT%az|iPK9BIg-g>8UrZy53_IR56#qjiYJhzBHb
z^}PGqd1Q#NKbP=}6)nD~e<C!O9O^w8g^W2K7W@8Ay{zRj9LUiTg$M+!-#-c-!yT6{
zfB9<xsws8T^DyibDquhQ^#!rg*Sl`hcelYe>J(~lBhL>s(y{tA!E}knZ9~vK6#0T>
z=2rh+UZ9U`-AHbXwfbwTCc%9-&3yuF(Aklp5HgJy0<&@$1^)j2wmH0ojQ`HJS1axJ
zYb{1{iy^vqaF7i3ao{#98PgVkwF3B;Kdc6e6y$o2%!6R002iJ0(w(fYik42q2|Mk^
zpiFynCo{x|yk?EE<|CYE8DPSw;}ofFebN%lL4*!|^vd-X*^EY?|2uqUAn@+(;KlGG
zH#>!3t<#!_G9pKZD}_5I1n7SHzycoyJ|cJ{ODbrKb=2&))vQ9WO#+C&UV)KT0<5RY
zz((&BLXo^kJTV4w3&@ZIy)@v~v8zOZW*&Ib;7e~myCVBn#LGnS>}_I32xQiWc?typ
ze~nvcKj(ZAT|Rn*?psgXn~=jjG^vz|HK&n#I5Ieg$&w1(t-rxbKLxhaQ&lUjoXAxO
zO&uDrf}?&YoO<PwxupT@^CU!MA!4B(8lEUt7WtlapxqE_TAVdw;lI=9RVuw!=anG)
z3)SFqu`c8OLni)8JwY+tbVPyX;pYDMPgVtNuGakSPa*Nn4y+AQF{v;v(Mph0HbVg$
zpsQjDvtTZx3q%N%f@2JgQh3SpZDy5_Qw>tWp)Biyg#ds5RWNgjLK)R?R%jekJ)^>m
z!2F7A?Qe!&+_gQw@to%h+lA}g8+q>x@g9)c3li`|_vMe}=M8oN)MzJ3YZLc2QFM~c
zpw*bKs$n9d^jwfqwZxNKQ~o`wt!%o@{0>Fse!tS&C(?uNk*eLoS-7Ifis#!W!e+SG
z!dB!Wopode*QoWF!fx1XuS!Y48J?QHCWp)$Z}9z26-H5q&2@91jUnrJlL=L=ajB5c
z9pCOo62I0$Ohsu}s5TA$ts>*#7s3%S%3ng23e2s8n~&|^=(4I**&=*9C>iP=?L?5x
zSD18}sepo5E}NIZ>dr{Re!5pfs-M`^%Jsz!4r!fGb@FvKy0wV%E|bzQPx+fSx{FSR
zYQ)zYB?~a{VXMh`ylimmpPK2-zWPvo^9!3=iBA)8#?zbRB)+obyNWaciPIcD?*R%@
zBHz7zeeAt4!<{grTr;k?+ZkV&Y?3UT>~apG<z-!@BLg+w`TY>`xSi+q!g%QzZtp1(
zrmgNs&W-Av;{Dka=egZ_%Wqy=H}~eZcCyo~Z5ODnX}rABp_0>)JMCMzMO1TmV}FyG
z?6CFDuil#>hdPeWj_<h5mo)f&_MjTC@>ZL4)c$mnO!qF?>Y0Zx+^N4PrQ;W_>rh1|
zv{A*U&2d!pSj~8iw%-==R}HJwd&Wx@Y50AvhhpOUDOGF8BcRz(zo~n}c}88TMQ9#u
zy@FZD1_Iw%7d^h^BHfYZKZK*ns!vLoXu2JCa+q!-!a|Ks)e}#q&1lgqpfFdU(Q0%<
zgYYL&B6+c$bT2DbxyKPK{79~l{`3^Z@}vt@difg~y(xX>r<3@Es8s~G&*J#aO5FL@
zFeG`En2U3uEnkmT$Vj%5p&>-glTfiEtuol3C%Qj_(T+mns=iv~)(fULWHVfvlT~R#
zuG*|poiDI3EBRyG!7JYaUP@_4o!jn^Ii$TsXgB0DGn@0xSA^J}kKm@<JV~AQo~G&5
z8TQVKguzY|Na<5KVX$js6frZyH9+F?i7yXtn1XcJ@V;4G{euy8?Iuexs{_5+Qlru0
zN)w*lxNr30l0kuv=_P*gp?lG4i}d;AG|KeetjD;eO;dYth_fk9`yNaw*7FiOYuubS
zHgdacM6F_VvP;lQ2;iT5@@bZ!JH9P<;`<(W>5Q8M#9E~7t{VZ8A#F+&$=x5LUvOJS
zx$Bo*uTK0HO_0pnRfqD4Bcm!Ij#iwfpI?5k5QR;vy6T&{zL?zNXMf6%`40AhuuOt4
z7fY&3qqn81%}EB0@NPq2335nU^lAIVqSV}BRN{;IP7K4etIxAJ);pz4DC?Ay`qYu#
ztPAU}m)_2j-#6ll|2mi#N&2q2^_I$KUuKUNClh+4KCoXcuT1P>E01HC6}`i1|1u`V
zVtiIIoO?zM4#rs`Ep#(MSH<*P=44qiCPvKQKF`OTc`hqLqPeHmQE7rKA-(hl8sQK7
zlP@(Fvd1*!W#@}We_&B;nMxmG!jH{k8R&Y-`u3@L_5r|0SWyr8)yH^)e&s35bKq(c
zUl;zqkM~JL$DRL_U)sE8pSIt(e#YV=pADLoD;b5!0(FjFcPN5nE{)0welpF?w!y4G
zluvl0!r?u7|CU&9Ar-+<CS{eEVYg)^$*+=-(=1$(W<51_QzGOmIkSOR{8U3f3$a{)
zbVM*OeG|zrZ~LM1U{gUGwUZKO*TFR7q0X%1n=&s?d53Fm^i^!hQ~0>DgR>7swZ^+X
zHZpcP_ss2DedkhOm4duXu2ia-mS}G&E7~T|2$B*yc1RKtl;c!{ssJ<!z@a?E?-sGS
zh>7|1MxF_|pYHGrS7)D7XClW|&@2&Kso3#>QDBHf;V_1#BxzBxF`Dw{<Cy4%bp^H0
zT*=fGCO5Y<(sD>{J&j)OWpeck$!b+sJx*tNAuzx=zdr2yx@W|?XhcAPowZ=w4I3fu
z7!ZFmxwHn_`hRRZ(JwSJ<8;EVV#WW~o)AR}%$hhIValIa*4e(k$?je!m!j^qWhK>G
z=`($LLo@9Xd)@^>i6CDSeex1J&t5tEyJz%uDz1p;n9)FIDKRj;)P7H(SGue7iH)ii
z{b{>rzFmXyiaf8BxYa4&hNsj^4qjGfV&uW@V1H|>Mqzq-<;qUpXyf#A@8dgSF_%*!
z_~(LJZE=r3ENQrDgO@@A^Ka3^Lx#s@<z4Vf#|T_3JhwpoOn{^=DkuFzR=S`1aTbz{
z_s~5lR-7+!HTP;_Q5?IT+{x1)bK#TcSbF_py5Hsax;G}<VKMTL9oRJXoR)q`*bUXU
zYfuQh4Y~NEUaFtue7@{??7<aE(Q)DCj}Kn1Jtz)k#(FD2@|-M%q+jHrX{ggst*{Jt
z>)F^~7s3!hl{7^qZRvFUe1d37dmGPasj)gYbMkdXw##~b{U$@lp2861h<zn2e@Q;S
zlje?P`Q(LBm1hx(J6$RKRFMvi0W%i7Er+k2o;jA?lh!0<(&T?N_sh_6OEjXfP(A%L
zU!_s=XGi^qf($=fZAE&Pui>1!qQ2sA&6GPWlg9d9zRvs7v`tZ53S#z4UDNF&7tWPF
z6}ofQSy_km-U)VbThh%`;zd*ciK^Ms9+wBWQu=rdIUkd=dBgcxqR5{56g6JYl%8P+
zfI7tK<um;=2P)(%SCU#tU_JtL^$q@*b9wEN(~0+HVlG^2>9r)a!*{l8luXDQ?_YIz
z5Pjr;t@h}(9LdkcEOSxC@7m=!zHgEn2!EZE<<3$iT8Wa9O%P0WkS;Q+vE-ho;M2bG
zzOy^ip!~M1cj8Nf1hYKO*u-x{nwL{d)nWwb^EHe``Fd$-g_%wkeKhHSmaWZ%1z?Vs
zsgqQ0xttccl9MsWbWTR436?WIg%IKb@4>p{XUw-f4stq6vw25i3@dM^i0LSYt72BR
zyK`%#DYGC4n99!=)_!}vFt-fl))+>Cg|OIY?zN%Xt73z#R_nuB+tRgu<J|k>a<%H3
zM(Ny*w|c(+Fe>BH*UL{huEOam+$9=s(4JNqb$f7BRQ6;2l&x$>3Ba&D4(FsuRkz4V
z2QI9U-yU8p3dd2LDD12vOA&E394#4|loy4zkwKzZxvle}>KbdE25C@QIs=Jsy;g}L
zeoql=Q~2yzmTRNMWb*IYeJWt&s6n1XgkoqH^f%a7Vw#Sx>EAGszSYRkg1z~k2^vpx
zfv3XiM19(ZmQo2?Dvfxu;Rjy`>$(mQIoY;{$&(e^+4pVg#WT=eSGb6u^(5Y2tPe+~
zV*D1%5rd)0E3qL$ey^7|p4GW^#$EiBy0OJ;))LpCpj;b0M>#<kM<cSQ6rZrO|H*Cl
z6RD9Gjas`u|EONmgT_m|Mi!~rr;6xpmWrsHs%ftrDpavFIr{ei>^|bcc~nKP(Dqqs
z7g;_(4kd%XKq{A0IgMh_%ig*DlSQUMK1DNwJybZgmuI3ba9EP(+r7@lw`=@zBSWcT
zIc;Oaq<4;4*pb&&nb#mEdIiI~Wn9Hj7|~~RY{>TYbfrEgG?v;=mv0Y0xn4i8PcLTP
zy?tP|9k4=CGT<4dF#n_RqGrwkHEX`co=dOeW>S}tw8@byoA~m6+K6c5)~WEf^J<&Q
zk$d5K;+Be<F3Af*%T-(Vki}n+)9v=((Q&9af40fAH|FfN>Oq?a_6?m{Om|^Zn<jo>
znwpx5&<be?;L{^h&HT)AZa97mC`jF!8v&Ls9+z@Tp94}7n#zhQNwa~R6oA7Z*$kk|
z@v@TQbJ=!jH#>nY1c5M1&Bo9Fg(NT!0|$pWo-B{FAP}Oy{uq|rD_gO)X={fSWckR|
zXuC;J0W>FdsOksqyI%n~xuy-_VTZr=$f>D~8hlAjvH_Q=wi)ElDjsp!=)gkS<sjEZ
z160y;$_K>;Qc)ThCj6DbG*HGM5c=rw*8#HTH2PNiGS0e{^}334xn1Vx3xNixED~FN
zb_NMzk>e)Hr4cY6s)wr;nZ<z08_7Zk!4VkAfM*lj#Xx>b9U2<S7pn-Aq;_Pl2SnwT
z0)HsHfycJ*JGW!y9RC-XHjuz@j#U#b2m<(^q>h0adKIz91i^;ycX|-`EoW)H4@AL~
z02LfTAzdgT63iGs2{nP>6cDg!z`Bor1riP2z&&RUR9GR~zR?voK}*Lrt?lnJ;x`B@
z+3vhL^#`N!zK<otE_+VuEcj2(0_yV>g4UP<4a9a@lzA|iX5<K1b#E4;&L)6(gR2W6
zhv-H^&Q7q=k|NRnFPz+QR>$Ul5Ki91KdWh`p~U(xKI8U4hVNAo%6}4?iRv^HsM<pO
zBVddnpdVD}bzmpDGc-_*^8t8--ih82l$C4~7z#ce0NMn%O%sT`=f*K;52+ykbB9i>
z#aLbUFw`_dSz+MX324`OVaxDJ<irec7db60)b5?%zybR>@=kGCwn><T^vG4c*|jm+
z7Q?>+6bDKfs+rcn^i|ivAqsj6PXcvpyQL67gq$tm1W;&_y;y9toqc-r*O-4_2f{;|
z!0k(PxHvcv4mgQoW8oX$-ekGS7XI725yq23nf8OVi3Y5K+`ob;F&6@Fkj#S{=C`jO
zJroob?E>Qg#}8|pP70y=>-vPB)EmFTM*SGbd0RAUf~HQG>K|b3DwUOt5t9aCk%3KT
zSzVy~UlsjTDtsBcUd}Hze<z$$$cEkc_m?Y|<q(Pt2E}!#o<`23k(?I~tk}X&0qb38
zvFq2tp8IP#(titc@rm?DBZrzKp7el*qFxH}*LGNOPpx?Z4I8UoMZtqOwZB3=-H1-U
zI-qo1vFg<G7h2o~g*OKHusa}5|LDHMc{pv9aEL)>pTI)h(Q5*BW|zp=TA=%TzQ3MK
zf<lDWFG|)FF2E!v0TvSj>3yA$3{}&959a|<fV2ndVr_qk(6+a;%ZG+n-7nylZq2$<
z{(IUJggl{(X{%#NVe|W<{QM0lu@a$jAY{izC-fV`giZ{?acF;iqs9E}QxhrdtlSdc
zXy^I&VOUMIC=?{<GD+Bf{eTUl4XYFc=nmlLzB40O(*lgUoqA@mkKjpPd$)R>)XCWy
z8B91#g~+6Ten@8UNz=feUQrFE<-m37zvh;^7a(I*g0%s8e>;zD&_H_u@{Omp#&^%1
z*Ref|2(5@@iTw6~LI*J4w>XsW3lUs^Zyfmm_+VXl#tTl;f4Bg5@3nwaY8POuLI3r8
z*2{G)=Bk=B<q4@?A4egM$ndullI9@DIN!L71|ijCo{;StPIm&eV9?tG7N}nlxS|2_
z3M@n-MGihNONX@?LBof`(($z5m;(a{I%OvB{M86hUMt6c#hN0FT>!J~Yk0-iARlra
zsKK$wVgaR>U011x#to<&^tP5q+3#1!oxWW96nXC=SZfLx!Ao%{nT`e;H0&t54v3vd
zBCqYT)_~lL<Hw^<R{nhxBL~b;wAtzaz>zKo3-{B~(!2o70Z#P5DiGmgY7X~m(%}CQ
zaNW`(2faj**dK?xxrf3iAbPOKlo&1oVZz|XeAbM^eHGUo>jOgC;d7#r&%peF`(Gnr
zZGNNNZ^&T<8^Lpq&~pQajOtn#0q8!tZ+p2nBd#y_&HfHn1-J{DWK2#@UROExVXcX_
z%UEoKJf?#QEssdggD^yQz%T3I|J(=d;%wr{&Ah7V{{|(?E0-U@9fd~TZ^M0*rwOAW
zNi$oOg#?|wrBiQg>HND)<o9rYyY_q(7>fX^b_yK&7Z+LWfv0N?AJyCeR+3ZkHnin&
zs;n2E{k6j0C(NygHpGIgSma!d7}<eODm}pdH`0sNEefJ2!=XoW>go7i)%awj`%WZ!
zhr3S(3(X48u8yg(kUZ|x%RU}F9ro`_SM%oH6W$gq--Zj84t12s0A{BzO{^hy><7~X
z21mCyRS``d(3)9$y;bS;_ANaK9>`xY&asyr6E3W0<odU*y5_fHNB+$x(Ho(hutY<u
zw-wa}aQfF#{5G)I+W;EPDkkxhn(0dDTAyw+noIsOLD!oNE?7fH4hGXbc<MW#V=#lo
zzG-TT5ru|=R_Wrw!9o1^Fu2J;H<=4BNdMtqBe|=0K|mH~_3ar~Z$QbR4FtVx(Z#|C
zg)Qy}xrgCkVPpjh15FCaU&}VbH#=&^qmUQyHZ2YJDSI1NBNi-Q+mOfb#W88Rn@eyp
zRhA$i@tIMp2>WYwa?P=cRg7phdxKNcO>_TxqaWw2S#C21U6Ag*K1qrZ4Rru&)N-E0
zsIkIMn5)LijAY5IYn$)n6b7%!ItK|TI!BulH$|E8M1Ri5H0nsV8I#yv?@jT)cw05H
zJMW}VM;=&T^*6_*kVbs(;l_?3R7}ScBIXGf@K%thV_JNIzpEhIZ6+Fg7iacL2G5k;
z3p;{LDo4^hzFu1`8ItKiig3pgp8Cqt^GQRQUTVfZ3KA_--#<H^BkZ2#vzsQr>v@^<
zESn;MhYA3hnm6>qq~D$>qVK^z1?<~svzjRSn>g7d=agrtJDsl@<&AY)v--4RahViB
zkplo)s{9;}W2^90Y;FqN<S`YSljmfFntQYLoldMtsA))><Ky~goJjkK9ojovI99w-
zFT_J<X0pG{<>nAI^vv5zD=pa|^^u1Jd0<=23rp=ljq<=pE2lA;>18kCe41VM0HZz`
zM!y~tyeZG7dK^FU)#<BQHbswa)TOA$C%rlH;CRaB%D1^q?5cac23H&(rBTVWsx~v5
zEcqOjulY$*rGF<v<pYmGy~7)_{s`X!#@^}5IgX4c4x%^#l4n`3^%}o;r=uJInPS7!
zw><7F8J}ASetwF}#8)^`%r)Ug1~l1aX4s`oFHjTCKgCky(a9xB2+@>($WEi2d_<IH
zL1(sn%$Z4tIMV>nSB$h<`8pQfB}K`mTem|cC@Z%<$UBoi=&TMS7*@L5dCY{r@W@z*
zkn#*iQcg=0tA2vpQ{9eO;hlL(;5=#yEXB!pyJhAy9rkTDOscSGYTT$}QW(OKHl3I&
z7gU?ce!^Gy!Zr7kmT1-jX64I_mZ%PG;-%(ZYu2|XqaJ-`m2uP*NFl60)*4*uA0mA`
zn=iZwAXE8r9CeZ{02AaM^MzRqFvosN&Ak$8+*z@CpqOS^9yRgwX7NI{N=&;}+*W+0
z@}u2?BD3s|?-Nos0t<XJ-koYM_KdTPJcxLHLNV(Ik!3f_RX!t2Z<RN?t&}l005`7m
z88Ve1ULXvp{U*1@LMQQLh&;%+o%zk%cX-Ehqf{g6HFzvDcEhaVxh*ctxAw+dcGhH7
zT`^^m((#btwHR{bdP~i~mNZ8_bj?C|DTKh=BY5<)hp2cN?Sf8P6}yyBmQAv{bPTC<
zdhnosPU(mBTJ2M8hEIe$uir5!5|Uz&hj)Tx`nz8nplsPKj2@pul*Dj{ZU$~FW>M+#
zy?!{pVLM8>84BQyKJ_aHc}Xu85{mJRyS(O?pZ+9Raor2(uDn$;YD@Xuu7Zj(CI80c
ztvzF`1@<8pqF5kaN~^r*@D;wGhT~`X{!oH=csZf+>Lsbg4u!Rel)IY}4|)B2s7#bT
z6|ic_`$g|<)7(g#jk@CQTkN4D%BAz<fq<z=#-^R{&l{9P!^@lxmhl}j4!M^E<XHI+
zDuEAOhmF$^rfj;6C*rE)m1dy@fQo=>SOO*8R@}4?OVhphPsYCN%uQ<=q3XtbI%mxR
z)Av*~DSeUgf#(qk|8z#?rlaxDi3Msq9z3Y%*-Bb&^}jJC)+bI3uYJ>;#QKJNOzkH9
z4qf6y;3^8bNYC?BZcV{7F(vGp(oofxs~Bk)l=dd66H~uuKRjZ4QZd$kK6dMM#7Q~r
z#>aP3QW@A^ciZ~56e&I6up(3qTwgQ`46P@B#zUgeURfGx9Smp8LL#2EGCeovW|Vb1
zrRQDfAW0H_Mu&GT70tL@s2yi<g--`O&<7D*>a_HB5iXw~FMfKctAemTea`-EmH|<M
z%2xJVd3IV`{<U7WO<p@VeIE0*u~4UB@+aymEK#MskqyF(cI2QWXfNK&jXTBij_^32
zt90E(#>+`D{A;~S$s2Dz%e#iE#9So|9rR~5^AOFft9yM~!te->Sxa7gK}=~@ue8WY
zcduPR^hJjiABCqIy>`?xGaQX(ujck!oG}7Fv#fpP)C}EkPJE*?jHFF_7|U0X9LlN5
zI&$gJ%bo&O_-@5pWSyj}4kxM}*!MLQhSLM^kUIRXU4zXuUQeyz*wUI#g8(*{=9BrX
ztKCJg$8E|n49n5Ju5r9C&(`bsLygc^EoT#192>IQi6*1^Qgtlz_;h@?maJYhC~m35
z$U^6(Pvc#+T)JW!{;eA-%37F}^fV5AYWv13J9M>9GD=d(VoB4gz?PTt@}KjTh_x5@
zOqx-9c$xlYOpLrw>Xur`Sn)7+g0(~no+r0l|K&?>y(C!KFpVuHra=vA@ie?KbHCW%
zj^4*U#-b|vb8fRG-Tr?1P0Wh$!S1E#Ife%mBaei2#)_MCm+U8ktp<MtO0q6(TD89*
zsnpjl>-_n4w8?9zXyF|1L3(4?zTkm>Ko4g6MDbfg>mN~6Gr?lUsTSMOB`Q^XEY|7v
z$9p)7r+HnoHr&d^N|mVb!{8(drLeEb$&e_FWb}Ue^()*-CLpX_K4TCkqfOXIMtRGJ
z*RDx<>#djD`6Iw8?@eJ;U;Hu3=W=7lyrO-OiKgq;5MBP#=1y6LYy8G_Cf(%QdiM<~
zuMpTTS(R<PR`+7J_PmVCbmBYz@l9Sy!y|#0+Kuw~8yzx-9IFc_F7MV7O?H6L0*}dH
z@kd0xQ9QTSCHd$Z6!BsSbCvSc@oSQXLz9=p<KOd|*R~h--LPxy<*i8yySg{$(&Hr>
zJLlG^_+Epz$+-1}>M7A8FPHf|d%e_0WZ5FF7k{PX|7wpa9_{L8rQ{u02y?yXIO6u*
zq53z1>y$(^O&1isECbYX+rbR+vEqd8w{tJ%s&8CU9k$pr9wJ(x#*d^b$qutS=iIzi
z8f|cIuG4Y!#}k4i{?^QBN_swQCnipZxeDougtd5{xvI1|-ex?O0N?dvZ8JQNEdvy@
zDV_<$o6+vZx0G*gz3(ht{j_q7cSppPGoN6tU+BSty#hmBXfM5Dr0NE_^Q~ngN4N9=
z*WksQl8+ti_Y*dkuax`FFHUYh>6&`Z_#&3#^iZC}NxJhMN1oRUu=OOW8l1mO$}#q?
zf5_LQ>SnR(LczH#--(pdl=YfAk`pORn&wAKNYyUgP$7Jjp?)er``9Z9>&v`*2L>|=
zL?^D^54qmCGO7QTceZ`;%EqGcrehp$+LhCW{|F(iza~$6?ERxxyRNAk@<&hH^V;{%
zkG<Nw#rQ{tT$Fa<PqI|<71EY}glCLP8M1!_v8KvWe-y}a&TOy#5xU9_sr@&;VzW@4
zy%B()p$hVU&U{_G*<=NWiXDkFQT`01RUiXCFrq{(9ov;LYBP?;s7;GMK{kued<_$N
zfn|ZR<h>j8QQ_r+ve&*CC!|0ER$Lqv!aZEZo9spf#_vNs3fL=V8iAZ0H*xNtZzWO<
zS+oJ!@C9-41a5=I$55^wy{V~r1Eid^Zrk)9XighqA<sl82tQ$)X8xm}EmD*!a5c;}
z-w4bbt8(Ur3~n%%+yN@ZnGZnd0tXxv8ici79W+^Vu>BsC^>ECghZ^7=2Wrw^N1XnI
zR++1WJfbBxfDoVYEl^7Ss^xB?x(1j1<;xc&=R`?m3J{(SY>o9G_k9l_E3OZj<yt|;
z`K!oCuUSX)F_nnONSS;0?uq=6{pZ1{-eMf~(4j54fB$|LNM_U1)4w(|1RX=Rs7Qea
zzOI2+Hr0{h6t($5Ac+#25agzos-U+#QuYJ946DljxkVS8yEmU;^=iM9*aWm@6@+U?
zP}mM=1+U${O$vX|3BDh9i$w-`Nit_p2n<Npgdmk3GJn9`5aEy3*4Cx_>unkD-uagN
zQSDOu`5KyYHLe11XN|(V!4Uz1@EkzQ`5Oa)6s}g#Q|<sH^#KS<z?!6lI}65~A>rW-
zQ0$zMU_Zye%<P2>cac}@$jFGv@IPnWlb<{~z>!@HCY;jJ3poRQeUHImQu@r7mF^^5
zP<*?8AAZjn9UYB|Ykz+{7y^Ukc9)QD6X1cgygb6E&z~Ow(X=BlF@)eLw6wP$`<k8<
zH~|^oLSGO5I3&78+CO<kZjgB73p{u%2cQ+B+!0Wi(nTvAL5_%?c;vNEx8i*$wVFhA
zE;8tKgrcYvWV+yOh~i&>TXo-dX(-%kve61~xk<+{i<$>>j$_!1DgO+(ea@a`r2?=g
zhWME_s5@oeTa14HoOvEWwV(}TMn;;|)xSfMzzP%;$23=x-NCgJ#f!(pP^@okI073;
zqx3O-^@`2j()*)aZu<_6LA2vDZo|7q^8a`X%#roXC>8V_-MH_vCMCANdJS+A8Hm^d
z7UV24q5$_nZ2<ld642MzhXQ+mpw`mbdPMir8%boM%*PL#2e5ZoU$^cQUSLPGRUCYZ
zzK7yYn415*mgY$Lp%J+U{(30C2JRK|_4~0f5CrKf!cUd~>%#{tN)#6ZgC4=v1mgRT
zk+&l90qE+&MWYyO|Lm7X<bbuEWS;igAY$l`_o_=-a!0%&7P7Q}2MdrD7o<4-&Vd8o
ziprfkNNtN11^J1fOieB*;9(2isirlr)0Br}t+|Rjbq?lq(f*kkIU|e8Lh<NejKMHQ
z6zApANdkxnZ0_vDd4B&r6`u*AqoX4NY!Czq6iMQsrlef$s~hq9Y=7(-s6>0-^@PHZ
zHBF0`hOYsMH?xRHECev2oU&EWMaO{jR<sKZ<$G05RzJFB78XMP8ONq{*~^Sa5nFb!
z+EoNG%)ow{6Y0|-T4(~$?L_0Q*WC-ymrzE>Zf}3o)c?ma`dP5O6wgyr$Wt>P7Jyd^
z1O6mh*=7$PP5}Ma5FlSrJZI_*MLP9s^f1Z=(^;@rTdl6HVfo8Y(RML%2R7K@<9I9_
z2~-|o95mJ6CbOUDH54V<-d2D0&Dmx0tl&ve_SbR~FIE&eEKjYxk)QZ_H4i`Z!xK?<
zu@`&F6u4cCAJ<x%x%8Yr@PFH+3QCqE5+o2KI4SBfsYm5E3DWDPS6nCUiRxm$&NSYt
zFkqC6Va%cY!D&JCh?iS*Qq2;tdhM~!ssho%HYTS5n@aF=6GlV<hrID^i+YSg9Ok6U
zB-Ul_=WZ9}aGHpTsf?JvCr#qi+?v<S3zTWEw|ea3wX`lfVX(HASV4d<>CP^oZmHVz
zfsj|%YV{eTm|W<Iz?%Ge#W?As=e$Et1m((y`V<n=(t1Zb^bPpqc}!~BN0AlZd)1hY
zKUmIGvfc7sRFnC8+XLoK&s`>-hk_I4!yoU6%k}Uj*eq<9=6vLpvyUOuiu)?$V%KOH
zc!oGGGg%I2=>xqUk*-UOzaFj_Ipz`Xr!R8X*WzBv#2v$tPkNTkB;8I%#Vp4lrWKIV
zNoG$LB<(`f*i=WvEO&%cl;9lmeyH#RF*m}l#wLBa2^p)X`5HfZQzA*)fb-Q|`qbjy
zA6Lna?a2n$H!)zUzZ%5%{lb18Wv!ja9p-C({#t1MVm;2}MStenC|ey5yXV`F6)|`4
zbwBEJ;ZU8YV%|5>ru4_BN_Zi6Q^x<NR!QjExN_p;P+t_C=_9^~L~bMcI<n18hKfcq
zDuc&dD(*x~-4hN>T2o}{g53S@2gUV7cl9dM8lBmgg*;6}C$`lMu*5Zusr>9jvtGWC
z`(!e~xavdvsLk%<;Ho+{IhnsD7su{Qa$}t7vIn<7`|)q3#h(5G%O~vPbwx4^B4Hdv
zk=Zp~#9_qeBa>(F;gWIC%c@w!A|kUVnu<Cm>HqB*_ax;R+S6qEOiB!dyeD5VJ+orR
zxbX#2<QtNoKXZ+tg|5qUHZcTa*VscglS4)mpdQDSP~J!=_?2??8O~&{w@mwXxA}w*
z*BML>ejx70pSAQZwofVRV<!jcc0R>LhBC?368_|{Q?OFVG?SipNeL`8d;T+rc8bNL
zfsCtBdS%U@-t>8#1(#%dwBg22dVT^t-RA;(B|jZ)<5%~w1t@qsatj48U4f=zT^$3-
zbhN}XlVmi2SL66y^R$xFHDz>gZOf+$5Y&zJSsOQe6-R9~U&T1A$lfZLu@QS>l8!tu
zw6C};=1q3T00HqdZd0+I=#HNAWNBkl10l0yK?6T0hFMa{jElPjBUiAN!RsmgycGT>
z5R1`ZY7yWW)O*y&*~R!p&fjS=zgOLUf|%E8!r>c#chl(&Z8qk#%}C#!KK>E@%WGWB
z#|%b-hG@<-`B*)#oQ^hpjF+cwu&UXvb|m6D+2lPk%6onec#WEbA|LsOU+fHhm0WNk
zi7$MkC@_+0@Tlktm+ryWTrT%Ai(qkWW-39w*MaNV%2C75MW@PN*bpX)&dLnXhz}{x
z?{=>{g|<ds)>}S3DH5*bdi0zh|Gn)?^I}Yk75G7Z=R5tV-tX*2_|hKj7^v&pDeP#v
zOlEw}z(ZUqicocyd_iWm!HOe?)?cx<Mzv}1#Zey8SxkpbAIk$gT~~#NhfF3~RX+Kj
z)eETQtE|!mJuz$P9T}7JF>kgo9YXZYA|10#WBT{VnFCjjZAdjejcvUc=%b%M9f$Fa
zeXNZ$TP-wZz@&sDmVeG>s1NgwmnK1&Mq|5=GLo7p+{<HAi?aS+Eu;6#g+www-8uoO
zvM+?4Ti*8!YA0h`Gcqu;KN-lnF3TMac;U(`=WkVe5c7Ta>r2P#?)nVgU?X-M6Rx~a
zM()sE#;;%XOnKZg%L=*0td4RU$mz>^PKuWj;2d7J$yqHA7PjP=Wwj`6|5#M2bZmpu
zfc=}v$r6+2U2?Cv8U5)WXj9bgSJiFT`^eZdu*`ZUgsVPsXpt}eX!qPPX0o7-s-QtY
z;W2(OUr<M>Twa6KmptA;^BZ|*nad-DEFLds9Um946v5yUh&+>?lv%@e<4fJ=%GI2_
zJ}K~%ynTpH=J?&387otK_0Sj@90LJT$_!WGsF{sP2T7%m(l%EfVP30Ut1pgv{L5Tz
zk1p!81Cw05&o{4)l<T$uh1r2x!3@tw7AO<Vw7GEc_RKvPVq-Q^W|}1{aR;o-JZz?f
zKmMfsfZv7tP5q*a*1LoLC=(^3_Uu@L5A+yLPu*CK4LP=R_)NZ}*B*|IP+gOx;8uAp
z+o7hmpjhgzYSXz|qd1s*%dLpxUQK3SC7qq3qN|v~bXK47p1=Q$eJKIiuoG<#D3dYO
zL`-L0Jo)e0{!-g`Rf0(^JIt8uUw6Rj=I1ItPU~S+yXn=GIc*$>FM_Y}&DihyHUB6d
z+~LRG)>^g?YlX7#gN)iHsg`OQex^=qQ>bpN8b&dD?C|pq=-3-&jSkCXc--fSuE&>2
z;r2w*nO&*2%~j|pLpbD={qUbH%|{@Fzn%aR_Mmy0uOT%Uixc~u<Itb)%MgWT8I@K2
z*`-m&5YFrWQ2^YDYm-#8x;!WG?*~_;YXdbZJaEsXLPBswwsKo8#9GD++v!7qFBUpN
zh3myzQ@Tl*NZU96_2Z!-Dfy9RAmwwe);}x1cahZpsa&&JLGgD(6=!=&78f7E#!AZG
z-riYIYU=d==FqmO_uv27>K3K_`vw2cWuFEvL{io4PjjNM2ng8Jw*O)Txs26Su@?)v
zXk1n$!O>e*MMdR@F?g1A?;=_{sIRym4RRr)e_jwO+*z@N4GY|JWQ>CchUpn<X$8<(
zI68_SatF$;GGyWc>7C%BRFOSSYjw~g&uIm_c3#`>HzAg>Io7gz=jEoH;az9H4Jds)
z38;tIYI5&Y32WAH+C;%N3<WbIBNoVog!*K+cJE?O4@d!oCjEKQaiJzK+Wj#HoxkGG
zW?{Y4Z5H<@w1spG&t%!MfR$4e2P<nnbn5c)@$omIm0k>XUa3DtbE{&R3X~zM!JHF9
z8U+SE4-5=Lbu<hq(Vqmsi49~21A%1Rwbf7+b(^8|M=ecF$Q&_Sf=2TqQs+2>goP@7
z#;*t(bi1JkK!q?q&#tZu5}@&iqrbctKoBxSf6a3cti8eY>5-#j8MVNpD4iO&1MG#Q
zX5)sk;(lluk3-h);@W|8Y#lH#Iw91x59-BXb@m6DnVCzFda(p-g6cdmpT+qx=aph;
z@#d{Lj983Sez|Aea&Z$#>;C<h(%L@iT3T6TTbDP#{;pI&@Y~h~^ml2N<JD7BQ>)N!
z^9~Z*^@cMldD44b>^+2Ukb=7oIDNHrcFw?fGk{WLA9mo!pC9t>L9^0ksd(U3LV_|V
z9LnZWfA{}Dlx8y9FE4!Q@@16YIR&-WkU6?eC?(Ja9&Ni@TlL{UR5tD0U^h?o$?b;I
zUKVEjJX{>oN&HwDPk?}2s1Bv%<W52nfuf}M3Mnrnn4<6%3O?`lb_J)+#TG=Whn^j2
z_Fbh&kB?&$F85i|<jYk<vW{lt<{E?1JfwnRG*Mj*k)Dw1Eh$o>Fy+4SYj4E{r4*>D
z*8Uv0-v#gm7P9t(gyLGO7_}}SDj>Zg7LbK5f>)klbFhgkJPgcK@HIZ1<hmhRtEz{e
z-`y#5A29c}B|*(ih?-6;DuP;R-DIb_U5{k8=WG=;0>O=mZr5_!R)}ngVwSaa?m|Ty
zAS1GcP$1FCN^scC^<30s?&%Jo6@QJyc~CRtL!F7}Pe^|UqQT>^)&CCi$;^S->1nC8
z_+96L0TVn~nG*4d^0#mG-j5_iMmB-z%?~Mlttofl_@Km$$5dWWC|Z<$!-0Iz^AJ=a
z)}e<=hTZQ;TG^IDcdio@K|&`AnLx}2lW6O5sz*b<HaC^de<2=~>wh*iGjl|8x;kZd
zch@9dR%W2a-Q8(xIe^=9f8PZprI!)o2r6}X3yWoo^PM2GV5#QEnQAj}ueU@0w4h~i
zEtvtCF~OX7;nF2PWWwafSsewWq_a2U+9H(!k`V`P4`K{pp#DsnZk@B{`l1%+fre$e
zZ2vRcxt^DR#-QgU(*~|YJ77oE0eYO_e4`78?u}&r-0EiJn8A{ZSXj6ja@QqqK`_px
zyE*k+2POx;q?ZzhYxaK3U4z*$lRNK02L(7T4OdEOCo>GvT3ylo@bu2YlTlAunJElG
zzozc^441FX91FS~V7cW0HJcEqezdaHU_><-H#lmA^-3LKdu7K3d?y;DS%tfOR^-Vt
zip2Rb*XspB5tssDhIsex-8M+wum=l_P9xdi6{jC4iUeXXsJk}MwSwCmDrFIO3V~Cf
zGSrxn#s#IoKu7%x6otITol|D__b9*s&Z5ggE(SdA)!C&lgkaH~^Mj1kRaf_#`kzRR
z3Eo(_HNS+xT?+y-*1vhtssbuy;TrK{L&K^XvVdU*xB{^iErzdvw%`HPF)+pazoWSW
z2_uon5q|HLETaPm0Fk~3uI*mMkGnm+kQfpT={|6s4zU|%Xyh((bcXjzZMF~IDpMCe
zHok?z2poV3VEX9M)b=EpI{A}RIZlNvjq-)KT7SNMHy<>zaKl^L+6?Nw&!WU&5)u+D
z2z<VQ5?Nu=lmH?`8C1dBK#e1=s)599M2CR>^zp6@h?GaRQ??+ot14%_3U~zs{<K-N
z%|gFBT_tH#Uc2T;k35QrfNX0>eTWis6$Y0wF}efdXI+6C)FSp=5Dx*9ua!<Ds=(ns
z%e*`~_+F^})kS&4Vc3CF7Mk#8K)?YwykX!MQ7#1|1A_-7_}*y)54Je~WAfZ**jnd$
zUT>b81^-caah>1%jjXzks{Vt}{f}U=A$&;1&I6gmfKvhDWU?}YAQoGeMoqKfin9no
zSZ~+`P7q3MY6BpUOyCijPPy${*kCM;vIFK5LcowzMfuy!?tW#l*}w#b0F;rMu?xgL
z#{e6{ExnBnLhZHu8!Q%T9tAognrX8mAeaf+1;Vc9hK2??%-y;me1>3(Fyy5LKnR2>
z0}F5p!3^I&z3p~Befo46kY+R>bs2avsbC?AsEu*9J#r(Uo4R%PZZ9B*DBTD!DcUMa
zl<ExZN{|^7`!%2h5b*mmhJhh*Gd&gt8Zu3U&}>#fkkD`$fyESeMgkydS<U}Iqo?TZ
zg6kv<1UT6<Xv`CeCM<#;-{{BpZWOEifvgVrhKle)z{`Mr6bHhO;43GAGdPJn>_#h|
z;6A_b1p$zV{RrNotd}k&B_=)(VUXbg-_k%B=!i@K45{DQ+7(aIU_heD7<6P-ux<y#
z+5zCG%}hH`?(V(0U*iUrrU-sQ`-=JWB*eA^;S>Q(ZdM{RSKv{|op*YSwv)}0IY8)v
zsqFaEIsy@Y_~2nRS%I$0?IU<I!mX&c(}DNC>BUK6gkUyN+2A4kuU`|T6ulpsfq!S=
z;7ED%<|z>QydVh-#ybi_N;C=x2mq^8Je0Ex^Ex7DA@f*3^jv^SgfdCca)||dv)5v-
zyD0a_>F0a{>=14v57v;Q(p|p?ft(twj{sGreJ%inD+)A;;LvJ>PhNk|;9xWCqcp25
zAls?CX6t+R(E&SFc6R>=I{-If?fuQja$c!w0#q?6_fBg4<l1-KV@NS*({eGA%VGor
zZ*d)Ds9cBjTSG%*uXr4dKfuUQ6N6Ea+Pay(f};X>Jm1HvL;!LYLD@Ua?d?;*Or#5u
zEppP4Zd~qd)`=uHkiGJyZp#(&QJR1`Ry4e)Dpm08DDHZb_^%My@~wgaVmO*WkU$(Z
z7?5doIQ|>d{c`?4dK)LOV2n<KV<2<Q?&n;XVQ}7*ln9hlE^d&CI&i_B?g(Iv0F>JV
z<+io7=#7+G`BMr8BC#Q4sI9|3(5jU+&1G2ty^${G{-ddK>_qj2FEGy0s&Thw_crYG
z(11u1sS*o9PLE)r_!tJyV+bwjPE)2u0zE)O;eJmR^(q+vFi+*S{SA)_6RZ?3tfHae
zvrrC8$?V1xCAd3=od4O;p>qg6L@DQ8(Dy~D<gn94ky}8N+Z`G1gzXrB_*1Y%&PHfi
z=)sH$ot1d-36y@4Jv}pnQj?(Z;#V9|?YhSU4|jKSX|HSx<&hT%Hm1^${ApagN{!pJ
q_+v#W>-^v80{t@^{$CnRhq!0z4Y^w4Zb)FkA2}(d>p53-pZ_nrB4)7w

literal 0
HcmV?d00001

diff --git a/_images/7d06f5baca09a6bf3099c8ae1060ddf3eba8874e355ff047933d9a606da9318e.png b/_images/7d06f5baca09a6bf3099c8ae1060ddf3eba8874e355ff047933d9a606da9318e.png
new file mode 100644
index 0000000000000000000000000000000000000000..9b5df2c6ea3942345f7aa252dd51ba870db69ab9
GIT binary patch
literal 14548
zcmcJ01z478yDg55V~a@)3Me3;QUZ#KB7Sstj3O-|DM&b@=tziuq=1xkN+{9_3M1W!
zl!$aGDSg)C>_7H@_MG$2*=NsnaWVMjb3b>id#(GuFE1;xnRY)d4Gqm^%B2g6G&Dcl
zp`lq-v|%0o<@Kl1eEbr!6<4=avNW`HxNL1eBXil->YAnPHIpkp*&A5fm{?lya`SWZ
z9{tJK*4D~Kn1{#w?<a6uS{w0Hapu(HCL66TY1q)v?64(2t2zytZqU%2JV&{3PWeXY
zXva-$WxG!+Uj_H9y&%rK?il~^s!g5zGF&`Ou_*~bk*c~J4PGjl&5{f5Yhp|^OxQKd
zISuT;20za8w|E~M+-K0HtQqrWlkVBIo9QlmTU;Jk)y1}L`(?wRNheXpQ;gL{qDIv>
zMSq+cs}#C@?|A73yiP#x;0S)vSaH)b)6m>{tFU1e4b9O9VmS8m^<82(;`wyj4>UCD
z^zPeeXkL9f0;|yMmSA2(L-XvUrO|pTaSaWP`Mwa##)xyP=e~VADY6BpvG(@$4b?`T
zA8O5v8){5YN-c}n7Vt5|vQXjb)#yxp+`_LlF5ETom(}Y7%(Q4PZ`{vk{L^5zyLo-|
zecuKg*4C7i^f-9vkmtOVqx=Rb`ELT(ec5DoZQtJgnGc5p9r)vNt}AT(%WwQw7TG^K
zLUW7LgwLQ#<;=yQN>{pvSaezUhQDs7fA6W<x!Vklj5q}ZR9zNkRDS+B;?${AenCO%
zpW1)@^@oDFz7W%v)F+FJi@SF1iq^6$P&u>s<+lF$)igB9$2+UH$OrE{EPF&$RCn_M
z0p+$A=FeT1=B27#1|sie-S~RNn_ft<r_86NHPh(1O}}hRT$~8^F&dh)bf4?pj~zWK
zE#bqcVr(4$!n`f^fMtIA8%L>tqjfjO3*zWFRgG&S&TWx}t>vHROlg&hii*;1-dr$}
zucqgs@<ha;IpW-HWqJ9COrsj%pBS(sH?0Q6)OQ<Iht@TxXxa663l2Rt3R5_H_LgaL
zazd&10kd~h*1!(Y)2DUv9L8TPNKz=*&st+sTRzrX?hAZdb7tkJzL)6oH+{#OH-}>a
zbro@GU-gru%O!Uc>_<NgwPla9X3Jf=B(JR<x_T$p|1y9>Z$oBgCM)HzTv($+mrESC
zPVOb@x*zvOeEDKk@TGWD%<kfW0b_A-@mBAc=;$LSPGr=dAQw_`;nLzF3#GfeTiigN
zO5SjJX<_*GR?*tGo_l&nbtaUaw_eL=y|(2YQo~1&#6_Vy(?d<ZEbe#i?8>V#%w*@`
z@$r<#+e>!uc0b0)r)*+k@^(zb`txtjE-orbs$7Rz@SynpiO$S^(UIrxk*Ltjvp27M
ze}CuUEYl{A4yWOii78vpIAUOS?PdC6$H_i+PEIL_2lS6vQgjP6?`>rYV37$rc5c<G
zRl&k`i2+A7qF&!#KYagim^@6g`>1+Ee*USdpyMTvzan_G!z%U^W1oj#{jyp-TP;~V
z5_@=wdUq>R>aANec<+HW%c7>w?|i2}n5E(M<#h7wo7OWqM_}`XIfdaA{vRjYkMZ+A
zm!8}&XrZmV3pbUw57l1(@1%r(&(;4^7Wp5$<j(?XNWjt2(X=%q1_8!*bLLvvnzh)X
zVw@g!^19g9uU{o)WmOSE6<JjXhRECNx2js7z!uMT|2<fmkh!KNDtPqlU7ehAKlSuX
zPc_Ji{N4QrdU~GL*4Cb=2|r=r5zKEE+g)-`?Ft9E$0uYJI)mZn6sPfm*{RX&uJV$S
zr<HApNX|otN-U1Dvc^R9z~V2DW*@mzx|i(x`FmSbT2giVgM&5u%y;hCv1`vBHOFq;
zOgmI-SMi%S7nZ*D_-MYk)=Fo2fPo<;KK@8vx}V=J@rxH#<*}bMEvmhnx*ziz@!1V2
zpPu{dYg!xen$FTPv-N6bD+lvF?0Azr&KPOUwCgTjyO}{aYar!Q$-S*o)DyhCcxs~4
z^iTqRWToMwA9J<mIeIcq-uM!MqjcN1M<T}D@7)^?(6V^M0xvv|So)A@<m0P^vhgyN
z!|jgMr#JWCR?kYw$+aBvITf{@@y6NJ>$fQ1xbf1a7%Np?#jUrYzbYgyIr%7qko5&>
z7M>NAedn-rP{bAn;dm_HzHL{b9sTz0uir9bJ%Qa5?p8A+t-@9x&(@8S4E5H7jv5Uc
zbN=+xD~&pw(*7%gfOAgzwV3yQe(LeqPbK)rmysN9q-M~hZuu{Y*<W$P{t35=y+@KR
zDm9-J!75dDNc@$u`B$FwAC)$c`%Y_XtB1F@N>`x^-QK-Nwr<_Z%EtDgiv7qD`72kV
zNUmGAdH?TSU3w{+&+9kw<VS>t()X`H%*KoTZ;mY$YBSa@bo$%J+vUIiUi<QT51&yr
zz1>*5MtXXB?qs!yf!B2_Q)KeQw{5MfBzNxK9s5+mhXctCLDAqX*r|R^7D@2=^`47%
zaAf0al|jc%8{;2c(bu0&b6b+i?sCmLq>r?x%+Bp{r?#w2A~G^EraD2lAXjJKQ9k3^
zpD2pf*48msaf;O*obtt#Uc_FSYFowM;wQ|5!mw}OK6`z3c6R@OO+XV(FU^^q2V(-2
zkvBUW`=7*AOZ}1kZ!3~OVaz^yhu5&OHrH<0@8QD;VY?weVL@SGO$7yo?>F)A@=DIX
zg>C14GfLZiINUkPDBMLAD{6k~BkI&TClDed5K^`39nR49mC{qTj%iNSNqTfhB2ll%
zRrA@;zhw5{kn%<gr48BFTBFY0UYDR0C##up^?rUq$Jocih@TiMPqCoT&}7uy1cj#;
z=NBhydi(qPM=s$)mAe&87efRs!!j~5Y(M|Lol-y4lw|kq%VkRaosBy`4-O`Xy38Np
z<yDdsor`Wx*3i#hnr_icOV!B}J|&2ItNf|AOHmOXER^6A29oKc65jj82`W3{4$p6F
zY^<p69;btW{lc|E&jZcfn)q`Z_g5T}pduO%v;TyOeqKxe2O#__H~J6H{`W5VvsC`6
z{qyIm)1z%_{ASI{fJ_&?_C|Vn{aCgRus#0Q{3i?%e1~O2<(-}L_X}Jj@Wym$ewwvf
z#QEDZr>TKR6hxA-5L&^fZ@zF{{3>DIk*m^_q~>tV3MS)t6SL{Ex_Zz9#?y(lwesd|
zFJuHeZ|L8jG^`AY1l}=Sm>D&F|16Rv!KS}@z9D=$8p+t3L8SRnIvg``SD<Uf`#-!t
z@JY7woOOp?iw=Pq7DY?;9abfq;X@RIi`caI3GMj}mo=heV%XvRPlW9_?DhJqpG^C>
zEk11=ZD?rld}0<vr7A}5!+o_i{CVyGu=!85wN<$7LUh6%*#!mFfXrAe!Vv;5U%r&A
zGG6N)o18OLcIioL+7DE$hEoMeer~d#HQ`Nh|M0LDU%#UJv44q^_Yv?|?!@vvVrR~r
zp(t8dSfno?R7d*!-WX=SGt5+xJxT2HDu#!H-Lp8`V^f##asf#W$BuoWO037*eO%}M
z1;9eZ%IX=+tvXVeqLrNpi1A=6imbf5@|@Ta3My)If4EyJGCQ63(y-o2eP_W<&nHZz
zA-T!zwxv7YanjG<Kjx6cgG*G^G~JZ%#fS~{02_9CvXWOb*EYdvb}Xfg@uoC3&uOkV
zuqoHhjAXQGudR5A3GKpd)nTW#)^DLdl3Q@)Ep3W=x)_zu`J2_+HEZG#!lY%Hdb3l`
zxZICD5s7JjdN@u^*U^2Ftbx0vchhB4Q`5(N;ukIqRPYofw*xHGSp3daP5k`qoPXXn
zQM|_dxZ8@>BMlt-xTcr;4zE#lBf{3hBK(x2qOPv){QPhlC(B+Orhh8dbwZ$2`Ev>G
z{%Y09d;io+IjAs+_@c-%h}d7E%7%+_Q9fi_hLf)8(<0(qyJ`$Kkl&b^xUKZfn<H#&
zr4~npg>&DZ()S|$L3p?ue;&g8Eb9>mX6=_&m&N;V`Ym6-yloHm^Cuoet4Z^U@-iN;
z{pwTdyyZToGp^c`eN_!W?=j(S%Nj+?^GT*ni7{+4!SR+|g&q$c1XLL&{g!NH1NXUZ
zkL!*OCemJ3RHTmO^fMY58kX18$O*Pv-6mLhWaJvb9v^b;lKN`G)BHKrhG5s!_)0uA
z{%o&n#kV*2ckVxN^|wToq}Z)z7UkpQ!kep~h&;bsvTn3JM~VFT>8)okk|05}0Ht(~
z58zt5=yvxnPaHbyfLgNbM>WyYNOyM5tUpjtP+$k#G?bK9sAy;yyF8O!q+_R!E0puC
z_Y|iU&Py&W%&MU%Nj+(KVXn1#@3FIb&R@kH9pHSoB$da$XK9O$Tz?ab*|84)&`|A8
z9c)R&4(zC*fdTu;lP{%hZvXl#yiesHM|59Wox^?_1P}N2#^mScPmB?sz<(f7)ZP#!
zabkKJ?a+A9%Cf0Fw&Iqgt%|!<b*NA<uaWHIKBpTuw5w8azC5K|XM-=5%5ef!nI;rV
zd}dS=R{ri?9jue(w&J>W?b^ZAyfH;}b#?K+?|aJWz$`5%r<QHeL5}a!XQ3=FFKgyl
z#~_2u&~74Wy<BB1$qO6l7!w;i^tFtsg#?z6P2cY=EowfeCNBN$PP5y;n+GA)2%zU#
zDkdgo+vPGl=(e&D4uq=8yb5-Z&nZVW{`mTKF51n}Qj5Yl!+rX;i{HK`Z6Q2J8GNj=
zs)}o7-yQdC9^L#Kb_wN|0Z5QrgN9nt()RP4#-n@j@bPKMvFWD|gV7^6auREprDbHA
z9x#X?GDKZ8>n(eeWuFMY?Q<BY?KZV{pxoQSponZow{2U**Pb%YW5?uCE>*8xOMYS5
zm5LrprQ8+o*8W<Bt~e-Y@7p(Tb{+6}@Zd$Bx!0q6_wG$hPNo?;1ZEl69W*d77({cU
za^b=q7L$?p=^RIn)Gmw{rBzl+Hz5rv&TPgKo3Pc~+Ap6Sw9)oQ#9o>ojSA@6v}u!X
zTwGkDTH2TwaVHhAoK4;F3Qsvsl&K6hB?TH*1;^AqmEh#&mOXdw+=na^-%jUA8JW<N
z3fPdqc7cD!EDS)@uMD)Cv`SN>5K&IcbN&4Kr9ZmNv>J)`tc<TLo%!C9T7>+OsmmW=
z8$V{WM{Sc-e0tHl@u|et&5Wni@fgiaLvK=PP<p(X&MZ^g#3dv+IXKEmlOX=+(Ifv*
z<d#?bFU6S`5rq8qBkH3ad1}+cErBc@F0&nE^$l`oQqab$;cl!JX5SG%{k}c8D#mZY
zW`hnzOzc-=8$IQ9n@Ff&6B-bmPyBww8Phr9%%@xzZ`@}x@eK?-{7#6;kC;ydkIV1A
zdmsxm0v^|5u?Q4Jh3#OJp6IKvU1|-|r1QkP!_zfx%Szep)}Ez4Olc7jKe@mX^ZjR5
z67Dg%ssljU*k8cufxWxkpP`%)qlVoa6Jljy89*YS-MY19YOrCtQAtingb|J|W#fs`
zK?sL4%U`z|mEqEO>-JCdgro9f`fmdIe-K(U?J`f<%|aPgp<otoCpZ5;5Mcivc=<27
z-5eF>_H_Mw1h&6=b+##9;XHxu!)ZlhBgFl<s{p(?%1Mp}9gze+$m9Nfg{olw>B*Wi
z!(X~fsvjU8Ki0!(!h~!tBB+&%T$i%0wVrTv%5C~ZjwGSp-Bp7nThBBir1)LtZ5^GQ
zh7kC03M>1gIL`5!i?$n(nSkn;n3x2OOv6Nt%vf+tDH2RB%0fayF%}_VVW}N?4s9+V
zO->#W!c~(wZ?#H+dK$ov@T2|mW<0IQQ}`{aIZZDO(C9oN0|%b^aOarAvp#+LL?86x
z3V(1$Un0{XJHsN{OK6r7IndamXbU@xg@O%IvsL-p;n0<`X%F{6N_M9rpjxnix#p)A
zb#-+f9v&G7q$m`yhF1;Buwtu?77Bk5HnW{K(>lJv%j2>;ourU23a>I}eP?R?!?weF
zRdU|`gQ<D%y=Qj;lE6J9JU^ZTR<j>@uR!$+2rv|MC6HVZXr+V~Nk*9t54iW^zJnB}
z*$#)%(iIe#cWrH21y0k}V(8y*CBBApU+XF?V00Q_URoM=Yk?)2GL7U?^on$)0=b(|
zw>>)*@vdjD<5sLDuYX-jAmVOr?RZiibB9wK%+d;CDZXq%BKvVeDVq>6^%PD1x22^V
z0s_sg3fTU{UD@9u0!{B<0?Ge{pZ!7SR>PY<I{#76m4Eg4Rv247qPB{HP<?Rd4xaZn
zKKK9Go&S_I4%OOSJlGIdQBp!2+}eH@o)*||@BRfo9DH2=?&8wY69!TyB){0(-|`S}
zoVeOy+aLps(E@%flmV6URc}ty_F|N&E>Ikg`zjyzj2yvTRPMIXGgsm~G*MGuO137U
z6*cJO{-;Dt#{eYw${<IESsP8<7W$JBI7O$<e&fcC9-f{_(aL9=nvKhYlU%(TA1`tI
zw2>y_1W9?06Vb6UA+e}5YuB&WtNT|!T+0SXWTk_>4R@K*Si5eWY4M$nYWXQuhQ7c}
zca+Jiod-ZCL4sP8f>J#2`=dkRJ&36CiV7uD(?k+ui2B2^-|wWd%+ZsT#YeYmmoaz}
zq&=Sb4;coYGh-b|U{>Cim5GS3!A~j@CnPp#JvWVk7wtaahGH^}bYVKyo|C&gKm6n&
zQg(!C{Dk|LV6#;AuRlC2t~vPm6JX}8dX>3sFSxdqUE<>QVkwOx*MiJLcuVDt|I<y|
zpZep!DKGr9J60oG&33t^5rA;;pctwyiK@Fl?u#n8IV=1<k}4*N3mr3SK5E|n63xGY
zX+x|xP)95vNdj7p`LVq5#zYlPBMD-64YGqbK5caQ)_o5RfK%OjFtenIs0tNQVsifS
z%gzQE?WF*cRY$HJr$BR-i4wYjBSJ#08}Gw>iCb`l^?2#Px34`Ozx{T^P8USri4!MA
zN);6qIe5Q^_VqTgmmuGfUfD%O3mYB(@H`uNj+-PFfolAGI2S7sjQt}UXxbklR>EE9
zDQ<HWdX3M`)WBRNHs<_G8f5+Mehe0B8+@=b@D1!g{dBe>fD0vBEn(7JAmo}r$nARa
z=8V}KcL=@)j7|WvX8!)!x$iZX1uq}iFW`Rc@L>sn<rcJF&n-HHpI>XOZ+%1-wgwBE
zo15D&Y#T3N*_jLgmSEPBTKxc4lQ$q>0}WeZjp%|jxf)c6hs(BUg!4a8!y@YbA&O>!
z({uE0IwP)RA78~FwzU85dm3TY(1~bh`lJ3-T>9tT|GzoRcT&Pi=#LB+sIcAak#MHo
zT9#ipKTUsi)9#(UsG=eF;Qz~${yZUr9<s5qAv|yr+zF^uDUe%-G%rSFs5(}TIL>tA
z<En!Bp`_fE<pq9+F>TP$gRoU*;nIwjkpvLGl?hRDw!g6GNc1_gtj<%kn879&c9gPT
z$U0gRu^$)aBpD)LZf#9)NlWi;i4B2VT5LQ#JRluj7*_)F6I!dUD#UJ}micfjxudzo
z^*dGocDcZ8q}Jzs1;Ue}E>DK9kUp68dyxlal9zt^=FBB3abD)pm7IsyBNUmKtky!J
zDI6;MXcTir<<0Dx)P#ess_nhg-5f=N^yy*lF06ur+3h}%KE4dpJ*C~WiBmxLr2XgL
zA4&ZU-R>Co1pW5V`Ii6l`Mk`XAAP=sMeyZ|as(V|nIr-sesW&ed5YawUUY<fxa+{|
z^fXuQiel8$r*w3527>l|mHP;)1$^iJdOhW*KalS{tlSh$*=baEeP6QF^<>IB^}tW+
zzSMN?x9M!Gd5yWB|E@n8<uJ8z(ME=c>)pRcv9Y%H7%SQpQpFZhrFtzNZcVr68Lb&q
z$AbqChBB1&KKp(HeH{;d8OhSzU1h<3$<cP~<Bwj8SzK43{6Ph2Udy*i#uh*LdqiA=
zFh}w=+KKnxmJk*OT(LiXmB?AAXWO4HE;<)1&9sT@=j7yQzO;(WwH>s6$4PSXN4EZL
z_Jj4}NLEDIsc&c?t11AvkwnLoH(oeAazllH6Hu#k^z<VmN4VYX*dOq-?;BR6H&Z;(
z9CNZc`l^2pH)H8urEZ_0H7GJbigiTJ0p59T(r^)STK@dhV9$s#UWvQPoP-rkdD-8X
zjQ_H7Yqle8^8NgZ2IVdBqW6CA1OB>hOLmlcv~_%L%g!Cw%NVfRN~r0Lg)TRVIEj+&
zU_DW;^db0Cz1Kco6mdax)9QtVg{B{}62LA*!^sZEraA{|1$I@Gms3$N(Y*z5Y4v@K
zeqfWsPueUn8Xhj7^IGmFaO!F3a_-^QuEot7^BryG=I1$BST2CRYXH&z#4p#{{{>q@
zLIP`{ENxM#*S?<5Q~v$dY5vekxzwMX9sTfJ33!=q$BrnG@q8r`G*^bo?Q;0vp$0=Z
zgtB#%iHV8dqFtRQ|I4r3Xl7Lw>^y6jFU&blH>nME6&0Z|N&^aLO4d+1v$AMY$Madw
zbxwRIn{=&tc1P^tup45}&03lR^;{F6Edq%;5Je|+@kfr%-sTHb(z`;fG<pB*GTQq9
zs;}>KwaB<J-f5nsM>jXO!=ks#6$0$$3QQVe2itROw#W{u8|e6gi5h;pM=$GuP5+rV
zjQ~6H!UWu#G_ZN^<GpD9Axi<w2h^UZhB&z-Q~+7$b#&v_W;^KVlf^xE)3uFx@)XvB
zRk15$a5yQPao(@@`NU^mxuGUCy(Z*oI!HvQ2L;p3T11`~uxO7*MiBYBPdPs$`5jJp
zZab)WdaCw3C@9~;g)tE{+ykHl?A;=IJoc&A1J21@y7b#VKI4b)YHE0{%U-;Aw-x0o
z{VTk#_R<LepQMqXEh+*0IerEe0sAO@Xb=;lZL4i!h9`P_&g{3ewA{(WWXbY4Mk-MF
zrx<;II~k{$5felpeNbgic`kRRVbzJIL=|4G7uT-;EZqMxxQjEp!#*TIC+~$2oR2le
zCM+XaN>?}R>*OTwY`cw5$2VlkD=d!z8HEofmiQUSAcLB+JjU+jeJxs^PAe@bDOm(a
z?Cm^bxlgMR{bxMXa5Kck#MD$H#8MS``JV;*0i6^sT)2>KzJBw577A2A`%9`4>=nV<
zl4#6XmS$8F@^b3!97od2h=|c)UBR_QdQe?0mu(M^uCl%Ti`6@YB{Z62A|lR$4{m}k
z-k4)!q?3Q6F2#o}aCWvF41&L(Uo_%c3vsdK(vZ4=Mj*7?CuiJrnVFgGdmu%Gg@vJZ
z*j)WAsTRQB`t0&;7UDwoJtcQH`yRg-?mR)snU!PCiU*b@YPh77loITvoM|Z2l&YiU
zBRU@szE-kh=6$+8)s@q0?6YL^o4rpl8!&)>`GEVe$%Y<5#Q+wl5bEZv^G~j*kC8&A
zU@Mramy(qn-m8|uc5`;jzo@8)@PZ&8-~hOyhVqV6O{3YMChW8@<ZZO;Pt*0w%Aq(B
zFb1earBaFa>pHr%fJrYS^uy9@*Z8Zf2TB2U*eX&FG0zZ>hG`i7M99-5q6K0lO?bB_
zi$^(BHJVvKWK%u-yTK)an(>RZ=A2CzS2ls$ALlo1JVt&Hg2yTXc?8LCYUQ)bCEPhW
z78a={NW2@jZWTmT-6E|$mwHtOyuB$4ol8@iUKimbWF>Wpd!ZQef>7iGYtCP~*VoVQ
z^9#%!*vXKNZnE8`=5cRDU<j~g*N=OTy}ESb{M{d~Uzsn*8Z^?cxIeo3>#x5`H^j<R
z6}xXDKQj1#85lEdh>J%}xwGpzcwD~AmoKX=3fQP9G~2a;%pI(ayxlS0hHf_m0Po%X
zK~A{I)wzk@i&Tv)<3}AYHtjxoj4GX^no?fiMjI)Ptn~iUiJnsL%mZ$pA4zzzWIBv@
z?#MpCYv}dpDkP$w--5Q-v6@jsLd+mlnE_>)LxPscvmdP_HDloSEqL%_0rTLEn>O(h
zg7<i7!KTZOZmwQ)Md6WEA%bf5()Tzka`S;hE>ejuCv@Y>&vl_tWrMKikHMLB)unT)
zIO=ku?OO3-nFPNi!Vz2IZ>NLGU<j4v^72GrQ4qA(pgVWgG(5X}z5zMf_tB$nuL1wq
zqh{79=QfZ#9zT4zEfdNI2yRo+xz9Uy?%H)Vzo4KB3YXs4lF|k)j)Mn_FP-Rep7iec
zCfMby&+?db_U{FF3Z|%!u7-ivlz3|!Kaos{^gbZCpJE70`!h6jmuGM<%#I7Xy1Mqk
zjc(R0jcIR4>w7){zAI<6MgeX*SRZqdItcb6eH7koUR+YbK-a=tKB-CU^5`nGiN5%t
zEo$HMD=+HU2_%H<5Exn)RgHg=&`u|;V>pM+3?RM7B9VzpNV_;YS0Yh+R~hm`3*wqx
z7P?m(MSEr@D%#hMm7SNj!Q#~M;|g+Tma?$baZnkXk25ojNVa^Q1<?}w?Ab9c-Tapm
zKL}Iseu$sknGa^rns~t6a$wwIK)V4Y5`P5D{qQ95qp2?E=KXvRS)c|bQtO4}G@5%A
zZD{-bU)ZZCU}KnEzeSSC0_=yX>)FHX6}gMpD!4-)DcpHNG9dygp1ZlSAkvoIv3r<N
zsEP!eW!A!t$|!X};94yNPCmoRy-zUm06PEtA6cbXj}HSZ_B~|yWU8(I3LQ|VKcnln
z%Y;x2KI2x{BPRUbKbb;7lCeelEZ^9F0XYnU4hN7Kwn!DKK5R`$dPyn)ag2<qfJG+-
zZm{MCx0EAKTek|Ss31!s>=RW}G}DFLC)>?+Qm3A07*y;A(e!fkEsHz$x6FqrFnY(z
zHZfh@P8GLpnMpa+77ri>EW>uPN|4aR6bgl884OYAF4L&l+}Nqo0&{q=*nrq>_3i@k
zS=!dt232jJzN?7x=<u)^L?fVQ>(K2@m&-HQz(Vzb6MOrDYmvVAjH`g3$dQWNYww@E
zPS(z;9~++Q4;Q3XD#@KTP)*e~0cZR%+DQH@u0vbJJ;SSnnsIjv1M19ZTeipv6x!+8
zatz&g2L=XGWM#R7p%60m^!6IS3=F|J8K`r#V2tZ7oq)3az-!P7hJzdJqZl^EE7;or
zFj$;AeHu!iT9eM8SxhakmNckwX`oJ!2-4p-=-E9+r}zPnGRSx#nDaQ4Fu+mQ*r>d7
z+qR!E8CJPCH(3=W?m1G}K5BYapSy@n#9@pN+%KE3@P~ZR1csgYj(xwZXFS%~*%=F=
zx)1V<tm``e9_wMn9S)-RYJ)1$D3E$KZr;p~2IS>q?D%6Sql29Vd5^A^K42II%8EOh
z=8w1vf%gT&rzbyo@&xbEbFRDN?*91W$DeH=Z6A(Jqc>QZ%enD23O(;}!46wrU#c*&
zd#{|^yzir{7*)7R2y|Za4}4157CJrS<KqEp`vVPRqN1WI^?gnob~+7nO=7?%WH*=Q
zUI4XBh&D+d1l?<ku@65Jp7?uuA=wJpGN?uOX3D^F@k)f=5C-U)O<FlCdYf4=P~!zP
zmd~=Y{oBGq2z2)lP&>ik-f}wVEQ=g+?AY(;?J2HEc}MUA9xaVJ@0FI81`+@6;QREa
z8X%4^VPU2FMN+l7rF>?ztqRU&0oF<eDWejbEma_#$H6%q?ELlZo{LmRNcRj(^&8>7
zq{?8Fh-8k^nwpw5Yt{tijpbI1wC6OiKjzdz{zoDw)I@1ZOG{&lRtu_4gjjP)$zJ{P
zmAkwE>50PJ9Ox0Ao%}1E@W1w_ep<jKB!J!T-o>QRc}!7S+q1+?eECvqQ**k1WG{Cl
z;0hK~53-#hw`=XD-SHT-E3c}u8*u$45skOHI_66N%(LLD*5+rRWMJ@Qs7`{Z37*Dx
z<;`!jvc{Wl&WzOIesS<cBZ-7#OfFN8pj}3Ttr4*;+Is5CE$mtZnA(cSg|&&ow9L1x
z7a)$>ktPB2j*xfN${sL0VMvf9lSAN*+Y`4HE>2VsVd;k-exT3Q!++pIB%$yndL7@=
zAt>*U&c#m)oC{G(FZ-xr0x>i+l=kMSlVyOj(_p@he}Bj<+q&|Q`MmANSF2cAS%nN%
z&Z=BsUh{K5LO>o!+&>^7_V?c}5rqHx_3PkMjv0Vntd!xFv_K@9nenbPj1~M2t1&V%
zic7q-wy~i*a3Bf(_B&L4B<EV>29=R~ZVcl<pwac*!=no6$=TU?3IP5RwZ6U{FnTLH
z+7?LjAc<<bR+b&|MmO0U)yTL2^sa_(rgwNasm&r!207tvSR2}*DJUcwNR9-65s-y6
z@S$pMFfQD-=Jf0lWIaGf0?kDiMl!8adNGK$3y2;970|dGVx5rYB0GWrw2&YuisPLH
zp3T;HkhsnAT#efh8Poa?Xoudx1S+Wg0!PJciu}2AzmO?Q%xqSGM}R#rXlYrOuo|=g
znfsA)mW9QXAmB;5bOOaBKr2gSJni(}`wSpB@h1>T{C^1SN6`2kKYsi~=>XdM{U@y=
z2;&JLSujz~`Id136GZZ`lWAjoB<f-!aM(Ps`DmXlNcmrW`2{6MB|&J_A`{%?3O1xE
zPL2uEs0k*%NW-G**h@#+X=*f5gK=Gycy3>?8FY6MF3qe<sCcAZrlqANRj>*Z0EDC>
zQ+(Lx0@PcuTE^)vE{IbnU{rqF0r{#YB7|zIhp!TUGy&VU|BP!su?2S8cxs@I=(y=+
zNW!tG0}1bZ51YaMgiW+vS(<ID(#W$<A;Tmv)IfR}V*_VeeiUpm45@}EIP>PL1!1(&
zOf>tkD@~Pirja2nsCSTYd+E}82&su$O+daA-U`s~B$<d%Ow%o}#gG<7(TDLg-}O%y
zDD}KGr<Ae+u*!!lWU1Iv(&55Gj{=|@1C|rf*0bQPbI!O{Bw7v78WfEB{X(=$cBNP%
z2%|1OVKI60+xG9^8&er{H$~kIQu)vxD*}BgBM<BtX9Dp@42YUz2X2D2$6(!`o>>}{
zBZ!4Okl&>Kg70Cu=AxCQ@u<;G$9@JPg>q@VXbcsy!FXaE>3qRTzMZLyJii`vY<goT
zjcXarvt;*8d-hZM05Ey>AK*9jdvq1Rb+iw)uM@jpXE6eOgcmEzJfH<&{B>%I4^ZXV
z0X^j5M_19Rs`IhEyW!*%OjM~VhSN@21mR-B=h*=!q>7CjHXI{)_dBY;%JLWv64c(_
z9_5I3jGtPGie~I}ybrB0+o@CS+72CdY`|<|$2ZZ^dMvxNnP)$S^M_z=E^i^vZrBhq
zlqt6a^}e?^SFTRb<SfY0Yk;R+Zbin|zN1Tuq+~|O0$6x(H$y?k^-JG-Gl|$&BGte5
z4xC<?0mwZ-F(ku}8FZ=Qs;U8Br>FS|I_y|NKgNf)iA_XgVF%}z5cK>Z7=YMwz{lGg
zFN$u+vlNtsF+_Dsq~6NP$Go1Uyu7^L;o;%?_)V4T-lO!$pjAy+;O2ww+tc5#7Mnxk
zh2ivDJAlHf$7Z;)+oQl&^k%ei>EtTIVdOFOVTYHAN+ZMLRF5YuM79F$n=$a>_(k+n
zYAef2243kmy)ch9fNJmo0%eE(Da_Q60T~@t3dShL+7l2^#IxR3J5GK1@*SQ^>bWgF
zCsaK7Z5)mvY-Oa0EXyuYFyA^eb1PwB-|C7cCnjtsK2wPA*bOzZ&(KR|fE^{8edo>b
z=UVI0l)~R)-q?KRo^{bGbk4!bcO5vOb*-)V`GyS}YO&;LlpQ*qphw~1X_%I~fGC=n
zo#o-5MNFz-NJ6S}1&L(OQd>oZ6ks#;_(jm6-22m%;vN937zP$#>_8Z!Uxny<6lZxT
zr1J!l@_g{-ha*q2Ggs4Cop$f{U(!Vi@PGW+vDJKbv@P!Q=g(K#_DIUfW#=(4GFmu@
z96Cf~PN#*@Y|nxw$AmItgl>8=)nj5AV}I0WU|O9M4)}b*>o^f@rKKxbXK#)^Z_%AN
zzB~Yc<^-hh+^Sm)wrSOwb^?g5uqcA9C>$IOZr>T&HiNmsCEN=cHtdJhtED<-As|h^
zI~;Bd5@4wA={>k<9NLw@xu@>Z>}ubZzb;^Q;wO|y9h)i$iE?45G;9}8F`AKH2?aq+
z?$Cby9#WR5-FW8`b%Ez-4vi$Rg1i@WgPd$GOp#(}W*T!f0XN)kuiHFgv9f$bvwP<I
z$P^JG-jo@m#qSuE>g7uDTU6+F#V$L)^JEL=Gg0g-^NrrPlg+rh*d2dTbL~eHZ{4DC
zym5nAMtp%&Di*vHJ0D-x`y~wV6cwlSEzC%`%ugwRMiTw@arN{--Fc$8h@C%g2dnDY
zw?#xmD6aHJMu!^AF}^Ua6C+*d6yW$Up|&P^EBrA!0oIJ2o4eW~k4$%9RO9{drnB!L
zIdL?S@kz8t=Rg{1<vgffvwl5tn<fZK9;LY>W^9xX;O$tMGOzcPQlkMNRo=UTMJI5$
zRfj9Y*Y_#du{_5>&A8%rfshojmQNS}1ICDxJS{nZ`JaWcJh}V#@57-2M3<(Nz{$l*
z$7BM?eJFBsbN{?Vs6EV3D?@Q=LbkI7F-V9e!tdgf0MEWrP>Tb_^fZBGLgZxuPoP7`
zI&?b5y?5!a2`8UuNOg3=gUOT}*6OL#%ds>QFyRlb{5D!EetcPR#UT012EQspB~*#2
zW-ZG>u<vvkdOSR7S0vcGZuEFxVOp(pL$;GCb9%5rg3y#;QT?#h%#YV?Wjv}coqotv
z@#~?)Ll?c|K_)Ufe%@|iWTZDU^hu0-4S*(+W3i@eX?awiLSi4nD?6^4Zbe3bOQ}XK
ztZ(0kk!R)j?T0QJ<8uzMsiB~M8VhbZVD5DopZ3&KS*50W1NBICj}a|J0wc^9zRNHs
z!@^){5Q;INN{kU}$exXK284;j7eeIwizK{+)0fxl=`h<!6fuYpglA=^kPeg#B|)&#
z)Gzf?07YWA`0cuRXTAo$l4!<+nU(lL!EkCGnFb=0en8(J@*Oitz+eC+&k@Zpl(O&R
zJiuQSfN=n&NzlwlH{!kbxEO=5Z7mq3T_96oe^V6LA@n{_R2;+~lZittcmV#@Qs}bK
zh`|}6+LEB`$jj~s4>rSG06{Odk@9HV%Cmd`&W7N&EnqgDij2MN@tE-Mdmu6<k56rK
z9mZ2Zj<6m+Y=))`i9VdTBA+GbQ`p0A-4=TF%M0t;8Qiy>btk4FpG%?3k-?Z*A%P)e
z)(Rf%1TxaLWV4|5%yCR~Wnpum7@J`pUJ@hH^70SSE%TW*CtLRpHN^Q~i%4<CoTE4Z
z7153PF<%rVL<%<u5poOztWPk#2qF?qZ3{@})EnWbp8hb}ZXVra6h|Zw9pnm5`v+IB
z4D$5}nm+|3T2)uqLOcN<6_CL)uu4wBfK0lsa|YyVFwBHDk(vaO_?fm@IuamJC&mjG
zQp3P$lD&eburewffYcS{G?c(OFVY16PzE%~y84bv%XJ6SY`>9rAm1IKbH{XcA-K$V
zLbF(Og2QaWzDenNt&TXk)rQqbXEn`H1E}K!=8-QaMutAt9|de8>Z&b9+-R33?*{sY
zgd{=VFu@QgF(my9y@<pfOcb5H?{%CEfS^LwaB|$H5}-%EKZ5_cvc_<>g7;GaR=N^i
zwjsin$ms8dZq6hFk@KR>7$k#>;QR>!h4v?ZpXlHskSd_<tD!6rxDFCsL}bS;$%}u?
jO@ALN{ckjSOKzuGb*Cgu7GI)oq@jq*UPu-D`Pcsi?}8`4

literal 0
HcmV?d00001

diff --git a/_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png b/_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png
new file mode 100644
index 0000000000000000000000000000000000000000..02e19e11cdfdad69a174666b2aba075d952c3749
GIT binary patch
literal 169439
zcmZU4by!nx_%|h?NJvP7bhp5W(Ty}HAmCSGpmc)_L?oo8Mh^jLLApyiM<WuWM<X3$
z@E)(<8-Kii>^j#u8|NB3_jBK$dcNrEsy}{6_YeyU>#?SWiXj#j4#~e4AwF<ryd?|-
z96(;Gue^+0?Y(^e^RUBu@t>EQldG4L<69OVI}cAsR~IqC=YnGVEDl~?Zl2OYLeBs9
z1A?v|5TQ}Nrb*x?L~a^po>*9qDE_^$u`;q~u&}YPG*uLh{j&BkX7T)*9?PT__9mN#
zCXk)O^qibk{U;B|D22R~p`=h3%P%b79<WH!y`>(_Is|JUXbT>96kmS1=}od{GAKgD
zUqVCH6~2kGu`P~{I{Pm!F51j`TRycOf}MO1<oVwRilN*41-$=z`tRz!{N^LKE5r6=
zp-t`Ue+N5{YaE9Ece<&a{<<A2;{Uw+??aq0q0ZI+`SAbz&EWrc3y=%u>U_>xUFEZn
z)8QpL4H5Awbk4rU?CMY#3Lmw{Tw;QlE$39{yhY#$8;Q*G^L?L|_Vy(gUwzo_sxRi?
z5{)Yz7avb>&{mt^l5!o$gvNjmDlap=Bh@bclgR9STNQcWL*KGg5$pnnGfzXkyr%Ct
z+rL1T3x_<KVkBLcdm>1g-!?4tPD~g8xBh)}NdVn$7Sz?%ecSZ9Z++cnacPP1??*32
zM#l4X(D%mugqH=%>$J!0xM2LXP}{AI4O`*9n{wnJ-%zz5Gx_n~zvGwdS67Y9(M}B6
zb5@mya$qpi`03=Yi!009C$6uBS{+9EZ#r2syfGUa>~V2%Yj=0;M?(I)&;qAordWc_
zSkcWJ<DKRp&o&>>YFn$`N$yc{RZRRO!JK80CT6oPDrbFtT}oOys->mHgLtr8#G)I?
z5xUlNwtS%5M{jw$m)>^KTrHA(YuoYVh=FuIc(CD0;jTJ<c+*3nXhCvPJ>x7UKKsxY
z#Bp$XEem{sw=GMB;KA$dlNZeWH}gW1tKOdbQo3kNJB$`pV~yy0bslwJeQ)N7LjX~+
zyRtOn-Gs=A2T*F%EA~sx{7occC>UvCWM=O7X>p#N*sa_KgFs*U`}+;(mi`VToHWaf
z$4G97?f9x1+-n~Bw8#lJdZpFO^SijZ+JNEx$H&KtZH8}oW8**cx)xR15!V><6yI>!
ztPUxf1K;iDXieRYrFzXq+0IkWVelZW<`SHEnK@c`4-T`Rw6NqNoo2?sMjB7u3HH71
zMORJ^Fw>VqioTX}r~j2pX&7gE1qpVW8Qk~x4IjLWt^g__K0cn~&9m&x(XBUuS-v=(
zz&Yk<<0<1eJ!;XLGq{It(?!q;6FBMkB|JVZE+CUlCUx*>Hm{=a0r5fR{_^j}`hB8c
zM5q1(-E8JD1hHFHV=M~tiEW|;?9#9hiWZ$eK}?6jk~!V{N(_8MhIbc^F5NvmL=}>=
zo@xEZuF<*`UaS-ySaA-@msHMMmba7#gX5Btl3rELRV^GCk}0Y8)?dLx$H0DgoJOYV
zWuk{FTauEJq1&;3nhMX<Yv-N*)hbZb=K)LVN*j1`zrNB&4SluMFgr8G=l9gvrKgHi
zm|hr7wmLy+<)azDplNO}`>?#%^NkS9d#aP)`S@l~D~%R}v=n`E=(FM|Km5prgA0Xr
z9#^I>((>~0(b3UX^?dz%0jRmww#fh36EDZ?5fd^XySu6!;Cl{IJh@ot|5I5PwUAPU
z&PJx#Bsw+bG>KL0=S)rFh|v}ggB$ic)Uo`RvOVy{&s!gErUtBu`y=~4iXS|R9<J^r
zKByYbbmDT>?q$PlP0-D3dhE5bR}x+LF<*Cnxusm87ZF(#=3EJKoo$p>E!yl*?|C!3
z{ibKf9se&_&SErZh8jM)YhYc?f&Xj^SOH$bPz5n{wGg*IwrMpSth0&{zy3Dv+S4pA
zZ2FGc&h2?48-QBGCdFe_QMQ+=uF4#>2Oh_h)9Y5aqGRARcwyps9OMC0KLZJR=q`pw
z2ws^b9$P|R%cJX9Ub{<SN!xLj9WG^oV~b>W#W@sgpd}@Aax6GJIwF`8WwM7(*6DDy
z9o>~<jkz@J!EjXAj&_;_fp@hD%qAoya%kmjE?YcDZX_Ng=)^|-3>u8H?QzyAw4h33
z5xD(vF2hDZR~GqAU?fkQ)^fS<PpR$LIq$mfaE?Ev<lYCL${QJzr^QAU9SrJFd&;~-
z`HPq2%r}KRK4nA<wGzK{o5&r1bV+rCJZD|g4O_ZwmhRf@Y#oeikTv7&$d^RLczvD}
zt+gkl)j=&{k&!|3AJ@Nq5M36{NEqFMUMpQX?x<ag_3R9dq}!T`1_6z<i+>3A-L{oM
zVcQBWa&(E#5t`BYH6dMDbG084UW>!Jtd1CTH`|axAK74K4~rCNUo+x(Tj9CQ2gkMC
zlt7+WxFnpJk>iKxpvt09{gGC1!k@;@_5p)tR;CesUz5mC(NnPFG3T^~DPDo~18;K5
zmaJL^ExTxcy)xTt`GRP8(ZE5Pz~%{cSF}vviK1#!dL#`$viPxe;+kuD%X6QUb#Ks0
zA`yDiOGX&dk`kZKTpXConVfh!k{WRD*w2?y_z5wnO_feQg2!!>mj4`NK`V|yH7C>4
zy6zPvSE2aAyvYp0`m?uEzn-Da%;FZWS8H6t?|fP;;XO0nOPyUZ8b<UcmgMAfb913t
zk@NeYlCsUuCbE==IJ)W2^sX}<3Y06!ytg_R^_xDqYjV9UPDruIS`%+A7Ru2#!JaZ@
zEU>AWCsv^9f7OYA$gXh&LpsdI6hLxjS2>q2FP^o)p~JHlOBDm*jkrS_^XQyoP>tmp
zykQSv;%!=scIE%EU5wQ8uQ+q%<xXT6pnZ9>h8frhqZoB6+1TFZp55~XX3=9#4zJ(o
zw?Cl-8l@4#4`xnFL%*iRKz>|cz#Scl-l$N8d;1#DcD#_7vFKrE)yebLLvqrSx9{4O
zDWr@e({nqC7|^<FXDBf>Ixoxj6IYXlL|eZv%Lf!|^aO85CQ)MP;~sqnox>EWj+9W&
zQe>UiDq>gq2(tb6?k?zOLBR)_)L~-goiY7JHn)c@H>>!fnVO)tIiul3rnhWu;=)0d
zd~r8pw$dkWb-G^9?@Sl5exSDQw@X$bSwe4ml>92z-74q(cyV<_LrO{t?U)Xip+~t#
zTa%0jE^2tzw`ydx)iRV_&g_kPq-vbLy0^l4*q{2O=A!I<dd<6;E(XrK*{FftmOT_O
zNd`>x<)V98n=7J$Mqo8(Oc*F8Q(pgXZC<2S&*J<%)T;)UfMNcyXW%0&nc<FO?!G|q
zk?(G8C!udzLj29>0s8?3_G>j6|8>lDG!_BLxHd7P%o<7TcHdm47<O^Oq4tiueVa;^
z#bpVzS+P`Jwx#^{Mx4F-cXf0dLRZArViQiBY#cM5<(ay1Y6_TvWDDmn%RDS2Az=!s
zeAgaL`ys?fs0{1EfeqE(V4n0pZg@gAl9{6cs_@;IvyK#)$VDrCgcg2FMf9e}JiX9X
zo~=#qF`}~>He{gtzyNN<u&se$=q%CiFPmIj)!tscEJ@0OYZ>`i@6EsaWIS_l4q7?8
zr`mo=IkGToSd6QpMri_5$7c&4tYWZH<G)Sk|5cK>Y5WxeR`j&A9o!!ED-U31Vj|f+
zkru;v)4#%Ik%NUiWA3l1srl{F0Nu#96EabQ@tPCLw;+kivrHuj1wQfFH`SDnv(MSf
z7u6CcBAQ<XE{#qU%o$BskN)gCgyH_z@w}oU@Fv!PRN6P}57m}wu;_n<1`eQ(7<*<T
z4)G^`_G8*!AB+pvkig6uwWk2k+2P?KzP5cS`h0KBJvf-Tp`n35bk?w91gi=GzPNLG
z6j){T<65ofO^jIg4@lDmA}gKRf!VW<KfV35hW@JD1y+^1=x0HE&cw{jTpl4KP|$Ut
z{oR)w@rE;Vp!Ku0F5=Y!m7Z~-PLya+0CQSe+Td4fgp=_3(dz1dZWX!`?gdL8(%kTJ
z_;8GFW=@Hps>0BRY@z0g&GmBI`Z(2dZr<+pNh{u)#7$V!lrNK5zX0yTa=mlp;v%u1
zyxm*z3<glRea-unV|ToLwT_Ic_`0%y4JiOG-juu4HM|xiAhI+op={vrYi%%8yB7@o
z{V*&5Uc;)c*p-XL-5)}G^6(Ti<U^)^kOFgSHaHGgBuju9#zpDX%%7TSLdWtbHb4vi
zao3p0_6)jzeG+_o<(?aPC~m%<bPnCW117K^`og(2_&&&(S)AJZ;qW1XA(Nr>YA?Jn
zZvs2E9gQKmt!{_BfgOx@zQA@|hFIpuffW8;h~>+NaRnCIS+{~az!y%lE_NRZKcz)N
zP}*drRqN*fqy#@@Ez8tSDNqmKxV{*)c+Pp#vE)KPlIa`(ATGJR4u?UXrzS*WBz>Ye
z(g|7S41bM%!jeGLbC$@Q-D(z#q3-HdJ)1Mh5L`%3I>lMKDqqtVjR6FmKGft`*S1rc
z@#*hlv%gDAmV+ZacBP6eass)lWAP-Kp}Q!VpH?Sk_vvn>NA0uxUC)kE{`vtZz0$hn
zzoB2hRf=h2Z%6Lz&dKzPUm()ma_qFl%R@2t)U{eZf*}ZLKMm4h?d%@P^rYK&X7*Yi
zHXdB(e}MYHu%*Ca_@*?2#l+R2Mn|Fa*D-oRGkQ2SP{f}R21R>e%q4Mw{aH)Q&8ZEP
zg*5PFjbrbnnvoh?@M#v_<g<t$K;z?zNSgT7W3p4|6pb4jC3rCieK9LPDB*0fY262|
z0TbZpviTPhrKn?xT9=?(D)vG;{MHa;A9;57di&wSd>OCZ=x+ns4)rK!z2i^nhPFW@
zKa|2JpXvM05J_Xxi9pD@)y8O;H*_@Fawxo7V`%e-XiDYo4?FxNoJ?CuHy(${<Gv$D
zSR~fku3&w^8N`hX%#MBt4GLlcnx3GL5NV=&<5<1<4*<Vk;Ilkhm~d}g9l4@&_xr>}
zCCHb)m=Mgx#Z_hgV{Lc0wDNv`pTDBA@<Y3EZS8DXKe7KgT03jSr2&9@(V^}2ySY=o
zovOk|Vtxds=`oP+514w8LlgdkS1M*NOHSyjAr_taiM=J&<`aiDM<mb@vX>nB5FMuK
zdWX#0>k;y@tEwGfXy%I>fsGq<Ys(#qWThX|j?UZmMjy>)DwWANe}caY1i26T*p<ZX
zD+tZIshL<v6ieUgAfdEM!9hxg_BA`OIMy6411A5{AqXRu38=%#wDbeCOeg#>kW93{
zL`ht#E`~Pp;fuIu!gOyX&5ahmv6%LS$F)*P$NP<aTanvni{Q}H(_7!#DlAgXRqnF~
z*u}fLD4?>hj{BKOu_7aWrZ>Ji{)EMWn{Jh*r!w)n>?}Xfx5e_crRDQ{PI3|biqgin
zGR}vAn4HoBk?YBO9U<FfkmqWQP1P%665AoK2#aXOS`#l*_+UJLy2yJrKC`XH8(NVp
zmY*)Dl>5}3Iz*#ti_)Hed6Qh-J8e6s)ND)m+Yg1wBYi)E5LA&gTW>X8MBk*2R-4It
zFYT<TR7+omv)BSG6KIVQ5fNP{Vq9d^o>{_n^}o581xW*DPx+;~CQlb72Afq{;6-`&
z^Cj>AHDpskE<~<cE1{c%<3iKRwPE4vY{w{AZq0wONXIJ!rF}D^qB0ONF%RiW{!dVJ
z=Bq*YM(Y*mT8PI8KAc4-|FQ#e$z7hj#FgVPsvl;N*kjFzIEZ2o+q$W3AO0$p4Ni@+
za+1hoA|@j&ek@?q8ibLy>Ilkq;g~sIpajchVsptMJUOpE^sENi_&&iAy@hrlnHl0o
zXjOWWzPWOM?7|J^c)Dkx`|H{t1s3d=jQvm%8(~oPtI7ROXYaH4KBjV*mI&VK6s_SW
ziPzcYu7w`1Zzy>oP<k>}?4HR~C?oqAv4oe$Qc1oS-E5^PI<~`-^BAMfJ=EetGquR;
z49<Ji1|i=nu(jZ{cU<;qHX;g<>|ONI1HC&_c7^V^NBf;_y&%e(HcjnMh;3t#IiQ9e
z+mhB4XUCWAZk;Y$sn+Cp`p41ewu>!odJ0v^kj#pL&PD0m-Pn`El%|X6eMn!{)GUM4
zs|B;>|CBYrcv!mS%UhIkNmAMMjk2+>H<a>)X92a~cK2mRR9RUWm|uWPw|AKMXmC)o
z!y5ikr#LOJ;+HiZKPA8<eCW1&L3!2$N<?Sd{%}S6F>o}=m#0Aq)cu;!vq^8uj;U{@
z)|YnSqGb)w+XZOXr5a)a9(7K>5%G`F9ZgCrjuzd$$r5<_*Nj;RYCkK3F1T;^8A4h5
zmY~54L$=|%eMxd=9UFF4uX-{iUDs}In&tI1E~^iy70%X*Hg6{8(_?@!39xMU#pPv^
z!$H3auxDGdU4dl|FR1+Cb*B=&0XBE6_ZFqyQ_CaKqLjg}mpoJS1U7V+OD<34@2+g_
z?F}Y#cC>r`7O2CzMvlmY$IZ0W$~N(=rR~?ML%)?HnV$V8*9@o7nI{g;sqcabERyb;
z;S`8iOY-z9hHj?+ozG38ftc3S>x4@6g5sB^G8k<&Qjt8lbcA8YRMf-CuupwRQDs0l
zS?!LvBM#TPIwy{8CXUl(Jk8h3=0slbjI%t@w+ztb(o^f{ECB?BtG}ztLd9B%Z_&n+
zPI5CThT>BRH-5EQ?icCKetGf9FLX_h)@a0jd5{Dw2d_hDQ(!DXR0xb2KXvFl4fcaM
zR{!m>ogGI2N`Eh!P*sti-A&4wa}BtBm~Sf1OLAvcoxm58ayq*CZ7(No{B9l*ME&&X
zQ+I#=`A`Tfw&H_%UQ<&O?Z7@Nrprkp`t<bl_kj<tu=!vZqLt4r5oDnDPV56rE??rT
zWltI3p6|xdvuyAWKn@XjB_8#ncY@T>T^^-VS=ci%hx^j~_ZJuS)~4s0hCO7I6p`<$
z$LU_i{<zH^Hhl2WI50hYaGCe6Rj*<(rv;vpb9HsBV*khZ&)ax6v9~X7`i2VFik6Ni
zXl&ztGr}vr5(OF3P?(uNya(lWQ4Kz5e861`9z)N=sZY+>4IUn!pL@USq;+{4gbib(
zj#g!0Wb7Rs9mNiwXIhxnN&IF7W<#-ma}I;B<yk`(4}97>6V(k_CRYez^usp>t`A91
z(gJ;RATpd%=riKKsV6Nl+kfWg&E#Evtl<~hwX(4SJz@#oS6n;=Tm{;Pl#I+?{I(n*
z1zjE?FE5=N7XGcutnBRGnHdwH7W>}S?RFgib;(a1eK*caOB5TH>fe0m$lP#se{;U~
z>W68lEH$&xMRz~jURi3`qJf8L!H4gz0B6|I`BKw3??iohc^S$dG*i?p3fYV++iTt`
zVDUWUpU+gT`4v!oYU?1sbYXfP8<Iu+IOQM*feKa5V$DHv?;hB%r^i0`UJ5A=RZ<&c
zZ_%w(=$Jw{eEhTKtNNa={f5S^WXOAgZP`h?HA&9qvsjIYli1rdR~FEPL)Og@QMki6
zbWHN(O8z$Do{SkB7)i<uCJrjAU#rovD*c<z78{AzNvjyDWW1@XdoR}On6l3lY>Jyj
zq+Or2;;kwhLIAp?b|yZq4x%0Vr#%7^r2PFgf=3TDRn!eV-!)NDdHwozPart{JU~wB
ziMcrW=N&qj&~$j^(oRNVtJl_p$3pp(IM~R*(-(p0gNEba`=k4Vu*zO-HDO~y{j<yy
z!RZRW5)m%$RnMEH?C^(=O(m@I)Jf*hl}{yk-YIwYN>odj*?#mkr-y)|*vn1Jf@Uvw
zriy|VsU8EW=W<6d@`%zuFpxp&=e*PGWyq@g$?n}9SR(TUa}|pQHTOvB3*#OFT8$-T
z78V61n>|!~vEXLWYb&b@ZI8QDUF;G}5E9-F!&djT5@7j>JVJ92{izZ`DG3|JUZy!a
z%4IGc9W+nbW!t>TyRjtE+`G*j<pk7Qn>H)vEMbkH2?INpa-2OgbhG@PwaS^O@PnTv
zOfQ;~jm=c0%1~{!oLRLmntfx`Fus3Ry^WBB+m;>Qr(T<K3LGIj;ujVcmdAG5dD?x&
z*H*Z*bqz?=xVe<|1<nQqJefw*3~k-c%a15oI=|)VraRs!cu<d^KpzlRm0Fa%%BX|k
zMx+0(MA^qYo=<hOwH?G(N9Fuy9CEM|WdG%n!Fy^x0<p}|8?jBI_5oNtvwYvu(o(MV
z{2x01l;eoYU9jHOx+m@L5qq%J!0GPu9(8r~<D+{_M?4a9=!?+_>_GN!Zf;_^sr0Oj
zCKo*O<WKGzwV%ol2<uD7b^GvPfh)+;$_lurWvl-wXZzjF9x&3UI*+vo@UW)CxwJ|i
z$9BXEEX-zL0>R7!;&zy;D=DkCfEl3%qQ(GmZSg2cycfGQxOO-Wb$aPOY=UpbR2Jr)
z{C+1ah=8>}O5xLqnhoU)X0dOgJ)`_o5>n@@5kIWOhnU8h0Ns`}E(l^V{jvxKw|oie
zDhb5vAKkR6Gk=_#`kMfQI#HgJ+f=pFwszrL>J!FE+zk3ziMP9!UH|jLM#x=qyJiye
zMdn3<xKyL*!afB8HeLgC%TREFo(=e_7bWGEAin<Y+5U*8gMF`CH2s{&w9bE-pYY9!
z8qOzwg~Uk7`jwPk0*H+O+eKg2qf;^UIO$nRRqE7SO$MXe4{bA)^hebilTQLOLXfIg
z{5NLY*=1M+T~U-u&Qn+puM2>O_|EzvyYYy7LC8imbu?cc7pIBM>c`J>05@Eh?6@Aq
zdhP#&CIn|QMn;3d3yon?E5e>X@B!L6El_!YRAAHaL3&~J-`otcxU|omkF7Yg7KA|y
zR8DHlf85NsTn^7^sOG1N{}bwWcX{n=P?ujcI2}Jqp?8~I5*CN&pD6#(xlQDM#Nq}D
zUuR`Bu2sRfjB3dl<;Z4<o}L_=_!D8JPsS}&l+@atnMki9t_-ogm*!SMA?x1Mu>{IT
zUIpamsS5B~6%|sqmJ03uG|{Z%;cKNP1NMX@rrla~J`k8`(+E$DUh)t0VoH$`twL9v
z--^73IXk-uh{UBlZDX9G?M=syO@>rfk4xEQKO3^kre(Ej8RHIDUy2-lQ+Ja^q$=^@
zjrCRQ_PXbfb=Q(yj|`7Or5QA4%5*qCKfkf9Ef!ep9j=P_CR*%t|JM3<DG-RZ_3I0@
zqzFyTbSlx8*&c6*-+zG_Q{v#nWoM<z7<?iKTQ56yS2+~j=z#P-Zbou(4M~X3pGpHn
zc<uJ~pRcesnkjjVM&1Xl#ITVd%>FhrXgD=2NC6u2@bGZg#HyBCU|`_BVvfSFea#nu
z*xo4!{dj>uoi6{K`s;8V26?5gp2qcTo$K1h?;a@lF$4llTe52;p8`|<S$R>vaPZ~{
zYiAjXa>5XeH-widx&;h}swEe+1ST*p24&_5+8z;^tZKp?uoW!}y@WWUZ!pP&q>3M>
z%+H-$mPqP9?RB!ivG+O@Uf16iDPeK9tkt+m8zR4SyT62z>z$w~15nUq>)qNvrMfl+
z<x(t&t1X1@lG`P0AzLfFc#4EK3m-22B|Zn;-7cK53LU8Uke;>PiKxKK`TX^n8x%1P
zuXLOs7a`}Rrx7WcF(6fcqH4d&>8f38ixrOW|6oLJDd8s?Wj<%xvT)$@w{QUX2>>iX
z;A*kHB8Fo?B^@~Cz?OrpM$I`3r8T4$z)|F<lU&O00OV}IEFe&-FpE1Jw*N9S^7dUx
zvJ`IOer47vFnc>xb-n<)^Nn0)oqi#$WuzYyspZqo>9*1c11^xEe1nqL;9#lcU@WeJ
zTz(){8x5rC^@WGGhsR_ng^SoUO-@b@la$9$g>8&_?l68aiCUdkrGkhjl~;#C(v8wz
z)7TH^3u#VGmhNmv&Ff<w)qICirN}ns+A=HixC8~|IIE+3gKWIZ41HmSio`aXPd?eI
zAHRP5)SAp@a`%W!u6b;;xfEH8m{gt_<#1GGYE0&2%x$;5orc~n=0)42s{SM@R;j$|
zIl@ELR*{TMQ}^dUX>bdYV7#IcopC!El`~fuU0EVipW?*kY0zWtg-J;G?sW`}P`9ab
z3)?BfZ$EeuPRNLZ-?w%5yAq(>y5dN}uf65tKNf~08l1z1uK53}Gd~|DjuYq6SPbG(
z77X$a!a;YR>#3@r$%a{D^43a`aS3H)d5?ndrLs{BkfEu!Rof8sS}Y&yURzVsj`c9q
z)03jox`XK>D7Gla+srH7NJg4o0dXC}pE>AdDf#N%Gv<Lg9Ri|7e@nzq(gGtH<+GUK
z6DNtlC669YroQg(ZpBOwEaLY@re~HX>~D5-7y92j2`oQ$>=|E0SHCcg-)nE^SJ~iw
zm=0wU+p;QH|MPYa>iuUsYsNM9t5f8WZSEa<vgAp*NhK#~sgbv1U)GC5Ji=;A?hNrN
zMrK@Qz8iARH_VKTk3FQ=C3oSr9XP?XN1$)tb;g@oDp?c^<jUbmOT>+hOpdR46%^#Y
z1;recrv=6-M{ACi#|V=r(~5ZOM>Hmw{Yf?9iN#+C(|wOFn<<5-ANU|F2<)6_@`_IX
zvtA^OrFzKdUwC99rxdqQ6lIN~t3J)d3SAf(a19PuV_DeU=u`A|v>)9hms>w!DN=b*
zP{WD=e81SdNc>}M+4IkObn=qY1z(RaEUwKJd&)H#91YcRoBOOeK73cA574=-pLi2z
zCq<UoKwBpS=)-Eo@l~wXbFwdcl^b338=pJvTe-%E|JcpHYoy@DJp<=Y{Z6bH3UWVa
z%L}cRlmL6l(vDir^#UsbFt`hx?Y-(1l=m8dsk`SP(aj#)Drg3*qR>b-Z*$p~YXzM6
zlUx@s){6<ru32Szvj{fv-FA$<c}olI_*IKk^0RU`(oPa~ezHD|daEJ}Pp3F$Sj~sJ
zM;uDeWqdrLh4D3fV=|j*6lP66HWR;U6y-G@jP;x2O=2TSe*;pE^m@hqhPHdB)w{Bx
z<+UIwNx?*Lww(elJ6Sb=kd_E^7${GlMyvi4_v_0E{db!gO6)Ioi?*rV+|^YJcHal%
z5E>IMWzrfv)lx4O9KYSaylS-L287_5;ebZ|^-p?DD;e}*mRdhhbzsX7{xNB)JJWFC
zt`Z^DrrT}*LvwP=Kg*3)mFtyvyOnF`qmCnpGw<-G=s(G_+riispJC8*M(hELpZgu;
z$iVI6z`*BiPyI58tZ0Ax&!GBM8@*)+i)xs2%cAF87$;MJ+N$dx;z1U>b`QVOq?61P
zVh6~W!X?OW)rF(J4S7U*F$QtA^{Q3Cfsgh{pmXMN{}_b42}jg?pQ+fxw@3qZ&s+(w
zrpR;}M;w0fvsxSAlxx=0_KGSmCFH-cwFks36imwD<uvP>n|k8M2OW^=Z$y0lnp~qq
zq}R$=qaw$9fjO!xik%VSQTI5HQ%3(&sl@rspG5MJ6?>U17B^++L91%ytXk1HtNoLh
z@mHY7PefhlWBqcdN;L)mE{t|d!5{X0q=9W;GtYul*t4sdFPB%1ZA}nKc_V83J;WPa
zW4S!M)>@bg((vrrSz%qTi3t}{<_^e3dsQL==I@<+x)I{Pg5+>H0ACh^F=?}seB23;
z8lHxlc_Tl=7TVUwTlls%C;wOsKlMLa+x6$Y8a>)LWytO4i`Xh#`RMv_1rTCU#IPX?
zmclk{`vp2zZg^(ee55uH>&;?1b^zEsz)<hsij7Tb2!(t<Wy8g%eRUaxb~QSyJGa0$
zHvewYFq}SB);=ilV{Q)+mZpVq6#PeScyh!#zW8gIRS<dd)>3xiZW&9Z;-%lwqnErb
z=|E?LF<uc13`-SN!RbFHA@hC>?|1X}qfoThp3Pfv**7C!D)hKt3(B$<B6pI^?I@*e
z)jA1O;85qV&miS<_@^F{YD<wh+ODpIE-5e=EESgwJx=J@&3*Hvu1bzN_>1A3cj$95
z_2TlEx~c?!ZAqdBvUDTD<4BgeP4<}|eD%^U0SN}o`hK>}{ccBA5(8ehu{8Dw;l)u(
zCEh3|LH$SFG}P|-G8~n#f2k+KA9S@EaYuU`CS&m1`UJDJ#p&YWGI=Ve(NFX1@~+us
z!m*T5V8(yoRdph5gp(5y1C8q6DmVSKkVmdBPhNq<Vg7%x8YJleQDnJiZ>|!E8eHa%
z+r|5QGY@d{Uz5v%8ik7$je_H`yJ1*S<+TnK3SYUV9J3@dU%a-q?gwll3Tm;BMFCC3
zoDqYTLQYd`BelLBaauTB@N0h{g#t3qVVhdhyl^4=qh9wSf**VwyhZUqiZy%~FCFTQ
z+_>l1t42s0T%4So<i_Jz3{`oUZBz@hMG7=wqBo?ZPM#@J6&LRgR!sOMxoxR78rVa{
zTmNyo5&N3Z{1CqNteO713BQQ}89fGUdu^d|Z7!@yno=eXy*edrGk4|Ei`pMU6Yr?7
zUkSp<kIxMGc-Uc<uZwx}HS+WH>8?k*p`znosU$xh>iEUm7MTbwV74+{0iBrX^Vc>w
zvG*jI0TWHh+VjaB@%WUHZ53kV<Gj2`^ZFiP)k2j@k^pHSUBVo6^<RQmQb>Gx4lzx`
zjSfq~ps5|uosl-Vo62+=hz*!g)~J2ksbeewcE(<yokd}c>vwtDt9L0QOm%-^BD)Gc
z%W@s)WEH*`>Qimj-*mBsKVTO>-SYctdZKduaPK@o_&O0&?vVWA?;=(X@jgWAuNfjg
zl1A`BfbsIC3UVI55^|>+IB_C#U^(ZF@o`*2SZd|PpGT$gi@sJBYtsq%Q6GFM52-lZ
zjOovIv6r0>sq$6_-V$o_Ki{3f9(qsPTuvG!zc)+_BW=WghHH9y?S=wo&&Ec9o+zNm
zV4~|<Nq9&IatP(Mqb8WUKAk!MtS34C(iSnE&}@{P2#SB+eU?shy(N^)Dnu(d+NT4j
z6)}K^jnsezsG|XaLz|5-rB&Fscat4AMle`c&qNgT_CtC_BRLSNU@%FfiB>A<xZM<5
zKf1p=T9>~^-7^6H^-tUHH$Cswf_|f7eDm5SIVKLjs$s_NlP~`%z$3$RW!Z2KaC$k!
z*~BJ6bs*4iB`hHQnD70<g71P-j&yJtB3*zQ35c>NwECHKrxH=Ze43rjZU-nhOJeIy
zHXMF=txwCG6+R+(gL){eF^A_O<>P`Y)|R(Sm>i5bpCcf7-7yjFZ8_HgQ}yY0j{S2^
zSiZn|_hoD59hEDxopIZ$taaPTurfx26TS4<T%tGytM1P^+6?d;)&U{Fw9F(}K+NGs
zyIrwM`QIS*V}{&Fl#UnG!=-=n<Jp-P*{?E(zlx@!ahFpFiobQ5lXXaWj^w+m_4rUH
zXm}&{V!(O*HaTg5ZQPrZ&%AeXo`uFz1kX3ZhN<;HUR*;*T8Se{A$M8ouSLmbJ!o*$
z=Y6y~^fX%oLwO#n3|E<YtnXDjAvZ7PI_|t>!Qejm`&0R%%Ge8^8n>#4%5(fCJzp>s
zLW|4)gr=RHofTTD3L_b*nKqzR9|JOnSBWd}c%9RiORng{<4Y$7u@BlT$=F%c_x)Na
znLQU47m+qwUdO&6#F#Nla{>BM>4MPc-n<6Ipc{U;<GkAo7fZ10CmDQV<Fx2*NSief
zCUKLZJizAqNF|?Un{jOb<z&n#ACEh@)N3)Xop_@){h#p=?*=D<=J#pW9}UGO+zHzQ
zVS>sUR_yxbnQV`Bk*|V5o|R<27Yy0qT=3c_f40eh(9w+~<50*u?UJjSg#-K#TQhxt
zOy;UGuXnDjsYc0yikKy?i^+W{&~iKluRL!#F~kqdq@Ek~ezEZ5l@PhN%9%FL{Hv?_
zwj~!yzpsxR*`0xi)4wqkTHh)wD|-eKha2@A0lZI2PA+z8YRZd98JV(@#*tg+rM0ex
z%Sc`#b|KyP_yB%?4PUpu+qAw}(wv_pa6Gl6uNy=|O*g{F#C9EZpdN~5GpGAC^tvY_
zjy}r}o3&oYv-S`4BrDcy;wecznqQGeGQ`vNobomtnX+FxQuh5{Ejot`<17qf8Pv~K
zG^Y?Kg6*5koP>tu+|p+~TM;n%#XgNn1_sE9YWjGK*l*n)VaYxW_W~y-X51c(z$QcU
zBSABIy1H@f6pRMU#!*J6eS`yl$_Ov9jgIlSwrGm2Fo?9ln<e3vRd34Ovz-q5hn1Oj
z;9xLv1%tkXO4P7YWVlyb5`9dSovuH8X)z=~_Cj{W1|UFU*3Ks1S-1V^c&){J@-cvs
zU`3d|REEdLJNQrJw9wBhp8e+kiVQunBmy2Qg&BMWe%yhFhz(^B7ud+{I`_3(YY}gz
z1GV%0ArjL9sF%mcdvq}KfHvn#quH4A^1j#532JSE2E3=ug;~@fI*wv@`~E-}KU{q^
zc!(1l_ZkQ#O?HsHt}Xynolo!EU+JBBlok5mI&ezbP3q5=IPHrGPkHnS(31OA`lQ&3
z+fC?94*wXTN-K$CV8&j31^-RL#X*9l5iPX3Ed%K^`5Fl1ckW~<rBx2;rUWHVhO9pq
zBOQ!HQPOI#AXH6`#z7JtOy)wh`T(oz2o^gW<&`9nc~g0FPQA?QD9x8Y_%$8d4wBYW
zyeFO89X3O5xV4CpSr|&Nm{ltjk6RQi+Q{uf>x-OmtVEt|gJ^Yfo<94_5;2_Kah2QF
zFPpgrG+FE21Y9_vyUy7@BHLM^rgbF}q*!71E45hf<hLD)B!=zI@i|?hSFIe*t%usI
zDgn*uyu$=VG5%#z*!|HQQ1kbKq&(EIpQc#<@O=bvJ~HZud))v1OV)uoaMjlD^bu{!
z_k_r&T=n`s=Z6gk@%x*@6*%LsW`E1j%*Y6{Ta<<>7_AEtR35Vv+EHps4<s?0$bM1<
ztp<667@*Xyy=(%CgKk4r#!26JlCpJ3T1oy;ElFm|F~iC|l}ejS^X{sS_@jQ;ajKXE
zYcvRDYxME9>>3ya^YJ|L*8v--?<{&*b~t`I9hF4>Lu;)cE-PRb_0O8fk!MKn2#C(r
z?@`ogS2{(a$R+&-eg2k~2UqZ*{G_VC7xLA0QRJ#^>abKI=>*@OWaYx@g@d)3vX{_;
z20F=9PB#`d=+E8({<CB%^{gMFYF@*GzLt`N>ornFVr;7{Y`9d-m3ea=5LET46<zK8
z{*Du2`<eQ~9_62`;h@LO>SDfhck@mX1G3aYUX{dcV^b^5!0DcKIY+yPsUn2$K{D<$
zefO|QsU7|+z`$s1X`%l}jKeCY{`tkGl_1O=aF~IXAxD;hfG?jPA1<}rv8#hjCVy28
zV1*7Do4Hb*43oy47!J~8i$n|k81&m)KJbxwb+Kac^1+%^zKtav1zcd)cPQg1^;-<u
z%0C|exCxMj)e)3!<83&klk|3}W${so;~`(+Bv1OBCX*BApQr|jr$<HpGYz_4d6(`a
z*YnN~R+yaIi&`y}9kXDKIIKCwwL32@q<T&j@Jwk+2Y}?}8-Zl7HT3+Sn{aVquwoK?
zb9g%m8M3h3qgu_i**uJKk|h-Vjnp|T)}AImNx9v^1iALv+#j~fvLdrhQU=K*cv3)$
z0TmwQ6$7^dd{iDj-6lLFA#qckv}EKzBVJgWPjf1p%ycP|R518Q;eT`BiN}9kJJk8|
zy)z{l2#+Tp3qNHB!fYt6UU5^-PJjHm!$2WEE+-CmjyfN3)v*V1YrYwlLDdh<b*}-3
z*7XhRTe9l0t1L@|?cC9aFZ^zOxhR?vn?bL<jPzpefEDSU`4Wb@cmeG1G9q=F-MrnL
zOiA@4;|3RPOkiy$RoTk9n=a>NYX6b?2duYCbcvdh%N6MKLlX{M;*Dn=??IF3?3pRv
zu}Z{)Y8%9k8%L~+C%1S#UdFViK-I%?HR7bZ?MA>LCm8;oH+mk|B7tX^I&23F%JWOj
zE&tZ%jD<4(^AC~c2N&_$K{00<Po?IK&Nj;`+jp%W7AY-N?mt57SuAIKEOk-)2~AYZ
zxnb{yWn0)yE{Mz%R_ZCX*dQ?tBC0e-)mq<LE#Z&thPl8L0ghjlPoeeVuRQYdzH(<_
zzOwjOK1HEWnA3;;9tXN=8EOj+FGuZbuKw9<3kMPa%J`s6f7~j|YV!;%qW-4Ms=oml
zM2`%<haP2F7T0)k6Xt$inF&WRZPTTsbsV1t-3PWxd5T{}Il&$gG^b2$VQ$+vhRelT
z<{)p=Wv6~u3kvOaR0_1{P`pWO$m>EMBA@3F&wJQwT9qI&UOeH~?-Ns2TbD~+P-KdF
zNo!qGbv_T%BXfwaHXM8s4B3c*&e7Z&N8E!D@$+7vq<8BohQR|QUY2v}n&-0dODIFy
zF5c@NI}$1+l#_rreYaV#9dohJ*xbAu(hYEkd{T4HH;q;Y{spGmS$FMJsA#<bV~6!Y
z8G)y}c3}d-yA;IM6)f9tEpKDnm2dyvi0O||Boqx8Q%L<}eVEL*k^EO}YhcL@jj@7I
zyPQbZG>e*V6V}`v`+{LaFSckF33qEqbFA53kz3pSyhKC)JxlJOHml1hrDtm+h8CR9
z#a6o5nB{|FURGM2yYAX77=DgieP73x<NKWZR_km;*t^H`FyxQvyi3DBbkub<YL5Vs
zzOE_PZ&J6j)$wDtdzJgkZf@_w=eB_-_nQgCltKJBU8jOCeK<>!Q8Jbd;gNf*?LygR
zXo<%({ooV9?AH*t2^3N0ViOzfR`rupn%@<+x;B~#svnb7OQwAZUTs9g2R}cg>fG>5
zwkbk=4?A!$)X$E3HK=F8@Yz89Hq0&KV6VqYkmCJd<oONWM<1TIep$h@q-w8RU0v4!
zJsFC(v9`9ha()5?zo?D5*^(9KS?UfXROe_N4VXE^T9yRwZc*}=2aWj<tFZW#_9>Jm
zsRjY23hbGu&|bw{OAxag)h_a329HFbI~$7fwno0Sq$;IlON+}sH+7IV(s_)J-rRn9
z`Qlqsp7tCR{$th(6Q)Uy5Q9>4$iG`jkuAZEF36Sjnd!=@_bFxKDb<~u4mb|#I8s#X
zaSN_#N7&a`J*(p<I}S9(3)2(IT`f0)?lz;?%_mx|DhY%dhLo-u)Om*hrw6W|5t}K4
zTREI`dX*V<(`klAja*d+={$~BuZ`E_{5Z8ADh3J>JAPrKX_I!Yc}<u#;U}(z2#9+z
z=M6P{({rj>2!xnPzdqT2mr5<N_>E(p<M%kRad^fX4*<|~{>e^}F`f~lR&tq_o#31Q
zakz?lcngBw?E}7Rzlts*+OOhNk_NkcL;dzyY>ZFy0ibPu0b-PCqa=FBa(k2u-}Cwa
zwG&p<(4BTYP+Ne-*Vp2Ev<B<AwF3mTo`5!{)(@NuDY<)0A#YrWVM?#2nXOYxZWFoe
z4fTxUg*NTIub(rLrKz%}dv<A5{4w&%kVbs=m_+g>S$S-R;iN3R6j4C5Q$`w}NjVJQ
zXPn!=`3SEuLr$+@e*7`SvAJmNAWwj@R8pC)nUSjGk;DPEgsTJlt0$!ZPYRV}Egfe8
zB58)HZ9)S2jx$BTMPPWyq@`S60kq)4;%?E;UUv3*UD0^KwjH~U0o}RwfegHCEqA?$
z|Gz3SDP!rRi9E=8_6qk~KWSFwR;HWp6I=Uz*(#=zcHvEUL!}6YnyW%1t(4lKx>q3t
zwAK~R(PIouR5}={fmx^Dv2CFQCngBUMPc>x8_V44D3;jR14?1u?85!V7wMxpy2_S-
zfgRd%n}sEZaE|d(o1{{GEJoH%E0}FE`}7aTyiAAndWEGtPI<FB1v_JIK~^|0hyjCt
z9Xl3$YU<SoE4%gwf<ZuTpex~~6l`4Gy|);v81$v6RPSRi8zH&!_%~FOa-7#!-FoeN
zHS@HVU%m_fEPbF`W<EmR7wH83@G&+2_ShPJjQw^{NrOK<>GI|p9^Omv>u7k^s^m?1
zJkr8za^OT}ZVv-Gw-)2DZ#3$f?(=Dh{b!@T@l;!hI&esJJ>GzdN_gQIidIlGho1h3
z-A!tbZ(GQY)BaTEBerbsk%x3#;Vn{7?msfi-q4$|_X$2Ht{e_=pBqTCahy@_?i2oT
z!=>K`tFjx8ndaKTxnACdm_D^ruC1~pSV(NR4qR6Mv2}XA(uzC7L(myCspePvUyQ`X
z661r-iU;o;<#}aNLMT7MMdpWp9SFu}P&(WuC&<DPXbMBIru?=<r*}o4^OOTjC-|YW
zwo12`n|x!9n?3V|cZ7y`H|aL{&GLa)WB^8(S&<`gc|`P3B{LPW^d;IA*xX{hcJ7|n
z8tEgS4U;`{r_|q?3~gAq`4v#+K+X2Z>gi^Wb1G+u)9GGi2iPIh?xeOvG>g_@Ad3rr
z{GrLkd`V1A?;@s>6eL$M{e<^Zcxh~^$Xj~K;+U5v9km6DxFtwOFG~>%RVDTRHvhZ?
z^oxr|C)2OAvKP@}S|`8zyB=2`!t(6iXOVZ4u{lmavV*@)?1{ahmUdx&gkza^OD%jX
zZQk*r1Ki+tyM(<HuV$^S@uPjOV=fz;ijH4i*8&5(<#}u3+>s<l=Gx;sx$6bG(jVOD
z^$t7PI_W=hnQQ&qu#eTcJ70CPSSW>%#Zg)4ce?^0_3Z)j&iQuk{>FM8AnW77bB||A
zHgv<UXTq~c8d@L9j@To_Y*ZPWl+}DRE|-wVexOs`dq{jz5%7NKR6K@8qpzmmcheoo
zx+K<teu058z`_IY;~o;tM*m%sr{!_}(vz#Pjj{Vh`-$ZpJr5nIFIjGLt&{dV+c@N{
zNM#F?t!U<XN{by3y1yOnxYg!O`lpi1_ZG*_fJ}YE$N1b<#RwlKkd{Y41_yLj=hZG+
z?)&bWmqn#rhy&l-_N>N{ayHnljCKJ0CJ3Z(P1!sKNe=BwV1VfmKJj}+2_X<4leUeF
zl7k~if5G#@TBKV^y(GL|{nKU*ay8bq);Gkk2t*kwtJd)IqdwVIi`6xX(RfOGJy{e*
z)kq6kx~Z!?n)Ya+-T8Wa>M1E@;e7**rpY$xIXxux{k1QaB-y542)Mh|`<=pK#mz!|
zIe%xHeaqcgR|4JVOWHjPN8MlE|1+O2I{?fwFw<W}6dSSIa>&DzsF1AvFe}Ygn_?c1
z3wLZ$ZdJ^8E2!<mrC`u;wDnmM3+GI!^9z0zK1P9%(m$IX;rm%yDGgmjiCWtCd^vm>
zzp}DNDvyJT9Cb)QTYz^fW8pB|m+B(_p(O7~ed@ROpoUa2m%989Y$pkM*047d-(*H_
zPA|OWw6mTG+-B?6c`)rq>(^EAwQ^_JwL+InxxE1$Gon_*z|W`cvmw`C0oZWYs>bkU
zwY*c&Ue2rpuy@Aw<9FL#Qql*LPRa+GX-y+aMc>Ckk888S`Cdqvil^Zi0I`dsqf@8*
zsLK@PoHe{@*Aou<x!Jm^8w=^)I>-7-NdR?;etP|3^UF8#iKCe!qnR`wy}r$vb2Sf@
zze;WSt1lBP?Guxgw<Lh(^*HV=to)dteD~Acwg!~C5Ayrz2_Jp*Y)`K|<m0E#@bBil
zgNOM>l)-@>N(t=kZNc|oAg>~?Q7~p-)=vLPcJs;SzL!NNL-tp3j8vm8jJ0Lff#DMq
zL$v=VG}d1;zrUjM0G$6*wqN!wq3+<LorKdx>_nj?kV1kb;?JA^3Pv#`=%f&H-QvK+
z^g1?w3g;|_xD?IKZ()Uw`(j}kh7B^aRr|;o{MlZVtgTi?x=PWWt<{V~K21Q}m{!eU
z46|AG6Mp%Lg!#WW6><}nbLU(F@9?(>t)JOU=GWtCu)baCOd;+W&-fLvn8|o79=f_J
z_zw91$Z!I&sF{$ef~2X_KgVCgipp4(`<zRxs*jxp_lyaU-Uu$?phKsrJ<p%wR6Q!h
z4e%R7JUa#1q`=4^I_FOfktlv>+%X1&dDEnQe)7NquSrLD>iN>JlTW;x#>ECpU}mMQ
zT9{@x6O}$-E$KWgJRMcKF(V8vw;c=9Z5SeAZ6gtu5-}XiiWVY*)Mqc8N|EO+*)DmV
z;IKu#Yj&QgiP7-3gsH_1WSV|;0uoz(etwZa@9wxAItU5n{i6{KA;cDh+vqL>fm`d_
zwWt3`qm>mOeWkCG+qII-6OovNhMTMq?x4bD`xa-yC`SwPpygqRI)lSm2l}w(>J)x2
zn}LV=&g#XlV<@NBUmUV(4%e#hRS!!?(OS+Br{OP?26_aT6ufC?+9iJ?p9&jDF`%?8
zdbW8UNn?u;UC0P_QK;J-ihaP?%J&(uJ0Ook3X-Iy;w>SbD0^8+Vx>5IOwJs9cM9*1
zm%sZ>PdYQ&VAf9gTbF0%6?!BI>$Ir;sX@n#-icuTQO&nVnsV(uCOVb5r-R>YU${D+
z(Z)Nh#~>QEOGf9NQpe6VG`4G1B5YMryA&I4G5)UgS^7Iaaq)4AMt&zo(9$@>GIumE
z6f-{<(+F8;(2*#uht#Z2=2cu{Zl`>l;c^ssk({bRfX;`H^7&aArMtapm$!64eo7HY
zs|J;qlRI#xW79p<8{kuW-pH<#zWzDxV9!4ULJcF`c?x@n_j3k6pNPrdCwCjZB>6Q$
zH+s)xtG{oEePvwKOX>Z+D~+v?fJ*W=2hLO=aNeA3pjz|nE(gR*S_O4Mes6rD!;=sr
zs2GB0Dnw#P^66$L{X|gX^_<ObgR>8b?EZZTsfR^X0}QPGT)pM9la{Y|dVp3F<V+x7
zRcm~@<cmB<4*}a%Dm(6^gU}1ezVq|*u9%VjV68!hV9xjr_+1~C$YZ0jO5S8;xXF!9
zTKJe^@*Vnk1kCp~8;uD{kW^AuXFRQPRz2g62#A|tz{Z>Czr-I`ymA7qL;EbwVbD_I
zx?MNOK)a5B_{E}q32RF-ZNO3Vo}goilUYwqTs4#+`v!4o0&p-H`QsxXK8d6)T&uh>
zwalmkvLm5(Tx!BiF{%X2$F2PRi`=R|rApkH48|N#yutlQ!ouL&<Nj&utD2}!k@tx&
zRJHM{2{)c!6u>#?Z7z6{sE*@A$D|9B;kArnYmWBEEl7+@ZW=8vUlE)L8lGtW-H@Br
zPtr=9w;&7--P-RZh^=ktq~2+4Z)YOrv{u!5kgzHXh^l~;E`R&C`y+}n=$qfeen5JG
zJhGg7OgZ16s>*PP|Ai)8_(6bA%hHN4A8%<W6M_3KEcD4NTf41IkB6`CQy{hE8o(w2
zpd2F!M@nN!#whw}9UrluNrUf9BPy5biIXnfqi$7Y=|$-Z=l8FFauH~V*1x8Ym(nrJ
z4>?^@YlU@qj%a3XlMByc3KqKeDzxviZFYv;y@<zzz}m|nuF~G`<LQxTY_C*b&|II2
zctbXFqt;Op{d@Yy(zf!8%Tb(jmpTNEUOiw@!(#UpMP&wSwWn-(deg?A6HZ!?M*wsX
zzuR)X<Gee!zJD$#D0subup|U(2XaLJh~`-9UXG`;h8W~@x*y221bOs|c(gs^wr>+)
zPEo}eOR}-ohSjDg@9^%99#`h+@4ME0&IiX`S5DyO_VKQi+iDGxRoz=b;cGD#&K_Ht
z&-^Aidepz=R*yS~>Yy&|zI*6W+UFd8r}gTY($4bwOQCm@nA^B&`1X4!O9iT=Qt6Id
z#ILus4;E4ACTAWjPPkU&BfU*~{DQ0Q*veV{Bq#wwGWq)+h|vU@p#ahCxCV~l2M68y
zxk<$l6XXl-LhP}M2bg$&D1;0@L2G8PttpJPjyw1b+Edd0kZlRQr`>`riX`^TR|u(!
zVkfQIThB&CT3MZ?qaN*Fq7{ox?{{Kqa3i^c69oq0zkbc^%>%n>)za@xUgH<E;L+a2
zga+1Y2Z%0gcYNWZb>sLwhUIMH7mV*>Ebh`OgzGrzJ~7<>GyZTz)_UvkP!Lc=&<zD1
zll^v405JdUB!T;8SKB4A%Lx{@{Y{75)-6kXHr;i|N#47P7mu5SRSJ8BsB5_bcf8!T
zIyH7UP49k?oPadnSodp>T~s&v%EHvOkUJ<oUwrnchM0<$8?&Xgv(WSU*)1&rCGFAT
zK!Pa`HjL&B1E9m7<<Ig5fo+<4V#feEv0C|3e9n-fLX@-^yP{RN=Ychbdvft_PvD;7
zScFB>020`d+|kHq<Z|+*MviV`(}NZXy|nyCN|^8dMuY}kl~+Hochr%ph}%^q(0)Yx
zGx8`(;;T@Kh_8D8&!WD_k1=XSy`H#$!x&yH@w_I<6BQG^x&lN77FK9Zs4X+RDSEiO
zQUg&r9(CWnJ){Negg(0a_{?$*JR}wa8TKm1%EUQ6>AW{%4+~he=|^45)r2`kfVj%F
zPr5!-7c=Z-utu~-<i3d0@<4F#eNuD=5*oAQvne1p0)`kIZLnV<*5;Q${dTGbWK@6R
zb0%AqnD6oL(|(nthlpAhB$3a6_IXj_1A^e{sam1e4_S*1FeF;E>2G8h7GkI`IJ8S`
z9q)$bJnX~pj0x30yDNwl$bQ#R9502naEj-G6?T=kjH?}RsnYstcEunZcl3`*2rO!p
z@3vO7`pk4^sd5g4{})MT9TnC4b#Y2cL=X@LDM3Iw2I+33kyJvz3?<z;gmibq3@r`P
z4JzG@1Jd2nIq=^3z5lZo3+}zobIv|{e~yT_w{4AxIJL0c_~tLhxICXo-HCQ!Vn}pB
zBq78^@UF3|h&(#7Ar%IP_rV|Uw^7HA?7P2Tin*hP6_eOnW3zdG2!U6SGiR(>j^XMv
zr)fjhZi^O9Q4f{_z0px3^3bhWr2&i^4#a1y-L<v<s<Gs6W^6Qi)6o>VevUB?NOJDZ
zZ3DLBKZUto&5KQwIwzs)gT5ABu+9%$pU{$Qb?AO6zuQ0a779Tmn9MW`DeE>B>X~jS
zt5cGHQ>0Z^qY@-`YL51ZdC6lHCps4tHLA+%4LqQsO381X{F&=11C6&CyTny@-PM$L
z>>2CuTImIag%Jx}2=Cpm{quKO6i%RU{t_Me<rpDdUfJWIvPyh(9qGFl4!NUF6-j}E
z`!-YYEii@*H>^1=!RN1kaK!45<63FFi;H;^UUTuvA~r+y_Kq{`^Cp$pLKa3*dIq6K
zJn>4+6Dvk&OgYkve>9!M-Jh7p@KNW{jIQ7h5Tf$EtcCXg#(1CmYaY2xf>^fcqZnnr
z8?ET8;GB`WKfLs#t(D&I-CRA#_N<{y&k-+ww-T1Ob3A6BvgGfsf3P((g%^IX(Q6j`
z8<2ZBIyy>7Pq0D$EbGA=OR^_5vG;nXYbKTPqd#vx7?bY6y0`HMk1o<+3B&K|u69kN
ziatanha%v=ARc-ze<p<1T0vI!o{xvej=PH-)u@+H$zN0^^8*QsZF2!hGDqAYfW3$T
z=5?@byY;%iKX)+p9qF3%*|8V)SlnC2Qls_tX&HWk;I%*UI&&VK{67XrRdJOmbXt`U
z?CahrDU<PX)8w5IHcdh<itCPE<p--}4h|b+u(iz7U`k*qlKpos4$fEw_mnI8FaELE
zNAe{BIxVzcfW+Z@yz2|OGl3BTiWG}}KN|sQa|*%y*$me*2%#=iQ}7NrrT*W--8TmO
zLs3m3%HP>8-q1FRe7rsWB0Tt1iq<m|baim((b9;a<uZCy*AA&QIaRpY8QXWDE89-q
z)<_<V54S^L9e<GTgao+m7{jbq+UOPkZBp-LH=ygbD2O?3i)?uye9cK%y}MfOg)qRz
z+qL4qhne2l%vsr=1a?~7ZbM0z$uwPjPdtU*z^NRBWUs)|g^VnS=X+%vK(Ettwd^PH
zXQ=~WgPo(T?Dn@J=klfldym%LtBEUL!ZJXCjPvALOx*6Y0`N}8d9{-@t?Ex)ULPB1
zx3DHAFo0jL`&;xkOHY~NF2b)ap+tQezPh}otFHU4M`vMp$7SQR!HGz>o;hdydNE+k
zI+rBHE=&=UC<RPZP5{73AL<koXIBpV?fReat5G!i{~V&YVndTB-|t%WTdFj5uho9w
zOn<xm6AW`DN{-04-fo`<6s^e4hyBi&{r!E9#|Nb3HOTef_M#OaX`4ofr?TOgKNt*r
ztGL)6c7)unwyfiQZPSZ^-(V6FFu+XgRXwNvXvjYiwsi26aFrc0@`{N`b}gp$t)dEy
z=_+Km_?&o}BKLe?p9fh~uBiuJ=`mw%=sR&AKOPlf`x@GnbUDb7GA#$aye;i0xsG{8
z=6>IjqKeNKlGj=4T#G;{fd9m&(B(xC0`lKB1M5^!W;Tx4VLS~@-?TiCOYz5-6R$1S
z-W@V*CiV0&vQ>5(^NRE}_|Q+Vw#DWwrn>#@_~*q4XXT0!gpjqams5d{$0Vg^bahze
z9J*mF2NY;ud~vevW^iWH1BfT;3^{?aqt3+ds(lFdZYY&_pv2vX9PlN&zefq&K+4Yp
zaDA-wa9|O-ySrodwF7Q(>gP^%(tm~jDuie*3)M2%L%2mtRPlSFJjB32afKPwtsGnl
zp59kdjrdFjgrvrR{s7Q2B9A*CeRFbLQuz82*Jrf^lWC`Q{<46C!_$`Z1E0c!I^8h5
zYDt1I*{NAZQf2)oH8v%f5fdz9c)?OWc%}YtS}Phel?L~#%X{PQGvNbZG?<l@)eRWL
zI;~M&+h$%DCoi=cfVktw?|;O&Qi6s^T8lBbxE|K;blk?HE?JQ3?Db+aP5<lgm#ww<
zv3x8*7CWles~_irD-xS`HC`^}<9biy`<Bkn-E_`fSWjRZY0*DiGFC6jjT@@&&X;H~
zL{-ypSy?}KZH|>czc&(2-c^1otA)+#mvQqtwUvN9(R^EnZH<BlgX$yhG;xGPB1^un
zGo~>}OTkK;6W$z^TRp6J_mRVDZ1%ownUC^T?X|lxP;MVntO&o#E;lwQyZAhs);#L&
z><q@0>GZy`s}4@AC9>7g^4mz82_X$$HQUq}-ahjc$xnj0vXJ(-VM^>MwLx4WYQ6CG
zfT*K=#|_g`ySKm}US*KOzuc@4b+|Y+g)*%>dR3Ukc|{QC$awSYd!Y*lk-jC1&*=xm
z1m+?1V7FTn5PI^W-JHnU&Hc^A)LPj_j8DPOi`yn}09{}+*l9&njBkwvg38ANZ^Xe(
z?%Q6k_L#?0lLzJd*vAuz+f)u7AJ(`samRATPgTb45n*Ad@Se0>VCn6&y|89%f@hNS
z94?LPh|I!Y{ECKqG>&-|D0#H0Q=A+zYpBe^bPd8Z*ni}w3^r-|GxxQ@=u@TVo|U-C
zT&)x0b+_9?64SN167NnT1Xb`MOCgwa_9FR_NQd;YymXsL$zV+kqpNP{m1@#!GES&^
zIuQ?LR`Nt4ewD-f{N$X7T+`ZSo4CK>5fh-Df85Pj@6RoAO(#{q=NyCLK>H$a&4FJq
zJ^~kg(vQIiQu+1Q%)AE%-ldt%ZzEhZ;_=*)SVq!SWY2M0J@O!K(pg;Of+jA}zx<q?
zr#c^IclEx|`~qwYZO}V@-?PkR0(0!j@)r2$E>OmI-0qbD$Ri};d1`9vuF6mVkIX+l
z8k%Q0R8Ce>SFr=c)G&{q%S|NR=xEPols9yOg{A%XL|3nI3wW194pCOQ$9woASls8D
z8!gWG`&D3wPCWPbu{$YHzK^%Wg^}9H;X<P$6iN-GwMu>u_C{X@pYu*``u8wX&io<j
z(!MHnD9XC1Wi{ryTj9;M{k;~={AW+xE^)>+cD{yM;x^+-aNJUrkUoTYdepqu_eN?V
z+pG`kGD|H0vRe(`N{<GcMY<WzC0nvQKwK;)x&i(Z;A{nYD>|N;0PLrI0WO-@i$y25
zKvWNS*p|s~pP0a~|MKbQt4@qxR6NFTUm&uf-452zo-9*q$y@09Afv!!--E|I9+4Fe
zSPa_o3UleNx4>)9&v1+D$`iwqK1aJY)BkI!+(iIau?O<j^?p~~F9H~Ej}h^iCM6uY
zwTk;3)ViL=98UG0G4dFu(Aj2PM7Y~^>kE(+cie1cqV7kOjJj7=wFY#r649dp`LUsa
z?@zVIoM!pp64cTc|Nc2tuS*7tx&^znYs#7fLr_53S=rPokF8OXDPA@TDI4qXc^x<C
z_2*^4t7YY`2V4Hk>)*5w+-U>exo7Lm#A{c8%8>^9nt>ilKso_~T&`?OJbD7=pKc@)
z`Zew*O9A8W=uMQ^-N9w#g?7zq54GMFuHmDi!Yse5`W!OMQP|OBGJOoL?lvq*p52fl
z*t{RmkZ@WkY_+|Z66V{UdjTXofO_YDW@0J-qr{Op{&UtzmW*9m(sUtUh<}GV9CBRw
zy))Y3z)(^-+~7RpvBVyDG``4vlRlv20%l0(%80yxp+V7LUd$pL$hSjYixM4QUI&LS
zbS-0XLK5CMZv^q9TTh1PxBPrBZ^8ehdfgV!cAY<v&QSxdS6d~#UCT0hbTev|7YXP9
zMd^szwR<%xu-9Z^sYzZXzw6C#*lI3#;jqD7JHs{1gzrJi14hJxNiP!*@`{N+6Sz|y
zv$hJr(KHJQ!amgNyw&l!{`VT_11W&}Lrm3h>~ig}^&6X><ftliZm5@4tlyt<0peFT
za=5i+9!D|cZNU8dp)_-S2zj$qv^VbIRTt!+QoCnUvl9P~k<7_DBhYMSePQD5RbGJF
zCpEgKKN64LUjaibl<GaSat%pk=Hh&no~3<QV|Th_apE%0abq)ceBK`eA;uJ@d>*Y_
zsY9VN<-RFmDJoi^eiNU)nt=T?@9NX}KXDl4`&LFo@?jzk@j<cpx709-9migKv05Pe
zlF0Fdsegaw>kO?*Z+J*n$64Z-yr{iFM`vd&YQS{4V80H$bd484@e2etsr~NPX`j<O
z8mGoSx%+)Gl2>WWv?-|lI{siOrP@4LdNmZJW2+0q7At=6E~MKw{bY`&NhymG^s_K2
z5j6E@*bjuo=dK5^40;n~?_ab?lteevsA#;PoW)CQJG#qpwQV<6=;jVkv>329=aicJ
zF~>`Et4>Lpy|QF?5xW3pfJv?c*W>c`57)u4@9nEVoN_0hmWm7X@>>dEKYb0TyUTu6
z=F*6?7WU$Q!&tO>?aAP&&~iwKNpPRJGt-8oW=wjmFGz3kH<sS^0x;f?xJyt0TsKRI
zGDSkGdzLG{wn2MT?i6+O0l=~K=YH2RD5d$-0!z4g4!$BaP;|@jG*HA(?CJ8&o_|*=
z-{S7nt`n7PV&mZz#PmSq1MfU;H*4nJm3boWpcQtK5YL1djiw`PAADhOCFG|=n~%3a
zer8-_gf?7HY>#q(4CyVAFJz2uJG#2MqG1)k!isPyuM1qlF(Rq(F^RpUZ*>7ajj$-*
znj^PBI+WA<=EhT2G#kwa3SDF-c|1++jJoim2Gj|Z4F`2i0TK#IANAoLJvzRZIL?h}
z1hl@DS^+MaSr-{YvZ}F(<FsEwEn^yuR|+Z@h%|6xo{X*14wOG}Vy`QAE%_?)+d#b*
zlSV1ch)2C|pF!2^b0g2D%MQSo`EJ0f7&#obTgYd~Je-xkv$~(Ut8b~FyvqLXmtLYC
zq3i#2X`QNnhELr##6t4K&X2~tUTxs@<(jp%YfH{rV1oQh_ELEAc(etl8biFqD3v8v
z<F7QU@YLvG490_jxu&R4X$SO`{ZiX52{YJ0Lo6eDqM^T}SR9?vPScI9d+wbc3EI2>
z`V05#w$x=L&D;fRM9*5+kQHRHJa%AwiF|SFc#UPuG<SFa=6E?=$jm9S15%}t#bLlm
z2rXBq$^H*3wMJQ^$D3~ni$6*>{4uNF(I-pc{23>0<@cI0!<yXt7ICwcfkSMwfOUkc
zdb3nS<zP%FFgpO=hdaCtF29?h>wTce222|$g@sR3PdsRd<HH6RWr7>6D|diKj+qxa
zbsLn-W@tZ$&xk#&F+Yq+Vg&bU7Kw;VZn`9@u@>6TSRpNtSd2&9mK$%<r1!C&iGRwO
z>X&@snTSU^=VEpsDB|#CWG9)|lM_FB^iO2Cc4w%<MKgwjG2GI0*@}v)sY%qV_V2KV
zUI^P%s)_{f@66p-=v2?1P;--vzkM{2PTg}g&}b)U$A`RSKd3{U_YZf+>!Gy1VSv-9
zgJv7EYN8pk@cRw;@1%;U{>7Mu=YfgYkiJDXH~N@o5j{?Z&>OA|hQ8LCvO1|cl1b;|
z%-Oa2HcxkV0krKr&G4Y_-@pIQBASJ<MQd52CBj_?**Rquc@xLnVYRD$^<Hf^iXm@#
zbfma!Z-xL0RR2!RRQ-jwqgT%z6+o%7-3*I;6<rBGhY9O?|D+zXA~#*tRDi<{1s!N;
z9NktQNI=Cx59j8$+7|8Bftuidzl~9#zK2mPnikrVbV{xobAFt@qA=YHZdJH4%eBTK
zbJ3u(v>$rb%m$JQ$^uv$tX}QLlPi2WJ|`N+*t~py{*o_I`rzTttr>r0)mW)b8=lzl
ztc><Ha(a<vwI?8U3=h9V$QKqDj7ULd=Tv?I3NFZFGWA4ljyK(bx&&5esB8}2&L#_a
zFXa8-8&)<@&hTsyHHERUv3&Pg3LmFj^>OQ>9k=(nGf6S6U7FFA63J#HZz8(Zw<uP~
zVnTY@wi!F;Go34~xeJ6vO<QBeO10hKsPAK)wH5GL0|i$W40dTBQ&>&|Xkdqjwt$7=
zGA3zMZm27Zdq;%=raa}k`1w<{4cXWgksle}lzy49{8Lg@ExzbPrjEkoQ2@Br0WaCl
zV1DFTO<k9JXssCDu*QG8A=oT>cO5Aqs-uH92?ULyFN!>hgs&%M8sKE}N_#6wW^3g3
zeZxlsRR5VDd{mn{{@{@<|E=@xv(D>lUB{m7`AoKa5osPH=t;f8q%@HvBs*71xTfbn
zo<FzUYu3pJA7<4r^?GNX#U(|i_Z>Xd^qLJxLPR&dh=I5i2#zwC2)Ol)G*}YVGBOC{
z!7LSMN+Ij8fM^xBT2PQ=_7EvNK43$a?G(1(OlG-mpG5t=hnjEfFVS=G_}0A>L)tD)
zK!N4y=9|ebHzQb&f?R$t{kM;lrZxK5N9S!>$qJ@_=2ZG9=)c<uz0DXQY|J+2-RP38
zwUQU9`7uk;EV|#yM7s9)`?n10L<b_!uGKz0Z%lD|H=gv}0emO!PkOX?<A3<o^K?X9
zW{s}eAX~~c<?4sKbUYXY@+LFP0aRjTErqYJid~Kf3He|h1)DkPw|j)bA{(4%u4OZ<
z03zp37yzwxZX#@#?F9aQ-Rn?b#1yw%=9Tz$O1`MGIwyzKsm>y;#!+~6*IqH~sa^Ym
zj9lv&Q|5vytegY+z1Asbop9WOZ5T$Lk>Y<ABsY_Lw#JepTOO6;U?eEh&37QX?FGY2
z{b;>~Aaap6<kHC$zSNPcRjDv_W(PaSiR<jUtOdrJ@`KGg!8fbdiK*!m4wSv`pRg8*
zX4a$3^o6Iv?~LJDKb+q0yb;yQ*>4di7<SvCnyz$f_E-jzbk7&ou1`HokW4ta+nZGP
zSd(A>g5%x@op<~oTG0$KP}haz03PZ!z}Odo3jemWQ2J+w?WAf#*-KmAnO<3DOE7oA
zsjCaNWxw0Z_4LRAk@T&VCuYNqpEndNvd1-R1ksAm1LqcVOQ7)2o`mrb{~H5J%z2$n
zo164!^X<$nZ|rL->Y8b8tZsNx$&P8Mh3?i9X~+<MhgqMh<{SaC6Bo`zM~Bx%>DX0X
zxrg`taa~J=ZL;P{oX&>4m~&pbR7KcemUNLm!X_X%s+`KRoBQ`4%~vvd#Gzl@DEX@9
z=5)b<+32Oij&K+ZmBa690_LaUSckO|AkG<uzA8*0prhJpKEVCP!Z@Q)X$Qn<D*Jj3
z(T^h8pz;frF`)ys>F}g)P=Jx^T##PVyl%0$r@qO@o3qpD*=|$D%PQWmX4+kwZ##k}
zy0P;fUSDB>H4d>6E9dHy3{+sI;qi$lGJ$N|mR!!z25akt`ZDxutW^>;Klp?&H;;3^
z7RI{_t>ZZk__UE$x68X3!}AJkWFh{%Zv_}Ppny!5m;9(p5ir>Tt<fSg={{)igLxeo
zR8BA9l;9I1ygBFk2nYs1WGUUfzAWg{pI-xQ`q*qi{PilWXPHuX@2scP-6&r4hK#M#
z0RxO47U9O4@Kcy2(O^`UfApM$hIkEW7b`Sdp;+UhXzJ|ZLJFMkNJ8RSh7}KQZ*T6c
zIx8$mgJ}=^&l&*L70&4GD3dPat)PQw-zye5A$YA5((>d#A4xHEC5(M+aZD95;qq&(
z3lZm_mQ-u&TlLqZtnPvRUfQWiAr^1@n??FhxvPW5=SD1j0B$d>*!b)aa47-mv|9e#
z1Qs~iq(_^qfA;r{?<Q4QBJkHxH5`i%#4>?%lVtS<o5JHKLjE=vjR{f{>dY&*Mb<P1
z4mV7S)3s~SB4`5d?K~EnaRp0^7q<yRu=z8oAgv$uQ^hTE&HO?-d4v_lf8Rod)?)qc
zP2pj0bB>kDKyUwbOB^>IMG9v9)i&L3bD+T+uhK08q7H`)+QWku?HGB42W|bHQzg1`
zQ1|vzi2dR||Iz{(9hvRC`KXtN#BSVDzm+pVjd)Y^z`f=ZStJdVyc0-D0x)2jwlrma
zJ=2^sW|!F*h+-$ug#6UJH7L1vD^nSC?9ZxW%Mp+-(v}@}9530uVRe$smsc=&RymdQ
z`pYpmyNS(gCghP>Qq$-lf6r6`GF$~5HNG!a!BM@w+5<sZhZb+=b{XcY;@Boy5=sC<
z%U{5K0~~aKwY>?j=7nuvEZca;O-ThiSV$E4b4EVA$_ALuV>$ztOED3zXqS$=q&p|j
z&8(+%;D2+4DOT8pox|?eqecQ03-^7LUixjTMDPCsb^d+rwG2j{I#fxdqKlpyhUlC(
zt*@p?Ux-<U9y-plwNKQdYEJu3B>w|o72*H_lYlS32XFy-jmIvu6A0=NDkO#uTWQ9&
zsbe(*FEElfh(6xJ!XnW2uduJa2szLdPRiOelKZewGK3T6Y@gkr#W~<<x23e3(vYSy
z^{Yp>$9!F~YYfbyl?MSd1>2&*s~pjN(8NH;q(U=cbC%g<qG<l{_%)s&N|Re?v|k5C
z6=?J4!$6!T0IDqklx;YC%~e!q>pJOe)KifnN-v5YY3xodcohb>!(cU+;I1yIb7;HW
zA71&IV`B?*m-}Y?FM4SG!J4_Iv01o0AFBW6wyZnwm9<cn{}ZCb6r$K}7O$Gymy)l8
z>0JvwwFWYGpC`EmX=_0u$Lo_HW<o?(RUE}`bFH~F*}qrFJ@wnJHGrlwoLNx3`)~0!
zdq_6vX9~Kb<XGM&_h7}DrwVhawbM&O@v{ecg6E(9!_ov>1$zv5*Y_NTP+DL(Ha=<F
z8mPJ@(HOVUR3@Whm`sc&scNSA%D)>*W7d|`?w~N_i;HU+ayO%AazTQ&ShXQPp%w52
z$}CYQ3$Wd2Lie-?qvmA#nC}#ceoZ;(K$Md|mug5bLy|CO_FI4RX2zcgO{Ej??LVeH
zBVMMVq4^I;wJo`>UhGkPO=UpH*bAI}Nsg`@w2Z+7rM=turYOD0BKVv4H#LuFlWDmO
zg0AVqXCzQo1RgvP@IitH7c#=FyZ-oX8-#q;d9wjL@pgJmw5_x4#(|pz?Bvu>?k?~Y
z=AdG^)TnVA{#THzRw<1MUz-t9mxfL?^{I~wz!SU^>pQ#gA(B2%e033F&xv#W$DDWd
zek7&E!~WeYVO^VA>X(?@>s`7=r|#0)cdf%aHbT6EH5<Y5nRO*lrt(t4rJ;QTECA;H
zEa4SnqElLdDrld0OsPKt%K;Rm95Sy{;PzM0yi3IlAie=**F7+mc9l3gH7MBAIyGQY
z-jI41{sjZT$;1onR+>kYe`=OCn11Ihn_TNIWF^Fy3V(kI{F~OMzN??mKCb&knLOUk
zov1&atGEB2X}T@w%C?n%b?7zQ!AxgPf+ffK=j9>8%|ZClv`9mVcO5$!G3KjuA;`oQ
z@$<=#H=-x>nfLLj#OkfAC=doJZ0+1{_%iwLdpH%(UqY3eB3p+Cm5cC1cO7l{dZ+m}
zVc&}O4~k#eGF3soZG7?rN?gS+h>phWXHFrawGCMcc9y-^J@NklXu!&}2upz*H!k~5
zOK2B>t;+(kenYq~bp4)YO9uiUEhXJ2q<jDg$8c6GMaQXL=i?fAXBdF?eCW{v?2hib
zhF>+ay$lEz){YT2;B~W2l^U%85#KG&j3{gY41-rxb1yxtCfL`vu%+83;O5aSl`>F>
zzMKL!6={tw66VsopZ<kDO-DkTIp5IWSHX^kzGvt1mH%sc!j$(eXz`NR>fF=Es5U~+
z??Mm7S2<p4-`}6q?MCq`I8)ngN<Byeh%qKuL0hBAQ35VaN!0RGf(_|1?jlVKhqsvi
z{UQTj3&~CWbzA{rdhfB{<6RS=S^)W7g4uyh|BH0B?3KmM=8UkWMfc7X5u1Jzq*)?&
zJ@XAi8bi}8bo#rEbeXWxm=M2s3PN_i@7IDR5+IA704Zz~9uE%IMUdNq3!?y=al>zC
zGl~+Mo7u4NR@OEMwodIc6Chhb+<4CohF|UZ8hz(=awyozpcWyry6B<Sx+F+$$uQH-
z`6exzDb|_yv9o!-A1g#Df`BLDIq)lx!uf%;&^crl;(C?YM&><W9L)>=qV<iey=eAE
zD(3|C9rxvzs=D8<HF_`${R^Uz7v!beV_dJrY#v;xKIRJn{kdRFdg)QT;~=>^cq}=;
z$qqIr9wU~l7;tWgA8~1#&i8JvTkTUyRHIQkxUkGlA8QF@kVqfjS`gXw!OFX7n{Nlk
z7Jhe)e*3Xog;^wjAl`e~S9!sn&u)!sirwl1!*h%;%;Nb@VYdgTt5;*@V`#?lpco#r
z9_k8A;?zt)GzH|@6Xex$AHbuUt1`J75}VYF`EmWzSxT^Dk^w{h(18fe2hdS~c(U)O
zRX|n;#3W!K1@)UR?u?2(e?L17+ZlIr<fw&&ZzLV7z<Cil&;Cn_>EwaaCzjci&l~}D
zB+&lQ{1Kk~9;B0PDn-YLwvgdq=Ejt$&XxGw|Gz>=N*WL;3wKn|X9;{ba}UATtZs8Q
z&H1^%xzci0K@~BVzZR1w2fk)6OCH?5U`b-{;AKmL+P*ebCpf)p6Q{($Pb1u!;0MbK
zuyzkGzv(8>F^?E9;OK*&JMw2wWS&UR)wm(#%@uFhJ@=7k!2ES?@#GF_H(t|H(=i@4
zuzPb{kdCI*vZtmpx?DST(HfPMg~q_(QjNu{7J4fq!)7s}h-Vuf#Vj+k2b7io>k{*y
zSmp?JB4H{6t^H($Zv#Qub-u6QKW6o|V-BGl<R(<DRRSJxz*1AmPynRLQGrz8Y(n*b
z_JcF}Z}BGYel=2^8D@m-kjUflU*19QZ|6FXHg`}lFmK|q6&VsX_lS*mH}rRdg&G49
zXj}my3Z9D-7HwK^mMM^=Gp)kyH91cb?FHPHD|JZxy^pKwyW)jDAcW-h<QLElUY+Qq
zC#~nKS<BxJL(+#=;=&D9KXZ#LG#UtHe0a%iA!ST^w;D3x_d$fvNuLWngg~jNob`nl
zaKV+~-zeqxZHtNabO<SyNgJ(sZ=qfE#qr*#<-a;%Ps8$1`_7hnm5yKOYw6Rgwvssl
z`#me(`(T)2gGivBEZ$G&`G=4-ccEfGy&6P=Ku0$cSCLV^WYM|)g)gYU8uVSg^(Gyd
zlJ`juxGKdOmv6^rUW~c59#UhOu)t4J@LWLLTL@+{jbZ;UJrcA6gS!AiH4*<wHD2YP
zCGA&|w{5oa%9gKc8PR9cM}iDGQ-Y^2H#f~5`Q`17Q!0RAD=^_VXF^h1%Y|7Ay!nIz
zgEiVO_o8A%z=-1{RO0>%ns^<1ffrv9M%h;r(N&f-;3IBl_f#|{VMovi>vC=j{n33i
zytH(u4s$-(I`ixdh5}cpf6UJ03;a}Z*#+SB2j;N=hVVBUMB9y#{#Gexh`}|`NjsUz
z;j=!o%==0W-b(}d8iE*tPddD&pGLl3WLMeR%_wIqR^N!Rf-u3MtLLtZXI~hIL#@}7
zf8^A@d~w-0bu#Qa+4qi(r4BaHPMDFE8s%_Z9yG?1AT4E9CyCVd4cwcCCQ(n+>$4MP
z(O3;%XB&+fCnV1>8*7Ic1Yg9;<&_?fDxdy6&X#T!=?J5sAhVT=^zdCMxc-%%5@kN5
zQN3IlSQ@w@WttXHH0X9Mc-okmnJKO=gDilJkC!^gmx}B8Tm+W#`nF0j*?SDPWTott
zqu~G}f}q3ZEXc*bk6V2G40B}1z%c*t81S@_<UoBqKEAz^hAH`y{`i7rkb~-L9<ClA
zmZ6j739V2>q-d(4vj}r4xtTStpbLKK?dv$!<?grejaD6P^_LZ{#^A}KIde9g!`oZ0
zoX9M0rBJd}AgJi*=r|Gh^fcok!A;`KTa=(W!q}n^kIJXFs8<O~ncq1q%*8x|>=-hn
zVclim=N9Yd{yP<PKma-g&Graej*A)^f?$IbK*T-Y8J*!mGc-U$lkfUq&ga3UK`6d#
zmXu7G^zeY;?5WF<L_oTmtKitpd%NIQ%fLhuHy;(*v<e`FjH+o1=8vL1``yJVJJu3Q
zry>euy~W--n<_q(+SaerakHg!j^hcUUHuX#Ixr*r1qTpLftxHsI;kH(<|kP(zn+tS
z5S;G|IVWNVO3Q#RZuzXj$yKtHEHYUR<#YQF7b^)?!j|wSw_1jg%b?G$3NQpiDz$C2
zsa@hwzQl3OTG!9xE9urv&1VhqCxLU&n_>(~!s`nIzliI2!if<xwc8gnB|AJ9uuG}<
z7jm@vRk6Xh-@9!Fze9#16$*Wt>0SxMD7`!^GyzB*F(^|WaRwl6qC5n2c<Hyrzc#Vj
zTYvq0Z{GK|)oUr$iM822d4dXuH+Xm`s*SCfWQ<#Ps9G3fOju4PIXIzaQtKB1Ps_JD
zf-VZh6F0Ew*45C&ho9jvG1)TV@TOw`xXx8G;kq2j!OYrofgn0QZmC?BP|JY!4FY)*
zfR=apT1KqQAw3nK?;O8zWZut7ZaW`KHq}`Bo;40M@RH&8p6*G2Ly}7Ron3($fuskm
zY~&OV(*iqhX}LORvml%>yH--*C_!js1tHL36N;7$owkW<bIVr|H(!AfY6`D5V;jD7
zxGC|cM2|e{bM1<*hPhK;zKxk10aTA`2lHKUf2oJ;#6syt(l&FLuKaT$y9{cJdspGC
zDzZaE<2b5@6cAz*SoZQ9<DN)Vjz`oYz}3kHQFJTu($=BF3AjZhoxFfH8j!KW0d@X=
zg9?F;sJThk$@O?>{kAesg!Vt%!t_oj-@3t#!!)~{vJg4_3!kSv`|cnRlUV#~uaNgf
z!QrLWrdM8%tpB>|r76krzhIO8ee0QtfsN0#FA?w}`W5w)#83|j;)>bY!r~YD7gFmz
z?7M*ck0O54GUb1PbWSe;N8NE&tKGF$h7C4De%Y@q+?k;_;-gWYunp$UUrI1S9iL67
zvY)H)f3ORX4bk0|6l*h{$P@xCRVfm?H$wbFPc=DYntq;kcrVWU*qW()fwa)>yV+Eu
z|1a&edP_|@u|;{3nLQWg-e@Gt{-n!bGWtsjD~`wNk-^+s$05+-MkE)rCuvNuS0xF4
z`qN(mN~duGa%9c<5&nH}a@zCl24{Zn7@uaJ0c<Ds9g=CT)852tG3NJU)r2%?TQP>F
z2cHF}L<TWa|D7)YUU5_zadot~qJmAU*LWyy*EaU^m6_yeJpe00P2}#k1Uf<eW={w|
zl0cH-=WfF<Z_QREPZ7h8-aA9YOvmorKbQNj@wuy)QvUN*bNo`81-e@^17M3lzgi_|
zaLs$Cl1t2y{yXo#$<LL#WtiyuE#C#+bO4M+RKbs$<I(l}3e>j5`h&9_xC4QTlWcy`
zy&c^y{Q>;o#@pa=;Nk|TOw{K_*xv>fHaFd*2WVspym|u|p1;fZ$lHwC!h^Cxcc47%
z0AzX{zzhUZ%iC%DgpXv027<)4O(%Hm?qShv#}itKS3_{dLrxNdlg-3~mOuf47V%l2
zHpWz7eYI7qSihdLoUKmAG89#}^AAtyi59Q1aGLw4fjiPo=ZB__(+nG|ebY3fp?6WD
z8qX~OIC;)r?5*Ht>Q9`hf%yf^(1al@N`^i5o6)=mJc2UmxYFbMXj|@t7sSYyUaCa^
z>qtt=YbeHyF;T4X2dXaalX!6Cer;h=3#e1#Z4`n{l6#amX~7tGFLHj~h7TK^fP?8P
zN<+c#CzpBsD|ccr7BxkLZSkyiy1Ddtp_dTq?bdm%nuG*s4Vd_?;79p{sj-IGrNzY_
zAS8I(*kC5F4&CxavOq6CJh@?)8P5~QB!&Q#D8M1>+fb1DQ*dXHiu-0LNG7(S>BqMI
zxFDn^ZdY3u@R6+pi7T3EWd-;yHhS*P*0AbW!>zI0y!P`{X29SkR|-GfS8_U#h$Y_2
zaN%z)`*qgR%cv#)!zZi%`G6w7!3lzSH{S83Q7KnG7bd51cP;zvYOTtol+kKX;OPru
z9JS;tt5?mI4%=BmloJ~h+_M#hii>4t+^55RvZm~k^cuh1HR7-tZSXC*^CUCJyla1K
z{271V+!0THj?RklNjUbG`jg9svSW6#exk*w7NDLhYM4hSaxK+iVpO>OEL^LoSjIh^
zC6r7MUj`K@=fG5Ey7p3^jlq0!j6>`_bH(FT^bH5+3xN_~R!6B)bK+tF(f)GVoZK#z
zUUdx>+nd@@>#{L`Rh!Ux?qxQkgd-(W2Y8u49p)tB@?YVjs5W?S0_h?T+Odufje?Kn
z_WM`%hC8XRkL{>GF$6bz;5ksSq3ZdQ^*$6v6kU-RZCczLotH6BSj_iQ3|v0rM@Az@
zb0Vrd1qE15sq5{AoOfpQAy(A)|8T7<jezvh75LqS5}{73+g}zU0_V$Ap?B^~w{Ub#
z3B?NDBNv*wFPsvfPAhv}kW<d)-1IRk{ilI9*)f4%HjNxCg7<UIeqKnmY$|0{9v`%(
zyu8WUEAkZ<G>>`GH@#coFJ|?f`pwQu|J51`*>kS%S|)G6sB-#q!cF2>J$ZT*<TQ7m
z#_BV*X*0FziUyvw**KK0{rVlNZefsi8^M=mCyPUfv|10veya%S<QORZYRcj>A2!HD
zF*XRsJRaT%p!trde_+AU7}kUYR`2h9h%)H=KTovs6w~Ep7VC$6!iIfQ-=_agbkerO
zEYq7GIJ4y2NlW$SLohkpxCR-kr&+ZSAtIyJgI%F~4gB83WH?3aBO_)xJD`31KW^8;
z_@NwjWe}C8NAtu9sySI|+;ilTD1PL9g-9NgzbLotKbjFskQwRs*ojRc#Ip1J59TrY
zgQ$8uN^8}%p>MNluvwOhW^?#$wmrxB+7Kt=PXsw3O&w-<pYoae?2tttz2S!D<?LwK
zw!nguPz-Ww<Mo;lvy5bglOaZ9dF%yPn20!#>ggV{Crj&^zpURLrYE&y@vDUJ;of4J
zfvEGiIVg*yEvr-$M#VvFtZiA7Eh!S=p*v=tSj_Y39X=KBmn)4$<*Zhct*{O)b^*x`
zJFv;`d%htrcdu(Zbk3bZ^61g6@vwn8c}4;wa0J4ey1lXqg)B$2lwJYbFvK@VcBkDW
zSey4&{W!dVar#BAoQWIVuL~eNpYjN+7`dU~X8vj6zr@v9^n?C+T5iQ7RxnLM<tgJw
z;Z^A8$Hji{>~t#UMSZ^bDi{@i<_IBG#F0Cu4CbCE2Z(4(P?-~)^PXAdyEM>*6VLrJ
zig|YgkUaxAx;Cs+G!Qj%tO?Px2khW}!n%7|$w@eZwSH((9f4I|!?V`TeA+TK>ukt0
z={U~WhauDgmsGQL8nA3OEaD;uDmbLzt|yasq8^s7cx6&m)FW0TLW?lJSez+IIl!e6
z$Id=oqq5`rg_GIuVlEE4RpFIO!OZKgZ8eN-x;Nz;iYD8dU%ofd6lj6ueJ=y9G&~Et
zlPC4A!o=&7StOz*vR{+nlt3zrD-B_|iMguy)X5RcjOAI$cQsP7=c|4J&+ubJV#B_>
z3UCGx@$LHJwI}p>z@t7E$?{3JE>=o7HAHAk+z?gY{t<lql%um^wnvcOX?6&n-f{6>
z;DmqZ;<!Vw`o_L`V_&-uOT~J#FkR2y*Q1V=l&`TAim3@dX^~kIe&;#5$%7oDN_cp4
zR{qHk_SD7OTV&kZsLJEWhmi<C-v%aXf<<y}!D5$VzN*OOLpLN_bKyEHDJ$Sfbl?L2
z9*gJb^0|SjBGGOudL5fgL&yY<9eRuh#?wJ1I$Vd$e5G(-i*xA(KZL~OuBP5;=&{y>
zU`^FuCEbDW*!bRV_h*wwu0{fa%HK%*Rkr%U80he%YHj|AhdM^C4<sWxOz5FJ>pCFC
z=QC;?2wv%;x0MCSJa*_*VyQxCoqZ(DrMIPS;Zd~~yE!vOT(<oy_<~`VHj6q}P9vXO
zAm;#g>%nq1!WV=)30qVMb`DbCd|6?4zM;L-RTHLrY4c9pS6uGNKVGzq`qe?jKH$8s
z%`XwA+gAO(b*?S9e$>UEpS=x*_!QU(RWor!4PG}teub|tekH)u@~ryRxlvCQdzR#>
z%}z*B>o@`lLy{V`Y^@A#T1B+4T?}s(iwW1_M`@UMKLzA+xNXyjxr@gAHnKDpht-TO
zM($Hrz3K;1z|2+$Ocso<=5|C`$W^{Yd5FvDPZ-36rZi6Dz&@8@sUn7Llzu|f7l;PA
zZd5ZCADRxz9#a22o$MZ?6o9`y5jS@Co?=a0jxiA0jl#Qc-}BjOyUAqxQYK?Z@bO#+
z`}Y;G_Z{d%dsXs(!m0<cQ%AUtfUPu89h24cKvJsi$gR~^cc5ka%o9G;r7;UoD3-hB
zWlGvGCri8epsD!@Jac${Ud?e)!5WRrTrG93@2&r$+$Mm0pwaf#hxMWC8n*XU6wFvu
zF8C~c!W|PlgEs!f{Lpi`9XDwgGR}&peLu|-<<+p?^0R5rc#t4E*D{x6MX)qGY?ZBg
z)4?(dL^w8UnLO0hco0FmmTtM$N>qGB$y~pVz2qLs-CgFEv@Fvn`(S3Og&x6^e&K5_
zDch53B`ANOV6F}}c&R_(Y-ddmyD(NuwUwv(CK%lcQF3X0*zsl5+3y(U0e^i>N#R9~
zBv$WQ1;Qr|qV-2#Vkn+3CK?JpqtU#=b!#}!=k1Z2dU<p4L)zE7?w6lM_Wc97Qj80P
zPxoG&{O!_|Q=Ng)QU|v1e(~5pq2hzohdWAFd3YcZeyj+Eee8QfG8`i5Grl5TL*DsY
zqlIS!_AKo|ieQOc&&uWjTtSL2d@@i?rw`p{F9qcAAl<IGD&7CYUgo8~wq!Z-eRP4{
zFC!P1dw}BB@4nCPz{F*04I~wxqUnJReVK46{rzyru@VBTbbvch38hrfcX)Up%$PF8
z;`lqB%OVF-Xp)$J&((UvT~u$-vs@zxO$Z$@M-7;wJGv^+0yY>UvktGun6Wy5Q6_2-
z^4&N=cX?B0T{YW0srxW+d5Hg>8ZW_s*8W=`fuZoVYaat*ouxElUdT2{52s~G0!ZuE
zl~qQG?{c^X)vO%W#L1@$%r6el3C5+2#D@9)w=UBA#+&#zZbNSfVRk1E-u%Uq6vk#u
zPd?OV)H4C<elujFv?&5c+c({9rk?SrAKWO6C?_%4Os_G!N&q&yWTRJ%nq8Uo`r2@{
z&+S58LM!ilZs;+EK3u1;;Exfk#CJLs5Fq4aA2U7%FIfJe52<}&J~Bl2!`5hu#xmq6
zjV)(CL$3G%5GK+_%%LuTpy0+mOU#4DvS%ubGD_V8a1X%RcoK_5JRaq|waAk4MM3c0
z{jS{qr`Z1wRzxGOxA7V5Q06#9@NJ0cnOWgej+`_h#^tf2+I5v!UW4!PgKowBAd77^
zk4L%8{cjPmp*zW=coe)XuQQNut<4WwZNhYe3w6lvR_;f$N;Z;mmOZ<?vaWzjsQuyk
zMc9#Z-w|W2L2^FPGgCLKVd5PD_5LCAo?+{iKf~MV3gq5c<2Rh%E0>y@ER!6chrFg~
zzcC`?jS=*!85H&OX?fz7R_nh*pX0Qhi1KWI;g@GU=Q-fIlku()*5MuBSl{+a3BbV_
z!m{z<#9U%kDH<Kz2$Si~4JgJ>m)P`W7JfzgJ@VTUr3w-zM|<6h%?D-z8F4UP0<znG
zVr^r=-$sX*rTc^_zQpBg@Yl=|+)xGUAyc{>x3a~C3>+i%C`sBcD)4{r*LI*b$ISU5
z=TM1$K5h*Hp@gb7Uq?q`pgn+QBlv~e3x*mke~c@7`u*+Kb&VQu)V1`Xz;mK8JJ+29
zBej-sXKzFp`105=IS`Lwq9uIb3=Dik+Np2A>xz;o7SCLnGXrCtKPG$mduv`zslYI)
z(`-c@3Nc};W(iFj8C`sx_}s>1VA7>yiEEq9HuBfCZ&EufQDs+W0s{mRW6g32^TR@O
z@euwU;saT(<@rz-gv>YWXL@@wX)FI!V+!4wX>1!2N<abeorF>?47;Rso_S98w6vQX
z1@@ks&%P^PKp+q~{|V1U>A%Q+{botR+@{R?V???ufK&0W(fHK8n)^n!CHbd=a5j#L
z#wofp7Mt9Jg_yc;ggIXG(Zyl<*F-;Koa3vLnRcI{|GJ&O71RuZeii%@_JbNi<@G>3
z(N7s(m`Y9&^WW^tcU2|Q`Y0FgZPupO)aQLPMVmJ>zz)NKMvIX<dfXP>R{h0XHs7@!
z<N=)YT#@xDir0u|l%53!Q4On_b{@S!uvNGJ>NP>iNS&3eI&IS>FFRoD3hcOBc_Kw&
zmgS`8EhBEtKk~-nK}f=k_D@LcC)uO7iun)Dh~MtL)ztE@JSFwI9TxJoW&gd~DUfrG
z;1Wf~$7o6dOH4+uh*tE(M9)W4r62<7@HpF>n6QU;3&7h9ogPCkhm9pL35jO~ow%#6
z^;@3?ou(uY;zTbVe+pSl)gA*j@@~C2*;(L0){2O^;}Wv>CV?4*WmG|`YI^%iUsWT)
zwiELiAoar!PbyjC;>k?T3Z1tOe$D;UX0|X_xz}8Hu>zP;U)fVe-S~*>tAz>rxHile
ztRjxD0<Y$$;wSufTW^_uuZO7P?*iC<o!}v6U`;JiKu!SjF*twNh=TaWO#|l(A>pUN
z(TB#P^|T4xYU+uE`V#()YY#gaY%0Fm@>BLUqZ9eI;?Wj5hm{xd{G1qMSxB>ip*{Bo
z<g~2}|IwP>(;e5v4~l=!xXwQ*SIW}vzu?j<oB8aW8_3zP3T${V+1fnr$m$X{q%CvX
zwe)F3@B4dUAPJ`ngq|z1o>^SSJe8MBuU)zF(pK&yZc{v3OnDnoSYU%rD6fQvMJ)(1
zmF>#CoF-pJN_Ui~#j--x92ecP<9z{ug{hfV3}&`1nOgqq?2q4%#>~-7d#7>A&4Q?T
zUZ`sAT>RCSx01(MU{bx$fAIojA<Zl^FhzlKk;bPuMC#J}vV^E*O}VKoa&6~6ShtZH
zt(7FKrILQy4;h?K1#4uA7ol(ss|jsiNM&7tT^@mmch4v{W4Z!O)oJ#OJH$m2pXnV8
zvp_<4oaHIiS$&mv7d0VtX`UFs!2Soq23TnD^&f%j2hk1J)A7w<d6kI*8~Hp$=h1pp
z!e@=En%EXSxKeCPWDzE*stc17_eP_Cz4XXZ<tXC`;-%*y94smOJrR9iSjfUZW}S=}
z+5Cc?N&tuI7Ffy#pVJA@bN8wwcX`9vhouhllf2!1t|;P}wwrOL=Cm!T$FI%XBa9Wb
zvtu=tO{?Ny_9{t7KA5#rJD?8Vfk?1bR5-RbOX#AnDxPy<ZGx^rs4xlxM<C1PoW3{G
z{GE<f?j1{Ei-3#;sgDU}0k`m)bEr=J=Aqf#eakEK+1=L)^bjH9k*(&#FzXoN3EH7D
z*_y#0gA<m+hCh=enT1iJz8oL;Kk)^PY(v%ow4gnmBpIk#7*D27i9Nyq6Icy_N;Kel
zeh|V98c6+m&ug)+tN{$gPf)PG=Mwkh!?|>w_pNL{ZVLsxZfor*aEUw|5<OOiEcn{8
zRz1HI-2elPe08iY-|;kj6Z(xCjHugYLhkvM3*du7QSPMT_8Vc_?CD<09VPaJE$6K}
zuFN_@%Y)*A$eCJW!@injyp^VRBABG<G(Vswr4Dz+kK&px7NZxw13)-2JgQ~XSNPw$
z7OUsaO->V5y3PSp_whEr#)XT%48Pm+k}5qMzvlANPAtdiXXqO8yl=16=19j^YTiGp
z+^y4TCk~X5K)Djf9RFDxY`k_3<LG()b;g-PV`Z%6tp8{mO%G}fTIKop5kFgKFl>8H
zM(@yK6aYcH2RwZSPjqw_cI6_zuxorJh<&QIDyfG{S!P3>IPr{<HYAIh%-Kq=R!uL$
zMpV(c*V(V($r9`DY!KTc5W8Cg)G?fir|z<9ohI%Yzf)sRU^a$O^5D+bI&-B#eW(?L
z&!yH!cQAG*42S^dWi|F8GR{%Ers5p-NdS)sm$NBV;y;gwP4ik%+i--YoV$lkJi(MF
zL|F)trbCPw(<l<gu97{_V3`N(iFHXQ_VZAk6Z^J)9%sk+pS8*cpKDrHM4nQ{xWl)c
zY-Y+&)E%V6?U=er)&Ls%c>S}{UZmytA0p8(H)>N=$>HVn<s#q6&a48WViYvhKf58;
zk2qBGE3MxJTjqUJln(Km&eYT0uE<UcH|g61^zSb<G}C=HT!!>ZJp`FDSASFj$z|)E
zWJj_oaFYHEHP5j1KNUpO%a;o6##Emj#5nUM*pi7LV<pC?+R*?0Ajo>=s<cBHHp@9!
zt!06Pj%<TCXF(#J<k)!nl%liS+YG_T{!xZ;sLHw1l6+lkS!Kjfn_nel;7q!0C$1bQ
zn0WJgygmhAaeIAW{OsxY`zQTMZ~w~x`26<MZ-(2{!|!!BZ45go_*2+%KjYFhacHc@
zJe@dJ6J<+M*t360^Id9ek5NqAV3pmt?i6N2z{}ioN9|E7k^C%?#LvTnB8z18iey^0
z`!hnTZRTt+QHxf{&8ZYnM1B}o-CcGc0|T)DE+X9Vg3$H6L&Pno(PWeLz*j)3O7Sct
z&j=gmeDRh?cx9rJ2W^jc*YvkijpZDYCYjnguxB%Bf?#I(^rvy}E$Afuxb$BL5=<uB
z(@6XSMi5{Y^H(H!q~jx+B%F~-x)r~wh4{h@q_EWnSL0F6bDD3r%we&4FgTwq)?E4f
zT7-ViPvEGn*Fvz?87Jp#Qz-L`H1Y0dC#Jg%(;iUV5U$nbwj~yCvy17n)2#7xt1>hh
zUEa{&z7`7z-><}e_1@nG)Q8(O-VAPLwv>r`2eCx5{~#3h*n`bG#jUpVv><4t%7#KD
zWQ?`l7@Ys{0r^&6EkesE@VI(rp9Npy=)rH&%--c%pYj9WcJ(zB`dfWeU~>0<=94lA
z>6jxuf3u}5U(#>gP}qXfpLcd{G(@(x;vN={Zj5M3WhVH(z5MnO{RPG_(2qS7yq~8W
zpxy;m<cwxsDG)gSfV6^i@*l^gJm*V#y^8*@Za3fkNKqfY-%Z+#1meAwKAit#3#jM>
zXx?z&9rgE%97!AU_+OaI%?Y)Z{92(38Wt?332D*(YoPP_8*YZJBU=HiZ0Y^(5J3wk
zWx<M}b<vNx@JKB?qiy{?Pik7#Co7!zgJ^eL8imUc{df5gmtUUUj08qy1RmHL1Ha#o
ze!QyB5~2Rc7|HIm$FXwqVG!te!ghW)!0fw!I5^~-OiYF-P`|Y%IQg3S9WsDU=c)8>
z?Xw1-qBxekUYQP{8d0}VPiME^mOT#I4`kKsXsxP5^Le_&jMaBRhDMt?&Mc5#PbEDD
z&c97O*?qCA^(yM4uwo+Bi4FN;-{xC2Dt!noi?dCln-;H=n;U8w7^XkHL#ek}Ed@wK
zNNCr+T%&{0Jv}?d!}vB~eGA=;)B7NQf#5Xyu<TJ-U!AXc*vvQ4G=pq@x4c^XBtxX3
z#?n)Vi5g8VPqFQ(exZs_Z?YstBI62VaBX;)bjkvUib(jYz*oaFt<{_u5Do@W=MmCX
z+ybL<;m|8|)GMjFUpRQQnWn^BDbl>)mXk4QRc}#?iZwn4xoE?W{P;bpbe1hWu^Mo?
zkT6^*7*7&aAF7yB5j1648tzvAu~1k|etmiJnyc13n<+Z@6id*K1Nj-g_$5rw6d&a?
z$WeEaG3EFu)Oso`NQe<C9L<-nEGJyPUcxe#z$4B~NwM_uv40Nvuv`CU^2yOZqUfkD
zO>?l=_O{Kz+p5yUTKPfKck3?~UAhlQZj8`z#MkRiJx^H{3K_kULw{u}*Fop&8FA7%
zng_zoN0lzZ^biSnR?aYI`q!XlfDTK=po&-tl5OEi*TN961@$E&*AA9abq20fW*I_`
zEhjGRmj{(3sfYtB3M(^9;dRY6;eG`J0VTcKJ^Af>Ra_+y|DK`9rI_^GOMJ0Aixhl<
zx-4tAPPT$q*C_#)`ybMJw*SO{VS(FkwYmy;v5db{l~xTnZc(4K)mH%iW_C|!=fqKV
zK7a_JhG*9TpCG~81~!)hXHJKzdD)pbCDEj!6thv3e!P0Yr>?Hl0Kf&EPv@nSN{$h(
zGONuixy_eQR1U0W37fC3#JAM574}cl;)5lCW0I#>s{2|t#|_dfq~}Gy=5@466%G$+
zI;3LteY5v1oIu6#oiX0(?Ny!mAIq<5$rhf0sz1>Fv6y8ERKF=a@)k*7*M8;I`QExA
zjl&y)XU9IRI+9_a_mX-xsORuwHyF)Df$3&EjTlQixa+592|#|IXWWa$Bt>$1%X>?C
z_x{j8mg?Owv)+%G`QJ}K0{BOF^P?F!5FR9G<DnW%c%+LRH$lHQ`)p++X}I9Y>npt+
zn51uP-s{h5HU=iM)K&n~NT8BOvbVb0o8>UTl814^uRiPuHtF~VvSE=8M(de<F8Ftp
zh6o{$SkTPots#{r!>hl|akJ|8;58g7(JJ}FW&dH0Pevy}TRGlE!Xlm0S)c{L@<QUS
z^ay1_TISvKCukg-D$4iDAf;w5Y`0)K1xj8o(CAG2(Y0lpWw%iN+G*$%7uxVG-hT|R
z>}Syuk&@xtA`rT`m~9|nirNDObn(?LnyMPtnfvt7RSQJuCs8Q`NpXa+5dix2fnsBr
z(_^{a?&bH+kNk%%+qPa#Fcqii-B80Tca2J4fR2>$(>Y{xGYI%o&u%_^D{wm<H{s66
z!IDQ;zL4<FXUpe8yoeZTk8P~<enYn=WD%xp0dPQpbrIOE&9&}1G9bF0T!~y?O<luE
z(TD^&C8Wr%OEcw=vj>ku&aZ0X=Xk=omHr=;nN~pkJ1yJfFi1#{>y2~rNeO`{j=?eu
z!9x@v=Xb~}=F|UXk?H2nn0M_~32>Shl=+MVNwljQdz4Zpi$tD7c;le(5m>5F@1R$^
zxM1>lU)2nZE&NL2|5!Q;wy4^+ZBtSr+(>sxcQZqSgn)o_gOUnLH$zHygT&As!cZcO
zbPORN-Q5jCeT(;d_yug+teJIP=Xo6a&b}9!68PM<10eK4>6e7xP3n_(U))_JG@EMJ
zgk@HxR9Ur{8w01y8bBW1H#Nbc%fA=4t^T`FjvScc0mu3q#D4X2*JtN=(xC1rvF*7}
z{)}3<QG42z?D4y^f8vyYAOym>RdR}sto(nW+@_;CcF@fwuWZjuHSFt^s}?in3!<h1
zMF=mCC1h~EQmrn~i<GR?Hs6BmH*Z`1eP(t6ezi4VEYVuIHbLOp$)NW|)-BvDn0Za-
zmvvsBY2hJ(0wVdrSMuMeSW&t8BOqLq5C-aNI7o^`p=1vd4V>)*yeBL*N4C44|5(dY
zW~ztyFCbgYZ<i&oCO>6Aq~vN=YX+@j7jM&_y(O7WthB79=3NdBH}l^|NHj@|W&Q#J
z8#G&Qg00Y`mVfEIYi^nku<uREg{w2WX8^IY$L|Ulv<tJ+Jivhi?=>=6TR}~HCmmeu
zVyIFtJ&24woxWKG)TB<D4&$&IVjqtPlcbnYabxd}3(!l29l&t16=0;D{h68P8t54U
ze1l1xBrF^ijIc0k)bcxlpsC^H2I+Pv6pCln%D4AL_bVV_yx?2Q{f3M#2l54&Q8#_S
zG!KZhORUtNFZO+RS`qRWbNQJHIvjR@`Of6xRP2z+8~NyuF%+%g%@6svdW>x1&Eh_c
ziV+om2GA4FF=hah0kligP}Xn+9)Qm9UY6KCz5m$54;S9tgfq1c=@Vv(Nmn&|CHEg~
z!>{pC|I0^Ap)-pTkq&tP(FlHnH5brH-ad-xEMI|#;v&=6`=hVpZQ7f#Mss;eFj#H4
z6gycN*Sz9!u%kKP>ZJ5(OI};1`@aNsZL1E*gNQTX)Byv(Dw*E6iUknc9Qcfar_!>1
zhMN4mARBZ4diM)p7F+X!^Rrg{*f^z;ufgR{M;oOZD7@)rkvdbAEjU2eG#D2w+rueW
zuI^FpFbpKa*H>3_p|5t_-I4TvKZ5cBaHqR^vY3LHb>NWW{y+Q0ZpKRBB(ucJ|3PTY
zGc+)z2{RK24x49V1p+`>?u=$rr=eMjR0e^V1WCDujr$jT0*?!|MPRv?mdX%7vKkJY
zYX36DT~peRbKiB`@i|B>sq!MInV3a@DWugt9|e7zh10H!F71!&{uC!eMppp;PX8R^
z6+d3gcA|f~*b3xFfo8UvKv|*272W);95dUARGSm~XE);T*vNBLuXTJvBGaNli|`pn
zZro{D!YnSvQeI;1iqm>^<+cVvzOCPBAvXSYR)5v+1dTZ3(}%VSllJN_!kPl~ePOGs
z6edtnDU<6v{%v*SY!wealRe?l|15dxIp6OGEXCNUTeVQOW{Sf5w!IzzSO%#J*anU+
z0qlH}OI%lgHv4WsOf1I!Uq28wfIVAJE1?P;#fdv@-q{$AQR?pU`rj@*27swNTKu;2
zSrE~=&^ojb8_XV&iEPp3lgnST_EX%xPul{Z2ntYvN28E}&eNS2)-B%_wEoZdySR9{
z4*aIv=~k70L&is;sfSZyJ$x<#aL2J~si-5p1AU#hB2c(@hc<F|g@xC_jMzAQb1c=6
zYHWvknmz95tCvZ%-&WI}sB>0HG4<?6$VPkK>oVCMcQuEFXJ~5{7<+NwSTsri!2!;o
zOSET8plQ3WJ|nsZ5F~0#mo?XnM1%h!`w6yH${<w%j)jAGX$ixCWgG{qO=d=g*QM%o
zz1fARX(+F26LL2P(-OAAR<Tk!lLXPx^cvauUZ%?Y^Hca+HC!MKAClVf&9iQEh=9cB
z0_A}RgR-$kQRgaFr5l%QG;0?n91X=dftcvrJr_DRbGQb`h9*>nOCYf*U{uEO+lng`
ziNe&9>xWND8f#4iyrb-n?W9|x*~wbI!}7hH&-O>*sEJ<(>bf=+P1?=?6w@qao1aCA
zHna*e(APRl@7n2QR&VUQGS$C`Mi(oh<=O;f@j#|P&9mBCX^EJ>`ZMk{vavEht14Q{
zM#8!`iDo{e%DMw%hj}bHEYe*qW0mh4H-k$h05-96g2+7*WTQ@j_JBrtAbC#UaW8WC
z?GMkg@<RenvrjqITzX10yaVHI-St)kyv)uMS05jsU2V7tWFlhz*8?`<3a?w@pIYYk
z>Qi*2XWtyA{8)s;*=<PV?wh{u$av~1EsX7yM@H{z07@J|D4IjZxpL1c1@E1|bDIy(
z3=a0oQ&PWwzkV!s{^(3BlYex-em%T%bF4}e=|m6oraUe9_dhYUQCwB{@kxkHr$lJd
zodPOos02hQOB~9GXH0j?ahpPxFpSl7|4Z7!(UbPoBUeDrr}1`EnOg>v>YE}=Vrxj5
zGOii0l!}qXeU{S~pWWhQD66e8+^Vgx^%V-~r-qt>Ql<>P3JKj;OIXvky)Y<0UElbw
zjD_oKM0~kyJ=?rY6<+1+ruOUH3yZ0?kU)O#4IZ0;0X>r0pNYo72(E$5-lb*#(vnFa
z8)~U~_NS0@mSCU(DPTvVHA!3UuK?~sO($mL<M|@Vl!r5pu?c;4K660^c+eK@&$mQv
z=F8>7W`!wJ;FS#AU>yPE9=@8r#U-QaF+2&W!}LvG0Dlf|F*idwS-IqTOvp^H%|KM!
zcMFcU>-?5Wi6CTQ>_?pW&mjCI?$yGCX1>W1<=YZceE58@B5>Wdt<tVG@vE$tm&bpb
z>hT6~k>1f83_L?Z<;xE*M1X_IGtmvkTr~KaO7$uV0LCVb7%Y4ofA_N$TgvMq%Dk#(
z&7eY3&7f7J|6|6+?_SQAu@Sz@^aB!(`hk#~3ZB<FccgPgr}=blW(T;_o;^Ep^oA~G
ze1A3~Goxnq&F9x#ZH?I4N}@SFmEMVCu-FcLjZw1NX?+!oP?D%yU})2M$XSvXg9fh>
z@KU1{QIc&t4wdd%$fIBnQc8z<^omhV^h>aF%RBXjzrbrEz+AOZ6H>cMwWpa*`NxK$
zG)I)>^b5%4ADhw~OVOdWnEd6Yk&L19>^cLSDbKnyzeo5}c-1}v?PDw<OeLH=Hk)#`
zOe$~{Go?3^YQ~yew7NUFHjfdN7*~o#itrQmt1P^%x|dSr9)1fRyVMYO^W)1fu1)3}
z2#OIx6I1;w%j@S;5lnpQT@X)(_@T2d#ame0n>@I{KVXJ=`fPj$a#wd`5s1=1U~x4w
zSss$pOzng9J$XgfMt`HL7E{gA*8^<Z9YTvMmS<^X>cpwLsf5*P>$L`zxK`)*yozF$
zZ#!J|Tqo0<isUgmb^@y2#iU(wZWnu7!?u4q^`ef$>%C*bMHOexiPcX5xQmgyPJF)B
z6XqL;=bLppOXaVx|J~|Y2`MzE$dN8=mA=0+jnTd0G2}YED$G#Hn@-!g?)ijod@WY9
zQX3-5c%?5;Vyx=(#@v87DRCw|uG#b~zlCg8Ow~zb$ohLg_zFs+o#u@AZq|K-$U^kn
zw%s2=sL|$(=Mq(1*AlY@V!UUiPtR;gq^8$K^%iawK5l<<osMSx6o#u5E#g9nr?3=C
z^&Jb($4Wwv7ss8|`7ctDDT~E5?Qn4=8~-bMBKaAAN{_e*NQnm!uB4O@rQEzOw%uy5
zh;mF|9sdDhYa|2kuF;^!2LeG{2Q7)0I#g^CZd&&_XQd9dX+xj3#`IYz0Q>9`x!R&E
z(%?2j$g0IM@`r%%ie;ao7kV2zApRtCZSl4ZsnA{Xf<mfIPQ_FrRjak!$MefM;~6S{
zQxH%$SPIe!{79WuwSdW0&V#1fQN0!p=R%#;YfNw?iR_X7RB>d!sI54kiY3B&sxe@z
z95d^dqD-e=#aq!K^4V=g>V%dyPZHU;)z=^qWw0@b{2I9+12^5j@3ynRK)$pjb`6!`
zS6y&z_cCBq1M+)4#cEK5&C7jF8UweC`yr1(@cs4PFO{bU3C<Ze;~8i!ecKnTiHFfN
zi6|h#eS)inGeFw8PwnOQXNuk<;#1=elT!g~2q&Auw_4J9;wY-MZ8-hpH_jL0!WCk$
zv9-<nL#LuXZ?Y=qIaoT=F`bOgDu?tkHy4-JA2k=apo$3tyyRdblHnYuBD!++>^DW}
z7cp_R4f7(2U14Eb0WaKx+W)|JN}w~}cp%7Ww51iwH}5KG?NO#_FFeu_3FR7WJ*Q-j
zsX;UWfUP^Kl@plM&xTcvZ1F(aA0q37q^nqYqw~X~VQ$1=SXgeGNOh~~V9>sjU<?-l
zS;^HMf3xfRc2TORt8gYR!=F@8JsyMZP3AV*Jpcq!4K>F<MQga&`zCLt*Irf1x!frO
zTESn-jPlqB_bT+EG2QY4q~wyN8iq~{xEi(L!EgldDZYt|mH3vv+`o5bslU`mSq&W2
z0xEQ9i(eS{o@oT@{t(nKzSj4b>;_XITffz%DoqN9-?}Sgx-5JWAySyRG!$YXM7bJ&
z@dufds*EBj%35F=`;|P-tF+J(YuN0H6wBhVs%8sqtXQn{s0mEkj~gX9#CB_mTQ>}d
zwa2ib8Ie7FQBHaECY|dG%b`o8v*&ky{q1e<MbQ-W0Hh`~6fu-)5ilg>@@7fbcsuQj
zicisji_`CdjUWLK4N#l{ij~(GXZ;S=xCZ{;*Vp;>JJg)OjMlvbVehJ~a*kH`eM>D=
zn<9#IkjAL~+uu6B_XCo@&yNyt4SDVo!3{6FcP|F8KCl0K`PT7W(h!kLLxkdIeJw~e
zy>qQ)`r(IGgNPFnn5C-Ev&=H&zwZ_;GyD-O@m%m@8)T8o`LC_9HQWT2?KKZy0t@a~
zm^Z5Tu`K^WX;88B(_6HpSF<M63rl9}dB@bp`KO8f+|5h(InW4aqeSzTs<V*^CrS}w
zr*Zt!Avc8InDlZPa=K0`u4Vy&bL~_o+PH<4e;=iB3hB%qOw`=?{Hl;v1-<fS@9QoM
z=q+;27HV28i+Oe5VTl+%6n@e@8`+;dDB9!Sb?A^EllM<UHk>}J9RU{;PHgXYC&!=J
zPT-v}m7Crd`zYWn3;kR3mJd3LPt05Q1D=Yfs+YjJaBqKKGrzzt)U4lc+dZ25Fx0~r
zc%eQArK{TzWq9q8R;Mb;IDa38f5k}}F=_m6%YS$MMZUTaU7pGX>hGT^#7whBbguM<
zLP0@|c^$Y_^Z}>@I6HWW<42)ebyXF*i|v}UCw(i?&WW!kmro0o%1JXSdNsqO<lYxX
zO|mSqrv7Xy_QZf7iYFaZB8_j!yz3jNtATs+Q|Zgt^C!N1f)t7kyI4#WaE$j~Z}>Bj
zn?AjxE8wB_(xh5or^0{OH@0jM(cGvh8l;^r^`;IU!%4XO5>m5#j4|**zzm$Laam@=
z{0jfx=dtdU=Dtf3sse&2wr^v2CL7d{&am!mSB>)ez*Np&hb<i{Ly-?cd)XXIwu-Bm
zX(}*{=7i#$5Q*dP^Q7Yq2F=mE80%EuHlW5zZ7&lS&mg-iv+z3fkMIfim>|n)P!iPu
zp(0~YnP>$F9zR2a)zqdgqrUAz&-G)j*4F3&G{Li$ee~~Gn!MO<?NWcK7h^l6AM~Gt
z3V-M#+g>c?1z;-fBy~;dh+WBS(cy~S8C?3f9&Zu-DwRbMSn5q+T@D{xKdh1UY1CP^
zg}Pc{%>formGPdE55(y%Aj0KLp};pZQ0$7IDk!*B>fgc);qS<rOfC}=$Zl(Jn|TL2
z{Ihi(>CU}wTj0GO_90rO|E1T&JJj}f?)kDt-~=S}UZl$j?0n7Da%okNx;!h85SGz0
zsB(5{jb4>wn*}aBc%Fa;NXPQCt+*E4aY5|$3M0i6j+qYpMn{LHeAvw60r!h*maA6I
z)IQ_QiL0M+@i4EKTu?hOG!>txK1({)!ZfM-LV|Zo%dp<74bz$VjG|30IT$5Xu8O7w
z)hd>F8OAa`^I83$0ix)q*w+{uP9u=R*<S_+V79W!Ud=yWV4XT-3C-LBzg@g8k=E{K
zhFDv?n_)qnYd=2YEbiwm$c%WxAHO!P&07|%GWznG=GK>FFSYjzz&jZkd2h<y*FjZv
zQ?)4WMNwt<A~n|d_@d7S`?gSqw)b?!buf0)kn=<HgM$Gm<(oo5zV@5M$6Q8kO80ST
zP#$DpXLL|a0yGhRHsNZFMajlwt?G@fPG@YP*Mo+4-4v(EQ`TJT9k%1riel&`RG!hR
zXxpM;WxUNVrnZ#Iw9_Qw9Tw&*J~2-O!L=jYN|L2pIVr*7_tmx-o3alkGRD$^MrDcu
z(BGAq@pL<!unl%kN7_}XB-<t~Z&hXDn*8p>^XbP^wzE}!snOSrK|l1rpl^OvlC)Gw
z7G}ui#qQw&BNzDlUx&+QVVv(sKvI@UO$Jg?4X&W0-*Nw@qJci2)X=t463_^bGC^8)
zZnUc@uk5bsru4&0^nD}LxO(m0GYQX%vpH2opzvLAI~>QqI8z&!l2g6c-ZuDVwh`F(
zonPRH-|W2si!vI#Hdm~F7bv>%Lg5(W>T|pU^vET9NgOa=2{peQH}NxR(W6I>#L5GM
zO*kL(F#+wI&*jL2e`BFU<D6AGU*}VWGbJB}LcVe+5q|4y0EE3;kzL^{{8rWUiX%9V
zC81sSQ1qtXP0XTeHCJeQdqrBF%dprf+xY|*$7}rnxNax=LR(gZ7>QW<m+2J>XRP5N
z+pSJI66QH{W;C8<bk){kju_OZI93KZDHqO^tOQA`#Of7DD(L_9keqVl$M#CeXX2wx
zRLo}M<SAHrn2i7}mrF0miU)9H{Qi!dnEd@|@MuOy=;2;-CZv~Ov`NtFzz^85nbRHO
z{GR;icxTq|R`zGtspGZ`Md#sH7y5#5thnaw)m1v*%W~7_=ommmd=Ap1J82D!rG12Z
z?8D=U1Lafwkn3<*YGcAy1H9IrfQ|zm^-T-C+QvK(@=GN4<mv}JL|>yl*gDt|S~GXo
z$9Nr)duu6>%k)}Bsu^6W*boP?B){Bcw&Hzd{BPciZOspZ9d28iRAE~F0Rm$&+CqJP
zL-e*LOkevdq8u%U4XIAW$2B8FJ8HZQk@97{jtbb+6Q-a6gdvB5=A$;J(zdTevLb_E
zVXKk4Y33^tMa=~pnRFINjBB3dxo7kim`Ry$4R&wnL+sxdu6_xSnHYGYWIuHQ>KRVI
z#QMkkdq1qN`cy?MC$gH!70=Kad`9!h8mFSqDy{+{Ir&b~vNd-LiBS8Hk-Otsz-$^y
z+g_7;oq|P}#5(^5;6Y<azx9~A$heIEeC^@s(tVf+7<uO0nQlcIBM>FJ@<1oDskN1A
ztmJC5MVoEDYdI5Lekc>b|K)^3CKHHa&2>e+Aa#4uVZQD!U9-7inyY+24kDKB=zsP&
ze*FGUfR#Sa8wRri-bZ)qrBg9BACr`k`gY&g{1~wcdKbGrXCahdi}CH8-#=|us$u<W
zljKU1b+FLH{--DX&|{Y2!41mLL0YTpCbI;Gs$Zu&6J;W@B5|`)H?CxA&=>Jc?l||y
z;p{Qg{B&pkB}nM5bnZubtgU{jC$G?DiyD#5`a=L8Unzy~Woo@88hH({R&XDv@t@Hl
zm-{oo1auwnAp`FNjP%h_0(seg+#uFPW!f&V0q&e<|DhI;SuMB(q4weKEJ!zZc|wvW
zPbFCD*kmi0%Qo>+ImZimbtW<>NCQctUqoA>@4RW`>+lIGwHUZP3YGnUJL`8DJ4A{r
zp^Zr@_AXvP*YKYaa3N7`AJEZiC?Hqci}uLxkG9v;_T&3?j)1-U9weL^$&pv?Ceqtt
z+R)kAIQejMBl53{we`{_ftxp?u?%|45~V(P1oYMMmehcp=@fUUL!<s5=Puz}*{3`s
zHo1I{xU_RcFIoxSRbdL6J?dL&{i5@ts3zl&-^2Vn`hGCNG+UNCKQrAyQR7TA*H>aD
zBH=j(mP(3mKJB2@UK@%)h~Bq-j~l7-Ay#87$M$p2=dw6_7(kD71$<_!bz!UQaiwp{
zpv*7!8i5jAfb{f_LhjACe>%@P?XFy|f|3k;TAG@kL0Vy90BQrISoRl!&&zXGk-9yo
z@b-8oGo+$U3e8&N3-m%uSrJ!;uW$FTu~+rlQu$MIAfF01PcI|3Fs_&USH<3|yW&MI
zFB7c#3S}hO`6*0wp9XSXpI0^KLwCT3hyKpkg>)k{?=3(W47v`BvmZVd0alfbH=bP;
zK~xxooatvaq<32(j~g3r08kA;*Tjnw1LAN2+D2WpHfD<e;HA##lT^C%tqbjd@45S6
z;SpV{c1r8e548wB&*`^oKp+;WtLYD2x_WJ6(CqWz@OS&ah-;ww)R};iI{a5sp0`f!
z@7*YZhML-ItSg)-Rj_5bwtJyw%e>25uDq?$E|j}^Jlj%8CD?q^f9~^)U+Qj|N{mR8
z<CH5iEdi54K3f&3T{R***9G_cjGR{LrF6UHpwQMo_BJgj`y796gvSyISLgEawIhI;
z`H9}^i}}l5D5G<(V$ZNrr`6w9<$J&J&n8u_zibk>Ke(RW3P<|@(r|ZzIJa2ZUj(#G
z&eV=A{c7a1DNdHb@vn>f=BJMHI-T(q36LqR*VnkIns!bQ#zKV~0!+-z%n-teYRat-
zqlhZ7%`jPEg35`>rgneHezVM?SV3ruIn>pzChtq$VrS$YVB<o~SvSw|9xy#yWFU+x
z0Ih5aN^C2mNK|>LnwRSSOOya>OjTMoV6z}@erGpxs9R|%5c|5#(lsSUtB>a@`2@QW
z0}kK17r0532R5Mh<~al7zzTBH36#bO6Cd32k)YuSP>uogig#HSRqlK32pRHQg#FX;
zmE6o{exo9g8(L_KFpME@e)BIR+(X5?G(ir0kr|sc^g&x<f1{r`@Ts)tRlrm4XhZfT
z*Z_0sRgEaQy9~AT$sOx|)Sdk{m6SiS&AZg5{qRoD$oD^V5sx#!f-WCF<jH>2;kRZx
zV4q)MO(kp65+yPf2`cd~kC?UvSVinna5g?q+X0EJ_(@p?4u1tOo4)QcKWdTCXS+0q
zenDe0-ERSeu{I|+qyLe)1X9$OZubzJj=)VyEWCj=0i1w^6|u4Z*XMT*d_lk1*BxVU
zk;G2(-tuivK&1x^i1e)Y#Ri?K|E`Es4Wu=7J>7J^d(aG|V<2Tq?;hFcR#uLsT=78b
z+=Rk+AS6DgNk`YS&TpT|(Ho==OrtD6=?RDXCY(}dN-F`d!!u3R3Ko9cc2~8{(5E{?
zn7rwsuKSh-Icomhc*d*?f{ca&fw#Q4@^q^*72=Y@*Nqtwv{B(pjKRaTte@`sO4$Rz
z&HJRink>~G<zrY9-RjzlM`?(k(hs0i2sMG?Y01Bw{{)4&ZTbh<nwb03$@=GI;V#`K
zLsED*TotM}kjF*#9i@`KcDpdT&|SY}eb@rM(NYZ&3NL}xxS~n{RM_N!QU*sF67CkL
zCpKAsj@TWsK%m%J`Y71&^(QNjKh81-z$+LTX*2udCx|<s(e1mdomJdg$+VWxyz=@(
zFhV@*%M^(g2EZNo0{I5(8~et*iO1=G=9*7=SU`S^WXHObAf0&$erprhj^`RIJ)iuF
zA)V;EEy&c^Ox;ZF(Uvf|BO1gvU*N0Mj)kz4+$a#2={d~<tX(jZHnNAL$9SuwmAu7#
zs_EJvRh7Wv@aEL~?b~*hAw`aSj!_!bY(8}b(Sc1as11*StbTr`XFDtRg7H@K@r9*?
zKca9GTH$jw&s(lRw~8%b@pPG<Ydmcl;2%LfrSQTLBmI@j73fom4cX4PAGjl|d#1=r
zDiHMmHc;h_yK2dgTF=mv^2tePMbRLT`NdkR%in^|tj6p~5!T!BM)PnC0lmDwb0M3G
ztUg=`BTs(M!M4Wf14yPp4}(g~%5De93Yh1OyBb3s;ez@x(ChaX4sJzvu2%kg2_JNP
zk#;K)z)jWZUF5Z7JqtNQpMj-^Z7;im)@){2w=-CTco6v)u$)7J9$^~hgw52y=}#2J
znn=S@?U+JA=ODAm9gUM1&xapf!18!kI8vNeRDZ99<JLHfuE5nLMY!<07q}MHPP)K~
zK`F*fTE+Bt10*t3RN_=;g@vT5SESl>P3umTv4)%Zb+EVXw#=Dr>~s)dQSeJLQ>4bh
ziP0-*c)(k+JXT*2eJLdsibnZ2F4Vb5Y!VR-tFS8?y)#_a?j1|!=b~35);M)Ydx@K6
z`t~tp=m3_lM4jIZB&hylj!k2d(UnPQkRLazfr^Fb?3YJi|CigmYNo;Pifv}+avn&(
z+uzR$kJc9a1W>Eh%r7B>pl3o7KqDBja&~BIjpck9pw0M_@?}-97ZvVXtcil0XX;lK
zy6H=|Hde8WhpN-P<|;{lg&Db*(rHlwn~CkeIYeUz-0Nl4nbX~<m@qe6*YI!mdIeCe
zzs$T6=<CV?1=LY4Tfsnj(+*?%za+TkH-vX}=rBpBhA+n`=(YW(tua2OzoNr~WEz>)
zS_ZK-_g*pDCNKv2p@6|84GIn`SAVI>n3>On{X$|AKZw-JiDnEi?rxsNRsrqD09e+D
z6$b5ZrzNs9x0S>7AvC{nInbbgD=LG8vcJ;r{<}++7nt{t?TRt7+H9y<$-_L_m$$SQ
zT^9+n??b})8B{RV`H|hPh#tPB@qzo$gR$2Q%YDj?|2SNC00(kFOeg)iV5U?I<4QPV
z_JhBmo03j~BjDZ{mn-KhT8wYv{d-*2C|D(6;<Lzmo1Z{u3r0Vi%lZyI>6`VJ$-ya}
z7>m>SYCJ?KHzVGPlDgg<$Fe&hQT|H2{K<%U3x7gILlg`qufVWXX(EC8Fytg^0_vZ`
z%|w{q(P?Tv!HP@kRtAM@^@e&(>IEWg$DI@Z10KM;dbl|n<9l#6Oq`|UP7<$LjKrHD
zsn3{2#SlLtT;wgcAxX~DPQXB^8(!iC2|1r|*@M^gRU<CP9q(R~9=?n%oA&rj)|a|M
z$ueGg*I<7(;knsvDP}FnPu`@m+K}aS2vRZeu`_M}LVg_g`do?s#qqq(L+Jl&=S2cd
zRmWsg4k%@jF7<J;OTdP{^OffZFZAvV<5Wm4^K}toldP1D4t9i53#+$msd5u=ohG7m
zH7hbVn(wjblMsTCEc=yNiKr(9v1?@(^Jb0-IE~h4+>dF9qI<voUhD-(zHIhKH<A#9
zR*b7{#l3GbPWrj`bP8ok2Xb^-rhJ3Wh*zzR`joqHU{ZKW%*j+QSB6iDxv!6p&(A&y
z)X@p&k-g6zkYQl6*hz-0A6SDFyL41`#cHc!u|eL9>Ng;0oeR_V-DB`$w%4xRL~B%=
zB|YzPe59=YfIds&TWlWQ-`GE)+H}u1Gq5>vCvWgKrT6sNPvg^}Yiv7I(Hph>6;rD0
z+c>;3fztV*>Zwkg%_|=2_Tw}2xSh}f?;9@*nqQHTeC!5wS+{$8W%h=U4Q0ezKXkMU
z{orlcrhBcNsNlBV&LYH)s61v7OZSHLRzB>i6&S&HzK9Mi)lY9X{s0H%YR;7%jYQe%
z+|Ta$lBnX!=rE@r^!}1_M<>2_Za8`yq$>Lkm0r$;=TH3yHkV5`6|n5L&PitkCjpo<
zW9Ko-FUB`ZcsA|B0NaalB`c9~OugwgnL-@<-!+zXIeWTRMh0iQv^7(b?S0WO-)#?h
zvF+>~f9BXn9<N+R{MNZNWui>!*R<N}yH302)!t1iMUG>~*+>e$R)-12mAw7grF$no
zTj5>&V;^?EH%eRQI*uKYp7kHwU&^Y^&RZFxn4wvFnb4}?nvS!z8Aav1`>hBZMdx%?
z6EKy~MAG20{%`)PMx@=p&nV&?e6HHUa|bd@nWO+&NDaWB4vRS`@FBnK?bEst#Vmwb
zQ=wG=e2H-WpK!{TyZp%L)y@_4XOnb`TXlm448_`|FD;oe_4|s!k1D(m*$1;Vd_A-l
zDnp2BI5zCV%>_k1wl>vlgE)7ECjJX3o$~GTxpi~xrOCgxUb$7vEDX4>rbnR&-@@}h
z(jA2}@>yPnStUTb)re;Mq0M^kb`$2DVuvoAT)m{aFEP5%^7W~rOK=>U<+qO(V@C3g
zy~`)QmV)5lxB}+e|2%0d6hRUK>Inp31+q6>G#31B;_TvPzTeL%NI{x3JAZs$WrV<#
z$~Mjs)s}oroQSBA$az_^T|1*J`Dz8PRS+@Rzp)#@qS(~ZBBn=Px#RMKf&Nvs^F1p`
z0^>uf>clF)Pj$0RB$@~?gnFJqUqv}}v_LtSLmtM3j5l3Jw$(6Xb{Ej)+g$^htslvC
zTRpXF9RP&Cf@%^waMnVP7hzi9n^Ps9mH)FUqyV+B;h?JXZbiHgkJ}@$WIg|E)NEDD
zbE}+PKUk1z^j>hdd1kr#NQtK~DtD{&I!as<qLV9K>94WHb742-7U{;ymP&my{g#{n
zVp2U5^j}qI2*_}?*<o|c1Ks~<o1f6CzR%IBB%x*Po#PsiVH7hOe9^`zoF3J1gV69?
zb^i9}CS!)3x_+9_aHcj7VKjSb>dBb7JD%3x&pY@gpN$7V)qA;C&&A=N;e4Ceg|UM_
za=8zP41J4(I>~cgJF!FvU~UW|^Zq%sT}r3=^j!{D)u*H*D7MsvJ45dmfg?pn&F2x6
zOPs8<RNj+;wm95qjDP^&Nm82y3z8=Du0u?Bz0zZx^YX8p!q0gqfUV>M1m7i1e*Bhp
z@k`m3@=RS^^KkMZ#2QIKe-q?YwV&N!6M->mx+fmMetUQ!8g{-s8*=>JETfitX!DGg
zxr?Fpe&qapW=g5Am&Tuu5$7)XY)oF$j+Sq!7`0<)_w}sC$~6`>jf;cM>veEdidK^G
z(@?@l_n66#n^&`KLjDn2CDL7P@Q?Cc8ecbk!u&4`0~kBDYV4A~!tMUc$D@yuKP{6*
zf<i{!Nq5YGm+n{u^TqP;z{+tQ%Yn>lE+voiQR`FS>XFWE3<u_p18I2Q4C@(JaXpt}
z+S!DFO>dQ*o#)KW!9j&-yUrgavk$7ehmj;n$v7twI)TcUsRL<tvD_PL;%p0KTENT2
ziDj=Ej>4}HH`U1!70cgOVX%HxquXlo2k$f)GO|~gJthP*Zs7Jl%yIbtIA;-<3cN+l
z2u$lA_k36Hb6<v#0B5r2^*Hc{aZ6E=dL;*0t$p=5Mf_q`WH;Yr$8)n!otU3lt<^kh
z%BgS&Jynm=XwTVF0l-brcK+3Tky(UIX>s-Q?3^RNBLEhp<R}9Bh(RBcBm348F1lJd
zo9iydZn32I8B21!L^L`=9;=6tb}}#2zMLV!<;Ug?C_hfx&AfN4Myem6#F(WNwtc=1
zj!LG)$x7La^3X8=i)_bRseD8F8-iLNA$(BIu45V2I>y_0WV^^9am!Jqz2G3ql004_
z6^yF0*9EXN+tCV_o4gr5dB}JAJho)euX(;oPYeO9*N=as1~6v4Z!4^%6W&1|BAJY5
z!t-3-I*})BPJIy+`5C$jiHXi$rxO80LcVPqbP~V2OCo{{rWe_op}&}u0&p&F-h>~5
zq~LQ-&HBt=iO=;h2Z2qdnSPR#<~0`KGxcJ>>Ebs~nJw4{=AMAJjJW;b(0+YyZ*N76
zY+Rv_4Gc*S@xv{Hh}}jVWFxva$+o_TJn=70Zh?QAYr&;IY?<=^fOR#nkZ{3Gnq2Z4
ziVziB1M%6r*+k;L04V;s2jm&qUc*ocuJ!&LS6s2JeHhIv$-aP|#7Zwo5bVTst*JDm
zZz(@+O>sYZuo30ESZ;$vlqH9c+U6O<uL>4e{*s()O+;t2ZTa?*k7Q{(*D=>&@fb4d
zaRM?=4rVI%GPu6|QO?ZKU{f&uxw5pD+Ei%>DKdz$q0&om3pc5>yrmi&FEMqt6$RS>
zSMyGcZ{9k0^o}i|a2t|^a6Wi(*PHoz%TdJR)9>_5tBzo?j5Q;prGAXPxV@+_iRO+H
zM(LshS>oLIpo{!cZM;-LSLhk$)!jqK(VlPsv1EdgfCEYJ4V+~USd1{s%++`I2WAtE
zeORk@`daq0EqiDq#fynt%i0Q+W#}9G;#T<suyubXnF5jM9y7qhI{pjCB|>=CDS5KM
zD8AxePBkX!<g8GM>ptB~P5P8*C1-pzGed!5;lGBt8P{Bq5Ci_^exKN%QT4l*H_l{F
zwdtsS*ClFyKK6PA%k!fu|E-u^0a$rJBo#0z9035NVgQMM;kJq65_|4{4dPz-$i7jz
zO~iO=Q4nQ6n`TjW=jxuI;Wf?Kx^j@#PCyB(9PZR{iBKae$*#2{a7;TH7^}i&AokUZ
z@k>EYHrHBaHEfHHqfHcWj5adM^g2#oL%{8>m1wX`isYG#lYnl?xD42G>*G$NO0Tw>
zcDu($&O+pz17sBG;~&^C%Wh0CzFhsg`F=(JUZq<ByTl=t2$P!_%4oX2V2p(i)MDJS
z_gWi6`==7DfQ0X@x#Df<%|3V~5$&;B%EU5OyB0GEaS2EV#8QSV@2j6K+tJ6t#R2O{
zw`wNkb|Hfs&Xm1JtH_*W>n$SV>5kuA>76^+X6Mo*g{Ot_I-@U{KZ)4p$>!E%@+A)V
z$6j~7U|83+;^;-nw0arA(eE)vsyU#lruO_(w)}Lg%rr_eksTcNPOUeH7|G(HYOL}s
z;ZBj6hzC{W(HoF+PI-1eYA=&b&m41R;eW-EEk8Lx9dv9U%qJ^^-|R{i$Yusuo9ACy
z{IMkbRj~#?U16mmF>&`y3exDRvHFRNcxNbx@2~`BS+$?wWO6w3aKJcBk9HQfpCyP1
zxbEpJoE*KwR82P$0b=%MF&9Jfib5M5;&XDpS$@=7qC(n%{QDcl*J-sTkQ7ey$CuXk
zD6|oI*c&&mGlDn=iSp6bYR&~C&q^WkCm!#`;2j}tl7MFg5DK*S+FeoB6mmKJ`fk2{
zc5r1qtoG2etexB2hZM97fAWPWc~`DQK*fjlEi4*Fv&mvB@ZR24#6}#`EvbmyLa^<P
zJ!-f0gz=gE5<7H9RRYU?GBw`)G$ylzK7e^jq>Pyk^d)Z=k`AlP3mIeAUX4TC+7<Bn
zo!!3!3;N+~wq`!+emc)Zgvlo8CpAbo&>m7HRVQizP6*qc#0EuLaJ@=Jxol}T*flvW
zkDzd$koyJqkEIL?O)EYTos><*a~E<d%4CIGb=Ny!%p9mqPm}8Dh|U)De2uzRI~#IL
zq13*TyLRyT#oiqtz`#S6cFKst8NKH&FV;fiOC5f$0*{mrG?Hwa>+5ul#4ngV0M%sr
zoHtO>mZ&{C339!JK}-in4kQjeZ22?*WCAxj?tw)^+osJh5y#yqJ@R_`T}2}b)|7oR
z|Cuc((kgoUCC<y`)i}_9+wl~yo<^jDCztp9(6bkiCdM^aEEXX9S%j*Pd;(Y5%*s*_
zjVol(e(8zLQ#i4k_~uJ;nVLx%f?amQ_6HYxi!Nf@)zM0KpqU1FwCfK)BZm`{9N9Ds
zg=}H1a2c!Og0itwJ16P`mfFoHj~@c3dKSVtDhIABWfD5;G7#IthJM6)`v*xmyvst%
zVz%b8yVAz0b=-v>X7?Q@J!+zlNul_Hfihrq6l`d~WNx!j6D1h0NrfLOc^tL!8vky<
z*|u(90Yo^A@n2>0(i#3nxv1Sl-yI}e%83iiPqW83U$`M5J1~@<`DIAxp*)rMPGpx7
zqkWn$ZFYOFjoANWpVM6J2a7<9B*k|ne5zj{J~+}iVGlflomM)4z4e@qB7>J_osmLW
zcRFnKrG-)a?n$P-Y9pMMom@){Cwqz$eBY>q#owk=$Ty5|mZ-nTFUyeE8C4GvG;BB{
zi+Z$%u@9DA_uw79hLZW}gbBY){~`vgwQNFg*W$-#2E4xn#gq^cCEzA(4F_PV@F2P|
zF*kMUYQJHWX{Xb3e?v9;1}nszNiQw5g!6gFkHazZC;PTZ@y&jJgVn8fxpF#3WqwYy
zUZUAlmlIRKz)WBb@{K4}`v?3S+`Uzz{{E+X5a_^5eYXUrD+<O#4W&>N4vEpc#Ka$8
zK!J#Az$iMW_p#0$h*UoxPw?^EpB|3yyi1lfFFsCyaNlbCzW?KKGq`U<B#djpW{t}U
zhIM`gr5?DOFr&gU!z)a=kzYxiJ6bYO5&37#wJ~U1h6h&N6(s3fHT%JPGh*O=@UDnt
z9}BS(C3(ZES!-zyO-fWt4(FwVjLC&P1#}hdvX-=AP+TzptLrL!PQv&Nq<Iu{xmNQr
zp#`a!SqQaqV`!}(fw|51I^JO{jd`B*ya%RznH(}j4I)e5(qLHXOf-LyQv<$yTD?E#
zBg%{-I(05{c4TH+dVLapjC<3sV7jFdgWPD3E%LMS;3or9Kih|R(5t*u4C_|khi1vh
z09q%Y<~c8p6+>8=254ryz2~`ufdhtK^-SyiYc1<{-;+KB>LFAY;U(mKgZmrT9N?9q
z_F+axvTbJo$xC6RIS)RF*#4wtIic}Fv`dgyXl}5dlK)=wh(ejvCwwy7pmug%t0$T?
z`(B$!w>gaRooe_Shmpt_HvUg!WAtopeSg`saF<awA^ly6nw#D8l&HR%M^ndB;wE_3
zyAp<jnVwtU^_Kq)HwrEt?RYLyAz5o?{`wWw*xQ;Hd&vKzLx}&;YKS<G{de_#vviT)
z%kMWFF+xhX<Iw7q-)|Y9GY@2>Stu*|_S$?r!c*w=9d)AQ_Uwkt)q6Dgp;B>>lg3`F
zd!m0@f0Sn+sT6*{fh;0HgR4awZr=Laa2K0jk$o9Mo&3_7UUDZ%S=oD!gMIOC>JZo4
zTw@2x{hUCA>aSrY|A-ZfjxJ|~vNL@UE)h8y|09<LS?$_&m@Km*xTsD)IaQ?G4!x7F
zkj4ZGj7G4Rj?0c^mVI&T<eY^xqeTe4$UR6bj`8l+dKxFS8Q0aUN_vkN4mBxH2<Jr1
zKM$C>si9k?k(P*WkRc7=d?>;@va78hCJu8!g-eLOF9>zLv-edSsyTBX@3{X_hI`6F
zj|))BGeLiOTd#brXe+0<CAP!;^}RU_E<9CKTiaEIrMW$~hja?2n2B($THb?fL4w<%
zkItUbGb{?3@pr3^5q33j;E?|8nY`vMTZ_S76wVNbqIU84&bVT{D>wQ^i&vktdiB9I
zzAsr~2e1pyo|y@?;i~9K24vhE9OA40Y%7c6(pE`#jUpl$O_9wS!Ckgr(x^>E2}>P2
zM!zUz%=mEYu4{E9S$x8STxsb6T*Q?Wio9>2?Jq5zUyaQ0vWCSZ%vSN0F8*vfvxJ;T
z#8jg4hjXtB-lo;Rf5RQb+u-Up+%IZ&Qm+K2sQQ(_w3rwW0!7K}L&$Zff7Q(V(S7>b
ztJYU*h(xu0sDP$(zcwGC728gbyP3xR6KqMzVmiwqNRK%DG0{DEsz&ARZs1~r`=ex=
zd+GymHuxPy+U@JJALE0As-^0<Ezwo`Ykf3+nH%<s%y{j<=E>2^3+%z05@Vc7b+PfT
z_<f7jqkqsEhtux;DclJA^buxyZxYRloWN~gH)!Kj06q$}iF-z8|2uTsjMV)*dms!o
zmfOgMw!ihnvWY`#wJ<b&Nh?C7^~Tq|XR5<1OW$;X*D6p8QiT)JxPHHKi+i+UIw)zM
zDAs?3zi9M&5K*Xc84ni$TN|oEC$}!O_+N;3U(2%0#Z4uTB$V$zUrtlsl<{u1W}gD5
zpV6&;H**!&G|i|F)67r9=5^Zkz04tbw^_CN$h=dU0AqAuu*O;}tmG@j*GdzFi@SSj
zn^CEDCj6+$1FvuSbME?(@AWZ}o|{R6Usbi{gRJY~y!~d?PfXGs%fJW?;F`=zzsG2_
z7c=MlU%4Cq$x26`B|yl!LP{`~#yE!S5D-F^k3+Ys;)uCAqjqfxL=<<3BYRf-nVI*F
zj;aJ$$UY&IlT<#Me(Ep!d-r4&PnS7oRkXs<cGc7Mh(It(TJsAioNHlXJN`LgSsT>X
zDw&Z~<Ha!t%)*s-F2?iMMj`C8gNPsE+r{_0CB~M~jUAn7D&F2u&6~{84ZTX8Al>)B
zoju+;qG~g0#16<zyKTKutX@9GClGE*+9meCS0;*aoe>XSCzCFxPt_MsdZDI1nrVVw
z)yidjUBk@MZ)2`1%_NLVZs5kS4pabhmE8Ub9}~cT#HmfjV9C=iPOIjo9^%_s)K+-E
zSfHwFinf&Go2_I@B07xn(FH)WZ+-%8($5~sPY_8y(KOeB%^k@;9_?ZK#d#+?$sVaX
z%Ba2VgKp%M<?iiPTwxY-6#P>k)m}Rx<wOguP-e)jJ6TZ$gf&-mB{fOs3AS9CWPYi&
z&ad;NtV~>_js7cU|0=F7Rp&QnKJ@b4q3-YRB5i|0*0{ZDFBrc8&Mfmb$3zIpNcO<d
zyfD>Imqq84wmmY#kO9gv(HB?g7oLuY>$`^)TKa15ia0L$RZ>B>C0*4J#Apk~o@8G~
zKvfWLyD%!YrvEr0b=2badhwzHur4t5bx7Z)vq5l6ZFtggY~){aD@QT;0R{?XS`~5E
zoFwF0j#WX_H)OH3eB_RCwUCF)8Q{f20TkBqV?I0?hD&j17sC*)?MIVye!(4VRu(W^
zU$<UWM@n=5LZ>fPK9$WZe-C6+=(0|u!sSPqeE|$}>+4_Au1JZX?J&huy<U%m#qNeo
zHSNWX?j&zNr2!}AvId~s22imt<I!7xIoZvqYhx;wiD9L_Y!7`R8cZUX%mt~SoHxME
zK(yv6a5l&)KM_&^rKnnx>;|7u&*%L{I9a=rB*WSD$PFx=VZt8K8llrPdBP)NDtLJR
zsnkbW+4SKSThaI1=V<+*ss!Q(EFK>F-hDUP=`-YXu`=KG8E5-iH-6e7Yd)ynrb<n=
zj;j8jm!pnNO--aD;toAkT~tKtt~NN!3ICk1mqPFWJu4r13%9we_N65wzt5M{h~e~>
zm*myFqSG_3{Obd@xU?;jvoC{p3iKLhpwY(icz#3vhMOC`*&|iww<(`ur%cDY?+mbV
zV6&P+jV3rk4VJv@`e;%}%hu#%mA)fTfXxQ!{nb$awQ>H@KnJds&#mtlo_KP(I_z7Q
zB6<Hi2p!{Zx^sm^ajh$-bp|l@(E=y#rccjk*LK(Pd5MM6O%~eaX)s=n-+DdF4OR%c
z1xzX}`+XTWQ14YZJq9EiV;G$EO^GwGwv4>}>{zv#*U&2ViT*H}zA*okzH8A#F_ods
zgcYcl4yS7zjXBcn(3_R9L0Ze=UZqW*-TNZyhKq@-V!mJ;$UDC$jlknB?Qxc+U-+vq
zqT1{;;}z*(x@MqzW@d$Lh%AaB0qq_l{#Tyq<Eo~ucBL{Ta@z`2Eimk<GUrkn2aZ)k
zng5hptSnnIPFw?v8o&?m9FMr0g=FqFnSR3Pt+Eb>=#ynomDY7=5_3BI2PQ&N*_7_>
zOO+5Ugn+mP!{#RUOAHgcYT71=Et(}=g(s5|#@A-}?*5fU(lUQ)`fc9s=aqGxA#A?}
zq^f`_C_?C83MIlv;@N@-D38E#hTrxHrT+6<h}Xr6`r8<RVUEzA{W{*M8?C!3XPu>*
z?s@v-``_T?ORVtSN(*xsaDid{{O8yY>uY6OG&1^92A%6w;~ah3xG^#@l#Vs(fcZy_
zI97ouY&u1#xJOoBB80qg@M8m-{xVF+z;aOB9NtzCL)>IDGGP!CHF#{E5y0!G3I)AH
zP<Ci8R@VR+zW*r`fD}rk6(M6jFU5Jqvbn%}-Y|1R>%IsI=Z3r2o~JbzzFKFg<3};M
zWt#23MoMg)MeBjd@HQ<^E+2*LaZr#>JCIvLvWDpckm(DRU?|^*6UNQqnH~3~Ar^v2
ztNT1}z$Us3s9k#ZXR0wC3WF@$f}9AN4j7EZB_)rY*%XV<vPtE(0{q)7C%z8b35klZ
zrZjVqQ`I@({M)WM^%x(B=jJ5b!C_>(Wm{uCEfs{(Cf2(~D|J*{RzS!D2mjL^&CxdS
zFO-=eQ&TjoZ`%d#dF>a8NbiEq>GNm&oPEJ~<!zSO8OGH%u1hx$nsbQE39l*?Ng}-;
z8YW8RLqd(Jx7oL>zvRfojrNy-Z$uK^(6HP3@x$$cqxve@>&i7g&D~2<Oh-sK>CwQG
z5zm$8jvJ^v#iDm^@}F1JVwSp<WS`48^aEt}Y~)$!3}D)2fih={o~c&*qxG4Pl{*aU
zhkfE3+HnXl+_VhLy;mcaj^i3Q1q`CNl&~jBAdr*#0=7uubW-;O=DP#(RuCm*S0|<1
z$C%BTw=E|d{cN|z<ORIUIRRq^)iawhmP((=LRllkdeW>*z|Ijwi@56lnD|{_clX}J
zG3WY=gxv}zlke1d`OLQJ-sP}fkl9Q+XS@6*-^O0BxPH(o)~+|MzXZw0ISE)@JV=~H
z%rDh`4d>47oBt%#HfelW%qCEwU5e;m?whkag}tC;o9m?%JI<i@P98p)EnT6tt*-r@
z{?*ZYw;>Q$se=J_SO@ZnZWNsXu-iiLJF<Pl9)h-#vKN$)+D&t%W<spku$vKkUjpE#
zo%KZhuFO<40h9bs?6a}GBAa{O%fmxa&FhlnVfQ{2coI8-Z#>YQaBHH77~fU$y$~7+
zbBobYlRcklmnyaT4x6p;tN;8*#Gq<$oAf}(GrZfeZYMB1*QdSxi>CbiD`jmqr1i_m
z0P6_WT<5<FPEJcq?(0!xUT0&=R{!>FH+g!e1^f0!Z0wSUSG3G|gnTvUlUE(BC5XaQ
zSd$3I`RwzZ4)pJm#bO5tEbBMx#PRY`(y~Q)i7ZMUi2Jj;+y_2S&(wf>ZK!O86Lm50
z%o8Ofim8MX2#~*X<Z29=nqQ6WmL23rsFqfqeeCt8p5GEYWHYGLd3F!h;`&azXSaC+
zvr~Y*Y1(yC&==Kfr}9_WFP!LH(}i2Dv%(JV(vX`c(!-?ebMo^}pUm21^A8_tC#sbP
z&G^UiCa1(~r4M`tCK@6X&g4D~Nx|r@)fA?FI+wT9pDq)hvS=#cjXIw4-!8?fE%|eB
z;(p>zCkN|+9@mmZE9^L{wb?>-7DWq3YZ_WaOmCi?ZBCYJOS=9AaB3<s_Sr4-4;Z?F
z7N>a6%v(Nq>$g%K52t9sb><1ng)WVo2W+QwhQH|3O?p-@e=dnoGwD4l46-&v441@N
zZ|+M->{l;bUHOoG7h83%WC<go{#?1H>`GQL>U(|aLgDS~OaL(6Fa$jOF4JQST)2Q@
zqBk#T=lXsRQTj=NJN);DK*3E<0E>eme|UZOIE6avF+g+k=KY~+4I}gg<-*TKN|&ol
zB_Y_!C54}acU7=T>tiX|Kl2`pKN#c51J+_ylzaT9%7&oR6z5^9mxLPA-D3<;cdvnE
zoCOb2grrp_KZYUb)3mr{W?#e(@WoSw{U}a~vQJ-y%EV@8q0kmABncZ^8*_7#Ri;x<
z{A;hPlagV=D~O)qXWn<UjOm{rMBHoeFpKJq6d1x)In4#M352$7EoDpLXtLk<=|e*4
zPD?IW$0g<fO2=W{;J;DWmwW#IhFhqMJ3Z6o%KB4rGTZ!~e9c}yR1N^M;q3yEO7ch*
z$reAamn?p8k3-SpzmLkjX0ILk8*V)HSTt>{WrA-?{&04vL!iLC8Q+F?IJjqKlJDcB
zd)2LTR$R+TWoh=R5ispNtoepzTVz+&pFBB~=~tAwA$@6RTXFHlkZ!NSPc=xwI)n=~
zxgh@sXIq^`$KX@!>EvBOmxQI6UDRU@@375^X?g6dUPF~Nx_d-aJ&tt)$JU|NO6RQ!
zsG&wf?h!bq?o}g2`Bi&l5J7)7!CVaAZx=opd0Cid71x0{%@V#oxD*eZnrq2^eWrnt
zITsyg767D`SE@8q$7I!+)#(*rDIllC?U_8Qj2PQ%ed9Il99@ga_2iBJb#@D7>6|1|
z#N)>Fl~zn7)n_`d*I8tQkB_M==>%l?!kW`MewZvq$ol`6-UxD^gjGc~lJS(F8SQHW
zV&{~hP1*o?`CJXR;+bme2ouH+I)*+*_?S+$GY&9#y2>7P+iZ=~KMne5*<SOf1F?RK
zp4(}Z7pedo8>t|v8H_>Cp3G0hP8~jI(T)5QyIquqPpa$Apr_9}s;%_l@{SL3=4Mb=
zw__iW!HY0xIAeUs968%m0vk$ba>~Wwf;%_>XS>*sL2sX5U6FhL0K?chAs<G`u@Sk$
z?*p;ZcYl&<9QCiad)lXo`*7k0LQ31&$Q(c*P&lw@qyEpKQQL~S^*<H9Qkb@U0BRO?
zxETs?Y5{e9sn>M<7Ex;zGIu=i+RflXxA!Z~hR9hZP8O>FFqUPYqWOqWrU?@cR;I;;
z*R;#j^{|5nM8)k=n5}utnr!r$OCzx>V%6}@KaR+Kx>Dm%DP6-L<nLwAF%2qyOD}Kf
z^lACT8vawGlriJ}lH{UnPJfispwFK%^%wvsa@HkOEHa@)(n>wz0GlLv@J(h--{D3E
zS1tR(38mJrej$!8bVc$0;YBlM<;6O_^l8`_@m%FM12<p`#G%Ko>8#WPLzcjX8|IT-
z2`je*L*W70#dWdIj9X=+!2MPtmoJfoD?a=xY{%W>=3O=JXf6F6Ur}sgfNYVAb{`CE
zFw=|nWw=8%moATL!j%v)(M*+7hxB6JS#~I0$Qp^PB!{-CGn}-o(g`+8`R1<{3aGvv
zO<1OHfw%|Z8k^M1pz|%NKeeI`rx=j|0&ly&NRnu@kZ*lKeSclibeggDrg*FrVL@1~
zvR6iR_IG!x6%E{$q7wK2I6BL)sJ<?WQ-XmYDxFG6OZNcMjdX(&1Jd0kf=K7UP=ho>
zNjJzJX_%qAq+w_nI^OI1jW0aU%$<AA*=Oy&eoN7yBr{~3zNy4IlIw32xI{0w9e%;+
znTur!Jw4^|7^-n834GN!I5%M5I(U?Zy^dN0d}E!}4RJ+5mD|kkzs%~+<=`Jxi_6LA
zGoQYEM_X;1uF$eq<=7xjE@4nnQ9<BJlu0P^pC1r!3^Z}>0H}q*pRRiMm{+n~S^?MZ
zocl6uQGx{NY%od+=4z~WTDA<&XAiY1<A|3eWM^x(77Ye4vNJRDf25Fq*bvWNWS`i1
z%2!ay5!&Tf{0(}##oH|RKIs@4RGKWl>TJEfHpatAz-Jgx9*r3VmPQ4hr#{iJ7hMZO
zeJBN>nuC|Tl;W#LR-y*f9H3&(AmJChUF1sWd%ZgcVWP3bennY3UBa99xGyr7AiOzh
zwc8Yw7fN2B-f-c^jA^fFn{cF7xQf<&A^o*CKXgq;(@a)8IF_GeHkTi{l&LinoA+gw
z&A$L`q4k0l>KAX)u0-*;wbpenwjFxc)vcnkIxLi@OqpO*25+(<A6Em)m%n8$<kaI&
zJa9qm%asbO7FfJzOPvf@KV<^+Y`af869JnNk0e!G^r0_A;0TPUX*PP9*N8<L9F>Z`
zX2DXu4jtF|>m}CQ5JaVj-v^^IAjr_ZVg9voYjSw1$Ya5|32vB`?+3@t2`Wj;aL9BI
z&!nw=?{6BZknx^l4HrqHW9zbbf$6NF1UHkm>V(x8=JBR-hmcp(KOoX4fu~Em<?vR$
zlhr5l2-5SPt1g=|ZI%{21$rDEnz|De1PaY9Ep{SiTWnrYznl&-SWSiZlkrXzI!?WJ
zbUWBgo3n*Ew36DNye*o!k9bl!>8P_x-pugLDq0cxOO-o5<ukBpc7~A5itCtgv1AYN
z;(GVs>!14#a=+9W#JKd#+?}nmbHp*UbJa2PPJh#z8io;4^GHviejf$52)I_p)Hokb
z72>&?<iWUbjtV69!ZAp+PD#>=>7IawQc90zhP}U|F<o<nJQebW6_F_%!WxnoZ5o(Z
z*bgl4n&Xt^)PW%(h~ZuNp4?!Gi0b`Nn)Es7Ce4Pb?o7XKIlTbP8z|f84Uv?qoc_6r
zwyy7vV?`dg7;T|75&mKZq;8*z*UCf(v0X;ej)QA|NvzxU2-3b<TuxkLxrc1X_9sA{
z$472hbhPm6@}^3{m&gwkq6kVS&u5&pVz_pkmn^^bf)^EbXw1Gl)wRv<;})gtl%29*
zjearvkS%;v_7DiOexfD$+KsXLL!jm3_TBFddYE*vSK9IUpaGkahzPavAFOuJJFEMz
zZ(k!0tIiEzySBtx=S3hpl9GDkxYEAo8{fh#G|9#B$V|K%+k};c@+`*O%EotWgh%#h
z1n;@Hi=m5@HQW|!gH(kO6>GH0Un1Kf$s{f#?;FPsHVSvHW%c9@ls9a~i^&Idf$#2N
zLJADhV_#)yihb$27ml$va?QTNs+?z5$d^bm3nC{*AuHtZk+VlPMRK2dm6Gy!zHodR
zjILWw6_!|R&}aA6p4W_4F|WC9t`Jqg%KzOQJTB!IK>iYJv>romxVROb!gs|>j#Xgs
zMtpW?BFyxmj!{c(A<Wo8G{!(-YzN(b4L&TIP8c0(*A5;5u7;9dLc({wI#O@7OU@;t
zwyuL*U0<TmB0vrx0nl3bh~P^yAlcF>sJza?(7JQRTUiM!LMN!va!vZVy1LTQPEf|?
zjdR^wUnlejPJz8d04(YH6N&|*W48?kllDRk^&p6BFsGXW`Sbh5t}_F(p6$Ui4!>-R
zf`@^cljZj7-O|e|WY7|A)B=T8Z0O+dYmnY~QzG`)*Kvm4Mmn4dxHgY>$tUw3`maX~
z9go0$cB&eWb<eWgb^HHXR3EP<(f6BsTms45cQ?Ien(pnvJoiRe@mYi`;nf?G6H`DO
zZHY?xm@B(M3jrsOwt7h3=!poGN3~-hGB{(p_Yc!kZxT~<yczD&`brO^j>qtpH-_hV
z<TLO3^^H#`d@7&*_1+uku{METH*$CD4&cJvp&T3L9fIz~rOK`|7}e@+oca&av{uE}
zi@0<m>g~&gcs;3j2Ov%^>j6a4<PTH^uI)N>@KvWx_kbSL9Vy+yna2-{Q4kv5_tGM4
zMM-ts8=ww73x)Savb;6mOQUHIAV-P_h;#;8OL?f~`{BOC8o<ZlQreh(`AyElS4~%!
z(4)3n8HA)JOHtFqh(5nA2n#sLB8TqxVJMWgJ{Gd1K5VWCa(NO&|KW5tx`J2POKF~q
zn8x!S2%4tL(VDj-w?RrG@!6$9rvIunJ$>Z*OB|}=a72Ondrfg@wb*<hr{k{oAL^%7
zC9+ea@U>dUWxBA3JOV1kNZ7g<jd$?+*0Au%)%V{=1fn7VX;p&WRF40dqMTUh_nVs=
zcap9>$2dOWEb`@ED(1uf19})6wj4H$L+j`iupR(qi;%c@3}C(PY;+ba>={b6Yt_ch
z5a9{KoAqv>dUFM->krfpJE;ZSLcp*3?_mI|1@DfrIPdOA0FLDBf~v$3dlf@luf4rj
zd3tq}eE^Wy@#W=={1U)Nb%K?i7m=ho90qv?d8)eW0l_^TlLR{n`vAgsOJ%e5F6`5i
zAZROCvvqK<WPl8#Y#@p-L+?9Tq@|{*4$oGLX(jzz+^wvJxhze?4-nxWx_3~b;iuA5
zlK~a(FLQL;Qff}i%%0N2D7A#b*KU2R-|O9-+|Ak>@mY<Au!KjQtisy7O^la51sIfU
z_t{ud124?~vmCp*0*QdJ=79UJxzC!-nwZ6cmF?;#tD|1BX7xnVUaJ|n{whTh6INft
z6<*ETa0qH}wp2uokB<YoG9J%x!Ao_z(jKca?Uh{CS$(jwmjznMDU$>*$7H&@zhCjT
zOX5}<Kw)**Ol_?xiCkX1+!84$b}*YO`~4%(%@m0LMB4(BuE1JwyXaEB(Fyrob(o)W
z9(qwlJDVu2Q&>B^+*bp)q#N(Jv3y%fpCJR7jMOr)MOu3L0Hrg95HR9IM4T+{a=flY
z0R|S;k<oOMU)fc5P`YsiD*|8q(Vc4`Lv`KeMZrMh$<omK`WIk<ZQ`^Me)$6O|ArGi
z7U-%ZpvRQbb6=ubSMYlX0TQH$n?JM)8SPmxvt-I7fiOB&H)5Ygc>u5(K<0|v_oM_~
zF90@)3+{BwN)P^-8@k_+$qY;C8lY-=y77K~@#hV%58BT#e8d*j?06K2AwoF%w|Qn)
z)Yir1q|cDQefDsvgT%%?<8P~T*S>=-{=eL6;5`DS!cy!1dMYV{IOeK(h%dE~Rn+J6
zJ~<|@9KFuw?3Q^-uf=v;zL1}Cr${^Wr*&}sm-;@fbssFD&9iwWQriW6?nVAGj`cx0
z3%$kDku;fi{5C6bu#7L2)li|EFvHt+2lzB^>Afbnh;q5Zj+UXapu$_<cDUYMA3g_4
ztHlC+3^F<02!DE{5S$jFb=*z;t+KM^gOFsOjmx6}{AXFjUGc;Za5oMf4Oxp6mIzTJ
z2;|oC_xZob0FeSu)}vo}_6wImDvrF^sUhq!0gou)pwlb%=48y*mArZ5H@$-G1Gc*6
zvksC|pk8P|1X(ubs#xE8=SJ0P!ua`1yi5RvMhj#Ma%GWqnm=0g&x9w=MzWmg)tfo^
znt(Wob1d41{CsRnu1!KT%+C9=E7#W<yR+&o?VF6{{+!Xb`5?*Y&IM}z>(7;ppWiR3
zJ~pVFM2M_x6h!Jg*>P>6qWW3VY^0V0kL#)W;-sRmAbeG`;HY=pcZzHL!no0=EnUjb
zyZj~HSU!FbUPyjt!#k^|J21A5Q$HI^F~tcf1)2O-k2yoTwWWvMY!z<B2K=lJe>pkR
zmZvs+XVxG@nc;8$XuwZ;66Y>MbztEIdS!E+_A*BY;g4U-+Z>;QFQ#G2Gm=iOhpt<t
z?G0^RJtTho*1_Vg{PJ8_fP0YKx!K*RG_XM!CQUd$OqIe6)Mm$^Jio|As{m-g8HihR
zr8<#Xq;P9i?pb^j$nwnl&i7}w`#Alh?B!%eh)9Q`AboSiJXUUgX}VbRg4IU1FabLU
zwmAA~=;){!@_{GxJe3{#vu_}6e4U@kujxO%cMA#dXP@Pi|2Pj8Ji`x5`NYe}Y^C<j
zsKpcfk&#oC92OlkD^vs0a~{?OSnZE8Ss!gVaafG9u3rHD2(<5FgNwgE6(INlkUN5G
zHxX(K9XbBB?~%yNnkoyUFI@9q`air_iIo<nnL8h1q)5dp>b3nFaI0<#4n|0+>e>uZ
z^J#Zm*_Bzk-;jY}MHA2%b>#|RnBkii)p=S>Wp=8{wuCYXaoM&fjdib_-r7O^T$~<l
zUWb*_tfwxmZSLY^ol>}whn_pQ);b_h;>7Al#HWfS{j0KwuH)llDJ;XV1;cJU2vTRe
z7T(Zpxp)Zv5ts6%KK`fY({DjYYvYPbgF77ZYqh-OkP^sW_lf3N`)f4V3nx2{PNev3
z^74j9!@Aag58c9SRsUZmpHfPaOEw$<5Tou@z5P|uX=}E36=)L<1DhCU07)?#&H|l7
zAfvsAi(P57eG~fWcSiV4!FTuad25mklo={p61_i;4N?Uk_K@Bc@*D38=b}<U88JZC
zA~v-d0U0)9#e7P+XNz=iYQJ8cv>e}n`g^C}4rHjx%r~~f5Pqr%1)7@BA%#2%tsRXz
zcDbgtZ!!MI1x#N-oOabFu>c*Dd>ni0`TXmSCVIq%vyv^=_ee=FBCRTGPVNzzwju8V
zT*$4Egm!2*8}bQ$;8m2dpV+tiWwQUUB26nMwUV%9yzj$2(RX>t!;72Fk99z}w7D1$
z#*9Z7V<dYe9k%k)7Tn3SR`v*B!ELzdnCP_K+yy!_{WvJ)NoET*3;7NodAIGSD{n1n
ze9(GGKv62Qx$ikIPr3b~_R*YA8oG05qm9GDU{WXGVr!wX-j&+$j?8T}@L^`p5nsiU
z&yr{DO1ZqbjkA7B*?gTGWynqvDq}uC`|7iiurO4Mo}*ULm>~IXr@1JLt$b0jtjp|Y
zcH5pRlDl7vi}5v#MU+M-Csx=wZ?SYecU7w(=>Sjr;rQ4k-E5PuegF)-nqeceR{+n0
zUhU-@1~}&Mi^Ocww-)oaV#H<I;g;|&rxfns1Qu@y)!LmcIW9z{YWt4?&!f?r*w#{f
z^xkkdd<!t(0;9r&|Cw`Q<B&Pp2y3OVI~T@>rPkDmn|gK60QO8^s27j`t+rhwsVoWR
zV!p<4+u94`LWX=q=vHb{{+=qtbAg+vn#HVnr7NPIGI$ecJj1ZEFboOf&RN2Gm%WwH
zk8_hay6fO)R)>;WHv616WBN?UN@VBanz!6i+shIMH`{(V(C>tgFl>y(^5pmYM+9M3
z`yBGWOK5gu^r^5!E9bc29jC_7<41&3181`v0^@8XQ|0;7d;7WYdHUzs8PX$NTbama
z#9PYYs{ZvHz$J-MjNfLBs~;%~-mohn4>@=bKWUQ8)yWqkHazny6?uEY7#<y@^X^x3
zN3;5LPjJG8vaE$H-3y|C?W|r;UYwHEq8Tf>$}bZ0y;+!1Wg7R5Xf8K*ctcomX&=eW
zdO`Z+RSicIvS&Dk3)R0p@Dkn{cvlD(vsYwRW?o%MY|aI5@UbrWncaFxq>OC%s%%vU
z>LEwoXkU>nh0CwA(jVlEN;I5_qwrwJMEx4@-4*wZLXiQZFdRmclCW@u|FW*+RJHb*
zqEn?92y}pSjt{WiF=1j0j1%>}tGqT<t-ghFW8~iQNc@5G<8pGH__V9+Den;k^3JBJ
z0cr#O5=?6+%gk=JHv$<UWU18k5bIjQ&(Z;5`S^^_a$mc){ST&w3s2XcTC3HIYJ3|<
zdH1}v?Md6@1nawU4Omd&`zmkm7=qB>U_jU%I8c=1DK=mjm>-rb9N+Ojoz1=G`kop#
z5|G&^NU~K&N#po~k_{SiavdOWfV4z><u!mkk0p+k+7z2;KM`L~t8SpNzH<lof(}=+
zy}E0<vQ*XWmS^#W(J-!(Y1F_$im|1KrT<&~fis;8-Fnuj>bkj8)U9})@R~BNoPv72
zXT{AT6D>7$_rDh1?k+1po&z3t^TYkok@E`!#I|#ryMc4ZACQQ~a8HIOJkeoE`Kln)
z2^0TVPHfa^B4p@Zr#+=Uee{h}5}&Z-)L|<Km*TBlaA?QtFG^utETCX0o!)>lp7!MF
zm}Xt~GEbN3^ZI43KZ{sqvFnS+EnlyzkWeYB$5W`yojBvSM;8|V5Tp4>U=$piZVgw{
z;l^OU94!>I<ntTEJ=gh*)<@Fx()T`jELRT=4j02tf_4tf4OS5CX=6z4`-p>Z_fii4
z9Xr;k{fY_bHvS{tN*=ymbd{9P__z~xK)t4pqL}}C?Wd~;<nJY?1p3kPZ1KxEF$F;@
zN<mCxx*mE+Yd-4&ilVCHK353etn=)wW5e;zcoB`ZsfkH903V_4ACgp%OSML7h3IvM
zez|?T%c4wh`}V{icMQcC^%-@)w<03Z73zNV5_*GCId~Uk=C#a2vJ*#!3Ec5n|Gyh>
z?K-EV+-e4MZ)$570ZO{L)q6uAaV+Bw65zFg%uKFqIF^|1=y_2jT1etB>z0C9u<{ra
z<phHLO9cn68~?%t>4e8(a&|RzPbV0t243zIXMUOlFQP`t%>0<dC)4UJLoND1S1&W<
zrLAkaNUwyr^z{U+Cx!fu^SLRgX`i)mg&;3OOvnirSga2<kNwalgGvKFW8y$@4^ZCz
zK@IkuaPh4qOF68dT8O~G_68+3rwG|W!PZ{%_x@g-=TxGLMFNI&KOb?H8t3epPHmj+
z3dRaz{#=Ju62FK?L+i))1eCAe^Up!h1!Kem-%#Q6>cThp?z%qG96nJIp_xSOodVcX
zt<zFJ{No|9jasU;fIt0=|6EJ?mC(Y{k^o!j2LuZ3>kMDg|JYj{nyp_MnDLd4x7o16
zN<r_hQDxqTKe=&ODbL}eoKCQ9uSlZ7ZM|ay<<cVqmWEy6jsbiP!WSKs$xp9PLG@)h
zTM_Nu*d|U{oV~M!?KLJ3dseI~Sx;v&1s2D#-oYg(+0^CStS*{&9#f^2m^=jf_y~J@
zmYH!n0Zoa+oe^)KikHGN)&V@8H0vO%Ue}shDu%9YEviqY+G_>Ka>j-|-)9k#=VcY#
zO{eX#bxD)m=i=&2tK{Ci5~{J^@oGvOkw(StAAOD5KucL0WEd-M(Y5X0NQvL;NU+Wi
z`SD1sI59A_;k4(s<?4v+)$@P0dDvD@!`a;jTP(5W&2!uST2vT4vQh!EtFhG8)R;sY
zAfIDXz6h+`<*exJTgt&uDxip1+^pEA2GxsCS-^LL8m`^+UYs=NlTsh=J-bvC7k*vW
zMV?1DEUBI2-=KJz&j+v8F)9+D7e(FnTS^T6bk^9ZoIuQ-{FmPui6m|-8+Rjx2U;PO
zdf4LG;>(xGONM&1AAmU(Li_^x?P4>WRGbt}2@6-PeoZ;2a}~CZ0hIenk;|aB>Tk8S
z;hqXCU#)c|iy~hm#W@1g8W3FVSS|Q1_J35c0Ys$JP}`?mGhOW>`N1BHZ(DICoJ{`X
zOx3hBo5(#vTY6e!h%G6Vx8>dD;l!hX*J$b1vfm9ib#VVFObv_B-_l{#;DxGV_rAz&
zmZ#~94^8IZA~@Ydv|YV0c%HIq@p=`VXDJ;Z!2xT^jc{MtJkdD7fZ(*;9+Q`s2i(Mz
zz^A!jduszUa~||eqyruFrEiQnKyF#vw`8Rqf9(Q{!H>$TaFhFQtMt4d+tf{*ay%q@
z+82U5lhq{uP3#-6U8adt!e@zNxKp#KGDoYd15R(?nzQ^{c2l0EfhQ^j#FHhNvW0lH
z95_TNw=a{u7w?y|;&MMYBDam9+WQXUe-jY{or5k~J;wEI8iAxyvezJnLpSsybI>df
z!y5E@{r6-@Wd+h2z#0FLFS=^=&X9f0v$R9e5EP3!z;e&C`snu?Y^=T8*Yc`8zRX#Q
z7W<;q1~)S97PSckOJ<C9zhdAO(f3Lp(X+%Q8<w^>txg7yTXNli-2<33umCJZ0M^Uw
z0@z7lcY9!$e?+yQhdYenfs$g5`^py%B$4G*T&TCjLzk;hcM-#w(u{-9yh;q(s|<2)
zIfre@(tFG9o0K~Q6582+z2OhW`4fU|wQR+j;o>8j5-3vK@6B~LJaudEJ+!0@whp#Z
z=DF83uffXU__d7`G3(u>#b$!4R$Q8KNP%Nn-r|1zVLt?KzwwC4n>*r2JGkTt_{nJ{
zw#rak{Yu9qrQOORrJpn;DU^4hnpUbj^Zl$oQuq@{VN;?sg1@tEUTn9bf!{hqZaH?g
z+4OlN1&m_gbgQ(pluBOKqHapKPAHFUP4!bjj_7dykB*N}4;x)bypo7eqlH`#b^6>(
zaU%6A_SbZ6W`BW`?;pm+qeY`ZvS7|DG#=d#LsESrnbnX;Glry;Ee*Yk>3?Qx+y4f<
zyJsQ5yVtzMuSP;aGBvqM{+!6l5(wBJFdM`OYbehLseg;Iq5YdJeEf7lC+xcPm)E93
zt100_V&PDsN`j)P+tKRXCVFp)+vMj*Cs{i~&#^1fNnC4zERpHRuSbGRYjxL3f!4}@
zxj2+|vyfW>*;8cij|*^%%6(%oq+FR7{eTc^V-u6-Z)tE$Y1+o4aH%mG?|p-)B792x
zpF6}q-S8_@i4?NLU-wMF-)a2M@+(Bx`VKLos@oGyQyIfTc(ZaxbF&w4qLEZ#<!;>*
zrdaU$^-}|@_m@Txvg9}F#GBPXCPX)2A~rTPr76>Dm_x?zzLqMxRQq;7KsU%(Oof56
zKj!sAp!zCqpqCXh2DB+cn;kk9yB8XItcxe@aB-N9ZrJY#3mzDfaLHx$OxOIs!@+|R
z_iqI}`3H}0q4xcGRt?%zzv}i(P=-^>Lu%D=v)&~z&I>3uVy}-ZOXs0_i^Z<um1ze?
z`FKXwnGk;~qIy|^x}*%}F2Z(HZ}eZ(Ne!dB3Gym`uyX?7!QJcu!$oMLCBc*M-(3L_
z(~Q}dE#`e0N)7HoB(`)x@QCDp<vc&&D>1mRxIhIJI?sv&vd^(#(5c|@{oT#KIQDJB
z^E5Yj;JCTY0E4tL{Z4EW_M*t2{73G^-uN!42JW?Sc+xU@d4A=PNd4>8(b34Et4k;<
z>@UBFuN*&vAbwK}vEa(C7K!4(xAqNp-fXMonrk#vL?dJ(+6?-MRN=FA-GNJdSKXM>
zIq$D-JCVLpS3?A0DlNI6h%>Doo?wdBX59dQ93Xd^RIA55bWp1@LSPFbmPm^>E-xD!
zk_^mPGgfBzU0(oG?25;8uNE_V4YUV43*r(&pP#TNJv#4Kf9GzV*-xfUUhXXgf<I(4
zOQr+mhTmct#B_!L9ii4i!F<>AxTqav<@wJ?2%af!3@AHOJ6Wjao^32vV{E_Hgk7eE
z$kqd)mx7M<@u?tnjsr1cbNaXcJVLbhe`#0u_Uw5~xHv5E6W{2G-bYhPEN|v^j<2XG
z(phHtlK-%=CHJgUMsBxk4^kycuytG9<(6Z~lgBhKKN1}jLtH0s_zEpP&GWmL+_1?+
zl0WpUe?>N6=%DQ<aZ_;~Cwj-sPx%jz@8s=`K5oJnaw=`#78GVTMXsj@h>M8q#<Y`W
zw~b)o?PXvqctN0U<A;dKXvb9XASUd@+X}&4&&6?>e;|CDi$LusuvU37e}Ih$7$|AK
z)z!4-lvzr0phA06Vv87gL7p@QbSoQ!BQskA?bhb;YwP#@%c-N)%h#$-1L%8af1de*
z2OxXupPx+fzkS$v((zR4*ES?5kU040-6Yv%QVei+yPMjDt6ZGB*z8q1;~i#wHP)Z_
z0T7xyZWa$$TMpZ!I<7K0go}G0z0qn-tu=U~%v7qfA9Z0QmvE*VnhP~97|NfWFS9v9
zI^ytRzw7GtGgh#AT;1iw#|yV{{}iCuzaDvay3RoXI)+5{@c)OOsSTSY9lnR(Vs}Wi
zY_<-~oxfnJbTPI)Tl{#^phEEY4{M$&<_i3f+<Aozh}dvu7=sS{@Lnu`$n~nyPwtfw
zMb@B~;fpE|wyk__PrMcMwZ5=P=_GtBb+Be$x*jEx*6&){h+nP1bV-@@7X6-$Lo^;!
z5VY1rw`XR3NcCDL=R<Yz;q2efj+6@(mXc8p-rxEw-J)yid)Bpy;s@j+g2;nyQVgc!
zA=8y+gc@&0vk0(P{hq-wr<@E*ULOJ|rNBFsSua3@SiU>(+TrE$5HWDs(ocD6{LAu9
zgMFNPef_|VhjT0};Gj60`Rhj(-bbY;>~I&CqqSNYD!JIR=;-tg_U|6C<cR~FB0S9M
z#VXL;vA*kcR_e5IBl1h`bY9T-Com#!?2SlP1le8!=?f3+Zqme^C3cyeH`aohTF<6U
zXM@g}jY_A}beN0PShC4btsBqgO${5KQU?w?TAB}08|YS=onC9FUY5~k@M3K5qgqgp
zZ|JjDYb^Z7Y7DR8frC9M78@wEfdr2^WVRbJH0CH2fyi{h1)V`}FFuI7#%Irp!6q8j
z895?l*zh#?YG}#HoEH(^J?GwYMX8Fy1imfM|K5L2&ztJNvC+u-7iB%ad|$6Py+wiH
zqy_8UE&VngwipC^DQiwdm*HB8Av4(Dl!U&R@*52?&X;g|JUbL$;5(nS{jrVa+M04N
z>0&enBgLXh|IEYhYwksW{S(`=5xPdA!ChB+ZQ&n~<Ec64dge^J2n63-&G$vz6GB27
zmW$=9v5C{YR0$X(Ge2seZZ3f*f&rUGT3NkX*C{b}gTlY=zUUgE*=|&k*D|WpHRn9-
z?^selrbKgTVNwMFN%zCwJPA)*GHVH#2W&ne)SihoW(q}QAYq>7+AS{nac34y#xyhP
zbEI>IV@l2f-J6yINM>4AUs)Wpd%;H-BKlBGCQQ7r&$Cs#I9xATadke$arc^#zwoGj
zIv)K~jmw>i1Jm92jz&aV0&aL4zY6yFxB|3hK6eBSya1^4*jod71@M42xNc<~S^*=K
zAVR~(T9}%yGh)u2c|9jgDUmU(>3u0uWrds#2gDlN%rELCVu(7j_>Nb%{TF;(EHcF6
zXXT=sItTt*f+C`BM<!mvw)iaVZky|kIsLlUkWDm~^=d(+5vbxV3+EvPwi^RsqR~ws
zAsi*BX`+OdgAA_Pt>9>u$;F`#H3zt0>(=xH>7bx{dmqV(`OFM=aCcl^*L4`mXU4JO
zptojx%$%oKFM~1YGivxbifr@1N2UkUZex7Lw7xDK?R}qquYh;e)sLZWt0RiFSlhLP
zge==<;qO|V_1V%+aJwVG^3#fP-wJfR>>T8WF6w^57oQ4SU!2h35ypNN?~HT`wYb~6
z0BqIv0vY$&K%EO%kna3YlU~9B?Q8e9w_;N$4}j@(QRD(Y(&ot;b#`?<c^En0>OA3?
zH=%8W?@qWQu2j8fmRp>%@!FmkKro)f$Mc&T1uvR=1+z=@%Do!-f={sa_<N^<K+g;T
zwquyI18<#89nn+-hN6w6?^p(U)9<k5^zfpkV`r#nS9|h60&#b6I=fusdMI+(PyX?v
zj^7r-xXE?HaM`$t#|tPhCjBB5Ll>e9wTPWH0<JZl;D7I$Sy+gUiHQ*s5fSksd9;{7
zNl96kb1h=5HxQg&5w*4MHr@P1!0uB=^-?0yXJLna3yCQo1g0deCly&ed0f364$e9(
z`U!E9wNN5UPhtc}oYI9g)V0Q1Nwp`8cQa8+*OXJ3r8tUnJ(`eor?TF^z<`2SyjKu&
zTm+qhpL%dCczVsC((ElnB>>4(#)JY+?kZ#WpIduDLM~Ti6n;}{+*3pC_A#REvBlY!
z;TJIcw)NnLt7z<jvi*rg#$P0(?rSZe&$RK=jCuP>B?9wPDD<wR(QNRsw=tgQ$rY!w
zG1ipbP8EWxvcP_E3=GSh6#8U<ooRp~i%ooyK#-!Rndneq%k`a?@j!~{xu)(*V~N02
z2Q)ctN?vxVk<LY?HHCZ%8|k_YAKaupP_79$Lbpay=+(J8J9HqJHDdCMj+3$`4v>HI
zqu<i?U~=Uwym%3dk=&V%wcUM3GrD(;LOVs$)Hb}Yv+TMC`m|$~Iq?(g6RsqWkJ&5|
zds=$m{0klEz}4(cBQt$ixCYtRdRN{PLKcr}*M-RNKMPT1=f%iRxPFg`ZXZZ4lHNbJ
z!sQ%|k;|JSd!m1hukLOmhQ7Hoj8(8G6zr1@n$_J5aV{`8!4h$5jZWl7Uh|)RvGyrp
zTabp57nM1}JgglM72_*U@U6{LU00N$#<}YZ(i_jecL9OJbqmE}J3?O{`9iiA&#a`T
zWDPX^_&5u?0@&(o=pX!97*KwP*yq@`>>C&Nq%p0B$Y%NR9rt!{HL|%xaNcv6+XD+v
zXSK$a@a<r9pumgy*WmC4&P4zB&E})#e%S&HN^1vK-y&<Suc9GT)&2%ACH`$L!0(Pr
zV$x^d1q`QO`J1-j$dWF`!h@>x|JVzZ4xYOSzw`F3y9OR^S65f=i>GKnUYf%DdHIr*
z7#PnkZJ4K_rY62Rl;ooHC(sY*`D>+AxRdC<mi*=czt%O_8l8L%Ig#?6x2)p0svdV1
zDhr)@*Oi`7+7aIM3HW8g{eN)b5JYw7;ue8$aTMMWt{pd?Yr6<rH5q3Qmb&ijjEVQ?
z%F47!e|6BLd=jC}jHG8dt{zmKpJL`aW6b;II3UT4L?4N9HZ+@IF`1x=oX>8+TXK(O
znq3NpiV?uF65)y|bWekP;-5|6PA$^8d-v;S6@1J49DAN6{^u6eg2G`{hfw$vp$X1u
zGlF-Er3;%JePVA6f-DaM-eAPb3*}zbYcl%I8+YsgVVz@U4z?U|((}*ky<=j$k|%qL
zJ#ZN-J?=a)63PEWhu8n{c$WgjWFnPe$se`RCWL5?LHOkcuX$NvAByb&qdpL+V&Udi
zgHY}s3QQ`J#0_O()59o?al#UDjPQta7|BU~#(#?SV~OJVh2L?xQ-t@jblc_nFl*p7
zlvP)USK_9}r{}@JTo5}IaR0y8!lLkp@F-3f46+Xn?Lv5pzNYLIX9G=_xpE(1qT`?y
z2MZ*QM1Q$(-p;;O*?_bAP$z2_;n~7XPn@dsGqD-OUPny>qxvdLNj*cl7OESO2Li~q
zH(%v?Z+_xn@c?PE^*hNbV9N`toNNA$1NV<iK8<M-Y|Z5EZ0#~K&wuyj0ne8#u6Ii^
zsx^JD0r<v)$ry@%Y&!s5by6~PiD%=}NP*Z-oL6HRM8nfT%Dp9vN&WeloIA%aM>f-_
z;(Dt(I)V^MiudF5tLFnTo~e3^n7lW@(l&AAj3SR`2&Ru|$LV)fvMndWBTyr$V6Ya&
zEEs5+%5Kr5e9?ab!BG3rULR5xyMZ6ZE6dc}$=`L7?BiN|xvg}*?#O7y=GJ^wFZ5ij
zLJ<BuHu{OF(uQb<685*?ufHRD7VxBWbhx)KR(tOyB|s<oT!oZG-<gRz&RvO@@8}2h
zb*_S0vN($#2xw;F7~d<)j^IkR;AUh0*b|Y^raDWO;ae6>3}Lvi9_f+^McnXFsKFBb
zN`&0B4S+#ZrX)of8Sy;#xU|&ay4@kwgS+D6Vo%E!#krTr_`c6Cl<o{QtA|Q?xYCWo
zOY}@2K2yQFc|_-1%98RwT%IOktBE8(e%q$rctgKhd&sXfm6&qELkHBd6R%#6%F&@r
zF+#MCW!65R9&)8?)r-d6!nJCFUvQ3Y(BjSezjv#?^w`nRz?Bo(RO+>Wb0~(Jz`X+L
z#@AcRlO)~gy7DX<JUvz!OFkwVC^DHG<q_RZx$YsUsN3M>S!k~O4p!nGBN?%x^!39;
zc3nsZIo~!wFV@AD=jE%n<D=l;B@IZwdLs6&iRZ=1X7Je1+xfSwUbVPI^#|5;sx_|=
ziL#*=A#)Y2%hx@D{@bIb_RrNTo#NgCPxwzmTNSCq)!CKc0@0ZUr`3?Y8Pjy;aC22o
zCnq@}9#itOv1D@A!l7qtxMojvlwWZYx^tt_krn)d78sQ$k!I0#z?%j%QH|NGhj-JE
zDQx1D+Pd~^;AkBW3JE?c0!khuok?LrvK?fPKIKSWs!VWCoBXQa;w);LKBnVk{JAz(
zX+>``mA-TJha4!mSe@Q9dY&-VRoErFk>>PR&SV=mDQEa}&}fq@(lyX6iY4`r|BiYR
zLl3-4g`?}F(V4c_zDU!0y!sw-9cT5CP;;re)xgzc{TzsE%K0zIt3WLdJE-(cJMpjQ
zF`}mo@q8*Ov%dArm0Yiwkm^>;vKA~UV$TXh(9uYmntX(Ipf6tp!hl99{>1k+Ukp~5
ztXR?sbKgjrWCd8l`CIFzNU^1bKnA~<29B(@w!19&uCw!^{X4yYs$mf&(P6;$Yas>9
ztAr|Oq8W439&eSCvR%4EjfmcP_V<B-tAr4B@+e?ABGb1Y<#QX9>vCtQ*~!B5h46;&
z&xSTPtu08HyV;L^8Wb6K^6cVJ;#AIYNo1+fzYQwMh_jQ8Ng^fJu*mRfQ~hok5MK^U
z-92;V=~8l1&~we&MK|qv2T-7hh8Z))$c|hm^mc<x))JGj|Erg2%l(sMxI>6VN`6>g
zUfu|BG*^3KC|#VLNAy7!E!1>ET(h7MJ=*b1ACw>_pNA!`Ya{+z$da1YpQ{aXrqrO6
zL)Y@_DaGVcfyLU2z&-zRPh1+RkQCOT3bb$0yrO`h;M!7~-`4pJScUV>sU<rkjkC~$
zQ5+Sby~M(py!#=O{OJ=;Y?gWa$op~1-$`@Y)BYfdq8EL%6rF#a$&8GQs%mPqQoa2B
zGZCBGf>%nxKU>90LzXZ$IGQ3-KtA-V&O|P@?w&#I(U%j$<6?m_^)ILCChx}*De?Fp
z{QOL^sfyoA+>AW`fZM3N4>r2QAJkUC9wd!VJQ-Ip8N^Gr<tTb${_fCSB*wH-?;=H#
zVNeIARy3ZbR<z}L`=W7Z>lFCY10yfE>MOkt1k1bWPGw`Prcflg5n@KMT!4c9wU)TD
zU%L4XOWCWkhxzyg4P0s<`}%e>PLhY$4PbsPX-aD$R*^7Z<|@YF@4EQ>)ZqoHycyvX
z8^9Q%P+tO9Yf~i@JCa;z_9*2vs{UQRb4GXbP0n|MIqyuQ`0got5%XX}sCB^5c^&oV
z?&c<{Ey)n8lMb#<tDH`B?8`PQiW;ccJp|*s2l(F9AUxbh{p0m)yP5cOcr)5Wv(Yq`
z+03wS0bOt+)OGlo2Ty9VobKyZvnAzoRzYl{Yis8MwzOz#FTt_*+rp11egu&_HPvU0
zZgiOZu;Rttcz}XSFVTAN%cH!u87jFmKO*_42w32zn74;J!p<bLoWU$v(Thc-4Ka=1
zXq~gr36=vx9<T72Z`YJW3mO-?gdSEbnpTnmcD?8602apKUE_Luf5?fe*S7=iO*y*I
zPAA82CWT5O^t@%{X1wv!1z;t#)rMCJvj}fx2+XDP1Ym!taXcsbE%fW?<nfCg*-W*f
zs=B(#2zk2r1mhj5$DB@ZYG1i(#$TmLWQxmeeiy4ruv^QuJ+iOL5|-+UXeXfZ+2ZKq
z-K>OE!fU*7ec7_&%IL9e+U?U+S*saa%OoBL8WSkfrCj(A9S8PHOD&cJeo-zzglo9B
zvvP7H$P-w$@qadH68~kd%$KPFL*sfvRZbD>B}PlSV6l&sc+PV4U@YySTUEn_#>U*i
z-!Wf@_+Iem2vIUDYHw0BLKC%d_~whtQ?$?ac=gtyMjUor^Sz+6n>JG}0o=RGhNWtv
zj*CUYKOGnA!)f+1{#}Gzt4V@u3%n|vc(F5nAG`nU!DbhIbRd_!NSLoL`Jj$lOa5TY
z#gEEKO@rn#zmdPqC;Jz74eePM_Sif0T<ekN!LoYZtCS6hDiZ=@hN3fprl`sJ9i*O)
zTL#?&xFuP}jU-7+vy9w{_m?`onzZpQJ?*@j^bfpNwf0j(+*WFh%L<Zu@o|^@kF@<&
zp6|UW`AXtxcF}5UP%^eoXd_ps24Pggz}k|EZz4{FjzAmvgvA#w@D<k6Wcide?w1vn
zY5sy?-jato*J@TRAlEAwb4iSZ(GTM4zP7MGuR`+Kg$zAP@M-Y3OL&ENq!F=d3cn(Z
zrj7pMyM01vXc;vz&ZUs$b$VLbDa_XF2|aNUsL8Kl$K+)F8L@fr$ON1di+5&oCFLmm
zh*2)He8XzvOO7jwd51{pBJjN2SWSYsz9XR=4(~B!N}+9|Y(Jk~(W#`BxU=UnO#@Ml
zce38Epd7nlcg_;`d9-Fhclyg7ez}pfH`t+<M;2`}3qmFYSKEVkmlY349+JTY3^&Q^
zAc8sk+m{-R+(<~gejv+LfZzbgdYh(zdm2hee&i+FXMt?1z;|X*#8vuTY}scFi|A3I
zwUlhp4z!lM3-lD9f4d0DFIL9evy{uXEYsf-sebdj^4_CsOt*fp9ABW<Zd2GZJ~!Xr
zIGiYkvx^;>Y7mf{5JTM!uA<Jw|HzfNRKK6f4pWp&TnRqeaBoQ2vJ&kZ*dFLfoTu)0
zc=EzZ`Da0`uBB|@I?vgfovOAbitK7ka`?n!6uOuc+?;zs{RZFt7%^}e-xSo0%ljNe
zs0nyjN(fz7L)7MWa<+UT;&60(RqUwaFKQuWarNRXl><7alf!E{Z9%oT0nW}Z04N>M
zebKH2s1U4)v-Y~QF6DO*>BPy&1@bF;I>e%)qQ|eiOlhNZ{#>+{lg611ScQfL*F653
z;}mpC_pemrjPd2(h6M&)Rt5T_C=u9(Sd15~mP(0YR=^O~1R->qG0ArHutvnGvQ-5q
zNAojWIaurVZ=hd~vZcjWNeyN(r9$%PQahzYuMG=YKcfk?tOzATzqXpT3sBb=?B|Rc
zTm6Hq0jWLB+lljf2d>i@2CSop)p_fKkL)wtsmw0x31(bCI~l#<Zh|o<@E~_>ai6~S
zye0S9#Jv+SyhSj6Dx|yE>CE1!W83CI;Fp}A<Zt=-H|@OMsjg{0hO?e{Bel;AB)H>3
zzu=Kk6}2-cWcxjMhKxo$^QEe`gKD~m-|FUbTJH6yL_EV0$DbvCWE$tEbCyxFR`x2>
zlVSyM|2!oPcx^m&SXnkXyKz=#6RZ{Sh(Rvz<JyA9m|W@`ybznjm&tGFa{f+%#OYsr
zFtvU5b8c;5piNRoC#yZm;`-pSWUKc$OMQ4%Jurth=xS?FnHphHXCX$@i$_!Xxp_2G
z*vV0*_2Op)?9w&=hu9kvM2CHjuV{%%ar_^rX7O19-?RxNzT^5|F||Wre5|P}hs?hC
zy0(7b72b9PBC*UAq5+Cv6GE@&SH9K1tMD;zO!<C2;22sOzhgc~M@G|*i^%a&mXfME
z?kP1_=YO9d*?kKuJNDs{7@UVp^d|B<7Ns;KoiG`oLs#8lgP$F}srl@koTj;X)5WSx
z#KuhdtQ{ciz!h9BSTOaSO&<%_tfodiU(bz05;5bn6BQRCuHya^6x8(M&bK=_^u#`K
z+**s$qhU$)@fm}hXzp?3VnoZ^oKLQ}M_qqB#><SyR)N5(1<l{azJKY4KNIV>hdv$K
zJrjM&s2Zs1(SPYWZ$aWO*SV2no$-jQma|{L<w*!S$&CMbs8n?JwWNLErQheX3B26z
z<Q1nsBvo%Uze#|ft*0{A*)<vW0K?jQdwYK*nyxHmWNt?<(5oknSH>lke=LU0CA_@o
zNd)9#Ns3zdl~(AJc<V}Vm}Et*WTn31;o$+-qvd!(7_`|XK`(;SZ-6D0O3Y@g?fD!N
zOS*4cO^p~;no^)mv^G<3dMkKH6skgD<g5d0*J(J6C)z)~5(93VcEK`To35*p_KSaR
z{g+}Qc2}Hsv0`1!G(WYps!eWqw6JTyniJmWyPH5p-z<rm5P{beipCBO4tjv&Z#Q+O
z<Y7m6OyB^Z9|{Qx72gcCVPy7g2N%!Ulbl|)x`6QazcSH3t<ZZ=j%x&~h>aXyCrbIC
zk_E9}<|9+V3|;Fw0=Vy<tv9tw7?Z4z{pj_))QlVqL8<N@3(S`W+T6ej5c_Fw|05=?
zEahoy{nPNWo@n1kzcTTyCE!~_W>K5wW<(64HiB4}hT_4uIZ|eVO_<K1Fme?b@vc!s
zHqkag{4F`2Ooh#Y!$ywcG+w9_yI$$c?fGF`%=cxgwiuT`pGl%J%wj`nA5f!FHkr3Q
zGF)+UoQF6l%b;K4&GkUyFc3Ibbu~e`I&Bv426P5cL@okCFAsz0@yg01-tL|fT;B!y
zpCjJlR$3Y8OxiJPR;(7mN;H>hxY`Km)8`u=C!dLdGj1E-kT~Vn%1T&qx$P@~ytk8D
zimbjp!k^&c+ugPM`gD||5;C_{!(X3Z<n3V&O$LGW28zmk7Njd52prvMWM%lK4rc|I
zB%lN*?nJ*5Vyqy4E#iCnz(KR8N__7F)mz8J&3vuM32LkVG|#B*QqpJ^p#$)EzrdJS
z+e|A1OQ<~{1AA!1)QXwUM#{|GoCx0MZJ!*_+8cL*K!(GvzwbFsnkQz$XHi*?0`2<a
zd-cYZl}tz4EdvUVS6kgh*H$$BnP%S2j;Qpc^fEiHD%RtCF!%^HS@~4%rDOTw0Do>~
zCgSF>L&9;Xb+Bq~8v_#qLs3Rb+rX0V2?zfSW$BGxrD^%}LjPs704*Qir6oRbNyA{J
z=x<vsnI)I(YRpAcR#WPv+-Tp5BDq0bCzu`~@!V!uY%#uC)r7aqS3~?hn+$YazwOXk
zj|)og>5+Ddy<a<wxkoRH^(yOS6Z{r>bv%y8FRA#if1j0=<-J={v^tn7uqHhz@k8Rc
zGw^m%UB@L8<K6inpb{cqY@>aTVgBnv0@p>odJmVWB*!d%c(}G3{oe`#i?Gi8lmgQd
z9uV0UL1XVk8$Vvr{9>B@6!cHEaKYXqJ%0uEJRABUFhFW%)4hR(^_oTY>)75M(|TpA
z#d_HKc<Al~qT)&T>15@VZ&m7AcwD4Ju5iW_nYFMzSDTb6bEQ><TS_Xq3goZd+3mS+
zK!5a#N<c;X_(@pYM_-wy_1zxTyF-CbAQ&tvDWQGMZduj{*Hod!MyJUzyWk*68Fsj=
z(X>YoVe_p)8M9i0W|dFE`o)pt`tMBEcpXW3k)$_ESEG=rz;aI`slkn%`GX3xjwQ*C
z+2GFu_*HrDo0wri_g9}Su3%rM<akk9_b<~=eZ7~7D+uWzRRrl4=09<weoIAGPopMz
zg-C1J-}hm^smWmyhqd)<{Rg5VuvkkFl*dm;rEydBu8#_KU#?uuEG_+ue-#oFv&Z+A
zSOMGFp^PzTQp|hgNBsb%*X6dn1jR)Fs-Y)q<L>y+`Tk``s}tM(GOZ|kPZ^-{0smfM
zA-UwW%-+5=%<R%yyc%~qE-U7yx0F$+`La!HX5dDcazssfJ_jAR?txUMxIF(b9RKcn
z<fl>I1w_y6bM<kz-#>k!dfAWbp%w<T{mrq(B*R$MtQi%bLI`l5H(;`j7B};?3?YbJ
z0>-Dj3a(p+X9aXOIY!GsMHQ~}D^;on(Dpxl^*mN|ENpCkoNvE4g<v(>S~4ipIJu6>
z1niop`nz&vw;C@AG(XVif+W8EPIO>0KTfS}a(G96P*H2lN%=kB$A=<lsS=9!0oi2S
z9J`vfms`gGR<YyIu~eFP6_ftTlL%K?2srq}PWtGNSEIPH{j%&R%h@xcJzuL9HcNTY
z<(*e)%~^Br*+aGlW)}6!*jB~Nk9?c5d!-5*h7)k*ny0IPBZZwdP~P4+kbQL$Y=Ist
zqKCsl@nKGoyc<Wp7V7VlJh^zs219FOp9M`XHU)cMsf!UIK{xim8N-MFIF&&r$!5YH
z5*$YJ#OVfbl=z?5z+!+I$DYuRwS>z!+9v9tb3juZ<#Oy^S_-rfFxAlspxB(6nkq5!
z<BsdN4atg#xI6?IcI$Z}kerL#r|E>}9?3k1P4i94{ov}?E&@@M_-b)3kpEt3SY83W
z0?+Az2+?6<3JvH&b_Tf~bt5lD%<A^`HdTC#(4#-VDOMM)2Iz}rahx`jva|G(&cD>=
z_}drO;+vLdnd(hCc$aEJ495<Y71f@df|nT|so!<_6I9sPPsoj#Vp(Q3RxJc}ET`>=
zVz^~Da?t%^Se}hDUed^dE)}(FEdfJ;;TJJUdV<XdL>C1(*)A)lhPxpG@jI_D9;W=b
z`_#>Ph9P8G1c!Kdj}0i%%Rn$Yte`HHwKc(>?9>XsPt$hA#pD7IDYlFg95$>anOE6$
zgY9w<3!3sZi;TZl_K_k<4+bP)Hebkhlnmdr*Q_UtG|eUh_TL{B^^;$Sb!6}Zrys<9
z3ER1GLqA26oa!%Kx>DV*6aF!OUrhKgMuLpcvT_E3PEWb+zThvTO{y}NNI&X)<gos@
z=*tw|&`x&Ifyz2gi1TMSSZEqwDVe%SYpnO)87ly~L2z}3GxC7KU_hZY?g$jWpNm_L
zT7IA<SRKx|IR?nM0k@%M!uI&dqc*QoIZc4fNv>XRu~^_}UPfQ4wH|irwASi%2i=u!
z#Tj!v&%~z_`oI-J$kmYVJ}cMJdt8)nFxk^1r&2V2okesq<71jNqtMMJyxK&QE@j&6
z-n^{SnR!40xwU<wP~Bs0%EFrRGe>2Bz1K!Z2QAbH^i?>%x3hh3Jq}hrNyu3wzG^Oc
zEnVekDDOCT-%@2he+x~dp-bz-c<%91TnGp~0`NMB{p#bj{hBAx-Gck1XlK0?_27j9
z+EFnH8P5;yjVcO({<($vPat{zU$C!~2{SK~RM}fQT9=dcvrk!*`lCCV3#&2;5;Mm|
zwcw>}9==q67(CKwPj~ROx~io#Uly0W^2QCR=pQyTlsIR6DUW3ATAuPTn9VLuE6$k#
zZB7Qs@}lSpBgUa*4fs|)Lp}piNRdp==kb7stuJ<VzAb)PSR)knmjRcsO*p*LFX+tC
zpA#U@?~hIIY5Q+q_`2*bk*-ahp)8!YX48UpnLh(-bIn{Rn#=$1GnciVn9BjZc1;-2
zbL5X38MglC7ek-3^=DngHJG5XTI547Z_MVRCD0XZm(uEsu|b!WNNX}I#(#3n9U^A(
z@8V^U|H`|!Z|}#D<*P^e0l%%MsUHjf8hY(_-ZCM8_k3s&qyu94Ob&K{jVqUGLl%b4
z*kv+(8ptNg?=^3H+X}%E0lSrBkPk_pMw{pmyeV%gzn(;p{`(jNX>gHedwE2;m~@Z9
zLBfPB$@IN0xKEXp100<f(CTf!$?H>Lqxx7P<lfG0yF4j7G9morpLSb;)pdQxWc=Z%
z;0Zb74!2L<TST-drk##XRa4V!gj<f+IHsBPk-r$+i7a!Qn})9*Yr$}gp0J<3-+c;Z
z(%;#Uy5nObZERi5!<E}H!GuS3km96V{#&Z%EAYyyiyg}VGta&|zaJ%r*4Sw@+Z9?G
zH_;fqx-~p=hdll~(~n|zOMdaGB(IuB29^wFk-)^9&&5=-daIq~0VkAUarbh$5<tM=
zxevclNkU;C@oBtSeFY`VD6ugqS+0Q900TgL<KTvKv*>O#Qm$3^;-8{}$Pxtv$#@D7
zUdk%*Vh*EKdHz;o>I2L)Na%jei|UUcsFv2L&FO_%s~f8*AQ`$m++9p_?Kk_JTHkIE
z9&}@8v|SdE!hsF+Ja^e?tcUD}N;Zpn;Bs#$c5s<E^cF1nrc~N*SGER{x6N`=Md7lk
zgjj&__^KFQf6RPVNyHv60aqVn#*ckv14-0_@gAvJ#32)d)GV+Um|PjMR+h@@*Lgph
zoGM0)%eAMvTh&{B%#AuR7;7T;rt|(Nkzls;;NRop?C$O!3-S$S5TqDPlpUx!{amHb
zM-Y;vQlrvOc*Vw5eOO30lSDPr1!;PC6hZ#jU>AF<23um$=~M$O(jAf&_vV~`U!Fy6
zcw@vCA;|&Tul`&9`;$mk&&FWt+WpO8+;<w^ciDaiO>WOmo>o5<Hm65)m?X;FM32EP
zZ1h2c;V2OJb89a+t#}~<CtyT`kjJW?)9LKkqM%NV<IPzxO+~ljrF`B;`j^?^8Z{xn
zPMMUY_m%`m>RXTg?h?7MRViEbtj@_kv)%dzh|i9ze7LTR19q;Cs>YXs^sAPh3oD`4
zEfgc^&m+uMcY1P2mrvGtmI5AFK|XnrLHGaB;Zhb=5d%uXmfGh$S4P@8g5{1g=iAjK
zMgPH+E#O+a$a=rfA+%T;us>J#uXq7!#zvPlxmHJCvY@@YHvNjX_p4lI7JL+Z+mR8t
zm$~@YuJIQVCi_^n3vbru*Q+<km>}wcY$k2_CRT*Fp*=sjwDa-Bv`;wlSJ~P;m%exX
z{S^y#e$VAYY$EcB5BtLbk^ACrzuUq$mVt=GC&AS~i}BX^joJT5y6$+Y{{NqdhRjN~
zjBK*9uapoXD+!m(Y*)6+C7<k3)=kDGNr)S=cet{%GcWhr>$=yt7x(fzeSd%Wi$Bi8
z`@GNV^?JUZV?W@f8YwgSLzVHxW`i8&raMIckkRsRA4YJA6!EHK6$(f17ra1!Z>!PU
zVcFzu%B$XJ(!K1#&uH9~zSp>IDy&bzwSdccaforj1eCl3#*y{&$n160=`NYrgZC(U
z$I+v7Ki&@0uWVuT%$FOv#Ut0H?B=g>Es#)(C;;bIJ16hs55|lXGa3IrV_QtUXh!G&
zvtT5YPeuc2Y4wP+eZ>X+Hy%wTery&kSFbc!ojO5AF>%=2V6P;Zy`=Xrn<g-c;($EJ
z%@w1b#f58?jps?sEv+0ZQN*x716t0J(>bHZ$WjiJgNutWIQ)1es7G~y*&>BERw=gr
zr$Q2bq71lg&p55GhY;c*w<+M8=X>Z1OlQh-86GVz$x!@;Dwf3B<|j$pIkuCka={s3
z>jLxgPQcWLw_eE3K44Xk6~-lDtSZ#lV0)hC<Rl{ldlv}D>>nJ_HX-*LhOVuMmy-fm
zgd=!+4(nM`)MVX61PGb47s@sdy7buf`(mW6zS*)P(nM!RVGLfhzge`NiGJ`3d-HW$
z3JN#6G*F8BNxy{3B^26Ob$<hP#~p9fU|I*I<PN((LB;VA8&HsUkfyG*W<t7(ivfe6
zGIcp$C&+OA{d$NqQ#0eS6_93BuRd}LUx8%i9jebFf@zyIs?&S9PZ4sfi8_toPr2Ii
z>TKhWTfZ;%JLny_sXQ2cZi8=+l#S#`$`6f~lm7Mpn{{pI^<#``e-0ds<O~6&;Rg?4
zCA|6<%sh*QCG@9LKlTltG(F8re;L_ncQ5u|$*F8=3iTRH>z&sq{U|p_Z`@Fp?6g<c
zmeakj{BZwt`n-0zLPf^|^jBA>q@wm@<Lb@f`c(-F&a|0Q_EvEz)Dzu5C@e&)`My5F
z0Mj}@+#}A|EGo7(5!xzn7a1BjM|{Vd(9x^ufZFd9M|J_{;1GOR85<i!>Y&545?iGI
z9d>;v(W{)?Knpf6HXcqbKvNo>po<?Y4Q$g|%ag=F<V6UX4PBqe+eFdbHnW1P^kv*m
z(gr{c-u-C^bdRtTBo9tXX;n(#J)JY=;^LX|oB?iXnFSKnl2%SynYF-h?xN!~&S#Nb
zDGR|?)~!(G-T5Xq;(VYjxwY;(CYuK)3K-qu8NHiDd4l-dqJ-*{#~fu#>VAlU9~IvU
z(eSyySXB}3XSvRi&%if<sQ4zB{jVH)Jf;0a+YC%U!UvVf9d^%fHYu?$4VB)c^Ou*7
z*(jPE9jtTn@=p2k+Z9-dy9YnYZ`Zsz^~Fhxi{!TLS{K`A?vdE{$>1&kVva99rM+yH
z=PiENDoQzO{h}$#?uWSI+-4qqV3_@IIKnQV?F%x<sM!J(LT_n!<%X`s$mP?>r!>DV
zexxy?(KBQWd?kb2+#sI%FP29+g-{EpJ^7cY1>68?yV9go^Ezdl==uIpIUy@Eu_;_r
zB(YWCN55l`P3X@GaQylO2E;jejdNC8Ds`^inK(;Za+{B8mA>!Yi}Tq@&lj~#XD{lg
znLt?TG1a@&>H|$uLHLeOxq^4!!3V#l_e??k?J>Z+RM&i14mnjv^Gp<oW`n!5C9r^f
zHFhZG`~`FjHHJUgC@D}@Qi@&iWe2JuGsSd--IE^ED~w3|cTnQU=20fcL3()qn8Rm4
zl?~)Wcemlk%Vg$1DwJ{`@VD-Fuwhksp)##L-U8tpDEaW?iE`5TAuz9dFT@e&0=Wto
zWLwuhB$3)${l(Az7=e?`Clda3vZhArH&fJ)URXE!iRNkKp%l3&KG5N`Kum^WtA=w~
zcr&3-9O9~ysBs=$EWnS!wGg7~6>&Y#@LLDHlEt49lycLnz*BV=|8@}L?$dIew0!w!
zaJd~Jpfuq8`rzsxrBsIAOvhZ!#?~*wQ9~bTxUUQ-I8O>VKzw~imFA5jMqNi6l~Sl~
zNyfvBIWG~+r2M0Te!eW<H1#Mkn9%6!-zi;I)NH7LJD0m980F2JjIU^t3HUyt5`RZj
z|InGsgBbL{Fo}lh5uFcxs@e3WTsP7O!~es=e9)rM-S^J4S%cae&33r&0HkznJy*!i
zB8(pJ@4b54SwmICQ-8GA`j_=u)@Y4~@8@O8`k_<p-L!qFVAp@$Q^&om{L|!^fuPLG
zN51Q4f#Y9fp?gi0%V7wo8Y<Q+*QSEQ!g^JQM!<1j{aNYWHqAm&nnK%KogU&-H5ste
zzA%{+6fTK^r^X-naWuuO38d>#Vl@tGUy%vo=1}~CQh31e5Q&nKf>lM!$G6PSpGX`V
zfnX})VFbJmHV!WjG{wzWZrp3FhdMdj3<1D&Q|C1wA4=Tq*2GPJM1pw6I;QjUBn3bK
zxVhl##O&SF7+YsEpMVRS1#?OAYe#)cdB;P%g&z`+!Shm%>bjBhO}C8H56yr#(K|_p
zCwSSf|1V)4f>R8?Q^5n`<mCt#1x><W%`1}Lamd+IJ+<af-TT%3c<De}-q~3?h58Tn
z7>Ss^oap_&3~`yc^`02yTtcAwk3=JDT(^;WOQN|sNO74z?!t&VAH%2Mrv!f74MxVa
zSF76w$H(P@o#Mx-7jHg$Pq^h{&P_0Te+CepqSv0}x2G+*cLd<)sr4=j%IeMC|J{j9
zhYv%(Gu10eSIb(r9oM_qx9oxgKh?`Uo}!7)@n+zG9)G^U`(b27tpHTu!sj0IU;@z9
zdH`!f+B9Dwfr?OqY%ix?{+%D9mZ?$!xELdnVAR~`h73JTM9vG9_Kadf+3!_)c@U!t
z7<az1h^YFo3wNR~S=`BYpc%0CEF;_W!Hb?RK4Xz@YfY|{Sh?d+?VIbmH8g2f8+DoT
zd<GyA_ByiEd80v!Z`Mu73ttUhOcs3jY0~bOR4nexDB~cbwBL6DWd|IqH$|M860UV%
zfAsQW`z+5);u?;p2j45E9Ti)*pplF_UUb0Q><5#hhPP1~XWQZA-)zqK?xr%5fH!ZC
z-QFiFVofc`2R#vlnF-*|*5uK91x%7EUQ3%m#Sg{3RI3ZRleak^Z|u#!@5W@S*2slE
ze^=L)U=$YGd_JQ|9$Iq21|DfJd(Q_Qa}8@lSkQQsQ}ga2ntO^Ts2&8>rQL7nf!SJ9
zM$w6kxn#Fya-(<B&V&~fyNi6R)iuljGqP^DAKSg@F?)-;PQ;dnPk|@J0V12wb{?97
zQ70bSH_F6zKJL=7-R%h^ccxuFfZe(<r{MLv{_m70k1WW_yx1#@o2J{6sUMEHl`CXf
z{Q0xBthVNP)@w&4hgE~o7QXm5OtC}-x0PbJ0LB3I`sWw@Z>n=QrOrl73)ZSzeZzw^
z9HVoGYiwK^Xy!n!uXv=W(x$Sgwl-}!Oc4rOpAuCN8O^<KI-TwwRkNH9Bq{k98{NGk
zxLj_Rb*8p4K>#q{W}BRRNFt=<5f#RZwd#s^4?7NDH!DCG73OAM3-q|Baohv)uC4}0
z_vju6TD{z=nHjo}BIfMGzS$e=>;j3XuQJJE=eBGC-PZSvSK7I~`{4JV+ilz%88b$F
zW}AuhLfTyombd8c1W)1+Ema?5XKasHA*Ja$=z}l_2hH>sY{AYeC{BkbdAOyD&B%+0
z32fczSSRwYb|#SocyFU-al9xPu?$#)D;i<i{_b=WTOGKNvztA7($8&jCtDgCEX7&G
zI71%)9Do3j1*u0Bro^L49t`sE5za&M#MzcCFeL=e^Xgy~QPM<Oe1AV<m#xhi9iDmu
zsO$Ck{ih3$yHt~50t+WA!>HXBWI`xoxKaQwS$9B@HyN1A|ArE*Fzd%{Y;YGv>|qZg
zX$p7t(!)#`KUc)qy+ophoRhmha`SduVY7nvCts<B)!%FjIrXKFcG2<mJ_v>QE2YF}
z_dv>FOpG2}s7Kh~W^5m%d4E&wq{UQED(N3BNh2JWeS#kwzVA}FjWxI>iW3s(nKrzu
zZ&C#?5i1L$A#$Kr>sp0QK;?HiaCO34)EQ>#1VA9$PxCkz0HNk#GJVFz1*;u__75X}
zEH46vq+sm~7ALNEGY=ns_f22F$uCVpL}ySQtLV!qCQ~0pQd8k0h&f)7mtezJsI0#V
z_cD%tQ&9$8+9rqbW5lT{>udrdaq7Jneyjt$owdwLD7~Xq^t5w^q|LYLhwEW{=PkU=
zj4LgFsx9ZzB0liC<Ry8xt@(%gHflu?vrnX|%OrYEfA>7@y<d)%+@WD89Qg+E=~vOM
zP1i$>jf6xmklA%6w89N+E=1vYMGBrU@OuY-RjW#;67w!D3=z`4bvmhBn{^dlg2weM
zb|`9VFSUg@3xiI`V6^~O0Wb_LSb^9^b!N8&z2?BLKi2?}{>q=B>XI;R$1AG%<5@sl
z1Oj*l>Y!Mo1!c*d7*Jhe8zC}W&G%T%HxEdQsIu<va1#Je^L7nJ4gR&d?y$D3H)pBF
z(L^D3o%!BQSl6M)9<VYOe(Zs@;OnrDDAHY>ElteKm7CWKEhOzHcu~+G<<h0+9=jTB
zQ(k0W>n40ZGc)8TrLms47=vA)=1ty0p6>>AvcW{o2H|JL>pEIDnGvmI$x8LP<Wr<g
zlE%#WNDyVwIq}3i=X(eww3QjrSa*}r$PlxHSL6AV)1_V43E(gQArKe#jZNdwKsm2i
z{tXlIx1t!<AN4VP+fmk)MnWw1!&(H%cs!>(G<8(ywpnRuDUa$W|5CNa>6@wrF3_Xm
zETc(n#7#Hi0Kf;bs`KgwCJCC5_(_14h4gmMA<@tG2xOFO2_{?2@5`W;Ln^-XD={-s
zcJscf`IOt*JqebjJePgFY~lL02s9$ghGnUG5Uglus6$)f_xHxe65$S&es5?~WhHZ3
zuV7xtW=XyO)^h+RY-wGK=yerJ`sBmx7s@$%_w&!2QdXt?9n4#Hxrr3JV~OG8oC$bM
z*0Jn8_7RHnK&*mc^1xMzg4M)lEFamOg<m?UUUad{;&YD5mlt?OxNKdV;q;6uBbC#4
zasw)cPxRbh<QL<(@{@l)*DFMlHcFE|036@FN?F_FIub}`D+^lPRm*EoO@SRBcwx^n
zAR0VP^07Tq*X{q{y1iRfQV>v?hj|HNCwK%*0>jwm%DlhA@5Liw@2wo)v>&Tl{bW1u
z;B!1!%V21td>ol6voq*0V}%0OFeBJuwjTf8c->VMPx8Kh_JSO#1CHw~7q%yEHaUd4
z%YEIStqj}3Rt9gSRr(X($jrk)Ur&d73Wv{N4mXSKBGvKS(v$}|T3y-)pEXpl=Rwc9
z9mdjvJVKC~Q@O|5b4onjhlNQ&D?S1hyo}h*J;Ad-oVzGTpCj@JKgLusTVLCM_Nx0A
z`CI1Z6U8y3^&9MT@dh;QF9Vog=NnHZy4|}iZ+O8`+008NRI^C{`dx3jWE)_nhvwcm
z<-^6i&cCzf8&v^Od^e*NCHc72|Mgo&eC{yZ84Uupz?2xO&-U4WYzf#7e+3DIWf_aH
z+bsypd#<rK#yK@sj>)x#z5@8zek$cFZ)JZg(Y%$y|JVPp`Rj?bqVI-5M7KJxThjv+
zz~5F4-0fOe02h|4DJ2q0&)9xXl>9y*Tz&9ocB|jTvMtm~y}{!!#6?D{VzyTtmPc7r
z>-GB&ZdKLYht5!O;0k(gpq*&(!u~eS-B~g)McCqft8~t<=by{XG+Z5<BN+Jtl4(R{
zYrCr&&}xM}s$!-FG{RChvV0|k*bej&6k!OilGyCrWn|jMh95rbn(AKi;6cDP4CibT
zF-7xfB*Nf)5Pm@oJfj|$J1rLATt=EJZ${Gaoh^(pU$5(67aGBt%WUVqO*huC|7-gg
zRuenBuQ}krRJPvpN+J2!jpux4&$~za`&$VvV77qaHME=m&Ma&D4t&RGBS)4?O8LWg
zQ32$!EykX*{U_EM99>m`3!cgvp!h%=<-|Z+_2l0C_`>cD@dnuHv|NMe;63-3ED@i|
zQ=F@w)+syb=#6J6Mn=u>Jy@z_1O<dieu`d*4A7GMm{1Mn5vp*x<ZHb;uTU1kZ&;TN
z*ifnN0%<lt%zToiMiSGOE47UmQ+>tmag|WasM5dz+i7DNqQtf;&3B%pnrj(;cK6;h
z-Z+a-29>ayv}l)V-N2LQ!Bl_oL35XX`yVqudkzNSc*<g{Ez0&|-IH@?>%0|*Wbgs{
zFxgwt()h7t#y;wfFq976r^X79$f2SuOUnG#kjiTfS!&yhaA@uI+m3z#kulE~`J#F)
zpIQ|NeZfVymJYQQNQn1;U@<OW%<)*X`ER2c8EU32RL5o-&^G?cvVr5CTOn<ITiv7J
z5+B)5qKRc`SkmrsTcWhhsUoJ~*qJLw%bJIXh3CztxOnZNu-Rp#uZlLwgBue{Dl;H+
zen3k~E^i5fxbRFdmexb&Qh3iFjtPw9hE$gFJDYcX=V6PQ0{Lft`^+ZX@}${i&*N8T
ztoZ?^zKy*TK-~{%nJJp|_BSh4;?8_|C@T^nw3ka?YXjV%uy7~BAiki`HWq?2=-`n&
z#g8fSCfc@w;!9s@cgcpdtd2AQt`eU$VsWPpr`Y*43yi15Dt^O4;2%@|Bq}MkJL{qo
z`p?rmQF5%aPu|+#g^C+X>?xm|ny9`S-yuDl7n(jyS&vgf<~R7K+C5r~h@RXSllx*i
z>{jwmcA97IjItpMXv~YaR7(&dgY~XX_I#>vTDU^R5TXLT$ah9^9GKP3?G}nkWsC9)
z&57QeGRaI;`zTK$@Ho*y@$cm`if^+~h;LlKH3`bK-kdp>R9+0`-L?&>>0+h2adDAF
zqajTU8`jHLO`j}qRCFGY{b!Y%aAJ)z^VZHLyirs3e)0$&=J0ZB=i}ABg8ZfePgdX0
zE*3%4)p<F}V4Wv2>(dpX?4AFi)NSwH+qw|%Kl9s(PQ6f(0Z?NoOmbV+(#Af>v*)h|
z2NHMZNm&STZ+*z21t#>o9r!nLgxv!jVZdY;jkBT`=EYV#=fvS16q>ZFI7$8H53a|-
z;x2x1l`s6b+mcKgqZ8T`u)k^a6b|{B_|by$QBuJ0dCQB((wQR(v4{T=(#jOOw`h@h
zO=p!8+1P1u^bNGqnXmOU<)H^RS_!`PdMi$;{%Gj$Y`8=7**@$vXX*~WJv)W85f#|I
zQfP$AzDu;1ebJ4tVnfVlp<LEddQd1sE-o(l8=$D43=9ky0Q12cF&biYjD{A_`E0LC
zjK+(}NObu8^(GB$HJ_N`#ZaD#&J*P{`}U80sXFajfkNN37GeVipUppjJQr_VPM(}J
zL~OiiD(Xu&fg6;C9dwm|Ya;vOZB7K5aD4x1$3aC!!zXjur2vJjm`(Nb4M!<Yrk{?{
zTJm^x)&&O2^B)3{dAtjcgTK&=rJ&xOr2Slf>CE=`E`=@VgPQ66x%^G#UV3<V^-5Q&
zx=$x!<yGH_3N^pbJ!3H#cj#Iibdg4D*f~F3GW&VWi=X;%sa#V(a2#uoAzrQM_LYU<
zXE4q;)tXE4O)SpAIUB*#$(S(8x9jPCLPI~yp4Y-Gkx%jlos)F^hUhQz96VPcGQ_n3
zgpP&%JYpAJ>zYyb`BBl`r`Yf=#JM6a0e$w>-QN~=veH5-NSRNNo&!nZKDc%r3V1)g
z30v}vfdu=6_$~qNFqB&;fTJB7TOaBze6(7oR2p4rn}ppcgg`?AS=w68OM)zL;`r<d
zQ~Z7zv6~WwB^4!UfP0j3YM0c`Gh2+g=7>W|&FAwFpPk%)KQie08|LQSro89w_{IM7
z<|B;*%)sehHGYXyAgy*=5!uGV+vuyf0NoG)!Mty`0UUBGxS?%R80F+}Mp7(md}Qk1
zr<8T=n@@EOE}`$SXvU3OjBhcpJfl0)nD9WGsh~+kr}q$9ODlP+#Mc~`!qXT;z1Rha
zUAo$rpcE~$l_gc1*FI#A3mj?{-c5!}xBUvgyZEC|Bb*%jO9p!4*!u6-fO(O<?`4t8
z!-v9%z`P&mE8%nPB0nbJzFO39)c9Qu-6V4U9T%7=BGg5s?Q@1~iHhJ9&F|&P<@7XL
z#Jf{jv#+zkxl}!5QmUnEzk{f<(wzvg6$js7KV``2N`weC)beFNb}`muWQJpL+RnX0
zz;&97sYp^}yWND&v9HX`Fkl!5-xQ0CPKb!3UpWeCQTg$-Sx`gJfa}7ITcR#fp3j&i
zZzYgM;jK3mVN13;YA3UA^CG58<A<Al6c>V>*iXw|ktxGmUD9f>c$SL9P&}^H>4t6K
zijvoMUYG6ZLF`xcysW2~M{uK?y&2*bZjI75jW^0BTz@P)LXNxvkQ&@aTsKVR+#9<y
zAt#e$2{gOghqZtNtY`0wk8Ynkx_{vw_#01hiuy@0_Kc24&1O`&<a3=;G9CDg?cceo
z-D)u=SI3+MLAR35@6Z}|iy5-%ga<<9!voGoNPYMkQc;>dJ^|^4Xmqy_1~6(XYVHtE
zfQW`&T+9!5p}ko{3@ka9Tb#|TBAZ>7H#4bH^LjZOQgr!euJN&1jB_l5h}qXnae*Oc
zlj$pIIUT;WeHLQ-f;vN7Hpp1RU>#^e$=l^F+e?`ve`I_{Px}H>9kB*KAzaS1E#t*S
zmak~VUSF>6ef~UV3={TYV$fY84V47unb?v3#y^N@aC{tQNGbGPUwR)C-_NsYHJ$0<
z8Y53SHRB)39G+{sGmT%N(TMtwGE>WmLO`0jw-$=C@KQnxfyx7fljTfLLONAn;_AV4
zsZdXmKAfWq2tooF{rJTwMpUk9pnYH<(fSf7!Asy{RBWuJzmrsOW`9Q&LzL{NM*1+$
z2R=qxlMD2so9sppyW;Hv3RcTfSHMT#tRy{u$KHEs5#{ILAaG*TEREqjd)<Yae&SmB
z64ex=h#?#WkY{#_UYND~UDqM+3ZLS4khmTKiSHi<|4%#$T;z4)DL$)-LRxaUE5mMv
zhokFa%f>XVTS62{H~lo)7T&;sH4HD|&DLaj(E4r<hR8$KO~jBtR3t3i167BJh5H{*
zl9DE@urG1<&Nla8pd{BINp7I-o;MP&Lk^a4vA~C`8~ARhDg8}BAFiCMMVDbo88Luo
zz-MA!&>ks^L{Wt)f-S*GFmll=a#md%N41mm*<26$Hez8q9`K?8c>9o`pcngkJe-%9
zM%8(YL1_!j&D3t5p3d#I7H-r6)&cLUxEW#&LxsP5H%A;bvR?3MQh?v6rzA(Zh09AH
z^E;roa)UX5wS~ZC6VLbG%jQ;mr<Hzv3>Mv||Hf(y&?L%#I@KRAu%MYgS5fW>jEsD(
z)(ep|tsQ?Qq>Nt^=##m6vP#n*yo}z|fInha1|q;UfJm`NJEW!2*nlr_2$lQR`d62)
z*v8+T3I35w*!PjCNm^iT^)s|H)!9HkFNWMpx9Ik947>YGkWar;v|Xan@2RH?y`k3l
zj&yLsU$_@iFCrq1I{DEI6}*97{uvMS^Ty$O1~31k@ePO%EwJsL$;kGPmR8sUbLR8?
z0wT65lnZB^Lxlu)S8N=guj>*%AM2wYMSb#f^=^EB?~<~<P`k-6DepXp9(^>(Ow_n3
z6Xbb;c!JOBE@W0M9&UB5XGI@2I;lZCIp~P15P*qQ5A3O}n*z#rkyG9h*g;@fO;}P2
zKiF^ru=7(H#4vV!T?7z)x}`*=pN)(d&ou|j@o1b`dnB&;1!<W-hC=twNcag{td(jL
zK4Ut7E3p<kyg=Mow;0q^*hJdnF+EsbGN%9EE+=SM6@3eaFF+)`;_6wW!^I++h+5k_
zZA^-^oun{|rkYAUTbnO=7IR&Z7N5&#o6>f6wuU)BE@DJoi@~aVT{_yC{r5sNtt&k$
z`XxPx(Q77dWesFTRGz_iAIrPILekujfdOZr9>I*#h3kMU(9Qe8Y-!b#Z0lc%Mt|O@
z?LE=-$YFYR#o5UX2XaYz2~LP@90E8Cw0{aPD~=D_d*N2gANFrw;4paGZ34?|E*pn)
z^<)(g&-(8gU|!AA*FOcl77^)W;L(Z-TK=h5(dDvv=LnuyF;mg5N6X#>e@9i9bFTzE
zBKG5s-SCRlB+~avo(OW886L0|chj-v>2RrSgM(;wm`52)OBhF#XPVKf1OM&c3`w73
z(x>!~q_W3DUT*z%Z44x<ZyDHVK9s2T8{RrV#ZgT^*sIc$i%iV)ca8vw6geI=eR-f4
zwUCkdUwVfQAZOz%@m&Xf*-ANNcB<xTV0|2TD1H9?0KK1U+(IF0(equM5R+2hm$rxQ
zy(0~Xr*vD?)&b}0*ZWYe0f2vjDhw=;ngfFc^5O!!djjv`cY(3XH$){0!1&Ou+(4NH
z$T*n2y?g)W7b{K*DK~|l16QU5KE(F4$|oQ~kJ&K5*&MzCwe3ZVg%g3lZ-AiCg=w+2
zFvu<GHnQ&=F#Md6IBk=Au-bEA4-{kgM%-=N;7X9?UVn;pLZE?n_h|)p51RbT0a62F
zLX#H3<YJ>RwX-?zR#j)tPM{Zc^YHNK0zR7m+Pg!(06S5z%H@Ctf~6zy_*oA{Wm1}6
zD8CtcKsVJ@&CVq-R$cS5M$=c2dV3M<w2^lMFg$&`R;B0a%jdAnBRgOC>OIe1F1OlZ
z{K9=?4_wMGcd_?WZHt#M)g~7;6wowsSJ$p>U2#8H{a~_@D?Rh!c)%g2@Qah*ri-&;
zlxLO4(1pL7zfWjL>mT?j=>@Uh`Z0#|Q^RT<#>J2Ijs&aIFFmclY$C$OVe|zd&FqKc
z=-}$#I}&#a%Wq{*(uN!pR_enkA)BtOh{Iixj5V2S44*7B_lJeQda_PSl;6ZYldTO;
zdAk<nl55#wE3CVBqPOz?N?)1ccYo2X+EbygA*a=Zs&W@g{?W{K-+WiQXL}92;J3T+
zDY|WywSTv|-5O%o8HYTr=B71b3#G$kz#$t&jb35EMc!PAMM77^A*V+>L4>7OYITQ4
zBw^`ftko?<{Dylw<_+T18ej&F$3N=%1*&_|rD!CKAL`eUVgHVKJ(6J5Jd<+1!e*O@
zoqpo5dfJK9f=)#-^Fm1n>8QfTUEbCyBtW$F&7Pe%c=6s@i*TU5`5^|@c-%o6XG0{E
zSp;GLxc{FNe=HP^<sP%be@)uLxAD&Np7(mIaluZv*6<Z&)g*E6GY@i{ZKw@bUJk<_
zV7sTI)ivJ~xaIUN;cKdgb6=a|v#NZtWTD=eljp;(u5aDRr8@Ce`Wy$hzJWx!*sljC
zEPr?pe;6!cbTI^U`oZv9X=!o`lXrtOx<iWfX7#3TYl-9>0@tT@P5uoJl6q5@t=~nt
z5h3b1I^cVtDhFcFp32?!W4{!s&|0V&ruU2=+aJncpjG@jePBa_sw`mZ_)TfgY2t&d
z2WEsayO+I<?NKv@N}TxrAgn{u!OD8%z9H_V0kpb~2V#AS0{BdzvUQ$ZgN51<#*G#0
zi@(!2H*G`U9_20Ea+qeFNZO?9HZgy<TbpQeA*JHkE6O&Obl($7tJoi93I=(0h<)jC
z;giVT!O%zSeg~GQ+HL5RsHmtJCU>7wo{E+$HFBvo?!yZ0^%Uy^w(8hBh5^<Cg^l|U
z%ja?@>3E?(z(!CY82BIIg6E}nGkKh~Qni`INSZgHXBVmrIxI2jF1M=NtZ^nd0^$Q+
zsMl>YSxb5W;>nDtZ%ssmJKyY|SW@PZ%49y@J(tE1)h1GiYA?ErZFjocf(m$n>AYhQ
zOlEA`I+$nHAZ%u1vbrBZf=C{KrcfMwelpb)#R$4>o2PM`dAO@dY6s=!FL|_9Mv`kI
z_>hrp*o>gH2`ovvZ8w8-h7OWoSWP$)GG;ao<w>uyj3Ii9mOCWQT&UZ86)ZcKnlU&y
zc=8m+>@tf8+W{{cIr;LoiYs_^XS%;wjG%pM71?0;bgGIWE!*Jkb)U3`wZ~HcVbP{I
z<%0+93v+V;&^xvCD#?aCS4Zw|O6Sc+VX0|@jwM8>^F}4pcevAg+DY`P;hI+bPgT-;
zVdLYuSxtw$3?#nb3R6ny-K4&)CRxY5u&GO11ygUahG{IlHBD<R$jc$W*%0YY%_}Ty
zEetd-4z^QW20HJj(XN%Rzj6yoE->vi@;+73cRxWFAzwyPQ^%SWu;~ddi#*eE%#Wm&
z%dSadKiiy^N?$er6LUAWBCZ;3(KISQ#FZwzRRqi=LPJ#nK|#ySXT&g=8Wqc>=k1zN
zRe&~WW6XFSssd0CMeQAKRDSH)oN%!RnkR=p8dlk>220D!Ul>HV7-I14)Qnx<Y`e1q
zG<Xng=h+xVI`25?xd!%AVYqz&QGW!dFr_0f<hXoJvcGx$v(&y)1Xq|zb)1Dfz__&f
z_N4*!A+1;GtOwn|nd0%7d{R~jjA!#~M$Nn9hk?LH<S!bXGSzl=h;L|dl6c+{v{h92
z@!(b^7d>WI4KqwZrQtQsYu7NGymJXS*l;<iV;ID;fBX$3<PF#wES0=ud*)0*2jn^|
z@mJeKDjt?JGtS+)6Crx323Vad#+8Pv_W-ew&vsCN-qeXM(meP`Fgvk+N{Kn>&%W&5
z#YVkx-is<B$y5Q$xyycl90@37jovmpOt}Wz76_=LN(@R|yz_Ky_G5;r?A|7h{&leW
zmq*9*O5dqv`G+RY8`6e^seWF&_#staaa8{MRe-f%a{o%&_Uq_DR?#bNX1wZ!0DAkg
z+JnQI_%?|5!;uH39V%3~U6sCa7$<~qXahV<j4WIUHW!5#;5H}aH<cU*>_6X&XJ<E+
z{ju~Pyl=IRw*9rdo0iE~(PQ}+V%6d`{vzj1pYdydpLoh$938UbYKaa3I1s=M<e*Zc
zq2;=um8<qPLt2bA6;6D0%nDOS8HL@8W>mj9G0#>=?DN*<{ng73WIV<HqiDn1$SZ(&
zBfi(s#M{DpH=O?ju5qJmN#5C;%q;_*pVB$su}$tokRxN4gCL^Rjg0z~ua>y~C-DZI
z&Z2qH#07I`|B(*XM<J1S;S^O5^}h|^KMDWjJ3!D>OG^tNjHuK=Vy-n|gQh01L2D}%
zQbu;Msu5UeE*5wGmla<JB16fuDd=Q2A{{;j4E;h&Z}J{+r7aiE)myEyjI%XGG_dh;
zt`{v=zm!fO^h%Cl1YG=5FJ%ZQZU?DVP+M0|Gi9`#=_`win1U}MTzyNlR(>lCNakAm
zzLzb=-;BtA#h%i0n2Z3P=}mhc>gTeiCsy&i?CHf<CIdq3Ux)Uf({F^B(t368mtenJ
z2G96q=Mi1Vw_k6tN78IuIcy2o8c$<2c>DX!JT}uNqvVt!pUk9rl!VH;C;@2rI``7y
zl{e&DReWDo^~nj!=51N_`PV|b;#l@YY;Qg5#c!rb)|mU^!ZsaJr>dH_1Lwu3L%$Be
zt&Km*URvqR_{4ZCKr8x3GnhfdW~#O-k&5j}5!q8Nb`+QTYxu%DLeo};rZcn4$g1pr
zH~8?w0`FZfSU<?<A<h!{?T>#6e-7C&E$$1HjRVW4++Mr((QB@rIdFH*m8Ju*h<O1b
z55@8<RLkUaEfy&Dpwe&&n?**7ua^^}see4((9RJncb}^0j>*gEYO6j(^78}iw!KDD
zMzL)yKwkuICemoS`OTW48s%UNzt_pw(}VLuHB|tjJ>HMq(#=1Oo76OFZ}-5C;$pDK
zu*DGaM%{d(1;Z)6dI~vt!My3=5s{dH935=d1sUEp%?ANCI&9h}A<xlS9j~&ou30}p
zWwl#*8vBU5+9N(NHs4}a@OhueD!-*M40$0noGkkFj1Tt2s#dDpn?pq?vwH^$&1{=J
ziBL)6dr$i$E|0f3t=B()%NqZ{Z;bSF!=W#_|6&K<R*|Oj={mk3=644eIi*UcbVXY?
zru}PTbANSR49}%aYKjYeL_}MA`~%-vL8Z{%Pefh~3|r4n={x?1PwwAu@jguStHT8u
zTY|ur&z)nq^50iTsgJW`7o&Sjdw)3Rw<l$>?yhL(z_3%mEdt)~$u+&-E#*lJ)?c=5
zMuN`n+qEYTr@uS8Qf5O&;}G1NX$Ah*of*{bgrO$|#-CCx$3@+cC}-`PACL@)58m6+
zFLgga#n>zTea$1)A`^cWhD=Cf15K|#x-KhL`rXLi=2+UcmCGAS_&{oPKUaxvF(-Q?
zxVz9~m*Kp$;Zl|P?mhDU@IQC}GnTLuThD46cy8?BVf`=VO@(LC<yJQS_dy5g4GyYs
z)_3aFHRQ{IsITLW>fqaLJ3mw5+XMRH$BbOA#E$Q0lLMAlV^Cu~k1c$CcJxhU^XM5&
zVJ3MZY|{*v!{x=teA!+0<r&QKQw4b3pGG~T*0OoT_q2P!i3!fiQ911qFvO}?SVqum
z%I$LKbG>iXpSOR?AR%YNe`4u<cWnB$D0`>Bu*48A$K#_kcya5Z`p7EzGiYs}VB#a1
zvEr<`GZ$Q_62|`Y!9(96xPr?1aH8}5?5VJ~N~7&~9UhY*{V<%rQto&Z8Bt`C#g+2C
z$Fw$QspO`T-{bVU>tAX$F~NVQ4T`8f465lpu%BOEH>TS-IHk-4x$Nu%HPC8JHXd5@
zaw0Uf=)2!fnL3Uq7v!xbw!LJt#zRTUf8}p0^Zj!tQURKtOB=1u;Vc2{pMdHGC>`89
z_RZ$G?oq64<Cg5hrT+Lrwi+6(@j(E5Qd?KIWy#*}5AF*BWJf_{d=K_CkF3`W3B6_8
z%!+3V4`3GVeUc{T;_dxE2IFmzpCo@?N}z3SH?R+EhSg@^0PFg2Qi2tJ{Xp=HqBxh(
zcG3zEiecO3c#KEcYEvS9XvmQI_K!AzoKKN$4I##^6<!0Sm3&Y7{s?_h<mI*Jppar&
zfiEi;d-+-tKU+oo7abw|q?DV6;o_3D$zm8G5yBFlQ#(^kJtR;2&EOJk33NguZ*t@&
z!}n^?oga%;k4xHKdvsAB74)<1>nF>5HN6gG(l=s?I!PBU%#N7aKW4wO$-63W*|^2r
zKj3Gp-<{(p$9A)TZ>Ti6j-)k>O{F?sri{=o<Qhrw=J2^Jwx3R%+7RbjomijIc|a%Y
z#8mlwIQ^MVgA1thp^HkgeeYSWd|?s4bh9z__xfVZw(7M9eSO0!hymbrJO%vqx&V$}
z)nsGJwZt*$84=wtI$k283oC?!dhE{5&Ks-QdarMhO)twiB0X9brmOzzGoP=XmH<Rh
z0_#cN%;c}}?bX}e)&7G7lg%W&LEM}`hf_LrQIbuujVpl-G_n{+|MqR6Jm*Tp;E9jB
zyLdQ}wFl1)fCbb1fJThB?5}F*%BnIuvn@F6aN>WCBKoSGLH!wNK~pusP!HvbU@0ET
zVsS-xXIdI*RfiGN7BURSo*x0SwfSk+R^Wx1s<24_nB@Taw6?5sUh-ZPzU@62VNF{v
zWYTQ^tXv!wTiveBBQ%dA{&QxkYy<aw@`6SV0fr%fSCN#F-+HVk`}OX=Z%Lm9OP&Js
z(VCpzen9_V$$34PrL@$It4gTBJUI1A;JdW8tIY>?`Zb#56`kh0yc@yB1j1yPxKGz{
zZ*r)(62oi97ogK@9$v*RaNzyyfWodPpF$Nb>p2Dq^3HwgEuX)bQEDWtJo=^}{9xr;
zO+w?-#+2ZR=fVB;G=jsqeP(>cGn;9n9%QB8MIBQLA$(kk4LKwG4{Ua#a|rqA!K2XN
z&ch$;fw|ZPekv=_nu?Cv&PAqN8HVqFj7_g(WH8asKfIA=(vTcG!*c1nID(pQvSSMf
zd2*k?#YeyGIepl+wM&RqdeiSD$1WBRX#MH~abSSCmKSP7wQs8Q1=^cpU{`-Dk6ph4
z$8$c;h&r+gw=%QfNx*Iy_4K~sP1&sJ=z*hOqqMrsAFg0#Re9Q2U`L0=-m53#=cVx}
zZB@FiM;%Z={}0g9lF@4Q9NrBCH-HQF3W^!}$^5zokmBwsVWcUjD12b+D`s4a_bnKO
zVjHJ%CT@fpA^h|N=%<w}EjH3?jt}Aw%eZN-VM3h~WrGXqMBa>4r|2fsX80Zz`F?qE
z>8hZ(35I!ajJv3w=IZ5J^$RmwxhUR|`ZVywN~PLWzg11mmub7i)5AL;z<3O#;H$<n
zrLE7WiA5~Fg9`rn`c_(~CQT&@$SDtuFySy|xK*{R5|1(<OH0<J`&nrn@9|Cbs#RoH
zmC63}v<xm1xPg5jotSxgU>gR6yKv-90@U%pyR$f=yN?Ldlcop3CdaLkAM_ts`aXLQ
z?9?vUVIm6naAlbw2#wTMHM6I>ZPl_Fb9Ya)CRcuKz1A-m=rO*={7O6ahC%Q9RO2DR
z`w_=-4!^Wj#QeVo5vFTkT0|eE#D9x^AXG~=>*I)dUvU)Z0CDvbbD!FZvEJCMME3xD
zR{E1$A0NZN^?g|S5{Zd94)ZDMLzez)Q5+685V76BfkiwkOWF-zeIQwce=Ts9fNbKi
z(C+^??*r;*3OSFEh{MbSeJn;Xunr*DHrdv(q4(Os2iWcFiKzvlbo(UJwjh{VTN^X3
zUN+&JK#}ftlR-|M1G&ze&Hk2$3a0p{c1!E)-<CZ2a>ECYm#EdU-{}^p>K48|{L8*u
zvvyLs{M#u~+ainu5+FzcI(<na`gF6b8HUoXc7>ryz#q=T({rLpaGAZgo#{Hw>91hz
z@C}any>ZCy=i&4k&h_n(BSiIUb)7|QV}{@Bfq<u(FBmrja;>FAlHMS#4zG3{r_v2y
z8VT!do~7m1k+eb8T`-l7mkRzX`)K#%9pMg#AbMMe+hK?#)yuk4d|lqLsyo-?rzD5P
zc4!8&N;dnO8V~G4<ap}!yCjKjqMzi1mfr<B1oP<c`mOQnE#eSwGiRmOf4=RJW)UTv
zqd3JxA1kpH<6}RVe@2Gr{(#6&p^Gw#x8zMQbp+S<>uGp-%y(Yy1~<F0$HJG&Z7!6;
z<IUnO+d7?^(5GE%q4k+OSIBJ<Yu&lde7>i;ia#qmg<$zGDU_~YBFFLeAz+OWLwAbz
zB(~jVW;U{U+06)Do*=2B{p4s{zB#D#8)hsVwmc?RBHE;PhfeTMQ@FZqL(n_xho1vV
z+_je}zckm{m8icNfWiSkBnpd?J^G2Uv#-E2F}uEbo+6Mb`iI&mze>@6Hph*x;DH@3
ztpE^-<-1pGDMmeeImq(v^-EZJzZ88pj3=$cZ`(j-n1X0_(FuK>q}1DJL}VXmSB}ZZ
zu3Zr9)PINm*WLL|tJa=KtFV<Tc5SK7HQg(*ZHKn9rCX`)xuX2)TJ7RlY=3%5U+#<6
zXD?_+-b%W@YFUgMdvarD_5M@OSF`isj$~u2s!*>`bw+>I4Y84#%*lAjVxZ_u`gJC}
z<s&ZhrhCCH?7u|{DhJ*>CT@gCeHLBzRMoS_`>1D?lnZd(JC>7wa3V>Rgv%`dny{k1
z`Tlp^AIL2si#v35DfCCRmFMYjW76mD`7RSXKLxi%cXhK}w+!^wp}oVO6kbXoKn+|M
z`^s57Wjl?;9E}_o6?y#l5c5A3QuyO3{@Q3}T*LhYZ|Jf}_*sVQQ^X8mdspmixvf$9
z4DQ{^jou#zH$&^0H4eJVYQ!_P)-BJgaC&mKcSFR|;doIn?W}mWM8nt{2Kn`wOse<&
z2p-`K$xAwq2Y|(`oyIAN|GLzpQaE1Gn#J7eGSdqUVd>?9U5J{oSGAJ`A_<=bnDWIv
zm|aOz>ubJ1gU`}$0MrbTg*5sc{jJY)shHnh5~ACYh}vQHv}s<uz3TWdm;=j+zr=o9
zh%aL*5^O4zWTjr%+^SWuqM7W!l9mJMmBQNgehWSKwGIJoya!s<H;wvXr39#fv!4j#
z@d|MsHNv?}wiYg@RK%xVn|wD&cV#~{QYz<0ywuOJiQo6f)2Xh}DL<@_H(1Lbm>J7x
zoWXkX%O=Hg9J^irAP4*GI>?6vx4gV2G@r=)RPmd*qvCBL?DsBqnOV)p`7Rq?g1Rao
zkW{PAGO@Weo>6@?;c2rCd2ouq_3(NNsl=tvJ#@N;sLmxb|Be5LrJrnU5O!6QoUrmu
zMvfJx5njdyTY@9I0MEHlyvstcVnt$RW#DK^N(&c01vCj|XJ)<W&iT~v4Q;K!Co0n{
z9i2WC9Yb+#KI%?+5MZBx|K!PWi~IAVBBZ6|<gG~ET}+&k#`Bu7M_#+#_9GW+<<Bm*
zUi~5{`N@+-O|>2Z|0!;W&uT39`vc!<Fo1+|H){{aN6}rBK+yl-rnxOIyQ|*&+Lt3e
z<@j*|=bb*fm0wB8%U{tNp$S4AN?Qj3p{jDB1-0&@nwCqG&TBP&0S`VVS%ru|KQBA{
zL05MKG`eK2_{^5<4{zApI`jSAI=n4%E}6f4G#xxygq$QOy}y!W-=7$N?X@dv*<ZPX
z{&N2x@N9Knm*d47It~&*wz^zEVBoNgH2Y}mp)Kf92+0<e6O-|$8~H_z*`-g3f74E2
zaM|yvd<j>upe+2pW_+T_YVIGooVa?<XXL%C;?#(&8AEwWUQFpnFRed+jTHG8PM>ud
zfx^OLBH@Dvy?{|-Lae^aeX-<Q=Fo$8V+z6DQoo-n1HJ)(esCDgg<^!---O|}0<ap;
zF)uS_$iZst`Qjia-kTETBw?W(Au+N%rGH`kX0TQ`M-joz^{M^~;1oKnj_9W18ldfm
zwuk$UvdVQu7TPn^YCuR7r}yxRM7o|kFf7E>L)7|iKZ_(Nd5BqMNV3E8`0aX8@*6_x
za|-z`UDu{K2Xs1j{2N+F-|TY#_G3DJpuFV!FyKbp){`Wx$94%iw^;0vLu+ZN_JeXG
z-OI^etmc%INfSz(fQe=M*X7j4am}>Mp9i(zuLiqwpl7(=t&-YJC(k}C6u9rXd5vQo
zRa3buZ#v5fGY9X8PPDh?`?Y@bl4(Y_z6{yiG!U|2qr1tesinKx(7vH)tVpMzI<Vo(
zA=*Cu=|>Y%cBTBmCx8iB6hGrSbuP7s$xryD+r8-iV#rEp70VC@4EcZu?BS<>Mz`17
zf$iMp$}7$TtE=YFaj*Ba0d9Yq_+L|po23(ua&mIIguTrmKi9hhI{)fGH=WZ00l)5H
zHLGr9n@M*JuX<?bs*Y_5M!40FB`Rqi$c0JrL}KOvsoB<kxsv4$>IdxVE8ru>8!nHT
zMXSZ~o{t85r<?TpRL;u_o<UK<-euy_7ZvETWd()8jkC1m>+YOLyjUS|sHV|#cIdaE
zkxpq5x(6fJJJffaid6umC#CsyK4E`3{_lp6Zr~ZK-$om19S|UDX*q;e^w+2OUktGg
zccuQ=XY?cgHZ>z@A7Xs<?Y=k1<IPJ+p^{N5dj-jvb^2RG`9*6Jp_Z|YpCfDI=R8K5
z$E-<-%*=eo7o^~mCDM=o1pxs)J@9)fuY(){)0fQ5&sT!v7I9$%tv<*o;vKlOUf9Nz
z2(gt$cLD5|#cw~DCcypF<wAP~$NW9ZE)YV%&Mw#NzuowKS{FT)LeWRAmL_3Er_kE+
zArti*ew3*TK92#5d(KF}kX6Xc7wtAFeU=1Fl4IEmH;}C-ZvSLQ3hF8)sXKNNq8a7R
zm$<Cux%ABJE7FzP59wXpX$pWz{%qPJ6F*|X4<f&wxOb$n@1eNqYzffSo{uEp&OVp}
zrd9v0lr+eH|AxW$(%2O!4`6x)ENA}>AfP)y0_r~#iDt^C8tr6Jab60T+y(ECp_kFV
zOHJhlqXo~-pw>#Z!=SF)5Y#v=$?Pt?Yd!gEhfgQ!f^~A(k(`U730R`DC+EZnJ#9R7
zaQRQHa)cN@ouE4nc`5rKpH2DL<nE!RGgBAiQ3)qKWfjIqlVRMiM(cqfh#S%B>9V#Y
z=@GyqrvtSgG%hk;d?7zd15Tx4S3o=E6SS&_j{YQOk@;ec8$9a9o<#g;YqRa)uJ^dt
zpg7R)AGYQ&vGW!FD0d56|Ha{>bqwjhv|8)q>Z)0fJQHyrJ8HdmP6LbkmS?FSq}lCx
zt$`w-CHQINQ%$9r^Oe`%sYNC?E(hejZ+Y<F+r84#U~*0RyQcLUaZ*m=2Q5uaoQUHA
zm(6-QFH3%wBAtiw8{L|I@7ymO53#Of$U1Gmt$>K<iM!J!mx*DB6s)Pkc^+NitL<vD
zixrstQWYBvskeBT@fV<E16(O&vIjTwpEh7M06}q~;*q_3&~tpr8g^R)t2~2B5WiI1
zLUYg6hh!V}&EJ-g;^ijh;niWParOH#e*Dv`D#H}_f>%<WLd%A9=*R6sl7t69@<j}v
z8LAVj49%TN<9t4`!DD`rYRU-2-gm){Hs4QOu8}<~d{8sLKb2Dla3XCFkea^7m;}AD
z=*Z*J`z|wIZ>LD7XLe~Zp`&0q;pA<|Ft$mzjLxf+G60uU{yD8l$CN5u_aO_A_$uRt
zssE$j^VqUjGp6;i_L@Z*b$#NYV}4@)?Y$n)6`{|D4ENhslY~3XqlbfXpXBL9w8N>Z
z)^zCtuj>A;8nRYw_+VN7srZW91z^6rPxD@TL10F1+}hlna-_F1!Fjj3Wy$Z#RV9(X
zak?P>A#P>KM}W45Q<(a>xj8TOcQ`w69{|>iOqPM$5B?t``5KPfQ&`bcez`?Y?Odoi
zlO3?n;%(r><21~>>3Eg*q2!r<bMk@QN)a30&!$oKgf#2Gl75*WFJNjr^2f#M>>!)&
zRWG0T1;ci8Anb3)z3fY0pPF9IvFO(hn=kSMKy3i?FUhjo41YG)f7u#tD6p)3*IlD6
z?5m^zF7H}0Oi{_hl$OJK*ol>EQ@hxb@^B9mGQRQ3Qb>L0aI4Nw&zW3(^agIwH62WQ
z_@CW`HJU9YC?F|{ySU#;JO5#Oy=P*R&8wViIOR4qNU%b+jpu_%zrB!+)?x#slRg8(
zMKk!a&d$-Pjq6UDK(4Z9ys~>f$ErI-aB_auP)CckMKwfhWYJG+87U_zq%->^I9H`5
z9F}=e@0~DsgQnd?oumHh4a`8QcJ_hDy*OftqwYQwghZRSTPAQ;b=WZhl92@%u9~gU
ztcVnR_w5PVMS2-x#Vf+t@d8y{3kr-XJV)bG_*-|9T7q!@?4Xm($ph8KblSe!X0nDJ
zwk1Ob6;FUQx8Kt!aM-s9nO3W}eNng@&r<K$zxULVPv)4?3EP^z*pT?RQ!?e!>~gTe
zo6(ZAP1VTjw&8y)sXm)iWu*q!%my0;i)<v^&+o0Ab^iFGc<XW!omA2LXai*Wrmqph
zbx~6*`lr2imE~uN7r0L!`F=HZpWhNxc~bqBrJGJlvJr$iW>PzEPihVNAUXJF=(1gv
zWS;;}(D&zGtIg?-yBn*+M_|NeS*Nf(syGwg5ZjQ2SS20do);FIGz>)wFD_CIi8b3d
z{-;%s%ux0{=;(vGiDd1HWJXnfxncdJNpm<+warF!FAqKvt<+oX@f<6y<{K!+YEyXQ
z>LC4<PdBW;t}hPQh%7X+y^V_Ye9q47Lfe@{>zYwOkaS{4K<`>&R9&RABr<3PGlDbg
z2aGTF3ti}H6&a~OXBn53R_7z7^1A>5I^=_NQfX;*xw{@ET%)$OmP7w0fNDK|>UCaT
zDpq0BbPM=$9G2BhfYtl`1B1WcR=o4bD6G9zs2#HrS_%Uil#W)v39)aM{*4xYB4sjE
zR+Op3X=1kF1}^ClZqDNkYK#7jXLqiT7My*b+S=(&Oo1b)!fAh&4j+cI_8FKf81<O5
zC?zTF`<PE0KS>CZd|7YS&#lUVrC~X^+GJ>&RjR|98Qx~=eWf^O9ejV)8hrWttYb$K
zbaEI|_c|crNz1Qr59&Oj*;Q<ZBE!$3q1!QuN=2u;P0}GPU_cx~nR_#&1q0LPU(3{s
zjkFYcw&avKd84}di+Gc*eO}<hjjY-#b;xyi^z&as{(5*Flo|BkL}|<SQfFjA$c+~p
zyDB3^Dpmdu5DJj#JCEuueJ#gu2EmzxKS6OkAZ&0(m6PyYl$&F8EZMHbv*dbv(6msa
z9$Aq00RQ{V=!5o6+4gW`sTk0n{<9hcx&uJ$JXNhtt3A46F#K>byie7(J;RgAW%e(8
z?*2Emi`~m<{U2w#6>uhC!l+XmHuLb*du3(2*3+OqC%D1f|L6=Z@$4S8z0iEGOG5w;
z#I4zXH=0n=y5SeXU>yQY>&v%}W87s<P)eAvSM`&S7cM;+6q4K!kWwr~Y%JYCTUM;6
zN!@7gy{l1*_P2ZDl|%F-6CwJ+{U^gtZ+*>2!x+SDkbxV>8ioiH0#$Qe-4x#!N#$R6
zI^^wHURqonz$G4Nmanb9buBDQc;YGQuKMn76OMb(=@J9@I^UIhMl=US0;13>CxAtw
z^5O?k#?hsyPh@O}KnJ(xyFQ>NyUtb}5yvS(7iPw1Sk`2v?9&_b9a&Aa<>}QgeUh`5
zZi;j73%cvz*`6-H=b4c=4CmC-S2N4`f*{JS47L{Yo06|ymm?nQXmKMS2_e@qbb222
z2L=6c-l%t6Dd<?$&*bu}oF8Y4KZ8@5KFHnw&3awL<y}RB-cOf2OE`Dt*d?i|6TRp%
z`LZv`J_5hc%hzm<M`w4LS4jUy(pksT{r_!Tebw~x=;>ydW=x-Bx;faK<1pRbo0@ho
zc}%yB=^n#TM{G=+?(W+C@w@-`cpT?^;{A&2dS1!SXU@Zdn_0(Kga;?6ro`$TI{y6R
z9I=_6^6PnaR=Q4KA3n*!;nV2z=6?>$;+VLDRmOiqj*|%>Li;2kji)$dmA_{7He=ko
z;)%+NNTA6S8KDo_+tj&M3wpc86+4tvEp+QY=;QEXJum1xg>hN5YUGxtSN)0=GdW&O
z`-f-MI&ajWqxW)daa!k%<r4c(n~sf-^7o{>f|?02>BG6Fw03(Qu0~7JT-pReX;j0o
z#M9IusJ>@vG(s=~yN6`M%64BE%!OKw=c8UpJW6nm7JKzzlnUoqI~%Et0;bn^styc=
zH0fbR&31q2LAKIeL56(CD(e4*Ofjr9^7~7eKki6tEPMQL2?D1?SR@Da46rw>)i}`3
z1mVHMUAkC}G*&33<A(F<uCmDduj@4Uq`j9E;_UT(31y|7)5dHLUaT)KTrScRtPQZS
z6i76-Y~6SS2$J{Gvzyub8HkhpGEHfz$QRFx^pP2NPji5Y<BZtbR1^^386X6#KR8eu
zY_w>2QVd|MaCJXFid9Rvx{5n$>PVq5NCDd(r@b2{Nv>yKu6Q6a2U!tfMn|R>2!Kx2
zQkj;Y85V7g8Ru6HNN-3*h&^b@mmRRMzSZ7c+)OYf`%6PpwzYEtCd;&&`k5TIrK=#l
z%8}t5rT=?bP?uU^N12UG=J$F{MtKX#-)prXmtSjgBmU^fptLKM8gB;6DAJq+qzk(R
zv_k)Mp#4u=qPX{?(yg6GuUN5-(IHoFmUV2GGv&Zqjfo6#z3z>I2Y#c8^%^HH#1flA
zmBQKWj$Uq(NzFb;->g#mt3x25^r2_|T4!<&(&uL;D*alR$LODsR+wr=eF7FAq&zgZ
zt~`~2?Nm?S48OaC9oi0nwC}a9fH%9R68mtbw^d5Fe9~UtCNf59#qq9a5b?sleCE;&
zhuo~P|4_n<O)2tRJ#>0|kq7ehNKEq*+!-H{m@t?~-nL^*RG{<MN{CV<lWyd{CoQ~w
zZ3=OdyBNetl0h`GXz?oBmP@cU4j)3SjBgUj_v)_jox27O;!qnO2N(1?mohR!sUr!$
zXj<W_F*(j9nW*W`_Ui#}G{|wv+=xM9%yHY*(_HeEf89X03tFd+ljgw<YKZDr{VY+Z
zyA`LS`IBGO36q@6d;P!7VTCRIy>pN8?EQih_J-DveIRV+!YhtJdUvZXMhv%fqf+D@
zm))w7`=<!=(d5dYw<zAy7(ab`Ef9`E8%u#>h#;?!#GmKILqS;A&BF2?c^lprehNYd
zS{UE!;@;nrzyf1*-(Et`nSY8Uu6>pCl$CJHYnnETeL1p2jvqD90U8){HGSut<}AIk
zjM_bWv6h5~LEZ9gM((6V_qdu&J#axQH2(dN@i3SN`b<nA^A!>QL#JUzF%JLT{>kiL
z327POSxGOme<Zkzv`^X~6kO4{`74kfa0zLvU_{JmnVL7PZY(Zu2umeH4V}HCGxEpp
zn!Us#rc*&d@xIygN)rDL&R4BZntFq|;VZbA!j~P3(oCmg$5S=4r(bNRc9eHdY>m>k
zN8V7k^2wB8W7<4fmx2Uxy>%VyOqV9Xg8YoVpq48w;z+4UBl;Cqn_n-<L-L;{(2I@A
zf|#)i-adlf@x(lysa)xdGcvE%@Zw))2<CRz&!BC1JL*_@xScJlUnp7aF-EiO@E-II
zTdXJ<8!fSC1j%*~o_Y$<g-8!Nu72~4?Y+=LegfJCKsgC?X|_hcQdZRT`vQL^z)ehA
z8EOLb`+H(elvUeP`<J~iDaM9Hp-lKQh~1&=BIRf}a?E&LjAL{)l}L<|JLZ+vGa41$
z4GvyQH4%%(woQD_ZIY+>fw8C8YK*vq{k>Aqo0B@LJ;kPfwKpBXOZ~5wb_!&<#y1lr
zO`_nQp7o=ZiEmSXoXG*}1)5VyN3`<lDF{MDQj0d6K$I4(%(MS_9%EK{9$MeoD^HV-
z@#+_ZW>;_XC_`}WvFY*?h24T_r-w_o?EgW*NKqW}Y|7`1%QMzBi}ES!76o{({bv;v
zn19T7zU|RTWi`)s%PCX<S^cBJOTG&7;||s{*G9Xg(OA}nM!H?%anRIH9RopX_~c`B
zI9z0!%)_!}lU-M*Z~Rp*S)4S@upfy!<Q^Uzd`<acR^|P=a&mh)SrXGn{vWf|QQ{{1
zmP*KSbk*HVFxvo(3;lTD;AZOOD*H%|XeICcn9&{Stt3WMay|bHCy0mR+lTLv@87XQ
z#u!hUs)K*t6v&}|4|!Y-n()C<1*G_)z^kELBlZ^!zj%)D=Y|-S;CSHjxtld5U*C=s
zdV!wjI`b-0kQ8eam$O<Ng^d^j+u_fc_4Q#ViefF=k6X#T^7ch0EqJhLEGruU)qolq
z^c7e08XD06-XlA$gDrRERnIM!X?XE89udM4@y#t$$~k2|oE*dTrDO*Y;1v*JKuR>2
z&%B-OD_RwA;9=I@N7(gm`fLwXvtwhLc-Wl?R>|F9C3idSkUJJ#5={WXZ==Sm-5~u*
z?=#}#tDToZ7V~@rFGzlQr160`yz{L8C?*|g?nBnSa^B%fT(<0u&%CDJdUcJo{3tun
z>ybe#&}erpU6g_G8cdNnA#9L-j_G6H)K5n7(tVq|8OC}(ZcHCeC2?kJ>TrPl*ja?x
z=|t({->kJ$aUm}ZbRDJ^xmGT%`;0XJ*%3-pyyNXIHtM7>zS_I^1Mnx)EJv)Br{Ova
zJnII)km8>FG13Hi>LDy&N2<f36Q)b%sPa?4k#Yf&FCl<By&J0!^&t=Q5=c`6<ZaFK
zqkjj4(zK7$J{Q?}f4cpQR>BvoULX@bSwO@&d{FRIEldl?)TjmXRm}NH`s-KGiRyOa
z;7}suYH;F+HS=yBk`$Ys{&Q)xvwO%|ymEwNZcr{2+L_~=>E1on7FY|(nUwa{uDrrL
ze4rTK|1islL2|!F#uazf--PQ^&o4MSrWyBRT?>e=Nx60Xd-JRWlr?9_*YB;K!tor{
zo+7f-ld*y**sJ@c8w1VQVY+VPyfdRby_L`Jq;16B-l*Y2>DiBKFZ4i|&tS%CJf4}B
zFTbAK%T~B8u)nD^H)Wf~^6aLjCvXp|;!b|VN}#zm?~ViQn|N^(J1rQhP5pi9w(d&m
z(oT$fA?D*Ggaus#YMH+Wr%FhGnODSh4AL>{r#8W44znYhZ{oS+NBaV#UM7diqWDEO
z>g$5<qm^=g#FQdu?L<BSt_}&~3<x(X*5WDR)_Ccq_u~0W+cmpaI<rX8xpRveyACVM
zpZ^ehR`UqW&)iyc9(lslG4E_`%S5u>{&Nky2GAp<XM}j&NmF(c5ph!h!WYn6x*}Y3
ze{cHK!jx5Sxn)#|9335daml4L2evvp<ytP5KAn=$_~fhvJQ3JCRj<dsW>NIx^afcQ
z(te<z@V?ub4K8O8%=se4Di6&aN4XmzZ!U*lg1kdy56ya--ypsPY&q?P{AE$7=3nHr
zL<O^<Ib$U1v&#Zmd#C%$BpUD_-e(1o5ykab%|q`3$IE`kdF+j>shU?L&C(xvJ_U%}
zAZO`xY?*d-8rhCFK86hNl6%tQ#hN&kQhjQkv14?lsyEEG>k%GxTQpmMkh8rS{3^#b
zPw(=C)4ptZkS(u5Hz4wiid@xK$u{h@(9?&((Hm#2n`z&E{rmY&i`y3TxqLSrN(F;`
zFWv#J3-zPRas`x#r)vrMRQBM3>quX`^kMhIdz>Qxx7rh*LQkvS!8r5EcmqqK>qM+v
zwDgL$lYrZFV7k?Ri$7TUm0{0jvNCu8;g?iMP^qZke^V3M3eXLKqZgkimEZiHn(MV$
zkV?(3e*BH6>6Z*|v(r>rX{k@+;PvZed3a>t>|eR5p%B%%XYQv}aR{8DK`-B_t9^(-
zc^v$E<7Y??@|40bQvbAB%Itl3-FB2H*sn8dLod)->f8rm@;3SEwk9ILn>X3q%wY0o
zfaV{U#&Y%-vnhSg_@|hH>2c!i%Hx$m3xf35N5&$~oUG^okuNNcGl`)QJKkFq^~d%l
zkBu988EqRrGaFd5{o)F@ogQ58?@gs`k-LHwxhV>te?SnnJvgmh)oQiRd49cr@=3jN
zxhdn#LV%wDw2Sqs<rosAdg*nC`%wBdo_~HqM7F!YS?7L<)Y#uBK&M$@lbT2qKGqQ#
zaj*Gr`FFctRi9bH_Zx+7A7CWN4Sn$sxnZ9-n$$zw%ZPd?b~1(Ao0okkBtgj2VsEUk
z*@}ZL*s&rynP$WDomVHyS&X)))^?J*de@JQZj8KbJLFdJ(v3p3X|`ZbsR=?3_cu&5
zAJaWDb+NN;X{>x6#K`-;x!o#La8|_1bh3;WfF-Mb2j(pk{#){M?EFBY`1`amWm!Ix
zx`(a87grj02@F?Gk)KvM88p5=F3$B&1ArBNezf=6k(53#gsH&RUc_SVaH!haL}(-x
zZKeOJ7Yoil9<v=37D2X{LkxY(jXuzNEB27+8Mqrz{5AMpx&R)mTIY6HhvLI+3-4iC
z$A^=?H6ryxE;X?$(*B;o4>>(KO<R9!)D-i1YK1A3e7!Jl=M>Y;)&4Q8<CA<b_H0aZ
zDIDW>Q|^QFV63kWp&}#1Wp9%u_qC?x;j2@fSC4_v(N_qVDQ9zYYKbjYlPN58LMP+K
z!QTGo#DxBRve#eZovy@n<<n8wP3C1;PkzJ+NrZd)irl*_8D?Z+A!V*}{<lGE_946y
zG9KCyiNF&naw?tc=_~ea1VLIkW}9t;yH;zmhA61=`eNBJrG4+2^T<psi;7Hh8_OKo
z*C`ECD|!<A(`J6;!h&2P>sji`NH7z&&}n-1`Na#?elOLk)Qg+n8|(yrzZ!Qe&#L0O
zqGhIaQj)O0xNC1~f5g%G)%XtY<1yeQxVH;7E2JrRQ|LNzhQuHTTpS(uXVrdaHHQ*2
zslRN4&Ax;HB=X|qR3RfX@~WEBVW!rZAjLD&6b<-Ar^rguutA?0byw2uueylB(UjaJ
z%JC)gn2@7VHR8uJ&ndor`!?$JHX7MD+ZNE`T;*98GjMHgKX|cW@%_UM1Z(`7iRK1S
zx~^}*v(j$h|BS-x6<3k?MR7^mxnoeCC#NR7fW@q<c5~67TFz4vo@62S{mf@*Rie^n
zyYct?--c$3S6?>@oJ&f_3y{s8`#k{n19*o|zkIX&CT;YK7D?3bl^ptUC5??J$I7cz
zRmkn&&Cb5v&bTJrOqAgHUY4b_8GEXpnA`wNO6Wan`dr@y74Bu<;opbYRpBlL5*yNe
z7gLV~_FZ@`wf{94gkc4<XP5c~gKDRO6+}cJSK(5t|Fp%8nyZi8%v;z9qr?IuK-k*&
zAN&U^<juLA`bapqr2!Ezuw@E+kbyHRVq4g0T%qp_I075|uRWF(x~Qo01inhSw=r0C
zN-ZCX(Ffif-gjL50f(PrZqeXJVkuq0dbyq3$Jpt$Hh@;^$tT5~{e9GOO|$^&82+3%
zECh{=!!$Sz{1J#F5JhV`1tl%?YY6k7=Zp{550{XT7ayM?T3lDx{NS(9_TCBI)w4qx
zAQI{2p3Y=r6JyVgYQCXgtC!<;F-M#BV;)eYi17=RMt!}<z`6%NN}c=qU8AXpdL9xd
zMXKSQIbSM}=Jb%VX!|c93~~lk&aMM8KruR>M(Hd<lmAZFtSMF1Rpg1ON1yMOz}bsr
z*FqG>gG|#6pUoTMNy>t6@u0%!`rO(9OI618-St2yU@`#QXn*?lB^0q^aC|{SQ!jfa
z2w6(~dir2ey1vjh!FWZZzmU;q`J36l2a6LYK?m=+&?+-8W!7b*THwBa^d~iNNCtcB
zyqm$oFWm8!n~LP<m~o!jb1^R`SB0uT4vBL^y=yvG!%9CKtth^g6f~|8C90Mpwo=mB
zq}~S-t~%v_Z(+N#ZVU);Z>OR|_+}%-ids%la0y~~F8lMDC?qH7KgrE@v$}$X_chv9
z$ZmsSb-$jjG<O^QMO^k}layJ-?WpAb5A(U7o&}P=!+EhD%KeH9pIR83#lbl`0-xg;
z&fyTw{AYda@c=MnG}de^Gv%utO-g<v0y8{6J9`%x*q-99F2>?Ye54E2KHK9R@M60h
zscDnijH+KL96kw(^Y@wXYxX)Tr=jH+%8PS%&P;_#2*KuMw%&M?5OMrs-@=W7X4Cyx
z|L?4YZulX#nfL2tjQ7yde?Fjmv<#aCOwx~}K{>klgCPpHtm<WQzl#9dor*fk`80ef
zs1KgtCV2pp28q4Ru5<gsTm(AJCW7HpfBu6`i-&Q<q92Q)@>0JHX#zo7UTbNH;oZFm
z>tXN*DTwfmKGFDl_p^!DUu^N!s1%SAlL>4Old0>wM;Q_RYunxP2ASAm2mME5BjR3K
zM``qrL%9vO%CueLGlF^9);S-mC<vRLfbXc-33QC4w^<0<(q_Mr*MJD-Cfr7Srf`VK
zbUAVPOoV@=b+SS6vPwVMuNb;#!-WjEARv*wEk+crli)X*V3eA^sWLn$F{jf#V=_s4
z{pR3C=5KDBtB>$-qgrCE;>l3~!pQtRq@lglQ!p)4@~vivrm}kKbU{FVQ7O_uFfVjl
zrRM50#<u!L)@%DZIG}yS&P+py(6sYa*)#HP??;n)C*DFCwru|^Ng)ZSLw1Y6{LUmM
zYHe+^XwgyDHxN8=>i&378;hfV5;0Cb7cI>0x5H-0AJq!J890$*LM+gzk_SV2dhUEU
zW_;fxtO*|seEB;Pr2jST=)s2#8ASfx@MHVgAKmgfv9Yn#?mB<Oj-Sxw2g~{3yf{%8
z!wBZEd^B9LNe%JL7D5I1wyMbpWSJczh1@s=Czvnay~iCW4fW?_%(0cpSaX%3=j6_3
z7g||K#)P!QMjUUKfp;Wv=)Av6S!m+(mxcPT632i4w!1~Jz*K~&`GcRD5RLE1T;-a`
zGR<L`7O+w~*x@vo3d@i*0xMyz)i4KeQ=W`$IcfG*g*&*rm4u<2_Bs)xs)#|EUY9E)
zec5e+*P|_l*q8mSQct<$C3LGd1mg?bcOK0ky%$^lHCo0AF1c6x73Hyr`pDppw>_Gy
zZE2pctu5gJtDZ}1u+6%WCU5b&htYiCK}if~y*0R;@8N^fQ=ZQJ0oy?Flr8x~rRn2r
zkrnn}N<^?TnA#60@nl|To<IGh*d@)Aey>wdhL-u+7c89_K7Y%QFICzwDLI3}U=Lu+
zlmC~u4>#m&xf#Z(`8<v9X<`=0CzQS9nIZVA=4zilQN#y~4h};$M?Xw%-de5`8T#?q
zk;-|^Y|WefJwp<bV{!|cHnO{2{rV@r4^otg8sSMOHqz(9R^r0yX{vtsG9cC*HT!jE
zOrgu&pe}knYMRuK#*j6TjuEDMK0B~&Ki(Vs2^J!Cd{!dWuY-P;P)Wssrs~Dt9+B>^
zBIr9>F+nGXzCH?(dfxunV%YT+QW=q?Lhh4M<1?#Le;5*!hQ5|IdDr(&Pq5?1p*y8-
zFWM1b7Yr!91;EYU-waC*JQku+P@Mf}T^-A4+4QWmNo;%*yd!1+HfoBXJXrTbD{UJ#
zvB~qns}^S4)y{W-*jRurkKLCh@VJr8ZF*tB@#G7B>_jI}7kkP(o#Un&B7w67Ab*y~
zxTH^h+*6zcf((HCA7wqg7#%jq?GHy{FPjk$klnDwEXAniwOs-RQA0%yn9eNnkb<&Z
z+v?%4prUid?#3yds>!ld@Gg=hlAIC|Hsra^fr@jdrjLCd0^p`S9Nw7z1AUTXRLA9}
z%TJ;N(d7E&#&vpd8Tg}lM>2_gT9S^X#zo0!!4?UAj~8z0txw?WRdLy@<ffjQ3bF}k
z+ro?~{bY7%$JxbOO`bTDP*(Cp#oKc)TJ)C|oJ5_|Pb(2c8{nr+>#6+pOI~Y#SS%vN
zPP0d@r-13rL*^shg(aQ3j2Mg-J|{P8Yw~8PnrB2DQZLoK2z4pwsFXUS^Zq}ww!&?~
zUK$ip6SFhXT$7SPBseH4mN3Q}EX-L8;v#<Qp;7Wk+_WEK3$00v5DF?#ftyz!OeK?q
zV>AiXb=LtI$-Rtnl~8j7^u>k#EH8LBo3$1WaF&4<6Ac?9jAvmG{uB|>1f9L?z=lk>
zMh}wC_PDMGdc1MTuOn8tB<Z@2B}Q-B<?-s!+Xop%31xr={=D~3ljKq0UwE2Z@f`>P
zk|B6~#76DL7<*AsU5h`^mhhsov^2n%9ajVm`RxrLxDA2w$;0r!ej>th{j%xnmuYHC
zV+D|y7xA1jr*!0R@?PgtA$St{ACg1BS-b8XnqV>s!Z4;`5;D??>1FpQ89siP&>K-Q
z87*g-vVVU>2uAJVNW{o(wap(=Q0X3Iu~H(|eIwY=gMh&0K2gk@Zy(gj9mBdA9U&o>
zloPp2*{}lp!(NkK?i{joR9NtcfgpP#OQ0e4t@&4TR_1~xA^~WMVckH?fQbq4lgA>3
zx^~W4u<V2>a|wB#ZzamP3uiSZPs!y*E`|tA@FYJ=c_EacCm3iW{V8%^dC@K#*Z2p-
zvx@sED^Yuuf%i(=2ModUnJrB7+aU_x?WLpFRx5LVz~aRV#R~owFk82Ie*1nY*TzqQ
zS9}%En_WEUzXQW8r-LS~yl<9i!fv$kGT9pmM^lo${~_#%S@Lop4dIQsJ#HS5&IR#o
zbtKI!__U6nTkY)aO~~1pV%QAF#-@QDo&u0_-<NZ)QJTWw*qaNxeS&=xf?}UozN2zf
zmO(+=dUqtS828BALIXj`>+X5ujJWKUuN`v$O|%~OD|Sm@xfvr7-u0GLX6h+WN}ava
z7HMD(l6}Lj7UgjFRF=}qkY{)v&18Im!*^&0guxDPP!1iXQ+<uVKWm)MYcj$Qbfs<Z
z6oeEMe9&%(pBGCuAlF2|7V4)~K0LU>m7YM5FZ8h)Hg_^3$B&fo&XM>8ft1Y{n=;gE
zzz9c?R2XqQBh@S++N{hvrBp<fEKfud>G0E6fCEix8OAZC^xVC`&W3F{7@3@HR3C#Q
zl}$@mBa;d`in<ie<J(Q?3;NVQp(Z5s#4c%|=J;i<_p^l1mjokFp(qkBZSRzg!OG@q
z-Cefg#4$$M9OuQv)JH|{bilHP_Bn!9MuI-tJLM)lD2aQ{+PydVoVwxfa=Ei$TKD~Q
zq+RL_8nIXTfG}66(Qbyd5K&#x2C_InKyD4G8O#cs$EN0a8fM&M;wWz?e5L$uoQOWi
zs{WhvRix<tLxYZp1A?{b0iGnb3o@0@ugAcK=Xmb%ycJI1>NpvZIumVtiC?f?Fe{EC
zN(y$y&oo(|KdO_lYXc!mso92NBuS?}z_WgN$I25W*SO$kYFY^Wmo-#aeQifbnq}_#
z^NXDP+kNUj`z-Y@cE0k|bKd#iJ`gH0;EsrS(H%bgDCuZ3Vkr8rnMS>3&18KeAfIVb
z*1Tf)9gQGMS*Iv<40Vl0V8R==6+pNIq(nUeZlZM@^{IIhU8k=IM&8^_xd9T9)vPe~
zW>oUR#DY-(f7wYix2>okiRo#Y1Ze^dTdFd?UY+;gN#@dfygwUzt2~;MXwUza1V`=F
z$Y<^e?m%njek~N5(?TnsmdZMg0Fq2iv99|%jnOrQ)iyl-vsUz@bt^vn<My=7^9#UT
zxV{H>*${3%{QX2}Q(E0DvnoH1eybcxS>(dHSiEX*S~mXE$Lk$TBo@!mf9m7;Z~e3X
zIvbA5DlnP&1fF*C2Dxe#$C=w9+DAmA67(|>FM~pf*8|z-HO%eg4?qVTx98gqEH!@9
zdW*t{1zjwAp$)43bKUgF6p{~C)-I^S3mu}8Drs8^XPWGG#A~<^4(j!Q3&^_gR2AeW
zowumM3fcg$Z5y9TOSeQuwwW0JN*(Zi0F0&V31c5I*TAB0tNxMU%$Ik1zh~2@*_z*@
zd@aG?j$paEHnr8Q<w2XHmMO8nPu}p!Z%PW4gJ&CM1rmAl;_0!Ce+tAad7>WvLXQ-y
zQRCNUTmnFFQsGD4O4-g!KO#drAe@C!yZy5pJLe<<5fh2wpD-h#i<SxShPyu5ul?wg
zcN6MEg*B||Rm|Lz$I(i8_u1Sj`cqF(-;>>>s}L<i7iiyTo6$i3XN-IHxo_PywX5X|
z0@_quzNY?+hNEp<ZqtLmxeqYew=I_<CF(7~6s9AzXo<hjnu>_RqL4jBN5+A3m$ZfY
z=?jaYX^}LwVSC4&TuBJDuO=Ul<A8E~&aADF9Rnz`<{pz+>-1Ym0<RGnm+86dz3OFs
zO}4`nT#fCIZEL+<?kW#rBrQ2JHGF!K^DDK)-GUAGxk)X8*&J9`j^9_|MnqdxRpD{@
zKN<8Lwhe>|C!r?T`Qxr$-xyddTR5Fbymcr*ngY`erT8)`>?o^@CAyg<3spvMKvw_|
z;6sILc;mJ-K*!sYRlR$A4!XaM77L=~t)A(wH4Vu2r#5a2EgGjs!{7*fXv>6Mx|GxY
zM{QL$!T9vi9mErro>ib*a-DxZn#X8$*tqaYQh(mNTs3}0*HXp5#1_wEZ|+!C9=~w*
z0oYO4ofGw?Z6z^<e?WzQoJNgaGSEA#W5hjfo9GaHn#@LaS3a}Ccbmj&zGjPScaMjP
zzdcJ!3fRsdWqz%%eDLZccvYuE55Jx0eV@uOADB1|p!nhE?p2AnOG7$&?Xg*QO~FoG
z%b^`}KEh)wdNyIQK}+asqp+Z&TJQFneYli$gphoduTj5s9zi**-22Tu{rtHRl7>Tr
zRsQF-QCgad)6XJbVCKa!16X1fzUQyB{N2fhPytS+m_v>rFz%W^ko%_UA5gmZ-u+1s
z=iTOpDfOp8S&EA){jxtZGi8^5ub7sZ$DPE~Jo~%?VOVi3XK8IkiP;Bdq~TrD6t3st
zTU@^}3l85QmPPrHe3PGu)yFw3L-*&$v-f`7Pde@86n+KAOQp9U3BCM#_8+S925x1i
z{4jM8zH6M_vNSjab}<0aAtkV`E5<ejEFw5QlV<01ZDXMQ4@Qq1oPc*QZS(gdVa&&B
zA{%>5$p3s@dJ;QnKa~C6Gy)1M^H@Mq`}^D|pl}7UipY}tr6ij$I&4ck{4Qe>o9G`R
zn5c)=iw@KfWnD?C6E9`0>(*>CO%R=ZdUtTwZ>7jJ+4&_9xiHBL*^bVCbq!6o5%<?k
zOm~LYWNULJ&lG-jfAuLC9i7cRq83||Miw_%Q1(hk(BV7LNFzVg4H5OZ2Yi$pFe3T(
zhK?Qhl{?o-E-2%RZ=2M;`GN&%U*&p_F?y?Nu%J!_o3-_oQo`G%=~>2|?Jb}T=fzaT
zlK!ks9H`k`l{wz{9MPpH4dK4T5`ab`rlI84SQ#Hpf*(rlX7vKRW9w6QVx9x6J`=0>
z4dOkQt^pw|;A*=KXvatAce+2kW?wEiLO{CA!j3k6zHKr~{HVrW&1)2haFJ_t5m@yp
z<#lhPtNEa-;${>1<G-}fYzG?y1zo(@u$BZ+11EMEpXGu_6$=&=Q;`8kg_vc6R<8%O
zLf@Qj)z7DF@JWDhNjdOQk>PFhl~*ra4+7wsAE}^U?ZQ_g5>5ddfeSom_JN#CLhO!2
z5;@aa^6Vu~PjU=-lbv)EZ{toKHhrq?SVqjJh?)!Rw$E?uU9nU!biLW3gbeleJ)^-0
zg<~<-6wImur!89axPd#Rh4}*AV^O#xZ9N0GhvIIRyulZQ_SxFcvx^d@UCa1Z#^8H4
z`t}D^^e=vSd-q13+hWCxa{`j{15WMufRraja9^m74OdsHx04<Js6Vn@O0O1~Igv=V
zMa@1RXQyKVg4#g6Pqr(=+t}i8Q?X&00H4gH{+csTLHGj#aPJRgi1JJ2S-REsd-gO7
z?oz85DJN&|Qh)AGY7|SG-Ce^l0J^1la-tT~^j5QDMLA;YN6>*)LE!Y%b=+OS`r8Qv
z^}YhEsCj@Z<2o}LGgr8Ez3D~`WX62sH+H|O?gv1wxNNO)uvz3r>FE^0O3a^ETlj$D
zY4YgV3}VRi31iUV9Vdv4%Pmm-Qjr<kLe<D8Yj7L0*}n3@qNBbsAx^ije;Y!xQr1rq
zw~_TFTHB$x!zNz0-Rf&@%zyyE(^&idG0tUf`*f@IYU(Iqn(Z>j&sk&$K;@uro5<cl
zoAeW<Z!v7jK;m8?LAz1bdshz(obOY@feSQ1uu<16&K4bToF3|YD9tn`qH_1FgB@CO
zof~<(dE4ERpH&xDJtLxY>p9urk!=OH^3Ra&I-Jlino;NPLM4~i71Y*Q0%8rXnVI<~
zl$7_9-Fw!MPixJFd(|LfP@AeCr$FR6LR-A8B}ZazXmyimLu%=r)_15%RT;D2L{SrP
z=t<}F88rI+$F+9s4a9yl#yTO)i2ebQYPbu+iyJQahe(qHz*%;v)D%&?ed4J{f?q^H
zbXlk8{kkvi(91ORS<0LLW`4d3!TMK7<%;|IX&XuY60h5<q*1$x*XPzI3#HLK$;gZV
zPxNyPc7o&-d81a^bXLx{oR%mZMzNFgFV;$IPirc5jE<=v3tWHLH+}o6T))V6Q;Juc
z-C~PH4qct{y3bRTxD*qdc=b)#3i@0!h{9Y*aTo|MB;en3aexSawP#F1mD>?uP2CP!
zI~&AqELImE_2;QfE#igF<MvFCjHmLBa>X&gcTXU_oDA2WZ#6VG=rwT$#_*VMKY_5e
zt3?ZB`ekvq5nRFXA{3BmQv;=P#5us|)digYuS}IiAP1=f@Z}35k78zHP!g4;Ze?>t
z$=5kUZ)K{63q}rib(^m>E#;PiC;CgzXLP{2ZFq9*?CdY-<1o9GPArwDs=+u0BGg}j
zY70o{8Tq3qYb+%3k`Zob&7nH*_wQfv10^5M_~;D`Q&d(UbTTs)$$;V<`|Pc$+aoHS
z^!tsOin`Dp+=3&QNZYUO@7@o)Sxs>OhbJZ@6AK9RfxtI+$wmbh$K(woo+tMq>t5}V
zL0sgnxdgH|ny4!ei#J@r>KG{N?=FTZT1R4hEbiUgfZ(#z73e4Nc~_Eg#YfrW<t-Tl
zQjs-Rovpmtzlx@&{Y@Gj3<c9H=y`dqyc-^<RBImjm1;EZ{^}J-rEmOR!n=ouH8oJP
zlE${n6)N%?7-Rp!%Ni`J_UE}s6y0{mgyTB<tj}7Nn4h*~*`-yIws`wj{u6@gfO#PP
zy@Kr%-Dh7JW@Kx}jAve)pPw%y0?!b(@Yg*DyKq`Nw^JlZAvxG{)!8X|gaNdOt8enN
zWw<;G>c=5HlRO<1>DZb7<+x6~io=b{Vf`g#K=)L7cJ|EdtX*bZ4{k;=#mjgF0}qn2
z=`RnLn!H%^3Rpjph`~b3Dyo{UQd;t_rF_Gpc$kJkp0~pzektHfL5t0>vV53fd%%Fe
z0L4gg@xJgdixA*g&Bk(SO&9WL3WpOG`%U3DW&I}ttwa;xur{IKx+~w_(KHss($@ms
z<N}oU^DO;SUia9<P=a%Z1~ZN~WTthKS~tQieua22WD$(VpO)e)V)>IW24K?-lpUR_
z6Y&TBe`YbWWBY<5&s(l6z-s~w5xS09y7t2Kn?6Rmpk-C;89e7z!#7)BjkQO+TA&85
zYt_G6&=hm9o|X7QXZl@&v=f`7WhRYS_g@=$XDO@eG#1M;d2l`V5U%&5<c(B)B$BMm
zG9b6BW*lyX(i3BDN+AQUR+o#cNwl&)D($mWB)E<aw*Fj0PTZ&|q1A|tZ6*?O(iT+d
zAF!x(|M+W};bhmInSc7?Gt?E{cO@p<Lc*EJvlgT@jPP9{53DCWDTjgxHv|cVzfumJ
z!r~4`!i%y>TPxK-qOwyQ*erXFp7H!m9KJX#xxqKOX06Av^ACQPtUl$fa=OmJyc5rU
zDMA=vI&9irKJ(PCPB=jDD@1}2Ics<TLJ>B6adP}p&wgwe$-c+Q$S_sEr#0pZJXxn>
zk5eW6U6>?<Bad}#R&TX|EM?`5<{R-ij(VWjT9-;67M*&2sORoVqB2>bU(_tJ;=Xa7
zQMBO0<qf2>UYFu&Lp}OvY9$8-x(_pgPs_<BivUR=kiXmAXgl^_N?2H!DaOir+7lPn
zUK2bV1`Kc9XhQ#Cz}wFZ4;5RMOqiI{YoPvOW$(-a^YXcEiBS)7w{i(Sm|=Od-rRUP
zqJmHq{svoAcFGgOlB`@5DeG8$tD4@$hHf#E2G%>%Sau><S!M)UMl&q!WIyy`uefmg
z^5t(Q3Qen5#nzc)ki9Qos2=ujC}+NMl-<>7E*!V<C!|>@ZZhQq;B&><hiK-WMK$aP
zK6EGLBxX+uEr8`x%%pNn;B0b6vdNYuHZOyLebUp(KnE$IkrdEmxw()dja-PYF(Bi@
zH8<{IHow_P8-uiPBtaoR^E<KOR-znFLC@F?bzV7h3P5C3ISa<{UY~mn6Ut@!9z#wg
zQL{3L98Gt`)p`9Ru?EJ&84$PV1<<<FSV!ZI&b1H9e;BZ(T%YU&Z%319R!=F@;7V7N
z1*s^b<G)H9Zy$GdPnlSKopWW5XPj3d+A;w{&&41kgP7DerV%Z6qd9j=DYByji(DY=
z3fpEouRugt3RYr|PLBwU%1+YYpKAYH)tXb0pM6b&vNK8qI-B085~rKuR+sif^CPnt
zhl^h#YB<~jI3QBDm|&vh=Bv>C3PuKZ^FyL5{-*l6x_1b~IxW9jx0}K*$2Ti`nS&}3
zY}7j>D%H|c>0WCUIpI}9DmAo+Y1LV$2vz!-6dzTZZOq=cHWETu6ZD83gR{zvyZQ?f
zP)vVXE3<dXf`N3TU(KIJR|0EiU~6aRRZQzf*#a}84y+`7=CzREaZc8-d$M3=MzeTu
zQ~4122cH1K+PIDL8i=nIOEBG!77OfYKvp_Bz9nd2EZG@jbPheLrS?IbD9Qz={?<2V
zGi*`ME|%)yEpDaN1)DM;-woj>F{=Yit7s|>uQ$OZJSL91FkiRQ%zA@jqj?S{?`+}C
ze&Z9(O)9L-K$Ae--&2y0Z;Z}DdxgrGAA%x0u3C8*zVrRN*{7#s#c4CEGH7>Gnfy(l
zD+|Dl?=c>NC+Y8wJ{;n#lzHVf;Bs-}Q1OPxTj5Lte~H54aQdI`NL`B$7R17gbYK4u
z_t5V_7PnfgKZaeL6MH1_3p*p)Es16<@oua^r&-W)SkNo$-;A5FmGF(xSWRv1W8g(P
z@BPba@cRnbG8yeO5wPd60miEW<*>z^h%}gSf>gHV>yU@c%Qglk)BOhY`&D%=LQUYK
zB-Y82Y`z;~doDv(CPk3=e{F}Y#M*@Of6dyqkimzkzhF@v50fJwz2$`Fhfe<Yj1`7T
zX3eI0)F}Dw%`<<tj_o_D$E0GY|Im7C^B_(~f8BcD5BYXPZl<6}vfGIg8$6Np<f|Ny
z0u*T!RTZsU!qb8V%G5HxmHp!ReNSm_v!wHFmP2`Z?%_W76Nm0Jf_tQxRgF=rQpKU-
zL66L1Yv$q=$Oy7q@~mLXU7kZjUN4KDd~mm!E49-eFOw^r7zzk#;u6<cRgSJr&vJ&H
zG>2Ra>yCV~qxp)=?J53AMP>ISSaOxs;DKAr3J06NBr(=zb{`g9(K$KV+bee4XcnZ1
zir=wvA6W^|D+)JuK3P0DMAfRE!m?jeTksBSe1zHoQ+L~z7G8*bSp6?V<$29IEl(34
z%K>X(OQnd95IKeW+ou-5oRzj=UWFmLUpeIZHelJ$8AWt(Npg5~q#PDs1L8XNl~$8k
zK3K`xp))V-t&%i8If@LMb85r_$N%xbi7k&I(I0<#u2$@T|E;<u+Jg%Z#8mk*_2O?-
z&oPWR?H^?S2^#-ts;jkbD}#Hxtx__YlYr-+?;YP1+Iqk$XiAJi<6=yh>#b-HB_iC3
za_1ZG5nDVP&I5z*^XDzYNv|P`{nt~_pMLYp>cyXh{Flz9y&1-2iN14?lk*Q;C7^p9
zGrm1{g8#b|Th|&ca1q*%-u^;j-F&W<9R?UQe4p8TF0R3otu5=v1U?s|aC}^**ycOa
za2z#nOYmk;9;(o)WK!rz=Us8s5Z$ReM*38h^`P$>oG48Y(Xv08IZPFl@8^9t_Ux_k
zHuNzchb~3hYh7lbB-sW<O?(G^MLvJcE@e_S>OO{i5#?S>O6^g&P?Y0y2`za5^`Tcx
zg;+NtbFja-Qa8-;FukAe7b}jtx$&kZ%eAZ31EakMCXc`eVv<zkmtoQ%hXn@p0592c
z?th8j+(jLF#?@XNiBqd#*aI{j3@h(km4+{{Ck${b)GgQ2`Bs0w91aq%v^I6O2g(FM
zGA2KoWR|kL>Juo&(73mM)I}8Pr6eIOF2)pt^v>RVO>WuquRNw@Z_@g6i$BR9JCm)J
zh=jM|tOv`MovWrd=##pBv^CG+YqYuks)^rJZ|0X&FN|Ox`xX{~K&ghGbJeHH>#@qM
z6D-To^`qAZ`svZ%q@fle{@row=tQ@or5L|iTR%@;?)Z7_z>C@{A~6`Bj)>{rE>&>m
zW9#@!5w1eMkVpri<(?}DdNiYcl)%_3M{rumxF#6&TSrf?HGl(G_<zo2C58Jc2SD@!
ztBP9R1p*-SvQ)!<g=ehZlGv|rw>JyV<?REGgNlSBjsq+XK2MVHE+ql#YZ+W<=F3at
zcW5V{SB=h0ST4*;%;sCR=O4t1-$@pK_R57S6{RN6xdNAZ5u=UmEk(SHlWz(nsQJHU
zN&a>m#Hhmu!{i+`m?-<Bv*^k0bPHFy>sKxl#_Uin^s|XZW0F%ZMf(cAnc%pydM~+7
zsfi@fO*-yeC%)BcQ^N;mKbqIcEvnA9m2L}l|Ar9^T5bBCKalmxzDf5;Zf+^Ko$MJM
z*41e_`5!)VLGvQgY_MgkuKH%S2MrwFAY)S0CK9wQQpW2}FFpi_jg6y1CZV7s3eq_3
zB9qz4plWR7f2}a|V}*<#`i%u=+zcPTeJbhK(SBki(}ecsIRJuif>@Q>$3NwSrmp<e
z3IO>iQEUr?eJGLUuem<vxI6Q?D|9!8esDvc0;;Vo7dKRnial*yFJK*{hR}j?emVBM
zZva9z6EQ>`uz5}{+e#ay%!+~Q=+H3XBn@+hK`iDO1*3H#3DfX>7UwiL4S<aTuFzF_
z`(~VDsjfhN&zPHp{Z87sJ(3Ic0zVxuLUh_+?x5%PKsYC~b91xMXSdExzG_k`HK<uZ
z;Xz!BJ{|3LX;OBIT3}-?54Ind)miO;Jeu%WoFM2%L)sXe7ND1fXOgCM;%leF*&Yz5
zlxLjnMHbpYe<`=0d!#J9V4K&HDE>w>U&w0#uDmRsHxufUg`w2ir_L)<@!zpXg-e)&
zZ{rU~dPOc*1^>b|tKb%&AR?7Xg7nE$t3TKU3hv2efKBkloR-+J_TR5M#X;^duTx$&
z2n+mix*h(!ao&voKBd~FdL_#{!R552MO?ypXn$NldvicuwqQ3(J!FvC)_no$IZd@5
zW`2HY0+ILIOqf<~d(Z+5L;w8wV~uzCqOp_R+R32sxUJiyt}8i_<P4aA`a>;-;L*2?
z*`;cH&KLegs{OIeWGMr1dqe5r@JQP>(}^)jdcr%POV?pO9UYqh_+{^#@RZ;qr%|vm
zj=SypI<U4}tWw^6Y0)=pEgS`C9Pf2Ez;5rJEC}#R|4dEUlsAHv?-zXIY+o5MksU=7
z3neDitRp<)78Dique7MU>>rX%3{|%}oNvK;SZzP0qBdz#RMVy6_TcnHTTXPByq5b-
zvI~iCL$G3)kA_pr-G+iusL)Q2;yIkF*@@xEh^<Mlcsqngh@?+I*|>y9_%4u9v<^P%
z^_lpOywnkz(Qu~yF^@%GNuDKGE3{TbUcTsGf%{mw`rYLIAO}U?MeEKCD1&r%P^?yZ
zVErN>yskSn$lI*)H=Ix>G!|>HGYwc}IPA*?`7`6@bylfZt}AwL%FY2}Py#EWw>5lC
z^=MH7tQFg^yQwDxRry_yUmkDW65~Pf-kLrA*SrHx`$i5CdPr}<EHmHeS|rErEuR)4
zc1&fRB7?$gF9+Eu=n<r5oKOj!rOCLF50uIMgybz@is^^+YZW|zpy^&9etEkIj}4E0
zu8&xX<C2ebA(AQWD9aEOG`tdBtvAN}UL)fz`Xs-9Mf8-i*IYHQb~O37!4+zgFV%i^
zoN{b7Mu<A*y)4N2i6fI!r-$|C?_Qs}pBauO1)_{gCs45yl8qe@8XG%=O#zQw>WeO%
z_%l_`*SMN2HhWcj^b$nScj7mH+u*9p;!%3?zRM_ED)j>3d)@tPw5r>^LraVs$6bI8
zTi$fRYz<rC*zx>J9RXD(7!`E-P`<Sp;IB8F<mn@QEi&%Ls?%gH*u;ejjkX9Kcx1*$
zswYC2*+vy(kDhb6+Py7ofrf^_#bQ9Wz}5)}`4~o%n;$K&COm1MT)Wit?=O-l-~bEt
zmD5t&H|>|Wv4pzumt_M|+G+VONBP2lPC%MiBEVCM9+i=-tbqIdz0%eRv5JZ^hm6v?
z7{y|uhABRi*ukm&QJ1CObTlXS!^^yd{^hV*UO3!2Y{#ThukGra?T=0mZ4WQrrNK|z
z5N-zxyfD+d%gXo&p`511Mi8KX189jlZDn??yT5|JT)i?eUef*uOYr`Wg2lF6=cx4P
z!~_DR9NJx`hZS%0c+&Z5;)MR%g<=qOkn?HTuBNk6JU-vn`No2xI^dbt<m&sVuC``}
zEM7qP@EAmVgIV~Wr4P-wsNGE*8=crpJTFks!O0cmc&@^Q;9}dCV$ggf7J@~fMSIY+
zKQk}<)mm($`jhq98}xS9%l*W|r4F~Y$Tc44o_J2kx}^7>a6|xo!p;q&h3>U)s~Mfw
zR%AbN0fbTrA5gaXKug#dY-4g=IC$dIs<fD6b#98&-;<2{W>l5+jUrS}k4|^{Aptrq
zI-4w?De(VAeUm2N_2N>P8g_n6hNNC)Hz^+NJd7v#BSCI0kl9<1mRDc#2JuxR%c1Tg
zeG%0GqEWr9AJ~}m=&m*au9~z=goOpLMg^(^W~Gme!`I*=9!m|gYEpW~9E8wUbxdkD
z;vYv`|Dq8qrL#lX3u(jL3@pbvBwCAxgM#$esZj^QK)S#s3)52MVrY=MqgdWZm-NA^
z2SB_I!J0<mx#xCP#bLRcfD4<kE5<WYD;CG5q1kIBta~Es<ArQOSq!#hn>u^2{w$D{
zTueUlH|Omj&SOs8FyZk5Q;PuO9E^hXXW<P(i$4T*?Bv*%GiY+YaIq@Ef*zMy4js8|
z(McU?XQRRjOiLV{N~2ox2w?WTv^0mgX;V@es*D!%G<_<FpJo0c(P%c*Hqg_qfmPhV
ztlql?oN@1feKCOLvh~(-y8zz0t&{Dg>WDn!;e65PZ_n&%RwgF@qCG`NSQY0dM1^T(
z)J+!-vX6Oj>D0Dp%%AKD5?6j|4Zx8rQZ%DmEeIY;^i9a}C@RA%XDy22-w^C$l{Ooj
zIx?fgg77D*J976%zyxw>9Syc@#}cMMYy78<2D3N=!#idf*_=}xXUh`RPrt2cD|TOY
zG}%>{o}<6>sQ;)x-B`~4Poz$`Y~p1KW&sw0*EKM7RQIB>?4eZhqke!a26zO3A)W*6
z?*Z)JPk6S7j{pOYb5ov7bk^c@_ML&Ao-eE5)9<E%3O}W2LQYCdkC2p?fLr?CETl$|
zbwvt;O!#+pCh4+{brNtWb0pP9Cm>Wg<=DO!l|`)7&uv2Oj8))*$MVvI0Z4kd^F95B
zEbY(d{(O^zV>QASBfPm(`5CZhmPY`TZ6&D?lCSDf$k$b+rDScZ7f9zM&yek*O5c`t
zXl(m55^X@Fo1AUHWS<Sj3O^SDIXOm5s#-41$B0Om*_!G<*ATDV_`BTQGqCn&wa&u3
zw!w5AU2mMak$OX!;}r;()#N;OyMVNVWto;6UyZ|s1cM;Af2<rQjqtUW9GdtnbiuZ`
z1pM(FQ}vZ_j%NFcKGHhomoJn7;^9Ku9LLN6C+HEiw9LNn3yz`>i`7??Z?w}Xi^5^f
zF}y-9+gzLsex?xEqm~EN_-$6%*_Z`sIp*}nh7BkOo%#=6<J?-|Wly66k7H5q&xg-_
z)lV)42RefX6gEi%L33sZ5y{2c@+`CLqZ)?)ur7)1Sy47rgkCwQ6c8OFSifOjjoc9E
zv3A^Yo`1pC{8z(E{9ecRrq~kSF)?+1m_!VDqHSW_n^R`jUU{ez3f1155_B**`nKw#
zIHwly$@55R)D;ph8^o423VMN;_-lv~adgjkn5oT5P1I^zP*}Bsya#Tz;;n$X=zNh&
z?bP$QJ?9Kh{AeMp4j}v#lQL+iL?cbMWcXxXlF_91k0<L{pn<V%2@{R%qbJH8tiKAj
zrzHIG=`VHwBacFd&y>1G<O|kL3j)o}WF&4?vP8^x7|yVi_DF2R9IotgkMwvy3R^gN
z;Nl^}ulJIzxTUXbo0<Po1$JwI@Yls{PTP2%A8ON4UUG*TJqfJ2ArXLKY67dJrk;&D
zqU79aKf%F&tNl-izwX!a$aXD4%(?PPhwJXPWVZ9a3<qWS5<xE;Y0M&zdKX3OP?8n1
zvYc@#qY4@IL)T2W7ElOHWpV3%b@lJc^1(V)-IQ6$HK{X~>d<sCok_8$ge(o3GJG0b
z9Qud6JTvZOr1i(IExrKhn1H-~mvZ1;-t04dz29Qg=@3}H$+G;Zmp9T9z33(7!1GB9
zg}l-)Vd2gl$M(iI^d7p_UoW?oa^*Dk8(Jxe*(1>{o7?u;;#+aHP^2Phr29=s%uVnr
zlE0O6coZpw6iQS(1h6YznX!QK;f?IL6${gZdGnHgU|3J`A)(%HKO)Vl$xp!JI*psN
ztN4Bf;6PgrB%U#sssChkN(<UO?$Y^i;yr=8-H(@8)Kmg3O|^}gwbxFEN^)?50!1n2
z|B-=f6qq+}^?brGf{sLwkd=sPU7zg!7)wZX7pNKIgRB~YRxZ$sQaH~q|EP#J<Le|L
zy>1FBncRLk+ifIs7(%ZAigQQ!=C<=T(!wpJle7ZO&LKHgSwNK1U{dgdF-(n7H+K*E
zV1^WIV@IKa#L9~<B@-J6`X0(mP=S|dgrXW$$i(I+N3xArT?<_riE1$bHTR!|g;$e&
zj`BV1M;C3^d!GQ#=zXlo&hGBZSh1X4Yus9>uu#B|`}>p$9I`33sNRk(W1E8w?;##R
zMUB;E2hi;k@bH;U^K<lvQWltQQ+CyfnuynI(8X0?kHOq48JS9gjCTxPIE*%u;^HNf
zk7`qmFdTn)mUfUTtcuMVY)OD@;u~Q07r_6Yhp-|E=xG!+K7|2Tv>y6^QA*17j;!AN
z`v$N>Pn9yY<8$RhcW%`%!45emS<7W-ejhg6Y;_d*)u6!GfmO*+`~iJE53unj68u>z
z<wAN~yL94}z_%7-?Ep$1<ZOp2bhb#4x`|}`Z1p{-F4BxL;0hF`{_S0RA{V44%6r)f
ze}H;!WujH0F|GYnmIrN}@oqRsNww4Sjg*uIe(t4II}w&(S3HP11XCw*({zz_zh{MG
z?(+jre%zmb%}Ml3Vs5R&b>6pt_?Su|Yq5#xcxPjZ&j~G9wspcX2YjzbCk_5z=<|lk
zO2>tS@1fNIt*FwhmHHl@dQT=F9v&7SVZExU2^Gv(6B!?P56S9y{*dfNPb4R)6siD*
z?Uy-MP1*LZzEnN#?b_h<CeIK>bzLM|tFnyFO=S>x^_zV4?P_)^!_XE`pHyT5YT0ek
zK^wXQAQkGbvA!9IoEmnhhi^`Bu>tD@{-0?m2ekstrWRfcIonS~!db$2irR-Q@oT^q
z6$rE~EU5^*SU`+GG2g1d17gUoa>ttO?G@fpt=vYKl(n6p0MCD{;`%0!TPs}}rFG_s
zv${9jTq;Z!yYhyBj|YOP|IE+F0=LyY#wKIvok+&Zc||54H*1P*s)$GSiI3YTwR~j#
zSvz(hSq=(5twNKv$Mr*hDsdT`DxoWv?nfyB%1CoOm24QkXcFr`jVY%)m9UL(IRq+^
z5Ovl#XT6tk&2eHGK~6vV+VpAgq-q+d55twlHua#O{kfh5a(nc$v!VK?w~xEn#2FyF
z2d+KvGLgDmN%|kw8}RK(@A=~>-eN?)qbX44MKSjn*I%RN`jJAasvSWwCw?-0bG0Rc
zEYF^be+_XHvzHo6*>w+gUD+UpMybm|GdpW&>rm%Yqrh5%G*A8NC$rhOFQKl;rf%Zc
z902$S5;!J%rG1$`k8cLb*ybH+UIR|3d++91#K_#6G%br+ptH9-hM8#upj`gCfbU=%
z;<(|g^sw7<1f-^v#OugY-?L;WlNG-DzL1YRGuG!x5UkzVJIb~@BO`;7vK9exc`iON
zx(FDiNflpasW0C@#zomaT$QGc+LHi17sTIk@VXU)dLfR}b$lsxpAQD?5a?}zmeM`Y
z9!Q1(euLe{k<U+gjl*0+J38967Z83PBY*wJ@E<{Ld)J$`hE^JR;XOQM;T=O&*UwTa
zo%6PjhE?%G&%1qN5Oe3wU$S1Ulwbu}>ouAxS<L09Hv13?2+{l$yaYQE<POhNc6(fS
zIsP9>=N(Vg|HpArq)=3Ji9{FItV`EkkutJp<~Od9J+8f#E%VaN%*=>f``Rg7<6e8O
z=vvp_+wXY%{`E(0?!BLL&inm(J)dMx3`_?F1z4-~dEWzx*eh>_hl53TTxs!0Ugo{;
zh=F$kbpv6cQ%t3wqdSe!Ps7`v5zQF&Zn2DZtOPseOlqWDE($(+NW+rrBskZad(8>h
zq;jPqPOCW_YgYgEaI>wBvzunVO{!~bOur>FV(LFfzhdQ2^L}8iOvr)gGw@A*6mlx8
z_1ieK#y-o+I(1aJLO{=bMs23RfNAtCxa}8I(CN1`XH&NzSRg*SG-JVK5XgN1Swi6N
zL_H25Y^(G>07V>NZ+dm(3D~)&8hvWE4|;){Cw=~uD<}6v@z#|x;%Y6@Vz9$;=Yi}f
zpozUwYXhn4H1NEDW>okI{KLN&@ZmdKShs6sKBB0<i1ZF0V}%UAKN!y+>*Lk#N7tW_
z_0lexS1XZ>9<x5HG2%A|nj+7!eYyrQeyS{<E;;UUBE33@)v85}1>d&P5X+_`;GV60
zGUQcA6X;QiyLzd1(;k=U*Tsw<Ww4Rf)AHOJqncxj-|M{q_Vp^p`kWi&#JiGEP5D-l
zUTNi}@`Lw9w=<LafjjCCqmfK+&OH{^S`$$zvgu_ml`4N#thF|9yo-F4VIe?f)S*n}
ziPJMU^IW*WEuB))ot*1GZ(6e94hHW1_^>WI`Jv++2PPB{<!jt_ov{(N1n!hQaW$bY
zUNO*@uU7025Tnr@8=&{h*YkTS99Bs`8?$Du8!UEPB-)RJ@$2n}P*n7j$!^8{xCE|D
z6RAG|0x?+_Q0oHBi2#Zpc<t?1*>s?{yI_2#vlem^zc~=c87VL|P(do$l;0amK{ye?
z1|KP1dZj}_wAACc6NOp-M(x+`Wj~l{?kd|mnX4UJ`YJ(Yl#abJy}f}?Auxty-;i->
zYr_mlOC$#Qb;`V34OKbRyrg|oYZTi7+$@b?Bf$G5%NhnOS%o2mPn%UUV+A)~Z!&x?
zqpE~wEG5W}LRrgP?UGixz!QoCT(PEx?|k%b++b%5xHbYjl)d|ge_Bsfv}!)R`dy{p
zYX){B9ux(uL?XB;qD{L2DOr?&1S@BD9ZW%(wT!5q*7MU_l1ksAY3+vObu!*Ci6xnN
znSIu3*ek^w9e#iPe;F=LZ|`eym%>=%`}5#=nkBDNqf5-6@!xuLRyVMVYEU9K1|`Dm
ziCt7hmZmxVT)-hPKvMnA#dCjH4=CC)?W;cKgHHo`nJX}mtLyHgv4e{#z;<9dh&w+O
zdZ&Fce@*fMIIHt{=YiEp2mH?i0I6;Ko|Is0_VVz%Biesgs<Z3qcgV_%MMQ>$7~7zu
zklXcb|IpjiRaDVx7vJl(^H0L8URsZ+c>9g#3XucDSLUp@PA{{%8o>UA>DO1!FUG~i
z(MkLDlP?A~Ol(I(9BK;>1zA_tJN*Ny6cLqPd+Z}dMe1D#p6FH=$rXd-LEWa?HA@=D
ztNJ6;`F+QkWbB5i2Q+&aN)f2+zg>oOtV<T8_U-5kr+tk+RECBmRo2jp-VF&7RkDyd
zIWjd_C&o3YC|3JOzbpWJWP12-K%4FBs1?v{y$Z_$<}K={rli*!QobGVW*!y~t{m3i
zM~qe>D0ak1U_PPUuWu>!2G%7N3W%>rk9t&n#sD&}KMbbPJbI<S0j3f#2fR{&vva(H
z9;_CZwHE+Npeqpm6`d3KwgKv{dYUYHYTyD7%w2#>&+ra6=F4@)ryfZuYt}AOOQolD
zKiyGtB6Cniy!5vcxXbXc2iR{;fsH3DW0V&zCX`BBAD{i7x!uiOtXA&m7x9NZG7lS>
zbJe<8iuCZ$q+?3D67`bJ-Z+cQN6ec*zji(*mEFN#<2`<5n0IDyhVuC;*!a>>z-AY%
zB=#d|q)M2W)UPu;K_P>Ti{*hxE>&cNFOp~!AfDJH7cTkAIJI~@FSWRj24(O_(s1dq
zXbXWfaU?ON{gW-ZmwT1ZYQuc5jyg|IPn?c2rf!P$Mho<fV8Cflyu?C;LpK3U<8FQ`
z@}{!_0LOUi)&a}vnMR+&Z5~Q=_Y*5+DJNwT*QyUq)u_X!5!o-jQcCtfJRli~t??`n
zl>BWQEZW5>oBRHWeu0Ht))p>Z{K~PX;FYU5x0jb<5}9J+4FOP%+B!NNJU`FBqM(ct
zt~o@E$)pO{m~;ar`=D7OeBd|*bVG6$Il;kCRAp4k`j#YZ+xCjCqkv-8(ywP)^NPzc
z2g$)jgj!o%azmcItN$|*_>fuxVL<+Q+~fK$<1-$wpJ<n9cMHGuhCN$f_v!M?{><@|
zsH8#N3$d%S((T?o<2vP7h+tns2IWubW}2N3T0KUBhq!xhwe9TZT5UR`pK9<RJr%?$
zOf4UAJ<W|rBxwRqkMfV|=Lu%H#VHk9>G+n>9P!K-><jL?i_SOJHr2luEgs<Q`n<YM
z)S*gd;5>rRp!&`+LjCmBgQ}47gn17TcYld|EHzmR5Hwk<bdt;R*KG)FVIdy;z>XhT
znru%?ny5tkGR)>_xjB8Y=x3kJV;|d4QM@f*Zv|!2Gj7&NDM;|GeJQXawS&uaEM_-!
z#apxT@$#Iy=e%VXX!GA+9lmnmSY`Ubh6f)x?&$CIiNq=n2#VFqy3CgwCqUNM=Don#
zhAcwWsa90SAA=ic1LI4?#`w3Zlavzgx1YqhXa0k`XpxeZYCRGv$>~75|0tg1cXP3*
ze$gId!!%!av*`WB+69I>*G!`Y3QlN@9HB5!jIY(~{Fpt&kNM-7i4Yt}xuyWVUgVPd
zJ)(t@KJAV}&#(yv4Gr5kuB0Klbbv}s>!;L~awPIRjrt$Lg6H&yRm0o0)Kcw)$>{v*
z9jH`z=EQ2dFSv|u_f7uTC4-`Gz8g#Ajz*A+-QC*VCL+Bz&oNbyw`FC$L;@}|YAX=a
z3h(S*CpFLeSI^$ypFGZSYF9jnmjGg(Gtl#2w*`I!Td5k-BC1t6lcoR6ljIJ$pg&U#
zZTv=)-DfP6+evx<-3f8UGMI|ddiH;b58`~|8aTnh05UDi!7fl8RfBw0M^J}s2XWXA
zz2Dw*LhobpB2~uHK-yzPzCdysI36Mcr=Oj?OXTU}&zt&G`bg1-V#^z1^GrUiH%&?A
z>FMTZLr%R|RFLCT#)?m@<ThejiH^qm{hN@S_k6Z}6}hr0RXNuKzl=_Q(RO=81!iwY
zS0D_M|6HCKYyQlr4PD(YW5N%RGh#>(shXY1*Of^$Sv!b&;^ZN+jg#LhlQY2F_J@IW
zbZ|pKzn1a;Qk{ot$VfAI2FPLcElYZdM#;*1mO%k${M*>$xn@N&gh{@kwY*MjcXlxQ
z9U2<O)?k{rF_W1bf9G+hvY;Kvg%aWBBTSQ%pZv4U{kuu%M$+;mGH*q6IiVWwrGRof
z{XT2QbUd>tdC-(YW9fzBr+phq&CM61Vd_3Ub^BUa37s%o<IKv}F~7rU1|TnBB7G3<
zsJE3TwrBj(nU>JHq12cwQQ^L8$XtT3XtJDG8FjoR+MVbFY*s7fDRS&y#5G)KntHK<
zG(J#@)Y%H;bO*6o9XnWyJ~Svzmc1D&($YB?&C&H5JBFYR)I3#MU#da?Onmm#{Z&i`
zHp9?Ex7GPORAkqD2qWNC_Z&OH>pdprV3Vln0&YL?N<Z{H4&EUs4SPKI5+C`i$DvFN
z4n^$1(MvN$nxaX_1kSdO|Ktrg1X=Tmp!BX{#Sk@#FJbn?eO$AyVd&&OI4HcK@z!?)
z*^je-rm$nITalV2S}_tR3`Jkh?B4TlZ-%k|csf%#E|a45SQ;E##+g2aX>Oh!dm-D%
zcL$;OVc6Fo2W#if|KJd~Tc4_RUZ^C!hc5ZgKJt00BphE&@c3w|NoQ|u3ncqkJm-Qu
zJH%&b^~y}%S}|z9+khgTa7^%08%=AGWGDGS7;C5{kgAO4qBJtlS9!DcLN{-Ir>Wyh
z-vsspFkxXlNSw`mhmA~f{4#nmx_SIEfiu#C`oNLlnKL0VhD;beuKRP~a=`O_j~uLN
zYO7OHrEl@p$K!K@ZHSf$_>Wpfg=8c$?vXP?*a;gOqs9U0Aj_qb2J4(XPgV1eHGVk_
z=scK(Hu|-YxnbwM?Wjx9U9hPi<hy(#*}Ie{j|U$Ws|vuuiLWN;R=Z2isBR{Ce#JJ|
zaMsV>plbNMW|AM0<N>fB&8M%yb;Dxsd7nt)o((PlD})U1)k4?T92RrMwo^u2?LqBO
z_PEFpq!p3F{+d5Z{Rjd3A|&eFLnFdXdEH!O;$xj{Z1H23oTn^X#`pMnOZz4#GC@9<
zy~2gt2K{m@{k@aP+VnZ;<IO@+7&RFgFHsQ@Yfq*x1-7iBmlGtpyY0vM@-2fS?qJdu
zem!=A7IgJi!R=@_2sjzLnc!ojfHGHX-j}+Uaoz&;HG<D1sTk}Uob%t%ub*-HnF=vZ
zNgXH!7bFtuUbj!q<0Bl638qi$cx(T2I+-A0x+nP%r<S~cuL5;fCxMn#@Lr~tLJ|JK
zg(@59q5MfD<v;cWOV&5|PQjj8%#xu3qOo`5M{#VL)XePocm&@6pkfKPsGh*D-?8xX
zK$K?Yujba%3<ny@s`^5krxI|oDP&f{$4ILT!Jm<M&mnuPbyBpKPP)S9S-?<h0=Qq?
zP_3i0d%o4w@rdutz<DU;qBA+{mR`ZV6m*Fvm-NzqIb)Fha&HHX2pA#QrbmV!m>W&Q
zbw2<*|H=M5z*K&Y;$L`KH(#al`3tm9WyT2=O2E9gU}|`47j=>-&glE-)YD89`UUu>
zX?`Q;yQ3>Pg~`DfM9jvK=EWMu5^pcaCz*O9T8wWarn`#>O=BzrshtHdJ;ZU`Jn1!-
z%PUvTGGTn3*R%^&3dD8LSm7kPJ2*+ffogYdUVGg5)_|K&t3KMYA&V+Gn6zTMcM1J(
z=gT!=iZpQCYae7)w~eMt3&?p9!I4Qh=WA2Khad&eot)f63C^AvhKPkVA#)i+&ba*2
z)-U}F%!$-o{V6|g_s6VzNhp3hEw5|Xr0uq6=FS=c%liumdN!H}NM?M+*QR{u6B)|9
z$Ca&NXeF6mcz<lE@GXrJ0!SY^QV$c`M8E4;(igC|CG}@lSETO+!fi*5Wpd@4yRgX&
ziXs2NId!v}!pX>+Mo&}i&3{HoRlgvUV}OJyh>^9;z(BCi0^GeNrLd4Zn8?k2Okw@a
zyUOhV#!*m87;0j9&>J!I>V$i69M$BMZm^@BED@I5cE`VK`F-ja3~3wtd>WN_GgmC?
zkyjH8Wi2>k@7}YX<MR=?HBd6f;7@~(M<Av)Kjz+$H$5j&c>Dy!Pzei7d;-4vP2%N8
zgAOOJ)@j0t=N>mmK`m%$Nyddk3setvyNJhyS4j5Y?g)Q3x^?$0%kKr-%@SdvnwoJ|
zgzq*gI@u^qKx(WiIR&|*^gV{gF|0dP8rTL8oEH`08heutr-wK9EzuB`73psFt*#`p
zpc}`8g$Sr}-9D%xgJ7s^P#_Z`Me5T<Y^?>q+LY+jOhIoZyv~EEsaUoMwjuj_oH=Lc
zJdvyJl7G+VX4<L`kP^>OQInEdJ*h#ik8}8~K-68z>5I^H0R8F*!$Lr7RGt=R{6X%H
zv7@I|90yjGZZ?7yNik~19{x^sZ6ObhDUc+}dfY1+yg+Dc`GW@=kdy+(S0W-JZ-7t5
zir5-kJ$*cTKdkM`^Xcih@q)Lkg%a$lNT3510c=a#v0(<;dMytkFK!1S=V_OR&9dUV
z_Ly(t<Qx-s)Ao$yI~!Q`EAC`p{qKi5yZ>JDw<Yv%8^#OU`sk$dg%$b{cfRLRRlG1r
z5$&>9Eced!hN+&Vnrqk(n0S-Vt|BI){;*xS1{#^biG#4Lxpj+k1`r-Kg*>r0ezuy(
zLU1H|%3H=)9pV;VLyQh7p*C<~C9<*#HVh~lSNJAt6iwO)T(idO4QS>0?>GHLsFqCK
zv~t-R4yHVOCdzDJh;cXZ@=uYzW2F7q>fVaoz#@cf5EQ<a?X#X2Y1hY9Yo9*;IyAB(
zNb98-xNWhkK<Mto{nwO2>f?idc(Yg)Zts@~-mS2OmY{eN1%~EbQQ58nZ*1<Sko1TZ
z_MUTL6z`2nS6OE4Z;O#+Rc6%TcEL@?xl}B{f1s1AyS89+nGENM$V1UkUx~Xan59-p
zCe1gwa^fEC3(i2b?qqoxFc^vj>X`S!+dI@!2M#vr1Rm}E<R{NS?pQ#(s;y>Z{ahQO
zYE0ZMuaCc$5RJ$D=yh80ib+BI>p@E5hs>sNDX>_|GAq9kau+lqwR(trtKsqSG`X%4
z#8w$%n$JFTkYIQ0*nS$1tM-<x3QDs>f15?|Wp#+~lnLcm4Am6}zf{SYdmw_z=K1v7
zDx`<v`GX`=M_UVquND$Amh$JRWHVMLzzDm_$;g6!HJ$hFy#4m9O4EocyXL!r|FQ>a
zZ(6Gx*!vy;0VpEO+i-HVlYCrER$Rq6(0LmA&j!O)rgZ+7U*SA|4SDj@LcPLP$d~4$
zPJa70lTVWAszmcZ$N$@;{}>CXrvB?2{o!XSo!Rd=TiWdznGo+_8!+i;n$fdU%Y@-W
zT62S_92pSASMW5vV2dyrGm)``|2%&G_9ilV@Rio6rz=bHnu1}MMfWx3<P>b|$fBn{
zr#`GcGbowv9Jfn%2jhP3HvDIZ^3fxqq%W(JERC{Y%OrBEh&qP!@l!u;R|n~BgK(y@
zwH8i~cqA1!ao9CiZ<eNk?W=8fNzQf8-&#ndJ*3;(ZDO*|ITgB(BAE?|Bk=5xJAP`#
zcbRgU@hCS*Bz^Pa);2yg)4MY=PbEvJt?%<9R7B7{wDfw^X5^pL+Q0J~VETlIU%`Dg
zHa0}niRNE@d)!4VHnfJZ_TBF|Gy)wa>A)#IP<54eX%r`$&&j1iCH)~<xZJ2S$e(<5
z@-eBMgwf+i9@Xmp5V^cBvElg+kf2LCV_c4A_Gi~7i?a`8K`O?W;BzcSY2RZw#|m9-
zB>}05A{@QAIDA;IX<~Aik|-)Y5%`)sTtZ>X)}&}|f|Ejr{H7Ag6P{5DN-fHjI#PY%
zIyS9=%^tz*Hh5*#_n{z+9;4k%Xmt{GTn`CGlLzvgV9&H=AF@+Sa7_u0|Eu4jHJB$f
zkzu%mLbN2@PptnB<W}rE&iL)nm|RSW-{i5O!Ta>anPX2jJm<uI53nP?njZoabfD<r
zu{je1NEheDIj8!{9Tu^{9XGv=^x_J56VW~1d0Wty-+2#$#dea++o+4x*d{b}n|Dox
z#A|mXc-2OYA_PsDtgQug!3*j8H8n_yK?g|sU6qAZ-vjp~%YP-x0ws~*?ZjThETKNP
zLkZe*IZi`0y_<w}kWZ@+82@~aj}s2}_9FIaU}{w7#~Q***Y`x8{nD3xZ@_wsBILGw
z=#=j>t3Y-+PlKo4daK+iv>S|pVvbqGG~DfWHgLfD<%3J7$Pm`zFknjaM_};pEBu}M
z;)jQUkS<9KqJ`vvwc)uEO5^N<7E=*^!sP2YdUm=Z5yVf^mq45NoEOA!!QcP;_sMjK
z?494hi3>!cXla2y4dPj`z-Gh<?A;hWBkgv7Ou^xw_vO|NE+{W%JIj{ZUY+%Gm2`ra
z(!Gx=isCKU_MD9>tPg~~gnDSFhA9vcNMd7KpLs3!>W>UO00ZZ^w;e{f$0OT<m5mIV
zpq74Zkhsqn7Ozn~v8R<<I_QU;1b#U%@&$mFD`VTgFTiD0uAkjpj1qBlce%%ZtOYQ%
zY#k(*MrD)|GF0bU+ATQQELFOLBaTdur_;ZHAe&!cIYXgMO1BKL6hdkq7NF7|G;6VD
zN0M01&_?ys|4e@9E{%+hJfGq%AU5sdR<>Ynlog&o<h*Ki1ho=ygHfytnv)|pGRIpS
z5<65-b*E|Mm9}x>eBeKSFLg5Up1z`XRbtcJ;+gmbKy3kx-xeN>+No8iC4llZn_s0#
zUxrk^F^E$2exRM<$eHK63TocO?#n0?bf@mXEB>kct-veK1lmpc$V}=bqG9C{Bcm|g
zigLV*Uh$xZ@*4*wg(;h@AIo{=-)$#2_%4(=M*0;^Vr!Hy5lx>-1Z|vHP0?X9pZV2g
zoKV;9s=oQRU$AbII1a0yOdc8jk-jAI^BMJtNtuU(LY}X7vYlP(b3NlVb@vR4cfh!V
zoIzY(pa^giISlc<L-ei?&gq_YWmowz8AeDkpj8S*OxbX%daivk-i^#at@xjk41W%`
z`CODkz!7*L2ug&S*`IB{uxCk3A12^yAYXteY=)$vH&#>cObNn;Z@hMqs+1U7FJCh_
zUoOI{HRlRTP1iE9tMAI8;1kqD6$F|^K%$}mg%*pn@Y9_BW4Q~Yrp-b3rIds9S=wtH
zabQP<{&nl6v6~MQbessiZZM0Cu1z7-N27PpKA?c}nq6$>%HipA#u(hX%(jaGCxLe`
zX2L|xXUX=`JJ}D}m@R|9M*<u%K|G&rQ+5*$bEqqkp8pyEco>IfvX~@|gc&yCsqCXU
z;zPP*4NL~*n4T6(=3^@I?d6=<_FcTRX@`I(i01CoM``2rXdcSa*SRT4a;g8hpG>{+
zGU;#sk!IpP+S4ARLD_0~D}NnOfmKtftdr{JD&v2ePAof<r35{5v{LL(A_=YU6<to3
zsvb87zY(zE-hW_vpgkXeeOTAuYZ9&V&fCpf@Y9=&XLq{S#FVC~?KlMNTLW;-fSte-
zp9G>9AnYTLj8-{6_tF7387xU85re5-4Mvs*3M(Jqbjd{s81Ljr3~k$UB5?p_S+wr$
zrj_>#Wdl<U56gvgRvt2=0WJHCV7e_Zb7>7`0tL?eY2o!}H+=VC=8B`5u;T=7X&3$3
z8E(N=#7k**V1zZZIxH%R1ZIk!+kY5{RnUS&jcFC2^@%8~=q<e|_)JIB1XG~4@=^Li
z0OSg0@DAX=-I){IyNt8u-)?vtS6uX=;r0_E&%jquaC9vA)4K0mg)yt*VTC`OIWLc`
zK&QeNsr(V&5Pc>93AJ<I%n)2kFFi%SYA;Yxu&zBf=EZr;>{bPg@)-Oq>3+{qv-71A
zGzrW3y9sg5h^G(Uw!?ypd7}Kfx-4f?y7*Q!_aj2c`@0%UGB)mchI8KXPr5y8<I3Cy
z^k1h=341=`$TIo*Rowx!5w^1$)RSPBwBz!~!X;yzenw!pa9iF)bPKo4gH<kkOI+@S
zO*1>U)|jI;O$#EeJ;J<1mBuH=#NciO>_-j6G8K8X;^hNom2bmqWARx$Rn;x|2~Zct
z$$7b(r5KrzK@?^1$05`B^^kNpoNQ3`zLp{h@)FRS`%Awxq>8C=F4xah>XVgY(&qx}
z<$C9@*Z$||%H)}4M;TXCIzSiwm_}iIx4QO;1DPT**RI43QeihjR1!37;)<btY-N{n
zc#biverq|tH15I^l%AqywrKYKw+{h7^FZsv?K1dleXH$<<6oR47j4YVb7*-t!~!#^
zvXcJ|t+=$I2rC2Ma7=3U=y-zt?kzC^;9k48fH%NUT*M(Q%&47r8~qk<x$dBx07Uv$
zoSpQZ(FRNg_K?FlaviduPlhT=UVHh{FGq3-K23Y>OrPYX{rz7!u0r-#&95rZk7#@z
zk>Q1^udYJpE*-=tE`GO0-n5;m;7N}1C36y-g@?l0_~qtYe$+-n>pY?>l1Y7MY{j*u
z<Eg;yoIV@tF?wzR4r%1)eYkn{Z=UlJ{pG{cC6l!Li#Yzi3C9(_>UktjD;>4aL!8`U
z<aZX+(kvGb=D7Z)6Nq@(FG7e=eMa_6M?v}w>t|EuG8cV!x5qw65+!g{s)Y{c`Dsz3
z?+!WDJ?<7<Y&p+~Jt2!P$%Kw28wq#PxLvC(!2wqyU>y9!5w&J@c=G~Fq?+2UnY>XD
z>-^ET^gd7PCU5L$Ka9RY(L8xXibUySO7O<WThP0Gl#$;1TpMzUFHn+~K}(XDktu6r
z*I{*p)Q<<sT$2)%#G$`k*iBJ=4IhyZzPgbHH6=Ir;W(@bQ%n7(Z*DbUMJ0?*Qc6U*
zGC4_GLCZZ>;q4(ZEX*~&#y&EBG>`S6%<zcGK{HQ^=7$BywUDWES`DATBq4r@K4Gmm
zqfqR9U4|nbWVa7@6#9pI5|yyc-AxIz{!;-m3V&&`)V%7x!6A;2yzPqea#|o$;GF*^
zU$DOY$8J16E+9?x^r8BDGg!B~vDUj0$aYJO;ECU*DPwakb{EKfc%Q<VYF*;yTN2F^
z>=Au`-3B+{Caao06D0bD>wx!SDrMkZOHhP3&zpzTb)|~saPp#$!O)@A5F_|WVPqG6
z6Bd#-ougix%>&|6yQ(_@`&%F{X$Mn&V9DU=#3;x*nD<C2@{J#|E>be~HNJwyGz4!Q
zY0a&Z3BT8gBt28I&sBlXVQqlM{~`M1(L?o8LC`)_a-*f2BcP!yP50W%Z62mX5VONw
zgeglVkf-p*zHt+&#kNzyKbmEPN@#_ann{v2_HwB!u+L9LkL|~&a5~{H;$?M@O)^sc
z?4Z|!t_#ly@GqgJte&e56oN(ZB*7U1%$UV<E$B`IzZBf@c>f3}5#{_g{me-bY=82t
zTV|(;!~KdXLHL$XQsF<)ISX#-xGd%5I>i|-(RP?ez($$DQk6lU5G2@%*7zdz;No$(
zLXt2{u-%?K>QXs|XZo<y$We`n)a`^F_gduC4oqZ);QTM`fPH*f$SRjiZa2l#8)d&A
zV_NH1-7?@I9Y_xd7c6__fV+1BrQ~o~Vw_w5R#mCN;I(r-Z}J4Tfj>%VUY5huAQk=3
zgpskMM`z7Vo#y5U5PJv2BLGb7D*GXoIqLYKHzDNOfGRkRFXub+jQ<Rg(J(iJwPinA
zdP#!b4~R4X-dW}=Hly93B{LYrvm#z4P4?0ohoIbq+m!1Iz5j6*_^roUbUB;qXdrf}
zJ6UF`Q)5wSVUsO1!^IdHY;0Z=MY9UOWAW6<Nvtv8*w9L*5?d_2)S3?diRQn_wQWAV
zbmnh(1qY!{i>}rD_@iqJH#f&e8Bvx8ExU7T_%G|9Ni914vgX{6P{t!4;5Glkw$||T
zX<wJRjUPcy(h+;_E35=mzrr1_yNP@^Yg}b|+=3cqORY;GH;V|!h`ltRY5l^J$=}k|
zP|_CrOwSMrjNAe4hKPiOO83r1*Zb$dqn?WZ3$6ZL72V-+=ZU6#rLXA3X9iZ3BvqgQ
zJ|9itq!Uq}nFs!sM(Ax7sl(HP4i4#~xe$QC>B^6TFy(;mpki%jFG>XT`r>Mai>3Np
zX&xG6?8ZxIvm|T0ld>`N|4Tm?F4(f{GGm$6DIQ#+IoaqaudMr8(f#VBDC7VLkZZq=
zj36?@URj&fBAkx}lawdgG$#!dCiPt=ChYWp$HE}3<dF%T{!m5~=X^tujhu#?M>!Q^
z;=na&ZxiMMqW#AE0lg2g9u0%4d08$~H)`Mg_^=1bTWc51>9y(n_&H=dyPN}@q`I&e
zDboFV;!L<Z|JP@csGzLudE3AW63)j?piKS&{{rkaJHt9{7Xv8@YLl9T`?{%gomBqq
zm(c>+7N}^8^9Qul44$d7uE#*TQY5)|JE1=bB0vGLy939!B5X=Tf5tIb1L$-hO@d1&
zCV>9(gPT6DiKy?Rp`wIFJB6NdB3wfnBT==2NGXGz|DBU_nM3%bzRC>Y=h)Y3zf({=
zMB|v499MCG&R8?3ZjK6@Ven0!45U;*rYQf(b3Uueh4t;fNMBG_a##k3Yd_pj=nh|S
z6gClJd;gUum#QDmY)ne4f?`8Vc=o0|dlS21CD6I!f^OY#6BR>*D1nI~Qxq=AqsZsf
z0jZ?8)ofC0!TUXJEpL(Q4#x-zD<*--4WG=@Mis0DR0}$ip`<ZqFPjX8>nfG@2|k-&
z@3mygh=mrM6khkEPs4t%$-7SD=PUm9VzPVB)(EPJy5cvPyw~MU9j&)sCZ#Ztl(PJO
zXMc4!T*T^VVlS)Pg8Y$^%TT^}VXAMM6|3N&+Kv%dX(-+`>;~Bjq*Bngf9wHKBja|-
z2Z+M+aZgvjHF4CKB~pufHt71~RDP85y1N)!OeYq%yamO$E$iKQlq6D*7weC)$-fV*
z5N@0MTYY}~_1md^J>MI1LrJ0=3I1^w;9qN2W!CO1)mB<}Z9(SIwaocZkmg~IOvz`b
zopP0r3%MTI*U#H1H~rI40<k)6VNWEHEic>KdPLB;YWHJ5y&dD0k<+)gI#bk@{J!4k
z(8w?jD(3CYH-_xKOI<s9F)OP6PIF#Z@-gbU$GfWohK0o)_p?Dt&7lvPJ~}c2LuL6o
zI697Q(r=&T>3!mH9v2?u1ceVeI;`_=|042Z?Et5u#|Cdow?bM_$P*l9vRx~}QH4ct
z5lh?{kc{5vsXI^(?6>*r=nYbGerb|yx|uC)OT?VO-oKq|QE*mZ3+lT5vHLF7+13jw
zx-!YC@#q4f+mRwZPmrrzPaKltzd7}5(6PQQ?^(Fj!rnEhNqIhCW}%cAGBYXjCoTw*
zZTQW&)coGxnqC=1CI!JW3BCFNuNb?_A^6U?(6+@&qX(lZg^X^j`2+8UudwZ*qkB3<
z-fYbpheGDh4hS$t{p3TR{NG2>0z4-PG?VgmZZo!LM~wg$`DT|pSqIucU|*Y~9%a&%
z1dXvCxqyw+N^55jDfFfcuEdBypxTdyB*FAUl#jy;egprMzXCy;{n?mhxJ%0%5I6mE
zm1_V}d*kDUR`pav`>>k+oa$P;D#D>vr(<3z!k@H|?xr<aXvw6S<o?p+-e+sVHSbb2
z=rkK9{eZ4nqKcH_^%Cy-g=6Y||JJdJw%NC$BF0zu;ApDntxPKf%(X1Kf07^xMh87h
z5H;AudCO`qnQ9)B{tCRZ2PbQtG%j)k(i1@I2^jgkgKR%K>S^nLWlq1A4@F;7c5D1h
zXRnOvho6J==POVzd%L=lUc&U1+cOCk3et%N$?L4$vf1r)$d63_!YiCr&E}lRW-Qf8
z@1SAHB+0R}C8c%;goU8d2DLiXjN!x^v&cY;kx`bGuYSA2r+;LbcP$qt?(lTAdzE>d
zyp(2s@K)3#a427I;=R%r&Tc(u6L5W^X0dF|R^<uGiKSSL;cxk=n%!tSc!L2ndv-8O
zBR`x`T1&GUq)8>uB!E7($0QcIigVmZzabdg1m+sQT@HP!klO>*HeH<0|IfkJXr+ng
zT|qB}dTK!%QbwV87ael}1HUE90F{)S+sP@~yiRK*sGi3^>E~CI)k(h-xlngO^F8U!
zpjb<8lF@bo&W?J7xTy&P=X3_8zA`T(d#Mysb^qfEA&G>noBkm_P#EckU51Con5V;&
zqHClL6Tk(%$8CAH9ep6GvU6G2xu?P{>;1mWT8(93TKI|V=llgI7H27Ff4DVY=HZ(}
zro**>5&qHt45UUuUB$LQej)~1$o?hPB~Sk*|Ewic!3(l~XVWJ;EDe_Rj~X#O0%qVp
zOzTl^LK7q%ZROLV?r3yzIMM(}@C8iyZBGQhPG$P)vVyFwo+>Zp_75JZmQLOmi`j;E
zCwIIHyvH5pDxehxfwz4WTa0QatT#2K>@x!WFWtngA;G428`#T}(N}BsteS^#b%6m#
z<HENFaFS<+iTExE<mRSrb2~~sq6yLBFoHwuSeQl62vxT|5ZUyat&mKM=W$$#sms$Q
zbt+|=n*7@mq%>j^8$N>Y-HACPzWc!DXpy$^{-DYJRBbMPII(IxAZ15nZga=lX^nWS
zc|4TfG(51+K~QN1CjvRk81VgIt({1*i5N2mr)`)U_#hTIndfYlD~PI&i^v^Hl&9eb
zLzt?ft6}{Yp?qWdQDbfsV%3C2@a^h{1?@^U$L}y1fCd*398CTP{r#4s^4osTLzCZQ
z$65T8vWCanzgNRcdcrFb*P%%&f4itW^<F2FL|hNKrHrtwESfcXBHo<F@Q<mC*hp?<
z&e>~F#~pI|8x$2M+2P-5%?tW8khM2!TnK~$b(`q2xWT5}=9h}`T$ww4;m7(}J$1q*
zPuox~JuUPPs85GFXf`X-{F@(N!My<O&J~f+qN&$h295y;?#p`oKkB}@UBp>ZaPljc
z?)vr+#V!YLT9E`B#vwB?f`hM6R7g)`{}A~tKNYwyB&MEMBiYU27a&G3=T%iz73+5K
zPm+kdP+{_9$4@PvI3Mq^Sa#2W0xc3GJ84$tJ@O}waBuV?^z5<3Js(ulCikxGgidRu
zad^de%?>u}C#~D4eA=DL>@5q055{2pj_)&zH~hAoiA->l6tlYjMm;Eh1B8yGx&r-t
z^6#7sG)o1ujtllo__53RPZzHAw9fuEk(G&7X)&6z;vY8N2?|sdUgwxN$u7nlJ$C~y
z=}iWDT$o;WO8!@-{H@qGSaTLP#&Pe0k!T0%>G*n*s-$yXmIpKS!A~qlHd`1$V^!Q6
zpp)RQ`@WPv2b2w!gwk^B`2w(CyiZ@raTxM<ukviyS~Q$_9Y7j0_Q-9}E3d7iOm-45
zxJGtVIghbz)GH88SU&)yO$>xd_P2*n4V5=&);Zj3oQle?_U3<YEAivOBgIhhU6e#i
zWaQh?3U(>K79Vpj{(#VV9o(w~9*AE`*MHL{nS<Ha&_Iq>f<^c!zP}zv*ZOLs?(ZJV
zG9q$JD?d)Bbo4JVMCz0=wqt+hFEFi4YNw-CRSxY@UT@_3*)uWD@i;fpmsPrfO4!nf
z8`3{#lFF~XKtPwm)Ds}Jbg{>?)J@&E4+xB)Oyw!+7D!|X$M-mZ>VkcJNOZi%7~WO!
zCql*Yjg6CW(@j$7v%fA}IfbYX)%5FVjn&NPcFp&vW2RDd8V2wD7p~do5g=uTXL?4#
zgRB=oq7tw!FYSEiC#BtvU>%mVFwB-c^w+3&+w)eOt)sFud?k2y&8@b<GaY2TWd&eQ
z02b7eX33y?LDVeoXqi>v-~ubWZ9IslX@CS7+*K_A4_>53yjpk2k9zot|At|9kUe%M
znVc#!-Zr6Sl`S%e)R1nu#7{LRpytNxE1@8>MLLhK=_VLP@NFQk4UFpE14%%@5SrpN
z)*iiylHJA#W~r|psvMV20?BdNz>M8Iae?MuYxR63d%bK@u;;eek)4fSs)BTsvidAK
zw!&GNY;|t(TG=jFv0-~>BUR~Gx{2c@Ft)!EL9(%xnl_uP83IAOH>%bRjT3CDeg->T
z2@J+0U-v8)raSJO)>f_y2-;m$x*}@R?k6w5pB5^be{iaAC2E2GMa0FHj!W1WbLtF_
z>kjt}_8!z<YYt!kmf<e`ch1UqvU=k@Dliza<3>tB`ThWRlV9Nb_Qp4-+%|<HJ9%5Q
z(O|F4OKjZ9ezoj!cw*1InO3S&nE*e;pWF8F&$LgLN@X7-8`HtBP<({S{tTpAPF%mC
z(wbnVBBjXp%*n31#!~8(3KF7f@{SCiTtP1v7bL3%1ec{2K8tKZIs@;XqAQz>vwv8)
zrG=qrtrN)tKI5$#x+RSBa?53z=P_+wYIyVZI~*R4iA!P)OVo~@auT~ef{<8f9qY5M
zWKO@H1po7c>DSj8P^zvd{uqIRRxL$}W4bt;MZCuslS-;23xufkHZ~_iR8-W)*d<0}
zx=A#(_k%YYOpK;5=-LBclT%cCR@USNeYPulv$I`8-X}UT_OHrleaFZOV)9o*6P405
z_r%1&CqOCL+H|9Ml$b8~8JMenO(8CKGYt%*vU^)!C9BymFRYf<RX5~e1i$sXzcm7i
zLj|r*+H@y{@l>m{!UK!<c5>*XF`jDQdZ%(t!UmN_5e#uQS`PN>%d56sB^MrlA}vOh
z{1}~Y<(~6Zvg{RD8U3}g$5~`(?2msLaFtTLd!)re9)7%*yt4u&w&0MZ6Fi`QRL#YX
z^&0>R>WZ!p0BSgUi^YbL=>WO5#p$MUdllbGQ^!X)Y@Gio6{?#ZELZH|!9^28pyP!H
za@Duz%0&v*ZyuiRZ>oern3<XLLVK=p$YOj7u$-g9G!qjI#b8kg0q;2n3pW4OUA(d2
z#LQ?MkUSu^$5RD_9kwN&E2o1kecHIPHgFbX>^t@lhnA_@oo1YK%ITzrSf;sONtEKR
z*MC-J#WF(M{#yVs-a{J9yDE|>v3~!Zh}f_?mZ*9%Rt-yPxpge{kZM49uxdD-GGH<b
zV!T_`U+P|!r(!vlVr=rDI^t4?X1WSR>jIu<k)Q}^P(8NO;sxpFy1*wzQw6F}m>?mA
zf6Djm^ei8)P`Vn+va4nkrvZByAmY?6YMhUjtR17aQlwW_h!XS&Xh#J4Y&4W4{58+C
zO<3~QJ#4(Vx;r?3UmgyisEHDMJb-VBchYMZnUqO*B~)X8Hr`O|<lxVqHMh_&7#z8w
zuH*}!w#Tpy4w{sanU<+#y=~YjS4@WjWZrFRkl9>7EMoq}=iU<Pj1m7I)sM$lKY4(p
zIiC_`Zjs!v#U@!nM%d)+L5-mElSP{Z?Ke#<n~=%mM2K*I8l`>&j^9EtMC*fqO{7S3
z{`#+fq=$w`Z=jZ36KeTu<5q>mn3AtKOPT5DsD`gEM^4qcuuUHV^3v<(Zy#wlu2j+T
zN5Z$b4Ue%*nSZ(91&|$-U8ovMl0@)KXm~|i1crob#f(Y9q=;wnPe&LrHr(3t-u1H=
z+jDD9%nQmGWGudB<uL`d&`ARI!G964B|J<_EFS+9!xXk>fU?>-_lLI`BJPSFmXEE|
zz7LMm2p558pyiKmcJe3a&wdz|z=~9eU+`*SVIs!kr+i0iV%`l!xS1_f|9JkW^@5Q=
zx#<Mp6rU5T`55}~^j)dmgnM*xxSo`0Y!8)^yZq0tA4P+Mo#??+XuBm<0g0Et)M&ea
z<YqM9s1+=B{*KZUK05<DPhs1unu)&X8T-jw_L=}ce|fThy1cwRwehv&3>LTO1<kPX
zs*os^7bHXVGD2rO4FxBZXG(bix26&=_#mj;X3_ExqpzA@E6R_S;SavPDA4}}%WvcI
zQ|==9pLTn4*k@W23)UG)VUzr(;f(}m%N+uJf9rvl>9IxLsQC0atANY_y<C|VdHNYg
zptCvAPfKTs?0Rf$ta-WqPzD8Z#la)!){0+$FN`O1qM*hPa~LFKc~WrwH(F7nP@s{;
zp$5U?ajgZGwIGx3vQi!DOW2R*7%?7bG+@~_Mkl;gZOVU0@i-0l@I{;;*E(xlpjU|3
z_g=BkAS`Wyza!AJ0XEv_=jX0_f~LvPg%(^-oGAn8Fv$psbKW$EuSv?~fje?SJv%jW
zaN)L6=pUl}!fW?ltFQLgH1fnoz7_DnmD;>2Kg${~QGWy(mo(p|N-jlI)$ll`oK+)<
zx9X|4G`w#A+>xf%H#9KFwAd4Og2a?pi_paQm?|u5wydd{X7jB)7JjE1<#R7r?(At_
zG(}!gkbq;rhzXe4wuHdcKS)fUbuMsR_r-SUNdw4Pwymel!|uWE(yV+7h<0{Wy9&T;
zZo{>|ggH>gTjI+<YkJY>H>M>xbWeN-f7T4}$g8>&=rYIpNhVhcAian@8tKmfgfi}a
zaOK-<x{p1G;(`&Kbu{;#-CfcI3ws|{8x!BZhaDZO<mXNeuVPmapNI%}NTa+oFgX@%
zO3YLCvZ}ONpOZX@z4J~DqcHkkyK5;^3e|!bIp#-9>69@}DKrw{_d*CDGk|hnWK+7;
zOgWl{`nn1M^kX_ZZWf`{KN=fe2dcIa6u47KYCVGXT$rT><#Go@bSvbdG4q&-z@kws
zsDXsntIP^5m2%s}YvS{*!}46KTB;IgqZG(J_KudR*BqE=X7nXN#uJzujP63r?HwWZ
zWVT5`DhJAQ(OC)Ij>ZeoCXe$GEy06}mxAeIH?$=)i38Py*LhV8pU`G8t!C==U%eZ(
z4gnAIqNqjjteAi0$5pHu>2<bNJ14%BZa`W0;d;NwssNaPs5~1kO^IO8zTYG-bFrPV
z!vlACdhQO!VvaK1M1M5k7781g?O!HQlY5}F`ZZ{it>2#m1jy|%^VmEg5!U@mnss1q
z78r}ayNr3>E#s_*xf!60FPT&oes3V^FU@oZ^=n5yoz|JY9FN+?NN`tuaw9g?6A{xX
zs*Mhe(vD=Na%z3g0I<b{y7X$|pL-Ve?w@eJjMC$1WQSqW9{r|IYU2egYego?`)h{P
z_znJmn~$8>++1AffNoh{J<!z!Ffn_|mUor2WV<E6CMCs770=PQ0l~@mw=jOqf{fi$
zW1-3z+U9oWlNWzPe!jdVZFTGE)>ZL+X9UQ$vS*fg^vBU*p|OH>MMWp_f{bFfoD2Kc
zWF9|7ks1iY$uzEXNWj#m_jYjTWq?n>!NR))i^E?E3UO4>P(e!_O9Y~(@9xt#!K~Lv
z_sWQ}eK{tWeOAigtW{k-gdPrM6&dk7@$kvU#P(#AWs`@+zQ<#!M`Wb34ku=S60my8
zcWgd{tkC7VjwG*|Z8WXg($MjI)`!i|Fv6Qbn~RYS`oO}Y!-g}HZ)Vqi*;#HWw=jFT
z)aeFRe0@-vCZ^GF2}>G1F7EU~h*~OyAZ$yvrYW0PHY)N;RFjca&s~$oyLYE>M@^iJ
zJ~Nj3w5QR2#EQ40q7&CIl2>ZW2o*g~n7dZpyfhAvFOqNVm9Nj26%~9)EVx&WuYKAP
zm0dkq(4C~;yt)t}yaQoszqo&77Qhp$Ce_G0@}s+U3b*2EZ5aUdvaktl(N)mN_~va;
z6Z=C&P0)~4R6StX%I_GH&EYh+_)5SDidvg8Rkc)fke77%`(3o#R8+Urr|z$Y2?QLI
z!Sk}}@4D#F*B~+LCz@920BhN2hE%|WfnwWYr&2D{qgUAk`O6ZG_f*eZ!RRY|dPf5w
zZRf3G*`{V@4mLiQ9XSNml(0Oup!taQZdWI7Dsl;pR)fDO5);gNW^q^4VQ&e?kzM8p
z{m18TVDiWDhS<39zqa+z0?s!MRq@BZpg!uN?u27MzO<S)u`-mZ#R5NB!NxPg!y=1)
zi&Gb9mmRtr%+!35p2~S_2xro+P6qplery5QJEb*gNyr>bX<1pfmK^`6G$OjD>pNim
z$_PIIYsh6X%OCl(IKIw5eefbE#C<^QSK&v6kO8RXxQu<!9n+2xD~|(IAW!b{UpER4
zzE(JSt46xm(&8cq@Nqj7Gw2%eO4!DZv~f#~={j;wE?+h}0D0lB<@SuV*Ez@kKtXU!
zqUct1#}f*ojbVFBfFtOjJPBJ{<O=mNnkdGh^%lasiz=Jsmn0xDX*`Mq&q|q-1=YCY
zz2<lK5XKS9n@{nUZKYNQhI>r=Qz`OCscE#we_w1(-B(=7TuGN{oWZTw?t%gr1U-tp
z9dE-0b{4We-0ALLS{s0zFD`HlYSqBG2Zd!U7uuQW5m@$O(6ba3=Dlg)E%~)b!)U~*
zBh23Ku;3?JNDxj7UF&nTri$zKx7|*@Mj;T^B>3Lrai9L(7e@&^ruTf$PY@-eL;!mr
zNzlnt!T&6{!~42&rbT6ZzhLW<K!D?K!U+-Wc1K$nL3Pb163K4~Dd-9<qdJfn*wgyc
ze^@eV#^Y%_nDLj--<qPWC>Uw`Z!;IE7+3zWPWt6HtQo%hb-N{KjmDj~|C>Zuk^hhf
z#V0v}w#```(?EF$EIVLmykkax3#P4C6%W{(_DZSoEg7*-;$!vr#EUk8cQ-Y@n##J8
zs&Ma#;UHrE%alW=N!8kd{DQDgh;3CvTZ5fkwW9uhx3ks=sIzEfvSK@fZxaq1k^SmO
zZjm|Yu<S7-OS^5MY#jzrb2Ir+O7!ED)E?fz@Z*Zl`*zH?t*B<O^;5Sw6~w-pf{>CV
zTrCwa!npf^^P1(NK&H6|h2!TvpYHSb93G`WSD{{4+cscE<AJ%6kMrwan_^YZuQL{n
z$4wx&wGe^i&%AcrEpEEk`+2`|(+#W9Nhi|AM>n)pxUsZmZD;lG$bLZ<G)$0mH(k|*
zfunq;Cl@{xmV7-_hWr#iSZ&z?YO$r35O(pO_C{J!wQL!rEn(B-59)UZi^K(~<U$8y
zYLEW@Zat=TZYSC&umHj?Ae>aL&pZG3@3+*TQ)kDNFwd+e8muwxjn&fEvSUm90*+Db
z#fKX73V056kk?b$dm;)DK~XD$?!uQu`m)JQTBpgm{aY6BM1}JvBmVk@G!}T`c~*di
zA8;)vyT+z!`B2diAPoZph(?pmt*-67+4~y!i5He@qMqG5)!!MLb;NcyVL3amOYfk1
z@iH5CsU1E8CBs-?IeK)oZ9X+U{ih@Lu?KnVeffHQ!0D8hp8n^mE(iSFyK&>=T1WU4
zr4=#bIyWi!tXH}?jXt<2Eg(7H#m&HA+%X}M>CTdls$$62>gtEShY4vEOX=hRT&Be2
zo?gqrNuMlKzpvD0>6ztZ`s3tdqffooH=Uxf!Z5ze8j{H`=+K?2XoF%2wWauCswwX}
zomnrR$H4xeWoXP@kpU32)=%$~(p54@B`ZdMs6NSe8FT9~Z+}Mk_B?f*kz&;xwxg>-
zi^Zle8*$KFf0Rodq%7twz4|4Hb~_hc_0-4Nxx9k*;_OVqvSJ*Ur=CQc^0@YCt9RPN
zYyG$Pglu{b(q736U;7tvyGD_3GH$&Jo96Qh>F<51ZN}=8Y+U_j$NNZco%%z++3&vx
z2e6sufXr;;&7@w-dwP0$jfb_(zQ@pw%}t*tZ(eAp1l&Eg;F7*4&ymzaa`)K9OsZq<
zw8aLZyKhJydGje(x1GV0oStXZN7%bu-M~ufqOrMo(?o2c;!*6pHX^QQzl*yuLr+QS
zX^m1d8VWVrdSs@FGzP7kl3IK*{ioVx#mRR!eLGJ&c`~xWf<}l(Z{>bnA9@04;PTQ(
zd0t-lsg<Gky&Y3SZszSvYM6vhES;o^`sO)@Ee1X%GekC6xBb4IqKe*IOHAIV49wo;
zFz`ZMzfY4q7JIVm<L<`x(Xn<f?`_TSxD9{ap(ISHY#Hx%;`!;QxA$#L%A3R|v-*mD
zg)QWh+|&4(n#2{;cIj&=qtEGE1~QKbW-KduVwX5&rcAlFb6KF=u-KzadG?BNyZN0R
zD8Q1cAI-*W$G6T2vS}h*M!r*dSuu0`7egEE@eZD5fRBs+CsSg<$@1W<Til4tca%n#
zI+(iep>`UwdD=mIMu_L%+~~T?4``9tyFKN=i7%MYkw!t$`u>B~Or+qRgl)HeU)uJn
zgkFsS4S~T~@uj&>50w~^NorT}{`@}vVhumvEzlg@d_ip9_=Uq?OVMfedbh!so#QC#
z^-3nmVJ#y_iM&+#Un_H33Ub}2)6?0X8+76i^G`O`{HUmE0GE`I*fXj9*YN`QYAm%p
zKn^ugV6{?pYbMumx7*W9vF+EsKFVjyCcRI3Jvt7){@C1gC6n8!6*~3$Zmupuq|S8S
z!?e`g05$E=1zVySkK%9wRKo)E=ck8_k~dJh(aSlndLOZ|Y>G;Yh%A$kGD9G)7cXCq
z15$J-Do)~eL8Y8b_NS+UNu6SnnPKB;uV2ra3)danr4umUJTB0%Wdr&N4gp`vyPGD5
zTrKL$CGb5%5`N{)eG%&&sGnsYqZum46-D)9-%3W<mM35r1oBEg3;UOS(QindUXYyQ
z)-*QqgmrY)w7Ns>zqcHTN5@N<3p+dLdPv{o#=iEByPpS{3874ecF@wjPu3gOUIf$b
z4jUp%G41a2F|@e@C6S%ENmaEZdm?8@Fn`S;TF$>;=J0@Cq_(bZ<?LfJ1JDu$4#6ur
zdQ?=@jB=Xi)5LW`lc1^Cf8t(-aCm`VwD;_*f2ph$<?F}L$Ir>REM)LR^?*A2ZCKwO
zNUTiUWEs69Em_Ooy*(k5MsI*#%O>S_GQ<<t9-^SKtev^EdyI?jIOI~V6SPQaq<Wh3
zA+U<vWKkvUU8wqT==Jr`QR>(W-{~L2GfIktK4w<9ujS`Ne7YWp#sS(53wt%>VpT`u
z4$V1zl<>$ax;y6W4aN@A>Ymo>5sKng33>k~G(`LLxn!z5iZec9;*%}n>vJe`kxb}4
zViB9I^!c6^P&;;DPv28DXRntFCfr$3tG*sbz&^5tDIAF=!mBUR@(O&B1;CtXk9X(Z
zBV0j~IIGjZ1-SXRf795^i>3#-&jn8I6iyy(9Pn)ezRp0am-%5lRpD)xAWeT3S_Xmf
z+#U$of2irlXypg`y#v~TW<8srJ`0OaD~+nABi~!BqrSHM1>Q43BbMa6ils{clK`xu
zV?S+fb^r&2@5x>vLxPKJ+<)B9k;R`#j}6{LhNR6l3Ko6tJo<OJ6zAcY|IjUs&k2}0
zQ>&D@JaW@PScYV>h6-z<{TF}8M08qF>tLP~k?SV6Q@rmk)MonGJ|x@4W-)GtBe`*!
z`MIqv9=Ut4<kB=Gx$3j%)ZoYMLG_afismB$Hwn2CUbH{z)~Qb~Q^ALld=_UsU%~`%
zkl{{ht)g7EE=pkl)VX($hew#FBRyr3-kvRY+QP&Es-3wk{6CVeIxMR8ZBkMq5)v!j
z4YG@LN=i4<DlChVF0pjTD<IOn(%m2>U4pbM-QC@t-{JTD&&371=iJXcGjq?hG!7)B
zusoel?rVU6n}5%AQ3bQ;^!~I^4hI{%E+!87?loAnSe+SI{M-WU$vAat_F7LUIlXht
zom+P8`fd~R{-Lc$Mo`UCyL)>#1$h0fdLcU?TZu7u=IHHBWA1zB;*b?cI#DC{=0|eW
zp@XVoC_@IHgHv1!_&=mh{xP&==d+$Y`C1eU>OlO4L|g8m_mdFr7%RWGzBv%c^B0H{
z0PwYI#l*k(;=Z-N|7@Lwi5ut#)`a8;bvaXx=kAe^jj`G|rgIe85Q8O}UL!0$Z;S#d
zOO0vI%ob=3!4$VD!|=}zZ-Dk`llk@c7Pi(l>mW9zP4s2GA>*LXA&H(|8NohBS+(|g
zBadD;D<rMyqVM#EzM-M#sbU=|+R<e;>o<=QzXkQAt>0ptzo%|Yk6jv26@>@OOfm9L
z5>3sJyA>O}#>dfNV}%jM-jKl933qyjF<P0s_^CR6@z?&CWGNF*6;I>h8f-N*G#=mj
zJmy&}qzDfud0Q|-Au!d5{X1q8l#<iJ>={;W&t><phd|p2J2kL!PAMcLB!-(3nA&aT
zVx2PR&eBcvtD(MryX2Yn_t=Pvku|Q7ErI`Evv=u~-6M)<3{1YEqil+lCDlNxhoU<m
zD`7ON0c!<(Zl38A30q^DQY!Ide<6mnd=V&Ps*c9lV#4D^8`Mi!vkcN31!B2i)1r0W
zAJcx`_X5gS-<$bm^k!Gc_fve$+*-8!q!BzvcErJl%va9mxw6$vW2$nVk?Xflfdow_
znLf+&eAQYu^Vh|A0jI4I?~c#=cY~pFz9cdt?gDCZDc#B9b|H$kv-8IgCIgFB_}bIP
zFujGV0}$?)hNvleIvYaKrUNjXCYvD8%UWJ-QJVF;QEg}T!YARXt-s$^u^A!Msv6RN
z6s!WkHeNRR%#44m%-G8M!mQ)tu@e<0iECUzT*@cQe}C-vp8zlT!Q9r%{z6#?=D1?F
zCqQz<cM3c4UUjKo5m#{TJyqAg6RMV?-s0!kq{6mYe*$6?FjL0AgM7!~9eR>uYYKm_
zi1ik$UU%OiTD~~#zPGJr;ZjsA7`?d7Do<5qo$IU$-n5FAR2WMD1^^B2=mDHCAt9mZ
zx=u#l{W5K>zOA{XB?4%ACe$Ty-qvV=F4<M91?qIIG@4Gy$)lnAj#2laQ0O94r8|-q
zdQ~aO+MQE*=w;}6)iR8TzUFR7e|vkZ`>nI#xYpvr)l#y(D{i~zy8ruDCBItNOX92u
z8hvR7rh`iU_b59hY<z0!I$HjP`+Fb2+#xI`Hd7#GQ`VWaTiLlp`HS=A6KcjxgJ0%{
zOnC<b-|%-rcckLJ8Z_dwkZ9^+mTg;0PP9^eh6%RP=vvZjN$2aizGkPTr3EG>Kxu>l
z_Tv<%uBr5Kf~CVlM;ybGHBC`<7iZ@xi~oT028T#6DKCi&XbF)MKWh@?1B#!|7zrz<
zxCx&>Zx9tOTC$5Kj*#BCysVAY8c&$=d#N&YPkGd?BsVTJ*8d;;1SVMtdKqkwn>aDE
ze-)eX!F&IiV^(qHsS(wB|42d{<=GE6(V&ANJ`9|#G2k0T6dQ>a_bsu;bg4xo{maE9
zU5-<aPK(o^c>1MZX+M;9t4_>UXxdpk_RAOy<Vaeb49`K4x$EZroYbsodZ;XqmWSe=
z&u#V5k~%7P76#QU_k!H-2IzJ+E^Ikql4%EW;RQvbPQ2yHeAYj7c`sZx4PDn--@)_D
z?-V4b5LIHxhyq)GOP-b`@P;sp6mUMyH+#?IJRB~`!XD&h!IV#^jdtD?CMzd^EC{EI
zsiWx{GTuNjKXxPYLsehFFQ@u5;r6As16{5SdwV?TvZ(1&7OZF3pxG(g@ur=x5dIcf
z|AC)!?{gP`t6l>_&22LZzYw_y%)bM#-#$*V1rlV2x<T+IZYTB^2gi3TqZqMc-($=%
zG~byE7P%2dB)37T`VO{r<l}Njx@zKvhqU=9f!J-gO?5GT#+-%m9T&cj>7cMuGT2OM
zX=y!J_W;Q03El+R*+s^z3_WoKo@CSi0xR=?Wt;tVU7Ehp0iUlgdf$q^w)(p09hXyN
z`{uRG=P{&64*#M8YEz)7i5bhp0yfSG>t=1Bv2xb_9YFjl-0m#lG=h|(TBWg1kiBAw
z&7+&IFXf*Vjqj+-wDeXMBWso7@rOd2ZtPm#$eLXzm3vviufL@-TW*=Ida#8Z2)SP)
zJpq(`U}@6V1ryk=o(?Ct6CDor$}4-RhOUPQ;Id=ocDd%`-{fg;MNL752U{=1L2kX5
z9~C+N%LcV%5(4v5f{1anyBeCRWi=-%RpLSUa@=pN)qqa^RH*bADtB__g+r}X;2vw+
zFn2!9EYG4Li|tm7Nh@pUTS?a)f!YKI{ehb*8t-fx`rs6fO_xn52K4waqC)6tXlXn5
z_Usyc?nHsztmW-(&d1MGG^f#|??Q+aj#=rfk;jmY;I-hXEa(s3M2nzqM%z0(b3ufg
z`vMjuplJv-_?DblV?xmhlHV5(|Aa1{d*Op(r2^qP-X~W$$E|z;ite)1#bL>{s$kmR
zGj=8>Ca0@8N2)O9R4mFDdW?xnB=f*ufP)KEY1gkLoI)_lfd98i5|)^<H@K!f3R3f#
zjHv%<gw}mE`oMr@H06BUfujB~bwd7j0V}^Fk9EptNxze{D|O$BF`r(;m&EZ9Wc$<<
z`YHPbbD(?>);dFj`5VJcJ(|N=HL$&<`0-+3{c_4ca0}o{ItyL6y1G)&ma{g?N7AQ%
zuC&5<$?Z(CtY|mH9ht=DvO<OaME0KmK{M<>zc++$PzAfqr3J1^`lv5|<^>ZX9`tjx
z9xMa1hZWEFDCe_(5aJryF>r9^OvPM_X0{O9iKe*Tl`cF72FDt^LJijYwObH+|C?j1
zz+qE~6;-9RFK>75?<Ilk*a_F_6M*tc!c$HK(&|&}IVEJI$wi#&o}DF4Z?3op2TIH4
ziXlzi2>DO>aCQWtsvu&G6%Dq=jtHw&WtF@QDibI~py07-lw9JbsChY!l@uT|Y%mj)
z-E?vkaRmZ74|%2J*tjC2uGr<rO(bJg`cfkUz1E%l%&?2WGkR(t^;#l=2-5H0wzg?C
zNSz$S1-#&|0xQ(jog+o!`E^TXEGr4`<EyilXP~he=eF1#9(Tj6O{Z*3rA+W=kx)po
zi?{b&iB=J>`Nc4INFd|#TM-wIzV>$63tIGN_GdBin&P4?vIhYXTDWv*DE?`lmoIT;
z^II^U%HD>JeKDUXx$NtWQy=zo@}4>irLqGM1t8>f{Gb4Y_vML}dlzaNn!iEM>6RM~
zns>JFQv#grCDRZG=#k-WSRv9$5NUs{EvjeBz(N-=+zA)0=uP1#uiZsgd$m;PDw`Qv
zO73&Q#oXLWPf)vmobuORbdH-xDGJr~7?bkh7{hm|N4B|e7BD!B>X+xKqT?5Ie^38Y
z!kJmBgbfVKj}2E#*06H(Bx$36D~{m0>%DZDZ)&0rP)<8q_DE7o?cCHP6$YHqT~}A`
zz+!4Y=>1<w_Qp`#x$#d*-I@9U#&1F<B@Fp>BYe*V8~9h!3(+DS{NTt6+qj;#;(WBV
zVwp|I69djMHrgtIKmUZV@FpfEnp{4Kl&KnmFiw5Q6Oq8#+#d_QgstDP2w+K@27!a5
zH~x5KkJCH97r>@=zhefOFUYRQNknGozj$-=qH0Kg2xq^H;b3HIrQ8HmdvbCzUo1dn
z1O5!;2Z}lHI)l_%W8RhMmz$M&5_5jFeAyG6hseXFgoZEdMUkdFX&*QA*lwvAu`|_w
z-gnDFKXv9<d<SPdEI#J)cyV;9j?eq!N|JgZtNoSOjp+@8Zf<Ap4w#8znBiU>Y2BlD
z=2RAR!FQs2GcN9+@4#gq|B#>_T<Fz|lervqCzmR|kVZhQODy#BayK$%iT(R(xTbrY
zqQVb$ejCd%cOHur&|G|^2!xm~{41AyFIxQc5>>pg4ZQ9xtFB<L;+1Z0)@{Grk9Vb^
zElgUuGxjwpAI9x6XOmfniT@S*IAn#EzKN-C_<BoXeXHF{Usp%W03ApP2=V5`G)h6z
za@ORefe!AcT-0_dn`4seTXQhBLhq~kzqq)NEqGs;Cb$BRRiKG@k<^R2$TmIp>4guB
zbHjn$TcPG6sJqw0{PMM<>O3yrg_>>Es5q<Xy4SX`8=g>el<Cubso+PR@)Z@a@4#?W
z4LuG#BglVGg9Yp<3nx6J5>i=X@|rk8gRNRMNYUorW*Pp<8%b<zYTAn*A7xcgo-<?T
zV-v-y_K5G%gzMD(q%5HGUS7AT9xEH`XyZ5`y;Ysr|CBh9b&lnW{C9k8>=IF33D70h
zvg4D%Y(a$spHrxVC-v|drPzROYbV<cXs3vZl)jSW&Ea*R+wirioYZtCGEG{Vi&NT6
zE_~3C54yA1j>CGTd*{K6<Ll%^N+;<x*&_>Gn?o!7K`0NNw5MwyoYgC1A?vOCxi@rK
z#g17!cA)oDwG_n?H}}(9U(UUrfq^pTB(7hcpbw7s5jb`P^R-H`V&dAC%JoIia+Zn*
z*(i{tR;SzB+l82D!_W320BLO7Dg`(+fZRfhN_Q{PvheI@XYyk>7M%ibZ$e*K<*D`f
zZ4g1#1xJ!>bKg~OB}AV3As|Wy9dvBG9>0MM6?AMf7xk;_W~wO+3`)cyP<bZAp><Px
zTGA8R>NmYe_2g4ct=G=oxr%o;W!-Z%%Xj274XZqnE5CzRR99o(IvDHkQ+~z>3T}3+
z0D&Kzu(@*?uOY6SBq<a_4OTS(F_+?(Du&ziY&B?gfN5qjA)__~%*33hoY^*!7Hd-Z
zdf?J9KFxzjWN`V3>b2>oTJCg#*@FHXT)`wYp*~Bu=p+hAB>P4d2sD!c9;=xu`pgwz
zCh{RKen59KEj+wMD-gxs8`rE~i5#yiCK#v}Xm<(#VuHB`!lI{h?{MJr+p<zT?eJQ>
zu(M!<I<JZO3!Xk{{=Nw<t++41trM6Fa|$Yah<?kjDC-2TU06keMPdB+RIwj6N8@ly
z$*ufe2BJKg_oC{Yo1s(s3(PdcVm7;l8fS+s*FXD;hPlk&yf?hv^-!l$Zk4BfQsoU_
zPHOBrPY_!zAN|=uu*S47Jrj9_tDx=0!Sv%D$MXw;YHK24xmAjw!B?U3FhkJ+$5@3e
zug<L{{(Jq5vtBo)$bD(1ObYaT0<U-nZtWcHBI>mA&14Ba&E?`S={_jieK(G*lTWgm
z-MX!2;7_#9zK;*OxJKSwu<WY(=hBsw3TCs0585AnKOb5Wij2J7b{R*Z)-uk#x+h_t
zkj>iJMgC)=A#;^@>M8;Dk5|L|`$iiCVNuvy>8F|&%7JqeEqKn8`l1-WWglFM)o=K{
z{zQkN$LakefEfQdzjVq2M1cMLOwueqi-c|;Jm{_H+kFgluLq*FMfLUz)+2YheTu3!
ztt&6r67e|I?4bIX_N-sUDi4)kwpSR8J9A1OYfX3)=5kNDr?03rmpb5Xwo|@Ny|%(<
zX)E9Nh7%ndw5}EiLaVJEBLWO|ZSv-=?IX81+rH6-hi@g1y|WM2HWZMa--`Qya<iDX
z)y(%67u~!e1@0IV{)QVjwm<BKV1%=POi)@31hfL*ZaKiv|8T6WV#+Zr(Kj{KCF$`5
zGEM=FMyQCQFcq~i=cy>ise>u_@#2ZWPpx1Ibaaeg!FeDSSbDOWl1L7zGjH;kWosGZ
z0p@^X#eI54&i;^QG~2oPdN8Mj_7=Kt-!bgH0GG~PF-tn|(agWaI?I#x<Fy)g5T(<|
z+ovLAN_Dm4T5cs0M9*6-m0X)V&Uqd#<51?8i~2Q6?kj?viH$e865kQ3p(*w4-}$Y@
zJRfxZYxpSSK=q+<Hx!o_cV2|*EZFghlCT-rkJWVXn`-d{Q~k2PpgkDrEcjhg;i|E$
za+A$v$Q?F!4ag2h$7}z)+P&SeE581e|6F(7DZbIvq1I4hCE$)|2!~v9e3tI`is+dF
zwY8F0%BWCW<QM**uLp-3sY9)%k^uYRo0qQx38Oo|hm?Z1a5+>(iwPXrnWumH>xy|o
zGQT<z5#jAX+JA}9CvtF4fVQL`#lQRDG@t7Np{Ik>ul%1z7@C@zx*ViLGA$XgFfqZo
zof%DhBArNZheS>Q6Fvy1Cx$qgfFpt>Fnr1!Nt|sv_l|`7`ncm=fEIchgY+t9LZB8<
zC!8+}jG~$z-Lp$~FA(D8vn^8dL{^$ZL*}d1e^Im6ipV0#uXb}-Jy6bVu>S^-RGFNb
zy3c(D^*&NL3xc}4A@bFGr<vGB?L;5?{X+{5>IhZ#mcEnnvWN!j(Cr}h`IsEDb(U#)
z)dZNI`b$m;ws3u2lsPxjB2Gh&bpBvF>N&+|u&-fa1dwE&dD6d?u<S;JrBW(r(1~t%
z`$zd|3^}h#t;4@~O^&qSN#t!kx9I>6Q2n7fHeC<d!Tc`_Mer)+HtlXI?Jt1ry)Nbu
z9E=DY<yX>B#>vQ!aPK0LaOo#ItRJaXq$>z5NB@p|el<Kf#PDjiD}3E(31A5uJ<d3R
zd_W2q3<g9BPI_Jhg>vwXe{h?KsvEwxH?0mcpo*dRtpT<6`%HhMQZ77`Vbb_u?A>V?
zc-(Mb*Y?4<kq_&)ct^oCxX_9Nh`@$z=%K__Rhq_N-}axZo}O!)HD)ta0XRA(DmH3-
za&dC}M%!|=47l)RIbtos6{rt@=XuUkMYr^HWp}suhtC$a&R310NBD_75gZJBlXl@*
zp=5P%oZa)x?H1K_!?aiCPq5yOf_N&e2`&qm)G5KKO{p7S879QS$rNsGg{H{Psu>Nj
zNUp;Uo-g^3&J*se3~FQ$<<}Cp^<}v+(TT5Re1FQcKv<=B58Qk^L*qGa5xjl1wBISV
zS0?W@eWnHPE}+a8n8}HHP?)IM6_+~p<T?2>Gqg1VD8XRV(1#O5gsSD*G3OwGUiIY(
zdvNL}&EZd$esXKj0j`e-@)!o7v*=nvtbFm)taQ)d4K1eBK<?oF*q9!0eRA&H9?8>?
zfU;8dSZ;b936l|XUYK?d!PBqMw3}WoA`pls!=OI@M!kxE)EM7u*y?5a-yq^nN>3TG
z2&g@ix*G!c*|zyFs5XV-d9n&N3H@<oAQ~8hNA<E<sko=a)>eCCsmXh;R6q=5Om$fr
z$)Hk~_gs8b@R<kA&RKmD2Yx8YFZ~ME#DC`ox}(E9xdpqpY3{ng6|TSLFnVMKqwX5!
z_UDO`sv{QFR}#Wiy0Z1khCko-@_vqPUE@A#cioNv5^Asnu-HY0v_{|9i2~Zi^dB8M
zA1r4FhcdBo6|-t;FGoirKC{MPz{kGe``}Gca>A0*Z?#>=XJd7zE#aURy4k*IAfmtp
z_x<C=dl7u(KuGqid9@Hsl5MM?3@s+$7*P<<oGxp*uivrdc(wye(PrHf8YHwCUpfo_
zg5@<JlYOQ~EGr#OnSz5L|BR_+vo3d;f_-^?Eei0C^-_CbZYS$%)YQ}+Hx~z<Wrli9
z`@)8KmOyDEq7@z1a;<J?gRSBNjlKN_A>gv=!<V(5Sc|^vl!^OhVMrI2m+DR{LD!dH
zX)M2CaxSS~vH6t>bAoEYfZm7M#(M(d+k>1au|`7_jvUqLSm~;`PhrRXU3-p!#_Qln
z5Te2J*O7IduYrV*O!}NZaS9WI*nFV>dZ=#=QmW`_7&0l2lMHufm_hta2>fyo-RDQj
zBdTN{Io?ER^jmlIyP)A2K0&bW)GBbW5O!<lZqtqRc@xeeWE-*!-23b=D-y*V*AgW0
z*>?)>0duC?&AA;Q8(iKU_%4_pmEGR%-ni+OZG(m*$m@8A>wXB28`Z1$bQj%FY!^yf
zk?$1%kU;ozk0x<kSXfR>#*y_H(~HyXu_Az?Qvn)1Y(Z%Y_vbZ_c>0kfr}%C6&73o1
zuy$Cp$XoY3m~9l$RfpulR}CL%EP2R(@V1EA{+6#86s@vK1gg(vhov5Foi0d7<4KQB
zv1Q|Tb`ps-s2cPCyX9V4+NdqauaBQhi0L-4X|1-*7nnK#Q%<4Tppy=Ot!KG>j4(sJ
zP5j$->w-9XWY<rr`+u*hOubTGwS#N63sJ#UV+_zQx16U%WciZbM<G0F`>D~0|090&
z<$-LbE8jvt=ol6_-dDm(k5n~Qm%{WhyuvC`U*^G_CDvdoU-dE#tm2D1Dd?o$D`Au;
zRA9<<x*i+09oLs<s61MPe8^#2e8JBsLp9+1=&(`jGcfli1GajP79MVVBQ6dW5Z7Fe
za!vG?1GiB#Z+~Y*%f#zLtAl-H(^%RGOltq7OK5rX_A$JirQRZoF}X6v*21mBcuuU5
z#|J>mnu3Uw146%xIZ|SUU9+WKFR0yczfOSMC!oKu(fx!~(&v^EcrPKg09z{t9ho!;
z?}Yg=mOd>C<<`w1C5h=_1}L44Y$@Q!fB~JNGeHoBZRX{;l4d`@f`{J|$)JCRc{Q@<
zup9=VJS!0UfMa#v1RpkJh7mx++htGovklWOMl?-TcNXJX-u~vr%=za>ex_{UM+G`-
zNN(!Y7X?uqJjGBH3F7w1LlJyAeuf;?Nnp4L)tpb3DBQJ;DN1PjlKnfMdb{Xkckk$^
zw4yUIfO{JAKxM2QS={G>>t#u(zeh$nlEjt*9#p%(5_P^?*Lo8#gZAgBOhVPjgZ%1v
z0vNJ$q)TX_Y}S@1UI8W3Ykq?%J{>E$-H%xkXWM#SK%PW-OAFn@iSIpXzex<psOYj3
z;u_qz>M_SAx^>dey)5%%&Z?=H8+FGCpx%i)_^h@m7m&n6!vYK7+Uc-GI~0m%TTaCS
zn`gYYOTTWXyZ;sNr${_IR7%KpW^V-#XMfX*s_^%wAqy3PR{$RO1|FXW3lTRlH%Cm!
zh>1kU!pj>!2h!t2KT~%I66`IIM5oyJ1a|kJfk594y<<uhS@=9C%6I`TtT%26`-IbX
zak@WA?R8xGG-sqDwUdzM*Qp`)M3(OuJjyI-;9vRjoGcK^D`u(dvK)}-R`I>Vs&dN;
z^|@v`!x-B%HTs)nJ1*WuQ0+YY%CDvFS^l++HAg$NZKc;%EbFW6D8_)1nPY=mq4Vs4
z(V_S23L*Tfr?XmRx~0SUcGqS1^uN|<1;+U5;M&VkB-L6TiE86m`TyI{C~~_xhI(DC
zB^-@tNd28K_oV@zsg!2^vW3$=(V`|a5WP#U^tw~G3>upQsCD8JYIfAh!f-DcST>|W
zHsEHHUYjoR%KO3*o7(ztHn_s9^fj^JwUI%Q#A*VBne+JW&Kp2J;ev#5UhkSw9WSC!
z9Nh&Wi%vaC2$?}Q7-m|yX@}$y@+D3=b!4yQDg)QTP1;!$%J!eD(txC&E(O8}<{mdN
z;VLiScO8~A1D00xsc2J7H6`bvUtY~{uhwbRz!n$N&i#E!VA^TyN!zjDEV{S%h3H_a
z;%z6zneg(bQ20u2gkAHssm)Nb>L6~RIdv!`>~ggyT=m>uyy@X@)<z#A4o~`)1#&BG
zN6zH=0btQU5LE^5Een5G*sg1BIaPM2@sbyBT9CVXP0MtRSh8N(1whkt;Y%>qf@cOl
zdrLYrre}34a4C4xthux5lm!?Az{dP?d(iU~^-rh(Y5c70nLR;XK0bt?snyEwXtA@<
z_rg1hY41X}`hxF!$Q5>|t4LP2_<mRtPGz>~ka4#zNE^KV&pNF;D(x+`-&E^{ug^Mx
zT40RL4HB);U;VG1XxfzJ6<As)3PC=R_u3}bwi-(orwDA)<J}%@xzdN_yz9K_jDc)4
znOELG>+nzS`^>Jk{!ZTTv<Rb|@K?8#|CUoIBG-4INNj|^a;Tru-Q3b;2ZHK?Y)h}0
zM_oto^+el(-zzqQMWTg$a7P=}_Wd45kJ}L=K<7Rx#5gZ>$-`<-I;?X7{4+pJ7f(`{
zZ(yM^W!+Fsk#q*n=rQxPqTY|bisJH-ry|R5g>Y$85IrV>YG*?OJC%avHldddEZ`(Y
zT3j%$#}lySK?TL3T}?#P3cy^s9WBcP<e$8pbEV4>`2Vwz02tpjLicdl<p!h+sLM4B
z`^PupE!MqAoFYzRhKlY=OMblEOE|4w?b)#zqIZ0c>GnZGuU3rVC9`1gvs|qYfSnPk
zy(NWW?n?|`(z+@*71dV=l4Uu%!<$n5DsDw6t)Tw!2ri#}Cm87ku}dX@s^Z-MTP!d=
z#W3Z&XlZ`^;(JRi{nFZ2YHiiKN38&EMiAum3lb9bulsUpCe|B)7OO>uf8KmL=x!tN
zQ(QC1TOnQ~TU7|uJU|_CB8^#c061*Tn!P}P-qL)vo<xXt$SCk|K*gdL?i?dtD;Xm+
zFo1{6PK8(x`Mjz353BK($xqXK_^lT2JZkt~{4O%hH}yCC9F?IOIA!^Xpssx}u7gLh
z^;-A`n**Vj$oMIhOb}mFq+;{kJBWAS1PhLN=kjD`D!l!5!pLc;_@LD7<x0fkLs?YT
z98y_XS=-pStDb9WV$v};XDZ}QN@qe_WNF@0*Sn%3CMh{dSoQ-g1ipSSg*AIqUT8xQ
zCK(PIdLHzvoSDNxuGH=JK$&UTHn4Zi=Jl!$yMD{f&*vjn(BL4zBlcKwc+h-PDr5my
zRLqJOiB^@AJd90aXa#UxnORc^V28@f%WG2m9PRJ-7f^11X<pmZ6joGJR7wzdeH+Y-
zFs(GTO~d9a1IK56*7z3BgGJ;lQEkPOTrkvnV)vDB-^*%8*FON?WH5>pO*nFGy%FE5
zcEHL-jQ;Fb`Zemrz+)Q<15eG&`~ZG+sYIE1OE_TLzg$mp@JKSLb658=H8;ONE56pv
zFqSY#JDE(nPFV}V#ptNC3bUueVyvS+T51nU^?6JCa>ZykU+|0jC<Ep`R?=vZZQ6@A
z<?)~gb8V{|6NB4qHGPHk!%5yW?s8s?hfty37*aZQ)`}xJ>C6M4?<vQdkh4}MHD>>C
z;E}!r4i}<FR113w@Z4f|%TM&9(Gw1n7d9}uxWCRa)<BWG4XG;Yad+LU@69G6!y`oX
zGX8*_S5j8-<DSiO?-oDsPO6qGnLxCSz9lynhp>%vtsRA#VLj6=RCceB(c}0&o(&YF
zECaQOrirRHHsW@f6}k_9|H{|A)vbKykeBRtY>g)NYLr_V^P74bwACUdrY9S6Je|&g
zRySo^&y@?9=REn`j{KEYOrG&_iiBCk2kMr5ujH9m$T?z{!<Nh;1x31wcs3-nnO(ll
zwKrwDPg=NTWn}cZXjpKmCncWKCYSBSz{hgHN`>EhtWRM55JC`bBhH>q^j6NzUKc+@
zQX9|MJ#^}A<J_RoL(C)BMdST}@W)|^eauf<MP-#$$R?W{6dG|?Iv<O{VvoDU4agg?
zCtQr@qt47|7T&)NqC3K(#nN=g_|Kb;*wnMpoYFA9WePAR{{1+#RmF$tc<kuwyI?%L
zKBSV<qmFjnk@a)2Dz(PvqWLgCOhY*~+_%b*2hH@#A#TqO@$$@7wahk{`iaiqJ{ly~
zF#JLPsC6*(&4zO=2;X8RR--1ne5GRZNfrmlttay1lXsiRZf#+qc4KCyzm~(p<usWE
zN~Y*=@Z8Bi!33Zga3Hb3qRzy1!2ja4W4tKP;{sUXpL(aIrn>floBf0E(?I9NEy+gj
z!~Obp^cor^ad6_=ds^W+3zCy_=rJ0IkEK#OXQGea;NKq<K&29e?*+nlV#_f{3?D}Y
zQBZ?H>AbqkftefPwe{!K+;M_GN#LL!?*T2n5)n~&u$e#!i*fZon)HhlG3DPu6|^ht
zf!7k4LZ3bRQhz6_Ku=ltdfs;zXl+#FaGyP0`0~f~3sNH{bkB9h)BIUhg6Lej#m@>-
ztbFbSVwvA6-o1E{yF^_t%%5t^#oduCe6BnlPhf47GQOV3BzG$sR#<%>ef^GS$)J9_
zc`IS?(B@a7u%WMA)bqmLtB^?!0=@dy(Kgz^ynu1P#rYupt}m^F@2uQsFQImQ)4k`{
z;Be_5`yU2dGqYlD|M%4KXnPaTJtZQ-)e;=w@}-b!lncZ(+VLY;A|4}s@d*-pIg!iU
z`yaSCv@(_S84Po>J#Lo?r{!90VAvlGx~9Ohvr_nMphl@)4RgTNDvxaTd(RSJv+R0=
zGLNd_mtMGT!s&ma`}II*Ktig%g(JYQeXib)7xvoy<*+IaxCEBZ&wuBojUt-X-OArN
zCuDtj``yoP$x-NXeu4S8yS%`a6!)p^mdH)1Jf}{$-h>s*m8UOhYP$E{%*<?mmIs1G
zJ8E{+T`oT;O|$`y->U~^Yn-Aq5%Js{A+T@`;8gRD|4}Uz=3YZ;oY@-DVcQ}AGd|yz
z`8FiErA7s@-{^fK$Qsi#P*nFfp>%ltJKoq^@BZBz%@!DGiurbW<?-nFu9*A3ZJJ?`
zg$C`R&0yKiVopTvoWDV+dfvzY`{pkH&aqR2{h~1Yn6<>k>*jvHL9~Xbj3HOW<*288
zd_^O^ebXn0heQVpEeBRnW3*`KPwnx4s?wT28Q6e_c#I_H8O)G10iN->XR);KZ0e33
zz^NeAE)HDO?IM?1^F~HS_Fd*<Wdq|O?sc=c>eLcW3XDF&Hy=FCv!>U<%Lh~1Apv;t
zV{dLHJ+10ysmG0;k5>!>=<k2$4pN*K=W6)rf@|G7ZjbAPsD=9tmmh*jU69rKkVH0d
z)|h)UEdM9yPxs5UX10DIQ=@`yZ-3;DYz=twiJM|Ka?0+wT)l(V7q{F236aDA^Fvc>
zQQO$76Svi(=Lu_Uy7wgCgWvF&j;yIPW-$OVeAXS`utm5`wSh_x20{G?{ix?h9cv_E
zYFiK&ojpc+KttpNxPLLKC*In?_6iZ8oZ`ILM9jL@?mg&8w`qc6#S+Xpjfdfv$h4CK
zPTyc`KELSJ+gkI#iM|f}<#Uov!8xA^9-LfTVel7TPg91g#}j*kg4N%qd9!Zz{zZmR
z<@CEd)=6W2#ZM=qTDVp0>V!KdEAiDG8|W#DGx)v~Gd~}aZmh3Y*2wvFw`@=u1h^Gy
zo16Db0xI``^J3DzrrfBc-XG<d%_TOIY<HZ||C-VZ*^TZmWsxGl-CV#|q=TAFh=ZKC
z)(wSBHf<Js#@QKjh?X#?UVoqDr(@?t!^)wZ4W8X3q3^}(#j{*-0oE9N6PyO)M)zx{
z79RCRCNC3pr}V!YQ^ceAgJO<OioI3aEw*(dDXKYa@Q*-PglU7@GSkY8J8NAiUTYzk
z9f9+V8L^n$T^pe1O}H7SkMcr=Ni!_CM3t0`jJwCi!0YPj9{Y`X3$qN+U!}}Q)R304
zky$$q9&4x-KGNC%3kpzK*(#cbcgx7zALjn<=TE(#0AYOr#5kG0`uiV}3I6FnW4e-d
z$hlC@iOO@#-kl<F!j0nK2@{-=)7`H$95Vp%-5D5~#8{+YDJ9n_?(QGva>u3yAeR(N
z(;)qd?tp|r_iHvd&v;DS_P}mwJ_7hK_cP;n+4Tc6IV=)#s90XAAHal439p46IrDl^
za+m7+xN}6qZFO2j)Z`~owNqMmC?x$!?>i^#&Iwvi4-p&QXO{2H-s^bUfu-e)GjhK7
z%jBb8yd7%Bi_0>g=NewKfcJOGvx&YKUsIU430@S$WJNyHWtOwqoQj)|yE_pL+})$3
zS4qj_i6Olh;i|wK8(sijEgY;ydyD}fk>t^gjj)~6Sj#6IW12h5uj*!VBPM<nen<);
zT*7b^^gi)5Iy?Znxq1aD;1oCl&fQ#33LPuY$p}!|=Str9ViFQG#SMVdEPcv9$U1MN
z%k+^qoogU5X{8O(bI12r5I2SG2F@ZUAdH_y366_S<zV5t`KOXuiw6RS_Tm+{Qy!GQ
zCj;q2y;e55&KxxX!oBGs9m1bdqR&VPhRXXG2J;Otzb}4+kRz<Xs)lqGjjHj^@q{L|
zg4-3)huT}bH;pY7&|m`HW5k)>`FdwD!4?RosrN9f4%EGf$LNo~t|Jo5!Su#U5-C2v
zzp4JtyYODu=<D;eZzPBC%B<^v>vd1nQPv|s3|Rp>VSz0UB04Fb-MzTM{%ELY3qh3S
z8{>5q!_cZ<Ueh7A1j;x<Noo}E^+dR7A)^i<Q7y9}`cyo(xRJj6MU{phVf=ZJisrO$
z`f-J~sOXp@z)Jx<6J~JphT^eqN3@EYp)tU)c->y|^YZdCGKwy&3Qn`}noR5bRJO9c
zks{MuG;OpMOnQ_qw~}s{*0Tn#E#+=BDZ@~#uK=eB(Un<1aPTAJRX%6G;_z_p;Hj$P
zpTZz9(t~aqv%AC;g`kIhBi3NBdMQDiQohim)~1)AbZf;v52Lw8;mVEY_JW3pwf*lp
zFvWg~eJiSe7o^Bx1l%cJm0T3me1pW>Mv{)|5l+vDcVu1UJZC*e`nm=!+I?A^i8d76
zENno37ZlVs#)@dQ$JLEbQ|gvBz-G>|Q0Akp4eWOfHc*Ih`!E7jp9OEqS6SX-2~hvK
z3XSMAZTTddar)W8f$vd>_4So%zB`{8C@0!%=nrz9huMVg*WG(g89r#z87&TtoV=^j
zC<gYlWEJd_K9PcHfL$QKgcxp!lffCp_T^N}McqZ&9JnSo5P7jOZK2bcxcK6vWXF4a
zGB9ZYebAF4O}12tZw?*HEZ=H;*4roec*}UWHh<s>8-O)OK+DC$T>M$DTMLszfU)`!
z^Ym(IiQ(}NN>MH}tY$*clr1~GbQb-u|C5+|(?D4_4#!E6>AboqzQLR0YZaKNvj@0)
zIjU=-Dot9!2vgfZB6g@h`CBM$gaH(uEu-1YoxE-5rS55J-<!4-el7Ks0um;*F&v7r
zpX0A(N>#m9QIq^Jz>b+=Th<$MD^VdGpuE!fj=MfMV#f)a{uF69Ful|D9+@DzEcP%(
zi(yn^(1&INL^egx);WvOyWenM1C+7P<x-HDv0L_NSJ}i;Mfd5{T5?YrxYSG#9Wz0?
z(Ce1F`3}@aAw3G2xC_o>fhtVdM_{0##R<HXlhKW6G%e*q^7in)f@dzwF*qf^!ca<a
zI(*Zx;u)80-W(d@YJAII;t*}sfEE5%C`(?l?d4N(`FY8s_%oOIh!!VqOXe$`KyK8I
zuCH}})oGJ_mR|sq-M0v@UwU$ZV9UM-`^Bm6X5=_*S7V4wy}NyD<W=L|uJia_)N?|=
zarieX7Y=Tr#B##s)h+}<Wyy5nZ@?Aea>eZrY0tO_%cr9((VE|-v~~Ry^=DFU-z4*}
z##w)ru<pq#M%RV?4Mn`l<1?urhXh+e-LL_VarpZwN^Ow%bv!!hKUMige_ip4IS<}|
zEjHLYGje_>cPkYI{J*dp#P6B=v5EbDcH{fJxgi9Pg`Bu4t)7|*wPiJR)keW}zC6;j
zaN3AC;*F<gr}&BkGSx~nlZ8oA$3r2k0?wj>Y8RECPxHw2ff|nL9IK>qdhXg;z1%RY
zFHkbn@8n|$m#CSOy)wH09?O#ky<0ghIF}90Xv+y9U$)CZVLQtC;xr`eec;*K&jWu6
zTL}9)QG_xK&KvvAd;2+1A!pG_Rk1Xjmg)Gy^c8jCA-8M}d=kFy+*q#^gmSYDfy1oB
zKdo)P?gRE+L~kA!m558)iextEqc8)#iv>7mpa+=Zo}~+v<Lgn<+i_wT2@*TR5y<u)
zf|oP-;7^gb)?O>`*58Y3Yoq4?$`no6<31IDs8afD0FPr<1r`>u1|L;)xKcPqS*V~j
z!L@B9pYh7%Y1W`}k-)M!BFe1++A&rZofRRA{&~K=tzr@<Zlv$c@Vu0*qLtI7*olYX
z7(z1PBs**1xWQrq@w+mFTeJmajulAkP2c0wyGz&C+sZcb_IY-I54<&j^Bv#K7#n0D
z&+TZX3E;YW3xAY?Z)82C&127EI>%gZszls^JHL`2Llg|G(-OO6jvsfxbXmsI^Iz(q
z<9-dmC)*a0wV>8icTO~Pi;F_rh&hAemVk+PZI4mvnHd9Ugkj&XiF_?)E}PMM0ykEn
za(}6C0*^OSagK}56nagmGM!Nhau0gH<|AhIev84`v}<pGep4y|;M@YHe<bZeNvq8G
zb5)g;?)}Y+DkNn50RApIf=a1Z>(i`tt&b&2;?EnjjV(x5?r#$_GuOLpi@N%8W8xwJ
zwd{jj-Uvy8m+8@z?lNSw7DhNmGWrf1OlkvFMYcoh6L`!Tn+R8NJByCwF*RtqJH<{e
zQ;B{T4AaH8Z(3N9LQ+bTBlDQ>xMKo$1$I6N#E7yjD%?oJ_Nd@^g5*E`s-J#eQDNf{
z|I_tl$>qij>+6TFs<{+?mY#JYzK@c3_+mAS3U9fdPq_Odk*LCL`0`$mT)>gQGCU1c
zcnmQuI!zkIBAlxKj(?7#RsJ{r9S~6JkNgh;zv{Xq#{PtOWPXuYA($<|AmQfQd4ZVp
z(Dx@7YJ5pxpYHXGQz~u_rmoyqc_UEqlcvxt=OpFirC;+_Xy6-^yR$(+pSgB>(yYlR
zk#2}QV#SO@Xhc{uKlJzZSlBn>AsX}0V0)Cmy<%fy<7{1Pm%Ec`Bm<MVt=KcfxFABh
z{!absQ>}M9Qy&p!3Th+eLoP$^3B|&7S$C|nx_5C^D19A-=z#!-RGfVeSkZIS7De21
zwlhI?9eM#*!*%KE=H4l4cOl;Y!u2MV<9=N*7krJJ>$6SG>$yA;kYsSK>lSqX&*1ht
zmA$5X^yyb(Q|y{4ThTi`+J(GUZ|dYtN1f73U96BUA&`Wwk=lEKdA(e>_2{%6Z5D@%
z?~?1ap~+`@3ib{(|5c=2dJ!nO8247mrZar(u&MU8%Y8Eq$ni5Ta%9xxZTWTX2C6@m
zX<S_{(3J4u2p>ClB)YulD|0)9i<_@83rxO{YyhCP{jyV6KYDuG>IVmN2fz$3mZyKZ
zkVQbOA+}T}w+zI@K3yw<o2t*&sTy7t*K03^Rs_q~MC#VZ9Px<I0p7$%5#0Bo8Dfh;
zXzp_beN@z19xfgf<o4c!Q+ks9Q9ckR-8dBTxr=Jc!rYvT3JAltP_Z)(wn=uTts8Q^
z>IJA_187fGBhzdxUc&sG6-+*7@)q%6v|)&2<uaR-uQ*l9sI%?}jVrEj8^&-pw83lp
zwfzZvP^Yrqi{8oAEi3`W_S-}!+SmXr6gx=FSKm0a%O$+EMgbfx1-h89a4<l80V>^&
z8kn69op9?XFHTJA5N*gra9&Lnye|HIunR{-@7RIKSDK$*!4*2eCC&7UEYN|uwxpd8
ztSuSWTr(q?S}J6m0+3ygW?(0E;Wo*abiwOH?I3RS=ikEi#I&yo9}h27rF*+vEa)bw
z5Bqn-Jm=cnXqYMT%dWNz1)Us^!{Y)BJj{syy<+TJZ9P@9`&@*518Ca7j@N$OAFJ51
z(7N$O0kkIjxGzplBVz3by=<VzSDHZ)oiJ_hRT?Q&HLpwZ$bTAxkZcGSScl$k{7>ar
zTvRmVIh|zo#FHEP_;%7SovK7Cg?<=)P*acn8>Pl%TfXm?F+<uQNkYK>a6g0P!hr%X
zw?<j*YA&$pOiuVhm!kdOFhm5+R}sB0bH%xJ@i_BY&+1$zL9JB`^KYaFN)+!Ucxte1
z#FaSK63wc3pHwr7IuzdcsaRNe7{-H?-wiit8c6Z{3^;%($23B%i5hKdX3VOjw`%)k
z>beK&2A;Mm$RX+yI}60DaqK|2x~8!X3RPq!_eT|Yg0lyg%;Rn8H%AQgv09|UuST=K
z8sdS5?A~u|Y@{%t2#7$IvXvliDII!}SStz1LqkH@UyO)TP$W=aHB~I+^l@reOZnf%
zG76Dw6In|TYYufZ4Ifa)46!+ji#qw0b{S1r+qUfUv?qG;+wrLCUnWK#kQh`UCW*8X
z9Ei>3@@@;5O^>z-&#A=KdSLN#dHH`t9<abWk8Bf5Jj-e5FF5qX9yGE$JO_LqnlQ29
zCf_F7=MOu`03CnyaGY)bUd@IIB&7_}8UU4m*mz*%+JqMmr)k&{fuySIt#+>!AAlBD
zoWh09oyH!JChRx8g3*DD&kS#H2|Kvh=!ow|v^)laqglPh|ESP{c!Gm@wO=Du>t`&V
z2<G!`d&afhk=XJH331Zd<zZ5ZI!`O4joQ{|P%K6nNlTxFMM+$BA2$%*ob(=v33vOc
zT0Jy|@c=$QfMkBW+lKKs$Cu&*>jBG%mK)Ouexbl{5B5g#5jl5(Rj<m_yu9*7r+e7E
zt&`UJX&qr2$BP+N3uO!HBY~S(NDf1PXoNZO-d+yV$zc|k^?G`2yf>rYBINl3WkDO8
zqcr9^q3)TYI6AAnwKGh`^hh)X5kI7TOAyyU%9_@j)o@%V6X($hhiX=8O0{9nnDE?L
z4k*+sIiirDncRH3vpz&5q$KrfM(Z!rP3)TTb1wx}i~TY`n2{*fynf-Uax0{XSysRG
z?h|WvPdZHJpJa!{?2thaJ;j1qshAbt4;7J2{&esP8z!TAg+!{WsCVPfRAONqtIuWE
z%AS6sk_ypV$#H2GX;joM_(2IG1%zeh^-C1^B#Xbo#0#Q$dWIF<I~a{gbkphv47_Ef
zh2;pv?}U04R!J!%ipz^RJ-`CwKU>>SBrpZREc~ekCi0Ot@n#7^uVlNpgHZBVVushj
zYyvaOV`LpeqC17j(TGecNl`|r_}C039P3r47zrpF5{9oP`;&+7mk+{HfZ17KUTEL3
zQj+az_(h6;cS1*C_gM#x-Ff|9v2Aa5;S7$Ym4J9y-=DLPz(_AkY-I(*GS_B+Lpv6I
zKjv~=+?D0pTM5`O)Csic3h1U!+5Ix^5l9Y_<-H{l&xE8-LI<PlQjx^|OimOFohNdo
zdpk#+<5|HI4~V8JsKK<?S-hyrC3(ekLg^+Ig7I2n<ZNg;Km}j9OGmE?uY&Ax&j&Do
z<mjj=bto43q;fjdMX}AlFZ93@G~ECoD_5fw_}2k=9f4RcX)a@><8EJzl9!T~%<vu%
zw#LBuw@HI*sUGkLr729Kr>`JLdAllG1kojb5I#vr_ZolG{_bA~L^Jxx_?h*F1(1e7
zkoBOBqlDAn)_*?IU0GI<gPw*?>}3hpdwJB&1h*{{WcV`G(U(I=ObsP2W?)(Oi?Zhj
z$3Oe;vnI)Z`6eyp<&4|q0b2K^6<iVpZnQgS9SFe_N3))^0BjJ-7B3*dzkqh&RY9e5
zu8H<>6Vm{lP|gs)25ZdzrnObE#Iznr9j6+}U%I->?73WTFoi48*Za@;KMPT=zCFdb
zw5=k+y3|5d#4lP?C)34uA3M}6Ooz~Gn1;pd)O`SZ%<bujM$^jB$kqspd+0T|1(WTZ
z|3<1(<=EUL@JK1eq2KO`Z_0gaDe4)@c#7>hY+LBaSd9@Z;k!Rf{^<}SpL~S)?u@TA
zML`aQ>t+}i6Ylmo8#Q0a@|ED7<}0sULj%F?idX7?%{y%roR{}U@$@SQC+yvNrLj}g
z!*_zF$;!|Z|IO+Og*6<cvyr&AqnxR|LRB?+v7T^4c6?uE+v4?&xYGq2^!#;6@Q_L0
zxl7o>I5E)uc9&dXx`pHh7Q&r%0&kX=Gco=3Bm~lTpO)engtJ2esk^S{n2hZQg~to{
zF!HA-WHw@XBI;uf*N`EcL~O@B(bL)2IRTPLcvJu7;R2^)mVjg`xMA=8e|ztz_D#Gq
zNr;=e!<`wD{07-BRa4G%CIDC7W3Dl=lsWvkhjhk3WJQ$_RIukGP@S`)ZZ=^n_+#ud
z8KARsICk$Ec2c;|+i0vefE{<wU6fFem*U|_4xXpuKec~z1YXq=?zGemP6~_ur!So9
zqJlK8v`!N#5AEl07op}hbjQNm4`$&4fxcZha|2N&9tLfIo(H3ULco}cy*xi}hRFWj
zX|frwG=F93si!D+YQMKaXw@r&`QxtOFsu2*?FBKyHOmxx9o+TLe;sXD*7cSf^lDc5
z9@5n*fyIL(E0Hx{A#od4(Ldc`2v@LY7niz8c(0+zCzO;IW-$f!-(jQ4V$=tMM9saI
z29mTO;bhIPM`qR*zq(w(s=vD%!7Zoz9SyDoB$*h!iS-2g&K^pc?3p?#9eRIZEFr^A
zk22K4M$>}llGH!-d1-h0^f4CyJiQE*KX`gqGO@GIE|BtG(B|TMf?GFQ+3+tOlXdNB
z0=F-90x|AZOw3&B))jZoAT5RFDmAes(s=SkEitcIa7jR9sDdK(Z6%*X?Vs{vRhigQ
zI7u;t>IJR0;|Hisg@E63gux4^GWj26+ah9PJ3WK+#Fwr&&i{B*xX}}0jR_5vYo}~I
zHWE@a?s{6U%+UNPdejeW4Vb^~#4h_EBixHj;owlsLFp@>Cke>QjQ-QmmK*w>OtRU*
z%{ycCRa?AoXTn2TT71SG*YX4<HN8Z&FAhYNm1$j8)kFcQEN-ds-W9kkh~`L@LUt?(
zW=dz+>OkQ)6fel$V};r+pKWfF=0aD8&D#e^rM7i&@C?(SX#NXL4wc8cEdfvsMV#2U
zE%9pSX2uB`#Ja4@k6DmPGw9biLSp3}8GJu4oU3IfUvQSR^A7-8|9fR`<g%&6150nK
zf;`kcHT=?#=nr?#+u4*2>GD-CEbbd5ZB{&f*RXK4V0dyjK+C9;N7k)|K5CPXp{z%3
zs^#si^@t;;U&a0tm{R@k&HE_=hLg<?HqYOzI}6BWw8I6ql#q7EuOfneNzASUr+^)u
zcOV3cIsJX2bB0f`h)PBrhtdn5R}7Z;OstDY#PD~c#i6Qt*wVDllmAxE;_q4cWqd1b
zqPyMXyeWiQj!8A=?z?k+^ufVMMG$mHCQqt^T<>!4VRU`L^9u`I0FdjFO4yqndiyZ3
z<AM@)Qq5NbubPmeuA}V+u8pO?`o+NkBiv4=!>Yf{;QRwG%*?^GC#|!38zw@t-}ce$
zoHT{HE6j(fWgvomB!NJ9r>2)9lp&4>w)@-6RarE1W42OyZn%I_L}Ol`)3DI@iT2sK
zB?iSc_PUdl))mH4skkkUi;4T1qsY8zf`pM8WAMLS+!MW(RF5sKC%w#=PO>p4y=FfS
zfKc9C@$acmsJq!|(u9T|ZdB7gd^sSZ2uy`_$DKi1#7MN)U*UeH=Ke+CD-yEZyDFV}
zxj>fGY1L$g;4A8V@lAz_u<+KDRG`WT{+@P2D_y30E-tcCicV?X9)oU9ysJ4X`K}wu
zRU@X1W3SE=u<fONes8jHl_Z9xAto0GfKgdJUf7f<%@<G4nWos#gQpp?xnB%MWmZN;
zoX)j*h}lUPt~QmD?IodN-Ur|9&sO4EhU<cYx@KT^f9>GZ$<mkq{bf?g-jx6fxqJ6@
z-8ZFEDk25<l1|>CPm)JEdggk}5$fOGxP6d1sTI2@*y{a)!Q`)f#v~;3xh0el8xUBU
zbjU(52Gxg3hw-*~pQ~$n<Ej3BBibPl+UsuVPb@NReUy5|QS(8$5>Pz(0P5?^V-9n1
z%nk9d=%FNO?5!-uMU!>+kq!nG7Z*1py1YKysqz}i&%qZQrwM$n!2hlUGwK5cA#qMn
zEK;F^h4?_6BU|dkk<}M9?$xON@_RXzDR`9gK0o>S^9UjE=d=K;`&A9aVdsWbz$El7
z5l^*a0f?-b!xBkDdA9x+pjOH8^7X=H{#20_)=O!m>=?liY>sz_yLjjL_;%H|o4@F}
zNkYXMM;-&tn`Q|8pVyQ8Q2oxfE7DiGQh=%8u@%~Z$w=@f8&bCI$h4+-)SV}dX>QU{
zZ4&YIZ35YFE-i$SzG>c8FvUV!GVZTnT?rgcpW!@(_21T927bO6^!-BL+1kqk9n^v7
z4Cl;=o_mF4cFd5|!1~zp`l7ZQ-Ef0EtJUnvHNw5841CF|!?Z^{jUKr@xO*+qKit-v
z1d;zT><iV^ON9w%VZeQKKw)8FA<v;Po3Q-3xAyv-=3SmW%orUE_>}SKTFC4KlP*{2
z9wt>GwIajMf+;g!abOJ)R@^cx_HE0}W5tV@ZfIR`uBMHdqm$rJ5PqfpH?+x4N2Y&f
z>GusERX-cqLj(Yed2FE)21PmGxZ7{^e6z96bbXM~{U#EhZdh7TJaxq2orbtW&CGtY
z+gZZ23|p~3OJv8|^8`g!zJ$07=y;msV#WOj!)CE$sE!ogKDb#aj`+x*XGPWVo&P|2
zCoAUrP!jk21TEaSC51Tf7kB(un@%p&#_pde1bjFnWgKNz5_<fgI@obcY(xX6-7bEj
zrc=KQdAA77x00JW`eyV%SNciJH4N}(_vuitom!q0z7%2fpwurtv95kzb$fvIWKW*6
zK%=>F+q<KuP`Y%LVFqd1@ppkKH9XdUiZad@5QO_U34Q){0oHQB$#yPCs^W8+v0D@B
zHA|o-n5dyEuP4#uvoRhJ0)Xa-KaQ`iZFAs%WIN!nO2gZ2gO(7$+4AU=o#>=-$HQx2
zLb*A1rHQ~nxWD-@LDwhw3Ua;%uP21LWLrt`lh;n_KYfc7NW^;uc=Ij>(2nV0YV*ne
z$I@AbHTnK+oD#$zph!77q!A>fQ&MRpq)`|H>FyFmkZvbk(jeWTBC*kpbd2t9o@@W(
zdF6pGc5J(M-`6+J^K()VNs8k3_=u+MrY0>+6EJ0`d|vnPP?I~lG<GgdqMOv4+FMcF
zuzH^Q#M=WqWkdLt^WbCRE%zb}_Amxnp-`^m#?4EH)?%ku7%VsR7_;_uVwo}nZUp@S
zYyVYfOVK#qb2P#e-f~}f`V;Iz2aOfClSSFf&s=QSi?NKy9CW1g{1AN28KY$vzH@1T
zq7o9b4ZTHB9?qMAVGC4!Urv;q__F}s^2^1A1+z76awO{07i5VmLXu&4<@4z@lG&=;
zvOo2CrT?t^hO^rOYwHuT10^q9juQ@_4DGnP5Ad&i-P;N3+~1;?bDeJ~4Ivto!&$ti
zx&lv}9G7}OD>U&@$|f+lD~fB&`Fk+HLgtB0$eqdX*>y8tqC~!vJK-*B;?1|`9lPf8
zWnZW9jqMRpc1Qoq^7b8hd=-|`#F)KyeKMc(gyRc#GSztKHPd^%?Ss+#&}l61>$1Dc
zUPt$kIDGQAUziTLYzL}GT^!V3n0H&=w?BUSg$i<8j!5mZY$=FpEjG@!EqrQIe%vD@
zPBJshu&%`tMdhK!m|m;dbi+tLj&@dxBD+)Tn6Gwm>lts<zGPrdkXw}yh}`V~`p0yg
zgL#d?U$pc1AHqs6(>v5CBgC}_NgiqlFm3G%?u<*;f2>wl(u9XbUhGnn-~<}bX8A1+
z<J--F^DbWuks;qHVp$bN=0=~A;4L<)`_E-OdE*@m$h@{A3h+pOAvxFVMFJQa{r&xb
zd2ojD4!5;Mvd4n$!;AW^cfFZNYG;-bE&*8vG^WdA-mmVBpof~CZ^N5?8MeitKyOMn
zQNl1PB2h*m;q7&mq4SiZ#J(iqoRPp>;_P_c?*ztgDhf#vd(N8v*&h!3Fkc0y@6b~`
zune%To2D^m@NcH!CJAD%G3WCVmn^E&(N&v)QQ^={dB512*#129<Zyc>c=)Q8Zy}MN
zLdsr)S;`LMC`J-FHuS1SZ{SJRprw3(d~IjuCVlAOkf15iqNh$e_^qYLb4Wl7l<+W#
zp{vW&zvOTQt5D3FzDX09CY_37>T!hAZL^ZC_xER;Jd6dP-4(E?<N*M7!LOFaoqj4Z
zMC|fv#7=Oh-{R0ULGQ|*F?(8B-Vj3E!mOIOrO3|S#x<-Usr};+rfIfG&=oOFpPYv{
z&ikAsRYdKYB*=8z?_G6RO^|a75`;s6-~g?Xotc^0KEMBOqL+3D=u5yh8&LLGqH`FD
z5`RIL*PDUgvrl&WXK(kgrwU6`a<g_71(B^HY`1!S{Ywlja<Mn<;%{BSKZ{j2s6r&L
zjcY2U9zJ!{fEZQA@Vn2g*XNlf;6qh3vJR<*EpCPxC$emTfNGCIgq}M`<96l~ipS!R
zZ}$1f`lCk&9(YEwYQdw<T-55RG&L6M-Z^(B(HRAq1i5<Py#lt|u$z0(gE^z3`3BOJ
zsFV|7PaSU)I<EKHssRR<+5Mi=RjsAlDwbj#Q=1SsZMmEGlH*9HB|mo`A1bhUoxz{>
zfzL(pUUs@S-ad@fWR21^X6maQ(GOEUd6p*LB|I%{m>l$O<GbsN_<O%Ne*K49BY`B7
z-8BWD&c9XH35iz^4C!`rE!ewnmaq%Fpy2}brb`wBTRM$E<p9V|yHl|d-`i!>LNeH;
z?L7~suxT=0n#hTj1k(pbW-%cN?%|bf=EuJ~?<Grgi{T4?zA98BzadT+Ss2AJl-1<4
zvTN}<atycji;^MYHyoQL-SP`ZZ~?rM<U<CqP4j(PDO5G#N`_tF|MaAtPv+~W`a<|u
zNXUlxTr=HU_Zhz@op!s<&S9(D&zH`|I+sQEcPi$BI7$W6yQw_`uA;Ldo2fNA-dge<
zN^=hraw6=;S$dv@_gg4TUgQ~fpFf7SBkevbsux7Z(M%`zM4?1^c1Pbv1!JgHDy5%^
z5Pk6-T&$1aZ@u<VU+y<{|MLq2Dkt@;MWa$If}8Cjs^1?b7vW}MHq5FtxBHsq1JfQo
zTEYfY6KMVw%|LIbSsv$%gNFP|csjJv-qH`*ou<@2Z&!i**5M9{pbmAZ8-eb<0?vFL
z;XkpP8hi985(5i!e3r>4iihx*T9D6<7b%~Vmz7X10v!JQ*lMH7P}alTilMWlpX_*)
z<vkX|O(r_kVBMK=HMxua>&O=j@Icfkx?dHS5*K$>6vQV2jE554;$06SGrKa<axftQ
z|FK7kClkg~2ON_23dMrXJUCYcC*wMb-DQTovIy57eEo@wi?H1iTIX$9H5TL^-?6i}
zo#nZ3CwOg&Tl=)@!L<@Q+Ofk|f$1W9+{qu_)L>(7gmiNL`ndPmTI5aL8y(7V3QdQv
zZnU#ykxrZN*s{%@bNew~oChI?@#O>-o&gtrc>2|!o;FxIQ0vlpnhs9Mi_2HOO=8KM
zbMV(3eP6L&cHohz_bj+^cBmI5e|h9HabTpJY=I^?H8oXEu47}vTCQ4MT|43A#l=d;
zD$sm+6i8cljx!)w$kbF9o<B9a>vxPM{WTQL#FiuzoGZ>XFh9x4Y%hXI$->SFI8TGb
zL(_(_3{Hut&VPhORCL6Vb#$1A@ed4g()%oD%S^JSz*#@nn6=rjW(g=?pfL{`0YdTe
z=z6un+u?u!o1(#nFQqi~Y-wsxjam-F9gXUjR`VyUR4V2K*2tbC4+RZvRwh1uHnE(d
zJ<i}55thx|&6#PR=Es*n6__?ddf}u|2QD2%r-S6a>(BbSx(VaCTt3e+fv)1H4Sb9@
zd%4Pt@`XT?FQ{`_NycP=vxSvX>a@BtVEx}W(>uQ$8%sSSqRYKn#2tyN4oYvwI%uUN
zf;_bQ*wor8079IXGjVf2f#gaV)lJU!zL2n{S>2M)>w7A1WoBc7ZsXR1&d$&+WT)`a
z0GcX5EF9nw<@%LGb7~BUM1p*@Z9Cx~v#u8vcdpepAifqzw(DSKN_u|RBZ8uisDK*Q
zCZ=HN`aJBSFL#_^RQo|;WzmpV0lK`M@$zIZk)AU!t65HdcRz7_D38hL_Dt}rRdeV!
z_O~4GLuL{~s-BedQ>{}57AeyQDHFrFd{-5qXM`~Dv}^e#|0w`;I3kC=?8s20{?LY;
z93D|eigY^w60`oX#rUc&$+T~NmF8>aMW@8A1L;jU(GfgvzUx8qWZBw!Hh|Y~0#a}E
zKmPu6#@&Cb_BQWA*X<1A7ts#ftv@<tD_!B=s(+}`B;_m!r>W)BJcP$5xWu&X3lN+n
zmOHa34K-XO9U(1xT90pCT27^e8Y^AW6)QExjHHV6Q>V*R);g;Pkj0(0z)Jqxzt=WP
zP}&=^?ZA!i#8i5oO?b%kdz%7tvDZ|I$8%W4!7p!43zj&4rr>Fqcnat&Ko$DF!6=EC
zV2|#wF#HUN$Q$^>iVmBM_<fcofe`TkC77$zAG1U0SM)Jza-x9xiN!_OQ$xaqMs{Lx
zv4~EEHZdzl*Cfcv2YMZakmiP9#3QCeHvUKYZ0VJw7)itvS~9reo^nx8E2bH+0MJiK
zkkvIVn{4IE+M8Kkz<3>+96WxpQu0kY;->dkIowLn%Fs&Gg{aJdOs8GTc^TGYNaPMg
z-o*sIoy9A8x87&F#^9vv05_zdKqEB|H@ExT&~p$V2=uXjEW>W?o1CJ%q3g#^RbI`$
zD(@CS&$mH2%|J6q=rgKg%+KYI{?H%R?in@8C)%-!+&|a~?s;>sn{SG6C*xqs;lwA;
zMW0R&70J7qIM>*hO{0|?7W)Y-e$%VK#%sNdBzj%_H)+fJ3~)HBwRK;CaTsOnFwKjI
z%X2@1CBxWzA!rw=yYx>+qw1<=JDp4!Z#I*9SD=AdUSGFbll-dUz_;%Kl*IDpB11)&
z+#{-&M0g?amSQ$+IyfCQ$t#+ZvIme0z|2Fd+0`>7Pj=^7_+9_CMwP-6|D0B?OoAUi
zxcldr6~_qD`$GqO_jMISu2h4IKRAS(SM4qR5(}y8@wz{#{+aDlvK_|bp`RSSC}jr3
zpQ_7kF~pw<_z7c4&x&>1XHQ#pKD>{X+&Hfi33QUqyv74)HJ2Q~#PbIsp)l)2yzNvm
zVy*EJ?`@|PL^n?i?D<~a{=3(=Z*R%=R7}X<c>(P>X!axqMln3$`}=ZO*xw<qS>1Bu
zmlcSKp1BnowY3T_Z6rKCIk~)SbTOcva)G>7g9rVfe$ukftl735?x)f7od5T;!FJy{
zp^|)glK>rV<QZ=xc<i){u7m9nHYScSR>*MMxQ`TL`Wz>aZ~1udR=J;4Nr$np5Z~mz
zN9Cs-uoX|gYdiGj^J+|1fd2H&s2-H01>_VUO~>;-F(eyi-A}kFWqDTW|HChAQsV1o
zG1g^iWuz<#W!(i?q<>GMs2`}CfhE)9f%~7-*qqef?yEytx#PLbwJ2fE6q?-)zR?|w
zF0r^R1T9Bt>NhdxHaz3p7pAm2IbfIf*G)=lZ5Q{ef??$XE%92i&r>52iD>FiK%rW~
zF7k@o*NYg{C$kx%8r6}^FRqUK=`*ji<_l+)oi3x(&p13&NrzYT4{JlceV!%*Z-XZw
zEBjvU)QKr1R(|iZ1QJbeewlJ*qVJhbk{Mm5*MChr&=a@Yv$kk7ywU{cG%XdICV0Gh
zY!3w1%qq-)*9i5*!Tg$_bXI)u8q$s7Bf4k4!6S53%kp5G6_3<DXC|gDHE7Et=?CM)
z-UAgMeU;>?SHv=#0E*TO@Uh@g(72q<Jsp~_oU{^Ar9_`xR=^6eVsNmZr#aU#dzL-d
z+f|wHg-YVWSk8Fj4EQ|IZ@I3c2qmMIZv61rSkV!*rHw|v{E?pdD*QKjyL!;V2o9+&
z+e8@vSnHSs%?A4I^T6D+S@;-PGw--9I9S?7FdPXhisq4TDEkRS%=oI45|KgOF9L`j
z)WD<IU5w=XC8doupNMKt(Z((W9z~|M>Lc;0{NJ?gYmuG?QN8X7bl5mn*3<s6g>?ng
zT&NDE@PSOf7;QxVZD}^v?)}LSUZwN_Kb9^Sd=WNVYK1-R=;1+u=9B0bnNz)J@&ax@
z^cYhN#Xd7|$UF~B$x^rH=4gNwC@#>F(hL4zW?Es+CSW8123OE-hXZKcZgoG{(XZw5
z!G$NY(g{n+nj(8n1&3ZW&^Y?6_zJh_Q~nF7kTdQ-E|-TB)q1-7RL$Bq+5EK5m;YX(
zeIq};q&EyRmttx&4e4q=@nMN)Xvr+?JB*hyq000H2jIc&%@trp9Gvb<1Fhg!FwjRE
zJnGhZp`2O*C~#mqxx3ez7jvZZ8Lp8$N-mzJ(-It_dlMj^!yU*NY?N`TfuG1L3zw7F
zOwIcBSEj%hqJx`uH~TiqiEB!Fbd6zgO9+B0F^3!$TS^N!Mp0Wo@<MT5oZK$X-aWfe
zVY$(u#l)0yp-DMGOxDwa<@4M<COEb$JaULES&WA9;yvX@^5o>RiZ`pDIzL8c%@?ns
z$$k>jU*Bp9f<6(j*_KlrS~o=Aebn=~cEv!Wa6CTsGFi+BuX&iR!_P-7=XOq=s!#xN
zgvRlKPY8Ca=$Ak#iqg)LQdIbSP~NC=b8V%GeL7dIqKjGkz^eL^)n&Bw_Ds?$=rIqg
z3zv!eNmFNyU^d6n5lBGyAd!li4ZV%^h<5E+9bANwgni)4hEbN!7$&N6Jvp7X&)~LV
zdT7rh@=CK>a4P`i5rF;~fN$E#%WI0J=K2uHvy6s8r74-y)-o_QM(#eNTXi1}$Pwq9
zEp6A-O+G4?e3CuX#d{Hf`2E_LZTFjqQl-GtJx|It2IoKDBwSeGHFcD(l%CoawY*I8
zT;1+hw^$D=L^H%iSFqt5MR!qV@MvGBpJMYLgqX0NN(?Znv}XLmQyK3flDkL~UGn0{
z*&fLFPsz#^RbYTU8+Nn92Vtaq(^va6nv0Kh_NY7d+!b3K?68;3ojGR296?Qi_5<IX
z&7Woer+>HpHf59IZA^L-w-R|{`ahY6&~0;sI->q4i+{n7?1K-K;+mi5^j*fea=THh
zJ{(SKX?^2ksa)lHogT#2g92Ntx^l2U>;O~7WZnFd8Id<EM!PHZt_Vg7`?Wkh@uj|z
zmro_)`_C=LbJ{8EuwHFrZ}ObHh2rO$0{vx*ki!xP5BIoaGVENku;*%+MZxN(uUihI
z6L}aid<89{af7h`8&hfbIBPn{dD*zn|HE(~v_zOBQ3m-tY%XLQLZ~%9$KmjoDMN3)
zACD#|*<_6J;FC(jFU<|lY?-emEfn^y)j>{GQ^)?S2CO`k+e^xe3J#{Lu5XhrL)UWP
zU*4(K@L)e@l-Ke#*Qup@cfZevCOx|SZKc^{s@erme~EoK*_~~gGf6-c#M60b!DG8a
zZ&)GwGvZgD%!g@wJGdux^aV1D;xE|@>6;dyLz0h1*M(vSRC<C32Kg4Y0%P62@|kfm
zW+X&*e$f4L{zF154KAsp<{y#v?jHk<Q+{ku*r`#!{z6p$8|ByketNpQj~j~S#pm`b
z(Y?@ZxcI(t;jXg5wfA}-Iu6<~phtL_cj=XoZ0-#{QeXrk`(}w(Wr{<=hNeKnAMh5?
z*R^YFq8vcWTq({QNXBtw%?gJnO4RvN5`tdiNk9S%!2EqpWAxo|69<^)ft#qlscFlE
zO>U`i;|Zp`xQLQ{rVXz;({N)?xSR8!g}6>|17rVs>JyW*)BXCVgngDr^^53Z7Fld=
zhoeZj-Wa3EAk=o3acz@BDT8noeKx<fp!dqgZy=T%)&$`0DNb+Fu^3x7;?1-cfA<az
z4Q>oD84-zAwYn|$3{06JTnE!){H3I%qPMoTewY^(IAiHwrOsf6s?!NgmAt$#MDdre
z^os&)IbDXN-zLN{DW$Km(&;Zgo{YTBLu{n=Gg+@2c(Wlg-u{nK5|CTD(9fxr^TfMs
z)$!g8|0^^-X$u%XPgl#I4WNlM;v%0#$)jJ#c(U`bg~q6?<R6>g5>BX{S%Q+_A^_is
z8RrXq;3$oses_}3<E23xk?tIzKu!!6)uC5Fzl`<~yjILCp{Vt{JZ6s!)sdr}9d8ha
zlxF#rh58#*X8F-zTd+<m6bU)Z9b3|-PC2tI;}*XBT}BU&FSR6t$9HdL51;os6y#`O
zzQC=>(RViyR2~aF(8D(c_P_D#JV$UMi?NxmWo{(A5L*}$Ip|cvO6(Wn(JXv1J#yJ#
z+PUIaVD9Y8PL3J+dt-MUNX8#J78vRIuJ$m>lbI&ciasjhKJGb{dY(35#!+<M?N9d_
zx%OZ~@i@9s^KO*rqsvB!m>Q-`vee0+iI|3gps0cm*sxF2J0s!c=+3B6jaB4rKrI0q
zE%kGeJ=>~a`hP-eiO#(dZZ+4hy@cCduOS7@t-8r8XnNx}TJmF~&<tU4>w92<^-ves
zTG(Ng9Ji}~CdSd~@@4#U!&XVd6}{{VY9&zcg3@<43QyIu^3?hxSZ4+2%%7y4aam+`
z=WJK60Wo|N<P7-Jrb)6U;=r?EoBCa!Z<K|`RAR}V!Te3b`S_<%lmZn9-h&XP7ww;L
zcp&(s?ISp|JZ1TU_xdN(oORRRycjv~7TJOXSEqgnWo43v7r5}5FK#*NPwD&oZ4)(J
ze1967W3U$98Q&qq%qRL=qa~Y+;EOs-G?KY7LEcQ(Zm`%m(Ns^(1wwCF=i0v3pR-xq
z8v{$|fs3X-%dogG^n0smV>{?sblpo2&~<a1Kk=!bXad<BY?&$B$Lryl<^9n0rPR>;
znjKrZNbdf{FvbmRXz?X}^?e0bNaE&^O@Csox4y>M{B1MYvs_8X-;3@+YEO|;a**yP
zHXfDUR`G_lW%nFUt^get9NmhFifMnA|GRjC$^1yeRUrFefxyP#Q~4o+F|q^KzGiGw
z6PQ|{8&Ie{Gi1B@f7dD)*tFwdGBHtLDaIwcCW;#M%%zoVqy4mCPv9<i0iM_?7UdIL
zb5Dvp`MqeJZy^r1RKM{2QaejnGKezyK|a?);>i1#!C?vrLasZPg;)^^92#*NKio&1
zfg!wG$ddWUuz}nACf(Rxhi=y&a2fTT2VW`Z8&+YKA5&|@_l$5K5V%(OiZaJV@Lzq~
zW21kpYfyu}xBlxX^mWKjI0Y9u1%3(fnIP`NDp=S~?!E8{@S<vwdKzCeQ}VD1b8o(i
zkI%oT7PicR%s)$TtfJc*x7`JiDQF|2P}_wn7y|td@P2*m`-QbHy?RYStP=p#)wH>m
zc%9K-2c}G^F-be01?M0a{J&lk0II2zx6J%bB~4?Y3G=(w@+V+`EvGLFVizX5xQmVR
zHB*yt$+p#a79om`|BdK2mbMy#5(@0~zNzIOpdbIwvc=iaoSpx$+`c#e^h+H)A6tvN
zmt*`@d*7Jjb-8i}2nawsEh%C)_ym&rJoFsl^5(yo9@>sPxmzUuv~heO5TA5N<5+F`
z7-<#|CN$OlfnBI7&0Y%F*U^*%w7KMN|KZ)B&r!u@PrHEvu+8XW{ZWdl2bv8q+o}gp
zYJ3UVO#-M>W)l;tN~If4?2xsy-5|C~&PXkmVuf-EUKJa=`D{RNL4v_1LDm!bUt|^c
zWGwH1#nRyr#u<u`2<Nb8{{CZNQu*KhN3Y!EFDT#9xJX;UtzE31{>?3^AlMj>z4meM
zRl-H_ympA#4+4THUXD_D@Whs#!%n*;t#J=0Q;&~Cf^OCV=xqVy^I%NR^l!QUNindA
zsRjK3INj@2p|*{_Z~oT@UG^cr4<W6(PxRH;viAzyJY+j^jP=VW(DRIN*H@N}PhR`+
zrl~b?=HnZ14x$drZNF|2)4ao7(60@!ge2~08XB}Sj)e8<3*VXBB`DIR;XJCg$b~Ak
z)X7x&#U01gO26~TAT4z0xwEMBiLACf^rybQy=@J5Wc$Ywhztw8t?$vbb}Ra#sg-O)
zevet_*1B6EqJl<8GWs%n!j}3xU}F}{!9Y+*o=^Zs-#6xJz&MC%=GV^nd{G67e{vWD
z#>~dszbg&eyzZN}4dFL1byMSy)Qodpw0w7~fAiiSq7dTAzt52SOkd=unOBi1zuHGW
ztcOH=lztGa@%+38wOz^P;=f%%xd?7#CtqJ`VCYNvTVU79B*Nba@$O&wc4C**iTNd%
zh|y8;SYHI9t~UBGzyuTW*ehpoO@*7$JK3{=ND|CfL8rOgmm)N4GpV0!@rx<_>tkc!
z@!bD2qT{TDer@oL_PxAkd2Xrg*9LZn_O&bjYApuBK66Jq(;iEuus4=)2Cpv~cyyGq
zSr)fuHKrO3@TI+xsCM%w70yYb4d<C{0Pk-`{t?ieEdiuc+8i{c8UI#pGhl9+@%iS}
zt4Wi&^4OXb(c~$MIBAVl0_!H>A4Or<Tl=k9{uc4^wliBIcY2}<RA~~hX%YCbKuRBr
z8bGIlTVmlkrwsv!Mxyh`UDgZsO344vOf536`utN|l{2*jv`v1SNbMu;jRGFKGhkD`
z0;nTbwwk~BocMU`WZ_vlqEi5s1JOwi=gb{E1@M(5JYVZT=kYxpT7roOZ$o*ca5woh
zXK7nWc~?CA{b}2-X7`r%_n)ID44^Fl`>zpww-=6+LS*enu=I(oEF+QNKFi*+D!5E|
z=-l=+9=8y=R5G903WI_IoH}#!4s5gN{LdMpqN0Xdm;rA^VHT)1c1NY<T#1_a{7YmT
z7rk@hrK=f=f|q|^#5OlX7sL%cax`!3wTgC9ki}59OHpy>F*p*0LM`ZLMh>dla5z9F
zz&D=0AqJa`Fwu7NDgTBy<1BZjlmTP|=F}WC0lMy%^DXv4nJW1P;+81Uo4EO|!2b?4
zg9z?+Ggrro^=3O&lYX(CKe`)e#3(YjTU^+Gj{YL--a_7g?aavwmGdT|wD?qyA3~(v
zuNm??lANX+EP;gh5vf)uy+RQcH&Y&m(vVtQ^nGrkRBck;BxvLJ&oltpT&6)VF{D|4
zWU=xoWzw@O4ck8;pp9P+?Q%Ph@^t<Q2txn46`8Wrtx5Tq5dMtdSYF{{bf)_g{xW;p
zd?S`;(KAi$vhUB~$`=Q<#4?LTn7zFODCE$vD?`d0T3{zZqi8)O(*oy%DzNQuCVnXY
zdr)e#HTNHPPLHSZguolOeN{&W#%{L;=#88o%RE^HkD>A-6QOCz1l^r|rLx01YUlCR
z-rD^<Pkp!~tpGWQ)mgf}Zca`zkEb7WXfF*+ou+&2mW-V0O|FT(^3l#`dF5LkGJIb7
z969mrOm!6!od6Z2trWi%JG!b=gR(Z4&PnZtziLzqTJB^zFt<8ZE>lWv5nDz(6S>|B
zdG%B{5+c$>$ud4VF$`5y?yA&70%*VYQ@4$`kQz+jR`hzpR-*o(lJW*#S|`SnqO}Sj
zd`P{Fcu3y;+Ed$7Tf0t-Mzn#vGVwRk$>|aA@s3}?!1oO$MGlAs(r_iqd7H~7w19tV
z^@V@=aX<L+`4Emt;Dq|>zg7#XauR)qI;ku4zdW{v8S7O80rb^gTQr#Ep)IOFnQqVX
zvXk>ldTXLNC<bXQb-gcZ1ek&Cn#z>JItShH;t)2E2Ey22T%me<_4Sz|f6E)%zTWM(
z;qxCNpkAGUs&rJNgkyReIh_@k(zSAa9o@M~Rqk~|WTDGYO?a@o;-;T;6SE(2C^>I7
zhn7KW>{#@BJs*wx^fbOTm&tZcW`*-(<UFv5jGp7F;l)rzpW+cd{Z>TM?r4jh)i&Bh
z?zgHkBJ1=%u={h`0@m90Xj(ldL1btKWPmifdswz#fOM%sD-J|Qe**?@svW9VGi(-7
zA})@^H;~+aJ&tWF+m&P&yIy)PYhH+LWZgmHP9POu`1v7Fzs}ahjDNddd$tVl|5?oV
zVp#Bhe$adI&phP#Vm8A^4tb|1`^8)FhAa}3myyU3!RDgEx3DdRm^*sZTSc%z3^TRY
zn2m{J<3Gr}UI`{<nQ%Ll$4D5HHAaMH3g!^R?0e;zR`LQu`)tEa$u7gbjN*!0Lz?yH
zvEoJ?N-)@-nmt=`5I%KMlhE+`z+?9b843MBz11M5D9<tTsxl33{zfj*x{8E&MekHQ
z5WW)M^I#oj7Ua*n^$<~uMWsK5qX3~RV*+mo?mXl`W?f>Va=YY1?P>imQEz?<9|uQ^
ziyeIyd!&Z*P&ln@A<FiepZDC8X0W7k4OM$l1)Ypt2%%YBn>OaH+<y=KCXUm^$1IeT
zzwFjp{?bbSCfnD)I0aA}d#kRZBvCVL>L7m1Ei_tpRKac~v5X@nI5<$_noWx4fy}~S
zp!>3l20y8UrZ-)v8l1sQU!YcBxTQ!6TX0NN$cy13ymMwh-T{x0SbeK?YHLuQfJu*E
zc{S(uYL<p<{|K5LA)srq&oM=S`Qo%;tu%>P$ys1DoZx;0LtT*hMS1#xkm$!=hK6*g
zw1^mOpX+rmK2@4{n!(?2#Qt$B*YE3HQhvo}H|~zg4Hxx5OGX%Z(904D@38PLyaesX
z9<C~4_9k(n)Rkzwx`gC_)kGwXK{%R2<NlQKjigWFVa*c~EYQP=^>b%S*}!Op;_Q8&
zAo>PH@Fcd(4B4Y2-m*;ZVn>Rvd&WeOs^*WNm%630Htmi#ny~Vk!Z>l~mT&zkDHFoj
zBb_g5O*4b3445v3+IXWc6Jq0le-4x8$Fk3czEI*g9sy?Tm;Y#<SYQtjdX9Zb^|Qbp
z#PdWQ`*~R^!T5guhS{%M(Hn7tOKa+{TKUFCwU*CwMS>qOD#MyrPEa$kby5+52-Eeo
zY#tK10Fp~R^9Yp4o&GE&T4%bllFzvR@R9Gi37N<7sL;(?glN?TSNQ2fSL!WHxiVdG
zzBo^4O#YUp-O@~;3iJ3)f{SHv5iggC2m>?)xS~Bh3^gN}PMJ~4ci-hJjMZ^g#4)5(
z42Bj8A3UD0t%`atDWv;8AtGGHFF9^qXy=((cfx}nnO}|FLS)=qPw1p=Wo@K&zoNb_
zpA@F0>($k;BOF2bkmjRj(j#5RtHu)DMFSp(%V73Fae5}WDKy)(=zE0XC64xM0p|be
zO>lmapy$T&fZM|YgdHFs%Xw|4=CldDtMzsP<u|t6>#kcQL6hf)9uX>wbV>eJ%D4KY
zoKy>HL<vterfdbFE1g%}8CzA@W`3v1$~s8g=E677N26aogGj=J6kiMCL{|UTnxK7v
zw92V7^|e&aCWF@gO&ijHY%A8un^U(=m<D}gz%UGAzU`c!8fWmYa#o6KQ2ftu(I>z(
ziseyR^e~CXhq8ySb9~5G2pXFEUSYAat^^?gtG=#aIC-#4bh6N*)0bAppQvVAC(-0&
zf0838Fj(-;Xbhx*H!=&&^=J=O2`Lsy&nIqlcSim!H{5_zISGB2Y${>H#zXAguSB)$
zm!DHoRU$~sUf7@_G`lifP10>tX%Ee*s(zN&$wD`bZa^_Cbl*!Q#xMVuk_jViG94(B
zO~C|~r*RLxqzSn03fPXg7u23?>JKsXJJCV+(jb4}`ub?#U^!aKonAKO>S-GRD0<7i
zCkIS)1#@Dq;Yk51DM5sL%0E!bi1X87#5;?HL{Y7cM75ewfDG|oM22?x(2OP<unYzW
z0ig!uZQqal5$igV6uVp|kiM#Y??k%|NnD@?VUJSYfn@R~TytO~Y2sF=&gkxYjnS1h
ze(Ty%MMCO{ywYw<_-@C}ihwV=j$L$D^|t2MI4ho!Br~|POsYMBQaCwbh?_+GQG2n`
zzvmlip2t>s(;K~W%6MZ$qW-3W<=eGq?3k)AFuPmzq#3r<r)q6CH0L?@m#*YH#F%Oh
zF>!jIQ8qjOOH@u<B}^hgQtqno^C0(`3534e@82{?^3QNBYT?6^f)!r(=)bX)tMrD?
zO>ZfU$1b@yMV@%Kn7<VX<>+E=fKvUr&eL0oU-Tq0E6o7G<>+VW*>46$KY95Cy*jM)
zw6q-p?%Bw)ynN8~xH)_sA^S5GlGnR07U+!4mTojIpna>7FssNgC3@efb!MN3JRD7n
z0!xwIxvyDn&Zm*_4z{FsqKWkRT#nMlgFhs8sfgDHnN?PdJG}ru0&EmL*Eq{`+M<lL
zl6AO~>wQBCQka;s>ylsQ=5a-lL}J(Re)Yd(dF~N{;Q#Tk$lqGTqQ~t%mv*2fc*f29
za#f$7)jqQFF_>MGnPPWFA%A0K;y{RpL_88178sg<Dgs^50@sjts{#pG+Uh$~T2Lbc
zn|nb-=v-gylJfGe>AyCGI_0M)jv0wasd$Dkcb@{OknWqj^E{dyLL`P+()`66TKTA!
zG%`Rio@pPxAIi(uWn4%0eU(44Yp|v}hm!i~*pP((0JT&a0RWcn40T)PKcx`OfZ+x!
zX<U1<9)>91O2Us#DH9;C`i@)tiC!k67-Sg7fN};6J@lwr0C4Dh6%KTYMC;w{6}mG7
z%%C}1qU2xTCk@bS?4TkaGG<x6F1Y=mD1P<juz!4^$O;?wRA5_D*f?PF;mV)35ibO9
zE8s}_Uro<<&fne4@AZ-%OL0a6fp0HHq2+_d>NT{O?^~6t@*&yV!Ac#LWb_qPI9pht
zM_b?3Rt=&=#Kj|khq=FjO~UQl;XALTH86R=PITq_`rYbH&8cXEzaBkbHXZX7vssq&
z<Iz}^`TO=QCB3&g3zmEv^TTY;J_6*=#k*pyQ?FLH(Z=1eaV}G2X2etLZ0Ue3XEMJx
zq{@xmwKib`JHNpvV<nL?d2OuWCmDh!58#LYK6=e#w&|CKIZ#Arkol|2SJysKit0{f
z?=)M!`XNpeTvAT}=vvcgKO+?$ad}$Fd2LwmN%MAuBZdNaOnuo2Cx+qCNEYHIv3iK8
zi?#`SM7s(KL!VqsZD+x78QkHGdmzu9)ZS$U9`wW>hwPy~PVj?#$0HCp0hl6b(~>Vp
z&YRk<n&!s-q9s!zTXe+e0Y+#jS}Eq?w*y9WB1&3$6my-nfzZy`fVDtKkrY+QC?4Jq
zBRbG&Mu6qq`9T*QlF)rl(!oZKj{YLh(#Imk%VH?fQcmiDJPir{rw{Be`>()u7Vr6d
zs~;7h8Kn_Nty4LN_!k;BQv99QT)f^q0@El^=>cNQhJy8kkHpdX9~Wv?FGTu?M!W&5
zSWzYQv6U|e!5#G+zhd%14CV^&Np}xPuLaoDOhnxGoI)_QWE+pl{YDk~qdH(9zGUG=
zd`+}5YT430{U3?hn`!a#!zR-w%BA?hSEH{z4BXWZ!za@)np!LD{Z3(Igxosy?Wp=|
zvO0nw4%<>9<|PYR^dtq?JEL5s#8R`wC@HWWG3qd~Q9gTsBx4W@H+HME)a8O%@q`eF
z3JF?b8@yWzoz$m*Q6B6ZIw}5v<`MiPw!kL^YdrzY@~E%=fK5%)_FY1Bv!$si9FXO>
zC4J8aOP?WlV=3$mmZAd6_53n0MOwCkQ_`tZYos%o?C)m1-a&BEDVcx*{a~kV-d%B_
zsInk~XqahJwb55$)+1l_@P=zHDzqyZl9ry{33emi4&vHr&I{<ZA|6Hy0bSd$*oLG~
zsn?`opkBLzvgNaM39Rq(O8irIT8GX0ZIMnGDqYo*7gi!}YPuoPj;qZl8N|l4bu13m
z#rh&9l0m`4UQ(8a2TtEMyzjNAQ71s7;%a~0girI=<_wpDY#JYN4&$D_8=C3d`Rd+|
zGpxvsaG)O!O7wu-&eu`<U!PZeu^+7mjCk~M1;)8VkJddN7*xT_TQ50UZ!foKoplV7
zR1uF}5Y{V_YST@V4k;$gB(NR6aHJGh+=QEFQ=^_w!KX}Od*MwVpFe}feSYEhh|Opn
zdk{81mZ|9-lH=7xJt3`+Th5RU+USQiX{DBJI*5MMck)>EIPAfsDtG;Qbf#?=7>G2n
z5lsI*gn#k5zwQW64w-}wf(4nnj;I2*l<)Z?Vq&o6s+c-94dRY*adL_k6%I%a>I?_d
z58u^}weN4M1r;nC7Xoq_Zp<3e$6?Oq1a?Sxc>P`OCevoVq)S-7P94vYrFgx$#Z{-M
zbw0s2SUU$Nzv0*3O}ejg@yAp`R^Dkdlge`(imFW;nhO;xwK4<R4}}6a)216s`g{ce
zYz<KF(o_+?2q}o8W5*NX*~S+%WRX!3q_TFzO)0&Mv_jSBrS`m+l#gzk1A$W(G+-qX
z1xy&7sYog;Y}(5R&qjpaz>KgD%-b*ZnG!tjDb9=CQ9nW>s7M@38;%~eEp3gXTskC*
zVC*Yi&kSbtnoCjtpQgh%8S4eP4s(jGEGxm-Qv8z7oMG~c*)_A^%&bymHN3NgL-yYZ
z1iNjjBXx<JDC~QBY^G$&7s_6iE^UD+>;E7k3EeO~tX3S{ug^u><Bgsqehnq%;!`Pk
z;aKX?r0Jfs7Q!=4AQ-}-xazk?#DqQRs_{(vd5Dxc;BC|DQ$#P-)lg4h#R8b0j~fQ2
z61j1Ynv{jLo7PJoR$<&0hkIXrt&_Vx&qiH5Idi_Zv@t>*O0rO3H2-JC?N00F(7)%J
z?k47LfJJH)dz^Lu4$NWEcJa>St?KRHF@&~Qcx|<B6O}}K@i+#8TvR3hOS!DPCnmpW
zvR8lzy@WmR*gklS>Kh;9K=hZ?Av~7<7z%h*41<x0Ag)u)m55JImzO(SJ2r(Yjp?BG
zqAJ6P7E6CBPP%xna?e+7Ek1E8`11-<$d)32j^B*0Yth$UFeQzm=bTI2#-|ex)pwc8
zN?Ws}71)Zr&q;$kU>4fU43rkUdP6HhA%SFMd|jNAg+%cQrpy}8R8x5$z%j5d9A&#|
zMc)n`MVu9-ZhO#~>pT&`emJ$(>s9CaZRMolXvq^&#Clb-hm|dR8Y!(I_-~Iuv%z39
zY8^ycjp`?}x@r3T9-%Eg4HG73tNJh{vm*jDMQUs!lYOo~<c~F{T>QoZq6qQi)gU~8
z=!&k%(PMI3!RS5V_bD{gM9+Q^aO-#mSB^ZCGY2zBF_s_4a`y1R4%-kdPFS*=#^(}#
z?Z0cT;<kp$f$(5E5y@RqnA|U!Pl9yT3%sKea24+>+WLgpp5PSO7jGN7-h)IWfB|N3
z!g_%v8$@*y3m%t@8s4Iev#|EV<Kz=PW+_6~vLV=_SW<lA-)j4r_2+&eZ#V2%Jo>g9
zV%4PwF%z~T?$nkCyw{GrK-bU-PNip<4b)HR<SE4?c_5Bl!Wm<e??oSm`7{tm`<Uc<
zq<2okX=GQC5qxV4c4oO>b`bxT={xLX?onEgUKm3gu?LRlP)X0~EIXW(@yhrNR<5CD
zphMIRLAF#jNWRNe?R8BKSP2p|jNP#18j?gjga-0$kzCe*ecn0jZad>I8scg%`*X{Z
z+{;;?{bq17-?k8mqh8V#BzH0mro>?702r^v<-Q*?$k~V`wy@X)19SI8|0{O6lmy9=
z1N&bca9vO%qShe=92zp_^r*$72by{zw!H@43jXS!5n~7ZB4dDL4=Re;zR1_b<{thy
zY7=T#ndADl2Teet#BZuMgL0ml#d8E9IN*fXIO%AVd#!K*^q_T3>FvP9Ko4fK@0<Jx
z$N=8&hfsV2zZ;C?pn6#f+jlNK!k6BXeiDy`@cWj7W#Fiy(Uu3(n+-dZ?X2~ys9~td
zC@Kpo)GBM199UFCA!A*2>Uk&ZCGbNBsNLf}Pm_VYg0aZ&2#4`yw<E7k)3GmOFuwun
z;&c0;kdk>ucUZWjpux`A<I>?;TMJzwb(ZH#>w%q24#QJiqDz6H))L>>#G&SI&7n00
zZe4taa$>osKn2q%M%LjcnK;wZ%z%gr_D%`G$Wm(NM0mKnlY_w7k$x8I&B#lu^~#!|
z_Mnezfr~YnASSBM#lXXy$-09rRb>*A?YyG&i!GoNz>J__)Kjl7tSzn&Ap;pA61PcX
z$VOqRHUP|c*@&fld9?fzG@~FG(Fe*=GO+b=Z+F+*?9<ILs6o$xK%?n?(0D_SX-7^F
zsO&eSyTOnb==;(X4t|~h<YK?S%^zFNxcDr~JqHh!F_qwfLKm*+%k7&L`eenSbEC_0
zH(7a$`s&{vpR{*$<P^=KZ94s{1(%$TG99#Za6T@mnkVe+b9Ro^H6CawM*wLEjZ`kE
zjLZBJvySx*L<zC_RcHw$kou2ED;$~wCRyjovmX##$jP9>=f@#;6>_sK7^9WY(!=$)
zezuMycEv5WDJ&2Bzhoy?kAMaG8a8A0mf`p-zjtj1CijSQQm~#5v+rRWYYW0v+{@mP
zm&b*x1s-tYkFlmX>s!wN5O|l=?Xpzo`wDd4gt6QS5V`CHPG<RM_c><wHXO(0S*v~<
zr47P}qKlr?Y<(@3JE6L;fX(g$=ucL$k6V1s%9<3C)Fy<I>mSWDT$v4w7v@oGV?Yq0
zp+G+N*w;clW`Z}t%}kkr)UwnPQ_>LV=T#l5Rs!;vFScy0C(+U8f^J1LGg4WYl_qF$
zh-=D2KNf~Ru{+ElL9E4aZd4D8dMXSXR0`6xQZQlyAq3fy2mCR8H7+agv=tnE;0!26
z96K#bPhkeOKfDTzg@}mHau>QBfp6oivLf`BvL;#&lknyT2_Vxm7z8n=|Ilg}b<4{b
zSd-Y#JV0{hU$j2-P$-3lK4^|+;D*0W-{#p&-Ac!vPny&25G+XLLTsK?^nZwmz+@Q4
z{y|ZYYp$80c+6_hTxN8l{KI*lL$9vn@P7Sf0LXWoRpXGBkw^ssf^Dy=sSeQ8N7Jbh
zO(xrWRP{utqaU{K!PI=3E|WxR&f#|IsxUA7HU-;aEzNqyfjIS5Ss5PA6G=%!f&0|Q
z8NDMSWfNQW$t4<*2|PBrh@#&e<~aQ|PglYBa<JN$a(+@CFgrWBhW*?{8i9;~z__MD
zdOoxl`HCoN)r5s~Yh^$`wGf0P&L|BoJb#^eynD_b55(Q*?V$W)-<>zudm#OGO<<CR
zC>1dL>g(&*lQ<^HPx~foj1GU1;6T1P<1Mw1<h(-gti4#VFgjx!37u>>DI+s^`n-Hl
zGAQ$(vzyx(W+QO9oP<%qf%F(1%7BJ|-kx_!bv~539&CyKs437Q20pSB$mgU&&w<X9
z81jF$g0*v!$`j`O@WFEhy)iU1BZa#P-b~#N7SDU19@xdS9u>GE&m-><VL?aZi&7s@
z*YDUK3}IE5WvI25$gHW0A#1BslaAY&Fby7`gjPi%hR4|sw@<f#Q=_g6arM3O@#HG5
zvqG~37N-wuASZ0Sb$xa9#|U$4YB+LGQdgu^B<RSo_3~XWUn)}K$|#S)G*xLZTC>&H
zL{*N>qo+n5F!sZmPgMBt>8yM_bkklvnP<FAVX6xHeK$snH|7I?eCRqGP+pSm&x(Bw
zkPLp;*lYIFu3rqb7Ca7nMS$$HtU`BD4uT5$_!qFP&^Y2W`m3OY@cDs(f#v9cYjZUF
zfA(?M_`bNelehOj{_Do{@e>r_vZ2*0=qJCI3!%|2MYlz<d&9OhTT4bqez0Kji#yoB
z@qBc$_e~kyN)A>CB<A#%VE{6SdYotp-DG|Cg7*e=cfhvWSsrk+1q>NsAUB9zFr_cb
zX#%#9befOWRfI^0Fq#<m)v~1>?NW5IOO#WRKLiEv?6gUtkXk})xesh3U0G755JMJf
zKI5+=__7VUagmXc$5SSB-6tfb*F3Dx)?!M^LJ!oV?kg+t;i2nMbC*v3yKpUm4e&j?
z%z!l`F)>>6CPqHMi&_Zs`QsU6>K+glOpE#7KUq9+Zo?#*pD`z>x7l|Gs5c1u?6!=L
zW^*?shy80uY6?16z!ratz6<!6lYTAGVm;7q)9HNE8$Q}VpkJxxkxu&HGlUpu6UhOF
zTkLCRBgdhB?H-2iZSRzqSe~}_C6~4mD@u{CVJ!$Z!T~;QpV2=O&y)Dae?-~|yI_xk
zg2L&4L`4p;LUeG?#N-&g1;GJs9xcFYdbWF?zH$s#`B}75r?Ww#??tMe%b*f-@*5qR
zQQ`@a)A+M5!`@aigl@R*MnC$QGDI*mB`F=Tcqd@1$*rP;F>^imZ9wG;PM2k%IDoY&
zbc__<YXPxR8gAXfHtf$_{7WGz$XN527gbKE%n6{V&k<W*(?Hc1^R=YuP*pm!cd7W>
z=7=RmSRtQmXiJKEk8f3};&ntkg<Aj3VZqy0ug$164%E|(1QMvxGmXwA#mktWdhmmp
zt1Jzc4#LZzhG|cQh5A>2Y|}$QNw5&3TdsbN856^pXHqz!qj>gTspI53_|<g^9Mg$1
z+XY7U9~Fqb;1KA!=Lj;c(Yvjb<Iw%})PnnxBfSwnH3~U-k=>7AxKH>JdWg={x=Fth
z@h)jD4)phX4~X-K&69<?X%XLh?&}b=me~A?x`RtGYqLbvRmJEi!qdRRB{P$m@@AE4
zj&C1@k=R`MkYjL_DEyK!PEBJiBtmwa4nDHKH%j-hyH{GHUyW-GeP?)FDu|U-Nho7^
zfYWhy<IA??%ehsht~a0mR51&w=UbtL*A8?9m@vUrowoVb4YM-7pNZR*l3EQ@T2GI3
zp%N1!(DsNn7_@+{dbFLewioGdb8x0CoqX8P+Vsa8Jf@%NTMtuIT>Y5-T7;6=cmn%Z
zi65aBH?PuWu=CaPYrlmr!m_5*XVsJ%#Jx)N1`0{Jyi5-=iS+n4E`={_5$Oj6$$S4^
zKcoStS0^%=XZZ`A|D;SK*+(txSxggS;~jx-Rj-5=$_iYY<>sbr-(*xd!URbQ2+YUB
zeEanqT7A0b#NOH!Ch$)O%e1~KtttNSv%X|(ZriJ4*M~L1#O@4p^Sa{Ui*CO+g}0n7
zmPE~cKjZR{VV$~ejpF!x{J~0d27y6*C)cT24ZGTY8&|xQm{M@@)=%3Gx2q2CYYa|I
z9;~1c40CgHodZ&L)WG|MP8rS_7kJ*w2|u`W!b`z1h3CX=gKM}1CNo|Cb=?f!77{dz
z;|W;!dO@Z<%UyqWb-T1<Bm0k&2PDeX=Tc#6zh|N@bD6Q96AAma%%+h*|AY+yTatGr
zqf55EO%c@QANxOyfQK;|huyaBEZE74s9?rJy)K^yqFa8k1Qb*GF&S;ohzs|Y+;Mc}
zWt?qzI++s?6NY>2&7BN>=Xr&;Tc9QK8tNgLM)=*=fZ&In<U`MAzPZS^Ixzl+|MTim
zM0x0>(`ys%LT6fTo}{hU3n-_$qbII!AuyBdcvRNAgCN@2jA+)Hhzyg^Jie2R{*;}8
zY$=@!tS{8A{v}*nJmo^uHp?NyLYJE4@B`Fg+QgnYbFmoHN#2>J;QEOX&#cz*pvUJC
zc2hv_f^w6Gm#IqLLVlk*alEgNJJIRa_IYQJ$%dSc+v8#INHQr)-pSa|xVd;c$G5QL
zUi}yEg85Yj1d@=ho*zwRKcx_)sCJ)>?2YLM>I*7X2@cf@_M%*IAiD&KFBu<IAEM^x
z=1ws%kUF!esT*&rT?N9&YXZ~qU8c42kgh@_k-EkRy-3KvN2Xt$ujJDX$Uf>C%H;%N
z!&ix^_<DZEQ27YuPGMx`bmQg<esJ3O_;Dy>LLJlDuS+Pg{ujA}z$rDx9IpLoangIn
z%c~bw`zQUFbSjKv#lL#>$s8Ok8va4cF9mGQ2%kGk<LJ|JyGnZUj#-XKE%y)5z8jQ2
zyVYyZKkk60i{Rd=(Z}|22rFyA?L7<l%P1nOo)dbVpgtGTz-1=BAqWDuh^CEO_LWwU
zimqmHB$2I3{4Z+_`H)ddmI{4cBrpXoMXy40j$prKg_r=7vq#XVWE6?T&^nRGmI@2R
z5t??HR_SzuL_n}JE3Zdo75v=9r&%pqZ3WU!c^JaiP3JnC+9M-P%BQX7Wi>axiQSc?
zjcVG}L%lYq$5p2N?m>Uw-HL?A!K6c*^+b&@qq2HZ&-LzfFf4lChCRU`0;CU{7|tNE
zya$OD*y2X3X+Hs?cm8%dfC^0_mb&e_3kBY<xe~*V$7f=mFy0o<yoo+#{>j_dLHM~G
z_qtO*xkqyD0XYOs*UIjDySJpn6I_8=zBRuzbNgwoXukkH9Xf8&R+{4C+@^?G$b^lQ
z7!~pz5d9$mJEI=3zrOISWqRwrL-|&ophL$ZOFr5q=oAr=5v!42i1fp0gl}T3(-bu1
z1(mElbt=r%yNGRM>%mjuM@-QE5JCD8Q~%T2%JeKji=6|QeL>aDTm3oS6P-Zb<?x;T
z<)7(gI&*KmJn6cJBJv0bVtD4r_;izx#jEm;rN}Zyb5PE36Y)-0fa_Nm|ET1%;Kd&^
zbr?S9u0=j%F8c5D&Z#xNE6`&D$-v%tFR)UI9dtNNWV1LH%>B1V_#-wq##do6<U6eV
zB9%Vyf{{_u@3Pv3%(-LNtv@P`{cYJ!4YkjUM^!R!Q8p0syyYF=m-*lz0^VJ3epi~7
zk;v$m!=k%N{{VZJSCl=vEaAuYD<Iu8>;ei~F`>V&7{Ibgh-%6{Bk2=$x0C9#rHz^D
zcIA6dj3jA4AGSO%Quh-6nHaz&B3u|KE!V2e`uL{{URqFk??o)$<7oh5e#US=Q1t$z
zpDPh?ViAXVaE&ZKM_7b)N4o}U3Y;doYEfuZNHJwM^{B^V)Tn}n`ROtffAfIbD)$vZ
zvo5MJ-YTf3BtRY>Jm!KL@^pTlx?%M<T56y(!&SFEs&?8<VMByv_J@o3!OFL8rbOZ!
z!r>JWY-OX}(SXfNWP&pJ_x%FDrJ)ms6t?v$7ISJB`#b(=qYeD&VJv$4k@e-uS6>m0
zpsb#fH|6d>?_1aRp-$!H?fNOm?lMSP=J63R5U~afo!5Qy*Q#f<7!PDO_#_KSiTwf(
zs1tU@YgU$>57kPkBXMUfk34Jjg`6*?bF=2`>t;pPdmSCyUCb`E9uwB!Vw>ixKn~%l
zi99aAT)VxEm<SxRi<4E5l-)2-dR7okq~JH$u**xzv$TL9D5FXW^)E=GXUCpb^Z%1F
z@D}Gam~`he4MXOsGO3vFoie7Iyz6fw9u)#9dM=b$p3MS^&X3v`G*vf(FRxBd1y4_D
z=Q#=*6LptkeC%+HzUey0NLrc?UZoCnX$TTCLUDaVi+IyT?)|$LQlpa`oXn7t_$;63
z0KuDC_igyc`N;`Z9*|D7_=rvF4fAh=vpgGx+^*%y27Yr_rymQ#X3C_bDl;S=dWHw<
zn&s3-kb9E0T<)3ciaQZR>-#&!{Pt$cFCk!6h~$siz;GiVoTG+Aj1<M1rGr<t6yCQQ
zOdRatKvMdCctr3~$-$S8Wn?5n0`ZU~o34lbw8~QiuMhK=6R(8>{i?KdYCV_#tFg$n
zx3?F~%sds59)iisXj4QpiO@Uc_o!+>?n6y}WXaMZQi2``{V1atRTX7S>Vaeo*s_`5
z^jwIBmL4m%T@V*XM2B|y;Q?abtl!Ou^<BbeY33kwfD?5?1ZU(D7*^G25<Sh9N+B)x
zJ*iFo$fIbz%|X*ea>wQ)GPqrae`?C2Z2zw$YEwsmoK*Y^U7}bux%$oFJip;z-e3Ru
zpBsjkmk_!pG>)q7@x;!W`R9duVtxJkv;v4Oy6n8fW0v7p9rq=*C!8B*L=1%CDxz%*
zhq+P27#C0UX8mh~?G(t#vemUjsx$OS1-8^#@&2>V7slnBrB}+^VYPC(6|fiXDX<#t
zi!JK@%n<Td&eBnbD6#Q#_HhYnv`Vxx$_!7-YHIJ+xBLFNgE<+KO|k|*&!}eocx)MN
ziWt49AfwT(Lf3A=8Sr?5=C(0q5u}KW#8WhGhczy)b@-{jWX|9A=^%N8l3pzdq0ql*
zFB-q@f&FkntWn+YNtFNVbj5Lf+SbD#j~Quh<I8qgJ1aTYULch8;B8Uc3q#U7urHcb
ziM0d0bttHJP44UMWiQg;bG_lyn=j6=YCjg#Zc$XA=+mZ03?v{ki9@Yd*KgQw=0I}J
zl<jyEcBl3Irx7o9OJQV@PPA&Sq??AK$3?pp?XN0FX<3asTtSC|`DV><Pi*ni@1T<N
zRa`^W(EINAw4}2dV&LHfnx#4G|7!Y}$S(&-sIm0KPhY{<sYr(WIk#i%^z`)`H?N*Z
z9P{R_ReMqMn0sm8#3ds9LG)96KEO8Hy5xF18j7ZVeUDUI?KI={U*A^$&dN}>v<{v$
z<`FS7$It&))mcYH*@f+%5)42v&_R%rF6l0Z7758ARbmk7Zshe95a}LikdV}&8w7_K
zU<m0H7&>I=K6~EpoU_j157ruGJ*;7#XYYOA_jUa)rZ>66*(}R^3nk<@SnNHc9ru{Q
z&*L+pAUx&l84D#}?CA#~ZO9qcwQeyCi1+_D@7<TUtBpoS6pCnmoL*&ud-=6}*!h+C
z7qF(fxV!&us$m@FRDQ!X;Ws|u@dYaoL@Lcn>}b4qFrtB3A{c=uJQ#i%)I|4>N#9pc
z{T%(nfzwL=4PSfc!EU0mGSWurdEM?>0`nVq^)sn9!LJr}&#Mjpi6zVIJ0Vd#*H!^W
z42VICb>>O$S-AD7w^hCUPeI|qS5hg_eks$UZqg?+58MS$Bge+UT%0eCk@4FUeSjbf
zqnZ?SfblVi$UvrN=q7gyu^UipseEa2OwL@7mG!f5l^BYStatdf2{kAn4ya*L7YnHk
zHfuDHTzJkYs|gEQZu!-~W(Q7AGpVuMv`wACMD`*Xq#f5b)*5x}-iPwMwQZa(v6n5@
zhO}$nH|mmXg^SyfjYUtOG-OA3;womx$}Kv93;-N8w<@2f&Im8Q@@u?Mj5EDLTNMUp
z(AV7$n%eSYCs)fm@ErVcx^w^i6Rjn~RA_N!y}HC_RqL(ph!0Y4Gi+Z!c@H{wEyaK9
zaoXkFq4KlW2aMQul(mg!&z@l=Z5?-&;$-#sFPXL72fC^Y$wP{LYCaz8N$J&`G7dNN
zDl}+xzT^(bE4S=%X4k`_Psf5VNo&(L3NvnI)ox7!f&$)I8TOw<@bdSAH3gf_3SHIh
z+KO(&5ySi9eUmX&7j*#)u?+D2Pim-4{?}Cg*Hy(ImaliUe*C}__u*c^ifcBp+Cj1w
z-#s&C=XZmvest5vG1U3FgME@8=3pUdU&wE^^IArQhKYZ8?H%7JiGlf<ycLvhn>sFX
z6_N;(z9$W`wJ=xbm_;D!l}hibm$+_%RwHLOym-cr75W%<`HA>4+5Bl`;oN;=cj85%
zpd|JHw*hYujX6xgpa;0G-^|025Roa(nY?Mb?=LT5GL)*<Wlb-aUMPpob*qdT%39`}
zPmn>5Jv;kYgKs*+Lq688q<SHgv!Z=&WbMn`YA{@kmAA+&M)#$0b<Q7nlel;3RJngI
z%GJ*(D5hI``0|AImT~x<@9~?){P!1KA0~gJf3hpyVzv7IAM2q*m5%<}>>ZcyYA2P<
zosOB;?SIphSA044_LHET7WK2JTRfCyXQ<q;=RO!pR(;r~SI_lkh5A*J*<fL^=}C~N
zD3sHZI`bX}G)9m1f!-)YSXtkwN)Wf%%hlTtpCIoXo5_Td|FWCU0MphtKRBS;PeRV7
z+pM<s4_D<9%LJz(T!+H4jN?Ii<~u6=hxaA-kmJc^Vj9EiAHuy<{`RwPtdL4u`1Rx9
zq#qs1juRcbQhW=O(FT+cZ){dVUcS|5gZ2`Y2=~j%PBR)_#7fj~G5-uS^x@-A{uGEe
zh4+6w%hy=O(Cxc@BciozCHY1REfb<{AShkGl@v?3TSA`x>g4aA_Wrow9T_SDee<5l
z<f-c<NGhqbWVCTOA7~Cirb-10r+9o$Ly_2*kJyu%V~r|}OTm!9AATE)ji;}<ZC5Nc
zBT%0d$vdf;v~ynHx9o*mI*j+=%pB7leAN*gS@w)oe{mwEm#N&K7_WK(E3&wp+#G`Q
zZ;uYGQ$O9e={qTpG`(s|HD7(PpZ{S73Qu5e#cSrfDp`wr2eg?0?au$=V63S@SX#^a
zJKcz@ya9py2Q&piy17A+^*C1sH?u{Ybw~W{Q`5PZ>%HH7<qTBk(0S(84se*Mx!e4J
zUAfAxDoODF<{bdh_RQyBziU9dX@*UNjE0$3EAd+Z4GAR}4t%=>nrBN3D@S<+&0|Uq
zU7l};3AHoo=|!fVsMi7J&#aBuNO$1EN>6YU>!Hpy)yC;?r+!i)GKC&g0`UM12EsYy
z43G@zjV^(8y%GrN8OVZD<o&0KvNq36zL2bPRC^DnQxrLgrh3Q^u+z?N0|yQ;tAo?T
z0%$_6&_ND7f%w*3>c-Q7qwgAHQ@zUu@>O|W5t6Q(Aev8J5HX-tv^CL@=9`O?UVr-4
z5YyZloB%NkQnp+tUt1jmCvuP<_^UuG)02*)#V=RJJy%+f#-ldlMugs<-w|{!>o6Lw
ze%z1q-k9qI^mKIlWf}Q8Po7!uC8smLZ5pSK{ux6=d>ZfO^3jV$A!gY@XR5Ap>3uBH
z>zDQ9<~?FK#delZdpcpEK~RJXt?)g*;+v**#9a|8w~TDgZZT)<r#Xc`Y5Jdh>>A?6
z8CxVT^^d!Quj{}-aLT#+95kJN_W8sicJ_8X?+Vxb(1lD-EiAGT55r`xTC!+f{lz3+
zO2{8c&zfIsyDo?aPy!j)wqR{EgWO?H87dZ)cl60VxM##m(j2!z`8n$kn0|v^B3{D0
z6ZMbe^?97SVh0wB?a%lp+^L#-14^dF7kZQYdl1rH|8Y<m1sZ?cdKo*Jv6YUNx3h-o
zB^@cCRjO8D&+lJKzY7a0BQIP9om70H8E}&U%~e_+2`YWlv2oQ3+4jYQ(aWSp@`vE5
z`x1CNV=!2#c3?AEaFnW{G@5{cj!v;-@zFER=)UDO*FtcLppfmOGyUdG$F(pByRPGR
z2DGQ}glp>Ol%K*@M~W9;Nj~dr9)!<jNmSA);x6v=?9o}S)&I$2DURP_t<4U8<BNN;
zg6I~Xg7k~1=CfYmds@pPLjc9*zXzJB6)d<#<y5TXUjvA2vAZ!34TvIth!@qP^k~ZZ
zH~-}~)X$1}aNtbtGbd5|*mb(}N2bJ9M3J-OvWAh9r-};cY-7uK0rFGDg#C)+$CB*`
z3{YHv=K|aC!Zfscs&ldBF-7;J>SZyOv>7i!$4d(F7z-!!<<H5AlFQ%J5|bu4bSKqt
ztCydjGY$Pv*W~%ISS12;k+3aKX6C~a5Cu@eK;3^9>DRxX`HDF;MZ)LmFGY5Qa3!0J
z=7$IJ*4OHu)hQejrhTMsl9qtd%Fn#%9MBof>J{H@z}TU(xUOc%<U3d<^P~kWUqYBX
zzxL4dB)#0*a@X(z517=Wa_)jpLp4*K4G`hhd-M{(rVS>on)wTcYaOWr{AbV(3eAZ6
z>mXo~<Avxu$Af~~FK79+2|Mhc)>vu2G`p-8m+*H=?pY92qT#~0e{$FphRiJ?7P(TD
zRF%_ioFY6S7JLpIk>#v=cerJS5f}L8_{0P(8O1-S<MX)QV4wR@Xi*TjPukmuDZY}}
z!u0HTst7vV&^<V7Y*9Jrh&FI9nn;reX9Cr)-`xA2o@1BJtnN@$L$m@BOhPzlEhFqX
z1@N5^<Tl$4fp{i7m>lG$;(a!GaA89<ygnmSb9%~9;w4T^4(o42U~=y4Z>@Ot^$<L`
z9f0tHADT1+Z#a-3N8zJ@wa&M_S~lo|^@CP>f*7<AeYU|+vawU1@4$~el>W?;;Ub<W
z^hZlW@N41=sr#FM{$z{Y&Cq0+8~ckN3*i-Al+?Z)0pt(6?y<Q2wX{erom=fP2Ee>J
zR2qeA8S$R4A(=)``4ymQ_z5H}art~wvAFC4eMW?CtAdY+(7-(Q6lk(`9EMnC0Ba2Z
z704WskX*D#{Ob*vjjfQ1M!L=g%o|D6UBXSwExfLuy4qObl6|%9UE$?dm`O4}4H0*?
z<bXw4XexY#-HzO_8j6<h%DPd4^L?}1x_TFGOg%XKf-jV#X+5UwWf>noS&=2%HS-52
zcGv?Kl^!~4(qoWoH1tDmsd3!g>v})KfyspoVt+^bmHW^X>T+!MBN#e0v2+D<dniB8
zZk$pic<*~>&?2y@<v(ae)Pptmj$t;{5v@=J@7dELrnwxJP4@`8&-=W*j40VY=4dW0
zfBfp|6u6gqeEaH;Tt~~9*A*ZBBfv*l0T97|!QvH|G^4tJ1!||2&jp6Y^-x)}2LaRM
zk=Yg`bjBZtf?W|z=tZVi`BRSnaTLJ8ZLz<qmN~=$mGva{p2P+TZOF}sKQ2pvh+~9X
z$wtNSwHD{yNcoQ8=O5@C_>$4v%bmck;j#0baX01uMz2jdt+zCaa|J`l!LmfzN}tV_
z)$=R>qIH^)ZBrni+1|#M`6Gb0eY>GLbp31W-hn!iG8rk+KR~6t=OcA;NlNPn2xC<d
zc{~0y^=OryhL~7=IjNm+5~{FMto%Y7G$+xFarNMmZYa^9KHT4wBJTjhJFsx|_Bq$z
zHrI0q`qdYK#^GS)&Bp@$t>MhlvY&d^f6H<cJ^JU?#4S1f_P-5*cfV78$I5c`1fLe3
z>WOJF)tw)gvE&VV@8x}1A<)?^Iw#^HdP!tH7kK7?pDP2*SLNdK-zrOO$>#o$lmp#N
zdOGW^d2_L7LCQhf`4~>a{DITV8zxw1-Q3%OTt%bXLV;f++&%R?%XQDS#~WFD$vjza
zJGF3^e4xyWXOaaE2heH?>-9^H@5)~1+Hb(8Ib59uF@VFsuEo_x&{o-0BC_s{cgyyc
zOtZL}R6@GA>H+(S-)c%YD2zQ@mGZyx35_xJrPWMbB9+*ncQ%+M4O0VXI{mr$iQ|NA
zCQ9S_Z^JWwyFlpvDFSdD&;n^_U?v_tZJCfZFY0(#mu}k#xI05HnjDWMrm`u&nSbn=
zpdm7jbPGar5yo#IhNxB5lxG==J|vgxFV@6;Uhm0h@IOvdp4sR^A@<}?>SydDb(PhM
zcH<?empw9?$R+H|YnK-#7T&(e{lWXFE00q1j=NP{qngS-BI>1=<N)gIhCuRNv1dq1
z<vMS|xPA2kg~rV!p}a_(FyLR9H~HSjAA|w|=&#n~E(*r&Me&<xCRjoqgWl%MR$8-M
z_Zezx_3r4X?0xK?@qXrZSTl`CFtM10RqN}pkQiicPVf7Mg)?6qD38Da#%Z?k6WF6Y
z-&wiaGcU7L-D44AOpJHQ$`b}n8*uG<?F0+^1+3ot2#hIdJ>wdlzR_PTg~Y7CNf#!f
z)+!epeYNMYMsxj@q*%=Mdy{%>au<LG^uN|&gVv)aBD*LwY!o@=yi70aSnn!#rulzJ
zP5N+8r_hQ8VU_i9?LOb=osTuqa}Pc%J%}CH?9DZQEQ_^Q^>8zIPL{oMe?FcMPMWY4
zd+7(W?8IrGGWEFgEhb`2H4|kcI2-hD%MG7I{+OZTj%XfyU9oAeb$6>gSuLwIPoGeJ
zjUv5%j8$$vUTZc7tri(=`Iue?G|MKeY984?IAh=79s-LTJ`+0id?*C(*1yEIoG0U_
z?Atd3#xQow3m));yL!5h<34SSQ5kg?-k`fqTNJ!aDk4<&@(x6in&QsKd1d~y5?Aba
z?PX$O{Ed&{I!Xl6LAS<h?G^DiD7X|}69gv4hY_~Q{T4R!6`o*J^y}(!?y5mdo5j;>
zu6q5q%pQ=zP5s^jD%^Cp=8*{Xs}o-<i8i~C-(NVXU+f)rt&57`+UAfi<ao_=WH0N8
z8LL{0!U7gZ5q@${DptfPM*j$Y<q>M{+QEl*g5Q|}v;siVaYO*ZLdeM)0DLHbfp9pc
z#cuQ$tPmYb-D$sPe(j-5pBu-c9e+_2hc2oA?2wYy%Dl)ek=qi%J)WOz3%JO-s{+UB
zN*vc-kbW6g2|WEbpzB9&bMobNRYMPLhWp>7KrH~?qu?FsrK}SRVW0qeG|;^fdYj?D
z=!)~hnig<#jqv_DBPkzon%I%8_PATPw?n|08+2!0JNA(|JWVxv!|+W65nL5b_Hy+4
z%|AdH4#wbfLL6VSt%KKVxB^D05vGb@aJzdOIG>yHVoPKaSxUnWh6$uxsPwVac`DKl
zb_xC_Q6jk$_B0m$yGok?Z&@=J*x=SF|M~W|!k|!<fRryvFqE#%jqc+HLwfd@Qv22#
zclvT5p@c);H*(MC7Ks>&S~<(dfd(CGV4$};gZUF8#S+=T^P3wt@ojHgT0bg@`nE%(
zdloh`^A=UeYMpNfW}Axssd!dl4{rC$^wM7|d_UHAd;#k3mRr8@>4SCk@Ko^y!58=M
znIp<4ca(1f1fkQX9OM3O`z+=P&?(2;e8E}mgG#JY;zhYt0USS)SQ{?DE8>eSF1Wxh
zSRZefV6n_yeA;M6tyQr9+mkFgvAHTIH%CWIr8;eusz@KwNb#9;{No>G<ed~Jl117g
zDesj$yU((4O4F;iVKPo^4`eLA{sL^=B?sr+Ox~93BL<Ac3-ND{A5hOIOW;^@gfQ}%
z1fBR&WArDiKv&cRE?0x2gv|C(Tenp33V(InM;JC5)`$u6MrQR4ZM!EMl1i&m3X-S`
z!m87_+lbq49X!LBv5ibtNY!hU2bk4|zZ@u3cbu$Hm?=c<xN%gFotI)uxToTS5hc+u
z<;0c+F*iN^vx*G_i)S)TQxkMC<q@;K`EuB_0oR)+0XT}f-)e_q0X?2)_wst59nAKF
z!Nd&fS8N>E^?8{J0}vPPc54wEE8vmErx%$#vx20hSV<=Giq&cqe+r$_W?1o`%~6<t
z`mbhDWhsI<63T7%tKFtzqK=XCxynb{Th9LSQZ7R2dz;EGVO+0Xa=ty3z7^#6C5g#_
zu!csW>y<ghEZAvgsi(3p^}jIX(W4h*1U%<WM$ifx^Q!DyA4WEu5=-aK4Z5;Dx%jVC
zO>6REZ6%(F{iRG&41ea5)WIrH-l-6=x0ES__-2}k7(9-}Fv<AxRiO?ZJ~HHe?Q^*@
z9#a&UwAkA3^S9a6PrzMc#3|>%T7j_9ZbkOsmmGXccY<pyHwN+^>*n>9?tvtsKqjW)
zE|fP$J92DY!kYW<JURjK;!;=M871cPZE2bZKK0NHFS9mmbA$Ag3sacBvq8k(GOT!)
z!QFq)mE4uxDe<mn=dLL{c|@*bimDJNBCwJ!1XWpuXlKz?85qqJCP}>R(VDCK^hr6W
zPz!lz<pC$1+8YRiD|eqfT-UndeV0t2l$2c0zsNp=6S-iLwI%1C8QeMN9d_D!o%M%d
zs7)Z?>rPoQSaz_-9*>cESq2xrhXu68Nmvgg<z#dbKhodw*#4kx{7@gQ2s2wsUYzvl
zcb6C@R%Qeihm(hkU8wzU$-2QY+3wKG9yy_MhMF;_=_hIs@{`+SuQ5!KBGm#Fgqi@y
z!iA=jj+%H`ZD5HoznKfsANi!u?)(Xwd2vmqXV}%|K1NpC*a0(|U3d>eNY!7FMiJc;
zD?Sg%wiE??!`Z&$+xCnGuiiAw<ZondO8rAK$1+U&<i0nWJYvsijcRH>l@!)o17;UE
z?d4L|nx5_-rNSi$9&Tcj-4nVjELNS2I{?DMe^u&QG=X~tz^Yx?Wl8PeJOKv=XyRgd
zPhc5u;nEr1%FD<^njNTKqMA$}v-%VBbkl>-J=O2oaM($rDN5F90+Edq{64S{eF#0_
zmgymwref^#M)zRsi4=ZOQ8X7nyQ+YsVvuw<Vwsx5{*5dfNKKBHNGy4bODnyoPDr9(
zVltHRA`{NU{FgTB?tTvs?mEgTyGf!yr*Gp?83hE}I?0=bskdz;IRXt+)5^E1gK`vF
ztRH!$geH9a6{>nJ{_Ieg7rGjdMyPIK@E2z`R$HbzVS_5ewoad{dCe9fW{3<Wy4B}G
zIpWt+eypx_lO>bFLIT|oyR$b}`V^Fwg>93WbEGksS_&g6eO=g}VQ~d+PxU1ZM)wRP
zcg9M1%{8%#W@OwZ`c>U$jv9}GJ_%uPHC6+0Dx|U%hqc28b95w?lZ*9<Co;;eCb9x=
zyrOTgOFDiYN6|$3O0nHKctGKDBm0gjcBw*`D_I$2P@)+7X#N@Yg4pGK`!|_0ggiCs
z0uNBbW1~Tp2I{S5x0_7F+z4j4^f4T_n7mZEMD@e38CE{xkvE!3U9sqZ;4yh?d~RjE
zdpcIqQh}RcT#Ih44%e)D|Fvg%-j2dw5?NxO>0oWi^!c&RTt#E|(%MG_#HCGC6RjUK
zf$(v%4weqwNQPV+Ehu#ZEo<o+PNk0istHZY{j#31xEI*0i<1G$119*OUHdpLWT++s
zu30SNsrdNQu%HoZ^vd7q1rfJ#T5k6oOKana@Fe=K$v?7R*DcKafB9Fw=CN?tmHqQ1
zc57^fDM;W+>R_t^f7XH&%Xr}ry&DOP|7yG|*Z>h9+(x2ZjXV(bo5Cic%R$lFl_cMy
z>KYp(S}vwqA|LRZyQ4{45>T{UUMajWafzIdPc#XWa&KjFH5(J>A2f~d?c6FOtdG*4
zA3oavwoSi)zn~bdi8a7dI-Cv>EXu}iiO$(BJGfB&*?QuWGIjt`m+O7p!8h>&-v*|Q
z?ldFP9Ag~a`Wctp+X|oh_K!!U;uu&^&H9Yv*^Uw9i$-Nn89|~JBUc93yeNsO@C`_O
z=gHeVi24=XtJuh2Hx+e{iV=>ZhGi8Kq^CV=`scN|huy0BLDqXAr+Evk!x9HF9-oou
z$uT0IvZ=tyf6w7MHlw@j44ry;seQwxp<soXnScoj=%EIj8dns+X=4G9TJe21U@diS
zJ*M>!xp#vtg#5J)Imd6{J_VWdK>hONeSVQ;n4op`9YfNTM!3hWcS+4y8&l*Lbe?r#
zD0zG?oYMJHpXr!Kn6DX+X91#hz&l?AR|XwSe4}ROemT_N$K<%~;3Z;|N5e;kT+aQc
zEr5lWH-jdsWID>ODje9M9(U2QY=oV^NaiIjm_wjgJY+SG%^vaP8Bz~6Qn+bc-2Yq7
zL7ZP&nbDuFB}!T3sb?8e-<LcWdF?g?V_qYNrt2Pey<v4AbH+o@pgZd}`HW!eBy>|8
zt8RvKd7314h@omp^LO5q5Xbcur%8VgB(o}u-ku8CbS5PN^Qt<GrZBDeU#>SqVZE;^
zBiM%}_V~yzRst>@V~id@D%W&-g3^3FW7!3jGmje6sj9N33RznDAgAjqaX}tttr?<T
zgeZN_?6VNHN)t&)p)boYZh5GtM{Hkb3RL$|fJX?F4)`IH5-cP_F0&zb<y`>6Q?-C=
z9wG6pTj9eoXfEyIu;^NFH{z+Hv(OxAOYD-$FF|&-z;hUz6MOT(3(LW?YN8hj!94}-
z52aqeZm|jxe)OpKInEy%U?#!IKb-Eu`hi@}Kwjt_vBSFdY0_OANi9s6V-?S`HD%T3
z=GOrAj4ymZ8x6=N>)Suq3<8x>tzBjfOMK##QF-#+$3b%%q{ioNRS6kN#HO7S8RWt~
z<}1JWH{AdpvQg)B);N>SL^Cmji8JJP#YBMx4M?=wft8b%?xj&2Mx3vk7l7Ii^t-&q
zg7Qrh(~RGLz8QDXFPI{#OPN26>~}vsMwS3^KE5&bqI4p*R-Lk}3))ycIYryS3(vY$
z0zhf_hK{DDCb%P!4ZLORJLD&FD&<0f$LamQ`88ePNDvhQTpxIiAegkwaQpZK0=4H#
z8stJ?>o2myMDgM>K5Vt=n;WpQJ-7J%$U?B&ia&1HsMG*o>w#a~CmM64LB`##eu>;2
zqAiEImTpE{B`AI8D)_SIyZqi*v|LI5QlQ!|BH$hN-G~s_CHSv!{EL7x4;D%D^;g!I
z8BfkF;L6yVdAFP8Y8;qzbI)%nN!sT`9kWwOZ=br!9PNt?pv*TWi1aZQUf)Ww%E&ai
zuHORh8Xoqu{rvN<bSa@~c9-@K`;cY&_g1YOXBi%nS6aYD4^Z=X>5Jb?^Qy;f>cre@
z{~)Ci=Umq!!)9quL9Yv7WgTt$T0n=-+uK{U+&ZxloO%7qUHS_J)#IXT5PE$!M$35q
zRC@Zai;Dof&ABt0VevD6a~N3L_Ge>O`k!H+Z#S2>filL$eKF73dbGwq^zwh0N+>-0
z1~5NX-s{^`?W1reDBZP?{!nyqO$7H;ZZZu;Q&agV-_UF};6fsE)D*aBmEOneSPLJd
zj|oJ(r08Ef0oiosmT3Xks6h76Z#O!x5mdG-9f7sOE8}R-Pem(BP_e;vPn9O=a0gE_
z!-+ty@D{gkT_>gPG8Gn(Eaxot%)-3S;Ye3g6fN;cV4VFGiI0nni^XV8o0oo}{DV8)
zWCh$EM#AP2Dx#y;%+2249gJ2ilu1M)`3y^Lo@_|9w*XHLcrOhNqZLt)p$Hb+Et|54
z+CVq<L{1d2NrDAU?&{pFhAGi{mws>iy|`-CN$F4~6P0dNw%{S16|<w>NL2@ZYAzc*
zn*XfI11w}fjL_6prX>4!dbP{eR)c)&KGn?hw7c}c!`*8zzITvzdmjD9DpkSd<6@lc
z(YbGvW7Jhoj70Za@EIgNcZTT9>NmkM!m3%^AY3o1{`8_fWw~Z(iKt~44>1VcLXHX#
z=-X+9_?i;FjCEXn(do%K5Q>1(WdYxmp~rIu55_ExP=T!FH>t!6SDa=u!Ua>m1beIx
zgEbRBn?0HfnwwA`iJw4fXkIpK27gEl?<}F`Dy|CF6P8?v(I2AaNR&76s8v8yi6q|;
zToItL)@{8%X&);k-2WHAnHQFqZP>|W^0yL3VobT4xni_AZ^wQYeZfV?lkqzVOsSR~
z;(Ft&Tng!7m2K94*nQ6{HAtE71^@m2{iuob#PjSpd<`&HxA!%c-wR2^iHRj8e98^P
z4qh^CIsL5(7K|qNqM{a(W_fE|FGTkuv#K4vtn;U2del3s`9%Ca18<H{KT3Z2IK3!L
z2|ZDj=9a|ow`+2(T4p9i(u9)Ztm&7~p4A16FeRM5hB|S0-ydvA|NVI`ii7U|=Waoc
zP|3gXfqJ+|QxLw0QnD}@bBo*nYYY%eRl<0udUn&dgkB;NDQwqP+|E~47gOHUe(APa
z>lFOsBh<P0%(+e9dQI*cXIq?MY-oTGLGBLSl18x<RS5aCTU*aKQZQ>-S<0;pOla4f
zde8wZ(AN%m@3xLHq?&BQH1pj($r(k`60`W>40-Sd=SCxFqXK_$S>wt>vwt9YmXL`d
zI<ebBuepTfKp8^O)6Db^py>x0o8HU@d$0!Wa>MGNu1#+x?&d$z&D5oFE$vb`)HbPb
zF%6c}@g&Mus*JVdaLYN}?xdVS9j83p@J9U=CUwVE_8R$RuHD#E9o?ovH~}IcxD6P<
z;zs{BZZ>M7w$xYO^YicM$#MfiO|p6Y0A~6>NAV8OLrn}_mg4V6YzHkvA~g+`EDO1g
zQBT|i!xI-DE)JM@=wbAN43aUbiZ>B9KmUrUz3j4Mi5`gI&yig2?s=>8>yeM^{)WF|
zC5xQC&x8}y){H+$r^|blk2p~(X^9Hbya&6n@#;i2WnWt89O2>~D3;n;qqrVLFc#c9
z{<$`T3;uJ?G#c4f0+*aYn(|qKfy&m_R{z7hDF83KzP|8CRLIPI1^PFxdrU6cMX$uG
zs}P@D_)Mh9ye{Xau$6p}24t4>fWfEhlrx@F!OW&Mu^(O7V?*z1)vPC;E<5WXgwRml
z;VoM48xq$7&7d0U@^ChNB(wZqPaD@av!H2kd^qLwA4E3TVXTiyCab+4WJwd1v?r4$
zW-KP6hapCJJAV)BnsOt~t5)o9ft&!<i}rxWvp|VV({j@11WF$ih))q)H+%Tv_Rh1r
zCn1l$(rAw@)8NX2jEZm0H|gUX-1jlPoxlU~25#K^uKA^M5XgTP*D^;@Sszs5A_Weq
zV>AAMZu?HUBV^$nxm*T2@vEiMp=-3|>SBUD<^;<JvB|x4dwz$R)>WgFCsJaA>f2x)
z01S|Y)m09F%c}R&!_#|lI6=Js5Z`SPjV?7Po_(CjG&_Mh@o#pkkrz2McHQ$e@nO#S
zTvD0z6Zs?ma_lI)NVLLU(rVpunWPBWWOkmprnk&k(LIJ&Ve?-u@cW9Y`2ajLNK5#7
zL?<o>aeCsV$j1_JC}V|TC3`=yqc6)x7D1SIIoJ6n|4~t>A+^<EUfwN4l!O!CyBflg
z6m&+7XrWz+cP79lbhI2lxMBwTxboBnoW9ea61mdq-_O%}HT=|8H<d-qGb6XDm+%MW
zG7$3rvER~nzUB^1yU4)qVW6rRq{GoGGRcivW}YA({K4biAuEi@#kNoAF}yDPI_I?E
z7PyaXu;a#>O~15QLyP(&RPq)*Dk7eM#6Tkur*nawVWTAu7sl5&+uPgq*Of#)4|?Hp
zX)k2q?H2h`mZ+v7{32d1=qOQl)9O^ig<Qb*%LM6)z=l(K^xJ95qBJHS8$4j8q9XR?
zTII=DdBYLXgmQAK6;i~{^wNO0PT#T&`9apwX}1e$ENin`V<35?v=qO!EXX$uKes-E
zE)B@^u=c#C#B$Wt1+IE!K5<*S1VdH4b^Ca|xP0Mu&@t3YMdS0A)@<7Mi9()lI94V4
zd8T3#91@suU!4WXj5up<Uw6t~x9xs?{YmZ5<C8P>jKn*iYm1w23Q5V822$}u`ELQ?
z%MLHQRev5<qv-`g;GydRk-D%QY=YNZu;9p))-`D&UzIkhWUD{7$EI)+&4%}L15w?q
zQqB*B$AT-*V@x4MH}9Q2%!NMoP5$?hK@DUH-MzuEC_g>#`=#VuPUOYGncMz0NC(D`
zl<F%4#))28E-BUNditVb>ZZo*$PJ^6*DjXE-3p0VO1vU2(6ic`r{l&p29kWChFKbG
zcq64~>OooTN(KX27VHUgh0?6(ks5XM;R3$>ZASl85#w9x$+}O@H~_2m%=29=f|J}3
z_HDkw%gqk;Mhr<tml2rroj{Dsb$xqsUUKDQ$SWMN@0hFt%UfgbMm|Fy_k60d$=!|#
zDO^A6`D3_M@VGwS{v&(PC#Le_Cl(_=ntNmHMYSFbJ~b`Bd`PQuy7lJ|6R3B9u{NRg
zUriw)Hj;HpDx{}SB=Lm6yZrm!CF_e`Wb?Dhh~7(1_aW<e39V%p`I9v3mXa;s!)f|`
zm=COphz!{A=PO(@p_cpK^vj4xY51X8ZE|n|)X|Mv%IZm@wRT_$Y&k7%32!+^w?teW
z&Rv@P2QUxJLGmZg&=%iT_i}LoV18fWSMP|(K3Ww043kf{?O;AlPAAvFYL9fSY?F1R
ztqrtNyKFVOCX~sW8(t$#krVQqSg}9(Ntl`};#t5WHdJ}4P8sz`CEZv#^=$~jO*C_+
z#u%5yTCw}KJ@T{&Ib&3bQn+Tc-?R`dbj#xGL6>u9UK4fTql`U;6D#TBsBDD(b4n72
zQrW_Y#-&Xs@{)BsiH)8NZy(FHvK~l9zuZ!XQF-8_F9E|oJ`EM8CMq04&Hg^4JD9G`
zAkhpi*OiTP!53%+8)1@D<An5*;k}iFR6>o(cQR98m;we&czyUo;1<0^2c7WGG(?*+
za8tb;r1FmN4!`b-uvFE25jkeV>hRAl#Yg5$wc<CIRdhNc+7h;)2=GdD5YZM&m~>rL
zhsp+oq)?FOIOd^6-33a+9;K!|P$?_5d>>agdVFGXt5;%x5znGAv<Mgx{x>dm^yRm^
z6W@|)2)bG+_v#gv*t%%G95!mmk+>TZJCHx^oivmwZv8inTC9He$bLcrtFr=T&Z(W1
zT!ByC(Uz9%n3hSujX!8`THkX~bj@8WDBxNtyy7bt$ZA7_!)g#;ovAgvIMM+jz#zB@
zgYYsY=3(pu^ZCm{OMe&;j*TAYI=Z+n;tK7#SOgK#K#v~V>jJp9kHQU$_1BjwC_ep8
z2esgFJHsW_)rs8I%JJz&`EcGd%*J{2Ae8Lm>2|XZd99r?9rCEa(CJ#lD#kSAXHhRV
zQ<W%ZHx=$|Tw&w->_&MjYMjutD*i*%!Rf6A!A^|GoAQ=tnWxhEi$e9ls>?@*^Q`?b
z%s?1VXI>3UBpyc}#?;9@(AVT)!J%D?W%~jF6D)}23-lB>4O2il<)?y{+Nt0N^>oYT
z?k7PbJ5(5`W*>d?qt*yPGC9v)_=&OnL9+HQG}UphyHPBCvr*ZC$zlXHvMRK<{0f#S
zv_-YrH95D?NtivW@T_6gVHBxGY9y2n3_+tV!;GCV=wH~<osowH7oJ!5e)uscv7*5V
z4H(dN$AT`#T)e!JaSNjMy(xt#LmQzKctm8tXyPzNL2{xv)Qk!o>H)lUM^Qdn;$JnB
zRphg8r=>S1TnTpF^x3YhYxa3%-E|%W&iLQhFHJWRt|YQ9Zf+;N^foay2X*f5pY~At
zn5?-8hBUJ{6whbv2bF_Ed)C<Rnd45SEIP9VP)%tTF%8fd!Dqn4cR!H@@Qo6)^Bi)L
z`o~mZvC&Bt^G=|Xa-jy-f%Iu;BDQT%vgn-xv7&w6f;FVw;KTL_yA8LuU2N-OV{KrN
zgR}NN@0INp?W~&Ao}is<BLBMXRt+1NN?4O037kZMGYDS&g-6I>wHW+{m|8+rM6)Kp
z!@9dO;l2ufsp*<wVY4$VWd-3$x9(YWwn7xjmR9Do)sth7c&C0NmF@$=r>E7K<8}m-
z*gWjGdl(dj_pHP3A$`hzI5DvqZucm}I$dpsKY4sZ+vql#M0&ncS&4Tp-2v?RJ%lM2
zOkeP|NeUlya+WXO3E~+{nH<SyMEg=tj1B6<p^r!uD8pGdR$xdrg+=L1BVB00%n43A
zVgDW8=ojSN&Fn@wu_?kr)|qiwH~+~$I(sD{^Wb2-dH07QVNAI3`$BAovA^%k@3+x7
z>l0W_4dzUSc-hji-M;D<dTiNlMv2)kX*X9ooR(^Jz*GNFRd|`zk<OA&E>^VQY$&Z}
zc&cm#cQ=C1YJRk{`O=RbTFpaD!TWt+l5(Is>ty4~<W;rwpD69O71tH2{Zf@vj)vfr
z^@6sr)6uNG)6hW7`fw2KLXXe)2n_`C#ap`ZQ_YnXr{C`+m@^Z_wQ=#(6ZK3d3)C8;
zEcdoz@90#q?L5~@h(^bqgiLBS?{^x?s3?lH$5W?jW}086jjZ&afu<N-;#VXL^;`)-
zoS$Js%GOYC>{I2~?{;C{Ql8guQ?Wg{4xVPgg+e%t!?VPV9-6@JV;ech!=EY9VsBWC
zK?Zu*jdbqZ$INCz<>ePWHrL3N$(sU3NC9(F`6QaqE%C*QP?G~76fdz4U;8JN$?s1+
zlwHk4acZN!c}7;b2cK1=DQ(R@PFIV!M&cVI_?@%q?AMgu8uye`CPs>2p;KJQd0>dM
zXGM#YlG*d-M-#{-Qx0QMO*;ieI1Sf+UfjyxQCAcq*lRJ4vu<f@0e14KuxSsQX5!i{
zx~S*yaCqB*mf<^M?@3ES;<HOICa_bkFy1PB9|8>hN5Qz!IIZeRg#i>YU|7GZF48*k
zq>1(2KlYjTHv&L<dI_;Q(8BoSw5@^TDL5ziVhwdj#d;GyK5ETAgm3lJ)!#kqN{%pP
zNa;kEqN<ZcgaZ@Url3R22M&sx{}v0G@29oYnV%j1&HJ{V@Uf->$qL<K(z<`VbY&os
zkw2-On&*;QQkj4!9!~?)m*{I>N&LXs$=|tLwPRTMRV2$~C@8;B0DuaV6G!WIBgs|q
zp?A1~25>rv#;>7JX>uYmv5rd~Xpq%Y?&pduDHoS?!W+sw4GMh}jVtKp4ak%x>pm=X
zFITfG#uj^hW-Zn$=?~UGDvJAq7)i-=Oax>fHicA>0u=_4udlD<X-b*yyW3HhhVK*f
z%zPIvHYJntoFtNyWN#;KQ8K$tf>k6cIvPJ*_%Y*%_cyAx-9!I50*k}({+h0#teDDh
z*Qs&YsSvkZ!<zQ*dR*!n%(0q3_uA8t(@zKR_k+RO)hqvIN@$um^c*kJKaL3TLGa29
zHQh2M>YdE!?zdjRT(lH?Dm;3nmR#fsm-dz7JhNC}AE&WsHU%_o!=MAh-9E9o4lqRa
z4Or7ra(H!cUBcnVlzOs|ll#G;<Us#vc&srWqB@79*JOt8%SkO@4`okw#Ol+;=K2O{
zE@+MgaoLG@bDT4~wo&XBh#oJi?zYbbysy_SO$!-vGx!<I$eR0M_^Nl{;-zb@KxK80
zB<tOe9ZrspL)hpa9%pVj7xq3p%)SN6DB5EI%9cbn;u=hLZ>?@;p<(d|7Z3qrp4M5G
zd(+H2>Kv%M(z1y)j#%Xk9o)QkvSM$c&AO-_a+!{`YOPpD7j2Wj*TbH?t7h@c{b8=W
z>^>gKBlqaD#@Lr%HJk3}8*eKXjRxkI0%Q}!h+$t4hKncQfQNTe)b!lsQGFb44yXt0
z(an^*Uv%=GLl14coIuG5RE-?v3CkHjnNp>_v6UP`*RAv_Zh0BZ7WR&NLj%}ykpX)e
zZ1XD|M+19Fu20QuvMONNHG^V%$2$m}T+^9{$c#-`wYdaIj5L>?1brU*H0BAU4<N7N
z?&5x4dfeHWC{0XA-r)5Rk*b&$w6h%W!ai+GGiU-Rf8w{G>IKaC_V#u>HDD6YtlpVz
z`i!ql_vEhj948bS>Z?$FRPRLmTtnIGo;Q=+B;o-ZnaoOo?>N~cRdVE9x78i+*ew>v
zQJlffbZcG8!e8k*?6c-GImas<q=`j3)!e^tK^+h?Fr+p{U!97k^Peh2Yx|+o;05H~
z<W-idF#XOHc;Jou8(A;z`{+elzvaPej__obS;1=u=Jo1-TYW2JOktV&iw$uNm#ar$
z|5e;CPX`KF7l2^6Ywa2^8i|CcY5hh4Pj3t%t)1kJo1ZmNA7drd3T{$M&Bx%R+0R9i
z0&jr8pt#WAg4VktdtuT!@|Fv&y%hYB(!*;<;A36#fIDCavdQcASQj>mVCHPaJTI!*
z*&f;2zZ1FD)zh8EhIX@qH1~8%7m&yD-G~cJ2Y^9E8WI#9h+q+TVtq7Jh^MOqA}%$4
ztKVJh>{n~?JK`JN@W3w>5jvu7@pQJ0tjU1Ukp3e^iG-OMUy(RDl`kgx6zLjFz%wE1
zVo98jD}z41HphvfqON;KjeM}b;YO~HX0lvg5?WJ33siX*)0tk|G|{WdrA>z1)|WdG
zFWn#A<2dbV5aOEJqHf?hl}AKf(|}g9qH`msX+O7I9+WR`Zt~z&7~Idr<nI0pNB^9B
z^$kbQX#Jan^36X?IuM&oLdE?tAHjlc%6$ED+`v=k{5IE<DY<LjjJGF*Ji`7>hlu3-
z)(}8yGf@j!cOy+mXCm9I1lK%A-`R9~9{0+mMsojN?c}t+C-iu2nk~LlqyyYSn&u<#
z;<;VFC-h0025MMTA*|z3AscDaw#TRL*PaxgL*9yJM6w4ogPf$6tFxA95T4<l+I%m3
zWLIp)y}(x}*D$!Ew-Eg8@>5+W0+q#zC{iI=edH2TypLjeuf60uE%CiSlV!Gd7;fmr
z1+%ag9o#4n^FhW-&UPkFH|`)=rHJBtsSV!;?>|SWC_`c}>q~BzaFQ3;lRM_O{-s#E
z;q%sE-|Dht#{@~cal2r+6O9D@TEEWmpv#jLdc67tPdvhdEF2?B_@Yz-4*jIkK!W=8
zYacA1IHfeDU0=wzX+@7sdrq>CnN3FvnKy+a);fN@p$4L6snz(Ck&Ml2Fi2PzsrR)I
zt^&kGS*~PU13EEPs=LAawXsuX0lWc1c@!y+81x`8*0EVlGbA&d)*A^Y(HHbDbNq3I
zQuIWg!!plkYd1bkGFoPjIs3V5uN75S_~YV(VJiq}iGh(T`{+-^ju@7vAyiqn)lvRL
z1YKA|8r`OIul{_#67NipeZIGU|0Q>zW8@<~-m1g=U6=@4v8Xo;lTM|lStYW}#hB|a
z%jQ3=d;7*V_t^@D%LH+7VWb@s;01315mXm9hWdb7<3-PsQ~q$LE_1|Y_zI9xrz)yb
zhTBAa5><?l)>r7r^H93UH8mv<&!uA0p;jOw#7jOX_uEoEtU~%P<T#<`yuv8YVp)fr
z*P%8QrM=&&Nvv<KCrCUNdpMh9N1U2*rjh0|&TlIEn!(xuD4oo@o?mP!n^+t2ciII%
z@iInO?yi)vAY5KOwqS6Auv`EihjGE!0UfQScrj9lQLBgiHMWc8l>Lc#wbBGG1oynf
zD4AINGJh_eY@L3?bwyW#W`b`w_2BI_=6wm`yGsFJ-Lt~gy2t0`YV6ll<xj6$#A3b+
zof%#*aHS?xHO2W|6Fg3ho?yx_u5KJ**I2y#`s+Cxca2{O+B#FovZ--}NaRF2v9xah
zuaPud0qt)fy2INQ#Ujs0ql$kN2r1@9TNV+8WPh}++f{3c{rAmkn4o3ZX&8|Ni)Y5P
z&ut3I9q0Z^iPgrvbN`^~z@1lxuE!%3_PhnKJal0BPgg`5az~FXdLJKgYFKufKKO+9
zPnWHYU@^J6=qnRSr`59<>CnL9h!5!4SC(LE2|Tk^*uz85zh9qyF7m8wwJCm4DhyFw
z^mc|9>`c2YrS37GM)hHPT)PX7SlZx8TK-n^zkD#07=LJLyRoJX=#21TT_$`Wy;GAs
z>AUiQ2N9`?GB3E-W7f`7s&D5Nd)uUscwtt2CPXdFd*!ojZ3TNyYlDAJ%@uJ^IZFv&
zqGFB*j6fiFBv9OB%;0@~nL(F6E0M{#w)&;WdDB1xuQqb}SO(LgE^S1LHf>-&3r%jb
zLZ^;Negpzq@VD#ku>Q^X#8Uci;mm)%u&?Q5wI}#8U*G0be>ZQ<A^Z*#^TC7IBKzH3
z$FIPGH{CY?vN)OI?S-kjr;iT|M>W{;CK2h}ju<rhV9>3wRl1ba{j{Njb<yNc8+alC
zvBm`uwt%k#D7S#tNEVz8zcj)R>K!8q6Tc==QVzx};*dumX#@<4!H(@oP_WkuYw5vZ
zku*;YGQwv3h^LJVPmD1YH$L|)Uy?_saO@l$9P~T9&rwe*odZYY&K=lEf|kM9!>r%F
z84?EU`7Ws)!m`cCIj&~G^a6U<OY#1q=E6;z&xS#w`A%lCybNrqt*kd4_7muBAVk-|
zdWj#n9{>FV^`e_)@3Y)QRcPv~?h5|aB2(B$0>!2!V`CkLTHsUct$sTRql;hq`eZmC
z&%I1iiNy9`^;`3bo96e6D@g<`yNBVZ-%?`4BI~y;rre+`Ic{+-1<%$V#vCL1Y<8#S
zE=8L!4yX3dB~A>FMg0&6;(M}Akdm9Kl-Q9Sm?VF)<eVURxG+GUjRW%~Sd>9{W*A`4
zd?QosjQb$o+_foD|HQ6uD2P+%F=z!heQqTAs%LSClv;y|_S2DAg{1owE+z^+dbg=2
z%d$5CVdti9*uf15Ri_S)_CjPrx!Bhn5~YLG%=AwZ(>+p;o)ZKWpWDr;8CcL{yAohs
zUHnpakJChDI&bS$bHuYp^SU{GP`N1Fz-5CFzSSf)cjUja$%gUCjgD8}c)H=I8kuWC
z0r{AH#N|~6$^m=ve_zWE!^1JHfBxq`!HZK3OxFMZI(JRT`S}0&dD~NRy8nLge=aHf
lMV2@eT=@U|Q)K=XSwIW3U&cREYXsm+O-WO+?Afb-{ui3%R}}yN

literal 0
HcmV?d00001

diff --git a/_images/b21d966bde1fb74635e96843f91a4ec1072e73e79320eae51395ff6e3a9fe551.png b/_images/b21d966bde1fb74635e96843f91a4ec1072e73e79320eae51395ff6e3a9fe551.png
new file mode 100644
index 0000000000000000000000000000000000000000..e54b231df3942483c7ee42de0869b1d58db07daa
GIT binary patch
literal 96489
zcmZs?1yq#Z7dASSbfa{GAR-~%4N8o(Aky6p(kUH+bUUOph=L%UlF}dz(%ndYXYlvG
z_gm}US-6nlecw6f?7g4;>}PMnm7#Lj7-Sd_2n74(3u#pd1epl}LC8jX0N$}ls3Qm8
zkewx8s-uCIH=0Qp_!-^dg|;&Uf=K}XhcGhddkx;?cahO@QL{I5aW`}_h1eRpIM~>`
z*jO6TxS2XRTiV-ku{~#d&Prq9;^H94&i?=Yp3UCLoIQy;DjWi#fxMKKP=A-PGv|>(
zbebu4`}*}M&+DO1PKtIV@js7H9<EgBStyQS!@{~*_UKFW{e~@;<1Ch08WyDI>cS()
ze)28jy9z4@7O6;}2>c=VZl6M$Vz`&pw0(J_qRTTUBEs?7b%?uSePyXVYiqCg71HS3
zY5BR~D7O+4{=b(>8rr5F<i8&>f0Ja;(A~daHAW7BVgFtu5v}O|-lda&`he-bcbUun
zKNpV_!(fQUT~Vze*yOBY6EeiQ!Ry(6O*3c7y<Z9O@lBuK-|QKOO6`pYF0(vcs5b2z
zY2@w`KI|Y&5paHNq7C)DOMI;tezo5$)>$%wO~L^w(JX7rd-)Lp(a_K!;WERTDAP$!
zI>p1sKh?y7!oNKqEl_GN)+kxNZMayD6#B}pG)nVQm9K*1`wNDV&+7xJT&O%eJX+tK
zOiQkK?PkM#uFebxGX!+s>aK^u!kT=p+Lj#hRxhNbrBegp8-?k0Eji@aFElyceSgaa
zX$e3=J=njy-A{RLicz4L8kl{vW6tk=QTccFeBKLpr7ODq_g8Dj)A85FZNcalYw7kJ
zzSr+c^%{SB&;IZ@hW~&ZBmoiUAnCX`Tsb_Q)JKA7R_H~i+ts6+XSzNb<X^$+Pv^tc
zskMvNSx@0Ku6%cBH0!?L!@m;6bx@*zg8?}hmZM&ZS72)mMkjQ*Ih`~<-<xCdZp@X7
z?Hc>6rlP4ivDH*v{XB|<6VW`&<1xd_w(PrW^S3U;GASH}h~+<y<bV8KOJ|y`9a?YC
zjTMa`<uWt4zBsHq`)yriHC|lTJvKhR^eL38O<+AE2%YfpVs#%|yU_j*q#Rf6?!W8V
zxR5$<Kc`tg3H+JN?CjlBRmN48qXl(0zE``o{JS-y%G%l<Z9D|{_?Z~@ZqbYPNg1tn
z!B?nli9^rlRO83xR>{HDO^TSG@7syM;^KE(p9>EbgK&-h9<9O``B*6}_^b6~+SPQm
zDdy49QKjRGEQI4#Q=wA&0|**2V%<^*`2jfF<pI8hj$faoIQ?$Dxw*MJ$~M2RXSF9h
zQ9HQ3+F!}?+G~xW6mnSbIbS;JWA6rz=09WCfQ|Wxm4DMZoy#1j>GslET|*;vCM`Yv
z@UWXYU7Ijg_y|q(=0xp!-S3u3^m2oK)^$=BzP|VeJXpA5*BqK<Z{9OydlOK5A0QM!
zGb7**FxBkW{MVe*eqMOz_gBnUBBw;~JB5#W=zDs5<?EZz7yPIp2kT;Yq;UTQXBjjy
zqMG5g_q}Z5@NC9@W!8D9Ge;r_dU+Wh7WSMnh!D76WB=vp_CSBw6SV@k1&|=Dy7fe-
zr>7Sy)OTTrE8WJxX7~T!+OqOK8J1f+=22=7!W9hz_ROiNA(D}i0q*l)qEs76oIR2G
z$&*&oWL>idtP|zBk>P|)X-uohZ=KM~>y|O#o&xutl~^$xrP2O=jP~a1zGj2@A08yn
zP0&E}hBV)vR-|xS66`hXi^8u?*7JN~*RcETW7C@U7c*KuzR7+q4!sH(a~&|Oik@B!
zqe9YRs(Dsxe=3*Cn>P_p%j*zp=RKe5dF_79yAt-k;1K=$hw%EWwz<<|Ju@_pUT&GM
z`7gF?ELC)}zVEcKtc(oGBeu6;&VvG|`T6;9FaUnrn|H;^!h#6-v%a1S0&;E0d9A<o
zXti(Y_%8&aS3#WZdujQ*R|O!FCz)40u{bzHN-oD`>Fe9`W<I}1GWd{6`vuV`3V{(}
z+6VrkK9>d;2TM~8?)I~yj}=q!&W}$vN2?r`B)R$cjCX&3<-gkb!s&DA!uqBL&7UfR
zL;evY=f@8Uzsu3=O4GhaAW9@gxj`^fh<MhFzcJojXr|^e>kpf8>moJJIN@n{%XWRy
zE!O3=-xvxkP8dj&Q(B6Ztm{UKfq_wJHb7>P7Lt8?Qs_AJ(RBLy_+<@P8|2wEO$dl2
z!@i_v5dYnoTH~=I6<|7nO|n;iV3BN+c6Mx{*9TJlnL@<Wz8AD0G#r-R6T4Y9|KE7E
z>t?+1>r<w06LW-FEJpjEpC6YT_$%$^1WR=r9>L8}@G3h7U@CQDV&c?yXC|t5iwMVu
zVpr2c%34}sa@0OVNDr|5+m~8qzdP&jOxl!Of8tu}PYrsl@AFXK?}jHLGP26J8%Oy1
zpdE5zg9o{|TDY6vO5sf}o0<5f?bbjkcC+~k;84hS=XDs^^Q4~FvmH0zhxBxGh*}d8
z($Wc>ve9JkZ?4XBo0{flHFb0%gHUmePBw;_yvs*7JK*s<ZJuq|9YYEC;E0F_4yQja
zgwJQ49r!6yggu@^az1@RxH{Wyr*xkUI35(j&6SM~>|2%8)t!4e$q%lzng1bRr(=}c
z^R0uoE6O$)9a5O>D+pNxKt&?xVlnF{6@q2y>?f`uhTRVQyZ;I*cH^=$Q_EgH*%?lv
zCZw)m>$<Qr{iVu!lI{9t->-Xbu7T6*Y{zZ)V_4XRddng*KE6sw-BxjlvbuUOurTpV
zt$p`5+Zl)b#&ek8-EG}+#8c?yH5^(uNAgKHjgY~i4cN-&z9<|;7vC?0-v%(6LiinC
zI-h+PyYC@p-Q3<ss+9>H`_80iItSfixA>^IR2^b>e_vgnR`}VTTW0{k(k#^q$X7`2
zI<*G-V7@)ozk3J=tRz<L!1hpF<H>T}ck7w%2O|pXjB}G8Xr&zft|qTkbyKxZ>ib25
z178Ve%4(T!^l}71jRf&M$)z@2YzerCy}Q8mJ|5tMgLxW{^|I%?4i;8cX*V}sINGJ9
zk#Y0zq@H7waif8!Q$1qU^8Yd<N)94(*LM!?c`^5y5^Kq=p~_;Ijz1F$Phos>E`-%z
zz8nJRO%Hfam@4WcXrkQ(Vr;+`#ECy^pQ8^QvP28WrFY9car765v3^l-E>Jt`gn4<k
z>osB=MJ-14O6-Ok97sq>c|8Vu;|T$k=f)?u<}NtwFfpOVGc0YDRKzsw1Mu#CRy!L=
ztoIJ#wVnr~-S^c;tlHIof+hS>FdrEJ_q|wU_Y2Ry{=+m?V@as@ZV3ed0fE0>($R10
z3Pa#~bs&E`c0PQ)2#!mNnMl4~Oakv<X5scxUoty9?ZsM`e}BApyItOZ13tBWyWINR
z`Y|1y;KM2I+!w$t!opZjFSjdJ;2Eel@#)M%4#1KGH7YRy*TIkLy!I6<E-tPF#{+^l
zIRMTqkXg6k{Hl5i0B;O@Wh6XSL~BFY*?X&SCd2>l*We03zt(`q-M{iRD)t5NG2U0&
z&hVb>&Je?x3%mq?4dAZ?aTTJmu`zfyf7u;-e!18IcgdA#z5*}Oe^*Gm5L7hZ3Kl4(
zPgr?X;6h2b#z928CL|=}dGP#qNkUO^F$UlKyIc>P?4n6jrYO~z_;>;=Dq;SC654;v
z5B9BQJSG5}6J#s+>;H%R!QZ}TiZJMr9OA|0uS*y9Ao!a_liR288%I_<u`7&i|31cg
zMiNYm5;W3C0e^P2Kf}wSq9So%fOAz(;o6z#)*#gOB?p&_?{I$Ssv6yl1NlX9Sm?nB
zU#>}FL<Gj4)m6i(O2f)&n@v4vY`!Jg*4CD?ib}vzd#L-xU-&6*F$@>PN)+RA7yR!w
zW#-`JWupg=p>1YgKGoILjgmgTe|~(yX9%R3=*6)-e6c<9Sc)~TUf*WGS<Dv0v&*|M
zf|uKZ@Lz>_3c>lS;LFGmCc21>wM1Yte0+qs7s$Ym*aro-Srt>CqrmBJz3cW!W2u`V
zV9?LyUtlcl>xhX<$FYhf!B`+E0ss!BrA68SqX<+==Us|qZ)#Z^NN@Pj3|0t-VmK*;
z(+ofsP@|(KPIQdn&*^OtN?N(eq+ThJnve_2x9AG)=)Y&>TL`^<eVisecz@Q`-1=^B
zx0`P<XTI6CkK`*<8FgYHBO(YNv>?q@?@NR2s>psuA47r!6LCL#{`~QOYuP;8*1sbg
z61x-TwVCpVvybwIz32OCL?~j{OSn}on<I7W!pHqwb(;n0@q6q<7=}ON887}0`NiDc
zT+xe(QR509Ks3wWh676Bc6GTs*Ca;=<5tZj<PKC_mnCtPmNvb2=Hbr;dm9hNG=0<K
zjZaBk9!H=;ttX9%0k!~dE7fnNs%W}off)3sa6%v|M2>T=lS{j`3tbwoby37|wSHXf
zHM~1$Mg6$95BMKiXgi>df!}Ta^uz(i)i#(Z#ObiuQvRbx6TU0bPPPrs*Y;E9f2;o6
zWMd~B-93)Y;0O$<1Siu|9i_nbondpuNA%{(MNdy}_Uy+G;p3(~_`&uvtJdI~Y2g$u
zYI$}A$4T59Ew$p^VvU<Q*9ijv*6#PjdSQU!lEZi3+uQ3fz%$vh>0$)ZOWjJ0P=EI=
zfFI5)O~5U3i;8-FRDA#bea7=m2RyJ%iMY~Rm;D2ZKYtjg#@S{Hx_2elacsU>e>jl>
z;HdT0>LI`|BuEOcEg7%<yv;42?Qd=@a$YR9WVn5~vYY+xd;uAWq7?czGuZ_alxad`
zMw%+7fS@c9L|sXsmHs|hJwxEgO%;m@vMKJM@UbEmrC`lKcm15*Tc6YMi_z@c;Fa#!
z+|QrUUy1%@7!<qn#v<b``ri{BxS0)IKgu&c=BF(w&hNelgDo81bZj^iJKLQZC=Dm{
zxL68x{B2pd)cdre1Dpz4a^HKn=OmRzi!g;T=Sr(42;h@vgf%L`JwJgB)Hd$3Lk;fx
zcBd(T6%!E?uPjHhi*DY|+VN5OT{{ZjUaZW0>-xsw43D4U=S+W6vL0dd-3MV2F{mQ6
zYAo>A1~V~3+vE0DJMP0QVi>dSD2tGc>%KfKI)AtOb<$x_V4X%th`e^rO{YG26#fL}
zT;{a-JK97MrhmmNEXxj%-V{E2z53+g$7b*VQBeb{oy7ZSN(o=Bv~C;xWQpjVCBN~f
zw@x)nxrndU;K0uT#`f=%Ni!<JZ_uE1(oFw8J*++qWXE1+S?$nqADSq1O19yjS#W|6
zo@Qd{unba5cPFua6xSd>(rDqUZs6oGnL_Sq=lfGtNFZ&Cz>6VFwmL{8G5a?s0cjTu
zj^|8KcRHOUj9WEA|5e>vbP7R)U(5#D8uptaqsjS}w=0^vKqiV5+N(n<E^b>tHoU$5
z3+K@j75cHDEPOB|dc`Vwdtv-ESz4M@ag@7A)}d{yKRl%xrv!=ies!sXBPuY!Dq^p+
z@sL4U$@%`BG~4f10OagU_jz;)JPwZ|lm09ba{C{L57?TokMgdLJsS@^8y#mrreoC+
z!#`!SQvA@ewHQoc8~C&O{+PC&(8Zy_2`4Ct$>JGk@z4wompfBLeLU-v-BgnvFhLAf
zdlT#KZm))CXR*=;Lj103{iJ+-McsCPzjs{elE8Zt&({@89a~*p&3k?dd~6X=#jl<}
zrN<OH%{a8?@<uiUmTJDxK8C~bDOR@k5jwB!?_l<(3q%ldKr&EFKC1;aHN2Vvj84MQ
zk+TU11$J*Nt&>oh*y<J$ZKf&_qNAg=s!h-zV39x|2Y<bL`uZZhO1oDAhSsf}_2C4i
z{QFv3oh=X{>JGce8~m1}$F@mQ31Kv8(VS&QCJOul@(K4zBJAkM35)bOGWhJd>So`Y
zOSm2Q0XCr%qSptlwxUZ(NFdlX{WZ(*SWAWTL?>rwmDjI{&q*pKz1qa2W4n=JyNyp(
z!;6xi2O>jT19B(=3oJpZOQPuO&%n$4l_Q}sUZ)7>8gwxAOTZMPaYZj65io<^geUy#
z8P5GzL^7>#aMHg$N`^Dm{@o$rV=TDz0r))6rf75$CLpa@UnAu(f(ZS6mudQ*<4}u6
z0frB+#vS!HYV8+dz)uBEFQpSC-XK2kbdVfYpR$E9c0aG?rhwh@vRGqeWDF_-r%5<$
z60Upn6vy{tys<9-HARM)w`>DdgF!1IU>F8Bm#1cxt<2kaGWe6a?v(SsSGHXIpO`V`
zea>}-FBXF)Set@aZyq-S>i~CyK(@BG52uV{TLFw7{rOpUGNJ%3hL?pe7y`?5YT=~?
z->gGWq5>TZ%j_h>Ap+;d98*Dvh+aX;<&7`v{gITXiFok>z>ESF1|kH|X9Vpe$DfSJ
zc^;;M3S3O^VwhOp8{fB?8dwM}J=}t#^wX<sQHVdNL`h>s&PepU4`qlRJz557u&q@4
zOX*tzd{d6x(sn~+CRQ7DOA{zRO2g_s7Q(Vtu9vueXAb=Oc(T7h4Y(*Re3S5u%l!0d
zJI~}>q{7i~Y;xYk7@<85>Z|1DYZM??DQsS{Vnc=Ru8*01_|4WizS}OVEq#lKpDAh7
zCkV0rLi(mxuvF|3+@V$6;nK&wq7g!F*8+5|gM)TlBnTjMAp_}rg|4<wc0s~3$xi}_
zMBy2@k-||7PC)Tb)y=bgc!jBO_r9-F-8-A0%LY0%d*5>+nE=+lU)<PB4kVC87r}<c
zL*a&6e3dTc@r!=+={8kNdWH75Kx6nr@Lf?sND4Tw8G!05h0A>KC3Q0>m=!bl@!@4i
z43$U(kQR)VJHm@fOLxw`a>6k?BwpF-)0>omoX6Y?>e}%7Pg>g2?R?I?`{sw&`JQI%
zlV{IlKp3)tICR@(cg)r_geS$Q4pq#jUq}gp2>b74NLbe3{<-lmOeRIhos9xVu0EMC
z6b`H;R)S<k7&c)q;m;LWu&!tffy)Nd)6;dk-)2S{Gc@5fCn-H8Hm5+@Jq7g)@;F{I
z1W3QBC!2oH4{{j|a=LY+4c-HvMxvddj)cc9^C%}jxG^`swEeOlo#(wa8T=Cx^wJ9Q
zz0eya|FJ??)F%es&o$iqpIXemNDxW)#md01+N6)5V|Xcl{@fGgWxPSHd$|`f5Ooqb
z2?||{lAznaXnsqF@p4{r{YzudRR9PSmy|rCRLe$J*N|~_<vBV&7CQCj?O=dK#qpO(
zeLFysR(XiVOD-)DMirTQpFUEHfSUrY+yaWz(Z&QSc2E!bT%K%>Ig@74fZMbtK53;0
ze6p~5yB3`I&p%J_kr6=Y`||wxb1HcLGbT}bLK%3vBBG)6o(8Q_&I?-k4>(oU@AuxG
zkz-gZ>MW_DfyRpvs1zo|`_Bs~?l%(^lJdDX1J4!_Iep+d6@Fx4g8xh+UlNisah`$K
zUybvhLM#}0Z>!=01V{#q(En>HVMM{JZ3Aso_ybiY?>~*+F`>xFutj{F8!Zqm4qOk%
zG`4%b8=~^Y@Bvflj=-lbbEp<6+I=0;>jF5eYmkp?sPeoE_go+UD&tv3rOTVpf=v;o
zDZKldtf+4XDB3EI-(7rJK!VI?ec>2*D)5yQ0WmaHz)-hEe3&SV7XhXe_5bV)B<=ba
zk-UbAKa@V3#}GU`tt&}{FIKV={XZWY!v;xri_X4o`Uj;Zo#k8fR8z{pEP+pNNLWX>
z9|qEh<NYrwvw+Rl_fVs!)FP>JL_eja7dIf@8P0yzpP{_M#Z<~l`+sdi<poZ9EosIJ
zb{P%BAf+LTO7x@$Sn5jed;aB3m|g24vI;SZaS$Q?1E!Ie!$e>DC2Qc<5GlQ9u2I}x
zWLCL9uR(BLMUl%~Op;(_q|;as8Z=lw0+NOX&6EwE4WdmYJSM@b2ot^wn1P}=U1{*E
zgBR6lu;Mj~SFL72#cw8^CCAm#x~h<^zi`8^{U=H45~T~!K5-&@EqQ243CkZ9j8cI?
zXtdr+O2R$9vD=4QET%Jk0XUZ@C4A$|1YDx(s(IYy!RlZ0OQmYi*DO>NVjk*98WH9=
zy@4Bk_z61=ipsnT(cRRsM}6|hwia2&yQ7~>j;o2Ujkd-<!}$u8?}Z^;+7&Saa!}pv
z<u<3~2*X>4Rurut@2UXwZ@I;#_y^}bcc#lEdR9;>g~z&k@AJn>Ab8*oWTkIgpZv>O
zn=Aw59XP24ErU>ajs(g3-NvWT{wp6*YEX5@rgE7ZffV&_E!F&B=gZJCsLH=NX-m(s
z@ahDn5^g@!nJ{$a4JxiKy$6xyzw`kv?4Ks9x%m$x^{)|u@*mW`Bo1)Qu72||s9)xP
z+>O2x17TGM6sjes@BbHD3Q6pv+tRF>Cg{(2QPEM2^Azbpl_{m>KiKgM^(W%CKW;Pa
zo>KFu<guPG*qN$Y8)|UdZL4!!>3^CDfj1En5)&DHnkKj3Pi{xgXliJf4j6_1eyM5#
zXlI*zu)Vltfki~>0Du_<FK7pq!heBAE9y+g9J+)@Y57)ZKfZdg15}uB9qI?r{?~Z3
zheCh+_)*vGAxw4Q-IRCr4wwOc5^w6Eu$XuvbZ7%%H4B$B-N3RZREvT`ndjM_)<0~f
z^HW@tR>o*s=~Y^>@=Q)ZW`avL3pac2T6GSzaB<Dq3rLrsE?=hFKJ|=CPrqhANYb%u
z1u2oRw(OtYV0_M=R+LFsj9j?r&mhPz&piPA=4C#O7CA@Qwmq%Gz02Su&<dE$<H#>@
zm*<^V3G>5cMz|W!{^I}<^iHDcuDW6<BUk&9yEBz-7MeuLPw0~XZ|h#E0u+&w?&)(1
z7$r_8&P!8mIl(7s1PiEv|5Oa-_iaucKPVG|nBFlkR@cz3Eq6v9j;Xye1PP}Mkr3jr
znV&4FLd5#E4iA+0pgSX@txalIxAfp?vdmZ+Saj%yKPU#w)}-Hj<}{`B976xk{l?Ir
zK5Ioc({PW<PbCDM2md*{ZqT)$sPzH84HA9_8bJ8q64|{<2de+Dwl+z4%OQowI_cg`
z;_2FHr<U+`!YbFnv)?FchYma#9|=ieAvXdI7JQYA7m7jGXz8o&Vfn4$a}Dm<C!Fvu
zl#YF)>FV<^R>l$nia<)Y;lciJZZ#v>aEw3;$D!~P<Nsg)B5l+wXrfoQgQoxz0u-gX
zSdvJZ9|e_eM<7l!qTo&!;`#Z#k?9b+(tA8F)_-6Q^)O^KiorEqlkVo=ml%_&W@!gc
zr;^Y}Oh|mQ(OsR|J@O9$1y^N=(7zAbK>Rxvr9<zpjVRs#G&yKgaD?^Sm}Jd6AU0W=
z9V?Y|C*5ymA5Vp@$5lFrkac?l?;btTpi?M9LPrIuuqS4p-y0A<BWEnOV&S%MOeNGz
z`T(z#RlrRQCPLK^hc|L^Ad!Ft6}Pa3IfTXJeb%}Aa>;(NEz9epgrX5mz5iK@yR03}
zGt7)8_+Jvu%O*Dw%Dp!b8c1ro4AvMsy!be+^DM3knkJ=XQ^B$4Wr{<~P)<*6Ifa&%
z;w(a08p$xsV6Yp2dM?EuHAvZpQT$U0DLdf+O=?PsGd$1Y3=^HQm`~EME?-u|Y%Faj
zVYJfa4T{IhF!(3rA)srI_h<aK#gvI9Jc_dCOy2m-9zuBNd7hSq8E{naf|7JYO6{BO
znN4F5jHMFOjek0}y!}4^h&W67>*DE{%3T;C5{QPkqL+MS+pc`9!^r8}W@bne499Ix
zT<yeDv%j_{xw+r&PeyhKT?Ow`4F<Hv1bWeS+fk^>7x`=(&z4SI>B}E}tCM|gqI-p2
zTKZX3HBO;{T(Bu6rO~T%L?Cx*YZwrp>N5%~$v27-*nA)8NAK%QazzV#KPm{5z@<&g
zZw*O-X9DD!1lK7UA(JFmm-{2lD~iyT|BxbpAK$*Vn?oeOSr3&72XL48hC~AQK1$@D
zDUIe>rlmf4IX1d+vvKc#EJ1{ntqWE#Ho#1a3nYn7s&O>bR-mfw@(K=gta2k&9cuF4
z+M%bp(kp59&*8=}`Hm`fV}EVed6x&|PTTl_m%cbXI|9SB0Hd}L>DAiU!W?IiRm3c#
z2Qtu}VnMaa#%pjQ<-)tpGVb9q?kj^#TWgmS<kDR93sX^{ipE1SZlCyU3wxfQvCydI
zVQF0&xdn%vK#K>=MhZfbP9!+P2*B{4bELm$LB#Q-x1TeOU071GZs)kX(XT>@<ZVjb
zGsVN#E@zt}ac-am1VEncY_^4#AJ1^@UpS!1A}J0(PHV(5N;e8wMH3!;d&3)H;_LRm
zSpY63)<O#m<aBKRVdu8ukQDS*^nhIHCq7wkZs<g(&l#<*T{rbG@Z<8gke(rg7brV}
z<eZQ}@lj(SB!#jY?vHR=3_LxnN+mQDK&)*?Dclg}FX#xW&o2J#8C05rPNtdc3X~Pk
zc#Qoj7xsX#9#!AZrWkK%)_7H34aW-@#`1QUtVzpNu7hWh46BH6e2Pt0ADaqE!3UY<
zts5~|p;^>9w-*%6lNwEF)F9GK&1In~DvNLeRWj^%3GGzE9wRsc7PlkPAi^<1QqnQt
z_7w2Qrujx&@JK`;=GpcKB6OTqoa7B@X|q3L@>XwiiWo~;BNl@X&)I$h8aMKkSM8oX
zx>O;c`!RgzO2lI{*Q=}9yxA}s!rb@+9Ci$)YIxYhlro;Boo%Z%{6iav9>%9%f7QC*
zg-VuWfeIprJG?Ps1N&;Ej)0DCyAy>8@O}|x8|_ft%JfzYD{A<PT3(D3%%n`vf+P`0
z32i{ahi+FUXj>Tq7MMH3(-*_YknjO7KTY52yAw%*6i5^RdWaJw*E(s=j$m0m{l$-D
zKv_&$nsa8TjjBS-yoO$dHt^`#wnd`Al7G4guqDG;0A6WdhK0Z*E+*#2>$s$+p;g*M
zpn<|%D*eIKoxW5WZ!JuAE8tIx*<UvaaPX+`s70v)T<hcL)JMaY<6cmH876D`XQ1AF
z!Ma^()KNHml3+@iU|02IkQT03g+wwFP{4dI+oIIB!lM+?hYR>R)JXu;B)Q_bN+{?~
zhxZA-g?&sVe56LuE7-pnj3R-=^VqnHKdHddOidl8&KXGfw_F(czAl2QGVr5>io!A%
zN{Ub;Y?CU?p*`2WQwVV0EwRI_LeGz{qhrR~Q6ogURI61|9b>WB5}}7zJRzI(;0qIT
zXVju}_6{r|*;V+98Ge&MHEn)3157PFug4(_#|MK9cmxt7ZwGSeU$(T*a8N^J{8<yz
z0U&lwI*b~YWTiXtDHiu@^?L9$-e?9^8!ol{v)s3(98bmaU*m5hVqdKFJOEEt>5X|Q
zO{8)P0Ba6NYh(03Ia3L{-aqH<v=O`aG`A$jzQoJmY7nnLlw(bv${Rs@q&@5`sWgJe
z<tC(3OhPCqy!RuasXKv>Qfay>EDS52uJwuF5|IE(v}b_{;{q16h8pBsR|y;4)Z*#o
zZwSv)3CXyn)nzZr@;Z<sAl66^WTg~2%@T?<y6FVtDk%bVE8A_|Y2({U!lG#_JcJnD
zFS^yKx!KSOx`Lbt9mY{Y>Ju!Ii^!h9f3|I)*6(J#hOE7}InKWH-L(SXS5lY{l056S
zvc|*Wg00O^7*a&Bmea#*7b_!>1a&{lELW9gU`e{rT@t%C3cO`$F6w^3v?ZnVM2HPP
z^EZP`pp=7jpCCy_P8O4gnBW7Z-yndh)oBPi>LUg&j(+`)y6)ueM(I4s;OG_1O9|n}
zX+?naXH_a2Nh#7B7|7K8#)6_vm|NmyzEE&zLzhugb8{_tGG;TsmSRFD|L|qpiIetl
zKv@62`q*(tl4}^q)*O97grIBP4tmC%u3PF+q+A8#H$rYZEnvhY_T|UNOM#edEr3*&
z=@H~Bn>r_YT-7!6MnA0zAt0yZ@q|iOWGM`b2TN6^bA4xw94Tj%X^ZQUS2&x~0SHE-
z#F7ibhh!n)9V_TMJyD{Q@959?<p&)|c5NOb_`vZv1YpcF7^o^vZRF#tV-$agTR#K~
z0y?+z`e0`}uGs9BD4@50kJpWXjIPz}E0X20Mk)+C{1+E1v1NLX@?Y#&G+c@slf5>I
zCREz~@K8j)IN6mSVS~Q1^dM9N)A5e!m4W^ot-{#?d5yMJT%aW-02P<;4(`xV_&QFf
z&fWZiLCq?U{iGlv+LUsPuoJosPIt-~cEgGb&2XRhxEPN04IUX8(LDFL+$_XHMgZ?I
zy=5bQ{3<JI`inW})6G|UYA4XiGI>A_WUejWhV%6E(dYoim!>tbRC&_zsm%*(L1Pyg
zVk7viZ03AHr$&3i&~ONy(=dTKh%mtl6cJrRpYakTnFraBA;=6M<W;PsfxAfzWq|r{
zE8Iw-N#{9}*2J)$W#$CDxgAB$hZFdVfSy#$R~U3H;zj4TPusRnyKGB}i{&r1wn&XC
z%01efxXU}ps+8>UMA(&f1OOMUjJ!nu@{kFKoTU~K;zXf6_1eAiemyMP9Arjs_~q5{
z04i@LmRHB*5CI6F8$%(3^NJ_A=NXml__m0pmSi@(K!%qwvHDs3?F+3LQyvdLnB(}6
zoJ?FhuyWY4uHZ{>^`E+{)Hza=9{i`$Lx?blu*%gm1Ry;Q<|_(bO}oy6HdQQk=Fa-w
zXN5bct|gX~Pe>Sht8wIO!{})H_MjdPbTpyLTcl{Mk1IjlrYbplQndpj6%2*Rc`4(a
z5*_ABPvd>b9U;r;85oe!)zzKlt(c6Qot@2#ZC*{%fwyrw;D^x$jAJoP_SrhBAD!F=
z&1`Y}>(p3ThwULCKiHff72%PZ9%I{{7e6X^jW9!WWb|rIqMyo-v3keO+wp*$JJjc(
z?ihOe2WzyxwA~`ia;^p^ce5Zazf}d6V1+Lu+4PjXpp>46XNpz;tG4|;g&ar`Va$x6
zx-YPFkR>-l6xRv|oKqyi!_I(k3hxUcSTc!_8*37~K8}XQz{v)DS_D*zFRAwr-z^@#
zQ~fTz1^Yk|VnFy|yI+>$aU`7@$!8sy5{FP-iU`v(nUUb~x6U^n7T9_hXFS*McJfpB
zHq$iEYx6~?!jk!#(6Pw|$+$_s26#eRI2R7pxcKpJoL}{uRl>&T_V(evfAaFRww!cK
z3!`A%Zcku#EBY1#eGa+Kyu#Qc<K<MT7e3jPu(Nm#6c`O?u`*&qEz%xn#L@iD+<apc
zYxcJ=0LBcub{`mS2Iwhpz?YPU##_(Ld}SgOOky@^27*>lro)Rna=W_LEmM^U<W;=t
z9X6BcIY&mSRJt?l7u>Hxa!n~YbERL6;W_Ky7(!ZZFQs-x_nPXZ>b$0Nr@T+R-U{tG
zVLc%U7M+%X_}61)M2t|CrKw{~pkV#Z!ZK2Fp!-u1kP(1w`g!}!tL+ym4s^f6xMot{
ziw9$KeqK@6=!7d8`hI7qU$Ejgq@xyt)m&2OfB&@)^(52Q=+6mG$t%tEUcZ{EKXOOo
zrB>*>^b`b`1D^|f49y6ev2Uh?aDee6xH`Jvd-d#pg<eTTQ(KMYr~^<ySN^9d!vW|8
zM*FS<_@XBorQlNQDMs|-{<?ss+*4NI+Hsa)aYOr?#0C*UajIE+RcGDHC#lN7T6cl)
zR&0q0;twlDgosNWjA21ZBxqMKGDumgbH>`H3WsOfds(9ke(O#(UE)rCyTnDioPwk*
z%NTWSZHnpWPkgc@Ym-NYnn9I`cRmY+^{%n(*z-qG@_3B`-IPe9^)lj_MVPOw+9hf0
zYdJ*+y7;`eNz9oh8X+Q=jT}|91qCccpS3VG5|!OujDJr|ebm;Qa0I=!WiV4=dwdJ}
zHA`i+^VOcD36DM3g30X~p3-wMC0MKlOAvK1fFS1Lr)B(JJdHQfeDSV*VAsyfqppg8
z9wXF1dKk}HCx(k0r!^+z!#VO$|6AFXmJK9BQy{>6#<P_ySqy&IU0inkh_@9-DNfnv
z!Wv7+$E}t+$des-suC`C-vk(3XesI2lTM<?<Z1eybVY>j`vk4<k5r3J_;SD~^yh;{
z7pE(uYLxSF>b8#f(UIB|2GQSN5G#Do+01YR2=5lAYxBJXf|JC9)oR*lHx^3Zj8VyE
zE%?xHWtN7b6e$M*ay=ixM#;qdzOv#L@A1j2d7MwRMsd{IEDWKG-x>g+qh&VNTMLG<
zn5h$m;X-@h&YIvfM=}<K9#PQ7{*_&%9z>g{fQ;je4KF!H_0d3oxV-(tCefOvpo*e%
zkpJ-rdTr`I!alFvB3O4PJxJ?S`E~B{=MSz#&%*5vS+92*T<Y9;Hn60sGJzldKs*CR
zzt^Zc$}N1jntkARdq{3Gb3}$???>JgqUQdCGgHfBROL^@k#d5Dzwtmf+{l`g6Jlb`
zVol^g*7o3eCaH?n{CDT2_YZf|`s8>g)#`rqvy2=n*>Fr~N3v`1^`%#)P;Fk{!low0
z^@CU5e+#3oC=DC}Ssx~}^*GgmUE{|;XLpv=T&?RzZ%_&@?KFpiS}<c6&2PB|6mY{N
zR?n|w-boGY3flQp;>+v>YR+KJ{k4cPGrAtyN<LV3y~DlQ!g1Y^IaM{2Qp_!t;wW&{
z`XKO$1<&YqzN|OnLq3G=(*qN)2d$|Z!7nO5>72i6Ooy=y51xMB6~O1*Zp!SbZWu7@
zZm`BRFdq;F$!~@{77^kdjRjS9l55l;`SV2LZnmlDY&_3Pz>&@gIW(O`B_%Po-~(~u
zfJkAUoo?qN5DEUxcAnb_=@0|6JPaIX3^F-x)mQW-)uah{guQ}+q9Xh%zxG{v@Tvc}
z=_M$35g$s8*3?DjrTW8J)(afK{MvVWfcUL#y(xeI`hu6}#bYu3==xK*GHRSP26&L1
zC*DVk(l2o6f&Q70$-p0%7#os9tVGM++vu5_snq?=bvf}=)as-+Nnn1o4G?1OP>C95
ze*A+irgxqWFg(>MOdhG`Fv=paD7fVMQd_bh3{VjTAk?goK!@=sUorD4HR}!g5x2E^
z_DxDP#uiu>S(LYlze11rg5mjAU^%WJSDEMGI<^QVM0tGHIgn{V)7N}nbPKePKB|Xx
z@O0QTBonI${^Vp_I18qE)8j{h^IFtNSPCZQe%r{%{C??M!}4<vE(x<nvOs_WqSp@*
z2jm89A9{(l#px=HR8!@y2m87{109avTLSn%JGH03`GqE%u!f5adDq_bk!sXpP-A1z
z-c|pT4*Ry&>nc*nB0=Uwla+a>o|=7P?_m~PBcOar@k^{`R^Vo_mj&5BsxJ$u0Sa!u
zfv9(X{SISAveQ40xj0oxAT}6;Vt-38fd(gDRO7r%q%Fa)LOOMr)+I?%i>kL!<^XMb
z#$k%HR%&#E%0gr+cfE!$57btbbM-KNCvtv5{K@=u1(OL20lz|Wx<dGM^(s?-IgoJe
zrM_8yjoN;4X=(dK4A~KNs8AvMTW!xS7lyIF`z~DqngxaqbnRo*Gf@$93XT>tGZ25}
zoJZrKnmKDdR$m66B!Y0V3+=NBoBp`P+aP5#N}jV@jnn$Lg>6cQrvpp2f}c#EJMH%&
z%D~1Gm6UNoew7ca0s0t5z;i*AI&|W)was5jp=BV>oeUgi4TD?TRABOwYr&?i;Rcj8
zfs{DOs8IJwu^vy@GmV+Ut_end>jRT1gZ908q)k253VanE)eKu+asUhYK$MhHyopqK
zf$<4|&T=W@iTgTCPbKc77{f}S<(P;S(q(fb$do<qtF+9a{d$i@S^)#E+hvPlp?Uyh
z1Em~u<Y^)eiLVt8%(2qY8VtIZ3d(uLwr!(Ev{1R2ico=&WbrfOS09H|09l>$W$V%z
z1d^w`UWkB_|KTrb#!Hfn9u6hf1_Vg!n*0lVYO<1MzWrmq#6oWo34u~`DF3W756@;z
z#QaV$D0gv(H^@J#xzu>x1Xwu0{OqV9yL}l&v1U+)r`sG|;dW)e!)Z~FIII3}7o|l|
zcAvX!xSq;}gB$O3G$ss@#UR_V;UP$_4yF%V%?SJZ!yQ60cYMD2Jv`F&s59@{P0@P(
zD3na~PIO$SlQ$Fz4;xzJ6S!{8A7+dy8amM_cv=Z6PIUK^VkKo9T88%@c6%hN+M{JP
zq%?2wq$Q(&v7+Ir?oSX+^zmcNQvGQ-hRt|F&!oSfGs_>Nf=<y!x^6~%KYBi8bMk4c
zEOS~!fX5AIk6PeSfpJnr34L#RjM3`>(Y~{VqpZp2uuKu4HibXkuKdc~k<g}B<b<}D
z;fuEmrY=zFmIhIoBmvu+r&%oXZ_n3Pu`i#A4B0B{86-zeUlh-Yq_-B7*EH!V#;VmF
z%7A=u%%_ZaBv#hWHZ_1N543u+3<rkq#iR+RQ6X>5ao)nd=#t``4)=$fU~@)yr48VV
zfdwz2cD657H&kv;oy9cO=oPwZhd#6Oa#z(D$MOYoZgFWYPI;*j#J_zN!A7#;<wyHh
zN$LQ+$zy$EF+;Mk8!Wmj1632eVja2P3ccifGI)8HX=?^WPLDM_uN_+1Xo^*Fw^2)u
zu{MK-K}`f;`@5=WhBAR5HOCkPVTikW=2&9nQMrt~6ba(YXIaQZjeSa%u|b<pnO3m$
z!Z=^sZkn&K=LqyFyqa3Baed#fFbTv|#rO(zl6;uK)o9A%T4lgPUv7$}?#|U0w)K6r
zX=kP_rzq90n656~m<<|pcUMHnK07~3d(HgN-JxPrHoW|^e?Qd!vS%V0fDUnEK~L}^
zC?q@g{1#g#Ch<2-B4X)K@Nvdq`-_`+(!^UawvsJ60d4U=!I_gxhQb=a;C8buTl>@W
z1<Qx0d}Ib@0<&fgbfalTy~P>7@YKRU5G-Wa41Y8!G~hFLA&;t#v5|XJO{u+$0p&(c
zhcTg)<>pJlqInPu-bp;6fFXTU=rS~3)l}#u_l!@fjsabF>s~=3t_y^kD1zNRV;RY4
z2ia``+d#C!zMvO9Y3x38mOh=2`n}HAFUqatX3tz7rdO?M^?gf3ofIEezwQfjMu13R
zK9Ev*NMXCyqb}~}cjK~<Jf)WYV~ml&`60e|x+!H^agO(SC~&e$bR6z+muFg};oAa0
ziullu)@KMLCMQ@knM8UM+moG6zyY;>tj^(>4_T%qIj;JT#WTk3pXJb`&F}2eO0YH)
zi92`1wq_AP#1}G*`uYTu_Aaq*UqBWprXAY4DPz&034cyI9rbt7x!8!Fq*Oodor&4!
zwYL|vQlhtWej}%-4M&f{VJ|3yUY)v5SJ>J{{w&UG-()c5v~!}^z9@~G^BDuj){AgF
zv{_ne{E<5KZL!)FI*JtnMB4-D!wM;u{Vp3s-0%+FVf|!VRc`9rPSJE^?3s6fsh22U
zItN)4D|+o5cJdrLq_Kdondt26(^GBDam`%m=%=UEbg6MZG=DXe=t)2sBbn5m30nD<
zEMqC_JjmoK2^t0uM2tM%em7fd$wOFyK+JK{f<u%E&nBrwH@7zj`c+`VV#GkG-xw8q
zV-9sOM-MNS&gqd5F{bNX>J26O)>Lu>>Jg^KKzF*nGjE)iaR+~I$Jco6r$P%~HwWKc
zZVesCoeXPlKmBH)m>tLbb5yJ-RJuZ)%5v*g+{(Tt*Nk0n{^GIk@!yX?5GdPAHu8i*
z%7O^f3YL2Y&#N|%Lx<ls;|9qNNglRvEtbtp<`dpd6}SK$z>pCU2&%j%3rKz^B&t)9
z%{JFUw^dhMtP2g*bKRC$9*2HOB7rDR?<V6H1@m7S_gI^0c*0C6qeU2;VFIOh<)oyP
zKF$#!XQrxIk`*_1AJQ+l===hMV#ik_UA<*Y-aC$dls1*0Ub*4RJ6%Z(tB-xd=Z~ZJ
z^}ajxGajnNR8xKP6|iDy3F)Lc+v)S@sTO9qnd5a(n50FV=^Q8mH=Z|73e07(on@f4
z4FK6rmq-9vX%y@tlG6G%tZQY4^Id;z+-q)(fzzvxqD*vh2TSFCuwnrZvTUO0CV^R$
zp211EF?gm))Wv%3M%@!<#ys~{v+D1Z8<<uQh+*}1HZh*{k<}DWYDf93T@i`(_-~-C
zvc^}rx3tF-yRGN?A^~#?qPS0-mZ|XMC8ZQq@CLNI;#tzw1R*zDW;!BFvu?y7Y}c~2
zOWTF!TPG^AIOLATQOQ5@wsV8h>-D=of!XP=a|>`ta9baDL{s-k16FVLB{cvMcm%WU
zNyf&Yco=@LL)QbQSx}Ays-ofgoq1h*Pv>S{`q*Ic8orLcrE;sR%J-ADp*-6AjfD1+
zmcpvEW%I_@z>Yd)k{YHYE)YZr@y~C#wVa>N%ClG8uEs&lOUWWs;%QJ;yT!)1%!<MP
zXeYZc@+B^l53iUYWb8c_#W8X#mHq^N9KFMD%AS#voN#tQXwC5JpaP4nt@@+0<OJ{U
z?kJ#jFy-mZH&NkBcB-Kx*l4z;ohGUGM@}qPc#r~ei5#2Nq9teNqU%$A>MQ2jD%*oE
zhT~mJvf3~o+I*;7OHsl0%io_UShsRT*%)Pv(A$jid=kc70&yFTiJxtP__OFAVZIud
zYYL%5U9uMp4A~+?6{&9@Ub9O0bJqDQP=f!r%|MtM7frfEDq&L?wLZl~fv$*Lk^2V?
zQX`u$*A_qB`IS+DRgafp?VABamU}1PW3vQWt8um-6w@s=B;JMZYu@5`@iif9gd6<%
zqciNc;Q;_T>_fBQrpDE)bHC`xCNEq+ty9U7k2x=jR!s_W_ANo7OyGl+B-DUQq$I`v
ztI_qN9NAOlG%>t&?97+}jhU5R-l!S7;qdzQrw9<9@t<lp+)C~Hn%B6uS0(17WI4(j
zq^QkzLflFL&7Ud2XS=ld$J~M$W+_DhHs>~4!TtnkDP@dJt_T#`@=_^`D`ow!D71Hu
zV;a7gG8j4Ej3^<X7ii&~+8+k{l=923^sjUIQ`D8qm@a^^B6JtVi4)UU6K5G=l#DEI
z@G`=rgYabK%pt?oF@ti`gqCS$W}GB$6FVq5s0L7Vy{ZN`F<u7rcV#70h;X$=n6&%m
zt63$Ro4(gJ4qSc$%&5voHf?4XMj0f61k(FAwOcE#{8|jSQ1x}itku?i>1s>2jEeB3
zO_-&7+O#FexC6G-%Y7dLmk8Hjm3PPxan16qVyY9@+drT3Y0V_HLkR+OTso1N=pZ&l
z5m`lWV}9zwI9})o2U&3+_p8bL*yom|-Te%11I0!uwIWiK;eO=gIg|zixnp!Ty}bCi
z{Dk$9Aj0><Lo*=;#&7*gs|vijT&OKax&v+<Hnh>6#u1*y6&+6I>X*-lT8MM6>}Zh_
zmjbdVr#H;ZKli>!^dx_Eny@=9)>vH659=f$r_EZ)q0M`Ua0kz}?PWJph2RgW=9}eB
zHy-j=$)JoQXZOBh(ag@v3-!bmsG`nPYIgXH$`qmVt-<Nj=jZDA&W!uC(CwT(x-gua
z6YoEdVk9@GN*pi0H%m)2k0NX^lBThwGm=$Bc{X9}bJ_@B`jlMdI}jQW8n~~~)jIG)
zg|77K<Of^Gb&?m^8(uGN%8DiJ-InU;E9mAR+&Et@mbQMReM0c9gIkY#gvfM3Wh$3C
zLTUa_=7szCUVHiPn<>?=NWfrj+hjtp!wWZ1RLUx4#;T~<%T@j*iJ{g;q4C&t2F0_N
zqJyO7pWy14iOiEA)nDGm;n+}$3p-96J0Hr`2GL*V+dUF3I_OHH&@_R#e%w%{ZV+s<
z*LxzQLf8^2iFJor&qSE&#b|O13=)dsXz}$YmI8`{8=fI;B|h6<JY#H*N?)rktIxN4
zXRUF>z5raoT^~4p<rc~q>T*R(1!#p!crd1@Itg+samg)cdX^}Iy4esVsH%34hF9J0
zy_$=G&4P;KP_xwzBmb;cK(@~XUeeBWo{pZm{Sr3MN;*CHr(%{S{z@{KQcLMDv4P33
zGKH1J+^;&*(yU=;SiZ{+fRd&nfvB&oG;W<y=C12!s1RhI%>C_Q)rlo;HTmd}6flDV
zqX?dRoWnVW)xd?SA$oJXGfQhOscNcTL(lbi%A081Q0!7B$XTGBEu>i5;rgH$T9NJ<
zGu4j{CKrXxal9${1BAiqx?kV27ti#NZG?3Ux#pJA4_P8Joea%f8W8?)^;+y_k^ELj
z6i`5~E6V#zwd<_ia1^1#JcChiuR>;l;^t35%Wi;z3#|8b&{08dO7oxOcUGs7VLkXg
zs(2vY|2&Lt^N%CWtzJU2Dk<GWLfFXm(HQ9F<fx#2;3fZy;!K+W?*nYLnm1j^o|*ag
zs}zAMmr3&}?Jv0GP*t{%P}kw=jI!}L!X4PjN5bmj-9R+$IDGVpbfrOrmZdSInoLWw
zXC)X68fHW2TU7IvxOeV-1QhIKp}tas<3SHqw7(-kAQJl5@0gWFD}iwa$vqz(We8|}
zRF*{8Xhjd@!u_@puIe@~er?;=pjCZLXjAkXh1?$(*Jq~8hnh@EkEl+BziglRkAQO}
z?_<+RW)-xtJ!!WS<h})2+do|?YcGG6<Gtx5x9Vc@Y7XWb!%|XHadkk+(3UwMik0=v
z<^<m)l>U+-(<59TeMGtF3*lSN$|j+FXiI1E%tJer>*boXI0o%Z777IH0Q|-2rIr00
zmB}i0VBru6vQPJ7EHcgKY7fT3#zAUcq~oDFDnrJ=$y&C}U(Q}ZK&W{*Ut9;KQf;AM
zX~19R_m5QOG|T9;hEHA9(=oq4*leoEv0@ROS#WA|udJfTjSQh(*llg)MbzdA@N~AZ
zvLUSLG9Yvt+v$uN`MgHFLDqaL<Ts}C9ym9Nt}q20JC!;gD^fYsPqR)Q0K+?OCrmqp
zLw}s!gcf0%Ol*vO>nRQl6}=M4(msAZKYK<iI^RH`h6)vLnh63ADLSFDb|o<Vk<I_{
zBx8z#uw@>9AOr!@Ku->eby`2-=S&G`HkGZkQv(SjIi<<2b<(Kn0>Q$en81)gVkt(#
zu51nYqELwV%X(B#JWBb0vjDHUbZSwfiz5yhjSYja^)w^^9&%IF?jc927IMzo#NeTv
zU^T~>Kz}qJ6ME}6VrK7bzl~)J5+i7-5Y_evblA|2Ya)((AcNUyyEXicK6cvnd8^tr
zlF4ck&4ztu6%bLMQcPWXhNdY~*_O+G5~3Dkrb1A``}lwk;U-WH6nr<Z{W9>k9*p08
zRK~q20G~wwJofaV&cXQ7DcZYJn<$$h1Ma$r!+1Xb1@A9C)guA*;z$e{UH-GaBq@cc
z;xjA~dl)a{`V~3FhjJ{QI!lgnQ^cBKQd?|%7paZ8I^m(RD^)y}FKS@m?W@Ifd*Ota
zR?uMz<%dlhtLCkMw!XID&?2%cQf(1msA5w6`rt0^6f$!9N)RDPd=@xY90fx!G6dc|
zZz{XS`FeXnc=MtMv#?<`qQ3a^)Z4F==GbXb3%VDW{ZCj!@>r&H(|;7o$o-6nS_`8_
z(WVLu45P+L58zo%@0axia@(wWj4#s3QM`DafTh(^Q+*JL1~-<}5vpO~d#-B28;`lW
zc9@d=@_X!x+^t1oG~U-nzl!<>Ag_k~pL}T3EB$@P7{jfx`~|6GdCSOD<0Tj?NE5+f
z(BOSV=D)!_++j)Sgf|V^)DlmCiaD=%=-Ex&fytxcu%0B2;iW=&<L(Fp122gnu1sg^
z!~t|E$Yp}RdT1m9#nZPu&d)?=1-2%OH`%N$YsNFnh!14%(4Trji_k{}OD)2lyl<>T
zcMBU{d<Z!{7aDm#tQG=lMB0+)muAF<Glz3%@19m!&pGpJ`H@IrDsYw>;NkE+zKefL
zKFZ4V&0{s~ND0Vn>o}<)q5_u*__^9Ti4_kRJ#IMa-|+`Jp#hcB3@a<m<?i+MBAxIV
z`#B*0-X0dV$@<pzm6r##Ybh;^LSg}Vy%#^H6c8XtA~#FPfos7?Qa0<J;?%pOF=pnw
zMf2fC65=m4B}&#@_GWK&2f_4y%toRDs<gxy(vM$@mD3Q59#@lAa=VcB#L&x1Y7{$M
zpj6NerNYO3p`KAI-R^Xy2uR1MwB4!L_I0ElGW2*yuid@&8RVbIXe?r0f$=2qA=MWX
zCtsf>DqWfzzP=ju*7-|`!>6%qpWfHMn9aT%G+-w6P0DZIRcIDy*N!!2eCmvWO+n@@
znWMN)d8AH?gw{%<%2i%Z$cn$^d@sLQFy2Cn-uS4C)?gv`*&KrNHBLZ<+{VAaawB=P
zdsZ`k5;t+41^yv}q|dS2F%=F6&rFN#e{m^03kMx8?F}J7u$_?8aVoZ_sFj6C^BIwj
zF0*ThWnW9H3yAX53jU@L(cLZK?0aN=E#iz^oA+XEYWz)P_qgg-P1Xc09f41UgL&;>
z#s{IpIRQK_7n)HH{<sGq9|{ca<T>Y{wBoqCEoh#$bUo~klaV^CSTU0b3aPEVqu=9H
zt$VV^V*Rz^k)c4;XCx+rx6Q*OWR#KPx&v!b?7gxpoqe*|xY%U8G8I##t>}dVA@LIC
zTh{6ir2`LNgW2!S+YovZ4PmUpQ_A;quBfUbBt^7S!DSkrNnP&U-&QsP9x`w6QOO_`
zq{JSckRr8&xN?5=U+B!MX9A-E=%$olR8Ml7(+}ZEaOnndj!h^zXmQ#HWd+Xpjg(EX
zKPkfCT}mnSvlJu~`N=`%!l>qB!N_vX*8XY#Zk|PpqkP|jxp-tZpUbHuc6UKIdbcDy
zI3eUWG<RChXi))-Kr3zI1y76$3VM~cM?d7>i)3GoZ`0#{n^pIj703uDO@rr|ph*=9
zzt?Z?1`!CnOVz`~Ih=nwn?(8?Yi=QkP9*Ihuq`B>^fq?KpP?2-n{n|=qxPo1DL=C*
zn-50-uFvlWxionwORW<dMj-+84$<lN{Lc!ajJH}2ZOgytLpei;lM1*$vScRh>jtO}
zL9S1?3LS=~j~y4j-^Ea$kRm`xh<x_T6^mBvfIjI?M<)l8)Y$}et3iW5{`-7>W$J9-
ztVFfI4;A&JJR6=X3g&&cJ9gZ%3)WHYKZMPBf6!8Au;HfV;bD9^GW;^lQT&eNEQ<k5
z<}*;({-PdbY9OI{&6d{gYqGwX8;_*NtGGMx!;qTWg84;%23jpU&@{+mWZ3syTz3w#
zQidey`KKt(=qFPNCtnx+%8cDAaQbsDfIXch;jn%tl?#K>7Sz!5=05wCAzYcR2#BoA
z@X(1NSmu;+R*hJ~A*rDW6e4bU3sIM2*Bc-aMyY+(NClKmZ;d8Ws5`9{Jwj>4RA8Un
zU@&3J7Lslsw`JtvIR~IWmup1rW?!@kGnDDfkPK+j1U@|;%HTj;@6>+Jzh|$@GDk_I
zG5kfhF){OUPmSEd{{eRXUnBq@XwpzoXqIOa7>ne542F&-4EcHm1&0-$5d9Yjk7cw`
zig3HK4M1<HZY2oyAa-p{6HEd96mQNjIvIG<^}c<mwKvH6U_NGb&6v)RodP~1q%mZO
za!Fhdo4xI+o>rdj#C4W3J_-JX20a<t43czj@>t%EZBjXZni}?hXu8JmING)yHA&+p
zX>8kWY-3|Lw(T})Y`d}T#z|v0=Ek;-Z=Ux!zMuPVcV_mwPTtdDS!f8Q6U|YD%tgWw
z%E}Rdrg-uV&j?Wd_`E!Hi0#%q?yOQpKxT8Yq)0ekp8?uf-nU^qJAJX`xc?Whs-P!0
zAGiJ;4Lg~e=#Gzo{=@U4+EOP#q%R$Vb;35Pg2U|&j${dZxuPTZgT5djMJiuYEh;hD
z6ff2xl;qe!L!rr=L_*!{bmN-)PfS%1GFQWqkw@6Cj9MFTPM$O*jWe2V_p4IU#XDNE
zPio7{u`P{q<X{G&zComr5oQHRnE8{42S|-GuUmq{B_D&50=QfJk4FJJ=dnbb0VNQ^
z2wppZ(ZY*ke3-nyBEZDkvl)GlAIsmSAVwyyWrZFzCiB|M`KDZz7AVX}g~?V8Sh()J
z?P?Znd=}UBkqjE{w+C=`AO>Mf0;5=Qk<f7<(WPl}3CtzgXLi#kc{L_^!R<o!hvg^*
zM67HE?vpu3N7cGl3vcTe9S2z+Y@Pj6rycr2saT1s{c}eM0$kXuot+$ou=Ie)ZKV-;
z(stiMOKm=bczWvqmlRz;v`uCp9}g1af8$~%wfG_Ab+`IdvG`W!Kj{!c*kw0vmf~L?
zq$ffuYcYek*8GN*g#}kb@aL!Qvxdx!eF?g6%YtEzg-~Uk8-rP2QYPnf6fCSSbRqq6
zvy3p81-wjo4S7Thf&y>eM$U9T#xJU5xf`vla_7GadD@GM;o0wJqc^MH+2TV)%~Nc}
zyzE4z;Fuy4eVcxOs}eAtDy|PcQuFi=R+2F$8CUtP=L7vwipQ3TBY)U85iwNsRt4*P
zE?Hi8|5gI=B$-{Ew+J1V{7Eo}{wjTTW7<KLR#Ml0eZ^cW#SDFqgRmUWxxFP;uSwWD
zXmh7jR`Hjb8B|YojEJzxZ@opoAK+}U+tzm6k&wFf7oILpoidkdFT~X|A}aPVW;qsf
z9^yXp#_hT`R$O`4l{aQYUcfxiEVM7q(i)|8yLAJ<gkvXa`Vxj?ZFjuA3+<3A`3Fw#
z<QQ-79x%PIa+u{wt_;;UTE3~jn$~hLY|!A9ahAge=Go<#bF4ycQ7-c}>UCG7V`w@J
zY_@pG)O=3x@!h9cmW@$e#t8={n0Lx^369RC*L)e1V!@YxwygMYeN4r<zIWOf?&HWr
z&WNU)wvKOXTrtZYV<3fx^0TKL`g;t&iTx?1KGkr^PEEQx0b>+f;UvN;0j%bY_2m%!
zuG1M;HO7_<dI|ihid`GWo@<!<!C6U2tvNZpgK!l=L^&X6f&AU$?sE6kWzFf8-gkF-
z3Z0&+Oxq7L!_D2@)?%2uM=x?tzZIg%7POj#;nrF3S!ch0fA?2$u5{6Xqc^OpC<2L&
z#<~ro+P^$3#d?MY{wK7uOI+b2o|J2Pm}iFp)2lc+8j;ZW^r4adkTQy)Ll<`z_uxta
zh>>JfwK16xpn}NvVgNbZzn??N_5w3&@!>T)pp!tMSffXxKgs5GjVPb<M3=sF8Peic
z#sW_^GnV_;YV>TDuQBI<O)k|g9XDEOCMlV<4V$c@U$ZL12A>?^JBtol^g^n<B88@+
zYNe)E*Ub0Wk>N~OhiQzBohF1&i(D{0iwq6m)tc_&Cg~&)&jbo=M`xid?(qa)x^s#v
z?^v9JaJPO5SSde=tmyYx2H{chdY>M~+Fkxs%haqwod($hMm?exPqSl!lEA|r*Y%<7
zww^fYK8lBa!`8;PQ$cYNy~VfLYU&J8xdt~&{qeYE*;HA<CL)_RR%-XkGXot3jry;9
z>};j@#4Ym@8clAP6u7b86D#wqVNv6VV+UM*LWt?S^$E^kc%D~2UZeUytlIpT4XFpj
z4yE9UipLhn=;7-|y$wAqN;DVx9958d_W3bRAt@&lZy<;Iq+UWt7#7B}YlXBynM6cW
zf`QFQy*Md*Xi-W54UVIC3$!C$H4TB(?3nju0T=Rw@kL>qNa>R#%_6-TzB=D&1hQiL
zxxHywB-T!vxQQ;weUtvn)8W)BZjS$$p^3`8JX}z_R;dOHZ~K22|K$9y!$UoNZK8NV
z-O*elG2B=yyngkM>e~~iPi`j@$>M$r`b9L<TvXIp4fHUl{?7sea~XH^T`_Ci_y0Zq
z>{qnT3F1XKMH*fF&LZvh=hyTUv)$@1Bi$0?Mc8h4J>8rfT(xc-8S3L)-S+r!Xmg~n
zx=n7r0k={e5Am_eR%*11^WkBdBjFW)K$oi>g*yA=atZzHF4V5fVCMpvRx~4{TcY|a
zqrc&_e5ZnMe;qh={%BOHS}*Vj-Eiy~C)DhY`X2u6^X9fAZIF+pFC25Iic#c4+78M7
zC%0u|(bea$aU=Dre8*@#T&<$`1H|wPk@Y$mYI%*#T&4erP#9Bu92Nt)SK2MwSyayX
z!Ws5*nctNcI*Q7`z*#hbf(pHX0RFr;4X+PkmjEH$#n_6oVr%!BH=ibFB(5gbZe0+#
zE)_7p*7UHw6RY-V)gom{_VM=my>+E<+f`H*X+F#mSqawd_zH-eT@$LYSa9qrlV)z)
zeQtMd%B;v5lvsSWGUqBjNjQ<q@$kR#bgt5vW5=1LKh8MuXTP9DYqZImcdiO#XB@y4
z#M)XlZkP`bh5`fd%8k(Y`^75H>q@ur)I@-=<*H3-%C;J}cI5?rKvFFKi}w3tUN==9
zmRtDVf#LPd5`2+&IvjP*Yh4wj$(7YPzXvLBQ(T;BS#}Z#JF~<*XGgB2tLVd@hpSiI
zI7E(AMUOIXUpGu_VxbO;*)6M-fis+uz-Vx@Fnqj91EXl1Ktzf<pCm4YLL)-~N|OBU
zl?_}?LKS4ycf>RT%0Ow4%`Je&O*f^jij6sWsQjIijUp6oEZb81WJgrEzc7N&sK$PU
zfi#dM$9z?G&^2BcoalRY3uYcO$%>5HtZQ2?E?-;tOM-XU_;Nd02788A5O0E2fGd-7
zOSWcrclds>U}C46wnRF{_7E-bH&xN!hOkV>ZlM-uVTo_<PB?RXx6Zqi^JESgOSV30
ztZ>cW@H;k2L-thGBi_ZI55K7Xg^EA2%ajPP^yWdITLywxPf1Ef4YIq=FlStJsqpbY
zi4OwLDuTDSLVDLY+9_V9$LWkE&vKw3XHwCkq$S7O2X)(+`9aCZs=m`)S~+j-OVGbz
zcsueOhN_m~K;uxq7r%==M4P}$PRrRESU+g3Xspgz!^&jGG~40&@zY=j*7sMTZQ=ZX
zw#VW_YLV*JI!(0mbXI8MSzW!1i5M$<8%CO8s|GFC;5jw|3YqOy?DK=iZ7Z_iKi-}8
zsQyH^cPFL@JwFB{p{4O{qXtTrv6aI%J3ALip-Vz9M7+ZbGNvRXACHgP@0!q(9-O#3
zdvc$>AdsG6wT7!I+udLPl*-*9vr%*TM-|zo2`VYcy-z3WN>ONU55FRwsz@144M*rS
zrFHVp*SJ@$Z*3F{4pr>-YdCW8kAHaWIz08)F$2-JR%?-s_71%`Gq28a9S)Gg+{#}T
z&NPNe#$~?IJA@7X55MBs8>AtsQjY2szUpm+K8KrxvfW#S^W|Zdu_ySCw~&n<W0r=w
zxG-ixDFsRv^gUPI$sZJ)hw(|dIf(QY+A=lIRGfo;yx3N&!#ac_udwlYRfVB2ojBVA
z6sSsqpL+aEVjI7D==|KWr)A-gd00LgqMeVhb0&}MjWVsRT&*08_>7b<a#4Pccxucb
zcvo-w$x>@liiVr+^caVNm^g1}`+HuTa0103278Co;ruUxY`rK2wCcq#KjWsy-5rSC
z%JMv@GC>WDntw?m<}%&hPaZZD$~5}t$b>lXF&zr;iDa!wn*m(^s+(*^a1Mz+fN39@
znA?yjPZ>IWOsPZrZ_)UOre!!2F<m?2xKtywow&UuM~6~@^_x<GMi0hPD0p5BPN6=T
zz~qW{{pC>V&&g}&8vFI8oHy~wJi)*z*?R8G)mL$7$p+{W5Bzr`%G27$USq~Fw%93x
z0gF53a!nq?s-0V8mGSXz2Exe$o#J6VCg3*CPWh$33=-|#P+x%jsb#|c0-bJQRN_jG
z^4Gh>oZTtnk9N8Bfc&JtM)lJd=jPhUd0hBk3(eicod23S8EWkRLZ7)8Qxv*4C@<pq
zoCsGluP|&%nv*d|gLlZXDkIC1sBNAJmZR+s6yadmJU0?-?+Pl3$H4PKqo*1vw~kq;
zHURc-N75WOY8|47(|K7p)mOo>!z(5B>$MUlT>GUmcZ?mOyj&LTl&ANl*jn7G*LMAy
zou%zE`rpDpNuRt<1_j2WC0okoi?jQ&Q6HW)erW^5O!>^H)dRXKT}qqA-mhNE0n<AJ
zXXBxxO9^0{F~`aa^>r4Y*7)eT&;vX7&?p2}`^eaPZaiK~;pqI&oalrx$ITXZSLjCj
z*5Z{8*mrvgdqomuv|gUIMOua9OD+AnTk1gBF6J<s;swc_wZ<~MJ!HX^j5&K`R&2~M
z?<f4N{u$gZ*JEAZ(_tYy+C+=}wd^Q@7oU(=-=qemcgnKMWk!LV91iq>x#C&m_oiI8
z$VFIVD9q-#Z}FOlUeV*LG~I|vQ<$6K=>_;*`@U7us@B_@nPfWBUUde4jha=Zr{ZF=
z2EcWhgh4H{NtADT`s2a4jWjmRDW6msOIUR~wa&6RjwLdz&lienW!LC7uz|@1Hk};_
zr?!Z_>Wqi`?|)gJrcb(ix^~;e@pwGy;w#9j)T*+Dmz;$II%y~Zg}C0XMXpcjs2zV5
zbUKaX=Xjzp6nYSNd$<Bc_*M<awe`Z{TqFNW%CKwhrvItaF#Vg?hy}DE%sZWwQMp?H
z;P1<_<8{}kA`8wciORIXwNgi)jV4d=pS8U)Bsff~&M6gGXrJQkHPJ-ShC6-@_I<>9
z8t>ykSdf$3i7yu^S8G0Qpa#Twd>o1~S8R(WP@3*?dbyIhmBmba(ev<0Ty832d8$h(
zK5M-U>xy+rJIy4#{$K#Gu@LE_Es?4A0`?K&w3B>(ye^#m_kW8@P|#(=xG>Yw-#67w
zPdHx}=077Bt+Anq3{t|6o00$0<*SZd{z@T3rXkzRXo{r^M<}_M-!+d`Ms1NJmSVQf
z=1HF}8{>&skx6?1jAlDTiEb$s;;0Ei5w;b_MoeO?g5YAyVjX*nkb+<s5uZ@O7$?K;
z)WJIOMX4g_r6Tl#9-eJwm?lWxZRs+MEIIncYRnBKE^v!59_)BIrg>>pmT2*GR<Lu`
zwODj3aT#_PWIF1YM<Be63T)wq-b$8ox}*G0A=MLErK4<I7Ga*W5<Tlb7muk_CqY~*
zWn<`mjM+!BtRa?-Wbm*N>0S}b5>e*VR2A5vx<|(1<~cAudmG{4`Jj?K%m70H7jD?;
zp4^(867$$@3J?iC@h8)cpKmAc=1p)AM%mGDNyxs9Xa`lR4iu(rb!7O5hh4uVJOAv(
z%<g?ZBe>lirDY;|!U-N<izND%oJFp!xYOXx)8zcqx<Jw;(TecLy2=8@-yTZULyRKK
zY<Ho~b9)Wv(Fc%jDS!?6b(`G)xX$YKEGK@DGq?qo{8(@KQkJr`f7Z9Rswh~r<nsy(
zjnTvhfDmw(AYCwg{hVoLP3-b5`b-8b1)g>r9}`FGRtWj#rK-(Bx(xGo_Oy68R@1VP
z)8KZRy%d^*R2hcu+M^rPrB!PGSDSnldgs@4)SI=xG6q7A2<sg$ZV*|Y4Yf~H8q-Mh
zM~=Z06KIaE)u;+~3#+e-ct&wd7HAM96S1?WJx29as`^U=slS)5Zr7aju&BB33ATH~
zsaN@On<SAV>joC4MMd%FE;5wfPBql>`$?<)4cqiY{LK$+9%onAMe;PZCmS9ryNaz-
zB&&prX^#k%QU_#ItwHu`j?b7+E~K4J${VVT<m%KInZz#Ed;F3j!Z0!8MD4qN3v}A`
zS6|i$eQNCkpc9q{02pvw|2`(ssr{jeoBndStJ<*>da*NB^1f#lP$@^MJ7UC@wlWCc
zK3A8mJCgk`kiEt01wy)PnYPI}V?@=^*lHUR_)C06hF~wd)5vX9MDqO1!*@IBT~{23
zadhYfGu{6W*xho;U>jU)0G0W-!?j!fQ#Wan#fzK|m8qI<7A3QZE#u6A5y@S}r2vJM
z9T!R6t%lsNcB*#c;@zQ(S|deTZmPJ6H=3uOHNH5dol&r282X+|<gDrdixT)ZA3oe(
zj<0<t%JK}<DHxQNuZy(?f2choWha1MDv@{IoazM8T;9Umm8&b<%0#+L;A#}K6RURz
z4J!bbrJt+dXPGa0B)RLdQD>_xQR?zf5suBKHaFuZI;QCK+m7{r1WE=v+2vcyPZ*~3
zyug~YAQ$o<TsHrMraoBFvce~>oliS9Dn7``lA+Jh9LHN%dL9nj`6gd%erHoOw%viQ
z^s<5Zzv*FVNEgjX9IjB??gHP?D;`TPF3GitNNE1@ZrFExTtG7BsayDM1?eOgEY_w$
zwKfBmPXbYlS!DpbpGO7_*|p+hQR`_>?-#1jO*%5QCjjE6TMrJoaSx?w^z*`pt&M&I
z$fn~nvN3NsuJn9G*nJP*p13N9GGoySRJt)RJS?cEpDb_$pPgXsBG7l%r$c#(aXZNn
zB7(=A@!R?0vj&WscT$oAeKfzr&6`IL73X~V?(iM%C8(WtK{l#*|A;M2P1ebKO&>n6
z$CzuJSm+i~eSJf)N=%H<vRTW6<Ew;1a#QNurKHqYBg&7q?>(k>2{grrj*kGHF{R^8
zsK0ibXA_4Q9iVa7M0NKd4n8F)S|da8=w^eR03*Vi?vST35HZ{<+Ex<<=8KAs7h|_}
zy+AnNcyoq>A};0NWheTjSaki`LOZ|a_x@TwE@~5DSu!CP5{L2K+PMv58@q$WD!Fc3
z2<bKW+z4yVX)UoI+AY?EX<>G~B}Y)Ssb>lU`E!H+bxgXx-_`neF+)>4$T)I9ERnJW
zR@cg9oI9{s^w23ZDX?67aU>Mkj!=WU$@*MEY{EJj$pyPxfr<BL%?XQ-{j(q6vq3M2
z>F<vcxYPjZJ<&ApCgIdSvhv_LAIAE%LQs?oO=GNR)$qX5pZ5pZ0@*r^{Mqx8zI_Qy
znEk&NBS{TmflF8vCA7pLCOeN&l9e$ep$0epfzT~=4_`0-ZsC%&I$O^!ZnCV_;W3^J
zQgyEZWUO_o9d}|N?xz^nH$Z@6viRXLJ&i5!?#xn&#a5*%&m}rMEz_K_D5IeKcT0zI
z5|1^7&Y~9nyE?aw$dLDZ@H{T?wH!Q&l8?<48FLQQp}DVDdUa{%$Lw~bzYa*YSBV9w
z*puK@d8&PysDEDO)%xIpUN`V}x|q70TXP}MoL4Bys!u7hG0T2+t{Bkl()r#6+LdmS
zA$!`+>#An}h743_`@Io*^mrk**0Al`a+bp6v;*1^S6GxC=#MpwJrgVJFi&%5NVnh(
zbXUR%*$CUn<^I8S(PbU^M?W$D<-3B^gMCOSb<Va>z(v3numWO3E4fd69kf)23H3d<
zBF4^`;nJ;%Vw?QX<7X@V2ALHVElq2%_AAPGQmKe3QVFQ}9x$KgkW|aOsxLICpcrlA
znu^?J{J2mBD?;UYD`-59%?POQ)0bF&Pl9ox_M0;pHAiOS@LZQVTyew&bn`S6O`4xr
zbQ|NRbV;c}@PR*CB%Bt`5bCoc*>beW^un%A-~vQGRvG;k4X)o)1}~%%jtLOXVUwce
zr4X;`z^hX7ETC~F+w{#@miVk*lCakc%{LF>x>x_JxfV}Ar~^-u1a<k>hqIb9`zDSr
z3Gm3fo6n?5DP|#~-YhElI}?7X$XmB<qy<I8?t~oFFRUHAc$~N$ub@2>!|6<%R+}x`
z*C|`84X1&VuE$d64@cAy+n$#!?)20ZZ6SfBro}pPw(0=6<V=2JYzx&rAFaaDQ3`HW
zp|9tE`e%0F!&yib@G>q0(YVY9Q}&aB9^t;}pDp2v*KBF(0x}AO*~We|&ocJffU$dU
zl<c*R4wc1Q<0c*N6-MA-f{n@je#6-5x$@wkTf6A4wM5(fo6jqF5%eYb@06Jq7)f2u
zNjxZ6Q8O%%p)TRKnD|J5H0CU0+2k2|577k#RBe7SXhZ0p#^P_E@5Nu#1%aXEIf5<Y
z2xxcgZ(=WBekS&9EFO4U&}nDKGwr9O@xPf-89XvDsT3Za7E#fQSm7~9vLCP+31&>H
z*ey3oML3DZplB2m<$<DR!hUZtwr})I2PeKx%z9j=D)|PY3i!cMr%ZNG<Qoe%^tT7x
za=%{n5&tX72B7r4>v^u8GFF;R8!GhW;cbIwm3?50A_~mI;ls3riwf-^Ae{o*3&Xy~
zZ34=e<o{;@C@goVc42<#88%H>UfbA_&Pz%zPn<Q!6zddH>HxZCSuaR^`PThgl8aP|
z8C@_Amz4BK>hQQK?V#gxGyT+n<$$rKU_fd>iKx~e@7hQx$fkDS9DH4#TrAeUt;pR3
z3|#}QBJ*fHXDA290Nwaygvf%#0~6KP`-gYw3y6uqK`(Q&v`#L#Z}+eKp)K7B8FM5H
zrO=|f+g__XGRRniT~phSZ+7_QRs0(Z^3n3XINJ8xh=gQ0zYR)uwIh#ip*qi*P(0`t
zl1Fwc*<@XVgyn@()0(j~Z7}{cgtJP<Zivvfi03C219NH$984i6j&LQHfI$@rXBs{j
zsnx(%_eI$JqvHKno>+G{N!YI&Ry<G;AUv)47pdk%WSQ42_1o27(Z6g&O_jKI%RZUe
z9}E1NJ^hbnJil2JuA^keN>WCP{n&ftiQmN}imW*6{5Kq~*Yrcze)IaQ$I|^@_XD=$
zpBKj`4;7|g+;>8Vfx<oK4yv}DYr`ex!1EvhpcV}VNPbvol=*kZQXAsFx^RWrEklG!
zNy)?7mlf%uXf)Sytc;YKa;mC2w^Hj6Q`%s4G?t28@t);6Ci&T8crrORFyB%{{1OFl
zw<bCl>!k=S_@-yF%d3wJR&=UwB+Fu^O&33ZQdqECHC@{X;JLgxy|mi_=qs7FtSXau
z=?Gh}x3=bFWda63_IyyK<jjb4mewQz0Vs#}o%8*C_#g~xdu`4qO(6T5r{%TG`ddNw
z*3t5gT}vuOIB+OW$+U3Ggu<UBZ4do))f>d$hJ13gEm4+@Yz6xR{c=___x1_&B}}?M
zQ<yAxv2lxNIlYV`y|8mJ5C4-f%A@IW0=5(d!o1m|21|xDZaSQX?B+7u25Y|4T<?)-
z>trkD^%mIwYu<*N(u`b^*ASrK*Q8Zu5|@#hL?`g%%%2{8-DZtwLJw`bH{#UMnyfO^
z1o8214cOeo)qJI$i;%AWC1b5zV}1pT=`2P~EsOi+J$H22!#oUdE@@DZH)rhf@$Bs5
ziSCkR^=>{OQ`oun#~F@fd0Sr7!ia1}i`DEi8@w>Bf2?>+yM)W=@A;+ycwThXeka!6
zhdSUrLuj|NrMuyb3&0cFbmo?6;PI^PoBHd>eRes3(;YblKW1{je*X5l>gj#XvlDK$
zUb~|n7FZhUf*9EIqai5$b}xAG{U-f_jUaRLTOX?J<3rQWH1p|xTl09S;@|e|*%~=_
zZS5=(uxnV<>GaZyfYk(;Fmo=Er>g3lIll!*R9kOD7Uw47D)yapvxY$!aP>q)eqS?9
zNDC@yefT^8PvwsGTV|fdi1IiCOHSI~4Tc%gQTfWO)?HS5RUtCKTtF;%Fj#Q0p_J9>
zo)v9S#K5gE=>*gYcTzdPT|sD)O^KZH|Fu;>OipOkVo<eP;My*V<mp|;c_L4J;7_NH
z;h3ED7*2Yqbq6_Rks5CrPTHN|TvK#lvk(b2IykjF<5iAn0uXGNwN6Zmba~Z5o8~XB
zt_d_nx1t=omekaG=R_U(Js|T>+9|wMvIbIp*skF%A&Il21MkS80ok40uS}2?k>UlC
z&U+NOuJrsG+ykdjswLg>tn9JvCfp($Ty%jVBdRRTi5PfPWU>!+%JiVh<V(D0=?Oa6
zf@D4Czzx1vIp6at;(czE&gimTh6A#K*SBfr%WAmy>v}#x`X?G<&ui5m^&-!`hq2Ph
z?7$$nhjUjQt@>X7=jg^^G3fOkScn}P=kx6e?fhWA#N~@~Bfi$Q&1!{RJF(6}hYpUB
zJ4f5VnyhCoZS}UArvX>zzn0{h6RLI#{}~l0LMqamm@wlB`5l&X*Btm}m6IeWoFW80
z;5yxp<s)^jjIKD)ciIHDpC5d#<lQddfK9+y`IQ9;(D9ocR2Hw}$mfF><6%nbxFXo!
zJ%bpq5Ven<EJeFM{47^I08X)erp4>|@SOj4gy+l1eeR6`k$5CAuED&wJ=k@*dwtW7
z9fO`3P<)o(N%`T5=#da(&|`Zu+i<c085eG_v`b=^$ex&CF)Z|Bb-UlF{_60S9k@nz
zpFs4D1nCP|zcgwp7{(Z)*5w4-dl0cN@|5XkjxJ|+KoOGwZqJ~DGx$~otuqvE(DsjR
zFNGz8K<l*Pn6}D?XY;7{{CvRDy`KMy#2{#x+TaE5_4sFtaT6SxIQgD0PGWCA{d3KM
z+GD=`nHI9+O&$Loc~<^Ubg~g&wT<tV@TK*>5H&O>3BJfEO2otHMX&pyB}J(rIgnzO
zo}kAvc{qDrr2Ow^s6F3b>-r8xU0Ein-vB=+Xs)tB=uf6~!OB8#ve{>=#757(i`4bX
z$wKTUFl!@iFP!XfMnPU&dIl7k`Y20smfZ=N^=P;}Z>R*?tpe!ZdYygJ?}gqn+qgbQ
zR<4+l@5UYnXhap6aqK;wawTO8l`U|_&85##AO%Y5Bo`4?x-@J2oEpjL9^&<9Ln{t^
z-@V%w+5__4=+U|rozk!0%bY{3nlOCok_x&;l7M{2)YbUwXT@ro%(s(v)awek^--pI
zmydqDRmsYvLMFLe9=_#uR%wZpu&Sd*VtpQ&wuHUvn+MQ1qmjoQ%dE4%Sta4zC~bsp
zZA4SPIwNZ(T@s+GGyi_23b;R%;=WZKs~;z~=#nT>*;8`xfitMcR3iQu<$F=5A4H!8
zNKVaZ0DB+_ONR*BQ8yO)07Z7p{db<|%B6^4dPDkS+^>Mns#ld;uAOUh>E_AzP>-Yf
zPVDMBvx9JioNjFXbD-|UaxS8+s?Wmhq<0-JoD^F_80`w?9#VEk3JnId7xhjMRE_2z
zB;BP!k*y24>%P~IWlK2rG~715>7#x!nAkA1M7$Kedb}-cStF<plk{Wfapp4{tJ!r3
zgWPzmKWf2<;^O8BYXmHc+nJ53=?}Iq(vnoTe_|v%$FKXLIq<)lPi~AXg-_|R{c~dK
zEY!7f_dvOVYk7_NjeXzNt4Vj^;Qqy9$R4IwxCO7Pb3TJnb@CX0VwN78%T{uJ0mj%m
z-O!Nlj@fRpC0Q>CB~yu7(NAM4WJlYOF2&_b8=>H6Fg#G`=&;_Dgosf;&PrcxGo#XE
z+-&|6HYpUOg_JlfWvKSm3!nzX-p?_NdkwD~qV4APH}^9qC`ax~(aJSgIK20tLzLD;
z8nX%(zvfSUHEm5>_<CrgQehg#OmE#NZ|f=cwa<LGnTB!Y9(sb#N`|UkWFc5ZpB}%7
zzwFW~Bbto_&Cr^9y6=ewT%BvrOsq7y$z@1M4JiG8?VGnazJ+0-kCzXma0S*3&HrAg
z%pP~Q;@s;3MVYw49M!&9LCMJE_nm9bT*=Yq2gAUM%<cK&*;d>0A(2HcQ|_~mnU;I4
z<qJzDgfleBi4B?Tcp||*w<W1$zP*dE;e-Rfa;`GIYS)Q`-Rb%!EI(X;yE{mq`<1)H
zy_-y^T62ZJsA}NUSkh}EV-38lhH$YCb`Y^vquIy0TJ7wx+n91bs9Ebh2-qTDTk~O@
zHWnprZ~40Bx8;xLoT$|KRO2b}aR+#ur9_1;ERCtoq)`*g(IoJ%LqXj?#!fda+}Em!
zKOTf$4mu8+J^r1XXj1-T^ZRX|u&3_c4iNKa_8*Yx&LT#jiDNSFsbMjG{yR*m{p1l$
zJX%r?n=qPPajgmZLuc5#d@Cm6l0K5-N?&cTnie6cUKi;bt6lZHe8f+`?(uW{jTotk
zUuY8KSee8*dV3S?hOsnZIO5QdkwM}Gfa0n}@bdFFf!-)%;;Oh{c6cmcp05)5FiMU}
zjMk5zL31ZCS%hP5SDI?pMv{S#eeV_9iMuDpJdN=x*u~9&!yC8+Lk8BvM~*4KYM*6j
zEWps&H{#BfRLsfkCMG_L!qu&Ot7~nO8PXUkSC^Bi4_8MNE6*ZT9XS%IL21z6AmHMb
zIb0l0&|;OzX~$PVTP@7kdu3~Kj<`{X56~k}-+&%ifRvq9tug(AjD>XEI*m|u>Brh?
zMB5Sup516Z(jR6bM<4^xG0gN2W><J^4TI3x0?h0{%btCB6_C1w4_`<rSNS+)aa>Vx
zb=IYJmZ?df>ylp*G8e)f8<Xmbbj1RUR}GD^OH&2(>AB!tt38v&MdJ&TE-!12_O^+X
z;1FES^KA;>@}6kYURh|m-~;mMxdBSYu@VamYC_fS65&!X11Dg(F6EWW7ONAx6=;|~
z9uDz7^c%+6@H%i*$_tGyQ?GYdL+9ryGomw7;ojakJbEZ26xCsP#O3@!)k1E5?SuOW
z)M9O=LQ`!W+&?Me+7th`EK{ZJa3Z^U>y&>&S=Tr1bYIRT_5(~9Bu`;+_uyQ0(SGN`
z#X*N7&s8aj=$;tuTo4HUp05#xmhKZ#boBx_t#*i_XYQZ8W+zv93A?1JTIAPDG4+5k
z9gl%gZf#Z4lA1-q-cWBKz>1!^nK;x{TA>24sDDS74cxUssGUaw9d1+j>GG2x8{DV%
zLUbDFE`Rr5nDuVP6Mi<=w~m5?AamUIiqPd-C;;+ZY72<tA&UojjqO&$DrjFH|IAF2
zogg1U|M1+l;pcl}P%Yyn0zsGQ#>+)HIXxk1K%%&xZ?;swe9aH$VXm|#a^}IxF>WtG
z?VgaUsfcVH-}~aBiJ(v>Q=c`h&WblK$xak$Be&HOYz>*Jlhi)X+7LWmSud!y28`_l
zjbt;(&>SKU?6KBTQLdcs#HGIt3`zpxZCL|m?%&zcF3S&=qOA3+EDjD&J3kkA+vAHv
z8$<#kf!ct4YCCqf^^vaf4I={=F`!EInl2_|0CH92^xa};@f<}~>~y;%XA^iaRtj9(
zVK#-&<%_y4H&w@StzSI#&feP31B@|1R)2Q?;2)HO**&{-Fwq<$F>HT?I2acJ#j{5V
zeZ&@$9{>Wh7EY%{-wq%^9?7PhbqT(C01s`~Yh=I>p~KUU1VmM4RmK6Q=a5>T=cA+w
zj-vgL=B?=5=qi73bLQ}31cBE(;tALBiL<*-;mvZ2a_^q6xS^^@-MKAh7%%XoR=Yf0
zUy<h=&~lobiMraVeo|`6GJl;*)f#0oqW{I3X6t&pKHhHsrr&ih2fX+Nf&Y(Wx*M&j
z`kpwOFJ35(a#Dt=fvFBS+u~wN&ol{+>i5$1&MSwJV8gFITR{3M7@8k&7OX$wBmV6b
zWnK{Iwdqc=*)z#q*YaT=Q9u6rI*lAVLR})9vgG!50jamZso!8D2B=c@bnfFD-5*m6
zgeGkC8+YEEN~8jj^Is@<F0B#Lc{{IE%2P+woNxb#+8x}f@Aa$eIl&MT_D9H2B{X<4
zFzG`PI=hpRplNVh=`aOX<uNpaRO7dG2q1l}TS(WF4jn@Pb6+}TMjX+JkC5=h7M9C`
zTOSrW-Z+9BadiJty^F~7JC(BvakvI!NjkFyt{OX*FfGaK1Qj5DH%JMIw5sIA6Z{9;
zUW!{z2C9LSoTf_0^g>EZMymwx7fmU5tD0#I86%gw%_B8FXuJaL3R7CM{YxG2YmIII
zcI5Qm<;&r==w<zb^F3%K|MKWOr6dr6AzvSUsDVN(eSZz1Cp~A`BpCa`BC{kF(gQnL
zRBeL=yyTO{{4q1ci>A}$;@TRx_Rj^aLg2?-C8%cq_MZ|m$j{||g`_g(WOnJja`bOZ
znM9oVNhxN57ECE3l>leQ>Q~!L|III-^h3&)iESs*BpZUj+;4)59oJwx_9Tg*&#3`Y
ztFZIi@RcTkexiwm4D}ZaLZa3Se?FdSE>Wcf#|CzC%&+ILhx=E=Ak4JM@bP=Wt=L9Y
zI)$J5T5*^}Hj(usf;Y$4So4dJex@Lnlfa*z6BzTe2F(){f^;r|Qq9th-62an&<rmj
zfahPq%|%5!G0gvbq?@9uoYel1w{%@{8UE`G`>?Yr>gDYxRM7aiYR!HUM3c+M_}me5
zd^kN!+J3`gn3xozHc8!;u~v0z@>P`tW!!>^Q@qvZR?sOHl=jQ9*W4bGF{ci2hqWiA
z%GB%A7I#}MS;|c3iwX-s4U(-meOu=?FK-=u23p^+q2F2pNw!bgu<DvH&X4o5ROi&i
ztXZFL!>62|>x9fqZv|m`l(i%ij$s7+cJVkP>Z8=Hi81Klp&V8p8kUf$+|v}=CZq!x
zu~1<P>{LY&lAQ8`b*}1XnmC~vl<CpIBsYa$gU1O9hJ+MM_nipN1>N;Oe1PFA!0{lw
zflLy%tEvh^KnUc>1T2}GJ(H|yQ38iwhaXlZSAkfc>HX(I^W?oUV!Ows71%4#?k~yy
zWtiFsSOH8q@?IS=Pu<VZkr3t;^~H^f>eJsjmRVO{Pr~x_D5reg)881Io56%)fOi2T
z7N*9tNi-Q7<0q<iVeZhW8W2U$;LBrAn%a^djJ6^3Z>DN_?){hqU*fr@UXTUgtYj4i
z>%Y|J<3frM2bH^p4sOd%zwdWOtQ_~x?bxM|Xs5KEygpp%2>3omUu<0fVegKQrVDEy
zKY-&~Nh7a6JZ>B%mYQQAKz_qq8&XV;%86EL1YgK|{}dxFiWJnZ`0jGgl4)T2kGm%~
z;K%Yfde=V@(DY3zXR#A&d*IZuk;8ZDdC6pH!NKsNhPAFYtou*y$H%c6rrYF|92DM2
zu-k}If65M6Af@0QVg;Iw;@9Ic<f$x{F_ihPSN-0`i<p@=wq->SlaP28V&_gCZdtlz
zsHEwy=gAZ!i?bX7=tf(3;8>0z>Ty6+pKvV*b4qyoCy+wUsvbx%j>_?Q|5W2GSXACs
zlbF0U`H^pV3clNK9y(J#sqmx?pD-P$preMsnnlzrUBDfqUBqvg7puznPq*19Al`vi
z@{>i4nX64h21I6WYPBF0j?3wsC-(RI=!@yaT8-g)dEYBu0Biz|hdGmPVo^9;EuZM4
zkNuxrA8G`$?ts|<*v-649$q=PJ<(s)z1;uWUl@~|<<z*8nXPkxuW!jkH;x(*didCh
zGyj}lx?l%bDkvVr?;)aCwwxZK<kTp6dM8o^1$O=RT{^ilrS_#+5U#=53}cnM`M2fP
z{UHBOq}}oi4brYxyptiJv_N(baO9s)g|33?q;QnvZ<Y_bCvya^jRYU(N}ejHg<6fE
zjjdkcDk-SS#@GlY$$jj7+q*PYCs6MBI>s>*GP#&_rsFa&Z7B7q=#xc%5vdn75b#P>
zsjMo0)uE-lJ%2?UOBHh3D}lUOHf#3s{sQwej$OFbV420W0~Cc}=p8fJrCZ9AeAB!!
z+xMewT`_dZt`sz40aS49{_wePov#(F+%=oiGPlgpW;lzcm{}%qe=|+aD*Z}>oh%SO
zCRE5R@x4hKyn|rGy?pzbLnrQ13I9LiO#~)88<_A^=<KKGddYVgNK{ZLLiEQYILt*X
zpUI5Xm!oKAQB5mjAl_4=D^`#DZSIs)@c{Q}QJW7!9t#`plu2A<v#U<(Dw6C)UCF6Z
zV@674a*~*wdc+$GnnoMPt?Xy(3zhS6aF3l+-O2ODfU)jbAkVnF>>Wapx0gOiafpgr
zp6;k58d~jF@;`6wEZY1y3Mha`3TQ?JQ+^5Ii8<uWmkg3Y@o&B0EIppyrxGD)V{KB+
zzfU+qI6l83>#mwj?@bV?0mUc>e5WI9^b~N`qg#i=_`1?-n5&~mGB)7`UW@+jMkC?e
zgz11kg%Fn(pusG8V#75m*_+FJyE2$Jr)-a}!Cjf*oS&xKPX<5t33|&HNd*lsYhCwX
zen?^7ZI^?>j9FG#%CWgp9pG1CWqM8wB0Cey%b_4C`{_v*hdGXukg=?PioK3q(@ZfL
zelLWCEr^bsjdd|Tit?9dZ5J*uSaXyO4FPqIOIye5o0IRQOR+%p%__xN&$+cWhlu6*
zQOMG+dt9k><<Ki@Cr&F@6k5$i0*ONagprLYAmB=v8`vHFc<%rQTC%YFyfXp5+ePac
zA;Ec@BrmazS;R6Wns}$l1u`Lw>CANGF1m>~&cStm&@d=5e3t<T9=IEJvq=<EqMEQ5
zchSlL;;8XJ;ro*r9q&25Pm&Vh3%b;P%+4iBTxr#B`W0z@iF}+qg9*B2bo?GuCvv>1
zB*e@Y2DzuNag=Xv)>^~WEV+VMG~k=Qnm0LE)K~A-v(suKygRaJ9U@)Y>R)Ra{H~~-
z;gu97*f0r^L%>Z1Rs-gVnR7x^S##IwEq>_wppS>1R4Dlmnr2-i;2l24NT78J3m6R-
zdL|1=zlt_%yUTfFcL{1Dye2YLi>`>}W1|bkfZ^Yn8)X&hlG}-CNL0-pNL!8%@p1}4
zUK=ovfE*AoHt2j|0o!r);e)P8Hbc+Y?IwO|yu7*OO-=3!xf852Em)dhh)?=eFcJ(_
zlA1zx#YeL1P5P%e*`#w|3yT4J_tA?y{8^dn%K+zYe{CDB!^GP1d>x^(Goql&woQ+9
z*QZ7ghvO!Gs-}~QW$d>YMZo95Aovq~vi@iUi~0@=n;3o6@KKp)>+{WHlWiHNYF<WW
zv=AlwUvqjT4vflF=byR<omVVsj3*cji>^`bN&i@v_%Di8df(Z2$1>rnDS4fU=RwBV
z`sNNL==^*D_M|WAR_I7GB^8KOMLWDi-LCU;#FGacn6H#q#tFt%{eb26rM6AmmV-7U
zE9uxQ#P`ki;l3^*b{mOE(pqgHzBTnPh)X`FM3?xBpYIV8a#^oJXJ~8Q0m)5z5tnWv
z!&tv@0R?5ah(}oe^=BAJz}3u5lb&=W<!e4?-Jgk6%jA4(a&Ym8XDE?riW*2Z4ZrHY
zP2o}(LgCm04o<Y@D%bP8?*n31Fwm2V`}K0W^<$}xvU@h4zzbXTJR|DPKp-m1KqA_L
z?HeubnU=(t_GQUbb|8m+J4!2@>+mo&TbUY$z3q)mX>kBCoU2#RjN0X%4JATO-Izs^
zI3Fa1iT3H|#dy|cFDO}w{{tBaE}BQCdQ*;B`hr?8UL6%mUXo<_o4CX|9?uNi%QxDa
zN-<5;B3PmPTI&5iYREXtIZn)A@84!9&y4V%`N6(Nb^)O@VGKU*PMh-qvKQ4keb{_3
z@c1zU^VQ#i)`8<E&`qe{fA(MB6#lgco213sOb{IHK+UXxYgq%bDMuLJ>aSD)>8OeH
z33Aj6ulA*O|KKXEvL6O<OcgaTI+o3OoBj3t;X|CzWUV|)>y5XBj(R4UVNWuc$Rlj&
zugYM|#?Ne^9<(ooeZJ?A^Mtg)^CCc$<|$hC&EQ7p>6O|4S5Th7xH%RpzC56w@^?~M
zHfxL|gh&57RhIshS@V(YcDJKwNX)5iY>$}oH_%$Hq5OCtt_=l<*KBsbZ3?^&><~NH
z{7H6}%ksT1z_g~!GXI9SyvJzP_{t%YK#5gP&nHv7cU+Y%`KgxejwRo+&|)Z4Yc%)^
z+M%1rQK5RJ&YGjKE~!A!EO!`5nCtzTbZ5D|#R`H>GKP-JA01he#yEWJ0u#Htf@Xt-
zQCjOg4h<>k7He$w(FaAu%Eh@~8Qy>0<`<m4YT^l0qs5D6Nt0HG#x%Vve8UBVD9DOL
z9m;41&O79!i7zv%{vs+2k8?jp9Z8F)S_~rv4TW0Tla1XnM}z~0(4e5oiYu`f3%-~o
z1}Hon?>9)TD#M~q`<-3T1Z;fJpUhYgn&Y#L4v#P7a!d#vvf-pN?rDukXiZAMFRkIk
zIl7(BmB?c@L;}I5X=U}#@ypcl8E*GpZ9x2^Fcba1k7ntiR5^3lyR%#B;qR@(G;hC&
zI$7Wz_iWw?4L|mjoJ}rB-~1ZCwTLQ~zMmS)ck@`r74bbK%}kF~EvxBHD0b3Udi5<j
zZvhg4EUP4dSSn%J&B{c0h{+ERduZ@nuc3MzoRk-B2?uC2EF9$qr0E&%Arcz-zK~%M
z8)d4rhIPGlJzS537w||xLZfkd)jq8CTVN-R#}*(mE8ER^TJDk}uzB(=VyolDOy(3a
zasC;!6}Q*8!bcL%ME|61o?GBYuwu7rzdyL%Hw+&&k1sE=66rGmXY+6iWKK!LB%NjU
z^yaxNKFS&*b=&Ui56mec82BELAgD}>Rc~I`A2ueOo$)fsf0bf!Ye>OI3J6nq|4u%V
zVp)AfB6?O*i%<bgyU0@%<6a}v$b1}J{9VgWDb--hT=}wnjC7k@fKfA{0LaRdE)R~j
zSC|_LA*`OES|d`ez2K1#q`RYR5@3rI=}g*|XQ9D-X>WK*;6+}!v8JS!y<@vFcJ@o1
zG&L;ouPcMmqivBF<W+1L5!Y)pK5I>{vrZ-|kOow!?h}@Gtdq1>m6h5{+az@SVm0U`
z0)1c?jj=T;jISt9PXLz*CtB%dA6zqJU*8mup)^jg6Cm^|vi`6%gcGUs1yF3vJI7=C
zx}*DMV@bjQq~GOxZ0KovjB9_?j#6ORbF?Z#+a5Mo$$ofZ!HQ3nnqHzKyLr)tqI8Le
zMCG&E7$WKoE5h=x>5cRFW087CsRRoi$Z+)x{WoWlZpI$lYx?#!WQIu^()8_5O26tx
z<Y_>8wwrJ|I;!(d8}#t5bpk3o9%FAOcEQe4l!hMzNzacG>_ki3_O1+FqOtjW;1J1>
z_0Kcao%)jyL2w`BZ$1H#DjeGvT`}2<Z`c%Ny|>FXv$kq1VX)PAFt|0`UV?0c=P~oX
z=ptyYQ2iEn-#K-7)?C6QWOhA0f4vM5?}#aVK%*)M{v0zc`Q{!Wf6I?(a?&?`YL3tS
zU{Fz!OO)Xok?fnLSfHYL@4DRCUqK08=c)>MMt*`jY+2H>8fB>PzaZU!Snc#1wR8+Z
zP8<-Ub3Ur&L{4C+OG2EzIGT2W1NuXA8x8p0D74*(H3{#ZWDfp63vixn9Q9<3j=Qws
zLwIK$BkCQZ>3-H{;dE7nC}jD<^wmX<pz~2cI9-%d23h{|-g`{eqZBZ6>ltoZ4=#4!
zVgs^J1>R8Yz(U$7hBIkaZ=AB|8-)IG7brr&t61gBGS}>gf+-Xjgp{hX^7Y_NB7Wxs
zlw$vIgWcn5HFMpfOJi;du4O<bPs6x*J=fBpDl{VhQ%f#d93F00TE@Ws@n0a*%`he&
z`5DLz?zv1dl0o1GIUBe(ahP+^STDlF`$j3Y8`jC_=((oZs;RRT(aHgwiB%gsSFTN^
zMs|w~IJvf-S}hpaI;}k|>R(GX_KWnw>Ur*i<Ps+F8k?p7wqN29_g3T04d)O*%=A$-
z6xICK&dHgkG4&v2^1DW+oOuOYszaFC?rGbUFB`913w;>a(j15?tI5f@gD*F(U{TT2
zRWRj&Y`J+Ral!IWaM6+(9DzU3H#EiAk<WgjI~vw5p7Qq-yS8F#5RaWNlRkb65;u+4
z;T-__ZE;ql8r-UimOGW2bJNv=t&`hJsE*H}ORTqirkkSLtqYyC&c6W{V`*<=a;w?~
z$O1of?xQ}f=?aZ=zrU3%?;q`j?!-0)6A$etTiJh#y!OBISse-Z@un}^wK#yxU~%M}
z0tDCRT|7DPBeV&5i9zh?F$YP{MpT_WCdWKlKl*2G3NsMmugVqop7}c@qi<$oEJ;P0
zRP-sS)u8y>B$4%o#}i|IQ#YQu%VTlNEwey|5=TzUm-NNTaKtC=x`{h?W)$ei9zJRP
z-t@1GLA&YOC~^*GeVC9;)-%?l&iX#TpfWt(L~AE>IajiZM+2Lvpkh&M0f7sQv60h$
zply|#fB|^1P=Q@F4P_^(8|y?2ud{(bJwgd>RhP-OebIKIO`r7dQyP=fiM#6$1{KB<
zn!VI7;usoIwTwsuCT=a!z|!M|jA+`C6bsr|zVS9NVc&6oKzDo`3D<Czkzf)^Mm9l%
zX>J*gZr816(UEej3hpdg-L!nt&NO!H)Fg!-G{%sZV&1mG#QdDB<b|sDuCIphC54Kp
z=ltB_A=;rtg+2};he}O@>-nZyq`U~`X`FuuOYu6Q>!h{m19mc43dpocHuP%wmNN>=
z1~sQMo9=&^f4Ca>Uhj#9Y)&`exoQkgf|&n%EKak5hJ!bC;CQC}Z<TE~EcADDaUekz
zY2NuuD!DWmk+n0|nv$5$ZXHWk%C>KJ13Rb~YPymzeM#b3Uqo~g0xKxJ<qW7IUEd3R
z?+*AO-Pvmi-MIIJe{jCFBPBHCfi-n=W=yEVMS~HJD-gR0^s2g?_g7B=9X6*;>*2Ab
z;h!?r!a66Ld*n1wJh@wK-8pzCfb=vf5-6-s(icXHUdqxBgT+w1aDNkEE(>l{o&bV3
z{l+(2pU<hk3Pw|K#>tJI*bbvxB)TjKRBU_$-f<m_lAxrdI?6aLa4YGX*Oyjhu)W1D
zUtttHaLQk_hORhue8XgljU(U>7fvbBds9r<auhJTTL#o#i7Iqc`@7gmU73bX6h@;F
zg1R72o1)&+@7vk|m!gnPTVULgd`V9r>BeZXT5q>|?PkdLb@*WQZy!LOqRch%d-I<B
z;rKgG&K(#5HY1#yrVAwJEdCK13$~Ag`-BQ<<vQFB&VW-Iq25}KY40HvMM#>c3`J5h
z)MP)~PDXcIXjp`qcZwG@JN+^0skU|8VFbFOv<+6nE7HD!A#<B6)o->p_eWdWfYP*C
zQPMq;TSFW=GMIgEPTA^=ptOl+%VY)`1wQU&3%$DdbKG%oB_;!zOyC+_+l!@S0s;qs
z?Wg1LqNWHX(2zWX;J!$Zcl{%+uQLMMt4U97Um|T@B9Vki>L;gY`TvYJ2~c8f85cK}
zY(cVvp?osv*>J;2b$|oI0`+hG*5ZzpBYVmS`(r)unr<`}pKX_0uMwVCQG5Up1)$;|
zfB#2>l>ZZCAMSb!1<-Gijh<qlnL5Z?Hjsw3QXXA@ZYUP^Q}oeZ=+Y+1Ir12WC|xt-
zQ%L1#M)9X{tAFrlXTINXR^(c>cX~lzqjlQR2oHX=W2m6P17Q?jXHW}#vwidW+wn)Q
zX226hhx4W+QI6nWV_vD$i7MUgXMTY~3K`UvEE&e{XvjcJ1RhjEBa>y&8YY7FHNE!v
z`JjRRu?;0rE^2m3*t`P^phDISlHKk2K;dG-a(<pvC(fkj{J_^#9T_l1SPdPF(UA1I
zcR0gf5#@G-Ot1S1*NjV9fzsZ87uq@+@6P&X*{q=vKX0TGCJi&8#q+qPQz|CZ)ud^Z
z2ax3L;=V>xnh=ocneLVLP57^*R?m=Fw(EIOK)lr%>t7gM`-gy^>xRL@up_6P`902J
zU+MC6?X&lpV2Yw&jIn#XtuJO~LuO}<C-&mGCnJQ)vM`^seV<--$Xuab&7re-%Ru<A
zad0p~fYUDG^I81_mAwi{YF4gp=lw4im8{=qTBne7u=L61*+iOQwze=T%9E^&r!z(d
zIjgnYwW{e^$%rTExaN_eOby0{os(r*nX_eH<5;NCIzN;3egXM1U{lY;;dQV)i(Mb#
z`a#Wnb9f@mxMTY-{$jfYKWouly=ucP#>5&kK-6Qb+I1nj5OR_*Wcy1KWlg#x%~sXq
zslBE^C*p(-Ub&u?DuErSXPMsr)!$W3ygl=xKabIOy$458%#V~5)0!VRc_y8t&wqgu
zDO@!+5o{R09cZuPg8~4h0^I=3)xiC~VIcMkU}O@6GzNQjxCb<53n$c$kj>^*c%oP2
zx8`fAydYjTlhWV}jeVlNeWbdbONTGYE&}`Wcoe$8%TzK<0vu5zcOKY*Knq8?*Q+~o
zAju^$$}7KID+}?rYbEIfkk1fS>=uBaii{Cz=D`aDb1Z0Wu&E?0^^(16mC66RZvbe$
z>@5*gC{mf1+RQll1Qa)*kLO}NRrAHv!m>k4e&;WgUj3)?fq{b}my$CaT9*3l>#trv
zEEZxGSmcE_DqHBNkLRIo`aDeP|9mc+_&K1LCg?_p{wW?{dX<*s7@hyjmuU(QARN7t
zL+WVomYvQVXlJ)Wx>uWH`mu9Np4Yg;CWyCQXNgU(Z&Sk>=Y&>TwTH(+1)Y$$|8^6-
zO%%|?<;_dhcN@zQEP--a@Hj&~fWV^BfX3BJ*dppWLhTk2aGo(x&CMaHR*~YtgNY}6
zzlUmC{aQFv3Yekc$qC-yK}|`yNxB+ZR@1o$Uw^Pgj4c{NYJHPb3!0e!Ki1wlD2}e{
z1B4JDc!C6n5Zv7zf&~li5^QjX!6Cujf;&TScXxM!ySuvu*yj1(_xZNAYPYs(_m7#H
zo|(RV&gpx8NBiDzD$Y5Oss4wzfR$Y?eUJO4MjhvSqbexFJJU+^xTX<ku=cxD;D_(B
z(HdKr*P22vvDXV9ukRJ~!R6xi4s&Vg&Ge3Xt?0L;l>9avQcN4KU9<&iY;!Pl5v!5Z
zyj@PMy6NvCmEvF-MhPF=j+VhL+S_q==cj1`kfdmFJShl6(NN`f1XAsOqKT516cjIN
zQ)rd6<R=6_BPn^wxo;ePejhd8k&_6$BShwyt6yi>q|s76Z5cH^czR;~BWXC~->s`K
zy6fsf<jCIW26FTqavgI=ZCUX$$&F{>Zip)FAWDAF%s;j2B}V*$h#T>{tm<=Wqf=@!
zLKm<Lv5^#kgQ~hsL}aS+Lm&f3S5I-Z8?Asmc2*`@xZ4#<pp?R%>B5huy7Z}LvAN;%
z`X?gIhsDpFH;{&6IS=3Ew;nKWJ)RQfKbPC*tp6rXe-?*>U)Vb=ZcpM#Yhq|l6V!di
zkB`?CR=xOgp%ZI*Wb<8Z@jJRBTs%dt_`-_ues%R@OK~Z(#C6?N!ujU(?O6->WbM>_
zbt8*@EXrDb{RVcG3_NI0G5SbU5D}d;_sftkpQkuD=McqX3AaVuDKihXik-ZYPKY>V
zq_7sRcG~*8K2=k1K2<llKMcv}RMOkOFjEO_KTMQ+u6%uo3=FEWZ;oX$!sUoXwf~Tz
zE9M5ek5bBJXKzD2XJ)+FeY=2Pn>m3u5tf`38a&c{kSL^cZ}Ikts8zs$T2S4Vl*t}i
z7Eg1AnSd)1R0Dw($VVD0G(!75dJFa^mJ9&N^Zx9RTpy~K!?-wygKO}f+^m1{wKStL
zba?8wENQBH&UeG;I<BkSkRwRNu@+sND39|g%aAG!ek&@xQnu8tJo{zmmzV#aDzA`<
zjW8bg9Ii^U5*=c848c8X7<R{Dp@p+^m*S6=b|$=<;T#v^XX5wf)-F%Z)^3-1hPl-z
z<|~o95JLk4qNIv~sT;P+EN5CYORo@13Ak6{^=KM?`4YxeySdrccUDaTrNhWs)Mn`{
z)&yQayx#R&3DLJ49>Z5)wOB%I)Nh2AE3%db%#G`~$g|2KB!W8hP4D15kgsMWMj3Mx
z&l*VgZof6X5W762xLWj8Il43^X0Tff-<A~8q>UKQZ8%KeALG!U8==;GAmZ2hi04$r
z_TWM2@WqyAI^9Egs3vjF5>rgtVfwM&9iMASL$s+WcAq--%>9wO{2rPH%r7gy2z{|p
zbrzv1Vn9UWytJq*Coyf_%KFIT<<ZOH)2~?NV83&|S!KQRcj?}+RYKyzD5h#G3-TR_
z-CggkJa66b9eP{S9cWl4G`#^>f(;qAJFii^@hyE-(f}dijr<gKcehDQi~|vmZybyk
zk^XG5UZulzq-n7Wr+EhU4nsSua-@l)LTA_=0i24dIl@?>iu1G|Ys4EObkTL0I5Z26
z*bgZox<!3h7^p`x&a*&J5?lyZz<)M~z=|3G6M$P$Kox>ct#8Gt$$ThF71C*>#_X;+
zL^?$8wlGwHN{z|mP>y3oW}F_}R0gerqG&K4ZBjjca<Wx(aeKG<Ab8eRGNOFvYm`#L
zrdE<$SM~E}u$p-9_}>1yKYy80YT{VSu2*1>S&yaFT+e~%>1F51Y{60C@5-Who3`sJ
z$O1d+;g#d1AhAIEDTe@2%`obkZ=tM=tx?ow|5g^gywc=k`<LnP*<}I8rD+U|1hFvw
zv|ur5;2GC?!h4^uYL^q&909ZOXtUfCF5*OA`JOu@b0r@28Bx=Iq~AQLBb!*lA*DeR
zGuvrUh5YOzXl>1%i`g?r0o&pcE>-PCFo^RpGYek4bj?e`IS_yn=`rzSwNn+_-%7w~
z;SG+XVL?icxnxqaV?bznm(I$i+wC)gm2JoGk*Z8$k0tZ6Hmp2Wh_o?AI}WWO5W7c(
zHg|O<CP_Uta%WHnvz|kx$!}TVyorhm=bW_+(gsX12n6!oXi*Q`CSF#Tfhz^;Sa(9s
zZ9yWH1bU<Iu$4WrXI60kzO4>h%YzH5Y~p}fD^}{W&^Hvt)059F`poC0)apmLjzs4~
z7ESoy6oS!E^}~lKys7FJh*M<6*U=xu)7+QRNMT8@4e1U|_pEyJO>J?kFo&C=OGYs@
zAZScIT<Gc}?09A+_fabmzHDlx5tf3ao;1?ZWtqjU#NaabD40bee9h10L0xI=3Q5Kg
z++hSGTv=sgr*8b4$fwis?fvf>`*Bw;k8|2}tcsyucg-j%I50vxWIDgKQ%e;Y!&w6u
z-eB=0T3oQu9sfi32E}b6Q=h<eW9ew#vT%{8K;lg7+vKuOHkA=_TA3^lDXN2IPE*$E
zqI6G0-?<yl0K+|^W?Ha5m87Abv~tr4TT^Jnw|A@hYRamH-e4@eaH3GpEr9{d*e_p4
z@2{K)bIot{e9pG?kDn9vt?v^XFVrRu!qf@5ufhlAlmiyJufZx9d>bB>Rc4H<o`=|M
zXivuhi^18ZOen@@Jv^ZZ56f3hM7gzjBsBektX7ELtp}}!M1dJj=&3jo&XBlQzyYU%
zwUm?b!is4hOij~XY`-|!X0ijg=yo<9M&pHJQ7s0N3p=+(B2`)7O>s70qs*Mfm8>A2
z*l~gslAsd0gpWqyNy+2qQXx{$kR?B8?O$DGPVV&<WAeJCcq&)KuiyS6%#aO{8>+pr
zt@cu#Any)Kjs|l5cCuV(?jj_11F;I2u%HDx5n)EWl4$bZYU^mQrYyZxRjz1=h)RkS
z;)wOfX8bOlA`a=#CUEeYpRxbMGOLnEL8VbI$qk+!f;zvE<o(Su<_Qsi`-V$k+9gPX
z;_Wgc1|vFMfkiOzTm8z!yzvkh*@YCF+6RC(rbAgWb&PWqXNeHYWV!6#?&vFVUN|KJ
zm}U0rX{#>^)nvJ29od``^dtiz(!kD3H`Vb<I2po?ZGneKA1|;{{PB^nnWzh*@`j#A
za~qF~%jWTjB1gMdZ<-^f_1702USOIAIP7Ct%w59|ar!8PvPPsf^}V+`CN~`})t9}z
zX4SE{qsxxXKwX&`O`K1`;G4iul-|Wj(xf*xes3J%@U5Ki_4l$fp%2~&D?40?rc4mX
z@6QX8f)AXch2r?LqG<H)D4M7QLQ8jaxU`ktA;TsfiPOHYgJ5y}LlIzGsyPEzUu^ZQ
zxwx1cO(YxnVGX@uepg6E2gV=_kefQ0xCJb%g}<c~DYhwF_vYl|F=;BF>o%ZaVt#*S
z|JaO`vk>E2#Mhk8jdRArHI5fY@@8F*(4>W0v1|Q{!O_0xH5BG(KQE^}tALYf!K|(L
z*j|KC&pkFeq6_uv#V3`fjJ&+3_Wl@x#|i3kAu`kTeVdNcWBx)8U-5}3{RoIvrmqlH
zKmK-6Ij+Ip#nqYHJ&lNYkk=|yQ}S&^+4Xx~{M&D-iIKvfPqYju^7>Jf;AWxZ<sQb-
z27ea^0@le+=-@)3p!)9M>Z`i*jTKIAK)nS84owi#8WgGw+#p!7(_%L6+y%8p$asV`
zzA)WQxnjD8PBQHV(vXS`LLR)jpLiiT0m%lPI;K3ywUB_ndmEuHMKu9*O`!#&1dW=`
z#MhsA$8`vGIam(BIXp-~6Q-XnrlCb7hAUSZ&hR}*XMWxXZJV2MV26tmA_bT^!2xsZ
z3anBIn)KwWyr(t@=h(@Shg7+Oa$=>WM;0WRJk?XV<A0=Uq8@w1>=dH)<5-f+>xX#g
zzDJG!)ff`l*eVuI`|?>v6`L?zYvN#-ODQ+j?TdvNPD3d&d{}WWj+gyS-Tiu!9;H~`
z=4w^$Syff5|Kw~iy8wLJN~hnz5R;ZtIS++?_()E{1%IPvXGR(xqPRH|l*~XfKTB>_
zN_$8z%THm!I31@gp31$$LNci$VP3~&!cwg43Df>CS!2+pu5Qsv0Q(``%#7uBLpYka
z;z%FMQUR%Fe*eXpyEGn_I*lwt@A0i|J_K)7A1zR8$n5+vF^x!oLD8P^6J!rDWY;6`
zsI$joI{Ul0qA-_IR!h9FN{?AEX#TH_63Yv`jIJ#&U;{3m?pN1-P6g9d;v{7V#Np?8
zC;6rjXO_1mBc}yOfLpr@2e0V{c|$YOpR8g^g85pXTy<A&ou}O2F(SAIq{+I^tJkGr
zJO5?BXqkC-Mg`B<n(W8`ZtaJ8f%%i}_2jSML!zJ1Xv{ZfPMQjsfY*|LV%BSF_h-`1
zF)*|);DDn$>8N_v=6fbdBdbV8UB*d=E*y;RA3_aBj?>P`*)Q450i{uRW@P{pcytx8
zj?dZ!)#pe>Yp92=go!&`9ovO<plkeO=gTq*k?O?VTSp8<O&|?rWd+&SS5YWNRk>!u
zXk})kH==Z#u~k#Sr5_yUt?rt5&EVd7K2>vC_9XFyPBn*C%uSk=>N-r_XO>mPT+pN0
zx$G+ClXy?Lu4iBwmNWIJ4EHu3M?*EL+Vn=VC5`T|hzJXA9|yV~w>N1=HTfX;l?4am
zUVeTYH@%)iN8CF^2C+owY8p6fMRVb_f(pBQ_(=$sq$JZKn#xENUGT&!Q8s;vitRA^
z?ToUi5FZ0W)||s>Xi5kzBk8PgJ*{y^Q}Y2Ipczi_k|E4d3pZ{1@DGi440oFUd{xD*
zsy~bpH7gK%e3%y7<5MG1+hsJ8xD<F^SrHQ7S3}s7VY{gzxXk*dfQh)lx{ve#?1!K=
zxwP852rsBFJcCO|-RSc+)Hm_~{Dp<`YNb>X%7e==kL{K&#Lysk7Jq(OF~TX|Yq2$*
z0-fH@RDqk88NWrmZ*%d*ZHc4v{V3N1U;ch6nx3!3wAx^p?C(ZM$P@AMSf9nI{`j)^
z=&$X$z<5H=D-H}~SV(-eNl)$s^IV|TghZoVzIjmk4NNyz<v!j~s8JaV8HkaL*^@zM
z#Q4fd<&S0)k=TSH-!5;x{V1vT1EPJNZXj5B)lF|)zMa!&c3E6s`Ms6BJxLwoUyAhh
zPu>gnZ<y>EZE73EPejwiOxfISY#t3=T-$Dfw(BGuB_8=E((cIf^r$d5z-^C&j4fB_
zM3I5}lX;gzde1S{)fi)Eo-<fHeY&JvsefK}I&8_n_Jd>x2ZEUpGv2IBPMShax*4NL
z9`#6I34GWPxT*I$Sx-5OI>>OIOS7AlI=p{<Jg?>yEu}1r7v;o2D5wrciPxy&*t`UU
z`So7C?An^P48eGbR)3Txi3KzwbhY^C=wGt)%VMiK&9**jAyQb-a6sZJg&9>^bhfxz
z%JEC)keRzvHs==I-WRQe2}8tYmHf@6Lvjua&4YToPXB0D33!mEI`<}KOsLB+OF6Qi
zb{+;Z4Ck@;cKUqn>flVuSJCiT0c*@-T7M`&=09?ONIAa=%aQv?zB8d(1uF&-LxC;l
zvF{Z2C1p_wc+&3(_nUf8=#=tJ6#AhN6)k)}7M>EyV54&Xkn0N-X8Sr#cn+|ea^tO=
z*@FxlJFQ*q@hAQnF_VeyrC*2JC)a2WtG#{DT!J8h7v@bo4`@AQ@v-s=&ZXSVROP(f
zU*&~W(c#;)4K9NTuGW2*Q*dvDeBe1=lC;MD%B!l_n1~w?(eq#bLZeTz{ryY)Cd;(i
zD8a-3M>-w@Thm<Q^M`mYCD-3A(JlNVWP}xA2mPCKK)|W;NwSgqqx__^h%j;Rz5|s$
zCvVEl#ifIF2QCI$w_$jSkj02w49l(|5u)EO_ES=E8;V~2;J3)q8>|RdN85i=k>(m$
zn?DM4Z0LV-Mb{Z2#z4LcVqYl7r#m`x+vCAW|2<`Ky~MY>8i`XDz+V{rw&bJEi1`jk
zX4l`-W^(;RWY-#c3c~WGu<@WjLRxSij~wvv0i5c_O9&sp9C}Ll`xC|r3zJeD{z^Q<
z;ZICRFtyIt?NSin0sU@rLbH=1H?8jNx6dZs4|lHSVLiSkR?vIqDa14A+_DIk$8ETm
z3zN$9%#7t6RGVO*>+)v2qc~NwTC0IMJ>TiLjilM&T=KGUL68Puvh0`kN7psU5Igj;
zX@$@a;WPcQyhIX3rgxEMYQ&e9!P2Rc4%c`h9l~Fa<(!>4FLwt=v<Wvq8}9*|ZC@A<
z)(L!B{-Gu`@ud;oZ;a3wNCODVEWOorK-*AJVUV{~sCZHU+&U>n?Vosve-`cDL%7pR
z<0B_<i=&58p_;0E_?F?Eoq!e%U<y6yr~pvf&4Xl@@8#BOx|mFq@{^KDGNp-?=363g
zts?|jzLsM%b9k-0mF6-CvhyFsOOlW7JQcwhW;}n1?qQ{~1<_*DdPBey`@@OH1&%cy
zVa}$w86B9*oewUUVxlL%wyUR#@p)8xQH_wX)rg*=m{b2~u2jS7JqrBUbl?iZGqbUj
z@9#>tWqpq-PB8Zu9te>`NnKf%tsW(1GKMbdxfZZU#r8>uBm+XjJ?B;>w&B67YYwp(
z8B*SA<Zx*S5C*@+)8Nvf4+*0OgV%pOSGE3nx<K3+Qx?2BJQJP~yKr`!7t+O3<L@ND
zxwt%?or&55kksNt=_a-wzWlN#J&-i@wlnl7z!w8AoPj0OijCK_K%>vnhSnX+A2Tf=
zL;FmLP8Bru1+IO#S}~Wa>fYtJ%$D5v!_l04`>Dys(y4uDlW!iUVDNHG`vtYy_i<b*
zx#3+P6uta=GHVJ3R~S67zkNfr!gWPrJn~pSZVs6_Tt6<1q?&+}l&0IbUpTvH-XGQW
zK*4Vy>kZR@e7~qZJagsQ$2<7K$`h50ogu3+DS(```)9VAUpvxh&t~<{YA>y5hiMgF
z=1uz$7oo<uAbPip$?Q;S96!8IHUg3YYOBoVdR#X(O`ynnl9kWJ526h=#Pr7NeDn7}
zIzbVxE93Zl;qvLOIOWv=$j+h*nGhDLV)u8I8u+`tspk5y7PtowhpXI?HT_5T<%5)_
zzi;{(%#}w~uoQxZW8E=;Mv}vc_$Rip+!+V_V@lp<PL>gT1{k406rfOXDTOIUf_b%R
zD;xY)<5>k{GHT?$X^mQ|rpV@KEiaiO1-g`LT)epuKDI{2qsQi++_PQ|pN|ijKwmZ^
zC%^SrYH{l32_>y+?Ro^VL=qvYqHzWxz^Z_gyW_fTNy#mY-}x`y_pUwyn#^Q@GRKan
zD#2mW5==eN^BP1Zd}RnX@3Ug>QE1t<*&4eW%XiP!wbA2(ZLJHo^iA^n6+KEGp@8cp
z1&2h={87<<6tvpwE;ZElR?CAZO?J`DX6{vOlpB)2jZ1Sd4dsK1X2+1?Qm7qkv$Bae
z)4a@Q2z0IKcArtajaK|v=O?Q1OT>r%P^)E9Ay!T1ozHaKp2K<I)VLvs{A6b!IO5?1
z5sm$NN}I<ReOnc5_Ee;P92jaqSbSg&;l=qx&w(YI>HiH({=)oZPBi~O2extXD+*!E
z);Bdc^5?f%tE8x8a_2CXeXIv}k#6;ZE|*<tvrp&)VShpoCg1ueloFC2x;6WYth+J9
z>hvfq>WJgJr`)od{>*e>Dan2Nm4Nh?x(l{k0y{A+nkT94#4FBC34JC$a)7EawKpWr
zW%<~0w)b-9unuPZ@~Ce2FFa341dtI0aM9?x7yc}t#-VCgzameJKhpn-c?~I7DA(q{
zy#S#2IyTEJEnJ#-K^p9}!r+8&G0N>K7>qmfI*e5X*1Y!-+B_-FUt~8-oy0G4^~O#X
zM)H>r$DNLTL%!o9m`z9%2}TNFbp-bjd#7@)xcGuHJma^vjJz&WR&BdbIL{7>$|7)E
zo<B>La|KPL>|b0laW}jVB#$_`;{(Qp`eNNG9?HIuZC%{wCgf+Ev#1}|!#U?~ya(sZ
zr7~YusEkrFfKSW9$0xn8drbL^W3fS)tB!7n&SZw7P*6#c77b(%JA5J$+z1oY!;G-C
z-4rJT#UMRN95K&){Kn`OMO~~EAT%sEMNI#goHsMmmL2(~iJ40It1m21gzy!VnEJJi
zrZ!9B5?%4O3pOXP(fJPp?*z3Ajx@lOh+hU5O#|a?9oekAjO@@8q9RLw8{DPrT`@ay
z9Jliq&fYc;ClwdXizM^IR*i^g>T=;suH%}a;qo550bEK=G5iA2amPCyHce8vNkKK_
zKD-3124=_l+=zR;1ojVN3g1-XqzYmyD0U~88w35SHP+dr(t>S<HAin%pAmg79k4Se
zO(zI9>!tAxIUMPfkUoB2X-lS$BW4+h?nrJV{DpHhoZ*4xz4Xj!?R5xxD<WK)QbXW9
z-d95F9sZ{NkBP)FB}YZjT=!?iAK0Z{i9X&@kt;2+?9(aSl)J;55n1!gVxFJ+-=qcR
z4o=TuAO;1GtbNtOt76R+`wdTFJd$TW%lewBnh~hvD;iIdR@#0<o(Zg?p)l=|kkNfW
zp1VggkacDYfdehzI<i^*F0tlkuFx0`tGM)1bJ6I=-B;4ktMUi*AY`~yq<+D{SRQ5;
zGc~--ld>TBtmk$)1a0DKCL1YV2Hz>-UBFLe@T}c)ej^U$hD?Sm)m9{3+=N-Rp=w;<
z8A?+Kk8&5}SIqrt-|qg__)Xuw<-!ko62ZcrW7G8r!U1GQH3;k6_&&YzsVB6O3_4y3
z5s*#Gn;2YZBm{<a5MNNv{r&4bsIsbQ8>(lN&wjJ&GcL(WW~4*lb1P)>6Sh(0+t17V
zPBP=vmf+j7*y&sQa!8DK%FXuX$Al1Rx~zr+&qhrFlk^#hPaMH%Baiv})?2l(oPJs!
z^OKSzO+gPAvwLEF7VxwB;|EEW2pOcx?2<*0AS+(oEk!kfofE`>-Ti|hJ^Tm0Y9RB}
z2_&*ks=Xum5{0?89~9M~o6IFl2%exBx3!4hbh6UJDOuK2M3eL2D6?T&t|&L94TUm>
zS<aeXJ0_)C5+O4UWQsPulp&YJ%Z;?WFEV^18@SrD%u*SDBM4&X`8{o$2}uP4D&|Ck
zf|W}?z{UB_#65lx=7FA@NLESl+_u?ja;K|b%p-A5tyHjl3qC+=jZn4e;OV_qm@pln
zxo4SnPtR+zA~vTIvN+~Z#7Y$0UdW%NK;ie8e*1IE(GTS4K%j9&I0qnODC6(lv)dLM
z7JetPCR@#~QbtMmexS?IQ0Xms_9hqmX+mvPM`(oLp2U&?m%l{LGnU^SmQhW9N2Yoo
zi_Vf(dHd+=M~U*Lx6iJ2J4z#`ETYC9SZ9izev-n#xm2GvMcI<kkh5JRZkUb}c0P|&
zx9Zb<$m3k7pxl~NZ9@nb52P31w6A)yiDiUT7?3|dn>cb{YnJ@5Nw*_Jb)b8#A$X3s
zo8j$!fwiZ1&c$V;OTI_(D!&n4$zKE>EW=KLd(w_#4v*j;Lq~kz;sgndUd^i?nJEyq
z4nuw(kxRig*Wb!|HG|j_G{bVHtUL89*Y_<<9A5gb?K0O%=`>+tDRfYr+QKaLw4M~m
z(s;U}{RJQC@Mu-hH!ml!5&ulhc3!&cg23+NtTGgC3pWPUWat<v{zG`rTEt~m`n9B_
zg&$}4j>M{_#fs-ULU3m<_G))OFFGjZ3e5_{RBNI>bls{|g==bBaQY$rz%yiC?A)aZ
zv_Pb1KQtwLU|eBGX=6<FRN}k+=;+3ZH&VeGEl_}4@O!Vujbp|-@-3z1xhu}_Ih%6H
zIwYWLYade+trc3xd?g!l^TAL)@2WV1pk~=X`nUcLN~%&-cpHIs*U;k%#d3P*@RVXD
zrB2uLNL!xU*sA0}i%-ZX&7W`+WD1L|Ds@Fik~Zy2-q!Pl<0~<qZVhgS1C|3puVhdB
z)1L1kl+rf(Q2q>H?CWXbKlsQ@B(9oMG0PJ*)D-m;e<o{*OJ;(yO{FAL{8#jQK4wUG
z-X99x*_efnb;U=?2losN1#(^tKF4`{tBD?-V?#ub5a>gumCW#$at$bKRV@@imH*=5
zG#CfOX=jfQ16P4C^g!JraG^Z89N8F6N70rc^kP(`Xkuex*udmklLFrnNAyX#2UOWo
z5`&q7l*jEmvv5t8Y|%jF0C$*tdwGcV(oVSR71>1?*Ia!o3-wo6(1uu3^c#v@eN{5=
z$_xEV96xLkNkDfP_g=WOvdT7?n7<?bfrTw3z=3jjvq-75vzT-J(otb$xxes0R$K7N
z+(34HBgUdoaHBbuRjPtKtI)h&vV?@<0vR3AfA9Q;dCfs$xrt!^8m7qQa?SiBb~<sA
zGFSX^3H<AitTk*g>)ED`i4TPcQ0WNR=Y52XYY51SGrB!!JFAkn`i!)+Gw2W?5lcB5
zBuh@qkT!jMos+2IT-t=xvl=<U9vioPm~@aVDd$1Kc<LS~QTO7HWRD$1u-UEgz<#3r
zqy@E8Z}MmVWGDPWMMjpIq+=t~K!1y1dbKtqIbD=0B>vAYsi~`)v%PJw*iX@ft&Fan
zhcv#*ZH<%0Gx!{A1`?9%AWiA>1~v{tA>k}C=8WVgM#Ipt9;RO!{NqXz5|hp>p`~k=
zTS^~NU7qiEH7X2Xxw+T9kLs44@h^8q3?FZfhm+ZF+?K|8!?V{9SrcCGwze*2D^54X
z$I1C>7+-DyMHP?g^}xS+>5u#ia1na*UUwV3sd7mfOh1quunjmh?)zH7R8F!4RiCDQ
z7SCH2FUtkyg?eQ?=AVk*ecybR_WA*5JiPYD*1c&!oPY9ijQ@ly(%8+Mi_0#6Jw9gz
z5y*FbXX;TvPZ(@Wwmt9iPH`oS_etpusoS?Eq~3~8_58RGo8Y;z%KZRGqfnH^j`NIs
znR-t}*!1r1AxM@60Zf4VptV=c1C)TWp4ag6v!dBXa0_A^%x>D0_J)H)6uld^6amjJ
z#y}R26=HN{X{!0$lTokK2ZZORhOmvF_zW{JH_6^3N0|mkIx6DQvoeGq=IQwYL7Uv)
z3Z%ylNt9omo7z**$N@Me#=}O^#~kTP&)6PWpF&wcw&Euu0@fw^O3Tjiak{}4d%2Hl
zyUQO>JiSuq8KBZuy16lS&S#g)>@!O~+mmu6bFLY-pYDwKo<fw^WS9>+_t$^x!FHla
zLSQi7dzrE)ZkQI_M-6Y#CRxd6dSs?JIC|mHuJjs9Z?*wXE;yfbVaA5_#I!GCKQmRz
zbmAq*T5f#7L-n8hO&O>xpvY##?V!%QsaZRfU^m3u)}7B9v)Qk&x}dBpct!+NJ{ZzG
z$e&6Yluz&RgEdiJ=3eBCw1b^^fnsbZ8VzS<sR>eR=1YAKc4W@Po~?0)r(zjBzX1i{
zZW&Z^Nio*{0@5zgsZKL?-YfcH#!!-yX;O+id^g@#?!c?5|JJa19aq+7yiWxdgn{T*
zfx>onD21I}=*5Od2*=N*JqaWOqlJPrY&HPPNJ>os4P-2Vuwc115{PyCS@-5ot{nhF
z&lsoODR$rIT^2v*81|8}=J7cK_im>6FdwP21^>C<iq?BXr3X72D<NPqrZWY`_vjMk
z{ecIT4;Lhy5m)rBN?gKit7rsiETgR6V8=*X-wEBPsyk`r<7Xr{NciwZ(Vpdku(slW
z0<J7xcm%aWD?-g`m?csH36NT(SUZhWjs?OgCuRDyu(?+^<t!mmPq=7;X<9|3kUpuy
zCjp*A(7%gdh`fRWQ;)GU^%7Wir8`efLV*pabC>~m&uQ!2>saXg0$VaFqu8tc4YS$}
z$;*Zjp&)NJ+4IED^$PhWg4=LkGFj%hwQbX>E;iA1M}giVP~*r<CKYT^zm}Q3hJVzt
zyWjbd*i-J0DckVZl+GE2o*l<ycibZziWT;_A^XwIjlWf-B^#dzBt4AIU&@PIiP3C4
z17J0(G?ZNg@ZC?go=)I6te@~fDY7|vzb?KWGb}u<vtRXOK|k=&GWT#HIQp@k(6z+<
zYI(sBA_2-aGLBp0e^*}4zUtvK^^@Q45?ne|_8K#9r|M;zR`(&li};wq5)RZW-~L#l
zo)xR5G#<^XG2w@={R`~_q@NRu2Zn8DH)?t+7nfzTbR`1hq=QK$%}4KZ$mF+TIxk(0
z+ufmLIPj8Iyx*ySa!qe!fg&0dk0^5Ook!TtUA|!r6>}ug<;O~(x?*#~QStpT6cgWZ
zm>E}na<`*!+Rp3GTjZ9SoR4}^Kdp>1dV=qdY<Rnq4{R%XT%-xU-f#o8jp!v!XkUTW
z%NxeeWf_Ug7@=kS_Vk1d4{5eB(yRJgtwKe0Uk+qaGk7x^)5TW=uAre@rTiVi$6ddX
zUihm+0{n8M*i=g2Oi9kX^?M1&9(n%w++odw2Od|(eWqg#DGkGH=ZhP5!<bZ}sK|B0
z0E&Jr+y(D_M};J`q2-VD!nIVA(z&3U*SB%kogtKCA2WtNYvG+vdk<eWbwSTjmj*BK
zoF3+ErH<vo*qpavY3;u&7-ruOmQE(*wx%S<w(lz5{%f_j3rsoN$4(pQ_q?|sHO*(&
zoxrvb(iH5X+RBo=XstTGrsl0&BL4l#>MO(rc%FelE?-~zCR3X^8IM8aD^dS#hp%m4
zEtY%p9ahUNDD_75+lS!e@HI5u?(I%n*zo9AlN$Y&w{2qa5+Zt&JGXFn1BcRc>oz~_
z2sv^R_)>GwWb%6G0u^td#)AVx^jbNo$irBm31**{9MB1}DNi|>X_1+{Va~hLeg?lh
zMh|4Wgp2&-j+dx=g2jr%2YF)o)}^d^|Aj+4)*9jd2y-HO8XMOZly}=AotKx=<G;I!
z2J2F{Av!Z_`8{O*C&`*g=xm@GY#rEV$q%8~X%D6un$e<&qr!49&eCRS1Tc0+J-N)B
z4sb}7qifj2OW0;Qp1PwRwNs}#z>HZqfeYgy)k}AzMrj%OYjt~w;{^AteS>ZVLrA8P
zw#CJApJ|S<h%eY{9@Mc({H1Jqf@l24^n20Pf#E(2Z~wvorXI3rf_EU&C+nff?&Kr7
zLBUY9o3gDAdT$fGMG~GRonaaJ@P`!Ykj^?2y}noI5fvEU*xA@Zpduj@X7BfwY2&ev
zyuWA#t1HzYB(O6pUg54qIH9pJO&#xr`;uON&0W}#QV@}GFku~6`7wWBu>Xs;A?K_)
zzZfe`wGX^ox;6teF;yIoUDg!nWpl2Pdr`%fE-y<t=+WX^)G0+V{RsXoJ>RrWRCwDe
z<T2=2&9gb~xhZ%{aAua9|ME35Un#S!EHa4|bg>c2sHn+f|EKBdk95PsLDBufZ0pDQ
znZ-qE8yiRH>YW{vczUhggXtnEd180j%UytbVA_O?L}SE#ewvEFp8Hmp*2mCxm~%Tl
zS(~wQz`ic>OJb?Nrir%#-^$QA!RKt`3}Z0^WN-h$fhG4v>GZcjU+i5ozf9`mviHHR
zXx94_d}l=h42h@KgfrBdJA?un;}s|nX6o3l?4eHnb&FZRFPqTBPoGY9J1HZDj*%BZ
zft{~7XoJRi=YN))+EM&7G>ElNvGw}HLldeJ*CDWwDZ-tv*K@@SAp@hPpKuYxV@UK4
zu1UjCa{Fg9EM_E4@wL?S(IaogXlH-ejUZ44H@|81nb;sB-<_K{s6%ay6rV^v>o_*5
zLDA8c`$exbuQ3#<wk#`zqL?##zKXpR(G!msG4B4@R~+NX;s)n_|J(!?52@<i8^NiD
zbJQes3y*LHgK}gy@3}$t(bFNMu740LVu1oYULoNd9$zdVQ1Dj6Dn;0hiE)Sag(<0?
zDeVKwXNmO?n76v}*nvaOVPliQ^YJ1c3k2xeO5>(%OBFldBWvnBvH~9<fBMk;1kCR9
zk@;CMlQW>xN5kC|g_z^S<i~tFoTYN40@-2KkQaY7!S+i;VSwCS)YiU5uG{Hqe^Mm@
zgd2#d6KWqufSeuB8c-t3doOiQfw}e=vG?hhxWy;TD;Sy(@jyJpOjE}_WK|#_LB#up
zDMYeH+`fx|JyYGgTd^E9+}-_Gk}w>@?si70>X!T)X2|H>#}IMpguFm_&X=S}ts4p5
zyxF~mPhe>25XQ{O<&m=4XMsN#llknneXe!U$C8w5x0w?Mk@rlVbT-6)eb5RyHg4SO
z-Tt6|G|CVLGP$P@O%n(7n4^t$!}@8AYny<(K2Vz#c!J~WABES6eFpSowaI3Y`eu5s
z<*(JPF7~~~IN4=sqfvuwhIb25y7hir#kL%B;6;IjL(W)*dY9p~`+Ze_Se|6(Z+yFp
z(VvT65bMT7g-|M+T!{%)EF>>$pb;^4m)(}00;k?%5l(NB8$lG<t{&m;@}1!$tzV?K
z*u`NPh7t$aBxp{0OWwk)8<P6K=^bca&^b41-9{yYNh(HdnR0>DW+10vI4i{un~~O&
zdM7@CU?4iaC_%$GO%AaTaU2JFjd-Wsp954uVZ+MF1~JTw+TC-kwz+A-K%wD{7bmV*
z0%T%u?-teeiMB3_3b8HuoZN$+cYsRsBKUw;*rgYIoyFeleJ5J@xHv(B_tQK`+KPi>
zUwI>jBAV5hJ0yGk*q3;XsgH%ykjJ7(U;_Z9o#G>I4+m$ix$HucW&zH+lExHW8)79L
za-gkeG;R0g6U{U)CgQnV4<1n7fB8cH5MXf`HvQZHD#tET9s5C#X5b{dUI#x@<v55X
z6Cs;x(1($q^~oRPahW)@0tgrb`9X3p^+jG@Jhi{$=Brop1JS3{b@TB9(rbJJOn-QW
zCvHOaLj^WXmJo)QJg-2`HHcGYn$F1MYo<KWQQY9l1W==MH2NVeb9RT7C7XzO^qOo)
za!#V3q3$DC#cGXHD}$8lO7e0qX7xR%c>HwvgRjlRtX)(yF74-z;Mfm}q7wAcJUhVi
zilj98MH73$(&;jaJ08xz_`7aSObGB$cCk|B-PY$YOi%(>a(}zglY_h>W$6Q5Oot|y
zSWXY4-EDy4|7*&mtm<dHP?YcU`~Bd1nQF?(`TFp_Cie+y#1}J8hze$*lU?)N3eNQM
zP-eS`JKyE8oCpaqyxXE{mQ9X(yhahJ7R=+*y3|h@NK(~ij7wROKZV|1GZS^MT<Dne
z;<s^?(9Rqj-X^%tTTO8>-|TjMDnB68$$!O}VOrrJC4zW6BQ~3$0F}2QnX)P!-4now
z*-40OA3QDd8F;u|%G8Mh&doenmTo<%6cyE3`8>AD+QdK+THew-Jr2ua!s7bX>tl~i
z2-7$`>N*%XGM^K(&_mi5ogYD}0$rBOzkIxR_f~I7y14l&kiQUASh(c2<GoSMbvM!&
zcy<ThVD*e*X>pebct@(f{;o-^6?-FB)yH`TyRl>@!=zPrgAlLVaxiMi`T!wmN$jyW
ztvIk$t~}b0EQY|LuVq^o8%Sn_+G2)1gVAGB5TN=z8;Mx%y)ZpH>#*XwOL#f&x#Wu>
zD=VAoc{5+E-4Jthbku)iQQ8pTkA!W%8O|wQ`?QYcz4im12**KC&u`5SC#5i8@0cqn
zuPn)@EkhmjK~=(f%3oVWW#Lc|QSA#6M7M9?z<fTOn!?UBt(p2_eA}sRa#jVa=4z{-
zxzKzML(tD7dis{5qSgZoNhxGNQqalbXaSB#U9T!Gy+c=|Fobc7@;&kz8tga?UMlms
zDW*Sd2xF&X%>I6tFcSckIm7ojGIsfeZeOh_n)vyvQ60an4QH)#3XSO#7u0Ss-|};f
ze+4PwGMiPoZnwK5-z{|~Tgu1`g4%bvO+Z2NyecXMT+PYp5u!QBSn##I#)jaJsPz1Z
zvpRbK+2Tc%I5=c$hiIUn0vZ|`WSxVfQ4uyxe!JjUxkM|6$~Re`;1_6s)>R;=?9Xnl
z%8n;?bait%J>#Ks!*KYMmr)C)a0ji(eNq{sPt|`CZ?6fzkbhDr0Xv$`w3~PxDSo;*
zxf%w}By1}~1A3S*?!EDV=7DeH*fq#I$_TJeKw?-JRQ_H5&IoP3%ejtV?P)Na4A?3-
zp((06%Z;j|{S&_xJPv^=IfLt;Z}_Y7kv^;8svwihCOk1nmw`O*_wtIYI$WD}aHnNs
zY)#$ur&Bd+eS3R_`gA`$9_s{p@9{|038J`$`BQu$6_w%u98I`|NUg+sN3M%xCJ!&3
zq=@f%ixIucj`dC4a|*$Wr&;4kUCJO-Ja$R9jk%%wXAO5v)ba{z_*Ne@F@^eUmk%iP
z?ozXUk^|f?83K7OFakXM%LE7c2H$UvFkaftatld#@M^O-d$KtffYzbUr}S%J48#Wo
zO!`KJA(Dudz>q??2~)@X!mqCG<o=_n5ZzP2Ha1zIRg7j!F27mT=Y*$G{ebuoJZV-j
zrzj<pAw47n55Iu6IiA9$T$N5;!N3NMW=`cJZXsk-t+0<4lf>a$+93dR<FrrqrZnkw
z6mZ7;MuQg@EgTFuQB8#-?O|SbPqzm%KEgFpS9dEKghP4`zdP9QVt40dzIhTBL~z4^
zhjYUL`rNFMK-;SB4-+(qCgJgw&kPhj-O)x&6NG0U%7`1%_pd$z@nAs%cII@_z9mCf
zG#f%&b_f2Mc)TcJuHYAGrE17Cri|hjPuGEr5Fkzsfo+=NUPyw+tTX3(MbT)fECCzA
zj&Em)!b~bE`{wgv+8Y%X)F_eHEg??%`HpBe<GY0I_-EhQ)%%0TvVjg?3kv$)9L}u5
zgK_jZoaS@}<lK5C^stw<n*uaBWRP4l;4aFf^Kn<|uu%HF7kNhlyakN~V!%@Ndg9HR
zR7J<f_onl^1K`}+DqwD5LCp}M+h<v~Sg{yc%0m0UXOzl|(T2Su0A;$%&%#2QC2P*x
zB?q+m`T5~Q=FM(Fy8r!awcS`Mlsb@!^7Qni(`m$v?CT1{Wp%k*4{R&9`r`jzA{b-j
z#cV|&6SjvXb@sq3;`ZBr$9Q+{eFS8*?$qA;t~;}hjErpZp}!LT5A|WjcS)XixX+Rq
z_iJCYwD6&Z;_2~sqyPEs^=-ptVPT<vaFECRvAz!dNeu-RwGuQ&gpPqRVcyrKb~S)!
z{o`TDVQk&Lq1T;j!$CZP(XMfTnL>_}ySICvtS^CUc?$<4ZxM8kg9PVWT~ots_J>S{
zVp3uCWf&xvC_`Z~n##S>b(hzmUCQj<jA%#Ge@^~CCsXt#a;0tEq35L#7;ZQ|784p8
z8p|NZI^+Bg<p6a!YUru|;o}`icYVy-f4O=`(i?ZPEBJCh*602n=q-Un&sTvA+3@<@
zsWu!W|8GbH9=&@U1|R<KfS?q*|I^X`W%#%MPJnmJrDR<H3N*-$yaVCiqcB<oxqoG8
zpe<$lFYWI)zy8Gd_vrtTlm9Q5$}k^FVBD1ue3)=S6Mdy5FFL!hkY~mBmCf$E{ey#z
ziasL4>w{@F4i1AF&4Gb|b)dzGB|*m-mnY3yyKi=RIb1HU%e*Gykn2^vb`N2v%K@)v
z_o^r<VMr%1h!suSUyg9H+HHJs<RRn*pKJwFre~IyM{R6u*x{y(?RM!iy~oDpWN6cS
zCEx`NY~%}kxO>yI9)LgB;?8;6e2L1DLVw`y)kfU58mU{h`}}yiJFTYoyQnC1cvz<O
zJlv+-<=g<uy5(B6v9ZzRVt^9j<<%2{)!owrMC!MSYHEEt;AJrl4O~D=j7&`IpRQ*e
zda%^K=(c&>+q1B+NPhjg@jF2m*xs4La<2ViduU*Icx$%pnH>tB!*VTJf+(xGIaQxY
z($0=~_2v1Nnw~!Vs0d&o<M}G@S6p0$`816s2R@)1n{CgBZJ)~!-oD|x=qGKy7^G$6
z<lMC7sz2*Qx*w4Ddeh$C4t2W>CZ<&@&8e$PK)|4|1x&pCx<7XMg5b|rx&4Rz?1PgN
zljDV&CC5ow30MlKD>r_(mh`5R#?S;kucVBZzq^=NSQYl$U#**tIa1{)9d<{z2WiW;
z4)tDm=;`UnC@Epa7ZwJ4!|`3OCM2M3o0f-@SOaZ554F_kKF*Zt450zuvNi*@!_IK<
z-_g2Nuaurp?8_-7`47l&eFSxj$2$uaWxw<DgY=pDmE`#WY0|1U*ps6Ml3p~z6MICR
zdhGJP(pIch7oqojk?<=%o&|iO4drzOAwFLZB);$kVi=$!ZgPHJ3J^S0VWB-xG>5WX
z^}LNvO&td4YS@m~ayhE#`~3LNWGf)~M=>^#3TtcQzfpbB)m>rM+dOx1*+~M`Kiw=g
z+--#Zh_(z3B=$;daM)$BS>XfNE^B+{p;ax0TDzXr>)UsH{q|Gpz&CYuoOxX)HpqW_
z0ZxH?I$v)`YSVl!u5A06nmYJZf~IVJrSqO0u-IQbJ^Kj3XD3akmDbBV@pKyK^Iikw
z31@L-9<V;fL-FTFRb!QQo8mj8Y0|Q?C}qtTh_|cWg4jBaAE1(wl4R0&uzcvXYQL$f
z_5zHN{zxe*YfY}KWJ!13!Tlc%&rAGAO@kQoV>RYAl$3l|PeC)q8VZY`kz}la5FnhT
z)2Q@Q)AM9UPj$JTR?}#9u_NnW1K2<7fWKe!MKiUq=sLaKyu453cXyuhlK%QN0zVNd
zzo=;Qw(W)A*x1<7(J`}-253X<?d{E<#lpf8^Gy5o>x;WPFO>W339SC}-+RO7h~*~m
z=Ix_Kz3ulY=eAXORn-~0zE`LUl9&4jZ@RISRaHOYg$IO$gm^bZbYiJ#cWg!qjy^v<
zLQ^p^?$)(ES#3RnS3Nn4)XH|+ULM;3Na(QS0GcBH20TAM=RYNm>|i{X;#=242X=MH
zdp{^CDk_d2UcVZrwpHS5o)%r=>zZsTN~|d8cm5)CmK?fn`@bQr;4QCLYjs{0`wvW?
z!(w4$_jURsxt=uas<(RFC&Dtbva&McAX_ds#{K;Hv;Vf%dikrDm)F+e-@hJTzkYp{
znu1DPOiWDv_UFY~E5q}R9&P&pa-qBRw#G)hyvh6E-lgejF)=aNNP!0`q>pq^nGFp|
z5fKsc)fq7{=uk6ddc+zt%nMrsvHQozJ%`ugEjF)6ol5jSW~8;<>!vEh?zm>wep*G%
z4!|@eZS64Kvb^HruvbI<s+-$G373znFX?x8cap-x(?x2Rd4f;W=6Y({c49^aeS4)i
zP#6?4`T(BNX;!}r?db#%gi=#WOVG~tGY3cf-JOfB$GI2_E9?Hr$xgcaNrw@OXZDkk
zv2g_ea9-!3dTZBiNgP&q%gf70OAWDyrA-ms4m(W1tha95r+9?|L?opB{r&H|qiHzd
zy^{1u|8Ri!Uk((Nl<xZkpF0dA`Fml|_=F4%NmYyDmX@?j+n(JnE-&v-z-^)8ko8gC
z)39pvmfdvs2pUA!dE*2-0JLCi_kSmdd!F7b+Ux>@FzY=o%$94Gcx=<^eV@n{o0^{X
z%sw|~!#inv@$$MZ(yK5UKnE`Ps^|5$EAE2l?k^cH_rydVXTq<40oV?Xj06E^3b`Qk
zccBRm0~EEVfmfs~ENCk#Dgb;FJ1~E>2^&)iQ?WW8)c@LVJ3N9mg4$Qu_~)x`Gm8J0
zM)iPowZzcYHj3Z<CMq#8NVSOYbp1br0GD~!p&wOS%RwQV%w~dgHs)b#FqY0wr&Vj&
z9QZ#MdQ}t*7hpyv0#a946uj&R?(Y8b=`WlP!z@>%_`f4bm|I($%vTymgYHN@cYHCr
zq`3j81@6lxAKt$s5e!ufGD`}}{QJfK|K5O{tL?@J=c>{EpWiYW0%@I{ofxgHrKJ&v
zWo_wZlex^mya6yzKv{)<x(oeZY35B)qsh%3JI6EV3<6f73!{|-2l*Afqmq*cAFdBw
zZ{|U+fC+U446Xci`M-DNoccHKG_Q8KJ~UJcr;Wh6MV-~6H~^R4PltNJ!NDYylz~{Z
zs%PC8^0vU)#MciiE4ps(?vPq`PR@+vAOB3%R}mU$r(UdeQ`)VyLswQ-mM<PhTg6hy
zX1BJcKR<gFt6EGQ0@y8h*$#vDR}W3;U!mts)WcCeJU&*;TPiFDz+f4|QkFKI@RR~p
zINu8JYk?bibcL$?pZ|B)U&4#-)%*c6sixY#7=RYMqL)?I5Jv>Ej1ZSCGtbb!0>SRO
z$z^!1uL|+bi0N6X_B$tv;OgmR-%?gq{sQm^I2F-W0LGH@^Lq~e>tqHo=H6N?eESiJ
z$Mfn}vT<#91nV;e%^vog-ogNCeD2obCz1ux<#pJR0TjLg&^5qFE32rS-)%-Z04S4w
z@cqA@;$0YGfO4Vx5)|7FO2nmU_rC3*FGgAG1I_Q>zoV`3fp@_CRyrN3;^E`J!UF(*
zFhWJsfC~~iK0e;w?0Tu*=tvDXZS#p~JpZ|0c^QkuWI2VK{KzEPY*+4Qu%bDO!;bDx
z^-mq=L@``Fa;_u6MWu})!0RA!b+y3(G&PHhi_3T3jgjZ}A1B+R)K59Cjo4Hu4~RoQ
zVTmco?LS+%FBcbEez!2hPI5*POw{`14=?W<Ab>{9?pVns2<z+7N?Fn%m`cyg+9T|l
z5eSHfi<8$ZRf1fO{C@URZ*@aO?5educWL~|IZ?=a5dyWqK55S~5TJtHQm=V%`sMnk
zAHYv{$O}tz15GL=%}(|HPsg6XvT5>}%g5WC_(S*gHG$Lv6;oG@^WRqOzIq3~b9a`f
zKUJ>n_q?H;+C|FX%p>c0Q(oBU#1}{^2ZJCN1gS<vzWAcSzMbaD*9-}{b~?WEm6%2T
z0N6!ecR`+R#+lQ+$xZyztao-IMTMHKw(jJSC&fgkSwUl&*G`1pDadmuJTo)38Gk%&
ztPVPf74P@|8d#^Mi{>s5Q&97`3|_lEDLbM-jWf)a$mh)B3qdPdQGDEOFa<xp^E0}i
z*bRP5*{$GQdzJY{FilS;WV$;8JbO{~GPdF_Ar-TtlFe!PP)v8a@JO_pj2j*;%!&?k
z>E|~t%y)d`{ti`B^-&$wjrU!3xu&xnJT4hmjH=rwk{@s1o%?teGgXqBHSg~;Tjg3g
zH8AMK39ugBJ_p@>Caq%mOv*x9RoJ+h&I+%r6NeRtd+ynG`h3#T);E?_iQ=KdNN;>T
zHyu~f8up$<#1ccY&;$$o-5wm2=CfOB1qC%)ykkeMwXjiyP!ao0flXVRapPGBrrs$B
zt*iw5&4_<}GQA-gQHL$xn}F}OL$<`FJ74n)we((e<Xt>!bH(TLY4{&oR$CgwVv+Fq
zA3c|agjJ)^UCWVuTl1mbWK!Lp@FP!n`fYs)l1&_qV}6gtoD{Myxay55=)I_0{EM8=
z!O8;3TjZU9`vv0j!4aM6;GcM1>7e%{weUy&%afNDpT7HO1z90WO*3gxZCX8wzh779
zBmBA}de=U!jqtWT>zOP9CHI>u+c2D|S;}ZaMINKc)2X2^_1@e1Y6avY@KXOLzqNt}
z_P;~hVoN-!y{4akibSH$;H%*3@$v?aWh5UAoF0BetPD}^4w5+veFx=}imAC(O8F1T
z>I_mi^5W1?pZhji2cW~qKXcX8*R<AOo69~pf2UufvcY^WVr%~$Mle;wpI{+Wgl+BD
z{%i}_04kImp^R>~?k+^c&$idG6*&#6FHn(-C93=W+LjS%olr@`&~KwvhM*5F7RqOK
z=I1~f6)nvS4a-f==FIHP4em=D@tYj7qWWrG<lqeo7dJ0LJvU=Ki~_3Xk&&hk+g!@i
zD^u*!zcrO?D7`sgrNlJZB!m6*qcF+2$teFrd;eEN7KM(7RphR<n+arwV(ePw1<r)B
z5msF)|8+1YVtPw$1i*f$n%vl$1qf94U@u}x`pXE*+xr<0YE4Iil8FHO9K5>=c3j8W
zc<61IeuD!A&w>y}2~nq&UVp;tWc~P61^Ejs8L(<geo)01GA4V2-<E7X;us$eE$n`D
z>C0RZaLK&u=*wvEi4<^vW<0We6yiMVN>Y&_gWR2YN-z^dej2Qba{D4zynxyF2i5v&
z%NUD#;uE<+vt3w;UoGSik%aq0Hfpk5R08JJVm|^Dmq6lAai1Z4cJHJ4I6iyDQx9Eb
z?hm`%bO|C_ec?azc4nB-BiB}fq50a%q3MwY5WgGL>SGQ+vf3f{e)-#-Wk-iuq&lu1
z^nMW`w1=jZy(TOra8fda?*a1JC@r%d<swz&2zjdz@d1*3-4=<GYirkn)8AUJ%6w4R
zp`LSc7+XNKgSgTk99QY*>M*ql?(!(-0p_swXuR#|7L>ZfUyEP6-=^Iz|6*Gm{M0RA
zZ~PIH{rP(ghC(__dDTn$`;-qc7$gVZRF&UDNnLj2am{5w^+D~*26z=H_vEFw`N<1<
zH-Khq6{0(skx_3b%Lz>6mflm*;*FMJ+wQu+)kWkes5p&@DFz)9SI@bvmbgPh)g?iC
zM7$k?j)RCtOGPN(o&ryiKxK|uzS(k@ya=3|@BbR&v<lU}{)3E+KGD-|T3HUcr>C(j
zk~~}eCjTsH6t_0^kpkY9TdR{ja`w7}R{BnU)fM(~_WW7S*H%w(!p4)nvx=WreNk#W
zjli~W|1F;uDH!T1Yz<)o|5ZhJevQD=GT1ulMP3lx7^BAv-VBGVAEu8cxI0@Z53<LH
z8T=-LjsI3+>V{L^Aj_P_`gdLeUXhDW|IUK37@l{Qzk3um4J<b>mD3NvOo5V%|GNqu
zra1cVdU^!7RhP2l1HAW{@3ZUrRkwMJX_v9b;PsEyOZtP(Q@1zs_98>RIsJ@*_eUFH
zciCOHD@Q)d32avrdDm)Z_5oJo|Dfrt<D&Y$woP|;4KWBqcXx-%NOws|r*wzXD1zkB
z-5ny`DJ3o4T@sS-;rn}^e}T_1XP<r6TKB!KwNKyl<=h7;Haetn6N7=SC%FuszCqaD
z*rkc~(QstO&z)iscqis}z}C5n_zFW)Lj~SX+wYeCbi5}uMZs)I6F6IjpGT}E`*w|s
zDq3n5{Q|z=GGPC^%|rjqc-Zo+Lj-xk!{RilVxVTP>f7EvE-8*M^io(g=``v+Ve>ck
zPZCdfOE{Q)ER5e8l+n~U5?_U%#g&%k<Q~k1A#~E>Fn&{Ot%V)lUd^!(pqOtIR%!59
z^pLzhd7$9A1B-p$z;6vOgIi70@0Y9}gx9Yco?XANt@OmMG7iEEsVnn-m`T+eB`s@q
za?L(jjUA0#`8~n#<Jy-+^yAnhda#1+FnXpXUG>}XUvzXnq5oZ8i!>htWHer#bS_%k
zjL+0wLHvh+yll+w1kGi;TBD5C%vme>BGG3-2g!+=M!tu=Gbw^(f$q=`Xo&K9^03`<
z8Dy>yfI5hpp{{sCC*u{IyU)a|Z!{-L!2*%Fm71kW|C565F<?!@;YcLfh+2`ft&)TT
zmozmX(YSyP2)3zL(W}1G94|3g(!|(x48Fx|HsPl_HycrD+-lRbs=DR_a5DK>$u7tG
z`GVFTN#P*DBZ|`dOtf+7S#-XPwJ(`J!>Tef_ZP1;`NtMn3qZD^*$XUG2BPf>LXEo#
zf6(Y=FLl*_OZ}p!vqc+<#B;A6M1RjK>{2a>9{1kal{<R7Hf%{H&{G@ZDaVjS>--04
z)o)fl%}^K_m*DVK6qXj`)`XOl>$}6desKr${h}(~{wJT#w>_g*J&4Lg7R1r-X>HNS
zw&d#ksNixM=zALx8fXg~j82uvc1XHWqc9q3W)cW*pg1y+&YPko-0W9jrWIiGVmd6s
z;&?he+$}lf+4Ab{u+GiZ<r*5Qo;6ES%K^<sq)bcaB_*1RS`16nci*ys?fg#M7wAVv
zdu<gi-{#S@fg}U(+eN@2xLlYERywD)vlagrnvX6PE&2P}c^`@@h0_*NYCFDw2O7ow
z<)+bpRS}op_z9TwkGv7qS-KUhVp|Vg4M3*3TPeFuuXDTQT$K!6y<7j0iG$X>^1pDT
z7K2)!>I&|zna_$epzfZFwS@CBpMKm(*GW3tFe5SX+h@^CV6Fg$&=8`7bz}PiT^)&^
zy<Fh40@hs(mOmbRX|-X}5=0c~oz2U&)@4ziw@d#K(EE2T)0i6rl<_rF87ylrN%Vbe
zt5=TaGupr`emYQx@LTJMf=ZUFYPd$(!*GQb90wBAU}VQVz!!d>GjE)2p*2Hi<FicW
z9n~HIhqCg!lp);}fs4XO5#^iEatLfbJ<fby)4`%DRCZ;k5Bu%z$nB4%>y2Rl9&KlH
z`n|d1sCd0E<_3>4vmfm`H2-EOLnkh!=!GXaaxQ;AP)M8~$v-TN=$vnCpak3u&7Gwt
zB%ga$jTx-UqSXvyWZ*qa7a$jhnQ}31i0UM(Y4T}mV!!hp)Indz{x2f|Nyk%H3HKUO
z@w}=IMS2SbssJAb+_?AEhDm2r>a(o^AhkGInKjy4tXY=Y?*V@*V|+_Ak;i$o6`vHu
z_YT6=XnvXWd~!(Kl^T#cIWCJQ%!_2gu^_lTnIRtC8M)o6BR$VK%}`LrVGMlL`%6Yz
zJD%upjwQ~rx`~sTB}c}db!&{D#q!W6+G<X7F~?}sBj*4q@b*X^CJ@>n>lLb%g`WQA
zM$vHvF@E<97~g3mCdo{W6{+OxcDfUcv&189Zc$5wa`lVNRt+hlCZup`r$4op5{_;S
z&u1y?#ui<H9Ea$`)|PaVxRWB*PR;}7P77u$`_yA$&1r`?H5?@^@>ExgDELI037Y6p
zFrg+}q@hg1(G53F==(qugqt)y=II|i0;N0wk2TY`gSfxI23pnXNvdk@tv7{6w`<Gc
z&K9(h&tALIRHLuXen0r=unoNXRk^<Uco{92_{jMRBJ(>KdgyHC%}nP4{AJi6J;O!j
zVC1D8a<ql@C%Zltfl>_WAKrrM;4d0G<IF8M!RB3DSa;@Nk|-_16{d|_Qw%9z`%87@
zZY8mh^=3NDRyF?vAnx7njzqt#ZquDQ+vt|<`xKAX%QK-`E#!{JUOS-<V>$VmvxC>W
z?@j4EsH1B2J?HbtROG5Dplj&(c+je?vzcjLQ2g8v#$j}QQ#(Q3&%bQ-y6x<cAI>iU
z@<iQ(=nqd9j~e{*{kih!7OR8G$zUq{=04*gGh8q7N-1uJR0Q{Ap3CAuJba#l9f?SQ
zcOIG^L`dvwv7U!#bKrBg5B~H&o|VQxB(|RB0M3dAW?}qwUBfYtaNQ~m=#j2#w<eV9
zdKA6V!PTq+yt!vefO`b$H`*lkJG%-$b8=jX4eMNg*;$^2Ni}L(FPm{8A=$-dnKthF
zs*gJ<W<&Ody6ZU4zu9RH4p^ztkJ48cy?<EX{;0k``)^|t%gctDch!uM{sv;6-`)!S
zZGk8*0+&Fqwz^~np5Mko^zG3Qt9xF|QFUExeY)$2GX=1Jk~|G((2*`%P;$Y5n#jRl
zQ}K-n%JG=Lo~aF6_v*dSn-X3Xq#%g`EKU3=mtt+z6jsfzB>!40T;ZC%g<`*mdtp9(
zG4R;jbzH*Cv}|Dv#j5d_M&ss%2#_8ia^rjJIV^=!fWtt6Ps3ONTy;?cAC{7jr}ou>
zrAT_;$<q)bYR!}ucO)RTSD`JlQ4MNV(Lf9mRp4DCTQj8GlFLE&wo-)csycGxOHX|)
z7Lm@StFs5n3$YPCEvxoc7~(ijvkg;hfbBws{TGUSBR^kTa^?5||AbKXjyZKG5rdo%
z2~fSK93;8QOjnF6oJ^%nJ0Lk64k`nMBKTz+K0QVvMqE&BAcTkG#;uRq0{TwB|CrwY
z7Pm??sO$$%`@;cWu2{6o(r*U1X1i7z@W|Sitd6jMGSydmPciX%Gr!#0H2w?5QnHZc
zmuI8-N2euN=1pLB249Mum$jDBfMS-!7sZZymH?r;sMh?of&w;14K=J@xue&8HQMoN
zTKr6DI4f`W3NAX(1~%g2CC=EUwe=l=$Pss@EpxSqt<T54)H1sCDZ{90kSxjxXBQy`
z;aTw-YZLn?3U^w>Zy-to2VV1oqM!#4V?>D)7hqW&{BCSw`MT@oglI}H7w%}x!>8!p
z+;W*MUAgAF+UO5t)~~z>V)1%J&esQ^O(3#mc0LV*fIy$VG#GKNFiR;;QprLv1^%{h
z3`1UuU1+(JLgoTKX>-yh#T%<j=E<*!ubkO{sFyw%Sz5XdNHm}{P)OXx<Ieihjbq%T
zGfZG?{A-=v_cN6wL=<eOA{-xyF+Z2wsjlp5zvYggH;un3?2J`s{+;kIbr}Sb(nx{-
zaYLbSNc_2T<$Vkz*pz&&;$Gpbr3P!o-GU0v`LwsQ&~zF|z(rq<fslI6nM~?$0$iDf
z|K;BYpX{oi8UH^YHWVg^z`jOeuseY2S#imZszF6buCoy(Rj|yiyj6eVWP#Ttn_@{3
zx>eF}XjB?ES4JgdNOAIkntwr~HcEK0XN8G_q2?)-VZJ9wY{{>HuYxB`LYl6`*@Qr;
z<vl%EZ$%a*?<4R_p!XV{-INlMJ6PWeT7Zdc96^bhc>Cvj)cTYHdY11whQ!mTe3{a9
zR4L;5c#XYIBH<hm-hL%z<FEPZEM;v>VDmvSQihVxlRsFULy8LZJ$9WefK&y@jZHJ+
z6s3OWioLAz@gWQ{K9hIwp)hOr@nnn5I=gyzI0X$mMxDybG0Bc<K%=bZJm$^LeFLw$
zssH=*3EL2}cu7TZs>8+8ZU#wJ0}4DcjNrAiph<@3LteLF3CvPxHa<f92Rhf$IPBhs
z&>~Jasf?DNMh1#%liB!Fd`Q|`x({=L)B}n%&Bz1cSO!x|xUrj_D}ID4zSAd#RX3gO
z@81^Y2D1uiY3Laz&NN<Fcz|VG!E&JCD60~1D=i_hnl<-bB+o^s73X?M9~%(nfgr&u
z;kSz4x;y`B#s4D$?g<D*RYQp_J%JBNL-_nI5}t}&X+wMf0#BA#`M?4xEyal?W0M@=
z%4B1iV@(<%=uc3iI(WA>`nsJpM<PH9txG}6rp1QOjEnK)7On#CNjkDOW?!Q*wDNt0
z*c4qQjkQP@5ZMhEnKr7%M>j0LkSnrU4mGS9@C-!)5BaQrhR>I_ZdC?Hjvro0!L!J&
zA^OYL`Q}1xX{*}u4t>{#!)bAI+IFuKt$rZUVR@-FLi2Nz7JuVjd<#j`!{nDWTsx3x
zvMP=hS@w$!NBu@y{Op&NYBHb;Hj51u2uNx&E?T1PC8PiM3`^@i75kS)+&GVJtUKQs
z{1zcpMWI$8RD8996TIdeNPv4x)8&SFc>55ogW&6E=>}E}GpDOQUZQQpERLeEUY(8S
zJ$$qpM!#!`z3S8DIk5Icg^PQSU}ATC-vb4~ACmqT`7Pz$VZSG_=nIbm78v~E3Z(|T
zE1{OsXkDo@Vvf8H<f!)vUch<u)Dl7{!DUTO6LUrCrNJo?xXZr1Y1C$437<4^XBBK~
zAC6o5ui#|RW`_wro8iOA#!PEoA)lmPMn?(HmUj$w)#0PZE6UF_X`O>|@3ZB0r3evN
z9VMnKECzuk2i{hS^>ha0n7wJzv$lO=zo*}c+j*WcjVzmj?SUwP*B4p?r+O^Prf|Rq
zuTA>mY~BQHF~XgR*VdFC;0&V|vJOZ7kRcZvd3o+t6SY`Py+k_$oRc)MqYGGa#em$h
zh~d_%rmCi&K)4yn1b35GPJZ$2z<M7phQHzW)a-gWOejS;$lO#<RC|N~a7<iIx1=%%
zUNKyV*SY)I*^*~*$#!aGd=#s&|6zdyUg6I?f8bS1zoG}>cXDB^xcQMj9gq1sJ5~e@
z6o#saPxkyAticl^8O9ldWm}KJtQ9N1@>3SY{;<mEq=g{(3d3Ec$_Wnl#{H+xI))>M
z5^L~{UC!%CEQs;r`jo)xMQ?SP-vIg_2$FG<={yXd>v=h<$D9`4YRW;?{oEhkxQdUg
z#_@0YNU2TI9`=v!0JVE3oL4}X!CZ|&_Dv+I0qU{B3R8?L8Y1ORniTq5QET|s(i9vQ
zf75o=t^_x`W=~hR=j)y2+wVVdWoFQgZMaJ>qEe%tj~eu6rH$1OWP(ZGXkqoqqftcz
zmRiiKTqQJ^6RX}#;+9~@2eLPfl&^L$U+kQuSE-ATCd4DVhWve9ChZysXyD`{8x{Of
zNYrAB1f?~vCJ!fb-20TENket_jm4Zoel1==ak$M-lQ!{GWVTt|?w+QS72mh#Ply@}
z5MQQx&BeQ*v4tiliKO3Mh3E+`3-3rs{KoO+_+-R<SGXM7?9CJ#C=`gz$+<0}Ei%#L
z3+#E<qc~up*HXuVQv!%PeqZko46k6=lO{S}+dzRmzVSIs8Y<+Z8k_RA@5S08t@`z%
zDSy{%=zZ19mkZF4=u6ZKI~wT~e8fT?;fgpfFKEwBC5%;y7SqhAPXsNh*`qn-&VH?E
z<oio;*eGi(0$!Up#o%PKx8J{)5`{QsP+{nw-Tyn1J@PBpA@236+z<Y8=jGY*kv~jO
zjQb}_rsdhn=FhXFww{-A18tyDY_D>@WyDhcQz7BQlg93OJPoX&|1$>(LrmJTH3p~j
zM6C0-EzgnJ8sm6wqUdZ%7BePB$8<J2)4B@t8r?o!aMW^%kQL^4N=Mvl49-TJ*L(L3
z#XH;D!|2m2yE3=>!6`W5ZS;uBZ!Jf@@`1)jt2UhjT{ZUHIe8)4ECh}q=;m>FzspCi
znHt;OPkq=xEuKcO1FbGoINHMN;kuP{S`K9z-{6ICZtAo@@;u#d?k@hC`^L77gc@w=
z@vt(;&jdkxcqd+(b}C)=VR+7+`|Vm@&ap?{0nYQERxnV+>rO<vG<z=UrUhG)Jq$00
zpP8d7xOnh?>F{@OoW9wz>=~J-AwC?8+&3gg7%|O#|LPsRnV8eIwh%L9uh4e9$@)_+
z`HE(pdjuehH?U`G2zOZij|FfjAbgE7hJAHCDZ!Rpz^)aqX|(f0FeG3+wwo!OtJ*j|
zU0)<fUyy~0!N+{<?`s@DH<C<_@?Zf5D|-bN@{h*;3LT=+tr6X_L8dZ_UzH@;gdqj$
z#G6AtuDP5nw7YKpe2IkVH=I4Q_z(VRb?)7`*sok{+L|439<Li(imjPgC{WD79tCDh
ztc%zsrhgGQ8{<C8!HskZdTS?Yy(wA5N`w^`|KitH6Hu7M#L-}h!^+Rb4{<Uy>@-vE
zrc4V{V3eo+R(O+NXQIt#HwIXNoegF)o)heSW;Ia@e#oA&zB(#=x$nwX$ASM70EoGM
z&Q*83mxA9fCuE>7dCA8?&p{P>BxMlp|449vz`bOrFyUUj(DH|)MH(7590`|wC?Wyi
zDjNW3GZ0A#Im7HY!%WG_@_cy8Q~a{gxa8Oa{IvY}SC%BrU&~KtK~bf_aP9UWIIgsa
zjiN6q8(scAUapQ6Py$RcqchDgMMiLTx$SK5x@}{O`2KBSyjW}mdihlvboEVib?cG8
ztfR%hrDLI^kE)EB-w07v(#N*hsU&q8cYZV(=dx}0Vk@vL@9A>9iOaL<6NvtfCyFbq
z+y$(~nCyN_)#-k8!%_Fr&{0E_VNpF$)6}RZW$z&o#%{?O45PP|_@#UZ$(qfv7I__I
z<z0e%1Bp-8S`|iG4lxpr=f7DB8+^~>&^kWnx8wbXsaOcl(Bq{65n;Np(ke9>6H{d6
zffZFegXCP}&QyA6_lGZ0#w2W@!&pxy@q^3U{Ej`!>#u*oo`6|!XzI&lzLjibGyU}U
zW_Opa_N~Dd5;2k>CLG{zI#TZUj8m<C3VlR3bZnN8L8DGLu1R|fEgpfHdKOm5*0CoT
z%F;5fb&#typryk(fUpq7*RKAwOCI%(SAdY9P``CdlTN|*b8J7LvEE1cPAoVw65VeN
zktXP5iu`t3@=*gO;y!$Lx;L38ZK92an~mBnbh<qI{zZ#SpgVgstn3<m)ykLa()V)H
zA{U_p*2#z+MC6jGXzZ8+!+T_f?3sbM1(%-$+R{Wpm!DX^ft}*{eok?tu}xnyf{-4V
zAi3|8A3lfX?(4E>ai`i=wwRHlAS|sMUE+yieq`xeX}@c;5TXt)->)P%&n?$iK0eeU
zXSGBJkC=*)_q}b8u|qajSKRbZio_Uv6Sa$HEED0jgGA|96nm2F51<Z%u+LI%5Ws_a
zvGE-BatyTG5Aj}W5x4{od=*|dZTNF%s)E*K(kkL$NosgXTUa|2ExF<ppgm+cj3WPg
z*u%BS{gt6{=m))rt$1`UxRpky-qN<C?74~PNKkE|0qPQBs*rABn|;zWt2a!4cpZpe
z;wofzz4wkw(WuurPU_bDW;!gN|H(AOIcIQcp6dXGG0+ied9s(U;P;b@`_zi47`aGn
z<tH_a9&c$_<aCQX0QTArgW6&zgNnUuP=gZ)Ux0|Sr<)EMR2(WPaC~RD#9z9%Dk|#$
zQ6l|ZiU;2VhAd)*FJ#08e;_Z@&CkAE=1t;}?rM-B{#IyiVq)hCX8bMkCN7mwQ+Pcu
zU7nPO1-FhBVs&BZ_wW0@2<7k<nhY9;15r&Cr=vHGcM`DX&31z}O={fa$5wrQ^m~jm
z*fxNl9|=SwkrLz!0ok9MeKJRE6GJjTDbSEgkqq4TX>~NQ^M=s8lDhQe`)&35O~pdm
z_Js=Fh(}3SOoPT6<09Z<JL?3Er#|BGK@5I!SA?E~Cf3ua$aQgds!!4yy!c-*Blfkl
ze;sXRZQxB&jS-NElU7BE$R2e%mmT0qRxjpaV!`qL!=TPG>LE35S-etK0h(NW1DWj;
zb<6}~w1-F{=L+*h^iPb&%_<AoXZf(N{3DEqVS>lh-89d8$ABGwUih5Jr(EW^3)o`^
zgOO<C?}1xWB5;toTUkNRd;ye`N(OVTt6Q{yg!VLglH-ke1uibuH#!;6{MEd2MG|H5
zTDK6~&)Y~SdUkE+2%Ckkmxb=Bpw~h>Js8|VbXhhsZhtg{N)1KI<OKcu=l#cZgnf2y
zYM<}=WH}lj>4cyD<b9Ed@;&*1DtV<L`Kby(7~)ow$9K5}1$Ay#H#W?^26kh|)(&Yl
zy@3MpBGQ8B@{@j-ELw2>LBvJ578^%0!{OSYKuKtK<(9KL^p!j{RuBw#J1qAQPq%*&
z@9u5Yj)GQyvT<fVP)gnQ7a)&UH4&aK>f^s*K=+Kfp7DQ_Vhm38^By7T24d?6a>dh9
z(8m>b!TxjPSe+y{z)Bfe+R#U*6ZObT_uR8y0_3M4iq)m%Uq$2*Q7E^?iJ&9AE6{*|
z?SAF9JS!tc59P+c=W*G3g*^VN%3MdJG4fS`EoRj$j-Zp9AOg4Bj{D?ixB(ur7<o1W
z)QZElI9c(gw1IyO2!M`6ei-{9EH;;;-vx!LQd%0Wb@E^ViA7tC31W50kNHcz#Mgru
zFg~vq^bG$nZJejhR6b^E7c;HQgrR2<az=4z2r%Uu9<J>J)kRcFr!aLd^N&dfSD8t4
zG+QIj&SGDnFWf#<31o6q9+xtbeUi^AlzHNVbB?Pni{tzmrg`Ad;PWR6U!c878}MES
zP_Bn?i#6_iI(6<Ljxhu?Xf>Jllm&%0?>bI{%qEpWp+$#nStXjocTEa@|3*yufmGF5
z%l3leJsea$v1n@b(SH%E-S>Td7=Nb9`ZxL^q;(5VU_z!WA+;l+gzQ(vEWsDcI&-3R
z-I!i_rART3a^kKUtZxV(*c%<IN60dGAiAHqcg^qBpkG^hIInEmYheShh1j1fA1VVc
zhY0`|@8;bsss`xb-G+h2h*!XS#GBJaD*YJ;P4WD18WX-|5G1k3p%P}Lv^by{^G9H6
z?hdsjg0tO^7~t#9a(v@Vzn~&ddvyEdAjG}rN^>@r1TUokPH7tO1T`PNrwtPvC1y$2
zB!uHK)kiWMLYG%e+?^mL{+g^+u4O?r_=;`H?+4p@#r<c&N5iq)Vz_;k=hlBW%}iZF
zj9%MF43MM!XZpwp_u}`RaI~3&4<nTFeiuL9xSQB%D&7UZe>YKKH?n9C3UYxna3AUi
zI^eGx4d;;bm>hA~u`%q37w-H3r|!H1L~%HLYV#?V|E~2aEuP35ptZujCM|n)W>v_#
zszHf}D>nHnbW#bgu~4G?Q<2&w>!SM1hzU7*J?Q~=7ihdAbo48zE=|9d@0W0kttHGy
z)?4Ode#Oar`&jhL;E;BJM`5Q|Ks_yqm>6(hU!a2Wkw;sI$x$ky!;(n-EICI&7q<}^
zL+Ps5mRxq$YaSb<VIqtx)E1c>(?<jP7pzGxPc0%hLa^iC*zJ$3d*O#o$tUyaQ~%Zb
zmGIMmo1gB_kJ~0smM>2-_Px@O2X|%+n3Y+(NU3&PgbOPaP%SWnk<BRu{Ci&Xca%ks
zUk~zOa50ItlBtAIg|hZuZ9|$^ZiR_Az)IKXeq)m#y82w3yV1}fc9KdqpRd|%DHK~H
zhxF2XhmZ>R1UL})iRP(PNEq9LigYe+qjid!-u{sK!S6fg#9FQP{rso_TT=_Pnr^%>
zWoLtW*@w|jfIkQQK3Ma${+RTezBb>>UAk?9^sMbdb5hKiJPQPAUID%#*|`3_KCdc4
z{X3#+&C=X_YqBWC>ZNO+CLAY5{CoRhbmi%Dz{@>vEBnQKSHDE%IS5l#u15X<RqujR
zT9|qc-$W6nvuj}xy*3N=vszJAZrz+93pkJSPkAeUDAE)F^am*D1h&I$4!P3t5Y_Mx
zIXs2~TUsL|z3#-?-P{*y%{Y=D7Bn7$cH*c`-fdhc)D+qB0y3V30F6!?B^2T+vzC3(
z>jp`Ma$H&*I_qk<%cy=^M&xdz<sf63cVva6#S@~AV%3+BXapP;mr}5lu4vDOnPBzB
zF8cKq64ic)=JOo~8t(S^gib5sNT&BC7|*M7JLjnhIs)4>{N2Wz=<o|7Ayo7jaQ37;
zVVW7bDvQT*EbfHC-v~=IF%l9Ig#7R7uNr_zpPEWeB^mH%3Av`+yv0K#qrcMV=fM*t
z<q`$Sy-)9JJ%!g>ha`3#*}W+2Xl2?1gI;4l2kzq&G}*qtqB?SN*oQlCV1%X;0&dpD
zcj27Fbe!$*du)Q^K~sXJRQY`^8p;`H_rC@(AHtuN{7_D3Hi?V=O+W)0>jUThEyj4C
ztenV7HpV@BuTucIafv`arj+o|TbIqz!S<=ZU3N$yHEN@;?<j8MFom=J#NI(?pGAD$
z-L(_*y=jf+CxxJ8oEw?fL{t!=$)GN#aXQd?1vrq4F$5@ix{rO&&$`s8XcqtA7oZ**
z&Qu4q)+hFuWMzbI(RzC(Y+bTZwpQeo(lTGTzwXINiHIsSAO}f}g`a$8T_A9zc)gyX
z$mo328>CD<__8xyY`HPvnp@SnHQeknD`N@x0sorE7Fpl#LoPy8?P4O*S555I)Y5n<
zAYRJFUH$c}$1Iik;a`tSv0wSAH}1pk-1&#9x0^j*GP;_IlGqu7mzv*Ul*+i&na;mg
z;c)YI2)69x{#B*NU)1)GTz(acYJ`@<*$4v3XWZ(w21n%&8r(FHs&2pI9E@Zde^1L|
z)<64aP@6dVZ~hIMxa@w|+QL8gKLX$Si)dOvU;eaG2B|4J70GgBeEU88%Ti~+Rb?>J
z!yXim^GkHhbY2935fTv}+YPr=Tey#cAgI)K({|cN+Horn)Mhz`9Gjao&qCM7OW~rY
zt!~pdnZj-;FKC^G-sjTNmsx`I0#1K)r@;HA(DrgXYOV}k1z(Slu3}|qiiG-b?X;SB
z)0gY_fsAp@o$D*gP9E^S(z~<FDvC%Z@=-akVu&d9AHkR`F`l@r4hK%gv6F?6@rmjT
z1d7j%Kb7b3xkw&lAC(OktsRFB7$3M68>B{OqXihMT9GP|V_WlsjMqQcH*DIHz9t}p
z!=)jUm7n4IzATJns5Miq9dC5W7<60)v*y2G5A#&@mFC?>l0rY+@kMB$>h~hCN&g;i
zrS(TtIKq=A!=8ngM2JK=SQe^&7NVjjuxaX9Yzh6#h{p?Ic2B~|Cc(4C72&8?r8o#9
zBe=0RegrOHH0^z&iQhjKB9#FeF@I_xk}VRhE1)libs9V!k&Ww})>)?7Oj&!`I)WJF
zd)}qYMx}pNSBid0BM6Wls-Q<y*zC50DnfMGw9h{!SLv62KQ%ULd{W@tLK1iWp~rQ^
z#LR+}<RV_ojTMSEE{tM!SCdhX&``h!XpG9u;I{6x*;&(L|B*~OVECd3^=Uh?C{p0J
zc6?-g74X8lO<87`(U<}1xxH57Y)5YSM$JC)!+=wJ7yy%bIWz>7=dZO7Ny~Zv5}Al*
z!_2VHf^pn6+1I08)f2MVCyYjVBoZzU!O&J_yXvsy9DNsDS|U4#A_12E@nv}G93Usx
zUexcygP6q4kMUqjy;fj^L&#+V35XdeaJu^X-Bpi^RqHQ|cH7NL^cP_6;dTvGG9v47
zAq)GiR;E4(P)bbN)lK6+$1BMJroz0dFr6NBGsuG#@VJt9JCE3;+o#StFS-{ZyMqK*
zrZkD()nRDdyH9)**O8l`G7RkebqG$ZITwkv{ua=$U^6b->Qyg^dMoE_1jGFcsn(D8
z%~Ur&$JdlYXXT6*SdN|(12zqI%ejKk>{+Q3T~d@5(B%2U14BVv`-3yAos*xol#gS2
zZv~x?zV~3*0Qi!pe^X!%r$jEd43?iq05R6P3d>m1d^=9cvZjYZf4*}gBLWtxG~z=A
zGdjBqzX_#w{R$^5&aG3t_TU45qk;nmw?j1;$phM)%P$+{#rho-y<o2;9atX_Sm4VG
z!WSnZG_c~U(6Sg5DH(CA?$G&GxY_L!+j{NFxi}y<9hzeVQm&wgO{o1Sb9P)dGQUbX
zAd35CvuVYP%4tb3ByoI*C?Gyrv)3C3h%;n=3-zRD*im`WUyS_OGK|bLgV0W%%t3Tc
zbID_-;j}W_ZXcsKH$Rm{GTT>ui_+>_>}8RlLt^shIYE<dLTW7$$7*{JHg*+_Bv({8
zZp=pIITqmFIJ-_R&!+L;Ts=`ULEzQ^g0Z+*1j=DAHTjW<cno~NTKf4zRfX&<`qSMJ
z6?M9vG;CL(#N1~2rx8NUo1HfE@!Aq53JK4P*BMdr+**iDSaYb4_36TN?sRS;mUw<U
zNOWuD=14~#|3+J4JExm<lloqyg(8jr$LWq`_c<dW_k+Z%4jKT87rR{xl{S$<`vQQI
z>rN)=1HN%8X%ecYPFaFwp9uGQF%MkwGDadC;Ma6fs17qKV(x%c{xO7<MBwaZ7aQRy
z@L?VlM@VyIo}U){o(w<^#S>hvDTI}6e&kH$ek{?Ix6Sr;d@H!-o++wIU#{;C4<zUS
zB|*zhXm;(nzrV`>j$Ku?E)T*M0r(5zBf6NU14z=vdXzuL_Gb{uiE$#~Qar5ML_=9E
ze*M86fM#xKAe`otjU10dJrJ?&o4*Ljv(dma?MO3*be)#va{`cYYrVSt2d0P$WIGl-
zJ3>Vgbk1?9(jNSwZN6!Ln@<4MCS<9<LJdql0T)3<tas8FxAvV-PkALj!cG%<1Rc=#
zIoy}_>sEkTP$1N0HK?5E{;k&0woiB?7bmYbb5IWV_ezlBHy$VZHUJ5nX+W++-;<RY
zz(|niCZ?fkUy1X<pO@Xa7=7+2MUnl#tOO27<kKU(AYQ!Q^dAOB+@J&WKHq;N6U`)m
z==<w_ZH1L|D@SM0Olg3>%rIhy4Sdk4c)n-kmCP~X8v%;Ho0H`M#vDQbqC$r0u{W3c
zRdoCerS+B+`jZ9Zt&pPKIhT1u%>7F}ZN73H3yWL?fDZt^+lzTAIMbP1kwYk3Jq+J5
zfnKAqkZC-$3AAEQfPTdGHkGmA@J3$&o5a7yC4hk~gTJbzCv>lxY_G_}97a!)p7@2S
z1=0RmvfANNl>5`oNygp|pi&tObvZXq$4z?5SOIL-`D1k*BGUj*zpg-nN|K@aa9g+W
zgfMZARBX-jy1m06(hT@ztf<1D-(GyW7_*${_rL%KEI+#nS(w`jDyw^R?z{FuA9aMw
zG<7yO1@KHV`yB7}(t{-kfv1n9l%)CZIIi|cErUs*=_yyzYl2LyY|0tTQBsx&XZttX
zGai7lz@ULd)>8t%BYC3^s+oU_(4lqvcWC5xbcnT5IxzsK11abPm#Z_!cUa?+a60z}
zlZnHr&<s!m+0;b+d>OnQ`G3ZT-x1D#k2B|D6rujMtPttShR7~eQdvs4#o(TRDq(4*
zTCQyo!4_I0M}$0-viD-wRw^dKs#zfvqMPr=tIXete10o;M)cMua8>?|u$u2!Ov&w_
zfKqFRr#m@NzE;`MwoTy&YP39nIZI}PeQ6mR6EjhRODh}+gh@PODFfYDWs*WzwAv~;
z8T9HDyjt2TqbouKP)cYM2}5!|3%7k6$ZF@aTn?{~{>7U?V^lth9{x*sglm3@)KgPd
zEl~>_D@}4~8Lo$`*}WMzSulzun3w2_lqoqNkQPcq?jfiyH(rw0A2x1l5cATFL#>(|
zP@|>Pw58!%@Ar^^n*ZT&n*pvkeV*6uTXB7Wou=v~oB0m6SCh*@FIr5j2c26kWpD~r
zkJKM4J-TH%CFq~*iH~N=KgO#_&yE`00Vf`j+#xv*O}pl;U@cVU!%^>NgAd6hc)N2^
z^LB9xupwb5HU-*3UpkSb2wDRzu@QPtf+P84LKJg+XRK+BD<$^gU-#lYb@n4C=93;u
zE*Etkgp^vy;1S(TH&&eRLPyyd^x;w~cb^mL(}DQlS5GGm|3Gx)urbO!dV>-%FjW)#
zPpoSx)~aB457q5K&Sh`Bj*q**L<qx~0kZNmJ5Q&CdJAE>P8wu-mjM}B&p|{X5!~;l
z`f=)Sd*ys0Aw7Ct^Yc05r&hw1YCF6R=X+znHTV7xTx&8=O-+zmri0HoN(C%?K_Ipf
z=kt3u86HC7_&Br)w638(w*Q!+gZ)&Mss`0^F#Co3w`Z%cYw5c5pIdZp-I*%10yXZh
zkZ{CCegk;;Ck%j%Jx^+rO{XAYo~%Ck7=1LhbRl||C$lU*&puJk(Sq9UzKRAk&@xZo
zmBO6U$|oaobA4*v<-BIk`Ab4*iGy6h%ohO4@wD}u7zmb#S%K#mfVSOH`38<)qk|ZP
zgYp{Be?pm9KM&U>m-YZ$43kLL$OwwwGz~ExsSrRE({gF5cX*je25df=t2ju7SccBu
zjiR@h{8f~Pk&Ud`RwC(5*XYMfk6a47jprSD<q`A_%AS)3f#f#p$>%w&plDQfdIpuA
zVOzb}QFeV8ETx|TFH;us8GCaU9~W^+ypvkS3fasf7k0)}*WmdPxMzxnYt>s*L%=JZ
zTe#t?MQpEA59Sby`{>qr=BT(@e0p5)N5o%2EHZTIY>WX2)aaWV??i6Ohg5bAs6}qB
zRzz}B#2Je~etR8ODA3H#$+z~h{Be_@_aFs9s{2=K9#IAY4@jwpqQAo=$Y6mKN7=BE
zSE$6Eam)NL02(G<9_@_hf$Z5-36|-tO97@1#GldOTvohorP=R0X|W~l%{|wa6(Iut
zS%kWuyV?>$XFseJ!irzai$eaI15`?mX~bd#GallG%x8!dcHeJ+NocM%eokd4DkF0m
z^dNsH7f&Q4=I8v|Rs`_sy8)6sKPiXc1W*)`PEACBtPjvgf}TI_6Q8k#3JSzu6v<JG
z1n0>VY7kk@_*4NHQtiwRO9CwbOP{fsZ!-hOy4;T6(Db#(2rjWR@Qb0E4I4Xveho*q
z=Ii#V7geMAX$^?teA(2miCXo&;D_s!(zLc)whv3F=!ttOMpK?PqC_X|`#H?ZM8yMH
z&fxq0Dz!KWD4oM^gHuGMEfz$4q*emg3z#3;|1Iz%I?&2_lK`N+H4<RYyYL%y*L3mm
z;KSGLtQLp>waKQ_MP6cga~P2)f-L&nCjenOe-u2ic)9r)KJ$J(mC}G=1&~Xbk-<{N
zKRLm=1RecSqaG7`F7EejL_iy>vJX7rmuO8Nhf~;DhxOEfIoGK9nTF4T*SvS6aFK@C
zSH!4$EbACaR(QSLdU~RuM(*RA!N{wqkf#(?dPt*^uv(&EQTyuudLV9de|kC^DatI#
ziM0S{YF6VDQXDco!cCwDhOhKI6xF%HEXFU+xq+Dwy2<1F2{3=2kI(CFqZa&Ezeml)
z{yaot-7)4T`0{%}$C~a0ct?B8Rog<Hw#^=l8*5t4aln)<B&1wG>rj(uH^?N@QYt9!
zhBO1OSfnPA5fRXLk2~wvrh!x6R<rcR%mH^exH#yYm>a%HfaPLZM73(UshHY3=qowu
zth>pj^OL-U+YXaJvQr_kxFv<V^0P^83sY(BuJi=DG-e2#`2{?Ko@U0{k5AaGj%?HQ
zn17D1>Pu(*_8-;oDS|hV$T)$r=2yjzQHSCMKH4(YN0BtrJ_xt?^NIs#%%skg-bDl%
zwCY<d8SF_#(Ib1W$cAveXUc*n596=sk<U@K8!vzJsG-)Na^utl1gpL1v%xgmQK4}<
zjx&-RK|J>+G5z64YF#8?u3|mSfqsGU4#YKA#LMb!$f^xM$iYVYWSIcrF=IcpXp**2
z^9bGj#`mGj2x6LQI_oA{J{n1Bt*^KioI>Y;lB80y2vi~_9j9nS8M-zr8DNexZBMx=
zzG^HHv*~45#-r*l)r;mnLCYdfucM{j-la+U+Y`LZvqT@tQ-S`7T?Rn9ap3`k$;+X_
za~Qv=#uvKgF#OVI1#+3!$jZ7%iDg&4YYYQ@3a4gvDrP_xAVh)bo~-RE4u{|dq~bA|
zCP_}Vp0!^!0b&wAZ9j(i7}|Ryi=X+Bd%wjb)0*qLM_W4<ILi0}D0IM19A3&^k1b%Q
zT{^lQ-X||h7tV~?XpEP5IQ%Y8KUXCC?$lK;WRPbb*fQJCfbTgdhdej=>F?Ls@+FHb
z3VbA#3S})<BLzfLYE#2O^iC$JJ`w?02%gN@J}F0DxwZDUM`yE3q=cqpN<a^|oc6+L
zJM=!X_t=PiG$nfW@x}*mI|ZPjhv*xwrgqiBog4eoPL-~UPLDfFvK$?wPD>`rms@!i
z8dIhA2@?P-%u1S#4Mu=ZZuV{M*gs@obCp3}bC`<KDQ>~hg3Q7ft{y2-vNT^GVFtQ+
z-~jc=rl3(Go>l`dXKMlM%sC>N4@5z?Qs{nGo@Yx3NN?<1C>Pb2wVPE}UV;#ptcSlf
z$^h~f;mTHNG16X(45Rg~cDJxkFz*LthSqL8*n1NGRCn3R7eg@vlYK6sSJI!k(Dbeb
z*~9pKKS?*@_R0C}9`K-HxxR)Vy||2RfR`#cT`qAhno+jMV1W98!&ZbYSl1toZbFIO
zEw|kM@xv_Uf7xUETUh-d{jgiGB+Ml<)b`FVRwe0BNcpLwbzS!j_@xVfBM1tI#>noP
zzU>7jdeG=p0z2>{7VcV3(y2P<e*Id%P}sktgJ^DYZGJjG{X#(wmS%6PP^}rNXp=D|
z2bM<#a=Yb5s;14C50){Kw$p*)k@eY^Dxw3n3_YZTEGk-?u1;$hy(r$v^UKYq?=-8I
zxA0*awEcT?8f)-})RujCXf@59h_tfW`QJYxC#(EqO5txO(S`a|KYD#PS^`i;_J*Lr
znl(Mf311r)gKofiI;Z#jELPzO$j%D|d|Vo;zAq<~@Ih=!NB_qHz&pQhL(3pDH@Wdq
z{n7kBk0no%5lp=9@;*<Y8^Eb6OVP6#7aV8lu=Ho&3`I6X)kngZSW5E{zLM0HuB5Sf
zb}MLg?Gxuwa@tVvsO8xFO6aVaO$>d%H253n^ybS^p9;5Q+7XijL8>n}20u=GXI&4`
zHuasivO7>MZ&7@g6qL{*9Ae6Z8OQ=yVYdtkcV^$1?3hCl_ob(M4<ke`Rj>bQvd6PM
z6mIbirTpb{64em8WL|wba6=;4HkE~?wZZ^H8Q(3@x-QMOEq6^RJovz4{`B4F_vQB<
z?Yi<a^@r~2I>Tu=C?~PJ6e34Syz1IBEq2Gm9!2Itb!klq7R*BwH>%UoOH6LLuZh`q
z_>YgHH@D14u&@f|p57v8X$^Q*av;J1R;Ibip(>CW{3Q#pe`jlPCYsDDaiQ8%R1T$2
zi^v!#MjTR%SfbNMp0E`s&XgBZyi)P}KIRSdC~pfxs>m7DxsP~$mLe9KRO_unBjVvs
z3&>(H#iqbcB`TGus>*eO+hJ&Jvdm>P^*z?pQlqR3->$qn7QbdmMOUj`wx3!uF<zPB
z_VqfL9%c_m-t|lcKOv#yB4nD(x6j7S%|;1*RrWgQ$82=5$yhu5uzM!ByL0v~nMA|~
zgrPTG(UZO^6AIPN4j28*{i>}X=%fHpcIWr+G6pU&DR1w72)~o8(+A*4Q%{!9nsQPi
zU|l_N>GD+M1+){DNU_&eXF2({&DzCgCYT#}BoDKO!a#SQmfE`$3bjuP+GXggnWjdH
zv}&3dB`qbpa@O6@L-CBW1Y)R=6@AW1voD&ZeBH^7IQ4~55XHj6!I~okvbXxjiz|~$
zxoE+UyeO%*+5gq%?Wjg=6D54l49dQw);^JlD$AslXp>lZ=?*7e3pu#8Ny^k0MTxc7
ztA}OoR%}0vmUT#_LQb0AZ1GCp@@+$i{cNEll42IWaR0x|m8em4xaJZ2)(K^f`pP}0
zctLoiCc(Slgx9j_HXIo)DtzxsC0TUGrn7`4B0WK=o#7|q8^viv-hrr|B4&%2%)6+G
zLg*}>A?-Nh6avlA)BB8&udrt68pEWP{_9229xbgyYjs7)o*-tnRk<g-?!Zz5HU#XR
z7waL?8YZpw*I=rDWeZwQ-&7cR157ObUTbd`suZ}TiY|QkcJ~+4^>h%PC5Jjjp2A|V
zO;Js|%k;7kKx|H^P6txIcQ?XZ)zmeCnF4oIW<jI^Saw&OG_n{=i=%S!qfY>4ZOc#y
zDQUA5n+$_qpGih4u*ZZ>GN?NJat@TQ(|PCSB+|!IQEDBKxS)y9H^7iygm=dDKMb@W
zfBzHsV0=>7+*~BlO$kEcPVP?qA5TuRk=%>C$Jwwo_ItJa8)gQ{eeu&zgQ!7)fk0(`
zGbT9GhO6)t6I5bI9?CtL@P{Ud{kiv_@$Zizv86Fxy1fs^1t%YvJyPjl&xacb!%?3_
z2M!0q7bnMp-j_%lC7gVP0RnBi(85RwR=W7F%|@3?zXQVXMtGgw4WT$*fXG`p7%K2C
zVjpFXm)-r=dn9dz2T#1aCc?Sgjy|sy`K8}beLpJOv<V>Bh{cvl-ZitT_qt6#&#oWy
z&NFkq*88|mIqXCL1mBEi6)>_1kPXfjoqDO8f^6Tuxk|rfst*O;k24HY9h+NEQ3Yr=
zk?`{ULGAC<ERe?jQzO<_0;KSNRoAo>mPqjpqboB|`cQ@APtN6n(*7)tGpU(BUE0bk
z#OyjBD$$Tc&$MimGnFYtwa(BTSTgP83L`l1;RgdzH^sEzHbB*#+|J9M3pp0gkY73T
zI1JNH9xof)HTQ9>wY*qY5O;MhYDNkJM~`<UFz&q{DTN)$kd`RYD>;d;zBcW;pGpZ^
z4+oJP$)U3J`P*dT*&c(#^Y${_nX-u=YN+9IQj=pt#o600YonrfG$S9$J-5RQ@K;Jc
zw&Sz)n9`Cg<dZNDvY-V?gyTm`f3?<Ah9Bsp;P*EB{g2BjoIAvcSNY!^`lji`X&xc;
z#z1gePWxL<!s4UveS}K}D#Z~tFp7$pip1fN_?5_;RWXmer9>_dl_l*NZ(TToZCXG$
zo_C-X;pF}ckrZ&+xoN$z$zGS}f|)C+>QVxwGH#2)br{T<R_Q_J`l~Ps*-i3vTLB?(
zc$bs%4w9-Q^7y##J1wPUZRotR8zi^fDkFC|Qb^*ill?YzGod{od5_mEL?Aq*$vtH_
zGH4JO23*ta*Mz=|NgfhWT6N8{Og0(yQX~>e7p?nzni;ZjEv_7UbIrJ~678obF@22E
z(&;_)r&{L5=l8zTho^eeMG;pmS&23X2r<|tJD({vPs4`An{x<&T#=dM6@dV^u4Tag
zT62o8?9C+F!MX{PW5?u=<T`nP{Z}7;_g+~?^lWAdzq<*iDCgi6o*g0R{ZUa$i*~VT
z8`uDJA+n7`6XP0kf*H!)r9R_L2eF3Cuhf{6p&+W=*eDxkAADx#WWOK*^nw-)819i`
zTrXD0#xqJJWawMEm6HTIi`*u~*`>B5%vNoZlRF?0IyeG$1O9HbRTnR+EWg8w=Amt_
zQ)H(_aU_&Zpq>_g#?sPl{#0ko`ddjK&TIT!+3zlADP3QHdfx71*lLT@ym~U<%lrw?
zvRW?x>tFD+mQV&8RLwr}b=|qgqS6D90Wz4{@`Pc*7j;1$%J-F;a<=8Szk%ZK?bK^T
zy=ifLDVI#Q)+TFu25RS<Hqb=;@Xe`h1}OtdwV;pmWFgcYY{H;|IER<^PFCWv(}~tf
znHx>2si6P_40EWm8=xWvn>Wfw*l^+j90qMI$;fg>?5mYx*gV)Ed^^~o5H|6!Afk>B
z&mfQak@`_P#PPK;$&TEY?<n6r96%J(+Uj2;(UQs+0tSD*=`XIUP=x}nn)|vx_sS2-
zX$MWMU8<C~nn8m+OFSl)EC&52A4f0$qUs-g+?dS2!Tp}*;WLkdHIp>x4S@_tMXmZg
zP(3Yeez3dFwg%pb-NI=B&ULtuvy%~kFh^V*72-Tdy{#Wi^pGCPu1~duhXbGkuJkG)
zs8e;1aoC%*I9&f5Er2=2=sb%;^)H^EBOL%pJI=sE%a`HRm)oE3A?N@SR+cu=gd5=s
z4tt1$<lX-HlGPc>?Sz5V$uw-<D!P9+rH(^|7;S1n-X9;<4xqv`zGoz@k83H1rv*P)
zmv!F(@lWk%fBs<JT4=tXPXP9dHUx;}Bz{*<ksT=1I-|@7OpS;Tt&O^XD?C{Y+jHz0
ziHj&$+4~ARBa@<ZQ-e-aImz$H(Z9^T^)#(S`7)yD2UHo5y!%?6umjyHb5Yn6<68yp
zl-_<E2x4szZUw-gY?!eh9krIxfgw*yJZ=0%%WYt6i#Byyq!Ga-6LEk`<!43!4=Eh2
zJYiQ86NJ|Kwx*%i3kS=m2nIB!h5AiYK~B)a!L;GX99<;9MT#f<_&as4OOfIgk&56T
zx<R#};o2evJvkEAw%p}-h9%kG0l7K2>TF=MuX$z)h5u;1Wc!$b_jY%fofqXO8(9|u
z+P^BrN($!ZaA}(owxLho<Zt^%_>sz0e&!q)s0NHxBem~%A?Uq8Jw-+~b)4TBD(&{h
zkluQ(9UoC`l7D*z6Qgm6hyEN*7v=W)H_wJxxZuh41YNz#VRN&wl(za=)9wyDSw>xU
z^t9-}U4=WVf!5hS(kjye3#vb`)-@nZE|hKvVdopw^I8dWIBP@N@fS*CAUoRUJK{Z=
z6`7A}PN3f@ou2ZM+77uerRdh8-H?sIOcNm2M$wFx>yP)Ge57KKxAO_g(GHG7<QAUQ
z;E+{T!{5|<ch9cHZV!gAVDx@=FEPk2l`QFT$^CZFGrN8!p-;ciG_-fD{h9~EF3}p-
z0R%<0ZL*4_@zTOrsAAu1MemfTpa)=;HWMeLs~3+msMtH~XBmu%HO&QY5p00k!=)sj
zw#04Qh~hLyLY(F_EJ|EjKxJd}6@anA2IrLjwnUch58d?SUzW$4!FazDQjM6tqOPf>
z=F7-tDw~C;W4;jAfC3{L0EHBX^ShSdk6-z+LRob~;-H`wr90;8&z6u>KV)EdCvG(%
znMf(^#3qbM9xuE-HY(<0zLlFUwv=#AinKj=m))(K8kU?qqw;M|$O+FV!^8GfLz3}$
z4Hf~yu5vrH$c8VM({^xGlrH>ey%}M^tSR;PQ)M1urpQQQcI`Wd-0(C<q=i_9yxT#_
zxb;okl_=a+N6A80%ZHf-J+vL`L#twDh)A`&((Ac2?0~P%=pX#u8E`Y;oECVAZck$`
z4Lu9p`C<r=AC3VrJ)}--sLyGZ$0BP%4vY?IH`~$grA^ehb|2Q`Ec^^%vgOY_F!5}&
zeS5H1e}9ZzpK1RAs}5w$3jtH@-u|?0j60TwdLj34;f=y79lN1vOD~cK%j`tQag0C`
zZ;WE%P-32a8>zt^sSw5`Ngwy#CyBTg^uL&=g^V9}tWP2b1mRBI=Cab0J*QLTDex-a
z;8c|>c&D2AR8w~=W}riod_oS6(;8CD(OsikrZj;jon@m#@lJt#TTWUaWDwBPdmv%y
zf>A{W-ZgAjGdoS8ZwHPIucJda7hH?7<suAz^(jzl^U-C?(tvD$6I9~9!v&omp;MW&
zK=etu`0CbT3zM7UvvV5DphK1YVD4>hr1_G;yezHC&~j?MGLs?zZF%&1KE4;Dan{oa
zX0U5<pZFcY{jH(A*oYYlQRfD_HpYUpUwlVRhaxS?ZQ$4^ylD&V2{?__?|#OHt=V&@
z{rV0Z95J3C_mFL1QDu2pJ@Z;QUU72sw58q=36q9g$*fWppeq#^S!ORZ(vIMh^z$~Q
z-kvtb<-c>ytx*Pq2G{R`eW&!w;8H;mO{iA`aDENml&@nM0=H=Jn#=P((ltf&Nsqi<
zUpS3uEN2@JcF-(+Mn|CIKbbMrZ$Imv41!bOLBt1%g!WEHLQC*;!Y+s*!%IK=TzP-{
zmRf7`JZjT1V5s2cL|NLMT~p6ju*@Out)B8@G0_3jDAJl*&u%>lVBD7RDcMO~o~ZA&
zNb=Ja%JgmO(6^RvZvSE?pIy1RdHHz|S7nhE(;II1aS6j8XMay)X1;JhZErxdHE<T~
zojM8$2tRS-MaqR0v+R8#>)<6ME}u~1bQZB6ra1pDER8#Z3nJ?R<@w)6%5J(p8*7xh
zU#F^_a>@_qHLXO#vWHutY%(N~_LUm+k!UBWRA&vIkyM!S(}xyo#9HCd5E}Zq2u~{N
z0YEndJY>67TNG(n?ztuGtA9?UX^w3$V$<f%jPfWr+0*Xi5)WI#tw9e4%Gwn+Jxg?~
zAOPE;vk?I}HQ=BS-5-i$vkSJA!4qR^-(gHc3aa&_6Z*fNT6$pz(-o6{)2q7b-Tx{H
zuQx*H7^9EHTxBd4DwhxeU(otwc>3}XZqy$V#2ccaTejIyqVH)BvhSQ0vED%%(T9J^
zaRv9fqaTc%GMXWT8_w)$J8gGo;H8tk(WWgl+3qC)F@=6E$FIglA97T}Xw6F&vaa83
zD(Itz`<og*6q<^J7Ib-td>ACBMXVfuNU<P;)|J)46RVSa`4Y@u;Qfn1x#gSQMrVCR
zEw@w6`r)3zHyAsZ-=!+$n_|m8iuO9zbC+_YFr)F@jUnB=)jac?czXJa@igse+$H03
zd7r;0f|PV6mYfkV4q!DVEtjT79_XfcMsOm(GOYia7K!;Bvf5ta5G|XFfFXWyevT+U
z&DAtk+GD<zORh(gazq{fKASg>FtlpysS4Mq9apMM;}Rg8deKZ(Z#ZVB15g8nl0(+p
ze`uP<%x?f|&Q8--YX%qz)eX<Jnlo&`Y0Y4TSg1HLb8<4Leb^2&B7XDVL_DbHd}ThW
zI5%G-x?--{L+~E-c&DwGEE){BCwGhNiCbf-gr7c()KYZL8^X1R{u*&Tn@$T#a|QoD
zn!duVsRrzu6cMCG!%%WaBV7_AA~8X_yPMG{-Q6H;q;z+T?(QyWq(P(v<UP;(eeWN@
zu8W<WJAQRD>wK&o4Yz02Z$Nij2a)8Om@+mW1uI1DmvL;XD;}W~ef7O$D&Te5(Wdo6
zyn(1*`t6Xk)zQ|O991QP{wTQst4X}c1u&av4K<PRh$%tYQc3^|`^DfG@}l`w2YScj
zUIzaikbsM&2GJq1WvGAcYs|60WF`E+R{xdH@E(P6q0VVmz(Y>FrbM*E(DysX{@**m
zf}&|+*Ooo==6OUxUhbiKrj+ljx6&^Ffl^lHYpseazi)v)S4GUCaf=a$))WK+f?gez
z6E8P8C~wK?G>r_?VNiDNp`zJXiZdOGPN+t(4j||&vPNc+$%h16iVRkZw0(x{y+eYN
zK-h^4me<t#KTi2R7;|jXh)CAfC(u~7E>aigcG{aQrH0d+x0}~`1*{iJnnF>^88<$B
zs4p`r2V7>kqB<XOm2xXJhOws@WJ|!Oh|FXVht{)%31Me|i$IE})ch#QeymI{&+`%I
zx<-sIZ2pK;D<K5Lj`N>1WIDC;q|ry<TtnC))z;UO^MqS1=(Il|!{JRzmlh6;rEmgI
zRaLv0QKKb`Jv2p=cKt!@50K1mt~a@&3r^e{Bqxi)f9*_lvV^46H#Aqdr$hIFJ8?`J
z!0tHzEsGfV+qS%l7EnP=f6gOdPZwPgUF6}wfMx%|kqZ0HzVYy{MJo-;xnHHMmYtZ+
zRtr*^>jl_H1|1Fw`kKA2{#b@pIR=Nzc6Navgjp_{bHcQ<PjdlCzGb<Ht2jP4%0lO3
zDV;E393X<Ta_xXvK~bHKP?Q0NmyYwz?+x;PcR*|Kp$4zfWyfh?7+bG0Lsufu8L15a
z`=RhVXbrFFI<TozlHN2C;rDK3EI=htLT)GL7sGj{bzZ4Xhx*@x$%+&++&*c*?Y12#
z^ZNJLdn*mb0x^sRkaW{L76pPpm3T}8OQ;*>di2LRb|~B&Biw*M>g$rQ-`=%E$CB9!
zw&;MHspp@71a_Rze;nSmaRkOJLg-4+=j&PWEsgH)g^c1ssBMiUK)b;-4*72_d}#(m
z<tLu52z!jPq&K`J;5nGwt$Ib&`numCg{VZhYr7Vhb8H}z^DgJ__=A`S+>S5)L;#q(
zNyB)O*vIWL?+_(*z6VD)lI`%KdZ$_bnvEg<J1I<2U{E@xejsfetXK~*E)n*Jt}^!6
zCt3l~EVRySK(vsfRo>#o)uhF~QU@XFVUmc^4&|+gtlM&YL-#nd5Fw$2UrM3^5D_rA
zcc#m}V4*r$iL)oD!<l(Lza!1nfY)qv%)@B9hKCxB5d+5UArFedcLr3CFTdVa^LThJ
zrLEFoDm|!pA`!X>o=<rSE+TOt1Y?y&=?jmPB~lPNyj$zQpZ1{o{TbOeRL!t-N$Z5~
zYDbeT;tnMh^sH}EtWQ(Vgu_0!Ua9|}54f`BtG&h~3Hy{qA%O3rH6^jZ>CHILMG?#b
z@gfPsPD_*(F8cUOa%rT6t|?A5q1f8S;`mg(r?k8Zlw=E<=9N0n05VT15gGOW+JNkx
zq$22MDDMxq#EM1X#nVM$!IWac7^A)ap2e*7<^XaH?>P923yzkVHk%zC#5NRzb3=<1
zZDbZ0#;epj0qaU2;5%f`E<K_7&xs%Vuo2a_$I@~tA3?sUWf|o1!D=$hnyLY>@Xddg
zvw4@mj1ZeV(Z}16M`M12F6p<2+Et4KC)Y%O{PxsuTNDeLoAi4DPcI;d0(cRyJC?c}
zkzh#IzbJ0H=^->AkL@SX0~DQ`YAD<%7AyOpH3u$~_Q{q`hj_(kCDplmSL-DUL2bza
zf!&APnMDZ^%Aa{Mrnv!}IsgENhO#jn!o4do@4{7d=dt2%Xa=>MhlytZJ5KuN+T>rn
zumEV%K_B^mdiI3nwvSHDVE8Ek^B<JiA=5PTB7WuXTBe4PZGjaa-d>Fj%_CvYl;FX)
zc*Eg_w%{byc)bGcNxmNZ>STLzzBYiT2a$ML|BQCF{Wv3Bri+gs&C~dXK!1*Gg%E84
z*wCj6;7WV}yEPd1Z#9yVQ{<qP=kK9Odc?kn0AYeayxyI$i%N+i_;V+4E+YQxNhimr
zRXLqdmC#nOkbYf`N!1r_;b%TX4~;Jjgv0{-5`ri$H-1?Evjw|c{%Y`&5@77|!Qlxp
zI$v>k7xVPD-z1~`-z$B0lrurz5I`cPWpK0o`415PqguFQ8AASQ^!8&%Tp4hv?EZ+C
z1CJ{SWqXJj0MeCQjXiHx`kb`@z8TQarU-){<n5xq9lN|hQ>&k>XEV_^Dc%1jtRWL*
z;$+GNXl`a*mCB~lY@^|h!%I~@1$~1e>tD;EXN+1iM`1fnp#&T1p&1I&Rcsvk68Nb$
zdoHBf5AM$A1^#Y;PY6)_q5|4KTIaI-uoXGLQJhy|8_AvP*G4g!4Og-(tPvRwr%gn8
zuHC0;H$9h+A-_%g3kIp&aL@WW_LsVrUFxd|URRGKrg6j`<(c`wkF3NjJm)MchrF~A
zjVcim*j@O>lqLHGPK$cK=)+n{d5=GQaaf@qy1{o@q(>kHge_&1n_Qmdl1{x|f1k-$
zfZ_F>SN%&%A>0t1hOtB|@yZ%_$w%DaCw_=K)nH31^&0ooWYF~oU)Dlb21F7l9WTZS
z@4HP1xq*Q~+4qDK1jB*O+X4RJ7I7Uc_Q@86i1OviUxW1iL~cO;B><$_Q}l_%D$K4W
z`XnOSfS58MelL(cv$83&tT98Wtc+yed;Sd!@ZTk=<$<JrTFAN8M*<;h@zOuc_G@$b
zUuhgX3o+`fRoEB~SEqQv)BF~AUcY$kUSH#tehaG3NL|9?#B`K8S5In|{BpztS(RmZ
zaU|xfx&6ojVSXd|M4FGuy^R?_VX2|!+KV|ZEMd_i2wlN%!aMo6#PLk=Z*XEn;jEd?
zfRZShzgd;iaTJ?f&=2bjS+q<u!rdidWAA*nO+eRG05by^V3lJ2CtJ!R?1++L(!)QU
zGT|~yfxn0FtxO4GWIU?>(~GKAi>&E5=gt6^79+vsg!W6|FYlof(16Yw^@7g=uWf_@
zd3y@02%%K6<HG1ZUy*Cq5&54l0?8n~I@cybvjr-*Jr$B%JJZs)<k?3EQt#@P{2Ad<
zj*r^`W-j!PBn8n_bnHJG?>&^gE%+U1h#!`PEpzBy@3<66IMxoUlwBg>T*AAZ5~B#X
ztZjUGASu0lqBcBX2iCCB1_-VIr!4?K9VCC%xyCb&Aq2;KB-o9{eFjW+zCgm#zi~H_
zX?6E8)#gdRj{-WO1t5Sy17uN+yAiYNAD6EGA%}ub;`rL~A4ew`;>{<q{mVeCo3GCK
zbBa=grXD=Uq%l<0)-ELjVmR`#I$oR3)GO!q4+>kKEOKCZ#h{YcIfxIdlbAO6i(xvB
zi|5?ouF(NboG#6`)?l}QfE_L%+!RfHMWnzkF|xumnk`d(J2o1dJ|io#CYP4U5e%}(
z4GobJX@~(*p*}#0iYO~cN++??`7v9<naQG@yCc;m?WClN;EvC}?O7>RK3)n7Tbe59
z4N+dV`D=37a5Ii;chT*(81ihRWzHK_E_GXhX)APWRY`u(w=y!GkRIzDmg}}69&#dN
zzD7@DL>aHq7=9cIw+~`@hVAX1NyDWU-Q+ewC5g|jkM<yp${7DekHz>6(Ypl1qhZ^Y
ze%u$$8Q~V*@mXv@XsJG+Tl`dJRb^z5gf;LM_znk5p2WQ!XxIm|etP=a9F$b7pSIhi
zvdK#&FzPf^S9d)Nh}~|*jsx_`@HEehR!(&E&7i7L5%l*7Ll|4Q=rIt}XU<SVDgECG
z*w{fqv_W%5JTjUoj9LwD2ZyKfXL?8XALmr`+B9y;*DHLoU@o2CkoWT^6<&v)?GraQ
zk9yj@kW5bs1Fzz{)#SO>5V*ZDU}7F{B9D8oh5)74qJs!C#39m|HlFT9U)`hKW2)D8
zEI3R;jqG2AHGvGy(0=@aTSDQM3_tK#Y0GT0Tcf~%?qxAS){P>jyyNg9gNp6G8nzoR
z(`|{kE$MYG+Q<Ob(Hkjuoa|rc++}+M^R4#ThG}B-b+bfzD-;p~X%WHb?$Fs67BO<h
z;V3;mXPVOe^+}IU8kdwi!ujv(hlSs!(+uwK2ty7{dFpBcN8P#?a40JT$D4sGhDXmM
z|8)o1bnQuxCAq5V43hm7g5{2tOScK}plP&J_w|;&5(IYy40lpR%W4kuahwcW3Jcn{
z(nI+I>=_Tom(mv}7jZFs)V>#l$9_9DX}Ui}I{o>j+<H~+&{uD>VEyiC76>EUiT~(0
z!rr*iet+6yA%4G<jURqqmA<g7KaEOUOM`NInzM#y>K{4?BN<vQS@IC5S@y05ON`WT
zoz1Q|Fr*cOJjgcmkuE%2xH@&Vrj6bBCs0?5s=vM2Jdm{H#wn=Z;AL1CS^}61@*oH)
zIZLZdt!MlC7~>?+xbtd!Y}=Q*!!K3+!Tr2x<2e1;v6~C4XRa^Y3rtH%xYaz*1L5td
zxcu|H`Mq3`{T{}O<XUE!gj2snrQje_Ag#YULjGTG_}_&?N6l_Is9l<Ya-<ZAQ?BN`
z@=wxnSUu)1$*g=Xzegbo(oR@kpIsi8l6ccNHbf<G=@>$})9peOzpuk`WSEO@leKCG
z{F9w(5fxU7A2oO6h2LO~xv)j;!YMUHro8M@w-bJ~5TJUmewL+cU?kQk(i|&w4U`)}
z>T_uyD5$;^faQ)~t;hjB;U^6kWkrE=945{7)VS**%JJxrm&vZY1Zr)wFMDV2HQ;d=
z?d4k-Bfb=BreUH6Z?insVz`gNG0jh6B$Jw=l=0Om3zYx#D8}qHeky1+8l=JKifZuY
z^h7bM{!a_QHcqkl^~b^jwIso<+SS2lUrOJn8{gQD+sKa2na|?U2M1QICw2~jT4n)h
z(8B*xnz#Q`{{zZ@`mYb$uLgiC3Bi>>?z!YE(#@*9Rj!rjh57Xu)J{(70a{9;Qb9<1
zM{$rCtpcm(ORGY04}Ae`ePS21GmJ1y?Gh*sTSt)IzI4WjaNT$q=-A!14%z`vrUpSM
zas05GT>@6H7Zs7~bdozl&#5(Np;gBy4VPb6W(hLAHlF+3%-z#EcJeG}Tm<uIe0wz#
ze1$8W33FewZYI*Mrs2WJbiWT0kXLVnaIN?Cc}Q1Kp$&I)K}<GF{1fa6DhqvPQw66C
zl?BRVClyM~AaPVlhfBNVj9)y5tcB@}?Qn#32Fpk&k-}dv?-vK?Yji4CnQ>?F-S&d?
zwCN(zuA>+AysvRT2>fZ6oG$4p{Kb3pavx*BDJ-Zrq9!KFx~@&tfyF$Qo=w9l&(d|7
zBXwm!!<jOqo2bCph|L^l5*JkJS$-Ph$iSd_5|pTn?hFl!9!M=Kpx?n>zC_oN)mPkg
zWc=QvpnGsj{Tchhbs%4+>4x}~$bNK->mM7fkFr&@9?RVB2*us=;w*g(!^lpw3E)1&
z0G?fsl4iHVPfyUh>%Tp>gO8QZj~^)17&Mt%(rx1B{xl@HprD9=`Im>HawwQQD#+o-
zdZD)FXdNZ*fWt7`7Xz;b0ILT<w9Li&O*<JGhi+%>#5<jf)yi)tHXEKa{~3a4N8ZLz
z3k3idDad{AQhh1@7!l%?VwTgpsW%APxZj1nGR_$WJI`@t9IB%xX=5+tIx|&t)_IAD
zEr_5R3%HEcrKkU?rE)|fU4MwE&Z91Fb;=z6DXmV>pv7se{PbB050fK?H+S)8+jaAV
zldjszv6p^eXR>gPM@)~ZLRaYNQmjZ}sH2tGk@vvHT`+^&OPRHBZk*3-S!{eP;3A;5
zuzx11QFG=n-+D2R<2B|Cm;WS!o4DUpl2f>@NAPniDAvZcm%<^HQeoIy*#%1cr2lE7
zE5~Us=!?aaP_8V2C!Pjk9&u&3@NOk|>>wnuvAS#7+dtH5sms{Ne_%~V<{W}g-Q_>k
z8?r7StL=L0GHf-EI~AuyD1n&&LnL4(Vh9#V`L@{dsB~<r`T_oY#Y6yxEZbAVeM76z
z-(6+sPOEKIMQADlVHt>zB8`gc>5uyR9W;jIC31cqxJkfKVd3N?cGTd1$}kmY8JWs{
z`e5OTj<#Q^qtoWpqR#Y5oM6jS(^4#~k><(Sl_JD!bG@7%s1UU6O-o(;GJ-Mhb&0rh
zG4@+1A{Wo*e!WYZib@%lepGUSVqE&9YiS42R1p3^E<IO;%jSP+ayy~KqY?e}E=881
z>f5~={N1;ftvicxc$YsqR;Ty7+dwS*E#QQT+E4NC3R}SGPsXRcy}f`35I<ju5nU+-
z_{~w6Je|+uhI;Ge;%UQEg^V;K;rANdFmtkfiQ?qAN6uPyEHg$sEZf(S^0_uxGa2kh
zj&4TF2%&g&*D%qA__Z+Z5B<~6bS6k#@^+90V^*%^t@^{|U5HMd4$WBWhNb#|kRa-k
zmJx2SMwJEgvzd}Lr=F2|RlM2*Q6NCZ<rP)jdpCZG?<fuNnLBh7^`dvf2QtBfyeG$$
z`lVl1?6pZN7+{O<3_s7WNhmn%Nk{spMfaCrTSyw59y0%W_V3^Bksvam<lc)(*sD?`
za~rOgzHFicEq#b;4Z@(&0lnZT<2ya<4r;lOUG#Ji%N`Bix*P7VHkbdqtXOIW{!plO
z#ui6MTW!53jdWw7cpoF!<?PG<LT6Uvh@YRh!QFD{wk=(Qn>vMyq<&(Jb%<^LH>C`H
zJ(<3ote>?u$xm<o>{0B@?bjE&Yop=TXl1V&Aaovke_2#}dQZE?MdBj5O3TCVh03Q%
zG;oHNdP^9ge;l+zAKle#E}E4&$}Os4uLFJx)`9)R+N<wnYcm4IxG0^Z|315HDKoFL
z6J=>2dO0q%*=F#D>R5G#6N3euKjYTsM5o+BPZ&F}mX$0nP7|9d3OxOicVsB<F27G~
z6c=W`_+{xJI&9OwGG@~d?YeWrOGirN)Mh)quB{c?mSL-%awKEeFa|gOqfmM0oF>Rk
zjMt1KUcV%v(CAa&7HBM2@4hHJwdKG@w;5Wi-M4W@Hpg|5>FYzYf<@<X#(INRg#h`;
z38Q#p9sc?9_`ziIm#x$41`<E=Yp}d|oBxsZVy&PhvEvm!=Y%AY$Lc%2op=3i6sa&4
zb7_%P`?sX_ELv^U@iuyJ!{R%A&5>3Ma}BN%S;n7Ny2AK#T-9UP_X*&A$#7Ueva1II
z<jNG9Vs@vU9s8jhc(rSx<XKi)8$x)#H`A^UGr^d$qM<)K&J4#kq!)yg&5bmq+wBGD
z#3M8G0+qks{3ddlDbt4j*>Zl7I$wKASM>|dl|wNWIk@X#@Us*Nmx`oIKZ0A<vgT0a
z(&k8IE7*h}OP4r|X;CAXF-Rkp28B=;wOHKbl|VfaLE)O@)PDP9i73b33hb?BDx+^K
zwNp;u&j$hdn3P_liHw8d^JY(mkbTm5R1{>W@ykrGsIq|4a%Ws@6WSW@g~fzkFRKiw
z$>OMpr{VrSigDk&L~VYma!b1UGB8-SQ)OYvg{CnIzv+qnH(bQRzA(c}0uN}N0>IX)
zr8rPEok_6JiF8d<56^yY24*LSTnbcWQ4`*T=6^Yz7|t@_kY-36Dw)9a8=&#V{Vyg!
z_xBkr>t0vlW5?Z_f|Hv)x)wcY#+>euH92<JA}n#(oAg1#K?Jf`{gO4|{z>O$uWciY
zPHU3jOsQ!_Q}tRcOEZ!zRfU<HX}sdcS~%N^<mG!QqWHP!nCr6Z6F9+L#AN<etGfHK
ztRTpG9<PQJb|9MB+#Ye?7)!fJC}Ivj4lgZc`_H!Ybb&0q*iEf)U?x!6@%+W^B8hU0
zt>Bq{W2o8P5ES5`r{CRiKmY0L(R8lOK*wM=7kD3IKgn(OaDH<?oh4;+_NiSuv&WP-
z%BGn5-^D8R`c?h@`RW5@JIoV%1w>d>5NxrpZquG=Hk8EKOQ|JW>t*6({3J_VVzqhf
zyju`xrUe8Gw+C^;n?qFOuLJB?wR+9J!;AxIBl1%g*L7tFrEJ!)rb`eMlkzw}e#2-C
z$HT8<j;IHW_6>XOJ~t`<t}xW_^x<(bX0L2PZ3b$*ssBI~-R2$Uy>7)3yeQ#Q;=)T!
zOU1BqnY>Gu6=UonRtp)f*O5&oUW1L~uNdN)5Fk!Ari<+(o7>ULRAnv&2DLT!gQ8y8
zhLRBUo+QSbq=6Px2cuJ)jQirMFc4W1Kj?J2T8qeO5*;vNqch3z7bH|YQ02pTw;`d>
zQ`QuE_1lY{8BI)n>q5}n+}!c^&ZnDdRhD4jDw@gDzekdZpN$T$pa@|#qLjOC98SZZ
z?caa~5HD*8-1=gCVuC;d|7!p?U^nQW7uZ5Ubp@%5Wp$+1-`l{f5%w4ZLQ_3*Oc@Wj
z!?XgeDx4}NRJ7-ANTgdMm-cY6{o_|81<V>uReW5Zc6B^iO;C3gaJn2x)kV687)2O@
z;$=WVRZ7j<4f>OZErDqx*~(%&0vzm88$zu{`yNBviBnusqM`CfLSIiF0K^Wdk+OPT
zqrU88#{i?#Fxh;pj6LqnaLDW*_h){0(@i3J5ty1y$)v9f1*TECgaHy|$JkfQr|q6(
zNmmtb8j+!~*lWOOZc{A!@1Ps)MM!vIrdVu%(Ut(7LyxIyKkV{|&p-!|tW|Z){9P0N
z6SCFYK83-M+4Gr<#rFvgX+3Wab`APKgZ&Y1`W++?0FxpNP9a*HDTzs&r6~MvkZb0j
zmJT#Zs?(cgs2jeF6MHg_>F@FW4(ot<*|c4qB26d$Z)wb8BsfTu4NHXAU)kh~gnCAN
zxxov1XdT_(envvy+vWEj4=G?O!IpWWl!xCNbJ`wfa}#5gd?dINPS=`x|AY<;FmdWd
zNUnshhSm7d?#r|!*!>dh!6&3Jro6X%hfRfSBfnjLS|_{J5nWZNbsJy3UK^9yuowao
zK?20@YsGtjo2zkS=D>}U&xQjr;HHy?6_!jONuw1t#AqWi>wy7C4*~K-&VxC>qIUc(
z?VtokzVkXTvI`EPkHDCb`h+7fSC#IJhq6BJC1W52ac)8Vv_Y6Gz2iFb^31j&DQEem
zuq@k1w==X{BnKy<`d5E1_SsPiES>e=%lp>$Mboy#t)N45npuO<BoSaqlh)=re379b
zI%sn>5x>8;;OOx+yO$4j5p7XHw`+lu!;rCZtJ9MD2os!|2+$q7ZD+-p<YCw-iPc;*
zt773dYxQIfLUrK%g}1j4#jJvD@b30_k55b=@3z+6*qly-$Shw&n^4oE$+BjgFj+lG
z-lz@u6w>#_hojocOP@znV~m%ZS)cATeFu@4f$I^V6cb4Y-l0THn){C+LFEtdG)0oY
zx2`i%x<q0k!{)Eogj*}mwDzSVkw#XEXT@9!ey%xOtFY*3hj2U>;w2vHK+KI@!21DN
zydkV`{`Y~AFh0(&&$eqOXXxL3nqYC5)Ea5!vFJ^U!Xh<eQH*tsl2+!v8GR&g=5`BH
z1Yd>_ch(z06P#GG)EUh&8$S8f5?XB0nsk3(QGxwTvU<7AznBX`gy-;>WHm)^X)k5w
zl=6B#xTEmiEK4FQITJ399Sn-W4kdP}Ww?>CibJp0dj7!X`EW6C0bE$yTJSkpa0KpL
zZ7<i`-0nccu`~BCfNTKrkFhj;$H#0a$g=I3tbe@imx!~C#cC6eeUYEx?ycM4|4y_8
zz5_fT<At40VfCN5kJV&K8J29=lU<4J7KPUA`-Yf8+TfWPGcTFx@IMe6<}3!bYdOgf
z-8KkAzLeN))0C?eQZZnE3Sv0d9BohsQu`)X&wLpRNg>7WMuau*JD?%PhZV3~<vr<x
zT;XDVQvS6eH-rss{-NfO4d44&JQ3HDL}dzb<ceFLp6*DUsK5Q35U|wn&IEjcOl9l~
zwpt%Or&;CmKhxLhx7P285BE>lPvSg1DI>dYxr9K8(Rrz=O|sL{pHg~OF0%AtC{KIy
zdMJM7;m=ij%9hNlf5DkKkBZtXIRsDMJ??;kz1m2qP~DWUitTmEZC0GrsXJoo*#5a>
z9g*3eLk^2lV5^w>Rosf5EqGuNdHQ1|powQ3+aG<L!$EYUu<5q!-F1}6^+m8PD&0gg
zhn)I+y#=$j`-T3PRYQWqDb4hAEn@42Cw>JHk<#SsP5Yj=kl~R~>*_RiLpm0d8zo%v
z?*{L?xH5(%;xTCk_7~0tILEgDh#Q4D^VOsFeAn=9z$vdl#yCCOZ$`9?U(I;SV$VLy
z^l+6e=cG{x7$O1HG(vn1_ZMu8N@mDnp81DYEx~P3@1^MaW#-L>*qP_&zKIyHNXYg8
zFc7){SEwH53?9`cLnT#>$_-2V+DuVTFv?D0LZxW**CxBJx<2P2Qlu^-Cqit0YiryS
z;jl7WKNAtH@}&AY6}TatWFZ^`bU;x$4Hv<4c^N(~FwXXa_Op>Yw4}JlyT9dKW{Fqz
zcoCp=#JyM0i0)e(&wF^W#ylaB2jsToy1r^pq1`6=6sY-vY$%PNU$#WTlUvoluuS0^
zN3v5c{oJrXyTtJ>L>;#WyWP(^4o!5^8BABGUz#!peO<I(drR;N)Y8_&ZTnf8izb<M
zyvq(6x4&!JaA=`bWYk=~4`*(8v};9go*~TRs4#OJk*<C}A_x5-MO5w06@p`<6~&w&
zQt+0RUUOgC;l-s<4&T1C+}v89W93G}q^mCY2oq>>{?1>th^tJvBFHpKLQ4b!b>@|~
zAE;sn1v^qCH3_&#<JZ75D}^E1l;)pa8}cQwz(Vg7P(?MozuJ;&ZP>Knq=ivSTf{->
zaYQj<N|&aa;lT>1STft>k;8%qLCH|*$Db;u{_X;eY1t*$&4N=L`@Ocpp~ad_rZJoB
z^d~_;PZH5kRBv9WKFZPC7aMTeXx@N>JlEK;$?+pad1p!}xZ1RSi4P@Ne!u$qn0_%f
ztfYyCWjJ);tB?EE+TKb35Y9zFipe}~az%84RD|;6gawFD82bme?EwXDh4gGl)1ns}
zB`qfO!Z{MOR99dkOxpjy&=~CA{C*5h{z9IhX?{dbTk=wCoFTRN(nsWxIaldM+we?f
z#Gf`2&F5di8<cC;lCah`<ib|&mz~=bChRen=Y$@34(*-=hZDzoF<UpK>3!o|Wp#P$
ze7o94-*$R{t?Ows^UzaKg&R{;LtADuxW1Rw0W=~^sM>QvyqU+i(h!7(=VobHESO`Y
z8NZ5oZ7Gd5;iM#*J1&FHU0P3c!k4DDP*YaGkMC7omS@@Q4hM}t7J$BkP#iBEaSg=8
zV_RXd_9W`_>LWtjgvl49S89&b8pcNZLU=d!V7uFYbJ{Nrv>W?+vbERdS&t001W%SF
z?IHgQXMHsTswlb*=F4qm3U5)whculL{4dE)@5?v&kI2$w30z+9U8&N94hIQUs?<J?
zV;7dY3F9lM{)4Zrtf&D{2UsM^5BCk`8r$7yL~cQN(`?p$<}3a)0~yt-`D~I6^I$EG
zR5-%3rzS8e9f88-)IARD8)2f-nW1#Q?K|%D8S<7s9UR{#Vye$i>5PX*+Gt5jcPSp7
zVa#v?ee9*sm_!$s5npwjVySR0Eg^lx4n|&aSO~^X0R#ViuChS{2}HqHvdYLm)L#2^
zYtf-F%S+_M4|3B#AZRM01OIzM3Yjc7hw^U2i$31%8;L+7vV7^Ca>=ust99xJVdTm`
z(is&>jMT*(%#{7C)70cm7Sr}EoY*3-Pn|t821p8#W7Wm9t=kQHw@YFTT=`*AI&e@M
zWtb2FURPB_K<q2r87XxhuL_04V3L}S;AB@9Iv5(tHQs=ccIQYb73}klD@au1{$2C8
z*JwBmO{;Plsvn2}uyqV=qy_+dB2U6M;9NC>3v5TsFRI!Y?e7?gh=4~@b${f+8<hcQ
zcHEDiuYbLCQHh*MaO$CWv9!P3I3CWAVO*dF{-nl9c;xrRDR(VPgMopLHk{FDL_ck%
z0Fg^Q;pFdnG)R?CXHy+*e^wzBkS$)xlFLY0kG47*CBv{F)Fp+Qu@=vo;!2DFp%5fW
z$&z@vXAfGsal+hMts2Wrw*Au}WD8D1Ryj?|9Ay?jO_-jl$)hM-oBI!)JoWX$%ktig
zI9u3rMVA}22R9y@LU~uz?bzsyxM#HYKCmu|)l&JM9qSr_s_Fxt;aM*Fl*bEhF8Ftm
zySqth*BgdtkHQT>H|ie<WjwzXnQJpHAlS}bYoD#2vY;yb>Zhv@uKN1YDL|;F#5I9g
z81kU_8<Ok%X)oAJ>y(fFH+75xZp%P+P{fvW?XJbHhs~k4YignyZrRRbs)t!RL8vfx
z?d!@tK2tuhfu|>)(<j`~my=;)J=|fUp(^rJBM4^lXQq-xp5iT5f~R*+)g=?fSQAd<
zc=Gnmoyn$n!0>dMv-h)<l&}yc$2TRGcjcgkWf+?9*9ltHkgyFs7;~Urj=^#lSlFEq
zvFC!U@4Y#C&H)Q(HW-n|`Ys3{Z|iCV05HZk;u={#v~mo84tTP%GZ^#DxIgaL$qb%;
z{CinKmoDSF?70?Z2KEC-3^AehVLZ&YxDw+cb+BqR`WUP8z+~x+<;!A-CB>u@5Bs_@
z`n<#2UY`B)P;o)zrU@H4Q$@b>BDazlGjJ>^_{lf36?d)f{A5Jl{_&fb!bpI*b{qyO
zu%llEg!I|w9K5c*2#e~|g^UO*80?GHcJRobdr7{R!c$xPH=-;RjR3{RZQL)LgdYa7
zKkLO7)!?Z|(wt(L2U1{06uu*t_+T?C8#p1)^N{MwYpfb?6B1PH<?58Xm{asc!pIK)
zF%o3NX8QXy=F7YjYc&U!et-+q1A!@p#nfNkDLu9e_sy9uN1Spy87txW?;m&UuB4`B
ze<#&%D|h<-UTF|5HY3DXusIC~In#jZ3mm1Oevm@@4}x12<2inI$3J2z-5)@|`0f5T
z5wIzls06;$-_2@bkP;{_(QAm<|C|*8%9-XLt2fL#lcD~;&K-umkda^Z{$*9!c6WN4
zvRY<fPYt1I{Mq2%46rgE9ekZT5??`8hVC>=44d)x1Q}@o&&D>eC>7ugCbKk1rg%r8
z-{@0ZQ95O%`vI)hJncgt(rO=r^{5KzTJ<(upA>Q|ZJ+m}jDF4dwRtmat3;)PCW!r9
z?k&!Ybp;;pa~UyEiVi~iR(_}-O5mVLiNl#OLY!`oDnJ?^KZpC1eTqNAUb}tRb9iO7
z{GJV41%PPwq-(MT7~|HFt{Cq(31Xmi{6>vP*E)hQ3Z-24yDfcIoULeJtnR&251zG-
z`e^opFYoYdWUgq!eWgd8Q0Ys_vrN`^g@lRR6eU(hbIoMz3kM2@k^Ei93fJ#ZctMcX
zlM29~mf*34Wx+~q{pKTe#EoVnCQw9~<u!*&(<X?yUWcCdx6OdXLRJ7US`0C8aW{Nj
z@AHOSW{gXrAArp^4c2P%cGorl97z{y4oq<yp|z}n{Cvc}V(e#U?h4@0+4aUmOW8m5
z-mP)2%VGA0Y&S0mMnBA-ZDAnsF)6}4jo1STKa@Q0FoVTFlL-|2k3H4uFX{6R2%4^a
z9E_BUR=O8BZQD=m{1vw_s*4@(Ei{($TL{SktiQ?zzGS$1yrsJ#BrCda3I)m*_s~zp
z8mgxGpK%*0`9HL68MQNoHtQ`LYeY_VbMt`AX`wC{^}3r2#AmbM&BL^h$gCsB+Tzg0
ziPNc`E}y}?Q6zmN-ywaV%;op<QMaC@yIQzP>8WGtOi>wqDV5@hra~X_!=??s(T$SL
ze(>NWJ5ZLvLLE3~+DY!0%F)xK_>2s%OBZNFbe-t@R8I-r-P?cbjvUC$4SBcOOP6~3
z+Ee!bNQ@Cb+kuwV%p=t*D9-lVrz8|mfRiB0p+CK^{9kr_Mur7r1bI}Hz_t!pM*!Hi
z|7RwG-LrM6CLQ`VS5|YVSaX;}c%E{3WQ~U&AWJVg#&vD()>S9R(v8w2ofe|j*7n`o
zy(uw!dTr3$(yCyA&G1&!zk&pyDr5(PT3e2YIjxH02)a(+QMOU!c)J7R;c93AZf|Q<
zN>>GoUH1|^Pi3DHHpHe*@tLQ@Usl$()Dld3SCSt$*3FOEjV1&MQ<R&0ti6B9FR49T
za%!BPoDs1m&<Y5LgdqIJw030rN1Q4VJmb>&^)+}7E49Xo2KIp>PZ}yhy^K@!7w!ow
z1tI4!0Y#(U;%^4$gi1p-<fgGOG|PW~jqAVrK5WC9s#2^D>&uk@SK$Y?UZ(n%jF((n
zd3~(Nch&&6^OEy6Lvlw@Q(|UQRcp<!)9&k5v4@lkDzdP3HVW4L=4jFK6ita3qF$n~
z=uA^RCx3V0tu;!TUL*He5C?<Fx+d78z9F1(OByBd=>$E##BSFMGiIBMvRL1(huXR@
z^$R2awLTdn1;&z6V{K^1ogoQ;GEQg2q<~{v3)=99=B`Hj%#DizXS?D%Au%o}FGAJT
zsFBdvmGmveoFCiyx0M*K&7?A3>+N@0R`&1`8P7fGaP-E{ec1lnojGY;4axvGuKrtF
z?S_@L@F+>JYma5FcfG8l@lo)&tQZ(>^_|!l&f_rKEd-<kbN}nnR_e|T$@2~wJ!Abv
zJeusgFR0`?@guLZvoqAg1&uV-Uk55rHo0!nPJuU8N5e^ng{_4kHSe4$51qsim9ieE
zy#F5`lZi_pyD6Z3&KMJ54%q@JQ&9T>H)6$o&-x1VS6=v_b8C4ICY3CVJ!3A^KN9Y=
z_pP6b8!|Uc%>BJK(`vubA?A(8$)SNiW`fY<Z$W->p%FydneBo`UDf0lai@ipITv0R
zs*yp4W3!C3NzG?GH5|`-{HO!i&h3?!oS`4W_De#_ydbfU!gmfSPM`nzrOvfRusdM_
z+|H`Ca;ItHj6RUV<X2q;P{<Lx@WXX)k+H67Q~<gL;gpAiFilL#Wgmb0uJQ(py;m0X
za@5&CHtWYX!^=QPYn4^MJFkR|b^rueTXN-mRf^^4w+b^M<nJZVjvyYbKmMpt7r%y^
zI^m+G*2hAACf!s+l4^lc3Oh2$arV(`I4BwDxovx|odOn#0TOFq+1JEp7l1G^VtSnC
zb>^o4$cvzA52L;sHE;b?y77XkB~ldMe>>tkh&{=bT-_;WzdW|0@qXnwFv5u;N`Wu^
zSv~3V-6~9vyYZQM|Gl?7f-wNi>GXu0Hx^=c>G@-pXyH0Yq3BLl>!V9sGuTrVff~2F
z!~3m=HgT+T{Nnd^bIR~$;pNfoOcuQIZ%bVXRt^0alL?zCUnLbID7wq>dAijc@eTi5
z^<+@0bQqr;mNRt!yY_&=yQy@=-MBg2W2@X+lmVx-s1W(yaG=7ryvA7;n!=Czcw*0@
zpZXbEyQ2PzS{_B>6YG3b!Fy-7DX*86mUBb%lllz)1MBgE<8^`sdp1G6&&I;&HXq&B
zj(r9szvyS~Gv-m)ooG(l6V-e>E>pAPGgrC3dVA75&>+>E$5Xiy(AMc%a4(IF^T10t
z+VT>Pz46`ld>EPAP5F5QEXHU9MmovrO}0@MKedp`QfPFvIfrc=z(-DKVUG&s@PyhQ
zz~9XG2)Rp6dAsxFvX|}5QqA+B3fdvzG2k*%h2q3sy~wIT<Yw-=Jy^ma0(7joq}XMH
zbqR|4)u-sruPK{*`r-u@A!PmqrE^R8!D2rXfg~4YJe_*Q#{X#nR<ql^n#{oCLVx^`
z*IRZU-5YqnKMFLSI#2Y8|J{`5TM>KaZ>>%6P0(Bl?W`#X_r~L)NlMesw0;_Iq{iBu
zsxDG{9NX>XEX&8w#a=xP#Xjfe{Nx3gbK6a55vYBfl|M>TGOLK{%q!=ttLB4Q&GsAZ
z(*g3m9hQ6i_Z>3f8OpI;O>k}_;@20%7uvJjVik5~<d{b<9<JUETf?H584Ib50+Iak
zcD`!c56~yh>K^)OXde<Hd`2E-$bqQTu~v@^Fq%D?pTd`x3OE@p-QpOu^!n*`h_VsF
zF;4=a=61*%$L5FD9z_d5{^|+76tZ93hS>mU((U+e{{MRnIz)1F6G72f<<E2)yNgt`
z)uHM<%k7dxiP8*ec0Kv)Wv0bn%Z*bL?j1zdV*lJEf4DX0t6Zz{1mPkPmR<GCjmAs3
zkpajJi%QNA=+(MmbrP$@j?e<yn@G_`$ED*RS;dVdM%{I6KUBxoNLIpvqfef>!H7nH
zWqR<){C#!b5%8X`z2{!#=bN2AqmCWEUDJ5gIfab`uN~xHH$!-5{+y(R3t9alTc{-%
zrlzSk1SMMMgMp8%61F(Oh(Ky5Zv|aG&TC#`0T_qd?VE_)z7F4R?C`ILOP99;uuGE4
z1zZr_#2p^OwpTV8n+ilz5(%2)fdFd@_v=u(mbF_WI?{eQXs~q}0KHMJ4{0Kg{&{2a
zr+6WBv_4yICe6^q63z-O<Mq?mi16DlZB$&zET)8Y2Am{Wr?lF;{V2;rR4IP9rfs4a
zlc;;Y23q!Kh0ZP!3j?4CF9R=FRd4*?lpxKwK6Ri@QK5K?*7b$|dnYahHROen>B;XJ
z))BnUn8AhXD4P6~b*=t;ev9&-*aNSA(YcXC20k|ET$8?j{iP|qD5}-o-lOO3w%bcq
zv2}gCP5cO|?TXKf<KY}ZFRzL62(jh0<`BSuBU>X+#kPNSy}P}P<Ly}0lrn6>Mu&`~
zIve+?EfJS@=OAaAF2S2&S$4mqvi#^Omc17Jn%u1--kSZfl{hdW;5!^Qgt7HxUM~)V
zruJYDGVL>!v3O!M--R<C6ZVQb@`P7?rPvgSf=~#x*a}i^2<&p!uzvhtHTHZ}Xp_mo
zn{=JWsSlj$`?gX0!fVXpNTUN+M31<&3CW|@P6jZ)IA1^P_2ysM4xKg`XfFJ?Q=qX@
zt7M0+hK*xeuw6a-ZGN@2Gs|($?Bi}#?%szTvwe<r^Q!mbsrcgD+HN_v4taCBt-S&t
zJ5I^e_nH{Do;|T$Q@*@P1ZXM_CG&mwOc5L);TA=xSE`yuVAQ88@Am9}?_7;-0@@)O
zp*2!cqJ%Jj?}&(1Kiq|D%XL@zy*H`v3xyiU(?=k@w_Q2YQs&}E!{deTpks06Q_6*@
z?{5QhJ1W__^Ytme=?lcoMZYW^cpML-qt8D)9V-(x?Z;n75MMumh;Ymg+4qCY(*%+K
z0q{-t<S=nrw4Z-o>iiajEdCs3nx%4rju!k53ki}S4Wu$Ff__Bj?2{PX2vN%12H--Z
zYt4y-opH5DnQT+S=vKG(Gw-hrzj2Y{;lr1-ji}01YSJ-65IM31+mG)n-`c0A1S?1&
zr)xR`(c#xU$uaEy`{!9_B4JreLNQA&?Zykk4#0OG#Y_Z<LOwZoGl|ar)hAYFZf?hE
z@o#paS<xndXrA!nxKS5Rnb#!l%T{uYwqSa~L8Eg-gePl1O(kIyyLHx0{-JHMx<tt7
z;-1L+@8%g)Dr!AJ!Sk!hZT*16h?%604TnPU?C3vEKQ2y6O<o-Zs7wQ2pnYSRcM~$$
z$Sov7h!%75LvcnB1|EbgQArU;yu<y4M73!pMP_?T6spjy>N8~|g)&3YxC|wIB~_JQ
zL#iurZ90F7Nc`M!h0*TiXppUiTK$)bdCoY&YHSa4H_VQ!D^Oawk)Cc_*Gkx1Kw}2*
zoPaFRYWmgXF>e(rGsAh<r(B)WDLjCPy_mVHFEGeUC&lv&XjmL~V~w<4sBuWMm%8@o
zbrZ2QC1}qaH(?>QxA|$+QDtLV-{r1_%kskQIU=03DSeo4?e8>x$_(1r8T8M5P{5Cp
z_k8TDLSn~9d|wcw;brrl`#~GejMXRLKsD=|c$+Tfk1Yk?5X3W@T1N~wkB3ip80-ky
zDl$(hmi`dq^y_rrUo@`PPP1c|^aG2VFjjV>r610A8FIv@YRoVWs6ZhT@%sMzf>A7A
z!Uf|mba{F<3D}4S3bHZk#uTd2RbWEz=5(Qt!$oxT#fOY&dDhy%2Uj9pQ5tOO)%SXD
zD%b<j9al_kc2nk_dy^s>Y708V#YtnpgqW??Bh3BuZ=nJduaQaY8^L`Hr)$Dn=Au_<
zw(V!GoK+WfApbN8?u;@uO-sAG)ik@Exe=WM_rJb7jOWqvk%N}ce7?3#lj15-DL=EZ
za<dX!`K`EwRN1YU#2?OS2nQ9IZHB*s_xXBXi+)kWl>weKAU^Pxhl8#ee-8hMY*Xh`
za95&MmI5GDEF(bwxI&K*MnQ^Eg!a3MHr(PZJNp>&#w3+&f2^IW+>JKW^uv7Gj2K`F
z%?=aAi|YTTbKCQ7SXj83mByMN40{nQE$3^VwIa#>py}-1F4IJ!R5sF~&BRq2LgAr>
zrR3$a?{a@uNupYk#28!a6cgOb*OP7g^iKs^F|1nm*VUI^7=p4AzQ=xao+8x;4yCa%
zL(D~I+R`uEM1A;?FfofCy(<0Kn?_H^Va8gku-qnVBY@r90^2IQw7D(B!4VYZ)||33
zp&!gt$RAUh?`nS%>rn>KBawno;Er%OC%paeUM$m1`0FaLqq7ezvdo>jB9>np?SHx}
zKv-!3C3&;?j?Z9;MoX{KtE`V}ah;VBKjL3|jlEI@LoH*y%rQpWmmZ<&En$fL@hG->
zK$ZR^_iS!tFVYaO)M51i0hO5LTY9nxgG&Q^4`(N^cc=*?^4_*)E@F$*vlm-^K0fY!
zoWLaioBYE;OPS9uhbXogZCPgt_1|$=isGDBb_g2qrO%U7E3y8z`xUb1e;$}LWs3rm
zQ=gx;e~Wd{h+NQ9fi}oAsfb);*6Jr@o-ZqC-C7M}k3?a)z>Sn0-$JahE}(tg1Wm&m
z?b*nTs3tS{zK*Qmz<CzJ=aB(ad)S?J`K!Y}44js=3RS`A&P97$!W`j?Jj8uZ-&`za
zI`Xn8VuPIcl6%w&Pen=(Ho!WZu#<eXH9g&Dt}*uJ{q%rnaIWC6Rv<xs&L&EDp;dsI
zE3XE;^Y7n15mW<%6Oc~)yT8C$B+@uP!fJf>_g;5Ds0)qTd*IFFi{pmzVQ!jbo^Yf^
z$VEL4PCGJ0Pjkdjhi)2Xww(tsyN48bk}58XN#pGQHdm?NJ*>@m{5LHaaO&IKzN`Pu
z9U<ht1Apxmlnjsf`Ry+*l!FO%v4N(6l`FMEU!yrerK<~X4H+LUPv<65e_5W=#2|_c
zUjH}$e)C!g(t3xbp#sQifyJd+A%`xT`PAyoGFy$s=6AMVfN0G~SwfRQdmQd4;`%2H
zh9)uR7P-6g=Q@Z5TQn5lPPkIv@a`sD1e)`oT`NgRsqFFtL&@;nWONN4|EJqf7|h=I
zv9BJNqs)x^AT0TSUj(9Q{=U2h4~4*!CUMF!7)Dn}<<Aip(mhI{3iqgah}Qnj8Nr<7
zQfaikDuDycs<;x5D7rwS_=mLbuO=1Q_`_PJx`*F>W6a&ESXU{@6k@@<_Bd|*joNbY
zO+Q@%ebN0-GgGZs8MH8Cwr*e{_4c+PwK_5Hbb@*W7oO?qg_xn9i8fGOkPZ}S2YKKT
zP0*GfX5^M_;}z;&x>>+U00reDA1aFA*$x_?F^68KblDwIhZ_hxTuC}dI%&ijLAYxz
zBpW-il)r>EKy=&R7o47d8x>`)>MBqT7jcFXh@scszLgA^3q1kYNJfos?uLAow6vje
zq%#66Wp3@sTmspPr{(ajs-ProN?tNzR+=f3af_2<&%%QHg3W0^I`fzIeKDG+aKp&N
zszD7q@IfdmG~%xUejTA1AW$sZJH$cPQgqNMnjV+DU$yOL+{PNKCVZyFq!e=uf#n8*
zeAK0!tltFxd`cxlMS-N+8r`&?n4m=95QZEvJTEH?bkGFUOFY>5xup_N9Cg4RNdx@a
zO$FYRFH4!QPnERVEdfZX=8lejQ}uQ6$}_JX_Br|0HZu%>`2u?3<@bPEn0Z!cU&>s_
z{WqmOX_GXQ=`$$T2D{|Z(;H`AhSvbQ>Fe9@zW(Nfi8j<wlXN<h`yqdh*9})%jz1wF
zU|X<vm*7iD%mWkiQIo4P!P8?`9wU3s0R3CRY`UkS|F9o9{XWy^->vpMU*=^R)osSJ
z<1uIwI#X9ITxWXgRcY`~(>tF$NALh53V_2aPVc##THs9#0qADHw=b<qMC-8(L1@X*
z(!Bu6-t)P-lIuCWp<;orpRl(1R%9+3@<c8|5Zml+KD=~+`+j+#{FtDjB3vA|v0D}K
zccBwmEkv3{w5;5~tNL|~3uR-=i<EA{sGw{>v^KE*OLQmcC!w)UXQ->)*WHFQ^QNRx
zV+m64;W-}al&bVf2lVj+^K3V7PC&Kfh{92EztVgzUuy;~-zXnqhm=wgcvj?<)3)JQ
z^7C_;so@UNwi?gkMU16{gN{R9sos$Hy*V1RekM4(9=#?;-Y2Ku<NBF($xYBjh8TO~
z4&X-Ce^RD?D0Hp*CRVmvpP~msId1`P14x?F!l>-XS+y4MO?9;T;-G9ydp*<oto50J
z!7s4wXw0qs6WkIPhQ~_}ts*LOi@rbvZ38_!A1tzXH#1Q#?Qi^x7Dsdr7X)lA74I9s
zRs1m&YG?N~<R~Qh#kBqdVRT0AT{4>LGZ*FZktiE21l})o;6*OD_4nS}LU`UGxgJK`
z2kQcr3@|P>933n=Kys?z=1MG{sZ-9%Wvbd{eUR4)3zMTuP@DJ8z$AkF#a&mn|Bl%I
z<NU(=qp@0|6-B9M{r+KB3+9AVlbv<ZE@r|Y?6$H}Jjxe#RFN>CmA$tAS+DQEqoDhd
zk-##@_+V;JV&HY~HzX(msConw0IW%Tc@R*!ycBt)1<Wrb_TSrL5{>r_F9jU5>Q?CX
zTy)@uS4TBUD?(0A?18>d<z%aGe>ndq&}5%VkIdJ@WhJVklzK&9s)o<<0!@pUi!hm$
z?H^Tsa!rgmp(ozgXunzAXgxo-Emv@rjnNyOPD^R$jt*xhW8n4X2>1U6WrNY3D45&q
zCanq!J;&?p(A7gV4DB;VcY~r_-{~a?<;<P^B}kRltb)k-i8YMczNR;SvZU^~KSU?#
zeq+RQ%2A+Z%GZOCy`&D6m6gz*({l$d2Yq!9ApC$9IxR6MGt1g__)#Wl`pXA*EUn|S
z%g~d>TE&c`15uZm?ikWN0UwIE_nL#m+27|2NU$Ken=M<?U0fH{uiBzI-kIz(*`8Dj
zQJe7684V>0m^AL~2yb}^@$&@*+jg-x`VLu5c_k*E0ly+H1)ST12d}vWn~?A_`zET0
zU4|da2y1?H2IwDJbm!+sm@gxj&;W$-I<|K$*tl8HNQ)Fe1V>Qgq#5V5?-lyG$#Fv+
zeb@u-urYxqf4hpKaXyr#+nH+65_2Yk6Qi!_+TL-XNv^fV{N}PK1eh)AGW*8;hpGNZ
z^F}@WNk_kJ7E4ro1@Ko&Qf@DQdP&0Ef7q^gx$gHia&j8c>U0+Rc;(Q&r>&$)li#my
z>m$l=>$2TnurY123<<we$SE)PY*ekjbx49HxjIr?M5@gJm?FdHMo^_-VTK#)Ts}O4
z`-2<VqgI=0koHiiJqG}e)r6@A&c_BCVK+kYwscZSl^xM|`DR|MW{T!6j=yF~D|m4g
z%i6V^QY@ST_Z}akZ=0F$sReWpxFPQ~he_k+{t!8_P!h#Y;eOuh-P5JU3?IBCe-T=>
zn2nxDrUCzTsesxaMwe`=k*^A-o4EQ`Iy)7H>yN%i?GzABtYWu1^eU2A<j_{#$q|r|
zyvwL;CcJ)Xcf?GcKdG6q!_Lodm_``BrIhL7iabf>19KNQQ<tkR?LIe6MR=~tms&-A
zG!>BbVyZ9%)V$5X^(B=S6=B*OeL>W8`xsL?xPx>6W;-tE3A+Di6A989gObpw7XcW6
zBXy|3x<DX6qZtTQVRqCx^gn8`F>E+Yj5knIHXpL3xQRrtjwGHtyD>&{ir_^9KQusm
z2M(Mgdjw5!^^%r95-K}Rh%m7%nt_uer+k}Rq6#ZHOaMVEpTgHo1uNl;T~oZ2{cm2u
z<=EaIZ{%J&ZHzj<0IEB?S-zNL4I@n#uyZ9&v_{pkCjWG&_Yp(}d=|ZQ4u9P{EWF%i
z^BPC$ss@jJaUA3WY$s0_nU5Ze%`{mCx&w01K1*q-!D1XU<RtFB7Y@`TLY@wI2aFpw
z%-d(TGRyP+6SS6_tumZyi$KcX0`?Kbku}#(Hqpbv$0pEX;c+2at6eRA>8dCeq7-NA
z58B?Vge{s-*^Y%)_lT*m)g%~QB^9uHcQXwgP^vTRxXieQGvZJZec5G_@1Xrn-B0({
zTtxw1)v;a8Z~oe&2Psf0C`gzbzfc-FXdM#%YP@3OzdrMy2cg4vt{spo7J%JpjslX|
ze!V>FlI(I8O=E6KpILpN>deeIrV8KAPgYQ~;6sWp5JMQ^(YV9c0X724nGY~nTof;n
zLv_1by)`Q`NpR4~-t~de!@qBpo$H%|Zg?^mRG|%<dIN6T-~l(2^6bq@sGePMs|jPF
zk!*YwV}%6KZAs*u_}?zAuLt@6sp$JZKxCahlI-`&ga=^=ly^wOnL+yM&Mf~PdKpnV
z5i{?bY|Q_sweOCGt84ovdha5l_Yl#e6O0m3A`v}mw1iQHF^m$O=p;%Ef<zKTO%T0i
zlxVl;21ArYucHL-cE9iQJoi7}de{2C@2pwYa?YGN`|N$rwXglVe!uHt0n8;&dpq`!
zleE2V7X`q}4AmH39uom`RVRO?lX=(9u?*l-%F>{x;gs5f#9W?ixi)$W<=Lz9)Dr_)
zElU1GU@diY(nW;x0wjoH@ujQgrC8L56eSF_hL3TkC5V!K5DQqLcKIx+vc~V+*Pj|t
z{E%;BKCk{2*ufUZJ|N@Ep9O<O5?{&HV~TXsM!C$d+M-bpe7_gLXzONwmsMW_!ij76
zz=Du@#Oa2OeA1ZyOi>884aw~XQ5=oyM&Bx?)hEs=Q;C{HLsKxeTb^uUWpCePYic`)
zV7mWQg?vWL<^68E#iZ5>KWho{uUv2j0iM;O$TMEAh^Zk1WgWAjlhVa7edf=tUFR0v
zS7^-VG5S$2iyzHg;Hc5bBMHT@{sXOr2)2SM>VjGTQ>Q%e{d068e3;N_HR=sX5r&hP
z<YW)iIvV((khO^mLq~XrlwI5NWn@Z0zSriOKmYiF%uJ8`62|T#ynPw`pFbTp(S>(D
zB%;#W1%=tzuZ~OIhct_tC_3G}9@o8jM}(sBI@h<l<!f%W0z3dI(zce}QBAaIt%f~A
zob*Z>i1Js;<Qr4{0sb{HkDO^&zP$x=utH0sh690nKQuE@<S55<3ta-5@<5UTvp+|*
z^zM~(RifE`(=!poZ}iIcxO17h6sjumHqNQGPocP$eq$^`m2Z3EHmz5+q1i73e_NBk
zfuuW90TWMja7Kk}X%Nb2ek<$2062blQ=;}vq%MB3U4d<8%qU&LOjTl76d>$=lbvtm
z`6E%R14xlz+0WM~3aq3FRVkCS-%SBlDJBYsrZ=yAay~|j8n}TpcQfW|8#SX_Y^;V{
z<=F^-neb!1<_&&O0N|QbqMRc0B#LJvu}x?b#nEnRom(R0(X{#!_E~D2R|qg^I<U!>
zVX4NxwN%c6ItUAlKnbDY+)p>pCS%8rl&?Hboea8p$Oq1lQuIfwrz;`nD|K7Vr+*6}
zQ?xnSvZS1Ut6Y334-%zjSBL<$Wa}@QkNZ7{gUqtYYaHD1<>X90r+dVTAc7@gt3__$
zB+?D<U=-^6vZ1-(5<^LIo2O86fE;OS;px5%yi>sO*-1OL&w1_V<?AFLDb;sloh`mb
zbJ!`pu49o5S6KmKCZ(xF(hd1dfIXL}Ph}TZXj>h+Up{En(N}V9=C(ZJkpywuW<JZZ
zWF)zEw^VhB&d2)ctBpB&kd6su#z@yTTv;CAy!ZCO?n5|ewzAtStUE59qq&u<Gxb&<
zF#}|39|AhQyI}NX^A02A4CK=LyuHHSH7AX_T$>@czWy-}v*}JieoL@AQ-6tWLEg`f
zj`6!@E$Shr%c5}r`aq8!?7X>R{Jm%|Q(zo2y;F0zx9N%Iqr0Y$4zTW+7cjSihke>P
zc^9xG-D~$tc0y)QT|)6|_zX02Z+F3hcrQy`dzU77Z6JHQa-6e_lx<E=3Am*<@dTN5
z39b*8JJM!!Hk&_OXMM2%A}i^R+@sh%6|4{Z28+PH@Uao^iwF?*`5k2qD;u?N6w-S#
zS-Kf}$N8hpaj)sNV92iE(GglTk@IX-yKW^HxeX+byg6kyQfG=Sdx{_}Q{CwnsF6go
zTXCCRmI_?~1UMhfShNL6FzBt!nyY$-DzJf1VmAFBY2~0JW=AJXiHn}6KWJ|W!ZxpT
z)7*TNxAAkJlZmo7pZ?t|eK}|jODCAgU|CnF6TS%2V7C!hhLfKIfg1*GR`7LPSppC`
zxnmKC3ub_)+#L*CZgwxMbM6DYR^?FH^{1cSJ!`n;tFraT@aO{=vD1ul2LXWH%|={J
zpV`_NeVjDXcLQZ@fK5QY-|5Z5rsh=|ufhd~`t0AOX-VllkuEH0@Lk_aCYtqKix;X=
z_V<Jc3Ct5tvi3JvF1%O%H3lKv%L3_!tvuqoKSU5hAowEw{ZuBUrqB*AHZ5LBz?}rl
zhJ825Q>S*zY^C`uJ#CFcD9AO*zXD-!^YV@;E2?VFo%_W;22V@Y1b6TX#U7FU;6Q&7
z6{7&Het7T36W<Qst}kg34^O^k*)4s7aa;n%z}=4x;@!#Z&k}OR*;4k&g97JQe%vP1
z8|M`!>wa3l7WP2^vB{nY)EUJEKanaATVr@e8Cx^EtIT17SC|C<mNH_aIlrawYIF5T
z&YmCE65utTAC!~qp7%VVH!ZgoJ(iV~z1oe$F({$2bC@70mB6j@)Tw%nMS`Q++6`}l
zQ9LA8fra|O=kY{ByZghM<o)%T;;r$U@^KGpkiA}PHR-<IwpYG2O>1G$Q(q9xzTTvq
z7g?6N7AaK_Sb5*PuB7E(!uNmr>kM!;z{FpmsXfp@mGH|Pd6dk+b(+70^1k=ZWRG=O
zl4z#E{j<Y3gXQzm+p9zEh+ytm+Cd=cE7a-c4%~ipy50GVMvWPP^7|G6kA5Ir@Z8EA
zX^<|A(NEbg!m`Z5R`d}om6%P3QX(l>(RWAkdAxWMxwZoX`?ESOAT9>$0l{eWG>n}_
zggzZcB&%99tnLShoc9c}My$y@+!Pg+xZ4Q$2gpvRM)jYK+`OfzrcP~lhdD@5@OD}?
zUd2khhTSxO`U79H)lD-tX&{I*P*lUDWlNJ3T-Vd;D6^%RP@4u6Fl9^adW3O*C9*rW
zgu`#<TMpx^-Ubo=;y;@Ng=BAK2r|a+MUH3zzwx@U{y=OL3Y428q2ajMEI+u|8cG0a
z75ZStm4N`EZZT1D8>Whb9sPK(LZn}EZWppFmDjdrjjK-$Nm4m+Ne)q;YxS}kC8A$j
zZGBp5;>{XqcSy?nB0h(JPRX>e+)E}Mr8)A!*+`0x%0a?(9n*+<2o|5#60B$k;Lgm=
z)+pY$4B&n8{x+^_N&0+cFT)D}hG2#U)*@YbFxM;z=Q4dmY%$m3fk{F8Q&B&7#aOB+
z>v-S`2Gdk?pbwZ)igzcV1I{^=gg!J}g#JsMi;kz17z^}JKf@1+OX&34(~%g8pKKBn
z+}MHx@V|nGXo{#UTVImdRf!pSwaD+)SDP?<NJ>TKBaHZv7LN2VOB2$Tvrf06@zdY$
zJMKl}tQ)PjfiyZD@z$xkG6UtrRd_AC?Chk1zFpVaH~g?z^ySGbcQV5sYZ4TLEIA-C
zf!71bD}a^0P*F|#L`HO~J$03s<8rhB1B|nODmzSV`K}BPy)ZEg!DtQx3F`SM=VP0$
z_h6C_IUBww_q8tZt6np7+sA5ZW`NA5aD+r<8$WO-&ZhKukAwq(XQ*<_2kHdV$3l7r
z7j_Vss>GzgpS*IO^LXu8aRZcGD6SdTWM*mEfdpknU#zGcPB<V&vS*EVCuU}5s6_4?
z^9sbW81aD&Up*z#S?k<zD$N~?NT0Z9-}vUKciHngzIqB|tF_sOM12*3SU}`qG2sI%
z2xLv-n1AhIEAfnK$n568PL9aE`w)Db+l@!#bKVF=;M6Y&sNp1MV^!DuLtK>!LaVDE
zD**CQ%)J5+FMEj+6U^jndJNEBP(G<VQ*P$vp6+!qPd{y#D6mzSwPZl|n6Xr~E&>F`
zub*-efJld*?HU`APAtH~<X_VN!a{iufp%{IYlUgJY%M<{%FGPyrm*$$Y&4Jj60+LM
zQUbGU3PribA637(_)_T0z}eYt+g_AQA{Yak78gI?4=WSd&hH*)-?Vg;yNUhoU2GTE
z9^7g9l-WmXFd?43xg~z<v87i-dDGJNVe^e&c<b^%4)Kxa1aEk=_~Ko3Al9;&W6Lk6
zZtB!nAb7d(vBOt#cD@_9kOjljjr`R+&J<oO%iG);OB1{ujsSgcy0?=0goe1%jPj##
z@lR(^3AC{A%T+zU8L#Nm3y-D)lg72jTkW8by3r$)4j(<!0*O%%P|)4$xhV~Rv7j(D
z1Rmbij=K;wV*~f;&FjNTfLL6-*A#!^DSOlL(RjcB3tX2Pxb*g^F-wh8qt~lsS$=^T
zi6x=?G6t}NW9Q&>$<}XT58OWwM_kAn)*q7r%qJ5*^%b7lCcJU01n4tN_MVPnj@<{K
zfhHyE1$5FW$)qDyQaUe_@qozjQ(qhlb#%Z$R~L*8VDg>-Oo=RgygZc1T$#i^OFa^{
zh4+P*Av_`PGd`TA%N8GTcp@AX`0}jhN|H9UoBNss-8%pPE;v$jXTIl)LoxJM^5@@?
z75)B>0!pkYV<v~GQ5gKC4e+jK{{5nKGkT69mdbNXvG3d3tb{JMs~)Qi0(WdI<XHtq
z!>F<Gb!aDyoypf6Pj0P|sU*c=Z&jpbVavKc5lql@bVPysB+mFh#PS*aIRGIv{X6t_
zV1HHNuuxQV^t1E0u5}tupq!KQ-+g<kFe2^s`^Sy-xt6&L`~w%N097DUHEuwCZx8zv
zOy8vDf3IBYyW=iYh^J7=AIbk7?mN8lrDe7@h>~Xg^=jN-30e@;bqbh))!pTaPZV~~
z0qJop@WUdQ8JUY~b7yATVsFF#5*Jzg^^URZ-7aFUIV*R@Yf_o9Yv~E|+*dchzx*|?
z3)m>34~(l6Mf7!x1GL@CI=`{($A2M^4D#r2ol+erTFZbd)m9{CnN5r@tEozq>Qvx0
zHl`AjxD{8uqSL$-TH4I2j%;X58jBA*1E>VZrT(;?m9O4I_q(TQ(8^JlPfN@W%ac;P
z&h7ijO%Zrk>rKXv3(Lhxw*eamXX`+l5Oo6ocCjcg6(A$?K`mAwZ**HPC1_%SSc=OU
zlR#xAsCs-CuXWAl;-q-61+l|(S-+Z*oG7yh$j=Psdj#`3HE5I>f7)R@#i#$&=JXso
zRSO<U9dyQKnEr+E`nsy53o_Cq?)n67AwNXL8fRQ(0q|nzu$<EZR)5s#J~P6zqAT`g
zF$h&ZK0YOO>@hV*7ck7o7a7kE3FO><okJ?cCQPx@SdmLaGVhX&!TaRhcf>_WaWoZ4
z#FoXf>SBGbHYS3wmZDQAmLVv-*2$dCmG}KMfdad26uY!(+{)E4F92vL%H~h|bkgw-
z+<XtT(XXCh_Z8XD*l^rBdT<@JdYJDf9K}S|QWZi`z>V)QAs?C|42ui4T>-UOmJXt~
zi?ZJ<o*)b9VFo{gNW93EqWv(<Bp&jOO;{L%_v*I+)OH`}gov&#%|@Wf>iqp%n3_%Q
zxnao3NLjT#9L`Lmyicog@X-(p)LJ?MX-7?Sr^;4ey@`Mt$KL+RWnzz*+UA2#Y8+!!
z=zd<WPJX(%mYOSgOp03&oGnkgZkAkS>cht(<A7p7+Aofwj1F?dYDn}zDQaGi&)}{Z
z)t5N6tW9v3RRnlDH$IX}4Cf*5UW>Z^!d&dlhsTtW8af{pq&rG8?>os5g!uj@U-0f`
zA}W>#d2%|onojO$!<Bn;lqs^5qT1V{^oV;l^MyD7Py`~@!LAqSP)-9hY21}Dih9dk
zjeQcZ_5kjPdvRqA1AyZf>8@4Fs7vVgwcF%<jRFjiyvfOI(K6XEMO=4DoqvSFCha-m
zY{07#jy9Xi==<?31G72zHS<wrH5G0W!oVJLlF-`nskU)R6wzln`{2NNWMt%Te||2S
zhVxrS^ZvUxG}tI@Kx%y2N7@_^%wf5pJ8elqf1X09IA1X%e`%aHFD|Bz{jCgFyr44S
zr!@&CKbpPSqVeS;e3Rsc-JjL8!YV}QWv_y`mWkAi^&Eyy@I_khF5hm8vSQqpJkQVj
za)fv1pA}X3^{UCRRw7|GFlPMA<&4G$3oKOww2{5mrp}c-WN+anm|A>K@5cz#(G8Bi
z$&8Pzoc|G$0Skmufy@Wz<%NpKnyv$_?EyNMju$Rb!B++9HGfnezNqD1<i8b}{NQ0i
z-bA!pE`+ThBKlBIT~>5PLZ4yK+5x}^ov~M$)9U=?qAFwJJ*=uCU*utw%}1iXd6TB^
zeV11oTZg}iX<WNZEhEA+G=^`9s;@{Fr4wBK%RlgR_ix}Gh#e|Z$19$RPaUv7N>2<5
zxIxrcRgrvS9?lvakO&CTUZ?xZ)#VHjQ2=Hc$S*!Pmh%TI)r#)Lnx5I4s?$QL4~ty#
z6?qiZW?T>$)*V*6%<J6TvE5I#I;bIM-ODG>NHZ#g=1r;Hyut@#X7jC3-LAwSk3*fS
zToegh{|=o$SW1|VeMGVYOuDZG_zUxQur3=sZ1l<F`e1A%uDxog3crM?fa7c_>fZBb
zbZ-wBsFn&paz90POWZ-sV_Z#-nw7!wC6-(Y9Wr27qkZhGmy*vaFjn%z)yAyP9(~8>
z;THWx`>ZP<tsFw{5URcD{4$%EYO(xH$D=)w9fU+<sHF;Y5KuezB1b?{Kv~}F2xB*q
z1N>KB9A}uTIX~8U3HXT!yBo1xZulB@AKxqBy9qEFRdk3`C>&DyX|ICl23SH~;$bbn
zQBnZCl&Tg1AQ>SLRfAEp2Y$M!vg9u%1&Nr8t$mq5T(2aaTm4UD7}rTs<p<FPF-Hng
zGPc~Kb91-5LWuPtxRt)6Y*5d44Wz)%{AR4aXe%wGe7<R53Tt8yjelF5;p06;nG6f#
zQutoMqb4i*)|egmFqbO$E2isJ^kihdkz{x*FEzT?1V9#&o`Oj&Mx)lJ1|U$><x`3b
z+S<Y>guQBvVxU;Vd;T!)F3`<8Jgi~i-c^F;acpJ{_0Vam>(TF|eSL)2YW4(%ks}LR
zRn^LY5>QNDX-4{!jsN9;nR#<itgNHrc@9DLoU3j-p_)%F`vz&J&luTUOFjJ&(l2{9
zvxmNuim)$QuCgMLCIly)2<l0_lHyI)yl8vpCKv#gdr#uw@20yx@NoY!TXtH&FoZ&{
zadz6uMU>~*Ja;sgQ)!9P(P`E1*!d>5c?Fm~Bu^L$dPV3KIcXSIeOC-mX!r)|2)iy!
z=$wE4u6i7QA_obMge|bwZnUq#<G=7HqulL-&laxe^^yNMYaokdARM+U%2E4WJ!lif
zLynje!2yA|`?KH50?KVMX-3`#pX0I+!=l6b(JxwIN0X1!5Uw%_3>0@SJw|M0Xzs0*
z$z?+S-YbUe;iO@!pcWMxB<a~X_O$}(d+wb&(K7p6fp7y%JPbl&A<97Lwdnc8o;HbX
zVk#YXxNPQjTbuFkCVVTj1GCU`zU?E^=-S01tl0=boM|LRGylE&;#iLw8{574d9#AG
z{6rk|dfNvc>__K8?!s$w3i3mvi>{rA5$+Cbg^grkAe3Jj6H)S!4MBaGVu0z5&E1@5
zpk2H#lL7Z|Sbx$k=86Pa0R<3pVL-q}Jx+0sXAt|TcD+@-2B>2<9l71$>Q|EH8iNc+
z$6z^HKm7rhPvm@G@Txc?HI~VtN%ZdtAM_-4OdqkxM_Qn2xADd!o2{xN!q0F4=pdkI
z-Y~A1V4&8YnLhaHstDM{T=kHWcEUpI5+fe70LN`rx!cc^F2385vW-ijjpLGAYx@Uu
zXbFeO(M3_0a*M$^X(Bv1X_e<1i%p{Y^n_~0h2txspVSu!)o3A{+`GmJ+LzysMw57E
zc7nl;$&GaS3Xc8j<HJNR2zLMLvntGI@b6>b#78=)e;?x_EhtHr`_HMg5L0f@-hZ9S
z6;t;2RsZ$Nw2vb@K_CD7a6LH~yZ@a1|8`+AW7N?(t|jR*MTg(4zaX^v*vYCX;0yMd
zi78N;`L$fHoedD#tXrfGT-D$Ng;>-$nE!KaT4lXA3lRlEh{fx3F_*Qwn{D={Th6At
zovM$PWm{s!LXP`@yUM4so03H6Z}yK??~8}NV9-uvYs;S97g(RHMvw26TQ|LUpBK!!
z6>@PZAuj$rHkRzqvd6i=Iru{ysm|T4!Taw%AT7|QYdG#3M9C9q8dV^1P+X-(G&Ofd
zAGSS`Vxrc=re`m$hcC6~Nsj{z24j~u`A7dvg^PJD5O5yhH6L|}jr}<$piz1*m={>}
z%O=<0JxF46kHg_2$Kbt4?nTyMAhDtHf$>O7^nbcPX0|IsZl{eVn05;+PM@AzrM^uM
z_P@zoLd*T&<IB@(Zw}<TMB^`g4`F{dD3tmVVz}3m_mA^87w@5Bv<kr|^kO&Y^pHEd
ze_`e;4C;#RNfV@_j;@5Q47^woq$lH}k8Mo~|L2)>E?h;fuZ0-Np(6JmJu(tvv}Ms6
zUGbQ#BErqUXmSw?WToZh4t93<m+R}<Kg+Tj=}TNR3IF&B4*vUI3t7&&ZP<?JOSS$L
zJPuGYn@uH<B?RXevZAuGI$U`4S4Q*Ca~xRI;hU#>U7X<-^P@T1LH?%gP?|ul?hc-R
z3_~>8XG?UB-)!tvo50FVSR$vY_3m6pN5@g8pbFhjyMMj{+JwLV`3{7+F#msiW9VH(
zhM~4B#CWz)Xjz}O`hPAAk(qd%^9kGc3+ZL=<wb|Ly;_(AKi2urJ%oKgFD&rXc+6Sv
z$T^#*k47$%#RsY+r3o6^ySY(@arvBj21?1wMuS@8jCrRsE8_C<lN%k}x3BpG(``fz
zNEBd(<h4FCD73`DHC9-Lp(i_RQj`R7adC2pZLW5#XV>8AtUMT3ul0#aD5qARo`Hcq
z$a{iOp|np^U%5IAU1j5)qRMfADCY0Qxybq?(rpyY4OBi&BqJuW2ZrchxAMQ)5IU7V
zSelT!X%eqOo<7OV#yx}Xs+kQ9?%Uk|hlbvsmLJ1m6#mf@s~#Up|M|fE?}mu|o3wu5
zx&vni#&0eGL?RYGc?5}ogl}UT9Xuxys^^QU-_FlYFB3cB<~+{EJie-``u=9*@c&6S
z<J~8E8P(XH91O1cd2k%a!Fe(i?TKdhn;*0E=xUfd+q`ILI;C*kMYiIdyVw!emuBk{
zMA}=_XTAQgZ%P_@nIi4R-M6X`Tig40=^K&@_}d~%DJ0%Pt+YaZ)LXE=rH?duZIP8l
zTHsjSzgz-cLDDs;W)z)!Zag{+gg)7tv>pgWuYB)_63_16v?dni`SRsU_rQRbkB_vv
zhQ|AZ2A+p9GBP^4x?v_Inopk!zkdDt{MT!jolFs%{QUf(r(X<m#zsf^lLnm!-Q3(D
z;2S)=yeTMDe0_a=Nb@ZVw*3I*2M-?5va=_!E5hGYRM5_~g!lzeE5<Y1t6XzP8dkSz
z31&y)HZ}G1C>ItMk~1=*6BDWZXBHLok}WZf{tx*C1YT@zLhZg-bb*8_ouDHlF;LTh
zr=JN~D2csRhw`tXZ!*W}=;+YCd-o1Jta_h6e~y`$FxxrWaX6OO)YRlrR(?!LNl98g
zc{7niSr}wvST6fD!n40VDYdgY%xfpj8fozf&I;2kvTh3Cmu5{6&(c#*<uiKph#%aG
zW=KfO^q5X>mZ*(7X+G!#7;MkRMs+-Qa(rCyv&E+m#l`fgd}P4Tf`44x-Orq#6?7H8
zeGB<qTN}5t<F#|JW}1?nEdq?ft)%J>m;B}4Oxkq`i%i2j6*ADI2-?eQSzF%%@7}F-
zc!M^ag+_4Zri6qMFo)W~N|KU1?HwImaaY9{QB0yt)$9hwrlx(4htsb_Ogn-7ilkMa
zaJ!v9>;ef2YEnAiS=0uNyy~u&aw;t=Yk0ai%#@{oJS^=V9Q3RA(9(($5*0l&y_c`)
zQWQugfy+MmdZ-uP#w95!iLy@1%M%;OQCum=8hKzZnFTR6e)%VpU90TS`CEi-Qp%c6
zDxa{haD>GtW~b^&Ud4EuHBa8SuCVj-U<d@#J#H+_6c!$yqtdJ)on2FNyAzG(m5`t#
zA|hH{TO+{pfx|Cf`0c~<xpH3!git|I5g#v4H6)7f?vRW0?$Nh|>y_S1y30vUKVRti
z!r=<Ovw1i!VIiSAR#t3(jf`9lhfUsKZcpH_vA(`ODFsCuXLU#S==t{v573=4Cno`J
z(=}AA3B(y<Hd;{VEt5|cODXI%R=>y)^d+O@^QJfSm$tUVlIZ#c<mIoCP_x~!wdEov
zAu)RJAZ%vFN(r%zXHvp+etvE+X!_CEyC(SJ&z}5u&a(s&Nyjf1?|pN1->DlK4z|3$
zEy`UB#i$z_Gd4Cg4bRNPOF0kNdVAAPPEPJkNqsUAx$QpnQ(8r)@TdIm-@k38S-nmV
z-H%xn<E>na;_*RS>y(j)^Gkgh;FYTjxlkD^x1nj3QjE_-dGe~Ns^Ya67`?ADySd9U
z?DpfSOjuah{L)e+NCsO_SI45Eq2c1{8k(0k_c>u5oy6%~lc{<V@|EXdaB%R*xA(K*
zN|x&CR|12Au#1b~j~n3F?i60c(Ka{Ua@=xxO%3x^F>8pq`8DuFEv>DKd%tRNq`Vwi
zLPA0=K|zwp$jGB@Z1{gag57g-oN8)nF3+B&m7CvW4m{o^%9iyo@pSrZ?C<XnYL8Q>
zlSV{FVlfSVo^ouv2qd!M&rU~jGHt0@W&4c~3<gs%=_VBnN$F<FBB!8$v0m(PKpy%d
zks74<w!uh&Jx#4bI)xl+S?J82z`({47SKU<HSzrs%32gyPdPBb#LmuIC#{Qf7}S|N
zISKUk^$m}Vgo8&$y?mUPkB?A7LPFOtA^mxMy#gf-4G2XAokkTD@RXWY6B`*BU79Qm
zR2$svH#B-@x&Jy#g4Pp9P@@cVPD*+@ad_MFwl=l+`1m{J;|{n@3@JJJ18?)UTp;uQ
z-BfW234FX{KC)E4HmqkI-$M{oiRgI(3?hMmfWZC7j}vfPbG*Q;RLZ2CeorqyJcKs}
zoiZ;jE@q-ooqH>T;QFC2Uosvaz&5g6U0qjy|5o?)l>z4t4GleiZ<xm;CG{}nAQkz@
z#N;ylqmY<b-cArBu!&F0;w#ZE`NAS164KIycq|+oQ3mhBlagp^pqSq&?92SYoEs^Y
za%|LA4R9*It=Y(aCk<dYeH7SlyG^{=|KUP*3He!ATG~}lzIpJG_WJd>KYs!VNk|Be
z4qB3Z71$LN6e<D^p8i}~;uaLt4nCWh0L60Sc7FCSOl-i{XTCCk5y8jD*A<Orgm29%
zqf|qJrw)31bvwJdzVcm81>!7}EG#T{JUzw1pr#}Pfc7Az%BgRBbyWzMdSYVY3b#o~
zLSkYb5s`brBShalF6V{LclR=erl;e4*T%5>zea#>nrAtt-AhhI)wR3S2cFeMpC<5h
zhK7dXSyhWwbtsKnU6abv1Mf;Q@IV$0*2Y=b*(3AvIHg&sKn6Abkf81nh4^R&V0SLJ
zw?e_6!ejn>W9;f6<N5FH0V9pz|LoN<QbKY4&u+~f&#?ciFCTE?818U%IfyObf$grQ
L0i;abHtfFuO@Foi

literal 0
HcmV?d00001

diff --git a/_images/b44ac03a6b1270ea102001c424a678d09aa630125b1a04b4b4429de3f8b4f1fc.png b/_images/b44ac03a6b1270ea102001c424a678d09aa630125b1a04b4b4429de3f8b4f1fc.png
new file mode 100644
index 0000000000000000000000000000000000000000..f87a653075dfbee32b2f6412193f812e7e2682ce
GIT binary patch
literal 75337
zcmb@tbySpp_cn@3Dh*24(2_&P0Me+0fRr>th;--B(jkI?#2}&|h?3GFHG|ZEbT=d2
zUGF_U&-eE{=RNEE@vigES|i>wH#eWydw=%c*LB4_(^Mt8OM4d!3ybKnnvxC{7WM=d
z*6m?@T;LNi+vPjJ4{Q&G$ItPBS0KK1G;mGes%GSYg>{b{^S<?aCg>9QP{LE$&{NmN
z*3-w*-3H6a($m$^#naK=irL%7-NW9+S(smlUx=6arKhK>lz_m0zUOyww-ZQXijTp<
zV#a!`B>&tudvnIGkzE<Kd!wc2bz-zFCEZR)7<Wq>8`mf-oEcvwnv7CDBE)K6aitaK
z$VP<L)z9S37iL!Z1V}A2Gk<M}iaFN;i_$Qj=of0GM;W0LXlJ{%&+8W#`4@-1_n$2h
ze%OVfHk!Aa3NJhj3}73^BckYOsFKpNC)}S?b^i6w)!^omDc1j7a(}*_UIS(Q`%BDi
zQ)*Nb|J)2*PpZ^+75wLVSnRj5*nj<Ybuxd&e;&^LBf&FS19Sa~&LnWJlOA?BnB)Vl
zZjk%D;lE)#*Ue+L7rW)(%bEmv`;@<K*=M`*)yTu+X%DTS!+a*EF%KT#SJz{>{TuIn
z!2Hdl!p(0kFdp->ZUe^s(<r%1%j@&f8~=uERt??Rvwk0UB!iFZn5j>>@19ljLC%h2
zI`nJQNvA+ivTdc&P0-<>>OmV-^P92e;~z)RI4XhF)hrj=z`)Bru8)`*U-$m`*VKo1
z+<xA|b6t2%)$i}<>FL?PFc+XjkIeS`{rGa3Q*Hq%cfIE=UG4Adf$qO*x{yE{mRYny
zpV49#LWb<ml=3^w<ARbT5{-Z_N=jAo_ugD>x7-B5&xZ@$r<`i~{4asgvjlAYp!#Z1
z8Z^FCGvzdUwe924&0SbFq$z#Xx<cESu)qGbbTyha(c<Rn1U*t3BuXddnDJxIpn3m(
zqRg@3@oaz#vGk^(?0$06GfL$q84cZj=^Y6Ctk=$cJ`~S=)(@R*Goje;GvhVpHtmMo
zi?@S>W0v})<>tz>VMdjA?iXinuiVv;oXqkbG*L1#DR7xJ_|Q!`LJPL83puGn+3Yl(
ztbWN)T-@&w$o-+EO<7`6_clLKDz@olHp@=#Vxkc(vr|;#a^*b;j5^_ByCLZvk3j_*
z9`N`5YNR*{*6=5ih|U+Y+ve9R<`scw?wYb^)<LK3Orui=5cJhfuv`>_H+8^HlXw)P
zKNmXNqo36*4zqx8iGQp|dcwXi$FSTAXUxQ-uLRk$`&yvPs`G9duTjMe%I{*U#$)4K
z&Hi3ea5iS-(|+ij?BMIOqhmUk#^m5jWbktBQsUBv`Hhqd;00%wXgO5(JDv@g>|q|s
z-j+*n)&WW`_v`oiSnr0u`vxZ94T$;me!3r8dhQpEU3Tlym^tbUor~hDnQj9feG+XG
zFc@dnQPX_%&FG%@&5`*{=Zw!>-yV|bMwRL0Aom7BW$H<02lwNcJ4J|}HndzwWByWK
z@kvd~)s~4{=lyKh@q|h6)e*)il4?Wz_ByF5t$&f50fy~<&?oKzyeob87Na)Dc1|R(
z4u{q}r2_X;a+fC^N>_H<Nzn1a{qtij*P>Xt7t`p|Rq>WzI8-J-fu}nGUPOKjj7GKb
z!Cz~1`yZ3u3un3vS?Nh1GY>v82tJCwXg>Tbu#z=*$ssH(jHRxj@pmQCflrg3OuS`V
zxWjac6|h;ZdevQgbBVessR>@B$es5?1u&NSk9-~X8Z(tV9*0U^Y*yB7R`hgL#5>pY
zyfZNSC5g^%*H&fd+-}(Em%D72E9>EJ=#|1OxZe`pTo<rPC{5|Dj!-<tt~xZl2XIDF
z4o*&{YDXl!2cdnYdpnW%aFQwD8k;q9L7Vj(-vNu1+l$fKtqH!UX#~b>lWyw0U?UJf
zBM>mrQ+xRP>8;Z~m!S3^V`e?T?jSLL0Pus}po^`OXePLztZbr3qIZL7i_W0r$(yV1
zE!Y0Gwzj8Ts<Lgu@IAsbK4XM~*#YhGTyXIB`t^rc`lV)l{0*CJ@^}m#iPGC$vcL>y
z<f`N`VJtZF(B)t8?TQij7*%2rB{O}y_h+&y<FDOaLGz2rF^QE_-P6gjW<*yPzAC+2
zN4`$=jQ3Om+vD{)Dt;4*jjfM>KT;tX12#+0^qvFqpX|G@PUf`WXlo<H{z{MC)<LhG
zYzknwWFZ<^V&(ovuVi;aNp@RFg34x9l|tOL9dcpEergjcs;ZLvU94y81})uSB)uXW
zZ2~;F@%pIrCO=79Shzh?)InnR=N%?@ja(oi{5kT!1neGfen64rib3>5&Ij<CO3U`h
zsBGz;igQIuO^F5Ueu*W*MA`F~@S_sDa=^b5B^INNoIU|jt%#Gu)LIvFWSq-n_Hp#2
z@m+);Cwz=fZ(L>vTv9aKy_*~Jv+=Z3fFqVM$KHS*Ss5gb7>;{B^%VaW*Ibz2ZY!Rg
z^#u^HcB4@;*QZMu%bBRMr>L3nrkioDxqd$CB)PrB67dr-9NCR;)px>`tDBA|<LI1g
zLyev5UK_hKy*p_>X)^H~R1RFYXKvx^ds@z4fe93his#<i7;`V$+U%gvT261dWH&JL
zW52nYyXlMJgrC;FTjAF7>n@S&C0Us=&+6ll-m0RI-f0weqnBHz(5SiIjXXO*11@K^
zU591_Oyc-WY19R2bf_YUOlo;K_<O^))c2<27r=txBM^w|KTE_UB$e&N(vVG2KGXVC
z{jZ%vx!0ap&dq15a<%Y92}Hn+^vMJe)9P-nFK8GUgVc<u_+C8U5f>5Z7*u8Ilb+op
z#`xbJ5R@{RF6Z%rU5);paOlV8zy6-DKXUktDkA>I*psH6l{bpSXW_EoLtrEzl*eeP
zIbbnBOlQE_Y`)l@I@E7BXj<26JRd1trN}+SR%L>T_)NRu@fdtx24V&CV{*Wums7M>
z+wKWi&o=sD+$wM{&U_gjux$xkAGyu$MvCty`ZR_Q4r;@JphB)p<}(Kq7oYK@*a<pm
zVA{LXrrZmoY7F5iZG4M~^GU!FU0HntM#>!Y3QOKx9&EJ0j)7pFi@CuAB|{Zr)04I*
z9(*9)xM8{j($Gy4Fz@f^zy^-aB)_vB^XoFCNxdgGX7@=TyI6g;DSLZ+QJa4cXszJ$
zp=`g^_rOsg3B(qD$LZSSo#2}b%v|ZchP5v@TCRESQSn=sHXVPQaC8R_xP|P1uDM!_
zg$`3B{W}MdxvMF)#&!)DK~4(rKa$CtUSX8=(fY{IPPo=w<3c2zVZG-XlWAf*iCi|Y
z$gn_6gv)2T;__JST}ai{TmW4B`-$H{)|`5pP2&om%dPowXwu5CuGWh2-1%t9J|PLo
z!7r_#w@(F{DADv7qXJg`aEGrw&g>Ax-*V}`+Mm;78O0#xFva9Ct*a^d+N>1$Io+(g
zVfKjMd)yLl_OL*_&%bm2V(unL_P9LBZBUu2+<8U{_`M5hxw6F{r7%9=H04JHUmq~-
zTN2$#Pzgr`9aoz8F1&8sa&mzltoHT37mfnXlPHX>JUV;Q&?C^24Oku(nI*8eOPJ|2
zY*kO~eB;D;$d#XPMbpj8D2&^rF2?eFpNBGCP>av{=3fs6{B<GDTQK&&3E;5z@S3M#
zr7ck|khNgKnrG<WTyFibg52q&zgeBMKPq6h9WX<rE_|Qz`g-T)dXGQx@3Z~m|H*?e
zc8y6K&!7B1GsSC-YB~9<pxoT6O*>x0GA#7;k|{lA2$)AdDUu7A4FK#MH5;&_Q8_Wa
zjT}=3?g!?Gk~kXHW!kag`k#Ac7BGh@z3b0gQ&+Y1UB{d+Tf=a$;yqU@IueAfL6NSM
zkL0l|Dy>WUIVZCF%9_QMrJV%}!5!OuuOWSe!fa`Y1SPDTs`&Utz8s3(s6vBs<Zj|7
zS;7)oP}(>2_IpuYn#`WWJ^IxNj>kuui@3N^d8!tAe+1V83<-(ogl*}t=UW5c2H1{*
zKpp%a)9tF3aPJxC^Tw;h#Ip)Lh+?^qnhcCRT4PhjKa7i)QB5y{82uzBUl4Q@3$=&9
zAan}3kLS?ox?eMf=xu9872vi=e0(`Rv4c6`uSNs(tU_rfkQAdR5H`rcXqXyppMb;5
zY^lzh6>oP(Qihvah>R?)N=j93`IF9knfe){Q5ON|Fz|^ZsNjH6i)rS)%^?1glH*4U
z*q`JWP9jMny0I;BYpX~{^+y>5l2}d@;4zVxmSxmBw@q(nvDs+IDo=65*F88;S5UEd
z;%=*e0FBwQ5G*bJp|+%u^tn&r*ju<p$*sJ5c{Fa8eUN%2LSmvYF8Hgalxd-|B=3!o
zKI5j#qHumg*~X38t^9*GHvxAH0$bg^{7H2E(s6!qQTRm>HQB?Zsw1d6q$`Z|EkbX`
z$}T%xRtr;F@(iqb=}fy8ed9ZwqhpHW@0rHsytE39q)uCDgK%T*QiE>iClK1*p$p(+
ztOB=Ed+$o(j=uzrc1FU<IQBdVd?WaAQdroDaQY=8i={L_TX+n9wuerNu^<+|o|@%2
zTXXDGmNbRrEP@-cft^NtRJ5k&%Ob^Br!?iLput)di!>rlcuTD+Z-FR0R7~s?T+U=e
zkkhI1!@W8wvxbQK4ZU3itP*)Ho)|!gphimJ`0dy@KFf4QDVhU3LjHa#%%p4suHZhd
zR%$K<X1;D}x6$A6R!Yp#^Xz3)1E5t^#ZNlB!QU_DD)&1IZg!X?u9{^PtA9;8hA}JD
zsI8$T!=6RlcHl)&hhZyWJ9G!};rB~mTXH#90E5Qa6FktEC~Tl~JccoLi3#VNlp*$W
zSkXh*g(So19ry`Jc96sTt~FkBx{MkN{E>yCRHt=IkL&L?+VJ$BuG62^$T=#B<s*jC
zybPJDR#xuZhU7`Y)=(?n6RM%<xw)U#led`Nz+tIgT(_WvkIsZwP~W&Ls|IZ~A@S|M
zQ&Nep><fPLj~X`m23??%q&7&&UTC+4jZ)uB0bjQzls}E8A+QJ$P2OdeJUO$!8<7@6
zT|=}=0R?*!IHsl&H6yEbz@~@PiJ9^B4~i&7)O#0SMxU|S%=L+t%jU4Tkp)yg5Js`_
zve_or7=~Wt|J<?+<DRoC^o@tvMB}_3D33I+&!)K>8BL7@gVMZ1l9}1$yAl>!sq0nn
ze`3SA-NnusPmV`ZBD4Jx$ao(l72=2TQmWgUzR*^1WauBShwaZB4bwv9MZ!eW63z%W
zem^$3uj|J}9b?=Z6B!f#naX|iP+U0Y$hA7Y(w1__^D%6CJgv018X>g)Lc~d7C>4D~
z{wE?^qmvM@Vl%kH$Y}H%KP~N*`{8-_+1a$8n<Zhpn-82w=5+_$DNl(0mQU76sH~xq
zFSeu|2cG&6n&0|CD|N|%LN&}1|LLs~c4)$mYjAyXpFJKEq+d!U<w3i4%1L>2A>11L
zxVKn9%Oi-`VW0ahM9YcSnuP$j9aktQ$Hf*Z)@MD+&J!!d&wcSZdid%$ZTqv|UwWiD
zUsVnh3&$#k^es>8s#kF5S_<t9{L)DyL_wu?_Rf!oCgwE2kW6><x#XpBdnJdb<j+J8
zwM;9r?H?LJ)OORR>gDAbt3?T~H(yQi^6@g<qjRzlg`F>M_&d@XBrT+{tb{s-4X4*R
zh&Wd=Gg{UZ$8)!#F0otIzb_OiN>}7N4Yif)pO<jk*0$7yF4X?Yewopcw$v;dsep|~
zIDwrsKp?3o#?e~AeMms=iWnr~27*9AT#T#lZJAuMsc4&DCO|l12_<UoX=*y^wU)`$
zZUYQ<g*>w${=@{rkSmmO()HN_LSy?_C@L~0gP5D7AnSc^{b5|mCe!EwyC9{gE8_Qo
z5rdCSMw-)vE7`Z&P%5n+QR*0A@6^RVdB&!8sKnO_$>EauzkPciw$`Y+9VY}2N7D2>
z`~6{sHXf<ifzYed>iWc-5*C!aP|`{B)QXjyLO0sZDfl+j!V;QY{-sZ041s03=g2E{
zsGuTeZQcBnI$ZdH@lg%LS0@)5Zz<v|y#AufQ61}?1v+^$k*Yb=H{`!rL2X4f=uV)m
zrzPx{t2v9LdCD4VUt7BwCx)O7kA4<ixMoG4sJzF8_Z}SCAqb7|DzNzWNj*Y#NV$~$
zqSBtQfESvqi*0`5XUE`?+1@dle4eeVlG{I9G$5Q7taTn}cOdY7i29@M;Tn~sEOsNd
zH>1(dx7NhcQ<g%veW)l(wq2~CF7=7Q4usOgeV@_7-mm(hej-UMyc=Anmi3u^Mp33O
z?)mzQq|XF)WKfZljjE>MWup@X`(&<;dMRZBNz;B%35vf0;sg*cl?I`Rljv{l2GBG0
zQC1T8!lz$KE$wd?{5r%GNTvL8*BX&((`__QAKXQ}W${)=O!N^8r$^^mWK@OP-_Gh9
zr#z=7JaA0W20_a7&pj1Ytmp$KwBeaFDjh@Z?pjv0bJo}v#AHybUxbdIEQDz&nzhXD
zsyK#gX9+4zq}y?2+@;A1Ob_ZJNQG(?sy&8j3ZS?Yw2?pO1FTsIBiUjzhN!K~AZ-~A
zgM!52c|4MJ5$>|2LcBK9#8bF&PFaPK5&9tk`9i@?hcaT6!k(~h>(S5YcKA9;EMd$b
zuS+GdkN7wTQ_$Y0Du$Fvw*@J5Mr~3&V`bl;oX+pbt9^_<>m&1a;GI8g>=QO0Q2d=u
z=bX`l?(>(DwefPHn1OT})KxKv15w6K8`2FKvmzr%@ns}rVC>>hQ&+MA`M)KnD0hJJ
z^U<4fk5ixbp31&oHsnp75V9Tf072IwgZF6cyTBtdQy+a(2VeC!avvI*_cDXn#2y~l
zh;<gs68@TJc#ar82pjyQF(npVY}jCD%cqRL39e>Ec0M)A#No(U8U2FP&~5pc@_AFT
z9@~L%s2Mdmi~V%!K+RF<!CieXv4MDUxKtzSYqgxA!UHpy%zJ^IcBO7fhfVs|B+D$E
z)!<|vq%)JkRd#E38NC7CS;-glEwkIo_(!tFNe#(Ated}V3|g#9dVE>zr;TeJz93xm
z+3=+eoO>(XEtx{X2vvv0f17d?R2BF-(@U#`yJsJtw_t^!RqWMLq4Cm)bW(T-KROf1
z+NTjau`5FiqiH%`eaIG|BSicJjHowGC1dA>HTLMT?#eXVR&R0mw+r-qM*sM}?B^Fr
zWh3;jTod)PNhkJc@lMXn+DgC$4C02<**)1``q)+JXiI|)CRT6GgRb`+LT{w~sN^H^
zYURyVPOBb{eP{}9xN&}e(77^B$G!=v0?tx=>G21+?Wxv|CHwP2;=LXf*(8b6rlk#z
zE`RgK-3Qt{iAv;|N>v(4iEj*AQ==_9Sz70tNO?(lV>$>2EtGgkUs}21Gs<^R#BlmX
zvwc(!D@|OI$(WC3FW`O*TN*5RfF#x8rP1u}j8X%Ub90)8qvJxt>meKxpqLo9j#MOp
zRrw<REa`1ubJ-S+@Z?v?e5}thK1vlLABUc-9CeZlvk^6tfnx9()yoW*1VKk`;nXD(
znQw=Wk{92%&UgN110im&NV=0{hqPOTy|&1cf*XdHu9t8VI8hAUf69v^-%S^;R_*yY
zBO{V+q|NrsP(i(it*TrCEJ1<4_?XS2bD{X$@vQ|u+wEj*+P4E@M3J4h6sW>e@WYhD
zSbSe5mq~b6Mbbq$_tuKIDn-0{MlxIW%qp#y?sd6ao2wC1ffQoY&AJl|B3i{YnB9zK
ze@nfS$AwMNJ(u;i_u<mB`(O=lFPUW!*HYn6QVMFSRx)O`Lv@9pf&zww3T0p!?j;d%
zMl!9?`SK2OE^H8eOlq8Cc(oTf+0!~fC|M&le%gHNV<koMy97#l494`yO1`%_48sIc
zU^InfV6w;%illeXkbV)Y2o@`FUI`C)r8<(0&G0Vw%V!9;c$UsZw&JO-fYvxtrRWfg
z?pq!Mcvz(PBNh*^!%|_aY19Id>~U0m-4ro#f~^bc_esdtJqQ#Qa2&9qUpOAF?5q@F
zI~u_)@ahWUTd;h2lPUFJwDF>w<I&WU#e_L35%*DVM7q!KlY+5bxbCt@{7fASt)$jv
z<cS_WL%fJ0Qs8wnP1n|Q3HK+k8A<h`{$%Du02&d&Y2-dooancae*Mrk=YA|RiIFvh
zv8<TyHZy19leRCZ!DXH*5f38OteA&k4Nm0Q)E&O$WT0Q=i`_!lQI%wx%%I{6SJmo6
z(fDs00k!8^SUhEqZ<lvSaKBV*ix#1#Vj)+&r^ufMWL|Hh$;t`u#m5eivLFRnQ(?E%
zG~^#fs_j1{)-XaN(wxY@T*MjPF%-0!rR&&Yj*@mX^#<dr`$0@-&@0DfpJ$4Ti3=J>
zTix)2aYw1+Dz#^JK#n5@pZlNAqFzATFDe+g+wZz`FS*Pl#VV}-Iu7x&_}cvXhga?^
zH3bTTtCL4Y%h`{5Qt!F6Tm1^<AbtE|>>Y{4rxIxj`U8!&M8yX~YF&ST`$qA-rs%*T
zANps)gI-`4acmoNpbh15j~ev+jCE!^a6i<HNRS5V8R#`7JPv=|5&r@eIq-oSr+k_4
zY4{n)FoM8-Ak|%6xb^LG$v|~p%?|>hxS`MZUfmWcru*a0NVX2fuU;+!?7D-D#Y>K{
zmwS=}l>7{>tUx2JExIkutM=1UjqUsOl9C!5M=ROmx>L#p*TtT5abgF0ot2>QZ&RaB
zem>Sm>IjBIp06@+lF1WtJilV4Ru@I8MMoEVq&n##-W2Uu80!8QOsnNts_GSIBsE~9
zfNtz{M+g;$@%212tt22;_>zRixoggmH(w)#u;}K$#S=!5_bR+1iKaOFVd~Xbs-&Hu
zKFx<Ww>fRL8LRSW4hQ*u{pi-U$ss&{g_TW+Kuti?pY`8#JYOm5envUsmGQ0Ga)VA>
zY27=X-8KTZ-3rTcQi7a7NhcNdFhVG)Trc;;_`K1uXoJ<ay-Xy7F5>$Sk>WkK4cVnT
zl=AD|F5u{_FK~*HrPp9Dm?U5+xb21oeQAE7DbZp?2sSItcj?0A>am<@L9lI7YkkI8
zb_@2eR^yh{-JfNjB^8BJ;}zA!@j$h!W9NXCWw%o5zFFSq7$A%8Ouxtfc0kC6>1jvG
zW9EX!0!b_Cd+@~a5-P<?+J}+RU;RWp6F&2h#vlofysDyI3+M~2k{|XfjL5Z@oIP9U
zDQ|4_D{XFYS0@?CfMhJO1L2@9<oOMFp;$8W>v`Z)a+S~W#ui_?&66p|=WSP}<x9rD
zttubgPT=Vg&7eY#WDkwh{Tzmdkm#ffr)Bu%wI*)e1`m6<@^qG%Xw?OXCd~?ot9@op
zD``wGUorVX|J77R^S+hc(H&znDPFiTF82k3S!;a0M>D$UV$1@5KTZI7Sez4fDz%C?
zGx_{L+(84DyW%?jiWuFV77!^({)m@o&+F!)N7}c8PG%BfxIi^M)0;9q1_KY3?rVk!
zf5$4v7>14N%p08Q7uIy2*oyXvMy_p$Pju26)IBPZrmV)Rq=;0cyMD2UBhA-udGI)5
z%wg*P2V`7_nwJcR5+*hr(ZinYgzt4^ep?($jS~w$E!k<e>ACC;Ak}4PcZ|^!Sdk=J
zsTGWFRU-e1^1mMSV&0z`Gg|dX-p|1LUSCR`%?g~IPWo`-kpo=C_?h2GZ%3NUktSuS
z)ZsdR*BxdSQrtSl>Z%i$lezk4(}}sp2_bR&Ct;lfRMDS)TD*MHHoabyG@85mn?&}%
zFvC}L|7QrH|8KsE9Pdf{zhKA2nf|vIoj}|1SFzq##!8Qe|I@@NLXM_5{|hbrcL!)o
z<9~x5wYrt7-@m89G|kF^zN*LRLX^kRh{4!Y(D_iV`=ouL8_;C-=pa+YKzBL-BzDEH
z<BNy#>-R8h4wM91RR^cjXj$=%@sk-JCV}9yo{8X_E6LRi=;<|piAnbeiHddt!?J~a
zGpe+`Ff03aD+=gG`9D`q^0~!Gg29E1>fG6>Ok8gP9l#o2pf}iC2|wc9hoJA5`r_;a
z%w0&F|H(IM@aD72@p=HX>&6;aMX@#;cA6Xa5?rhU1GBSSn%`IUOU0S`%)GnLqd(vD
z7u)Me7eHeGvJ5zU)vGjMW{0<@Q0~fJ|GUYVp-J_BjU*MqZT{;R=;h~zlgLs#@BvT)
zFDonS<swzfJO+DGW$>cJKq=ak<rqlmsMug{MIg8?Xd7sRcTLxMl#jZPTSo0|+R0@B
zZFQWGTmaIX2Kt4Vc4GdFbLGoHP`}hh$(&Id&_-XGc5NrBZ2n%m7^`}8{ukE#+n^gh
zKKc~MKGpX6p52wKh=@|*&;rn>TNyQWSpinStcfooBpQnuWa9pdJrF>BEN1GxsG)yO
zF5;7t?v#0Ny;$vi-@d!+dUg5vY5H}4MSNnSkd#z?6aU`^<G;3Fh8PaSfFx_6l}8_Z
zxrad-{l>R2h#D{$o?4(U4D@1U%0~Ay#~KUFB8;nEl4Xgz6b!8iiilwUTp#($RJo7A
zkN<5<NMy?X*2kn}3~)4exhj4Jz)9j2yzK!fm!JbW=lV4ctb)Qqw|=l)KfO}KP9R$P
z$XGe=GQbG~6QFa0Wt+^BemIN)6H14erRV?iTmXDvt?#OShiSlKutyWv(Fi@z*`*KM
zZEf84DqR9_F06R~f(GL;`Z8noibua>z97qau@05dDhT_B3jlMzy5x$*pjmY%vjIi`
z)LH7=$s@54LBUwsqy4WrVEUp-$(%D&%9t*w>~cA#Gk=5vy72-4BsS45TQlLfzd#av
z>cbd}!N`OG;HLzs3EAtA`}pZUk61^&Zhre;3ti;&DbDdfUq5kT{WsL}uciBcx%;2T
zC#`tFn&&KNxq6?4cpEdJwy>8G1u@hWU`BbRUGg%-+t*<+CGEv8!^+t#dl4Pf+0^b4
z*&&=*7C^(sQMg^K)k;k<kt9?z`E4f-W9xI9F0$Ki$hbOEqQVlj_Q?Nqw2L^!r$4e{
z^;cFQy~TVFXUH!pGJ&RoDAFlBgyosD7A!HEl&y?E%UTbTSnh=A0O6D2hO>wzb{WQV
z2b;f&@^hQDTezeNstsZ>&yjA^81gCq0@f_;B8>p~f;M)^Gu{wF31M_#J{%Y%OEC%W
zGuF?i%*=*_AO)4s^hacn8IX<C23bS&xKz-PIf_;~R5TKpk5Y?n+KAMusOn2qqRH>_
zW7dwui9ubD6*XDl9ZDq$lkpQAOV#qAG-4bWx2xefiyh*Vgz;$LL%4I7A^VW}L_AJ)
zOzS}VQt^+y*9qW?XJuN+f`Qj^iw|yaumlXRd;-x?9uuVJ)45Jo9>I}#=Zcq|u#=yc
z#Fj&plF-Xni??XCdly&fE1u1l*bIx(5Dz?31AA&%aD2j!7+03ittJU+7hDOrtf<Tk
zcyW)$hws=mfAb}4ee1%T{hri|D2rV~snN1Y9I^TgFi8x)=FbfG5jmkq{BWwMA}hpG
z=5(PT4}!x$M5O0khwgecM<X$T(3qeo>6KffY!njEUyAY%?t(U45jxmf>mK{_zXIjg
zJ;XYNy@G|POTcVmR^7)h>msThEWn2_O>r%nCz<L|sm+w}lN{yc>Nuz?%UGf_k=MS2
zxOTjw)Dc3XM3Kxo#}s6N_|f<XYRCHr)NdZ*qE!^D)H?_)`xW=w(fV6zj@v8O*SMgi
zvXSy8Yy~|7`P+IPPG_)z8rTFL&~7kCl~_7KS}ze?2l<vD-hRU{d=09h6Uk}@lvMCV
z+f&Jd_d5%jWubj%ge2Z#&;(2L0?wvPg)nh@u*O_e`Xg(4zg-bib(9?z<)WKTwfT?|
zc7$kNkJ4Zse$Z&$x>Nu)g1QmAppsE}mm10yR=9KEpF6HX*@~*amcQN`F+dhh+ewv(
zb8B1Euz-Xj)o(Y)wn~9-*O8r*|54NAxA8l8R#qU_>MyO_j@X=i5#2ujm7bn074|K5
zirzA1FPi6Kh{IiI?+}K7sS8o6Iz~h3BZKt31#}p+{?I9wMRw8N7PP$0u=MMzN|GQO
z2dJ9Np@BlZfC1<z#gc<INV$>8nxlmd-Eoy8C(Lb;p0NrLzN~?JG)<s<e7N^Cmq#<o
zkcYx1K>w;3zd3QO722P=m->1?kccY%im?0AOs_<+AwfB!@l}odMLLXY;Maxc0i|6H
zWR6ymbp3$jyrSsbg4X?<C7+4b8bLv4oC{`(PtvOksh`0D>8_H1S45Q%x8@xu(%JjC
z?mjga8vR5Ouf)f@avDd5yWsHEKY5)gWbWZC4Bvu<ulux5p;dj*?%}rZdZv;g#3uOV
zWJXuL(bZ>`0mkl^i7OF!M0D&c22G;Y8YrQL5G7{O$PPpDr9Th2eb8CVxF1=t>N%Ko
z%x;(;6ey;LJ)*r^@#?Y)F$#M^8vCl5w75R&k58@i3*@RVb(GD=KLbbhFjKmHZk^mE
z_5n`<zI0ML<R>?5Y&+iaa*HnrEG?E$k!3w|d{5Yu8a3`PyC9EL_;Lg@CAMu~F330b
z;n<v9Q>cDw*%z7qM&D2TKI<OmvV}}5>gS9_3Ga_#zu%;sXWLSyoQ07_jHh{};hCcc
z>WP(@A3f2W>?ubGQHr`D-eyF1r0H&76pU5&{84xz)~dfJpqE9w$`sx#8EqNG{@se7
zS?x_dQN^=o*eF~%y(A9=H7!3s^ciF$FshkSol0i(bBfaF$d&gq{=E1;lRyUDESt!+
zk5TE*R@7NYaAYHjTlAJPz+U}7Y@c2L4qAL>#t7**RC)U#*^N?XH42VWbL{xyP!7Jk
zX=vzu910V>d`PoG#lH}x997i#^u@9l!|QM7b<y+Tb{;N~DwIR9<UuS9VpX*2PrFgS
ze|R3GCVx`a|1o8up3&P-;xpv+9lm$vZi1h;zGRbKx}zmTUh{GyLu1(yOaF{U%y!7}
zk0zM>2eaAFH0^@Tdj<W^=HuC8>D#{>B?f>F6*Q42YiA{JBq2g5#94c_uJz#Fp4S=R
zTrqLcCHZJ!eY-4s5&AkOsAK$dSh)b4V2+gpCrdp_dyEZCR_E(Zz=m6pV5dmQ`kGaI
zUgK<<zqus+p)D>nSvU>AYd#3T3*4Aj1SP3eL;Rhg6GGKMofJp7Z}EtfJSL+*0ElbW
zkPgFP+eC*k6rQUgB(%18iV^N5Cg|8L<{9pxcY>X_{d5>T=`UO6Rd=q8sM;tWM06Bn
z_XH4FH*ea7yW2K%EitiJ2tpMzB-7Ysq0aez1y+$A%Icq2V40=oM<2Ga?sk`RQlJo?
zv`n|}8AF#_D`j)7D*pT~bCXP;F207?Fl|0oLD^C2<?<=e(E06B&QeIK4t4UrOk5!N
zv`A3E>(Sm3Jt1UQMGZo)Ux*Z)*OIlXYa<Gy1yfEoj<b3jE?YjeuiJ*p;+(ALN<i&Z
zz!b`OTpqCY&gr>{M>;dA%dNsGCq_9=9AL7hH-k*QMBxJHE0*UcwFQZ+c6^%M7t);7
zz0e08Pi4M4FO(zn%~zT_n2B*OGQX`?!J*G3Z3Js}b*gKkdo%A4ufD-Iq#CIMohZK_
z(|Dhbi=H_5iRV5kkAzSj9k;^TEc&hDTlM_RKQ$sWXA3Eg>?bqQIx4$`sii?1Ki`ds
z*4~hUy=HeFa1T|hdc3*#+(;r;Q18r*%#oJ^BL+%V6T+`H%8Orpt=R6V-l8&I4eOsg
zpY752nu;Y?F-eJZ8}BsUc46ZgC)JtM4qt0DNTl#BOEaX$-i<I?Id}3KcX#n{v17<e
zA#NJIy*kd}Cpu=Ht|63iX!M3Ru<M|P^WPZuztgn;J0!+`XOd2b5vJA+r;BltTUA4h
z|BXWzafWjCf~Ekj8lscR{L+_7@h_D9_a7Mo8OsL~jQ9NDa<j_+RZRtl!AavBF&Mwx
z`C!uNDN?Q<paB61pmLMbO7}Z@|37&cnozdSte@}BhEoeb@$g;YG5^(y$J9%vO5X|)
z)3h=Rj*c<JK7f1Yj-atC12}fyo6lcuHy&te%3XO9%br>gOUz>dv@4NqfC+al$#gM}
z^qoC`BYS%Zu{K7(K|5pb2d+Ie_a}U%@Lxk_#0(9SWJ=D5-0>axYSx()d<nf?lDobW
zGUeQmu&`hW5OQ|hM)VBau8zhIj!J_WLVyAd!y|$NbVJu64e`ARR8Sg*zjKej)?&20
z6rfmPxF)Uu3(~EPfD>yMNZ%u~oa<vSZ%j#abJTKeaP_bTQzBs~B-J%d2@n*(D)W%I
z+-=`LUso3uaRWu9=!QdIrf8DB-hT#Mq}obhmAsYA3s7<l8W-<(xb-?WY#~^p=-akx
z=7@hLh|l8x+%P+lqD{bkn$EvM<l=h=U=$L8beNXd1^fswnKXj>#kU6l?+q`C-W?w(
z-vCNr*6bk%h7|@onZXMIM*<TG$-TKmW9Xp0y}hmg&r-&UM(%1&>j0=OeMZNM3gPE}
z^kV;0vrTS)L^fb06}-Mj)qmbJ;|%cmuzZ11bpl{cxb=@r0efqAekGeF=41tsaURt@
ztNEe#ReulQe7(uNJ}gw%(t<kGY1X+dsd&%9e6SEmq$IG9waagxo}z=V?PwDatZA4c
zPg1~tFwRz{+P06@^yOc+*55Il3-3o5SqmdBla9?BeZd&I4NHW7-sbhiPT_>3kk9tC
ze@kn1C_vgtm-JZOeODUzy3PYFBq{j|;J}G&jJROtR(HAIBYC!*f}#3IF27OCkPE(U
zPlI}TrVb7c-e^z!S8bS^#@RciOvP*X6QGG+;O*RGqak=?3m7JfQ5uHP#$W#j1n|uw
z=ISiOWn?tIj#yb+N35-_MO_lr0Hn9Q^RX5<K>dmGY1%OTTSx(wbM{f_z5f&`>oF??
zm0jL|vAd#keB3c?U7)DFn!SpOio(zlPXSg=TV51nH?SAo7$%?b?0(wZStEvkISnua
z>3!!zvt8Q=AR9)g;+of_f9rVQbL3iKE^#VTcbR27E-=cEP7RO#r-bA_;@luO@E!yv
z)o{H$pt>JQ5@<2H6R=&!nQk6*#=3no1NaI{5@2dUfk$7>bB@N#k&MPSffF|AeUoAU
z^9vx70$qmxJn-{%ql~55zxZDN`gC|n`G4zq{GZXS;exsR=SqiF2!-u})`Gyu_a9MT
z9LVTYNm-&T;#+B+Ix-}16)ad`-HkB0-7oP;Dy^N=GA2HshTWIIlO1WHrr6&&iqDrm
zKqNUNhNOOJ;g-Q?j`&?gKx7QJ>*Aosuw62%?2$yEr3@icom}9&v9HH=Q?RCI70U}{
z1>20{MA$xOWTnv34Efg2jbG<`yUiDhZ^!EqGjSkE_ZWVY4%wNiA$K<H70n?yb`&Vs
z(9Px^ONe_WTPgsD?-_<>SIoK-s6Q~RcxI8nBJ{n)06VUgHXIz;B?T02ECaFRI=LyY
zZ`<)M{UVLaT>8Xf2bI#8LsLVk-8Fb4O(0mTdRYVtR`liNBIVdW<L241-i%jhLUe9C
z4A0*xhKvIl<u1q&ij_e}L_ycnZxm{LN#!*yf|AXFY2EOPzkM5-oujgPhuLAYb(O%s
zCBrJ8#xf`%)hu~X5S+}n)Xln#>rTiRPVsV*I<B=&zCr%UPgWo^=W`DdJ-|B1%HJHO
z1#wy-UX(lnDX`g4zTlz8zIUi$aVN|t?QB+%l3Ca$TgS}oN)AgN+phZXUeQN?XmDmP
zRn^A`@97U~-L+;yLP^yQ%=p5}xuDZ?p3-JIXctVo=3Dfe@#v6e*jm;M009I0?w*>S
zI99Tkjigev(?kKT;s?hbk>!#Njf$axz!!S&fx4Lt(*L3a>_wNk9`NQ7V+iOglLta*
z<$^aZsvHE#4(i(Qpae-kze0^~-zwUVXq_)%j|MPMhT9_f5ev1C9rO34*yoEK+z~=G
z;C@qK5-T(jN~*?<Oknaxmj5bQkdx82|5VU4x5C_kKr3HEk?^Er!)PvxtvV5oDnkgh
zfV(rx-rsp4U7lX`hmHly=Sh?T<ld6@d0sEm>l0tP^&oP#GQB}EY+W36GDs7s9iXV}
zPv-nMC1tGMu1<>z6Iu3TL$XqbPGhgW9WKv^<D}4f2>(tXLCDB`5D7LUEKm4zpN6hJ
zrPm>8k?d~7^X@CwB>?2KHl^7}g*7{L3)$+JIXuePH)ZtXbDy&A?Jx}Wb^R71ZllT5
zbE~;H>3MgvJ=jZ6)U+1>9<B434{(Z5J@sv*xB)S59!SdimXwPuMR*Teh`bsSBpWqE
zycMIckE-edYt!gv4bk}fiSeYtb^)%J`7)M>Nj$g!;(gy>G(#voy3`%^;focD=ILsj
zdwk^so=A{y;1`6<2KD0?Y~M?IKmb<&;04s1({$6zO2Av0LDqR61SDaPT@ACrB4u9m
zMn4}<z@Hx(x_m|83fyw&1{d!z{6+~XYP|TCo$TDUI!ov_l^n|{t}Rm_s-fPFkbtg5
zmGExpOKK_Mle_92KaI;Ss)`d$v+7z}0t(#Za$j4d*O&8~e&z14J9h?$ywODtOw1dG
ztDO0~o=yc|8(sr+TtfbWRY5GjgHw2&cAKFnAAM@wLpHMr?zykXQF$1?4eU@+eh&AX
z<Bh=-RW%aBs?bp9zyXLb?;BCB(<Nv!Bd+oBEL$^&i1OM#`Y{fn#P@H0lSmH9;^4(_
zA8a$LSYetj-SwXPRht{+02)5karpUzJ;f?pyNHPqT3T6yoC2VMbSMGL3Y497^M_q{
zGwRQIRpL@_`qI_2X=^rqq!-g|UHfzhrP3%_!xwt#uR*L{GlsGA<%s#RBMCBNWcKK;
z{+>PpC<i)}ZLVl<HP{61lB+3Bq-U1{qZN)_i>T_oYNj8HEcU+-aG=H?qOr-uqWA$K
z3cv;t*Lpuk`^{Zb8%@t+BMbC6s1;u<d+t*@5E4-dmr5$MBY*<pe3q@>@P)SDRruvS
zQkRjPNw+%!Ztv3-nfcribc%$k8o%ER2?zz^`%-it7g&_nZ2oYR`PK>dlgmJ^e)^?8
zSh!$C`z-^{R%~R+)z29p8?@1-7<$i!ciKQ*aBGrsaV;vlOKQ}Ivxm`%FprjVlUe+E
zotjd2Gg;?KVBV{*X7Isy@=T64E497DVs9FUs>h0n%)Yw)c0cDe-s@&^d|V;le>QwH
zn-caZGk_Q|6Pdu>kuo05Xq1D8>oL`uA<h;2(<<ScM^5LPSk7+<bw$aseoV`aX0G%+
zxzL^h1m$LX^H|;-VoCwFWu^|N$g^iC!ahHJMOblk`dftT^{Km=Wp!_I?0SRnij1nf
zBopxny-Be^<p|`<wdsef!cVfm3?uVxb6gZeg9#l+1h`(NJYzm2H4*(zxSujr?2LO5
z+Daebdi||OXr4XQ$)@N6&5Tz19^d3W{FA1KxVrP>r??qQ8ef#_Ar>nK;^~LAxBlBd
zjwEqUK<^2xvBU7UzQ~t${kdq&PFjXx=Q-%8EIx1=`{EIqw`rd7%A|PTl?DE#9t~8h
z0L^k-S>$WrRC*XZb!@EK?F<rO))nn2DM!<5P}59RL0zx><i)b_D&eETKb6H9R&8S4
zCJGjA9>luJNpO_mFtvb(JJ0I1Ck0^Sze<cp-SA7cGzJoBz39h_b4+GPhdtZ{Y-E&2
z-yMXl2le%&U4*ro`0RQtcKbDTrGa}4<l^v?c@CB=%w#1{%x2_N3yP8tC|ieO-}?6_
zpM~|}(Q4~#bKYW7ph1lQ5RRx36u*MKkfhtFzs^ZUfC-ZR%G_WkO3e!ZMn$_G8w(=_
zq8I(=#-)Vr?mrS8;yF;q4{b)Ry^as!ivBunRqW?zw91E6cU0FuLMDIGTLGgyXQ3X!
z1)FsIAT?%hGU(%mZN(-K5BeSXdc_%OS23ja;086DgY?rMKDiFo8<9je9M%NX7urGn
zsUvUEOfXav5Jo6#<Hy}9t$UgFcwa5ZkCZLMESzm$Z6H8EO}&w9;iJ3JM>c=G-gt{)
zNOlKvgVm#FZ!B}*d3GDyYby=epVv>2CsXDx<#BMgnVWMDPGxPI(b>hig_<LU>&70<
zGu3zpt5*$Uq~2oGJz2dxM*y*@qZ+C{OJkn1HZ2Wri0ZH3wkt*lBe+!w`VED0D}_rb
zlym&vp~gk(!Y3b{DZknq`OG<?^v}yX`N99si@^==<Sj*}b5f@BfhaTmOia;&;f;$v
zOpgBdt1B*GhQA`_pUcaN0Z`UI7hU|5-{Ks_=K_TbbG&IW%zKRJ2w<=055Dv~DYWlq
zPd919RBRX=2FL{Ec&>m<u^%Ha0<<w_O>-AAVlJ}~hv3SIWehJH5HeVg6zg3sFkQ0%
z!mmHuE$hW!PjlwZ7_PT>t`qZ<WHT{Jtc~vtiG5k(2Nx|j7Z`OGrfdRWhLWrTuDckm
zi>A!cb7Hwm*M8Xx$8CW2zkU6_wB&k7?uvpF?hA@y3X-5Q^JUT1)h+ct`e$NY8y(K3
z)>(4FvY6`VWG+|^Ac(VK1O;Pe0Zu+Q;w@(y80Fb^^M$i4AkA^z`Z2cuxA;q-6L$wB
z6#D@!CjrxdfS~c_Y6lId?BMTi0AW?D%*k{+TJFXVQ$mK{A(jY{I}Ml1_%VNk(I&|p
zkB3jWwB)p$Wwk6UG2KX{o46A&UGGL=5D|=);tbI2{Fb@SbojpDV17aScWeNVhpC-|
z&kKXQg;9_p8CyV`0s$lmmpv_)_snn3$Nc`}Mb)kqWSLbFb8v7_Gcv-DXlZEt^wtgj
zF`etVabV|Mf-j~f97>x65oIXZi{06i=5zr1scg6jxkEe;WcK2K03xlX@wHFO#Y`Yj
zf!~@3&@LN5nwR_I2G9v?7aYxc^zhYSphqD95i5GW_%V*ZEdb>MC=s{ao~H3G000(=
zHl^F;9#U9P@FyVJX$SNSPXS#a0FFWq2NY+PTpIUOt^u)CHc(^GIq-4j1`1(zsAksR
z;Ut4A_Vg#7?4vpp+l98**@64NE_b8k0tclqlf&riq&AMGTCS!_Uogb|W3CPYlR2C>
z2=V5>8nz)VISn{^*{^?o-ooeRE&Eejh86%Kyxy9-CJISTPNtfXgE~6K>wgt>$nr?c
z%gZw|=P|_y7$Fbe-912~a*vwLO7E1xF1!m03M6(K56X_~j($N<G_<tq%}hXRI~M+R
z^3{*-!^6Y4(N>K9XrrXk)`^aWW>DcDH|Y{co~iB<6B24urZVl|l-(o1XfyzI3WCmM
zndy3gsUGP#S!J)Ei4i{mP$bp^z84)p*I|ijGZ!o<_^9Z)m9K9GfVSbiyn*}x3?HMD
zGkCt*+${U#GdZCBWB|m3<2>C<o%@>`8v%Oz<YfPPCC@->V5Xk_{<J$XckbL_d)x+e
zO!mWai;e+7-2tZNg`hARrL*7Z&BIqu{s4twKStwXbb)EPcn`gBkOXFP3b<JS;HK(d
zM;23ZFWIn=Ou?5t4?W(pVYDhHek<=FfV3%s^rcu9#{3~b!CYF#@sFV`r_VSWTTj{L
z!fJQh?$!PI7$>n=)?T~c&4bYzN^WHMG%hg&0vPhrSkTs(1W;+8<;$J3VtP1O7*#B$
zspDL`fSU-}SCa*FGzajLMmzm;64)O_KxszSh!F~5f)0Q=&Yy3T7V3X(T68xDG=|y!
z>qRE6tvHywFl~!K059i29T;R6{u8|}-KW0ErEE=mpjLo)&W|nFP5{%ub>9bJK*q99
zr2!zY$d>t{s`rNj*M6^p_o#wf%`ZF68(*FS+Q}Q7DL|<qw{eBBX&^97!gjKuKy#+V
zsB%^L00IcT?5=J{=6W&ydoRfRmd3xem_GlXbxC*>=Lq`e)u8tOdGWI5dhyQL+3_bO
z)@{~uvLY0i8WM^n0p)(50sj)uInb*cI}enz@$6y`?pgclKyb?t7U6l;&{x_tQ7l9_
z?*;H#-kfLacj05l_TQW2DSr^&Zk2G05G&aY`?ep~%QM1|7tagUK(|Amqyhz3g@#&u
zyG56RJmcNic><RClFtIg19lX}tz?#5_=Z0fFJ6>qk6&;cf9P-#5q?|4^xF7{uEV&d
z;*U9}#7hhAzBjsV1VdsUHSakPmb-Nz775hfLcyPWV>S6W#M->ScT(Wz!TN-cw6T+f
z(_Yc*x`X=-SC#Zu5uy({3amix5LEqL=>w@Emt0mb*>~L=88MbN9a8*=ei_HHqP1x|
zD4`9-Ae4ffm%CZFB@J|{dB%JXPM<__i)qr!vup#Mr_(V7A9MbACkfp<NoCX^4A7nT
z_irC>Fm>`y=Wg~#`5s2y%ivmRbn4GK3xZ~hOT5y4FA39o&_Ema0*;OHrb@}ozGZBi
zJFC2@#>qWF9o1_2M-;r`OMK_*KBH57S-$Aw=b{}U4&n^Fw@8*>f8LD@X?=ZsRHbf_
z`I8<ReTS{}nHVk;eJi&6ZT6*T^@f@EXj3Q|<n_rR$9?~lyTSn;=7NRKKg+GRTK#de
z5y~Cf<<|8}>7_a)RV%qerR^RWs*^A-MNY9`e{#oxFxhQJ%p!gbE>0Sq&`zyrMNh2w
z0K&fXC+Ymlct-3~*PNSn-=OpN7f$gk5#H>tHRZA<(l5d~lzU4LJ@iF7S}wFTr2~8J
zEw<07mz9Vl?Px(B2mikF;K^qob^M4AU#oh|s#U~pgZo63oNkayhvT=Du6;^}@+W1@
zU!8|YnGUg!j?gC@AFb&rnmea!5sP1v&aOm1&cWvHb(bxbKTMM%$`yiZc=$TeD5e@^
zwtjk@v*sx+d*4W0qm%uVlnYvf-Tp!Rd7uR^cX{8K{aO4lnoaB#5+6#yVv&${YT`t-
z@DCTD=ztDdhFDZ&u87p3go3S`feItAFcDVB)I<$cugQI<p_J>Cj-pi@U5o1<N1y)y
z&&bXb7!ryae{=)^7F1oEv#HHskyM*2Yi!m0m<#Wz09lr4j0*fW#k(~y*@1dILF9|V
zZzA!W;Ka8HQ{ZliCx#}4Flj;?DEDjcTyDH@oEJH(BSN9JyOh{(HQ`&)GtQ{~A2k{U
zL=_-TQUI=5wu%Qo{|KPz@!;V_p-Ar;YswiZ^|<UO*>7s2LB3XWWa*&nAL<1}<W^W#
z8(o3&!YD!N?DCoP)`3HsS?n^>tmBOPj!Hw)pr!gV)O0>gC{U}u<I3A4@8&NsvZjWr
z#bq~Aw1D57(cAHHR3QbRgpuBJ9@ON}HBqwf!rPxvaW%~p4Uo_(VQ1yh!Sv^TPrbC@
zeVmfQFM1bRhT0hY2p5VteQgQhLRk{-uIBsmF$wJV7?bjl5|JOx&K+YOSyYL3T2ETI
zp-fgeVb2{E+6#eRQh-C4|18Lx>>~9684cc}Bpxp$g%<P1FgoZeZbJG!^3Au3^&6h0
z{!^ETn63D$(o)xpe!SC}GgU!a9sk>oUR!Qlo|7sg@3<y1<P9uCqHA@jaHGp|{tP>(
ziyCu?O4@gEhjEgi65iz<An*8b>hV1k7AV#hCg^)3%xLg(nY#!TBkeXyt>~g-pHx1s
zIdR*_&WubtDj$|3e8PI#FQEZQ4l2{er2_FOeY3%bnV%Q}0w?**e<5Zh<_Kx{a(^J-
z&by6zPkD55Y{~9`GX0+XLTvfIsX-ZGCy73Aa1?Jggc4vo7Q!y(xs|t?2li=F3M%^$
z8DC`N-q)2D9B@{AH%=s-+!8qf+9CGm@x?qp(3FV5yIXrrRuqJh9V*m><z6!m55w#n
z>MN&qbaE*anBBGv?;9bxbZ5XXR|AEY9IP1<AnX;1*E%G62J2>8U%R;KGOD^7PClen
zK_e@-{J8;R&?-k%G`g_BzIS^LE)V8psrj=Oe^+|$?KpnM`{R>(=CZo(?u>hHhj5$(
zWai1@-L%y-n-xDtO{5#5qRK}$#V?<ZlXD37s6=`D9xe~;c|Lj1!vo2h46PaC?F3z{
zex~W9Op?ISctLHEFXIs+%Ru;|iM^ny&?^C_sz-n$f0Ya-D$b@nki7GbR`qzhFD5Tn
z@dck)-CgqpA(a%`gC$w`E64VM#9u?BZ*ae4h1_{CAw&Vl|2mtL4Ba7G)zw~Te(B7V
z!*u`2gLP7_hKR_gYXV1tquEyWPO^AuRD*;#a<3c<9=tOeo<mi8hDo};J9A^A%aYDu
zI;j}fe-8z{dd(snDV6wXXqKFAp$v#E)~q1jWCEa-2!24N-Y^ywBEIEkheggSlViB1
z_~ESa+6=sPxr9#8wOtif<UFQoP3h}EYqk)v^u})W8)jF|lsC|`19m*Zk(!WCGbxO9
zkJ2~LE=Ww6QC0Xt{euiPBzRQKF22ik6kPNBX(#(3;=Z@7mr4O*yupRVhVEc^SRaUZ
z;g1E%-L~95c@X(&3zkkI&EB2SQ}J0VeyQYSKbCRIJe{FhEJPUAZ(nVmrLa5l5ml6n
zKG_6i8cjBtP5J#;dL~bMw*=gq;9#xY5EMAN9Sn1OH!cM@zAqV#Dr2m7l}G`DRss%`
zdnzN7#|PS-24Rgj_8t6&mvLm2_{hhkA142G$S$FL2BkItZ|ECVWYwD56Z3`JJG&GT
z;ncaz<#(t6vG}=@G%>38*6)v3lE1uH@no_6m+oRMt$B*WE@ZQ|10@e6>s*j&T~ZYw
zybq3V`@dM$AB0cWAlHFFPtINy8~>|HUX&kbIK}8U&CHf;bnG45MYRRE&%rmaDY^nR
zUsUbDmM6XB-idwbU<YkPXS1H*lzRc%8}!!R?%9^}Y7nZ!z`-#<^Y_H6^q$1du=2va
zY_FMXQpu>jf+?S3x0rF|tuxo%oO-94nD=%2y>+(7sMK~ZM9H+G=)=K{hldS9%|rZl
zi1<f;`g16#1`UdZ28(rL$XD56gq<Gasaw`U4rBI<AX>jCX)C-Xqtt4zQXSP9?=t4n
zw;J_)?8Y))UmAWZ)NeRDEv(eS=k{yyTSv>a`P5S9Zj#@kv5h{5M}}7KyL_#K0=P_u
zH-U#KiF6#Uu5t=rf?EZskdwQn6pZb^NG{{{zL-Z_U=kWCyd{x79$oZt!dXD9mAZY_
z<3R-%yuB@ZEER4|@lVz^q5R_i^2HqtBsu!cxTYbF<)8Gg<x^D`?zH#+V(l%%s@lS~
zQBVXyLPDfbq`Nz$1s08xQi~2LK{`~rQx+{qDc#*AU6Rt>DTpAPG57nr-RJtwb$)$6
z*1oRoTFe}C%<<gM9kaHO0=!tqyS=@=K&$la($bP9i}IgCar-^`-^$y)owcmcpg9d+
ziT~r}HRS378k*mMMFkQ}NeMDTIR=_-hfz5?2T-9Q4T<<y0S+S6r3CcKe+(;DBE{wh
ztliy{fDO%i(0haYQXGI2MSrj4pGB&7WrNYf=?KAX`YPMmE_`KS<QDQb-&@N}`)W=G
zg^+7N8LCxd{}_~?9wVoqr@9hA`f3q)8JK~V9`IG6+}f`PEwt$pqubDzKtE2XEP6q$
zTV>-Wy_S-g2<!}8C#+3ORv)a<{J9kM-J~wjg)hE1fD#t+wu#yh0HrV+RIo$EIMM>J
zjL~j%|1kvf0Zm8_fF6Jv=)irl9%?QkAI_J9v%KCNx()-4N40#78ql*ga1gKqT*rC9
zT^)4)eP(xa#x$t>=M=9OLS5b<>nlL?QD_HfWcEmsl9CWJbfum0nmsEk3-VTh92ms7
z_TJvPS9eqN;>-AsGZJv4kQ<8wCFA~eS{lW#%S!`Q;>iN-c5U-i9_!hjo*=@RKRfXb
zZut0c8tSJ)zA8}94RAGsR}!HHHpIz=#Qva$_+`2w@Md2xhT3Q>N0Hf}iOko%aq#EQ
zpG=H!0|QDgFRx<y!av1JgJ^?MGb@ID&*MPD!Zpg9BTC4c#P6(a)?D<27_`29HBV0w
zl?ebKh=pukz<?XI;Ql$)Eqa#OontBg<=b}CDED0U&Y$5GJ(<gQN(G8wFchLqiGhy?
z%CNpka!SfotytvV!&=aJg&4)T>E416Y@l6bLkz5dvNae71ot6#0N~cJ)dUci-IsNL
zYaTCmNuQSli@?+;?T7cp@tQE;371!aVcT?hK%<b*V0Z&@k8Qtvvx3y@bLX+Ku~0$D
ztl#j{eEr|Y!%oj`lWtm{2|N~1rI<yJ0ouV(%M*ARFrZO-9`3Uy{kbl?sCGZWZ|?zD
z70`p^5Eh2TP(%GxJ>ae)75!QLe{adEX7mexeLvt0a{~MZY~MWmcg*Y4ojR-jdr@9~
z0vqdgfBUh!1C|uzZlSu!@HUq{wfP7jzt2rTH^c-kakt{IMn#24K?$kF7>WI9IUrZ%
zTFo|NeG7CEHT`g(cv58kHnPQ%<r3gc;tHn?iz;tuz5BQ=uQetrtF!vgVtdXpU?aXA
zh$|CrMV1pPXAom-S3R<#l)*M}B@0MWoYI4phA=V>k+lDwv0F^9Nk_YJ31f>RZ+>;Q
z+Ugl6w~(Q2vgd@VY(nM)lNgZtc|T8mk)5BJh7>O7y!_72Akem0K*;W^;CaT~G?}#J
zuuh$^Mn2OOdhmjp`4RJrY@v4Imh#=bdkl9XIReajX5C07AE_WQq!l>r_weJ9@!*ei
z#QM7wA``eKJ?;jD;HSsAYh3Fy7&|$XE0+gz@E7%8n_O5ZQ};g_?_f&tq*?A|*f|hi
zS|2>3NEu2_yVlfE9)P3MoUt_r3$+nNR^(`D>~B~sv?qUoe?qLhB6ZpsR|c&tV=7_f
zz`eNUmsOVOM^e_L?dr3{pRw?eNN;5iI(MQl8^T&h>|Ny_tK8eGD?)jP6-E3^F+(nc
zeW}(OeXow4h<)!YmPi(8pLM`n#Du)j!M{Uh&*Q#uf(xnjhswDoWaAw`X^gKvDB1cR
z857PhohVE&2R~7Xy2woK8Wo**vu{g*kt9l2mN3QNOG~wX)hxa3lwt{NO{DAhNJIB*
z6ha$j4L}dW%dZ@?T;AW`3}<4N|8=2r#QcDi>skXLYBn3ET;d}fDe~oH^UZ@an^$mi
z!MXjc$A)1a`SMkUbUt<RH_G|<2+Fh4I=fY4{<HiyUI9(sy8?IoM0J;2znT*TcL`LO
zD_{L$Te;r3G><phJbxH{Q`~Jqz(9A_gdS?N&^EFnA{HSsKxKVS4!2Leda?paeCL^{
zdwVq+@1GKyZCkm#6iep_7_Mz1uEJM$d2UQ*s#lGbAY|&2dTGTmrzJRsChW&efO2cL
zTu!ISYXB~tx`1(lNOs?6xpRH6+TPDLu+-8dLfn_ja2MwPw#;GoQ-;5J2Ooky3AT;}
zmcYcaOID}T@0neDyP^#l?rLG)*=ve*{-rIx*%eOAbr5o>N9HmSLx{|ElB@3HMt}G*
z^#hLI(~LE)lD*R|BD{yu7sfA0&st+Q?mT|~$_LPD_wV8Hc_eY*znL|?H^(_5V2&&7
zMqlSJ>AYOlYAED2lep1)b#>)k*Qf=r?-se?!B_8loGEunQ86?kpuC~~uuIbSr<-NB
z?%36^%`!_o4cUlQH?GgdT-fhR6a(g=wfvF7UESGt9aSrS?V$9$>u1)3084zngYltf
zUSa06)f`$|rltp+gi{z}M~`rLc~u?IB^|`zP2!CfBgX;=jy;c&+E?Z?E83k>r+Cr7
zUGdGf_7dj>Fi&~%zV(euw(7-}JQ|u_qRN4_7;RRf^JvoYMuY^{xg{A|1iT&Sq!vYy
z=gr(I-zR(YgA1<TLZUF;wO13?Ikw*?S~06Rwy~#0&Z(CE20jN%4-bFBwReSa<#hha
z%yYkD3*n-QvLI@1?tOH1X@H_UwgHQAS>^%j=ViPMk-@?524loe%+{$N12Mh|n+`Q?
z(l<>!Qq?Ml>Z@$}SJh4nX*kkVsRyv6p2}JE&2e!Pa%=@7W)79avziw_&xB2L>;^mb
z5#FGP?GXeE^DvrwcLt)y%&G8JT4+t0#&B+*N<Tm%X`1Vz+R2kP#SCyCN%}@7>^Ht)
zB6^TkJ<HbaG^rmPvw!`WMHOgi9c+q~*;R_1$H4p;Z$2kT@99B7pgmcben&NY{CQ$g
zWuICp`fXL5uTML)1f~1T$0Y~T{7nyB-=_I9Ox!smW97=#SRZK+7qR9*@RsC|<2Wuv
z2>Cr#)-#LjMS6_#Lgt+fXF`2H^^Jf?zj%D7u$dEYAZ^wF!@3Ml6W=d>QP=h7zouf8
zW58mxy~;DD^mRyE^@h&`5%_2Nw;LaS(MA-WooMM9zhiR!RO&yzvCHkutMn?X*aaz+
zQ@M{sUX<OErSNPI5bL+7OsPy7gyutk4+uAKc$R8xEa=4UzDyg^&8)pDA?h{+SEH)R
zrA+yNDuKpHAd13e=57AyqpS<=cdFr3u3+72WS{XtI@_B!8y+W*tGJV|tsp>TjQJd2
zbrI2<JZHTf`ei(mGdQ?Kg3FbJHj4NiR=JqaC)XUIli?Qth&|2D<U*YYZ#AleNAcwR
zMo?9(Xqe31>)Jw;dYTEdHRK-NDGklrV95!gHT|H%B}YeCap`6=<J8ilbxQ?D0N%9}
zeCT_b@6bvjxg=7g9*|!lH&;PqtLQxYYb;h55O#AlO!f04MK!m&NGWIbm{=UasA1Fj
z{*`xs`UpRPMQ<Qr7H^&)Rk=rG!P8Z_$y%o8ow=-aOl>wvav3=h;An`K_slr?+(u6b
z3C>HtYt3vqvOOr7E{1dbu_X$Z^jz<y^Dom)Pu4^P7g)*Ho`eXlIN*AorxcO@NDCDh
z`jXWl-N8zA-dS+Ti4Pc*htivaJD7VOyJ0jxT?B6J)qd~a=tNw&)(>63PRr3HRIbNF
zoX<^(!D$2f`*Y2!2KsgDPTmfBr4t*>y6G5CUXvwWz6JJb1AU5o0sNC`c0cdiPvA*x
z4!$YN6q)~#@HtKo1}byGHJb;-gEQJc9tft#vJhGeo!mS5PGl6mHMoSaP#ZavJUQ6!
zC1LKd8O9}jI9Ddz>?3NRQz7OP2NV91{@}5%?FRXqZPt5xwJLq))V!I^<Lw#~pTI~z
zt7{Y$U%D6Ij!GQnUN^b%?jT}ifH|9>Gps7d)x{m4tZF%UH2iSHVDen;qRGtS);cfh
z7ln({9Fd>)xnZxqUNeh$(%oFoS$gNzz2zb-TAlX9^QB9k)za9`T{w;UkUoiO4}d3a
zdW#2p0x~aFxkJOsLW$`+`Qi0WzEA7-2wP8J)@Tb`ZbcFL_uOL?|ExVt9$H5K-~6B7
z@@)S|+X{56U;Ss<<o#8;z8<N5%(i;&?DIOGuy8+E?!N#{lrAE$Ku05geIK&?0vylI
zFYc^=mRD}Sf5v;q>{U8KR)$bN%;)!{Pc~TmgQfl6KGnayIqOZV>Ygg^-|i5DZvPTe
zlK~YOWP6Fs4k&w&ub#THNT(cKY6mbZyj77dp$pVR){{kRzzNI{P)ear1=OXOYXE|$
zf0nY9Ru4dRLB<EfUIoJ5I(Bw!kp5}9#^DKA@fXiY|JzS`^&>#nL})uNYpTkYR9W~4
z1qmt>K*_TL9QRNbAvZD5n_||j+LazF)~$|!yxV~CE!pj4{U|!p^Hco+Fp@NH55NK!
zECTXkgc>9^(`9Ss38tQ&DaGDDzF=Zwk4{dWXLiG2Fm_>K;T;}M&SQ*X{Xe(RC6Mp?
zkO^2ML(Evn0SzL1g67}{s9EEAR-V2F85jPal@ZeXkD@n6K<7t^bRYs<4bVn}4i*WL
zB|`l?H*pHcvT@k^*w(@#$hxRDfZqM+a}#K;{T$B~tyGv~{<kcfB;A+e^@SWWh4yRS
z0b!T#$B&Qa&#J1APREqF1O<Cj-^_mny^s^GF;=aTjLQS7O8G{8G($j^<W_{bW%~zP
zTOXU7Z|=B37vl1uIUUexv8-7@pZ6zXLl<~%EnJ^fn>Rv!a+;3Qe|B7!+_02pa$Bp>
z9nhV9M1X@IC9n^rjetW9p{A=fwsJrb1qObbA5u4f(oZhtJ=8@5Y+$7AAA&0@D`H#&
zpySvF3>3RA=g&ofyJ5{fzWvZ7xZYgN@{E6W(@ZuWrCdht*z;J+`l8$CN)~b(&1gJ4
zLkxKafUfPLbh_xJbU`IlpzAq*du}sV7cZ053aY76gTz<BUkxHG|B}!9-!i0!bc)n5
zdO)BXlsCX8@<{~$EekdmMJ1){@9EuVp^!!p)ORVMJDCd6Gr{Ny1N2ih#(3{nbr9DU
z^x?BF`-uJwfYtCdtP5vV-!H{Y?h}_#(XyEpIXDBJh+q57ADFd|-~RscqA;lvs+&6>
zXue)h1`xnnyk;3_6s9B4TA&s@Fnqh}Ws~`{N&QLagd1N31O{G!zOSq*0njNrfs(28
z^_Ci7PcP2B<o@%RPKVEzfZr01a<75jJ{&Y7spa^8F21Who3Re2dY<*?q>DKVu^TZV
zRVV0G&+M}X{3*&B1~+#y|6FRS`e$Hc5ndub3DwdH0vx9$j|gBWeJ<8hX2MT$I%Bt7
zlBDH*M%xp9@fVqBlMOB1U63peEMz_ON?HZ=2VEQImaf!9sgE;<?3B0T;!czQ*Bos<
z62oE(Yz6vgM9W!}D$bn~7);5q<wfGgXD!UErmx1j+5Q_NnO_sSp!Ii~2Z>v6o0S+q
zP!aGx*kxRJp(^bOYGqy+SHhim0h|2DvGh~+I86Jfd<VmOdcJ7#hc(6JwTY{;_@r{7
zDIL<ngpr4pvjI~(2BMt=0%F*@#1sqSc6vqyJkloQ{^z@6^<XjOvhG|ZgXR8`txuvY
z4&8>A<+Cc~bJR_q2tvQy5QZ+#cT@Z!i<XITJ%VS<16kb(TANnFX|K%1zjv3qu-aJ?
za)+;-Ir_ERn^<zJ$aN*!sc_=nv-L+&5N~<2zlJx~w>0n+X`i{BPz?{rz!GW%xFsX#
z?!&jZI9Sx#Q~3BR8qk=$;!P2`NNF(Uo24!b8sf;RDm5ol)}I@An%Ap1d1Unujf|7*
zzH6Alx#jvcFzXn{q+=t+I6JZpuEc5<T<M&yXAxIFvTxe_vmOMLJLrhQD>#CA0<7DV
z^M3v27oPblP-)xGhHX+vyw;j;o+m?-R%Eww*wi>3hyl-O(QeQG=B1@EADQ5{w<?S;
zSzO(p+i9krIY}DeuKTmCgl8w)64PTnz|^hl>G*PfX0N>sVg;*_pCZ&T$<^@|c6(di
z2k+RDcE0Jj@Y}<bKkR>9Ngh2esZIQ}#l;N7iG!|!15Y*JZsS9VETtUePdFp|<Gs&<
zep$Q^>n7DDwpEO|c$>ZEmh$SM+R0(XiEt`6(K*X!TC@6MP$UjeueuU;J0*Uus4=&>
zEql(oaj5d-zU%i-;|!Y|5%Aqdmp}TK@{L3GI=R7`)(*z5wM7FsZ^`Prs>uoldcZ8H
z(W2DNDo$o1^Dr%~<~>gW`2{sopCkdwAgSH)*wC=I$AIwQ&+jNARUPMDv(2?Nchg>^
zss6kL$9(+$to06e*+jwVw@)w;TRe#tBW>IPQ)&7U&O-tgir%W|OF;I0v+s#{Q;{6+
zy_+Jo`jc|PP1cz*?NprrG#0B9_ref@g=SrC@?tFM9PvC1^@+>`CIcy;0TF2?)U%9C
zAupR(hlj=<sh%m~8s}p0x{H%-AFd!cc-Do|<ixU+dtEg#+{=ackb|yB&6&p0pyb=G
ziHFx)D?LE#4Ou*D#h0*{#M*22bAg3iYS}RDYEt*J+{zj}#Q(raqQ*#ToJ4RtuC#+l
zkKNF?YIO1Xh%V2Zg-~Ooq@gY<B;4|67?Tc8Is4YeK>wO;<0Z##uEnpEB&KqXbi?Q`
z4xS@@hPWkKRWe?lY{Aop-CE=<tB2!J4*~-Cgz?=da>a3PDcqCodun9pw#{%yc}{Dd
zj5)TbZ44K4HjKQf^Szxh2bl&r)V_>5;(x-Og*PDFi4XXTp?$8mqAeFTz^X_dm8j%_
zF?>2||4}LKw;CxPM(AGo$%i*Go+Og~SY-dscZ5<XoJGfqfA^G51gy$>llz0eoYpd0
z_RPYc@pc-HDtzlWoq5kqz}+*2m?XAS1Z^LjJC@OnhBh{949RqFuD&0YlhLd;87irn
zs=E2!QDqaMrR4?C&4-{4t&saAErU^9xN3N0p)iCvjv}Vzp-ms{w6I7Bj$c7fz7+cT
zB;WDJnn1RK!jhlO=^YKRjGnt()r$vWakvH>(Ou=AL^p;_kewttlpL!bI`dY_ObheK
z8{}gMPqI}jMV<_7-E<JW<4Mp_GxK)CCMAZuarPLEP@KB8yX+9>biUhqk!BmAiZbMC
zos_yJcpmU}<;K}3V#ZM@UJ^C$Ec;4nMa3JBHb03GK6sS#qP3<xjJBjQFV_F9h#uRe
zn$hzyE+cYMymn7r<Bu#SVd-6l{Pal*;sBP?aSb?;1BWBzz@2M@zg=YbDku+kC9Km}
zYtcQAs9SifJ58i<H;>ybvohQ)nnPpT{NtA;bd+1#hgxvg@~u1fNLOgx4p&{fBtwN=
z38i@O$Uw0jJr&gW`Q!*CCzKXRgx4s7B^qsr*T9_9uB5Oww&_#vRZS!XzB&hPcodfz
zru1&GP5l5H0~Y7d^XaW**kuK|VB-cN-}WDE84n`|T|Os*cLrVsx)-?G-CFRF&dS=j
zmlBN@prXvHaS>t;JLuYcYN$FI*;>-`<M5d^6X}i+v9*#>cs5j0S*$aF9>IKG+w!M$
zZh{KW!F8LI4yjL%>2U636*4fYO8AURj+ul1fB)1@<D->Azr%xs50^n<k8%74ng#<V
zUT_+ygv-x~!r5)C_d5cg(#R7r{{T}F@F4usZ*IUBSldCMnjt{9wv}%S*6v6AWf8U6
z`n{}XjCP!<Ji!H^@j}=hNjk)d;7d#%=Pz%-v6EYR{~)>Mg0AmEY>DGKpiy9{&VbI~
z4U_V`Rqb{(jc4Triuwh1KxXZ!O#G`f?dVv4&ll<G@}$CCjaZboZtxrj^LM)``T7jc
z*Dj?zxfXi0L<bH}vD4_QR3eN$?~-ZhwsFq(=~1QSd=2O#OiQJBQdWc?D6E^)t*t@2
zIK9Qeds@Ly6FE)Cq?NGStvM}Eg7IQ#Aw962AkGTYRdN?rZIkSoO8JdXbx613ILzns
z1JDi|he-%XH5A?qXkp^4!mnK2@xk^Oq&WtJO{t1bh=z`}Y>}NZ8H0oA-<nIp#nM=2
z{%`^A4Qhchsun&QDd06?)X>drCY@9^&}H*%QEn{Lip(k4&51KzrwolajRIe_jr{?J
zj6|$HzSC=cGu1S8!^19i6|FS;w7c!-Za94>gwA9*31sJ7o%XyDTnP`39yuwt8CneG
zcf*vPtlG%EDRXcaLUM{>{4IwgqKfDd%qgb@6UKxk>6mu;Y;#wVI&;C<NzOLaStRNv
zR)j<B$um(Uk{S4gtEgfp?Jn++k^c@n`i+KHeG^|z%GMyiI?B-5I*RGT6(J|zwufv)
zY+)SKth5fep-fJQ9rqcyV$WAkr9SgGn|`nW!M98`&+UGu*AX)#OmT#71H04`6_nu6
zW;fcXyj1M5TxXcw5#s2&KYJ#-A6%fMSIUn$NW^)7IA%#H1;8WXjD<wHSM~Sh@e8h=
zr^s;ZePcT1?5y4X?OlWEcgVV#z&!E<{aA%vI@UDXjk)9jr||8n{<k%e+yr48V;x<i
zaN``wuC59Cm1%CLSP~}rd--)F?}~b3U4+6BHL)6rq?7Dj08?US!BXQw505(0uog<*
z$s!xAj|(jqVBSWo%Z3YlpVvL?EJaJS_$E+dF6W_7G;{uwukZ3u#WH3jbC7Fbk4#;6
zQ2b->3*AFkj$j<W+$7zj);9+VOF}2SaE5Ml>BH0b&zK4Ba<C4K5ffRbm^P1;N-g+T
zFL9eEI?^%|%y&PEG@<O+55`OdC49}3LPxN%6NPbbpAcBjr1u?X&dJc8$Oze*Hwnb^
zTFktDF%4tqE{psA@F$=JEO}kpS!0U20t$x%N6tx8y!l>5s@t!Az42l1^k0}eLZ$If
z1>`ps838t6r(vWj9H^hX=x2U{PLWRXF*=u;)ff<gG+qi7*(uOX&8M_XpYi<o;E~vy
zbSf*t9EmoSjG^9j@@~<MYPQYQn#7zLF;ws1)fI3wU}Ja^s#xv&(7Ov<m-@-^v`M0E
z?JqB7yEs!0IywU;zES49p7%k^cs-R?P}e-Tu@gDUtj4Px?y|G^m3_1@cVAc(1uQno
z2X%uLz`REH0Ch#5P^OWY(-h;@lcrXs1rxc-bT6B5{80u`t!199|Er|n{}-!x1h8Ti
zfD;dET@XaQ1#-)noUJe35r3G%&xiiyl$vyb-(Et_{J`CP=+}>(3!vtE8be)Q&d)ae
zzqR@Cy~xY-n-iZh<Ia!|uM7SZ2v^4tk(iY4mC4p<_PAB0W}IkWnP#n2F(|KV1EF#n
zc$FXQXa4zpk2UkmOS)^ocU71uda3X~Aq`M*dXKFk3NnKR6w*XM%Vfw+Y4d-m|14kQ
z){JsgD&SCh-drvKI_`<s1t6BT0DBw?$Oi@TB`<Z}c>xL<HqoE0IF2AUqj}RUH1nr_
zw+~ccL;Y1NtB_b_QN)Wc(lOLfyNYdV29QD_|B?$pLe*uhS$v!Bu?zgZ@$vD?tHQ&=
z8a|!;d%!#GfChO<d9&hE0w|vQjpzEJAZ&sIxbHy62zg^)pWj?B*vwXigVt&!bKg(E
zbp5V)?$74w2g<%ANo;}lX60XbCiU9Em91S*vYrCO-JfUG%>(JzTj^z>?U}&(^4Hsw
z>&%tOLfvlA2=4x;6DWN66m*;aao`5L>!X<BJ%C4SKu+I_?{KdDyXRkzDSB?Jr@Mc3
z`xo=r8QNK|4!s7%Pa!0qGjksqKtcT)LjyT*Ly}pl_roIl%~v08Ah}@(M}T~G8<3gy
zq~BbmmxE?4yl@TjU57#&0A1K?<m4)I1tOFl>o)(;g&+(m)sAWN^Z@|d{ER7H5n8yw
zhV+?`kC6W8k@ZBu(ky@i3ML2m7OrB3=>F7=^$V{&V)k-H0JReI2Z@09|LkbCf+cS<
z<fOyTf42p&d+RQ%?G^^d#|eN!l43%%T0e0^0CaQN_`4MZyn)~nBAb?vHfrg!nQaWF
zUVSt8{Q&X;GB{>)*xcF4DlR^tA!qut!(!#)Wa%o1q768=D~$)NzP6xG>sZ?la0bwY
z@Np<8E{;^o?|?2fFg!oIH*};HVE}pZ@Y*I-+5(&MUBGE>X1|f8aC;B>ug3sa=$^9$
zqr($Kz5EkH0ZCnbj41x>@wm_YLlY~B#3B$;!mX^VKHPHzdg1#LixBVDGO5gRj21E+
zPlH6FL~wm>#s5DJ?&i;htAT}81%qFBIHErvxuu;9flv!!ASh<*VH$GMDV!*n1T1{*
zW^V`^+G`>N;#DY8W9@@MQ=c>SpJOUkX4{dgiuBrrLx=^$1b=@Ya!7}|@**N4Pyz$s
z;2%GD_z8dyKoRY9vG&1fx3pWW@P%8q(0nW4JfAv^y@U{_>yz}2xA#~V!OaJK<A3sI
zKnh56=E{)p2~0*t24eprfEN@rdOpP}xwlS=u{EBmu{NHLYR>{9KPw<cb3;7srluxs
z<&}ToLDIfH9g|Cl@(q!zkzO_bcA@Y(Rc!*n3}wJr{vYkYkk`Hw+s-O*HM%Z1hx~{D
z`T7dcU2M>NeP=es3D9-}c&sNHERxEHM3ceil7`YQY)ScgRrzPE*hQ?ioAAvC$_6a~
zeI<nS|6(Wrl`#eM>dUjeIl4+{r%D2A1lS^{D&FUBd_I7izRJU}xDTnmM{Ia4D|Nnu
zCk+fXG00|rdam<dmiA=+hsqva;K2k)7N{*ytO`Jle9wl=m_nAnvCX|YhUS28Uo@S-
zqy*6$AWAftsEX%aGXFtn+)~}O4IYX>3TZGMK^9*ifMFe&Mcpw3$Rj2eRwM15@8tUX
zlluQ2yyk45d(0l&%`uz6a?q2hS|>Wqm(C9W$p8uB<wNZ0VmahLzn0QJ_-+8DzWh)S
z;g`en1~6?Zlzs!OtX!&r{r?D=+iG;N^o2K#A`Etjb90E3LX;4DzqH+*B2)xtn(zrN
zJmhWaW11%_=M2iqHl~&bo@dNM@Zc{-VQX?9j$AENIB>zj%$GrxhlAK}Z<?SlrWSB&
zsVj)i6_cb<pMIq^LDg}^WA39jVt2hA`#4v+RYJFmNZ|>=X>#IG)bZ&!FCei7Fuub~
z@W`(PAo*09GA<^jj0h&pZSbz<_UuYn=Z|)RnJQX2_`9j%AS%NICCxx_I3c+H;n$WN
zV;FZA=-TrYmD!E=8Oy(?r>(v}8zjAkjQ6lKTXgVj0>!dNprPuaonzxEN+?GFA+k9p
zYP;y;kA!3_8SIOGatl<P5zMRn)`&q}VU9}wTwNiHfH28E50%a%L%qWcg=q>Mn-eVF
zhq16<aY&y&E>6P4WOEY4jnk@#bv@tp@Ou+H2$RKCXlN(tt~k=L?T^&eB?@!bGVBfW
zk2{Jg9pn3Y$o!hecx7q}6Q@G}=Vz+$D3Ou6SKPo@mW{PiNoExqlkAhvMx<L^g69XD
z8-DCxyVhNJt7M2R!^+KvJe6L~JQ7@&%A24U@eosw#KP-*&?V7Z9?p?c33N!M!uxbq
zov?|%D|!y$wcM%iByejlI*)T!6V-7A6Y?J_>=dBr&lKZFZ7_0srLBSz0K<>$yfanr
zLzgYDl$ACNgC|WmswC24C0J78vI~=wQj}o1)IEJdAgsS@!p)7?VC?9^gKL?6?hkcO
zd}QvyD{k_nrG#qRTzg!$`OBe}wa~r3-@70B;RcNT9O&73C=(oFpQBx7R>6v}Wc=P!
zWewVzM6?0!8!?YJ6n2`=DLk*L&s^7q*5{LxlcU~0C2EV3lIkFkL-`biRH`u||G7yX
z+3n7kTOrx`7QV(iNO{lD?|-Zyl%UcO57N<R<U)>l+VgDR)2W;5z@gFSyZstT0Ln<z
z*~Pk(Xu8vczQ>-qvUTyLNqTuGqcQy%(Om^sDQ<)%758%qk}woDH?M-e>evdhfMty6
z)Qo%Sahw8FO2^1-IV@rMNxYH%UJZ!{h8a4K!k&b9`#*|!riC3~wAhKPPsXmoqLTcK
zi(#(iy%hFr1-&fAp1*g6t=eN_e&+Iymw^Hs?@PU@<`0Q<(_P#?Rgg<E^!f!-qIL+%
zF!)dW!n!2NCO{$yx7Dej*_I<Bq{-fDHP^b-SoV!BfiV=soHI6<ulSL@ZoMky4IAv<
zbS?F}IjeVbF%il!T_NG+3Fu?Ge&iGrYvF8_S2~uMBOSILrNP97uXy_YPH(iV_+^9G
zbl1ieGy2+LfvVZN$?gO-9FgqDbxe|Bt~6KtQdz3<ALH7%Va)AhbJ&71AM*l4n<CgT
z{Vj<;F+ZFvLwAM=oT|v<?znQhnC>PSa0rY&n~@*3F#TE1-`@Y|Q6YS^l!_Y$Yd4WX
zgU>SGlHg9@WcQ6LOH4?5>QxpHa>gJ{p}s39Ol4I-AUlKAzy`B#C~h4rc}zxSG*yvE
ztY^w;(oxf!X+_VQC_0-djNYDO;6YtT9geGk*@?~7p^8e(tk3zGI!?~|6A|x&B2-;#
zeHX>+K4z{wiP2uE1D#1)CIthHPkcRN8<9~JWF+62ErrBsPAwI`IZ0XJVYwuviu;Ad
zFG}WH!k%09GUhyCKHOb=VP>X_h_*2{SE1hc1em#9WcAVeH89a4wFr@zg!YQpT;(x^
z8h5vtaa1<UEF7#Dx<yMPh_U-t3#cvIX>9pxdOv--&8z9Y#S|G;^|FD!g&UK*StvrW
zUFxGfb8_&OV3JxTH?0$YLbzJUBW#X1WQ@1;C>cZ=M28rss829=FeDxPE0&Du-{IkM
zWT=?L$u&B%tBe~(kDx}Yy;kd_qq3$G7iY*bPNwxH<6L}@*akC`X3J>9Q=McNs@ry>
z`W~#tK2nt3YM+09jN-%864tOfMVm^<TLTIq3ZI+r;}*>ACgwClai=I&;ne8u9cHJ$
zrm{t3mKh@~M+^pHOKj8<6ljhS309qZ@|=gagtdyy$!P+>d&MClmeyI_`r`D3nhmd{
z#2sD6#tMN-m<H>+9P!-jBP8X{2E6!^IPtKYvK`?u-`pp0YVR4lc$wEtbDT2b+h3rv
zqE^m4;)b!qge7Q6<VY%{{el{>iGq7g3aMjvNs_`)SW%xF^J>&@od#?^7xgnz85Tw6
z2;d}&V64y);69XRe`e>L7#826f)P7%a&f;JLEd`5^sUa5)>fH_id3q-PekZ_H35pM
zmzYvC0a+LkMyhbyQq0h2-IyqMJRihz$)5k1FlG~=la{3FK84-mQ3|oz<<eq_slx9s
zL@15A1P3wPp)exD-AZgazN=5LoJfd9O=q7m*3#-ij^0l5BIIL<l(}h?+AaryCQ{Fg
z99dsmq$CwNoDqMr>?kWrLgix?XAtKDZLU`8`T!(FQO2?76swCiOlWjTNIdQoG8B^f
z(K>tsmcf>u{mSe#Wv*>`B{_$j3{0&ix!AKM)9l|PY;{~6o{hfFd&9|HqixLH?j7v7
zsT%C*s1#C^_sm?+%tOO)m~ByqH)O=S^l@}MCnk(Xa6aU9KNyvAeMZA<%RBegiOCYJ
z%>1~5r<;~l`Ig1d@L>_Q5qu<@WtiSx3)rX=H2E~E8T1{mlQ-3omCx+j*a_l?b6KN5
zGCY50t3wpvXGsK~M3E$>W<gU&2tV9i&3-Rky%+U@7C(%dfhpL|sy8~&xMFA%2g5ai
z%9gf=@!q7o;SB9X##^|~dxwf0oh<Z~yJL#O2mA%qhRI8+qsSLE@{I`NB*f#;){gtw
zk&IKNav$#KuH2gK6GC4}e7`olJE3Mx{fse`e2M*K>)qm?lJ7ezYMf2xsMD~6D!MdK
z<6Cm?zt}v#hZ#sCOpcdj>0&LHjGh!*sTCmCoWE6Ni;>gz6qiw2Q_oa4&RmvVO1qQH
z*hGN8tb}7K=kx1;tjQEye3DAVa_p`yQ%S})nI&S>#+*)C(yGsl+NQBQ_^X&7-eTwB
z#cB{@Y_??yF^)bb|1_cK^t3Km?iG1c#NkR5Zu9#ILwf1^;R%QUaUxj_$@Ji@x{*CN
z?N*pV#SY$c@>q6eyxz<_uPOrfO;#MES>CZ-oo#q*j*2=50b4$X?ptIR)U1^-61q+l
zzCg>Qj?Ge^X10(g+$08pS~*|ET|~GF$?*5cJ{paxZkbOpEy+IC--IJ*1?c2b<JS}E
zTTkIUW`&YN_gL8yMXQW(KNGQGyZlH^3X5Hh>ZgWb9+#r^J}P9!<RO_2Fh=7@lA-If
z*I3`Or`yp$gx5V~z_uv2PTb;-QRPi;e_Zz#S2QuP!^jM)J2@Pan4_7+)?B?Zl4VU4
zHr5e(8_dS){mik?V?LGTlVrqG%uzbT7wU^Wm-y^Zvg4^7Q?8C8m`UdLngDG|b>G}N
zJCY?^pLxvWg$o9l5v~x^ZM*og{F}p!_`FjUxz<~UsQGxqhNg$Fdlc4KT{bx%CmHLR
z_OO)}4Yw)?4*6i&nW>is`dRr+z(fK!Xq{AY{MSv$XWl>S?y$JBc<o8r$r0aUPED_l
z`B_8K=XIo2&}*rNFQR!bH*EMP=-iB+@{klB$Rewfg_Gk4a1QI>TPuEoXS$vW=Bn&&
zuXkDeu7_Kk{yr#LDAu2poek;kT2*aes(J}`k|b`fUq4{<*eku5IhUKI36e;{<00Er
zGW?7%qC!uKSLl6x1lwf9iYg%!488q29&L*ehjaJAQJe20j8a(WBWwr7<OwNqIa75t
zj^r#{UV>TTJOUTnCqB$YlT9{bZ8JQyW|OWzzd0OP6Zf{1+F+V4>YQ!XW<McwI&HL`
z{#1w}OGA6ysC`=tpNpXB5`TitD~?jk+>j~ud>Ws>_0`81rBQ8SYRlRzua8ED+k>uI
z%?@q+R699$SAyQDv5=Xq@>7?o2KW^dv=$wCymL(a_~ojN7%JM~d=ZVYXWL4=P{SUo
zQksCX&_8~$GnLj`yStv<@9-t=K7N1ypZ^&9lxWIT0dF4!AXGqBy%5Plm6;IvCkW&0
zgsML=dEdW3muQ&Xdmgg~$&&#MU;`L*bvUGy9z)y*zzJ?#X+{6DVTrq?VpCtJX8}Od
zV(|N$!QL79_h5Xw(S?PZ2vpkt-8<ND$F*^BjE#<U8~gXdhWh$s80|^K9h(1sZvwuD
zs2jjRYNLj8;T%GlTA(1x|Bvk${2KiL5yLM4?^)(P3W#RBx}BTihyf~ezze@xrl?-I
z6ooX`2SEFY5b@w`Iyi7%v|EE%9oJVE2d?%1IoGSEM)@*D1gAelmIvC<$hlgVb%c4D
z<C;9JQc6wto-oKBgml<YF$n~(F5e{y<_Fe?(ubh_`|fl(2$Gy{U+xTr<k|Bj*8rI6
z0YPSNU{g2mRWbg(QbA%o)^_F=78Z*@P5tW~;mrK}U>t+`{MkXx9dTdVSsxWlyGa0j
zXai#_BO*sA3KW#lzz6vv9fsgE)oX;PAy$=j%EknLkKRYe4q@T}4)?646gXZrn5Gtc
z?O7M-REE%pB!Ux%vgT?5Ssvnb?@Sic#mgT7TiwJ^wkD!UXTK4l9zr(#e!y240Hfv~
zxNA^ggA6#0@j^b-fW6X<j7HW1+>=cqhQ;K|R|L{_+BYCHmIGqTbQ(v;DhZ+wo!FBF
zLL7$bGv>Z!$QujNvjYMM(Fc>At*s>p2?BSO4yg2(DeT1ly&!i>ecbz~XTjDS;?yh(
zkyYf8ch86yf!~~G-yLsYNBt=~+aLVVTQFW#m<cZca@H>^dW`FR_>Qu~X+w#%;ZWl3
zqeb8_K$Ls{(gc?wnL^K#$WJ<jjw|7OWxzY^6;Po=esL7uKaBv8cMF)URnERNZPA@D
z0L&A};Q)#SgoNpUeX<V4s)8&&&$%CmUg_X`Z~Nw}=SIX^7yc7X3E9(GTkq(;iu4p!
z7EPzroERHBNjUlYC<MKUtcI0|03-%}!1wBV59BQ%3jD^n0Qny&&HI-nlpClCAhYTR
zEyX^+t&src^Se1QM^0Cl6r3{$^nr3WyJB4bQ6+=u(K}B$5v!0DHPAqt12H+$KTd6_
zLLZ>qLdb@$dadF4Z-Cw&fAG2Z0Ph2lRvb`B?yEhBO9au92tuXxCQYZyOvwNV20jL)
z?n)IC5KuLj^`jYt>D65S+Bvtk{_k02fwNFd;<nuV4I)%&pzJ>=^EEskqS0{k@En0V
zylE>kP!1mBiHR4mF*2rP0APQ!if?TdCm|t0vaq%$O`cy^P%hP41u2ak?B3_Q0qmIF
zo%MV_Jlx%({JnYmIRoeaVX_#bqM<>ge^>ClKtmr%$YpmCP3$M~4UlwHb?-pBe#ph!
z!4$w!P^{t93h#gBjiyRX98oO*c5#{YPZeHeyu+T!Dh_E%gC?PG-=xX9;pio2XD)cb
zC_jGufd9}>ChJ>stSRaPC}e|d^X+Q7AG$g^o4dPYniJo?@wjZv9(*Y--hkqYA-D<v
zq05kuhO)>H0w7a>%T;T>1mg1)?_XQs0-rj=!N%-rvHw1g%X59MJB-a^qoZvAgnPqe
z036m?SXk_sd~v|AwJ3y0Ie`b4uU%Y@ff2(8D5iMD`(Uq<ta}McF9fC(EI>yiZmn3U
zvj!rKO=hcX?Y>+i%+tPtBuVu&pKG@^_Eq3dZu<Mz4?7YMFE0_ujsb#-gEE7lxg7)#
z2Z4z!yb;*uZ!<zzSySQOlfM^i{x^7ORDtI|k<DvM0V1vSr-_jIf_W_g@&pHCr?C);
z6M<R%-SHm8hJz?*`5GZm;3Jf4tqAN@05A@D(miZCyi-?S5As1kdT1EPjXbse2En8d
zdI)4Saq={f9|D*sBUkwCNV*}MsT?G+Lb0TfgGbEQ)mP5jJ3IOL`P1zzRDX}~jL#$9
z<nl!9#wkZBwHHb%2Er6wBy50qgS6{n2Vj+d;2;EYiAQRNxXcHk$HS}r7r>XN3;Veh
zL6CAAK#cX@q;-ZqjJ%v#*Zrm`4-pTcPq_6;<<1Xx2a|GbFc$^$g?X)Xh1>Bf{`;Pk
zxEca4N?u=ZM!U8Pm1nQs09TB1!`54IYLEu20mz=Nz>%FaVJrOon})tU=NJ}DNhlT@
z7%E`4K!%pUtuMrM&sqII5DLll0nlOKL=u=zpip3t?ufR#Q}XweJL<pdn|8VEPPKyM
zOKo>|_pY?Mf3lz<;1wh&bGaR_ipfX$Fam^gYVz^lT?hRh))#qs3Q4dA1_q!P6!emS
zK2WD5Cx;dUJKOy({qGxN`K2v*(?*J2^8;gbVk|5yxM_HJxSsn~HpHYnaJBzCm?d^`
z&VB`s$EnrF6L;yaZFf-o3ZIlCkPMH{{MJ8eM*Y?5wkFdr%;c24`?#RSQVI8T(YQn;
z+nRtl3R4b4!HQ&##-E?!g0TqTl-(uI81eD_Q3*n>t<vUQ6Z7#&_@k(ilkO^Fa(AM3
zUvea;FyGh`TsWNDAyRBtH^)sd-ox0(#%VYcdo#WWKJyWXw4ScWCe5qn97o8J#PyY&
zZQYfjm!^mx?&h(5%bz&{<C^5N45C7X#X48za22r=XqLnUXTyvO(U*fNme@I}5(z9|
zbm4Zg!h&VQ-4#6^rqEl5_(=YWx7k0!<cT~3t+s$!K63!C>tm8-e||%L<`E0?uSBe~
z4XK!yG}vEZZ9F#ajYd{`+*~{6$9mxXm(iOlqHJ%dc&{9<keis*%K;Ic*Z;cSPx5(E
z0k5mvb2pub(~C)DQZWoQCW@i&d4u9|7&tA^jU{FclRq+S`Gz4fgnHE3y<&Qe&9zLZ
znbjkN=T|E1pIWf9oLX|(&J+m#O2vDp@7KPQi)62tvYvs~9HhE{(PA)ndj5KDydQ2!
z>Tm3n{lQG{%J#RGwa@KNs#!|M8j)AI%}HpFR;X3zJ;^yUj9qkSV<i-=P1xHDmX4q@
z@Ub~9M|_)1>Elgvr_yAuorsLNyIY@Nsau?4_S5n)q6!=|L6($*xL8b%GEYN-GL0OP
zE#a9x7Vipp`KOCZ=AWr@KgwTv)NcB$^L-2cl2RJ+ynL_Yh18sC!$MMv8P`CT|3Lo$
zEIUI|UXVJQnQ+zY*t61NT!iuu7a$oekhg?oOn$~5319=lqKBM3uaXj^SY&f$t6$Ea
zx#o|51u#{Fk%NXA1{y*@CX8CLR94?ZN-$^s*(jxEDQ0Y!6t=IR(7QrR+SrT&;}<*n
z`&BZL=A`cm5E(3UPd!T0D`GPi;~iL%<JMx>ZEumQqI)&4XNdKdt4H47an+*)2v!d<
z>q*#)lp#;A?-ggd?rzkl{QNx8&YdnX36eyz^fH$g-4*C77S+7S6%$!IQ9UZBU-)}7
z^##T)W6L;$hgY*WzYUwqo85YedHhk~yMt|wfmYXd-Rp?5xsFQ`9r~TS_|gDvcy;>$
zImdm?ye?$jP!Z1>Z^F<kg2=kZ-PyS@&k4tj0|PU?oBHnu60>L#fs$)*AS13#1C^n#
zGxSyiea5C{@M{=>tObwzoGLwDm1J85ICX$QG{@wUxgtRbqBMalytG@e>jfdcnsz_~
zRf5rS_v4#&)K$F?Pi47D<&ZuL<}mf+ZxJaVG%)YY!p2B7x5DLZ=quDuzhoWl*AWH?
zQd!ejmRi$qMj0ehB$%=jZ4lqZ7@sD_E=}tBLUfydI5O|?BU2iBU2S<mli221B|gQ_
zCA`}VwuYG&!>G|q>@4hwM5wzZHtx*tByUHqc&3{@@h>L0v3>sYxe>k*S#)JAHCABU
zmT|^r5jO4=E-{n%sN+=3fl=~yHV5K~WyHiR$phlxFT0q>Hht%uB9m=Q@u>~Bh<cx9
z<!?v*7-4@-&Phao*qkt8h;7a2sf#B6Wb>gQmfEy#s`pWlB$aUke@3Go)kKm2&ltb>
zHzZ5Z{fQ$J-et{rO$+;S*z4`hdoo1my{{jlm15$vogPV~M@Zc4==1Spbyuljgy(cA
z;&<WFO_2AZy~e&HY^O0rtQkB!PRyF1W>P7q#6)Hr9o#ySWFCgG)0;{pzeGgHkyH`;
zDCkL23=g*e53|jFm-hkl>+H5OI_ny~ZGdXpfV=`dV@`GB<fr1TcV#c9=sPq}6O8aW
zkF*+!(>_F4t|StA6BrZI#!s2k7e%S_^3)g2GVln#K3R#PkKl+GFISgjC~$n?==;?`
z;69HnYT#UH(%#KnhnG<LO12c;UZZUEQ;toU-LNm1GN!#B15sgETUya-A!Va+@#0Tm
z@8s`){K^;_Jkupv{=Czs=ya(Q`S!P4Y}yQX0RtvXvTDQ~BraXg)5zxAuIeNa1hc24
zxI$Ts7kTo%ueXY@M>fLmZzam`$u`Cth0H#UMXWb`{;E!0Gq)fSMr+9*2ajYm40x%W
z{89n!Hd5AaWDXr`URM#A6NwySsJrPH2_g=btVBMhO_s6wBL58+xB<f!E~suYZ#i|W
z!I0+BR?dq5`1RSbJi&0%XS#7*%P~{k+7E9H#2NhII+)KxuRfQ(Dq!<^K2ofb+_NZ+
z41g;F5Bf4Pa(-c3%d{gG`)7_+lyA0!e7jS9AHGFKZKv{D!{sLF+$`PkN0#e*JeqY`
zGuS(9Y-Yi&H`?2!5)v_rM0zh(cq?9R79B4^vP)?=s;5k3q1&;!qd5sXFk8og1Ce55
z67XE1v206)HgHmLY0dLtgb}Vb2QuyxBZM0JOfC&G+HL7nMB8&(InyOW@TSwJ*ji2u
zAcj?2xp8b#&ZdiWK90yw_sfD+fmT%jGKRn3@<p`fCO~aU;%RZG;*2AG$gRv}2Bd6E
z%Z5dK%Zy4yV7G@Ae2XQOUWvx&w3}M9cid|)X4hiD>l&%?CT~%(s?h6Bi5Gc4-4u-5
z>aKe-TM~Jl@l@F~$ckt&k?`B_#%7%*nbkKZB-Y<jrvZ0nLVtrd<*H_P(MRWCOmM#g
zr+$!;(tIU7Ec;xRJt}MGmUwLvac&mbB3nMjd6-4;J12(}M^^jO9!hqG{5+HTGYxyJ
z(7WMZnAty0$g9^|3-d2NBq~Z^2h9N+y}lfp=dz<@A`}h$NEu{qMD&cM%2F^JE81lA
z9Dmesqi2z(xYOtuGvI!v&OegO*U2$ddni6r_Kb+v(lq?+_EQDLXsd5#Uj~in6W=Px
zGZ|+H@N$HsfaL)P@gwSdX0_h7TiX@lozGM_BvDhiBg*n*MEy%7z}aRRbZVe_6D-QM
zPI-qSjj@mDjA>GQysxoYg38bHhQ53m1uJ@8b#@~^BYCe=bbVZD!@iyZzti`Of%vgk
zgkmbM#dB8v@CKDV+6j(`CC<f%n=Cky`srX^KRJ7DM8=`PvMh}pz-ducmbKxmM-!A)
z>P?)%^IX}7XNWleoj{El(ev0Rp0$x5Qgy20->AWJH|fu^h(u`dt0YH?m$#w{mN;8s
zs<}4ug2tiMv5u=p51nOWzR@;gU8>M4aV2kYOYlnHkuOzN5=hylv#5SB@Z#Q$&Rol$
zq&0ZbvyxZBI$jgb1gg0_5aYWi%wK}tT`@&)hyMrKm~l;tPNjuB%&;+fwc4M>hV5K+
zCC45!^ijm!mAjI}?*%hS?z|UV9wNc%e9A%gd`fRShh?NCv=h4tk=Fk8-eu`JMdxs{
za399xqMgWxzVdP{E%#`r@8~Jq`Jo?E#}cbgqzA2plVabZD3IfHS6Db$3p{puB-@>}
zQYh3;-k`UAE2CNX*Lkod5j{bb(Z@Kd9cqR^igN2*j#k(hWxQMIRBQA!Lx0(}8!MS1
zJN>cWs)-pIHcsLl3Kp9j9#!R3L}n`tZ<9!uxZsP6dax%+Dw$KNnHo_6tF5uD5nbto
zNh&qKv(?J{)HG<UY4f?%udFuQna5T$cBKRvZ>Q*c@Q*z!q1}c%)dpTWQjD?ZDVQWK
zuDUwZylb{d6LLves~k%5!y-$vhJ%%wE(PsI4!erU9<`ziR`%4&MeANo!T|CXdK?wb
zWC?hYnE!4XvCYB+0PaYy{p<GE`V2pek|(qber_*xEEWAuDtp<+-=;>Jqb6k`Un)D#
z6zd!VmK&Rz>~Z1*vh>+9-5xrzWiR!{C`5c#qU1N#&%V?r@Ccq&V#X%CCRcogW!Y|O
z9+g`x=>Aeco7J044O>`(oP$VHz^SlvDEC2Ly_3KJazxD%#`VxKBNHkglFCkwC2T$K
z{q?56m+_*LA?okn&XemC$hTYziYbZ+9asXp1Ncfsw7bQFX<>z%m?x)~Ze=fG-dm62
z)96=Iy5{aaciByZM=Sx#fT@XzNr4RzZ~!KX&noywxBmnEdsFt(`(h1pOIrtutVd9U
z6B4L2KNb~1oUgS=yo~=r3jbOCAGm-L41j4Q&7gjSG&<my$cs>c3gD3r=eg=c-T%}e
zH;6O(uV<iKj<^(d-qrwBo8Z|6Kuge~io^Z`f8E4GKLeGc<8Jqf_rez(2^INLqxOI|
zyn=rrp~q*pJ>pC*2SAcl-I@{6@xjgY0py@D{_R_1mfv5L;%;g4R90>%FCX8**$gnE
zNWc8P0hw3SF22174AIf9r%G?xgBUg7(Z0=>;yomC&hB%4RXPOKJW#?^Iu-ZdogOuh
zQ_xK7TZ5W91rXDq>a_tNhyR$*Ii+8cMXX$LEYA)=^tVju&}-^GcM)=O^8OTo*ocS-
zzQTtAk12tn<5^zO-%B{vymvGPq;vHb5S0u1OW-k4bI^3zC35vN>>OB7M9s|1c##ME
z{ga>gQ&~dO0LKy(y;{N<1L&Z#_c1;&zWnuuxfQx@F_7T{#MqlIH-~CchzvvFvCBco
zs@Ibu-~`cZ7zjg=`qRbI!C9EwTtfsU;FrPaab`0pa;{fbS2zFK?ZPvSskk8qxPC-~
z(6NLe@3ZoX>gq^Ews?6xFha;cU_gH|UnEc?9aQyKmjQJi6mtg!s5O2k`CA(r`&7uj
zDVy^=8KA~PSIZ}W6nCT(^=k>b^Dpj0p<X`$&e#JSBgvaC=MIh`S3)3#0{lRXiU~-z
z_uYM`5MmBd@)>w|2tlc2n)+@uP|zIYDROl(M2CPi?K@UXK<u=Nii#2d$AaA~0JRt7
z*Qf=AepR;RHq)G-;_e5w8(-HPA-UE3{Je<%!-aDQQG(nEznXB`X^W)~#SU$B$NxAz
z-90+~Pubm6Eh|?AtTgbm$0=EYAdCZC+S?G54=C^`kMfnhtaf)2=u+%&frFa01&Xu!
zNJ$)#I6x#K^@VIp0mg6w5ZWu}59PpSA8;1?z;)$F2&xCR8o<6#$Q7U+iCpH`K*qHU
zm`R^%Yl(dUQW6KKj~<beYY;LDZW1*mu){ptte%2^SLlX|yfXOvd^KF|s>bXAMCv*V
zmBXm*ed&3y)UtD*eD>(86L3bu6BE2eyXRuf6X3Q=+(!X7&rZ*cKpI$Avh(u?9bCcy
zlxL$Om^#0*@B%`L!6WT`l%p*6+Y5q@3d#pbASAOLaIAtwUP=Dl@2;-L)#AH1SqTY*
z`he!;V49kmYNyp)S{f}y)e44AfazY-wgcG;KQ{ORmM%)}3pjL#hldoMkc<s<0H7i7
zJ5ro`3-J09hdkH0pyZGGf!Myao;630@ut5+-}TK*78+%PZYsdlY^F;75n8fix?Nvv
zfP}&|(us7TH>_j1%I1JYiiwGd#11?+Z{r9*%~oq=$AnxpLFA#=uV26NG#K@q26&2R
z6Amt2)A-Jgz0#o|J7J({gAj2D-v==s$WWHY*Hp{DWBKOY-lSaB*H}6y0M0;>BM}1Y
zM9_l>!0;5l0=pzyAb#t1ehI-=W<zNMkkKvV?^3#u0vX{TM7;zd{}jmM6u1IsH#JdA
zI(Jt(Z|CPau13p7Gy;qrMjQJvJKJ>k+tk$EuzB%nIvnx8r+@!4*X!DO1i;*3m)|EQ
zy1K$iyW?L1F(Lc_IL~msnn9;aX#pShqTNd6$G?9s#D2-oUxJ)<!N+yZl@rll<A&jX
z3qlVI3k&Ut7D%U)>;9gf@lB<{!q>Cz;gOM6fDuX)NMqvQlmkIv4h#YL_?8w+{-;-T
zpmuqsRP}~*;anT|uz-n<KcXG#bbvIz_DL_5xBofrM{-rb*+4Wk{r9f<r67<%{rRqx
zGP*eQ@7#a)2rY&>`OlZ&XFiXzD<>`m?{etN%A(yPWL~@)jkc+V)h%P}E!}-59*rv0
zC&i?odfA|{ok^s@5eMpGdnrde%5yH^jW%BlfkI>Zr#UTlU!1Q(_ZLN!xOiqvp7~!j
z%L@tp7Uoi8;Z9-zKn`t4U^~TjKef43C0Y8`hM(gTHlpWcy~Ykai-O%7wQ54;YO$`x
zstVN@t_3q+4OL9U*B@aB8osPdMlNo7jF3w)<WbGS`lZqF;$@ByBRdg+rL)!6X&|U2
z$AyTg$YXh8$yT}9Rcqfr6%Q=<;W3qRu-;;GuQRXLJT~yyXMWPpYf4J+#yN=6qw>4K
z{9B1qmIZ~EobC3NQr2~8?n3RCb1$%Qn3LLra`8h7HRRx&TrLQ_j@=S0@4|rT?H^$K
z{mw~5BglKIe%)xixI09B?D@`PWXY*YS&j_&N3XMQBS|tuz$!*F$UMnU<n>l-+n6*Y
z(%)*DIlj7urIbLzCCzXK*hsS&jfuUBw|{V1@_@xnh+xNaS_YW0jPh5o;g5N+L0AkA
znQ<Y5RXVF(kc8IK;cHS!RfNJgf#`7yUVp-E<V~?1uB(9v%!{cfqa386FNMW&gX?CS
zLgmfSuNRcHy+;HS;wP4qc%vh>Sg^(tdpSg=`@1#nalQ4$H7%pXJ&v1f@AGHX8>2Ov
z;jXVt*@P_*4}WSO70x=LCy@QjjCL8t_kbrlM^Q81Dpqhd%hK47Ut@Q(sGWmKK3aM-
zs3%*v>Or}1-20G-XH2o??a0Tf{^&0V4P>PRY2o~@e@Y@z!vYa5w_mD8gR|*+X73xn
zVi<G!B`DdQa^;Dllh!Vc)&V20%KFmMhhYos!3s4vHT{mSz7L8d1BST&?HK$#K6Rp}
zZ?qoCoU{!y!G?J7w-SSu$WQuZc-3;4nA!0Xf`e4)*J*OuhsqK6(Zc9M&0kK*_=kIo
zHO{jO^*!m($*Pn&n5k}F*dyr$OrT7D+8tU4MbhpPdqq-TCpE6x<$Sxa<c)}|;?<q_
zjl=|-<>@8csQ@{eL-ds(t!}&NA^PlL9XJUs5}@Zq^{9a*cDy(b_7q?T=&wB9Jw$cg
zF3WNdb(@~LT2hES#_{>#BzUBCaL%#$(9+TRaM>42a*s@G=|1lVj~<<`BPX-+#u@b8
z6#oxtZvhom|MvZY3J4-9poD}d-9t&E(k;!<-61J0Hy|C--7vItNvU-APy*5|(hcXg
z@Beuoecp4{I_o_z>mJuE9qyUE_iyjn*Z2BfpYLmug-(j`t~FpD2(KWXd_sUa30BpE
z9_4Ir#kaY{=ac>)#iA=_%THFmz6dr{tO>E460z&cF%pR0){p@@5V*x3dYxk^R+u|6
zi(X?6qZ|ENrk2Gt{&ib~fd&l|(f(LQxDqMBZe;c)>=VtUj@Wm$=vlXLuFVA8#ASni
z=cm`~>qe~-e3-op>pxIAi0EYSU3W$a!kGDxNpHnP+6*sRh%scGx6C=BC|L_0mQRuE
z<)xe2XZ>@#n3$hbVVD9HMrUl7(8JcNX62lri6#?4YT{0&$rMk+6hSB(n?NXgD>|=<
znR=?aija<a@wd-_QuyuQ^3G#>@|CaIpk6J~6wiAZSb*J=*-5dXIuP`9gPl_|c2LRd
zfwePJo+#&78_YBA8*i(jFn1SrQ5-#tMY#bszJ4=%W+&0|a{A>MHywSXd`T)|c)-4q
z=bu(j(c+(&G|Ml?ot<Nx_vfaGC}TeQemm9+TnngCO7DxRj-r&s@RsOwv3vDU99#7s
z(5aT}G1#Ug%IWBT>+L#_jI)Mir<i<MWZMefIgAyTm}1r^QP@HNfm+Gp!OGoK1Ap>^
zWRbqO{I574K4s(LomH~7_vQECw65f-rd8*X9;)pFJ4|=!NK`><$gS{x(k?Y0O7T>9
zTZ+2;QKFpJq198Ndwg_aDw`$P(Z!5-F~ZnZL0E?YqFkX>z}YATY=>w|=eK97fW06k
z`MEWXMB=Y?XZBH0&huux?5b><6vyx6UR^Ks;-In&tX_@Y5Kbg}C5wNcE7|X`@X$%j
zXS5YtobLspzlk_q?_hLNXOj(<!be^PVWguuI~#`qyTsQr90^(=^rQNjrJcB~7h<z<
zj<?k)vf$xz%9J^9_kj2ag(9ncULzlS+O0tAz4h~3jxiWz_oJ&~UuB!yTX26Z)F+k<
zrbh@cSJwOpH&j1VI~e0Om+S8PsXmW@1*^vfy3|#D1R5xng0y>mdYlnyCsp20df!K+
zLyw-gxS8Fnwn;L0+kWr11)h0@Gp=;vvM(E2%Nw5G9Y^g)-Hf?km&&HQQ&wfD^X`*n
zf9=&F_nf=yg&*ah2)uorRMWqM>P>(0JmDniW=y3l<&f|2y*&BZ%kAIZ_$ASUzl*Jv
zexF>-KX2)ERF8>&Bz5L}lQS$e%O3wl{e~V}DNRrz!}+?Ak!j^i-#BEdkm~-Rey&5v
zDpmNm;^HM{Xkq8N(2<xZ)jwsyl7KUZR0gnR36n8e<+P+=WN<3UXGg!VuCppSgwZJs
z`g`!bCVAApH+l3t^aW#U+go(JC?3VQUXi?T#ch3qx0cl9v9BqPr!+RGEU7jMuwm+S
zZ=)rYaQG2x8VQ?RY~2zH&vIDkE2~FB$OA|c6W_@39Eym3X`~=d*<QDIY6Fdmi*uyC
zZU4pTPg~|^WZNP_2eir8b{H+emZuSQA$jOr(GE(6o*#hMhMTw{4z(^$_<Kdc`$}wA
zQEJ2{=)@Rd$9rfQQ!*~Kqpl1I)3p!m-vv5;4Cfbtkk_%97LQ6e%fn$YuZs<}dMTDD
z-(g!i%6!T<^hs@R%NQjG-4tAZ*^tDAP}bI%$<|TyWUBz5On5eiD|L&tRiBf(8?LMg
zF`A<|)w$23OYSyB>ekmm7VFt>{p>`_wvq$|CVULl32D)NpGSrllq)&L2e8K6(vxz#
zwd<f+{&vhFn!qXO(ax|{&Uy&LHugR4p6{yExs%zrXWSiT<onkmPdhKT(i9I9lPDhJ
z-PTqv*bIyzE+KfxPU4Pl7@=+nZX7uMLxL91EM9{5jagx17{+?_LCeH5Gu8g5J`uGS
zet1~GV)>Pc;Ztf1W6h0|gF_WDGnr+eqjV=?Zie1LiFjN)tVxXER2mF5V(07;y$wt1
zpBpIO%M5uP6xf_EV*6<=Hxoyzmu%*9X?SgnA~%AQORQ2R8ew#Q6i-%$v(lvvY<wEg
z6c^Vomm{~d3T?306BfCCs|dBibZyZmj7L?H>u{3u7)WDYM1;Q)!tA98Hng`1_GQE4
z8L`QM&%^|6wC6Q$*`lCbdYB1BD}`NF3*1*~Mej8gSI$Ddqg(ej4Jj?FjG{rp;53p0
zk_*IK#IjWT=F%Kxql1R@zA+mWbn~L2CI*G!Vtk#$19#*K5-zC=&YY@Ag^9p>n&kPb
zvI{F<@T*yLMhI*C7;Zkks>Xn|OH?YkUy7eZ_Z13xgS!wF44_@GnfM6f_2-mNK0IV6
z#4o5G<wzuQKJ~5h?+-}fN((%n(%D(N+9;reVFq^7CF3y!<HK}ScS7$N<053m&Aod*
z?_62vUQO?9S;GA#AATI8sa#IMN3a3ko_?vwaH-xdZ-?#<gAOT8i<?&qvdFqcY2#5Y
zKGsK*LqhkXgLS*ZRYRvyR#=RIl=Cgpc2qel{!rIK96ZtV;Po6Sl>n=KXSQVmqGKs=
zvi3N4o)`M)QtZaU?)$Qan@LTCAFCny@9>FtyrXLu^wp0HYfYtO3b1FL780`)k$o;Z
zT*_8tg9T@oRKciHJ16=aXC`4nhe7s2C&+IlbcZDuAISe5kHGOUZZI@#_0i1)R?H5$
zVWWxPl?O@hu1)Ar77iX9uza}E^=N1ew_ks>m9SueIm+5`9;|i|u2#FUyIV9Zg3f|v
zCO2*m^lx};rV3Ty5cZ~bTE0ktP*;YUkwKxqluaPi_N}R@d2jTgCGJP}-v%sli72Z-
zCBd7^e<=Tt{`nY~zMynSP<qe5JBQLRF3%1+1$CXLjJ3eyL-5dl7XuV+HwMUN96j3~
zV1VVq@eGs{jOu0HMr=WfSQAc(oR&|ZEc!%XU$Qw0X@4H_gej}CG8*J>{PGR__}@iC
zTl#k|Cz_cCkR(0Z2@3FDP=QGUqXIyPE^HzHolmSXAH8lH{v#3v3RQb18C`(coNg60
z?RF@p0ob+RWqPOp%1Ij`a16;=Mgu84DIc!SfKdhoGIfDXVduI52+!{z5+3r7L7^Pl
z+S;~X8a)IDW7TxJq0|#F$<+(>>ZfOBIuY?w|GH!#w#;gOawiDAe6P7`X<P}1xUAq6
zL?k36FRej}0OTJ4b#S_CL$df^g*MIcS*q2)n#p0FP(Oq`g7&Oh^ZsKf_foZ39iqJS
zfia@GBc$g0<kF_5R1j5=4y7D|nllQN^C1vE$hMB8gcBmFuGl9+mLOn-Vs%)R0X!pq
zpJym~zyoA8*O&+<jt;<xV1@Mzmz%r$d|5w;z69)8NnzpJ^;$N?9)K8R3xe)}n1VgZ
zG4FVFaK;Y~*pd#gb=a1V|8-Csx&?1!YyRIvli=xT4HUq1x&!fWXFN_ovI(%I&;pMM
z7m7H!Cr1bY%PJVwJlkUa7oeow5C^%ujMy;0LcuZ+iWY2ICy=xtl?Tde0p11_$f589
z9Q%gqYM{^P9S+n=MpN4XpO*o=WhUms&B=#BX!;QHUH}B{HDc=nl*<VW6fNeYe+tvR
zUp0f#R(RAhgzo7QlsdJ=%Je&2y7KZ5tgWq=dc1B9ESrF8bm#TgM%ROmSaAgr@K3(J
zzNgKDIg-&Gh}_YQa-(_65}l@`ni?)hdV+@W%3Kdj0ISGM?{EQoXJ^ObVrJf}a-6L4
z|8_3TB6sl8Pus})S^n(3<0j8EBliXk4GkAy^6QLsa&~Ukzv}1^14bpVFemF826P^G
z!ort^rlz_NH%3uu<9k8T-lCrjd`@~m2Z6LRQ^4cpc=_@G9OyDI<r*$@?Sk^1_0CT)
zAhuGZdqr=WV%Jf}WncfJt~BXSF&=m|2pzcEEnw}0_CK^!{q(!QiW&r#rTMaJFb4@q
z8X#8GSWgcH#Hl<YbNPu1m|y<-kTli|f^bGBCd6|U7l1+3wU--uUlXvpk7Ea+8C~GN
zB=|461;nYMmRXF`%L$$>zt`*k@3R^83*q*Ec>yTe8v4DGjgt9Y*?@?68{!N>vNj0#
zSpq_#YwJ0}yaH%4S`#ZEo5tjLdkSn7(dB-8iwGTBD0LS?K79a6I00^{zOe$Og3{7X
zr`eZbEBACyZ=L*bcmQ~2xd?=RTwGka8n8>Tx~<I?wV=`bITY8XQ+c`s+5llzHwab%
zfcDyHO@W2TcI*W<Y6+1F!ZlOCW2r~*=4>#cspWjLs8ja;LxCaxcLo_miQIg#1{!U7
zDPP2YU&;JV(+yjiqIV)Pv$K8k(%gVmlkoaxd)BqiSk!a<%66RixbY-$aKiy~+S#iz
zIJlnWKf7Z-k5cODQzDW}mH*SyySLw|E0$*P=y-IZjdt-Ji->4l<2kMerP!y3crR|@
zz$8Q-8OM-i<gq_+pR>#_Jc)2{lV)G58F~02eQj#c;dr{b(4s`8s1R8?u~}W$7009r
zooWD<-1qHs(w`6Ng6TfheEPGNBHXaXYU=y1*#^wkPtZIzbjcY!&ARjF#G*KkTpx~x
z97urwwWFu@myp=l*R_=O*H;h8Gt7!V7wNlp-Z2gXq}x(~S37e}$?OB%zFH*&h7>h)
zIGxkmcVF;|lcP4Q|IR+=DzpoEe5O<EK$1qfXr`^-=eA6uu9s&z`8t_ylrr1HWQ5m!
z=HSp@R3}#ZwImuqKw8wadaI#iWU;H0c+yjfjBds4H|SQYS3RO4Rit><4zr%w4m6$8
zNX@`s&F{3vjn0Ziq;yw4pYJU!OIP{w$|rjjmn_6|QX}~lUAs$WI$6sr1yl8f$J_&}
z8|6bozx}v=Jla_<KflvjPo@(~zSJTDmvd<mLqm<wQd+arNwcl4jpS{0L8RA&1!+sY
z*cp=HTfV|>-Y2{(h3$sM5v(gZ%HLR*yzu5h3FUY0-i6%9hb}xQCLj%<s^768*a63o
zX+<W@?T3l6owsnM?{e&5T2+UU=|AC;v}imy{W$v;r4ejNlCkgPIm#V>e1F~1#kS1P
zhO=5_46<&qLJYr&$>ealHIf8X%Cz01#Wze8tmx?!Rv4jR=w=>|@Pkienj7Mpo^gnV
zU!CPO?9iMYA9*ZiYi>A@y?kF&naA0QaUr+dWav&u@;ULQe)R+Lygv8%p{r0UW_|3X
zrX@{(Rjc{R3-0bG>o3D~(Zb!gZ?T)F*4GUbZlgp>(`_-rR}Af{w7v@;%^sRcEjNDB
zZ|U1|WMN_~>K{2i&h_6ble~EqEr9p<SJ!ggsElFCidjp<1cInz#82DnEAza9nm^eZ
zELZVY5OqF+NsPTo+40E|<0#@A^88M9zG%k-6jtP-`eqM(;;riZkGrm5-?|F<=5bV}
z5|>Nv9gegx-I>mgrY>f_udsJ;rzbn9#-*6++vdnaaif6WZ%=z~o(#XkB#n<!5fm!f
z(R0!dZ1oY8I%TX&&^C|NWfuIWR3M8(pH?@MhW4jrL|jmUjaq@2Y}JUmsVn01#3&gO
z{)Q!k(1fRgyQfTTxhIWl>Ai<tOV!Pl-%Urvh`FNa)YE?Ui2}pocHSA;XboQn9O!N|
zj}z;oopXH}tcaoY5M2};oh6+h!N(8v6U!f+)SkREbN*DaL+7=p)*}9HAopT?q1jNV
zt=5PCK^;0y=k1-QP~&sL6Dj=Pr*>5VlL(n=0-naITMuMM9TUO}F9&}V!6gGqjc<98
z$g{1+^T<3-e6h+`End)RO>CFktQ91s{&fLthj}|G|BlnvP&$Lc?@F^=v*d*Kd`X%h
z1u6Yh(W2a?lQGY<Jz<0Bx83+Io`a7L*Ri7!-L3rs%1ciI@G*G!E6315z(S<-?;+`{
zINi)FuUQPawve$AE_Y&wi|X#yo)^W~7lfm(tJ-~^jUIUm`{<l2-+e_yP97MOZCgJp
zKaJx)LQLb|il0jTJyKjjEe^pRidVs=Lg|eWnv03vemrDaTBEEYHSD@P+*cGsM1Yqy
zWxdHi@<W=_=20%DMuCm?NjEwLj}T?G|HKi=H(u>;!sW3~gz^PC-ZcE97)+}@-FY!h
z;QhvFX1>|@;<#QBtAwLz>Un33s(sG9aJy^HR~=rN1ZqCM%I)I8x7d}+eEl}p@`#w{
z1dOQ@3<tMKhz~Qp(PkDFR>_`ZQ=_DE%$3?qtymMY;Jg!CU_~ZktL#h`^7hpSrtA|b
z!2OHG(M3C6V~tfCVvsg#yZdR3eOF2nmTu)3!6{`aUj1;0To)3ENGv?)-;LPI3wa;?
zfQ>{=qrZ;%UH-jp-gY^bmCikUrM>6!V+bO;L)=8aGw0Nyd%vrwrJ{dvo%eda{&}YS
zoqsF1h<k(qkGq`&cG9WDKbL@r6**1ogvAOc1QVPMGxRE7&+zT(UQsU&eu#O0vrGNJ
z@t6(w_#u6S4;DBWjNzj3Z!2M#aiRN!6RyalPZHm%*!mu^8}gxm102cTH_a7MP5s5t
zbo~sRt;KiU;cF)_0|9>g7G<Vu^M%&~+n!X}Myx%1OGZhFbz)XZynu^ESHL;D<RIbo
zNzLyzHe>qJlU#=}*#P~PHYx3A*&KK^tu377rZw7I$94sr_&$uKJA6K)%0j+kuqB@6
z-PKdL6Gpk@MmIipfeWoCecMkL4sAVm%iuFP-!Wp3ok$H~`I=X@>Xb#dovTeSs|NLI
zLf^oijwhI_(E0OZaeS-0x|s3w+-`R0nIOwtV^>|*CHBoC9?uu0>w`t@P}upWh7qyp
zys|||=$0wlY%=oAu*=jIp7wgCss@keQgnRR!8z!m)f=YbCy}iz5tL-#czd@(g!&ur
zgn;ew(HTyF6OKI7k^V&o`W9o_@w<V@2W;4@^u+@>Lsc*b$<z(!*u6JGS6)FiVQ`sb
z#tzv{g{J{8wgQYAy_;ns+lY_3)`q2E8taCX+Y6@mW_YSNWJ-1QzV<p4(;M3TNRwfW
zA3O8?zHrHyYD!&TM-@;epuQ<2jwk%2_lhW;$he+i4xHBcG}eW7_pWmG7}+otk{nNB
z6@wpt)R|M}PC>neOERM!LazMbqf?yNkIa~Er<u7ThJcFi42(xdi0)Diyf!XTI#Gt8
zz!5)>Pnk4IdN7N9L~}L|4fER?|8x_9>(7;`x+*7Zit8)eBatM=Z|FN<YjlVz3bfV0
zsU&8$=&`%8gVTKveR+jvbFE~btWPF*wRAO_k4W&1MLnr3AzqOo3&?({WF(3p2+K0W
zLu85Kns%*wXQdl+Y{NHCta=i3vqA^UseN}dR=*k28%#Tuz1qeumzGSSvO_Py&)nzf
z*j;%r9UcWh#}pC%s7xyfNPjv4b~%6kq$hE3+$Y-K2OJ|S1}Y3(y{zA<lfe1eq(<M-
zMdYbmTmORH8q7--nRPHu_gwp=N-S>~TD|Oxo&8Y|99Om;Q*}W8b1dtd!HIA}Bv>%v
zx(V8D3M4IQ|D$u;y4rD;UH(l52RoO+_c7P}8q0z~w;qy>M3-q{hL7b>0Ns~WS{ETQ
zHK?CrO#K-)T3k`cFJW>2sg$^VK^WY93gz?ga=--b=WdRvC;hBbI(p@m9Mr~_D7p9y
z_ii_`HTYbwG_Tt!nerCwxfmu5zah`uH%|(ru^7cBa3kzKW_zN%UKjJRN_4hrWAC1F
zBgqx7G-f(sZ;my4ERuDAjbI|tf|)I*w2b~`m;P%O)L?7#*e~XlyxWE!2g;f6Pp!bo
z-Z?e%U7+v?8I_N=jV86__~{Ih3M(~?saB0(uP}gB-dJ-ZA$Nez!{+~?KZo&)UxP)9
z0j*MIT+Ie%+_dPnVxeS+$rp|^@U1+Fi1FnYHwmEI`b0BAe==abb8IL@$153HD&SUu
z#-A%$oQb<?o}A$J<Ja+3?lW-#bESxxJJVpHhpt_KzmE%gH!uv7e4iRhJ|y=%SU-H|
zz|N+CWt3LKT%_ZkSsn~W+-l_NT;JOn7tc*_hhxq*ZRwp^Zo`hLE_p%B^D)8ZV{r3j
z?-i&T?Dj*fe5}JDdPIgdhW*S&lLI*-^&ET%`h^irHv^>#9SPmHITO93Wse>pofX28
zjIKU`%WO;|i7V;b+D9Rh(TQ@ga!H>OVic*xtrXE69wQ%j5RJNO46y1aUuHE&f0XlW
z%USlp-l5zaXuUN^^ZU|9C?P5Ln*Gq@&0yZ6_g$ybhH&J)^}TzH2<fXSnWA8)aK(hy
z1$7m|oI#d~*AMn}KXd#13_XA~kgo>42Po}C0g|YvMuO(7#K*t&3}1#B70gmnw=+j*
zUUX=d9FA0<@4Gba5=%3L<#9Bx!BV#LIYy;78S#d1p>#e`-4AawbU4!obPpU5e34$M
zdn!cn{(s&XQ+E9=;h;u6`@4yXA>s44qVxZ_t^Qwr*=^X@>R+FQw$nKKwexP%2{3n@
z-|o4Ars8&{t7ms1QV4j(HR=l8pYo{9!Buj$+Zh0B-pzZSkZxP>mqLRgKquk1{Q^yW
zpyfUOb92w4%Tn-c<^$#Y<?71~M_!DEzbqq*qVL}NScQrC{+sd$2FReaP>7!LwNhx)
z^=Mqp0f3#}Ee6q4fO<L<Hp>CwgCO+<46d1iK@EPOXbi%g@kJpC1sI3!H?DG)LYXT6
zCXUWWQ!YR}9jFTr&A`<!1~a%!00L8SM4e!Ccef-2X4)$U#2d8Izvf|dnsKB)?Entc
zYQ*(=L><82K-sL{K}9qPHj^E3H8toZ1_qg+`7$V~_2x9<rVG%p(HYUUn7^z)heBzg
zkru(D0{J~KOnJo93#xK<2+i|v((JtFOGx*CWSMG>``%E-EtCuk;iMsQHn=7HF!7>z
zfal+=RA?+`czAeCR<~$;vdR6#E4{nBn?Nh=PesOk3Vqx@_Q7R9(s^Uq4<rxS0_Evl
zn+lqz<m78mxa;eOWnl76=j6-8Ua#&ot~ZE&#RG;IFt!7VH)vD`%#Hno_iC%F|BTbL
z(FZ&uAS{4+N#OU&%F7FA`1R!;1S<XrIgJ22=b=9ZG?)emGAu4L&4Bl`n6k73CUWjc
z-9V-uP_G%b3?gKcb>6xDSq6(g4<3kR0z}h=Cks9nN3txKX-Ru&EtsA+WpztB!Wj6b
zvW`j(KLV2TpvJF(R6f+nI=dyW-e^jF3yZ?!waH2Klb0o&VBE!pBDo#>`xzKF8uGk6
z`Uq%T_gs24=PKqaz%j+F`cp_YhLXfQo3ZWb>bmQG@~d5A=6==AROK3c^Z~Gp{4TNe
z3<1GQru_%quIiSF2?;K*kDxSOxR#cO#>3qxY32oBI%9`IjHY>kfM#ER>9bGpXNy>5
zS+;6Jz!cvG$XkU*pBA}I03yiatXJ0yd1z>82>DjP?3UwgbFjjpu}c|w`P=V+gXA61
zpehYJaTd7&-RrZym)EtqE0yFMe9>8X5SZ(PgnOJc%&&kcrX4So>qVApA81mn<&5;%
zpL0+LR|omNB36p(bJMW~REa}uDlnVY10sw%!TjW&srK4v`ZWh+0Qq<1H7yPCmiG6i
zT1yhd5=VjDY}u+s<p4BQTn(iXK{dP#b)~K#q@H@s1C0+7>FTw9nHwJ;9Vrwjd)?M#
z{OkRLbq=4++<*Dwca<^l@R*+L0vWIB;soHgkE^PHwncM%(F~AG&<yf+?RGbdkc$P`
zlvP|WbHg^#0s37ZZ%7yd4-ynZY@m2-df8%m+LNXaw$POPS=%nIZg3&y^O?7$6aTWS
zqbR|``P{fK1pw7x>{wt@<u8{Civr75r|5^L#nlkE+XIY*idaHJtPs3<W#!JF1Gfm5
z{Io+yAKNoZCiK^ZYO(%*6(v?u955Pi4L&ry-(3ZK`Jboj5;bBftI_C3IkFb_!FTOv
zZBr-Lz@vIQdlwE#*E;rgyndoLgQm^eG>5msceMMv7w!h+3i4nGJy0Z*SUcOhG@PyQ
z*jY$v<qpZ?NJF+!Q;{2~$C4XabAOFoCSna<7+XZ0??+s=DBX1QETa5#?Bb~ZU@eqn
zu={h{o@9XkK|N3MKFbcx<Ra<oc4C_W4KkTE7hfS6St75Yxo+VDJ|g<Vw_zICYwqf?
zq#4ER%Q5l_XMSc$gpA=$h}VnQo#?l<_R<;rcF#*GYcDz9U!O=an_}dx;4I9!1eY9P
zz8UOPN@`4BDE_{Ht(l&p#8H*~k1@-mNOrT_N#a2m*<qRLV_45#e!)R6(k8Uh3_p8H
z5ZBZt=Sy;REbP4TR@#UeKIe@7a}B-EN;E@c9dLtagJLQ!OyQ4YZr^WVi3(VQJB2pt
z+$f-_%t*HnDzRIm_+PQ1J<jkB-Oq3E6X%O=<1*$SmCxrV=E|}*d=gFV{`8>nL;F^u
z-J{)2qq8fF>k4<7(asM8rt@y>g6|BgQrmO7^VmukZ?5IZ?x5~9&74iQggeN7NsSBQ
zrheIRRWjRGRb+JOeUrDwjU&$~Y~F($TGN?bOeY{0V-Th58eSqGJPbM9Vzf3rom+fC
ziG1hiN=1Y5qOp95@Q_a?4~~nE#Z5W$>@GHv-Xz(T42SKISiCr+;Bixj!bI?ePXf`?
zOz*-F+QLtTJ3kpBP{LG$<7CYM!7^xl((jdfOcQ^k^yEyMsfrgwlW+<zsFLHfj^q`d
z>wRZr5*QA6QT=5_u#i##4M{(zCm)|b_2XV6WIALbl}~1yDTHaQ1GRy59Bx#j^5VBg
ziUWQtV-RP`m^-()t3aqmgrxQ47VnSwIbkcMV~*9k{ONMIEV0?(E2K5^?VYp;r#LCZ
zS5PFC@)Mf}Ov!~wpUbx`s9W>TxzES_g4LAJe{kE>-HUbsdO^M?Vu^mHDEiPO#8KuA
zAGn2c>OH8&@NU&F_SDb@wbzj_n@*Wyitnu;S#axGOz_(`x-#Z-y;n=K#bEN>Al$u3
zeT|!quMcUKl!8-m>WXWcAic@2ke(@|X23lvtP{Xt-Bxo>`C+Ifyo(4$E<>XDCk}hR
z?~%6Obmt=N8ILok36N;G#HcBxqOUd|x(IqsJtt!M$}7&kGbHsqv^oc=m+GlOUKEq{
z&?rc@C+JH2UEREI#I1~v`qMRpBd+1Ku02mcypV~O$S5_X`YjGB(kMgo^cooDKm}ik
z{mu6!5;zf+lmFw_4@!T#yQmOnkX}yi{$?rZG*1i^b`Vx4DVHs9;c;_%g3&kq<UrCF
z6hgOHHMW<=&tielMq$*QH7OV1vrptK2n)BaqFWpdke4O0U^cd7i^${XfCuvmtDbF3
zJpl)srW)OlNg(l73CHtt|5xr+qlCKdG2cmeSlZOcly*d`6iGA)O*rx&^0hi@n993{
zPpO6}QU^GNj}hZAgs~FFaE~#Je#7Ny{Z-Xbvr1`+)q1H@=A)B0m8<m2LTKQ48ejY<
z)InmXH=~E`%_}AN@s~jHEORf|4Yy&t7btJVF#6h^Htn9uZ`hEAY4Q$)<@n}tI+SSC
zYMbS5492=x+ubJ*6u`}^XIeXmHi`?`VZC2cH@CA;rRNts+D9jCQ%~Ag5QEasrDSTW
zf%a9HJ$|fwz)nm5u{`CyMqTpnd~;vC$;heC&Tu|o4rAGDZD+91+E>j><H$bk&DkmA
zOj-z9#nzxJOf{)xvdFa?c;q73&N@RHKVvQvLmk8JdTR&ZrIeC_(`mAAE1IfHB*|1i
zfx}7kC~SVGyN9n;(Hi8B{7iEbx*@3|r0TU;+T&Ge)#x`$TG6dcD*4>~ewCb33eU|!
zR0mF&mm&OyLv$s&f<vJAOM<QwVp26V2Y(fmOrMz83FF`k@q~(S14gBq{;~m>lz&`>
zRjsUZ$HBJE;IN!wh0fPEdEd4$biYl_-K8ozt)E!Xot9@MJ1(DQ*n_UJU$|e(*Ld5-
zNvUUSuc<#e{-h#EERkYtt6?l?!#V@!ol{8qNhfRjIncyh=*VA`UMH>im@eP@ohyke
zN18EnK_xX`8xM9xzkH(oqj|VXpi~`4$6Dj3iuk34bDL@N(DEnqn!NLPKxO`Q-BcQ(
zAmy~ClFKUo^Q?H_etD%pt3(aK7UTU<xP`P7`%HnIh#kH!l@t0J$pfZe&$j}qq5@OM
znXA!AZcniZTr19^96PTDSxX`B!PN*&_09<SBUPvH`YZEbU}cpX2*o}d^xvkxmFMDv
zQ_I3H!qXRh@FvN;ZD5+ComHHD!|_65jU?^(u0Z<G0Z>-dKR?iQ?W1x!=2|6UK770H
zP$eTW({`svm_gP+iC8jVGH2FvESf1J7C~sVaX)x$luLdr&vo{zhN>Ht*CSa_vAW7{
z5ONu>5%ElAFz`I_8d}bC_djE(P7Ww+=)xchtsYy-t&i3rYi09#Wg&Y4W<YRY4`UK!
zH-V7gn}|jFR7qBYBO1r>%*!O3xP&q*wFAX3O`a#3X4m)~H~0V=W6!J616~XX-acca
z5)z+j>wCY6|D7>lO*}5VVgIz{@EJUBojsh>+5Cgx`r0Xt_KI369b!7<bb9I2<uhXz
zX$4cW20!iLkv$J%ub4JCVubhs%P;CJ49Af>CHO!a6cv(|3!of$D?h_1Cab#a-o5Va
zHjOCvz@EVO=Mf+@9<n%gp;L*se{V`Kp7Rk@bNl&sTzhoFyAaDxujJ6oa=GPHxG7ac
zuzsBw)+7?UFQV~MXKt?%V)e8QyhRfEtHhCA{<^F;dC|iw+?i_>D?dZUn4ntwql=o+
zk(gtVtDQBVRDi!}454HhC>yL8{L0-$$}qKc!Hfpf&jyO&J3=u|iI32gUp?(TnUOE6
z<&?KYbUnqg<PK$IVie-4<X4{uN{_k<Nlqur!-kl|{4X3Vqk`pUM;_lt$u#KD?TxX#
z;)-$1W)j>+5UCxrFd8;Yr4rFwgKPIaOMeoCrCzH1QlI-`Ojy18xa#Iy?7Ahrc)7A)
zmo;u@ZLYq=EEgO_^vcI>KDQ#Z>EBc4vlbp4;*BW#V?^<I)s&^G9T#U6SN0`SGvN(y
z5zx8AEfB53)!N)^>%DPh1NIU&pLAHn<<7Qc3S~M~aC+}6C{qx(`&b`HIItX~qL)jm
zYm=F<GG@Q+qvpi5>~cc|dW2-kN_@XIwlO^k?%}@e5h<Z5Tur0hYFe+bGO%6Yu(HqD
z;?N|R7GHx-)%cvJXIJz#p0hcQhq4WtkjCX>)1&67XxTK3%iNvoybeRZ>9v#nqA5d7
zqu-OUs%INNxT+L$kQuSmhaOn3<5I|-Bb2O^6fM8(7|(MS*~Q7>M(^E5*?;XHO7Ji_
zW*~ZCZ!)JUDxXhYo!-)%WbQpvM$B{gua+PC+;z5}8|Ss`c}g5U^L5TmAb_qubMp4{
zo9;mcSq<<JNwndTaMLL~ma<CNXwx;9v$+mga!j*T<oFo{u4$mvcSq&iSP>vxDI~GR
z6ur~3NzsqhpQGBAD7y5p!-$lOJO94&WoM?w^DGbfY45hr$~m)kC4$h@ZeTc?$+6VV
zAbQcf-yp9W^D4kWhV`q|8)v+O5Tm%xnNejQ{vb|{QJftP1R-O;zM>^iS-{!XagX*K
zzM75hutUrc?AchsI#ASzMSd|-aJu#3gL(1k`P%6ujp^Bhqkq8D2565XkMS#aG-@Lx
zN)`&zJW48U)0D9*Lzmq!s|WLSGj&a2+|)9@;+lAE)GJkUt#v{Ngb*4A6n65xC$G9w
z|92FM|3i7h=l_?yac?iGE_%lgPwpoe@8FvpchLdZT;tg<PQ_3PFPM^nxXpm5a6&YS
zFz<oVub?px&<fRE{=2D8Fe?4L>)76Yr+UI3=p=hk?3G&6rpje^`Xy$V*pm4M#Nrqk
z9d&qULHk!_WyNCgsbtvfcjg1nKrk>84&hgmZN6Z-9E%H-0wOE~pOxd^69Ha>P#7Y4
z70{J`{%{zC_#aRq2jOUWdc(5<q4^4^1f2k)h}Cpe5HMWiU${q6dhAE1FGDC3kcY~o
zre%i)h3cjqJ({rM>YbT8<^pj_0Jt)_TMGhLKWW-F#R^?i3c1gDD*mNFtZB>tajP^s
zp0b$eL+BheT~E&R^Rc{tY1IKjm-;$4=MzGP)6g^k)T4)p8Qas<Ay5<&v@^lfPXGv4
zf^0ZIFR0t=csf;WNf1w>E<^NRUI37w<Pg^c<R9<@y<lT^cMXwjvj{LU%Mb{OF1{Bc
z6aN5NyI|hvKFBtDzpKu-krSN;Yf&v~pTFLj$MvTG)%fg+9SMXit_1(D;P%CXU(ux?
z8B5FaXnZhP?(dCh%70&x3euk-u#Q;?51>%f#Y$bw${rhmCEN%`m#FXFLW20;Eu9ly
zFHr2?dIuQPi-0br<3<N4O_TZocHhtl_$}-2-Mjb0zLyEI*Z@b3$Gj;Jroar&ziDRg
zKvwg+VNrr|Aj|-M%VG!%0x+q~U9X_U(MkS+A1GXFH8VC#V9IBXWp+*Qga?9q1BCd#
z&v7?IZw98A^7%C|Lgev!>n|%uyg$K_Tnz{-zG^1GLQCpjWxt*`mX}Q)Tmt}YC^4Fi
zmp2&M68<PWB0|Wrp6~syJ9<D#yI_)!@&n)%n>6|67qXW9!H*fz;wA`MPDLSP%ki(-
ze}lE4sWo6kLT8-yJOPSNa?8y{%Ln8aewBJHgcftl&%?ip?A$fc`xRG+(9qDb_q#g9
z)w8c=(gfVoIIq5bMM0Sd5zT~TWL`-A^pt?=xS*gQ(|!k#Aja^y&PNEqyNrOAM*?WW
zm+Rs}2F}iLKvTMnvPM#d|D)me_j|547xOo}GjIQR4fZO2S9mn61J{QfSl|oybe{o#
zg6;eOG^A!KE5yas;eY?>WQDQx%TGuUjtr901`iK8P}Wi$aG<a;u-?Hhe6C+5UR_&D
zuOtf~btVM!hROB7`Ge{L!MvccHJhd*Rp<dR<WYa;vUmo81hOmi8m+%p4}KG)uT`#M
z5J0xa0huit(6EA_F&SvkX|R4EPw2`VO4*(RLu=3=-2mWWH(wu>zTf2nQ1rRZr}L*`
zT-MWVfE5Em)gadkMRYg6X*0C8AnVF|ARYq77WnyPI#23&`}S>?b*j66?e&!nQFpCF
zfW2~=t^n^}d2s@U(x$I2K%AGm#<V=3C0%&x*4mKHpLV=F);wO0C+Fvk04F}A+l0E_
z1~w-^1s?zp9B-IEL=OVuqiOMyAcNZHWniT8=-}@J!nEDr{wea_-hH{6d4CXLebRg)
z5Jj0eWd&uDMIq=QVdsC^M}pkpkbUHvMla;Q15l^eRmfss_X7J%b?)*nkvT2?&QzQV
zeQch97yX}$W7|E7h`*<ZfKz1V#7%ZDs@(m=2J+S^S9jIT%Zr+BC{UmVdtz92IJrEB
z>v0vr$T`7H1h`WcHx}m>i>2ijmF1xhlmrDWMaM`LOsK?WEpUXkGoTfPKcUaz!;zgT
zK>+u2k{tPur&9VUtx8y7&+X#0Yb0mIM3{3sdtt=~Mnl`>Of}AkNtJZts>x1e%Sew;
z`o>;RbNkK8Bj!$rJnLkEy<6^@2dC49nWH~XuUomP)W0Ph4Smh2>5Clt&T7inhqzBx
z_{)G*Z6W`XmS&v{#mj=gHKRT`E#uW9b^$x+Acj`EXs~~xb1-sF@QTfh@K%SLX!oV;
zynTyGv@&c*$N3ceQ7)?>Ri3~RWc9{M2{#(oMewMuzlgp_eM2oj-<yX3W}W0{<<qJ#
zByF3iLW{+lQ!LC8jF2uHh*S|Ye*9KZkS9+fCVNxSM62zy5vn(8i$v*wK}z-pu0cHL
z9I_R^IWHC$=RruE#=?!{IpnS4H$4>2)UUS?gu93L>o#UDmwh9CmuHBHn!?LS$t-2d
zc-}MaUcN&2{N7zB{%b8DnWnzaCO55$sQ0My69?$L**h9jt)Ulj(@X7nH&iCMZ}SVK
zc>0YIhi?-Z^KTo9tR>c11J6tL>yNR#5$y_&WTi43z8x9ynwH+V+cZY(n0WoqTd;0v
z5O?4O1Z3xO%A%F~pjj8t3kPSfDy~`uAi{>FB9&0F0=0T%6hVz?&r7LHR?bd2!1@Rf
ziIPKCE{N{Aqvj{#fR-&#Xc|d6;DzQP8Y9oyFk{QUzD2yz2|(eH6*0o&&!uyN+w(qa
zd(lzGo@7{~1SA+81;T7K({B2R8tL*^s;(}y)6B;1{GD%pWJEfG)>gSB%I|7%impY=
z2V0D!(T+1#$a^Sd1Ntc|_nRz1Lv3DuzwzMurd(HO@1)Zsh6U}w%32;(FD3U$tr^v$
z4vpf<!Kp!)>D$l9jkt4IyWYMvX=8YpWp0?jC0kRSCz(pYMW;FMerHPP#_GMNX8Kan
z=rsXfQ?B)gzOjvAqHDWr)F*uR{jmhvsM{C1E{J+wr_}^WTJ;!=E&txfLLt-s(K_73
zK(yBz<Jc8rUhz^xudpf5F0L{CHG?RpMPx#iW;3ux;3Ij!eIZ@nFdqKz80s3j1BFyz
zK=^vlaDo4oCjdxG+$6w;6sg4`;PnrzrB5m~dy#)%4pxe-)CN3O*x^|kytNl&iV?_n
z{_qL?Z&d|}Tn=`rjq#c5rl3G=AFS*tWzBzMqScqZhmDttX0Ov_ccv1$&d5+Q_({QR
z&4u#aWm}&Dr?o0ZQ;JAX=GgF1u<kibCI0PqQ*Uq7B03FPP3d8c=ee%@LshY@^a%=d
zz*x{|T?vj+VRcuzKoQY#Fr~T$mjQFZ0cf|rNf6Bj{wRMRthAB7y!${3E0-Y`*7?VL
z)abd+S2Z$GmR=dH+j_<|>8)2}#LO$1W+17QuEOh;fQ5t0AQt#N(s<U#Op1F4%T3rp
z?7%pb&+cbYK+2n;|6+9%eLP1Ll+S~=4R^dc=N7=x|9J)^tp$l<3YdBA4omd|+4Zy2
z=@;ZipgwfFy!DtxJYwKe`->Kj0S%H49yTdFURg5HP{Wsx@DRfSCwNe?v%)Gb|Muqx
z3{Mg}I~Mg`!wKcHvK`iRF7ufLAU2Dspbyd^0$sTijd=^%Mn40I6_mnD-J`6z%C*g)
zA#1B{^p&NehhaGHR*X{L!vntpSoByz$R?w6Qt=LvAy?sEeIKu!pa)C6U3<+3x`T<d
z!K*#|PQTWaNTY7e&f?OuM1u0!u}M6Z-#yz{$iiCzc}f2MP1F0_)*8PiF8E75`0*pz
zm9^W_T>dd3Mjaw}LUi0qviX>fFG7Onm-O1H=n<yeI=nT3`I8b^gLO!E)xhUt#<Y<`
zG|^i-(c`Fd`y}6#lE_lO2`jyaO|96Q%0!RBD|j;u@m8aF(=;?_ocvXZ6>#=R5=jx_
zA01g`lAp}XDD^pQ5?XejqXIk5;VGIuwD+mA&@AjioQc^JipoB-4%7`OInc@Hs2qzN
z$kY~zn7<KJCthRMd6gmXJbDO*xkLG#N1V9}NH)j^;9@jfjtOd00(_uVnnV$Y*x}{k
zw<f0jY7UDV9bbGx8vTjb*Y3%`!d?5b-dF0>MR6SgZA8!Ghqz!&mC7~k3v{vr_auw;
z1q+ufBZjreMOmcbrtzH!KSY>hN~SrlV_YXyiE^#HOi8A-!L)fF@CypF`xcL(J!ibf
zY@o@}iWf$d#X+%zF}w88ECQT@bRX^FOhz)vs!?hE;Gooed$HVDHbc|NMKwf5M9PY*
z%-#O<>+RK8{n9N~YYh%-kxs^OR|=bliGFRw!2F_KIq#~}jel2wJZy1f5k^8vgz!vn
zIVrq+x1a#N%@V>afRM4cBDnIq2McM04m>~81HbjydIyl$l_eVQQ&0h1cnAjD3JG{u
zu}QIoAF&>|%T)^-*I6zbxog|?d6^aOZOtdUa3XnF?)?Xo#1Cp|@~N0o97q&MjE7i0
zeKSPX6eK-4s*;`S1=q8*8hB_@4)X25a-#J(DCVC7-DUtrFaiD+NTJQZJOd7J=*OnU
z@4NbwXNOPQeVqF&T0=t;Ib=z~QCZ|$u<X9-%pmk%1oO6*7q)%5+cQv>AwY)nxB<@F
ziE($`fecgnO!iA4&kVvEkDgb$bl|N5Dr#_{H2mEokhbN&@d?N(l#Vhoz-q86EdT4G
zY6P5w1b(xw@D{aad2!*U{h|pguCo1(tm;_2!r;qm<FSh|ag<%zXzezo!AMPlOrdi0
zlY0uj$bHwsnD4y3N-Q4)cJzxgg>+4xDVea+X2)h38zQpl`>TYlMr;}|t2$qO?khjD
zyH<)>;an4~bMC9P-C4RfhYhskrX5!x$Zo}*pIJR*(@{@KKi(i=aEx<}FJmv+6EXE~
zjHg)p2;5<r$M9&PjEq|vy%Y51(&h<rwWWtcZOl71pkLIy&#JCLon;vJS%KawO+}yF
zm~=5Szus@7s@&rWH|uTN!0)%71Vk))v`XL9$e5nO$X^onC=+7R$0$pLE_9At134`q
zUhOxcj{<+VuQ0}aiLQo=S^_e2(4p}dBrnp*iecmoICh@Ax**XNj3G<Sv@u+j2`TBv
z^w`N%7+d1%nELgDX^89zke~3e<`3dl-KoKqr-^^{_JyZ>2|S5~u{51n*y4Vgj+Ev0
zhzG@?hoWWKfYacaXN!<)imLqiPIoW9MZIq21(O4J)yOPYZE5L#wYWw6=Kk_!TGCDP
zK~_vRIIWCu#q67~10JJ0rhx_t(I7kUzgSM}9V(uCIxHBu-kbtgPFNmiQHU@$hOvL+
ztw{Fmz_o;RM#r@&sREmpUvVe(c}>5#hi6~*&Q#eBtfah9Pjfh%ZsnG?zz&j)FLjMO
zo(O)!@?uIy_0;&<L`h=xcKTPUM>`Q--a=nymm*hnfU$T^CbEl~do57*OU(Sfp!=rX
z6mW;+?s=VQZ9R}G{w=i}Bi>3S!(4k&{aWbhQlJonPzN-paC*w!Vs20Kf8fga-_<;h
zHSGZq@&Kgb?EO!Z_1*u)?s1-qQz5F*28Bg&SxtVm)HxuAY7Ky|XakcZ+~@1rpb0+7
zas{=J!QD9;r~tM~Thb~ActgPaL>d^sv4?uXq5Xdvv=1?+b{#)K2}lr9;fKpo7&JSO
z0^+wINC$|!wg(K?gZBV-Ia>w5rAyZryK6wJK>G<08T_@g&QAd;W9?sdC}8~dL9-s9
zvwx%Qls|5<J=f&kGPu0FZ0pGP-%|n<ZMN|r${602F)}j#TS5aSGCMy%_E#VuoF=!N
zJ%z%9=FSG*+k<rEcimL+KV8Wno-Zgw=wG~efdVz>S0R+^knjVvpR%@Ycu=hkQfzLZ
zL6d!-tv$%W0LaDXFI)cktE>_|;CpH`^S*NmfEZ3R(<m{Ol$09R1a=poSe-M7YXV5Y
zZcuUfxo!~)#l+OqoP!vePcggNN=6;!Ft)PVo>zhRyxz5qD@+$m$tfvDKn}Q}N&X9Y
zjpH*v^=#(9>f<l(D+I<NK?ce#1ELJTw7hIwxQ`OkYYc{*0Ii!8z_EXzj%RBzD>27E
zpndSC9JYIRHE^A7^~ZxyW;#d)5=GGGZ>~>3@Dln@P`&dB0jLbr9OnU*WyAZDkNRDF
zb9*?@27u+@?+`AR*9!4wfEF~&)XF1{FM8oQ!r{61QuR;0d#QRgfIfu+F^wI@6vX!c
z%QpUKYr+9QMfQ(MKaj4Gpn#P16#E`JDDSMcX~+7a<N}1H`$3CJ=xS7G1(MmqY&~88
zKy<&?>ndPz0ZC3@t^~<%PCIqStp5UAw|KTxMugk~q*mY|U;~c|Q(1OF!GA|RhK7c`
zdIbiTpwKDsaOpVWviWgZb+h|7$hGs3&y!~60Fsy>Gt$|T(Djn9_TxXhZ4dVtdWFNk
ziE&UZJGU499<m>iZEybB=D>JX%nE@J2$O@+fc4iRF@8fv@VB1;Me@^MZy$J@G9ske
zFYL}^EP9b_GsAqIkoHIgGIR{@_DJGZ16;i1n73z#M#ASR5;&r&=I?&u|IGL&8-KZr
zHr<tA)Xu-#CGwR!_=-@{@f~7`eoux%hUJYD=O0#pV8V{mn9d@lbdge$)vS`0F;JSO
zrmps<l~tEXI_-H;D%`{C8$taFkM@DCy9hP7DnA*ecK5*<=zaT6+HO8&zii+UeMAZl
z3>n2MEO%Bc$?0lYU?L^sM+K^)>kN_EDpZM*Z=*?B@y6=LMAR9B7KrMsf5hP*mt!LW
z-q4v9-r2!V8==L|JhtP;52lrsD1lECNM*-<|M}%Y=q(xe8k-`lZS2#7sTVsa%RUXs
z-d=en>_g?-!7YAIc(lXCj9w=^xMxWCRWT!G9M5P@hVU(XRTfg;LTS)JZzJ|t8LolR
z{Y)bL@twoQEeZa(LQezV34MvNd7u0FPDX?Pnt&I+$)Gc@%o0OHF#*^HE2WVYes&z6
z73-jPRn~8i5Awz7Qg70m_#&dScu$0`1yp9)Z)5vz`%zguyY*-&QZ_6r-;853Oec(1
zf|8jq?Cv*8^jfpU;JOS24h#IT<Una<4wL%)?y&q*hu1?QFlO)Ppv|p0es-mx%{g0%
zqVe^a$6T|^YqbfGzv__Ya9-}#(OWn$*khM7n=(S{b6P=P`*~VZ+}fi^bwA5X%YeVy
zCN^?Oy&490I>asfojc&A8eXhgD3VT{OTT29y%$s7%n#2k&MQ7LGG6IX7`fy5!Y~Xs
zoWan?eU(_8DvHE#nefGnY(<&^BsFcyT1-BYDX3?um)=ys75O0C374K-l4_aOuqi{p
zN)A34%~Di+q-=oOvl@YU@g_Ar@<q3URBTN4tjrrymFMCk7SSq}dCHc`xqeGM(-}e^
zJt@~Pj(X11>~)58qr(|GxfSRfnpMQokf(GdPBR?A4=^oPG>fVTLWPCqO{qjDrtmFG
zLNU)>!s3zWd2*lal#|Ha^#w+~R137k3f?1&ihrK|ApcN=CRgHNS<QV_&t+t(5^450
z^xn(N`x_duR8jjSDqAAv55*fpk-3E{gyLIl;}*$QxdIGd3g1^`^CsXL<5$^E57IBK
zENki(c%<4l)6Ac<b#!SYM_9gg2Ne{t8kO9e@KF1XXxkDgJgN#L{gUz1@p(I+$@SI2
zVI<S|OBu_feXV-c9^!6JF)Djy3)~)c`EZz2kAmhFJzY0pHvJa8zE*64pGG|uy+f4!
zO8%H6z4?!f2i-jq_A4TG)W!}Pmh4e4a%JC0GN=&A^PrV~o;IoO*o$=}9VafTI$EJ!
zUxHhzSQe9XcJ&OScPb1>Y?11wuWAmLD_Tas4U^1olg#Ic9P{Cn_JQeN<?F`9$I!+0
z(r!>z@m4D^Y;n=9xfjw}mk5?Lba$;UZS$54j;~ivJH2K+7I3l}SGC49F5DY+vsr4X
z5>&|!jgO6twDqiJvd(?*((Cg$)zj>~FM^uc7A(2WLC7e(*zz<DzHT$Cx(>_^KV5R6
z;%WumBM<qGJ6GQhJb60eoqW7@iN(|Gmb+OdJEKiW&XDvQb%mVfU<XWdQ}t-*N-oh_
zdXD$dGw4d9d47MXSm1Qfq;T{>v#2=!UPp5Ed5%f!lepc_0`oV@Hnq;1KboX&t2&O)
zJAC_LO1VP05~_kcS4hPTe>x5;PNpP6D=%r21)G|+1K%p@S<Twhj!nnlf7-|Xc_|+$
z-CNl_#{jdq@KTFy+TAFLV#`&|#<6*Nq>y-?oTSIRH`DdLgbk^b8m~~&U)Zx(HNsg4
zQ>~`vwteU>+Z<+ZZyZye#^pw5-G!uIbEkLmy;$qXk0q&&doiI=%_Uha6_ej1UDOlP
z6CYE-V|<ljB}iRm*SqX@BGt^l)q8M7JD>rNt^byQbJ9JTqUa@-cBHa+$CGbqZMnXg
z%}b8KVZ6{zYPs7&wnmC`D#ku7Z$w1VI=bW#Zjs8C%M;5h3I6jescM?0)+hiQ(`BZx
zW8ajVu7*SuRo8r$YCc+7-Gvt<GX|HY?e=zc5rTuKWd6KBnK?VRx+wNYpqhngu(L|f
z6UnOp?oh<VN>ZZquJlG<%k=0@upb%D_ViRXE9!{xbVgTyOTizQNjn;!W(r%a04INO
zx{^u_^QNIObtV@zovtcfvQla~=C6kMxN23*&RAQj&O3HoDNi+Nq^c=2i)tODkJ6$U
zzpzMx*;3ylQ@L+-j-d)2gQBqM!SosMCiOC`x-`D!OT^+_<Xc)>v`cH0k|o7sEIZO{
z(?4s5`MOolzZQ&(JEvMyn2)cpDWuzlMyB7q&z!c>P*r8I%r5MmmfDZ43fG{LI-Ki}
z_jP3wqFMHke_C0sv7y>4duOYmio9`+SB`^#n8+2WR**a<p2%59%b=hMP7wZT=^9m8
z=3M()8A&3QQ^mZLJjJs@cF7XwP`3T7=H01tPJ{1u$_83QM8=--<6Kmo3F~ub$U?5<
zI9pGZHARJ)dg}ZLQ_t8$Tw`lFbD<&&p^_u6q8v;E$&zxZUbV3P!j3^&eUkB0cq}*g
z2u8(uyvgHh)J3`7ibs7?&7o1MQJBcItr(?gOl0#oR}1yPpeYgBK`bSekurmPHP?Ew
z<dT>}n(dXDj%_duFesyZT!-YrCAoZQ=I8k^FXK&$gz9S#tu5;DX<}=RW{U&E*6zXk
zk~Fjd@e!-xc1Eo?SV+ER%wKl(nuysS6<PZq4O>?YvtdW%nVyMyC_G|kfe%ZVE-y(|
z*9=t)qLshbNfC-m5>>|)+2@Ir%qr$yS$W<Jbj}eQt8AH84;LvR2Sq(;w08XsX;Oso
zI>M0gQ89ZCQfF}E<?debQr={h0k_xbdx{kmg*`~Rr4{A)UNd0?7yXBz<e|+&F;CT6
z=0NdgmLffmcwWh)eQlM~sAP%!rWuV6!q_@Y&q}A4OtPipUmTqrKk&t><_lM7Z7GgR
zaet08L2kBsFy31dRYAzf2_lGz=;oc_LJ2Gm&3U*l>BOlu$o;IJ%=P{}@W!*gw+8QA
zo$02ldwEt}nH~XOI4+YXG|ahFXJr$$qe!_2t`T80a8ZZ%btX<=HKlXOT(|rtS*SrE
zcRI#wOM8v0G|it@v(Tx{h%nT%?YFw8=^DBs<BxpZmDT0wgw%M%;4FEYT2i@kY=n47
zH6*gM#K%*>6)k<~yET$~*wu4O9F**s3|U5A1I#Hb->YTS(_{pnAgtti>Z9z8q%`P;
z8<M|V+lJ}LO^8@)a36o?d^2#8<xcELO>5R$#+)msxRb)><m@x}j0YaeBbj?yY+)tW
zE3-#9KkX8AALZYlz#z(}|9SQhPm=3D&z@07S^ekPD;dhE|2*?fkf=cY&r?{W(e3{{
zGkD%z6Dt5X`<oM`>GTW?DsXs#FUYR|*tl`fv(ort_Mg{Hib+&h-=1xFKk?MIF%uZc
zg&zoAmd$w_$ZCJMI%3jh{Ljm*UNjg*F{R3Ffgb4yV6<gs3I|y9CBWQ3kMa;SaMnS)
zI!+MuG_$Vc`F1#F0Z227a`3<rG^jTKqYwG>o39hN;k5vnIQ=>x0Y!&*`IB(fyM^!8
z8Z`fThpH59wr*!oUYYGJwu6k{0FDpW9;6O^FkTRdGhXjtdY_D69ncias&+xYD5G}H
zO&S>RD=0TWUd#~4l_>!+D#8HpHUxagFH%8U>>Tu>lNyo0Kb!;Ngv`Mhfz|#MC^jts
zo<Q^iU>%enG5lvwds|ZvkDyJ?flg>1-~+t*aJ8j7UTvu+Ig$V~JOl`XO~4qP0KKz9
zV5WZ!wm3~|Vx0W?H%1sC><#GOM^nxM-o%%y<C*N7seFZO4wv70{i%Fv&?_l7z-Jr(
zL?Im0ZxSn&%V9Oi3gJ;@<JLfbeQ2k4zM0el=ifcxZEYMcUGY&qB^3llh|W|_6${>6
zI)sr6>45fk_4QuFhxM<wAB=$n=>q6uL{+&RS*WO~Z2^L1{L+}vorK<%v25{(b*GbG
zvz)eb{AbOjV8lWPwD<}@TR9sPz%S_sf#dr&G-fV319m_r%&^0%n)g2|MsH?%CCVz5
z*C~op0POMMXsi3Fpi_Mg0HPU%h4j@nvl?K(mr)jEWxdJG&HeI!wfCK2QC@AgCcbLa
zV2KJykwg&-0;1BJMv5RsM0!&J=^|y2jxn(zMJb9@1?db3QUqj(4Fp0N%FqO&fRv#~
zmojkH#`i1Ncb)G#?ccdx{wNI0JkNgizU#Wz!nj0Ux8H)}j8SmU0Hm5*Kw=>VQdUq3
zSR5(eKBA+o{UjhDATs~^mAgBBy1ma&z&K==oZH*2Dg#x5*-c=t*sI#y-0THjf_#u}
zsy3hvo#4ON!#9sKG6Od%<z+{G`cx5j#Q-GyF-9GZGmMhX9uP8kcAO7FJI`B)T2emv
z1+m;$JM_1F+e#gs!VfLOo(?vqWcV*H%_EIYt*TP<?U(7MgX+PO8nNA{v;O)0k$=Mi
zOhU`0fov*tRnrDqN$o%yRSU|;8yt4;@BJb@{OpYz@?t;&#Jo!b0ye|~a!?F{jEI^`
zg|?mFKJyuC#KoK=;%RQ~5u)2tjC3)oGsU_tj!VkrX)&%auajcRg4|3!0o@CP$c(?;
z#?5*up(+hBRSNhgD+DFKy_5T?lHM?fXu_Jv!b5MuxPnpf{%@~+-O#!hQZ3<NkYqyr
zseXW-C<ZLX1L1F5Xcg`IQ@LVSP_`*_+xqP<zU+9@qV&x=T#NJwbzIl$8OMLX+NCNo
zlV6MG93CcO9vYYu=rR$sug9P$>pst(mI$8;oho=IL33rwM7TIDcOv3`N=#*?q+VR*
z0mma%?;rA*2Mbn?CyoSvcznpE_mtWrFlLQg_Eoq4&BKt?milS^wulNLN%LCCDozy|
zk5yl#L>x~`)%{+sN@-eyK%$W<L$XRWJuEwXJX~8(DlCk-w!bzqlM$WQP_FN|Wa8p8
zR~|<3k;;m&-&>JC5tJewrXA$7H!zW_M=&Y#NL7_6&17d%-vRsF33bgl*_}HLE5dqD
zZ)kt#eBG73KtuVXC~fhJ<eIWr#!)FNjW6vIaXxK}rGLE_&C2BQ%vGz_skQ8j{(CIY
z`J8olnf|M}qXHS>qUjOYd6^9Yd+K`mD;}z!spAS`q~seGG1ILilJjbx_Hk?N+?E&1
zW1bNk_#nIO=alJ#b4>d@hHABud9I(f0<MHx!^l47@4~+soGNjZmCi|uuTKSM$5bA$
zI)XrfVY#;_Lo+MTIs9HCrMxO?zPrFU;cs~^L&w8ejGZb{dm<#Iq;id{Gt-04S@<+p
z>NCDhSVm}TN`*!2(~`1?5EW1mzAio)s9XEuNclq%bE&pekz3Ou9~53n_%-!YGj-4H
ztN8kgvwC)K-!kvig~~9FSOuPw7rG0MsO>b*&pi@)wyLr^L`-GcqDjp~<FGu>zV2nK
zUg?1ODE>;dtf!fkRqt&Q8;q9P`jwn7Qp5CWEv<M2EiwZ$dGuXai}}OJJ%JX1fub&!
z38LX*hC6$zs}80-9^<s{D{inh)Dl}`^<FP2Tzu!;oS$&T!!Zx-ojlu`ROiGClI7(k
z1#C_-J#zJ`BGsE_-^eqILv!W~b|3g<PtQ85P0Y`|3x7nH?R(`bK4<WAkzOui|NZXl
zd(lpND%dWbVOZ)~G_>28NQJs+$Nz-~UR9}6e(He5xy17mnGxB6mH)`edCcb1qu3+w
zcqBb6ymt0v&p%kl>~qz)(=_#(%XNM|8|y+Z8mjI|tu;teS2mned;idar^uzZdVg59
zQSGVbInI8;uL?GXhE^JSjP)0@m}_OYA3Yv3+}70nc_FT5wsetI`6wn({hUNggeAX}
zK)Q<W;=P~B-@kt^noQ$+v0*-C{nDn8@jzr|k9_zF=Vn5C`YSYi4Z{_jJeCusdn<yp
z&#coA6;r;@QYikSPN173$f8fi>{wy3fK6IvrprWM<^71nn82UPb@lXYtrA?;_*N-S
zCiSQ>HH_kuU0m#sq))h{>u&L<jwg6(oYfxQRA<4nr&~N#hm-Hae8GnPknr_(?6sTe
zG5vyeg?3>Ku6>Ab+ie1I&SEOl+n?rFS@xbb%KqbQGR@2|Oqbe_ek9H=i0w0Xn6`49
z$)m@PfqV3o4L!9Tj^--mB<k4~mFo(0kG+<BB)YRm;DzQ#&L0?CrY(*}YE80!e(^iZ
zMy9$St)^0Rdc0gZ&L=GXIaQ=@wtD9tt@IF;tUIG+djyT^*KGu;^53$`xpQ-3_kS3@
z&)i;NYcXNZX|Bzy?lBhLU6=Z}tUn-YTbI68kqzw*L#e7lKIJi=WSq|)Lqo%!HIs#l
z1G@|oQ#I*af@MO(3bYNM-}pFeUh7XA72v#6zSf_oum8c~3$}L27b3yj;X2x#%!<g$
z&F@p=`<FL%Z>|=M7CGrB7k1Q8_fZ~0`V}R8f>{uFMElVjcZ$ZDg~#_NMLQ!?yYx;t
zq=^nLMpi01SBah|ng}~}LZPOry3U$z62bRSKtbzZcQ;n?51*t-X8jkQd^P*u$9pB?
z)Ozu+ZrIpDd5-o=HpZP-JokiN?Ay>2tJ7WnabD0tX6KsO3TjB&1%)44((-fDk7U^e
z>Be54zGv-nq`cBxBGv!JQVmyHtz{nl6f-b8Ml@YKfZ_GZ`uh2is3U%P?zIQ@S@>85
zhKuVJU8hi<EwM5F6c1#su@672%u!oe)lzunL+^y|VFjb!{+3Ijy*m>#Mb)_Vn_YS<
zQ}v1#U&aUxNn}Nrg{+zUU1<OBOwKkR|NUks+24*VM%A!~AGOFe>UlArC%z}T$}Uz_
zWzxfAEHh$p@x#{g>OALjf@#9t3a=%%_^lm}J+ELNlPWwN|LQvD4!6j8XWPk~>Z-Lp
z&5YTgFHynjESB@jpOfD!_nZ`0h<o$Pr_VjHm1B?i%s=)g1PKUrpK2Dky%22Aoo~El
z&+FC<o8XO%YbQ#ret*DbVRJaA|6~{&BZ#qf_C1APVmb0&#=Mlm%QC&`zQXcei{B^g
zne+00@Ao%Xe<pA+aNnzi6Y=X8)beNt22NfWUY~9LLNiBPK_qgXQWWOzI58a>R;K)F
ziR}jS`uR>CzXEsecbSR_@?FKjmJZ`$AFlYGJM6Ts=eZxRwZ`Fov3H`=)3a3Drd^>4
z3mkE;^q(=-Gp6Pi!uc!=!<tj_mG{oov9}H7H3)3;Ro-9`y{(HYn$Lf(;E+TA!2AW%
zjpatMM?$YtJZgIrD3#?tzrB80&A61=`9Ve0!AXSnHG%LrKM-EK;(Pzp>>n)|WBhW9
zC%3t-I?O@lv015?YA6Vdl6J{?<z2D8=4kOQU<cRHOM$z*27aM{`8(<EYKLu8uEQG!
z`dfF@*X0O4j<?V!Jkf|-g{w*&9QoVakMG^PH}omqAr;YukO|lMjv)q>o5dp|H!tPB
z=0E@O50k>i++|<C1p=%7=jp^OfwGAh7jLo2%*#9X=GJC?TU*=H9BW!H{pU+Ku27M6
z%`e#WRl#t7xsiS0yx~7X3ez9VIo5pp0Aw8B?t9f+#s8GTbZjc0_3gy$tu(|fNfrVY
zbRAGHISpR?&bSv)M4ja&7Jj~yMfx>Al`ext%w}qU-i~2bH>3e}2ws&cQqCSAVKbMw
zL*z_{lLABoRe@ye?rL7e1xTf}8wmJq%UCly-85hB(lW968Bj}6K#6Tv|M-M7vLG>V
z@*~xF{G?`jlIFWWeyk+Agto=isig;*7tFi{_wxELj<0+u8V-Y)cJ2H;{r#`#r8Gjc
z)D(fbpY*`v-;W`}1=U!V4c{jz+EBsM_-%FLx>wOK8MwVCxE@)Kn9tWM1m)?4mR91d
z@v#TS!)HkINCt7z0urUxp&>i(!K8=n;5BD>#kDjjtroX058R1vNR9?gUsu;^GgKtK
z!6?>t?x0n-M-Q8f+rXIx0Q?TL_XBPB931i2K_PX+w9Kzm4hh!?cHHyQfX&C47zHiy
zogYjYL6AEP4SL<FF2+E|GNd0BWUl;uTn0Hov}Vc(2<I-&wpsI)_{>d==&797VSO;N
zpKZ4nS(vC6k?>yp%s31}uKeHEvgL-zPTbZ=X>`Y&u{&0522zF}E>XVUpb;<MDQxz#
z_0o}@CCEZ#{1!cAz!yK@`$@`uZ8Wo=@l)M+^DN@%iY)W8E+czK>DU5Hz-F@`QQ#<7
zXFqPBiAeM)SrRnNl(HopTF>Kkrv<do3eKlp6MCM;-+={UUqh=$;C~(z@^1Fjd<uo)
z{`a51%^>o<-DRF@d8ZTbK7+`Mx5mkptI0bA(O75ze!!&Z4W)iLfOUs4KV3jg|NdA;
zn&0J-&kEs=m)VY5l>g}6pt^4duPXxlc1%Kh2uARHy$C`*Ye7N5#Fn?N!+~5guJ(8O
zTZ^7H+AfrIvua24Tg=>ngcjOWJy$%i>1lep8uoC@weMCZIPTc7qaEGJySse<;cf{|
z!3K0Av+wUwv8xI{U{vgu#y@DjJpYjB<5fBjy{X0DOhJ_Mbm$aah3`PDraf+nzJLam
z9=Q7lmRp}V;x_OL8T)4>WY;}Qj67l$aC0+X);3Ma6L=t}@_bJ-4&#{HfI~1FbM;Eu
zxX_(q!~MUIEs1+y!?yE?;_7vqnt+N<8SkmcFfDP%b#F6HbZM1N&&b%j=0xThbWIay
zukp4)`ZEdGb8S)q04bZ8@4OpHIkn{Hp5j-V(XmTl2_8I{K;vENyuvrH;O=Ie=%lZ&
zFA>1Pea|EDE7Z;TO=KANtZ!wN#8W}X@NFh0x=$6EXr{!vx3g0(kT*)vicgr`Sh3-m
z9NDx{N3DZ*`Fwe0!#&O;6P-L?q$pca`w@Gg8Ba7BojzuW|6(E$VN|u7GpL2_yf-o7
zqWWrCZvQ!5-5sUthTo&XC^9`jKUvO68EMxgDKgX}l1KrgudgpipbHjL**AJABs-l-
zFRZSv<}Nk&*mPcnC?cjJhJg{~6215b!FI|`H|d08md9?zlAr=H029;%=}TV6eQ_4n
z4JHZ?Z(}PA!&Y%gT#kY2NcwEIf7v?=Db)hrQ&`C6M3v1#MtS?&^>lTUuyTs5g;}Rp
zq2HMYiz=6M8rnB!n=6z8+&|%~;N_3%2EB&~qPAOVr%&IVUnY4c^fCnaoY%Q;$J)-q
zt_`qm_OEhR4DNYvBxqZ`^0a^W{MVyI=q{Xv5M>h<+k9#lmhP^`^kniDn4taznoCiP
zJS-i1NFSWUVUjbK&UR&>nsX*UKR-&`F$#(KW)x9gB;8%>zyPmh251M=LP>1^)6?p_
z(r#z)0E4l^JmXS;@zr^xt_;@7CRh_U2n?u&R#jC^m5%vpffukITO}Vl6aXq~+~RS$
z?FEA!+Wb9o%Z9<c+_p@9>}$bT&FWbewV|*KG=BClAtcX7$~+Cyz(HJEBXkDpFH*|0
ztaVsEq~?R3dqkyWcy-@Yo&S<f84&SPgk?b5&(%JX%$m<be1TfA%Uiw8L67)aqEQ;O
zORiXH=&orU_~6_?(Kv+1=IG+6f$enOl@;R=9Qz8d<bfDawE;8RL<u$>6$Qv^INhyq
zv3L_M=Kce1j0si-1^~_J#l_O3%mjguBrov)5#!7RuAn*EiI#@fbkCi$8K{3!46JxQ
z-r{s9ls<V)J^LbvwQr{RdK1f!J!0bKb1<TzS>K$BDv4%X{~&5e3=n#zp)N*EO8>{H
z&|UtE&a6)yWj1-fcmQ1TK40|)J6!u~n*(<q&4YBw2u4Jrq$%l;6uvfcR^fQNn?iP4
z8Xdf&<kh0A42*>^F{&QKp4sq#y5he~y+`urzxISQQ;j8ve<U1x?rD7Wg-+&se_9_A
z9=_etAFi^g)cdNcu{HHqS(Vf6CJONqPS{B1z-C;H>+)g>H8(Y-zI<sy%m<EO&p-xU
zXlJ$Rzt<PXDcaw>{r^cJ@_!Z<ptAX=<}l_H^oYx5k)!vmCMgr4c$$NJMh62`GdIjL
zY<J8^P2KMmivpUuK%sRGUUNSvOM;DOVW$PchT!Vy@j&?hidm4GP|cu%dO{sHa1PG^
z+@D%R>w^RiYd7!8%(1Aj`}i8(f!R@*O@0^Dp0+4bHiP3JojfdEUEL@cTvSXckbh-j
zC%V#|QLtnHNRo*;1@nlA2rfi#?1v1tl*ZRwqF@Zi?(OZZT!+tJ+ax{kvqqe(XGtE4
zFUL*3UE*D%7gwG@P&tv*C(AkahN(6)7(;k;x@CSW92lbI8VhcQd3dhOnUM~a3ed#r
zKy>!LfL2^)H*56jJcvP(!9p?!*gpO3oo%4XIK*1)S7>>5;Y|}Z9SUMFcAD@Y45D9)
zO*(+^rTOc4&ooM)=~u_Pyr5t{jSxZ=;fzkUSt%(UmO_qMsP$JYrg5!-z2j*PQm5-E
zC$|YyiyBIN^a5ei;?&&Sv;LqNox-C^$8oROHMT$_?ugh7fLvZa{adxbxu<*(ycYNQ
zsiq}*>v*nQD-Y()Aiofr2O|N2Zgpa{-rwIdjs3|47YH4sr{SlXmi@c%r)f^zv|`BE
zOabd+gz=|1LIS$eqCm+NSLAT?pHG_mn9V(+STm5%EGDJaX0RtHuP;bC{m>5KlNV}U
z`Ivb@vs7O{1@S_j4VHIaX?e8%$?qmDy4Ylij-Z~=Ir69}O3a>L)v1hCpqFh%MID<x
zc3)T}@d=?*-OH?QE(O1r%0XO7(0w`6y);uOhA;)E$07D6E~oY8O09jk7sfCTCP1RS
zD_u83xr3-4<vVqoLHNFixUB<^)Q9pKHFd<Jh1#>f*yepM*lDrd&-V4K@APP=F@j0&
zu)eUTORNQ^Klo&)8%~Q1!amT6$-rooE(yG<s;diMzQ$(?U0dk`@T($0XoTXh8;lJT
z7Ucmy)gWiL7ne}*cdOS9fR*%|r)QBg0?A=N$M!rrDMZ^f?bnDrd}1B1+dVViX<c6T
z)(bh+c*|!(D~*C{PyZH-gW<1-jh_Op#uUD%1vVyOU(F{;_>yT9oCiRSO7E=<%ZKH)
z0ZpkFW}UG1PQ1=<w~``oQ_^XU>9KBeD=RB69OB;MEcls_uMwy!x0bIg*749@$zXzE
zV$HXdF~%$SvWfB9r!aR1qUq(W{lY>*&0tHW;HVE^GUx^=!{Z>>BjTj%xWbzkFE1@j
zBUq|5${>WnQKA+%3IE4inmmk1LAlevXCN*{#ke{0^`}1X|GH&zEw(+^s=5gVvJ}yT
zkZJhH5aN6Wmg^}bv#9R1hz$+MfSI05cd(`Pz&@v!<2N_SO`S<PoAy9>!>Uy!{G=Q6
zx8TQrw__;j1kOnH1RiyRvnxD*;3^N8Qm9ZBKZF8$anjw~YOjgPea(eV-6QU$q$myX
z`6dL9g`HXJSFIY_`|SpdN!hs=3ReLIQnkUX<sXke3RF{5D`1Lo2%Ds0_ZXwU(eGY5
z`oQMSOzTbu+mcwHiOS-k^bEI|HfzC0>X&bYez)pceQ@bPBxn2YsSBWi(x0HfE8`x%
zU3O3$ofSFMpXoZ6vq5^|HcSG)3$On%f}lVeqVnXFa>RP~RYgR}dgk8zac9Db-`5U;
z#J2^6LZ*3j;NPZz2HF;Zt28*JrVzEK;|!BK+uQ8l_XVAj23Q~}?n@tinvBf+VsHXc
zQ1`bX=@>qN0%4>BVU>#gGzi0!T}pyNX|OKk*U2voyqv7np-}Qc)m=7_nizZ*VTpid
zOlo?1x<)}GY4zaHrX#3T>fjJEdP?$1gbt+1ML}3UNG<FN=<-Li$?Y%E3Y$N#eMjQF
z%e1*meJ(|pvMs31BrF2WwaMzWKh~30*1}j>nN2)#8kjYFPDbzSzJB>g(sF)V*kaNO
z5JUb%%$P-aARQxY;)dBTNZdpSh)G&YUgo7VX?r20kN!|tlpPYxC8h=(qVZI5k=IL1
zdl4HhpsMYE$fwE$fyI37Ra)Y2tJe*IP`L@sD#Ip*c|qqUsopJA^D?u^{<@Iv^VnD$
zKax%SOy>oxV;=eel+h1{4^3R|@TcMKjfz~7_-*H>28q8SK`9B1NVf+Ik&lB}Wt0J?
zWphMvFuvyyrc&VJhcG?)WC_kA<K(SP^JpDWKq)nZ1KW(;E)%*{gAn<1%sqrSI+2i0
zk)%8Hy5+fwU|y0MP=;Hz0re}RAWGhU|CG(Aw|C$j^{5Si)v44(ik{?_xZFfsl=CM=
z1S!Onp}`X+>6~z>HYP28X)2|r;2p%e+}{6kYrCU2B+td%hYg5X@`99JQ-Px~;^<Sf
zvD2X>kI{&g)(sQTXu#}g8Ya~AF;sdZP!u9#XTw@N;hiZ=YT*$44*ko=H%Rb(Od-i-
z*O(kcIG}CHAh2e?_;q}qHRdl;H<<@MX;bHSJBmS#bPO3a+=M*>f#{cI#5I{-?aZ(A
z6;O|S3_kF}%GtAN;lzcZnbQgEDVXOUBDA(}Y<+eitp7u}Msp&(A~E$@#X|>Ka2oOa
zIQZ&k@15}lE|qk_kiuwz!`a`BEL5x}NB;Cf+roy1v*hPEa{<bMoa0GMgfKM8ZziM%
zreA1&r$gv5pRZr`i#|{nKs`8_H0_8*&yoHdX~C}w4y7j~XhD3c1mV3DGru`2A(Izp
z>SQMau@UyD7Jt0S?S@Iw!H>827>M62<Z#%AT5}S~h2nw8`_9`;^j1pav~imPl%Pq6
zIX2I8#Lwtu-ZR5#aHm7~TVb_at0OzHDwHAkPyUq_F3JnuB)twSAYIvLUo~4aQil*!
z(+eOnXN(-Tg@OoN**vEcse3?|0**7WDyj{jeM`gEB70RDih5Ct>7P2Uc2g0H4dD{*
z)Xw4)?U3InsTPnk0ifk=vt_s={0)_K3rRp4D!%Z%S~+AF6=Awp*upU%m+Y0UuT}Mf
zL6R&NQ<ws(cu*|@b%j##0168*w_X+;4nfU;cqum&RFqVop#W#w@Kkl4_&SI`=8nME
zNC=r0AJan{?@taDQn&K4`jP9{DU5IW(hI%}KCT&T$u^&IYN0geSg3%G-RSI>@D>nH
zv+r%?PeUq!I!ObZGQTP`?&!$vdy6mUX}$Tw{DKs>7Yb0vtZR9M%xNaLj;cPFTn76{
zBhs$@j|5cFYCW5x85>YQPBkXMJ1tFh(!xnnSRiEf(g1nkmFEg?E?0)`)&A)S=^H0C
z+Tv2rA<h%_$}TQ80%uu${gad2yhjlb;+T;&4&^GWKw~)-H#$g~vgH3HYE4uQI+$*K
zvSg$n8m;+44TiT4wBj<c9~9RYP(M~T_29Sk5isNMS_R%I@QzDylcwTIl`H()+y#Y%
zbSac+y$tl~TaOUj!Jp_E7(Q}G<SasQc*FO-Q=n~lZ+pz^r|*;hs7$KIN#6(Zzm-t>
z<|f{BB=1l=a4nwu9DR4@G=IRuOXZOQ)x3W1o=YM8xIHcQo=UyojmSe8l=P8i{wzu<
zJO(+s*|v?zgUEMDsU=EQ<wN#eKkb^mdp7!CZhx4j(%O%>f|0NfOi)@qzHRQtD2k!k
zn){9dzh}H|#d<cL5(t}+9oxt^3)(du65-%+C8V?+c@zje%Tx@b>V`I@Xz|IhrWz+?
zMZBU$NSP~PvJ5f7P<L5bc?ch$qcP2AjzhTr0p^QoNhJHX=|HAQotx}Ojw`V3l<)i$
z9d_*)&Z<oZaNg;;xw!&mdWL3ZS@8R|udc5_AJ~#EDhGwk@SpA8Q%u=-rO&|zH)Fvm
z#8HiYD5?=i`Q&Hnh7g#66O;^<J_mZE!^g%8AM2*VuA9ev0Sv*(S?={eY%j#4xgCPN
zg5Gz7vVw%uX$RF3_%(0RtEPP^`fk;6*?&GJ@#tnms9cU?O=)3)qC;)ZQ_MX3cKE_q
z9~2SMhg&Edcu73qS~jky2XvALP)Z{|rJ$w>!v)n!@C0a(J0^~C2y|ujc)Nssbn)3g
zI>^BWzTzB75b_Y{3=*6Kk=PJ1SP;4j8$W<FR<PZjVQ4*w(P|p;D5ScCK)mtOySpS)
zg_Tp*`VDj6om^se;a$G7rkxnMM-1*0rVyeLih4Q^0iuv{DLDhB(6yK+^Z5fD^2hp-
z?nHw>G5MF1H&d~eMlU^b0BfKR2&Ub@LQWC4hHaO~#~M(XAJp<sC&wQ~%c$_OX<~Mz
z+AEzf5;%-n?rGQhH!J`M7JYlTyUz$eY{WJ&X=0DgP^pWP+pCIMghteNb&ZVDG2LME
z`j7R+gUQjBf%Naz@%&sjIfd;o2v6SDE0Wh5>b?Nb3#mHz4Wdx~Er9=8<$+w$t^UM1
zv37ZQE6PCq2?CX$CVyGB#4I~!Y@A{E{8E}q$gXI)Y*FR1xf&-ol3wDosw|IiApt@^
zfi`d_Im@(Uh}5XWT1*lNV1@VsvbM1xJ_zz#GCY|Q;}U?kl_MS$b+|n$!A+R$e{9Yg
z(l^igMZ5=!p^Z3gQ}fJhdz8(?wl7{=guh%t_h8sv*dmaI4<ABW60x2CGZ|nfPlyEG
z?S)R+WJiDsFeJLP7E}AlD2&t~WhF%uSm9)BqV24P-;?v~TQr|vs%iM^(NSF;okYTv
zIZqGq%WLcTWYe4;f%4x8Qa%dXr={oUNJRzd39de+ZDm{@y17;)%f?wwAhpDJgtzQ1
z7Kk261x~nZ(S>B1q!doXGh-pDGSrZCm%JtMSqOj{A%*Y+71Il06g&-Y)UT$<Xu^yL
z1<8SVOpf`~t^4@j@z8{9eR<UXqFE!69UYj>=O(rnydo;-fpn;9Xatdh80)h{868`>
zqv73mtIj`coYHcGeL22|K&6PW0vn?`sYnTzPMsw|w2aAhjSaM2#W;`|(8x);dws+b
z`wo%IPwNrT8U3I|3Mvjy0J+j8MIZPrf^g!6o^bI%!d)cw3lD_3aD7Ic|B=PeNX9|F
z-oAC>==_erU=S+-_ws;QEM&^F^3t_zS%f;+cWj)8k0T>9W-XGA5h+Ro%4Af`x$;}9
zWMV<uaq8sFAKQld_f{6HMhWFP|6ti{Y(jz6NJQ5at+0hou}cAJ%S)d0f};=QQ%`oY
zUf6Ujvix;@OqA3Hx0;H5d@`uif8Yb-3(t-IN6rp1sts?8xWmbx_WhvI+A4Wewx`wo
zAZwxD>s(*x(;nqK`X9HFC5%U}EKl9envXxd&u-MYLM>EaP`<{DqJJ*_wQJl{yCfC2
zX1oo*>bjRchGCA{2JyeLF(*^2C?9T%YsV%%?u^QGrLjq*<aHV<wQR07GRV3)+jC>7
zgrieL)#yu=jc<zn?7)k$L#$xwEUsO#iBn_o%w;u^CZYKYd3F4qDNo+`^OgF|kBNm|
z_s;nn90W6awXV|B%Hp)AX#A|Yq-}`>hlb;C(H3&%c?w;gQ_Lsa9x5*gc^01r*U<b}
zR)f2>m|grA$_4H}UN5kBExh)Po=B}qq~#`U$llr&|HbU)hHUPcN|ByyZr9MKjWSyG
zSG_1LeVt5xRej%fwK^fzvEo}79?^#LUyB-~etaRHHYLQyy`(rLedvwc>6_wG>;ko&
zk4p}or88fP3kOQvaMb72qN$r|o~6*4LQ*Oo>aEtV4YnQ(JX!ND>`IK%^!t|lE6Z~W
z%jx?ch;i7z5$$SPn4DeKO`O;foOsh>|I15Hk}n!xklAi`*+(p2%he+Dz>R*#gUNpP
z7t9StHF$L-Yd0>KzL{}I{X!Fi7ToPDtm44_f}5^qFZ^|ISdgvmc@jl%Smepgp57N*
z9l74ks}^g`9@@#(JUvyrEurV3Y^M@FFJd(P)L>xJW;MGHEgN=}&)v*<l9sbAuH>W9
z$%Da9G!A?yyJB`|xVmQgoyNNyNE&>RW$QWT+OsD{W`jA$L7(i8QqdiZCeICZ`Fs1E
zJmRhZQn{IXSI0A94O@?qD;o7ChTIR5ilo&&Y`mI8UQ^D$$v+{POdDuDV6yp^v#uSF
z`D@>kH8m$L@J2~HedNkd%h!<H_3I60<7dWWwrswx>MyLD;$qp`Yu9K5-A`_(ee`Mg
z;&Y*m8|nR2x16-on~CxuUk$2QRJ7om319lOqK+@cmGv=Geq&V3x~P&6wwARQ3^!)C
zJX@;q{cNN8hSRh)_=Y|6=7N)|tc^$4vyqJphC23BH(TuNg0`9|)V6zCH(x2`nm;n0
zlP3G;=iu&!X~qLHg+=vclk9%}W`92+6CI}@=CS<X$S?J7Y#y_FOcUnOM^;8pY`dRQ
zo};0+Zr%_ihTD4m3UoNqGzx5tI~txAbLr$!6Lc#+t?p>5bRSA6NUBVj+<w{KBlz`X
zt0wP6_BOSoP>lzBqy;+C@3>sO*KXOYv16{qQNnQaihoDi-D0l9iHY-nPKT!i3<YcZ
zZm2sD@g8pzpm!%>{+Qj}>+{<SwVi#7(|#?`=}4`8<*Xrc{I&Rpm|TJO=NF2&x<bh3
zE$mg5tCoAzIn<HmW9nHiobFT-XBAhq@q1k@Po=&wM)JWjf8*&xvk4=*<#_Y6Q+@_<
z&xAHESnFE^1UqT!Zd=e5d$H_zI_#i*bB^|c;%vraw%v38JW4-JJq*1xW)-#}3uh^J
zEi3&?@wD&6-Hi8ZTQ-b2S1ekoQ?=_TYi19dCyRJful&gFB)F@k&#I3;y|H@BbmdJe
zr-*y)O*`U6&TVnp(2;dG&Znq2e0ynou4}krUxrOYt0j+}aNHj~hqg0tY66~#I0-Or
zj<hu`?4U}Ex0ikNa5R`v9;e=E)oa~!BZ}KxW3=%vPJ6YFp3#A6<^xMD54{^JnnEUD
z`?WNg=&zg@K`g7+gO&=$70rWzhBthZ#ytx%FW+GO(sU)JRnw9C-F)0cz)<_M>~nb=
zKG%JoH+u3~viq?3F){bot`bJ}r<6qV4%}GasJ*M?Tl9$bB2ALfJIxi7av=ui^$%B-
zaptenV`8$_t{qYGTCAdir;BWL+iKgCCNKPM!OciIz?RLMwLVJFz9;vX>8;mudra1e
zn{X>y3~daKeNC~V%Qg?l?p0zX&q<{|O^s3N7cxi`&o5YCaq^=W7u9xym@^9oYfK_9
z%`@qWu#H{dvE~`J{;-?<8>_pnm^mEl+E~RcF(W3nR(wYp&7<r9&E4`S>q(pE<3|gv
zR~)xQ<p1q;VP}*>oe=$JWsU-iHUY+3<DWKj#TC+JA1atXx4JAWpfaNSll{oc$%AIu
zH)-lV4hJ_so9iibE6ulR*{slb3)I?m**ENmsLrv;jIG!eL;QOxFF)t2c_)I(-pbd1
z11k7`!bSf#?FjOD|93#^f4$+V{{~(1zYkXb*Bbt74gW7@9R9U!{<Uq$w)uOx@akBO
z?!#HWKNXVLzgVAeesr?_gyoaNPmHv;j7icJId(@I*4yrJKKcC6WYisLH-4@{T?wVc
z+M3%+%fDSn-qvJkk*jDmy8BH@W10?o)A#S#y%SkWOMkjbrf?jOzi97S`cp!`w`c97
zPhJ0lPdrTumBKrJtjyxvJN15L8~*EC;zIwm8voi1|34oMn|RvjD%1SJLKQ{Dlarrs
z-?p)}WzM_^E#ADl;PSzCMtuYOn6)9jAS6nhJ>G!9vlo&07uAv?_YH-V%ZbloLyH~b
zUv~W@B;?pRV^JY>vSe_!;5Ie-;>kR}iDKInRxPjjS$*;r$Hc{z72kSm7qcR@%5_>V
zDO!X#MZ8ndFb*hh>@(UPm~tv*+~GE3f0}(&i06VQpPS6*o4e$*YmG{Z;$jz?=O%6T
zzv#lJZ{0d>W?^eNEt6#m;*K8&-A1}P^tgiPkqVzn_X`az-=rEVNYM^-*&nKL@>i%%
zcslxf0;_1G$llN(9+&Z>pk&#;PTrQeNhcfIyNu)@(=4r}$%8M}ddB-)E~=w_X&rUQ
zepsG0GOYf}yH`G*p-IvGN|UjhXk19O^_D5B^XOT2?2c=;##Ows{;C=;&-Csaef7!l
z-7KZ2puE-K@Lwi;6Jl*iW5&CyqBgb*z9?TasqryiFh)zh!+`6q*@zi;(A<)5Q=V1I
z<Qc6=)lvEC%Aou&Uqf8-U*0VezbaSKv8>%#E?lvh=~m`G>iuwz8eK&FBE5ZRaA{~J
za@T0i==Oq+k(l{3`DI>K*PN8h=;Z5S=BOcZI9H|*Rc)y1Q)xM`tNrMsbCT9O7-}u<
z?oJLjrb8R&<PW6T4;^VK<497n7?CaPkBgUUMh@pmoBUAUS*r8Z;qt)e$(q=f*w;N9
z<L@fve>f>L>adGB!;;Q#^o+yjzn7B<Fz0B`$##vFb7azDc^yYImt_jWyf1OcZI+4c
z@I2HxxY$)Bc0$W{r2XM&#hI^*@2RV0vf1RBr8+ON4tZ%(Y*TVlFMIi|6=R=%p4xd`
z%<igv^as|n9=WRGEZYr~$)%#YBZXzpW99TW^e2su-XG0k)qXCImE$P+Ft&!J5xY=q
zY(MO<G(SbrS?uyGNxiizezlxWaBR!kIU4&*^Q_ICW_6u(+hF0B!BiPT?%`UW`l3$Z
zx+1N1rn6&bV@yn$S&*uI|8E%{rBRcHpDRXPu9?P{l`yHD-pe2Ari$V(xGI#fXQg@f
z_!b+h9Lvj7GIj9pHW_W`P|J1T5LB1H3iMPkCNQTd62w2OjRD261ZmNlIkiZ%%|8K8
z(6QaS?d<I>03!7Lz}<QjEJ9ld@Tmu&tN;LMXDC<Chn(pjW^0o2UQMa(Iv)F6!l2C0
z_nf|dzw5&;0ZZmBV7qv>ZUyKu_~a(3ry<^}y%~TTHA}6IleTtxT-(N->N>M_D5ms;
z@YO-brH?BkkEi%ohGnlQ{<;!)o;o)VWt*uK-_G^LTt_TVsYi<o0sGL<V)~n$GCuLY
zTtI6W#Vpjykin!9;MAXHZ?o@`{~X`_%Sv*}|I-R)4&NFpG}7r?3Cf<7-R}a$@B%g7
zivSb*b_%r$ky^8QjHD=jKH(3NoPfn8be5&i4&A20<E%(HZnTY!jotiQU0qGk2@$0X
zHAxPoCVm1K{O<hxe4RIEJ*Ef*KLIsyH!)cHAij!+$Aa4ZhI1-D0kjbUs8Xbuy`C(Z
zzTR%?^=(;0FB1}Wqey=D^H-wdQ|5n_aTqOR^}nLT1cN0%-IfG?<s#4@-aaPxKuEOB
zI+mix19*coq`BS#2+bWW;Up|w?o-RrIoeb-7b8*(DJMZqnUOqTmDR|)p&60j0j%Z3
z!h)M{-GIIyCTgvL8-0(#;+o8y9MJ_}{X+@(<u9m~FBIE9JB=E3RjtDE;^kSfW5+BT
zQ?%*;EeXynDw6WXbmVjB2nl!-(x(b2y8j$4=j~nqNL(h3CcgkYLVQiM#MtDJ{L}!^
zYM%kf;TPyD8H)QYeogQ#)$Rc4k3uenCauv?t}-$*%0-|15?S5?fW+z?1D`#YuFE~d
zjjmGx#xr)+bvJW;MvJ-zi}alQs*sRS8`@sESV{*pue=@9&kR6EsRnZcNtyZiVi_43
z)t_r)qfLFr-_5Z;yRN-%+U)t@QQh391t18+np;||4IfyMcR?HFA(=EV^!E0Suj~7@
z!0!^^Yi&3L;r^v)W)GulTZg_{k^JzE;keiv4uF~0Y57Q%fzvw|JWoGY=N_`Rv#SC)
zFopn^=U4zGDdR>xHtNg=0UGGGp3f_>08CRu!)V_K+06pKRY{goPHg*t_Ek!pRmsv5
z2YvfUrMv7_AS%|RO93=u3rx_W*^rcvG+80C7qe^ne^iWR%2xl?|KlNY<!BJcqj@Kv
z0Yd4?A~a%W_OP+B<zN9LfynUd00_eZ6X8nczJDwY14bB#mYw*ezaqDyF=X`;SUET%
zAce|X%9f@%kJ(h@9YrhN17d7@c7FowDj2%ik+PmHe#<YMdoH4}th+pAk4HzS$>|@u
zi`WK4pwLtmDeh?WSDc*817XvsZkmieCq>hL5^~+;f2`m3@7G*Gwk8_8JJ6hTQ;(CC
z0;f{^JK#hufH=X$`^0x~@{}fGAMP&Dx_ATg@DIAzvD$W8$804G(ifXjwZ}25@fJ`b
zF7#)Dz<tio#m#L2{D)G-W<GgY^NylrcX?%2WB0VOqlCD)`=f<hoB4J*p&1=FB<^De
z-T7cFB5v~x;Ku?$U!=DJ+(OuzOY(2pIyeN9c9_z_#<3k=WvM1bN6^J&$9VI3zv#J`
z-;ztDLk`CuJxtF2I6%*GLri85{`KVnWPU4s*0snfcX|(iaa(g(PMbI2y!6hV@F$y{
zQ_jmRJUslZK5Zi61=^FJfO85YQhWdtmdJW20LtX!nBh-L#TEf3zGPU$VfFm)=p|PG
zNMP~h?^n)1P`$%^1t*!bS4?%K8K9xM6>8}x2?3ikKsWt`(RDjdd8JSOFe$aV8qLow
zfF3xbO<7KG-tzI)Hyl{+flR#!Yd#KK+~-1E-&=oBYez~tU*wzlQ9e2}G<5%g19uKp
zUqok<%x<XQ%WoMHf8opX^Q}1+7J$DgO)mqLtql89XzqM#DM7-{8^5izAXE4qi~jh;
z@4KsWZ?~3;1xlwV7~rTMGI?<rP^LIZpbziPJgVEd-(9c2Q#5{3sRh;-4eh&F`;@wV
z7uaJUeqvsX(H1@ts91=?lBAuT-9><GM6IgAX`@`j4Zy=$04lq6YYyP42G7o);6Z0j
z8-oWa@qS);g@qNc4ZKt;H3z+Y^5HQ!uUof@x((Dz!pK_C6O>lA4<2ela<TyNb>zXA
zZ`znpS}F(glR%{k?cPoPla#bqT3XsGsi15!cI+)o%*E-UW;%nxj$<D}D8vEx>9n@E
zH1_<tv_Hm%K9SiM27^I%lPmxW&e+fuhSsqW3jR3Kr1glwpL_$Fk7r(f{yX#)dtiTD
z(Sl9UDzENJDJT%XT-5h~+<dXAF&^X(?z;WNTmkR+6_#hu?u)qZ9Pl!5%6eSk@$9|t
zEC2AJWzqX6KS;>$@te*te>dhgTVi8p7l1zRTg?3T+&j5xkeT&L*wm$RZ;ty&J9+Kn
zn_Cs46+GLBWdWjHVE1l7mHG@`T)9BNaH1FA-X=aU0i9!i;@5~YBk+h?iZF^XU_T%i
z0O0%TdAn0vwr)k2^$J+r#7qlX7>shXt+8Bu7@#o7v(d)z`!BGc+u+9>)#K$^t~>b@
zhhNqfUIv^%Oe_a3V!x13kanW-JM4{U^JmJ#U0gt^#@zv;UkVt?SWdt)Owet1#!vl(
zdw#!BJ*B#8)fyR9#S`a1U(kvvw2L?wN*($*2SCvT-0m{Wh^!ktU6MWtybVnJ>xEaF
z{k8!8RCW;nU2QY7U1(_sgC646oO#e_b&<(Gs{@PQhWP&@F!(QV!hieikLL*>5Emm%
zSS#o)rmI(Ro|zeBV&?yuH80rgoq_1t;m9rT<3VSl-Mn8=a6z_Y&8owXzY1`#`g4`O
zo*p{^W!zmNN(RmtdMXD3<`bD7fJ;=t`mgLHhzq=JO4G5x)RHsdBS0@dpdHBj+j~zx
zcncnPc345n6_Z4s6ciUtF!(?Y1>t=E1o-hG>Flg5VN50Q0g32<Ih}Wa6@Ma-bU%MY
zFdyLhSIo;()v!;X=evzTp+(ms<yEWFe8SjPofjbp*;e4PKcS}|Kx_me6mmL&t^N>s
z_`<K`-~lz#14er@*}2{@vW*st04vfW!hFuaC0D>l@eyQ##ZXfkd7|Y+lrWV_4FV)b
z8W>5gJ<5T%KY#w5nw1rdqy3JM@pEF`GGT(@zq!eLRUk$!fh2h4foLNU#}|O#`qYe!
zyZ%d)@xC56FnnOH`oPN2{Ob3suE|+|&}1A~k8;c^OmtuX;4xsT0pt>Lyrzf9-FV5C
z;pBivk{#qEw}J0adOPEk3;1CL#<MuKZiN_T05QOT&hU}nxP2RF0`eew5Cz=XdGyj-
z*l-?#CCxO-KL~hlMfaqhzP^f*k`fW9a0cYA9CVtyAYpnOrhoWO^>?eDRs8eGW2_Pr
z5OtH;r%#{m0b(Q&H%(dJ3YbkNIXz&=7YCssr{M!7*x5xO`~;;h!^+MrD*)A+n3s19
z!Q@28Ot=3sc^G7*O7ZPRk(XP5W^)E;X?uQSdpj49H$k^=<2iSTULE}qDdl>J2SKef
zmahV5pPZgj*I!nxn)w-#@t><CU?n+^UQ&NJLWXc2Is(tC8ptWSn_&ATLH3!LnRy6T
z`qIYUUQjygONWh|qad8c2N4Z7H@CNl4L;%3;R64yTet1g3{(I^<Tyl1-xB!YipRrv
zXb3_!Wlzt!(aoHz9&i2ua5bzdBE$;dEw%!a;Y4f<j61BuVj=ykdepH45I4EMUqmEC
z#CPzt+%6!j7v8~WS|AE{29Ub~qe@#qAY&5aHI!EA3lLlwDg7zFoiQJm#Y$NK+(9H+
z#o(yPmRZ(#z8%n)fOWe(e#TfK2LgJ}Kwg=ZwYFzu-XU8L$QgFXl{oc)d?et`wteNq
z10#^vHZtO}s0h~J<5~6RE&*gHKkwcRY}R%%GeC$?MezolVs__D0A_UtWS+lJ&hq^|
zgIr714?y-g*NC;e#Yb9TBt;tty=8`pr>Ey3vzL-xfK|QEJ+w(?_!)WhJiy#mj%{A`
zEM@K+KgiGY$0%%&FCY5)6=tEI@84VD$NB01=v$UTxY76f7bmjW$$P0PX)0#@dj9(V
E01T4H)&Kwi

literal 0
HcmV?d00001

diff --git a/_images/b63ef4ea81a463798c2aa2540688b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png b/_images/b63ef4ea81a463798c2aa2540688b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png
new file mode 100644
index 0000000000000000000000000000000000000000..e12a87875218f9987593c389519ca6bf18ff0988
GIT binary patch
literal 41697
zcmbTe2RPRM-#2_(k<9EpA~MP@PBXJY_9lB}k26JCq3pe~DZA{E63L8E_Nat}lPELd
zet-49uH$~5=f0lfc%J`p^#A>SbbiNYyx*_&IUj3js*(~j5~ENk(rao;Iw%w_8vgJR
zor0gdHH)%?e@OZ$8~NzE+xy(H^0Gr|SowIkxcj&`S+is8yu6*<-9#^nUlbK!cl7b`
z@RquC$@Rbg!bNv4hf4$e%^%?+r#;k+y-_H-v&bJDhun{Fs{+kyN(y>@S=iaTW_or9
zavOsg!8m3{@664#l=@}{^zc(gs5NF%>daOyyzl4!ElkCoCP9<NCMzrRrN+!lp|g)$
z{LEuLvE912V#;EAX7c!y>5^~c9J?e%@klx{g7I>9S3AqEEcM@C7cKGjoE^XaC4@JU
z^AY@yk5}LLP7wS{PYyknf|61@tjjV2`MKCBw33pNUcT@I2PbC%aZEcIB_*Z4BMAn6
zS#s5;N)dib!x+lK$;nwZ^nd=SWQs2#U(`$ARZf2XGck<fe5=pl4j1}H%MG>bo%Ls$
zJUgC^i`LGw8yg$<ynoNX>W2JU)53zarl!W_fs81ctYGj*mRw+KwEtBlr6K`A(*}Xj
z)r`si-%s@a`ZAf;q1jYCnqLd!WZ%YfD1M!v$5WtSCkh!@n{By6uKP?~VZJM7?%TJJ
zt{5hpm61}ln>W*5e1c(E@PN^@Hoo(gHuMzf+sj`9Y%nva-Ocw8BwqR~heXj`DscMr
z?8ij)xevK=ale-aJHC8jQ;lU-fFZd)dh2g&<;->#465z*1BSARe$BLt<}ePy?r;9m
zmX?-!czRYhSC;A)fBV)MMK5AYl$e-UW%r&o3w?(yiRWfl*cm#Z`+s~G{59c8Vv8pF
zV=ovQ8cN^WDtVWYJu;%Vx3|~U+k4vZg@tmqlrPB=tIWNJ;iOF8)@NJx4t~DTZo#-d
zIq2?(*-dcebG28!=)b!jrdh10r4`RLgoC1^qmw>9{5|AcKT+#&lVLj^rvG|JX$4wq
z>_e6$9?J1c>+8<nYtxP0nc}xScCf~(m~x}Xa^3S&FCy}Fmh%P-IksOCqU0%jO1=fq
zX_n1gz4c_#n8hh(6(@|^r<{Oi&+%~X*>n4}*6BWt`G;*1+jDDk24m(h#RR2W?H^#R
z;Gkq>WoZXj<9K*@Oh<EMN9|WWKE2^MR*^v}Q->WJm7AN!UPOIu$Gd!ZG_D)&Etz5Q
zBkL6&AGK`J?qAW-P1D-9q2ZtI#LW-C2o5SqtsL~4X!=RspRqxb<7`k{(E>}0SbI65
znpN5lg(}f`8rk_<zubTTWovIQRdI8S!Z$9^hp%4KYdyafT_et_+}b+JHYrcJt-&|$
za%J?jPMPeKqJgID_oJUgWVwG^{_BFGXfOT>zp|?6xXPK=Y&XO1=q%-Du1nQ~nGBs~
zM71scA=TDMKJp{bR!S<rgF=am);T5^5H{X6NHa?cRktm5!$Fw^T%2Ei$&#<2Z+zf(
zMw>dp?;=g}zSP3v>c;TPTSA^{w;y7*1y`TLwv{wo4$XERC0@~~sjU^v8SLxBMd`h;
znEke2^Rd~~;SjSxfodBaIQ=P2L#aRQ&DXeF?kH3a^9F8+|G|^j>*KrMza2G)1Yx;H
zRDN$<OZ;uzJn46PP})}Z9x6MYJnnm%@Nz~q<^khs*6}JCZG~iHMZ{eSWy|fI+AF_q
z^Uav?{bjS^<H5&qG%A+LarqWQ8}nLgbfCGSMM=rBDpVzg5eN0$yv2U=)sY|UB9<|p
z+a)6~u2Qb7O4>>TQ*mb^-M<7_Zu<HjY)Rmw8>z_JcNaH3L>S+emz9QS-z0r^oIOE=
zdh{mo(bD0|j=`44EI9?XCKB6S-R>C0h2z+-7!@+q*R8NRk#(AcQauzw@Xh&`>m|#*
zA@^!3OlTaWN9)c{iM@C%C0>^7Oo!q#c}k9%s&m3aUAg<KsPp%q`AXA9j8SeyzRvCF
zRw|B_U)=2<vc=xF$KFYE_^MaQ_=`<CXnpwH%+a#L$lA23mxkJ>=M#5?Sq(*><DtfK
zrXu$TKk(88QZkMEcJm~K3jaKgWnZ-4FDv8}IUm`UGIG$Ac<&7?jm3#}DuOp!<2zm*
zx&y=^IgIPiOXlW?U#wj8&`Epq(S)a-c>A_~sEPhg?&-0zt#-{Xt@rbFO5@|>gG?LU
zuF;YOZ2h3btk0aTH;EY<8WO=il}x#w$vVFL>#O;}rq=YDDLMO!64CiUdVNMx=WQaD
zmFFS8p>9|?OImD+zH{{L%Ot<apQBJ*v-@Y{j(mw+S9?wG=;>(kiO-mq6{e;!6+An2
zZgYmT$l-gkQ|_ba#RX^2xfPj1&&7)(wZHpbqKnuY*Ja+tb?f%MWHRtCxJDXqXy53y
zd8104Pos*8g(c?44@VMKX?nH?d--#IukUSBu(HO!X#GR%{%Z4-g!^X_p5l;3yYKxO
zYBz3lOjO(4sJ~YuL*$%fsrn=$v!0Fa?9!sMr`$P0R*UZ!wxsUgV+zTxxRJK;^~D>l
zN+;c3@tdQk+dh#_eN9^>Sf#js?~uQlt^K_0`u2v)CwctKcTdSonOa;7DydlZ;e1{!
z(-u~~8~>XZMqY6-p6KgTCE*o+VbTj>$t=kZZ*Db5)%lZaw}dZd^e@Ih^@lod(;h+)
z{?pt04DImhOmpp=k&+VrpWUD2d6u<<MMYdi&z@!0*9Yu-ah*SJeXz4eL_{Ryf`4y)
zMpDfEGqbj~Hs(QQA{D+kdkL!Aa;jM~jc}arL&^GjWd8xdWbAY!s}TvNOv3Hqir`^F
zh7S(P!}}OV@5x@-r0=B$mPp;wrGyv!a>quuT5==niyP<4Yj4dn<Tg02j5bx(Q<maV
zvA2C1V5u+jE3_YF-0<;@xXvrL{2=AJe23KNaFynM=7jv>r`ZF^D+dmY7mx3=-TArn
zb^R~%Ot0Ugt$x*^lv>hkRz`xKcTLPG4O1^gl43r7{OGauL%VjqtNY`}3Vw@9<J$AF
zEa>o_9<_kK+v6}lgwW|i8$aJW^YhKIth9{G;!1-HzCzITMUw0tN|b!DAkogN+lHd}
z=sRvQ^0l<}0<Eo@xfbD;Jc6<~8eMT=Q?{a`FGroN%v!9B0__P|7v*Zk3n#d4rNt>4
z#?23n&fCq>@DXO|X?kdTf2T@u7ccW`|I%{7WW<Fy(EZj(xdmYedtP`8|2-eaN31M;
zFC#jce*Mhuc|U%fX#Rn-no}=5M*cn8A12}Z*MpuxLAAh!A6w_k?n-p!tiAcRH^+|(
ziZy@2?&HWG^<q>u;E~O|MexO9cA>JVic(vHRAEZaNTcXUD=*)>O}j#P%aXhA^&1{P
z>PaSIFU(|g63>eNz8F27quFyO@mW5BLJ;wyocP*zY`C!NsdGh6?BO4#!e*oeQS5)R
za~#daXF2Zkt~r!gi=w#%b!dGFI^z?Uml}!L2n|}A^9^>{XJ&R~Q0%K8Z(&xRzM47O
z=j7zub^DH!Mto$*gKGY;C|6a@HE!JYcwvAjgOO;y*Xnv7R%k+7C`Eh90!AA1DyBG>
zjuMNK|BLfCftHTk+->ej_*qKr<x-_KaYM=WVBv8y%9|{gjEukd?W81jh0$Yq(>G1f
ze=Lg)vb)<gzyD>H?&}SWDql1S=Zj%17^vayIXriI(K;*gQAx<q;v&a^GjYvDbCt0a
z{vjE@TfTAJi&>^x#f2ionDj}G2qD(4kITe0qN6fqa~0a>E;U-JZ|_tJh&e78J2L49
zFy=+x#B13}dQ>7pRd2E|({$Q3wq}f=<xe-m%)ysasGyCTW`!aqEgHpnBV?g01hrz(
z+(fv}W6#4!ZZn{Q(2XKd2)^uT#jj^{BV9@Jy?niYzg@87_{=fBL}GtzNJ#dtS|Ngz
zbgqeE8H>u%DNZ}Urf*C1oI8%?K47I~UldshoI(?yFX<OW2i2)%AfR>f>eB`aEN`8B
zMtf|CF`w2ccoEKmk5pXknhLfyufA}Kq8BJqiIZ`hs=RkGv~|QLAD=vW)JGWAVMLrA
z(`l*NE<nSJCm)lkPkNxdT*vwGr#<@RZGfX5T^&YzloZMd1bqvBVFYN+56*s!YfAp6
z!RQs3lSv&W2Mp?mkCaoeNGOXJ{<4rk&I0o)ykCh}bVN05ozd%KliFGp9hR!S(PxsC
zE5#-*^fpiejC*$S`|?lM8=@=+$Rr<f)fT3#yLTcWEhRiJF`P)sytT+)^}S^TbAFXF
z0nTqrRXv^71bFH!6Tw+qh4O*R3|%@OmWUgpw5s01tL52euY?~D-c66lZ9S&a)Qon0
z{7C(DNKo0Rc!|L*4NRx@>@0!ihmRc0JoVRWy43C1OP=O36c`ZBH_Bfl=p(^USCrr`
z^6870=;()HD?;Wf_wHXKTVtZsRxU6o|7zSpn)6xzt9~JIS#!%(UOQ|d{Lnxd&+xKC
zlfGwoRC!X-`L45orc3Zkj`sQ1?>WRUj+V7djL4;2XXO0h^RYiR(HR4a)yRsUcU9ck
z1TXOLUPppCaEtgl6PEQNyE#o01LHeS-%0s6MIUvTr0w}Sf5Nq^ycS&JWzBL5ef@mN
znM+!yf|lOqghZq<iHLeJPYQB!Di1p=+uF%qMb>@Oamm|Cslb&9{fA_(`YyCR$;%_&
z+U5FK;(xDaC4k!4@apikx+z(6LrE#}d?cqA>vx56niz86FLcXyYCM0sNvzL)c-!If
zJJ9<EpWLmeh{?M>y&pO{BAWe@liZb$iD`GKe1i;%@hTo&_A#L>cyyM$rF0;?+DAa^
ze7ERDEn44w3S}1jVC5bqrG~TkIBxo4_TA;pEd`n~yCOdlT;aK^T#k0be=27aC0Dgg
z1!QLjblPL~dP^+%QnoG?X__9o696mQ%&_-#>3y4i3O#c%r0=1pJt0czjG{68W<Oms
zsge84Ir!C7KqoD)C{&M6mfboDBEa5c`I0QVQ~P7tfkd!tm+N9jypi2P#7FdYzGwPB
zf9IL#3l~y#pROpPm7?0}WWBt8OVznZ6Xgsbg*<Sn?@jNnObpk^XY@Vtu~Hzj0a|ET
z6XDSABEa+`T2#`6mRFbIf?FjJ!?J*W;$&A6b)`>M0@ns&H8fIB!|ixH=%XGzQZG@D
zsCHrT5;={5yTFztw6X`@sotxfC%{N}*+g)pVluy`4Hq`Jh@Dil-R-j~QqfP7@U^pI
zl*u}fAJ-_wTe0zf>W?jw&&Xi$qDG!1f!j-+iY}%Af4ENGSh5;w%+q-xT1rY|YqJI$
z@8BBku$q!j*YJ4BY!V99?#MbIESA8Qp^J~Nw(z7Rg(Z>L2-m_-kK@TCd$nr|=P{Ie
z5=MFIN=m6}_L{vy+h!w9+=@0P_Lv@RE^#z-7ZjjCqC}_p1_|q#aG!B+0&Zj-vqV%w
z9Xp#MAdWt-J7}DzZWmm8@)KCF%(c}@EN7BSU5Lya;FSg9&z(6}c4a%O4~LeQ3SObc
z78A*<Wp5wB9Nq!&!}hTc8~^ka1@{>$FZ!rHqQ)q8PR<*)r`4lt70(m&5i@pLR~7p7
z19?y2PF&-(HPX`>d0Zo)XG9Cn!7K_a-uAIT21~Df!84{{kqH>+U|0*roITa`p}j5w
zFkKl?)<a4{OFpHVJ3w`siMNw5(AFJfz{VHSliT1h_rW{jdR62JWZ!!pnmMhkWegv{
zH7bHZA;7;z+tnb!MbKxGU8g~yjD|{~sU>->^<-#+;QErZGz&0>2KWj>lOiF?eegIs
zwI_3tZfTO&-%loac6<uho^FjgmEkogShcXSB5f;SQv>u$DqX6<#&Bm+%S_LRN7z*9
z;;4;H&AAA0a)vks*GSh^+d9>f2u+lt^j5l+`^=1ho}!$ZV+M@r3sw3haP_sEBic0y
z@^nt-RHufT_sNGY>tK--nt5Ei1aagJJ+)KSo<4<(Bi?+9uuI`Ve!0+bz?+_hS+{#g
zfIRT^)EIe<DNFbq!^sp~SdLGsj8SWlOsgyX&n?(5XnjXknnF&0tfe&5d)pSGC$-9U
zL2!3=`WJcqRws{aulh`xfcr7)DfEIo>|9vy*CcRLoZJPB?fp`4PN8W~<A}JS>owSQ
z)>rD<6XHl;daNG)$9}Ri$pA`hvsp}UV~UnJ>@g+Sf7_@xVUiMdl2IPkkDN{R-AuFh
z+|Qrnpa6`#xWz_wUcGI5b^ON6*zT9WxTC)Zg$9)-9vDn`W8>vNTg#y=Xv{?Z(yv4T
zzMhlktF!OkDP4lL#@QI<QJ0m){P5vJO8~@v8((qWzI{7XWu+Lf*-7U({+#9L&+4V}
zX_M9fN#sI5TMx*Nj*jT1e0XtjaXr4a;gB$h;QxHRo5PY@w1qTWcqdycSg|3?DtXLZ
zHE%$+;1m|7G=J&S4*!Y>8q6~HsBffQR5meTJg=T^sgHJYa!MD!O`DaSO-V-=p%_7C
z_51t46`vmzWUSIepi2bo%|&n%=P(a6$s5}u#e;pExUXv!n?8c&<r&kquS$CIBnsFF
zs;x}FT&vPVyrX&fQ*oj1>e!>WxbDYPJdv)eKZXnafKXk%dUfcv|DE3EuU{>tcW1mQ
zlqtL8*wS`6Y~P(!pr?sX?NvqGI@dUvmxuBddwP20YwZVyo|~g<e|LnF;-Rp>6($<Y
z%*-lY7B@ED>6KcRM@j>yXbtwB8}<1|QYHFNbLaGFsTr_YlHR9KL#_Ayqn5V22Xkah
zT}(lk2`Mh-d2qO?`mVeC#)rRKpKeYxy0c|oc8|@=%hTDJdT>&Lo)w9s8|pc-W4QZj
zwX7IC_>K~PEv<}}qfl&YY{=!3l9Hla-*<H2<pv%}TKLToY+)y<%v&&FadG58zr#Bs
z$j=E0QLK(thN=`^Xb?m>PS!l_oCn3}D!fmHb77IXDEb-t6Tp{6o1d?%lf|N+_E;{p
zhZ3TK`1H$+q;G+4hmyCqf2O}i3N!rW{w|A3G<`5k4nu59OH1L+Tv^AVJcSGXF1JFo
z?FlJ3B29=wI0NgHl(?vt&q_XJ>pBa)(4g+=jkE%gq=?AKx+ZHKHbK0}lJTeI;NVC{
z-?<o0!eH%dx-^*ExM%mGf(%tM^09&{3O0Zm#e#*NmY$2fYC?PLgB^VgrlIpUh=-^k
z7zDcMF?dd4n=X<%*J)ZH%gW^jm0Iw;bs&pOH@GUopKAYIw}9Q*JHKRt`VovpdjSe=
zFL9rQJDrtTo=fJWRh5k)$a1(Svlh(2x4VBgLF<DH|Jh!}L4^{MhhOKuj~;QZ=P3XE
z>89M|%q@T5dZ>BFCjd0H?TQnvzUu)R?8i<RK9O)mp?GfQ)cIKizJev!+3=J+T3T9K
zAZMKI&7GeXj}Ml4L)g{AYpz5qD=B@n5@4DZ%D8t?6?L*k@K2q33$qCG@)nltzz6r~
z22#|<;qO7ZX@l99zBqY9%6NhMi*zXYr1Kgss}-RodU0nGm0ce&>^uY%h$p#nyQn=D
zdM<z?yHjY!m{abp`{z-XAB(PqGqRu_MMl1bVIPJi@AG4*-9uVRiV+6~#}7t^+m)12
zj0m+cYIw_P{JD9e_m_a*q!Mm3L@C$9V0fpOofbHHu|JS4U52$arYbI(G$M|~P^0DU
zf*yFq<131cP1eO|>yo4-gK8YBa-~UqmHhx??{=g67Zn|yvpTxE+}(tvj1T8JB5YtC
z`W@{r4OiQ!?)~}m?bqV_aF@XX<*1<o<ueg7PXwDHp;Iv=J(V{$-}`JNuh3?bwRYhJ
zb;cwP3M$;Y&Q7Q+xcX&!?c=usV-^-{qnR#~!I&$3dZtn7tV)6JGFd}_?1YfeP|wY8
z%Gil&Jj3T^3XYDvtq1Gq{v4U<4>xu1CF&<LK=;(Yhg2vBo6`k%Een(NlWCs7Z7yHF
z+@CAQ2BPhyohcy2N-8QQ(mX`CNQK(gTwbI+>GSFA9V&Tz^?x=mpRcVZCVQ+^c&O`Q
zA5nzAJ>Qg)0Z1s(jn2x3Ntv1#FUWwy7n&p=QA*#H+76J~OKOsAT8QSc@jw$S{V5hf
zVdo<{=`zB1C2++Vi)!L0wfmxh*3jT1YP-n!K`C@xD2u|wuztWdUhK<4v`BN{L*Era
zCwATg$YorsNcI%E9E@FpjYKL}t&Noi%TpJIcFW?Bcw2xc$jg3Y>2r1IsI!kC=+jG$
zzKv1(0QIVDYoQlQe+hX2-N}DfNV0Jk=q(6IZHUrRJ87mIEJz1ES?)4$Wdb{ZLJ3?M
zEC~M6zZU|$Bo=ouL?TAD1{O)gH>4>tBssVXIN1{l_0nHBH)d6(tF4BSae$It@-fLq
zTVdIyBIKAz1p93L<mH2r5fkAQ)M}_E)gR6PAh4jtCyu=14`E3IErV+fQF}ZSRs&=y
zG$)u(`0h?u;`3U(bOD1NfD?&6+U)_43Z9er1B|oN$icNclfgnj%ZsK7JbfV;fyx)N
zEzT%d3fxdi)%V<0m_0AUr5jahoj_(YITpUz8U@2ib9N0t1$(q88Wo?gW6H4o@KG}R
z-!Vz)dwz5=xy<MHHwq#ZMWXF~!&C%18Gv|<fCc5J7zLW$aJyMjPMvu4c--928wa40
z+42B&&FV1f$7&f<;^PdUUxjQvd2{tiuI|_@M}Zgi^FVD+UlTE^X>*+1V@ywVO(ocF
zhClI%wvI}wFNT$lH?7rN6`)V>6-h${ja&z~q>|o;yO;eKuXP`VW>q{v1trV11}8n^
zd$m$`(dq_nL|dP*<g=VJ&SwB0(X)D=SUD-tN!~PKKQ}IMu|%o3X-c6e(*_@ez27T)
zD0rGePgAQji+JS}0nK&lKSX?fqYsEl4+(`FYw!TUF@>J3U-3-OJ9}m-fh=zFvz)3m
z;AXwkPG&MguR7jrbXV&SeYcyQh$PZZBdokgK6Mx&QSTwT&*KM==r+1CH9^Q1&Ad8C
zKaQoYD@?M*)Dib$tK>5tl_YvqnlKGazb2{oP7zsCeek@pot8%V<%d{T*An&A^FvO*
zxfPNt8P2KNAZz7p`N||wWs54i+^#~e1#cvUvY^k)fFC~l6!mwxP5QT9)G|e{m&R99
zgi+c~j43=#{H|0t6<|!9%Yv9!+T6uDk_z)H%R~NUqZfNr6{<RDzN{&7D-OhVS_!z(
zDFDtRD5Uw!fSrga`e9lp_*Iw`1D{W#Np@Y&Q9N;uBB<!G6WOqYs$8uUQm#v(iavek
z_gaxe_NCXA-p}_*Fxaa?r!g)7&c`Ql`W<X8JO$YSXuBk<C>rDN5|be4`PjsjqqeY(
z=A`H(ba>LgKT<!7ZXmVEAV33oq5^ELXd+!6O|Je_n5nCVLrD9KGP%tKF4#A)pksN6
z;vDuL6*Q3e>v(3+@J@>{y$_(jVTs_M9%5p%Fhi$)XN(JDmx!^0>WIR?Brpi2PMd+)
zzZrvFHBU*CtOP1)KT@ilxO=qMigUc_-?Qq8*kVA_`v96z?`NYE;KCOTq66{_35D_u
zD5ZJz69oFGqQgqyfJ(hMXPRD*@lBZ}eH5};3#>f}><*~#apLn^Rm46l{P+co2gZ}s
z9>c5{`-gE^D79w|u^TJiT6S|puVP#-Xq^Jo`ir~6=HK1$MbWVwr`27eZ*X-OnPiKZ
z7DXFH{2RF2aE)~ip6pnEglZZ`WUxdC<w)B77D_FWpY-`R_?%H!q8I5C+unW0wJQv|
zR-R%XbpkqT)L*W`?D)fi2vrwhx-oKXC8!ReN|eUuXScK3V_&P;+f16|xtJ9e+6}-w
zCHTUHC~EM}Cvs)=3Z;r5+5#ASejn#(PN8==v$qWXWU^#y{`1Sk0YVtn+Q@nJXNly*
z*a4e48SSuye2>&fp6(IvbX!c(85+KMhi=4EW!9A1HZu*EDV?y*jj_|Ig`tF@yQ*fW
zWza?V>}>DZbhR4AkCR5%_lN(jw)ujPSdRB2-bu>%{~1p*8_9=0YbqHc=o6~Kmm_vE
zN!&WU+C&$P^@dKUFu>0DO{>i|L|)jKr-?-o|3-og#!(^O=96Ype{3egYLq&u^hhww
z?S3oGK$n1MdD%uWpV+*TStz(sh%J63!dV|NN9Yc+D6U%xXcsRMBOf%F6G~O=sVT|Z
z9K2o}`d=3V3PRxp)M)^#2J8O6mTPzEQmj>+r7GxHFuF9+a`Nq#-4;b1Mhz02#>!ec
zc@=iZvbN1&FCjg?G<A4Ey))spys;QGnJSw*|0AVwb^h6x!k>P{lQZD{eZ-Rm_pt!j
zgV8I$>V6p>&a#;+Oz(yDkeRulW#HLD)ZN?LtDvSvL@QtvL`crcE+9Zc$|UiQ?p8oB
zW^HQj#hW8OZGQyLnl!jPKK}bNaARku3tV6XsGBu=5#&DjQ<|L<@~Fg1<oN+Ac{8|F
zDTKtBepw3LH*@hV_@S2N{~M(Ns-+Et10}DgNB#N1fvbAD5Dwh6uB=k-Xb(>@oJ3)J
zWpwlJ?w7*b<IlsrrLqy&QrCSB8nZ>BmB9Kvw@Zn*c1*QGAwE_tNd*p1TE1&V_G1^m
zf9C|KhN^NN<x|YVLwzc~sR|g1l8r46jBBIO&s)o%5|WZaFTFX`Ir{?@bp7eo?pT(^
zRcT7fYjQ#vlhsUj$0mi5;yWo!#v=7FI+{f6_87_9RJ|x#&SgRf)%GD99p%c;%}oH_
zh9bZyN=8OL|64#knTvDK;nKQ;F%JyvSFyb3CbSoX#xg`5P$=+z8v*1VVkc`m8+<pu
zDuCqavo=MuzrX(|CZ;oAF(Pl&d2>Yfrqx1E!p5&(i7$Qw8`wwL1=svlv%YRRfA*6y
z0gX{!1rUw@2bWoG{PF+~rB`l1s-U2trmmix6$q@d%5_=-n9ZlM)_Rx8kKZ<7;`5Db
z?G5S`l=Ox@B~opAs-220rI>C66ZAzgcijW}ipsm5!bdx3SZP9oiqFU}b}`*upT$E3
z>`XfB&Bw`AdCsfMy!NNNcI_Iys6EO0=hsmP+Lg}Ee)Oncc*2uJa%_@Pqd_z66k6iW
zKt48gviHM>uFd)Ggv`uHz_~Q{W^Y`2ZTsS)k@D`KwE?{Q9&k8e>vu$`ptLl)u@|@E
zz&$$r_5RY{VBk@xauhAi^ymHGOM<MI1m>ZcGF01l3orD%rFX#O4Xtrv&tm(t(RA-q
z1r;$EO#b*I>-H^i(-4-w*z&hi|LyIrGyt&;2O#|<n@#S4^aCISKp)eZazO_qg3^06
zwsc9zNijBGST#vzVt*m&W|ha>)sj)?t)E}mfUX-5uy7BTXrDv1)!!aRh>1HyGOAsd
zUpS4FB=m9`5@d=w;i8&+S5pPexj8w{AT_q8;&EZ(QAJUqNs8FS{+r_ixu)i3;BEOs
zlR8gFpJ_0mczvqn{%(#{n$k_DK3I9#is>E!UkyjTCY2oIYqE%HXbHX;{7*6R#(8E{
zZW=HTYqUR5@|KqDz>hqZKXPZu-k-uAHn>isP=j*E0oAU)u$(@B{@mm_e+tqM{U25%
zUI<h!LD^_Xz(-qRpq}Us-gtAF)~7%kU5ZY}wrc`OBCXBGDi;oa9k~1jAEllt)~i*B
zN9{uA2^j^SpPyM+q<n;!<6>h~<mFMH@9&ylGu718OqabcrKJ6?=vtD;@zMQ`W;{Z&
zukSCtdB{%Ww^l!n{R2%cg*cQYBHpfrD8;9m71X?|BIp`&8{mw@2-ASz3YKxJKSPv|
zkc_zvre^rnU2l8>0*?Qg4h>A*Tgtve0TKt?tqr0f+2}i6u&EN#(}^k#qF|}y>lSNb
z{AoE0`ntqwJMsk%;%x}*+HJKJ>T%B`+YMrYlze;4lhf7~jEB&i<|mgue0`P8%$T7Q
zXg$47gE)C=nwniOejdLTu8Fx$jjS|)tWxbT$Qp3C*<Avr-_F`}#I3-;G%!`&Py~ei
zcU)Rz2m$54vl8fzcW|%hmg868=@S+EZdv1ixvHp`|7qC+HoNrZJ1WPSCgFg9fIOXy
zZv7DUuMunys`f41+}zUT0%gAs=H9~ACnO|1ijEf8dR1F1YX2eYF%*ue88*;x&mx0)
zk<ZbFDP-A2V9^d^OO63~$8j$ZUj5C{pEH4fx6?Xr3u<hi=PZaE#C{13vn5LI+H-Aa
zZhcmQYZrDn0iU>dt5x7m60G`mq7ai`uX&x{m+;U$xgMLzQos3D>U;)E7=CFDA$m+;
zK@}$O$=!ii5m~VmmDtzST#B*wx`Qd3-nt7ad5EWTzQd?uvYPxZElskfOxiiq1xL&@
z(bo5YCq}jQq-W2bB|(Q@b{s;rx3|;7mIH6p=7lL062FL_q(mM5w(7MKXo<I5X$J0H
zM5e}(!P0_{o^EvSet4Q{15}RQCkLx9-dKaFzxSgsZexEb*XGAi{!Wt&1fPQ0>4}q4
z5;?ROC4tb`W17tIQD65q=ZVP4RSgViPm!}mgJAb{q*V9H{Xb%@M}KZ3A{#lYblaIr
zuY%^g<KEx*=j7BOay#W5m{=4&&|##~x$vwdhyUDIf?VzIxd^ta<)hZ1g3wF*RV|o7
z;}G`vDvyw2kZ(Fw+3wp;bVq{7Pp6>OzG2*J*?n<=SBti2I;$2C;De)m2dL_!0z<+Q
zejAn`P;hc^D6CI|esWc))+P{i_sf;gRWQjShrIq(NeKzT5c;6;H!oC)iBCx2*j)qp
z<^lqr*%d9VnXyC#B4?7<YF76l*%bMD9U51OY2)*mMjlSiFiZRHqlJlW<}Pt`0nF9e
z8Z=Dapv*2aY<?kFn2z<|bx$j6zXK(0qnj--&;3iQ*O#0}kEpNKRI!J$w8W2%PmF~Z
z*>@MoGiY^bvenS>!o0IfNYJtc9tumiPCcYNr;>Lq={zb3L<5)Emo%UX1cN3Dsf2U?
z9yEQj8Vx01JHsBM_t8?7CsHdv;HU2e|El@<`B1nDWua<ZSF`t0mEWdy$CJvyLeqNZ
zFeK#wQKf&+64GI`!_y<OJ!4B?m*Qk!d{~UZ)|6O+CWWM=n5F$FPM<!FB(QK$AknI7
zX;IwU`e79}eU>sI!~-~i?Ih7ill&xn8a1?{bY*&^W`4r8g$M<TB_SRYY@J6LEC)$;
z*(}@h+wu6ruk2M%i`Fzz@<NRzJ%z^bbZ1@(fkimwLjV4mJxke0Mm9j0)2N|RU7n_w
zFBw}K4Lm$NoLb}|{KO@gV^?KE?ggb?r_^e1`A3o5WqF1Jn-r;uWw;x@LK}GhWPyQ!
z9lK*SVlvPY`D7J@KF|wIR1-={;D&)bU|(BJEflT4V!8A*%_4-IXk%j|KHw(oNU=p%
zbe@>k2)0Ds-X`P0JHyES4{5g^7%%JskW8f?{6(p3zrUscV3I@}9TJww!flH=(K_v}
zN=54*$qA*X#jbJcInS3aeI@Se+y~6$d31PJ%w>XNQ2KZDId=^urNw9zE${RbW^puM
zyAQ%h5ZwoQ(iI57E$pFxz6veQxkr3XHC71}En!%_v^-Rl`t~40qnAG$H9W69$I?|(
zXRdUY$G@u9s^b*&A4UWdF{1<agi<e>M*=Ld87a|r{8)6l_3w^p!1k!&e->PjXO-^B
zLuMvd-`7^?I6nULkhxwymdn4&8$uy<ZnKPxF<l-mfB$S#;Nu7j3-7Izzu175l4nxy
zyoaqFTxdnMK7M0VmE8QF%#Rhdg*=B;6D77rZ@LqpvboK^3_C^2s21DV*_jVO^~a}T
za#SAdEJQi&jG|jS`FPkSai1kJgCI6DCo_tmMewzeN0|TqJ^WzzCsXT94UJBKrzaY>
zZ%$55)199=|Mprua0DCpXbs$1BJ=&$%{DHl#=~x@Uxa9lClbn6ui}3E^a%%b>;5(k
zI5@VxTW&K=!Bh3O|5Iz`eKBqQvF1I1<N40l76v)0?pQhiif9TvmX-4P5d`SZ8g#)D
zgGxp)5J<)s78mEBq;2f(rXB1+y`}hw3;N^^JFuMjNuy(YbuZ+}52Vn?f5zV+JrNq=
zJm`WYCiQ%>f46k~{Qa4pPuAb26A=-iID3}!-)_AUZFi?j{43Zyt8ol;y!`C7=iueB
z*k}lKeFL5VQ6oiY^JRuFE;e_KkJF(sSox(7+5ZObA()IUy1iO?0t47X?+WN{TY!5)
z&TC{ENsl+Unu3}<X3+@75W&|G?}Ma5Af4_4^3P^RSRNsaR6FBDQkw{4jyk?3g8pNp
z(jMCYd>R@bxdMfd(?>2?Id~u-HvXv(VF&B3pw1lrpBte!H7@whP7<kQTOBsBdMlPZ
zA|hu)*ewAjmuMG;&-veG%|V2VMecMLp=-aA93;l(wG_B<rLv#@qTvk$BL)?8Z+qoh
zE7VNB$gR+Hrr@d2z9J*#k;jQrtgBP;5(S@}_-_UOf)r2GA=Jkn&N)v(qKIm2hgj9{
zW6lv?P`mYN@laBgV*n<m50-clR#_mz(cP+{CpN!JUsdVW?U&7hLWGz#J`QA5y?=k#
z2sl`4Y=R6_{<v9d0K=CD2ZFFnxlg8o0~`w2SA;1}RW^!RPPEu`GMfwRS}8IUk3n<(
z+8zq@ME8d{!t3Bas;xVTo86&I5QE`nf$XCmC|BXbSHUo@)a1XQ)!>_&cCMEdi)i*>
zRdDC{eS3WM#ob@MJi|41H?W^yo1Sza9gmgf*-}Ks82R|iB3mLiBO3~Pxx}8!=U0-?
zEnf5ed$y*MyuTS^*a@9mg$7XAea^SXy59HZnha&B9DC_$s6estWBA;^=V=EFKFBpz
zeWCVTI*H9=0j*P+60VU@LfGZx-VELT_4~KyLJy^w^N8BN2Y?oM#~w(nEv6sbel5ct
z%O&XT-7!p;g>H{kTs_xw=gWP%%Wg9WM*Fwy@L?VXKT4sw?L4!y`x?w2mIBpbf~MS2
zWA}ao%G~^7+AS4uWl$)@23X1sjP+jnkPjp62XOH}Wdpf}%StnpPN{4I5$R_SaqT8U
zvl+#mEsu}>LYo(Q@ORr#{fUHzmj$)*-ob$l&^kbS&;Gm5;viy?V?la<tl+0f7L6TC
zLKKYgX~<l&R##Uea=wlZ<kNvFE4{FIEo9q6QKIo#8iZoolU7bd@vvJVR&ZO~-rwf&
zvuS0$xBam>CY4<qa~;9~dmjF?U~y7MXRvfr=;124D#%;kz8|B>JF&@JRdw0G?nJ|=
zIfV#3v299X&xeOEypW%CYmaSpD^>_S1<B`jfK4Pz2sQ;9WFERu!#LoGl2<nNAR{^0
zPQI)U^{hJQtU;X2f7{-nFZsfc@fU2Z``?)$yUTg$(oL^ci`Vyzz@%(&nY4~WzzVL1
zJK)*%i|(dO1)i9mB6;KTpxkV1`v-`rZyX<eIR-4mf}rG?H-TTUnLuL-tvbRdn!Stv
zQ!icY+RO*nRHVGG$65(w+d8V_RL33?^pzXF2nP!f^Ye2IRCbzs<}j0+0QLI*VKN0j
zY5p{5*a`2FC0YjwGO`0Bv+7Ms^ge`pr+Cpr($M<{6xD=aGL)a6-&cT}C;cV)Y9fo}
zK<sNGC`xtbGqSJKV^QB*DolQXo!c=mFu;nyC4%@d@Pgl4ED)(fAVDgd%e1I6IwKpd
zF~888WUt)YB_b;7`De?i&Sg^PA7E1}&Ph84Z7<PZB6X6uk9cnq_81A9EHmWI!E!ao
zAAfbQ-U2RX8x*2{ON{{d|3l{BYueE0DOP)K<w^Ul=)v}4c%OEu8+Z;SE$yj1OR6Ee
z_o-H(Iqbpe9|q~<aP`)4;=f97rZ6eYr90?)|4B@_u2S|l1Fck<2~ThDj$k}OA-9<;
z$R59f!P2AHIXG|t8=-<ezxKBQcXh~7>_3o-pBEh-01it!Ew5o8ELG>xvM?Yl^T4+z
z-UP}$H*3bt)+7WkbZdQ<2~5?RZpb;QDWp$&k{AP+yB1N+#m~!c$$SHvR8KrdiQx-Y
zFvP<E5#~X&^ZEj-l@Iq6nvA2TIxfW%qgI^69puF|c_w)mI~j_;v)(f+03L@>KlVzv
zG*d9+_+Q=^^2R%s-Oy6Nh4i`%u^}1~WI4YSjBs^L2VikJkD6kwhZXxUBI2B^EGtxi
z_!cMFK5FnxMDyt%zjN^PN&pgh+w!wDDld;6VW!ZXc)dIo;BUPGTjf#Y1V;X`9)h4i
zHM`D_jGEJM0ReKB0FD+|PgGfXczY`Xz!<J`(ye`AKaiQSV(~u;XE#RIfv27ZUyCRt
z2sTGb8N{S&)ETisipChu9<y2-iRh<6{ju2EY8Xz=eW)8PBu<M93+!B6_>UhyM#c;}
zxyNE34H5;IfL45#iVC^)lc@)Qi#DNzX7~e7?otnBp)9anlS+*6(N-{pp+^T_T{G)F
zlk6*H{<zhp9;65e(ITQ<lgAt`Y|s$E)Yi)%pCZ9v*!M2ZyT1qJx<Q6%ekn?sFq?so
ze!8cg_qJDf+YBrM%}hZZ`8u04*o=agFVhVV52J!0wx18XFWme8DvTwU%s#p7{w@2?
zXN8}iUsJNOUc8+P>*Bgr2APfLlDB_6E!FhIBikh(*F{0C&$SNZ7WcRHw^ql=k>eb&
zH9hx#*)?aTA<+YX9YYnyqII{&KVT25tU6FASQY}>EpR%d9nAI3+~Y&l)CXXn&k2y)
zBpia6Epds@QJVVMD+scTV53=O7~xS)ZS2Lsb$L<+j6(*VT(<ez7VI_C^%a&eDhP_5
zsq_oE2S5_ar)S_{kdXnUsAF)D9N;VeeN`xcM9f`Y_)g6lU_23rWPfa2R(nc@kv%^2
zcA2U>4LXJ&OwBN)6QB%Q!-uYTEf87UTju8C;pyn;P>emoML}02Mr~|vR)cgWEh#x!
zwgb||Mi-NN9taM5yKj!o|2J_h<=zehB>k)<q6z*3)*yDptY||9pv3`@vR0|y-g?W=
z4ZyREv2UPsAlw*AXT-#WQ8b%eYh9xV5+V7&X>d;Z{@&mF6}-GezJ2@FM1`hI2qil2
z{i3NV@4g!3#s=Pkd~VCY=$)rShz2QuUkXP0w9l6PU``C^Ro}q#6<l-0eg`;d)a=EL
zyx$g(qmi|2xxmlgHs6ZK0@qJ<0{*CjPt0}WevGF_VG1qpZ5j$}@=N)BkZ8lm27knT
zc@y}it}$U~u(j{QhXy~3H<t}Tn~ieek&|OXK<L-y<;UDNQhTOU0cDveVA_5tCU+QN
zIBwW?b0rm)xn!_tQ|B9$3+Cwl6=gNe-1ZRw^^6ilZK6~mhbUH@)6JMQM$KH=80aw8
zdmD3!-pk9&s2~t*=|yA_QTR2SA+lAW(4*Ma{U8}_N0)eC-ZkciIvB3K)n8KadRs$k
z&eb`6I)y^o5|Z+OY=f)5KhzRlpOmGoEgum5riX68<HdSbb!4&}#M@~w&vU3&I@Xv!
ztd#L)hT3P`f)Rw=Lgys~3Sya`3$~^eH1zZ!ERdVM^f_I8Ggkur1~gSv8|=7SwrGt^
zv0&av+BYySAi>EVRLs#+QuaopbEmTfI^QKSn1jtty|K@Gqe|DPO=~8tvcvTML&V_5
zZWRc?OucVI5W6h4KbDtTmg&ancFod=>V@as)nmZxkq8m9Yyds%4ul#fCM9`7OnszG
z{|vyX1hc?{9ak=HZVqwrdi6h0b-&}=W3ZZ;lhxIfASUy#Vkl$3udGBiHC+Lgi73B&
ze^y^4WMxHr&Uej0SCsha2gjwz-dT39FXY4!fXA1j^*z4_*G}gDWcbfw;#g&<H}SYQ
zbwuDkH$lFPHf*8wl`7%4L0t9(L2!WrKQ<c0;@VYvomhC2AqYFCrg{$v=NnJpM&SSS
zfnDE{WOq&rJUDyg7+nPyOBdfUTjq7H9tx})ri4zplLi8tkWLmkmrYwK*pZ`Wmbg1H
zwve$Yl$xqL_!vIIBmLPl0i*>E!Nvq9`Wqr0@5gc^;leQcqQIpH0-X8cfe1hAF#3nC
z64@scpJcH3#~0cYCeEp_j1co{8Eh4Vq!3)@eA;)N0)IJzL;ngc07*@5RobpAwKiar
zozA^Ug6X@Etr-Ch!3-RJgZJlKU8hK4fY_Da-Mj6OP2r<S8m5W8))lnC15Uj!{rjd3
z)#PHbo?v=xvSI=6HwexEfZ&BAe<{r%kk29i4HgUtvl+{!mQ5iwf1w?_xOQBrJ(p=4
z<+7GBFEnu>0-y?MuqKq0x%G9jdK2^h@<)8EAO!-8ori(~NA6Td%k<%w7GtxV42bKP
z6f55=#B@UpqT2mrlV4$5qk;ecw@pvCMEfi1Vf6TDUcG*ukeeIVa_1+fsAy8Vrg8_&
zd(mx%$t%&^Y7x~<AtEUz7>ooe57?C53{}A|h2uLM{QTu`>d*_2=8>&2N{gGtEt#5^
zKhGZWgp%hXIqRlE^p(ub%=(%xD=mf|{R}ZE*MHYfUc-7D76?RaL3$Az);&7b6q&f-
z5Pi&ELy=Sx#C&X|yZ^2~_U~zknyBw~>0BuPXAh@<%?Ogs7+DAyN-#tAh$m(jj;x_`
zwYpnKOcxe%m^IkA*Dwq55Ey(~3Ax;pMFqw(Pq&M}?-+Tu!1rJ&X3T<*0fH-X;PBO}
zS6{z>KXvQYEszdr;mqNMEnhf(70OPu2Udh&PFze(m`Y)AQW6z76qh7nb|ZkX7XbJM
z&;?%2)hJp)Apr79s;X%RDbMe$&pd{6gB`Q8EbHs*h!O>O*ZdoX^V~U7e|_UWV8R*X
zzmS1G2?DWpp$aKt+yD-H3#IS+ZzDYgk8$sVT6oxb)b;Wp`jjZ*MC-$i4sv?gfGcoa
zLU(s}(8}|X{R#xA+HF><!nBbLIeDq4r}ubAYpTJO335>!k(`L3va#XO(F~o@=JTt&
zs34vkKU`3RNI=(4NKSqToyP@2Qvjxvfa$y5dj@US8o(q1ST3411wq?}CBg)iu02m7
zROsH8BLpVDJZr$Q-2b(hF5$|Um6bJAZg{Dwsp*7E30b0m92^At{T{e{W5yoLio?Uh
z{@6PZpZd0q2)sQB=RgvtU}lbnQTPTLcpK1e#B2~St|@wVKQB`8<!W;SR!S$hH*r#!
z_&R~MNp0an_LO$A3va=%3mF>H0&&FZ=y3l(p-0lCK+s$PINEeZ(rA^B8ol%>)nDs+
z_s$a994ZK?rS-c<l;PeynpyI;wp{SulvGrqFm=}~{O7w!T_&n<K<VLrfNUS=>2pap
zWfee1UKxK82iD5lPoJnDV7a%}_(jP3yE@`FouRw<4*JaXSEubEsAb=ujsxofLRR?z
z(NIC~+dMEUk)ipsvvyn9<3b<*i&r5~P>p?&1G2(aZ}A=LO<caTvVgydqeZa*sW(#i
zXlD=(-Vi{$v)jRZWXsQ&7LOHmA9HOMnGy7P3F!TRoMeI5O=@Lifi^ccr(bqw1>1rP
zn2-;A9`=1ic{zWoplSHKckisHXJ6eFMRJV*UZFMIz-7s^bi4r{rWbxIFoY9?wAij_
z1~`?)JN(<$!2v;5$Q6+Q81QwoHxF2#3UnFXx*zuUVD}vOoSU1QAi(3n+XlmW`f<G}
z9Y&iU97|!(Z%T;A3O<=1SdQ^fKxp_4l1l~>{{YNE?4gHYVGOt|s1n^`GTHqFDpZw}
z&{N{bmK<o3Ou=QqNhP<l)GQEB`N8quJ4hb4)?x7C)=OliLTnGt3_}i&uVw(5$^LX<
z08x(~KW>M3>A>@AH*Vmepxc%EkXu?G?rl!A+_{8|fsFr-5i>3XEx><JQ&VdL&lQ<;
zAjfGxWdFR(0;-yk{fTWF=ql+=_M2?T$u&sIF<<dONlHp$)(BH>eawWU)?}za@MA5%
z;2Pf4;|k`8YHAf~W!uFC<iJfuS$Wa`|J7AFp$Kv|x71-_ao8II<}ay{sn9I0y|RT!
zq)5`vxt<ZsgbsiV?$woWq!^dCMDi|>xRWErf81T5iX$&nockaW2E2Y^l|Z2l!hIe~
zA6P2Pn$LnUP*Wz$D<i`KK~MtJ38W5y9jaYxuK_@H9x6Hal-X|(_}{_%fo_C{5WCg!
z7en{5MIAn%P*aWW?ifvQ_)3d&-hvhDIro+T1lOq*S*=bh>1aFeso+SiZYeb1_Iz%V
zrARaZwSYm%#qOvhI1`axkrnW8a1s(&yww_K6txE`f5<--`a|9ihu>)ZT|uSK8+HWW
zA{gjxAy~_azfBul+ToGrKvjg=@-RG{c4`bN;3n8AoUo1_Ja~W{&V~gT4$gWWtbHw}
zzl)734<5*V4m`dOj_a+nKM=45rg}%5+!0ZXO6XFKKN(Os6;M`&n!OzY_7~HTWF|a9
zmCv#c_!9LK{>ykQmo8n>t27z?CJx?06C5fBFf?4I&tIO~;nQFO33j+)Bvg>I&=^IT
z-vl1FT0P5*i@TYB|H^~IEAW>YuzCiF)(ugvjAuFJ8X?&s{hIqR7)YKjeg?A<Wpp`f
zVPOFu?t^Lry>;Nj&)F6_$RxPG-Znra2`~()p<XJ%<RoNdJVwCxt>Zs*F#d5!t&6$*
zS>D;XY;-UI4ynR;&3`#QYMh$@?UjO>`XSulTQ5(p5<m#VF9hT0`m;L2QXK-sLCKNv
zk55Z8#J+~J*Oq?fR$B^i?%|pK-`(H8I}*8Xc!1I`?YDutaNz=?n?tY|52-JgTK5V%
zWvNHw5_f|6SfA&)1oBBbWBN~`4avVbK~UB!h5O*5kSznp=o=<hbMNmMo4xX*1W^kK
zjliRm!)fxO;$j;B`;OBMf?zFGyxc`57z`c|*%)H)Z?qGtfo=}^u_tyy_+3{Q=Y<QW
zkvZ_34`0}Z_fi4S0KG&jgeP?{UD#%0XWD%pc~u0Vz*Xoh{BYhJ{yGEjV{?1_mZ?6y
zrfoDo?{qGpxk~UVQXpXTHUxI#1$BKx9v2fMAC&^>vK8k8HgRZ#sJJuXQHfA>;5<GK
z%6Y8933HTUXrTbsA;h49!NIZw=iCqS3sEk*g~OKME+T?oP{_S>gi`KcC&9R<{T%RH
zBjEV&Zs*P#sOX67ey9RF?ILJ_H^y2#@7(DI-|r!$-g{TKb!R?(`V{p#?ZK4`Xmrca
zZ6Jgw6r8`m)bP4sq}-4k9L8>#lqyIsD{3Tc#v35<>zC}!f8_hPd$B?7u`)fv+su4U
z$~AWf0FfppuoFp)xAN*k6x!p*5C^l|17@49=nKE5K*!_&%FfEly0N`o?>7K=A0io#
zZ_4iDTUlA{9Uc7S)~$`22jT@G{elO7{}7!yqX>33oNSL`7QrPR%L3Cdc3}dZ4P+<#
z)UyqGdKGLUwf5$@CkM>|N<>st2t@g!f&1d2#K5wc*opiEPW&~t;E=olK(7o83{61G
zasqx`YuO1-?&F=fF!zhtmPK;xop$L+xK^Dd#8UUYawg`4Xj20JOl7)GH_({AsDYRW
zN<LTiex%vMpAZ}A0);OYZUyDw8(5g!>xK|93JM{>t?X^^yLT_sym;7=t6uP4YAiQM
z9Uxc13^cA-{0UkD7uSACS4YPKm<;lON}|=og@eyBPU*W&)#0GBr0;&jir>9^_sX4h
zW;1j1%6sR_(xY!H0MTHpaiO!4=1?U3-LEMb2Weid=J6M|<nZzFzd}qMv6YEmUxbwk
ziU<XMdzrWsB5|cz+&c}l6S0qu_U6fZ_rD!!H~DRH289H5evHp31#~`txc@uR2J+fh
zt*r};DeBz6$oB3+YRah5jYTDv*%-U!{~GoO9xu^ZjpaBRiM3WPh)O`QLZ#mnzEnou
zf?2z&sYwPHn!wO-TxT~l`LQ?}EYK>&=4A}!v&kR7T0g+D;M)a64p5#D3gF$}-{(Tf
zW#u2V`c&qFd{TJl=V!mY4UT7;@KpvTV&H`gWQrqY6^<z*U<`s-ogT7~`TqIpH`zt=
zm$o}o&=2P8$1U(sB+OSR7-ODI9E3f5_!SJhM1hYWBE{$A#DY3G58Aq-z5RJa8irPj
zhl+=zA;=eOKp*vhP4G7COw{YR&&d~b^9;&2p?(?aTsZ{E-Fmn{*~820E&QTCho7Cj
z%|-u5U7c8kNqwKo*a#d5*Z^KR4{ZQ3uWeIo^YT>s030Kkd21UR(r$OK?;z`di$X$R
z<++dfIDtQEfZ188J2?7rJ07BWHKEWT83C0b^6_W?pEAf+gKa=B?o17MH>#~ozNo0k
z^4CHy_mq)NrGX7JkRQ{Hm!98x6YG$BpEmIL@8zHn@?W4)w1Gq<Xj0b(&I*WJaB_ql
zN!0@k-GFim+rRRqA)GhSZgevjFsf?vz=FZh1#rU?xHf3RX93dFi#bvtm;t<AL`j0j
zfEb)gK!7w1sKa(`z2h)0G^{Qt-AHK%$Vvi-1z^FpLp7wJp|NYYWB}jZf@E~;)9vi+
z<{*2c1j;1k*|Q<Tg{mi~L%>5lG4k%+L!t=wzBk03N7|r`<mr?OK?DssaRefXBPg5~
zEnYo>&hLGz;<~0tbSJoUNK(X<sml&0gy{FA(V}!M90b&f1{HG-azgK*P*-~|X(HA-
zG@GlyJ2t*WUHS&4_mXygb&g7*f{Tj)Y>|k)JwHSs2I!YA>Odxi!9Kcwurmkh8$c>~
z6_rynv$MlslLNk+gWksLGc2q7{Q2{A8Gnh-_f`cfZoOfJ4#Ww|xWR2U62JgpWCjcm
z8&CuwLO4h^<i})fu?D}^%_kD101P0MC=AgN6t%#ID}c30=!M!=Fi2PU8F<VDnku4y
z!l?<8dJ|HFsR6WtIaUPmQJ@tA;B;@fI2Q?Sfm+H5WfRII7gE*%FgLs#RHBKpJValH
zEX7qYLn;aSAn<WxFm0DDn>}?mxPiC;1bLLDL9NyG;!g|@%H3c=Kfd$w*)NUaPaI-m
z)Sy4TgLUpU;rW^u5MoN}v#`KRe9s}gKHs-5hzfGH&=~%<JpR5(m1c0ee0O6%G8YG>
zU*;q<|99#7OjpLk*AOK+Mxn^dB=ZNu2k5WNZ<^TUz8+`tb*7d%S1(ramqd`{xRc$B
z9nOz{#-4A}9g76)prAa4Hemcp0uF(J<)RG6l;DRCoMjYTwbf0OB{DtL<_|$P=6x=o
z0f!*TzAr9@A+Zba!>;=Hh*?`(e_dUT;>is-3ta@<Y*x)|sW7I?ZV0D?Z(&Fz`16SG
zZ15_?O_*1)L=fuvgV|n7*+tA|B|G`TVRk+~V&vO%z;vmGe#VT(kU)7v7CUq+<lA?u
zZMsd&sg#P85=)nOzp$lW=LRI1Xg+xCbd{TE=<jb)MERlV1LYs6cDS`|!FAPN3UJEL
zp8{J&$;&>gQ1HqDV}zU}wNs&Bbs*{UpN-JzrA$_H#%#f1%u|ID;;TZ)VfT`LqL1c$
z@1LKeN%uK9-p5La<}1u(QEXY1);h_?+NsnaA?6k#GG5opO6_X~S`zMtrXO_S7zunF
zs}^1jUR<c{ZNlLOfoRCGGc@FTP$PnYo_&!hd;`El=ziuEZg<6}5121Xvv!HNB_YA;
z1m&6SnIrYfYjscERG&&@zH$|=wb31C4o7~O!U8sF_Hx6$hLJUT_u+z?D@-RrZ_X$q
zKwBXK87%OG)L26v)w$Y2p%cXl35JWF<kczFCBBS0)z`Qv&he69)U=T>B^;gc!A?+G
zSXh8r5&GiArLI_(D6mth|GWeSsX&2`fM43d?yjz@g~Q1A6+vB3%ZiGM>VN_@47krx
zj1h#Vd?2!jSWr+{2*b{?bgs`0*!_>suhFn)K^2OHZ#P0<kwjx!T3W>7`?Nxc&Ow1!
zt#=;%`txVpX{z%`lcc@n4agEPyx`{{Po6M=X%Gf~ThZ_{;h)}h2>-~z;c`=4s%rye
zs7eVo=l=|85xv~u1|AUq4ZtY#P;ZEciLV+N(ZkIVu>wiG!u<p{34Z{?7V=t9gDC8O
z%0<E44D?r|5JJHKkI%9#2nXrLGcz+3OiWR|Nj&pl>NW0XxP5;07&#s1Fqi`;>2RR;
zBd-r|)_U_>Cz2O(bz?%MUb+<vw%IrES&2wUlmK)<9}dUCJH;*~_43^sq5vT$D}l{>
ztWAk;t&E<z|7R;0h;-}Tw=-bcPh!PE3~GlnY%u_A^5N^L=&yLNyM1Ymfq%7vz{27Y
z+qD`|M3lM_Tr)i4jiRVGxhAM}_xN>~zfA%s5@aPJB~^xE4>^Z8AV}?#lan4tP)bz%
z{H_>O7~#P8)4*$(jJm0-lc>b8lA<P>JkP;dL8~iwzP&pQk;S)@lME1uftC}$P>)#S
zKp{bFH~zLX>|AeD=SYc=YFj(Id;ow5LWl1jNk~l%hqwVL&_Cp)Yc&)_#N5_lcZ`8_
zEz;O8UcC4yDoPO?bNF_(i)Qjj;Ky}JY-??r4hpeBJx?Sj+$J2vcS}%>kn*3E#d=Kj
z3(SP}*;(^nnXg_+BPI>_XG)OWex~u{6bi8}kVCFOXgHx{f_}z7T&vyL4mw^2<c@wF
zgiXQsQLG?m9=Le4sf96EBM53+!}USqjmMWl4ON<6#(sIj22I4M!G!_*vMAR(aF*3?
zE{u+y2p7H-2Yij7kbFV-FeHX+<q1(MW0g@1qW0|U?0HIqN!i&kFx$u#k=RedIaQCp
ze|#bRz=Xtdc7A?|OHQT+6QB?ZDew^F=-`d4YLjYff;tny^O{+XL_R1aSPC)*D<l_?
z)J6I9^h;lU(DzT&Jr$MfAY?dB)m;FdaY3@drke~F2?I2CVR(Gx=&YiB?^Ihd-+8A2
z&HbY`JwDhTPwZ_<pSH(pmfHo3BxE56>JDL7l-znF3l5F>58s>8oA?l-!Plmyq@mHb
zeFD|_9XuWMc?ys$KxWJ|PfATag(?BpZke_19r!p%DHbq=qb)mF1=!dBtGO=^%W>b=
ze^4Z)PzhP3L?wz8i3Sx44HOA!mP(U~21+z&q6kSEOieT?kxCjRC7P2aQItdyiQoNc
zt#kH1zrEKvd+&3ubDj0iTGyht_kEw|`3(2;eR>;qss9@3sOukV`!y!W(2SHZ6o+0;
zp9CUS|0O$b8?-Lcmo6=wu4Z|-Pub#rLM&&Iflo7R6TakI_*CO%jL!Q~@~WAp(!wmS
zUq9E&ESM)}`w(Jo7Tm$uhIdysGs@U}CoqKa;U2K<%d17(AlwBD2nE5F7)h~B<XdWZ
zvcbrM6TkD0gc&P?c9P@%MS_Wi{nqj=UNZngjxc1vUkBQ-QQ<!O0Y4{(%k?+4qDc}6
zfVUl!Y!S+Y<D1GkL8&hUA#6mr+V?Khf~<G<5<0`oT^W3ihmx6Q5K$rz;YII46+O4q
zJhQZ`((z)#OIGc6+;{z3Q0%XVkae1+H)O;53T6h|4quSPT@Z<AYmZP}K{1b0JyyFh
z%K4!HLgmr^_N_nBBbn8TMcA(Zfo=XF03rJktdBJXUS1Erzf1NbtMDUeSjgQ#x~7hf
z4%(w&`sT=zMB~N9JYT1(_+7c|$&+cguvte!U_EAp<mK-*4wGjbKm+w?Qnq7Zpt0r$
zrl6t+ZlB3b{a8BUfYQ)ciLHVOrYkrRITIGMwncrD|1FnhPmOJ%tQ8BB`yTl77j*S!
zeAsa6=<*}tN!Ce?L(M=Zm^8c9*jA>Z#pJksUPQGFRJg`Cmce|i-))a^=T3Px$#^iV
z5~#LGL!*%TziYoC<Bj1Pj}2|sSYUR25r6;z5~qwfPSU^g`S|#pfzY2ObpQaXk9-uG
zIB8N?RFqB5ZGhGM>)cn;Obh&$H>v|}gX0}8zBN-+A|oS9hR2{{VAzlG!6Lux)sZgF
z2V=u6P)9yQOyCJreKWw#G@e6pp7LR%#&>1jvu<fuS#^eD>?Pz?MEU(vZJXI@&G|R4
zd`yo21c9j>94rhyBq)Rj(;JjV`!Z>f!@5C;CAPNT!2g0LDZ|KK1Nj<gvRy6%Ks$@d
z%BFyQkr5beKfy(xCSu7>S`J{+PzT-%0k;ES<HzsuT?Y5{+3v}4h-1-~<Qh9;3=E6L
zW@6XT9xx=8VzVOUEAtRIchATCut)H&b&slUPt=@uF+uOm`@RHdI2Q{36^vnDzW!nr
z$3*X)MsBRr_#~C_nq(ZOl(0rQ*}0sgZpP!(ReWj|@&9=99T@q*bkYuS=bbzx-W=FU
zG&J2%(h0}7L+$>w;`j2tFW<#K;Aty)m5esduv;GEU4*<zMkmkRCAno1J9`Rf@)lZ`
zu(}wwo(>;6F0tsv0>0bvwt~-nlMW^{=|-_!zp+H@!e?WDWd3X~Y2w_+vE!9qt3;KZ
zTtPObR$~4`1(iZ!F6VLUhwK;Y<1~<`Sg#BpK+i5;URha;VbmurO`6b_@WiH)!J(mI
zd>Gw=V+0g0Shj3BI;8saITKi?eX2{`zR{@<pAo=@%;=Q1ZMu#C&^msW5OwVB({b}}
z(2L;I1JAFWYWfBfek^35bWZ}LvM|WK37UoGUvl?jtG<xLrMG;l&*P757#PL+S2`A^
zDDRrvckXL2l0eu+*e2q65&Q6@@9y21&l>89ApypqZj}2k;2In*OG_o^Z{D(H9Pb(!
zu&>J-noy0(QCW5K+O>A*7~l&?Z(8qto`1^e@0F3uqqX)Qk?WGw`FCr4<L#e0*Odb#
zre=2Tg04WP`JvkNZG|KHYcj{pkrA2&$Hut1xxFz9?iP4evsqJ<W`7oyH0!OGa5Hv5
zYP{aveGT<y*s4^W@^ienhl*q&xb6JMAPrl`3?m5TCh)MFx-nE1cqzov8iWDkjm7>_
zG<`BVyJZ;A$c~=;-QiuW_zx_TN8Tmgl(Y;=aFUL3uQD$CE!eB$js^cT#yPxw!;-|N
zn)9Cp=lu!HaV-D!rT-7+aEw#j{x>+s<4#vpBGUPh9VWBE2)spV-tb&+A0g~6JRSvS
z&T#L^bJq`7za(u!0$mc@{Q&$(_aPM)j1COrbm-6^=pEtJtNF3o8{{~cq<^i6T4BLn
zYc2b8wC$H>{o8K<@6)YngjdVUn?OBhS<TDMZS1JugR6)+9K)pH^H=DBoHmm~1eJcn
zaOHQ1QV=k|<EB4>)pAbgehZ5^mIw33vOihQ4)<jDnctP7D*vliPpo%(H?k@&O{R>T
z@CF(+!4egLcA_x^_r5<KmSEUsFo3FqPD2NhcKc)wwygM_Yn`<$$@KG&j&`E*5Xi|H
zWv>V%c=LJ5<1sOUhIg!vLl`8ocC7$>VL%Z}(AcPfVTj-Dk_uQk9*>7Xn(^_DMu@G0
zF>Q3BYmtN$v-)zYXF`8ipc?NSm<2ggdw+ijP=Vx|_a-TW@}$~^1XVV?Hm2MX5QNX*
zY{JAj1uf2U)U+ExY2xCKJ2^5j;1~lzrV$Yw69p;|4uvd*5r8Y?9rT31CcI&_f&wYo
zu=^ANCaeM2A1ok06F6-dhB<;gu*}!L+aJ%a2^qJ2sj61?<VEq0t=joJptnOQ4XgqB
z^#htZQo11h@f5B?ClD3lo{T&P)gPd>ib|}ke_^$f$~iqZ@C2gskm!s=Hjw0{1JDD~
zIvJe}#JqQIuEH9ZcPDWX9t@Vw^Ig`wzoM>=7YtSDZu>lo#=U#5Z!>|0kzNka>mO_=
zpbG|z7vl}Ym<)I~?%za5C<N|AJO;$#-aaLB_QT+VRpeYw9@ec`{5fHglrp&1bA%h!
zv;+bSis=X%P)98v#Py+BoSG`#RA|~AP-+zda?{kWR_8q-l)#|^+;CLdNWH1vBqe@@
zLU*3!M(!sb9}<1&f-)EDY%f0Vb@ppZ+I`1Lw=?>7qUVcVXS#5>MF#Pu&C9+krU_^g
zyVU$=xu1f7aAa>+dN3tA+jz#)#kpJjce}ioxo95U{kfoQELebf*r>6$CMWUmsg@DP
zM(Lc_yeB!bUd0@Lx|rq1r<SeE!-C1{e2=R<4Rm))G#H&@$gln@Xz*i)DsiY_3(0#0
z%!~L?@QtnS<nVl^qrQU4EdpGCaSVN~Os0D63Gj(L3ZwoB3jrn3?T#Ieh(LaMr-hha
z_YXD~7prm1U&DhZ@a}4bpoDnoW!o$*%M@Ri_bF#c$32$bA^$5~+?9XH;>A;7cYr-|
zWvxA#Rrh%gvZ71);^)EL+-Y<rp&3?Z?Aer!LNqxAKN2Vp;Dbm!JZXV@Rej41F+Pn9
zmI*Gmsh0%lFS2OS<ox{nBM@9;(mM_+q65G+2$0EuuQ1k&MpkXr>wphLLEi5&vNKK@
zBG_uaNyj5!ycspjiVoo9$c+I7NC`rFX@SE{VNj<9j$}EoFnYJJz7|8RI#|S|)Q&w&
z9mL>A)TIknlR3g<SI*^(kB>hCMjv1?IUq<PO1K*q?Ks>_1`ykV;~IRvSt`Qi@Toq{
z6=M%Bds#_Dn7_Ms%ib|8GWyflq3!0Di~rCFve#=a|9R^$dWHccQJq~~j@U*Kf)ZDr
zO)kKoGUx<otYDLR@Nptt6fVUP<@_l^-~CbaNas*et`-sI%)gn8+&SivAr=pw+<OGA
zmbv9yKoxkhC>cm{S@OU8X*atFj=f9$o`}dd^C`r`$@`_N!GsNG&t!#93WwfsFtONK
zDBw6|okzCb*KZ&d7zPaDQhfq(R>M5xNuOvfYsKkbJ8OH*cO^X2<&h6|`SMPm@jiX}
zG_432eXrlI26fFaKy7yvY*1QPRf`4uSxV&Cjc!;>A`tZf-H_0&Ch?<|8*6)^74BZQ
zN#ERZqY~%cv^m?p`#y#E=3suA*RRoiL>NOn042itr~TYBbZZC_k+;->U@~2KR|omv
zmgL^vl{p#KCHNYAmmbkGD5W?L1P!!Kg^@@E`GY%-X9%Al%POYHptqU-#wIzoV6{^P
zYJ&ch?$oV`u-L}oP)H`zuc=CscVJg`khGnujbY~Y{1fxlr<(yFW@8A)AnhBr$UX6{
zQv33Q6_0*zM}>p0uowkpbAv7<Q0Ijb1iiLb$*=&ft4xgFZ`RZ-Q>J5{b~R?^vA+_}
zxkN$e+|!!73TZUUEh=~_qgJI3&4<Nx2?Q!0XDzbK&Q7$*^09VBToaI&zaPc3yAHqo
z>&KUesylFJlj0F{<D&hCNuKFbQ+iu?9F<GZ@!MOieoJf3${kfQPJtY*xS<KXWgcSU
z=a5~)05<nU)s4rq6<&At>eZ_=(1No~oxm*ui9b!AOu!}&e__Yf_^G4w&FSlhy+BXm
z0i;5pL*Nx0@Ps)C$U9wo)&McVba*gPRPP*7gyRk62K#_P!Lhv8NT(7Yt{8qIOc4*U
zpb=RB2x$gVdm;3a7u0nnniu#Psjm(nJ`ADBBrp@p92I<oct&|=&NLm&f;@gb0(7iv
z6OMz=B5NuLqP+(MLnPP=N3JG3kZO$A$*67E#2)T)Ii#Bn&8HC9g~P~46H3S>7dM=!
zXkEI=JC>&G%G}^IhVn*=9b}b19{l>sgA2&ial>1F5G5HM`(_dJbxEIUxWHRd)Zo23
z4!l|i6Of3Q7#juuz!DIOC&dJd%|%Q$C>62Wb_teIUJ?oxV!SXrZai`%XB>UwRJ}x?
z#1*$UbI_1Qz>$FtZiJ?YC;(L4mZ2@H>|YfFAW8#8iUnu_1W)7(BtH;@H758zL@#k#
zS^wt}A-{cdQ*c2V*H`xgqM_U)WT`jZ!^bp(wdWaH)AZ>ZKx|RLGsd32+GAJ}1={ol
zK&1td?p5Tc0{V@eFWI*uGBdre3XR#((QyPyDQYBS=Y=kRGB`Vb>{2}U+hY^QNn)Tq
zNNO_Zae;WR{{NPyi09!1<^NJtG)+lKDY<z`-+(fvdWwo>KotbPe)$ini`LE4mzL04
zcXt)C`Jla*k{B7;Bd^^jc8Gs-_0y-EFkqFzFGuopWH^Ht%1`6(!E>~pM?KRk4K)4c
zfOvzx*%46l{!ajSVvQTj|9e8Di?8<m8Gy%`eDXpmmOc6~&L3XR!5n$vGH`%DRV4*J
zIl{mBwe<PD<ttuaL*gdFCO0YmZwiwH#!7QaUpGt6X5?%e8HVKPn``~SCKut>hErY{
zH7apbr9!l68GN&@3#wlklpeIYx0faC@{r&~tW2w(^jv(-rP1G3Q1EO^IiApuX!t5%
zL%|#IWy&!WO~|myF1~Bmnvwo>8-!}=z@2xb8iin$;q%B7Vk1P^$fI8t(xc}|!Mr-s
zcdoUmVr%|eaIQv%KNrXynp``?J{Zt1MN3(~FalpqHd3%;;Ol&z9*t%9hL|H7Z9H-q
zq{{XumD3+<@Tk95D&O{U=K;tTrdVR?4$8~MstPqFCWD9APLSn>aR|lR7xfKALihD2
zmj48UJMS6M@Z*Q)gI_;Kd)f+=$e2(>aTg}oQ^{0GWf;0Oq}6d^>EK^wC*bX~%MYwZ
zhI{!&!5&a1bA|(DM9ci{5mbz*eXIspc9zN?rhJ1fvL5!$n=2O40QY!}0#fjFi!oGI
z_ladI-^+MzY0=fwa{+NvqQjLt-?!^M{p4@w_~GTfNe}OLK%uo5uik;d4Vt9vCaupX
zL9EP!FZH`Ic`wbl1?b!qjh^}<R=YHSY$<zl`52PZ;|-H68@%m~bsMoX0Nb&(n#5pZ
z^95&3k0%&UKj4N)0tf*A4K#i3?c{CO*})@(Xe*8J0JA8u%n59foPB10(aU(hw4tM7
zNrFy8$In&6uGpe<Dp8EAusbqEnwF40da}&_o+8Ng-keYJZl4w&8|_~O35a2y_hRI-
zPV9ENvOq`TUPF_vAtc1pBRsx$%s5);pbI>n^U2iOlwB|SevqCL1sxVcL2CJ2N<VaL
zqtL*y?3H3+rb+eM_3I|sH_`bY0g6sBN`CWivK~Lkvc_dS%_L|bH~+Bv7|++Cp&`=&
z{g;;alzZhWS%2H_;>RU8efI1N>qm9;lr&qWBniQz@{+Cgw_uxx2RX%_-n_D8F+?xt
zA0;I3D3sYJZr{DjrS6gc_CQe5^90s7vmqCA)$X#A+W#&JW>bYt;x8!J?yQz=y6MMM
zw%F8vsj)xrfi@5BXIDx~x=YcXSmWr}NA>kJhYlb1_xGOy6Aq1oYmsP@0R`A}aFsCv
zE$aNK{w$c=wAL-&HPwf?X|2!uRsYh`)k9gb*l<*_*n~B_o9)*JZ%|m!$D`QGkD!r;
zMMU^Q#}MR^!n?6R7Ao#6iI&dZ-UF`D8?J`cX7}t`bN=GR04OmZwY9Y+L5_C?6HjEc
zHn3wf^qL$;vz*-A*)3NRfn-j=R7sduiT>Fyfz5YTNl9_x2{@6QoU9L?{({g}g>!p1
zN=Wa@kX9ep`AbBD+FBqa#6Z7uclCp-7cd1OuawQlxbD(QNDv=^g*XX+2OxtfUOxu?
zUwb+xzQUqCQH;$%k~1>W8o3*=fl0~NLk<Xaz$I1{l*6;0F)%vNP=sltgm3ZtFGzcN
z*cIX!65*V=XZ!aBDFkEADd(&6NA+8P-4Ta0_~1=$9gAgOzillv<Q`|tqUXm2@Sc0(
z5nN&HU$GA4Jzh`+GBa%IBZMEv#7NzHgL#am=-&{lk1}hRwKD>B7C9LR@brO9hImIL
zD=fc!Iw0U7+3-=gBJ+j`(LtID!yxbEBw|R1k70QQSCX_aXeE|r66HYwLds-Pg(0|d
zvW#UV!=MUuw%8Y_J2*%a--^xBbl~bLOz#M^`2oXHL`wDNza!a+E!U6B%*+HnG@UuW
z-(-$Qcf`(iK*#jD82|jb&@vuD`!RtbrC=Q{)LDw6s-9>uddAYk)<$(@Q}<<1MJqhM
z_ui!ZP(0L~d0$_t($N-te15=R;{ah14XdfizDWDZp7;>azs0Wx!R}5H<JUCIFrNwn
znIJ_vB?Ch$IAkB}mRSBLWB!rcm;W-!;|pS2@r9S59B`D;6ihGx^-1t1gnMHcjRRQ)
z9J~y57H?xRpms2TyVpIP&k#6<QKt(-Hu2LK)-my)#4g*885aW66)m%u;MPRck$RQ%
zC<IItzq%Hgvk={=haIKSvOZ<kBwTaIi5I~3hyJ?}sF(w?Wwyj9Mvt0?J(lK!&p`#T
z$f8ju2Y7iY&X4CRbi5-3C}OEY9_}pCmY}ui+|-)?V5BpMp0>?;Rw<Cg5#vp_KENrg
zP^2$H^2P~*KcH@eM1}*s7vkFqSV5fM)shV9NLR=z_Puwi&w;}x0~j4hfziGUa~vg;
z5;cou2Sc)7>*FaL{NB(TFv_@7+=u%PY;S!5BAn1cd@CX_mns)_1`~x0VY=qa--PBr
z${-okARx+rLoR^(k~Hwaw>MGhAFgRd_&am*lk<%<AnTo?6OcTkVqW!Z0m8zjGM2LQ
zZaA-r?zsC79N@VlH*F-^oY%~U1Hut{S5M@d`t>gj1V*vF*T!(<YrJW$s|lB#(J2VL
zhkAocwJ#0mZ@iv4@cz*1OFC&uKV($$kJvriE^61+BEP+T%%+f|Vsyd5IQKWAYWC|m
z8LR%Cvx1Dc&fO}w@lR`O#Au6G@r&G!xc%wYtXt!ejUILa&9s9=)>y{4u?Swg$WN*<
z=vV<*O(7{I3Ld$4NF1Z5ch3B{Vtk+<Df_;<&Oj8{xzASUXS#DR;*fY5t2hyAeNYf%
z`ccIo52UCDy#pPemYp39EHlLUX7k=rKtjMFk<-8SEtas3rj$L4^1p~_DgsC+>pW;S
z&p%!7dCb}H6$&YIYQqZ1TSh4|^hLjLE&C;$&OjCz9vQrw&3rck`h_b-i|xjxFWK0>
zqXmwi8kFW!(=~jwy~@jWTrRT(Di^QkM{n8M7MqK0(OB|_->L+*xw95;&cU`Li3lmM
zCqNUnBgl80f4nQd^ZD-{9KCED8}aSuGd&57PnB=fS*KNlTZsYQPk$ZYg;E5^i(*mM
ziZ#CnIc^0<J8t80O35?>`Aj~p@xVl3IJQ5SN6PCLyUyqydy6yIc)Iaj1I$pv*9Pi~
z6-G2%2UFS)AtZ@Q1@Iw&Bo<_+92oS3kPSEjK-sioWA2ad!RJN=0?EDJaPuH<j7k7x
zk0J<*1NZBs7~e;rF@C{+wC)_-zMJHOghz~g`&L*|!it?*1*q?G`&Vn}_^WGNdV}QI
zIa|xuGQu~_`b(F+bG)RgS?D9Y;V!J?-T>oC-%dClq2<%2odB?m-NYZGwJEsqXAuq`
zkD&kZ1JuM*LB1Vi?$tQh!RW#^bt*~nkoqNU{E(q8htMbh&IAw-Lg~`-V|*V6x>AF)
z$5tYC2?`1l#ejTG2Son^CLy`IoD5mef$n2LHUU~AjEaTtun12#?&rPn**|}L)+X2!
z9Tok3+>U~mBT*&ZLl1(xOCD-h!81@y_rQL+^N^j1vMRh}JIB+3Z=2GBKYjWn?e){M
z@{;Vds;Vl#&q}|3$N&@~AqAQb%n`Lou4bmD<RW4o;!z<LN6|fqOoWi=fuA571W1I1
zsd+$7X#>*MuAZrM8gB-Eh>v*Ggg{tQ3J|`i4h+cXoy6u`ls~lSO;vOdl3Pr1R+Ubo
z=5VgYKe$DYq9cajbqL%i(xCR&kMkGcX8_Pkx$EFDh}ZA0pOSQurNgn*I}*0h{n7Ie
zxR);+7H_!N6!(9{lk;)r{uI&^!NmkM_*@?jhPs}|K%PJLq7;MlNAHG@eeYe&Ajs(C
z0vuQeTDJ1C{48Y^9qVdE9aa>K+*#&L%i=$pNFKyyBi$J1Gyt=G5g|vIL`W#_YxK2v
zmNtcB12Dojqn>>aFiY?pnaL6n>?BcI<DAeW3%3;G4Td@5!-Op`HUEZvrvy<QX$Von
z65iA|SB0D0UnftpAXs`Hehgu9?x#LnO+7V7&h}!H{8&142ZpUzf`T@uHvqzF$pTeN
z>8~uiZd{q2pCE*UV@-`;$K}E0Wd?_ekDaN-f;{c%Npf;Y<qgdQZI^A|X1CmC66JBQ
zFl2@Jhxc6uV^tTH2*w*2dStYHW7+~}eNG`Jg@P`U<VJ^jCov$*CfsQ*dvQW$#R|GD
z8HRGM8|c~KJ=d>&pNSn$H7M#AF1N!jSyi;Q<Y>gAcS_?JHg?()qZUk}km{M6o3~>z
zv3&c_X-*aj4%b&ySC`@0KaJ1cg%EOCfdc2=nUkjr1Q6Ot!AD$NWmt8H7lvA80t3iT
z2T)?NP1ul%A4*@+j$77)b72)?l%1S1^ba1xDFIe>NxVNl$NVCJG-x;~4bQG=%%Y+N
zVU99H(f!R+l47Z4NJ6^+>lW{wy)1}<exs|Us2G`fHzilsRA%<DZ}RS%U6B+CFvZ)Y
zNCTh+En<LQ97)A*Q@_RO*uX2FuODvReTA{hD(s-_@6H+0Hid<U``~?>!WG5f`PW8w
z6@ii+%UtF=dCWTw2pLeF$2b7?XF*<|tSm1!-Ib)u_}w)F>2HI>-UtbOsxqFKMd`#$
zxuP$a-HP6`U*%E`ytKkD1v@kmX;hrNejURHEOD?y(gpa96Ro4lytB|LZ-y6(Uylh9
z?NQUnbCN&pf@i%#6o&#ZX|_!(DB7pQ=ohdgI&P9`F*FPm<N=0w0jyexIoW_B1-#b0
zE)Pm3E_~7O={<GmBOSlBwZx9wdrsj_Sb%8^1116n3V}#3OdWW6PX_i4ShRd+u1)u>
zy0R_^1E~-oSklVVlmW=b1h^{Y2ypZBTYd3RF%PCGL;Gpy(A$L@qp0OntUFqhoNt^X
z23Zn`9++6Lg_dvkxpnhG(azxi@-H`Kp(}3MQfsx6{4fLoQ6XNsl#?P_aDx!c0FL}c
z(pw08R{R(n6aA&!*PJyC)Lpn7HjBZUje?PRMoJ~?)~(BH^Z<Ipsa#l@`Uvaj1d5O{
zvh~2o#4l&d|MTalJ#0ayz+yfQ4XHt}Esb?Va!(F7d<-SHOQ^i&(9i`3{_-m!qJcV3
z(j_uaIPh;a*|R5F`jAN*$(ZitH2>Uq>tB(X9H{v-nMnh~nyl3DR}1#pAF3<{%X+qM
z-75RCVD<PB2W0HtwQCojjU{^ZWCok7V|qV)DxDD1tAWHK`WnYu_;ZzE?K*hzirs0{
zQ%ixq*l{)rCiBv+9X?uDm<rvCO}hLl-0kn)q~u7w;QAc?H_{ST8UH}Tf7mg@k@dIA
zlEPAKn18J-S&mB;bWmIH9K^H+2+X)$oz!%dOV3dQ`AdXnycS!u19@C)5NzD7(2nfv
zZm1~!NjXVD&TJ_o79H+1gSjhY>ago07s<#e1-Arot<!v<aasIL5t=@XOhqsN2QAgA
ze-D@rI*pohKMiy<Yl7>LgQ{HUVVii-px~l0Y@dB}Pav?&%ST{)bx`BH_c5y7eE(g&
z2*=fc0JaC8s}ypad&8S-QXzICEg~b0fJdv9mrMs{LwJn)Uo&G9&V(U&9WYQRNok=`
z8+j9DFAYAPQO4Rs`F1Q*rqG-S$FJnDqC4!@pkqpj7JU4DEj-DKvF5M<{@W6{qNML0
zq<@t=d_O$DKFzw}b_j$^v^3#Nm9UPNZC<}fPdq_--`Z#A6?H@rC7OE{k!PQ6C9fh)
zdQH(Zu{L!azO`w_m1Exzke(HsBbjY=QjOL@G(|#WQvQ<G4%L&?a*xABcic_hk(_?u
z#Uc{*0y!yzV9}=0^E1ePgZ~T5^;wDEmut(+BnS=$0cDcn5_gCyo+)WaPw0RdIFuhE
zb2<=rkQtWk>mMfRE8fUVulVc`u<h}R;Opo4#G>}QRt!{!uvT2-2h{KgsXeEl*dt7p
zDm)$I3;jlW$?Sc_;ERpo*PE>aLJe^_1=o<P8Ndwa?dAKc1W0)gV5kUABsxbFR39)a
zQpRIkFPdlnK)1)o?FQZ}g}xGi>(^SR|JUZ0LcD)^3p7#BnKP9k{t7+wV3G&alyiY*
zkX~lt!i9{O!DUESBl)F`l}|O#g_j5`y&sH8?GR2rf+z=`+GlX#!L2<Jzuh+~Drzs9
z(%FTL6PN(THubZ}p$th0vAro}eQgV)rp3%x=AN@&K@FJ4goi1XMd!3`%KX^TijjHd
zfb-VwC>&CZ*_<$96s(tNdFr}OPEPIb-o5jF6|wJ!hi-cJy8EXy2T#jf>32WM77<tw
z^hCV-(_YURN7)#ir>(3n-hL5xoVW37vCyv<u(|}#z1_n%#=4<;tjgiolXx9H$E<^c
zjh*8PkDoy*uaz+jLb^tY+RH5~`ANWoYms^6psA@a(W(^Ob2(aF&vBDdVDb9lERmL}
z67uZlLSWZuS~gLv4X5ZupAx79?FzA)2pX3!us)dDA7bq#PY;D31NFST#)TbfcyG{-
z7;QMF4b9M3^VxMHv@rA~tTDE?n!;*#Xq<1Ipp!q1P7sKgfUBO6AhzYo+Bs;UCE%pV
zt%_-{2r^rUKX9%A3(OWaE?c?7woP61MgH=!WlE3zx{*sXf`gON=zKT-0Lj5GU)0Ub
zuf_34CTyEo8*V?RuUs3)uD!RFJR72{e?!Ik8|NQ(mI_<`pu*6;*Nu0!-}hNjX=xVm
zHu3H6yERR-dI&W%bUvPuEx=_)G4}Bp78&IG`=-&*lzxk|H{RzBYeo0!bUOPapkha%
zLxAI(<MiQ{h&?&iG=oJYgHzpY*&hiQRnAY?!Y`QHOU|)&J?S-Zk-T4-!0~Z|iu?yp
zyM30s6&krO;vnJ04%6b@atR_3mOwsp?uw{Ic2Gzoap6EN4vd3}f(54c3&z=5st05g
zSW|ay{yKT-T{;LwH}dRsvxqM3t-j%A{!WMkebbMl?KUk!2gmoZ5X>+ee9_3bF2rS`
z5f^+A`_WcNV4&R32W=Si*38T-uEwhU_3M*#)>ajgI!!Qd(VmJO9SI!cSRcG5Fl2%R
zI`esk#a-CcDCrBbJTj-it(FF0F{=5#$Jb@A+}<KRZG0KWxLAseDF=vEg!di(`;L_|
zPC*UrnSEAI(nNUA3cM%|T0-ZLg>hgAnu3?<z=~XkvO#t|NWJjfn}I4v*4z?!EDbxk
zo&mZfoEWS$Gy>~LOE#XVA%Z0u=qL%5X;{9gEl4rn80yG0h1JRKvkambW&zMuM<pPK
z3PNQmHv~9nF(yZXJ6m4LF@Q4#0P_|DjF`dC9{}AVDUK#h=1_Kac0Sdt34b*y15qPr
z1%_;DJaiuL;QlLalaV-ZcjP`?LhUG`rLfdIL(CZI>%gg(0@9PVt`+;qr_I2VWRLJg
zIUL{fJHPUO&Z5F8kr^EqbRPIVl$j5KvjC~=LCa*?;XEC`8e60Mu$n)ruC7iRCYFzs
zrW9rfsgreM><~&kVU)`;`_rE!80FpXAbT)L&k25~3IMYqLmATqq|otKD2~7)cE<0&
z8n_W+nqx{UBQ!&grqoK{xX`CK*32y+J^-*dc?w}hfm7`y4t)2<cx(bqEL3c?-JZF4
z(ZDtrny6A_OfxFnX>Ok_HDV|6a2z}p9{>;fMOOGjp#LE64G2h4_bU+l&?TZ*1pXQ-
zAnya5)(n~wa`2~NDFN^z1mG9>>kqMF<8)}F&fvQLEOW!4ZU>NctgWmVI#P`C3^Qhp
zQf3RT{>%KM^C%GNfgP$cjq`v>%;0vRbX~v7@T}wF`nHs9RADQaysf@3fg>~hw-~AF
zsq%;j_;3_i6_hhW5^b#PxzFBH^d*wwr@~4hA(2!rhdLYpvPAj%wgnKv!CfPpy9kQ}
zgN5_f^L5^*5ISLoL^(3`07f_gcs*(_G}cgpQWOGC*dqNL_$H~EqeCw8*gRzvZYpKt
z;ZJT}1B%ewXKyoLJzLj&7(5tnFcPKj6iLBPB1!OT6tb86U4@q)?#>El{wLX&=Jy((
z*%2j$*r+n4RrDEp|2?MTx?p%STpDqaX8;bgQwlSNHuR*WmIs+w0QrnJ_$ta!k{dR%
z{v(<Lc))U(ceSRevsoqp;D2P3n&*dNVHlA<(Hp!IB*whV?}yw`04O9Y6rqKua(0Um
zl4KZ2h@N!1k;wlcMvH}IU)u0szUq8GREC<c+ZG;<E+pn|qj;u_ReX%@&9k>Nw};xm
zvMv@Z7Fb^Gk&u)6{Kt(Of9qBu>qXPIrMt5UGA08y28r?G&&D%1IuL6N#}mq%33e7r
zQ6_D+RhFoB=G(V#yKqT2;tn3f{WdprA0jtUfYd;V>mZN_TXz>cA9QSYZHAh<dK}F0
zJGj2}jHhfoK67^YAHacAdVg=`Zr{DhUa7diL7*}&eo^&GeGcz{oJvMNb7KCz$GWkS
z7i&2s#L0#1Tl3~^`<9n4X`jpLl^4{4bs%pTHw2S1PYYv-4=tW>_Q6B+dz{Gs6kA@j
zWij3|d7h6SKVJUysqCJeNc<-Y8{)pj0bB+?b?BWG3f};adx#=iMP;RA_5htX7!)*(
z-Vpegc-V~m4#kUU-Wb|%X&H}4XUCVT3!!h&I&xXoI88*U5D~MStQFQs<y~6K9!Wu4
z`V#ncj5ieE9as*R4u2;rI+i6h;JI*|1>nXj!qf;`hC!aYEcw`B&`bW1hv$D}WF!{l
z$<IvC^Wy{v@Q5y=rhs;{NmJ7Yv_Iky7PuY7a1{pQ8@7;h?Wej)e!A+_kx7g+q;9w=
zD8T`L{Snxzg)as+udxYcLu~vjWK_R0c$J-(zU8M~km0R6%J+VhuHkS^`0=9QM9<$1
z)WGqU(<s~r#f(la5X;_*feO0>1%s@1>|}vZHqHHuHEA92@KF5IE%#6$&P8o)`#zoi
zj@22)QmF&9FbN(V3TygHj)r1pcy#pM9@+0}1iBJC)xSZ$wH#9Km-k(!AtC8F97oul
zKBAg$wse`JT_SZ`P_%!N{vI4ZprkLJDJ=nws8({f^m!4|O@m%u07+CZmbiGJV3Ytr
zN7;bcJC|?^H`4+j&Cm+U;=4OeDM)ApAQSE!n9u8>t8g!VYxOsa){vRo749|vXeKHU
z&4l7%pi$cXItyR5Dt13ipm0^&3|^8y^hQWVW>$>nr>PLfQyB5)8<w-npxp<QGH&Px
z@d?ra2}`r(%~^PAGyemZAH`>#n6O$_w$9;KCnA#BlnZ0t|COU9C^K0S{jU&TCzVOu
zfx;-Ns4&I2m>gpbsVTNEDfOfS52_WWl|<|a7akIht2gnZ_8BucQk$LS&M%8$tr$L!
z({3<S5c-3I5R{=RoeQTO?D-Q>;)?KvSh(ya1s*(hX}HhG=YSL$w9hrbl#kPpfk7{;
z>+HE;AvOmm*X=Ja^!mGL^w`*7+VI2dCY*{<PlSoZxt8kFZ<3h*FQS2Grgs(6u=cPh
z7v^NiHt_+}%Y46bZpgLRoz>6Np}p3_VKa2hO6z99j~MGK`aIFc$e4z<jQb_8aX6Tj
zmDL2S>j)wzFb=bkL=v$}lNl%tRh<r!Q)caz`*xaMFSiM)MQ~6`-+fiL$NTIbR86O&
z#c`~lJsuD>tRVNDdwuYq#@LpSdoM<@^??6z@KE8qY|MIr&<&UZ38ZHRI#maR)6>cT
zkF40>_eBd+|5S-kZ40UPk>2YUC`yAgw+Jsfb6Lu6%P7V?hJ7O6W>w1@b642~Lt(c$
z6lInji3XM`0FOo&wxMCuyb1EhJiX+xJ6vc}bhsB@dM>!cTJBO7;zJ;;U>J(oO!+{j
z!;jE^cOcGbJ%AF;R<3PwnA&3o2QFuCUAJm}X@%66XBV%Et!R9VI+v1vpH4x7&ljBZ
zd=S7yBWqn=t!uj*G7R!{VEP^heING@avg$s8s0=ckS{_lYnJFaVVT29j#E$zVLe3B
zk04Yx`<$L)POZs3Y)?-dHq@*fofvTpJvhLUj@keN&{LOu!Q3(Te*$e`^{OxU7r++D
z+w4an{LU<xLK(lwsriA}N~jO^)Fz&N{R7z_9Y{t1y=~l-Vd2*WfbZhU12dU+$SJbF
zM<#T3zQVZ_ODEp70MQkVqs$hIBXVh@CCPdbeF2ZEWqNkp!sH@gvXlV@QqS@W51c=A
zP-k47HJusbwnkgd?H^P_ApY{aLF;YIA`mJ(yKBFXGT10m)a3J1bX8*jI<P;gV)~HA
z!3+e7L)i<<0m=LcSwPHu@`pO`b+rJ}U>uYd@IX?)a_1f$(x_uXI0?Ry^bOz(P?7~{
zLq`=LdxD^#{g;)88Hy6&lQIg%nzaD09TyizDJK{dDbX1@5EZujQsBlVpX6neVTAh7
z8biZ5WmG`A{0P2D#7E7!;_*Qq(`R#^EEqkcE>D6pO#0||XHsBdM^VGXOBr(<^W->i
zMY8Z+PsG6rhp*pCouP-MdFbo9zm?rU#;E*Gi>w>BWNqpld>fo%|93?bPttPlYhHhZ
z(mMl77D_lOVjSKAAe;XfUgDw+!Pg0}yY$t8(_&bIWMI#kh1E?=cQ+01*i5k<1Yo=^
z!a{@GphsXimLK~uvVOE7%VsA<1W@!I;WM}!W04U+hBhGN=eE@y(k(H$-1InSY|5Q)
z@)nIzfr%Pdmy|zQYQX*FQDW<+pI6Nnc#8$*wDmjIf4wk}9h<qq|7H{qi`J@{)eDAu
zx(cd&l4nmo?XlPI%7rWN7~mU?@bGOwT{B#Jj2v^U9}-?uzzW4QpnFNxyr)(D@uR~x
zaM6fR$?T#iD%Z&d8Z_@yCSbwOFH#A55;Qr%kB0@`a`aH|iD+=_+gMouFg>Wtl1Eyk
z_{zLia9l{od3-|=sX5P0Tc?LzaZEe(YgjPHYxK16?M1s+%0>@a6(wx%mn&`S=5w9H
zv=5y7AzZBxE{uTNJrNO09)yb|>Pb+-4)Q6<6i-<ZmlY0Ib!MS>(ka@MZRhC9r>h?C
znE7B}ZkPZLy%n$CvcIK=_xxa71-YyL2^49e`iL5@?wI9KAvWvR-R9?9c1omdJpIXW
zbjKoR!<K2MK0666kWzRewqzmSnvgNwV>LFX<->pfUc1>-_4rUpfklx$k2IV_d}41z
zKbVf$AA9&@m=*aITvO^~Rn#<ZxH|cs_WG6aVY0ps><&Xg*L%U@>4)p@>b4C^xCXPr
zj5%`-Bvb>c`6ixWxv@%r*S$weaE?aFjdJ6~;Grnc1t@lj#m~(dYrt_HjoX$8*wQ%L
zxEX4qG;B}c+>6e$LQrx|a@<f6VsfDer}@lzXU$Nq3zjK$?3*=ls#A&vfw56CR_o!m
zh&%cOE2F|-;kAD-=x1*63Tq3^9=I1&Ffa7~(8G(Ga|Na}O<;Gut@!(4#=&o%Mp03M
z!-g6wWzOT--E_NvqWHqYITSX|+l><jkm`_rbkr5UEEbRdYOkMbG5!0~&!WWWo)+Du
zecf2xPQCQ1TE_F<-=v;asI#3)xW+T80l@OlgNl)4kXU!HE@js&k~BaU0n>R9ZW<vG
zkx7W><0F6-X}67o8rLxT_GOx1P8Uw99P*ltO9oMj#B!B87OSKsmXEDb@V!-J)`tf_
z=g`~fppv1@IDziFqDVC&97rnmIjjRyZ|#_dYdT$gmi~ItZyZ@_=|u*-k$@tR1$XA`
zS+mA-;RzXsTwHu{#`&(pCjxi^kcNA0M>3rvN31B=gz~3L-G9x958QIKhVpcw8uBx|
zwd^M@*CHeXZ%1e(V66@Kwig7N#zo1@)HnALO55NcpczsaI`N~RC&E8l(%(fOIzwXo
z+*&ogxbf34<Jrg0rktUVGlKDl3G2-$BUg=g?%g1VGgxFrfVyKtU4p=2xSc9seIoUC
zW#yJAqw31g6EusVZf9h>p*|poQKMdNi9-I(Po1OUN`L&z<H1e*4V*NS^fYz9<;FtI
z+~sp4yF4VM@?$>z@vkMP2gGG4iub>IQ~T+xQeEHDN`&eB?OP_fkZ{s7!XN1`5fb`$
zKCPDgr$2KQ<`o}=DT0fVaS=yC=U-x^q$mm)|1T7Dbzc-1I6xi>;1>{ELq`NRFI|lz
zZzA`nLA8RUf~LW#6|?1k1<xIYWDgtlWEF0TSc`N|=5Q*w3E5^Lm|u(%(+~m+dGajE
zlflGi@xkre!ndPz4%1aA5NW1D@@xnEvzQHyag)0CxP5bmQrUB;W+O$nfx!k+3e)zz
zycMIiGZXh9;4Q8Z4%ivq26!GFkzs3!>0FI5MGGkgpt!4bhoRx6p0{roPvk~7mk8xw
zz8)6B>ukzHU%$=<k!)7BSX}9~yLKDG(q?#zvH3J2IRdRW%yiSHfS%4y;fdV2F7Fpv
zT39rFJ5X`uve$;KUocMdu&A&IDZ(bPV)(x2+0k{anfp&0U{}`VW=RLM(~0Gy80dKr
zbeFtQ7=v!!6jOTbsdQoKZp#smHgvS{UjfG|r<WE=%t3k=*pAc~%YE5mPpTJ9R_H$Y
zb5a&QfvKhCjgAq>med(j6rx*pHW=rH8ApuYrbb3rI!5put230nVA1FHO|-$|JLXSn
zl00Y6o)s;V+LI$Bhm&Pkgb=Ed>*SiZC;W4DLOO<~E%;wZ;fPq)81{MH*x0YThYM76
z_3Z^77Cf0aZNdaq7BgWjA@e{lU7oWOHc0kz`Gje*bOuX3KX*yPOk$SN>Px<}<hzW0
zXKuMY*HR)dp~R<CEivxJx$XdN7Pg7<?81+(v?aI3%u5#bFc~n)kBK~*{`TcM#jlYr
z)kR~6J_vfTVV=jOUMjI<$=2`qrx+$9Vg1GoQGLbThKA7c&X_)Z`udwMB8MU_T$mQJ
zK}1g?LY+1LR^tP=OtPRMN8yZ>we^ahVXOG-FGm^ll|O_g1`}K^BLpFHTdEf{I1{~L
zi0V@<v$nP6!Qq#WYHD;>u38l}<Z}2h)h1-^m!N_y4B8iZt+Mw7Si{{6Hw|1?0{7-6
zdCIL?H4Tdx(pf|*Dk?VUrHi;|&6YF}VgfLuZJM97^~{PGY*>6uF>ZVN>|cn~oWam_
zUmDs{&yn4jKK!!S#??GFBV#7eKV_U6h2^O^^%tZltPB$AN8-#-emUIpenWhjFU*%3
zG1wu<W1gR{D6r05?^3&kX2oU2a$MsHOh`xwUx>(I)re}qX#`dR|GBVGYqzk4*8BJG
zbqx*2BSnCGgqK-QyH<H~!__&@3VOS{zgVe+8rNSL|AwIcVi*i#GA;I21xE%u(?Ttt
z_$}d*m6c@}#Ki@bhULbUzlGSB(sRft0to)j+D-FBEVNwK%pY=J2Qf%ZOG^+urljDY
zt^wyI$cP1d+gFHMD8jBi6{4vhI0jWrN@`VaazcXgfdk?wDE`184f(&6oAlp%d~DN(
z>l^9ZnMn*>7zeZ4o*C`jxhlo1DWju872aw_XxXx=c*oaT?>uqQOuH*0z!H{1rnPwM
z8MzG=m!XvyU>IZ(lS^7@Z0M}UuPK6_-tqCmK;d@H%t+CQ)P~#_g$d63sSsh2&IiGY
zHP4=T(=vnEQ%p?E8~#XVg;)oCHAl=hP6k%Z${guSFRL8R&B<XHbTIkxw^UAWhSrs}
z^ueG-;&2z(E##tz@GjNbx;kW!9xm46-}a`~;q@CVP>;|nh<~|r_ih(SA8;T=b4v@Y
z4}&u&I4$~ecqKH$qN1Xdhi+jZI{cwEsG>p>d<cbgQyRpzsh%g(aNn5vTX9nk9j-Md
zNs>TQ#<A-rxh~{1k09-bSZpYhZTzpY;Ra%!;i}+fGRB!Fq#waK2kg_Pkux6KoSB-@
zJXr%ohR_U+Zb9N}LXHSyLkNcA_a0C*3k?+1+*u@D%^$)>#XPI4v&b&U%$&s_^IeX<
zhq%fjh}sFMfel0mN`LUeC%9@iY>KFU^~wn1xrrFI=%Cu`oX<-+N`86Z`oKpp{9YUR
z)85A#V#H3J`&QfQd;~~2`D95Q1kvVkV2A71uh-GjJ1M^XMmvNLV&dXqc2jt&nh(uN
zZ3;V%&%iKq=gzGhp4U7!Hs<dlxm!-pjpC3fEi5#2I-N@beC<4|waB%K{i>>yl9H3r
zYg+8L$)&`)1DoiYAsDT-Y9!V|t70I*BnmHhac?W^Qt6EM`Sa%sPg>8qR=MBAWC24|
zf~)I|G8T+9P*LgZ-5U}Z$WHl|;b|K=Y^t}?x8pGS2|4jk@opAlgdhKA%2l^-pM@B~
zXSKCtv4F;|@RRbvleOP$?d&!jI+WRz4`uAmnvFsn*D4b`cDx^Lb*X#xYI^%Guf0pc
z&V5_C<pIb$%TdmLk&b1n5)F>p%?+>f=m~q7%U=&(zd6+_Er|W+VaFR`Q2kz)=UpE#
zJnV|}H(@3Ks7zeLd(lvXSJ$d;%8R(^s%$>w4pT2gzoMv@n0>bu{6sVW8)FJR)V7n$
zHg9_)BPy!Sa3coUauh9sl_B}V=Avx*ZITjJA@@#|gl^$KFOPeDHuQiivfJz<KZDmh
zg^&h>HPqNk7KOjWg1ur_xCbUAkxz~@&cC^nnTbf#b*P;P7P-{Z-#<U}qs2yi+43T{
z8fUC;z6>C-=3)Wp3H}!C%c~a%2nd|iygpN3Utcv_%1KCH5y!kJOi;!sPO{sx`QQYN
z8k?G?Dkvyq`d)v&E~PQ-J)Ve-6vVU?ZF0+R#`}-hf0*D3!ClV4g{xUv`WThj7-1nH
zWl5t423vsKFx(hs=IL@xj1ay_W{g39DRFTP_`xn66^}mGrh90EX*#N!hr4?*KpPeY
zA<H6&Fr#xoPa57rNvE21AT?!UlIOcG?l?c!%CpD9$|?b39UiRxrP*O;v=^KdTegf3
zj)Kw)9V!W5vw3IFo_F`c_*>m@|Ni}ANJ7<>mAxIDoq2#lm;ytI$G_XMxuP~TH#Zk!
z(DaRoiJ?D6o}4MZPDBx!Ssh+)d9C~aGJFQ7d{9XZQ?3~3@e0AGXFG7H`Nazn#h)W)
zln2rA`t`!!ynZ;saNRo4b23~bLOah$3Lrfu%9G$)xrc{`=_FrnbcFb8*Ub5c)qVr-
z_2pSxKCC)(=zU(Bxjx|dz786F&3ETi(Re97H;CrRZ7ca`GQDOcN%Qfh8Uvax!{XDy
z!3ztko2Ao_1eTi8tjQBn@+UvI3Fl=*JbqGDH66D%9YlbENd^aFH#9Gpz6qI4NUA_e
zz}99N85vs3Vq)wgasoEY^{|<utmdo);EnM{AFQpf?}Vi$At`CSqvLYyD3d)rex96B
zQ#5$NC&2uD!A8ZTq$Eg5PLXMf!rq|NbeSk39QEjhfttAtZWq!bGH64IE+-ZUM-}2-
z+g2cSieZ*6T{`ytB=<Axtbou^ip<4=E`99B5mc1;<m4xPM~Jvb@h7wf%CLtB83a>7
zky2=8-$U}uAJ2rmlBDi*KrsZ?kkn3x+Ha-87B#-E1CqiS;_!g8Fk~)U)Q>8SwGY3H
zLOsBzB7=~Qbbu5v?%FjLz++e#N~@h+UCDjI%gc)+>z4Da&C1FmO%=I}kY78Afo1Xp
z=ntY0n)b((U}Jt(U++g<4bUo`l;&2rJZuYHs3<3BHyAq4*ZBFB=x9!ep3yIGtO7>(
ziHy6uJAv5LBSDsfl$ATS2pt{27uQr-S9b;=kC3C7P*o2eyfv7Ym$z=Zz-n*E?P)A4
z9j>q60kvW$QOVHe;>KCdle<-ouN@ycd|jT=X(ZRtVcF=%taz$O(_`jkZ!d-OrzT<t
zKnp7~z?t$_uZjR0rz1t^{IAj%TQTF5S6BNZUROosxJKM01_^1DYn6ZZ?p+F2-pW3o
z!4^pPAHx9hJLlPUs@!R9hN@b|^=*mYvu>dcA&*YFR{=XS#dMJaWn^S%G#H0P-_VW;
zB>@25Fy;wAEUNqV2^!(3Cj^*tt-kT@owWdB#a&!ndfvVB;eK?)nqtmRG;S82YmLWX
zQ#Qx@KIg-SyYMzpS7#Ck1OQ;(K*dDkwXL+JWU@Zbqu+^AS=-Psi2)-snP;=MHW#p@
z$+vcRv>oJ7=bG_&8lJH5TfdAxjzSe}3`Cn^M*xWgZ#3t)c?0@1$gD0;3B}|ln_vLS
zo*fqsv;-xe5R8V~6!nw9?F9&9S|6YivQ<qr2+v?sFh<!<jEcm?V8<qOgL#OtX99ze
zmP+q|0Qf^EG0^c^D7Od#@MI9n7jm~?UTtkHhZTo##&o0DEoyQl4y*Lh41pvt47_rb
z?NjmONzJQQim{H!EWyZ(QPX<u^)Ev6=g((;?lWk$0j@HMGbSL#o8HAaK<}oI=9Vg?
zF^Gns4@en$=%<yx9l-tRrDI@V<*5Y19prs%ZtgrXI@;LCH058v&We^$`W4rn@qYPo
z&)NRKz(5@x9igV5Gs1>Yg2J9IUn}bgg2~FxZZbo<a5^r)%}u`av>@e%nqc@bfHfPW
zAsg&(V3eYQ(c(^EZ9NweQ3ymAwUBguaD8#&FPseuYe0B-mLVlpAa<QH`SFpJ!GJ(H
z6K&G!CG5Tum%jWB(Ho?L`|yFh@)H=0rI;Weg4(0#M(}8+Ac^TDZL-n8veYH3I2koT
zY{?RuXCZv|1_RImp3Lgt!4f<)p`e_*vhmo^&!Fn-Zr^^Qv(t3-nl<!7xFPD&J}Avt
vH#efae-hPUW4guvfs*?#*Xlq0!C|EpLVGVrPaV`{@Q;SNmfG$0#y<ZK2bQEU

literal 0
HcmV?d00001

diff --git a/_images/bonsai-connection.jpg b/_images/bonsai-connection.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..32b725416e2ed32f12977cdf50c73b53ce9b3d9c
GIT binary patch
literal 113805
zcmeFZ2UJwgmp|A9MTwFXl$<0=R#J;dmZ;F=77@v`l5-<S&LAKlpkyR886>r2$r&Uk
z$*G&zK!@$`&QAC`vpchMcK^F)&ade@R8_xM@7`PWs&4pH;eO$k0k<?%)Kvfk1Oxyt
z{2u@}2T%eK5fc9O#eazL-)lFnT_YyGMoL0*{RTNHIXM|A85sp7%`FN_YDzM)TMW0T
zY3b<c>B*@WZ!^%{rlF&!`>PQGBK$qX*KS_Bc9V{RjDqg}_6PR^Kz)NSiugAX0SADP
znt+I!0M`j%1po-H;i>(L;Q!-8K!~UDI?0Wjq-6LDYHtAu35bXYiHZKA8h^Dv{&xT|
z^)(uHQN`=DdX^*{@94zdCw#rZ`J}RqUVjw9CH~4K=q4!x<Lx_4+&sK|_xU9xrKBG|
zl2Ll9tfH!>uJQbZfg#Aq*u?6!^&1;oJ9}3*cMnf5Z=c`~A)#U65s`^WACps3Kc%JT
z<mTlU6c!bKtE#T4t*dYN-q_yJ+11_iv$t<-d}4BH`q#|t%Iezh^^MJ~?H%ON@yY4g
zIqKr_FTMx>ME}Irzi{>kU(|TM2#JY_h)Mq9i-6D*pNOc5ud$0>r%}`+v3y6%A@=?T
z-IIi`m2EdU#q|;NuUtk+8Mq`?xRHOc_P=xXUt=ujzs1?VF!rx}%>u}Y2=JXpL=6A}
zt~6JNW8LB8sMy(Qt(GeZD%+~63$=?^2jVIEJ_KeUv=E#DIs(E0xN~rTRRPtr52g`v
z%1kr;bc{{~@5!q>U&pZeWoWWl3lelcw#?R>YIQ?MRjy^4cC)ZmOuENWX%%jzCd~HL
z`EVinU07xx1q;-!g(Iwqk{$Q=v#ym?xP3naBkIt`IDjG;of3rugy&)zmZE@L0yse8
zC=^Nm`%<LykOCEKg6*$_Cz_^+-7w$hi6u+S_EQvRi+$Db=xC-gIO}6Nu=9}8;j(R9
zP$Vy9ySa_5*G}TC*(WS2=#Aptxwo$C)m9|({AP(DqABQ#6RH{;1iuhs!~v*P|6p+g
zd=onoA<?LdWD+F2>hLi6sU@P__&r`-V1S$3iiSSoson8|9}E310wa>JNQ*louPPJQ
zbQ4!ZoUk)D$M<=faRAk|^Znnsdr=8j-R)PCy)K6s2=*Cd!4GyhhaTPiPuTSd>w=Qn
zxoT2$S2Fb_m<$vpsBO44_{U4Izqn|dTW7I2@L6r_dzu;cce?p-fROst;Y(5+;CBre
zt^DppSG?knI@Z4nx?CX=um}9V6h>T5pez;2G(Zps_yWUfD&qj1dN=?Xq$~S`9tY^P
zg`*<Zw!Tny{H@FApzU)=8_YH6k%<VFqv4W|=zkU=@~JD*_;-fZD|TaAzvb%LP1~_=
z7iyngZiy%B`p{Z5IXIlb9;DrqsvBWs2|Q;xUCt&1{&vQX3mAU9jV4Twn1JWAv5nBZ
zlE)-xCOE*bHx#*q$h{JRZZ^YFhbRy`GNUIwtOHE?Yj?`y>|E;I&KbjOeCEAq57+rT
zn!XGj@~l^_+A8!7FzJl5w3acnJ5pmwzqTT(J>c%xp@>`5uQ9O9j^>S3nZWo>hifSt
zT)Kg`^p4c!bfnG}zz309SZF{<QQeriW2@7Qh(&99)*&VZe!K`qd!qTU{osp35v(kh
ze$xrrq>T^>K?`T!5~iq+g47hjYRT@AXf(HKx={OF7_+OJFuS{`rrjqM4Vt-Lgf_wC
zoKO*?96IfCz&s_Z8DKIquMC~~*zS#;F*Rb^Uh=U)nhOWo+9vWeMj#)FaN2j@va)^u
zW*AoI;PrFN{&$}_Wc>xKg(d$#B4-uM2)7pkwt=Z6O&17;V5`2&nc_MPnocZNx;0hb
zn^Ppn?$;}&JayT)wKDVy{Be+F7fc1UErFsy+a+GB%lAyQJfDx2mDNy?UTQ`yRn7M^
zp!R5o8%SPTes-+}O5LBGhU=hMH{jYi@P@)s93TmrD<AhSy`uLoGzmmzoCTtw*Tae&
zL{=X{s|6*jR>`80gI*QCVtH!!@L9D$=gw$!V>}Tbmred?GbG0=*ra95ZFj{TYGdc*
zD`!W$ujVZM$<&)!-I6x*%K?D!F}b+x4!jKiMqvh^)9YA=L-T%5!&kGkY$+;iMIBam
z<K^Xdm|BXVJsHsE?8{r<BmjEKbpNj<&=BzNGX*TSWEslru*enHopl#+9tM*rliOiP
zt$%K|<KtKV>Bt1|?ORe;^hFO>(#ShiQ>b~j?;k72pVN~ULom`uUTWGPxGU)gL{Ks}
zN7XfVMn}+d+&|NX@FDGBf~lSL{gjLNK=p^^NmvHYP#j=|1KWA!mhE)h1xq_qZyn&}
z<e+_5-A8i-^n`mp?6%!^hL^zy{a1Y;7^tC5DGu;&r=hsN&tjk$6Tdfzn6rk<7cw6Q
zUGMGM!PqOlo4Pd`@>$vq1n9bS5NKLgiGGgWuJ62e%{{4rx#voC_S)Cfy)jJoFlTr(
zrWKqv;<)hnvz44~?c61{ycOYNt`JuIx4G}nhX(H~-K$}ichwjdw49KXZk%FP$+C6Z
zh4arE7}sMFPxShljrf@Z=Xz(wP3nMj5)V)AaU5(umx)?3CY#LA8W&dcdMhp(@-&Y8
zH4fl+8U2|_8icKb|CW{5$3P$hXTZNR`R7!9Bm`Zp4?&Bat)A;8Bzn<o>DbxHPl*25
zyeU}qG`doy^R8_2^P1zf9@FUr?csO#{?jE!|9{xb{*SfpZ`$5JR{uYh!~ba#*%_v=
z&?@qB)o5&hih1vIq^SJ3^X9SokG0~QuLw_cqzo&@s|^RpnTiA)QN`i_N78sjeL51l
z)ho~`C%)p4$=aDE3U5f{sh@&bet-Y!&EmJZigu;94?I;mUuNC71dT4yqe)<1C7mAU
zCWcwl@!UDx166DtB(BjkR8It=@d~Q!8tWyYIt~D}!W$UHp2y>tS7EN3U^28gk|F{T
zno!*Osgdf#Li*6rkT=ET0bAmY&V)~smhBpv^q%vNw<uT`Q4Xs@^3CjdJ39I0VJo~g
zRn0Z&Dvhn3G>W$q#hFSUs;fJ}cveXsJ8Z#F&s#<Dsw`Lb`56$)ANTiS#yKn?2is~M
z_o<l|vDRZjkgDk~EJ9_^1EiR-t!H&?&0-a{wRj`OX2k3haN<{J`A`n#e@|c~F1(es
z{Ij39aB4~8mf`HZ;j`u{(`!fMA-q}%y-hrz0afR>aG=ShIF{}(8oG6p7lVnz0U~q%
zUd*vLJ)Fi;UBI_FK4&<c*RMDPXkI#A1VG-;PS$RQ2=BopW(pDJ$+hKSXZgTwp}=~r
zw|;e^p~|H1$((@F@+5T_hw4A&S@dVcx6c_%gP5y?jdveGHaGyOXE*SKI0=jG!2!DP
z%f;WyFQV7@%7_0><sU?Up%jbU!>`QRs2*%!B6R5&4uJXw{Cg4bk5>L;xlEv_aR5&D
zO^g>D8y(vkG=}|+1DuZI0M;f*cnHh2X+MEgho~6Cw$fE!l^EkZr4-??+r(Lc(^Ym*
z7wF|S4p3DFutGmWm;tF<Sx`wGv=uRuP6`g9_lFN=MnVD}_0`?B`W%TUGd(j!^3)@$
zSL&i%%j`;4wWMD-4AU)&q>t<{oV(P_<~^un#sR!J37hJI=xpu@Y=7f3k-ZFH-JEU!
zdK+J0WdiQWvZ${?iM*ydR+Z1A(6kJ}#agAd6<euwS!~VeGfvHEHi-h^=?|@sfomI&
zN8;bt6)!$27s1|vg2w!qSK}z{(#w(uTMjz}bclY<FI0E$yOJ*6Pd6}dO31G14$$wk
z41IX#QX$fCv<Hp@IN=px;I@F>oj$P*(`}T?91frk9=92c#}3Tm2Q-?u2*ugK=wlXe
zGh25|vP9EiE^XjUIv(8i&X=iDRZN*LBu{^^LF0g`??*mHWH~OzkGj;zD_2p55(YE2
z$8w}ziR-s!O-=a<X`{@FkEd>$+|NDA8RB^9vYQnHexe5&INyb>SzmCqkSEqWVwi>x
z4F;55W{ZOZ`h14I+AO=J8dokc!2Eqzi$1DVEX*!|%k#Nw8<(TWKD~^F5Q(O#w>g6R
zuO9l{n+%{tfnQWkORBlYXl$Gaf*w@I*FAWeGNqUK!*BJ5l}M~~b;9S37UxJkzvBFQ
zwDd)QNE35&<{}P&zyUH(Ps|fQ`Zz$G6ux_aC+sim&kCN*RLW8G;{abt2d|!jk4J%r
zU)65%6=6pi(I+}rTY6{V@XZ|P(vv{0oZNrDKy{V_2gr)R0sKh;iIH*p)X?FlI6!L%
zt8En4hZl3gN2tLMtifCN^0e4XI{zGzi>>VQkuU#f_W##HK(Tf&k)7ZGEMDUs>7ljx
z+a^PC1+x}i+*yUr)f>hpgt0|^p!c-De^WPDJ(4_!>)zP%ravl!hnzMQai-s6)K9-d
zC!>DvV`*vL>swEbZseBr__DZ{c@c78$%B<@E5~XE$qcR>Oq6biN=EG+_tZ%!|Bx+s
z9(!GU7$&hqj(#%86Uh1u2l!xqx>BC0G@P{Gfe|$SBnK0LT(;E@9=q0~#4Q$ZfE&<o
z;1<C*9AKFoj@DvDgrS15flM$g06&U5$F+xW0AN2lKI_}kH2BzNNP!x$8IHEj+nH5O
z&7Va%dbOS%Y2pAiYSUr86N?Z&#q|#tSr6of#RGMIYMDeWTYQDA7OPW(Z4-BGOs<rN
z!`&rBlYrUlQN)%+V!h|Dg-M?1g+IEa#oodJZb3(g3~+!Scc7?r;0hIz1Ot!+qX|-R
zfQWonk3lya;5}CJvSq&Iz+rC=>W>2q;sAzn&$0Ab-)*e(GHk{-8!M648;w>e-khrS
zndK@A?NaQM&IH}_G<nW|gL2~lLWfz4?3D|l{7@`*TCslAd4CPbkek<DM(ts2uperD
zC*i%S!PX7y-r=$AzYAjYPeBRTL3SMilw6$Zm(Uk*@gH(}2#5)GaMTek$&gmK@XIdV
zZp1<Avg|aZD`ntB`B-7{){x<y&j3%t!G?sr(Go8E(ry(1VJATSbhQrpcW!h4ltw+z
zIcH^pvIE{|W?iMa1lYq-L?So<tty%w%ZAsOu9x5dfnqoSQj$}QuMCP(LM!3`k16nu
zVSqJWZ?naZ0$?ZW31AY6xDLkh1pecqe~928qu_u1i=YtqcPmln`U42b)>AjBTde5N
z8eJ05R~(=;dka2QTf(b^0~iBGp=A!b1j=;(t#n=ZI}X4a51p36zA7AnPMATb!I)%w
z3Rf%;KSz~2p}Tax-~dGp_&ofIf8=^5o!;YX^lfNe)c!J65W*B5BWF;!r9V$Nb(Z)&
zac7ybfbW(>q7`d<L9;h}Ifs=@VYJp!?ll5B{1Y)aVrEr7h9@66f2m6{UIR);3t;jT
znEmY1_Wd|lo8*<A$H?`%rFG0Mb+*iu<r{nicn!__PUrszoY9Asd2Mzq-b=`ts4ZcB
zlzJ!BctTBl{P)n|SdzzN^6ft7QP7)p`e8fPTg~FIXYs!rX8i{a<_zf@^f+0qd3le@
zLqCj966f;CsvC#U={<|jiR1(PKahx$<1kNw2O>wXxBzxPOC&H1D~A?GNGOORbwd3(
z{g?HM=T%pzD3wbVTgDzLJB4;1JkIdqVk06-$O>rx=X>uVhJP&W1EaDCP_>x)bsXT0
zm!uo(9<<74;Fz>`<6;igqt!&|vr#adB~nn2Uwy41-AiepbFRj12>K=%@jVnyFC@oY
zj25kgBZoW7>~$C&#*Jp~S#?Y?x&G{BcqXmeZ|NlX_N|eG?cJpF$Yl^cg6yol5gc;q
z-V^IANp;(69E=Pt{Icbvo#5bfZ(&wZn6QZFwgKl8GIdT7;@61|bFCtN6nM*_WNkPE
z%Z;7G;x&&CydZzj9{lxj>s+IJmc{<=`day(^jNb77{%+HD*YZUpFg<jAf#zJBBxs9
zFeU8+4_j*)|E<|`-N7{K4tHm9i0*Bf3jW)#!ogWba|TD4rL{Z^{bkE@jb4%M*}@f1
zs%^Wk$f&cQ={jNCR?7mWF+{8P&ZC{=$hcD@gsAgIA-r5^c~28^Ki=s6{W>9`{(dhY
z@&rbKvRUovfZxeKs9ak|g)CF&%kGHon3(DAAS^f%nZ{%V+>EUPLZz?uFl(|<(ZDOt
zlKc@cBgEhVEZ;7DVW^t2rEpi_7NRyKf+5wNV6ZZB@j>_Fw6}iST(9m0TL0{^EJRZw
z4=PuSPvLja<`W`8TjO?8>5ve|yivhbZ`~3L^*asOgAI53MA>ztzeScbbelC{SqVoy
z&SB?e<{!qKc=3Jc1jiCVzHh-wnLPKc3!v@l%L}}n76eH6`|+-G3Y%Q8S8Vsx?ZVH5
zPj&8kibx3mc*`9_RW6W<H){4TmiJwitC8eba%5f1a(^ZSi#LN&8t!BQik9@FF2P>0
z+w0Va%zkZmbhU3@Zx|+rw#hk5urm5RoGL^@!?=^$8q;@Nz{V)qpOdGB&$CbFHkuRi
ztaj*xz30a7LiRSV_)wdxAiP_m5T)5;GQUWNco87fI9ZMs0)}zbPaGu579>8L-7v1G
zXbMT<py+(dO#fnU)6VJAg6yz5Wg+E5-;Pi|4nSoGZ`_eR_sNP`8J_#LtBFkQG^M)h
z_2hC`ZPd@kZbhEopP!FO>-^3QVNdhVS<(#APNb^LRRWdB&HdixY{1|evu`F*7gO?d
z$mx{`rP<teUCNQVtEW4&oW*0iX_u87vUQ&m0)3{1*1*5?54)I_V$1jC{24MFR#|C{
zy&9(&M+U}>?HPXExQ89dF}ktz@kUG{9N4F%5B()#JjH_)!UxM0cod5rn9ieX;EZNw
zvnbEKQTFP^YmKUB1s6sS3HoWIj&6URdQrYc|GWl*1AM%IpYyIX@fRc}g*M@>YTkP6
z4C+K)1cWJ*Xb~}+IJxb2a}{IS>BI9;GSgReJh9#Q#KAQ&$m7Z==J#vnsfR2efmrb_
zA9x`8U9O)7!Z~g=n{ts1Lu**4*`yg0BHi7=$^V$`j4e9o1}qyZ9MI_l?WBq_Iqan1
zzv_6flYaWa&?lrm3e8Ru{zaYRae|rYP3!0mw#NjlVp0VB?<`Icn{84{lyYP_UiPD3
z6YojQygU}#ds(H<T=Fnzfrjr~l-Hh*NV3#%#T(DgHEaqtGlAPq7~~Ea4QO5p8_vs#
z<bTYU6AAcVmDbTzH~@#mdmP~96%@lzqqhvbpgEzsBEd^{t2QT4;vmViPa&qrXolr!
ziT|r}_Dd&D$I4{Rtj$MfMZncMylQ=bzd7g;4san>yN?4*W|zI$v3M}z=8)@D<ly5J
z(--(mTKZl0GD{e_;+GrnKW>RbmBF`Uzd)B`q1b07GCGkGPClw}&n8RV?{G3_3JL4$
z=T3z15>_?8LyB}PQLa^CXB!(BrH6X$EDL|lgBQ(`nX~)q>Q$55V~xJWe~Dg`fI0QF
zb`AgAW^=IukW0!fRupG8J}AnEfW*}3gfsC?x{;m2tBx&5WSJhZCsR&G8qJf$D$9vI
zB&X69a4B)>Z8>{@Hx2&I;@{{k>$^|hf%k;YVS^h<us1gG(1pG~Wm3=gg&g>QPK|&|
zB;X%S{8P{TS7i}p<PH;EpK>Q-p88>w(%ris<Svvs*4vOJ8*Q~%f5Ljo$a*^cmlrbs
z9k$fp{va5evxaTN0q}SQCii2{{dZUD_F!^gbu-m+Fd15Lnq8XARqlyeO_WfV$^Ho9
zz(}k+EU2%m^#vP-pxhyTOUgl_nCnQLEt1&_<d*o;hd<8Z=4K#)Bx1M=@}X5|1j~bJ
z>RIFsF>Y4&DzWWld*@(BLmafkwtVYLWshOP@S{P=5+3cJM}tH^>P)oOr;5D=FPFk0
zu~+*J&*O(Z-LNoD7u7FYvVe!>akgZE0G8xkEG-I>YqlljDP18saZX*DA07SW^=klK
z>}2sf1iYjAN$<Uk$G_ikJo_CNpicjN$ZvWTN?^yrAIq4<P}f=y5gXU7cBL3B&CH8b
zW={zF7Vof~J;QjGI4!4GX6E+*73O+&NIx?txfoaOc*DI!;>SC-+j{R<))%5l49NlP
z$c`;2rDcXv^Q4F{c@&e@x)ST%uh#vn-%eOFy*xW(oS}f896!a%`Nq`X)7$PJL_dpn
zFhld6>YTs+(D%Vd{#kNgbm{Q;@C{UZPj|XwYRVI}PFD#rC9(a4+bnBxuU}#%zgyU>
z=%)j_scspY#tYVJK0M`XKM{(*$*Mo(FfKs*<zFYjRs&1{90CaiUpsX;BwE_tcUrak
zDk>4*!4+ZGv_8M_KqLNUPs6YVY2=2TMR!;?Bo34wg6WP}DXFWi2&vP8I(N$L$$YRJ
z6`yzSF>^<BzZy_~#G9xkw&>;jHPp-6xU4`q@lg(Y@@OFY$8+La%Kp+wfetX0AeL71
zT|rkNR}j6FwBfqn1(*|1agibSo8|LKm)}53!Tlk9kM*s9Vr?VGZ*+G0!KpQ4yXCWR
z_<_P3*h`cma`CFYe#9>QbRf0y?gtg_kEw~{6>O(`5BtvU#(q*zt2>r#8mg3i99?vD
zP1Ydx?ae)HN`=sza(e?C6|+%!GeeqKsf<Dc@UcbM)+@Ctj>ZRWmhpWqH%<#d=>fwE
zT&bO43LyrULE&rxqg1Q&<R?`D8ar?V(@EpyPjzj6&0^Gok66cK@_h)yh@LY2oByuW
z@^9}W5n#Oxht7l7E$&k`!BEvEW6k$+QRNV#Z5h@ON6Ac6P*q)B%7GFvJ-&z8tcOc4
zC(15uCij&~AblI}>>3WBM|Bm2F6qxjIh5Hy@Os|0NQvZzg*K}+(zj-KV)FAdpY&b+
z;&U^8W&AZ>&Ur{L(V*wkawxH-J9t_ZuXay`_@^%E41y8x(_+;hP-el>W?sQ6uW40O
z4c8P%brjNk-_=2K<IDbyk4kSVt4w=jLef5!aNHhIxRqI3_6<9lu8>+dQ;+6cWNV%(
zmt0#Z7!k6pclafO9XQv#`(c2Cbd<j$UQPEVT?#<dLfRCQFdXgX{}u)7G=W6=DHeGJ
zEp;XgfAjLV@Am6}quNx5doMa{uwJwTQ1Ez--><#Dc^@TxW>bgAObr=hE+f&-RznJS
z#d`TkuY8ioxF1oSKB07%`Vk-Clhg$Zh+A!lG^Ezf&jNhqgae!%WVX7TpnQSw#d>pC
zKMGx6GdxSCHZ74!V=n$=>ae!O(09%avBO;jethQgsk}3P!{~)F&m@<IqHvSvn<(fk
zmGVR_kYxK(JF0A#F^lWU`RD$tkbt(cN2@`5W@+4y3wS>U(FXX{ku`jz=s>~F<j3XI
z9jqnBQBct2o@A!gbNpKOp`^pE#I9RN<@Yd`m`Tk~<M=%?{-0s|+7G%M$`P!1t<-7R
zT(gbQqkifEEKFiwyLaEMg^5CD(X%Dxrtw<jy<*a@!31vyln9567+z<kLZ{+%G-o`I
z>vEI7r^n7>radlZhflb|&fdq=#>liE{s<5mV>A-)$uYM#tg9#}l+f)3CRS#|$&lEt
zCVVtky>anYO%5cBfSsi`AyYfDNq0b4D!eZIyvSB>H$65jx2$~5)Z>jo_Iaw&3@2B{
zEm^fEcZ)dbBbH&mkj)}Ul6DK;#;ImmU-y-{;zD!nrS|rFkQboZtR=?TD6^mA@WbSO
z;>X4chn&u@vnx)Uwk_@&A2u_?>}M5Xvsf4{_z=&0@~-#sdeqylB>ujAK&RH0N2zx_
zphQZ%lj>QxE~4GsQnUYEf4!+bbraS%=ianaBQIiv$#-LR!0^v0j+`DcMeTFUfyKOh
zZgbfR^1d-De9$l9szKQ!%cg$!+PcaMmc#B=VJOMA?t<c%T^8=9hBT+;@fuV2L>m_}
z>vUDM99_EgbvLtR8&^IBWu)vG2zX0Q{HvGqk|6?95LIA@`1NFZk+wels4XejwP29!
zN8R%WJ?!`Ba_E!+p$Aw|^!<$0xDO}$981sZYotI43%A{)cqSZ0lC8W3q*GRnqLB$a
z`vr=}Iif=kob$_UJ<%dq3-Iq}o}hiy92)P8C@Ad6j_`m|m!5jWJ)5WzQdMn{@yO4&
zAMF+v;-&84(;t;fcqrP>d_897?SjjVyRaar9R*c)x)!#<P005tBEL0FxRNSD^2SSF
zvJy4yRY{vf8S}M^=Hiq`qS;C7FS$@82r=OMOVx<ImBul(X-V+DPBjXuti{te7*liW
z(@H&!8oet)fcm2V7Z)AQTx_q$u~$@=Ok9M!7Rv$aEn80$`W=*O$8PRm5qF#Ux%5G`
zTD>$tq;G}OK!p9~F@Uad?gDz`4903|VKL)jLM;b4fGqgn*8x8I5IZaE46UzOc;2=u
z#NrenUdHt>tNmW9TmSEV9x3mH1J&8J0dQ*hql~!OLai~awrM$vH7Y8V#%p+|yVLIp
zp}n((nOQ+6b{*mMzMoHv!`E+gpn?pkf^YyUb1J2Av{Iq{J1=cnA&^q`V=*tywy8$$
zIYmCwqv5w@0=>e3t1L6`UeWASXPfX=Qq)}M(ybZ_vhDyL1+Gz;+#T4Kn0=$d2S<ic
z8wS0w<aFa;Gc}@wKK7&7U)@X*CTDRr7h15agrlXF#Qf6SeF+d=@8QO)?mWkHo4Y3G
zJrGJSXB^-~O>5)@SH+m4`zu3kOR+^A?v`fil`3nyS93vx>(d*%(fh7vk)zEbIlJ{f
zQ^7F1#*`@GtW*<bCF1jZ5$9_yHI?CSycSHf*4AA*`XQ+gNfIW3)|&9UMyM2#d#LjM
zPMTdNsf@gDCbpUoCuc|J{Dx2+e&d2;0y{>ZN9=VKE2j2Yte7OHnryONnYS73CPEwe
zMRuTQEEU3$N%9+a^1LTIxr0DI=c6b2S#Jz}@&|8FSR6_=f)N|&eT-SMzsUN902=1&
zE)$2c6XKC)*fB<ntZG83+saf|R#sFZCu{FOEa)4a8X6C_juUw|4YU!KybcmI8AN`Y
z+w&pi4e?X+Q!17G8P=gcr*sYdY9VOoz&hFTXrzit)Y=*<O3Fx1@zN{x%)e=(S-A`q
zr;e1_wqVch^l0T87v>pn9i_;twA+4(=x0qYu>BpbnF*t2dD(X1;@Zj)Fi?fo9d8!#
zUUdR>Eb=2gLZ^M5(K;(ekdtDo@77|fXRMN>ACA_j>p#ENg^Txouf9jSTlgCWx$-$b
znj2d9UNgHC3(tqMrsm_L5Oi|D5<lQ~WLl$x5ap0S8pL9sv?Z<1M)PBW%G=Qz{rlZ2
zFP0O6{HlB4HS0qei1BXEOK5Pj@>f(&e+S418)H+ZWJ|H%Y)|lAw_flohaHDaq>B2J
zdg{c-<(nN##BTl~%VpCYzBScO{Fi+@L&sOFTEUthqGGU?kvCp&J%-2zAeR|z94mj<
z>dY*Gkts=KN^4Zia}F{e7H`85b*Y{<IT^||q5IU<4QUjeS$>;+QQXheLtavQzJ4}2
zpE!plhFgG>FLyF|K7LuRJw0i}Jc1q7+f?Ux^__QMsZb`ZToZm;m}(ffEH~q)+ft2|
zfXV#u`hqvv?^4TGT+Gl6f!l(Lek-=DwdL8XQT{YqA}dQV2if59>=&b!SCWxZ`NVI4
za$kKbDlY=;Uk_MOC9J23iEpv+qwH{in-W=skbRMAXuD8xK|VGYI$Jk0Z|Zo}VRMO@
zxoABQ9a4R?%o+$s1mgomc(h5x;T&{@3~@SzFS7)I=j0Bnu;&=dqccqE0eWU>!YY$r
z_IvHZ+<^z$u@x#V<MV{nbOt<ypFTpOX#M<x&O0u7YFFUbykOlj0mb>sr4cn5re2AQ
zi;p!9ebZJ{l7c_w*u=lIer(O1%37lm=Y2hlH&hO|O2yn-lBWa<cT;0mts7q-PL|8r
zuzlAMMku+!{_B@g<ng}J*&|;#B}!8z$rN$WVx6-kb$K^sQ3z3gR$C#;hy!TUgp?-;
zZCbLE4{-lf>wZw^VOYDo|NP7^tUV=wc|!8I$ONS_+>z>&92ZnB+2o`yBpv=l)k%0)
zHvV;d$#5gNW^faPS}(8A`Cu_68P2!e)MVNs)MKi=1R0XX0p^<Pw&6uv__(D<+^Rp-
zlL<E}N>s2}>WfhhB-t;MF2_^`L;cq($2WPE=exKFs52vC8dft(1I8LDYO%>qcwgpu
zjPhB66D&<mo}5O4fwo%k`6ui2NBJaHnekk)gh#pjR3B}ZDuZ?{2vFYc-;7aue&z)z
z?P0}<ugI{FWhAla#miPpu$1WCpGRuap|35YzqYT{qB5@pvcp;Kj5JF_O1uKbmB%Ge
zj_eMHH|GYuU-q_rAAb|FY<mBbReWru9w|ql62!LKN&X|8tp8;Fc&RlpJ!Q@m3iK?*
zUPF0IgU;fShmn2)?+}M-h)$*fo6amYQ?fx#Wi9#5Ao}|BSK%VymmWveiP40rBC4XH
z0by6#MqWwnt;~}(uxnoB*yGN~<E>f#o>!@<b)!AZ{ch3=f_j|=%KFv`w=0<qg6}(z
zcP~-)M2nzlT_hQ5S(qlfi|E(HYKMe`h0YiRrLs?A=DuHj>UnfmA@xdOr<Dc)>nM^F
zU2QUDrSp>3@MjWwt~+M3Gj=ws9IjC{Cy1=*SEMaA4jpSgAdc*oxzD?#qX|t1^DM5*
zDu8~sUiF{UuWXn#Ve)I<gmnnypzsXwPWUNa|0>j)QD8ckhp7En+2u~x?VY2x!1AQI
zZcH!mmB#gY#L`*lXaLVHlYC}MMEQ2g>_*CEcU;j<yb_&EH1($xu#ll)vN1XGw?;(6
z-GcRvuL#JGhs~@oFOZ?Ri{@M=&zv!1jAR7`)V<iO!&UBrBK$`pPrGF+h{&@)(GQ5i
zJFagC_=!RGhwGv3j4M}VWLSAQc%Ao>HRf<h!D6Wjvy>rw8SSc9KZ|V@!A>7Sx%1($
zp5c+>IEH{R3(-ncuItyfl^Au2eUTo9SsG<4io9z3siUoNb8jWLO@})lTu2l&O0A_I
za6Fa=KEr&eiE&jQ<sO$1a=tR1v}vi=q0BMmKB2O?AGxPtVIA^R-%0;z9?fpOhlAAd
zGR4AlSDKuOoI0sp*(x)H2KwfnL|QYdre}Oc4zDCI-??));E=6^JiY@Wz6$?!{tU@o
z_HkTi1p@|EOu5Hs&tYj`6a{64w`-Y`<l=kn$JyT~RS1@F+%10CL0=sABWXz!GYXDs
z3NST3bX|9Pcm}@mg<jcToNn)&WivF(yCa5mk#kxSOe~oY&v!oD;TKcg6OHrTTjIyv
zqiT7CiUMt%#K_}-u_+ly;3<UOe(?q}Aj!^>J5tq&yR2*=wck@riEOr|rL&^{xM4m=
zYj?q0cLDew%NPeh3+UXi1J(Ihx3`Ivoh9U&%qY^0iHmyeQ_JNWyjd|4WuqAQouYu(
zXJ&Fi2w=w&qnYivole>AWlz|Qpxm96l<}G?NcGO!CUa&{xy9EFZ@jq##S(;{gm5i2
z4Na>|Z^BmQ!}nk<cxUqt9`Pe{bq%&pk=s#-dVh)`F-N2n=c+nLGzyo|H<@dvirp?V
zp(P&c#yWnpA_`C<e_&B7i39XtXq86M>aOK>GTjlywxYAW?{dtP>82B-3PnvWES^@W
zY#1BlngPiRJw^vzE!Z$=;8m5F)lno-Cqn?+6k5^2%6jvc?XH8B6aDA?cg_q`#x`OB
z7c+S@YHe+^gs}oZMNDGp$uJ2j^%)}z<A|iWzNtWl=`b-$VNSWeI(^7ezB!cbs8{sZ
z!1zNFJ)7lI?^jlMhG06fo+B37(`d5A1rHe~(3+`GR<PuBsGQBpg2|eRznDCsgDHLb
zk)@S<U-wuIjOghuD~*yg%{4=7!pCxRXa1O+^77^zST^KY{;GGa_lZ2v)L-Fj{F$k)
zi13@BMOkWlx%Sryg#1J66;|^vQegU!4_Kb*0z}JUB-##AYHuPE4y^0><z{K)P`&3m
zN9L6)F8c*@5#OAsq3KLyAI}&TcP2S@&`iY}>hAx}T~h^7m$b4B*3xQq3K^IdotJPv
zC@($Yvo*S9peE?~cIxxsh5yvlq9@+QOUb_~mpZw+$i$lA-C~CFq|i7So}>a)q?4!C
zq#v+oOWlEcD9-z?{({W>_t@#aBd!E4vA!~<z>E33do2#kE5lvDTd4lEu!tIYohMUr
z0tLTNp4u%@s#A#rebt(-UU7Vhv&PB6Tzb*oVa!y0d7oFjAZJ<?V}87=S~BW#A~`!I
zd37L$bBmjZn7W8-pQjQd_ScU(`nd+U4c^dbym%k-qMvpK4Mc%gwMbTZ>eqmDsPdjp
znXJNm6FacYxSMO?oM|4-yxJZ5!nC2on!BN{Ke=w*Fvxet13*jzJTIKAWLwpR8=>18
z_>~$DQY6%w&DF$CUfqKS`GLLZ8V5JSoJ1^ZYmXq30&hM^*7%ka`L*A?t+7)HP_Vld
zN~A3^E3OG2l`uZRZ!HW1O`n=(?5d$f>{CjLl(;_^2*4%t8cf}dK3@~sbb7U2QMaOA
zfWC$u#+%PjEN`pLnNIX66(h7KOKz#wzZ%1evFWeh&xU@!fVw%a$xzK48X!(5%d1!7
z=x4VV`D!O6OMb#6DcH-)SmyK_ub$P7{ivs|cCwCgC@mTA<|H@#>IlQw!VM39*G&8K
zn;a&=SG)rvMW%Gt$83Qt&62BRnWPWbT5G~U)$qpHjU(>Zhg?+&R*ki(8}rpFT~!ZW
zgwKaQEQe}ga^+;-3QP6}uaXHol6zF(pu6pvQn#XEVKzun6wNN?p93;5kG=wk`cH|x
zkON=w&!DPTgfo}%@h$bkQpwwrnlUDu;2^r=gV#;o@ES(vw^0(FAnIt^3NhACUxrdU
zVKdn(s7&uw@IG0?+PJ$vmDU(X#TAX_y`~xSL07sDj|4avl3#T82NfxDNt3`%b8!HP
z)Le*flqPU@Uj$RPe>rmx9*pIf^QtzStnscXQ+hmI)3B1YsWE|066=t{0r+)Cdbu1+
zUnP6FD8{@eibi+XTs`&bgHocOP1Z6G)7#4Y_D~SaRvZ^DbMqV*0zG@jRQKxG#x5cK
zfnSg9O!EL?28A!|$KiX!C6e{2@(}okvO`HAHxrNqrA={tlIFp$zLI{yXw5{;zP<-i
z-_Ud4ovhBDW2Y)6%it?~N3BbCt0m@3GZ!<MS{bb-7KnoOfK|-JaPL@zpZ1&!ymdA}
zUW>0!I!<A}|DYP)cQ&$9OBL4$3Uh+Z9%@gzGs|C}T7=wkkb2i~=JcSKrtjI;JY#0+
z#-HtcOZF9)pl<@#b^8>FUsEr(zC+gUS?8cix<QwF>MM6xL7D~=1xTu3C-<lc+gECy
zGsEKh!EW!$4b?WyJKpLgWq)z&H|<w#(u~-DU3*OOE2qBh!6(ZbxmJmL_v%SxKeEmC
zHM!(xs2eZ9fV=4C`ib?BiC5XyQ1GtHDV?5i`nA8%2LDOA3&#h0ert{Z_gL}qD?F@T
z?(g9vT@qCYwiM&gjTJ<dm?Ea%D+nn>+*Rjj^A@bf0U(L9tHVs@turrg(n8*az9j1j
zDmivd*h|3yc7~@K4#Qa(>*ExWl=YT2-mYc&AM0<;%`Td&)YIlYjCuC<XH&$ls2(nW
z_*glOZekLsQp~J3N5I_IWD;$e+j5XrI1A2CUcF$L{gBl4J8jPs_-TZF@%j<1^GyTh
z&{a?*YTnHqNfInlyF|SMCPyn%Wd5o`x+nd(v@QWAY|iL#54_<y_Q>T;W2(Nv>>xl$
zRbq>!8at64_q0zHa$W!QAr+KlMz62P?$_h=o&S4}dG1)<H||5(G9V0oxu-vJ!J>25
zf!2#fy|%?-sK4#VYsN`(DcpiVDX-_-HvHAkW^q+^Th&<xW@BOf$AIW$YgTu`R_@-F
zet9yG&h15Ddp}k9tJ&es%h+PBl;dr!y#8#3BNYYzh89;QO@WFxFPI}p9d^_2cOBJC
zr>@`;06*c{HU;OhOWzz0H>Rf-E$S=`FKAa_;in?uOVns4E~GI0U0&}uud%a-9FxLj
ziN8(lr`kukN&5!r?0*@}t<P%2zepBE3n6Rjp_J}rsvV2cNXBsNLn^WzMw9vw=fgKm
zq0<I>33uya%kD*S&BXG>KjAh*S2@kY!K3=q-+wllV0?B_CFfO>My(KNX2~v8=+J5%
zIuq41+Y{N!t_*>!Qc&rSjz6)Pbaa%nD)Ne5l^qsiXC81CT-Ce-5HB%3)OYQKHgWmm
z{nL5*gHFH_{3s#9UJn(GipASlh+h17)HV7qY!xJ+@42izEor2>jEcM5{R>#Gk4MoM
zD%5%^;(gBVZ~$}YnZXy*R=)A_17<CFMTE2P$!@UNT+=>H2`48XZ{RP-Am3<DSbFQj
zfDsekxSA{cUV`lh+V$xcU#2c2=f4;Rc^#Fp5(iHlQOQ{agN;FL@H*8Fhqw<w+5%LG
zmGj^-j5_=!l3BHHd%BN-ulyI3T>nd~lyfBVE6^{!=-?U5iAo1y-s!18KQ_xt$*^#1
z`Q#;l^O(MxD~d_HiCFoLU=M)G<RWqit>v6Jr_;#DoSf@S)uJ&?Qx!k<ouo+6!Rpf)
z73FS<bSv3}5c_wP@3FE4w$-6+HZLC^JW$m@=tO3dp`F%5LS#;}!_cf{JJmbsZV9vA
zCVpSJY1m2I`a-&I-`%L4wIoa<yRcLh`uwUDiNf!O;exl}eb3kN#_;G(Jzgud?oFzk
zU4~gmr-SV9-XR`z`udk`<u7HI*=rYKG6~&snF}|6$4+j65!DjVi<4=5FMXbXqPiT_
zM<ySYaJpop1qWHN{&>v}%`3@@6;=JCY)%13`*vZsum%jp-;8a2oECG77TOB8&l2R$
z89nt&-zhN#4IJz}HUQA5KLqH{8Oooiu>R4;pKY3Let_CS3BBgb#%nUA^(NSxCON&`
zx+f(}_hl&54LmDr(l>6eeZNqEE{Ly-t*sMJ=NZtr{>ecADBu^kf-w>+?B_StY_mLb
z_-3rOn8xA5goKvLBes{{J}I8L!b9%TJUSd^E%T7+3N4o0GX&M0yUb{vIphTycm$b$
z!<q&L7|w3ZX;a>2SuGp8b8))D`T@&^@ceM1ML5@y#nBL~Q5o9Hlce^Yf-Q#dfYatq
z_q(}=Fs(+EsvYKip=3+<cTB|?RUKE>-3h;B70v@{NFr$I(4$A`w5FPiTCS?Ghly}t
z?=2mrI&Vdh#Dd1K$#|JNg~ATmE?H5EuWyMRae@{zz+m)aJboRaZoA%~kguVwpS&>q
z)%Za8GrM+XPVNlE`P}hXNNTQ4#`Aq1qGIgyLh#KqElw|2-D9pKtPkyP|K-P=<Akkm
zC*_D>#yg+~Gg~sJm~W4$&B~)Zf|U2>OdhD-i}8n)`k5l)!uHTwzTX_)+MTg5+pbMQ
zoD<gSgH#QyqcuQQKb5{5jZ_nH5Q^<k=n$hF5!sy-E>4WkzE6~%7L{RVFlpsm^^M9C
z@y@evA_W>=d0UY`NqyXx|B;4_eBqgp%#?YP2ewwU-R5>F^XDYE_!I_v3R4P45tYRe
zqofXcOvY-jWQ#=>-Q%tZFLmO3n1Yi(#C~%a!*A(%gZHK#Z^Q7pM@4uPT;#TG`6tEa
zk&#>8pCXSsCzI_Cd0o@-+F=@A2N^EF@7nmVfzNgG{rf6`e3(2E+B19&#oNhfa&I5+
zyR<)hTY8@2b3F`N(re8eq*t!|S?q!!klSytDJW5l;M;?0UuO6pZwTujWvvw;%a+0;
z>=d?R!Vm|z;z#_la8lLbw`9_+-ymY^5p=w5t@c(?GQuC(f$ykxyH*}E2Mh4~m3qAI
z!#~AEF8)e7yxluG?}S|&uQA-R>t6t+WCoabvHGg**yH_M19(XG<~SIibJfEE*6_y~
z{82;Cz1b;3`D?3@1;$fMyPZCNRyXTDK|0psa!!0h>LW?m{DYoY7q#-x%=DH^yW@S-
zSDvgzxbP<a<t7tX#R0Olpm>+tAU+q3f54wrhLf?Cmfy5|>67<c=IfsOr*E?*n}12Q
zho>yFCf7AejR%L4G<;*ZgG`BAyFNLs)_i3==6=(yX=y>m8zdSc!a8t$kHiARwi&G1
z--iPTNuEslZ}~#D@pJ$iivRud{sR*5Sb2Icm<B>S*tN_71b#Ix6nWF;Xi=+aK`-D1
z@#kv3WzV7$_#NBJ;fUH{s+mRa+1ALqeJEsI=7PihP1FXPfU|+sQBt&PB;ddHK#-;~
z;NMLwe|EeZ9$5Nc6k*J8q{JMM+SvB_yo5~Vo|ujkbLR@ri*7`fmE~JX4;zmW)7!)`
z8gd;!H{kzn!~_3b&9?0qN(U^%n>ZXm0YCP!wCVWVZTx{Hf7FzxL<0U-G#Glg+WOwo
zL2_`wWTB;oZ%k8Sk#&NeQ&-G5A{YR;$u$i!tM7tbaR0UGu|jea8}0+x-71H+DLkk{
z)`c<T+vmSZ`XV{<?;xxH5~hC9d=n3E0*1l&?m_VSYD#$zHiz|`@xnla5c)H2<?xEc
z6pwSt|3cBnmlH_(|8aT*CBiFpsb8SK6qaxR<uW=iR_xOjB5d0k4sa=i@4DC}jPd1k
zte-iOGBX$Td{zFSkR`up(WK2Sg~1lG*!f|6C0_Nd&@T>^VfOhk`kz9_mfG_FbW?_(
zHX?M10yT?0I3pPm5iD$&aC}|qZS9pZp~}x$inMZjha3sg8IU!*`ywg5kq2Zq=W-TN
zW~YZr((&ie^7}P!YStsclxOljzOULu(2?Uib9<b8)ed2xQPPKpzv1?W#CU>8_Kw%g
z!KR3zv@&zM#XA!)iT)U?XQ~aA9<^_lZYN8(U%cRDHln}t8<b-kKszDm{_9j;Rlly;
zt1PcBR@PK1kbqNVz^De6eX1F4RlIA0<v|=|u2E6$q*KBmXDtFRYE$|SmV9~B$iwxf
zUE(|V)qb)?BOLf6)Fo+s#Lrx#^GCBpznSp10m=O2Mz9Ph-}V67jIQ(vDQO){T-}VO
zk=e8HPfm-hV$bHjtcR+os1m>Qs>b_)r9;r}PNe2rjYBcK5l%HXB`_i#>7~0g{T_As
zzo%3$xm-CFHtKCrFRd0$$^Fk?AS_}^TLaHi0$gpSHGK<NC8vLWQsQE7uxvVa`PtwY
zm+|l?@w@Y->g>Q;A-A1+Cg(3lnHushjmy6W)RFll39ha<cCKDa8q{6?@Ul`++f&M8
zL%+082nJ1UDxA+=I(IFaO$>XHcyToiz0{e<`mu&!rM(W$yJqEPKU=Us$n3MkCYw1a
zGBZne{j%Rkhcc(672N+>GdmiN14Mbf%M2cFwnSF1ibN>f_Io~B+tk9_uE4HCvSSi$
zqSX^t%PcU*6Y-8;>Fx0+kr7UMMyVq<Rnf|rU?)~BB}8g_yijC8qSUWlX&E%>c;6GS
z&ZhU+(fq;HvnBSv7&PCHZi=|Q6+9GHCkzzgmeC*Xt|jKgU<7lIn)?{qtrsCwmSX(L
zr2Da69YyeOor7|*s(0(D4gsv_l-3MWt1@lTZv}OYP`3;mV9c=vG;l;g(z7UpvOyen
z!V$4ED5?GqR`OCy?D)9QiFfIofverUy^U$7UQ@;Op)ZWa=w9ih8m0wgVUG^Wwr4iA
z^NDS<%&bk=X(prc_`b;A0r7FKpWEt~7XJVtF{4W$gP1y>&OAFtwv&~hnQAr0UjDh>
z7*FD|D!I54!&g^|+8t$Vv2)p&a^w~-M4t7YfQ}yO`HH#e6=dHBoPnt=GO@=SCEGM{
zGCR|~M_end=I}RR>t;mN@h+|<iOJgDc<9D8@F8eC|Kz}B%q&6ApYZ}};W>fyjf-KW
zZGKeb^}+rarH$=rZ{-=QkykFQDJkNf$ZeJK@=`K?SZ5ouJhb*8{wodu0&7fj7oqPX
z%F8hIXDuRw*x3moN&NhD5}}&u?(qA9=IxU3Qnk`iIl;`&M~`;x8S&0#EFtDORBQp}
zMyIFD{-0Uz)Mn5UR+FZGv&UdkAnw#^7v%P@m^r}sw{!eyJ!LUn$@?|ZDW23d;->R#
zUh68it*zy5DY6B(4I06Ks3_#SW|CDcqK-3*ThnN((p|%IPhFEO$g9wYWdj^49!=lh
zt4dP`(j|JYMaulLlde1-j)lLx>E`<#)_Mz7;blR>Jsc|e)8bf5U;7<VD9s$~3fQbS
z`hxVh8<P&TEjqXc!-I6LXGHtA$f>?%A9_4AZUb-i@lcpNGt5W6f>L===^`y!mb0mC
zrA$0Be2W+7%+(9~<!(AQol#RvL=JNt+^_xQ7v`-L=$j#%*IFI4@t}<-=C$s^TNM?k
zOqo7=m+dTmmud*?B6p371pCA1FtQ*P5|t$xyQ_uwkY?RMMD{un!}kU~ers&0DBeh*
zoj-#WDKJ2LMG#QT`%-+jJORm;1&pMa1bi*|YtsO*i-m8bMtyM(grtmuIB{owOvqj;
z=*bioCRaL)vVGG=9xNunw$=K#CSEM!9p}@q+^x$Kfz(t@D0rjsOy<4}Q=z0;(${aU
zy4TD8*<4!7u3axp1rr~ntEKxhOdF2DB!9LakZ49?Z;E&l^eU378BwiJSoGl6ITj?e
z5tyv5X<E=If!*}?Yr3Gi#5P~CFkb1{PsyBj-~f1#5cht|X1FAbdTW~H?iJU!L(#~i
zJL_3ARD3BV;?=Q)v`N9kBt?E&E#p|)fpJv9&OHlo)b7`I6HHfIG%Ebnh~HctCKcGU
zz3wpwR`demeT0EnmYr0e><~ZNm0=2dP^A9GvFe!9K2N5mGy9^%wa{hCo4bSR?ud*Z
z=dFC-kwQJeen4ROfgcQLl&bl@eEle0I?Am}x~@Bt$1M9}_lK87UE}sNTKm56RwnOl
zX0R5zmbuv*T9((Ux!6)W=kIm_v>4rQWk$<l3YrzaqOv=gDD&-BjrQ?T%cV}~df+RS
zV6BAwn-W9se|TK<9n5?f`oYl18U<opB6p{{ZI?xkj0-<io(N#t_1QmKbnuk<kdi7&
z%mAu8lC7*0t4Re2%5p+@`GWg@o&zJGM1Jx}B26Sg-T28MW2euAA;R1%R7gs?Ifd^{
zta7X>v-$M-V}L19@NA<930eXX)=?QCC$M5p+rg<Duj!46D^#*OOTE*fzFxHGE%*6U
zGC4CRRjmj!`VPG>Vz4#&L__lKY|Hu}Y6*ezmxD2`Do7%^>thtSP*&|FtD3!~lB}V#
z<t3hW`8&s^HQ(NhceV6-aTyY0QG$WLgV{)~mDFJpYnX*d&YVy#b&QPlqUKvt($VL$
z(dUJ}b1G^oY4dpj_#+D8?Kw>%H_*HYU?3Nb7r3pJ6)_(?*s;xI#!z_P3eFg?Rlvif
z84kfh;+eTj3iJqYSo6(7^+J?`;<%i=CD*XaS);*ck4lNMm@)d8$AWLg7PJKdtnG~P
zBd=F$4i7%T|M!@F=Is^)@4)JTPS2tDPNvtt^9C)g)c=4{?>Z;ou}j_6I~KeRjTf(i
zjuib8dLKDQrXF=RWe^y#KkQ-oimY0d*IeUmGw0Pv=0ZxAe^)4x)2%rEI+oJ6ArQgZ
z*g2-|{24Lyf>M<wA_px!CLx8C5G|SCW?{=baIP`Ui`A4!ROVRDaU9~|sxD#B?+OQ0
zJ45k7F_id9wDen)Y`YBOhm|I4-|!Q8DQ51pkGJwWHJZYS*W|?<9&S6F*rE4o(9%0K
zRl+7qgAdU#A%FkFq^Z3NU|V|a0=x|lKtHKLMXbuZ_Apuq1*>TK#yETLJ5QcU2l+Bq
zQT?urogctj@us(`x$;b|4u_U2@cU`wk1$wm(!@xsHjnu^2&Zl=pssZ5BD<%^R%mAW
zbGg%M6E=d0C1PK@y2O@eH{CphlR|CcZup7Tj+NuJ-xDp(t|?J<Dl|86=2jP!L-x_!
z26_xYUuWgL96&z~M)M#Piq{07jEgAgj`J+R`nlh`(@bVm{tTlbbC0yv0Lw3t8mSGr
z0ofSXg~Af;@doc9PwHuFcH|)f#?AWaR3OGxZPXna&fCqx+b~&k7%7l%G9y*_utG~g
z%S2XP$Yeq~nS(Cn36rwT6Oa|3E=e)A-zT#yD`IV5<auKx@U{`&Xa9YAC9{1fN@xOR
zClPqOUi+o~XpUKfjp*RB&-@|<0@D>;r@v+5nBuZX@kG<JI=cJ(i6Y<U@&~Vv0r=g@
zzHdn$ZB)rwBV2y~soe$4Rl5R65K9rKWG{~P4^A_A77vVGh&mM=vGa?m|B%<VE%bUG
zEmsEp$hgU{KiRZ+HL)+Xl!*h}G4>y;@2^}2-fhO85<rIbT}{YY;GajEj+E74sMj33
zXUEj)qBxRwOcoCowW;Q-Tt^)$P)MtlLlW$eN#Xk%-;$?JBsybXwM2QIOGxGH@O|}&
zz8eC$ANcBlhLjA`*`Q8`k9Mtl!TgSlc`~Lu4LIxgrm<d~Wz4BG>$UQGzB%Jmm*ti4
zEqe+j{M$8o01+0c5t2GvJ-c0Ykl#=f`@gaG-ce2cd%7?R0!mX*kggO_X-XAoD$+zi
zL~4MjbRtBg1qlQ}dXp|71nDgxHS|bVdJVmV-bttdLVoW)Gw00A`JH=b=8t#Hy7#X0
zN3vG3l3l)g?{9fNPf<_1{jFZFhF>IavpXG6$9zwl)htpGSmJv6Y*el-YaOM|?(ggF
zXext_l=KLLQT|Hz@IlSve0K2rdBh<tESz7ABJ?_W4CRp%1v4j&jgmdJDq)}JnHbFA
zuZXGh+R!%l#1p3|&nzO=T)d}_aQNNvT0{|BvtW6mnwBJlzlWOEp6l2061TasJIh9c
zZOcRD;OwESk%|;<9c`H<=)1Jt@mn(JOSC2#B8(m9g6XUIg_1ikMD00m{dPmLcmp09
zw4-1GMgJZxma67C5zVBl`jFCo?q}0{LCRv~!3d`LB7BQe>1bp-Ie$lH^cCjtl#~)r
zN}hzkj(&2pv?w!?90^E6B&e~17yq^R8`zY<J#~1RUH-#v)Vg(xH+fLw8i@1PZQrd+
zl`9Ve%bm0AW&}0r^*nL3;*GN2fg3MNt1m0Z^0lJa-(JEIEgm=BIG`9E@c<xD&V0z|
z$LZP66~ZV7{1Y*^I~8LV^Ka-1n|YqprC4T2ERwRXhF=zD>6oRK(&aQ+s|tYt=%X`a
zV52smc8tFx|ACq&tMxB>E5mcCXe;{54tzyxYxum5&dyhp2x;nxy{ZPyVrzM)sq8)H
zWUa-;B2umE@y7gHOu!~t5+<%SD(&dfeBv1S$*bIScw1VoV&iy+Het$aO1@J?^;&%+
zNZxVpgG|u*B9g(dw6=IFZ$(3USfTmq$Mh<E=i|Jw#^0#%vbwkvTojFe|02f;NQDTR
z>!T;yfEiZFSNdI%j9ucXw_+`GR2VXU$IEv=`%Y*5YK8q+JsVsPT$Y22(n7WU`;&>o
zr0?({Q?=zC06Q^Qj#2VPw$#acy$unA)|~)80gK#D6@O3!o_YAZJFa_01o6X-2s@wk
zJ1i7~S`r#rDn=c*G6~kiG=?W8(6{!3D?NLD0qB=E#9mKn?w(jC6QHdOM)O837^a6>
z1?b@YzMw-3A*^)lf>#NheHFxo4yk;`AOHzLH}Stq;J;LGHV1-u`%bVXpTzg3Pq-C=
zrzTNZM`{=LtOK+pT6@CFDEYr+&9OoCjX4sVdCx=GIJl0;m+)KrC-e~3Si#<Ss)$^^
z1&7I)piWAMxmy%J_p^o;opmju&>g^Nqh1yL?1W#(x;NWV`7G>=RCe5XvVlO7c{hNu
z2d;^0kZlK!ZofL6=#fc`yzOfK)|+c)dMZhEdQ!T)eNe>0O!!AEM^eb$57l!d=>-fI
zY@uoPf`a%P6gpG{uFBZpsZ^2&tKE5&Vwy04N7X~TZK+}@m*jhvo?ez{8|J$T%^-`y
z=@L7zCl>-^ZSlL48Ut^}t?lQds2FTFpHp1IR@Z&Han9h;i=geu_xH^i6Y<JT{cy=4
zg^NY2RTg6X?|j{ZS`EL<(1jdHR_yZM>bsvh{reJy$a_G4l%gLSV`$_wgGPH7H!cln
zEL^W1mqoab?o}8q*pQ*cC7tbw^HEGBen2KDI9yZ-X0dO4P#&g}*7Yh@-|n@e`Y$`4
zU5(5b6aLB=n1$M-?rZ26t@|fxHJso8AD*EDCI&4_Tb`60*atc4`0ryaS6NwO9+t7k
z2s{qDQXecgK~=K5xLmt#bY!}RnM+(lAV1X3b{#-<1-%@J#(+#CI^aJP4S9wCplFf>
z3_ZPo>r$SOm=8BDBd@PM>{&#eWfqJ%9t+ikm-Gv?p`TZ~Tg3G20013kF#&L2`Z<a`
zxw(;F44OxiVrH-ZK{3X$_XovOMG6<*awsVl#GC!G(o=U6m|h5AJJu{v|4^#rDnhK^
z{DZ<75&se|L<%PJ0E4C*h8${I_gBfnpRpic*?-rV{LcyQ{=CosN`UuJ9&up1VSymf
zm=4o?R=R2N+tb$m`>J~)oy(QMw$0tvDoup|mG0=$*3VRKcb)gFn1%aA|3!ue>ucMg
zh+?+)k@DB;&#Xq*2YQN1|7PxKPipMFo%Y%$o}JxPwAHVkFSUisOwKb=skQSEUYC{x
zbuL<Kyj3wIEnmmEx{VSO)7ui6nQDIvGoo=VbE4@zY@az^=>3Djj+YG<)_gR>p#?f~
zX=%J%eX0+!;KR>^*JdB26m%TAUb$|2kuxE#B9Mk|$@iI4&pV=W(~F6Xgh>TnDDI4B
z=Hj<WPu6_SDXFU77^;A;PT+F?YTw7NI&`Gr+H;N;<|<Dy_{8AE772UzK@$cUh>A;g
zALZsxFRK_n9%WHDTHXA6s>bUv_TP~0tus$D#;2{Ldk#)Vcx`TH!IH4kahcdYYCw>x
z)k!KMI>{sU$FNw>&mTi2&Rq>#(%^b@bvu&(GyUqlgT9I8+1A;MB+UY%mnwlNppq_%
z+CfPWznzUT1<`J|)!!&r3lj-)g^HM%=&4Y~&(x2CEKe1`9yrRI){VwELdndT^9|4J
zM8aQ~Yo8$nL))&Z8Q#l_$dnVkDah!9pA7Lly?a)In++t1z;)55yiFf-KP!!#A(8T1
z(V&jQh%{(d`LDjj*gB7y>&t1S-oOH;3bAVlj)<_^Rm<|M0t%X+l+UWB`yLZP1WO&_
zcZTtSh1y{e*MrE5l}k@qWsM$Pk$NGsHF$mO&A3q){j)%k;=&oDRJb%zk*yhp(D$Z!
z5joU65m0fVUjEuw4aXh5CyL$loL7p1^BYC8oF2{j_0<hHEYD9boFRU0EkQP15|U#g
zvd<9a@Eu1H09vrI1Kue9J1^BgYwlhZE008>%>T{qz>ZR<UUz<_MVc5w=)(;(gn`4<
z$_FY-lulJf#4B2^mfS)vWlouDaxyeo-DLkUd9i7-qe|}yF?I{lCVdjm{$LZ1IaPzp
z48JK4dyd~uDux&W`~gN1D7QnTY}&!4Vowkm%hD&TDG(!PbcsDoW{}#sBN^6$>kC^X
zelOATc&Fc*C|mV`2r-SqjVzbujZYk3z8yJg1p4K;X<%j2(juE6F(~O7CI;HBy_+e-
ziaa35d7qrke3x?>e3cLaCZKV-2bL=sYU29QoKr|KQOf}D5h@+wCLN)3s!47c5GxUN
zv)kUeRqcCL{9uC4rh5i-mOVfAs~h2NFgZm2yY;^j#LTF<?^uw`^cRF1Ne+I7GzZ9M
z9T=V87q_>tJ3D9cZrkGne<`jOET%H(Cx7Co0;*X%Y^RyWhDPBe%PhDOzBqSY@#9j_
zzPAjEVmF~lU^7@gc)`$JVgH5ZJ_>{&9)+C^kP3feAOX=Bn_1pqbIY%5t^69<aWV%@
zH?Bmdjq~bh&P>qcj}dw2-J=MXv?>NSNh%JGPm+|;3etINH@16aIIH=XUFF8VIe)V<
zQpty~a){j5%CndpQ<WN+R6CL=a@hUV#2?PeFVg)afGg(FwFU51M>r%M&Pi~}ou3Wn
zFh`M7xjqKQ=fiB(>jvu!gHzzn+UyP0khp}$TsBuCaO3aW0u}UIS<B%dqIwQ2rTrwb
zE(eygGcOizpHK7|Z8fTjwqumOegkyX*gC>PM3uLfQfxQm#az3ps*Cm$#d!?=j^N3O
zsMvwg%cIxJr&@p#A`FpRII7L5G?~#TNr*#SA$by*<^v<r!INuu=He?W2AnFHf|C=z
zoVcoW%Egf9CxlI6j(PNNS!_)kKU%AJtL78i`eFl}ILj+Nt<uJ^_%SG8|I=BLea;4(
z;h_wgDWkCKZfj)0u}jTU-_LIvp45n%yvoZTa(gI<h%avRKJ`p)W;D2_RHB=PTsaax
z#Ee|iHy%MrG4P$f!X*AGE&H_XBsc-Gvdd2TaZZSW#a9x1mzPBg&)aG$mYYw^ZA)H`
zFMvNBKA85{N=}*4IFflg(AHL=FB?x4j%Hzq4xSgXF<R<znXt&7S7soVw>PpjdZFZ>
z50g389}Wz-F>zVIlGkMnUj=$<cOo+CM9K;UB!078olE_cT*@_^DeM&2AkO0jpqq@-
z?vo&G;a7{X6FSN+OQf2_1IU?Y4y@}r)@grLq;dSgun2W7^h2%p&nml%W;P$vOlxLn
zBfiS9%fG+K35aR~-GYsb2P?KbKn+6*g1pUjkjYsg{p3#h_gojnsG=WM=v?NVxkFDe
z7rNJ=K}^K;{rrQ1jr0OfM$K!{!qq);$5*jFcL|1>&a4qxT5eah5t0^)9nm7ub=|#|
zP(gajH^Mi(bFXF%yVrq_71B#44o_pfMJ#o6oVB;Ni0u;Y`<TVr%q`PB-RnLBjg)xa
z;7SHg(6BUMTOsviB4Fwx<9EdfzfMU`GYfUf$Cti*Fx9+zd-02K3h^S2my@X6Hp7P=
zEiT&HzqpCAsT#oGq=yysT&PZ$n3R#;P_EJrzJfgM?v(FuziYw?X#)Llmj!@u!3DpA
z^HOF~$$=y!J0TG$hoilaqL-js>M#_)FRF>Q`;k5FxHLAj-6NvL4VpVUnp<LN{NeX$
zZx?3z2>C*tPn$!NvnaZnz9{#Gf4~2lPDyNK@B!w$a{HN!+KBX);RbFymdb~tA3(H<
zFT{Nii((f!{Dl+xeXdHQ==&$@b6+c6IdoEJd`Td@%L{x-1W9$ikEoTKAgXn$+sej~
z+L&+kPOuuX7^-fUJ}L1?1tg+DHF-bqh18(#^hZSM)-mK{e8PK)YO&HDQv={Oit?Cv
za&$)H9lQ<70}2qml`o@|_+U|HgtpPJ*8?##3!pSFP^Gu_Tw2?v|FMWAT2lGz@8yyh
zuHTiga$I3lD3YaK6GbcrD)8C^J5orFuIyv9kJ7McMroxV@oCkE)!#ObR-Wli4E#x4
zQ<;Ui+s)d_GXkSvGQcE>1xKEkcTEHMD;V+zMKoe$>yQ;yIJWxGu{RR{c-JPpi)>7`
zx0j4FLTC&<-k$%ot_`B$=Z^+oBR#{166+2>5>MxKFT<i*N71YvCkNSCl&wbC!-T5-
zDLD;NaE|fRc_uN`E(!VGvABFbp`n0uail~+K7IY9r<pj~g11j&BM*;=RkBuMSr<VG
z%zgZrZghX=hw-+;Z-nNEC0gnF82gVMN#Yi1wG3-AX}>FGGd1hnK^Rfpg4+eAJPEZt
zVQ+Fp+j=I(Cp(<P#BbT>Bsv643xsRQb6J%N#Ar&p=a0+ty`9voe4P<`hyKrV!T<PK
zIDv=)=DO&1<wI(~dVhe);@lq;RI`N9wjEvorLX`%@UGLs^bOe<^kHP-8}wq@vIOQ)
zEbFucZv3N`EQSP73SvMDyr+~Fh}EEUM$1D-;0uW5eu-hpowF3gV)EFT@JQaDZFv68
zQm2TC!GR+u3O9YX8A`vdEse^0A+Ij_{blDTP4;ij(I5g~R1|d9MmAtL3jSVR>*Y*!
z#C%Kk`0@L!@K1a9R;YLsAsYJe>UP3Oy;=5!J!Af4W?*S>!YU0uJq3h19jgH0A`~b=
z2SgZ=$rA{|i8Ai4?g`BdVo{td7E5OS)p4sC&$x1Ad#ngOFrDo>zV7W1wdeB|U^A!!
ztb-@*g@EO*qcZV~7at8!9QsQO{-Bth)Fmr}?mbVu+#%vvD|huFI*<Rx&_<#o`$%IR
zfO`LVhkO{ohSCxs31F&_Bp>wM%Hm9@x8mkK%|<P`&n(4*GI9|Dl+k@%d|Z)o2h?A>
z`-mDPK)Kj>1hKP~3phDt{^^`H{?*Lkzu3*c7()DWLw|qm|8yz-*+&2M?fTC)`oFV{
z7`>VgCD+o}2Sv8^yK2zofp_$8Y*{~|{WZ5MuV-3@D1xG{o>Rn}lphoe&5Knh2Yn`4
z<{+$Q%#B+t6;WHO^_3NV+}Nf`*K!RRiWp(Gu_l*qTvyC1|MHKLi^LuU^b9vK={pD(
zbb9OXgfl)!+kT7B{ApM2V-EhL{x1RXi611qmygTNvC-yh-r5X_wU>$BH00q+STKmv
ztsmH_NC|lpGcR6QK`-#i%BU&SsEPXNMr6IKbRYaFF#>^$uvEeC1XXB|)aOxFy3&=Y
zZ98LI-_f7NQ=Ao=1gyfPqJ@5Qmf9Q9g-6HYyfro&MTc-fZD^hwUS9I~Jg07%zd4eM
zetkZ^fA_*0&PXG7cKlCY!;)5}iyJ4rQT5yl(@a{`g!q}LoEf|10KcgwsHZdZeZS^P
zt69gzI8BEnHM`-W@W-4_z9z&xJMplHoQ+8Q?mT`(#^|=LD##Nabl_5<F9|)LmCkuH
z&P}dMuCpNsV(r7$X>>Krf(gvrYP)$V)GyOR2UF&A{ZJ%v+j$c>IK2Zm*)wZ1GOO=v
zs#|G3X*F3T!zy=u4u2);5u{rIoK1Q{XiUWYLBYJP#PT_nrtQiGYa;_F_`*I`%TeH_
zO<cEQ@|tyAZ@LCQk14}x;CfRjQDxo{()v+07+Nn;>xf!u)tk?+kWboL^p==*O%i=t
zw&@<;Jyf9a^f^)D{KcEW&PS<fDTpo1@@%4L18eA+D|BVY9vw4UB0IbFP`Ry-k{J84
zY{(VC4nr6Ym!KHVvg0H3sfNrZ_fA&_%o3yg_k7i9l2;n($^AT6iNA5v91u}kHUHMp
zWUGhqqk-uoZ>#G|ja1)h_%l*WVeUU3cJA@q2_iGV!qk;mVfznI)|JF06fAHnBjFlf
zJI+lo^sNefGQrEa!rJ#WEvL>sB|6f5LL3hVP(y1_!A+&nAkKG=)<&Wfm%^hhBu)A#
z+nhwRqsOy6TYgy?cl`2n2~9aI^L8brAlwG`suCJPW0q=l4l#2RBuBT&Byio4Fe1W&
zv2<|X1pat|_o6-9$v4*BHu{ew36=XP@1uT2lgcG(G|%6ATxa?EJwzW&%%8(tf(3Rw
z!ji!0t0!BBvt{-l19Pd^WH}>8cIVycj}_=Cd*Hu;t292_hKJuHP^Wt1@^s3Tn}xxb
zCJLuObqG80$>#~Uqc54R2`lbTSp4gj9(Z)j1}a~Ln`FTLplF$p4LX4GTMRA)&ge2m
zPz!56d2G$~PFfQxYc7@X)XLD8g8#~!CaU5xEo8i-$B}vVvBk*_`sI|7bSxRg_6G%8
z8oJ%!<+Yu7`UuAK606{uiwWMJRv_M5mUpz7KRKcO?hn_%b$KH1o?n&#L;E;O@Ys=#
zwkRqSqG4a_D31i5#KYMjV7kHV@J007qVWBT_q`(mBKl}D?r%bxNh<ihc6T83t8&_j
z>5>`bqg!t1{wzzAL!x_Va+cnjuob>z`B|N9sNTx$I04(YZ=at{euOqdudX)=4%Dl=
z;rcdHy!wJM_&DthU1z6V`-HA6P5Wi7(g{i7URE{ZB)l~hOQ7`SiXvn@5S9QUe*o+b
z;rs0;V4#Q~uQ`bad!tRy1d7;I?oeI%qW;nJcGkWB!pHIaKQK)G7p_dMe+8GWwZTPz
zSt=MH6P!{4-1ODY#n|eS0OYO~i5GAvgD9U}jyv}szz|<)U{bp<BYvd5TTr|0p99MT
zum!i&heJk#CH(iz#%3RtH$CJjSp8W@xvF+|<Zo08HHi0{Y?o_ip-+vsjLO_^xKkas
z2+f{PIJq~r1Xsk<;DSC|N!WoI75FoL*@P~*Bfp#egpYl#cxlbgE5ZB>Z6FXGIyutF
zP1Luj!fqWX)7wo%G*(+aEHe5Cf0>hRX<*!qiwkiNcp<xT?h*IhmGR2H6WTREM0HLG
zAZKK5POO|!0bb1{ideuJU4PodZ(#bm=}r?OXy$hHpb!>*5tnM=%_9(`9din}G#6Pu
zxg;R?<=7#M{=(1e*ZQ?^5j<`qpOdx2{f*3%gCSPRj^=3=w-%24U(@Z`E-hqB{{0U0
z4jo;{eh(333Qb_%o%gX&dHvLH=oMgZG$2+inwz3Fv;=s`Yh4$N8DY;r$3l}X#YmJ^
zwRT1_w*8tj6bd9W5^0*MVoX<gXz<HH5CQ#1__bQ?hFYl)n|m(`Zd?F&MZI60WRv@e
zDk+B`_pb~Al7RENWJUEqop!amd1!J;3+57}N@Co>;`+b=o!q-b?U<SM<~x>N2CuuE
z4C4GR-?)r)z4FaaJ-~+=E{hiiqXQdxVMZ-I2q$c!W42utwtbledq<&yv;Rmx+MMlA
z`}otD&z_LcuWgC^IHW(BePU{YsN9^)8z5TzO;h0^Pbx$5v$@xj!SO1WAN)oYI`Xym
zl6OyLI%nu6r;!ddk%A^dV=_61l%f^yCHq|qFgBygtNg&Ue~t=C`WW?hr>b`P$zfFE
zM`=SteN`D6UDs&sRT!LM5QB?<awpoi>vH;Sp$8O~n(!sffN-46XxI(RKvaH&Kg`c(
zhC(Y)5_I<4A#Xc+TwnC<G5fsrltbzM-RA$?lD6-6ALs@smv?+tx<0XHgc$pHD|>$9
zdR7pb8Mzt!78A|=ZEMfRp5oHv{d&CerL__=3yk*_As!R91Lwq@_OXJtcHK52#Aq7B
z?3%Z}<q&+|--DsR|JgpX;!#d1G>H2fxCm~kX%!vwTWzjqk8`}b^xmZzbZH@_etm68
z{5_>1U*?iLha(C)LKk%D%t4LO*)ZW|AFC6#fC_JJN9LJrjkH@yi_BN8FN#d+@kLQT
z(R_QsL-{{xT5Wsu!Q}SZ7*8F*kel7vm>S6_YwRgT^sDOvebYr{;sNj9{o7aCuIN9#
z57QY;g@ylgpZxt3LTMP<c`5@~^*-iS2K>PVfG%d(JH-sJI*ddDEGK3JVV&atlB0n9
z8AUKf{#|R;U_!pE=i^7)mX{BEjS;0K9s-7u;fVLd7IqNV=?iN0e=oRM=%2~}rB><E
z6QjcaJ>lAnH_DKZHQW>DtR>Oyj%q73Z|0hqmx;OHDv|Na^6jU%Z{AC%E%kr(-YsV8
zK<}-UIk+@2>lnU-=NDUGh2B8Ejx4L%ggno<XKoPOHw~@z!n@b+kV^K{W0Me!yz`)~
zQG^&-Jq^aIODN)HkjhhdobR33c6q|IC2yUBF}W(*Y${5!$1Py)>jw^n>IllXU&e_?
zkkwl5k^XV(_tw=icr}K_^=8e<w@WEk1Uv=4ZT{@TmpoohV<s4)3rXUGB_xw{R3lmt
zl|NiDed~lX?{&S>n=b<BVz%p7tSK9t$Be%C5E|%O3azDP9U?xn4S`yRjDH`xNKs_G
zF>Cww<^DhtqSiU|Ab&F2^bZQjwy?y69Y?g6r@JU|Ve16>$yJ<jt=~+lb}n$xD>z0w
z;gD>xmb7S`^=hhcDPn>BiyOtrPC(<0k$M&U1Xlaw;CdgZDq?d-X#^rdZO?xUUu3>p
zs($^wGs$?)j~d)sxCYo51sF0}e(x!9*j1;02p6_G(o=a)xMq^`l;%i-j#-00_rqO9
zmPf-C7{8ARL32sVRRo6YtS{s1JU7Y{1Zx{;d3)lyWxuAWR0f-EO|e|pw@`FA(g~<?
zoJGx&=)7lahX&H)EWs&;D1be1+}O%n1|}tsHuIidttdGYh^*7Sg!tvs4nAblJ-bm+
zOn3Mg-xpTTc`5^=R3%gga{F{?`O#FM%|x^AaAe&}lQ2^YsgG&tmT;7J|1?9?C#$4M
z0Ob!ru6-ffnOjI2HFhi8+nK^d0h{{u4~nedvZs-+7#R^GG3x0nH=>Q*8(D@H1kI(X
zB2`)w=}qG1&LzNzy%Y~D6BKu~9oHqbAo=E>sI#tj-B2=2ap8+ze)gD&=iK$!S+ujs
zpt+f+<kPC!mW%Ah2}9zb?JBkDrK-Kv+m!uq7JjX{=?h~^$<0%-+m@yiMF{F9(Gvgt
z2R0GkljUizEJQym`w{V&isv)OSo-<U+yA4lOCSDaVW>$iZLD}fb@(re*)jzWhNaO@
zBo#)avtTwU*{vBg^+Z7YYTzx}PJa4h8me~}s^?ffnf~jq%i&<A2?Zfk+ma{QI(%cv
z)5nI?-omL+<6%406{{#CJAH%WAaX`*ke~1tW?gd4t)mohjbXpeiZ-EohbyhXB|T>U
zHT9tBYPmV2Yx&sJ=^Jr&X=hl#d~DC%_A;04JNAtz37tmM<Fsq1ueA$^YD{#*jz=cS
zRg#`KjTzC!LI-Est+4=_8|H2H!yX$OVkaQ*NCPv_i=%e$2<&(`p*IsyXQSi_MR_$M
zVGh7?CAA3_sU_LVpYk1?R}T}NC!fP7bCH{C*R@mHt!_B+I}KsveH?ZQ7lHa8>^98k
zFub44W-@735?Ik5)WN#h6s}}uNU42!dy$nTAcclTNa3(hHbN{Y!d4ld=+Lroo;dz4
z6Yp_YG^077G?ia>D(?~ZT1M{c^fAbP<tT|0qS`t2Hto#EISO}cXR&T|&oOIpe|Np{
z4uE7%=EG+ljfRs~!lX(VX6&kJ<44DHu~0<$rhb?EUhZyIqRp$z7y${cDw?kFMc?;I
zf^aK<TK~W?i=dXNh@9Z7u8x#9S6V*Qy>)>}q3Da+W@&p-cB-#lwbichQWT3Ru|*G{
zi-THByU{Z;bwm`(cx@1tqE}P>qe6qlIP#i?$F&GO$>!UlpAF>=+;@Sji=ut#b)wPp
z5hDYF_Ds1Y{;2)ELn(w~YrMu^?~`La(TUH5RU||(_I4hh)oxWuLiB}zBIw@nhdm!o
z<yUTZA(Lvc8@WBt^nm5rYF|GcBzQ-3K8<zuY92Qym9@q%h0|B@1WYTGT`OlAHtfK2
z4=csXy@r*s#chAGF3@&LPJ5SOeSf^L09-!JZW(Q*DRVcfTJApS2{Ca23=k%<?-uK^
zjL~!6kBAEfR@DjC)dSn`;GTTt-GR$p1rODGXG@;|DdEM2;m02l;7}t+*K*L<V$RZo
zJEEYq-7}D;2ga(e{B|+hoOe|@Q*#NE_B{J`?x=YB!-$nhA^WsDPTV>70-eHmla>m?
zFTA3lzqFkfX9fzJzM33UI|g9ue@9gLUvcICd;IvnEqIEBakiuwfJ^-su1X)kRS|Tr
z{D<{RI%B9#V1AF0z^cQy;d6!1zmc=wz#kK^iTI(W31o;{LgFUt?du&*Da*-fucKQy
z7M-G}qpy9C=)T|t@Lf9U!P`+w7LZd_BEK<Vv;{2Ga!8nkd^9Y~g;XU-JmX`Fvh}>y
z7u`^;e<L{E=I4b#4R$G{)@pIdVoBk#s2ZkZ_c)da>flj`(MK$RFRc17;YY^>YQ07X
z$`i%!{Ct0!w>oBqDu`|MaP1b9&9Q>NLCoN`M1(EwOvN18VZq`5bf}oz<dv^kIEku6
z_3>86JUM--o8+T*gVYiU>4RGnOlRjY%?OUwWbpt62(-u?=0K`UL*XDBymwWcI&ba?
zhFB&)*}N*kbW4|Cc+ZHAApi4ZH0VIwFS#MRBTPXKr_8j9mSJO<YhY%0cBEaO!0CJy
zWiOcB6?cb=z!u58YQ=uHcscB}N0GUzv3`~mQL#SST@}`d2@*e^wdg*fF(Wq2f$520
z9fDZt0+<nSaU8Dl4l`~Xc^%lh@gYI?uFd<X_zQu9AI?el5l0WBi7tSC#Tf5|M8V26
zDwRY8*3jblAOTUfxLuK@<i}r;cB*bubhV5J^Ub6BM6YzBSO>tWRU9G-6-+@M#vg?$
z<LE&<ms>nUn}0<MX1S}1WTd8fvS*dIUwd`zG+v<0uy5}16T#?n0Sr-MCy40T`3Pep
zzrT|TtZ66<bsNt7w1ln*CuMbAp1$Py=mGyle^Wtku-hLL3>%2%35Vd=vY~PSnmQ6!
zHfPE$Q4)g5>k?AYpT+7;zE8rc))an8u~*)HNjA7qwdh-Qc->HrG6VUni&#Ps0|&V$
zQ?YWqz*g>DUJuD!b9XU2OPDU^kQ40Q+jQZ2`@l+|K%SZ6)mz9in2{unfigoHuM(<i
z2uh+)jyjkj_V$@Xhe2<i1MS@@lW7JG``)+43&sYY6E*of=|j&Szdqw8jMNJ`&#hL>
za>Brd1hr3jQ^5jJ>I2)AcO0!~cLh#G^>Tv?Z_r6uevrS}dB6^(?P|bim6)3ztX7#i
z%BfcZ9xck+$XufuR3@zD)u6(id<gFql$7`dv9!%0GtdTro;fz>?<LP0J9g41P)X1`
zu0N(OvO`=YvCShGfeO&$Oc>_zNCj^N3VE0vnM;Lj^m>@lwp@5^#$@2hia>3+Rp@bG
zpEe=2J>t8=Xst(BJ>W&r!Xt04);ZKF!`xM3X0K(n8uQ!nu4s4j#EGuNeLl`04TI)i
zO>|&lK*l=eGVi>4WW6NL(b?5Lwnc(YIQ5nnw~9fE_Hgq0y4{wL#<ED&62EpQhYHUK
z)deAKv676T#neiKmu?2cdrM~ROxd$gT>EtL`#$;m0t1akb)DZSXs9T+T*Im+a(?^4
z#MPwk_vL;*di;fVV~>(F*!gO<R2Fhu9Qj*hc=#KM0XPs4+{j$d`mwUT(vds6J6OXt
z^5rRaT;ch10*-QP;0GkN%3-oN1bMAw6V5)blD=*V{=8W<`D#GSkyp>y4U||jNtIWY
z7Xe;o3?1VnltxbukPKV^SA)L8&_;p{2G`?af}cJLK<%gU>^<(^c7txN@mD8pqb4|w
z7h~=}@UmglxAaCr?%hpqsT}~o3;NlGj6W!3t=6)4Kp+rd2{Y|Z21~C`9JN8Ih(<35
z_5s(AhtyGsqAhZFC1!Ze546>K_xRE#eUeDJBaL7X3IfY)P2gd|FXt9J#CjujhER*O
z)v@e1uV`^Gu0+zfw81|uK+jodF6@LRDBZyj6SG!%Bm)PePQ__qhC})lbwlLVFW#Y=
zjK~kzGW8oN!UEB13UliptlsSNb~v(bDN${>S6wk%k6AmcIufO{(cGR#C&*ca2e5m%
zMkG>LhrM>o75*S+F^g?9ih5~qJZy*=tgPJ_A|3V)D@nn=S1M9APpE&|+JPhT>~oD9
z2X~}np>HxL*ltI-xc`v%jQp9=!6lxcDR(KQ7vt9`sh+1j&n?`ZAk#=o5DMxR%`@%E
zD>`(f>h6NNT<8+<v3hyUsw?XqH~X`wI5cm*!z{j}QOX?tEJ|ktt~rNfhxs)BTIb;!
z${AiPOBml4449aiwy+Jgs(lb6^%`}3y5fQK5|T$MWrqhQyFk+lDPdS=onRmbWD`BK
z2HhYo_8$3d45wNmeN=T87ov+5Doy-kE|=f+%Zh>LS7DOLh7E~x3AGhdRpWMth%Tu7
z#Lkk(xLv~OnE_5Wn9M?0!sa^SO0L>9@>LHm?TX_D-DsWC2A}^Jza0^x*7>95H7R$n
zzz)e)e{tSslueyb669vsAD?2#`ml(F8yGFx7a+*n3{k1XcI>|NUSF#y3y3(0+HiqO
z<H?mcb^zcCnkWex)dN`N5o|qmHFiF~?V0;H*o@&JDcvtD0W)x*W1-W;D<}!zEy@+M
zN87ZuS~-+`E46P8sS?CLZ&lPm1e0lp$JOxm*!iWSr=JYjZ|P23@w}?g(Yf~llr>#9
znX*^r6Jxs7Q}j0bc3Z=AgAzCSXfJZK7W@K4dW#p_1gaW+?WPPT#1GUpSk^ARK#!pX
z0~^tYnJm|v?;rb~=F}+D!*z(U^OGP&ry*Q?05V`_!5&4OJ+H!8vJ_`Y*JW3Qy85QK
z*=UP?H7cgADY8Cl-u9=XN;@4+nbwX&uD__Hk#&QID`x9mvOGl2%-B1bJ;`XX<sz+q
zAM<M-$hZ;@2@1Y~uv-0y^oPGmhvf&2@4Ry&DSI5sn_Rxa&+th7P?P^y(@-<ywGpNB
zgv>XQskZ>&5eML}RHPErc2c-COideFxH~8+-gYwXSm2s<mVwG8vf;qNtQ^AAv&2Yq
zrOv*`UP!UWUM>$GVYhCz__HL*R?Sd}IlqL%&c5FuUMYl0OEX;Z)>6ef*=XzP`@P4E
zvku%%sD&NI3OJa1%=rbgA74^^&%=Oc#n_*!XyNWu2zNn)$ebhM_4c_Afy`Bjva-J`
zMdNEF+oVq70uZ0CGdY_(HQEr92=ch4%`{f5V+F*dUL|L?qtJd?dW`=vjZ<c4iv9gk
z{*ON%tX%t+;;bJ!Ju_n}W>hz`)`*4F80sYoe-stRo@`7Sk56XSdPhBu?ZEj6o5Doj
z`J>u{#LC25rx;>8*b>X3a(l8))05X%q_xOw9W5bJv&`U0*iMTPmdyF&q^1%pnsrKj
z(M&g}^=VmI1aa{SSV9+%p0`Ga)O+JkwX23%6~%Ge*_kk{w&z7Zw5j%<zS~Nh!;8JS
zL}1o=*1?5rb;iOu@s2q)1@`FZb!C=fw^WwqBUAIn#Q9meiPb~X56^xSnu?U#A9l$x
zFgsq+60z(Q{dVn^z-;Mq;9>0EjQh~<+d1_|J9^-TrL<A4E;}$-W^puUx*@Rz%&j23
z(C5#_dxdnz16pkFY#-ymYa6(J3%01CfUQc(v^8Eb3$XruSNCH~%R=I3Q4&o$F{PD!
zb3lQYCp&Axp^p(&&_aujbbDFjdCA(N>gO0m^hTQ})R&~|>3Mp8HSHpp7FNapAzub!
zeaNHV%|Pmo9*(wGb)A+CUiR*YZ+B8_D%4ymois}fIDqg1IrP`aY_qVa=H!uSED~Qf
z*x!@8>cJ5)lX5>n+7$xh_Z)S{zU=&JUj0_qa7511;(geWQxVw!g!bA6gh(r&%^Kna
z1yQM^y~(3PNx)F<&6Wq(Ee~p&xiIKToEs@pFkl#IYs`0a<q(X#ZD`;+K^txMpoyMZ
zZbdq14l%*f^6;qnOt?Mm)^G)q3It9mQZ00dRtx=MrXVt7m2!^vZs|hAoEK7rs63~L
zP#saaVs1usKUFJ}B?}F=e<LtUd4Ya&K(y?umfpVMEZFS5_DS!X0k6L@<AL$(GvuTZ
zetQ3<orjV@b;?%4%R8kp>TyqZY7p)wW?Mq+3~jc;?b_l*M-1Q4Xzl3A%AS({5cxig
z&JgSB&DYUR=4u$JOo3L}C&-*s*2m614pe$~C1{%O4cGPywJSAq!viEk{PQ#!D78?y
z6b->I933>O@4HU9xF?66UiR)w+1_8thVNvPL~91f*MmjiLW-$JBr$62(-&*qF|-J@
z?#YXDA6PcW&gK@ncB%l3ksV;44Yiv-m&GcH1fFUU!|Z`Un)WajC*`kNHMB5H5%t3B
zSQHo@U>Y6}ujnYLzz_pL>1^rVxvDy?K~m%&6k<@M=ze2l_S)XTlJ4?QGRafPtjk;c
zfVGkk_S1rsz?H^3LW1IY#KDog6-x15g`gJKQpf4~*52!4{W{L!E{65m1o@^JI;>O!
zf2)#6|30^o!Ruu2$lQDtsNw?aB`S?g4wV|;X8+%m4e(C{H3U(OOhGch>xL*?L0P~G
zc3RQ3c@5^wHbdK3M^&532DV9`zS;+-SFa>dr8i%Bk(#~*SI08~vSXG;absAWF)<#J
zLG0RGv*h)M--i^Mtg|XxYA|HQ${;@+4hm%W2*zCh`Sszqr0Y-JmiE9t6XJ?s-LidM
zRh`b4x>kU#Q^}fD+L@8IZe9hLqw*-C5<@pT7NPv2EL48d&?G#>%VsI=SylmVWFUTF
z#mQ1W_=}OAg}$Yg&{?`v-n*sbd2t4!4oK@XL}$gewi+$7z7@_kX)F2dgY0=VZkP2D
z(+@cS<BEE8J_e!GDm<OLM9cyhX?s|l>eZLG3$xa-4-XS<^Vbp+P9jshE}2ctm_93O
zQPBAw^VCg1qU?(br5`Qd4<%v5Z&5<j*{2faeU;GwM3W-IR-A*}qkL{_A4p+}S<o9U
zv620aBk*b2rw1*e=D(abNec12H)FU9FFjgHoc8&T8|bIG|L^FlQ5?Akv#b5iCw_gH
z;!oyS=Q{uDaA72hf-U?E%b!y9&g<e{6nm4{73wA*_vP`OrY3Gk{@=pS{FENRpAJ`u
zM<|I7A#fu268{P>X}^W7H2w;E`0=(FG_lL*)a=r|sw7IPZ>3vC?42p~+)2!zZBT-)
z-u{65OGit|37xr*(JMqV6u;6~_@QMWVbz~#)Q+I9cg}~Uv?liiU>ksvEX!K|8rJCW
zV{Q=2^t#nvPJQT8fvM-55K~4KMy>M?ie(S8&eAKPzK(OJVRXWrU7|XKg+?~#x`)jv
z7k4{reaCFpY1i5y2L+>=J?mRF$;|^E6>L*c15&`s@9T(xo4bLYZ$|G63LqEQ9)Wug
z6DRwN186<4Kqp8amjkx>>2>(o4si-1Z%e6tZ9rhnp}lx<u91iF$`@KvMlprF#8lvE
zm-$Q>VY21Qs?1SxP{`|%?0QU1S%tZfTAij1EUGu;ox-?T`e`rq49Vt4!^kXhZV-`#
z28jzD<p5kgZ-Bx`=ih|bXGf5z5&?7R0#ih5$@w+3X=<zKRQE~1&zd;JqH6k^ks{*1
zhu%H&>=QVnX^(-s;iY`84k@Tf4~Qk^Pn|_}H$fE2`;;{9i(CqPIp-fg89h_WOG6xe
zN;nLyt@`AX91x~4z-C}&6|KY0eVtm3d&HN2Fyhh4ESZf4u$^C(04j|afpZ7KPsNx3
zoBDohpd!Z%q?$HbW_&{GAe!g!;4^CQ9~9>V|26tFplJ?sLXRPz*8&1<N4s5tyg($r
zS^+3$KL$?Yr{=$;&HjCh?r~o|x4-hI?MD&Kc<6vKNf%Hg69A%e;T@_J-g19xl*~UL
zCHVfpBi?@=)j!+nFHh<}@2mgApTRL^%Wk;q?~EATgNTESFDw8+S*Q%;EWnlR@M&na
z_?6=X$&&b`<CHy4`ggla(@0&h=lf%`u4|zG<_~=#112o6j9w;L*M}jyN895$W>xOT
zHSPTJm|>;+-<B0@d!jtzeEQQa)DlL6>u)Dul{g7SA6?lJ*#K3sw)Y}vJ$UA&D0WuN
zz}<K0QOvF9eMevBlAN$)Ovy(2XwufOQ4cRYng8+JVn0bTeHBgSfgrgW{R?4G^Hedl
zP+si>ys8zx=JC04mh`R*%(Ve#7oMAttQQb~w+;vVit~4E+?kMZWn%kBjoRUND>djy
z?5JLO0Qi??bXIhNRNs!?hwf>=)T^G!S_K&Lv+R?SZjmid`!nFeM%!YgjJnx)E8MWj
zu4i5r_?cLbyw#7gVaOM0z0{KIu7zQ+R7ZigrnhtXU4*HXMLdOUv>VtC(?^RcQW5$i
z^~(7UCqAh-FU;l2DWA(D``2$IUobpnk2$#1X=Nm6_5AhI)?9v;!Hr6?#Svh91xU&#
z8a{Ncfeg3yYPI1h;6dBLIHPVrdLa!CuSGO1czMr@Us@YI;wbjTo<s&%d$LFM4&R6b
z*@jOo^+TU`3WBBSgWpEo+*q?~DY6r=w!bGUQQWWlfm(-V-l4`+yD@Us4^groKJ@aL
zmhBk=9GRK--@w`bg1!F_V(prG|AwL)gD=C~vH^1r_uP4N>+I7Dq15*$bB=SOzY4r>
zui?z{uu?S(FF&VQ#~@tC&2Vjnhn3I~B@j1D9Bul@>rawC)D%lQfa@L>rrb!s$L7pw
zHy=e;=c2AqP*i;(afM@)mz~TzPDH#IwIu+Iy8)>U%nyt8mkx&;ju4lQZGkV1!?DZY
z=wRi03(1JlRngw|(8Ytku$c?^zJUE5<_#Wc{B}I9N6U7sGHlC`_T4u6j@53CE$#6O
zV@IJIEC2-|0%6Y^q|9^>MZI=18YUH{Ki{mA^5G=D?8$(qE?=)n7JJCyXCw2Ub?G}g
zy@wm<z8@voC(vC*iO}W$E{gn%*hFA5AbMok#n>earZp5oINBPc9U*xR;+E&@zegU$
z-%Rr5d`R2tqVHWl0hZZ1j-UMvUIYLO{`!pi+$Ic9(Y7*9sFbj23ifrwFwBT%pY#+7
zU>SKRU9NsH`}HMJpVsE)f*M$K_Z}rrkSP&|Uuk*2TJAtr#y7^ZqAyL^f3Vn)=|_>Y
zBg!6`bR_3~m@9rfrD(r{KJZqpD%!FuhPGi?6mHI4UyN-f3lAvBaXA<DPE=(w05O$^
z!m<4HcMx$rp9$Dh6iFSA420id3YA7&wW~D>%rTx3wK;te$SG@be<?oc>rIB}3jbcU
zJ7BGS#;S*Rh2ut)L}8pN{UVjq>g_(fKL*6pPYrUi-3lIaSGLMMURLLSVEFv^^{I=z
z4^Kju;Y_DGa0z0-JcEns0LeN-E8pHUX3g>3h^W?7XBWF+fZ-E!I?W}2U#-kyiM!}T
zaWK*P<h!iho?_I~feLN;mfClwKCy2gImS<oJ0P}~T9{eic0b9Zu>Km`ArL!L2rF!^
zHHxU<Ol32zJatO?@S=<6D`9(VePiz?)th)-no1Kic7LPc<=%5YGS=}~4}Z*ZuUvf2
zC#&hD@^rZulOU={_iwHx68Mg`K33SNMPp)%62p*KJ#X%YzFnee!uJxzs&tEs^(jej
z%RxqJs_#!#`Z(`BaStyx+(|qkS8@)xxRQ~pA4jfrIsjP(Oy7ykL(Lx(!|RoEi8*_(
ztNOPqHTW7+Ja5W8>g7gXcK+1+00k;xvon-A=|1AXrb4O1dh<oldIe)Xn_8#4I(%o7
z!lbcdW1b)>XaQr@Iz-QmguJ3~hbaVAFDDdI0VAonZgFvFfm_9v0Q=C+Qm2_#wa&$J
z%|D!ir1e`cto01@r#%)!PYB4MgE|x!Oa9|rhc9M|U#R?7?#kVtD`ZOHbv=1WS(s6?
zLq$>p%tld4UchP^hEyXMgfkyAPunGbHi}lib#APO^B#jABhv%ES`L)q93X7Hap^Z8
z5HWfNy!{|1Zej@XGT`x9hdo}hBB@W(W@_uSL$2MMU^=WfqP&4#N_RycO#Mx-GY4&?
zChEQd%I0A&%pv;qkz3FhAk~32(<N76{X>jscx;OG?Vl2F#iQd*-v7KZtEb8-X7Iyl
z`{=IyiDF>PsH3Gd)Mhj`B_#nJVk!Cj5w-fy<F5gF)q4Rv@lJn@%KisoJap}0Wn@G4
z`1(idrN+#hndU?pbV5wgk3N<>t7vXb`l77MKj-{ti<m-ZceUV_f1ywIon78P*X{GS
z#pB45^1n&z|3TaPq4aM8AA(RD6E!iB%%&SkQo(H@o(=g-<iAm@%FWX9BxfwIjf})B
zbJ=c*(B3#3=<!DXRc#LzBMy<(MZL(FZRLH-7c|K*m1d6R#Ca41JbHjqHmDyKDcaPN
zn!dl88Tx={sp;m**Yf){PIDBQp^J?VNlb|F+9SW7jxnqL{hH#)OsH!&kH%8a+vC@F
zIQ?HQ#<!oNk=_{Tgw=I)G@oGOvnK60N`j@ipWRsrWw?(@eEY<bv3sb8`5^w@IU&^Z
zjBEflD0FxQ0`We8F&)YPet!tu)%r$kMUR%r5L^?QH$AN>)$P*+y&id;-zG{h&UQ)u
z)p?2s*M`@W%X5P(fbRCAK~GSi+uD{}*Oqqu)|10dyBM2kPSA97%KUrh>UXbO!ozJE
z(vCV})_B<JM%{E798SzDQ&W4kx+N7eyO^8ns*qytFMB17?D?1ELkVfgSez3oBlX-x
zKALm2bKxU(4d4SbFlL8L^qmX><j}-L_#V?wFt183k<O@=Aeb1eV2*QqxxRIC#L33X
zqw3p*dz>*5!V(uat*rWl_&-_9&K(eL47?)d?^lgXsM%<93$hY&d|ol!oLXc=1@!u?
z7M=0V+eD#&jsWipw(~a`WR??PX=jxiiRP-%j;(v0aBu27UCCPp2bgW^VWV()ZB0d_
z-9lAE%@1b@-{)VR4ai*=S`rE3CWL<)48V^&Viqwa%3q~tbC4-#AaZmG0FV4xPWJbo
z8!&T-pb8YI*~0~FpWqRL+DMO?r;7GF21Lb?NkV%iD!|hR#Qg4`8&bwIDwA*dDrsjl
zmi_6=DB=?mV3zZUTM{7WE&OE#OSw29ydnM%U;pWU|J>L=+w5<5%>Qe*#`#U;2Iqz<
z^M4Qz<vR%f#X00<|2ebSFf#Wlk6d!|f?>tinYAVI?>yt_9%~OLu5)}@_=g$=b7D&j
zffF?}l-z<>&o#}$8SpmQxUc(nDkh^L+((BQ=BoMA>OSK4jJ9|kO;kHlD&oYKoo>6Y
zK(9qPpWcJZ#f(1lW8i-=q+FKMkk;uaWIiln`)Nne#3{6k3J`kMyHWX#aeeaDD0!6;
zwXuIjc|!VwqUUULS4-C~+L)w)->mwm69x2hkMZR1*-e&QQ39j-7Fv|g1w%3?%imEY
zO2}&rmcG8H<pcxGDf=n#3T}+8Nltqg^i*J@jE0>mZ!eTv?tz{rO8olK{9f|vn-^Pp
zQE`@-jL3uo5(rVqG<0&tbg%R**@wq|hI>Rxqr$7BZ20BGovdmsG^a*mYqlF5$r<#u
z>-^J&4+1&`L4uD2MzbfItMKZ~Y<D<5Q>%-OQF4_E#BS?oYpTkZnuz*^{@2Qfzw~|o
zQ|gBj>9Y*dl_J;%t|S(B3InoYg$;%(C$-oGuh4ar*N3kBnwh%u<>?E_*XLUV=#%#g
z|7r?R@JzZE;yPrJKkdW6UAK0$+DA`FY=a;GW$r~1B*PA1hM*ct-Ef7w*pKLdk|KCM
zL*4D5?gYU|d%7Plekdd<gqQS~iMQ@hZdmXRH(qi$Oym*4uY4vcXH1Qm$@=YUNDEKP
z>>4|^zO2kjNECaK<<e^LK<oP#tAe`Zld4cHO*zqN45Rh27QWhIO$ZqGq5$BM|AbeN
z&RiF_?LSw?%QOQ!(#5A0Iv1@Q&D;ha-22da6^1D^9pBT0=<%y%q3)iGHZgtI4fL1l
zZ1PC518%eDNOHt$-bFwFS}72WSYrAx;Kv}x@?qzl(U*}l2{VI|yRjlFu6LFw=E83=
zcn7JPx*L*<YP~1CC!nyHWs=X4x6ct`m{{I|VSrAyChr6&I2OZo!(KEHT9YSC?Z_Mz
zNc9GNIG{0CdCB4OHQ@pF+8e<^jWiWL-$rAlty2w{kr5EmHRmPDvGQtY+mc}YdDp>I
zvED`3nj=J3JLGM`J!&<r%-em^8x^lW(|%+l0BZG}m_2!RCSq3Ug6Pf$wD1;yzxwgi
z@k%@L{MP=RdfI4!*d#yM#|?r!+%74+U6b}Up*%zDcBoY7?S;upcQXVa0Sh9UCAZR*
zxZoPZS_Saj7NWIqel6Ym){Z8gbdn{D6}IN0)Q!ePU>RaB_KMLvBP$Z=U2cLi^YWMK
zt!XMsj-;>Us^Sab*p}rHz}0JRa`X?1^Q5#@;!YDG$`5l*iCtwun^GTp;LU;y=Xf^}
z7?!PDq;2^iRWDB6DdxEW5DavM_SokwB|Q#AdPqP7iUNj<2?V$`;9vIGWnMV@2R<Q+
zx-id<n!M9jW+F0!T@7F*TZt~Vn%+uuGN6k+1e5q`jYxt)tr!5a4jC^a_nsO76kx*6
zvS*6S3Uo%US8^v5i;~OQ5+Ybwr0oNRnLBXgf}&MI?@X<#`cnB&$sD`|e}r=^PF#rk
zJX-6G45{?yqalbDoCJkRRaR8{EiW5?Rru0DbqBh9KR%k?x$63oitWWh7+4=Gs-y_}
zs3@uTgTQvU=_A}P?o?zG+3tCxqBLfFa0eQL7!(bafX-07E0!zC-JtB?p&nG?hxs@%
znSp-0{{FT1y?j_qB)rh?lKf*nNwuF}erX!g(9`C@2kHVNu@<H|$O30DURK*=5@`<I
zXc*00Mj%U&<Zhdl*0hVmN;lS7OKcGfg-4^T7j-N+B|;ctcy;&fGCs&lMOip5$f1OY
zsgNSL1b!zHE=S;lUVaQIY&!Et*nVHfmQHzkNLq9WfZxP<f72jL@gJ{TE&yimMqs@O
z)5p0JEf5I3JGMWg<LbTiR}mCH4JdsnC@Ailj#Qtj-bP(3s*f*1x>?&!eMa)X^nmUm
zu&Gf|2ioE<Dv4SIBb<Bw+_cwA!cnNSxdlk^2{)r>ku-H}>O?zXXHXJK>nC|WS}H^U
zVSUah_(*brkLcfG!F`rde@}sb9&u$O?f%f<f@-Y#z)DS31C3@N)@+LJYeqEBjN$!X
zotu|vN5)5V$6Q~v?Ve$eg+W8s0RIhS*V7K)s6)tXlg;Zqlpt6ic48tZ=rBtAe5a#u
zYvYX}$0Aqnmv-GeE(~`r(=Oehzep!G=VRFSM3`;<DH30au>}HR=fKpk@)o0H>EZ;j
z{g8SYEK7E{?7fC|q{FDQcN&hAXa3}S0^0;7bFr+qx$3jp$*vtEbeECavH9$SiTmMR
zK2;cUOwoxq7T@Rl@mj!2go5V$SmjQz#f{<OB2lT0FbpTrUGa(uzh|0M<jprWgA(k#
zVg_*+DHx7?j~?DGfMzh8gAayw0bltq&Ho#^IMY|eTpFf4F#=(O^gTSGiB!;Z&RaJs
z=Ul`SqXjGaRd%rijdLR1t@U4_!k$XTr@LL-nE(GR)8KUry2o6VJAgA_HDERo0oGIh
zSEGERu0|;Ux9kFXqmt98v+Kj>!hm}bsE#tR!FBPTHBkoI4`+3T>nBkXhs%9k?hy0P
zp+~DN?~R%hnkHZm7O-%|+>S!XiFJ5ef^bUSclRfh7i=k_(}>cWv2ZrLcVj*zrwe7S
znm4U3fIWn|(I#7k=S4?fvi3`S{^g!*@yF1Ixo7X%Q=@_!@4;kt<zbr@VWPS((LsO_
zae;jOQCX6e7D7xWzN=3D@}$arA$!W~4`~naqan)#v5J_?ti&%9VnH5E0Z!8Lg|Ba@
z@^NMXv}<(}|I$5oKb@`C_sXh3K*%GQ7a1X~e0mCq3^_B&S80MUD|`BYEd=lc5o+xx
zW=9SGpm+hD)&iJ|oQP=r=FGW~18=YPlZcrAkG=PfYHHorMp0B$nsg}wN>i#zlNwwI
zh=_<v4G<OS2BZfmvCyl4fKru?lz@~-OOVhJ=|pNs5TpeXN(d0*oonyA_BqRS?m7G1
zF~0AfZ`}O{Fvc7&bH4MPZ~2wy_dFTw99GK|?4JC4PcCsq^(6|qfs1y#J0gyp>1FWc
zrI5N6#5II<SR6~^5|;@mb5r%glsj8I{o-dLF$f!76&34~FR1$xVs`QgJfMFOOFPfb
z4)0$8s?6Yk$3Eu+Z+%AEq0wF8I}{LI73dNp3AwAVyc0AbF-c25FzFXL6r`TfW=iGl
zZ4VJn{mF2v=4IOniMsbiU8&@f=9ty4fiB3y)7VQLU%s8Zcx%7~B}j!=EC~ydzD%&C
zc^AO)Zj&6}d%)r1^2y3;y&l!fQG!{^lN1lN+oxF=^g<&4+yR4M`!`cku|o?$?<%y2
z$(v`@0Y#!Y$AY*9w`Dx|BsyPD1m~-RuLihz;5Fg|_@(5-ZYG&L0f^&L{D1+ny6)e*
zQNMEoh*MK<3j=I@IejK{dM6;k0$7Y;hw$sWw2_|->#J~@`-eLV^rkhZ(;Gk<UaaLj
z0{&}*gkSEkd_(ha^gnUqXaAF%u|E^pkeq>I_kon|r<Gxueoa+;ug9@()G)3#i>La@
zMK*6G(jAe<BjpH*-~C}-42QkE7?`(}`<HOSblw$#u-dSQ&K%Ymob8Ru&rlD9%Vi_k
zSe<)%%&8BrGq4#d$=RmjKQ?N`M|aRaqW~h5gx_bpADZU)u_kaugahK5YHg7qa<^${
zVPE;6Idkkzx0Af$d`~IIr8>2VZ%T)-rk1#Gs<^F7%kZ5wz*2}EJa`%WlR;M)n7tuH
zIJJ-A@!-E&{A)b^zn&kq$p`kq&bfcfS3Z+M0{EejS~<PnWeQ!b8U6_2{&)XI&mCCq
zjZQNi=ye|W0(=AJC4bRY7&w;o2QjKAgc<ZbjS8Z-jKTNL%GCa~bcDg@j*s#!*Lo@4
zdkgPa|KRNV5IZAWn{JU<fMQv-3>}d&?SIGqOe@R|c!>lUdRnCClSzWGIDnN6;aO2f
zgqF>0T0LuJN_O5!!Z6>A(ImIzDD`t2i^rV)N_ckv$1w*99zUh}ti(}xtD+M?_EpPH
z<F|p6dq8;p>YohR6A^Tl3J`TZ=%6C#Mn-E6gKs~e$VmMhOqRK!s_vgy((6)W)BTwB
zyP>y0c}m%Xw|U7fad|F05vQdXQMCv8N>^lqY^5TV6i&hPcQA%f<*#o)fc4}?zgWoL
z86}QQScMi^@#NiOML#7<7x9?rV;>mIj0#5p0s6ii8f64Ta_|JIU4D1*9Q)~kt|HXS
zfGskYGQSr=R4t;qW9-t}wNH{QLOp7-B#r&|Rqq1sZDUo`s8lK+vLRJ$p3iu3Naoq{
zZ?E*A%eWem&>9FoECJ(xd73$k3Denz^>;h9uQ0A}l-mV9rU#Mh$$a1#`y-pnReHUj
z)Cl*@nd%e?)H|@lEugmi(4^RLKo<-OL?n+{h|s(esQQlq@MGQkpA0W$66k7nfn;Q7
z!+J!>i3DM4kc|tm;2^;UJ^bT^nnc<y3qu9b2b*D@?hb0YuevUNThZ-=b1Gk_1Fz=f
z((sxR!Z#2k5kQdh10TPi8epyav0H5OUYMw^zv$OPm)@z_L`xoI?bZOs#!>jKgoS2#
z16+0dS6qzIHsq<60i7cQv`Z|8&-MUjSiI>kqbKuK2^W8Aa*7&=*>evg^q+C%bm0|9
zeP4W)HQLb0!|R5wrw>b|jN+U5hjmpb&AMG~D3dKp-T4AuEl3BdtWae@fKc^rhTIEr
znn2g+=b#12^SecU-i=Ng-Wy}_tVYS@MC#<GpM*wdMqy5L15H_j{g<IJpoyu#Hz#ic
zM8GM)aV}J$U_+f02@iVUlJ7dhMpntmvvW0d;iX?lmAPA_HIk{{Xg)jjp(a#c-1#uC
z*urO)04)(h@WB_=rLvXu^6H$(kFbyqW9f;3oC}txiSwKGV%?&btBW_EgqSlWW<r!+
zTR=9WkqA^)V^GEXS|QG4BN-^qgY+*1GX2t%yg!efMYc^)P<<u!Z7x-MACHJBtJseE
zVPt%DVdznv^MlT}KD}o)Z@zlgy{UA6;hJsZ8#x7Ft~!^CwOKF3xMvGvS7{)GjiE~}
z?8AbLXW_E9v_@F0Y@a%lWx@FkjKCF>l=!Wwu@cbA9_{l|;FcEipP&AhFPBEuB|n{<
zlaHPlOT$ll-xxf@QLfBtB{yf(>d(ZHdvk*|w!qNWOnja>-mUqa`Zs`*<4+6a^)IV>
zH22wM^q*I@$S<?H*7+R^;=lNO<<{pJW&Vq<n@2PJ|DrcGxVApFiGE<NEW{gi_zohT
zNHnMGEieg{sxgsy*jv!?*{2S2Q9?^I$|UnV3Q>Ax#|v|w+0)?)Lvz#OiE-&Qetyvx
zWd(^T@xSVtEr1+xyOMRF8{4iMIpavWVR^z(v-Mvz$<!==GK}cb34jptu75|)hMF60
z8z8;@$&e?2%KZb>nHJ;%1kqBP0dsjrWbaZj0|`Qlw`M<p12~Aqc7z2vf|mIMqp;5h
zY%vFdh1evjUOt4^xsco9wOy;JZFGspr6(sCW}Q&%L(TmiKuy|Ke^iVk>LN8@cAY2Y
z^1?T_=N_7I-=?ekYN7QLe^bH@%YWSnAsdI}**VFO+1u>S0c@VjA$*nOGBo=ZQ4r&~
z((CWx<|c#d+Ps;tAQ+i{`1rN+Hu@ahSs}is`^9o;x5@DKm+#B6rQzWkLCp4q@${jg
z36t7>wO2O;iKe|V&O!OIe2;R_HVxz)lM!vsMfJM4yd^-%rHLaTf3dvrX62`7)fa7h
zpZRMi`F}F_oJ13Fhpnl*k1g8hM@Tu5?V%{&ccX7?gR^gG!)U#Ct3C<E3QG$GW4Y;Y
zDla;Ecpv53>mTSxQ}fsvUpXMwVB4BB$n2dWe42&@Z#Chn7%(q2pa7P;@U^40<2U)D
zz(m=wwcJ-_C&g|em;^(7G?0NsK>iR1)VmtY;<#K>$&;?usAOoMC-)7rBT4#Y?N`5F
zkP`L&ER<c_L47%EhKxmBu%Tkw=-i~Ds62Js>>QV(<t}->{8z@F3ORz^?%(VwaZTsN
zobG0_-{iA9%gAUn44_*pU+2&gz-m&#00;cX1A%>%q4Iiw6@2LunvjivU|pj-_AMge
zi+qK^31LDv$RpaF1t6JrZ-ImB0%%wGons$Q90hR-fQ2gD1h(kk?`AO31e|j^&5FQb
zcL%<q^*6)amCSx{c&ner53t5`Z3UOMfwq`|w48$!FtrAb{U3FcCpo-;YIP#b5g0w_
z4n0E*ap@pp3mxgUuR%LspVltS?{V@c!^6qSVLpsMbMX9Q4GSW~<H9E)ho)E|k=GWr
z6S^hKNwDmm6}=&ka`%d$(o(n54(BZEF`h@FExIK(PNKeIr@t=(YPNu+f=`UGi~(sT
zEPu(*GZI?%*nQ&k*{g%5yi?N87mr=g^QzaTp(9jx<Lx%0iSW)eT0&@0@V-Tm4uAoA
zPhJmi@+O$@I^#&Gg>NYt2zKW^kBEx*in|_e+$Lv7Vnn+(y}a~q9W^PS7IkFvkPV1u
zq@Hb;i|Jw4wIwab6Noa6V3)Ny*sc9QsfI5iu1~bDo~yj#exg0tp*W;}=EP`iNo~yf
z%N#a@uS?<nHRKNOMMI6=%yk#mY{K{t@rBxLg|I#CDyOmu2<-~VcDthX3G!(^-}xj7
zTJMfSmT?K^rxuLL_4b<&hb(5sXfD7WL+spH2?l%lX9O$259Hv)-nFFy)K#(xF?g>3
z80`$%p*gj=beXO=9OUMNZ(K<@wfH+5TEuKU%T=cr_V!O&I?NpGIUPDk#}wCcs&m>P
zk?Ni82tnB0c12r=vx<s2a*UJuVpCUrx+VeB?#Y&>D3yu}QDbrGm^FIhHtd5AZ?KEN
zYXG_vXf~Q)5!G?se?kd-JJHIxQ4rZ$_g?S{f7cHl)fY{S4y3Um0=8BAFexpp++D~}
zN&Ny?DW#>w(UN$NEB>;ZrMB(%@J|M)(fkLm=E<@CQcnqAj}?N)#?Yj%ZCZey{hwFC
z;4jPQj-um#vG|^;IG1_U!XoI;pE}W5fZ+A_r`n?gk-ei!In>L5)snc#_^(_1Yb^eQ
zdGVhnc=fpdLdbLIpSX6!LS(sftYQ(gEDu^i(0gic{OZ&`Q(DaZyEKMO^sidsw&j2J
zYR@y~{;pVO*8RJ99fRw?K`!Y^X>FCT5Eo4ew*tGBC~<uasOtjQ&E@VFh_bciTw<QR
z@Ie_z%VUS@eUHI^&>h^lt+-IGO9>{Y#jMeS5w0qz)1-<jg43XBO&l);s2S46_(=se
z#)Q{}<$v;#6g?uw%zeMn>H9Zo$-epu=u}fMA?K8}>zIH`{cw#~zR_jP4w(ZLdSIJk
zo5}*Of<P9cQE>0R6V5r$G1I4R=4!BbcGKi8k#4=SQa@J*Vr?=db9KldXxE2*SmkGi
zuQd-4Yl4~YWSGYLG&vi$RU6CI){Dq;<aV5zlQ$B2ruh*|(Ip4>V}LqoE7v_roG$@Q
ze3&gUV(M^dW|1hwI|S}*HsTka>FL*~WG)q&LyXnjO=vj4t=M_=p;413rt@=5HjvxX
z`P1WI(_v^WRAt4aG0dsDA?jT^AHPN1>(UFtp`0?G>72kWexBsi3OVC!CpK25jNo=l
z9dhw&h;_Uw@D1Ze0UvplGW|*Pvf(p>w@Poi+JhPOoX%g+Hw`qj`pICsgDhRU_lPhj
zL9xYd??49qw0+5tmK@#{i|k|3tr$!?bm6%};bc*~>^8X6762pmxI@Rpk(2&VQJa7}
zJ&9T0YaSK%8AE0vZEw%jlTbOHcwbT!c|`c4@g2g#9C4%7_saT0wVNk<YOF}#<i(Vi
zNAxbe_gdO2=Rs;vvWLspHyhFx(v}^QVztsIx`VqB-RNPkh?Hy2M?`dOg3#nLP`mqL
zn|U9^uS*ao*A;kr`d1h|MCSTd#@VrNDaDJ;7sorrg~GmNyItPnq_Jc5`?s1k=BeQ`
zaz7cYq-altXgS^o9$MP#`)0%~bI+Xs=rE*%{U^iYgt@5_I=LH#{*VN?hx1fFO~%jK
zN2%p*WU+Q@fW?KKX*m)0Vr~^OVUCh&;96=>PdZok;@i@DBi4Sg<jd=c-CehzKS3R#
z_v+$#i|~^tLIi^3U`h<+t8=A-x6_FuuCcEgF%%uK%oo{@a_4tWeLs&2+w+P(S8vu-
zz60GHMXOBup9sXFE#NC8yD00iD+w>`m{3$GDG#H4dSo7%{gdHl%_(DvJjH8YPN?1K
z$&U|wZrjX#_s<xr{~HeY2MNi)`^(slb^wi=v;y2rnAfBq5nRNH^0DsPtPt3XDOc}j
zIu9@VU2A+}a7tWMD%H;)uJa4OxUI1l?%yYt78&3CgC@MR2mTN5<S!w!dHbIXrtJGM
zw9<K;0K97+0mmM6(qmosqCwk$nfVFbgR&(T0S?H{0elL+*zu0>GkinM83gz(9)RD%
z+mJJ%XgN%zD+H+y`>x;w^zi%TidN}vprgHICl=CGt2mUR3Ubs^bh~?5^WY^s-Bf)C
zB-t@W=9qL^W!3Sp$-{b@=Z_?>9&?-VmC<6qM;;3!bnAa=hzDp<>^(SL(hl<)r9RJ>
zXHk=7ZLmMW1{NiUQIZ4Qc0nV>jlvT6W|MYn08jQJci%r+E=Uee=6E;wkj#yhDeedI
z@b**n@jv=Kw33Mz!~hMs*>hZjp2N=Q+IwkeZl#1HxTO$2<fv~{g<QE$GpOSt#I^3O
z!I9+W#nuMAqIaq;r$$?~R+)bFwu>Y5$}x9XM}&JQSlGmyurI?W>PnBjeZn)puQ>os
zzL;4tJ~ey)yTX+&?O;6Ug`F@dz6gKeAz3xvqbBG0lvYQ?yQ!tfI={g31?2df)g&vO
zct3_-u{s@%gueQUp{EKPH|PzJYX8X+8D-yQv|#taM8V?2px+Mi6g1699@g~Mp*5do
zySbL9Uy*Rnqx9|?!ALP@bJ3o6H95{!uP<ImsBR3ju*ZwUW?w)yRSpK3&8Ae2Ah!6o
z<Ak*|!0{BfQ6$HSvv0bEE`3(pB#7-Kh<q7OoA3UveTF7CRI9~K_`XMpA%IxWZ2=Yc
zGj6FQNPNwCv~&(?5wCX6A^<3~Y;n4c6-s^~lUGGm!$(1tk82v01EsG&8C(V{G(jxs
z^D(7X-)?=V);m>K8|u(1`RK{enV8MFXbZpQTGyWpD*3C(@lWOc)giim3(<=i`^}^9
ztbDqQ{DM>j`f;4Rv0Iz9Uo&5Lez1s6Gnu>dGfFzf#`$dRrKkt>5NR%!AAPnTR7)Sw
zJ&~z2FcmQeRha$y9sw)GVDR7l8p2-!KWP;QRUv(n1@2PO`j&)5amyXqNRB)iNnV5A
zLsf+|KGHfR#ugUS7NSDVhkx_W76KdF8(H253nsH?oHjB#uXfP-3>SFhd-f$QL@SiP
zNmV7L42C2yMcsF)Svs3+hV+6a_2A{t#&K#~OGt}lN=z*9#9V*!Y0&D*E|`?OwA5QT
zNy9r2)9Vt_s11I@NM^F?90N6O79Nfg`-CJ*KA&y&ZYLZ!2=_pIEiGOmw31WuV{RRW
z?HCCllVVqSxVx*6=iGv1$Yy0AF6G50Muj+K1e=RGOZAGnLQKW>A(--Ea&1ye1Ey?n
zH&L!j_D0K-w`xb?Td{cm(6S%{;Zi6lT5dL?O{$ILV)#IHT(sHE<=vZ(LY!AVDzIx=
z@R-RP=oG$2(_t2JocaSYqHKFZE&zOIKt4#h7z|$8xzrxur0B#M^26`Dq;c$N=w|*(
zW9&?NXa7%z>mSkZz+v!>2}`7la&QWItB!8Bo<R&LN~GpcT1Gu+s$?dE!MmS<TE#a@
z*<2&FD;H?0K)(IWYFwqBZ=;9<MB3qUscfLQ!TBDWL+PVrc4Ad7jEP7I5w9neXBXl3
zWek0LWfHgzj;<zkAHKEnz@1cT3pM0BvI~Bb!YDdMpaBAuo3<)wLccKVoiTo8o#mXX
zz@sPmycTvA>BV(p7vtu-Z>k;H@Xo&E8N#{r%7(uTP$^`;>Gr97ZVXht8p>LxbiOLT
z`tta(a1%Km=7~2a%<N`4azB0yfn=hHa)G`Rlt)@={bbV(teFgCyK3#?m*1rbZVIb$
zU;ely8vD|u%#_f3>e6>tI90mg_xvuJlnMlN`dLkYSg!?y=Kk&mWn?_J3XPYo0s`^^
zIrNhiXzH}cFRh+uD4DC$BvL_}n7e=rZw4gafV4GdugEUG1Sq}yumaSH{LMuNP+SfP
zG!l?UC7JpDJ+m76v6Bw_$MX+W>)FFwt$-a|QyE0#v4xWxK!1KVLqLBgyTCGt4PbwE
zAnV7k@RsI}+e}YfZ#k7z5Go<aRKdo^UJx`*%Dat4@zWp&Z^lG;smZnp1dDJFN3xLm
zLX~$QPkYaVfgh5iQ|4pq=WoRvVS=#j{!wJ#dsw3}IXkTE8)Rll9aCH0ho(TgaV)8?
zI+i_!?<MBw4(uFj>co4j;_;yH5P8xVS~ao>N~qVADKyHXdJWqQ_v+<O8b0e4NpTU>
zczpYLzT;Wg)@PIH;_no|F?CGeMT#`+=ksVS``N(*J*#)E=wOv#vH$X&e@|TgGUf?X
zBU$*5@FL)#C`Hg!bp-!6eu95>=RcdE{~DeDY;^t)&VH7LQpp4pPO1UkuKEM@M&G*f
z8y3+3m4pD<i#1)<{X!=;Z<4PKU9XB=JAaJrNnqn2LT4f!Nz_z!_CBf5a-Y&^Pw~i3
zXt_-<6e=AMvruXz<fZi39#{L~3Ttc$e@W^D98%smY1eMx{n@ua>Mrc=EL%O9vGpgz
zX$6#oi~f&W8_H@>rIojG&zqv;?ae;c#wsd#K`+%n&sEMHne`t-*MvtLO^C@S9}6R}
zAD^MR$Q9NPD;qf%G}K4WM_V^HOd4H$-FvRa^*M6gw%{J&lidGuuhwy|bTRwjad~+o
zYNlcNe1Q|-Ta=o}O&`eD;OuW76m9_#{cmnV)S}OJ(xf;_a(*|!>*BcVVlR&LyQH4K
z(Ms95(-xGw)GVO)<jnOu$IF|+l$WH`m^lRws`~=?z4tpRjGDeid?@0uf4<wtiurz5
z>3sd&SlPkM%?_;uj0bnnHF7j?2yrS+Ga4al1{HiXpd!DG6pq+uDF|?R$_^Vr-3@w=
zI9G*U=}qbsWtXizctwN!&d7)1wMI4P*}@R}mPJv%+*UXJ8#D}#)})fx?{pX-KURNT
zJ-pm`Ww~rVZ`i~Z?DEvgtD82KAzqdgxbr(hyg14gJH5}|$$n;e0<^kL5aC(mvRM>B
zQ|{PKbSvHN*2>c!QWZNPCf`_jb*lbMySLrt<WA6GC*aAE4rEHn+&atDGUe&5szPdX
z>u0HOjnYUW;ZtpD&kW8@Jn6kp$|JVWux}WtX6n*xun7UwD#8tO(sR4MZLI$V^VlQb
z`5o@lsWT~OQeewGsPOuX^`k*LR7CM&C2Xx1W5-YWno@lAgHgev_3OQwhOz<V+r!;T
zbz`l^=ThuneaH|0?VWp^E>nSrbGo%HWN44hH)+@4UI0lf@uNZM@N)|V`+ClM?)6Gy
zn27^yII&i5P@ZP%!Bamhhp;wff8Fv${<Q3u#F{8({j%#$f<lP1q#Z(MEKQdTB8)DS
zyrb6pfS7IK6mO9ZYSfKBKLN7Qb*~>JnP*Pbzp0cvr&FcAK7p^pQy~<Sn9nG6)T<yR
zDwvQZztX1Y<_SYu*D~7H7*?>hCBC6|5nROleB&WC24Nc;XsNzYSmP`_dfSAJW?4H_
z3kDmZd0_GVmNnPc*W`|jRe`^CD(7#up&p%k;HKn|#~%7xNZ7nk(3S!m$~1(G%dpYf
zw%>y15*!2RHzIDRJ~+tR$_NI5gFP_X=SXAac@UR6i*)&dGo@3-=q%B?EV~om!w61(
z-ByMWxz9GwSWV(Z_#H@JqvmX-bdnZ2+ZtHo(~4{7bOkZ<wlkX2e15BSqG@V23_Vw+
ziqe-rA4Aag1vI^2MSYUUvjuFLZqtwti1iwwbz=Qiec-b+w#wfgS=s2b^a@m(i#^!*
z96ap}7{zcAyC8#6-(5|`f*Dwb(coiQAEGveupk4fvYj#(P4a)CC{I3rzg7A3!>}jW
z_QN;&(X^v`xV@^JZypGf=8!av1ZLWOG8+=mhCsIX3p0<%)o4j}d|hC5ZZ7=N<9>PL
z;bNC5Erla2=85!v>x;sA_V^r5JPFqk$AsC#c9>A_X`lH-I_RT*+T6;3*G|iQxn8p1
z%DeG0)L-K+Q!E9cGYl{!8uQJ~)DvQ0Ta9QIw+6kqtX@NZS&O=4>)K)YuF#uFc?a*k
zZYHKoz=Fm+5`tvrdLgj_kT^||e5$w23%QfQ!h~U)u{EVJ^T5ag3;pP?vlrVQo%(IN
zc>mtM8=Q8`9Q}0$d@+Tn;juQk9!K0j6QU~Um#wJ|H74wfbs^`d)ur9}?H`N+a+wbk
z9XlWSCv^>h!8yt)lrqh_G3hrvT|wg1zQ8$!6G2KNwXAn-n?`+?d>bR7vSsC>3Af_`
zn8=eU*)9afiFx8TdkySPgkUw&e-;;|ZqjBy*UW58YX4eL;hfrFUZ1<HKHYQGNusG*
zQ_*@#E<>p|=Iu+dcHY&qEKSkxN$6RND4Ga+Yd<m0$Ozp?YFsfXF>w};a~Ie!@NvFf
z(mAu6k}<!?>??0wxQE+@%as8|f_hM8X!)MP*a|8~l@F6M;kR!^bqq2fAdak9JC=A>
z*es^or8(9m^_AX2T5uS!HWZ9JQx|9I3rN56*-%I|Rh4xgF#)+Y7^Fdf>b7YE?4{wg
zB5cZUN0ek1<>s3#-3vIt{rt#uK*KTvwkz&<OzMa`gRn~HEGT}fkMBA-KvUzOYIdC$
zLXE*}8U{RhFe=)ZwAqJjgoq28Zy@*?JDd)^)*jN<4QzuPX>uZS;@Y1sH2KJhwe9sR
zj;ZXq=L9Q{W|bGjksBl`1dh)bieCnZ2V`H%a|Fysp>RVVv12=E_nwI{9CjxQTO^kr
zyhow*pIk|^6CZ%q3idIZlBU}xqlSI&9;%N#Cd#)Vk*cl3!r-MNc14ZR`V*QHpx<n#
zPc#DH12~-{VKuH!#F0L<Z!)18{2n!pcIr*2CT;-Ei<}~Qz8wlPp>ry%P9{)+B;rRp
zj?$YWwJ-H$1Jw%zF?IqnceBm=sJs4alN}e<Kd_<^Z)^XkAw(!sr|0_YTPbutU~>)^
z9%JW$HJS<N=L7^FN}u_1+~HF;evSp@*meKa2Y12DcT)U_8wgEmA8FE*ypZ*NSzCAn
zx`p7QI^7mP#drsEq>dJD<cHz2?uB16<o<Y8T33D+qv4lQm^QIYPs8@FVe1Oy>t?>r
z)D4r@i@c+$Ss!3HiZ9h<PK56>!8VXk(3cgFaWu8osh~|xT&L}eaf8)@Z17eunz{Ou
zmKM6xULj`7<0R;L(Cut-bXCNiPgU(Ptm@+nNnElU<%0>w?WfFD{mx=ttETuAJM6fr
z$z*RlB)&<3a7LpQ#q(*@C#tN5GOQe!Wh&rrerUnJ7J2ENhW@C9l=ydhCi~BHW-^E-
z`-7Yl3Ej%l1}CsiE-6~Nqzz4%d3SnihL$B90l&(R>NZoGy7cTs0a;HA@GfUBBWH#-
z$0qpEVYO)5X~btpul_ba3DUW<hil1FZ3lQCgMTAKD=ORz;>Gn7?Nd!#i_N}{FxK9q
z&XXnA4Gz+2x^oEt3#qdO=v#O31jJbACgpu=rJogJXNvicCk}JXcL#toL`%?|%2hyR
z>a<pQPB2`}`?QhYRAsa%SD08Zsr1{U)4o>f&x(rs%&=(3EUDiPvXy}%ydcm3P3^sF
zLwp8L^%QtqGV3U@RXRb%o8|F1v`8*OOl#!|F4bnnoaO=0T|jTq*A=n=8lez0B1?pw
z=1lT`-o!zG^A&Xj0<hCP{WoJ()15`V`H<d+4zKmY?%qO%Cbe(Y#&Dim<)5u4ND&R;
zcz}%(#No>-(_+1y6GM}7I5X$mFkzK2b@StC_wM+0`<9z`4wp)$pS<{u6-vcmK%8`G
zDxcYaFOjh|TQFtqvAmq)aFBs_aap$U_q*4Hv_b{7uYVm<eUy17FTI~yLgKO|uho;S
z*ZOY{m!Bm=Udjb&Jsf_^pIaYo)TtmHMo^?gcHB$Gc((LyV?8!3Xm&ZEtL-w{dm%0%
zCy*Q4Xu}UdK554E{;CL|tI@R{-2L_kMgI75&80s7BCy9JE2+eaG2D+w_oI_F0Fxmr
zp-xK`;5(-o5<qOubiqMo9A|HZi{HeQ1@A>&ZWW5+b@J{o|9oswYWHU6ODR$5ODs&8
z)D?>NC_rVh0uccsYOHjAD$Wu$_+Y3O#7_M%|D*0BOd^Gl{&jn6_jI^b)+2r7P5g0>
zWY(n+M(SOEyhtSP9#H=D)I3EbOm^#1p>yJC(HF3ybc<8%HPV><0lrLf7dN&IF^yYG
zLERc$Z^Y&%R24aRn9so*Hi&io{>w_Hdgor$d;F;)eR~8eM+VakJ!0XjtN}k6a)qC%
zCHw;b(^T|tx|M(7+h;TXk0Z>8;=KPjB+c+Ur<GXhO8^)H#0T~+xnZ@X0NLDEK{rVr
zp_(Gai|T4)N(VRS3p+MLP9vTa$6B*%*DK1&IW<+{Rr`6-8=#{A7-Ak^vcn#r>9N+^
z{ZT-L)K)@_2$!Z{87xhAAjlBk$7eF~F5?zfj&ntEO1UKZTKu&~yVp2hw7pPeagt&t
z$`BFXiPUe!lO^3$_*`c+TYV7>elO-)GdiNv7%!)LZ}@$ZhTEI)w-W0aEKr%6%Q8pe
zQnGbUe(LV;omzy~fS%C|%D_u?0tMfbW@O7BcuFP&@>CSezxU_$XFndj7;64lXYEFb
znCRUgQ&Rk~WPCXy?6D7meG$cjh1X=5)P{lq9^Lr0Gp{-YpsXnTEp5|hoochYkgyO2
znm!~dn~}^Ht;}phAuYVu8l``r%a5$H6$<w(Oc@HlAxjiwxW7GJ7FemuKnv<~%Y%cG
zutPh3n*$PVwtkaeH~jYN&HBx;bPd>V*jAPFazqrg+`@doef!`t1*%P027SDGvRmTI
zt3vGg7uUOg>*~wOV;vMfc;Ce2L<l2|cGSzbUGR3Z2c;EEXTaR$#o~xVR-umPg9fD^
z&&+vBT$Md8e1?R>%XPMi9H+SjGh0@UD#=;*-XWo1tcQk96qNO8ZhY+e;C-sTe#2i9
zU)SfEX%YM(h>?7hk~|l9uW9&N`j5yOR=aOCA0M2pvb}N>v5z`^5O4eL(V6a~`YXlO
z#%j)4IWs>Ql==J%caEj(2C6AErjgL9-996$0A6ZBu|=4ld?a1_ed%;ToBzO)mklQ0
zuK!V5Qo}j0^3;=>!*_;6FQrY@cQZIPOQWR7b<LhdI6IoufHIubl^8I$4>ehp^L>(i
zHrBATJ@`ra{RjLnKr^68&DL<`(|31DiP3NNM_%s(@>`eS8St?Rhqb=WTKY;2Y9DTL
z?H9Jcf6xAR=y_q;x-X%iHtC9c+$@KolH)aleQ`mf^A56yI&>c7n2w=BND1u<vFSgE
z7YI3FNWij|o`eM&eH<+q$t4xc1{P#hKD?*?p5ql0n-RPbqPqy%;3kV4z>DZfpY0$(
zQet8zY~P%^PJp(LW$&>yzqefS8zXluEJ=hW&5yjhT~x(lQduDQAu(7PO=pyWe;)%*
z?h|k}Lysc*YXsHC7((fy8opl^zF5;p<{^0;zy1(g6rnTXh#a0$nf`byb>X7nEJLRs
zo1(22mdHe+blQz>1){mY;}92h*^qN6K2sx>+b(NQl9kGyUo;evC(exGt&yv~gy29t
zk|>Ce+Rl<3@`w`HO0P@t<WDtw0h<0;Hk(U)X#aV<iNU-7#&_lQZwUlYI}uD!qN|TI
z$>I0m958hZ#+(s@O;hb>83V|^4HYsxj!XA*blK_LXA<*_tXKXp)=iswgqxO;dAe6x
z6M!6ki%3(t1E;bp?kePrK!EI`v%=w^Ti9d2vyFo$)*q&4st6RO`4{$kI(9O9p0sJ_
zPdx|Q6k};tTWID*L8!kgqvW-sxs*UW)JEfh_Zhl_IFs6jTCyLyVp_A}aQA%7IIj%s
zM(pkb3AK*X_XYM|G)U&6L_mF$q^j6}+Jy~T3J~)R<2JQQ1-MK8Lf-NJ=Gxu=c`$i|
z*vpO>(hkLU(07wBd+aGsV!UFeQXIGQvwimmIj%!lYUY_PpKXy2%U|5g!cKaXn@(IZ
zM(dSoN&c?608ybSjF@yBtKkdLA}7H<u76E}`9KF<_MGo@fRvyUx(t_Ic5B>Xx5EO4
zUU@KC0i!yw+-CG)$wR>K$p^l#2CP+)Lhie#g`P6bG?$s1S}>pJYuo^5=$HEqx$YMN
zlUpkxyXD8orP>R_rpOw!OoRWtVa@86ixP1eX|C0SqiX>$O&_>)Qr<uC`b|G5k!17c
zS&=X#U3NfTK$Ua=P~{fuROX{yhjUY`DYBQ7P9Qh0%Y?I(z8q^3#$o&CNA^N920ad3
zHnpUq$vJ0e*GI*uc?VgA%Xi1AkgKGWACa}=kI%lans;IA6;J59!~L9f>gAOq9na>y
zq*rKeF}g!R8bvPFQqd@}6{LWyK+}K=`J2SDbpzhr{bthSWa;(K&#r&9N}))OTrpgN
zuLLL@)Y59d*+LJ3)2N#YsLAYVy@12?5fpO<CPW-7H+;9gYS2dz#PLwWVG~haaSe52
zvdvQMb&~>LD*kvfGfsI7W<ltPnlpC0HC$6rpM~%dgmZ|`-!?c?vub{7&Tn?T?zdw@
zAir_ZS%K;WPasePte`IRCyU{XBAV*&QJl@{+$QxEN;IkSd59|Y2PQ<F`W%mAaiP>%
z+jgWlRk*6u$p*FO^gp`H<8thd=7!GP={^f;@V=hqvNkK3Ju-g;5O7C+BKyun;k$xB
zlmMKm=}Q*gwX!O7&&NxNQ|IqF`X4gn%u)PEgtyB)ohB};<~FW-d~V#cau4j|!8iQq
z1UI{a*z^E;oe-|EbL6blunOmfHWi{bi&xvws+?dimrR~M;d;J%zB>!QUX}%=Xi%wM
z)9?x@Cwgs5fyLF!C7sfIH1lx3v2&06TjbY|MyD*;I*xxo|1ORB8bl>#Z%Ce=y*DD#
zu=<ms{89PJTJ?l;Vdr1;;0ym59M9j|1mQ>+!ZmGaY99eAV@{nK+nbMwQYd?j63*Yd
z%wPV{=G{gda`H5EYCp7o*E^HZR_a-h3;FOI<rUc@EXXPVFq!PHzMEWbaLmbTp*|OH
zZj8EF-MwifV{pWwvN+nZ*nF~iL`3~11ZoSP)x@O{)B*vP+-W~Jl}}+I(1F|Z=2e?t
zK+Vg4kss3$>iq92iv5{L$YKP*&Hr6^d}BQh^xX@KUZKzZWYAY;P;9-}=}qY4<<GNv
zm_*@*O<y&?d+x<?t*bmMIj3mGfLVQ*nxR8F$iWz)`Bv0Nn@2}@8_v{aX4{xAaENKx
znjDUIN?_nL(e$)4o<;8p)rbJqu)xY4jIoO&d$&yN#L(4))N-zmZLc#=T08FzO}Sa0
zDX&3+HB$x8`qd+|C2&<;g1REIv$F!=ZUsm;`U_y=4bs#`o6eHY_20Wl)l;%HF>F!~
zDad!NQ(5tLx^nH6%&xKc<CE88`Omx#Kdt440Nm<-Pk_^8K=5y6hM$BP+=wG?BZL+$
zLZlKb$PLEqlOD>EAz#SHk|vI~_6l6Db128h96j5j+odU4MWUQB!iQ8JAMlOiYB46v
zk7#PVH^HB%b#s;->%Ltu7Jay%Vo<+1egB?$+=Nu?`7RgSyxZPBk_L8bFezUxUk3xQ
zPRfdcEzf~P3O$DWD1?h-8#Yt)f$BPThOMLH$8FT;ogTjMp)2<knXc}SIoh@zA4HGV
zJ-HFyT;MEl>;qO-^Mael<1o#bl5#nUZJs^?buqt^q3VwKy~iajVk{xLK(26QMuDpL
z&c%bYVRSvY1$FTwa_IeV#kCV+y{=<dUd%R5DxoZsH2fb9&Ytf~dtT}=`&Q6sOhWkN
zC>jq|jrNJjGQtcl@|~((NKNNHMFevRbjzAfACGCNO1gHl@z%+Vc?;M5Ak<`ha}BoJ
z?w|wPkUld62-w}JZFVQ(j6?vaSt}oYeUAaa8!`d0V=kMu0;!i<XaHBmm4-<zFJ9sG
zPaE4Ph?V3qa(w=X&E#y<r<?aVFpuIDH|No<XzC3hGMSBHLq*{8aU7`0LRiR-aq98a
zE%d6(9Q^WVig9!GbikW%J<}*1nFwZ!CU*S6UA$d6{*2!bjKDV%GGZIR`Ib5rx_I04
zJ#g?0`Dmr-H&~KF1Rz|H3xbEQcL}O-32Lah{A5@y+6d}XNd6{(%-L3;8;}heCJkrz
z5s~T%FNZOyd3URZ<!Y@Y+)f82b8qx>O6N3xQf~at!+7+GiKv^aF#o`6ezhe{!tH5t
z-+Zq)Zyg?*X0MtdxshLSPyUpm19rZHpzTgAz&BgW@pQ&D89L#ceSMh8cw>RDGUeI}
zJ(MN6?U$bZ3ggqfu(@lQL94|fhHMubOTc)+MLI}Zcvw?_f;#q+AZwwr2SvTn&pMNd
z5=xZv)j?T?Rh&tw+5n_8Xl>D10fN9*e+-n5K4zA?oBmW{SAYE{gJgqtI4KM9qRA7V
zrPwAQnO11GB#_czYi$E}Dhqk^_;Pvez{Rdr9p!SM(g_lQUStG-?Lu&XT=*oK`P>kl
zn|j2WrZrL#WKaNmI`dgm^#k?fo|}upT(&+SSxKZ74a-~KO>6QJOAIQygFhjX%ezWE
z;&_dt>z=#L<cmx%McfMb)+OuqCI1~C{}uSgnv(*+27*TPgC7H&VSk<>2DGywK%w$G
zHULcj45+yvD_841I;Q&TRVDbJ{|&5k`+svIrYu)K6hA3GVJ^=9_5f5~R_U(9WIJo-
zP5M1Q&l^3@D=&vUJf_pV)At0&N|$lT;ZZun2pOOsfxV*ogRlRn8+AW7aOW&@Vw5eY
zhULA^_h(_1SMI<ndI=6;W2K}{0x>0}2Ah_qqWbEyhkMxflu6@QY3X>YC{YGAacl;?
z6Mk5(Ns6S3FMMCB5bJ~HvR*n>`!%f?mTVq>+~J$9*f*`tq<2mSR*{DmFWu!^H*QW-
znCpXgvq4y(jSDn2=x#OOfZ>@*I|mFt{WQh|fD0s#ehSSFA7BD%7Fw<VCFAL|O==)X
z5^|K@sh&do@tHZ{6s*^bd~fb<ZF5Z%vPlL0SVJ=Vw91q%Skw}?+Pi_C`TSY=hcJi=
znTs<rICxjne-g%Qk$`DHhHvG$+2e5$dub6q`;&{?f)5vFTD`^id9`*Qy!fqZf}uMb
z%9C1WvL2mAW5W(NBhqm59!8VE+?_|SxZ?u$08qf6rltrfj2cLPrtU;P)cHH3_5v;z
z^hi*{*n3Xa+N8B{w6cck;h+$4XVNIM{h;!!;KbCShz6ECamQQW{Z17CuGLRqD<TXR
z2kDb21qgtlmSKpf<cK1j*@=Es`H43%rYkRghQM3ZtT-R~(F?YVb(b-(bpxDJHec!|
zm+#p3ueDYJ0PI>FeB&ulDxZ^=^XCo?{?gwdR<hgdT8Sqf5#HIZ5~4&-tu2H}yWJIu
z(qvYRRO;<!=B(^LnlhDa<$8o0QX`Z2oL(*ZoB~-TG_=7F)3}7u_<GH=r3Qt1LK8x1
zUen5jxRl%7GEEWlC!ef-sh_xYwa6o;dlG9ULl^WxHOnJNyQOqylTOk$()O3G{4#W7
zeLNDUasPSm9}nvMs|hg8e*LQd8kfH>zyDxdxW5D9Z#e7oe{*@?ogn^X7|kg+G6hTm
zfXh1dFGY6%fy+O2?Ho|AY_47Fe-(&60>^_boMQIj7mR()p`t&K`E~~EFPYjD#)tI=
zSaO+~2FjZ1*`9oqHErOEpM32E@HYRVe4bLEo|JcG<Yo1(mC@5<1uS}Mro{+;$(+2w
z9C1L?kBq|rY61c*Pc5RMzQ)%lI*+8wee8C*oF9kodvn1-&~mQ#K~@_`$O1|KiazOq
z5`a1GK`oEh+M3P<$c?@XIRCuvf><0*gDc^-hDA;Hz^Gd3;-&HV$IuIRJWYSN^h`fa
zg2=I__QKrQ(=5V#)r=CAx%X`A=uc5I;pKlQs|yztf!zdrN5j6$csluVT~(65ToHTn
zW6r}9JQ9DjkqG*do1ba-W&g;O?cnB4tI7C&yD$sYXs(>O?dfQEEg{Zp!dM<_8rvlX
zJ`D}&y`RL{h&*!4@Z#}J$z=Z16@l%LkD0@;qTk48NvarJRGJJv;|$~NA=={tF=P_f
z(<sL1rHhx2hl(ETi&5ia=qIRjNN#S)G*WWgkH!TFT<{Oxp0Wm~$SKpT;hAPAV!{zJ
z(<%}E0alF<TgVWJs9|L`sEx+9J=M=}Cw;yQ?&X=z^oz`M&G$Fu3h(R|vrh&?s{211
z8{kinO-B`E5l5$&Cta8y<ulu!W5C%XgNk>PzocF1+6_PP9a}kgi((FTr9-EGGT1#f
zKLD|>kZd_HRHv4mJ%x6-5G*?YI)kyP8-7q1uVOE)-0)TClQaWUpUwU3h%1P5<R7cm
zfpe^E_=+;xC4%5SNXQme(Wy^nG1h(?d}Odj&clnvxT>jMDB^^u!98fjAszr~&%0CK
zFoBDPXWriP0>B>6u=A2I$q16b8dit2sv^uxlERcc4zhBquvrS>=pWBMZZB>4%KG!d
z7elRwI>-2=_%-DdW&ub9`cw@l7R5H0fvJR9w1vow%F1sC_P)(S1JZ39bM+kHNIQ)P
z=ib+jsiI|bXJ2e?mva;9AI%RI6(OXSjuz1_x)s#86y_AEuBbe5aU=6~T7m;F;UD0?
z`AWodT&P!tjKAm?nw#G@n5d$6rqHK-Q0#OZ-~oYe0Kba%Rq322)l!rM3g|5b1QG!y
zF9hu?X1f14<7O9Kib7*)AX6(Di2;DRLHD0op;r>ZP)ul&I<*F3<e@-8MtUeaYk&PR
z`QS~7+2djii_r7LHP?a!%(VU_<=L2~*RXY#<R(VNTIorIho&*>=$9|;e#CF)J`8JL
zug*R(d2f0{CM^l)VD;2%gRc9uL}9@naD`PDfY{rmqko3duax|Sy_dmT<R9F-0Ve<E
z6YKn{ODa?bUp@;|_`*~Ed5$KA{^K1()M`Dw#P^QPp(joq?ztJ$)g>l2#R1WwR$B~f
zGUlt465x>_QoM^RK{2@0#&V+N^p@?$H9je(SnM+@;AF@#(b2%ZxCMl0gB}cgttRWX
zR67ex4l9{Nst0Ems3h7~EPb@IvTJEQTECwlC|j~$cG>Wg49fI%rcv_{1PfVN7ueL8
z7FH(nra?<w>@GKqY^~&TUJY;@^TJ4H$5vFn!b}M?{z4fOl>YG3Xp`w|=_O~)6J;)N
z9)|?W7>u{vSVpleDnV)8^TelTuUR7ObHgy2ZSy@{Dw=%cv<@%Yb*hUL1skP|W2O!*
zESwE4U24$FPVqcdl|3(B8P3m2vvj^2Q$FV=l+|vmQR?jj*h(8_3LSgUK%a5R%)4kk
zczrkMKyuw;J{UJ7P*5_cG+Az7O7@)(&vD^0@$~hjTqlGpAALJLaf(OhNYX6^o?e(L
z!Qus7%$4&S%!ELBO9))d+k8|zSgWP7cJJn#r{!QI|2pI>Y_Q2<jy)WZ+w+g4`2tkx
z;Z>xfcI#M5xzmHAAJ=NCk4bP$G95RP;k2pyep)*J{hrA+l1GB3KZ2%#4Z530^}vh?
zb14*q;Sb&T1X89qyPYIHg}#-O3%wBKU&oRWeZGZJ8R|9NTeRB^CpPk3*(h*F(&Tn=
zArSzm>pl)X&<Q|-ag7kVqFJqs{-Dofx$bS?y~4uEjJ4WF`^UwiPml0A`8n&DF>o4Q
zVXz1dQl-8ptkhBO5^N3Hag(MnIv@lbbeU2&IJt<2N_~=?9#YQOVLypc$P#}m?^!cz
zA>~GnASi^<)aIPXd{MQE1WA&15g?6tz;&Z>NHXat3-T$nuaYx>p<7t2Dempdvr|td
zQ?g6e5L8Jb6>e@HYLlPlhvstwD4-y6X7ua^%T=%mpRi&*Qav=6`I3wkV&c*Hubi5g
z&AK+fmhRT%0fFXU=S?-RU!8&_x8ekAyJyK~4!&!jC9h>U1UqOo7j{Nc-G~~u%cOh!
z#f`FXeiCaoNMjO?;2*P|0Z1XP(<~<lK|1LKnSK^Qqel_fnO!u0tga@OjX%d*KHAJr
zmiT%&^Q6?8!Q0)C3mWjx`>PIv+RFgK0T5Me_%R^zpK>4t?eA;RbjV{xB|)+|ZFKpJ
zwgKv7tA%xQU)dOM-Qf6?&RC!|7mMpRhk7V}{Q3Qpt0Hp2s%Gaut^c5^pq1{L`&~j;
zdY~8Wf9JxQ2Es%9dHcuq>ET)?^n(e&15F>#eW$NVVcj`kd75*u56tU1CHS8{``^6J
z|K$ys&H%;uz|1<NEjn_RV@7^w9J&e5`^{hFLCfr1EXyIBI02?efV>P5bK^@l2@8=O
zhb7#rzO%#o)QpnjDSJcE#qnP3E5=ib!{Se|s`~(3#2Jp!4DzS<ZlZ7p$$|3$B5=_7
zpp!7Q55Ssl?km!HcUA#l%W*^^l76aOmxz1NNqc$_4%&W0{IRz3^<a8a4-}FYvq7^5
zBFc3ks!JsxKcU{fgRkmOQBcj*h8&+i50k;4h6r%%8SMYnVD4Yz@b@+KKV=;L2PWzM
zSuU&62<kLh;0GR#*^dL8)JE@H)i>4*)iegaxTX-`>v^V0_<>L)rl;(iDDiH*r)~Qm
z?m~AEnCi{fZo)XdX#8}-b~c@Pg;h_CGee^{;Er2B9VcU0s*=PB!DnJ(Y6TPfwvc8-
zc@q)&2D`m7M7O|o;e$Y$GyK%7+{LFm#KY<R(LXG*S3avIF7dAT#IZ!E)REhep1GAX
z1%*JN(MvuxPj91xf||HnD>ncL6@Zken_(YC3>yzOIcV}?ZZ@PFY~ZDHy}hb>B{?Ku
zyehJxDp+$hq&5>Zjm+0>cOtoC&na}PToL2%jWZ5>x1wrs8P`Qo<;^gk0AFuA88S2k
zuXtV2g4Xtrg|G*4kZHn4Nja_4Rx}kr1ZMhcwmtc1!F(qrg;*X-?t?O0)d<_%x_L>a
zOJ3+yh1sDeoO6z&l!kpo=~_BizPaI~xseIgsUj<Dt8ubm$89J?-O??Av|}~u!+v^2
z_*C<dtUMysSdy7v4<jzylN3|x5<Z0>K{d18UQ*#?K{FBvgFa%rlHMN^B11)YJoRlp
zD2zE;n8v*n<CBWdyKpA|QY-4=d{J_#{)hHVnLzdgC7{3xCfH^zm)Ch_EU(CByls<{
z!|wW9?ka{?m-U~p4>e`bxp6T~7erduPr+Qaj1AhgIY@ZY^xE4Gw#HD0;3HAch0|+y
zE(l9n0kRkN%;B~@za{Z|hbQHz^rrz8q0OkEl3iC$Z0a;OnlcFif^5jCrxuu;ND7Ek
zX>xbAazqJgaI-mB<Ia$><)D0-JC7L+xiLNWI6twW#(lA8Syx~K4Iq}ZsBaNL1`XBJ
zDg^cMk}#lgyw#^Zd2^_!GENNne7Mg_j%DLQfZ!9ZJdTz}nh95_#<Nf`<}$3vhRik2
z&Oyz)E09I3$PnvwbLN~8guRb9xaM-W^Dx)hYg1o3U1JY*0B{;=?JR-=HgksY2I<kB
zMI-=Y1qa)kj;+!5m6J}}ORV1Me+;<kb<A&ezOyckSWIGfK^tzETev>Q>2hpwPHPz*
zi+2gBTiO8suG{B4S-O5`(c}Zwebb73FuImkU<Bkm%dQq>x8i3Tv&FoXJUbbr%UC|X
z&>Xp0{fhk(*?{tl%oHW3EjSJnG$j8{@p3{E?U}!vCJ>~4d&Fl6`o7+!T*a<uAnN60
z66A5>o53D6#{?f?`GW)~y?<6?#h9x#hf`UBy}ekk#AQaK?q1F4BUFmJy!&YCiFP2n
z=WM(Znfcg1VICg=EePUiszy@pJ@#QgK~+&;SwUJi(_`@Ym&Wyz#%FXmai=;Lv{xc~
z6~zN9iO*X{JFra+pj5z262Ou++^sL1r=*~Mc%=J567m+P>-kibPSd6aI5#z=q6s1u
zF86q>XLLS#)nEg2JksFJ3Eha-sYkp0j$cMV;;rE^O3hc^&`f8W?|^v0RKFdgz(C%{
z@WtZSx@wZnv7FbsI<DK)jq|x(_<8|D0@=pzeYN*5dK^LErpw+JPnI0SryOQaK5%~|
z6g<VaX$DSsh7!%_RA2k3pfu<fJlChu&T%%?m@qceD(IvycxizzE|MYTGyi>9S@CqS
z_fH0|pA5TzOPXc07ra)ONoy|fpp|&^h`^^aJoa_no{lAu@4L~3$O<offb#$4-olBM
zr>;v$Q|BFOCf*AZd|i+>e7$<}yN~nMU)0XPs^{_nRn}wWc`n*)WP`XL19(H<m;`o=
zt9|86_QSllYiBR_*kg8|oDh<cyyu}a<sYSeg4%0VORdcX*M<7Pd0p!!#&X~itKFX5
z(RGI6oUNkLe7qH{dash}uU_pnFPmYYM$IY+z~*j~#O?S;YtW}{J4^Zw(7~QgDrqpe
z+&X`~=o5M)4x&+Q4(~)uS-)R7kC3;V3^ChGNg1B#*})p8HcTlsts@nISUY&Bd8;mc
zV!bxT`Ug?%{RG5n0v3qL7AH^bWQ)YK^I|Y{13YOae)q*$J(X?<&NKCjYq$%VMczlf
z#e=wAW>iS1M8cVK_TB}qb%_w`oA*j!`8tbF0{D+SaqFcE(wII`-O9)1vbI#8Do4C1
zYf@lI<^@MdzA&^Z@MW%%IdKRd^;Q4kA>|NWDi@&$sLKft;UUH4&#-@rxi+{Jb4a$&
z=gUUt1zVl(((^&jP$mq!Z)a4-x(&>J6cs69=}EbCxL4q`4{}~)w0CVo3l-f?%Atuv
z#*rc<fN;g(08LEOHldijoexgy^Qd0iRNK@n&meECPa|W^M|FHF|DO<77Jb0C))v<#
zAYx}EhiX0I`aCKuVM%)(z4&F5ytC<J5<TqaCxffPr~5w{<)o>#C~t*%iZBPU!)JZT
z&ZyKG2+`xzjO##f*4sMFr8^(SO~X&CF^WwMC6L`I^YGv~58z#8LeUWp&Dd#Rlc2H5
zv}h|)pL+uA@{O(8GEIZ%s@Va#ZCgc?!Py|4B)jRhaI5^?f|R@$m%rW9lf8FgvKeoy
z+djicotrH`JlBauDZgxcwamF}Ew>%_TGu%GJf>i@9{F7Sy18xTL;p1Gr>yPI<;Esy
zTBA^u)=IXTVcFM=0DPd9z-cSoSH4qIQ(wOHA_b3UF`TV_HN|#<0B9ab5#Y~#k7cxL
z^O2wwL}~8=HKKhCb)hkvP7EhJ7Y$ZE+?G+-$&|^dEOjKQJsQr!H4W_OH&W2dZm5}(
z5o+OF_54qUA$xeoo<FXCrZ_kH;mLW_zZbs6-u)F+!0_K^N+SclAwU)33x(fBg|C}f
z0*L3gJl|hkH22cEa_f`wD#ss?f{<sLs(Ms0!Wj{(O7y<|Th(!Qv7Fw`@3xbcDvI;u
zOpEu_G)xk8_?7@XQd11y=ht%Owe_;aF@osqDM7bMj>P2t`U+Gu!|ZnwYvu%Ne1z1q
zCJD)y2kN%gE@eH?@&$9*Bm<Fgzr14-FBy(YDH|A24d21wZJcfs5R3xjQp5LT<=$mK
zPXnIb8=pt>FP9fTY%>&<%X~##b{i_rD}6B3nd1s!-RiS99##U9sS(+1sNt<oGlI-U
zuRiLz2)>US5?9GT_mMOIJVQz6jV1dGyKtp=EwLN%N{7rWOkXklU+leiSX0}!HylI(
zMFf>1ARtOnY0{Mz5$Ph*J5icIgeWx#7)9w#KtMoxl`6f3pi}{oP9W3-=`EoK2=QC)
zbM`)GyWhL-zUO)G{qFNV=PwAxRpwfA%rVFK73!PRqx*)lj1*1w<fE1N-C#)(hDFPx
z_>=@3oi-U<X(n~r!ooFmva>2Z^)>w$ZH?S)($zi*DKxtl5jveLUpayiGAXiFi%0`#
zB~Pjbz0%em1%CD9GTnN5dF;VB7K}Yw0i;dA?m&l5QI7wxE=NpKt9ezaOGWKF$a}|P
z>Z;lj63_8`I#D8$0t4K6VtkNW;}N*F7J6Gj<Zq|n{HN2eEoAN=blji=H@i51M+4fM
z6xhpUEb2{-eN$YyM!B8pGeViCE^-Np6+HJ4Nh|tETWK`o#oMVr(=i+FW~BJf#L#|e
zqx_E?<ckSrd8eA~U3ku-OVJk2Dm+i!pHu${P=(X!*+-JE7eha#(ztO2%CycEVIcvi
zllZy!@N8ibiL|PX&g^O?JI#A)f)wp{UMyNBDATICu@{K5|D^sF`FruH(TwM!sNV}v
z{Z2ftDDg7ozYp;@A-#W(#Xm#^|HsDSN(Z?VRiVSb-?ra(^Y31>aTx$8>xYK(zYKd-
zenKwvbzFN%&X^$syS$YG{bKM?WRXQLV$d9ch8)J)Q{Mg6ui@t=90tvVFDI)b01m+l
z-9vz`aOSzpl^>E<r<a*o;?Uxsd|?vIUkTcRE(L;jUj-Uoz-qj9p}*s*qvS=E{*Z2{
zkT8zvVe4Y-t#OMwXKP~oSn()>s(CNf(69Sl71-H@oeiY!(y(Y}m_gLqJ*G>3G(A)$
zKy$Jk30aGr)xAdrSfnI!9Z~`5Ch0^Z$sUNf<w^i9<!kD9&<h<v{K2vBCMyIVXAfAB
z4*=Q^SHT289-^C`{89!%KBh{REKmjkHXA1CfC%J6Z6L|Z1GFIkfddKf{>eokax-xe
zhl>~~VuAgHT_9iro(u-Ca{xY&4bX-F7(WZ*xd6LJA2DNE@$XCde{fWaVYr7u6$R@7
zC=tIgz4Drb3g|>>>*KI?h8x#G`zO}^Z=gDjmR|kZ*Mj_k*K_38$@f3_^MuPU-f8MT
z;?2FI06S=@XE-~r=x@k?KXm1O?$Z57pZ^t?<8|uqpb@s7@1Wbf9+h)!%z^GW*y!e+
zz7^QF{t|cklgbXqNG4NGmoUdnsAz8JDo#fv+f&Z1hwM{_$X%K#;Ub=G>R-W^M67Lf
zHE7;*xaq`fp|V<JclFhrwtQ4zS90yV{T!;SPl6px3rA@6xh;H=d6o=iO0KYpsk+>I
z@nlNP34xE!V4He*$0@TxVG_r5GP~Xzhm5W}1Gvpq`cY?7FU>E*3mrTD4`^4FUft79
z*o6)$FcSIJKh_Dqhq*fBEb2@`DrZrVnX>Y6$*dulNqNB$asl>}dO^L;2{mgs<z;@u
z#&1kvTwC@_fKA<-zIERIb;)puW^Y8c^O$W}$i@!4!^5y0cXi_e52xJ9r?J0=$oLt0
zY5;kNZ_C84GZ?qOx8HxI!XD~y+@jO#sMzbykIy?hm9BvNXsBWq-QC@D-F=O^?Il(Z
zo#r>jH&PjWgz+LPi}?9glpG%H=O!f5%0?YnQfxJd>U`D*w|<e;I=*zfnma|(PE1z(
z+C=h{q$y#y1;WV1;Gn?TJ3k;Xx*bWq^t89Jz9Me=(-X?ks}IiDcigHLvvH>8rg5f`
zNd$ka3cS<MFQ+`S&H|rr-ZgcrGjm`g4L>uAx#;CoXXb5h7h83vK0B)Ci&@-WelFnQ
z1kgaqmrDt!GI8sS#76znK4&@kqlw!?wfd#8>DHVh7lN3nC=UAN+9o~Au~JuBzlw$L
z-kl<zjGw~+0CMMdnba&!DX6FITe?$q+lGyHv?XVjz#t^8FRsCH%5q3V0xFwYr34ie
z?rT6k%CP3?uPEE_)QmlO*F)%OxC`V2nWG(_`$@*rzo@c7&thO}WO%aE!u+G1-vpEH
zbtNUn2eC{yPh#HzFj-0W%05#Kcizh@UmbNG`-N32Dl;1LrT>a*`A>dQzabB#@DKTj
z;|p^ly@cbv3LW<?U%3l-j0wFlivKXCDfh5EU!&Em+(|wk*h!QFs>LWaqN~BsGY%q5
z5uX~09k2@!VMvzGsg1u%oAmyS+FYHWzNiDQpgA?qhwi7~FgdPFA|NEe78!_CwC2pt
zTsnQuStyIw^U61dP9=dOlmb^y0^-vb*uIfC@FBa+Vl>+#K`0SIKw?~x!>QyFO*?Dl
z3&L~JwDy4h?l<Q|b$Z?iv$?`tmaDF>++rG1JBJU}$gt8Zo6PGSOR*Ju>xUkja^BdI
z7W$Bxj|j>wn#!J2j)~|{IfdQL<+_|$cm7da$Yft{2iKUixrPt=6xqtz&ai3q>O^i6
zgmIK8(hQ8LyM+pe%G@LuyQ9WpmV>8E*thaFsu7{gXhT*x>|JX97l{ibN<94?JOlnQ
zgkiaEigeL1p6yxeVct$lJM`);KW0%y*2keoj@fl>j3p4;&8DX)M~EEmrPe|%jTlUJ
zTM-z;YW8S=!NlmqwM<ITw}ZrmceMo?Sssp%u(bYzn7!Q_Gc0V^UQS4U2W53Y)}gZ_
zo00xp06_@@#TJocix-~0Ds^N=OJQPj(y`qtO`UKylZzwK<1O#GFH_x;6?=!Tndv}0
z9DA?8=c60g#K}K|B~ZoS-`gNg^~lsZxwmYZDZfASIe!&;-szDEEv2{`>DfwZ_B!gk
zadq?9I`sk?6X?I<{2-AL*>A_?ShZ=AfZ>m|p$NA7bYVu$n(|!xw03~tJpL#~%pC5b
zee}vpO&PO`YZdk_E*K3CGoQOmtaS<tLobhrS#hrHH7FcD0`#B87E!AO5}|P)W`Vsj
z9k3^bxsTa;XZ8Vz5h|uGAH(mUXtNov14>K8LSS*uPrTHB!72G!-rrrb4EFz^k_eph
zfBbK`zvRugf3XGyY5&dfBxLH)WPh;YpP7UImES3t+V@Mh3+m|iLnHTrQ7tRd{Uhit
zjZWrQ#2gOAjOY--BMx5_0$`F`Mo>!?Ic@jp)X`u$Wvka+-dRLN*tDj<5W6~4-1bR|
zqwlYZ`<;>v0WFXW2uhz&v?xt324cBVFU%(BP8=pBn+_Sy%0qo(<+!dvz*8G8jhLGR
z5C0i01<iGOsHsQ2?fC$f+5zF2ns6J?8UNR>&r#_`)I3;*urC{-L<#&s3QQy=qW^SH
zfXrcJm@`0%ZJDxbsHyQAUj4FskdSpvAWzI^TKvdJhb|P|Xd_GOB%nzwwY7DkOHQtC
z5wbidwA#z_M5O?vZ*cx9;Gp*ylt!QJ1oYe|MRrv;-d~mdRf4Ds;m~t-3B^h@)Kf*q
zYa)jweMN4`#MV2-bqaKiIAy1Zj}Gx#6^&M<3_4;$CMx0!JRm`&$u3B};nPZ{8JX(l
z9{<#Gu9xoXL5Zq<ZWChb%^2nG*01TgWB!98nfjB9^6mRw0UpY{IS+a9E%S+?T7;gK
z5^7@Z$5Qwbgw~M_5?Y*kml;DB-i^d)p+{|N!|(L>T&RzHbVkKomE!C0n4i8x@k(QH
z*E~SK>R;9`CXr{MjX=u(d8bZb4t9#<LQEl3;M-fAoX46no6s1fjKy5AHT202mJ!|d
zRUmK8)c1KH{Ow08u<-iiJjksMkli`$h_*bBA2V|Zow-6XLS!uU6`t{8-Z4zE>3RHq
z&Y4P$E~nr4Ia*Ea(rh-DFJEI(BfS@-@?hd<!;t5p{X~NsV!&uu<JWiRuly9~{L3Rf
zWQrIC)bgh2d~W0~0aD@8X+-T-{9?HhOVHhv0rRfdW8H(nJgyeZSI<>M#|cwL8{1{*
ze#^aDVvA_t`3^ETyg0+@&_?SCYaQEqSkea(@{uIIAeh7Hr=PuGuEloinGzlbPLh2+
z_4|vn4&-#qFfy!{4(yhLwG**ZU$nk-j#y-yRS-3CN#J8d1kSl2uQ_jgta-*s-9jQ<
zcwXmS9Q@<hu?iQBX8w0qK7PZCn|8{@+G;h#ka*7l!EEY9m~Ho6Nfr#0pfv?(V#spN
zDEo{c91Fr-cN_@q;Sxn6dM{h{IfL^YPZ#Y5O0v_+6C?^ws|#D{zX=a9(HRN5ChP5q
zpa%mg_<^Xi7W5X{^u|jl(P~;({;(3OGs`TlWv6+{EYEj-YQ8pFe_!td==o)J`AN*P
z#@R!t<5S)r2u51iZb^G5xfGD;IA`<|ZNDz@_odrjYS9BJfN&#Ox^Glq^oIPa?fl@+
z@rI(a{cmQ22K(uMcaJw(_^Yb7|DE4E^6pPumH(4O=RYFu{g*_yKeq=6Bb^RJ`2!zW
zo-1*O(FnadV=ihutp2F#iWLR7;)8cbo{XT2js8%QP;f3I(i6&c%ZZRQV!--GO>FA1
zC5**>M^2abPKR|lJBW9{TJ+Y$8|L;4Mhzrcq8=9Za{uJM^eeZ<m2fHVIAk<eg{IV^
zi4LuE)X3pT@BQ;nwu|FiCNK9PFXdFF7cULV2tB_au&h(7cJbs>y0x1fTit5ar!B_X
z<<9W*eopUK8xgCEdacNz$gx0a^;^2%MF@Pkh2SDu>)>~T1g!@vd~aQ%Ypr-yJS*qN
zs%CrtrmZdao9Bv0DNTXSAz_{6AahUuO=4X+$7;S@voew8n*;i(cCef$AmK@Tifhu6
ztbm>Fu<LzmAtklzhq;pDlwdKSHXq$Cw5bcYa41&?Y&Wpq8$?2Iuyd;E0>#0$xZ_1x
zp;|($L3I-MRC)Tn?(_u<a~ff~H*Mdh-p$Bg+a;rXcb|r4H<m1G>1@V-LX7(PiU6wH
zfvuSNpgU}L)Arj*+~pRSt{l~;{Bdg8f~BnQ1Y|l(6TWC*)f6iz`hCQSPVTORuC}SO
z+b}wFxRv*Ty;}jTz}Q>6<e?*>p;Zj(JPz|TMfX5sRhakpvv3VL0O3DQ6QFA_!Me1y
zWCZBVrivP=5j_xykxR(szsvxo`fo~oOKxYx#?<+$dS=&gT2FI5fBNuw;7dPZpv5pi
z(K%C5Gq6>%s#?V-Yog1@P#Qwh^~#t*KnfB!-qR*CMMdOn|F~Qf%IeHZ9v%{wbRhS3
zj@iBS>#Wi7KX;Cre;lo(Y%mrtJ_Tz9ck^k*);@*peFr6py7e4eNL2M7YmVD@6H2es
zWFHx~v!CF#b6|QIQ)^srb3W|6Td}}Lzz!4gQQFjp&9jB;To2awMsPX4VYR!k>UI0q
z&&W;KPIb7LRKfVQru=t#txtU!<06*#S9jf3dtnCme+Xy)0hQ}6`BPSycFOu6`XeP*
zeiOvb`#J1T#{S)Wr(pZmuc9O+z5h`V_@Dgyu^ZrF>ERGy1^dBxIO6?_Df}Ri`~E0!
z{L7^mnx87Bj<G-Ig@CNHGUG3y-Y6L^ZCQ#h`FKceI>dh<m5>YfnA})$3L~5Aj0w~`
z-BWM!drZN9-Y5ed6?xJEKDJ|}eQE?LYPmGYU`E4JV^vxhJ^Dzu<;X3+=IcjiO4I01
z6AVLVz~K-Mwi$I|QXU+v-9IA}AFx$0MKh6kY~xZjGOBg(+NIY6*k=wJp)bSV&3o<`
z1h$}P2}Uj20W#<iJHp-g>Canl-D>Bmj7_i42bh>p3%06$P80ud@40Yas|VZe3k5!6
zvt_6bOC8s95h1tG0>F{{POYz<ax}Eyiu$!~rz1|I4i{Zo=5F4^WfP_GMyAAvI2ao|
zu*DMSo4@KYAxMuP4^mgMK8)_zL`&}JjNUaEe9W@u$@i9X0`?umNkG(rW4)+XpH~$f
zTqm|zjvW?%ZhypIy>XIkgK!LV*}H#ckCpu<&-^*b<(*%np&?5PJ>>{V4IjQ6=zhcb
zLK#vtfv3XOgsOvC(XX>P|0H2R@{D<(y~F7Ep11m-#CiZHA3-}<;r!xoBMpfz&ve8g
zZWsWL`ojcQw>cnRV!uJ<neMQN_k24VNfVWD_nOT{P*DK_gJ@$(J-O!2S6w<#yfww7
zp0+f3xdGpn%Mq7qU)3}VoA3YTR_9Lx&v?L-ZY?q=5g^)A@1sGqbs}7xQn4;vO5qrC
zGh1KfkTvZ4!ew8T)EUjp{V>G%t%S?xEIFrUoBPjI1+vIr7qLYGI@P@j7l>WA2PK89
zw48Sh1H5HJLleHe;Eu9z7P1NLf(b$F!yY-GwF3FYb%WzeBp?&cw%-a1w~jj}E;o{o
zpr`PSW~)WSBnTsXe|M(+g)jmuhacMYFZN+~&!1?|T8tMPow}&~5M^&{8QT>s&gc*6
zGIT#&TWyxcimWQ*rh=dG;62h$1$qLST#OnXBdXKkj>B44M{!;kYGVu%o^?FVb)?e2
z&4ps;yqSQ|2Ux@p8>pM>0{xsTgtQI&==}@%`ey=0r=w*L)Ge>@2F%(9Z?Iwwg}CCT
z<C;NFw2273ut-4lu+Ac<1VIsBm9kM8_-c*6(pRC+{MzY)&?hc^IXk20W6FdXDPG;I
zW{!<nM6~H5LL7u7%Ew%Ii!jTuLtc?jIg`6}lii$2%8dGqe^uebOMCSO&ey0Je4-O@
zg<SMZY%9rLP}Qw=DjQYy@}{dOZ}XPPDm18<8o(52AApIn*pfg=3nT1Qz!ukWB8o%G
z-imO100T{MNvo}k;)#G}v??)2>$ZM$(L9OHBsPEA2S<`b^N8Lp%%V%C3hAc<aF|<0
zE-8v9v&24J5G#A#dGmrFMX~m$DVh==@e7f<#89AMb<W|v<M8W6%vP11gzk|whYM99
z4yn#6i&?KXjAv1foljhHudAxdj*^OCsP(x6HWdx=u$1>^+JHp@8YbS3Vk6xo7RyoN
z!8b$AfhK4XVDr6>P^iy$Fw*r+xm>i$aWd+#+`T2Y`|<i5-<<6u;r=3w4=+J0)sF-w
zNE-}37<6{R5~{L_6IJ3}^S)}z-FNcC+~9o>^ObSh1<qO=`4J4->Y<>gK1MiI57dC0
zcpaYPE<!MokM4MsWnJdBW;`8F^WlQwqeh|H)G&7x9bP{21=@^RY+UoODBs+1;4nS5
zs$29}?BmQ6*RGw?t}PDkO039NwaX@U(uvf9Q-4T|XVS<IK=$v%l6O!Rkr(hi;Szwl
z(EKzkh;+l#p}-+Apdfc*OkKV`u#!V&l21(iyqJ>E7mbJo11_2sgZS5?FG8`YOco^X
zVKTMp(sj*nzD@`rx;b>ic$?N;N~dW`^S!Y7N%xAKk_^B&xIwl_Mvl!e2DYO~V}iNv
zSw3A~UM;9Tn-FQa-QgokyaYUP1rb7Z(3VGo_jeGprO+zQ;tuC!8sWC<jF(4QqfN2}
zV&W5(QXXHF(OB;w4Xrw3>Z=b2f#OQF1?uiNdOf3T``}>za>fh){|@pCB>$g48h?fD
z^WXiQ|Bao=WB8q(1EuNR03n}6A=2uJVXzc4W--Z~{mss$N=maO_a7d!&f9qUK$JLf
z9bim^x@-GMSqcugdlkFeuL{ap_}+Ny&~at2nTm(yXyluB0_$8-DhfC};?*jN9-Ac3
zIXSm<@LAUuiF(r|-xB7sq1Ij?eg>iVk*PVbRZzL#u!^i*ov>=hYF;+Waw8|Z&}mqn
z)F3_8c^Lj{?VeZlA8ViV>XW0MkaAcgHc>K;RRR}zK7`doGw;I7!P6F)Ejg{qhsAvm
zX&-Yy0h-It5Ot1_+KixEOrb&-`6y?Pbg={!4z~#z4lnJndleNw@HgOR=e4`Fbmvj}
zoA+Oz%p4ZT>6H-caP!THLX+9pF34#q|HDZ77u)s0Yq*f*m1-7ffl8xwDD*hb>9T8M
z%xt&qE=qGf9bQ9|0c2Q*NV}%ME1)h>O#-@~7fE%N`&{~ckjsds*(5BdKMwA8;3X{C
zc$%a>SW&=iadlna%;l1}x~wVYO5)?Frui<d>|j<Ki?0<(6M`NtH_V5^Zuse|XDp2$
z2&l-MaL3IiwyId>QkZ;e)v%^)QjlFZQx<&o)`npID{3u3uyf!L_YwjHE`)!UR2yG%
zw~Bx51|(OStl7P{(WqUszY($rY5xwY*ai6YjKmTZe0T7WF!DTiC9=92&sUo-T<hbM
zSrZM^zNMA#OMMQ?(UU#0XNyVe*rrPMaHmrkUQCM{%?Y--cDLRp@-a2^RLjGo-UdQO
zVq>Ho1eoZ3d|_o50B@4liB1VtA(ykR{gczo&&lskpZ9XH>mRiP0Zlx0g6BPg>Wd0G
zs1br?!mPN+HMa$ZQ-ngRJ8LPsgWp6OM7_}KeC?H0a^?Dt*5Oxy?;z*nrdJf!rd}e_
zw<?(kQ-KKk#?wTlPm^cy=0@nclAxSMbdg>Q;Jq<OQUQFIF-`}j?SO{zO)a?yHbyb=
zjBm2#%-oP7p#N3b7-XqH6wN0Fs@3#u+a$F}l?h^;hTjPrULHC#CkxF!dpkPM=s2Io
znm}!7O#@oO<4^?FlY<p;-{|rln?DrV8LLx>|Mx!Oe{P=;041Lz1co)*IZ><P70O|0
zP_(K1E2|xB!mXAE+SjNItO|`TTw~?<;Key_Zq(GX$4jOMEFC3N$w3;yt#^e*-WRb&
zgsTmPXyy94L!Qiofq%8-ItmxZ3Qitx97rD?KD?cnOzNnDCbT_XYJM{sQ%*g*%@gen
zp|F>k@^|A}@wL{saF*rUD6G74{NT>;G#v`wCf9n9{X(`W;5tvp%C`l_cSG+k59B&4
zL+#xR+l*RVy&YU#WtXSwR>!I)q6&TsZmv^jkkqySYcdh-@6KLO&n{#*KGx-G^|`$5
zIJFClk+~oBq^^8HI)G*Qo0WH7cY%efK*mCKe}3ta+wg0XGh#1at{ro#ZF#)U|LGi-
zz?@J-%xqKOA-?d<>U`KAfEs?HXf3L$63bm)T^*r*qLyOTgq1?}gfpOVmU`p?P$CQ|
z-Ud8}EHC#~m`&VS+I_}SmN>XLU|vog`@-9!CfMdl!%T>cir{yU$k&7VlGUW1eb5ru
z?#2}0E8TZk?gdzaaww6>0Puu>DN2OylM7MP6DELm=FrsS!8_sd19>9CsBhg#PiI(S
zdC^CaKX_F#^vaB8gUlXLghKSqSVb9Ql!b{!>`Tg!7(?O(Io0VKNZBIzSSwq@vZLXC
zg#vd?bhk)R@rAe>+@e_?UCvCVOnWDJ%><1_t)NxpQ2>FSb_xKgAcO!=$|}%6OWmRv
z1h3CFeg|>9*No+<Gn<8)P9&D0eOVqfSWKnPc$1PRlN@`%g9@O+l#8h0ibzf%HIz=Q
zxTabZ&Zo(%WvZ5WI(XJZ2-1=8`j~?dGWu4DR&e!@DkW3wPd5y%f`Jz(^AJ;A2(qlf
z>>&d38rf|sAaG!}2`GC69+H3uH&cOlGzn;u-1`n{j03PZ9<SLd42^cW=MO>d^ap#u
zyU=d`^Id=(`%ECeyOYe{ACjU#+4iTyM*i;6^imc5`PJV1f9NVK)kRkJE-p`UBR>$M
zX5`s5(A!jl?sW8bW3)q#bM&22RwsXU&Nx|`H*tPxxx-M;PoDkv!}-%t#>TF1>^bC4
zzsyD<z>DR9Xz)Xz|1q(MI`gu*x-Rwkx*R{PsXZn>(2`rSvOFW}wb;0-pl%zMA?hml
zOUQHpMS+<%QO^nX9<Hg|h;~nQHNa0JMInx2O*VDVs9?7HQCA5h0k(`h5XJLjZh!Wm
zD^2UC@;o*q?F;{W=G}DsSs%CZckkv8PwKmKGk_=cOpH3RN~N#FRj}l^!C~|E7&rL!
z`c2mVGSU3sx%MAWfT-(u*vkrfhC|f6LGRYYU946Aj7f>mrP!u$&X4}bUR>|s=bS^e
zm%LG6115cOCV4I4J1ECVp@Ed*;;&IG{Da!$96UXT55uafTeu=t%Y+e4rFhT76HmFw
zv=6?6Z~%zU-+Fp|pK%LJVg@h)Eoe<;?<wOV()#*C|Lv#X&!}&%5PYEz$9K@Jlt93R
zXaqq_^xMVmzNi?xv9>NJiCs3+gkwx{6Pu1kNa)1bS=E3*>ijtZg(X)HbK;cn84j$R
zT36|`@74^TW~_4ihV)<CrRY#(aG6>O=wZ!Exr#K{o0{BeH2PNkOu3o%mab-$Z`XdK
zl?cre<T%MaPx)ac(P~&vPDs9TsK=kbvZ5}`dd?4L<{oli@uXa}Q`hFnh#wk6zrTxN
zNUO+Mvvz!7OI<{4ep|Fwx}}8!+Qah#OKZJM2krZL%9E_x=_Tw0<;Gw^NpK)1egAri
zyPA)xwR-_%+>v|@)j!*aRGYqdt~g!YCg_+K?}6M!$%|regZsUN*ui(_Cx+@!$f+#I
z*NHA;nKK$1v}RitLW|?Okwim<Q{O@9+5j-}(<3lR%b*U=|GcocNT^s&9{X8Xgd?*m
zz)Zm1YvhVc-ig^D1G@_zM|Vm%lE`O7(}<9dn2LDdQHJov=rz1N5O1sYVE!;V_9c&b
zT{lFcHUW2*2Yy=APS<8u6D^Lt&VqA}!JjoiS0-!h6SjSITVaB7ik@se=dVrCyj4G|
zJXdNQBhF|T<ys$Ln{F+j>7bAdfegvZ!t&%e&gG1=ScU@Se`@EdErsYm3?M96P5?ju
zX`f@RylCMWdw)fN#3nzouJpyFaMuWR%Wu!Ab=4?iEZMPp<cp-6=t{^LGs_W&cvDy?
z&p=qbf_Q_UtR~#K<<_v)6(_CwiVxF%YZhS(SoAmA#YQe?T>myCATg)l>L;uj*O{!k
z!p-xFr<`&%9r)y#j&&(+N>hZK?bSJN>-9-ptDtgdyQ(pAY8Vqfou}d~d>nJgqPe(P
zl^TC*_a%<#FcX@T<SX68Fg9-1R#su!_-cN@Br2HVwfR^((#FN;`sc4=;)6gj<#+Er
z^zC7qj|2GCGfwjk>-WVma%e-tU71(Tep(7&p?T>kE6zvhg*bvd*!8DV-WF50zaICw
zHYBH*MTA1o#e!?d(*=&6I*x2Z!|A+l@OxGm+>P9*Z$XU-*P1>UirSn^my{G8df(d7
zT@NW(c-P1a-QDtV!%H}-^ByaS-hcouf(D22WTwcMjSD}-?D1wm%6L|lB)~<D=*<OE
zkLfBdG9Qw<MI6ver$@Lu;1CRu?jn-hkh?%fC7rHAskd|7oa|eJB&Ygk<U57jENi4%
z9W4T1Jr4YGm18crVEGx(;I;_J(G=P1xVtgYE_9ezWj`Om$7|J-%C~@{d39_Lo*ySw
zasExt;n<YXK<anUYsAyd(vx&-?~tXII~b8QK&t}sVvG8xiz;xmq{K6aKak%^$4;Fx
z9yLpkh9)Jk!)L-(4ehRiIOmXO`Vm1WN_<?wjF8k$^Sx@s@kz(iKxjE0yB@@@Ut>HI
z&0g~$1ta0FU(RfNBK=I>Xp^-zt{V6Tee;1du|OxVm3#q>*=(MIK|4q~UV@%*ttRX}
zt<uw>Hdzm>xSx+bPorlEGBW?V><dZlS!r0^8&#za^g%6ERP}iwQWI93rc)_}D>;GS
zNAoPv*``|hqGCEF?87zPlVqbfE}jRp&X;&j9&z0L>DG|(?N4_~@y+)L63^IDbq1*-
zr&%j$*=3g(*Q?D1H<+TridQeu8}huCnm_k`>);%c9jEQTn{uAuQH#?is%J;b0{wD2
zXQx{nx6muPcVK6vJt8>D3V9@G*{qyunlB$|8-^w&Zvz?*i8_$|yN&pr)c1Xrd(7xj
zw=f$()|X!I=ULeI9dx4o)6~|L1q4wDh`zxkM~E<~k_;2lX#yjepFmZT2RI@*f%UYM
z^b?^D=}^st%p+9RY!yyN@2ll=Ja_ss7F?2-j}0lhK9zj9UWSOd_5M}nU2|k8s;eH7
z_z-|y{r>owKfjX}LQIHXRs%bPQ6?#LP4%z*7e-&J>dnT6z@Y`8gmDnn{TJ38u3t{_
z;6GY%z*u9VEZJWkm?p<Q02>3z{yx=V0fg+qh_7YjhDhys^iR+74{_M7LT5W<>MMDN
z>;LilD2BCJp6Zp8-(gP{P#SRS3JBE?&nd5n;J@nTy`diqnz?Y^`@~k_42l=gUYtx`
z19Vz)oNfhL7QP?UO@hW*AnHEd%Vr&T;-{S`&$4-G_#YO$8>zqfBcNywKKxXf`FAVw
zPg;@foOFbU7ufacF~_>Oq#SbzQ}a-|+3_bg^fE_4A4REIk3Jkhk8HR40K4L72v8zD
zpagcj3$IF4kA*}q>$*gWIgX9x`=(m=v^}s53DdXj96+HL(C*^|E~3fgkA&j~6XVkx
zFv)-mUycN_HKiXEikFuW#QgnRKV-=*<9D*T_O^hvh8F-^rFibx!uIwX@hm{ekxL#}
z9lruuAzdOD`6RIV-~eKVe=?92vP*zTtiXX=jRvHCYXwS2AfIHAr0^-(ILK*>%81Zh
z6`tCfh)MNP{SP`pGlm{KAI=%@Cu&Qi#ZK@8%n%XsZE!hIV(S?KG{G9lYI~~61LWgy
zXXsu<5(Kp9vc%BP@GVYfqzNs0qpO|!o^*ky-fO5)iz&Kkip9Pla~V)t8Lo)E+Vvo>
z8!87XIdR)rgD@5$u)uYMWP@*P(K0<y2W&HnD!ngXs!P%u-hxi>E!<k^EGv|L+_KJU
zV>_8vtRkMc>VUAvuEMNV$v*}l@&L>*aah(!gaE?dKCXF6_=kpl5)*8Da>u>09K!3`
zwQHDl%`jZXnc%COh*u{n^QK6A!x8{L;UJRAa`A~dd(Db7XTRDRmoIbo0%z%K6RNFm
zSb3w-Z*&f`mB_we(s9HF_hAx*{H%m@3|LB4+jnDpE)2>U&<7W%ev($;353NK-gz{+
zA)iTGWxW}8ii)18)|){65fl4vCq9K2zde-Q;>6gig`jM|c&%CB$|DBF3op@<+q{ej
zhX=piN8pZ=_Y~&WQ=rblUQQmbY^cI7;uT&v`5c*O&OVG76$TxpxNwIG`_?rjcQ}&1
z`qTRHMWcPY03Z5Hhq~t%H+r-W&NGbr)vHCFrP#=dP@Kl0no6#t0Cx;JYaBHBxI3NF
zz54QcGxn}zk=rO#X;}>q{Z-rsjU^3-=M~#%<;&-oF@PIPe{(g=3P>A}je88_uPwX?
z$G^Q_oC01r?SRBk73Xcp3G=tBHl<3Et>vrLuP;ZRj%j(s<kE1eFRJy9oGqRJD-uzu
z#YTBOq%|`$_bs0K62LcBC-G=jf8|cz^eRLVV$i33pzFWBl1e_>xW^^E8+ssxnjAt*
zivaUR0gV4LMI*<OX9<COk<O+~St8oIzrKC6`YkniN|uk-MSZ?OHKXKH<0;|%IyJmh
zTwU3lLFL!wuPm=jL`jW%JzsNX2gCQ#BX9Hp4Fc>+e5o&rx0-y!XvgCc;?dvd3a|as
zU4SA@+wPnwm5Jrx6`DQ}6tI>rym;A@Wn=KlIjxqezBYAv$C5y2h0}1bE^5$?Rqkwt
zBg?3?r>U#SSl5#{L7(<JEZ<@}<ldpt+rfG80kwgxZcCzhM_zkY$G6E(S!Nx1=FNOV
zFu_kpkp4S$QXjKFe4L>M`g9cwmgJ5_L<I`*aGhnx$+saNeNZyH6PaiQ*a)ay7~qaH
zz-K*Pgv}YNR5O2aVu$iA1Uyr!9CfL|2CYH$$jqKV!ujY&!f6Yvq)d>z0?zkT9;<kr
zlc@M{@RB{ZzsCn)SnJWhzE)o#KZg`B_90xB9MBCMgbJua5e&9g8NY)nF9VZ2_qQ33
z_-(?6{(NFplHfX!>5l%$bUV4oS5e6sk*gZuO@IY9t@#}kMZeNHwYmNGr)u?cFaoG%
zw6Z`}BxR2viUCnBq$Jlq^FTL}w2vU3`40M~bRbquI)o5uf!`zFHvV+2`iS4hx`IW_
zW%6L=caXv_?-T?13($E58<CFy>I;|VME>*e>Jb9|_nrSe%KynTH=hbaTBpGSSEMk8
zwK+?kw9M}>DLwQ&U(9H~J6(E0y-kgEmuZLY_g&mYO^NI-Ubefdk_1mJgveXLvJwgN
z`~tqzrhQ;KBz^urlk!n+`waiiPJTx%{ZpXS&p6I{)1UD3_o$WmmGtKdqJc)PBWMVy
zLI16JfVAG}GhyF(Kp62Kz7bx*&894vj_PAR4hjp~`Me*p<NbGFdVW=9hQs8g0setX
zD6NH__B#rp>=gkCU63M0)C=mPl%JQ5@f8<RcRdp!JPbu~z>{##N^z_bhpxDQ$Z7A;
z#^YMq#+AwbtI?#*l`1d3n4So%)uv+XV0!gchWe09tUN!C^CfmwwPn6V>I&e(M>{BD
zX|KS&#L1ncm)fCz0aY2{t)@G8#g2dFODWEVT|s|r+jlq6tw`H&z+JOj?CZMTz@31$
z4R1HEqC25gN@t_qj3YU+Y#u{X@w=Z&KyZ~A!<t?s9kzkl*HkK0CtkZ>Xc{_br776T
zz)HdMP0BS!=X$N)W3vi$_Fkjx#9qNpRaCT|ZdJGRyfdULUYp-ZhD5+N2JWIcC5W2Y
z7R23l7DEL=y#H_}?p#s}ItW(mG~Ck8<6K5vzE^Diq0;^(DF0)zV#ze@wBw{y3>NZS
zP92x*L3d)EeCB$p-b<szu6slC10|>)2gltgu{&m;H~Cy5O1NJNY6Q}4`&gdue^58D
zX5GIsQ!|hTHM<fwK{<A7Oj?z-O@}_7)8O%`)-9&^LVyp22ttbU2TDRt*e9k2W+fCu
z)R6<}iGA4)C&Rge-=tfud^KYoS6_l=xZE|;)U1XX2_UAW2m0Kuy)wU_F<DhfhLk`y
z6<i`)WlEYCCAEy<NydZ7v(qA+ISIVo<V?wq_&aswem7=P46V>d`q!2{6RvnD(q4J=
zz8L7yvb1P3TKY)!AB{nBF`O#JIS5ILut4DM2}(`FGCtCVy<HM2vPPSxG_SnB=!&1G
zHrk)$q98g>HF4pCTd@;-U|QdI7{r6@uv=XnH@;Y^OeCGs?h12N$2C$AAGRkl^vP*r
zC-_@Tuh~AybsQG8Fbjax`P4ENt6E10D4@}>lwjHQp)11G4P&)A$r4ku4UJR-i%=7R
z^N3I30&Q&IreL>qhNZZtDfKY<Eva&hW0_;h9V?Tg3?1%*I>`RpqU^G1R(f;eW5M&^
zK{mdwYSn;y;ia>diZI{I(W1zpGq{`4m?dC`#r58`W|mjHyxjQ{Z?oHdC>jaIt_n0g
zmCT9CaEN;<6=RoTQ5leWMDggA!@+Qn0p~{}`S&DtJXPfkjgLtlk-gcIeARy@TF7Nt
ztvY&NDw5MoaK-<V`AG)&$5w;G`Oz`7=JF=0VWP4(B?YJ%MQX?2#Vo$xbS}-eghyie
zTOq-_3Zew&n3;7tM@(g9K+?nFapi|V!A+JoJZW(OcRXzAgXme6b$(3Eby0r>K`sVM
zQ*!S!0T?qeQv-!}kIl)<aKJn(m9H#{{qppYCw(2y{qjpS6d7;iY^b=E#LsSh&BPac
zGHnlY@-ah9A0F9l^?0F-s(+HB$h*buT5U7ySuFPCWt*16BwY2@;DurCx<0pqx9)E?
zwy)1sh7a#~(CmDrIh-Cjf-T<&b1U4QRd;Tbm)!Oy7fmF-SH-wWW>1YztS-cbb~Fg3
zZy;J}!(H_Uo`H$2au0C6GRsYd%ujRtt#hGg$4pJe`4X7An1jy=c3iqIp&Jku-g4Q_
z3JNdRCngX)+o#T2_lpR+smqCQWGsrC!}o#GHk*~YEl1Oo)<f7sL8B)XMf}o&eNoPR
zugq`2istdKz_lVtGm{D-DX4|1D0Ak?xF=<n7s6DSZ7-dnd(Ovx$(}OX7M_Wve<25G
z&ejD6A(^lf&%T4S?o~;f=ENTRQkBN>@%GnOX-6qHqCV&MlOge`CE7}o?lidtW_zo-
zIpO<_dhJM2C+Sv^5XkAXZK$h&16dFCgu~9TjAdZS_+nh|$KlssVK!=6ijT|hZeG}I
zPBZgOqnvY|MttDL)B6MYYBKcZu#Y0NKoPg>ttdbYA<uLfh2C+wr7nKib?(|!_!!Nd
z6$;fs6fe<ix`z##XjcgshcC#imF6%(>6CN_xNB-ExT{|>Mr6MZpa|%{!J0ck636?J
zE8jy|EE)TR-JLYft*)2M4Wv7)WvCe5j=NQ#te?Sf?)3>jnG+7n-ewcf^(UTt8%s#?
z?M`XKP3<);vsquRxVj~Ha%)jzt9DQ-%M1Y$CZ~vX$Gd$PTO6(f2Nk>JtvteX-T<;s
z8#Nw9sXfyt2`y1CoGLE2B-O%>x8oq;m?}S%AORdAp$BdTaGmv~z1!+r$(UDVU+6kE
zSXs|IfiG~Mn&vRR`eG|ERDlUD(2j{q%_F9M+NTcMrn^4Zxo$Xcp)PK4jqad2O8xFl
z@Rjq=p>7Lx-i|IyDouk+VjqQU>NkwK;Y!S><J&ZjeDRoPwelXmE7aDt?^70jXcir_
zX&Hx3GCUXqG=xH{vtuA&Gd#c%%?+*O;`BL2^u`HUTD#^inKC7JkADN%#-{78KL@ei
zkUO=#5UZ!!w<0H>jd4gc2(L=W!Ki!~8y#&bRXXW-EiWkT*gYALsNr%()|>e&yf&k6
zj0AN709pCvKV$6)#1Vl6m8ZH5yh8<as{80Iw0@s6El|Gw4KQBN`H%9=A9#SX)_=FF
z%s;&MD~x>WXO_(WAD=1iYHx>cX9WL(AP1%VLLIDRYTPg1esBL1=;!iZ{d`0qa@%^@
z$c<~8g=-rP05nXqf0pC8$>#QVQ#+<V6k>{pe^TUlL(!)Wh}GwROQzFHd@kwqD;uv6
z#sQyECOTo|HCs>@>!BPhcf-_O0iDU1viJ8hPq!7WsNcVquPs>)`xT1~Z(q$QlU`|_
zDWxrRWw*a25n`g`n<03Hf|YeOYMV){PAhF&7WQAObwIzZvk=Q$r%oZY2w-Ev$;4;1
z1SM6xeCyP!oc==7uT480&%Acb<YLt4+aH&DrFAad8;hl}XnnA_%f+N{9?ow}*oj5t
zEX9?<g};OJBgTJ3$(Gf^2|=9>nAfp})^|D|^G!m9N{v4!7(JvhEUxuJ76VldnD%V>
z%GPjqzD`VHpy3K?VI9~G`c70SbZ4NP`nN{psBtAP>mvQ-p5R27QdzVFP2+b`gM=kz
zeRaQ#Xg!8yhN#-7jfve<G80rgIwl%$?|LLnRZin^1$xc*#P-%ZQw;7&-dqe&-DR}m
z%nPTrEWb(SO;JbN=<D2KN@c1OUGWzJaD%lnlT)%|!&3rVm0K5uG~^D2x;KUn_vPhf
zWK&HR$%6fr+NYQH>59-v`HP~Tgk~G#W;#t?zKy_1Rbkl=`=3NSX@wd!PYM%`?S887
ziChNc7RQ&)jsjE<Da}KeiC3y}fEO#j*kot9+R-Op`n}46O|MM@Vt>Qy!4h1-;8P^p
znI}?6)cMjHHvSQ(M?iTA9!LJ_CCM0~`QnK;jIp#;&_*QAr1xzS$l15c*JwKfQKGrL
z8&HMbAD+S=IBckn#Llx(uhLG#n4mlRhLL%WZ1EUj%YHh;vD+wLK9l&zHd?opb_}^=
zV}iB6p73205SZcj+qR(Iu?DbfNG64H-y*srWO)V6qFXpJ*(8Pk`q?|{mu?8&3L|_p
z2Z0uSC~@nwmOhv9JMXcy1+=NBIvy~}Xqa{uoL$ulKJxUMK>5Ra`hH?GS(r!(00E5#
zz&LS9Q?o@Tm5kM%d!}QDhg;>_77a{<UH@%?Q4}3XidYVv0W&Odf(caR*{Z(xF{>?A
zXS-ZIq96`kx$#ncL6;oR--2x-SH==TsjZo_pA<W=36~I#5}R9g*{YHW%5DW_e8c?A
z$8$RhR2_H=OanHuZbg4qQY&20Sg}0;s$^ce3iBS_pQN2@*xr?ws(Y%M;DcHO#{<Qf
zg0U4xfth5D8S{P#3kdEwAuxDN(n%H<vS?VPRd`qMtTx~P)URn+NjIzJ&-*$c%A_}T
zq?3_?I-pqLs?S{lUp1~?^H5q9+3y1tj-Q(#9^9)6X&a`+o*C`R<V~oIntORvBErVi
zC89XB{y4Dl^g7qofHJy=ICx6(DW0pnflPu(d~4i6@0Q5=CK<cJ$5bs8hRKt2T83Ze
zwj36vBAtZxgOSZ=sW+H*uF90vS23MPo_!2$Q8?}J+Q+z#AQCF4hbv()gZuTg%0z@8
zMq1-E>)ExW&OL++t(p$A>hgJUZ_I7-C~-?u#6Ca0id8tXs6Gw$e?}9dIJEuwYc@e@
zu5)vhZFLrdRDTa?vu-J?xpymk+S;40r?xgjW@BpgdI{-x`&N8vw@D?~GwA?1wRbgs
z`RtEzo3PeL61bYQLmJPn`B>wG_}eGhhAN>@I;kFunvkBmas3{e>_E!ja+Y=Iz7=2i
zq;}=w$JQr&j1?{*D-Y91hLb`Bl~~{DWdSYK+XEGcR^CuEfwx;DYg}$>pX^j0Qz~DE
ze?ECDz>e2~CutQCgPz~>8Xd2szOo^-hVGUH$4*W73K#nND;x9|e@)LGoe*1fNLN3x
zZ>!k8Y?z<*&g@R7m|jTq62LvU==W#REI%>n|F1YamkiN-2MsEbao<5#8~<fJmj6hn
z<zM;x{|XbCpJ9-z&=4Zfu;6L@1UUFnl>=$X@tYisKbFSgAFE^@rQvT}%%URuM9-kD
zj;&SGCba&*JV?{Q`$RY6bB3EX_Z2^ZHe_BH{<fIri<>Re3A`;?(Y@I2*bq}n`GNl{
zs{)Y`w7h<5(KMwfXDL3>HhD5PCxI^f+s?wKX#&fm0RBr?YT>eKpQndKUOhVt=laN&
zP0q$fQV|i&Pa}YeBUd~?zc^Ni21~6^%GuMBT2j=%%~d|@E=KgKY6GfN$HByHE;}@M
z;Sx~GoIaXLdKytpe@UdN@r2LWX(x!zkdO5AA!YM^PL_HDs=wm1P;OlpvjQ|Dz^9ad
zjz1=`)^{3wa>?>6jMftFsa}WgQsH)D_CA;LIgQ)zt;Wj73+Js_nXFeZ^M^-)aw`c0
zfUP5ScAQ+vet=9CNF4z3yf-nOpdLQdFr{xs9ETfe)(l8MXzlA1lAeGn6TdcV2hiBO
za&CDPl*Y(uP0F&Khw172$l_riIqCb>wvwqoPO|q&)OT3d)RqoZgnGmIqs7>y-EN=w
zlCAHjqDaXh0Q=LEZ^LTk&#XA2xgG$AWHSoOCb-AHKuSNvby&>CSXTE7_3ur5I@R_#
zspGyHzudRED)OHW9hWAn2lxb%DTnxVrgi5TeXo#@iOV_d62+XG@f&_1*<9L&G7vR?
z8dIk9OCRtPYoRxF%t^fOs$brVo>-X?ZRmZi!7mWTD)*5KF*#20)hqWf$s7P^MgWL~
z=VuUvC57Leo!xW-WHU(z_}hk6=`NT)igTHyyQS$_9@e&dOY!o%;%hdWAYVWe@-g*|
zxpEj8cs=w9v5leYKJ3_o9%i%E4fo}6VQN&Uebef@9K!R**Wg8va%k$-=>ecJengur
zuR2=UvrhvQs)-c*Kn9lyHA{neZ;LwvLa}0-ZLi^NGfgrl$-*ygwU%_`+(HwcXWumQ
zL)~Xrw`5PhnEDXk^J(8X(>u=;u6Mh#v7tdrV`+&)hB3~&a-s}7B5bNt`(b{LJyH(n
zmHp{yof?Z9Q_{A_y9S>DTOMOVcu6RX#Fb`AEbhHrjdqs!%!7I*?6LB$c)RWHHF!sn
zY!A(X=_|GtX;a};*HdLFmbvGLOY5eDKXHzJ;zF{6%%z^`M7(MI-D@BD`zz3-FWOIh
z()PM7sx9=q{gM2<k03YDQ@sL#KYN?<P6N17-#>d|zaUfp<h;MYN`D&k-=Fc`4`k}^
zCQ#miR4pJN#{vEh0s#!vnZY0E*8dGx@w2Mz;7=Di)J?h*XrK*9ph-(txilY!d)Hp9
zaI-!LY&+LEGLLyK2DY2p<EsK|Y&-r_H~tSUVK<Ebv1sXE`fZo@FDn=RRodTQ`QHBm
zFM-DpdUtjUE}+dGAW~BNpscL2x5zGK4%?*42zAa*ev8is@x**QQjggLpCVeK+onKn
zC8ws;C8SKN10H-x(R4W)dcG?C-1`g9N36sA#7BCm-a$@R&{;P<EeZ0G&r`&W3CY|r
zK{s!<!#C8~VApbnF*X;98KZXNa~uPkAwl~_tvwWrmO_1J=H$h7P1tylH>2OHvq_z#
zxWpY5cr^0qruI=ES%PPi$QgJ{t4swt<-!0mTz78T#1U<T#(e2{LTh@XbVHXWI|9RX
zX;<!i9_TVidPfw`Sv1Y{f|+4CZa2{*H*a!$>CB>M>D%z2ha;s|IdwGJcrad6Drxi?
ztH145MS>J$`pIl?Sv`VqS^uG5E9JS>0iOZ2FK4b>n}_$ymR&4XF?n7z_N&57H&AJc
zq+FVmBf`GfUlW?E%yvxlsCTg15VYFRl&vXNM(uFl`+V^-YY^xI&x;t%%ALB#XjI9!
z%4(gGWfIB;#YBL$tZJGK+^xrNs`kl&;l@Zj{N02jv*v7hZYpf&cIb^0VHa!O2!M|b
zG8ZKw!Z<1G%ZEiS5UK<DD*DnQmi1~P4Xf=}9z<!X@_gYsoq3r@{652RrkQg1TapT~
z55H`TA5JD*Qb7f*S>psata?lbDn1vshIhV#&yGD1^utcG2GxsF#3tx}n3I5jmL^$<
zw0N;O$O%Ul8A!nPn<0`qcG*`x3whtFmKHo3w8;y1<_+)-Vbh-Q3;B|NWe1h=tOTL?
zR7BR1d8(iiz-`G}G%}JRvn0SF&u9tYKm_(e9}<$bgrUb6_8jQ8>-j+;^2s_@n+754
zf(z-c(H1IV%Jxf~*LvL&N*<twx0bIwvI_CeTN9oMFPe!Rz34>&zFL3jYRvOiHP#F0
zE<i=P!omngNm3)A1AK&vg~P?*c8a#HSv}Tyj^z$@F%}kD99fI!TjR#F9BP%+tTAcN
zvo=%DOA17CET(d7Z{wdKQEUzPp1?<s8mqQC+-CACp8Ag_N<_@+3{c&BrIF`F7akGT
z+<(f#;5&#H)-XARpD8Q%TqRZ_>h|}!%m7a&@GFv+`};9Cw3SUHs92)bCzR-xe2YD|
zB1TY3<C2w5S>d%2@RhCN5G4Jl&Gwmf45bo*B@ii81b0=$?*<506f8GDlxzCs_V`jJ
zidHX%*!OmR2;#F^I0>E0Gvtks-$-s4JODa+knmz;s>QfFbboAXY_q&hnNdgW!M~t+
za?FC?uoPWUFRy(ESv}}BHljN>ywPtE$BJ`58GY$JckCO#bLSZpnO+dg$m*~##0s4k
zAgaoV6d)g8KDYr#?>G?YVQKK0mUy;^Y;RGNHhu6JGp(awAsFNHGhe=Fb_8rx$IM1H
z3Zj;Oe}v~ZB7Z|kA#?ws;_wH_gHwjT2{!zDkFfw=5(Y1qkURg$sPd0{3_qfG9tOY^
zQj<h|eO>8dV~hQv61&o;&Rowla~E7`_rFKMlLBI-82Y6$mm5HOkVYKSjy`Cvn@lb-
zQRVhubL+}myJP}7nr_2G)qJJh{PePUY7-<hQM;|@xH?YjRF`f=KY2)~P|We3pvsJ$
zxZr(p+vB`0k6prO&ZW`c`K|GpmG@f+Er#H7BN12=qVNplfPN=F#?=9~gOKg)Np)pp
zuXcjf&#krH7>f{31T9}K)@<Wc08|?ES|@4x6q<KRi0ZKU(gW4x@|8)8EHwMhs_f>s
z)ixFBhjRBj(n*GR?GQ;o@K=bvO@^wd=<uy2!fs^gwU+C*^HvZc<Eoy7XOFC-dL+Ck
z28XpnkOJB8F~!<G0Ks0h7|(n1gI6TyYD96P7=sF{KHl3<0(=g6><P}hgsyc4aY`L0
zU}88qqo!p#KGA&XnzCx(vu&#{`4p_Y&)1EF$h`1do91<W_X&o<Ql$<@tEJg|o`wlM
zIVpN0!i4kMSgg(`L#u1&?oapk6>M>tad0rT6Jay_&*dIwF4Z*#l3cM3o*Mo4I`XC*
z>KYnanC%%$9g@!%T053@oo<T3RKw)-G8Q*^q)i+AE6PkYpG9@M1a&)cPKn(=J|{7m
zl$thKD;EOOYN;F<$fymc#KpgqGs4j}P~Rqayex#qy$%fx7uOz4PnkRSS(gK+BAD-)
z%VRyqnNj!Dss^x52>3rIBIbdu^ubb%Cw;)jmKS$ThjS60oXOb>JG{r*(mnY+Zu^hw
zzzsd@E?m7-I<3oQB(pEj?xRfTG6pn#CcIn0%v&7PO;3eWoy*O-Gl=mAHz(9L?Z#<C
zz3*r|mb$t|x9L*Dx*|lD@KLC%?&l9}w`7OA2LwyZTUd^V4#!ShG`{;CG|2-0W<dL{
z>>BHZA+u6x!0{W>yOLK7#zwXnqqu#p4<OCt%j|nCR0YeFc|W@QTSxWAa2uR!dU*X@
zw<<vz7nh2k!G4<*WRR+K?slSkXQ<0J_TJepg8eZMzhkD{`*?l|b-68Z#4;L&%&f>j
zrkWD7XCM?rX+yM}?Q}PDN0GDO{i?{@1jg#VyTZb{RO1^aAN$H+hHY&pPf4G8(kR_J
zMGt*OH|*X+MT}`m;bD8!XcujSo(X9a2+P~Ztqr!@NSZtUh@V&GM((&-ayJt;h3Bvk
zDS({{lZp4U8zKofQTb<wYg#l6A~op>cUoU%DIKd-nwQzT`I!SZQr7m&>7!8^NtD=t
zg@mTk7y3NJc3bvu-ma}S`4;7D<!V1^THks@)mEM1+GI9L@79p3+tjQ_3Gv!f%~+tA
zOaVr9+k5C6iO;}-F#OL+)ZVz|gO^V-$3DiW3|~)J3-dQThXr9LRowJJq*inaGzDK0
zs=)8~bTaAi*Nnr=r}x*4qkVqNH|pZV>%1m`;9KRS>_O&SC(~2$@|*zmnBwPTed*r{
z#DBR0_Fq#VE)mDC`>*;q;i)z7t6MQjwgqrI_T%5GFQ}DvIi3!QrlAYC{`r!8_*2^_
z5L)oJcp{kmaTu}BeVMY|=Mw&G>%Ma(?(??AGuwgfE+H$-jA5WfZOYn)SwaSLypW?+
z#N?#0Q07KC{rSs4qO5hm{wW_=MJTpX>7gl{gGkx@g&1>FtyAfC!#MLV8^<qA><$J~
z89z3b^xpcZyHg?$fFgwo;O*1ohTH|iDd*u!QoF-d-mzL=;&xAu(<oXovTRiOsewLp
zl-&BqYLU(BCRsb(56{;QhcKbePuN>>5QI~+iH^_)BYhNs`K4D@p%d@Bsy?%}Y8@53
zs6-DdtMQPFH}0EA18IpLifx&bgXno6rJMt|o7WhUm~dQjx(&#|h&9BgK0wTMTpxQo
zGz;HlZZppHF+xk|@MIr)ji0R~4M1K(3LdsdvD^a+7AMI^;4FH0sp<za^e2hcYN0wc
z4SlwL!;_1=?lpA*gh3_O`O>Iz%FittCjVc1-yPM|nzxGw6qP1Lst^?v1p%c>jg2M;
zBcODMij)uoA{_!z5Rk4Qa0DdMrABI~iF5%0X+c6w=slqZ2=VSS^UZuS=X_;m?z(ra
z``z<b*4lfsdG}j?<@r6&^FyzeE6^c(fWkfa)pn(rP*qJWj!NIFXT1J)$sO+Z#4qca
z!SQgf!XPWB;d?wM8%9#EU|ha2j?f_2srR(RwFwoPrej%g!&^Pv&T3hvv8SchSX4~f
zPs!eYr=pX8qxQlX^Wcg57ZrCI{fJ$gb<JA;$mwNCs;{))LeCH$kN0$+i4C{ZUTa%1
zXig3bs#7ie-iqg{z;kxFVVs846eetJUKwVWo*Q*jFLY2f$BjkKJ{P(4+1@9<c?W+&
z>AR;D<d{$H?cVa95XR}UTKrJCLA0t}-D`8~aOKyCmY~))pCih1cDS~Yt(kiS#S}x7
z9<FbVys_;1z#sk!dvZb~&(u2T=og8{rwB@CXYSmMg`Ha*03d-XJ96qvcaWvQNxA@x
z7)W_ZQ`fN_6Kr&}u<pJlK-EmqsvzZ1kRX>qIXOk&Ld2z_?o^6Wxe9xbbf!lCt%)jc
z<Fk&BjtcGd<u9n&6o+LGFH~Of<u`qrUB~SrQ>u73UtV95iBrGJvbd(qBsHt1lChDP
zYJEl!7cN)JI0dAJn+N`*{IK*Fl#?9Sl8oCVSa?3!u<ext#k8DQ)E#YyT`}`2wY0YE
znm%$`<0Dq=!3~{*kn5w@wCc5m=fCaemKg>twgu5`sUK)v6alhBJPE={%O~&eADPtU
z&n0z^$;>6ZscbrK)53duj&=2j=%LfEIRGn+m>UUpa-|ub>Gizev7jM<fWrdvVJiH_
zx{@O<>~Hwe6g`kE=ggVQ^IDHe66Z5vIyM+up5*Sr!a!1?wf_&MR48MFhSsD>k}%;F
zJ;RQ-DF?VR3{BE9hOnu@??njLm`Qv3sRNt19-=dvE5%)d{A}7CNdnPK(TCy<X~8P5
z80TT778b+M`5`r>oUCnXX&c|0bM*<OiqtdeDKiP%>8C5DiSr+g^lyW^;1eBynAQ)b
z&Z4D4==uZ|yNj{}@jKgc4<hv`J7mcaq-M5Hr#0=q^t{nRvUilnKFRykgbSo{q(OB0
z(aH+{;gGTm_QsEoj@%F01o-#sxQeZPM#vM}x?z~?l7TuUQh7*Wp-pCWRhUurc>evT
zt8e(e2uantzlhi?jp$uE@U_}w{8FRuO;>_Itl&%Q8i8QW#dxF_Eyb%Q;7civ5)%)D
zd-R?Rg(ccRkj`^6mooI)g(lVJgid~?aW$#UHFFKCH0~Wq$@js0_FGN*a5H<12xPKv
z!rWw<O&>{OJqKj01tZa0B}|-J7FUw@gd@VWZ}SEd^#cU}C@h+^fpc6DH*&#u0BC#!
zG`4fj5I*S|Wp|L~@gBtB;jn5MKi@kC*pD^}7Ro1HD82l4=ZRh1_|bc;m)~-oLGm0!
zJ|1}HlfvLvX~l%8wS|_ig$v6p3&f$h&6oK72=8;W9WJ@t6K|6pIrdn9i|MQO@GJJh
zj_k1BP1Ekox1(s&iLNZX1Q6~H!(EApQWbE4CgKxDY|vkT>HPK6AFO!4a}j^~ud-c#
zbMK#G#rv1Rp4E^@5O)!dk0!j)?5%9#taGtw<w5z21ok1`?$<X$zK3Yb-+o?A(t%uU
zSo~Y>D0?>2anc3?sH$_>&~mS*>LLQNhhdO>jS~H{nY_*sJyNnOUr(B^7+<@Zm#~t%
z#jveE9<!H#R3}gHeuvcn5vOJwP!o>xQ$I;S{BA$}#ts2@m)#!rIGHfg@tu!HPdc=F
z^=N?~YYT>%hSA3kdO(C|DD%EmIWxo19mwG6G$Gj(kq2hFQueBX*St@~NY56j+@*$v
zAmyfLI5Ak7^dJsoB^y_BOY3T{-0JG)u3;+x^(nQqd0Af_%7&79tUI><*;M{BA742@
zVI)C)`-4eD4k6OAS2Wm-P6s&WhBO{#v}9XD7x^J>islj0R*%x{V=jD&J9EBI;jL3U
z*4|f>2a7RSz!;2V`m6IgQi?{fNl7jk+p#AB_W87>CO<yE+C|&beVd!TlLqhgi!qO0
znK$lN>MPHRSymF7ln)Wz`m!jPC`FLy{OctxVZ!FxiTRC0^9l~BRrLbj<0YnS5|8s5
zSWbx@Ra!TTpmR)tczw8<CQi5@Gz`-LZO1XBN$~<v`Tfc661zv0Yw$;hvzvLs#h5C~
zG_*`=u)loC9)ypA9Z?+LVU>Z>gyjl|;1VyKm@ztL>(iqft6!CN4;vwOMJBOtAb;hS
zP0xu-c9Ieg5;L}p$2>Or=|tR+rAn`I`OGj7!U58N`@@S#7L2n<zY0L_^JZfpcffgr
zrGWcWHjsNUh`g+0)0c-8!^>BBN2PD6is-d>c>$z{NhB_oMWSgE01R<T=FBYsC*)cz
z`b8=C+iK0&e9HVJFZwm}IM*dszun3;5~u|fRpZ)X=saQFry+?Tcea&(DgkN$J<n$&
z5RY5cRe7M5jc32U+BsIkFXb@m%^YFh7NMbE0LSrCV-iRhUceqYvDGE5-@moSwq(mw
zouXI&LiXH70*2+`lbn#ODo@{YN6j8y(XT$59Jn5lxPqzCFR#B}xU(tcr)rfun`}5y
zh650zd7E-E3?RVX$VQH3hLtT$IDMij=bew+VFdP+bZx^NBBje$IUim*n7k>!)pKg?
zCEkfcr3EZRM@@2lQp%;DO(g7qh|$XPYbvK<gC<UGakl3kv9Y(u9Bnb&H}G*arm{91
zSOy#?kawEZBr5l|lU!kFHF%0@9O>+}L+$aMSEf#_wfXJVwm~_OerQ%CH^ymPbdWU)
z>3BXbn-<C!!WhiCS&UEa|2bd&a6xuW+*9s*sqAiS@5W@dpvv_<^yHoOPn-Q%=&g|?
zh5_7nbWa)pRFtEGW$V~T9H<t?X#nc>9HU|6ge3+h6~XT|UHVjgFbq-YbUbHq?2wOK
zj9=P>EiH`XdVJ9|*`<tZ1JWOGm5305t1r<$B>SFoFn`1S3%}#zrytl$Oqs7LO(q-G
z=Qf}Pk&r(6&B<tYn@?1x@~|68+LX<`Q~|#yrrcKNlh@c?D#@y*fwJsw<mpd#?~aes
zcqw>mK%u@)Rzo*eUpYM!snoaWJwPv$(0>-+Id>!8j{P%BH>Y01yz_K_;A#(0MDn(1
zvD(;w@jy=aBUWaxe76PKZ%QD_S$Q)=4$_eqJt>_{0Nz{&ljU>8vr|>>blISiUgkMF
z<8>C8vMVvu?KC^*ox^|zCo9TXf`;XKv0AiQf6AOzv1Umyl(ZEN9&Tgd{>n3b^seqx
zWEbYB!}=GHx+60XuC+(A($FN!;oN8>xWE~kCeh32$%Mmmx_dITg`<gv&DXhrQaOcV
zF@p)egq?}IJm%PhVx>j&G@;6Fab!amcKpmjTx2ko{<_fR-R}3trMizZR<J?Jm7Ld4
zs+3D=t{^l<5{8Q6Y$|GM%Dm65UC2dvulGX@13U=b3;xl$z&^MrrkK*OjgzMw^hBnQ
zN_;yrcpAP*tWEJaL~jV0p7A+Fh{i|L)jb_^Y??3&?v~C;N4gt**2Evzrn-JLy8J{v
zRu+=*EKp}E7%Jmqb{mhQDFJZ@bBaTK4*kYArD&-&BV0|dWK8R^6^wY=&Rv_Ej>Fjs
z`Q1&O30dy$p?80N;f^$<<hIEBpj(ysaiGD+eSxBjDq@!@brX&<V%dQ5@?lEE_taIN
zuio3!hZ@J8A3Bd2yVJE$TrX^0yfmm{KiC%7U-TZkGHAO1or8~L;yTur5n%IAgUKsT
zToG%+-R_8ImFtC}LeE)RX<%ul4zIMQO@yMMmvkknX*#8el!lfqnVjHvwwx__m`7{6
z8*JF$_4Rf4dSqzHlZ2KAfuF9D+q4X#qAQlfgs84`i6P=3c3T#cjdHNob)9qe!u{;+
zPl~ajoO(a6(e;uVrfWN#+p+DRjhP0vj*a;>*E{Ht=*0^eS${gE^We!dVWA#2H_ZlR
zGfVmv0hYGn;+(5`kF+nXwtc%GNhzpk0=Vr_Zd5=dBN{0Ra}&)Tug<dmeqCeudz`+A
zRN`kz7rpL7SG#VF7$H#TNgKt~zQi_75;!7Fl`7#ScDjzj`JS|P$R@8!g@}1oS8qB#
z^w7R4xo5eRkA<lvaKOT-^x6YEUK;Cm8qJO>(CBR|-j8%lZ^*H<>2Z9zV_B5HD<^VW
ze#ZVm{sK!}y5ww>ORGXCU6VGBJIiSQwgo;yYq~}?;VK*dK1K^%TlOkFeO5>KdF*nW
z?s{BdJkRIiBU9x*Yhg&0fI^zxXNEAIf-BoDVkUdetY&ho#y)+0iOW6DkfZT-`%EV{
z7n^JJyL&aEr&CQ!4c_UfD<{xX<M0j%dluF$DxF>D9O4naAKOQ22Ca${w*aX0BqSmc
zd_u#BY@=epzhmi=AR0v=R%8=<I5xd6`*zIVvvau;Boxe?CS6!DxQt0$DRLpz^2aGE
zM0)cGX5D2IKin<&la@)(**bg^AU<2Hb&>wiKFRpo*+Jf)4P+dMEThMXCt;@*bg1g$
zC|{fFlfiu-k=!(Wa!=UaP(%fXdz;pILQ%12WWg(s5n4I=JXKVj?P=?HpDO>2d(r8=
zK0**{jQ6=x{aX59(kP&PgYehZ99yAjqA=^Y*hWmGtqd&>o<WQ2Z>X_BHyt=ASCtK*
zp^gItp2A0QR#)INIphx~t-|cb<>du|!nC7-T0T&6YeJTUTICyia>1CUr%;`2a+b-2
z)h)fIay$6SOyHyE$R<KvSVO5LC~{n2K#R`@)g~rTI#}#jO$~if<fImwl%dN!QUV*4
z%`i7!TM^R(%6`FhW71k%hg%?F6C7J|>@?7}y#a&`AU{cVsCC!M&&w^<cPjf1_3Z(7
z{?hb?x+)V@a)(yLGYD-*t|9g?ia7n~g|6w8{dljfKEeiMsFoJJZvy+l#OVQH%^n*H
zLO0{21|FC7ujEkuuIIcm={iqZ6P^eh)yZ4`{MqnqTm0*`a+)I*z|l<b({GXyY)!Uj
z`YU9tnsyurs|4cSvG>pQzIO??F1McU^Y#A%Qd66XScm*XyJ=fSK`$?{@14(NuYtZ=
zo5RN#=$|YShl?8CpJ54HG0+!0{OOpdXp0w$<H;5a726uidAnvO6tgAsbkR$AK=|S1
z6z@-E;SND$^T(CG{I}d@DrXNLd)y-BH<40?Qg7{Cokaz<JLSVVrU(<UK*{rk>_sdN
z*90#ZhRTdsjx~A7NH{t2uS#6e`@pZmmU@{5nVY8Vo79Mo6qBJ9xn}sp{M*<%E2s4s
z2Lzx^%x{-O=s#)s80~#h=zg`A&6N0==Up^$N_-2NYYTJIS$Hl6X;@@sy(D0EVgXBq
zX&I8$bofcp^;yp@z04E2nL&|Hs}l&C2JliL4J5=(5+sc7MuN*&F8ZSdozw`yA7OqH
zPc_)P)UpQ*j@uk$wu9y!cT%z)$3Ik^DTv<7IuNlz6{!_@oTGHM<C7s90qUC@duhLV
z27h4TEE9hRi}**$S<n7nTPEKOwAe-{{>SoE|1cIVF7vN3!jD~~C5vodVXMjRmRE*1
zJWuU0(U`yAJ(6%T0oX0N(DdyAFvZ>tK!;rb5=ag`41Z@3O&)0I)Z0H%w>2s_0;%W<
zF1nC4irN7t2?3KTjTpvYF9y?Q{)DliL;==$58@_Jvug&TKZ~X50AfXRl>j0xSUQ%W
z`X`(XU0fNpF#?qIHYBwDU^<n_;7$SWu6zWfNT&dHzCQtN{%FnroWA*Oh64z2{@v6{
z0N_a9Rtg!yG@R3PouiHL&QDguS+)<{I?7(7xc}gI`;XI%+D!k9(raBtTPpz%#8VwL
zRayi-b$YsH(VWMZM;c<~pH+64D)xxH`;@8q-K=2ft%c9|A+qlBOb*R<-52p-uujR<
z$>l-YyF+k$HYvyGFW(=gEnZ+&YXg8y(e&#%IR#|b2(|5SvTFR-eNbpzpQg&4N8mx7
z`0L_#_$amxkD7T{9)t*k+3EV2#k)&>k`Ys33?7-Ec9u<Si8p$!7B81xdTrDEvhv$J
z$-?_YsjgJejbLlmW+A#em6^2X17xwq2#W9}eLKd$V4cd$*zn@hS7XQ#502S4v%kDA
zn!fi`<*+<r)C2f7by28cpC{Zx1MEq{u#J>IT~PLWYcF^1bak@ti-yse_$Y*6BSqd<
z)T<z0qecrqG#Ah<5FXc@e=1oCzt!KpKkwi57cS>HaCuT>h1?<JF>3BB#Q?~@h;je2
zu#+dp8~kwhHM#KnuP1o|?ZpSoGOd8^@=xNm*1B!3EY2t%trN7|CN%}3ZvLWasE7kD
zAu~8|QfK>~mr>i@!hmRxP6ux>UlP!LVw@C}APoq4lh%WxqV@WP3F=jwyB?9dPkt_a
z^(`Mppv0Nuyh<M35DnH4>udt{%8a4dh~<d}+=YIPo22=N)$vU`Ppy<@rWL~0C7!pa
zH=F9jXzxzn$QB9wtYD72OzCOHN8k<*dWia8qvVFp_@}T%T9i)!g(lLH@xFY0tXvO|
z^u#HYM-NM~I?a~gWRa{?*Qh$1sc86(!J<nhmc1S}T7rfLU8(U0F?;@zk-iQg-c5Ye
z+h;e#FMU?0zkCWC|G`8^cHIF~2{l>y3MXu|{4jlMGyO(r9=y{?3N!?CLT<&Q$o(_!
z66IrQ@*0GIC@Pi}5Arp6+`jEAdnUo=R)?~}<;M@Moi?2<qi&AEa58-<O+X<~5Yhln
zTIz!9F~);N1Srifh0@n&-r5T`TV4#(%eXa~n(CW?Sw@qsQ0(;kmA&P$hmcw+Ed;-x
zExi&*AEC=?NJ5$3x<>YcupXIeE1&3N-HlAwR>lAUvB1+v-JuE_c*_#0OW5C)ZBY!=
zTj*i6yE>H!?@IF=Y>u;P?u{M2R=`uR^7Un{=m$F+{Z?mC2ZLdTMJpxs%n4@<X;^R9
z>=)zNpsQ<Hl|>9ygB%}JBjG8{-G(|Bv8E`3c(m8nwBQtXo$Z4dGL-Ce_U=0C@rai1
z1uHkWUv0Xw`e1*i&V_iOxt;nex)buP&W)AFis-bQdhBA#RKOXc$8zHOv{^S!>fr5c
zN{gC6IM)lD82#sJDEs$$;n1Wh`-`G=iM0{dmewK$am>D@(yfndM(hwxFq2J4(54_=
zjwA}wkA?IrCX`jg@Y-d3K3)1ED7uryN5qa*T}SOvt<<w>!?s#%_QZngl6DW|ozAgs
zXaHaYi2Pu(tWUm>QKIyp+A-_si1Euqt98C4XqXu9z^}z2o$w;#eQWsRs<<{trIYc^
zpY-?3)Z5*7(mg${*S`ZDf)tqjLcd4z#w?~uQ-)%Y0`9qO<dq|;%ea<8h3v#^yQjTE
zE7OTejv?5ayjl6N<vGVM+lA!H&yHY5S7=0r8AuoLg9+2Oy;0JE)Tf3;;~mk*7vitG
zVUSiy_o$rcN9Xa_zR9YJ2pBnh>MO!gL}2um<3r-ge(P|*ZVMEY^9G4XcLG?}IdIja
zgYyQ!Fz79J6(>lWALM|pM9qDDbf>H25?(%){tFO{5vA#p&6;DwmPEJTlWaiEKw(?X
zB0(5|D{}XAlqZg9t4m-V2r4(vx85}!t9%-WoU{lgxxSK3r&TLWEmZCVDprKcqz94p
z?Xa(5?N3Of%~>Dc+!2($J-Na<BV;<aeUgV(fn_j@Sk0l-w{cj5q>=0RFoV4fbp}5c
zVkOFfZq^1OVhz!ft-@<)ADsl4w-+K4BCq>KeDubQSt4@kg*mi5EV=^!lg~|kT)(S}
zqNW(a_^Q$xx5Y49`zAu8O<c22=gUG<q0?gzkLbsa_Gk1@YPjty`q+_^8UuKn#H4!_
zMGY+a6{{{Gw%{wW(zE6E=FOqJr`gk^d?o=VqXnCjYcD;yLT!cmHz!`&a#B2_3%#Sy
zbw?^jXIcslgWOxU3F{&8U)S5e-nu;Y<*L#kG+`h<W~grrv#aR-ZT22HyS@VN9|%kH
zB%@y?CpE<$p@Rp@F=C5g-pzXdmb+HEA4iDat$R>#nYt0g6x`e{<VCX5*=u>_pm&W0
zN*-`(Cxq&dl*<c>&=APiC3m7Y#}`~(9}PVnJIV9PQKE}GNbj9sJ6qz&*%f=$HaQY_
zc&<4)U<HTqP~_ktFSWtROZBWe$-19PqAcpm3V8(W<U6Iwj&`T6mB+T~%=J~#*vQA~
z+>4jaZW!`o$2I!vjCh0|kINvI&s2ImEE9X^^ZofTYwkqWh;EVniLdC4gTbL>+aFBv
z75fwVPj>yZ=@V-*KbR^RHnEyift8a~(Df?X0{InPj;#L#$=?PnK8qwBMwbqM<;FpV
zm(?4O^}tOW&$Tp@55u#eY?}`0R3RjoGDEj1p0~`2_|#lF3Nt%0Zcz1%^b=aj@m1RO
z-N)jr#XcA?RpoQNt1ZQ`4I!p}pfx_#=51^Zi!1TcQ^+BXTu4Q*1g?*TkJQeb-kOBb
z^=MpK^Z=SK!k)_9m?i+_lQ?p9HM$gjEH`e^+^S3QiYhy|dkCaWNX|*yv)pwFfjW`Y
zN*saqO}ud3eo5Hs5^oX{#Znlxo}}6?v*u)n1ldeGj=Hio_Wqag52_-}cTQ?ra7rJm
z0144Sw0D#B(x%jaxmctmc_hyRqi$JR9(K1MlQ<54HTJkX^k^EZp2Ur4{w+kIgw*ZG
z32v?9MX}GzWk4iUC|x1_J*Zh(ZQAl0KstE;WscCxIN2p1iSJ#pb>c35gVA!Iaq`r*
z7$lhN3TjD*@sFSK&FR<B{4mz}QQW3oC~_rktdpH3Ax8hrL-T7Qok`E`-_3=q|6o$s
zueRG8CoC0wGp_$tT#%l;4*a;zq%Q%c)R{jdJY@c5i2grga{i{2Ie|Fz)|^HhBH@?k
za%dUbLM}y{Xb!8f6rU@zcT9(Gx4xTlSLSo9J<hbLuO^(U<XaKj031iOU_(Q*<_SK}
z72ald$#>gej&oUlE~(9^{OI#RI;j^Y>%O+Uk~rFMTrp6H77rarCu~AawsFxVEDx`J
z&#5mD$-oXgd~9*LHvAYOy|=M+^~~d#`z#?xo@*-qwOH(L9OSy4?I<8xKiLx`c9~Ev
zvfZ~8seX2Q=uOM1XtqS7s!I`H^m?cl#RgdO<_FAzeL5hg=*Q4aj5A4X#N60A1^0sC
z1+~@`vNRh@n&4`&I5vVzu#)x<D2%%u(s%#08jpY-`+F44W?M`w@}5%$RX0jWz2~I0
zlW<`8Ddmf*RT1J6uSZIp>`oNA8Jyn_S4q7*dZ#7bf5ex;38u%b58<A#R-g#cz-%tu
zmXhuQ>aOGOQpSTH=1Vq#g`GQP;-TZj?=6qc_Z#_THCmNhNRV)P>{%;Q`Kwb~kk}nK
z)m5jQvYk@sGjfAa2NJ8zh53mVvPrZQGi`hwk~^-WQ3*}dH&boQbGZl|H(EnA=rc5m
z$J0J+ngv254FH;Fv#m|d*$8ShcnV5C&$vu81BVs6$au*3IQK4E2F%o)Yc?UaGGD)Z
z`j_d5yNeB0Lad2`N|~eyqO=~BzdmC`m$VYt)3mqMM084%s)-nd)ueXBknGqL?z|ve
zyshym<0>Q}W^<G=B)4^-^?r{n8#whBJXZ#`oPQPH;WPU0p<yRyz$fb-V#a(n8u6SO
zl6mKY3ziaxEVH+(2<>j?V$e^%WrecD?>KnJk<Z;&oDcSw*ZL{VoeYW6K<sPK%zjR0
zi}ojv*5tnO*ymO{z$G1SDKvrbLB#CfSHI)P;_943uH(wx0H6tRZqB=NcVYd(CO)oq
zy%ayQn>p@;5ulp2Eo4DOoGj%+HEy~V`=IpW7WePQsVl&<Zxg_<u&;Ki8-np=PYmR~
z=7P_k+Nwlx3jmh=5k<5h92Y<mzMnldTp6v=E_N%{jV-~}Vk71-+xRTjAW&n19!Y08
z7`NKOlUz4+y|sP4SB@1eCF_Q2sJoI5yq220lQH3nT~anWts=gTF+#b=&NsO6m8?qn
zm7X%?b=Fd#fvE}&ex^m_eL!In(t<T!u=T32hjFIoLS^cPxNF*B?A#ihuk}KNWfn)w
z_3kCG1gs!eHcC<XyD_#)*aU_dYMP;o1)#4?|G(*@c5|>A66#=iKw-)P0n0E&!)6Zh
z563$^uI^T0j+zX-Q4R=3vQUN&*obC~L)rkEV89Yek4eTK1TO>H(Ds8Kiig60pq6AP
ze%`hZPdP5fI3G*9t^okAZ)^i*_j4l@RjFE==^>EiJx0Rp36x{#KSct^VkYARgah@h
zJ{vX32c;Xj7+n)YQG$HnjF1&+pzbW=&+O_id7Ge{axf_EFo40~wR(H^2#9_<O&DP8
z^uG-R$uHn($>6_l9$|*>j|TnU>Y-CFzIluZv?e=ld%cH0oL$%nbR2$QzC$*6;^-vQ
z3Ymptzo7>H%*zi?90X5$^nsT_046h22A=oVJ5J1Z{({D-%jj%S#`Az3WmFn>;yd9-
zibw_iH{FkDd||V?U}2)+qSM5_V`Pi5H6sDYNCqqANTg#PP0pGldx2sFk>L4WlNh69
zD$R(miVBWVSh*&4ol4jt>&9pS#D<Nk7$1;jl5zj)*?@<2G6BL?@XAJ2zKYD{#9Ido
zA;e=3Ii7c9f1ZFGMA0RSe4y8<N2|U8bURS4pD6)hx2fkQ(<Xmr$074sPAuyj;SNfD
zE2%WrzJJ#TdBub@!TZxy@cdJN%yIF>4<;Cf3rz21oCj^sZ~;}W7&Y??o%|)%-3yjP
zR1>`xZ3NNEgPc!;EDv4E4usdPrVY*D5@#z`(pOFY!V_!8O&w_vMmNX4)VfINj@o{+
zSd)+*I*iG7nw+TOu1^qs>oy<_va|L;Rp^FgMw>CXdw4}ZaM3n_#TaXXC*{+D9L<f1
za6E+v#L%=2UuR@Tt{p~RqhUVIfAU%F1k!wz?u3um^9{1+E#K~y*A#xkzWb>i5M^pS
zI+bAeIzKD)#s`<^o2-pgfd<jQsm4HYAbMfTLoHOpM_`mn?28tYltGRW+6Q01jkX}9
zo7_d1Zl0ZfobX1k2ugFtbD_Rr$*y#kUz=m}gNeh8w6@1jHv~3>*S~g>bwTcyH$Sj}
zB<yjnVYV`%`|cVW&G^!_Xr-k9u^wA30a_-ztx%ZGMxCdpn;U<;U(rQ>T9L4j@v5nX
zxQRqGd)Bo>@jKBmkxi(hf#>PU19fb9G+zzrB3$HTB$N%hP?MPH?hp?Vkrv8$FOw}5
zSw1PaK2kj#;|KchTNmI`J3)z~=g{e!=M7Q<2|t+nP`f8-JNBR>jGmhAaJd(fgP!W5
zHx@e(?susypIpQ3YX&;?PgmOH?l`Y@sGRW1kB>8SM;`*v3?-DXdz6`|2@Wm(dVE+@
z3u{av;szXRlrMi{jcb@G=Gk%IxX7f@o3Aqi1xkRD$^y#H0bAgCU~|($0hRQW!9X!!
zw{vZtpdI^lF6ETR$1Uw4c05FO64Gk0J8L^zn)wmbf1t~+`QJZD#{c$xnG9h})5N9>
z4FDb;I`gMu-<5l25kS=IKFyyIg6Bm|N`UEpvA^lGyQ~#Pw!l99+i?Z|X5w8F8`So}
z(3PNJ_<1X;3gagrf715xcc01hKcQ{aR1?xT9zd5i*{V`(y3wG>a>2RMeg03~iIxe`
zO}8Aeapr?*{A28Q;_|ine!0*PBd_620)<0)v_OU4MDGv}6c21~Hl@HER=HQvHJI87
z_vBlS8%r@CV^0yfZsj%Hu)Fa>;J@w6<J5{8b&RAx!*2w!IUrWM)w^3@IpO`ap0#n}
zgw0LrGk9A+VwDbq^H|+H+#J{Ddb&SuxT!i^-{@@WW$T0nB3H+0YTI%zDK-LnY!kxn
zgSqw1Q^xV<zR42oAf~(uliZtN$0PXjbNSFKQ$f{&J=xd)`-z~hEj{e&+PDoYrC*%;
z@CFQJ{B$Jr|FG9=qVl`hjr#j}oiEe2hVO0<)c13#;MHH6aZMPZk8F+D5kF#1{%)53
zZZT#l5DR|GCuSS)`_5>gVpqg7Xz>lnj5i3<8n>?dN7xR3i6Wtdg`HX&)O`Dnyv|b&
zCNFfED?Q!{l<imH>%F`Lm&t8PxiA}@`Bc?ZP-#)|w_IX}I@%Z+pdEo-GwF1gr85n5
zhv11Yxf^sPSJ&Iy*R9C9MVRY?dM!v%>f<y}k6i;SyMLo%*+79%?FF<jM)OGQ!CB{y
z@-ek%$@3{n*<3K=%K?_!9O_dl4IFlND?fb#3XMiamhwymWs2w?5{VXq&p$@UKKrWA
zbg2D&4I7w;t~!7!!?jPPSYcK~+7<WZBwL<=iN%>em@HV#(Q|M7GiJivZ1wj!7oaDG
zwtZ+u?s5?t@{E&=7VN_u$<S5){f_Tm7$R=m<muH;U%{*U892XQ+(jQWnNYp21`m`T
zz*lTt)w-}`dZlaIC%yS{o_YTh=gb$aZt<hZ<JozG+YlTsl^V+!!*1?oN=n%d0WkI5
zOMrV-`Gd&@d;?G7`1FGb9c>06(F9O*uM!QgTrkoipuoZ@O;W9E_6>O0QzQV~9N%;9
z?XY;~JL5-o$O2?haKk1h9LctCiiA;<!j_S4yrUxqa#5TKZlsmlb4>(GKCcWr?lLaT
zX20ik%(I7a8&Acht``fCV#21l!cc+?Fw%pfZVTt7xwh>^1C)jrg$^qjo~n*BeTC>j
z^MbDUPG1zuUoY~QUM@mGf`-<21}6+{zYvYwYAoB;{pj9Ubk)j^Cs?OWohwA+A>7UL
z11%KcLL4WpoKLw!J=xNzEq_YmXX8iAd5gJk#w$?2`AobT*a06EPo$p&m^VnmREKKP
zK1L9!N60Ge*8rzmPB_<;Kh~8@;7f|#wt93jSfN8?L{QNua=Vh&&J~8#CGCR+Y+-=M
zPnEk(`Km_9&RH;7Jm%=lK0|&l(?nFrY{e_u{vz@0VvFjt0_TCi-Dj|Mcc@&##;j%S
zf^R_Y+`(xEN<9!T^czJU0))`Q7zQUmg$C%sQ9{3?dRhyBw0R_+p$<jc%F_0!WuGlw
zTSZ+N{O<`ESIS1Ka9|kXp>O&^eoYi_N2DZC>?&Vy7?WV79joe>5j?qvh?+WrB}1ak
zEJeF!gOFT>#qV=qYqBp~Gkx`%c9K(*Us7!jr-Bkl7d~*8m$2Mx$<sDO!0AN-s2ZnQ
zB}p+XwxBL)jz6Fa@LUjKbh<O1cC{8b{K`?kX|5%Okm+9N<9-`ohcz&$EzrwN9#=KJ
zA3n@1A7*_2xkU8m*KV&3VfQ;^Wm-1R${PbjOTg#;<|V)RW~;aN-u?p*{p#wRsRV6M
z^8tJpdF@wuzu&aB;@|wL|FVbw>@)rGqW@|Z{?SK&^3VU6NB{e4ja6lGKis8GJi|~i
zR<Y0g^O{7I#k(kHeTglJULzWaajTx0(EwOMdooTRj&0(1MU_zdtN^lTzB7Lwp4JBx
z8hH!?HR~Wpz)6-ZNdK3MS~&gh0U+*vJwRzSV2%-0QJi}nAhH1fP&TOT`^8rO#mx^K
zY8Nv84E}ojEsOg0&CN5599j`fodCMU(Lm!M6tq7Hcyvc`Ky*jz@(-qopDHeVV8_$6
zH!-`^Fb0he)hL&Z>WKpHd@j0ufmx2h9U#p>0?Q9~7(|I&fNo529Z;l=|NKHI5XC_Y
zqSjvM($DUh(P==y5P5tc<GDO=Y6<k==UV^xXaCQ1<5_G)MBkNDR|=ifq@s=|M*NOk
z_OBX8g_vXCt(@_k9&qUqC)Z1>9v>g3jgy$$q!guugdpr$JV!N7!2e3{QK})AJ-meC
z@Q7A7Di8X>1gY(N7(^y4D#cyQbh=ThcKOWPpzACTg7p^*m8wh9=lxtmJt7UIb^BL5
z-ph7h;i!m?Vq-t?Vg6&pPanmeKl-%H;PVl`!|2|Z46Xyi58nHpPq&|5cq%#Uk_A|;
z8s?BVoz<mK%`Ts|&j45h#Yxws#*)Xj6DH-@VM#Sw)o`&CcgIYtmU9ufc&F}9r5~<@
z%IoP_Kdu#<t%?l-3u_os!X~jnBF)NsrUj%*ny)XJjm<`O+-dC$qbFwcmzjht`=l<5
zhK|;hqix+<&fcbTt{Eu5XLME9rFB`w(fTiYRzr}1y@xms+9CQQ4a@N(lo6jo{W%Vg
z1%4n^v2i7u5!kg~((0Z({0#~E;A4xpgAf43s=_M_66^4BjeS>zSH1T8I__~@>G66J
zzhrAjaRmzgk#6KAgt$qcCtHei(Zvep{8(0J;<@_N4r*+f!g%p;XJJUfH_TpQ2fay1
z>k@5js)(y~TrO@L%!O(cC!u+r_D&H>E$<BNxJ{gU?PzhU@hVZJc2q$mjM?QxbQcoy
z&IB-0+AlD?0I+4Z9MLBCz>1EZ_wzvi#1Jtgxsh^d0VFfZ9MN!Y`=(Q6Yk!aU5vZg?
z*EEV>Va*7JX9^T_BAo4n(A-EOB3Ham(y*1^itJ)nt{mZpRM;_pz1CH+S3DJ@U~530
zAl__F6Qs`6U9oW@wuVcFBg$F(%X>JTv1~k@K9$+igs{D{KS%8+k7bfx#QI)+^c|;!
zG#;q1(OeN}JOQi0M{Pk^Et-_k9vfB%?=kG&aaXGjQ7T^wo-}h)<;n3Xys;6H>c>c}
zq))kwn|kGGx6;SJ9fpkKW@~*rX50H~A!5e{R<`^Bl5r=2r*MIb{(0J3Nz6r?^Vjtq
z4qp+^V+wup=0)&bpUBA}U`G@53F7d#rKp5%)jykf5N0yq5$PZK%l4_9VH-A%+&95(
z&QyJrGe;nyr^!-trp}EPvvGIm62?<gx^XYD#5G!e*-D|!&bdc^H+#x8`9=x-a~V(d
zCPHLNn@m46$T+iv6CF@osuyczEKdDgU7FTv(USL;RWJ9t1YEvV;_gxc50iV46V0x>
z(h|U4K$D)%#3!AraG6kF*_rsV<t#k|jhZ-tlUM+AAni#j?k1GQ#hSJkN(g_*K)E{G
zFCot>^<HkTa`Y=k8BhCqOZ7hQjOQlZ!M^fv<wj~!hL-o!ec<<ck($nXbT49bWqg(a
z9OfxvZu9lar|d7pTsxk^K8G7|psM?CF@n|LbF^~CEG|H?V#cCH-_sa%AzQ8moMf_a
zTY3uhFRY%e*kF_h^5#^mfE7$QE&ikRuno-cqv$;!b<<v%<yo_+fOlzIUfnj2FBONN
z1}e>F1qYkP-E2|8aX*+wpqqoELuYG87$Wqc&3CJOX5C_1v@xo{O{!!}PA{2K>m(*f
z<$RMN^M0$Sy{3NO!|SHodwVvn#2x>#nf7{}(K8~+Lv&fpso`-bP1%cbFcgg<T?%V5
zt@-MQq%7pe!3sY5u03%P>b&&m)_U`l|KT%ld04bnC5Eg5HSqK}M!ZO*J}w-(YKZTo
z$EG`B4u?&}hHt70gHGsX-cDwXETSurWpuj7%|FCR7dWL67Z1ar?OKSzQVo~B2mI<a
z!gI1<r^pfY0M<5Pd{}u(rx0C*9LsI4@9NyR>#<}ZV>!rU+=cB4iPOJaCD}RoWY(pq
zt8WNcr@nX@)a3ltM&r03OZ*0W1Dd20pswwMYQRV1&QRMTi^`)R`UA30ipz5bIP*0`
z3!D<S-KAT#zV`UlI?9&vo87=KT1juP``QF83WHaS$c9j}G79FEtQ^u6{=w3Abl&wK
zJ!j<o;o`Seqe@M?K1|YwuL-3&ejYz*jS+Ah##eu2xNd!aMKpSw*$buMXEqk>mVr4S
z7lHyQfsX<dCy2=dF(O$#63PtQtHW~B4DJoRT)7x`mBjW4{8WKCUObL-^5h+1+$Rbs
zTo(Co@;$3fU7yP)Ti5MYk-W_U8c}t5+JpUDO=%u`R)Q*X?sd!C%Fim>N2u^kRKMA?
z9s3Rd-1i(k4UBxPHlOLE{}dPY7jx)e{8dFU16&75A5h=5CVnt!7xDT0KV}C0i{Jgv
z_UGTwE@n?ccDaAeU_4kt9q?6n{e|l*&xe<B3&n-{uJ&tB$}kyZ(IB_KZP|DmmsG6F
zd3L_68928gTDQQ%7h=Wox++G(`U3MzkFnt+7>czHlu4uo5=U;uE)zN5X+`HYDVE-T
zWARm-r?F%C){B={f}n1@e(UbTM!E(lE30*cALlZO&Ow1TLODJ4l|*@3sjk8k%Wj<?
zos+yL9I&GzU$6o^RY8|R4o0oaW^A>LBK=|;(Z?4jMo{oR`ov1njsE-iA5400eTPhO
zkT;SX>nr<;V5v#b)ZqBE$_Q_7{1$-X(G2k9+40io+uwGl@MjmlH=QyZ5HrcsEo4&~
zk|Q*@8r!E@%lwjht<*RF)Tb!^bK%{y-CKCz{}54*yOy>+pWrHDXMk9j-rHq7x_cvM
zrh07q9nK^{gvQJ)ir|y;l&C=gcFeh*GnHd&LkzB6eErT<8Iwt9&=%s#6R2TLx%rkz
zk*Rc%`aM9TE_8NQCyq2#9eOYMinzMV{ij){?3j+vBf*maY>f=Q*d|M_ckGtJw<sy9
z3Q}K+_2a+F>c6gbxYVuxs5wp_Nc0+(yDc1?@tj@sZH?d8@L|*G^DI&pJ-bNm2UK$P
zs;v%8*H`*l8aUu=UD8T*SHX*<3z<yMj|$GyA(XZ2)R35HSRU~!)oI5KYOYi<z1=s`
zC0r_f>5Uq8%-i@_ll0{W*U#46zF6;p8iMv1Sy)a0IvDj5F)9W$O<=niYT70cB~T9E
zJOHeqmQZ%wy@>9If{FZQk0jX7MDgs$b`Hh=g8XO3G=}`^KtXA!4V1sg4pvo0mjns%
zbX~z$ECjl46VhS%z)Z}#t+7-9mHZMPy6bh23`%&)*?qnJtMFVy&rzQUO;6YI9B%^V
z;Xv_R=&Fyy{=F-@d?qbYSA8Vzg)=!th#gzh*UaXt!xy8caWmh&{ZFaVBqyu2L>I@8
zptS|k0F~tHA4kS*3^j(L&#Sa6HB~3;bF!`pbw(Eie{N01b0DuM5#FfJZ!deiR|`1X
zAzs@a(f5$2lPl?cRN?7Y$j_6qN95i6dJQ`FhB56G^WPJbF^l^Lb6k_IK@1+6d%ph1
z`9zgSAHokNyBM3B_iAzbO1Q~!?W(GTX70ywt7IZxin<<UnL+}GxM<win@%=eF{-#n
zhyox3Zyl&I9Lb&7!12@Pl&4&gF0(a@7wf$l75F*KM>((UB_I~QD~3SrtI>`F3$|A=
zJ$wAS<vamAoQTL?5~x&&yaVo?PCgc=|7cP)PT``)l+y4#PM{yDX5{fjtb1Ht@5!UG
z7~jK%ms_fIN*ZcAqn<bO-#`5PuKw{VLdN@X9BOE>$=mw?WqMtk&-fn|v-oG9V@{%g
z>EV8?2f-^q#?rD^_XXgl8+-#&*{%TVAt8bZZSbqF4TP9K1)v#GHN~OTD{>57Lnz?s
zqxR$%YlMF%9`YYi(ELZ&<|-h&qKdKmtmCNpPRqr=(chT)Pwz1G|3<y-pZ(r{_W8g1
zb@{d|Ku3U@Cbl`T0ExMQncZK=bCX_xr*(nMW^fVEgT~*E%={yWr~kvxoWlQR95sLh
z>sedi;)8wE0e|^jmc746x+n#I1SrV|XDu+Bwd%tOUdLo*`gGXh+~U+8qq~79Z6g<{
zXO9Vnq8$bSV3lj4+9!dS$=_eprJrV;0{FL&B>f7F0W#r=hX#zS@xQx@`ujio<6Zwl
z`iplFy{-lWZFBvqJ(O6{=d|<Ms!gp^(G8oDs{TB1aY!?l=B?j%9#*JWtA6pFmypa#
zpRzQNV^Dj{)lu#ZvSgZI^Z}48^heYhP^TmeE-wPlErRxG;MPyxAQPRbMJHze{oDPF
z!1NzX_+UC9EBiN3H;e`tip1u9;8<IchV1TzF&JFb%)faW&g!<M=%SVLTm}E{?~VQU
zXA4E!270UZl9|T;?zMk3;*a6@V@^>2x;h@C7SH@(aziB;QzYrH7yy3flP8`LVTq6P
zBS`*ktN-Hw_~XF*!zlQRIrN9&^JnYvf107l`Wst&DCp?Q!x%w3^YrK&5^f*w-)ClG
NQe(RCf$PWc{{rTTu?heH

literal 0
HcmV?d00001

diff --git a/_images/bonsai-filecapture.jpg b/_images/bonsai-filecapture.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..7a809d67ac10f043046af7dc5960e00b5640a2d0
GIT binary patch
literal 68990
zcmeFZcU%))w?7(+qSCv7fYKEKr7A5dT|@+>mplq65FsMH2C34UfP#QhrABJ#kuFU-
z2nZpe_XGq2gyhb1&i$P8`kZp#-+Ruz_g74^|Cq_l-fOQ}v-T?AwaMSfb0DU>nmU>w
z3JMC47w`p<r$K5UDoV;ff53+t_?<d)>J&BgsdF?mr_a!xqoboeM@vi3z`{h&z|25P
z%f!LN%*w{j&Q8b3d7gvqJPR8;+n=3KPyzQ)pE`T$)LAxqT6(ts;Sc!>i1`d<Gxa<b
zg#d_>nSzR$g4_<`2Z1O~0o49U@c;dxpaf_<O>^e#Ia=U?Y9<gR1r-$~HPxR~16TV4
z|AVNRPqAFOe)}}5!DAW$7dGiv2_Mf0-YIWkHypwV$vk-(boLwv=Xow}VG&WW%U5LO
z<P~n*R8&*f(A3h_x%=>uk+F%XnYoRv-P33G4vwyF?jD|A-af%0p<&^#BO()%-Xy1_
zzI~ULos*lFU+}50sG_o}x&~QWSKr#!-qG3hwYz6{WOQtN;``(j20Op7xU{^ox`yA`
z{kgY)KsY@5b6yl6s(+r=zs&3}^I`_(MM+IfMNRYPyeKF=fkwqled^Nn(=4|QXdb(;
z3P`^?!*(a(V|mM2K^a3F`;(VL=QxC9F~ayir}j@X`}-yq^bgJKUncgi^O^$DQBeR7
zkBS)t1|1)FWU*8YoL)1Sv~KadTA?mN?;L63a8>E^@}gW68AOY|MFwT5=#oyf$e@l1
z*5f<YB3W3mtvoaH8*W2_SDBS7nE3A8$)@b{XTL4_;dG-^8yUpur-Y4h_*tqVSJM=$
zf5oU4xBt;OzU~&r*88GSbn?ZRcH3)R2AW%66)lg7Nu15E$RJgp)RUt)GAJtNFUN>8
zAKuetzHut1njwQQzAOB@%Y|_CCo+go5B~M=u(+*Ragb;xbU6>=UzlGWZg_i9`yDbl
z!v_9N>md(*)Fxh9+$SULhjN@J85Fwx0X~qKM+ODtkU`j$zn)lc70ab*AcI;D2Eoe%
zhGY=RfDFRB!GAq=$FM#4ebt|w`TvO%!>X)=<b1DQ!^EB*+S<HW!RyeF2MxCkCgZ@@
z%BqcZ6uezzLT?B$-CX}M5#o1-3`(fqB{f3GAWbdEZ3p;E_|YmER8a~t-jM_^0|K@U
z2n4rsRNUbK{CE}KnU2-xbl_u(yPss*$YHR?Ik;N$?Q0(&_F;{z@a^eZ+}SVPRLYS(
z)r3-v?f{2ZR@;Mh=4K=&-QIZ*3X$lVe3DF?!y#VXY>>#7_@V6959zWJ^j>zuaY}P=
zWfQ8@Ds}ef-r&T)KXwKN&gw8LiyKcytn5~gCl#4@_&O-5<GC&+sO3C=!Syot!Zm77
zwq_B$NXFbY{aXD#!ew7iRJS^1e{0?boJ?nLHSHMGbuM>~^8O8W!}}724v)3aNJ%8u
zf+c(a5N)(BN!FW_e|L)vIv6H{VvO-7ak5gUyzatSp$cf*p(oluzJ2Z%RCwEyP}CV|
zHS=!Fh+UWE3rHsy2ac14yq;nnm*t7rx~kDW9=D%<vs5_N-qwCG2K$bAEP3orn8DZV
z0i*AVm)-0T0`nw)H0cO}&-q#lXcOS=YfdFX%_DEwG`{R?$mG!Y$`kQ$biO<r;RwMm
zp|?dzAy;BRYz7+taC=IKvPR;*#kmw(2_@b>J=1&URP55Z>xndmN)zmV)oSCY4%DG&
zJ>sN-3>qqg@I(1I8|6N(O~q!sJSstQyBIb4sxTa_@w^-=3HBShGZdGQ9BUEi-EnEM
zDXs|hs}lN)PQvl<tos{w&hC8k+H#T<G<%#TLYF8=S4^oGbVg3acuLxHjBop|^bGy=
z#6*aJ_Ot(IWwhJB)MR6$oU#pm=g!@}4Rwx@?Mb|;?#fNsLVNnK{8e?t#mfRGAYT+w
z%Abnlw<mSH#}6oLtzQqVnKgc2@|!RJEE~Ut;rC1OFN)&-SYhyo=A*lc9BPo|r%lVA
z=p<NzT1+WrVu?eKlzGV@f!J^|$od+5r8+xMId~pLj2VaTVY7~;?;b9HQ3{W37Igw4
z_p56tZg7zXsu3TXmB^s4GtYo~E|WnyLWmnVs5B%N_0>}qPVmL*l<gtT@aQu^)&j0K
zBtFJAabrn3hF{MJKSe(`E&XKHt{rNm5KJ`d=V$cOkyxt6Fjjg6KNxMa@)T<8NK@(f
znV6UOTKMVA^FHHIp;n<gMd>Z1t4qy++4;Kf<j_YWUm+7je#lx^xAnOOTu0+951$C$
zY9j0jjLPBZq?5aY?0#*$u6wlX^HM=ZOq`Ise$V&w8+Lf1-P<1+Bls?jHj2eyilI?Y
zBaigt%H#wf1N4g(-ie=>5-7eZO)sd>48AeCzk|vkgZ8&~iJZwr`f2yAHVGEH$oT_e
zZ?RW#yI!H!Q#~!qv8rj25w~vT7oTt5F3@g2e>eWibX95c+lyG`>(iqrauW#}?0V<|
zW73X9DKT{pFb1?+BgH;Y{vC}`i@;0<3A!(zc%ewqvCTolqy;i)Z-fl8Ge10-(;~)9
zWbI+epd)F;BU;_#Br<41|0LO-D)=NvS-~9}7eo>X*S~XA9eufYbuI6ef^nko=tcHr
zVT?f+A|QDO8Ie1RP~}|#Gs%A6PT8J6)miYNDuD(~r&yIxmM-J+hS_3Tvu?V@b>t?Z
z@lzgK5^C=mUh3G}43bwLgS1W63xL}ZGte2tB-Lkxjv{+i9zqmf_&oUhnl%V#lp6LC
zMqiU&(x00*`$0Ha<Q87hKihO(*D9sY%2d0n*Ll`^UKhbeXlMzLufu14EU|Fh2cMGI
z6VKq4dRkr09fDiaQBJZDydo3-f!+O*_60j8#I&11@%?QRcvm_Zlt>13o7h1MgzxL*
z<!1$lFZYl^IMHQzh@TZh4i+3<PX>iI3qiTFz1AdPZ3>^iuPS!TWTu}ai{v@6G1`_Z
z`v$!e*5Q2>f3IjEP-1`6-X}i`bF3nxi-$zYG13rak~#vUxWmf%>Rj@j8{-~6enxi`
zq@Ac)lD}B@#ygHXNMCI~k3A=OJO6-~-FzYiU-I0MzB@Xm+7E6i_JQu0duOpYW{ML+
zW;t+888vlS7S)TwWKf7~hEYMX=qpY`8@o^QRJL7u3etx^IClhu5;c7JHxSruCgR<4
zJY!__QZv#HYDM98xCj?fiV~}_{+fUURsTpm14-<qbPW`r6)TKd)Kh_DwvT$p*AE{o
zZUgUQ8{AxTrsSx`XcRTykDAp#Ib+1#R0uE(-k5d7R3sZHmK_QDN83C}0${qz7<@C8
z6myrLq%-XIu*mNY;T>N@*{Y9T)SB^7AJXSu>ccHX{~EIkmYT&a=jmd;c7Me-;|WMI
zNa(<b461<~$%!v>8{a=zD?NlK=uF$RNmlWywQ?^Za=je4AJ)R@2qNECrr=XQG=#q@
zBoueE#3$CBX_SuFrr|oL$-W$cK@KY4CDGxZ7M7WbhRD%2-Cwi14=t!bJ<`7z{?rC`
zy>BwqgW@x94rOzE37XZOgCvfdEzY)TET394u9)n>#q%1fi*@KwQw6!i2fVqv)3A=1
z?<c{7b}FOpy1?vUZK=epSyBL3E+NR*)gyqH5cdEd^A$I?ae#IhwH-B-&`^K+op&*X
z%Pjg!Yr}2Y9y%)1dRWX}`T!lF-b%^3+}^2C*jJ@zbym1>%=wSTN2f1y<szfawZYzR
z+o&?{5a03ZO-+}URz0{REO;vrX-N^8SY(deJ%JD4tc|~=yBIXDp;b`5_&rpgwOCbf
zOXkfq3oDpWlu^TQ6&0%$)I?hM1;4St{J9<8-s(5(RE`m6DAeMeyXN&-b?__}TzkCI
zZB~4O*ShRnu(YTuoj*idLq}L9b8PKz?jrwSZOV9Y>*!9Tqp`yw@}P<g((hW}x9cj)
z3UMNXsBuz^UJ;wkH?^J|5$iWQ#!lAcno{a+q>NuI>q#}YjI@aqnkuU7d{p<{o8`js
zCYxb%ga^;K-oZJs?D$h`2DkoaYWxqqlt>!9jHJdt0FMZ&!N&kN&RS|@5aK-er}@w=
zPC0t1HW<{u*ee#TXQ4#vdUayzwMOijM=~<XF|FBwIzn?Vp)rF2m;6BZdlN49&5{J)
z;$8<Q;mvjr?xr`J^o=&s_b`^58W&QODMm-Lsu+psaUFf|!NNlLs?b;*!jYd%m7hq7
zN4&1q=k#(WIVrT7nADpMHOk4oshMD&mgy3#nm}H@$P=)gLZRX{SG+2+55d8ENoR|~
zWeV6E(;=QxWDuY>C^8cYyiC5NwI9Trn%bvUn$>r?^450<-s+1ID1LRIqooPjGk!Q2
znPiOf43g!QUw$z9v!DO|=NdJk0~3b}2I@J*GBuQOFQH2+7TECDL{=>9f_#L-Cu@jT
zT$j0gs*6i*)Fn@rZp^WnPWWbQ3RcT2<@!<vs&S5EioSQ7%PYJi(Mz83wfvLm)+zmV
zyYgljla)i_H^9Yv)|hhV@cGlk*_*~k+D0ae-5i@$9Lx8OXHrw=dy()6GY7aJPnnVE
zxH38<s}}<$4g}!*Q`@9o`)KVWLiNM^^zyc^`!{H%(AFwM+j(~0U9fe@+^Zvl&ds-?
ze@X#y4FUrSW9@{O>fmF+1$>;})Kwc78f^63cM_N|YN{MrY_?;yl$F=M^wh%j#gYy)
zSFDMf(?$DHD(qBN=XOLU|1e2-MFu*x6#8jp8xwK0TA$n8zS5Bcur|`BTm@D)oYJHu
zq(3GL@nnf^U3=Z^X}3sdfbBMjC+B&hiKxD#UJ`p(C~5@)EW9{gcxxQNgpY#^`Uux2
zgTA-}bRO!@hjvApcXhxS5vl@?Q$$6pgAh6O3Hto^sWm%4)+t45;<cV%YCSZ#|1KJQ
z@w%H3-@H0u2z%3_#W;x&b`v4T7?HLdUZ55ELq{d_(`aMue#rUh1&hTur{`>mT_W$x
ztl_Cr>r!8wmCIlHgAX~SI;1*8$7cc7{K6wGqIeuFd=OPY2F)n~fj~bb1{DoqL4P5G
zPGQpx@Zcyp2(Lq`V*lb&Fzn;>Oe<c^*>HTunV9|}yf0C6X6den{13@H{<Jv8j;ZZf
z)$^-I1rZlx=`A)njfcHG<tyK@h(`#-%#_-tFke=GD3BA6Xwtm?%2`rDzksN)TtKAP
zNv+?20l~&v9P)(r2u`9x5P+cCFGN}ivx8jWxH4s6ExAL=B7*|gNo?@Hy~9LD<M$PR
z!@)oK@lx6gqCMzNopgW1>3OF1JF^d=j}j6(Z>^Hp5!tO3YBz<*pwce>tx_3FGAQmV
zcs(6{{ox9DO%lA0AoWyHcy1J~jru36nCE8)UOK+wQ_XLuZ(R84!<m4d%u=n%hhs&K
zVwC(91;qyYkM<nXB0<G21zj%8G4DoM7#>CATvv@ypP!IHnY(bnFVJ%sy2X|C{W^TW
z8-+)cMB+Biqm~=d1Rwy!YVP*n^s{|B$nGqR2*k%Ca*j_+9d0X-0NbQ}JcEtm-AOPf
z@c^*j0EYu>Pil!G`Sz1Ru|VWiu0R9y6#hpC8KkmA26e#+Y2e>O{&z9fdm^ri0FyN?
ze=l3J3o*Yfuyw_Ze)O)U=Jj~}yErS(t^j&GJW}~_wwKA=c1+^~68jiLYt(rIv2iiP
z@*;J}v1T4?#awVk)A|aS2i{c=|8Cez2I&eQQsK`~lPA@P2o&EQnB9ypa$6V8{X@S%
z#}45*C8UtDB-j3_X16`*@=d;{Joz#mcN+?w-<J}%*^kN_#2}@^L_l!WJ)4zBmAP<T
zUji=T!y`ysgsi(mZhF!xnwb8*s&9+0>7Ao^cb`6z*kyf92Gws`2V=;fFzYjTk>LEb
z@hRnQC%6;usIJ71me%^y6oK*4bJ25Y1i*7kAy{BHEwD73*F};rR=y9b#+0e@zON=t
ze;;@um9JysW}rH<^cB<%O4)*D5mIr8PFLdHQD?trn1&`;YpONP$cCS?2#$9<5?;Ar
zVQcZ}aqJY9<8y;_P^-y$*S7|jyDUo$$b#-s<HMcK*(U4l)JKRD^Q^_b?Unu2J^l^}
z3rs;9Bb%3JYg?=NawHLgIE;Jz!OKT0W%1)Z@1zD@Wfu*WM2ctb3;6dJObK9bT-&^|
z4^a^oUhTwVjyqW$*^6N`_x{*ep3e0EW9Rlq@2<FDMZ(}j6an1ACqod$yJAG!;VcfM
ztN5A((N~Nidr(ZS%6MshpXjl=^bd23ZkJ0xYGb0EB+s*SrLB?_ypB5{T&iaYiDo3~
zb(5nDiYKKtSp{pmvu!T^_B_j|+7YKQNp801@a)9?iwmEoBZrNH{N16|O(dg(Qul4&
z+;``8p!pF8s0Pmi71>Rw{LoJ+;^izFGH6g3g_DMpB#WT@w2fy!tSZ!J?7ts0oxF7*
z)4N|BHpzYY>2n&PIjS#Qh<^SEKF*0oB@BE@6AGdGLGb;8Y-av!3uUB+5KOdcn{~fa
z*q2IX47_E2%h8j<rKI(6-I|L4wgZe72SLoUHVKP1YdTmsXA0kCy@$85?p2OU<_4Z5
z@=GqVZpxIPuD_)@FTR`BV;JzwFm_depbvj9vD_CU=xm!gfg~jWJ`5SOP;yN3C4iM+
zhgWJjGKut)bDl~GF1jOmb&I%O+xWd(P)W${@mFOgdKkR*9S(j%X+Z{2i2^!neYX<B
z7~+l&wZYV&MB5m-)Eu7S=8bTqJXk*AF75{z1nm~rX2&pv3Ac%e3ijG`l0nJksAE~u
zxBA#}Jk@@HJ|FQ$bM3I3J_s$*>J_Y+v@!09o??~Nk?u)1wu-AW>hz|g0gVqMWpts{
z2d*baw=JN%QkfTq4H{<A#PFGtnd6tJ)Y&Xtz8?_Yy_Eu3A;5(GOe5YSfJv`_b@dex
z3c$M`uFk=O;mkRs&#w~o9+x?gLF`z3!E`>}_#KY4wnH-7)!AjJhkdMuazvdQjK8L8
z!1V`xz=t82i!P8xIGSE?5bw+hcSgyhbdO{r0DD4a4u~~SBT1(L!>0qV{LVdVA%lLN
zg%Rn1Wjv~EDl+~?7U7{PBN_BT7d~f81`)0J|4QWMrw<IGB({NIGH8VbP2d9mQ{es=
z32JZ@qx!ps17YF6P8mpR{`#zHvIBpU^qWh6hv@#6rN8F||IMYpi%W|{79E^*puXs_
zDw+(^C}s2FC#g44ky`cv1u9|xfwsQ6FlXw!mBZ0ihD+gU{TnHmT%(UA64%ScV|~+5
z)}=alp6;5ahBROY^RbV~rpH`4&>54GZFtQ+1r?8@Ln}+S_g3fFUURzgDE?72lZ&|m
z=g?{m%a0zegfm}nCv{UDjhhff<@t9z=ShbU(v?>;Z>0A4!|Th^i&rO_S<`m)$K}Z&
z2XjlWJDb?1sVMHluTJG6%*>i|FPOD%U6!5R4)>7?83+Y}LuYkuFo`i-Lr#tS?z}0T
zH|=(ezl)p|N{Dqe=f}PO4j<r&J6?nniolqgp8*F!f&Tyq#Nr>peD^rUk!n2J;TU*0
zQP>k>CG}-JpYUYkM(xmZBBv|+o9Rw|Kk?_xNQSx#2W3hfaC#EQS^!l&4i!ce%cQSt
zXdEe%gjW#Hx?XVYg8HQ1;f%R>tEa=(+6y8V%b=~A43O3zPQ(L=_+}F_C^Cn{n$CDQ
z??MLoA%K<c{t<cv_uXOIhDN`f@GzI4xZP@OO&K=Onrd6!848<A4=I0}cvcEj^L_H7
zJ8!7!FdE|&{+h?S8h6D48jb%t^<MQ{oke0dk9xEn$8<$`!^N8z(^oA5H2`?$mx;n#
zT_2M{%Ff_zfFf+6du_ex*Ksh$%qyzZgTyiB$9__PiTWdmo@2KqRmM!6b?a1UX{6-r
z_X{4S{i0pgoP%!DR!aAV_pFrkt79G`tEzDQKMn3gUMqV28B|1b>+}^V!u{vq<(RIM
z9Td^%7nb`I+g#O19AUs>sd}_cVx0;ngU;vt^+X9^8~jH-+_Hza4i15r(N=)d?)!iY
zng?9*Ur(?DYx{r3=v&<UxGGuw!-EMpE^`!2==jUwGoa_c<xr6>jLXc*QsbP)3vR75
zv{&YZ=Enmi|2pcwB_j{;4uP(ICGLOM%j~+a7Wyofsg+?Q*;m*4eIKio>GDxV80h-X
zy1cjb8D9z}WzUm<6vsyKqcZSVT&dL{eAE&?0r@lib3Z2&t7K|zW+~(-Z939@?yT}T
zY@<<W*C)#;3x{FT$4mMWuTBkI<p0pMym>1^PU`%RdwnU67vnx81WNk227lR}0Ea%!
z@O<Fl*XAj?v0O&8I9wY8#mXmTLv$Q2)@GU#;Fl=Q`Acgr6<H8XhMvAN$Y}_<-gzi!
zm9riu5?H<_&v&f>`G;g+LtGt5JTCtR_#R`a`##bDVv&oxktoVV2Ocorq1SCU4@eba
z)_>9Lwo!aRXII8{?t+i+gc$~~v@6#5pSqIbTz4vx0(ma*cNH{i<x!OX5I?<eA@;rD
zL>f`b!A}7TWAV~7<JNW(;`FflrvHYRkmb}%NU~18_4g*&{-Y!u17jx|zm>Jm1^{V>
zzZNHq2PkPclsS%biw60b%+oq?!z$N&GQ(ldoi^=2GM&n+*hEjGixtzkx%{cye$^9R
z(<GG<Qfmt2$c){~sJXC^C_PAgI2ZLv**DiQ33_l;mRv>vx8*z}4Tf!P=_c#n+}x7D
zys*0to>KRktiATI{$mZo867i0<mEu*cQ`-~5%mL*CSlL~g27B4JYV`eBjalQQjACH
zk#P=;cdWGUk5@wXKPG>^C3rFR47F$J6@qUEH>^F2$($f=fuD(+o^Qa}Rk5@c`KdPG
zFVq}0A9AmsaQAoxbjcK}MMQ=iUgNsHYfqS;i!H?J33iz92@bl0IlSU@33>%>d26Xt
zzC0=80ty>^a$3n6Yvr$(Zt{+H@02s{J|EU=jzFo@HGz{oW*Dw7rAizJNQ&c@fY1-E
zb(XyWgoS8<3y+i2=kqE=7#nyNwfvzhQ)odrw`+l_?5MhnojOwb`gmkP_tn+#f$zDW
zpY*93Z7S>G_I0h)=1~!nL|UoWd(eK}do@8ePA?LzOi$wObBaib9#q~b<*ABS*Uwlu
zAG?*zsE&q`9L#~xkuNi&f8H!QZm+;~T2+39bPmtoW~Hx52xpv97B13tdhE%tI??Ct
zEEvl+bC!j!XDB-8dZsp4*|*M4Ug^%+Z<ERmP_J>ZB%w_s5Jgh=oF^g>jxp{n@`7VE
zD!o{gw$EIh<rsC6?H9%L?}yt7b`<LFF9++^I@o%L*0x7Sa;smFXS(g&pofFS@L<a=
za(yKnh=0h@>fIl~Y;GXolVE#Bwr8f$ubbJ(V=<<t1~GT59s!RQXtarf<OG};11w&z
zbQiMT=5KNw2NaP(+4|pl%lDhucl3ja=Z`oybr|vrysWn$)-{Z|U#{l0FveBSgzjh1
zMO>m@2XKK;Q8f~axyJ2E0<wOEuQx|d3YC>J1@WUC6Pf2K$1sX_qW-+aBM1AHZb?V&
zHhoi71aEYaa*cvykELQXR#G~BF~l}`n+&3HCWFd)w$-d5J_gO`Wmqa)&nl&0)Xii7
zf{R%TxOft!ER{pZ1~ZG4ioD7#`&7QV2@k0F@G6g!1Cd&LuaH%(_iN^pK+}mg<^-*d
zoVhgfl>oN#A?2IT<_+9^p)|qn=QrI|6gJvr+uDZjSLQyM>(py}uJATGGF-{RK||9u
zhQtZ$5I=;$Q<I=KW?J-%n=>Q#t<bvgFRDMXn-wZ?DYaSjSF=op-NDR6)uhhsT_fCP
zYKWv;E~)qUQk*TfZof*~)Cr>Z-A6k*bfbX4+iz0A8*WXE9ckaHG5V*WiY6g${1F~L
zAyay>p)VwzfT%(%|8?1)YSJGQy#`Xs!;JjiJYW#GvV$z>Ap#N0Imn^-?&l;=<kJ(U
z{(|B~2!Uk>_2ZFAJ9k!?pPg#Vhc!<{w((4MRWZEY?8m%y5?urS^h7P{S$U?O{H*<A
z{CVO`g(NQX5243MDLo7%sDjXWjFHTm&p5Z<Va{F?ht*6`fvWrJdU043YgGkVJWX>s
zZ`#7Fbv1PCiQ%0_+p8y;uyqV<dqjVgy#Ornvw_HCLTbpe9D=Z{WHO_3z0Sd-z>JMX
z^GbF&jx3wpX^T}81y99*qBzm1+tXnc_3rZN&bjs;hgt{4RcYmTL`$YzGa+?eF7alL
zzuwi<*7^{>Q;idQv}_h$qft(Gl)iuboFGXZdvDPwibiq{<K|-d1GzTwWAk7S>D*v#
z9aAujPK0~Zpxg$kiW3Mic!3E_e_6EM@SX%9Jf@=$bw7_(RY!fx^oZ5D{w_CGj`BwC
z6SFU^bGGq}JjWXLo{C*zepir3Rud7*63~)1g@~W&JPLKb)0e1wgP*@j^BVUE(hN1~
zy}}&U$rv+9yikdkES+n}BA9IWvBL1oWj}eP9u4-?B<E!a+!plDd_?a+HQUK+<9<&e
z#Utp^+3+SM6$HT=oH5hH|I8)_J_B2Ja5%_=8_hu0UvKC88(XLq{OEFy7b^Da=S<w_
z`_c~bKYk-7^oBOv^14!B9pox$NR@J2*=&A*^C~xA!6%1_DM*MDaaz~=ZucT|r~5_G
zieeWljZ?vu-0NSKK35|GYpgGfD*O6l1}IK~mUZx`@MT<^!_i}-iW-~z^oepKOD!H&
z*X8qp^brcn4lR1FQ%6?QWROTC88nU9ku<1uP(?>SB+j6U`5zBK7`%4d6iW{}-pcWc
ze*t<EM#XH*T6|aLDYAA<{2p`ny6_h~#hY6-+s?W?b|$9tPvbzBulPL#;G&ODvS-C`
z_w&5*ijdbFtW$)-Du+y&g63*xCu0>gSW3A>U942)Sp^#ELwJXV(3+S^T?XoBMgtKN
zyj#_Oqb3C~b2otmv~klIX+FDIG8{-j1D;QZ3tvwCbi2g7%qNSreQSz_x_T$6O{ICY
z(vxxtL-a;lx0jh`(Ut*g`KcNr|6JD~%2C^0LaFt@C!>G1xKQuY+7vRhY(uJQ#a8Ex
zsHxVm+(|i1LZIf-By-N<JuPT{-KfRz`cM76O8=`{EC-f}t0WjP$!e>_pBda9Wkkqg
zd{z3h^zQKk(twrW*8Zv2=HI_;3#i*YOR$P+lon&oNkB?exmvmVg&wQ<F(dK%9SDx~
zYhLW`m$%)v&#AlXI*I#LRIEGge-JVHVdj)5B|{@1tM~oE6>ubqzO-a7j(@dqPUX%)
zhvBq^1yOcP)}LFeX2?cYdp`1pq>UO!#-e3Oh3e94Y3cbFhw5FXA_omb{zGs^`W1t!
zG<`kr`GP~_ihr<nAG?kUjoc1RJ1hu>m2kzzqTDgMrE2Zr9i$7SzOj#0Z!9G~#p%ZG
z=xRaUF@2~t<CB*D)bfor&YYlx+Zt%CM|KR*wP4=5Wt|!*QFq(C{G3|RFTOY5Jfi*j
z<{29YznK=z;p5$Jj#ek{bCGueREZU82i6iHKAGVe6mXs3ish?K3kVkOSa@ll;zLyE
zKO4(c>TLT?V2X!<=i1Fn@qe&y$16n=_i?Z<H{2ORj&^i$*dw&jPymO8g~R88kNVGX
z>*czi8k#C(q4f*e1}Ua3HcONx>hGLUSagW8cahg_hn`PGRFg3+1t#Mz`in{0%t@w1
zwh70?(5xCmmVo71Th)LL;b4x(J7wRlFI7R{^;_fH*~itq+Ml{7SIHns3G}j5OU`DI
zIFRsh*rrv(#)a2w)a3DQ{*a$~V<KI`#|>vP``XJENZT(jZCe|da$LSplKF6f`l>42
zin9%6PvCln=fpI%t$5T!B9u`KX7&n4A&*bZybUrQ(N6Q4x4YEQ5YDuH>&^zxW1*wH
z6P^<Zr=!RzjROEU##y1dlHGh|ypeI<NdX9AGEOb)>?kukz{_q9?jg=ud1Y<L?^<pR
zr#LB2xxYE|I399Ab<a|Xt)<b=MU)c7ai57gxTU)J0ZQ}y3TL-`T-7d*!{R#&ed$#P
z;nD0Hm|bj~$I@$MUt$f$M3)SLsB9cd*0YXDK?v?M-beg*{nQDmFx-UZsJoS})|h08
zgz?yu#IJXjwy;<iYL6lDyCp)iLkqRqUlQ(<c$~Mxs-v1bbs~2T)KIw1wo0mkHGSvZ
zi`0QSL1s2B7ou<8P@cm>1pLeh>dH>;SG6!(h2Gg^;d9?Nex7)AS96sTn;hY)q8<*x
zI=odS!iV2q&mehAT=mUyjFCUBWYnqDBd1$D0>tpTR;y{@!IH$=&iNMm(6eMvrBhy{
zj^GY)Fj36?_Qe}*r_EQkXApIWZzTW5eqnaoiPU*MSljM#T+Mn04jlVj**_mWyNdgv
z#-38(o2QMuQq*Yy4G+@fF(u~Z2<sc3`AQRcm^t0zY23_6umOxjo;%~pE^|C~V=-HC
zQlSNJ+hi+qRy&<83)R1AW4mKNZ#+A6`gN@)jWaS-ng5g8F}I)b&~rbDqE%Kjy!DaW
z%sFiT@OE#SZ$45zD$!Fsf3WVGrp#UIwWN)#CFBONdhpTw<^}dmZ(+Dd9wI-=im^}(
z5$)%Fd}EaZPt=XmA;9ZcF*8kDZ$3M0VfHZl8$YyK&;i$B$3ySxAn8j>%N<fQvySzM
zIDRu+Ky?{WeY8{pj-i;?bSo^dk|1@l1g1uY%I@oy%O-O1sjRhaGro~8**tL`&-8(e
z6s9$0ACtH>)&ZN&e>Ll|k)He!bj8m5=W)!<L<a%WEpzE_MT^F_7aQoyKz7t&Ej-Eh
zZymOEf*Jg@T<|qg=i0LM+~@Q;A*O^Zi_wp99PvL)%S1!FY+m42zeBK;bD3eaU>1T*
zr+m$nQGe$fForF>BS1XsLB&`juQ&EabkZ8#xQRr8Mel`*kNOI>MZ6|`^wHPslF<{m
z3gT9(uT!Cw>UM_?De?r-9DPjHF)Q{Nrxhxf#q3yyUrp0{GjX8L&H0#3+v=-9x6&iL
zWmp#eoltP=b}>P>osTD}JuPk3gK15(f974OT^$QD!By<hG%Z5ZW5cg@sj+n(Ln<r@
zXls_h%~n44#9G~1Rv)wOy5BrR-$;CyTN$Y&pN1LiI{wHa^NvmEk5HGYm$N(sNTJs=
zEJw!aIE-!#zh*VyhPhLkug0*K$gg8OLh~pdiACS@t7xP0o_>#ilw2TmWU*U$2w+fS
z%tn@WvgVST_eswts!Tf2AV2<jC#kMkdhV6P-gY008B~Px^F}If`-8(UjmPdKI$nP?
zX(i+&D8HRSUhzu^oDS$3-V>{iksl{f*T&MmO|i8o_Lf*Y*LG3a@JhXca<SMSUKhC~
zvOPsN<xb!b9PJ|j%K6`ccDjIi1>~c22f-WsFu>j)k5py^z_`}H$v=+aCmdA|P7N>v
zNZSU_J;=#fC^jz4L%~`MuUY9@XwT2QcPCSPbk6GKM5pxayp9Vd!zIvtgF-=X`y#dv
z$cv#XZHQ^}pVF50oCfVNf{C?{sFQZ39Crq2hzc<3)q|Q8`KgfdnKe$Ir1X$(7R&p)
zH_jXEguMi<N84srY3aCn2Yh-texVt<RY?ZPddc`%hwI>zGb>$0atgeBb!+f`^`SYA
zEHh{OlNaj`Yh7c5?j-z-R_*xmX&VLIlwWS0y_|F}tF1n^!3I;@6!49g$~w+iW9LpV
zZ)dw2Pq5H4Sl@Y?HowNB_uxwBc%;5WM`p1<T~vwCI1+R|Pw&26$|8>=Rp5gwb3qh0
ze|Q&X@(?b$XXpLk&aA}7v47XIoqFkTG_5&~w?Lkf*Psh{eDTz-@UBp&=9a}}zbjJ6
z{N@8ln(WIsMF)pI(t2?KC#rYMh5*4>ZHMjAvl><5W1D>4ir_VzrcG%(5_;y+s;1_*
z*^AH0*Ivtxrp(h;c2(rx6p~UEc2H#=-S7iLy^99q8s!x|BQK)Hu4dk_O+S$y6QpGl
zXQk`9`r*0(T5ihAbKg%<q&U~g+Ob&_&t4YMa4e`lPWQ5{pgg*($?CD0-D4i<ur!@B
zbT3}Inl<yDyftZ4Xy*2+MzGB~WOBV*8u|$hsj;xI8q_`A*Y!M6oATZlF5dKuyl7vc
z>9Nw14V09=hMNuhmMlq<1fTG~4{bs5`{`k=_*>0ua^`$hIQ8(3vgXQJ)A?Bas*_Be
zd5aK^IW3+N{qTpEE;!%1e;V8dp5nJo!pH*Xo~m@;O{fk|T~$2O*Y^v@0etF3$QdT}
zb_>cExm1@xOm!*>U$!TE&2R(3L^_gq9iDwZ=cAeC_z~~?t?@Lwp^(l^uCR;4ObCHs
zp~O=R9`ljD0bSz|TwF~QN5ecVl*{cY?xgk0_}T;qS9}Mp!eoUq;^uyZ(^DE2+2F`J
zhs(k{M7VKdN%Kjn@ZJt_@9-sYn_x1IvaiI9;qyNh+NY#g^WjD&r)oZ`7BLsp*F{cZ
z^};TechZ03(}cefQ?d|s-fo4(a)UecEG&ngBiwY?Pehcw7t`D7TOCti7p7aWcIuZ}
zBHAt)L>G3?SUmKj7}=`YW;4@wD%NFuEm~7w`#jl8B<TE;Fm1C(>~_b>_s<UW$&5sf
z<6v!8kk<Ttiy3Hp^X10+%k}eSJ0y(<cZxKq4b(I4Yv1bc#J;$8GULy=cYy90Uwe$W
zvORG`s@_pndw_q6+!u#MUM@o4&NU@e_$qBGY~GX5OsuF5Zcx8JXsE#{TpxJ;*#+8W
zA$;DtS6ip5P^R9U?KW<StprTEgzA7{%D2m9i;aOlek5`AvdOgYF;Z(>;o~hHQkKBQ
z?oYIluHyoVvwv86K_4_Wrc7o~U9C9?%}cxb^0to9o0r#vLij##LpHV_=lRK$<J28W
zbql7g)jp|gw9C!(w|hc8Y?JpNNA=V_7n_T}-YJk7E3r|!UYc~A2*0V|f6k{#2~{UQ
z*5r05eqkUiK)D~QA2~70Na`NmG@Qb&&CI7-J{rq0b&Z&9!Ky9jMCT^xh^T$MNa?Xe
z0MFkq#WK<poY5^L7Axb;Mipq8_4L{l@x%1(>+X@lwO)C?5#Ku`6+&p7%{nza!{@Fd
z`D90kw?8?~ii~BCNIq*Esfxm|PEGb-QIyGft-cwWc&liq;?r{wWtlYQ<ek9q&r*&7
z^|8i>m{qr#D=I9y{ow9h#Ph+#g#M*l>+xn^y~ZxUo?q)14bW3S&O5+%Yi3mKk-gp)
zJ?H>b`23{fOP}RUDYqI|$AAmCv~_b~L};m>mAkuVVU!Nb@I%#=@#}2q`9^{9c|l3F
zoR;)62W=?U0l-u#FvV>o>ABj-m1QAJyJcj&9b~%M5-5Gd?5RRGmlD2wp|3xA|MOkd
zjEz3BV(jWwIVpM#j2i5@s_E4nR_g0pcRXJy#q^oJXmtNp(1s4H&rHSZ8sX*kvs|3{
zczF5RwLc&yB_@A-?eIFakanN(R;1Ur)0x}y>Cgix>;#DyfSjv2^52gEa4kT_|KLb0
z`#&rqIP;g7uw!!9QTbkRJ>yYS+!0`m0EMEqUc0>sZeCK5-l63t=GpQah7A|8>o)Rl
z1(K+mTuI8qKQs7%)Rm%LQ&VQn)nh*KM~`+T_OCzr;ttwuQTe7ckz;Am;Yr)3tRpK;
zkGK%YW$MG^Gm!<gW*H&M`7b++Uz@u9^5)J1$@6f=`=@VGovcy&En^XfT-A`HBLJRo
zM=sgAx71H|#63n6+9r!rE^sZeb&~k8qco=7%)II#P&s3z{b8Z-ip=KltM>urKpn&)
z;&?ynNK)~DSAoW`VrFC$SQS@n*uObuGLE?G7W?K$b=SX2NHHadI-I8&eL+*|!N*g<
zX`a|GG6Zuvp5%%Jb<Lb;!skR@Nh4+m1V03J2<{vAXW{I)`OnB*aM-<BF+4n?(V{BK
z?1!0I`S}@c8&UjJ!Ft-yp40!8BF;bWeGvd8+~VBGpwD@!q!bM+de_;%ocMQZMA;K^
z2>N#i-JM(?ey-k;s%{8rK&AG1?B;3Pgj!I?*T#xXoC6)YjwZUV{99>gJZz>U@ac0h
zvdAaDI6ym*MmS0fXL!MzS-ZRYQVJWh9sw|gUUb9KH*du4>{#R)Pq9kSU+YpTq?AST
z9Qi~?l+{!a?@AnM4meU$Rc+!VFZXF0O)-qrws8uv$m65muMcTi1R7|tf&O1^6upFB
zvZ*-$ymd!-g@2zB`A;Bkkede?v^FrF_o<W&8p|rx{f8mx%$D$Ce)EP8WY8xd%UFgw
zzA`2rw`a3%;>N$n$-g)8XB`e`<P+>Fwi)Rqs)t7!oHzp5mC(^fac|qc!NP{%M6Lh=
z-};Xe#iMfrr2Z!Hh7FS&Ku;1`z3i)x9tZ}F5M})Rddy{@%nOZeG^E4ztTOS+k)H&&
z2BV-;t+BfNja_x36U3jzj~-AP?|+_b7YGsxx~6lMJCsXQXwD~5f8KCNg9?D&2mV*u
zjb9*rWnp#Oe+%&|3;Qp1{$C;Nf8^eqa4Nt^PW=G?{_p498Dh~VW=Esd?5Ly4y+V6X
zcBI%p(kTCp1mwApO8gzS{xu_ai`#uye^X)(O}HO$!%q@Gjbo^BsqHHDstpSJ`q9ff
z&qr7s+WP7Db+cya<`}QO=-*`sUv>Q2h$~BmWBp~?Wu^|(I@6t892{JiZtP&+w-1*M
zdSrB-sk@kfjJ92Xu)y+wm?6q=Ueb7Sm!`8<7X~DL>9!h$o#v!y2wIzPMrA8+VMO>J
zjJzQcyfq6oZ`EkJMbqo=wdcKKJ+6~?20LuCnG%wWDu-nOb4j$#)A-0@W#9VjxSUeV
zqeG^~M6E)LVCZwLj8_+4FCUma)xY-qh_w?!=O>4!)bxkz9sA0^^UX<G?&Uuc{vK`F
zwR7zlD@KAf$`f;Kms!nEvWT=;Mke`k-@F;woihPrP=;b}`WSr3sn;{+LfN#5MW5=@
z6qF1oFt<3locd3E`w=Zu13&M48ZFdho^kcthBZLZ#K|Ro#sC(A;Jj#9qds=~!fJ{S
zDl*-ATq5qZ6lK&2gObH6CP+yscz5#fGK=)v$j7}p2N`d%UOD@*_=!-zdlf4lH31R?
zr#wfW`w^aH9ncF)HZPsx`-L%$!7T7R!A)dEZCo<nlN9&x$ws4NWddtw+Sq<Ukwa?3
z9%bpL$w??@O^RZ|15SmM_`Y~6mdH=DFI5+eG<yP1!B_+O&<u@M2)*JpbGdnaxy71~
za-}tu{!VrhW2n@HZ~F=NJ}VbpdJVIccs<H;t3yoHglD6)Ku-d0C0Xz0)W8b%M$Q83
z!QzRT;DA$rG*S0$Z<(G#r~t07`FayBMBZRERo|CGN~Ef$!69In9e<^xnVEZ0;d6qQ
zfyurVAp?6HM@3{FBB~Wm6N2E(UKS6=ocC^Kr*GaH-?qO*w=J{fgp_KEo4urc_Nxk_
zl?<9hYT;6<KY0aPMCxf-CAFKERofRA&Fb3FOgTUE%CpdNp|m{n^t-L;0o`YY+uc9n
z_FZs5?aEM<w(N}N#QmZC6zG}F>S;pe41nu110_6T-1j(t`bQE6yJjCn^q#>klnio0
zF8x64Ogq`mZ;Hagw$Ogn$90C*CNt}pBmGUg`35u<A}a6mAa|P{jyEo<nqHMwJS=nH
zd!aFE19#)?p?Lk7SZw0w!?5wI=Z@fjR*#w$#ivF}xv2zv+aU>89J~2UdGVXz#;Tke
z1<flQfmFp|aJbgO=~JA}7Uj5!FDOcamX4FuJ=_NKDk57A(^Z+NBPrmoD0o@d!;C9f
zY<IC+Ii4UCcwHfo#NcItX<9f)lWWeA*!V1gnA(eZvim6W2F2H~%aVD?*>N{i&Djla
z3E7?-xet?oTlmjz%{0L>O!~Iz<%v1)rd|J&P3Ngdu|lSQaJi#sA^c62*iBB~u)Jn;
zp|++G8M!t2V=K&!AtpE>-b*<Uehx4`Ls;7&KcyTXgj1XV$o>}38)JUr#Y}FX3dI))
z|9%le0cd$Z%S@cmPMMG>Gyt&u;ORwnKrNgR4MVM152JABqyV5k;kL3IV0-NYP?bOf
zI9`^bP6Gh1w4TKvjyS6jM9Qc?5&I-2eR}v{QyzSd4MNldIUAV5aY0^CQs|~Q83fAy
z6Sx1Hpx+$&EknQGvfp~>ze6#e0{?2dvRCpp8kd(`>x^$Z|2?7eqtT-mAS%GU169LF
zB`1#Nk^f}LR{f)n76toXqXCA1dt<Cx?RlFFG5`RSR`7kJ5A=%vMX%zo#{11f@cfbf
z!SvBYu%9*|IoD5N(8=B3k+v{9b6LOgS^3b@v1f_U?r5<N0i|T4Fr9_+E3dgV)#2Vr
zBycVn<c})#(rTCD_EP(D#E@0!)$nMzN}PAx<H>>XDBm}Q6wEt;Xtsw9r)-~!QZOkc
zvMEUwh9t#YiIp7C3)pS%Gkmj}UsgS06|m{o9~H+mZMJ4C55~n+A6woVOfNRB(w=zU
z&-F5~uYyT_>k|7nhST#MZa!`)(mD?A`ArRsRqC1Ug@$J&KSU$SPoN?^W&vXbQM%rj
zs*w(tk{>5TKTf<;nzU555#J?z{{6z-d!tQCo^)Sv*?Kc$NOfC#pBZXg)b(oC5xWm@
z*kx>XIY~c1OM0L^F3K3f=-yA+GuJvadoO=4Q-dDnJbYDLgq9NSd)1%f(D?ohVLL+q
zd;_qXK>+jTG1>1<1+}#kKb<b8teVAZcotfi_k0njsuk*cv`i5w#2oLfAIoulN=Kqm
z{=ph@_S&(0WZLry1)EI^b1nZ45u-4%`Kt@rixqPxTi#OBCRoW=IL|qB$TyVAsJv*0
zD!+k8dG`yNNp~KLLaLWLERT^HuaqimF4inPKr0!#<((1J>~}uiL-oSc<4R)`2a}Su
za%+evgi=}fuzMdt=0OIDYeixA?gXc(7BXY;U6c$ezQiofP-h`+<?0swVn$O-n69g=
z?xirNhIt`MqngAzu9g<T*t2E2txm|f*~k|R%3z8XASA=rYuOoUrQ7MVSB2&A)sxo&
z36JXrmNYTAG<>k?Cr}I~R&V&ZU^$>?B};zdVRC&OyXP%cTem*ldVFa>Vnxa2UP82k
z3(r(zUGm%0W)ErlJ$`TriXOjTAe~=UJ%XWE&xdKx>aH7Hdzw7T?9S0A8+mpP31Y|(
zRC)5(2+02g$LC(cOJJpJI;8UEst43S{O|w6_{KlzDE~*!H5J91nvWN$M15a?x##4z
zLtNK~)K0a#=uu11mMtIm(*<c?DAxQU>r7R6Aq=!f#j6#apE!MIB0o$s(sa|ASLlwK
zIQ5e=f*(&eUxQ6Ks1CW)DGo$$tM<Fd<tBb#sju!+eQqfG@=EuGt1;C>Z<2-V^kDZZ
z^V6zFwp@+dTASXd>p3}b{X7|HG2Mp!?48aK)oZ99`K~N8qjh08oWYhs$cy$egD!pE
zi~E}5*&}}Z#jS%93&qk}S9*pkPrFuyQPFZUWH&#-p*s52Iu&Jo%961XPb_W<r7%59
zT$<HLyv%)>IVTBPES^(HNAb1fE!;lJH}3UA0+;f$Pg(40u>FFOz6sEw7tN)w-@X2L
z!K3<vd*x_hUq|rNs{$US2Ep|v7$zkS%gC%&Hfuf`W6on-5xF)_TAE1|aq9b^B@n2y
z_PzVd@+;##Dauu!&QQA?>(%Uay~>nv`-`cr)V|&<+B`xFZ@h+HDKFkeL}E@p$sQxv
zptBdM@v+tZWh*Y;6Z`T8vjalVWEy*Q)yKVvp7c1;Siqfz^R6E-X<dQVf7W5U#$wqQ
z6C~4<ct*9|)+95rdg9?=13|wmU_gqv3+wr;%#T@5m!h!uxu4_MEvsnDfxNbS<1tHd
z*`uQ%X?i~(=5ZGB8or`Z2*BoSv;M@brxC(Q!BbhJ(>g%$m=VR#C_mX@)5O6w>dd;)
zGN9W3>zjB#>7mlCOBv6%rnGI%9`-fWy3>_#-5*}Kz{Dt6?FUp+{HJ|BpczL!OC@<D
z7quoAfl36YLz8aEufrPrwF{6xbl)4koF4GVKNE7egiiJYtlHnB{Og1I&9PtP_P31v
zeqVlTnBUs+|IGD8@1nS1(6yKLxk)EB{`R^5i`$`L_%32prU0-Rlv36S$Jk+E&YI!s
zwKaJY?(KF)dVPk<;s*1TS=Pj@S&gKJ^=UWN5e46P++JSlY5rjoy$sF^05W+AuI-S$
zWmf>THPvkRfq{Rnrxysh-mBDux<A{`cxKuVE57a^k-w8ST4o!Qc9Y%8Q?MBH<-#>Y
zqN<{P2|-~$6K~QU!^ufMN2qFuYG6%1)@U3$G?d)nZDdm9Z_2{79E*3&0%f2f{90i^
zwHMrJSJYG)rO?}Hlzy{?w7qW0Nt9JMxdq<wB!j9a8iK#dTy14;$6TJh+yrA^9Ls6x
zd^YI}F*JQ{rK^*dm-`d#_HKtUeA7>D&UYX}@#j(KOnAgz+$kUid|P&ZM^<{e*VUT`
zW6IQXCjos^r0F3xH3m`Te+N39#XI}$$&l=1_Mp?=bSaNPq^5@Orq&xAnu|e~hCxd_
z!Fr4gqNgW=9xxJtO%pfW;B9K~Q6R1%3O}~;VGP;L$hk8<JsXzRdpwf|OZZY-Dbnx~
z;c;+ahiW)MZ;`<8G_T^Wv+j5sv+%B2DZE19PvOB5XFo7|#*Z4;ds}RMoi|>weyiXO
zla4_zY(orI^_LtxoJBI{WX1f}KN0I7i~KVAV^dv)JSH@PRC(UuzhTVKFQmCc*79AU
zzKp~)Ui3j@HyLEU14RQU{DsPm)SW}fk}o^^JPJ36-U9(UMW9I&T7T2iC>d1bJwPM?
z|9k`*JE_JNy-Nlz6b8y91^%-;he(CX;N5v3aR)@@h4|N>>XAWUAaiX7G|_ScYo%Xk
z{Dpn|YngzEse*Ux;XAi~kNCIG<2P6T*OeS=%(1Ei3fENS4;1e8k=l-_*J)Cbn>+_V
z(dp&5@miO63^u95`-C*h@~ML;V!G02?|rMcM*HU-8;57d6N$rOL)QI3;(MJ8dPqak
zf$#T#F&Xf+KY;gAiZp3Wv7ITS;j2KDL7d#<KNJQkY{Y<FgWSoW!eHP9L>=3Ky!zQU
z(711?y=(A=R5Iv!4vA}ml+Q>6wiexdLoqLFQVDNUB!fm<fZawRClzsoXON?};2m9H
z6UzrQ(f`@txA=Y!{l7EA!{+w`m@333tQ~A`XwR3QibI%PqpADA$e-N=Cvh$9lHks8
z9ETJMtc|Mt86|5lP^<@Cx~A@Zb+FK8(C+CCqCM4jAdhZ}HW&LRTk6Bv&40!!8CZ}(
z#bEQ4e;9Hj@NYx;MJD`fUE^hcvm0E22HOLQ|Hei`KOc9<sjLCO@IApt62*&Gj*wma
zJCa^eq{c1yPYaWtM-)pae(-{RC17s-1YlY(AI{@*ylgO>ZC_2gG9_{8I?sHt(`|#>
zkq`GL!<NC&aQf*jBo)GOUD*m(X-%lh@5&+g+z@ICGVs7+4Y&^-OTv@Tl8we3ACE5*
zt;b|%r1b0c<Sq216(Wbh)t$=K-rSm8v`HyXm6s8D0j2TtpP73d9<+}exTGiPSEsF|
ztpz;HoGahyE1GB<dTAJ>Q;Kjq&{IiO6KzX-UQWE(H2QQDfyf_I=7-C4dAlrW(bAjC
zshWb_uJfb|ExwEM`3NlKK|bJM0Mf^F`=l7kP?)2|WhRV271H$}>E+e(yfl7A76lpl
zj@szpAWbDx_*4Fvnj^(kgYNv&pzq!JuLFk8l0jODkCpO0p1lJ-Ul#mu16$^Zszi1<
zLDiG)Hl}JPPT9{#dK;A{E@1>aJJK*VEIq}cUpxdC`GT5OmcE<2*I^=qqLN@HKwd*p
zyx0lRhlf|-WpcV7o>dU4^03&+J@B_i&RC3b5DoHL@U;6u!4F3^=i}O~x$(TY<{6E*
zs63xY3Yu;OZVfGH0_NqR)5*Lge~2p6%B;fphbV)Z(YbWfF;S$Rrk-_Y>v>>njj#DU
z-j%{jjI-Y#EQ?VHik4KKqz5Rj{<*zhNYB(-D<jKczsvs0RyK7V_4Uihr)PwiqnUh+
z!@_Z*^UL8Hh2=?iSf=hBa3DFE%?9KuD^p=t%mkiAa|mr3o;JmImuWAZcVtUeJYVa;
zmQpdI<)1Q932`qeIT8+RAAet9aaP67KYOd*N&z~pwdy-l<BbecW%E)!>T2DedNYOI
zFGJcDTFLd743=;J;5^^xC-Tu%>Em=ZK6CFEmMQA>Tl#4_==0a=ZD!3@DpyPt8&ecx
z3!T`eG4=AZruA;ILdH*G>XKe1ba(ptLMBt^L*Lii0(EyetDh|`%(JTQ_|-1&DlabH
zoLTDb?&hiOJk6<llBRlN1vPnR*h*S<<YLPBlt?CLTzB&1Ld;NnNm1X%-P*?kAJ`|A
z(Tv`A&wrgkdu;9#8*WI6D|zq2ZO7~NPe}sL?j#nulT|v=+eQ*5fbZjyx89wk!(r~c
zn;$YQOx=~7Zf`6w^gQt{{nq)_=*6uIH>hp7=ZjY%tR%pCJx_e>P5-TsFG=wB8DyBa
z_eW&d`4%A|7tP4m(mJiLx<w01;Djgq5oFNrYCaH-eF+qD`g{+#*!xjOvrI(N8bL&<
z=<~3~Kb0w3)S?&WRg|3C*essVQ?ju6@(JM&teKYxfN`su1|!i({kfedof;lY@|Q&u
z86dc_xV@+=z<yt^-=o*IpulcqJId!G572)dj~~|#W+j7`M*ySu3*bDT1OH#_y=PdH
z-L@`_BBBDKRH;!wP^u!mMMau4>7A%_Fd#iZAOZ@~1q2iXL_|bNqy(fT(gmb92|e_l
zPy-3^JZtUqeb;{7@7a5ub*=CH&b5z!`0+5G%w&u)$DDJFd)#AKw2%_U2m1ZeZ$bPk
zEdo6Qcx5<{O~}Ue90CY30K%S5!2D{C!3-o(vLZpK!A7=OL8a<FYGVES=ZTL4i_eDa
zgqLL5edk`fpA0nnM$q$k((2Hi%KPbMNrlG9<4S{sh6UjgzU0_PyH0%-rm}mx`$u%?
z*+*m)K%Fy)@PHtCs|$dRM4b?dG9(UEmLc_@kGaY9d%)jh^;-`AW;6c^IY^rW=5GRC
z?M57(QRtyeq5+Vgw-M0zEfqr!EVOaOzKl&8TT+OXJs-yU23zyO^F`-kY&@-JFF@Q&
z7M)^1G~POjJjw%LhWhIXYmEHg(kNgIaBi!B7moqR`ZnaItdj?R(ZAD@!cLIMnDsch
z%*)s39x|G{(*6M9;5}av@`$O<gKPahhgZ?r3&(GKV>KeVHDh*YKtx(vApT)AGlZ&2
zzY*}xM(d11`#iP<%C=yX_$4Ft8*SG0n_PzfeNx@Ogc!F{f~|f%C{yk9JOORV+agAQ
zTj|ob)jFVfzst;)7)j9ALDy`r`RF<h_cuz2vA0L<j9I+)q7r+3*{_o+<dF{*qXH51
z$zoYWv^M>4gHA5Jy;sYVuQjjkTeqaXHT8jMg@Sk~;0~WpiV$ccC_+T>9!&zz5o&ih
z^V3p~zmlV(Ids&}XOQn6$NdwI{9zLWJuowRnCy<xPRLa_d6XmNKo9|J>7iTXEnRbG
z1{_A1ajowo#|enjPb#a$#cn={8|_>B<k}?<EVDX<36-!X_{E21DB{Va$ktm*7iZ`T
zwLQcxIs}g6+g8e!Ky;$N>=W(sWpV^lRRo-8XV$VV@?M70t`UyP*CDR#5tsZI*Gcl#
zypBEmh|BQr02>*-&%&TUe_L}LH(WWnnYy|7?NJuEGPrR4HojS$e1vh5<^Mi9=P`$T
z9et0aG-<Z8@9~qWYOK#=*kP|as5UTY_<$_C8{{+j`6%BfX2Nno2%i@HXot%tP@ka(
z=a%8-Fwnp2u3Ls=ejP1UYphwyLi6=u{$t*Dh!9*JAG{d-$hVg99BlD!KluXOyTFZK
z)r}^a;>}ikxvQETsxnnxD`4n(a>1zi3Fo05g&D$#=nc{VPB`BH%C%2!)4H4aa~Laf
zkufGps!|tOjXJP89CM6r1<uMNO;bC(17N%)m?E4&)P*bmq`I|!1E84Y{~L+isnK_(
zm>FsO{t<0O51BsU-{|fB`0-S~$NV>xPmxuK(t}UghmlysQfE5kSIa2mn`*(blP_Do
z%(on9*0*zwxm1(j^5AmAMef~WY1_Og$@<JE2kd(;#4a2%W|B@<wFS(oa@Arrd6nHy
z1QbE0rRh$8p2olpSr*CFy>M`I+!B(OIri9XGUQ^&VC57&l&l+209PP}0u(m%Wo-qp
zbY)2_myTzR`LGKv6*dCC(Znoa&1@Lb6+(^QU)#pFC|v>&6Roe2>x1H(=+Rf}n33=O
zk46_T6BDtEZ=9&}U<s`fKc+N^#f7lLdCFCoo&&*}^2i@8*lRZ|t;})lSXk>k(%^nV
z{PC7APMiMk$8#@yEJT|iti%pCgH{(HHH(6K3kct-WjALZe*~%*@81TF_CKUlD^E3I
z7>Qu5ks@f(+WEN&uMX;*Qvwf3y${A*7loWpB<KwMq;ma9byj1#hXv@@G|5c>hgL{1
zVpW61DAi8PW43Bn4j77CD#}$A*@SCZ_fOtGg;X3sIPri1P6~rh%V2h&F0jF{(xneD
zOOE&q7q7~NE>`$nybG_q>CN|2I=RQ<L2?g5^C>M9nz{kBz$h`iEd&(S)0z%NX9}iG
z*Q@Z*=~Ot|*lVX$lI}F*Z=ctC`1Sl3Xp>P2?u{L?#e;lCH+VHIxUx4oaH}KG2Jj0+
zdS-G!y)7EOI1R#TXcCZdKAWefh;Gfc^dl%60k=RK<nv2a$0-9BZ?x`hf6qx1sgXY;
zM}KzQti-Gna~x>K0sLM#KOrysxrdSweqq6iM{x}xPC%Yqt`X?hgt<kPuxU9siM%kC
zi8`xiJkkdS05U-U+{1}-QUk!{m=4U7R1g_eGaf{)D+w}nlaM6LEP|+5pDx1ntA7Lk
z(Loh;$~YH!{A#ctfAMuseh>OB0^z?}1mf*~15O$U9kw3!-DCz*eSZtFzN-R2&mkzW
zKdrm{N6LNGel7g_6DrPsF?RGU*B`^GCc7T-;5lmo`%wTuJik<leEE@jK!pnNZ9V{q
z#WDhI`S^?LRdI_69(4}0aUzYACwrxXWgFA3jp=(Az_Q##GSY%H-4oKz{OEWi+MvFy
zuDR%3r2p>^q$@26Ynv`O4<LNRyo+IJL#jR~a`i*nv1Q|eRz<8jeFlni5HeWoJU?Pm
zgYUes!;?0V=(O)+RZja8R=eU#80RJUY0_<Hid-EN9bzqxp-w2DSWeP6{9<1B#XW2K
zQKBX`g9fd=d2wThL)kQ(;P!P=xZlXQ4`{-$Zl@De@!6`A!SM8qTiQi|OMTxYuU|aT
zt*1^CAlvNmEQG&`pb@>PBs2pR-vqKUPwi9(d^)e@pkV%%E9^||+}AW4rBmf>IGF{1
zG-uiBv6i;DU{2N9m1q#s%n0{JDq4of%0NiWVP*1qh6ej%TMMlK1;se2Qz?9jZy%eN
z@)N&ctFAU&)m>)w->N+V_ZWD|tYx8BNP`~4%BfsSpMCu(5H}o*pEB$dLrUy4u@0Bi
zIOZ!I#y%ELDt7YpDc66qP6H*<n_b#>isjiQ=1o`PT?jCOUl?5Qq+Xr0Tu4*gnD^qw
z=>C`0k?4bReY|WmTmwHc(~}^C7lCCoE9A#U#_iU#aTFVxPX{T!Do)XDj^a|fj!J)G
z4)1&uuQ@fSL_0e_0A#h8&0$(a623@wwc-yk9!D9zMJ}eR`Nf6$k4DZuN+P_Ey}`8c
zPF{d3GybZU+@vHCNO-yUF4mzao8!r`<JNwzl$wsoG%c6KDTGe~E+PCf-i}zWWRnAa
z+UpJt@V%lN7B}YQH1QeUe70Us<TB-&i!84sK0f7xwo{=qg|FY2!IH<Xf0uW{uF~y6
zitw45&(+=2Ul}sD+j`LKL@<F?yS$#jn>$l%1gWbeMzKdnBos}!XcxGeVnz+H9)e??
z6&ABE)6Q71yiz$3xIReCBdGuZMXnMhem<ffIMOuH+^yWk+{$`Pmlm5}U6DGeNSx(z
zvcL)j%*NIvap=UKwXCS|pm2?2y1T~oK27-W69j;!IF{rj1XwIs!VowTDF?Mp;S<OT
zy-GK5W5;qV-7sUa6rrT<)9=0jgkk3n*v$BDQR1!qrcXVH9d3K3BJHxz%R>;`kL}K!
z4$7u?rPUd>+%!@z4Krdf%+>H&z2J<M$6YMrcX%#*s)YJ35nKq@`&+E^v{Rgqu75+j
z(9??5lNG&n7Tyo}>f#p&`7yrFzB?Mz3FySmUhr(J+~qh&zD}45rGPqfT+->&yIPZR
z2OiK};Qmm)fvO$jxa5`gxqkHk;vU;kxh~c~lN+s0m1=NrlWhb^NljYcLonQ06_WwZ
zIhwVbdlg%OyoJS-J_w_aG2XhVMZq-J0k^JC61t22JSa9_m1@9Ms!0BNX8uHgZW}zY
zM8$R02g3vNuo{t&vQcbSFrkz?)UI0w7q1yw->P^9B^0t*GadUjod}U2-k6bSRu)`Q
z_PE#U+eu0#CS_gn9dxfqET~Ujde@%+!b-}ow%<%KMPB%6!nJ8sb{O~sJZL6?2@W1m
z(w2?cSXa8Re^#0<W1T@Nz~XJ{x#3lss<FK76YtBcFW|je<{WIWYGT>D8~Gl|&;asl
z$v&4c!a@=`X{WxYXz5*a=(GQw&T<dLn*9KBH>Z>^HQo%@jkeEB)=6%D)nhT&3t-{4
z6^>WLrD~mH&W>fxIqiP@IQMp3cFFkBTy>?22i{}7<s&gkp9&TYT;pCWBN*WRxYJ6C
z?^Yq$1?`z@PMQwsBj3(MiL`cH7dbz;*E4K#K$?~q9mA!O`9RcYW>*SBhVB4`huBH-
zBehJqqk_0<%`c}rE4{SPX_=FJCfpbIr8{tT=S-!m5_%O~GBQ!9-nCx5*%!z`8prM(
z*lzWyd>RHyXWqX37vt`azsX+ZT2*8EdvxWP^53iAT}2SPPOFB!|7<Ss#4nje)ZN)5
z^TVMD+9S2LBfuw(nm&;NwD(Upocyy!l{0mo0McghIcBwc5s2=h;CT>J;w(k2=R9I0
zvxgGndT{rNGUIeeHd`L9^UH^TKDFe}GW$%oN4NPw0T;Nv2johUn<4Bfyf8>JXLnqH
z|LoOGDaSj5_5(g$frr@M3his#;uoWEkcpb>op?sMKW>MsicN>K3DQ%sAg&yITXY48
zaoKCt)WBLiXXe0Tql$;`8+)`#T@>>?|FL(EPqlwLaD_-?qeo?ld53AN>;&H8oRzF|
z1i{jy*LiOKBl}LeRm1Gxh9o#;%QN4bJ6g!D-EX~mJxrg2|AZqDxk0nT+*+`A4oz$A
zWKenC#+`N_Cwr8bSnqdM^K`tCEToS=MT>&Z0Gt6j#7iZxLXFQg!_bOgS|I8Zu`^gR
zP?B3!V{T?p_q6?TgYM@K{gxIfx9)87_?3mpBO@u|1feJ~6){VqQFFLKcvFpIB2J8x
zhptT)EfnVVG<|vD31nP7Nuj&4khxL2wOX!0fB_^np1g0eyJZ-7ggxA27fnOf?xR4%
z5PyHnh#6Kq?L(&nOG0sJZN?fQIBV<B#Pmm3+LdPYFRI(<p>~sQi(3HD|5RoIc^|-0
z;oSfWX%ktPI!j!&B*3$d6DH8rg&MiUk+!nnu;vNP7jeR3l6!LJ)X&P?9mY+^&-Z2-
z+a(ea#jvraEE>Ex68M?tU@^^TVFDPSYBCNAb~4I35*mc|cNN>p**|<f{k?PR%C@BJ
zhDUlzx^Aa7VoTigz#@NfKmNRtr$NyA7|xZFZW92yk7yLK1GuAi!DP+}1Ob4NFqZ-a
z1c%ua26qf`9zK9iYvH|6L)@MYlAs~0kJLe2{AcP7gWZSsn`JFuiufhiq4MI*9ID=U
zq+KiNk?=o6^?=QW?uuW96#t|;IkN11q4Tx%-@-n6ZIqxD1W~#MNEf{PW!BMmZ_BwT
z=M_N*n?MR-dUzFRiw0!dei4^rG}V9oK~Jt@{`eC>_^2O0sd|F-fv-~qA$wIizoqve
z-dRH|qVs=Jg<Byh@)~45Kk)XH(NC&|;*p<J+{NivsDFQ*f4Ov0lA$v(G6ct5_RNp#
zRwAY+T+J?(*?*xGTd<F(Ju&Sa|Doov704zfb{Ydv2|E2F4)2@1iUlFP00EU$Hk;Z^
zP$81H4R<v+$yus&W|OI6yET$gz*($4@fbzAgEUftC3@)Bp(1edDBpNv3l{B<lj_sw
z6RSCzuI`>Vu5ccrK)(5t>SY}K>QAcD8^9cW@4g^ek<ZCCB%_et)KXY(>v1qWOksHY
zcArmN;2svM*e+%sMN=m!{EhVDf#$_j`4h7jNP3o@FRNUEXe*T$c~IM6q2Nk!ngm6+
z^FSW-@E5!OWi^F(lI4%yD_P)>=i$tKb(dzLsx7>d<wTF79w03JvGcLeP(N1Hz3O>S
zlpvFk*4ddhoiJaFBP5zvk{}U>K39Y|SDa5!8?T~nskyu#;|sOiB3#qyR7u@a(R+o*
zOkA4Pli7HDt0a%EGwqO=$8{J~6{72I*R>%(b~qmkBtrr_yA&%z!vFJc`KZq#)$YF<
za`VE}FY7Adk88?G>fTA&U)ELbFH79Ci{)6{uge@js{O%Nq?D^s{r)=t&z|vHI{$Nf
z_?zwM>F>tvzKa46%MM3VW&L#=M?L_l**_8``A5e*o|O9qPme5w|9aHB$s37RQEXST
z@Dp_kn}n3gi+NJt&4Yg<e)V0HTLeD~LX5&yfmDhvkV>H`Lu{$RdcHCqa3jI14M3+l
zd*4ONk+2F0R<bK`Sb}mQA*v?B)$j3`TinUS7(Sm2<$^^UYYn9!^UVn+8CeIfapP*~
zAhBeGxAw01b;F5llcTg|@D2-*wy_cdAf%?Uu}VTX4>lMOWY&yniFqrH3vjyEUokCi
zXC2o-U#1YIZl>r(&GcU8)HVa1IWg^RTMk*a9|b+Be;x*DH06w_E+U3a1`Q1gB4k*j
z#mX8x=qx)d{iovEU$qN<u6sbzfz>N1&M>mMSo$VJkS`Z5v84~Dr<T8ixTd0g@_i}?
ztSvsvT;+TOc1gX;?{Pzp>y(0mtJh=8!m&=|^nq>m0`k0<<(P@n0r;pIp*1$K17=vs
z;_7=Ol_|4-7d5aZHjAnZ^;!1hcFr+TcUPpM(JpX4+$+eRu;?tZXyIwrW!8c^DM^03
zf9uNfV-Bt?4$X^O3~<6Mm<tAAWM|c$B}8oOcrcDO5zh2C9=cg^SnArc<M*+MejA}%
z;wR_0LaFU(K{ZWNL_30AYnB2Y`E&<KV*%4>EVqIwG&P%*+PEZ3Cz4d}CR$oHaT5+s
zsq_fii(R~T?~%K!{lVKsaCW!S2no5kU20vII=_}ZC2m5bI0a2-hjA`GCEr8y4b~xo
zXT=`Uu?>ip_>!))msdJzBuTL?Tx`M0#b{j&U`)lxS|1AiVioYkl8^yjx#0@Msq;kB
z&ZfhVxO_6WlrYp%QVvyaX-0cs)3xuw=Htvn++{jUJ_r1mK|}5kg?uE?f^diGGj{Bp
z!`TXjzF4esD&xW_GJ{K}$QLvaQRy!xqMa5U567)54}e`3GW`Z29=nrJ2GqWWr8`mR
z0L<rHv`Qxqba=)kn<$nH0E<NyAdxC^gg5M!Asl!e3J3)`=Qb%lm~Ak;E?8^cQc}Vy
zqruY0M(^l*@pBtJV<!r)<Uv5TsHotJ<4b3T`jz;LT@~2FeYH@UYxAi~M@aLet719@
zQaZgN^iPFS0$%>4D(d+DaIw?9lI(Gqzt4WfDg*KiwXdB<P^l(K07bf2KOm+n^(mUU
zYiJ;s^Y+7Q075egKH#hKXhLwnU4Vl}4VVs=<5N$kSA(uCFaKQIi!7e2iI;MW2!d*V
zPwJm}Z@kTb=1*jO<q+%SkM6IGw)s98*RpdR{fA$A<qxwJr(Zo5w+A@~bPhEFa;6QB
zvdM`Wzv$^vAF#zb5EzeWq<Y9~?EtWwCHRj^@Nb*v?+^JG$b>Z+3M~Q3fGj?Seq6>u
zN-b0E^4lwruRWj8W;f9K%P;bW#s?lYR{=g6V*;>P2Io^^tMi2Hf1Le;R)4i2G|f>Y
z<?I-7U`Et3Zm|LUZ1Q?VS~F*~E)teBhj3wGcjwyI(0L`(@Z#<F8=?C_fm4wP=c(t1
z^}y9oajFWgh~Iol;aMqn9u_HX@PBmK!Jte#Chv8sv`zcgdr^DvmC+Z2)oH@X6kCtf
zY;uyC9NM3iJdRNW+97;9MEqjVzpO)IH)1TP4G8cF9P0k)CF`<Ig-b)uEg!z*Wg6PK
zVz|;h@nN)b&uH%pm-~rC*3~h-k1)_)8Id$AW;QeYinb{zD1j4Zy4Z4E$0t+O;pO8E
zK0!#Oqv@lI->$HGgEE}y@7`6wtc*{QpQ1OV1ga}3Xb?pY9I!X_i<N`^VBxF;|7Qq~
z9v1SQm>Ku956+cX-cp?Y*>PC7=^#ITArE(@cvvwCnkx6GDAQ5#{qv*^(!Fb->(ZlL
zByG6!pb9%o(W;UN33#_+W?4L(iQ_6vIO{6g+9(&f;hr;aENsR6gTOiOS6?qiN_$zS
zVKhA2QYhfX((CZyDicIg#@2es*ix|af3c<>0|82p7OU)7$eFdeTZjS}!esDS^Kg|B
z5$c|cVp5-jUbo@Xr3<a{aX)Psm>6!Gt7>{;!|soZ*Wi~)tfUM%elO-f%INR6gJ0zI
z+e`jKPM>A&|E>ayB?R@xd^y2ktjR%{#3jUQQ>%V~QBozu&mf^qX{5+`@0znhWp%+1
zqP9^O$?A)C;Xz0rbwL07t7j@18t9(m)Eb|tSnmpCUn3p^$d*$Uk6>U*Y<yXWC|qxZ
ze1kyU{Me16Fi2rrN-bn0^VdA-eOTS+v3;f@2{KW~^!)3qvVezHj@Tz*{wJB{0FprD
z>iqniq3-xj0uNcS7cF2rX6u&4IFi0~%JPS`%OU26K<deh6-=o5aYu74j_X?0r`L_)
zX86d)EGf4K1u%9?Z5Z@nsZ5|>@gA2|jC1?-2BmX0@4cIAU!9say~h!8t(hANo?U!i
zgub+53mSq8c@|oj6OD9+L&j#Dhj+y{HXKt0HNn9PQlj6E-%t;|<E0c^fQnDqT)&aG
zR0DZ_c|?gH7sqJc2hw9#VE8UmZc_uX4$i`AV(FqT`MOYtov4jy?bzwc>sSXZPax6z
z+rvj%s|fuTL5%cvZ{HAw6XtJ?$vRwYQ0TDj`qWTd-%|2HY>xfKSaF?~h=xy|M$$XU
zm}d^67kJv_fo<~yoadwSA#W<24ryuWTnQdc%^p1*v((yC0r%x4akT`@$Z^CQuILMR
z+iK1BU?I&_S$x@A!yXdo)6S+Ll4X4xzN^bpp`xb!OItP-hV7zh_e8TECoG&{w|f^M
zR?Cytgo!`DR0i8PTx?^8>*G%*c;utAanCQJQmrzU=aOVUM{CRFhRrn5J(js5JaWu@
zpDD?63rHPlnUZ<ZCVmtMaY4#RCH-XI0VpXe@*|Me+{-_Uq@O&7HkY<JuO#nJ&wj0m
z%jMhU)9#<rsMM*l?wz6f;a*2?KmavnpCx9S*yznR+Fl;m1(mnBT%L?Xw&!(v(g)Rj
zS#Y!Z@x7xq<GAjRanHd4cd+w1mnI$vZzECb)Ykan!kplQb<30Q#X^P~Hz0z2NtzqH
z<3(l1Lu0>QHJvm%G2js*9e_ndikH9@XBin*WYgEujV-zNhWV_jQ@ewnslVvJSDY=K
zw-56`pVxjOK$YCrzLK_kNKtucR+Wkg)8C5TDzk~p6ltgtjP!l0pATc&YuVwtx^pBA
zU8uTK6=uhga+@x$_)8Oq=+`%4-tFDgaAv%nvskV~(~+$R^#===V-JSl)lAra$MhjX
zwc4|B-0|H4ri^dk<MdvLGXTm+(|41443^unSA$M0kbSl@VY{dift^^l*(h-eV|L{a
zxa=aBc>HOS4%NdM^Ljt_sfrJYIYfjFdPM0uPM6W7&v5|g(0J6$whdXU|8maiZq}VG
zHNG3EAc1F*sb5Q`0iMp5Eg*Ikhb3)HO)YS?Yz2nIg*>eRHzF9EZAB{yJjihPZC??K
z%ju)yzO7$Sm8tg{Egqth2s!Z@XG!nx(8JpV$m>}@sYrn&%Cjzg3dIuW(!27F=uHqF
z!-*<%Fo2Bkdj+|+N8$JSeM1?sVsqCuB$FPbnvYt_+ub+mvF1K^bM1<MKTxK&%WlN=
z$BzM5Au&tVKumvg9tlu)6{ic|Z2bjzoUp#SdB`peCLI7t-L}HNg|UB+`2SoUWn#O9
zJ;b}aLX@wA+R`bXUNxtVy*w>04@S41)mP1axybGGxx^2LcR^nqDH+v{2^Pm1=AuWc
z<&K0ME?=nDwl{t_H)y)Hh~^~9suOR`gwAZ9A_^Jx`*?av_g0eZjY9U!B>3Hv<}*?S
zX61NF&9m`&WELL;w~HZMwmw^rfEt@}+Z$o`>_{vBOois0J8+seRH}b_?o`79+6vFU
z&P>pzoM#N3V)I3C=oSEX&GSkQN6&(1K%cRBad|~sUG%)#MYr!0UWK!{oL!sc@jj$6
z6AoEjgiN@hUP2_?p!4~OXF~RYd`ew*usw1=e-Zh5dmK@g-()h(W%>nA8!fG@R#6+r
zV!eVEu0i{hr3*%zvv*^HKN<Xhf8m$EpfqwT2Ywc>5rHJ%2>L))7+f!b%jUMfT{c9x
zoFpm7W`~GM_c3N~s#x`olo@EPVs<e==!ddD5f@a8utFq}((GZvl<ydZY5d}|3T;Tl
zVRxq6jPGlypg2XVq~X1;%Z@Gv?zno3j#9_V+KD}PONG`J@pxiTlY%s+DGs7ji60!W
zc9G>Y{XA@zTBdXAb<3b^l)THagHHL0R_{>Tc{Q1_p|E`)jV!V(-Y*6p(?}o&HLB2{
z`*vQYEM6DDYi`|Lln_W#Zc>VGD^z59ERU<n{&lh4F0;BYH8wJ}BkqdW-vZJ*2?j^?
zKxx<Q1t7tJ5JM1mfvfg+k3YQj@t5`a_ngu15&v%~#cz51v9tt`$A2f9!v9#i5hp}O
zXi-QApgLQ{|359gFf%E(oI<#{Z7qRWrQD!5EFTEegw#?EKFbKuW~NhTHwe&dPz`Vw
zIM#XX>qFG)prNO}=ZE2s2ieBN^KWn2r#nTU1zZlr<FnWY;8J26p57v2aZS3^QZb&q
zOK&s@U2Id6{H&(&WpG)N4AD&$!mL1S`+BIjnzZEA?lz-*oJX)jZ~&RU`uM4hrbQn8
zxpH}aj=2XKLxUy-(&Fw_0HpW3jeMXB(L$n~mo@-Drd2~M7mCqn=3$i@CNNY*fy?Qy
zIBw$xK>GIEl)kvw*m$u23zG=V2URu1*B&zoQMPAmdgloy@fkOCX(ozyWbc)aSvN@P
zpUBZ_f+k5_*1vaZR(i?Uc%7AXfI>U(;6BU+twj$XxKSSApTpVrzQIjCxgnU~03HWp
zTY(zLT9*0}k(%na$zNP+q3tcApy`<L{d>bJaEIMj>BwntD1{4elNAx(#>*XA@so-b
z9EELDEst-~N$`DHneB)|kyMW|!9fem_4_K@p|$gedeI8fQ{>Hqx3#Ex1gl4<zIbG3
z2-i{8j<QITmy2a&^Mvq<rj)LP#(;-t@oZI)`T)|gw&KhEfo}qj`_(VNbLz^U<b19@
zjBMi4XT4Pnf#u>%!#~2ZjlcBmPq_@oxVg}jWCn`QMkxAMr-b@TZuq_2ntpc7FFVKi
z(}KSqNs+u6bhy7c$mIn^WAcg*B))HFpRxk{Hd>U>3S0(GJH7_M52s~vXn=UyqI$OD
z*4lLCXl=k$HoxEF>$MkVg#h5dABsdOoeP%I+$3sm0;&D~yf9?+T_}*3`~$2TfDxQV
zMq-wffk>}Cf^72lxa#jAEq|F@Os^1RBTL{~g{M(*^2JUA>p7w$2J48dS+7Oq$4y{^
zRhvwfbVE}+-FzRo%k<Q!hCY5giweIEK=L!WjEQd3`scK8V(e-pJ`K2|=qA09P4L^o
z_xk+Tn(ykXj)5!?jczlWhm1f@fdd4r4=X7X8|>7Gn*YR7R2fEgw-%A%C`%SZ^8(@h
zJs`Xvng|^JjsL?-K=|z!L-`wR=f5ux__tj3A8hIO5TI=0w_N_svHbsO4F6pc;oe(_
zww>(Fs@($DtN)WReej>=djDrR;{TcV{@Xwierw?0VUPczhFAY;FvxEW{MNuf+gIuL
zKJZUU?w|4bfA-Grec;ce@>f3lt%1K%+J83JZw>sJRQ}3mzcuhzO8d{|`mKRKlgeND
z?6(H~N@@SuT)#E&XHxkqpZ(UrUn%WBo9nj*{!A)=<+I-!_$#ITXLJ45z@JIwuYC4f
z1AnEo|7@<`8u&A*{FTpsYv8Yx_Mgr5TLXV4mA~@YpK3s_`;bHYL%CVkcg>`l`}h*`
zt37&mW#@MaU;eXS?LRizc9?{y@(jv6B37sH8td?yoQaTMZtx{jPz7&)QhogLld1$m
z67D^j`M<+-^bd^XCTA0lgh=Z^-|(JcY3*R1{F)jkbJKg-wQ=tsexEU>xu<pc)mhp=
z00??IC6Y{9!4QRNf#P0c0CWBTC<t@_{-_)xDW^98ddeXX{<S%x2VDr1hz4OO(l`I!
z0AA1(w7vNW;4I171R&~v1J(Tgpx-k3ZAZVC$Zt3Df1rot{5uC6D^2E(LS~U{YR`g|
zHD4$5bX-T%#px|M<)ycbnHVUN$(6!*luZ$L<2+j!^v-p-hWE|z{-JNvSkGzMq%gxf
zmtIaMHZ19y;o07#L!M1F`%`F%s1NlK;WF<=C>M#5(p@Q?bPw+yc&_Zz)6J#WF@Ad-
zVtTw@5tr5&%s-nNH>*$WHX>V@u2-68-sCCsW-0ljRZJuoN8d~Ck1P4HBmO}%T#B0~
zdS(Iy-neYLtbeG8?p%2O-PYpbEZ672S&nt{cHP@ZiRS*Q35g-(V4vK!k>RK;f}BQv
zO#p>soJWO-d3d`P{Q#u1%84&TXrbHwA^qX}@W<=aH#}126-(?*R3klWQT_-9#B@##
zC?2jcZ6#8C<91aA;Z0IQep4>-6OV>*>C5B2A9C#ZRNWn@<B;ORm#jFt0@<GdwD_A;
zM90~NuDv*jNhQ7~VxvxkCM#D;f?*=1!9%u(jw$uwh1ddv_@R9K*xti3;)8&hodL>4
z*v5X-B#22cZKklx^!kJ2*De{5KN?!}9cvWb(qyNOr}1&|L*`A?b1zUxa$`N-C$fdI
zvJdZ=O?=O!m@3Kc1>g}e7}tc*{cPqEiZHRZiC1=HSFxa?T57`3V{M&DPTW@REoWfO
zEd^-8B2$55&m?X7&QGdXP(}80H0>mhtYJT#Ur8qLuJunUC^5e^6(~yo@x|n5&d=|i
z{G8B5fGy%jwH!@pvN)2&Ow_+O0M~kJK`d-J1cfQ7!Lbnq(?^nrCk8$w{*d}q)GI~M
z;HR(f)QR?j9WNm!?<wP3^Ko5P&78DRN?Oy}6ISYxx6lwq<<}O_OZwtcM&8gHcWyuM
zW){53v6$YiXI-GGFSfpcMpusb*IJs7P3*-Rj+pJs3whK~v>ob47-xVcazPfvWSmFD
zq0kc*?y*&NfOK>`i*H*Rc@E)pIa))cNdQ`geMt2A>cG9heJn|C-D&ST0h1uT8D<Dm
z)vm(#4%OKN%^+EvE#v0AuGLf;ELa7+8zo$zg{8=p${D`3j{$fBASFbh*^wqqC9V=!
z+D$y@Ssj=Edq!sLgiBBGV!O}Qa~r{DL*b`Sk6-+X#a!d7R~g&8j8{pUL7Z5&R4>IT
z^X@kWHe+``R5J`RL#!t|PNlnrD@kfxg2$4>FI@PFL#hv^j!%UZ9C(S>)GOE$N#pSn
zAvWX*<q-Jhd`TB~&oi*Q2T_ZB1K(r@0KbKRwuBil^lW36{$3@(({-S?ycc^w{k{A9
zTH`DjHcXR1X5rjo43|YPOK(Q*S&E0NR8DgD0UUNulfS3#?PwbLv=kdEI#X$J<T*W5
zD&OJ`3ld}BJA;QrlC@@_>OC!4k~ShuoO9?){dEfamBsmE6L(2d{A(1q0`TOE7pEII
z<R{CIV(VyK@Fpr88O5$p;VIK#uQp@1_qyth2OnHaVGf!@CW)^r=k;l6eeX!v&jVeZ
zDj`3qJlsAM<`flwU$4?9XHAhD4wu#!FjPQ^WS7XSSwfw`bbUI<@Q$Lh$Fs#}<XgN|
zA8f7KTbDN2<X$O&JC^ziPK2GNm0skPtRd*HA0z3)P4R3Avpi0XSf}th5KnR1o_MbG
zYszpvL{jlJy<Q#}rFKm3W~Yz9NH5}S2%GLSkGAQc1l$AH!wGTalXg+k2<B>W|9Fj!
z3&t0HVSDJ6Ms<Rz^j8<~8DP0%#`tXrlNk@I(2Rua$Qjg7zE-YADW>A$?eDP-nF7yK
zXsFM1cAsQ0ewTlxW!=Sg;=ln49Mj#bZmH`D8(K}u034yYh><yXwb1(YS;iQE6amxN
z^Grb<Q?Lx;sJn$@B&zv#t}0eVQyP5sP3x*9)%{Oif8H+772ZB_vfdqnZ5)`f$!>{S
zs{@s|wlz))SKxPBhI4UNwPW6BaSulqCRK;yq1(?54lYh7r|;^Y4r!xM;UVj+IL6pR
zxc_2H;GjNt<+e7L#n+8<TB2I>krto_?cy5UL~3avvLrgjC1_Uhh^vKi4L@0c%M1aX
zx1L%$eEC?5rdhi}J-p)z{`1Vc*Y>8Hz4sSh)-UF!mQ`)ra!QC?Ro)K?Gl+H-nX#16
z);3%8XI<v~z6G=z3mHyanL2?Z*cQyh1ieab8MYJv<0IiWx1l8`;*VR4&A+mci6swj
zt3Ta&5IsonUoS3Lh7x4W@%v4=IM4~Q{2;j8cf!`j<<Ul#aT`{p>rwCLmv>HeAJ|Qh
z<%>suQa!*>PdO{B$LACilkW}sPMx<TW*85D_)@7|SXk&M#+0JT#uDpL#&lUoInbCQ
zeiRMo!POhJN~+;NtT2QsA*J6I7Soh2y>sn{t2gzw;>NQ)RY4Vm@STat+{yAyA|#b-
z-^>Wx17+%ycANM~b$4s5yITgGu{(s%QaY+bL1~+LWfGbo3~vdbNPt*+48&}^G7@r$
zM2~dyWzvj(5uC2`g&rnO+u?QVB2SkDZn!X?GBQbD8E%ZO#e-Wwv;eL0s7qSX4Y9cS
zvMAZs3p3P>(~|k7RN0oRzCAw{zhdXxnuXXcTeBJCd_?rmPGU!U&a}9=K|yKkqfK}2
zR*oOH{xbWvNdGLQ=dgw{Fv&>}n<x!$05QV(c|&K~O9bOKBuX6amFqPyb}(H&<u5${
zD&<1)19#tG6sgz<LC*J34&>0zP8{8V1TRlcAhPFzL24|<=`??GSLEI+080Ho0M!8Z
zNj*m#ZZ>-1ly|cfpqg^6(7N(}{lA->|7yVJ;}BpmCHMfuGI?o~R5fdfbrHhw;xuEp
zkM6^sCIT#J({$v~&K0N-M#i9U{D$zi{!@8!C+MF%ogVR>^&sx+lg?NmmhDYu5tTiS
zJVCQ&O_<Y5foG%GFnb8L<+Q_RZ?<kHW?Y%5+iEmDGwU)RAiz!KYO~3LNtfMOFe6JK
zZ!NpXjvnp#k@Xr$ZBysrH*!`EG0}DMK%IJ1jYC|aPE|P(H)j<kd*!%`K>zG{jWD2(
z-t`m5Di-<D!hcdZJJ;zFFfAZPr~Y;5rQSNQLkYV2<a@!Y*X}3pe!PzSmYXQJ`N|vR
zmU@(faD<+LjP~I4BPgtJW&B<OP8XrWR@(5++|s4XuTjlG<HZSU##eD&p|(P^J)>MW
zNP<L&s0G$aBc}}qip289Jou;`#M|z^t^-vrKQ&%cc*41p6|0IkSKtN5im?MKHQI7j
z^w%yaR#Oz*NBGy%9LNWE?F_7By6r{ajo|VGWHkITPMNk%Y$jNE*62%FpYL-aa81<Q
zsdWy7gkPiNtwV0J7jQr&VWBW@T;y8QtVDE;X;hUZKC>cuC5t)v9<RNQq+RnQ<a$kb
zloI+=6NizlryEHDeurRhiOwm2L0a|#hq9Y8Rh!tvkZxJL=6L`Gv*y~v82w5kfxAS9
z)BroI=hoIxH(BQ>5`AeRq%(veyFn&<6vQS57h!xx*2)nZKd$_+{WB-~3rv!I6xug<
z;uwdVfr%->Xny2QT5{bGc`t56d&;<eQLq2;!hjh<`_>pCB`eLfdqSl>d@qiO_&|g-
zDP+ee+07tnCRfvK&;`61#cqCkTf*9$X!h^+;>YC1sk#+S^P|v;aA~}Cr2#Aw2YM!T
z)dQNkH6*R<jveozW@0lwJL|>GeeIi^(nv>$AW@qTqs?T#JV}=#wVFV^e0Zw{@52kS
zxxKe8W-1RwDQ4a&y_QQa@uEm(mR7n2e);G{h~Uy;PpG((#`IFUGXIQe-pr^a?ywud
zZ<w^O{4UD#>Rg*Jd3EbaGaGmGr+%&OdzON0^(<4foh;e=X(?N|{ui#+Tbk{{fgQ#@
z*)3mWyJ{kYjo30R!M)0~tinaWL@M3-)B?)99G_NbHQpW7_T7N{v#o|)8(;R!R7}=M
zLIlP)$5}-RzgJa^b(iII&-~Qj>3wzpRn;1A@far0%}GpX1kM>#&Dy%4Y4uEKPScE4
zF{mNrQlIa-A$Db4COaU2x=3`=I3rKRE_H%;jXO!*+Y)me?3n%BQl%IV(Ix;#3llf?
zrcuX4>8sN9CzXYNtP4*$SOuxqXA1Rnad&ZWbjsgnaD7*OU`45ZkK7knAIS*W?h@B?
z3jj#b3dS~}ki1H#Nx-q#JO87cfq&IFFACt+6D0uR)X%{fN{r<dJeV96NP83x6lm>O
z_TA)WZ_Q?P4LRShr0SV#oe4HV+c0W|-#b;jcf<dq)+48;<L_GqMX8TV=3b2c7T?n}
zbs8-UgJ_}%Rq;?1&u!C)g{fv$(_yJD$Ti)IYBkxusG{jshojE?Qyq2}OQ!A&*IoSp
zTf9e54=)<aD3%1at%?ZPuzK8Qhw6)k1@y6}PDsAEMHP0cB(QBS2tw4XlP_}1KBwe>
zYs0UPV0SugEOCAJpM96S?O(*!#}TDv?*6hi_HmO~wfge%6heFonYrG_fbl&9WAb-v
zmaD(PWsfS9_z7Gswp{&1n`C*>kjqK1ov~oWW&@iQPnM_!NTtOmYL99@_jr=|r^?ea
zO8g{CYg-sXkH8*7z65$k27iK?ZnyZ!vekS>XTC+5>kv9Y$rsd{m)awtO`DHAQh^X7
z4V<wOFPLW!czrflx;~>zC*LxfJ#~e3U@fJA%gkO|>eJg7QA%aPo)rv*(TDzPi!+Km
z%#GMP>%kwF(IEWpV#LGBx|A~$7UM-CS2A3$ikeQ`6sdrBwdu%~S3#AbV&u)vIcs_O
zqa);z73K6I7P4Oa*45JK`Q%A_H>v=bte$X(&6b0pWBJN?Jw<zLT;UGJ>E&x?kVwJ7
zHz6m*5T&DcK%Hh}&DjJ*v9m2bo={&Q>rS~$7_{`Qb;xk2uSJzB4XmB%rgz~?J7y1O
zQjkM95gqX(&!u?m+EgS6XI~dO4>)U`&GmO@>lik^GHKXyTr`R1D^vI77j~QQu9}`G
zG!1q&i$BtOpw0xxNaqR~^>iNbask<SRASY;+qqt^&tDA}#YhRt92_W#pGijohZYFn
z;@yJD$YK7ng(-k#9cWjcg<Yyxp3^5;O}St<@^BNkvE2hRMd;JLyJM#$q)*L!iC>^B
znCEWmr=X~1hWu_ws+ZCptNLu4)U@}t52a5}#BS;)s1n2rW}ZJP5rojWH5m`2i@dg!
zPTa_i^l>|{#pyAg%<Wc3+bt5vIBM2ddFY;<FJ4%}MF&1>PIPN0-MQGw`X<6kF~XB=
z_e=C`8%r4z0dmI=DxVreqhhBGyD6%c7&$9uWQ2j)_!h?0!b-%Ha%-;B-)rD0i(Pzt
zt?JcTfyuYWyyk1?u^W?bx~CNVm)+l4eIR2_y{G6-K|O#@LSY!@d1CqWfp4|gj<HE?
z5>}q9SkhAT!}`-RKFyNz=VligEcn$;4NtY6o>r1288fKjh9jqgt@Q9iDSIey<syZU
zG~8;ihS}#NIf3jWg)V-&Q&*Xf_jO&PH+G_x2H5^X*hb;fsNuSLeS;B{vikuFxK0eg
zLp^=3$MMsoo_ZH|Q?u{p9fMXqEdc_B`RKI|=G02o`BVvZSo*#eKBZ~gIgqu}LM$_D
zHK2IUQPDp29)IQ&$#3f~&d#QeyR0)Wd$eK?Hn&1<lbAM(F<c>84j!>n6#y&b8M5aG
zq8|^JW@gE{UA5LOz-%raD`7G3+=YtKqEl0Dd^T<u@Ni~!ZsS`EF^A*`)1OI*ChNYR
zoPzmXnn`)OVtSbNJg01ML4xVKX-D6SjHJ?<U7zvhx9unEsd3p2<|?8ZzAnKy%=u+Y
z@f=rClrooI`Z#u~kt0fj`4;~jO-Dn#l-?~4JCtUm4Uyfc02WO$?yob#>7Q0BH~?d7
zLPEXiudSREFz<VfWlv|@>a&yZ5YRC(6S{aV*G>Isip-d_HH2>I7#XPA-JWW9ss2t@
zKgvOEp2ZS34e{CGqGXeIFu~rQ!ikC1h(?X%wL;%{tBX}HwJNXjIXb}zm^0jA!jv46
zYH*}3!C{2VPXe5+Zx@&P=-xQ6L3prfeD6uyIzT{4bsJlUB^#0|ij?kTSa~Ccsf2vB
z#O=_r5C`RimFzGmtYAfyZP%feS2#^mB?+a+1?;#>MTD=wV;7AKkJ6=(=O?v;eUZdS
zorHVYI)f-4&enxvF3RhPPnvG9MZND9d#A4ULjCrl`<n6<OUmBl=He#2DBZ@_taAC$
zyCcNjSfD%5mc+JAQDQxBnh1Y<cy$miyJt^0h3~6U;eYA|ak&xWePy81shyeIeOvYs
z%^nU_2PUbzV)J7}p-3{%$&8tpO{SF))~cb_(uHy9p`2<yRgbwRr3zSZ*Mht0tc$h$
z3N(av%q2@W&1e8_frE-@WEespBItiisPa4=c{Dcn;&aUUF`MqaD97{Hg|4VxlN=>8
z^gdF5!L(`c-~%juyXD?S5;%Qjab4x1Z2NcBb+7yM*Ra)g;VGbJkRZes_^6yzu_SLb
zS(S-iwj#`7dWN=D*pJ@sLbV%Y(clp{`%w&jlf*e|TQ)Q^WVo<gY-h8)7xOS`Ua6G4
zR9e>*#_?kIIvI3SpadpZwLT_3q}!yQaP?QPTIrRDfNQaRE^STVRPyIKuk`xX>r;@|
zR*Vq`0i$^2+Uh%)oqm1a`;j`eV%S|Px;XK%#>12q9+i%MJJL4;yExukRdc49^Y+$H
zg->@+Jqcm!nhw4>-ZKkff#vX5vYDpV3=N*7)4WA{=7lQzt100G84pRF<Fe$g^aKf+
zTC2sqs%@g*fW~$dbYl~-1@w9n+#Z|nfX&Ws<j-Vkx5{!tp3V-##in0GHU}oJIMTgp
z4;2@%vht)WT9C8XfNrVnf3AHl{TlCYOpu6DQG`P<!{#c<c@twUSVODT{TsKq8JwVU
z$G<nfD3nY7E;;|^M#cS?_qqm%J$u1@WS8PqSueNrh1VZop?10m)q#c3N4rm+^Yi)<
z(~ZBomX=0n#t1(dZSh5O<9V}Zf`eKtQSYl5JaLL&km9*1FW-CcRG4<~AG<|Y#-fzm
zt~5*KP@J*aG18;o6cpgR=~ag*$sRbcY<sedK&celO<QTN(|v_tIc+nhgTNdRbUrSC
z(-~?#PW#M0>9oL?H)5s2{j7}1)xDv$b*030<ZHN*!$+d{aF1F4POX@G2UaZb2Fsqm
zWEW?BbAKDBRfQw13oE>r1s->`F@w`-t!V}-)EJl+nS&YT?7WAMRiRJSNaw#;9&1gL
zs2RiSf2i~N;Tb$ef>uFA{3sjX=aNTRAU&k7e8BYYtl#(^ripnh>TbJ$wP)a>=IbEE
z<G+m=v$WnC&dW%X$X=A7j3B^9T3+6^DiD&xPpUm2pfU}YAU$-!PNvN=>;<xwe_8^M
z7BCSdniE%LWXmDl!+X7AJt^n4l!lrL)ny7rj*qrj4Fh-Xb^V#|lM+?L$Qc`nretRY
z$IOqHTOyx_47qrDi04&C>6DFqeQ1(&y?)~H+tLQQ<e+9q`E)sOyNi=PEe$#c*Cg;(
z+1Wkn{ItTB%9LH&BQU(~zPH!cM*p$D{HicZnIKNG%X+E_6Pi8}C*8xdl6{W3<+M|=
z3)cNLz*=<r@=@qe%DPb2j62l;rY~nyp!7ns)8%)Lp2;149(vjH1KmHV#45%+u(Hb@
z^|pslw{#g$<;l2DFsIg$z(;|gC|`YgRR>Q=qLnS@hc|uJ9eqU!zTAqr=Fyr1A#<;0
zZKhQ`Z5A6b(dh{-5|Nhi7zewVW#qe5m}qw#?#^IEZL1m$hX;R*SAme5$5RDXkHiQ%
z{d_Wi9jedQ%?k58L~=-4asSzf<(Jy3Xxu=DV(z&X`txH~y8Wr?0-oS>SE$(0oMf>w
zqMJ)`K;Vbml9j-TZC8cV;!g+1aVvJ5>Zi^5qAwrgS37oIqcb^4OW5ymx)mzFBpZ*=
zp1S3;O9pq4tZXR=c}h11q@*@86`5K_EZ349-lH%lv?%~&ZzXjG-iYfF54@-%PTxOf
zonb(FY9B>i*{|7j=vY9y<3?fV_bnhln2res-HU8ur(Eq{7e(dzR6WM>NmXvkS0#LV
z6Hsb@CFNRkB=n0i`%{b)_%u;pJ57fn_>;>3?BQ7BjWJiIT!G<hufI){vTQ!^zWDIe
z1!^iUU4U2Iv7T>LYk18U!vafd?)m6!rotJvX`J<=E8kr2tv1U^xpQoa{DN9f<8|Cp
zmBLLxQrY5I;8X<|))>4fyGP#}`+>p*B)#XVfG0nx7){z&Ar)<LN?=eF!(x+-GU!(C
zJAz3+GAar*>aqCvRK>ztaF+gfY)jiA<@F(3|BB+6xjUagi=^mt&9)QvX2)<6p*8GL
zh?y%SlgV5VSwMv&p)t)?CowLyU@ulFKs5@{Z$CE=gQSIL?X~tvkwk*bJ1@+{*%?k~
zQd6stjo>w7heMQ9vW{;B0#pM=%67Dh|F7dU|5+^Oo<Y7E%e{ZL1JIwnM@#$fA~ME5
zsW@TD_wbcTY3csYP#)}h4(Czvabj=0`Ce$f$S9c6&82$WQr1Kj`=W;n(N=@_Vc7YT
zN}I3vARzAR-1M+7S1gxRPH_(i-?J{_gd{6g66#07lnma&{C0PCO1}*1Wp%+M#QDPn
z?}Tlr-zyfr5fFOH=<CDlwk;5Gt!ZV5!BHfcW+teZ(NG~NKa#I*fbF#5%4OZ#eILf7
zsXsKEIV!Q9(Q3V6Jrak$K#XbJ3%TA$2F(z65{{3;7CI=pJaBoLoCo4>u*NE$G5Qgj
z=?j<qOeU{brD`aeN0=@p-8GC}THmT;r;JaKYgjvqumG(V%|O>$aH1Lvv)1%(rN#`P
z=G!mR5tWd3aT<6u)?@?=p>u|Rd|6!oa-Uh3MVIf}wAOrP_uNIp_e5k9=xI$P5WIq`
zp!@1*K7J?<!q1w9Ln^DbD@Yln7VR5QsJGWOf_An%zvHD`uCV*!C%yK84#)tLc(oP2
z>ig9;tRe(BT0(Y@^ea9hdsZRTu3K8O75KLB?u_><5Qg1Viy|@@zi))c>W8RdMscsK
zYq?fqrE&z$7sw~ml-=ITQapO&69c1MZhKD#k4NFRCik_?rS=~m;z@q){L`BlE+7gX
zctciX8$Lc%l#Cc@0UC}kPp%H~^ST9KD7Qkq4lE8qj(rejxD`GjL0X>|_HnZ4-KI&{
zSD2e&6`pmW3gHyvoYI2Y@qf{j|1Isf1!H!p*8P@O4rY@G=ghN<X85*erhpaH5QLr;
z`@V7{s3hAt?n?yOUwr)-`r?K@XLqk=@26^q+t+?H=vs&cjuAHJN2+l4dA`2DvVFF`
zJ?gJ0@qCI&EQA}o)M_JPFuaM1Q<JJ?P}DY$9q8_B7rY^w*!ot3C01MRf;D~8v4UBT
z*eq^5PuEc9PMR!d+{Xc1HxY9!pZLS=HNP$QhaWZ4xh`J6f9{RfjqSYbhA}YH$|QX?
zFjYufgXz?H&dpo>ySns+E|wPk0!*@_^EZcID6c&jmNjk)aB{wUGkGtbXvS7ch-o%^
zOcdXutXlQg-8k-w^-GY7#qsz^<4}WIq)Kz82*pe-%?6J1<oD7(13>a4Iogjru)Zbj
z;x?By)s6c}#em>R_<RK!M~33tnr+$33GoTjfIDh?4`b_WlQG+JUY%)2(Vg~L=+D`X
zwS-IF8M#hZRzkQvUM7!imaO_5@?Al^Fb^i%ED<Eo*tl()8Z6UQCoLN}T^ZRCb$4u3
z-&3mZPQPgUi7bopjoUQBI#c5>Ng|Z<eq_jads<EZB+!Bb(H^|oLFrS2+u#%STze@!
zc*iL#qF*u1Vz`G{Gr?8C?d6wRBYS8PSq|tag}5r%-Iafz3#QB12qEaem>@B>G$Z5(
zAM!&*MSW^-<XEp5xg3^psPrw|NvyT9R^Yl{b@rS@>ere#3;7#@`zQ(mV6@@S|Fl{!
zx2s=;>nDqkZ9)98CWhUIZW*;p_5&MY{8fpDB2np*X*b`M!WwI@Q6)vfj8f!i9=FX+
zvB71DF?Mf>AYG^e$!Lh|JCh+nOYAW~A+8tM$L1=J_fxl@)B32ffMAI@Prm({p!iiq
zv<gk=>#Yd)n5iI?q<Q<hsf4N&K>a^xylXf;=D?DuJg%kC_6qmV`}|BLi&&vn;Q?Z<
z9n#;WrCc0>y%RjYg_#Kwuc{v^4rWWwhLqD7>fh=LkVzn^yVO*L+7U8{X3c&;KdG!%
z9`L;yEAY;4ZH>0``*^8p7gm4Gc=(VycKo;(Dv_i>)LYPx3=t_Cc|h_DFLb?el%`dS
zFRI~5x2(I~CSc&Q&)mg#b3*`{^z?F;{)?+8$eYuM@W7r370@!9B_b4Ak(S&Pa#~hq
z53AZ}xgqAKRnqzC?e~vQwFE=%=X>Ax);nLX#$G;8cAuJXV=0=RAEmHs?_yShw*loc
zhQqx9>i-Cw@*PywbFNmWdsRb4Wa<EKA1nmx-x_H}dT5!48I4TBmu;GizY}^$7aW(j
z@?w>(Sn0GUr)b%2wV{gC&1&x4b7jJ%hQ{lUpbsJpjl&!j8uT#-LWg5R%wmGGW8|r&
zlDgOK(O2I8*-r9TJfoGEMjSBhFp_A2jeQCAE`b{GW@&E`KU9Fwidr$tu?0v4V{Dzt
zv@)0R5S%v^T9}c|XvrK5PRcUIW;Cv53qtBZ4-~yVt$p3N$Eb{toYjchskiXiWO4z!
z^!g@ntsY)3sg)@kuvV2aV^7vNSLOVr%6uIw?Q!j#F{7wBXT4%8nyQ?bDe)W*4h{<V
z&>{`k9M3|Q9dm;eySF|NZR5`o5paIo@pbmZdG$2BkZS+>j$8{R-wUfBC81=*lB~tH
z?7pGa^9oscO$oD8{rJ<d3$}V^HEnDGvepBHi0t;ZLwmk{8bI%VAX-eUcnoGhJlZ@(
z)+BgD_+%&o9iz2DQSZ~mWTx~8EN1C%F7pMLt*5#^8Qajz$VCqe7l(~Q;f_bSLPWFK
zIN7GpOF#t9-;Jk3A4v@Lc@Wqmq7E&x)U>?1!6H7()yd8X67ti~m`(_jnLIWmYV|+b
zyUwVlx^#^)ir*kr0RbsOC{n}*A_l@ms!^hR(gFcdfnb1$7>Tr4P<k(76i|vFp?s7O
znuH#jBK?CvfFRPOgc_0%?)lceKW1h*Yu1{3*UbFj=Q(HXz0P^}d!GI7_j&hzU<~*0
z>=iB^gt*YShA^teBvdLQhebcD<$=^-8WFFh4b>`fEs$rw1=_{T_p)1p6?@AOM*a5p
z{B_c_tBB-~bD}ZI&rHMOx^gCvZ=ao|+>1x0tn^uuSG-7R;k~QmPA*^PCn6B!Nd))E
zs;qEp7)OV(2Zye&ke-ZO7$DkXCJ!{1Nb445>ul04^F6e%wEh-t*%JzitHOzpy`1?t
z`V5IMPnCYK-fvicU8R2c{#p~Ulw93}qZ#5ho*|<mN5dXtSO3VQ6kKNO(U@;2<42w0
zgTzJYhLN2epzP6Qj-*5IAPj#<L-7{mc6!jIvl7lKCMa}VY?B{C7YFi}^>a*u^5>jR
zGE_n`2f!JNUJ%2omxnOCq8iAqMSdv%V`NDRDM49{kQjF;b?42da}bft5%3lq58d{f
zsbhIsJ4eaN4{~t%bF$(qcZ20uxZ!>tck<^y6LQKIG#F-QbiST<b+e=6RUWK12j3P!
zJmH%mM>vaZEC6{Vhi;FY@#5zXkzm)nd2t=!3~$d<%Wi$%DQ9V(XV5nry6%UAiEF7Y
z#wW>hn$U^p<e}J8wbxwfI8h`t59Wo*_mD!&n|OS@=AB;x&Te(POjI-#fl9UIi8UR5
z9{Vx4OICb9V?+8edgBFj0VnKWZiRfxUvRZ_OsorG9d~H$?ct^I!Gu{(#<-V(Ew$+L
zLPgT(uZ}wdI~vPt)OFounc$IhQKlaO1rHjV34I3>q*5*NM&{F%B8G0h>Uyf8?tX6;
zK3niWOR3Ly+SFWIi_#rMQ>|VRY9ZkEBYzvUXSy3#L4gl7DBrvwKBVQc8KQ%XBjbB0
z8kY9ip$DE=nd>fVqn}jx`40@f&%u(Kxe}4rp_qY!b!eVTtkvnv)>y@IO%2#|=KFxh
z3b*smV*g~#3};`Qa~1|74={`l6ZoB(+4lp#l#%J^kgXSUV%5g%Y!fJqewvV)#`~x1
zll3hZ;TAbFJN{_z8pkh@-;$13!vP7#@qrp~jvwQa&59flgUTkqR|&`9M4g+m0iodc
zo5mJWy&jxl(CW8KX7Mi@Q<_A1g~h7Shk|=Xz7q-9AnoD<!?SqPoFZrJZHM29WxgQ2
z<owasIwqSBg*=1X<?Zd`S50s8;@5WqCTVZzzXYC8YThyrJQG|Uosl#TZWkXx-3N8N
zEovWSlz$WSiM#)UdvI0<GQ989hPae-crn;oa8B`35uk+gT`}-&v%!!uPg5WvMfc*z
zOC8rR=%qfc(1fv7v??V(cFH5THbea_B?KTkHJzJ(#oDkI8pT<DFTrlTW|L`VY`b>w
zl6(SP#E&oTKxK+5A}Kpjr2f_~Ew_PXQnMbVX?Xee%0P1gSUMjY3NH_}4X`gCtmwLi
zR0+Qj4tGsIqwquh+a*Cmu@UU$&A=X@8UnfHgYy5Lm4npMzzaSN(hzXXDi7-^3B3zd
zr(Xn_tYs(OuHuhTQ^mQF55Y$`*I$kik3l9~C>n4=Vf`KN1T%$d>Om{T1K-m<g`2uw
zcVB+`X2Gif5Rf@F=FLJT5Qm96Nqqr`prN6a9!!ll3v|CuZ-)q0T3NOkEN9nr3mK-%
zkSY>y5C_0ev7(P!ulHNpd>Qyk?6Ldi1pP465fz7W-^2qCeIbpX)tjOOlZ&y*nj4bx
zLrV|CKZL*?ElV^88Rwr$cbvLX`nEN^qpih1+F?R7jYh=ec{ph@ATQ|sqGK48zUch1
zJN;3rC0T6%5**n^<~h;KQ}d(J_sO_Nf`4(DQs3n{JVyKMdo*3HpG?#=7JLULyW<5)
zT`g@rm#__OPCjvt8@4f9<f-jPoU1O_*6wZicI9v|BqtVMtHm~hV{K?Nt<_1S3Ij=?
zK8Uwz2p|@63)_8<OGwxb#eGcp?%NJ(mi3#t=k2-J5+Kxo8xb;QwiIHYo0O1OcDOm2
zEnYt7X~%|+*p^^(*Y<m*-8gi=NB6Gk=|3Nj8@1_!xRUt*Hik>Nphx_i_w_swj%IZ}
zhy?iz!$6yT1cj4F3QCto4IDEAv-tF3FAb#UWo2G6ioikiw1Df?=`;&%0V}9c=kaS9
zN+)8*<f9$EZC%668|t{$f1Vw62qKbzm@{{}BB#GB@%}1DA(4~OjfkN|s1`K;jQ#T+
z$>qf0&32wIL)KjjF?vF*$XFJDf~AEI{_YbVj~<hw|7@NA&)V&4U{bmIecZ+rXO_L#
zL+x(tUtXU7S(E>?d;2<>a$Mt2yxayLC&$u6X4UWVtSZgQZ3+PykRFr?4=m@8MIT|8
z=XZndT^>_}?PbT!SAP_f04vd3Y{-rz1|jfI>%;Ah;{T6B{=-q!u~<HzafCKT$sU^*
z-x7bsP#>vPL@9o;69Gl|n}0e-E5tS40O%K^FThCSG$G6#S7sj_{i05W!jFxFllr7w
zQt7vPuT1HVitO`ruRCM*jH`@wrh|MTp_o+db3$iL-)F@i^ht=}y`#^F=%oHSvl-a=
zhPc6w(*JY{=lZ3p2I0J<4VhG}-Q{`j4hTz;0U4*@SmqTi$gKSUZh0EmmLgZEr~a$&
zIj(4AH%lDj%-#kPu9R5-%QDxvFVbuQDFYmX)tC9?+&}C_m3Oa~pPBbxXP47i>1#F&
z{p3;nHChH5?e23#zpv5nT<rlvkv2-U*N_Y$r-U3J7H<rdcOOgm_K5c@%i3R+*NpuD
zw!Ib0<z{e|V;9>YGc*XBg=6BsWIj2MBY}2#GO&R7#M~|q9myR{<OWRvjs$p4c)|Q$
z+dW(Tec#Y(i63>EV=orszu&E`s9aTaJ>2n1q^(|D?HmzibrQ(W{}nEk&%Qeh$Z7l~
zHf8-aEG<|@@O1qRH2p0s?e+3MkiBOY9?G7-{O=sU_a6Zh^Byw%Uyz}5rSR<t`sMim
zXwvGE=8wq;<F-ro*Y}^*1;_N;<Q(0%LXI?$WxkmTZH<*{xZ{UX&i#OBE3q5v=xNt1
zg-&ZF=_EH6H9jz-wS7T(dpW~aLKj6QaEGx5{McJxi@eqYsBEm5Yrm48b@gMr!@o(&
zA8fPAj!_@<=86H4hqT5+xo{`RK@4m^3g4O`(C0XufJiX!e=;d!xtM)Yt^W7}0c0R(
z=k+@JrQQ_k6+DTiFxa}RHjfRl5b<2P^(OU|e0*ieJhnKA+g(IJl9ES@d$*_oTk8fY
zY%O5|yM$)zONNe{d}6Z{B{n<Ci2Y!RiAdDAAh1c>!9y!*J?GWs)rrK#4n&zaHPk$2
zDK-8X3qM(_I8~?)2qtq?n3N{kF<@>&T}^Z-?eVOWrBdTqk3L*|)YVIcYeY(4>1|S`
zqEA~$nJJ`r_&8f&r|R`^v57_qBUGwsEWrV#0}$!1h(5#S%)04|MlktXED=Xa6vQ3i
zrLa^uL-~D>&&T>Y`4SSbj#ySD-0Vh53?j}8*GyG~O=v_ky)gW*kxEu-K)%I>7pG9D
zJGt1wAe35j#V06%GtKqbCKOfQcamUk(9u6|QWwIO)^~KM(2{5Koh_{K5DZ#ij-XN0
zld4ov7qnV^+En%?xpvJc_(xQQQ#kuNHbPUPwSqL5g0J}W0pTHzH|b{`%$jT~%I?o)
zgB2X0y=w{-RX2AVR}8!ImWjXDOnvJRSTk>yCeN6O;FxH34Zac|NnRK>G04y?&qRWr
z$`fN`8Vw7s#@8*r)493trtPNtSMS1FhaF-uHnbTY8$MAM!@k*n$^{xWUG5&>F|m3(
z`b4~Be{`6{nUB<j>Edf>AhTXnH-jp!8e8xff-P`mFgxO*Z?Tb9%(O|ZD5<fH{`BJv
zxv=u(&@EE+-Reha`q>gDXn8kMfLmqq^T1R=<t9nKjvC8l#EMzVNdpBeYn5uOQ^vf~
zZCqLxFc#&SGQa$=#I!&fY5mmtcQMd5V<JlUv6k2@UggJBE;DlKLPV_@eO%gUPD;^k
z%;KSESL{7^GydZuCg?-Ac6pljCmdVSe9@Zm<Nd@X2royv&s+0d%oxXT5vbK@>(w~(
z<tWuqvPwbPl@x*-QaiDdM=s5nI!SIHZ!aP#6+b3!;yH=}!5e7pQz`lp0EJ`sZ$5bV
zKP`#!llQp~heRxbZ+v9RU4LVC<8hwyJFje13gqtUKr)kuPeY=T{^U~70W?NNnLpZy
zCZopKLYJYElPGJ#2Cm00^2mA8BW8jnnTfE(20fIWhsfL}xpvso@iS>FIwLK{(h^qT
KW5R>m9sX~}(<E5{

literal 0
HcmV?d00001

diff --git a/_images/bonsai-predictcentroids.jpg b/_images/bonsai-predictcentroids.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..e284f2338ae73e228178ef7866f24bf4e114d118
GIT binary patch
literal 76547
zcmeFZcU)87wl5l_3yAa%A|NVFk)l!pB29jXpi+g1NR1HbK?1SSMFa#C6r>9XkrG0W
zbOGrl5+EU=Nl&N&LUOb3d*|$Z?|b*&bN6}w-1pflnV-*^pS9*%W6U|`n4^5>qW`AP
z0#4sDF*gA)F);z4j30nL1uz0IpE&WiKjUCw{8>3zSy@<EPqMSKad4mH=H@!d#l>@q
z|1{4jzEfOWr-e`R2?z=a332m^hzbje@(T(H{_P=5%#1cHtemW@oPs=DJc9q*2mJ?t
zkK;rF3!a(jEZ_tm6EhzZy#pWv05GvKsQnwk|M$Ukf<Yr2I|t`UF2)VW(|{99%*-cP
znEysK<L(H?bpQ(=E5GbzLpA|xSN601g6c1lb2;R0RJID)43p%qJPLTpc~V$JR7_mq
zoZ|Tl7d14sw69*fZgkVw#MI3E*4=yeZSCwG9Niwfdw6<z`ve9BhlE1Ip1pb<6&({B
z7oU>)J}o`t!^h0L{DQ)w;;$v8Rn;}fT2x(qLtA@CXIJ;np5BqsvGEDa?@26yI5)qr
zxb$avg}k-Bv%9xXIXL_qUrYez|BJ1E=j^}Zi;uzA2^JP+7WTjK#dIQ+QJDEySY<D>
z@f%vRyZQ^9Re!-Dcq2KtvXxWriVaEVQNZv?VR;RL0{L&O{a4QZV~oB0r#Sm}#{SM1
z7QoHS#27qgJ^&DK<YrGH9H*Ay?dgC=2Wa<fGuR+<E@VRP>SUMyJvxB!1-QaY2c+g4
zS2QINi>gC&DX$#ss)u%YPu~i@@E*H7=$CRHw-k1zPkh!*QY%a{xLcOANo0oI943rE
zB5joCEn`l1`-A;UkQJ(YU559g!$cpt>9p-iQ-o1l%zw}nFW$rFbin))9RSX!i8Ir_
z01vLw0oc`_N^+epib&GicbXDE7t?uboThwV82XO>^)7e*EwdTt(igpF*7NL*@OA_E
z7&{%H2t0C@Aq~mwGSLA&gEWkb03k;JOQ{%GSsb~R{}#Kq*xJ#{q;9BmbN*q@`zp!M
zE_L=>m@n6Bp3wnR4fbOL_*NhtfE=a+co<K(BSZZ}2doFr3*KQz&;kFWcNiK2{?+`S
z?z#xgS)eGq&9}fZV$Z5FEU-o&l(p!2viRB$+WxY$Zb)^%xLKL>1XJ{brJS}V%&aLj
zUOQ)(N(UUOUf)fSp=_b(fC?}*3UK#`XRrXYKS>7=z5ad*)>`d8^iAsEf^b+#bvPyt
zq-i0z*Iun{a@%xEmP%a171{TQc1Z?4FpUO!Be<cBgpC;D5s!}Jk#uEwd#;mxzfWAq
zxpY@?v!9QDB)JNB`s+IkOjB!2amWTbAV-asR{RwhtaS*_&|WB62D;YoB+!n4R7;tE
zKl8?1BMwU0LxDD==Cn0!{yw6V)zP`}FZVB|9A|s)o2QF9f4};(r^n@T3h$YXBk;bD
zS48yEh}zK_9YEIj`>AF-^)ZE9KoxmAwT@M-Mp$V^?Uha(YCIK=9_Z)WI2Wi<sVZnL
ziWRLE7QH*5$KOTP<WdU%+O8bfRrS^H%Ow_1e5P>e>LKkLZBeTnynB2S{_iyY1GU2j
zZ{p-qMDD*kugoc7pZ#G^x+cxdh7JI+A0em{WkXEHf_MJe!Y)H*CbCn{pdZo!21lXT
zrojW>h4m@aaq6dyW2JX=Kr}-OeC4IxhVM04VL8^t`RD*d(s40J=?`8~1!)O19ruGN
z5qV?`#|xr$d|md}eA$*iC_AMq#UwLdIp1TzBzHB}XANDpCZ!s2_rICMLI0>C!dL~F
zR_TC$P!y*BF#9p@tPP~m|Krxs+o?pb^D_t~!&Gyj)Y9Kq6Z_SSkSV5%N_=&CKAoHC
zzS5SoiQV-b_T+#5CI2}R|JC~6R)GJVf&Evn{j1mhZUp?V)WHAQ(t?<?I*`!ZJUq^!
z157lbQg2_BBi07Lc>CMqy1yE^K0;+aD>DutDJN%rFt{p<l|4IGGN6NLp>)$s)&pti
zVme?r(S9d%8%2O`OAY4Clcq=iv|OeGUaPh5AF91BPO{`GFm*kbxv!w;f!<P?ar5f?
zasNsUMDES**9w^S(lYLhJ|AgjFPUJD(2(;j;!$-{!SLbBru(JJ&er(IZK1S&NVw|G
zW*!R1gH3Y4z}3Djk=^6S=3Ra75vtM?Y}%m%)Z#vhyiA5+p03{Y4(v*1Ed5>?Ax<c3
z(YP--T_ofadRnL2aRllIp_nVkkdp8@v8Ek)3%bU8oVt}o$6vMZwZ6a49;v!rg7iG0
z)?Y(>IbSFff&&$+(E)N4v=aaRr<@Qx9==T#`AiidBt@$R4aTRepsq`oKnQ(Z%?xfs
za769Ng370UV#M>hFP<7g_oYRYwl=3+c8vvng```XJ~c)OS{tzb2aEJ3NseYLF?dLW
zAB#vm>rA4CIne<aP1=p7od2=q165p)Ajt6hVAY~`lh}VdgTUlL|1UZYHy)>RS>9r+
zO^M{VSZT2r_F;#MPw_=^S^@RqSda`h+;T)YNbCzG-2Ei93P-z2221BkCGFW@UtVGk
zb8-hP<?054EQW7JV1hAQe_E_?yPV;@INKu}`q}V-%$fe#to%9eZt7X_N=gS#+vyM0
z{Ww+oorV?aTNb<~<lqLA>j%r?thyF>Xlg^1X3~omxiJkLc2Z{=<`vAf6lOh9;RpLG
z(=c-p_bbufX<C047#!D~{wsv}s(TMSYu)jJ@%c{ELxwT+@&GlTwhavLN9~u;9vnVX
z(}D@jL0(;gX0=^+Y9&Sc0z0k)U%;FSW@3hMD|bs-QCkJ^%fs2TgYp^bovOsGu#udW
z$INAvkF((keVdX+C9}<7C(4TmcaI!i<do$(Z_KT;z4G@hI&Wl#n}(cAF_P<UwvQle
zB~^VbTZF(@Z8sYAe3x_e&X$w#5uvq7E%+<M&Dd-%$}7TPY-WE~5@E*3X1Gc0Yfh|q
z=w3vl@7P9#ZtTXBFOr+@zmEU@^2cvh&kRx6!bUOKYz3Z#=@j{xm3-4++qc1T^sl$q
z>?@{NsrM*8+xSx`(s#;g)e@+rb`%`kys)x)97>(eIc_?%9~?ttT4@Y}LbG)DAMCLa
zQEOvCz&}}-=IDSL#pxkhhRo6J79ENOsb~Wd)4~nnTWvV*&<fc@fk;WXZeJFy?QY~8
zH$TgVwTB#b@3!(oz-u5ZGwmU8kIZ<&>d*6zI){-WG8QsGnG;+<-WeAh)Q`B>yVNCu
zxxV`?#7d#__<8nf&##9fnP;|)71Np7oI99PKSB7s8iK@c$7aUFY^Jt~NNeyfV@390
zZN@d?LAmc8S+|9+T<pc&{<W^S(Eg|;vsAq3mD{&z?`oXkl11@@<BD-q)j_}kM0$f`
zuutxd<l6{?JpEOWT6+c^kW2-o#p!!q2o9x&p%2vR)pDM%WQ$DLwPhd(GiNQ8TKkD+
z6~y6+9dA!&E8*P7z{y_nPd9B~)0W8uO(mzzeT2#rjPQ<>M2_t&YdW6<iD(Y~eoUe^
z1b-98Km5ocT-7Jn&6M%U0;M8*@U$Ix)B3hr28pUd2mC~IfaWoO(;7)1HLRFcGHr`X
zCn+GqEW@!8ppd-w<>f&zAl-Wwti`{6@8KQQ36|#qGHnLvSrE@kARPQPo({08(U;C1
z*uvZ-6=0Q<bbL?V%n(sltb1Tyd(#&5v5NW?Xku;dm9m9eqQR4I+l6jJHoa*viIIa1
zI$ohTIaKYiY~{$_RM@lE;yeSdRSTHfnLRIe?MeiyUyPY%JB8CrkH1)`>KSsw_T>4z
z0@we9MgPxw0{^pi?BkJ_C<%m|=x66L!im6InRO-n+L2k!`ngq`y1GvVDJ*8bc_~I)
z#&Vv1fV=&cvubhD{FL|DH;%G(BAJ0_ao&eTsewf}r%jfwK}eurK<RgOuu!}2RYkTY
z{9ZI_!mmP~yOVHqB&B8#5vK$4XI5<0!XaX`{;>cT!A#*u^{dHG&2;D3vnkD_qmFD|
z5+n`=Dx>sl=(;0Tq$P&k$6!}{4s~s7&C71)U9}s?H1-g=yXldJVq*qy^J}Mnp2--@
zV0fq%DEZWQXB+sz6;R62U?Hs?LI>!|(E%;29?%@yVVDG=*$y<p9rEmU23Tc%p*scC
z8?B+bjx=<OvU$p+D0z}HHz)G82{em{bx|YDv@{n{BIm54yYQ_xl$h@8<;bz+Juf;S
zOz3q;2D@G0)5i|lH7Sy;XC7a%I=TZ*y;V0tyJ}B0AV(JyW&-dx(t#X*w4ROkzt`T|
zEMiMBThDiQBKnC3!7s~iO=$i!>~ZSO)Q4jb!K=T6h=n$nGMplT6Li3fJ?Yf1-kGBh
z5}dp_P>*d>z!zjq@!~&=w1Q3;W9O7#KMyE~wqZtS@(5UsovCc7|6bdiE#@a%n|ae*
zK3=5-T$-9v(yYz*16SW=3`-?K#8<RX*>W`#P4~<G=(8krW$L{}mS9(|IMM+g-u+*k
zdIcZIvDHLv{F0qF`75k^o2LVFnOxR(2z(1C@FujjXkvmVgW>EHv>9o!xC|Vsw09J(
zMdX4`P71FHU13Rn9epoRUQI2%Le*!`5`A<VGA`e+x&Y1J+`0gaQ{2Z>a|lb2QkVp}
zk+7eH%yPbB+>#Q8aPxkmSU2KVULGuHU7tBZv`<(1BC_~Pl@3tTyN0-N#Dk`il4um6
zxomL?frKFn7tYQ|7{p^b_0Pn9B#wST#Z@$E^*ovF3lm6Ih-<yHmD`iyWy9sxP6v46
z<7nDs#U!ejXX$FDytHC+#b81Zh+?rTWqJo-bm4J^Iau~oRu$xY87JS##Q2f0S3x~r
zaWeu3Q>fM71M2(Na9Q3{eVD`1z_oH4U!|L)e!Eg87CJooBj>V-8_Vfea}y0cLPLJs
z2k@I;1>7S~Ia-RFOWx^`c;4Il<t6V8-oi*Y&T&;Fr<(RgwaJ>;90?bP$zhOEwt7a+
zjdj}<<%DXewEXsEKIy=Gjvw3iV^hQAv-@sNsKE(5lwu+;Ck5RhUXfKe{m0QcZ6>=h
zu*BG{z{0(-dAGkoOx{6(Ph*QS*3(9`gf%PVQ*A&(IDG3iFnHp);?N#vy|93CqDi6`
zX>)6qw&6VDgX`MDDdsXCC7pbq%-2N>ti*?ye48?KP45zP{}9CzpM4RtBg3-Hd~YIK
zM@gn;3LDm9WPUWcXm<9&#nf%{v#S0tK79gM!b&3M;8k8Nmq0NtTH`QN5)E%d%KW{W
z52_nKfwL?t2%Q+h_sFN*ik6XkCMK$u+4o%Z>k{oECDmj+f1F6RdlPK0uh=<vtdGOD
z<xJUKZp-umrxV99x9zqvg;9Jxd+7-Z(@B>k&)=IHJYTVaSL&t%QjU<jgI?QVQzXbv
z_1ezHU|q}KXSIwgBNmm-h*{uIw3w41Zx$+#%Vb9i$ggS2O=FtAx3#NBsP{t@N7vhK
z^$6#F8Hg4=6MOco^Nrz}#04$UTcHu*w{UCjGpAhEro&H1MB`~tQ#t^4QvdI7urfA_
z{fFv5ys$X$`I&!{QNh2r&5j4WTGOHfw%dS5v(|eO@B@O(G3ny7^Cx(}PtgH`9CX0u
z3)A1ebimSU`2LTQKVLZhzj;3M$2#M;TMPGC?@*bBo+^QFM@d*A6~eYM(BHp>BPcXR
z@T?+d$QGAH2b@_29nPbF-eEy|CLN(m^4AzPMd+@kf*_O3<Q#2t#p_5HK67lTFg~Vr
zNn>BY6ms%-S>a*nwQl(dt+csKZZPJ?lqH#=q^70cVe!UFv|AlRY?yy+YHkcL-GkBr
zjfa`~gEIT@yEfJaLY||(8z(=WPO0ybiUH;BeWe2`eSvj!!2T{B(1oxz5QP6jJ!#F!
zY6EzD3@4Y3-C>pa_gemu@aRzz9RLSWeRc(zO#fc>qIK5L81@w&LkEaA8w0HWu@3x)
zNB`?p|L+~>ow}@5*)CUl$=$|uQ*`feBGEXQ4v?;f*Wr7g>pIN|*Sis<<nEDGXh{%+
z(H9EDc04*KutvF`<Hg>_#VK<@#d@*WpJ#T_{YS=wgw3cHcFJcW{3TAXiX;-}QE6H~
z5{#||{WcALA(`UTQ`}a;bWv2k;0^hc9TFBo!sXAYu~HBQ!|SxkGM{YTiM4O*`=w#k
zX7BP;7iCQxp7)&L-U3Hac#mW$U{i_}X+TZtPeV~I5yCs^c>B<!lBhFd($!15`a?yk
z>_KyBn!?L9%Y4ey3!{{*^q%gGrg}6>v&P0XDpy8m|AA=+8G>=?E^vaD&5kd{t5P)~
z-=M;9K5STy*YI3J75FFYyQ3xPRddL5i^ory`t3<B)uarfku%~oqDh39Uek?bOVBJN
z#OP@sxYsVRBMM>tUn6^Hx@2H`4z~El+X>!ND>OZ_gv%e))`PCdDX${!dm(WjlVzU7
zA%Ap#=&0cQ70G-2iFSt4=n{VHWJ!icX@&!taL`ACc9`9AuAZpH*sgT4Q6V-?AzG^=
zK|aN#*ObFt^V8!QeGlZqZJeLV)Mkw@diL|x3NYxlN6Rg9&|(%`EpsO)7E}XjgPf!q
zkwjiK>ABAiiu>T?$tV7VrXPCbdW|kSu&ej*%T>1XG~ONv#{}k4+6f(zoONViyaEN1
zw|ocR_b6CG<q>wJ)xpUp|3Q!7#VYc)Y0`Ug`YK470-E*gw5pE$g%i%2B@g-1B*E6x
zt1hyZAj|MDHzQ<W6u;Y*IN*fVg6#m?=$J79IO$2ucpW4YEj}~&8i+baN=gd;?6cn{
zXI;Y^3J&^V7%b7$#w>5IEHq=)KNjeS9a(@_G{X><ijYq^Jv)x>A-l*L$1rR@l4cFk
z#Dns6gUPG2@aITK%tWYBJNAwA)dsMyn)PU8yGmZv+(62WSq>B|D<kxm{{yWo^5I-p
zqy&Zvo`pQO%~qpC6yTfRhfZLgR+N>>+k^%7yFaQOP`EE!8~ss?{9{lkXKwR#<Q1|Y
zo(*PMKozefm3m35G>#+TC!ygs-;WGl+Q&!a{{C>r`@@@tvKDpykR^Vq90}8fpDZS@
z<&qlP@TbbDcO!~@v}H(dq1j9Kk(L+AJwF%nhKYSty+5}ts%|R5*3E;<)84N~wPZgt
zrj-Y^G&*+n6*e&vUa9!4Bpj-FFaV?LKuT!SdZvrJP*2KuwK)F8DZvTIFX61e&<+SW
z2o&=D@twu(49OdJi5%LR6oay@#a*0e;v|%1eQg<roT?0V_u4kO&HUxv%f5FerjmC3
zIRa=;Iv}25UP?9v&xx*UqJ_M1>ZIn1G=JV#8#`cs!Q?i%?2&upa~Zxhi2C*$7ChYn
zm8~TjypzCAahVQCOaLkxyWf7u-mWwQ=r8dJL2WtW%{JUn+gdjs%SwJpcD+B<6>)EV
zEGz}%tS2~w-nxgtWQ}ZG9#`4{B`MH0JcEyC_EF!#Inw$ZlyD*{rinGAMMea>m~uE?
z`!%ooNY8gn*m*01MdXgdr2_Srf$_^K`f@N%^biEu2NIy_)DJ&-Z05InKA4NG78N?s
znPM<?ul5wvnNN~-Kc&TvKj^DMLG8ek#@#PBf+d=@2^;kUXA>;uZl$_|^Mj@q!$5_$
zGE*exoK^r}8QaK6y?W$GQ$h*wwdPRv*SQmrNQp*E1XNY0&kGx-@a96Mwl`~IpTLHH
zeV0XwM1>!A3$<4cR}E8|>b;<k+X@#mu1y68hlhPNp6cE51tXL4G&cpPdRDBzDOv3?
zgo*M(XS}hebUwVy32SfWT$}cmaQ35%OP@5`T-QQH_}*(uA1P}OPJ+j(?Q&4X0*Mh?
ztq9*HmEdr1@9^;Bv77Fm2lcHSGuOorFdd*ORLACNifK3YGvXCtSQTbR7P&<hLcJ!A
zsaX7WbgC4vR4R+r?LJy?^Uw`;y9+a+u#%ljhv437+{<|Gq@)dfd7q@O-ueo!Y`;f6
zjQ-0v+CNbJRuMq|PU*eWRu|-(`p?&#Onvpj9<EP?S%5as>7<3F2U8%@VhVaXJ?p4%
zX$O}>R+`<p&K{*owFsMqbF8$S8TAd8h>*MAR2>cbE57xPIr5Wx?Yx?N(7H>CE*Bw1
zU01IlnGSHVhLbf(OG#)?nb?rM&=|%f3DzKM<q~`aEbE$*?hOa?@ZQVDam+z?b+~pF
zSadk-P}|zvrwxfMYa}G{I43$NA>#g%(Yx^jq8k}jHnnXJpI?nV&0#r-<yeki&Qa4?
zA6{M#1;bJ>zZPF*C<~&zwO8HYoz1!}#&C+Et|lFDV=s*kc+Q|h1K!3-KpcwHAgCu2
z=zzHLUZpE{d1<Guc;G|Lh45KH5F?dHdI<&7qSlq@06-oiy9o}62Qa*f*8=UJ9W`$d
zg;j(O5HO{l$!7Snx#73yfFI}4l+5Y8=a(3%dlnMrm<wx1V=|@#I)>>04#jBnAD0m{
z>7*L%9P0cRZs7ltqVvrEBItiUX8-D;|CLn#tB3y2?;(G*sIFq}a?Yt;aF(8gRZYE>
z;*%lAyz7gP*;%;Xa>^t=3x0~1fX{ze1?@3hWz%j0DLH7>k$G~t8>zPZYgdd?u(!&W
z#8$Co`9D8in$3q4FqGJlHA891f#?AAgbX#!62q8iYggCN{!0g67{>mzn{@X?6`G(=
z2NcUt6X}58Lu_cOA~xxmlcDU6Y2O*3p;910e+>S|jNxZ(?#9po3bv+HMf?uK^-4w`
zi8C(#tL^`f`|3~4XQQLLR`E-HTTQ>H^#hm8Rw;K_tteJVnL~btFdZGcGN54h;B65M
z2Y)D;Vb$VSd1<%Gea<s0(gfB`X|EIEG&;Z@zI9)Ql;Z>3kq0wE0EdkAHKoik9?<ju
z5K)G3{kP&bk@s93{@;lpL#R3T@)_&A2?Iw0{{}p`i2VA6JLf+_mHNaAn1Wp)_WE8o
z|A3nG6BEBF_G8-mV}QXR_cM$|5hKBH9Yd>E+Y(Cfr+tr50$b7n?}8D>0igNwC;DL#
zM^7Vhmm>$N4QJsjzuSpc@s98HpM;D^O<*6G7GMCUE%<tiOl7W|z9g}{87qBtlqOBN
zYcsJqe?!H$FjMIf!G$o+#WAYq5DSO^lW+sw9oXXdCet}JDaU^HH&pX*mEGC;8|C8F
zDPN!O{1q|nUvqv<DTmMQcl(F^s=vw?e|N_C4bPwzrhX-LbIqwGKFjyP3sU{dC=I()
z&}wZqrSrV|kD{J_1x|B~k?jRwBBZSITTk4p-%EZQn(k?~`a0banRdu@q(f_W6zSZ2
zRokLGPBs4Oup(akdGGw%O^@>@%*cy(dM#3Z-ey*1V!nn*)+|@q(R)M(+)f*DP~MP7
zdCOh1J>O~0J{nK<vEmafzTg+xoH4s~eWdJlu8DfkOd)h;Z|0yV2+qcU8`1%9WU89i
ztIf-CR0~R|KvJz&=C2~BMz6}36Wg^8A#X#D2hit_+IOc2GM!00u&0#pe9CB>T^k*6
zbNuY~*++mBElFJ))!P^oR{Ljf=h)D~X51WdXPY-^oZ;l_@MvHSFt&xiPiv5eQZ>9=
zesOvy{!oF<*vep7{`HZOR4C!HYW{roQis1O??@l95^C(F?a`>87$#E|WNrM+bII^O
zlbx}4hIfC%Z)k6o_$VPLyIIFww0lhRK!GNAI^g6+_kFV6Z$|zO_?@a*<^y`X4MPBX
zPMcRpPE38S8#r}kXu>3!4DY%Qk86@GqKvj?>L<GsOR_Om%|kFl)gl{$?|dDnMu3CU
z`@51huwD@BQv-HQwQeh(75&R3NPLjgS^WC5iV^~kPwxo+S@B1WxRYl$9tj2fVlxML
zrrJ@tND}S(r!)Q<x7B;J|DYPV^`#C2GlZ9AoqNnB`@*j{&GnIbVLl*j?kX{*VKJly
zU(T3Og?{!^q5X6~?bLQ;!%Fa>=Ub{~IXI(1^)pfGB4Z*~&g_-zPwZV#J%crXsB1pz
zduecunFaH~2aoi8<w5y8yU7B*+u<TW)-SH8LFg36Jv8(_$?;am%|YC}fd7@G{T+hj
zCcrL}DnyZ;!*f$&eq6VUQy@po6u)v@z}%^C85%WB%j9(B@%!!+(Ez%1%R-^u^i{f=
z+{KO5zmBtE?*61vLU;7Yur4snjkM`R+T9%)<a3fZ`noouyY|RQMY5pEPv*k49<lK7
zsY1%K;?|_<VVY={jlUl*0zEL5WAq0ttjkBV1Uj^g=(-b*ZTgI-gDtXDY&i43V>uyu
zX+IN~>_7khad-Ki2W6Rr#p>8)5qYB?SrprsPwtjIJQpaR%z6W~c+Sk-fSqhISNcvD
z*a_l+d6FdzG2-|i{LqU|GBKHmD3zT_#FMF`+lz#=rwh)BIosl{aQS+_@%`9BOdKyK
zbIoSgS7>)cSCWpK`b%J{G40|G_!H@kY!qa}LUiZx0$Py+y0mZ=VhjhhXKh#QEUr#?
zS3tWCwJBg$N}%>q@7@G%+>?B_^;vXI1qe$)8`f72ea;`Ra2=b4*6cmlaEj)Mjcxr2
za}&xXfW*9{uMfLT=xSDYx5$pjd)z~hRLs@C_;_2CQ%=F0{nellLH{yEyA%hOKI=tQ
z3k=WeCs^_Ro|^h|_(oX!#76JoqT60Tw^#b<;p?1Q9UUynOix1K)#w<iQywL-Bazsr
zli;w4^j3W`P;x3e^V`seJS+8s8aiOd^F!4o?vFacdfnIBQx@J}L%nsvCa1`_nTD3+
zVf{Voqnz29oXq9QCcF$yq`Y~zm6A;0V|`<78F~TZ8#2|cs$2Z-;KTihzM4;LKf3~*
zC0jH}JEvzX&NUaA3YEWH$>7(VcXh3P*y;CR<KU6L6VYhvbK)RdlL&$L$Bbtinv{pG
zfn5W8Gn}*T*TE{lyX^*e_vV~6C^J~9liK*E>T$$cPF6Ng<SFW3Bz#-n=TrFBK=Ac+
zB&l?&jr;pbh<{N|#0i|cr{{Zj(Fb|?#)~|?Mo&PQaS9S;c=$161xf}{__tcfuOIJ*
zI7WH{>z%VFzebL^2$AT3BtPBD6&8@fKE>E_s!q`~a>vPUKI8V=om)m}*jN4bEza|e
z-N)2ZTQXu{c%H#{{oya4yvyNBMU{=Wi0VfzQw#8txsOdPA}S~()ELv2BevXf%{fHv
zl;u|^qqzue+b!RU_16fm78zmw-aPYru++=FDWav9@8HjC1>q-;?!Vf*e_f-tp^9=u
zJAI*sset#i)t*d^W`38z8WCRQH*K&Ta<s8|%l1wh%J$$01%J!HZ7N%8iuHcpX|UV!
zZ9MMnuGQ|eVs~1dc*crJfC1TKD?S+srv#ftj>;Elq&_mS*O5w*d#D=H@abXvlN=FU
z9^P0@6;`S{VHXqhQ+nmd;(CFM4>C&dN{_PB{@JKA=??)c;$$AD*^D?pWtyxd+bg~?
zq?I5AWL@>a5|Cm6Rv3jX;Tx3}KBzt|c=>e7aQX@`h9*gAH>WJP>t_*W?pl&+p0yN~
zfTTQ@{z8wun^x`^EMjC9tfIpnCtH+7#Tos|m-F^i$giFl++47XR6%I9tj$q1wOcYH
z;h3Fcu6B6Memf8RhqM?usyi`Ah1s0U?cyFgfpl{6mWtsOQBPBsJ#e`HmY1yoaT<1y
z1bV&c`)zkn<j-}@8BE%|(~9_~)8=ZK_QsXYFMJoZxssZ!D?f=Jykb~2y!sdZNVAiQ
z;$&VUXT&W)T&-W6T^Qj7zGbz-Op!DPzoM2f!z%Y6=!Am$iLLbe@)vT9kF#~1dE3@K
z|7s$QseJ1rfvresRZV-p@L22au*E6%nJA`gX0$9P)(F<5r?;$<I@%wOpGT8a8dtv5
zA3&za{-Bq>bih>O-r=jelqiC;NT)bF8p+H5XHQyg%q}85DHtvonqe_`x$T!!cklF9
zJ>EhySTLcI2&b3~#s=jQmHlWxkDh?QT{Dwgp^z?oR`G3cXAPQ_5!wQtZnBhGYEOWR
zd(+ONM}nEkon5dh&CgM&m**^^-hH%aBWVnWS(|a0sSkhpK1ZO)%>97{`xm%OLm6wh
zVOCdhC!fnC$m^*$UzcC!h=^xxgi-%r9KXQtpXT0-O^mX(_-NdGNX0CK&adZJH!Lmg
zUt6!pgk@1joOcgo{-U&ZyNFZ53ssAWh||k1Vxx}MCvjkrO2-?I&%Apv|4{ohuh<~l
zM+D4}go^TA^PFh1DDXZF+V)mBrruHM2GG{3d8ntMgVqn&J2!b(C-D4DymP*aZK+Tk
z#-wKKo8pLys!Co%T*XaR)kjBvC3ihN1D#T5u63)?KHD_fopTZv>Phsy;zFXOmjA)}
zVqUNBh&5fFl}P07F20pwA$Rk3v7emQ{4J3nS*Ocyy4wAwH`d{+o(GdYwdmlAT4y?-
ztdMdJ|2gRk)tqe9+Hz{_%y$gcDc36$!zRm9QCMOGQI5XRclv$^%jV^uB2u@suDp8C
ziCDit!dab$qRpDfAxZhuP{%_(eVQdB7WNXTS>8bhTu}F9x)~qJDx3S>ZT?(S!oB0q
zR72h<s#}rMA3XMo1u1>B&GGSm#`g!73iFTlgQWfKtNW&Cq;By_?9)pIwP?;JSG6uB
zr^Xx}3;lO3mq%eYvH8<e8fQg!*sI|0`xX!4o|aKX+siIo^=%BOh_u~dY}8<w2L(km
z@OWkbzL$ZM5^aP{EgS%iH_@ajU3u8U76Iw=!*IeyucqB0hSei`<ewygiq(cpy$yWx
z=<5TR4QVrmsvTAV&F_em7}6+<xDn0_KUXa?c|XNA(R$&XNr`q}bi(JiVgbK<$YJ^-
zl-7be(o<6P`;2Zdad<8VITi8AK$Eo*d-vCoDfXt(+LtgFmm&R#CL1p}57m9viqDIt
z4B?!HehgX<?^KsM)wMC~_xf4XewJ@IXQmf%TJFZT^0VI^rkKx^4CrrJE;AtLA_uiC
zb`ejqGgdtz8O?~<oPim{8lIDO&Ix8m8s9*9a$7qZm4TMO_ptt~dmB_+pTeYG2XO5O
z{~^}3_IUT3^Oe1qq~f@i^UKmaq=CwO%IAu2yIcu<&5HNqXMdaf?5!@CC2zEAV|i@~
z!aQGPUjs5xmB@OSwbeO1C*Hwyb_U*I7bzQV?^b7?GUPHnt;^G*db#$#B7koYfRQEg
zM$@iR3QLx2{h_Bo(R&m6tP|8r4e$3SY?SS^FL=F+wZ%Ee{FJ_%_DHDi+Rdv!cfy|_
zRt}u-RC8ba{IdL#$4)>-&jtAw#_;j#AbiLo-8m<q;Xks^lYO9B7UMez9&GbO?r{d3
zWh2W_Nd|vXKA`cwec~^JavflT3s;LJM{)d)afb6trO)c@V-NXVOckfSGT@s`x+m&c
zb`Y80A~_iV80{FCop3RjNGL%(-<9FESBul+VfA&-3^J&6>MnLF(=g1lk*|YJ8g^%F
zUv_L!rd%e=bckfD#7;u7Qi@%B=}PCa3Rk3MUP{QQjwpr6Hm4Jn;_|V>HSv2MuG=Ys
zR~0JyGBN81J&|z-zU8A~II<N(-<@YzsI~fK<NC4`J7cm*4x!w%1FIJkyDNlu&z}Kx
zO3n0c|9V=@?!Se)$ZNFxptwCJDN>Yz{Q7}<xpY~(c8GyX(&oSxm~8jge3VlW<lE%G
z+1kFTT6mwJWA$;1Tr!J4O@R<I-ta?Wo<f&yk!?!~SNOoF(DVe>6sJ}eV?SFf+-BUY
zcivXsJCXP9`PW?AWM1g&V1j+aEogmNhA+KzAz-$8h4IFlw@ZRqCIZfO%iUs|A)wAy
zeVBTC@6+h&g5tKh2HHvw<7QT<bZ$MQWt9<Ljt7nEza5TLhRSthMpj1@K>OMc3)2(o
zRx-343j|DEYxr^DO3A_}Ds?KK7mrj>K;7A5WHbY9K#(E}r!qWn&tJZqz^cu>`Xi0}
zZvju-Dx2T(z?O-K=$0###4oFfkXcS$A(f=|e!I5x!zY+p-{a!Ir^y^P0^Aj45V5G*
znl7zhcz=_XT^S~-{ut~sIox{uu0?#iVQQPWewu{AaA`yy4%|jwyPI68ki;(W^f&uU
z)=4m8H!RXy1=xffeK2~A&2Yg}*m4f8V2%)$>lvi|y)p)3qXsPueaQ%v>kACQ<nqD%
zUVDc%G=&#$Hr_6`+_`VVcdmb!_wujUP*}6l2$DB8TW|zk*)l_n5Fb1OL-stenjHB@
zA&bT!IhCVdHI+=LYwBlcvN>$SkrI@?AK7O|bnS8cb!7Er?0)>?>r#m`c{3bHu1ziM
zmVm}^hrLOZgQM}o>RL9#yZNNXSxhYT%A8uv_4L9*r&e&W6D6Anb<Nbays*plM^$D0
z?x)({)DGq6Uw^k>RZ5f4zsJ}*6zPjF4u|9hpmbBP8@P-Rct%$9pdVTLElhhBT}219
z84-EZ43F*#p=hJc@%RBbG0x^_qWbk`@g;L=@zOk$&^<NEEr$Asx~+e+#T<FsM2`+K
zZd$xF9VVq`3`KXV#YkVPqm*mM!*sOc5kcjq?LB$sZUl!Q<MeC%-Ua7eOK<kg|5Uj1
zmnrn})?2Oh9x{77PzDOJBu&J8S$3TJ-AYi><F~xq4#}PNF|lvz1c$Gd-%{XrEUgl{
zWLQ`EG5$^Ir-K|tbw@me%0*3Y*sR5bJeWo?l;KaB9XxbqM<h;Pj$+kX<!!x6J3kzB
zP19SqTrXJVhD0Eg3i}@8x+wDz`rw$b`d}c5JldWC=hrozEx&6$vZLo_HI_G8a|G_c
zrCB#_G1(f>&#i)t1J$rzY;4&GxblM7!bx$M1IAnjM)5GV`(1@WF~CzbgPbnj0U!tE
zXA%AA4<7Ney;F`@&^4o)vmVFYhqV&mFJ#kJ_-VW)S$$=&eQ!0qjMCWdn4Kek<%Zb$
z6FQ(8Heo9-#CM>5)fRjE)bKX<PC)0qh30WABP=z7o(Y{qSR{)?$hhxM0{f9$=#+gG
zc;{ZE#CUn`nA%Q+e(SdiM8&x=O$Aiu)!D5+vr2UW@y>jYY%m}7lN24Gs!df(iq3XI
zQ<Z@ypI8hRRl|L!T>S?8{ch6~t^yTtbA@61>nf!74)kd?sU4{+>wjQ;L|++w7$Zbg
z=-lm7OQx>rAH-W_Z6u;Ei-~oT@eH4YvE3^w`8b2nc=x!F)WFN_9qD8!y^1^v@x>@>
zo!?eDxO)2&wOC9odh>7slxIQ1S@o>{%u#KmiFd-+`wxu|5l=9*i%_+;?6Z|anrMby
zs^1uaTFZ3FEoo&P@0vrx>OQ~Hw_*m&kLo&+-G0!R##GjHt>)BtDFKIz=d!t+o$kGy
zaBzFD5$qr-Z2c<bs;K6+RX08wC<t@PgEGbjs}{2hk#iZDi^f@<N@P>AG%<9Se4=R%
z(RS9zyZPAmIY5rjwUy-?P-ec76#5$O6)V#|fm23dc3y`e&m$$yXas1CMeBg;3f}!y
z{7lT1&-C0LBRtn{)R1_%dK~s_eL44&i0+r#MN}VTd#Z2au=oynKUyQvvHcdB0X1?j
z^<i=x=p`PWU##a2DM{q=d-^wS#t!!yj}yc4+T2I;#iKSS{o&zzp^zE#b7!0rmn@E&
z4#Gh1b^H!4McPtckhF<x`4!NSfk~<}{0n6S+o@ZA;>f7B2E~UphRBBoxJ>l~p0;_j
zvX^JoVs`(jML4<>$i708oro0aFLpw%IjY5W?otsKFT7<c#u+V2Cf6OLnq0~UYWY8Z
zy0lSN9!x_M4tK7+uFr>Wptg}1Vo;>W-@+e~J|e>Ul4Oq<1a8V$s$&o9Qpocw*<@BQ
z<ldwE_NNVXCEwV60^^aWmqAJ551=7>H#!SPb#G(}BZGrW47nYlzWU(yr?(dmM0&g!
z&Y0S%DsJkhS(pIj8d2s%56NKqkH7b;b{g-PMzd#6y3F*<Uwf)wLoRP?Yuvr;KCRBF
zDDP2$B;|A*W^{+C6_VPWp)kr~dq-bMk6xMKylHOp$K%P={%wB3ow>e0y}h?3d;<!t
zO%A~nOSLX-IO48nZ@%{YI>R37rit!fpeaCaVA_Qv<+C~+<@KeSuqwzfO;OvCh{>s{
zM^h44ueL6+)t2vC=rn8TnQ+ye8=a56o#FqpT)@O+meAVzAn?<i&{ldb$&Pp~#y=OF
zX+~yE+Ovw)wU!-)De1q1BqEQ=nK-D)7ldzO!uEy7PZh3Sn@(JbWTt>Z3nB-uC%q1K
zzPNp_e#|#m&Od(z`|YfE<hcBZ25%sk#lOR;UvA&LR`0h>tC1+~ltm=IIA;eXl=3w<
zLUo*)iezBmjy`WjM_zzl>B8#K$_aQ8uj_jm#pxm$SP#ai2zRIw<%1rTM9t5?y;k-N
z__j%#sKUSGFw$gH==03Db`yA~x;fqnT3fL=bGGt++q>IVfW;O@z&lEttk#KRp~Mu4
z?tYecR(%)rk$NY{t$aGmBI&xBG_mJ#)a>ul<cKe3t=X2dxvkltSNghSyC`Zz=>se@
zbJbxi$3&@!LS*a>;$M09+gTz`rPur2H$G0mCPSR5O5-WMdmi(7S>}7knxk8eWnw!R
z`08Q(&Z$@Ae8PVGx5b8GSSaDM%Y*=bk6v^o<hpD+=;E1Q6$WaGKkJYBx87tcc`!hB
z+0rBHyG{u+p?_7QMx5Qo-0PYWh5J9=PxDZy6ubDx;K}3b*~AtbatH%79G4>)DT~Z`
zp{lD_!Nb+Y0DJ2lFXT+A*^JsVbcwEd)5W##4d<542-@l=LYAR91iOwT5oc&>r`DlK
zCBdSWTpkm<9@Ho?qwZ`%oKQKZV{-3G*hF4%s<N=n-OG$XNXV_ID#+n}%CXad$m36`
z3<T>%TdHs1pQFJz+^JEjjxS;cT5VK2<mZsiW8-^E5c$5e@sfju?<3)BoUOBR(p|Y@
zG-Zm6G3j$oe6WldG_t*2yU-Tn*f4(PZEwINmfz#)o7<yrq@PA8wZYU#cRNOV2$oj&
zMm5Xoort_tdF-3W^UWiR4bcjt=W5Mol=I%sNc=jTqbhfSB}OHs)v17bnVj?-X848j
z5hkD(|8QhCN9aYagNs99$DI`sTrhG>Kd3}lx8`E+V_UALkv5?xpTS+?rWR)qBN%W{
zDl!!`QAMq{qp>t0cFr}E+1qeJsLC8ivTAaHA-umJW1nn*l4L%r`lU@WoYOP*Udil?
zj`L96l~@IH-Ie#}+$gjkg}p5i-+094FnRCP!gI;YgM@#=|HN%~$#7}*!6$PJNYP1P
zt&(qE{(q`eK6g?BEU#8I)P5w2a3{;v+l$?wbhCXT@Lk}Qh4VhDcU^m5X&FDVx_-W*
zInrX+7uK?AfoYM0TD47F|Ga@DCW@F<k1SY(A0-^$5S6P*9(Ulk^*EvB9&|>Fle-{T
z$i^z}86T|={R%`{Ymsaktq2CLn@q?#4;d|_Y8aeIxxt*WwRtB;5r<DBMOY9v-)OET
ziBVGMfL3X_Q8ft6&a-@(@6Yv#%*@V|qEsQor<V@m!_$}2wd3tBASd6qh=gb%7)r#X
znY45p*)*hrMxqHk<j4;s!#yQi5s^V=hk`yXWocOao5#?Yv_;4@z)J>ft(FlLQDfLS
zR@+aod~b?ba7l0#IARrEv;$=9X&`H6mj&ljijBw1gEG8!#nys3JS@)Lu~7c??dD^E
zInNs=Yl=28>6N}*7@ibIX}m!OrX<Lte`DWkhYMDp_t={J&S?V3UNLoZVJ15f&J{AU
z9h@p_Aq_6cU*K(;gsl4(t92?*HqTX#6=hn8$P-wSYo12C)*R#pm>36vICjg>f{~;6
z)4LLglU9TaL{e7*MDO!L=QjqPdpNS!<47aLM6^Wogo)0HRMeT13lAo&(1eJ{2oxyr
z+Qz{HbUS5h23A4a$dazb3+al;yoq`3)OmQ4@_kD~pyY>IQeWe-$vOk;oBjFPH&7f<
zV3Vpygm6>r+7DHVP(=7?g|WSND~=g^5qZW><kYKwRo*&6hl}1f`}8R#<hdnO`)#Bc
z)sn0jM;Of`(aOOEQ&8fd)9GcUFH?U69J@=puhsibj*0*2;QAS9Fv8wt$kbdmXBV|d
zcJmH#%xOPXBbvVY=&s%Hpo9tia^81K@%gk_vM+PL?ZJ5K<BL|mloUd5tB0ddJwP3w
z^{wSmQ?CK*8MSRNjrU;pF;HAnCzOQDuAW_O7O)*}GNVb>qw4B@S*v5M0#tn$ezX2k
zmSwtqn%Z&6_68D6wCX?zKq1}w!ldFTm}bH1VK&M$0yY8`A8UAW$@Iy@ih<X<AJ-L$
z<gOvM(Lv4mOH>VV_iGI_H|RIi(|Ka2T(Mix;XA2pWdd{k?Z+Z1^>8<_?8QGOc=WrU
zasro(GXU?jDv~;rxLv4s$T^Hap8zyr)})v?BVt)SzV6p_t&9$Eb+&?NAUqy^C_ojR
z;D~;B@~3?$vt|TZhROsyRw9qgmj6N^bMXCYx8owkeV+B>lY;30HtJOj)pG8$fcc1D
z(9eQQ9VuS>j5`)lC$-L}I5`E#hw7Yot(!KV`l8qJ=mw7AVkAPu+X=Dkv4runLlx9(
z{-uI%Ck~J-RZ!f5s~gg}J(|xw)thVk<btMJw05=z2_FL!BY>h&covv4cFU)sX(HL{
zT|>mjn?t&H7$y)`^|V5r!VvX)t;xyEb~0k^Xs*tw3ny7SrkNx`3GEEclt!tRhlP2%
zhWUm=zB;L3U-_g@895pWly&94GN5j{|IE-RM`DlFDry@Rn-2?p*A#pu4yZFtwZqH9
zjqAZ{E<%)Cq8nL!s+B=aZC0+2-QknN!l;3mZ#^Yb7LxqNl09<fV87wID;=i2y@tR`
zazZ03AfYCKd}y>D^BwZr7j6|jvQZk1pKUlc91Rj=4FxVenv#+b9hkT~6kjrnenk@{
zohs8WD{B2ZvodP5=k|3qvvaXL=-tNjBlC*Iw7ECmxvkA?4CMJ)!`PGbIaVXjLK2j-
zrBF=}J}9c4f%B95VL6^wychb(De!5@{4_P`zlNRFO>zDSqm6U`DusG>bL(N+6jgt7
zTd#DGh<X_;!#+)^8toEi#YB}SaPIyLzAD{ywt$W2s&4`#Bc|)|O|8S~B=A*If?dbq
z73dw~rO!d2jwi71PeFq-)qc<c|4cHkAnYRfvHn%EOq{O<^EZ1(3bDH*9jhFANzNyo
z{D<s#Eh)TIxw8MLlfUZ3&?%Gqr8PNRR9E>xXg7_|^At|UN|+%J+?Wu&mzRi@aFoSC
zj^~k8rv7-6cHeaV(oDa?rc)6<NrjPvDQ3Xo4jTFQVLNEX7d2*)wEgs!tl69-oCvv?
zdJZ$Rg1zY?N}5>{P9O@7pH1mqc1OO1U%GWEp4n>Qc(Lcf(kC9Na3U_tP4NZD6bw4O
z%)p=CP8K>&BsN_pVi(=cRz>qaoD#%xEL3M-f-$>L%*i61BGQUv%c7`Kiq?jeqysb^
ze`^$>QLiac=C$*6UBWw~zeXTV{m5P6idg&cq*7>`3<t~^q8+uHP{w=OGKi)J=Ucf<
ziKhepO4&LOtSUzGH#nFIHCzF{*R6Yl4i9^h&ShPV+(9w;+j61>Oo1ahzNOYSSZvmm
zQwmqZpB?KM-yhjj7}J<?Dp~W>=vj<=9IjN9Sok8g$MkYzuItdc$Ah<=lH;XL9Ov`m
z4Fs)?0smHn+F`y7go0Z+d`jk6Mk-YEzpVrOLC7WQJD`Mr4y9<<>iC=w$qE3|egkc6
zb?lC!3Gzn0^UgCzIvaUpJqf-lJ~`~W=U{muL%k*-az_=fc4p%$qP#_cq6(CzJaZOa
z8C=Kuu78`zco$ZrM>c!m5X#2#m|vaaY*jAFB_9Hj<eOZN*Wf5yENOf3Tz|jN;`WID
z3}fM>1Jnx+@<sR1bC4}~?$%<)krA|*#s`c+iC0Y)mCbQoK)Pp`kh=0qy4@lCh944}
zo062@4qnptu)g=y@T$tFH(Z$NJf{TbSQdB5f)JAURxht;3rtfI%p3TOc|P^<1=WPU
zwDH(>3$_hD|1GfYZG_h5p8trN?&Ul$X5+e=??XeTMLn*igWzDV+z_J9heC(<LbHsG
zd6!l*Ee0piCQYP`w7kSq;X0NVuof3;j%~meU&Xc;)u%fz_Ge6Y4p8eAL%=4Z5KA>m
z&{559ZDXp+4h@a?anRTb=i6wJ`0ebn^2{Wo6z70SWEirs58bmOYQMa3d(q=286bS{
zf@fbKCnFn03@IeLbQ2sJn=mkKw{We(JPi@7=I0Sn>eL?_rw}9m48!^>XWoUyJ~k46
ztQjsGbEH9mw|~dByAhynFR?m*+@hV*)Ace^lrwJiZWwWj3BLO=Co0MkU+8&sf#TC`
zI2IJU!-c9rJ3$x1cFR8c@tuhv$=_{SG_%n@NOp5S8DPLN^tm2ZPSK*-+0qctAh?E2
z#r<ae1vq*bAxBDju6u>d)!K4_41a6OB_nVSQEI9fSCA~%Ff3Qwkbd3~>#ZG;W^+Ct
z#(hD)+HX1#+?7*fg2<!fcV734Da5Zbx5GV?)<RlD#*ku~-qIXPva0K;Jt&38_a)i`
zw@(Yk@0MHbxeWy;x;UGx-lz+!kiMomYv~AGM|>8#@3J!E!S~=hM}AQ0(t_CXT}u8Q
zWSHvyO~1JGPikMFdftO>NU0@+btSh~<=5p~mD5Z^ocFs<UH3Ql!nhxUXzWA_oE2j)
zz6h%CuyUTUgQ0UyMY~g4AseT{2UGd7dU!io;oDwn)=}&x|7akm?oe`#Z*aCXiV4MG
zviO7IG>4Cs>zI(1X_%<_d`sduu%mj~dB%HRP1$zB@wIjABewauvvypmv6zS0sZ$KF
z8l;8gw+{$QoHN>HD~|w|xG(}K-?#l#s5!C}cm`yWKzMp;z_90+xZ>JD5j3=;3cM$-
zJhTE)nQoo3PB$B_aMZ55AaKiCyT+{J;){1ywoj5ra7m2K6AZxpxnq6e)a?tw4b;qI
zs|zGZ=b(^J%#Tc5x%3;F1i2$oI>*%=auMe}S)0Bnr%`0<u=VcUlv|S9rK2C>zbIg`
zM5q3>AzhUMsO?m{z6c{t?pBWbhJ8}yxL)UNqXYJG%LHC9eRmKrAaYSF8$8KC(|4Pa
zx|)I$b#A09A=N*~rP`!BfsJJKT~Vm8zibz(0vFCMDe76?Z#d4N0t1Ww)G`~^A!~=%
zS)&lYk>2Y$Yxy#NwweN!Ect%?3P`^@ea}2lQJM}|P(#}`!>_d@wd7=@qjM4mKf&b4
zcRTUCk=#^`BY7=x&zQUI!gc~ahi?<ZzdH74o6Dwv?fT<OSt0uNGrS*%o?P1-Oc3d|
z8PjDY{BTRPBD2+$g#^|=e#sT)R~sJV<;GrV^J8ApwjZ;WLrCaXvK-}&smL9DqD&l3
z8NfKyzZ5IaSL}->@IOUy)IU*f?GlS2ORiIn2&zb;g|GOM6*5}I$#~3l?1`iIOVczS
z$=J>S1L)(ltLzdlTDz3Odezg`_YhkQu{woU%}Yn2Jol%lhm63J%oXTh+dHTqP>|B@
zNIFI&)a0ylmEVNjBn=dcVrrl^kN)zJcj2FMeN$lb-Bp75l*I+?Sjz=;J3<s@klA}d
zRaJ9P#QQKvgwGpwJ@&PfIL=LNay=$Xtw#V*^46Z;fUS9An=t_p5WfxiVMVh|Ri5$j
z*VhY%?WcRqJi>f!thu9N`}~aRwk9<_?T>SI`a|^GAsSkq;}t}7L#EPFJgCN4+*oqh
zEQKv<T+TpTy}(V+Tv0T<(xMmt25#*4x3FK}E$7Ys9XLtTnco5~mox{8Rc#Uo4Ivxm
z`tX5)lj(=iD8kHbCOBWr06%v5P2QsCQO|sg%*y;zip1?x8RX3_+S$=3JIAV0Qo&2)
z=cGRTN`vO{Yc<YM?DpW1;6Qc9T-5IYBj2A7C586z6psiI8eiwINfNYnHT*bFl`%d}
zjIXTuUi3VAn6)nLdgE^B$Ir!nA>;?eadg1rt7@`C$~9olb!9oSC??q31EZR_(9R|1
z`#LUgu_n*O_25;>0q4N8G2ebf25ky{+{j=mU^%$C@Zc8y$nUs?rqqE0Vtj*~Z|&P3
zv1BWdfblh}VroG@GjX^*+}ad7mt>00IoGB$xuFtsHk?<xf8zugAc;r}jg>LKGfAuX
z$&T*AKZnw3MTHeH(&fU45@hGxU~stdU-EByl-~?0iP@YTloP1E)HLtA59h^NxaCi^
zUVoAn>K?#rtB9J2B28{GK3~@v>Ag&Qyi03Duaa7#69ui!T>m>T7i)OZ?yF@dM+o&*
zIeeAjZOYoCWLDWnYHRQZ?2e2te<;vyJO?shjI)(8+%$P87T=1B6#G7gL~u{V32${<
zCN4m(rfvQ1s#i(M`7-x(XKtPjh@r_~nq)|D(_!j;!HI_d!rpraG}UG6!%<XJKosc+
zs30g+r70~c(nPw{08vp|43XZ6bfgLh2nbOD0Vxq7AT`njM5Ie7fzW$G4Uoj|yfb%-
z&fF>9xpRN_e*Td_4m*c)_St*wwby!{XI0{dEzC_voRqTD1aAxEP04BS2J)#3lk>O0
zjkWwGB-=1I?c;hXXF{`K;_3Bybr(K{K<JC>6bHt|TY@mi03=Jb7BREwW=H;Pva?Kj
z?;9gG2DQ6$7j1KId$>&qx*5iZ-KR#rQ^tlUC2tnW85=unuDN4s7bxq>FUWvuN-TFy
zJ(9)Blg_aBfv+UUMz&ZBv%;Sh*n3oG?DCl#lTmvi5gXq?gTk94rG2hkiY6p}Z7GRN
zN7CGm;#gjcoPpvtz1}4M6C^X!04H&YV{?#PI6flOqt(4(sobDj8hs}a$0RahAWa7X
zu@oU2yeRj8rx@7t4YWTAQ1w&wdOEKc{k&*ti~rTcd9OatzEEDn*RQv<^B^U@l~{0G
z1MLfRpPMA0Ja)8bou_7i+cgEHQ!B8Su-8+#D|z|t24C>;&$M=@PM2LTU-p>Rvf)ez
z77cIOtl1~Xn%)GdD2On#Bt9V-naU_TD%C6IxLCkjWSzFCS5Kt8zsq~;XqFpS-G_Cy
zd4A#zWH3jDH2E<B9b=xJ^<<VfR0b%BTt2w=?!&N7#iH!uKBp8x-dp|eP0yLQgw@9?
zJicILcaPLyZSMBEyQeUHyUa4V1C{esU)9HUBH)LywCA7s{8xd24RVxNZYVVf4PJ*A
zy1IS-?xX9^ij4f#@0myyQZ5iZ9{XzMkRwfq8DVe8LApbmO6I%Tp|E^-j_6cdo7s6C
zwS&`@4rdo0t^zt0kN*56A{2XvuTVBO$#@~J<1j?iyZCGhLCr>Eko5jdNyTBEJ?;W|
z0D$s4;T<-NxNN6c&+-Eq22;rbM{BzrOhuF?Yi<K$LBYY>>if9ciO5G3DdLJ2T&lj7
zx7%0hb!n(gN!Q}t(%=dZ`#J7_gn-XdO^#h^AO7Hn|06&Dp0Dms0x%fRZ22_)9n6kl
zF6dRn*`Imst!de*TQ4k>qh0OCZnPqw6hMBoEmzoKGg#ch<^>{Z$$EUq&_^+YpOlnT
zNR1;vWImNmKHqkhP)|AL=stPW`Bv0(>LIK91Sn3b&Weq6q~sOg(${Nx_+?&{c!kpq
z#=zAWjP2rY^ZCHc1nAr2IHQLTa5YJu&g}4MdAY4bV%W5#zN|ess);C|fGbF}I39SU
z1u-zK1i)x|z0jXN*JAlGb9keON4JC=jkfShP$gfU=(x7)-oUXbIGa1{H+|4O#v~!i
z2J7M4q0?gyP`j9v=Xr)0-Y5C^q<Ju(D1OUWRB%b;>b_$GG7c5@gR(mc9{6eViFDZa
z=tV<{qO<V$ff<+8*fA94xB@}FY0047klZ6}b%^L|AXYdImmj~6t)YAJ?qhJl4bNpr
zpzO;A=3XR&CsdQD8zwzMceQ7%yA&?JzpV6KnRamlchr75N<weHAF5HCUXR3{{Ss$k
zVd>xvU-x!f)HD6O0GCdxpV}95+#xSKZ&c@l*{w1H>(YE`cE;n^BlGgn!|dFCS=vfX
zBfXI)^qOCkC{!I~yRt7{2gKIJ+5FsZ<k%Ydxtai%3GS}ol<+V{Rec|6MCHVnkM9~z
zhiMAjZfjG!y3cLjtwf&_Y5$C%)L#fs{fR@Jom+&R=tpevZ0pf8;#U>zk()=V0a>py
zOxM7r??3;SL?8btyY$c4us?5WA=ixG7Lp&J76W%c8gTjk^A~!=?<Q9s@Ymm_+IObp
z&&|r;`(1PxS-^h-*Mpqo1!Acd$=ipU09>D`aw6tWxO~<<Hns3CrOe^-GfgYeM6zL1
zEgx|?O$cMlx{yb9hO5nZbu9*XPC~xw`x~6+mYcjc$%|P}^s=cS={9+zBHc#l#{t%X
zG@b<<hPxyK8*zao+z4<Kd@Ka<JZwHJT_g}M>)aMz>j3)X5p1kox={%uL8lNYSrE;y
zB<KlGbt|%uA0hS#nZbxe*Et3aO@)zNbb|DrT_3oneT`pP%d=MN@l7XmTUTX?2ueKS
z#R=udBmtuY&E5eGQql8Kkbti8X)JN3QQV}-g!_S{7q^8DE~Kf{rr|sn-K#b(RxTrJ
z+$wWXCV6mNfmgI4neS{L+?PHN*3w?m<!j`@6teb-6OeGOqLG4*g5Av^&WU1Nnq9rh
zqGVwWl2Uo-3dQ<`nP<d2@?Bk)g~Jnd53wPA3MRO0)9;<y(3Q`;x@v45j<23#239=a
z_{u&}FK@^>%%!r@F5kp>Y(iJnHU<bY@_I5iy6+$w5l68Nh~xN$9i^E@1ygY}k10yu
z@+y5uwiAy<tANZWs|OBOZJnN4@ISrV(X<kx=(MXq{)`(SNJap7H@>Y8=1*yY#Ffpd
zsn$?3Bs9(ue)WFOJ<Tn(x9Xb@$3BJ#8r4&U$0czt<2jMlas{wf!V<q$w`<|umFw>-
zdpAVwtk;&>mrqgZtIP!KO$3i~jy#gG<c)3C1<hwjD6<xmT0&P84fY@wFUk0L)O^t1
zUZqQJ?gZeA-o{d~O2P~)Y-orx00mIO`$!tTytV0eO17{yyHYmm8%R7E?ypI7gbPiK
zs|VlvvT-bbPFSm)ed`q!UEOyq0Hu_U({a9Nnjt@6ZKV#_+jt_nfj<OEe7caj71CpZ
zQhnc7I&zx-Xj6%?=gW(n1=DWE>6*Bg{a^w;;!;tjHo0Z2y#t{^+-Qj?6_w<0FfJZZ
z8E*w(v!iXK?VtzygQH{vojG15ShAg2645OB21*AI3)B`d`^xiCP(&MkI;K0~Y(e|q
zn==3D-y;@iN<bFS37e?(M{nc>vqGN?jF-Qe`{uw_(&l+g2;q>#ebe6Zlm|}n>X+zO
z3nj@q6u=)=Ig$t{6BQ##MtKYcFafFkMqFZmErrrpm?UX|#>g!yxVh;3(^@v?GnZNj
z?&pBD%LCxqa=M}^OpySDDU)^_3s7tu0E+DcD8<MEm~cT_-#{-CQB(k@OUjom-yCRb
z#t}px)!fwK;ZKy)aWZkb8?B)^D19IKH&5lLZ<fl5B&L&Lv`}8rZy>t60A&~89tE!e
zblvlFP%;ld*JXn|)jPmX5%%W>T-7iDQ}y91mJATt6acm^qr3_2K+`u+lkpVm9)Qnw
z2oErJ?*NQlSxK5m?2or!p$Y=>y?^J5uKc>7Up@3|4E>sA|99=sdwdk1*q04(m^TJU
z%I1s=V-w0aCE`YkFCKq%L>GjruaS=cJwVP6%;~&tp%wZ+od*F)m7jW{f8<??vQJQN
z?-=+0nU>Z5RFVyp69{^?L?fiJbn{lqU5dJXeCP+R3hkW2#qZhQ4z)8bcsPO1j$V^j
zd@wWPcF*z+6la%d>xRRN^G_#iGq<7_b}F|~0-jsr&?eXlheIlL(f?#+T9XxTS@#AH
zf7llHNlaP1?E4Xff&)qFnfoxK7QNW@ta|K8If+^jxkQ(Uh!rc84)Dmwdj{2)%X%1$
zUX>m&-aIx3la7`;xb;AGOMT|Lm<rp~YPt<Lo4#)}rXL?%kvt4wv6Zgc^;U`cL(}8d
zghTz0lix8xR7-1#!;fLvzJ`5@HwWS^Ac|6^!%ex0hRyr3SJ6&~+3uS##qZItOw0D<
zXvSL;@c}|arKM_fa8%Yn7Aei@D<<l#Uztx0>BaNm?e2nFL(yv>N#W0GoIoDT7U6t;
zj>rG-IWQ%0zxtBJ^U4wAn2z>0(3~!D4cmw(X43>bO~#UH#)HXTauG5JR^Q7d#4qO^
zWDiza4zq7vC^cb{ZhzKpalkmNgC}U2p+U|FMnpI7m^6wjD+x4ke@?S@DzLAN$=S@W
z5zr_0btaq;74K0$zw1{s7Nhh={6m!13EheScd$v9`5M?^egh7!C`-;B9Zh#*(;byC
znIUJTQIyRKO96N0#omasKlgf72@tN#u)~NcgI*ITX+dr`nvzQwMWUw!*ziWZ4O(V=
za}donpe*Dd4~<=vmj1Glvgla4q_%14PG8FH_Pl>JI1~T};*X!4oOu3jk@1yr&|vjd
zKtatOxlqc#BLi!8>wGNFmqrrSCN4ESk1m?$_M;UdPR|-@fnwT>q7zrHevN62x_30?
z%2)eSEIYltxsu+aTeGq0Eo(kVH|WR;42Eqlhor;N;WCJ&s#%0<?y2q<^zD^EzQKMH
zKF{g}wK1yO7jLZAtPmB~NA?SX#mNWK;XZop<FDbn$lSZJfmVPN3Y!mX)QGejb|GX8
zcU#5oR+ae#--@H-5?D1$rl*@83A{4)MEwQXM(Ri*j0?n85~pEd9UA$Cc~UtMZ#??h
z*rO{&AAxJ0$Du!c=<z>iG3;xed_CjP@vGHq1+KfV1G=nMzaDsoILlV<n5f3N=0G`8
zs9Z%DG4I?ACfbJHsTzFcz{lN>VMv3DGp8tW$U9uy4Ih>*{L}z?<gt3>z-1V(P}?q{
zqc%EcXimL>#+|t0xXg32)U<HLHsL%3a*WSFf3-Ba%4u7ky~v}*3fwCV;B@lG(P5dy
z3;-xU%HEf$ZFNSd!#c_5&>4>eM4R=lz&X8xBTfzslBw>VMsEyrdvbWpAj=DD3)6ll
z&fR*InkR~fVj)ONF&{)Mc9Py6Kx6nxeFJHY<&O$hXMFxFRGuzAl7+jMz;osBCl}O~
zL+tm>a>#>x@QY_Nr?R|`G6p=alenzB@!8H<&U`CEC?&cG?O43LPYuA|All<-1v}tN
zWAD6uaWvMxjTA*q-q$aYyi?Wv&YuGGuNtdA_BzFZdLCuH&|M-qoq8WsN4MWb#POwO
zamC{gLQX?;#0ff(_P7}xdHcH4!P~q0>tfEEOz|uQ>}Cag6eexGU{?Lf5Y^9wwf@oX
z?D>za{l6oMX7`f|$6s4;ME=0bHWL3+Xy(!7iuJC1GhIh$`$%c%$FZO%UJD=DJ`MYx
z3!R@G)^c)KWgk{+IjH08oZ;A$Fc9PH(*GWdi)>69Jwnz*qDQmd4bJIo^$3EvxGxn>
zyB|ITDr21xWZYs~Evq5<7!W^)<t}h1Hz3&{_Bg-lWDhB?9m8!Wwe0k=!a3o?hd%7v
zF;SJ@IM<RSseAn!NGhp}J=&wgK+bl!+gBk&EhQ#D=)(T8>JN-Z{q=;lz1>@ohUPoq
zQ0UF~YuQy=VWyO<q18uOfi6jcsHZL&eXCa|QrDP60m_w0h_bMqvT)xz$>$Dyyu}S`
zc}-f~$Vlwo9?Jj&wzrz}Q;J5qT*mqix>}aAWo(MPXEkqJ=%<d(%<h98p~xI~g$wrQ
z>&nE!g4_qR><>QutaCHFIDK-hst{J2ye~P<vQ$r|SEDPd#c{*<<h!7+EcvJx=+F2N
zFmEY7DuchQpK|%df`R=iR4J<4qOjEF>zjP86JFbAJwURK4`!xXWaf-I+S{{ZvbM`?
z2b6`s<n)(*6(5Ud^;{`f){MW&RBX{CdPJ)A)=8g<9P_GNyZ%ZHHqH+Y^E#m7WpX%P
zLR}@O$=5}dPPqht;3CI6cR+UEKx77HvQdp(eWh`%g7%%)_M1`{7cW9oPdqR=IdSsH
zMfxe|$*wE>w%fwHt=~X?5sC>^Acs@i<a<XI<cFF>tq4;$&#vljCD3-+NFl6yHj(kj
zYSh-(%vSKI?(U*VGf<ppkfupnMdI(q3If4Z@Jcb1Q~%c)fA!=4^P|TSRfpd`i&Lip
zIM^=9EBspOzw^w0@1|${Y~c7GN~(%={k&}cUu3jy{lS@gWB#Y87RYyS*VZn%R*0HG
z(aoHDOU=>m6?cS-zZ}h*);g0lA6XGjCw8Uvr1wCxztR^O=rK=3laoE3x%Q+)alZXP
z|Ch)W)~jdY+npHWqWfnI?ds`-L2B2=;B<RyH_1!iK&RnoDiY9HdNqZnf3Xiy>7TgW
zH;xYkv$(=M;jVs?rls#l8%@n|=q8u@r>TLgRr_?D4>Z|5Zcx5@;EFP_WkC+$KZop?
zM3*u|E-XZT%#N9R$>2E4VR!xVl<s9~<CjE4{ZK3?IQSK*sWd}GRRYKD(EplzmP@cM
znoO7RG?Rby9Aaw3jdEaU5;jAFKHiQcmX9OCeWl#S6l93~#qj&D$-8=KQf4<4^L5X#
zUp9*%hH)P8ytDwp9;xQ?8OKIgalh}ToGWlBk<If&xG^8K^N{z?%xP>4^}_5+Ufg=)
zd?RmEqxn2+gO0{ZABLS-GX!*^fQ6n^2_C=V=|e-xfOmKKzkzsOx_S)I_L)qyFW?$6
znOf*=z?3evOk+P_57xa}NU{fr4138*D+F_xhl{0WDvy2Uqx=&EXpahk_`T2ck4k<7
zSE1j)@Bl^eyVJhGCD(xu*;T^(U*<Y&C9$&?OK)+Q$enLty^t&RdGX`kTu{8`hrMrT
z^Oo4QCSL;tb{oH;C={`!wGv2U?p{xeHdr&SEbS2=hozuxD??7joU!Z>xUBv1TA<eP
zN0;BELhDyR3*OC{S#zU8Fsh-pX~t%AKW$WV$=GP*SR$L&NjIM4hdf~d($yi|Gm-jM
zb+!OgVZ7|6*O_c|lt4vQ47D^TkXL7L1hJQksaUUqCD_=Se-wA9c|QUO3O*ukQD_dx
ztxYgZPW_n|-yT)(fiMaWvAijDiT3y<nJ{G5?@B{RFM*9Q8H6Ev`cKh&?#;wpDN6pa
zyXA$4<-ZaSi;gZU(~)huM>S`9-45P~)~b%zWqAXPi1_Cqg?>q3ai++TgREs%LaMDY
z&Xk1eia??T49hB``S+BSsAbUw9$~(FmaZii?OH;X9HU-a-suMj5*3xjad)ahd2%h6
zM&c}uRS!MA8&j~S?o`06DgYA+*)zVfP3Z;1y%v!a$482LBesAwRgVuq8KF9KX*6I4
zG|&F;Z?>R}2-sourlyBi_4?)&=A=Ap<s17*d_>>c!y-3%mag$E8u{a0#{Sq1bRT}T
z`u%|T)f0c18ozqt|NWj2Eg>ra!i78<fUFgM=gdOS56fF2n2EyQRV`i!aY#y*>$bRB
z<<N1ORpDvg$Aiq*eE!|p?VpW){k8A?^@vzgtpf668jxX!lD>gvD`0gCqsl)VnWY<{
z9rVkjv27M^-t@~~GT4SxB33hpP%UhAzJN;*%s}Hx9xtby0RxnSAy)uy=$G8Ft4k5A
z_1><kMveMqF8SN0pf<NCY55p3zid8#@N!dhdU_)i?rct2BHSLe@V$L~*!si_vGA)u
z@7yNs0)o8>Sbgh{kqe(@!tzR?Yy$hZ07lt9r~dHSmFXIz^z|iwDz|QSns@UkHWs;1
zIqw;bYS9f(O^%XuTw3?LFTD%p;b*GmkdtStjS){;gk;>SfJ``dp=7Rb+T1%Fm33S>
zdIHUF1*<fo;$tKQdoXBTSnV<HTjqZ8>h|!1y-d{~r&*>wJKivsu-+ff6c+P)jR@P6
z<GUEy)J>+Zzep|;T%8!O$g5SYvwY<Ot7c@mkP%wcGG1|NSGn^bK-}&05z0I=NSYea
z#C}RgcD+xm9=td)t!#ad(wvjD-pda|&WW!3JM3<pYd4b_K!~3~RA7L@Yfxbd11Tq3
zk%L%Pw^*&2JHKNa;?FinPQqV{@9yL%H}I<#GthW@NHP0p<KCAh%H~sP68r$1p%KOH
zEB^&kT27*5$fWL$Pv;Y={L~<#GV=F3Dpe<|PjkkZkE%$)&4Qz!`R_9IEsb&!wE-?x
zsxGr9PkrU_PgQ5&7n&;KJgT4FO==MEjh6#AE_x>zxk5wDbwk&wHZ2-tm$4CvcJpBw
zK$c0v8asRtp=8{9$eAl(<awYUOtU>h$8v?d=5NssUEZGE&Rk+$dIi&^&{I&f{Zr^5
zUj#`v+}DXXFp*M4Y8++lsg@j{c#wKicGKS~OF|m+tiyTjdN({?(_HXqwM8S|lX&@=
zuRe1fG&<ulIh7<esO^FnA5lH(A}SkV?PQHIoN4@mu9>Ov0~2(CON>GnO1c4nOe>}V
zN%4~~h)=Yp0YT%=Z56es&8#;^HAF6*OvTCtH#K@(Xdmljype2jbmTKy;(E!oqbj07
z@HNG|XE%$dd;z_S2<a-A)*N{OA&Z_&!!<i~Ux>oILn5mX70uMaHQ~BlKj>Izv;;-L
z*8T19JpSl|v4JepKH8wzD9^5C`~3jX$|;Yjhn~$&akl^m^0u!@FO6v~0T|7e9t1e7
z8PDYAXC<nC2M*s<_$pAgiu>BYU2EnDIW_ohjP<Ls`A#84dez|om2zslycU*SE2ph|
zUygw}zQl1?*0+vhxxwww>dXm)IT}xbChe2#Z<YO0x|eX>4U22fXsrzJ5mwW(?W5cF
zi57Ile*1-b6ST$SSIL*%mwdo+r9v;ywqt7HvMXy}rlgz67O9AOn;!PoUQvE}iRkGE
z0*kTJ2PvrJwp9<uHLyo|#dg`i9wMIL5xJ5!t{c8WF~t!8Xf;8!laWX3UYG`hz2eN`
zVqC#sz_{tXwXB3QpZ7jLyn$w>DD+T-(GJ|BVZY5{Zdpt&qm_fa|5phKug(=gExpc8
z#e@#cEYBl74K%@CfMHEOUhnOQ%_I4}8GIWtzw(^6X@p^<Jo3NmeXF;68)ZU9rtfas
zxu_`?>n|351JO^_lr8(F0QHRkfr3V8$vpp2&&^#UT^?VFeoX$e9-ZO<xKLTdU|>TT
z<5Hrq>yyjYt#+T(K^{BcMoG>&amhfFrY-$?+PQ9u4E<w?f<;<a&XCqCRrecj>fff^
zOWU%ulGU@}Klg!cxpOSdOL`Y}fG8JYUf<tlDa&B|@NPu6-YbY3wB*j+hv8iGZP$vf
ze#mj)V+=FyakEt3s^ep??%9B2akAws^buPQB9hNThgj|$V|Aq8!|J-XblL<i&Oavv
zP_dHR5a-SbR^p)B7d63&MZRsdiX>q`X3O{L+pZ0`Nwdcex={(NQqlh)`KIO7TeEL3
zGfZ#G-i4FI!18aVV%wm0N)|{KnjkEx@<5nSp5M4=6`D^jY?L-@&rz9DOlwb_N^&jp
zT6*=~jDs#<wcx==m5b%YIEqwvfv=q3_z)79F6Bh(>Kn_s3nJo$^>5ysjXwp$Ji4XI
zMZf3D?0)ku$}z<FlB8CTE1J&J@MceGUi4>_tqtMqgDHUg9K!zf;Md0|9QBUyhPm93
zHCOWIuT%A{1wa=8ek5VPDsWAn84bW??Afi>0%(%Qs^Jq1D~CKqCR$v`-ULgP#{L1T
zv*qvJ?ZUr-Ht&xGa`v4GS?U@&(JFL$^4gI5UMI;Y<qMu&6tf&pk89O^idrz{;M)Vp
z@zX?u+2X6YZL69*It}_2U&VW`M~9n*uX#O`wf!oC58XjTNUL{YP+_&qed}RWKASx&
zft`*65}!IRL>Jyj;Yx4yy_Y1cnCcNx%R4tJ|E_0|qDkDXqn#@4`6`oMn_o7qa2Es=
zrg6A-0(7ZmZ>DiFvQrH`l|N{o<B<t>(afHkTCu{rC(pUqj>m?(NwQNe6Kl&0Am%s+
zTxbfy&4Ver(Sv<Xw7GzAPQXL`W&$Sh*|8X1^SOMM%Nyj(TiU`}*Y!h;dhdbud_`vV
z)b<hhclOVr$QJ=C5x1a;IVYPKMR<NULDsGSf$<EZ1$f)i7y+2lq_^^SP3I5)y%w`{
zHWAIcCgI3gj@QVsHeH%}=91MZ0DNPW@(mPp4!q6xNEXJuKvNf&q*Y4*Bl13m-+{zT
z-MijN;nD<jU^Wt5TmU>$y;;uj9D4$^*mA#qu4XAZ1re_sS4WTFoDq(Txgx(K%Jai%
zW52&{(BP}$?{9PVyQ_M^l=|xw|M7%=^~Ik?>#x4}+kNr6<bD6W3e~E&#i980ZWZs<
zOA}G&Laxs6Hi7JZ554$(fcNKe=)f-G@YatUnPNw9KXW$R9=m!31Bj|<Q9?jw|H>rg
zAEDB>Fl+-9?+)4oQUpXo#Nbzj0PyAj1-T<(PW0Z2{ow#;3mM1(y+ihUGrVZ;n^tk}
zYYpU>+C#9&q(3uR(i{*0+N$GcQID>^YD*Y#<WX{-wZQ8tr)_$bwVw&^q4bsCI#%)c
z-l6Hd!&A?K(h<RYieCT+H7!lsV0W>nX6Q!V<i~A;aH&KC`JinRu<ba1Xn(S?drLFz
z<a=Ut;?BH#Eps~C`bS3=sQks8MZDeD_&Z6*FFd}Hdng2StlMOL0E0$C=3am9_m-6_
zfG?~q%^~`Cvs`0$ug){%YQAAUT<_1<LfA9@dH{=I9#Zbxd}(u^Jwpdy5}7cyvbIHp
zv^Tj}?IG6?v|wJ<N?CIYK%A9*-d~s`1yL__=r)HyOb6$6xg;+|co;!t4|cK3&eONO
z(Req?OLnB4fvu2lV(MPh7RYQ8bCSR%IxhrSsF^e+3T69)*3|YWK&(u=e{rG#dL%P+
z)KZ{%XDiGtAFgZAyR=cG%nEy^O__$}x_7_?#gAXG5N&Ssvb#DgXAl!PI*QLi-lAD%
zgeRujBV`71usfG1REdp3tlg4|c`Y$ICVG5d#;%dJ#r#9OgTMNw^<4RdtFA%!E_!uT
zst-oLEPL%9dPb`DpMOMYG>&32bevZgATe-HZJ}P^=GnWr;?;MxtU6c+q;zhY*>h`@
z-(%x)pWvk9?yLth6-pE=#YR)EI*s#)ZuBns^2>|PJsB3Pb-k}USld-B*dQik&622h
z{4gjNC%fW?hFek{3v(c5vtP3d9#qr(YGH*n#(BH~v@hO)^DGWQ^>0yO#pFkx_L!9=
zmCa6!U#?k-YEDA=B4KjdheD;Wuaa)9gK?skebxH5Fv-Bef^{Ci37Yq3d)J$0j+ag@
zL)nXTfPyANuR2V)ueP=Da6Oz+s#YM5P;Lz1i7Bw4tA3`irEy&`q)qx2jWfII6rb(w
zwMWVAYt;Z*Qnsz;ehx*dnAWlih_<5^vB#}Ei+H=_I^0|~Q5Sf&6dRQXgg^~U`AL^9
z>kYafq0AE`nI2eCM0(AF*U&eRP&sW7nO;><0Cv!p_hZ)WD=M`Y7e7Vp*m|sZMY)@I
zJ-~q=T-IDS{n5tki*pc_w-e)%3SBB3r)u3K@qD0h^=~N2)o<0F4Qy{vJ{xJs#QQGK
z$0tl9d2u!2wDpu5`0~ER8E6BImkb_XSP3^dHfmwnGLd7Qj~gjiVNPlr9}nNhXYpW7
z_Y{b_J$ZR@j}9+=ouMbJk^2amq|!t?)guGvgC9E8Yj$nWye)*M&9qSe-V+H>l9N~v
zzt1$h-8g;MUwLcF6RjYZy(rU`g$B$Kv&Y(2jYNFyMD_h$Jy-^PZ?)3v`3zPfEj@QW
zNgsdEO_!a4b9rCF#WR?vtGEIaj;m~z1l#a)aUc(RH@E|md*~|3Hewwv+_2{0e<q=^
zER`uFd%EQG(1@;33fqK+mO!LN{L!gkjqS4#dwV7uUsa%XF-brhF0w%2e=OZ`V7gU$
zNnQQzby-|U<;;|mOhNRqQz>R+cFzJY`WuO)HU|(}Zl#%51vTyzCFad@W6H`Axu(t6
zE8)ZJaX7Wj@v(a?30wy63L-|W1J6H{ag-avj&$%-%V=mHJ6oRB1*{}Z8?lCehnpH&
z!HGgJp)EIJ>)h8o2Pry=q8?f4w(zXun4r-X`Qc}BntP(uT&6@Hb*!?ygnX$zPPvjy
z&<zIUCtnxE<__Vl?34R14Ew!2^}g(V{Gn@F7`c~rt;Ql*ffx4>s@pZY{JG}}Gl#Va
zg%HEO2X50O!JX_$+s<A|?R?zwbHT#_*Ckj5H97grZ!;eCM$?gVK3MJPK`fS1oe6em
zH-!;nwTeKsJK`M;ou{LVW1mU~pPdJB39jk$I1jGm;O+2J{)h&6L}Eb&vhe<nGfiRH
zJDaT;Sx+E_Q2}b`syTvoI+ko<brf(`u{14hV4r7jY_QA1mLef}5cvfP#PD7NXz#N?
zpuTCGd4~b;j!W?rRgk-@3D-9eLu5VwI<OCEwa1bpz~fGyDzw0FplxwEU;<v@P5(Rp
zNRQrD)d@v>Uk!N5SCN#Yv~QsBOd1DAFnCrHkb>wzl34frBbVY94MZdal9GWsSRBZm
zi@0Ls_g7+6NIPZcJrF6D{k9@w<IcMZv;8&Cm)p!W4pUxA#Xen&^1u5?p-q&B#?sEe
zJ)Q?pQ{OCn14VZMK=u2v6e5tQAsPYyq4{@SJ`jg{*Fy7jK$1lO=-Y%Vvfj)ZsDUg4
zrm$HeNtM5O;Mj37AfuPTjiMbc!jeb$iAknFT~a1pfWB!01rk2L?Iy@tV#t|LKqDkD
z9}T3kC})zuYYjjyu+C@zNdEoKANzGVzk27_*!f#COWT$p3R!p>cnO_2bV<s7>zFW|
zV#!Im9G5HH7yf+=)c=0$98emzobtE~yYUUgPm4^d=Bt_b;lQI~F@3_gYQqoE6Eav9
z3P-P^ieJuGbmZ;(fjm497ZB1t!7#-GcZW>1k0R1?{XNmJI~5w6Xl^31jtLMaLj?(C
z?0g!uue3Pokn|<_j+2LDQQ2~e{%WkUfo5HsYRd_&nk;NlA<ftd&elye*+A%E1rQBj
zZt00JFoQ#_U~+3827XatyRbB$Co=Kzjp$Xj=zDgAeG@)U=nGv_XBtt=v*5dC=rfh^
zKq-U|sqp6m1$QQnAKQD`beq5Wfn?3Y3@)-Q0l#N%M`e!xYh^SLsoG!m*pGMPQ?cst
zYxnM@3Yk8C|Ecp?+v|te^V4X1FcTm@PlJn7q=eGmY<si44s)TRrIU8|S2!oM6bp`^
zWY##^j&K~gU}R__0aZXw;;Cr-f;}cZIe88UY%Z_Y#s(`(nmX|9Cw@j3YL@8tNq^kn
zg0Bx<UfySSi`_Z5sFhDg?s5x#hA0O?y;*dHv9%!m;*5GsJzl39al037log;lvhZQ`
zvYDPVj60O`{<7BHidouT*p5~D_@MYykX|p`mBEjF^&sFj&YcepuR5f^6gp>e@%-wD
z`dv<!^s1FW@4zOuGL$T8e-<D*W6kO_y+=E7O;<<`aXb4UL$Y<oA(}{e+sqWQer3IU
z)OcXq-w8__p(66eSHh#+vhWTUfeN(9aQG=ezhrHlS61DW&n$7T2FvPQt>wcxQ-^gX
z-tQIFZNRdgOw)S*zF#-lg<70z-`|h(E9hw#fedSM-r=TqzV<@w!|mfkCymP+5sV|A
zRSg0<CGng&yKae}=UN%ccyeDmeNaq{3TQEqcSB$5t39~DB9rlQU)L}m=irU0$Eg&K
z&e@A`BrDr>G`(G8>lL~0{w_oPc0GdXhbLZZPyP(J4k!v8v<1w4hB#d^E?OwT;Ac`_
z-oKTcTG$=72DrL*dTl`hRLG5+5e*M-Ifgv+K-Zpu%+k*HEy*xMno9GC^L6TJVHiKJ
zD6TcK-cH|MojClaLhLKoN4B^z;Ouf%yuH4n3Wf2IxlQlq9%ZEAVGGBkWxM;^R?aIY
z+mxr>_a!{Zyu|6wzn`Kz?x)fO{VKnB%)V#PY;ZQNRUoVXZIhE=w|Qv*&jX+PcHOG`
z<21&e?kh~JL0aL3YW+3Xp8KzKz_x(R8*y=IV1grfK&#j7(8i&~Q#~2`s$%temf^Nx
zpC86asffW&8e2aaawmF?5fs4DR>x|qBl;B<J}&c3iH#~mi!2{UNF^T2+9mK^b#5>D
zG=vp!gay7O3lrp$8#y?b?Co#~vmGan!v=6A!h@3LUM@;)2ezL^zX5?JZ~IO3!tji4
z+J-ciZy*sL8QRwY=?)#H<%tKz_<L(~l<Et+3S*2hCW0;znwHQVGO~kOzw8Dl(O<Q^
zHy28xb!`UN-QShQ+OB;hNP-c`)2=*nw!)m9Y+WXgRf&>%<)m9#;pkFLw~|7?S2K<B
zshq8-&ef*C^8qRMcHeQ(y*%4^Vi$gcZ0G}U`=Eu5G@&Fs5juDKKMP8rZ<6-_Zdu|C
z)e}sMimCowBQ-`(i7`QK0MBgaqC8NN><0ye|CCqxFKEZ}Z)HpWwY>1p9{VrklZR*r
z$;cahiozaBv@}~Uk<rS>5+^k|&&Eha#q}g?_p61GU>_<caWVrGQR=mpdfZgE*_^US
zWq0BJkj3~5rf>5FEUv#hU$6M3xxnn9|NTx)BIkRe@(XL#c-^vA3|CpG@-5GOV`KgY
zs>7~)*<d=dR=+~r(jv74_+G;$zL}>TtkueoJpzHGgbnEnCCQdWvx8+aVVXy!9@IZi
zd?nwj?_*EKwz?s0-JfM=@jw>HNQk|Ew4>iSpl}d03_D8)#A;6<DcS;_QVnfc>GU^0
z0Gp23OF*E!7D=`q_Y=@2SLC%5_(9~*8G^NEtU(2zcuc)(xx7mXQ_+%&bMq#Fj`3QL
zPh#BF)zFcJX)Kek)SDJS(YHB2$MXc)cxLz7lCLUnP7si5{it6cQLDA&Ot4H%^{NLy
z7dY@Riqp61W!Z!qrJMrh>ns_5T?iO6K!HU8x^`osrt?iLUwN?d8BhLeB;%Cvb`5Fj
z-Oma`p394O%#3(dK;s^uhdwMy3qT?Cs8U^K@+RWo9QY6h%sCH~0Fcjam@V76rLV&H
zuH(e*w5r@gcfV-X2L;4<q0^3kpk3^yoFb09b@0eDTJ0<887uBnq(c(S9$?PcGtP@I
z6`l6cOko6RpI)Ml%AJCy^(*ocZ|pXrSPG!cCRXR;wIo4yNxB8*^8-5qBbHfPq4(1Z
zipr{Au-X`krD;v_rwMzhURHRJ2KA8JmaI8q9D8xJ)n5WkNwCV>R2ottjz;F&vnMB8
z9VzQ75{gnx_1h@S*LBn-UJl8Mila5W-IDR49mO&uztG<Vi~;aG$f?`d&8Nht^2w3X
z)QiXiR|{||uhfsA3ShOYeTw&@&wH`G)OeX<+<jZ|wVd;Nd&Q>xu}?Lor16JR*0Et*
zP<9F(0n5|q&u{OZ`r!OrvPrqIeA}{|PegjZha>k8hk-Q2_w}bYw^GA*w|rm-K-zQ5
zea*@wMVP+=xrCkZ*;Xo8+7OC`jivh-4@kE&N#4xZz5ZTHi;OMl^G&g0O))tGde9;w
zdn<0uI`<?1!>H#NU)ccx$sl?FS#caHL{1woECX0ZEj^uMyw@fgl1IgQWS~cZu#~}P
zH$UOsvud|W?dpkVwpPxR8w!;1M)2Zg77!uUO~=T#z{W0EV#%MYgT%8`31|;7iD<kk
z>28izbi7yUc3^a5xw`&z94oW*-b_zcmC>nf<Wl)JP&bOUcR<}%TyhRsuQOUHS>s`{
zlkj7wg2x$%TU3y1CxtHdw^GU>dT-dSfEYmv2HnR17wNk>!cPn#`_LZA+tT&`0K&%p
z4%R*Y_nX`sCaYhAvfF=&=FmMRlh{*ww_hncIBAUKLhk{@r=0<)=>T-~vORKR+~XT4
zQ#=<0JSxcs{Q90AF!2bOCKOA)Mp60(y22|3RHB~vtrE3O1gQB}qhDR{YZUzNm<X>+
zhydx)BD?8zf;H@6%8;1s$CwJvODFF=)S$q0{=Ee0e}SL+M<4i4tcyR_5@G+5CV}?+
z@yH8i1X<wqu$8CIvIYj$zK^4b(S6~N;qLowg$Zj^%ZXdEU_NDldYxFCexC#^blxG2
z`TJy_z;IgaVo~mIME1Ad)d1~@*M|85#xhQJjfTa41F5#$az$>-uK?=DUsJM>bB>uh
z@~y(<+d@F(78_`=glNj%8D8gK<XPF_+yL)x0V`~?L;AXaR%)AQC#R*kmY>{Uq4o2@
zObe}IRC^SszYQ*CNKJ<=_6n@-Hh_qXTvim#$6JtRAha)izKA?eAI%<DIKOVv)ZMa*
z>O=eUZ|1Yy=&DOfo&A`0=Y!r-^%t@aDZfPyz=34WC*y(*a1>PplLZofHHY^a47_M2
zsaTb8&kFyEC2{-3(CVweJ;RK`KDDi2ED$A*H~9t%t$J?3qnjf~)9L`=7gZj+OY1h-
zb~uSNw_FE4iR!m+F^_ZJS61u_I;n66ep>E0c=*L$O<~aCR3+|ZyKW7Nc5W-rqqerB
ztfSsXeKQhQjZ;Xm1gj8Yjwb005IcU|zrlQrBHiz+5h<Hy%j)g6ktN&7_#`--oxW3{
z-(cSCs$S4!5z3UDII#<NX{ucJzmTdV+Au0aTux?xROy!!=fbpT{_s6Z+ue&Za*nh5
zPx5UQ1`bz=Akq?c$Q8&7ZA_5gHrCkAorU$-O{Og>qc@7W3#4WyWI7y1&l9JE2$7OZ
z^T^{%r7Opvx6S7`JSu%nJh<kj>bt@o?9y4CBH$leNyC{9N$J?ebFTCl(l8%~?xDiO
zNTBAML$P_HO}xP63{x%biu*YBtC&k85M^LH7c50%ooR|C0|D*3ofJX&7Rs6T_|0gP
z1M(eW73)*Ah`qO3<~!tD^PEr;TV1h({1CZYGgVYM2Y1vmu3YekqJWgzM<C4oWefY=
zV6mgs)dIpXu|NnM>yD+^0RDX(Q1HWDoCLM(QM8l{TcYtcIQA=_i0?im(AQwr@{oHe
zJ36NkEpR16t64?OrQ+RwB)hK~X(~3-*QKzIP!%c8-{3F$Mf#bd+$Mtx;}@%z9PxVX
zb5U&?a*iMMd-HA23>+>x*LFj+5X&A<8OYJo#4hYC`^bbvNb?#tnfP<PBPTM=JQk;u
zX^usB+}?T;P=03o*ovN<gflqYJ7O057J;3${*A>kg8l~LLdVqddk!_B4vdkGy&aHY
zX1fAy!8O=dhG1Uos`f=w?h1|UJ63C29rfV^%j8**NURyrB)nQblVV5Yi}WBVBY5Wg
zPZK+6B~G?7>A?>10)ye%0WU&MT)EyIa>U~KiK=4_5!3qRM#y@=-Gn3i7Uv~2WoCe}
z_NWwl%g6S+As~&sjYocI0Wul*Dkx0~B7G~~`ujub-?f)1Q3d^PY+xoQFFr=Hd5#*|
zd2I8p^0W$X+0<Sw-+<lcw<q1Ww$(0t_TFg*vNUn+THiou1}v2eeue(KJO7KG(mNpc
zs{t@#e)YreXU?yF_{n~dk;zDKgE_~y7D6pw&7A-AWS`@O&tISJ5`B!*SwUaFf%?$h
z>DsYWbLzR?zhIy0|6@$nJyt)e)EJBZAQ}3CLHf^%46yzee`ot6R4VCDVYw@OKppbB
ztqj>Or;3Y{4g#c88}WyqIPBCp$M~G<tA_yQVK{PT4TzLr-vJrotw;nRZ<g*6899zP
zO6E{YZCKp7zmoE>XB!2NSG;fJ^c&^fW1}$ZU2kUJ>%QX+IPQ6uYf0T!g$su=1a0m`
zqZYSC3YM=WJhBX*hh@KdbV(cX!9C+7kvbak#*?|;JL$2<!vPO&_|PHv>y+moyQ|?2
zM$aoGOrw$nH*(hjIU!^y99~$845J<Gh8H@jP9FWb6mt5x>f~cR@!Z1%$&uv~OWQ6n
zqS#llr-sK-b+W7WoBd2<h{KqW!l#uF*bBy**s@+;FPq2O)HjYFy~zAD#3Crbuyx#8
z+QSJORDF>&s>{^ZRyufG`$(sdI={_!iehfr??L}j?zar9@H@-&pZU&G(<qt~l8<2o
zB@43bbgLVWpUP)TUV#C#lZis19hfswHeWTleP8oVnb<$sK<o7ZAmW{^2r*x+JPMc?
znn@64VS0`BD1%h_x~@*c3>ISV0+$@WWe_p<Hlwh-cU%U&*PNWtXl{ee{|)#4x+?n{
zs3kPdyB?OW<FcD)-uup_vrQH9o;aM0a$KVbj91V`oR~)NDVGtYy6l}IW>WAqc`tpq
zb$qS4HIJIY;j=@x*DiTg3%GD=38jr=nP+h+T4(`-lEmwq-4<`XCa?8pIeP7_x_5N(
z#ao9p{$OOu=NjTli*S=kf9;`on&{=lv9Nla$)?~%u2JC(!j#edwIcqH;fVdfe9%pc
zeVO}8d12ZLh1w44GXsc10~=Kr{uY?GF!e0m-@yY9zIyb(^b~no%;ER?;RvAG<a*sC
zq=P=0`chZ{C;QnqW)!-xWGMqgw>HrLYUPV0FtO^MqA0QIxM~$3lv4wQat;_ESCt-g
zUE#!khq7Mx%Wnj)GXSxl`@%qK;&&mkUnlw1C%?wYzhv$hnhnD)TD%(hYTAJKJa?+b
z>x5G?YyFG07((gniv$0caMJZ<G?9!KzrFpm?(!c4J^#v}(SLHT|7*$5!$agZ-hscp
z{gm3_UlXnR?>zQ<Zot1sqwzl(0Y5d`{%Q!4Q50}r=NW$UVgFGfv;WC<|F=Zc&KF`M
za&pCix1VO3?EP6B!cTnfm(Bma#K%v~!k?Vczij^h=-Tkh=KoyHmw&}t@yq7_gsuKp
ztqZ?w{x6&V&pRi7il+X#jPx%*pXv;q{Ht2*Kj{_xRS)EspZ~}5WdBgfvHyO;$1gws
zcXcBFM>z1mVmp7M4A{TYvG`^4|F}c>$LeqVviaZdDSr9+|H|q3PoC?Spa09x|K;cZ
zx7S?yW%Ga8{C|F${gbx*AC3JmvynWSW0@)7>MppVE`jRZVno`%`4W+3?=2g;6IH$!
zXJdRThCvfGr5(Jy(7Nc9LtG(vgei+S;QABBS9oC^1+%X&<j$FAJds`)x^sG24Jdqk
zY4UJGCcJI~oq@HB^t-#T`GlHs`io?NYlkJKlWm*BJTYMu((KjhX<(N^{zy=*JGmCc
zPU`{0IVk$X<Z<0*eAGa#QpOrVDEpP!T-m@VZ>b+QuYAA4vzU*!bj7Nk#cP#jtzP*+
zD4EgEUa~E(A;YQm_|}Eij!X3lZTP5EWZ-V+TkJS57Qi2)i~y9DF^Y0GmaGn7PiAfV
zkSl$FN(U!8#9wT|N{DPSu@b_^UrmzDLlB#H8{mFU`JOo~xS}Vip>NwmH8<>%?4N|!
ze|+M<d|JzT7n9;b(rrmXG0+Z3M)?Z>4B~oBU$v$mqKG`BT3LgMD=_509Ru4wS(ain
zTq!*tA$X7zEL3?yX<1jn%hv*%kzZME$!J)8tI`~hRgsiMiSuuj_Br_ZAV?++$+(us
zuKE1j<YYj+Qx16`CEt9mlEG3Frr2J2{<^Vt;)m8a=Sa#;(J$a1ZVoF3s0Z_^zQFk+
z`uur3d(7T<7by#SI``u(!b~fY2=gORk*8(a!6_mq*xIEI=v@{#P}8GDehl{^tr6g;
zCfx{sS)yE|n4%@ZejLo~>AyK9-kcV@ZWF0{Jyjt)4Rs+^xcR^hV^=E8sBefAEm!W-
z>Dazv;9K1ZAk+_O`}K!XaVt5!v{U5Baq!_;f5Cnk-H5Is`%xYz_TuyJ-@4p6)qFb9
zn5F&H$2|v!H9>T#tz;xArw*X%RbR&)2qDp1Sb8i*<_e|c+kFEKCQCu&Gzk&&pJkK4
zn_rIgWWdHGc|5Je<p^@k3gJ+_UiHvgkvV07eqUw8vP6-iv}x#Ek$>Cg4iSMH!L6m^
z52Z@?Ky5q|HbaS$jR+Q~3K4w+Lyjf!XcKE_FS9+<G>gk?Fx`{q%M;;ex^7#0p&YhF
zTWKdTM-yk?pe-fu)AmbtV0vuIWu8JyH(}av51KNx-3rU%Elypa9RpOLcDOd6yBqcZ
zD*{l`33mSmqJ{dI(5U7}QUqm}nyP%3+}@ZWMOG`OsFbBOaX4GVvd<=uiuY~6+8gVn
z#C9O2+wu^gxP@z2^rfMSx&sw}tfapLDWn!fRB4jqKdNY(hbNa1WM7U`4e<IT-pS2)
z)8+V6EcM+FuD$dS4S0Xyqaq`$ZM&C!1>iqK`7gA>YS1sAX1CMJmhEmP41DhL>&LAi
zzI+2QjhK<!>Q;ba&&Od&`kyveTwb@|apG<ZI}#GlLU(Q~=EX)K^vB83&L0q6ev524
z#;eerA=t%ZK#oNZkz+NxDEeM7=Q~l3AwsvFDc;$X>sT$Jv?$a%tSAdnH2O<<Vq?B|
zh7shV6vat$SNiUjl6}Uj#BG-LPo*xgNAu>|K-T1xVISY|Wk>X%Pfe6Dti+-k4lK>A
z<Vy+<S)MRY3tB-?QZ<2^QXc-WT9p4#>zRRYwDffn$7AEwe%1`ld)@ltW{?+JNu@_W
zj=qU8_1OE)4Z&%*!{4aH@L`}<zV9|krvk1tin_CT3DOyHECuq3y@`Xa<h*bhYyuJf
z4aDJpah7(D{1qooD;O2+sXbw5LYgIU05X~t<`;aVWe3l`6G?%bL^AuDoXaAo)ybDH
zT_rZpeBfu@=w!jGk1^f#p*pM?R+0AYxV4*AiRClCzIo%ISU)KyH}m$#&*>{YseshQ
zN@;l;ab#=v)m8)yEsma9TtO=6A|$cv$R9?JXbQ>XaVd7+f)!7<4C=&uQe{r~3R7sJ
zP10K}@lSOOZx1JZGBUa<wiKct6&D8;^lt$lFDg=&@kBOsPFNZeuBV2DPF7U*yXta`
zEN~ket(HJdM^&cn2gIZHY3>cIb3ZbgG9eVdMmwxNrs2j(Y3Jbw;fLpL=(xCU@d)9V
z_f)^2@2TOr!GM*c-2m#*x5Fz(NA}Vd@uov7j_3S_2iD4{I&XG2Gs$;ek65&Tn=FkU
z6r!;uw*4j=lG!S5&%Jd0%hU!TN>PDT9g@xRvL&SBCPZ;m%!Ve#Fl(?O*UB{Vl6vAk
z>^%OH>=2J~L}SdDNrQKCq@n<EWegoYL3WwAEDpcjo1;ksQM9`rY;Y}0WK4kX#S9|5
zjVwdg8`|EtpWtmmP_C0qVoSl1D1HonuU(J5^5xj^vLI=0%Ytk64;bXV7P!Y8?J_S~
zGxQx`T`$<~)<+^~d(PSSxYgh*H_D=C$F_$#d?YDHE(}c4sA;{_--x5%P1_%BYAqjm
z4@nZzTkS0P#n)d3&`47s%Ui}2@?R&}<Pkcq4~raU8`k!^TWmJM-bPKy(rb(zNw40<
z2YdtBF<{8qL{2iK**o(a2(f^=+vhq0B`9@<%ZNsrU1Z)2ecmOGdHN_W;P~F0M>QI_
z8+~lb=#7Nf5w``$<jM=qrM&9z29_g|@#yRw+$SnP3i<w8XB34A;&LtAVc}w~y1Wi1
z$(8%p>~ZPEW#utiX7!=Cw#^S?`FlC<n>rste`el8vFM#9gEJjy{A3KF^Qf<9T1)8X
zf%<N8#xhygLBh<O%XKsE31oZE^N+1{CoauF%_s`iz#>_vrYwS+dt6q(TZg+pnm=W2
zJ7jz<jsv)*HeudNo%DrgB^=c$t<-rDq+3yqA4A@5`I&xMfC@bfR`OGF-2wkF*#E**
z4!?Q=lzol?_-J1`-vH#6d|lcFb&5baAgC)fkKLLFoPwzTC^3C-=^Kd3xPzwdlmeV6
zR%E>b5-72)oc|5vUkD~QIqLsbK>F{Wsl!gn(l6f(Q(2Sr4YZgX$?Bq0oWFElMd(PA
z{#JUG2?7vD`zSv^tB_n}YqZ0!a+Bn!!N<RWG{qOndytd701wV0dHXP^M3q(VXNy1m
z#rOVLTRhL<52f~Q?P+k}z@av$V$ySTVelD9r;}i^@siw5+2bpF?%lNQP@G(61AnCe
zJeaF+Q9y{y8PI-Hy@Fp=v`20psRn9Qm0^g<OH;eoJh5ZoB_CprQIDISn=h~R>lQz>
z-6n^WUp&Oxo!Bn;CGTsaXf9*J%7d9pS*dQPw5d4=Z5}GUZR@#0FQX;=twJ)&+b^~e
zyUq?2_%@OJ1`^cY1Ot*KUm!qH@Hz`DB@X($ABm~O5Om3B1!<=uo8(J;S$o9XGzwq7
zt&LkBT7Uc2%Td_WsO+Y1vlu_UfDQ8rzjI^wAqC?;iCRu8fT`a~oC&50>cC5x4Db+0
zdq`Y>-TuDv_zb<c17rG=49`y(#@2AQZ)1(Scb8rXX{#CN=FBD_-8Pg%cKIJ8XT^aM
z=|xbQM&_bFFjZcZB?3BQ70B(=wH(w}SePUaMTd%lHY?M0Nhperd5|oNQtS!O6jjd@
z<vDR%9vq>ypAdMUQ<reS@9X?~Ld|i2ZBNukFnfX<{gulwJ+2CcQgw0@I=uqD)%t(}
z0#u{0%q8cmqWaQaWiNZ{(W?v@gg(nTV5Q|iN<MZp)Yfb%pYak!P8RGW{BS|B#pGmK
zAkKt#cnZ+ni=Y5QI*TUb`LbGdR|QIjHOnuBCeS}}6*p0XD5=>Q_vK`qwh1wM6g|kZ
zKY3qJAZ%1|e~wV{xCe<BkP4@X+oSKi=*R^;uyEy59WM8s>%Yd=T6otq?A<<N+OT?p
zKk-rSbgA%VyG$R-^JE&XPQqU8!ay3L5_pD=>RtTSBuVQ2QZ$9F9T5Cu{r(Y~G}a!W
zW(uMR$#p0;rRka%WR{8XvO)p_^?;xA!Y$5#l=8OK^nr^ac@6KyaaEn6NTOaka|`VZ
zQ7GLd$4=HgLpBKlzbuVnQ^}ooXL)o}G}r9llc5yQ1gstNwlJ!%ZBBW9$NPpp#lkXt
z+IV{IHFk48c`Pe7r?!Ov-C?kYl0|=GcAow20i!$6gUp++$RT6UFsW9ONuwODJh0jh
zmyQingycfo6>&B>p!>UmYv8l*mPWWn9^IPAf3S&qgfGnj7Mxv4%IRK=XKD29c^tg4
zd{lxQFyZMMiIi~^JajDnda76UN#<?aQpWpL^f{6;>sv_Hks6~~591Lw(dL{IpI!X6
zC@|t<+d4~>tG~OfzZiGA-Pp00$+hQ;$+jH9v=c)M@>@}ubggSJO^U`_98TAEio>-6
z;<c~q;tkGAs+UgCrIrAK|C`(WzWPpMLp8ARU~lFyMNLAAB428h@@YuLE9}jw>e!Ok
zvmSfpl8wjGpEVowk0OcS?K4|1HFJw>slaxE5>-3xr~AXyRR7UL_@)6CuI{Gdx!iX9
zH-#~wwZ|(5nnQzVEaB?mYz*5l?;ZW{V+btvxy8AW61Gh@1!W_NSV{sVm1pbT*cNnp
zU^*r2@BI%)eGNwN_d{LzyQeDZ906+hn_XCb_QkJL{D&F))fay<>A(8o{~vu(Pbw=X
zKo5|uwoG>&6TLR!s0vzVqMV}}v}WeH-{Sg`WyeE24|4L^0(28tuhm<waBlH)RLc=@
zc-<aMQeSqjj8=y+L`#WZrM6Ef?o`zs(Wj-M#zux>zHjNK*#^k`1R(0bqnW1vMPjld
z@DQ$_0J!FGNAcGd>=AZC*W-mjdARu_oF{Ha5Ifn;$f>&%hz3)NXuiF?M9wCdo&9~k
zbcNf&n@B`STX<&0qflXfSqdZH5zkDL%T<zk9daKM)&%G^C1(hc?#Ydw@8?@aFb}H0
zHw{?`p2l7emg>qeDnc`ND#|Y%F}r2;xE(#y*;Z>@C86I1>7yhdUzKklau=Qx(T(B^
z&A8BV+gHP#o;xx<W>(spmnX`P=ZTP7q+Y}rs~Hps*nIu}+I#PyCfn{^lvfl~L<9r`
zq&Jl&9i@o~2na}*8WE(2SSSGkQ4x@?AfO;1y#}O4ASBX7dX0n<5_(Ul0Yco*_nm!a
zzW3W_mviQv-<;X=`-5S^K=N=u&%N$zt?OFX(&J@}IfUq70Vu%dp?RUIf$VKSn6@GE
z{<e33>UZ3Z7(Y?>w2{|9?&s-3q<;l1_ln;AO}LklclZp&<&sOe$Q#0PiBjW4s%-Pd
zb3n_zf}6WyWj8*~Xcu3uj<jF=qJDna_iD-q^%W7`o~pxA+!{~<`4CJkfT&<mzkXe>
z08XEDD&aujg&F8Ng~goA7zeW_C17o_)VqFd-bC-o7s#Cnbcr_@?$zjXa93GJ=CLaz
zL#Qotg6*a{7BCrK<-dlc$XzU%gY!-o#Zq}gO4jsgY_ap+(X+~$IwGPuUniH)55yHD
z_?^REYPn}%Ns9o25+H0B@K&xJ?QVwV>B3lum>8Z)sw6yeMtmIQ|6C?V<*nlJ>zo#1
zMl-!-Q<GFjg;Wab?3hII`AD4QQ}eP+bfA_Y41VlZ=1(dzjItV6RC9Fcko1P1q=kxo
z+5X|THo?Q0B!1y9SBDM3H-DL5kz_6~?wCBD!o*`{fm4*-0a4;-eDZ6HikbBa5p$zt
zv<C-<Q*V`e<C?oB6cIAx_LEKVaWqT0oIr3X)72Zvug<B?+WeZKPqrDXgPivACT6pH
zDfEz;>_{0pCEkH5H#IX(t=_$oCGjHZb_O(K;MoWps^O~5U!_#rj1J71UGN1;;g<Cn
zBJ46WQ{Bj2Y(5DU{MLFuwweUwC<15gK&hsgfyu9UzXZlU-+)t7dbSN|kVUXb%4+FL
zptAD0YbO?&(#$nUqKxRFz9-BK?`-VUwc2&_eKdUr&WFNw+22~s)ANrr%M9&&?RV53
z%&?4PHXm<EnjLt4H7T#4Hva5c%bP)WKkBJ<a(v=!12cNN%&}aUVz~f=p{j<c+959G
zt1~Ji%zbgVy7nIFJc^sXb%<LAgMh}BXLOo|Uj^^wgvjEU+WFN<@nza2(CS-Ca<j5r
zP8oo!oN0mfVjIV2<okPA1y<1zy!Y16spm-X3hRWj0aBt8Mu$hv|EY;UZ}RaZsOa;a
zt|fY0O0l6l2i1(MO+A9t*mbFw#y^+ogv;DC?VU86^yqG_T4sjjRc2bWvdm#vCaKzE
zyG`6E*{^G|WzZ-qRD*p;bfvFLgSfwo*cS!S8wupg_vyJWi{s`MLtow}>m`#$%tz}o
z4g4~BiJo<dHlVAHUX899yAv_7j=Izm86kzsaY2*s0#h&bmRn+_GnytUPp<3IWKAkd
z@0}#T0WXzp6MvfNp5}4|dzS#it}H)vu0>O}F?GAGU6ZXbHH4{$OV?8*RX*AN7dNk;
ziQM~}KZ=E-ZXMbaiaI{m$;Jy@WWPt2ZeK%qaRuP|QuF4z_Z)3JzBPW*56uq}#+rP1
z8et;nDwuO8pR;|20jmxsV4`Se#}!#vp~>xGH^^FT(k_-sd^YCUX0~;_&~@Ax@y*?z
z2^V_5>+-wlo%HQkhsXpDfk?^J-0|l9s?x#*6(h|2bloxP#5c5A7ZxQEuXro~WE!Jk
zm!N<L8NbFTp>`N=P8%GTshq1}xs900UC8^qQM7#lAu5;Te754-^-kK-z9TuZfY|r4
zV}r9P$6J#lQccb+T_C0sX3W>z&31?<-~3+xPI=){GF#K9UrBcl!&Q39^Cpjun>5k-
zx09Y(Y@LLj{Gqd0zJ2zmFxnNck&Tx$r__~pR6=*!FTqSqSyuzP31>##z2WP-@=SO3
zdF@m;(*L@UyVY}d_$I&Sbk+C7S8#T9Zz8bZw+Af8cOSuwMI5s^mNBF>FP?<2bdh(v
z#XpWe;Zt*+F|%e9PT}Rq-(h&s_-rw>k^-U$Ag3IfXCM;5V7Vhtd{5C3=q&9r#bw`<
z;{H6>^9b<^8rC1uY#stQwhKD5sToVE#}t>wCU;-`t9#1^m@g|rI(S7jo|C=2lH>qd
z)4bVibfWXH>4U4AWt`~EPOCetdJVW+T5_Uauw6E(xMuthjVq4Oj1Ifq8F`9vh1c!;
zz@vpJn*SI|Ig!2-?It49@8^7l|49`aL^S#La%0^VRp<}hWu>k26eMU&aaw*(XGCgq
zL#ZcIPf}XkUPErHLVEP)nrN<gmiF}vt2!KlHNM!il)BIqw8Ip}pIEn1X~biqjLd*+
z)?*G8G~36N?)!z9*7m2d=`ZSYPTG6Q2*|9)1n8AdYCETPUVi#1;&e>db+TK^Y_q6W
za3A*fde{)ebMesk4_$MXYou7~n=nsN-Ts>jKzOT6kwt;`jd?qINIRYBMek7bfcKA;
zPhCr;X-Nbs!}|UfI2WB{z;Ai;kazHB3vAqhuEU#=nD(wV)SAyu_A&)y8AWt;()fl9
zZ^cZiu(qea@D3o3VNGnd(>al>-TSGDVpEAXuS(zW&N6eHM%yF;hV}>EYgt2!JIGo5
ziu=TJXp8P;%UA!^ap14M|9%mNs0cY^hGx)O&;6mxPu-hb`41h<Ldc-$$IH@Mu-Qk-
z<KMMnK-Ej(G+8KaseMK;+rqmnk=YSyZS+Vi9^6;HZi{D-pYc#H^ZoJT&^QpV0Tk&3
zvg5~V11G%RPG;n`uJoly^$;+)1ey#YvApX1RMSP2Wm$f{Dr`R|85Ox|Aic>sBO34W
zxbLcST>SA6lH<#vpZOZRLy@-Uzat}KpxkJ`V||>N{c92Q7XCOpQj-~|Ir6B1HoVKU
z&aE^iq%mjClhtj=AVw}}zteY`SVjK&_LZLg%9S!2(@88<WcI|qWu%X&#X(6lB)KLT
z6Z#l!(UC@ca+9jAtgT7+AF^HQ-3IeX>KLDVl~L4S{1KpH%?>jmOGG1NWqdZNbl(h~
z1~E5vUO3*Ah!>tp)@$dw!5bW@C9@gRGq)dK?iqZc(d0w9S7T?D*p%7wrLNO9-!$?L
z1D`;YEu+;xpU(^1IbFRmbylgVdFuptFeTuoBBIGg15^a1U<YaO%RiAKfq7FWeTJ&w
zJ!g8=W9A%ff;Ju7njU)c^;_gz7**!jEv;vI$4CCsxt}eIO&S*arbF57=xbvzjX4kx
zloQ{<<q0oe3A-}tT#syOI1TxvaChYHHXVH}o1XnzpD!qEi09@oRYN8=$9v}bSX~#&
zD3r-C@(d~tr<E4C_<1o#blbHZ{P~vK!;BrL5c1^yn=wBP27oxabPVE<Y4&&d+M9}S
zc467~Q#pE@O7ln{brP56J@$<nIO0SL@k(#Y%5;On3}?KjjgU?A9+lz?ClHES_u=X>
zFlK>O%LRUh`d~~w1YWY5CHMx!?J*wA|5dsoSNNKAzD<0t{Jp+Kw!D|OU^_q5u@z1f
zz5e4C1&Wh_1P=@s;)xKhKFn3%O1-bzxz7K+GsGqxDlunO9wsp2Ba+w=c9YCEo>%g_
zAZ6Fi$1XPGdasR4r+DOVeO{l3Ki&iu&4`X6?DszoZuo7n80CA|Ek*Ac`+!2Tbw@bI
z)!LO`oUIs(XeSFn^B+gBPJG&a{Yt#>p+g04NtjC0i9M!RKjX7fBW9)?vp;rlMUxA7
zn&=4`iHs^ZrBLh1Hcj3}QaNXE+#pINz~&4JuP$Y1`|gQ+fgX$DFy-a(%?X%7<Dl>e
zO%S@d)7E(EXS#c2rP8C0w`XS~^tW>i-0v}-t?3qYxTwleM9w49%4P-LHo{|fSuQ%Q
zB5l%e4)i7s5rXO0HY;+X-+t(m5)+*iy2Avkmfz&nVeNs5oaD$V8M_&?v>#d5@*j5u
zk4G1+RZrUMyP#)?3%NDF?u*KNy8{ne`1s^4H`Bb;M0dBA+R@AwIV-U#U@R%gax@k2
zDPQQ}gC(J3jd%ps-ATv#Cz*RJ;Q(Iq^yX(rg+#|4#Z_s{%l8PqH{vb#DqUW>5%BzV
zX4|b_0D0$(KT9#h{}>=(E<=mjj=I8k8+JpOy>Ld?E1bq{?W}D%Lu`eoxF;`u$fJ9n
zzN^JV^&f&+6fc|U7Otk+xxaIIw5#E^BE;F4C#?P{))gVVnQ%)Kq{RLrr|a=!GP3It
zN@A`XCFDF?$pLM$K@JngC1|a0Mmsoax}_f13_QPtZ{$2=u&q{(Y<uFA3OTCwCdF?p
zpd-|pkV_Px5YKJDR9}=$;o4^>X}$6{J-up;`$HGwFNr_5W*6D)t(+I=yYcd6@Q3`F
zu-{xSXzQ#<&+c0i^^wWP9261#LD=>Dwo>Tp4U`ITLLG4GcDUnY)aB-K_0wDWiw@MU
z3@bY}Pm5vx^~naq{-zfwIvxC~BmR8K8jilZMjj%OlJ2KV+#A!+xc|5nBfl-fVCm~{
z$k?8UYqR$*>m(JX1*5rd>qq)eGQW_#pJo59a!)t%J~i!q_wmmcG3fixdW1yX<Q*&)
z&5SHJmNzkvWT}jI9fFe4_(gzgKw_w8H`V2iMVGVAq7qmxLh+q60aAW&jrTbMpMqrq
z-o(~6R?#_?^@Wg|Se!8UZE~k-rC$Ya3ovt++?h<g<xJYFDp<2S(`M(i%#%0~LF^SW
z_@E;?aBe24(+V|>Q4xFgd=Y<9%gcL!9d?V9P!Ku`4zVPg6y)kxc-c9<<|wh3w05&<
z(0n?`?+}+@t`ztwLQ7oxaz=Gcu%uS9o%f1Bl4Fc#{rn#~1l*biK5!qjv`L~mQYN(`
zG%tEOsgZm?%}9k?HVvnaIkUbNtDEkX+iI_Fdt>e(V;|)wr5q`eLD4_H>*ST*n=TN&
zi}27mh}lE1n_{2V*C%v^Hh1*sx_^9s-|_J)+D^~4j<fPyKTFL5Jy*PGoPX$)c1E6u
zMrbBdGl1IQ{^-G~WFJ+QjMXKZ<mKpcW|Z1`W(ftk4DMy*zqMP9d$;{WI;l3cQFZ<U
zY}EKy6yNJ1GNi)@bb<1ESIwB+nk-(%oVcGkR`LFuZ(Y6A&!k)Fk6Na>1+DJ(!s-v9
zCBb$C_Yk-vMnl6-Em*R1FA-PS(r8kW@7WD5^HNDaDTv>Sa;d4;IcetTIm3W`St;kL
zm}@HAbN33hNbynFNTcp1;w3x9s6#oJxuRJja*3?eS698V5{wj|u!-*N;BhP!eH3$Q
zR?g!4`2{d;8oXmH8M+~B^Ta7;tA{)mb`F!QkrcB%ev+g3lq~rVuRQN4ZR1BC%tXu$
zmi3`x?$aRmKXmp>OZ~#K3f<cW)TXthag682T{s8S#gw%4+P?`Qv(u@<&xszBJ(GU7
z@04S7{C5>^ajT8+O;?=DMF!A$am_shmd1EknQU5OdT$bN5eAa#*yC&${BorVsjCaj
z)xi@PaW&Yt<Y@sVV!@cjw&}iZMX?#b=`=Vc-*Ko*G<Gr9*HU$zg{6cxh^2`Gijr$o
zrCjeucmz-vVCB^&GX0d+>?p&<N_NX#+Hr!mK=b_FW29Tc%(WNtCyg+BC;se1%17d~
z1^D36shC3~5#PQB=A9c`ic$hB5b>lEMOB!}(_%Z_{dj)-Mj)HRbbe9jNo~QmD&5vg
z3_)x5T1;pSuNn_D!!S%cZ*@kH%<^twG(PhrB44>%DWU*(54lrbttVnDl+OL_)#mAS
zhmUvjwgC3Z4n#$7eVT6oNp4fE{?N&GnRKCsoQ!h9h8k_&Rg_y}+nj_FO~*OsXsjF7
zL^ansVE*ZhIl$myW+IYQTveFPH<k!~=mR*g|NHx?qI?qIZWy;~Al0MNCveA^*rt(}
z%Qqwe$bgg8c>@+x*UEv0cQJ+L>k=8^7j5$~Pe%$=V+i90BJTLQ{f<jX77O(vb7_>O
z{m3DdNSW98R?ZqOD=kOV+#9zLkJ54-O1;M>n<bhe{zT;IhqrP+%^q>mD_MKPS8-v%
z!OO;2)3wU^kzMbGRX>|AR5fxyNhRfgSdkqRAS31H=rlfQ5~mm59lbkg``eYyUfysc
zRh0UBnXd8+ngixfRyv5Jaj&scjyr>~@6>eyizY#yXmQfV;D9%+fo59C!HR4CG9{>X
zPNMPp<ul2Od0UE2+d%1L65g_I+5G#~E{L>*^RxLhMGYIm8r3w`4(-L?4Z&XmQQXG+
zPrrVex;^9@yS;Pc^h<X^YqzV_zUR+L?j-)e;4HfA3aeboq1$kdwy(00K6?{F%IVjT
zjSc8555b7jQX_{_D?E1^*tTTczy%(OsC5j27Zj}X(ZK>zbhu3m7VtH50tPt^@#shT
zfdiIoNVI0>(&gcmm6th+0oyaxGgC|<p6orJbNxM45=4J)Pl~8AGR`fvCvq*<!!P*w
zF-x@HRvL(@`@X(+PLW*D6XJZOK0P4(MA$mKPh8XXUP=9}XOnRuQAIOp0Kd#qjK~X3
zhQGj2RbEalSAL<gAM4!nKj-hg0kXo7qNqZ4;+3BKIi5bq11_7uz=tw~Q>8(}A73k9
zXN~MUHJ`rE!U(%pPj*9;F7J2-N4~CDdETbcpH><qZWRfpS?1&<-Jf{Waf37;&U=cH
zyNDAb&@#DuE(B3DjE^zhJMOQJ*@<VG^G}ObfBHOkg~m-5>r6XCJ+gpiO3i^)Hr|IX
zw%>o>n`7&E0m0VMY#1f_$_r|=|8@xFi11V%T25cg%wI#?@VI@lds8~~+fd%mR=%X#
zl23CVu>A=dt8Xiy#f?Z6W55(~<a<3hfQGr)oM@hlem;)zF$x)oygl%)mW6@x2DgX0
z9z+B)9>18usVZHWwcI(&^z!@J^hZ5qK2o3MxMs6z9-ZV)w$cP{_g9gQJ6S!gl`rAF
z?eYWaF>WIqxwZng8t~6p?WMrh&u-d?eSB>lh}^^O8EX#Q{o(*#g16!q6QAUw0_`Y&
z=n9~w?RFins>v$R5PZAx5p#WpJJ{>)R8Jv-D0;dU-gxbt`{(M4jijnLgyaCurLF3>
zE#-{|%E09Uk7jeR<8roJom70G0h^<)O0Q{FoZ;65Q)%R_Cwi@1MnJ5683sz<RN@j(
zT0{2lAFprB3{-9G+I*Ox4dWR1;qPSp!7T0kS2OO{)u8RAO~yr@hn|^==P3K}!SGQ2
zSLDsas<unqHcB<K0&&rEkSks^E)*&49ZIZXB|!NoZoH9b_kOT{JO8OD^84p2i;Vre
z34bjKDyIhkvob(wR)hVk*j*|Zf?l284Cdq5&U$@p2E|&F`(ybA-qdwVXRUQ88-HE7
zIiY^=rga;4O>E{s>|tl(4fKkOHEe_V)6Uxaut7WVDwli*O5pu*m^m&omh0UG9@l45
zsrBYdJxOMYRdH3ebh=XQoz&+>>m|Ez68s*%d*3ZVTDeIXy#yKXT`D$n;EF{9N}yKw
zc`!{#G!VBHwVCwG$SmLf5XmL!FJ;#15+9SCEPvI%6Dkq2xTWP8FR44^-;)Vx4U-tG
z0}JhvWnKp|m(BF>ZTvvir^?<S5dAz=uAeL}{Qgm_dS;?<#BK(!p}031QEKcn;GBy>
zTAzGX>|<8C1%atmw0Y;WoTdpiFnO)yd={~N5XkE;HX-64c&z&IjNtrYgWRzxXvImB
z9QL||B4CG37->(P;Ri5nwQhz%h-jk+>M(Ztl|1Wj1)qmW_jpem)EaK~9F58Ll?5)V
zdBU^zR<KA5d~v1h7+T$a`S!XpE-hT$#wDkxR5uTx*`)=IWb2HsMJjFg;yteoSAYF@
z$3-oi@miquWq7SycHxwE;iF|Bb16iG7JGlN$KfU)MeFZ};@WQe%D3@!7Tyt&36w9s
z@Zc@O#}S7fFq}z+qIjajq=Bix%nyKbQD_z%)4g#izu4@HD?9@+_78`ikpBvQkZ>kx
z{FlSw3q$c#xl~EPvjv|%;S;Xo61)1uJr$ua+1xd!+u0VwxmXUr%SZwDZ)w7k(agPr
zakrT3F4so6IIAj}P8p-K3joIbWT1wtau0b}*&L`}Wah;56)?M<X*d?3Kss+fc;W4&
zJgVr;Tas@ingp6rgnN9J#y&wsvl|Gv%Qd{e#;B^E1alePKrxdwciKRgGdc!xq2|Zt
z_HZ`O=Z~P_2IvG8kIx&r<>w!(YOM*Kmj5g5!rjr6?q4iWJV{1w@wri%(bmYGW>GFG
zWdAj}vxdO(${jSKWZx%Y5w!a$`{77=(&nOK);{if+{0S)vIc{zHx7MX=w1F|DZog*
zgT$Sz8O#ktI|)VdDu38OdCwB^?DSM*w>Vu2?0lb(Kt4gG0=~F1+jeWFfnGFx0V#?o
zjqb-t{s?1)Jpv*E%J~^uw9(TK!sQuH1fA2{C%1#whh2Cl>j$|rekJ|#L7TY0cRx?J
zogD#C8g<5jt)Pqkj*@e)0m{D5Y0bpg<_C0GoV82w9rA65r&rp|vm)-vAq`9mgq?IY
zaPdZnYz8=tS+R51JQVW_$VdQn<>;|_%&ev`^>H`>5KPo+oRYn;HgkQvF2lR2$w=d(
z+6zPbEBBgaf^zQ>{Y75VJ<|*1uwH^m0vp@Dy-FuUGXvql!#5M5^={oC@#JhC5>@?5
z&zFzmsi$SvlK8z3g|ja*x|CEJF?;{Kx4jiM1&vv4x<S4X$9Fxjg$Kv*dsy|0g2QfD
zBIHuXsaS-jG#cC*#MV<QPC3vQ?`GKhWh4BDu2zGZeOX;qNrT;yq#thI7Tk{vIHGc9
z_~s&&!F;Wso~GV!6|GL@J%sGUcfY@RG;v7d!qAeirJ<kY4iLUgK^w>!^xm|<9Wy5(
zGSl3AJs?Ygd5a}{E`#cHaMIX-IbwNm?>3)KC?QJo#pxugLq0<9iwMY!jAh*1?^?BJ
zSw%<9pbS{`Was8@dMQgb=t1d1FusV&K~WspVV24cp>e{GUsMIeTr2s%w05ZAor6%5
z`Mha>Bj_?^t>N#lm+2ubTH9xUnj07X$VS?fkj?4obhD<haaab8J%iZj%vLz`&2WEX
z9pl<YXM)w^zIxS_?N;SkN=BuvXJgS@DXKlu9>YjgZEV8wnvWssg8bg9l=x|OpGH=x
zL(a?B_0WkrZt#>0OlwN#QkKj}m^Uy*TtJa8hQ*g^K}4d))pdzY0~g!3<8^7BGD=sz
zCRs`$<jx=8i_3fIO4sf`27u9mN{ATNWI-u644$9896F&~<qRpcTdAIz>?oer7aiY+
z+=wo`|0|mN(wFKIzixW^;H6@KK|&XlL*I9kMmmzbB}*vNC`sZnN=2-qGI$($z$7YN
z)LJI5z29)bb|D8SZhsU#Y%OEfio&qNz@;lQQfV!?#l&*FE#=1k!tS&}M|xC@OlNs%
z<V3!Gtf-LzGN=?BcxpaQ|6Mozkml+ix+s+ufeyc>y9V16RJ*Y2p}3Wqfr_XJ0FpW9
z;6lsHn~@_Mu%CpTih@VUu!Itx%_j_t0ao-SQgrh}g{$vZ2YaN>J&V0Me5Z90cmGQD
zEj=MuE2;eOx=yP|ZiCZ26l9kI>^52AN2s+UJrNR(vMIF1LwyfyzN5%u^Mj8t`*Fga
zPRr=WX+1K-_J=vn+jGw^-($GFpQW0x&CI6`o(na-NAXsPR=48lN1+9?S{@9po>4sZ
zR63o)e+^Zs{*~>T3!ALSd9QJ>2C(KALH(Q4)U3DS1;&<+c2o;^VKZvH==WIcoGdpW
z2xp<H1Xg-u;m*Sw0q4XXWm!vY7QZWw>EG=CRa(j<_n1x;+pEb#o1vtX!)^}w53oZE
zC-qPxR12w>k5<!o!2Jd8XZA`xM|#Qf_w9fBJ)3<qS#7|k>WT6gp_-q(qg`Ef62^?m
zDDjq1<up2`3#j`q`b6+Jz0EDA9MLGZdD5|ibIN^#TcTL&G`xhs(6Z+hgzQOhkml0Z
zZQmTuvLf6}YmKN1tJF{AO?9@2$7fltzINozvU(**x})TVz7yM#{F_qVA`rWmM}EP=
zU6{!dzA#=P8ec01tBO*-u)xpm{5gJ34-{sQECLQ*T3=$GUQC+{_1dx=^9SFZ>$p9Z
zE|XVcs(48UQtXf~r^7CH|D!Co?&()wlQ`~{mFV_?F$6%CiIhzbvx!w*K4`}5R<=zC
z>^TMO-<Bi-mf`~cU_-P7DDnUsQtJV*A=3i?U_(+@{~I=BhxL4X%y3ZnEjV+S)RY&9
zNAs#oIPun|NnNqyq&Fu|cA|`u<HOg^{_<V${Gd}<<M&4A$@JIFBTl`4L|?iF1U~|%
z^1a68Y#_;Kp#p6QZxU-!#zYOZ7TVP@r46UK^>phrZ|RNc&QAD^D}zn-V`maL41S#c
z^6bW5U3*|E*^~(qV|g+DjeNc<e`p!eh=tu_M(`vH1(zO3r0SN#46}cv_o8o*78a=3
zn~$asr9^GPU*sPt`q~QL+rAh5eIwb|1{cO;=Lj_-jyDnB?IGN)hFwV&iRD4T{dv-J
z`u<Yb55I;!UpX}Fxd3lU^zpJ;uv)RsYlM?D$4+dT@@OdrFFK5{J8wX0J)6|fO5-u1
zv7FRLYy{3*`L84TYTwJbzKFZ~v((mr(L1i}f_&0hC13Pi(wY~Zl_u<Qn*l+rWY;M$
zE`iPuT+FC&v{ml`mri#1L_W%NWn!y3%Mr=gb6|9RfD%Z|e5p+uXajO{M`&6nPv25B
z`l=TWUGWXCG`~G%FP5w$mRZ~MHO8OqXIcHttKW9$v%D2g`r(o^Hu$kOm02U><+H>9
z^peXuPLcpR1+yfxLwi7g<h}w&dA&%IsVcFXn1JrDb<lhZ=qH(6WO2ew-ZytXT@iYL
zXnId)?w3QZYp#ry)*?Iv$UP@=e?9FCwo+oG(*GD?E>hZA?}qP<wDF6aRr{_yE7SLT
zdr(EjCT1}-<G2BH$aqkx_LQSfj04!c+iJjwT~PQoW|G5vz%ue^^9ZN--BxP;QGq8S
zKl;i?8wQo9Ou~dnTP8YfbDit-ivkl-U5#BaXdzt)gviHkKFogCeXQIE=2A>aAgT?>
z3-R{)N)e5?>2}5t2PL<NTGL5_<iZjJG@Q7fL%@<{nw<+AVE3y>vNTjI(<dVGpdS7E
z%gZf!dha1>XN3}Gb(|x<Yd+-%Dkw>bYj9JIGYo|VGToR)bv)dZ!Z>=ARpg$#g0zdw
z2|N=KQEsf!F*X`|^<n4N8nv2B9CuD9nb2`$=TO@8$>Ig&zdh5+0{!y_`07?mZ?YPQ
z%JsjJqq`~B!~4`gKw06vE^MP%KZFSR-A3F`4uzL`RTj#|C^p$gR~ChI3)OvZH1Z1y
z4@$aJm7CO_M$g<Np&9tK%ezo#PKyU>(GlEEUYWT;%B;v1%fhJd+dMP|=eNjx3X1SO
z=#rFKiF^B)<HifaHx?O*Mtb=D9Rtu74N;MlD3b{h+le(IH)$^BqGh6p?(qwhj^_oA
z-nYe*i3Q}(^-;zsazM_0ckPFlbsUPS>oelf8+y9^>kQ(0AI`tWc#L+<yeV!31eWkh
z{DRv~T9e8}{33SwU;VUS+Nn?@Q9w@Vd>W8+8wGFd;tv@Cogr3?cK3TcFHOe@Ah^H1
zfv<=)2OL4D;RyIz(jPjZJDmUD+i9q}0rHbMAPxNHzeJ0#E61=xwK`;p4gUBGlm^9t
zTAaS;^MW;Hc`wde@tLU}E(QdT0<~Wa7p&eC+DN94LJs*AF~^Kf@a0Ln<z@QC*nu~+
zYb4Wl)5!X;vtur`>Emtffx(dtr5@FwM;RGCO}FOtV{YYchc_7x6@R=WLQ_}vQ#U*2
z_v9BGq^gR2A5t2)w4xFmzn<oFleBJE7cr0A3Bu~GQ-9XeNc7wWm%m;SYw<Z@(dA>?
zpn+*k7q^b6C;q`(C4_x=O(sCJ?PfFd%nn>yrty#kqMV!SkMD(XNBH%8@|_mV()u>p
z&G17sM*MEItIuI?#HY_x<Kx@Ylu}yA8qm(t7Z~!~`Rl}h|MQ}*MZ=?jTr5KJ?8GH*
zBOhGpDgQ(CMu;(op|O%o34ZBBqn>TZ7H(H`m=D_^VtyYyAWI`Yj+hcoKsHt5{_!T0
zG`33SLPXJ(I^JaU_uXD2LT)<Wq{@V^mZ-LIlpSrYw_@?<cR^<!O;Hqo+eYISdu&Kb
zh|V<8>MHncCRNw)pC8UWj^q38g6V~4HLu1b*7S@V(VG3ID%=+O53pqXjyk4Y;LOue
zG~)`0WgVrX7-;7Z44J4yn6N_!q`wUI@%FyCC&T@f_m#h@%Nx~F9@8I;)pexxz}vfl
zrUjI?hh*0RA4Ac_Whb|(>mS#d_6M3%2hs+8=djCVFJAbXmNikex<lI$B0mA#`wBZh
zvD4KG`h9SZVm!wzVfHP&RcoU0dfGah9YuMZ*|V+eMr`K28pRB&#J;P~njS;k4#PoY
zt#DLODQyTx%OQZ>@NLTMP*dZYG_%oKiBX}?5z~jR5pRcX^u8WKy9oZ(IWHJ<5OV{B
zN=p4lmf^i8U=TD(@}x!e&)hh9@3UtlaM3e3lokd5x_l(c!{dooCXui%@lRx9GyD}g
zuBu2~je4IKUaj@<OpB{lwefUT3ZYpZ-+2j<mGRYsM-2NppEGHG#iq>va@7Y_)}U;Z
zwRvUyFYUwoki2W9=Fee0A^4>VOf0RO8Hp*HhHUGL5VYvvt+K?*t7}@2V;idI(OZ=n
z{p6bU8OJZ2lPbPyJM+H%3@R>HT@T&g!rky3RE=T3VwBR{iDs8Y{ZKAxVQji9D$8S2
z>TwXy*tk|*MbD@9;Y7t65#L3jSB!&-mWt`fB#>x1hOVg&gHTfMYR{(M*rc6Z$-Y?2
zY;-KlC!{!3gm<z(sr0rry)eY>Vw2c=VNhn4``pwsU3X6{I{shSwyLSMp=r&uIcA-X
zq?D>_$r(988P5tWzR6B%?%QjEw3`w^xh3_JkP1xNq3m~=(V*M#AKu$OhjRqk^^jbc
zI5{%8__sDeV14>>&;B`S7_T=&EnzH0w?Ef5-MF-CF+tVwn@SC=r7lb<tfvvw0pV#Q
zexD`fH+Jp6Sg}atqF74@vxqs)>qqcEiZd^}|A$U_(fTxlPA%)ErFk&!<jnK|eup}6
zeoV{l*5;lxmsvI0b3_&;>+bGNFs)vRW=5yFbm1~L2vB0$#trp)t4CuZe?7c@n0Sk{
zG&>NBGAmjO;1?Rl8j(E7{RQPf1g(EZ!uLNJ%WpsmZ!Avk2fD_X2u=NhCNrQv<!0hl
zR0fv(whnS^Uu++RovR!r2cBdrF^q~1cg5rgu(*=x$dVpf6Yd+?UdInna!eRC$VA`I
z9pxb2L&?P_`N{jC4N(t<2-W1vtkWxK1vJ^q$iMLOlMlXG%iTUh(I3ftbH&MlGgeRX
zpL$PRD>3|q$C}b~LpzkGHxDp#07Tm@Iye$ggIzyK@RJ4FVu!cNH>iR*>%=7{u6h!<
z{C4l_u@AcA|6JU8-vrgn|6_6eSSBXhd8D^q`M8q%n!CbLEXARf`ood2^&2^yBa>as
z9YbHDwufPoqCgsU^GP6!UoOI+EGR7d6A}XiJwEPkTME;9Hb)(VC$tGZx$CVVvaV^O
z9K=4=vj~~)->RMS9g6?l-B0g75VeM&8j`TF{JX*)M*%rS8SEDcs*T=m)0n^><ofT<
zk<1glTWiZD>rEhXGjKQ6ni~A!xklUf$#4eR04^;GqAmhpi)v~6R?Q0^=7E_^{GT2-
zy!#}IjNPzdaI<YUd$8Ih99so$w^sRov@OiN(+}1#S;r^qAnep8h7UYNR{cZg9&g%m
zz@tOGO@2>GpQMCg+f>|+4UKgd>+9Wb8neplRZUf9sPwN%xkG#PulCHDPW+*>By$|Z
zBK+`Ng(-wY9%0#N#lu>R=IRqcBTxSES;JU*g9{$<jo0;TZd^{^55AEBo<5|>Y%PRk
zoEn<e@T0OBk@Wdy@lghVK9LPNX&FN@7I?Sy6=nZ4>Tw;bZHpJXS_1Wuw<e-<-6bwt
zo|L%0Mm$kzp|Oxa@6?0xR<kaWM#mM}AO#d(BgwVYniMglt*>bFXXUMUrj2ML7rF{>
zd~uF)dME<G4jTuI6AQ~w|FreBvI$DlgC>6Y2$WJGv~;H<pwV<+Pv!#FZ2DYo>QaB(
zwUByONuEm#<1_c2V+Gc|3#Os`rWaO^h^wjbgiwjLAL#T5SP#=@Uy<(~)6<T&^A88V
z36{Qpv^lrjcwJiL7YcLcDFtpL{mZyP1#`F)7M!gKh3^~J55r{3S6U(GJdo+BIR#D%
z6s?wT9^<8{HqWQt%9U42#Vwpq4f<>&SjzHArIfPsLx7awg=TI(hW`c{#DT7*`XEhr
zoYs)9G!?bU3!j?{*bG~g>(_VGWjjcdoP_sKKQK)j%OI(K?)@KVosEvz*Ym)9@d}v|
zj~T9Em^#x+u}$wcpu<`%GJbk&EK8S}>D2j>GYOk&px5|Q2sBF%!jTDn&fdu@%uq3m
zh4h7oxm47-D#a$%_DY&)TFS@AV5>u9a2<z3|Hi?IW2@6K3R5PE5lxgNCAMm@LKWJv
z(JpiMu|JSJ1&bAa^!Y-K@>2jeFCE}2xer&?e}AL3`qZ5r!#5W};c)k9qc&s~Th%Dc
zmJ>k`U{LEOOiL@ao$0;TB8CEaeOHk07mjhanz-y3>nq%x8xt$R@00;gn;!Y5VIR6h
z**e(Ugi7J7QGxt_kuZ^BWQ8lWAO^3nOWtp*+Dv4Hk#SRhE%HTZ4rFF42c8fTf<Vy%
zRS9|Iy93tY4rGZ4|66s&nrix(G4-nu-`WYzX=i*PDlOkm&+y!NaKOwk-=Q;d)$qn)
zwu%afB25FQ4M;0Z0`*UkpV_kvTLSP_qm$-DN~tu}U>{WVhi*VxbY;e5O92h*qs(n?
zXazTHxa?cd#=ngrFobK{XuKz}%n?6vIB`;Q+qj7zqG{6W5`Q?*J=}a016qKjBvF&j
zcDJRat>f^LfLhrQ9z=k1YBF^x0ira$13Sbn{&%o(N?k?tlVm4t&FmntQ5&<epQrQH
zV>++8zWaG5_Ag5<cZ#|V5frPbKrVVG(TJ+bzCYp@Yt)l?VTuwY1}$hz)#k_?scX%R
zp6mH}-_32ZbvBeyS4kq^B<!GxhLPl+c(Gk6+hB)3G6X6%qh!o!+Q2~=8|7V=a9aEO
znctIBYe>%e>xG{Rv`ePio(p%&ercg1qhASe-I!%f5&Q0Fv@bloGz=518Xy?EJQ{*&
zFh)r<k>TOQalzCpM)JlNWaY1jF^lXpkG`Ro1fLGo(A6#f7!n3Z!{)PRKnmCa8{i;Q
zH=0Xqn_SbHSZr$z<BZgAtU?cdt@q|r)|xKA9iQWg9yl-MRa)HtE2fyk^gQFaV%mTm
zCYJrtD9nPm2G(gm`q+58ZZRXD*GN^`@0bvDRc^9&TXX59pu4C2R@^nF+GBqXSU+_d
z0?*1`IPx(Zz9H!f+aZhWYYhSXB#Rwc{UPNZ8a1eyPZJ{fSJZ^C>+*Co_6L2DQp?jl
z<C^evO63f<;I2a0K$<l9A*E23)xNC&-m2|Q<QR_p1RG12mfPLA^63GWvz|WvuU{$7
zNms!8b~ju$C)0|kEZa`6_L`5%VD*|lWP$bI8By}^8um6e%2&OTwS{;^Npg!RHZxdO
zhAb2IQ;DnL@G-BQm}PW8j@i!R%}sRy5(|yhl*bcLt?X~%(B038+_Zsx4M0&<CoA00
zxpi~y*LMAqfaW^&dq?o}<)no~;Ym;jFwf_7%*QVo5a6uQ+Hvo_owlx9Fpy6%aLH~>
z@2twD1hZT;4Fu{gj2-3LQa4jw)2}$W-ZbRVUkWKX&0vi?yEQrMR6=Ohd4ML#Mz6yZ
zp7Te83CDX4gSlC}u^!wwZlOr7H>%<mQon`lq%_rY=_&z0bc{7c;PD|e$WfK-Odxy=
z0qG9y`b-I#oyJDe7%4A*C_71xN<ntjjP6VC9iP7RbKAaJzNRAy!TS+MW{=c+`tzp<
zOGh@{9^~E7+JWYEvc}t^j-^O{mJ5#@Ja4O~zQ^f6+5403&n#MhuDN`Vd6~|l(?q}4
z>|RFli;yPTjRDf&^W<1#KjiEthP?x(O^PB)#AymlQQRHK*vPA0GTG0Irafuw-^068
zhgW$c6hv0dgGy)ao0qzXJXdS7>I)$ZbjWf;L5`&TCNg82+;6XOd0Tay#M@1<%o8?e
z$CUZe0*8+>c?&<Kc(r6$D$R5bDzZVNW(RC$!;}09&yRuD0$nWXdIg264ad@U@=|-0
za~|)}D8;I=P!&x~AAj<iNishw$pmZZbM^57=p2>t58ZhTAv;oCWPfWzYu!<kR{#k3
zHknm-w<a3um-ytv>c-7<r_(nj#j1X;h0)P|Q$7<axk54Bw9b(_X07!wlP$FkW1q=`
zP2PZ$jKYh7dsmkF*Z<HxEn_65g~M)@u9+5lHgdhKV>vhVZUa_?O4yJdDzs=u<>pS!
z!lp2wwCoqTX-066_dil?cuv0iwc3&2WT4#Q-lx+60t>LhtXN`(P+Eb6GVLNyca!Z6
z-?vYKp7xFerCkt;T>?onjJ)ggipQ=mR8V>>BdGAX5D1&+drF8oX|QHw*Yulu>cN{c
zLHX|oWZX_&w7r`ndMP?QJ0Ld-F|tF2z)!B9%<E4fWsOllLL>^CAYguv_vnxgVij9d
z!u=-v6E99f17?ofG#TrLm+oaMb8ouZaxcB7%P&=w%X;w{XM?9|72#I`D#1ZPkigR+
zY$HGGJ?vCd70pG(^)7a_)z4mXP6=2rQZC?d=mTm<m3vRB4P@{LO!t|~X683@|L*nk
zPr~5uguwrVAb{BX!qi&Vox~D0E(THE-3iSaGJ8d7CZp<RAq40aenY}Vzr>zCI`6Q4
z^83@|-X}H(L06IOvay=N$nPW=u8v4~*~T@)REzrdotzdggHm{Q6q~m5HVJ3fI5kyz
zmgRkmes71%t;a<Z|I`oDfIz@hhd^|W@1B^SKn}7!hd^RpS&GMc4TKPuI89Xysh|d`
z+<9P%8umBpW8<^8aeMFh5hs{xYsvJaK1m!7d&#PPk=~SoErGhvHgEJd&pDc7&Bnr*
z?HoO7fcd2|vPw$!bBgtcYFPwrq`T|ksE^wZx=5_7{6^6<u!KLgnuGE!5^E=qF^4L$
z3w@P&+nT$%XkkhTS*H`{9*;717H`V;;7OYF1lp9}^^{RC4@VZV{c9h%h8rVmKKsl2
z+FALlWxUF;bj~Ow*X=ifiCgbz68sX_Sv1gO8n3S8>{8THHYJmJS><{`L9g5In+dw_
z9SlAQe_?O9TlK_KZ`+K9G};N?R|vN->T@`ni8)LrWXC8XfO={fXY<^^C2jI8b^Ve_
ziML#4nE(!~8T=_t{~PQvSMPCJJ{BX*bpJ<CQ9vgU|J?D>VTxo4Am^W0O&^bCiFOUi
zBldC8*rs*NFm$>?uE~bpau+kJA=k=+-WBL|c&fL2j>15c;A<n5knb-Jf+~k5H^E`b
zBbYSPn&nDB@Z-r6lLS*Eb9ESnVA!WU^tK3?48ro6(GT2pr@?oA?Z;fSbG7j8ci~OB
ztMHgkUxT>C>)-+Bt7XXU<l<=_8_&zv1{0>WI#hzgN_-M88wv7Q<k@>Em)kBR&2G3d
z3$lFD<5NA>OIj-;VpmQy_$(7^nvW!Dtng!h^n;_d|D{^6HtAp=*`D~$j`MXXrLukh
z$N`s@09)Sed$HKJU=OyQQFE2Qw@P?oEp+PJkHfH)r;6$d<cqTG9w3&@up7`V8&c<o
zpxIE(1Ec1e@uJ~`k}`$+GA~6hMiTw4>Lq;xBX=Ghf+y8z8>T?f;VnkM|75Ly?Amri
zkTPc57evtly5%@_0;AL+{@%yT6pQxY{>sWJhrS$7L{KSi=3S+IFEnK#b<xU&%{e;N
z$M-(9%bZ)A8zAOQ6J$`@WM1OM50DDmRT^*2NNrP(*a+&WKe5SUA!Uor&FHu|>8@$=
z6}dOd5m3RJzsh2*Bv=*{g0)SLmN6Dsf1tJJN3aP-Gua<u#5+07#1yaXw^ld;-ZVHb
zm`b?^&P#iFV?Bue)DIU4j`3xh)NtK~y~Bu<4C;7;jUX*c+@6MgcDk}?)lH}DlB)bG
zNz=#wG|NX61o0bcV?-_?1;$$A%`(g|No%uP@jlWOf9UFEvI6AAQ1@apkq-hpSGPN&
zbxwC$20GDGnrGpssXi1;J|(-8Ux3tK6I!vd1u68x)iu|(=hS9!@Ul*E9lm~OBlzL9
znJM1z7})z-fe>jB0cj&GCN>1mcET^!1gJ4v>_?C^iyOy3@KAmoI%h2zHZ~5Tgepuv
z^zkZMNCjH&7X*qh@Xm#u{i$y4pBhU5h5w<u1XBnh4hTTQ+q15X=Y}%2Z4rgoKjpc_
z&qs&#f!X5M-_uJSx;gV~^*pBF+Y+xr^>x0JGwUml4J+Gveo~)S)MoL5joL`>v`32s
zQ<`d%o}+-{g#7$<%orsxA38F(6^j;{p@=*fRLYbvsts>DQbv?`@BD5q%%d+@PqSX(
z`_dyLaV(tV_?PmhmA-4W4H651D`4=!c2O8c!b)VzB-?Em69(3jaci8^!K*uy+gj0#
zWWao*F=~Hfv07>l6XM9;7Mxn@@_KrRude`A`q?9&PL6RwX@sKRfzwZ*s<~Gzs|>B)
zw!)=G5^vh5J}EDF#9(cG#_-#j`+~3ZBo`rc(5?;)w<cgW&}uP@aM2~9u(VUdqNCFO
zCl^Bva=tjNUKgd)5A-q1O?j%<zn-!zH;_Z=Y1x5souq^b*?HP^%%|6(V-a2h<@&Kf
zed|3My~*OeLD1B@mh>rq=^1br$xe4P_*PPsG1{G)!=*+1sHVdXt-cLP)}$`XWv-9W
zoU{&S8y@MLQR<8)ZQ?uXBTd<6*pzhk*ZWOBAFxc!^?mQafTsOA24b8v8WOF@3!No%
z%=X%p*Pfo<vTY})x-8Ri8pi4U5*8du-;K5pz=|)VqQwEP5dB`yu9Uh-DsH5yF*pF*
zXe(_Af^WxS>ytoghbntf(J1{Qs8(BgVRPA3XtLq(#?$7e@4p8Yba;SPuCG<CC8~7E
z31^(@zb9<ZTql6ZhOdXd=~>2w57fb2jvqjiAABFCDL<kdcQmF&`S*lgyhq9k_e40@
zwuSm0d3cH@;3TZj;)Ejo9X4&APF3yh0<sR#CT<sWg_8V$h8sfNf^!sRp;^E-9k|F$
z(O=_LY$VUr{g*O@VS}zVIt|~?)#-7+s{jk6&B42R3iCpAE0Fk}H0nULBAnCHZXZ$M
znabr9x*z>4HH{qPi+)gf;pL6_SG@m~_PG6qp&5#71-p%yQX0s?_=l*vnq~}HERK_8
zaMMD)?Ox4?B*EV@ck}A$`o|=$Ff<pY)W&Wm9}hnU7zV+An3mr{{C61y2fcWJN2p|n
z|JK&pO`2&6RiSB*WA?h?@Bbr)e_mC9|0Dx6=I`4(f9UjZSbc}NfBV550Neck`|Izm
z(f>`WG;i3sKXmV<@cWm55#7TL-ofZ6Sk3Fu?6xeyalZ{Ui(`#1w<EA=0uw|ZH+Pqw
z>ly}ImKZ_+WrP;4btnmdYSudfmerf)EO+G`{?N7IPd}n6k(46{j`H@=yC?kpYLY=R
z6L!}39)4}GdT7G<n6VDF0pKcRXZ9{|%yd7}qQq$Z?d|liJ11I%A4jDzhc+}WNW9jv
z0)Usl!O4Kn&jz{ZgaN*7%|a7?rCFMb2K2!ZlXIp=7l0(UT!5-8Kx%>kP)IwQngYb#
zcVAgbiz`#$e|vj<stle7KIjg#iaq{AN4!SAT1r5Lk~-%)4@zEo&cQE3?`OExe2>-n
zj@+(zBBAQBDU@P%DAM;%^pooJD>(xG?;nx(_e=a@iPSMhBo<I*2?({8EZ*T-W`6uX
zS#FM$0yJ^P!L0b6LzU?4U){G5xBrL#UkOqou;Kc(GLh9%$psL_5v>Tg*$mJO0h_#b
zvwvGao&WdN;(z*CbpL@j=6~U0xb%Rh7aNUP7tB5~fH<OoY+Y~-bTp6oX_E-!z{<k(
z2`8pzqi)B#^EiNc^|9&~iq=0*x6J6Z8(|*9@$|Fh1>iwoT)V@^!7%{)L$`y290OK7
zbNT~+=rjreQTYH(@vp%lhl784{V%_EhmZL`{|s27|Gk!E)CNReQ#p>Oma#1Tq01V9
z<AHBHdegPEeU~mBNmlAIJu_w=It;jm@df>%%L9AoRbHh`6M^1x6&&+ovj4tOTEe0@
zTpD!afoR-Kk<!FpsJXufiof=(IJPVQEf)P-G@}c((uI}%+g|^Bu)HH+Wd^zdV5)zK
zjER2_77rQ6EP>4Ve;Bs@^J@O()cCiM{-0~}pKBxhpL^>6#CrdG#+$!`WGB9V3%Yp-
lsySKf-EKqehQ7tI?w3i}Puc+6(uFH8N?9JN4Kn_j_%D@$`o;hN

literal 0
HcmV?d00001

diff --git a/_images/bonsai-predictposeidentities.jpg b/_images/bonsai-predictposeidentities.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..8582fd70727fbfff46555a0c463506627fa05ab4
GIT binary patch
literal 85845
zcmeFZXIN9+w=Wt5ML?v3fPm5k=}iHNh;#t~r9+e=orts`fhb595D*ZMCLJkKBSj*;
ziAaY~LPGBeB?JiJ=H2_8{r>m4XWzZ=zW2lV@K3UyXRWL?%b1yC%-<N}H*-F9z5rm<
z)6&%fP*6|+AmlgTd<LKapuBM5uMhb}MgCm8eDNX`)kRur>Pwfe&|bMhM@vUXf0db$
z{wmW|Iyy#nMkW?kHa4~^3>??kS+6m(va$YE2?Zs&4An)Nix+8F>FMZM|6jk(e*%~;
zUwBWoOi95HxWGg~$wYD93E%?&C@zwF`&Wnm$BW_uxsR8qFVoP{ksqjM1YDq?q`W{y
z`B$&WPY09l1E`oTGT)R@zr<qvoSNU8RrYOS-erLY)opAh!&pH%`&XefwCo($IJty`
zMMQ6j$tx%--MOo*@laDsTSr&V^vP2*a|=r=2S=wD&MvNQKE8hb0g%AgVei5tBBP>X
zl9E5Bq^5mJ&-jvGP*_x4Qd(A1TUXxz|K8Zt-qG3B-P7CGKQcNtJ~25pjYMM>mzGyn
z*VZ?1d;156_#?vc$zO6&04V=etbdd2f02ucEY}4pDoQHqzvQC05I{aCnW!$_l)1#L
zZcP2$n}uKY?Pb;niFws+Gy-xaST_4t!?f&z@@OI4U!whsWdAk6LjPNm{hMHa%Y_77
zp`;-Hc$7>4AmDTtL*2<hP!DM%y~Xd=2j=X}49vj6ry_UfPRwtf1L!Hv0gw5JpU(l?
zz9D~r`{YY_!dWp$YKmy_qZMZgffA@Y8;*v%*g6MqH-XbwWmUfP>&-K3IfSsXAHRMj
zD`6W}I8*{e=K?n<&jHCfXO#&A7lxBJ=YXZvbAWr*jW2KR!AM+`q+H<f-E#nP?Hmw6
zzFERRGKArqjt4XeS%sTT%MB@DFeHtGc|%5ACyJ5#)(Gb-4VKW~w7A~i-Xl#L$|AEt
zD!hn)-(OhAeb^ZFIY0z>Y8!Zw_4464;2)dL|A$sLdqr~@K>x1kf9!HgXbj+=rT^sz
z`@fA^05+kUfuIM~a4kFsI0mN;p-xeCyEFBiO&*-<@jo_41mbnGQl$$@sCMj?SEo+Q
zfAi4*m-1Xe_{npC=DlEvqH{o1&e<5r^Zu1n`k?~Q(eyb0<MQ{ZTVyBj`V>?SYgmoR
zGT|;gm=mM)@ojM7)r?~L@l)^ut5d=~*28s+eWLxJ1p>keyVL<ij?aJeb8NCuucP^q
zVlV48l+k|oy5y{9T(rO&!v^}4b3kS6p|%m8J4XkxTYC=Z9RB-s0^EV%ogzNMmaE+u
z^{m|Xc=7#v=*`zL2{$4F-t}InrH^;GU3^!pUx)hbh@7X1(S14HdN%({;3=<Q{&PU<
z&^drb`ybA2Cem<W>f>tk+Bo^b<>H7LEl=-M$lLGnbM~^IzsOx^Q(s^idhOhfh&4m7
zu~=Ii@Hlk9o9_KmoHVoQh3{+VgW1{{pJhC~(Id<E<HM`Vn5vYyo-))Pw}BuEG)m&(
z`Nzio@!ZPyNmPH*1jHeb(*c=V-SHVmFcsH2C8<^dH#1=ZmSX^8P1gVJL1APu2CVUo
z-Q1NLJiX0qu)NCMP<m&`1JhFbk7_kcbT|j_$2Xn>?Bvb?YYkP=M0FgOfk>ALB&4Gq
zIT=;;r8fTL#i-pN^PQzYCUv>lhq}Td3zW#W2aN{_M0x77`>;Kqb3py@zf*0$1rig^
zHZ4wG)K8y6-dm-QrQ>C$P=v*7L1NPqIwj?8n?O9x^@+iI?_Qq+{^MGfSWPI&lkh*Z
zO?2Z>eb+)%M=@LK>eKHH*&7qxY}`BSJelpajBi#RQU-}0H!UzLFvap$o4k9U;^7$N
zt^J9&?ca6P_kXcM6r1M&>>s{EvLWpoB28-lL+>U(7~lU-Y}h|b|8vy-herH!w)~Au
z{;6aCf^hy5>%eCU(eC8Kek8meeR2w%hMt`RnotO@NvF-X+3D-^F+Bv`bAaHH-Z`KF
zaiXx6dsPEZMq-~B&RSsG%Ex6LMomU>Tyy)@j~Yx|*rf~m-XDAN&)TNmpQ8ix>ugXm
zSe=;vLxDY7<yn|h&kw|<oHZyJghi~J1Ky>op93zR9TVO`^(0L@LC=!=xxmOU&(jxk
zp;9#~a+ipyLj%bfg4kIUHmMhg=0gA(6QVWyA+cM`ff@#<k3A6{Gv74vHr>xqNz=!#
zd_eb&9|q3^hI~d46EMUI*`q2lsH7&I11_$@GQ5Z~&V14Gx-Hh|d#2`p^2b{A_S!QP
zx(=QV1hd>9(cqGCbh?-(rvCH!sDL3KU7veC!J5R3phvld&_o0qM?_38yU}~p7;<&|
zu2&b3El?SgoY)(X^o~~x`x5>|AAWd*-+GcG2OdPUl96Na07)1NjH)-fvSv1+tXRu-
zSZDQaQpFaXWI(@jttizI8Sn^q{G8=x{#2mN{y|a9#9+P!x5|r}+C70i&umJ%hnZSq
zG9PFer4>KOynot%tAY(twRzaQ(f7Lf8~~bJcN)IPYOM2Ill3vhf8n5hj!h=%v6GRG
zC+xs<as&xJIX(x>8Lf=(8?_r9i&~kV)t>`~jX?UYh)wdK`2jN@Q`w1Qmy6%4N|n1P
ztqE3--(IQt{($1rUB>(?d@mGmMd@gUD5TNP&77zjh^a{nA?M4c8D>(K{TJMv9iIxD
zmZ+u?>^|M8OrvttxyGwr^<A`d+3qN-YA*m(cHqf22q*W3?M3VB;NIh$9vI!h5cN9n
zavhL`;I`WhyV9akU{LI)Y+ISASQ&9_57)PjF8`v->qH@W|K>G;{&Mf52;iF#HaF$s
z*j6TBjH*n4U!kjFt8nwTFf+L8IiLm;bgC==y-4CivLJ0UTdSjt_r7-GVK|#<%S`ed
z3WvZ{6+7iI*_2A6VBnsrfYM_e8WzSYINEw)Q&}or^=2t#^vPr^_aH~)N{(id)}O-A
znE<@0^@sU~ar|v!`fo;cIj?AF(@qA6wD+~B@I{qpwsGWOE1%>1lYPEIf(%Bk_XHOe
zg;PBesE-4V@arRP_p!Rnv=npeUNoBi__Pr&B8evo4W0ui5w?etCM!+BfrvfFG_<G6
zYr%wA1u!)P-qFhT0|)BJp>b0*`v3{u=xIk@$?#1bwQ^0vGa{Sm%xEX@n<1sI>>X_r
zy9etoh$GHEk+?SJnQ_^r#qO^?$6FH@^eNGxs~|f8SdmrV%SmC2sn5z$7vBk0XJK{Z
zx`J<MKDswx{RKYTXuhy~4tT59KPnYyeKLOz$U<P*=H}^k347nk@Ys_W=WG^tL=3cn
zO)8SKYD@pSNb#?76bXxb$E<!sC%7Ck=C6k`eI)=bCrEN`0jaN7>e0=>E$%xZ<3%$w
zvW6JU+D8d*s(dj{Lk{Xy&JFQh*tK&&Z+<0g^8GpBMB?jmhBE48OJ!{mwr_Nxn!1Yi
z&nYIH{!IO_p5(^$BGb9dCiCl<5-mK*z*B5MfGI(yVE?-B5_7KSwUrQFf|vC@_jbM)
zSW;dQp=m^yJnXTtoc#LZ(s~t5%O55F1im+XYl){I-suXaP+xj>P@3eW`gQ2=9B?ma
znhv&kat?@8vnTN7uGhMP=p~_2=z%#Un@}0W#`3{peJY<Bg}N{Yn>%zbD+C@qWy<kF
zEzFpQuo40~C$hz`8yn9n>NDkD`lVM8#l5{IE#ul)Hl$wKH=4bQZ?&rKuj+quJ(x0g
z%}njuut!~%EJLj;-uO17#*Z9((Vv|5x{G7uCyl3r$0%|y-Y_3koxzNT{v3;)q91V?
zb?nj@aD*0;rZd}`>dW`k^IPB#Nct`hADEQb+JvV(X_Yxr=LGB!HaqM@GCP|_^x~83
zG+df^X4TlfkL^^vdCMB*m4vyo&GX`F&mfXU1<v&I#6AX!M4ikbCU?Y4R){ceSmFY&
zG<LR{XkR-nabS7)!_448@5<4G9BHkxn7tV{Z^d4ITFwlb0Iz3@Ue}#bYhPkd%LEU<
z8*8%NU^UhR{GW6L^N`J>?#{oONB3cEvPl$a`d5=!r3nalZR4siZ2oxj+h^(C4RPx&
z#dE-ZJMeVD%rMp4`{s4NTZm6p@vs*tt{Jxm&?PU`yEuVPUPWAYK_f0ZJs86sr|ao4
z`2M#YW_-5CwZX69NwS;D&gqDk;E=4I!2x^`wm12WG>}G~R}9mB&NDgB;pc#CrgH%5
zLc+J1bHLE$bHMIf5ZyGPqjR2xpoCR1ZJk_h!X8D{HP-VnZe~m4R`T-igMYDOrumF@
zR$LgXKAQ>cjcC~<*$~e>;c>f_L>_Eos|aLaEgLY#VIDNZ?5cdYlSQ>BxF+8#G&U_E
zO#4h01NvYiulEYwz3s%U#QPgHIaeI4@~(oz`3z<o#P<jS>(1uJJ?%*;I8OWC_@g#3
z+3=Ib^@-#_C8m}L>6rFMZ?46iS`;l5#08seYX9*dO_mxs6<ZxAcC5}RKg@QFukOyw
zr1Vh{dpnj2c;${h2LK50&euwj-q>@%lks}af;vJ%aAAk|fH=ygtMq}VV3HDN#(kUZ
z`GTy~(;lT3Kc$WYW+LTyA^G_(2381MzZQWV->LQv#JErAeuHg`J{<Wm4V@WSPIHy$
z%zM7K_(7A2qF*1qzl`{80Mxf6@%q2Aseyk^BVE`g%!R1wp3(c9>O<vfTgPsE&om32
zF=|aSXdW-XRu(lmwczyZ6YU5y!C6loFY2c&hAK~%(}}+TMDYupwSh#SQUqtrP%Ixk
zF%Z2*bpJ9%RGuK3V#M`h-GgZ@4|OF#Cf_X|1`I;%2SQ)2%zfi-JT8NGCvYf6VqnpX
zc90GW>~4JBC(lWAPo#&RWkvp2&$NK!UU}lgz573gQ*W_9Ot#NG4O@|`(cN_<0~o?)
zJs#otBn=_j<A~fFo_0TO4tX8yhv$YJqxsITqNnr-A~u0UVB3@GOyHU&1mKfKFHXk3
z7z2Cx#YrMj?I%VtI|ti(eLVb@O!BE(4PzD8Y6pbcXnTw9^GUyV=)eJc<e*O9N4)AQ
z7<6#<gs~a*B0EBd@u?&7eiJe#gVo=NPE#niw-NQ`m!m6VDV=wMTnq`kA9D_vNt@jc
zmMq@U?86%kJwdG#s#Jml&{;5bTuRpXd^~R68aY3^J_`KM*MONfi*W1^6>-m()TZX7
zig&^OM&uIue{bw+Lv|&$!Ct=WIlYcE-e7SJ+2<cw>W9+=!rq5!E9}{&S53RRJ2_*>
z@f16V61rkCmgtB~t#B_l$j=7}m!TX(-KVmZ!a1#$-!O~%$30oy$=24ytrVg`l!WJX
z+a_}%JoAvRT~&eQgygQY*6YzLT(M{GZQNwk1xswL{Ao`gzR&kzzED?T0xmOJEjnvS
zvsFpEQ_`>Nn+*DEmF}(T(b&!TTOBRAQW3f90X~-fxIq|gn_w#YONGG^ZTTY6Wfyxh
zZ_bw;zr^dPUg=!!gf^pr1RatH2+BW9)L!IyV&SH&6zx@z*50&zRyN<kHkUCiXLIa7
z`p6+(J-VSPJ#^YJ5q@l!iP?m=U{_<oP~*}W$fzCG{Oy$fS-m$4$ZL|yJp36m)${5P
zYU4lmr*Nx^o9_dQuO0cH&Zd!&yEDIZMGOW5LuB<2)$lcw2RtSx;?~p&kDcA4eoZ`<
z$K(1aqgG;c6CZi^R{9@%wMCl)c;ez<e+vh?4B@bxfU=ZBGPMc8U){JOqqc`S^{ck-
z9tOo?z>7Z`6xAGMdkL@N%X1TQe!p1i{q-97?i|3Rs(}F$bTFf;-ucs0Jqw$S3fyrb
zU0gY_yYkx_ON?{6tP{qumJYvWDm1>{UFX~B<LsG_N!sK-2dsn0n7zqvTy(F0^3)LM
zn{);~ItFbG{oN_;DbI7j>O0uc&#FwA!w`pX{a47OUYI4AS=~O>5LuMDsGq>Pwk=+s
zmr9qJvY+?q9lcQNF%zYhI(K|X`q$5aJnxemssF8#{C_j!PY#}33H!Gz<Nv8sVy{In
zF#PYDFI~&MtkwZO2dD~=BYYS6{x;XeU*6pKpPI-T9DyQ1w9f%h+Uyv>|Gs`oZYy{U
zU+t&GCqlg{_xajg<t-+3cLJ^uONq$B$<}T7DH)5a%MUV~1rME)zk8#gHpF)XGu576
zB<*gv>4{zd*7lsC#)k2bb3pS+M$$J0sVH8K;a&Qyb3n9wEO%>;b7O1FUFF0heabN=
z!sAP*e<|s5stTJsGZ6amOX)+wAHC%EMR7q7PNP3Dsc!`S^%kFV!YQJ(R5bUpV_l5S
zG^cF2w3d-x)mg0TS<g8DQAJXC9RtY!`%)fekA!)X*C}#9E6(oA8_{#XKYYPA)sQm_
zJ0&BPMe~weTHxQ8+IWfLBR&!dK^Wf<y^!#4SHpi-{m+*E9<KkV(W49_MAKRmOgoLc
zFKdpjcIyK7dXE$qP}o|j7H!-NhM@yYP2|UV6^n%BRc9<do~TcLj(_WSIpu*iZ;smK
zrR(3;7@!PL0vQGI4@CpF_<*Enh5e%NQ`rq=a>B<Mm(xWOt)G-X2heV&Ha&W+xf6>C
zs!-bYWP_*O@vUu)>oo%~$U7v%OA9Jjy;5Nzsp5vZO0793CJ3|-2o2h%$3}Le=qrdu
zm>DS=qI@%HxM_H&ZkKUxy0bLUqwPiftsMJb`LDA2sjtXhe9f~OOQKtE>dq>#-3+0u
zANLc*q;_|^Hw$L33vot&d7t(Bdhkf@YJ+<pNg!mP+7-rxaCV19lCIV^V1c2BnKRp3
zqDxUiB@nGXos^^gmB&}njWrZ%^-F4Z&QeH1#Zj$FE+gx_BDjD^wVUo5CNTyK=6|+t
zFNc3(i`d@XUTF%i!0pvhbYIwhD!(Y{l~dHO*6?G-dL|XQlO~lzq-%U_*C$>f6|_FN
z$rn8@Nie|18YF>PZ3uD?THIZ|ByQISZr@&No}!t#mVB}1<sIr|8||MtvAkS_>A$eg
z208E>1rfhb4uj<DeIG;ANGj+Qj0@~7lQr4L6YsQ-IhGyqGALMd-L!M9BC-);v8#(5
zoggL)$R%*h2{;DI?VRU23Ui+pJc0hPj<Vy78|}REs=?7Y`l^u%#hpqybmUe7&ait3
zZI=Tp%%SfqBHo#VD#1!64w!kAPc|&yU)Q|lj2?$;f4HY!u*-^^BwZ*i9_;4YE`w*x
z8(VvZG&jN0<XczwPQGknoZBuRBrv0<H+iWis3SdvevEC`4SeNXfyb-v%;&XgB*IsI
z&fXt{484Q7B``o8apV|?ssa{{j*L)Po*>H33!Z|a6m$;u-Po6NgIrDKM>d&5Sr-Ge
zR<A9GP!ol4@hRBahgjiyEKuu@LNHx>8d8=pvT0JCs+IjPNm_!6U)TSUg-SZtbuX`L
zSm-G}@GO}acn+WwN4Lx?V~Y@ITZRr0LuLouya^UP(Qhf0dbs>Jq1Y~PkV#i=|Cv)D
z=Y>-3qe`WgpzUC>kTPX(+1jDhY4Z#SQ4Z9W@E>hOsY3+_^$WY)#hv0H4uVn#>B2aX
zy5ua<nuDe7-aEH})>|X%U!qu@R#zX2|B2Zp0xO7uIIp+Cm1_Ycx(Q!>#_nB{(UrPv
zB#KhxS0c@mpG!;e;@v=(s%d5#(iP+yjFt%WI#O&EufrJ@7ucb6_8?K+O9O0Ozjg)%
zTGi*|m_Bj}lwevkPgMxb81tA`MM#es;c){OFuFy>DIZDT_Y9f%z5gVeCdu^G5&OKa
zbklY_`-{HE*UFNko81l00Z&<X*Oq<L44c@%fuI-U(3iONGLP3m5T7AI+lruzyU=}d
z)P6*w!wF?25QAr3S*&ce!EL}6mp+=GBA;K13rk&x3Qe|%V{9GTO$k!no_Ut~f!^RR
zvsN}K)7*E=Wp<~VT{#7?3j6d&TSGV)h9fjY7V^?cKi3tB4|i}X*szl#^{d@z%I0K3
za>rIkGYK_nLfX>PBXHeHW~jF$3doTqP6WTwh_2#TN2aDN=p(Ko2dmue_?UD?I*o6C
zi8QP?y{$Lpl;rAw^xpfN5tD}Zc{w~iGcj^zGCDM}lU2p*wLwZtIQ{``n4j;#Zeotw
zcUhoJ*nmzL`!qxwy`2*+@ow09)7-4OJ6`t;Imq9AeZl#ST1q-re-Ykc6oa_pH^f7Z
z@Pw;!pwE=-nubkCCp*&>&KBtp9*H={-A{gbGY~J_VoKJYBb+vxw-Kw<HQ7o5<CKaZ
zasVU!80cr*{Pleb=jBIq@ln4<Z(rrKbP(e6zxPr+5nHvDj7u;g12w-%?BQ-?J(z)T
zWQMy!VjDiZC$lp#zirPr=qX5Y%G9ONYVW_KerL+tRze%ukGhJG3KrVjnZwskx_{Cl
z{pLXFAD_n6FYj8ZwKk|B8*%1s&A`j*n1MKlpWh%>7$aexSNfSkHA#jcnxC`6TTOr7
zT{tnP$BE;k2=v&=pZ;pnfkZis!PWhzvD)#A9aYbLsdhtk!7}D<x2_PEpRBSiv<i|n
z$OmH-`8qNgeb4ETCjeuXg^|#IgcGhU7`ok1&0IUz$5S9Z^YC|pS2UN&6!l#}V;m4~
zMMy8RA-ZI0V@tl*kGBNoL1MLQBTh!-g4{j6e|PE7D~tQ=ee)5f+Wgg|Cgg0g7<MLR
z+sd)(o-#6Odk(;#!q#PWaZ&<EwHqPBAevT|b>)aEjJ4DyJJ;w6>ot5>hjMt~S>Jxz
z%cPZC%xrKZLPa!CI!DnRLH7HiR4<llV+J<YNW2|+Yo=nYdBq#4M-Q&WTw)0PWbA28
z#n?RiEUc;36o=Rv#Tj)X`ML&9!IzBKCK_=Jx;VqE!oiZM*x1-R57M<Bt_URta$DWM
zdX?syFpxK-AIPQ(%Ol$SXc!h578!4W6+_@gBa#YqGK$5nR|M}hKHI;abm`gsPx4Or
zoEHO2i)QTh8c^b80f%t;Q>esiQ#*V%Ohq~M;72}UUt|+6`L-M?jB}3M_1k7XYRl*j
zcWB<W%FdA}z)zfNS^r9*c}lhr?`Z=<{ea;>=G>;tc8-v%xF?F`%oqb^ef_TRsNSm%
z5_KO`CVv^7-i2g#Tfw?*!JXL`aOP183@onfRsneP7^rc+rH#7;<=3wXTLVVIH2gE~
zl60>(T&OAAyGMz)B?RDD+H+V{L8XKQ&Bkxo!rz(B?hZkil7nbfZEyK!7iw+RbIZSp
z&T{;=KL-TxY-*O5pTcsEC9`zXnz_j_;<RH#^8|_<sBfo1SDkRwRS_@5lCfnziP4B2
z_UA_-Y=IR-H1IOkdBI1#k<E1u$O|H;41S6t2pK>$6?XEB4rxxJ&?G0-Cze?$BFO=N
zkqE?^-3S6pD@77cAgHS<oC6-<$)MwnF$|~Bs0n!fpSE@X8Vnl>aVLdui<4K<FMkmP
z$jc~82k1b`6-2nmLG}@?KR&+n&kp@_g#N?%`p-G^pK0{}c_VbaDog@XKkO87ZRDB*
z;>W{y1=_MN8;=2$xff!R+hKe3WM;{?xFb@^fF{6o)-8=oFM9i+xxxm%gXC6<c6_WA
z5OO%I(@R+ySm}<Qt-ucb%wTwTfS`tC2dik+#&^$K55(Oqa@3@yOZooTn({N?-q&*X
zgc_vH5Ii6?9KXu>Q-o~@)*gHgFa#cxV;0r<?#HAY>=r$YU{ud{!c4Ygr)OmI)%S$_
zP{}#qPhu50h`Y%^(ko9O7$Mflp0fZ9FaM7<<o>hlKga6-pShthIz1a)HbR5n-e&@N
z`MiIRc|&Gdq1DD<YdU0J)9xX8@f9*Y2hia2$t$iFnH~u%2Of*nm*if_`OoHWg_|g)
z2#IA|oAfRYMmh&r!1kW<VRPJo2ZHYBfTNR+b3oEps~A8Zh{<jfPIG;pXRI#j%f(-l
zBi6FjYpj=xe6n?yD7rk>V4FLzztYZEP#egiKj(nSSu*V)XRw6Q{<~2eimOHH6bMxs
zhk9a8hiY6E={mg^27PXwe&{TS``wc}R4NG2*>Oy?kIS0k7afqh4xsh{eMDeQq*&eg
zI73uixNfuW%|R8gaaBnAS%KjCiF`>Vujg&ZW>(5|Q9j6+q26bk9E~cPh-^8gEBEd#
zji@#+a%|h;67a<>-&6NNEd-whPMb`lt1hP3iZdj;G0!7DB!-3UEjixUw7bX92BQ&#
zq!}fx7#3N2InElDc2Jy{j{<9kel~*=P|O8?{OI*;<qYd;f=(mhGpjmfYnd!BZ(Q%a
zd*cnoVxeouL{9BE*{xWVknPbKmsF6DGO1T$j2$e(@K@BQ#JsuYd*hADd+%hlVa4l2
zJiKNP=+d@4ElkH)GkH<D!fwwTf%<Zw{hLJDh#UI(L)Q-!Zg+RQMdoGPiejeViaT8b
z-Z8wU;?T*%zikC!2qPP!nT%2Vr}`FbRg4EecCw3}mq}<RBa#vHP@2B0RrgLG_lIQo
zs}B~uK*i0RaiVb<*`b9+gxrF7MAegB>-iq<HIX-2H4h0J<>gPvDfMkDw71F1<IE=K
z^iCUv2#Y{_toQOqKUa@QTV;jT$lZFj6f}^<f-u^SREqL#RNa;{tBD{@80~h6QEL#s
z(vDYo5TT-9VgUbN4s@#!thW|pbV^O;R(>%E{GEsM>4DG7;<{7NU`iLR)=aeiCr{_b
zu@)nRl)-@?1xnfj;T}$g?9cLj^96nC%=4jSfSTysnS*B_=bs;D?78zz&yRMZSV*Go
zA-59jB04O_w(YtlfSpGZX7JSYVykwNXMKIY0BP?o<8f&!<@HK2kesLF&~6ia!zM6f
zc-{rZ4dp9^aJ09IjVUzq(N95=`o(819}ax4aP;#vM^b*9{VDx>_V9qej0hr}Ex<xU
zc*g4)m?6p9HLW$@l1M!Ko2H@jJNq;O{N!NA?U8F)=9EbVS0h^B<a|dvh{F}ZBrpNx
z=8L9|6vqBi<1yfCi`00BJT~cXvsJj0p#XPEF+UP{xz}ynJy!>A1rRcN+UMm)w(1(C
zoN~|-#qa9YE5QrdmS6$vg+H7zT{(x}La;qtit0bxE)=_#>T!X=#0*r@>6zrZrccQr
z3Qe>zJ*^~1sJKsFuZF296CA9ERfSJhqoD!?n<Sw(M#bx9Y61~kwx#wQTRQ_RMx)ej
z&lK<fmVM2mjmZf;2)5hRL*PQfoXRZ2l@yh@J~Itnei8koiG8+F?&l?KVV549B$T5Y
zMN4!%fpwzjtqAJ%i;Jf!7M@TI{J@h)ZO59J8PDE8cH3`#+(w*lg$8s&)|cjS#ce>w
z%~Tj$E~zWQg!cvx6e;!0bt&vj@%NbP7xZWLpL9z$(74c-{$bypwOyao48t<@vCRRS
zp7LYuDYNE^_;Y~N5Xi8&Y8}Si%Hy8wD&sn@hHcH0tP*c47OJfE6tL@kN3HXqai8;*
zEQg?T#&9hm4Nb4z&7sCphck=z!+d=rYn0bf(tb273P^=ef{i0glT;F{a3BsS7mofA
z)umDa)QH`-LcNFHEt`6j)SjSFkvIBi#OpTgmbA@tYpR89>X*MlxCsH6Grq-&@a|0X
zoT&k}&Z@hx6vgvm^$}v^&+l3mwT#6xHO}QEP031jAqLs^afK=J1!;;oq+g&F<e}NY
zgx;_xa1%t}%Sqf05hmm<pe}AOPC+8On%KjFWui1aOE(_R=-*7ta-gmJnYD8F!eXfw
zWnQr>f<=|D!phy;3Knyqa%#(jlrVGm)Z%1r8)F_&5ocloT#x9McJFPKh2ET0eU6ce
z4Ztf4D)m?sUaH<J{I&Oh%ev-h>(7N>w3I6|NrT~cuM>+|iB8zsu2dvET1pj&j#I|Z
z5pSW-)Fwt&R4nZSx1}nD(gh;4n2v|L`0FHu1(*Y2ij)rb0>P6cJ|bezslL8cQzU<z
z%rYYr?)6{<%HyC*DY&qoR=W{9P|YF&sv{#Ia=Xw%WW2)9J<!uWG5x7=p}APO<wt;_
z?lYAMrFW#ekf6@V`CC#YuBld%$)y(S_}2q&#`O&aUKJ8MH(gNS$AkS}-VC<9Q9rH*
zsa3VXqO!$WarnHiF7r29xkd@b&&b|mQ!t~QUaD1jvaNV!P6J~!%O&nuIm-9Rd-3_^
z1Od#p_~LF9No@VZqiJzw6JKLN2*K=7TpUkcpJ;$9nW3YYY`c3;t+Z_WxVsgj19<q;
zlzx!2__ZYP$l#{Evi<;W7)am^(f(81W5>-aLT=E55UIz@-@Ut)(9nf%@tn&M@h5s7
zH~FqK8U!sCM|wb?qUR{J$K(-AZd~z=XVJkI;|sY5ij)!qB##|zJ9#EvY1cRGaevNZ
z!$=bua0hD2-$$gpRG-6g2Nz8-QB@vQb$8^W7M1MR>1t=x@>YtdUMTe6dHm&}R{J5B
zR2$hxo~jD3Er@llCuEirVs>;^BE_BVjJ;E$UOA3t`Kh?q@v@iOkYBl5HpB|Hpd2E#
zg9y57=~Fm@KP`4XDmlYXR()R&QO3?UR@%8V1w{FNJ$o?_fUMiJo+|hr=1M!WXm<O%
zAZ213e;EXZNeEN5EODFQ5{a#ocyX#TgpX^;Db-MIun^?CYaE?$(E>WP7A(HX7mA>Q
z2BUKrA=LV~>)Db&>ah+sO{h84TH@NiocisXx$OAv6Ogois&yFeXsVQe*iPC{1U1pp
zQxX2f1|nwyqN;kgE3b0^V|BCgc+d6Vtw-JVgFoE$bz+rtwou~32ccKR`nFDhtwTFA
zo_0T!qs0+Vk2gsHQzyW-9Hy3AINslw-S!ZX;TB_eut>7}iDVudQd3y2Iz#apH`R4s
zcg4*`^Ns-?y<NlaxYY>HL%KUf*((#a_H$05cPx{feWg;KI6B#Dv6fO_Q}`M3o5X>O
zkHko^%|RkM%<D#kL;F!inO*LMx)nP1%)6GtqtUuI?)M7pTNY6DC5Orlx$;pSmhiPF
zc%x@L@d#Ew7|XnzDpjm5tWGsHc^~=<I^#i+21>i(dFNHks<Dv)kUec<uBBonZOU^8
zam;=Wkk8-PT?*<>SWLJI3FVQr74PJ-D`fvlm1Y$*J>MfX{~-Kf(!-|-pJHq&oO-#_
z?r8=_#v$BxFA;$gl4NFm)on6lyq;h<0<NeZ2!6mGv$=hgY<<JvSBWu|u-VJirTP2E
zcQFSHk#@>hzSue`Ms@V9`jQGiFnha)r_{DXlBge#mP)L3sAHme{|u8PYdQy{)8iF2
zcOmg2A-oWJZ!|Z#LrN{*0S&y2y!W1CqHwus^w3qZvSK*nsX@E!_ZLoDw=eB{|1|v_
z{-p_#zk--eIz}WG-?lHxY_^6i=z>-%&lL96;Ml6D<jxGgE~HVBjW*9*S5sdhD7f;r
zw)l<NQx1<DsUhC0-LuV^oOeU65t<i3-0sY-YCO2ANWU|aJDMF+>#xf><jS?Fbd0Zm
zGi?m<a7f&=5org#^O;&dQcjD?o$$>v<73*i$|}riZT7xGPd|7t`p`La9>#BPJchYX
z_<g5J@GKK5Q%UGSSN*goC$fmVJE)@TRtO|&N_6F<h16IK(OA2n6AL8-X2ttf1vQSG
zcR!o#&ROFSAaTzA`NJk9unK5@Wa#(SNvt*@4sFCpNH)Rlq+o)%O>vjf&jEbMr`dU|
zaI==hqiS(>LmfV89@lIx0-1Kn<<^v)h9<Ko*D`hr3mXPb?<pBaVBEF(Z1kc!^LhnV
zC_lKjw{jC@aN)5~QS9>3yZ9IcLlGI_b`A!5{Bh^55P1{peLYUmIWW}~y0+M+5x08j
zo2ocgyC=Y`RS4UanqEfG?x=x~^M3cgWp={3oc7lznYS(MM=Q5|d_6g?SZ`bSH1M-Z
z-%(a;Fj{p)^`Bb945W;8cecxpZ8tZaD)2<3VpWC3Dtf|jWn?5#qtLx*Jam>`5igz~
z_<2FCTSd2RbImOqfirCjVa2h<zpWr3)Q1mewh@jahxCS*B`cM9zAkreaqdB`Q46=Y
zF917ou1!JJzfBO08Q$L@_59(}ifhpb$_q}1_tCL7#2UcOA(0<w`&7F7FJAy=#q6B}
z>fjx)=n(k?j8`Ptl$xC~M0qq=4xl^#<kpT^Z$<Me=z@KG9k0HRdG?sf{>F7iy|E;d
z7;E6Z`!d<XA9{6*w0ATL9^I`cwrUR_!B~hq#e{4I`Z-9gMgxv@KR8>0-&2pRImtFw
zxa(cR-V6^n!1-!%9@_x%$ysj8E$r`yqt+I~hYW1>D>+`OihdSfT4iSEI>0N(Ah~U0
zva&UY!8tj5_p4Zl95^0g4)achuV^;ex55QP)T_|Fy-g?B(BNROpX2ROm9{jCwg`PW
zqdqH_j7f+f-AU*;=?SfXQVTbLReh+V$x186fp_)5_n5_OrJ8><z(;)niG$PX$~D{%
z^ll`7WZ9`&-vNzcU<8I)(5%VEiT^NSSSqLQOenB2IIZ*KMas(h0Dd4Sn%G$7ZwpVg
z=Zv9`{H4AE-EH(|xz+(*j2Z>HNYO$vG4Y8Ds4JV~q!e;v=<)ZK<`F-OO8R%HRbH)6
zqItLEf^!eWv{NB>7pWH!?Fc5Q4(@5xaL$(E9+_=!CASgr)Y#p}zaf@FJZp(hOe2Mw
zJBV4X4HAY%#H-zMd+qtzjxr7XqdwuSAL5E+;l0C;v226=R)YL&@Ko(R>W|A7;a*pb
zugcM^hxDzJ`KmK}2^FL>RtLz}T|_Esg|yqeW}t*|4m->OK0%BU&GS$e>pD(jk`?vW
z9AgFn_Es;<_@ysY4>;`P=q*E^li6R2Sn#`DsHRfvfv5JPYI0UH7Cu_9;u<pKE&uhJ
zF3krjv1`3QKRnRd4r^!;V+chsLN(Daj-o<`MH&SQ;>A+Pi*`Q74dn~gM)J5^3(rxb
zw{)6KLNh`g-A@=*5$Q1pIn1qB2w(KKU#xbJxH#Bj^B^Xox3`KL-p!Kv%_BC07pN0i
zWjpF_y($R+aEHZ7unASBf&AvkoUfySiu$PO?rpM%y#?DdIMF<w8QulDR$U;#x-vkK
zP<0HL!F3Xuu^~4zt5MM=RHor7{XW{!hf~*D<mkDDxTfrmnw?RRB)nMhn{NuDUO5-e
zq1E}Q74ojySoqh@lm%1swcal15ZRK1k_0k4Pjn}Vx4aeKmfAvokD5k%RajP3_~}or
zgz8df?F(pgUl5q@w|w<OXN|nGkB7tN2X?25?4GcxD!pFb?8>u7X|;9Ym0xB0ev}r8
zcIGfma7`^zbOi<3p6wc?HE&ilJXaAu4yq@topBkB8rTlKP*cWE;ebQ|=T7CvICdqN
zV{n61{i4^_u?l<<JUCjckR#GHM5%$H*H?#eLW2F~WT~54M}C)5BP*V3UfI)UKkLoY
z5&8t6xYX(&xd2PZ!;N<*26->iom~QR<*;f7IR)pn6=-udbLiM+PaiTUNDIhR8n=ft
z^DA=Xk&{Vei{c8rUS2#3RVj0AWlu04s~2A^pIuKoG2ssXlzuSDw~igUopj6ZWwmJF
zM@LTy<^CMS!Q!K71wup1_G-Srp{~30(bpfX%j+;9m<A!}^dUq8qZF}4G<K+5^UvmL
z9<2bISpTqNYks?ujwbs+{o1FUOcDFmzL}$d;=#s3q8Guh(x9NL7;@Tv@~)|+>eA=b
zrop<nAbv)ko3yLHoO*k060_Zx=P2<ZxaHNip1#4RZMzikKdK^M2El;=&irIIyk2nf
z(S=_>dzP4P%!Rf!Y|o|eCS!KP>~6L;6N>$c*Sv$jTH_KT6|8BJER0v`>y%wH47%Rh
zM$o+%fA*OtcZU->>@(_4x;|`vnhBFK?ibq@nb`xCZVi%cvoWC;qS*q<t%~XG{S#jd
zXa!2{9&*Fa=3kR>>QE$*m%PTf8_N(Jo{oS$^P0!Xy_o8C;#M<VWwoqC1?B~#!y@Nx
z^}e?I=g2LSNBlZbj*y8lV(BQ>M4wqr>NeoIvKwxWn5)D82-hc=#{0aQsVz|$j^g-0
zd|SZqK9o|GNQ0Yi{_2hd;cbvtSkxDTEEM`AE0Fwt`X$1(iR^c|9xLwBH?n=~o_$q-
zzBJ2P_@l5mL$~TjL&fr*c&aKfZDbuZFppTu`Q0;&l<G>Lf+QPFfa?-CAuzIg7{Ft)
zGVmnSWlJ&CddTS&|4%a!e`(f}rzkqYeutXUDU2x4%0J>q*1zMpAFFuV19RYukACq6
z@BNjk+p4lupJFGwVN7*A-W5a&k<h|wLSPxe1Z9wcA4Jsccro9keegVG+;dU7Vd7Kw
zbmoreoL&9`FcO&%p(f>~T;}R+;UWvs&>qQm;ZiWFvmS)l-HUmpC|JF)=*11d-X9sT
zg?$~^*L6KxuFSyX?2I*${nwvl{!L#tV(vBA0}+X+2ZceM3JJBH9Hg7r>@YMBw?(ox
ze7qQU`--z~fWcm9*@tLFuc*LyM~&v?b-R=j=tFE#@>v{ymt}1NHe(H|RU0nR<eP!Z
ziXC%;=bZ|p=o|YIsc}~bim_N!w;dIZJrUE5P3ehFGn{NOz4>R$Yny#dvjh~_F!>?9
zD4xGvUaWsa{9@KOS=ubVaMYz1ZFCMN^zK-_gxEZttum_fh^6DoaFeT12}s==?Dl!N
zp=af$<3+7*&NiEp>Am6^Q1ia$Yw?+2x|GJH)Jgw?=2~LQ+)hwHbc+ufaq%pT#LUO&
z`kDdxsRDt|+A<?~U+Xp4QesLR9eZlT34M$uZ_llUwVPw+=sNg3?Hdp?${!W81F!WO
z$wVynu5XBUk?un8;0AlTT**28>6WKKg>GuR<oqfh?ZB7Y)xF6EH|6L}gh9gUVxjgb
zZ3!Q};a)~jYMcbJLV_de?osPa@^&-OSgs|xWpf8#T{MyHS`oKK*}55vCXGM+P#Q|#
z_5Rlrrzfkwj_)Q-oWd4BOZAh%4b3$x%LpTq_FnU8>uzp~E{?A~gg#j>0ff_v&FlNB
zyX?AI>c%*e*|WI(PK805IMBYj%V4$Q3yaUb-b_fu057^_#HZzLa)lJKXBE%bxfxe<
zm)ZTMOLckfS*~}Md+X`kkx$}h1~#lw`w%Db?l4glOycZV>~19Z6+!g&^^;Ee97cCM
zi5>Fhm&}k8Bgg$bc`Qmh!0~T0w$w3dbcBc#aCDkk2p`;f4%9WzHFj`(GlOx+-T5qY
zrKb8uNkU&^d&9^}|AbG@`Ke;X5<8WXB*PTaX6xW<o2cfsQ>oTIl4n&4F|E;^Bvp%@
zukgH@2g_f_(>wVEdn@LZO0*33P^VgHMyD|_M+)oGl_V@J>>hg~&`_xS9(i#?KPQ;l
z+r=qlo_fvD`TNAa3c%fJJ61d55g$Uhf+Ki^2yY>~x$8M#jfDgjm*ag*!kO5P%2Rm4
z>hq&tCQy{G=W$f;Z>f0FrA_5D;_ZYj1U)imTPpky#6f=#dHKk;YNpn!@p-j&I=HXh
zcXjzrADQ|zQFF!&!taF`c|AP#Lt!V#*<&y8h}3w{i~k}HO~;^Od8x5i^T5wB<RuB#
zkv1tKkG0xB7O>Y0aIN~3^)A%Bzn=KBttul#jVMFN$IP+xka&loz^{Q^__XU}6y)n7
zW924GC%X8&Oh2$`y^vXezYaigE%xeVq2~m0fF2drB5?ZfK8d#h=hTWQ5Ev}gESaXr
zPb?~snX&B>40TfZGjKWPu2LO2Fc{*-GX|anZ4OL=4X!LSE?<Nzv|7Wm=RpDE%8>~z
zG6iIZ`ze_;agZrt`j~dY-<)GuTcOvX{*jBdPv~E`x_Gi4wdFp@&LWw0ZA>a;Aodhh
z+Z?i?D|2BBtPGl49q1?K%>?v??ZG=4-Xa(kczn83{f97{XVl%P6StOKIFLN`h2P^x
z!~+GIlkoHUpF%%6o7Xc3?>+b$YCT7~MF5)PuGV5(J(cIG?1!ORQdP@=#*bdotL^bg
zFU9rz%8&`?Jg$9D842{Wb+0-k?|+!g`z$m_PObSNvMLcQ`B+5fE;VHF@%L&n5nBFU
z2p%cABhoN(DC*@HqXfvW$!9M68A`LTWZ$xAcUIBDQUDoQ?5N&gmyk@yu>Pt4`n5vu
zvi)d>>%JX#Xn<00tNz1)V~F1e=cQH3g`Bn_MgqH0=%rPS*hjp4*pR%!O|?hgNGlu8
z)qWb)3F(br(qkpjA9U{uDEL*qBZ-W)s`Ld*MR09)toA7K3fmg$Zd)<%-=p6e+WD0Y
z{**==$qUjlGkH1Vo+1{bGzw~W4UuU+d!4EFloKHWwpIxNVtX2maN_Okt=DRCg7ad-
zH?i9*!=7LJO4=KecYe<$vrkR8bu-*is0Wu>&aGE(NZYn4<_JFUZI0J;$&gi|zIv5*
z@$>4ns*m+#B+4YO5=Un<q8)V`Rzt>u(N{CkmIq-_&#K+`6Z`E77lS-JK<WAdZ*R5z
zUILq9`Pxr^2;jZPpaNy<<{5>*^W96i<M_<RpR?%WONA42gEB1y;`!%Y`I-!>ciK~F
zuXSTzS>w-C&jE$s%TqMz*8QuNvZ~GkW_O`-IQ_P4#o^sU@Kr*y`efY?)o(4WqY9#4
z&l+w9FbU+dJ(3a;^d7VxCi<ZfOfG54zf&+g;-(`K@W7MstSyt7uKEDQb+(vF*3efF
zYK46U;w?mUh$L2|^PY51`8xl)Z&<+)CM=)FYYAx+ICB4pU0+b@QrDB%?F(90_P`gf
zUfCmZ;#Ol?q!zbLRz^D~t$r7r$k;0NaGXe|J&9;vo7MF)eH8vIN~aKWb?>g{)XSKB
zB_UR%5q!NeCu!VbsHmB|hjM%<*OxDjkD=wMv%*B|BI3#h{@bTi-7SUKz~L)3$HteO
zUQq<9X%5YUKRV*9l}Pefd>XiH)^4-ODXTSUlQ&<YYssa-qrd#Cj6_tQ`2$Lan`1|8
zdug$CIV{jegnS%*Ig!R2z5UX~j(yC2cz5?UY5H|)Mi<$sMW>+6ox+Ee^t5RjKd`{{
zJIS2dRzXP9W@q*-_=rLRtt~vxLsi?oaB6-%X+w#<+EoTAo0s{!o~1nfSxr)ULgZ=9
z6NzNlWJ}rxs<&BiDRYLL(lCK9y60%mLk+S1IppwB$YxgtPea(NU*Fu4-5wW~7e1X1
z?YDMktnM{rpty!txLS;J{&^zxZi{GILKx`4^K$+WYiJsuD5df@P5u2RS?-;^|L14>
zVe57*Xjo{raJCVx+v}qA3fq1ZBMdD(=77`ONmCe;kxW;pF<f_I^<SHR_U_o5Q%CTw
zM=I@{2s+_CloCTv-8rz34mG`$R!%6Vr|sVxl;n?m+`3;;nYUl4Q@xyLX3GD~I&3fb
zyXol3fgwt9J0!K)x*)4OgH)fk+>gKtEyMB@cB9nTu@yhwpG-#k#_`jUJOvSmn2X;R
z_~ufdBSIO-8v~7=Ox%{&mkB6ASzB183q%?DiFIc#`RPphFv|Y!&keW@SepE0r%Etg
zYzDHpW*kKxWP!Cty;?-Txr68gz^u!Qd&*CeEV=moRJ=Li{EwUSXs6%5P~`SM%9bWr
zR$7yOolw7%53;N`aL`E5O!U~1T1;~7N*S*kV*KtH&C~z0zjA=9_F2}O&s<i!KU&RQ
z5se^2QX0rr({FPR^woX&=(XJs9Cv#Nd-d3^UlprMlV*0Bm^Eq^w87ZMQxIR)n^Y^!
zR4blG@lD4`prkAbs{bbeZ%ZUIsv@1R%PUzvrRM-Q#cHmdP`=ygUjVF0jMvlBUcWn5
z!xO8*Hb4x}=T$b4q0ThYWgKsMP!Oj19#|(qeWGq`;&zJsrwZs2bMHhv`dxL4NuxpK
zuR?5dbv-N;!K|<wsq9y=^49aLg3QeV#~$>AGE`fA``!xS8Wgs$e73*z#6h4_wMs6q
zmx?(!^AJv&4uUOM!7^X`dGB;b_#o)m{S0P(rj$q2!KOu*A6i%Nbty|$xEC~Qmkc?S
zXvkX1hFZRGvKP2<LC516Z9*G@70QX@{!Na68+9Q$qKO|Zg-G3FdCgOW=^9F@gKzYt
zsg6ukxUhUNC>jNE5J4C0P(N}IM@eF;x`HeN*|xbm2P(uD%8ytyid+_q%YP?FUpz>c
zBrE=fpj}~0$j4_BB1`5VCNLT!jtF_{z^z?L38y9$J{E7PG+mG+-KF!BS!Rkgs}W!>
zwo*fm^w4eYZ#SK$4+ZoullXd4H<tsBe5Vc@2&D_(iH@rL^HP7zyCM^FvkWPqA-1Ru
zCGcs>!TO**QJ0VdBd0qd&h4#y!=4HV7Q$P};<`waF+}Y)v-oq)lx(LM^SRTh_Ogdd
zOUZQ?+}m1wutOMdw5rk7ERM)ScU$3^9`@H$k-D~*Q1>$??it=49Ps*`JE?WUPGj%V
z2Q}Ht^xGhEN?a+HTfb(E#J3J(R=qxO>&M8l4n2`W0Gz0L3G*=4h1U)KyOQpvNyu|c
z@h0!&xgV{%$1~>OG{X!Ptyl5uuuKpZHuUu-VYqY#5ng_nLv6d3(NKZW#x4*3=<2qQ
zP8|tK^Z`r$lzjgDGVROAm7Ae8Y+pglAwrPMj*z=pc(|%Kx<UjrcG#@4xsq%#Q8@ZM
zROJQSgi_o6rRAsDc*Qqlz+H7S5?xH*HC;%sCv|iv@4yRP)$S?}MuvmT8XCuZ2c_4O
zXRiG^QhK0xMvPFumL7WhJJs8I-;mo7`y))`dlDZZvlDbd0m#1A<e6RN?uQ-(oziXu
z1r`R$QFRIZkY;=+2QYik@tyPQHW?)?(EPPwC_e3?ms4;#iRAj)S06iSI^jhOsv;<%
zyAc@h^|cJcWS=5U!QrX3*PTtFHloJW50j8Hi2{vNtT&gc1V3Lj0VCOuNUZLV3BD`$
zRkg2{vWEwt2jU(V_`?TFgp#|D#OZj-cCJgu-p<qkeKNmQ3YBTW6-N+N7KUhv3atbI
z^oz;lr{gPiu@A#XxB7+OVj{>0$Tk}-(Iy$W99|>+WP&da=8WQLpI5{#H{_K=P#vw}
zEq@Lh{oDrzROklWXM!J0dyR2FX+Uz3Q@(ZzSGcHGgCrj|7J61rt_S<oPb%z=i-Vtw
zyP$?vwwD_c4C7GlRrc7bE(R99sEPS2xZ?1-IXV->sPp)3N#A=f3=EML3rV6Yx?K9!
z{&8*D{Ay+<G{<kRRi>E`Di*T1J7w0mBH#Pc@%Zarxm7s3@j?652$d!6<@7N2skNzY
z|C;P~PmL}T`LI22{45V`73sBh%`*6Ed9-aNq!io>BK;Sc4Fcq=t~<T4k9m+papQw{
zuWBGx+Zx9LwRFU8|5${?cD3>#>Rov+c$HTGYg;fgnc+slt4mOp2;^0ClG^@zyDSuW
zYaN;cmP}QNLLb4>x6NQ|Ei5GjwbicFZ0SFeEfO9UMwJ?+Kk+hbH>6!VJZzXL`W6+Q
zhwwv;iPxc4ts=vj?V(btn_ox*4$EPoq>$}QpKhP`lTna-w0Z1JTu&5H-!+%VzJ0QX
zT}`lI*rD+)_Y-)6cFgdvpgCiwc<{Vv$_Ms0uVgEp77##-Gy<93Ky2(Ce;g75Co#Tg
zZaCGr-vp@>1N}M$BB8_h^1xQRWj?hXSE<)?io`(jwy#1Uk$P|{t$CIF7gq-rz>8H!
zE%gT|Dq;v05eGyownmzb@MC)-cAwRO7>i0-Q_b}y%2|#>jo&@yf=TTv>aD6SxNM@>
z4`w)3X-`zF2$6<3vl`#1sx2@ZI)Qysj+5=9bn>OCqCeUapK)OeTu1FjQLUoialjZ)
zc?t72_$vrzptIq`EGw|EuVl_zyHML@=25F>?q}V!nOnC${U&d#vu<T1Xka!2TcpZu
zEg+%p#kfsEf^kBp`H5p%qWW3Ny^@rX{yP=J>BPL{PN&nE0mpAvaamP`jb!nXQ2VJf
z);mIuS@k2J;6iYhl=aNyI#R4MEl@JA2svlu;9X=SuP-odEAM&xik9BpxH-D3k?yif
zFLhQbOs2Hu(7@wMWLnRZ)+sroyZj}AtgO9?tE6@%)QPzS>zgt1|6_b0?^ZZ_2|J^j
zY(v7|`GQ%X)`ag;A@mGUufxcToO<~+-(OZNMONEJiWuJvRhDL%e22T7UgEp@QFe<)
zEip2U^TJPgWOZ(sXRgEOP?}?haSw(b?ST~nM^^Agbzb|9I!Qtv>^EN~=C;4NGh||5
zBo=^pfnXzPcH1=b-9r-gyMeTxj>CSjSwQm|Ea<JLUxmE`>!_WZ5`|>q$gO~q$5zwU
zL0UbWy#r3R`60u~U6ncI8(a;sC2m&136WlYlg+X1$m}>)*#~7PCn^6zC&_ZXrbJRj
zw4G+q7J`O((FQTTGMt}5B6Z(WTwngEtnU#gaN8riVNp(ZM#TT>eYS9P&NM;``~FCF
z%Ejl`C^D0|DYaDLq|3w5phaJ`l0RzU#hvz2WXnsLSINoALCt5RqvM{P!<N%(w*gIX
zBYB72NAgJOLw9j1=D3hZ4Bb6&mv;NfQQ88C8Kho3<6857vG?9#O?7R$IH+_G=^dnr
zNRuvIMVg3!bOHoLdWq6|5TrK&0i{LhJ<^qu&=C-kE}hUj2{k~7XTNjKocZ4O%*;3E
z{eHi>X3q8gvG)aJC#<Z!*4odypZmU_icjrkat&FEd&jR8FLm_KNbbi?6s%iGVH#os
zElbU@!D3e@*Oj?GS64UHAS0l)mbLX5U7@a9%0C%+v+tM0=xo1syX!8W37!;aA4dWl
z>lpvtrNNO(sj%Y#x6P&g>a3e-b1D+v`+3|<<eB%ER;E+T&C|ChK5owOg;{(-pV#5Q
z_QF-0v^|g)KiFGGTC1aq9KMs+mUekplD2kP$&11+YM@FSNRA=-&uU6zd)T09l`Tk$
ztLt8~!4mwcK!TzeS2t1O6U=Mlv*+G2<<^>eJ4qndi;48Q7xzP2ORI}Xh!Tsib}BT1
zf<x+=`Rnz;j49omH)YO<ql1kDEXil4Sh&6=kdP(27QLi=vNL-tWN8)dGNfWZAlZT7
zPXT`#vWqnEb6dSS6<4hu?>^MhrkU|1x+Pbg+mJHUzsBCA3r5<LAQe`)B@eOy`Nc{|
znW&R}OnPjrSr9+A^uk}_Ht%?@^7gala{S@^3Swp8K+eAx4wFV~<#3k<nqmfEkzP%p
zAo<~Q@DH!$VJ{`!N5i6G6~nU?PZlgSp0c;wtF**Ol0W=b5NjLKWj}DP2Cj@<h~mu|
z-!q!vMW(Gi8N$hxI=$!=d9FDb+`$?fHB%e?<vDNPtv+pz{BxR%%)>;i2G5XzW^?G2
zLSYN2Ea=&y&|a-Xs$=}(H&M#C{1O{({>Q_bi<z?9Bf0&k7BN*F+STEaM@g$SpR31;
z>YF39$Rs{DCfLF}!{2=oOYfb{fZEM^;>v#!aJYy_HNdtYBWGjS$0Zo24KFw;cFXJs
z+(pzANrBB>4N_hlc9y%rxIY=P#ooskZ)@67{4({bWR_j_BAn3|-ujE+TC>XHJhM*T
zCdSlP1w3X``=fld;W?@Ba|i9_>eChiT7vh>+YE=q+2<;b@`nEAAYXTG&P+!y8@FT&
z7N*`W&z(b1S*H=|@!FER&)hu@*jIg_$Cl`mh`I-(u2N*HjS1gfm1Rby9ruhC%)B8D
z4|42{VwWw{35?oWS>%-D1t2gLz{Uu%hsfjKJrHTYsc4_g?_#h>2QXa+YLIrWK1pfI
z%-8C`BJ^yj-M=u!9^yex402>$C%hdrNyR4EsnhzES71@26C5@F^DHM?6rQ``-w8nt
z7|#KdulkFTrs$Cv_fA$&zT}QWw*><&uf+DF<$%iUa}jL7qv|nN-XE2fB1a<1I=7y;
zjJvrtH_0*w_)^x0hf0X=t9=SA>**8ETcn=cdWgA;zz_XcPK65tCWFaBr`*t6bH*72
z<;j=glEtkzRqff);wBco1!KWf$bGOF=o?7CX8sq!<OMzAX#do{3w%vci)v+Mr#6@?
z>_N>)pxX(83*Y~Caued{9*yv%3-^BW#<UvNB8xjb4S>=H5BYZkd$Q0TwI6$8x{=T7
zR11D$O$<ia^pkd$vhxyFXCKMT3zx?5QGM-gUu^NwzexLy(o~Q=7`e+zBzrnpHp8na
zrX}mNT|?~kCM)&XU2aW+VxGY$6}xQeF4m})Q7OSGRr(i$42TlNYgQUuXLVm9<ct;G
z<b~WYSxR`DyPX`Qvap4nnr4tGC<PB->mVK8enm0CqP}GSos@n@YPz`GQ;b5_&)yXx
zQ<7)Z2+r^O_|Az&E@DcbmwB9JiKve{3&OmPvkkf^!i2R=QFZV*)~7MeT~GhEK63v@
z$JL<p9a<HoF<t$n@l9AwPD;6Nj*|bL#+T$H(Pz!KE6e1~dS5c@vk#k>=aTVRMn^su
zmJw6A*#&Ih`NxmMn6NgT#xg%DcGE^ck8ixYnfBd_)9I4r0Z9TA0ot3U+g(6|84m8(
z2`qr?3c09mts@sMwp3Ta)L7$rRC5m9Sl~U>76>l+q?Qb0(C(k*Zm}7!_>tH7DCt2w
zWw)BKk>3xJD-7i24b4gzx3$(6H^OY`dp*j1VVx%;ew<+e<)`{>#h-<aoptAE2PANv
z;d)oyH5m@Crk~5MpDE4ekNfjs?%B&uiFoeqXl1ld-PJOV`O#+`SHf30%BIUp9ItT<
z89EYCa+z2ATu%mlS)$?bv~<wXi=lIOP@(4~ec3G%dzaErPp*oG5J~4Y!12>~%ZlL+
zw_gPQQ1|rHygqoav~R*zTQGaeg9dbE1(bhL2U6<mVSsE&wNCG+o!)p!x=ZPPshZK7
z)70~doJ-swCC&x2)~SXTNekl)%bVL0tO!Ev8IQQl+ME-OdZU)=B~*FRg|%YTDKkHr
zrgFC=9~|2cxg7@a>gaGD=tgSMR%A_C)p*!=Fu#d-%NBJ*%R;kNa@VFM#IBd6(!lWI
z%|d2p&NWN<_8`d)d6D|EMm7%j&2hAa<=~a+Vc!B;>uQ~*ld!#4p?5Ad$d*A<fDCc&
z55|5ZP6bD0RU}}Gq1GRjR*B5Gn2{MjYo6mBRpcn+e=kJU8fTFxal6NRUS^2HaOJH+
z+`iVifW~iOO8UhM-3!2Np{SJw*gBDRJuR4a#IvI(U|!C)lzARjn(@p@)qd^Ta>816
zeiD9H8k#mfwIE>5^kL5V56k5jLJ0GUjxHcR2Ld~e7f!&6A(r$k=4Srw+?bmLz(JS-
zEpy6L1<)uD{3cYU7!Yk5kI6*h5o`Zsfy%D2BmEKB8Nmo~u7y}-`>;+QR+rxALs_!#
zYnMo>zEh*-kXP{pI5F(21#q;<6g;}t(%{|@cWqOwQMzGtuJ!;%zWgC5+O$tE^h+CC
zO&9dO*Qq^_T7HN%fC74A>d0V8bG{K@uz7<elLcvczvBz!W})({(VH9NbhBOdPjHE;
zPChG}DP6-x1^t}1KGO!@5F{IOcgBMGHjL+~Z#33pgsNy}Kr%!0K^*x?SFAbRNtJGE
zU218cKL%srn3ScRxl|18{=RALkvd{xSs^#T0w&^`a4Gv2febxi?XEE46)r}s*7<$k
zhJ~H%2^{x?w@jGcG+%>`E#>~h4}p0Bf(ij4SLJTBT*k8vE(u&vGQzOJ0JCmUXsx6#
zImQ4o^6|W0KEeJ;ZK?H=jsTmW!h!!zgYCh-lIrpBMbhCOkpm+_5-pk~;#tOQQ$Hq!
zb@#p%QsOSzJ4iTafl4%%Myqs2Ca4f@Md&sin_SLWwhsObFNQ<iAhC_9S;y(-WAOtO
zov!qs`U^nSnsK-LT8plSstc`&P^0i4&ou+B(DUs<L}2d%?{xXGvBnh_Vf`@8ouAh3
zgqzbO81nHS6;`6j4ITqt;(m{HjpedNGj!Anr+v%?JnL~S)IY_3&QX*>CRI?EtYkKj
zD25T!mvP(;_ruWl^U!dc(nhOv=Uhi`f3G+SS6h2>P#dKU1-rvXBaKf4<aOY?!OZ`t
zjna#GTI4by%Do6%0}SA9G?bAnn&4uU`2ZzNAmFgYH-7XVa%$r6Rg5Xg-;j~N`rf&4
zBhC2C8|KU$GdDcUA9GWWlIVfnr4R+en1+Ad_Q`lm3P7uL8b|8bG{p|jKW9+Hf{MpA
z3*%oIYp7q5c%yke?wZ4h=d=!+Zr!Jmou^e*b<x%hb@g!?^y0;Y`t$@0cL@mY5Ey0r
z)13NkUj3U#r9B91v#>{tP0__L!cj`pMi^pP!?u6jaVF)2@$_Sp{?xbC)>&~h1D>P$
z<EZ67cCVohXS<TQtG;l>gh2*<ee%&WeaX(aZ%H*vFOBTo>9a0vYZw@>=Z28q{Fg(>
z$=o%krfBA;6XPea=}h;KuCoVovY>thi<#-Q)&x@N^o3spgjjpjT-ua<c%VW))?UYv
za=<?CXO|<HJAv>M>B|C_v+QGLbK^t)J?8vspujJJ>(j7v(s}5HMXmlH_F*BiJII)#
zViDj3=%sG{se@PtUs3A%&vqmqGcQ1fQ(R}~uoHvRIszACf%s;Ik2rO7P8kYDyYb2C
zjIrc0$a2zxGCwLfEA%R<^}Q>XNUyU_SDGv`b?#h71;dJFxAx~tH<?1)hdCs*8%Ny)
zvby)LXu9SmtM04u+Pe`Lq}&~b*c{L;$>Qit7Sr3Z)pyTPSB7x+(KedH8mRNragK+*
zy8QY-OCORHUx)3|B?kw#)MF#_EGU#f3ww#>V!gMn%d@Y&WD#(@DNA|Do#eX7&U^@7
z&^KUF_c+v{dcMc2sF8mx0bXFak^W+jMZ-CuI4|9XU>fW;eyf$%C}hhyqmVIccDYVD
zmZ;M#ff!bu_XK0fk=-JXOY8ziLaP&Zs4?C93PG~1HMl!N&s}SrgGGp^G$+1*yg3~O
z>aQihB-W(QY|n4td+;<DDb2h}IdArzm2ZrOBB~Y@x;jrP<yu*kEv%F8U-Dm0Q=EYN
z#&(;lyxy(Vb~WDHu5mW$c=QP@IRE{#xYkYGw;V{mC+zhk{c+17kPZ1P`SJJ8P`7lN
zHV@XmSRYsZ+7C2uI0YFj*~ZN7adbI8ek)k?HBp7*-QodD$ztKkbndB0V=z&pPzuJU
z3rvpXQXiKF>@Za!d4mvV=v5ef`owJ=1@%u4tY2AK?%KsVh4)@9VwNk-A;tEh5~EnV
zx>E{e;P$0$1dTHC$6?TXT|R~S_q^I)c^nO8E=v%To2lBkuNm#R_$AJo1aP#_>)!=@
zD*yDsF53w|Wen8%MPLnRkq@;hHrsyhPtx+IKl&eO*8lB$*#`l0q-8e1y80>G-i1(F
zjN<QRsvhx01IF-(2sj=7X=BA+<XD}~N5DND5J3X=oraRXi!44+#bCsy%;(J;#}L=l
zP+R9zVx|xl>YCY2Z?OOf3%-9;5<#EA7%Xuo>4m1^5($vQ#PlSnpeS3=1<0R>$&03P
zjr`pMZ$Gw=X}XeZkGK6yr+enU`z|*=i9d%UHs=FwiKPO!$s2s8-E`^(C=nn8*vxk2
zYZ!r8FBl6OjM*8Wy&T$)DoWblNMxvsajFtarG3BpMlz^$DS$=OdchUA#gI5fgGCrI
z&Je>G&J^w9+2Wx+>gpOO;JnBcW!`S3`|MfdQyp%RO5$=p@6S|1wbi@A1$pA4^949t
zfO}=qp>nNRd?Awxq%n@uEp^u4kZn{$R4`3@jyjta(e)C{U@wN_ZiA0LtbI^YhGBeK
z8|!+au2*4-A|dwZ{iZE((B>!2sb_^dtIH(cxV@Eqme_PZ7Lf&2jLokw6%|4Eq-Snq
zCi2eW8?;oKApAqT%UaF%^0vVA!HkHK#fygri!nIWW=oq66V|cL=%<%Jn?fWw?fz5a
z-W5{vhgBtSWOnu;d^x0;<IGYlq=Pol{5Lif+1)wPVYs4#7X&`@Q@eOh{-N7SVt?SK
z?8AOH<w2f-lav#kRUlr{#Eu;2W1lX-NH=qM)4R<@%ZfJq<Lgs{F49ZQza2NwJ>nq~
z*t>iGi@kN{wWO~vypx{z)*Q#66GL!n`Hud_Re7CWE$$rDI?hVQEbLnRA3jsNCm8gS
z-1;E+CP|6}Z@zHTk%?#wp0c8^<=rv@9^p8Ji;1<1>dYgKz3b<BR})zs>!*TSFvs&c
z<94x~CkSD**KD)@*jQILub*d2gUIvYqTV%#!6%Iq%BOFyQ*3k5DUx9>tD@&)@YgpH
zg4Gj>;5@9;iSCd|2kVtI3WpRacZ)~7$*UVHnb$OGC;`N<Tg0lUI1r46hX<;N)-3KZ
z_<}mjUyN2t{oE6k{u*AR#h#EWDI9wHMyD%6^q~Y$31U}PdM}of6xg2t-*F<YStW1N
zzk7arTH)de=?%n8n@O+6i|75<h<~_i(raiIYXrmY6gc~f1kAXnOqqkAsVNg<ez?IT
zoXL_HRhe_zJOx60&}(6J*i~SxF5njdvy*&Xc*VrDp4yehHyMt#Lob=8XjU|IZ{AWS
zTj&VF2$@4(n+rF(Rn)X;FK)$hfXrL4Nxo;G>7cj0q+y*&)0uSupnQvk;~s6^-RGo+
zT#XJN9WsDz`;;!)YI8lUKVZofauFzlJA!zh?Le+yq<ig{baGMcib_)%=#IPOjTS{}
zH;+hzDy$f5^qT0wGxsRwTwS-;LfS$Inm*ZBA=T)oTM*zkYG%%>K(N9S`X`?3SJ{3a
z^Nk=;3V?G1rMwBAxC<~_G#&@m6>g?8_&$UAl1Wv~n;xZ|W{ImMe5FPEH%i6dX6IHu
z%Fl%h*eKxXkbo;h2jB|f0>%Lf(QK?UK7s?_*<1$PB^FE9aD1%biz@)Ow08!LR~CUh
z&QVJcIyd992{(GLrE)I=X{#TUenPY3h^$5IN8$v6e>0Sb{LN6}ClXt#@`q-vFeH^k
z0)p0gz)|8U;3y#@hQF1FeT<X-MWA{$71;b$K`}=E%TdAyaIawYgyX1U0T?sy_w97k
z|4r-u`?eMX3CcadQR4On7|RAYN-+FSM~T1JqJOfu_^Uzx&OQ69hyEHv|9eIFYnJ`(
zdimdW9U>80=#Q$bHM5I&kijChXUP{ICM4MSHk5ReG`EoqIWA34a1L24*`xqY>;LB7
z{NLNR1F!uZtc;DJHP8@^C|cApK;fpm8m4%MaavRtOMWp3Lkh@?aQrjL1I@6@0y$Z%
zd+YgQGhjQ|%*X}FP~`@^1$3mW-i*~*YU*vKKoOx0kH_&`n6$hKOWz@gOMLSui>_kV
z96q|B5B-^+J7b1QD2tr$oR+1G{35W2MVo-HQilV|c3Tf0KuJOBl);g_nUNluH*B8<
zVyxHP#`n54Jn7!4ik!-z`e+365Ea@rzZp|&%e6Yc+pzp3#UO4RGrhNZI=dGXIQSMX
z7D$H)E-2or@%8D5eQAwgL$|b46qFzTcwaPf<kOfjcYF7vscJ|;Ot&`SXKO%!5shHM
zeuqRQ`&LeOFqwFlVJLJk1T_h}LvE}Cp8VCR-5a5Ha_=IZEWC*`@#${@GP!Rv0_S`H
z8u?qKRC>|e&E9TN3VTj)`mMxe#K984Bm=mG&9aNmVhoN|*g}$61B{f2<HG5oI?`0k
zOSkUc9oCGxFY!xiGYrW_%3w-aqJc&%^%9a|-Fpfe87R4EGzLG6a;X%3KIW-ppw)cI
z&o+U|Zfcd(J73x3@&%0Zky~H6w+7e=DAq7%ghQe|4}G4GUM#Q3Hp8KrNR+F}pnfrV
zZdlw>c8lPszNWUCHE{Qt|73Y0!ti3~S2cVYF5I`ME0J~Hm%1InJMOu|r3c8hQ(KO<
zSsy14iHvPm^_3<GJtXaBPpLCQ!+4sWvP~UPpOPZSpl{&x*e^@*Vkm%`tG3ADG^L#_
zrV(*&_*qkG6`7hQ__B`sNi$4xuja~KmaNhU+rt-?KQdeDM==K!=g~r2^%G-`Q1P{{
zspX7EvaH4l4ar~oSm#X|#dR)AiGQ_8s^I<{S6$PXTm<R6J(xA2>*vEU7LHYWD|K_>
zJ&)&Uh+>Jg<kw9Jq3E$@ebl;?rF_7ww${+LQRf@wayspq3R-P^?y8LMjNb<dj`AXf
zPr>?QYK6T;Q(KQQouM9lF*cLQSC>-irPgv*62EK)WyqYk>{Tsk^{N$CAp3*F#-}Zi
znAjfkqA<?@{f}n-#ZLQ7jpvGbT7u5L^jf!1X{wr*qi9JaubS+UbFnFNTLNj|8gN1&
z1O>tnt-!<K(m|qF^<)>XI=^Qr&Y?wD1UoJ%-75@qPv+okr@G%aKDF0*GK5GhGH27|
zPs6O*60+p0KfM&c$CE(5oGfl3u%pk7^*E=&;zx9c)6GqBN<KvvJs+p}1ZmRbcrv^q
zq^C;Q6D)h!^q)V|vXK4WS0MK!j+af#?R*|r+c0pXTrX9s{<ti4@z}Bg4;}9rlBs-N
zAN_9SJ)r=)5(-R6T!<TVTc`ve3<Y&*0UDG}CQc!=a#;DSqpUgoS$LL@fDVGlO(ZP7
zFS+Eh%d3aGR^Igs6oe^UKoaVd;5QzFiNP;|;q#a^DY;qIPV*%-u1RTU)91Ex<{<aY
z4<*x|OBdXY5_IoPD@1s$MJbxt6j;qF!^r|1@-rP9CciN8JvHY}iM!+xcV<^<-usf{
zGmqG99a>nXv(RPwyDN`kFS<7s(Eef3i;gl_dwmKBgdOo{I=><2&a1f24#!LC0?E`x
zmlJMYo-F<coiqMD?C{UNRwjGaog4k1w8H-pjQvM{e>zXO`ABKvlzdeh9{F;MRxaS(
z107yXhF3EM38us*2BCgEpAH#DoVfOBprybL2n?G5)43rVjw-~o(SV)l-gf5e+AmF2
z$?bOn$3q2=;n`knS$vq~oo+4m#rh&+Pk8@mdmXvgSg8=qu><{Wdb;O|Eao1o9CViD
zS`Fv+gGoq_5=nu1WUE9`R<+YZ%;U~Jr1%}Vua?8N^#`BEVCUO{rNGRUm5KRItQ@PU
z{=#nMN4$hd&u^>Ux!xfhWthM2>#}zJRm@tmC1#`pL59snAuny}gGEc=qFbPMOR6sU
zHMh#DUbLoJPzr7eeYA0&Ra3H>65E%<=S6BEESWNY5vWM+T{yQj%0{i$Wf~8A&OQNI
z&f6y+ldd`ddc>#J-fNfdY`uEYfy$PdSC@Yv!+3i`T_$#0ygT-4_jeQPB@%%+ANU}U
zYzaZ{q<8aS9XSf~0v2WJWaSz(n_op`=(E_pxwc|x6BpQ@EZ6hV(Imx*HIPY`k%{&B
zu;^+{qr;SG;)!YZE+BaJiaM!eaP|<}I;)zmSyBE83MwTK54l5dyBtojZm{0?yndvY
zqG@i<B8>e#o9&l0Ble{?m4qjQR(1Q*^sQ%OI3m;rNA{rz_xl8G>GB6DTf@NDBmYh3
zcStY<n3*tgq2Vb&Wycy^%EdC^+}*{gh2MwHKSSKk4N)H4*F8zWpVRvZocI1B2#UeB
z#}J&W^s3D1o^R_f!}bI&j)j)UZ(iO8*aW`_{8aX;G?Jmsu#=zgxW7K~ufF{6G>F<i
zTh`oOxVRdiZ;GG+bOYKz{uxdIy`Zmsa@l<vTk0SB+v!i6G(G>lwO1+2?}2GO;;w%n
zA@Hvu@66_rxYyuK<X4kbw<^?tx(c)Jmt7q95|#1y6C>>{p+PF&P$hZUxSwWs+uG;v
zZny2s1a97CFFB15e*Z6PJDVN}Av#fW)N^UEKl6f1Z-@su_AFy~+NddEb-jZVW)DGV
z(K`Bw`*NkZ_pm5tputleD6S|x56*0)j^pZxfz~0Esz1m!$lh&$YnmJ8f8T-<@-*w_
z^)-^5$Q>A=Z}5DVvT)V#?wsk)+JUST9|@a{9Wx)z=+&`xWpOMNJ&l@!aRA^H?kQJ<
z%X6rRXkGl{so@)t-H4>S0#gMQhAA~5|H*>WRHi)RmR|$|DnD=)=%NlT_W3}yA)L=5
z_0`QVD~xD%#jMlAv?o#9y$!w<AFhY3B$M=&`*5_nt2qy#7grD*I7LhF-H8HZ6}s~W
zhnOC^rMsd?Dneb^K#ih&d_sVi78Ps}IP(l?SRPE`Yhq9nx<6=$XsZ1jZyT{l5tSoH
z6JeAZY8^^?JqCuSCy7L^9RTqWvg#KBrqMkis|lp5?E&2fzY_O9usXSL#cQl8m!N_k
zpBrPHp-b}wn-g=C4d~JFLDV#}s5JP~{7PtTlx%NWtW%jBKUv%LQ0I-r$ok{)9I_!)
z&G3;3?U2Z*YSHEai(ADA_KCWLraSQ+(HFPxbd!v7p+6`p0S~YY*u_O6P<8-V<H;{~
z%paTr!QT8&puF!aD0L60GiMCYD?VqsE+Zcxb(_)TLbJkn3apA=6QZteyt2}*{JE(c
zDq@o<z~{M%QPO=|=Dae37n;`oG-m$kA;GNc6Zfa}Fo5Eh3lgin@8+fLFu#-h!MEa*
zA(!+}YMM%OYbYUceqPTy+<6pFkEK%^lNylK$MB`GCKT^+glCT((hPWiZOVS8uJ`;_
z^P3L~?oQuR(BO8~<M+<R8yN$0c}278{0&XDjS;4W_a#_3^dheKkkAvNJWVQmYub<$
z&6i)=l0@TPLQuVXnYQ`n^%c%e6I5Bx#IF-21W8)n#9l@XONe<%sBD!(ac&9e)V!@X
z;m}_M;e@EYQ`{(kwQd7{|FQ+Ygp;Vk*doI#1MR7h(Yy3g^1_W3rE_0N?TDTVHHR`J
zUe1L*R0Ik%bLX}UpD!T84!@M=Hjh_U@FOy}$Eyat$fY}*Sa6@)bl=zK)eohF+x%b!
zm>7TPa{^x?hCSeE`}gk6f6wd2p;5sk>*kMNmCUR96xY^|4}Xi<S2wKwc%5C~Q49fw
zTnNDSQFna<@K9-$^K-6W1Z;VKoFxRn82sz1-`0!2dg7l~&cAx%|B;^94CDpsr=$UC
zDJ#H@>rBw0f`UU?=|=Z>H|+pJtsuGN8`IKaf-Ilb+z?R^ltEYTx6_{{F;M=FfukyW
zR$o%9a&)4yTXf(=@bm8gs{gk@>}3i)fD##dVRA~2okH}Esx0#WH7D@*5i%dH4leBB
z-n#PQr=jcVm3@O>f$|^qV}Pk2qJIq#stecy!07yx_^Ru<&fsxk3gB{tSy{Eo@-tgv
z=)m1_j=}{M)aC2zygZ++8sLtn-i66omus+WJ5M&-F1%CbaqKO^!dvkS7^T;cRsPEA
z#a({L9#Qhokf05x4Cg_NPB2os2$Z<}`J`cC<bs-M0>O6=Rnkm`g1*l1Zt_6IyKjL$
zJ2p$WgtvWe$YSAckS~v>erucNb1F+!6};}!OZfxmhv8b3-s^~G!0F0{7L5rjs^0kM
zQP@*3KcjOj)Z|bB^S{WQ<(uX{V4%2Zm-HfEP}dPd^2#;P4D~rsp%O6v^dF2w{p204
zBc_`wY<oRTH<ik!@#w08Q4`nI{z8u1lUVJJv`Fm+X9a4Wg`&ypu9>;9jPpEPqKc&p
z9I@*+>*K{#dJ8sOr6$}^K?Zo_!WoQn8mNz6SrUxACX6%lD!_)Bqjkd=pT~S3)^?cM
zGfgIYua6QLDzK)=Vu2Nt)r>a#EbXyigQpi0iXcn<WinW2SC{%JquJCHs-XBRgVTxE
zciB$%74C=rY`@HFO}C$Dk%CG~k!p0Q4wOSX_ZGIM)r<DMo@!4Pj%L@`M)C;d-Z5&W
zEY-Mf6z|-chTy|6zK3t(ub^{EDmS1QRH@acjLnfO{HAf8aMr<olDV8ft|jPDXEtbR
zqaLB@?%3Nmi@cHsm|pkdR53PP6P!{LIO(gkQ$%%Z@qmp9bz5mr21$x$L%MpxPAEiO
zJ$3b)({OnV>(jRO3SW34f~$WCN#mQ|Zs3~%rrD$m)M*H%az|Y^8{So+P*CNKqT!`g
zyTyH~@Ao)lVv@D!wSh&|xbP7d5f-dhGj0YlXlP9TmVhlS70fpO8BJ=a##R#&BTr~l
z3y!SkWQ&;M;055p*HXn5_Ml?k>S^*WiO#2H*F=h^7I|za<;cAc7Yk*q9`~b{p;@2<
zWLm|;c`SuhMTroy#c^e=XyCSQo${SPJ*+C`1_;VuJL^t$i4=slgtKY2z@FUO<=^L4
zj(6!8Paat!>71ttRBFUfS}DLDWvXE<pDvMchiO;8ga+$uycG0({4?8i%*^JeFOiwD
z>%4v&%EKwGmpZIjnigcGRoq%NP^;#3;}?O1S64$nuV1xo&$qzt-qZMd)S<3B<#yGP
zt#b@9ea=L~1q~XPWNR$XUEizKQ>uPTp@F@HxXDSK)_U(j`L_3p-0TCcO2qg$ovHav
z0y*`7gvjhb(dJ<<>L?Xfw`TK3cPE3E{PG5nc05__I;Hg*jQ6`Na(nMU^1HLle5Y>t
zra2@WN3{eRjKq@;H8VKE9Zl@U<?N8*Q-)xv`7Yil?gZx8IKHAmi$)Ag7o*V;FWW=w
zxVNYhiJ-<@-sohzF?MMt^O-zz|5SE1$!yQV%r3;0I?Vo0cr=a-r9u-+7Pxh<>^1;D
zX9Z$T(~4|f$T2G#YI)qqD@;32(L(Hj&gM^R2dw%tO#lwFEaXONi^w!Mr*1F~MuXEw
zWgB#kEAywfb4`LN4u27_X7fxuu1}MLXIUw9?<l%05?{NO`UD=IFMn<NJxa;B(~1du
zx=3-ZOA9FM=lC!?@u|n}m0I4%e}r5OD$zHyHX}6FOnMCXw6U%UTV`&6fDt=Tsug^K
zphNF0Z8F;EHao|LuC9(kI%(~u*Ah>X10#9Y+~lNLv8Dx!b7Eo55)IDo4Y><C=WQ=O
zol=PI)!6&So;+=Qed>`Q^w{>>aw>A652z(es|KEJNC7JwrksyJCD$kbk6VPq=L?vk
zV}<XO>%+B9D?aBztQzCocLirAR#*J`2O-+;<~$GFY?E~}6~!*qDO2D)u#rnZ;tO9=
zykXwS3W?!dx9TagSbu5S%E<DZ!urh&+l23&w7QmlCghzdx0O=q1Tgb$79DXKuIR0{
zdOUXyv`)M1L2-loBZl;Ej%t<r=a$`z9uZGxMjOE$TO8C~viKV#&4efpl}`T<o&(c5
zA&!-=cfKau)?Gbq-PFCOhi6PnNA_*)1;d}%?zv8ibs^KvT>|$|;CSHXi2P=^6)ZFs
zs4WmzS+^JI506gB^eURZN7S;C<vqxhy38YILifqb-X-bnPi6CMMSToK7Z-Uj%~r;Y
z3QfwQaRq-`iWfcch;GDYm*_4dQ)6QP?Ve-z@_}nv)>1kMJByuTSCtNTNAM)<89Ffn
zK<TP?=7u}A58j^hQ8jq#Kbu}wT`4qrakcA^L8Ajo?V`w5xGC@bYJr;6!9O_v+ev+f
z+q7@x`-pLgT=S6S1Tt&)Tlpg`=||o>CsJ@kPFji7R4||(5>i3#tp<}|N4m}?Vw=rA
zVv~D}8)`;&wkut^I(^I)HlG$8Jct(dcljR6$*Ug!Uax+GWT4h&QYaD)3H<<=^Yj+^
z8URl+bT^DI!oGGW%lyIjyF@V>E+4W8ze{#OF3LJp4k<=d(6$1=x;Z4^AsyuAGGsdM
zgrUB;bU27(0XTMF4}*Sp)#%&5cf}9l0fzWv_%)y;f!L#81Z{6&xI~SqW&roO+amdk
zfcOb?Ux~nFTsB^1S!V*d)xPvwU1qM|{m-M6#jR|wW`PQ_+8rex83U5%qkMK0EA^Vh
zgeeJ=T;>E@>*bI--$?u`e|tP-K0tT$R~}pBIwRVEVI@)klxh94T}=16NfdOQsSvu*
z2bd$}?O1XDBKTkhK)+#vDyN(wG7f)z#9w{!*SPqH=Ex=7{1$7B-3E){x0vKqqA}*`
zh@R`T#OeVHK{r2;Rs8=f0r-<y`hSb{{4d%2Kv(ed&HtT<1Hf8)v=9EFOwum`SLC7g
zgykZI8RAx;_jof`u~Xx2Dr)}fy0LuZq^Hi4rAN-wv0o}p0#l|Nb?T*S33xs~dH$^J
z)8|DzgJnTtV(ph@Gc%@vuMq;Nr_nv+A5tbh!L>2Sa8D%T)_4X!y8_$PhUaOhcNW{x
zq0o7D07{!*(Pro>|MB(m^Ad-CLAxpS)xOPaWSPpu?yOT1*tP%x%Yf{efYIdXoHsDK
z9d~e&3p$Ej7^ZId(7Bmui!CPmM=?suWOS=vvU#)L|ImdnuEf1(TjdV6v9KyQzuS%%
zMyJQ@f%JwvbyPa$Ze>~yr3IlxJ<nq`3!Z!kPfbd*j=TFcFup0OGDG(@g0XhUlfsqc
z2?i2=<jSaB;)=UhxUt8781H$$`wi((mnImv)=|IyoyD+nHjla<tGaO0wKmhtRr<VN
zn?u%ccqMeRd!=jNNkW5vIRz%*p}EK6Qldwh*z<^W?qmXt7!Vr?g23OXrfPYJsOTQ-
z47nJMsW`+#+ZrF&sl+MLK=L-PVF7w)m%L>`TCH2!F53h3v^jej%qd->jBq$@A@iG^
z1`-AB=5ag_E;s*BG(0PMeK^b3aV%fvsLX;P8u<JrV?W?SJparkvgiZ_!;7gmnFVj?
zt{XWuakW9IW*y~c6B-t|*nQt0QIW58ys+F%;~OY4ef&Y2o><|*lOC7ny|<Hk<J~)-
zbL`P$g)}GXL}ZPJ>js&9RjsP3@!u^$4@OLTrR!Bm3lbK`O`j9Kx@E4^XI7s&GYR(z
z0H38IN!X4Z>%3D@;3T1~2_V6T+Yv)rO6gnpW_GDiE};isNgV*CH<>pn0z6zj$}O`c
zZXRO!bTd{x6TI#3DK&%AB%VXwg`dXTz@`*H6F1QH*(k^r)e`5aPV?3;HDl6Awv4cs
zg)?eIow__7f-kJp$@{WUWO~eR>j2Mb3Uh}>qWZ8B--h<H`dI$adi?bf&y(ac^~RtR
zJ8-ReKRv<&HSK9nS!&`o6wgJ8Q$R<)J`$sxHx-L&z91gaua6pd>i&zs(Rm{`W7RJ@
zQiNBFaXU0mbjF52dv&i88I+AMsoSeK`dK_FU37#G&H$+rt4!4oH4BI~<2lmqxCQb%
zFUE7u%qGg}`+@qL6<yjjQ<NIrh+duF=ae^wbm~qkbbzS<k1;ASlP~Cj?4F3l23M=T
z^UfiV6E)wP{O@TpgPd;ld6i`1ZZ5?v#C6jP&=CG%WfsSSo<&d;o~zll;Onso`2_hZ
z@`CR_36d1Qen@v9mu+KNzwOzqfj-e1Z`Lo~lJ|)Ye2LO8dj~SD`cc6bA)=LM^uXSy
z&hH~c#E4?LmyW@x_v;mhQ@`E%rgL~v&i6qdPnhwmX^=PIk%X(PH=msQnzsm^5;7ce
zQ|8rmMlpL)QMB!wgE;OwxpA(7qFilF9Fk|hd1DItrK(qA_&}k`n&yD@%n4UqE$ot0
zdWLr<&t{Hbie|(?J6_~GIP3nT#lZ0^shc_e&tz1u3>8kC#^9oKb!1;3>WjsR3VZ1?
zZZP3*AR0#PT0T{0^VelCbB9x!HQX|BZDiDX5Ho4rS~6`K%!SoJS<)`{oh?1V*wvJ)
zF48Rq#6LNIBqY5vSsQSd-_(^i?s_<yQR=ZFt(M3a_3?=SF{Cyg8D8fv>jqX(x!^C@
zZ$VnhX0{<~84DJL2Fbuq%5<?>=)}0$CG>7_#`~PrC)bk~#p*!sjI3$9D5|&9tlH{R
z^a9x~KGr2hn<u``8xQ25jwlBLR-H;mI@Pj^dUaECqBOab-g_*7vOxWQ`FXlRsBFJ3
zI=vlA>8oKr>}E5j2rk}`r+avI^;MAdj%{9T6s9WDwzLnMphXk7!S}QB{!+{jG_+%$
z+?T(Z+~qljuc+-zsXO}ValE|EYKLsnmx9jAqRQXi3DvsKu*cEV%^m}kY>}OD9OC{6
zc<aonac|jwg&`gwaN8#0&_JRBZRGWfpnDzIlvnrQeVw4#Ta7@)XfK}M;zF#pFmx@V
z2q34G!SS*hIQ}%1gAO3^)Mf-sShf#T8A5D$+J3H66p;K`DBb#>>7YM+@NW_5A3od@
z+HPh8RPvTq!7*4u0YrK-w+e|82C`W>tS~%zB@#Q!g-$aC{J^Pnfjy@Q80Yp+Vd~#)
zPB*3q7u|Sr;cEe1uh08K)5djUn|1A9qLiHYBlRhd4Q`4R&g3q4cch7jkbJrG;VR-k
zubuzJy6vy#{5Klof8DHq`({xbK+By3bKGZc^?eL&j3E4xXC1ene~G$^B{$?oJ^ohC
zhu==W(+U5rX87;@%<?;^KLT>(YT)p{YO9SBR;khJ5&nO85Ho)KXZ;VRE-2-n3AKT)
zcR65f^QSkbG%o4^@wk{(c4?qRYzLTL(IDSBolzbVAqA%TxuqL)^C?fS^uWyHa*t5)
zD*=Y0i{CWl2>cTN_Pv(iF0ohsJ&8coAxpbKy{Z}}^T)A_u(iEFE620AiP9T)Z$EV(
z5^NLL!?gAurvi_XkY5C43`N3C8xR~B82v>o66z?zYk_^W*d)hkDb_y4_{yeKOSvlF
zYNMe<YJy3>x>ojPf??FNpI7{zbzYhz@REgbD$*2SJ&Y!#smxtPJ051tv8{}HZk2&z
z?LH^jgf!EtSBHDQxV$s<Vg?1`)L-HXQC7sZ_$}$Ec1f%&v%>(oT&#}r^!+h0fqj80
zWX85%`XwPhnI?Ivq1VrQdv3Ahjr=08L(*6{D03&TPUsAK4FGVW>dQ6X22;%Cpp5nl
z(RWI-CyI%(&mGCbg?$v<`i8Y190CmNhx5A|!=1?3U4VIN1U%oi#<DBn=TCgC=C<0!
zH;fzWiyPOX5+g+!Y6GwWBX*2~%66}IaGiH(u2s?ra=+n->aI}~IXA^Rop&*sU;XHz
zG*Ejcnp&c^7~o2apKIUM-k2kO%iFz3B+T2PCAg|?u`1i4sD@VUk_^1xj8T7oBhjO&
zJ1v+eVtjYk)CY7ULF&}#{nMaoVmmKAQtJ<lmYr~tmuTfZRAh(3GA5_)!^RvNIMO(n
zc~lzOpv5)`25Bj%N|FC~b=^_az|D+V{w8_X=LCmt#r+M)1<wv(ASp1f;bFoc=1py!
z+G248kkNg95lGfv<RKk1Vd2Lz6$@AwujvcfPVY0Hs0z`FwHp+=p9X^Ngug9)kBbTu
z6D71}@na~#&@B~KeZalII-<C!ryLumh`(^-r$z;??%z$Gp3$<qKGkYFkiVQpDf40D
zjWcyCl1NdlGEfQY7_caAr+9acJF_e2X2g>JUO<`l=P4pKKZmYw;!o@2t2o^WZevM~
z?J;8EE`gkl5j91dK2riQL<ckJZ|Dj@8z+o1dL@D>w|e*<owo1#Y1(u<>Az^UEz?~-
z)T%s2R2}+k+kPwJbU2DV-xFZkdY*D*EWZVvifv_8j?U{X$L51AE!U6FM$`S>!h$~q
zB^Y~x>f`td=QxHC<|dtikf)z%&j;DDTRqCRF?N{?Y>8Aq9p@M`6ofM4-M$C(itU+S
z`Rs5_O_Aq{>FR60{N6P0)7wZ{5+I>58QUeV+YuDdaR6T534o%%PqZazraZJ^N{an4
zEH1tVAkJEel!R_<D+~734cGtx>x-0MixO?7B8l)4QlYuXS>5MwIu6jNLTEee<oHhA
z;XfEkX4ZY%&ifVuTW$X`DF^yKUVkzYBxC8>WHP*3z4CZy`mh$PD>IQB^f=KFyON8Q
z)%r0QSFu;ibq)M8cDjQ0`}ZJoXCeYJbGzDvyKnMn3+<-4I5ZwYVtx@g3`XWqAQMIm
zC#DC2lsUjb@!<o@S~f!R1tYzbZqdXxE`6Q1a0OA6U)H>+BYqLkD9SEsHmiPc0%}Pe
zWvh17Hxk2NE^2{wzkO2t-jGy&O|!_uLc4LEChT=*7>VKuPLdOzk5%c!bBsu9E4&j^
zW;Ms&=v&;$k#T-!jCoXDY&^+w{=j}kSe@DCRq9v+7CJo*9+pR3?7lV7YLYF4C5r5A
z_yx^2G=0GTI0@##9#mrW=BFJRIp3G7ew?13!-Dl4kF+DIWkpuz(lvS3SMF)dB<)=(
zF9yo2T@=(h!`Xd#JMc_tI-1VP*Mn|t?k{OIYs>LygGwx>W~+ti<a;j(?$6yNUQqXA
zd4V51JHzgC*zQTqSHWY2kH*z9?g>mT{z%B#RDR@_<2Xb7vL-8Re~jxYjt(1M*n@89
zkkvuKunz(+7e_xE{m}gRJUGuK%|~X4G3+ZDWyqt<D&bdO?TEx{hqc@*b|5)1_>6g-
zz=~&}M#Mzo%#MhX_XJ|C!U4>zD6ptwCnBjknxDMrAu-j@{}uzXVy;=G@SDA2kV+W6
z|NgVF5}|bB0&+?LHQ#w@O&vq%OPqDqIBWHDW)jrz5dI)VLkg3tS(T=D&V2eDb(|yA
zhqxBpFW=KKPlJ8TanE({5JL7oKB7z|_Nb}1=mraST;+S-SZ#GOA>v!M7H;IxO)Foh
zx0@=Q08?fvl&gkORlSZ-YX<!F5k;AL^8Fi431f_=hNXLI`|R0W)VIWYJC!o`ajs%x
z{1Bx9p4>G@1wd)qarTSAGEo=3w?nibM(=P|Sz=$IO;q?E!ZAB|oR;lnBan3Cfq@1)
zhPR(R{9#@<i$y4XN|$Fa9cDLn$*v_)J7>XSQi38=VM}nk^64Czx`X|LH$nv7RTdl-
zp)X#pDBr+ZVwe^u+OuuibHp<{{3+Xn*&ai#&dnj;zjT&=Y@2NTNl}njjWueKD39&k
z!yX4q@GBM4Bfw$Cif6*BiNcvm+c0Ko?1>7)<{VFY-|Md(L1U)|@gvHkn(1yJ;d2D2
z79u*UG6-1P!)>mt^kvSUA#uus5Xql_03fveZ}q)Z&5i)szztvnS}}2*iImy>;Sa^Y
zj_1!AeO&qz?<21PLY|wM7dHBLssEpeaERPZxag$Hz|rFfqBNn}u}aUS61A5r@X1M^
z>eGM*`)}<EHf}N{a!U;36Hz$5A6}l=@DgmE&M-u5*HcOI5uyg<YZl+H_R?EaWa&<a
z=8uP4wI1hPNAz$zcmwWIg>T>V;rY<I6&Ln>fp2Dy6(71H%LEMhVa(W=qd9taRZr7&
z+S{~})Qy;$eZmYZKZ9lJ^nS4wAay?J2A5KjSkgb2sLo7xRbLD^_nnVJrt?<*@QBc#
zF{($;95~F0E0g~aJJcl`<ji2wN@&-S-enQ6L+*>r+Bo#L8cEhdo_70>nfD4H!3T%@
zxX06ukPL7_QLXA?fEmur2ep#$G{D)G>inyrpPihQq25)GCUF_xZqnsrMWe+4FzLo+
z{pJKm>vOf1`k(ZSvr@g}oSF+#Avyk+Ckg=jpkiNh?Gl`;KnIe177NjZX^!wF$PV#;
zidnN=9(0d(n=_z&#>a3<u@-hfT(TXgP&Vu90S-`b%xQPkcyN2;SegDapv_uB#IL3)
z_!an(O*RAo!Y8MDE;8&!Q@V{fAQ#5@i@;|&(P$ViXz84OBJPMF!%*#{w&11$Hnl`^
z-auju=;Ii6UuUWc;wj(UVQ2!-hS%bN`_GSEVTcF>lGl{MCE0(|ronf#>ed<|Zq%0J
zWp<k`Lua5E7IH<>sS%i4$45CI0#!>adM-Pi`~X$6eR<@1{PbpY>f5jrQkS)?31$QX
zM%&8b7lEyrmFcDqS8RB}=$l$+_8QCXaiHoXQ?RU{z#9n(BDL+*b6G6-Kx*Z&EBLW<
zFmyQp9;0{<f>NQ{Y@D6a7OBj16j?4IAXeA>(L{75<S}8l=Gi#C3jHAnPWaQee#Sz#
z&`KL}sk`cxl)JWkTh(Y%U9!g1k{lbTgc*KqBv7)#0&CyZ>u4`^9sBtBdysrjJCppw
z8A}OjVFm3MtE<+kAsC@c9Ee~;&h$l%u%lUGHh~|W{qY(a6LnOgI#1f?B!<l_)jTe<
z)#>9XlTy=vl*;-@=f<pRwXlIR!7*DKGv80N&hYxUO8nY{>Rd7W9nK(ke%AL)sGo~6
zlgf^Z4pf^i?(2QFc)An(^Hl+B)l1s;A=*c`n0aX55k8g{LC{UjfgNQVa+v)fqBzG=
zs{4Mf7R?y}ncaqO04SgHdthEIekb;FSkt@s6$ViN5>gpN-_lOQgI5I30(XOG)B?v!
zd_h*jhZVF$BTu*}9H$U<bdD!E?4QRzMzgM8#D~8XO&E$gd2LHs%KS2v{G6%nLK(1K
z!$J$&iU(c_PZxSJ8QTpkMHhI-E2ap?#S|MEne`KqlU}*TI%e@cpmGlRYH~=8{@*lO
zI_t+@LsCP3kN|nq);j>nI70<Rfj+q)@EpWl23W|^6&W`#v;B3=U;Xpf*!ZtLFD|90
za8#Gy?ygc}eXTik*GTU+!DYY}E%3JlDnZbHf+f@6n8f`(z~?{m`t7=ZRxHPtZD*_E
z?~5IQmxY#`2-X#eQ0Gv3Y;})zlb=lk?Ay$bH|(zQ@2uZw=qZ?hmUMU*);fY7iYRm;
zX%6QpHt8M+Buto3xNNWZ_TFOWd!=PjM)@i(Y{L{u0$t8?<T?f*iU%#uTT_v7+fXMM
z4YZy6gM<85LzpMLw(?8GKw(tRv=>zqrLhm&!FP8iW)6?`&2;$k{JK%gq;?$=s9Eck
zVc(`g=4>8RhtDpX1OJ@q<`3N*UuEG*CH-n^d!LzR#t4Sh;{=}WumN2>z6Zz)Y<;Wr
z6r%~ej)uL)QT;}S1YdJ9*qr*-2zgaIP5dQ2=xfe{hI9i+jVgq&(XgZPl<<Dj?bbz{
z95#DtFBXihsGY=F7W+B?*~%;@6Gj<xkJ%Xx6}^Zls+DA4hQ1oV2;q;y(=TLFADhdh
zP2lJt`_7}r(}Q>_#9^X=3V9$Y;sz(XGT$5D@CxNQDOkh3;~H0_efEm${L1dB&l1a$
zTxmKNHJ;6E(>x+uC5|^tZ)Bix&RnZxTDZR3o6L>x{uf>GD8m`z#k(x<6g*uAAe}hg
zL%}<>TOI~<q2aHZ6+}P!c~%J~YBx2GUjL%Hl1qJyhVRL~WURGmRj*QIY`N1M1qL3c
z=uz}_3KA6vwn14MZ)!wacFJZPqWK<Gk`lRm)A#&3rP;RHpm(ttepYvYfZa%$X!QQP
z#%aPoosL;S&UzKuB^DxoFHxv;;v&Y0e8z-fABwSqv=;N9?0hV&aDs%sN`AyVA0oG5
zVHJs?SY!h1d(<n$Vpm38jvyRc2>Qf($?f9`4iyXdA2R!7jN*c+=Ht>cU`K<{rs_2V
zz||0U9gO~lq(Y(u0W<yJ>7t9WZ3JWrisC<*7?@qgAM9{BA1HqNMbLm;o;B;4tJ&B+
zJD+W;D=KpYgL3nOGG7LiojFWRtipO_e;k7UFd%?L(eE`b&3(EtoML98iq8}cQ9PJo
ztfnv9cG|rCZD8*y0q#H7n0lnKbN3WsHLR@DJRY371m&7{Hn-gwOxT0g91XH>A+7;M
z59GUOzfW-&v-vG~zsfF(n<}J8i%*Y;PcX~NoLK{pCwZS9Twkf%8>&0;0Pil_`=MZf
zum)3b{LvP*$I?+bx3N>dKmTaV)s3%S;r&aV+~i32PiyQY9A#!e*)?Syb5{}RvcgeR
zxajk0Q1PQ)x|&;>y)`e4pTA&^qm3g`1xp9AVbd1}BUFfLTVw`1B`kk-EWoKOr*%uF
zKz{xP^FX;zofm<{OCs$0;nwU8XU$b|&-_)`d}}(nh=RFs`lBv5edyAxQ-3?2qv720
z(A?kBT%^0mhmDbjznSI1+ph>~wezNFHyziT`~B%-fdEY#oc>PB{fp>0&JU^;bU_{T
z$Bs?>L(gyZNW){nbW%tvY+p}s$p<SlE3CH?5NCA`<N~Gff@<_N1<YKxZ}{T!ILRe%
ziO=7l?j!vYgrV!6dR!5-z!LcK#SLZoI`k&CeYfwcRhzTS+lIFL!a{C#jw&zWp7PRC
zmROUeG9H4*tH9^(7evPvO_sRi7W@*>*58lZ|3h4J0A&q)gn>qSaS{4zXiiA$a%q+H
zmWhnFpzhTA!dC(`)IfGj?iIzU4-!|m^*URpZSH**D-9&Y*nL@oMED06E#Bb`Gd|d@
znD75mm-RKtLW|vP$65GIG&xVYp*zD#arbRn6BF>n(S?IRVH}W;{Db`IkD|1RLVZnY
zfMfUjk@|mh=m;Pp{x5J<4NeuUa(~#9>hT#{i&{Z_ecE%AWQl@QG6IzRbCXKr-!lXW
zxPHpcABt~6Pj-PMY~0XA7%*~FIe|_I-hpBVT!!v4&;NDBU;Xjd`1sG94fM_R(MGK5
z?^Ewe+#!7TYIK1xWt$)u_M1^h2$w6)Na@~hr~i<v^S@@WA@R4l#xcs@i@l34Kl>BF
z;lBqqQi}Xb9D6SCHv*&(`T202mt2Xp^|djAl_~-Ca{1yPFGkjc8KRby;sg53ob)$G
zoT5=XR9<o)wH?Kbk6sVbN=m2?9tWlx`*3TuigtQQqhNm<&_V)VjNkx*HQ7$x1wMUA
zCjfQm|18%WOJNcq)TF&9{VnDDo53q1-8Mj7U8xX;z^<>-q81fUOO=96<-{{Y8J=s1
z^VO2dWl*3-RcL<~f(Ts{Bm5!dd(6D~OCNSfsO+s?msS}v7WRWD>v$Sn>Wiq2U<x!M
z^Ozx9B)CT`_1Nou$2Z<-7nx`dqfF+Ry8?VBnZDp9nG&vWMT%Ts^?MlIPXBN6uX~G~
ztLr0p>ucU=2sU-n%3JXEQW6CwHQrs0LaR9j>Ro)O14lHvCHWLPSh1nk8>>KNiC7(+
zt#kgZZyt@q#1CNK-@oi_*NCu?QwSZ+C<@qt9)R<}A?gRyXZT7ld1!}{kk+SB7nfiT
zUzxss{qPF7YsDCE!sZQo?*t>p&t4+JaYof-Nr@QNp=NZ|XrW&vjsulhVs(ed@kVBV
z#69jbBED(Qg!`5x=&G-6tIhWScG`f7)_5Hx$|qgcc#f_S%@tYatR#xZJdF^9L)@0`
zFIOc<Xy1p6UBC5w`DQshTBWuwy>l=;SbP|#QZi-b?5~Cou}eYi#Tak6iR>QT%oV>E
zz|tVz*Tx^88W*5iG8^%=(oC(mpw#<tA|1WI&W03l4SE7Cov%pzn1`Nua{=oIbO9sV
zC)G}RJ6VahHg)i}?Y{`}UJCt-phhnekl;g>7%?tjY8*FL%;ZCv#!NNLliT-T+f8T<
z4(CKPI2%whv?KT@wgDkbi|8?Zym3#Dl`Lp|)pF)*wN{B`k6U!|ux#J7Mql+c#!re0
z=Qd`}tSWI2H_ooMl&`ByR&_9iDe5p*Jq?h??Lb@!8pR*jqzqWxalji{cLcvQ81MHw
z+bCMWQW$^989c*HZrZn|RV;BYE-mPi!5E^NlPY()-1adI_ON5i^OxO{aYmnUQW!r%
zMQ&6rBk<*J@4z@MM>BD~6BlgJpn(*R#CNChDF>_vVr+m`DKBqxo@!@TQ7j+pZmHS@
zo6N|q{482gxue{nc>W=M8)2t>|1lP2jV5Bzke~t+<A0c^N5K{OgNm__MFiTFC4CY*
z8grfO0^VzeYDXc)+@2bRXw!@K{}1-wJF2PeTN@6?f(i(NbOoh~G(n0ACMwcJK<O<<
z>BNBa0D*YuML<A6KtMW3iAagG<j@5KL`tM3BuE!Xr~yL!_V14GecyY$-|x0@@3{BA
zpZ`F{9y{54&$ZTEbI&=S=Xq}#bbP`<(vfFhp2kk44}WJj%*6GX?n9){XXl2^4#RT@
z!>x9_b(o;D!vdEaPsF=q)my)iefL<%o&;VbOnJE_Ml$%9?AWRsYCeO{BUCkN8p6KU
zOfJ1mqxPM)Y_2z{V@g-NK7r`Nl$C`rBeh0DWRa?0)NRWaRS4%3%_G&C)UGKXK`=D(
zUwlzb<ujYIo)e7xTyi0~oM_r}4a1M@_D8MkF9ytqQ3mxW^<fe*`ScqU_oM>dr|BLg
z7ME|k4;M%m7Ax1Cvb8)DW`E&r(r43t`_^hiXuEm!sQMWaT(8f<C*?tX^9%j+o0s2~
z->ZCbH(bVr>B)>pLR(YdWji3V3>^)x{kplTPr_zYBD5ykGNalDnV|@~kqcpW8(73&
zT7&1m3XJ5>)-!Xr6iqB=3T44dmH_cUpF{ANawLnjH9q!Ik`}iy&(txH59o2@4Jjf@
z!EZwW`&S^vUqBR<nNqC(JglPL<qhjUi_BPYiUs|L|3pPDugyv&>86W8iR;pt_pj6>
z@<gvdW#8Yr!us0p<31j0K>#3tNvd{#;~q&o<QW0|71T2H6I{hw$>_N1Pmq<L5GpmA
zAp2iG@>kT(FPi|+reE{qKW6P5ll`JtmccE4ndPm5BU{2PG1lk1f9@L_{$Ez|J@3(0
zURzDwNmq}P7TIgy=-2>A*+?dwfjitw1QE(WdO4yB%DH4U7^f~(Wn-mn@9$ga`h3No
z^~ow{4*(3Qo?}gZG^+9Fql^}46BUNrFs@BJfd2sH+co7*@OBDwG&AS7q7nL*l>VyR
zofe2;a<|oIcX@GFIgiKniZKN(;m3>>>)XPUqvi#%rlaCbeg^=Y^_g;2Wbi_pFpFHq
z8_m4s9C=DHI=?4CMqn+aq@g*rYBHFZDOr~+rM{%n<OJq7DgMNyUyhU>s;aY=)xSqI
zPgG*NBP(&9b8T&`zp7*e{@B^Ua;~X1(_lR9fkS-Kz-IE(4=ddrI`m_nB=l*-tPN&_
z4y{rmItZK&^FL>PDp{+dO*|-t^G&Agzy=&tll@lkiQT!wH|lKU{=h+<VP#nQ2i)fI
zTyM`J+sSb^@v}|A$;&!vg0C;vgv=2y7bY+%>Gk_t{Q#xxxvkZY##xlNKmdZ8a<s<t
zVa9R9>F)NWf)sLGmIYm{7~!s(HZY@q&Wy9U9&>8xx1<{%4=}+~`SgO<i>mF}S07Lg
zcQVMrJc!LsMO(sGySlX!2h&-V^pl>P+Glx^{*1L#Xs)<DC1gR;>+v`+MjX)|QVLMc
zJsY=D-XcV_N?tek>@mh(-8&HHHx&MIvb|BYPmh)u|6TPrKjqwn9BztVQH5p|EmD<Z
znX<UElXgRMDy8UCn!ZmOfClCij5)7*<JAazhJCB&rmLi!vm{U?7V;Qa-`WtC#!iOl
z%C|~~I1-~9{lchM+f;LeM)1}c01}enF6-0K@vfh>OzD9f&GMwBs8)NqR2LwaZO#nj
zksh`OhJ&TxsE&~9RFyokpqilWw)6_S2n<!-FEjT3mTU!Chj3kd>hzbBnjax5YXnpx
zfjO$JQ9GLJVRO^_(c&~h??kt5a|13@Cr;+0%xp~65iR4eFg~f;wN7tdh%hf%9H3>W
zg+}=AJ5Dodl~3lHj8&|u$MMM&>B#X-I`NC6z!tM7u?oAB{=G*uv!@>5sX)hm2~_?9
zHZ2GRef!lCEU<)0yP-3JgwC#7n-BZe;vbE9Jc%1zkdbG8DZ=x3T0}u-(d=4^THHc_
z{Jno8m|#gJ1b|;0qhO2tsW;SY!7#P$>N?!e>bF$o7t6(x^)*_VzDXGw>WSH#{1@K8
zaK%GwC7*pqUVT(GSYaccy^}w2_X^5!I%^Pff?)W*t~72?x4x@G>m%RXt)uTlvx|Hi
zQE&C#Gp!9q_Z5Rtg;k7IA`F`6O<JvrqTv7B=Z)~`99tOP3cB{a@>|oeoPgO~qc;r}
zCR~4fWm>gm{|Rw`K2A$Nu$ZuNV{p|2dN<sWiEmvqi>rglY=G8VaRGOs%Dpl%IIEq$
z)-uESmDtnA8*T5yE;|n)9S^h+>?C|nCZpP?cmU|+Ta!_OO@5HcbCf|prD%#wu7Of5
zE%p8SZL&h>jks>;hNG_1m*D%uKVb-le#T9#m`#w?`Vb`%n9j@0TCMyK6xqR};*!1v
z;!2}K|1!#=M)Tq;;!jYA#J|EE{vihY$qxSn+R%TN>0fg(ko`}{3&>wl7ypnv{I}l@
z5OyMR1^^|;5&hHo{NIa>pau<^J@HX8_8@jIPupO7#4zLTbP%Ii4zgu!#PqWb{|WZw
zZ}HQA%8mXvv)ixR0kYQr04@3-Fv))n);}7oWAfMCxgCvKZ8boP#_zNO9xyR~eLnPm
zRZQxDYU3rX?}Gp;1!jOzj_Z~X(_g^P?~qQEn?Ui+@JPI%K^lQ2@9k3h<XT3miZ!&=
z^FnW#TI}6FJ)b8fGmE1Bg~p@;r(A#*f9+-hszc~FZZzZyLZkC(zk6m#Vv$?g)voUZ
zZ(XmQc>dd%qfG95J?G6qn~8jsZh&ordLAo|;O!8S7?G7HwqrJ@h;C@|>V3*`5WtMl
zk+hU}%F42SDJizQ>Bv*I?*1(Z{bC4E+<$W#Ab)t~34CAar)Sjb9ovAd!x`x9arTT4
zA9le=Ao2FSlbkI{!nTn1&|7{jIGQtT=3}95kwV<Z&P7+gSDSOF!tO{AhHZm#^|KZI
z2=%BihwOG&^RPi-h+GZ6_oTxnc<^+%&yGz1g5bFKY(T~PMTpf7gm#AxnD5qkpa2H~
z+TqDA07y{2ow&sW@Nk@lFeGXhkOQ{RA0YAmy@Ur|E&Ih-S}Cqobe_oNvKcZMSRX-d
z(e4o9t{zFKA#1SOiH|rB*EUr83U)T0!V~Vsc{~lh_O2r8(;>O++7#DJ?jJzf3SoYW
zmYUaH&_hmoON)h8FNXvd<rEbYm)Wk3%k_wbU)4BT{n0+<(tY!V-a|=%JqsFzTRZ{~
zXPiUP4TLEXv>?W_8_o=-ZeZ9)G(VL4nc3;1iTKgj2lFlaO*mWB;D&mW4b7j*+mY#>
zgm$};7UxMm6z&kU9j?799WO_af8XQw1Uqtp6Z^o`VRM30<6TFP*5!l@Kn-&jeBley
z;O&&F%6<nySx9PB=1EDKeh|W3FS+$%s8-B9bfQ$;d+pb_BOqbGDt|Fm;?K5p?c8WW
zuK>7&Z|1p#b-A_O<;*>>2Ls5m>sl{V2JJke-Y+YLrBR`a%e!M><@)E%U!DnCH0Z+h
zAIrtQz58n5w95L5jcRwL*DMK}O9hy!@7IvBTCBSQ@BO|cP&C>h&}0kFKA+iLGI;{f
z^ZO|B{I}!_F@0+TDtJ8cuXX|=roWg`F7~FZM_LszNO>nQ!|*MfK54+voaPz*UK_*`
zpfIm-)U*DE!=>Xu)$?DPtfo@eRV7Cj!c9c-8wHjA<?|m{)@TA@#a|Y+_CL7@(|^m8
ze$Lzd4}FaPvG)IU<Nqy}#jjxr@z{>nyxNm6ZpLSCksp6Y%;nAHQa%vVXC_UM!T;R4
zv((7+|675U8u@!PHUFKCGnxHu9Ju}l#bMeZc0L^80sa7E@b7~J9|3p=|B)iOQU5>I
z|9z(N@0AVtd*5sO-=w112LI>dzndQQ|BEgDKkFE`(+`lvi1GjQHj7`={nvE=*DR4=
z)BWs!h1dV(-T(b!;{Tdy{rArKHQj$r_rDjk{JrD9f|>u-Kl)$QGx{r-`4!CkJ&*p6
zkN@)Szr6e3^WuN+_%HAN%e!m-o0iS5*!@@R{%>dF`QHGTU%||;VCEkOEr0LxuVCii
z<dy%UvN*rI`+vQ6zs`ORktWZ2(qo&=ZT8ORt;5T^Cb-7>2yw>*_Dg9K_FlKlyd(I*
zhtmDExQQUrd?{Aw74jC<F(NA3Id)QSWZUD^)K?2T-3t=R3CqDZFN*)U7ZMU=R<f!F
zkizGZnGd6LE&xQQ3g&T0;lmvIL+Z+oIqiBrt)y$;M(DTDZoN@Kogw;7F<$)1Q_`JB
z)4gmzei-=j=|$ePSr>9PB+xL+VRlzl`mo34`oTK`_dV=Zfk)?ZFhG5IV)X~;kaZt@
zxL*rUUY|v^si6QUHMksrQddG~ok&4&169P9d8dXN6@k<s_rtpE_^IdiJ)*XVt3l;)
z`YawP11?XR2c@#@leKI=PU=Y+otQTaLvc03g2^mPSRoqmKGibrPk(cNA>p{rMogS|
zma$Gu-3O<J_jhYDirSQ`Q&o?&Il`{>=;cCoO|Z0v5Iq7`x9&cBCXwKIP*|{sYw&*$
zZl{(!S3*TKmPGmxIS?zR!|O=>Qn~aZ&7Erw2L?Syqc_qzBti6nTf3DWY(0{(6+#<*
z9d_cOVi_~iWv){85_8wYv~T7sK(&s*PTv;6sVG4O5zbv;@xzq0LRwrEt%mG77@Plr
zDwkRg^H+sD&K1)W=>OX3_CTZBl>4L&MO!qe9mlKg7(iJkmD6gv?L@yYgosG3u?nOy
z8L=G{?Ji47Fx--MZays*dK})X;|KyNs;%^nGzjgw9V+>ZuCJ)B)E~_CkHQ`3JEc48
zF7SojkZ_o4n%#NquuL5^n#(mkv#^-ZM%b|CD{Okf`ZC6f!{PVqZQsbvq4=E!+M@tc
z9aRi<M%_mJiqi&rv=}8~D$os=F-NMJyw6d}zL9JL3e4PFnLUa<*^E2Pq-#0?Vn$$b
ziYewfS{5KMkfLgp`MQWE_D#K&v;2LNLowB|$LT3Q5)X>4KDdAsq~><Be+|U9>Ca{Z
z#E~b)s?e)zZ!naV!LmIF{6%k^4?f3;7<zTb%S6>n)#T2`LqD1oJs;oz99+O__X(<w
z5f1<y{#Nd0^{y{A&foW_w+6XI&&9!LNzB!$w^!@-(c2$I0EM1#zaJn*WUx8o&<aJ3
zrFoFMqh~0pT^{FX8r@J1qyc5N6?~ooe;oockhlCW{9qpdyuCjj5j9c8scCZN%9F&I
zn}!OrhB4q9DWu)DX>&4fM?<rGW&cchR=-BuL{0;WbQ0}o$tOO=yF=ys*7Q|G*zx6+
zKKG5TS-l6biM0vD5*+Oyvnt(u#gVUfS&;yc2z){9$!F)L{s7$p40Rs&1C*C*3MfOG
z<<cKFKzV669g0v3DVW!mI#zC9M$4pDD7^xAv*x;Hn|m7cJh|gJU{kt_jzc<9Yr4!Q
za{t_`D_?bP7iQfyJlW$d#!#f%;v*fJM~~b%+B_-)-%a0071xyF>}m3soGW(SUx+~h
zNLALX0J0%x>+H$I1m$Dz#`0BV+@EyZy<?{z2lLN()(g;wh@Il&O0m-QYC;77tyzC;
zKyppL5IxxrUK6K107u~kl}mIERv)YbizC{#X%K*(Q%Bu)w$fbp3wW5VWCFNhVoPe9
zR+D|b(_A!v7c<h_{AUlqy?2vpPgX|9gMaIn?=Tcv!BLtDsZ&OF2*q{thFBJz<VASA
z&u4~SNA<BsQuopz$Q<@!<O5nA^*AI(nEy%h?b%mI7IGY$+qLEk7>3Nq)-c-(&mC9U
z+Boa$<dHG=#HJulownZvjK`#~xYRdWYllBT2EuTkmsYReG`eoiXX5*&Hv_BoLja1_
zpPZ4m6(**(QfS5OeM)y`>RNf>r1<n&V$tLz(M{pb!igUshtTo)i;4Fx4!Y4`vwkhJ
zFaRDA@;NYqn5b@9K%}Yt*?7O~k^P&M-Fg)csP3uS_b0ehF1*i89eLXAwNS!Ez1qbK
z7NVAP{s8e}Vzy<EbY*t8Rq~n)S3Mf5l9gy^ND_6Om_AzY)acx*(Otyj-!@?UhG}xP
zFLB3=(T6I4Oj~kdi?|RS7Z<@;cfkqgF4ir?ZN))WPvA&5M%}k%!2@asEdpHwf|U>W
z#Jaj+e3h9q&NtPYw`ap_Wvdc{0$^+%ArfO|*Nq)a(<UZVLC@nqj*(vOhccZ%f&P>m
zJ3-H3rtWq@IW)=H)*4h_VDby>>NPP44WaYPeirokD8;agMio!rgzrV4@`AmmrBm&@
z0O$y3-FZkC7+|#vH*eRU2?o!$t2U;``wFvZtFJ$a_x4M+n(W^>2!ECIt#6V~Ys$cF
z>t+5vl;X?F7@up|nRpa1G5r7)2=u=8>@7BKAD;Yq2}4Lie_1iuleD_oK^{vQj8wPq
z9DK#v3W^CWZf*t|;p*$4%*ExG>;a|iJ@~`mA*c+C?rfK}e|ltpuxKG|(vm9q{kkvN
znk6`p%?o;b+Rz;<=6fS16<&B^qfz9z(`CMEnp3ChkGoy&8IAy&`2b|e&K&TwxciOM
zy!XLhd`Os^I6+kIR-OEG$md-4O7*ME(SqcJB8d#&w+024b=cP_`Y3;{C&_9S$)9_?
zBagbr*cv`|l|{K2M(qM*tj^Kh9n*GfkTwKkI&xi?XSdP)KJP$na=)|93(o`WQ_s)%
z{t&@o#yMA1jmP(zV6>Z%DCOQxS|nVOd_1y##WA0t%j0Cd@LI_&1#CVaAI5#*{qxLu
ztQ2D90hPUtIN_|Dwhj=}et)8_3Tq-4#}%|Y%5D+dnCv}euE&xJ9=`6nWy4{@72!*)
zHhnPKojKhyirO5S^mQUEr?209UH<@h+@3=G9NgS`S{$<|mtwEdP&B@*Vka}|RMQ+G
zcOm6toh`L7@?n_!iI1Y8SIz#3L&K?uTtMEZx~EWoTi2pOVnrz);c)uN;vSOjI?}Ne
z@t8TO(*EjH$5`4X%lK^i+=ke9(22=-Kk#*0dD&N2-0|;76gk^sB`_-TTT{cRU>C35
z3Cif~k;EI;5q0`qjndWO$Ih~ikN3lkwb&+M0B4sTJ($=%Hr}@J+<H0XU<AsTUMbkX
zP2wmCL_ivD@GYR5UT(-xuvth@*DR*141f1%)}O7^*L~lDjk+H>Zw>QH989h6s_J`k
z=eWBn_$()rs0dEAS(Hk&raFg(oREATav3%uA0a^{V2a!XM$s{AaDku`1ram4>$l^-
z-5SL7rvRFy^pcW17~x`iSMGcCsx{0T?Phm{1}&P#&@b-qVEn~rKlO5Fi!`Rj4W9Eq
z6`NpN^A)4l*_Xyyx)Y1*oAMZ%u=02t9RK)b5g*jwBYPe_j~152_ieE-&H&H2&%X^l
z*mea7!!G^2KKf~p;L_#XlSuA>ij%x%ABJR#nz?|j^7@-Q2OrL@Nh;&bMlXI}xCBn<
zQd(HO$iHiT<(~{%B263c@r%x*sP$1`CGu7FN?+AT#fBLTZ5y0D@3pfpSY97$;ckyP
zW&Sa>wfAn}`Hw9tm7jlr@KZlPjk*2huc_Mf@MfUZ_z@4V-RZ|JZvE#8_9jt_i{uiE
zmv<>|mB#Ad-;cW>t}o;G?3(EF*vcIUCotyLdVqwA4UV2<Km#^ik--?qcVP%4M)?~G
z;~%4pIuX)?I^!G9Y}zuU470wLN%#S}{~;)VBtgEQrxULDn)BMtlNHN`<E{N$N+rUs
z%BL65!HFHXEuJqpin%<n7aIR`8R*N9n}-;67@7TRRvUqRX#3EwefC)VVJAA;U8FgU
z&5Y37$%^(AUE{3Ub}^~pu;^CY@P^tT!G{%eLtoM~q+RGN#UAkkWJAG(`m}g=WTv6f
zg1<TB42fIJIz>KIA87Eh$X_UTs!ujvXre(06-3qI#D(H{1$Ub=+2K89alyc8bn|}7
z_YlBky{QKtI<x_`Cz4oAOJ{^mm@`fR%lca^g-v*8)G$UMv3**QgbvH?0C0<OK2qvh
z@a-1u14ZKf8eIV%#ZT`BEI)nx$jy;(^~DMF85-|_3z4o!v*bifEF7435+taM>SD`8
zE<|)|$<n*FA^xNIje@<Woz0gcp0A0)jWe%9A)-SoyJeRq*F$1I`R|&fVdxDu=w11a
z1I#~vgS7Nn?bTpNr(vkSDwt=)x;^9q&>x>zo6Y+mw_RX&VpZ;OvP~4GIb1UPXju-w
z38lBMQ<s7>!mE2zVY9ETuu)WKciO(;b9b2v!8XSAV8yXEQ0(2`Bt`)xF`bEUkY~>5
zYwHvkej|(?fS*!zz*TKR5A4BX2f{I^McE?MOcj)&Q}EMGVwDaHuyXn&S*}@hxVo}%
ztU9EM25l7x-V)ft%3&K)?<R{xSmbC~%p1QLE>92fS6HzXsyy&pVjOFZYo)ZmA;JRL
zBPx7t%~m=I(usEuOukHtm3z22#A<k}NxNUa^yb~M4aC_3>G4&Dx8d41RMA#bO`{p3
zWpFTC*yLZBc{wN^h^t{ti27jxSO76R)3R0+oCkS^B3*8z7>s_4;VEMEzgqN3=#Gb0
z!{kOe=)S$6?u8=&BJ4EEO*%1}Seu@JG?<C&*Y679OC(L_l-ZL|t0r|3a*WSBnx0B=
zteohJ(+y09gc(81rfWsY#@swzSX6(InxW=aXP8DnPT|rOLvq!E<9=Eb+F$AVKR~zP
z(tyAapyRcniEIE|%(34>hdZotqRL$Cct?Css4J8rBByAOuA~*C7}<-i4mAzF^@$%}
zyyf6y=4a-w005*6f)E%8M#7%J!7Q#TLFConP@JyO3`N0Xh1MAUtnW}!jsmw*<hsr3
zUTk}EY!d@3&(MN;>OwvKp+CQ1_v`-PPXqeXve<JXGEQA3O~VXmUX6Yk1xq(@akVjZ
zWe}1gRlsH6FPqh?=k!?9m7aN>FCEKot2d6%*0&7rc+;pwAw#*Wt909<36-3XVdePg
zj1LJj8lGHmUtN1O2aR^=1b&NitmL|H{;>KVAW3K1qPF6Dz%`qq`X0XjXXhN|Qt74k
zf8mTbtdPI1Z*Vsmy3gVYrAep%03}O%#e#nQ#4lTXi~Th({?8ux|7%;YvoG`B$jSLa
z;LDtobC(@|b0?Z$Gst&QUxEL`X~n=HBZML2HU_eD1`7mn4K8m^2}w~UvS$vBRwi1y
z$Ca|>?~Zu5uc_Cg<oG70r|U~)zM4vcI6eP-3+ATn&aCsn#%?GVHjd?%5(|19d74Pm
z+Pb`t%)L)otr5giL+PDH$skF-`38>(7(*ADi2JSy<VLpcnMy`5cxbo|UcxE7x+}Sb
z*AhdSIS5HFaS3G2e5xw%Mme6N>%OKRk>GlTLwl3s{?lV(42c~xx(7p3*Gd8A-%s<J
z+{AC(%x@UvMBe$_ELpX9L$1L4+Nf`joivM@_4PG}6sPI6HZc!5B|=tVaA{Mdg@bvo
zFo+?6z<1cGQ+#vEDi`0;GPRn((nBbc+=%k|TTHAG$p%st9XI@Xn)55fG%u@2>&~-o
z3JFo>-}^wrbFO~zg`P!3k-3TO@&4xZL4bw}bSh^3i}aCG(E|4T4+g{d4Bm4TwR4;Z
za@(0c{1S+oQA_39>}kE}2WZnU>cM;YSlJPEPY-93yg(0Sn0&l}oUH<h3hZPwpggOl
zpgqNIL*4v!k3^dU)gq_83>pI-A>aAQ4Vi_AE$%10#Dc92`4PR-$g7lIV#f`HA*Wp;
z?Nx>{KU+u1r%%mY1$~bnU*^{1zI$roebXlWI{F8wuqm*zu68)CjPJwtp$!j?CV(p-
z>K~?PB2&ngz7Q50#DNBex>H_gIYqO-Ggdh2yZT^*tQM4kc7O0~uXpZyUWSN?bJXKo
zCzI~L(He&-NHeSig*RF~tGx2gSf2%;G&M`MRM&L8vQc2aBYVe0*E#&^nS=gL8=bsp
zqa!Z^?5{)K=gt7Y%JYqTA~z%n#OEgIGuxYMLD}3?|29A{K<E;Nrf|A!79}`c5=R#d
zDqGQJaK?Ru#muOHHAKbo-b^lF?@3&h6nYfmomh^vNVlK`QKp-3lda2x{We?E%$HQq
zvnz?GET^Zy+s+u6P7Z9cIX^w7{Q}#WQ>;-}AMUym3xH4<x5vC&;NdjAilLvyJzPa`
zArkLXP*KVTLFOM0BvOke`hxsDt8;rJmE>0IZrevZc*-Ps?o<oh^ZZ=%!?sCs>JdQr
ztc;&I*Nx_MU^~VTMM_b5s>y69i{>22A8CP2Rz>u8RtA8uYH7uAp9jCiHz&~3UadQ?
z9nV&6(uic{NX9w3M*WEVngp%H<8+o^J)pPVUEKXfeJC%d)l5Ztt_#m1)25o?V9+i7
zt;qG#hMsW{dK0#Vd->4AUbR+WWxrJU2dK!pe_YtfOe>MxFKstk*RR;NGKd|L<NU&q
zm&Ny_k`f$kUL0|{ZvUiAbiUi@7HtP2gb0Juf@ht+ARqls5Mqo{MvICD!;y|CbW6ps
zo2Rx9@54-j%z8}0adH3gP4`n`tu?dp*48xdGUVN;_KjCt*N`5C<14BYRXAVAZ%eax
z>f5K>^im8tGt7<!3|P;ry*lp11-JbHs{9?h9<;wqZb*R@0bw6%<<r5PSlSTXv0_E|
znEU#A$VD3J-e=y<peG)pRKHTCm8^^Kw3?I-5dyz4@^B3!-J)w6=6$ZxiI&wJWkeIE
z9GHKgR!khvtX__WBlTyqrX>%|9a=>)+58B~t~;0PPwDfkp80tG<x5$O_*7y#8JO=!
z^!F!?o(E*=R|0DAKR|u^f$|s^bS^^-k#0JIbA+-kr&&csBX5>2KS1~jER%hsOk$jI
zHM;6}O_q7ZrsI3J)S{(B&6+o@sz;E5rHGvsL#lI}81w{<*uEp6*|HwJ?*3RmsHQmU
z)jX%3K`(B?5p>_->xqq>$K^=}+6G&G^vAfJIa-Qv;3@sG9j+^;Vq53%qJ?xLN^YDw
z#I@A5F?M;!Dk8h?o8M0Tm;R?a7&-e$v-%T`7UN;Refr~;w#+Rv(S>+n@}~_%O&ftL
zy;C)jY{8IRxw0(q&`!r)bohDT>$I{P&eBIN{E=_o_D)mkMKfYBwCW8dC!55kH*7ni
zlpSP5(GIn*i@oVYmao<F#a*7W5BqegSJY)RwCe1=@$SsY{MV+|lC`hR8OMN^cXUQY
z!S|`u?L;78?fU^@UITxlU8h88(az6QlMGK!L3Q&q{5`#0VL>9gJ4$I97ehtfmh_rA
zO$d5}o?1|CJCUOYnr;H?fG#TsR+ScWI*bJ5FZc>oZ=^@2&E9r-0&WkS7mVp2DGcoq
zT4IW3b!XCHZAQ~2u6B~FM=;coTH<jWNjY+u!F7CeO~`a;AV(}I<$?320dA7YI}x5?
z>h#+^*6>`2tq*ix57<C{vt4(dLl(XSFUTKCd)S#5XQ9=CQJTnr3PgTTZp41ySGood
z_Gm5Dt21u}kY0^OKdIafR=1RHX<B(&V{rD)T+z6XE#{S(%!eoE%Gcurwoq)0vv!AN
z6?87&!*Zd#w={IO`3(n7gR`wNyD%ebez6GrfQ#*@`kX~y^%D=6#b#j0<^8=a3<{8h
z9;A5CaV0^)SE)y4s$!_Z-62d=ry}3d+od$Qji70;wB?ycsdI^akFGd8{_}K}-saS3
zVjN)RE%|r`Pgr}1co63E7=y4s-#ejsw`Idd=UsoQSciqw?@hWB*H>(y?TOqgl?PV{
zu>G+D?{cWdqtYpB0?o0osWEL_$le`JSSE<7=U)`4NaPu})S&n7L6j!w79*U7=UYi0
zyoLMLbX9zvhC~XLbF-6EGYEv)a0-x!*L=Awftqf;2z_;q?X4coX(kPy-VcDJc8FL)
z(H*%GX)I&CcjpIa61|)I=2V`^<!2y1kYB|sq-Bu};{AHN@wEHcs!ZO9MBT;|Jk&h-
z8qrVXb*&=nktECuCjLE-jvI+F*z^6faLKz=$ryRK7-b>KH?FT~a<5LXNjZms@D(ga
zUBRS<^b3MN{&`pGrRkC_FR)Me&a4bhh<2f8JkCxdnlyyCznHjjt)vU|2IRs<SY-UY
zbw#i7^D}*w`ZyDM{B_QD3yOuoYX!z?ncF1j*FqV``u6qL*Ak0e@?$?0<o?|p)q=V)
zAO$JU`UCV;6StL2ZoJ8U7(94TMxXeP2ZNEY;;4@`WWxjIL4Z8Vz=7^WGb}`U(HyI2
zf_5KTYQOkKrwPW?F5Jz&)A-@*!1FhA+NSDGb$|VqJ<?^Y*$}{=M>cl;0EJcL96%F$
z&S-PvCd70m94w1;F!Gty4hP>v$khur=Vm(gY^>5c+m>~^y?c}szjBch>d1-bXrY$M
zrM?i}Mg2E+J!~BpZlBOeDhlU|wzRpEm4P|^;fMs<E~FEDHi6{s!_JLpYQ?9}LWu2F
z*#lt#aKYvJr&`r_oi}BtPW~oGRJ<q%dV1n5;TUZiUeG>!-V`e`MFV8ZutJ{~M(fVY
zTMdg2H|swVvwRHht>1amwphiSA;7T=_G9(inP6Y#FMtBPiIe-x(dsOxmbg~NHW3LY
zM6H|n?qn3-X8rR1C8w6v^*cI864R@c0?RSOlwI8YU^QsonQ7~mbaWHLc*}3uI~`h`
z8~S8Toi~6aYY6Oe<p{%j!Muhe&1$~vuXo%nGx;Mr9*R9{y85kZ_0xSApx^W^@ejK6
z%%PypaI+@)V_iJ2%?Ik~Y;ryJk2)BvtgV9P&78J$;TFF`e3*ZKpmJ_^Hm7nE9CdHB
z`NF%!VZ<MhdeP{;9gfw_;@d})XPs!fqHq*Jz7NlbhYCH$Qjq(_c_g$T;v(mY9!JY)
zL`jzS`ROU64&&yYyopbf4)L*{qC388U2<z#^MH)@ZO$8v!+Z`Dq%GxgDYBcp4(SE@
zZ!aUFc8mdjrJ0AMTu#JbSBNNERV$f!Wc#L`l>0r(2Zwmx9*$0S;bn3R^K8E_Q0N1n
zTFgBO@cz4HYL4Z31N;b%+fCaoN&KSOFQqI(^IyC=5-AhB(#7BaXi$MfVXW_=b~gqq
z`B`Rcw&TI%5@u}Qe8;vjEz9$qmqL)Ycuy+#OO>TlH33<-jUro2Auep31H(oi@LPaV
zTt5D8Zih~0p;3*^*IQ@%Wghc8%&B(23`QSx?gf&$ljm(s8D`MJ8z}h{Ib(Mlv{^RR
zwQjEm(6Y93hETb7v?-{LmLRfF<y$waZ@l1ZL+~wGO}J#lgG(2me*7F66Wbt~+!^(*
z*e&<HQ#j#biXDX*N4K6aa9Yt0171bculqkqENwk?f)o%wwo#gT_L1XrQ4r#$bQ8^$
zdL1ivJ3y3X4y6pfsMZ(WX|q=MH`XaEe9;eA8oD+Unzt)2_|n0-&4+y=uB#D~;A63d
z<gU_vvTG;x9?q;A*|X5h!<#mATW+`Ea->p#>8`#&&4SC5?v(WLz7)9Np4&8_<0R=j
zzjmg%Wt%{Zz%yP{&{GplfHIU5@w5w$r}-xJjIwRWE_4W~1{6X+K_+;AfZTLf>FpQO
zoRjXozh{|X(Ua&#s`>%?T{yzrPMnreE_S2nHHt$9+xd2>KK)E5jD4N+q0A}KsRW1T
z*ZUN|<)$-pFJ1}xY$Z``UkLDE=!_vBOhm@<MzabyJe5hSf3VP(Y^+rMyxJ0^DbvwK
z_p}O;qm2S?;v`)PLkmwL!mz@P`ckr5Ngnf4HkBjQ+9wlKF5kcR(G$-<lXK~%+_o*A
zadaBrPn^$vBt0$H2%#U74(=sN`>R*2><Cikt1GSD0v8h1UD5ikq?JL|@pfN{Y*v+e
z|0_zICuT!?ubh*1P#C=6?5AMM8xE9Aq0BUh8AYYQu?&_x2q<T;^^<~*<skI>VDf|j
zPmWJH5qjtYiN^Divvqab*cR@EFuG|WB1}7yKsG#WW2NDVQdd3Qb2N#?wJ1I*`@F6q
zlTer96C(r9Tn?0*tbRLzZ*(Ag=6v8_=Y%OEk$dcq3KoeMRz$lKoQ}!T?JVw-igFU-
z4ySDIIDE#d%>@z->8DnHfIL$7m)9;WVGJ2##Ql44AUO~0`0WgK3OHq^v{{<$!HsaX
zrevpfWjT^%!Sf?TLsuoV^yJh@SpvV2S{kS`_1cB`mMqeG=tI(a1YGIvSRMKdAIsI{
zkc+sWzMk<{)y_%pZsc&^dHWpn#ypuWdYC7#40khN@7Oo1n};IMRzpgMe%3IT@BWih
zHnnB}x<@0%POz`5zkhC&cm=EJhRE6p%W8%2s++%oS+Pxp@D9--kIFjBCJP_vdU|_%
zE1mr9QdSWEdHV$AjaI^Y_ltU5?hjJ44uE|ADluVad3$UrzB_GY&xalWxv5_NW$b2g
zEE!;q0kfZT6R1k{&RZt%P>-j*=<_dt-8phYL*)lZ!f0oskE`I-H&8`*=ySd!cI<CL
zM(mVFo4xb9y=N-D^qm{21}QE#?x`5woO#z=AA8cjrN=~}pcma}F`BEU;iTSe9iY0)
zCNYW+3SnbJjlMKbZ#}r}lG&M%<V~6uwWR913CP%8xZ;}h=X<Mh{aiWSyI;B#I~|zL
z)^G_@JADGvJ0Yi+fV?!E{uBcbARROX1!Y}_7l^M#v}$5>MrPKM3gd5Gn~U`9S9$Fb
zl1nk42RK(cwlUkobUY!UK6azhc_(_PS)mwEDBX#ua;b5FJ4;kLd!BDj{%SeZBm%u2
z>#X5kBu4o7>7%HZ2~%bdbfQ3eBrz)e7R5hK{XRn^_}1b|v9(ocOK3k?SI%_)L;M3?
z#W2B-<_La<7{S^aWewdp3{n8-#Br5P1rv+t%aB^F#DM%oa{-k<(6D_%wd@oJ&AnF|
zu}%)s-dO7(Acxw8lWAHKHk`d2^`uu$hlNU6c_f!wc{W5;M`Kj<BU~~I3twe^+IU4#
zf?(+yTL)DJel8g<>Up#PWZ2i}u;<Tcj0u`VLGkuuivT;H`N2I~S|4dJ!WYP^GcV&q
zq(+o3ZTD;68ydsIE1RxfHazI7c#@QC5PiPz;%SpJvCG@SiJAAlxKL<N&9P=*a7mG7
z)#XJgcktjNxx40^f*E;yUV?5?uBW@&p1O`4POlGd;r1X6JC9Zl;SAW=O5@ENv3mt3
zv{thCVGLbrwhw*wJhRP+Y^^Ge%Ax`1$LMGExvWV@%fL@PzAkpqQx`ZJ8`cgvN21_k
z>L#aE$&TG!iA=6wJ}h22SZ~~CGDB&6bwVZKXl57RBY7pK$08~7hhwXM3Y5uN>Tw3C
zVCVd4XY;?MSP|)84h>66(Gtyvc_J8VK6o<=xnJKb^Ur1c4TC~6t@*Z&k3Hu*CaX*p
zB)X&ouJ+z}VSh~L+$Lirt)8JiI<(c4mCK|-g3CjPx0`fk{acbZ0v<;y8LEAfcp#!P
z(ouLV#S@fwd4P_?LpNak#hAMinec!RS^2{Th*nN_i?N*vvG#s^i3?b)8DD!osNJe6
zG;PqAbU^z(bN{y2yqcl(_f7q?IX=(2njA=fuwl4wc6VC<Q4Cw<PSFhU?2^?;W3np`
zR-myiY!#nwDEzAYXq-ZQ(!EEoKT3YhJEtaP#^iz7M$vqCt4JX@JgRYGYHL}#qs}lM
zFR)WKbD|cp-VE`EJiq@3nN9d%D3|wfrOB+^i@mim1Cs$qI=*pe@r7!7LwUg$-RdDi
zXmwS|c%&~Wo|D+_XBDo(I7!Ke$!zuj{TE8z2P)JT@W?A~zB98u-_*|sDiD2HRBPVe
zFCOm=*VE;R!GndRXnvYBwV~Pk4NY#)${SUQ2*sN=*J0}tl;c+`){-v_mwHSO^eBlJ
z7)PKKbPSqcZD8q;Q9Ct!U348g(-Kx@SY1^mB9CaLT#1eC(i%a>H>4SRU%lc8o$e@q
zCh@@WxQs$M*pVA=IJj+a)LhA?QGT>&Ki7T8?qNRNYgl&nJ*MeBzGV3EhzNzwki_a6
zURV|%<2oD4f5c;R@a5n+JM&jM%pS_sB;mN2*(?Z;%@UXoFLeI5wRJ;F37+bBz4-~t
z$7yO$1`iiMxFT!8kLgp_Or^Ettc%cd4A{DDkNHZ8VGiyaW8v$1Df)Tyv4cB^b`zQ*
ztbADjA5Ef$<ncx+>;66}Q=9jQt&&kC`^xGt+j6H-ru<;XndKcFow?7ARdXpTSJE~Z
zg;eue2pmDgIY1uq&Kt9*B&rSo)!Vyj^vW#S^z%%tc93ri){R)MkH6jHML2}?{s5(6
zGjozOJ%!JJFHh1TBT8^NHfBqz8LoV9(6wOEl_@(v!lYkXeWR08#N#<~;H2UuBi)K%
z@7d|JEV*2GG0m_OEHdJ*)G_8>J1KbSO~Y`-*HlY6(Z)2+3QGrL&b1e>h$lQPH>H8(
zi!DkUo$|f&FeEoiXKP=&lWuTeig3RDY3zn%w7NmDHfhC!+r};MW~~H)c<)v3vGxq7
z_=F)Ag?N*XxA#onO+ErD`!2Fho1M95l*W-KtK8}*c&cH=u17RO3A*xW)|aYS)UvXZ
z5;Jl?y=wbINL=!hxLd{OFmUDL>GP{AA?@|)jC#Wj*FA}*g2_^}NDXbZh!(o8feC`1
zp}|an4u)5>z)kHO!snjzpR3arHJ`Q?X`G3c`D68)q~~+@E~?ETZuQX{L{_W$K6*z!
zu36UJlLQ;&Txxu|llB<aWbN}%&017nWUo5oA<K=+Dz^oPGjBlmC;n#S2rW4SO&|_d
z!6^l;a?fIfYq4=9wB!Ztly-(#&hY&ar9){WJY&#Zrqa1Nfw^7|IrE)ws;i_gzFuK3
zw9HJJ%syeS_t~RL4`JA`U;GEWqWz5lS)WhVAQpPF*P4&A%c5{n<`1jeXKtB3d#ok3
z(NsiaFz=yT{IB8~i-72_osqca{B8&e2^of-Yko-S+J_6b@89i|Qx`<9Xjq&4!RYpO
zW`(mHOEyk@DEX;avL{)_;iST67f5a7iwU_(iIxIb<MF)0af|7?sv?gQSw;z~6Vk31
zEap(#HKJOahO(p4r^_A+2dek&kD`GKn#&n`>SA#e$R40+zJ+@?WRn|)6|qLtn2q`b
z)V65c!uxz6<t*rrs;69T2e$glMxT4co>V+eMR?tg|KrGW%8qdDT2fHCO$URMYLF8K
zL+g?n=my^KCyjx2l|A*QWN!tnbhj8Dk!K%RP1cykiRoy%B7MTHs-uM~WgOMc&|_rg
z^k@cX=B{!V5<5$q!GcsKj+OT`_f9J{#~XF@2mR-&Cyzkx@~q9xZQwn;G%n8~N_OI!
z>PG3jj(IewQA4f~)sjMJa`cgn=z7&Qp#Pd{=S4(pz=KE)kyq`BSW~^tEjq9tt=o1D
z#}6-|+IH!=QwKsM%>dpgluOtw`?19k(a+d(h)L5p>xaq_g=Ra(J+dF}+N<e`e`)8O
z+lffKo77u<*HN$>e6W*F842{uB^%O8^J)xu?MBv=Qfo0s&#UMZG$s7L;GZj)cK1xr
zT)j=&qj#UT8iOI4sA0QFg;I?8**X+{dpBoNGe9o)K-cA}<;t*IBhb3puvHFI>!=xp
zlembt{Y5ohC^B9+#oov($!C3^Fq>8iCMF6<q!iO);ob(LNJ#0TE=+tNG2q;T=^IlG
z3HaU(u_Q6Y-b#UIdsBtyn+4$4fQ}SAU0Ak<=&^i&%+;=5QZi3KUZ<6ih53<Bsd7<e
zi&otcJ}bhfm+;jt`(G?G%v{|)&dhOI>npx<h*e0uGW%%HBqdjk%G`-Qy9AaW%flms
z<pFIzj_wM5c-|uT!cej>o103gQT;*Pkt0|2m5{k|b9z_*_?F#d8C=?p+nZe9I2rWT
zunxzKY8r&m6U!FSYCtn>zquVPf@*jV*SPt;*7ot(Sev{oZ%_Qzq-BN|NH#+}>HRoF
z<pzj3g62=Qt0Ws(cQTvNGN{mdAL^)b<s}-<q(mlhMlLG{q3`VFbe8-9!4u^lGG2wf
zp859BZd1R}a2>Pv5-$^w8x*{R`@Bz%rN}4bayJ{0pgbDDOVsG3T;j|DAJ8lQF=X&u
z<b0PU9ewGloN*c-T5`|tbvxexUsw{=np~8VMg6jHYm9M?J)9wI$oh5J>@r5+=7*LX
z>#W+OFyf>h^yFEaiy-Hl%O3f`lTa@r#c$nxB6dP`TynP-XF}A!x}+!F4J7H!M=Eg6
zS?lL%J?`@9x|k8P&^+PuZIuFhxER5pw{6nY%0*NcM7mGWV&eORJFHNemb+<LMQ57T
z_bzfBTJeX8*S%73^_YJQWJ8Lc<~F7%2yZ36WGsZ4!k^&K%Y)yt7TfyE_Zb2qaCwzc
zjCRA2ojDoEd1+9J%ip#i##b)yYek)Vu+q4beU{C!P$pXJp~{|_dB>CYS0b<LmH+m&
z-y7&d#ksJcParhNvbgZto8IQ5bhUgN8=I9t2Nkqyp^czfCcBd1<=PV0gge#Ra@^@p
z`ODltKD8Sv%voqOwlgn9BEK}>goWS({mM(n7gg#v5n&cJ3Bon5kveG)9N9!ZevKc~
zIwQJjCam`vm;}=gJA^qxq+Ics$Mfx)W3q4`T89hvo?-FCr@mBBy+i7xjP73bXC^DJ
zrVl%~nJ<U;X60PU<Qj8gG|!rx{a*mfpZ|M<xqg6Nzy}Y*J#br9R+lKb^e}hScTpf(
zDe@Y+#j!jQZe<u#=~1Ff5QIzYJ6l@^NWOgAQhiZ0-g5fgHPd4;Grz&?JHV2|NNG~M
ze#ecvcFQFJ-@<hK!`l1zjJhJGV?h=#-ekXVer~TkFI=lq5|B^7NzKVCnI3$Eg3Yw}
z1qGE`3S57bWFL__^fI&pP*3(cfj$b5!5b9WAs&*qOA^%`%Uv=*fzL^kWChsV=KGZe
z{lARnnAFFQdSp}y_@;c)J?a*=miTG_yWNOtJs_sTvhJMhJ$YD+pYCHgPnzM+5Ue}Z
zWwBC5H&jmaCcOyQDFgPKSYgg)=X_dsN2c@6SUjnZR;jd2339{AjVLWinAJt6huy=}
zWcXg)2@`2N-~Xq<+bjLUw21>74As8_#g)iK3uRenrwzB8@!SZ;qMv5^7)=$Yr7L9p
z-YRB!tmCy#$<Z~V3XALp53nh5TYnp0J!r7m3H3_>zsl++2!&oRhMOI6bWiv2E|stv
zz<>Y5Gh_IaDVFDsj<l%Mp74=xPdv0m3<PuqE0R92R+L-_`y93e&7LSK<e*uT<-?U)
zye+bY2FXej3CWq;Q(sQbHez5<1=h}&ufASk7VH?fzpfk?n3I8aJ435!hq57+$-(wU
zKwsDMAE5Vjhw?UImYLMbtrbK~sV4WZ&U$s`*fhd^;LX?4h+?4sKSRE2cTb$>kW1-p
zsv33xwI{8Mob=lU|Cx_P9KKbDRnfFNz{ia}OI!)84ZI~L*wJQVZ&H(HgK6CNGTFms
zMfKk!Ux%xij^C1f!QI^66>?>)kfuCiVBUc{qm;#$;cI@ThrIf*eQdu_<ab$to3?Or
zvTlAH(*vE@sdd+<tr{jJz9`5FgPE8w5dW=UYX1l50o5tBXj<P@7N3ne)hd>z;@Q&Q
ztF2Ug>6?o^D_M~gxzTh>TkKhdRtr*;?7>l)n;_Gy&x0o4Wa-ExhSPZ{)_}lkwcS8s
zW%ODqN!wyP9L|^G@H_kUfzeG1Rb%@Sx6FZZ(&3%it`K=k-@%y7^(eX7s19D^Hl-1`
zZW`g*Xj^xfhmiO!@!M@5-&_*dxY;<3Zq$Qh;TJZ|Y^s8}aa#%gG>H~?{leGL^a@&m
z{%4Es0Kfe2lLi6jpI;uf9XxySUgg;ntjVi{XW?eoefj!8A-BiC7bKEqf(3Q(bloyd
zm<L(@%u+pOVj-bQ{kpxFU;66VGX%Gx@@e(9tI?koHh<G|m@+!EAb^fT%QrKVp!_hf
z)QBun15Sbpe{~If9wO4nR9ydo#r2$*lYFg?h}K1df><bb(jP9ZHYttsz?=R=JvLYt
zD7t4Y7^RzgWWs|dZlaEC*<C+6R8abR#Syk%lg(xJ3kL$eNN=iqH!qJr;ac#+uE2q~
z*#kjoRT?U9V!7s4W0{KQjrW%2;Ufb2HA-G5ISpb9fCWeZq)VGT$m`LBzxB&zb@1RB
z!qk$;iYGo_l>_kiBL!c6`7$r+&ZjN%QBLfm?-Q;Dzv{<6#x3^^SIp9hRpohqTO#0<
z{jH6)U}>^pv=wjo_B6BRh`X~uRfBZrBWIZN%}>csCR@JVji2kv8um%7<9*u48oRhM
z51lZ)J7PYo8mYX_7O#imwV+9NU<6Dnu?1c)dVJZxemVvJa{F9Hz-iN$F<1VmsaB!v
zM<vL9r;E=5LIlWLc=d<H-IR3;9A17+c`|NKso(0_tV)HLzrFL*$tqopp^SXm>COph
zae2S>VAN@#2ZBDa2@ARJ+9ti~N&{-G7-|2S4ebdN^C-Fr>I~3h`fAOQ626~MP~OQ_
zkv#1t5D#$S-ou>9>bW=PnB>>7dF4{&+#eefXUv+>?WnWzxqm{ql&_MlVe8CR-No8R
zu`JZ=vBoiPT?PHf8!glMGM;A@`2DW7G|D8X*E=eE@7mxu0F@|zg1y0Jmm<UdyWU5<
zx@8f<eN(e;qzc;)t_eEKeVk*B$%dqgH7UM{*t_^*?j6|1S?mZE5*2a@w)%!{Rr?uf
z<VabKFNs<UnIsMJ+0roId4>stJVl<z4nK+5uE4nzOl7M1`lQ%isO5540xg4nz^k?o
zk(svH0lq{T!9vck8TdBE)kiVJc_Ve;z7i6a9GmHu*-6=>wx=Srd)Lx4=~f46jS#y=
zcEi9!E|gC|I-Q%g4UVJkMr^gf4p^*52z2&SL>0`@K%~Ld9+64IlMgvXo5LK|?r5p<
zusq+odC%))fwobr#c{!xWeX!hw|A5`ebZx<vxLWa$<1YzZV#4)grR4@y6zY5ChG>B
z^Q)QecJJ~c*z3eb9}QpIyDgRji=_HA&V46xY;RH1W1pb663-%>?++K$j+G=#ObC|x
z6gCJuPtWN#+vuPItGE;9PIbnJG!*p-{|%C(g!}l#*tKxi1_=89w4%$6?>KUA@WwXZ
z$FXk_g=t;~9pDR4c2}%o&AYm$5u~l><20&4VY`=o%jle<&EzD1#=FJI-Da25W70P3
zp&{aK6Y0;CEqo1m3|waf4TZalm#yXjv64#NCrXjSBCnr}iRo0a%KD1Axu0Z>lVaL=
z)@HxHIu6051KowTJ0a#VB}cP1QC=QDKybS8&f$4!QdVMgW?E_<t)I?IZW`pIohEjh
zsiZ9LFT}3fhvN0&5>kQ!Z*+?zr>0&W_nm(6ExkQm=<whQEybLw8iv&H4=P^@9C+DS
zTUKb}dZTl!n$?ed38N*yCQE|75m!tSSqG`~f6HTngg*R^rfL2y+ZndenBGXvh0-8%
zG_+daTFP6EO4l^Z&XVUj-ZL@A?yH6MFMm@!(kW7&_XFhk^32o&H-Y}h6!F*on6Vu%
zyZ)UZlN?j69NXs%o8NOY>I>?JQX=d>3(%|$#*n&XkBG*|2LD+ow<anNDKy-y?E?4h
zdu3Y3;d}hxUaL*OC*(`k_wMGx!PYF8DU?g}*7Tj(89PiBvA=k-o%}(-`qlRJir>vr
z3>&|mD~v7i$rlr>$)3B;saR)b*AJ2O=<MD?MfK@+C)MstFc!;Dt3AOC^PTmRv~qoF
zccjlChf6vX{zX~Zr?K|K1_7d|!f!S)bLHdrTgsKdMk<xk31JW6qpM}L2IkbfT?t%o
zXR~6+%I)u4mpftW6F!aa@qEAM$$eGvv9j$<{6W@y)h0kqy;e;|M;lvD(xUE;)P{e~
z*oFxT^DTo#KO}5^iq_f`jJuZHbxVC47QB2Pk+|y+f6iPjbP3hkgIYze+xf-DJk;F<
zCL;qCwt*W01Mc-jdL-(*2ytK6(^e>ofn7KpVZdz`Vj<teNx26nFpOjyb8W!*=mwyk
z3&}FO^B>Ny`BO<%GND<)m%M(u?~#2K6yZhF=kCN7lwgIfizYc3^&*m0r0ZtY6|b^B
z|0dso-s3^)0tJ{opb<^Cav8L@IV;&Fd<F(Uq1L0fv$cpFSjogztFD;BaT-b|?eiIL
z4(EsbIsTbih)1WGK);i8uM3z7*D#7^gSaF+63?`wMQL6u?e+D6b}xN~Jm7vwagpgh
z!~BvWe1ATwK0X}XsvZ=7wj#4;(&O;*4-o3AbOu4ux*T^XxiP0bfIWGbpk&DYl@{6l
zAg)cy`R$_j;&H>1G{s@fCKm}NJIe(&X<a$cL5@v8k3tH@!@!dQ?XnZ6_%sT~g<PCm
zQFE^?tF#-uZLKqFE&ZX2<wiy}!4>(Yf;7r}>@lBOOL8pM*;n4Mbvi`sQW4DOZ=<L~
zsXt8JI=lmWRR?!nukHItH$2u~;fJ+W5A6KpeZ`zGdGGeAu&@uWNuFWWB#-aayzOe0
zdTW3fiGV3R(<IXtC?Q55*&U0fipda^=SA0h8_4ohsGR5~7yq8$!78`(^pGE>54!27
zcmMu+6G~+bp!)xaWu_m4oY)@ZTHdf-gn=8rNiQ+DKPSW$ch7kou8%kGVeZdsU~Stk
za{O-pN}cNfGELJbiy&g5oEky)9^O?`R$q%VAEzEK@n-9Y6T^BlZxAFv;Wp*Zv`o@=
z^e&}0uN_!_+hX6u4UEw<HcuBFoN@JD0&`j{P>QQ7T(Cl6?J3hVTq_mEq0cIwArZ-&
z!O9t#v`1LJ&6HXnswSKOVfU!`0dme_KS{IV7h3K*xDz^F;q&ze$nV{<N<{I2`laf(
zZwu5eUw&h%e(78g{On#f!<D{{8XyJrWv>G}6GMl3({KlcN1B!(s@6NeN|CdjjjNTZ
z&v0LAOWs$9PK^2GTr<E-2<dWaC*Qu}%JA`e^QKB$`!GINYnTpn&q3QJ*ci=9{{K&V
z*B%Y^yY6+Bgj{kBMRJQtxfG&8nlz+D3?@>O2BU<*FhyuwL+&xTCHGurWH1wElv_lu
z<2o+IxD8=2rt|&nz5m(!>~;3pYn`>uS?m1%n7`h4&Ajt`-}m`G@8|hEpC`uf5bL5V
z(YtEPdve579;()(NpiYx)1LLA?(`SlU%ev9wbg6=tGu<gzqz28czuAW!>yjq9Yz`2
zvWMidntyXqOPK>HeLh67)MtB?2Si2#*#tBx#FvMuIDUwH7!5CAxYCbwPmsEc-Pn2M
ztMK5tR_4;F#e1*5i0n)J${LhHaqTzvV*I)_zl}*GgI0&7fNlW*qu+ikoy^|z3}E2>
zum-|;ES@aF%5lUq1+Z&S1cE!=yeF-(oA-{Sb<DX5zjt@dI#thVhM$Pqx93O)N^u=i
zyPQlhW{QwbgsnG^{&4t+FNbJSRrt&Sa*!w(JtHt(cYW7^8XK;t0=4<qR<sqw)P`7c
z24@=CGO!tbbLVAM%`Y6!DKFNoJAIlkflCuDi)`393;$9fOsGA<h_^kA$LK11$@|!|
zR@x<*qDjzmJF#d3p0Xxr`N3IByP*c*l29V0%;P6hgJ5;X2N`W<eE9g@QZB7o&D%Qd
zX@L$LU*zh&+L<3f!jO+++G!^+pS2S=5$$Ul@kg&znOKUjT~lFMaGf+iR%8o(KZ?*+
z0e`LEkZNa>*jS(^$fIT}CcB^QHBQFD_^Ud3P!S3}Re^?%5lxo8Rm#Ycs_pX%A5Pz~
zAkfcD-*&i%(croBOoes_m#>t~+0a?TlLb~e@Tc*A^ki2Ugi_rx@OslG?XCSe6Btob
z_@tV)!!<-S)ViC~j|kk!I`^^GRxKw<sPMkA3(v8^4|~Xpx{iD-Zn2OxpoU_ut)wNu
zE0@NQW9AWdl`Y~`Q#=%L$IHcqj0djkw2s54Th<~}+<T^19(Atn`11L3fUXPeo6fL-
zpiu~@De}Tbi1!ZE^jl9;BZDf#6KnNs(0H+L;piFr5a~v&xaRX#1HYF}7k`DZ#0pW!
z8$BQ(R}&z6XxKCxiVE8Cc&d~iMjzVm6L!@pbUcRVb({79>qvoErDv-BKl6!|oN55A
z6&Ur!^o>oCCR^NtcBf;WkwjUKyeE$XKy(F}{>bgq*`0_Pp5I)(+?&3eQb9~wt^Oo8
zft5VI#JYpyN8g%&;eq4&o6E!1*A6+vF4JjQ&3?Es|79e4u=&d4?>1nQ^E!H~-tJ@@
z@|2a!cu&mFl`<jp@qVBZP7V^#z){05j}|=6Y08>|1m>pyTs#(<`!E-4%oKlTq&e{5
zjJoVTsogahM(dV%O9VWJ6<XCH*WW~R#3GL-T}Hfz_mtYrv@SeYJ@Vv)v&)sJrB4m!
zUQ16ceS;7T@(-Y*QqO0$;<&2%Z2~S7;3DRhePmuRTlf3UcF1R?tx{#w!<L$LD_mRL
zi)JNl2Gy#5bA7Pa>d1;20>8V59n9Fc<{R?-gLBoI#fwaJrc7l1Fv36=U&G8yC@L+_
zo%0zw9BxthnD?U7HYn32UNHvj8}?D%?(*%feU=Q(n|AIy80*b7*)<k)rHi=$8t0?<
zC&iV_Wy(X0D3+{zj!=gAqO@U4*y5%97Ux$+f{9TY#qKgX`r%4X%;zIXDy(~x=M9t{
zSe<~Gx~gpIzb#Znre8od^*p-x^5n-v7t@&aF^{0GOM7)jj`kN?ni|&vZbFUQXDNsn
zq(7EPUKS9n+{w@a@t3_QW$R4}!b=}66z)`PTXy%&g7)WmR*SjCor-&j>$rID1o7=e
zyg5EPm(#*+wmh1eCo3uuofJK>Am6F$x!+>G{!*EKI}d8r*Nw&&{|e0u6fikgEt<B7
z8^Nq%P!59tm*NMeeRF7R{d^Z-MXVPprhBY$e`BCKe$#QIflzJNm65lad2UTYeBouw
zmz|clUB_MQm0?g8a$JCA1EBBuK;{8p2=sYNL)%x-^|(jGcnnA?#>X$+_wgI}o^z$d
zi1dKW&b*S99`-9kKf+cm8#rOw=+$>YErx@m(mhpV)-`r#`Oi_AZuHpvJd)G)U2?Vk
zbNc7>+6Tkz6kyS3{@vAOKP;L-507|^byST_1w%Tt)XCY^Q+_`vxG-WTwhjysJVa1X
ziuI@@Q#+QgHx2UMHU9=avv0EP%;yuTnrg3>FV8hz3eG?NoO+KT%Z<J*Q;sZU1$-_C
z0*%Ql;l_R5d{7fJa`k87tIV%)O{aJd&GkS+!mb+}g72>Ph={T3SEaF+ILIHR6FWi5
z99JOWGvDBZm>3g$XlIRANXjTc35|VPXJ~x@nhCyi*!84c8r^K@I_+7P^<!pIj|!{g
z&nq24J)@zD%sHLC-n`O~s(KsafLd@~Fx72T^<&?hhL^`r?6Jrxnw+Yr=k9*8<!!}w
zS|A2Us%7TPVEVIRuWi=6evF0VD$qfDwuMcKW<`s*Bgqr0+I9lr{gCo2x=z4F`cviU
z=qJ}}vG=|oO(5pg?j#2FE^jnBiUANn4su;5c+KJfz)dnV@TJLB$1X_ek0rz7GE()b
z11Dkv*CV9dp<CVh<Rsj(sl+v*n=uSFjZM?{y0McPH1dvB(Z7v>udmnSAoMs?oc#Rm
zFMg<11LWCxSU@xheudWg3f0qOpjgA)_Zh(1`S^mQg&56A#J1VwOu-aWA<RQ+fR`>~
zx6vbatgXu(AMD2I>#?;kCrb#ppxL3Avk2pufXq8@8hn<Utsd^5^8<YbJ+H=y<yR-Q
zQBME7u6Y4=@uTo;^E`MT+cM{j^TTyrO^pwVp7NfbzBrlE1?JoG*w>JUPnNweCZv^y
zyn9F=nUB96W2RbAdM`Dx#r8<137@d~!GgUR!vQ8ocukOuvQ2V(Dcprz(!-{cCR>62
z(+3GGLv{R9HMWAgU%{r*H<n?e=R@uHabeB39c^8l?c*J-EuZ6UW=4LD@dq7ZRa3Qf
zN5Wrf;k()5Xs-f<ypg*go&1Jc1orYUp+TzW7h6Q$T&j7^Jhq>doL#t-YF7B0OA`>m
zP+MHt99u@L>4T`S3Q{V)lZza_##^s<>ti@ohon*wc+b>~rjSPWn_7Y6eJS5we0+ND
z>4|5bvi1o6#8mR4kWCz!5b$@jR@NL5hBbXj+ZJf_oZUhIjnBbuU24T=iUdZk4a9?U
zJqJ!yJPR3BDj-Dr2Z>fsk&6;fPSFC9EFiyvIlJA7TZ<ihfanZ26hN!V)MU@ap)4$P
zpAtyQ%iE9q&9e_S@9nnbO$kp61I0x}iGYtajKtyIaXxFzl6Dc6DF!F2gOxOq!mbDU
z(#|v-<a6|Xedcn-f=P%+*RzSR^Vgk8ud)jWw7?}yMh|=N+7t6L7Vb7{tNO5i>uI^5
zkn4~S-DpKRL>H_N%AT{!CD@2=IPxq<6R`W<^u=QrrdOKUP82KLejs$s>UpSDoAA;v
zOJASu-&`3>jG%WugFe;c@vvaiRJ0)rMu%|IPgYMFim*yQQ*<-)^Xyz<@Sdk$`7~&*
z(C}4>Bl)Rcm#U72d3`x?_`vOUrfGwrAPUzMB)1&eA5C2f#`k{&YAq}vhzxU8zu_tZ
z(_wQiYyZrhw#I3(hi&fr>w_ae)yeF0YxP^RELG5_6+OdHN1f%*6af)LkT^3ueEt{g
zu;b9Ey(_UgTpgq?zfy9SppwjkIlP&~_2gSTuNSd^<vU?PS=vLuBPPf)EOTi&pa%iD
zmmYq5Ik2~ba&m44JfxkdbuMu|^}t@aDgML-k(&#%$RylUc=w1VTi9&bjrj$*?34Gh
zBJG&!HB9a%jx;Nz3}KV$OT$ABiXrs8v2qbZvS-^}Ob$Sg*T5glJ?d{DKle2}=&c%1
zR3;3j3%gIyKzxlob_a4-)ZjHJ2U>s!Qhh>R|4>RKM)Vs)^Oi+2a;~MUg<NN0)x*JC
zteP#F-An)P2PDGu&S&K1*O*3E;65}~iD?9nf2orPQaS0I&dLSYtSn<cd#7me@rJ>M
zzJISsEpS0ifEB5kTjGhBG|U^DkvUW>mJD~dxcvvd3F&6TWD)^__azYn`5qv9W#4W+
zqLv$AOKBy{Ty-{fXLf8nsRe6}y~I?0t){DDq-RND%*3+LDv*?IidF#l7mUyvzmw>W
zrn%~~`YwDk6(SPB8L2m4%+xnVxQ}5NZ9Q$y$G2;r{Gx4bqC4N{yKhVUB&jeHIkiHl
zD>l8|z23@oM&mRG=QT<3c<FVo_U*ePM-hqf=aW5$MO=RL>VuZ)tBnSynFi%fg+u9v
zhabHfvmXTC^)(T(7PKq2B)0C(Sk6svjjO6~|9aH-=`FzPw{Tv96{-<{B;3u@*%3-y
z|KzhL_3zU}rU2`3N(xAcIY}+;s3t{l1z^4~0Z?>6cl9^dm7#u9M=z$j?F9aNX8-Nv
zOGFso9Lc(fFUR3RI^aAOQZB!(3gwr2I5l}>RW>Z>1S+O~0!dY3OBA-y5j&?{K_`py
zP@_u}|C`hk+v-ESHCX|Xt}k|tUhV%XD4^EmG7-97SF%Z)inAvwj?e&9)R1AeLp4D=
z<*{!tjmZKA@{e?7v6{58Lr{Y_iMr){>{8v$Eqe>E>YJH*+mUxRj^K~8t6UE$@V=@I
z^U^}fxcZypCkG8=&{z7=mnQhMn;QF<fP5H0L@aTCm~ZzArtf>x?6}5GhCrH8MW~JS
zF=b2Zjn@GjNynwO&i-=MQAinx3b1S*+nj}u>jhwB$56ycOY(QnOchiLib#r`-$I|a
zj=O7Rf*jSA*Sq=^Y{FIe48QU@nsgT7UxZ><FB9rTi^dx!qqX}OH*pRhdy3Vnn!k<5
zrGm0ztO?bK*i)(54*Bhhp8WQZELI4`;9M74?7mLq04u<X6}B$g?<mOBTKkcjAFVQD
zO5pUwi=U{mIDZs-%V+e6i@;%(w{OxAfrcyhWxFxn+quIfM~aA*7~iwoJ7aBM$L@WD
ziRVWNjdlY4fMe@?OHlz6)0x%uj5C4$5Py23Exd}P;kv$|8$oO}^jZ(rxpt`JVA!bX
zS+pH1q3EIRM)=!wO5Er1)H=m|%?YI_GPRl{l(zn>F4)uy>2iHzw}!_U{HX=$YIFq+
zo@#5B1JEGX=6}&T^k@$zU9W-O;hxq!T_l?)8+^&GlcFnP>nw6PZT|b~3+KxAleyRV
z`pOGkvjc7mMz-DZQ$7^ea`XyF>y7To@9Gj4_u8be$aH5^v+T>Qm<*;|zXkBZIQ_v4
zTD#-FYSz8DfGHpMsdg|QSggcO-&5(3pvk|7rFYB>yDMJ-w0<_JFkmrk@q#kfBkI4o
zro9*~_^Wq&)!Zc14{ZpxsOUnkc0CI<;(o1ce6P;Q_PLvOmk&X$ziF$CG6M9JQ^yA6
zAsppCM6vcXCw$gwCTzq*6e?*Y^{o#3qJ~w1o29~UOp9xrceXI2s~J(q0{5AOxNKAc
zrJ4|r@@gVdr$?sQL!l#S!>~^N9B<@|s79RoUrRG;j`-htTv@$VmsFpay|iPyGITJA
z91_8vBFn1o_?n={whdV!P5ooi`uFous-GqEqv~V@!ZmAO+A%0Py1m6VATR%l6pxef
z5Xy@vZRrm-NGYG#Q_LUK)MC=9PeUNAx*C3fjfCe756qGD-aJx$wDsC&MUEA84pKH~
zr?@lx)%tQwyr}w>2g<FU;j0)NXl?3xnRqL#4cQC?bo4;q$yjWr49izKcD=XvQw_<-
z%#6>Vna)|6dOys$jVJ1tiaeoOh3<t*6sFlUgSZ(!x(Xvn6_wQDIZ&=0C{Cqioe~3<
z8MRW8lsMv)@T65F_~W?~dz#n*l6wtKEoaQ+)1#piU(1n5e<Ct!Ty)YR#4o&~sz&GF
z^309*t)*{q#|gTxF0Dx6UcQ<w7%vHEmaGmd`Gt_hh8j+s$4980wdkXBs5O>8)4BJD
zkM6u7trR;+J>IySzoci#7&Tj4`Gv4-v32mlynw=z*=G%sm`l_pKDPGvsuDe0cRL3o
zs~A`JQ@;ZFx2@8QUo<N`NqW`VT6jM;i3{CW{1YRrc_E;2EBWP(N5lsS`8<+y5MPt}
z?kN*v{-tRLBdLmghA3t3y?b6=1h7sJ<u}(}R3yE>u|BA2@=7W|(H0;jmR-zL>DF>k
z8;`d<iA#U1_TB3uc+QG%47ArP740>NKuV!6bl35WgRr773GXst2Etvdx39#xJnywi
zwT@?pb_SwCX-J7(oMKz~R5$pv>PJ`+`vL>jJyahyU%Z7$T;Hx+yFi3RkMhM%gpw)8
z3n4n8>iStndPz91Q(wEUe~{z#j_RmuZP0cbmI~g_2OcL!Wr{Hu@dmN)ItuL`EjO&^
z<TKI~785*|7Z2XQx;TN&y+~2Mzjyb=6F_0_s5Y5t5sEfs?wR1=^Jitmdy#dQvQTAB
zOj}1t;ee_1*UE><DiU-8e9-5eP1Iic!bPV(Qo?dsBT&SPX>1@canWeVJH{3Xmv50b
z+x0U|b(YX&q8|r7OKkD-TsnT*co%;mm%OW@mStsCCU8mFy5_T5g;Q00NGrrlmk=Eg
zwSe!xo=q}h;dUx@um&m&PgE0C(t9*=62w27=Ay0J&a@k9dXnaBzsp0aH39YU72U+M
z*jLH=lo`I`!A8S|%&BE+e}oo8Ae;z`ahqRrc))-X<Cn~P9(2onypPyCan?lGq3+TB
z)m_o|?@E}Cq-?hstOIoKO|6YH(^4n5PjaTl*cxlSk9C)ONJ@P0i`{jW#Se8qhc8W-
zrFRIjQx`lADW6dmYVK+;5&3eBcs1Nt{t0uHhU%L|i=B?84wbVE^xcvSms5bEh4vo4
z!vow8Bx9B|8`hVGcgi}3593DfdX%bigt?a&VkuHfT!`pm))}PCY!{G3DxgEZ$U7Kr
zOYU}9yOCi2YP0R&&RdU7i?M^$#^}7Fp_VP}3B-D}&L}RiucPcFXVLuvOtq)Ok2crI
z`!h&}rIfU;<3F19o*|>&2Mx+Eq(d{V?Ub~kIoA;)056qGMQ+z&@s=N5jL9(`RJs-O
z7z-Kj9wBL6o|8{05ChkA<I_u=4=jX5qwF*=kCD6%yFh8iA*fvhq{0DNdsqLGq?&g_
zbGq)+B!K)usichzwN_9y-OiN8<c8fma4aO4&lTyz!6~$+!J?~rAP0OP&_Z%&8kl2J
zsFeNI3Ftb;ZAg#Z<UzqXjcNd%lYHdcPYh!bGh#8*`xTYLx$GzbKsb?@Sp|UCF1xjj
z+qkuf=W?_B_ms8&m4@rCCiJ!6Tyby;Ze0NPo69(Dv{)7kTRFu_;4}cy&d&ts_kZ;j
z{};x0wX^<Cbs(n^v%*&c*e`m6lh|g$b%8vE>IS;xO8WsJnmZz=6m#pF9cOE3Mfx(s
zye9a(oPe90USg~0-m6i%&Gj0Vn0t>5f1RDfZOrum9FrT|X7&qZ&*#;Gj?EcpfIay0
z1K&YiZyL9gu14^{6D6*U&?4;(z|N<|xSp69mg8jsr7erUDQ$6@zheMREW^KQV&#-H
z<p2y<;Iyad+Nq}f!{D}*<sY(+R)3RqeD*&zrIl$D-cNOut?y31@m;g5>SS+ycPXU4
z{nP=KpOaa8`46aEG1|?;ZTJtrA3N`E3aDV-`l|{iH1?hj01`ER(dR2dNOpWB7o*B3
zx)qoB`pQc1A5)t6fmq{+8OjMOc)db4=``5ql!f>L<6r#6l$fN58S;N+vA6vHuV(O|
zfi3V)t4M2Hme~t|a4+8y*w{|I(J9aN4=Nr?EeRD0hTgP#w*nPPk*;vVz<domn#)_n
z=Il(%Z>|dufUVTY(e$gpblQ@afCC-P_4NCnV}IJ^@At%m|MU`whFU=2*XznNrdCR`
zP|}GN6Smjlc{<Nsn2TN42JSdU1_gf;VCH&SWuvtd{r)d9iI~5rB%1y?g+=|-E`M%@
zKM%>Do`IwBzvm7W{{hLY$(cpu3`E{fgW^WPsu1mw#cn-4;i_F5DZh3{aQz<q5Aln)
ASO5S3

literal 0
HcmV?d00001

diff --git a/_images/bonsai-predictposes.jpg b/_images/bonsai-predictposes.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..2e4f04a22020342a8f45bde98ab45eb6de607a1f
GIT binary patch
literal 79334
zcmeFZ2UJsQ*De|aMSAZdDov$HS1D1ECL0lyUZT>85J5_)Q52*jARr*1bfiRjCqk$q
zz1NV?q$iXRAmwJCbMF6td!Kv0^WSsE9rup=e~YY)u?8#ensdJMo#mPHU6d)xB7psY
zj=m0nhK2?JrTzgZvj8ms?U^%wey9f>^>g<8*|T(XXBp|~&z)yxWM*b!WMX1r<z#1J
z<zQuEV&`S&;Ns@t;bCUu<LBk(=j7(${&Nr-TIz4;&N7@m%fQXT#KQf5`=I;)aGXCA
zL5HEGxdJ%DK|{+yL+JpB001;+snq^S@P9jK&QNJQM}MAyk%@Xk4Ljfr4K3{%I@&*}
zrrsS){Xc+?<1FW8`Mc-1jGxh8@#I!`lkoMt=)H<I9+P3LnBsG<PzFX`K7Ii~aS6$*
z*RCrmtEk?*bzAGcwvMiz{)5L)OwG(KEUoMv9G#q9T-{!I`}q1n{R7^<3k!cA5gC=3
z^f5Ul^;24UPHtX)L19txx5}#On%cVhhQ@YeM`u@ePw&r>(XsJ~$*JiXG-hdeWp(ZM
z`UY-q|KJdRL^wYAlP?+o?SEzK-#GglUmR4v&d|}(($W9P7tI+z>PgE%clNUUInKMr
z^v^uGt|+`Y&wVf9YegG_sG<p$=egG~Bd?eeS{(N$YyX9_{}^MT|4W?x8)N^@*9?G}
zmWDccv>X5+;FK04*<r-~+fIH=`h^eMl?gdHhmb$qW4iM8o@C`s-l71QAf4F<JQP5$
z3xW`pi<=OB0OQ+?%(I=D+Mw^<D{QF&e@Z>_T=NEtIo9`}yBO%R-gmYN5aK%n#><Pp
zL6gjf9;Vv+2RjT;yL!!hSJ`h40Y5l00pwkN-UJ<7SW0R+VVwaDbT?4|?Xx&%(plJ^
zl^pp>Ji$6dngaL&CmU!}03GH@1Q^CIocPc~JDHe3%3`Zgoh_57nx`k?jc#%MiYBK^
zm?sP>N0tyBMm*GoFu_K5G7HDNTO@)Xo>vwUigwI2QE^P5iwd`h2B>yD*wggTme?l2
zY(MEPxXL0XH{xGW0AY2Q!4pObV7VGZ)E2?k5-ll!yC7mpGzAa=&sIXkr8{Ovs%z&H
zLZ3M)9dD=Haxe7|i@Ta%6K;FN`K|JP)ACD!pUiSWT()KnI5lL5@C2|5vb!9$a!+Yc
zIB%rw#RW9;{uO>S%j_@5{>Nq&7k*A^IFor%(ZJtQ6hPuI495-oM-eoCZIyt!`n=7{
zG!QJrzaG`>TsBw>{?fhrrpm+j7zRu`y1_7St09{z)w2aeyA!I&p@?G{J_>+Cmneha
zg$<cg0OC0m07fc4=d~K!Uj=IXw~1Z){e>0zPl%~Mxk~%}pUbFRGYasZjF5RU8==Vz
zXM{)K^HDR+@sfR{8HxGX{i8dBcPB?$R3?7ew2LY9<fZW2*oW#C)<7q>C;$pTe)rUm
z0+>nOZiEp1;irQXz*j+lGv0~<81#qXP}tm48Q5Agg0M?4zl`JRPJbU7WQiqDyKgTK
z9g?gJublYv{Q96bG1BMp^X?B8F*Vjj{{+t{Tp!y^Qb%KUC|MMQEL>+Ou|ioFzT?RG
zj0^<(L-8t{4^_`}u7>M~Gsec$cd2p&&ISY>j7~mLvxcpM;rZh&MtO+JNpfHx1rSG_
zX_09#1)#A)0d&LK>vy?`URdYQ<~vFnV6zmKd&-gJ%WsVuvL-t8JOvu-<5Z%?<WE~)
z73f}RfJcF<HPg_r?N~|V2+ReVtvjxiA#-nG>0u+-+M%hbDkl8%_0Cm(4`j}9niyC3
z`9+SSyqr1@R4l?+#>kA^2Rh!BBdd|QVUBR(Q7Q!>OaU-rccVqB4-1k-!b_W3UXk+s
zOQFfl>L(T0o!oC0v(+zq#X2<y*dP9Ur2Fn8AZSE?Pn;=tF?$BA1qlnGn{2j!IWcn-
zp=41L7%jv0pt*%z=ZSxk(#yOuX2$M|4tZC9TSxI$c>PBFqyRq5c>2O7%0gUor-1(h
zX*4iVC6q-s`T`D-Ro6Z<V>v%%X>hw!^lReL&r^vm)5{}9x7GF19zN8YnOgT2;dQl>
zXq<4bo`u&K<St}t9&t35`khhw@Z~w+g>-RZ-fMk<o=a*PuyWLD>)~9+?`+uS0qo^z
z38?beL2tLA_QTYH<ad10#zOJYB?_QC`&7#GffnP{oY()W=gsVB%`0Q-is(LYnaa}E
zXv<7Te+B5e@xYBAabL`n{%VL@M3xZ5C>7WtG5bXJ0Ri@i5W55xVIUNCiFjal@tV=F
z;PWGSqmue-Q444cqio1%wI2GWrm;{2e{R`(XD|@~+|x|^=avI5mWqiF--i1H8bq(l
ztPNu<YiPx*bEMpR;=YY+Pd7($G4iLZ0Twam2yJM+AYjp^`R$kq-rL>@y3o1_Ro>Ab
z4o@pmL-i)^@P9}0v@Ndzh8|2zYy|TC)ec_$8x{VtX*eyURc%HwNH<uf`nHQ%Qyh69
zP(R1)^j+t5sSoqqVMvZQN&22R1Eg7i2ENcIP&vOQ=0m4fs@daCbwjCl(Kjj?Zv6T|
zTSuP-y-DBNUG`Vm{oRR${ssz4s~cjXGbMN$M>PAtX+0Adt1hXZIr>$lD#U$Tcex%J
zOhR4vzW6+SLkSUK#adIWYwsGct5k58F5X@G<sA1ZM!@o(K@8WP{36$o;L`&QgMVVQ
zo3oKXE_Jmf7X2-yriPUUi;$5lpO08E><-gv``(q(-b788KDJM<ddhR6Z)^-t{l)$K
zop=%LOaF2r|8lGU3N-#bD*R6v@%;ZQ2GnfLiMmR=CK3(XdsG8Db)*1t5}S39=<2;7
zO&loPhbb@q)MM?5r$({NO1F<vY2gRWr<KfcX><;8;B{aGh=r(+VdHn>I|=zTsgigc
zC}#P{nWIRZw`W;!4TgS5=qA5TgrTM=0Kzw0;!hK|iQl`&6hPX3*(4IkPFUI&?UXBU
zedaCH!bqEIB-RD0K93xjU5a;7&+pKJE)!S3OtcCSs=87H6DD94FXSF0Q)hUh){hi=
zFATmcVjf9}{KJh|KVnQ%m8*Wyd<2Ls46>&H?&t$|w_rq5k-uI73BcIhW5l}D)m8)R
z0@V9sx%H~mSOaujOaVm5RVDYHpk$QEO^4R>9Z2<z<Wx@WhX^`r2cRg=#pk|aS;>%&
zY-V*W1wxhXh{LGZ5Ybdn*WWK#Df*Q?IWA?uK1Qcc#YFiMUp$y=%zQZeuaEv4C2ACa
z-95YQ#PvEfS$Cihi?E5ud=9x~>5UBJDo;r%OJq&tOLW*f3eNpSl~&W6D)kh&S@|HD
zdp8F5n?aI9il+dga{qdnKZ2S(oE(p$eji<M<e8>A;!IpTiAIb9KvYlwxF@i`T{ACT
z{hti^g;gPn8m|9!8i9Yk1pI51|Ek`9m&+#5nF2VDhs`3$V{rFT*it`i0Y*AMnH(g7
zjf10gCiqONRNPgQzAKh+#UB4QDuSE-+f76COt@3@P8m)S9a5MF_1&8_#N9W!3(G~>
z@mZAZUP^Y1{QO0l&&&NL9Xg&q-;<j1R+)BaDiQAb<hk+IE?mGS$M_1ieEzoJeqwzI
zeC8c&@K&1_{}RjGWXtgV;O>6mo$P)I_z`fUM|@W|dE2S=;Y<kqYvMhE)RM;YO^jLW
zm)<Xk@DS1Ku@5N#lZG@o?;Lm0s*gy<_H{32W2vjIm#3CVXk)@A-gAuj>}!zW@XTuG
zR*)%1n9n%Op&^am%1Sn?Dle|2!Qq*!?0g<+xli_?W@t`WeyhQ9x@%Br)T-+|a1*{9
z4;UiT5HgS>P~hF#mXT#F<So&NRGj0V_)``$YSpzNQh9spdOvp^W6^A@%j@~@KU$sN
zw9*G@5e#w(UPi1`EtfBIL$<4E@{K<mhm>Ecd$XPH6t_A*v7nRu%JY&-n*DqY>)~o^
zVos7xZ|5OuP*<dPx8-n7Znh7bs&JZZ*y#RsZ=mXwZ?Fi~el15ALkDBDfSM{98j7)2
z&D&X94MH4WM*sX))Kc%I_+hv$W_xLxe2&0pK?uaIc0a_4AGQ`^q9Y5(e4=L+RvF-u
z$@w3<o6k)#xj(woUbJWq)3#1>ucKl+Dii=*4}V|7JtTVtvN?X_DD!ydE(Kr<#G8Ff
z2Ae*uXt#K+K2S)S_y%-|CI?884zBCyYwO}85NkOw)V=s>Dhi~*MLyxC5$6$^xk3SC
zMN$C4jGDuI!~>(#-;d}%K7swWA6`AZ5859F?tZ<)X>3f&kvsmKeKeH2%>f&{PXV-s
z0sfN>d(j_<B5Ml)Psm0jKJlQi#q7|Nb76t761S|2LlAW>l?wl&&TnNOAumg0kiC4n
zmbBrfju$S<zt)3)Z&jHY3`S=#1zLzV)+dtgY0){(S8(So(j@(m#KHwBfOo(SkpsXu
z40{elmWa(n`t*@}C;-Lo8hF@ATxpidH9s(OV{M83xAm>rG}G<e+L<a@dKU2pdVynW
z4QRO$#{7Mdw#fTK6qnnuYf5_m;lz7K$TP=R;+@T4)|{Bqy=u=YUe^rZr5OON@&SZ+
znM_TAznm+f0A4Flt-2DdT?E5+0&qhRXyqsXu6rX3W<;42Db}?xXf#?Afjc+Z;g`NO
zaa*&b$l7AoYG&Q<RUh}g33NF}R&a%$LDL7{>mLE^`m4p$M1k?ukwd?9CN-9a_)UXn
ztKgIH(x43Mn#{LGuLea!#w^Cp_}e$FoY0fmDFAlZ5UnW%@IwGbNC%?XaP%aAGKfe+
zMJpl;`TD-sp-#x9(EK%|u-6(!sj(;+cq)82rJP9txO}jA=Dk+!f<KR6^PONpd@!P=
zy@_HAIk%z7F`>l2^G?+&B-BAAbszfnaf82p?9x5y%no+p^1CxO_Loilf)Mxq#<K2z
zCzBMa6LdUcHSB|J!uHk8(QsrYtUcs<;)bzfu8V;SHdD@ET(-QdTtYVQkCgpO2Fj0w
zI65qX2x8lGCz({t=dU9D@5i5Q2^+7v5;gg^*^eh^5veQs4vRSj0H8)v7+0DE?f<}V
z|GMk{ROWw4;$M>ZyMg<cPW)%r3FA2^$hsy}zhKzs>a$pb=0wW}-@cvG7lQ*B0Vm~A
zSHG~rFwPxnh%f}JoDT`8|K1uQ7#17b65Efij?$O)_)=@K7qNWxn-jC1&P}`Iv|OUo
zq!LI-goUV!B}bzrQ*|3BWeyGHK*9GqtgY1zJlA@a%PQU#W(qkJn=l-|;V)`cgR&VD
zQnBIt$Qj4xNmV|<fh6fjcVBmLcI_goKE=GR!wD%@=iivIBp-*oFl7@V56oz!w#P@{
zTX{wgW$ZiCQQ?q-XzNTr+orLdj9+ist@Aq%_V{yM856%IGXS#EiVeGD4MFCGq(&-&
zngr|@?2=E4v96h!{JDA1@LMtHDOR{P*}k}c3BuZ0fYO&%NsTYDj_?qRjnYeK2vQS0
zQ%WPo@wVRz7vB-jK-7r(X~Yf9)yB@C^p~eIL65?RY<fvO#4e%Pgde@DtuR6{HlsVm
z?q=m=vnmC^@JFL*_q%PPnr!Ogv65Kgf?i&7&quce1%Fm?Qwgqm{x0LL`9>=?WIQ+V
z%0x9&ACP{5a570|&tPMW&?$D>Ur(H^E?u>BKE2&bm-B>8%rjMD-YD@d;1=B125#$4
z<;Z%sa`JS+AcO9}5D$u%Tk+fM8#xJ(Y)8e`X0ta7<rC6RPety0Z^CZFBb-BRpXz%$
zUTA&hl*a7B%PxBJQ!Rh#Teq!Vr<&70f>cJHVck14Wrq*Ja=aU5x$CJj9)<48%}J#m
zYYGfEo^V?X%y$P>lxw;jZ)%e=h{1eF14D%+WC{vgh1Csv0W!$K)3U5gWELCwyn1SB
z0Jx}7bm8^89?ci$x{IJm_i#p0<;2UFKv0+o!E3d<D`CTL#>;Mf&7`r8JdiasZ^ND1
zJ6Cp{CC6XrXAH;wb9tZA&kbti3loU>Bdc9quN8-t^+LEf1#nOXWa%P+@E>uRN1@F^
zOR+FCIl6!#*HJZfSGX$`;|hmww@zQzrqzFeXD=37zB4QST=9bt&sG^uwB2Co?&3Zs
zlN(yyxfV;e!A@%Kj-4gOE;!A-`fSg$L{0bxx7u={s%XS(p-vUS&hE&AKp|I&(HamZ
z%1uQjHRA1B)cB5AjgHPhSL}`KwkTvO8WLk=5Py!1@m3t@B0Rb~jyws*dwM9JZJv4j
z>r{8<zSh<ke>=M&&0vJglhSWq-mPTwD5A#N8X>13^&<=(iNYGJ9i8@PckI%;*(o$)
zV+*5|)ZjcbUlW(^y*CNl{rZST8hdUbm$^{P?_!O!0d6har$6B6{%DYP-iF|K%5IJ8
zMBT7~=eZG{x~@{omwNP<o+j9xMN<IbtqLO{Y8L(Zgg~>AAgPP+;tjvSQk(UF@Ki1v
zv5q-cD&TSXyoY$AwzDo1=lirf&+7bcVRaWF9c)a5Xf!z)V;S@;5WHB>;4>zB$ik{2
z=;R{)PPx9WR(0<<)WlMjIq;!Hb4*F&Z8-xX1t1jQ(H^py+zOgJgeb1Vn1dc-{bFm}
zjUG^|Y)Tri$0M_jq?FWRtyjvq9#WCU{FN6%HJ{$1GdBb&kp~g1LC+RZ5fivR^t`ne
z)-LU9Z{c0~$a1BoV;M1p4*R-M*)qgP`q1@TkeUzOFT}2@RUFhjNz_+5q5!Cija4!)
zFQw;dPSpRjdUZqD=$JvDjG+LmhKFIx7qWMO<Z!Us6KaX>WA={RVVN)=`ED<Hp91*p
z^jgkXn;T$kIdc@Ll+8FLv<-ezIU@aWTUGUlf$r&v$Mdr<@f1LHYI{@0U}a&42pfZ!
zxm=M>ardc17|DNIzq(Gf@Lri|z)N?3i9!9GlJ%3=f-+~8mDpdZwMnwScVw11b(Y30
zE^r)_aQ`|(wYmShC!5BkMR^&z@^<tw8~K5a`P$$+3INQ;vw^A{Bi~HLhq;Exc6%z!
zKzXDOEc-X}r{k__b_#!KBw=XwrD8nhKQof(C;5CU(D~~b$Mk=WaQRf8$V^owF9NUt
z%6AGObBa{h*bGwHt_;qAWh(Z?%LOh8;f@-*pz9B=Ef}^&E!0i+dHim5$SJL9JT0XF
ztS{6ipW3Wb0NBsAP&sofETYV(pG~WJf@oNzlTuvZGHh_>=eAg-jr2PS3!W!VYC?=t
z%HTFo%~4phV=+|anmQ0;ORJB|{Ub|oQW*^PsTvo&VeNMOsyTu0ktH3jyPowHrZjsx
zvHcW-=!EdPH6l^2+3ZSc!Hk=_adRCLI9;y$Tav7Pyyn36reH&V3ZM-&Wj2a7iU^W=
zIg|a)heZ(?vkehPyq1<Zj2;-u$h@1bBO@ZBuFq8=AT#kzVRhocBn5EG^~h^>e1W9s
zxe}YR3n6R6w&x~~t!6*J-xfIabmJ3YlvN&lH{%!a^?P(jZKDEVacWvWzdv$ZQg6ai
z(CqR4rKFzOrd)63byZN!<lCTgD6U;)cZ1G=DrjKG2N#CCK}QEW>uGcOnf1Q&I?2{a
zX|sOxs{nKQ5?C9sSz8k6NN~<2T1G!4K+d%)_t~m!BJZu9+z{UnYqAY4roEITeLhB2
zJ@KXnFdFzQ5XLRnAr(*fW3_U;SqrlX5^z-=T3@69sE;V0LS9bgV+x(7ced+r`sn?D
z<hKj2M5Gdlbw&uQ?5;WA^;xo=_24lz^`+G#tA|WwiDT)Jddl1MXn2HonG=w6+9fpr
z7G1p@dH<CvD*S!}hhcrtx8VmHW#lRsxWq8bHQ3q)6!!`W$G^)!*EXQx9B#gvBG}o^
zuL9~3sW-mXB~Iv<TWCH$^5X1#$)G;jLewLr1v%ni=(4r<L|dG`vhdT*lz{QD<%<`h
z7P^pII|9$F9W(lUJpB0Toz_FYum@yMFD+`I+Gv+%W3@US5u#2z7ewoE+<VrpT|cji
zeY)y0%si6#rT9|H0IQnAujeR@`D)+OV)C~@Vx3e*YI1(C`QjQP6?wAtGK+>(E>}im
z!*(0sAa4)n?J<*)2U;aIP!O{ALs`e8%a`VU)^Xkckh@KHEoi0owWrm4{4xRNgU&v`
z5nvnMQnGPTa?Ha!Lxxex#U1{<sSYrASH2<T>T~a(v1PIr-=`bvG_QsXre{ZK-f#hO
z2MJ@Qh~i7ass<VN({>N4iY@1-oA}NTMSi{hbNRa??D_d}IYUzPj=Cy#P!C&>hS`=g
zV=9^74A=^}fz2$hDBLtxrIqe~VW;xe9`N{QvS*L`@wbnkXOqi(4=8}TYJ<Mikh<A>
zqC7%6N0lh@{QU90`^fB(nxPT?+mI5Bg`l~&+fJ6L#)(AB<&9Qn)7!ZggG&vSoR8A<
z^YoIrpT&gRetN2PVZyHX)Q!NWh4m;da|;M*DcrEKLx_aSso#3#&Ir0&r2orP(I4S_
zzHgr)X2DpbJ4<R1hMo_&LjW(9oh2Hx7MBss%^SyG>?pTzs%?vv4WMohtq)3^ZRPNV
zXEgBhvh@238f_&6e<3PQ1gRAe-nR!h<(9Y&gXT%^(K{Ke2^y6zrSrdwC?&8P4UDGU
zy=89v!4R|Cntcj^KlGpg-Zzi8ps3Kt9vHUW(s%;5A%Jj-WL}R{m^YaRC;85~vJq@4
zP1JYzuKo!Sc?G(lR}k|S*|o*WraWUFP40(PMTHFdl4WK~XHkfV8o6!W(SuCYQ^g0P
zWSjHbGU~lSUtaA8t=+Wy1a~PM4v=e|A&cpppqhvWM5>w2d-Pvi!=LWq|HacxgxXTG
ztcznKy)6eGJT_Z-M%X3ahf$HFsoT_OMq3#G%89yqrfdWn=E#!gmhPV4B-hV_|9r%P
zkQ7rGyhy{p_;wVo%ttgs9S9ErvARGCK*9u?VUley@;rUP=-RI)19$ou`5lcCBoTOO
z34PTuvi^L~bo^`GxR_okq=t<JWjQ@Z1!7B{G^9^oc}|6SXoKdf?u}2=8yskKlcgXh
ztiMGFqS@4FmD(o}HA{ZmNKZb$My2}a3bhC#Jq}yjq(&iy1Zw9PgA)~i1A>lOfbFku
z#}mpBztz6L7UN*#KXU&vkkmjKSzczfA~1h;gKG|CX_p^1_Hum(!;nSu^6Qjt&6BR5
zeo1xGd}<LqB}LG^UH$J3t*v9|%F5c<f~?i0zPOvYj||l0`Aq?w`inHW25|#;Y74^u
z_;H;+@R^MJLq4%ae`G(J@NgbwWSXwtWk%V}TJJ_W+U_G?;MnFHEBaotK9p)04GQ)t
zU7x7K-~R09Yi=x!wYgW+EB1oM%!K|@=y@gY`ndn6q>@TS&Z?;BVK$r?o{fH@42tCX
zU1_1~Uvz8&vMI~2l%|={e^BL4Lk)Ep6*-|a`eHP9{4)`MnS1Ii@YIp$z1fHygVf&4
z<fTUQZPhPetMS@2R}6!dK|}h?$ey8sODRNLsOkcrJA0_pvGq#qRA%AOq!BgbjAFxN
zSZ8F)JwAP+a`bBIC(3!c*2Tdk{||+jZSQ9ZZC!Uo(^bc5<|%+qY65e2o|XcjqxMMd
zIf~${VMBg*sJ{6E1wc%*K@i2$;Y1Q_S&2$5@E`Q#5&hq2{0RHoP@Zx)L5zl*jQF8&
z`nw_JUoWdKL1HaU(;m)U;QQmXm#S<))IphwbN%SH<A)u;NXt0+#5)r4H>kibrk08V
z@4TS^LT^#4ObUNd!LpxMY1#gwhJV-Rxt!NG{y#E?LfPf9aXUk+y$SsK?E9)3Fm(J5
zzmS#gt$~Mre(RmyR&vC*NhInj9O`$RqX1akTo8M#Mn}|Ds-;8|&@6>7A{D%#0I0P#
zW+cz&mncG&@ngKuHHK+HR^~mLfgvdErHaLu48rl<a`AR-q^ux5$;f8MB2xGdm7eh@
z)roaSrn+Jz1@PYJl3$k;a7MK&y{<!;>xtpxd;W=C{7>`}Ew00RxpIrHTeBVum>4V`
z^7Ka(U-zENZ8-R>5sF;G;qoJsb6XH;cygm?8m_y8jf1E^DK<_tESU%M?S9sOdCU>|
zs)jE|?({mKJfL-k6iVd7R&^|};4YA_R}T9aad!I}mmTKE5=)F^7@J)sX{8<d;&cE@
zF~k-eNPAdli9VtlekRvVP|?lkT$*tCdh-u!8qqF;*p9Hxgt!LynfzN`-(nZmwm+!{
z3p{sQyf*6jBynOQzVBd1ADg=LKAOz98Ce7cKS?}q5MDo;|9qnEj&Gv+A}*o+Svd3x
z!c6lV(Pg;xHk5U;!%9zOZ>_hb!nbc`03iMD=t8Jaz3A5-k=Pgc7A0r+FR%who#c`k
zP)nQHY(XyBNbR&zbuH&BpWZAQGFBH+)sexROuubwaeFUjT;;%d5NBqj%N3TT@Y&r&
z5Y@N|GaNk@$*(rn4(vvwwYeYB{BNA}S(~NiX4q+AZQ^@spRcUcibxo>SMt!%uN}y$
zY%8_&21fzJ`vjJ$Xqf+ij!1BHD@D}3>vVNXy>VXDIVrG#X$8HXeuA*Eo#_rwEr2YG
z9AEQ%M3lse5U*3yh!B2k{4AoH0;upNya<_IKDP7iWI{Nt&VKgpJsB1jTR!c^Qhj?!
z<VQ07ML01=g|MAdP&zEE`h?tJ`_zR6qw#_C4wQCZ^PxIQjpWI`yH^WY@>Kx;B#V=W
zvo7P#6{1`C+BAi6zNyasOV^r}12Iw-)t?(``Rsyx`+Q!`rPR*1pl4U!1CBpgjn<Lk
zyk&qGn}sXYb=db)l9NK9<lP`4uumDo4CvNUym3O@RLtl2xiof8UhVee%@-knvb+FB
z%~}aVxh|{t=m#0<3}KVAz~NSE_8T&NbUc`Pq_cB=TR`IgK{i07cJ5UkibH~R2T<PI
z`v^%4Nt&RGe>dE0UFynP@jHDrJ0Dzj-Q_mv;D$`f_u<^0cHzY8;!jHnlLK)#Hr|y(
zOIalG<=-a{5KOk~FgBtqOf^b+$A{(0&zmpIz!^~9x!!9szMRSwfF8g<CO^hKzUAVn
z9SbSjM~-y?s?%+ZbFiv=7sS6S`QE#Fl$lSSi!KEh{B7}9-_0n2OmhcgUzz%-mPL`*
z2<%&`<F2AsnRaJkXeR981oir22xwM)w~~BBdbWorrS1@?QDgQW@xI5wBd7slNiD)t
z2n5<BpIh+TTZnsqY6dm87^MJeJd|PswTB9@q`}BiAKM2L;!)(E^Y-BINKUhlN+E7R
zl{y(!9~YV@9kGL$S3!E=d%D>27IggO=P8P%;#HkK+))qonI}dKx!;B|zMpIMFiyx)
zL&B<By2HKW0ai26VoW18HC!%8jb}L^fH%32_lFIgucK}Ftfn<R3nEdzA5OL3G)C<*
zuR2beQv(DC%b9%K)!G|5hP|hnsW4#6scN>$`dp7+U%kT>X69FQ4SodJ;zLxBJK=Cz
zUo1Hbfkk0K$39;rw~g~D05?(`s!da_Im*^)A6c9FeP-rZH9fiN55XCYNt)JYbTQ&S
z7OzE=BD824oN*WpF!HuXVGdNJGyRY>-#`DkQ8~Hg+n}RzX_wHu`gH4eHTNp70sNN{
z;;OIpN(zW)$y6`<4n~%2b;cXT9J2Ajy0g?#HNjOR5t4IX-F7zY^D)fZW9bu4ZVRqb
zQE(T0u_0$eUbl2pu(Z)RDoAstAnB#tPV5g(E2d9x`xuT*OqKy7L2v!|H-Dzu0*%HI
ziFV~HeZP)<n@(FcLl{MF0kuch5!E5F)Nom6P#;U(gCMuttOpsE8Cf_#YXp{x*xuC~
zaY(y&Y@+w-d+T;=mFHEVbOQaUfh$>_V4DxxRIL-tUrk2W*MPVQQCCF`gczEWlXE^6
zgHYTOMzUh=QRhe*Ln`s@py?c7m={#z>%{iX73{(7t!oh1{IWb*Wucy(JNM0{I6GIi
z7-{G@jy$7&Vqsfv@}X+&wPeBS30@8AW(lMM6w)!FGQq?Mz7<pI7e$+W1?|_HdVExZ
zt~3>Jv(m9?Y|UytYz49Pmo!OS-P@L>HsMkLYA~d*&*U~oo~xh^Jh;q75dRK;n`JQa
zSygM?Tz|0rnYU8yqG@9ayD{7!a%rzD`*h!O#+?Kj2@^t(Y-+|1K+LlQdxJ3<Z|feJ
zesn_vqPgx5|A)t}%}p<h2D?_~IO^WKM-NU-^l4#_qLtcI-+E8`#Iw#reHTuVn)Edz
zYxgr%1a=%O1=c<Q8lJe=163ZYJvfB1Bb;pIFpmjul!LV$tH-^C6?J)c9TH1TrsfS_
zUHEk&Il*$}(XhheH+KE`PR%Df2zWG(5CkvJl(^kK2qRFVifyu0z~%<3S|SqN&(~od
zu$oZf(wGu1-QN}MkB3y(H>LL0ICy7%nHBr`LGi(vWcL+7zc%dj?##L-PvzL5$T>o`
z;W(rQwq;<%W+(nx)lXGI<)gn>O;3%teC~}8M%CNbSt)>&@9I3KiglqxMNEc)KoA!-
zW5Bd=!P?ap4Qzpb$m{6K8-Z8VQ<*L~nX}wHHrgRpr6@N8Pi5x_5Og^l9c^gd{;RQ|
zbS85k3HZ*$i1mehDa^@GzFhHwzjo@%XD(Tb#Jv2AI|#Xc5VjTJwIgK-Y9dw>Vd#^r
zWMHR(ArP%nxU_9>f1)~CL)EBZ%;O;A)HI4L{>G+PAf|N$)#RRcOQ+{@)Ccez1gX?}
zURJ}hoJg}|{VLyec+(FoUh_%b6|(xAKmp{x%x+AQ7deRkNM3tubt3YEXBg&eXpMfp
z`IK5LUvCVH922dcW~pf0_qtTO&z8y@;xQHtV~lEQ8f~`3@{u{?35GCCARa={R~O)W
z2U{A1P2+=0NTQ^tsO*~H0nN3AwJ@ny4`#iIFl#v-lBu?KQL$#@iJjD*S4eSyhQ{Ni
z?)TO8K?YK@Zd|%w+)g}VUk-H(aP|kt{2_PI#*u#Z%G|^A$fp;W37@m8*CG$3w(lV=
zD#vz}9;v(fyn?$vvJC###iD@i281~#n5N(Cp%0?&i{SL5xh6ar33Mt0nU5=nRp3f*
zhrBIyv(l^xwwH)svr+DUQ^Uv!xpl8&-I-r%C5735zPrbTkc&5JnyB9F1IzIt9VUx0
z(Sh%?^%dDfEmQj)@5pk2+xsr^C-R>=Hh+1&`XQN`go&HM2t24j9>cC-Q`KhcgDS+5
zeGLw@ExZV$heDpT2)09$j7L~L`psO7RN}KLm3gJ~sB_g?Dm1UTw~Idb%V*jPz-qHt
zZ~-(5!v6cuhK?EyP9e+HVIRGxi96682YGA$kU^!2-qoLF>@<59o`sHsW>ni7TZ^%G
zm24S2WR>*}HJG)<Yl<_IevXw{7Ra-Hi;CGEI8)pUaQe0oGDvm4mv9%qVs?3TVVHgs
z8BHN~1Kkla(!G!rjxH|~RP#%5a^h<6laulgxjJ=i9S*M9s|=7mIUH(N#e*|o1Zwsn
z+?cK^LN$fC1Xzdp4E{|y6l^-w+Kn%xHQ~($Kl77Ia_JTGWAiYxq`A}e^qXnj$2R4!
zY~4(b*zwlfBB)VEx@L*bH<pW|$fF_avX%K>txN2Uzb5uB$0#00tNh^D`%%o3%+>~B
zU(mvoJ+{W#r4~aGcGWebS?lVW=TcI9cSXyZEDh94*Dt3(lW<~gBR{@KQ+^#dh0xDH
zH@9AFt;Zhrj}YFIISthAV@KH7X8yRm7xZpq!TF-rWv?D_X4}5-$1&nnh}2a$0d!!y
zYY<Fb!l`?!+hqtg53FvybkHksF~3ZC_3XARu2fn0Ht|x{$jeBH?yCdu<$Hy~b)IU)
zz4~wq!A$fU4l*us@$4?xNU;J%2tFA)EUA8HwGpo5&-RhiEWuCrLZV!HWW%GS?}|Zu
z1T9x=Qv1=2mDU%r)FX(Wr${&vcrNxi)AcX+>tZT?n2IPR>MbYj+q=O(6R(lt4nSd=
zvLiJtnsVn0GIRY=)BBbLl@S*Ax{@oTLSyOGdvD_VFL)|Gg*XYl6%GN#))0Ls#}6g#
z%onI2j}bvU(;ojV*Z{2>qYgqt=m?<}-V^}aJUCr5AI3t3)T4A*g@VlQ7igNh6v+Sa
zNM%g=Jnt6FRZ8QQqiVWHv@(mwzD3n=(w(@4&a9a|cNvHof0{Vc%18QFy;APJkWV@a
z4b(2O%KLRzsWH1XM(i#)8$vX4vQ;E~3E3)SD{Bp}el55|EZwT}V~S4>?fBCC$+-UP
z+z`-N?No%_0-RnL<qoeoAEXV&2y|wBoWSV~t<xLsC0qG3o*T?a?2m7Hd-N?;X6sar
zBn2X-+P2)zwe;HTBW(r`uMj<4pt|^!ks!;`-?ftM3lhWVab>1V9=XG-GAZYMq<qo}
zDua`%QjH;UyBt=W=4b9fOW4Un^0gI(7PbVo&Xaqa=b#4$Wz=mc++(Xz0+M@Xik@qC
zwlA0{L>Hk&WEtqA;*ALOc%C78{0;@M1sB-Rq;>#agv_)~ljD!bw%IREiEDkQxqcT^
zLlO2}SVV^i12k*X3cI~M`QD=-(?<=)bl#`M*lr>)2;?U7O7tPM*Uem-gT~>sIkrKq
z*LM_G_j@O`ppLu2TvZAIQM+Zm`#|QT$vTgkwj<p1e0FJPp}C#-#@2*l)v$wkyEWbL
znfYzoopR?;E;mtsaaW0iGw%}r06a%qwD|I&stXdn5@5Eilj`yqLpIZ4r_rS*EUANc
zAGFm{&u(oXlAxZT7qjxd_wuuqdZ~JVzX(f6PvyFuP||}jb-Owf>)+{i@dXQ(t}y37
zol&?`Q+55*Q9ZGjBRc#+Uz@99zhawC-ftb1K)<oLopxT7t{GD|`ex&Dx?8<{xy=1i
z%_v-Z?q#>f5$AW=Roj7)EF1POvZJL9j^gfIFfYWK`-$(?HS5e-cc3^d1-}Vv-U<@N
z*{Zwd5K;nRESpv#b?D=Tk$q$x1>k5seB+JeeKY+T+EuZNWQZ!Hf1jtVJC4Xk%2RGa
zh1-d=Odg2vL;E_@Yd@=N1v@*s=fm+nG4I{<-blp$aW?U;AaCMIt>%-ZcnCK8H4x2*
zGmrDht|JwaIfL&B?TYqJ;~}yGOzs(V!pD26&P=gW-2$3mG-|pVM1bWILgeV7n8^p@
zGU{qzblHQw)BCE%3wJ#nTneI;>tm1#X$43{z4{>T7mU5NNnqbploN_<=(t|zjgpgT
zo?L^^!69mPYtCPYUc=3nOYTz08P<$%V0kw|ad?DImHPz)dzQ@Z&zJoa^eP==#N3}a
z9gi*1{H*KD7FL2VxZOT5Sx+w`XQIbr59jUrFRRR0cyopvD3yr1N2ciPC16V#)}kQD
z>`)aE<{({`o#bHA_1Xpg(I7>Lv(nR^Vpnlwq}81eCm1lH+ADAEGb<BDcPluDu!*@?
zgMog<Mh5$T3DzfSEpFO6;U%J!qkTT_-?Z@4iDK-%&{yxlsBm|}rg>u)eB2Zw+?aqG
z4tYQU#1i9{Fl^mm*n2OxI=&DVoY`0O_C@jRCpezu(lb9t6y8k#NO4H<d2*?uF7DAS
z5hxEQD#)|qZMV6m(0A-$r1I}Ar<%I4!`SP+;?<-1^`<U~j~uveP2rmOa}w+=c|I5Q
zUIgQ+@EX)m6j=;STOKFg!ppdgyQ6~x+mWWN!x3+uiA?X;{1mTr<h_@3#jonakMFY!
z@OBYSmj%sn<#-n8w2~p4O=ETIgwo^)CkCVcCF-Wd?7+y`FvlqV1Z3wmbd-WT-5)ZA
zs)w5^;O(;Q`uvqy+$f~Gj|$>qv+<!#V>Rp5Y@F+jzheC{{Sd9u<~tJ;xGBZiIh$&g
zx8O)7CMc`O&#MpE9pn;hB|F&S1XZb9xi<TXVGtA*?s1)YvRi0j)g}#EPaT=LkMmB>
zS76ID;9)HO?jg;jgnEsjHo#-4RpXAMU*uj~7S;a}_PaCi@LE_d1yFkIORC!4&!*-h
zb0QqAXQ(M93{z%qMV>#UpDn94fQfE}I1O%*^mnV8ehKWW9z%c3#-dc`b5xGF_KbGS
zVZr3xw9;e$f~Bf)*dmM=mW^J1R7)!IUMezlE`o}7rYl-(Of;&%Wwy+8WUgbD+hv+p
z<JLQ(C1<X*z3rd6;GZD{wjkgjXi|ujt!`WE^~#~V@hiir9iuaMQgk--@5>j(7Vv2c
znmyr+P_L@L@x0LI36AdvXq&o`!jOvG1unzd4KTuzc2tBz1LbteHihsZe}|y(axGmb
z#a#EbFr3I%z=9ZVWOMR1)jX=hIodG%%2K;bOql_4vN+k09%a6U<EKJa;!h%dzZ&m9
z-ML))wZAs?pc+E0%vcXYCLB^lJp&G?F(b?>Gdlw48Y^|Twc#2X$hFaUwzi7(jcTOy
zr5crMK}XkK*dynpJH#r_y3%RWn9)S?m!<p%v5>YPFglxSju5(pKBB&};PS9+vn#l?
zHPs(2@Xk{bA7fspSpHp11>q#_nDA^ya;{B5bK8&V0xJf>3W)|89cC$H_EWl(CGcL$
z?pt?{ny+dGYwDgG3i7iYNeQP(SI*q)sW-`&QWh|9M$EJArv@rJ`=-BoJiltZu(ObL
z*czaV+bR1k!W|;x3Sw3+@vqUZc^}2bwsEqYVilya%Z!~aWx4j808?TB#7JGw97N<K
zK+uz2LUIuS1BFz-@pRMf!dPjUd(GL}2~4<ZLv_8vc@I&JJ_()}lQZy2ZK-hW(!rQ)
zIvCpBMhP(snzn9ME+qs)rh|5=k-r|_!09F#-ZOy0%#V8xEWE)^p0Z(-*P;;BLDX8k
z3-4`_Y4zBqkbaqtEDut*TI+CZ{gU(Jtn)JqM~9_|Q&0=B+-%u~TGD8XU;jlm7$)da
zfiO2Y!HtDm!}bkWvpVsrd1ygzwXlQ`U)KH>U&`=Z*`tQ*)L6;%$Uv}a&E-XRw(Wa-
z=&*9M@Rc=Wqh=K*WlYJU`Oec((Kik?m8qIGiIMp+j!`{lE;~!8E=BcWji~!U{Lpd{
zk(g^i`YhYAH3Lna;OFVCQKo{Kzf?J;6OVt1^g%zJ;qWS_5iqe18$&!7mUWkGH@HCN
zo=}eELk>Z`0ug)!O&KSKXYZ=ShT8Wphdvqss?F<NQm7reUqJe7T2qODx4>-uL-0x4
zCd(d%e!d5*)GCbc@x?)mQH<RM>~8y1P>4j536?a*TQ%MC#We7%7cRdMp5F7#Hb2Hm
zN85J#`4>*bX9V<%R>8V4A2tR;lJ2-i&E4D?XvhhZPjgelRfjWtmgx)k<-=j}PE%(z
zEqJa}@3vm31~G0Eg>vmIT{qpGSw4E%tn=<$=|ylWc};(bQ{;|ftoA&HGT-A~jK30(
zmm6^+K;+)T%hferdADAsrVfb%!|F<Jk@%$ZLD`p;>e9MD^w?A~Xz_WWKQ0Ms+PPYT
zO9E^|sMQbIW8Fn_LJc*ebg%e52@<Kcwmk>V&D-SMOxReWmKNSQ-9+Dh>vm3F;Ri5q
zd7puESK;`Xf&udd6YSdd<vehDM}RgeoT!Z^*$~3im1W|JS?Zx4>22`0irQ)=x+}7(
zxhfVhXEfc`s<WrP34!gm9Rx2HQ>z{<ghGQ!1o~2C^N_mKx$OLI`LMDXiHfK@Blm}z
zqYMTd`W*BCrWe-2$&7?i8>%0dQ)XlDcE-f2cKHV2Gs<mTWR*5+1(j3=O5E3^lrB4_
z$k%$c<qBm+JbmKc4^LfVn%zxBgJ-t$15bd6g*x&qf!LxW5}&U5#eMsrOvqwFIWX41
zWc2BXyV<;C(iYbz52s&AV$a2c>!la7<uz?SGn2INABS_}k&(Fsa7WhFF`p?PhNq+6
z8G)Wka`w`mww?0nF2lkX?={zs7>8Df(|THUZ-Jf5TpEq=ak*chsU5dHN=i25=Lqn2
zoSRipNs&$3N}K^uS<~^dN>ActAtV0hHM3AYd`K-;C4Vux$$Im=IWpb1*a4sJ6n!q+
zU2UO{t`yhFIu`b*?RCA=<u}e$hXU?fJK00D;}OShYURPasW6}UR^Pe}71)fuR>NDZ
zR3up#Cl_-rKL?2s;JQ~Cr!W0NTufKGSgh#RFS9#W-sxYH*JNLGe%wJtE0sV@3*cyp
z)Dn<umWW>7qz~94hxg<Wv?F)GJi+mKg+fpJ&AD&i2i47ms4-GqM+nU*Q9X;5@P;=U
z21fURm^b`V1qJ7WzX|#JhC|tZ(J2<EzYtJgM#gl_`sd@g@M^=&?n~RCj*7x}D@isg
z+r|ADqjNK0)gGKq0;DR-jDf~w*sD@o!9yrC=7V=o^NhRS@%mof00rQc>M&nBg7QKv
z!KAF$wid48CbH4OTJ3m0&pAWMh#KHliB6rVqa`&XKK-1+Z~Nh9Pj9cCeBw>v$N2PN
z_mDnF0Tm6qcx8MCerSDODfXsi{x8Jz=rNxOJlj)H;k{IiRvQ1!%Qs(mBe+9^?4UE<
zNE4Uzmex0=&{t35pNy_GepmHvnh<z;O<GFqj~@F>T<!f!VYe=^2Q8VBQhhYE(3^zk
z5DtPL$RO7I@wjYvw7TfUd{?nlKe?L;m$|xz`#F3~Jtam2D%095Y(rrA2+*d<q)aX>
zn`{Up)uI%?;DFl<UkR#@#^m5%aQ7>axxbIPb!~578je1lmiam><eoi<>PV<)|0X*=
zUoT`FGMuhI3~{Q*PDZ>SYJ*E9UaLz<`}P<02Vl!OJiqGCoDt;R_e%>s%hHhQ%Jata
zRJt_9ccK^5RBXl~ikx|rjz<lq8*Me=oA+&rZ4LSJRs2E$M3iqykQu>VGHj;?A!NDW
zR=II>QL<{X)M><9a%>=Fy|C=<@kB+Bcw<;D=Q#_{OZz(XJyzv5kuBa$rwzG~w!Yxf
z7QuN<{ODu?Y<oUlkKl-VK8tZ!5UmwFSA%ZRMIIFwqNc!}?Ui|DbgyCy1RQj26XCx~
z@80UtVdxzutIcj)ygsZH{Ac8W9TPklnf-lIP~}SZXrU)Ty;D6ut^AVx1(O?m9NO&(
zZ9mkh=~!<bAx?+z9s}aYs7>k!*s!%S^vi2a+}9FLu0ec~UbdG$Zzsu^m?SZ7(y`P3
zDukXCpV6Mpp)N|@>hFxKtFseoN<sTb{gV`}SS~<))~n>C8oq{@pIC14dN61j7zE#|
zA(K;7LT(cFZLo-V_p%x}PB57U%RkBs{nPxhYTVZka(s!M*B5VTpY2G9m{Qc`w-*gL
zaPPFhns;P#Y%C}c%<W%cH26T@Ck}+2@EfPduXcULE)rGNu1n9EE>A0_Xg~j&8E$bs
zeTS&MxXX=*TA2(|#2k!jj=Boxccqq2^%hw;srXucjg<_``M7(1S>B8>&88gu<~`nf
zQeDc2x<J~pWJkUG&2<c)&g+-`_5!u}OXR=e=||>y+WvTU#$sTc`G?fg<(Ra>o!N}R
zy<`M0f<;Z$S&E7tyOUwDhYxy?gc=-^X$9fP58a>+UV=0KP9Y0bj$N+98ijFLx=JRl
zZi~6b<K34o%`!i`JH;v_aH~AJV?mDko(^t;vp#k(O8GRK7g}O0XqQn=NYaiQee`}c
z(XudQc>Ag8-j{IquU`Nd_3-&5uYEMA14{G_0_8*3+9w)27b3yA8`u2NYH8amE|T@h
zj$wQzIuRn(?562PjjTJK564;o*%9O`!!<@x`vge$#EfZdXR{!XDEeHAt=izae8XO0
z*3ueBS4B)x?7&^2?>ulG`gzpI_Q^(Rqjg{I$Cjh3c(cP5gHc`IBShK{=;%^Pu4+3l
zETFt_$r$@H*q{S6Wp{17t>Uh;I`3rkEB?UxyfB)ET7bNrj(qI+bPx@0<tQr17DKw-
z*R_U$y_I&~${!_K^t$-X^*HhI=74YJ)@AfDFgbE79M4)y<|+ZFA%@ua=(1h)TSZpC
zHF;}rVww`Q*#4AH3Dz?MO2?(EdQl4R^woQoe;6evoE)ctJj>*osc}>yk*fu<m?WDJ
z*9+SoA-aQGjJT5UMI+Q=QAjsrF`I8eTeiv{J!>%K?jDrU|6sa+#r@*uYt)TU9nQwm
zJ8bFqFTWm*#>b589vJcl%ON!{j~|lER!>HNOf&i4HrRWeuSM<lyUOy;n@Q3b_|IB2
zzL2RmaYfL;s1+@=$e&fMG*Z%%hE^vEw-&Fu4Wd@IbtwSLz3o$H)S4z}UZ{THEk1BO
zNW%wQO0e(rZS;$h@iBDUlj_&f1DEL5j+M8L%2a=s>3A_~`Fv*c`u;^dBu|kwRne^E
z#;N_O$_pD-yGaPNe@K+vgwYDMm54rP-X%+<ap=Gn(lJw8KdBE!lQgf2wwl$zWVa0P
zk<X)m5`T#DemK){c0M7?8V%~+oHs{;NhJkfK~!Xo9Q)?PgrIP<%J=zBSLZ-6Xp(nA
zM~>Kyug{k&9NWBJS(zcKsNFN+?&`N-Xb3mbeu+i=P%_(Wyn3+K8bk%SAIi$n)#Cb@
z08URFX0J|vr)_pt&ni9@wAvJ8jZ@Yzx`We=IH0~PYRqJ}GlVzVcy!J7-r}HwzVm9~
zuTm~>UWCqQ7k?QO;;?bHujO!|W4#XoY#%{kXN~H&4p7k}Py{9wIUCHmW6`WpvTn7N
zI-&R@`uhCzoRwC^0E1cNg7;->*IU3FjNmD7SF_xyM&Du*S}2pvY((jFmv!B$EXH`2
z5FiES%fefT?`zjIbrcIK@$a?GwY9pcp6TC;&dl|sTS`24eZDktpE!?TV#?u-x>^LF
zHZ%R(Sqc=U0D9svBB%5=u-^q`Kan^TF@j8u7{B{;lcU@K2x=MpoMaqu85B=$q*~iD
z&;=qJ!4m_!K;$C0L*USi2QY@&S=X750Up^o`Yf}b1SG<5$=Yq)cVvyQesSHJU+5Lc
z-EBq}KaqXmA|$EiXlJ&mf`3B;EJS7(ISHXY9Y{*%II<4Ork0+5`5pNiMbwl9b_7Vi
z@evL|nYJ{A?slh#g;#i%K%;hkPV^auZ{VIA-<Hm{kGL6Tr8jhey|nGlHD@DSZG}G;
z&s6rf1CJ%}LU@DI_pOI~W}7-IN65b{;P|``?b`y4j0I?hsY;&xb0g;4Mua=C^-~Jq
z5(R*#Zb7GZ(HNnQ>8U7Y&TT0TU*auTA1=gC;y7<+<cR2x;DNRITchD;HhoiCf!SmZ
zHg?cPfbd#}6xgRJ$t3-^-92@6Yx{>={px|kdQQ&tZXX`fd^2_v)QxF9OBfJ}7bGN^
z6W)5Fg&#mTpg|kvvMF^@?<(~Hn4Y}XT?vIB#qa$n{c<6gs*&f}NI=-K2>ECHKN4tQ
zQGaYrEh^Lz`SxZeWffLFJETkyg<xff5)r{5k&$L?42yO5<|0JkckO~u%F@FKn+4Z>
zFdr9fo@Km(8oahh(&L7$!?Hulx*5^Zo*Hn6pL8Xm**FMIrr7wmN)TpiY6ugd?C-tW
z(xQ_fJ})LClC6+qo&0*oY@dJ21NHPDMU?vwlgFb$svg}ll-z5Xw7+^<VZ04#gAYs`
zXmSuTmuz7MLqw5MXl_T61-J~|Bv6Q*6C@41DQa04&r2Lo(g`O!-_!zM;koziCg&?Z
z%u$CC?>d6L+4{MUp&WAZXDtF7O-&pLS@?Ey7DjoO>din)^GBNg3ASF{uRj@whEVx<
zM^Da1?6FbNLS5n|-~ox+gbq8sKKyfM-3PKuuG`YgmJ$(w?sq2rhFvVBBTFPbYOEhu
zB3#1aGcalyLZ_w<u@0uA0?d399aTO1A4<R3oI7_p-UN4l03{M6%#B_c(V#W0uZ@Iq
z+C<85FNvLLSA9vNA%I{D{Ksco3ZmMYY`so(TT)HZnObDmwpI4;a)z%6TvX}3rDJv}
z{^?!1Gm59zHQ<Dj5TRqljbKnBQJLtgd1uHcP~+TayP)6O=094mj11)3pC>qD0;T+Y
zG$eqx{9cH@7-}i{m6qM#oqY73#bV}tbWJwXv_(-?JF?yHq7?4#7(^M!G3UbD_p<RV
zDO0Q^^qK(8x#!&5=GSR%(O6R<>wnavzs@4t5@hcX@n!cUgpmrXqgR6rN(uN?V0i0U
z&FeVkryFcV>Vofy5`1<dBhHZ$XXc&9S_g%HB}RPm4}2uu=I~)h$f{wNm2e^qen_wj
z3dYes!s6fDz$uxlJ|ZLrUiAnWgmJ4B%lzJRcv;-uGb?<$x5b8nVP8`KqetYsRC`jw
z^oSrz?NIO2!AUoM-h|~_*;*n6yILQMBkv7Z_=xdy@r#x?T)Z*a8L~|_3?kiq(F7*|
zy(Y1mc0)DpR=b7-EkBw5!<9B%EGa*19&7LEmv^j+xh5tP{&Yx8t|i?1e0Ng%>07F=
z`e2w!dTB(WGiUR``i>lRq}-l6d6zgh)h4DgT&>j?(PJL;+F7mVQG?{05AqFF3)8ct
zl#%nQkiFr@zVe@$oub#_hm|C`jcE-C$pf_sxgS}S6ZL=jvad@GWWlk2-xndfMgTw7
zrV`)<Yo-8B1ir9+LM(m-RXV{>sXL-at5IwqZP&7ORoIKVxRr<~`<hhn#X0#ky8nZ{
z_Y7-lP1}Z}fQW!dS3sgv6#)xH5eVu=L<9t-3sC_nAw;Bw8bkqUA|N0jM5Rd=A%vPp
z2NCJLgoNG`YJd>m<(}vL-tC^UXZGx8j+y!V5f5FidaZTe_j#4`yjHisEf99v;nJO?
zA_6zk$OIp_U8flLJdLIhd<0u$tQd3SOO|<GQ`Du=T~+7hxAUQFz63&H%zB!quLRYt
zJ<l)TUeCfC{SMG$7UyTPJWLKQdo$8W=TBdAjJ~0o5op1C|JKOy!A&n`7yMdfgxpR7
z>Yn-KG5#RO3JsE0Hk!k>@MCk@rY#$mi*xz20e<2+<kHSlA#^Il8r(N2>%S2vEK0dN
zXzzPsMJ%wjQT$C&sJMIP5o>@4<elZzh4FPrFMU^^mHp2A#Ji6M@qP@Qxw7l)89>RR
zAby)p7_(4O>dNe2wy$XIF4#1*d}eUGLO)EhUBh~S*4vwp(&os+BV_w?kVXd4B*h?U
zh3SpuF+*&yuNf>*`8sGiN*MPHH30G;%=4$*o<$d7gBI_ECrc|0EYA{ye!A>P*^L@~
z_?iZ*Mq%VH%$p^amq*Aw;q=HEY??lu3|qynAKgiE#Boqnr=S$3{C?Wayc3Jq=icjX
zH8H`Jha*1M6!Qq*>gl?d-8&WX!8F;;&eTLf`=o?OzVi4h%d|vYZC5WDTu|pNpkyqA
zbo8Gn7yDLf#*ChWEDMlZcb))POH)QFEe~b_d@;R&sCi(0+R}6~mj?0ZssCI5c!S39
zo1C=E-=dEZJ3QThrrCwBph$2nWPN|hS5UjQ8vpd$5D<oA3zN|aKn%m8EfByq`dvT&
zUyZ4pax^hl`<JXEePN_>i0O*%L2ug<@>0Zo^0UZMFgLaJ7Pj+Fw+#5FIPH*QM-xm^
z!cLc0_{B;yzB+b<&ACPRa$%tlo`_fQz}KPr6p2%{W~FO~kLix?=--ZopT0Fv&b^Vq
z@Nw(tMGFHkjeF1OW3rT^3ChdaZ~UCrwlC5R)+K8Q!gd6%3MqKrOyrPRooF*-Ph3d0
zN?NVH+|7%}^qX<#IolTQ>l>Z;yjL<c&x4qG&FF=z3Jr4L{!4q`FEQ?j3y>Cthyzsl
zZ7@xhde(z4!^TDdo3G)Yket;lTH7~X>&K*P>0fhh*n)5yK2<w-)f<HZayXLuU6@1_
zg;s+|_!Fr^{-nvPyV%9DKOMAk0}q2`){nvpByGi3d2?=AvNova6H}C5>4@M?WG9R0
z=&!7n(GC_FMcbo_?#GDvvY}Y02q@K0+2pQs=1j_EGcD(JTehR)DC})B_#{bLR=W35
z$Sh9CTeFB7ZkhSPt}O6z){J8K-OTBM_dyGWHyKTMZ`X)Lp0XW(rq(37Tj=55Bev0~
z-Xy$=ijJM6AT)n%>#z@x-qnl$#Hjf(V*as_^V=iF=m{Xqfe-f9+t!>bi%lf8a#KaM
zg?T!q>*&EYmSRQ^bBDL(95!O3^TziFu@pTzBK<1YtCGi6cWa=|+3Q1umoGb{ZL_?V
zP`H=`A0q?_NXLdsVA5txC7;R|>Gj-Ibjr(=uM49LeVMrC(W;nuYV*n=<G0~{9DB%!
z8ZfwqSRSgq@5|UVN*2<0nNZ7ojm$Lp*gbRT<+b<aJ&v+rMk{jV@sfU5wF(}zrxY4Z
z8}PWmdRL-%9JW)7dC+G<7N(|9X6rsO+vvS5KvAA%83h}>NIyh7)l(rw^vJ0Y75Iqb
zDeDkTK+3O3*PP^17vy^;owI~}yD0Xktdf<8engW62o#nkNIWsQG}Bu@(WL833dv5+
zF6?Z7W9O!}Li(i|CtLW)!Mcoi2}0!cBlXvOOw=W!g9^R8MsO>-tMufQNjJA;nn()v
z6CG!y0mM`p=m%hy7pWYnGpr;;!1A=OP|t#VXp8J6|9(=%S!2g%7d`AwYP49paWU;o
z#a=4~O|{`9y}|7?K}(`@-b_MoZ9c0xGQ{U*-B*x9$Cm=9cc0gF*(+B11}_&Ik6Ass
zq`s%CUvPd?D<bV(_CvS~kXuA78_kTDy^fxwp`a_Yt+S)D+}@`<m*h$oA;kS%^A%!c
zfjM*m@_ISBXQNi&x?$9&C}Q1e&xIsa=QCPj`n=(W=gl8zf?xpmR5oNy_JAEfPJ&E3
zby4w2QJ;&K8J=<0LLu%-7pK+sb+_Id=2v>$cMWi_$2(0rgjW!T58$2K^{ktkI(5yt
zp}u8!X>5g5=6UJ0)$WW9JM0(+Cssg3W!j+x<qlsPs~Lz3uDCN08`~S_D@L~d1XbhO
zn^ICzL}5x@f-}8WTNXSoto=nIG&M_Vq3l@pB#<sFN5yTTt~Fir4=r0jw9>wUGTeZL
zQS<Kam`Z>tq}|*n`$`$+Lo9$r<HAoq|Ja_ra7tVF%DUc0l2<pHdNm*5uo`!xmKFfq
z0?uWMm+mmJf)~^W^jCt$6M(q%lMiPb-&cqh?DWwQW;aum5~>b%<?i#;2HX}F-mjn?
zR)PGT9iUcwY!y|lj`|9^RG<A7gn~n;J798*<PO6UWSRdfi2r6EHAJvE6?;k-Av$G~
zz|W@E1eOwhz%td5XKY`fT@opzeO&IvQwUK64=55_0t^*=F`DKZ@#`D>B=vFzD0Td;
z-Bfh-vF)pa<=ailiCV_V9;pLn_9tR3{aDU{R<q@H1^wpDJL;(&M3EzbI7wSU)>jZ=
zP?rwTpotj)-13w5_>Fa-iaDZZg6GGXe%R^9SFrPAwEUP0sDJS*Fu{p;$!lBIl;u8A
zJ=7Ho%htSeCCW_gP5eH8r}JYHWwXb2$MEQ>@xA%kU9y7A_Mp|6fDOOrci{H-EH(K!
zXPSKGTOb5`9s&qJvb8bI{u9dOe@XHDT#F@L%@e)j*$i+36DYb*4RO+lgVg(QUYDAb
zs_a8+C78pU`4bhL^k&Rmf3=!VA`Zn$<Jsd`MK`xC3V_h8iY~p<8dzhrkr}nJywu0)
zl!!n$P-owu*WoB{$%k?B%DHXaWRtNI_pq7KDU+)pQB+d$h{W!4E#{0XBULL<-=Kg9
zd~rZ7Y6;%!Y(K*8*lLJLFu$p9n>*h<-Weil)_mlN|Dn3<^P7Q7vG<(&B3`VU9<CXN
z+Hp03Uo1#Pp%uYp(c7OgR(6-jcSVf>6dAqFyp&`ZNCv|KDeO7;$gl8hHXa=|;q*Sw
zGrUxOi2CpX($U2V69WL%ZLzQ;H@8cb9Co3OTI4J=U2sE1PU#_%6J|1CJ4v)J7wUnw
z0>RS0VguA?&Dve*?Q)^hMRxF`A4)xBj>NOx)`@(&f7NJtud5G!dZFus0U?~QtnUj1
zs{6*AzJk)F4GSy05PbO*`)-Lg*IVr_K4md$-FeZ<Em_0!23YJf^-FA5_e>pPEjgwl
zP>8aVWu{%HNY=^v$0f6rbdI&H4*LRCMde{dvGxv&_Y^tr?rjS6da`SSLvGN!{HOTM
zEzHh`5KO57KL5fntu-rv>B_P`UI7=eMdP?ljsFS~Bh;bRS;_&=Ue?rCTu&nAj1X}=
zTt{iN@ok~6q{32=>dABUF4;Kd9VgE=%9W-0coWMg^QhX1d*Mm%KV+UWU`*aMx^GHO
zx%CEl4Oeg|DT-VtlEqOtTp$%A?w<ALgj=byO;mD-dL@Haac70W>!!U4$44fwj%$K$
z+#38`<=2-w{w&p%xl^vEBp$BMx&OVP{5}Oc_x5N%eyXvX-VC9uYd~>esp9VIMz}oF
zGv+(Xx_6bYWgU6ByH(<$=7Tdj9<Mp{ewhyWH~IU2^pM}VZ1BEaZacI^A1YBVzguIu
zd1rFp9%f0gkTC$4DpH*X4NCQ<vt2;66TNllP056Y4)D(5!@k}~4N`q6Nn5fe!kwZ>
z?1|2s<1=)Zdef3tTJ}uqY1B_ok9NhxN4$LP{kr%fLqZR)jqYgl73Yk7XrY^`K+ylm
z7_oPUDGx7|mE1~_teInZ(l#5l`TCCB^P%%#?T;j!E)_=7LLa0lk&E%3k?I~<Gltks
zG0@W<&AUc-UPoP;cy>m!!P!EEJhh%sN?~q52vfOhIcJLbOl6a|d_JF;5esJZC0{jZ
zv@D$5R3n-u)^^909<NL!yd8=OO5^?V8mr})Z@K!I&d=-*I1K<5n#uzBliV6K84f03
z4g=;fG8Z7liI>4~(0fgTYT;6K0?CE$DM+YJL3hslyd?b0<aA?n#Usl|eTU5_B7S50
z_G{{&%ZF7^DWh-*6fW%J5L$#YKRsb~Aqb;Y*s-UC^JT1#@5}=Y&t`LeBf2L2U9Yw`
zeiYCNP;Kc|6{GUk3P`+38-h8DSFTmWcE}liVBR-->-ja#SHh#+W0asO8}tHD^Yu{^
zkjkXS8B>==&6#VXr;WB|t566z!wvRzpb~Aw1k#;a_~i7Jmgz6-?nN#oTYX#3zx9wf
z%eQ{2uYe-ns!{wF8=k&f3n+MP2O8Tvu0pLILeUt3%LeDaoPN!b0U<s>%F-VLrt*vW
zlWyoJ8M?Cva4mL;0|lk9Nu^^9Gp*Fy!mPj|0xTq>VtAT1m|UU^_)Qo8@7-@c3jzJO
zs~@)d@m>ADiRQ-`{5G%s*Ub!F>rh^0(RU*k?GV|@aomjsSJ2ik9PgiF5Pxu9Wp3ir
z02xYqX3DXH8U$9My4M_;K#oPtG-$`PbLKgqF=OiUyOT&4zxNk<%v;Gl>7f8jAO8yS
z03ttU)%byw=<mws{IxvLGCCBptScXp3fWRpqJ~!cGEpCmF72BhV2TR^s*96jtl7h4
zA%Tu;;X+qAwK(5_%=Vr2+vBxw+nv-(i*B);;h;ZzhC{q&gY2f-&=2G@BE0}*vu&Go
zog~EFb+l~riub!1#f7H9q|$nwvt7NPAqVi5vImfOJ&HnZ#_YiTm^^z6k%VPd>SX+F
zyH53qX10JETn*c=GMNI2CB4(=k3RgmJrJxa;V#T;VpCAHeP0t+^}f?{r&k=Gl1Q>4
zBh2KT5$rK4$|n<vP)o{PL-Ux-#G?DRyk74y9~?JH<D8nZ6?uM?jfKgW<-^039T_}(
z+?U~N<yyKk{rM~O$jxn!ZE;v^HYrhCWKN^F6M@MUS+ACiH8?#uaNN*n_Zf!BFv}w=
zQBtWC$tn(k5$eEK&@D1;OZLe`dG}a_kF!Qs{_6yob=B!1<2zN#8P9K*Fh67Rjn@ER
z0(*;qrf_hYrl8-aW;;$s;U_{aK$x?dpSZLInn6>U!2o<Q%Cf8Mr2zn@IKQDZDJH=T
zXjU=V`UT8JJ55=t!5>1FLD_CoANp|ACiJdX#F-W(<>k5qy|BVA3j{<LcUa$c-&6Sk
zmq>2O_xHYQ<La7f=V?8}hs6qbx9%io79+)N)32Y|?N)ZGadhlRWD5-SKbH@p4*`I2
zpP9N0A*P7NisDB2Am)5f+~c!SvvAoOrNi_Xgu63Q%3Q|Il1n9KIJaDTm#DZ<y*Vq#
zx=GL*`3Y4P>KM79!9t0g5#T^OE}i~{gkaAIo|mZTl<wgW@BjEH-ux+=vvFHZ3Ue@U
zjAy~!ddHS9jvhKJR3^KEootnTTbZa%4IhPaduviM9<3snv@+*ZxfP&}$fE6^{B?_7
z-It6~AD#f!Ump!mm090}bwU8)hm9SOODl~b0suH^?OG{(Ee%<!+HWiGw~~4v$t>n9
z>iQw1Y2seKW!Cv&lAUnw>j}h(Wy^iagdzuzQ`x7V{8aZMsETo`m3Guu7(hJx0Z+>h
zaQ7&3j(Pa}Io)X8cPfYDm*Ko?p-9=9tkfjt>mR4+%7Cp%upp`tiG!VH+Md#rqbjVM
zI&E}Xx*nZrAL}M{n08&KH(WGk{DvBPu-^WIPOto=FQ>Kcd5e2q99^{mJ>Uvv&Ippe
zl3|%{X8++@&#6s`;}gkZ1`w|YF!#n2d-jbct4-s-f4r2ms%e0Akt@4-c6>l;PL_I!
z2Jp7j4xa^7#=qT<Er7~pYY3~mIw9o!5W0jPU<cfXUu}%|v!fXPO-}Z~HGl84p1tYr
z^w5mAO`eUj#RY!N>vmr_gu?uMU*BN<>|)ddB?nNw=zdBz^j!^qLb!XOLHKEgbkwL2
zilz<nPE_{+fd6+J!QUU##@tl&<3vAf^y54Ff8*wVafo0rqxemsBJdW#CImuPq0sC9
zcg6Xi1y47qT8X!)y=8xBM%_t=!(v*-HBt=EFEaSEgG{~&0d{>0!e0M|r>6hEI@*6Q
zaI~BC`)ZXBi>N@19CVXea=R5s?<IgLe@8a+TN%+``!T=x)-gBABgY4*yJ>IlZCP6l
z`bF@?o1iA#vaJLZ;8uMFIerCg8<w^*QL~MxU?R{@gDOY8P@615InuLsVVGyoZly9X
zE6)Al@J|+-qDc6qO}bh9r{|XUcpu%K^JadZyuIz*)RkAXgS+DUrQq=9iqg`RnT2PV
z0sE0t<jKdRP+9g_6!${W5-0SI=?uSnnYWSqp_z&5_8^CChD&&PAf8w=GnFxwai1=z
znm~swc5RlAp&rwmGQD*v*a6Yw3hz!f1gr^ycaOk&W3L?vH#W8B9Ij%2oc1vW5WNz;
zS#th9W7a%!ka)OdDikM*EJ0e#Zq&un<;X(=s(G<z<eD8b)6%+P8<-uef0{e4voD2N
zOcJb%x<f#<HDm(0%yKCG?5%mCK)~=?Zhn|do@0)cu_N(GfB9g!NLfzE(NCu}3_+vM
zKdu77%E)X$NZ}m5zQ<>g-yI(VEYNu8iHIizCYM2ISb};5ngdn1AeUrQx6vxy5Hunv
zo}hi{i@DOBN%@eQ<5@5JwcyL|_u7)npnH}1oieD2#J&&Y!2GwgpE@T?AP1Dy-P5!Y
z!s&rMu09(%{ue0X%@@Tk3hPQ6y=4-?y`fj(fUhOuCXftv!_)4U=9<+D3$=OcSraRD
zNj}9%B&m5;bCJaj-H^pG73C7sQw!)jFjWiND{(P+$V^_QOqJy43CV8k1|ZqTI=!W$
zo2bk&ij>2%)0A|97Cs!92p-$_NKxIw4wo9@=9=vMu!-=^2V#%(!i&CyJMekzKJm1V
zU&r|oQ8vg|mfRZPEk*J|?VYnRC)+%#IO_wu({qM3coJKJH*4l$biQEPD&jE9mKoWY
zj{)k-5%_7#uCJh_;ry?lLn+&Pw&{UDw0{Rcuat!Drs9EiEDBxGJ0_bHmoyj^C*NCV
zwWBhu>WvTe$z4{>CFj>EdB-hC1iUk!RSHZZUYOg=%@vCj-mK72nV8Mq)%rsGWDWJk
z@pHVnM8acdJNl6}Wpy|)v;|U&#RD~Jy!7|Q4_RY6GU%QCm5q(X_fIK36Fu2jGwFbS
zhyR=RmL%N_m)}k_^)xYS%kwbVaO-WkSPu~6cXauvkTG)@aifO7K$90{p|DSd78%3k
zE~GeHDhnX`iGw_GH8++MhM(658iYX-YO{E6zIw2Nl$oaX0Qhe2vmD=3is7;Akb^)e
zv-Ws6_rj-bRFT`}rCpXUwKEb+a47favOjz2nW-Y$rf_NHyDkH-PNjX0z2@+_iM#N`
z)u!>qeV?B(aQf@{P1~)UzqKCKV}vK6hZpB&X}QrM6BDMgRCd$?!G8t+xlFd|U`kJs
zwmpimNg_AX`wE2}m+$JFAs3+#ey)omH27lj%cVPaGBzB(<m`b;J_8y|x!3ygQ|)_9
zWgok;bGMKs1FvLQTi)wAYr)f%XVsP4<riP-^^h2Mq4nkK<&V9<OYKu6te&1lZ7lYF
z1%<WmJoW)(5`_V0!*>BiZ8?TBu%4hw<j&*{$iEsIN}K~^GgnX_S2Y?E8vG`4WmL3c
zRqL)<Vb%_riVJUQc=VC?d`_?~N$HhFZJo;n&(D9pb1w&7Zl`z$oOHVhT(g6yg>qED
znR3I2w)j%AsP^YpeMCRTjqqXE`7*eNb=h+^&ghnZlA656C0BdxvO<q2aK(ASrLBW&
zknIiFS5Up}uf|q_8|)L|ab+b$R26Fh_19YMos5W~X^|_C_s43><b3zy;)p8f2#5T9
zW%*P)eNDJ?hC=iKXFxC3w|zH3?x7}+?dXT|16S?iNQ!|n`1d%WA*xX|))ZW$bzc6W
z(`#yIRWI&{u~=m1d{mNzU_g?{M|AUT-_dKY>%e3mjbGUd{GBzs=3t(_EYe;JuKl!J
z?UCC+pQT|yna9_{eGG4-Eo#_CKQP^s&{8*1*XR{2!ao4B>riKC6}`;gKj4udcD3Jt
zb6n=#yMt1@Q#!H*zaNN@?}tKG^rZXQkPiig>XzS6mm2Hr8#`=3rf(Y<+GFzKf^vq6
z-))lnet)&!-a$S*AJurX2E9fH>R+`~zJKH%9n9MiiL}=Y`Li;nYkv2i4IcAtIp0{T
z@?h_jDC0&I=&%oO_}?^I+n$ymd;DTZ{;<dQU)T?O{1Nt0J<PX!JjJbd?NE#{w_t+<
zub;~Pqs+{GkJb2W|8=F{f3Ujvt_cVqYCIVRzKU9Pab5B8IZCmiYfcud0%da=8c@YH
z=0e{u_?1)N8f<pwUH%9m2coK9Q40W&1)c|B_>&-G;18@Iy}fWjcU_xqwB+s+H=k_S
znuv%id}Sn^PH*!>eMFyK*wX%zxsx4^wu!uH6Z2|H6?cc04L01#Q{{(^94ny6WRR14
z^(ef-dpeEEF__pz!eQmoc-B0Xr};J<QIiMD>@oXfZbaFC1<7yamiH61y}hrv#t|i^
zjdD`X6)h~IQc6d>uTtYk<$ToqyqRQ~xSDmK3kVSl1Mk<gV~m_7=8ToFyLHrel8ho>
z>xz#ZO*uc*&!99{SN0{EBGusO81oJiC{9s~kgJp!Gg*Ect9@Qj4I4MkC8aT0Ui&`N
z)tdd-rY>kjpm*UBxfO6BI^}nuL@TBA^0cp@hQLfO1YEv5<NQunPFi%bul@=6&GpYz
z_dMe{KF^36@vf6=v4&j>i*JjgE|9X`rMp^-T;k=tGCwNv%+KgL(o0_FE66Brl{A8L
zTQ$#l2x#Y+XeSsyolThQw#$g6`{-oNk(i&!<mV_KhL4%0MsYlakVBhD?@l@l%r19s
zi)hue<+AwcM3K%k9k^Yq3av<=s-cY5Q57k-y{{@b5XjE~%aLQz=ml|K0H|6KCVV&C
z`+b`x95#zLz0TJ;3D_*ekmuAp<-!$psrO<9+#)2C(g#fI;6jG_aBIT86ORPpk@dG2
zg6<A=!_C%9sOI^|w+*#r=xQ%>jaox7L2Zl0o*uJuZ<pmsE18mYk${@gEKMKth5HK=
zVAVY7!Sc%0p0rvMEldHYVXylLG&L4yX6Cp;(x}Dr_F?OM+^<tya9}LeoxR+c`cUht
zScDO~&Eb!*;#dVVS4H@457|TPhrUcU^3qHwf(@~%GL(8ObI9V!RiJ@Ep<2glGNDhJ
z?d(M_;a6A|x%9Kgj|`U|)zHs<6?gjtW$W*JXTuE(xub%uSRnqtUjQHoTy}S#6fgIA
zi!AogZ;IJWQhrxl&;*q}9iFmZNF|~Enl6V7IsvhyxhXX*UTXp<O2tv=cqtF}3$5%l
zLyBP<#Zxwr%YI3<^vxXqR&4$Q={H=rZT-cgnw$#l<Gjj|*msOOu?`0w!d_8jDXk=M
zxV&!AR(eVr?0_BanAaOG;aul8b(SDw>*vz`mmeU^$B^BJ#IL|{ckcus-+&i-N2gtb
z*KBDl&cfMF`^;LrbN41+eO@T$KpisBD&42E>wd#c<1OfdD;a;80_Yh97cB+w4WOQ`
zfz&4pU~t&siN<q{QPyfpulA>|@`7g<;mwG<QUe5){FRQKq=6xQ(#f=uzNY}3<e(~M
z#6<V?bOJ3oERCtM<R$ZBN^D_{y+BW-0ybMFh%ab=?!L_%ds`~_olbI@zOgwozF9sI
z0SmaWC%JI@GBVtyIrs9_-g{-8dEQ`R;<nGVIc+m!Zd#W-OKwJ&u!z=)oV!-b{ryJG
zi;CP0FapM?8jzDk>9k?;AFm7Gg3`z4P2oE+sp~x@f*LX0M{j)`n_vj(Lkz+`hHdO8
z>`zK2wYY4)%sCX>?GAs0v#=B-?Df^FEE(@<4O-E{uiKe!Y{Js$Fvs4?1y~%XwY1&S
zh&3^w?5Fq3pv4ZyGu5NWxsMO1gz6q)1++P?W0Kqzn=IDlaB^UBXjUMue-fK;?p_Fg
zt?gkaf%hh*=LfoHAxuEmC2eOCIM8+2ZLGGHSZOb%Wv;)z_h?D%AqRNgrN;LfLRQ(=
zXZO7w-<5n(a;lPly_jl3%-;`iitNBFaF<-x1y=W+IBOV;J^xi+-iOU=kKBWPy3^Nq
zBI6w|)@Iusd0Dp{SB4?WcVoI&L*<9sDvoXJfA=m|bXVM+r&%mbyVRzczD*t>cEkz5
z^MoPFT9+hu?tOwI5a7A3(zRjU<sh1}U=5JYIz9`favlbg$p{Xhd?O<hHTsh)WOfm7
z57d_Vx9Vb0Z9nN|Pv^gY5JyLVOvxKveV~-$=`<j?_!x<z2dHR$1??71*)eZHO$%S7
z7g$4AIk{)Iq$W71NJU_&n7-MCCyOZk+0jPM&@JZYs2mfc9mb!wa3PzQ>B8(l9A=AW
zO_sLzS;Y58cMtW_AB}_O_zZw*2u{`fx40X>x`FSHUUQcRk_sNozx)2{d-3DPiLF5R
z47DSkjI;axYI&hV!h*I9CA)q=2aw6}^tv$U7P~6H){*0o&HCncRsr)y{IsW`{D<oe
z9NOL4>hE&<F+xi+wYP3%8tUj;`3S08`efHaR#}0RfwkmUkZAjQ{sF-HveyAEsTXKk
zgvFu7Z@$Ao#bw|%R05fXJ^H9+6|g<fzGL@lEA5abP|moz3vf{W_Gx~0Kz^L$JJbHK
z$)A62|39$FWv6Ud@qE_S{a5C=Sq;?Lf(8qwQn_7tb!YUT>Bi13K;?I74G2q%bge6;
z7rMGVZ%XgKq`I<L9)DZX&A^u5jk#ni<Y%D>rgy0BT-Eq*{LZ$&7K#3??Tq}88q<Wo
z*;?Zd=<&Au7w6zOPAvphwH4znU`k0z$xob{;rK<-!6wd6*?m3mE9fQjP6cW~X;#^q
z+^b1(iy*gdpQ!KEij{(^wweO<h+##}rs5yVq8VbV&8;IO@1%R*0xAwsPf`6Sa`-gX
zw4skiDDp|VxV$kx_$qnZY&i`}#y7B<nO_?LGwSlVZ;1);yjDX7igD(I#*wrWUtq6a
zr7|e{)<usUt2t3~jC7Qx2CX56jUI*;!o4FJ3cU}TO!j3cD^KazpP+<#9h>i%?wt$)
z^HzX~t(?fHjoR4)s$%tn86}|Z+or>1HQ}=S=L;#}A_Re`OBc@agI$s}>zp>gd*Ln~
zYl$tE7J&L{-(?#)rGS_?w0vKASd$|cUnEGJfp2KcvHHxAO=VCO`rtuLpcJ6Spu4EQ
zXj?@IuCREB@r?8broc;<@m40F))S+LKc9XaVo#MF;~9u8P)H$<+Hd)03LR)o9afvM
z4+XJ0iY}*VayRQ09oVcoKOAi_EL5Mg*XtbVja5t~RYhfFYsK4aCTj&7VI~_Z;@e@Y
zIlQ!}4YL{4fd(PGP)P_*8dpS}DRw|^uk50nkD^+KnDd+IbkgFi$tWYFzq@Pqpm|a#
z*{eV#mnutCwm8r$Tjb-e6S!qdyG+C%U&`@sgf9vkT3J1~#;2b}NuIs<<*cEsvMlBC
zEqnz~e@OJHoQ!w3fmtg&>*>(2Hyn2R*gWz1gXpo=6JtvUuIxWioH&XPGgU}&b-ha-
zxgdwW@xDtvw-O!4rfX|l8<KrwQ)jHC;N-BDRn&v6<rv9p#$|3r@THoRP|uN#4s+@I
z;o^_%EjrF!d({IQMK3qmCV0YoHjLI9dKKrA<vUzhp`28S8H9WvoJ(4?x<Fk%@BTo}
z#Qa?@5%NC7(Xs<qw6!FwHkEBMkcOQ)7=#kxdRM+HOHr<zw(CQYF7aTuO99>uVs8UZ
zf)0N)ZSBDAqnWhR?%I&9dUJU8_F$Xdi9J+*85Vlv`ECXqM~ytEkuoh5GtDL-Cx#A|
zZ%M1n?S4`0Y*rAm$TDR&uIY3Hf$L}57m4v`Y(d;S6qt7Kq{4}WVmy!HMAQsy5xW^F
zB<<Pxj&jg67H&3-uKA2M8`N4@K1z<LxgZ;;G=hbMx*T8g5g4ofV9z-)+q0)*SzOWS
z;?d<HZ1kD%3FFhm7gLNO)#X$hI1$fG#R&$(?_&Ti-qmQW-tkBAreB<$f?s6Cf&3J+
z{4XE9%~g`?ffMvLG?Bk}ms~9^((@IhMQ?(c2^!wI8kTwdPK(rqMib^1(eqUeYNru3
zbROyyaM}Ut!@;VJhbAVo4`z#_bU_~QiQv$zH^UJp3r^nhKcjiD>GO^!x}M@P=%~0s
zD$-V6kSKO*A6~MOeb=S8uEnA!mS)o=nH`R`Q%LM~HHBwBcHLHjM5YW4j?Bfzx({@x
zKkJ0ygnc0H1FTF5AZx_=JsnrYIiwnt)vd`IHJ(orY=i2t4sip`NM8kdZPoUj<;^Pe
zj>Z!6mj?jM{JcIN5k!|Ud-){^@aL}~7QD<I4YAKAuOyRpPbd~)_q!!6#OSLOv{>2!
z&FIhAq5FM}{H74|3$7Ix`KX%154<a~eOF%WbdXPZC~uvcbktn2Jc3`C^(kA#@+>zt
z5UT(`W)O1$uiYz6cU!y%v`>NcD(j`q+}t$9CS|qC@pTx_aG9G_TRyKVn;bp0e5*0%
z3@<swDgS7Rzm~~(puzEz4tiZ{j6$5yizZYs>9gCZ!C5e?G^QRCYnOAsYeD$(ed~LR
zS=lF2Ln~rgPc-h5e5QlEvXkg5S;*zhpXuDn`Hqg6COX^ib4G?4D3p1%FVeP6H|7ea
z#?B<<oi@HL{nOUSnfWr)!I2&>#KhzaBOpFdBTv%>K(#E8-NNzHfKrk`7!?ihCoO=a
zfwwHRH*Kc?_)yxeL&)d8g5oDomXpv`b_i`ZblfA2cLqgnL2&~ZTkBhjDc~ze076Wl
z19-A6E)&S`vC`#G4%YZpZVuXh$RxOLGj4|gxELsIQ`k=36G}S^1(>^ATtG5_h^mjE
z-p2<6?@1iUUQ`2#9`N8H#113iATUI12Cb`F130u30RG~1jRZXlpdOj6mja<bsnn$d
zz4TJ(hw@3N9Tp%mPR5UEtk;_XDE$<$_Za=|PDDJaO%4b#u%IIUu19YF<rBaB!yedT
zdmDg(O>RYBLHVp`Kp#c$`)9~E-HG$%N0t-Ob<HZ!THN6-mr9k~j7QeC0+%|@SRK37
z%pG9)ahe@@e%<C#Zmdy;PKCHc(KStc*;M88iI=nIllH|N(>bkcNb1<0_vr;v2h}E)
zUqQ<95E`#Jic$tNCIdoa$pUElzC1j27zF{hNdU2VfP!irK_dh2k8I!yWMlxp>dA)&
zI?^6RJy8ZE8C<@B{J(uFkVCsMfgB4&+gjOy*UMxCr8!1WwN9d@0)danKRchU_%ptv
ziN``#YJd-|$^ght38v7wn!%ey;6oYB0v!49pF_bvZs&(}{wuy6gYXan^u5@b!5V+1
zu!?y4rFXv?L>bPcmv7>yFh#gmjiEx%TQbb+_cA`eyjPRaYJEx$G#C*qER4Y>lMp<1
z&#M>}<G7k7-U#G4nKX)2L<T!B`CsBPRK8NXr;3Cf4hzpH=W1CG=64z&YkWSo$~)D|
z=POi(O&&o64LG#vvQJbLXr>Q+XyP6BH*4$@ROU@q6E;x3`jM|`vdjlkTfMbDEcv0d
z$iR6b)y>_J^}~zi3)PR**gPo&<r{l)|8t3v${F94C+DZCo&=^~<sQ}Nz?wBZH@O8`
zNA&m7?{jRse+3O->ShUtaFTXt4w^fKf7_@kqXu`V8)IhOCebdFQdZ{wW%256RsxoX
zt5NhI#75M47dBx5VgucL30WWj@@XC3r*`6L;zV;@1InIqpze2{y{c>pg)c3qDqkj^
z4I7!d@vwz?Y2@?~m+)?+GN<3w$p8wpc@3J@Ty|rnY&%zquAoTM+P;JYR}H*aSNf9T
z0h3RtA9&I>W8|)oaDUB6&%>2Xm%U)%6QR;|(^&BmyED)O6J;N<&s(0}g9#+uqU<k7
zCCscVu6w#g`^dQgHx)Z{+>V5eQW-J-)R_6I@k-rh;j^>rK~xrsAgL?)Y-f%ut63%3
z<b%T2ZLF|N*XoOoLPt^LctO$AI?voAHqvrQVp2s}e#6Y6xIM_YTOGb);i^6oTEkwo
z0t%(+b<R#F^&;=Z3Z}&ONSt1RDej7a?s_qH-t=h4&B6S%o}~54?ZJztEm8#C2HjR@
zNd`G?JswZtA@J*lot!V4L+!N~8Jg)Z(Ga^or4{J)l;P@!(-L>ZQ{NOkTG3bz>-0T9
z?+}R17&TLYkz9Ksh;Y*pX<wh4*E(0X<NfQ>hjJb^Mg{rc^dl>;6+ACFUUZ!K0{sxG
z*4=1;eW)xp?!NBRHCUzRUaMI3@(#x}4s7m+Xo1O#CvO}d0u8KyJPw6g?7Z~8nPrzS
zQDL;DRJj%r4o{Jld%dSeM&C@oFZ69zq{P|B;pJPMrqb>cH2uz=Ey>BG+|su<Vzd-(
zR{%}&{KQ2}(X#!Nd%kaH@<9aioIp@O8r;UDIHCSir*4;Yp37QE^449Z=)i;{rX!jh
zThld}w;rYygp80!GAXPzTjK>tA`Oy>^vSWRfO~0R;Y!j<S3ga%q$E!oMw%RYc|A>S
zx$lh3#XLwop>QN~jN^d)mS2~HLsHRUuEFaiRvzKFv<!<@ccBj&0qlGmPw$+%rx+3U
z*whpI$XCSMpxs;ZEwy9i3K<uWX=~(p>OFa)a~5nH>&miTDxNUH7tw59byxf268y^p
zCaxkx6Al+!oP<G`U&wn^uUAxhtH@$ud$=DT)$NTo6ExLb;2f2<)a+S1=jxU1r2^K8
zHZFzjoLFv8sM~eFEI=aIlqAHl_B`n3^Ic+L6fids_|@f@ChT|hv8-Aiwohomj3n6z
z8L5K1Qg87`KP8o&%Fz;Q<K;S|Y$cZCFX!|5br(gk3i=enwVGs0jiYlWN!|>tO2`Os
zdAFL_G#NN)R{K=`^L*QtyK3;U+7se~rO(!n7%V*ngJ>5?xxQOl9V^<;j;|8fO|n`}
zK6*SekUgTKBdRRtu-g;Mcx*Bu1TL{!wq1&z;whK+-C{iaHg#hGDEmges`T(hq_xqN
zraf)^=P`#-wNiP|+xGmX4<#NGlP+Bdp7Phww2aUd_GvopaJ0{gjc@ZJuP(85fioQ$
zUDc91cjk;r4?b(g6P_NIZm#R>#Gcz#AmtO!J2@m8em%%`pYG9zOm}8bj5M3kq<b~B
zTbymKaJN=N3J<<NA$+x`xX{Gjy?R17`Tf=Q7H|xo6E{QL?N$4UJcu>nU{l7lX=O_{
zfUbjAcg!rx6LmY4wX$saum|jo?u|`U6x}={=Owy%O$qy%L;X&^57b~oMfEGlB5v+(
zQB22-<ch@xjiFzDX@LKu#o-Tai;k|ri~MH(q$!;s?M|7cjFPK|_`KB_b%-fv*r#uu
zAujK!ukpUPt?sQtZpn*q_Q;QyUNhU3ZPsFG>5L4%=SEtzWOI}r-3(f??t9EUIXV_Y
zbt3cYlXbA8f=kpmv&$)VTjQPqm6dA??TU&U3M&)hWp?5(7<dxLBfAO@SMexxPtjP&
zx+J6O3$pB5#CSls#y*rN1!kMIG%PK>3Kz$CSX_9i@RKQkN3X`YNRK-Xpet^_QjyI^
z(Y0tPc>_DmI$CQ0vcxUOEG)0BEqIObKk-(~(!7hxQqtU~F5XhB;M@wc1xnzvVWCgk
zfnK-chYlg9GJs~<pCQ~R>q<nOg(AggEu$afBUul&fnSvMHp`v4F>uDL0dY)0L4R+)
zUdJi7Cb8>hZk67)oGC{dj_f7fUI@Pw7NrAfxY<OrlX^(VU)zCZs=#+p(74avRJQK^
zx-isuV1pK*eEus4y7U#4TG-QK2_r1?kf(M~czRST0Ez#FLUa26#U-_cWUzYYTEs#t
zw$F5A)Avei(WMXY*7#ReLN*M`sokp5NPex7ZpVn4g-cV_3EJ}sZYiW@pA%g(>xB)H
zamTFAu*O~`oahjf7Ff7vw(n}?cNMx7g!+WTJ7LL^UCO+vK%C=SLm6CYRHs%4-|R-s
z7N+X?rXu%!iGrVA!9C12^z)|JssgV;1#}TqDf<2ax+rRYB0T}{M7w8JY`3lhRFYIH
zThqJXwW)8UA0T!09c96L^BvL>NWlbb<`a;Fk4hWC-s-OJW*$<_%sO%;6Qo#g1s)aA
z@J!Qh(RVG4y5%8yCS}Ya(I0%CG8O1*wiJ5GUdiaumuQj}O0d5I`yQ!ytD;-HK;XsQ
zMdqh$Z<(-O4~+*;lr;#-0UfnX{U=|D(g=Z>zWlwI<SmRW@gX4nNh_u&MQ@=$E+Zig
z1Mr<iG_l$E9?<L0C6jNmp4U<uQ&T0%Ub?Au<yvwo=W`C;018L#2$(DzT*SfAPm?RB
z_!j1x4Vzzty&1G?ny7QU%XJphMLsr(B<~ViY2Y2T5GNzT;L=6eiQ&gk)lwWGG$YkF
z*4ap<m0Bl$^o{!;O5sW)#uG97Jj%sy%;N4i+<h#25r9&U%kqOUh$nm_;wWCtjQ$x5
zk;wcUz<CW4FEVjl5a=1Ex4Ly)@;_pD&gx}x<BKzamw45kfno)yH9j`mu|BHE{+f!c
zY*kUE?i*#vlo?-<cFW2xWstJ$!_S{whD}H6rquQ`aKlfwkFFhj=%Y4c_N&tP-Y33o
zBkTp8F0D+{aH5^E?rXE$KBb6Yxp1@k_XAolKf~moJOqMBqfNyg#+#SUypA}_{MlP-
zbns;b1G&Xl5-Dj)f%wbK<#SzjeG{?Yd{~TIehzj2(}2y=#DUzh&yl-@eiDnAWMEg{
z%v<jv^d?T16_;^!hJ*rB2Q#^Ct4t8eIE4JLRHHZJX$f<Sy)nye3WS^A4Kn77IpG!(
zbxaTK+0{7M2nqvS>F=3kP%-KlF)=?B0|Z4lkPPOX6I~U&Z`9JY&P#Mzbg*9>)GLVE
zG*y!kX#Erc`)sPld#jib3wI|D32X!^!(pQchFMQ=P57maYRVuLwY_gli)*MME1S^D
z8`>ZgDf#^F2;DSrx&OnZ*bG=voDqVS(7U>W=+v`ofz4!)qJ4|u>+SBU?|+MrWWOGr
z+BdCao}QR)Vm342`bvulla#8(bLz@#rk~V7C;UJlHqeP-jz%`bpHZ29b?#!@)IExS
z#_ReO|G-S{#(&XH^3l0SjZ2~WZM<^kDtuZ!8U~oKJ<vj(2gq?NhPB;gTi@rC{WyNx
zhOhq~L;u*x_J;|6n4tLIzyx+=s>Y?_g!6COU8rf{k5hwT2d(b_lCTdh=Gy;9%Uk_+
z6lV?Xx6VyKIeI9?^q{egt^GZscvTzo>0y<Eo+1q58Jf-Iu;J46F8T$eyG~cmgsqIj
z;@a?u<8LB$4jn#p(U9@L#Y2p@poRIg&){`n#l_eY!XV>Qs)ta0mH9xZXjtLtoV@S`
zjT`m4r@==!++EI2%vYkzD*U6hwU+)~mA$IE*K)Hi$z9OJ*jt<`>s;_|B(CIzztY5M
z+yZt!;%H-iLn-Rxer;syQ`IBL*oIp0L7;1xlSitZDPLxWZGu5KE&}c*0yx!shF)it
zf9}v}I^>~Tbe1pD{1y}7*~)ugZj+>Z3G(qFv3i9I&sBGrX_Fe|+`3|9*DfO-gzvmE
z9jb57XntYB?*(2Ic(5^`l`l)zhO0Pq-XW3l0P-;2$Z{4jFV)^e5l&~`amIY(4;X3s
z!3M1JQqqytOM0D)DO`iY1A^!8#qVvja(xoW=+4YLn#@NH8pBZTKh7plad3@m-QGUA
z)Em>%cVUMs4Brc``(G5P7eDg`qdo+M9k0^vcdi}qzC%qU$p&F9j)pck&qUkeS#y-{
zik-4^(~(_js+19zWf|noEz3UX#R1BeY$<t`otTc6%8+exp?SjQ(>@csQOmc{=usb!
zUl9wOElyJ0gS;7>teJ!9l4jJ0UX%THe7Rm31DG0{gJGHnPdd*4;%6af?g~pE;!>E!
z(N1L_Nx#I81Gk-x?Vk7dYx52^Wu$#-ec&kKeqy=tUf7lPuv+Je3D`TTu2Z@Na&AnF
zej2W_X6IYKEZ=wWXxo>hU6+6e+eu}B!|1Z<zL_+0p&`EVyuq3q%@vChcZZ*}6x3AD
z$6$$c*wmZ(u}hVrqO}R};I%K`*f$+OG2Jh@ZWf#fl4Osov|=@V4<&aB$ZRp%j_fPy
zEy^y02^B)C7f#3~AV`t52O8Ma>$z@D<(VT}!Fk?R)Cf`p+ibE#0p%p_0R0mcnnCTF
z;VW}3DroU0QAQN)#@r8_dUPu*#v6XLEpXVdJ2wjWgb@h*@|idMijHNz%Eezn9qp^q
zy&-afU_LXd=@jXf0$sA?M%(m<;XdhkzrZ_&+``Ly4VCwY_(;9|6(t8}Ku*SxWLat6
z6sDbwU{}FV1Q&9>{?qpL<dM(yZP*~yu6H@_Px@txUClrZhdam0v=k5_!NO;|#Hf;_
z(EO(#ba2YuBhr_vuY9V5=UQ&_cI7g%9pwkri0P#Uj6XC+fi7HqyCAr@und>LtD+yB
zp<82R6`P^UHXYFK2spo{GdZ;D@A;ebm!Bct-{Qyod+q>WIa_RSrY&f0D>Sz9&Y_gz
zPvAnGho?Di?>^sHr9#$#{wlK2w(D0z<$F#+PYchkzs=J5vBwXKeNPqqVUK^*9)FdC
zW7>6t=>(woDSerqYq#3?t#;4vFR6I`J==b(<q2!o9~@Kkm<m^hIFZn-gc-L89(V31
z1D#w4;E~4Te!6O2MlVp+eiRNG-44C6u1i<6SVGX>@0_Rud7x|ZcT<BtLI%L)owD@a
ziLapMp~!{7a{9|xQ|>e7$~~qr^_j3tE#ks{K-m2JQEY`*&)G{+yk}0F$9@Vt77rSo
zzrD)a(XJ-sx)f-pkLi)lqk)Op*>D^Br3WpltGG=*2;nh2SXJF%vvB#{VNi|_D^WFo
z=*8mTf(Z0w={T8gYpMl1nBuQ^s?qvMjM)tqO%~VI;d@uBCMzXE_!3cu;K+ssVn%2q
z3z38UHMTNZFUWzqyqHRu)WJ^=T^n&oeC6Hn*#Z5s+t!;K0k@QYp}L2hV|de8CPupK
z9A{<1(xAg9>G$`9tzL%FQoNj+@+L!eFQMk}Q<*Cf^}78HwP=1!Y4RNXKx2_t=;^!k
z_>{Ds><qO}`nIdvOVaEmk(ytq6IJ(&7(kvj!8k!XoPv@a1zfYlObu8a5|r#N9+odS
zSQXY}?<UNA-6j9h+km&&&l)Tw!fsRQrh`!}oso=wTK5F!`nvn2xlgi=Dw5WGv!FyM
zO!x$l(+HaE7Q=fHSg3B4HP+J^Bl2}s&d&CVB$l)Am4@vm)yt79SfPZucdr%t#*C`#
z5KJWMg&HY0tdyq=BcXzymE-w`z-<Nhr+7EPce-)sNEM+q&_Yx&P)8GDsCudgLS)A!
zRIkG?T?m)e&{ovPmmt#uz&FaTHkKY2>aeuZW3N3u6SAcmVpp^oOYDq$6aK6Ea*baV
zg*&ugWkX<)LG(wSHgvXBAhYJF+D!s)F#-BWXQF}jCUf^gK=s>3Pf^qSfauF7UAjif
zg2hpw#@-?xhz=fyYzqJEgZ(?6A`Qjy6p>fdk1t%U-9B7}Myx*x^1T<OjCwb=b05Xw
zIr9z(A^bbu_<zkC`0*hBg~RF(c&XlsN^{x=6LmT}L_S`Scq4l1_OjY3wpVVATRj+_
zUDtr>b=Dg{Yx$i4ufcR*|FWF@ANl)UneTOHn!jsYX1jascPfqlHFa{K3H(;gYBsM0
zVdw6a!5^y>_}3H!|Jm(nK=$-Mli0BT;h)T$|3Bd8-@Sj>xBL6gO$VR<la-CIUqQyc
zG53BqlKy+p_8+y+aex`W%l_LGlK!&~0Q#Lw`CUxERsuC~|DMJJvH<e@zgI5nFCEA5
z>VFKKJqR$rA=e6iS4)JyHc|a=ZRdISpR0QC1xV9=GqeGK;(Nep#=m>j`A1f$e|Fz}
zw}46IKUMs8O8QTh&RYKpn&arm`K$77zeTG1SN{KhZb-p7_D^Ou@ItKx%^3cn)74*^
z!v2VD|H-W)&Nu&T=M$AbgF>F)yZHPokNW|=|5-Twj~xF4djBgS`w!^-N5Sg9=R7~4
z_a7Wv_$|o(1A6~L^WdL_)ql_NKcM%&5_W&P)(`0YM?mVo_Z&Z<_rDaf|A5~A2z35u
z_x%CAe?aeF)5iY7znTAIquwi14ND=hC}wG5H;2C*{?IJIg$m12mWT7DJt)rPzHNF%
zZvZNCf8dbEF4R->y=HJotg0ZjHG`&1jU@M0C-WrUYF~5>hnRF=#JWz#OE24KCYWit
zJ;`Dp42cn~w4!$tm@70WDH?4q`291T_K^OsAkNj?n|j+STm8*UFWhV;V#HQBA0Kmy
zkj9Qrc<>G*Q`W7>d2vLv;H<Au`w|8!WS=Zl(Y}C@b~Y>TT3lDw;bKf)_h7qfUb-vB
zqbX<T)$vbm<vOOQ!4Sgc+0EH-k6UR6C1Gjqg9Lyy-eff10)sq7?_N8+XZv;$MX{zi
zm#g|Jsl4kA{=FB>-Fx0dj2**zXOxSjst(>i;w{tVEh-tZAW07B4M@2Xxz=dWJs-nE
zeECIHZP%=Gop*e;Z!?6;!IzExdFpEC;`~~E&`-@dNH6vB&W@EYF(AbQDvePsli@Xr
zOJ@W>JDgJP2J`qf;SVTZB%^Dw38<ho%Ul9=y|peb`5@&=X?l}HDQ}608tKzs_z{m_
z!($fe5~s8n%-Ln^q~;e;QgzVQD<es~<%0)@>(9Vky}VNw##gV=eU!G+@C&kJLM@bA
z*^Y={wF7ftd}M2Ihl2o>t0!4w=EXtR+&y9y{OwZw4W7u1YE5iCDzcHEGi?aLaUwNc
zY`t|esjM{vf!VvYVQ>#Gn}en7k6yARvSmj5Re3ay+VCzz>Nxzqg7{QrkTqlYr_#Q>
zw~_jUTHM~^wxZArm1f#@`7)Lz!C$_DB;2lk>Au6NaO&}7)5p0xRMD}eAmBh{O_CHk
zfDg@I6v2apTy5{vt>|*N7kWKm;MiC~kJH2PgNI)<nu<k7Zm>)iMgaY-a<x?_z*>3v
zdNh6PS{kIgiH3$NaXiBpGgl)XV^IfI%^-Vce2-F~k8|L%>7DVp(r?PSTZHZ(8{e(5
zBFYuydiufXxdZ}^nyL7R<^)L6Ql_dQ`|M1i&t$$N-OZw+3<V<nr|X5!O@^2WK4b_B
z|75)Td_LDs2I@XEMEE3yxz37=uUV@spdg+UsYvOXi|T&}>G<TO^)fY<VM>eji9y&C
z>j*|T(Su=*W3O+E>fSb26bJR$m~OW`l-qWF;$$8`b7sG6q`)@KcX{`)?ulBR(=R}y
z`)xB)&u3(7&kRefuW6bP@-6bbOjQs7=d>!nwA7`i#jb<qD?wF}4dZwcZ9u+2&4e~9
z2@IJej`n>ah+bHE<`g2UtW9>;rabZGF7Ni{fRW<OEw%T<LgjF7cVqN+U1$=0e!9gQ
z4%<FP(}zCc1MjsW%a_}`;S}j3h~N+QhB&Qi8Rey-fE9Maz&`qtV-=k)l|||O2zmOR
zrq!FL%tBrt(-DXZBle`joo}?Sp;$+b9BFzdceY`D&@6hVk7q%2;^FzXN+r<T{FU+E
zPeZLfKEv2Tba|^|qQ{&p-s&r8VPO%o6>3Idt_HJF69#J%>eEJ(q-qt-7zMOf*B=Zw
zR}d9yYcDveoI0OQ&8|f=SFq%m&Zvq5k3(EAoZf>i=yp80VDpl@h(~kpeMZX1Ij9J7
zY~6?&n=C>dPSHw5qU%P4r~$d`*N)4nKra^+#`4`U6vX1O3A23IhEHwG2U#(<dX^f%
z{e%X|T8m@oDq*G`SxmUDddKQYj%sd-y{SSmStNs|?SR=E^P-+V=#eqb9TYT?Sw0??
zJBu^IHqH1T`Z9Q1#F9h4f)b22GrU|D*I=`V;G0NpBD7`)Rj)J|kk*iDTdZ#}ba4LO
z$xorrvradku6{HWX(7to@J`;`i->>XLU>%ZkGfGWz)TYhPq!}%GmJIlOMNAEdnabZ
z<iYvyqhij<du9iAFv?&OnA5jL^=t>@1E3&$a5}b^|AF1HpV)eqcnVXmUGc8CLOuB9
z^lHu$zwHX`UZ?NjSw%2Q#d(S|Q6_`fXJrs_!MdktFrM>}Rq$d)2t+KM=jo@$y^_yC
zvUb`QGx8gzo+GVpadZZ_>_h&_y()>i<Oe%~+gk&5x;(yZhRJ)W_voS&@JIYSyW?P;
z!pOnW4tbtu=fuvQ({qQpj&nM3D4$YqV)U-m#lU=zP%jhF+-68;<E{$sG-Tl1oyVHE
z<S&Ay?yn^JUIxJ%(wXI#buim!X%^kM{V)L`0rgeS>&Yse@`(LZja$9<D$CkkO2j^w
zFIvUCm<r@P`2k+POYBwESVTXRWQq(Oa|@S=B%%0`&LqAf`D3YVYd-rsz8IT4GIkuH
zkHFrU62MYwLU&}kZ|cZw=g-9^C&9C0%VrZ+-Fi`T37)IC5GA`5-QGFWA!^}hFF=O`
zO6g&$OytV>_RWGpn317jl#-2LjMa&&jk70Yxsim?IDgYxWGIxp-ygBiv)E?u9;-mG
zeX<&>GOze%zDA^ipVUy@8p?;0p)lbd5xue(_N-L*Y~5DE28gslF!ClHU};jfBSxOB
z-G2|;=~WdTiiIpWy4qKoDoq(1-TyeB2C1MZh7=)!04Z(TSR*$Ml|sRp;>Hi!N|zqw
zWG^?pI$!@lU2Jh-vUeyQNNr(_=!}w>g?30~6)-m-ASGKHBuJesFqLw&)+Fgb*r!dG
z9(&)-mUxB}>+}x3gNe0<+(*UiYZ5Grkr9HAt_=1<h>Yak9o7LO`Yv^V2;Dn{ujP15
zm!!IHgQrpZF`<G!XX`DLsS)fwk;K8{>b0hqT3jTT86-a$iiL-snYc2xmfs=Yc!Q;B
z(Z1t$_$H827JrOQdX<3JH``d~AZ)wD(Cs8Gx#<UV*->qkX{b>i=(irk{zw=-U_lB{
z4tx8p{occns4xiiqOv?7Q??!h=r`!0C~~N37GnMue^EYdcPOBgko)Nv1IKTk2LzBX
zo^aPdS7q&>R3`qPyGS2N76A0nya37UkH_%TgdspZ0FW#E?z!6m+O}jMe%hh~Pu?d>
zm%tu-y<bDC!@&i0(7vtD#H_EnhtJWnIM`Bu#JcedNNAMfQ+Xvn{bCaxkPMl=Ujk?s
zM2_78n)A4#>EUI0HGnOjKSWgkT4;w}|1VVVK7FhBf7pBPpr*U1T@)2XL_p~s1q7u^
z?;!F50%9oAOH^tIAs{VCAOg}m2q*}sRHaD`B_v7}klv(*gbo4;H9&|r@11k!e&^or
zn>lCZ{Q1rB4>KhEva{FPYd`B*&qLKi`I3X@@|Jmj=S;Dq{nzK2z)LT0_%=U1+_pvE
zK@k8EJN2#58?XuwYqcV*+Zb*fWun-+V4SMa6ghZ3Da)+kWw1a)c?DJV7u=zJbOUiN
z(o`L*y;#1$I7JREgR_soC7;5g9d*s~J2~$@{x;il|C(=ypbW#TFZV4!+dfYw<>`9#
zk;98M|5oK=i2@h!OiVagq7tO0g~U&c96}t)ZE952x6Uyg&8=m|Z{7RRsVi@v>#W37
z`U?`Ge<Oyv%c(Ta1PnzTn9fRen8J8MX8?<cuGv;wg?ZK|M2-a9qly)-eVNS51Ak&K
z+>&{>czy3_iH{EZ&Jm^|s7sWj^PKn!z&Be}8nA@;!$EQt%>^$Pi7xe?-G@p#o`00t
zRfcNsyl4@olL;VOBQta36HfL^S0=kz@~Wld#MVL#8}^nx&dMD|$?CWyp<T<#fP^+N
z<NCyYztGTXhYs4iYIUuT*&iH=mevp!Qoh>EE=+q;y{+;(!3xSZz(D}MvhVBx&!w-W
z>d^Y=REPhi@eoW9kR-JcdX9>DkSRTp^(j{tj6d4Kj20$`AP2^;@8lceL2(~Gh`rrN
z^lfdMmsvXQoNe<ZGvLnh+&fO)qrwE%kB?Y95<Y!t$+5pXH}BY3POYSs`fhsba6tx3
z+1~;c98ei2Dt*9oU|7j$g87OQ==6)e_6N+*Gi&Q+0x;&#a`X_u_Ls(O(*z`SI=ukP
zf|V$ZB=+9de<vtKqh7$pKk(8ns2du=)BOqZ{5dhC1Bsk6uefh-Pg*yWtd+fB#^$i{
zE}iC5x(Pgqc&VFbHw7&W*CsN@eH`09+}2&h#=4;#MRz31b;Yu8$*|QJ3dt!cGgfsl
zCK8>_&zLrVH>QqYBr`vmcnvuUJEaTxyhJQ|hvTg8cVP7ldcnd+r9YTU<?QoU`o@zM
z#GXY8WjeIweY3}hSAS{t?q^fxNdicBJTOKr<tS)c!w(#AUKlJ|;JQqzGB~pwM{M0n
zU?uO#_&a>tk@!jtb)^nUm?DtO=SX$Gnjs&Kpzpy7@RQAFYTz)5aZJOZiq_pNE7^e~
z*r(3>Uu?#_JPX6SeKY2ACI~Xyw#3qIC&3zGC%_LPiN3F9*%Q+sQNV2LVKx(&sQ*jj
z_~zzC0`ohLvw`o4U#^rS9Q6jz?szhW$)DJV9a<d27p%|5a>iCJZC5k*FJd`>R97dl
zxWT&MBEXTOJ{CX}+3$5#dbGFt^xZqGI9B2rAygd7)PVzV+08k6Xx+}WZ$vhQ7?zBo
z^n88N9>*}Vvb}q~`RdDZQ1u`%T1_xfi=XUFIwy{O-wv(-;4p%lN&-TJA{Vv^g_o;_
zcDIpQ@W=43wbY_l<$uBk5Zq!7BnhfIU%MiHz^~)j={_l~($KX8hIrV3o(S;xnG6uW
zCnNZ$u~XuT`&;G4`yM6Z<3MY+nf;8UHjCA<0maG~vk+C4gH=z8nK@?Q<*9Ay#;FW7
zFG6C-WKRr9rIk8hSdVDueg2n5o2Qz>B>k6$v>oOTOQ0&v_cF)84RB}LoK7IM?p&gb
z5#g8GWO2!c#U{Ob(DfC3o>a$-d|@|6!RejIO9|sHtYDZDjKPG|QAvH^v6hK#iU6zw
zFuYwMqoAp1Z59!1=5nUnb1~It@e;?cXKSjw4xuOd5|tPM0L50Y^F7)I*;Y-W4)X1>
zkeP>pBwa7<wIxfc3}F{lwTr)ya$}bHgfxL5_)_n4Jz!h`N7&e38ZJc)3*571cn7M$
zx)|!}zby|4_Gbv??b8=TjW3Uj#ji^)lB?<UXp)L!h0p}TWT9UL%+CmeLK+*lmOm=Z
zSx~jIjGMkB7h`VuS|sUvs&A@^N|ndGQ|&fK+P3MJrxYH!YVz$y!F?iiXD;G*|I#G?
zd#?P)^=oL>yZsLFKc_!~|IceCSJ*C4kMRHLb^pJ0_cnXL_22LFU+>HRUZ4NoKmVh=
z`0x1o5Bd5(i3?jTNy2>Ndz_=&-psCL^6Ek#=mlaMT-cDMbZfCU^5?E&=5x|6<sZZi
zsu`i$(q9@qMH=Ae_`8?-;b0=C4T0$sdsX-+o2^q!EXqP4jaDtI<cNb<@aKnW+R7s_
z@2ur#w|k%etOQ;n+l&V=v%m^E^5sXoG^8rlIjm;#Ivdr>fuLQG>^sQ_{G(^C;*gc=
z)jUfy5eC8;q+-J0E<UjNqg@yKK4>N=sSq>nHWr}jzZEOholM*{{G6_8uzM%Ym)HAt
z3UAQU0yb@T`(Ujbq=J745>9ro81wwT9Y+Jx@f9?>?M$)rWvtWY!QiiJg`XvJ3(d(@
zqu-1E<Wl_<8Ke-}jtS}3@s8QpUsPr6u<Hl2h`0r&R>B|3t;nt%zM1PR6)c-+o2CsF
zFXhF*Dtu~;zB5eo#L$($E4U{T;kkR6gzeT+YWPMJe3SA<L*lwr0A3dFAsF%~4Nc_v
zqK5>^KYw2=h_W(4xuF&Sky?t>`Fs4?xxo{n<7#PzrvRTWWJzNC#*k)P*5{1ARVg)*
zaCT?VeLL%5n(m32Pb~m;xCHS9X}>f7UgI!R0|>Gd!hF+year8nRyl+tg}*R3WZujr
z`Gt#)#4RbBWmmh(I|Ef%${^-M9wBvHzm5}|01_stW;<&k0t3E8=N}wl>~%!No<x*3
zUG%OYM@wf;5rezxwO<b(1D&5c`~6r?4B;vFCD!5QOFjiW*G_@UG9DILI5^n#R@hg3
z>tU69yi>gQH8HUQaP4ViA`kIg8K)c{K*h8cgj)z}N0x1Coj{V2=!jSoGFYMCbyTE&
z|IY7i?#wlY+cxhBu5@D8LH;S>r{l~xgC@du;~6ks4HLJ^oqtiU2s0y~*L#9?(}Fb8
z^jG(%Dt)+nHi^zg_K`8!@`o!MBfTP34V2Q8y<+N}SLy&$vx(u>xbO+){)8x6yp+PY
zLE#@yo|SUivPQ;_t}lMjlUFfV6l?+L+Mi}q)tz`BIa-@eNLKha%7%yf`J{b^cUz{c
zz2IDRGgV#wcBvBHQ3)3&hGkQA!L-IYoJ7=?dDh~JYg!3a#l^vdnFVZpFJI&F_7O;G
zRfku}FzwpiYs#_$;Y-kHv@p@&&lC<r@3emEIN8Rqu05mXoyTYT`|Iq&Oz+E@!W$8a
zkCxk$uR2`#$u?>pAL7`vod1_5$Pr@-&hk`7Q}b0VLUVA-m^4q=a>zAIctSjmI;0N>
zc4AbO3djZV9Q|`k?IHWnyZ*G)W?_wqNz#|aCbEl$(1OgcbyrN<$Gj&O;9-u~O=ajA
z>R@94Hk5#7ghik3;`1g8K6M>cLc6v2CcH!qpBv;%4sVozlsPU3a!KE>jPMDJY;sR0
z4s~e9p+$%qYW_}w6P5nCP(ItO7yNhpc|$LPZ4On|l%-nLA|1KHgPDKwJRtqNB^T3n
zSFD(dIhRo!hW^rA*r<$?(CO9Q0+JX`Yj$#udjL$oN#z|aS5nGAVM61I{C;OFrwsf?
z%Gr;hLOYIo-L0|e70W4L$g;BPLy5u`(t8{O|Me|vE%6bl1K*cN2Bz6&Q4q>-;!k35
zRO7CdfqC;!aUDLM6KrQe9U>Oxx;P-yQP+d`Xy8i}w3|F+I2SnL^8Wm-i6=HWzy}~*
z_r2(P6LX?8J$*S`o|G$)L71e>KAG^B7`m)m=3((0#(v<RzQcK=h;uRxTW(+El}a8_
zP5x(}`)R6HfMAT{0ofcgcW{FBs2MksuMeCeysbPgDsU~`7hbQl#|yrxFFPn6r^<28
zRMpdrS#46-IM3q6V}q1;UygPbdcSnhA?s>bw-3cTT^XY=k~nSN!a{J3nU%i{0-k?v
zo8$PVi=y+u;}yL)duIxF-T9_C3CKk>28UpQ%eJA$s!h<DrLd%3kqKy}w72ArOjTt`
z5`Dzo!u|`&H+pJRAN*3P3to6nP{s#E)@xs$`ym(~Ssh+MRzwuRtkxcaAXP%aj-5T`
zNlW*e-d-G*_3G(snr*-YXlsuZkh9R9{VnlNlo6AS<OXEc+C;t>={;4IjdqjAH6|8i
z{n<nLG}r4}!-OIiq`1{9s?+EX!-v{L>0sf>f5Dir8>3`L5Y8CO?vWLC7iD@hbEmuT
zn$uvZ2;1w%BS8s&{*+?)Eb-D}L81bJVV%eQ4OzPi3-%$|^*jg}+SZgX9h1(wcm_H2
zfI27kUU&U{(pW>!r%o<=7&%PUHXBM@CSe$lW8j)!M3iSHbIUw~EQU%PNr5ZR9+~a9
zL@b@lF%&&&>^XXll<n*~fqQ-uaOFJ&uY?NDS5gcEg#R=L1J_Clrl4s%`7)#=oOaB&
z9_4sApb+MH?`DyqV$wqiJy$C|w=4SG7m5;Df43y#zr7&3x6j*mAy4OC4ZV?}%E4_E
zS1Yo&G=$8n{fK$@INcp?dNXD@O4Bnt5ycQv4fiK-#zyC|FJu(1pAMj=%(;3aLZa(#
zUba%N?dR(dSD}f}#}(W_o_`=e!(JuR<8mQ9v&QHK5~zygO;+qQX5*alwY!?u$J!*{
zBf;_DJG;n-v=1C}LqmI1v3?;8K?VbRDEIcod`1Ihe=badwAG<?Gor^S<H;yo7tb%Z
z+OBSsV-q!Bq!zUt81_nO(qDgW?)9Y;J-_icBwylghag9TFlW@EiRjS`Z<|p&j`wNo
zLU3b-i0kJFP5h&GDxoXmUr}P9eJvpO6osGskQDxPv!w`qrFEFaM7=%B#}4bMWd3;D
zxuR`g1S;;X$8d8&%el!}P{|ZT5?dUKTP3}XtMaf*&i$wpgx{5aT0dbvU6NvkX>4xv
z5J7%$&^MBZ@JXzQsWG1ssqUv}w@;oOQqt1v;-Miu#02P0R!lRg6+ab@x)97?8b@}g
z=36GmpJf|Ys=w-hg(b4CKK!|u+Qr9go|GJ20V~HFyg@Mt6o|#aEk2W8X+8vq4>UMF
z<TZOqhS$93f@&mlYU;r_zPQRnAVB;t`gI8$r9|4EMH97d<y?!1kOcpb0bQYxgp<w`
z)<wG8MTpEW{Us&>wjIL-za8^%A1)du>i)gai8V)^x7DNT13{tWA=Wcpa@Oe1gDz$5
z9nNF-*K#k9`O<gyNthG%>SI*HORJ0&Z(x@mruF^l>=57;QM+@9jwEI}CaIx^Z)4TL
zjL*78RedSzfJn`{NDEm%bD-j1e!T(>&Q`>!=mUQ5e*JN>ZZT-4Z3kPsaU4{m=c{G!
zG0l75t>5Q9vxap)MS;Y>ly~U|nrBpjQP^@41%To-wL7NNii;QC`xXKfcrkuo>^v{|
zBJ~5G9c_-u{ajr9x{62;qz$_T7LX%FLZLG8hnjTV{T|m~FZchcuYG-|K5);#|G05)
z{+~<2k1D-HZmG7%&^8Xrnj0NYRPP<MOrH6R6_c|!&h10BjJNFWWf9*!OX8l<x4h;!
zX#6Zdg9&jh)a}mKaVkyX%D;_%KWa7*?xrgKKikqi=M8X-rv83w`7aIC)$l;xfr{z;
zBz{HwR->K{`t#P$sbKF<RqX}D`Ssz;n^XaX{}sOH|LWgp{H3v_qOMcXh(EWexT~SB
z9S`!&fo90l(L*VK{@5YEfl*rV*sk=pxoyRKHdsOW`@3E~mLeCcu)4oAx2vyHi(4J)
zL#N|Wvqx_TQd7w&J={NyH3V#u^VmmJQ!@#H2k#1DpX%?Wc{iTCEU%{%r%jV{L!e@4
z`$G`Gslzd|Dx~I>DI9aSB*cFU4805sT)>*Nsbwb0f=tEj<})&UZ_vMaTd?Gy+$KV`
zj-lgBWKOX7p~VMamp_Sda1@;ba&#*PpDC~LxhpPe@sZ-bk_<Xdp-~ip4dD)+6>Ku^
z1N1PQ+Bt(X4hL=QIk7dIDSTAmHCGjWKssL%+ab{-tHVYGgb5}|M=_PNwhOSYQ(dMc
zkIfwEn9yL)EveTS`>1bG%S|tY`na<6PolTa{1O@mg6-G(scmlq3b#ka5x5XKqKe3%
z$_=<z!0HwO9Y8@ab2+R4w-yCA_}U@~2`%@?nzG0QwiGWv{LLQ?`+_nd*yDDSDi3~Z
z=q+kr-x}UXO$gyS)D&1)fe6uD;L{oY<8YyU(MO`MGE58BV>j%-t<|CLDBt<2G~OV_
zOEKjp3}HIvRUGptDR3&HZ?Q-9f^gPt!FS4!-ZLjl<enFy0861_KdkMn3I(IKQMNJ>
z8W#Qwx4zZp8^lR<Rg**9o~F+Y@V$4ziuTwHZqOn$TDv@`$orlXw12JIf2307%%~>n
z@D;pjYPDc~HRbaznxAB0jzYGg2;E6R0Z^Ay7x|&m&RzDxq-yU9YavlSHo<%RdZ6`J
z0=c+6p)sS8gEL1l<GMs$h_my>e_>&9q1RvOK5|v3lf{UJXj+mt)pSJ-5lz-Cbwaco
zlzPZIk}TST^u<8_vDGfkS=|%MIYHO>X{>vB4OOfZ=LU#lJ*P1pt@8%Pgtk+sX>f;0
zfmI73u;f$4<BLF~UeQmYW8bCP@6HAy_uVhQg;d_*Rmy4&Q$m~t;m_2a1z*WRjmyau
ze`!v5;pu-~ofovH(1NhjSQkQ2%r>H~no6rc<w?_?NsSk5i6zkQJ({2Q>1i&rU&Ou^
z>=y2-T!b<r<2&<xrk<ohk9Os`uxo#?t$gmP)v`ojCnTf`TZ2hNS(a>_x@d{SJQ&BG
zSr{N&qE(8+U9X$pvs2w(y>RvSn!kAf7P-2Qp(oAuAeiB1OWVP_3XH}kKo-)=ep`c@
zJZ~9+&5p;7pj%LlSS4;t0ws{5Z^0YKh4EOd_rxZc0ZVavtxorjF*qp2cwei07%un`
z<^*{H;nW@Tk83c_LSfR|M^$)s0?$VKo-)f01$t#{IUE#Rr0|;Hd~v0uP@FYkZN_Bn
zNY|sWdu&Mf3Q){#U>U(7zw2b^d}a!Mg-oovJyy>MFde%^)FlNU%mV2)N1Ni+t84pV
z0gAdF7w>}}yZFREZSu~;b^*9y_#wP|1d4yLnXLURpq1a~XxDVAqO^67W4iL;iLai5
zeXla@4aX2Za;qK7YNFjVB_VeYxzfV;bk2UliPBPfO5FzXP!rN6LWh10haI9L<?8S0
zgr@4Qw#Fzh9|9D6KR&daOW?&7NcBVz-_4vcG|6l+?+Ml#i6R7s2VJRoncec%V#7xb
z`{WH<)-wwhwbZL_y<A+1OH)8W6d+*n{C4`;gJH7HdEr-3(3yg!fn!#%q_;_HuY1Xi
z+%KZ+h4(Xnd2hYj$1bq7&Y^@b1ybeWT}AzIsrn@8t?vt}$nk@yZ>?_}`&~95Yr!x3
zk{I|T#Ep41?28}Tpe_BprvM{<i0$$rh=O>Y@;!ZK<5u|27_cDgbTc#rFB1u5CO^kt
zRoo8wzQ8+q*dH3f4=@{*+L3nt{W90pUs5SwpP9eBrdp7sz64|+z;7G>KubV)R_#p~
zOeVdm6hc7HZSYOt*hH4#bnSOPA=DAQs_#EA6D6`vU_u>Uu(Q8B5_c7@^LYUlX-O1}
z+Yz2@2$fq636YTCCDKPdJ9`E)3{Dk_Gd1Aq7W9xR`n)7&5{zk{rXUYCT?!P~u<2*`
zWeoGsr$bAsDzO|X6EDLV_LE8;($IC)(BdaPd0(;>_|nn#B!w@C1E3*i?A^D%)ztng
zsr9a{Un_`i5GA$*<tJ+r4N@A`P{WE+ig06Gzm%&$cR&*=q?}Q1x+omAck8#Sb;rr<
z3HM`fafW<vt%3x@FxS2HYE0W-n)I11BANMHS#_=?!jlndMa=9}3k?6V$>Sn$7s0HR
z3tl|8v%YB%aAXT*Dk+1|=ri6+OulK9%{W#^!V<r3;fCTh#Bc*z+@re=@p~K9p9w3J
z*5xkf%Eix<{l{<Wo)oA)`^vT^Qubv*Hhqg?P0ssE!#MV^Zq;WKGf+K%cyx44VkLz)
zY(hYZkV4(b_$!2lOjN#b{?>yOz|`%g(b5=mRyn(c76-O~nl#g&o2G_b{wKL&JZKS8
z8_~j`iS(sxwtw+NK_<qx<IV&s2ZEhxaBT7C47!#q9r*9Kuur0u(^oEd@%veX%t-Y<
zVB}-`%RuUQtMPuziTsAR#2S{H{E`@oLfnAs;Yj`}D5*s*=WMgFLt~KB+0%g!*6Fcs
zwwY2jY(Kx+sUn-8Evi)(OH^0pIGzJ7^~jE#ofKT2a0f+{aFh$VJ?346F4&d>)u!*=
z8P_r8&9ZucQDbR-V$;B}TnO1F)n|}R3wK2(hbVL>g_|=HtCh+}1yc<`PLs=7QC~c*
zI4Gh}rvB1tlCpIgs%a;zUvcdy{Y!(T)&E%b`7uHKaU#l<$ojn{x`y_YJtg7~8CvSG
zRov?=zbVJ-7`Q<9VeaGytAAR}O)mP`NQ@Xw{F+bWD46I=m9xn{$cx!&EFPI%t138=
zpyXgC>a7O~>a!Axwy=C?PJ6I6BLWASVH9-q{T!MLC~&45BPz{qd`?*D<z{SMU369~
zz?~s@;e4Y9^U579N*Rmg2p-aOWqmdC=19U%e+G^*m(;TON!KKY0nfc)Gb87V=9exS
zvNFF~KolrTVSs7oNtx%sUQz6FsDU1wg5^cMKtmJYZP!UWGVCeW_{ILzAv|TzU}FQr
z1%tn}!-B4Aq#H3GzE|ucOOj%Ua+`~!hk`dekomVJ;0E}R*uzTT<L3N@zP@hq!{;_y
zzj@4^T&{k8db3+LW`^M=s*zF;*h@!5pVF7I2aQm7gVFhmsihifwC%6v5?i|$)S&dR
zkdEPX9<g{udd2$j21fIlkXbHF?M~JxP)H<;BV!=zAJ<<Pp`F1!38)Pb+ZhVzDj>vr
z6t=fWw-%AYX6+~4#&0)a8uD$6aukq_5_ej~-Db?)d}eiT?ez|bjywN4I%nUZ>SwPV
zmuJ1&Xrw&Zu-|pBHv?*uxrp@$7PNdn{%56Aw<Z)f8D7R<4e6FqS~CsJ{if`!GN{@O
z{vMEWaR5Tt#bXi|Nub3BupkR5dIg1m2G?~?-V-R)I*Vof1AZ@LI!iY*b=yE#-E+>t
zwO>ZN-J#HTWN)j&nsQkHN}|J#TSt*HmhxLa&ysb(_7E*I+%D_7-JM($wl}?(*T=|H
z8BB6Z+vyNt4_@nAi)$-+ynqR~AW<A3@UI01kVP~zXbF15>7RU_w~mheB@6$=E`aBI
z<8JUb?pUG+qsF20O-2j5dTzVxZ!u!0>|ClpjMN%OjVF{a!Uvx5C((j4(5`PJV3+mw
z=Q&_Q%o`|57h2{(QnnfKf_QH)F*VjUEMD|Ei(}@!-3i!B`ko{>V}$KGJ^s@VV%#bJ
zB>m-!XN!hi^1OIZUCu(NKJA`Hb;A{jxSI}jX+jl|p23l>D!mG*(X&}#4kCZxCa-a}
zBR>@zDA+_J<kx2;miT4xm#bd~3<!3Lb+l;Zx#RdHdiu|^ak<RjW*;|NHzvG|<@`N*
zwYEnrOSCOy5!;QuL0amXoY5~W+T>KSwzKf0vlL*Os5cH#&0G?6b8mC3%eG~Z_yjzZ
z$~=jtZLaK}0RFqbgs`}_Dnk{SkXE3=C?bcFq1hh4cUJ>S1)(|MrbMMYTrpuD6Qdzj
zHUmO7w$xi+**h<akccaacM1JQo9sVRc^&O6eDNm}Zm^fSL$8g&>h(TRJC5|633O3U
zQXy`XaqW#IfHKfyKad&AzhID-9(nHeMBI7zW$`37DgA5*<CvF$qIMQ5bIJb}#WlMJ
z)&r-2?xEP6=(ZR8$mnOo4N#6Pts8am5)3gd5ysD7J!?%H_--w@9`1DG>ysdQPIb-%
z0&_P;&;vnB5uiue7H;zVTHr=bNQ*!jzyd<!RYBSGYdWBNFjo4=0O6*z4!pu_q(-m7
z`|VPEE;*ADdZ6Gx<E?3Z(nJAO_dozOcdOq@V?a_bca0Ld?8Kts+ThB~d{#n!A@72L
zuwkCDdHNM(R)F{b+w<vE2(N8ktm>>IACMjO&w}izhB7{ZnQ_Z*Fng)#z-oNEep)Ej
zKr}&W69DDQcr^U$Qm`JfNgT+94W}NjVctpA%~atog@@KARaAo8iH!1cD7PM*Dh4_e
z(%k}$JOHqhM*DTpV~W+Zo&r+Hnqi201lYaKILIU4@J{L@5!=11dV4>A?9&Z6Vx!=G
zL@@xV5+}n1YusvYiGGil;`mGRl<_m@m}B40_>x^X(}}c*tu?3Eakxbw-Oo0x#Olf+
zS+$x<d&F#fOkY73Y=9}?)5k}B5=%1NPNXwa(pnIRGuD6h6PT#fht-ch9mIVa>wtNK
zxH8Xl7CyPqcIR82Txwj|zn_;_(7=QNyW{c%_GVPzh9QV5TcD>MAUZ(G?Kh*`y1jgt
ztLMOdmh2=n=xU-1OJ~&!wJzf}j0m+m@Rug@u%VAhN@mLmO3|Y<Qe|MDm^o=+xD}_h
z-Kts053+T#9}%x)6nh34J}>|6Fe6!O7l&$o%u4%S=&gS3Wv~0TG=uu!6rx!7P@Ev6
zf}=i=HHE<wc^r0q0uCrn675^PS}9hWET*)7lF^*UJM{Pw+U$hOi<?$?r?CLA2i*fu
z_&+;yDzhOokn`00pF*gx9()!Lm|^Hb^G;QuWPF$Vt3@LPS_6WLND?*^+bi+|azDJ%
zX4LD|{}E9bm1TNjNhgjfh}BYpKN=%b)p2t=XE#ipNY?G6T{xK=HuINC=dhdU_C|Nk
zPhb2qf6Hy}ZR79oOa1CtiI--M+xt_8CSfxA)#3xdk@Fqxq71U~<I~gllb9ugZN?>;
zH#OiXhOe*?TLhy`kLFQsG0DOj4Q{4okc<Go<Nq{@^uEIZ)xK&r{ZoZAU?_pAdrJzh
z@W^K$CCis=s-+}(9f*y~NV>ix_I^z0(7>zS7!YTB>hzRN?{PJjRS<!r!c#I+E0bNe
zl?Nl3bZbP=(&*Z#g7vSLVdHA&l?BoMA^jdW;+<_uvuo%VkQVoNL+&$${!d;*?3v{X
z-;JD<PKtzNQ;k`g*A)tqoD_)@$cTCqVWUPuN_@cY2IJg>YBrClh`Y~f@>L$kx$Ht^
z<xfT)#qSi~Ob)u)-*A|WnWRR3S<wf_;#D;G^gwej{$vWrro10UWisEaJQTb+O|pKP
z?LhoTjug#t|HpHA;JnO?wO_>7#Gpj9zSi-pHHskVML$lvv-*n77LaKhEx0+g8mfi`
z61+?KPNA`)9M!i|Os$etXCBsQeBBdH-AGZ`xs5kp!F&#Dal&TjqGQdEq07tYHSk{=
zI|<LBuVvNQ2hbD!KPY#KR9g=vRnv>iPEyI|o;=`xPPqaXBcx<Nngghan$t<nTfNW1
zcQMnZFQ>(^ZG7^fJ2AEfS$>fSX0qTOU$&&hE}|W~8p^B(w^-CpcojsJVHCsb#5hI4
zT`T5DFE_y{%$~B;>*cS0fiyWM4WW6Y;YDY!7%?-r|C0P;l09^3mzk{JIE-eHTu^Bv
zI$<~ph_V?*NDg@_A!<;#>Fhzlc1Q2`vv28v5N?Z{*t0&FmZ^TrQt)&MWmmROqq5`(
z6uN15P7H&9mqW+&f>XK$xnYGrdKWe2J7kI*zu@_O{XyP`Wg<CuO{eF6xze+C+eSVp
z^<Lz1ec3#RqKgbuhCj%)B6KbYQ3*lb5J6L-e(~bjJB3z;*d3k_J(LiUf9+Z9>X4wj
z_%hv4PS`d-8MBDUIV0y~=EP?>%A8Qa0y%_?BBIcvnuZ|Lz02j9MA+?Ao3rC3S>S0?
z1aeD3lXX&*EXIl{pZ6tlEpu)v+zujUUT%W0OAbj>|CW%~&K8xE0ze0M;WAaQOz0nl
z5MJ)6?GMmmZ2b%SNO_T8rsN;bQ^0%|fA<E~)z{Y<o;au}Pb+Uxqq(o=-@*|=j{pt$
z;oZLARI_!)OGn_3MctCMuX;|#74|HnzV%8M%7x~vJn2e&cf>gQ#r(V90;QMw+z`R*
zm_HXle`!8<X<55rjl-156};*K)p~4CWB;>JCJE*q#GWuIhVo-d@bq5!%Wq^VMMg+F
z7V(FlTGM{cc`Y+SRW=%$2IoMdv%6Zds>|#)>;ljqqrv^tSzOLnCVyWX?7Yx>DPy4m
z2!BkVw#jbebFu8}f%#k_7b*c|?OC=`^~@@bLv4G9TV280-PIRJ+1-s20l1JhBi`*x
zU6ZHvS*N<wl`Ea#(hAq-%CSe37w;{{jV`7a1V5SNYbb4(Vd+>@A28b2C(f_Mdc&>Y
zIt-%7&j!`DkEFDINjpEnCp?+!yH<aFr`5%MZfNf^e{gj%<_x%x2&ybMF2+)rPh>E=
zhcawoqfT_g4G_*pjYQF?fHLrm;!Z=c!yNdAZs`DH%a^%hnwqCRzjQyV%w8WLb7?9L
zbq&#Ov~)Ell)?Ht8twl45jK%Cv7=JuS1O7vxg%wNub;~|yj`NT{<R=V(k3dKlc1mN
z2XK~c6RQMdf<LJ85AT!u5D@ei)Z<t@-QYr6r%h(%51?Z2C6J^V;s*KZ)M2O~{xJa8
z8j_pj*NNsvMgB^k%gt3+9Q$V_4%(8hH`MoAjIo}-L+x(K33Tc&jW-g<s~(kExQ6%_
zJg?Op#p2h(y-|YN3Ac!)F(rfwCg*I26wB}m=VuSZ$6|h86}iqjKFQR4rbpHs5hge+
zb*Vimbj5E!-mtSgOfG~<p=x92<MuUvv0tVQHtB(<RcU$dr*OkHiGu~zuqDe84GpMp
z8!<Gi9CRlgwC^c9E1kwzpRII}*m-k~KEuteXNNp@DRb!F;?NuHje~@;!vWonnT2AX
zFp<*Od`+V*L!P&dfX@j^Ic8oX_E!vTcAl<zvS@TfOgq+N-vY7y0VK>5pgoI`_%YR#
z_MSb_cyevHk{H>ZBNnZpSq>}g3gZ~oua7%2WNK!$bA0qb9a5TVu$HWU)%4M=MXQuJ
zt-CPgA2V+_sC?c8e<`NZ#m?Ie;|+HqE2%$iUtFvDp0}!cap{YGw-0zNB?fby3?U$O
z37yCJG}+x21B>c1409WS$Bjm(uAKl0BN6$>vBSn|w*upeyuM?EuRli$))mBip?L9a
zLXKKkr}11Id%+v824*FhgQSOwsC^;b+t<sRw|~0NF!veueE><^Gp>cM5fOV5z};Z5
z(Y=B0QxK_Xp4?nOxr9DC+)W1ZE9^!};EW9g4KnPAN@XpsQH>21qVMVE+=HFW)9>re
zl<yU~uEfC>7z0kVJE7NLk)5xyW9o-k*ITk<6@KKNiw~K`+m}Cipyg%Mm61;Oi_UI|
zavLV&K)l@3Ze5QA+5a<D*fw6YZ>n&UQEi9iFOBTU(0fT|hKzl8F50ONaAlG+{)~fU
zhNB|9^o`L4%&SdV`<<&h)40h?c7O|KYDGy(?!j3gjrFA;x)D8BK09q@Tje5iYRTy1
zP$jvxai^(xgnB=qH`sd!K|m&<)Rz+ORck0$gTijrN9(X5c^sF6WDDO~U;CB#K%JkL
zuI}Eto_q&dgD{lHL0H!k8V#s%d(f;Fa9}N<quPC*7<fLpwH_+bPi7@#J4PO9){^{0
zv5cfFAU!(uPgRAKcM;M>(oXI>ZdT-r>T>)(_0~ZJS<lCFDp$)(z*Da>QzEN1w4Fgq
z<$2c}MK^8#33<kG3Tdftoc-ibzsq}B<SHdl!s6Ifr78iSDY%)9J>GB)sM_6`2unK0
z1`RJTlZqFIV$B!2a6nEEC(-oIaQ_^wV#aBt*e!ZmufPLM-{C14e98=K`zu5^R0Dvk
z1nUtqi?hBs7tjBtarGST<!_yH?C^&73NNsI%nch$-0tE2Nr<IfCqE{I;~{ns{i)8J
z$I4C@C*jJMhvI<`lhxJBvl{l&X~!<9L>@zbCH}eSZ~k&VCYH>M7mNd5UmJBHX@ha<
zO+|#m!s3aQH+4?-Gn+RCG<5sifwf&cKYxm(%6t2+-0NN~tnfag-it1RDA$hmv3l^a
z3E=6d_Uhr@tfS^}J6e8NBcL$JzQb-c7tb%1S7^U2zNh|<LFPKz#idGTQ@+d3hwMf0
zE+|^BgT-_^-K37nsYtEFvrXaBxf5sKg#DAU@;<Ha(ai|PS>8*LmSKD_?Wcs@WVD3S
zylsM`y8UG3uK$tsLs6ZQ8C~IRi%cz-Crpd<v-xj$2-?h)LCSq335e%cI_@ydb6dh)
zu8xwj+>x42{)q9GvyWMLz@2r`{x-9q?2NCb%yI$6XcqX&cnxWQdd}^r48d?@QVi!$
z{k8JP!bD-(cJ-50^;9R+^`N5`;d;A@$(cmCJCS8~vI4DyglJS56Q7OX3gXbr#FChx
zN>WHyjzC+X?atYg!V=H>bM_q#K}sh#<E!Yz#KH$)$V4;;tajFlcxNq!EL-X{i;2?E
zE}fZi=mARtG>Z>A>qLBhwri`Y6j?<Tb-n<|tbX?jKK$3QoM&E#sN1P^m#4V6suYYH
zs_R|a+U8iWKdK!U&9c9j`{W|6A$wt*ahBL1O%1nYx#z?ImBRr(IY#1tg$*f}<bP3z
zSG$q(6H3ROxJhk`KpxwLc~VF>nt!U&>(iOZ_^Dg@<#rtoo8aW1%i=47t15!RuXv^i
zQwjAl+#B1WxN|UI$}97iiC@XQfiIWO1#w$??jLjlc(frA8Ur^Z_TxFbS~bNwj<x8f
zVq)M*CdCsI3G;mfPi1QXhsv8x;dYOfW<Dh*)PM4=&#VTLE-hkMKxlDdP`rjwzI6PS
zWklZq0w>)%vao6yIjBzU8n8#v?8isTsefDidf6e(+KyUpQh)r}+NPj@F0oSzouzUo
zRS4Y8c<2T!*NFg)ffj`%Tq;A?(8$fsIw(9B0MWW|Ej$yD{>R+NTI54+0V`S#W{ne!
z_(%lC2B_Q`5o@V$h+pv*vP_*bN-Hc^|C)K>SE3H?=5KfQW%Nao44zRN9~fB<=D98g
zs6`X=e&m3jW%P5NzLB&0b|LOpYMk2jG7+UwANLGgN`lv48fWYk(qoU>c~ZGv%)|6h
zA&GYy-AX}OW5z6xMz1D#MMbQ%XM{H~f%NBUpl8Sf2zG}$2O{rQI5?FH^s6On<AxrP
zT!~v9+BX{*w(kUJiuQB{o+cN9IARr5id*v;?7u)LqcuKff_?Mtsq#rZ+*Pf{)b{P+
zz{`ETF_dmGa7)P^JyA;ao4eD=`^XB*Q{(#M7GE+w`yz#hq>eY1#i>&Zw2}o~qJ%Ax
zf4`!U;%{YrudKpPULv)#;%M_Jhu5>4Q7Yd_!UTgJG$&E4<6OS84u;+t`;p?aWZH}#
z9y7?yp?^N_{mnkK1$f`c_|A_MI!Pye%Q<{TZ77CUl&Zu&unoxykm8hf3<#tSxfEqD
zW}H+WBWy?X*oFR|_Ni*ue?6j7rHNREO=$n&cmGRM=w1G81hLFlu&Ygp{y#5dD}bQ(
zt^VJMlVT#GqwP|Uyp4nBWCriN24-53fw#G$YP;jL$kTi}Lw1MDbDzG=QbDxmiUlW(
zFn3BjGZhZVW*;4<OCmow&r`s#+Q8%u$ArO}VvmWprxaCrpSC8SE-3}wf6uY)FgUiG
zC8?Tv|0xb_Dp3Xu&!-2__2SZs4A1lKOdUlILaP|}9V4!`DCfJ+Xb@l1F1QhBc)v-Y
zKBsK${iP9MwkmLic45w&^Iu_U$joAq*bYFkgW<E|4h!&mMfJvUt>0I>hGr8np#9gT
zp8@hdW=M4y6)}kpBQSV;+mQO0U7G?JSNNs5-C-!g*=4f!{&ogD-N?+gI8e@AEYMJF
zc)jqHEOi3f_$U)lf;3?M0%s?9SGrf&JAfFGmMsydt2R)vc-N)~E?l%nu>mt}9G_X4
z$9n+xprPnj@@0ZU7`G#nI{!m1j>85)q;4fpw8;DYHHB`{V`<~J>I~d3EmpK;*w>$$
zoreLJ(aXGc^GwzzBp@zH3`eHWH;#_O-0%w_wtrS~3?FYOEy<?qI}6ukza0E!o51qq
z30SUU5g2iB&T?jGLT$vTZ?#ICRX2JDS+fjietd2JXtY;C_1=+(cEalPzPp&_Hrvk0
zhZfA3C25oHk}pB#hqV)R(CST=iS*K$?l!orp=HzB_}MWy2`S}r=S=<aRk{J$?7qkk
zuff9jm>+j3cc~W&ZzRIkmnr4*#%U%BGd)wekoQWoDMXi7PpVEh>LUFm_>%hTj*GsV
z_2x7ER5pP`n9c-xt-w?80{mu~M~i5-!25tA`}J~h8{<!#$XBkN68ElH?r#?-Td~NM
ztGKEEzT)+5z_j4VjpXZ#JLA8xZqZXR-BZ@Zc7GF@mcf7GhB63vrw}J0`=aU*?L8WH
zbIphGS8}b8x~u?#Q&KT09`b5tsBgmUTcmMFisB&TQ?S6@qpaDz`cb|mu^^E}=M40D
zKAExX&M-vVv@w0jJlC3&Lv8-B;^^4IeJ~{W#e1hY!*oZXyH$pg^rCG7@d7QFEqltV
z(~OmKj|vQ<rh@9vFn-d@+3)EjIVn*tvaD25mK<o)M!4hFbD^3+yuBaCiy)THjEo#A
z0L;1Aru`3k)PjoI51#JyQMR^o@+La#03NqEMipv}28}{<nDHl#@af*65IqKi0L+Oe
zdadlJ){*JT67`F%LOrKeg_1drJx4J!Q)b|R@v5vJJ5#*gc0Y}e0^i+W&T*E>q*JAt
z#3YCDdI;Y9SZm_Bur*kd+7AjT1<rwv1ApmH?(Ik|CqKQe@~q+#6BD@fb8@`qjezF0
z8p;rbCV4=LImS6_LLnjVx~^vDXz8@HU}$^lpiZ#Om0-XwUwCtpZw}|5a@;P-1pfRh
zEQ0<}Q^pKc8^X)HU49g89}>draS`~y<>vfNk?S3UEXy|xdxhF2<2jP4y~AVj?lBrS
zik$-N0bi<Uv%wh&z(kbZIkFriHw4>aap++@|HXmq=l=C#b4HbeH?{4}PZA%K_+%>q
z(H<sJLsTrLych1mtCz_Hh(_<hT(V#sccP#A`;5pjz<iv=bsoH0o^sWbklETjzu>Wx
zLN#EHqF;=}#~$vNIP#7+8!ug$@E&-Z{fNf?y_5!4$grF3WPlr92AQUr2=zM(Akf7r
zo-~B&>NjQK!M_7krIsdSgzCUQ^Z2Drz8c-$kJNvk;kJ5-HuaHKhMu{%l;Rk*jV053
zt^bsQ!Y;Thz$}c$H2@{B=n$a1%@|}yN=h2uBj8~1F0s6xztlHc`jH4^iCa~1=o-jL
z1U2H_AOWl&Ds-ZPE;g8VTgyW$h8*O37tn@9TwW=;DnDt#X2!^Dvi4fp;7v3gNeblx
zH}(I5sUH}|P=nON;rz|~h47bN<e|Fv-E_OCwiBA20gM05&1_}WfT`5(4ekGi02Z8J
z+VT+<@XG|P2HZTeUWP$d-;l1f|Av3w!`1f9S(uOsQfW^Fi<%1}|D{n?h4kdOHJ@pd
zSb;Soe^y3ZzadAdr+Xd5+F)GTA1|dsB@m^~_Ht(lTzQdN%i7DAbD(y&mV(A{sR6?M
z`57D`&QHc()S@A3Si|`Xu1SzvARex)`+;{YB;S3fqvNN!l#oG<dW6!aP(Q`w8qK!D
zgO;SpOyxMmdkW2KSr(FsM^ZkJcM5N6bOq+}GdG`F%wtb?Pd2r(HYi5@J9e)Gg7E;V
z62<p?DWb^C&X+=Ctxp<9NDIkVQR7N`Ltg8X`zZQbam4`f`yzLNo^J-RS@n`=CXzZ}
zoa0D}MJew|5h7F}y%6DHROb|}|Dr;~-&}|G`g0$8_URGh;ubJz_xO+>d$@v}3$DsO
zQzg+EjvXf4J;?3Pg;<+p>*pylzBt<Lt4V#nUHNNFwlOvlp87{noVZt|_Lt@cOh%um
zAI_BUW%|7IaWoh-%B(oX3>7z_pR3?qfu$b)=w{LpYFz)?N}-aBf7wIMAr#{s18!GB
z(2tq0kH{L2wEE2e;j@jROy0AHmN^G<^6bxrKc`>4<;?K<Q5lLFS?xfk07FH7mTn@H
z2m8mjzsxf`2<wKN<Nb-`WNC_!j5?pQw_b75GE2tC8=GLNH+OgGcA@71DkdR|&<K>t
zk_b`1un^VK1QW^bZ)=L5*;s@OYj)K>y4?gDyrxjr6<Bq3=DzaxzciPTi7D|+naW4+
zTR2w*w;2k)<$k;`ZBp?5$WP__`EfSp7XNhBPCQ2z60mgu^vMDEr6WlIoW0pIq0z<!
zez&DgYFA-{B`<sIWWj$9qmCX^?t58?pn2w;9dq=Y6aoghLcq@}Y<{?3s85na8Gf1F
zMJ^0VEl>Ms{AO$2jye{id-nBXnp3%+giv$dE@JbTJR4TAKZ|&+b+KWrQc`F*(9X^V
z)Z!wW&`dFgXd$!)HBR1ZIm844q(%~$B+Q|2DC|u{hOKmveS`jlg`NEQXS3YSdL^**
z{2A=^+u5WxVk_0B<twD`c<XfQr1+SD!b?D|lzwWL`GH{s*-2-B3PDQWj6ba(t4Sv#
zyTQ3Isa2bf9uYU4wcC3yemXZr-MZgftw54L)7%ocNy>J`4}bxq^_HUt!3VkOQp%lb
z$qlQ)qYr=e9mhvcvfZG`8j051=h|6xU-!Ig^2HAe+{aLh8_L;knel=ZViLQYEDDX(
zy4ggmgEg8cIaetz+r_;F&c%j}QDsC#79M2uE!<8kl&TYf3<g%+n$WsTlCc`&=sK7!
z*Gn;LnRwAM)?nPK-sJ7=0+JR`f9Gmg$RgjkVtq}g`#w$8QM{wn*eUh@Da~R}v1IJR
zkUBBD-QGSWT;MJ#1ZNxqk=E4x2R-IBi8qj~sOu8i%^X~N;q-8D&5y<AL&gAri82^Q
zKT^+V3@CWqJkD!5nrnd0Hl3)JaXAyN5<x~77Q08VemL=EbGq>Tmh;$pVL7y#EV%!i
z+<2yp9Hf3Y!EWGX@|~rg+Ki{QZSnZQx#=FiVrq5ZzwPb&P9%NCX68{G&55_I-=MtZ
zA@p_r<;m&_?;QpGxQa?C$y?EbNBEgJ9kAM*y|6A@0&i@wX@9e!!#6(pOY3YFtxvYv
zD6h7*%Fm{ifK&bC?70*rBJ$7nA|N*Vv|+Nj)RPtArMJ9k)tIw#UR+GL8T%uDfqV3x
z%`Ho&bsO3zVV2!v@H>P;Fm~2}dNGMBI4XKIRB#lBSlUwu)D*P_EyXeoG|kQ>%}oLI
z=x3o@r?ukSyBm8dQWn$`{177R1vF`?w>jjkw34V)f7)8t8oj!zpN?B<j((aWE3EEq
z=~Kk#Omoe6H(cu)nlPC3F-aQV-ER<-)gG#B-cpM~22tk;g06qj+EAAD!^p81?+gTu
zn{Pn3Ox_#@h5V(-U%Q7xM1`>uppo7vv0{%bBFTHB%4x}@Ny_&+)*~>uBhJ{{TKdGI
z=a;^0mnC8C2bKY5Z&8n_ZN%i@`;_8`&J3(k=;Q(0z%*LFzAfhdy_Ji3Mb*0T+a8~`
zLJk>^HLCT_-)mgPoilCyQy<eQ^=WQnA7l<vc9NW^@`|b`bB2x5OXn!?8q$1?z_ja`
zI8(EuE`ZnkxZ6T^xvjael{kp>y^xgp$$~9A@eRZ|F<C{J_L5Vc9`nN7zn=g&86#5;
zmnC#JFLIGtLyCH2qLz@+lbV{2!v=hJ9^ZDfk$Zdh&HY;uAJ-HMaRu}nPE2(z^`lZ@
zy->ek-!db#b8k!V3djv!BjO$89R7Wf=2HCB=UhnR!k@S4MiMSATls9GD8t~-^PW~1
zP=S6MWw-_*JJnmzmo7yzzXZ>Cbbne<$v5@tPbA3#XAk&8EnAJJj~J<$N(uXKqT2sA
zQ6c|BqN@CyKKAc7eHP#9IrM+W%-ILA=rB5x%n!;P%KT{kmyc-u%*v%~)8?GXpM8dJ
zXJ!qBb-jRMzEP~Iw+<BzOwRM*7m7%^fT>D<@Ei0PR0K1IhnN$$CEWtjT)-i2Z#%9&
zm)Nlxv5f1beQi)Q^6(hq{bm}x>Z2t{VqGj?AiY<dS2u2o-s8+M_7EAbDz2*yet8EB
zHhBBP`xke(tWY{P2IbefxJ$E+zA#p)aDI&;K~rnAk5Ox3q}n})>r+DD_b<Y=(I<lr
zjK3~Mi+L`6qPktxQZVHyn0Nn`SAvxwwXRz>ATpmvR%sttSgri$V`o`V&UL5DkzKEP
z;uPD|J5C>~y#?Q)ln3$GU$vW@yDO)9jb7}MX#Gx@2+U}xLn%zSiTEUeE~Yy0JliV6
zGzua%NAH-P^?RCTiS7f>J|G^=*+!Hu2$9Y!+G@%AR7^^i*aDGd%k1nnt;`uUg;^g+
z6QTDy6O;HlrKh<PYhS|`pl|-tuvNm<31d#(HP6dE+->5IjS28q?C9d!9aor}OLO9H
zg;|T^Zn@X>Rn8(VDvgk1vJ~Z2Fx%+Z>Fzng?py-T9E)ZnWJlwOOwIUEUj1Rw=@rkG
z=ud1MwNP%8+SbeJF~W5D<L8_`3tU8*-+yUpH4KV2F<!Pavwy5dOx9v;&0pl&?A)vq
zZ`9g&#v~8ahPmTn618{?2t&Yx_ETg+hJlHP)4JT|EB~C#imCmIEuWMG`-e{@(`s*<
zMg%^UyO8mV%sWiBaQ3A#?>RsM*OW&#PInXQTL(Bz(2a?|djq}-3wvi4>c=s*mcQHU
zxb;<A#eAgIA3&I$JoT&}H*^wr3MBKMD`^#s|BK5Tjw!SmoprUa=jVNuK4WN2GGdG0
z_F;PQ@^|X9$V{=a!p;IV4c^7=hYXB}t`?4%P@fk!W;X5KD((Grk*v9F{p`s;U%2mh
zp|gtOUP%T)9gm1S5WMWr#Tjp9Nl+|^ZeSE_53*x5hWN1>?bJ3F^L7toaxQpqH9B)!
z^9CNAmj2nrUxmvo@ylWLDI+<+H;wr$V`8Yb2s}ten6vfcetcKH6ZSY-Wg>YjJmN}{
z2sQ(u(2P-+LbNScd&>H)pmV^Rcnm{p_0Db^I)0prvs>#C^$T#7x}7bWF|mj2Sl7aN
zU>E<&=>9LCIy&Dyq&TzghPfa=%nMdPxmkzjzY9y?#nfuea?TKuf$+RpB64+OEQbh@
z`4J{c80wT^H3%l)vE4a{c$tH|2ViR|Q3m=<YI`v*9p@ryPCzj#I6XDLIC{O_T~dWa
z4fj()eg_aIQs6#j*p2vw_>Xm?^$Z`+Kfbv&{!!E01*ZEJ7v*_Y`0kns2|s^ozG?;y
z!zi9Oo~V2L+4;75th!y51J5OoIv$<~J{_vCvD*xs@tT?(@#Hk?Rp3ZBDwnM3e)?$r
z%}S<m$s;b_DNQ}Uoa}Cy3m#jaa{tosl3sR#KiL~9t2LkyGc>6eko_qNEPYpg9SSt`
z!OchD#$ckw_XVk8l&W&vw*MKUg>iF3Lu&;6b8tz6$0}5HnCtEA{m|kM-i~GDrFLmF
zJzQsm!a-7hGD^NvE|$pJuB67lWrqbRzxrp0RXsHKX(MWIT^PjuCM$gTxp1;ll~G`3
zb1SvQ30!X2@bk8W93r>Qsf|viPy+-}UWc&VT?^O4Gc&DcS#cVB);WN0mS0h7U#!@~
z(aL!(;FOh|%|suF&<DuGFVK&Wg&}zAdvSSiXmPa6E`x*AKHQw9)+~WfY=qtExHRx6
zXLFJ=9%77NJL;NDxICFn<$YZ2aoSHzeS4<iB&by4leiK+oBEZLH~hlu@AEdN){$f#
z!cZ55bE0&Z)1etTNw&RH5CsX~y<6=Pu}1>1uY4U+@NN7V&2L;T?vO8^dC#lrEX<H1
z)v)3o)@WX9SS{0&yH`uSQDn^kH%#cd)MVo8)4qD-v_;pHaQJgJa^BNWPtEjgqj&g1
zD*(H-TBjjM)csL#Q?xCi9Zg+!6R9+Y0D+V*x`D$ik1p!)^0zTN_$w}H(|A$YU}y5A
zE`>Z+{WG`9ucY@C3c1@z9&mm3l#Wn_2ZP8RwAJ?xY2P1TwXeb%mQJ)ZJkSzEj`#Qw
zh-goT@{F052!#GJYSh8DMGqp7o3&@uC8{y?Ldkh$&?dg^iS_3ON}xCP6>?~K)i_h#
zn~I|%$%1LvEg(1C(07RnH{~-CT&v?U+TGYD&=Zvm5?@5JF6TMdztO#wi9$Ycg*~nL
z?i^vDi-X34@hJiT-1x{&{ULza28=gee9;7KF!ss+Czbt|cY*twWbljoO3NqMu9O%#
zSbygXYT-7?JYS+Y@d6IZ26H=N)HbV+?Uv^|W^w4y@T!B}YPU6#BYz6Nk`jG@iN!F$
zL!G~RqFJW2?l>4P<_YVgClz=r)wJcP0#S{5dEPsV&aH#`*Dw5EJLm5=W#!xVz@cvM
zuil?_%WNxq3p`%Q{Y(CzseSwWyW#~*&&yRDcl<EDX^-iO8$a?saZBcxt`~`k@HUwg
z<y<_q%hs+yg!9moeV)Mwmi;ODvE@E*oy-mIbroKhkL+q&F1>E?!`IdOtr&A<_)R~z
zRXEPFc<go{<9cOEkZ`?R=IR>Vjk|tq{Ky{vG5^g&^&_p<^>yBD`Rf%O&wtm(VBP#&
z;X}vXeU$Y&Ei;pg>pRPH(<{I=Y)e1NAMx)nyLEi#x*DYq|C%5E*wQQRcRAR7R*<^l
zyNi!2);;VMaL$(C?EjkI6w)=fQs{MR{jPWad|$lX`sH@i_mK6M?mvCK>VDj9sqM3_
z$a9s>=lG%jkp1v`j@kJ_QSsSa8}IW99S!2&%=Gxpwc?*Wj~*J7^$D+hwA;S5UP$jx
zrF_SmueFjUvh{~|O}z9~+q-&mdG51HL%;Pr&n!+ab~j0X=91va>R!+DAstvUch60J
z_^$EMjj!{6xc?E(UiaeLu6NTPNt(Rw-t9K+%*%6c(@lb&9F|sI#`LaXl0m@yMSo-;
z$nrn@)@?R>!A;XG#$MlrPn#Z@ZZ2+Yz_@f)W#p1+P7JJ1(sDE;4~Q1mY%Gd;@%Daw
z+26hQT1vWJU;ntjZvXWfw;%r*KJ3kXzmES%KSzg7Qdjm)zuno}?pYQ-wP`<dX!X0;
zB@fKpjAvc{D1Vfn?T6K6?_JmbL@#_%wre%(<!7LI<x3wXa0nW25EA0B*Q@FRZtk{c
zE}hQ>4E(Nt7e72*yZn!`oZyu$fBkuxYo~py-}r!cO}Y4|cSa{Ce&1A=SNKo<kLkz#
zhyL;(`YwKtarUvvXCG+GbKS|D5Fg2YS-o{>rkJzIiQasTsmwEGx&1x#r^m3V?p4{f
z>p%b0&({JL47*-d7e+6?{V}tS>;2Qq#V^)N)VO_Ee)z|lB_HiOx772k{W8yY;nl4i
zYh91bztevvapz7Mo_X0bKg1u;=gq#)16=<N92&@~3U6WW%t_j_9k?LZV)8VHqk{L-
zXSh8)!Mx%@_$=U1@cN!Q{`(bq4}RqNck=Axxco<3JM!*rmz;IFz*)YHJ1ait>d$DL
zX@2r|>zUkRz)e=}Kg2&;AFPx6;PS!oQO8{y?(~PV|4JVUw-WZuF68z&2d<hE3N?a&
zJ5G-otuuRDw54+Gx?MFp|1(&-$4$TVH_a?NUjO*C{|u$~?>=9CWtyGc@ARX&(Fc=`
zXKc8b^R<3^<;~AFKi?&N`j-1R|K{f1h3EbxB&)Am&+#Mv!{hiLiI1hd+b+KUYHM`=
zU==&?ESCM-cg*l*J$hkw@%D*(Tqe7{d1P{r@kZ-DukTewU45PR>(Ac)u+e;X?OOT2
zs_`Fp1iFecgC<1==Y+_*X*<`Z|8tI$1fC_<3>?Q$JzBDbp7*qnWfzeRcVpm`>L;O(
z=6_ql|1)^%zFq*l(o1XE?8P@SHk{@IPP8P7Z1yZxDZ8kX_36aZClgwkb<P}}xoeWA
z%`D)pjU~X%hM*0HAS~MYSe~&W{gK+VUHM&mtbrXJzetwV>1!g6mU9$v@CdmoE33~b
zU(^--j&0rSWw$bZ)o$%=U1Mfub`f;Ym78*(N?>bf07C!+1LJ|!B7vd%)|dfP!~}d7
zA;<vp67p8Y&?Wby4<~i19`SfMU+Wy>?_*sWO!vP{UH<2<^M8hV_8%AjE(Nw#>;wKY
t+&}owB+!9<RBkj}(9(pA+h)|Ar@Nvw2LIh~=qm<+3I8wfMiTk|HvtpCChq_M

literal 0
HcmV?d00001

diff --git a/_images/bonsai-workflow.jpg b/_images/bonsai-workflow.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..0481c3dcfdc1105ea5676c4cc55f907580168188
GIT binary patch
literal 33481
zcmeFY2UwHMx;7f5cah#jK#(p~YNU$@2r9j(^e!b3YEYy%0RaI8=`{rDy%%ZHLocC9
zC)5BToP7H~d$09<XYaN4+1Eb%I@f>wZ!$?{-pR~7GxN?o&;8tUGkvoRxcyX3Lk)m~
zg989#Ux1rAfHDC0)~!F@*b5K)#wWtZ$HT)XAtWRqA|oLqBPAgvC8xM^o1B7*f|T?&
z-EAsr8d_RfGD>;|IvR#MG_*8-6oP|`{SO{KF+M&q4LK<}&HwA`rVT(vgky0l3>W7f
z;1(4QE)~vA7l0iAz`@71_Kybt-xtm;Y#Rv(iHJ!^u^Fmw18(8q;@-l;{iD^`+yU78
z06Z%EI~)%c38;175Z-g65q%$@MZ~G}qn%c7^pH#Jt@{UJ5;}SYMken2JiHJ1#3dx9
z9!bk6tEj4}Khb#l;^ixS14AQYODk&|TRVFP4^NPnH`vEF_+v<DSa?KaLgMG7<diR|
zY1uisdHDr}Ma31BRn;}Mb@dG$on75My+8Z<$Hpfn;ZxHyvxt?|we^k7-&@<rqvMm)
zvvbtN<sWk40C4}8SpO#3U*w{~%5@754;PQ{54mt|d1C`E6&^mvLxMYsx`c1ssPBos
zC!$e`&-&3$%qgaKNc+}(l!T5;9KntJL$rU9>^~>ihyRjf|0dYq<eCML;o@Ki50?r6
z1YC||=v!KE01CdY*O!SmfT*0m#9<_;)D@^ly8=@){Nx+64qUM6%IH9tvuPr?@?CsB
zTD^u?IRO{MA5EbR{*i;qKAY-%;m9T&rG7Zs^!llF*em<xC*chC+{p||EMG10JU74U
zE}SusE!x)H06Na6ZUCFZS~q}JJqXeh^5-~cjkAY|6zq+;g9F{_Rh-#?I0TddEB1a#
z`P9!vBfrAU^N@EsXb-C{O_OIALYoT%BKZeo^T3~ds5LkEc{zIby_Tt_U<?l$U4(%<
z%JZ!VznvSv@n-%FV5ta*Y5@Kvey9`1PAASDKC7M1U6WIw3N@Wk_fd$VJ@GJ|<L7;3
z&EQ1dnUGTmErGCr-H-JqJW+Crc|(x=BELDq15117mmvF>lFupiZxx&GR8n$IL;h0!
zU+X(c2j1Q?PU;ddcJT6kG2~PmuTJ)RkBjVGMIqs5o*WAK?d!Yp?0*@K|Htq=$`1NJ
znf|A{Kq2GMAY4dJ6~V8ix5cTkeuvIbk@TpMw?{)BaA2l>=G1z;06dtv0j!jKL>b%w
zxaPEP0A8-wOs89(IGNY!!sv6It1}3y7+7h21IXCA0R#@;0BFs<R}SLn7D*=+IE^f}
zR5TFsB;<i)X1p?=^&%#971=p`YWCP=>#gOHubOy}`#QA3osZrgqM~$>ubVL=Raan-
z+<<rn)3uM+X9ICU6cK~(>#lN*_pV(r+Sd)i<LcLTnQN1cC-PEPk`@>HndiGV0OU)E
z&A<&nejIb1iQ%ghheWX5K@vtt04ePQc}GFQx9z_!ZRz`d6kRa(h+FI`HIlyT-v%l(
zA75Px1r4+PmLYgwY_GsEe#*?MbQEBzHF+<CU$UAmX1!&oQy?YsLO=xYh~xdODOZcB
z6UK!bz~5CCw>tybY=A6P-T=A{-h+?Nh7U3Ckpu&s=8O;uR5jC*W@CL_Sl18r2eFq(
z!Idj<-Cp60S7tU&6*$a2in85)FZ`bzCZK+TRhx9PVUcJU@QP^r_I2Yx&mKl?@5=3$
zTeR+(PlF@Mg(n|l5vdhO&IFmMxB(1%V%Hnn8$e{v-*v+){obBMwuxV90;hYTyW9G(
z0LwLVwHts#tybek6!~#a<G4UQuJ7?rC^`SDiYd<lEp@G?+6nQix@14UMk9U0{Fu;J
zqZ6O{74dGrQ_RPIz-f2?VI`rgvK1MKXK6MeujK|1H+};!ow}%2eR3<3_xpcmpbS&E
z|NA_H1Le8`71jycqd%X0HCf$c^5GzU#roC2Z2}FJ1bj=>L1)9a#YBW`i7amMK-N-c
zfWqerwHWsjviLovwfiUbbgJ|1?!ocxu$MQ0kILq(o?7ZR035Cx00|I5H3j)iT6hCk
z=E_8<Erl&h71zR6t!@CEn`ehQ;bvAy`ZH<*1zs}<hALQ&%|6_i)Sqr?-DmCw0ErW`
zd3;ehq-nWq7Pw<r4!6*7mXw9jeRI%~*bHy}%sw;hV$}abZ|^=s_)OS=Oy-xVX6)X;
zS#|>$q1U|uv}pi4;!rA}rOR8}Kol_#O0OpQs1}3WtbS(6yF!3HHvn{^f*Z<pIg?5~
z8~tds49$h1b0w{>YntGytM>dB)AU1Bi6eeG#a{X0yWn*UXJj=3LIh6Ksy)<BmkPD2
zbtO#QNPii-F<(}!)%3$T$S1u$b<F@GZ3&1i2)0my{DZc<NOW($ya8BDDefO!LyRvk
zGAIv)5sR?HP2I_q=VOA1D8oou?DjhQe4wCZ;|37%4!`eVchhHHN%YveCfjL83k(&j
zHzued4s%NbL-Uua(!DXQu$=%^_zhs{^xDZv*Z2l-7{1ebf<FN{yrXlN^QS)t{mibu
z0sLvC)Y1234z6)P>}b41G%xPGNbH9DQQhnY@Lee91`w1031R>H^1hQr^8dpI1&Q)P
zA%8y>*zqBXZM*?IHv8A90{Gvh6_lQTS-L@+KyoP050u{<{6SHZ!4k>e>q@0!objIm
z!eMIPDTz@#hhgT(;cnruhWt2}gx^ClGpH|BKhvEMb)lC8v-7P7P=9F2FCk_*tB<HR
zhjCpB%qZmQkgeVa27f^&O>9qgce$7JrC;dp==Qrmj|v}WY`c3XGu1fy^nt1o#P0@h
zf9dl0jj$g?xJ4Fdo)IAwHV7dPq(^mhw@{;8u+NNERzc@X8Fv#>v9SgW3W6ujWmfV%
z8MpCHwMiU~m7{qi@Xj%=^*YgCcU3qHNZ>E#&5Qyy$s?Grv@|T!hy`2bhu$Jh%+zwP
zvlrDvmG9*%<8!2w56^sl(b$-D3i+1SiqX38-xhVl&P!;3PO_ZE;nE6am!Q&)Y5k%1
z8Ta~GBI1e7c*^sd_}3W&rU@H+RPbk_TwF4>%t=AGfzM465#G^EnsEeiB=s8p;1(e^
znJY8cOn2P`-7&u0a8*FAib6kO()u3m=3IAxw5ncGnqyAuDr2CV%D|#2Y~V1CXU+e`
zit9|{%4Y-2-Nuo7AEZ6gXR0hnZvc5JsiUU^4URM6zo`4A%2y;AU3Y&TQXn!b#O2xM
zIva#C^NiH)RsN`rg%eEx=)|8F1K5H{npK|#Odr8^E}dJA6cEC`rU#d;XuOnc&8*ab
z8vuI&rXF_e5c#Lo`!PO^?tg2zvz@^grWmJLsma{@LboX<jg+5TYqyZ5kuE--hJ4`s
zi&Z(Ugm-5KXAM#!cU+O?Y;0~2eoj}bQDMmaoLC<aKk|Wy>$W0Z<VKUVrLE6(B#2yC
zFJXowAgv{it7H3@5<LkW)dsEG*Z8D<K<6jN&u{_%7k&r{{p20pG3aR?$k{;KFDi8x
z&+z)R!pHS4HOpSx+hSrumb}ethm_Ov#kLy&tiE+;|Hp2XL1bBqH~R-0&6tO$@-nN2
zi^)g7y+!6`yVotj<8b|C!=`o&v<Cgl#16oegE$~Ib$D(xz4&`bD7J;~IAQAQzKN`^
zlN0lXi9IR9is&689_EOjjNN($UMFa03?xV%wu7=jx3ae`GLsN@+k#!qYLRNeZ%T}F
z<LAsfDtjscEc|3akGX7-F!{rpcx(HWxta-o*&>Y<kD`Sq5}vhg>@(@!N$Jrcb|0;&
z`{uhu*m)RfFbpVpv;=550x@>KJS~6%e5&z`taL0GC&q>=0C5~b{uA)Nt($kJHDzsS
z>-B!@@l)jDw-WGx@!_EEyn5jlcK56*#B(nxbY1wQA?q#VP9Q~5)k2Ace_IWM6MSRs
zC`<J{6P5c(YXt)AKch~PUj7=uKP;!2yEAJM>94%P%B)qNaSfua2^Y}1nzpG36<uCU
z(nyee`{<VBNnzfh|E$zm%-1yQg@Z7`(<or|T(8R!bK}eMSh<gY)Xz_ER}?z!$+1wW
zfSz8bEbd+KTpPL{ouVZ#DrZZio=rhd%{28ZLb!}P`x1y#+eijSxq`|lIVogTzfPsF
zQn>C)@#id35<pyuZva13b0Wy&KGDQgT6(9P#bx%*eYIWRq*d1$r`Me1o)kZz*p>+l
zaa}QV8Cw(-ed$Jc<`e$&6aG@|;G2#zHSNdk?X2EMXdlSXb;rr^HDmxOAGl*)TZJUZ
z%$0fK;`$~HJo9*(j+5cFwnyfo(4$>n{UmTmM$`&(#kl<$eFNxZ6dXR+=SHgbHo-A`
z_Ev9p@-~H^m1hB1Q@J_uN<}~_S$|VCd3Jm0>ETI(NfWddPj0+C4b-^@!Vk1eUb8as
zhh5wNQdZ=Jk84*_(A0B|UniPC1zRP)&!nWtT|JYZBqPHdqe7oAzO<D4MAY{P_iN_o
z)z2uYe91%~XeUJyEi`mm(!4u&I5#Q`Cm7e6qi0!iLzRU29n0QIX!;v(ip0Eadq+1P
zk8E%<#_!4mv>+bEInVJL7#)(zj+>B!`<29*YcBw}4BN)Uz42^4t%OWne1pQe1ZA%U
zCHVcI6#J9QM$W?VOpose#EU!>E8zzYeCowxS*flL@Lf`E82kQ>%XY%0rH?B57A?9q
zQ7}Q|A>&rYlQP^D{2Amyi10Xus-zJrG`8Jiw|LZ{E52s>Al<sMhpVGE{;U7o7l69y
z5PSQP(yZBJO_G}>$Ze&uZ==CPsiKtGrAvD*<s<Wy3!+wdd}CeE`v7CHfmyy1yu@f>
z!b%)xyVzJqPRaGxx%~W?eT1wGgM?nu{2(F5eZ=~+G)B0)AvBLmv55m})P9eio_fIQ
zxqt5ak99VE%M>#4U&+$eB;vv@;VFnE_QA`uJSgi^IO2Qd5p+OXD_f0}Xjy-$o(P8s
z4}Uqq;XK%+^Xz=qcTPAO|E%pjhzqV&C;QHuxc4)msmrok?1W(8I%=LR!v6-K+&Cto
z8Q}4(rnUO*i1z!bcb$~a)!$MmvW82z)E$I4tahFsd=Si=JXot?ZOZYuPqh4+Izc|h
z>!nYaCpS&i^IN)_s(?TJAmo<1Ophzg9Go&(FE}?iD${d4m0~vnrk2$NG5<?S&VbH7
zY+XD1Pf_KDp1jl#g=>{8IL|^in4>TBavObjU%OpA=fD_4KoCB;D$F+K<u)(Fk|405
zHHqzC5<xULWXWE~l3RCUhM!g7KEn)E-2gHTAqUe~rMXfSp#P@?V5Ku~6V85h2J9#u
zeN5Z>w`kn%oX*8*80I?eutfYhzU$wjSR-W|6YvIWvRI<|h7=`En;UnN!8w7AW^LlT
ziyh6i%-8AmSOv2ji)#G9Dv&m_kl|9``7R{HDEm|Y(o9X}@ebwn6+1%x%((!4*@-nZ
zdr_~_cyHDIU0nC?%b>N11VH~j;m?Hz@&8>N?0+j?-|jofe^{O4&mTp$uZwfIzbumg
zg8xn=-amCv0&}5D<=PO-A3Z;fE;jaaVcz`*UHIfbuJC_YzW>jwQ#HyCdej4NGI1Cp
zV`7VP!?P-N?<*Oi7R5bER&;1%vAp~6atzX;x+;80?<tIC!(wVbLD&gs`Ij^-$o7{!
zrbAO*2ZV;G7IqR-RHbq)@>!d>F-yZ|e&6Nj59M@=?$XP;@b<CwCDfB9ELfqmA94c-
z?rXiEi2G|ojoo)L*S{wI&*v(hvq0tlI@;f6C;M$LF~>is*6bE${rvmt?iWG;+52Av
zz9heY|Cd{uF4Vlrm<^umoV}}$uf5<4lDnj`k3XfYRb~F&*7jr4-ynac-QF4TjeyyX
z$}(1B>V8Kd0|1hB`Euzn9z1{vS-|S5go%(NR;SU&41Y~y|Nnm;MiW~T>a)BUA!E_8
zQUN1essx%(#DEN9+~54=WyV+gLzpehtBDaDb46tBYLx{A76f~dTk7q55>X#@%vzig
z66PhJf|P%MBIw->pOvdH?#74b6bhKT&In}L8&Y`f-+iOV89mkQYY_%=4bP;Zem1e@
zsP#gx(W^e;rF2{ga9|gw?d(?lj<&9bdJ)2Ax_IJ2&o}%7t=a3=<x2XljE8WvndcN!
zPk~3FwCsoe&9^szhE|Yz`K?N)10nINj~Q?u!2;!^1bW`i->Ly-?7Q)ZyFm&CP%+Dd
zN+rVebpukQ$(eKQAv84812H$3UR&Q3Ij}o5m@3FI`%#31aOnNdgQ>`sB&MiE+dln%
zx2lyLE%QZh%fLIBv9#DkKZ4Yi#?T4z86wrz#MJNS9%#p%%rAZ?u)OLx8hc;zJm}=<
zin1%{o+?>WMWvLsIzzAY)_V&Uv@S|os$+pvHRJ5UrJ8i}lO_x2xh_7*W4xo@jh)`H
zW)lS4_drg6^2rna)+Mtojsya^Z&}{m!#~Jc70twLxrJ@T7W;p^+Jb`x^v@bvuJ_K_
zl}1flJ6Y|~ip%nBNBT<-!uq}Tjfj<Wrd%LQ+x+F1ZP}5$fImX%4dAvTT!DVH?9@yf
z;q2u6yiXuz+LuyK2jI;`G6GA1IK?qNt+d0?z3=aOv3GtKLe=-A7ZLX8oGod~dRi*B
zt0%PY;I~hG?8DdZ#$b-swIz>$5jTLUf(SWRZ;6&9H_71aO=9xKy)?xg&R8GJ??;;~
zr!bBL^)xka8*j^G8)>e~w3FeP%{CdSE**AOSyL~`c;n`dxzZ5wP6FLZ&aZEx-gz8d
zJDl%vA8sn5JXQvn1MsRu7EFA`jH>dLiOC+{{lP=X+V|V>)KuP-jq-NIrAQStpz5M|
zKk!v}UUbi!#W%*@F?%ho+|{wbb_h+A?lSY2wb4_1M2L=L!b32NFT*f<c>9UO8Kk{6
z+8ELuL%XJ8B5yu{<89seXjon6CGd+JTNj$x4Awr(IIwIH-k^{htG|$M;+MBOQctpE
z`mS?9)nnsAOTd{?o2pv1H~^hG`q^gk-K!kp{*%QRuVeM!S~R*ApV(m`|7JnzdxuvS
zHc5YXFfLscnV!>dUj}i|)Bv9rcJ~!RXG9Ys-TgTG6=X3Z&(RWC^hhB#ei0n1GT(g>
zJF279*+})%LR;r`pbAndrh19dth%SRc)>xGx=vt>H3G98Fv-?Nl2_(v5O@zR4AR_^
zn;Ym$vp2cUmwt4@7ChGSbrfRTio1xW4Kqqpmx*q$olAdprHmheZtnK7b(sV{z_Shh
zDRBwqGn5NWlG2m<`mHPpi{YyN4t%9wCZ<sGVy1s2vN^q?cKcc<Mc|;!bLV4SRb4XN
zQ`+c_dvvRkH0?nvA;dvYN4=$XtjXftY-|b#Raq8~n8{{i=BkA$x^v<Lmhl_rP5&V)
zCOE`yX}I{rHsoW^RJ)dTG<1FRGUsHUF=2%(na79|L{=8Y@qX>LUt|EdU^iz~+F#>v
z>trrP)YR@c;G5y8NcjT$`myk+s-Ib)U{vP-vAHLAlhKtnH5H2)kwNU3qAHC62b@M%
z74VP3yUg3Uc1^e|25P)J`KC$MkM4%2$D3P}In2t3bvyHD)S}+jGMU(?eU5XbjZI4L
zjrHS>T#>YY|MFFWVs+%h<CHa7?P2T5s-GYz-?>ixyz@T7&&TU4YgwK)rDd==#0Nne
z*K4dA>}Z^ul4xIB*&8@I<MSz7(*bt-2B7wf<!AJ}$Ag-mz&%xK*$2GysrS#kq~r*_
zoKMDWo%08s40M}g`NN-S;^{^;elMm6zot=(3TizXefWu`;Fp5Pcq@V#<%(oB9nI{7
z+)<#fsV>}3+XyLAbSv1oZO+~2_yy6Re5}b*6#gnZrf0e>h1s|TIwEe^B%jx7NmlWz
z_~^;lprW_8m$zjV&cH)C0Kd=`BJ-n!&K>m?=3x@M6koJcY+C81r?6+~muZze?^|^Y
z(yO$QV-6PZEllR&Q{SFpEOblIYJnmQK|e)huFTmh2!O-1kyaTjVTcpSBO|bbrmT0I
z{(?Pgo0Qs<S*NE*;zS(apLV~zS>8)Uw@84a)GZ=7B);tOjLTUfhB&o97=B+Oqj$`&
zIt(({6yT{_TMyM9lZ^I@tv=l8$lyZKrD82HhS6#*Z!CZoN3M6S6X{r@?QvCPq{$<P
z&VA`y+X*DxS-flJ_u?==tZ*la#Z9QdyFgj+r^=@%k=)jTAs_=>IT1W}zMTC&shZ4S
zt?RQ|>j3f5D4m`0qp7K)<6X$@CXmk?qLkhg8CNt{CY39ipZs82pyv0P*YT3iTfoSt
z@qsgN)`CcUpmm2!pEHoj{}~bz24cqnt3R2N&|%jo>MIe1HzOWZfnn=@>Z&F-^zP8Y
zB$c?0-RE6xbC!yA!j0Q(w+?^75LeeESDk5E`(Y6NVefoX!*1^Ct~j1A2VAONjbh){
zK54k<NW9QPnJhOuc1SMLOvrJ3Ujmy?+znQ#ALmHu7tHi<BFm#^Eq_#G%<mhh5$J&N
zo`P;`N!JTjqvtGwHTxG_^MP$~bdDL4DBFVVK+X@3BwQ7?7toFAMuQQwUGY8p@QOa7
zPuUMZ>Kqon(U!O)d}S$lMs~i=HJSM{b6b}IrH*8DH9rJ&RU$=-!@uJLJOOW^ulg1i
z7<*N)kCq1ZL3@t#PXa#kL8H`F!G*yK)9Uku^y_hSNzr!-k?m11kKxq)$-8xQ-id+}
zz8@tIM3>dVEdzp*_@1wFGRS+R9e$(lF?Njx+7Ai8EWafy&1E+_{kr~$Q4y_k3@ciT
zv5twKkP+$RC;LF@fO1^v?VMW*Srkp+{rF>AgCkSBL0@{7;_<@u`Xj<t<#_`MpGUdy
z-eFtX;*k*P9;wI#*R}nY<9*-dz=swlTeIq6n++MVV9u#s*0rP$#l?e+5j^F9%<r&x
zEX?kV8pnjjyto0}N=G~Ln<GQ(S3)8{0mjFAh8ho{gM$c-?Z@|jBy&qx#Ywh6wDY8u
zZhPB~uvf43x#gntpM0~5S^JhWu_nULMkL>DaQ*q^G}ybxH+qIHNJZ4wKZ0NC_kcOR
zy(APS7#iksPG7PS?`XwPPJtq{+xtp(&A<H{!@iwCgpzkg#N3-KtCm+&)!bU;baK8g
z*CT%xbjy~NK#S%;4<;?`pBFdC_E4=)xCj0Nvb+=~bRPvN7`lcQ8kIshOGU)rJYMVs
zic<d)n`-u2I*gu&jI@_Pmfm4~gAL8!x2Wux1Y^GuJ~Zz5fxGkm1B7~b)nKRT>IEr=
zDn0I!LBjQFpB*Emj^>nNKM`?QI-dukU23t0fG|qU1^D~*Wa~j41hv5a=W-XagG5Wk
zHA-Fd^(RkBs{z5ZoodgKOGK1A7Mz$VxO+QNY1W&3U8g`rxAEOR5&LFyK)Si~)iZy=
z%oW0qS*_!5#2lG4?C)@2jjKFf4DxotVOJu&-;9o<B@>bnE!@q>grKS0hD9p7Ut2Rd
z>^Yu8+7w=(ES3d5RtO`VcdH~5W}k#7n4EI!PkNjWP+)@Gbyn2N?JO#O)ff6dt+${y
z*YL|SNlLB8<q)-aa8=aDt-~(q>-9?!Mls?KS`j9m8*t9qk!mp2Q@G%)`P{?!Bl;^y
zqUZ)dn%E0Gk#tx>(cA!5Epp%f-k`PRQF4nmuK9(j1xj`34{RwZEZ+H?*s{HtjA>c@
zEj(+o3~D|L&EPsbO;7~ocl?;2HOU|4ZXhFgF8iJR{O1k8%Ah~ZIkzD+%h3^V-)G%0
zhjgOyHUW^@Oas-snvBOzyWItu8i&OA((kq&)@*j03&wSwk5p{QSY=pgciwV*-~WD_
z{(${mc736rQx3i`bVx(p1(SBM>p68M-6F~UGB_&vy4mw+{_;w1eCc_FuCR;K(;Ce=
zTD&XPesJl*p9a!p)Z$YnjKSj1;|WAC*4+TIg7t3zEeNdSphi`|8%T$k8y4UpfZhNC
zu&_hVZ>;OdW)%}kZi2ZAy8*<&_h@jiwj|L|?==}y?e#h34S=~65@~FD<j;+i`0AnM
z{b=0wo`#5({{tug?-aqT{MGq=Ohe3^g`RJ!lR^~<s~w*fR1`6NOM|#283)3@7M*Ja
zY;Lgn*`!>_gSWAdXz(_b>Ad<zT^f*5JUZX-@_<i!<eErNEcT+U%w*-fGacKaz(<}R
zHP|w_@>jJQSkT|s$Zh~W6gX%~H>0J)dN~iQ05(_w-s@l_uH=oTdh`km@Ofu;%36hw
zw*wzOr4a!+==Jm%7~2;N4-aIW0y~yw#?=p7KO6{v=D^kDFZ6>4niC5nw25NH=J0Uq
zK(8s^SUf=Mpu7VP2f9;a<TtFs1CM4m8K}E-`j^bRyzWl4{-s)_t1bl|R8+v5H$DLY
zWecftjfOTJ{4P#pLcX@RM`Z#Ol&m2znd!)4-Q=sSjYh8%S-0sy3cioLh^<Hcvd%CV
z<X9o(I&zO&3)r2a!hTP~Is3(&djx4<Z1qXCn*||;ou^ahmf;p&iMgb06G)m|45yKB
z1VH1*TLy2H7<x$6z3QMJI<$49e@!`7(Ii*?h0oKr!5EEI1fOR|^7qSQKmV5z@gIES
z#vnuU6k-spuFq>YT$}}LGwyZX73~=ml|91uX{lU=k#Dh6m5Q5WM(U}k#$41lB)%HY
zi5;`cuU2#wFjrf8cwUcR-hcNKPN=`HZKQb0RFhbVeSAYV_g*f%e{WPX>$y&?#AeE@
zWIa#ZiDUiiaRcM1XDyMl^A43ht}aKjj9)S<PW3h4DZDwK3!Vh!FK4!tJ*#i&n?@OQ
zeHwB-W8TwetX65NuYQ@ds<Hku1Hs8n&r)IX6a1Vqd5!ErZwz<`;q?u4r3zNomN}1c
zKH2kl2Crt`k=0>#lss%2)wEv!{A7Rf?W^N;98QutC(HM7?+ydwJ+gk=JA+5l^D|af
zbhxALiVhy_kv-;NMt&t*{a$>T@_tuBht<wpAlW2j!qZQ7fEgVBS*g!`L_DP3G?`%G
z;RixNtRAR4=Ae^z1K_!w9+ek(=izTX^#)1UB`>RU7*f5FvOM)+WXW`=&*fZ9!`1nb
z0C%*2+db_b4JM^FHObuQIfJmV8-P<`H7j9v#q1UvJ5mUSxHth122vtl`sQc4OOOs%
z^ptqLiOwx-Pt5_h^$e$z$8WRULs5SRrw#PfoJ$#hwCy$y$uIH{&|%v#Nw?dB6Mhu8
z4t(KAz4u8bnecAWoMAj>NXB~NauQy<azG7E)SHY=Jd4^>+#T)MnsNrY6n-*MXZo<O
z{6^YBZGfJ9N)}^+VOfG9oE4Ae7BE;A*oW6UOI7<t2q4)`qGKT4*h#aW+SJMj41Mjc
zv+JlRSZt;al0a20&~UR-dnk5A7lUO>B9XN{Ly#9pgN_>j^)?-R*8BbPh(so}!>8YY
z!U-y0XPR(*<x5S3?-8}<#4PArNPmoAj3zi2)LfpYc2KxCags@c3hOkksV!12*7D06
zB6yo1`(#f^z(bi!hcQ;<X@6X(Z0Hz0Ba`s%!16VZ&e(3g<F(%6Sf~E@KHb?9WT8+V
zm;0i|BNg4|GzGcl$gD5;JtGtmrxuw!!+VZ_X-4#55B|NW%ac)IlNMA%!5$S1C8Yr`
zK!$w~JPA-@C$m5mXjWCgW5yOb?ZUDCyp2WRJ-3bSPuVX!@37V^Z=f>IJ|p8o4GfMI
zSnbDaA`szdlVWEi2Ptj3hNoQN@QfH2#;n{qm(Rt<LIhtwFatOP(`Xn2HL|*PvvDKE
zwwY*$Lzf#Td1}P$;nan9x^ZIh^m?7F`b6UUV%)o|SX*zSDMp&GIx{hFw(OJvyIOJ_
z4hGUR99cy>RqIr}S?y=h#3b_Y@OPk_SYqMhSbmaAEoKg)zP<7dN-KYPec0S$dUoN<
z8_44Sa?;;umE5v3?I7&#;q__RH`~(LfOI{nY)9FYZ`6a=6K|!HpmXjI;?>k<nA1+u
z)=%QXd|jeeH8ZK(<;R!i?dj6^)j5m8m0h{UH~}=QYrig}du}aLcuc%VY8gH)%M6L0
zyu1i}irq9YSoQMd=YM3v{c%sX<*E|W07(K#3|Iurv@QVq8`1#$JY9>lV`%H#f&A?=
zdmnoqS(6XYna{B=uOi_v(dz3QIp#vgDJHGY<_{)>1W$>@>s<u=nM+NLpk3}?axJR4
zi^52@`z-=TgA?DUS<p8;P<=|7Y8FVC4o*K)N#CR72|KMD6bLkwpt%b}lj96uAMx{R
zUB}C*a<|c6!c(0IeEeKRZLmA!(}H7I_G{!>gjBebC#QTDQ}xlJhO?uGwVWWIaUY9E
zoluZ7xqzR^qj&9}DvpiQHtGw&w(8B3)qXnxVwx2pWkRmP+=f}95{$OqZ*LXia;3G8
z6d)zCRx~-=`@Po!Q*1d7mzI0OFify-w#?=ZPL1Z6O?>1$@9&(g;UbB}DdOeKm<(zp
zrhHaA)vajOAw#hta=>&{xGJ^m4IV)<USlYy=mTrA(S6cpDXM0Vg(;=Dd)1}B6Gbm_
z*ljh8?q(M`Fu}j&YFW}=Uqe6Fjz^iO)TLTFMnqPIui~K(s#$BIzw)#7l#Mw3ZJ-QW
zM3Bi*%j}f%JArp6?v0C2OO|Nf*YJKz%c=6*Sc5hERSb@;4axTeeEvDNgCCm^>Q<Xp
z3V)@B6y0jCk!l}`Eg``XaraDvI8L7V+W{Wt(x59QRJF|;*HXKR&Y8RfD@zp7@G&e0
zO^)>F1X2mZ1UyquNdvUUB+{IvjK_q<OWLXL(CNB*)98t-2J?P;px56ijYrYnj8^cD
z?PymLYuEyLv~{h0VR<>fNvcpYa?#ih`7xCAZP1&IXY$wgVt1#~mpw6<lduiWX+Xfv
z?*J*Ss?pNBKd_8~v0qq1wMgMBRn{O{he$;CD)Q0BH=o4?nSfp|YMjuvlIJ0?DcB;7
zt`QqA(mZLonaja<Xli;lW~ljl%7FLv?lIRxAsXsqjeA0F@4tQpX!%?_`4Tk$8%m`6
zm;u9g7@N52%w{HRToHTrOWUgz-=vbC0y}vxhPy3(e7ynK`N&DbRNh!7)lQ6j$XI-f
zn|ZZn;0C&Su4wC6wmp1!&_mcE!<Oa!rVyNE@=Wp_snINR1R7Rzo!b>1+v~&~S;V39
z^z2KXMveyOYh)c+X~Es#k;qdKpJe`9FeNXY0-gF@pD**3q|r$~Zl5Qo%xpC@?E1&$
z2lfPVCohvLn~p(W4+}eP3@)7y@KM5>R}l~_ae%^K&_bbKvU$?Tt0fEc0kwbIQ}JZ&
z{C+b<-4ZTZs|xj=d(PfhJ6B*&LQwr_T~podQa2r2x3x!ipPEK6NGlbCl^fsiQB0Sv
zGRK(r=bkJMXvtwtEJOOr!BTtnon~dHQwEn}71=-h%~o02+kBZFF87wEJ=Zf9r-#<%
zGMmz@Lb*F`1=@3|{t^*O3?haWBXz8Jvm*ijEUbgBjXYUN^(Gd@_>4?6+((PrUq_k%
zu2$-0cVa++ckO;?=G&2$)TH>s55H;uZhXvLMJ?jyTn-ol@U{Jz&wE0$e8MJOnf@G0
zj}vsxu`lBp-@g|r!6-CD4Ir$)-xO?gh?c4C(j*Uvi_)7EfMwSiuEAG7X^F$jGW6$t
z+*TO9@6noGKpxCCg7UvFwe0?(v+S|4P3FR@mZ~&w%{@yTcIMLHh*$3B@0v4w|LN%+
zV%!gpXqQLOKFHAKRrWiEb}1PYOc_x{!&aFTr*wV+-p^|zwKDBB7CoPI6_dpWyZlsq
zY|Y3W;MX>*n5d<tv?c7kpWEC}XK0mdXW(dRT0!^mleAGEM+(mx&u6WcH%Qs#spvpn
zu&v=?a;q!C;<nL(=`+s*y2a!$y|K3mN~2Z!11>*pgKihy8)93D?6v>pX2G;~1Hgy0
zFX^biUFvk*p2hO{<}8wa4H%;~5fB20v_Y90K=|zorY0FVI(=i+`4p-J4XavBC;uND
zb@AgKMKO;rlI7^xB?UM4ShFA<>R9%ovt9=~u7j^aWs!wLTC}@yp9B1bao4Mljby6E
z_OHfN<{2L^f1vPRx>db_%J5eaF17lBRrLF(x*jDM9M3Kal5Oqg5}jRyzG!5etNj{$
zG{pf8>`hS+pY()J9qhfWE}xLoDw>_2Hhn52DCAmAR90%M%1b-TE>Ow8t{l?l_A}3!
z87ox{?7F_up=PSyE3WbA4`P(DJR6n|xn)LaRD-08vo1D@3aM%eu?;sf(R3DqGcso<
zy&74*hCf$#Gbo{M=j_r%$P4-#U4Qj|QqH&3lJLvf>hn8u`E3S*p1oYw`W_SJ*ac}L
zjl3VxE@O4Ns<hnyS>6$iYs>D^e&)KFmiyF_m+QICQCWTv$;I56&!IxK*Wx|+3Mynp
zQ}CgjGwjFVJl*A+@^=k+ZG2|Cv#Vrk+V=^$mfyd(P#M)0PFlKL0FF2>vSip__rWGm
z-d!1d$P#lihYZ$+rimsqYeyuVyx^Lkhd`iyi9<Y>LwnzJ<gL5<SH)Q-A<}6I^~I=8
zm5H2@Q$gXL461SN(bHxf&BODUY8~czy4s1o{G_Ix^p_8T@hmZCs_y%5neX*EUGJzq
z>g#U50SLC7%uO{k&h=l-PU@MDT?@@!NcjmZz-|C|7jr9Nd$h~Kv|x3!l(vS?fpSQK
zq-=}rhYsfC$%|)Z)VD=1VjgLzR>dV0x0w8#_E&3XXP2JCTJ|HJt>F!lsmLt<SBNV=
zSJEY5IM$ljrq0h8-%(H<%Ntra-A#|?P{vDo!NTxfu<=B{eu)jt*1N(J))QAa;D%q`
zT-mbZ&=1=%G3|C^7<j;}+3_9umHGoEhg!c|ik9$Z36@s5x89N+_4{h+2+P=OW<-WX
z2+2tlgO}Q;(kx;?W_}2pL?rGr8Ay1UoRJU6nYOC|G+h@WJnL=~{j?^CF21ABP2h3R
z5UZCi+B^eD2bS%iTqO-fnnGwnPe803AVSB&6S&z^6T{8NGlopS$|?tx9<5|tN*wHW
zVP!Q8OBTJvk})|@7Y_5yi!U2YZU9&{Zw+&L9terM?btxs*%~}`xfxl2eJEVJ0$H3j
zvHF`ClK<u58ufOgV3+-M7Rl4mc&g9Px`$k0e({`HoV)Bn4K=l8{+x#k=i#o-r-J%+
zj2TB$za9>=f9c>M7Z+*V5DD{%z`b5)He@>Z4K;-YNI^=xj;nW@5>#{+Gl8eqpny3i
z?2;Qqa%2&gNKSEu*o=fF>mN>qfzF!YT(MH0OqAdTBHRPTAN=dqk|k|8iR$CSiJqJm
z^G>OSjTTHwoR>f>B9O2i%D8fHls?iACDjv1fpq>DNQA`i4y{J`5wkxZS&M~BQDIrz
zy&x`=g`Q)kTo*jIAA~q@^j8{=iT-kNMv-$XEqTjxLzEpFQ2Xbo#4eey$By0=0om$T
zG;Zc|7iBOe0;85mCW09KMc#AbKumW25Wniq=-JY#_l1LT+SvA)>t>(@MirL1bo9}8
zDN49S#J-h6yBZo0*G=zbF0G~Ll27iK-%WB|(Z@}mr0h*wH%$_gGWgwDk3aK}FxGVs
z^Ju|v8r6(;g_+Lp;P#KBO~6x<+eHHL>P29$4Ip37qtrXJ8FXsiN7@UOBCsM)SlHRD
zcU*NG{cw@Rwr|pPmHu;8@Y^HL9Z&9PeFUc-ZvEo!`b=cuLN%@WD+7K_#Ml;8#=Oty
zLY4pAI{?%pWi0RP`mNU4r)DXyk(~i$utFHN7f8K53y(vx_n32_ve-Wg9muJEpWP8N
z8?yK1RDPHSH{6WVOnabNtQM~t-$aym)Mf!3^*250Tcsp3^4A*2@YsnAEbnOud!}FQ
z>^S<ROpQrX*)+8@dpQ@eDLpL|0eV5}GDo)gR)vfs^QyGZwsKq+pPtobIX=m2O<w|`
zs?q`**{Q&UEA-JWEuzM=+iV=8&?bjP-ny!ZD#tkHZI8uo;-*h%3B!_ypa`AH@Kl$i
zpt}Mw5AP@()Y@fN#(vVp)L3}U`J6YppQfikP{2(Tcf}vzvj`T4z}<)oYXwN?Fn}RF
zbmsSwkZAEGaxBeqify%nA3vu0+d^(y^e)b)*#KC7D$!egw|dh(jwiHzXj7DQ@eX|}
zBj{SV(*b0QWJ|0#TUATA)NE!HI60_kCwq!)9u!`e8Nkzf@I`GDJ@iolYP;$h$$mQr
z-lV)7kMP5uoM;GH%(Ttf(KooWup{Uw*2y1GCi_uSBH_uJ%6-PeEGwt|qp70A)z%OJ
ziGcvhmucrZaH99ZmP`TN-TCQq#icdp1|vOZ<xjo8lN^2U?om_D->5O$Z4MkT_njSb
zDI2dIvX&Z_Rwy$#nzGQ8&onVJ?;U8M3^HDv9SiX2lwvw>|9x7w)^b=De+WNUAD?XL
zlPN9A))Y7wQ$GuRpLvcW*H#hY47<0E@<3!pr*tHCE)NqsVwmFh&|vWro0>$e<XPDm
zwb4CJgMe&;1a4P;ZnxUIF}ZQYi0fPo`vN%9jAH3!OWg!#Lu2_hJj#DWexdO5utR0I
zT+KHXQIQbCMqaYw_Rn8|p*MhP+Yo=}Wnmo0*^ruoYzK3Z>V%a6H8Z};ZHt8_`-k|>
z?GLlkM6LOXs#1PU+7JVMs^NnExlo9F%ihn(cjJ{8TSEn)#@4i!ut>BT%JFcu8$!^e
z>9`F1aM_b;i{BIC?Pm-8DCG2}{s=_0=*If-N6EY?ck8znyJgFJj}*DSOe+kHrxA?r
zX_cyxTHg7RZTF0u8ke*@=mjqO13>H-+I3OPxH-Y3c!0T<lv%?#SFlFoq%%m%qSH6h
za2Sv3z%QI}HNw^M`Kz<JpPE_yUvE#wAv4di#uV7LYz9=mZnxamQr4)%VtKm_{y@iu
z+L#2B&ldL?qs8bAb3X%z7-y9M6R?Wysl}74*5!SEI~W#|DLA?Iwezc{-Mpmijl&10
zckdveK)uEk-2@FpZJHL}K0kjw2bUIW^J<sirswR8VeVT=AUnV12%j&{yxHHGRsX&_
z)1!8(+`h(u-dmK&8M<n39pAuL*vIo5r%dSByfVF)NxN78ds0AXO1r6=%Z}bpOrWMz
z)XRnNoColLaHG$Z*BKVRbfn|2a|3wY2)m}4+|0OJkA(L2$W!Z9wTz9y1vy_uGDnDO
zti>*H)s<UE>Int;FsPA~^FE}><QVRQK6=n7_12n)Ek)i@b5gK6imS283v84c2%eQB
zD}9xe6vU_e0Pi-@-KkV?)LG=19MBuuFDJXgs@ds_x93`l%Kk+JtUNh?Ox9RCSoxH5
z_}7T)T_uJ)B;~9+@3l^cqEo0$O?uzXe!CJUcNuzc19;Z35Vi>Ddq{Qk70>P0;fp?h
z`W2|fG9(PGyga^qa0eXK6>Ktb;Q7M2;MW;bvDv}^h{|s*27#qnT)5csDaM5B<9kG+
zTA^N;((0WG>+AH3>;7SjW??~{tep7!F<de5C-VgQFFq;$Ry-*%z=HCOP0VgZsGJq)
zNjc^c$3WmjcI^!y1X4C_pJ=9@7w^ba-0fSa*OoV(rop5ZHK<Rx!OjaqGB2m2qD)a)
zum!N*&T}LO)=LPilrVJWriK>AX~rIDs#~Pq{$6LqCMsIk{);*?u9>TG4rw2-eR&DO
zYmmBzv~L|NWFReK1?1R^z;FAHm?Nx_1O?Yg4kLU?8iB*&SGfaZLenf(R%<YZKBn`j
zabqp3czdk($W1DV>y!8!=M+n+)cfap27L5Hj1@aat8VPf;A4bMCSt*6M?D3BDUj;a
z6po-vSydhJ<d1-cq=WrS$JAY2J~j~`!sB(HR*Dso83vDlE!r+aTB;Exag^W}dLvE5
zSl7pG2PVmcOQ+<_*FP{_84s)_^OTKM&efEtCZ#{ImNel?nbw{_d_yxY#7-+OUat0^
zVDvGLndp?f%hSuTN`I{tdD(Uh>!ie<`RZ<(eu|^u-a=0gM2KOev1SrKmR6$;@9R7M
z)f+%4kQhW67I+605Rd~V)zz{&NAy4mTAf`t?a!soBa>}rXE)Z~tfVNos|_lvN#V7b
zQV95qG$V14&9Pz2j$m=E7l@5oyoSJ}!vo#JU1|qKXxb<by-Ua0y3QBN^tWHb?j~qY
zN%YSfsPde91X89zB-Z_&n~pY)R<#&dY@&SUPDo}~WCy?nA+_Cnglg@KdZLsHlidCq
zE0nmiX7=XyBiX|mCKmac#|3gw&V?pmmkrucTBmH4819ESc~#z;;t9mhWRt=c*?G@h
zKQ7{!0aw+#Vbt5c_wr$&!kM{QM`~kQduVH=z{hODH<;1^Mlzb^s@E)^8AiK3s%i|-
z@Qj8XEMyq${=wC&b>Pzc8>KZCi5KnoC#T9tr2f?!sCWEV|LcG0AOjnx<8o%C1tkX@
zEx!ywsluvQpJEQ4S4^G`?1a0Vcr~PMBq<EAHs;OtawW*v*oo#@IyoORe?ao=ro)!N
zm0eFORaXS*yHeg+SU51HM1!E>-FG7y@{h$!=wjgq-ib}DhXL2Qw}WaYJzqqv+e6PT
z><gBw7fzjCm~Itd^;<uz^TAjsf{oJgkh4qj5T%QMmQnB&6;gyMTP_{*hkNrppJ76O
z?Z;p9`-}iXzAW2{At1ZQD-a5-6N;j$*>_b#6GDn|G&r0;+N#wXJSTn!|G+pQ$4iq`
z>;6+?_G!>BQT0*Jm+mFvXApFp@cKo%pOMEg))BPahjF(ZklAb#j%ay=g5;xsCOX7h
zH7%@%@T{DzhS<zvGp+kMZju>iB7Ndo4w*J_R?)j9T$^4aVzu`-qg!sFRhq^nhDm=r
zBO?5zpy1$VNSe#!tN;(%^-4vF^dXmW?nbWnEqzI!f7UHk6)+EvztnF!a`*6SMxSn=
z$9%o^Jn(55dPr62LF?HT3AIC+3Si3R%xHfxq`SN`=a|1(7U|sF021qkqjC^6-cfL`
zdp;}PcV;3b)F(DX6T6G7<$8*`q=@9Bi`U)DGBa)b8yD%{Kh?sZ*Ol0V{29gNLQ)~}
z^R@2>*~sMsInwXFt1<WCKuSB!{DA^<P4WBv)3W^)v*FVZ%+~OI6#0*01{xEpm-N|+
ze9u~V$M1+IATr?obKQJJoWE*b+=r(s5-}wB*WR@v&}-)gOji7;8F#j<sj7{H&p!MZ
zN<$PW4`3k~Na<x`3Dzv1SVIz|u%$!ZC3~gB_%VcTcn3t#xHeMkwiQ0t;Cf9q2?f5a
zZ;GdLW;<&)l1ixq6TPSrtJ;l^b>D5eB}RJuAq4Xf-bC%UDohP-M3ATa#85O$$r%(_
z>wfg6w@cXiNM`FW;O7@Bmi~^<C&4Qq{X)><*OcREQvd9DgGJ!bg?Zm)roA$BYf7dW
zo<x3K+K`v4JM8v0@j(W|ngFNwV{g76F>PG0cbz;Xyjq?ewsvdA=pc0BFyeN1+Ff}f
zt7D|J%O}d-VqJBS64i-=tDNqg;wcq~a$wWLv{7Vl7ds2;4m&eSf1u~wcwsSBD%5<c
ztF}f`ZGU>3DP?A}HtAmIBZlyC-xZN|o3+I15_P)VY<Stg#$G``*3s!z8s6e;a0+b&
zZqH_F&g{@Jg)E#xp?l3klhIM+^CJCaXZt_PO6~=Jej8hg)6Vh!x2hG^C_c<DS}AEI
zn*nF4H6gjb+tiKjof2gn-IiQy;;1cTPLXEsl2`R_&ILDivyoqC%e_L>OE`WL?wDHS
z9d`y%GMlM8a(6Fmp2^20+C0A-N&W7pver8OK&e}Z4{c)fp&1-@Skgt0;jEpgrli}J
z@wA|VLH9sJ&=mg>2fAL286rH9^RCS=AFjK)A;czXXX~!}kr0@Py=~>QnHSE5&^gTm
zy&C|(*F}C@vmGqCNpTex&PM)m)?Pt2v$xh@kt(yh?_qfbIj-@L!{W<(yzyGigYE-N
z7wK!v?iE&dc){3HHArAw%RQ6d4S@x<T1fLoiY!FQ7w8*TRJ=)bhxzNIi=^3~QeWKL
zK@2tVAsmIZ+*kxBw;rMaeA0I!qNmL^UVz5C0laKB-Vd<<&p00bL+8UkLmU4)?st_&
z71f|BwW5Z4c9`>zng(6El2>Kr+P}K&tk+0c`uo-N!3=%#^9V9mVnj?AdYxI4D+`Tm
ziCAYHtrV;JrSAMH271lc;f#CD0HpXn%iRCwT(aO$K5g7xY}U;}Pv1upu@NiUBkqms
z!I#Psq)l+v`$BgM!!-u7zaz`4*MFu#J)}}wr$Ksudg&eDNOFkOcI+jvOYCSf)}=`i
z%UJ5dD#xenY$%64kVcs}jK%I;)B@$+(>X-2DEm$3fl`QJn^A^iVBYkzN7fow+w$1u
zHc$rh8bfyjC_6tHKd3Ff3?1&2QOL}yiE_-Hp0gWFP@DCa{F%INny_t&>%vfZ=i#r_
zjOvtZwD6&8Xp`*q`&_W5(pZ&nmEh8uS#mm=WQ%)4!@xRU$RaCMu;tz2jPC~YE0l4;
zN)}A<*`)D^*4y?>bgCPp9%lzX0<ZR2(-r2<1DzKwbQ|M<C~ssz4;y2}lyLPh<qg0d
zB?@HSmf^>`j8^4O%w3?|RcvNi5$``pf^#;ktpR({#nBT6V){F-*ELI*dX@=QH0Zta
zd@P~hH>+~HIol-kRMT&d@L5yvtN8m9_v$KR;72M%g-0Li;y<f1;8#3W?N!vNJPZvk
zuu%80&$frx##eXdbxFqOMov+1)o}#<6m2L0o`>VDMINRP&lYqp64%sJ)(7w?w@g%D
zdyPGivslTox1sAnW+y9yM5WDk`_|jSl|uRIUP1#n&%P%1UP^tW>^z{HMrE2H0TWh7
zE)sdQ<+ZUry7+^c4zUlSBdx)3(I1}>ova@cChxaGf&78S9}}b80+tXU3?yqigBInA
z5Wa==JW!IOOjHjKC#;TLAr)09=avaHddx|;>Bz_FRKii~&^7GDYizbtY23*~v&NC1
zpr$HT%-&_rTbT}UG_w1wc$qb-_g2$DJC<qZR)!g2$`3BFr@NF2<2+1GFD6%ToL+CN
zYE1L{{BF@y)i9AtIU@CO1f#pv>y`oSA?)#VD6_@p5%p-@W#9wgq(4>-%Kc{qv_J0u
zyCk$*MccE@$sq+p0flC&wzfnIT;r0H#<k;#7Bv3vP<bFEbN8J&lV`okm7jX;$*lH1
z2M0em6`UIh&~_`*ePJNi)XB<1E@T7asgU>XTB@#zDVUd&xu57(g8%4B<tQLSY(Iyr
zD?kaZpMLea#1y~vu<r@+`d-<&2o^T{%!uv=`$NdlCie(+NFB9HMHxLW!hxtg8wU*A
zq(AW9{i1wx`!aO#;$tM;B>}EP$Ih#HRNux-r2pQGul)tPCyIW>HVkubwyHsX<06MK
zdF)MvT}#Q<J~7+rq`_wF`%;Gt+6h4ksHBFoO{mVoJg=-K#nX(g*GKZ?0p4y%W?iII
zhaZs6j^dMP+W_}o*+?gmn}Gij!8?yBS9rs3_*t6#aAE&3u&h8kyJ~7B?nQQyT|=(S
zMxt$@N#^{5N(HQ(^U-BwqZv=j-#mZjCRa-SFCn4k6c*(TF^;-6AI7~E3~@J8SkKlh
z7eYMW93FSSieYjzVnIc#h*aen=Zs=E4K(Pmwo`*?_Hz9IBqDZ3D#E6i5fX7%c*P?&
zXew?Ya5m9r=nO*qJ*+m}Uz0tMbEPF<qzMQGADh$pIyYUG6}cO$F;@)wR*F4rJLfvj
zL)r#|GI6)Vw5J5P#t$UJsOrfWwcqvkKY8=Xhwuk%Tx5pAOl_h>!DserZeL^XJTrAg
zY`3R+d(`w(bCto>oXK@BoQ?Dv+VpUR`O98)lo481>?^UAn>gT-9zb=<ANRt1m27mr
zBcb!&08Jz#ml8?2+y|m<{^L6>;nztmLZ({{6Xp~Xo1*qpek~niEMKg5%Gvj8<sg0x
z+m8K1b-v%2v|Fh2p{@n-+v{@5deB$O^iwKlQ-hbYk-V@V1@zx`h{S^XK@WDAhT_6b
zW=_eUVzB~VwJo0155;lEvrWw17B!L)1*UQ`ba8H_f1Y(y(P)2Ey$6!N`-RDzwtCpE
zmEGPwwF~RJ%Yy-$UMCEv(i?l2JE(@8j+hn0D%r5nS7iCiw7=RXJ8a#|uY!7{VIzmq
z6ii9HlY?v(Gv)*1jkcR#cCV=QON*>`p6u^=@62${pA64w87~XvAoM;_4<^^9r$Uwm
zg4hQQYSBPc5n?#P%ULMOuf?us%V~p|Cu+O7w~otiYBRtd<<Vsv_VKjD%s{i=Pjk+e
zBH;nLowfb9;x_+i&%4DtrR<YaT+JpHvP*|@yX<$~ZfOO)Iqn_hIjU}@J}HY_Z(&4j
zQ-&{+j#vMDWtm_-C&%Zp{<KNvTI}c9y@oV&OP;iDdVz#ZmF+gowptbK6WM}d65i}e
zaqLIXKMD{3=y?7C;TwQ$^Zlbvga50&^9*aMPuD$)ihzhT=|m}_bg5FKC|y8$2}J~H
zp$JHaL`8~J0f83~kS;ZJB!nPMs(^Gt2>~fm6Ql$Z@T_;vHM3{tynANPK6~%;<$T}@
zS!;cP)t={m?)&%umQT3Q;)0{ISqzQ=y50`aW()n*rX-vkN>s_KXJ9!=-k5Ha`Q{yj
zQC@r_JzuYq5lPk=c9U?q@MBi6?Z{@m@D|tz_3io5f{y%lc?RXV;^(5*j@<K9qZcJv
z75(i>TXHrh3x6Dkf3FS5@8W29DcdObhO<452}64o9i04}jpgMHq1r3`oVv@)##J6#
z2MV{`q0a$}+K~mMKBa+YeaMqY2h%;-X<EpkKfSQdfq3oit8BSyL4+@{Ov+T4jx0UQ
z7=-k525(H?klppviE)t=Ofq+W#(vMVFQT&1g2(_ZbvflEfn$|SiUPv!L0&fyq~+cY
zebh{2lcN^LTu-`cCKTzJdX{ape~OuAZ~Yb;JGqZ-HBff=e&`PmZfYmOaoh<JWG9zJ
z9bKqp(mFRIH%)-JSn3bY>)~^YsV)NQ>4m1Lw!djodMbuJwnyeg`@<(II);wFL3RNJ
z*YHn~6hOuON1tco{%jontI!eEb(xCE*nLsQ>3!h}#fP=Yn=svU<0RADWl~aCG})d}
zu>Hv1|G63o17QLL50XZtjlKL`xGJP!j?|z}F03jGO9XaAfeeD?tB;otykOm*L#0sK
z&U*18q0iV(C%=T9TkaAUjD^q;vy`RZ-pX=qa(`+WIVMt><+11!xz@9ky}*=E=ws?-
z)wmsU;BFWvr!^jr>bL2YffMbwkw1sxES0kx^VVb=Cw_RH%h-y2SaCA>j_*>ux=Ga1
zMAHt@b?7_*s1XHR#b?q07*{1|5wUf>kQk;rdL6%C*kOgic7!;fV}&T0JWM`6(KEla
zdxwWCKOQMOHp=?j<7HKyEHE#2%f1NX!a@IfbJIPqu0C38;*Q<jsC?>S8t_w#_!68x
zyYo**cevk-zW#hSZ{rU(a#<L$ivUbUio!l1rV6UY+`hk?)yVpm)%MH)d7#A5-D<f9
zdL`=)NUo>X3EaaH{U5hUcRb`M3i-e>wJ=JV*S0dyZu{<z!8@^7%p?D}*2@r>lRw_S
zOUo1gi$d=eL|4`Eb<ZnJk?)C`^PP6@>|>Ykm7}MO8}MVQbu3W~@L|3w1=H;L52McH
zTlO?Rl{gxte`J-`cI-ikudHkys+VZRi2(xJ2N=+TzxvTi%mDx0cfe8nA2~*{TuJS6
zu6LDaQ7)|nzf9Nr#A3>3hL^Siuqlprjkb)NTkM&=^`lEuEsau5_C#_%9#K=aEIuRL
z?xR-WjOLG*g}UUDE)7s!T-PZIosxoMS@y{CH@hVg<czo0LxvdCz5AE@zm*;HSm4bT
zUy*MT!dz>nIX5h}FcZ$h;m-<F8gD7L-D}(U>IRG&!hH!#Uoh6_xB^{<IF){=$COKf
zqg%DxGpfeN)Pf_CZCTLIr>qm;1|<vb<^8YClrB!^k*Ir-D)vbo{5)Z7+@ZHOkxtzS
zuaV#wak7?Efvd#&;YUW7)N%O213KdQPqP|Up*{xzy}?|NPQwmuAA_1SU=-Nldjw()
zOK4~pQMP($PbK_mA9yzF=Op6b%xDHEh{n?`cg#ZUar0{hYgPI*)ETr&a0xi7ZRx}J
zJ!@@iPV?&4m(jXVC7*opM}2LP>%u>P34Tr;RflKPJWKO8&xP?62hQC0me-x1f_%G&
zX7<m;@BiHC5U18zvIRLGKM)$}7VWDW$lq=+7MsW31jh^4Qbt!8(cZGp?OO_-vyj1(
z2Os)fjZ-qVf0-khY+Bo_-*^Pd<wz~cu+rh{@VVoNxr({)F2S3bi)Ij*VIr}oFOM5f
zX@Syu5$YFBb~t$lxB~xXv-9752M(k7&m034Uo`!6&`LwrD6xF=;hoDU8@Xq|n~z24
z@@P++eC3y#5Q|*#H)QcLm&-Hf_sbhdtw;+^D!iaf3$?n8-L1`x)_FS{$&_1PIO~9P
zUdH~OdJ4%zt(Yttb5DUkp$sw=ry}2hnC_vjwu*21JUII}am){|RW~_WA6YdPHr0D>
zh4Y1x&x8ARXsg)7A$CNr^v)aG3vMC6L*wH64AT;yC*QS1cHCaM+@L)1SU95l5l^vH
z%yPW#LStgD4;Voz_I|AXqM$4}0-oS?g+@s=#f9}7%RnAAcg5xmk#i9o1>+o-DZBXE
zZ{j0tu!b_T_slZWn$*z4g`mXM8b0;S_(E<ynyNOc;mIka)0P2#M}2f_)6K135u*5<
zq3e=g6btin%2h-TOg12<-A0^nZ-OkA=C|XY+S2&xwL02~yk~WrG}rxcC}`+R+?R<?
zy*n9V67b8JfneK#QP8V4gN}A{*FGcLtIbCU{hkBQsYb(8YfzW$`8N^N3t47WE-d4%
znv=VTf{k{#gFf}wJmTd;(@$~vp_U8M_ada2AD1>RcKxEbWY%o9YB83t)Ls@bNjB6v
z6<li)ak_*~%03{^K=Huq9(8}_AHDQ{-}k2%|G!mT|J*V3*T?3E9s*WB;I%BY4o&^D
z8&R6V>^0esNF^9k$xon<G`LW1c5-9jZQd%^1;YUP_@;%EAW}UG)2obHd9OQ`0Z5M&
zliZ`PB`*FM%HLDd4r(pt)Sux8!?|G}#w|}+#-)|ztYS4vTn>=0q2I(lSllJEaTFH^
z2zDTqKWBH`8x<*{;?)Vhvt&mMuLPGiuMH)oUGC4ED=E(1&1tutCYcf{zO!PwP#Bw6
z1&Y!i3K}U%AL*}wc6j8MWcF2R7ee{0M(R#vgS}I)N~+^$+PfZ|W1}(4wem{P*T&`{
zOGb?+B1aaUF;__}M^tQg{`kCL_)X0<==T}ypZYJr@`sPlTFSXR4>Soj?uC5|DY(Q)
zq4i)1rC${LWv<bm^o+k3Q#nSxw>K1HG<!&7TJ&$DXCO<VQYuBWe^F>s9R_6~nb^4-
z6Bbmu6I7u?!h3r!KEIL_&O#EsaKh6W&;;7G+JV`0fl-_z_>CnACC8#IVbE#RyxrPE
ze5g<J#b~8ttY-5#WN4_BE<oR$5RuoQAeN<)?~U!C`F?ZXaV>xLt?(B8OYpgCM6;q^
z{~*E~F=FfNS4r=xVp~!AqPzTu*MiK7&0QKYOg=T%)JF`S>4~0~hz3*KHih;M{9K<R
z2H(_3<qmE&n&_A0xI;~{K)dX1wwwiO_B53be|<$9kq%y+WGxd?2D16^BiANq0A!H`
zBeT`gi*@}ogh?fk{p9yKBkh)No&~jIqBm-qV{y1ovH6HGjuWX;pbRYBAY80hmK`;C
z6C^N8dlM8!X=|P4#bU)Rqh1%d*s$F?aKzOe>0&RWF!r=pu+`~QV<j$FybU|uZ%!EB
zZ#k{^lugDhU!N{COqxyD&da_Fo>%hu0^Mkl`;6xJ=5g*A9$8mL-PV(Y>AN`v6Uu9Z
z*0Ogol~7%hmWQm*xAG-Wwd0FrlpY=>tu-9IHJ_UKk!pJ2_qMR50tIC@P?z{vD)Bo|
zDLYyim+3y2<DI141N`~-|DW7teSmW`v^-$6o1+C?NDy3f&ze;c8LL}+{#CfmKr2t=
zimeW_fec}U{}I~qroWv0xdsb;{nD}|A~04angyv8+a2s|(Kl=;#3NP$RZNHKt^2)}
zm%Ijtq&&8zVi%m4mkKILu(DHX!`gzyMkJ|1e4;yP7B*~p42qN9Xm0U0NN8(lm={Si
zRY)ngpf9D5ZU}$ygn8CLZlL<*Ph`!`EO{a=7VKQ1KG{(_g$Xb-z(>Z53j!LrG=PDu
zsEhjH@6S>S3_0?wv=LnHVwO_8ntm@-sZ2|c5t}JR#OV3mezQpj<ftbfLByzyB_p3@
ze22HGM`WPsbtmhJso9_=rH~Iwki_YtV3S#X?<2fjpkQvU&6A~m2Z-Waup|aS0`0QI
zc$=KW6?=vHnIE!xckJTw-0<r2i>dvv9aD^rUrj4np)}5#q_RDq_N^W>C$~3!cohSV
zQ(^h4QeQ$`G9`xLqTGry4Qt2hnjOdN0*33{pQK)`4$O>Sb}lfpe<!q2pQ@+W0dJ&w
z8KRGCWmuocfE6oYm@|TaX4rQ=vs4cVCp29hW`=V1E)J!?wRgFC_k!Q!N<WkaUhWH)
zu};}un=siKC06cHSC*wPKhu89_v6`4x-9#45~M7lE(PSLQ~;OSHWkngT<wCtXuZOz
z53nnDl#J{b!Hll9__!ZeQ_&ziDz(bI3tHa&%xPD5Z0nkVaPzCr=VRZ0T$`wk4ptP`
zwOwjBw%EzNt#yoDK7iBOx_h7o%Q^w)!i_SgMKF~!68=J_ncI@o^~l+BBDp*?W=r<1
zj+0EMk2+zV?Kg1R+rHgP_Z8l931uIE6Kp_9$}C;kg>jHBRv2U0p#izkZ{@q@)=vTh
ziT^h~(Vt?*4oh!aPTrI6RX|NugvLXhif4(7Seu&hxJ?eGuHvJs6F+APOD!{t67NCE
zBIlj#S5&;}&K1Ku<Tk+g-&|OAmW0j8!Ue>d)vksN1?TQ9QRQ~4H~Fb+vse08tQD`D
zzx{ZmcY1LM0s)F;dhcHpB>2#>-%;T6d2GTC*hW8XAm{E$$tKyNV9m@D)Um_n+K61v
z9jluYx?Qa1<MpJBhlJ`6NbKf7mGJQU`q{rzziZQC+s-30it@rWbNgte6jn7GKI#eQ
zwlFWtzqM>(VWFu^lW0SSfPF#wmEb^;rWb5!p&6}Vd+<%Q%r7>~x{mGZXZiY1Qi|j9
z-brhljMz{rRE4}kkK-plPhjan=n9A)qnl{mST*?YbE3$#*#Qewa{!Bdod50im+$Qb
zGDUntkTb0~f6w;GpRKBb1X45H7Vo|cU4Oo3pmz6s;N%w&^M<8;fLRBO3zeWf5n@u(
zkjYaf`YC7gwg769>0x3u&4}CUQjgi(Ig6BW?Hskr`-^5BKbN+o^kp*&GOo51*ofs9
ztXHchhh{7rjIPFgN|Mv!j~f-zRy^<ErcgZkU1svz$-(F`nWsCggP=|S8pca_8ec(h
z?O0T92+N&=a-_xt`xJ}6DYrRip&36va+3iU5bwpyC+w~`?Qn*d_>$y8w8gn8JVy)T
zyQ7rA+{?{?=|oJr`5`3YngTWA(QLH=s>6L{;N@E-=fUU+B?u5=8(p-w)m;Kb1-TE@
ztO2U1yJud2dwGQ&WGYoQD^;hdAdvav(pQsfXX?1BZKL6}kx?k;5ED>jh^HNH2F?=T
zDo*G61%NXuyzXa)DwNG=iapt_5d(Q6dkS8tk0$BVb7Hn&ihw28c6DSVtbT7CJ33d7
zgi9TTo;=K{(~mUqm4?t+>X1MT7@`wN`AL+%a##C(e#tMFUa&osqV{LJXdF^$<}L-;
zHS2DqFX?)$(28FoMuX|qyn>z4LP^3-qn1$hr&*x>tl}3D@pmr~)Ttj*raqeW3&W#a
zZR^ixH!%n0x!5q__pX{5+L&v(yM&f(B`F4eaC#(g8>0KXl8|a}5H;!LaNQl~GHS?4
zWK)&f6Z`HL_S>52f(waHrRSiJ;+-uS=uY;G+?(F;dA+`(ks5t<{ry-9n(nz@Yf2LG
zr%P|3mfs{>!(Lj-FACYT!jb)lYbZI=eM~^i!%Q<CDnemSBvv=gbkx^>WMom)aE{{f
zaJMIat@?I(dEyY4-U7K)yNbt0ZY!oBL>q5YE%?eQS$pz%7UW*^GSA_|vB+tJ%pLcc
z!&HB{N3VKUubVzi^yYN#tM`3$Is_9&5q&;}fG1#l71*#}6jepQ74BBY|7!C=Pj#2*
z_aTv(XN0_$QA;mDK0AgNX>+_tQUUp$^StBxLoCD{9W=V7@tY_wfYV+wpPA_@cD41O
z;G65~?9o!y!TG$s$mzP&zhzHugdFI$gv=#D-F4!7`{AOEX4t~YN$bm_nkOIX;_PiA
z=9Ot7ksXZ+F>41fk#y0BOy8g+Y4vjkw?9!0Xz^Xfm%H_LWc5JCpsV0Zr(cf6!(n_H
z2ulc2oFvO_qwC0?*(QrXys=SF>vuPpt7S_ZWV<VHktJj{Zgs715Pa4XPNdy$-7iZc
z)B%aE$P-cGF)!doumeg#s*Yk4g_VS*BWY-B3`b;F6iUAW;;jD^95th+wKR4^=Yf4Q
zN#gEH`^%Mn=S?q3CEBzt4PlaYaC57%Ge-E+l*^I0EY*Is&?mnrtgpH#@6TSXQMI<b
zU@kf~n<HuZ_MLHG!Rsxa3Ou$02@2oKG9A4))>^l<GT}=*lYWWN_XL@}qE#GolfpVl
zv%ik~nSxp{W2J~_gadKS1Yn@t9xIF1J{Hjj!6%ZtH<BNYl`ln{laMT-=GS!%PrCby
zLc;f%@w;4QUZe$ltFf`~#BHt;D#&j9zPSmqGGhj~sG?UXad6C$k}p3Dgs+bbM;R@Z
zh>v`A-veKz{mNv2Pqgjb`YkuBGhAXcq3>JMU&a*=R4U@Sx2XIV=A;b9l|1drY}c<I
z&P$^LDtt>kOy?Dqwb>>H@65OX#O1sP+f4n&pD%L1ILJi25PVeInEDg2Od^!@p+Vwq
zWCM|dC)?j%>oRD5xY|Cz2VAjqO>@mYjD4{vfQl;xjCW|)E`S)ILD{)Q{U;SJJ}S*0
z)j*gWGi3M*Jukn5rlaPBa=51W%jCwhBksQ_d=Z)26Gw<q;|DfxbXrbLlO}CtYApkQ
zYh_1^hS=kC+M7=_1EkdOgN;1-`2zKGjZxwI1&LO+U6-nAm6MP<5j5B)2Wqjs#4K4e
z(4wyGgK-s+5J)}2WT*=oXWDl#!Yw)4#;S57Soo>C(1^Y$U4*DfY~>bdxL|vB%E+d=
zGK6+ij4b7k!(Jup5=^UAnnbfIa{|k3E!3o$EPFh1A?Egt@l@}m)m4X><2TqZS(s4{
zqxcp$!Bp#~{9_9!R7(rt{Fq!?jrDK8C?+(ZbX7tO%oK*_pMX&iHZ}^a;K_|(l~QuE
zN}R>9t<GE;yLUxBSpmiYCUWxV6p$^6IfcZaQFTBsxUWM1$Al>2%IGw@yj?k_NA)p<
zZ~E`MRbd8`dVk{=6y<pOR``k@W*&4>nMZKy{3_m&;gOJk|LI4W+{Q>V0+%1h++0f+
za<I`al~I*(W$ykJOU1b9=toy-+?7`o(C0grtl<XZ<stTqwp9A1Q#Ue8yFb+Dq<vJ)
zJfI?})pfBysJR&s>mKnvtE{0Qr?x2lMqyDdLH`Y;JtK6v@T4_tC41goLB67ajRy5r
zW<*HI;8UES@ocOA`Iq%#uJ1=pbx$jS@wp||r(sn(Z!FO68@Ly;8a#jOYaP$p(b_N>
z0T5-POflD*@pVDI+x0hILeIf&MeD-1bfi>y+{ODY+_8M%CXt>4RUe3EZwM8%amgjq
zM|Pn3ScBlPAqG&~a%-|-Li@AXHl8f5T`A4)3hiPK9X0%SAFha0bIRg03db}S%&@0t
zPds<4^IH`guz(s^i1z<SEx~_B(`}%}Oz|)NoY|cZuY?+8+S0BalTVp=1c+}Cm|3Xl
zvzdM$#!e3>230Y>q-(ouFZzMk>&=i$xU_y4V5IwU9NfLoN{0=|F^d3V+oEGm6J%Z_
z$GX`_Tj`NK2=lO^`Sz{AbD9vZgfpzyPD(1s^7x!Lb+mJ|=6G~m=~`NCW?6+SNk0PW
zvLcS^A=IQeiU0Jt_WtKYt*`hi_~reCOC7j=KpYNuLh2Zv<@{8anZ+}&qq!<SP{yG8
za&YD7#n3<~RKe?+fqsE;==12-)an*w+ASZSz~25*3zt=b$G)`!NOB#^rgAeNx&2gh
zBDo8JV<$^3Ff8g<xk$LA_4)>Le<@^V?2mc9JLk@a<uZI~Zs%ZKfz;*fL|?<VM#1EW
z#kvIULyRXGg1e)JIyk6cHki{WEXchZ>3>{;x!#mXQwlq^I<cC+fS<w4LgXU%8XB5L
zC$+Mfo?h*BdnW;L<a{>#lJ6z^k@~?>vdMQM6vtjK9H**Jvxb0saiff^WwWSLAH3|C
zDU+Q|<SeZ@3p4x>%K%X;G|<}mMKNQ2@<YHQCw-u9wEk3E_@^Pvf9%K;=N}cKgP-gm
z#DT5GWNJ&v<^=AnuKKx_YL~?aW6Qf~{lH^R6y>40>Zgj+!{p;^<8!Vel=uoZlB|b{
zdKX(RFK*NRqTsR(@SA*4+e`vyvQDY=Dq2gcs$Pi+PfyX7x?&O*>JG&pwl^pg9H%r!
zJP!_455{iY^5m_0<CwJ5?~EQEWW8Z4IxUyn3}txFC={QLhkp(dE3R0#^nVqiX!mjB
z4S!I@TxGH+7xGn)5AWdmi$1m_O>b{~nIv7K?;Cw-h54Rc8_U!21x>3%+%ldMKfTFW
zDAMLLP;G0BlFmoK9S__=F`g*n<G~!3tm#p*ty7`sT%Aw(3&~j*_k+MZ$MAy^wG{2!
zZe>Nv4Ncf{zh7S+)(Y<6b{@8GC}BHkbVvj8Mh?q+L#-Fh)AlCkt0;O`Da_O$T1DHj
zPu`pS2IP#+|D&7{+q-{4Q2MvI@Be8{{=-=AUuz`)UB7q!lVQ@oJZL(j#_}D+2~$+V
z@0<!KXm|S~G&6r|y^SeZGxxWjhBHmEzkA0rJrvbx)IRDf)XX2W@$cb9%~prrDA`vK
zkv}SB;`CM@dHbJF83ZIuU)XxW<MWQw^Hwg0)XL?%5WPrYeldf{;pnB~2_)b0(@w9I
zLl>8gUlbH4+<9nYq7F&hRY7-?!zp^3^U<io^{c0Pj$v~{1VkmwZf69YI-z7xxF(z0
zXg2wVu4U{Ev*d2yz??jnJ=bc{d0(TCrNa4Z_<(iuDIcRB$n`kOn|XDL?SC<p3WGa2
z_7lSQpe<%bpCAxUV@Vx8Kc29$<j>r_XG?E;yWTc)OV?79*P7;2{g81yKX;&`WYB6|
zxxfK<q~Tk)7WFyjzIql~>3&#iYQOJ6a(o*E49noi%ktgI&>n}xmBg&+tf}SncrBG8
z^bB45c<DXu7Cke}Bx&Eagm5SOR=%Ex^sS#_hmCXjVEvnjLF{?5AwlDXhkdffcyR=a
z5=OIY!E4Q43C0I#)4;+T7fx@x`~oQ7W{C`L`@xUS&!uiIDRV17I(>`}W?FW%;0{5n
zjF5-7kwN+6{XTu{uUoR3oTm5poN~XZvHXnu_BXD`XJ7`jN{Ad$y&cR>T)l^%{q6yA
zS3SM6Y8U(cf!@#V58HrGPK{&&!|>cDmJn!7iMn`2Qx2ZAg5K`|3EifvVo44Wjdgsv
z=dM`mE%y1Pn9R;I;c}+SJ|A--0zdxPT8y)Cl(bq`*vL#5^!6U;!-{DQxg96l9M~o;
z`dlPw23KzbzM-}Osd8CcF)h-_W~<R&SqG@W|8pywE5p<eTb2-Unc29SJC0+F7yOjZ
z9))|uQZ}o5lGSU}`_?{7*1Ug3uSy5I4+Wor@jx4j=9k)zAKrvV4^F$khkwq;z%3CA
z;HVr%xZTWDeZoc8w0UWw8Lg8M-lIg`$Av-O#oZQkoh$o-VwGLb`j=-1!qikIGW?A|
zsqfVvw#X$W=9;W0x)qB4d2l&;8O9o-i$DCLPvMM~#8>RDD_liTWY61mT-J2(B8!!?
zTLeocTAPG4?f*$6{<B#8?|fbm_^S}~KYB3z(Y`+nLQ2$QSs~85y-;RW%lHq~fvpZT
z;f0UoXFk0QKNl}$%kl%pJ&xuLd88|4lvp97tG$I;(R#F^0uU0#tEecaXpV{hmzb(b
z*KdZ)&Fditp6Xei4y8I<0v?C1tG@q94&`ytCs3@}#Iau@?r6I>e*WS<i7HUwacODo
zg(q5@6oP9vvu=8~&~`ojO-5Dsp19PCcr=*m1`(3$Tsw~yoyZN@rM7<!Tfdg8Ns(<%
zo(niOD+}Qpqwn;5iA0C{lQvrTCGU18*NG9E4z8!I34t)X4xdIsg!%Pq)OYpz*1g$S
z63+1~mYwzqxhW|wDXY81+~s9gR(X(;x%agEag(rjw|!5CQ9P^cf%NmZ)`OW-r+V`3
ziJVRz%5OUY?z_K*6niYceaC;|w@RqYL)HFke5xIlvAMje*1w|<0gT(fBM(vfMgKnT
z`)m5px4$C~eW1e}`5*KU3d&>?aL!!iYART7DgxKLSQwj6g4YjQ^(p6M^V#x>vUOJ1
zGZ%y-Vsn?!S3nbJjT1!qa@b;c-8N@mj$Ty0A8z&vcqr)DgS(g}{dFMn20)NgQZszM
zd*r?E0>1zLBs+vMKjZ;{Fk0Vj@4`|xD~jxSG%N9xQ&Rp5TO$=%d;$R*vAz5p6saOo
z-8igspKiP{OCVF*LE>%aSz5+tUbEQ|P2%CFXD3yz1r*Kq-**u$h<YQ?YG~qa1`es(
zS)<}V<qy6PP7>M3i0mv&A;^4BL$WkkWImE)64=c_zE3hH{B$aI`_4g%HRa7_<=8Ok
zv=jnU?N1VpK=k1qV_lC^ILw24`<G6UFp$(L%@*u(z@Y6}vf8(Oe(#1Y$(cEE$M7E0
z<E03mbzv;vgY<kkafsmvNbZ=>IzN#~)9vHpk@+BXO@jUI_6s<c?<CPy$fY4^og=)W
zoFaWS<w>l#;lN!svaX+9^>ZmV{(=U7UB5eoJK9HjVz+~fGD3$&ErqfRi_y7^c{a0C
zDQ;;ELevL}?+5^p>2Kg<WaE>M|Al8CkC+Kw5*YfNFzw$B6CN#whT4gArDh3*u5CN{
zhV$D#pSL|dV*@0~{{~ovNNQau=b)R#R@;;lw3OFfGAhEyJCi6Z@_(XQbDSwT3-#vT
zo!RYfqd(Jy-@#QT6kR;9>1CX=l_w2(UAuIoOlVJd%5o?cLsBO494?eA42;N+Jsnk<
z#UL<jBKphQ!nKNiB}bZkN<8-v$#tYPUmaR#zds{n_Cc-sQob<7J#t)d_J06)g_yd%
zhf|N^c)mg6&(Ik<CN<|IEC}h<yN(MzI;beZ0_!{+DCs3VuweKdpx0)~elCcE1jTu7
z{>UfVw)v+jY@sY8O3Lmn&-ac!C_bmbzV+(NtU?u472@5gAXyPje7e}5#q(400XWdz
zCzg6qaM=^_MDYqVpDNRiFkY;hIiVO>xpDkc;>*)Jv2<)^AM1jTR0aTr`dlaBX?Mp?
z43<TOx?dzNU;WKu9L-gEb6&p(S*$aXpKq_n1}EF*n;eCaSRrU2ws!ss5cth~8mWoD
zuMyC5d1S}yjIsUt{RqmF1lJQ4MDI_UXZbN@dYreU_CwGG8+Nyhv$utJfu<AnbNC;%
zp1<$=?`S`#`7G=md@sjJ^mSYd!=O2c)%b%$U>%f#=<6bWjG9iPl_~LlVQ=N&fqG|l
zh7VZ|l^uwg?MU0$y7cFBHeba>`B$OOOqwMV`*OU@$BVaE)4w~KpN}s}@=Kb|0gTA)
zGlb$3I*#Jg^K!%q)0QPVpDtR)_2c@Lf~q8YNP2pE^4JUAV8okN2QBpQZU7t-+upq#
zW}pfTAU_Xt0oZunnmX(>Q>UTT8F+<D>abH<!_OO@r0=D0g@xzqn?o@rX)C%U`5UJ}
zFsC?6-7;vuUG7Xv*ZNU`<@5y+O7t0wq7x-p`xpt1JT88vLJtYB*Q|$D=o=&&9tZTy
zaB3yoP6+)#k=}hT*;>2<9>xPAx}tsGy9bukYz33vvvxnr&r2v)DVPkle(rJyDmqls
zx7%wcZd!|)&yLFdV*=q?9x@^CFDH{Qgt5oD0SIFy0lL#n>P+(y9nD1X`3&9*dz8)3
zE*@Wzm)U%7X3EN>B;Y(*jFk5>cYd<g7p|Lr)cS$A(P_ifm^z~|;raYiLFlfA`>j|h
zd-gl)P@_{;!Fc$Z+juMf7J#oqR}!;GG5~m(Wegct)(~FS<FDoA@iPD2F)EN<Wg%=M
zgV2wdn(Qi*q7AiIOXc}0lFSGCtPanDW!{(TBN(ifzI!^cvENX<;=3C=N-L(V+ovo7
zMDy1IbN@w-Uvtj+({5Q)|J$>G&Y<wBq_?Kph(nGv0{SiC@tH$J?%tg2sCw0_FJI;r
zb?z|pKWfi4VfLdiOH5x4a5hZ2bauj0Hl^;wV&iGLn2@i%EzewjsV>%bxE>(nuCA{A
zr7E*Z>q10eY`=&{z;596YMY+V`L#pX9q!V0wLGg43z&_IVBgL|Vs26r8W*-}D*&&P
z%uc&F$M#5w5AGH!)Aa3v&CIR2OT)>BFLVp+iGJB%6LMQ|-T|wp5nll&!Py)#^amJ|
zd~gS^v1L$zQK>zMqwNX;tU7OOqoq*-%SD$L3Vu=aYaTjmDN_q*C+UV5EYeSV9+VO_
zo=r4ZnuZf%)Dn1wHxxUoTc%WgBnmiiwxkV!w>U_sW~C0&&2Qab$jDok$ASAQOt*Uo
zJQF5Wj5Vp1at&x}`UgN*vv(=ym{Z1+OamuWQZC7Ly8sIznN6VVE3I8`TVA+YW%U)U
z<zRPODG~hs((#$(&*e~1`0sJ%*aC>P_(E4sgj`nKM<{ZDj#>Gh`^)j6YfG+;25sE1
zS0_2}g-kDJ6|pM9&X-n34FafZ!fBv(#4ffnu^4^t(#Oy2nxC(wm<$P+zZjp69(pFt
z7sADqqtWlsYLq~mXur`n)mwkLEMx05qMOX5u^`)sKg409RCtMlU&~Q0gbP1>sokGi
z$rL6<kdz+q7-wyF`gUC76M}A*0-aHzpuD`+T<xBDm#WZhQ%#_yPI~p1KJJf~ox7y{
zPdu=HxzqmOxSQBN6SkHSf98%*u&kK2)&Y)dIv{CR6Vs&--i6M+ktW)e5at^bBl9WV
zh&kq<pk4xJZK1P+k??*dWs_;E2pdwNny8%Vw4AlQg3@|gBp6qJpu+hI`%v~WtMRAv
zIxU%2c2T{ZpTEDqB5|wR!eYyF9G-4yTCmcHv<*r=p?3=F#_-QyuO5y)yy809A1F{?
z@%0_Pt^~^OHMoF0Qzu=$<%a9HpxXb{6`HY<h0EeOG#%fm*!wEhclA>f3YGFCf^F<+
z)?5_Lhf7J&#a70EvM&{ZzNG<v>pu+%zA1=K^S#Em^#j#MmKCKB63eIrpWM9o60%(y
z?!mm`e7sV&M8{;9aDS|>l_NmoVVcIozOs17yey#bKeNOdHDs9ntaN|WdHTRxZknaP
zDatKIphnHdTNpz^$C#DvYQc3(O!ziuHXJ7_vdU4|{cqq(6ceGg9R{A@)EX6&QPef+
zku$J>W(XW4%M4t5G4eU>S*N3#NN$F5n}5-wTxBrwe+-P$v?D-=z^)#Of@W3CCI_*{
z+DFa3Aqm5A_bo^3-jqLSZLP0KJ@$e+r!k4Rd)+qm;(DBub&%iiEjiqNY8*pU>`dd~
z0RAxx8(+o0AS+w&wOYiNMcm*{GaeBXvLSlWT4B>lHEl~?N?ibQmsS(!@G14$UB~$5
zo2TOXpco}?92FtFBcKu-4a5)P-y*~6$AX|@g`(v?AKc36K+_nRkXtmF8vS2F`sV!`
zPZNJp^qzFx!6;4sfS<?%OJjs?9-l^&e^KyUn*=ItI28hzw&??<9X5)52a{S-N^D*<
zEy48&j{8BPMoFs7Dfzs7w;<1+Xk9+!)lk)#0-y19_p~Bv;F@=?%+IV_mJdrQFC*UD
zXM=CQ*5h!t^B=Ig{F9N2)F)ddQx=E#zI7<us)G;)^9pY3Cm)~IPu3z8k8kFEI9nZJ
zw<OB?Sq0QILB97<t;nUy-8A)K(srkiden!yB*iOb9A^gySnl!DU=RYtBZ=W1G*3Y3
z(Nj;l0DoKP+!M{E!1HyYPkqi>AwLTC$7|U;K9Pt?XM6Gb^>Ii9@)+EIe8QxVKpA3%
zAL`WE=;yu>s){wLi*wA>ov4fXDujQ~8U7A3*To#i@VJkk#<-?x@E&CiA`T`I5*pTN
z*%*UuLrdZPW8C=g8c|{*G6>o!&Etqje!uEZ2_#d+65jc3x!=b5p7F&+-MJLdiqWyk
z#C;DMFvYU$i(i;`hB4|wwIe&TuaB)}LRyly7{OEd(FfSCz5&CMUk^SGOS?7@G(Ktj
z2=F-NZ4gh6#dk%Gyz8^K5Y1`u_lbx5EdL{jm12IJuh^Rw8KQe|kRrmom5;4g>jrX$
z0I;fr!5Jz;<q6CI2w?D=B%|!OvFWG2GC!=d+<bZvn8cPwLw!_`$a?4VpdQ5yv?d<W
zNl%X<3y#-UfEK~AsQUy}@I~wSfS7yr6OPRpBYWx*_A=$wLp*8HhH1ueyfJ(M3_8m%
z(C|EB({l0`9<~}{W-(cB*e<ONlVtNl?r?Hq;!qq>;LZCrFf?d$#!%^X%A?1yEr^$b
zx?CBD?)J#&`24)7-q3d|Uu|opz2L&e;CkQTlgoK%UOauLINMx^Lw)_2(*1l$z&Ogl
z%caD5l1FZAwMD`C(_0~WdeKZsdyW%><eV_YT?>vU{GnKn39rKs#aMFk3B1s8`G?$P
zEFRv!yYW2wX5P!+H8CC|FOq%;xPDB$q7XTmM9X<3J=5}t2r)SwXv~&>C`m1j{(9j1
z;4O>GrA*dysT7Va#$3DdEW0@`VC=-v#mJYiYXQC}yF3>d<%E*iqOkA`8(CQ-BeyYB
z59jm%C6{ej8~6EkgL^KLQOX<;@WfEMg>xZKChFXwW+va1)Cyc+53;YJYN|%-(nac4
zTOMC$ym(>4dy<w3b-f^whWo3&j_{7d>8Z(~jpa@u30j^@>iwdq)*CvAl_gm~J&r?T
ztR^E)d4accQxZ4t;^JXHvZXr7xcdElu7ji~r?~Go8O-HB_j2O9<j(Vlgnl63O<l)~
zU~LFnsX2J}`cF92q>}v7iN104aHiJmeeo!PXzHtSpRPQsWi0P!QYeq%^qHq3z@wgn
zsIc&M<V1bkT}_p1b3eA2dV^(VG)gc>Oh=wMb=oC-@5I=Dt0v9F8xm)2H5(84_e13X
zxB;(hhUr{0jBy5C{o_Z|;Y4+X%RH!B4q13ond!(M&*v2=*~N3jZ74I_d8xMMBX>WA
zyv*vU06?!~0f1~93Zhstf0{o=ig#VKfT&|S_PE~h&5bjw4fRP9Ax#WXbD!<rn|-xq
zh2zYLF$+#_T5<HmtVIy(+JRX%sAX(fMHD?ID&ker$)Ap}RCjjRZLIBBJs+#Vg*JAe
zQc1*dGj#)XNERVIgkcGMc@4vSToLQ2t8p*K^i#55oc)_8S)O_NNHfl*<isR%u1Cos
z^T3FF#%u)}igYzLwroCaI00ejJoh5fOE6GN_sZPjx%<I({)|JOJ|4KGRM|5r<B{hz
zuf0-OsW9udQE0^Lx3}DXgg1*(xen#I_>C-wqosvkD2Ltr^evn4O;D>q?Ui}z!Gglw
zOKu?0jVsib$)fa&7aFO$b}qF>p{&Kq7RzR~Zpf46>=*mLikppQ=mexLQhv~4sBfvi
zH1hE9F5g9rYPltgy^Ff9{I$K7w5%DO?Z%cG>m3K5K0O9hv)my+cD3kmZ;TWjO@sho
zNtBad=W4`ZC&ngc2~O0k+^VOOAa)p|$ISFB3TR4j1vLQ-YM<VW;4X90Yzb3?{GCXA
z#1;fO)TbbUm7eSCNwgDsd=JiHz2WC6IS^fAJvCQ!a%1K0rYYtYjh&+*_<2NL&W@-D
zw(xYrzZTtFq4N<~jBD=~g(BPrGQAe+4k)F+BaZ%PW<kmLPb$5$|6u<=#zXj^mGH_5
zo%-}~AcNFJTDxGUQ1p0;?Lb3Q^y2MQL)o<OdrNl&&+%xesc}k^=oDRlFFNo2ovKSE
zTo?jw7)1fzOmrxKb2S=6Y~|Iw67FD4E{5caJZYrnc*{L_=p;5^qQ3R9RH5`{ex7w$
nStzjnB#|fMk0joIW!wLsUH$Ksnf~si{H5XhFWgS`Yx;iyo?w)7

literal 0
HcmV?d00001

diff --git a/_images/ca67ed7d78a0a607f47eea0f22d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png b/_images/ca67ed7d78a0a607f47eea0f22d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png
new file mode 100644
index 0000000000000000000000000000000000000000..995d38f33820311c866368c6b499289e9629568b
GIT binary patch
literal 48824
zcmX_nbySpHytVq$AkqlZ-L2By-AD`|jUYL6Dc#*EF?7g)#8A>PbT`s5bceuw{Jy*H
zAG4Ts7>4tl*k}Lt-X}s`RqizgDaMNzFJ3FiOKZM(@e=#r3k?Nm@vOar06zr4GM~U&
zj#glg&n}iPR6c{9>>a`Owq{iBmM*Tgjt;yW?>TtcsBFMsCs#pE&aeOX0S-qOYt9kY
z=1Jg7(4FM<U0=Mw#Q*nt`64s>?TZ)f^a|1v+Md}*k9L*zhVE-&F*<P|otEp(Y#)a;
zYca<t3oX2bA{oudAJT#Li3JuCIHn}n(MVpmQld;ac<&@=eyg8%#^W+_(Y%fSj)ZNB
z92?C~X%<0aI|#*>!=_U-O(HO~@K|g2dvF^Q8S}fY_;TW1>0A0#^Jw@?>Z+5ogcA6|
z`ql1XxsKgPfBNPbrFpD?p3a;bdCgG8d3h@fr-t`)$MNs~_x|}=&zLz@Z(GMp#~9YC
zN%sFg0``LN8>jqh-9J}!Ze{t`I=P*);i>bl6$1Z(=$|w8VXrqZ)X_=Sb@0l`)2q-q
z%LorQt|4{<5igaNm%FK~={xG^R0)Q4e`X_!ng!nXI`P|!RM-_xFjd#on70}Cd$+~H
zsyF}qp`xL2aBwK7Mil3*ih<167V?a1=BAerR|Vzedp8dwzplK=AN|GdAN{9Bfxnsg
zGIN<yUy2X}8eBm)9x3JJnNQX?R_6FQGbca~;M+XFZ=|Iw=MXy{*Huw^ihepv1xnZ1
zd$o4ihgv<M#Hn!03e}i7+lqFJH61>O$IXksUw;|=+7YJ7<i7N7oAq#{-LyOZF%@$o
z6r+HD_O_qF`+~IJ-PKjA$b)f%i8;9P=NZt8%yi9lszJ|H)JGwH{_xH}s&md};J9Rz
z6B6vOs%!qvi@R1^z2ND!K?}T~&L0`}D<uun?j38646*p55gJQoEG!r^K0ZG0qI%6L
z{2SVgYs|Uc@&(T77^P3EO@_{a-cFmKQ}(+$ImNShU%yAhEnZzKUUb++`>uM$)?d0C
ztzS?+ATEdQhXtlv5AqZ{*3|NDU^$^AV|R|moCxg&dOK~>haz`38f;q)-0pW8&WKH-
z@Sw|=#|Lx)9nH-PG~b@AW^<EtbadF#;JN|KP33x<U^{xH6VluebC0j1?t}G3Ime6`
zPKz^cE73u7t?VW5`>E@bVyLAU3i6YUb#%D%#v0#_>?Se|v*+B*OoZMyz+W`u4I~pk
ztPc!4!)8YJ4$xF1)BLtQEe>@nbTGNS^Lj8z)+~5>eWt_k0tfqs*IL^^E(>M++C4@j
zNj5(-#C-<q%|38LycLt045}LHg{n%?RAnTnA(9E(u;%87ZI@;Ue_8lIQHGi2qwnZ1
zAzQhDRm7DUFmf1`m=Qlgast@4a)lZ}!}3Br3@0Ua+L@Gk4lEF7L0W596^X+Qd`chp
zwxXWaaN>ar{f>E}6=oJo7H?1?_kkwt!5E(W{&m7yiI-DwVFyoo3&FK<=Y+jeur2Nc
zUZuUbD8KF3yOEZ38d^!|rIX=YbLAlE#4)*ix~_7BU}*PexXy(2+;O!$8u_+KB8Rp*
zSyW}hy)0OKF~z3*{^8Ldzl6Fi+=<6y-E!wUcqmzKU!b9|`R2*eZ~Zo#lHAKX=OH@g
zMqNTzXJ%N4@CQ=V1sihyi8K)l$>PR*PTFPb&7j7#S?&A=V*$lEWlVV?JBiXqJY&HC
zw~@nkBX_m5VwU`r=vtnN?Z=_}3W=$_v_FwQIa<9N)=v7tf-w^|(`E;qrsJBTurx)r
zBW#qQow}!lW^fHs*3r)Z+^qC4wm%$u@~|f8`fV^3e@Q)80U=3@xv7Z1-<p7CLdvS1
zR|?W7T(o}ZzS2&*wA$25<p$W=sXr|vrxU&?G<=H79W*j~&UrlZ;Va>hKcw6=n%W0f
z<ZAB!)e&V!7k{0K{v=8?@!78U^9{2q9bOby&P#=7_%$$aO4Rb{6vCe4lH~Vk-2bsx
z<LBWs7Hx*hKOV8r%fK7Zn`hoIqthVs_zZG+8XDaV#=m$d5+oQO84o``+mL71G6SJl
zt!@16b{0%D9n|IKu^2~OZD{4)Ye&l#rn(IDNp^4f6%Bj6<0Dm#;<zKIVNZiM!C;Ht
z4chK777Hj=>iw@tI*TiBkxHrOdYY~c999}L!JH3W%K5SLXb(1gU&h6)$eY<%g7dvc
z$HvV0$r=cv7OBlnw*?y|r3qX<^8|01UzIi+vU>D_pG~=JbPGxkrL{0No%o}Jf`R}H
z<EDt4l5fKd;bq=;0fVQ#{Q>y%6n1F#CyT=PZ&6wnOsTUxvYEIIXNixd)Xrn=VO!lE
z6AV~gU2WJgn%lqjR*pPs`5X^lhu8njkq<tuIfjoDrC93iW5v2m3KMmHS6{NU>113J
zI!UdDM|$M|Xba4EMMXtuZIgpYj;eXHQQ!Ld=jG+)dUXtRPWm6daambevhO&FqTXVM
zo%OwQ?CI^5!T&TpSz6h2kev9ZD77zrn@N6;I7$W>!>aA^XzWM-4h>H?(7%2Mdx+-s
z+kc4a?(VjVNF39f>2o`{@^*H0oof10w@?|vmQ0(p4H^HKI_~-x6%{q+)X6v2qRnxu
zxx9Si`uf8*>UC0Ua~#d!@zkdttz#aTnN!8F&8J5zN*!!$Y_m4w#yyE{sbtCRuDgqN
z4cqaVR9~(+X~mt@HlMT0{i!yil$^+JcBYVE45^%vSCTBp*G;pV7KlbRjsEw%`0YJE
zp}IgX2K!CL`DhHnC7Kz(7)E@~tOWk~`8gCqfc7z?rOS>qO)9|PEeleNk1$2QOS3(F
z{vAf*ZiGDdtQ&b_W23X5-wKn+b~qPPRfBAy#_;aN&f?XTThPzxYfs@=U{-+z6uIrP
zAn7R0c(Ci;=76MXBNsjNf=jR7Mw+MnG!b3cOTMF7XEYKs&xSg=%wN$f(PPGu28p;8
zIHL+DN+IIl4kY~nE#a~vp)&~QW_ezDlCOjSUJiW+EX}3fxMp*q&e{;8LtBtAj23+k
z3Zb8`)CyLjPv24ua!N-s7dwE(-ubp$u7pr<aB^lvXguH*Y`c)i$;pir$-f-DXE~*}
zkm&UU<^(OL8kpx#Y(&mZPUBP|;VUHd9CQ^kEF2shn)JM6*qcv}54gBzw`ZH7tH#eR
z&CyFsOHFNUU-vfEpLTYNw+n15EjEB7nTuf}JhkFiV0SZ4pawvrKJ^?OeRZwb`Tgi=
z;_UQ!wIn|5(th77?24_ra^F9^LqV%rK7K6XC(VaG#lPYbC6j}D>MnBA4Ev0Q^xL`2
zS}X<1>$l6E0Q&plK_!}MlsqB3(tNnZ09@JdMTWQZaC92~>ZD^=Ee+V&O1K3kg+1n-
zGb#x^S8x-+A^_aL+)+UD-H&(pzD%gEQ+?6Wwf(*TZ<3uR6Jl!FaBvSi3Zq(X53qk{
z<o0ULl)`C`15FERQcp_NWq9>J(ey0X&r<P;FusgeAD8E1V*SL1TuHNC&-!+G+o}}p
zSS$Y8FBLq&>jwJRsZ5d?9u6QVRTwZhd%#);hO)=FUb44|aKEUQg*FXM$`40Hz5>f@
zY6Oy4@V9BH;4aq2@#BwWnM2)K&bC2zd?c3Wa2vKXI~R@z0Dsa)tbo(gFu|XYgSw0z
zMJ!{Bo`yfk!Vi92(&vUL(d8-=s;DtkAPxTF{Ke%>HFrFVG}y&ua+y=61t1C^vU)v*
zR|5-HQ73eirRYysiM(YMzhh+mG`KjpWKeY8l%&_x(NShdBBB)7*dnu%aV_$ZsR8xA
zR=dveMJQE8%KVJ4$tY0!-qze~yJ2;BVgJDnE88dn7?9`$w+c!OD<7-+<h0~BE4CwQ
zL^h1xL%I%7sGgJd>ts_(5Rx9tf&~RRIRqnUtR;`gxv^*94<kDts*+fOQXfru<O76;
zG(+=iVl5P+1bb;g`l~zn>-9>ANvE;F`(~BwIA92MmonVwUk{4cUuqj4YIUr|0Z=0Q
zmaIJKgSKP_4Ha3QG6Oeh45^3m5qj?(f1PCM5!3vM4+)Z_0f1qAOy4Ug?{kE^mPx1u
zI|6tz_#}KuwW}K<8YF3t8<H*9X$%H5ZZMh^-JAN$ENpJ_r<B<H6RDvaJ)7eCL;#}o
zZ&s}>pIbVhN1<_qw|!zGwMwH|560=eeLLXmBH$S_&^r-|gIdK$5;QQlLR_`6=iiGm
zFD55k9`;?Pp_5yK+l9@@Wpi0`u=Q%z@a%3ZK=kTaz`vn?qk~yqFmG)0?qavfJ;wuf
z8~wPa$yW58HLiM>Nf>PDl@r<-v;-c}-USYcTbB}bhDd=Zxjms39TpFg3SjUrwL+6o
zQskwSZd;5}+Hg@bWM1D<ya|{MkzQ+N%WM|CKU_@h%}CUZBhne;<zep{Un8DQIt3XU
z%S!W|y?$%~c5SZpDX*Ei1|P}r{iU0qM@;Ek;EA*I@*4M=FazPxRsihJ9!~-~dVenX
zzu{g0Hiu2o+ZZ!s;K(SHi`&?rDlv>y3y-c)X919#{n_Yercy1g2R98-jU(|zxq4X=
zqka=l!GlsQA&XH=K4y-kLk6q4)lQ;`UrjtKHD`2CsRJ`9Gg`y0VxBWjLubRa^WF90
z)=uR2ps8=X^WE6Zc+_OuNT(E(^>b==GDkJy#?n56I>!^HT;XS)%b!p>CD|n+%K)~R
z#6Np-lQplRv+|X-#8KoeRwlhX!(SVkt7Wj@SAs48jHT^SIa`7m%qwTU5US;gvZEyO
z3dZ$sT5-@FCJh#qseQMYprSE=pTW|GVrC;>smRWC0Kh$q#}gtx<}va8>9Mm@0{`9V
zhD|Y?YTsWjx}64si1u+F*2XSJT+Nxpb(-Q>TU+Bg?;ha1j=YwR!zwQifE~LY3_yc1
z1pFtZ?mo1<n)RtFw@BJeg4MX;Oled!_vtzx?S)h@M^nMw@-5TolQrXiLmvUi2daVn
zFb4ZJ<Bt=UcfJ%+y#@NQG4p6iue8ebh6WhF@R7d%&ckCJ;ooA)ZVPF@(633W&{Ot_
zYr_0erxv^lwrpDEq&^c(DqHzhsar8C!F`alh6ttRRj+RVg`k;I`qL-x7E4sUiUlZ7
zmcg7GBJ;lAbf?Yd=HH21U0t1NL|)xB=p<mU-ZbMuZ3KyYXxekf`|O(kdsB-q7G!x6
z<oox&9jj^sCobqvk})IHB4<Q=_y#|p@2-2DOI20Xfa!DFYLNsclvP5Jh^@fyG#0I@
zr?S^<BJd_HElnj`rkU!nf<bok!jBeIzvMQ;t}^#`&zSr5O>lTZ^A{OG7X{tNIZs=X
zjF5^C=(Uja6`Llgo8AZ4E3z`H`pH(PAF(D8jG-XnRz7e~zo{TEnO86I@^6XgRar5o
zba`~jzJ(x)iVpuOuua~w%qY-{!+1x2cOyyD5Az(c<lPQWRGtq%vp!NdGj7qj%<-tB
zv@s)1jv5haoKM`{-4Six^4IgpG9=LxD5@Jq_0sHIc)@$eMKs0P%wExl?|U@!1g~%?
zk79w)hSNUf?34qPsln(Y_X$7^aws))bnt7OiBpuK3lJkCD#KQFy1sH~Diz`xGrLAE
zWE}I-VHxK|i35AmpXbg+lV<qRG|QF$;*a#alm4%LKf#$=<auwR@DoSch%wbDq>tTL
z<2hTGOe?id{Qi`5EO;=Q5J8TGa`vY)h<&lPHF;xzArr#dl}kt&Tr*&SHRhg^z-*Dr
zt1;(RXM8*q^waQJSWWs$i8-P<gMM#NMVd&OU2%KoZK*K7DhyBcO$j&qPs~_MstnD?
zkq<95!)G}BJgyQ1Fb;VC19R*}TQ;eW^F$UqxcME|N#+jf{@rrO&CTs=IcZ<IkEx#%
zJEqs@TJaTig!rzlu6~gJQGGI61L8<KJ6-Sg{Cn$c{;5cP`yi4*Qxi9V2D#+>{l2Td
zB;^(=OJ#V={N}~(n2>FYo>Ou!I}JOUP_rYby|RZ&TY=UpRsRNG_S(CRUr_LiKQXOc
z8pGVsuR<-!{69N`g9@zSEC_F3-(@Ch?jMtLgq?Z;{IWCTwd9vW$<`511uGnuT?Oqn
zt-!wCerC}z#t^SiQJmf2p{S#mSpxKdsr~vg6G^TNWm2?J6_;W9v*zgw)hN6GON{nQ
zHpHhY-GB7vU+<MlEpVoZM@Uw*CS!-<A|oUBc6Jyj#Qe#CZ+1=!HZH5T{Ii^bxbjxX
zW-rtUrDMwbGS=un9?oo1f_W#`TC@T1W$ql+tkTmrhVt45d2BZ6$cmt($noVSSxJQq
zE{D$FDZzoW`v(W*)Erv#``%8xi3CCSozs`$Hm@ucn**Ny-lc96e&wM)^YJt(=2KM6
z5#W-RZQoaO8~<AP!^yrfPRLt_plS{SJD>NpEWmFTcwBV`-c&Zie>giB*_6`F6yKLT
z%EL^!dC%S<C!ggO78W)&HO)(cyhH1$5>E+ScUeVt_xCd-tbC+diURb!n0srDI1B|)
z4&Dlt#vW~!TVPea6!`qfT0tH9-WU0zzAlMKPv4EsTSrI5B3r2skXjg$6fRhhsozN2
za`Wk#r*}|oUjV}gJh?L%Ou%W}Iv=UTUuf<8=4s^Ejc#Bq#5Y~6;9(F<L6%~+o&kZ?
zOsh3!Quw;wR@%`wriJ$2F&3f{Mqfl>Vjw3CMKQ1APE-y>h`NYesVQ9rby6Ao9R0dI
z_UKqT{b42#RwF~FprCL--g^I$#L<Rh4?8lC#M7-ed(b8g*KEN~AQghymCR6D!C!zg
zg+|h(*5?D|GtGY?A%^_IP0UXRy~7b(2v}jD8#3gXlHZw7YU?_EZJn9Q)h5ZJ-b`hN
zv#r4-hCi9q=py}|fl^M^FLE2>uO!@<)uiOGnZntoi$EPZI`84d!!8gw^)~!Z{&FX5
z^)MY&^Kt3W4J6ss?UNH`l<=jpowFo(6t*mKW)wk5TaZ#dVp@FP^_F<QYR$pXaZ_mR
zPPHQ7A7z~j{9Ab)!8LCc+7PkGUpyS6L;a>U7ppgANQpVtzQCjb+_@h;xPsCAYW-rm
zVL{c<){k7h5OKaua$!nj+feY^U;Z03V!RE@X{E&j7<<o&+wyws<xCaKZvKd;NzMKC
zLrd+$6492^mdP}pvl(h;)ltlpww;*$v7$OhWG7Ll>39;35w}pTWwXSO%fFkrA~~`@
z?zBov*993U_)C%*na#AKC)`i|GV%lA*OZBA9%}ZrlXjh*)jf|kUpU`o-=_zZB5~#M
zt@&Om5XxJ9zSSR2>lGQ`qak=a3Ag!CNFdX*0^J4b&2V@3+;GTB30C?!(~HNcRQD?h
z3^_N}K{$yoKUrUvV>=0OoQ|B334N4peh-#W^tPOpjFjv^)Y#CPWXsEnwj66!$}Yhb
zFfCWhIjCQMjohSo!yqW&LRIHMjTXnq-_eT2IVoc<M!V=E{`cePFC5z**_UWB2b4E2
z%T}+v>nsA?-DjD+Ky$u;pk+0f8#%lw%J%~D;zATWsLnj~M`;_Pr!zedC&K-<m^a7E
zzK0Y2K<}A%!l8{vls6F!)Ofn4vDL~wD~e(Fg$ffF2X$}N^n+uOpakovZzy`|tj98W
zKMW)r%`^kz!Q#7hqQo^X5v{lNR2GsLZ#1N&`D)2T)Z{s&sl7X%?h*e=)uz_gWkuh}
zaDEYyuMJW70z9TA?Q-_IZF?V>b?FS$fACHxrX;!y#iIof-xuMyL<Qa#3VIIzusKp%
zj10gfsW<}4E(^j78`MXJ)eR4ni#9ScGW&e>kL~t`H+P{5oSY&2*X)7KB+1&&2zi+T
ziCYVy`sq3&YG($V3k{zrYO-tY1cfm~w(3~hX%D;PH5?9~&Bd0jwzMprHx78b0*b;B
zUGEYx<WMY6prY{(fnc2A9vD>Vs#JczbPA;giF_*c9|u(-1Sv$krxiPIypyMXJjf8o
zVa6NR-?rN$6#HYamPXAixA1ql&y%ZNe<|G9dS&?1KWQ*<Z@xB4N!WNB*YE7>OSZea
zTh$|TL|qU_{KmHVF95ngslRc0%9Se?kV{$#XrwnMYmPghv!<{&l*Kl|ir_if#`5EF
z?7-FiAe5tdE{$MJYmlR_AV)-^_(D;lt^V?4g}k9w9c!BS91^Nqc&%#thAby{5cHAe
zLu&fVx<QM{LlX_<KU-TxDSl;=0coP#X;`4M=9(cy5$nE*=FmahhqMT<t9L1x%a{f}
z<YYaTh%-;Jyyw5Wn`fK-XB)jyeFiK-n)`R89e+FlDf>@h-?ll2z9uvNxjwqzIAqF%
zBqchIkZe}7g;C2kH{-PbZD^m$HdAMs1_ZvjL>fLnAhl{uDole0>IiuD*Phv_T%I!t
z5GkIrC8MC|OJY>ZWYdlLC{+nj3Bt`A*P$-7_AjPIAzYzFBzG#_ti$~E22IoO(?l8T
z$o6;SBV_D5kl`O8W$Ip*M1hxs*rm=Uf>QceMiZ|G_cn(F>oB2;B>0KWca`a>Pi=vr
zZzx6Wb(8ln_WdzXR&wzQz=hqOX?-e<KL`d2k!N{Tcgq)hEWO)|jf{-=_!8bxQ%rn8
z^LUQ*D8(tmv7jqU4i@#CE_^58&-ln&c<6u{Q*EY=xDo_3gN@BNvDkg^;ueI9GeWVZ
zs8U}vVZ-UjIM8pwM5V~lYI1GZ8|R?qL8|e4$fA+-dS~~!#`|?;+LFmB>vngEBh380
zXl#}!WmADs<w46@3Zgg_&)@UNqFMOAIz1}(I5&x;<8nWh`1G1a%Ar}|e^pQ~EjoZX
z+^w4FP8xr$P-pU_8U@Hd90bS7dg3<DUjxhsBV*$}Kr;Y}$Q>uU9b0Mp{wA751u#sn
z9wz|(7t3++3<EO>wF$D6<#e+37RRr`6Ap7`z3bD{pAq`ZQwwiybypmc0eHCJpI0E>
z2|v=K^2`qy!Jd+F4n1pEFRgn(3SvGD_8ZawN`DonWKR1;p?n=e?p{D)c(^%{1I>-t
zy}G@BCE=wHvNoOQpTM*2<F*+7Z7uMT1RtmD)E_-R88xMXwm#fuSr}werCXJ3XB9D&
zdC;P3*E;rP%$DBR$xaEz<fVE>H<=jXyCFOHdJSM;X4v(!k*r=glXn`MGk(Kwi(lb$
z9`NF39HiQ%F{2s7yO4_VjWo|3daNwtguZ~zETAbV&wV&@Ju6n+vLjv{tjH^fPR_<`
zdZj~M<_+vGO5LstQyH2`17r4C`|Wsq`dZcP7MpAnigN=mz(QzJ<hphIF~0lbyL||!
z8@Z$LuCkM54Mn<6flap=xb=MX2D9cL0)paWpDf7`Nzbud5xeV-Qp#$uG@}Td@J-1P
zpdyTthzJTsPS_Ao2K;rNc3TlV(B#AO&WGrpCYlr8Z)dCLyHrY2c*hQC^08dyT~w|C
z+ac-bC~`GL)^-#+S*FWF$!acrQW+dEW<fd};QAUtl=M|Y3U)mg>>PCqXpiGzKQb8v
zX$+ROdM#*p9u3L6TwGo8Pud?GluM>U7aqMRwWNjq(+Jo$v>NJ}$iVootvY-eCb&#k
zj^{d5@TJlX<cN&q#`);F!kU+XV-#5ZqSSFFF5DAQ)wofY_h|3Rxtn6X>yh7eadBbA
z_@$4xsIdji<}d-724psKKou_eRTkbTA21<U$ox=rO@g|SfMTS-Ufft|fs*t*O%%rq
zSW_~*MIgN_HX&TOP!1m1>rR~22yu!Y8VLx9e6f~1&$FvD8qK#8$Apv)lRht7!>O5R
zdI+YPzJT~@OjoHVTv`Q4B>cj{%NYJ%f2Ng+*FjXmrRgIHtaOdx_c0Dgu6#5F`Z9^X
zrXmx94R4}imw#^$e*_BT?N9UzWt?%8Vh*~;J$j<p7_<*enbs)mt|!_7w#>74*1rsT
z<tY881@Batg1#%A1$vgwEs>p9&i8hQy)hK-cF)j1z?hogcu0+jaVr;F6lV?*%1}07
zteP6It6u)fk&?rWZNNJB1PTuvi+CvwmW3@0CeB_A&YpmR=nl6Z>&5N!u38W{xGHN`
z+4arU6T6JdsZeLDs~G~A=a+8w(y3`o<ZqHYT=nJ0x^&XS+-a-~)E}>z3vvjGAHw;V
z+1d}wp`X|`29p>^nwICCI27LGO7195?T+tGB6B`e3wr+H!++G~+-K|t|I=rCqN1Wg
z{dtUYl}qT=s&Nt+|GFh56P`y$|42;`Kpl?qy1(4l6~)<&Uuzi}MdFIni*K@5?(wiH
z2^@?yprz)4{80=N&Fcm~MTJ|fWWwXTSzF#3VCB~_OPlHotN`w&ouedR*j;C7&-88Q
zQyvm&n}7BVahv!xw3jpyF2i@W4a3TuI{IOmwzTup(2Kp`wcaajFek1IXih~t4@Ii$
z<U&vK)$%ZtUK0FgEYY-{1M_&@7hXHVi<8v6D{rHy%c&b+nE~`Y9Dpt&vu&~jQSGLU
z|0Yt;laRf7eZ2K1fO^U>!p+#|B`!$I`R{n=gyw-QN(;F`^9N33fjU^-SjG-SW}VkX
zhi38Yg+96p!_B)=4(7Kr-O08MZU7zngarE>Jg)5m*hq-06}-<4-_^(Bxd%ND)leoX
zzPRFy0k(j!&vm73mQ&Vnm3PLj<oY<jyc{}pP^L0uXHZ{XeM|#y_xp`Aa-i26k|Ga(
zCnVXRt@>0$hV9vtwYK!g8&=vmTXu*nU9Z*^c}hZP?!2sbysq?b;5d&fQOj$Hb#j`U
zn<!oU2MfTnvO{WF`C)oPUJQpY+0Jpd`wa*3GWNQA9)=(EE{%amXYK@PrtK(8M0{yx
z`T!5X(NPBjX35>17Xnl@^w<0tg}ODcRNtt3;@i_X*ln_Hu4O4bb%IWc^67DJ&xZAh
zu6XkR$JIU@tw>wMrt-=5zf2qX(h>4pVFjf}i{Wn=Ltb$h(;=fN@i>vW>&A3?<&{!>
zvUy&z<$E;KUDpG4(cIy)8xkn-BuH*g8Xsy3)ksYzv?Kd?o$7X(3^h#w1}h`zp3=6e
zYr+D-K5%6ihrr6d@_H}?L8^uM2Wg5;uK>*@lPWLV7pP{rh^PyZB%T3#8(@ywHE)BQ
z-_5XANHr>MI(?7FIj6Y4`@2H2(t+@rEUiX({#&g}c)o7}54?`AwM~kC8x2Y?Gt(N+
z&EFj-4<9Jx<DkbIE0${FQ6lmsomNd{OAIHX-CN~0ujY^~`d>RCPaxcqH;!3tE6<#K
z?Zf^S!BW!JQL=ltBhdODj{2IkuMKn}=85L*@mgJDaXdRPh4ap+bQL5lkDzOsva1w@
z15lPr$KD?>J*)u{3}`!6Fi)N4O)aM?@q~IagLY<p>1EwYfz@8G?UtgHWbdDXdnF~6
zA_CkPDqKG&-}WJhcZ!W{6dQbG<Xb_6p*%{4Ku0x|TTNtNf76?1Zco+BBuBsnb)|By
zX<W#n+oDb$kctwHN_AzEVPA~836_<{^tc(ozh|RQBuDD?8r_H?dDP}F0c}HY<WMZ3
zgHyX$23i$%rqkajmd)En)g9Ve$yHo>axBTPP-2!k37p9X;6>bRYk^&)wYXH|C&e!+
zkfbrTsv1`TqlN@PMb-E1Zzs)P2PbeGsy-(|)!Q*LLcw}e$sKT&!apv^L0>UbxGOZB
zU*_XAe9Ta8c(oZlWpij@U~;_G*NT+@qqnB^+mkJ&Bo#X;0}>A20(Fu5p+FfWk$54|
z81vn~FO*L;r$<*uM~V?yu8iqZ$Nc~zW718~lHTxk9>OmuRYLL?UY})HUma_WCabLa
z%KAtkctjLY!<KA>Qq`Tr?1uL?$VU*GJL?Ikge_*4gce`3acMrxS_<Yg6wE{y{-cf7
zVh;<t;?bO_E~FY|g2TfYxbeC9zXx<aztG2G^LA2Do4XY9nBWPGr)F<6PUNP5eI`wN
z$4rw4tJZ?7M<$hy;F2L|+4)dLl6E2k%aA@!B$vjF$w}j@ZD#KaX(a7{22_>^i<}~Q
zT5&`y*OW>Ng2#C?R(0GRDi|*<q*|7P&4P+YY`*pyWjcKsSN`OQ!||}`WcXdQ)4t15
zpO6NxIslOuA97$x%8fqO{h8#!2|^!w0d|O;nIKtVD=8!HdrK$%S7lCzbCV5Qk5{^+
zIbkwuY}asr&X9<JoJT&JT3ga#Vk{+$;QM&NPUni;dAd!()5JF%WF-l!`k~2z=9yfF
zvMJRwbW&wzX8&EW(eGoQWwSsbzME(aG!4exzo9p*R35$Lt+We{8HjE>JU(l-C)fOu
zdr7p?#yVSMi?((pfnCI3K>m|ry4(8r@uRb|^Mv3#Z47xu=|@dNGoyfq3`4y)VCOFF
zQm1Sbqunu#|5BDMimyGph^XHazN2n@3*!4Hw4XD-MSPA=_faaUPho<5fZc|_${}RN
znn}nPBNdHRxO7j8t$wjY#j<fyf3MB4LCBD>@6W<S=(pn>7+urcIS}9SZlTh`SUU6;
zfRm5SZT7toYhFa5WPr-6jZSiIK|P93rO{2p@)fG=BQUp`N)-Imf#G%xYIzR3t&GG2
z-Unc%;gh*8#&eKt;3d<8NQ|Z)-9U{lI{domak^?Y%sOb@%>vO%<Ir3Vt^?M_#3YLB
z7OaV;tHM|AeI9hX7uuSt+B+`I1B}a31$Q$X2SB9d?Hh7aw%&|3u}@=grn^)W2OIN)
zZ;aNpx9?-qF2U_nSBEE}LyQ~rgs8|ILBp$q(cFIpH%YOg6JI#sPE(?b>ox|aWV7)!
zweT#5a>BPz`&iKWYF!p|_^2lbW}j%!)3ER6<sYJ{6|YiosBMdAalN2|`OO4{i<UUr
z1Athtx#=Q!Cw<fO^NYQxJIzl#+7>s)mHBpSf5Dn?REzzbMw&(IhRJPnCoo|8$g9Gi
z1k2+L@q5y`*MAB;NWJu#(V6o&vHC1#boZsb*IZLve--1d2%H10tvGBePt?(M!G~Hu
z!VI&AGxz+hDKK#J=`)x&9K>XRC`;BzroyBcM=uLM5Qu@R%@zS7FUsffHjU?n_EFF1
zaQ42LL8ba3<AHJdb!Wa*ZoRT_1?m5qO-LPkR@Q~h{6{lBbZT^R<wsu(-h`LtX{$<q
z6FnioG9et=lkCi&$j2l0<JB(jQZ}B5sAW2Tw}S0pfPWrt(=RE{h@vvX_^Vg?<GB=Q
zjypp+)0EqGs2Nge;Y>|*&ROG@(gu_zrV7k^LML!R@z<l)t(&5r#vX79_=xumb{zI)
zYxtG-N_IFMqI+1u1l{TKH@dD<RAh`7RlWTK!4Y8sy_QcTyM^Q~--YB{5cr2uiJh`k
zoK5%V$G&>MsvqGD<x<uyf;`?gD;}FbZMW)D!Xy1|LUq|p>`FkYDBHT)(_(D;Y9*t<
zg&QJnTNFOtgAZ6k|B8CxFl3)oJw_$`c3`2N+EL+{DV0m6D~PE*cDDRBPQ_(;RIH%6
z@V4^M*U%3gW-;d{$;(F*N_*iHQL&#cq3>7x#}%*a({%OQjJvmj9#V{H6|q1pL02>s
zCqz}rjHm#r=7mMR7QB-hJUf(>u6Q}^C>;9(thRG(arr|sO<W^BT$&$+nHlDcagScb
zJ(gd7{NZHZKcA~l2=I^ou^JitT}A@~6gFOOCx<3HiOY@pKTx&+^WlE>4{PFT-X=iV
zxV>%3su%8%Mu4^{`S`T*;5m(LGG27%`mAJy!{h-=<0)=X>|3yfVmcpcliAa?=mv_@
zTTWfI*3?c-dNo+s;VAU9RDZO47RC?is{Bp#C?#o5Y_vj*TQyi{SY-M(ZK3pvS4N52
zjWpt=*D7j8zn*<|wKjuMEuL%D9AcI`^6)=)mN&cW<>8BZWKvx;#Htl6!gLWsKh{n@
za#-}SE_?{Mj(Hq0k}7FCzTl5OE@7=RT5S`e-xx)+O_5iWnrm95yOdsxlnd9L_CVg^
zE&Q_=6dBLA!K>dYxi^TYE>}4(ba%(U1E;99mT#PL1ohE3(e8YnUJlqT60B`~6ATIr
zmFDs9Fw}35Anr|wXJx!rSVLFRFY?{Pb`s!r?fqZ(Fui}@Kyxqq9SC20_e;bcf%EQ?
zb$s?QMD=Bo=D09VMdDt_$+=~RJm#o~ZE6fWU<^y8SY5wkr`!C6_K*5jO$?)nO<7qA
z@*iu-&8-GinaNjx+t=!w8AcgZtSohM^H7!2nE0STQLw&cWF;{Hh}z*cyLL@R$wQ2S
zuE{YLuSU=<^3`z~1P*YtQuuxvQsF5OLDtsnLv#(Hco+x!0{TS7sTcHc7Qou++jIu3
zZ+<bcwM^;iQk!nFs2Y~rt7zTwBrq*+{IO=bv7BrPuw^P=;e2SpjTe#?<Z6(-q|xpd
zXYz<nz0W+uMiw8m)UF;Srm)#D!AQi=gS~~Dnt)?$lo^{^mbB;GlH{3?s9-LZvbCsZ
zlqr=q0u*tLtJw4J^UCA?$|TDOA_R_q<tyFrY(ogc)avD2J#4+PyF$jQI>g3yq2Br(
z;pIL>XYZ*AuuvY}PFVe9qRC7lyh^H>HP0cV4H2#TLC-XLJwrhvtR+q&GK+9S{0cT>
zV`q0Qi^q~%XKaH^*Lk^0Y+*#+Z%e3iddD`?9H{mE(dOGyVbK#UK`=5h0%+ENH*dtO
z#5MD=x4YhCP_M4ZM;Cm0w~N<)APl4xfoKp4<PH~SgS&t&??1ghQOBa%`*oC|?H29p
zSm)AY({oXbpd`6KE-P`$WQd)(f6lnh)5m;h`uCS2`dGP<^z<8!v?U>gmuc~PtC}yy
zI;;cb;lrhV?&EPaf8rbNW5Q<2{ky=8^96=Po>pzxVZV^};Sk{pDVa5L*(ozwX;%4t
zUESFg5CKQP0ck&{Hzhgwd)EV3Sz_6RpMqxj)d<}~W#6KhAOaWk`q2l18uiat506CU
zlf1FeLE2GN>F=#O!ZP<RqHb|JK-NgtC#U0>Nk`|Gsr?|HU@(}tpK&d%$mab({AzBx
ztQ&KT(GV?7rdHma)rYE5t_~WSOmSkuR~ZUeFe3+5)ZGk<c0K7$hNDYE;x3na=V5xw
zgM~~e%XuOF7Adiw`mJxxoE^G|EnLsorB{I?KRIcy&~8~IOTzhaM=^{@$4|Yl_#(Ja
zv_u0m=j`uKF_tIhxTw{@h^))$uNjd3G(w98R_Pk^Ofw?8VwIB5kaU_$gs#g6c37`|
zV<kWTWHm4RAw7js(K-A(U2?{IH^}&xWwc?1KjO;;PzT9RaRzqKJ?PIoJSyf_9wO7R
zs^L3#g__tbqRnKK`x~2ZHq?;ch_U1_T4a^hbVJ2&XbKJ)_kDjTmv)B&t1CF-$ZZyc
z3LZIh&>LHf$cU-=^?pNLzr6pK$lc$!MR@D-JZU}V?n>Nfhhf?Z`*!@ir+V6WIYbHv
zq5-l0jSs!u1>y*>Vg#S?7P(y7#vRt3ojqwq;ZF|lBP~&9r{r<@9_!KXysoxYxvZF=
zueGYiw9~~d{}{4WXmVI8zE^Q+a^Qiw2{QXH2~{k~-Xdp&sliIglcxR|DI*}m!#kaY
zBApB#G@W$g#!(SJ^3;@M2C8Vpa=+UUOCu||-H!c>8HI**@7TZpc8O~;p^3Q)0Cz*O
z>&`s@kW!w*2StCF<4e~{m&?Qn#&G>o_!y>@qzlJ1<#CTxa<ysbX%q6fKE_HPki4<=
zbzpO6PEo7m=NUL$_>uX$8i`V*X7ESqWyC1TiGjdF4qPP@OP^Pz;R7>s!N~PhW+&VD
zZuEhYp(r+wDiIU7SEBERdyjV?eyCXzl3bdc&z6v!V|*na)!Xd_KDxekw}Z%-7vN2Z
z<yv<3@Wxq|i9u%Y9|>RI`KCH1k<iN3_$$66#1FXzl*eBsa*wn}^aGkGKv-=*h+t=8
zLc_?!q!A}|ulE)egmNCg^SzWmiBOI@<FM&`U$fHKSkW0rB$=UCsKGJ+{pDW!Ll?~C
z(U~$X5B>dBHiJS4+`&S=x|%DVk@|G7$f?vqR}52B;J^h@;6t-`!j~{U)TMA%Mk(B>
zBuT%+pfyEk_RIq@!KmhzrGczlN1PjDp!O?l<2ZX*VFcV5p)vnr8&Ebj?RYP0LSgBo
z=f*^|wF%bVX%^ei$}z9-sM_<IWT45E3NrFaoC7(Bdyq(dE)Vj|FsF7DGH_nB+=dXK
zGWyT?{r#)p%Y5&&;ASwP`EagKGG$x1?d4vC$#{X&bRtReP0>0vk5z(jYJkVq&nT(l
zk+mJxscCi_Zqtj;x8RN6+{Y^OyW97@NxPxuD@I~vAVQEq{JZWw&2ezZPveXAXeCIF
zg#C4rq*8sCB>Z38t<r2j5z&a1uQ73u$?E1m=L^I8{CMoXFMvS{LLLj=z6HfAaiq1^
zt+kt_P#e+k_MP(>FP9uv@hd)Q=`J5wU)_p`eiDTbx+sZnTzIMU0Re!vWxcLpX@;=H
zFL7tDRVi+o_G*xrTyqG%-f+832(l`>?g#;j9KA61SN)p4;WVXv5@uVBkIybX!n=!0
zMXxrjIQ2DYnEG(v3Js{!h5W2W%sR)z#`Y@&w@C#%ZX;2a60-`D@5*69_MfUb^eU!N
zd%N~fs&p%MR#=q3>gBXw`3vR%L}h!{g*{K}*$;b9($ZyHzT47;QtEk<c{k&SH#K<d
z;Ox6#lI!^`7OG4!+m)WAL}Q??^bPuf!tKf{Y+(0TDr0ym#Cz88{**hq(I6%H`iN8(
zZIzLCuqN<94;*RE<u&MMpTkbwv;8<*o|Ft7C)v+KMxf1<H9>xAL+4TVt9DmX&Ufn&
z6dwyt2fxV+L~%soD`@O}^~c5i=Lx&~XH};k<PKp1gNB0P%K#<c_Wq)LOsdIYlyH!#
z&N77qRx9>(Pf5d$J(k35JxQ9*k!V)_!=tn=)0JODUhtp9ZMYB<{5KOjnu&?{_eog+
zbRc40wUVLwE@tGHaDcUQXTtVyybKZ7g|+Pju#VvLpC4JEiC{f`D{j7i)_g+0?P6yK
zrM|E|N6H-tH_q$q<AXRxNNs!xgvv`CHzwrRnsT}i2oJ@g|H)+>iMTE}CGd+C*f<XM
zrna1nGPg1*{XuiC1*bi~S-Q=xI0l_Bd3~zA_Wo(Jkm$H&#yBM_D5#++NC)H1m`h`h
z<@gTnbo<KAs$c{ELtD5ist`PlyQHgn!{l_wgOU0wD@IMZNYuKzNbXF=P^~#N(V~T9
zg^%bu0ltPi;bJHpeu$A=AxNgTLlYs!hiz4O->Rh!U6Uui(vLJmgO4@+X(x!0xGAHo
zm|4-MFz7XpH}p#_z+FS1xMd+HtV5ly<4$aUeA@X%{?F*~pYyCs4xenM#LfGo;*imG
zx?wj4(Y$95GRz1mpqhxhGg&nWJg0*)#B11h_5)_y+}s>H3Lfq4IzPzVgMAGd_wF4D
zb&JU(^JCvGNuw$DOT9OHAJpT8ozj?1CWd#zBEMI#=E^0G;4WE!@Q$NBZrN-Vw~x7r
zW|TaV=C&it#>H{`n3+bo&*=otNX+qkqjIO_f^rJnh~(W0n~XR%*So`Pf6v|7>)lCv
zk!6JFeGXI9m2}(7oE!0Ob?BIj*yF0S@KaE?ChCw=_6MY0Z&xDFe<u&s=u9y}tQ?2O
zU7q9wo|Gjxft7`Q@$A$sLIa{h$oVd&*;E^&AyiOPtUy+QgEU@0hSMkgtK@N{4%gLe
zR1vD-S{$ADClWIVbjDAz%!q=xd`dC<z%-i4?4fkn-YC}0W_+)`OiVfi2tCBc#ZA$Y
zDQ@E<%XzG9wws2Ex=UQ9g`AGgSC{GM*?m2+V@RK#q=%GfO1@us1S!s^B(ciZ@ZvXs
zu&6$b>@K*6G7QXf{9M-C2o3b5wyW%#Nx>UB<e{Sn>yvgclH$D-e9a@RwB7x!>?VrU
zQJO^1NB5yfn=|RicYnHH!SavN?*dB(A2+|o@?z_~R3LS&`jOsDE?P=1^A`mdQ%~1R
z=bTn>fG&(}xP2^<Dx>S7Qsz1#@b~TjS8&YVmMD&eCPj?jdGP9pbvG5P4EASt!Y)vZ
z8V-#0hK_ZC-#1%1amr0^hrV-cPl=%LMALpD8;tJRbG^Vo;u0H9Bsrvu(>6wymFh&&
z{fA?v92Vs)PXgp7FrtM7IY<sTwq!&4oM+uu4zkizd#KqJO;j@{h1SX8sqNP~6nXVc
zaf)W=nfT%60~@@#ms?mnLw#_@$C%!faKkMBj<v1-;x%};;wSpku7PX`R#>;^(_b(6
zhTsSfsW<hsl?ha^OdLkQ@o;cKlb?jSdDDV>DqdM_+4(t+GR3oRe|^kmGb>KvIO<g`
zpA%O25Wn5u)l{4KwwACPZ`sRsloD4DwhYaX{SNFH-YPko;{)Z{eLr@PZd{{^tUB(Q
zUlyO9@Y~->GN->z{rYZr9@pMGFu&&R0;F#^=V*F}9~e{d^Guzh|B^KaFGPZ%UzGQ!
zGO~Q%7?zr}eKV4;>@#p<2soM@&+wnETz3?Emu|e@F0;i1<Zaj}qeX^cD>W$)PYi19
z%!pNExL}cxXCJpHGH#61CG!Og!lb1!HT@Ddh{~pOZ_@Bjx<F>HVv^=WD5>5lF?Ys1
z?vp-q$mqjz^UMCFxk1nkA~D0d8br$`9^i}_oQO(x5fV0>)MB%+#>IK9LECsqa(JgD
zU#`!GkXkrfUb%^-6}LIMt+*4fjO-tnC@k#l4Z*6jV#2gv9aPv*bBb?nxOga(r9{i`
z`*RA7ZJtK#zPQrM%i!M(%+jx}Os?OAKk1$etRyzMoCin~x2T8kTIqS|({1}%OjNJh
ztKPP@xR6B8-T;xsI0}(Spl+xV^|)&5`SDbV>7GV|V4)9LKk}H_?^6Z4XdewI`)wnr
z-AR(L9&sk>PJnRFs{g~$P7%)_PC^(S>4Ion<Q5*R{o)rzJDaM^qK^BH^Pzd?Ct(aW
zGYf0&J)bg;r(?7taGt1j*QS9nlFQC$-|pV;!e?9$2|a_E$=hs_J0IKV7Nl}M<Eylz
zW0p)_$!(61<ZWXc^vQ=&7>`9p=eA~tM?S3Dr;QT#j^X`z_&KPlMo!^7J;B6#BI<1A
zW(qR09-!vOry5be1Q_n>W1)43YAIH|W&xn^y}MicZv+CJ((%gJC>y6Q9XRKimGLX!
zN+VD#rn}Yqdhs{h?_PHP0Hd_Nq{uE>tdh$Da(M_haE+5nUS#g21+!}r=!?|}OkaVR
z`d7mIj@`@)JG6<0Lk3<o;!T4>P;px9ehVu6i~6Fq+(Wo0Ur%_@uu^6vNBO69;hcPV
zqql!Z#_2;zo_^<P?2c}`|Kv9$dXpHhpV#&V`+_GeBqX{hVmNr&1pkM%05<G3c9aqX
z39hsk>34>JoK!X{QQtmNTrQ)wr4WQyTT1?{IMMez9OIxr)~|2PtN55pwp*boijrsO
zq4C$FdaD)`*EWz+)RlmD#*SN_h9d3vW5WC$1e&&CYJDlfnnz%EvJ{5(<-BIC+orwx
zf&_Qs#=yeQqx6`0^NZX<vaf#JTnabTFj<+|!l8>JGcB0^T@B`Mj^4bhGVoY?7yKG}
zX1zHG^?39lb{V=Kf{Z6hBvU=GWZj6F`GnAFHGsaNk?fJ`rICT+<-FvR{CAiBBezwU
z^+;(mBWod$y^a<=fi)uT{1(kRvP~_Bx_G*}tE=uH{urcI7|iC{_n9mZ%NkJE5Dd&h
zxMoznrX6Q(11tFRzWAH#yT+TS(4K_EMOSH^0z34K+XKb@I#wgfw2M<w#VGUZN&`yH
z6p|W&$y4V8w%N+HRn$HsKI0O{zK;t-#r=H3XHnm$tD$oYAKfGM0~Q`1FhtPPm`f_F
z6R&RC@2z;Bb9CPu^038#u&V49poSm(d1k9^el>z5#R-g=rLJPKa-o*nnpyQ0hjp06
zb5#NQUq1A&KhBq^LAst&aP+T-tG(}p2!s27dH%a3ITF~aCcM*{(ELGPOp&BAgL{&U
zAL>rwl)TQjAQvh`7-lHLt-JFqBkhV^#F}c~Rm2B9)NW|ROtnJm4_V=TAiK8gus-5d
zFFp}P|2q<87?p-g#*9kwpL{Un(jq|V(+4&BWNk^>f(UiA{XgNdkArN!3}AYhtO+Va
z)|#HCa>gHUK4%OkjhI#Z<`e^WTYvwZ7;nAk#|+KS&XAm#iY?NLr%9vRrfUgHPC~M(
z@^V56clo}f-}<QNh|3(g)mr%<e(EH_?(+kp)(Zu(qDDAM7~vg04yaX?CqAcD8}t^5
z6XM&zMbA~w&6sr6l7Gqmb#1=C_xCV2<-EoW=D0-Gt{04mfj|h5!gQkmNYDKHhBO%r
zmW+)YAh8Do?ERHaTMjs=bvt&zTg!+>Vou|tL6a6mh0mTHEjZ2!r{uV)Hnp5$&{i;o
zoC5c>8@ZmI9{Z<c{r3T}dAwgw3?_ehq!De4Jsv;8uPLOL+l15Y4oUDnuT&QeG(jRR
zU_7k(-w;v6oCaa-IAuq5bfnJKB6Sdbxi=Z7Ew$tRn7uNJ=0~ud$d2GyXu<C8g_p4J
zE!TUdRpY@)Yd2TyOJ(NQc_-wnjq#}q#<+4bq7D=A-6p<<Xb#X^)MK@ftQT<WYvD8H
zk9I9458KDYox?-$uN*Y|TJ}=9$`+Dg2f`rWQi3et$w_r;ib2x(GZyXxY-^NcO*joC
zDdPH7sG<z(a#R<7VI9j#ld4;Z@%A=V9P&`LsvSKH+B}FKgRVRtkHJF(u}}A3Q0|8H
zxOjTgW0u6UY9B6=%Z9mR?OCxuUehCOT={L_|5j8C3NkHR)gL2UhKTbN4+JOA!Rjmg
zJ$5|8D;pl^Y#)nF&51lG+Z-Yk3N=MhZA%I9^4P3Zxl32=;Ga<Xp>L{`l=Z&|ZE$cF
z0i}!rLf_|UiDRJhZXv`43dI!gP3(FS97%r`|4#e*k%aO;T=Urm+6b%=S6}~{8<`va
z6R!?GM^)Wx(E&zt9F!Os8wqL&B17|hO)1kPH8m*#cLu*}p6CtjHFvomI@-nGtLP`o
zeq<}sT=QXubUDI9sIoE!3vvN3tk&e=Q|Iso^KH&H`)%Dqxt79dMfn%{a?8?&OCwt{
zT#u!`$JJqR%hscTs3AjPsBuje`=Ov#w_dt7<E|;O<ZyiHljZKO1JN^Q^zU^?iHFCU
z4LM(W*?)ahoj<Xr0jJpiCfPD`eubd`?i{N0ms6}qvtVjQ0^*@^tGqUQSnJ>oexr<^
zuUX=i&k~R6FH`iBK&3&CrY*qrF5EwCXv}?=Nz0ZYH!>&>KfZAoZ@g8yvF!TE`IRRm
zo8Eq(Fhj5-*xVH|MfmrS)O_y!xv6YeBF#`Vkg1wm%79Nq@IyzGf-N#+aKNLAb95&x
z+8+qC^-WA9sK1jPxR~9s&26}qRvPbYMR1$ZajL#sJ(<ng{c7d6U00cxr-YTx{J9Ur
z^RmRrVj+IDI&TZO0bbXeRbfJwbRy24g2!@trVuf~)%5FXSBZ(QFwf?X0B-@rObL|H
zRoM|-06hbwpuV8FO89FmOZrVhk_qM7BbJ<S+KUeN&i+KpE#8(N*2N_HJ#7AD?S^fY
zw>|fd!NYCFoUyU71|?q-KDSVn=#m(TXXCo-=h{O-%ofh*j!J7v$3^MN?`p~REp$hh
zW?H4|P5f2(e!Y2f`-!aiZWpWi!Pi!}rP7~aTy%^SpOE&#^-?XgufOxjf?g`01nO>F
ziD>x^ZO0Tz-bx&U=74;gJ$6d%#NOQQ!&=xL#G#-gA?i<#p*)_-{tEJSmY<kki+BJw
z1)#4G1~BO-3=2ZnK;Lyda(ss&?gGVZ#JvxC&~o9t??9wO?56qZTis30RddSleOAm2
zV!^W3FWTR6|B0igM07G_%A|lFg{1C|5N4^G3@1&=W_8?#12NMLm)466RC|3WIgl{{
zZVN{*)Sa|G@9i-Gy7|RH59rETy}CXdp`N_BgibMGygF1_d!Wgq)5EZS@b|B9NUKFF
zUVd8+hmzsFNF9&=B8--ahwp@%ZDtoQNx<*6s}RIAh^&_CK^eM?8d7RATGvVun^()#
zw#S69E<4fBo3_$hjXiMX+wj|de+H!7x30CA(Ixg)*^K+0xsIq@A=z<t!Go2D%C7KA
zcRdKqnkS<t?)f(EBt>C2o)_O?_K3t8cgb*Xn*pp(^mJwCjYvJ1a_f@olQr_H6;(A3
z6`c1+SDbUZdmpNbI?E|Dq^wEI5!ZFm7=Dde)n?)^TCsS;z8U@c=d-pIrxxFvg{9Mh
zcW!MFz1i8>1Ri|Wq47Yzw{A~el1Ft6R&DqW-|uiyLFu!VucFeK?<%1<aU6f(XG$9j
z*pR9oy^liO#~ta&aL@BC1*vap2C!Uut)?A<d=c61!I4T7WW8Y%!1DAhlyS0QTpxJV
zZSe`cOm*GLlv4;Mo+|g1AC&!)hnNTMO>pGfit91{KO~)HT$AtH#(yfJgwi10T?$Bt
zbV|dhfdZqYJBEUEN_TfkiR47-2I*!rqXvwIJ-7eISLzEs+s&?P=XspR@ja8H)3*-7
zS#sj#n0M5`LEVl?^>wF`4@0|o)|C~JfWwM~Icv7TPjBh1#d1LmC(r6`ucXezRXtbV
z(b>*f%GN<yyi~`_oIexCm&VVoZPZp)YwGzQ4UaL+H3Mvce;rI^4WTwWdtriy$Hx(t
zri!B#{aQBIS7+cE15P*$;Bwp^*-&}>7b8xgP{HT5q3U?*y(Pv<-&$E@*t<#^g=Xp?
z(Z}QW<T4-DCn=VRFXYwU%N3>KB85=$0#?7ic|-|1Zgs1wm7Os!N8<?}GjFY}vM;F@
z67rtWkUu&#*j~Z<bb20m+_0TAnEqRtS=pE1sXI5f4trbvhh-W%o(X8~WQ1ZXMss%=
z*u!Jm?1r+gT3VSZw=CR7Fn}FLM=cOg%h+%~^&9h~0s{m1&5bDF$^CH@6RuFc+g%LQ
zsBB5$BUl>^^ISmIvao6@7g`%>c4JP0u9RAc`}db%sMviL5ydxNO3&q5KNIi%wZRMG
zAnnN3-UT4K^(MRqn-E7^M-w;-h!+<|^`=~;F9vm?PgV~U5LIM^9wTC^EC=eeFsZ^Q
z6?RY32`D6V`>)m)&C<5~q6SF>?@mljCM+ZPh&CpIle@rvB{3#3z+L`Zek~Yc?cAs;
z<|T3Vhl3>LpjgzG$Bq2%m!8pN>q=D%EftjjtOY{C&(>rdlz$VTD!}S0OeHwfd~ZsE
zjteZ0l)6(V*^usbn2!p)>JWYCL3<M*M4MBn7}2NAT>9`J8P$Q+=RO{_HCTG=>+c(9
zf6k@ghv-|%fv}<%bfABLu<zmimbal3J6#S^6-&ZWA!L`{FZNQ~2S~SAot0cXA9&FT
zunK^@@%ee<098k4XDkqtdPRU70z94gTyF1xkq$7RnlZ7x{X*ROmjmGzAxXiBhjRI<
z4c`K)PE!)=6*c2)5o&P!Yy05Uws9<aN)3Cf6O9?QvAr9>1BrScBtR1fB$Nc(C@byG
zf(n7D=?r~?IcC_uKkvcrX3&Q-0;w}klCVD`-X(v+bNyK18t8sH=r!Vy`8&5!EeKiK
zITw}dTS;c-W6)C@&=5;;*}6s7A#(lF3-IXtdv$XRg5-t&Vru(sz?%WGuWaPpXkcc^
zB1uH>LM>H~g!egQ_ZJK12cM$svl+?b_b$xykR`rNEp57XUN=>~SqAgO<L*1G{HVWw
z|Ne{Hi^_%^MFnLqG_uhU4?<R_YNH*QMfQ!D^P>z4K9%1IolfpmEMVMvc9sc#QLZvZ
z>AY6?grH!JS<v&ad{!bHG+{|#rBV8~x88Ed+x+<r!ZJQQoEJv!VndvR%K(Mz5Cq=Z
zFH)9LeASh|%SInb->rN3z9CmI;bwr|-y*K~1KsN0mNVPa(|c?(K*VU`GLbx~Ln$G|
z_g@K>YJn)XimLVZBOqfw;@|%%UA>eWW4)2pEUCuYzwjZMIMR4h``c*}z6-gofH;sZ
ziPw=6*xdkR{2pt-7X1<rFLNXeKD6%qY6IzADU~?D1f)PHy<|};6DjzteHjB7QA6sX
zO8l~DS^Af`X*@RlUSdh2W9lN7=I_22%?EtO#hjQeL{kIb5EQtWBesU88&6eR9DIOt
z`<&v*mquNtcEv$NYwYUX&qxHM8`-ryM}`Pr^=+i?Vhsq~i|V*yJieePa`neYx*eax
z$yO`j$Nm<#-jX--MYOhkE#~^ATudYty$m7j#dFb*Dj|e?zdufBxoz}ZS4@GRWBL}l
zu<hT;;SYpwerjvo^=Uo2<=!MFupL$q(Ccu0n<L>jb1oY$bE)(N_zi>}$=SBX`70RG
z*UBf6$SEMy2vVaB$`5d3Qwnau&L!^4&x+wY0>1Z#0Y^KC4E)cJKf6MJxcXub!+taR
zoo?WXM&B-u6p_{^f9FFxhr6+GgLY(b*~40H>`Ez&s)P2O^NJYN%S>M&y0h`8-YoFa
zm;ilM7(TQpIwO@9hLG>B4dB<%oP?ntG%hUoY~p<Q!mmy<Kh1{fEC1#wIfJwTl2fz>
zY0M*)9-&uhiNERYX3liebQg5ls3<{R{i@{0mhE~$TmH65v+qBi{*3q<>`1&gYfVmu
z_jRziFbLpZ@QFXKnhUuKQ+-JH_g*vEojD$|8M9{w6&E=JZXNAW!@5^xi0t|)o73#&
z)1*e4#f7)=2Zr~>DKx+@5^-s9b%s8B)qZ`;fcb2~ctznDDYUUIb>EyJ!MlMJ_5}_Y
zblog&h0-kA?scTE4-qO6557J7;KdDpiTBd{e&#rC-99%4=ru_joE|n_#mrRy>}rj@
zazhV@Lf3H!0ymeTcg498KZR=G4W%7dC-TP6r;cH#eR%LO^xTDk^-%cgov;Qcs|l(0
zAcnNda}H8&<<+I7Zc9E}_OBIoxhKV*Tb;l_70BX60%|aj-+UpX<*bLqds^7H;X<@Z
zm(7=Njoh-`d9;%mEy)ZD`@M})`meU8hJu+vUUN1xl?s?nD9|e|J^k3?s(7hZhtnBj
z`i9YQi0pAnmX?dQXVEyHk?=kTJ064eWK$*JoJd#D&Wnb%@Rme;pc_^t{e0?)T7KJx
zgLzw!m&5hXNB>|A<-l~h>nsbEGI;YUw(xR0u6up#UiHl_3hk#WRaSX<q#=hMX}hfm
zR-VfB<@L>d7${poqheO2X!_R7u5#s%AFLTMpS`&$&lL8(LY1MTV5E9-Ly^-S)UXTb
z?!b#`QRu;l^4<K{*oN7yY~d9rdg~K^3Xu6I{i}7i7<kU{a7RW~SHalnLyEj;j2-w;
zDDW_GIg6}3-oa0fUbkJhZ2jj>Davh3Y-8tC?zNgs%j?+IRjsMexvVgqkm>)dcJv#6
z0{XP@O#fneEU%bO^Ua9}rl(KL?|ysi6BR>O@?<2z=u@3XcUvZMW&o!MxFee3s!M75
zqHZCFHy?H63QYO7)G}NpRe&$Jr0n>M&Bexu*X}S(S7&E@K(c2Kb`EN9c-b)5DXYvL
zHk0>S3rLtz;#@9RGlV{?zAX?659USB3ws0U>?=R?(e`BUm^|)5XP5Ljtv^ri>g^#A
zUqFtI-N}dt+;^1uN*T0>q;+>KN>L1Hpe@g7pl=nP1z!;zbHQ0xtKiMFP9`ZH;GV#f
zWIf=aDc)T^eBgb^Hp`U@bjF9FFVPR)XWsaM*ByZrQj*|qR0Ub!9@>KncJsjZFqheI
zh4hQs1$l7kBd&TbE3TTH6sl$*2o=b?Eq0d4)7+{~I5ohkh-uP^yaq5Sfc?2TIUdOD
zU?c&=C?Z~X1qrA=4Od|b?xsrPO#FL{=Za>>xtU3)z&Gb83|;cT_E$SZTKkw6@L<*`
zcPY>Ln2US)*h6X%HrOPS>4U#~7-fl$xGh&?L3_GXfDWmyp<yAJD}-+llf;q{o!{kF
zo)S#g&iSIT=&6V|U!*z3R6B5w>uYq9{Mq`{?YvY|D;0h<<>zwJ63F8SLunL<-a6`f
znsGT<+@CK}_WB?4BM%w^8TL&x-8;JI9yWI>ub$MJ_2&9GQJV=@oAX7daWgy&^bw#3
zd1<?y%xL<&YD~MWOPY~V4;xaXu;vF7{|9gKz&#Y&%m3y?kY*GUDRuu053aKTxuXp9
zCPAMg1s*p2W|2C%M!lBfaflvFpSyIJ59N@a4}4&p=Reo?mL%i{e@&c^&a?K-BiqL@
z%{=HAR9^w5+LJBvsiDb}Www*c;VtL$WcK_7b63$iM|xIv+%2>3cxnd)O{tj*Msf}&
z#4OM3p7zAhyMYB9IjCd*x|uF{TX%hc#X-mhKLAfI_jYLqz;JD!+gz5<VZqe<cX=&5
zg?B^72m|8F`II<ANf>}kB_QYs&Pv^6myro+q3{?g?9Mcb1Mv%UuR^ZA6)ye+%7>jO
zYnM0PVg(fiM$5hHcP{Oy1*TExDR0k6X5j0YOw3GC-?~_|^A%`g2YTmap*bz(>?f#4
zOD%VIu&8&}V|J&V^>6^AT5Wj<?6s~zH(agWT^-Alz?xfx7$nnjyUnhbX=7k_VNjt?
zcYpH22TVg;UoRqG@yM&f+@l`efVxN+s>x<Ir-he`SxCX%tvO*>Evmd&mRm^WI=n^V
zO1U{iwSDwN_nf>iaZH#6J2Va)z}w>3+Bd<aMFD<&AuIm=QVvpQAW%a^gp<A_P2t9G
z1)hEJ$*>*9VQ6kSAn$Tb|AIefUQbj*1pS+=6Cdz^tVSB$eimj^eA+%1v4z5?cdvfq
zcR$LdqC^1%X@LVRGz5u8Ps3>RYHOeUMM5vw;5)q<a48>GDTW&YlxU|(06wq&#+Cl@
zVR!}Ev@c1i29%Qr^U}iSPQQDN?!odt4!%l>Mx|YxHmA%IxOkkKn)u(u^`TIZ_RB-`
z;_@r#$!#G@SlxnP!}QWm45%LPM)$Ao(GTyrK#Q8`A<u&cztn^*J>Mq!sbM<n!FdRk
z%vA*BNG_E(om}N<+6uu#>0dW1a!J#av|*_qykr%1_{6dEy5Y=B+In7<?5;aszkA0K
zlY!n7U+(?OMd@DFYXFp>0dc)tlH>4j1?jdS21}iUFdj>3S>uGaDH{HR*MCbjjDHXP
zmypU8_L+L~KfWl!x@<4$=infasU`RG3A~3n80iF}6uS^4Pepxh-elU>%*)sraP8|L
zC&Ey{0p(u-ZWS$g>ReNQR(04J9vr^mq=?=Zt<jL86$;CZlyW{b>$$qzQeiYNq>AN%
zkt@P(N8ra!7uzAo#x}5?UV8xX|7u^@ebXn*(`$x~a2h5IBW+95AdJNZS7s(%hKke~
z-N8sd0UHRj+<xu;iOcwn^#*|i)t3qwycCQ%vIx18_E!R>ajLoZ$*<xnyDedV#qOE#
zAUalMw?^$c30~ga^0ECsKKw;QDe8W|TZiS8hY_>nG_>ZT-_O-3NO1iSj0**rYBv6F
z9F9?!krpPi1Uo`UJbD*i)W;XR@Us2BQ#UEH20%TxdmAnvNYxP)x8cPbXKqT{-PjtV
zp}oHi`?Q=gaxWd2_!l{3=EUjt#-dlV{(Q;5<eK?$5uL&kJw~(Df{R*e!sQ?zK>G4N
zY?=iaY2E+AKlRR|AY9Sb+pUd(uerIi2oFxJ>dlLvehd^^*Ger`Zzv?jmT2<X^SKYB
zBIhh;T_ZBchxRG?bLXABb_`mR==%<NMp(K1I>ctghF<RaZrk5t?D5hQ;gHAa4}vMD
z!9kX~f&g78RHKvpeq;9U8=CXVtMcYP<cUAQsw8hmODEYXI9s9HR!)jx#g?-GMvL~p
zQ@96h!1U1S=LOsc1;4VfBUx&?>J30hsjJ@H$iR&eG)Jd<V}#3aD#2=J861bzBl?Zb
zY*cuycsq`%>s=A0ZNcqk@DSe+Skq1<WF^>MrH(d6A?8%eSk4Fjw0cQ%?Wd(NycRrD
zd+yikH18%ib~h4?syJI)%QK8Du3YbT1Yi#rX-d~8rtB|TTn@s5DwhhzS_b1=IUkS3
z?X`m3k|$hKwfCZ309m7vVQ(dU_;^_JU1e8AK<6XG#BrlSjfZMR@+Hc*gQXJ)+gljU
zzIgzledTIl1aqmAjVQba2)xF4|5Z01^m>iAFYcl<Ro_gJWcB(yHwtg{S_j$pFo7<L
z_|%PO@(qm$bLus_U4G^g-~&{8sHm&Dzl&G<4bpczv^5)2SBYduekXaVmqH@c=o7!n
zt77yn%l^-^%4|%_|1jAZ7p&_CwK=dNI8BCAzl~L2>ttff{SC-}=fhHop^Pebwv}Au
zQZBhn$X5rkxHD4iX|4|}3^A`yI#~_d2m_x6*=HQA#d%S@5p0_ka&m=P;jMZwd}5Lp
zpici$)bduCz+V*Z05BLP-`Lfq>g(&{?sZZ9f!CU+ppQ(Y_mOD(0Gbihie}L?gq~fC
z!<Tvcu5!^)TBzo@I7-wEa`j=Dt@Aaw8;z{gCiF5`KNOCVayc2fS~+Bcx*<K=UoAPY
z=bBb+IN_r0?($cTd)f5M;Rdq==+xr?!H1QtxZc34Vi*~qFqu{DheMr+?5~>xH<%tY
zVEmClarQ6Iyf}tHaXd`<bpW5bTV!%r<yEp%RR-||xCgAzb;K1BC9S(HRQ9qrrzqY=
zPlOEDlh9?nXMhW~-><*uZ}Q^)<L7XSrSB69t?)IwIIiZeo$0baL$LkpVUbwZE3~~*
zyFcWXAnO%S4wVVLUAgBW=S2>XqqN@l>E_~-xy+vcOPp>pC-@2uH|NdYUXvGAeAQ{E
zUBPKzijy!oJplt+M8vb)c_r(%g79k}{|hg)Le_AgI4LVDa_dGI+Hf;jQe&VfRdL03
z)Z;Xr8=FfW@G3RcEN*r8-hJv`4D*U;+NoOu)k&~nH(ikdZZnm;8+7s)q5%g=!Bq)p
zO=95I=&?#+5s?V=gvJZ_>fLGKmD$6&kN=%JHTpry(D4v_wdGU=PleTWz!@I#bg0D7
zeEmJ5CV$W3)w4nG4|$w?ZeC}D5w0JP9oD5MwW;vDmqgxHjpt5)cr#}Br-&S@c0Zj2
zN%W0k&)^oM0C&3wzitmoO}cfI*UPn~Ch!L-3Tva<>kzDvjYyLTeGn(3Gc7r2sy5yG
zAxVB`I_Y^cx+9Jcsu-IHYyJZ#l_PDOpYuR!eP#BXu9W9OBVYt(e&^FyqhYk&_qUf8
z*F=Fl@TPEHlLvMG^_xpL{N(Ve@2?agtihWx#{}Gc3U@i*@ZsvaWVrbgKF#;=tE<_Z
zYVpfNre1g6!cQ{v)yMAIPFO7LWFvDGdc0@l^n0T(Ze#U&%<l~>qZ@`&+(8~c;B<H&
zH%798g0BYuc-#%%RvM5c350wV_+at{Ht{LngR#TksT&9#pHYQ`r6ktig~0z^9ogSd
zlXo_TY0?UK;fb{Uyx@~ei$lKGHLt?U(!=P#vNZF+c_1b#L4$k@6i@SR^y1X-rGSPj
z+$NR}&_nEQn0u<D!Ev7~>&T({V$mW<6b-ZddM^|N5YfbU_l(-69<3m-f0L-{hs`F_
zN8tu3EJ@35aR9@uN<|oDW#4M_O6kgCQ4+6HA@o!_IwwG4-;@`~wTSziGWiDk0Z6xX
z(1W!9&N_&Z0lpTs8k?E>0D^DqkuknpjWSqPFp-SCx^R-aziUHk!0yzCSH!~$f7;%7
zHHBYYJ}6T+SBgS!kyeTaK;t<X9Fuo9+UG@y28{K0v=4~NSC@Djt9O5dkYs^Y-v{o0
z##||`u6Fb5BU~SXGz5eF{1v+1)4Y41o8#18O0vUlqb=(0)%dKpIm{mNWqhu>eIJ=j
zJj%MPX=H(E6;RcxuQCO{-*jRBkKu*Yq8*hAm%Hw75lFP0e?{US0aA|9a548B_5n3$
zS`&x-81mLYUPy791KgVqIo_J9a)0|<nzQnCuxk3nFw4)O$g6vXUQ50RARG!%)+o_u
zuq*pI=p`$|QNaHS{L148eJ<lKb$YmXtKs3{J$<X;D2E>epIPxN+Kn71|5kFC15Npj
zuSvk=r8z3QwtMwKf%hRS@FO95DHx*WTbaf!_<W=GA#mTIdQoXl(QNVVe!uXrdJT3z
z@qVd_LE)aGV$L%1VM*-%s+jkJQi!?XS5P;upp_fo94Go5^>dN*o%}BWduozj4YJcE
z_!!)Dx6DI(^F?-wtW_<GT7-@qxV_+dmjoUeY;sqR5QS#tL0Lt9{ABM830rou_{jLk
z5bUw!1QT*@gcUp1=o9V7MahAltycY9{3l>YZQ+xKSWBC{sYiBju#LQY_?JNz9bFzd
zf{Ax;i)B^IW`u;^++2XnPp>WoVb!y5ZoBe3^lHie0@Zriv%<0+G8g1#;j4$Wykr#6
zJ}x7gW_1A(Bk<nM<*Z98c8d>?7)JE&CXXq*Q;(|QLQ<xEtM~fTY%fp^Ao;r~N=ngo
z!J5LzHaF3Vib$!e4<x}AWF{RVR0{sr)ux5QCSNXYGYkKNjD%w-F=9nN_HN~|^0<5h
zKMENH9E0;!N~%5A(_Z-*t&Yz~{W?tDSjq)y60X_DTh%SQelcHb!^&xIGO?twV#QpG
z^V;l71^p>Q7PeHvlEQdjDt%KXAb6`p&;m!25a-IqMeGX2X1aB?dhqVJgZ82$Et4(f
zyq->mVBC=}9|s>@-n#huE(B>YL;lbug`4Q<{d3>WtfUS>020WIOE3We@K~O~PHGiV
zVGqAcCWHT^-A$qy9Lb_$^!zUl!+YMht#5s*%7uE*Zi~W?p>uKQ``Sx5Xr<Por9K)~
zw&8RiV%jdYY=SmJd)*+LAGV<96=J@3+l8Vv+`T>>;a4MJi?=UzyQ|a~%oSLx8+P+7
z@O*xB1Nw4OqYXb0NFt%(z@f2`8yanyBCN$Qjg<2kCz-U~7PFkm)i*+)gFk6vO;8{V
zIk2!e*G1Rch`9mF9r*@%t47@mydrFuGZX#O+S(7_VySA!U^nYZ!9zo%Lb`)caruZo
zAZjoCGxYGg>g+YIEOU+A1v1wi4KOKwFl<Bd(nGqfr;HOzt%;NDtLEFJ2>u8KuCmOL
z0}q7unbW-uuGH<{Ou<gHh6QT(zRYPyYZg&Pc4i>;V8sPH%hb|4?7$xVy1&8O3F;Z^
z)Cma<bl2<|dsq{Gw~VeluHBf0!D5FNW1`%AjQeT>1BvLk-){nMx&xb`kT+fKej}l5
zcZ+h`-%(zX{UiPJmJlFq^nxE@&<;Q+BA=Cw@$}1K9W)eTjdn2NxAK_`{JpzMV;3&U
zuBBn;q2H>~b?J}^-j%Lh0Qyeb;?zXe`5JS%g;ZC-RhX5O!~OzgtlAhWoTX3%W}lHY
zo!$Yv-b~gi(&?fSxIg)nUM+*Osz+a7j!Y(Kyskyo3ix${qZP=bRl3P>F<o3;3D2(m
z-JC}#kL7U{4|~I&z41jvX-VAr0;9m656S3HxRSRzZ@ll_>Yx|dyApSHx-MP?Lbftr
z;(P7~garH0byP1Ipr<h$P}DO;tQy4srVr$>D^n7wYXuV6y?3En!M=QLvtMB<HhE}@
zv5NYD>CJoGt=f0-3@88P>YF{U1aKEb9XlVTfEO|pvL^1j!q<QJiT~Gr&?w0h7CUx=
zaWDY~_oUv&-?!6$+OeA|U0g0JqshNTUj4F=LZDjbKkSD0?1UX*FJrSS{wQzd&mgk(
z`MCJMdjm_TL+1Eh?<+@c-RG3|t+yt+;><D1GL>u{pc&mUjvH}c=zzLwGgfV?Ymf1M
zSnf@jLU?dV-NhI{uPEQsrU{VL`Me=o9qe^M&lHZ#Q|WA?hf@e(_dg|Kr(0uTXECDC
zt$-NoMbzTk^7&)jVJwnyjnp8zO)~Ag<K^SajjOAlO-sAiq^^ENN&;iVYSTR~6DL>g
zs_SQ(YlOVo^ga&nbN2?nJL&4%8CyGyznF9t;Q#dRmbm)P>2R;!tTMKjyVA!8gkMa>
z_AlF(UQ;G}Xg|`*$u=`7@dwqUcJN7={BixyPS`?!_J~PFnP8XS%sX+~aGy0F=nSi|
z^iRG6@KF>Bwfv}uHt_8}E3I6|)fY}zvAF5NOAFNUs)MxM#OIU5=<KIuy3oT+kGns|
zv!8@E3q^r{d}30V{j4T4wB-^-xU8%Jf*;X*f6r5yaQ$^t{A%~Pw2u^xS3INy@s8fh
zNX!gB13@l%95hB<dV^4i-q_Zz!0zj!E3wH-@<q~&F&C?r4r`#2#RtLdk+5$MMNz#u
z>~y4~<-O_1h(Ir5GY&DGQA3RThDh=&cd~BscWR5+ni+bKj;VcSA*G-mx?Da}z@)+-
z0fap4Ydc5@CfY~eFqjL_^sSQx3eJ=QbD6@+jhlP3IB6IekQaKNE&9_7;NSxKu)pE^
zi+>6BHO^}FzkM9QC;`!v`cDp|`IOXam#x@Y?ku8zpK^Wo03KFn&Oa8Nh(xJe^{)|3
zEKIS?e(N8_qn=UVJZk#@YgwEkNjJ9Lp}d&4P=@m2copo^T#O--C)e4*Rjc5{@HV31
zfU}I6+pDtSkd(5j7L3swhJ{<*t1&Rq>ElH=FS&a-ik4$53j%~a6`zZqvr8*p{<UeD
zw>4$_qV%1II}MAP|7MQt(45KAtX6*leU;mDx@ZPyEgo_mv8WrKFX6Z>`5yww^SHZy
zi!wk_nVEm0gvR7uOvqj7srRsx-C?}75p_vv#cm3W1;xbE)|e*8t1wzX4F^WP8vrX<
zE5%4;0)fjM-?75-NKGY65k3Jf_#Xa)sxmE|Q@$vjGD|IeT(f0z8RF4IIEz40ES%ro
zTb09R5pM@vBKMS}`TJzW83Y(Cn3#)5SO-$;9jR?CkA70x@g9}59B3gA80`uCFctM8
z3U;B{3moQR`JvQ%k;cw44c7bue;9<m?np+`4JyX-O{k_ejeS;mH~R{gC^$ZUAe!nK
zN#(<7h{o+ZnQ7&WB%7TS+1V_8OUuc*YKaYUiXDLFjEfn}azx+tjc^?M@!w4OK+xxd
zW^=OrT71gn_*rQyFN*Kfld2nb9i8UOru)<U1c!I{F*>E+qC)p+ujMkY`gO3O`_LE=
zC%fD$|I81e@Cb^z)#x}r=C$_0);axwSLFxU|LFsq$rn2-Z}8mr8gPE5B|RBTdKWQ~
zWy=@U#dZOJ5~Er&t1qO!rvbLEbfslr=4Q-^hmAqFGW478M_1vBlD~}JMWT^p@S>!r
z>DbXBK$Q>bJ@{@=YnSu^7_qZ}rObnsigfkD2oL4>$KvP<&gin2L~_jb(+7<LcFivm
z5QD4XMz28mlX=sh9>YDBWu8}Q_;-u`jED=FFVbuBQAPpa2u^8rkMD;LT3;^rZY7nL
zV%mDY(Z50}gZA!{5%ak2ST!w>t#^EBhnEK&>EC^jKa{Dcm&XL3#(D69Oqi(LUHbda
zg{~!EC>k{ywKj8l<PIjyG8DZ1a`Z#wF(*l&Mr01;Abh;_tC6KgNd3<V)h`a%Fq=y1
z#^}3B-Ynjl-<m$c93OJqYe#vRX^JR+{k{g>4h}P$Z$^!Hn8`IRkuR$<>gwrbWoQ4J
zf&ih&iozNpMNGm&@A(i+6B>r2tC{5YR#n|+#TR-Pu6kv-{c6nT@4Aq!x+b?C1fOSh
zHCSF7JtN~21v)-pNXz9EG0&Q){DkwVR1>Ls?vGX2qFv1`V2ipSvU#dV`WNp{#9S=D
zr3HJ;s%@XW?sgGbbwOH)z6!R?<xp$?v?hK!_BU+ZE-6#8Y|D_Qir?PR-XW50yoPOg
z2(P2J^xU3?My<mI-}=;)oIq+k4u(=4V-CUhnqb`LZG>an8kPJjn8)+-&*CN^;ZbP5
z8}IqO3Vibx!J)<`{>GZh!n+l{C0|RTOq3AZ(ePq_p9Am<6;vFhz<|Dj=+D|$Kov6}
zr7i%L9*n<ptIpY70R{sK;l%ivYAoCf@sAqbSVtm(5pDG>dqoZ588XK~-P=h{z_<+$
z56{?#;{TlrvbEuW7Wv)8F=Lm5;yI-K_NWcHo$1s&KX2%NmbvPHVJZnTky$nV{p;-u
z_0ZS;Cb98+&?z${1`!71k0-ap72z+8EI1@V$L>-axUs_#j)t|-#3YqaYupmm&-}$&
zaDLFZ+>oQcN^fj2QQS|M0&VPgZg#>JG91Pq<TY2=?U4HW+jV1w8AXPE@+E$T_17vy
ztMV&1`MsIM_|#cGD_*R3g2>CG<we(ktk4>+EK*aZrLVm-argnd`dC`Z5w;{DT^16x
zo5cWit}!WdA8!99pe6N8(N{$FHcj~1u#3xrrsk9rkB;>x>=!_-`m?jsG3W;e@a};G
z4(FlihHjPWm(gqgm3EBmjifgPRA7Md6&V>ivm=`Llys1lBy$L(R_Dt<Ki3}}lk^tF
zL%GTl<vl{W8N*r1K;(!JaL%P3yzQ~x>bwF6i;0Unurv2?`v)g;1*Kpy9wh`^5v~I!
z?vV&K_u3DwoHC?}&x!%o6ZPuNOmoH?YY~3p&k4mNHZ_tqd-2#BZ}34Bom6E$nv;oX
zt^j}tXum0mWt&5<Uk*<Zk7-J~4Ky(^y=>q&vb@W`y{MaH{)10YU2N&q(GS`WOZs8M
z5c2+9#U1|0_`)d__b0)0mOZX63o~?pSXJMXM*m=V?+BCiH#d`5B0a9eEO2jk1wR`p
z0pfcszbu2`hr1&8y%Cm}?CB|q#RFSB3PvXM?9$H5Ls*56^>TOZ|Ei>76KIwQl>by!
zQJ&S!r4wBBkg<91d*V#XATpw-x7-39#LDOUc)>$ZewkeNA7;vchP&T1+4i`10FC8#
ziNtcH$mPv6X72i))xXUqz=X7iUcO(C%<JQMsxj@kSgLu;PK0V}F(Z8dG{8TDlKlNn
zX~M_S_1WJt`WFk8gcI3;U90?7vjOQigmna?ig7eeM<b8TafcxRn{IkdhLR)}oCd64
z^Ex<MLFpmxdfjx|RDio3lAL}b*IS%in5XIE%&-@<D}vc8Jb!-752;>dVlu({7H`ca
z6Pb+5^OgBVYj2kZlW$(|_OMdDh`n#rGn=NMGA_}7w+{GdEFT|=((X|R^?`xrXNi|&
zhEtZI7@*HLr?|K9F>V$5HDQ?Y{h|s(Z{O47K|<VzEn6!ONT^0QJP>O}wYs^CFJ5wF
zw&T|OWD^_~!o4LqNP?2S_}>HWDup>9qs1sbIR&^OTR_s|;Wl(Fclu4mBW5@1a=+{)
zNlU>cOv+E#l_ZUW5W0LO)Dx75Cbi0BVm{#pWuKR&HapK~fK+e~;L2qvsAdAv5F31U
zB#MJ+k&*f)k7c%Go4cPf^Q-7;c1o*<C_CCEQZiiE+cIdgo7aRY+1ij@051Q+pCY-@
zb$~&bW|yV)4zeqz^Ew^LgT%S6)GbXNU1P(lJ!ZhBq!SoX5iBQf9K}?uPUFr;l4Ht0
zy`KAVlS$pqr4VbAOGdxFm4?!0ybRd0^4VOmO^->(t-8r#A<v$RMfd)AD;UUHP8_bU
zHPr{$e6qv{T%O&&UF~#r_VJl#d%o-kG!4h?XQQOr`5(&6Bw6L^XbRWbwUu>*ZT+Xr
z3douYmvDp}jI<_@iQ3o%1pipd`yDxkJgF{#`2$0Ei*(eg;h$V|symf4{(am9C{#sG
z*4TCgWzP$qqEF8xdzT4}%MXS(*lf8mDMYZCcT>B5*EG4)Dq(&E(_=mCk@NTrh}OR>
zi5fBjw%f4aXs@PYauLgJC+C%E46<si%h_-_?oIK|>^%tWFj@d?jPIY~UcCOW?TFdn
zM2-_bnU4^Cg$oCOwgb8J0cHR-QoxjP!7e9@RF1M8hEpd3XK<w-xi{^6$k)vc8r1d8
zVg!BGCS0ojb)j_3^=i0<Y0SE}+lN61A6hFZR^WA$G-z5A8F?p$UQSOlii+O*!plp{
z!UyeTqN?I#z!dX{kygw0VGlH$1?(S*J6+3p`<!?`2%<r2><kp4C(ZHJg91AJ(e(X;
zsVZb{WjvF5uxmdLkN1t5hPAUJ1osu0&(%=j*N`{UKOXz9-mY9`wB-G#DXhdOp&_}u
zAV1>b6I%0k&PRK;?7&8oIJZBCGDg9+fJ!StlkCY*&AaK?_==0$aXRZ>1|kQZCTo?(
z%}EgnfPrIIc{W>Q4y)PB11Or+1MWB)pf4(NA*8iaj^%+IS>MAi5EySoj5NqB`q7pZ
zbw@joeu*mvv3)I{HT)3kK4EQUHyz-JGDtNm)E?cH*+N9CAY=^GO8$EMpwAd^#WrN`
z3m!C5ZI>%MXy62yeQ~{UUfPd>4w*w5&x?IKBmo|ieS0fNwIYXyr|^#tafck^kFKi$
z-+*T6*j=ksD$VgW{SN|GLLaI!&l)S;Ibyp$oRuH}G_ppHxa}MKXWi8oTol5yb+$g9
zwr{6%^P{xa^&KJ7`$k}=pdeTIDc0;5c4j~fvL{G7C^{2oVglNNe_IZy#u<X_2WH2D
zhE2EwSlsx%_bzYwi)9w9X5`Q4xB9M=KI0$0wjrNT;38JQ4g@=0SpUNm*)`e0X@ZXZ
zQwF=Q`<{f4TI_>D9U0A|ex$>hKMAXw?%eaK>Ed=qev{ypJpb=I)WWI(R~ze-b+M`-
zpF42FXw!&+N#c(D;>WmkF-KH0UI)*B<IC;GnSR_9frwAnW`Q75iU(j^bO8=o)?hVt
zy$rc7jrrZb|Bf%u4MH$0rN5!qifH=+7T|mnX^clEn*GQZAacF4J{l8|lyp4RBVH?v
zVp*_@0I1rY|CX9yAON?Xh>T_peVqQVcFlAks)H$9SF|RNeBATAn*?V@IjKVMT%6yU
znaE)_ndMs$<;y`l+x%9ag7!=POnb#kUSpNJb6NIS;F<r+WYRzNNO1Yem7xKvm?DsA
zJ;>3JA$G(zFUsO%Tq#-js2&O}+#3B*=yZEQmm6uUYe;_a&5?xYGv9)b8DHRBHN}o)
zQ>&PcF3P*)_3gXje&3?+P5?%1%bvN<w#5irD2l|=hEJ!n_&7QWa=A|tLk0@27h$d+
z$J!*wgm}luvKeD4UUD3DRiCS=f^u63;va!ff5B@GqYW~WZcxq%@xo)gV&S=;IN>Rb
zEt1D7!LQqJ6k5;-jJ1x~cT~SQH5lP92P2gQlu!S7{EVxL_?iUeNG{YKYN1GEHg`EF
z#`@+u?sc?!GX1o)Qq~W>e5^+xTWQLUosY5VTia<lljihQhvElZ9v%iz-&jBj(!FY0
zCw>v-0W~O4P%@UXl{eqhb;s0BklysqCON_kF*80dpbH@3E_N<TQkIoiZJOZS;3nbF
zETBp)#vKr=bx-byeaS@h_kR3~&^#yZ-Sy-zc4scI{tQg#^DEhJ#cPrzD2alAg~XTJ
z`H->-JA1~2qs!vPidN&`RS3Zp2=r3U229`#s;Ma>AKvJ2huL7E)RTa$Tw;fATrXSY
zs2*l3rnY!Ni@A~=zlBLJmI3W8CvpGhMjY7R-{355N0|sSc|UsY9y=SRbKe)eb~pEJ
z4T`!_I_DnC8~(b&8JP))2fi|KS;wMeveE#MPdU-AIcXkc-xSIwmS=1FA_e}Y?9U7s
z)JvSZ9ZrX)6;bkPN(SYD&;#m*C)9R>e4O<>3Wcv*##>3I0we_=dWA!vjIOQ_nO(#=
z>Ro<sZK}f14R(yyfAUh3#)H!R?*lx2k*#KKD<>@^;Pelvs0jD??+G)+qqyY05;rxL
zQXoe-35E8N%VdX1%*I(p*zl`zQw@V56|DE=&{G-`?eZF){{URp#~!9miVu+vn1%CS
zw5x(IAw<oU8qi$Vf1`vmY-P+HAp8iRHG!pWSQ-0V?^`&rz3rYVkSAr#0lv;OHI{!T
z(31G#_I~@Eg@LS>$KQldmv#p@P#q$smoDw2SDgoCe<Fp2u$4GEyRwKoe4_oB|NTgo
z#1w)xPJyS*K)WLDeUj&*G3CWkI!s;sbmw?JO=OjWCdz!~J6iqnkL_;YwT5FYhM`J&
zC5?7CpNdJWr=(uT5--!%?od<b@5KBcJcUEvlzVW14?lsoRUurL83v8N96M05UULZf
z^)2{$b7f+cb*kX`ZYzeu#6IrLptdkU6YZZTC4_tz%tO-bPgRt0CSz?PYAM}SH}O+p
z5iY0<Rjl00FK3KB95;1wCZRKKR3Kna34nX;3kG~Tz}KNqz(>7L7p>2c?^zP7xdyqO
zX0^jpTI?HAa6)U#P8ui$c2pla+S?sYIRW!gCe%K$=y4&C9r-c^Y=Z!7x-ZQ2%=5SP
z@5)4G-nI<Ou%}OaWN#tjae6QgT~eluTegI-!uw~&(@XK|I<&%rHaauBl7i=pm9hw)
zPct6X%kJq^IvOfs&A9qpMn0KN=H6(TCVsE-T@(GPd6%LRI{?y2ZW=%szz@hn`Q5LG
z-?e@u&kpH#?`3PC1uRxsIXQwb0HYz;DY5*ld9x^&H_Xan{fuX=#*!-CJzoz*bwC*s
z=Dw4_pQ?OMXSK}vP++p_CBbTB8_S<#70BQHZ6tI5+?wC%mu(ZyR^`D~g*POlJ3;$H
zsb9x?^Q7=2kBP~fO|iz|8`;o}K_xi{X1+aMA#;~x`Nvyga<<$8Eo$M~TA%ecO5+SO
z+_Aqrv>a;`!=Lv)iPny_l)G8Zz-!_5=n$bx4dTjBY5D%bNNe(}Fiz5rz72>GI4@1u
z?>Y4@)mUbSQkm&fC0`3uSD#4I)G9sNxwUCBl9%uDj;5rToSKO5``n(JJ~8rOWl@mt
zC7PhY|7RC+yvFv{a7KA$D3ru)mm?kDM~W}>wBWsdXXt2lmAIEofL^}d4yhkX`jXE$
zX{L|=u@`}q#aD-=D%HXFhxiZHXVP0e-ORn3RLm9Bd~e1?)>(M@^7U=KKjWsp3UCC6
z`&37T7K`2)U2th-k`?f_gljiCWL8QLP!@@0s~Zaxw+$~bFc??oIU5>(g48_jW5i%H
zayuwE0|>@`mq%}%xQ6jx@BKzdox3_<(qpY>T?gQsH$q{#9QYuN$lP6_dylOyC867*
zPisJF?W2>9A#WpdjG-0V1^j@2UTXj+3T#9HQZHD<`%5nv5&&DBfk}+=EsjdSSnr^H
zqxyVD@ic3}H?{Ea&t>lCnF~cBNiz7Z^x6MN`~Y_L-_|67#X*nJD*R`CMWRnfNDPt^
z*pV-N*f!D+asr5(IOm4x5*`|RTaPLQ7u0zz3J$MKz~;tsl(^VuMLd}E*^=6uSr*H(
z8(sjK7f9kEZxFS8fF$)lBo~g53a`$^%I1$KX@21zAbr$zOg0rjj*bCT6#^)WnDQ6E
zUEtf1<39`N{Xa#S9)PRlhOszrC%HM#mo;%fzr>m>Y1n3SW7~;FcJSy-FnkigLu(&t
z0Eev36bydbht38*+%|xkEPMpWd1Pc{KqXD`cTWI%QyM4=<j?OK0x!nk_m``2;UhVD
z0BTBDOe_lE$5(m+n0i1k0_H$M$2rN6;`+gx%n@6CmC(o_(}K7*g;*=Ujy5?dGji<@
z*^xOSEOI&883Yp;9eVz66FB*DoDF42s4_X%0pmX$&Y(A#v^8AJK2SM0TKI(jP!d=g
zVKMb{6XY9u{bJ=h{Eqo2NAW%7poyhleZp5%g|W1B4R$}X7J4cW^{0XGsYLdgBUpmM
zG{fE6y+v{!8Sb9U$E?M12Yw#ZrqA9ztS=4-KFuq>|1Oup8pOZ1urrq12Q)oYw6ws!
zzfU=dkGI~$$h=yy#!l&MgF7>=6%UnJ@@Vb(Ovudc4wfGqu>+7EA8I*pb0SVMzdIDV
zVnqHIumdo~amJMa__4h^f9|kt4&!9bm|3l_l5RV`AaKzbP&CwI0s#g9x$VgQ{NZQ}
zqEe`<LONS`Vu*gu$asWdM#R3Z+nf8Lc!|*nlZ4?p{VNVyFp)s?&qdx1eJ;+Uprk5?
z0rQ(n!fce7*$-|=*VmTt7;Q=5^#iHSZs3KRKGB@@@ZaJzi-?oFVf(l7;Q16Vd;hm|
z<%2Fy)~<Tb67$j#)GB0y8~M2JP^o3e=qw$87gG6^(fM>QUo*p>tWx&eSZX%P8aM8E
zgJVssRCq!#vLaH{t((@{4fx>guuAtFSLpFdN3}hZFKNW(m*iL7>mumn=j1Lz1f9_(
zaMb{8b=v|%p);&-83FR>kuq>T5d)>nky@GRIXej%b^(<sSz*fdm%}%nzpD%+G^E5E
z)a5*=?5*~mcg!m`5fN}F?T9BJKb!Hk*r#iMpkB$rA@Z>U6!w2={69c5HFx=1Tj57?
zTozZ`J6xX1Pvp$YWADVy^!Xq|B36>oHivG<M`?kqliE_PFB2DVT%3PTe*?s@ls5Qq
z)_bK;P5=9)s3oi;X8Ly0V1*_2{{mOiXKsIWB77wNT&9Z_>g17Ak0n8RyLr}_w`8o=
zBHt+k=CmY?Wu1W@%VJATi9dq=&q|Gjv7dk_>kafk^C3p6`T{z3aC#~Dw=Zx2m<j;A
zku&vd{PO7{)d)EnlX?%?jXhjnNq$U@8aZJ>%;xNuo!u%qv~hheuv6FX@`2R~&zsWM
zI&tbm_l(}pO4JQL`(_vdh0y}due$&FgyKj-0(8gioK{(UY`Xtw{pVbq!($U0gE%hr
z3kL<AnMFNu>|<&9Wu5(Iv%)pa&xo5*FKbWkT*j}SJ<?WZ%h_@i1YW=YeEtAJH%x}u
zEGGcVsIHZfH`%)aI{Uh>soM5E1>}Q)2)t&CRlpE#P|n!Jk=)Y6ijUkUU?<Y{PZ^;b
z`Gr~P26tayqec^4CJ~d<m)1zEWd!L|gfyF?`^KQm7(2zNyDL`v{()=6dHSUr7jB8(
zG+3bmrqk268Im$q%lU2)D*Aqk)#>kLNz!;67vql1<GcVi+?OdBx?@_xe3fJb#zzG=
z>+w>uVWv`jfLa!^q?Piwu88g@cK3so%U6d3-Nyn2VE+VE)P{S`K#ql0F7-rPohO9D
zS(8Hyq&Me6y@nWvVTh2^>9P(~sk?(-XR&!HssxQmun|lgP_G>`rAap29|(b%I1VUt
zJ$mBbdk=8uyAnpa;EX=Biv7Bv)>BUByG`EEHcwdLKLeJXa2Ivmf;IVXA|5&mlTcPN
z3GC{zG52G=2|b#8oyhC)w?f)goH}>ptw>*pvsZo?u8U>=(xMft{)S&RdE&<t4Y_U_
z*3aLZO>*3$YGbX4cC&ZP_{85f@oNcwkACED>yauCHvkLYk&}kqq@QvH@Jz7Sr*5#L
z*k#?Ra6Z2M7Gf+*&z9YFb9a5lB9ZI%?Rv=4z-jC@gZ(eqV%S6rJ1P!Vv>UQZb$#Ns
zqczE<Bg1^YduX*AJchifY2{nLu<od?X$?@5N`lidUNO}n37L*65?2R^Y2*kCuTNPP
zhaPKGET;bfSuBmsx=H@18kZpAPswv2lW5(Qqrm5G>2YQH{l@jkr0&EtJI50^^DIt`
zIwdvYY`Vb8T_2QH$&$*s%TN{;@ry?>pk5;I`NbQ^AS=I`>oM`F+xb2!yjCJMDgTk=
zpPQ_gS(gD<Q5NQAJ~5~=r*sBuz#(m@9w^AxOhL2uQ87S9#d(AgRfA$YGJURQxry8F
z$*0<cAddw+sKRC9p3*m9W+FNe265JbPibY_lj5{Q8za5U=s#Sa(?`g5IQ+_qR^M;j
zyQ}W>gbeRg5i1QJEkep0?0IyfU$}6Q?~f1SG~|QHA>yJ`FO8ij&!H|`#W?9n_<A6%
z`?jOS+hMJM5(U>!N<p4)g3M;Iv=`JzeU}F7H4HsV+C*9vdbQ92$zhibNsYBQrRvIK
zEYU;1!9BJcWxNP`t88$c2No^~OkF%x#<_6q@z!G%uav>5;|>4%Rq_~ciwC=3;14Eo
z@Yso2clK>%;u>{F>!WEen&LDhWQHtGq7ENrd1?8kZ!)wqfLQ6D>3&}2D=0_0oaa><
zmcj)lPwsQ|SOM2iC0#|VmkpiT(JaO<25)M1&NC<1_ys1nI+M8cyMIvX6TY72<}7{n
zq}F<71WbR?Fco__kSvkDU_rrLv>Cv#;AR9F`{q^8yb|3hNj8e3LS=dJI(R)@+g0-X
znw5gLrHh=_JtmdeZQ2ae#d4RM6F5^0KZx}1C_-9sY?C65X*hBnY9m=Rj3kU9A6E^U
zjX9dfgZAgKaS%z96PrWWSCgri^cGw(15Y&rM!72asW;SOyP4D)PYt8L!rTy@`(qbr
zmXQlu@qcRtcog3)MQzHgiSOAYoc|T!n6G3pk*+uGTppXSQI2qHn%J51S@cP)Oeq4D
zXv$l!Y`E`gPnw1oT(I*di)2XijZv=AiQp%L7X*lOSv7$r5;fBZ1VX@2@qRSNdYRm1
z;Lfn4Jq~uQZF7+ow@kiD9v8_i>;KT=9=Cs_t-he=@9CM6!%Z-;a+S*d$oj*vq7`I7
zO0@nL_+#laSf#gV)MfEkx%yWryIqYw_c^Fl#KJRnq)!h0AIaM*%m1>rZ%_7XPWiB{
zcU-yB=Tc9m=y{G0Bi+0ORsyUEKPPU<l)FhablrL}d~t}DRECZaB=5t0aO9{RnHwan
zqSr~QioBO(6y-Sc^y@>jGPxdA)d;na{!^Q3bn@=Vf5`TL2pvnyQ-Nv-*ugUIJXee{
zUt?BiUWsUJ8e_`TXkN+Ekl5g{nc-dD@hjki^h$E7f-2t{inTV+B0f!&6b<vy(7m|w
zzi6p+XDLc$F@U-ZJbsP8iAuWrQ*~CgXu&EP)7}_OF#(PFF!0z{s_eOL7<6M@@F4eS
z3;8Hr&sqvFP}>xByFa84<*o=GyA2)dM=Ii-khfYbjjf%J^!7?7rx{uVoNf;2bNs}?
zOxBH{A?&ZHFsl7t^w<*3xPbX*uh;Lo>ZVjeFX_can=0KbqU{BwEj|K$^t@+Jx4^?Q
zn|ER*MAMVwrRQh458wOleeIl=Yhg6PoVJ7DjM}2)W=tCq1}KaxX*@xFqa;W(pLybn
ze|Pi&61}0{iUUwgSS`XQ&u)K>R^eq><TBRaLS_<W-pg<I=2;s0;ClkQ0EwAY0lk<1
ztuCwAbDjL8B>v$3^hu!dn~yZ}UY2}Wn>psmJ{yoETdpZ0oP8&7KkC~Gb&ulx?bjYW
z!`XnVi_tcd1D3ACU$~C?u8Pu*fZTh^m$<R2gewhd3@e73OL<KsUqU!(iYGnzlFWvd
zFz-jUe=voQ_NUkLqC%_-d_Vt~a>5q&;X%k^i4d+^_J6guwG162p@$Jol!ZHOR@hZ7
zG1S%m<{g1OD>Hm%+anyln)>O$?eq7z^%X=2OI3FS6+yMdx=la;pFyiY1=;ct2B*5o
z0GPdLede8fyq0Q*;c7s_R4vc?09WMY8;c5Ds8@a?mj;ssjp9>D^<x|m;Z<CtIH1_I
zJfmu}*{p@1Z#kQuU8w%n)fS`*qkM~!ijt3Kp&=UFmVG9ItPY(wp!x{0iSoF)-yf62
ztRT)h<Wo{_14Ep8q!rFE)54@nPTrNzHaj7)KvI0`M|NRsAu5B_$><|gOTsH4dYL)g
z;0cWKNpZPiU|%!ZPkf}lL^>A54OCkx2!YhI&<-zwQDwU~vY_m|#r+6N%fSoS2@)oG
zJJ$xs%z`<HlkX4Kv;lO#<NM@%0<g!221*io_#~m!o?7wI20MMzL8rr(y#U20B#UhZ
z{h{_e`361#^-T-VKu-UsikJYCer0ahY{OSeomrZ;!s1T)P-pY&Ok<uD9f4)0Z=3Rg
zeXm`$?mG|GSDY_6W-Q+&1B-%dJK20x1|5JnaQP08oSnUBHYxZGwp?hHe>=G=8-1nr
zT|r~L>_Rx{3h(ltp8pi-VmTOotkMr5A$0=xz4pc88Q@y>^)rB9;N-ed&nQ;;#f`y#
zo}<C}-xs7*rSX62o-B<76Ok@9zau2m?$f1R&{@ggqc~$$m_oj*wBNi8CMUZRpNpXW
zXezC8YCa)!tR~rJO(`gh0bvR)L7W21<Erh+soQg^0*e$`<apX~;}w@W7_ERO7d`<3
zx98??mFzH)9ngkko-g&Jsy|6N@YOe)eM{dDF4k%CW}YgzIgsS!W+0LO`=PeE$!!u#
z@o-;$hLe-|_hIlmzAuX*FKz0)Mbj_)q+kwv6sdpbo;Fl=R)rtc1IUrKTwUVM#yuY%
za^le{CtEH<F+*zc<S5qA1~7-_Q%4V4BJ67@agU8Zn)vC;2GP7}C86ka!0k!o>W14l
z_T&U}hLN{kZG)=Wt~G0$05UAGs^CDB18;WtARgnrhuwI(9tRf9F(@DdS4KWy>_b*u
zwI!qhYcL1JcmedKvHfE&1GwMk0l4K97HDcR2deEnTY%}Y`z71ZEP;mby{*Uc+i)9B
zCa1(t*Q8$^R;cbzoLR{@K=pQiZ-Kkb+5uqCYDQ@H1mt)}Qp7NV>inf9c=AWQ>102u
zryf94xI(A&fI%Ys1qZW1PqT^G{+gxLF|G?lJE+2xcw&ZqQ`fYDpE2*%JJyd7$p*5S
zJyf0Z(h=h^;~H06ymqLZC%0D{-?q1Sxw1za&StK%A9g9DmJmJ_-jncjx?Qoqz4JK>
zS`!j-&EUZWTcgN7TI3ZH^u+9HNRl|3;j8#ht(7gjGTh`{3F|s-?xhA}JP%s)sxEAd
zvH0pSVgz6xn|eTPY5-?u^S-yT+N1WU()T)ESn@c1g`i{L*!rw_izpTWu=OuJAb|}s
z`Acuf4Fl5GEM@v6t^2v``$AUMT`_&`#q#U^fTc2%FMs}ywBc<c+HP}hv5%jSITVho
zi?hr1Av+31>XFwdF&0QCx$m(P`AU=&MofzQY)g_S!-M$O;n)h-A7KCRoA-&2`;bPF
zq7XC3mh_Khb}4E}p;tSRXvO?x;8%@+k~8f%;6$7myBA?^i*~L_1~wwuv;&b{AgO5q
zxR6?CZc_qSeuw;UkK@El;n_<6AN=8|oPEZj0c4d)Eknqxq?C^o@85>ja-(^w6M%8f
zkVm)g)A=JWs@FfaX7|tBSI#lO<soXLcuZDuTfDPdQnq{eg@<>3$uplBCgi7He7;QL
z!rm}7H0nXt(G#g^8~GX!GkB)8?bSH<x0&OCIg@&~e6QVmzK|{n-t_UZ!^xM6tt(v3
zrhsZ;G1p`=4SAsop%or^leEwe<(<$U<f{~tB{MXQoel8wqXbf?D}Kivuj2Y*3I=Yr
zZRBRN;sm4y$;<=Wb9EoIRcf5)R=G6GU9>*_UrS#d7UlE3{V_m5lol63x=TW(bE&15
zMwV`slx9IG=~_Caq(K3ZMp{~0x^sc0Q{WxGzn8zfVD)+!=b1U@K6k{ZQrYksQ#g|U
zBW)&>^yq8b9FmzPCZ5+yn?T}UB-R$?4Hp`MT-c?5{~BC0#f2Tbyrv1ImE>i=u>X?$
zvsC{Pb8oUit<7p(LiSykD6F@0`<Z&Q;k?{K?Ja5T3f%SXojNOOoQ$#coiq1}3mD7b
zx*ki+3vmsl++H9OOFfh77F|5t4vU%yoWCp_ghq?0^dqb3P@?&70Z~|iS%<3UGYx`t
zB09{~-Tf&5nfj5r`48Y00rfp%b^(XwBdlT6wSF^lL^9A*FRkYy?u0}`a7wzbua0jI
zZ`u-}75kbV?8UzBPuq0F575Wfhq2Kyp?r-c_kWb-R~;IYQTw0rRAM9YigtM)SwJdR
z2Z?7#Eel_S6q@G+V9!on+J=02*_EMZLr^u$8DskUF(kW}j@F!Y=k#~94)$fbUEZK)
zW_Tx?VQ39VX49(yd0^1N;iAexgJj3+!Hcn%zCK1l+Gl$2>*)s67T&!IgKug&_pUmj
zKN6)s`}O|UV;h%Jc3nC2rhhN2+w}4)Nbq<@Zu=_rhjG^Uv4^`I$R$chJU-b21B|>X
z;9h`JL--_FdfA{<`^`D)oG)e_(61EL9{5YLo3KegL=*%C;;GBj;yDR2hQp&NCxG=`
zfS%b`__%IuuO^H(Ogy+c&p)b`f76V^f)E3z*t&Nfi4E&ee~@?*PuBPTTr<s1v!=Fm
zR5h|oy>5i@GUK?O|5bj3dmnl=tO`k>s_okJyk#ih@+^(=v!~sD17FF6Ns$jXk>6>|
z+0>VD^l^&LH0x~b1iAy*9FK5guRzsin}IPXaLS>}u5|+uKLF-Sl0OA=1mF<HDm2Ns
zL24ZNo*FAgkDQFn!SD^p0$NDI1AWyY`)bj5xFp*v7qH<PkjU@4Roe4F^{OF0-;Cac
zbY-u)3`gvCa-|5AyVK|#pAjK15DoMJD~fxbLF^vz6g$$|M=fR?-itrY6^;%*+7+4y
zX}Tm3hL(pP)B-U$2s;KLch%vZ+CpTX%h`ov9g3II2PoVu7Vl3umz4L;67f|U`Nitv
zp-dN;a)X7dXtTE{+%7|twG!{k+#(G@k|K5$y*~#xmds{cGR1S|D_?=a5nrbiU*1?%
zMa2|gj3CqrKz2Z!5J5K%=YL(exzQPHoiGX1SN>}BbRcJp-gFrgW(eOx_zhW@xRJSS
z@|tPsi#i*UQw2x_N{zYE9R&Z+!hH-T`1>i)@dw=m*m*VfKg`T~!A~tjA8&In&#|sA
zUaa_xxMW3fKMARruRk=b4;2nA4Z@v0s}1wubMJ1kqR<!=q}UfFCvsDKdz*v;Aix*4
z(91tmo&Wti;*dHAY1tsTa<+oF;?H(-9NuNpvFJ1Mo1JXELi3(fq}sn<@t0#`TDMkc
zAjR=HI40*j(WTg01v}m5v0v@Onkrf;+Q5J7R5Y>n<?}D2Y4LXxBgIZ+OebA%oyntO
z$YpcXfzRm^OfY7phcHBba(D+tXDm#?XTj)|*u#~W6^0$z!*}uLoBFuhko$Ua4qiAh
zR>47ED=|DFyNZ#sXTa)Of9>GUXR^Kz*e1Us%J56EOp2x!rW&?rN0&Vt2nRr@I2hJv
z-|g!xI&o!949|d?BVcl;ID$M2oSH5GoCX>-u-vKSZg2a5zWxyvgodG&o}(Es$4d(T
z;4Z!h%ZVCJs?hi)Sp9B$MY&mJF}s&V%)u<_{mE!$`1(07@A@aj`hhpla){`^O7g^b
zuIpZHiBqjUBCq>$R0ob89I5z3J3uRB(Oqert}1|DUaU_ucXO$bNDcV(d5-rny?91K
zG@i@^*)))@MdUHWY*qu}BOr9mDeS>%D;;@dgt>F?h56=xv|4W*W>KSk@i_Jv<+sM7
zc|$h^$pViLC`Z-W3=KB}&dH`M5z~k_Is?zyV|YIpnl;I{V<Y)=`+OWrG&wXSLaA|+
z>xx?<J^(uXn>Tz0Lybx%z7?^oQ3x<-`ZGTqsmmNk^Q^RSm%dUEMvDT_OGVj3vPCD5
zjgmQ-x{&7m7PpBo$h-Fzr3F4V|0a$vc8#Bu_OY-O9l7%S0-CYAW$-xLRMe0bFGH8V
z5ieTb^a~ez?&Ycv%QwYUX3|#WFs0hHqTn^?vXz>~DVs{J7RTV`##a3s-PyK71`!qo
z{_()3hW7^ymkG5QD~Q$iHN;C#H^%FVC64qb+=|_Ss|sKNfxyb#g5J$m;4bMDPWgsh
zXjbmi_VKQ+KAY<w&d8^`D=Q3OyhvgOiYgo_VJXSS65*wo>-uY_R+~o_lGghdS3Z=0
zR%HCbdf*;ygSIH#vx!mvE?w(E-0|}A@)Q4>)!qvgiXx1Z$N&vvz@kgN>5v;|4v50;
z1aa{>Gh~0`gtG+W>gm{7G{45PSOkv$&up;N>!q)g&RHyb=7^GwKq+q=+vso48k_IM
z)@dnLTy5ygIELlNP9vn|t{vHA@)jD3lEq@8gBoohtAcMGFvo5u73=HkL_=aavdHmc
zVnhQW(f;#wW>5V3nY3V0kn*AeP~N+|bx&{357#5ctk%Ji*KjiRmdMd<h9(i~IO>jo
zO5|M!i24ztm|2OAjubd2M!3WiLff^)`Ohg(Ngus5%_Br>R?7avfT{pT+B&FelKd<D
zPP|&keVJ3+>l(Rs2;#Kx@mlJhQ2e)m%Zm%o-l=nMuRrNXy4jhXFN&m&<oO4!Pj$eC
z03kKH1eYrX_P|nMmpLZ)z7b5v`&asS?&oKuSSvr$=F#tm?>FrQa)X@`1?K-xLgDlA
zOy8o2_V-a`D?|eb=yUk*9>aPwDpO1OUq6y@RAfzJV^k^1WA<)~Ejv4Xhu0**N;tuG
z=AyTXHsl4rW#yzc-GJ?n(0%tbF0vk}A+}y(x1q?6%M876_yTvT{Ljbrr1%KD*?e6>
zjjflm*kOWpY~jJeKR->Zetq$=$0OF%<Xf~O)(=bqzw^rqL5J(uH(yR_HQpS{oAg9{
zQD@*y__TBgf4m|(Wb)!EKTZmcUM6O?p8Pgh^7e!iyFObU4^60dWxSr4kNu?~>E)C%
zVoM!ZmAJ%WVO<pc#oU`ngM|GwaMyoAZmA5gm+u<+ReNfa3~<O=nh@Ld>M+r6+QZv=
z*O~8iSJ?Q`XRo%AG)6z(db_rcigD%X5o@+^(bgK%H`Nh_i4jxjL8<l4bSy<wwNAmP
zoM}xO-F;mk%0_i%$Oa(@ea&YEZ5kFK#M)1OSEApYB?SiY{WV(u8@rRp!qf{Q_!FeG
zn$1hV@vsE$*xN{y+ZjHa*u_Q;C*%6^#9vp&2*-C~c8DC>9X4S;IYo*-=EQH|Dcpyb
zmFN;VA0FnUp01bdoJs$-{x4zNx)HInx97l4#}(Mth=Z3a3++AF&|`i6^-XW%-#`|a
z-)^*MwlH3qZ<%?+BOjoy;sI(AV1)+UJutpEm&y*PvZP2N-SP>}iy>m(oTQ!r7O8QB
zOF?vtU?zoz&9d36YDBb9&g<EVl1D$JJ7n9kFOO5*Gfyw@sAdJ~e1N3<_Su7Pvy>x>
zrnZUHr4sDTes`++@5q{vE7xgn*Y(VCr=J8NoMql|Uv4}%N*Tcm^l>&>Mn5s&WBac&
z+x4XlI6-b_oZzf26(yT`j**-Sw%)NDh4cEG;C(svOf_fT8S~g%kVIWdFeVNu`SQJ2
zeTzQ7JMi>5iNn`!YZ9XVzqc<cruoK4T8@dGr#6J;BaUJ*)4|}Z+URlTjSfe<8G~W-
zNMIGP=m_KHsduO>5VbS!U6ap5;`#vcD=52m+=x@g8o)HmvO&Z*&Ws4ti>z&IjKCl<
zUdq%R#m|*?*$`Zz^Ku>XhEAjNTCUh~Hj@?I`~N*RVW!_<;7Y9UUZ|LZiN%h8Wpbmd
zvG=|W905tJtVsYn1Ux~IbvOAUfP;fGJ79E~XX-`17bq+4b^OSdx?@iS`DAXWF-u9X
zxr75fy~N<Ve<v{aymAE>;Vn^4KgI)7shvm32z2b&KZr_Y-UUe5Y+dC4$43Umj;Q3h
zxT2)r%=?w?H<u^n)BiZZvc5>J?7=eihfm4STc1VyhJ9K8rd;#DnHtS`^zqX5)unQ(
z{yseNVd7G~T2Tf!BJp79Y<9zZqpEjAZ4*j)fPmyWnP4is(>SD>gHN<=FuCkmC7fi$
zey~k&%9VbMPiWbQa?bx!^0BHh>eG0V($y8l_coqhaY2x4VjB{Bpd|HqZSv(w|E8p@
zTjK8R(TFL#G(UgJ!32HMf0iQs1EhH7sIi*EuYvZ7jA7L~)_z87?^k)0^88qW%A%8}
z`({O-W<&Bc%W8QebDa-pk-()2t{Grbe#@&~)A!PXzK)Rm&Q#{WBiq^(`<z<?_?xIq
zS5{VXCaN6+OZ43KLkqq7sV$KxX~ldOL>EX^xxK?18k9H#vJWSVLLSBjGe0fH{d09h
zb^Po{&I4`8TL<e5tbtMS>QW!z&VbEI6ymPL9JXHb%V`Nxb-eK6GOMYAi|+3A?q;{G
z#wj1pdSIy9dyP+yTK-k0)tnH-)oRH6l;}n0v;V1*fzrc$e@P~kQbD{<E=?G5h_B>J
zT@+{hbtNq!?OtaO|2Wrc(Iu`&hId&Ql^6%jg?m1_k_%%1bW329IK2>9;q2rzO!Iy?
zICJ)Dq7sq(2OI}l-Jr4yvsR(B@si!s^0tX+H?wk?P3$y~u6rXPpO33~C3NWLK_!&0
ziT}ird~fCDNveFHAb$Atw-?$i-+3}WR`t--cP6xW;}yS=Q~j$U+;>0kmwbqPG6@yD
zAaC!76r=dOKa&3mFx`1B?lZ$erlZRSq1qU5Y;uYCw{dr|7+-TG8wr}6_vfc2MhAXA
z6K5_Ftal?d$K*10vYSE%-f6INN6zpa<t+szE=L4|1?U*(%gy@vle=gVLP>V#-yQ)k
z57y*8lO_d~LEEJU!>FdqUm}fq&AmOFl=oR5j+G|`GO0_^Nm|lq_DFu7ba$9YB<xVS
z4sCpAgX-}MR}zA?SbG<Jz9TAUk>*hMNz)H-b|QEi!K=FH@E7(?<K~L0ppI+GkqqV%
z{G61bpG>DDxQdd~A{1pU9mu7KtG*+ol%lxzftfW}gblu9-2f-e=KMM1OxE+L)DNxn
zl)m)Ce_E{8n|_=Ys77o**}7cP8Dop1ft~@8$5sxUT5J}nQagLGcLaVuhVus4Uz@H4
z%cGgq&+6}|u2V-Bz%>K>eKOdv{1U=gLbMH+`{K-A?I*jQ;!@#b?-!~f*}&kO2o&}E
z3oZV-x6Yv5DA}_HQL=noh=d)J#xF9KxG<&`b%&`^K3Y+t@U_UvuapEvSxG$K4^=`O
z`Kq>#R*(?bny6vBT=FmW1qBNZPte=|11V*1fA4HTBJrX~P+D739-cR3Rci0Zqu{=^
zFX^9!h1Lf}8eCl?EY=nCabwzzZx03MbTy#$W*!(s(BN}f1bSw5IBgWZ3SbJ5+Ndv@
z_teg>A%WQZ2;7eV77k1}1x}!`JD?V{WNT6iasL)wm2i+lmgS}8SEk9v#h@8JbIdX2
z{6oS9^Fc766;$s(2eL;SDVY;TC&g^qTLou6e?oTrCQ?Xz5x-E=$>lGXBUP2_JEQ0X
z4oXmAy|E9^@sjm#(~X%rnVqT2IG*lG>ozJpGcfnADE@#gHxF8vaZi`=#S;3X>HF+j
z21J|OSS3~X@#hz8RY~Um)$-6qlk{+s)4n|kkISQ{mQ=>}rZ??1Rg1AGhGj#LAADTf
z7m;p~W1{a+!O{~+?TknHLZp0kATe_0Nt1Q3f@wuj+2162<_^OYe}t6kl&~LtDWB%@
zK95*Vvy+QrcMzUK?{WN}Fzn_18bK3fSQ7s@C0?L>A&V;XG^7mwiEylw`7WchZgTeT
zl{&QMHx(V3y^wEV*Br_G{;L5VO;fE}WaG6Dzg}43r34oXL8#VwzmbM~$!E%V0e9h9
zHMp`A;(_p_ie&!fvfi+UcA#A(*@vK+f<~tD<Mfh?NFlChr8)GmDF_|?lN#`OpQ*F%
z3}Lnj_ep*^`qEzR)V!F7(Q89C+*jsdM6XWQQ#5foPRP(IFuwbr%BG9jcV>?!<^_V1
z&F*U1V2E*m5Ja<2_4FO@Z}Qz6FPZ2SD=@>jbq!BB8%NTx&~QK+EQPqPED(o6?UZw`
z*$RR0=>5|bb|+RC`l_ml$W7SVf)@BIzL9HE7qr2AO!!2fJ!zU*%C_lI2NSb_Qdiv#
zXtz^uBxJNWZ1exTlJL9LBTb6%jeRLeQS2V<4z5fv?aL=+<7Y_>{42%ycO3BR=Fi*K
znWZUKGKcLU@S&%q_~S@;Fg%5ygUHSPt}!t^+5Jb1<qC^7O$vNE_=w-Ii7URLb*^%g
z#Dl+HS~6r8A^Jqqm=Qt~DtPT+F*=-4jn9dO@hJ>DmT@HX4lZ41VyqK&nqn*|_G-R-
zcW_9zz`s=GAl0-B#Zpyl4aeTi@VJ(D<E5C23@H<(<uwt<IRSz#a)P;475$qo8anm=
zE)|+sX}N!XTj4WGcO+jyJ<)+)a7=~x{uLiW@n3D^MCtRZsf5M7o+TX8{xFg7fI<QJ
z^|FZU=`4ztCBBa^u^|A=+yI#wrIOyFKH+JuiTgN<J~@FRi7HkCE`T7aF+FGz_8xKJ
z4BISuRL1~3<Q#!GTBzH$ac(<K(~^3+0c=*zwoMTg?B+O2Wna<CBT8}`^akh1!E<@b
zXV-f3jRlXs&el`!Bp7@W>iqn)WR<sWD_=dv&otDhJDq;Y*df+Cuabo~uv?=;F*qAa
zLVa&t)&nAPuPq{NYd9eQyym6mY@L(sQBWQi#Opy%>Z9;iKE&r05yPkWSFzZWhv;KZ
zw1F<S7xTv*Kli%)iRo=B-T%4N>MIcrsi^c|uXBa@<g{jK9r<)C6IJ5kIlZZ5QoqF6
zu;TNIZ4w~DrC+MRPcB}C>5R*TvY8qD9B}$s5S@0Sa&LUqxkIRcR7RLAgStgGIzwZ&
zc+66+g1Hqq8d_pxzWzOHU>}9=JHkHRR(j}TrlFGJhH#mL)-UQrjaAHYE;e={MIBa7
zZ#z&1>GV?BbUmj0`~kLd);R%H-q&M6afJSvW#MkD3+WgwR!IxuesT1DdfEKVURTL?
z2<K7-+BIht_ZHah%n>=^W@SDwEr+hGJGy^+p*V1$`yd7@FwAUBK(F1}+i+sGbq_+v
z>qh@WO6qOrzeitHH9qm^ZuDsLn;WYav|TFPS!oyv-8I#}L3A+-lncSwhjdT9E|=&J
z(w-4wTh=<Y*2HBn5)j9r=8{_++u#&!TtbDBL)N4K;sQdkEiT~1KpyTApC0TVopO(a
zWKXP)oo%-Z?ehqFOhR0F75-F{xXe8_c{7o17YJfgPmvD&X`fglM%3|W>G-f~9|XP_
z8EMLUR~L;vJ#@ypz!<j{s;dO6LqGSOHG{rG8kJ*)dje+)tP$sq@t>X~vfqz%;{BxM
zD6=syc<O@4uSf9Z#f%|5We<NR1x&3;surQ9Nm#;HIQQe@?h~ZIy<+oPj0s0Jt9V*f
zqKc*{KNWmWBxMvYzUiUl>hRyHoAWTeF)U^qs4|x|!XvqJhM$;j@ZCK&A|ubjsUJ<!
zE`-Gme4YV`5GPApG<Q};0_&Aa`CQmEs7!&-9P43WDSfsAB~X9#v#Qwf6sYkAO7`yD
z&Fued<VL!L7tkaBrWPnroafkPNm!vx_^B*P%ti{jGQv&Y<PZBfkNwap4+vas=VV?Z
zPIXhG20eOgG}w|`hrJ0ch{wIcB_&l4*aI`|epvT9R&Y{0pTjVJR4)Su*-t~kMSdE4
zKyvIucA9xw>hxR>XvKp!7MRJdCWZ)*29)H`g=#ZWzCuY-wptFkM<tr8MWSc|J9frP
zqzn8c**>3}lJ35t9VydJow$F?qyA<Q{Cl(KdxdK9v2nnx&9BQ35iUobsm_pFY>Q<0
zXa>}B&elXx57ff_c<A3#Cw?h-oSv~;?yE_d;3p%`jh!b+J6oZfaFapx8#Pp~NjYTw
z9Q^8NvIM*%TMzvSvh`R80&KT(ig>5^769UZ$Ax5p{LF3D#acR`It2AN&qvYB!DTS#
z=dp$OIaRIl<d<w$^8O%S+LOp4Nj-&ssp#%~YLr>)uhkJ4sy}Pea6h110953Hc&ajZ
zYGpX{V})KJMFm$LzamIgzf%6S&y<jR0IA46UarieNwVe=iY*iR0b?E7It3h8tHF%-
zwFJOye$Hb({cORXASF=&$OuYctb-PSVh3OvuqdE-j|3n55l~l#F%<oE{nk5UBD)${
z__qsIg@%o9{wNj*6%!P+Q3)kMR9&Sb!-Gp1curU~+#y?ZOCRPEhCrV<aSPVw$VKVP
zOIT%msAWx_)`5d$61-kRt4LT*JdK$wVapqR_s`0PQAt7mn)|qm#t>CPDVkOyhVr5v
ze#1^(&8HF!9O2fP2pY7MOt-G;_<VRL(RVVA-p)d|tQVj$_YVw~D4<(<Jc}?ilm+GP
zE9;Pf+p2BxGj%oL%>KS35C;FaU?BAbEPGweG0Bb6rPL+fjs=`pvmA#l!~C&cmPhtF
z|2#o)b33i26?FF5kaZ|qO(0Z=AwN&UPrzRU1FhO{%A7ypk!_<ySSvgDE?YnG)6v4c
zW%G@Ix9y<tXPNwQGE$H;OzWF}nnTosZ&c&U@0~56C=Pao;4Y3|fD>`LH}=49ya`~9
zwXw_l&mt-x>TTAIv}`JL0$wL|ELQAGA0I3AU#N24T+=_I-U+N@cmgEpn?HlCt$VgL
zGfb8?XrzrB26c`|M~}x)?@IK#NoyW<c{X+%>6X7{PM)uK;|8ev5)k|Kd&gj>N!K#Z
zQ0E1dw;2EF!5E7LZLnBK`KbAlnLPv6i9p4pzm(_bnxZymeUy%!T;U)%=UKM<<pS{X
z^`V2<@yb%zv*(nhj@WrHgy{uuvd&#fB4xi3urC1c_Y!6TKr8zfW^a@zm+->p;AOeY
zl2Gg`eiNARnqFa*@vdRqm)x9u;=10azp!|vT|M6>4R5`Ct|!KG3jT#koj~p^VqSIL
z1P)HJD8D|x3j@(Tt(Qn9DSFPd!HrnYK8~FCvddgeV`<9yBQ<UqLypAM36Eed7bJPH
zZaK!4EbOMJ$W|L&bHtfPV=gObtX*!spx{7JIh72IG?SH<*4CJA2M`0FShLEBhx4<-
zYovof(DIuq%}F#_^y3I##bj2po75JY@s9t#@I|yiMd`6kn*sykk@Ws=Tlpaw{<O_A
z1XhdF3+63Rk&zTnRc}ybT}^=>n1*|-Whu<7dP0eRoY~ETM5kM3<_uW+rt9PKF54d-
ztYh_<re~lY*xwk+c~TS{%!FNKz?G__{`+__=DvZW@vgL`8ZR2N+~;#x(fe{mG&wu=
zO|^dM(&k8h>6O6Jos4uMBIWGF`)_2GdNhiuLt_*q@~Y6DmH!IeNG>a?FPH)=bU9K`
zjdg3Hd}7z2HFpHLC@hV!!}}FFYb<7W83jr&1&KJs6n5FS>bfYNCXE@}y!U<?l&|w0
zBZb9~l0E6Ek`UmK#?CL6Q8cx3q{@U^BQ(aYHpseQsH<N(vS7^;;Y~j6f0(Ro(Kweo
z+k8C!rqt(_5Dm2Tpku$qN5H_$!Zux;{Q~_Cz?*Xo9D3!i<FbdJiUf7|uDp(%-uxn#
zun~voo$11l2Q`8F$+n(9F#sV+9)t#CBQ?jEqNEO6BjPByM|ScES59QA<{uTWL)vvR
zD;yK=Pd{>0@UWiIzZ$|*SJ3kU?2bR?wtwDfqLk1D176yQwFE*ZPY!6eKHjg|8<`um
z`D2c#2j)m`avU8cd8iHYSpEH5s{+!{i+_z6WI4p1{Chnqj!h|=QKmn(_NC!jDsx`p
ztyRG5ru_)ZGGNHO<WC?Lj6$Kt!xA+KPPYN9beo>wd<_kf7mZ#IZHdvAeeL7Mo=W?f
znA3Xzk^`oZj}L!9@Y(9K=-}%0nnjd9)xL9=7E`ZC(=LYVRE*^k6MR58GaT~f@pbM4
zfFbwJYb#3|6)AbH-rBG)>jqx2lcY78ABPhg&nW4qm1g6!&3>Wm!i~)Kn=JLBMNqLM
z5)5B2Ab)aY)NWF7@4WAjm|Uqpm<fdICbE0eNt|R>X;Dqi5&vA^$J5K)93Ik74Bd(!
zTXXc7%R~^So?52?S77B=jO*G*hwrg+pP`Al2{#X3)437cEq)Xc$o$7*xtH`aIfdN(
z{SD_N)_GORWry}JZ|y|+0vDd3P#D5l+$C*=9idO=tfO+r(##d;KgmFq%OU!ln)>x|
zjJg4isId!B<N}!a78(UI9`%F+%;4bQ8~u6)=*4#UO5Zl&6gr;JakUHMxR$M2=9T-U
zP;Iq&gt(a6z$q`<V?lq%A~eEz`#T@?1M#=uJ{$DtQtY&6(^oK=Pct;=Ufi^q?1|f8
z4=8>maAOd}NFacyC)^febUwGPN?U`j`H18urh8CsNp$-`SE7@{+kNZG-&O=ZfsyP=
z$u21_gGs|HA|KW$@YMHDNbiqbO^voyy`}1z0owLixA_R5@NmC4vhlt`8%5QXfjgJM
zXGPA3p{%@o`K{5Vepm~oenm57Ws>-1VnlN2JAlC6x_<gqHAGz*YZ^|Lnd|E3f;lI!
zZu|}caqgi!I)gTfQROS>x0%s#Wfw7d++H(1m+52&sKZ{dj=%uyF0^q4?Q7-G_01)g
zSDie3nbNH+Q`R#VN>ixJS>1%V%#jy4VY?;*EjFx2KYMo~-9sAL09SkGR~UBxC#UGh
zSG<c0e56bhDG(?FLmVONx}*z_dNK4V&o6M7yUlqp4-5>XeAu!_qTlQf2xF>@4}b;(
z-ALw{fgDJ9RIhE2W|<q&$qO{d#}h_T0DNh1jvGYz%;b3>kyOQFMe4Djhrl#w{ik-+
z0M1+JJ8f)`IB`4U`Ttsu-U_$~xEx&;DyM`51gsl7X0z%p$1ZCi4dnr9ymO8n`bCPT
z9=0`WuHa@k((<gS?ol*>5m{^H0V5Y#rS%5u=VUY589{zWK6AL2Ys+8KT(ijL)~yph
z{Au|MfDWe;>;hwDg#8jhet!P#(b?&%Q-ydjM;5^pdyi>!)Odq;qn@4R$%W$<UsUo)
zC|+rx82ev^&L{tF_Sy(iUjb`;;8k}vHPKYf$YFG~rSk@P_+Res`gd;%Vb15wG1rcr
ze@(|RKBx6>#^rY`r3&UD<Z8kOJ79q9=<50eyfFQhQzKF-(07C>@X2mIATk?t77K(T
zl6T+K%{<1Wk{f->{^S15pxNzsZ1x3&@dULKcacCFq$C)t^Y15DN=^1c&5tdaz?!JL
z^XNu4hU$PYD}jyI7PB_djDL6j#2x=RU2aT%5o2uI#z(6qPWD;6_lptsdVvFP95l$R
z<g@P>IbP{a|3=&XaOv$-JKoe!>&cfVQz+m5%bH>69rvRRb>K>v3rVFZIlu`%*z#PS
z+bgvO&uJaR@6f(+IH~oo7JtBPOyM;7bjW=_<<%=V#9AX|N~bKeUJOKHauj>6PC4em
zWqQBa8NamZL;nUwNU`i;O5aN+Qga)9nwx(FHWzQ_36K&*6F$QB2jJ_Begusq4jG?7
zZZm_dB_e&l!55B0SK>sLtI%<D@Y3Ty9(BK&WB`_z*pS$6jphRM9J_}KNSDE;Joh#f
zA#WkldNm}br<q_*Gpu(<RMl{#(H7;ypGe8-fIqH2R3}3gVW7KtK~!6)f*cp+t~}DI
z_=xyztAm83IIl|1(QW-Mi{WZ0`^9(3DkG6o@-*1*c;yhdf7&;lTi#Mo1T&(M^MSnH
z7G0Pj#ft8?;vd4>IIkQ8-re5-We}&S(_QbeU(7ne9jZS~yT)9**_`*hGm??OH{@+e
zLu?X&s8M(#z$EVI0LVftz5cC>-<*R1Pt##KrX5(F0I9z_m}LC_tj!V3<_8L}JHdXb
zcG>>8vH1CidEZvJ%oIsXNW#>VicwNHKS*UW(%-QbZgeusH2^}Zaaa2erdw&%*^g9h
zyTpMHYNqK1`mEOPRdrPEn{n!?Rq_868$^gK<lUc_8Hr`5lg<)~H!UBTKMs#3A8G{y
zVb_rh1Kt`AGx8xHIi5#$?VV)a;Bu<O%4Gjkzp-YA>pS9RadAL;N<6FUO7~dbg6AVU
zscQf#w)TnEk4e6kkuc4fcp66}jovgAinkC#Q&nR?eB7Qs`R>VW#QK0TqWpDlHK>i0
z*(#>Y6Buw+0v506%{H21vHE7H^mn#5{D4@H%m03LQPC|x4wghaVWGQQq2YWjbzivo
zP`H1R0nTRMeC3KB!vXVu^$E9g?(7KSX6PZPJkw(PA33QI+-x(}mp@Fg>R}<dTQ|-7
zs!Zf4s5!WvPiX6cD@iHP!@~m-)<Fk)Am;!{z}4-yuFI_VEBt4HANP?Zp!{Qr%1V#*
z?<IM9K7AE}y1}N!@0x%xTqYdGT4^2&kNl5dXy_OKso1vl#VvP;@nUIlI;d3G;mU7I
z3SP|;w<zin{<!7Iv=NBEoBNh`4Mw3=l6s!k;No)8q-9n{uH+JC3#TkSRUk2%V2+^F
zKdejoIH)KefXAN+T4%|RY(CpfDzfxf{7extj)ZRNA}qUmjheh)2Yt*MRr0qD>u{9Z
z55+=m0$#aC1Zgaftig7%meSn+4Lo8z_b)SDJ4hsyte(ka9o?>5@n7bMALtwUIqy?X
z^djpKrw(B2l+ySA4Z%+l&aoot_^6{Dl^Baf*SDV1z%Fvq%^V#1@h;8VMLYQue3B|&
z97qiHTI+QP7sT(-wL6!R8x@me|Lc79D#eP`sv|@s>Jw~Yq%iPo24N97&`a7pAK2y5
z=TZ5g-zJ|2C&c;&M6Ot?1o6=6FHfY7<;vSYUCO7H6l_IoGA{K5s__K{rXsD&`D611
z3I0U{C3(*thX$mFwh1TMa6b;B(l4I?#U^Ryd7stN?P9)Ckg4+bJbInZGWR@CbR@J%
zRPP&n=fxkzJHw3!n0SV~y4;WsCGQOF2~MhsWkCff7Duwn30Y>JzJQ3!6c0;RZO;sY
zetKWexmAP$ONe4aG9$SG@K52^{3T}jFHsbYmpbuJDC{i8LvAt@fBMja1J|S4G(T6Q
zpKo&+PuUxTV{lr5=%k=xtPnTY45?Xe0dw*=v0LKEEF0>Dsp+m(i1owUu%X!C3Kr22
zMa$DMijxrE#%92Ea|Fpv?>vmS68mBDTJK$vp#?(EuK}pBc?f6?kx^b?vpWmbC9D<T
zAL&VLrtZ^CmA_v|0&mn+q>wlDz4C{+L!zeX?^JnkcJ@{Vw6f{X=fFk!J)qIh{_R9Q
z2kRxr+q~IaJPxNw-eu{G1p!6Jw2~UfUQ&eReFY6ND#=2D$c<KK7{f=A5vaA~cf9V;
zBX1@EFS_#6a6}Nh;6dz~h=w`afEVNG-|4YRa7B7)dHZ#zJ>KCRb+{cio)5UeKMg+p
z=u)edP2BsXj4!;6PnpBWu+J9{%E}^3eRy$E_g&;d`R-p}swSykzZJ!;eCReS4UN$g
z&`-f<G=DEpR8Z8PoTYz(YOMu&*SfhClizi>Bn<#{uxkm?QLOg5Yq!r<(3^;;t6P3x
zv4$YLDY-RZIBg!ZX!2g5Cqyev23zNi6^l`smN@b_0Ghh@;|!0Mx7mpsOTT%Sb|d~v
zRzt1xb|hz=<39>b^!92_Y&<F2yX2~oS6r8hZh_*3|4i1q*CN--S<R`KHBwf^Yz8oM
znsSa{(vNhD=;O5r<H(dMr!QjW$owbyf6!E!DIBG0@z%;mF?#A_7;WaTp>HZY0;=Zg
zut~rxmcEx>Cz`JE?2Lm#b8Pl4Sxt7kGt<8h9K2P#tgP3GoQyIp^U!i)-_3gM%YbUo
zkW@C?{Lp_1Nzl)Da$!tWmw5Try)(>O39IBv!mcD_TkvL*EuP(m0(uZtuLz;OQcIoD
z{Ojn(_qXL;dd83yUR?f~B?(M+*lONy2PXG^|LA>`!d?@H9;d%+Q*4fvpyf$?F<r%I
zZiH$sy+B#di{MqfL`nXiyl^;0*j@Vuqhh&jf0~!GGy^Y*j1UuVFq;b^jt%(m=02FD
zjt2x9yU*nT_kw#i&LwztoVyABj{d^#eLe7=!e{*R79{}P>OX$yC+jHtZi&-&sp_eA
zj!6j{SlAH8G*rHH$57^+_7JG?YZK{zmL1?PiQ3tB2K)S*XF{rYud-}~Cov&H!>FcR
z)=ypiVmP^8_3@!MrKhVAdh2%^M6QQULv!E{{69de0iAR<uyh2=yPR=GS+kV==P)G{
z)4zA2nG1C>3aW;M>zfWDHxg+>hecRgV6|-CbBQPC&t~yDJ^^R4lpXTy^n`1Bd;2%N
zb4{XBBI|hF@j4-MOGdu${SLG1Z8a}!mq{|a79V*rdlT2RsoD#$XAvb3H>^mwfcT+_
z+_B@?kQrD_Of>6_xybX^Jopi+Q;1J0cF9jZ#UIPfYvI$@z${M++-*uA{pXbNk@Wa<
z<UfNhij`y|{6lq=m_`{Nt|NnkEiXlkco{5~tYI4}eyxJOkbqxe=HCo_$$q<N8e~60
z=g=85@s(Ekwanx&nn}-X9KTwxEbuWj0-VWkh23gv*??r);cz#8x7$+jdz8`5#YCzP
z@5g?}#5D=_D9X1wlV}=hl`n6_9z!3YU+2BPED9rOy?Qx|hRG(Uh&gny6&T<&yC}CJ
zLAS1yd-p@tTE?HexOeB(Rh@nB)o$ub>^uK%WnNpNI)zwi%(-}I$nu*%Gt#A?Dd=xv
zW5~ixhLd*5oXl)r5@jqgr#w8ZMEst7trPh#b!b@QN^zqjUN8{WqCBL&3~!3fIAna!
zDxa;<ct-{1SN>$^FAD+XVOf<EMH~n!62Ghze<pr&C{9i7y(j67elu_TO5Cy`&(Y)K
zD?0sWRBojO&}|2;Iq$>%1j{Iv@;sD-Uc3HX7x3I^j#=w7nCv|dDj$ryNF-<@-mP0*
zGGbpWkvIRfwytLU6hoI?j!p<g!ENP8vU0x+Lmn2TPRx{Qyip;jX3VSB<)H9I{Fx?e
z##>j;?AUwk7RVo$-n)zg&>Y5O$CB{(ETy6k?Gk&MPR&Iou+Ud2?51U|n})mG`$<8-
z$4^7f;$PPoC+Fhh&u!yR+<MaK9UG_JK_-Kh*kST5h<OX5u<Fxh+D0OGsc=MmuN&ks
z>Al|}S%Q$d&+_!>fTv2@!7f&xD$1l}2`Ok$0(TeI0cs3n?Dj~vyNbpjzy9YymKp>g
zWT3wV1(NE$>UYzsPq-sGE+xssVgCD1bgHFP0lHIaExx(D8!W9Sxd8f(QYAT2X`L>t
zU}uRe;!3A@PLj1)$1OI-Rm4T9qDjGvJ4&bLYcGc+Ttcgz%+i@(mgAR%1vSpiPA~s^
zL50BprytxCpzZhG(#Q-25R0~PT|kf)^DJ)y;={bKGk<aTJVH7a0{QR2+NH4OC1+MN
zV(_Bfi|X+k1W&?6A^EFw2zE*N(ZuxN1xT9n&6gq~*~6OL)@XONEGnn*pq@PCI2?sU
zNm{NG?0-6xz-5iIHmdu1lzznud5QeGcj%$xGg;%!HH$@F%3HhWEK*P(#)Y2^lK9rJ
zM)Ptg8lS&@J}0Z=aFnMp(0EXW36P>o<EC~ptw29MbT7bFw0SCKGKsXiGy33ozDpXb
zZNC1ZMPUh9Wmw-SO*?+d@s|U8olj4^{$l5_We`H=K)Nz$-ddx|x{uYh%D)YVB5;`B
zejcn@AhB_OwvtSoZ8FcB;sW6w7DUCDq`ym><hz*56L6K4#L0oR=Npe$jVyfVU%ulR
zq+Ohgh4v(2Kl@I7`5n0RPkLswZX?;1!><)=UTi2fi<r6pJmDzwc8?(07|$R&E-_y|
z^cd0&5pd5m8oWqgjoE%Z1XFCvvwE7=YwW1g{YQH96ydU6mw42>JLrxQfl3|;z;#%&
zxNf!&76va&u`d;-th{4u6F_!ixU?dQ$fJ6`EW1R3IE$h0pXf*=C}Eq9OYL{O2|kmv
zy^<rlP+H^qF|XC^%gvM3g#NV%#zw%@Na?h;Kz!?`eN|6Ds59eFndIyWnn<_+dGepQ
zB;C`?q7ZTE&T)Ij-Wme7{qL5CwICofmi4&n#uD#a085=VtD)?Eq1t{urNxd#r|hYt
z$_H+uw}smVk*nPjc}q~Mt{+Dzm)}{gxYdOl4CTS~KNCmvt<-pmixnkC<z#-QKcB3r
zS9|Qc?t(pcdpb*yXXeOk5HlDxC3K*Hn>X0@An6I93Gh44Q<dfTN43RSG+A@*+H4{_
z@8^fHhV)#AM5R>_1<9H9Q=C+#iCG`P3v>%A0348G#Im6_Dt<67`_P-*l%~}p6%kaR
zDTE+%WjaKRNxY&Vv%@j%s?F$!LNGPb<~#?WkEo>7f<Mj^bB)kpQS2J+wM9$l&&g$b
z)k6?xfR^44WDLvu%UuAlZdlDfJZnT~|J|hbwB9@%%8~1zWVPc_gvE!pq7v%jybA5P
z-qlB=9v*O^%LMn~#Nj2Gh~wh8DPvKtK0uMV4bG;c+p--zH$r}ijoV)gRn^XamQZ#A
zv#aTuo8CH}6<*z&7oN#GcaM6Y-1ulB6808T!B$evUptuu45KUcW7Cxv=lQ<&s>Maw
zgE6m}2WZ4R9l}&Ndrdi|?lbDhpIY1K^)X=;aH!|YSnh~&(NS2))0=4nDFu(^k%s61
zdEP5|RYUe#em&7k7Rh5)z6_40R(D57qRi`~{AV(Tysd^cS*XAe^KA`WnTw}|4s{KI
zD*B^TU(ZIsC!TB>KcK5-4x@4?U3FkH^JK#vG}-X`#$ZU!Q}+i*!mh5@e#C$dZ1WC3
zV~4KkSvXjZ+%%6;#RUPe9cA@27$h8c$O@+YwNae%nz5iB{rV3p!z*RT#b{-CY)x!X
zYmA9n&P(n?3OS$5l^{8EsBPIOZymWB(W!;!t8;T+&s}PT?V=MtpySWG3I{1bz!VhF
znwP<uyT3hIX*%!PDKJ_{e9uMC&V`pw{|-M<Y`Z2R`NR?#DLaE_75=rr(S~*6awk+B
zlg@cOl<Afy$;2l43t39twP_V9`;PjV8yKOd+{+?NKN;XNL1wtsKG-G^ceW3oA9*<;
zCOg+0HOw~F!HM;6+sv#!8!$?8U%=#Ly(a=C8kNiaCyNn>IUW<*_xkT3`bwI`5<o|o
zZsiEwKF5=H<$CrGR}}j0p%3d)Do-3K{<Nd6eo7T_&h9-D(WeFLfef&OpdF9^kNO$)
z!txZiv)D*)v7%<8<-<lv6l1$4g*4NDjL@oLbN<P^Z;E+k=(PX`BsYYX4H4OW<4CEO
zN%`Fz$=2_M8=shA;p^FG*l=|+y`RU8WWBzgzqx!fSh)+o^eo-v(#LpX<x(6$3xeOj
zX7rzeh&a`(qw-=~F2a)@H#QX5L!U9qPCWmw(V?en%Hj^y)wW`4iTSBlJz@g*gG4gB
zRX)y^$hVQbjEy#*fGHVW%=U7bgawbM?wn1F$@dTumt*)a021<YZ+I@3n#@rHA2>_1
zxzwAd2K=ztXLf)4ItOCAWedT`O5lzFMC7!BpEfyxoi<_{;ycb+a(`~*t3?DWetooT
zJ7OU|fn}OJdp=E&_b8tBiPdi|><J>QnL<6dm>n?U&NMkt#Vy=)PH5V#`fHKr%ClTH
zuLt{++Iy4tJ??&6yo^pJFF$<3ZI>QP81-np%@?9-oySe5pZb7)*EQ(v1R7{-yf3Hc
z?K(tG)bDd#-j^39ldC=!%1*~gX!KLd>m~NYa}PQHeIL|bQv7<!SR;pwt6!TOhmn*4
zJCpTNqu4oXb1LNT1-RDo3+%PD-ltc@55*rb@YQ>Lbylb{o0tZnYW><5#C6khdNss@
zCDs$pnJdpPjv3<?J}?yZg?E(Q=NdKzVxZ0)yq3TuNAdmWNY@1e+n>CTK+_|b3#e+q
zmMb}9vB)dWyh0###4Pb(<KNiCMdJ{%N_R5@u{kz*SQh+G^!df^C#B=$b2@L84uTh!
z^^vcF&CQAG?v^Ka=pZ#oueEtjZY*qMI_djF#l#{9z1R3SJsrm+9vs-di=2ZU(U<Q*
zPFj}Pa>MeWR?AZ|5B4Rsn?uiNDz|f`6R1)mUYdGEsRXeh?fw}-Z_`{-PbXhsLOyd?
z;B0Mgm;9*hAByJ2|MtAD#Jj0@+hxrfJJS_c6c_V(OF*Nz<`tJD$4}$OhrqnJk=<{i
zR(a+nFY`Q+O6S6d2s+bV%u3~4b6>j2re9)}kRIU#di)$hl*)N6#I^t&f&uwBb@?9f
zsY%Pse2t%=XD7~)a$~;Z?w7)@n9q5++0s26^c(;Ep8kmt_MDuw6l1AaEnL&LQt&`*
zua2GQy7P+m-%dA*x#J&&D)Gu0uX*leP1G8&6uR}J2_I?sShgcYDeRx~yltV~Y7cd4
zt^h~ibzueK=P@!>x18`jl$Y%@xW^MS5z_&ixx4AM%3+Rc0JUxMTwFHyLcEynYVPGN
z*w~{)Uhld!PdB*Vt_K;^xfn!*<lwuXd}V)s9O#&)^cUGAg85V|lkDf>)@_68|8&|C
zuh}WiVHe;X3B&2`;;h~Kdyfq7a&;r*v|@+LaKRDBlQku>!CXx7t2fNTzI`zu?`B;)
z<!<4fcyD(5wnO3uE`-@E1OJWZ)toH$Rl{|#waUM71~$S~>!JWIO&Hz?Ruo7w4)qbD
z36k6My3wl<RJj_J6i@z7;L6~8h>ys@Kw#)H%|UqZc+ba;B2PtJC4MQtaghR+XKt>u
zZlc9ITv%zxRuh<H^?pK!@p6vl5MS0Xz2-*oo|&=7Ylilf5mLRQ*|m|$q{bU;!epPZ
zQ*F$jXJNXT7$|8`IfPN_zl(NyMjTd$j_YATG&6NYgpwN;t+$QK^lbae%&&Ob2l>h5
z`6SKgz`xs<fLsnNvpPV32Ow2C5K~+Iav*r;bjVe&y=<@yKk%D%t&L=d`{dtQWS=RO
zEy{cO=!22rsQ{w!x9;940@jAvDe~1JLKy5(l?vU}n4dfm+d*-IL(tjoT|1MJG_Np4
zgel+@tgfym@N&fu#2xCaYI$fG#+Pib|7l<=?<qbg`qBWW{NEoMK6vB_7{=dzV)Fpz
x_<uhhUX5RJ2Y*3lHLsVwW54M3yNZc7+}~c`m{%|{>HG)26l7GOWm3j|{|`uXKuZ7s

literal 0
HcmV?d00001

diff --git a/_images/d2b61f464140b9d880c782c1347c1961ff1a43b6b62847e515dadb3047025028.png b/_images/d2b61f464140b9d880c782c1347c1961ff1a43b6b62847e515dadb3047025028.png
new file mode 100644
index 0000000000000000000000000000000000000000..56cb41af0305eaf6dfdea4389a3a101464be82db
GIT binary patch
literal 29473
zcmce;1z1+=)-F5=1r!t{rAt9TLZw@g66ux_FaVK8q(h`ZR0LGITT(z$2>}I_6c9;4
zknTQXzOeUR`#WE(|G&;TT-V-fEq$5qoX>ooG465S_ZUH{%5p@<sE;8CLZl!sqmCdr
zya<Bz7XK*xi52r3_ya`P@v5$)hRtoq2RH3+A=hs@+TOEqyk}v`=zPoG!NTVLMQ(oX
zi(HI%9UW~QM0j|t|M>xK8~ZyvpE)bO!9@sc<@Fp8gz^;nKdd_m-{8vj;R-U+nl7(r
ze>^lG*V@Nh=y+q`EPdn_sX6%vsV?<Pf}hUpGR>)2p4%$PBHz2K=AiI7zE#QVMQl`Z
zT13<+xBf5ebHT^cGtY4_2mAinZjbxP;8-wJ&ffNh(0!<Yk|)ODcKc>KUDoCWhjtPf
zH8N=#^oJEkB#aIrL4VYo_<Wd{(Vy=o*jF%rTU3wxI1v3Y_ah+0e5mCpDHY~3Pl?X*
z9$bM&$prHmd4ZTZV)R3i_jv!tg&rH3;Yn!~mzBwSdP>mJ(o&0(VJ_S<SmA(o_Uu`C
zW##<ELU-)WM-!8iDJ3N(g&!Ev4?NYhusFB0ylmyfl~8x-W;Pp-PQHP~xIE?_QpbM(
z)|)Zg{6Bu%S9%DPZr)^~IIoZU;K2hbYU=krJrtc8D$Uhy>)AhK_4H0t@EC=Xu*lYZ
z|9<-R?c2PUE<Jknisp9v>!3%EaPyjv96frO_z_2KPu>l-Zi`dsGJXBILM`gu&%wdr
zn~=beoSbZ8X6EPZjSSX!(9j6pZ`j-2CLt$3a_7#ZWofZ_%c+=`FW<Md9slv}ft-Q@
z0k=W9w@uCF(f!?dFT3%NBr8*2ugol72{_?jTI!Uxq#P&Gy-6E{E{goKd!#Y9+Fmem
za2%_tsS$FVB38P79Sh;q$!`#yuk!odnZY6^wi7f9Pc_;QL^}1om~CZ+-jjA>VnX+g
z36{j}ibmC96OGV=xz{~olkIUKSU5-J?G!O@e+9S5-9I-q71ydK7YlLUUG1G7KZ5xB
z`Hf6W_y-2!%gD(|7M>Ru7hn0^DaY~<3o$e_bYJ-@2kRr_=y<89xHwIm0Q26n`Sc`w
zf<i-OoSg*>3=D?H#x$}#2g~gOgM#X7Jhs2ULhG-y;A1;FIWgg5lO4Nic=;^)i`g%p
zKPNN4b*p=d>iF^FC04!pp4gVA=CO4KRW2mRm%JN-d8^q=F)=av>rBDS65D?+j!#a0
z=;$~pD=VwNK2YJn=KQD6XKBe^NLaZ3=T8QDdiv72XmsK9^Wn<nm6c3_f*G>Sv9YD^
zKfe;xmpR4KhWo2MP831h+}zM>XWu5yX?-p1urOSYGd@1fc=qg3J3G6n!HSDL1;)k`
ze8wKDT4@5eKgIZT=jiCTp3mcmkB`@H8yg)R9vxNRCB<xOy|&iY%lGe}_{FtNuifL^
zEpnEBlV0M+0!L9%QB2(>t%naEB8ZupS>7s~4a9qJaFDw6omumVhQ`LMx7AfuZ*H|w
zz!PZo_ck}<!E0VvUiPc57JvHencc#$G?(rh1XjlXa#CHkR*q$tB)S1f5M8~2`lgyC
zarPERVlR)1hs$SW)MP9zIXrh)b*5%!Qry<(Hbw#%IA&?c$R70?d(qv$f1mfl1>a}S
zNYR}=)*NvqN!8@@Oa5}V_4}n;)%Fv?Jv<T~m?-j%wGCI6e@0hY+NWo0XU84!j*ZpJ
z>wjBP((one3c6)=Z`?3z3O(JeEGs51&SyU^|2!&c<YOdD>2w0R{DCFUZefoN4<m@Y
zl2Vs`0mlPJ$6O7`U%!4Wtgp+#Vf5VJUAJVI{PpW7BCn{Jt+9{Uk*NwV2twb!6-0)2
zHkb8hv?h#<jGX4$zI{7|JbC$2f7a~6#fw;Cv>^u(?%R`F*l>&)VqU#s;pRR;$!(CS
z$v_P$0KTKZ!5Cd6)!9Y%#>Phc<HvDSS1)y7zC=^BNXdlk*pVUk<?BdYu3mAMnGs?#
z(HaH2ReFDKm!G20p!hbCn3x!^&5&4F+57kJZI{Md6g|ku$V~QjHsR^KOH2756VasS
zzjL0Y7O)u#9UV0dGgH^lSR4)E8J+v_ZeeXrHbKIJRoM98d@#0IJjXJr_dO!yzQL;;
zFGe6JD7Z2^KuJJAAmp)iDK|Hl_|&Q7LN9%NYEv^a&ai93ZfjgGxb(0^MMWKF`^jP9
zAf`6?AE$g08HwfK;BarC?7Fxeg0Ql&nYBJYhkxwY<p_FllUuj^x7X)cL`0}7Dk|Q*
zecN1SJ-~>6+0UH<^XTXO?~zhnxpIY8)R{2BeTks#{*PdYNh#M~3E&eFHpnq}XihLk
z?_aw8iE8}UuhEUgF(J5*3l}b!n49D3=;&1KZOypM<h_Mxmtq(C?w#oEj`*<5Om>4x
zM}ptKfAjM4dcz;??Cz4lo`Z1N+iXFKZr0MfzU;xKW_Z&54_#fcGj-W8ndxQiOm9Jo
zMuxI(;ht0}Ws#C%X!E-QLb-eQxCVdxm=V5g(KR8V^TO;PMP=SH{eO_z{*R@||C<*o
zQZ_ku{CMHrugvA|-hKG;g}AJ&EJngZ1U8<|tUzmP>l>Fv(@-it?DqC{-isIgl9HGp
z=I}cFQetIg)qHsG;X`5VJbnL+1W#%Rq;HCf-ajHBwOyUoI;R|gIlX(*WdurUY9u>5
zI}2as7?^l?PGVtUO?^qi5x75y&%nSiSm~70Kg*?8gzL4lh<ESaJ>GkLyo7WjWM|Ht
z+4%L!W*9Pv!(_Xbh;o*u>CVO?{)rQ^rl!oYp;TBZDk_g2KTej5pciuZ721`dgzn6)
z(G~$8y|R0!A(al#%!I(k7>hRctcU!z*hp#2$j?s>C%r&h^zq}zLUyCL5C=$3pAMQG
zcx%=XFJA66Ljh4e<n?P>G}%RkA15HF&o_KG{Oi|ah@4E9E|HCljP%ubh#ot3jQ8P^
zIRq;OkN$%r@sf))=IPT@5C@JT3#+}xAG*3U4Jm1933+&U;1i4@B3nV$u(^{|5+uHK
zc6PG&c6T?zNx_A41R0V=RP;uQCqxg#d!j8m#bbMIda2N?*>wD4WI4p-wLjlAW*k`2
zD3II=iA_dUmRVGkhMSvvVR11_Q}X_x5K;#za=I&m5&uhfw?8AkDm&Tq(#D31LA5J6
zydEk(Q<>^j70k`P{2x-I*>If?#Q5wkuV`+=#DW52o?XoOQ@iD=GL-ha!eR0RQa3bY
zWL2HLV|Ovo;^LilS}~UeL|9lD4Sog&v?34YLRV*cd#(2Wn&*BjV9>}*alNy=e>0i!
z%O8aF6&HxM(`}G_u@DOji@@;kCW!4{tKA;BLkfryak8vlb(?4<&zsN^kHsV(%Pe!@
zqkk*=pqb0>aB2EmF8J8l+CoG;d)6{fCrVUAq$i63BIU;RcFLPK#I^f-Zhq_kT1FQ)
zHyorrL9(FKNNa9lLTj>y>->2vBr7W`YOB_L^Mc~FYcr)@nCR^)|AIgXqPfZa`{Cu~
zLI&lwN$MVV@3JFxKY!M?dt8-~LA;ypnJ6nOqqqT1CJ6<FcEQ2Qf4%a*B^%|xCL44i
zoHk$bTMdtl_&i+xWl=&-LGkz*n~L^yN@60iJlP>9BZHfpm&XSOENp9d*aWHDp0AAn
zbmO!MK&J4)yb-`0W=6)uIksz<O=4LT=YQ|-rkcp-&r{Vr*Z%0CXW7}=d8)sZ^V8?g
zk*{Be9;f7HWMRR9)YDgH{mON?rWJ6Y-EbY&pZ?P0fV8MC-6j6^?HiO0SS&0o_r4b$
zadj=)bvf86tN;kHki%31S9(-f$mcvi6b67A_vq14*aEiB&iUG+03xJqZF%omTid}+
zKvpBV(-H5VlEU)q^Q(Xyoq}YWpEXh8onOD!ugwmip`Z@1JEXYgv}@74&VO`Yi#|9Z
z?zZ-P>pA9o^~I#gt`-*hG0tkv)M=I~V+tRBp56juxeVpFH>SH-eSM<hn+T#1wCJ)P
zH4+#W=Kt=Uo46c%Smoek>)K(ZR4VrWuP*w(8T9|3-VxIbZaN&A5Q_7@kkAQ<h;CY0
zg+MSK8>)8ezR$wRdHlxP`;Aa=jEs-t`d`Mnc=6)OpYNpb!elKhvU<j_ad4RMFGm^A
z^yEF8zJz%-Q#P()bW+yj<mB}L;MLUB-V_!}OGzR3f7XPB(FmFX@(K(N_AM@6CLPc#
zG(94*w_%3w>kVN%8McJ{_3Ky7%~{)D3;TWfqPnv-V8bgQ(BrxTaP+3NwP98jzz8P^
zIm~(e#yWs|`+LzX^l}9SJSXTxa5XeEkh<0=&c$}od6TV`-}om_dYOs>8h3#p#AnsR
zB_t%2*=93X?lYO-MbD=Ex~}L}+d{(rj@-;5<grF?ESyxKGY8QCTd(G_wz|46zf}){
z)a-3pIW1u$_4V}&OG}sGtR|<VB=q%EJzQS+ekZ|#l?Oqykc#9k8M4r+BB?+njshZ*
z?iY}DK7N!pzjNn9W8)DC35lrb<}eyOr0=bbaopD8+S*ZMu-2<qdG??zzhu67?0aJf
zB^L6d(y0>wbwp55mWJfp`#-ReJTn7L=efZkI#K7usw!HDLjZsS0s`dXwhnH27SaY<
znIk4PHfX$yfAT~%!&AG!hzP;Q$IrA+D5%*|G;@Q+jo~vpivGE{xSnSfCYoG-!8QM8
zd1(nrY>ll~Qc^+?K$fL*W^fxQmgVB&;(ahTsJIGGu&}!757qFW*;zDv_~D&2g(#%^
z&iOomZNM8^c{NZgl)h+gY^-l=OzWAHLcgGE@S{DUl8310j*zdfFC#wo<;#~%tgLQW
zBqSy>qIlER!ND&wlH~R4*C>FYmGq=N=8p{(Ydrv08YWltoh1Ld_Cm!(vpAfbjO=b+
zM@I)95)&KST|V0u&BMaTh@hZksf~xvb!Cd0ZvOZ0phy-uW&j0p;k42wCc~r=74{Q~
zQP!9j0|DC|JLTn)5Fei!q{+drUkhbto0^-=K0aak_~}!_Xk*A?&kdVy4hewph?Kgz
zdR{miW@DhMoS7~t;k8%Q<Em8=9j#M3prokyez@KbqU<Ah9Cs+lO2c_BUL>viP$-01
zUREMVsUBM^J{1+hqobp?Bk%ncSCf*G5X8%C-?B71HZ~~fN&qvIH=XI%bJm94tb;>B
zAio?<oVJL^oU2S{_+EChV^y2KULoFk#kaGXt1x?=gp|||cnGu*oL+*>TKC}t&hoF%
zop*B}v9QR7)}bNv=6hT;4-?~#6z&?ob0&HvXd`Q9S19`sDoG>v#jAtWZj?hqLsNjh
z3{|2q;E<lFO?n4;V^WVzE8$MXrV@{mC0lE=ArRYNOL$OgYir{Z5Pa>EJb#|3u&^+y
zDeY5&Bm;zB-~4>;m6er+d9VE&R$i|Jtkw77t%jb{NM~ec#}{{9@s5fjpZi%GF*S7;
z%8&O$L-Ye>)^4WEn3a;@XBoh40<ha0PWOQ=!QD6S##__%xgU2s(}^J5ECNo|s`xk|
znbnl@JUZG8B9GSG`d#d?xw-J~#dlbMKmkm2+0|8qkdSbAZtjGamly7lBS}s(Jx(H$
zm=~C>M0{24-XBJJ(e6vFTGxv)$n<j?z@e{M8f>|`UWmY&G8Fgc_~_?rF3`fW**<ts
zFo~b07QJoI(p(2AG`01280M`6PA40kqGDrX1NdP!KUA&C5vBg%fVMmTe4OADgqPvr
zVedC@cth#Msp&)<>)!hxx7%K`&U%Nrmt+drCEQ<rCncl+ltBc0v^dtxA}&tH;I+eN
zU}%V<fj@nPIJa)yQh#ue_RmXI=iw~xEO(4S2x|hy<&A0MF~{jHf`AirD87fY9{~Vx
zv?+{cwfi6l@n2c3d2$QZn1q%#6ey)(2<S0FcFeG;7d9soKEQ_YJ4*B(Lh!UGA?CI$
zk4}sGs0D?Dd;rX93p{{RED#mpbfS;##+pVT;5DMBl7klW;ipNi*G&LNEiQLRnte%>
zhSh4mli=~VsOSPN9^P1gsg=CEyc9%rBbSjQXd(wr?6$|_gW~=|h7LoR5kTOdUi;!`
zi5(hB7?>Og`DZLlaHtMS)0-wHi9@};y_5h4<O+<cAyBF@j9|j0+C`>^*vo?s9V1;?
z8pKyg^7KpJhn~K;&_rXCcrBVcLcG1RlN7Q_l*qxITFSDjm_S(`bpJ=?2zvQcg=Z%B
z?iu8b89kgJUzzDunYP93{mj#mVm@eXUJ1BsY|P!mV`OHA)9K(WcZJGJ1WIyo5yFY{
z81>v<oBi-D?ONJgJmyzl^0ko0Km<2~J*zTk1A9WZ$ULa)#{odU*kdwNx1#}-nWH0h
zp4;L2$+j`(8&vp0B`v7Eeh@FNJS7|jELz#L7DUR%5tRj{43mHW1yFk*GSyO2rw+bM
zA79~A3{>@5UKF@kf;K}9fWH~sLZhQ8CH8k5&Q)R}Rx-uqr?|iR3e6BijcjCcGT`Qi
z$2TDf=&x(Blf5Y|1sIk>W_WOa$C(wdJ8QkXRy!ckUj13)!QYB`s=!pn&C4%<@_b&Y
zfVpXABuwb|X-p_pyWYJ?5L#S(5m&ly<U;@o6;MKyffFAaLfteqHs({M#7@d400M{*
z_Veev`!Kv9^HcpeQmQod%|7dci)Psp{?{C~6Fp8!PR@-Pvm??{dU|?>S1_bpGDUJb
zff8(fMtongLd>=x5r1Lc$u9W#{rk)AOD$|0u*dBt+v!RyI*)#MO!NUQ>O@7~S(Zjd
zMfv04oiKsN1Pt*W&XS3ViFZc4>#=8NufP8kdoih?pkTERct`imiD(FR^{~7*EiHq6
zaPUl9o}Pg+p!j*l0m=Ggkt!xRnR#P#b0j4~;<AwuJvKHrpccG)gXJWUG@*9YF6~DF
z&smlDP-D-{BM{e`AP+D6Def4BY!f5yCIEG~D*XHwN>dL)1A&kbRZvh6RED^aQAwz&
zgXzUxhvB}|6E2`@susfbZwM&gxHcTUa)wGm6Q!}c4^+5N&s6Kr0<${2yZ*Bo>S(*Y
zorgORJ%sGX@wp8v;eg`37P7-`dvQL;(Li1v-+8wG2xL41$Ptj906#SYAXigW^=^5_
z=HPx3{o2R|Ts5-B%1gGerigc3939ZT-2>ed!^5g$64w+JzvLT=0^S(|@DIiB*UWVw
z$&{1`p^6Cr=AyUUBPzjjivX7RGK6>%28IH)^)3ySk?HQ7zEUen*)ZxfHA-S)Z+IPO
z@-}M@&vQbnRKy!XS63wwx>KG;{EJdT?){0&c2ZI|6&v!nB?PqE$lRP!2Z>?|_j(0v
zna)Bpxv8nC;ey&d6yu6OF)=H>TN>vNeAlV(-@n`1*=0S4Oj>R?Mhux03j6MJKL5aF
z#wlVF67)I+M*aXzfYnGU-G|#Ug$?7Rc&Mmo#F55$?rupF<5JPkSl#Q*2%+S`Mk<|V
z(gA_dlagv=cxsN_h03mU?PuCG$~>H%v_tOnQn#=V0RYrL&K|1O>9_wQW=kB6deby@
zK(nGBwID=G@<h++WtD%>%lhku_kW0>>6wCpB1ZJV{pz&>SG<J;ICdy){o@HHN4=~*
zs$haLstA0lNOxrkf+`S+0K#BP`{d<Wn*|(X-ppp)+i1K3Rum1AaC{JNcy8DAxiG*R
zTCV4F&Z4S>)XWwa@+2YQBm$&Uj@>{skFhKOd__gYg^6fmQ&2fphrG0kc{EqeG5lX<
z+EaHrDm5}bk1a<u3kIBIkd>Ckg7WhwRNB727<_>ih$beCmZE;emeUgr4Kl#m+snx{
zQ<u^(Fhnq2dCUlK3YNC%hD~*M;$hJ}rVwA3YgA1Q7{fO#jHqVw7Y(QkOi<09K7IOQ
zdwUbS22`<uW1EB5()--fbqp$PE1(VB9zN81_c=9{m5^SHN<l#Z*5iEugXc!W9*=QN
z=(lg`3R?%UlkJ(9&*t2_c}NqXz@?7?b0n9eo!8|A0+P#YX|L_hh8@-E(<ATZYjU(0
z{w~7%01Q9<yISDrQ6dHw`jNkXcf|n>kxdT?1b87r>Pz^_%31}q)8pfhr2UR+>Hnyy
z{C7V4-zbLv$Y-M*m}vt(%Gr{hRjOP6>1CW3bQEMCEM&fRUjmYU*M#VUxduqJBh%CR
z8FmtwU7#5|?zfSbo13W09$%hNQzHR<iPet|-y_FuHx_Qb7JGOK1cs@Bw>+m#owD6p
zvC!4iV`62MG21UGxzy0q#C-0YLW(=k=R!`?WDppVGBN_Z_P2lPZx57MJZoC}8{ER{
zw3lek$!vMLbHWXFQ(w`o<FI|dRytYbZJV2$4^Mqfi4l8fpSR6-=~4rba0@ML2_^sp
z`>Ndp`%5hR>guFW{Ys1b!lp|0x%jrWwnBh^9P8V6up!m?p={#nbT=B$0XN??HT6Y_
zJc#Em(^(mo9``ZJkQ&F><IM4f0|ls1#{vM`0M37Saxz2B>rWqF)%M)GFThrOO_8(W
z@3y;^^isAW+u{X!-ZCFnc<zdbI88qypE+6AR5Q(VIzsV3qrBg?tC;M?CL-i!novyg
zxh$AKhLRPJTZj{V@FYI|1Txi~gN8$Y7P;^)eq2n3v}BroUH2#;Ku3=)nE6|0XSYII
z3h0PhdG$*NXP5)BtflBPDl|}&KxP*98L?c({8g7Kt1;ezRGeAUwSD;v<j{|wKVQ{)
z7AGtu)b#-W;8RLn*RY=fTt9N0!UTeWKHfqOl;@yuZQ=W1pukHe*q?B6#Fjr@Ece=%
z`19uvsz(4*wXwA|3K+&^VHgWn+8avpmDSb1z#>4-H_8at=YKW>T(VoAGpt&8f1DT>
zD>5<?w4uu`E<&KiKxKG>UX0!800wo1!U^Bmnklfgw|@^?u%u^wruQU@bQ8Sxf+8fg
z<frEjs-x3zPAeYF(^oprW)JmrcOOB>C@FK=;O2O3how*%<*%Q#=+2I6D)aY0;vL-Z
zOAW~8s6&`TFSmr-R9}w;wXT}YgWundAhy7SIq5;XVrF5vX=cXquRD`C9t&kKl+j_3
zXyF*#1TL7DkFVZszB&K`Sl{>WHzBQ}s{y(q)4zT%zD_A#?DT{Ap+;k`U4OK60|d7L
zR0p4R|0OQt8d_gQDYC0BuCAkF5t1(|4?v=26F}Fh>gq)x7c8D%e=S4;v8En?$dlMu
z0+72yX@#H89BBJ$Ve;{HnvL)jHdTwVSDb)ZgrCo<eyqH)v9SopZB|m;B&Ny(vii3C
zlHNznwYeW55M%4Qy2xI<c=4vB<WX$w>BoepeSvf9(s_HZGthSd&>ZL?GdPtfD*zz|
z$kG$QzJg{pyfE^C>f&ueloJY0{wLr)`w+X>b=C4-e+hwrfB-5QJOS7RC?8HTBT(dU
zZv7y3qtG9u*K0oxOBS=IY`i}`*6BjS!t5Z#q71=ZcxjdGpFR;5RIgEoGf18QI?{dp
z#{=~T2Xyb}m~jFncXzR)MAW!!Y#AKucCcv@rTvH*7<dj>zev+4zoDh}!pKE2DoRpC
zyNyqD233*4-!9xT3+PFehc^nPDcP|`()G^X-hi~Uw8O_}QFe7^lK6QljPvcco0DHp
z#%?luUrxPE&delCSuK|LH9ySGEZ{q<{2uOYbZJ~yGF_+5twB(;U71n~rRG;W+<fK}
zt$B{0j86N&bxM0CDPN3!taw-MnK8XR^Fn3`!)5_r`WUuVIK8-j)*gU%Dlr#c(AqI1
z)jzJnA}r2uO;Co3vq0C-SyJHC{hrE&sXpom?j#opfg`7duZ9ud>QtZ9`3QiLBkCO4
zF%U4!K>9{&Xj(ctECjZ<w5#jY+uw6^c;NPq0Yj2ybr1lGG_>;@g7;-p5;F;r#DWYz
zwd#SJ)2$tR7AsPHeQ*7)xLu^JwPo;cyhVI<xGRm7aCQP;{D<p}_QIAN2W(P$dZSx$
z?7{LXDmrZEBOnTrJ(AN<L}OPnh53w4FY&8|h0-@~_a|^}bYh2j^uMI3U$>y46@Ng9
zJb7m98zQLw^+$?_UsM0{E(l#prkJhA?;y-glq6@9`4&&S6AP#E``Fg%Ow^k7&n7DR
z@I1TOdGpDwI4&XGx4u#SYJQCa=<ljk)5~IhS><dnv%sHs*lM|bWyKd7XHu%^p+d?z
z^Yq(tWSkB*vN5AAvh|?2U|#hVt6r^`j;?Ot)ewsMA3x|IE<kzIh^8!vgw&-_BKe%T
z_QIqgkOaguz)w?sg=1zB1cZc-81c0{0nq24xP$r9&)qFCtb<jr(uU%9lTU^c((fBc
z>*(Ma{us!5JD)MZvK{HOe?Dk@rY;keWtf<lggkf3MUvhW7hkFM+~tKhP-0{Jc6}{4
zI2ck+07$QWeSM5TalN(q*#wn+q1V2r%+;&)-@jMrTOT=cWCRL`(qJ=C?%tG@<!cUQ
zX=Lb^V$S_%ore{tUY*=a8(Q<|KT1zGpgE)as;r4uP%AUxy+L)HkCEZM=`-!=xkjgA
zfLKIH(~oU!ewnTUM~5LyZ_K{^x$rw9!4wK7DpBWqdE2`GQt!aYQV$UWc>}as)pENi
zXp>uJY*a2?PnfRodhWQO83M%VfQ*bgrDFh3xsMW3sfJuc%eLefzZSlWmPfxliF_gP
zjy9e$m_^IF$#838C*jk`lPAw{czR#F<*fn^1=tK;yHOeTC`A-K^3?8J01%HaO>7G^
z%Dthg(;&{jheDnhghHr~y2>N|D)Z$$qeMWJY&t2q<9*<DL5VVzdp4U(6VnOLo;|ao
z#S}GPpFMvOHI{LuaC%m{Ln!fzTQPO%$lV{u_yjvuRW*;M;WrGmJSr{hiWIPZ8sW1!
z{!B&H$S5431PjzCAew>`;4zflvf;ECwGv1HD03S@dezU7fXopY7e_=%Nx8AHp_$Df
zc%K$Qi%usIN=#DE+F6_wPrakVC6fi`jWVFi{i9EKxVLh6UiW}q(06v7;0WR5&Bz^s
zPrwCydH28u)}gZ1LpFjw3|4#?;C$)y6@H^M;ObE`1s@+@sb%fi|CB1OUdOOVOJK8E
z+abvqqVKE5=`YRykwQu-dc0v!{z`?qj799J(+`W{qjS)d1MZ)Z;o*gbU9f-Ig377C
zik8xdcXjnz)|lo57#q;<s@a<644V~p_+<iS5#4H6{*{$!>u?NYc^cVf_dBpQA!UgE
z`&-v%*DX>CGH|w?)8*yck07bz>@&0pujM@htgWrXB=>iggSHLJZKY@a7<NZSN8=&9
z0zL7iFJofBd*|R_)$Qp7D_93ykf*07>?}EZ`wM^;0#U3o+g}Qbs&1*JR%w5><u`9o
z@XE`}<xdNsK!`sP;0crlBek_dTG15)uM&j>Pxm@-MpWm|F)xM-<aH4!x#pj84SrFa
znKs_>>hTG>;J!Xx_l0^qb{?JrO`c$83NBr55SefGy}7CW;HL?Y*b9J5K79U61mZ0U
zPoOSGGZj#@sKyNG-NPP3admN}rKZ+(cN0m7Z;vg_2B;VdcnRahi8y`RN~@IQ=Hha{
zH_&5!J@HFHd6Z7Dlv{NC>I3HMq2nqh6x;@ZfCWd3J0w1UvH;}X*-T1*C3*QesL4?h
zvBY|S;y49YrWQlCUNa=)lAg&)9i{7cZfh6o(WMUmw+wis7+_kkP2p+$%DW<KceUtd
zE;5v`Pl7A`#Z~Fg#9jA&uj6V4I~!LyoSbkS@7g>rZ@355zMQdzpUr#N7(oBqL2^T(
z3>1J(Kt|@*NM-ZijQ8BK*jcaL&$B0NdzRVM_ZOWud$$taA`ZfO<_y0gMmFxcG&ObZ
ziN=;~-v<ruS(AYzr>O0WvsSgW-RUV#oKDzgzx&7^`rj%&sVO^s>-i>NpRkx1D!{*E
zu<23oiJ&?m)a-f7?<0V}M!f%3-*smt^tXQcgu7tNbpqW3rAh+%)v9ajRBncJSajC<
z3O!|7RxDJfr;^^36N-pbmb-A`{ir%)RnziR%;z)!8L&H`Mgg-rZ>-fKMUQ_|hOrFa
zId^BCBqBP30P}loX#E+$O%~t;rMf%CP%2HWZFpgYc(nue79>$ly#Zgwy83>?LIQFx
zv3*V^m7=m&7h3QZ8H8*RFde81=13>04u2N>iHm%C$$w4!*AYZJ-yjGcc4hu21CXsc
z(FL%pK#q+Ov@wd?u01d$8T|gy0N}F<Y7-F2xd9)31@cE^9kgm0nWg71wL;72Dl^|2
z{`quf{zH4K=I-+6?IY|Q7JCJe3VXOsEiDQxNtrybg;3(oZfRVhyrLorq*x$q>j3B7
z1YG~*331qNC+NR4G&xQNt&n&NJ$GFZDR6jt2QP9~yRNFw7dy}DuaZWMkDDvSUN|Bx
zEe#bK9+;hM_xC(d?}VTr<=WaB$jdjP{E9KP-DXJTi?qG(gPJ~NzaBqbzj^B>t|48b
z0m1XAkaF7SaQe7`YeOm;%nz|ILgdfYEyQkVX(=_DOcMc((hlk`l-KL4bmBy{W&!Jd
zMqXZG29FiBu(SAJL~DRs=Whi*;!>-<J-xEW<o~D*WaHgIw}{%Zwtjj;Mc=*EvXHo|
z)jM`K<m9BRNU<y`$e-Pc+vTkI&c^GT`I1_R8K5B*za`oIp{gKM^V?sR+{f|o@HqGG
zE}v#6yAJ<n3_-_#J%jtqh~%qw1i{6k60ajax}`$Df68G_2wa2I6uEuD?S}5z-MyRF
z1W8_L7Veq86KK-iuV22tcRdkwYQ;U=bf95DQ$SxXq<uVTVw7B?qSp2_KZ{{kQ%_)u
zvP?C)5YO_&inLhe(x=h=mS)*FiOfC<E$mXJiu58SHXEzNG{KNcAz$&AT~(8@w=e4Y
zf#FK|6@)j>2yOR%*fzFuXqo){%_NbBaYmY!tN$H>v>SSS$t9&~k-VO!u9w1tL2<vv
zH~ih$ISZ)L)SgHC^rWxui|8o2muyWs=rDA-d4*+pc1mMDNbzf`LUcM|-f$a7gou~^
z%$dE{Ta$o$)l@b0=RE`{T8x-PZHE!`4ZeFY|Gusc0T*D5fHgIUMXA%im~Eo)Zkv&^
zciuefE$P0<QIpHn>#w5RNo9|UuybiA8=M<0*a*EXN)yXU@M*l|8LB_vWKu0-BPb;w
zOwPH03KCG1sXHD7;#6@Fm!D;MSoWQ#sMF2V2y|#j4JzY~BY_pMO}ObCGy?bP;AP^1
z6Sit+sSUVG>HGJ&*XDmlz)QFZ?mIXY-W3%qoDmprFh99sTA<n2py^rqxnCRkSlbT@
zFKV;#M&C~tFzF!5X(#vum;<4sqXX*=G3YDm(H*Y@PlCMnAwqHwjq9esRXEKCV;;1n
zk}UdhN$zEY{J7FqxxgQW1w-qwZpr2!ZL&tv1WIA#MATQVWG*xAfvk}Pd?a;CNGl$L
zI}rMIzrUqbIv+kDo%s5~YT^C3{-mhFan2WFFPyA!VqR&pawX#%)I<#>?g~B^VdDiG
z5U`$K^)om+T3WKK1TirAy7o=G)5=YQIyT!*^IwMfpy<yq<mw&(07NbJ__69Ft}NNr
zscB1NJx#fc<cIm|m(gGQ{5|WM0z1b!P-3a3N4|6Sgq;eAp80_L>(rqK{oAiTmVyZJ
z-tH4exBB<%jDJ=Q=5|T`RBn}aHnK};P)lG{F!9_N!JC?%w%k31o7CzmUj5jE=Lxzh
z@|G5Q{^c7VbB%FmMppUAZ(JiSvAC5vR72Lg6aQZRDO97nIXpG`7|3n^U}ey7r<#np
z|1DpWhO#yFc?xObOGYdq74^93r(pkPVD>L_#i^?#0UbkXHYZ)1uzVRWnaW5-%N<?X
zC;z!~n|X{1<D5^{uXGi~QLq35i8OL_^=?&fv!+-ag9MhMeEU@+X$2Np7V$%CeW~Nd
zrw<uCM)@!9_BL(sxd{@Aa%z-Zcv$W|(LHcFj4Ev9&~e{=Sq;|Lpb9Y8O6V09+2tpU
z_|#*z`Im~1oR2*y+`HemTw7PuH7HyQDNLewPbjk(52Rmmyq~kTS>PgZ;MccYJdyCQ
zeC^uLd}m(GOL1o|$J@Kc)D4!u@_+hmNj6kg|Ij~T$u2BzY!((Z`Re108ArSGFV&su
znA>_4rv&G2^WT>`b!f8Gydu{w;l5>h=T<PlO&Y$Jtt*aB!;A7tpDjdg9X=v|qGqs=
z>)gm6vku9D4XfI_otu$9JD)9R2@l?JGIhD*MWS=D{s=;cMb-DCCX(e_`*Nfmug+L!
zl4dZhf#W~cpl(DizQ%F+ZbaEH*)Xc>-6FzekK2Eqe7qFWYC9=?f9xnL$>EoFKR4D5
ze#WOTJ)17zvai<kc+0lvqK=JV^s33rb_1J(Ej!3SIhU2eDsZu<xj257MmV|It-hDw
z;gK^gty$*{L}Mg3k<*8L5}HKG?4!diE~{(ab99iA2LegNzke|b52N!P-#CU5veft+
z&YCmtd4{{X>0VSiG46ff5>n%TB)+sVVL^Ef3QWs}yZYREY`p&d{<<Z1Nx;8)^X^^z
z8#iu%&_V}_Uyhn6<PlK(2v83sCnqNyBalrc7Ry0J%GEDD1_9<0Xq+fx25@c|M2}K2
z7BpU31{W;{+l{{WSY10_Eh=Edd;ayk<~p+f(MID?+7HO@Jt&S=UA5|24FHS^t$Mk2
zb#>7?RsPzwD_~(mz0Hs{#kT(xmkRS^zRvVSLWN(E{TU>&G#1O%sSx&>KKFNtzv0ew
zzH#L-VLfV`8L+7x0k;ik<*cl@P=@(}LjY9d$6(z7jAn)wLG0{x5SrD%EHKcCK4<b-
zG;R-P4NT86iPI$&uni2{{r+r%&G@fQ8)&1*p82;11)ZE~uw9w~uS|8p93KLwMxGe}
zq%4UtkL`fY$uK%m3Sdc5`Lhu8-s$**2(5M}pLI-H77LM*Lmsb9_myLn+JaB-zpBSp
zzj*<Dly6jh445VG)h5?(qar}^<KoE(q1BEWqp~-a)?I;LpKy^<b-P4?h-?8;LUThy
zm({*urK3+xjpw2DkZ5cx4ubX{xzz_24R`UM{ASz(svSW9G?(_|w341WW&X7`BO{|{
z2~(ylL_dOXmsu#bwbJ;ZkY9{8M{v-4Is%$!#wRAQ5a7T{dtMzDL9cFMkhA8#`o-fU
z4;M$eT2i*&37B3=Wo~2R{d`^PVuIG8?lS#zXv_|ZN5tFNI`nH!h!cyPjJHxkUFK~%
zPImS$`U!^;j@1i{oT9ed+RoV%Bl6JP_|DXH>rzj}_s9t?+kf|33bgHEIBw6`O_mdl
zXC@GD?X$-SXIr)_(=Y$;p8nPTeRykX`uhFDhhNdc$zH!lx;Q<nwrgl|w=R<Mx<44C
zpo*;<9HfPgjB?wN%V}G%e-Q85TCeJrdC)#UI7FSDH*ek~8p1S1bSLaaF0ylR<3Q5+
zZ6?EzAsZe@e+F6^{vAT&{x=Xho|O-zPaI^(Yu_U^H8ryh7_lIb_TG1QlS4KCbOnQ;
ziCAjikhw9$ihSP}zC%T$ZL}Ui9dR$g|2(v6fQr=wn&=3)zQ7~F%*2ER$onzShhS{F
zw59^*4*cy;pFZU%Ed{)sYg|i@bQL@TixQf?r#*V|4RZ=@fBg+P<iKjFpK3@V7<oh0
zPV%mKVo2ofYZ_tZEw;bmzRcf~ECV(gb?m?mYDXBKJ9kbkIu8C5U;uS9NTVijPEm0K
z>{ubY+v{UM`vBF$h3d8O;%<trHE6KatSzz+DP$WR>5TUF%;&wR4T&=~ESG@<#IKtG
zL3Aq}f0M2oR=My&la;|(W%JkW?yI)8Z?>E+pyB3efwf$I%G7*p^QTa9K*DK}EX>yC
z+Xp>h@*F2G9J5^*{}9`JR&2h#kmvgI+=PRa2&FS4CufdqGsLq*sJW(RKjY;m=NjN5
zy82W%UFO42$;zuA=CY!wr%$nUbOb@N5qmf(7jw`IgaT%Runwt)gp-u)31TIrWLuT$
z0g9(?Vq-8<Q53Mzu=hUct%>L6&VrC^xg_phHmn6~LH)OH$9~R+tMv*G#G09$4A?pr
z&02SG%jO-e{U-$0_>FCsP#!BbCU3~ox~ua`H%TgC7;BUC4!UMt%;Xf&X3IV~)3ct2
zv%|j!%Sc#~rw<d-dG)?nGV=*?90YC%e}|z1HgWc03ty=ag+QF`bOG1uT5bG;04=MO
zrsILxVFw)&nOD_<nX8U=XxfPCzPh=B^vZ_HQ!g$qsto-m*?+C5&wY?=)k5qA4TH2_
zgip`vja5OQ{Ym-E-k}J=rxLiQ8y4>7$YCLTxCZBCosIbJXS~RH8?fozprO1(Uv8}V
z1W&3VRM@z#G0&ki{2_e$Ffm_$qJEBf+WwSGq3OG*Li+R*1w%xVs)cuNLx?Y3FZj)A
z=+xcQny#&(&_;BC4=hRYhuWoNl?z?nE_>0$P;+KG-BVONW?=Z^$*WfyW1i3wX5~((
z6?Twr`Q2iLuR;x&K2+nd$tzO3r(GJ+@gMuH*6c6U?AQ9Ub8~wxwQE`*V4h4pPEL){
z@ibQUeNU%PQYSY!#c-uoNvX&=t=?JIrboSetvj~%yJcMT2gZ@2#{arD-uLy+sa0pT
z8{*HS$Y4inQXXFFdnFVez`ae+w$lwHhMlwVnF!UvG)=rr+eub`UGX$`=zeD_qVi5(
zz*L6yNH_+aCEu1i7aUmm{L#}3XAmhB9jW;Y>!ci(%=N66OU*>$XD<BnAw_k@E<Sm>
zX}dK#M?ZV>wmRP}duyG1JfSo88izaVWfYVH2Cn9<=6F^Yg}#@m+w|U^a8(W(bUf(X
zOun6xnMxQ*Ie%rxt<{|Linv&3g*dKsW{J&p?fl8-AFt!#En<j>;zY8mui>72M=v1W
zj0j{g)u|Z)dvn$^2XC~C%b+4Ee0(Kf=AEMxhCyeMm09&Ec!E~Mk$-NDpR?Oku5=x!
zi907ETyno5FtcXi;8nG}Ov*a5SE-XeB*5Pzr=ns7J}UXWgY6f2FDfczYE5!JIk(F6
zhK^jF!w~KIx7Ua5tKW}_#`y}A(L?PNH;_^qdlUW2!wt})KHj*!F{b|5b>MqDEg;u0
zDiVI&gL`EaT~}XzvlT7bJ#h9YcBNa172HnJM-iWMTwD#wS3^+4ADGem`uk0R$Sw!-
zR@eY=Kr*VT#GreClMIoQllv4eP6OJf{`$$k5clDCFm!o88F}<bCqsMk!u+a{V#xA{
zMFF0;q2AAab^aD->WKq~589;(S{8U}>!6@6v>TKC^Rt!#tj-v(`weJ+0E=uRV3CnO
ze?rw#WKoQSlA<V838iEb$bI?|;00s{R|sd+gXyj;<w?x9M+GGZcGvC%4JZgt7lqV%
z_}5E>Q1fwE$`@TXd2#;6W6%LmjTLM~y!`w}kX(a`Q}b2JC(>0C!WR}6EKB*_*6s~f
zJtPG)MLqE3*y7c}5fi}RH|ZFA_zw@8fC#5BQ+Q}}W);HVPS>Y#&G$ze8n>rs_l&IH
zc-_*3cR_UhMaOp_P~P1BOpVk*S7DyA5yA*|Bq-4VsYp3Gu8<DE-@ixeW&g`QrB=P!
zK(udwu?Xdi&fh3wJagvi7Ae~3k!&f-@XD9r<PJy>&czLZ@vslReZ26W^^qdpK;?-S
z=7BB>#q}xpkE@WI_@_}T=lpyGKxkSaTd9}BC<K`*1P6od?v_(n!LvI&W1S4jSI%kw
z>1<MBFmrp`6!&N#jGB{Q3_RS<i=$T&DR5Ai0P-?3HD$HnWMxflE&I>-@qiu=%(WNc
z+QZY+j|T?>%nTe?iM~!Ys(fwDW<e6GN~qq|#D_6N4{FbA=<4>>MXH-c-ol18#gO9F
zU<rN*2%!4yYPXKD#l4Q$i+C4rf2sqV&u=8>Mhjh0c@uzhP}h+*4VolAOKWRW04>UU
z=llAbqB1}B;Efq0VReV+(;k(Mmzok)da)5x9=WmY{TqSc>uEX`OC7^y%mk(e90YY7
z<q7ifVIh$7dT!hRpPBC4`*f%e3~V*fymcH_k{`U=Z-AhyS|4)j7Cnt-<Lm}Z_EMJT
zxt|tU<swujtD~X=9TW9ULj=}b1Cr_Fmf_}`xanxW;kQ2L19JN-*a97AdWgW2jka=#
zI8H4zl!0wAxwI;q?IQXp=cRXc?*Y${KU8+#iqJ+dM>aI@B{pCbwa+5js;Y^l5|WY(
zh{^5SJ&D#><O<EuM+$|(V*u`&Gnl&j+w-V`hFH#)=~+OPD~m<`;o)Bb)zmZ?+c2U$
z@+2lE6Yb-fn(9lm&H}CyD>EiJ7B)kmWm9VDdG4R*H9CZDCl-Y4NjRfH?!HMEE9^9k
zo9nft7upXs*w&UKJ)qu#A%=G_W<xC%qnoGW>8kN>d7X6o+D-DMqxnR7Q*aYe=u5)?
z#<k5>#t-gX)LI55R}@7K7dLkkAi>0(oDd*WbvVnQSrlwNSedStw=s*Jn_5Ck<8|#(
zkbh}qt_~#)AQ*mA&8{xELp63Xm*vmUI#qo8vwDOEmUk4V4m0#9100iss<L|m7TqTx
z01vulGQPe5x%{m1>p+mXKD`zu1sZ?465Kru(9`7dJ1zQ4{*Mn@Mg>rYK?Kryna=}F
zbv!j2%yUE4$8BtE0CCEFTSo1dmieu%QMxsowD%=l%Hu=czB)sjn<IE4R{zBJ_RD2|
z*Y~a8oVN3Ge-~|ze*?YFE-o&K(DPSrKXHuPxP}m{gb=0F&G!FiZkPtw$L|yRF5_jg
z$QCDfyhc}8{MPc$JASD!dnFqh{F38upk3#D=NsT)fR;nTA=jDn&!0ax>CHDptzb~A
zhCv@(;&qzO#`IvAK(hddFz>+~|Hfn68TBQDv1w!B!$~mI<G=^9RM4H3`JLn{$<O(k
zM`mA7vS}*TJfxlzsN6B*wKy}t#*3Jkn?HgJm;wM+8GNuf+NfLQVi&b)-Ct4<ZH++?
zgO`G=ze6_a&eq~5h(e!k4?36JYbP(P$u;G)8CZwenXDeG5-PG`k(Iq3d+z}DM!IJ`
zjrM4^$!Z%E(KYE;bz^V4M#Fc(a{3V3`Elq3Z-}fU%Px)f*JEEZUiMG&wW|Iy*0M!@
zfVQKbezVWru(_mc=qgnB)C%vo`^rn6Ue_ecL&QhK%vOt=uw#2PdhA>GtvJ!L`ed(l
z^QFt(O$YFH`bXS2tvi`Lb<6J~(?yV@&Dx*S^J*@<c%B`xua7=Prcqp_Bg6jsasoTM
zYq?3eal{E1tIIl5zwPA(x(|t=f3*E;x$OhdaAG$$jOFV|%1koc;R#9+v)2hUy$_T0
z5AR6j`9HzCrq#Z-w(<h~7R9FlaRY<Fk5ZKN&ni=NC|_+Id!~;ZAb}N`1N&0_>svoR
zwD#w>?mkx}A+VSzf0q7eL%!mfHRq?;CBi?sN;8tWmOY2-R;q_K(+_2-SFi_z_8cH-
zrV*)=*_Ao7t7%nrG3_MTmcf|+boc?f=~mF!W-`5>J|ZpaBU|AkLA=-*ura!Btz;yB
z;SeP@K`TCdD8Jzj<GJ(3Daypi+Y@<T+l5Yw&KcZG#m}lZtW3JoHG|9Jp8NGZQZgSW
z)5?B(-zpSH+WNsRXi(eK*JsS8IeaQzV5vM^NkrCPJ$@#rsHPpSX^{X5iA#UusUGVv
zVet#yoClvw)w*ndBHXw<fRu7>%*c}7z!f(!*T5mcIWO#Mg_%{K?n_>~q@#zkK@em@
zkwkFLd~~oIPsmA_;K5Ue(z%Qr7F3a%gF8--x``vFaGnl>*OMu@L<^`U%`|N1rg3&o
zzw#aQ{l>BWxyp;E!+H6U&K<UP4%x8Bm^F(0bXcG1Z*yEF3Cg+}sB-SgpC9f9wXX=&
za$UW8Pf$$FSmO<7H!jTb9ZN}&tQ+^ICO@bOPG3_{@X5)cfz+2hDEIRVnRdA0&F(^)
zeqkp+pTetA+ta^~p)?+Vyny*mfj5Phun>b<Ps8aiL6-gp-yV9XU^=|;ILT=x>hw^$
zFmOfe%s2a}kK8VwSSU0<G4`Z5xH+!h#j_FQq@tPbeb6UUda%VaPrtcmb6ot{8c>}d
zcDvli$92SC96`R377TkGbxnEu0I#ovZ^P|qp&08>OJ^kzL+;`-*EJmu-n@Bsu0fF1
zgQa2d@$;{%jDo){z0;br%KYLVmPT;z{PoCaXHtd&18<E1`h9cRSHQkiK2p-WbNIUu
z-t5ju&^7N@1$>Oz{@B7J9l-6Us&Dkr__zj6U+G2vxF~}Dk_(yzou{6~{G|%p3ah3K
zo8Px#g4UojVl>Gw%`hcwR|B9%ng_)T%?azaJMYC|hgtkkA?BcXQDSrF*bPpN7;14(
zu4^|qX&}REaX#)EZ=i0zuE}t4Cis~ml!IPBH+m#FW^c>HjrUu)zqw=zziyXCjjFHq
zWW2r2%bDUX^BuzG{x8oCDi2GvapX(d4ThM3?Ygo~hr73aXF5ry713He;2zy89g~M3
zdF#v(e5$IHmNPsR!g>fUA&o$8Q?q?8k=DVhQ)9W-!X*VbU&xW;g!q1Ga_jD=up=c_
zX*fCBTdIz~j}Xpxr#7dPk=y1HDS*KS)#g~_KXb{LfAV-;?sBq_AXaOxH$94p*80=%
zmmwu^B->NHdS58;kD`W1CaOfns4BoJLP1}N5Vzd%?r|uiVoq~X^IlF)KfIzig!aXC
z2Xvpm-sRjZR0x6BCnic|I`XGwW1-mTO`*+zcjT#-0A({gRm%U8^0m7GwmQ^M4ehKb
zQk|L#gifTKJ6K=Sp-m6W!tX)1CIU(k^$0?XPC3}&=-N@;XPgVdFq_%jSdKL%arp|d
zul8fJ<@y}`XOuM8NJAOMA9vT|EX)o{yNrBy{&>B%6JL$AB>-~7-(rQKBgYRsm>^LS
z+_`fHb$P)Mo_YYH=&+I3B2LF3OMMlRJ!srVo%d*u9~g#H)6&pRdZ62coc~b8d#rum
z5^`c@E|tr;k;nbd@hwk-(zh!3yUV8*FF1tQPH2ndH>8a%<v9*2w?6!{_abj)WcJ&!
zq4qqJfgZ${F_CK@>;9li-ML3W*)VeeZHpp?xjDwh#<U`i`045C2#R<{;~osv13n=E
z6IosvI&%Wr)J#9c@NIyvJTXIOv9M})ZD3<Mr+^q5nV^cYpamw?7igh+uwgVuFvvl*
zfQ5)U|3TgS=l}rdt=IZNK@!>ZPdwlJlWX@`tR5$nNhys(D+1jed09<YPr?!UG-%pz
z<~GxuQQ4N8<*%`fZSt#Uik{|=NxNTCzaMu+dHUHiGG!(uO@6&0!)U-2Fj58ahOPqr
zxGF>n05Pg!0SJac0@R?_d)>PMw(<?IPJxjch2LS1#z2m2?vUIMh4RP*>_^o54!RXU
zz@z_;PUL9{rwf6OMN{ya^%dVa3IEC2B222*rk+oxW}yB!`Kdiw)bphBRN6+beS?+t
zj`caV`l>4rlxRhXZs+M=!DS3SIY5jnrRuT2JvsP?@Xuh;(G1g5Uz2bWGdQJ^GLGZ;
za|6}H`^9Uc__7KfNMGOBRK3913oMMEa9d44ROt`1b@Hnu%-&t&+qxAi=UK(&0}8W_
zv_>&MPsUGEf26Kgm|%HBi!FHL-Cm@}{-$A2Z~}z=pmmdPd*_t)exAO2Ye?2Z{3=)?
z-c4_1w=UZkk?ake_#@;@YTa*+jh&d2f<q_mou7}u|A&VohIF6jT}4CysY~t8&*Xmb
za`{ZYRr$p$RY@mfbC<^DwS8LPpDNSTTjb1Vijj5oC0ib?)u@P_%d;U0d(Zr{utdrP
z>zeqlEsQoq-yHjWr0#fKNi%&jY_fPIXyu2FGt|8_I*QG}kf69a))aOH+EBncv)+LF
z3u;<_^(+~1Vey|LLC}dNwY8E^xS@7m=v;^4B}{8@R$$7)`#=(wlC@gk68;Rlg~-7V
z1nN0?CB-XhK}++m^sz)PH*&Q5DmQjyU+7-%$vgk0#>36&wANtmizrSt^=ugh3t`1p
zjvC3A24eD`rEWWOE4DpnKS#blB&WAnt>nJp@cZY$Eyd?iIB65i!HTVF`sumaT!sNh
zMm#0Fca|(=ZNl%!es13Kdw@0Q@h9HbyVl0iDXHN2)HJf?icRJ!@}-v)^eh~YAGNhJ
z1YdeRbn`Pf>jppJTHkn;WAlh;#VP4auTOmb30$l%Umv>o<+0#LwaSpNbLu1;=q)~s
zj^<R;=1w-)Twkw!+twPzeqJNFBzI@N<3`Z8CHv=lyQgznukF_;$La~YIDcYhr6|3!
zBpw(1=t`6eJLeWBPXW7I%t`$b|F@n^zSn!P3<`eTF?0}@ZNF+t?_D)*q@v+CCrFOh
zo{BA>7aim1sA^~J<R6?Ww`(JX7hGGg{$c*)47sZaugxv&wla$82HoX%@l9pn4MFN)
zTl7CO!Op>b<))op<NO^v;bYfYpMWR`jNSNnG$UXXa4vjoJuj>8%ygkBL^Yo*VX$5@
z=G9dF_a-WCfhOBR{>M=j*r;Q&D2L=tVcZAU7_2by=ogqJ5$}JgH}c=K<>doqZ)BSR
zhmGyA+qzuXo9dPPIN4Rav8RpUzIbzQHc_{+ZMSiy$aI1e{bt4JH9V0@>=pX&DwB~S
zhCc$l2*7c^bG*>f>)ffcYOB$<g`<ykmT#|qoqy?+{pyDt>2o<R3M)@97SUw<)n^I6
z(hZyMf8-{{BI9>vzaFnnK*kxx&V58x(`%nXQ^m{C$zN`-oOx%rnOw_sd!0eY@lo*4
zZn;zXX&hg*IMnpD7acs#wO*53^Skj;Iq1ropOAS*Lkc@jUhD3r2vuYWrVCtyPc)#q
z`k!<}$(dXG2pGb$N_U_i^Gk1UK~W78^v|YLRc-#LDv<$Q0(t?YzjddSKbVbZIuT$w
zHIifEyXC~I2EoUG(8I(3c2UOGB7=f<V~#vC%$|X<BPXSa_dL%lxNlghdnCwTT<>_k
zogiGE*nef229bIc`0eSbso%eGd{wi~TgL?|lALMu=M~_493CE?og`PQ)vwJF#eVLb
zPkw&9thZ^mc$vJ2mn2E<21_UssF<40nxe(I9NaNSa#nVWm#-+ROIAUF&%XHL8~a5i
zc0$7JZW}wbt3}JJZ2N6y*$<U8!g8`0G_<uBMaJR0CaPZA(Z2#bi#`LcW&kKmRaIh$
zH+1s$xRR2RB2rN4RijH1!GeD|k0+N5h!q5ZVJ%@4I0|y0B4uU7XZO@pt=XHNwXlMJ
zbY-`nU#QekxpBPrC4>1L!w)!0H?6F@O`2h}R}y1EWpB+*Gb??MOSNWjC1@i2!rpQm
zf$3a1^RZTn5Q-ceU~){;32DymX9?-l4BZSy+b=?TYi^A?*;Kg9-ryECcGKfF&wA;U
zo0uqN<pI<7kh)KQTD)R&uisE#JFvHRv;C6=$@EA+p;eKRVupDLlzNfT^Z;QK{9jne
zWJf|T3=UXdTFQ<t(`+`VaX+1w7B%Ttpl~p$1@lQ@M{&Tm54R7U=*Od`fi%TJ_V)Hv
z?pcAs5$$Y&{=>+pPm|Sj3=Benhy(jF<%fWiI?<xS!b!k(v2$_BfF=`n*d{d%bHGwm
z*%!dxkB!{^{4$N+0UMoOCj$|8Ykj_;FmU3p!%uEOQx<qxlT%Z1p(ck?F1B(G8WiQw
zZgkh*XJA5{@>n?#ma7sGXC?mCq!bLJ_kJs5x`P541Czy~%{VzY-owxav=Ip$(ux0W
zT?4vQNK`au-569U7_#LHu4q&6-%wkFdka<J0z*RlLPCzixBwrtD|%_74Tcb0(bOaZ
zeHrFnftU_3;th!DCHH?&gM;TXjO_y6#N_VXK<G+L0->2`Q4QmwDy4#H0yt7}tXu}=
z#7#|2XeSc%M-fv|1)_#TRvr~o3yUCl)Zx6hLziJRAhcQf|M_zd#WTQFD5XLAcXV__
zi3oR?euBErp)UliJ!e8e*@mWHGj!$zyycVs(98n$=|QZ=4AMz{5pYnklKc;oA4?ej
zKVn!I0CH`sE~`DYd=08?)Ncp86FzjigUJ;pFQ6-c^4jPOokhY-fz+KiuP&JJ0jvnF
zH1PtE!Pm&nD=17`d4X3Io!A66SzjL?Ea)hMcIps-LI$HBv+`4`)?jLn=%HZ(ooQ!1
zP=?C<ut-`bBQ*@cLO2Exx)Hi%p}QTT4vad>8?ahhf5=>l507LDUEwlFa#&cH5X?DP
zg!XiDtU!>XOrdY0bU^X?b#j<!coj)5E-pqt@U^p31DCP#-MbKQUaq>F)znZo+}>QC
z3_)b1iDh8w)uQqKR`k^5BpWrnKcrzY-c9*aM@PYAankhDr3ZTnU}I;5_DO7FJR)k=
zn_a9&vI=&g6GsL{|EUq<7J|PV)YKYwp!;=8%E~AV4GgA#bHVadtwAxMzAp$^VraB>
zZ<Pn8`V39~+&w9=kx1Y1$dZGRQ3_Y}caFfgZ9+3{y1(iV_pPn1U>L+Py#x&2Wn4xu
zqt*6XUym2l!(|LVeNh%xQY*T4wdV#iVh1w9DOqKwGu@^89bS7ZHT!$prk5}K%5K77
zgR9WyOX4jAbMDSv$q;!051|GPw3JwZXe$kA%vr@&D=(M}_7#NplA@whP^<7yx3;x4
z7T@VmBo|PS%C<^)`ji=64!9vx7ngEbLm8PzccEePn6GMv<fA7g=mbhTh$W-YPS^<D
zVXF+3R##4}KU6Z&_XZ;L>)@aPE@Ql;SB-jgbaXVnlo$x^qyv>+`!$(RHgS+0;}aIv
zOUcj7lt*vC$;qh^?<BoEg=Ba9^WHr>7+}YNO$@+ZH^W(;de!L%821~Y)hq<2k5EL=
ziGqEP;OSO>Ei{xdBYa$3O4F(ers_u95V(Sn2AHg2na3b(pNcQF1C_)ytzdUV|5;x9
zijFa=7z)OJ_5u4uM5v@~!Gr@W1cMUG&>4%+c7RUs1EHrk^Wp(agq1<(BU$)nh(s>+
zONL`68NzvWvJ6O7Gwv0KDHrHLd*__TYWH*nUOU?BWmCI*6zVWS1_^^DT?Ku0BR3oz
zKg-qcFuCvL5rymV26M|XbtMQTj}a~cH2Zbns>RNpO_v@>-s)BDg)XTZFyGc{Mx^x0
zx6FtcPxyZDDDrF!`<(=#LVukx%Sis=ER1qU(&YFGEdnqv;3iCbK>J!>+0+C<|0C1I
ziyD>aVXj*1kl2P|j|**1f}nD@Yj?4k0qyiaN7w*`-NR>TX*q76kkOl2t>IVxmuX^R
zOiW3KcVq=El!!O{pm+3St8#L3((Ko#*r~28QuqhU^(UG<@J(L3tCR@ZR)%UhnwrmL
zccCjB?PQ1Z((}J+JM*Bd%e9X`f+6Jw?ur_@A((;-nH~g0&;SwH#4r}2MZrBW6$QjJ
zz=&gsBA6ngYyxr)xC17enkgV#prCNv9>636**Txf&b%|t`_Gwq=k1TinZ};y_j~U9
zy1&=*{oIp3{XY0mMGf6|8NI~9Ay&OVUOdoZSXjn*8LwcR`Q5Iz`fx`3OMX-szAeAm
zP<}H=bqXxGXEST^^7<;sV5(!*+jaW{H)Q_R5S>!}-t&)kKN!jIKG>48t1f&u_kEjV
zhbr7!zW%JcqvaxF)B;Duj)r%&?(`8iv&D<mFjlxeU3<xLjD2rS_~_YctVDC6XJ9Z9
zdp&+djL*54Ts*$`LdWJ|Y21y+#!^x!tK!fEiAeAR^Qf|FRdvCjkV*8J<_SL_a@J?>
z5IKq}A1zn4*MlPmA7fW%(*+@MxLnvUkdm{wS?(IW8bI1IZ%We94#J2;(Npd@{=#ho
zD&XnqDTW{nN~_6>!^Ta>^k9c6gnMNA7mDZJ^A;^>#%UGI-)Lf{7fTFA=~GL>Q+LNq
z(oIbD1BdGd^prE)BR_O_iJE0gtZ|RuM(74Sgmcmf25}XRV<s*PTXf3@;Oci@)$Sfi
zrk&5uX#V4#Z7%O!X4%XZEh^6Rai-Dn2msFi{6_4nYZws!+Oe_gPEKnW{=e+pP8Uoq
z?d@G(Go=D^txtJ*dBg?}r{tNmYLidgW{|AHc+K};AIf=^v>OvjM(VQ%So#&Fv>(@-
zO%Ac#(FFKnfMV2wopYeo_Lis6fSKu4@DN?Y!Z9Hv=$B4j=k3XYCaAu(+-9>M3<R;4
z<Lppde?aNJ`;#;x)?c0AD__f(a(_+V0`{g)3h385_lMCMx@lS7dNoO(YyEuVhT`Q*
z{qv<4S>{R6j;Ja5f8H^#=(_R5IL*%a<Vm&WXLTzAdfLA@-BC5;)AqX--$s?FEOA_T
z%SS&nHuB)X_XwKzh}_7HSo_37A9{WtR9fMxHRID?=kQe4*4Ft1htMGHU|JZ-j=j`9
zLKQOGZml`L?BJmae@2q|+x*+5!5!748Q6vqy3}x8l8*iRpZ<F2Hwh>j=nvvNqfH~A
zJC#0m-_^WN-IRWHPC*aQ+lWe#H=AGFzZ6jwWwhfT>ZTw=&#rWFNtzMB(zH<20aj^f
zn3FIf?s98;dwVQE40Y3E2CIWsu2@m7@yKDiW&7XK1`drpV|~k~mnf2)_T{#1=P!5B
zj~dz1(qgetBM+odMfIiSzkKA|M9kM!(U9OyI9OZSRFk!;w76K)ck9-SDd$V)80qdg
z*tvT=`#bjBx#f+&{xe?PG$TD-Q)g*pOwD<Y1#rA`T6<~HBzG-|ii*;^eDPwi!W>`K
zn-jYdyn<%nvnNlcoQaR0TlBG=FCU3SN$GjFir|um{PXUPJ+tsu?T+=}dR;jm*cQ#u
zL3lyL_mQe9DtU~e1z!fVr)XyFY>4@1tn%;5s%a5cSbLf$C^Q2rPc0cNfA$2!TKd`5
zR-2y*{Q1mDCFQW~`>Z+CRGZ_Tm2>~i%h$%5SI0WE;49jlTprBQ_Jn4$otB<er}6X8
zrz}|Dx(m)~9^{|kclMbz7M<KHM&JV!6kGg;I8L+wq?2!*)LYncQQvz$`l?rFih)h^
zp^EIdIP)&ZA?7rc<<wq>Hx|X>{9ON4=^P$A!s_(J)wycvM2=X_qJ;Q(9TyjuP!#;g
z!Q+i86<?ScGPh8m)6=_yx`?v-vHuC3;Ml1v0`mUkuh`W;n*V=ehyS;Xyx0-b;Wa>Z
zAnGnC!sE|C*-tlbzRjRSzC63p2K+a;#L44Rx48@HL0etAY!6g7Y9~dis;Xv|7Zwy;
zzjJ3$koH5EK7mt12WojNScT??M{~v|8fZC)YE`_@YS2*;@9IQL@&X}!iJ6fCX<~Z7
zqI|cqoQC~fjn{K?-|hVEP;k^p8ZD)s<K}JYo0^&`6TL#PlTklxm;xG-DSpI`7p6EI
z8e!wSe=M!|<pGlr1|HXh>Gm#iKW}dBF4ct^lbTM=`_m-~P$jK79Rqt1q9MDxb=7}B
z*Yv=@hOo5k*oS+7|H3_S?1=d+Rd;fX`J{1K1%m>E6J{P;;5UNbZwbbm!!uk%aYdR@
zKXnbL%X@n%c5s+{`vGrdGs!pq^}L;btocJ()}L0W?}p@bkl*gWrgKqCFIiqfKL{IV
zpCv74AsJsUC^%E>Nk*6Vvwn`yD8%>1J5DdiDT4~lMikilm_kFxMw>^qE!fcx@M+iY
z-@jVysU5pNm!6@^Soj{kN=Iq9N*nazrq!o+ANPAPPwKlc<OFyjmQcP_^0Mxz=;*lz
zw9b-rU4yMaISq&A;YpW#lF+kS3mzFkex+c`>G};X`J#PRNXXk$Y@!?!ub%YW_)c#s
zLLbg+al%XuyWB_9K7}ef7K?>YT2fM`8zYJIUN=DKBn^GbZdm%3$G7xeSHqBi<unM1
zPI5U&78fB=uj%#0nFh4n@M#L*U5;Tg;jJ(3kCp`~2dY0|SRgA#ukU|(?%X-yp$L}?
zhb~!(YQpyQYhke}0zrphCP>Kfo;~6zdE}@MH)-<(FKhgMb2@re%<(_i8Y3?U!%q=_
z5Wp*#-Tp<D`ntMrBdG<si>k~#p9v<2k#Sl0Kabx%u7NXk`-@9YJ@`8)KrhaGB#TSL
z({npseJ)=KPyBGkgN5X7Yuw(~!onh=HAOi*Zb_jj(Nzm~RuC_b14h#A39#<G0v2kN
zJ~c*Yg*Cj4+w92DfI`89W#J3w=*och$0NnG3v+QlTpt;Z<p)(LC~AlUSKb`Stvnn$
zM6bo+Ub?G5r>Liq&-M55@gXm)tgaSP^+o^g%jI5u7n;KOv17KIE;&0(RZUGzd{c_1
zG)s`~n%0Nk@1s~6{eeyW?7^Uh@a$--U-#<OGWhg{2CsoE+o|DxaKdUDPq8C(u-7SM
z5rsK#&NyncV^(13O!caQ>tYXcIJ0z5Vxlh0B@cKW(9<+|^rWprR}BZHgpTwoiL(^O
zHF$_nrZXloh{9+8bZw8S-O7~-o0n2?IUdlngZHZnU))qMV;UL?etWSp<}-HD$I-}g
z3?o~LYhWB%K28C?I8I%f6Z7)&JXivNEAxRtp4-(9mxQjUm{i~0^|B9ecNv_4EjE}_
zd=VLw%q6sBGe0(J-j+#2h7N6RNXzkF{~NT|eu}HDRX!2Q^BZfw<C=mPsw0$Sq!_?Z
zeGyjuPuP<sJ@g#rpILe|2z;JoCLsR66AoeBt3sc0I{N;K&-iAeFd5@BFxb_$iD~*e
z?3qvnP81kM&-c|MEmkNnZ<k#hMC&EI=@frf1mY}u#3YMh=x>{aX>r7D=iBgtm=+y{
z(h^QwJrhm&r1x*W2|^7I?j~Pf3%DI>gT@9Ul1me7$u+$ghPZL}ZsF{^Mtf;z#5yIf
zsJyyk_VFE0MyFT~S>Eh7%e!={U@oY)F+#V!v4MjqY_B9yF|e1%oVWya9f$<edVzde
zEIz^###@^|1UNweFAeCc9XrKr(gg3-6tO0%Xs`mi@^<r6S?n`=Peo%R;d};Za*n!b
zDfGA8>(@&kQ{Ak?g<zrZG&K!eyb6rjus*?T*^`TDz5I;Ip@fu&1oYhj6R}OWg?+W}
z9G$(Sew5P={P=m8hr=$T1ZV2$S!M0UNI@>`t1Q(y{+mgn=L>X-SQ|A18D76$IDR&-
zTr+Uguwj!%3?IHWuMEMU83+Yrp>I2hXokOA?9|Y?8}4iL5f51T)D!c4lhjT7YWqX7
zP;NPGuPYziS6gZy^^u(Hnt^M4e5Se4`F1@X5i!}AO^-OCQ&#-plia~3JP40dI@Q@V
zfY{c|Q!0NaW=mxOZHQ~%#mjhGV`S^#)OYBX*E!fOnEa;#V}S(|fQK;zm?xNsa6kJ-
zRzn@F%d*F`t~}|biF{)W`~da7;V6o!p~Fu!a74VXqA_y1OI)#=Q*UarvKNV12m3%T
zUu(#a2^2*Jja`|z%r}LaOFo;!s>P{WD@{)=!t#eH<l>|dPD#<!0m#MCFB#*l&BMO2
z^$1SLx3snGw>-;v3+PCZ0SZ$D%jphGG9GO&A1T-4u9OT$ifdQC(r1{dv$7RYE*nnp
zX0pIDGczwQqk5a1mL0?z%I<vex0azXU<19V`qARK#%Z$hxlz@Hjo}D3<L3L`b#xm+
zRTRpRMub{%sg~1MT0jy7r8!4GcTQag{3nH?cGs)0%U$e)Z<3Rz(bE4rNaf-h{IGo#
za8;Aad*EcH*y*UN7<W9INEEC`_`xv$`tl22Szm=fK`Adf9EpyOu4l{d*H#y)mziy~
zBBHmuws#-WdS`WI<tPxdia?W|wo_G2M&EIw^HJjYr&urkV*cy1+8pSE?;CI%15-<>
z1x<GtWdf_Yd?P+233QMUxeS4c{KTCr=23{_>#HazTq?I3%YpF24-Th)UwiM2kj~K~
zOAzekUMB7T8Va$dEU|Y<7L_{1c;m+oy88Ns2q-H3cN33>L}|<9iSJ7U!wNm0>hsg#
z!>HH|Gu#l9IWM5c2t>4e%?wJg_F0^goVMTHzP(hWmuY@y4;Ph4C}9fs^t^*Qf}aL5
zyXnd8(b(qWZ*>d2S0N|E)~$NI`il*xP(zrloW+-~@<}^+_Ur%_x#N)|hYu$W%>*R6
zrpdPYDK9KsiZy{95y=p{;9>OH-sbX<mc~`#(o^y$ojiH#RcAbC8%Mn_zxt|Qpw-Ql
zu3I^H$l-&({WgyqL5hgZM9A2TjA5+!y>!e;x>lU}AT$3(W8;7r&ttbu+|wM(EH}G>
zUz2{W0WkHe9m-_&0Rzkl=U6<nLnm-T>W$;(6YfN4xc~YTnAH9#b0HUv30U_bHdn}Z
z-%!1l%65ryNWu%IFYFH7BWx%J_giAJjBH#8ZKe(1E_0f!0f?jT^5%?xZFW?}<lg;N
z*usKv|HOjEgyV|D4H=GJdSgUB;+!3b{*_P)XXxJL6Utpr?w)Xe+>U9;X=jun8fDa|
zQ69tqWxm1drs{79(H@*|N}pESvhFSz_Z-YB*495;-5c2ZYSj9n29EhPHQDwgAEOm9
zz!R#P9Gz)KH}lXFjEu%oEv?K6Fgd<csVldD2nmgG;LPfh)=7&D$3NZd>$@P00>gD~
z*AadqkyV`HU8HMrLCL0u^676HFQ<hF8sBVgt0xE9A*;!BT}wYp(wXJ7pnYO&LH^o9
zAqn}6Y%ZpMD;^XMqUY(IZnQB7;|6o*R&-QZyC%Bz|B|x@PyYBT8=dAh4}8qVlv^+a
z7A{42QRs~PrKE=|ccDWsjhVa>F)Y3osrMkaAMzIh>KeT4lYg0Ttm^nI!(q%Z%c)EG
zmg!uX$_0X#*m5)UgYW}c*>|Q<a)P5~N_~hMhQ5Zcs7WV_{xJDS<earb{5ziO_zvlu
zx@b}E&XBbzAOW86Xy<S=Q9-CjYJiKGs5W@~C;GdG*gbe7X}%z|2|1Z6isYdt+ot@3
zNUD;s{lVgD8&@+k71nabVb#EtlqPD+-EY<A=ANob-k1FO>5|qco84L{2>-6;C0G9_
zu@-*h?c2A79pjs&+$G$wuWEM0FJ7O+c9zmT1!2F7SFY4wsP*@&8?Zd-?xmzHWlw8a
zf~Ynb`Y<tZZ*Uk@c5k9?;mSFEQFEe$i;Ih03fCG=mgx$hx|q=8kJq)?CFvLpyh-6s
z21|L!H%>q*d%!XZr=?xBPmVhrtbEt~y@|0Sp6(mz7RczTYyb*I8MnGUy}Z(yD_0MS
z86VkgTU9Vlm~PCp11Ao#brPK*H@!@lpaR6NEHr@hj|&r|e%hL6sS+5PQ>t1t-)Y@?
zNqTDiJExo5<q_~hN#K(k90Ss8s78fVJB2Ng9?W5V_dy6iXYj7+g=;#y<iKD^RWaN)
zC@^&M%%puEQJ#5f^`=Qi<@M{t#I<4ZM^aO(-1@DzojBeJ#0w-og^fwOeb#r0O=VXY
zND-+0_pQzk?tWqz*~=2k=Y|-Dev=PI4z;^?{Y@@}>H9xLOPMEaKGEA1Uo!9xGE^XM
z!$yqoBxh7c*9LS@jh0uJNay|5%becM91%{vU3Ysao;^xWXZs4ba|t$dxGJYw^oyt*
zIHWbGI2t)gIT1XN0ZCWOS#Ld(C~la^DCjY@Fq;y=1HVPD!3@r<;6{O*Dik|g8$*8D
z+ixP}Zl;XQz=xmq(s#SM@^O^DY(x=xY-wj7wS=KM%hr7B3Mg#!6>7gM!6IPgzL&ha
zkBZB5*Bu%Xk~+eWaihMw)6XgKe8p-eDIyJ4tFUK2PMm?fX6jw`3dDcI&&_>89fyu>
zyno5TYk&(ItA`7*xGv1Gq$jEKs0l~yv+@)z-pO8GHonWc5L?@jgyxqFvw9zt^+0t|
z2B#uT0*vKC6SD5y8#mO^(+>JUZBV*bQ^9mP|B9pH5v$D<$tKiC3QrTEcXDY8L`$1Y
z&!cQeai=mCeF{Z`zkd$^MxGnA#etljj{Zcm?G$Ru3PotpESC8v4y@udcase7vUlVX
z9}YTTZYPghT_QJu1b>}0TnMh$k(!i^MACx*`nch)uyW|A=VMBdkR%DJs@i~SxH#kr
zm9wr_MnLPa;1PtlPbUzi^pGtg_4;$GbM-P4$Ak+CHf3}SpSMSTPg(0IcY&0m+hUN_
z5pFj*7RS2zH5m#A?ODm`{wO&iRla<=DelafSW@EfA17Q+<y1=F`rKZig>9^-vLDQL
zwf_1=__*3QN4<1<j_Yc($5ERS;sK3@gHchTydU}nLrMG5?*HrvS>zx(R!XGOs>Aj?
zlOl&=`0R#{U0kFf;@8@E@NAAC<REv;$QyNMo~%CyI63f15~eYo*#shGzJ2EOC@XWD
zeoQ*s9cZDtjri9RFExgOK)!b)X)7m?{15^pk;-}axS}pJ%mhrv09$2WZ`rctCEbsD
z<eQW6?)ImCElE0Ne21UNmm$j15!xn@Yi=;hgU09lTz4;Ln&7KJ%;oY^5zPu&Me?NV
zO6RC;+euSJelQiXPqIfE5`X`f?~ka#x~$hVi|n`~r@MPIU1mAJyMJBKQtHsgIG1!X
zV6;$8lHo4rG~M5Axkp1*bD(ly%?b7cWWD|<VX2{j@)gSvxxHMRwzijuHUgbDk<n6`
z$a^S>o<tqDQM4AY4&IK6Ce<C^vM9U+BM6&v)3c)Y)>WSC!<Q?XUtfNTz_TB+xHOTa
zfFRe+(K!`&66PWqNyD9?Vq%iB^MI?Byc`0TxO1oiO<+h(Lbu##$1PoJ-Pt5EbGEZZ
zdD`FQ0+quQhlR<;CUSe)O+!t?aY1cyM|?tpKN^Ax9w_&U@%z^EW_%hqMa2oegNhv=
zdX3X7X{`A;8ii63;iWqC4nL7F!x(ibMI{M`>D!y@7%&MQzZcmbJSNyX^N;<X<XCKw
z4x$j<%g=s%tWW5erOTEL608)5Q?~KqNvdO27nwWW&F=c$ntMsCyN?I>Ry<#$%UL1^
zr|ffF2m6?K*|}nBY)1y3EIL~c5&V)HGfMc&Po3touIZp)TN<I$z<V|$Vo4jXCG?5o
zdz4|M2gPqw#BL|vG|!GzW9oE|JO;uq60p8kypvAsnZ5tHr|P2*+WwrC|Nppn@xLmb
acA3l>x$VsCjJ!Sy{;{-JW}a&15%h19`s2s|

literal 0
HcmV?d00001

diff --git a/_images/d8517178d972bc0703903cfab70470e9c921794a84a0af190ce345f8321a14e9.png b/_images/d8517178d972bc0703903cfab70470e9c921794a84a0af190ce345f8321a14e9.png
new file mode 100644
index 0000000000000000000000000000000000000000..83409406a6b86f8bc3cecc03c2db8ff61b4d0bad
GIT binary patch
literal 27526
zcmeFZXH->P(k{BlL6RVn1w>Iv5+q0zFd!L3KtO^Z2uP4D8Hp-E$r%Mn0wNhCgCq$`
zMskoKNRA4^t+jo-zwXn0`@7$`=l-~39LMPKwz64!%{gmUJ@wR6^;A{)1}QN;F@hkZ
zH|4IXBM1&Jf?$;q;={jKo_-JifLwOCrtNUc#?-;t$nHL(WaMD`(8l4R`8^h=`*!x`
zHr4_>f;<9TEDszUZ0$uaUbOoA2Y77k%r16vR`kF_h-~HV*dquHCHg-svy>iqvc=}j
ztI`@SiOXY;_4@h`rPc;{c^f%`yuG;!vm!Yy-n|Skx6o=T>G+U3_DQ}oWvuFJKmWF=
zxlXiBSq;zFkGI0q_MbJ*sMmYQHVxfXy>jwvOx)(auc?WNHx_;Xc~x>4L6~XRPxr3J
zUXt$g^PThdT`!4Paf2{_&eV1EVNo1M|Ij@nBEkHhoA^|;r_n#*B<#Fb(LZPC6b&&S
zm*bDE!+iRs39k<;=0k0U*s_=ph1cT-{CTGD|C0xGC@~?oJNx^ofByWrwz0wQcpmeV
z>k;AM*bEE|qt$L@1N$3fid4}lDPcW5>bC-=(GQqtq^GC1e*T=_TT_M8>u0`F_3`7!
zSP@FhyUctat*}qpon`vl``<jr>+kQsH|b0MrNH=xwvJB2*cel9rb^55NSS57^3|)j
zlD`(%#CJ#SZJnF~6xn6>ca|+jRnpa`Ha|<<7<_lv=>B~?Gcz;A_)A3ax11Tw1|3&(
zG?5`RT4Y&F&B`KFk@|fpH8pi|ZZ6=}E0!~7&rbDa--5?o*U&iO?%_ddk#i&LRJqG~
zs&ZUenQ(Y`_}kGNd78!zenf=Fj`{lf;slV;JqZpbIoKJod0$v4<KiOnYwP>16xqIU
zof$S0bkTi@ee-VbOjhf-?tVC-rlz(qQ7dIVTJB>tQpzeSO3$hF?y9S+XogCPwXJSr
z0BK`Sy25(<#adc-O5uk?-tXV@IxKv_8q7Cn*;yWW^kazcV5?W<{Q2{Bkr!^MYHDf@
zHnL&1PvXpDf{ZtBaCvxmTHD&{E9~bfXlb)<O;o!%>wRz_GV4l|&eyND=-(Nube@`<
z)0|<7ZDICegpHBa)~1z_k#SgT4yN;kpBlqH(n@){=l#kLa>kfW)@6%F&W54PQlW{7
zjC((S-j>%#3}FutZwa@ZOL9j~pFS;jTBgs<&0QO{|19Wwgs%H4chngw+D@0KWC^#$
zyunC=4-VPV`>Q?a$GEt-=&QZZw@!_g4_~T#^X3hLh>MGF4Oo7Ph=@4b?R9(8u%)G?
z5iEu5I3;>)f-2!vI1w>11jzf+(g9r?V`ChowX@SI`ZpE&;}Q+{aqp;xpi8zSXXKyX
zxB>rUY-xE6fzA4&o05`3j=V1_viz^#P&BNnszQ)5>#-i){0qA~JJ$dBvk%UzWQd6K
zO5(Vy(qs4xeen?-)vMy-;_m%(=(}5FpNXp@VHBa*+1UvS3v2N|M#Ir>Gf~5Y@LKd`
z;eVfKX(0%c*r6ppcB}yY`TA;YwtBWkP)JDsVU`BRtw78kbq2#8`OEmk37BCc<E~5G
z%du35c8LWw&)pAx(!Thwwet0cHv}yQSXfw)&H2y5@+9{2U3f@Gf-t$cx%u!$l9=<o
z-IcMm-QAX0KI7n?7>$ojO-8>}jt^1Ad2dTyW8vZH>SIOsNKJt4>z&q<T-x4^f#fqE
zpHeM+d5!br$&(j}iNOmC4@AVojKTawN6bu2ScHU5W@l$d4gab=+Lt<*IQq?uJYXhL
zP*S>9>-oEQBv-dg*=(xgdvB)k(cy2n39?tNAb;HiSXSMqPs9nrHUSkCCMq^b73HoC
zW4$eXMyT&m6FA!@#)}x5n?GG!v*pq&rx+L*7%H)(36F?4PEY>~7FRzwI5=5qJ@&yL
z-s<X&8<s8kk&%%P+Edfg)+g(6k~}&(I@Wqsq>N@iMcFz#w~SZ2J^HzEKbTqS6l&xX
zH69_ZRhZ4iZS|vQ=#zfo8SiNLR_9zs_S2A%rsP)Ka<^ZTU7c5hi8rogDkmv2okX`s
zkZ=2EGUr#ZKTfl6<NvLT{NLtl|IdTa<2n963~VnPuy1K=2a!weAxvUUe(C9FN5{ta
zJbpQh)_7D7Slz#We`>jGA^@Cvr2N^c>S{i-E=ChGGp&lxAq>L2R>PNE+}vbetJZaO
zk*SEU<4PVK*n<`J&9wdrrOa*Yyw|Q>BP1hh4nM2#A|@t~&$zj+x0i}R*!rrWAqyoX
zC5bgB8yk+tudVtR15YA*dwV}WKlE`qTKU((NMFXq-7!z0I;YIS%S(|UWc4H@gcQ6m
zuVEbmLD6t!w1QUXA?>G6pMoE^Vyu}v*@P-KT_NT)2M6)*{oNLL&-Yd%LU2Qj;U>tb
zsI+W-?<KXi>`J^G^!#~a`J)*M28QPlWu`twamET-o<xHVgayrEv$>8qG=db~Z@1lE
zynV3hb<}vV_TWi*xyatHU*OuXr4H77`N-qYwPrO%m|)*hSAPOd2tfwk-94tOt7|m|
zvAR1?R~X)G^2gUA!P2Ps_-EX@rB`078jgK%<P{bk9LUoy!0zhm8mf9M(9_dXJp1_Z
zV-c@I@!Y&TKAUmzq#g7??B^7bI`AV)j)|B>zEKD#>R6v?c^MIbBPAt;i;u7QHig#+
z8!`NxEFoCxxHc)B)o+1|hxZy5Ch91kPMLN+<cWNG!{P7!lG&3dPx1=n{Ol_+@6Bko
z!kkpAZMdQ#kA4*)vhcMi6MXqk*h2a@-0A91Gm~A3@$m4pYuqckcAQs6>xWA$jf{+Z
z;J7N951*pH-fBB1Xw$RxzXT|T+&r6_pZKzYWK9P@H`h~+4hZgkaH#hsAYU7?sjctq
zB!vhn+Y-t|Dw^DtB(}sI78Vu=0Z7^k1HVde;gvQvHefdg23(h{9zH~nii!%-0Zs5m
zYx}FU#y^G%Ao4K2yIYBmyl69V`Sm?BM^PTLo~#v!I+ySOl{D?G&rE&){xt1ISgP~t
zc)8b+<UqbbSeVoy?Tt{z`s`bolRtjw8SC=#@wv9g;bJb)_waADw4J%eux!_*&CN|L
z#L3Btv`5p_l$~8UF$j<vqlhhnf}$TketfC%aO1jDh-*{5eG-t*+E=p_I!{&1F=?>f
zWarh_*GIf{%dEdcaCdWa6PD9Omkp2lwY?PRv^;!!*lhaaQ^b41b3gCz`w2_`m&p2t
zTS?v6#wOpHS?Mt#nvg#+{pHkuim&t|^pfr(5DF;h=(3Bn&mt#iXpF%|UbN5BUbcHv
zItt(vu{}I^oTc$r-pb<S`-c-X)vp8leq{A?oH=s_jc^481+Jel@IlS$Cclr``4hwa
zD-&o87NP!6F?i6t9gaBSoXYi^H;KmGH<cozqRf}(Iuio7wj8RqdK7FQJ?f4oXO?6H
z^8{$5U2e<TXHinh1K-K)zJ}!;c+8^oDxG=u-eqxox6@Sb4>JiNXvc}3!N3aqGh)ZS
zKEXR?ySHYPDDnuKNz~zrswxE<Aeop*0na`hdM|CZ51{j7bMt9I!4h>{1NU`2K(X(Q
zTgY&6abJNo1U-FPAH}Kl<+Ut9u1<-(nVYb%u=~zevsei?p=fU102F+J-`QQAXlaW$
zr>3VD{F3d)=}VW2SyiMX2AjV!tt3lkkolq}7k~WbxDOVhprGIv7-&4z;139_8B7z!
z4PU;fCW<+e)zsABm(70r=65So*=Qj5&iM-$qybdg+1c^B?L15rbtDcA4UO7~yKK)M
zdP?Hiix(6pPMnyWPw{I015<-aEC*2(!z&!~&{Gbzu)Z^G|M#)G;k%snd8;oKx+<d;
zaY1U*H4A;&5!tfn_NB7FlAN$V>qhV=o6$M-|6vmJ|A)ZLE9{8$6=rL2fPtA$?C<Z(
z^28+YqIP!L;!|whi<?b<(4y_5?W@CM2`=|ff62ae?pEM6^bnLzP|@;VrqNOXz!WP=
zg8ouC68bNxqtO=bTw0i9L!NF~@E8}y&!{z?;gyt>q!qL{;j}tVTB2k1gXfR?^5=KO
zE^_#;?(+Dt5kx^+F()o@35f#RDH%w<=qfqGS)%0qhlaG5Rb*ta{S3nu&=(2dbH&~~
z++Ssrlspwk#?%1d_0iFx$H9`p(UaO*sQ_~3Pzds`a&tR|#=@^Y$WR&w@GJ3>Xz0<}
zrNScnzYyNdtG=YuUUP9~?;V#k0R}Lnc<j)LIxhNwBRA@Kb8~HNt<G(=I#B9xFYWb*
zOm;;9tHR1ma>k@FMJ6%~1d;RoEHjXy#L$r>Mgwkj@AvKm-{TFFH|gm{P0y$a2?^_7
zu*y=LJUQ^i=mqW{NArOB;;n3r3^GhOiI2UWaQ!kY!foeAJzTir`=?HwA|xko@gt)8
zlBvQxTJ^Z7py4(9c{XlpslXOP6%*`%qZ2+5m;Xu^loWm0c_D`$`ryck;%!*=G(h2N
zcka;5btOGt8q8<Ae3`B#jJYM~IGgQU2VF<v<s#NbR-y+pKggU_n9Ivu<3&$kPX9><
zLu$9>;Jc~H)tW|fsRQqh4n;y%6n=EX^5Fn3@l((4vmUAS^7=7WN%Q;HHcAx8`u(8`
zI{mt-V$Y?xQ=U^v_M^bg(<y0Su&MH)6EJ<H5W~~IGFV}6mMG~lJT!%gcxnRoz(&xN
z<7?45(~ejl8Uy#^5d83$cgt#jld4E;M;0_<7>QKVj|)U-V$lsSNoREszJ6<UV&Wg3
zBYgl-rg3Gg(sI-q{JnOSOF=<prs<Pg9H=XQDS2JR@Fcg6-`z^5{_-*S*n0O3X1Yo@
z3wDuoe)JH(_p%hdpp!{gukXn1j(*Kb)T5ZKh#J`&qe7C??Ckhp#zMDu(SNymM*4aM
zjsSqhq}|%1L$`GPkitTK@Ve6-aRP)yMA=%kKu~CxKBR+m<ymCp2@z4zdwXkB+Jz=W
zfF-nYN(&1Mi^FQis@-Vby?ck+I=mV#0YM<gZhL!sBjB<EaM=)!uJFV(0(kVrCvC_~
z@g&yO9dTu1L*Mqegh?c>X>W$o@F0I2$)iV)1VisE4CFCr=4z|WOFn=0jDnFdI4kRX
zc}0a0oSU4yd_!{xLxJ<E{?%*O&`j!8DL+5|7uZH{0PuGjR#sLtHZ1n%JesV-N)&%g
ziE+&eii#h>c3;HC5*ip71k>_igJXbPwdsT760g^x`(JMKI=l_LxHu!i;&;W?mKVKv
zUI#PGu4T8+pMYE}u&m6v`WMg*7a$s;U&<)zAeV3+A`tOv)g~Em9HGW7p>%x%iI?pu
zXlMc-%`_X0lv?o$UzFl5GW*=DW#Hit5n=I+Iid<7={VbUma}K^Cfqkk^}T*OCjEj0
zd_l?eFB}vcUPnTINp9bDu`ipDn7HSHKxue5W$v9KeB`~;vQAPZBP5eeyAxhvK|zFx
z5^kl<g68VCZWY{bKY9KHF`B|AR3UZGs5w!0@Nm3}u07^r;PdCtORcvT`_ZMt#n8R*
z7|R<{@xEQ&1k%pPR^-Kt7YJ(~dno1|-h5i{-4nJR)t=p>F~011AOZnza(+IruC7jg
zn~Gf#QLPzJR3bYiPFSK70`yX`)d)bmY>j(;S!$};8bh^SwaR186CMNzz;n&J<Y`S=
zB4;;3eKTJsB^fWiK)*vSRZ%F!&O@)t<v8-C$V{brJAMOh7+rcEDTB~LvE$~P5_+(W
zjX254$s+Iy;BG@ul4ftu@L+EpXo38GK0)XB3om<D#{Z(}&KPYg8v?n-8qtPg4EQLm
zxb3O)%xlcd%mz!6fb|gs{F3XE;A~r48)?r#e~uOwva+(`vXoc$-g-<l$pd8(R3v}m
z!l^3%9j+}V)qxMkLgW<`9LI&<R8@&1-X$d^<F3F6B3G=ft$mB5qN4CnQiTzc6yIyl
zPfkT?>HAbvTt3`PIhq6<eeUpZx0Vg4j?WUidLeucE2DR89=p4X9~>N5jHaL)l9it2
zjGd;LljAYLXdYUSn@eI-wT>@&uqJCYR`q!E#mkpI)SQ}=GhtFf=l_aSjdFG*!KMQR
zPo7{`Zv$b8AYx*b_~tp&xn^*&a*96}oA<uI*F+4NR2oo%(b3WU%@kBrX=PP!#RM>Z
zfIF(&jMb9lFTlG+PB<9JP&lA?clhnz2ao-osGaUe2(c!nrUSG%m|IJCJ)B4p01Z}T
zm~{7zTjStU@Xa|{SgyEfT@yH!k^1bWLQ|z6=Ghuja2nt29w&Km@X+63z#$;>tFD%q
zo|$<BP)mM$q|Al@38oi(+gu8_jZ8p5;Oh13r@6V6GT)V#(?T5a(c@`peaTib3t;oJ
z<USWNIaFXAH5?oigs4i&>gH(Xkw|c%9{==P>1rxk6%`c-z5U-?dV7nMJvdGfLxiA0
z)__toZ^=gL-rml!{oZpHE*4?6>vm6cK3pd(EUZ{5uNq}tIKo)*$Ncj0@@sPmHmSL}
zin_v($-T+UoJo(-e*$->%X4$C69+-F2DlerqR+9j--Bf!-Vho3zVD$27$hf^rj#~1
z^gkq`rk`J=pr-bRFpn}})o#1j6%|R|=~X--V-hv&e|tM>hmO|}k=&ca;nJ_7_3`nc
z6|rUcW_KASEk{T7P>%-g!0P?XAMWnynalY22p2c^@l&Tlfw^lK9;UB!T9&c1<Cl?@
zeFWLJ4tLh~{|3>z?(Q$7g`uc>Yq9?rP`j`?@01i~lq~o$Ud>=vF{&Mi0WeozEL{3T
z54jD5E54H#_4Q^-pP%GzNWJ#_ZLYI}!$(LtpF##+sL2=1AWSyX9Fo&7`Nukh!8+8@
z8@W7O;;Y9qSR)*1N^gug&(!!nF(31Z8VSgP&^+GZ$Ix-0#sd}E1s)j0DP{;fNXBe<
zmMA<U*>&+ks|mlhuI^G5W(!X9m%5(g4SVOCYI7HZ2u`0ie({6L1oO3h8<#$DncTT^
z2SEr32qg8dWB%1~z|h$cbudPa0rdj|CzX|zZSCxyBqcEgkTL|McpX-}{%{t3wN%rN
z1)^XIY?Q1AYkUM=1^|EE`1mQ{N>S!vsxgqydQ`+OAONWNA1fGeltm{1(9m_;;^A;H
z5t!QEucdr&gdu=S;AWjZca8uJ1f!#)YXDaa!ACc*{8?W-U)nidNHV#BSvXEd_jJ&}
ziwO`(&%oEOU#Ch2^<_;=&Y;vZ@ICaEZkVmJXbw)UI|Ihq`hs<BeLdTb0el|`9X~Fl
zYyb3E;4f6qpAfV*Gcv+DckZ0AvkGP>I%5I<K<u7g=uR8Xj@WMN==j*3cH=lD<-LDe
z4FFw~Q4tZLg@ip%gBgtNwQOL+=g;KKUI&0%Fei~}nmc~%4n6^a7RRmgCq&$KxB|$S
zh;Q7u0TdJt0*7x4SQohh%qs;Y<rB6Wp?$)*e^$W?E|*cvDfg2oEAGlfZN8EYhv?&t
zGba3ZJpOw0$v5LfBxaPD5<zOS`YP`AO&3~>i&B#n_{HLhc}{A8`l=YOU|Uz$b!%%L
zKML2S>Z&SwQQUBm5*>d(Kip+EFE6jwSv#sfW{}atLev3uf&#;OY%p`(Rm77tPB8$m
zP5_}#{LA2CjRVFQoTHH||B_I!8*bsU3kp)Bj2c`qUmywUzka0yA{#A=0B2^qIxYcL
z3XqQ(xEwf%_-vuB72Dk{K1W(~<4x|77QUDT&ilQ~`aM9)RBYFSy%pITQN{=eC;%6%
z;2i)d5s)$RI{g80eZh1RK9^nASlr#+-@_|cZv);u0dB(&_;lm8@U#DP;AJ)w*TYUp
z6t4>U=7B>ixexcQ5g>NR(eG8SfBGb9VvkoRCy`g;1=+1VJ%#FYe>T<;YzurW>)Er{
z_G=(mBi=yQS+)}s6Wbo_Jp!Z#aP)6hD0%D-;3&;}{Se>|Pzru)q5FM7K|NrP=RmEz
z1{jO-XMke^g~weQhsOT!6#}MPxTc>XFO)ytWPhh${R}Dv_1)dak05&T0e8t|P{RnD
zAVUZNl<w%)U~TZ2gdjg8R**>`SJ1>tH9_RhD|ErjI;@T}>6Sk1zK{Uk9iNOb2;P^C
zlk*rj+mB%26=aG6C|ZMRLS0io)!@aH>$>D##hd(&8t9qy>B0VHKZ}W>0c78;J*LKH
z_h{P>5~rTOBaXffL3!*5i2bkJuk*(#4$$7klxWlhOt$W{Uiv>C1$Oa&mq*DoRnF4{
zkdFoI7JyiP`E-8AKa5JyV)cvrwzjtE{7`QKw0#fiKei4Ien5dED9i>fhydAI8YG9`
z7P0##=OO#*<QX&(ZJDh*5_;jZ%dQ_s=&;z=)@Ov;xf)xxGqyc2l^jcc6hJ=K2$^Ru
zd~1sUzH)mxS?3kz-sZZ76QXuzQ3IaJmYwg9&2d3Ix)c_8Tu`Xfg_iW6d<$tF_^X%M
zw&G|tTdlp<kf;M<{d=QEe1JXEPyt1}!D&ooXC^lQYWq@ZH4?BGuR{?Uh-Qx#cj(Xf
zd@V30KobpspM<Q%voRM9no({WMXQh=w*YC%kowp4%LuL^gG|nh7jpP-^YbX}3bm}p
zpPTc5z3_evmmGsgEUT}7O5c6`8l)`E|8V`pc-y17eSvs}cj9yUp`G*rDpp2td(kJr
zm#nv)OEH8@wVBKZ6NKY0q)qtzL<s}KI&b+lg=^ren(BOTAYhsQW5|H|C(gx-`SfL*
z8yhkZV5@&|aB+PkV0QP0T)&~GhZ1mjtfa>WrylR&)=!@rfN&<Yv9UQHB@fZU7+`g4
zSC`O(Bgllr5BFx^FnERMzgeUA^eSmR1m&C@&aM<GgBd3p9zE}z9LuEr3mhCW;De&M
z?mU4#3#Jpmhg4#y{1F>A4vymhLH5n-*YTmI?F+j|IX5>4d9M#deAH#tc<i4R5GZOE
zl04iq)~ofb?rQl%4+uge-~3VZ`7|(|tF=dx8XULNU0HEa$OmU&3@NhXirM#`^q^<Y
zWT1cuX&~vtky6s`?r!Z0yR$vLy=bAvcO+iGjBI*(nir4)B;EHQQ!KZeB{uDh_lGJA
zS|RCQ@md}h<kBrA$*=ikkrbABmI_-nyFZK9hzT{(m-;9fz@tY9vD6(nbhJn+b+}IM
z{(E~6MHlPiZmabZ6U=auvQUA6;P6CrTkU^X>JSa{`sYH&*2+b{I?McO2<Qav8*S$t
zMU$i7mcJo!KKAf<glWhT(;?1aQPQKI$%39w<*(~!H}p-RFRMmnvQ<wHd6$Um*7*#@
zxF?Y88Um{XM~J+-OO*js&E9$o^VHOoTBhWGmEwWLeXsr-_#+gyLnVd)VEfzBJg7y;
z<Y}^<KcCez9sj3F(f}asz1NW^fG*A9qlu$KVb7!A&If>|9zew=DKs-(?HCLDl;pd!
z_fv~IKFEZS$RKg#&t4D_e2a{(uYBKU@(A_XYQ4QIX`O(TpFwDw0(2J?9Q+X?GJ<A^
zkcT%+*848?A2vOw!9qYC&<CaUT8Ho9YQw|tjeN&TAp=LZ-3Wk}%lGscl4sTFw~nz~
z?xn=r<+m8!kuJEr!>r<$6Gdqi8A8E`Zj9h#I7J{e2!O)Qy<2$_miErMFjYlfo*2=1
zOmO2?a$@4>E2~b(W1u3ZU1oi^Q<|#J!e`_sK+P!2_6`$^;u7v_DRnKjvkm5Tqx9r`
zp>m&jaFMJ&?6OT=>Fo3Tw};UC(dS1i;q$fDOA5+x!DG3o2l-zU*ph9KmKe`>CGEks
zb%QFU5j$T7{<<Cgms~;LN|;&($;=;<)&Jh8dtrG@M_P@7T!GV4TUX}_t5&_t>t?FE
z66+)Dz2#a%#2l<REVAmQXGQ1upJE)?_h88&R5t)wYO{6e&9=_YCU8*b7j>jaP4rF4
zHB^fqJ2UE8z7iJrocov4BmdFdjXq<D^*f*L?G^Jp8O?P+h}3DNc(ig-{`U1+6BJp?
z$pm{6m~MP)+k-1<3t6da-ljU!->N`@M3G5JcYTYufnS{&$kP>Z-{Xz^^ozU;XrgXi
z{r)7n><i}wZs;5~-f7C^)>xS6CpGm_UQwmC^ry$Eteh#^EJIpVwR0Qt_}v`dbC^#C
zq2HO6&HpQ<X#R6b81NowHVK8b^W1`mT<ZT?HA0hEsLz4}N@beA^{6ULKE0==)^_TN
zbW5LJc0acBu~Y0`#6|qYHg|FDD~w|F#msceb)?+g^X2aQ1V6!#b9;ii*fVtt6S!iI
zi?~3xfcoGb=u_bCKrIFyem~vnK0h5K)s|y7(6k(ZGwTJ*T!jVumfY6R@Ph|%jP&>2
zH!~Zgm9;<V%*oKiactR5I6*Hqq2D?rX5iE56UR7BsH4m5g;jjSmGe=~u<tZLpPTaX
zJ&)na10oI?ZUL8WWMdPyx96f)Zi^-1y7eU+Xl^mY5I7&fQbG4U+t|1`R(^g;0KI6P
zXn^{k>sANd#<_Cr5z;&{X^^ux5%iNpz4f+yM4qXz^T-CTz2Op4V(_M?^BpeU{pj;N
zx1|Z`y7Vdu_CGb6p?i=OC-95S#G>Sbrj+Ar)$_cV@~W;)b-IB0Kcx0@X2;wANJ?Jy
zyh1BOipk=1XV0E>+3XZ{qzrq!Imap?!E~{DD_r@WY(Paq&K2`JKUaqo*}qi#+9!?h
z3*)%riytt(G5y9?7WupprFcs&g)=Fl=$cW|o%Wn^;uWCjp~&O+?VAbEDvbd08H65w
z`6>w&aR4TAN=kWs13&`%0Cfpb7Y8VYg|SNEd-v|y&VHgs1y8_RSmZWZ2obk;yu7`&
z*3G9iLaBXW$+_Zj<?vKb$(e64QRjHTrA`?(W=EkGr$d=GhPl1^Zp(sbx(_f0)s6wB
z`5<?Wv!J;6Du}btXkck+iN<!|wg^c{n}BbeTw1cU)rDXy;<$JM4KENq{rvq|A>~9_
zNwku>urOrU(ycdk(Y5?tsJi|$p@D6$H_Gf0_5GqW{E<>(8I*RT7vE<YUBsXr<zPDk
z6rQ6rz4r#+#VbD*@1n&9lqQBWks&QX9u61HVSubCZ?>WV!W!7H+jmwM7J|??YvB3w
ze)YW`MlXdOmfEZH>pW)d?yV-xH!q#A6HI(4|MGCYxa0fJlfx~2m{b3m<BVP8`P=p3
z*Fro9(Jnt{V~*3(1_AGZa`(W|e>6DyrFG`qxyJGFaoY88Ts?(PxT-bH3Y-jh54q(L
z+EwMJMM@V6pGBU(MU>@7BCKBcC0_U*`fMFPS1Fktz%7j(9cWfn2gO~qY7hR18Lcyb
zT<&E;!gB}!^lX~p-s({#H=XT@SD7tWgXjlxas}-LUZrVSF7Gehj5~j8y^98oAPTCh
zxN{q`?F)nX%uvNp&t8G&2U1`IR(oe7eCGJ~pL425aSGb8*;C5{Z>4_`tvUJ8ePU+f
zyl`vAiJh4Z&1Nl#08xs&tRbGBo~}!mX30;9XTk4cag>n6`ID#2+7({#AMI!@GdO&s
z(b_#+sJefsnL+H|^cMg7b97ztp}>B~5JFb_J0u}W*kEv5Urwh`A@|kES)%80iljGk
z0`7Jcg%l#Dg&QLh_bpgNg({`@F?n>az%QBTK7JzB7k3|KdrauuA$*KuaMw=2M`$2@
zc(*0rL^}OmA;vu<(#>+t0Rv55q4#e027AQqJ2uulUmY)fKmPrW|9Lg30}s*j#5rGw
zU52$|%+C?Ee&UqX=oq{HJqCS4#m4i%p`<EHxfkdD466NYJI6ODG@#nbFz=l6;+RWG
zNpH=>Z(aAaEeykyp$55qp61|E;G2Anv$mW1?{=yU!~(A`*GB9%UWm$PH=PbkxDzrg
zg$b#rLVaGlDzvM~ZqJXl*~MKEqe*~IhE!AQ+>uCJe&}o>leU9lIzj@8Ncx{sU}t|N
zUwu<j%HcbjyMB?=IWRDv?em0;tEp#N%M~<sl7_+O+LwdxeWL-d*+N|OMDT5c+l#a+
zn<)AryU8l^X%;xNRQx8~fW`FbVNT5rO6~aPZ-uR|@Y78`+<(7%tKjl*%P_j`te+rC
zb=1sIpSgF?Siev9=`_`KoZ?cBe&y`M2!jCo!PU4+r_gK0FUiiUVMb&9Che6r9>OaS
zQLn7_u*Ze#n|I*zyq(_{)!NvDF=b$SY%1EfMM0Zo#rb1N4_RX34qtLhkrOO0^=hq7
zdn(UxW!*&A5CYQI{+nty=~DCGSWLK8mhu~2BCinDY3xZoDA7S`ljZrWFpD+OP&CX^
z2EQBFAjQlak$|<*6;ipQaNC5RPLJu^@NQsg#N}9&t<It|f~9PX{2Un`w0hWS$YUcK
z6P{B|*Kx?$k&=8nXtTgPgA`qJP;ka~_oXv!trxkP;w;I{_UsC5bYfoi$ON=hGn36p
zo#?DTKwq)*TJNv9o8h(_b{z8UeV}emXUwxQVGqqJtid819?Z-iZ}L5$?8U>t@av7Z
z0T$dF4UH`=ANGgq6$iE`V?$2~J+R8eMIy6tJ#%=C=}~LANdwyxRrQ2Y?||*@eodf3
zZ>+};H9G;i<dxZM53g^R%>LXsve-J+yu=T#Dq=Y?zIl|A?=<0yE$UUo6%D7YXIU<9
z`w*pJig?CwfF+c-jh16-?&=wt4s4xb?p<@T!t+ugSl(wdvySJ%NPPK;*?CRgr-t(<
zRv+5RPkY`SRiq6u3fS^=p<zGxp_HXzL3JIokZrIU5+_L^rnBj@&$V<ebaq$!Mm%ME
z`}LO3Dc+cviL;-&xSM}6d>8zIQ5h@ILD_n!t=+jeKgFero2z5C|KZrB;zG5_r!99&
zR>|#o(M$014WAFIZ)A3#h~xIubky%kq2Dbnmy#bhDC~>=`gHdjwXG87Ewk@o%i>p1
zruZB*s3YF~6?8gn%&2;Vh=_<H)jW*zJKZF<n=YRB#?>EX#7jeHFnn1I;z8fUssr1d
z0<x8Xp(;*FROT@=X>MXRAIn=9AJ^W=TRz`de{e&^UWgyHW%OGF)pPv2w%YExKb)Fg
zB!X(Kh^WGhS|Ka-fZtQ4wuVl(iofi65yam83(cKeCX(mzk=QoxuJKYh3RjB~Vt6aa
z`ma5F$fdzS?LN9ZwPc+?Hb+RXiSkA_DdC{iH2hmyME6DOn|nt6y8F2PrPPdES)-yJ
z^sB&&(uw7`ioV=iHn?cqthmY=a)xjC_q7?!oIw##*R*Za!+-t@j$P4)pb<awiK)|r
ze(3p!Gs4I7A}cH8I$H;8RtK-q69r3>ImvH%=EAjnDc~Lq#fcy}3oxe=Wlh?u)z$CF
z{D3BVaK<4A*<9F%#Aj7fnE~wRH=Bbc^}iMI*g&3&OKK*KhgCFmeQ;Nb4d)m1oR7yy
zHNzWl1Kv7ZEJBFNnBT3Ro`pl?>geRnG80>N>$;9ImS?!(y2?|p*0_`pH5BC~5vB1Z
zn<cOH)YK#=`s<qOQK3#kjsspNii1c0Le$rmQ|UHJv=dG{4^{`{D?7Udgq3J9654(c
z*Y|LWmC<TZty)ef2>NuFu?GYF_=4EkC^WgEP0$Fv%V$!}^X25&shI3(ra!N4DU)fg
z?b`C1i*LWaLv`I=7#|V;(as_qA)S7*Jmr?(bFVjPX6I<AZpfn-TyTnswx@lp4Iihn
zqEVgoMlH`f0I?Zl+2b7JAwMU^X9(Ady9byqH~sMk>UCngZ?v?$&N*y5%RI^^zSeX<
zgRJGjC1+!#uBG=BU)y8k(_m#ACzpR-l>Y^cQaqJ@jf~{Bf;nqobzV=O(WM|nMYp){
zf&o=Vv0{kaJ0kSP^?gI{2B?&m74=a&m$6!WvDm4{(>R%YYIyh0hGl&>hKunOl!m`j
zPQO3Lk0nNyeBH@Oxl`=sm8P$)_acgHH&6XB2#nICUQbVAYUMl54Ks7)?A$OLF}J;P
z`@gOjoHF-cpA9Gb53>If=HR5LDF3nvPX+)osvJ>kJ3Gxl%)EdM#(V@-uZcPk3Yc|W
zjp4a#Xk|sp$;m0cGi1ECKXDWa{3;tGlVEhcganH)X9~OMru1>1@yS)F4nwvQ{t){X
zWC4?thG^{yl!~+hrX;8^<<ZP_5FuL7a9sid>Ob-zx6JmT&@8_Hqe#2biF;#nlkeU~
zLJ}s?6LfTRD6IjN^Z*dH7&G$=Qn=}!O?Iiu+mZNlF$M32D;Y-f+<gXc80Eg9c2)<}
z76lVi2#B~&0~m)&ig0<}W&1hMbEEYkS-9aS`Gt}MAh-gtCq4*3a*C?s;^^3w;{8B6
zjg~h+gE&;}CJZFX8svV{KqJQrS;ZHS$I+`cwX`bh-@K+T{q&F5PJJuK?{ne8h5C3w
zO9(83@BH)f@=C3NWkK}@a2IWNete}Lkylb`0s+}8)$?d2I5t5_Z^Ry03TT;V1_j*9
z=;#b?auCJByP<-Xq|zVD%LTyRA$1@#dzqN1mp}8Sxp%*y^nIB}IlJP=3f;eI>02*N
zu<d~!|Mu-$9h66)fETrRPB9iihM?*lV#OrxLSa!tFD)(oV5BrpIY~YHW0qRlEA?!Y
z(eS-z588vGA^~I)<d*o$vKeMP!eFb(!@C4&;jFlSZL;9MQ-YRoBfywl*VH_T)b;oG
zzw*d3cPULN+i_H1#5s5*Fyc^;@#y;SyXxwn2=2JKxGaF)kzs={hHnJ>6SFzkm@G6=
zOwG&~f069F<WgE9Yc?SS->kwO_b1Nvv{%h3#V7v@ozR){mZ2yQEkXblhIoVOvbcF!
zP}F%Pw@e%~LI7;>%Vf8Ot5SsTQLf-1o|U!?UJSi>1JNhhL!Y6J<`UZSaRKC%r+^*_
zOieuvTmb=8M0g$LGzf4$R5T8}6*s!TqxJ|B@YJZ-Kzbj4rv(ZTTLbg+^Ee2IJn7T(
z(DEqbWy+g3Z^nu|I<5Im?~ZXTb<@X>3*7mOT_l2x{{D>hDlYeSZDyAUnZ)CQr+$yf
zvPZ;;U;-ER8CQbl_8j-lP~EbvBS*bkE6dAyW#h3{)YQ}{9f}tF(duAlf^cJpVt8iM
z;iW1Ax8)^2vCE7dD8%gZliNse`j@1<^tPsE04M}cfeANyW8u8FpI1JA;#oYocYN^;
zf=y<TmnI)Dy(w~mXQ*O*a0nWpy12f+UJgpK^j}a0#X+uz(BC%Jg=U99xD=<^*c?yh
z;CfF_-$Pk4y9!3Mkd{p)V>n!Fj@}X1tpy*b<(m^6L~N1f^QPN<{Hfd8#wlcU3H8fw
zSle)~gSHuz(};K3xT}JgG9IFtr{fPKn9-M4(hL1LF{SRE@q(xb3`D_>!*=%ejX;N^
zbfWNtI~54dhd=yRW=*N5r{}Pf00gIy&3Lf4H-dUDGBPq!+czd0IF)6-zDbmbs!m4p
z>j#Y)-DoUK4LI8-eSrLL&ar#G9-=Z#RH2TnZEbyox>JsD^X|ACIi%#*MekWCo}4p<
z3n5(k+Z6|Se{_|ynRF*mkEbmWgKxjSc!+KMS1$*!uxO2}z;ciWbWorrJ)scIm2GDv
z?~$08XuH^VflH^D5UsI;GIY4q>iO3qvl);LzIAXu%gL#*{|m~td4*%3C}roR6iW4S
zdnmO*pkd(?{6pqy%V?9g^-2aAnz?0Ni+G5QgZ#%@dUkb@dJ;SdsLp}Y>fxzEJDRj#
zI+<S5SmPi&htL6nM*P-~d*UQ&f`=P#!gjW2_ejp9OY;6X?2otsS@dBg&VCT2miMb`
z&&Bf6qi6&$ng&|wRHI&LV3e?@x>IqD9fdMMS@$6pa);~W0dt6b&)E1l1H-W>`P;Ll
zDtNB19Q`ju?H(o%X&a&RX*V0-jOURrqo1WT8lz!z$?Mq_FoviR4z6#s)ZcPEYFAxO
z6F+Z%JBq1p)00urS`-Y6R5p(F67alB7PJ<-=S`j)IUBJvOA~Epq=3E{6{Hy^h%V<b
zlGO?QeNU~c;-54udht-h?UUj2d42Npmv!DWUw0#(7%YUp@L0(bhO|+ef#2g!em!H9
z$e7EUl^MA6J2cJfO5Vv3YIb<l<37*|h%~k|1~dG~_egRVra8wmVZQJMvtn#egq2Lr
z4if_hXgP7T8>5}&LW6E=RyTm(x9M$-qSBg@C-8+=XQ!P*{W0w3C1^ynXd5Y^e&SjT
z$*bLhs7B&TbOW>0(BJW;32=JyHO1M;I&+?F`hdF^|IY%@pW%EibrKZ_w~}zcNxxrp
z3<FZWfB?6@`&grB{g2NhyxF^(k0!6CKE<Ty7KiY&+lC9N(a7YA$#Wx*>XOwm#%I^8
zrEePhp~+s~JwP~b>%Ls$7R(T5Kr9vPkt_N}7fKnQSq1ZA=p-&!U>L81ngU|`{?@&o
z7oo$u`(fv%HFd0I18$&CNJXEJ*{$+3H7`vNxsrZz=>56okLbP~7c_vCF19u3g_?VP
z&-L*7wJ4ET;q=+>zkV2Octu#)VaF=Xqt?3R+KH1faTrl<>fBQ(e;tPscokm_L&xlF
zFGr>s^}S6W%ReqZ>nD2BXU5;<@v`nsc^^vDo?{WZ-22yqrQloNswM~;U1FE$#X7}P
ziXg1s`G?L}>Gu}{$fD4VZUU&K;XnM22q%(!A?0#G;Eu<}nBK$RQI;5spgTrIn`+T#
zsze*oRb#~qw{UqsC3Li|Nvel|MHDP8H5v#3HJn(Xw*sR{50OYI`O?K&vzHNC52v%>
z`9W`h2T;R#uO2@$7UO1VcDCJJHMoDIOP0`Rx)t;D5cKEkyS7&r*pBH1Fg+JGLJ#?B
zPYc|+Xbp}iC)2`s({l_=-=2?$mGsvfc^Z%?8&r?bVlI!M0%}1Q00P0S;eFr)wd8=y
z7jYZXOQ?}6+}_j)pHE-jmNxcB8IitRjNO2G+1Kn`QQ7T#^;<RX)jwe&{!RC_t61s3
zRxtLV@kc`!lN~Te$;-)kQ*b*=H$Qu>Gcc>SS3WYcddpQ-Q!<0uJPuF0N|G4wS~@8O
zHyoncCS;^<HP+R#nFv(a#Kc<qSmP9%u5mB9^89jey0&0XPOyFSNPyva&2t%}M{#_z
zs4B){08Mx<g~?wQt&9^r5z)(KaVN!|&*?pLGQ&cehZpw+Dun@1p!wZIYo9M!QR|j%
zy3uYC|B~c9bKjAUoEFK_QXY3DV~Wq&rL><SsNCQs)<1kq5h83r8N0&`+tolSMDJO9
zotvds%kRX701}i@lG{K(8!^{i_)H+YzgROkE{>!qh5RE5>I*-Lk*}?2y_BnHzYvYu
z|653_vao0iHG@YL1CE3%&Vk<_j7|q*9M^g&^7yB+aR9o@nC4>oAXdw!*tj-cD-E-%
zlRdY^n(RE)=p@G-=qHvx1osq1i#ZCS?<xtW7w-*v{aC4OBJB&ECHE`YsMf}Z<wmPw
zEE_QP5q}b{ozm5#cPo1bN2D@tWtzi<3>Ut}R&UUj9@4j0^1*jr1sm?0v=HdGgF=A&
z&dx2-UapwzzP<HH$Aj9>VN=r~Tg`-M?^0G9fBs{t({IE6jU*+LqfASumCsV0eCu%)
zcg^@?qFwuE2-4<kUobgw?u$SjUhiSgoq1(a^tctEJGxcwlCuuN0+n?carbWu0{t6l
zV6I&x_YV&bq3HziuDLt<rRg)##A@v`oO9vDBZ-yC0YLDefrl!21@Te|47iLT!PtY~
zj=$8X>b3j5VP3qwc0(+GakpEs?8o$C_U=k>NEDtv`S&MeJlJljJuh6!++QoOde6|T
zPVO~{EsWO_rKN1h=6jdENNA(qpg@mGF%<O~<}1Z-#;WhHNydMV<{YobIyD}GCTID_
zGbg_e;G|dFIbZPLdGQa*Wha`WR=CjolO{G#Uwc&ZeIvS9>{u+c_y|^dT#y>hh`u}R
zrhL<l&bJJlX*ccsRMUzAue`n^##n93EZqCN*y{eb6GrW?wI5zS@myIyZ`~x!2^LAp
zlNZ+kSXpfoHY@A+ZAEqOPhs)51vN9K58bBiCMVSJbEZv#p<(GG7pLCMPyRaMqdY=q
zj}=_Fqg}gCJX~2KlaJ7B5kGhyI-d8s_=WH!rt?=2^}`{#_rlft(@X;t2|<T&vP-WD
z|0pzRYLJ78NmL^A`BVS$q~y$AO@QnW#Op%WL}J_TuUXD}x41K|m#Z*{ugnm->E6d-
zx&SMDxqj1?DYl_O#(k}xU}30$4Mc!bD=VR(IiG=Aeu*^z`eQb=zlESF@U;fV)ber&
zns$L?^pgj0DyX*4VR@Ju@CTanytx@+3ko4uM(XYqXNv|*ME{crhmJ2_Y74G(6T{to
z;G2JRJh)<JBrW9rsoWD$WfyM|eatX1f;bJAg!rO|3q}4#v0#-$dwN~3kS1WRY|pWg
zFFq@?htmy;r)V4Z`Kw3|vTC$d9eNE}E?gjjD)>j}emzc2jR&G_-{DWKtyiF&ew>De
z0R9f`hAbdw2OWJgz+>BuZ#My!E&c%QEZ?K)t5Dyludl}foh>5>#NT`DJN+0bBZunp
z0;HO<jG`qJC2<t5rkN5_6U7!r$FY`y&Ei<ZCHCj;dp01~&gETG^-cM@<ci6+8L-P?
z{&X2K6&fHCPS_R~d8Z}eeUKbDwW93wV8*cNYZt7d)#OaVKX$-_{t6cw2{DS(wJxz$
zLy#DVIQ@u~w(5i=288ak!q$wCMdI6UZuIg)7m@pRzgDc6Gmntf@SCE~kWxUO1?c2N
zS)i5%-As50T1NuWHQE;sM<QsYdzCc%?S({YJL-p#_ZLPxsmg9qe%UN=YNaVHkB@z7
z`E>W?&sW+uZcS%7K92QNFKyONtY9PGO!-AFYu6V<oz*E*Z|Lr#%+tNoP!NUrPnQms
zrNZT&)$u3k4J4-gab3TJP<OP=+GaBd<umOfQ&PkbtYP>Y9)h-Gfb`)MSRAV8a|L~-
zl`iyD4;9`&c7hlW>W-Nawz<$D2`aB9$eat@f7+Fomp@p7qyi5?dlH2}dJeiaR1t>e
zC!U8t&7jA5{Cta|MnwF|5aH|VH&>V(Hn#n}7#M;V?}sE}3tDH;yrvbBt4Meb>|1$4
z9JyL2l?g{lU3CZcXge40c^-i{`n=`osRuSVVwQLh^SBI+?_<HqrTf!u*cH9z754RK
zRE=#n%)S)Z(?H@WIsU#t{Yv(WMs*`HNCM~V=u8SXDy90lu`M2kd__w+w^jaqyC&Mb
zju!u8>lpu@xQjj|L_nqjT4rLX1=c~sCNz8#BXZDr0PSCf_wW19C%Y-z99256Hp9pO
z2$m7+3wP=5(*+O2oDS_j%}MVqwbmfssnh&#M%%A#`7;E42*;0Wz&AIKJFnUH>RscF
z0G-&WLtph#uNQaQC2nm+$G3MH8u9p&%zAaaNN8u(xNlpm;+GdPo3cII7V2NYI5)?y
z-8#!Ffb-4N6zT_2N3E(;B_wv8mI8F&_~Vqve(%$XIP#G><>|0p6h4OIW(vKX)K?yh
z&0dZoF*Th%@>iYFAC_ok@m0hJS)YtGJAsLJoT)C|tCQHys~Iu@8$!bsmC$h|%P#w~
z#`H@mDDF>P@OqOwEi}Aw;}b^}0a7F-9Byu$|GHCZrmu=e>eSc9nIn7OXCtt;bP15y
zKBu{ejkxao$S%{r<p~OpwBizAQ<6$A?K^z$!2^xt_F_sF-|Z>|<JZO6c37RE`gyFx
zOtu%fsaBF#%56nYos?&ASUjMG()ZN(Fp1Dkm)wWAfP^@UR%)Tv$1li_9fP)JtjBoi
z53<Y=NE#xuFRgw%&xwcFk9Kn47j8*Wnla|t=BwhpqINp)B(JO1`Qes1r-UE@7ri;C
ze}$i;L*l+RUM28iSFkA2(bY_Djj~Bo&v|vzzoWC$$TZ{Tt5c<!5hCUh`GHnTRqrTL
z5LI<O>^Kq%uW{GNHWLb&2fIEy7M^=LMIm#7s)si&#$`ArDdRqS{+v@7uTI_kr`g2$
ziMh|B3h9*b&r+xFS=mQN%P&9DcyJ1|9>e_5l=%D6t52tw*H2JW|NfeBI1|$bor;`E
zhl2aCIg%uD84PJr>C`5!&0TF<*V$B63%JF_TQmnWUr6I)`JC<)LZA>^QoA3RM+ge#
zNM@feAAim73=D)s8y;g-)m~c*Ey-AYHha4%(_SA7#M3cHQ?F0sc(w#ETF#o@k6QGn
zfN>^+r?i?%tsYu&G-X^4s1on|5i65v#h|MjD`9tc;egElivD3^oS>DJ6(_ToKdm1{
zYkOyZG9sp1xglCTpS7_rQ7D~WCPVmbZq8!=?T|B+zB!W2IxoAev1sX<9<@3J=mq>5
z&g<NX-xGW^|Lxm@V?e4+4K$yeN#4o7{DS=|^cu*^|1>a_`1=SAhBw^CJLCvdWe0RU
zfB$mS*4EazGw}tK`!Av*$nEWqLAR_7yMV9E>r2Cy2x-qd4SnJC8LgbY_o~`gZfgc;
zIt_sA8ExbAG|C>Is5RXW55E_izzMXUnx-Z(1Lv_MkIrwiBhZW8+Sxu5)7II}TuV6{
zW*X(y&F8C=ceq(XBE$XZC~n4J4Im1AWRFa!e`rYUf}^{8)i?uIzzY-Xw6<?~dD|%u
z(e98hrftpE?XkbLph(*Plg5$+`U-Fnm^6a*K2Oy+cd+?kj(K%x*u|rzD&3O#cjg~;
zE@#*g`(u*_#`ITJ8|~Ax17YSpH=H&+S&<`_T=`+9B8NG-7X6P{XT>IecSX_r_g_5C
z-f=_Ua;K7_VsN(X7X|zI%<SgD$o0{F^3Mii@E@knchUd+^|z$R&q-%!3i~=lHC11E
zojlXI8HeeXj^_I_0OWKdbPca>Y+xg!AJV<*axz9oMk=L;`Az?fKQR%3?*dT>0Rk<u
z$~M&*Rp9DLd!kKK{+c7A|3B?>ep+07De1#MGg#64dwY`wEvICSj93A_XyyN{aUBRt
zh%=^Li9`t6w^MAajG%oI<qqE|@k#056a~!!(Mv%=;SF-frQ%sY6~NG@!2lU(ZD)tE
z188qir>v{1sJ`p>vk;jUeh53S=q(FTQ>B1e{o~9L=Pzg!yoww5CHEs6AKXh?aTi{g
z)&WyyFhh0DonwPJH=v{kx!;rg{CuZ<wDSk;7zOb_fq5@`5~fEj-$3dkX66g|KX-Sq
zz`O(KQ4`$<V^fUf#R4S}2-XDA_DAS}hA*xIRU`^rpjE~=kr?GeQcuc;I~yVoBK-bc
zYsclJZ|Us+N$tUL_99|vY8sH@`HKk7%CE9gY-eTcIrN7aLr;~6{TxXI4d!d`S^jHY
zh};g@p9xGQP!C0MThM<qZz%)s!oO2?ids+9++9LD2iIVxP~*r51IYKF+7Sy}VPR2G
z2y;0R^@$UPKsme2Mronly69x2=4N~_wU+6o=PzSov~qTa%WTNw6g6i*Z!ZlpEXXOE
ztxeYVpW1{O1(yB#3KcsbCe>Vi+<O4}lc85T@iQgPJ3sEAYDE-wiHnODx2Q)bB?!G%
zNs)YCQsSSM#&$iF@hNnL_FcdS9r<UeBSw&>`$t7lCnhEFnfIL0t8~g!UKY$4PLu-u
zRcospI-v%%h-h~ROxS=1E?(ni5~xWGd^q~@<%>mr8`xP!P2iBBh1)W>PI1<-vBzrl
zt_72?Mt=2hvY4zS0gP87T`~AHc-NyQ;!@dXiOuUvPZi%nhZX==)RYnu7|@ADZ{D1F
zI9h((ZFeQFc+Jf5ziD?=Umd=zzBO@lNCT2l4g|eAji7NyC(EGC_ROAptfddXf}o5L
zY`*X97>Iusxy#DR@|`=B`a#!gWNE3JT?K2$FY69}C!wV^$dOM4I|BSR4g%A3Ec&hO
zKkT4lTC^WVb04MT0fGm>PN6h7kn{(jJQZ<RpeVJO2!&2~-x2kkew27`XlR(6ob(0|
zL<SY3ArOtheGY^vIWX2EGx|oqi-r9mv`DP=-BQ6iUn>N=!y4262zb>4C~q};9qk+Z
zn5bn2owG^aj0?2-Pc1I$WLH5)eu;HXLEr2!jyKAZPfXORKgJA#3XMcJy7H}L5n>1u
zwbs_wtnGOgmd>Sj?<Q`>&(0^mIynU`d6@{yuie#)KxdcC!euyte0(MCOJZ#a@+NF=
zu$IL0+}!=qpW_7;5DNG*!F%5_VEhLA`mkmGQ&{+7<V1!24Je@zBlkdph+j6p497iu
z<OIF0l*m<3d{vKF+#J+*+*zHth!8@1+H1`m(&5a@`fe%i8|dtW*MNS@Kru!wpI$g@
ze6T%efQNKL6E><Ew)zP+=_m7=0;u!ToSf~Ghqe1FVpvx|K_(4D$h34pqU?lqWp96f
zNiS1L$*RjhPcM;H)FBH5?m=CGS{fQ>pxp}(nVOwF3&D|O#S5BAvAv-sJAd`{r4PgU
zZeS3v5f&lq(Fc)->F&$@JEONzy1LS)=_tkW^z70*9@wm<dHu~_gTY%IJUl0%!wK6P
zsvdgjzd#XAgrL2`Ft}vCi^<5;l>G#;f0sY!)t^!(;16YZVxEK9(s3VVEJAZMwl`EX
zQ_IVTUU|gN#I(lKa{5e_?0$KEvF1Dy1eE;eNe`Wd60KgE*U%VCh)jbNHD5y*QX{M@
z=x7K0-Q6MMFe2nM7gw(O7}zu(QUI&>9T6z`;Q5jZhF~2sfL@=^cQ;~#rtOBH6LxRn
zyXp`6&ggs&IOOdmt(@QSy^dOHw=qjP)lyfCfW+&DvN9Q5UzjGQo$dLzp<jJN1K>;G
zsE&ik57IO+!#~iTBZ_!`cU4$<O-UP5(}6Z(v^n5un)#v2a#TBjB}vcqjrNNPBJ|(S
zgZ~3L9G3V&M^2VdiZ23My&$MO*2DZLv;nkwB!VBZakQ@uv4st!y3u>EKFpc<I5;}R
zcXrF;oOxG@8zW9H_P=kBrR|IEJA38~g222n3yGC`P^N@AU4iP^S3&)WYWAVuM)+#1
zQZp<a9VY<7%-5jGM+NRNF){V~o5P%CP<4XN05%wv0drMc@^)>W4xU0lj1I>{3nvI8
z1F|#>5<W{$PHhjRE6@PW>Z_Eva4V-@PFWcrsWv9ID1=9#sVQvCUlS3qv9W`hOe1W!
z7SvFUC<IJ&@&GjKTSy%G51x7v9{vRUCp>EM=X^^3Er!nJ552odZ!UYS@cEcyq0N9-
zR-1GS^ft*=jZ4lt!X?8;NHEhz3N`n;5`<Oee;@689g!P&Y*V9S<%;*><6-E{V$~5^
z2ZxA=wq)Wjf40om^Eg@@qRf|gHMi6P5d<y~v=qod4;T~%7tuIz3d<10@p1|ZzEEAo
zC!tkc`d#~Lkuz!cXs<<Tt(n0F?Ze|0w$V4Z<+(RSkhFV@hUX)ANi?cgM<gyZT9Tup
zbyh!gd0#S=5=2Y7N8gEgfPrn!pJHaU#JbeGcn5kC^U4Yeun<7NLKeE}rBKt&HK;uW
z2p(GLpVse(oRXjyb6&X$-I-_?6-Y$U2@GIbdm|HvnwfX3;y<9nd|+x34kS8is;b^l
zXva*4fZk%XCv-W`TFGS@3-WKc_=M0F%C3^Em*b&3U81F?G)P|ex#0uaN(I$D90Xc4
zALSj!Qh|UHo$>><_$$)VSm1flu5u9(5p=$h%kG#n!}-PL`PJ1hXu2hVVGym*&<{ai
z9XgLw(XK7fZ08-a)LZ6zdbOH|4&FsC)%Lu;Z&)E~gd1=fc*y!^uftn5zo8TGb&^^t
zJ4~UVMm|D2cxIWtz(^B}ZdszXzXjbcg54h+E#Z!@H_==-za>Ch0B-iWxeBr8&siF*
zEAU1^5O?Uy=1X>M2hh$+hP=)>hvnPCw0kh%5BxPql5<9D;hnI(%fTUpc>IRB8I;IJ
zAWF6dVxpso5ePjE?<W(yhsiR!+S{oBf1O0G%E<+FPea4SGspxQphgTskZAM90yGAF
z7GPM?ajW6t<H$$AjNq&*?z)voIGcS=?mrw+)6g&}7J1_`Mx^%JIehM*qD73C+k39m
zL0Z~p{MW&OJCe0N@Va0mJQc-j&<cWfvA{H`bwE1|>0?^@kMUrHAu3{0SBI{Fhl^;~
z1eObEb_Qq=5wxD@0fRNd&nu0z%+FXNm^3r8ij0oF2cw2iWR~o{aq|GCH3;gA08DKL
zHwXz!0MvTn-7{cfAStwb;~>5?+g9HPljrPh;nzw$-77`at5z(2Dfc8w?Z$%Ji}&gI
zhp~WtrEk~;1gKy-o-9}(SaATHdB+2*gt)kR&?;Q{X)MEi4Ft~!ty2g0PYAT=SPduw
z19HmiPc;@azc8q;G^pa~DG2}$Jxt%jb*r|)G@ZO=K)~o62H<#bP#HY*$E)pUb(;D1
z_3vowZj@w#zyYQcAax2Ct4+3c(jCU=HvaAjF5Oy!CH^h52pNu!{5Csl^fBmox$8Cu
zst8!E*(c+V>&WsjxqlzaWsk&MK{dhHI(slVz;taDBqMJ<k}G&1qznRx1ZD>9;|9nP
zwP@+~`@y@Gzp|9^<(q9VRN+00O+X!S?ZL(|MwmG<enX+0UC_cU7ZxlAcMAGCP^uKv
z{%K_wqn5h9te)xoWuITqtj-pZQ~qKWvJB4tK}Xj18#E*j-#sTfNlpC})HYaGP{9oT
zMV&hgTJQ9ZLqP|UY+<>5B{dzLcxZ$K6N5`-Ia+g8-;Ev^hDwUr{Mti{zS4x=k`{li
zrIr^3ghvPSQh!1?%t&H56ABm$=3-qyf&kLse{URG6vYQ+>~nwu#K?3YvnQ`hE!>a(
zxzXx_jrRV~?|J0edd?-km3#Cshffjj<p8dZ1txSA&&EsF15kk~c}{)}Ws<04Pi4<8
z+#6hk&-1_q-tjH>;6WI#<6>WcXz+C{O-=UZP$o_!2)5y6N{YdZ<n5$N5EVQH`_mXc
z1av-3?gw>m2JW4VNpO=8suf{))b~i}lYa*23@{Q{f~aHF4hNLFDUhpBF87m9e`{A6
z?YZ2)bxls78*Nnj2&3PMXF+98A)l^}-v8WOt|Sl0xX^KiSJ4)~of?2a%ed#Mspg@H
zlMn&eXD}~nwP+#g1?&PP;+K<?)Ach)`S$!O3@VwHI{K-SE<FS-38&ATxr$IgDnSEH
z1Um%xu=|rLA?qwK?WD&HU)%skunqv$^7Hq!qnhGK%g<kM*`=fmhK7e>b0LuSTaW1F
zSAK9mXZ|k5P0>&%2$=R|gU8Ue{SmnG;?<<2q-($dbnO7JGKw-F?CtIKhtX&V3qXvI
zfM_m$aQL2D{c3fhc0%q(|60x7^f9a}2nEDtYH)k~`w_{>@5VCU+GnBzJ5R#wvy&IA
z)}8=q9R@S!XpoQ4eMA}`q96~5E$es7<8Q4mr<-UDPQX}&6EN5T+j{|AO}ZF6>~L?#
znZ1?DRn5w~3Y9yM-*NuhS<d<WvhDxs?cCpL&euNvt&ud@T8RpqgexXPRK`dS*Ak{w
zl$wlF38TZrv?<l-FqEkrdPw6i6Qia=DRyVIs%a#LEuuswiD^Q@>a64Q`q<Zgp1<IE
z{MO}awbu9h{oMEae&6@~T3b}Z>>m+u{5k`BTXNN;O+G#|s56+W(Kp1@ieL%9SFa3R
z-je^_F;-PVPEw9$pP032FLm!Fru9m1-n4<fGUx7*am#%w>Ur|BJJ_hrXw7_AS-jrG
zZ@|0}9)zxl_a;T%3F5Z^lj=y|wSFM(`J}CO@XnsMn^knxFha;^F&Gp;@E0i|J2Pe;
zH@rS{P~m8ByQ4Ps)?lvO@@I$UsHvyVU5@!HsIIzkew6HX+s+sbTTRDRt0uDXFpX;~
zL^2E-#xk7VKA&G595qTFiBLQ6Rp!?kLUga17M_xl0v^mi>C=s`%$v?8p8-Q?(`y;@
z#{TEe4bc%ZulTnZGb}S%T!h5LMDf1G1vIWp@)*GqvhkCE=YOkYTC3spYbG#nM<^5u
zCj8GZ`<C>?<aS}_E=GbpQ?3`X_K^KJC(C<f>_nHYOa|-}y{L#49DGey=4P`Z<I5nJ
zBG{=*D__<as=a|1;I?I6%E}rgvJNJ{jD)+)^4@-jpW0I9X=N{{caAS6ogAQ36N%rY
z;o{<gO<?$qkIzyVL1i5JX{Kgo<C>;aIrvo6u|alK>;ca9xI3{jeZzd(JQ09O#NP;L
zir;;eyVKrTxjsX2Wgs*)aFiGCZjSg_Jl2&S+s}MwY8Woz(9%pA@~!TNF{uAr+%kPz
z?J-AejrM_d@YDDEEa06RBLcuyULIp#z39gH8dwAqV{gdDj`e!AYHb#4-XLxYd*2is
zda|4A!0DUyk3X+3d;3$QE&Y4ieAeJ6_A<FjGP523vGeAEzW1d06DYaFF5s9EOe!^?
z8`wqkNbFi^mHYt@?X~OA>)I6hx2rb2;KXSGf!n|J){iGGj_EdYtVGGWF|ctovihef
z`yqam&|qS`ZLam8RaLk63=W4xAt95wy_&U)+buvZfF6=ZB$f3LQ&gl-l!=}0K5FG_
zF$OW5cacSn4(y8@t0kz`It)-52@(*iu1n54m6nyMF$(!R<#B4=Hl`CrHxE0PWt~XQ
zuNW>W<yL1sPD)O8f8W}uZ|)%@+Nnp#RVdrXKCW3_udi?SmEWSV@?m0ghCrff;s1IV
zhUCLXkA(HXGU50tFTbRmMS!NI5;l%^)mr+$)(?ACUP9b%ASuYHrQN)}1r(DyUEMXb
ze~#F8ZJB<!{}EVQBBDC4N9Rpczj@^HixqkjJPUWOctD@BTrjGECjy1q{Pu|=F2~Ns
zM%)*|n25JR=2A6Wx^$_8G1pxcxiVFu>IFfOEhcislE)Q<7A40pfq23R3CF4tSSEgY
zdC0*i_AKcKANMzMkG4!EfxQG_G&aV>#6$}_QmB@@6c9EYXV0Fkt*;>mz56qEWhrnE
z4#71v!ncsWY+Bv}jx&lCz_z&BULojXZ`_6ciHqD^rRaWkKnkN{nsY5L>f9I^lm>Gt
z{H(!6hHyt>vk`N_^93|S=v`%wZ-}&s`T|bDpmS{#gS<Kv9jaj-z!~vC61s1>4ovFn
z?^lz2K{vMKa5=McTF|rytYeL$N=g>1ET%4sWIZa$vG+d?LV|rh+^MAJLE`KpsHXP|
zO`#e#%gpTl_H}F4P(9N$6ggSh*LLT?w$p*jM68kAsrBP(-#k$iNE4wRoV!0jA1YiX
zCk{8)<wCb%D$-Gs!Fd?TE}1)ct<s7L`+w&m<2akQzIY7<qK2muKA@pvB+QF{TMx7n
zoH|m|rhN9<mjydm0xtv2OkTd!8Q_w(=4@uLyL$o!cl*o93>_M-&D}nOxv{HU?5hA_
zc7>>CC^}DZ%Vl>g+YSZ{yl*O_n23<c|G46}0-6YN{no7&mV<q5a7Ace{X=CjfUOIT
zKT9Hl>(6!%7rF_>F$%kGl-bw+I;Yg96i1bS*K*#`&lRHf0O0fSNn4}S0nap)rJ5#@
zlnU!5fwF#98lh>?3mD9imu`6WY!XAmDKpr%9m#*PfbO`S-lU@|;4_K_c|cIm1O2($
z)&}ByPBu0F-XXO*0r{&MMl3nKDz&=y3bUA+V2d!$rZY^OIHMRlfY$kg5S5Vj72WtZ
zyNcSur&_Nk%sbYrYfZo-kz@V8Y@cq`GEzSne)iHc3xkJ9qDATx=;?W$6?+}}trm|{
zl9Hl0>6#L5tUTonTo}q7ly66;v$JX&yiyG*0qT#Kd#-JzfQ+<{x4@C%*xvi@5BPJ4
zf1TP^Y7!sEeA33)7}v<wAvSFV8~BJ7Z(maK%Q$`9;^sZtHB)9uR&>{w()myFpHuu1
zI2c87O>tkNiE5Zg&r5xMFXvdAnmUSyi-oV)ryncd1$A_F=}}6UzUVd|P{xE}+pw`;
zy)MJrZ9*NY)Vr$%2PCN<F5DC;QoNXam?3i|EK*aUu_5WO)e;vKr;JDB;^BRfzl!*M
zM0?DslD{tqa_T+ay1ckb(@;D-Tp(~WK92y%8B39znVR_(pR7*iFG7^JXzZ|a6HdO$
z;zCsK(DK&ijAG;_cf_L6&rXoghe(RuCtz`0>@VGT<gdKJ_jf`;1@rU<^`5s`v*v^}
zk)WKB8-jVl-G!%?ljg3vv!~rt(d!wa9`i(@EstdFRb??+FdAv$a^cMU-RaWAfzI6G
zm&d$&`d+4-8E%mijgC4yS6Ov;%B+uu4nODQ;^;UD!+D4V3E})|A)$rcC<!EJG_D2j
z@!_RSaU<VEo7c26-R&LxDGMEc?O*f_s~y)yjv1pNoiO(tiN#}FO>J`PVgAk(|M`;G
zVwRrb70%5&L&HRLq0z>2+8Nk<lEkLrP>JYk@PdR8ACs7X#u~-aumCeFL9)H9`-@{B
zTv|#B`~=Bxh%i7eSGlltjspW(_aG1?=L>@bELBts5Vi)MvTyzTs?XQtu4HH#^w%e*
z;a?Zt;MTN1P*+n^k$C!?&sR``Qeda&^mF4~jC)yqRFTAxa$@4<q#g!w$Dq*f<)1cZ
zn@mhF`<&^4FR3CCCN{MAlcVHGrka|Xq7x#pugo<oiP&Dvm>Ob~c*269r|_KT)4{=S
z8HY1|xi&k}ygdH&>5a{s4JUarKxnl^l?q;E)xufp4D8->KC`!qe%q>rlTR4VbS*k4
z-q_k>|D0sky2=%QvJ51#7XQT3ECokO5lte}1(OUb_FYSumqBwPsG0R@24)1W2_5EQ
z<HjvDZrQwTo4OP?^svj)okb}i4R{=#@;5A7wk*k1FfJa~<l1dWpZYgFz2i2;CQBIo
zmjVJpE;v@dmFvVBbw<Oe6OVLZe<#=vzvc(O)zv9#L)K<{sfn5WbpQF2j<Ny>r*u`Y
zi$YKC2E$VT5J6aoUIU?0<S};R1%iT|g`tL4ip?A*V}bLa-sSV83bGd0yOjUF3$amv
zbC_GOYJPAL{rEitHopk7C8|tT9ly>jvu9<+`I8Plm=<gy_{2S)H*cP^ab{Ll+U4LS
zIe|`o<?D}kN&<5grabyIy2@(~9)EgpC>&0{gW`N`u(?JYwteMu!-V7raf~%Q22L>?
zotC??&7}{g$MY5|C%*J+X@u2uKdhm7BYtXsQ@5_*%jHX1#qm41zJ`Q9+pB6CW|=$h
zGbrW0n)sFosYHARwwZ@UshxXY_+?sZ!1hDdJN24&g!Yc(<kn9Nw!AR2W_56WUY;}!
zHE-cu_Uk7#ySFt8fGCKpp;{9TA{^7!=Z_?Lh(ap$^;Bk|B-@-{jI`%WNdHJYIH!DN
zi~8GRpZHnxw54T^mn_@<WhM<@*DP<_YF|Iu0cKb?$IjcJeKL>WEGCV>TLRWl6N#*O
z)Ey8-ixj=6Q=diqsj&Bg4TAUKOPyxZL9AKAA_n=t<H;wc`ypki2As&6GxWCpQFPp^
z;BcGz4ky<SEmbVoFAL&}>UTgs^?&_PMY4GADhU#KH(-Va1R$=OwxFuZRE4Na4YuTN
zdrQA$&K$vaNr(oWts5;dj{kIlw*F1s;DzPF(M%NKsEDVAA*h-k2U~W8?u(7p8S(ee
z!$hTaZ?X%Jx7zQk8Tmt8{LVS?8(OZ%q678eoPp{;WUPxXW6PXa&2slMuiWb4;e%w_
zT{n^GhlNiO)K%HHwJ6!-dJNrHT}ePpEIzt|@;iX>rzd4?o<6n|V?f#cJmxtm`6_UE
z<3sV<a^V($cg(BoY47ML+j(cf4;13!V2JL^gLp8WLU|GcIEph8JI-70_gXuugVp15
z?6X}r1s@Ur?iVDVg|#ykfk*|oW+E9y(B2fyRBETE3?c<QKQy<!6BjR>ey4zvk^ws`
zO(%J^+RjxO%Xr_uMJY+dj2&@{>Vv!DWeI!Ye?(r9W@x49K*iRj)~=elC7+=zn-)r1
zZ@n$;N~LzJk)2?*-z)g9M)I$*Gz@27$+v|>p`0S=Abd<dWgzIn?r8TVDGv75)|C^Q
z`OqCRJ-+g-ZJ;05vNu{}LjJ;o^ubmH<OWG5yqTA6m{vlVpjSlkDA+N|o6|EQLCtvI
zQM<6_x}#ZH2F0j8q%_R^%G0?>U}L?1?rG192>85AnW%JbLNt_?moJxOc#qFcZac00
zp)&vMn-ee^@Sxa{*qFM|>(tF22DfS#@?vG0CWRBZYzYhAVw9z2#eDtv3ijgYoSdAv
z<X<1$wN3299v_Cp5;_mN`BSmjvtdg+g_1>Yq4L@|k|?`a_JIx#NBn)_MW4yh^#?5j
zU1|8*u-FfS)Z}^3qAHiJwuT=JymD9C+w0P^yug~qD4g-q8#k;O@r!)%;)QYV7hilK
zL+Q!5FM6ma%8lrail8VUEC&!&9=Tj=9nri)(d2z7qaWvtUY%^K&CZW=`ek<y)z!Xr
zM_+{qdfcz9p$PG>%dCTYNtlY*A*&7v9Be%9DxBmGTVc0xKMbv^Sx;}&+(zlYR`mBM
zNQ6}q)b(Zt-x4sL({Jpv<Mohb!whjf%g`XA6c@iP-hg<cn5b#2%3b8E3o@Jy_NSG_
z9ywz8B(Q0GLsQdKL?A|%4A1p`fze@E#|pRbfD=>BSomFVj5qpTLrY5!zf_x0sE3YY
z_h%zfe;JO&i*Jf|)Gwsv*V7YC*=y4xq$R<=60)UDaSFAfru2+B;kQtC($NtoWrxgL
zG%4igOoxZGm@jT%U?PgEY?CsR1r}_U1*O0Lo;rSys}Q^pZkKa4z_8c3^~)UY|2CWC
zW92jz%0~%GQtY|4bf;eHygj#`dC_rnPR#zN^woBV=j#!M!){(vv?nmJYXfCW<E`VH
zv1_^D+uH}CuAAq5m>{S=2|^N;VtX%_LKt7w?%rJOlOxKhpM~SF<(>BbljHE=uQ5a~
jMja;o?>9v$2UYSC)eqae9kY<{#FDI++FPZ|-S+-Bp>z4s

literal 0
HcmV?d00001

diff --git a/_images/dc4bfaee117747c1eeba0b62c375704e720b95ba9694ecd512a2cc28d501fbab.png b/_images/dc4bfaee117747c1eeba0b62c375704e720b95ba9694ecd512a2cc28d501fbab.png
new file mode 100644
index 0000000000000000000000000000000000000000..aca9f710e43c877a54ca4717c1263d3cf55f6de8
GIT binary patch
literal 68307
zcmce-Wmwc-)HMu(bf=Oc-6%bjw1|{2fOJSV3|%5fhk(-E-JK#O4bt7+-S8a#_j6zG
zhxg0->GishVHkdMe*2ug_S$Q&2~w1o#zZ4QgM)*^lzA`l2@Vdy2M!M20|f~@W5?g}
z8hk@^5R*|s0slNuz664wQElIUc7TIJ$9wvM|2OY-2cG1ENvgw?ZH!^g`u0X}R{Aho
zOB<M_nZYY3BYOuk8*5Isw`^}&UYWvRwh(sq|L6B?Hufg$F_fV}aB#2SWF$mYT$1+Y
zojrc7Utb*$IXHFkPrKW^qWV!5nV$BVAv`!3;k>Zwq}b{gL__+vkbKO>s*qeov-&`(
z?At3Qlst@S0jeL(mka#6*Y~)5B=J6ei1y98?7AC*r*86p_=nEtE<NXi3ksOTP^kQH
zz<>UT1MiSu{rAuhXXQ^MjpG0Q0!QZA_or9*;TWL&e|;K*iPN}$iSqHP`(3~Yb9Qz%
z4J~cC`{jWT*7L7fSv0~oTbVvs;wV%dud+$E;J~|LV9LnI<Z4qA5p{Kik+Lgt#rnG+
zHpeb4>7{sHOjGpLT26oKG<cbo_w-t)C`2?Wzq(g7meVUYu8S?6cURpa#+|{0h=_=A
zY2)U%=Xa`wguZRVrJFBU!dr8};yZfYJ1c@)AeMOd?(41yJ3Bj^ho`4@-_w7yGnPDF
zx3i0Hz3yD-6jJ)5Xl2W0_k|`Nk0ZRl?2P4SWo2=g43epq>ddctfMqsaZzKwy{G|?k
z$?4xs=}q=^Gu1m~y3&Nxe2gJSA$5@5`&w2*LqqHKen+j<peytx7cXyPjL>lhet!aY
zNKH*mWy`|dorl+vFXfMz7%2@6k`N-+jJ&+OZl0GfU(R_y-e=|JM#sc-HMySh*)GAs
z(bCf1ZApt5^!{WFCE*UPci613U;i-KBJy}WG+kpw%B))Sijgr01lDP9YPsRU#u>S|
zq{R4mr8AUVAZBb#g@{G{ce;3h)~@s3l<22Vc$&4=>w^M^U&iv3`r|nRN?UFUq?_2A
z&+s@+2Ek5!55lLnf4sjOot}1`I1IuO3v3gOdIzHCJ2-38DyOISsxO9hbvrk8BiVUD
zAH==n-234)^>d2{zk={JO`%%p#?5Z&W_dS-;cShSP<SP5`_sRF|J2J3o@0yLallDs
zMlRlM@D_cp^fyWMj7b)7KABy7n5Z7YR}l0G2tZa4x#xldS1~@{85exKJ9Tk!NrK)b
zCnrC}2IlB^JfUgx^ffXu(e>v$dbvh1+}hfj?WDHz^7?uwsZCQ`Fd=i@<$}8%BpI%%
zwwBXwMMB4Qoe=EcKymFtZ?uAN&&B>Myvt60?h64YmP(T$V{k{7(-q&Vs;ck}l%T(m
zDLsCGJ<@U8rj23O4Npy_yqtHYJU>4ds=04{ycMBvS%L?rm|7}~q;sV+#LV3MncjSZ
z<K3?A<4S&RiUD}sbTMP`F&T1Rw1E8-x=<!?bhY$mYi-o<@$tulgM;xWwCy?xcMz}z
zhYhFiFBePMbehzmP^guSjaqbc^my%}S7SGi$lWsb_;YlkIvsjCx&j5k<BpexpKGm?
z6OxjOAY@pR6?1c1j$k8qK(<M9v$D#0-k(n@xQ9kX;c>rtqcmS_5nE7D&|o?JmYbU!
zT3lG@V5O|AoEc8W|4F;t;5+yoNXCc)ht0u#y9LhARi=?QW6BEkI*SfNLOVlV7nS%m
zvL8RHsFi5tgV%b2i<=u5WF$mKuWBrgk7w&RO&jejK3woZp{>`RMcR!u(oI1@K|48e
zi7HU&_@s`zT=UINA+&JZZQ>YY?Te%@AJf2z5;eIxS{BklMnd|ep@H^JNeP!HG0bkY
z>wL<1uFmc$jyvTn+}!7WhNYUd;k*kjz8M*>XJ%$>A<0vgJm%)+C}?QUUa_%-gA7bE
zNpbZz8_lsB61qHNdH?}gSy^#|*_~frf|&UyC6U3s<>ZWqI5Lq8^a!lOV6ZZ9B3rIj
z(7|=erY+TrRp~y)4Hv50+uI*qT+j&%FI4yIar{U~_+V{q9bwga-*S6<n>lWtUs#Bs
zcd<A9gaBG#?>@%A{V!M6f2pl(Z%-GxJfV}|HWt`wgW-|gb0Hn(;BN&5&BEb{iJB3U
z$XH}c=>Zf5dsF2s+6_NJF8ly1N7|w=xw2|>+7~Mv|MdB@jGCHS>5ym3!%=`pnfW*q
zg~yTTg2$<xYKc})mD#9#U?scfr6;o1yt41NZ&!PjL+TY@P+MABnyywtv>XPwRa8{Y
zGz9_V-2~A1fNyL}N=iyEFp!+~XT}Y$OSBu`zkkoS-{O9;*Vf$~&>2Gf1X~6M26;tA
zg8Gf?eX*5xt5OSIw^kt7;L<Xs2B8rV^z7`V-EoaGGa6`UXj*UR=;<r1W_i@2&b1qz
z4~-Az>Yrdlzug}f7gtnYUq8a?%uNU@V7028hK6Q%YHBKTe8{}o{%EO<vqg8&h~q^Z
zj}=LT$O9ip@|?=b==0sl%bgcsu}CTZcg-N~Eb3)nTOZC#J#Myh2?z*|{>ieJ>$f8j
zv1-6(JPpBuOG{0zPu73_{8<LjaKZC(o?>=wbF=bG4_=DLiOj#z(d*IFhYpa*zX3pd
zO}rUQ79iHu)g@xr`DkoR2cVWkrztKi?bSwqJi>2)BaBQ;D<jgB>%-~d1jNLr0N*NX
z7KPP{NO>%mmfC#RHa1Q|bnm|H?Cel%qGDsyb8<$ntQhR<?kdcD-~$lzNm<!-EmC@_
z(gb^YdRp+|Y87CW-R*vD>Nz*7ZcCC;UyR+_U)<i_UiE5oLMtmPp@#UUI>z@Nj{(%&
zj7pbd1MB*Aq8OIXf9C2PK7$v515Fl9CkNU0Qq!_~4$8#J^=7IGh&RB_E^bFeSnZGd
zvs@7)P^IRaX070|va^p@DIeo<Q@w7A8if-+(&>1fLotvLoaY^h(TP}o&(9~T>9}@s
ztLcL>VAFg~1R(7x(4^LNJ|I^IxOG<m2#cz#zi?jk<OLAQE-ESt%2r{w>p|n8u!o07
z-LgMv9G4j$sOCacM3PGC4KT*?@^ZnW-w6GQyb<7tJlr38U%sexq0`aP@wi$Jc)C8n
z!^RP)38x^hZ$c02pGu0f(vZ=Q|H~}I0z;yzo{IVt(dfW=EdwCpKmLJ1^<S<3e?Q&e
zgOyh5=;Ra?87T@%df}gh|NRR>tN`DH1d>mmK5@i<q%Ue+i~1cJigD0%I`|*u`0p){
z#sm(V+Na7571X_YXno+};on_8M@B}*w&nVt_eHVOhpMXDZzel8*k>L#fNQL3zw`C=
z1^Z69WcuGVK#uDrnY$zr_~ZW1-vm-g{8uHPetIEBgG%uK%hSBW@PQ$|?F%xR%8ab%
z^;O(NcmxjoIuGqg5SIUmrv$qU9-klop7z+UL-?dC|E$c>##ER96w26tg&j8jvc!lL
z)Aj0mc-vm$S5=b^W<A8<*O-0(v-l^jrL#7ZAe+xPE%koy*H)M@SHkGy2Pjn1amLhw
ziOOWCFRuXMYZNtQ`lx#vZ^E`|aYwrVT9}lArXj9;E+$;1ji9K~C`X(J!(W%!h7+Og
zsP{)_;Zz1!VReb(x0m0*3DClk&yBXqcArvq3cUSi(>}WOW|#_<%5RH03K7d-f50eC
z&{sZTD!(9lXQlWjh4%Az+SBq?ZqBa*<^vTJnXm$W?PLtp#rwXw?NHKeK_A~vYj8n#
zvS5aTG$Th+w*+L2jHo0Bp~0#r;fE*G{ZtRb-S#|MsLUj=QUg>?;PkFKI#r0R#7wwG
zM?T3zm({*lxu?`sJ+=wxNwl`|f<O$)i;APR-tUyo#k`FzT+K~B_bT1j?3$S?g`3i7
zO6<;IF|brO_x{cZ()d{vjqMM!Ji+&lJ(GOosxyhQ!=_;pVQ~RAe@!QO3PwfsH{R|r
zUwq4;dgbv9%boa~@Rot^jLoP%6$Eg2LCb!|K{6ugkYw_5zhhu#`n|L)ZHn-luix4i
zeA?E=4Rs}zP=Zqy?qAc8KOP0G?YYMP5kHLfn9ke^i)nEc2PZQr_K{3OQdW(*bk8+H
z?V?HgM}Wu!MGK-UlC9Amfgr{KL-04Lku>G1t2$SG=b3461AL1NC2~nU7VeMjvW*KX
zlx?ycrsoKrGEcW2P2=MmSO8({yG9IX5ALDHfy-_G(3b9;S2`uT7MQ5Sr4erV$UHHK
zbSKu0>uCA%-B_UOSj>JJ%O+<L`MGJ8)W?GdQ<&7=1W!hkP1`)4lpz#ppLJdRE=ryx
zplvOzZEfGpv+4X0242?@uJJo^-#fcv3H`PI{4nxiAOMQfUyF+Ub!AF?Nmv)EUxw$#
zgB#nz!mc%^<66(FZWjoh@s%^W^m*Kl8cQF#VxQ|eINqJ16Ffo8!)uwPSD&J)qe{Xd
z8PW^ky6+*w+P^Bf+%c=OXZfpKQAxtEz|FX<I3h~P?JI>!W{%^(;D65kxLh^1xq>f<
zRy+v%<LV(Ub!qx}kC`II7Ccnetcm`em;Ykk<MxgW#7BP2diVD-Y{~5Cof5(1$xXNv
zi9q5@+#kQIWjK3E2Iwc~@sd@gl?)lP2MGC+ndHZVf_r11&+=(SBmBOLoaAJ6dkvQ_
z+8%Tsv=(;%Xe6QcdO79?e7!tKT~G0--=Z0X%`pl8thO*xt`Ngvtdlz*CBo`nF8<(R
zLT`Z^57OOI_YnUl@5++dNGavqijI4MDrQick<R3sBvgu%Y|ossFoJ*=0CKFCOdfgH
zo_A!utfKh~`!!}K91SwvNys*~&$C1{XWJ~g%c+xt6w087@)AP{*8u!=@`cs9*zS*i
zmSJf^6YXn!l1f)SS+_18`-fN44T}z^fzue+fhI2R#JqKKK(M`@f~{!3MonUY<(r{@
z)e$g~^rh+hG(AzJ!k)|ph1N$3c(C;NKcScW4GxTUnTKrx97bvFcw&LFQlA}YhrI96
zI^~SuObIYYGwDCgtI8TX@;o-)$Y4MapvBy99XAqq>*P)o-H+DKT(t7|19i0?aJ(08
z-X2o!uzeu%WQlB|U57jqLfTl)1j|{7rm_9%SOZnoZhG!4%jbg|t-$cx{>=$Gxo;xB
z-(2$}oy{YuP(v-KIefeW+QoMcle2qF0&u^J86E1`?2t6Vh?-R+3JUHVZ(H+C^Q8(+
z1JTe|Th3-VNZp-qg(zhn?3(zTQNxl4JZ!J<KP2>CH+;cn1=q3M!M@%5kAKC|_#hOO
zc(AA8q+W=`sz}N?J1|hwv2JXY5HkwjQ7t^QkL=9<hthQn)w2*6<JD9~vyxXM>@nXB
zH}LoIM+`|Nm0V2!-QQD$NM*K7&L>iaHY0om**)Nm?fr*-nzWwB%(yjG7nICvaJz5-
zAa8<&)bH@v^AV5ulG}W@B2W1%?%98SDW!v4PxxTT06K^Uuy*Fc?U~U1^pD!_-UnXD
zUjgQ>?$nu?awTw*E&r+Kcg0FXt3B;{z|ONQUVeCUiVUHl8vNW?>`Tq<I+g6SCohnK
zgxw@Z^xAzSTO%;U2i#{~b2gcU33dAMBnFJ(zN;NC+<VH!ywAyC-}DuqF&$Csl@FFI
zc=s2%qd^%A%{xrj7j1+d9x$Pyp$G3K&n59Q4WWpS5AfURzBZxN@|o*xsFz2D>?tNz
z2Wn5p?2}7^#m#FfHY8-`%BfFPA^X|x+f~lVWx})^LC9V6GCxdy!%5IfEIx&yn%7Zb
z<6xVd`*-A7X`nuD27|^<uXA+eX6johMU*#R#%jh-|C+SIKTk=zppLp*9W-yX>U4)w
zzuV@7yVKiC_I(SFSmwlo_r+$pZS6D5fL0&+td8X7zE98E_Z^ttv1|vu3m4I&1Q$vD
zK8|mMvg66!B>SlG?yC;6zgdB>t^h|heO<>d<dSkmWF<YzBh{^|mumMPbKc{|B3^S*
zyaOfo#hk~RI_(+4aV`xZP@$yX&;25szs3&c6#3of{Q}Z_`|=VBC2TClOdga=VM3>3
zgQTu|Nd`3Ie~pN+<}0KrD@*^`5RTOJktjhx^N6dCQ`l0)J79*npuk|)GAl!m+R>;{
z5sIAhTP50OAX!yavs$ZgU*fRplWu*)O|NHNy!iX~x-MPI*vS`c7#cz(Mn^DL!dRrj
zcK>A>!l~NWz3hRBwy<^+JL#P;p$Q>DS1P!dQ0dPV+lO}^^M^0-6ME0vOI{CN{__fC
zXR`uT8mF;1dE-XwTSg9cyt%ns=JzQj^uIF?&uBk0qXyTLIXjp}Sb-Aow)Fwu$pRfN
zSSnzric+%>C)tD9OW%;)c1;Xr#9g)V7r^y+oD~0nwq)>#(UO5bJ*}_SaSC$XI`H_e
zbo}&|+1o#N1VM!`dSQ(H&5YH46>sbcYzmGL3KfMaDJ@SSo*-pD&ukQN%OozwyoNcB
zDMi*&SePKd5w%J@%goV-ozhAc=9=A-k`+P3oErZC40h0-BiLbbQjWPteW{sBO=FC7
z&}K%!16{JJP+T05u=a$vaNr#oo5uL6I>N!}gofU|u)sda>YoCYo;benArurg!ZmC?
zS2Qx3&{z<v!wVUSp8d3uV*mIvjppfy2@Bpa&VINkEB=Cw-rc{sbG1$U+JovQ38YO{
zIRrt4UbrnEfxP$mx;1KhpLgAEPgy5BE+OG&`NO4^TxW?DVOYxVzkOwznt6p&$OX}V
z`i)#%qzz6`$ns=y?paz}0ykT|zEg%aMt}1+ppy#K2g|UQtH-h#%~pFSqt#GDaF{VZ
z)3xW<S1v$q*}vGkx*KB)_Ejs958>WfgQf_Jl-<((7dTE7i4`c`*o$64ve0A|LN7^o
z*_UC^mw5%Ndrm4`X^=*5l`2F@-Ee2UUQ}_Cl$JXxELG#<nHHla=Pl|B=$EC<gZ6Qm
z#deLdn5aGOrbg%IMoyX$_j541s$hSKM?MxQg^6`512G8_YG424$#Wi66UEP0grEpN
zmZ4txhYFXj28BGOlgtG%+#Z|A+J|S(vE)nhzRW>oen)?eMEKUp;=8jEX(RD|O1R1r
zKGtqiHdSnT{$ZsGMftc_K5xRW8}&IkOx*D;4~$0nbvL};@XMR!9RTn^>D+mTJQDjs
zRyLmBfev(ZuLK1>(%Sj6lZ4nur*&j|f2Lx`$NXMZ$@Pu#_h35VT`F@<8`%QXt>m(e
z9mcZY)5AbQ-dld~4lp=1Y4U*D=C7T+DOYE=NGEs_A`O@AssF5Z`Sh>nvY0<}+-$%H
zz7JebhW_fFWEF)%QxeY!f-@8sw0bCOnx2+!4Z)9)cl{bE+HPAQA3m{hoGM<(uNB=6
zicfs3V6N`Uj%ofY=S2&30n*`ZA=^`ouFI*p611-wqN4DUcAgKagjP8<Faft{UgK20
zXVi3tXDIKS{mr~VjelIYrNmO9$Z04J#~FI(^>&%U9-a5pe8I^8rf~*A!2dHEI7VtW
zMkWR&kP7w2DeejAOcfTlaz|hkmFbsBYKt3=Y;IYY=n@M*xI>2c(rXmMTc~m7kERM2
zE~6M6neAB871G7Vl<<%h`0mrXZ<jfq+!lIy%>>s2gt$JBWrxEtz*_SucLpMN%A&_L
zac^Rwy4$?Xh?)B&8z}4D5yD;xN^R}+tm|$t5>|n}i%MY|bWlf4xivyH#3K$XckxW@
zooAXH8eOQ|wm<SmRv}`2O`jq>n<WS%3ekN=)as|RZ+L=2!*TMq0IfBvo{cm_cO~AL
zfS9nbU8=DdK~kxXeV<45i>H*QG&LOHT?k~!I3}f8`AHdFpNPWi=g;{m%~3b7D^m=N
zCQLkjT%&qV-MwblY4w9j3!8xKNxRjT_NM(>ZPdsxc!Q(z3)B!C9J}eyp2(7bz~tdJ
zd&6apn_fSZ*CV$6eM}&BZtlb<O+k2<J8~Rdm6*pi6<B%j%Bew<Sc#l~%#FCYt-qbx
z<a9B^bFWyiKfQoD6pKRC{K$5LW^n6`j-sQZ?sx8K()?Z1rJN}luAi3KB_6yCBzWDz
zUw5jatM8wmU^htuj#{Ugwf5MV(NrG=Ui~?>Hv883o}6_;wMl+d`HS;3Jizfu?8ad1
zw<61Gx`GihCRw8koyqyU=MZ>Al(Qp4os(*hANBqSzdo0&Q>FNLuzT`w8nPoSoi+`?
zri;mndJQ}uh^)!^iK7Rnc27}9CTUQ~;0CIv=e0!<kWI2{+gP{`%*|z;6*U_GF_rx>
zuK#@LK#7-=Ss$X-t<athakY5#2J4ESDB-EtMRRezY!rm7!%wE&`G?bi#Ze%+Hn1ZF
zv~%;rWg%f7md~Y)J6QB;t#}%`IGuiaC_iPrQ2X+hhwGR7s15CKrZ?JmC^+S3TTG##
zG00hPY{O4uHlW8}#cw<0WzSB}3}QB*!zB{%UHaQI)u1uF^Kt9!K!wOl#&N_LqS2dv
zz`zOpwQS7GQ{!w`cfxVYjD33X@R$Kl28$qv<uL_kq!q}kxhX1^1*v3)bEKMnQnv0I
z;6(Yl6!-G!Dias2Ke!$OT->B}hc89XbO1vYdA$4Fu%``ZvpsZ<%m2DNU5TJ3_F#3<
z{ZTxmdcR6WPtt;0e?!k&rX{@3kgoELH@bTtXW&KnQbSh!CfV%x@#xhn2TZt!gO<5O
zCv-zG>5DB@2)!K=T>G<?D{s=r{(}mHk3l5fgPTop)5TV>8mOV|X`9}O@Ve8)Al(^~
z0wc=Q(9^E4p7WbJ(z&9!Vgdf(m2Q~Ncumq+jhWnLm?n<r`S)Op?OB${g^H>Y0W3HM
zgz4C9khn;&BBr>r@#&b^I%2&M#x3fb?`XYz2w*t01^6`f&Y#_{Zz4n6Pg4x#<Va-u
z=4_atBsj5m(=`#k&TDG?J_gRIdQv&z{td-~5>5CuJjwp=P}$&%pe6Q7Z%L_hr7l0G
zhyc)^m^v;JG@}M3%h>i4a;Z@4u?c1}_{|8sEH$;`BeyK<$<4?N3e906vC<A5w%}ky
zLrcr(gx}k)h%~B<HmVfg#eLutqdJ*7ZpSlTC5FTlk$Y#;SkTJgd(n)lV%X2hUi>pU
zTmgjY&d)rFYUAJru2|bDcW!YR_G3{|?1SFx;0&Ct<?g#kUJmay%uVvd+MYiSl__G-
z@lhg=XYWrU+Os&Lv^9rKWYP*Cb`Z-cE32_z`(4&H$hnobK7l(DQ1X&gCGtG^+`=v8
zh*gh9$8$gb3C>WG+rhkw(Pne2zs~o2gg-(Twt_l&;Zv6m-Xv)6pQ{jAX4P0CFNo%M
zglX8`IE0-3%T4tFiW=+s`g+ME;M{|Q@(&M}tq$W*waA}8{YFeB3h(JE!9e3u=EDbj
ztH(d+>_=_Ll*1z<qHb=GC*czny%82g3w8Xb9{qiqS2V>oiTbhZScTXor^Q~xCo&cj
z#+QLTojm`X36_xZoHN7o^$T*rZvexbSqm+TjUOtIwM#=tWb>^G4{Hrb#F8+f(9f}r
zL79%%bCGVVW{sK)sluu6OF5V{3S~|r_G4WVb}M?OZ<+GsGyI20%RmbONLJ?C)gFf6
zjh~&G*8DudzYk!;Mpn|hTZk%2Sfs&HAyS6F!^Ip;&`S?><yGa_`rZm}j4)z*;QLC$
zEeB10^0;axFk+#Gtn*w8Q)DF-M3<t~MsE?j|J#1ikE8&X(5Ru{8d_bi3)ofv+TVwK
z7K(LGr*nF)V5G*;Hj{eYlg|gSTh(qQ3Jt-tQch|#td<Px$BwP5K2@$-L_ASl3GGjN
zr+)~eb;#Q(e-yAQ2*;1&51R;-%G`Z$SLI>|Mv@$%xNUtFOth}K>goPT{${m@h=Jbv
zD0x7kWno1G16!8QN{$xr#1WNf8LMq9zN-t=SsqpLtb983(yz<({83xzo(YI0pDS~^
zcX<~*qO)aV-&)OZO;*I}dc~vS(<Q+kfui~ysM&E~V3P>+WH7P(3{>h4X-ZGn416nb
z^u@-fy$K?ZZ*C&LwP6mN$eFxuBNOU~Kw0}_*`0$hlDy`>cQHmft?T;YD_<(Z#ZO9z
zeqt-&WQ$tY>W@f_9gJf&lcy_#y!<MTKNTHtDhCxII+-IJ%;}H4eFEe>Xjsh=swS_6
z_q}a7kuPWK*|?IWLI%Ge_0tz|K6`T@EXcWvyD)VL4|)ih+~&e=J{CZ@xK#FXTkI(Z
z&iUg!SKrj%;*&Xaex>hSAqAKP`ab=Kzq<+z3zmBam}BNbj<QQRf5-teyuklgd#Md%
zt6$G`mte4w5smK|3@$E`%P{rw&xyyFFMX$Z57eJccw)P2|NA(3rJLB^Rgjt5$n?Cm
zY@ZPBq5crTKBCB=sOPJcMF%|tq2#HY81e__*SRC<rZ7`3H-Kl>8#w*K_op2Btw262
z2&w%PbIO>O!Rv;b>P0EVtNWHmME0O%@`dm*)i(d6!c8@H2bdIhgIOsD$&fYN-tA%Q
z5D@c&!uB7iz4pqx;{;vbf?<a~m@3`B+<87aq4Ml&u0lVYw7nhYZO1<qXM)^OHf>in
zDjsR)9+!I=r0wo23ES19&lsl>CbA>)CnUjeb+p$Smhu9>Sh@EYMFQO{J$~R2gSpNZ
zRZ7|}`lwBwRtZr<)Ja<H9@d`8l?erh+vMVglliG)){b0uvI@h+Prd==bnXa==+5@g
zJC#X$Nu@yY>aRe|<Qk_)uck`%@4Z)9E0r#EXxBgdy>*}fBs7<sOOb+)|LX<lNXl53
zIQ#CTiCk@U&IhH(8V_pU{+$jf6Cn5aSxqG3OiAaLn+(%EEvKSl&E3M986J`F5`R;h
z-M>HDg$M*uj#VL&LQeO7lR2%}{h6E=1Kpuy0?c*-R`2)8Ve@#R0maFSdoGE$s=_w|
z=L~491-&`)0=3!~hAaz_v1j6_VH58FoFhsq93H5;HFe^>C%bHYl1ZEHZGL!68=AIo
zW{2~6-#cA4fw>SGI(oc_w^zh^gy0DZ80!rWD_XbSJ6SiMtHOc7I0G{?HSKz$0V7e!
z_q;rYii!&P1VkQQ-qE?aIm6FJMn)o9a*Jj*jMBN6?`+OUwiMaSTd20j)DAN9>0~va
z>{J(qx>+~MG#AcU3n=fbJ*nV?t##&-1d+osmRB=Lr*s-(c{Bx5-m0o-y*i}O>XDY7
zsMld9sr~x0JQ|rJ!y3T#PqJ$i-k~2<+7cEj!zj<@Jg1c!N~Q5o(l}jkNW8hzkJS){
z)ji_?3nsh3GMB9#9oLb)IEX#Ml%Y4y`lzuO8%)~wr2HmaxH<jYb3A{>AKxgw=+S$9
z8w#Oz3k#1?u0lL0{EMhJ_p;&Dr9V*T7g{FG%>GipGI$%=Zj+HMOsT2H37QHZ$?qJb
z6oJy%pf%469e;wLoew;X;lB)qBlY)NoHCz@gvsX&dUJU%p7S7U?!;{nFPhk48wY3b
zr=u4nn&2q))GEdu3u8i&kI?s151Pb3x6=9?HeXUG=gQN*UP<(+C~bX!n3(~2hor`4
zaWFGNIC$QAGhE;xt-88Vq_*>4Wkd9|rBm^3$<vgz_3{4j-VxPvaJy<0h~@##KdV%d
z=mG5pt-v{<J1uIHp9d%1(HW&&x$wLZ1)XH5-LK$F$6c?oinkW5hT_UWXISm`7^iWl
zcEX@1tiyZ9n;g^=X^1TO3`H@a84c=)BV_;Z<Dq55lq(dSasbc^g)O`r?FAp|BP=rb
zcP=YhdlSYV!izFm4A>7gh7Tst>=yR01X>rXe)<p6-OV-0oJAOkxsbG2FKlsXa0L^E
zWu#5+aF}9txTH2L1p_o95i;fPjQf4(|2k9yVOfW2lf1TlLR4uYfNT8AG7N@G=I}iI
zxDuY@)@u}#^VP~&Dm;~)+6dTT=rWzkc#nAc>fgb|z#itPqu;nLnqp?<H3&zd8zg^N
z+SEf+5R-CDRjk~F29jA}{6%BcyeE!56m<T%xclVcrHQ;LU85odQ0Ol<cjDlTHkX*_
zX<p|Vy?==uNjZ*#`H*}oo8S2m)C&y0!d0(WoA(gHYf^@Ot?=qfRA+}56bq4WhfFB9
zV`<6?3W707dBB+ZA4MO%T*BIo>r9o|900l^jp|^5gE|C-p@X(&R#sM_3$V3zVDsD$
z>H`w38}hp=dV4=SO}kp^LWR|U_yJU;$Fk^%AOaJG5z}K#narFki@h4B>vwe$|92+=
z6FnL2S)^xOOH4u7PNxRz(;JT*`s}#oAT}U+cM=?M>?q*)a>r+Qc%%c!jxv}OgMZWX
z5y->&q_v%xy=QLNVI<^_O%IoGVf2qF4Z|wX?1C+3063_jN#7dzVQJzdwyW(@r!p5(
zPQNJh%+IEwMq@h|rjGxk_>+5Qa@`toLGa<1PwqWig!}HJfEwW?d>)V$kn$+5RoQl$
z#p;8V>r0+}3*LLdP^V|XehM^a8luWpL|B>vbP=!q0JcaTH@9nO#s!%Z^bL})AFsAd
z($$KvVxi-74r>Ui#V%S38cKcM(A^dghwx@6v~F9jU#96BTjQ@*>qp+_n6&gsCZLT<
zyKRJtU77C5BfAw-fU~8<RRybFQAuV9_V0}`@6{}5FhnI?H`gAwaO6vtZy-}ufjSZb
zG19T3@$MBkl%d#^o-DY6s;O!tD~jam`#cw_WEl;`SYOY65R}LzOz$2F0{AC&<ozDs
z$1rqXben*YB0<0CulCi$JEhc1=2<MOrj)I%ZNxVY`pgm2hR3ok>tl3ubZTm9-$?34
z5rQWKWn(kGI*L|9s!(U4uOh=TKL2&gaog3H!WDeFO8nLzjKcu)e)Y=WU%WGpZ_2)a
zvP2TdjE|x356k<rI8XM*6^?FN)(LoiDo&~QW<V9dU4p6tR<BOZQ-sCL6OjBO6(uF^
zV##bX2BHf%_sL>=>PaIrWp~=R@;{Y9o7OvK%w)NuzS)kDN!GUcSH-`GGv{u^fR=>b
zJtv%onpRyIRFR}$TGDq=^4eujG<alwas0!ZF--Kpc7wAu#qr2W`DYQMrsp(ax4Pkb
zS|55Me*qrCQJ?BDgKcLF_%i0TCMq}bxy&<^`A+==NP_8(Wlml>M&oGs{I;I-wr~O@
zF`h!GT&;W8H2=4?1*n`^4Qmz}JN8<;PqQ8yTaDJ*65T|;XqrbpjfNS3)*=6L(Cl~7
zw<|Qwwk%gxGk28~6_vn2xO4f_90!vMg&+u9F92~^c|xn_=57F#dcTOC5GQx~8N;5k
z+%|+}sQ__D#ZjF@h>fHyH%|a&;%NAQzIf({!SGJ2Oa^qEJK1mE%tUz^%Tm5^S4Rh$
zYzhF1p@gkTM~YM0;`DFP;D*iQJ<_3$*JeY!?@OKs(HTwVE%U-`aQf{pn?Jm}qIvDV
z$d%;|Nh%dqc((5(09;9iZQq^?AuSsl8?V41pj8md94RyG#?o=$=LXi0M7+b(&7sY4
zHC>~Xjv&P6&yVLEhH%Nr2jbos-az`%OA0gLEu@7+BbA4m%K4bBHxxfoh=NiYOLK&$
zFVYwDTiI^1@woli8DSo}t}}lt%-x+LL5EDYQsUg6Ha@C<XyCbE>~+%fXz28ZFt4Cj
zuH8g`r(KL9xkmos?z|7K$=j50l16Kgo%aD1P(3R~hiio{b0G7rbS+oj>?N)IQH1`A
zsF*MqBZ<S~=8)_I&;W+BM*OL1ko`K%KX$wU=_4<J#)T4`;3h--a|<&pgK9<q=*a&Z
zW^@aksjbh1zIfeBo6FYVr@*u+rTJi%j0=*W8^;F5zM=5=CF1tpqG}e6dzx@VakOw*
zOp?JkPMSn8H0Dwv@6g{(E3Wsu*hulFU+`fB3LYrF4SLPtRV1u35A&>khV7r+i~8~)
zy%+zQr;@P%5vZqIwim<@T15kjiG<!C)Kxw9pNB-I%r^N3$GJmgOcvzMxdXfB$3Tk$
ztpk_|vUs5qd}&gU+vb^T{&S4Io5)WbkO6j&*K+Eep&|9@#`zts8Wj4pFG`<32dNg)
za&i3(p!ABHwRtoyI|o!10QmA#InK=VePh^lv=YLugEuNp<{kCrn6d%k>^1?^v}^24
z5H|S$IVNfCXM-v)a(MrVE5FCOqm@e@hzm=hS24%nMn2hicsyBNR)ANJ9G@ST5QTrv
z#Q=_=)N{<ua>fw1^Qv+%HZW<qs@y&AmYzI49yxzejg%69jAGZK^ksalVq*)bBq((u
zw<mYYTbIbv(0g6KM5fN;cUSU++FmwLVqi!L22u+PLYGB?xFXpMhaAFv*LVS=cD=8u
zbAclFdyib)!Sf=Xmxz8{CArDL_Iad^{>S~yo`3LCh%`_90AP5k{DUx1K<1SLO<{f-
z-`rr&LAeZ{Kr12-1AHG0YDTnD(F^dC7Z8dUFk8!xufAu&qtQjcoprU1ud>iG1pQ3Q
z^1)iIYPtPPV2)*AKISs9VYQNjJDH?*hze$2S5=Qx4!aIxKP{tajs(wK@%StzSd0fC
z$2}LounE$l%j|O8l<UdB`Sj&UKJk_Fwi7@3EoHRy(z^hAnK!pxO3*Vx;E8a#mVG=5
znLj>AYv6<-3u#Iv`@9>9cAjrw+y3xP5CP1{Et#i7YI|08bP$;2q3fnr6pek@pT^Wb
zR1M{a;CpB$RwvUYK1d}N`%;FtCaR~t_j<Ul1VD?eptF^5bOlo63J9i0V>cqU+W&rO
zQE<vwudqKWC^!LIBcU!<=MFneHNBL>kY_rX1X2*K>k0M06=x!`exIo%CopF|cDUrv
zHH^E*V;s6Cu#iSTs&x?PQoN9SnB;<S+1>sny*XR<`k;tc8`0z=gS(BGsNfl~4x3>V
zJfi@7v9(&Ijoy?-KWNebcDi2~*AY}Ao74mPJP+-r&8AIhC;=u=8I-_CcP6p#3q_j)
z4StdKE6{>nf89nlUUaE~@b2_pp2^~El{qfKW=2RAl73u$S}t=;8+nP`DX6|VOI}qQ
z9*apy=y77#02&v;yY12|j$KYK$n|ZmkZNPKZ|lCx0C&*6p)yf@_~XjlqQ2vln~IAp
z>IVbRetwQQe#3s?hVG56%ShmW%V5cpomKzsyP>tTGZyOy&-8kp#F7h*4CM{kXD%^K
za`WT9q>zL`e$+F}mvXKLCO{~m5;U{c^>^;6gjj;pNNc+@H0zK(YFcmB!JRuV1%-mf
zS=w-1)Z?2JE`bncSl3t8L;p4!WhZSErG;e)m+uw%wzsbIKEvAlRlqdhxZXI#h-z8u
z1cauKRSGVpIln$7V?i{Sm0Cy7r?w6P4qhjwe298yrv&O6u-O2sK&C>WAI~^Nc0OH5
zLR2x_)%P%Zu7e%foxVTYWlE8)DZSsqUVIb=6I&%EU|v}vhH$LfT6s>9SFvBYU%h4a
zf)qxS(ASpyaB1C&ww^S$+T2VPk%d}d!imIRi3n|EL!ig;hhcF8)j^!~x9tJ^TFKNh
zOOkpD*Hi=qbSVZWI}a1ns^SlAwIeSDk5NAeonk%PN>L_AE3e#6iU6~)PG&tmsAM8X
z!)>4Xna<SbhE!DxQU!&_pj~IX^t^%(+au##s^l~r_Gq~6)^%x{#PbA_%O_hhO^!>g
zWW@sxVXlH|kY<4t9VHO$ww2VeLVvj2owfR6_DC>@5iK;40>i4t&v68A{C|vA(ajcH
ztsS_B)#PO0Pq?~BzP0=lF|@6L)VowSink(o6%dx4^Ti?+in?!FOyj)RD(aYFdX^_o
z>S!cjHw6hK^>n6`5ad$<s&IuKHQYoA9k3c0*4BJdtHI|TdNs$oOe~nwLWV!3W=$p>
z0z~rbjHO))AoCT#rUahK92ixp)J~j$YRCEJ^?j8?q-U&>31E&NyxN0wnGBBjRenhp
zDbNs&e%~es^B;rS8s*P*krz3iW;A=@z}we``W`%|RFl;^btR`aUfU;|K6VbS6h{0Z
zal}Gi)c?oetSQZKT4h{Dd)k@_jYn@kuX_)lp0y5P?<6kB1NMcrb?Qj^n-BtrIcCST
z<Z1KrB<Nh;m}52<#O)KteUufasCU<u%;b*x;q_QvU#GBGxvghJ6hA)w<kWfNQju?p
zykCR0AX0^<%WDJZidz`P_k%U4Bg7WcQ0B;!o@Q@!#gNl^o&?`Ez`3Qh5CzZea!}>R
zCvCmKfk8*9v2!|ji(X%x@3%MqD51nfqj3EaWb>qQ$Ic3CY=4qpLBS`r7IcMF^{B4<
zQj7ZenXWdu*37HF(io(VOJ?E2B3kpJx27NUbU8?cWX?jC!Q_C;v1Vz1V!l9aUZd}!
zw>>K${EN*sRgR?G(n*}d?QO_;82C9IpQ#|It<$_VN?Gl1QU~+Ry%un)aj6(7DmP%N
z(8F1$v;q4W^(*_g6bMH@HVlm4#49P9@<qhFgDW1CSeZipyKu@50-aTVz;F<Q0t_!-
z{&q8h<0-EV4}|PZZr#Oc+o@teYdR;UKN`x>k+ZIEh3WXw_ERyK&!oM$;P@FImU4G;
z5&*^%-#1E#E>ylKj;VBY-4&Nut$1dlUZUxP=?J2i1xjZRqNH3WJ592<W=f~o;o4S-
z_mVLd70m^c$2WOx!V6nDEDg!_#gh9li$Av&8y{L2rlcH9WG_`db?3C0+A!u8R>H&i
zs>=Ed4BpZPyol!tFij1)b#OVFGefltCy;C#rH{!;Xy+rJ&<Tq)E8rJ1S|;Y(Bj?*!
zZGh~&8~O3$!E=(hA%o!xG53>pQ~XAQrH8Nd<z%)N$3V68mU$^H%|$OopCb~sn1RWs
zl-rjBj??OQY?j=bYWi`aV|XY>!&$f2eKl&+P_N(Il>hA9Bf{XD=$20$guRj!2tZNK
zENO=K#;%VfE2`K%DHqreMR~t9d0NFo6xn>M`!#rQDahu?)Z?4<>Nfd!>1OlKZ7vOu
z`KElq38PyNVE*$Dd&K?z@^Dnk{b$$CNAnN0oCg0CMlcntr`jwfCro};tGsQbq)6f0
zQ?=9%0}L!9{RED@n6SiABbn4kNf|XP@p}z%GpEF3@EO)>%bQyGVRI{LFi+&M5CBa_
zQ34qniR@K{9;qZ|wxH*=!l1g0ipkL{#~J0RVym&3!;H7y)Yr%h9Rdu$0<(4%AfdTs
z>3+7dfy`jURf*^@nU}a_nx7_ULq0KBb5QOL5s1lCLhSDEn8*9zD9ssBs`-cXHum(f
zM18K-X7D*r-GABfu)LaIIwfw@M(o#>I^Sf$dhWV*aLY?l3}0zOrjd+Wsrw9m8DHVW
zm%U}eAo)X4Ww{C3NW>W9QqkN?zU{b-uYo;)Z9KKG{t8+iqMA8CDntvm5SwlLGc+Hd
zJLEyo*gm+ExobC?7)K~VE~&()9xrKVa?03^o)Wf7D1J&342C)#PO;=2Y2%ztsZs9&
znxli8hVq$a{4z{sTK1Ge-@b884LU*1%}dWx2+tmW8ALiSG^wpTfcw-o$n{L%Qxsu3
zhLOv$3~7mnDj<&D*lM6a9X}~ogn=f^4$=TOOxlbRScd^_XaSyD5h>3$n$JB2DS~2k
zGh)c%%JsllXW^6$+v|=CNQRyH1&D4pZ`fdCGR%GEj&xNJeEs7|96Vrmre@A(o|SRX
z=IM#11cRxAa!)10m(;!rWa8?&u3v!@F}(n*-wX8$F?(g3T=HGVmq{KfV|?C-xPeAw
zwNK+3;y+B@={A!S1MkC1l-@;9=ZZ*Z!oPjL4jtDnY;Uk)FkXWAB7>_78A{Sijw~$`
zy8fZ5AkHiacZ19ppvcx;WElp-+<5=sm85g5iU37tg*?u_&7e|%e5<NKbe8s0?s&4|
zcQ$U9D_2V1e*zfA8C0b$fv5;7wh~F7owVWYT1M#m`KO63wFjOYHz4acToG{--C}9J
zda`s0d~040U@&8%|HE806uwKyfdSAExB>8hEwEUA<6iI}I*Hwuk6e&z`$0Gv`}_(A
z4BbAy-@}*_6z905O?&yJ`F%%yJ(Cxc4jt#?+&z}f-sA6Tt<rJ)m6w4T3|rg~qVRZy
zBAX!!Jd)y+tu@UfbQyY+LuRs>Gl3?A=ws|uzqU2GfMQ=Dbv-+?25FhjK%zl<lFvFT
zW1`8e(JXMBdDuBhzwz@}E}X0n1ht~_49+1yy9wKCO*f?O3!=ow*EWk5ToYu%R=T!d
z2W!lM0TorD>7YpVCc$&nw095~3NJER_$^9=&|r&oNx?R@B<rrJ%C<+^22%6io!>3S
zoCy;)w)5y2T_Q(`7SjJ?4!yfs@xhWz60<XWS%3b0F5#cZcUm;g*wJw%X_b>=P_}nt
z9WWfh5O}q{Na0%0hVwFeaU!=HE2ud*N$4@Fe1pS_YTMIoF9a&5NnRM?F}Ssa>z8m_
zVk%|S7W+gT`n>t<%pN&SDM^A)IfXsAfQ-m!$wF1z5(w0pKV5i`eQkjDzz_IjQ~~~5
z{Jrg=xZyDhZ*=2DzSmJDxqQ3gKy5NB8bnO%T@b4FZT1k5ws3B&ncNl{Y`C>e3!ofX
zSl(lqWe3J`U!T}V4wxG@_=CYXL>^U2q~Es~v!1iZ$99;|0F?0U$3Hmt*(C<pKvnuk
zk^AsTl}iJ3J@SJG-jTJ3+-Gz)cVhogf|?6-nek?lE<~X(&%}by!wlTFbr`WYQSKck
z1uXwAepAidA7!(`(nK!U$;!V#0q~6y9#4g;CvswU`kMb$<lfAkut-g=GXYXG@ArU(
z%=!xkQ6*)&Q;e7YwDEX}_dC)d^#WEpspOTfy%E-dI~YtrqTO;mRCQLkNHEJ=l<2E2
z;60+6BWo`nV|gR02@+sZz+pomDGt1mr2g6i1%EvEtW2N3>caDTwpObG&%ty}_qyLF
zuK+(xDj?SxR%-qs6rHpaIX?AC*6Z%}59*g~xwvD$c}9}ptho}E#XutHX#kKh1QZ2T
zi+AgZuPI0*XYVzj`OJ&H0pJ1!_3?tf$4B8AIcpZ*GIDi)Lc2%`u}3G=CDi{UaeBGB
zxk+={->~QX)s(g{qyiJ8VZ~pf|LP>S!st^k4?L!mvCb0GKVJwB@&*I7hrybOy7zXa
zgwKg_8W2(N%#riu`oRb<{?;^aGhFank+d4ifZgi0HF3qWbx(UsJ9AX8H${+&aJELm
z%!56ft3e51;N`v1`QR|2s9PR;6Wk11qbH-Al9yUi>AwxM7!!|vjz>OV81l65-^pP8
zx#{Jf$RC<z7m_vAhr<g-@sl08@`=UC3Jxt<ImK8|sM(IWzOE$FEz?^!%SFpdJag1<
z8%)4<7xkJxT_hr42H0uTQ|kRqT3k|2Vwj>3@M%#A3|@Vw)e@KWWE+=E3};n>JNMfs
zkdy@>(IMAVB6voft7v;fuCe35?y}y0x?bk2<niS>aLfT$u}O7={Gn7w87UA=04D=h
zcgc$b?>8PR;`COydsdDUhZX$$;uQi6FlSF!t3z*fKxb~d`zg0<JKhL<xYQpbxtgtR
zm710IPF%0w&10SQ84>#WJFo}%ubb`yWQ}^zU$p@X#!gmRdYnG|v8S7*v+Ro?L$7!4
z4ll=PH37Cux*li%DfM#O6{!*A-eygcpVWx0uG^wV3>1vGSvHkJt-YQ!I~lm=!}Fg^
zdED$RdL2uQoC1<fT0DGgWl?xBQzf;f5rE`HwegBTGQB(#HMbx|4aj(g*EX{8eJB0g
zE-we}zN~m|y7T`}Apl;sFK3~;sl(YN1*V*6Xz-D-$!eFlhhM+^*J3*9syvZ+0&<=3
z;>yUapV!S4f~5>-AX`SZ=NbE5QY+?$0{M8hO9~3+p>;bBW<?M`HNCeZ)EntS0i+s@
z?ZSB*C`~^=a)4=tK9J`Q-Zz))21f)gm>rm}??wrIL{ZqSL@Z0+!6Cd(U9}E+Ww-ab
z%5#c@c93ORopA9I%d?}j<jSlSMh~m?!{uHf+S4uu1OE>2d5O+a8~njGOCXXlX%t-d
zbll#2p#Wrz6NLZiQx5w5e@@+o8Q44_#o>$n9%JV>jm^AG^YTe=SG4An1Os*hw7-9|
zS6XP)^1wy6BWk(rq=3b+>Y8MX>>Q)YXSWzo7L3}BgkG9j{mlhT3It}E7_yNR0gMs;
zfJTlH*mSNWf^jCxujizBEK^%EIpsUA>L*n3x;$DZdvXTMNGlAVr8nAk=46mtvd#1W
z3knh>iK?`&oFHJdZA_R1d1^sR>cNM1WM}TF!J+wz@VoDezRNqo&)oG-?Yiw`V?88<
zj8fS4?h#m9yabQu&>|Ru#m@d-lRexh{3QF_7T6Pk6?AL?B-=-dho3;?U<)C0Kj^ko
zx0?bURC=Rn3*bZrQ>ypR7?6@rt(3^#i(gvH20E;5lZ%6C7i88POs%nDd?wj|-ZPN=
zai4yE@kMO4G(G=<q5f9Kk@fT_oQ~@L*3=P+3B5Io#T_$V^8p;{h>IPm$paWhL(mF)
z7GmoQgC=TqU~ec~$-xRr>}o5@$aXh4W+*M=v~)EF7|gPXPY@RVlTk!qx1b~AP>B1U
zzR6wh`8q3f@rL(epq2Z%X?aa~whJ+kGvzb9_xO{gbDcrAZXS%~{Z4${%R`*)ep~m1
zG}x@V#_ln0*6se<Xx|dJ&jLy=NProCZC0HB?oXHgNuA54A-(|zNrSVdSQ0gMDGFSI
zIX3nTO<8k&=*e6)qJ)nBG<}SJ0fr#UvgZlG_!{o-zxfJ^y7iO+&FGnl?82hd+Qk{*
zGc!tx3<D2w2)cJsk)tVMHe9LW<@@piHz8%RlDbL+-#?>{x|&VSf;N2j^o?wWGq7=t
zSe#!qxTm_Va^^8FiUXSF6RAuKPo{@doB~W$Li;@AHS!1dN4a~Yf*yl1uf8fAF$7nt
zD*iYYF4+0aem1As&4vVUBwdLs6hrWwK!SdBTD3ir+`Ci%5b>SK(UT4U*cmWvrY?=#
z<Cfrv&kQ%u;;oRqoJX*&^;1@Bi3RSdd$*Fg1~CvQAkQBcF;DX+JDp%P1yyIJ<jk-T
zNQM+ZP)O_?QaiK3to4QKw9F)9u-Ll}nh%vCQc&`z542<*iwG8?{<Ok!PgUJ(eDOqd
z`WhFFY$W}k4D}V+33}MxT%L;a<<(T!VS^GbpC0=;8E#?{7yLT6WDKQ8UCE^jIU~6a
z0@lb~qF0K3xgdu(F@z-WK{p>w4uvyU3rYgQM4TIP%(!-%)zo^(L)%^$La@?f;e2>8
zqj2!VNT^VKeq1Stz0dQ(F*il+(nC0?b0+2e95wD9Yb^E8bjDNUFOe$EY!vs-W^GK&
ze!Z1!443nw!{7p0p0HTsY*6=R+v@4k`Snq+jz>j=d_)NUn`sZF3g5lIGI5Ug?nlT=
z=X{KUM<hBC(SAL=`lyND#PV@DsSx!8gJ=|Sklshpe0!av6yJ&$TlqZ9`Y0386J6#w
z^?$tp)1MkbOq`m~zRvD$^qxuHLrO|u=q2HF8J}VlHUB|rxp`JKirS}stEZ3s;6Ie+
zQawm7vE-Dly^jy&zf7&J_5;gH8^%`tnHepgO}l_6E^2uEKH%*#bapq83%jMKnx(I5
zg0dYiiqYD#8z*G&W5_7a*69ltI~uv>u%`K;5v^NVs*#6XZhkFc!JC(d6c1OSw@Gs?
z9LG&wsgt{QQLUSjNmkVzS-PAD(jT7dbK%M?aTM5%olg1Z)*IeT#*+$o@UDG({47?P
z*RLz%Fc9lUueEh~sy1!mYCTEnM=EYtH#4PLldCI}#?m+T%cc|Og|x&<#zBiA0_*S!
zwwp4+O{fKj`U<1)6`iBeY@mGeUTF8nHRI#l52*5b%02QP!jx#89lLLMNXGa{@~S3n
z^;hOZQwuAb;UgL`Gr6a+BT?Hs@$_=0`RgKoM`Oew!u&Z%n4H4uY08SSqp-t^X!sdY
zZZ{XcYov>1E2gV3D8ARGgkETP)n6ze=E2q^<hCbWn-8!~E}Gs4G)KjqS=_1qmc#LT
zz;z~WH<dgv-<eIkI<4_8+ql|@#FBo2m`oK4r%V7e4=<70?#=abOkkMX82OG}Mq>p3
zT#oi`t#a{H>0pPGVs?7r>+H0c?HC4)j99v2*!yQLOcotn&&1!F<xTkvZSt?S;MYrb
z<bNs-;uA_%nJSQm{6w`-7#D>b3=z6CZ=mPb_yEVDfrsSDrhlal^Qq#qs%jLlZAkWo
z#L~$4Netx_SO3J4E>3+-=BASxD`M@1_44XqPau#-8pjx;t)LbskJIoU$s5zNEF}60
z%N}-Dk7I?QnceDnoz8`EY&*U7((N)eJTf6!ny=qDo=M>Gebk~Lo@Kt!AA5nqiFg{(
zY~xVY^@!@n;ep{}S5^IUSikgmmfsmlwiJ%-_0CY5bF+JZM#5%2!_O7&XSM3dMcX=8
z6>svzRnCs*R|&oMIg17-0Y9edUFNLFW{)u_oNvu8Pd_d_?JcGZ;*yJU&tIgsKEJrs
zz(Ox~-Z--D$$e*l!e=xrzc3t$H?y?aXi@O$<bAlMrqF&Q*hQ-?tFc8FV7^#Ac14B>
zTdaY&%80L*Pz6stI;reswloLbZq7+2t@lIiS6Sb5#8D5`MGWn}_8tnEe+-vzjJhh?
zy~itjg@Qm$QI8$1{R`RDq55<;gyIzygFj`F$4}zRrq0EZyJX#9ER^;kk~hiz!Q^D~
z4+L%wEaqXj#lmrMXN8f2%a2vyV5wcSi2E^ymYQg{HQ_7!Q-9PKtiLj;^Q7=<K45&P
z*f%<eyKsCf<}S~=TQ;evT}9+zOOo(4ZB(aw^Hkq~pTi=wooAebgC_cK7Wlkk@$ypF
zqJCKizFsiRYkew`d9^Uj#vI@9`4oW%JTr{;w2|mV#g0xHPvRw9e^onK{-BsP-i8Y|
zt*S9{ItGIs{(Bu9sc^~3YXACyad}HUg8K9Bo??-O=KBYJ=kizj@~uO;6C|h@i4x+S
z@~`pO1b0ccBPVvMj)$5L@vJfSy+lZg8Y-+@=JzG&Sm^}yMO)GQQNM9;bqM_92tp6t
z$mvnzWMsk2?D{RlY5jjF`|GGIx9xo#Mg$2JknVg)0ZHjbx;vyq8fm0Mx{>admXhx7
zl9n#%?uOrb_SxrrzJI-Aykp27jxF5W&suZMHRm<2Ij`)NlBmvV7Yd?en+Tcd&*)T(
z4r_WWUrLN7<G!>nX)vwWSJ%(o90auL?^T^W87~anUu)9H<Gqpo)e}P;J;b0~mOHID
zs#o*m>!BKWa^t>u2mAs&Da*|`tL|Pc!Ba;qziVR3a%<$4Q?pQGv0>W@N-RWDH20&F
z`AD-y2i5vnL*cu5DWmp$zDo=X*hY!yuGsQdmuf^EJZ)ny432qqQL$>w(S1H4c$7}x
z=3F<W!HI|X;26p(w5g*Oq&Mup(ba4S+bZ=;-70k@%oG>?qZbwY^2CJ)q(?iooPh%a
z_ZNS3zTE2{F7ho!j>PSsDJeG*i|F%4^@o0dr9q^SS@`lghEv)hvJkT}J0Bx+x+3S3
zfHH;H%p_mdlTVHbI^m+D{nlB-jorGnSt(x-7sNwwOH;WHudz5>PwyN?)1WYa(b{Yi
z9i5)}-e97VxBf=+(@+#dd;Lah+K4kblvz`mNo!5st!Z>CC$qxOMg%b`kLf+8wW0J5
zBUNC2j`9wZ>G{VrPo${4@@L)4dYj+R56fh4DzLx~h!sx1N%_hiE$ou5*_4Qa%=C^i
zkO;J;TGA1zWUv^l41Ae#_RvMA_mzhwDxsu_#Sarg@<o-wRcpUZXQh+TqM#-q`7AC+
zUY2Fkwl$ez$OPA8V9`fG=TbqEO^8pNk?^{EJ_(OKZgNGmH5Cx>yk{naH16FMDb>-$
zo<+e$poJ^R+9~jM4;xVsg{)8`=O(xPmeJacKud7ZaXLG%G;?w~9;=v=3GMrAcRd@^
zyAu`T{FqFwVeiiBR)li;p&#WEL=IF$G8t?~VZGB;Q{c4MQLQUs=YMRAIQ4M%ynm^c
z_SKyCLS=sM&&P()`8-C-ebH=A6(j?P-3?By5MUc2OSm5-hAeq?YiVBDck~mKt!>&x
z!J$0YjK7)P6|gZI@HsR~3HHQAs^Y-buYv>>W*RDd)SvO4r(JC($jA?bRJPtCGRSFP
zRgFEez~?I|R}n`RzmDr!Nzi!s;&QlE#r=g|D*Y(otp8VJ^4e>#h_Jp0m)}*;{IbXn
zYs3^jDbSN_m2^+<wEbj{aNzVIf%BKms^5f(h%Z&Fe{|j(L^ZRou4FqKI>3-KdR=Za
z5nEhuHZZkc#M!2-t@qamCsohW&#fchaPrc_up-xi108)fDh4KzAY~#)M^e(BE7JOw
z=V{@J5e8YZy^QgmKjy9u37_6yNwz2T$lm#kghrr-u$hv$Y#(DnNFEUdv7>-lVO)uM
zJWB?AEP`%^ji$OS21e<6#55`CYtqn;IvteMFOJz)v$$B9(<fb&s2E1Oi*H$|?U0WM
zwb~zxI_YJ0z4UltEUwwiL#D@6nls9spCFo#HPsFr{nO5N<kYCQA1^lM3mFz=Ox*aU
z`v_xqVFW~J5IgjP#bYQ;zYaOTyvMdCSe1bDVh;RFN;}CmL(IL`!av)aqUYI{+%cSr
z!x1BCV7BEjtT{Gu{Z#|A@EKK&+WwaRqILSEt6uEBzMV5S*$az|I|y9&Y3*CD-VDYE
zc!9<h{xE#L*7<97-Pt>oyCRhJ)4CDbo{1<3e6X@Lf@J$xniql~Wwc<Xg^vCwg;$ir
z8x%wXehSOee;%&hSwuwRr*dv*>wGEEVdZ=EEk7fhA<T-l9ci1rIVPS{8q?Jg#;tht
zvGLIhK9PJZ#nANU+<GM1!^qZ|;)L4h-|}-I`FVP-Q(~j~r|$J%7?{jyev%xwv{FK7
zFQg&gC}<csgAomBicdL$)=<~|lyb?KAZa~C&CUxZjCH@jzP^npp`s!E&Xb?v7QBz*
zM)}Q(0QIkX+Mv?APwE@bYGMx$tC!OUd?#z!e=(g5DT_JMF(3(5fS6_TDwLR<50W$*
z>D3;SlChCB6?r!Ih<fZix^QGo8W!)YWU>&NP@3}EI^QdqH4@<@USmM(ZTC^2!OCev
z<CE?|!Me|NJIA?|1ueYfH!_=JRWbju4`j57BpjhSSP0j{w*qHNtC40-j--Qjtr4|i
z#uHX$U9)NeT~R0>J=`cjykWDxqLut71=;N9LO7)*x-vGe@c88XZL^FpI0n?j^^v5K
z(RGfyN{>wkB$n@%d%-P0K@`1$U|@S@R>&hGJErw%yo~?AQ!m!CH*@Ul>Me;#3GY=6
zqAWXtpb{G9`=mW#mRW7bpw<KTxIb+YYx1<VPm)Vou#U#Y;_^pV_2A0Azg$Bov}Y>^
z?w5o_yK~^)aguxYW(R(mZ!7UBVJup3VE@iwDyK^2@NXsq$?Vn#rGg-dk3qtP?HiU^
zE$#&T83-A@lKR^a^!$YorQ8#3!W`LbO6=NEk+|mgKWzn3C?+3y*pXr(!DnkP1pFt)
zK87vf1`Bg}wu)2#n$On@oS2uida&i^{1pNM&5mqVd!6Zxf8vfF9vXK#EEkdqEicTp
z)eeKO+x_=(4yuf@HR*SVHKTiZ^jEDQ)af$P5h(SNf7ziyDF!S4=jXPWAHL7c-H6vo
z@9<~^i*=gwu0T)H&we<6`P>=@)|#Fc=^&0OcH6r58E=(vQ0Sh)nU)9y%qIIt8MbA!
zp8e?-5Xx0JIlol5shjgIT_f+sN3JKe!cKnHg7+2z|0a(q-k5Q9BX=+nGhXZjO-AuN
z%~|;8MB;JXe3a3<!x6QCNQV6LJ~tJd(w=p@OipH<s_I8XRsW0|zKAghS;5<2l9006
z-!bzuad3=M$)hAVPlUsyUW{pg$E+2?tQ9lbQprTMqoQ?~VUOe0ddlqiwcX=mf+HT0
z6*bgdfpf}+P!ySZsS!TnGlCXS!t8}hahTER=+c!;fuM?8__wc-bAC5SvK7BtY?2qv
zZ%iT#E`ljnxIb;8Z0o6}rEjT7WWR)4R7ab_X)DO(%uw^=rX#186fOTbcFdN*LNUqs
z?OJKNCr+A?`R1(i0ugu`B67y~L$a#%>Sdwc`qaZwMXAl(na}i_H#%iGiba~qQ%`HG
z9#N2BQ)M`#5>;PrdA$W{u}|pZWuILwQKEl{Ue@neS_@}QlG4V49ixe*`<xNj@w`k~
zy(aE^!7N-N`91NdDVh<oc!`SgxKcap?Q*OZZe7f>mwNs?@&VrKqx_n9B^crC?pKi~
za?udGY^oRx_N3gT1gK+wci<B0%U&-dpy*fpZu7gt?qJR8g+sQASs?S;#MzBICPzZX
znOh9~ASRJ}O#Ga$U(XQ6@{<o!NN`k}qQ3FD%HwjFcD>O+wG6WKdT2D!gRVMhm>gaB
zH{gd>Y`I*qf=_One!iRQIq#zve%<cdAZLFdBuK@?R@}Rnkk0_&XG-?17U!C_NsW$2
zF72<Ie5Y#4F6DUjx>BHC?kDRc4Lc1c9W;HJMAHuuEUwKSg9k779RiP>8}Mf+S0XyP
ziEngbc5MI3S?qbPu;fSmSQ-@8S8{R)Q`cs@#wjJR{z{=EOl%Bleq)mP<93p#?&u5^
zDsvX2b=LY4FMjQJk16rWjHT&FO2f;^Q`1iSRUYoy)hKC}hm2E@D5PI8A3`YuDQ1Oy
z5)$XBtc-zZa0Y*&A6}4*RwLE!8H`x?Hwwp_*jlv<Q^t2;e0eKuRkvu#MOPQj_^x$~
zk$Dl=ElKthnIlJME&#j!rbZI7w9t6m^`o3k4a+jK+f2L+Wo0f*8o{wSW>O>;E<H73
zW_dT0+XK;zwe`c}Ar5EL4HB$O6kNdPELOj41#<<9FkE^h!xmawAIf+TSbaBg8S-Mx
zneB%t9OU!IMZ}}O+Qc0rvGHLs3$Lo58?BR9f0CfQ>{*8kxNdztTK`9>;bpF{OV1k_
z@7?SR(3y+!3pH-eyLVAcXNL;rRlM^j)>0wnKH4cC(pL<vtvnA0)?O5>tTFrh`-hQW
z{!zKx-B^xzJs4V*1m-T`=s3IqlyKL`^)xwZXrECMpWAybE-vT<A29yxnrBGyEwwp|
zd2X`$WwY<-7Qv2<pU`&*lC&u{cvJV5vLZ8VHf#zqW=Ak!kb7VCx`|QPaPvHgFqPg?
zaY?H?m-glFTQh?et-_wSclRw{rxy9pC!SJlawxx5X)cqoUAix%p_qrvDB4{Ug|xV*
z0Z75Xn(o*l3_Ox1;JrFe_`$k;m3@^>1`9f$2C)+l(Is64u?_5m1?)&N99=BF%rvIf
zE()r$iU~+ZD)LNUf`~fJc3Q$hgjls#;6u)9h~WroV$CtfkKud9D9O;^;kx)Lyk+d}
z?{EgMaHFHzl2}bndkC!0T^CnvYyef>Zm*>Fbk3qIDkWvGw&jAMqobp4|5s_X6qvRK
z7l=x-xgIGXmOpXwTLkswxD)jZF!WG96y(U2^9S+d>GTR*H4jGJy63$Tim@`d#-%^$
z%T3uXR|UD*Go26o<gy0alMR)G9>xbXrBrM2$AunN@-HRVtkoYQ<RS;l;wa^*%&amH
za!Cd~GEz~ve+zSoNyCCa6+7zdU%8H5betE72<I<2_GLdL`i_=7t(ivY@6e4XEGT)-
zJw|k_)HmtyF(GIdAZ6wrB9da@y5H`8@3=Nr*sE}yyIJ4dN+jZYsCvCIOtz!)cVpGA
zCUZ<PfNLO~e?UhzKmRPOQtWd^LP%4jV&Uu0X;=DJu1+!RNHOeYb;4M1?Wh5c=>xk{
z#fmhT?niFRBE?fI9=8rwb*nF+6T&hrFm1pDu6r=448qHs+SVpW?0IeQ-fBTYHSX!9
zVwf5Z<<N&hg;S&9u_%-exIbcKvx!P{K9GZK3*0gYs;+ksH)S%q@RKkOl+_-EHTp5n
zG_e9ThegPpo78syPA2uJfI_rzuozCz2wF%xdZ3m?&!X9xNe~732zt05gw79zxpp*M
zSSj<g(>HuG`MK!@+4l3%=7wyhnxv{1`=x+;O|*8iXkA`nX}x?0??@>Zu^FX<vwh++
zs{)!Hm?uWBA14*+Uz%t$jG3ul(m7KhcKl(1NUp4Q1}5=1F}}oM{M#$=$h@(k!||>+
zf{5bn+n4@p#e;I`yh(<=5ziEql!Q%8sFm~FRd<-UxRSiRVeIVf-E?}5E5XpUvGtt1
zJYyRET!UQYy*^}F1Fxy=L~`;}o^VFA@DbtpV94>{+{gVil+ErD+)0;|BGlg>-Owz$
z&JFaazuNui+BkxgyKxtIO-B_}f2&J1^g_-BfZaG|PsMG^d^qCc>S^5r3-b0^Lj?!c
zClpB>3r^~CT>H;g%g?K+miVf&M<0L_qEX=n5J@MNRt7+1%N@BBNt|xY(@P_a=}y(N
zS;f!@lUT#XK;DjdpAN^U`O6OsaP&KvuRR0fAv0Cz>hBI4%`7ZBS8eoJl3e$#->2Ae
z62dHcJs$B>#T2^D7eZ;spx143eZ6Mh>g+0?jeOiErm%cw<-~OaT<k@7^hkg3Pi>sI
zosWL^Cb9)oGByZcXkIF6mLq==WQvkZxMm)I<VyN+;|BS??9cV;TLl1(E;S;#)P}>-
zdsZ8GPw~Yr#cX-)Zx%dVU=;N#o@-uyLh3&kQR&cNXCTL@RvO;l-ual+5@aq;gka$+
zFcc4gH@2WF33s7D+fY~AC=;ipm3YbZ5Ia)T!G7&-6`i*f@wYb@#)mZQRV(#3z`Yd}
zcajx&iHf>)(o2%x(vtpgcV)<i??!3=KIL^zg~hYeeu0CpZjm1QU)`>1lowsL(tc{!
zmiG2)TQ_eU$(xCa;)+BWCE<%dNCc9<Qqowsapsdl`CIa%E;n_*uirP>PEDr@C808@
zU~*8w?3HI&m6sfq*&_)OZ!P<`e8XRPD`)j)77S;DH)7-C7#0p)m?v55m#)84*H!+;
ze4YjTMBS=eg(0#}d@G{Z1*~V|?DD6+F3W#UC;f3JpIOo@O>ed2WkME4us@DO(fX(_
zPxqx4`O-_)qge<p55A;(sWzrjo$`Hq77iB>1cimr0y@5zf%p_jGWxqw5}1q@LCiOo
zv%3*RA>P-wP3IZ?{kykX@y79LS)LF!Hum_~*jV#u;gB&aGcz+EKRJ1i_{8?)cTt##
z>8(f`NQXWi0}{CC?CR_UU}n+`JHLX>jc-id?U=&#jYdLFvAgsQ5WGqj6o@QO`-s8$
zhu|az06>Br%pDUy@7oL5zgvczcuhq`p%^XLx#;*3`<M3JC4#hOI%vl;L<)EC6+Rz+
zE`EtrXY8>|Nb;>GsZn{V2G5cbcj~I6+f$#^XU)yiDt`~FqkYrtJZFl4Hs?>{rTClM
za6-6rp7OBjC0ZYpl`&w#!^5F$Uoh$MS7l}W`MC6nYcq7TXx&Ff6k^^!Xwx7=?E<W~
zT)yGXcrNeR5bNaB)Y#nIz%Qf0y5o+Q!4c^id}rk027cQm`xtw$z5XlPnwVcX5vTEk
z*k13`7J>NuZ{Ho^pJSSf4Q-@X{Sfx8^fxIHtmhL#4>52>ldd4dn3GdcB-&#P%8)g%
zr=1_T40E7akDlM-+PXv|RwwAadJ1j$#Q+u)J&E^zqRM}aVF6cI$BiPLUuzF8v0)k&
zo5VdPASA7mo7{BIp}fMGhc!D74M1`B8Q|wgB_Xnn&vT`CEGxiqt^&uoRHMS@ZE2kk
zn2cEg=3R30@VHz|sWan;<A#grdfphl|1#GF<~H7duH;cLJ+XHqL2t?9{+;K^OE4|7
z{&e82W}OY0R!O|T!;btTiv45^BsjINf4`Vt##3ZWB9P)$G$mUocZQEmKY|??i8}Br
zvg7$vc7+|z3;S0Uc+!$1Ef%6F$0$<{3ZubM9~=czx~~G;2}K%9#;vyqBoiRR)C*`G
z`S1h)!?xY|C=SarPrzAqgQT?=sUMyyG+s6T!Phw};j;xV$Q;+hDVu9(G3j7M8mEI3
zh#$N=Q!;*38#qrR#-|o=^Rj0jvH#R{LU@hUSS}_blJk-<J`q>|vF8!SHE66(wRKZd
z!#--cJ2b1^bKoX++h>M;c?G`TT;ScnZ=qT2zg_#%jBRlz2SG=+tz@REs!ywOUX+=T
z05~HB@1sv8F~uOuhRfS>_Dzhas-W#y`xLQVYY}#$$qmOG6&g1cRATyOuH?MrhdQ2v
zg6h0V5~dI{6?|h+Sp%IhHFy3Dy;eEcs>U?j$=hc{>z2h^ebYwXGHa9b6=E_3$t!-$
z)&Q$OM-Mc&??;q9tmZSDJE-IN!bwyeOX!&};1dUj(%CSMFHQu*ntiLQtDxROlu;W~
zehn`|97>AVDZTI^?2}l@H_5C*%==M^mH6?+`|XM4)5h2G6oaRTAE;l6N=nZFVum42
zH6=mRT_XB<D*$Db^lGr|ZMYTt^R1V77Vi0osw$e3oKQ;i4NtySZE+GEImlg_*;mD0
zUo=rrlF`J=m4)tNa9H^~Rn^ar6+gdiZXMmq;<bh+e};M6S-|`q6SFs}p=IXgEhb&n
zGc{NdRwVhlAQU%tMw7koz18`rq*<<S$y_bUz7SpRVp#l+VM;U%;Kc?WDM12Oy#D<A
zX8<%>(#;|Xf>he%>wS#58<F_lOHPgX#=liex0oDGCDMa(+CGgFU#AQCYkJ&^kZ1H?
zNs)8il`66~%dGeEf1A4GCf0p`d*ZvL1u)kP%LC7x8Df}-QbCVAy6KLTxjerL+I{}w
z)HmIUCrJ7JHo_nLh&Sl0>JV??R>+HPC97ioJ<rV*mCQX4eHxc0Vw9N;RvxNGb@@RQ
zx7RL>83>Z(jAspG;UZ{$QOvG~^#1%h$>Z$shF?~VVd6<12>pMo%ocmfgTN3Kr3^FY
zo1vhsJzo1Wp)S#lOeF%-QZAc05A8jB+CNvw_&TJL%145$z3g_6nV0JlE8_JThI<aj
zpt!K&Z}Hexo%Bd_qC%sx{VI!dzNfcOQ$`;`=U;S}<zsXBQJ16mbmlCBqS*P%69OBL
zHuGiWhLqtHF@EvoCuBKF{=GdB!a^op(WK4_C+>2`%R@K!A{b=|ewzR0oUaJJynhLP
z<fzCm_LkngU;8=RD+C~jSF%6EVgN<RRD2z$1mSlt!qZdoKX&p2sww=|80v{W7z!UI
zn?xI$+DSMrdpH7ltI#KfeThJBD$b8z99%!b?fC{=G<+B~0$>}+dlP@sg8cq&p{$xe
zVg2q2#Bidw`*#W!Zdbj8W{Y4{+zi~;x-?6M<-zXYEoLtF?dg{7R|Zwp5s^`>al(?8
z`j9}2NU}k@s|mNVQR$RoqMT{cX{0L=d-1oCrCvpisKr`^R6T_KKG(A*eADYz`S*X2
z9Rh5WezkH<W+AUj#1<C&tLB#A+nes1UBy7)0lsK1^&Vx%bu)XjG#)E|NdTn!gv(bf
zBQ@OURLc(8QxT+y>roQ0z1}yrmCjqgk#*<J-owoNc=w_~s;Bj2!)sQzMIG|(Va2BU
zBn?ZTcf-+Mv4a=g_frxUUpSh`Q8n75%FMI`Fx~Ecoo;=Ql_guWc5PS+<?PKS*4AH2
z!zkGu;#J4}VGD_w64LSI9t9AlgDbaYEblUjL0bF>nwb5tkh?ApvOpOZA<_f|)rrBO
z$-UM3L?WA*4=gN~g1sL&5cY=l&iYGy<ZpOG<zK#6STEQr-6yDH+YpY8&Yp3oYx$EB
z%Gq*d<j(J;9=D|UUh3UH;3N$MKF~Ze$|V4Twk-u4Iy0eS?|ppza@*3nZCb;d{;u3l
zmiYY6mwxc)+Xdl*C8-~@Kkur?DUQabwZxn|gnveH`8EWQBIcwxBM?0A2qOb^2U<(I
z?#wYKYImxVC`^kEGo0j?d|2<}zLpuzgV;Kpa{gxK@#^OC=X^cqP7_(9m7=ox^OdmQ
zn3=^cyxuwWxE7zkUU~kaUpJ-gB#dS*%3m&MoPkTflK?ee_ryEAlW$``G%)TlHJ4$-
zqP%4SbFf8L{&;EFt^X}9aGb0xaknn}!ZUDXYk$KZ?E|K#uv#Cm4@<*Djt&N_9Snsf
z4`7+$T;0`;T)sfQe0=`5RG29*^G78K1Tmk<A{^jtw3gg%%GWT4oUY1j;c<;-wW_~W
zn#<g@MK%~&Hr+`il~T`KPXNk<$rfj`H<`PL8ooPNJFeuE2$n5avT-eS<Sob*$fDbr
zKPhS8`^)AoFzrai=CdJ<zLH&CH?**E<sJ*BDmD2mP=T_(Ge|<&fhh)XzsH6LEZsj3
zFLuZzMN=pO{x)<EZ-C9I_@)T6Evr~C`GW3LfQ-i7+-c%!z(;<^vW%M3eDizP>N_ca
zbpIc?lgVwvi_MB1|K$SYHiCMj{mePREjuhL?kz@mBG3NRIysUGT`aH7QuBgkYXCqx
zptamm)3<t~4XKf5)2E=6g2|q6a>g!iCpqqJEGkb)ZFl<JE3G#K6xU20=U%YEE>4wh
z7=pa;=b0yO5y?1Wj&>|p>UjKP$<FKGBw5YFKnKnuP;m;Ds*kOiwxc^$EB87WeTJlL
zgDfUZ9?({&%RouUKna|R%=DzLz!NAf0<=t88JqO3Nd>DkFVC3|6iCMk6DSZnuBpKq
zuI*{6GjUJ_#;*7gmzf0QmGfTN^S}MmH$Fs3fw6$Q%Fk?<t5l>YflTF}dlCwNbv7|2
zUtai`x>lz7t3;_b@I_%<zZ#43ekh<PIUsGgkSJbn5On3~`Il69&qcpZ^if8V$9;H6
zoRqrAjBcCi<#=QD%E<GcUO;S$L)IJ1Ur8h|$GESIJ(XO>Kr}V6$I)pd9A@GDfdz1S
zLa@3E57-u=HRUy?FHzI4o)QV(=6${SOhU5h8NlAwaUS&3lVr=3F|YcO1_0Hd-1jO3
z{O@nO6c6(p<S8KSNc4%8Om!uQh8R>qRxFEi3vXfgt*Ik6KkbLZ!-(eVe&gZ{Z#XRh
zna`?aF}z^bBuLSJ!)a6m$ai9<3R@j8f0AA0EV@$G`1K~C60eynU^3OMfkXz`hp8nP
zViSg_u;)tY4HMHj4i<Ug7B&hly}MtSYFBys1o&d~0@jRbt$F=ewggs6@|<3aEZZ#u
z3vM}!hc{Y#jaG<k4W!Zei(FLGcxIS%A6vdrfBQv$c60Ks&zgRXV4`)&P{pl4A{w@Y
zAJfF`;k%v%8z&e1xpt+Mp^kw%`t^8vtt4_t&YWWJNe?>M97lQibP2#j0^VU`x=<S7
zyxGK>4hE8~bU)uAn&>Qip_ZC1mprFkzRO^;pk9cgzv~})o_vaz;mCKQIpg#+zjwIu
zr?gcWfO!KxsxY_=!G68(U?hWkH`?}gGOPJ<v^$wU{*v*OCQz<2ABgk_lKMhQ^<dty
zwi#LURTSV~C1YN-w5d^Ag~p2G?$bD5a!X61m_&{x0@=)6ZY5x|EYbo>Cio#UmbP=I
zUu(WzIq^tFV?yK`eK#9A`>5Ag;CiwMDU@BYkSU8Jt=EdqrOUU6nrfb&OKA;&0^D;_
zNpy52r#tjGCHP-57WqWl`M#t&q&QSVL^aNi0Sa?kg8>JwIYPhK(?Q&Nwr+Bk^4e?j
zF)s4_&m&0OJ|W?zbq59vd`r%ONyXtY6zr|8j>ph{MZzkO(Xq|WXNO)!bnTsnbyv^(
zWmWU)NM0#V1>rmMou?K-#50}l^7GAY-8+{>A^>0CFnK*cjQ9PR+T!;kQ)7q>k4Dw7
z)ifT;RETb~UHKLJQd7GwNe4WN?jaf!B_?mpxlQ#XSAK`V0kTpx)poIK@ehTFS*nQ=
zIRY2~jT;L{;oDj1@~-cn!^OY(q2g)AN6A0Z4w4E%RuikK4o6zD8rz0!e&Y$7mENFd
z%!Gf!E#txe?4NtO-9d7qc(?@;Bbng7m3Q;Hn^70?CxOz$Jp~85;p)>$k_v*Pd5VAH
zyGRy5+!E;}WNt%l(DU7sgnE$WON|}9K};YMw?Q_PgJO=jsUc#}8r}A&RP@{x6Q5_1
z|63Lw;hzT-=1mD>qmN|M$S*tKsK6&A9YL1`&!v3$Xvwd5!A^wPQ1mmtS_P$_9d-I|
zwhE18RoH_MB)@^&k*pZ`J&catsGFUY@hX~{2GP$lI=|_#&^4y#?6maZpn@o<u4E5<
ze3OR_7iyX3nrLWjDeVRfm!k6Z>Vh?T=5)Az*W;C0yV&j=e!a(><8$tjsj;A$#RxJc
zs~~a=s$@}V%d6)dtJA-yMZxv2o%iuAx}Z?v3Sh^`K{P(mn$jPR32JTkuiH~dDwS0W
zim;>846@7*TtyRc;WbmW9c__F*2>=DbI8}fNYmn=YS}u^D;@YrbXkOy1P2-lA)o@&
zo;eLA*C;P9iH31S|7oLvt)>D+?==#)r(x91CBE^jO9i8{5BZbnjzp?@THIPLPsh|h
z0wJh%f^ymxGsCx{Pc3imsX-9G7^gSB`W5A_Grq*g4TjSTH^dI)iw(e$q$Xm-48Lq#
z*0<kb#D)yv?bwJ>*3L{=V4^idw*g#sFSKU$vOzQ2HA}Blee$E2viQrH0G-{0vv4Z&
z+y*!9EmL`zXKc@e0OT}xpbBcH1r?kdu+b&9rm^^+U@02)yg;z}b4V{x)8g;Q&P-(T
z(=vkE!wc^#E7z90YJ73NBf|Tj2X^r~&J%ZR%8W=UrEC=k55PR&UXX(ZV`;iLPA1Eo
zD|<E7@np1<aJZtX3he@b5YgASPqWOi&q{NOj~vY6AA0-DhcpOc(^N<5i486C^}1Mm
zKzGKsTHc8k#9(20x9|DfdvXOX`u_Lly9uAy+bUX{!Q=N7a?zOE;ZYc+oI|tc7apGT
zlesFh7MgLb>6;O1EdpDmx_hSL<EIOwTmn{9dm%N)%Q+d2tt~8GL~qY#(k|=vsC9&*
z9aogOzl6i%`l5@@y0=@ih$sW^e0?+~pM=E<xG+rgGjT>q_dVplD7tUegwTMASQW&X
zZ+011f*TJ|sqhtdig6-me2XHpNvXb!a;!iv>`s@fSG!uEAjUIzWiNEUkE9N`RlodX
zv&o+@xCyJXXXfz=V(Ey}pLBGW?yRS=zho6QC+U5>_L#i>C&K$4$Zk$G=bgzPnP7w$
zIT%3?DiBGaG6X8IC$6ay$dpoX5q*EV9sU#-yqyn`D(S^f5>dFu#2l!6T5`Qxv0LJ=
zBlO0!fL%^@A>f}C&n<b8EBvi@t9$)y?ZQQoF17%Rt|P1+YHimRWOq%R^LrJ3viQdp
zJ~rGSi?}mh?tl8?5tZS7>DI!!_w|>svQV_bTNRP_x4JXa-X#-4S~la~Q}xjNDn3BI
ztxKqx(n0e#!0fQqU1*cal!wbFKzcRq5%(h(uq>)AQQ0G0OXc*2!U||jN6E65bXyWY
zVj7CR^_tGO-3x@ODwsdBw+Ck=Q!aESlTS3^CrB`n;cN5KBWKh`0s_b%D|QV=P_K8l
zGK*~3S5ICBNi^Y3J%xRWCDE4ixTzs7qn&aEF(<$Z_Z@pd+4+^XmPOv8+P4zsjMFA8
z8SbOrg6d7(lk3kU=;)!!jQYltZ0Uc_Ibb|nx5+R=O3`n(SO%QlgTV8<hv~&b85JGk
z12c~?_tj{9Xu@SvH%7|i(u1pwAX=o}>4d%3o$Gt}@NO^>k^1&w0f3Tx<Vb|#fUtvM
zj%a*w@-7iSeN5+8XWOm;!bAb_zd5PoQt<E#-118BI`GOE&TVvdy&D9X>*;lN28uKH
z0-(Ey;&*nX9MEJ>Cj%%XztEMWTdzfOAk0ni{N#Zts-OQI!-Y6t`upV62h9prNcBg^
zap<Mky}tD(g<t}-Bdf*%<8@qPkW2%d^fji6Ip0AIG6;PQ${}dz>=+QFI4ya73F9jV
zDRFz|-mBISEempw>8tNXB;bb%xF7NS5p=0NF0WiK#Oz(UqoR;l{h0VzLh~8j;BtR#
z%fSSOlrWRPDF1E~e*?d23EYf<>tBfOpNkh@zhQs>85;tCP~DfgnS0M>k!lyX0>`fO
zUoSM!{DP$yo`q9@q@Ggp?B*&f@_O?$gNcDWRFA&SYZCJ-u9?n+kek=wTTEn6rmMq4
zx`eCI=1}XAx?s>c13g<PlY|Dw4o`PCAgezm2u;G_5ph2dPXheg9Hl+59379<0tkZE
zz^Aie%x>k@Uy@D-S$3?w6U-&n!_tu#j|cq#rxolQ%6ctCvs*$+1?6JK7;#N-g}EWr
z90I+i5ppf*lhrNkG@s~7!1~f($A;X?7xX>-xN`@m3j}NmMYpA{rmv(v&vh)rC|!+;
ze;y|n%-+4u|C177h$6RVE7REc9OfbTDF(0zbhjM2w71LECzPE00tV7x{JQnLmk52e
z3_Cu{9;&Y&D9@%u<h(`{E3QS^<`mj~S2^7m9#fWASD%{LCV?646A4Hn3MQLH=N~+q
zrz>vPWl*G#RLQi#GD?CYHWX)u1GO)&3xoFwhijk){7GsJ37j(Cok>4gP}~Gzk4E|(
zm*bc4=M5uEyCQ%?nn=X@o}fL_BpUK2?w226vtH({v#B2CFRBSO9uyO&OntDhsltqT
ze>P8noey9kV0Qqa0hPYb8FyWKx)2!2XMLoD9>mL6#SSDC6rI8BZh*7=dG@8Z8i|x>
z#|MoLYOoQ{`jiM1XjPyI;EXXNASRGYB7^t<;Hf{VnNcGLzx7U`2<n^mm|?WV+)DL;
zuus|Xo;3uuMb_?82bTe9aUR?iE5(pwh>12|9DMO{9S?Xb-~GLc=28dytL5rjg)pID
z4Pn+3S8m0kDgaVQD9vp7{t%3YfQqD{MPKih{RGH)O~bAqXkV51W#Xo{ZX&^0rlEMY
zY)a}196tl}9-O;ESJS|G$?BYlm!hkjLTGYmk}I4!ie`x<0-^QHvw+un0^(v|`QNeP
zxyQ4@IcsMY7pF4?0PRRhQV7tgGxsd%m*l=PelJ{2dj47w_XRMK@7tPypG$5#z#uqd
ziu>zTv`=z&H8feuvcn-8s|;jdbjl|Zw|{e)CyZb8mEw+<2Gjg}6Bm<?p)XUZSgvdD
zKuU|VQ{v1_JWD>TElt7|)RfP^F#{TiND2Fe)NcoKqO0MX7ha2xqoq-P$LD1uKNLP8
z4dd_Uqj&?B5uFhd*uwt#3(KoL;1oI45)cG0#;&+*?kRu|pbO7xoGTuSmQeWf*(b+1
zi$mOrx|~=Z0SdXsHI{P_<pbCkq|qbzp;CsBHe5-4Q)~mIm-nYPNWxx;)nf5Fx6+=~
zR(Rm*_2ocuZ;K6*+3b>`mcZ#d$y4GV(H^|oTIik(dTi~*WOsV_38JO|ItL-N9$df<
zRHn-f2^@KP->PvZ?pe7y3W4u_V}!hQ+pmspgzs)78aWzj9wtl6H7~L9q$Gh1$Ua)!
zw;6DUCEQ(r`tdQwwlL9-HNYVp&hded*OpAS8i9830nL{y0!F#(yWaAK<;M1<z`CPT
zc;3#dRZ)nps6yAdVuckmj~xquHcXtFdz2IS#w|og^vGkd4Z{2K^pB(+zB+`r&BR+}
zPff7b;NeF{OzCF<myuSYGar+`98CH1DEbQUf2(^WWai>voEHuN={B!22MGx&S1OR+
zt&uM2*4;ZyiAd0N-z8^lk5NX(RTieT{#s}n+iFunhZrCm6*=yDksj#__k^P#Z?uJP
zxl~=GcdV+UUdMEdgVGu)B8HE10zoeTW9}cLGfq#_z1A=5XO(JF7>pW`CX;S9e2!w!
z9IWKwP^XC3BU3WgL5h7`{d4f1@Z$H0GvD@<NqXL8D?K+W!^rew+Va*$?xMB9G=BY%
z{Oc|M)qI5liT;+9ocTwYv38EW<g1nMS3$*$AAM&Q<p5<15b`ns;pE))d9FIb@j+7U
zi5DOtW{rBvYe}GEl>n;dP{&X9EG3$;LI;iKu)$0L=my9<TNY*)YE*j1mBeL^@dw`y
z%DyKy_<>mV(?nVOfnDZwGD6j%TJALJ-gYTz>RaH`95oZ5^e6x@0#eJ_pPp9O@8+_Z
znT38Co{w<caq$3AC}*?pHyfSWG&eEq|2h(xBH0GB{2X2gw-tm|BOh=@Zl*p6z%+;z
zrqA7?m&j?WYwmWVqF%d$9Cu*J$;M^z{_eN<$5_tDOy+Tf(Sx&0{FMn0(DPs@F5P;$
zsL>Q|Z1Pqh?P<|dAPq~5`ipm5U07}`p-TdA%6>h!CORX90I{piP1)Hu{ZV254jfdZ
z5ZvZH36ox#t#2+@r}u{(f3#{4yvL^p_^;CN(<L)U{DhTA2U)<||17Vvx6+w-6tIA$
z^hXc3vb<k<<JZn&{z?XEbj7W*RY=H?Sv_B`<J7QNQ?CjuusGa{<Tj{y;@yt@W$>`s
z*4S&Wp1m`G;oQFo(gnmPDabdBJ`fc(d2}vcL#!FfNUl+PcOzroB(r4NvpH6Q1i{Gm
z6#gge!qCLL-YTGZqnjfVXBRy59@wLra?=PVlKFW|1=c&=h<p^-kz$l=$s9E+CoT(g
zLiGa5b!@Xy*%jx``r!jv4W<l^8}kq8a5d9gV8$y^)8W@xb3TBiy$Qy-yTBJ&ZK}f?
zibxle&Id8Cx02Ed-&pG6s#QK<fjjzZgN_WP^@2{P>@_P7ZHnTT9Dse6kOWwS^%Y!0
zVd){+8B^TyY6w_4xE-Qj71x>$g{?Vc8lnN`DUt0(i1t9gH9a|N)z#e7Mz5N(S@G)O
zJ^{kL=Q{u;fy5AL8?+WQc^p$b{(3M<*E;R<zD4qMK4*miWjEgsCYuiwZ|jd@^eRX_
zQMS5=A=E`j)m%C4J6fWtl7NEyTeadK<8H}FXBC@8Sh<eQ<IQT%@56_$YKQq#X6ucW
zhiDehdXXfR^uv>d99Qu_j$pGjv~yTnl?aG!oXCULh%A*o$>ht|J{|ejWOJ)IN=2?=
z4jJ?{pC|Eym;zWRd2?o%<!WODEIs3+H$x81b_BSObT_*%tv{D{9*7DHa}4D5BX>w3
zO(UEvOhrx{9<HQspjPCc>X$*7TC;-WPp4T}UGV`6i!e(97<maW@@II}FIqV6P|{~l
z=*qlqNDhbV@xrAh$_&+UtYAso(PhO<SpXB*<?0yuecU0WtfF99GIM4N@ag4$(7OL|
z-EqVU*zs~y@CHpv`#=ATf-?t&e>k<&+qQaJ+YZgwfgQsrcxEG8TnCFycU-?AOcDW~
z9Z+{-0G|g>r2;LoF>U=$?-?oJFWPGA9hH9KTH2Z9_N+V>f=FhAb;GfE?6c?5>&4e6
z<NIMkaT#z)rMvt@f}A3K($CM!|L))r1$MH*Gm;RwLo`brfr*Pg4N6k2#!L7#odmOe
zUU$8Bss6k8c^&mu-kS&|D8~<&`sDrfER-bLvvyGBP<2`fe|6E&PG?d0`O7M{`!!&t
zrqI}{tB(E6batT5Gs=adJ|2AO)i#fiaCTC`Zc>6zH^>5tD3t)zi>twhK+@|^Yx%^c
z0tN(&eTCKIYBCt=P`#_ZLtu309<a3szOKi3NJ#KDp0R`OBGmLJ2HP0!5z~uy0N=?O
zwbN;tqZ)({@o(9y;*>rkH5g!k6W_dk+r_;8w=-QZE$L{uK4#hK`eo$LI{|EJy|$)5
zgydlc?$NdJqc1sDFI<uUxw3#CH%>TMs*t=pndw)CH))s>)avz0w0lCDsu%bVEMP#E
zEAHWjA;9%C52`I<oS>L5MiB)kr;j?+7znycIzbV0;eN@Z0F!P~3Jc_mFC1aMP*1PI
z{N!&bygTBMt1p_V2Wa{60RLCD{q_4gO%Gjbk1;h6*JK>#-@>r4EbHA^Jlfslh^7SA
zU_RB$ARI`W+kA3D*ET55E2s}^thC-w=cWsLw~B<Lq}LJ-c*$F-aNJr_Aa9fc*CtXD
z%wrh<R6t+L*1hF0iH34r61%@1yQ6#s)p7XF3~)v60j!8VHcH@{m>_4-OlplN9%t;K
zaMM&RPnS@}t;`m;uAl`UD_#PYnhB6QYK<4g09*u`uRv65r_B!N+ELiiq9N?xUapz=
zLMkfV1$65zVy&e}WPjU}Q|G9fxi~Jn4TbmbR5OlH-(53Z*~2<gm1a1=tN+jq3mYoF
zTQ<Bnk%js|ommdj3>pYC=7<rb{M`h3<#$MM+|AxBXb=!A3HFzKEF-2318mMVd@kSQ
zwTrsclINvTw1X^^fy*LU<6p0eF-ZIIdo5-j{6Qh+_u<!U^*xE#a{KnOS)_<2U!6u|
zdCPY>+%3A*4^$5aS`7ZgYa{wFT>@OfM_B?Ho6<}5y;i)D7SDl(!zS^aPKwJ8-YAG6
zueCPOm$<5Yq_kh##%~{fVJu6@uW6Ae$pH;%3*A0`T#NIEX|^xfFI}gxzI^!BHJ8oc
zZhgs-NU*m&hMqARD1VuC*(s69-X<(2!2%E)*y9!dWgPgoJ@qREc$1iAIUq296a4&b
zxxO`BW@U5wH6Wko)(trWn9CMiwG>1Fq+u?Kc{Wz*cCYm?k1_$W%Z?XX&#Q{*`vz|H
zToj$=oD`r~aOJI=U7ycJvarAroJoT#eK$7%pd&bt4XbQkL@jIpzPbJv#*DvZ&sUO&
z%A#>G_huY@0L@Lm%>!1RnwchI;dY${OG?Q%=H5=u;CviUm{|kev|vHjx#Pog*9Kk9
zV_LfEXL~e<&p=1OAE#gSO{9pzUf1kMA8Rb}2so^MG&JBh>yG-szjWTt3PeP{HCMC)
zbJMj-uyap~cYKbT&QNxCb_8dFh9~3&h5-puP+S}VupI9Xomwflxf^CK{KhB;Fy*vt
z&@=v)Z0v?f+c?xXQhNgsZnMf-E+L=svy}4-t1^UynQm*2d;b~K`@q0WuU_kldJij@
zRv83%iSjNOCI;TKFn=4r6AFkccCPr+R@<#OVtA6Y!(5z+?29kqH&yd=L9(sb{=dAi
zoZ*Gg;6SlESuaK~SE(NmOC)4vyNxUVp%tEGJf-63I%oBI0?lNBL{<~Ij4kJzjf9)O
zG4jr5!yG_&aI2p><1f$h4GM|J`5R>af+o&n$>^@M*_pG!YOl2`f{d)HYV_^oQy0*y
z!b*%ofD4Gr#jM!S<QGi6ytVY?Q9!V;1H?CsHtUJ^TV0<gCdWFmf1Ewo@f(~27U24>
zY5R1d_?I0VQ&=Cw(7n5-<s}tSi%!MlBGgQza&&Yar^FmA@Co3AIQd$QRZ;n`Ni{V!
z3#QC{e0}TJBY6H)Tg-TmoB-w+lvEiFhTOO7L-kIx>*^Al-L7s~RSbeaO|LIFSs2ux
zL#MmmVJ~>~z?HvK&(-q*nBeHVZqKSY&z1(1wHH`Sym6(++nE=g27Fy#@Tt7vVT?r*
z?5OSTPo_(nSo3TpMoN!M4^Sb10ym(GbaB2pUGC$E8pt^#j=cQmqTc~f;zX0<RZ(ju
zDWVVf)B1WR4J!QG-?y!^a{wjwS=FY13Di*0O{=C8F_`_YJaKVxGU@lKowk*;GVboY
z_4V~b$y1dV-owV7#+4V&C4kX~$d8fL3Xxnh!Ev`%*aGFm1%WL_00B9vZI*21(W;|f
z6Mm4*y*C>?+n##1!0kvYt;D!*B~6CtOD2gNowqnrp);zKUJDD_`(VUtpYTAJZwxzD
z#>|JFRFX`N-k@D~=*yCZaf`&mm4b1<SIb@*VDfxr@7~kxOl#lBh**{abOfm9X-vk-
ziF4h>v^HSn397F-(Z%#nHXo>zHDBU;J>4w_?kb~+{XXxL{3|NA*;=2$!ee~v?Rwn*
zYsdnDU$8{sWHMU3eD}i#vG&$UEmou$D~Q+VOTwj7aZD+|*?9R|^)(yBm@z<bj6(@(
zxqf4264<jPAM2yRWcygC=(S`F)y1Fp)=Rt&+(h`{0YIe~D#BdxsT&z4OpM|;JGnDm
z<Khx{Bc~-7>b>cby%VTx^Xu+>NSL?vXaiB2)Sy?yPyFAu00R3vO|sR~uO1AifXhN<
z!J<f;uxeWi=&ERGUnwOg4oP`eg|n5*aCeUXu|PDr!6S-qN3Bo%P$_DE99?~&vu(1d
zm!$w=>)-5aIA8=Sl*GxBYjF_2XqEr-XV}ot((8d?WBW2i5rca~WW>*s1OD!@jmdgt
zcLI!Z+!+8tRGgXa+e6^b{Fz&PJk%|@Rmo<L^qU>|dNbjPQZU!h_Oxx99JxdA0~G@~
z#Pz*yBt}|RajvAo_2iofDd-Bptq$EM!?ir#?(yF3S2h6()uQA=A&^@38Qr4b<0JAH
zZU-W?P!?7_AR{$gK`GfTKs-1Ra7|ZNSLwBD!u<UFJSk&4uI<0{<XXO+@2m(Aj$p_9
z1;}223Iij7Dw+2q*QXYhky_LmTEIk3k>Z+G=wnbW#j=3$H^`b!R<MwTI&rAM3%A~|
zBe~z7^ePsrxUUl|Z=c)_#Jvp;3CXUhi3e)5ipt8qN`(+Qx)20nXK!t*`dqhVFk%{~
zad8C5WJ>{u2~}|iinBNeWK=Z1VxuJ3Phql@JT6Cl)R=78tQ_wH2bCDf-_k>KlZ=wb
z=TCD_#Nh7GUA$rh^l?gBoCt^Qe{31td~}kh{q>3Kq8~m$b!?#m*<D>-VbD%DP)bEt
zDl-oi-#fh-4rm1+olN-7f1D+pK@*R*93)vC$uK~8*Nu(yEqypOqf25#&5Mp#WIy1j
zUs>z__ZI{NK@@OtlzcEmPpSVo?xcthq}X#v|Nc9v3@JGh=6{R_PAE|j++_cLmk-l7
z@T33iAJ{?z*f)~@`U_IT4mk2?!GGK~)c%n~A>aP{G(?Ob3D|#MQV5L-Eea0#zyI|A
z>lYP>al`!gd2leb^Ri$$<nrx*h~QnsN>PG=_}{;yVTODQKY&gG_|MDwDQ0D7!^ZT3
z^Z;m!hC6WP#M2MfEn1mdSQOS(`wjk=3s6u1<gl#-dqa0J|NACkP_e(275ylfbbp8>
z7yzH0lLP3y1ji@;w-g4GXh`}`uO0ABuWUUdME<)BLb%=k?+OVrCWMkaY>N-2^QG+X
zTLN+A*?RjeiNL9M|M#TTr%I<~TcxS3wO)ZOnT*AxgJzroMxuB+#2|PJQC|&-bGl$N
zDk@#W(w&n$og?-~nNN1Uo5=;iW2XDabeZGuYPTI~HCNo)d1+^Nh6V;k-xiH(+Rr?j
zosn4cc`FZxQn%bQj?S>2S6MA`0~PMhfY$_l!^3<>>yAJZK8NHX2&fD**>B1mwmx|#
zr=&bb+u8X)72aobKz0Cdv{zPEya|tRTCd@ua@^q@OZw|Q;Xo4n7_j@{|NhCK{6FpC
zwyV}B52%(j(9Z4#!(*?3`&~LqT|U|ZTGSihN4Kj-L2JMWPRp@Bf8@;6RCjS%%K)^4
z0O)BWA|b6p#l9!<WL4DEUhsNqQY-u;X^hBB$F6di-Le*Az@`4V!;*Z>{<ivH)a%jy
zlc6D4Rq9dxJ9-!lDjApt9NX57cx~sCZk&4HQ@$N`9RrT$rrYwJZbt80TH5x5udGG`
zvFW`mEG$6AoRHr=@%sAu%*8A}DoXG_+P{8?nQ?E`qf=6(ZEcxhK-|=|bq0jX)z9mj
z2RC5bR&CB*7r=jbaU@uvBGf?8)DJff`bK=*9LeB-LWR<G-A_IcyKa-|y6#Y_{nYq2
zYiT@Fs^h*Rt+zW1R;HTo40(QT4iOdg$fdNTghAgPMV^uZDm@N;-W&be#jgZqEw^;J
zxw)20O-ZW7YR|sM#E1Ze@6ik(%??zvD>Uc{Ic+jco;o^&0ZTQB)Am!FVGQKo<syT%
ze7wH_v(};8YCM`QD<w66S6i)Bvf#WP&U?~@F*!XwK0e+7gbfDPx<e%;BoLh~|EC%)
z>P;KcQKVYT>3pm|O0zP{dw<;d*~rNDpnB#P_-yppFP|>-p4cH3{i*+5GTcnSdM}%4
zjU?o<n|m)QDJd!{Y6D`ij5BAo;y7_)7eDX?(PvH%4*RcUwWxP>ZSC3l1d#qXs5N9s
z65~$Q{Xf%#aXqZx(kRz|_O&ss@u<b^M<-W!tna^S*P<{by3NT`#p;m`xF&3L;;`(p
z>lMAk(6Qe`iOjyxvHl?;AyH8E%zyT$r%5phj{jm(87e~0d$nK>UDWGq=a$pBvP`o%
zi9o9VY$KZCT!9R%p7SdF^iNF@5E=s^Z{F(_Um$H>3ADSTP-Or6z$!%!>hhA3@Q@H>
zD*5z<v&4xV2qmR|XCNvo5Kc+&)z;QF`|{<W+RM@LqP!bu+yAq*hQkHYusV!5ztgVM
z2B0S7)S>ryisO0Jxcpk%^55NOo(>0vD^5<_ZQ?h)s^CGR_S6ed2E9f__Ke}w)YO7)
zFOdkrkAJnh!xVEEksLS)foi<XO53wMnIu>TYHwgTV8<@zrgt3D-+%hoO3_4y*H55l
zw!&x-yGi#Q5fRb-MVXfo&_<^qn&S!|=wYf|a(yc=@W9sH-3>k8+c}ZZ$;ksiIbDpK
z?cdX(53Zs8*~&nmweUS9#ciqg4hWPFfJ+kED3ze|RTGAsw^ES<9kJYa7=LbVuI}M-
zerJFGGg#95Ra7rQRdJ({e;1^#FE%bNDmB%8F1EToBqG9iIGG(23+n{93i!W?CM<L@
zf~u-6E8h^D!Iu~}=D%;Sy>Ykv#1D<wJX<sHSdGi|+kMiv?mY#h6Bsvu9@Jl;nttR0
z?r;*5tHVWTSm8lz51MXZpwWd00dEf5|8&MtLu)H5ocGF8EzIU%vB8w`nHqtoo6+F#
zaHFk}G(O2iKO_toCfrvGc7qH}*K47Cl9|Aj{p(Sde!~d}D(L|Q>wmWHS(V*}6zFB4
zWn_E-b&%KY{H1kkA#Wmnru@(7#d5pN&i?;#6o3EzeP!YP{~XOlpji)XJA-Q1L(LeB
zMaT6{O-=Q7z~CeKf4*-xLx3cIqW|G)c@*pxIFVd?m;4#1t;><4TOu=#3sxL3#+7&<
z6k+jGhW@u4pho2noQ>`AijZ1~mgC656X;W7a@<w%yzV4X($wsqEL2Y7cBI3?!g>}d
zD+c?o6#XY}AaviN`^0bex6f=I9E3M--n2uh@n!2Cmx&@D{`ZaKZFBSUF1I`R^}r~Y
zlVm7}<lg<?JHzmSqe-t(4g-RzMW<!H@zqqZ0il0?2f4ReneH&?s}ko9V~zfwudWM{
z_@5JI2gB}n(0an!AAkPOk%S??@O@S`!!vFWhL@I>=4;MP@!!kplM3SN*SUFlFJvdg
zxMLM%XB@e?xeGu@z|HO2b|-C!`yX@bDnP@-lXP)$$*?#KUXr2008Jg7X(P{$-hv0=
zBwofVfpL+xViCRgY{xtz|5yUW2UXSRm>BS;FVi94gg*-M>M;C0@<eVe!(d+)?FnnN
z5ep;PDgVcAx@gGlVg2i1uA9^9lBH?*?vXw9aA;@WGACN9{B30&&jO<EK$}@n3@jpU
zr^9gM%tTLCN#yF4BP?1HXJlU&muYgsuO@Y~pG(rt`&DluRQ~Z%Vc1|@nFs2VL(IPP
z_3Ho1rg_n^ZoIN%-ZP~ks(SLmJ^ek5im!iS?1wHBU3fuJcN9k>ZQ7+PLRJ60{$wYC
zB7+qMzkim#%l|{PAek7mygIA6Q`v_ZxzYsX_$%|WFJXD4<9SP-a|;5#V;$G2WeRpH
z@PKrs6!r985<`FkV(YqCnhsG_j3wg&V%p%X5MN1x&yd?hjDt{n&(W|+;nw~fMg_Ym
zwA75I=#B4~u<7|{;T-C;6I`lL-=QQ6nx!jVP$xh_k*20MmJ(dwcTlxA-Ft%3e;Gkx
z@%TU!BEx|VeJ*YZ*zQlUCk;BWg!c#~7cMYAWE$YLZRla}CXVorPcT1_8I5o4m@?!$
zFQ0A4sJ|iRSZ`XNWEV&~V8=XiYuT!hQ^!jCqWUKdY0Ja?pIv8W`{$9{mc>99JRvzc
z97bb5N|gE?Ig7!i3(35G6$|5*?x|B*uJhBfpyFt;rVJQ0N{OKNBd7Y6cU57rtzT}%
zQKz_RArIs~=GH|+1^yx;GvnB#Gwe{MI;K*I6T^jD*6TluE#}^%3g*K(_EXVCw7`28
zB?n3B$b9U<dd5&(^{1mR*^LCD_7Rm5L03@qwV#;0*o>pYaA4^TjDpI#CPCk|htA(4
z7v!==|L=YDYvb>zX{y{5XI9Ez0SzX!ICmtm7(V@qm;V@1TNFegi0sE6HzmTOoS~{!
z#;7`u$L}*I3k5c?QeF6%e4kHnW!{hHFA0scY|+kL@!IYrfDUY|vc`+p63lF6?5rXx
zd^)Er8oG634<xW-;GMy*@h**ezi~M+Ahpg~y0Tp=pk#Z_5M3$oPWID3CN4$>-FMiu
zQv9M&jx}?}o{3oru+`*iCK#B}66$V5MeU->7rN#V9&myZpRvu9)eV`F19@v!%Gs!f
zZtxNKA1UMDk3O*>V8k?F#&99(%oH#z3Ll)*$#K?Qv<l5UBLW}b1Wz3(TV7Nd@^r8w
zf|`~`w_Iuz24X0e2f=@K^cUDs@;3b9E2X9*E*QeqB+;&6az+ol|Bt4x465?|x~7p3
zB&0(ck?uz6Zlpo!?vzwO8h&)Qw4`)OOE-t^u0wac7yoyj56n0t=;6Mv*n6$D*9x<O
z`D@|`_Xamsem(F&idr(9DLr21!>N}*sG{~%=^-brD~hinUFu4B|N7r)^`UY7XSCg4
z3i>`Hje*5+njn=<TL1RSm*)F%)HksF*Y~V?2xCI21pka4HO*UFmdr(uPDXJU2~w!w
zAYY{ZZ%c{ufCs{h*|gmUb%#d{%bgh4k?jBO_oYhWKR)Z%pS?DvgFPf;P@M0)^;F`I
zZ@v2p5&1hg0Hesn*&e(KC@?Zj6?ICS9aW8n%$c8gvl*cT5uZ7;h^H+&pa;2tR0Vd-
z4GRamE*h7Iy7~{WZswGdair5<Uj7Qt3%A57ds1WRkG}0*pLq8urRfAyHOc?q>M%0c
z^vtMCuAVMUC86@^IAcF)g0h~167RZN$vBb^e<?k$7b(IZTV*c^Q{kAYX<P{3HbXLf
z#3X10?U2U%!&w(Pix~8xoAY3^>T#f_zeY_Q0S8uzB!QE{i|;N4pAZBcS8gPh;EIX$
zIsamq_;M_COUMBK|BlHAGE}E5r0gaS_>q67S~fA=1^R9V&L=I$pLO;z!9tPO8d7)B
zM@H;4E;O$DiNx^0)^GFLRCGjhcfY6)dX(K}M8_>9i06eMsrH7}nUz-Vm02my;QM|j
zrx}ITsEs`<;YV{q*qBh~_k1Umms<iMiaP$oSUe;wFB$8+P34ovN7;i1NZC1mLNzg3
z!*l6}WRevXA8;c|u%fqnHm=tvy?2|U3aodVM%q#Z(ZdEbHt_OfO@F>I1|ECsKX$B`
zVgC{&_Q&N1H_ZD5Qv0wVfkf|``~VR<wQI*XTJLi6Py|O+2Q3|?TMPO9^-rW5^8Gcr
zMbru{-?!)CagV&BLY>>W7!_Ot->i$~?h&=$oVXvB@-6P1dof&J-)p7G)}5Z4uJl;^
zCnrbI`@_FY?fGFx<QxoTE9>Zl^%zmm(FH?6DM?LBdu8K1XPTs<sTn-(aW)k#bVVSS
zl-WYyx1D0{0lqA(tSCQx2+w-hq?1qQtxNj|E*`)}C_2qV5=6|zVC>{h%xJv|la=en
zO$`$^4PWWBrNC0xIHVQa!<I*y%uZx?;7ZCalC&AQp$O4Q4<9$7bDfY4yAqQ6bez?n
ze1)jqbt<mBmznqe95LukuDwBfQdxfUoThVf{Fr$8KdpZ`O>lLI$CPxw4E6GDQ3G{E
zAk=c)<#n7@rw!{@^~j=<A3wrM=x;d5!NA{%Opih4{5hOqO6#{@hN*9I$>CBFzO@?o
zo&=DydY+kpI#%R<ZuDkfvweWrzGd9HIqT`7aM5K_!G~-EKr-tu5U`Tp9S*=<f9p3V
za#_!ufNBMsgoNauLvV_y@wDCVPb%H{EW~?Xn6<zeh3fPiF}gH!o?PJR7-+6~aKP-!
zH>287p%W!?dsZJ(#edszall3}tQC-pf}tR^<N46WOvE5U!Ap&(k<gq;MZ@SubV>@+
zO<M++6-4$f({qUkdCtkuB^Ug6I8hycD4m}#7P6enYzdRKT@X|A8cXq0Q;aNxBQLLR
zi9xdu__^Z}3=N*%NCw1w;J`n><TLCx**!*2BubpB*kF=R>3Uwc;v^2M1K(*l(n#8M
z8Z|4bS4700TAA{DR-q0|g(|~+yZi>=<g14zqpFuGyc}2;faV_M_0bARF@Vg!gZ}nq
zC>^Hls<z_vB}Mh=9_qXi$%Dkc`1LJG>JKU-5<pgmD(Es$;fnoQ-_`ZtCk~a>0ap_Y
zsRS2x3Tu)Qph+CAhcp_RQ;Qjw#mg7O6f{2|1{?O2F!*<F2CW=%DV4EAYIvR8H-!u~
zc<&m<^OiPTn9h=(x-?n&AJ?M0>KB&mK5oeAe(b#vK#9y0%ykT_+PiGIY{_WH@-K9I
z&LC%bym}=?-F-W!Rg>UH=AGt<WvLgNE~cZUT7f-+gR=V3$8`kX@xB!ZgJey0UyEox
zC@`uPurMo=Wd&EXow8!cvHCR}?<DuFLht@Fqg<%7rQ+tkszEpzmm)VEN$1mZUvdRY
z0<QwHaIR-yKo&=r)uIIg$N^Y`K`z{C&{|uYaCUZP;8#*ws<dEN6C#@&USFEhRDA@5
zx`_Qt;~(XG$YMgzr9r3~xH_cM9fVLe%6pA4`?f6U{+SrxHqdz64Mck1?i)HST#Um1
z$9r#WGTcm#8w5<?%!Q&MN4eEsgutQI@s+UabfTLtP}~$Ntu7Fp8sPAg)VpKH?$r-i
zn)AVLKACJ+uG|>~CX*dIM+9pYJz6R&%l`(ESIkr|JU%P5WdeJVVyjY8P*n1=*pofN
z_}dRRr&;$JN`9Vd@F#)Tu%52cTss}zn35E;Y(mII=9X9$LeL^0(5=EWGUV*xqxFO?
zXq1C^z+^?tU#Jh%>j89i{2S+*`mS$*rGt1y-wUAgg522qo~Ic^KZE^7DIh>xIz7<l
zbMFCScrEShwn6iS@OR5;&4u6VXv2Zmp;50e<C?|q7Onk$33_*qEg3YH{C<|?6a=YK
z^Lutk2RnBFn|8P`QJZ;FA7SXJXJsZk`Q(>sEXS5oljapAS#h|pY-&Y(S6H;x^lb0L
z6$O4mj)2pkj`uRgM4^wPvNKB0Nr;TLxcmSmu%#~^@=Gd6ZLOIi@BLB}%em$p*f$#c
z?3bJ5riA?5eBfC=pf~hu_w-@oFY%3w$n{Xm+vd66aO4#e<I~S}S^RFLs@*-d=C_3)
zlEG<k^B@ER<AGf=EZ1QtV$_&J`(~%3eoplc1x@YFAntR>xb!5KpfDy8|8Dm95~3!u
zi+<P9`N_i3vlK`Xg+v-k)|N(HB9c3*T}>w=^Mwhk-X$PBfp%)MdFAn#FpDS63-kN&
zY*m0gfSF(BtQobde?gp`x6>b3llh!!S01)`JTJzjykh?XG!pdD-;O3{W+Y8E_sBoL
zPy-)6|K7dSa&_f$Uv!l6GnlDV!%e8%?N`TZ3kLPZI0@$jg+434{xtBaqhd5e#naWw
zTJYdpvD6<_j4$!1Js#ivg~9{ox024ND%4V*G4>|g{8yl~|58GF$YlDzw*wnn8QCjY
zlC?dt`&{vlT1PRzlyoB@KJcJAP{(VP>hJJVcDs0;i@k$GlK8D9Gy*o9iR7Hn*RMOM
zY;d9H-0QjLs#?f~WuXhdU}-3rW}%VN7%P-wi6`8wSHVo}lpCyCgKdFOwms;*p_mFT
zUf~Gw#QE=kIwN(}LW1?zhC>SJY`fTM6h9}MZ__-kCX-IB+MMmYmI4QF*DuoMG<}lw
z6Yf2>#ugU&x177)n66%YKtV@ErKB(NLm32J@jrM>!M`t)+2x?^cebl2Y5kx|r<nfl
zJ3O*hn>Qa!YJubK7)5#GcIxDnD}P2|BvF*xNC=8LUU^}g(T*pR(z|aTMDFllLO05V
zHn<Ji@z2du8<N%4Mx9}3!{ds>{iOBh+)Bbkvfo{g!XN|{#Qx{`w(PZ3s9bycKk>#t
zd^5?6il9b&L&lSF7p4R(&BR<!k(Irjhy%tYuYN%FE)E0@A~GK~wd<0OV4I{kGMDbz
zVGNVqszc`{Y7h_07e|4LbSNUrX`YR9IC$GR@4y$GwwW^$J5SbHlNbob7KJY*bAM2d
z9B4I_Z(O?qaRGc*J*e_%8606H|73u<YS0=i3U48iEuGP|J&q{Jpoi|9&q4kpj3G+K
zr|R~caBVFlDgs+SWijdIW>yi#F2f0)wzk>1<(qdJj`f$4Fwr8eK)%qWRyo=}`F9;y
zgE4oxSRQY8=(1Oo0X4O;7EHeVg1eorPeKp@xSCU59#yPXXvChv>#z#*;sgWIyw5rx
z0fDt;E^<!@LZ4}H=oC!=01Bh$dmA+F-$fQH?x4l5Awm8ID=VOXonY_3Ucq5#&$PO{
z$;08FsCAFW<;i2oOoFEC8>Qt-HtlrKuF8OJ)l|x$F#q`7(J)OssV7bzU-GW-fOyYr
z4$+V}ouR1u1RSWnPsrOXpia}u-);;+D=ju3zaYsk&o+cHLw0m~pJcW)lI12e8as8X
zr9db5&%9Tu!xF{J_>|iW{0W@x!&T8Q(*9h6F;~9|W1WCpnT#6Q8LYQ?o!>XwH?&n!
zc3ji<*CSsxTzSUsPBsj`MiBIS?vHZV?A%If&M+!6a!|M#bLlkrJqhi&+EKve%xH}I
z*gu9q_3yX}dd$vCqqC^nbCOdW<Fq*rxG4iV?|m3GRBpH4Vn+7=jD}T!^LsonoSd6`
z0dZCiP61U6Km{>t>kls(!KS9Bpl6hqM+Y<~2f$6v#MXcvPd<190Evw*D4+$V4W9eO
zg}patZ|Hzs^h6*rZ_A@2!`TY0>sd(*hUE~5v+&<`$06dl{Jr2=I~n(J^7L#Bhw6J_
z^=wyvx+DJOTLC_|6B_;n2I&+RG|)ymbcAE&(WBO_#vMJ8e%%9IELd|HyT9Z%J0B_6
zFTKX23>yq$szL@i@SHq*J16TnN|5jhhQ8`Q1(Wbypq7V`3gNe`IdW9ZQn3;+(TB-u
zN(lio;(rUb0!FGy|FPQLCbvkx!H<gU(XApZGYQN`zaw~i3su%{fsN~tQJ-M{AsP#i
zlRmIGtc_utw!17aBsC(z$6&2r$Jy55G&tAgQ}ub{Zazs^X!D9q?`oPvQfdb*?hv#n
zTJWHC3m5+?%xPlodlW7+lE{HN(r&hL?f0ls27dy?(C;z5ukoO~xdD8^;ZE-wyVJe^
zEI$YIsTCUxV#ChW4R2A*tY`Ulc0C`2hkkOz2V~|y3G8|V%S>C)Q2NezN`8?16X>Wp
zsnwX)6m|rn%Ky6Po9oVzJ-OcKTj^iA+%>^y3E3h-6nMR+`(vko<jw{DS+F`Lp<hip
z^@Q*Dr*R){!Y~<w?kch8Uj-&q2JUE}!IYt-mz4LEd;15cAcLFdNv~=CfET<Lz2T+W
z^G`YNO-34!H{M(4AqMATktDj!w}BEQALyS?gCTJc)zLy-a^v|Hd!yH3{OvD${)DXJ
zsRYDlmXsw9POyBHm^s6WvbyOH48p=r;EldJI|BDbT+ha6Kw)W0-A}mniKLA4^?`$H
z?28mcp;2yZj~`sTi8b%*aY#}b@E!1Z^ZFIF7SU5nL})A@E(v;f;VSXdA|-Ue;5Rbj
z)(r{9TM3m{wfvzNVXO0}gRDhJbqPn?7Cvl%m<5u=jv^xML|>Gzb#HPHu^6Q@r#YY-
z^1p{~<=Kem4nU9{XWD58oB8654H}j6Hjykw%SSM;&)J>^02b5oYyvm3_N3ksGb5x~
ztf|Sl?i}0pz4wn1ijDEzj*#q&KWwq%CRDVhSkCfaZEkL^)jbwx-5yz9fQtp~WdTw2
zd^P)rip;tn#I%{s!TmwfZK0AOODX=KRiEx54wY%d)!;XOEi1NV_Yn$trElmAr}VH;
zn!$&njvI8+K)Ca^kU#+3e&fnjaZO>Q-382G_^X?*Me~Hs+Wq;p|3#$<r@q*vYl|;S
zR>M_p-ueF8rF>N8->b9ZUY64+g2a55O|(h7PP=2j+!cT!NIW$!#&ZV!oR@uSd*)(R
z$ncsO2@cEskKglciPbOA+<Go#ed$cE_{thJd`7{>JzdX}ori<7_>9HdKH4?D<R=R8
zmQP<0`o=T6vd}-ip@NI3DbBr|*hMmbw<6xX1I|}52d&h_FK@dB+yzk{hNnenI|#j^
z%2@Pb5u&n-e#u$b^2&P2r7-faz5?B_a%1>vr3>Sg^UBhNMcZWD-wpRoC0leVX1h-m
zGT^#Xj(NtYTh$V!q>FNqUPsqiVt!(|mguyhc-W~%Ryyz~I^c)m+if?tR#dinV-l{t
z7xdoaanM2%?pr)?RH=b_%Htv96fAcvMdm;31S<i%Da@pT5ut;&u-w@6M^DD9SKz3k
zu<x^%$!H17e)FCWuglm_fo#kN{#Sdd)z$iO5gng3BKRyjuD*Z9<>Q6YCEK%pd#-?z
zniS~kfh;O$DMfmJQ=@g;0*x(R(3b^uz<(JNpHc=+O4-5UdecVJxr+LGz?fDovb&T|
zd(!-_b?B=AtB)E}e-Z=2!j<MX7x>Ql;{ZUIDU;-_*>n@7x-fXPI_WJ~b5iP4*P@9V
z^no3}iry7tSiekst_!)gEo*c^=IOZhMv$3y=!rqn`AwWL?HgzaH~<TrX}{AMg(@_o
z|KW--4mQ~SN9XsB!kG^sGbJ4nOuU~NM7#exx%FxB-X+COulf~y`P?h!JnTDE=KyL5
zWa_A?O6s$H)Jnv@7`}}a2X{`8I+u|es{mfE2#h%GP!Ay++4L7tu4EMr9NJQ~#3L-&
zErB*W7enVp7FzIr-{tYWvQ|bf-v)E7rdnJOvR~kcfGO$tVa1wb?fU{JS_2fR`_~IL
z_t8s^Fr(bcNR_$ALb&mBera(s_{6>2Ayk5PX6+*jntb#wNr6}9UzHtCD`&f8vg%T)
z1pJi1MlqU9jBt_8#SGp*L#YpnCG)+q3;qmH5ZL$*pU)Cc3odmU;N9M4ZsV8aI%#Yh
zy{s5iBH6pvCvA^K^dF+C^#2|?wxm`LJr_-0(+F?KQsS6l@+(pZ%RyX0d4$TOU*>Im
zoKaXd>`m)yPX~qu!NNv36!nxmq;^a6la<ir3LW8PymG+qGZX_f@%})as%}%tT^Q+<
z=HYH@UT!Sg5gI^HF{DXEJ?;o_8)T$@6a1eRfS_OLOCTRrugd?7Vm5vpVVB$r)o{Jd
zE1~6|A9ycvh7@MiX@n+q-aO7p9%l4WBN71bpxXfRCQ<^XnFg+>NbKf|0gy>apB{|)
zwCJy_b23?XLB2fNyQL(!e}5C1QY)v^(o++L3&q}rWq_BWdi%O3BAQ^H0aYIYa6&1q
zLzi4^ujm^l#k9Va>-OHJADGBBuj+cp6lYrO$L9<Jww*Vuge7hF%%VuWV<yB{z!mi$
z=V|Xba|aG3xZm8*cwz}N2jY{TlIkD{_F|JQfqmCdT5G>_-siUcE^7XbZ)EjI5;2^o
z&eT!u7)qKr0My*RUcUxiI{$ltGk%Htr{2Ih!-VU=SFf^Gf>jt;K**XU7RsL&!(r`Z
zYka-?lb<C*H9SemnssH64S=+1pF^R4TErj^V1lsgz-~`5>}}A<jO=_W>id5_L@*j$
zSQ#z<B~Q^Lmr*=M^&MPi)`=sKkOMC*W$&N4MdCCgc@Mj3pkbq+cIuQ4rdAwZoCE$%
z@Piyl=zE@m8Id=wq`aLRie4jj)m6N%WB~U#39{r_tGQ~o(c(&e<wOg&`fBtYw<{Z{
zRxGR;3X1R4&fEmQ0{~;>L_Rkb&xkb@{wlS*XXCMRSSu3Gw_#Iv=g`}$WV=ZWEO><A
z0wl)7-BoE-F`PLtxlr22BxD<?qVn2j1X@3z_iy18!BfCb%4S}5eX?TL-CdR9ZhhLD
z9XkZL4v!p%&~chE0F0w)@SPf7q>J=US)B`7ZN9F9`_!c%UIJ&<*zS7*1#ZT=5RSTK
zO?CLtjS9XaM|UZqjB*`xNe*r)(TdB^yRM4cahEc4{l97{>QrR!k>C2f3eQk<?GMKF
zy~C%~KZ9!7tN}+~Ugq{H1b;`Y!U_dE>GDM5zx|h@A_pao;b26C;yLM_awXMoG<!6!
z2nqbwZxU)?+_#+u&J6Ne^W*D&dNZKTV37$#?rsLnlA<`lGT!BPE%et_FX_dy6EDEv
z{kifK07;Bg+_iq^N671E5fXy6_s~Q7QBwoeXGucaD6Q&kzqDMRQ5F3Q*z#V=iNdm8
z6GtRLVMoW4yH<DumUB^;HFSZ<uX}2CboA6Ipf0vB79>3NoG<?82aZ?LD$BuS-W}oI
z;ija`-)W$T>pGSvnVPr8^P<ifp960gIUT06-xX$*mrhv{PJ(=B=1uSD9d^?;@Z%i>
zzFCTn+1sxqr=`EmjwJ?jK@R&&9F3g2FK@kbIim6{(LnS?@IYOvfBK|R^gHM6C3)Su
zxMMC^F;$D-12o_c;cMAI0a<wPPGpW>bq!Cj?An6|C-BfM+P705v#r62fa)PHH@S_o
z6S7^~atS{_N`xzM^<a;a)kvc0%!bVUoC>lQV561vNLo@WMn8QIkEH&d=buXyqM-Y8
zbW&1R3_ux^_MrZEv^X%{Jh~x`oc4k(qW3@^hJNi#w+E#y-=~~@UMhAe&XL6O&$7A<
z-{a&&(+5mgWsM0OCAp<Ya>mubmBqDfKcPqOotg#kQjmLyc>X3)X1__&T1=QiWzL*w
zy4`bvi$DXO<<L)Pl`XJu_@y9-X<c0bNS&;?E=_RVBPK|WgcV6@aHw&sJsrq#Pkxia
zwmN>+TcXE!MSs{gm;m2z!dfNhj1m;tchg|9<9{73XDGQn#FbqHAm~JJ&==&@iVCOo
z_yhQrmJlW0X>7cckPQPjVdf`Qpm^kyYyTqY>zAfyoFs5`BiCbvZJy{1So(j`I=sl-
zVv&p;oE*t=e-2>EZyp<ml}1tEwH5foK=pn75EM%jdd*qDjZK{9^<jC_tCp4~Z=FTp
z@t(r*YsV3sw6{w;>G*F+aXS(Slg_rsG3jl~-!(>$rktCifn2CXu0U}!N0)0I1~f<<
zxCP;S$=I&;nno&wePfxk(wfbFlPJ3@4MTIv8k3Xb<${0vJXM3COv4+j_Jvz?;FSLY
z9`^F_KV&jcJHbujJ^z>6I;sLX7*O>l+phINTr`?CKC4I23XJzBp?~}C<+yfs)9QnP
zs%NuF$0}QdghLU7Q_QmR$$MKP?@6q*$pi<~Dnic-H$HYzhYv#F1Y%A75hB-Y^Q$Xw
zaA&)Zm@E%OSjdcTXvgxC7jh;4E^;iiwPpu)ID%Af@(DhWU~N<iU<1mRMTZzoCc^~{
zDbMs*SCd7C)(wAi^tJQ5X_9+_fbp4%>V~b2r0VijF*3bp=Lylc<Rs2G;y>kJ(RO^b
z?EYmiFT_()++g01N}=e%WG{pmNwdg)+XPFa33J4hpKZ<C`Lq0>e9mIKBgk;=v`p!A
zJLcN+{NMeo{AWOJTzX&^(xaB-uCpSa_U!R++;U_nXauB05W0^qtIAqGs#wYKOll=~
zJ!q4*SD^Wq<X)$+-dj&-s9&s~-y`WstRGu!{-aaMJs|4H=@0B1`N*a75n18f2ra4u
z5Q^ehn6E*_6$3C$FwgfwKw+H%F-bbd!~+>%60u_b16_Z&$#=kge6roc1I5l)v?#*8
zH*;!@S=EWMpOthwUgJ~<9@OFG5hfM-BvaxT$K)273^jzG<=Ky`KfZs9hun{2PqJ6Z
z%z$ha^@T$_6g6yhy6XU}JnC6&mt#`oqQ)!m987N>?R55=L_o5v8Z$rr4|3+ltY@Ix
z4NH}MZ^f|Xj8j1pJ071L-!qj2&1Qsh(T=n#5%q+SrA7O6u3WXse$zVRYrnsnm{CWk
zM?kFnU}3mU%Ju-!SShUflX8tJTrdd<QaaB$;BP9=%gM7O`X60`={Lx;voiefYZ)RD
zmWxf2D(hguaGKI;XzgXqH!T63oLBz1Ryg{f(6d163=tkaPsKKs^wCPSO;b7G$<p?7
zF;ALwUttVpgXcC}i|l2KPm<aj$Y)3#<`5hKjv#4eX?CQ%$o=)#f#QlJwd7jh_uNg5
z9xIA-?_;`>j-~arL&)1d8m1zDV~2L6dG8V45wXDXvKfbFlwW9GEPhZc%aFJ;mHg)5
zwDGOmbA++Wn$PeCeGz4LAH+3-S@U&}R~Dka6jw4^YCx!bve$7&bQ(cIhkb=lU^~<=
z(*2xAW!A&y;{ozj66m+MkJtyITNIx%3xSs*iZyxJWlzt}hOq!jE&px5=}L!S2#83z
zqngcthyE4PTB}q3tkC!rF=`mo7-Hyi;pO}1^1T6yPtsOOcnVQk(B%4Gr^J}_cjlCU
zpUy3-`Y*5fNqTu?g+q8}8J6h-4YMBn9>%<SYqT-4p|}eAFGU>*aZ!xSa<i6WZkHuY
z`XZ#s<U=xMnN|8+dIoU)Zp`K+@zy*#issR-MUW_2s9#A3&xDsY68LAUutF<XG;C{e
zITIarE=I4d8&-^Jl&3!r{>)|4OZ4q$6{qj5{ujsx2DtNUDZte}e!1L2@L;=ADj!BR
zS~J`6e7Hb_W?{`J3gL2DiY~vkuw~31)OZ~>uo;B;wMq(oe-;osqe@MQ;E6dF9q;}}
z{T*h!A8pU7BEQXB0Ga2RNXmmY>O|g~4F6M3abS;pWA;qO+nsI5!{+qcN_ta_nNRVh
z_da&>y1{0aAss~cF}?N`UV(tC9BjTNzFpBDy%yx5F{9L0+4zU9^!1zG3e7^)3A$>v
z$SA-QsO|f3Ojta!y<QtRsvH>5Rr>MBq#j@xaU!z0_D6pcRvVrf6NWUAS%Zqt6zoEP
zW=W7MgToQnWJZiO;!EB9bDI_X1mOan-vCS?tR?>i+|l4J?kP{ZmnKAf&q(~Kf2{xh
z(KVr>PE}se3HWefd*r`GKJ5`)sHoQ?220Q*SN*;n$FH2opyQCmXp>gEEbPN9052xY
z=&waJRzi(GXsZJhVU!$ne#kg<pN)(w=p{7%8pR67i9|utWiFvdeLu&8lfav_AB$;O
zJepT_>fydOF+3HN&*3|hl&^k42{hjyGTN96dd*&2J>#6i26U;y6VG~9r`h}ljec5n
zqH|47vY4xs=j9217*TXpJ1wS!ipSOVvwLsM=)dkpAbtqD5|&#l^ozL;`g?GFX-%yR
zd8pwgp`m6!_|Z~K3sw?t3U3nl)|(bmlA9Y%kgQI?CG*8g?mcLkV*kMDhP|b4DnGYc
z{$}(eFU0-I-p)ux&2n_*`KI4l98rRm#_Yn~;w_b!MkE4X9H3FtT?O>d)wD0j0Op|b
z7T%Fi6FZ~&jRt>gM$&WZT2kQro#sUTtq?Bx4e~J8eWd^n+d`*h`Ae_B8B4p!`YT*v
zMR1QObN0#HEq=$za=$-THF{<6L-6u;f*3RwE3<Pz2ExYSdzL*4nuS>5OAHwAWkq0l
zI0hsKpksUKu^l^IcOG@RXHQP6G6EMwQwT*&mG9mC3f?0GM~pC#1B&FL!KR?k^g2GM
z9UxYtn{DDrqdqfdkjmuZZgb|s>Q#LV=nd0`a~mViHnWkz@B*NafGh!&?twk%PaNMk
z<4R*xhAgUaGO9yLIxiaph^l6#+Ljwy^U?jsb~}9f(Q9M3ovid+7t@@Ptwf;ndH_Ux
z_-T~c^0Xs(bL@Af_(UrM*Y{2VLT_K(X~7Ea9o7;glEg2IFZ2tNL{BzRCDD+U!VbiT
zhnMl?G;9#quexBvmxFMHT|EPrYaar6H<w{0i6>p~E<2z{-B9<9SEJN|ZbV@1nV=B9
zg+j+=A+ET%pG{bx?r>TfH)NoWKk-~(OZ~A6bVGTy>B&<))KTPj-4>tD7&BL{SwO!z
z{{h9@3bHb%Uj5p?v)A}xe`DpmRl?9%$A<goIY|a<s*O0EyC%EG5;M3$amaY!n>kM8
z0CCh)c>CT)uy>0voE78jx~Ntdvzdyli**<o9pT`#K*27Hn6blri5ncHF2hn9<M)Tp
zyZC@@W=jc>A@DmYvaq208LcS*Z2!s+YU)Yy4*wAoD0~VO$FOtpsfnX@h!2-fR?t}U
z;#Bt)^)_2wh{RMk*F}{410UA|qG>MnHeyNX>4%8^W{lBTo@>W)llz+zhI>#!g`#Hg
z*@T^Tbxgwl-di7wfX-)wub!g_X`61?$Y33Pn_7{4s*w_BTh;lH3W#;YUxXEzlbqx|
z_nC#@js+LHOiB&5{|gwPYYbar)wvyVpuJL(Cb}KrM@7Bg`H;(JXfPZc71cA=p12Go
z7{!cw(Zg6$fICHrPWuFY)8w+R0L&m;w`eX78;IWxJuF!aF0d{N8z>EiR27>LCl#`6
z4*r2=Pahc#mtvj)0Nq4#k@{(4!XDx9p<nyMiyr8`@Cn@9c{_)RbzA88nwB#kKnYO-
z9+uBsxF$hC)7el8Pck=FO~z8|c5R$OC!rED3IL}6Z~=!8xYg|ZC)V<9Y|!oyhcnp}
zb-m3Cgg^y0_`ff$toh|Nx{1K_&+!;mZgqkF-O0?F9-Od5C0B2f*f@`Rk#qCPu$!rT
z#f^D!Dpm#s5Yv)LINDb75VKrOxAOrqh$V8fs`BzQw|-dzG6v1z7v&hh2M^FhkXG+d
zlHv$s2CBAEqo(kkWSlUNdi)I`Czis(%yS*v#sYKRHgxX8TK&@-O=<5yZO-ps4)5}u
zam<&O<+GxFxCSnad>~Z|bcB56H677Bsu~8_KkxWGDME3|p>ytwzquMxqrbXSf8lhG
zmbdi9Ox@12fc^ovNIdlky5jE*&pJDQI*+BY_^sBkrIhY&K11m3S=1<jNAtru7ZY>1
z0?h)ix)*Nv`2VUQ&~`POm%}yfU@U_@`@*d{m13+r)g#-?p@j?m2prmqC-;fFN;KZl
zJ28l`mD)#xB_L;O0D+$QUOItP-WHH$bk6frVljPT_x5X)&{Sz@b@$q605#<`KX>Oa
zMPq~OpmMf&_Q%Gr^K(QXoAXY^oFsfAhk)t~yoEK#KOu7MHbZ5PJa538fCa1q;3_15
z5>wLU>zDS5a7E466vcUe)X!Qwqff1U%gK0ddiL`IC_PN(+RqP{ljc6xr7f1gKWy)l
z<9tmL2+Y&AoBrgT8Xe5?D5;dxFk1fJwe!qT4wUK-eO~SVW5|k3kS|FiYnt?Z>_mjP
z)9i=agWowrZ>fN!$P%!7&cni#*iJAG?%XQ+)iJA;7n6zdYy@SXFx$PoE)K^J?orHZ
zG$U`?ZvO)1&5UODyGAtdlY|>QGZ35f{-tQZwJtI^cRKlseeMRb=33*$UrE69@Nfpb
zVi40wton~v3SDIGh5)pF8C-nzh=&j0zb$6ek`n7-mVsk&kU9ttsSJ93*E}$-9+68>
z$=5rhFbc?=%umecxFm^^cT)oNgI*b_#^hl4ih=4cmUda%!exb2s%Lh4(i_H{1jVLg
zIh=_%3V{0<=2}$ioz||-u&CiFVOEtO8QNKV!{YZ7u1n;9MdQ~UMSDc{q)Q)lv?lyz
zKi$WYr2f6e_lpD14oP2_$Ey?*kzRJ(5U4Pi?KRu$S2U_HgZYY9ocV?&IYb%{1%_nO
z6+Nxr+GPa?y?1V~9oT8`*^d<R!=oNvpWGJ{taD{*D!=BJbg&=-VzswH2PgM};LhNy
zKEDQ6N_W1Ya5jV(n-2@T<r$nMQVEtvQk07(ea>mH+;3x5*!ucxIGv{(Xbvm-^NNb&
zP4|+%92g5*xn1y@(a9V8tVy*vUo6rLKbo@YpoC{gicQ-sl19P`loRN?=<g`&3u})D
z5cddF5x6;M%_m#ynew&aN@|J>d>W<Z{Kg)K=bLoOXGkIJ%m>O$R1u$_K(BJJ7x0dm
z#Nhdqx-vTEqdsKUqVz9TCfnlI$dIZ_yLM^92)xdX0syI}-3)C})y-Oh<I&8T;1fkZ
zmDAH7Pa&w#{ycDS8h8!+AteKw14~1t*tP=m->LYsi-*AN)q+SqaOSNTzNG}{CfH|{
ztfm-_QE{e(!x>BsEu;X~qqk$Gq7pRy7|8{0#iz$uuhB;R!zT*^z6%Vn^->62tK}+k
z>ATJKKAk?btLQcnDw)F@C=r7zn|+n#zyFFD(P;&M($*2kl25Cl<Gn}MaC9s;BfrXu
z-?&7cG^BK4eQsY&_u&F>w>T%1Q8B~4_o$8Sn%TE{jqR!KnJe+vFqA`^Bm#7KoYO8W
zZi2p7ns83?Y)wE(F|jx;>o&C-`*(g~_{Jv#obN){y?H~S`!MzG&s*6(DjHPa5G7e7
z1Yn%w;h2b}J@X4>Ba`X@uOgH5Z|4lLv((iu>w*h=ch5%r+^Z5Xx(w^IBNn)L0tuhA
z&#fYc-4j;|2bAoJh6<<!Q%jn4OQY*BMYC4=7NmIodz;Wt)BoN2?u+v4K%ERuw>pHw
z<n_fX4f2BG(tGo#&5`yi!R4vB&Ls1dj`FZp8{sPy&_GYs@vsVfCE8+27~<_R>exBx
z{axxo=SKy7>$~^?1>psvQ|J!{A}@Fso(RGrQgIC~K%iyquVY$R5A_kU3BRi{q+e`%
zN(iXqNy;uMP~ozlGhuavL?rAij+j+DjsDiX?LdyMu64m5WXD6=e3;3$0KlEHTD7J6
z=pJ4`ed7g=x`e4>XALa$;Ig6UpI>nH7;dLgA;u>cu|anKUe?F0S4!q*olqbi{RDJ(
z&2mGJ!T^9rI3zLJ^iYZqxEd#|$ty$sGL#^JnHV1sg~ct!2<qoNOM{YdcunRhLr~zu
z@Uy*7!nki7HMO3z10uY}_xJn$bVZ<1VTRr=eZ_@Wo6w8OrGKGxyvWe1czI`3Klw;e
zJ=fkw+b}IC*m}XkWT_waR9MSLkfffyE(OEU@q9q@A=X@&2XO}MmGGmYXl`{Vu<!TY
zUQLSzGP|xX)oR~vG+gurRw`NOk2lVJhIXXk^oMC?#bTwrnODC=hD%fR6ZQ@ENoB&*
z0QxS>D7#=Y&!1E+0wcBj6x68l;lsU_q#lstusr)*YhTrB&8YsKN+oq<1a=g`j$`>R
zXSvURkL7{W9}`gfB0Z28-DJ=sb6v4v&+e7L#jCJoeBq5~b$^gxWOrl+DqBl?CJ|fT
z3FQec5rt_la3su>ru0|<d7VA!t-rWAiJCFoiuSlw&Dh?-^CG9CFQ@(nz!yT?yIF<V
zGo&i+ghdSQY`1BKU%31gyJr#q%<56T&`B8#nRiT_p~Y2mVtA1CDNheBa)Ch+p(DO<
zridZJXUY9d-4W|k$LTMWfCoB`;-4piPlC~VrKq>6J<3iucHlTZuJhd(A<aC4MM9;)
zcm3^@JLg#!<`d{*27sc=#1T7&HLbTq2#_kso5@Bl|4Kh(=j4S-ymv-wsVKjFnFlag
zPiONQw|=e6x0_N_5(CbyUYpl>j8F~l<tS$;wPbDF3SFTR3sAXt8bymA?Kp`WXMmN$
z1|(?!%mUB@tFBn^R{Bxd)=kS1KuL4vHTnuQkQ>=IRMlNKp%l1W395^Fy-Ch?<gx4;
zY2*CaNHpsYBt?8b3ao`*8>ounz<y)?i3xd|&A=dH;3VR=^33n9U?^iUHpWt&Nth`s
zuS~r#;P!{pr{lj2I!fW-um|^|irG*OBb!<)a#=e5=akV{gt{qGI2SM%0Md)ZjAiD%
zuhdDuc+HVreI}=UPtCfOR%Iz?S+v-bbJR}pKgwu3V*b)VVg($jH<c#H5Tyo^@Ns#&
zX;=fH7Ba9x4?o5K;8s8S1o1X*dU)tI@4V&w$%h{>F6%Ud3bdc{KDTa0lSnJyy?@o#
z$FJ|%2yewK!=rvDHW(9e{);2WvX^a{O%EQ%UTw>S1D|78cEv!fq9#{ec#(+#ux0M`
zZBcCeQ}%5(_N51c0noNq+C3OKuER2^{zLD&>P`mF)Sg3-ZLaPA@lWb+b3Oi#2+IIt
zKu8PF5-PifJM1WA9L?S%!pUR5-7{SwtZTrP_$q6Irz<P+X|jf*88(+l9w5Ija%a$B
zOU6Ae3g{z%MU!Exm|fb45YX@Y$@{yXc}LRC9q|hE)*B50=-C}PAV5B7$$++;54yE%
zFEZ=vd5bCn`71UE>6y{l!?Isht<CkRY$&Ji?Kl*+fznu|eP&EBq2RVi=n88F(7aM#
z`=C#%Z{Pmu;M}F5y!$}SfF^arTM{Pxj@@xN#E1K!38U#}f#>+ca^~~5r83UXJL}hO
zU=KBw=iIkiG+{15L6rk~<{Lp<Dr|Sdp}I?Ig$CbQ`^Ajn>TaMtJ+^pfgv*6TY;|TS
z(yMid1=Aov{Bz6O!2JVP4>1a=LWc>)@m~CG=XZ$X{gPfPKj!%><`NxnI(3nV2ijW9
zfZl^|w{pOq-t3#rm1VoyIkveI^`8(|L~{1`-%O5mltNSf-UpF8R~}omC*=}E03-(}
z9rCO8)pFzRMfPW>4T1p<Ijz3v1Dkz+Uz-J_NSD_CG8+C0ECks}$?<&}=~jU>SXK5)
z{Bq><mwD&1@-NM8ZWQDKiUMNHqbU`v+fP0{QJJ>5)6!Z)xHcW97QQ?$*v3Z{bzF^}
zW9e=z;ReM1o?AW@tka3_JjJ{QGXl(%k!f`L9N8N#wUAhMZ@#p6nfi!1z8X3$^lexs
zJP|E`S#6zM_02V(_Q@MukZak9sRd2!l@v^B1e6ws%>48g&8t?9#2a%3l=%3xfmg3j
z*{5+ZOe3KJh~N?{Nv?SS$COxgTQ~w88oUPI*4cAS^w3}5pI1-vpOtYF3g#cDaP4z}
zyz~u`P$-~gg3H*~<ei6_^|<_f$2o&wLElZ1H}}26g1D~M8jO16S_;R~mBVB&Eugq0
z0mBB!%FSRQ3kMS|T3m;;nFj+%RQ}<Z+@>5xB=^#>#O2bV&jn=-TaYHQ;KKqT_+1n&
zURE`FHwut9dzmc_fa>hcSK>^plhM?-Jcw!i13)Pl%!yJXfK$o04g$RAKA?a>9brR2
z_6C7icBCfg_CYF`VpA|*Slx*>1!~%6j<ptJa)x8)2m~u%Ofb?DD2}<*GPcG|!5&a%
z+B~ORlQ0OU_2YMmBsfEW$|bTnk4%E=Wef#+)CoQK1sc9cjmD41grGa3{p#M0mFCMw
z8!je^8T{SwTy6X_{6|15t6Vu8r&q;-V;$N1ci2Py_=5p^!HR_R)U4mARzxC&CHlFx
zn>pObztEbQ9l$BX1v(DG`$p9jP)7Nl1sg<9la%g=^W|GrzQGb9)^$J+vjSXwQ=Cu$
z7GULPvWxutIB|{ClHR1S2LErCYoKRZZ{Q}~E7-=a1a$dc{3jL${LxCMw};YSf)6~*
z$5yKh6($VN`h|Mw!(Qat<psGEB1tPdAh-TQ&UoB6<~KsZm}s%i4dASvIld##;il8D
z<pMT~APy0~`NQ>jUj8+vrwxsQ>BDM(6k7=rkYn2Y8!_h6Hc?z<Ew;7)C`HG8Gh}FF
z`IQQi{LhAy15`te3CLqEuU%gW>`BlL6#$SNEC_6%Zye~n<cx=k3$(Ri`7g!z8?ojx
zNP_9Gj}(k%44~@p%?cjdJ#<#IgK&v}c@b6ZhD>k)a+_4_csD+h;Y#fNNhAI!WsSH?
z^l4;#x?$g&QNi$UJ4X8;xrk>XV8U)XV(pPGZo+ae^jLKii+_d48U8H49e*HG&wI;j
zpi^K8)ak_xPcTa6RN&eITvTW-!AnyQXlMK(`a_GUbvv<1u9=~x$e{i|RVlv%?mXRV
zK7-(}|7L4U10|!E*vHhgBgPkdCH_wfu>DmA4z_o=CHO%|=|u^1GxPy|LS=esZpxV>
zKgS3k1_(t#%N7JUu-u|P0m%>eQ(gezisRmL>oh7*tAX;Jxj%;<B0firn;g_;7@q$D
zNMlkVJV;(!fOZsS{B}`MKDiNk5hBBgBTG1Rb$?KK8-$P9658{nxA5L_*J9swakN}u
zVeWH~Rpl459O7ZBPvCN)|Cb^T`b;oEN9%m@Uq)exgAf=^Vk-{!B0D&1t|A&Tk&52n
zfjr!7ZnQm*F8c)csThQlKaX7Lmwj%%Hu$~1>`-5|myE&9#DkGU-Tnp}j^gMib_4Yy
z3@jXigXpW|C`;~Mg6n0b!C0HQ{Z{v~x?w2h92~`9qKV80Xj|6KDg1m&2H!!g+s@E^
z+m6fW;Ru9JJw-v^t1ATmK+DSrhdVg35C17lYo!QFAX+iMYrpbzJ(EW(u?p9a=gxt^
z79i{}JMwTDOjPtiWfETWR*Y0LJWkys8^r7}6D>;bhjkS!m#ttpzZt?&ndeM+tnT3i
zOYU$n?x2Qc%JuZSPQ@R-V?B%g%D3%sP^~@2!1ZVV1MX?qFRVe)W?EFTApG(=FwMgb
zB3oVJTx?lNHMT3lu#<=@8z07jZ?rU7uk2FK5H_rZ0kr$V+gC(ZvTn+m?Rc%`uMD23
zctnpCNidM1B#-egR&xj>MZ%A_yQSBi+2dLFFgwb+b;L3`l=R;YYrTjVGE^5hdh9uG
zf)xwi9Wdz}oU`51>5o-4@e}tiUs21}e)m7w1Mo_h0BdB!sr%(T)W~LqO?mlyad%{U
zTpY5dYWySFDbDfnoLj>g!Sn1j_hm-pBl)VA)%=0y<6i+Va8-24IH(o&xftk&%mg`+
zyf0eY#RL`@#=Vf5FIFzEUt43}fEnlWg_S3~)~5x`qF=wLmq931W|o|soabhY6qw(C
z_pLQE&8{gFgmty=41(;&lSSvzQ=r?O0J75VP#n+0s=ilf#La36k_#!^uSthZJrE;?
zlFEA)<z4%^<$Xh2XV6^d7=4a2KVBmazAjF3E@DsXESSJjfD>TMIHX@1&25<Lj{E}m
zYRMW7whr%TO<jyliRJdONWtBe&U84z+r@?nro%i1ga5M$QXz-}xc^Ws$%&#~{F+)`
z@6%HU<n$z}lf?uZ7K=l2Q}27sABz$`Ny}wAi_10-M0CH>1hnb->7Oba<It<9ZPh8A
z$x51D<DQ{^{RCWOFLpsB#RD{ByL}CRf3cksnw-h~-AwZLyWy<n7LoM16^=nLfYDAY
zwyY2XdA&)tENZjrVOh(CJEGHS8=ows--P0iVNNfqNcr^mdjDY8FWHTk%+ImLO|tbs
zj1nv6R@SV7{)qjURvWP4Wr*<C?g-dlQ^8~kQ%#<`F`l0((Xv``bM-7(*wX)<<iSf@
z>Xq|a5f?S`>~#e>u~LOoxf+(hnOFLv+^i%X7etnT&4H0d7(V<XdO3|-PJqlxh`$!Y
ze-L{cwiUuKP%Uq6O(Wb9P*Ls0Mtl|XcWg`<<U)GP*a8a#U@?;vibHqya6YQ0uRk32
z_5*8*{7Z-=!s;c&a3`FlZ57NqQPR_g6lQ)4d--lK#qT0tLrs9zRA?q!Qy(9tg{`I8
zhHR1&5`XYbXxa2&zn)ht-rZg0=Z~MU6bYqdMN6i^mGfncA%u;pkLCnz985Va`C-VA
z$I#MPyy|yX*h#60YmBZ(uOH6G?hjd@cAW=#*PZ7i`1d2*`5ah#Z@GVK5Z@c0DH%<y
z7fL+{$|fWR%6#~@e1$8~w+!s9?NxZIW^q31OTlMtmlqD&<qVopG;j09>{uj>Evtrg
z1Qq=7)pYtO2>hz8kw3$}SttBQngZeQL=Z7X7Rymy!{-m3pU|IWJSyrvfkL&eM#hzH
zsjo{jvnrHY({(4x$r7O;fT5UqUfvYAuT}?S7ir^r4^@d5PE45Ck=35R$1OQ#7%}zw
zsn^q3wf_oH+a8`>Wy7wQhd1tL2^IKvsoc8kf1^d}%KP-d%0~N%4qO;ka{Ey!Y~%TI
z>=eqsBJ@C1Nl=}RJNdNo!^r$TquPzjxZd?jebpU5kDt`#`RNh(M7)GRuCA?pvaw+V
z6B=K(1`>p1R>5%mAb4dVkw@jiSE~V-?7(mZgM{~Q{ChP?jzu=P;)$H|SJ}paVO`*J
z-<>^r-1X*hcnduKGpqlaL~3qZiL%sgfzrly$Rsi}Y^fQk6<d>MLQc{97mFMbAsLG=
zG=1LNj=&u;4b`psc+>2u>e+Tcr=-EUGm^^JS99Be*TZbbI^Ble$m+%euh&2{U1=&=
zsodCya?UU&h;h(Zc;>rm@#}6A#*NL61ogLel$^d*lSY<DRvV$iEXvC|L0QYlk+P9l
z%W2v%=FboH+5&s~MO`c|XY^A1{G75}a%zbVzwdZ$XD|kIOSC*!h1f?_B_P@lodp2T
z$XHzRqw<{hDXOKam+#Q9{V>09{v6Hn@Ky$&b^ThOl*u%n>^Z%z7p~?fOm5%Ors`w?
ztY9sf353Mle(tLchCu98tD-!`du3^CR2dnoFK}5qKEwEZ?O$cJXy3!0n*98=Zci}(
zo921!{Z!&%cFx&YcHy{gvkpCP6Y77V@{MHAmB*uX9M7(rz%lwH@WcC&TYsEF?vv&5
zQJJ%7Dh`j$aD3Og%=f4u1OyE@t91R=UUE+X?kEk<Pxms4$7E##gM(BI45+{;?2n(0
z4iQk^*VNWhfNZ<1tzY>*X{TxOw7I#tz*grd4xz55W(@r7K~+gf$(pZ>acr<A!mw+f
z)TU@nlLZ={GgZ#1+&3XR;iWI<daX^{l1e+Z@TK$?jD#`N!xRGlDp*YHE-q9WVIajd
z4=y^_iw_Iz<y!?cr5-Mme{2{Q`!hD|9a5;j)Br;=UF}l7IpuCl+Qp%*(DMw^34fZK
zp2}heI;j#TLUN~5%^88N&7{C2KUA)Kfmcx}(3<hTJMPcnXEO}G?3aq$J?YO;9({Y4
z{$$K<U+?wn3k>1+mY)48*XfmPy+Uj0xT`r$`x~=(WzjQp0%`OVJ)e?HY(3+5YjjW1
zGByt@nfb|x;8_eGP1<>MM*G$(RUDm!GRn53W3iiCOAABvK6cAf#+EnkE__Ol{9N2&
zRe$Cu@Wht=SdRz&YJExuhu-#J+7CS$L-(bwXby8GaNW8`N<bLE5TL_+!ttLXY<DSj
zLx5ky=T2!l6FL-bL@Dj+UBa;=8mF`|ce$Yc-a+MCUS;x8X0;1bWC$2T+%L76^kZlG
zJ$V8Xluq+3A4zp}+_JK=mb?AZ7XuaVr`sI_M8xr#nV*H`P~msx*qt0QtM23iU6Ye>
zb#-m8o1Y({e*a4<-lS<D33&hGE(~64Si$0Fm6U;Xm(0L9#E3O$11EmjI<03$Kjp;6
z8g8_K+KJts<jG$FQ9z0s<+n{)hs^=yjQQq_aYndoj)`QqcoZZfx3=_koSn<vd_$~Q
z!rX&vU#ZsNXZ8TKZ`(VDjAr8Un|bNAf{mAm$0;r85{~6;wIIL*F3)Vz`(08sj&+E<
zHvc{wxaiiTH88+dL#KxgJf=N`L1KZVW8C@Rh)sfZY-~NA$rgtIhtS9PG4DE$6N(e&
zEZn(^vrxKUG$zvA1-OTl{P*F(=`g7R`@B`qx4-p=cLjemEx1Wz3x7%p>}6@R&_PD;
zNg<wuP#X@k>=N>tAhWY$Xctu1{@ihEjjA8%=IRemAr^|nu%<(1h&t=d!Spa1L=$XU
zO)3g@m>dM>=-ESUKq`{HZc}*FjMEQeyby#FMrxT)(ueu+_Rj*}^3A3gez@0t&K|$d
zRmW`Io#?RuZ;$-x1Q}Q*<@nZHb1t>}V_~hdX)iA?3to(g+00gcA5P<O9%QQj!bx;b
z?QAZ279nY@ggp!9;jFV}W<E#swoS?%HytpSH<w?h)|QkoB(fL^|Ci*dBA`Z6(AsD+
z8~btYYvIL>Iek14Hr>)>VavluNx0WULX(|zmEaI{S8Q;cE&M=QHUj^yL&QugT7Uqn
z5JI~(ok64;O{+O&X1VjHr2-U^dHsu`NlQ-O7^2oMBowakopdz?K!lmJG-GJipe-T9
zz$LF3><7PecM-q2CsZieV|?TI<WCn^HayR}-i^2a$=dE=JaiRAe1rbQ`~z7vKUr&v
z|9-t<^_&04xIk|{s($+DuQ8X;raq!zyere#Vn5&BftDzPr|<eH1Lxj=(sO<k_|)zA
z<)&G4yVv`jkgmHoE}5-O<@kNAt-;Q4Lu;dmV{RIqMswU~_>h8~DI@z)G|QGv7;7lv
zP>A8En9E6G9d4hW=yv@E>B+uN#KYrC3Hi}|du_=luer{KhoTA%`69ibiKGU}Pjn-=
zy>GQ=0Ng)uN3$f{;1qU?eiwFxO~HT?T!z*|Pi+RyZ34li3Yabb*pHZ)n1X^o>t=Oh
ze`Rhm1H-WObt4^JT`x(5Xi@=54-XHeSbRJ@NjEp{y}hM8Ww0mX<KzE}ov1g;=r`Ht
zu-aw5$jp&O<(lr%4K7Pn(eLr)*ZmMG-E{TaUF;hX_q7_qcEl|k$E-n1lFwFf<F71N
zL9yC=2cGfewZOD=cVs4-cd&6Y|59;?PiInCEhCMaa<`<d4{f&BDxO~^M1DlV-t4xr
zSw@02?G_w>=lp>o2wO%qTf;J$lq+F@qXMdyd257yNcit41#&u7m-MepNJjDmWV`U;
zzJxb?!;z{N`s8~0*=WL^xYpTyjxoK6W%F)oaNN&a<(V&Qa@@bMHTC3|wIpv-70u;E
z&#{-3h<VTExK}h%=Ps$lVqi0$>+b97I>9K}4H=Klw>Lpx>~MH(Y-|p}KkMrknd~LA
zg4!II=GuV7I~-1RQX~^_6RuZV8;^p#z$`9U!7Ksxiaw+I%BRiGne$ifzPzV1pK8Cd
zeUmc2h{{>t<R3bk7LRNu3=);_a(>SciIp(_QSK4NCb4dE(t(CQb?^j5*1eSa@m{BS
zp4ULrUo>`B+sVJI$7`f%g+_;|r*9cL$!y#(_dCPIb!G#f+f{s&1Ml@g>faZWr_4;n
zFJIo-+S;~U&0BMFa<ZB$x{IXt$I;;s68671#e(x!9t@&Zv_J9JFI^oa&4N<&e}+)P
zKL2yjdc2qvmy+VJjMXwc(MJsm`x}y?j&I*FG3U{d_Z`m}$CR$^Ba~3|qsaNfX#5x9
z6_*6Z@1Y0;-gD}e)>?6YC$>=)x$vApTle!QSqvEDrx5a`xZbASKJqBNm>1vf(~W3%
zE^F|SU}P@T#ukj=9cNYLj*@C>ChkJ?TS^*=7foY7!qm=$819<v54hnF@Rgd+bVz@&
zE&nJ<Hg%CT(;N5Ey&29ub``%1Y(Z}PBO~qJ3ABA;9_$u)RhiT?W1W2^(EJ7G8biTe
z!@?DUl(G2UI~O3&&)tm;R7Wtn^YnhESJ!IR0XdR1P5cL5S8TU5*l*Z?$nCqPGH#Ro
zw?{~zt^+TL4gKuMDC8+d3SaPJuHPX{R>eEJ&ZYQGZ+WpTid5_KvE5d$un6hc>Ydhm
z=Dg^0Gj={*HFJ*VKc^K!KHp(G1p`v>IpAwmgH%;>|HIv#)}iEQDEhGUNzUg<TV%s@
z4nUBwplidQA}jh+3-@Kpye$H;3(UDK)I|aVwqNFVW}twvZ#hNE$yu7*0rrU%EoZ1N
z<AfaMBjObhXJDIqt&W7KRb_+-cIk_`=E#U5twL(|AWJ*ryjXgW;VyMac>UPCmVcJ-
zF8)e&m*P`L!aodRw!Oi-$5?Jm2Wn-C_sPT@0hrNcYh_gA&R*M>Vo^uUb+|{8@892*
zVe_Uk%O&?UJq&+zcWP<%Sdpl&I}%MUGxM!qV*&6KD62prvhf+^oxjS#ca@D#rn-@T
z#0cHKq9myqI{Jf3ATJRCkH;0?BlLpo2LJyR_tjBVMD4pMf;7@C0!KPUq(jm}rywca
z-Hmi3-QCg+(hXA5Al*oJ$KCjS_ujRBYu&%^Sq{tL3}@!d?ES{`KJR|4)pW^dYhH3i
zWJDsFN=89~=&<l?GFGWAUq5^wz}h<PKDGq1Nq0szkLfGf{o7bRqg3#%<%R2{PO&F<
zY<AT#n9<g^34tTCHm0Y>e@(LUO41&QuMu(uq2(2qKY5dLQ&TiXboO^FJdYM~$a#)w
zn_$P~e%M>E;keY(U?UwQR;Dztj62&5gz!|H!#%kDwS?zs%x|nbv9Zg4Zd5uXB46+w
zn^rFXwb_vR*~&P82*?J|1VjY|OLmj@8bcy%T5>=qp4xaB83i?5n7mXy&3f_l)dy2R
z_L>D3i4)rs*l8h>shBNm$JAP$kjybo!^~mu4BQQLnmpF-?k>CE$Y)@cwiB?W*`@~u
zB781YpFMN{tEN7+KZA`qC9?ri(oXsvER({2*s>+>po6+uOG_SA#%rjb<6z1S^BVT5
zBhix!@s9q`G)|wsd3i_LAs`iiGYPMf_$Y%bd|lv=9~hWmk9N;xPR<L>>4O~88|SHr
zf@#eEw#+#@ceuq-Mx;K+$)@ST7f>qNeftfpLcO$%LP(2_mF9%{>vaB#01aq%0?n}J
zym#<-DYbbzJWbgPo0kWTI_`}jy5{ur{XHG2yb)rzzEvC^rBxj}LPeL_^MhJ)3HOX)
zN&%4(O;D#XHVd`)d3hz_RCO(%L`>~RbVb8n)){I}8xq=mc~AxM%vFGlQ1)rN#jarC
znD$wC$N2O;VQ@qa<dMUJ<y7n`z)G9HAs^8=?(-^BgFr!+15FR!5wyfj_S^4vO~m2!
z+f^7zFE5;ckII{uFrU>loB9wVPJ-6x*OQbzQ$iwe$>k8RAg;6`YWMn^&=2wXF)Pfl
z;a&!{xlogI0c|Im7?W!F$FxR#A??3rpn;Qn@tg4r(X_^2+G@&|OsR6N55^JCUVG4w
z)Q8tqeU4UkjhbldBI`x4%At*?v_(G6%GdgBE!vT5ZR=$JIEVcq?iDvuGjgwY(!~K0
zx{vwlj)^xfY=7a9qXRxg2frENInK)hP^ephDo1QL08w`k9$L4^!L}S#e)$6uH2b+l
zRrs2;<nhuxKts<PHo)y2*{A{wY{cKgo$45%NysrKFNl8BMKn+Yye|wPMnMXin68{9
z|AjwL1*s2GY_nH>(&SwGQ_n~{Evnvd7lTS8U{DDPZD<oaAV<<n>z;&=%ByFWlxBZR
z_q;}M2q4K7>|aCn@DtA@AH<%>uUAx4z&PJ<Jw{KLR@_0A4xVWaMTc%I(l{RdYI!{R
zctL0ZC7m5<MxNBOXwPA}2+gur;r^I6eReee$L*p1nBbka{LycGgtH@Ei$S~hNNg?e
zq_nC(wky?C;#(8VyIrr_eUlSkCHW6&3*}FYY45M^7f;%$4gu8ck2@M9e3CeS;ZMK*
z@hQT)n{4IX0^_p$@jp;sWonYCVND6W%A!*zo!yWb5(nHaw5DADsbv|bRh;J6A(9Pf
z0Sn#>79UI4R?5}-7w+l1d-(wR(?$1&z)xO6OoCwJ+SHoE3?kifS*#$+ut^F0!z*ky
z2aA+2N&q>n^SG}W6>x>_m7^Cv%IL3%dOgGWY*b?x@q+Lr$2myw-Y<?kq_ploP|47{
zA$Ue)vgrebJ2tA8c7?U}E|vtVlpvb2gP1wUx6~==3|I3gZscPW9kpZ4pJs8V($Ot5
zPEKOFClXn;ZcHGXyir7cEuA}rqgXp-Wj~e+>me;6bm=0H?RbVD!gjI1QF_@O3Egei
zO@8dTiB_wpH6Dc{eO#Mox;~mwh0$rExVmq;^-n_(N5Z5m$;jnPPl^KFa>Q-ydgPR?
zi|hz$J2GQ`oG?w5>G~Y7X-QHUa6nJo1>fRd8b>Phdn~3$JvVRGBh=S(lS!BHhT4uZ
z4P3?;)4A$Xhp0+l9Aw@0U}H>RwX=fP3z9(YV2M<FDmSc(ru=$MEZUN2^zfY359=+o
z<Dk3>B(Anud6}F8mqDE?87)<jOsDJ$f7FK~f2cRvQ%2pdvHT4!=5N69O2mj}HsL9s
zDKTUG11Ip=-eSa>Z%A#)lIM}@j1O!NU+V5hdRi}>&2-z*CXjIC7FPiFu~TDiGUEim
z3aTAs7YpL06bGC&Mk`qXd1^V@GhR59SXxxXXb<6sZ__lq4>Y;G{QIXV!3w)UuczzM
zgq7MwA&}?fJ&Q1%MnJeqWO$oIN@Lbyki>pF%tsAxd~#F(?79h2rrDZEJJ`Pl?k6xS
z@7)~=UqKA)t>Ie9<?2f;%gxFAB9+v}a8fDF4+U2+T=c2Mzixc9C!Ob0945U+c$3^v
z7eSoLF)O>bCo3wIKZN3?$yQDqUP}9L*4-ZYgE-z4w69{evz&5PJmLWJx<bll-ZiW=
zHs8$5Z~clL2Mogu#lNn<s*#x)nS)}yUhr|%#!0Mgsvm^~`QXf6Z8s%gCk^|1@AK50
zAM&&uwIi!$TLR3qlg?Y0`yoOh;8Dt~ha}-H8Z6Mm3HRmkZ{iLd_&{rppRN;0chSeF
zuOo)>eTQE?MY{?COF4y`)KY9>#>HHcvS)6<$>oZNsP2X>Smf2}y+l<C0LRN1-C@RL
zpEcM6jGJ2m?P#sgarrw*@vw&2_W*v!Q2yJ#v`&j|6XV=cHzt6_XN_ADSv_)UI{}%4
z2j)w@ho*b|@X3_r4R_1qrRaDN?IE!Hq3lka-rWd#JPwWbRSupP#$#C^J>A+F1&4%}
zh6^F0g->3hc_ZM9tquiX?D39-o$YA0PU%^!eBF?vcQ75P?dXwLW!N`TBJ7ZC@46C}
z|D)diHZt1yLtP~pg)TK{xjApf=t8@X7t(RWp-|9#zk9-twSgsV?#ip3)~+7OK}s;A
zS{=aH=odt0To6kO!3I=_%4|J;snj81=K$Rae2&ky&%Hnc%N`mxSk}8xd#%BBRqBe!
z{{y}ipK3D`&G=z<?!yZ>u?JW$)h(&B`s6~IMSxgqHr+)tyYya2-tO&~<$#;wp{?+H
zM~t$x!eUQD4%tWNXpiyd_^2x1Hl)1N4gF23EsMFfE{%+t^D_b!Z0?Z17+5du!$qvC
z`ir=`FQVM%Sswg9vUM;R({4p1r)4gTeKfGz$#4k>t94p3l6F62OrYB>Pa36VqYhxc
z*%Kf0i2@najC%Fl28fumW$3vg1cGh(^QxNRfe4qQ01~{yRe4%|Rt*v~3V}elDE1<c
zJalpN6JGt|Ozfwqu9CmXr@dh#M|Xg<U1jW;P4u-f?2yi~Fe{_N*EAZhiszIIxG<;P
zK9i@XHyD)Gvs%w7Bn=-+OTzNE&RXY@pTpVcEo~T;ON4+bk<7CN5gGbTN?>3{ILCHZ
zaG*CSOQR~OBD#|+;|fv{=Z&6H{SM#&9o9qQSo<6q3wI1JHP+J1@`@Oc^&O<zxTfC|
z7KQ&-bPdf*A|}qiJtlP5>AmaAD;g`bzq5W|nz1q=Dm&16QmY{F;z+0Vu<nn5ud!(0
zN1JWea%}p;ig{%{9M3x!uDEdo1ec6=cQVQg&+O2lmKKDLZ(VA_U$o*!-K|lZ7D9?>
z<KG_JM(QVO<i1SA;6$8sZizokveIpx!@So`B;`IWp0gOJuN~3te`c+lN1Z00+et<3
z0}V+(IIkBy6uMm;J6M;MfAxwFH9Z4mhYtN2N_}rO9VRd~(pwmTIkKa(^3TX%WquQ*
z5!nNmiuCEDZMe7+>6TlNP@MC{$s^gsgy(32?H7DkM75BMBXHUCaUKtw%c!YR_!BM<
zEn7>~YFt<S6Woa;Z1$2=#K$e8dvyGQ$+-&*sv>qO+$^Dju+tUNg|hFcTnJ6WJ%y?o
z;FwZP&lzGNK?$K7*!rzvQr4G>*}oKakCE`j)Hg|llDs`Wyp_Uv;e2)Z1#glxIhVxN
zb?AFAO#M&2M<?8FApcNkez!fYWvetsUMlrvehXe590M2ZF(&~G-wVVbKkEh4?lS40
zG5*5_Pqm};mrHA46UtW1mM=J(Aawq^$t!$|a<}P~Z%J%i<so(5&jhTEU{NkkwhS=_
zR2lt;Vj~241#h@YlmZ29>$Rnyfeyn^jWfuhx!koApr7^xf5~`2?MHFZJnmm<BPwT|
zUe!{mJJP|w>_gL=4X5WWPf9i}?(eZ+1(`G@!~sLq=g|x66Cxr(BCs)w9B0pfpdIIu
z+VrBVb6`hy5s2o`dcGEGFw7K_bQ#d&v2&!DY46cv-s@a58#3GuR8gBSu1)WzuRq4!
zd9?A8jPw}PS=aC2gU0}#2LpyC*`59^9So2g($fo9w3k^zIXhZ=lxj@IPw(TSxLwKn
zuE`s1DAuTz-uUe|!5uX9SR!fnq=0D&h>Td6NHpg6!wYNL44Zyc1{&}oDLJxP?C|5(
zKS%lnl<>@Z42&t>xQVvjRwt;<Dtq|O$|(TAPw-L%S}X%eU-$6yo!GUI2f!5(QZ3f@
zN^jTq_lq2NWJJggXJ%R~e?M$?pOLPqSm()WSu0$_jX1XAgJ=rYVtaq*38)4**sNw#
z1YNZHVR%+MYYaeW+f0nJ0jv#H;Q?gx?8>j5<%*rV)k&P{_Mr{1^f!=!cSGmB-wsAG
zGufe0^@b<!D1AfM_9=?4sSO#zHWw*m(E@yy-0$%KER{gYUDb#mG0)4jB9PZV-|=ul
zp&spKEHn2U{H1G$)I8gRs6)&i=%~a@?vz$AL{<yAJn9Ye+?XAB3@p=L##2=FAq83J
z@rg{&dc0fv<TXzVtKj}$A{nPKrRLfIuoRk@1Y8N~b<9>7!Ectgen=&YK$6bqz2a-g
z8#nmK!YxZzaJ@*-4#t7In9WkP|IGyeXf4b<tF!Y6b=c6Kk8&5c;H!c(gG|5YK5McQ
zp8|>*xu|_-vl-!afbo;6wyBLBj=P?7|Hmm$BU8ZMvP^P1RpX}9h^_?YrjG2c^J856
z?3c=PjwR|p^dxS!Cfu4_Skt+hALvEW=2KO7gFI)j0ZRC_%Hg-H^Q&3y-VrI}kW{iL
z0lcqAD7h7sL|@`w+67b=QU(M^YiXZIfS9yl>0bH+_R4pXduddBYG{;`-UHK9U<<r#
z@7+uy@cJB+vR&zQt&;gQuI`C~^1uW0j!2-38%R%dTN~0XpYdG!hf`9Cz2o94NhL?*
zVd$WX0oV5dcW*hk3yyKFRG@?fGtKEO4An)exKuwHfTk~9SN=?n!f?o%Hp`~0ksHX&
zlF>WR*=F5p*&4C10|zC29mJL}XZnmFY!MH=tPj~4%4iM>Hdx-h$AiPuPe|Ut>Q+US
z9X~3GR6=!6+Cv<7@TEOJP&xh-e8B%H6kU2F&)S{I?k!Yj>XI)$Uqtdr6X~SAg6-}y
zX2T~~AqsLQNQJ&Hf)*v1Nv=H>VKp<TGANbrT6g}!8$uHa5g8$cz|q!Zy4FgWa_V23
zf_3IZ6wJ{mTb<mzdnw4yd9(}>1PVW7$eI)!8g(&p%YH(-L*)Z5Y51;yW+L`5yVeem
zgdrE^${%P`TyrVv7<4DZ6Qqb6SkbB>Iik`a7VjGE#lkFd90Y5!niz;WJidPM%-TX;
zZ<s*Qq|>=*r4c#u#ZMTJ_XBfX;lPG&q|*4)h*dl`t?x7@m2Xlu)K|E8#81lz$iBtt
zs}GQ}^L4tLKIBi=f>~=^XZKZhw;w^efm-yTW#rcG?~15xRWt<wN$O`-_{bL*_4FRE
zJ(q4R(G>MqIO)e{S3>KT6X4cCmssy*i}0_r^h@y*E>*e$j%%Eu8crA^J3fe{>bK#w
zz1@3YXvzBU1+ugb<#=sf8I=5z%jdXc%1)y`$}i(_5ybWY(8^lp=e#q}$Hwg%*7Fp>
zMKAHpnyx*|fOxQPwII`;*!0AdOB#sD=z;)yvv=^#sOMhzZ}BK!`4TbyzMO%tNK_sj
zo^lFJVLdz9AeYE!S#A4rIl0|LqvQ}_^1$Kn&N@$EuU5LU_5m?qC{!@Bhu@5r?;`=P
zh_!D>r|)-bQ0sWAJ(qtR{%*3nQ(ErfzOyl(<na)^=gmAOjJf^1g#k^$dJ!TUM*`NO
zKq(*`dkCi2aU=IP@1lD9<{K3*Z&xc?i%es#52#frLl~ETPE@EJ>1UjoEd1r;v>!&(
zSG1cMWUPDG7aR;u>ImqaIvVjgj3)d(bI3)ftbGzfMea<K>zpndbqOsSo5P`+9#>EH
z5#91}T;?zta8USujvEwxxX)@b^Lw<bA_;4r`5-tALs7zz0^|+Zg_J}vQ2Eu3bjhz1
z^uSq6s2(3$FlRNjrjk1^osll*vT8eWd_tAPk9%<xS`>|2(O`u~hu$=h9)IUB6V9DN
z$d&A5w^jvwGSCTEPh(CuURh%)I8U%kbHAwjQ*!5W{}{DRI?$|OA(knNLCl)IoNx_N
zbKNQjSxbv{v#PIkj-x8IL=s}WX=sU*wy-d@{#X<5Qw0hqvL<ga@SPF)KK{-AoT-Sf
z87;(mb(x7iD_@<gPayH~e7<(vycX1@M9JTO#DD*v;fwGFL>}Re0w%4uoWvi3E0;v#
z_P6#+l*SaLLEz~qvpuK0AqctZK3dOheLbZc4hH~pZKT}dhOybqP^79!o&xc*QV>?y
zO#2dD4_qh%YR;}C+Q6INo$1G9Ybhju0ue)-d@nXJcLpvousL_nwqzX>J(s=*3Fj{a
zL!kJQ+E1a!)MoDy;;=A?gk)slxguZyxhFC*(cx#MiTlP!a8a;;qKF-?EPzB-5z-a7
z1b4)KmG${YWX8z*;T7x^oL{Jg%10pYeDddYhU)B_syAjd^HN~ZP-Gngo8k}Y0FOC!
zm@>(@UDX7^$LAj)BJS%0%?ol<^3X35Zx^{_P)vALbK_#Xoc1bxB5YfL#mnT|lB4-k
z&S?fCnSAM5M~%D5_&n^P>qgY3sr2jb5#Hz*E+&j-`$cb1;w#P(8Bl`cbA?QgvKt;a
z$gJIy8Mui*&`OsA_eu9=Z6n>=vd`D{wJn%zNBhL~D|h5gg0pH4Ms237Oa>nU+y^z5
zIp@~PC4nC{lo+)@Q?{4!3q1{5ZKYfg?09ptq#@aX$(Y~pAppRYKonM%x<;iji6x=K
z#p5L~li5=)3eG0HLp2S++vR|A0VuUQu(N%YHAWoUW^tU515_l!lAr?Xy00f8jweO7
zs8NvP>9?(<HbsH$^<2&vFX_T7FS<m(>mQ^!od%}cGK2dcbR4w2Sm^k*Lao1)Pg+Ba
z7qAcH<`w<L3p-5eM)h~I&d=buz2b+vdO3H8T?Rp<Tq>>_=<7SA-D}2LRq@qOX3b1a
zWcD0#+gr8S#XhrF`_Ub}HZ2X!WFqjnK*!9R(p7ceMb?<S9eS{|=Wjxn_~v2$BSQ1A
z1;-MAUXuLwp3!Ba3OqyVRFk|WDn}@L)j~5PZL3@~@Yp=7$4zUzSG#8wmoD8Hk{B61
zqvr4e8ub+JlWw^6cUF0>?)?_^72T3D+f02wN!<Ns(!~WSAJobZy%i{1uhMc#O9~Tq
zvWEI9CFS%n$i(4#Yx&{41PkASDqL60h&C=y;6n&rN5<6Y*4x&NhSajE51xSgmP#~U
z3MOO#4#T7#yyHmRDD|WyiWKZu>$`cr+qn0!r$uECz5%Kurqd6At1=5xNC0wR9N>7D
zEhhh6!3yM+coqSYHxOpU|ET?W@T<sM0*Wg6btuU4YVT@VJ^hbo`o<x*t?tKA6l7gL
z&^94#G@tuk$#7IXS?udX9aK|emhlfEGD2{g40vO!c+Am%$YRZDfd;oHcn9i(u`=2G
zp{lw;XHfkt6uYt0W6{6*z(nv9A(r4zCPpIe<bquzI~!q&0YIYaXL{yU+an0c{kcTW
z3?q6-9lCaym6aRpkL9X@&x&dp>oe8!yaTHY8!-d<>X4>xT+c%LzOclM_D}aDG^C!^
znoM+mnbCtz&2+mu3Go9wig9X5w}nWy;`r30GyomLHVp)&KEk?K6wu)NsE&}VH4>Rh
zM0L$4{2dG<P@#*COmU&(m<Ht-9e46;>Me;)VX4lg?8h6VO8RmAT|qJ$qir+6vaz5q
z?>iF69^U>$XlB6=Ta2ITQD}~8y9xFI^<Kd>ylh#Ny?wq}+0bqa`w$Pu&cBV&uP(BW
zCG6_o0BA@P)pmL^<!oKq(l@IeN|-&;mfxIKzyuT-S95(zE2%=wL5e4adl!|Ggw<60
zNWwRK9VKJiy4wEb@*T^2cY$7&)T(gD2oL|z(9eY{ac4yh0A2Ia?JMMLr$|oQi0bPH
zDVBG)1zgL03Lj~#U_a3`KkM=-;(-tu%~Le|`r(EqZT6ZFuwf>@w4$5-*>Q5YrNd^A
zzvQ7TE%L1Z+F6WU>+xWrQ47MNN?-spz^J^~)bY;RLgje&Vn)mQ1*lY*2y`Hb?z%Kp
zr#YGMP}}>gGCKrlK)%5)m#~hGO;SFz43jtBZ~(FyS<h(V59-jCi8kT1`ZKbr^!Z;u
zie;Qk<|GC~65AKnEV*_89J8_<dv95hfA%W%1M4$*tvoTw_0u&Xr_&=bGme^6KwD$b
zFdqy`>Pf}2=(3B0KQ$;yoW5v<Z3~aO7k2W?QvM6@*-z-$3<SQ&c)LX1B#G=t=@g_D
zCvA=Mr_B!u0Bo<T{Co-VGVu$3qYHB>-LOEgg4FRP2=d#{w?{)9lr4WryM62D4+7bi
z><>HNE=J-B>&l{t=si;Q?LzJY=_m+!b0>%W&o^F0QThapdjf2MM92rf#4H6x^Ky$&
zNt@%VItf_8g4i=k4&NK(#{FxgzA;~bBseaQo@|jlsrLha?`DSv_4j(m9TNA81+B<#
zup7$15aVinM?4|Uh9x)#NK6`nk{>}H1*$^2fUQ!Cd$^L;KcIrm-nfNtuCafov!MGv
zS?$Vm{J=o^ta{X~!JDW7SV}+%c&hQm91W)F!+<i%<HCg_K(M}ko=`kUet@MCqeU>J
z0+uaM)s&om1qrEYarLNY3HDX8^=-xtMxwa}0Fo-+>mPse`#aY2GGYxN+VAZ@tnFvv
z-2;4Gpcs)D#!}<5^&6#a$fxJ6Yv9QM&YpT?b=zzvLN<hfI_NRcksqJ}{VgXHcXe*A
zWwgKY34^caN@K4G+^){|($5GvWst>JXXzehPHpoi2bHPn%4v78G#6Tts!Bp8yo5n6
zaiafV5D9~nz1gc^Vd#3S`QW&_bUf+0W_a|3Ac>f=DRlN!>8zFvqshIxc?)urHLL_t
z*LC{hxDNpOK1LiSo;XD&47Tn=!@OS=r52oyTgF2b&{m+qQa{UDqI9JCCN4*E)-;Tb
zV6taM*+bRQKN?EXP;(uMED4#%Ufq-Rm5afR5e6rkW$UG}EZK2>v^70-w%J{>eq#$l
zs1(PYYRwW|<6oE{`dp3-8#&oriEycPHbN;*uuJ{~<0qH-`9}le#=Fm3r4*z5aSjqM
zMZ>;vj*i!Z-^E+vy;YwFQKSyUFDh`|G~n*8Q;C!MEnLt;>Fl&wj-Qi*a$1o^k*y7c
zVoMTBj|J1o!%);G9hBWE{d10g7n3&LpY=)?C3fV=m&y=08QZtp10(MP`&}tkqMpVV
zl&fDa-SNQJg!41pJ%P`#d~v`F9&Rm&Ih<Rg5b}bVXx%buq?Cl=2Iu{Ij=@+N03?Du
zBb9tguX^4AQM~4qVkYt)D*3Yt0&>e##1e6dL}a`GPF4~{Y%b2$!&)RxmBb-2PSXaO
zeOfBfbvJP5fS<DQN-$pzzdPegs1~jKampbd+nM=Gw*ZIswb~I(rceRquHg}Z)g1gs
zd_*KaL5Np}NPn{>6xm&|`ZIccMAYwi2s9M=)0@W4dQfj=aR`8YMez+bQ|?D1aU3*(
zCIO&hAF$01p;-%)wX4leMkrY~<;hVHL;?;K>lmTfUYM)#tN{NQm?ud;Au7@53n`r|
zUx7nV*$1$KWIj+;kd*<okwB+be6-I7y0uB8g)cQ@wsem7?PQOSSze(J{lo(<QOzTv
z^}s2XS3e+ufOB?Mn^OdgQ81PV!_vZ43mqIdyrX@c;UlLfdn~u^LXXJ`htb4dy=Z3$
zD&k<T13nla@2|WbiT`S<lkPJhsi+6CLhYnnc|nRU9d2+H)WPgS1;vaOWJO@UI&qky
zl8i_Q6trm#!Y3tG^XazGx2R5N*&2SGV$ZZcB@JjJ!)PMl)6E-}lmdTu>eN+oIg-#k
zLj(mT@k&o(+wA;kjF`Q>-;5?7MLl7D%xP3{gsNb1yr=bxntK-~Kc^|<aNf<XQQtZC
zG#w{u*imj8fK5>Pl77^Ej0&P)$SA+V3vz5m+%(nfPcT3lF-0=;#m@@1-qy^TqA;Rf
zjgKy}?*QeCgBue6YTcR((QXte^TQ~WRbQTeM6XbZA+SvEycJSNhBNL?mtD;m7JcM1
zA*|GX{n`YB5s5*9@?RZiD*7EDkiieELz|z&KX0uRvdNEY_TJ8F!MxhS>XsQEgQ!0V
z7y@bEC;|2hsdfiY6Z(WKY2!f6*e#^nmNBKUmIci)!sWLk%uTA!(376P;oy$>*k8ok
zJ+;c~>jaOvzA`ep!vmh`>CS@^EME_ZQ(rf&w~8#$Kgl-Y!-Bb%hxS=dVO(<rosu3K
zPrnmX>b-gHG#t1Jq$FQ9zadNew#UTpJ*3@AjcJ_RIlSah(p7N0?1{%4`Z2p#r>eZ_
z45nuPT;Q@5&M#-V=#C-h1tl{ZJr;~#jzIp{s^Jltq@w4OPNgr45agSM$$n56jATf{
z^7<!rg^bZ9eOI{PoeVt6DJsmyy*I5bUNIpVQXP$A{JEgY1j?g58KbZna>faGtl1+i
zI^qpaLZIrz!rJ)k6hQC1l8>y2D?YhLBdBa~ZkDVe=S?<0+rB-AUceJC!2uXD@0^!!
zv%h15tkv3qdMj2bcEHeCYpsOU+WnXQzu^fkOdE(|;!OiK+2V*t=YZ>@kdzvUR{EbF
zss^pdZM=afU*}=85q&t9D_#2;qa8EwL7mdYh)v9ZQij+vs;9KU$sKe0$I65oJOo3g
zM+zjrtDpaLCzZ)IZ5!$V+aecCF)2UpS;cxRFQY>%FzdbQ3y3~0)dm%GPcN%xmt=8m
z`$?cmu8%jJ1HOfwNorN+-|0esa!UWawM-TcZtQ;|oBok~&(&vwem<W{o9P2-KH$=L
zgIWQ3bB9WrbLuA;u>`uhpjUp{jOyY7JZQ5YJ%l2wg#Ic6(*~|B@G{K7f%>16CQ6ky
z38FgSkpRia?dvab`dr-H*F%;boj}<OwGC)c7nGEU<lbCNjPRVIoNkRAUS9)s$R}~h
zjEaiLm>BHW&g&Ra(IT*??N)tgi^Hm6Aa0JxUj&?x_Q5{rFxr6m6_Hycr2lPnn`ORF
zaL@}OoI^a+wnBfMF(1*(s*;7VcJ43fHeWh;I+_Jt%z;JU(0aMl%Q)9xMAxQZx;|Dv
zn@yPhPxO_-Ql)HF)ARmZliw3CVD5mf%2qI4>Gt{6lRzRM*n3Xt0u4O>cukieWi9tu
zaCMrl`)okivHhq0E;C_RCqbzMw|b(G#Xi(Rx{)xur*n3to+MUYBqM--gFy)Z4l}8R
zXtbiw1BMHZ-^}0(Xf71W@m~TKOfi2=?2t&~6B&;HFCh>DRoMp&uN%yb5~Z$dQIJV%
z121=`>-6)-zbh4D&8j<5E>Z2y5(z;-K_Nb}1sxw67C-+|_C%1Px5;>%U_JffgV{{m
zSgwQ^tz!HrfG7aKVFgGr0BMwn#^dIO)Zy%6K`qSr0GlBfGmmxhqY@Y!a3IsrQ&^Es
z8f0+VmCUvot}S%{{u8!Mr|x8Cb1U#X2m)XuIR%>761F?L9gV@TL_XT%e|oUX#1ys6
z#W^{?B-{?=X0w%ib7>wo#(C1Izm$|Ux=qWP{U=TKXXE1I#Z*=Q)IFoc&dJLI`kHfF
z91IK$KmzdmMnP%Ag@VcF_{*XT@G@d0iq_LtM6xv9F}#KifME+b_){qd{oAy{+6gNa
z)DF0n^iR|t(K8<KpYajT-l4<$gvXn8Wx%U_sJyCEPi{)EQb53MjE%~(u@h&g%h6!K
z{3l9l1q=J@)oYI{ZJ;K%aCq@nBJK$ER9w33r8+K>@$sdB<?An8?h5nM$l2I%Ql-T{
zf5v?h?Us^~dJ=+!0mN_Zcjqm12Y>%Ufiz?K=>w1=-0a5FBJPv0LjK$s3pjO&uUR6s
zhVLS)I;~v+Ialrz5?}v)FdZgdz2q1KsS_o_snzHS2M1J)m^wupjbHkZn_jeAXF!q4
ztIf{b+8V&x#M;ODSS|BnfVGV1;`E;!FY^T|HR4?-=|h_kI`71xLzw9J?pA3jBj3Z7
zF3_`S;k)%<r%B<kCcj*8Bz<cBkii7vwNE0H=P7_Fdv||-OAJ;vKUn>_>!Q`AsH9Zk
zc71Z%N{v@t;zVj52RIPwAn<zeQ|#9hIb3u86yDLpj^9)D>P<yU<S^Q<_pAsehuHz4
zCk2@FV78+v5^o&)wUYPqTYTyTD~juHd~;01U<oQ*Nd%7W)nA+*_WW8b;2KVq`JM<G
zC5q^$MgJ_;iVkZA<acK$EPBK@JiISe2B<*87Yve%rdy9zrH?i??`CIb)g5N_VK|7<
zY;0_9u;ZVBZWnh^)CDCEcQQ+2m-Mf?Fa+7^=Ca#M#eFpI9X>o^H+=jLZ2alRLWTi$
z?Zepq7?U5F)C*`5mWG-mDmTSFnjfG~>_C7G`j&ZEHNOoZl(=4Cm(i^OYsSZR(%nA5
z3V2(-zeW|APjlu4M6wtE(GvdZJM~Xc+_bbIPljzID~lY6p^GUi<AKUWaWVZ9+x$si
zBF){R`TR5A)iP{lWo5Yvebe2^01UlqwGl+H*~l%e?i)AX9Ve-tr}fl+*?}ujnEViI
z{5*vVOgI^mpSP(3ZWZXSxlktFClF~&NTor^j_GQxTERTP=>Qh*X~hK@OINt6$@#+;
zU`<j?sPihS2m(6W73MToGtXsoYSf&8{YU`g4$`V}huX5vj=hq5uIDirD(A-b&l-o?
zcY4tAb8{zWi$$DtVac4$0{zzxAm<`&Y)k_XJ8bMORQCBcbkD2(?Y>xLO{lp!13!>e
zGBPuRu(3U@j=xYcCj^{yhiACki#Fi3tbt+Wsabk(SFD+momU?&`8)k;gXEx?+0PcA
zH04*1H;EbD;y_c#o%j4^d3Fr;r>ONkS@mDV1%-C1jZXV*v=6nn>31#{GB<kwI#>ZE
zvo4)O#=H&7qg=oz0tC=`&IeI|=_cSOp$UK(&|hIL5j$>i_f+uTAkv<RZmX_-A5O~e
zF=b9bK;U*UYq-VR1DwmkxcXvhjYAE<qK3sdwN~>lSH4Lctke-b3_kgTk(dMf2sqTY
zJU_3$z~lwUSQ*rJa7rSlKOt2szjPTRUj!FsP^^tz%}!KL9~&>a1QQMfUi@`epvtEt
z4Os}5&J__9%vP-8ut1}uT$<Wa0T4k^^nxqA-#-HllDqXDYEbtn<JNqe57yS=?p+dS
z(s~~+EN!4L%GhK_x$2hf<S@MpSk<rV*Ix@twM&PqZbv|%dhoec<?o1+7aZj34y<Zg
z0(6IJXhN=c)^YjzZ-Z}JtNg308@!*W_z-BZ`c1XU(m%=Wj_(>x%2|vDDnp1iCZ|wz
z@zt9Nx=)CKFPCf2z#5aXgCJ3O25PC-3n?!H&%<oM*M%7jO&J8STT}TKst1d$A4r!$
z+7V8l-QxpT)?YkL=l~;(D;C#_9o-Y54OHmKpZya7U-1qEi>~}puSW)eO_|{%_K+jv
z)$_20+(pMY;*9bN-?xai$Lv>JEb`hJJumaK<O^n$T)a#P8Q&g|OU;|3B@dkO5EoJU
z<(=w3P}-PCGqb!;E&ibnit9G+JNi!m|Br!xqZMI`95lNYfUr5OQV2x;i*(QKN&$)o
zQUc=`3M@un!OZ265YJ8_P-i8*U0h|7@Yve>exo$5DFcd|llwJpPG$_a$#9m<c*wd;
zWZ*HxLZaFtWA;b`G;`Mq&6%u0MfxJ1v^|FnnBR?IG^u}M3hxa#YhLG{{w0S;tdNCY
zcL|9ge8V@u(9E2Fe;~fddqcoxZvTStK9xis9VE+PaNHb<&Y=`aoXxdiAa%{BOkcCl
zhXI`ONdmH;sA1xr_pF|Mo@kzfc)jc}M~`_6VV`rYZS?+q$NJ}yFSwqjw(Bo6;P$2g
zHJvb*klzO#kf4sC2g-_d(=1GYdFe4ajHZPDOrf*)^k8)4ivDo7>a#sDgHj0_Po3%8
zT!vUMko3O53zUsL9j?&93Q7eaC|@nQ1Xa|hR)R`;tUZG5>iUIl-|z2x;3$vhho=+q
zlZI2ajfc^o|M}u%QEhU1jx*PHSMZD6Hv|r|-_qLyg2?^?go8uYI%fV)G|ZjR*gM9s
z^!;@szer3AByH)l_bq%iVYxICN0(N^Ny8Dg({jI6I&gxJe;2+7<?XEIPLMe|Pu(}$
zQ#RS>$%T(K8^@TuPP*4M3?Ta#HH2V*)f8{ud18SO``Op-95uj9Ky6liDWIgrupbwm
z7KzthqX8T_V50tOG-=sZ22G#Y4i-fV;vmrafOoJ8Vz_zUl{`7_x_^_;d)bW~k@zj}
zW`JP#+)Hfp5il%$HRpUlB8MPh6@^eHd(hYP4KREId)e>9sF?203u-j3`nOuM8;Wp1
z2Z7XDDRHdre^u#q13o<$>wh?*AFO2r(KKBY>}m>0I5WI@Z1wk_;qWbkOI;VM?oVM*
z#+G}XBe56#05*H9c7+<^1xi-|%Wv&{;SyQEU~2~<DH!9-bSAW%Y*a=ECUDC#$fR9r
zhiv3!{~356@f(FPIm7|S?FkhG$O|t7x_W@fZe7h4Q6_EiH~Ph|%E7!24NwvSnHl9D
zsuVEvu6iHp5=mL38+(@mrAu@{RY3F4t-;8tcjz<4y2%U9B~TNCi9vb`I<#{BYb(5$
z^PbKFC^#Y_BAMvWD3w>-(O{Abe^z27Xs3C?vnGG?j~cV;R;amNgdHK}RyREN0_U2M
zYeM%#*8<&rN8`6v=D&0n`){zpV*n&Bk|}B2E=@rLE2K14ib->OqZmSPqg=uu|GkR`
zB9L$1%X-MVhR2}eZdScIS?`e}ZRX@TP)#sWQU!!#eV3~jA8-%WLFaI0+Q+BdT3Z(p
z%nFcv0XpkbC<mS!AdOESI%hR?iXDpli-+DI=*E3|4%gm)W<ml0yQhjWMQpE8O|`Ai
z4$eHN%BJbC47`iGaFGr)e8WmX#Z969FTZ)QLY{Bs0OS3p?JO_|=8DkmYBWcGKFZ^~
zKnWQKq>$j>d<Z@wwlB+9gLTNNd$!sa^tUiDh@(*9kBSKIZ@B-PC%tIz4vG}`h)*H>
zociBzeiI4`v;Y0l-f!wL{{Qd9R>9(1I5d)Yfwm|rG4lUhC4x8e&G5RQp-uLoZ3Cpd
zwiefGV<4WJ6Z1cJ2mIVBEGc<QOZ#JFEsB1i5^I%!#h@w?XyQG8{#=$eb{H*ROi2l6
zAf9P!zix?=gJWiP1tI%W)8&E0|JaO+TMj|%@}EDdpds@2Upx2O#`#Yf()qe4>;1!M
z*;?)G?Y;p4+}tSbh+VFSE&pQ!)UaERT8uVUkca=DOJfIF^#5Zg%MEnAzqWRGbs$(Y
z`Of^$`)bvaDj^KfM5n`r`2VqucVD$Ms@u!{pO4D))Zr@lj|l-ki7=oU|IeLeYS6~P
zb2`X6n8F#2h(&n>Lh|n?CpLxwZ`CuNv=u>eNGX}{`*bV5^-h1}A|-|RDKvb?%9_Y-
z`P#DO`f;wY^?TkTGupBi!{ArO-Mle@%!D|LJOkQ$96>Un_Yg&K%)jUu)cA@~-HbU#
zP-e1SL6kzG_)*fD(m41+;mp{)&jys<MzdyB&J+9h8p8Q;aSuihb1msav$rX>Zdz%l
zJO>MCDhE~%eSHYSX}lqzg^q%p{CPt|0|=w5no*MccXmwl^z`0<Ujl`DtkvN@d+<nC
z*VZyejlRdk`1<?%6KG&bng3{d2?zHvdz9nMD?A*HlP`5qOG_&PEHmKi=SN0H_Kb~<
zjkSd_DM>nCNJobpyvpr&Kd*0Xt%1>j)?0p{F>Yrw^2bq1xm3NqRHG5-Xa{X?8|fJu
zW|x#mg*U1XAtzZ+mug^HTU%#URVk>@<63VH5>yzDyw<Lo-E=WILqBtAo@w=b%wegT
zrPtBb6|uIiz$YNU0O=SxEv-0@(~$sQUPgvMPEJl>OADX&yTq<cp@6QH&HzKxn=^hM
zk2{t5R!=Q_6!=W}LSmiazH@*y_^FoQie}4wih|J6(h94nBy8{Rr`E8I340DSz7i}@
zq#d1^x#~U5-?``I<rT~x4Ga!OB7;VOH2-O5b90kG4NFQ!PA<~O+#Dx6J9~9JUk2!@
zN)U6}2!~f4XiO_M;Ca^^{#b1FWRj$e1{$}}C}Csc!kJ3_kn&kG#xJMyFNVGA>p84X
zH>Ix6fB*iDfQX18lg^jg`DL{AW{$GFvQplHm|wlgNsgGqN&qNaPnzJE)TmBOP5nA{
zJX}PK74K$;IaSfsUGp9CVzqHUS%*eON3U5eOiU=ssjKrSNMt^0S2iKDoD;Pdy0mJR
z;T0AZqHO^!`9OtyX-?zKb$W0Ez~mb!`*L}(b~%sS{>SkpNO~$O-+>$D+plNSpR~1>
z!8&O6A-k>)FEcaqBtgDVSU|w5Q+cqOxu&}M4J9Q!cx-(G141$~=r9@0%_Kx0%)Iiq
zMlg|4QAJ!`xnY(&1JJ)i5W$6qmjRVciFn2h_lh<Sis}9ReKg+HOVi(XvGMT%U_^k&
zmywkv_&Qzs&~Oni+}|Jm#mkrf``RAczzC|TshO!Z#+oo;V^pvA0*M!BO&WY_>T#TM
z_nBR%LIuXn5~lU!_&8Fxn)xjR^7U}B`RMdi$kUTQI5_xwY^-0S<6(zRMO$xg?}}5d
zJk8c!@5oZbihIQoxE|aPeZdG|+O~Igk_wHZvD8~VNLE%>j?T`6b#*DEQ#igIA6tX@
zB3X2)Mvq&pRyR!cu-hzjDkLR^iUJR7F<(y>QE{-)sn_sQ<+jvpYdF>EVvhhQ#UIQ!
z_)Hcl!&(g`ZLxo2mCTfdxq9gTT{dV?MV##i&6RIoS|WY@`n5`f&08d7WEN&-FCg4T
zjU9|QbhkTQW;5)pRgY%4u9_(;8Tjaenv@Qvz{>tUDiafv*-{J1OZ3-W&90X)%!j2}
zKYtQ}^9n(sh=PKG>(*1vySEAl&!$?qkIv6U3=AkwPEU_7_SNExoh_alCPYM_13+MO
zY>Yf2amc{!?g|5(F1<4`vdY=XsSTv&f(YOVtex*pm(T6nG}Y87dE`KtnO{l7(OcIu
z0VGemMwQBa&v%o_aGeXd0J4W$HY06t!+c*oZ##40KjmvBK!G=^QB`kt!C_%xL1UR}
zHoa9lpy~wA5_qJwwYBYo0|zk3D6MdFa&t4QtFbjSG_Ef8iR`#Yp2Nao5D;)qD`b1h
zDk%80w||6b+Zsuila=+tq5J%Xhllw1_!!J2Y8smM?QLq!7B@tncCWKDI~cFJI(DFu
zmpQ#Vy0jFLnMt;7$F(<80sHXqVB0{S<U3^G11Kd#M6fWQq5vP@Yu#LG)UNyOMZ3}a
ztkP06U0vOR;^MAkcFXTkQSS^*^IErn1}sjQW^)HPPfu^}@p?ZNz^lkuSmgLTe=Me_
zq=W*)sQCG_j+q&Xy}iBD*)|4ncAI+5X9~{FoQxVjeZemD@$-Mog5iE+DR31iC@3&6
z@GoCxPMWN2Z1{x{vH#Q?at$OU)%YD+F}GjucKyy`;U`?tWN=^KCvX!1^k`6;FZ#H=
zJ(_J~VFAE2KQI|Vb2|V>UcGt+vwZs)J!(>8WNO-ZvM~^uni@7ZD0zJs8Hs_7i@OSp
zA-GpFwdO?Cstj)K?pr3%vzK+2OFe(b^Ql7P<Kro~xc<;Il>YePtyw0fr9}$-{P*<q
zh?$w0FJDNd4lJF)5cLNp9=wIip!$NJpFbx*f4~yK>+ku)&5cV6zvt_Pg$0!w6C4{`
z+X~mq1Mp|}Uu0!p;WBBWfT0?24BXAt_4P9==?KGx5q^b@MWgqX7a9{4y1hrM-N-2D
z=wN*hfs)T_SI287Yi*^Ufq}%tczQKp@S%wdKfTsP>k%POE{5EO;{U#D@jNK-={1kQ
z{_Wpk!WO>C0{;#pXYY`D^Y23<iB>4T|2wQN%nIw%e?RU4c80uCRMhhuIt+M0MWlsG
I1;2X#FHQjpS^xk5

literal 0
HcmV?d00001

diff --git a/_sources/api.rst b/_sources/api.rst
index 309a47690..27375ba93 100644
--- a/_sources/api.rst
+++ b/_sources/api.rst
@@ -57,6 +57,12 @@ Developer API
    sleap.nn.data.resizing
    sleap.nn.data.training
    sleap.nn.data.utils
+   sleap.nn.config.data
+   sleap.nn.config.model
+   sleap.nn.config.optimization
+   sleap.nn.config.outputs
+   sleap.nn.config.training_job
+   sleap.nn.config.utils
    sleap.nn.architectures.common
    sleap.nn.architectures.encoder_decoder
    sleap.nn.architectures.hourglass
@@ -66,15 +72,8 @@ Developer API
    sleap.nn.architectures.resnet
    sleap.nn.architectures.unet
    sleap.nn.architectures.upsampling
-   sleap.nn.config.data
-   sleap.nn.config.model
-   sleap.nn.config.optimization
-   sleap.nn.config.outputs
-   sleap.nn.config.training_job
-   sleap.nn.config.utils
    sleap.nn.tracker.components
    sleap.nn.tracker.kalman
-   sleap.io.asyncvideo
    sleap.io.convert
    sleap.io.dataset
    sleap.io.legacy
diff --git a/_sources/guides/bonsai.md b/_sources/guides/bonsai.md
new file mode 100644
index 000000000..d262873b6
--- /dev/null
+++ b/_sources/guides/bonsai.md
@@ -0,0 +1,75 @@
+(bonsai)=
+
+# Using Bonsai with SLEAP
+
+Bonsai is a visual language for reactive programming and currently supports SLEAP models.
+
+:::{note}
+Currently Bonsai supports only single instance, top-down and top-down-id SLEAP models.
+:::
+
+### Exporting a SLEAP trained model
+
+Before we can import a trained model into Bonsai, we need to use the {code}`sleap-export` command to convert the model to a format supported by Bonsai. For example, to export a top-down-id model, the command is as follows:
+
+```bash
+sleap-export -m centroid/model/folder/path -m top_down_id/model/folder/path -e exported/model/path
+```
+
+Please refer to the {ref}`sleap-export` docs for more details on using the command.
+
+This will generate the necessary `.pb` file and other information files required by Bonsai. In this example, these files were saved to the specified `exported/model/path` folder.
+
+The `exported/model/path` folder will have a structure like the following:
+
+```plaintext
+exported/model/path
+├── centroid_config.json
+├── confmap_config.json
+├── frozen_graph.pb
+└── info.json
+```
+
+### Installing Bonsai and necessary packages
+
+1. Install Bonsai. See the [Bonsai installation instructions](https://bonsai-rx.org/docs/articles/installation.html). 
+
+2. Download and add the necessary packages for Bonsai to run with SLEAP. See the official [Bonsai SLEAP documentation](https://github.com/bonsai-rx/sleap?tab=readme-ov-file#bonsai---sleap) for more information.
+
+### Using Bonsai SLEAP modules
+
+Once you have Bonsai installed with the required packages, you should be able to open the Bonsai application. The workflow must have a source module `FileCapture` which can be found in the toolbox search in the workflow editor. Provide the path to the video that was used to train the SLEAP model in the `FileName` field of the module.
+
+![Bonsai FileCapture module](../_static/bonsai-filecapture.jpg)
+
+#### Top-down model
+The top-down model requires both the `PredictCentroids` and the `PredictPoses` modules.
+
+The `PredictCentroids` module will predict the centroids of detections. There are two fields inside the `PredictCentroids` module: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the centroid model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictCentroids module](../_static/bonsai-predictcentroids.jpg)
+
+The `PredictPoses` module will predict the instances of detections. Similar to the `PredictCentroid` module, there are two fields inside the `PredictPoses` module: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the centered instance model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictPoses module](../_static/bonsai-predictposes.jpg)
+
+#### Top-Down-ID model
+The `PredictPoseIdentities` module will predict the instances with identities. This module has two fields: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the top-down-id model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictPoseIdentities module](../_static/bonsai-predictposeidentities.jpg)
+
+#### Single instance model
+The `PredictSinglePose` module will predict the poses for single instance models. This module also has two fields: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the single instance model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+### Connecting the modules
+Right-click on the `FileCapture` module and select **Create Connection**. Now click on the required SLEAP module to complete the connection. 
+
+![Bonsai module connection ](../_static/bonsai-connection.jpg)
+
+Once it is done, the workflow in Bonsai will look something like the following:
+
+![Bonsai.SLEAP workflow](../_static/bonsai-workflow.jpg)
+
+Now you can click the green start button to run the workflow and you can add more modules to analyze and visualize the results in Bonsai.
+
+For more documentation on various modules and workflows, please refer to the [official Bonsai docs](https://bonsai-rx.org/docs/articles/editor.html).
diff --git a/_sources/guides/cli.md b/_sources/guides/cli.md
index 35ea52171..134461c60 100644
--- a/_sources/guides/cli.md
+++ b/_sources/guides/cli.md
@@ -36,8 +36,8 @@ optional arguments:
 
 ```none
 usage: sleap-train [-h] [--video-paths VIDEO_PATHS] [--val_labels VAL_LABELS]
-                   [--test_labels TEST_LABELS] [--tensorboard] [--save_viz]
-                   [--zmq] [--run_name RUN_NAME] [--prefix PREFIX]
+                   [--test_labels TEST_LABELS] [--tensorboard] [--save_viz] 
+                   [--keep_viz] [--zmq] [--run_name RUN_NAME] [--prefix PREFIX]
                    [--suffix SUFFIX]
                    training_job_path [labels_path]
 
@@ -68,6 +68,8 @@ optional arguments:
   --save_viz            Enable saving of prediction visualizations to the run
                         folder if not already specified in the training job
                         config.
+  --keep_viz            Keep prediction visualization images in the run
+                        folder after training if --save_viz is enabled.
   --zmq                 Enable ZMQ logging (for GUI) if not already specified
                         in the training job config.
   --run_name RUN_NAME   Run name to use when saving file, overrides other run
@@ -99,9 +101,9 @@ optional arguments:
   -e [EXPORT_PATH], --export_path [EXPORT_PATH]
                         Path to output directory where the frozen model will be exported to.
                         Defaults to a folder named 'exported_model'.
-  -u, --unrag UNRAG
-                        Convert ragged tensors into regular tensors with NaN padding.
-                        Defaults to True.
+  -r, --ragged RAGGED
+                        Keep tensors ragged if present. If ommited, convert
+                        ragged tensors into regular tensors with NaN padding.
   -n, --max_instances MAX_INSTANCES
                         Limit maximum number of instances in multi-instance models.
                         Not available for ID models. Defaults to None.
@@ -136,7 +138,10 @@ usage: sleap-track [-h] [-m MODELS] [--frames FRAMES] [--only-labeled-frames] [-
                    [data_path]
 
 positional arguments:
-  data_path             Path to data to predict on. This can be a labels (.slp) file or any supported video format.
+  data_path             Path to data to predict on. This can be one of the following: A .slp file containing labeled data; A folder containing multiple
+                        video files in supported formats; An individual video file in a supported format; A CSV file with a column of video file paths. 
+                        If more than one column is provided in the CSV file, the first will be used for the input data paths and the next column will be
+                        used as the output paths; A text file with a path to a video file on each line
 
 optional arguments:
   -h, --help            show this help message and exit
@@ -151,7 +156,7 @@ optional arguments:
                         Only run inference on unlabeled suggested frames when running on labels dataset. This is useful for generating predictions for
                         initialization during labeling.
   -o OUTPUT, --output OUTPUT
-                        The output filename to use for the predicted data. If not provided, defaults to '[data_path].predictions.slp'.
+                        The output filename or directory path to use for the predicted data. If not provided, defaults to '[data_path].predictions.slp'.
   --no-empty-frames     Clear any empty frames that did not have any detected instances before saving to output.
   --verbosity {none,rich,json}
                         Verbosity of inference progress reporting. 'none' does not output anything during inference, 'rich' displays an updating
@@ -202,7 +207,7 @@ optional arguments:
   --tracking.clean_iou_threshold TRACKING.CLEAN_IOU_THRESHOLD
                         IOU to use when culling instances *after* tracking. (default: 0)
   --tracking.similarity TRACKING.SIMILARITY
-                        Options: instance, centroid, iou (default: instance)
+                        Options: instance, normalized_instance, object_keypoint, centroid, iou (default: instance)
   --tracking.match TRACKING.MATCH
                         Options: hungarian, greedy (default: greedy)
   --tracking.robust TRACKING.ROBUST
@@ -322,7 +327,8 @@ optional arguments:
                               analysis file for the latter video is given a default name.
   --format FORMAT       Output format. Default ('slp') is SLEAP dataset;
                         'analysis' results in analysis.h5 file; 'analysis.nix' results
-                        in an analysis nix file; 'h5' or 'json' results in SLEAP dataset
+                        in an analysis nix file; 'analysis.csv' results
+                        in an analysis csv file; 'h5' or 'json' results in SLEAP dataset
                         with specified file format.
   --video VIDEO         Path to video (if needed for conversion).
 ```
@@ -389,6 +395,9 @@ optional arguments:
   --distinctly_color DISTINCTLY_COLOR
                         Specify how to color instances. Options include: "instances",
                         "edges", and "nodes" (default: "instances")
+  --background BACKGROUND
+                        Specify the type of background to be used to save the videos.
+                        Options: original, black, white and grey. (default: "original")
 ```
 
 ## Debugging
diff --git a/_sources/guides/gui.md b/_sources/guides/gui.md
index 88cf3f656..813ed68fa 100644
--- a/_sources/guides/gui.md
+++ b/_sources/guides/gui.md
@@ -60,7 +60,7 @@ Note that many of the menu command have keyboard shortcuts which can be configur
 
 "**Edge Style**" controls whether edges are drawn as thin lines or as wedges which indicate the {ref}`orientation` of the instance (as well as the direction of the part affinity field which would be used to predict the connection between nodes when using a "bottom-up" approach).
 
-"**Trail Length**" allows you to show a trail of where each instance was located in prior frames (the length of the trail is the number of prior frames). This can be useful when proofreading predictions since it can help you detect swaps in the identities of animals across frames.
+"**Trail Length**" allows you to show a trail of where each instance was located in prior frames (the length of the trail is the number of prior frames). This can be useful when proofreading predictions since it can help you detect swaps in the identities of animals across frames. By default, you can only select trail lengths of up to 250 frames. You can use a custom trail length by modifying the default length in the `preferences.yaml` file. However, using trail lengths longer than about 500 frames can result in significant lag.
 
 "**Fit Instances to View**" allows you to toggle whether the view is auto-zoomed to the instances in each frame. This can be useful when proofreading predictions.
 
diff --git a/_sources/guides/index.md b/_sources/guides/index.md
index 7eb55b2b2..6d773d9de 100644
--- a/_sources/guides/index.md
+++ b/_sources/guides/index.md
@@ -30,6 +30,10 @@
 
 {ref}`remote-inference` when you trained models and you want to run inference on a different machine using a **command-line interface**.
 
+## SLEAP with Bonsai
+
+{ref}`bonsai` when you want to analyze the trained SLEAP model to visualize the poses, centroids and identities for further visual analysis.
+
 ```{toctree}
 :hidden: true
 :maxdepth: 2
@@ -44,4 +48,5 @@ proofreading
 colab
 custom-training
 remote
+bonsai
 ```
diff --git a/_sources/guides/proofreading.md b/_sources/guides/proofreading.md
index fea1c5ebc..941b85154 100644
--- a/_sources/guides/proofreading.md
+++ b/_sources/guides/proofreading.md
@@ -50,6 +50,8 @@ There are currently three methods for matching instances in frame N against thes
 - “**centroid**” measures similarity by the distance between the instance centroids
 - “**iou**” measures similarity by the intersection/overlap of the instance bounding boxes
 - “**instance**” measures similarity by looking at the distances between corresponding nodes in the instances, normalized by the number of valid nodes in the candidate instance.
+- “**normalized_instance**” measures similarity by looking at the distances between corresponding nodes in the instances, normalized by the number of valid nodes in the candidate instance and the keypoints normalized by the image size.
+- “**object_keypoint**” measures similarity by measuring the distance between each keypoints from a reference instance and a query instance, takes the exp(-d**2), sum for all the keypoints and divide by the number of visible keypoints in the reference instance.
 
 Once SLEAP has measured the similarity between all the candidates and the instances in frame N, you need to choose a way to pair them up. You can do this either by picking the best match, and the picking the best remaining match for each remaining instance in turn—this is “**greedy**” matching—or you can find the way of matching identities which minimizes the total cost (or: maximizes the total similarity)—this is “**Hungarian**” matching.
 
diff --git a/_sources/installation.md b/_sources/installation.md
index eea65cc31..2c1ef41be 100644
--- a/_sources/installation.md
+++ b/_sources/installation.md
@@ -1,10 +1,21 @@
 # Installation
 
-SLEAP can be installed as a Python package on Windows, Linux, Mac OS X, and Mac OS Apple Silicon.
+SLEAP can be installed as a Python package on Windows, Linux, and Mac OS. For quick install using conda, see below:
 
-SLEAP requires many complex dependencies, so we **strongly** recommend using [Mambaforge](https://mamba.readthedocs.io/en/latest/installation.html) to install it in its own isolated environment. See {ref}`Installing Mambaforge<mambaforge>` below for more instructions.
+````{tabs}
+   ```{group-tab} Windows and Linux
+      ```bash
+      conda create -y -n sleap -c conda-forge -c nvidia -c sleap/label/dev -c sleap -c anaconda sleap=1.4.1
+      ```
+   ```
+   ```{group-tab} Mac OS
+      ```bash
+      conda create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.4.1
+      ```
+   ```
+````
 
-The newest version of SLEAP can always be found in the [Releases page](https://github.com/talmolab/sleap/releases).
+. For more in-depth installation instructions, see the [installation methods](installation-methods). The newest version of SLEAP can always be found in the [Releases page](https://github.com/talmolab/sleap/releases).
 
 ```{contents} Contents
 ---
@@ -12,66 +23,30 @@ local:
 ---
 ```
 
-````{hint}
-Installation requires entering commands in a terminal. To open one:
-
-**Windows:** Open the *Start menu* and search for the *Miniforge Prompt* (if using Mambaforge) or the *Command Prompt* if not.
-```{note}
-On Windows, our personal preference is to use alternative terminal apps like [Cmder](https://cmder.net) or [Windows Terminal](https://aka.ms/terminal).
-```
-
-**Linux:** Launch a new terminal by pressing <kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>.
-
-**Mac:** Launch a new terminal by pressing <kbd>Cmd</kbd> + <kbd>Space</kbd> and searching for _Terminal_.
-
-````
-
-(apple-silicon)=
-
-### Macs Pre-M1 (Pre-Installation)
-
-SLEAP can be installed on Macs by following these instructions:
-
-1. Make sure you're on **macOS Monterey** or later, i.e., version 12+.
-
-2. If you don't have it yet, [install **homebrew**](https://brew.sh/), a convenient package manager for Macs (skip this if you can run `brew` from the terminal):
-
-   ```bash
-   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
-   ```
-
-   This might take a little while since it'll also install Xcode (which we'll need later). Once it's finished, your terminal should give you two extra commands to run listed under **Next Steps**.
-
-   ````{note}
-   We recommend running the commands given in your terminal which will be similar to (but may differ slightly) from the commands below:
-   ```bash
-   echo 'eval "$(/opt/homebrew/bin/brew shellenv)"' >> ~/.zprofile
-   ```
-
-   ```bash
-   eval "$(/opt/homebrew/bin/brew shellenv)"
-   ```
-
+`````{hint}
+   Installation requires entering commands in a terminal. To open one:
+   ````{tabs}
+      ```{tab} Windows
+         Open the *Start menu* and search for the *Anaconda Prompt* (if using Miniconda) or the *Command Prompt* if not.
+         ```{note}
+         On Windows, our personal preference is to use alternative terminal apps like [Cmder](https://cmder.net) or [Windows Terminal](https://aka.ms/terminal).
+         ```
+      ```
+      ```{tab} Linux
+         Launch a new terminal by pressing <kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>.
+      ```
+      ```{group-tab} Mac OS
+         Launch a new terminal by pressing <kbd>Cmd</kbd> + <kbd>Space</kbd> and searching for _Terminal_.
+      ```
    ````
+`````
 
-   Then, close and re-open the terminal for it to take effect.
+## Package Manager
 
-3. Install wget, a CLI downloading utility (also makes sure your homebrew setup worked):
-
-   ```bash
-   brew install wget
-   ```
-
-(mambaforge)=
-
-## Installing Mambaforge
-
-**Anaconda** is a Python environment manager that makes it easy to install SLEAP and its necessary dependencies without affecting other Python software on your computer.
-
-[**Mambaforge**](https://mamba.readthedocs.io/en/latest/installation.html) is a lightweight installer of Anaconda with speedy package resolution that we recommend.
+SLEAP requires many complex dependencies, so we **strongly** recommend using a package manager such as [Miniforge](https://github.com/conda-forge/miniforge) or [Miniconda](https://docs.anaconda.com/free/miniconda/) to install SLEAP in its own isolated environment.
 
 ````{note}
-If you already have Anaconda on your computer, then you can [set the solver to `libmamba`](https://www.anaconda.com/blog/a-faster-conda-for-a-growing-community) in the `base` environment (and skip the Mambaforge installation):
+If you already have Anaconda on your computer (and it is an [older installation](https://conda.org/blog/2023-11-06-conda-23-10-0-release/)), then make sure to [set the solver to `libmamba`](https://www.anaconda.com/blog/a-faster-conda-for-a-growing-community) in the `base` environment.
 
 ```bash
 conda update -n base conda
@@ -80,195 +55,220 @@ conda config --set solver libmamba
 ```
 
 ```{warning}
-Any subsequent `mamba` commands in the docs will need to be replaced with `conda` if you choose to use your existing Anaconda installation.
+Any subsequent `conda` commands in the docs will need to be replaced with `mamba` if you have [Mamba](https://mamba.readthedocs.io/en/latest/) installed instead of Anaconda or Miniconda.
 ```
 
 ````
 
-Otherwise, to install Mamba:
-
-**On Windows**, just click through the installation steps.
-
-1.  Go to: https://github.com/conda-forge/miniforge#mambaforge
-2.  Download the latest version for your OS.
-3.  Follow the installer instructions.
-
-We recommend using the following settings:
-
-- Install for: All Users (requires admin privileges)
-- Destination folder: `C:\mambaforge`
-- Advanced Options: Add MambaForge to the system PATH environment variable
-- Advanced Options: Register MambaForge as the system Python 3.X
-  These will make sure that MambaForge is easily accessible from most places on your computer.
-
-**On Linux**, it might be easier to do this straight from the terminal (<kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>) with this one-liner:
-
-```bash
-wget -nc https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-Linux-x86_64.sh && bash Mambaforge-Linux-x86_64.sh -b && ~/mambaforge/bin/conda init bash
-```
-
-Restart the terminal after running this command.
-
-```{note}
-For other Linux architectures (arm64 and POWER8/9), replace the `.sh` filenames above with the correct installer name for your architecture. See the Download column in [this table](https://github.com/conda-forge/miniforge#mambaforge) for the correct filename.
-
-```
-
-**On Macs (pre-M1)**, you can run the installer using this terminal command:
+If you don't have a `conda` package manager installation, here are some quick install options:
 
-```bash
-wget -nc https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-MacOSX-x86_64.sh && bash Mambaforge-MacOSX-x86_64.sh -b && ~/mambaforge/bin/conda init zsh
-```
+### Miniforge (recommended)
 
-**On Macs (Apple Silicon)**, use this terminal command:
+Miniforge is a minimal installer for conda that includes the `conda` package manager and is maintained by the [conda-forge](https://conda-forge.org) community. The only difference between Miniforge and Miniconda is that Miniforge uses the `conda-forge` channel by default, which provides a much wider selection of community-maintained packages.
 
-```bash
-curl -fsSL --compressed https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-MacOSX-arm64.sh -o Mambaforge3-MacOSX-arm64.sh && chmod +x Mambaforge3-MacOSX-arm64.sh && ./Mambaforge3-MacOSX-arm64.sh -b -p ~/mambaforge3 && rm Mambaforge3-MacOSX-arm64.sh && ~/mambaforge3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
-```
+````{tabs}
+   ```{group-tab} Windows
+      Open a new PowerShell terminal (does not need to be admin) and enter:
 
-## Installation methods
-
-SLEAP can be installed three different ways: via {ref}`conda package<condapackage>`, {ref}`conda from source<condasource>`, or {ref}`pip package<pippackage>`. Select one of the methods below to install SLEAP. We recommend {ref}`conda package<condapackage>`.
-
-(condapackage)=
-
-### `conda` package
-
-**Windows** and **Linux**
-
-```bash
-mamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.3.3
-```
-
-**Mac OS X** and **Apple Silicon**
-
-```bash
-mamba create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.3.3
-```
-
-**This is the recommended installation method**.
-
-```{note}
-- This comes with CUDA to enable GPU support. All you need is to have an NVIDIA GPU and [updated drivers](https://nvidia.com/drivers).
-- If you already have CUDA installed on your system, this will not conflict with it.
-- This will also work in CPU mode if you don't have a GPU on your machine.
-```
-
-(condasource)=
-
-### `conda` from source
-
-1. First, ensure git is installed:
-
-   ```bash
-   git --version
+      ```bash
+      Invoke-WebRequest -Uri "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Windows-x86_64.exe" -OutFile "$env:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe"; Start-Process -FilePath "$env:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe" -ArgumentList "/InstallationType=JustMe /RegisterPython=1 /S" -Wait; Remove-Item -Path "$env:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe"
+      ```
    ```
+   ```{group-tab} Linux
+      Open a new terminal and enter:
 
-   If 'git' is not recognized, then [install git](https://git-scm.com/book/en/v2/Getting-Started-Installing-Git).
-
-2. Then, clone the repository:
-
-   ```bash
-   git clone https://github.com/talmolab/sleap && cd sleap
+      ```bash
+      curl -fsSL --compressed https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh -o "~/Downloads/Miniforge3-Linux-x86_64.sh" && chmod +x "~/Downloads/Miniforge3-Linux-x86_64.sh" && "~/Downloads/Miniforge3-Linux-x86_64.sh" -b -p ~/miniforge3 && rm "~/Downloads/Miniforge3-Linux-x86_64.sh" && ~/miniforge3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+   ```{group-tab} Mac (Apple Silicon)
+      Open a new terminal and enter:
 
-3. Finally, install from the environment file (differs based on OS and GPU):
-
-   **Windows** and **Linux**
-
-   ```bash
-   mamba env create -f environment.yml -n sleap
-   ```
-
-   If you do not have a NVIDIA GPU, then you should use the no CUDA environment file:
-
-   ```bash
-   mamba env create -f environment_no_cuda.yml -n sleap
+      ```bash
+      curl -fsSL --compressed https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh -o "~/Downloads/Miniforge3-MacOSX-arm64.sh" && chmod +x "~/Downloads/Miniforge3-MacOSX-arm64.sh" && "~/Downloads/Miniforge3-MacOSX-arm64.sh" -b -p ~/miniforge3 && rm "~/Downloads/Miniforge3-MacOSX-arm64.sh" && ~/miniforge3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+   ```{group-tab} Mac (Intel)
+      Open a new terminal and enter:
 
-   **Mac OS X** and **Apple Silicon**
-
-   ```bash
-   mamba env create -f environment_mac.yml -n sleap
+      ```bash
+      curl -fsSL --compressed https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-x86_64.sh -o "~/Downloads/Miniforge3-MacOSX-x86_64.sh" && chmod +x "~/Downloads/Miniforge3-MacOSX-x86_64.sh" && "~/Downloads/Miniforge3-MacOSX-x86_64.sh" -b -p ~/miniforge3 && rm "~/Downloads/Miniforge3-MacOSX-x86_64.sh" && ~/miniforge3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+````
 
-   This is the **recommended method for development**.
-
-```{note}
-- This installs SLEAP in development mode, which means that edits to the source code will be applied the next time you run SLEAP.
-- Change the `-n sleap` in the command to create an environment with a different name (e.g., `-n sleap_develop`).
-```
-
-(pippackage)=
-
-### `pip` package
-
-Although you do not need Mambaforge installed to perform a `pip install`, we recommend {ref}`installing Mambaforge<mambaforge>` to create a new environment where we can isolate the `pip install`. Alternatively, you can use a venv if you have an existing python installation. If you are working on **Google Colab**, skip to step 3 to perform the `pip install` without using a conda environment.
+### Miniconda
 
-1. Otherwise, create a new conda environment where we will `pip install sleap`:
+This is a minimal installer for conda that includes the `conda` package manager and is maintained by the [Anaconda](https://www.anaconda.com) company.
 
-   either without GPU support:
+````{tabs}
+   ```{group-tab} Windows
+      Open a new PowerShell terminal (does not need to be admin) and enter:
 
-   ```bash
-   mamba create --name sleap pip python=3.7.12
+      ```bash
+      curl https://repo.anaconda.com/miniconda/Miniconda3-latest-Windows-x86_64.exe -o miniconda.exe; Start-Process -FilePath ".\miniconda.exe" -ArgumentList "/S" -Wait; del miniconda.exe
+      ```
    ```
+   ```{group-tab} Linux
+      Open a new terminal and enter:
 
-   or with GPU support:
-
-   ```bash
-   mamba create --name sleap pip python=3.7.12 cudatoolkit=11.3 cudnn=8.2
+      ```bash
+      mkdir -p ~/miniconda3 && wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh && bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3 && rm ~/miniconda3/miniconda.sh && ~/miniconda3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+   ```{group-tab} Mac (Apple Silicon)
+      Open a new terminal and enter:
 
-2. Then activate the environment to isolate the `pip install` from other environments on your computer:
-
-   ```bash
-   mamba activate sleap
+      ```bash
+      curl -fsSL --compressed https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-arm64.sh -o "~/Downloads/Miniconda3-latest-MacOSX-arm64.sh" && chmod +x "~/Downloads/Miniconda3-latest-MacOSX-arm64.sh" && "~/Downloads/Miniconda3-latest-MacOSX-arm64.sh" -b -u -p ~/miniconda3 && rm "~/Downloads/Miniconda3-latest-MacOSX-arm64.sh" && ~/miniconda3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+   ```{group-tab} Mac (Intel)
+      Open a new terminal and enter:
 
-   ```{warning}
-   Refrain from installing anything into the `base` environment. Always create a new environment to install new packages.
+      ```bash
+      curl -fsSL --compressed https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh -o "~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh" && chmod +x "~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh" && "~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh" -b -u -p ~/miniconda3 && rm "~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh" && ~/miniconda3/bin/conda init "$(basename "${SHELL}")" && source "$HOME/.$(basename "${SHELL}")rc"
+      ```
    ```
+````
 
-3. Finally, we can perform the `pip install`:
-
-   ```bash
-   pip install sleap[pypi]==1.3.3
-   ```
+See the [Miniconda website](https://docs.anaconda.com/free/miniconda/) for up-to-date installation instructions if the above instructions don't work for your system.
 
-   This works on **any OS except Apple silicon** and on **Google Colab**.
+(installation-methods)=
 
-   ```{note}
-   The pypi distributed package of SLEAP ships with the following extras:
-   - **pypi**: For installation without an mamba environment file. All dependencies come from PyPI.
-   - **jupyter**: This installs all *pypi* and jupyter lab dependencies.
-   - **dev**: This installs all *jupyter* dependencies and developement tools for testing and building docs.
-   - **conda_jupyter**: For installation using a mamba environment file included in the source code. Most dependencies are listed as conda packages in the environment file and only a few come from PyPI to allow jupyter lab support.
-   - **conda_dev**: For installation using [a mamba environment](https://github.com/search?q=repo%3Atalmolab%2Fsleap+path%3Aenvironment*.yml&type=code) with a few PyPI dependencies for development tools.
-   ```
+## Installation methods
 
-   ```{note}
-   - Requires Python 3.7
-   - To enable GPU support, make sure that you have **CUDA Toolkit v11.3** and **cuDNN v8.2** installed.
-   ```
+SLEAP can be installed three different ways: via {ref}`conda package<condapackage>`, {ref}`conda from source<condasource>`, or {ref}`pip package<pippackage>`. Select one of the methods below to install SLEAP. We recommend {ref}`conda package<condapackage>`.
 
-   ```{warning}
-   This will uninstall existing libraries and potentially install conflicting ones.
+`````{tabs}
+   ```{tab} conda package
+      **This is the recommended installation method**.
+      ````{tabs}
+         ```{group-tab} Windows and Linux
+            ```bash
+            conda create -y -n sleap -c conda-forge -c nvidia -c sleap/label/dev -c sleap -c anaconda sleap=1.4.1
+            ```
+            ```{note}
+            - This comes with CUDA to enable GPU support. All you need is to have an NVIDIA GPU and [updated drivers](https://nvidia.com/drivers).
+            - If you already have CUDA installed on your system, this will not conflict with it.
+            - This will also work in CPU mode if you don't have a GPU on your machine.
+            ```
+         ```
+         ```{group-tab} Mac OS
+            ```bash
+            conda create -y -n sleap -c conda-forge -c anaconda -c sleap sleap=1.4.1
+            ```
+            ```{note}
+            This will also work in CPU mode if you don't have a GPU on your machine.
+            ```
+         ```
+      ````
 
-   We strongly recommend that you **only use this method if you know what you're doing**!
    ```
+   ```{tab} conda from source
+      This is the **recommended method for development**.
+      1. First, ensure git is installed:
+         ```bash
+         git --version
+         ```
+         If `git` is not recognized, then [install git](https://git-scm.com/book/en/v2/Getting-Started-Installing-Git).
+      2. Then, clone the repository:
+         ```bash
+         git clone https://github.com/talmolab/sleap && cd sleap
+         ```
+      3. Finally, install SLEAP from the environment file:
+         ````{tabs}
+            ```{group-tab} Windows and Linux
+               ````{tabs}
+                  ```{group-tab} NVIDIA GPU
+                     ```bash
+                     conda env create -f environment.yml -n sleap
+                     ```
+                  ```
+                  ```{group-tab} CPU or other GPU
+                     ```bash
+                     conda env create -f environment_no_cuda.yml -n sleap
+                     ```
+                  ```
+               ````
+            ```
+            ```{group-tab} Mac OS
+               ```bash
+               conda env create -f environment_mac.yml -n sleap
+               ```
+            ```
+         ````
+      ```{note}
+      - This installs SLEAP in development mode, which means that edits to the source code will be applied the next time you run SLEAP.
+      - Change the `-n sleap` in the command to create an environment with a different name (e.g., `-n sleap_develop`).
+      ```
+   ```
+   ```{tab} pip package
+      This is the **recommended method for Google Colab only**.
+      ```{warning}
+      This will uninstall existing libraries and potentially install conflicting ones.
+
+      We strongly recommend that you **only use this method if you know what you're doing**!
+      ```
+      ````{tabs}
+         ```{group-tab} Windows and Linux
+            ```{note}
+            - Requires Python 3.7
+            - To enable GPU support, make sure that you have **CUDA Toolkit v11.3** and **cuDNN v8.2** installed.
+            ```
+            Although you do not need Miniconda installed to perform a `pip install`, we recommend [installing Miniconda](https://docs.anaconda.com/free/miniconda/) to create a new environment where we can isolate the `pip install`. Alternatively, you can use a venv if you have an existing Python 3.7 installation. If you are working on **Google Colab**, skip to step 3 to perform the `pip install` without using a conda environment.
+            1. Otherwise, create a new conda environment where we will `pip install sleap`:
+               ````{tabs}
+                  ```{group-tab} NVIDIA GPU
+                     ```bash
+                     conda create --name sleap pip python=3.7.12 cudatoolkit=11.3 cudnn=8.2 -c conda-forge -c nvidia
+                     ```
+                  ```
+                  ```{group-tab} CPU or other GPU
+                     ```bash
+                     conda create --name sleap pip python=3.7.12
+                     ```
+                  ```
+               ````
+            2. Then activate the environment to isolate the `pip install` from other environments on your computer:
+               ```bash
+               conda activate sleap
+               ```
+               ```{warning}
+               Refrain from installing anything into the `base` environment. Always create a new environment to install new packages.
+               ```
+            3. Finally, we can perform the `pip install`:
+               ```bash
+               pip install sleap[pypi]==1.4.1
+               ```
+               ```{note}
+               The pypi distributed package of SLEAP ships with the following extras:
+               - **pypi**: For installation without an conda environment file. All dependencies come from PyPI.
+               - **jupyter**: This installs all *pypi* and jupyter lab dependencies.
+               - **dev**: This installs all *jupyter* dependencies and developement tools for testing and building docs.
+               - **conda_jupyter**: For installation using a conda environment file included in the source code. Most dependencies are listed as conda packages in the environment file and only a few come from PyPI to allow jupyter lab support.
+               - **conda_dev**: For installation using [a conda environment](https://github.com/search?q=repo%3Atalmolab%2Fsleap+path%3Aenvironment*.yml&type=code) with a few PyPI dependencies for development tools.
+               ```
+         ```
+         ```{group-tab} Mac OS
+            Not supported.
+         ```
+         ````
+   ```
+`````
 
 ## Testing that things are working
 
-If you installed using `mamba`, first activate the `sleap` environment by opening a terminal and typing:
+If you installed using `conda`, first activate the `sleap` environment by opening a terminal and typing:
 
 ```bash
-mamba activate sleap
+conda activate sleap
 ```
 
 ````{hint}
-Not sure what `mamba` environments you already installed? You can get a list of the environments on your system with:
+Not sure what `conda` environments you already installed? You can get a list of the environments on your system with:
 ```
-mamba env list
+conda env list
 ```
 ````
 
@@ -301,7 +301,7 @@ python -c "import sleap; sleap.versions()"
 
 ### GPU support
 
-Assuming you installed using either of the `mamba`-based methods on Windows or Linux, SLEAP should automatically have GPU support enabled.
+Assuming you installed using either of the `conda`-based methods on Windows or Linux, SLEAP should automatically have GPU support enabled.
 
 To check, verify that SLEAP can detect the GPUs on your system:
 
@@ -362,7 +362,7 @@ file: No such file or directory
 then activate the environment:
 
 ```bash
-mamba activate sleap
+conda activate sleap
 ```
 
 and run the commands:
@@ -391,13 +391,13 @@ We **strongly recommend** installing SLEAP in a fresh environment when updating.
 To uninstall an existing environment named `sleap`:
 
 ```bash
-mamba env remove -n sleap
+conda env remove -n sleap
 ```
 
 ````{hint}
-Not sure what `mamba` environments you already installed? You can get a list of the environments on your system with:
+Not sure what `conda` environments you already installed? You can get a list of the environments on your system with:
 ```bash
-mamba env list
+conda env list
 ```
 ````
 
@@ -413,10 +413,10 @@ If you get any errors or the GUI fails to launch, try running the diagnostics to
 sleap-diagnostic
 ```
 
-If you were not able to get SLEAP installed, activate the mamba environment it is in and generate a list of the package versions installed:
+If you were not able to get SLEAP installed, activate the conda environment it is in and generate a list of the package versions installed:
 
 ```bash
-mamba list
+conda list
 ```
 
 Then, [open a new Issue](https://github.com/talmolab/sleap/issues) providing the versions from either command above, as well as any errors you saw in the console during the installation. Or [start a discussion](https://github.com/talmolab/sleap/discussions) to get help from the community.
diff --git a/_sources/notebooks/Training_and_inference_on_an_example_dataset.ipynb b/_sources/notebooks/Training_and_inference_on_an_example_dataset.ipynb
index b0211bbca..4e26cb286 100644
--- a/_sources/notebooks/Training_and_inference_on_an_example_dataset.ipynb
+++ b/_sources/notebooks/Training_and_inference_on_an_example_dataset.ipynb
@@ -335,7 +335,7 @@
       "        \"runs_folder\": \"models\",\n",
       "        \"tags\": [],\n",
       "        \"save_visualizations\": true,\n",
-      "        \"delete_viz_images\": true,\n",
+      "        \"keep_viz_images\": true,\n",
       "        \"zip_outputs\": false,\n",
       "        \"log_to_csv\": true,\n",
       "        \"checkpointing\": {\n",
@@ -727,7 +727,7 @@
       "        \"runs_folder\": \"models\",\n",
       "        \"tags\": [],\n",
       "        \"save_visualizations\": true,\n",
-      "        \"delete_viz_images\": true,\n",
+      "        \"keep_viz_images\": true,\n",
       "        \"zip_outputs\": false,\n",
       "        \"log_to_csv\": true,\n",
       "        \"checkpointing\": {\n",
diff --git a/_static/_sphinx_javascript_frameworks_compat.js b/_static/_sphinx_javascript_frameworks_compat.js
new file mode 100644
index 000000000..8549469dc
--- /dev/null
+++ b/_static/_sphinx_javascript_frameworks_compat.js
@@ -0,0 +1,134 @@
+/*
+ * _sphinx_javascript_frameworks_compat.js
+ * ~~~~~~~~~~
+ *
+ * Compatability shim for jQuery and underscores.js.
+ *
+ * WILL BE REMOVED IN Sphinx 6.0
+ * xref RemovedInSphinx60Warning
+ *
+ */
+
+/**
+ * select a different prefix for underscore
+ */
+$u = _.noConflict();
+
+
+/**
+ * small helper function to urldecode strings
+ *
+ * See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/decodeURIComponent#Decoding_query_parameters_from_a_URL
+ */
+jQuery.urldecode = function(x) {
+    if (!x) {
+        return x
+    }
+    return decodeURIComponent(x.replace(/\+/g, ' '));
+};
+
+/**
+ * small helper function to urlencode strings
+ */
+jQuery.urlencode = encodeURIComponent;
+
+/**
+ * This function returns the parsed url parameters of the
+ * current request. Multiple values per key are supported,
+ * it will always return arrays of strings for the value parts.
+ */
+jQuery.getQueryParameters = function(s) {
+    if (typeof s === 'undefined')
+        s = document.location.search;
+    var parts = s.substr(s.indexOf('?') + 1).split('&');
+    var result = {};
+    for (var i = 0; i < parts.length; i++) {
+        var tmp = parts[i].split('=', 2);
+        var key = jQuery.urldecode(tmp[0]);
+        var value = jQuery.urldecode(tmp[1]);
+        if (key in result)
+            result[key].push(value);
+        else
+            result[key] = [value];
+    }
+    return result;
+};
+
+/**
+ * highlight a given string on a jquery object by wrapping it in
+ * span elements with the given class name.
+ */
+jQuery.fn.highlightText = function(text, className) {
+    function highlight(node, addItems) {
+        if (node.nodeType === 3) {
+            var val = node.nodeValue;
+            var pos = val.toLowerCase().indexOf(text);
+            if (pos >= 0 &&
+                !jQuery(node.parentNode).hasClass(className) &&
+                !jQuery(node.parentNode).hasClass("nohighlight")) {
+                var span;
+                var isInSVG = jQuery(node).closest("body, svg, foreignObject").is("svg");
+                if (isInSVG) {
+                    span = document.createElementNS("http://www.w3.org/2000/svg", "tspan");
+                } else {
+                    span = document.createElement("span");
+                    span.className = className;
+                }
+                span.appendChild(document.createTextNode(val.substr(pos, text.length)));
+                node.parentNode.insertBefore(span, node.parentNode.insertBefore(
+                    document.createTextNode(val.substr(pos + text.length)),
+                    node.nextSibling));
+                node.nodeValue = val.substr(0, pos);
+                if (isInSVG) {
+                    var rect = document.createElementNS("http://www.w3.org/2000/svg", "rect");
+                    var bbox = node.parentElement.getBBox();
+                    rect.x.baseVal.value = bbox.x;
+                    rect.y.baseVal.value = bbox.y;
+                    rect.width.baseVal.value = bbox.width;
+                    rect.height.baseVal.value = bbox.height;
+                    rect.setAttribute('class', className);
+                    addItems.push({
+                        "parent": node.parentNode,
+                        "target": rect});
+                }
+            }
+        }
+        else if (!jQuery(node).is("button, select, textarea")) {
+            jQuery.each(node.childNodes, function() {
+                highlight(this, addItems);
+            });
+        }
+    }
+    var addItems = [];
+    var result = this.each(function() {
+        highlight(this, addItems);
+    });
+    for (var i = 0; i < addItems.length; ++i) {
+        jQuery(addItems[i].parent).before(addItems[i].target);
+    }
+    return result;
+};
+
+/*
+ * backward compatibility for jQuery.browser
+ * This will be supported until firefox bug is fixed.
+ */
+if (!jQuery.browser) {
+    jQuery.uaMatch = function(ua) {
+        ua = ua.toLowerCase();
+
+        var match = /(chrome)[ \/]([\w.]+)/.exec(ua) ||
+            /(webkit)[ \/]([\w.]+)/.exec(ua) ||
+            /(opera)(?:.*version|)[ \/]([\w.]+)/.exec(ua) ||
+            /(msie) ([\w.]+)/.exec(ua) ||
+            ua.indexOf("compatible") < 0 && /(mozilla)(?:.*? rv:([\w.]+)|)/.exec(ua) ||
+            [];
+
+        return {
+            browser: match[ 1 ] || "",
+            version: match[ 2 ] || "0"
+        };
+    };
+    jQuery.browser = {};
+    jQuery.browser[jQuery.uaMatch(navigator.userAgent).browser] = true;
+}
diff --git a/_static/basic.css b/_static/basic.css
index d54be8067..18495ea08 100644
--- a/_static/basic.css
+++ b/_static/basic.css
@@ -222,7 +222,7 @@ table.modindextable td {
 /* -- general body styles --------------------------------------------------- */
 
 div.body {
-    min-width: 450px;
+    min-width: 360px;
     max-width: 800px;
 }
 
@@ -237,16 +237,6 @@ a.headerlink {
     visibility: hidden;
 }
 
-a.brackets:before,
-span.brackets > a:before{
-    content: "[";
-}
-
-a.brackets:after,
-span.brackets > a:after {
-    content: "]";
-}
-
 h1:hover > a.headerlink,
 h2:hover > a.headerlink,
 h3:hover > a.headerlink,
@@ -334,13 +324,15 @@ aside.sidebar {
 p.sidebar-title {
     font-weight: bold;
 }
-
+nav.contents,
+aside.topic,
 div.admonition, div.topic, blockquote {
     clear: left;
 }
 
 /* -- topics ---------------------------------------------------------------- */
-
+nav.contents,
+aside.topic,
 div.topic {
     border: 1px solid #ccc;
     padding: 7px;
@@ -379,6 +371,8 @@ div.body p.centered {
 
 div.sidebar > :last-child,
 aside.sidebar > :last-child,
+nav.contents > :last-child,
+aside.topic > :last-child,
 div.topic > :last-child,
 div.admonition > :last-child {
     margin-bottom: 0;
@@ -386,6 +380,8 @@ div.admonition > :last-child {
 
 div.sidebar::after,
 aside.sidebar::after,
+nav.contents::after,
+aside.topic::after,
 div.topic::after,
 div.admonition::after,
 blockquote::after {
@@ -428,10 +424,6 @@ table.docutils td, table.docutils th {
     border-bottom: 1px solid #aaa;
 }
 
-table.footnote td, table.footnote th {
-    border: 0 !important;
-}
-
 th {
     text-align: left;
     padding-right: 5px;
@@ -614,20 +606,26 @@ ol.simple p,
 ul.simple p {
     margin-bottom: 0;
 }
-
-dl.footnote > dt,
-dl.citation > dt {
+aside.footnote > span,
+div.citation > span {
     float: left;
-    margin-right: 0.5em;
 }
-
-dl.footnote > dd,
-dl.citation > dd {
+aside.footnote > span:last-of-type,
+div.citation > span:last-of-type {
+  padding-right: 0.5em;
+}
+aside.footnote > p {
+  margin-left: 2em;
+}
+div.citation > p {
+  margin-left: 4em;
+}
+aside.footnote > p:last-of-type,
+div.citation > p:last-of-type {
     margin-bottom: 0em;
 }
-
-dl.footnote > dd:after,
-dl.citation > dd:after {
+aside.footnote > p:last-of-type:after,
+div.citation > p:last-of-type:after {
     content: "";
     clear: both;
 }
@@ -644,10 +642,6 @@ dl.field-list > dt {
     padding-right: 5px;
 }
 
-dl.field-list > dt:after {
-    content: ":";
-}
-
 dl.field-list > dd {
     padding-left: 0.5em;
     margin-top: 0em;
diff --git a/_static/bonsai-connection.jpg b/_static/bonsai-connection.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..32b725416e2ed32f12977cdf50c73b53ce9b3d9c
GIT binary patch
literal 113805
zcmeFZ2UJwgmp|A9MTwFXl$<0=R#J;dmZ;F=77@v`l5-<S&LAKlpkyR886>r2$r&Uk
z$*G&zK!@$`&QAC`vpchMcK^F)&ade@R8_xM@7`PWs&4pH;eO$k0k<?%)Kvfk1Oxyt
z{2u@}2T%eK5fc9O#eazL-)lFnT_YyGMoL0*{RTNHIXM|A85sp7%`FN_YDzM)TMW0T
zY3b<c>B*@WZ!^%{rlF&!`>PQGBK$qX*KS_Bc9V{RjDqg}_6PR^Kz)NSiugAX0SADP
znt+I!0M`j%1po-H;i>(L;Q!-8K!~UDI?0Wjq-6LDYHtAu35bXYiHZKA8h^Dv{&xT|
z^)(uHQN`=DdX^*{@94zdCw#rZ`J}RqUVjw9CH~4K=q4!x<Lx_4+&sK|_xU9xrKBG|
zl2Ll9tfH!>uJQbZfg#Aq*u?6!^&1;oJ9}3*cMnf5Z=c`~A)#U65s`^WACps3Kc%JT
z<mTlU6c!bKtE#T4t*dYN-q_yJ+11_iv$t<-d}4BH`q#|t%Iezh^^MJ~?H%ON@yY4g
zIqKr_FTMx>ME}Irzi{>kU(|TM2#JY_h)Mq9i-6D*pNOc5ud$0>r%}`+v3y6%A@=?T
z-IIi`m2EdU#q|;NuUtk+8Mq`?xRHOc_P=xXUt=ujzs1?VF!rx}%>u}Y2=JXpL=6A}
zt~6JNW8LB8sMy(Qt(GeZD%+~63$=?^2jVIEJ_KeUv=E#DIs(E0xN~rTRRPtr52g`v
z%1kr;bc{{~@5!q>U&pZeWoWWl3lelcw#?R>YIQ?MRjy^4cC)ZmOuENWX%%jzCd~HL
z`EVinU07xx1q;-!g(Iwqk{$Q=v#ym?xP3naBkIt`IDjG;of3rugy&)zmZE@L0yse8
zC=^Nm`%<LykOCEKg6*$_Cz_^+-7w$hi6u+S_EQvRi+$Db=xC-gIO}6Nu=9}8;j(R9
zP$Vy9ySa_5*G}TC*(WS2=#Aptxwo$C)m9|({AP(DqABQ#6RH{;1iuhs!~v*P|6p+g
zd=onoA<?LdWD+F2>hLi6sU@P__&r`-V1S$3iiSSoson8|9}E310wa>JNQ*louPPJQ
zbQ4!ZoUk)D$M<=faRAk|^Znnsdr=8j-R)PCy)K6s2=*Cd!4GyhhaTPiPuTSd>w=Qn
zxoT2$S2Fb_m<$vpsBO44_{U4Izqn|dTW7I2@L6r_dzu;cce?p-fROst;Y(5+;CBre
zt^DppSG?knI@Z4nx?CX=um}9V6h>T5pez;2G(Zps_yWUfD&qj1dN=?Xq$~S`9tY^P
zg`*<Zw!Tny{H@FApzU)=8_YH6k%<VFqv4W|=zkU=@~JD*_;-fZD|TaAzvb%LP1~_=
z7iyngZiy%B`p{Z5IXIlb9;DrqsvBWs2|Q;xUCt&1{&vQX3mAU9jV4Twn1JWAv5nBZ
zlE)-xCOE*bHx#*q$h{JRZZ^YFhbRy`GNUIwtOHE?Yj?`y>|E;I&KbjOeCEAq57+rT
zn!XGj@~l^_+A8!7FzJl5w3acnJ5pmwzqTT(J>c%xp@>`5uQ9O9j^>S3nZWo>hifSt
zT)Kg`^p4c!bfnG}zz309SZF{<QQeriW2@7Qh(&99)*&VZe!K`qd!qTU{osp35v(kh
ze$xrrq>T^>K?`T!5~iq+g47hjYRT@AXf(HKx={OF7_+OJFuS{`rrjqM4Vt-Lgf_wC
zoKO*?96IfCz&s_Z8DKIquMC~~*zS#;F*Rb^Uh=U)nhOWo+9vWeMj#)FaN2j@va)^u
zW*AoI;PrFN{&$}_Wc>xKg(d$#B4-uM2)7pkwt=Z6O&17;V5`2&nc_MPnocZNx;0hb
zn^Ppn?$;}&JayT)wKDVy{Be+F7fc1UErFsy+a+GB%lAyQJfDx2mDNy?UTQ`yRn7M^
zp!R5o8%SPTes-+}O5LBGhU=hMH{jYi@P@)s93TmrD<AhSy`uLoGzmmzoCTtw*Tae&
zL{=X{s|6*jR>`80gI*QCVtH!!@L9D$=gw$!V>}Tbmred?GbG0=*ra95ZFj{TYGdc*
zD`!W$ujVZM$<&)!-I6x*%K?D!F}b+x4!jKiMqvh^)9YA=L-T%5!&kGkY$+;iMIBam
z<K^Xdm|BXVJsHsE?8{r<BmjEKbpNj<&=BzNGX*TSWEslru*enHopl#+9tM*rliOiP
zt$%K|<KtKV>Bt1|?ORe;^hFO>(#ShiQ>b~j?;k72pVN~ULom`uUTWGPxGU)gL{Ks}
zN7XfVMn}+d+&|NX@FDGBf~lSL{gjLNK=p^^NmvHYP#j=|1KWA!mhE)h1xq_qZyn&}
z<e+_5-A8i-^n`mp?6%!^hL^zy{a1Y;7^tC5DGu;&r=hsN&tjk$6Tdfzn6rk<7cw6Q
zUGMGM!PqOlo4Pd`@>$vq1n9bS5NKLgiGGgWuJ62e%{{4rx#voC_S)Cfy)jJoFlTr(
zrWKqv;<)hnvz44~?c61{ycOYNt`JuIx4G}nhX(H~-K$}ichwjdw49KXZk%FP$+C6Z
zh4arE7}sMFPxShljrf@Z=Xz(wP3nMj5)V)AaU5(umx)?3CY#LA8W&dcdMhp(@-&Y8
zH4fl+8U2|_8icKb|CW{5$3P$hXTZNR`R7!9Bm`Zp4?&Bat)A;8Bzn<o>DbxHPl*25
zyeU}qG`doy^R8_2^P1zf9@FUr?csO#{?jE!|9{xb{*SfpZ`$5JR{uYh!~ba#*%_v=
z&?@qB)o5&hih1vIq^SJ3^X9SokG0~QuLw_cqzo&@s|^RpnTiA)QN`i_N78sjeL51l
z)ho~`C%)p4$=aDE3U5f{sh@&bet-Y!&EmJZigu;94?I;mUuNC71dT4yqe)<1C7mAU
zCWcwl@!UDx166DtB(BjkR8It=@d~Q!8tWyYIt~D}!W$UHp2y>tS7EN3U^28gk|F{T
zno!*Osgdf#Li*6rkT=ET0bAmY&V)~smhBpv^q%vNw<uT`Q4Xs@^3CjdJ39I0VJo~g
zRn0Z&Dvhn3G>W$q#hFSUs;fJ}cveXsJ8Z#F&s#<Dsw`Lb`56$)ANTiS#yKn?2is~M
z_o<l|vDRZjkgDk~EJ9_^1EiR-t!H&?&0-a{wRj`OX2k3haN<{J`A`n#e@|c~F1(es
z{Ij39aB4~8mf`HZ;j`u{(`!fMA-q}%y-hrz0afR>aG=ShIF{}(8oG6p7lVnz0U~q%
zUd*vLJ)Fi;UBI_FK4&<c*RMDPXkI#A1VG-;PS$RQ2=BopW(pDJ$+hKSXZgTwp}=~r
zw|;e^p~|H1$((@F@+5T_hw4A&S@dVcx6c_%gP5y?jdveGHaGyOXE*SKI0=jG!2!DP
z%f;WyFQV7@%7_0><sU?Up%jbU!>`QRs2*%!B6R5&4uJXw{Cg4bk5>L;xlEv_aR5&D
zO^g>D8y(vkG=}|+1DuZI0M;f*cnHh2X+MEgho~6Cw$fE!l^EkZr4-??+r(Lc(^Ym*
z7wF|S4p3DFutGmWm;tF<Sx`wGv=uRuP6`g9_lFN=MnVD}_0`?B`W%TUGd(j!^3)@$
zSL&i%%j`;4wWMD-4AU)&q>t<{oV(P_<~^un#sR!J37hJI=xpu@Y=7f3k-ZFH-JEU!
zdK+J0WdiQWvZ${?iM*ydR+Z1A(6kJ}#agAd6<euwS!~VeGfvHEHi-h^=?|@sfomI&
zN8;bt6)!$27s1|vg2w!qSK}z{(#w(uTMjz}bclY<FI0E$yOJ*6Pd6}dO31G14$$wk
z41IX#QX$fCv<Hp@IN=px;I@F>oj$P*(`}T?91frk9=92c#}3Tm2Q-?u2*ugK=wlXe
zGh25|vP9EiE^XjUIv(8i&X=iDRZN*LBu{^^LF0g`??*mHWH~OzkGj;zD_2p55(YE2
z$8w}ziR-s!O-=a<X`{@FkEd>$+|NDA8RB^9vYQnHexe5&INyb>SzmCqkSEqWVwi>x
z4F;55W{ZOZ`h14I+AO=J8dokc!2Eqzi$1DVEX*!|%k#Nw8<(TWKD~^F5Q(O#w>g6R
zuO9l{n+%{tfnQWkORBlYXl$Gaf*w@I*FAWeGNqUK!*BJ5l}M~~b;9S37UxJkzvBFQ
zwDd)QNE35&<{}P&zyUH(Ps|fQ`Zz$G6ux_aC+sim&kCN*RLW8G;{abt2d|!jk4J%r
zU)65%6=6pi(I+}rTY6{V@XZ|P(vv{0oZNrDKy{V_2gr)R0sKh;iIH*p)X?FlI6!L%
zt8En4hZl3gN2tLMtifCN^0e4XI{zGzi>>VQkuU#f_W##HK(Tf&k)7ZGEMDUs>7ljx
z+a^PC1+x}i+*yUr)f>hpgt0|^p!c-De^WPDJ(4_!>)zP%ravl!hnzMQai-s6)K9-d
zC!>DvV`*vL>swEbZseBr__DZ{c@c78$%B<@E5~XE$qcR>Oq6biN=EG+_tZ%!|Bx+s
z9(!GU7$&hqj(#%86Uh1u2l!xqx>BC0G@P{Gfe|$SBnK0LT(;E@9=q0~#4Q$ZfE&<o
z;1<C*9AKFoj@DvDgrS15flM$g06&U5$F+xW0AN2lKI_}kH2BzNNP!x$8IHEj+nH5O
z&7Va%dbOS%Y2pAiYSUr86N?Z&#q|#tSr6of#RGMIYMDeWTYQDA7OPW(Z4-BGOs<rN
z!`&rBlYrUlQN)%+V!h|Dg-M?1g+IEa#oodJZb3(g3~+!Scc7?r;0hIz1Ot!+qX|-R
zfQWonk3lya;5}CJvSq&Iz+rC=>W>2q;sAzn&$0Ab-)*e(GHk{-8!M648;w>e-khrS
zndK@A?NaQM&IH}_G<nW|gL2~lLWfz4?3D|l{7@`*TCslAd4CPbkek<DM(ts2uperD
zC*i%S!PX7y-r=$AzYAjYPeBRTL3SMilw6$Zm(Uk*@gH(}2#5)GaMTek$&gmK@XIdV
zZp1<Avg|aZD`ntB`B-7{){x<y&j3%t!G?sr(Go8E(ry(1VJATSbhQrpcW!h4ltw+z
zIcH^pvIE{|W?iMa1lYq-L?So<tty%w%ZAsOu9x5dfnqoSQj$}QuMCP(LM!3`k16nu
zVSqJWZ?naZ0$?ZW31AY6xDLkh1pecqe~928qu_u1i=YtqcPmln`U42b)>AjBTde5N
z8eJ05R~(=;dka2QTf(b^0~iBGp=A!b1j=;(t#n=ZI}X4a51p36zA7AnPMATb!I)%w
z3Rf%;KSz~2p}Tax-~dGp_&ofIf8=^5o!;YX^lfNe)c!J65W*B5BWF;!r9V$Nb(Z)&
zac7ybfbW(>q7`d<L9;h}Ifs=@VYJp!?ll5B{1Y)aVrEr7h9@66f2m6{UIR);3t;jT
znEmY1_Wd|lo8*<A$H?`%rFG0Mb+*iu<r{nicn!__PUrszoY9Asd2Mzq-b=`ts4ZcB
zlzJ!BctTBl{P)n|SdzzN^6ft7QP7)p`e8fPTg~FIXYs!rX8i{a<_zf@^f+0qd3le@
zLqCj966f;CsvC#U={<|jiR1(PKahx$<1kNw2O>wXxBzxPOC&H1D~A?GNGOORbwd3(
z{g?HM=T%pzD3wbVTgDzLJB4;1JkIdqVk06-$O>rx=X>uVhJP&W1EaDCP_>x)bsXT0
zm!uo(9<<74;Fz>`<6;igqt!&|vr#adB~nn2Uwy41-AiepbFRj12>K=%@jVnyFC@oY
zj25kgBZoW7>~$C&#*Jp~S#?Y?x&G{BcqXmeZ|NlX_N|eG?cJpF$Yl^cg6yol5gc;q
z-V^IANp;(69E=Pt{Icbvo#5bfZ(&wZn6QZFwgKl8GIdT7;@61|bFCtN6nM*_WNkPE
z%Z;7G;x&&CydZzj9{lxj>s+IJmc{<=`day(^jNb77{%+HD*YZUpFg<jAf#zJBBxs9
zFeU8+4_j*)|E<|`-N7{K4tHm9i0*Bf3jW)#!ogWba|TD4rL{Z^{bkE@jb4%M*}@f1
zs%^Wk$f&cQ={jNCR?7mWF+{8P&ZC{=$hcD@gsAgIA-r5^c~28^Ki=s6{W>9`{(dhY
z@&rbKvRUovfZxeKs9ak|g)CF&%kGHon3(DAAS^f%nZ{%V+>EUPLZz?uFl(|<(ZDOt
zlKc@cBgEhVEZ;7DVW^t2rEpi_7NRyKf+5wNV6ZZB@j>_Fw6}iST(9m0TL0{^EJRZw
z4=PuSPvLja<`W`8TjO?8>5ve|yivhbZ`~3L^*asOgAI53MA>ztzeScbbelC{SqVoy
z&SB?e<{!qKc=3Jc1jiCVzHh-wnLPKc3!v@l%L}}n76eH6`|+-G3Y%Q8S8Vsx?ZVH5
zPj&8kibx3mc*`9_RW6W<H){4TmiJwitC8eba%5f1a(^ZSi#LN&8t!BQik9@FF2P>0
z+w0Va%zkZmbhU3@Zx|+rw#hk5urm5RoGL^@!?=^$8q;@Nz{V)qpOdGB&$CbFHkuRi
ztaj*xz30a7LiRSV_)wdxAiP_m5T)5;GQUWNco87fI9ZMs0)}zbPaGu579>8L-7v1G
zXbMT<py+(dO#fnU)6VJAg6yz5Wg+E5-;Pi|4nSoGZ`_eR_sNP`8J_#LtBFkQG^M)h
z_2hC`ZPd@kZbhEopP!FO>-^3QVNdhVS<(#APNb^LRRWdB&HdixY{1|evu`F*7gO?d
z$mx{`rP<teUCNQVtEW4&oW*0iX_u87vUQ&m0)3{1*1*5?54)I_V$1jC{24MFR#|C{
zy&9(&M+U}>?HPXExQ89dF}ktz@kUG{9N4F%5B()#JjH_)!UxM0cod5rn9ieX;EZNw
zvnbEKQTFP^YmKUB1s6sS3HoWIj&6URdQrYc|GWl*1AM%IpYyIX@fRc}g*M@>YTkP6
z4C+K)1cWJ*Xb~}+IJxb2a}{IS>BI9;GSgReJh9#Q#KAQ&$m7Z==J#vnsfR2efmrb_
zA9x`8U9O)7!Z~g=n{ts1Lu**4*`yg0BHi7=$^V$`j4e9o1}qyZ9MI_l?WBq_Iqan1
zzv_6flYaWa&?lrm3e8Ru{zaYRae|rYP3!0mw#NjlVp0VB?<`Icn{84{lyYP_UiPD3
z6YojQygU}#ds(H<T=Fnzfrjr~l-Hh*NV3#%#T(DgHEaqtGlAPq7~~Ea4QO5p8_vs#
z<bTYU6AAcVmDbTzH~@#mdmP~96%@lzqqhvbpgEzsBEd^{t2QT4;vmViPa&qrXolr!
ziT|r}_Dd&D$I4{Rtj$MfMZncMylQ=bzd7g;4san>yN?4*W|zI$v3M}z=8)@D<ly5J
z(--(mTKZl0GD{e_;+GrnKW>RbmBF`Uzd)B`q1b07GCGkGPClw}&n8RV?{G3_3JL4$
z=T3z15>_?8LyB}PQLa^CXB!(BrH6X$EDL|lgBQ(`nX~)q>Q$55V~xJWe~Dg`fI0QF
zb`AgAW^=IukW0!fRupG8J}AnEfW*}3gfsC?x{;m2tBx&5WSJhZCsR&G8qJf$D$9vI
zB&X69a4B)>Z8>{@Hx2&I;@{{k>$^|hf%k;YVS^h<us1gG(1pG~Wm3=gg&g>QPK|&|
zB;X%S{8P{TS7i}p<PH;EpK>Q-p88>w(%ris<Svvs*4vOJ8*Q~%f5Ljo$a*^cmlrbs
z9k$fp{va5evxaTN0q}SQCii2{{dZUD_F!^gbu-m+Fd15Lnq8XARqlyeO_WfV$^Ho9
zz(}k+EU2%m^#vP-pxhyTOUgl_nCnQLEt1&_<d*o;hd<8Z=4K#)Bx1M=@}X5|1j~bJ
z>RIFsF>Y4&DzWWld*@(BLmafkwtVYLWshOP@S{P=5+3cJM}tH^>P)oOr;5D=FPFk0
zu~+*J&*O(Z-LNoD7u7FYvVe!>akgZE0G8xkEG-I>YqlljDP18saZX*DA07SW^=klK
z>}2sf1iYjAN$<Uk$G_ikJo_CNpicjN$ZvWTN?^yrAIq4<P}f=y5gXU7cBL3B&CH8b
zW={zF7Vof~J;QjGI4!4GX6E+*73O+&NIx?txfoaOc*DI!;>SC-+j{R<))%5l49NlP
z$c`;2rDcXv^Q4F{c@&e@x)ST%uh#vn-%eOFy*xW(oS}f896!a%`Nq`X)7$PJL_dpn
zFhld6>YTs+(D%Vd{#kNgbm{Q;@C{UZPj|XwYRVI}PFD#rC9(a4+bnBxuU}#%zgyU>
z=%)j_scspY#tYVJK0M`XKM{(*$*Mo(FfKs*<zFYjRs&1{90CaiUpsX;BwE_tcUrak
zDk>4*!4+ZGv_8M_KqLNUPs6YVY2=2TMR!;?Bo34wg6WP}DXFWi2&vP8I(N$L$$YRJ
z6`yzSF>^<BzZy_~#G9xkw&>;jHPp-6xU4`q@lg(Y@@OFY$8+La%Kp+wfetX0AeL71
zT|rkNR}j6FwBfqn1(*|1agibSo8|LKm)}53!Tlk9kM*s9Vr?VGZ*+G0!KpQ4yXCWR
z_<_P3*h`cma`CFYe#9>QbRf0y?gtg_kEw~{6>O(`5BtvU#(q*zt2>r#8mg3i99?vD
zP1Ydx?ae)HN`=sza(e?C6|+%!GeeqKsf<Dc@UcbM)+@Ctj>ZRWmhpWqH%<#d=>fwE
zT&bO43LyrULE&rxqg1Q&<R?`D8ar?V(@EpyPjzj6&0^Gok66cK@_h)yh@LY2oByuW
z@^9}W5n#Oxht7l7E$&k`!BEvEW6k$+QRNV#Z5h@ON6Ac6P*q)B%7GFvJ-&z8tcOc4
zC(15uCij&~AblI}>>3WBM|Bm2F6qxjIh5Hy@Os|0NQvZzg*K}+(zj-KV)FAdpY&b+
z;&U^8W&AZ>&Ur{L(V*wkawxH-J9t_ZuXay`_@^%E41y8x(_+;hP-el>W?sQ6uW40O
z4c8P%brjNk-_=2K<IDbyk4kSVt4w=jLef5!aNHhIxRqI3_6<9lu8>+dQ;+6cWNV%(
zmt0#Z7!k6pclafO9XQv#`(c2Cbd<j$UQPEVT?#<dLfRCQFdXgX{}u)7G=W6=DHeGJ
zEp;XgfAjLV@Am6}quNx5doMa{uwJwTQ1Ez--><#Dc^@TxW>bgAObr=hE+f&-RznJS
z#d`TkuY8ioxF1oSKB07%`Vk-Clhg$Zh+A!lG^Ezf&jNhqgae!%WVX7TpnQSw#d>pC
zKMGx6GdxSCHZ74!V=n$=>ae!O(09%avBO;jethQgsk}3P!{~)F&m@<IqHvSvn<(fk
zmGVR_kYxK(JF0A#F^lWU`RD$tkbt(cN2@`5W@+4y3wS>U(FXX{ku`jz=s>~F<j3XI
z9jqnBQBct2o@A!gbNpKOp`^pE#I9RN<@Yd`m`Tk~<M=%?{-0s|+7G%M$`P!1t<-7R
zT(gbQqkifEEKFiwyLaEMg^5CD(X%Dxrtw<jy<*a@!31vyln9567+z<kLZ{+%G-o`I
z>vEI7r^n7>radlZhflb|&fdq=#>liE{s<5mV>A-)$uYM#tg9#}l+f)3CRS#|$&lEt
zCVVtky>anYO%5cBfSsi`AyYfDNq0b4D!eZIyvSB>H$65jx2$~5)Z>jo_Iaw&3@2B{
zEm^fEcZ)dbBbH&mkj)}Ul6DK;#;ImmU-y-{;zD!nrS|rFkQboZtR=?TD6^mA@WbSO
z;>X4chn&u@vnx)Uwk_@&A2u_?>}M5Xvsf4{_z=&0@~-#sdeqylB>ujAK&RH0N2zx_
zphQZ%lj>QxE~4GsQnUYEf4!+bbraS%=ianaBQIiv$#-LR!0^v0j+`DcMeTFUfyKOh
zZgbfR^1d-De9$l9szKQ!%cg$!+PcaMmc#B=VJOMA?t<c%T^8=9hBT+;@fuV2L>m_}
z>vUDM99_EgbvLtR8&^IBWu)vG2zX0Q{HvGqk|6?95LIA@`1NFZk+wels4XejwP29!
zN8R%WJ?!`Ba_E!+p$Aw|^!<$0xDO}$981sZYotI43%A{)cqSZ0lC8W3q*GRnqLB$a
z`vr=}Iif=kob$_UJ<%dq3-Iq}o}hiy92)P8C@Ad6j_`m|m!5jWJ)5WzQdMn{@yO4&
zAMF+v;-&84(;t;fcqrP>d_897?SjjVyRaar9R*c)x)!#<P005tBEL0FxRNSD^2SSF
zvJy4yRY{vf8S}M^=Hiq`qS;C7FS$@82r=OMOVx<ImBul(X-V+DPBjXuti{te7*liW
z(@H&!8oet)fcm2V7Z)AQTx_q$u~$@=Ok9M!7Rv$aEn80$`W=*O$8PRm5qF#Ux%5G`
zTD>$tq;G}OK!p9~F@Uad?gDz`4903|VKL)jLM;b4fGqgn*8x8I5IZaE46UzOc;2=u
z#NrenUdHt>tNmW9TmSEV9x3mH1J&8J0dQ*hql~!OLai~awrM$vH7Y8V#%p+|yVLIp
zp}n((nOQ+6b{*mMzMoHv!`E+gpn?pkf^YyUb1J2Av{Iq{J1=cnA&^q`V=*tywy8$$
zIYmCwqv5w@0=>e3t1L6`UeWASXPfX=Qq)}M(ybZ_vhDyL1+Gz;+#T4Kn0=$d2S<ic
z8wS0w<aFa;Gc}@wKK7&7U)@X*CTDRr7h15agrlXF#Qf6SeF+d=@8QO)?mWkHo4Y3G
zJrGJSXB^-~O>5)@SH+m4`zu3kOR+^A?v`fil`3nyS93vx>(d*%(fh7vk)zEbIlJ{f
zQ^7F1#*`@GtW*<bCF1jZ5$9_yHI?CSycSHf*4AA*`XQ+gNfIW3)|&9UMyM2#d#LjM
zPMTdNsf@gDCbpUoCuc|J{Dx2+e&d2;0y{>ZN9=VKE2j2Yte7OHnryONnYS73CPEwe
zMRuTQEEU3$N%9+a^1LTIxr0DI=c6b2S#Jz}@&|8FSR6_=f)N|&eT-SMzsUN902=1&
zE)$2c6XKC)*fB<ntZG83+saf|R#sFZCu{FOEa)4a8X6C_juUw|4YU!KybcmI8AN`Y
z+w&pi4e?X+Q!17G8P=gcr*sYdY9VOoz&hFTXrzit)Y=*<O3Fx1@zN{x%)e=(S-A`q
zr;e1_wqVch^l0T87v>pn9i_;twA+4(=x0qYu>BpbnF*t2dD(X1;@Zj)Fi?fo9d8!#
zUUdR>Eb=2gLZ^M5(K;(ekdtDo@77|fXRMN>ACA_j>p#ENg^Txouf9jSTlgCWx$-$b
znj2d9UNgHC3(tqMrsm_L5Oi|D5<lQ~WLl$x5ap0S8pL9sv?Z<1M)PBW%G=Qz{rlZ2
zFP0O6{HlB4HS0qei1BXEOK5Pj@>f(&e+S418)H+ZWJ|H%Y)|lAw_flohaHDaq>B2J
zdg{c-<(nN##BTl~%VpCYzBScO{Fi+@L&sOFTEUthqGGU?kvCp&J%-2zAeR|z94mj<
z>dY*Gkts=KN^4Zia}F{e7H`85b*Y{<IT^||q5IU<4QUjeS$>;+QQXheLtavQzJ4}2
zpE!plhFgG>FLyF|K7LuRJw0i}Jc1q7+f?Ux^__QMsZb`ZToZm;m}(ffEH~q)+ft2|
zfXV#u`hqvv?^4TGT+Gl6f!l(Lek-=DwdL8XQT{YqA}dQV2if59>=&b!SCWxZ`NVI4
za$kKbDlY=;Uk_MOC9J23iEpv+qwH{in-W=skbRMAXuD8xK|VGYI$Jk0Z|Zo}VRMO@
zxoABQ9a4R?%o+$s1mgomc(h5x;T&{@3~@SzFS7)I=j0Bnu;&=dqccqE0eWU>!YY$r
z_IvHZ+<^z$u@x#V<MV{nbOt<ypFTpOX#M<x&O0u7YFFUbykOlj0mb>sr4cn5re2AQ
zi;p!9ebZJ{l7c_w*u=lIer(O1%37lm=Y2hlH&hO|O2yn-lBWa<cT;0mts7q-PL|8r
zuzlAMMku+!{_B@g<ng}J*&|;#B}!8z$rN$WVx6-kb$K^sQ3z3gR$C#;hy!TUgp?-;
zZCbLE4{-lf>wZw^VOYDo|NP7^tUV=wc|!8I$ONS_+>z>&92ZnB+2o`yBpv=l)k%0)
zHvV;d$#5gNW^faPS}(8A`Cu_68P2!e)MVNs)MKi=1R0XX0p^<Pw&6uv__(D<+^Rp-
zlL<E}N>s2}>WfhhB-t;MF2_^`L;cq($2WPE=exKFs52vC8dft(1I8LDYO%>qcwgpu
zjPhB66D&<mo}5O4fwo%k`6ui2NBJaHnekk)gh#pjR3B}ZDuZ?{2vFYc-;7aue&z)z
z?P0}<ugI{FWhAla#miPpu$1WCpGRuap|35YzqYT{qB5@pvcp;Kj5JF_O1uKbmB%Ge
zj_eMHH|GYuU-q_rAAb|FY<mBbReWru9w|ql62!LKN&X|8tp8;Fc&RlpJ!Q@m3iK?*
zUPF0IgU;fShmn2)?+}M-h)$*fo6amYQ?fx#Wi9#5Ao}|BSK%VymmWveiP40rBC4XH
z0by6#MqWwnt;~}(uxnoB*yGN~<E>f#o>!@<b)!AZ{ch3=f_j|=%KFv`w=0<qg6}(z
zcP~-)M2nzlT_hQ5S(qlfi|E(HYKMe`h0YiRrLs?A=DuHj>UnfmA@xdOr<Dc)>nM^F
zU2QUDrSp>3@MjWwt~+M3Gj=ws9IjC{Cy1=*SEMaA4jpSgAdc*oxzD?#qX|t1^DM5*
zDu8~sUiF{UuWXn#Ve)I<gmnnypzsXwPWUNa|0>j)QD8ckhp7En+2u~x?VY2x!1AQI
zZcH!mmB#gY#L`*lXaLVHlYC}MMEQ2g>_*CEcU;j<yb_&EH1($xu#ll)vN1XGw?;(6
z-GcRvuL#JGhs~@oFOZ?Ri{@M=&zv!1jAR7`)V<iO!&UBrBK$`pPrGF+h{&@)(GQ5i
zJFagC_=!RGhwGv3j4M}VWLSAQc%Ao>HRf<h!D6Wjvy>rw8SSc9KZ|V@!A>7Sx%1($
zp5c+>IEH{R3(-ncuItyfl^Au2eUTo9SsG<4io9z3siUoNb8jWLO@})lTu2l&O0A_I
za6Fa=KEr&eiE&jQ<sO$1a=tR1v}vi=q0BMmKB2O?AGxPtVIA^R-%0;z9?fpOhlAAd
zGR4AlSDKuOoI0sp*(x)H2KwfnL|QYdre}Oc4zDCI-??));E=6^JiY@Wz6$?!{tU@o
z_HkTi1p@|EOu5Hs&tYj`6a{64w`-Y`<l=kn$JyT~RS1@F+%10CL0=sABWXz!GYXDs
z3NST3bX|9Pcm}@mg<jcToNn)&WivF(yCa5mk#kxSOe~oY&v!oD;TKcg6OHrTTjIyv
zqiT7CiUMt%#K_}-u_+ly;3<UOe(?q}Aj!^>J5tq&yR2*=wck@riEOr|rL&^{xM4m=
zYj?q0cLDew%NPeh3+UXi1J(Ihx3`Ivoh9U&%qY^0iHmyeQ_JNWyjd|4WuqAQouYu(
zXJ&Fi2w=w&qnYivole>AWlz|Qpxm96l<}G?NcGO!CUa&{xy9EFZ@jq##S(;{gm5i2
z4Na>|Z^BmQ!}nk<cxUqt9`Pe{bq%&pk=s#-dVh)`F-N2n=c+nLGzyo|H<@dvirp?V
zp(P&c#yWnpA_`C<e_&B7i39XtXq86M>aOK>GTjlywxYAW?{dtP>82B-3PnvWES^@W
zY#1BlngPiRJw^vzE!Z$=;8m5F)lno-Cqn?+6k5^2%6jvc?XH8B6aDA?cg_q`#x`OB
z7c+S@YHe+^gs}oZMNDGp$uJ2j^%)}z<A|iWzNtWl=`b-$VNSWeI(^7ezB!cbs8{sZ
z!1zNFJ)7lI?^jlMhG06fo+B37(`d5A1rHe~(3+`GR<PuBsGQBpg2|eRznDCsgDHLb
zk)@S<U-wuIjOghuD~*yg%{4=7!pCxRXa1O+^77^zST^KY{;GGa_lZ2v)L-Fj{F$k)
zi13@BMOkWlx%Sryg#1J66;|^vQegU!4_Kb*0z}JUB-##AYHuPE4y^0><z{K)P`&3m
zN9L6)F8c*@5#OAsq3KLyAI}&TcP2S@&`iY}>hAx}T~h^7m$b4B*3xQq3K^IdotJPv
zC@($Yvo*S9peE?~cIxxsh5yvlq9@+QOUb_~mpZw+$i$lA-C~CFq|i7So}>a)q?4!C
zq#v+oOWlEcD9-z?{({W>_t@#aBd!E4vA!~<z>E33do2#kE5lvDTd4lEu!tIYohMUr
z0tLTNp4u%@s#A#rebt(-UU7Vhv&PB6Tzb*oVa!y0d7oFjAZJ<?V}87=S~BW#A~`!I
zd37L$bBmjZn7W8-pQjQd_ScU(`nd+U4c^dbym%k-qMvpK4Mc%gwMbTZ>eqmDsPdjp
znXJNm6FacYxSMO?oM|4-yxJZ5!nC2on!BN{Ke=w*Fvxet13*jzJTIKAWLwpR8=>18
z_>~$DQY6%w&DF$CUfqKS`GLLZ8V5JSoJ1^ZYmXq30&hM^*7%ka`L*A?t+7)HP_Vld
zN~A3^E3OG2l`uZRZ!HW1O`n=(?5d$f>{CjLl(;_^2*4%t8cf}dK3@~sbb7U2QMaOA
zfWC$u#+%PjEN`pLnNIX66(h7KOKz#wzZ%1evFWeh&xU@!fVw%a$xzK48X!(5%d1!7
z=x4VV`D!O6OMb#6DcH-)SmyK_ub$P7{ivs|cCwCgC@mTA<|H@#>IlQw!VM39*G&8K
zn;a&=SG)rvMW%Gt$83Qt&62BRnWPWbT5G~U)$qpHjU(>Zhg?+&R*ki(8}rpFT~!ZW
zgwKaQEQe}ga^+;-3QP6}uaXHol6zF(pu6pvQn#XEVKzun6wNN?p93;5kG=wk`cH|x
zkON=w&!DPTgfo}%@h$bkQpwwrnlUDu;2^r=gV#;o@ES(vw^0(FAnIt^3NhACUxrdU
zVKdn(s7&uw@IG0?+PJ$vmDU(X#TAX_y`~xSL07sDj|4avl3#T82NfxDNt3`%b8!HP
z)Le*flqPU@Uj$RPe>rmx9*pIf^QtzStnscXQ+hmI)3B1YsWE|066=t{0r+)Cdbu1+
zUnP6FD8{@eibi+XTs`&bgHocOP1Z6G)7#4Y_D~SaRvZ^DbMqV*0zG@jRQKxG#x5cK
zfnSg9O!EL?28A!|$KiX!C6e{2@(}okvO`HAHxrNqrA={tlIFp$zLI{yXw5{;zP<-i
z-_Ud4ovhBDW2Y)6%it?~N3BbCt0m@3GZ!<MS{bb-7KnoOfK|-JaPL@zpZ1&!ymdA}
zUW>0!I!<A}|DYP)cQ&$9OBL4$3Uh+Z9%@gzGs|C}T7=wkkb2i~=JcSKrtjI;JY#0+
z#-HtcOZF9)pl<@#b^8>FUsEr(zC+gUS?8cix<QwF>MM6xL7D~=1xTu3C-<lc+gECy
zGsEKh!EW!$4b?WyJKpLgWq)z&H|<w#(u~-DU3*OOE2qBh!6(ZbxmJmL_v%SxKeEmC
zHM!(xs2eZ9fV=4C`ib?BiC5XyQ1GtHDV?5i`nA8%2LDOA3&#h0ert{Z_gL}qD?F@T
z?(g9vT@qCYwiM&gjTJ<dm?Ea%D+nn>+*Rjj^A@bf0U(L9tHVs@turrg(n8*az9j1j
zDmivd*h|3yc7~@K4#Qa(>*ExWl=YT2-mYc&AM0<;%`Td&)YIlYjCuC<XH&$ls2(nW
z_*glOZekLsQp~J3N5I_IWD;$e+j5XrI1A2CUcF$L{gBl4J8jPs_-TZF@%j<1^GyTh
z&{a?*YTnHqNfInlyF|SMCPyn%Wd5o`x+nd(v@QWAY|iL#54_<y_Q>T;W2(Nv>>xl$
zRbq>!8at64_q0zHa$W!QAr+KlMz62P?$_h=o&S4}dG1)<H||5(G9V0oxu-vJ!J>25
zf!2#fy|%?-sK4#VYsN`(DcpiVDX-_-HvHAkW^q+^Th&<xW@BOf$AIW$YgTu`R_@-F
zet9yG&h15Ddp}k9tJ&es%h+PBl;dr!y#8#3BNYYzh89;QO@WFxFPI}p9d^_2cOBJC
zr>@`;06*c{HU;OhOWzz0H>Rf-E$S=`FKAa_;in?uOVns4E~GI0U0&}uud%a-9FxLj
ziN8(lr`kukN&5!r?0*@}t<P%2zepBE3n6Rjp_J}rsvV2cNXBsNLn^WzMw9vw=fgKm
zq0<I>33uya%kD*S&BXG>KjAh*S2@kY!K3=q-+wllV0?B_CFfO>My(KNX2~v8=+J5%
zIuq41+Y{N!t_*>!Qc&rSjz6)Pbaa%nD)Ne5l^qsiXC81CT-Ce-5HB%3)OYQKHgWmm
z{nL5*gHFH_{3s#9UJn(GipASlh+h17)HV7qY!xJ+@42izEor2>jEcM5{R>#Gk4MoM
zD%5%^;(gBVZ~$}YnZXy*R=)A_17<CFMTE2P$!@UNT+=>H2`48XZ{RP-Am3<DSbFQj
zfDsekxSA{cUV`lh+V$xcU#2c2=f4;Rc^#Fp5(iHlQOQ{agN;FL@H*8Fhqw<w+5%LG
zmGj^-j5_=!l3BHHd%BN-ulyI3T>nd~lyfBVE6^{!=-?U5iAo1y-s!18KQ_xt$*^#1
z`Q#;l^O(MxD~d_HiCFoLU=M)G<RWqit>v6Jr_;#DoSf@S)uJ&?Qx!k<ouo+6!Rpf)
z73FS<bSv3}5c_wP@3FE4w$-6+HZLC^JW$m@=tO3dp`F%5LS#;}!_cf{JJmbsZV9vA
zCVpSJY1m2I`a-&I-`%L4wIoa<yRcLh`uwUDiNf!O;exl}eb3kN#_;G(Jzgud?oFzk
zU4~gmr-SV9-XR`z`udk`<u7HI*=rYKG6~&snF}|6$4+j65!DjVi<4=5FMXbXqPiT_
zM<ySYaJpop1qWHN{&>v}%`3@@6;=JCY)%13`*vZsum%jp-;8a2oECG77TOB8&l2R$
z89nt&-zhN#4IJz}HUQA5KLqH{8Oooiu>R4;pKY3Let_CS3BBgb#%nUA^(NSxCON&`
zx+f(}_hl&54LmDr(l>6eeZNqEE{Ly-t*sMJ=NZtr{>ecADBu^kf-w>+?B_StY_mLb
z_-3rOn8xA5goKvLBes{{J}I8L!b9%TJUSd^E%T7+3N4o0GX&M0yUb{vIphTycm$b$
z!<q&L7|w3ZX;a>2SuGp8b8))D`T@&^@ceM1ML5@y#nBL~Q5o9Hlce^Yf-Q#dfYatq
z_q(}=Fs(+EsvYKip=3+<cTB|?RUKE>-3h;B70v@{NFr$I(4$A`w5FPiTCS?Ghly}t
z?=2mrI&Vdh#Dd1K$#|JNg~ATmE?H5EuWyMRae@{zz+m)aJboRaZoA%~kguVwpS&>q
z)%Za8GrM+XPVNlE`P}hXNNTQ4#`Aq1qGIgyLh#KqElw|2-D9pKtPkyP|K-P=<Akkm
zC*_D>#yg+~Gg~sJm~W4$&B~)Zf|U2>OdhD-i}8n)`k5l)!uHTwzTX_)+MTg5+pbMQ
zoD<gSgH#QyqcuQQKb5{5jZ_nH5Q^<k=n$hF5!sy-E>4WkzE6~%7L{RVFlpsm^^M9C
z@y@evA_W>=d0UY`NqyXx|B;4_eBqgp%#?YP2ewwU-R5>F^XDYE_!I_v3R4P45tYRe
zqofXcOvY-jWQ#=>-Q%tZFLmO3n1Yi(#C~%a!*A(%gZHK#Z^Q7pM@4uPT;#TG`6tEa
zk&#>8pCXSsCzI_Cd0o@-+F=@A2N^EF@7nmVfzNgG{rf6`e3(2E+B19&#oNhfa&I5+
zyR<)hTY8@2b3F`N(re8eq*t!|S?q!!klSytDJW5l;M;?0UuO6pZwTujWvvw;%a+0;
z>=d?R!Vm|z;z#_la8lLbw`9_+-ymY^5p=w5t@c(?GQuC(f$ykxyH*}E2Mh4~m3qAI
z!#~AEF8)e7yxluG?}S|&uQA-R>t6t+WCoabvHGg**yH_M19(XG<~SIibJfEE*6_y~
z{82;Cz1b;3`D?3@1;$fMyPZCNRyXTDK|0psa!!0h>LW?m{DYoY7q#-x%=DH^yW@S-
zSDvgzxbP<a<t7tX#R0Olpm>+tAU+q3f54wrhLf?Cmfy5|>67<c=IfsOr*E?*n}12Q
zho>yFCf7AejR%L4G<;*ZgG`BAyFNLs)_i3==6=(yX=y>m8zdSc!a8t$kHiARwi&G1
z--iPTNuEslZ}~#D@pJ$iivRud{sR*5Sb2Icm<B>S*tN_71b#Ix6nWF;Xi=+aK`-D1
z@#kv3WzV7$_#NBJ;fUH{s+mRa+1ALqeJEsI=7PihP1FXPfU|+sQBt&PB;ddHK#-;~
z;NMLwe|EeZ9$5Nc6k*J8q{JMM+SvB_yo5~Vo|ujkbLR@ri*7`fmE~JX4;zmW)7!)`
z8gd;!H{kzn!~_3b&9?0qN(U^%n>ZXm0YCP!wCVWVZTx{Hf7FzxL<0U-G#Glg+WOwo
zL2_`wWTB;oZ%k8Sk#&NeQ&-G5A{YR;$u$i!tM7tbaR0UGu|jea8}0+x-71H+DLkk{
z)`c<T+vmSZ`XV{<?;xxH5~hC9d=n3E0*1l&?m_VSYD#$zHiz|`@xnla5c)H2<?xEc
z6pwSt|3cBnmlH_(|8aT*CBiFpsb8SK6qaxR<uW=iR_xOjB5d0k4sa=i@4DC}jPd1k
zte-iOGBX$Td{zFSkR`up(WK2Sg~1lG*!f|6C0_Nd&@T>^VfOhk`kz9_mfG_FbW?_(
zHX?M10yT?0I3pPm5iD$&aC}|qZS9pZp~}x$inMZjha3sg8IU!*`ywg5kq2Zq=W-TN
zW~YZr((&ie^7}P!YStsclxOljzOULu(2?Uib9<b8)ed2xQPPKpzv1?W#CU>8_Kw%g
z!KR3zv@&zM#XA!)iT)U?XQ~aA9<^_lZYN8(U%cRDHln}t8<b-kKszDm{_9j;Rlly;
zt1PcBR@PK1kbqNVz^De6eX1F4RlIA0<v|=|u2E6$q*KBmXDtFRYE$|SmV9~B$iwxf
zUE(|V)qb)?BOLf6)Fo+s#Lrx#^GCBpznSp10m=O2Mz9Ph-}V67jIQ(vDQO){T-}VO
zk=e8HPfm-hV$bHjtcR+os1m>Qs>b_)r9;r}PNe2rjYBcK5l%HXB`_i#>7~0g{T_As
zzo%3$xm-CFHtKCrFRd0$$^Fk?AS_}^TLaHi0$gpSHGK<NC8vLWQsQE7uxvVa`PtwY
zm+|l?@w@Y->g>Q;A-A1+Cg(3lnHushjmy6W)RFll39ha<cCKDa8q{6?@Ul`++f&M8
zL%+082nJ1UDxA+=I(IFaO$>XHcyToiz0{e<`mu&!rM(W$yJqEPKU=Us$n3MkCYw1a
zGBZne{j%Rkhcc(672N+>GdmiN14Mbf%M2cFwnSF1ibN>f_Io~B+tk9_uE4HCvSSi$
zqSX^t%PcU*6Y-8;>Fx0+kr7UMMyVq<Rnf|rU?)~BB}8g_yijC8qSUWlX&E%>c;6GS
z&ZhU+(fq;HvnBSv7&PCHZi=|Q6+9GHCkzzgmeC*Xt|jKgU<7lIn)?{qtrsCwmSX(L
zr2Da69YyeOor7|*s(0(D4gsv_l-3MWt1@lTZv}OYP`3;mV9c=vG;l;g(z7UpvOyen
z!V$4ED5?GqR`OCy?D)9QiFfIofverUy^U$7UQ@;Op)ZWa=w9ih8m0wgVUG^Wwr4iA
z^NDS<%&bk=X(prc_`b;A0r7FKpWEt~7XJVtF{4W$gP1y>&OAFtwv&~hnQAr0UjDh>
z7*FD|D!I54!&g^|+8t$Vv2)p&a^w~-M4t7YfQ}yO`HH#e6=dHBoPnt=GO@=SCEGM{
zGCR|~M_end=I}RR>t;mN@h+|<iOJgDc<9D8@F8eC|Kz}B%q&6ApYZ}};W>fyjf-KW
zZGKeb^}+rarH$=rZ{-=QkykFQDJkNf$ZeJK@=`K?SZ5ouJhb*8{wodu0&7fj7oqPX
z%F8hIXDuRw*x3moN&NhD5}}&u?(qA9=IxU3Qnk`iIl;`&M~`;x8S&0#EFtDORBQp}
zMyIFD{-0Uz)Mn5UR+FZGv&UdkAnw#^7v%P@m^r}sw{!eyJ!LUn$@?|ZDW23d;->R#
zUh68it*zy5DY6B(4I06Ks3_#SW|CDcqK-3*ThnN((p|%IPhFEO$g9wYWdj^49!=lh
zt4dP`(j|JYMaulLlde1-j)lLx>E`<#)_Mz7;blR>Jsc|e)8bf5U;7<VD9s$~3fQbS
z`hxVh8<P&TEjqXc!-I6LXGHtA$f>?%A9_4AZUb-i@lcpNGt5W6f>L===^`y!mb0mC
zrA$0Be2W+7%+(9~<!(AQol#RvL=JNt+^_xQ7v`-L=$j#%*IFI4@t}<-=C$s^TNM?k
zOqo7=m+dTmmud*?B6p371pCA1FtQ*P5|t$xyQ_uwkY?RMMD{un!}kU~ers&0DBeh*
zoj-#WDKJ2LMG#QT`%-+jJORm;1&pMa1bi*|YtsO*i-m8bMtyM(grtmuIB{owOvqj;
z=*bioCRaL)vVGG=9xNunw$=K#CSEM!9p}@q+^x$Kfz(t@D0rjsOy<4}Q=z0;(${aU
zy4TD8*<4!7u3axp1rr~ntEKxhOdF2DB!9LakZ49?Z;E&l^eU378BwiJSoGl6ITj?e
z5tyv5X<E=If!*}?Yr3Gi#5P~CFkb1{PsyBj-~f1#5cht|X1FAbdTW~H?iJU!L(#~i
zJL_3ARD3BV;?=Q)v`N9kBt?E&E#p|)fpJv9&OHlo)b7`I6HHfIG%Ebnh~HctCKcGU
zz3wpwR`demeT0EnmYr0e><~ZNm0=2dP^A9GvFe!9K2N5mGy9^%wa{hCo4bSR?ud*Z
z=dFC-kwQJeen4ROfgcQLl&bl@eEle0I?Am}x~@Bt$1M9}_lK87UE}sNTKm56RwnOl
zX0R5zmbuv*T9((Ux!6)W=kIm_v>4rQWk$<l3YrzaqOv=gDD&-BjrQ?T%cV}~df+RS
zV6BAwn-W9se|TK<9n5?f`oYl18U<opB6p{{ZI?xkj0-<io(N#t_1QmKbnuk<kdi7&
z%mAu8lC7*0t4Re2%5p+@`GWg@o&zJGM1Jx}B26Sg-T28MW2euAA;R1%R7gs?Ifd^{
zta7X>v-$M-V}L19@NA<930eXX)=?QCC$M5p+rg<Duj!46D^#*OOTE*fzFxHGE%*6U
zGC4CRRjmj!`VPG>Vz4#&L__lKY|Hu}Y6*ezmxD2`Do7%^>thtSP*&|FtD3!~lB}V#
z<t3hW`8&s^HQ(NhceV6-aTyY0QG$WLgV{)~mDFJpYnX*d&YVy#b&QPlqUKvt($VL$
z(dUJ}b1G^oY4dpj_#+D8?Kw>%H_*HYU?3Nb7r3pJ6)_(?*s;xI#!z_P3eFg?Rlvif
z84kfh;+eTj3iJqYSo6(7^+J?`;<%i=CD*XaS);*ck4lNMm@)d8$AWLg7PJKdtnG~P
zBd=F$4i7%T|M!@F=Is^)@4)JTPS2tDPNvtt^9C)g)c=4{?>Z;ou}j_6I~KeRjTf(i
zjuib8dLKDQrXF=RWe^y#KkQ-oimY0d*IeUmGw0Pv=0ZxAe^)4x)2%rEI+oJ6ArQgZ
z*g2-|{24Lyf>M<wA_px!CLx8C5G|SCW?{=baIP`Ui`A4!ROVRDaU9~|sxD#B?+OQ0
zJ45k7F_id9wDen)Y`YBOhm|I4-|!Q8DQ51pkGJwWHJZYS*W|?<9&S6F*rE4o(9%0K
zRl+7qgAdU#A%FkFq^Z3NU|V|a0=x|lKtHKLMXbuZ_Apuq1*>TK#yETLJ5QcU2l+Bq
zQT?urogctj@us(`x$;b|4u_U2@cU`wk1$wm(!@xsHjnu^2&Zl=pssZ5BD<%^R%mAW
zbGg%M6E=d0C1PK@y2O@eH{CphlR|CcZup7Tj+NuJ-xDp(t|?J<Dl|86=2jP!L-x_!
z26_xYUuWgL96&z~M)M#Piq{07jEgAgj`J+R`nlh`(@bVm{tTlbbC0yv0Lw3t8mSGr
z0ofSXg~Af;@doc9PwHuFcH|)f#?AWaR3OGxZPXna&fCqx+b~&k7%7l%G9y*_utG~g
z%S2XP$Yeq~nS(Cn36rwT6Oa|3E=e)A-zT#yD`IV5<auKx@U{`&Xa9YAC9{1fN@xOR
zClPqOUi+o~XpUKfjp*RB&-@|<0@D>;r@v+5nBuZX@kG<JI=cJ(i6Y<U@&~Vv0r=g@
zzHdn$ZB)rwBV2y~soe$4Rl5R65K9rKWG{~P4^A_A77vVGh&mM=vGa?m|B%<VE%bUG
zEmsEp$hgU{KiRZ+HL)+Xl!*h}G4>y;@2^}2-fhO85<rIbT}{YY;GajEj+E74sMj33
zXUEj)qBxRwOcoCowW;Q-Tt^)$P)MtlLlW$eN#Xk%-;$?JBsybXwM2QIOGxGH@O|}&
zz8eC$ANcBlhLjA`*`Q8`k9Mtl!TgSlc`~Lu4LIxgrm<d~Wz4BG>$UQGzB%Jmm*ti4
zEqe+j{M$8o01+0c5t2GvJ-c0Ykl#=f`@gaG-ce2cd%7?R0!mX*kggO_X-XAoD$+zi
zL~4MjbRtBg1qlQ}dXp|71nDgxHS|bVdJVmV-bttdLVoW)Gw00A`JH=b=8t#Hy7#X0
zN3vG3l3l)g?{9fNPf<_1{jFZFhF>IavpXG6$9zwl)htpGSmJv6Y*el-YaOM|?(ggF
zXext_l=KLLQT|Hz@IlSve0K2rdBh<tESz7ABJ?_W4CRp%1v4j&jgmdJDq)}JnHbFA
zuZXGh+R!%l#1p3|&nzO=T)d}_aQNNvT0{|BvtW6mnwBJlzlWOEp6l2061TasJIh9c
zZOcRD;OwESk%|;<9c`H<=)1Jt@mn(JOSC2#B8(m9g6XUIg_1ikMD00m{dPmLcmp09
zw4-1GMgJZxma67C5zVBl`jFCo?q}0{LCRv~!3d`LB7BQe>1bp-Ie$lH^cCjtl#~)r
zN}hzkj(&2pv?w!?90^E6B&e~17yq^R8`zY<J#~1RUH-#v)Vg(xH+fLw8i@1PZQrd+
zl`9Ve%bm0AW&}0r^*nL3;*GN2fg3MNt1m0Z^0lJa-(JEIEgm=BIG`9E@c<xD&V0z|
z$LZP66~ZV7{1Y*^I~8LV^Ka-1n|YqprC4T2ERwRXhF=zD>6oRK(&aQ+s|tYt=%X`a
zV52smc8tFx|ACq&tMxB>E5mcCXe;{54tzyxYxum5&dyhp2x;nxy{ZPyVrzM)sq8)H
zWUa-;B2umE@y7gHOu!~t5+<%SD(&dfeBv1S$*bIScw1VoV&iy+Het$aO1@J?^;&%+
zNZxVpgG|u*B9g(dw6=IFZ$(3USfTmq$Mh<E=i|Jw#^0#%vbwkvTojFe|02f;NQDTR
z>!T;yfEiZFSNdI%j9ucXw_+`GR2VXU$IEv=`%Y*5YK8q+JsVsPT$Y22(n7WU`;&>o
zr0?({Q?=zC06Q^Qj#2VPw$#acy$unA)|~)80gK#D6@O3!o_YAZJFa_01o6X-2s@wk
zJ1i7~S`r#rDn=c*G6~kiG=?W8(6{!3D?NLD0qB=E#9mKn?w(jC6QHdOM)O837^a6>
z1?b@YzMw-3A*^)lf>#NheHFxo4yk;`AOHzLH}Stq;J;LGHV1-u`%bVXpTzg3Pq-C=
zrzTNZM`{=LtOK+pT6@CFDEYr+&9OoCjX4sVdCx=GIJl0;m+)KrC-e~3Si#<Ss)$^^
z1&7I)piWAMxmy%J_p^o;opmju&>g^Nqh1yL?1W#(x;NWV`7G>=RCe5XvVlO7c{hNu
z2d;^0kZlK!ZofL6=#fc`yzOfK)|+c)dMZhEdQ!T)eNe>0O!!AEM^eb$57l!d=>-fI
zY@uoPf`a%P6gpG{uFBZpsZ^2&tKE5&Vwy04N7X~TZK+}@m*jhvo?ez{8|J$T%^-`y
z=@L7zCl>-^ZSlL48Ut^}t?lQds2FTFpHp1IR@Z&Han9h;i=geu_xH^i6Y<JT{cy=4
zg^NY2RTg6X?|j{ZS`EL<(1jdHR_yZM>bsvh{reJy$a_G4l%gLSV`$_wgGPH7H!cln
zEL^W1mqoab?o}8q*pQ*cC7tbw^HEGBen2KDI9yZ-X0dO4P#&g}*7Yh@-|n@e`Y$`4
zU5(5b6aLB=n1$M-?rZ26t@|fxHJso8AD*EDCI&4_Tb`60*atc4`0ryaS6NwO9+t7k
z2s{qDQXecgK~=K5xLmt#bY!}RnM+(lAV1X3b{#-<1-%@J#(+#CI^aJP4S9wCplFf>
z3_ZPo>r$SOm=8BDBd@PM>{&#eWfqJ%9t+ikm-Gv?p`TZ~Tg3G20013kF#&L2`Z<a`
zxw(;F44OxiVrH-ZK{3X$_XovOMG6<*awsVl#GC!G(o=U6m|h5AJJu{v|4^#rDnhK^
z{DZ<75&se|L<%PJ0E4C*h8${I_gBfnpRpic*?-rV{LcyQ{=CosN`UuJ9&up1VSymf
zm=4o?R=R2N+tb$m`>J~)oy(QMw$0tvDoup|mG0=$*3VRKcb)gFn1%aA|3!ue>ucMg
zh+?+)k@DB;&#Xq*2YQN1|7PxKPipMFo%Y%$o}JxPwAHVkFSUisOwKb=skQSEUYC{x
zbuL<Kyj3wIEnmmEx{VSO)7ui6nQDIvGoo=VbE4@zY@az^=>3Djj+YG<)_gR>p#?f~
zX=%J%eX0+!;KR>^*JdB26m%TAUb$|2kuxE#B9Mk|$@iI4&pV=W(~F6Xgh>TnDDI4B
z=Hj<WPu6_SDXFU77^;A;PT+F?YTw7NI&`Gr+H;N;<|<Dy_{8AE772UzK@$cUh>A;g
zALZsxFRK_n9%WHDTHXA6s>bUv_TP~0tus$D#;2{Ldk#)Vcx`TH!IH4kahcdYYCw>x
z)k!KMI>{sU$FNw>&mTi2&Rq>#(%^b@bvu&(GyUqlgT9I8+1A;MB+UY%mnwlNppq_%
z+CfPWznzUT1<`J|)!!&r3lj-)g^HM%=&4Y~&(x2CEKe1`9yrRI){VwELdndT^9|4J
zM8aQ~Yo8$nL))&Z8Q#l_$dnVkDah!9pA7Lly?a)In++t1z;)55yiFf-KP!!#A(8T1
z(V&jQh%{(d`LDjj*gB7y>&t1S-oOH;3bAVlj)<_^Rm<|M0t%X+l+UWB`yLZP1WO&_
zcZTtSh1y{e*MrE5l}k@qWsM$Pk$NGsHF$mO&A3q){j)%k;=&oDRJb%zk*yhp(D$Z!
z5joU65m0fVUjEuw4aXh5CyL$loL7p1^BYC8oF2{j_0<hHEYD9boFRU0EkQP15|U#g
zvd<9a@Eu1H09vrI1Kue9J1^BgYwlhZE008>%>T{qz>ZR<UUz<_MVc5w=)(;(gn`4<
z$_FY-lulJf#4B2^mfS)vWlouDaxyeo-DLkUd9i7-qe|}yF?I{lCVdjm{$LZ1IaPzp
z48JK4dyd~uDux&W`~gN1D7QnTY}&!4Vowkm%hD&TDG(!PbcsDoW{}#sBN^6$>kC^X
zelOATc&Fc*C|mV`2r-SqjVzbujZYk3z8yJg1p4K;X<%j2(juE6F(~O7CI;HBy_+e-
ziaa35d7qrke3x?>e3cLaCZKV-2bL=sYU29QoKr|KQOf}D5h@+wCLN)3s!47c5GxUN
zv)kUeRqcCL{9uC4rh5i-mOVfAs~h2NFgZm2yY;^j#LTF<?^uw`^cRF1Ne+I7GzZ9M
z9T=V87q_>tJ3D9cZrkGne<`jOET%H(Cx7Co0;*X%Y^RyWhDPBe%PhDOzBqSY@#9j_
zzPAjEVmF~lU^7@gc)`$JVgH5ZJ_>{&9)+C^kP3feAOX=Bn_1pqbIY%5t^69<aWV%@
zH?Bmdjq~bh&P>qcj}dw2-J=MXv?>NSNh%JGPm+|;3etINH@16aIIH=XUFF8VIe)V<
zQpty~a){j5%CndpQ<WN+R6CL=a@hUV#2?PeFVg)afGg(FwFU51M>r%M&Pi~}ou3Wn
zFh`M7xjqKQ=fiB(>jvu!gHzzn+UyP0khp}$TsBuCaO3aW0u}UIS<B%dqIwQ2rTrwb
zE(eygGcOizpHK7|Z8fTjwqumOegkyX*gC>PM3uLfQfxQm#az3ps*Cm$#d!?=j^N3O
zsMvwg%cIxJr&@p#A`FpRII7L5G?~#TNr*#SA$by*<^v<r!INuu=He?W2AnFHf|C=z
zoVcoW%Egf9CxlI6j(PNNS!_)kKU%AJtL78i`eFl}ILj+Nt<uJ^_%SG8|I=BLea;4(
z;h_wgDWkCKZfj)0u}jTU-_LIvp45n%yvoZTa(gI<h%avRKJ`p)W;D2_RHB=PTsaax
z#Ee|iHy%MrG4P$f!X*AGE&H_XBsc-Gvdd2TaZZSW#a9x1mzPBg&)aG$mYYw^ZA)H`
zFMvNBKA85{N=}*4IFflg(AHL=FB?x4j%Hzq4xSgXF<R<znXt&7S7soVw>PpjdZFZ>
z50g389}Wz-F>zVIlGkMnUj=$<cOo+CM9K;UB!078olE_cT*@_^DeM&2AkO0jpqq@-
z?vo&G;a7{X6FSN+OQf2_1IU?Y4y@}r)@grLq;dSgun2W7^h2%p&nml%W;P$vOlxLn
zBfiS9%fG+K35aR~-GYsb2P?KbKn+6*g1pUjkjYsg{p3#h_gojnsG=WM=v?NVxkFDe
z7rNJ=K}^K;{rrQ1jr0OfM$K!{!qq);$5*jFcL|1>&a4qxT5eah5t0^)9nm7ub=|#|
zP(gajH^Mi(bFXF%yVrq_71B#44o_pfMJ#o6oVB;Ni0u;Y`<TVr%q`PB-RnLBjg)xa
z;7SHg(6BUMTOsviB4Fwx<9EdfzfMU`GYfUf$Cti*Fx9+zd-02K3h^S2my@X6Hp7P=
zEiT&HzqpCAsT#oGq=yysT&PZ$n3R#;P_EJrzJfgM?v(FuziYw?X#)Llmj!@u!3DpA
z^HOF~$$=y!J0TG$hoilaqL-js>M#_)FRF>Q`;k5FxHLAj-6NvL4VpVUnp<LN{NeX$
zZx?3z2>C*tPn$!NvnaZnz9{#Gf4~2lPDyNK@B!w$a{HN!+KBX);RbFymdb~tA3(H<
zFT{Nii((f!{Dl+xeXdHQ==&$@b6+c6IdoEJd`Td@%L{x-1W9$ikEoTKAgXn$+sej~
z+L&+kPOuuX7^-fUJ}L1?1tg+DHF-bqh18(#^hZSM)-mK{e8PK)YO&HDQv={Oit?Cv
za&$)H9lQ<70}2qml`o@|_+U|HgtpPJ*8?##3!pSFP^Gu_Tw2?v|FMWAT2lGz@8yyh
zuHTiga$I3lD3YaK6GbcrD)8C^J5orFuIyv9kJ7McMroxV@oCkE)!#ObR-Wli4E#x4
zQ<;Ui+s)d_GXkSvGQcE>1xKEkcTEHMD;V+zMKoe$>yQ;yIJWxGu{RR{c-JPpi)>7`
zx0j4FLTC&<-k$%ot_`B$=Z^+oBR#{166+2>5>MxKFT<i*N71YvCkNSCl&wbC!-T5-
zDLD;NaE|fRc_uN`E(!VGvABFbp`n0uail~+K7IY9r<pj~g11j&BM*;=RkBuMSr<VG
z%zgZrZghX=hw-+;Z-nNEC0gnF82gVMN#Yi1wG3-AX}>FGGd1hnK^Rfpg4+eAJPEZt
zVQ+Fp+j=I(Cp(<P#BbT>Bsv643xsRQb6J%N#Ar&p=a0+ty`9voe4P<`hyKrV!T<PK
zIDv=)=DO&1<wI(~dVhe);@lq;RI`N9wjEvorLX`%@UGLs^bOe<^kHP-8}wq@vIOQ)
zEbFucZv3N`EQSP73SvMDyr+~Fh}EEUM$1D-;0uW5eu-hpowF3gV)EFT@JQaDZFv68
zQm2TC!GR+u3O9YX8A`vdEse^0A+Ij_{blDTP4;ij(I5g~R1|d9MmAtL3jSVR>*Y*!
z#C%Kk`0@L!@K1a9R;YLsAsYJe>UP3Oy;=5!J!Af4W?*S>!YU0uJq3h19jgH0A`~b=
z2SgZ=$rA{|i8Ai4?g`BdVo{td7E5OS)p4sC&$x1Ad#ngOFrDo>zV7W1wdeB|U^A!!
ztb-@*g@EO*qcZV~7at8!9QsQO{-Bth)Fmr}?mbVu+#%vvD|huFI*<Rx&_<#o`$%IR
zfO`LVhkO{ohSCxs31F&_Bp>wM%Hm9@x8mkK%|<P`&n(4*GI9|Dl+k@%d|Z)o2h?A>
z`-mDPK)Kj>1hKP~3phDt{^^`H{?*Lkzu3*c7()DWLw|qm|8yz-*+&2M?fTC)`oFV{
z7`>VgCD+o}2Sv8^yK2zofp_$8Y*{~|{WZ5MuV-3@D1xG{o>Rn}lphoe&5Knh2Yn`4
z<{+$Q%#B+t6;WHO^_3NV+}Nf`*K!RRiWp(Gu_l*qTvyC1|MHKLi^LuU^b9vK={pD(
zbb9OXgfl)!+kT7B{ApM2V-EhL{x1RXi611qmygTNvC-yh-r5X_wU>$BH00q+STKmv
ztsmH_NC|lpGcR6QK`-#i%BU&SsEPXNMr6IKbRYaFF#>^$uvEeC1XXB|)aOxFy3&=Y
zZ98LI-_f7NQ=Ao=1gyfPqJ@5Qmf9Q9g-6HYyfro&MTc-fZD^hwUS9I~Jg07%zd4eM
zetkZ^fA_*0&PXG7cKlCY!;)5}iyJ4rQT5yl(@a{`g!q}LoEf|10KcgwsHZdZeZS^P
zt69gzI8BEnHM`-W@W-4_z9z&xJMplHoQ+8Q?mT`(#^|=LD##Nabl_5<F9|)LmCkuH
z&P}dMuCpNsV(r7$X>>Krf(gvrYP)$V)GyOR2UF&A{ZJ%v+j$c>IK2Zm*)wZ1GOO=v
zs#|G3X*F3T!zy=u4u2);5u{rIoK1Q{XiUWYLBYJP#PT_nrtQiGYa;_F_`*I`%TeH_
zO<cEQ@|tyAZ@LCQk14}x;CfRjQDxo{()v+07+Nn;>xf!u)tk?+kWboL^p==*O%i=t
zw&@<;Jyf9a^f^)D{KcEW&PS<fDTpo1@@%4L18eA+D|BVY9vw4UB0IbFP`Ry-k{J84
zY{(VC4nr6Ym!KHVvg0H3sfNrZ_fA&_%o3yg_k7i9l2;n($^AT6iNA5v91u}kHUHMp
zWUGhqqk-uoZ>#G|ja1)h_%l*WVeUU3cJA@q2_iGV!qk;mVfznI)|JF06fAHnBjFlf
zJI+lo^sNefGQrEa!rJ#WEvL>sB|6f5LL3hVP(y1_!A+&nAkKG=)<&Wfm%^hhBu)A#
z+nhwRqsOy6TYgy?cl`2n2~9aI^L8brAlwG`suCJPW0q=l4l#2RBuBT&Byio4Fe1W&
zv2<|X1pat|_o6-9$v4*BHu{ew36=XP@1uT2lgcG(G|%6ATxa?EJwzW&%%8(tf(3Rw
z!ji!0t0!BBvt{-l19Pd^WH}>8cIVycj}_=Cd*Hu;t292_hKJuHP^Wt1@^s3Tn}xxb
zCJLuObqG80$>#~Uqc54R2`lbTSp4gj9(Z)j1}a~Ln`FTLplF$p4LX4GTMRA)&ge2m
zPz!56d2G$~PFfQxYc7@X)XLD8g8#~!CaU5xEo8i-$B}vVvBk*_`sI|7bSxRg_6G%8
z8oJ%!<+Yu7`UuAK606{uiwWMJRv_M5mUpz7KRKcO?hn_%b$KH1o?n&#L;E;O@Ys=#
zwkRqSqG4a_D31i5#KYMjV7kHV@J007qVWBT_q`(mBKl}D?r%bxNh<ihc6T83t8&_j
z>5>`bqg!t1{wzzAL!x_Va+cnjuob>z`B|N9sNTx$I04(YZ=at{euOqdudX)=4%Dl=
z;rcdHy!wJM_&DthU1z6V`-HA6P5Wi7(g{i7URE{ZB)l~hOQ7`SiXvn@5S9QUe*o+b
z;rs0;V4#Q~uQ`bad!tRy1d7;I?oeI%qW;nJcGkWB!pHIaKQK)G7p_dMe+8GWwZTPz
zSt=MH6P!{4-1ODY#n|eS0OYO~i5GAvgD9U}jyv}szz|<)U{bp<BYvd5TTr|0p99MT
zum!i&heJk#CH(iz#%3RtH$CJjSp8W@xvF+|<Zo08HHi0{Y?o_ip-+vsjLO_^xKkas
z2+f{PIJq~r1Xsk<;DSC|N!WoI75FoL*@P~*Bfp#egpYl#cxlbgE5ZB>Z6FXGIyutF
zP1Luj!fqWX)7wo%G*(+aEHe5Cf0>hRX<*!qiwkiNcp<xT?h*IhmGR2H6WTREM0HLG
zAZKK5POO|!0bb1{ideuJU4PodZ(#bm=}r?OXy$hHpb!>*5tnM=%_9(`9din}G#6Pu
zxg;R?<=7#M{=(1e*ZQ?^5j<`qpOdx2{f*3%gCSPRj^=3=w-%24U(@Z`E-hqB{{0U0
z4jo;{eh(333Qb_%o%gX&dHvLH=oMgZG$2+inwz3Fv;=s`Yh4$N8DY;r$3l}X#YmJ^
zwRT1_w*8tj6bd9W5^0*MVoX<gXz<HH5CQ#1__bQ?hFYl)n|m(`Zd?F&MZI60WRv@e
zDk+B`_pb~Al7RENWJUEqop!amd1!J;3+57}N@Co>;`+b=o!q-b?U<SM<~x>N2CuuE
z4C4GR-?)r)z4FaaJ-~+=E{hiiqXQdxVMZ-I2q$c!W42utwtbledq<&yv;Rmx+MMlA
z`}otD&z_LcuWgC^IHW(BePU{YsN9^)8z5TzO;h0^Pbx$5v$@xj!SO1WAN)oYI`Xym
zl6OyLI%nu6r;!ddk%A^dV=_61l%f^yCHq|qFgBygtNg&Ue~t=C`WW?hr>b`P$zfFE
zM`=SteN`D6UDs&sRT!LM5QB?<awpoi>vH;Sp$8O~n(!sffN-46XxI(RKvaH&Kg`c(
zhC(Y)5_I<4A#Xc+TwnC<G5fsrltbzM-RA$?lD6-6ALs@smv?+tx<0XHgc$pHD|>$9
zdR7pb8Mzt!78A|=ZEMfRp5oHv{d&CerL__=3yk*_As!R91Lwq@_OXJtcHK52#Aq7B
z?3%Z}<q&+|--DsR|JgpX;!#d1G>H2fxCm~kX%!vwTWzjqk8`}b^xmZzbZH@_etm68
z{5_>1U*?iLha(C)LKk%D%t4LO*)ZW|AFC6#fC_JJN9LJrjkH@yi_BN8FN#d+@kLQT
z(R_QsL-{{xT5Wsu!Q}SZ7*8F*kel7vm>S6_YwRgT^sDOvebYr{;sNj9{o7aCuIN9#
z57QY;g@ylgpZxt3LTMP<c`5@~^*-iS2K>PVfG%d(JH-sJI*ddDEGK3JVV&atlB0n9
z8AUKf{#|R;U_!pE=i^7)mX{BEjS;0K9s-7u;fVLd7IqNV=?iN0e=oRM=%2~}rB><E
z6QjcaJ>lAnH_DKZHQW>DtR>Oyj%q73Z|0hqmx;OHDv|Na^6jU%Z{AC%E%kr(-YsV8
zK<}-UIk+@2>lnU-=NDUGh2B8Ejx4L%ggno<XKoPOHw~@z!n@b+kV^K{W0Me!yz`)~
zQG^&-Jq^aIODN)HkjhhdobR33c6q|IC2yUBF}W(*Y${5!$1Py)>jw^n>IllXU&e_?
zkkwl5k^XV(_tw=icr}K_^=8e<w@WEk1Uv=4ZT{@TmpoohV<s4)3rXUGB_xw{R3lmt
zl|NiDed~lX?{&S>n=b<BVz%p7tSK9t$Be%C5E|%O3azDP9U?xn4S`yRjDH`xNKs_G
zF>Cww<^DhtqSiU|Ab&F2^bZQjwy?y69Y?g6r@JU|Ve16>$yJ<jt=~+lb}n$xD>z0w
z;gD>xmb7S`^=hhcDPn>BiyOtrPC(<0k$M&U1Xlaw;CdgZDq?d-X#^rdZO?xUUu3>p
zs($^wGs$?)j~d)sxCYo51sF0}e(x!9*j1;02p6_G(o=a)xMq^`l;%i-j#-00_rqO9
zmPf-C7{8ARL32sVRRo6YtS{s1JU7Y{1Zx{;d3)lyWxuAWR0f-EO|e|pw@`FA(g~<?
zoJGx&=)7lahX&H)EWs&;D1be1+}O%n1|}tsHuIidttdGYh^*7Sg!tvs4nAblJ-bm+
zOn3Mg-xpTTc`5^=R3%gga{F{?`O#FM%|x^AaAe&}lQ2^YsgG&tmT;7J|1?9?C#$4M
z0Ob!ru6-ffnOjI2HFhi8+nK^d0h{{u4~nedvZs-+7#R^GG3x0nH=>Q*8(D@H1kI(X
zB2`)w=}qG1&LzNzy%Y~D6BKu~9oHqbAo=E>sI#tj-B2=2ap8+ze)gD&=iK$!S+ujs
zpt+f+<kPC!mW%Ah2}9zb?JBkDrK-Kv+m!uq7JjX{=?h~^$<0%-+m@yiMF{F9(Gvgt
z2R0GkljUizEJQym`w{V&isv)OSo-<U+yA4lOCSDaVW>$iZLD}fb@(re*)jzWhNaO@
zBo#)avtTwU*{vBg^+Z7YYTzx}PJa4h8me~}s^?ffnf~jq%i&<A2?Zfk+ma{QI(%cv
z)5nI?-omL+<6%406{{#CJAH%WAaX`*ke~1tW?gd4t)mohjbXpeiZ-EohbyhXB|T>U
zHT9tBYPmV2Yx&sJ=^Jr&X=hl#d~DC%_A;04JNAtz37tmM<Fsq1ueA$^YD{#*jz=cS
zRg#`KjTzC!LI-Est+4=_8|H2H!yX$OVkaQ*NCPv_i=%e$2<&(`p*IsyXQSi_MR_$M
zVGh7?CAA3_sU_LVpYk1?R}T}NC!fP7bCH{C*R@mHt!_B+I}KsveH?ZQ7lHa8>^98k
zFub44W-@735?Ik5)WN#h6s}}uNU42!dy$nTAcclTNa3(hHbN{Y!d4ld=+Lroo;dz4
z6Yp_YG^077G?ia>D(?~ZT1M{c^fAbP<tT|0qS`t2Hto#EISO}cXR&T|&oOIpe|Np{
z4uE7%=EG+ljfRs~!lX(VX6&kJ<44DHu~0<$rhb?EUhZyIqRp$z7y${cDw?kFMc?;I
zf^aK<TK~W?i=dXNh@9Z7u8x#9S6V*Qy>)>}q3Da+W@&p-cB-#lwbichQWT3Ru|*G{
zi-THByU{Z;bwm`(cx@1tqE}P>qe6qlIP#i?$F&GO$>!UlpAF>=+;@Sji=ut#b)wPp
z5hDYF_Ds1Y{;2)ELn(w~YrMu^?~`La(TUH5RU||(_I4hh)oxWuLiB}zBIw@nhdm!o
z<yUTZA(Lvc8@WBt^nm5rYF|GcBzQ-3K8<zuY92Qym9@q%h0|B@1WYTGT`OlAHtfK2
z4=csXy@r*s#chAGF3@&LPJ5SOeSf^L09-!JZW(Q*DRVcfTJApS2{Ca23=k%<?-uK^
zjL~!6kBAEfR@DjC)dSn`;GTTt-GR$p1rODGXG@;|DdEM2;m02l;7}t+*K*L<V$RZo
zJEEYq-7}D;2ga(e{B|+hoOe|@Q*#NE_B{J`?x=YB!-$nhA^WsDPTV>70-eHmla>m?
zFTA3lzqFkfX9fzJzM33UI|g9ue@9gLUvcICd;IvnEqIEBakiuwfJ^-su1X)kRS|Tr
z{D<{RI%B9#V1AF0z^cQy;d6!1zmc=wz#kK^iTI(W31o;{LgFUt?du&*Da*-fucKQy
z7M-G}qpy9C=)T|t@Lf9U!P`+w7LZd_BEK<Vv;{2Ga!8nkd^9Y~g;XU-JmX`Fvh}>y
z7u`^;e<L{E=I4b#4R$G{)@pIdVoBk#s2ZkZ_c)da>flj`(MK$RFRc17;YY^>YQ07X
z$`i%!{Ct0!w>oBqDu`|MaP1b9&9Q>NLCoN`M1(EwOvN18VZq`5bf}oz<dv^kIEku6
z_3>86JUM--o8+T*gVYiU>4RGnOlRjY%?OUwWbpt62(-u?=0K`UL*XDBymwWcI&ba?
zhFB&)*}N*kbW4|Cc+ZHAApi4ZH0VIwFS#MRBTPXKr_8j9mSJO<YhY%0cBEaO!0CJy
zWiOcB6?cb=z!u58YQ=uHcscB}N0GUzv3`~mQL#SST@}`d2@*e^wdg*fF(Wq2f$520
z9fDZt0+<nSaU8Dl4l`~Xc^%lh@gYI?uFd<X_zQu9AI?el5l0WBi7tSC#Tf5|M8V26
zDwRY8*3jblAOTUfxLuK@<i}r;cB*bubhV5J^Ub6BM6YzBSO>tWRU9G-6-+@M#vg?$
z<LE&<ms>nUn}0<MX1S}1WTd8fvS*dIUwd`zG+v<0uy5}16T#?n0Sr-MCy40T`3Pep
zzrT|TtZ66<bsNt7w1ln*CuMbAp1$Py=mGyle^Wtku-hLL3>%2%35Vd=vY~PSnmQ6!
zHfPE$Q4)g5>k?AYpT+7;zE8rc))an8u~*)HNjA7qwdh-Qc->HrG6VUni&#Ps0|&V$
zQ?YWqz*g>DUJuD!b9XU2OPDU^kQ40Q+jQZ2`@l+|K%SZ6)mz9in2{unfigoHuM(<i
z2uh+)jyjkj_V$@Xhe2<i1MS@@lW7JG``)+43&sYY6E*of=|j&Szdqw8jMNJ`&#hL>
za>Brd1hr3jQ^5jJ>I2)AcO0!~cLh#G^>Tv?Z_r6uevrS}dB6^(?P|bim6)3ztX7#i
z%BfcZ9xck+$XufuR3@zD)u6(id<gFql$7`dv9!%0GtdTro;fz>?<LP0J9g41P)X1`
zu0N(OvO`=YvCShGfeO&$Oc>_zNCj^N3VE0vnM;Lj^m>@lwp@5^#$@2hia>3+Rp@bG
zpEe=2J>t8=Xst(BJ>W&r!Xt04);ZKF!`xM3X0K(n8uQ!nu4s4j#EGuNeLl`04TI)i
zO>|&lK*l=eGVi>4WW6NL(b?5Lwnc(YIQ5nnw~9fE_Hgq0y4{wL#<ED&62EpQhYHUK
z)deAKv676T#neiKmu?2cdrM~ROxd$gT>EtL`#$;m0t1akb)DZSXs9T+T*Im+a(?^4
z#MPwk_vL;*di;fVV~>(F*!gO<R2Fhu9Qj*hc=#KM0XPs4+{j$d`mwUT(vds6J6OXt
z^5rRaT;ch10*-QP;0GkN%3-oN1bMAw6V5)blD=*V{=8W<`D#GSkyp>y4U||jNtIWY
z7Xe;o3?1VnltxbukPKV^SA)L8&_;p{2G`?af}cJLK<%gU>^<(^c7txN@mD8pqb4|w
z7h~=}@UmglxAaCr?%hpqsT}~o3;NlGj6W!3t=6)4Kp+rd2{Y|Z21~C`9JN8Ih(<35
z_5s(AhtyGsqAhZFC1!Ze546>K_xRE#eUeDJBaL7X3IfY)P2gd|FXt9J#CjujhER*O
z)v@e1uV`^Gu0+zfw81|uK+jodF6@LRDBZyj6SG!%Bm)PePQ__qhC})lbwlLVFW#Y=
zjK~kzGW8oN!UEB13UliptlsSNb~v(bDN${>S6wk%k6AmcIufO{(cGR#C&*ca2e5m%
zMkG>LhrM>o75*S+F^g?9ih5~qJZy*=tgPJ_A|3V)D@nn=S1M9APpE&|+JPhT>~oD9
z2X~}np>HxL*ltI-xc`v%jQp9=!6lxcDR(KQ7vt9`sh+1j&n?`ZAk#=o5DMxR%`@%E
zD>`(f>h6NNT<8+<v3hyUsw?XqH~X`wI5cm*!z{j}QOX?tEJ|ktt~rNfhxs)BTIb;!
z${AiPOBml4449aiwy+Jgs(lb6^%`}3y5fQK5|T$MWrqhQyFk+lDPdS=onRmbWD`BK
z2HhYo_8$3d45wNmeN=T87ov+5Doy-kE|=f+%Zh>LS7DOLh7E~x3AGhdRpWMth%Tu7
z#Lkk(xLv~OnE_5Wn9M?0!sa^SO0L>9@>LHm?TX_D-DsWC2A}^Jza0^x*7>95H7R$n
zzz)e)e{tSslueyb669vsAD?2#`ml(F8yGFx7a+*n3{k1XcI>|NUSF#y3y3(0+HiqO
z<H?mcb^zcCnkWex)dN`N5o|qmHFiF~?V0;H*o@&JDcvtD0W)x*W1-W;D<}!zEy@+M
zN87ZuS~-+`E46P8sS?CLZ&lPm1e0lp$JOxm*!iWSr=JYjZ|P23@w}?g(Yf~llr>#9
znX*^r6Jxs7Q}j0bc3Z=AgAzCSXfJZK7W@K4dW#p_1gaW+?WPPT#1GUpSk^ARK#!pX
z0~^tYnJm|v?;rb~=F}+D!*z(U^OGP&ry*Q?05V`_!5&4OJ+H!8vJ_`Y*JW3Qy85QK
z*=UP?H7cgADY8Cl-u9=XN;@4+nbwX&uD__Hk#&QID`x9mvOGl2%-B1bJ;`XX<sz+q
zAM<M-$hZ;@2@1Y~uv-0y^oPGmhvf&2@4Ry&DSI5sn_Rxa&+th7P?P^y(@-<ywGpNB
zgv>XQskZ>&5eML}RHPErc2c-COideFxH~8+-gYwXSm2s<mVwG8vf;qNtQ^AAv&2Yq
zrOv*`UP!UWUM>$GVYhCz__HL*R?Sd}IlqL%&c5FuUMYl0OEX;Z)>6ef*=XzP`@P4E
zvku%%sD&NI3OJa1%=rbgA74^^&%=Oc#n_*!XyNWu2zNn)$ebhM_4c_Afy`Bjva-J`
zMdNEF+oVq70uZ0CGdY_(HQEr92=ch4%`{f5V+F*dUL|L?qtJd?dW`=vjZ<c4iv9gk
z{*ON%tX%t+;;bJ!Ju_n}W>hz`)`*4F80sYoe-stRo@`7Sk56XSdPhBu?ZEj6o5Doj
z`J>u{#LC25rx;>8*b>X3a(l8))05X%q_xOw9W5bJv&`U0*iMTPmdyF&q^1%pnsrKj
z(M&g}^=VmI1aa{SSV9+%p0`Ga)O+JkwX23%6~%Ge*_kk{w&z7Zw5j%<zS~Nh!;8JS
zL}1o=*1?5rb;iOu@s2q)1@`FZb!C=fw^WwqBUAIn#Q9meiPb~X56^xSnu?U#A9l$x
zFgsq+60z(Q{dVn^z-;Mq;9>0EjQh~<+d1_|J9^-TrL<A4E;}$-W^puUx*@Rz%&j23
z(C5#_dxdnz16pkFY#-ymYa6(J3%01CfUQc(v^8Eb3$XruSNCH~%R=I3Q4&o$F{PD!
zb3lQYCp&Axp^p(&&_aujbbDFjdCA(N>gO0m^hTQ})R&~|>3Mp8HSHpp7FNapAzub!
zeaNHV%|Pmo9*(wGb)A+CUiR*YZ+B8_D%4ymois}fIDqg1IrP`aY_qVa=H!uSED~Qf
z*x!@8>cJ5)lX5>n+7$xh_Z)S{zU=&JUj0_qa7511;(geWQxVw!g!bA6gh(r&%^Kna
z1yQM^y~(3PNx)F<&6Wq(Ee~p&xiIKToEs@pFkl#IYs`0a<q(X#ZD`;+K^txMpoyMZ
zZbdq14l%*f^6;qnOt?Mm)^G)q3It9mQZ00dRtx=MrXVt7m2!^vZs|hAoEK7rs63~L
zP#saaVs1usKUFJ}B?}F=e<LtUd4Ya&K(y?umfpVMEZFS5_DS!X0k6L@<AL$(GvuTZ
zetQ3<orjV@b;?%4%R8kp>TyqZY7p)wW?Mq+3~jc;?b_l*M-1Q4Xzl3A%AS({5cxig
z&JgSB&DYUR=4u$JOo3L}C&-*s*2m614pe$~C1{%O4cGPywJSAq!viEk{PQ#!D78?y
z6b->I933>O@4HU9xF?66UiR)w+1_8thVNvPL~91f*MmjiLW-$JBr$62(-&*qF|-J@
z?#YXDA6PcW&gK@ncB%l3ksV;44Yiv-m&GcH1fFUU!|Z`Un)WajC*`kNHMB5H5%t3B
zSQHo@U>Y6}ujnYLzz_pL>1^rVxvDy?K~m%&6k<@M=ze2l_S)XTlJ4?QGRafPtjk;c
zfVGkk_S1rsz?H^3LW1IY#KDog6-x15g`gJKQpf4~*52!4{W{L!E{65m1o@^JI;>O!
zf2)#6|30^o!Ruu2$lQDtsNw?aB`S?g4wV|;X8+%m4e(C{H3U(OOhGch>xL*?L0P~G
zc3RQ3c@5^wHbdK3M^&532DV9`zS;+-SFa>dr8i%Bk(#~*SI08~vSXG;absAWF)<#J
zLG0RGv*h)M--i^Mtg|XxYA|HQ${;@+4hm%W2*zCh`Sszqr0Y-JmiE9t6XJ?s-LidM
zRh`b4x>kU#Q^}fD+L@8IZe9hLqw*-C5<@pT7NPv2EL48d&?G#>%VsI=SylmVWFUTF
z#mQ1W_=}OAg}$Yg&{?`v-n*sbd2t4!4oK@XL}$gewi+$7z7@_kX)F2dgY0=VZkP2D
z(+@cS<BEE8J_e!GDm<OLM9cyhX?s|l>eZLG3$xa-4-XS<^Vbp+P9jshE}2ctm_93O
zQPBAw^VCg1qU?(br5`Qd4<%v5Z&5<j*{2faeU;GwM3W-IR-A*}qkL{_A4p+}S<o9U
zv620aBk*b2rw1*e=D(abNec12H)FU9FFjgHoc8&T8|bIG|L^FlQ5?Akv#b5iCw_gH
z;!oyS=Q{uDaA72hf-U?E%b!y9&g<e{6nm4{73wA*_vP`OrY3Gk{@=pS{FENRpAJ`u
zM<|I7A#fu268{P>X}^W7H2w;E`0=(FG_lL*)a=r|sw7IPZ>3vC?42p~+)2!zZBT-)
z-u{65OGit|37xr*(JMqV6u;6~_@QMWVbz~#)Q+I9cg}~Uv?liiU>ksvEX!K|8rJCW
zV{Q=2^t#nvPJQT8fvM-55K~4KMy>M?ie(S8&eAKPzK(OJVRXWrU7|XKg+?~#x`)jv
z7k4{reaCFpY1i5y2L+>=J?mRF$;|^E6>L*c15&`s@9T(xo4bLYZ$|G63LqEQ9)Wug
z6DRwN186<4Kqp8amjkx>>2>(o4si-1Z%e6tZ9rhnp}lx<u91iF$`@KvMlprF#8lvE
zm-$Q>VY21Qs?1SxP{`|%?0QU1S%tZfTAij1EUGu;ox-?T`e`rq49Vt4!^kXhZV-`#
z28jzD<p5kgZ-Bx`=ih|bXGf5z5&?7R0#ih5$@w+3X=<zKRQE~1&zd;JqH6k^ks{*1
zhu%H&>=QVnX^(-s;iY`84k@Tf4~Qk^Pn|_}H$fE2`;;{9i(CqPIp-fg89h_WOG6xe
zN;nLyt@`AX91x~4z-C}&6|KY0eVtm3d&HN2Fyhh4ESZf4u$^C(04j|afpZ7KPsNx3
zoBDohpd!Z%q?$HbW_&{GAe!g!;4^CQ9~9>V|26tFplJ?sLXRPz*8&1<N4s5tyg($r
zS^+3$KL$?Yr{=$;&HjCh?r~o|x4-hI?MD&Kc<6vKNf%Hg69A%e;T@_J-g19xl*~UL
zCHVfpBi?@=)j!+nFHh<}@2mgApTRL^%Wk;q?~EATgNTESFDw8+S*Q%;EWnlR@M&na
z_?6=X$&&b`<CHy4`ggla(@0&h=lf%`u4|zG<_~=#112o6j9w;L*M}jyN895$W>xOT
zHSPTJm|>;+-<B0@d!jtzeEQQa)DlL6>u)Dul{g7SA6?lJ*#K3sw)Y}vJ$UA&D0WuN
zz}<K0QOvF9eMevBlAN$)Ovy(2XwufOQ4cRYng8+JVn0bTeHBgSfgrgW{R?4G^Hedl
zP+si>ys8zx=JC04mh`R*%(Ve#7oMAttQQb~w+;vVit~4E+?kMZWn%kBjoRUND>djy
z?5JLO0Qi??bXIhNRNs!?hwf>=)T^G!S_K&Lv+R?SZjmid`!nFeM%!YgjJnx)E8MWj
zu4i5r_?cLbyw#7gVaOM0z0{KIu7zQ+R7ZigrnhtXU4*HXMLdOUv>VtC(?^RcQW5$i
z^~(7UCqAh-FU;l2DWA(D``2$IUobpnk2$#1X=Nm6_5AhI)?9v;!Hr6?#Svh91xU&#
z8a{Ncfeg3yYPI1h;6dBLIHPVrdLa!CuSGO1czMr@Us@YI;wbjTo<s&%d$LFM4&R6b
z*@jOo^+TU`3WBBSgWpEo+*q?~DY6r=w!bGUQQWWlfm(-V-l4`+yD@Us4^groKJ@aL
zmhBk=9GRK--@w`bg1!F_V(prG|AwL)gD=C~vH^1r_uP4N>+I7Dq15*$bB=SOzY4r>
zui?z{uu?S(FF&VQ#~@tC&2Vjnhn3I~B@j1D9Bul@>rawC)D%lQfa@L>rrb!s$L7pw
zHy=e;=c2AqP*i;(afM@)mz~TzPDH#IwIu+Iy8)>U%nyt8mkx&;ju4lQZGkV1!?DZY
z=wRi03(1JlRngw|(8Ytku$c?^zJUE5<_#Wc{B}I9N6U7sGHlC`_T4u6j@53CE$#6O
zV@IJIEC2-|0%6Y^q|9^>MZI=18YUH{Ki{mA^5G=D?8$(qE?=)n7JJCyXCw2Ub?G}g
zy@wm<z8@voC(vC*iO}W$E{gn%*hFA5AbMok#n>earZp5oINBPc9U*xR;+E&@zegU$
z-%Rr5d`R2tqVHWl0hZZ1j-UMvUIYLO{`!pi+$Ic9(Y7*9sFbj23ifrwFwBT%pY#+7
zU>SKRU9NsH`}HMJpVsE)f*M$K_Z}rrkSP&|Uuk*2TJAtr#y7^ZqAyL^f3Vn)=|_>Y
zBg!6`bR_3~m@9rfrD(r{KJZqpD%!FuhPGi?6mHI4UyN-f3lAvBaXA<DPE=(w05O$^
z!m<4HcMx$rp9$Dh6iFSA420id3YA7&wW~D>%rTx3wK;te$SG@be<?oc>rIB}3jbcU
zJ7BGS#;S*Rh2ut)L}8pN{UVjq>g_(fKL*6pPYrUi-3lIaSGLMMURLLSVEFv^^{I=z
z4^Kju;Y_DGa0z0-JcEns0LeN-E8pHUX3g>3h^W?7XBWF+fZ-E!I?W}2U#-kyiM!}T
zaWK*P<h!iho?_I~feLN;mfClwKCy2gImS<oJ0P}~T9{eic0b9Zu>Km`ArL!L2rF!^
zHHxU<Ol32zJatO?@S=<6D`9(VePiz?)th)-no1Kic7LPc<=%5YGS=}~4}Z*ZuUvf2
zC#&hD@^rZulOU={_iwHx68Mg`K33SNMPp)%62p*KJ#X%YzFnee!uJxzs&tEs^(jej
z%RxqJs_#!#`Z(`BaStyx+(|qkS8@)xxRQ~pA4jfrIsjP(Oy7ykL(Lx(!|RoEi8*_(
ztNOPqHTW7+Ja5W8>g7gXcK+1+00k;xvon-A=|1AXrb4O1dh<oldIe)Xn_8#4I(%o7
z!lbcdW1b)>XaQr@Iz-QmguJ3~hbaVAFDDdI0VAonZgFvFfm_9v0Q=C+Qm2_#wa&$J
z%|D!ir1e`cto01@r#%)!PYB4MgE|x!Oa9|rhc9M|U#R?7?#kVtD`ZOHbv=1WS(s6?
zLq$>p%tld4UchP^hEyXMgfkyAPunGbHi}lib#APO^B#jABhv%ES`L)q93X7Hap^Z8
z5HWfNy!{|1Zej@XGT`x9hdo}hBB@W(W@_uSL$2MMU^=WfqP&4#N_RycO#Mx-GY4&?
zChEQd%I0A&%pv;qkz3FhAk~32(<N76{X>jscx;OG?Vl2F#iQd*-v7KZtEb8-X7Iyl
z`{=IyiDF>PsH3Gd)Mhj`B_#nJVk!Cj5w-fy<F5gF)q4Rv@lJn@%KisoJap}0Wn@G4
z`1(idrN+#hndU?pbV5wgk3N<>t7vXb`l77MKj-{ti<m-ZceUV_f1ywIon78P*X{GS
z#pB45^1n&z|3TaPq4aM8AA(RD6E!iB%%&SkQo(H@o(=g-<iAm@%FWX9BxfwIjf})B
zbJ=c*(B3#3=<!DXRc#LzBMy<(MZL(FZRLH-7c|K*m1d6R#Ca41JbHjqHmDyKDcaPN
zn!dl88Tx={sp;m**Yf){PIDBQp^J?VNlb|F+9SW7jxnqL{hH#)OsH!&kH%8a+vC@F
zIQ?HQ#<!oNk=_{Tgw=I)G@oGOvnK60N`j@ipWRsrWw?(@eEY<bv3sb8`5^w@IU&^Z
zjBEflD0FxQ0`We8F&)YPet!tu)%r$kMUR%r5L^?QH$AN>)$P*+y&id;-zG{h&UQ)u
z)p?2s*M`@W%X5P(fbRCAK~GSi+uD{}*Oqqu)|10dyBM2kPSA97%KUrh>UXbO!ozJE
z(vCV})_B<JM%{E798SzDQ&W4kx+N7eyO^8ns*qytFMB17?D?1ELkVfgSez3oBlX-x
zKALm2bKxU(4d4SbFlL8L^qmX><j}-L_#V?wFt183k<O@=Aeb1eV2*QqxxRIC#L33X
zqw3p*dz>*5!V(uat*rWl_&-_9&K(eL47?)d?^lgXsM%<93$hY&d|ol!oLXc=1@!u?
z7M=0V+eD#&jsWipw(~a`WR??PX=jxiiRP-%j;(v0aBu27UCCPp2bgW^VWV()ZB0d_
z-9lAE%@1b@-{)VR4ai*=S`rE3CWL<)48V^&Viqwa%3q~tbC4-#AaZmG0FV4xPWJbo
z8!&T-pb8YI*~0~FpWqRL+DMO?r;7GF21Lb?NkV%iD!|hR#Qg4`8&bwIDwA*dDrsjl
zmi_6=DB=?mV3zZUTM{7WE&OE#OSw29ydnM%U;pWU|J>L=+w5<5%>Qe*#`#U;2Iqz<
z^M4Qz<vR%f#X00<|2ebSFf#Wlk6d!|f?>tinYAVI?>yt_9%~OLu5)}@_=g$=b7D&j
zffF?}l-z<>&o#}$8SpmQxUc(nDkh^L+((BQ=BoMA>OSK4jJ9|kO;kHlD&oYKoo>6Y
zK(9qPpWcJZ#f(1lW8i-=q+FKMkk;uaWIiln`)Nne#3{6k3J`kMyHWX#aeeaDD0!6;
zwXuIjc|!VwqUUULS4-C~+L)w)->mwm69x2hkMZR1*-e&QQ39j-7Fv|g1w%3?%imEY
zO2}&rmcG8H<pcxGDf=n#3T}+8Nltqg^i*J@jE0>mZ!eTv?tz{rO8olK{9f|vn-^Pp
zQE`@-jL3uo5(rVqG<0&tbg%R**@wq|hI>Rxqr$7BZ20BGovdmsG^a*mYqlF5$r<#u
z>-^J&4+1&`L4uD2MzbfItMKZ~Y<D<5Q>%-OQF4_E#BS?oYpTkZnuz*^{@2Qfzw~|o
zQ|gBj>9Y*dl_J;%t|S(B3InoYg$;%(C$-oGuh4ar*N3kBnwh%u<>?E_*XLUV=#%#g
z|7r?R@JzZE;yPrJKkdW6UAK0$+DA`FY=a;GW$r~1B*PA1hM*ct-Ef7w*pKLdk|KCM
zL*4D5?gYU|d%7Plekdd<gqQS~iMQ@hZdmXRH(qi$Oym*4uY4vcXH1Qm$@=YUNDEKP
z>>4|^zO2kjNECaK<<e^LK<oP#tAe`Zld4cHO*zqN45Rh27QWhIO$ZqGq5$BM|AbeN
z&RiF_?LSw?%QOQ!(#5A0Iv1@Q&D;ha-22da6^1D^9pBT0=<%y%q3)iGHZgtI4fL1l
zZ1PC518%eDNOHt$-bFwFS}72WSYrAx;Kv}x@?qzl(U*}l2{VI|yRjlFu6LFw=E83=
zcn7JPx*L*<YP~1CC!nyHWs=X4x6ct`m{{I|VSrAyChr6&I2OZo!(KEHT9YSC?Z_Mz
zNc9GNIG{0CdCB4OHQ@pF+8e<^jWiWL-$rAlty2w{kr5EmHRmPDvGQtY+mc}YdDp>I
zvED`3nj=J3JLGM`J!&<r%-em^8x^lW(|%+l0BZG}m_2!RCSq3Ug6Pf$wD1;yzxwgi
z@k%@L{MP=RdfI4!*d#yM#|?r!+%74+U6b}Up*%zDcBoY7?S;upcQXVa0Sh9UCAZR*
zxZoPZS_Saj7NWIqel6Ym){Z8gbdn{D6}IN0)Q!ePU>RaB_KMLvBP$Z=U2cLi^YWMK
zt!XMsj-;>Us^Sab*p}rHz}0JRa`X?1^Q5#@;!YDG$`5l*iCtwun^GTp;LU;y=Xf^}
z7?!PDq;2^iRWDB6DdxEW5DavM_SokwB|Q#AdPqP7iUNj<2?V$`;9vIGWnMV@2R<Q+
zx-id<n!M9jW+F0!T@7F*TZt~Vn%+uuGN6k+1e5q`jYxt)tr!5a4jC^a_nsO76kx*6
zvS*6S3Uo%US8^v5i;~OQ5+Ybwr0oNRnLBXgf}&MI?@X<#`cnB&$sD`|e}r=^PF#rk
zJX-6G45{?yqalbDoCJkRRaR8{EiW5?Rru0DbqBh9KR%k?x$63oitWWh7+4=Gs-y_}
zs3@uTgTQvU=_A}P?o?zG+3tCxqBLfFa0eQL7!(bafX-07E0!zC-JtB?p&nG?hxs@%
znSp-0{{FT1y?j_qB)rh?lKf*nNwuF}erX!g(9`C@2kHVNu@<H|$O30DURK*=5@`<I
zXc*00Mj%U&<Zhdl*0hVmN;lS7OKcGfg-4^T7j-N+B|;ctcy;&fGCs&lMOip5$f1OY
zsgNSL1b!zHE=S;lUVaQIY&!Et*nVHfmQHzkNLq9WfZxP<f72jL@gJ{TE&yimMqs@O
z)5p0JEf5I3JGMWg<LbTiR}mCH4JdsnC@Ailj#Qtj-bP(3s*f*1x>?&!eMa)X^nmUm
zu&Gf|2ioE<Dv4SIBb<Bw+_cwA!cnNSxdlk^2{)r>ku-H}>O?zXXHXJK>nC|WS}H^U
zVSUah_(*brkLcfG!F`rde@}sb9&u$O?f%f<f@-Y#z)DS31C3@N)@+LJYeqEBjN$!X
zotu|vN5)5V$6Q~v?Ve$eg+W8s0RIhS*V7K)s6)tXlg;Zqlpt6ic48tZ=rBtAe5a#u
zYvYX}$0Aqnmv-GeE(~`r(=Oehzep!G=VRFSM3`;<DH30au>}HR=fKpk@)o0H>EZ;j
z{g8SYEK7E{?7fC|q{FDQcN&hAXa3}S0^0;7bFr+qx$3jp$*vtEbeECavH9$SiTmMR
zK2;cUOwoxq7T@Rl@mj!2go5V$SmjQz#f{<OB2lT0FbpTrUGa(uzh|0M<jprWgA(k#
zVg_*+DHx7?j~?DGfMzh8gAayw0bltq&Ho#^IMY|eTpFf4F#=(O^gTSGiB!;Z&RaJs
z=Ul`SqXjGaRd%rijdLR1t@U4_!k$XTr@LL-nE(GR)8KUry2o6VJAgA_HDERo0oGIh
zSEGERu0|;Ux9kFXqmt98v+Kj>!hm}bsE#tR!FBPTHBkoI4`+3T>nBkXhs%9k?hy0P
zp+~DN?~R%hnkHZm7O-%|+>S!XiFJ5ef^bUSclRfh7i=k_(}>cWv2ZrLcVj*zrwe7S
znm4U3fIWn|(I#7k=S4?fvi3`S{^g!*@yF1Ixo7X%Q=@_!@4;kt<zbr@VWPS((LsO_
zae;jOQCX6e7D7xWzN=3D@}$arA$!W~4`~naqan)#v5J_?ti&%9VnH5E0Z!8Lg|Ba@
z@^NMXv}<(}|I$5oKb@`C_sXh3K*%GQ7a1X~e0mCq3^_B&S80MUD|`BYEd=lc5o+xx
zW=9SGpm+hD)&iJ|oQP=r=FGW~18=YPlZcrAkG=PfYHHorMp0B$nsg}wN>i#zlNwwI
zh=_<v4G<OS2BZfmvCyl4fKru?lz@~-OOVhJ=|pNs5TpeXN(d0*oonyA_BqRS?m7G1
zF~0AfZ`}O{Fvc7&bH4MPZ~2wy_dFTw99GK|?4JC4PcCsq^(6|qfs1y#J0gyp>1FWc
zrI5N6#5II<SR6~^5|;@mb5r%glsj8I{o-dLF$f!76&34~FR1$xVs`QgJfMFOOFPfb
z4)0$8s?6Yk$3Eu+Z+%AEq0wF8I}{LI73dNp3AwAVyc0AbF-c25FzFXL6r`TfW=iGl
zZ4VJn{mF2v=4IOniMsbiU8&@f=9ty4fiB3y)7VQLU%s8Zcx%7~B}j!=EC~ydzD%&C
zc^AO)Zj&6}d%)r1^2y3;y&l!fQG!{^lN1lN+oxF=^g<&4+yR4M`!`cku|o?$?<%y2
z$(v`@0Y#!Y$AY*9w`Dx|BsyPD1m~-RuLihz;5Fg|_@(5-ZYG&L0f^&L{D1+ny6)e*
zQNMEoh*MK<3j=I@IejK{dM6;k0$7Y;hw$sWw2_|->#J~@`-eLV^rkhZ(;Gk<UaaLj
z0{&}*gkSEkd_(ha^gnUqXaAF%u|E^pkeq>I_kon|r<Gxueoa+;ug9@()G)3#i>La@
zMK*6G(jAe<BjpH*-~C}-42QkE7?`(}`<HOSblw$#u-dSQ&K%Ymob8Ru&rlD9%Vi_k
zSe<)%%&8BrGq4#d$=RmjKQ?N`M|aRaqW~h5gx_bpADZU)u_kaugahK5YHg7qa<^${
zVPE;6Idkkzx0Af$d`~IIr8>2VZ%T)-rk1#Gs<^F7%kZ5wz*2}EJa`%WlR;M)n7tuH
zIJJ-A@!-E&{A)b^zn&kq$p`kq&bfcfS3Z+M0{EejS~<PnWeQ!b8U6_2{&)XI&mCCq
zjZQNi=ye|W0(=AJC4bRY7&w;o2QjKAgc<ZbjS8Z-jKTNL%GCa~bcDg@j*s#!*Lo@4
zdkgPa|KRNV5IZAWn{JU<fMQv-3>}d&?SIGqOe@R|c!>lUdRnCClSzWGIDnN6;aO2f
zgqF>0T0LuJN_O5!!Z6>A(ImIzDD`t2i^rV)N_ckv$1w*99zUh}ti(}xtD+M?_EpPH
z<F|p6dq8;p>YohR6A^Tl3J`TZ=%6C#Mn-E6gKs~e$VmMhOqRK!s_vgy((6)W)BTwB
zyP>y0c}m%Xw|U7fad|F05vQdXQMCv8N>^lqY^5TV6i&hPcQA%f<*#o)fc4}?zgWoL
z86}QQScMi^@#NiOML#7<7x9?rV;>mIj0#5p0s6ii8f64Ta_|JIU4D1*9Q)~kt|HXS
zfGskYGQSr=R4t;qW9-t}wNH{QLOp7-B#r&|Rqq1sZDUo`s8lK+vLRJ$p3iu3Naoq{
zZ?E*A%eWem&>9FoECJ(xd73$k3Denz^>;h9uQ0A}l-mV9rU#Mh$$a1#`y-pnReHUj
z)Cl*@nd%e?)H|@lEugmi(4^RLKo<-OL?n+{h|s(esQQlq@MGQkpA0W$66k7nfn;Q7
z!+J!>i3DM4kc|tm;2^;UJ^bT^nnc<y3qu9b2b*D@?hb0YuevUNThZ-=b1Gk_1Fz=f
z((sxR!Z#2k5kQdh10TPi8epyav0H5OUYMw^zv$OPm)@z_L`xoI?bZOs#!>jKgoS2#
z16+0dS6qzIHsq<60i7cQv`Z|8&-MUjSiI>kqbKuK2^W8Aa*7&=*>evg^q+C%bm0|9
zeP4W)HQLb0!|R5wrw>b|jN+U5hjmpb&AMG~D3dKp-T4AuEl3BdtWae@fKc^rhTIEr
znn2g+=b#12^SecU-i=Ng-Wy}_tVYS@MC#<GpM*wdMqy5L15H_j{g<IJpoyu#Hz#ic
zM8GM)aV}J$U_+f02@iVUlJ7dhMpntmvvW0d;iX?lmAPA_HIk{{Xg)jjp(a#c-1#uC
z*urO)04)(h@WB_=rLvXu^6H$(kFbyqW9f;3oC}txiSwKGV%?&btBW_EgqSlWW<r!+
zTR=9WkqA^)V^GEXS|QG4BN-^qgY+*1GX2t%yg!efMYc^)P<<u!Z7x-MACHJBtJseE
zVPt%DVdznv^MlT}KD}o)Z@zlgy{UA6;hJsZ8#x7Ft~!^CwOKF3xMvGvS7{)GjiE~}
z?8AbLXW_E9v_@F0Y@a%lWx@FkjKCF>l=!Wwu@cbA9_{l|;FcEipP&AhFPBEuB|n{<
zlaHPlOT$ll-xxf@QLfBtB{yf(>d(ZHdvk*|w!qNWOnja>-mUqa`Zs`*<4+6a^)IV>
zH22wM^q*I@$S<?H*7+R^;=lNO<<{pJW&Vq<n@2PJ|DrcGxVApFiGE<NEW{gi_zohT
zNHnMGEieg{sxgsy*jv!?*{2S2Q9?^I$|UnV3Q>Ax#|v|w+0)?)Lvz#OiE-&Qetyvx
zWd(^T@xSVtEr1+xyOMRF8{4iMIpavWVR^z(v-Mvz$<!==GK}cb34jptu75|)hMF60
z8z8;@$&e?2%KZb>nHJ;%1kqBP0dsjrWbaZj0|`Qlw`M<p12~Aqc7z2vf|mIMqp;5h
zY%vFdh1evjUOt4^xsco9wOy;JZFGspr6(sCW}Q&%L(TmiKuy|Ke^iVk>LN8@cAY2Y
z^1?T_=N_7I-=?ekYN7QLe^bH@%YWSnAsdI}**VFO+1u>S0c@VjA$*nOGBo=ZQ4r&~
z((CWx<|c#d+Ps;tAQ+i{`1rN+Hu@ahSs}is`^9o;x5@DKm+#B6rQzWkLCp4q@${jg
z36t7>wO2O;iKe|V&O!OIe2;R_HVxz)lM!vsMfJM4yd^-%rHLaTf3dvrX62`7)fa7h
zpZRMi`F}F_oJ13Fhpnl*k1g8hM@Tu5?V%{&ccX7?gR^gG!)U#Ct3C<E3QG$GW4Y;Y
zDla;Ecpv53>mTSxQ}fsvUpXMwVB4BB$n2dWe42&@Z#Chn7%(q2pa7P;@U^40<2U)D
zz(m=wwcJ-_C&g|em;^(7G?0NsK>iR1)VmtY;<#K>$&;?usAOoMC-)7rBT4#Y?N`5F
zkP`L&ER<c_L47%EhKxmBu%Tkw=-i~Ds62Js>>QV(<t}->{8z@F3ORz^?%(VwaZTsN
zobG0_-{iA9%gAUn44_*pU+2&gz-m&#00;cX1A%>%q4Iiw6@2LunvjivU|pj-_AMge
zi+qK^31LDv$RpaF1t6JrZ-ImB0%%wGons$Q90hR-fQ2gD1h(kk?`AO31e|j^&5FQb
zcL%<q^*6)amCSx{c&ner53t5`Z3UOMfwq`|w48$!FtrAb{U3FcCpo-;YIP#b5g0w_
z4n0E*ap@pp3mxgUuR%LspVltS?{V@c!^6qSVLpsMbMX9Q4GSW~<H9E)ho)E|k=GWr
z6S^hKNwDmm6}=&ka`%d$(o(n54(BZEF`h@FExIK(PNKeIr@t=(YPNu+f=`UGi~(sT
zEPu(*GZI?%*nQ&k*{g%5yi?N87mr=g^QzaTp(9jx<Lx%0iSW)eT0&@0@V-Tm4uAoA
zPhJmi@+O$@I^#&Gg>NYt2zKW^kBEx*in|_e+$Lv7Vnn+(y}a~q9W^PS7IkFvkPV1u
zq@Hb;i|Jw4wIwab6Noa6V3)Ny*sc9QsfI5iu1~bDo~yj#exg0tp*W;}=EP`iNo~yf
z%N#a@uS?<nHRKNOMMI6=%yk#mY{K{t@rBxLg|I#CDyOmu2<-~VcDthX3G!(^-}xj7
zTJMfSmT?K^rxuLL_4b<&hb(5sXfD7WL+spH2?l%lX9O$259Hv)-nFFy)K#(xF?g>3
z80`$%p*gj=beXO=9OUMNZ(K<@wfH+5TEuKU%T=cr_V!O&I?NpGIUPDk#}wCcs&m>P
zk?Ni82tnB0c12r=vx<s2a*UJuVpCUrx+VeB?#Y&>D3yu}QDbrGm^FIhHtd5AZ?KEN
zYXG_vXf~Q)5!G?se?kd-JJHIxQ4rZ$_g?S{f7cHl)fY{S4y3Um0=8BAFexpp++D~}
zN&Ny?DW#>w(UN$NEB>;ZrMB(%@J|M)(fkLm=E<@CQcnqAj}?N)#?Yj%ZCZey{hwFC
z;4jPQj-um#vG|^;IG1_U!XoI;pE}W5fZ+A_r`n?gk-ei!In>L5)snc#_^(_1Yb^eQ
zdGVhnc=fpdLdbLIpSX6!LS(sftYQ(gEDu^i(0gic{OZ&`Q(DaZyEKMO^sidsw&j2J
zYR@y~{;pVO*8RJ99fRw?K`!Y^X>FCT5Eo4ew*tGBC~<uasOtjQ&E@VFh_bciTw<QR
z@Ie_z%VUS@eUHI^&>h^lt+-IGO9>{Y#jMeS5w0qz)1-<jg43XBO&l);s2S46_(=se
z#)Q{}<$v;#6g?uw%zeMn>H9Zo$-epu=u}fMA?K8}>zIH`{cw#~zR_jP4w(ZLdSIJk
zo5}*Of<P9cQE>0R6V5r$G1I4R=4!BbcGKi8k#4=SQa@J*Vr?=db9KldXxE2*SmkGi
zuQd-4Yl4~YWSGYLG&vi$RU6CI){Dq;<aV5zlQ$B2ruh*|(Ip4>V}LqoE7v_roG$@Q
ze3&gUV(M^dW|1hwI|S}*HsTka>FL*~WG)q&LyXnjO=vj4t=M_=p;413rt@=5HjvxX
z`P1WI(_v^WRAt4aG0dsDA?jT^AHPN1>(UFtp`0?G>72kWexBsi3OVC!CpK25jNo=l
z9dhw&h;_Uw@D1Ze0UvplGW|*Pvf(p>w@Poi+JhPOoX%g+Hw`qj`pICsgDhRU_lPhj
zL9xYd??49qw0+5tmK@#{i|k|3tr$!?bm6%};bc*~>^8X6762pmxI@Rpk(2&VQJa7}
zJ&9T0YaSK%8AE0vZEw%jlTbOHcwbT!c|`c4@g2g#9C4%7_saT0wVNk<YOF}#<i(Vi
zNAxbe_gdO2=Rs;vvWLspHyhFx(v}^QVztsIx`VqB-RNPkh?Hy2M?`dOg3#nLP`mqL
zn|U9^uS*ao*A;kr`d1h|MCSTd#@VrNDaDJ;7sorrg~GmNyItPnq_Jc5`?s1k=BeQ`
zaz7cYq-altXgS^o9$MP#`)0%~bI+Xs=rE*%{U^iYgt@5_I=LH#{*VN?hx1fFO~%jK
zN2%p*WU+Q@fW?KKX*m)0Vr~^OVUCh&;96=>PdZok;@i@DBi4Sg<jd=c-CehzKS3R#
z_v+$#i|~^tLIi^3U`h<+t8=A-x6_FuuCcEgF%%uK%oo{@a_4tWeLs&2+w+P(S8vu-
zz60GHMXOBup9sXFE#NC8yD00iD+w>`m{3$GDG#H4dSo7%{gdHl%_(DvJjH8YPN?1K
z$&U|wZrjX#_s<xr{~HeY2MNi)`^(slb^wi=v;y2rnAfBq5nRNH^0DsPtPt3XDOc}j
zIu9@VU2A+}a7tWMD%H;)uJa4OxUI1l?%yYt78&3CgC@MR2mTN5<S!w!dHbIXrtJGM
zw9<K;0K97+0mmM6(qmosqCwk$nfVFbgR&(T0S?H{0elL+*zu0>GkinM83gz(9)RD%
z+mJJ%XgN%zD+H+y`>x;w^zi%TidN}vprgHICl=CGt2mUR3Ubs^bh~?5^WY^s-Bf)C
zB-t@W=9qL^W!3Sp$-{b@=Z_?>9&?-VmC<6qM;;3!bnAa=hzDp<>^(SL(hl<)r9RJ>
zXHk=7ZLmMW1{NiUQIZ4Qc0nV>jlvT6W|MYn08jQJci%r+E=Uee=6E;wkj#yhDeedI
z@b**n@jv=Kw33Mz!~hMs*>hZjp2N=Q+IwkeZl#1HxTO$2<fv~{g<QE$GpOSt#I^3O
z!I9+W#nuMAqIaq;r$$?~R+)bFwu>Y5$}x9XM}&JQSlGmyurI?W>PnBjeZn)puQ>os
zzL;4tJ~ey)yTX+&?O;6Ug`F@dz6gKeAz3xvqbBG0lvYQ?yQ!tfI={g31?2df)g&vO
zct3_-u{s@%gueQUp{EKPH|PzJYX8X+8D-yQv|#taM8V?2px+Mi6g1699@g~Mp*5do
zySbL9Uy*Rnqx9|?!ALP@bJ3o6H95{!uP<ImsBR3ju*ZwUW?w)yRSpK3&8Ae2Ah!6o
z<Ak*|!0{BfQ6$HSvv0bEE`3(pB#7-Kh<q7OoA3UveTF7CRI9~K_`XMpA%IxWZ2=Yc
zGj6FQNPNwCv~&(?5wCX6A^<3~Y;n4c6-s^~lUGGm!$(1tk82v01EsG&8C(V{G(jxs
z^D(7X-)?=V);m>K8|u(1`RK{enV8MFXbZpQTGyWpD*3C(@lWOc)giim3(<=i`^}^9
ztbDqQ{DM>j`f;4Rv0Iz9Uo&5Lez1s6Gnu>dGfFzf#`$dRrKkt>5NR%!AAPnTR7)Sw
zJ&~z2FcmQeRha$y9sw)GVDR7l8p2-!KWP;QRUv(n1@2PO`j&)5amyXqNRB)iNnV5A
zLsf+|KGHfR#ugUS7NSDVhkx_W76KdF8(H253nsH?oHjB#uXfP-3>SFhd-f$QL@SiP
zNmV7L42C2yMcsF)Svs3+hV+6a_2A{t#&K#~OGt}lN=z*9#9V*!Y0&D*E|`?OwA5QT
zNy9r2)9Vt_s11I@NM^F?90N6O79Nfg`-CJ*KA&y&ZYLZ!2=_pIEiGOmw31WuV{RRW
z?HCCllVVqSxVx*6=iGv1$Yy0AF6G50Muj+K1e=RGOZAGnLQKW>A(--Ea&1ye1Ey?n
zH&L!j_D0K-w`xb?Td{cm(6S%{;Zi6lT5dL?O{$ILV)#IHT(sHE<=vZ(LY!AVDzIx=
z@R-RP=oG$2(_t2JocaSYqHKFZE&zOIKt4#h7z|$8xzrxur0B#M^26`Dq;c$N=w|*(
zW9&?NXa7%z>mSkZz+v!>2}`7la&QWItB!8Bo<R&LN~GpcT1Gu+s$?dE!MmS<TE#a@
z*<2&FD;H?0K)(IWYFwqBZ=;9<MB3qUscfLQ!TBDWL+PVrc4Ad7jEP7I5w9neXBXl3
zWek0LWfHgzj;<zkAHKEnz@1cT3pM0BvI~Bb!YDdMpaBAuo3<)wLccKVoiTo8o#mXX
zz@sPmycTvA>BV(p7vtu-Z>k;H@Xo&E8N#{r%7(uTP$^`;>Gr97ZVXht8p>LxbiOLT
z`tta(a1%Km=7~2a%<N`4azB0yfn=hHa)G`Rlt)@={bbV(teFgCyK3#?m*1rbZVIb$
zU;ely8vD|u%#_f3>e6>tI90mg_xvuJlnMlN`dLkYSg!?y=Kk&mWn?_J3XPYo0s`^^
zIrNhiXzH}cFRh+uD4DC$BvL_}n7e=rZw4gafV4GdugEUG1Sq}yumaSH{LMuNP+SfP
zG!l?UC7JpDJ+m76v6Bw_$MX+W>)FFwt$-a|QyE0#v4xWxK!1KVLqLBgyTCGt4PbwE
zAnV7k@RsI}+e}YfZ#k7z5Go<aRKdo^UJx`*%Dat4@zWp&Z^lG;smZnp1dDJFN3xLm
zLX~$QPkYaVfgh5iQ|4pq=WoRvVS=#j{!wJ#dsw3}IXkTE8)Rll9aCH0ho(TgaV)8?
zI+i_!?<MBw4(uFj>co4j;_;yH5P8xVS~ao>N~qVADKyHXdJWqQ_v+<O8b0e4NpTU>
zczpYLzT;Wg)@PIH;_no|F?CGeMT#`+=ksVS``N(*J*#)E=wOv#vH$X&e@|TgGUf?X
zBU$*5@FL)#C`Hg!bp-!6eu95>=RcdE{~DeDY;^t)&VH7LQpp4pPO1UkuKEM@M&G*f
z8y3+3m4pD<i#1)<{X!=;Z<4PKU9XB=JAaJrNnqn2LT4f!Nz_z!_CBf5a-Y&^Pw~i3
zXt_-<6e=AMvruXz<fZi39#{L~3Ttc$e@W^D98%smY1eMx{n@ua>Mrc=EL%O9vGpgz
zX$6#oi~f&W8_H@>rIojG&zqv;?ae;c#wsd#K`+%n&sEMHne`t-*MvtLO^C@S9}6R}
zAD^MR$Q9NPD;qf%G}K4WM_V^HOd4H$-FvRa^*M6gw%{J&lidGuuhwy|bTRwjad~+o
zYNlcNe1Q|-Ta=o}O&`eD;OuW76m9_#{cmnV)S}OJ(xf;_a(*|!>*BcVVlR&LyQH4K
z(Ms95(-xGw)GVO)<jnOu$IF|+l$WH`m^lRws`~=?z4tpRjGDeid?@0uf4<wtiurz5
z>3sd&SlPkM%?_;uj0bnnHF7j?2yrS+Ga4al1{HiXpd!DG6pq+uDF|?R$_^Vr-3@w=
zI9G*U=}qbsWtXizctwN!&d7)1wMI4P*}@R}mPJv%+*UXJ8#D}#)})fx?{pX-KURNT
zJ-pm`Ww~rVZ`i~Z?DEvgtD82KAzqdgxbr(hyg14gJH5}|$$n;e0<^kL5aC(mvRM>B
zQ|{PKbSvHN*2>c!QWZNPCf`_jb*lbMySLrt<WA6GC*aAE4rEHn+&atDGUe&5szPdX
z>u0HOjnYUW;ZtpD&kW8@Jn6kp$|JVWux}WtX6n*xun7UwD#8tO(sR4MZLI$V^VlQb
z`5o@lsWT~OQeewGsPOuX^`k*LR7CM&C2Xx1W5-YWno@lAgHgev_3OQwhOz<V+r!;T
zbz`l^=ThuneaH|0?VWp^E>nSrbGo%HWN44hH)+@4UI0lf@uNZM@N)|V`+ClM?)6Gy
zn27^yII&i5P@ZP%!Bamhhp;wff8Fv${<Q3u#F{8({j%#$f<lP1q#Z(MEKQdTB8)DS
zyrb6pfS7IK6mO9ZYSfKBKLN7Qb*~>JnP*Pbzp0cvr&FcAK7p^pQy~<Sn9nG6)T<yR
zDwvQZztX1Y<_SYu*D~7H7*?>hCBC6|5nROleB&WC24Nc;XsNzYSmP`_dfSAJW?4H_
z3kDmZd0_GVmNnPc*W`|jRe`^CD(7#up&p%k;HKn|#~%7xNZ7nk(3S!m$~1(G%dpYf
zw%>y15*!2RHzIDRJ~+tR$_NI5gFP_X=SXAac@UR6i*)&dGo@3-=q%B?EV~om!w61(
z-ByMWxz9GwSWV(Z_#H@JqvmX-bdnZ2+ZtHo(~4{7bOkZ<wlkX2e15BSqG@V23_Vw+
ziqe-rA4Aag1vI^2MSYUUvjuFLZqtwti1iwwbz=Qiec-b+w#wfgS=s2b^a@m(i#^!*
z96ap}7{zcAyC8#6-(5|`f*Dwb(coiQAEGveupk4fvYj#(P4a)CC{I3rzg7A3!>}jW
z_QN;&(X^v`xV@^JZypGf=8!av1ZLWOG8+=mhCsIX3p0<%)o4j}d|hC5ZZ7=N<9>PL
z;bNC5Erla2=85!v>x;sA_V^r5JPFqk$AsC#c9>A_X`lH-I_RT*+T6;3*G|iQxn8p1
z%DeG0)L-K+Q!E9cGYl{!8uQJ~)DvQ0Ta9QIw+6kqtX@NZS&O=4>)K)YuF#uFc?a*k
zZYHKoz=Fm+5`tvrdLgj_kT^||e5$w23%QfQ!h~U)u{EVJ^T5ag3;pP?vlrVQo%(IN
zc>mtM8=Q8`9Q}0$d@+Tn;juQk9!K0j6QU~Um#wJ|H74wfbs^`d)ur9}?H`N+a+wbk
z9XlWSCv^>h!8yt)lrqh_G3hrvT|wg1zQ8$!6G2KNwXAn-n?`+?d>bR7vSsC>3Af_`
zn8=eU*)9afiFx8TdkySPgkUw&e-;;|ZqjBy*UW58YX4eL;hfrFUZ1<HKHYQGNusG*
zQ_*@#E<>p|=Iu+dcHY&qEKSkxN$6RND4Ga+Yd<m0$Ozp?YFsfXF>w};a~Ie!@NvFf
z(mAu6k}<!?>??0wxQE+@%as8|f_hM8X!)MP*a|8~l@F6M;kR!^bqq2fAdak9JC=A>
z*es^or8(9m^_AX2T5uS!HWZ9JQx|9I3rN56*-%I|Rh4xgF#)+Y7^Fdf>b7YE?4{wg
zB5cZUN0ek1<>s3#-3vIt{rt#uK*KTvwkz&<OzMa`gRn~HEGT}fkMBA-KvUzOYIdC$
zLXE*}8U{RhFe=)ZwAqJjgoq28Zy@*?JDd)^)*jN<4QzuPX>uZS;@Y1sH2KJhwe9sR
zj;ZXq=L9Q{W|bGjksBl`1dh)bieCnZ2V`H%a|Fysp>RVVv12=E_nwI{9CjxQTO^kr
zyhow*pIk|^6CZ%q3idIZlBU}xqlSI&9;%N#Cd#)Vk*cl3!r-MNc14ZR`V*QHpx<n#
zPc#DH12~-{VKuH!#F0L<Z!)18{2n!pcIr*2CT;-Ei<}~Qz8wlPp>ry%P9{)+B;rRp
zj?$YWwJ-H$1Jw%zF?IqnceBm=sJs4alN}e<Kd_<^Z)^XkAw(!sr|0_YTPbutU~>)^
z9%JW$HJS<N=L7^FN}u_1+~HF;evSp@*meKa2Y12DcT)U_8wgEmA8FE*ypZ*NSzCAn
zx`p7QI^7mP#drsEq>dJD<cHz2?uB16<o<Y8T33D+qv4lQm^QIYPs8@FVe1Oy>t?>r
z)D4r@i@c+$Ss!3HiZ9h<PK56>!8VXk(3cgFaWu8osh~|xT&L}eaf8)@Z17eunz{Ou
zmKM6xULj`7<0R;L(Cut-bXCNiPgU(Ptm@+nNnElU<%0>w?WfFD{mx=ttETuAJM6fr
z$z*RlB)&<3a7LpQ#q(*@C#tN5GOQe!Wh&rrerUnJ7J2ENhW@C9l=ydhCi~BHW-^E-
z`-7Yl3Ej%l1}CsiE-6~Nqzz4%d3SnihL$B90l&(R>NZoGy7cTs0a;HA@GfUBBWH#-
z$0qpEVYO)5X~btpul_ba3DUW<hil1FZ3lQCgMTAKD=ORz;>Gn7?Nd!#i_N}{FxK9q
z&XXnA4Gz+2x^oEt3#qdO=v#O31jJbACgpu=rJogJXNvicCk}JXcL#toL`%?|%2hyR
z>a<pQPB2`}`?QhYRAsa%SD08Zsr1{U)4o>f&x(rs%&=(3EUDiPvXy}%ydcm3P3^sF
zLwp8L^%QtqGV3U@RXRb%o8|F1v`8*OOl#!|F4bnnoaO=0T|jTq*A=n=8lez0B1?pw
z=1lT`-o!zG^A&Xj0<hCP{WoJ()15`V`H<d+4zKmY?%qO%Cbe(Y#&Dim<)5u4ND&R;
zcz}%(#No>-(_+1y6GM}7I5X$mFkzK2b@StC_wM+0`<9z`4wp)$pS<{u6-vcmK%8`G
zDxcYaFOjh|TQFtqvAmq)aFBs_aap$U_q*4Hv_b{7uYVm<eUy17FTI~yLgKO|uho;S
z*ZOY{m!Bm=Udjb&Jsf_^pIaYo)TtmHMo^?gcHB$Gc((LyV?8!3Xm&ZEtL-w{dm%0%
zCy*Q4Xu}UdK554E{;CL|tI@R{-2L_kMgI75&80s7BCy9JE2+eaG2D+w_oI_F0Fxmr
zp-xK`;5(-o5<qOubiqMo9A|HZi{HeQ1@A>&ZWW5+b@J{o|9oswYWHU6ODR$5ODs&8
z)D?>NC_rVh0uccsYOHjAD$Wu$_+Y3O#7_M%|D*0BOd^Gl{&jn6_jI^b)+2r7P5g0>
zWY(n+M(SOEyhtSP9#H=D)I3EbOm^#1p>yJC(HF3ybc<8%HPV><0lrLf7dN&IF^yYG
zLERc$Z^Y&%R24aRn9so*Hi&io{>w_Hdgor$d;F;)eR~8eM+VakJ!0XjtN}k6a)qC%
zCHw;b(^T|tx|M(7+h;TXk0Z>8;=KPjB+c+Ur<GXhO8^)H#0T~+xnZ@X0NLDEK{rVr
zp_(Gai|T4)N(VRS3p+MLP9vTa$6B*%*DK1&IW<+{Rr`6-8=#{A7-Ak^vcn#r>9N+^
z{ZT-L)K)@_2$!Z{87xhAAjlBk$7eF~F5?zfj&ntEO1UKZTKu&~yVp2hw7pPeagt&t
z$`BFXiPUe!lO^3$_*`c+TYV7>elO-)GdiNv7%!)LZ}@$ZhTEI)w-W0aEKr%6%Q8pe
zQnGbUe(LV;omzy~fS%C|%D_u?0tMfbW@O7BcuFP&@>CSezxU_$XFndj7;64lXYEFb
znCRUgQ&Rk~WPCXy?6D7meG$cjh1X=5)P{lq9^Lr0Gp{-YpsXnTEp5|hoochYkgyO2
znm!~dn~}^Ht;}phAuYVu8l``r%a5$H6$<w(Oc@HlAxjiwxW7GJ7FemuKnv<~%Y%cG
zutPh3n*$PVwtkaeH~jYN&HBx;bPd>V*jAPFazqrg+`@doef!`t1*%P027SDGvRmTI
zt3vGg7uUOg>*~wOV;vMfc;Ce2L<l2|cGSzbUGR3Z2c;EEXTaR$#o~xVR-umPg9fD^
z&&+vBT$Md8e1?R>%XPMi9H+SjGh0@UD#=;*-XWo1tcQk96qNO8ZhY+e;C-sTe#2i9
zU)SfEX%YM(h>?7hk~|l9uW9&N`j5yOR=aOCA0M2pvb}N>v5z`^5O4eL(V6a~`YXlO
z#%j)4IWs>Ql==J%caEj(2C6AErjgL9-996$0A6ZBu|=4ld?a1_ed%;ToBzO)mklQ0
zuK!V5Qo}j0^3;=>!*_;6FQrY@cQZIPOQWR7b<LhdI6IoufHIubl^8I$4>ehp^L>(i
zHrBATJ@`ra{RjLnKr^68&DL<`(|31DiP3NNM_%s(@>`eS8St?Rhqb=WTKY;2Y9DTL
z?H9Jcf6xAR=y_q;x-X%iHtC9c+$@KolH)aleQ`mf^A56yI&>c7n2w=BND1u<vFSgE
z7YI3FNWij|o`eM&eH<+q$t4xc1{P#hKD?*?p5ql0n-RPbqPqy%;3kV4z>DZfpY0$(
zQet8zY~P%^PJp(LW$&>yzqefS8zXluEJ=hW&5yjhT~x(lQduDQAu(7PO=pyWe;)%*
z?h|k}Lysc*YXsHC7((fy8opl^zF5;p<{^0;zy1(g6rnTXh#a0$nf`byb>X7nEJLRs
zo1(22mdHe+blQz>1){mY;}92h*^qN6K2sx>+b(NQl9kGyUo;evC(exGt&yv~gy29t
zk|>Ce+Rl<3@`w`HO0P@t<WDtw0h<0;Hk(U)X#aV<iNU-7#&_lQZwUlYI}uD!qN|TI
z$>I0m958hZ#+(s@O;hb>83V|^4HYsxj!XA*blK_LXA<*_tXKXp)=iswgqxO;dAe6x
z6M!6ki%3(t1E;bp?kePrK!EI`v%=w^Ti9d2vyFo$)*q&4st6RO`4{$kI(9O9p0sJ_
zPdx|Q6k};tTWID*L8!kgqvW-sxs*UW)JEfh_Zhl_IFs6jTCyLyVp_A}aQA%7IIj%s
zM(pkb3AK*X_XYM|G)U&6L_mF$q^j6}+Jy~T3J~)R<2JQQ1-MK8Lf-NJ=Gxu=c`$i|
z*vpO>(hkLU(07wBd+aGsV!UFeQXIGQvwimmIj%!lYUY_PpKXy2%U|5g!cKaXn@(IZ
zM(dSoN&c?608ybSjF@yBtKkdLA}7H<u76E}`9KF<_MGo@fRvyUx(t_Ic5B>Xx5EO4
zUU@KC0i!yw+-CG)$wR>K$p^l#2CP+)Lhie#g`P6bG?$s1S}>pJYuo^5=$HEqx$YMN
zlUpkxyXD8orP>R_rpOw!OoRWtVa@86ixP1eX|C0SqiX>$O&_>)Qr<uC`b|G5k!17c
zS&=X#U3NfTK$Ua=P~{fuROX{yhjUY`DYBQ7P9Qh0%Y?I(z8q^3#$o&CNA^N920ad3
zHnpUq$vJ0e*GI*uc?VgA%Xi1AkgKGWACa}=kI%lans;IA6;J59!~L9f>gAOq9na>y
zq*rKeF}g!R8bvPFQqd@}6{LWyK+}K=`J2SDbpzhr{bthSWa;(K&#r&9N}))OTrpgN
zuLLL@)Y59d*+LJ3)2N#YsLAYVy@12?5fpO<CPW-7H+;9gYS2dz#PLwWVG~haaSe52
zvdvQMb&~>LD*kvfGfsI7W<ltPnlpC0HC$6rpM~%dgmZ|`-!?c?vub{7&Tn?T?zdw@
zAir_ZS%K;WPasePte`IRCyU{XBAV*&QJl@{+$QxEN;IkSd59|Y2PQ<F`W%mAaiP>%
z+jgWlRk*6u$p*FO^gp`H<8thd=7!GP={^f;@V=hqvNkK3Ju-g;5O7C+BKyun;k$xB
zlmMKm=}Q*gwX!O7&&NxNQ|IqF`X4gn%u)PEgtyB)ohB};<~FW-d~V#cau4j|!8iQq
z1UI{a*z^E;oe-|EbL6blunOmfHWi{bi&xvws+?dimrR~M;d;J%zB>!QUX}%=Xi%wM
z)9?x@Cwgs5fyLF!C7sfIH1lx3v2&06TjbY|MyD*;I*xxo|1ORB8bl>#Z%Ce=y*DD#
zu=<ms{89PJTJ?l;Vdr1;;0ym59M9j|1mQ>+!ZmGaY99eAV@{nK+nbMwQYd?j63*Yd
z%wPV{=G{gda`H5EYCp7o*E^HZR_a-h3;FOI<rUc@EXXPVFq!PHzMEWbaLmbTp*|OH
zZj8EF-MwifV{pWwvN+nZ*nF~iL`3~11ZoSP)x@O{)B*vP+-W~Jl}}+I(1F|Z=2e?t
zK+Vg4kss3$>iq92iv5{L$YKP*&Hr6^d}BQh^xX@KUZKzZWYAY;P;9-}=}qY4<<GNv
zm_*@*O<y&?d+x<?t*bmMIj3mGfLVQ*nxR8F$iWz)`Bv0Nn@2}@8_v{aX4{xAaENKx
znjDUIN?_nL(e$)4o<;8p)rbJqu)xY4jIoO&d$&yN#L(4))N-zmZLc#=T08FzO}Sa0
zDX&3+HB$x8`qd+|C2&<;g1REIv$F!=ZUsm;`U_y=4bs#`o6eHY_20Wl)l;%HF>F!~
zDad!NQ(5tLx^nH6%&xKc<CE88`Omx#Kdt440Nm<-Pk_^8K=5y6hM$BP+=wG?BZL+$
zLZlKb$PLEqlOD>EAz#SHk|vI~_6l6Db128h96j5j+odU4MWUQB!iQ8JAMlOiYB46v
zk7#PVH^HB%b#s;->%Ltu7Jay%Vo<+1egB?$+=Nu?`7RgSyxZPBk_L8bFezUxUk3xQ
zPRfdcEzf~P3O$DWD1?h-8#Yt)f$BPThOMLH$8FT;ogTjMp)2<knXc}SIoh@zA4HGV
zJ-HFyT;MEl>;qO-^Mael<1o#bl5#nUZJs^?buqt^q3VwKy~iajVk{xLK(26QMuDpL
z&c%bYVRSvY1$FTwa_IeV#kCV+y{=<dUd%R5DxoZsH2fb9&Ytf~dtT}=`&Q6sOhWkN
zC>jq|jrNJjGQtcl@|~((NKNNHMFevRbjzAfACGCNO1gHl@z%+Vc?;M5Ak<`ha}BoJ
z?w|wPkUld62-w}JZFVQ(j6?vaSt}oYeUAaa8!`d0V=kMu0;!i<XaHBmm4-<zFJ9sG
zPaE4Ph?V3qa(w=X&E#y<r<?aVFpuIDH|No<XzC3hGMSBHLq*{8aU7`0LRiR-aq98a
zE%d6(9Q^WVig9!GbikW%J<}*1nFwZ!CU*S6UA$d6{*2!bjKDV%GGZIR`Ib5rx_I04
zJ#g?0`Dmr-H&~KF1Rz|H3xbEQcL}O-32Lah{A5@y+6d}XNd6{(%-L3;8;}heCJkrz
z5s~T%FNZOyd3URZ<!Y@Y+)f82b8qx>O6N3xQf~at!+7+GiKv^aF#o`6ezhe{!tH5t
z-+Zq)Zyg?*X0MtdxshLSPyUpm19rZHpzTgAz&BgW@pQ&D89L#ceSMh8cw>RDGUeI}
zJ(MN6?U$bZ3ggqfu(@lQL94|fhHMubOTc)+MLI}Zcvw?_f;#q+AZwwr2SvTn&pMNd
z5=xZv)j?T?Rh&tw+5n_8Xl>D10fN9*e+-n5K4zA?oBmW{SAYE{gJgqtI4KM9qRA7V
zrPwAQnO11GB#_czYi$E}Dhqk^_;Pvez{Rdr9p!SM(g_lQUStG-?Lu&XT=*oK`P>kl
zn|j2WrZrL#WKaNmI`dgm^#k?fo|}upT(&+SSxKZ74a-~KO>6QJOAIQygFhjX%ezWE
z;&_dt>z=#L<cmx%McfMb)+OuqCI1~C{}uSgnv(*+27*TPgC7H&VSk<>2DGywK%w$G
zHULcj45+yvD_841I;Q&TRVDbJ{|&5k`+svIrYu)K6hA3GVJ^=9_5f5~R_U(9WIJo-
zP5M1Q&l^3@D=&vUJf_pV)At0&N|$lT;ZZun2pOOsfxV*ogRlRn8+AW7aOW&@Vw5eY
zhULA^_h(_1SMI<ndI=6;W2K}{0x>0}2Ah_qqWbEyhkMxflu6@QY3X>YC{YGAacl;?
z6Mk5(Ns6S3FMMCB5bJ~HvR*n>`!%f?mTVq>+~J$9*f*`tq<2mSR*{DmFWu!^H*QW-
znCpXgvq4y(jSDn2=x#OOfZ>@*I|mFt{WQh|fD0s#ehSSFA7BD%7Fw<VCFAL|O==)X
z5^|K@sh&do@tHZ{6s*^bd~fb<ZF5Z%vPlL0SVJ=Vw91q%Skw}?+Pi_C`TSY=hcJi=
znTs<rICxjne-g%Qk$`DHhHvG$+2e5$dub6q`;&{?f)5vFTD`^id9`*Qy!fqZf}uMb
z%9C1WvL2mAW5W(NBhqm59!8VE+?_|SxZ?u$08qf6rltrfj2cLPrtU;P)cHH3_5v;z
z^hi*{*n3Xa+N8B{w6cck;h+$4XVNIM{h;!!;KbCShz6ECamQQW{Z17CuGLRqD<TXR
z2kDb21qgtlmSKpf<cK1j*@=Es`H43%rYkRghQM3ZtT-R~(F?YVb(b-(bpxDJHec!|
zm+#p3ueDYJ0PI>FeB&ulDxZ^=^XCo?{?gwdR<hgdT8Sqf5#HIZ5~4&-tu2H}yWJIu
z(qvYRRO;<!=B(^LnlhDa<$8o0QX`Z2oL(*ZoB~-TG_=7F)3}7u_<GH=r3Qt1LK8x1
zUen5jxRl%7GEEWlC!ef-sh_xYwa6o;dlG9ULl^WxHOnJNyQOqylTOk$()O3G{4#W7
zeLNDUasPSm9}nvMs|hg8e*LQd8kfH>zyDxdxW5D9Z#e7oe{*@?ogn^X7|kg+G6hTm
zfXh1dFGY6%fy+O2?Ho|AY_47Fe-(&60>^_boMQIj7mR()p`t&K`E~~EFPYjD#)tI=
zSaO+~2FjZ1*`9oqHErOEpM32E@HYRVe4bLEo|JcG<Yo1(mC@5<1uS}Mro{+;$(+2w
z9C1L?kBq|rY61c*Pc5RMzQ)%lI*+8wee8C*oF9kodvn1-&~mQ#K~@_`$O1|KiazOq
z5`a1GK`oEh+M3P<$c?@XIRCuvf><0*gDc^-hDA;Hz^Gd3;-&HV$IuIRJWYSN^h`fa
zg2=I__QKrQ(=5V#)r=CAx%X`A=uc5I;pKlQs|yztf!zdrN5j6$csluVT~(65ToHTn
zW6r}9JQ9DjkqG*do1ba-W&g;O?cnB4tI7C&yD$sYXs(>O?dfQEEg{Zp!dM<_8rvlX
zJ`D}&y`RL{h&*!4@Z#}J$z=Z16@l%LkD0@;qTk48NvarJRGJJv;|$~NA=={tF=P_f
z(<sL1rHhx2hl(ETi&5ia=qIRjNN#S)G*WWgkH!TFT<{Oxp0Wm~$SKpT;hAPAV!{zJ
z(<%}E0alF<TgVWJs9|L`sEx+9J=M=}Cw;yQ?&X=z^oz`M&G$Fu3h(R|vrh&?s{211
z8{kinO-B`E5l5$&Cta8y<ulu!W5C%XgNk>PzocF1+6_PP9a}kgi((FTr9-EGGT1#f
zKLD|>kZd_HRHv4mJ%x6-5G*?YI)kyP8-7q1uVOE)-0)TClQaWUpUwU3h%1P5<R7cm
zfpe^E_=+;xC4%5SNXQme(Wy^nG1h(?d}Odj&clnvxT>jMDB^^u!98fjAszr~&%0CK
zFoBDPXWriP0>B>6u=A2I$q16b8dit2sv^uxlERcc4zhBquvrS>=pWBMZZB>4%KG!d
z7elRwI>-2=_%-DdW&ub9`cw@l7R5H0fvJR9w1vow%F1sC_P)(S1JZ39bM+kHNIQ)P
z=ib+jsiI|bXJ2e?mva;9AI%RI6(OXSjuz1_x)s#86y_AEuBbe5aU=6~T7m;F;UD0?
z`AWodT&P!tjKAm?nw#G@n5d$6rqHK-Q0#OZ-~oYe0Kba%Rq322)l!rM3g|5b1QG!y
zF9hu?X1f14<7O9Kib7*)AX6(Di2;DRLHD0op;r>ZP)ul&I<*F3<e@-8MtUeaYk&PR
z`QS~7+2djii_r7LHP?a!%(VU_<=L2~*RXY#<R(VNTIorIho&*>=$9|;e#CF)J`8JL
zug*R(d2f0{CM^l)VD;2%gRc9uL}9@naD`PDfY{rmqko3duax|Sy_dmT<R9F-0Ve<E
z6YKn{ODa?bUp@;|_`*~Ed5$KA{^K1()M`Dw#P^QPp(joq?ztJ$)g>l2#R1WwR$B~f
zGUlt465x>_QoM^RK{2@0#&V+N^p@?$H9je(SnM+@;AF@#(b2%ZxCMl0gB}cgttRWX
zR67ex4l9{Nst0Ems3h7~EPb@IvTJEQTECwlC|j~$cG>Wg49fI%rcv_{1PfVN7ueL8
z7FH(nra?<w>@GKqY^~&TUJY;@^TJ4H$5vFn!b}M?{z4fOl>YG3Xp`w|=_O~)6J;)N
z9)|?W7>u{vSVpleDnV)8^TelTuUR7ObHgy2ZSy@{Dw=%cv<@%Yb*hUL1skP|W2O!*
zESwE4U24$FPVqcdl|3(B8P3m2vvj^2Q$FV=l+|vmQR?jj*h(8_3LSgUK%a5R%)4kk
zczrkMKyuw;J{UJ7P*5_cG+Az7O7@)(&vD^0@$~hjTqlGpAALJLaf(OhNYX6^o?e(L
z!Qus7%$4&S%!ELBO9))d+k8|zSgWP7cJJn#r{!QI|2pI>Y_Q2<jy)WZ+w+g4`2tkx
z;Z>xfcI#M5xzmHAAJ=NCk4bP$G95RP;k2pyep)*J{hrA+l1GB3KZ2%#4Z530^}vh?
zb14*q;Sb&T1X89qyPYIHg}#-O3%wBKU&oRWeZGZJ8R|9NTeRB^CpPk3*(h*F(&Tn=
zArSzm>pl)X&<Q|-ag7kVqFJqs{-Dofx$bS?y~4uEjJ4WF`^UwiPml0A`8n&DF>o4Q
zVXz1dQl-8ptkhBO5^N3Hag(MnIv@lbbeU2&IJt<2N_~=?9#YQOVLypc$P#}m?^!cz
zA>~GnASi^<)aIPXd{MQE1WA&15g?6tz;&Z>NHXat3-T$nuaYx>p<7t2Dempdvr|td
zQ?g6e5L8Jb6>e@HYLlPlhvstwD4-y6X7ua^%T=%mpRi&*Qav=6`I3wkV&c*Hubi5g
z&AK+fmhRT%0fFXU=S?-RU!8&_x8ekAyJyK~4!&!jC9h>U1UqOo7j{Nc-G~~u%cOh!
z#f`FXeiCaoNMjO?;2*P|0Z1XP(<~<lK|1LKnSK^Qqel_fnO!u0tga@OjX%d*KHAJr
zmiT%&^Q6?8!Q0)C3mWjx`>PIv+RFgK0T5Me_%R^zpK>4t?eA;RbjV{xB|)+|ZFKpJ
zwgKv7tA%xQU)dOM-Qf6?&RC!|7mMpRhk7V}{Q3Qpt0Hp2s%Gaut^c5^pq1{L`&~j;
zdY~8Wf9JxQ2Es%9dHcuq>ET)?^n(e&15F>#eW$NVVcj`kd75*u56tU1CHS8{``^6J
z|K$ys&H%;uz|1<NEjn_RV@7^w9J&e5`^{hFLCfr1EXyIBI02?efV>P5bK^@l2@8=O
zhb7#rzO%#o)QpnjDSJcE#qnP3E5=ib!{Se|s`~(3#2Jp!4DzS<ZlZ7p$$|3$B5=_7
zpp!7Q55Ssl?km!HcUA#l%W*^^l76aOmxz1NNqc$_4%&W0{IRz3^<a8a4-}FYvq7^5
zBFc3ks!JsxKcU{fgRkmOQBcj*h8&+i50k;4h6r%%8SMYnVD4Yz@b@+KKV=;L2PWzM
zSuU&62<kLh;0GR#*^dL8)JE@H)i>4*)iegaxTX-`>v^V0_<>L)rl;(iDDiH*r)~Qm
z?m~AEnCi{fZo)XdX#8}-b~c@Pg;h_CGee^{;Er2B9VcU0s*=PB!DnJ(Y6TPfwvc8-
zc@q)&2D`m7M7O|o;e$Y$GyK%7+{LFm#KY<R(LXG*S3avIF7dAT#IZ!E)REhep1GAX
z1%*JN(MvuxPj91xf||HnD>ncL6@Zken_(YC3>yzOIcV}?ZZ@PFY~ZDHy}hb>B{?Ku
zyehJxDp+$hq&5>Zjm+0>cOtoC&na}PToL2%jWZ5>x1wrs8P`Qo<;^gk0AFuA88S2k
zuXtV2g4Xtrg|G*4kZHn4Nja_4Rx}kr1ZMhcwmtc1!F(qrg;*X-?t?O0)d<_%x_L>a
zOJ3+yh1sDeoO6z&l!kpo=~_BizPaI~xseIgsUj<Dt8ubm$89J?-O??Av|}~u!+v^2
z_*C<dtUMysSdy7v4<jzylN3|x5<Z0>K{d18UQ*#?K{FBvgFa%rlHMN^B11)YJoRlp
zD2zE;n8v*n<CBWdyKpA|QY-4=d{J_#{)hHVnLzdgC7{3xCfH^zm)Ch_EU(CByls<{
z!|wW9?ka{?m-U~p4>e`bxp6T~7erduPr+Qaj1AhgIY@ZY^xE4Gw#HD0;3HAch0|+y
zE(l9n0kRkN%;B~@za{Z|hbQHz^rrz8q0OkEl3iC$Z0a;OnlcFif^5jCrxuu;ND7Ek
zX>xbAazqJgaI-mB<Ia$><)D0-JC7L+xiLNWI6twW#(lA8Syx~K4Iq}ZsBaNL1`XBJ
zDg^cMk}#lgyw#^Zd2^_!GENNne7Mg_j%DLQfZ!9ZJdTz}nh95_#<Nf`<}$3vhRik2
z&Oyz)E09I3$PnvwbLN~8guRb9xaM-W^Dx)hYg1o3U1JY*0B{;=?JR-=HgksY2I<kB
zMI-=Y1qa)kj;+!5m6J}}ORV1Me+;<kb<A&ezOyckSWIGfK^tzETev>Q>2hpwPHPz*
zi+2gBTiO8suG{B4S-O5`(c}Zwebb73FuImkU<Bkm%dQq>x8i3Tv&FoXJUbbr%UC|X
z&>Xp0{fhk(*?{tl%oHW3EjSJnG$j8{@p3{E?U}!vCJ>~4d&Fl6`o7+!T*a<uAnN60
z66A5>o53D6#{?f?`GW)~y?<6?#h9x#hf`UBy}ekk#AQaK?q1F4BUFmJy!&YCiFP2n
z=WM(Znfcg1VICg=EePUiszy@pJ@#QgK~+&;SwUJi(_`@Ym&Wyz#%FXmai=;Lv{xc~
z6~zN9iO*X{JFra+pj5z262Ou++^sL1r=*~Mc%=J567m+P>-kibPSd6aI5#z=q6s1u
zF86q>XLLS#)nEg2JksFJ3Eha-sYkp0j$cMV;;rE^O3hc^&`f8W?|^v0RKFdgz(C%{
z@WtZSx@wZnv7FbsI<DK)jq|x(_<8|D0@=pzeYN*5dK^LErpw+JPnI0SryOQaK5%~|
z6g<VaX$DSsh7!%_RA2k3pfu<fJlChu&T%%?m@qceD(IvycxizzE|MYTGyi>9S@CqS
z_fH0|pA5TzOPXc07ra)ONoy|fpp|&^h`^^aJoa_no{lAu@4L~3$O<offb#$4-olBM
zr>;v$Q|BFOCf*AZd|i+>e7$<}yN~nMU)0XPs^{_nRn}wWc`n*)WP`XL19(H<m;`o=
zt9|86_QSllYiBR_*kg8|oDh<cyyu}a<sYSeg4%0VORdcX*M<7Pd0p!!#&X~itKFX5
z(RGI6oUNkLe7qH{dash}uU_pnFPmYYM$IY+z~*j~#O?S;YtW}{J4^Zw(7~QgDrqpe
z+&X`~=o5M)4x&+Q4(~)uS-)R7kC3;V3^ChGNg1B#*})p8HcTlsts@nISUY&Bd8;mc
zV!bxT`Ug?%{RG5n0v3qL7AH^bWQ)YK^I|Y{13YOae)q*$J(X?<&NKCjYq$%VMczlf
z#e=wAW>iS1M8cVK_TB}qb%_w`oA*j!`8tbF0{D+SaqFcE(wII`-O9)1vbI#8Do4C1
zYf@lI<^@MdzA&^Z@MW%%IdKRd^;Q4kA>|NWDi@&$sLKft;UUH4&#-@rxi+{Jb4a$&
z=gUUt1zVl(((^&jP$mq!Z)a4-x(&>J6cs69=}EbCxL4q`4{}~)w0CVo3l-f?%Atuv
z#*rc<fN;g(08LEOHldijoexgy^Qd0iRNK@n&meECPa|W^M|FHF|DO<77Jb0C))v<#
zAYx}EhiX0I`aCKuVM%)(z4&F5ytC<J5<TqaCxffPr~5w{<)o>#C~t*%iZBPU!)JZT
z&ZyKG2+`xzjO##f*4sMFr8^(SO~X&CF^WwMC6L`I^YGv~58z#8LeUWp&Dd#Rlc2H5
zv}h|)pL+uA@{O(8GEIZ%s@Va#ZCgc?!Py|4B)jRhaI5^?f|R@$m%rW9lf8FgvKeoy
z+djicotrH`JlBauDZgxcwamF}Ew>%_TGu%GJf>i@9{F7Sy18xTL;p1Gr>yPI<;Esy
zTBA^u)=IXTVcFM=0DPd9z-cSoSH4qIQ(wOHA_b3UF`TV_HN|#<0B9ab5#Y~#k7cxL
z^O2wwL}~8=HKKhCb)hkvP7EhJ7Y$ZE+?G+-$&|^dEOjKQJsQr!H4W_OH&W2dZm5}(
z5o+OF_54qUA$xeoo<FXCrZ_kH;mLW_zZbs6-u)F+!0_K^N+SclAwU)33x(fBg|C}f
z0*L3gJl|hkH22cEa_f`wD#ss?f{<sLs(Ms0!Wj{(O7y<|Th(!Qv7Fw`@3xbcDvI;u
zOpEu_G)xk8_?7@XQd11y=ht%Owe_;aF@osqDM7bMj>P2t`U+Gu!|ZnwYvu%Ne1z1q
zCJD)y2kN%gE@eH?@&$9*Bm<Fgzr14-FBy(YDH|A24d21wZJcfs5R3xjQp5LT<=$mK
zPXnIb8=pt>FP9fTY%>&<%X~##b{i_rD}6B3nd1s!-RiS99##U9sS(+1sNt<oGlI-U
zuRiLz2)>US5?9GT_mMOIJVQz6jV1dGyKtp=EwLN%N{7rWOkXklU+leiSX0}!HylI(
zMFf>1ARtOnY0{Mz5$Ph*J5icIgeWx#7)9w#KtMoxl`6f3pi}{oP9W3-=`EoK2=QC)
zbM`)GyWhL-zUO)G{qFNV=PwAxRpwfA%rVFK73!PRqx*)lj1*1w<fE1N-C#)(hDFPx
z_>=@3oi-U<X(n~r!ooFmva>2Z^)>w$ZH?S)($zi*DKxtl5jveLUpayiGAXiFi%0`#
zB~Pjbz0%em1%CD9GTnN5dF;VB7K}Yw0i;dA?m&l5QI7wxE=NpKt9ezaOGWKF$a}|P
z>Z;lj63_8`I#D8$0t4K6VtkNW;}N*F7J6Gj<Zq|n{HN2eEoAN=blji=H@i51M+4fM
z6xhpUEb2{-eN$YyM!B8pGeViCE^-Np6+HJ4Nh|tETWK`o#oMVr(=i+FW~BJf#L#|e
zqx_E?<ckSrd8eA~U3ku-OVJk2Dm+i!pHu${P=(X!*+-JE7eha#(ztO2%CycEVIcvi
zllZy!@N8ibiL|PX&g^O?JI#A)f)wp{UMyNBDATICu@{K5|D^sF`FruH(TwM!sNV}v
z{Z2ftDDg7ozYp;@A-#W(#Xm#^|HsDSN(Z?VRiVSb-?ra(^Y31>aTx$8>xYK(zYKd-
zenKwvbzFN%&X^$syS$YG{bKM?WRXQLV$d9ch8)J)Q{Mg6ui@t=90tvVFDI)b01m+l
z-9vz`aOSzpl^>E<r<a*o;?Uxsd|?vIUkTcRE(L;jUj-Uoz-qj9p}*s*qvS=E{*Z2{
zkT8zvVe4Y-t#OMwXKP~oSn()>s(CNf(69Sl71-H@oeiY!(y(Y}m_gLqJ*G>3G(A)$
zKy$Jk30aGr)xAdrSfnI!9Z~`5Ch0^Z$sUNf<w^i9<!kD9&<h<v{K2vBCMyIVXAfAB
z4*=Q^SHT289-^C`{89!%KBh{REKmjkHXA1CfC%J6Z6L|Z1GFIkfddKf{>eokax-xe
zhl>~~VuAgHT_9iro(u-Ca{xY&4bX-F7(WZ*xd6LJA2DNE@$XCde{fWaVYr7u6$R@7
zC=tIgz4Drb3g|>>>*KI?h8x#G`zO}^Z=gDjmR|kZ*Mj_k*K_38$@f3_^MuPU-f8MT
z;?2FI06S=@XE-~r=x@k?KXm1O?$Z57pZ^t?<8|uqpb@s7@1Wbf9+h)!%z^GW*y!e+
zz7^QF{t|cklgbXqNG4NGmoUdnsAz8JDo#fv+f&Z1hwM{_$X%K#;Ub=G>R-W^M67Lf
zHE7;*xaq`fp|V<JclFhrwtQ4zS90yV{T!;SPl6px3rA@6xh;H=d6o=iO0KYpsk+>I
z@nlNP34xE!V4He*$0@TxVG_r5GP~Xzhm5W}1Gvpq`cY?7FU>E*3mrTD4`^4FUft79
z*o6)$FcSIJKh_Dqhq*fBEb2@`DrZrVnX>Y6$*dulNqNB$asl>}dO^L;2{mgs<z;@u
z#&1kvTwC@_fKA<-zIERIb;)puW^Y8c^O$W}$i@!4!^5y0cXi_e52xJ9r?J0=$oLt0
zY5;kNZ_C84GZ?qOx8HxI!XD~y+@jO#sMzbykIy?hm9BvNXsBWq-QC@D-F=O^?Il(Z
zo#r>jH&PjWgz+LPi}?9glpG%H=O!f5%0?YnQfxJd>U`D*w|<e;I=*zfnma|(PE1z(
z+C=h{q$y#y1;WV1;Gn?TJ3k;Xx*bWq^t89Jz9Me=(-X?ks}IiDcigHLvvH>8rg5f`
zNd$ka3cS<MFQ+`S&H|rr-ZgcrGjm`g4L>uAx#;CoXXb5h7h83vK0B)Ci&@-WelFnQ
z1kgaqmrDt!GI8sS#76znK4&@kqlw!?wfd#8>DHVh7lN3nC=UAN+9o~Au~JuBzlw$L
z-kl<zjGw~+0CMMdnba&!DX6FITe?$q+lGyHv?XVjz#t^8FRsCH%5q3V0xFwYr34ie
z?rT6k%CP3?uPEE_)QmlO*F)%OxC`V2nWG(_`$@*rzo@c7&thO}WO%aE!u+G1-vpEH
zbtNUn2eC{yPh#HzFj-0W%05#Kcizh@UmbNG`-N32Dl;1LrT>a*`A>dQzabB#@DKTj
z;|p^ly@cbv3LW<?U%3l-j0wFlivKXCDfh5EU!&Em+(|wk*h!QFs>LWaqN~BsGY%q5
z5uX~09k2@!VMvzGsg1u%oAmyS+FYHWzNiDQpgA?qhwi7~FgdPFA|NEe78!_CwC2pt
zTsnQuStyIw^U61dP9=dOlmb^y0^-vb*uIfC@FBa+Vl>+#K`0SIKw?~x!>QyFO*?Dl
z3&L~JwDy4h?l<Q|b$Z?iv$?`tmaDF>++rG1JBJU}$gt8Zo6PGSOR*Ju>xUkja^BdI
z7W$Bxj|j>wn#!J2j)~|{IfdQL<+_|$cm7da$Yft{2iKUixrPt=6xqtz&ai3q>O^i6
zgmIK8(hQ8LyM+pe%G@LuyQ9WpmV>8E*thaFsu7{gXhT*x>|JX97l{ibN<94?JOlnQ
zgkiaEigeL1p6yxeVct$lJM`);KW0%y*2keoj@fl>j3p4;&8DX)M~EEmrPe|%jTlUJ
zTM-z;YW8S=!NlmqwM<ITw}ZrmceMo?Sssp%u(bYzn7!Q_Gc0V^UQS4U2W53Y)}gZ_
zo00xp06_@@#TJocix-~0Ds^N=OJQPj(y`qtO`UKylZzwK<1O#GFH_x;6?=!Tndv}0
z9DA?8=c60g#K}K|B~ZoS-`gNg^~lsZxwmYZDZfASIe!&;-szDEEv2{`>DfwZ_B!gk
zadq?9I`sk?6X?I<{2-AL*>A_?ShZ=AfZ>m|p$NA7bYVu$n(|!xw03~tJpL#~%pC5b
zee}vpO&PO`YZdk_E*K3CGoQOmtaS<tLobhrS#hrHH7FcD0`#B87E!AO5}|P)W`Vsj
z9k3^bxsTa;XZ8Vz5h|uGAH(mUXtNov14>K8LSS*uPrTHB!72G!-rrrb4EFz^k_eph
zfBbK`zvRugf3XGyY5&dfBxLH)WPh;YpP7UImES3t+V@Mh3+m|iLnHTrQ7tRd{Uhit
zjZWrQ#2gOAjOY--BMx5_0$`F`Mo>!?Ic@jp)X`u$Wvka+-dRLN*tDj<5W6~4-1bR|
zqwlYZ`<;>v0WFXW2uhz&v?xt324cBVFU%(BP8=pBn+_Sy%0qo(<+!dvz*8G8jhLGR
z5C0i01<iGOsHsQ2?fC$f+5zF2ns6J?8UNR>&r#_`)I3;*urC{-L<#&s3QQy=qW^SH
zfXrcJm@`0%ZJDxbsHyQAUj4FskdSpvAWzI^TKvdJhb|P|Xd_GOB%nzwwY7DkOHQtC
z5wbidwA#z_M5O?vZ*cx9;Gp*ylt!QJ1oYe|MRrv;-d~mdRf4Ds;m~t-3B^h@)Kf*q
zYa)jweMN4`#MV2-bqaKiIAy1Zj}Gx#6^&M<3_4;$CMx0!JRm`&$u3B};nPZ{8JX(l
z9{<#Gu9xoXL5Zq<ZWChb%^2nG*01TgWB!98nfjB9^6mRw0UpY{IS+a9E%S+?T7;gK
z5^7@Z$5Qwbgw~M_5?Y*kml;DB-i^d)p+{|N!|(L>T&RzHbVkKomE!C0n4i8x@k(QH
z*E~SK>R;9`CXr{MjX=u(d8bZb4t9#<LQEl3;M-fAoX46no6s1fjKy5AHT202mJ!|d
zRUmK8)c1KH{Ow08u<-iiJjksMkli`$h_*bBA2V|Zow-6XLS!uU6`t{8-Z4zE>3RHq
z&Y4P$E~nr4Ia*Ea(rh-DFJEI(BfS@-@?hd<!;t5p{X~NsV!&uu<JWiRuly9~{L3Rf
zWQrIC)bgh2d~W0~0aD@8X+-T-{9?HhOVHhv0rRfdW8H(nJgyeZSI<>M#|cwL8{1{*
ze#^aDVvA_t`3^ETyg0+@&_?SCYaQEqSkea(@{uIIAeh7Hr=PuGuEloinGzlbPLh2+
z_4|vn4&-#qFfy!{4(yhLwG**ZU$nk-j#y-yRS-3CN#J8d1kSl2uQ_jgta-*s-9jQ<
zcwXmS9Q@<hu?iQBX8w0qK7PZCn|8{@+G;h#ka*7l!EEY9m~Ho6Nfr#0pfv?(V#spN
zDEo{c91Fr-cN_@q;Sxn6dM{h{IfL^YPZ#Y5O0v_+6C?^ws|#D{zX=a9(HRN5ChP5q
zpa%mg_<^Xi7W5X{^u|jl(P~;({;(3OGs`TlWv6+{EYEj-YQ8pFe_!td==o)J`AN*P
z#@R!t<5S)r2u51iZb^G5xfGD;IA`<|ZNDz@_odrjYS9BJfN&#Ox^Glq^oIPa?fl@+
z@rI(a{cmQ22K(uMcaJw(_^Yb7|DE4E^6pPumH(4O=RYFu{g*_yKeq=6Bb^RJ`2!zW
zo-1*O(FnadV=ihutp2F#iWLR7;)8cbo{XT2js8%QP;f3I(i6&c%ZZRQV!--GO>FA1
zC5**>M^2abPKR|lJBW9{TJ+Y$8|L;4Mhzrcq8=9Za{uJM^eeZ<m2fHVIAk<eg{IV^
zi4LuE)X3pT@BQ;nwu|FiCNK9PFXdFF7cULV2tB_au&h(7cJbs>y0x1fTit5ar!B_X
z<<9W*eopUK8xgCEdacNz$gx0a^;^2%MF@Pkh2SDu>)>~T1g!@vd~aQ%Ypr-yJS*qN
zs%CrtrmZdao9Bv0DNTXSAz_{6AahUuO=4X+$7;S@voew8n*;i(cCef$AmK@Tifhu6
ztbm>Fu<LzmAtklzhq;pDlwdKSHXq$Cw5bcYa41&?Y&Wpq8$?2Iuyd;E0>#0$xZ_1x
zp;|($L3I-MRC)Tn?(_u<a~ff~H*Mdh-p$Bg+a;rXcb|r4H<m1G>1@V-LX7(PiU6wH
zfvuSNpgU}L)Arj*+~pRSt{l~;{Bdg8f~BnQ1Y|l(6TWC*)f6iz`hCQSPVTORuC}SO
z+b}wFxRv*Ty;}jTz}Q>6<e?*>p;Zj(JPz|TMfX5sRhakpvv3VL0O3DQ6QFA_!Me1y
zWCZBVrivP=5j_xykxR(szsvxo`fo~oOKxYx#?<+$dS=&gT2FI5fBNuw;7dPZpv5pi
z(K%C5Gq6>%s#?V-Yog1@P#Qwh^~#t*KnfB!-qR*CMMdOn|F~Qf%IeHZ9v%{wbRhS3
zj@iBS>#Wi7KX;Cre;lo(Y%mrtJ_Tz9ck^k*);@*peFr6py7e4eNL2M7YmVD@6H2es
zWFHx~v!CF#b6|QIQ)^srb3W|6Td}}Lzz!4gQQFjp&9jB;To2awMsPX4VYR!k>UI0q
z&&W;KPIb7LRKfVQru=t#txtU!<06*#S9jf3dtnCme+Xy)0hQ}6`BPSycFOu6`XeP*
zeiOvb`#J1T#{S)Wr(pZmuc9O+z5h`V_@Dgyu^ZrF>ERGy1^dBxIO6?_Df}Ri`~E0!
z{L7^mnx87Bj<G-Ig@CNHGUG3y-Y6L^ZCQ#h`FKceI>dh<m5>YfnA})$3L~5Aj0w~`
z-BWM!drZN9-Y5ed6?xJEKDJ|}eQE?LYPmGYU`E4JV^vxhJ^Dzu<;X3+=IcjiO4I01
z6AVLVz~K-Mwi$I|QXU+v-9IA}AFx$0MKh6kY~xZjGOBg(+NIY6*k=wJp)bSV&3o<`
z1h$}P2}Uj20W#<iJHp-g>Canl-D>Bmj7_i42bh>p3%06$P80ud@40Yas|VZe3k5!6
zvt_6bOC8s95h1tG0>F{{POYz<ax}Eyiu$!~rz1|I4i{Zo=5F4^WfP_GMyAAvI2ao|
zu*DMSo4@KYAxMuP4^mgMK8)_zL`&}JjNUaEe9W@u$@i9X0`?umNkG(rW4)+XpH~$f
zTqm|zjvW?%ZhypIy>XIkgK!LV*}H#ckCpu<&-^*b<(*%np&?5PJ>>{V4IjQ6=zhcb
zLK#vtfv3XOgsOvC(XX>P|0H2R@{D<(y~F7Ep11m-#CiZHA3-}<;r!xoBMpfz&ve8g
zZWsWL`ojcQw>cnRV!uJ<neMQN_k24VNfVWD_nOT{P*DK_gJ@$(J-O!2S6w<#yfww7
zp0+f3xdGpn%Mq7qU)3}VoA3YTR_9Lx&v?L-ZY?q=5g^)A@1sGqbs}7xQn4;vO5qrC
zGh1KfkTvZ4!ew8T)EUjp{V>G%t%S?xEIFrUoBPjI1+vIr7qLYGI@P@j7l>WA2PK89
zw48Sh1H5HJLleHe;Eu9z7P1NLf(b$F!yY-GwF3FYb%WzeBp?&cw%-a1w~jj}E;o{o
zpr`PSW~)WSBnTsXe|M(+g)jmuhacMYFZN+~&!1?|T8tMPow}&~5M^&{8QT>s&gc*6
zGIT#&TWyxcimWQ*rh=dG;62h$1$qLST#OnXBdXKkj>B44M{!;kYGVu%o^?FVb)?e2
z&4ps;yqSQ|2Ux@p8>pM>0{xsTgtQI&==}@%`ey=0r=w*L)Ge>@2F%(9Z?Iwwg}CCT
z<C;NFw2273ut-4lu+Ac<1VIsBm9kM8_-c*6(pRC+{MzY)&?hc^IXk20W6FdXDPG;I
zW{!<nM6~H5LL7u7%Ew%Ii!jTuLtc?jIg`6}lii$2%8dGqe^uebOMCSO&ey0Je4-O@
zg<SMZY%9rLP}Qw=DjQYy@}{dOZ}XPPDm18<8o(52AApIn*pfg=3nT1Qz!ukWB8o%G
z-imO100T{MNvo}k;)#G}v??)2>$ZM$(L9OHBsPEA2S<`b^N8Lp%%V%C3hAc<aF|<0
zE-8v9v&24J5G#A#dGmrFMX~m$DVh==@e7f<#89AMb<W|v<M8W6%vP11gzk|whYM99
z4yn#6i&?KXjAv1foljhHudAxdj*^OCsP(x6HWdx=u$1>^+JHp@8YbS3Vk6xo7RyoN
z!8b$AfhK4XVDr6>P^iy$Fw*r+xm>i$aWd+#+`T2Y`|<i5-<<6u;r=3w4=+J0)sF-w
zNE-}37<6{R5~{L_6IJ3}^S)}z-FNcC+~9o>^ObSh1<qO=`4J4->Y<>gK1MiI57dC0
zcpaYPE<!MokM4MsWnJdBW;`8F^WlQwqeh|H)G&7x9bP{21=@^RY+UoODBs+1;4nS5
zs$29}?BmQ6*RGw?t}PDkO039NwaX@U(uvf9Q-4T|XVS<IK=$v%l6O!Rkr(hi;Szwl
z(EKzkh;+l#p}-+Apdfc*OkKV`u#!V&l21(iyqJ>E7mbJo11_2sgZS5?FG8`YOco^X
zVKTMp(sj*nzD@`rx;b>ic$?N;N~dW`^S!Y7N%xAKk_^B&xIwl_Mvl!e2DYO~V}iNv
zSw3A~UM;9Tn-FQa-QgokyaYUP1rb7Z(3VGo_jeGprO+zQ;tuC!8sWC<jF(4QqfN2}
zV&W5(QXXHF(OB;w4Xrw3>Z=b2f#OQF1?uiNdOf3T``}>za>fh){|@pCB>$g48h?fD
z^WXiQ|Bao=WB8q(1EuNR03n}6A=2uJVXzc4W--Z~{mss$N=maO_a7d!&f9qUK$JLf
z9bim^x@-GMSqcugdlkFeuL{ap_}+Ny&~at2nTm(yXyluB0_$8-DhfC};?*jN9-Ac3
zIXSm<@LAUuiF(r|-xB7sq1Ij?eg>iVk*PVbRZzL#u!^i*ov>=hYF;+Waw8|Z&}mqn
z)F3_8c^Lj{?VeZlA8ViV>XW0MkaAcgHc>K;RRR}zK7`doGw;I7!P6F)Ejg{qhsAvm
zX&-Yy0h-It5Ot1_+KixEOrb&-`6y?Pbg={!4z~#z4lnJndleNw@HgOR=e4`Fbmvj}
zoA+Oz%p4ZT>6H-caP!THLX+9pF34#q|HDZ77u)s0Yq*f*m1-7ffl8xwDD*hb>9T8M
z%xt&qE=qGf9bQ9|0c2Q*NV}%ME1)h>O#-@~7fE%N`&{~ckjsds*(5BdKMwA8;3X{C
zc$%a>SW&=iadlna%;l1}x~wVYO5)?Frui<d>|j<Ki?0<(6M`NtH_V5^Zuse|XDp2$
z2&l-MaL3IiwyId>QkZ;e)v%^)QjlFZQx<&o)`npID{3u3uyf!L_YwjHE`)!UR2yG%
zw~Bx51|(OStl7P{(WqUszY($rY5xwY*ai6YjKmTZe0T7WF!DTiC9=92&sUo-T<hbM
zSrZM^zNMA#OMMQ?(UU#0XNyVe*rrPMaHmrkUQCM{%?Y--cDLRp@-a2^RLjGo-UdQO
zVq>Ho1eoZ3d|_o50B@4liB1VtA(ykR{gczo&&lskpZ9XH>mRiP0Zlx0g6BPg>Wd0G
zs1br?!mPN+HMa$ZQ-ngRJ8LPsgWp6OM7_}KeC?H0a^?Dt*5Oxy?;z*nrdJf!rd}e_
zw<?(kQ-KKk#?wTlPm^cy=0@nclAxSMbdg>Q;Jq<OQUQFIF-`}j?SO{zO)a?yHbyb=
zjBm2#%-oP7p#N3b7-XqH6wN0Fs@3#u+a$F}l?h^;hTjPrULHC#CkxF!dpkPM=s2Io
znm}!7O#@oO<4^?FlY<p;-{|rln?DrV8LLx>|Mx!Oe{P=;041Lz1co)*IZ><P70O|0
zP_(K1E2|xB!mXAE+SjNItO|`TTw~?<;Key_Zq(GX$4jOMEFC3N$w3;yt#^e*-WRb&
zgsTmPXyy94L!Qiofq%8-ItmxZ3Qitx97rD?KD?cnOzNnDCbT_XYJM{sQ%*g*%@gen
zp|F>k@^|A}@wL{saF*rUD6G74{NT>;G#v`wCf9n9{X(`W;5tvp%C`l_cSG+k59B&4
zL+#xR+l*RVy&YU#WtXSwR>!I)q6&TsZmv^jkkqySYcdh-@6KLO&n{#*KGx-G^|`$5
zIJFClk+~oBq^^8HI)G*Qo0WH7cY%efK*mCKe}3ta+wg0XGh#1at{ro#ZF#)U|LGi-
zz?@J-%xqKOA-?d<>U`KAfEs?HXf3L$63bm)T^*r*qLyOTgq1?}gfpOVmU`p?P$CQ|
z-Ud8}EHC#~m`&VS+I_}SmN>XLU|vog`@-9!CfMdl!%T>cir{yU$k&7VlGUW1eb5ru
z?#2}0E8TZk?gdzaaww6>0Puu>DN2OylM7MP6DELm=FrsS!8_sd19>9CsBhg#PiI(S
zdC^CaKX_F#^vaB8gUlXLghKSqSVb9Ql!b{!>`Tg!7(?O(Io0VKNZBIzSSwq@vZLXC
zg#vd?bhk)R@rAe>+@e_?UCvCVOnWDJ%><1_t)NxpQ2>FSb_xKgAcO!=$|}%6OWmRv
z1h3CFeg|>9*No+<Gn<8)P9&D0eOVqfSWKnPc$1PRlN@`%g9@O+l#8h0ibzf%HIz=Q
zxTabZ&Zo(%WvZ5WI(XJZ2-1=8`j~?dGWu4DR&e!@DkW3wPd5y%f`Jz(^AJ;A2(qlf
z>>&d38rf|sAaG!}2`GC69+H3uH&cOlGzn;u-1`n{j03PZ9<SLd42^cW=MO>d^ap#u
zyU=d`^Id=(`%ECeyOYe{ACjU#+4iTyM*i;6^imc5`PJV1f9NVK)kRkJE-p`UBR>$M
zX5`s5(A!jl?sW8bW3)q#bM&22RwsXU&Nx|`H*tPxxx-M;PoDkv!}-%t#>TF1>^bC4
zzsyD<z>DR9Xz)Xz|1q(MI`gu*x-Rwkx*R{PsXZn>(2`rSvOFW}wb;0-pl%zMA?hml
zOUQHpMS+<%QO^nX9<Hg|h;~nQHNa0JMInx2O*VDVs9?7HQCA5h0k(`h5XJLjZh!Wm
zD^2UC@;o*q?F;{W=G}DsSs%CZckkv8PwKmKGk_=cOpH3RN~N#FRj}l^!C~|E7&rL!
z`c2mVGSU3sx%MAWfT-(u*vkrfhC|f6LGRYYU946Aj7f>mrP!u$&X4}bUR>|s=bS^e
zm%LG6115cOCV4I4J1ECVp@Ed*;;&IG{Da!$96UXT55uafTeu=t%Y+e4rFhT76HmFw
zv=6?6Z~%zU-+Fp|pK%LJVg@h)Eoe<;?<wOV()#*C|Lv#X&!}&%5PYEz$9K@Jlt93R
zXaqq_^xMVmzNi?xv9>NJiCs3+gkwx{6Pu1kNa)1bS=E3*>ijtZg(X)HbK;cn84j$R
zT36|`@74^TW~_4ihV)<CrRY#(aG6>O=wZ!Exr#K{o0{BeH2PNkOu3o%mab-$Z`XdK
zl?cre<T%MaPx)ac(P~&vPDs9TsK=kbvZ5}`dd?4L<{oli@uXa}Q`hFnh#wk6zrTxN
zNUO+Mvvz!7OI<{4ep|Fwx}}8!+Qah#OKZJM2krZL%9E_x=_Tw0<;Gw^NpK)1egAri
zyPA)xwR-_%+>v|@)j!*aRGYqdt~g!YCg_+K?}6M!$%|regZsUN*ui(_Cx+@!$f+#I
z*NHA;nKK$1v}RitLW|?Okwim<Q{O@9+5j-}(<3lR%b*U=|GcocNT^s&9{X8Xgd?*m
zz)Zm1YvhVc-ig^D1G@_zM|Vm%lE`O7(}<9dn2LDdQHJov=rz1N5O1sYVE!;V_9c&b
zT{lFcHUW2*2Yy=APS<8u6D^Lt&VqA}!JjoiS0-!h6SjSITVaB7ik@se=dVrCyj4G|
zJXdNQBhF|T<ys$Ln{F+j>7bAdfegvZ!t&%e&gG1=ScU@Se`@EdErsYm3?M96P5?ju
zX`f@RylCMWdw)fN#3nzouJpyFaMuWR%Wu!Ab=4?iEZMPp<cp-6=t{^LGs_W&cvDy?
z&p=qbf_Q_UtR~#K<<_v)6(_CwiVxF%YZhS(SoAmA#YQe?T>myCATg)l>L;uj*O{!k
z!p-xFr<`&%9r)y#j&&(+N>hZK?bSJN>-9-ptDtgdyQ(pAY8Vqfou}d~d>nJgqPe(P
zl^TC*_a%<#FcX@T<SX68Fg9-1R#su!_-cN@Br2HVwfR^((#FN;`sc4=;)6gj<#+Er
z^zC7qj|2GCGfwjk>-WVma%e-tU71(Tep(7&p?T>kE6zvhg*bvd*!8DV-WF50zaICw
zHYBH*MTA1o#e!?d(*=&6I*x2Z!|A+l@OxGm+>P9*Z$XU-*P1>UirSn^my{G8df(d7
zT@NW(c-P1a-QDtV!%H}-^ByaS-hcouf(D22WTwcMjSD}-?D1wm%6L|lB)~<D=*<OE
zkLfBdG9Qw<MI6ver$@Lu;1CRu?jn-hkh?%fC7rHAskd|7oa|eJB&Ygk<U57jENi4%
z9W4T1Jr4YGm18crVEGx(;I;_J(G=P1xVtgYE_9ezWj`Om$7|J-%C~@{d39_Lo*ySw
zasExt;n<YXK<anUYsAyd(vx&-?~tXII~b8QK&t}sVvG8xiz;xmq{K6aKak%^$4;Fx
z9yLpkh9)Jk!)L-(4ehRiIOmXO`Vm1WN_<?wjF8k$^Sx@s@kz(iKxjE0yB@@@Ut>HI
z&0g~$1ta0FU(RfNBK=I>Xp^-zt{V6Tee;1du|OxVm3#q>*=(MIK|4q~UV@%*ttRX}
zt<uw>Hdzm>xSx+bPorlEGBW?V><dZlS!r0^8&#za^g%6ERP}iwQWI93rc)_}D>;GS
zNAoPv*``|hqGCEF?87zPlVqbfE}jRp&X;&j9&z0L>DG|(?N4_~@y+)L63^IDbq1*-
zr&%j$*=3g(*Q?D1H<+TridQeu8}huCnm_k`>);%c9jEQTn{uAuQH#?is%J;b0{wD2
zXQx{nx6muPcVK6vJt8>D3V9@G*{qyunlB$|8-^w&Zvz?*i8_$|yN&pr)c1Xrd(7xj
zw=f$()|X!I=ULeI9dx4o)6~|L1q4wDh`zxkM~E<~k_;2lX#yjepFmZT2RI@*f%UYM
z^b?^D=}^st%p+9RY!yyN@2ll=Ja_ss7F?2-j}0lhK9zj9UWSOd_5M}nU2|k8s;eH7
z_z-|y{r>owKfjX}LQIHXRs%bPQ6?#LP4%z*7e-&J>dnT6z@Y`8gmDnn{TJ38u3t{_
z;6GY%z*u9VEZJWkm?p<Q02>3z{yx=V0fg+qh_7YjhDhys^iR+74{_M7LT5W<>MMDN
z>;LilD2BCJp6Zp8-(gP{P#SRS3JBE?&nd5n;J@nTy`diqnz?Y^`@~k_42l=gUYtx`
z19Vz)oNfhL7QP?UO@hW*AnHEd%Vr&T;-{S`&$4-G_#YO$8>zqfBcNywKKxXf`FAVw
zPg;@foOFbU7ufacF~_>Oq#SbzQ}a-|+3_bg^fE_4A4REIk3Jkhk8HR40K4L72v8zD
zpagcj3$IF4kA*}q>$*gWIgX9x`=(m=v^}s53DdXj96+HL(C*^|E~3fgkA&j~6XVkx
zFv)-mUycN_HKiXEikFuW#QgnRKV-=*<9D*T_O^hvh8F-^rFibx!uIwX@hm{ekxL#}
z9lruuAzdOD`6RIV-~eKVe=?92vP*zTtiXX=jRvHCYXwS2AfIHAr0^-(ILK*>%81Zh
z6`tCfh)MNP{SP`pGlm{KAI=%@Cu&Qi#ZK@8%n%XsZE!hIV(S?KG{G9lYI~~61LWgy
zXXsu<5(Kp9vc%BP@GVYfqzNs0qpO|!o^*ky-fO5)iz&Kkip9Pla~V)t8Lo)E+Vvo>
z8!87XIdR)rgD@5$u)uYMWP@*P(K0<y2W&HnD!ngXs!P%u-hxi>E!<k^EGv|L+_KJU
zV>_8vtRkMc>VUAvuEMNV$v*}l@&L>*aah(!gaE?dKCXF6_=kpl5)*8Da>u>09K!3`
zwQHDl%`jZXnc%COh*u{n^QK6A!x8{L;UJRAa`A~dd(Db7XTRDRmoIbo0%z%K6RNFm
zSb3w-Z*&f`mB_we(s9HF_hAx*{H%m@3|LB4+jnDpE)2>U&<7W%ev($;353NK-gz{+
zA)iTGWxW}8ii)18)|){65fl4vCq9K2zde-Q;>6gig`jM|c&%CB$|DBF3op@<+q{ej
zhX=piN8pZ=_Y~&WQ=rblUQQmbY^cI7;uT&v`5c*O&OVG76$TxpxNwIG`_?rjcQ}&1
z`qTRHMWcPY03Z5Hhq~t%H+r-W&NGbr)vHCFrP#=dP@Kl0no6#t0Cx;JYaBHBxI3NF
zz54QcGxn}zk=rO#X;}>q{Z-rsjU^3-=M~#%<;&-oF@PIPe{(g=3P>A}je88_uPwX?
z$G^Q_oC01r?SRBk73Xcp3G=tBHl<3Et>vrLuP;ZRj%j(s<kE1eFRJy9oGqRJD-uzu
z#YTBOq%|`$_bs0K62LcBC-G=jf8|cz^eRLVV$i33pzFWBl1e_>xW^^E8+ssxnjAt*
zivaUR0gV4LMI*<OX9<COk<O+~St8oIzrKC6`YkniN|uk-MSZ?OHKXKH<0;|%IyJmh
zTwU3lLFL!wuPm=jL`jW%JzsNX2gCQ#BX9Hp4Fc>+e5o&rx0-y!XvgCc;?dvd3a|as
zU4SA@+wPnwm5Jrx6`DQ}6tI>rym;A@Wn=KlIjxqezBYAv$C5y2h0}1bE^5$?Rqkwt
zBg?3?r>U#SSl5#{L7(<JEZ<@}<ldpt+rfG80kwgxZcCzhM_zkY$G6E(S!Nx1=FNOV
zFu_kpkp4S$QXjKFe4L>M`g9cwmgJ5_L<I`*aGhnx$+saNeNZyH6PaiQ*a)ay7~qaH
zz-K*Pgv}YNR5O2aVu$iA1Uyr!9CfL|2CYH$$jqKV!ujY&!f6Yvq)d>z0?zkT9;<kr
zlc@M{@RB{ZzsCn)SnJWhzE)o#KZg`B_90xB9MBCMgbJua5e&9g8NY)nF9VZ2_qQ33
z_-(?6{(NFplHfX!>5l%$bUV4oS5e6sk*gZuO@IY9t@#}kMZeNHwYmNGr)u?cFaoG%
zw6Z`}BxR2viUCnBq$Jlq^FTL}w2vU3`40M~bRbquI)o5uf!`zFHvV+2`iS4hx`IW_
zW%6L=caXv_?-T?13($E58<CFy>I;|VME>*e>Jb9|_nrSe%KynTH=hbaTBpGSSEMk8
zwK+?kw9M}>DLwQ&U(9H~J6(E0y-kgEmuZLY_g&mYO^NI-Ubefdk_1mJgveXLvJwgN
z`~tqzrhQ;KBz^urlk!n+`waiiPJTx%{ZpXS&p6I{)1UD3_o$WmmGtKdqJc)PBWMVy
zLI16JfVAG}GhyF(Kp62Kz7bx*&894vj_PAR4hjp~`Me*p<NbGFdVW=9hQs8g0setX
zD6NH__B#rp>=gkCU63M0)C=mPl%JQ5@f8<RcRdp!JPbu~z>{##N^z_bhpxDQ$Z7A;
z#^YMq#+AwbtI?#*l`1d3n4So%)uv+XV0!gchWe09tUN!C^CfmwwPn6V>I&e(M>{BD
zX|KS&#L1ncm)fCz0aY2{t)@G8#g2dFODWEVT|s|r+jlq6tw`H&z+JOj?CZMTz@31$
z4R1HEqC25gN@t_qj3YU+Y#u{X@w=Z&KyZ~A!<t?s9kzkl*HkK0CtkZ>Xc{_br776T
zz)HdMP0BS!=X$N)W3vi$_Fkjx#9qNpRaCT|ZdJGRyfdULUYp-ZhD5+N2JWIcC5W2Y
z7R23l7DEL=y#H_}?p#s}ItW(mG~Ck8<6K5vzE^Diq0;^(DF0)zV#ze@wBw{y3>NZS
zP92x*L3d)EeCB$p-b<szu6slC10|>)2gltgu{&m;H~Cy5O1NJNY6Q}4`&gdue^58D
zX5GIsQ!|hTHM<fwK{<A7Oj?z-O@}_7)8O%`)-9&^LVyp22ttbU2TDRt*e9k2W+fCu
z)R6<}iGA4)C&Rge-=tfud^KYoS6_l=xZE|;)U1XX2_UAW2m0Kuy)wU_F<DhfhLk`y
z6<i`)WlEYCCAEy<NydZ7v(qA+ISIVo<V?wq_&aswem7=P46V>d`q!2{6RvnD(q4J=
zz8L7yvb1P3TKY)!AB{nBF`O#JIS5ILut4DM2}(`FGCtCVy<HM2vPPSxG_SnB=!&1G
zHrk)$q98g>HF4pCTd@;-U|QdI7{r6@uv=XnH@;Y^OeCGs?h12N$2C$AAGRkl^vP*r
zC-_@Tuh~AybsQG8Fbjax`P4ENt6E10D4@}>lwjHQp)11G4P&)A$r4ku4UJR-i%=7R
z^N3I30&Q&IreL>qhNZZtDfKY<Eva&hW0_;h9V?Tg3?1%*I>`RpqU^G1R(f;eW5M&^
zK{mdwYSn;y;ia>diZI{I(W1zpGq{`4m?dC`#r58`W|mjHyxjQ{Z?oHdC>jaIt_n0g
zmCT9CaEN;<6=RoTQ5leWMDggA!@+Qn0p~{}`S&DtJXPfkjgLtlk-gcIeARy@TF7Nt
ztvY&NDw5MoaK-<V`AG)&$5w;G`Oz`7=JF=0VWP4(B?YJ%MQX?2#Vo$xbS}-eghyie
zTOq-_3Zew&n3;7tM@(g9K+?nFapi|V!A+JoJZW(OcRXzAgXme6b$(3Eby0r>K`sVM
zQ*!S!0T?qeQv-!}kIl)<aKJn(m9H#{{qppYCw(2y{qjpS6d7;iY^b=E#LsSh&BPac
zGHnlY@-ah9A0F9l^?0F-s(+HB$h*buT5U7ySuFPCWt*16BwY2@;DurCx<0pqx9)E?
zwy)1sh7a#~(CmDrIh-Cjf-T<&b1U4QRd;Tbm)!Oy7fmF-SH-wWW>1YztS-cbb~Fg3
zZy;J}!(H_Uo`H$2au0C6GRsYd%ujRtt#hGg$4pJe`4X7An1jy=c3iqIp&Jku-g4Q_
z3JNdRCngX)+o#T2_lpR+smqCQWGsrC!}o#GHk*~YEl1Oo)<f7sL8B)XMf}o&eNoPR
zugq`2istdKz_lVtGm{D-DX4|1D0Ak?xF=<n7s6DSZ7-dnd(Ovx$(}OX7M_Wve<25G
z&ejD6A(^lf&%T4S?o~;f=ENTRQkBN>@%GnOX-6qHqCV&MlOge`CE7}o?lidtW_zo-
zIpO<_dhJM2C+Sv^5XkAXZK$h&16dFCgu~9TjAdZS_+nh|$KlssVK!=6ijT|hZeG}I
zPBZgOqnvY|MttDL)B6MYYBKcZu#Y0NKoPg>ttdbYA<uLfh2C+wr7nKib?(|!_!!Nd
z6$;fs6fe<ix`z##XjcgshcC#imF6%(>6CN_xNB-ExT{|>Mr6MZpa|%{!J0ck636?J
zE8jy|EE)TR-JLYft*)2M4Wv7)WvCe5j=NQ#te?Sf?)3>jnG+7n-ewcf^(UTt8%s#?
z?M`XKP3<);vsquRxVj~Ha%)jzt9DQ-%M1Y$CZ~vX$Gd$PTO6(f2Nk>JtvteX-T<;s
z8#Nw9sXfyt2`y1CoGLE2B-O%>x8oq;m?}S%AORdAp$BdTaGmv~z1!+r$(UDVU+6kE
zSXs|IfiG~Mn&vRR`eG|ERDlUD(2j{q%_F9M+NTcMrn^4Zxo$Xcp)PK4jqad2O8xFl
z@Rjq=p>7Lx-i|IyDouk+VjqQU>NkwK;Y!S><J&ZjeDRoPwelXmE7aDt?^70jXcir_
zX&Hx3GCUXqG=xH{vtuA&Gd#c%%?+*O;`BL2^u`HUTD#^inKC7JkADN%#-{78KL@ei
zkUO=#5UZ!!w<0H>jd4gc2(L=W!Ki!~8y#&bRXXW-EiWkT*gYALsNr%()|>e&yf&k6
zj0AN709pCvKV$6)#1Vl6m8ZH5yh8<as{80Iw0@s6El|Gw4KQBN`H%9=A9#SX)_=FF
z%s;&MD~x>WXO_(WAD=1iYHx>cX9WL(AP1%VLLIDRYTPg1esBL1=;!iZ{d`0qa@%^@
z$c<~8g=-rP05nXqf0pC8$>#QVQ#+<V6k>{pe^TUlL(!)Wh}GwROQzFHd@kwqD;uv6
z#sQyECOTo|HCs>@>!BPhcf-_O0iDU1viJ8hPq!7WsNcVquPs>)`xT1~Z(q$QlU`|_
zDWxrRWw*a25n`g`n<03Hf|YeOYMV){PAhF&7WQAObwIzZvk=Q$r%oZY2w-Ev$;4;1
z1SM6xeCyP!oc==7uT480&%Acb<YLt4+aH&DrFAad8;hl}XnnA_%f+N{9?ow}*oj5t
zEX9?<g};OJBgTJ3$(Gf^2|=9>nAfp})^|D|^G!m9N{v4!7(JvhEUxuJ76VldnD%V>
z%GPjqzD`VHpy3K?VI9~G`c70SbZ4NP`nN{psBtAP>mvQ-p5R27QdzVFP2+b`gM=kz
zeRaQ#Xg!8yhN#-7jfve<G80rgIwl%$?|LLnRZin^1$xc*#P-%ZQw;7&-dqe&-DR}m
z%nPTrEWb(SO;JbN=<D2KN@c1OUGWzJaD%lnlT)%|!&3rVm0K5uG~^D2x;KUn_vPhf
zWK&HR$%6fr+NYQH>59-v`HP~Tgk~G#W;#t?zKy_1Rbkl=`=3NSX@wd!PYM%`?S887
ziChNc7RQ&)jsjE<Da}KeiC3y}fEO#j*kot9+R-Op`n}46O|MM@Vt>Qy!4h1-;8P^p
znI}?6)cMjHHvSQ(M?iTA9!LJ_CCM0~`QnK;jIp#;&_*QAr1xzS$l15c*JwKfQKGrL
z8&HMbAD+S=IBckn#Llx(uhLG#n4mlRhLL%WZ1EUj%YHh;vD+wLK9l&zHd?opb_}^=
zV}iB6p73205SZcj+qR(Iu?DbfNG64H-y*srWO)V6qFXpJ*(8Pk`q?|{mu?8&3L|_p
z2Z0uSC~@nwmOhv9JMXcy1+=NBIvy~}Xqa{uoL$ulKJxUMK>5Ra`hH?GS(r!(00E5#
zz&LS9Q?o@Tm5kM%d!}QDhg;>_77a{<UH@%?Q4}3XidYVv0W&Odf(caR*{Z(xF{>?A
zXS-ZIq96`kx$#ncL6;oR--2x-SH==TsjZo_pA<W=36~I#5}R9g*{YHW%5DW_e8c?A
z$8$RhR2_H=OanHuZbg4qQY&20Sg}0;s$^ce3iBS_pQN2@*xr?ws(Y%M;DcHO#{<Qf
zg0U4xfth5D8S{P#3kdEwAuxDN(n%H<vS?VPRd`qMtTx~P)URn+NjIzJ&-*$c%A_}T
zq?3_?I-pqLs?S{lUp1~?^H5q9+3y1tj-Q(#9^9)6X&a`+o*C`R<V~oIntORvBErVi
zC89XB{y4Dl^g7qofHJy=ICx6(DW0pnflPu(d~4i6@0Q5=CK<cJ$5bs8hRKt2T83Ze
zwj36vBAtZxgOSZ=sW+H*uF90vS23MPo_!2$Q8?}J+Q+z#AQCF4hbv()gZuTg%0z@8
zMq1-E>)ExW&OL++t(p$A>hgJUZ_I7-C~-?u#6Ca0id8tXs6Gw$e?}9dIJEuwYc@e@
zu5)vhZFLrdRDTa?vu-J?xpymk+S;40r?xgjW@BpgdI{-x`&N8vw@D?~GwA?1wRbgs
z`RtEzo3PeL61bYQLmJPn`B>wG_}eGhhAN>@I;kFunvkBmas3{e>_E!ja+Y=Iz7=2i
zq;}=w$JQr&j1?{*D-Y91hLb`Bl~~{DWdSYK+XEGcR^CuEfwx;DYg}$>pX^j0Qz~DE
ze?ECDz>e2~CutQCgPz~>8Xd2szOo^-hVGUH$4*W73K#nND;x9|e@)LGoe*1fNLN3x
zZ>!k8Y?z<*&g@R7m|jTq62LvU==W#REI%>n|F1YamkiN-2MsEbao<5#8~<fJmj6hn
z<zM;x{|XbCpJ9-z&=4Zfu;6L@1UUFnl>=$X@tYisKbFSgAFE^@rQvT}%%URuM9-kD
zj;&SGCba&*JV?{Q`$RY6bB3EX_Z2^ZHe_BH{<fIri<>Re3A`;?(Y@I2*bq}n`GNl{
zs{)Y`w7h<5(KMwfXDL3>HhD5PCxI^f+s?wKX#&fm0RBr?YT>eKpQndKUOhVt=laN&
zP0q$fQV|i&Pa}YeBUd~?zc^Ni21~6^%GuMBT2j=%%~d|@E=KgKY6GfN$HByHE;}@M
z;Sx~GoIaXLdKytpe@UdN@r2LWX(x!zkdO5AA!YM^PL_HDs=wm1P;OlpvjQ|Dz^9ad
zjz1=`)^{3wa>?>6jMftFsa}WgQsH)D_CA;LIgQ)zt;Wj73+Js_nXFeZ^M^-)aw`c0
zfUP5ScAQ+vet=9CNF4z3yf-nOpdLQdFr{xs9ETfe)(l8MXzlA1lAeGn6TdcV2hiBO
za&CDPl*Y(uP0F&Khw172$l_riIqCb>wvwqoPO|q&)OT3d)RqoZgnGmIqs7>y-EN=w
zlCAHjqDaXh0Q=LEZ^LTk&#XA2xgG$AWHSoOCb-AHKuSNvby&>CSXTE7_3ur5I@R_#
zspGyHzudRED)OHW9hWAn2lxb%DTnxVrgi5TeXo#@iOV_d62+XG@f&_1*<9L&G7vR?
z8dIk9OCRtPYoRxF%t^fOs$brVo>-X?ZRmZi!7mWTD)*5KF*#20)hqWf$s7P^MgWL~
z=VuUvC57Leo!xW-WHU(z_}hk6=`NT)igTHyyQS$_9@e&dOY!o%;%hdWAYVWe@-g*|
zxpEj8cs=w9v5leYKJ3_o9%i%E4fo}6VQN&Uebef@9K!R**Wg8va%k$-=>ecJengur
zuR2=UvrhvQs)-c*Kn9lyHA{neZ;LwvLa}0-ZLi^NGfgrl$-*ygwU%_`+(HwcXWumQ
zL)~Xrw`5PhnEDXk^J(8X(>u=;u6Mh#v7tdrV`+&)hB3~&a-s}7B5bNt`(b{LJyH(n
zmHp{yof?Z9Q_{A_y9S>DTOMOVcu6RX#Fb`AEbhHrjdqs!%!7I*?6LB$c)RWHHF!sn
zY!A(X=_|GtX;a};*HdLFmbvGLOY5eDKXHzJ;zF{6%%z^`M7(MI-D@BD`zz3-FWOIh
z()PM7sx9=q{gM2<k03YDQ@sL#KYN?<P6N17-#>d|zaUfp<h;MYN`D&k-=Fc`4`k}^
zCQ#miR4pJN#{vEh0s#!vnZY0E*8dGx@w2Mz;7=Di)J?h*XrK*9ph-(txilY!d)Hp9
zaI-!LY&+LEGLLyK2DY2p<EsK|Y&-r_H~tSUVK<Ebv1sXE`fZo@FDn=RRodTQ`QHBm
zFM-DpdUtjUE}+dGAW~BNpscL2x5zGK4%?*42zAa*ev8is@x**QQjggLpCVeK+onKn
zC8ws;C8SKN10H-x(R4W)dcG?C-1`g9N36sA#7BCm-a$@R&{;P<EeZ0G&r`&W3CY|r
zK{s!<!#C8~VApbnF*X;98KZXNa~uPkAwl~_tvwWrmO_1J=H$h7P1tylH>2OHvq_z#
zxWpY5cr^0qruI=ES%PPi$QgJ{t4swt<-!0mTz78T#1U<T#(e2{LTh@XbVHXWI|9RX
zX;<!i9_TVidPfw`Sv1Y{f|+4CZa2{*H*a!$>CB>M>D%z2ha;s|IdwGJcrad6Drxi?
ztH145MS>J$`pIl?Sv`VqS^uG5E9JS>0iOZ2FK4b>n}_$ymR&4XF?n7z_N&57H&AJc
zq+FVmBf`GfUlW?E%yvxlsCTg15VYFRl&vXNM(uFl`+V^-YY^xI&x;t%%ALB#XjI9!
z%4(gGWfIB;#YBL$tZJGK+^xrNs`kl&;l@Zj{N02jv*v7hZYpf&cIb^0VHa!O2!M|b
zG8ZKw!Z<1G%ZEiS5UK<DD*DnQmi1~P4Xf=}9z<!X@_gYsoq3r@{652RrkQg1TapT~
z55H`TA5JD*Qb7f*S>psata?lbDn1vshIhV#&yGD1^utcG2GxsF#3tx}n3I5jmL^$<
zw0N;O$O%Ul8A!nPn<0`qcG*`x3whtFmKHo3w8;y1<_+)-Vbh-Q3;B|NWe1h=tOTL?
zR7BR1d8(iiz-`G}G%}JRvn0SF&u9tYKm_(e9}<$bgrUb6_8jQ8>-j+;^2s_@n+754
zf(z-c(H1IV%Jxf~*LvL&N*<twx0bIwvI_CeTN9oMFPe!Rz34>&zFL3jYRvOiHP#F0
zE<i=P!omngNm3)A1AK&vg~P?*c8a#HSv}Tyj^z$@F%}kD99fI!TjR#F9BP%+tTAcN
zvo=%DOA17CET(d7Z{wdKQEUzPp1?<s8mqQC+-CACp8Ag_N<_@+3{c&BrIF`F7akGT
z+<(f#;5&#H)-XARpD8Q%TqRZ_>h|}!%m7a&@GFv+`};9Cw3SUHs92)bCzR-xe2YD|
zB1TY3<C2w5S>d%2@RhCN5G4Jl&Gwmf45bo*B@ii81b0=$?*<506f8GDlxzCs_V`jJ
zidHX%*!OmR2;#F^I0>E0Gvtks-$-s4JODa+knmz;s>QfFbboAXY_q&hnNdgW!M~t+
za?FC?uoPWUFRy(ESv}}BHljN>ywPtE$BJ`58GY$JckCO#bLSZpnO+dg$m*~##0s4k
zAgaoV6d)g8KDYr#?>G?YVQKK0mUy;^Y;RGNHhu6JGp(awAsFNHGhe=Fb_8rx$IM1H
z3Zj;Oe}v~ZB7Z|kA#?ws;_wH_gHwjT2{!zDkFfw=5(Y1qkURg$sPd0{3_qfG9tOY^
zQj<h|eO>8dV~hQv61&o;&Rowla~E7`_rFKMlLBI-82Y6$mm5HOkVYKSjy`Cvn@lb-
zQRVhubL+}myJP}7nr_2G)qJJh{PePUY7-<hQM;|@xH?YjRF`f=KY2)~P|We3pvsJ$
zxZr(p+vB`0k6prO&ZW`c`K|GpmG@f+Er#H7BN12=qVNplfPN=F#?=9~gOKg)Np)pp
zuXcjf&#krH7>f{31T9}K)@<Wc08|?ES|@4x6q<KRi0ZKU(gW4x@|8)8EHwMhs_f>s
z)ixFBhjRBj(n*GR?GQ;o@K=bvO@^wd=<uy2!fs^gwU+C*^HvZc<Eoy7XOFC-dL+Ck
z28XpnkOJB8F~!<G0Ks0h7|(n1gI6TyYD96P7=sF{KHl3<0(=g6><P}hgsyc4aY`L0
zU}88qqo!p#KGA&XnzCx(vu&#{`4p_Y&)1EF$h`1do91<W_X&o<Ql$<@tEJg|o`wlM
zIVpN0!i4kMSgg(`L#u1&?oapk6>M>tad0rT6Jay_&*dIwF4Z*#l3cM3o*Mo4I`XC*
z>KYnanC%%$9g@!%T053@oo<T3RKw)-G8Q*^q)i+AE6PkYpG9@M1a&)cPKn(=J|{7m
zl$thKD;EOOYN;F<$fymc#KpgqGs4j}P~Rqayex#qy$%fx7uOz4PnkRSS(gK+BAD-)
z%VRyqnNj!Dss^x52>3rIBIbdu^ubb%Cw;)jmKS$ThjS60oXOb>JG{r*(mnY+Zu^hw
zzzsd@E?m7-I<3oQB(pEj?xRfTG6pn#CcIn0%v&7PO;3eWoy*O-Gl=mAHz(9L?Z#<C
zz3*r|mb$t|x9L*Dx*|lD@KLC%?&l9}w`7OA2LwyZTUd^V4#!ShG`{;CG|2-0W<dL{
z>>BHZA+u6x!0{W>yOLK7#zwXnqqu#p4<OCt%j|nCR0YeFc|W@QTSxWAa2uR!dU*X@
zw<<vz7nh2k!G4<*WRR+K?slSkXQ<0J_TJepg8eZMzhkD{`*?l|b-68Z#4;L&%&f>j
zrkWD7XCM?rX+yM}?Q}PDN0GDO{i?{@1jg#VyTZb{RO1^aAN$H+hHY&pPf4G8(kR_J
zMGt*OH|*X+MT}`m;bD8!XcujSo(X9a2+P~Ztqr!@NSZtUh@V&GM((&-ayJt;h3Bvk
zDS({{lZp4U8zKofQTb<wYg#l6A~op>cUoU%DIKd-nwQzT`I!SZQr7m&>7!8^NtD=t
zg@mTk7y3NJc3bvu-ma}S`4;7D<!V1^THks@)mEM1+GI9L@79p3+tjQ_3Gv!f%~+tA
zOaVr9+k5C6iO;}-F#OL+)ZVz|gO^V-$3DiW3|~)J3-dQThXr9LRowJJq*inaGzDK0
zs=)8~bTaAi*Nnr=r}x*4qkVqNH|pZV>%1m`;9KRS>_O&SC(~2$@|*zmnBwPTed*r{
z#DBR0_Fq#VE)mDC`>*;q;i)z7t6MQjwgqrI_T%5GFQ}DvIi3!QrlAYC{`r!8_*2^_
z5L)oJcp{kmaTu}BeVMY|=Mw&G>%Ma(?(??AGuwgfE+H$-jA5WfZOYn)SwaSLypW?+
z#N?#0Q07KC{rSs4qO5hm{wW_=MJTpX>7gl{gGkx@g&1>FtyAfC!#MLV8^<qA><$J~
z89z3b^xpcZyHg?$fFgwo;O*1ohTH|iDd*u!QoF-d-mzL=;&xAu(<oXovTRiOsewLp
zl-&BqYLU(BCRsb(56{;QhcKbePuN>>5QI~+iH^_)BYhNs`K4D@p%d@Bsy?%}Y8@53
zs6-DdtMQPFH}0EA18IpLifx&bgXno6rJMt|o7WhUm~dQjx(&#|h&9BgK0wTMTpxQo
zGz;HlZZppHF+xk|@MIr)ji0R~4M1K(3LdsdvD^a+7AMI^;4FH0sp<za^e2hcYN0wc
z4SlwL!;_1=?lpA*gh3_O`O>Iz%FittCjVc1-yPM|nzxGw6qP1Lst^?v1p%c>jg2M;
zBcODMij)uoA{_!z5Rk4Qa0DdMrABI~iF5%0X+c6w=slqZ2=VSS^UZuS=X_;m?z(ra
z``z<b*4lfsdG}j?<@r6&^FyzeE6^c(fWkfa)pn(rP*qJWj!NIFXT1J)$sO+Z#4qca
z!SQgf!XPWB;d?wM8%9#EU|ha2j?f_2srR(RwFwoPrej%g!&^Pv&T3hvv8SchSX4~f
zPs!eYr=pX8qxQlX^Wcg57ZrCI{fJ$gb<JA;$mwNCs;{))LeCH$kN0$+i4C{ZUTa%1
zXig3bs#7ie-iqg{z;kxFVVs846eetJUKwVWo*Q*jFLY2f$BjkKJ{P(4+1@9<c?W+&
z>AR;D<d{$H?cVa95XR}UTKrJCLA0t}-D`8~aOKyCmY~))pCih1cDS~Yt(kiS#S}x7
z9<FbVys_;1z#sk!dvZb~&(u2T=og8{rwB@CXYSmMg`Ha*03d-XJ96qvcaWvQNxA@x
z7)W_ZQ`fN_6Kr&}u<pJlK-EmqsvzZ1kRX>qIXOk&Ld2z_?o^6Wxe9xbbf!lCt%)jc
z<Fk&BjtcGd<u9n&6o+LGFH~Of<u`qrUB~SrQ>u73UtV95iBrGJvbd(qBsHt1lChDP
zYJEl!7cN)JI0dAJn+N`*{IK*Fl#?9Sl8oCVSa?3!u<ext#k8DQ)E#YyT`}`2wY0YE
znm%$`<0Dq=!3~{*kn5w@wCc5m=fCaemKg>twgu5`sUK)v6alhBJPE={%O~&eADPtU
z&n0z^$;>6ZscbrK)53duj&=2j=%LfEIRGn+m>UUpa-|ub>Gizev7jM<fWrdvVJiH_
zx{@O<>~Hwe6g`kE=ggVQ^IDHe66Z5vIyM+up5*Sr!a!1?wf_&MR48MFhSsD>k}%;F
zJ;RQ-DF?VR3{BE9hOnu@??njLm`Qv3sRNt19-=dvE5%)d{A}7CNdnPK(TCy<X~8P5
z80TT778b+M`5`r>oUCnXX&c|0bM*<OiqtdeDKiP%>8C5DiSr+g^lyW^;1eBynAQ)b
z&Z4D4==uZ|yNj{}@jKgc4<hv`J7mcaq-M5Hr#0=q^t{nRvUilnKFRykgbSo{q(OB0
z(aH+{;gGTm_QsEoj@%F01o-#sxQeZPM#vM}x?z~?l7TuUQh7*Wp-pCWRhUurc>evT
zt8e(e2uantzlhi?jp$uE@U_}w{8FRuO;>_Itl&%Q8i8QW#dxF_Eyb%Q;7civ5)%)D
zd-R?Rg(ccRkj`^6mooI)g(lVJgid~?aW$#UHFFKCH0~Wq$@js0_FGN*a5H<12xPKv
z!rWw<O&>{OJqKj01tZa0B}|-J7FUw@gd@VWZ}SEd^#cU}C@h+^fpc6DH*&#u0BC#!
zG`4fj5I*S|Wp|L~@gBtB;jn5MKi@kC*pD^}7Ro1HD82l4=ZRh1_|bc;m)~-oLGm0!
zJ|1}HlfvLvX~l%8wS|_ig$v6p3&f$h&6oK72=8;W9WJ@t6K|6pIrdn9i|MQO@GJJh
zj_k1BP1Ekox1(s&iLNZX1Q6~H!(EApQWbE4CgKxDY|vkT>HPK6AFO!4a}j^~ud-c#
zbMK#G#rv1Rp4E^@5O)!dk0!j)?5%9#taGtw<w5z21ok1`?$<X$zK3Yb-+o?A(t%uU
zSo~Y>D0?>2anc3?sH$_>&~mS*>LLQNhhdO>jS~H{nY_*sJyNnOUr(B^7+<@Zm#~t%
z#jveE9<!H#R3}gHeuvcn5vOJwP!o>xQ$I;S{BA$}#ts2@m)#!rIGHfg@tu!HPdc=F
z^=N?~YYT>%hSA3kdO(C|DD%EmIWxo19mwG6G$Gj(kq2hFQueBX*St@~NY56j+@*$v
zAmyfLI5Ak7^dJsoB^y_BOY3T{-0JG)u3;+x^(nQqd0Af_%7&79tUI><*;M{BA742@
zVI)C)`-4eD4k6OAS2Wm-P6s&WhBO{#v}9XD7x^J>islj0R*%x{V=jD&J9EBI;jL3U
z*4|f>2a7RSz!;2V`m6IgQi?{fNl7jk+p#AB_W87>CO<yE+C|&beVd!TlLqhgi!qO0
znK$lN>MPHRSymF7ln)Wz`m!jPC`FLy{OctxVZ!FxiTRC0^9l~BRrLbj<0YnS5|8s5
zSWbx@Ra!TTpmR)tczw8<CQi5@Gz`-LZO1XBN$~<v`Tfc661zv0Yw$;hvzvLs#h5C~
zG_*`=u)loC9)ypA9Z?+LVU>Z>gyjl|;1VyKm@ztL>(iqft6!CN4;vwOMJBOtAb;hS
zP0xu-c9Ieg5;L}p$2>Or=|tR+rAn`I`OGj7!U58N`@@S#7L2n<zY0L_^JZfpcffgr
zrGWcWHjsNUh`g+0)0c-8!^>BBN2PD6is-d>c>$z{NhB_oMWSgE01R<T=FBYsC*)cz
z`b8=C+iK0&e9HVJFZwm}IM*dszun3;5~u|fRpZ)X=saQFry+?Tcea&(DgkN$J<n$&
z5RY5cRe7M5jc32U+BsIkFXb@m%^YFh7NMbE0LSrCV-iRhUceqYvDGE5-@moSwq(mw
zouXI&LiXH70*2+`lbn#ODo@{YN6j8y(XT$59Jn5lxPqzCFR#B}xU(tcr)rfun`}5y
zh650zd7E-E3?RVX$VQH3hLtT$IDMij=bew+VFdP+bZx^NBBje$IUim*n7k>!)pKg?
zCEkfcr3EZRM@@2lQp%;DO(g7qh|$XPYbvK<gC<UGakl3kv9Y(u9Bnb&H}G*arm{91
zSOy#?kawEZBr5l|lU!kFHF%0@9O>+}L+$aMSEf#_wfXJVwm~_OerQ%CH^ymPbdWU)
z>3BXbn-<C!!WhiCS&UEa|2bd&a6xuW+*9s*sqAiS@5W@dpvv_<^yHoOPn-Q%=&g|?
zh5_7nbWa)pRFtEGW$V~T9H<t?X#nc>9HU|6ge3+h6~XT|UHVjgFbq-YbUbHq?2wOK
zj9=P>EiH`XdVJ9|*`<tZ1JWOGm5305t1r<$B>SFoFn`1S3%}#zrytl$Oqs7LO(q-G
z=Qf}Pk&r(6&B<tYn@?1x@~|68+LX<`Q~|#yrrcKNlh@c?D#@y*fwJsw<mpd#?~aes
zcqw>mK%u@)Rzo*eUpYM!snoaWJwPv$(0>-+Id>!8j{P%BH>Y01yz_K_;A#(0MDn(1
zvD(;w@jy=aBUWaxe76PKZ%QD_S$Q)=4$_eqJt>_{0Nz{&ljU>8vr|>>blISiUgkMF
z<8>C8vMVvu?KC^*ox^|zCo9TXf`;XKv0AiQf6AOzv1Umyl(ZEN9&Tgd{>n3b^seqx
zWEbYB!}=GHx+60XuC+(A($FN!;oN8>xWE~kCeh32$%Mmmx_dITg`<gv&DXhrQaOcV
zF@p)egq?}IJm%PhVx>j&G@;6Fab!amcKpmjTx2ko{<_fR-R}3trMizZR<J?Jm7Ld4
zs+3D=t{^l<5{8Q6Y$|GM%Dm65UC2dvulGX@13U=b3;xl$z&^MrrkK*OjgzMw^hBnQ
zN_;yrcpAP*tWEJaL~jV0p7A+Fh{i|L)jb_^Y??3&?v~C;N4gt**2Evzrn-JLy8J{v
zRu+=*EKp}E7%Jmqb{mhQDFJZ@bBaTK4*kYArD&-&BV0|dWK8R^6^wY=&Rv_Ej>Fjs
z`Q1&O30dy$p?80N;f^$<<hIEBpj(ysaiGD+eSxBjDq@!@brX&<V%dQ5@?lEE_taIN
zuio3!hZ@J8A3Bd2yVJE$TrX^0yfmm{KiC%7U-TZkGHAO1or8~L;yTur5n%IAgUKsT
zToG%+-R_8ImFtC}LeE)RX<%ul4zIMQO@yMMmvkknX*#8el!lfqnVjHvwwx__m`7{6
z8*JF$_4Rf4dSqzHlZ2KAfuF9D+q4X#qAQlfgs84`i6P=3c3T#cjdHNob)9qe!u{;+
zPl~ajoO(a6(e;uVrfWN#+p+DRjhP0vj*a;>*E{Ht=*0^eS${gE^We!dVWA#2H_ZlR
zGfVmv0hYGn;+(5`kF+nXwtc%GNhzpk0=Vr_Zd5=dBN{0Ra}&)Tug<dmeqCeudz`+A
zRN`kz7rpL7SG#VF7$H#TNgKt~zQi_75;!7Fl`7#ScDjzj`JS|P$R@8!g@}1oS8qB#
z^w7R4xo5eRkA<lvaKOT-^x6YEUK;Cm8qJO>(CBR|-j8%lZ^*H<>2Z9zV_B5HD<^VW
ze#ZVm{sK!}y5ww>ORGXCU6VGBJIiSQwgo;yYq~}?;VK*dK1K^%TlOkFeO5>KdF*nW
z?s{BdJkRIiBU9x*Yhg&0fI^zxXNEAIf-BoDVkUdetY&ho#y)+0iOW6DkfZT-`%EV{
z7n^JJyL&aEr&CQ!4c_UfD<{xX<M0j%dluF$DxF>D9O4naAKOQ22Ca${w*aX0BqSmc
zd_u#BY@=epzhmi=AR0v=R%8=<I5xd6`*zIVvvau;Boxe?CS6!DxQt0$DRLpz^2aGE
zM0)cGX5D2IKin<&la@)(**bg^AU<2Hb&>wiKFRpo*+Jf)4P+dMEThMXCt;@*bg1g$
zC|{fFlfiu-k=!(Wa!=UaP(%fXdz;pILQ%12WWg(s5n4I=JXKVj?P=?HpDO>2d(r8=
zK0**{jQ6=x{aX59(kP&PgYehZ99yAjqA=^Y*hWmGtqd&>o<WQ2Z>X_BHyt=ASCtK*
zp^gItp2A0QR#)INIphx~t-|cb<>du|!nC7-T0T&6YeJTUTICyia>1CUr%;`2a+b-2
z)h)fIay$6SOyHyE$R<KvSVO5LC~{n2K#R`@)g~rTI#}#jO$~if<fImwl%dN!QUV*4
z%`i7!TM^R(%6`FhW71k%hg%?F6C7J|>@?7}y#a&`AU{cVsCC!M&&w^<cPjf1_3Z(7
z{?hb?x+)V@a)(yLGYD-*t|9g?ia7n~g|6w8{dljfKEeiMsFoJJZvy+l#OVQH%^n*H
zLO0{21|FC7ujEkuuIIcm={iqZ6P^eh)yZ4`{MqnqTm0*`a+)I*z|l<b({GXyY)!Uj
z`YU9tnsyurs|4cSvG>pQzIO??F1McU^Y#A%Qd66XScm*XyJ=fSK`$?{@14(NuYtZ=
zo5RN#=$|YShl?8CpJ54HG0+!0{OOpdXp0w$<H;5a726uidAnvO6tgAsbkR$AK=|S1
z6z@-E;SND$^T(CG{I}d@DrXNLd)y-BH<40?Qg7{Cokaz<JLSVVrU(<UK*{rk>_sdN
z*90#ZhRTdsjx~A7NH{t2uS#6e`@pZmmU@{5nVY8Vo79Mo6qBJ9xn}sp{M*<%E2s4s
z2Lzx^%x{-O=s#)s80~#h=zg`A&6N0==Up^$N_-2NYYTJIS$Hl6X;@@sy(D0EVgXBq
zX&I8$bofcp^;yp@z04E2nL&|Hs}l&C2JliL4J5=(5+sc7MuN*&F8ZSdozw`yA7OqH
zPc_)P)UpQ*j@uk$wu9y!cT%z)$3Ik^DTv<7IuNlz6{!_@oTGHM<C7s90qUC@duhLV
z27h4TEE9hRi}**$S<n7nTPEKOwAe-{{>SoE|1cIVF7vN3!jD~~C5vodVXMjRmRE*1
zJWuU0(U`yAJ(6%T0oX0N(DdyAFvZ>tK!;rb5=ag`41Z@3O&)0I)Z0H%w>2s_0;%W<
zF1nC4irN7t2?3KTjTpvYF9y?Q{)DliL;==$58@_Jvug&TKZ~X50AfXRl>j0xSUQ%W
z`X`(XU0fNpF#?qIHYBwDU^<n_;7$SWu6zWfNT&dHzCQtN{%FnroWA*Oh64z2{@v6{
z0N_a9Rtg!yG@R3PouiHL&QDguS+)<{I?7(7xc}gI`;XI%+D!k9(raBtTPpz%#8VwL
zRayi-b$YsH(VWMZM;c<~pH+64D)xxH`;@8q-K=2ft%c9|A+qlBOb*R<-52p-uujR<
z$>l-YyF+k$HYvyGFW(=gEnZ+&YXg8y(e&#%IR#|b2(|5SvTFR-eNbpzpQg&4N8mx7
z`0L_#_$amxkD7T{9)t*k+3EV2#k)&>k`Ys33?7-Ec9u<Si8p$!7B81xdTrDEvhv$J
z$-?_YsjgJejbLlmW+A#em6^2X17xwq2#W9}eLKd$V4cd$*zn@hS7XQ#502S4v%kDA
zn!fi`<*+<r)C2f7by28cpC{Zx1MEq{u#J>IT~PLWYcF^1bak@ti-yse_$Y*6BSqd<
z)T<z0qecrqG#Ah<5FXc@e=1oCzt!KpKkwi57cS>HaCuT>h1?<JF>3BB#Q?~@h;je2
zu#+dp8~kwhHM#KnuP1o|?ZpSoGOd8^@=xNm*1B!3EY2t%trN7|CN%}3ZvLWasE7kD
zAu~8|QfK>~mr>i@!hmRxP6ux>UlP!LVw@C}APoq4lh%WxqV@WP3F=jwyB?9dPkt_a
z^(`Mppv0Nuyh<M35DnH4>udt{%8a4dh~<d}+=YIPo22=N)$vU`Ppy<@rWL~0C7!pa
zH=F9jXzxzn$QB9wtYD72OzCOHN8k<*dWia8qvVFp_@}T%T9i)!g(lLH@xFY0tXvO|
z^u#HYM-NM~I?a~gWRa{?*Qh$1sc86(!J<nhmc1S}T7rfLU8(U0F?;@zk-iQg-c5Ye
z+h;e#FMU?0zkCWC|G`8^cHIF~2{l>y3MXu|{4jlMGyO(r9=y{?3N!?CLT<&Q$o(_!
z66IrQ@*0GIC@Pi}5Arp6+`jEAdnUo=R)?~}<;M@Moi?2<qi&AEa58-<O+X<~5Yhln
zTIz!9F~);N1Srifh0@n&-r5T`TV4#(%eXa~n(CW?Sw@qsQ0(;kmA&P$hmcw+Ed;-x
zExi&*AEC=?NJ5$3x<>YcupXIeE1&3N-HlAwR>lAUvB1+v-JuE_c*_#0OW5C)ZBY!=
zTj*i6yE>H!?@IF=Y>u;P?u{M2R=`uR^7Un{=m$F+{Z?mC2ZLdTMJpxs%n4@<X;^R9
z>=)zNpsQ<Hl|>9ygB%}JBjG8{-G(|Bv8E`3c(m8nwBQtXo$Z4dGL-Ce_U=0C@rai1
z1uHkWUv0Xw`e1*i&V_iOxt;nex)buP&W)AFis-bQdhBA#RKOXc$8zHOv{^S!>fr5c
zN{gC6IM)lD82#sJDEs$$;n1Wh`-`G=iM0{dmewK$am>D@(yfndM(hwxFq2J4(54_=
zjwA}wkA?IrCX`jg@Y-d3K3)1ED7uryN5qa*T}SOvt<<w>!?s#%_QZngl6DW|ozAgs
zXaHaYi2Pu(tWUm>QKIyp+A-_si1Euqt98C4XqXu9z^}z2o$w;#eQWsRs<<{trIYc^
zpY-?3)Z5*7(mg${*S`ZDf)tqjLcd4z#w?~uQ-)%Y0`9qO<dq|;%ea<8h3v#^yQjTE
zE7OTejv?5ayjl6N<vGVM+lA!H&yHY5S7=0r8AuoLg9+2Oy;0JE)Tf3;;~mk*7vitG
zVUSiy_o$rcN9Xa_zR9YJ2pBnh>MO!gL}2um<3r-ge(P|*ZVMEY^9G4XcLG?}IdIja
zgYyQ!Fz79J6(>lWALM|pM9qDDbf>H25?(%){tFO{5vA#p&6;DwmPEJTlWaiEKw(?X
zB0(5|D{}XAlqZg9t4m-V2r4(vx85}!t9%-WoU{lgxxSK3r&TLWEmZCVDprKcqz94p
z?Xa(5?N3Of%~>Dc+!2($J-Na<BV;<aeUgV(fn_j@Sk0l-w{cj5q>=0RFoV4fbp}5c
zVkOFfZq^1OVhz!ft-@<)ADsl4w-+K4BCq>KeDubQSt4@kg*mi5EV=^!lg~|kT)(S}
zqNW(a_^Q$xx5Y49`zAu8O<c22=gUG<q0?gzkLbsa_Gk1@YPjty`q+_^8UuKn#H4!_
zMGY+a6{{{Gw%{wW(zE6E=FOqJr`gk^d?o=VqXnCjYcD;yLT!cmHz!`&a#B2_3%#Sy
zbw?^jXIcslgWOxU3F{&8U)S5e-nu;Y<*L#kG+`h<W~grrv#aR-ZT22HyS@VN9|%kH
zB%@y?CpE<$p@Rp@F=C5g-pzXdmb+HEA4iDat$R>#nYt0g6x`e{<VCX5*=u>_pm&W0
zN*-`(Cxq&dl*<c>&=APiC3m7Y#}`~(9}PVnJIV9PQKE}GNbj9sJ6qz&*%f=$HaQY_
zc&<4)U<HTqP~_ktFSWtROZBWe$-19PqAcpm3V8(W<U6Iwj&`T6mB+T~%=J~#*vQA~
z+>4jaZW!`o$2I!vjCh0|kINvI&s2ImEE9X^^ZofTYwkqWh;EVniLdC4gTbL>+aFBv
z75fwVPj>yZ=@V-*KbR^RHnEyift8a~(Df?X0{InPj;#L#$=?PnK8qwBMwbqM<;FpV
zm(?4O^}tOW&$Tp@55u#eY?}`0R3RjoGDEj1p0~`2_|#lF3Nt%0Zcz1%^b=aj@m1RO
z-N)jr#XcA?RpoQNt1ZQ`4I!p}pfx_#=51^Zi!1TcQ^+BXTu4Q*1g?*TkJQeb-kOBb
z^=MpK^Z=SK!k)_9m?i+_lQ?p9HM$gjEH`e^+^S3QiYhy|dkCaWNX|*yv)pwFfjW`Y
zN*saqO}ud3eo5Hs5^oX{#Znlxo}}6?v*u)n1ldeGj=Hio_Wqag52_-}cTQ?ra7rJm
z0144Sw0D#B(x%jaxmctmc_hyRqi$JR9(K1MlQ<54HTJkX^k^EZp2Ur4{w+kIgw*ZG
z32v?9MX}GzWk4iUC|x1_J*Zh(ZQAl0KstE;WscCxIN2p1iSJ#pb>c35gVA!Iaq`r*
z7$lhN3TjD*@sFSK&FR<B{4mz}QQW3oC~_rktdpH3Ax8hrL-T7Qok`E`-_3=q|6o$s
zueRG8CoC0wGp_$tT#%l;4*a;zq%Q%c)R{jdJY@c5i2grga{i{2Ie|Fz)|^HhBH@?k
za%dUbLM}y{Xb!8f6rU@zcT9(Gx4xTlSLSo9J<hbLuO^(U<XaKj031iOU_(Q*<_SK}
z72ald$#>gej&oUlE~(9^{OI#RI;j^Y>%O+Uk~rFMTrp6H77rarCu~AawsFxVEDx`J
z&#5mD$-oXgd~9*LHvAYOy|=M+^~~d#`z#?xo@*-qwOH(L9OSy4?I<8xKiLx`c9~Ev
zvfZ~8seX2Q=uOM1XtqS7s!I`H^m?cl#RgdO<_FAzeL5hg=*Q4aj5A4X#N60A1^0sC
z1+~@`vNRh@n&4`&I5vVzu#)x<D2%%u(s%#08jpY-`+F44W?M`w@}5%$RX0jWz2~I0
zlW<`8Ddmf*RT1J6uSZIp>`oNA8Jyn_S4q7*dZ#7bf5ex;38u%b58<A#R-g#cz-%tu
zmXhuQ>aOGOQpSTH=1Vq#g`GQP;-TZj?=6qc_Z#_THCmNhNRV)P>{%;Q`Kwb~kk}nK
z)m5jQvYk@sGjfAa2NJ8zh53mVvPrZQGi`hwk~^-WQ3*}dH&boQbGZl|H(EnA=rc5m
z$J0J+ngv254FH;Fv#m|d*$8ShcnV5C&$vu81BVs6$au*3IQK4E2F%o)Yc?UaGGD)Z
z`j_d5yNeB0Lad2`N|~eyqO=~BzdmC`m$VYt)3mqMM084%s)-nd)ueXBknGqL?z|ve
zyshym<0>Q}W^<G=B)4^-^?r{n8#whBJXZ#`oPQPH;WPU0p<yRyz$fb-V#a(n8u6SO
zl6mKY3ziaxEVH+(2<>j?V$e^%WrecD?>KnJk<Z;&oDcSw*ZL{VoeYW6K<sPK%zjR0
zi}ojv*5tnO*ymO{z$G1SDKvrbLB#CfSHI)P;_943uH(wx0H6tRZqB=NcVYd(CO)oq
zy%ayQn>p@;5ulp2Eo4DOoGj%+HEy~V`=IpW7WePQsVl&<Zxg_<u&;Ki8-np=PYmR~
z=7P_k+Nwlx3jmh=5k<5h92Y<mzMnldTp6v=E_N%{jV-~}Vk71-+xRTjAW&n19!Y08
z7`NKOlUz4+y|sP4SB@1eCF_Q2sJoI5yq220lQH3nT~anWts=gTF+#b=&NsO6m8?qn
zm7X%?b=Fd#fvE}&ex^m_eL!In(t<T!u=T32hjFIoLS^cPxNF*B?A#ihuk}KNWfn)w
z_3kCG1gs!eHcC<XyD_#)*aU_dYMP;o1)#4?|G(*@c5|>A66#=iKw-)P0n0E&!)6Zh
z563$^uI^T0j+zX-Q4R=3vQUN&*obC~L)rkEV89Yek4eTK1TO>H(Ds8Kiig60pq6AP
ze%`hZPdP5fI3G*9t^okAZ)^i*_j4l@RjFE==^>EiJx0Rp36x{#KSct^VkYARgah@h
zJ{vX32c;Xj7+n)YQG$HnjF1&+pzbW=&+O_id7Ge{axf_EFo40~wR(H^2#9_<O&DP8
z^uG-R$uHn($>6_l9$|*>j|TnU>Y-CFzIluZv?e=ld%cH0oL$%nbR2$QzC$*6;^-vQ
z3Ymptzo7>H%*zi?90X5$^nsT_046h22A=oVJ5J1Z{({D-%jj%S#`Az3WmFn>;yd9-
zibw_iH{FkDd||V?U}2)+qSM5_V`Pi5H6sDYNCqqANTg#PP0pGldx2sFk>L4WlNh69
zD$R(miVBWVSh*&4ol4jt>&9pS#D<Nk7$1;jl5zj)*?@<2G6BL?@XAJ2zKYD{#9Ido
zA;e=3Ii7c9f1ZFGMA0RSe4y8<N2|U8bURS4pD6)hx2fkQ(<Xmr$074sPAuyj;SNfD
zE2%WrzJJ#TdBub@!TZxy@cdJN%yIF>4<;Cf3rz21oCj^sZ~;}W7&Y??o%|)%-3yjP
zR1>`xZ3NNEgPc!;EDv4E4usdPrVY*D5@#z`(pOFY!V_!8O&w_vMmNX4)VfINj@o{+
zSd)+*I*iG7nw+TOu1^qs>oy<_va|L;Rp^FgMw>CXdw4}ZaM3n_#TaXXC*{+D9L<f1
za6E+v#L%=2UuR@Tt{p~RqhUVIfAU%F1k!wz?u3um^9{1+E#K~y*A#xkzWb>i5M^pS
zI+bAeIzKD)#s`<^o2-pgfd<jQsm4HYAbMfTLoHOpM_`mn?28tYltGRW+6Q01jkX}9
zo7_d1Zl0ZfobX1k2ugFtbD_Rr$*y#kUz=m}gNeh8w6@1jHv~3>*S~g>bwTcyH$Sj}
zB<yjnVYV`%`|cVW&G^!_Xr-k9u^wA30a_-ztx%ZGMxCdpn;U<;U(rQ>T9L4j@v5nX
zxQRqGd)Bo>@jKBmkxi(hf#>PU19fb9G+zzrB3$HTB$N%hP?MPH?hp?Vkrv8$FOw}5
zSw1PaK2kj#;|KchTNmI`J3)z~=g{e!=M7Q<2|t+nP`f8-JNBR>jGmhAaJd(fgP!W5
zHx@e(?susypIpQ3YX&;?PgmOH?l`Y@sGRW1kB>8SM;`*v3?-DXdz6`|2@Wm(dVE+@
z3u{av;szXRlrMi{jcb@G=Gk%IxX7f@o3Aqi1xkRD$^y#H0bAgCU~|($0hRQW!9X!!
zw{vZtpdI^lF6ETR$1Uw4c05FO64Gk0J8L^zn)wmbf1t~+`QJZD#{c$xnG9h})5N9>
z4FDb;I`gMu-<5l25kS=IKFyyIg6Bm|N`UEpvA^lGyQ~#Pw!l99+i?Z|X5w8F8`So}
z(3PNJ_<1X;3gagrf715xcc01hKcQ{aR1?xT9zd5i*{V`(y3wG>a>2RMeg03~iIxe`
zO}8Aeapr?*{A28Q;_|ine!0*PBd_620)<0)v_OU4MDGv}6c21~Hl@HER=HQvHJI87
z_vBlS8%r@CV^0yfZsj%Hu)Fa>;J@w6<J5{8b&RAx!*2w!IUrWM)w^3@IpO`ap0#n}
zgw0LrGk9A+VwDbq^H|+H+#J{Ddb&SuxT!i^-{@@WW$T0nB3H+0YTI%zDK-LnY!kxn
zgSqw1Q^xV<zR42oAf~(uliZtN$0PXjbNSFKQ$f{&J=xd)`-z~hEj{e&+PDoYrC*%;
z@CFQJ{B$Jr|FG9=qVl`hjr#j}oiEe2hVO0<)c13#;MHH6aZMPZk8F+D5kF#1{%)53
zZZT#l5DR|GCuSS)`_5>gVpqg7Xz>lnj5i3<8n>?dN7xR3i6Wtdg`HX&)O`Dnyv|b&
zCNFfED?Q!{l<imH>%F`Lm&t8PxiA}@`Bc?ZP-#)|w_IX}I@%Z+pdEo-GwF1gr85n5
zhv11Yxf^sPSJ&Iy*R9C9MVRY?dM!v%>f<y}k6i;SyMLo%*+79%?FF<jM)OGQ!CB{y
z@-ek%$@3{n*<3K=%K?_!9O_dl4IFlND?fb#3XMiamhwymWs2w?5{VXq&p$@UKKrWA
zbg2D&4I7w;t~!7!!?jPPSYcK~+7<WZBwL<=iN%>em@HV#(Q|M7GiJivZ1wj!7oaDG
zwtZ+u?s5?t@{E&=7VN_u$<S5){f_Tm7$R=m<muH;U%{*U892XQ+(jQWnNYp21`m`T
zz*lTt)w-}`dZlaIC%yS{o_YTh=gb$aZt<hZ<JozG+YlTsl^V+!!*1?oN=n%d0WkI5
zOMrV-`Gd&@d;?G7`1FGb9c>06(F9O*uM!QgTrkoipuoZ@O;W9E_6>O0QzQV~9N%;9
z?XY;~JL5-o$O2?haKk1h9LctCiiA;<!j_S4yrUxqa#5TKZlsmlb4>(GKCcWr?lLaT
zX20ik%(I7a8&Acht``fCV#21l!cc+?Fw%pfZVTt7xwh>^1C)jrg$^qjo~n*BeTC>j
z^MbDUPG1zuUoY~QUM@mGf`-<21}6+{zYvYwYAoB;{pj9Ubk)j^Cs?OWohwA+A>7UL
z11%KcLL4WpoKLw!J=xNzEq_YmXX8iAd5gJk#w$?2`AobT*a06EPo$p&m^VnmREKKP
zK1L9!N60Ge*8rzmPB_<;Kh~8@;7f|#wt93jSfN8?L{QNua=Vh&&J~8#CGCR+Y+-=M
zPnEk(`Km_9&RH;7Jm%=lK0|&l(?nFrY{e_u{vz@0VvFjt0_TCi-Dj|Mcc@&##;j%S
zf^R_Y+`(xEN<9!T^czJU0))`Q7zQUmg$C%sQ9{3?dRhyBw0R_+p$<jc%F_0!WuGlw
zTSZ+N{O<`ESIS1Ka9|kXp>O&^eoYi_N2DZC>?&Vy7?WV79joe>5j?qvh?+WrB}1ak
zEJeF!gOFT>#qV=qYqBp~Gkx`%c9K(*Us7!jr-Bkl7d~*8m$2Mx$<sDO!0AN-s2ZnQ
zB}p+XwxBL)jz6Fa@LUjKbh<O1cC{8b{K`?kX|5%Okm+9N<9-`ohcz&$EzrwN9#=KJ
zA3n@1A7*_2xkU8m*KV&3VfQ;^Wm-1R${PbjOTg#;<|V)RW~;aN-u?p*{p#wRsRV6M
z^8tJpdF@wuzu&aB;@|wL|FVbw>@)rGqW@|Z{?SK&^3VU6NB{e4ja6lGKis8GJi|~i
zR<Y0g^O{7I#k(kHeTglJULzWaajTx0(EwOMdooTRj&0(1MU_zdtN^lTzB7Lwp4JBx
z8hH!?HR~Wpz)6-ZNdK3MS~&gh0U+*vJwRzSV2%-0QJi}nAhH1fP&TOT`^8rO#mx^K
zY8Nv84E}ojEsOg0&CN5599j`fodCMU(Lm!M6tq7Hcyvc`Ky*jz@(-qopDHeVV8_$6
zH!-`^Fb0he)hL&Z>WKpHd@j0ufmx2h9U#p>0?Q9~7(|I&fNo529Z;l=|NKHI5XC_Y
zqSjvM($DUh(P==y5P5tc<GDO=Y6<k==UV^xXaCQ1<5_G)MBkNDR|=ifq@s=|M*NOk
z_OBX8g_vXCt(@_k9&qUqC)Z1>9v>g3jgy$$q!guugdpr$JV!N7!2e3{QK})AJ-meC
z@Q7A7Di8X>1gY(N7(^y4D#cyQbh=ThcKOWPpzACTg7p^*m8wh9=lxtmJt7UIb^BL5
z-ph7h;i!m?Vq-t?Vg6&pPanmeKl-%H;PVl`!|2|Z46Xyi58nHpPq&|5cq%#Uk_A|;
z8s?BVoz<mK%`Ts|&j45h#Yxws#*)Xj6DH-@VM#Sw)o`&CcgIYtmU9ufc&F}9r5~<@
z%IoP_Kdu#<t%?l-3u_os!X~jnBF)NsrUj%*ny)XJjm<`O+-dC$qbFwcmzjht`=l<5
zhK|;hqix+<&fcbTt{Eu5XLME9rFB`w(fTiYRzr}1y@xms+9CQQ4a@N(lo6jo{W%Vg
z1%4n^v2i7u5!kg~((0Z({0#~E;A4xpgAf43s=_M_66^4BjeS>zSH1T8I__~@>G66J
zzhrAjaRmzgk#6KAgt$qcCtHei(Zvep{8(0J;<@_N4r*+f!g%p;XJJUfH_TpQ2fay1
z>k@5js)(y~TrO@L%!O(cC!u+r_D&H>E$<BNxJ{gU?PzhU@hVZJc2q$mjM?QxbQcoy
z&IB-0+AlD?0I+4Z9MLBCz>1EZ_wzvi#1Jtgxsh^d0VFfZ9MN!Y`=(Q6Yk!aU5vZg?
z*EEV>Va*7JX9^T_BAo4n(A-EOB3Ham(y*1^itJ)nt{mZpRM;_pz1CH+S3DJ@U~530
zAl__F6Qs`6U9oW@wuVcFBg$F(%X>JTv1~k@K9$+igs{D{KS%8+k7bfx#QI)+^c|;!
zG#;q1(OeN}JOQi0M{Pk^Et-_k9vfB%?=kG&aaXGjQ7T^wo-}h)<;n3Xys;6H>c>c}
zq))kwn|kGGx6;SJ9fpkKW@~*rX50H~A!5e{R<`^Bl5r=2r*MIb{(0J3Nz6r?^Vjtq
z4qp+^V+wup=0)&bpUBA}U`G@53F7d#rKp5%)jykf5N0yq5$PZK%l4_9VH-A%+&95(
z&QyJrGe;nyr^!-trp}EPvvGIm62?<gx^XYD#5G!e*-D|!&bdc^H+#x8`9=x-a~V(d
zCPHLNn@m46$T+iv6CF@osuyczEKdDgU7FTv(USL;RWJ9t1YEvV;_gxc50iV46V0x>
z(h|U4K$D)%#3!AraG6kF*_rsV<t#k|jhZ-tlUM+AAni#j?k1GQ#hSJkN(g_*K)E{G
zFCot>^<HkTa`Y=k8BhCqOZ7hQjOQlZ!M^fv<wj~!hL-o!ec<<ck($nXbT49bWqg(a
z9OfxvZu9lar|d7pTsxk^K8G7|psM?CF@n|LbF^~CEG|H?V#cCH-_sa%AzQ8moMf_a
zTY3uhFRY%e*kF_h^5#^mfE7$QE&ikRuno-cqv$;!b<<v%<yo_+fOlzIUfnj2FBONN
z1}e>F1qYkP-E2|8aX*+wpqqoELuYG87$Wqc&3CJOX5C_1v@xo{O{!!}PA{2K>m(*f
z<$RMN^M0$Sy{3NO!|SHodwVvn#2x>#nf7{}(K8~+Lv&fpso`-bP1%cbFcgg<T?%V5
zt@-MQq%7pe!3sY5u03%P>b&&m)_U`l|KT%ld04bnC5Eg5HSqK}M!ZO*J}w-(YKZTo
z$EG`B4u?&}hHt70gHGsX-cDwXETSurWpuj7%|FCR7dWL67Z1ar?OKSzQVo~B2mI<a
z!gI1<r^pfY0M<5Pd{}u(rx0C*9LsI4@9NyR>#<}ZV>!rU+=cB4iPOJaCD}RoWY(pq
zt8WNcr@nX@)a3ltM&r03OZ*0W1Dd20pswwMYQRV1&QRMTi^`)R`UA30ipz5bIP*0`
z3!D<S-KAT#zV`UlI?9&vo87=KT1juP``QF83WHaS$c9j}G79FEtQ^u6{=w3Abl&wK
zJ!j<o;o`Seqe@M?K1|YwuL-3&ejYz*jS+Ah##eu2xNd!aMKpSw*$buMXEqk>mVr4S
z7lHyQfsX<dCy2=dF(O$#63PtQtHW~B4DJoRT)7x`mBjW4{8WKCUObL-^5h+1+$Rbs
zTo(Co@;$3fU7yP)Ti5MYk-W_U8c}t5+JpUDO=%u`R)Q*X?sd!C%Fim>N2u^kRKMA?
z9s3Rd-1i(k4UBxPHlOLE{}dPY7jx)e{8dFU16&75A5h=5CVnt!7xDT0KV}C0i{Jgv
z_UGTwE@n?ccDaAeU_4kt9q?6n{e|l*&xe<B3&n-{uJ&tB$}kyZ(IB_KZP|DmmsG6F
zd3L_68928gTDQQ%7h=Wox++G(`U3MzkFnt+7>czHlu4uo5=U;uE)zN5X+`HYDVE-T
zWARm-r?F%C){B={f}n1@e(UbTM!E(lE30*cALlZO&Ow1TLODJ4l|*@3sjk8k%Wj<?
zos+yL9I&GzU$6o^RY8|R4o0oaW^A>LBK=|;(Z?4jMo{oR`ov1njsE-iA5400eTPhO
zkT;SX>nr<;V5v#b)ZqBE$_Q_7{1$-X(G2k9+40io+uwGl@MjmlH=QyZ5HrcsEo4&~
zk|Q*@8r!E@%lwjht<*RF)Tb!^bK%{y-CKCz{}54*yOy>+pWrHDXMk9j-rHq7x_cvM
zrh07q9nK^{gvQJ)ir|y;l&C=gcFeh*GnHd&LkzB6eErT<8Iwt9&=%s#6R2TLx%rkz
zk*Rc%`aM9TE_8NQCyq2#9eOYMinzMV{ij){?3j+vBf*maY>f=Q*d|M_ckGtJw<sy9
z3Q}K+_2a+F>c6gbxYVuxs5wp_Nc0+(yDc1?@tj@sZH?d8@L|*G^DI&pJ-bNm2UK$P
zs;v%8*H`*l8aUu=UD8T*SHX*<3z<yMj|$GyA(XZ2)R35HSRU~!)oI5KYOYi<z1=s`
zC0r_f>5Uq8%-i@_ll0{W*U#46zF6;p8iMv1Sy)a0IvDj5F)9W$O<=niYT70cB~T9E
zJOHeqmQZ%wy@>9If{FZQk0jX7MDgs$b`Hh=g8XO3G=}`^KtXA!4V1sg4pvo0mjns%
zbX~z$ECjl46VhS%z)Z}#t+7-9mHZMPy6bh23`%&)*?qnJtMFVy&rzQUO;6YI9B%^V
z;Xv_R=&Fyy{=F-@d?qbYSA8Vzg)=!th#gzh*UaXt!xy8caWmh&{ZFaVBqyu2L>I@8
zptS|k0F~tHA4kS*3^j(L&#Sa6HB~3;bF!`pbw(Eie{N01b0DuM5#FfJZ!deiR|`1X
zAzs@a(f5$2lPl?cRN?7Y$j_6qN95i6dJQ`FhB56G^WPJbF^l^Lb6k_IK@1+6d%ph1
z`9zgSAHokNyBM3B_iAzbO1Q~!?W(GTX70ywt7IZxin<<UnL+}GxM<win@%=eF{-#n
zhyox3Zyl&I9Lb&7!12@Pl&4&gF0(a@7wf$l75F*KM>((UB_I~QD~3SrtI>`F3$|A=
zJ$wAS<vamAoQTL?5~x&&yaVo?PCgc=|7cP)PT``)l+y4#PM{yDX5{fjtb1Ht@5!UG
z7~jK%ms_fIN*ZcAqn<bO-#`5PuKw{VLdN@X9BOE>$=mw?WqMtk&-fn|v-oG9V@{%g
z>EV8?2f-^q#?rD^_XXgl8+-#&*{%TVAt8bZZSbqF4TP9K1)v#GHN~OTD{>57Lnz?s
zqxR$%YlMF%9`YYi(ELZ&<|-h&qKdKmtmCNpPRqr=(chT)Pwz1G|3<y-pZ(r{_W8g1
zb@{d|Ku3U@Cbl`T0ExMQncZK=bCX_xr*(nMW^fVEgT~*E%={yWr~kvxoWlQR95sLh
z>sedi;)8wE0e|^jmc746x+n#I1SrV|XDu+Bwd%tOUdLo*`gGXh+~U+8qq~79Z6g<{
zXO9Vnq8$bSV3lj4+9!dS$=_eprJrV;0{FL&B>f7F0W#r=hX#zS@xQx@`ujio<6Zwl
z`iplFy{-lWZFBvqJ(O6{=d|<Ms!gp^(G8oDs{TB1aY!?l=B?j%9#*JWtA6pFmypa#
zpRzQNV^Dj{)lu#ZvSgZI^Z}48^heYhP^TmeE-wPlErRxG;MPyxAQPRbMJHze{oDPF
z!1NzX_+UC9EBiN3H;e`tip1u9;8<IchV1TzF&JFb%)faW&g!<M=%SVLTm}E{?~VQU
zXA4E!270UZl9|T;?zMk3;*a6@V@^>2x;h@C7SH@(aziB;QzYrH7yy3flP8`LVTq6P
zBS`*ktN-Hw_~XF*!zlQRIrN9&^JnYvf107l`Wst&DCp?Q!x%w3^YrK&5^f*w-)ClG
NQe(RCf$PWc{{rTTu?heH

literal 0
HcmV?d00001

diff --git a/_static/bonsai-filecapture.jpg b/_static/bonsai-filecapture.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..7a809d67ac10f043046af7dc5960e00b5640a2d0
GIT binary patch
literal 68990
zcmeFZcU%))w?7(+qSCv7fYKEKr7A5dT|@+>mplq65FsMH2C34UfP#QhrABJ#kuFU-
z2nZpe_XGq2gyhb1&i$P8`kZp#-+Ruz_g74^|Cq_l-fOQ}v-T?AwaMSfb0DU>nmU>w
z3JMC47w`p<r$K5UDoV;ff53+t_?<d)>J&BgsdF?mr_a!xqoboeM@vi3z`{h&z|25P
z%f!LN%*w{j&Q8b3d7gvqJPR8;+n=3KPyzQ)pE`T$)LAxqT6(ts;Sc!>i1`d<Gxa<b
zg#d_>nSzR$g4_<`2Z1O~0o49U@c;dxpaf_<O>^e#Ia=U?Y9<gR1r-$~HPxR~16TV4
z|AVNRPqAFOe)}}5!DAW$7dGiv2_Mf0-YIWkHypwV$vk-(boLwv=Xow}VG&WW%U5LO
z<P~n*R8&*f(A3h_x%=>uk+F%XnYoRv-P33G4vwyF?jD|A-af%0p<&^#BO()%-Xy1_
zzI~ULos*lFU+}50sG_o}x&~QWSKr#!-qG3hwYz6{WOQtN;``(j20Op7xU{^ox`yA`
z{kgY)KsY@5b6yl6s(+r=zs&3}^I`_(MM+IfMNRYPyeKF=fkwqled^Nn(=4|QXdb(;
z3P`^?!*(a(V|mM2K^a3F`;(VL=QxC9F~ayir}j@X`}-yq^bgJKUncgi^O^$DQBeR7
zkBS)t1|1)FWU*8YoL)1Sv~KadTA?mN?;L63a8>E^@}gW68AOY|MFwT5=#oyf$e@l1
z*5f<YB3W3mtvoaH8*W2_SDBS7nE3A8$)@b{XTL4_;dG-^8yUpur-Y4h_*tqVSJM=$
zf5oU4xBt;OzU~&r*88GSbn?ZRcH3)R2AW%66)lg7Nu15E$RJgp)RUt)GAJtNFUN>8
zAKuetzHut1njwQQzAOB@%Y|_CCo+go5B~M=u(+*Ragb;xbU6>=UzlGWZg_i9`yDbl
z!v_9N>md(*)Fxh9+$SULhjN@J85Fwx0X~qKM+ODtkU`j$zn)lc70ab*AcI;D2Eoe%
zhGY=RfDFRB!GAq=$FM#4ebt|w`TvO%!>X)=<b1DQ!^EB*+S<HW!RyeF2MxCkCgZ@@
z%BqcZ6uezzLT?B$-CX}M5#o1-3`(fqB{f3GAWbdEZ3p;E_|YmER8a~t-jM_^0|K@U
z2n4rsRNUbK{CE}KnU2-xbl_u(yPss*$YHR?Ik;N$?Q0(&_F;{z@a^eZ+}SVPRLYS(
z)r3-v?f{2ZR@;Mh=4K=&-QIZ*3X$lVe3DF?!y#VXY>>#7_@V6959zWJ^j>zuaY}P=
zWfQ8@Ds}ef-r&T)KXwKN&gw8LiyKcytn5~gCl#4@_&O-5<GC&+sO3C=!Syot!Zm77
zwq_B$NXFbY{aXD#!ew7iRJS^1e{0?boJ?nLHSHMGbuM>~^8O8W!}}724v)3aNJ%8u
zf+c(a5N)(BN!FW_e|L)vIv6H{VvO-7ak5gUyzatSp$cf*p(oluzJ2Z%RCwEyP}CV|
zHS=!Fh+UWE3rHsy2ac14yq;nnm*t7rx~kDW9=D%<vs5_N-qwCG2K$bAEP3orn8DZV
z0i*AVm)-0T0`nw)H0cO}&-q#lXcOS=YfdFX%_DEwG`{R?$mG!Y$`kQ$biO<r;RwMm
zp|?dzAy;BRYz7+taC=IKvPR;*#kmw(2_@b>J=1&URP55Z>xndmN)zmV)oSCY4%DG&
zJ>sN-3>qqg@I(1I8|6N(O~q!sJSstQyBIb4sxTa_@w^-=3HBShGZdGQ9BUEi-EnEM
zDXs|hs}lN)PQvl<tos{w&hC8k+H#T<G<%#TLYF8=S4^oGbVg3acuLxHjBop|^bGy=
z#6*aJ_Ot(IWwhJB)MR6$oU#pm=g!@}4Rwx@?Mb|;?#fNsLVNnK{8e?t#mfRGAYT+w
z%Abnlw<mSH#}6oLtzQqVnKgc2@|!RJEE~Ut;rC1OFN)&-SYhyo=A*lc9BPo|r%lVA
z=p<NzT1+WrVu?eKlzGV@f!J^|$od+5r8+xMId~pLj2VaTVY7~;?;b9HQ3{W37Igw4
z_p56tZg7zXsu3TXmB^s4GtYo~E|WnyLWmnVs5B%N_0>}qPVmL*l<gtT@aQu^)&j0K
zBtFJAabrn3hF{MJKSe(`E&XKHt{rNm5KJ`d=V$cOkyxt6Fjjg6KNxMa@)T<8NK@(f
znV6UOTKMVA^FHHIp;n<gMd>Z1t4qy++4;Kf<j_YWUm+7je#lx^xAnOOTu0+951$C$
zY9j0jjLPBZq?5aY?0#*$u6wlX^HM=ZOq`Ise$V&w8+Lf1-P<1+Bls?jHj2eyilI?Y
zBaigt%H#wf1N4g(-ie=>5-7eZO)sd>48AeCzk|vkgZ8&~iJZwr`f2yAHVGEH$oT_e
zZ?RW#yI!H!Q#~!qv8rj25w~vT7oTt5F3@g2e>eWibX95c+lyG`>(iqrauW#}?0V<|
zW73X9DKT{pFb1?+BgH;Y{vC}`i@;0<3A!(zc%ewqvCTolqy;i)Z-fl8Ge10-(;~)9
zWbI+epd)F;BU;_#Br<41|0LO-D)=NvS-~9}7eo>X*S~XA9eufYbuI6ef^nko=tcHr
zVT?f+A|QDO8Ie1RP~}|#Gs%A6PT8J6)miYNDuD(~r&yIxmM-J+hS_3Tvu?V@b>t?Z
z@lzgK5^C=mUh3G}43bwLgS1W63xL}ZGte2tB-Lkxjv{+i9zqmf_&oUhnl%V#lp6LC
zMqiU&(x00*`$0Ha<Q87hKihO(*D9sY%2d0n*Ll`^UKhbeXlMzLufu14EU|Fh2cMGI
z6VKq4dRkr09fDiaQBJZDydo3-f!+O*_60j8#I&11@%?QRcvm_Zlt>13o7h1MgzxL*
z<!1$lFZYl^IMHQzh@TZh4i+3<PX>iI3qiTFz1AdPZ3>^iuPS!TWTu}ai{v@6G1`_Z
z`v$!e*5Q2>f3IjEP-1`6-X}i`bF3nxi-$zYG13rak~#vUxWmf%>Rj@j8{-~6enxi`
zq@Ac)lD}B@#ygHXNMCI~k3A=OJO6-~-FzYiU-I0MzB@Xm+7E6i_JQu0duOpYW{ML+
zW;t+888vlS7S)TwWKf7~hEYMX=qpY`8@o^QRJL7u3etx^IClhu5;c7JHxSruCgR<4
zJY!__QZv#HYDM98xCj?fiV~}_{+fUURsTpm14-<qbPW`r6)TKd)Kh_DwvT$p*AE{o
zZUgUQ8{AxTrsSx`XcRTykDAp#Ib+1#R0uE(-k5d7R3sZHmK_QDN83C}0${qz7<@C8
z6myrLq%-XIu*mNY;T>N@*{Y9T)SB^7AJXSu>ccHX{~EIkmYT&a=jmd;c7Me-;|WMI
zNa(<b461<~$%!v>8{a=zD?NlK=uF$RNmlWywQ?^Za=je4AJ)R@2qNECrr=XQG=#q@
zBoueE#3$CBX_SuFrr|oL$-W$cK@KY4CDGxZ7M7WbhRD%2-Cwi14=t!bJ<`7z{?rC`
zy>BwqgW@x94rOzE37XZOgCvfdEzY)TET394u9)n>#q%1fi*@KwQw6!i2fVqv)3A=1
z?<c{7b}FOpy1?vUZK=epSyBL3E+NR*)gyqH5cdEd^A$I?ae#IhwH-B-&`^K+op&*X
z%Pjg!Yr}2Y9y%)1dRWX}`T!lF-b%^3+}^2C*jJ@zbym1>%=wSTN2f1y<szfawZYzR
z+o&?{5a03ZO-+}URz0{REO;vrX-N^8SY(deJ%JD4tc|~=yBIXDp;b`5_&rpgwOCbf
zOXkfq3oDpWlu^TQ6&0%$)I?hM1;4St{J9<8-s(5(RE`m6DAeMeyXN&-b?__}TzkCI
zZB~4O*ShRnu(YTuoj*idLq}L9b8PKz?jrwSZOV9Y>*!9Tqp`yw@}P<g((hW}x9cj)
z3UMNXsBuz^UJ;wkH?^J|5$iWQ#!lAcno{a+q>NuI>q#}YjI@aqnkuU7d{p<{o8`js
zCYxb%ga^;K-oZJs?D$h`2DkoaYWxqqlt>!9jHJdt0FMZ&!N&kN&RS|@5aK-er}@w=
zPC0t1HW<{u*ee#TXQ4#vdUayzwMOijM=~<XF|FBwIzn?Vp)rF2m;6BZdlN49&5{J)
z;$8<Q;mvjr?xr`J^o=&s_b`^58W&QODMm-Lsu+psaUFf|!NNlLs?b;*!jYd%m7hq7
zN4&1q=k#(WIVrT7nADpMHOk4oshMD&mgy3#nm}H@$P=)gLZRX{SG+2+55d8ENoR|~
zWeV6E(;=QxWDuY>C^8cYyiC5NwI9Trn%bvUn$>r?^450<-s+1ID1LRIqooPjGk!Q2
znPiOf43g!QUw$z9v!DO|=NdJk0~3b}2I@J*GBuQOFQH2+7TECDL{=>9f_#L-Cu@jT
zT$j0gs*6i*)Fn@rZp^WnPWWbQ3RcT2<@!<vs&S5EioSQ7%PYJi(Mz83wfvLm)+zmV
zyYgljla)i_H^9Yv)|hhV@cGlk*_*~k+D0ae-5i@$9Lx8OXHrw=dy()6GY7aJPnnVE
zxH38<s}}<$4g}!*Q`@9o`)KVWLiNM^^zyc^`!{H%(AFwM+j(~0U9fe@+^Zvl&ds-?
ze@X#y4FUrSW9@{O>fmF+1$>;})Kwc78f^63cM_N|YN{MrY_?;yl$F=M^wh%j#gYy)
zSFDMf(?$DHD(qBN=XOLU|1e2-MFu*x6#8jp8xwK0TA$n8zS5Bcur|`BTm@D)oYJHu
zq(3GL@nnf^U3=Z^X}3sdfbBMjC+B&hiKxD#UJ`p(C~5@)EW9{gcxxQNgpY#^`Uux2
zgTA-}bRO!@hjvApcXhxS5vl@?Q$$6pgAh6O3Hto^sWm%4)+t45;<cV%YCSZ#|1KJQ
z@w%H3-@H0u2z%3_#W;x&b`v4T7?HLdUZ55ELq{d_(`aMue#rUh1&hTur{`>mT_W$x
ztl_Cr>r!8wmCIlHgAX~SI;1*8$7cc7{K6wGqIeuFd=OPY2F)n~fj~bb1{DoqL4P5G
zPGQpx@Zcyp2(Lq`V*lb&Fzn;>Oe<c^*>HTunV9|}yf0C6X6den{13@H{<Jv8j;ZZf
z)$^-I1rZlx=`A)njfcHG<tyK@h(`#-%#_-tFke=GD3BA6Xwtm?%2`rDzksN)TtKAP
zNv+?20l~&v9P)(r2u`9x5P+cCFGN}ivx8jWxH4s6ExAL=B7*|gNo?@Hy~9LD<M$PR
z!@)oK@lx6gqCMzNopgW1>3OF1JF^d=j}j6(Z>^Hp5!tO3YBz<*pwce>tx_3FGAQmV
zcs(6{{ox9DO%lA0AoWyHcy1J~jru36nCE8)UOK+wQ_XLuZ(R84!<m4d%u=n%hhs&K
zVwC(91;qyYkM<nXB0<G21zj%8G4DoM7#>CATvv@ypP!IHnY(bnFVJ%sy2X|C{W^TW
z8-+)cMB+Biqm~=d1Rwy!YVP*n^s{|B$nGqR2*k%Ca*j_+9d0X-0NbQ}JcEtm-AOPf
z@c^*j0EYu>Pil!G`Sz1Ru|VWiu0R9y6#hpC8KkmA26e#+Y2e>O{&z9fdm^ri0FyN?
ze=l3J3o*Yfuyw_Ze)O)U=Jj~}yErS(t^j&GJW}~_wwKA=c1+^~68jiLYt(rIv2iiP
z@*;J}v1T4?#awVk)A|aS2i{c=|8Cez2I&eQQsK`~lPA@P2o&EQnB9ypa$6V8{X@S%
z#}45*C8UtDB-j3_X16`*@=d;{Joz#mcN+?w-<J}%*^kN_#2}@^L_l!WJ)4zBmAP<T
zUji=T!y`ysgsi(mZhF!xnwb8*s&9+0>7Ao^cb`6z*kyf92Gws`2V=;fFzYjTk>LEb
z@hRnQC%6;usIJ71me%^y6oK*4bJ25Y1i*7kAy{BHEwD73*F};rR=y9b#+0e@zON=t
ze;;@um9JysW}rH<^cB<%O4)*D5mIr8PFLdHQD?trn1&`;YpONP$cCS?2#$9<5?;Ar
zVQcZ}aqJY9<8y;_P^-y$*S7|jyDUo$$b#-s<HMcK*(U4l)JKRD^Q^_b?Unu2J^l^}
z3rs;9Bb%3JYg?=NawHLgIE;Jz!OKT0W%1)Z@1zD@Wfu*WM2ctb3;6dJObK9bT-&^|
z4^a^oUhTwVjyqW$*^6N`_x{*ep3e0EW9Rlq@2<FDMZ(}j6an1ACqod$yJAG!;VcfM
ztN5A((N~Nidr(ZS%6MshpXjl=^bd23ZkJ0xYGb0EB+s*SrLB?_ypB5{T&iaYiDo3~
zb(5nDiYKKtSp{pmvu!T^_B_j|+7YKQNp801@a)9?iwmEoBZrNH{N16|O(dg(Qul4&
z+;``8p!pF8s0Pmi71>Rw{LoJ+;^izFGH6g3g_DMpB#WT@w2fy!tSZ!J?7ts0oxF7*
z)4N|BHpzYY>2n&PIjS#Qh<^SEKF*0oB@BE@6AGdGLGb;8Y-av!3uUB+5KOdcn{~fa
z*q2IX47_E2%h8j<rKI(6-I|L4wgZe72SLoUHVKP1YdTmsXA0kCy@$85?p2OU<_4Z5
z@=GqVZpxIPuD_)@FTR`BV;JzwFm_depbvj9vD_CU=xm!gfg~jWJ`5SOP;yN3C4iM+
zhgWJjGKut)bDl~GF1jOmb&I%O+xWd(P)W${@mFOgdKkR*9S(j%X+Z{2i2^!neYX<B
z7~+l&wZYV&MB5m-)Eu7S=8bTqJXk*AF75{z1nm~rX2&pv3Ac%e3ijG`l0nJksAE~u
zxBA#}Jk@@HJ|FQ$bM3I3J_s$*>J_Y+v@!09o??~Nk?u)1wu-AW>hz|g0gVqMWpts{
z2d*baw=JN%QkfTq4H{<A#PFGtnd6tJ)Y&Xtz8?_Yy_Eu3A;5(GOe5YSfJv`_b@dex
z3c$M`uFk=O;mkRs&#w~o9+x?gLF`z3!E`>}_#KY4wnH-7)!AjJhkdMuazvdQjK8L8
z!1V`xz=t82i!P8xIGSE?5bw+hcSgyhbdO{r0DD4a4u~~SBT1(L!>0qV{LVdVA%lLN
zg%Rn1Wjv~EDl+~?7U7{PBN_BT7d~f81`)0J|4QWMrw<IGB({NIGH8VbP2d9mQ{es=
z32JZ@qx!ps17YF6P8mpR{`#zHvIBpU^qWh6hv@#6rN8F||IMYpi%W|{79E^*puXs_
zDw+(^C}s2FC#g44ky`cv1u9|xfwsQ6FlXw!mBZ0ihD+gU{TnHmT%(UA64%ScV|~+5
z)}=alp6;5ahBROY^RbV~rpH`4&>54GZFtQ+1r?8@Ln}+S_g3fFUURzgDE?72lZ&|m
z=g?{m%a0zegfm}nCv{UDjhhff<@t9z=ShbU(v?>;Z>0A4!|Th^i&rO_S<`m)$K}Z&
z2XjlWJDb?1sVMHluTJG6%*>i|FPOD%U6!5R4)>7?83+Y}LuYkuFo`i-Lr#tS?z}0T
zH|=(ezl)p|N{Dqe=f}PO4j<r&J6?nniolqgp8*F!f&Tyq#Nr>peD^rUk!n2J;TU*0
zQP>k>CG}-JpYUYkM(xmZBBv|+o9Rw|Kk?_xNQSx#2W3hfaC#EQS^!l&4i!ce%cQSt
zXdEe%gjW#Hx?XVYg8HQ1;f%R>tEa=(+6y8V%b=~A43O3zPQ(L=_+}F_C^Cn{n$CDQ
z??MLoA%K<c{t<cv_uXOIhDN`f@GzI4xZP@OO&K=Onrd6!848<A4=I0}cvcEj^L_H7
zJ8!7!FdE|&{+h?S8h6D48jb%t^<MQ{oke0dk9xEn$8<$`!^N8z(^oA5H2`?$mx;n#
zT_2M{%Ff_zfFf+6du_ex*Ksh$%qyzZgTyiB$9__PiTWdmo@2KqRmM!6b?a1UX{6-r
z_X{4S{i0pgoP%!DR!aAV_pFrkt79G`tEzDQKMn3gUMqV28B|1b>+}^V!u{vq<(RIM
z9Td^%7nb`I+g#O19AUs>sd}_cVx0;ngU;vt^+X9^8~jH-+_Hza4i15r(N=)d?)!iY
zng?9*Ur(?DYx{r3=v&<UxGGuw!-EMpE^`!2==jUwGoa_c<xr6>jLXc*QsbP)3vR75
zv{&YZ=Enmi|2pcwB_j{;4uP(ICGLOM%j~+a7Wyofsg+?Q*;m*4eIKio>GDxV80h-X
zy1cjb8D9z}WzUm<6vsyKqcZSVT&dL{eAE&?0r@lib3Z2&t7K|zW+~(-Z939@?yT}T
zY@<<W*C)#;3x{FT$4mMWuTBkI<p0pMym>1^PU`%RdwnU67vnx81WNk227lR}0Ea%!
z@O<Fl*XAj?v0O&8I9wY8#mXmTLv$Q2)@GU#;Fl=Q`Acgr6<H8XhMvAN$Y}_<-gzi!
zm9riu5?H<_&v&f>`G;g+LtGt5JTCtR_#R`a`##bDVv&oxktoVV2Ocorq1SCU4@eba
z)_>9Lwo!aRXII8{?t+i+gc$~~v@6#5pSqIbTz4vx0(ma*cNH{i<x!OX5I?<eA@;rD
zL>f`b!A}7TWAV~7<JNW(;`FflrvHYRkmb}%NU~18_4g*&{-Y!u17jx|zm>Jm1^{V>
zzZNHq2PkPclsS%biw60b%+oq?!z$N&GQ(ldoi^=2GM&n+*hEjGixtzkx%{cye$^9R
z(<GG<Qfmt2$c){~sJXC^C_PAgI2ZLv**DiQ33_l;mRv>vx8*z}4Tf!P=_c#n+}x7D
zys*0to>KRktiATI{$mZo867i0<mEu*cQ`-~5%mL*CSlL~g27B4JYV`eBjalQQjACH
zk#P=;cdWGUk5@wXKPG>^C3rFR47F$J6@qUEH>^F2$($f=fuD(+o^Qa}Rk5@c`KdPG
zFVq}0A9AmsaQAoxbjcK}MMQ=iUgNsHYfqS;i!H?J33iz92@bl0IlSU@33>%>d26Xt
zzC0=80ty>^a$3n6Yvr$(Zt{+H@02s{J|EU=jzFo@HGz{oW*Dw7rAizJNQ&c@fY1-E
zb(XyWgoS8<3y+i2=kqE=7#nyNwfvzhQ)odrw`+l_?5MhnojOwb`gmkP_tn+#f$zDW
zpY*93Z7S>G_I0h)=1~!nL|UoWd(eK}do@8ePA?LzOi$wObBaib9#q~b<*ABS*Uwlu
zAG?*zsE&q`9L#~xkuNi&f8H!QZm+;~T2+39bPmtoW~Hx52xpv97B13tdhE%tI??Ct
zEEvl+bC!j!XDB-8dZsp4*|*M4Ug^%+Z<ERmP_J>ZB%w_s5Jgh=oF^g>jxp{n@`7VE
zD!o{gw$EIh<rsC6?H9%L?}yt7b`<LFF9++^I@o%L*0x7Sa;smFXS(g&pofFS@L<a=
za(yKnh=0h@>fIl~Y;GXolVE#Bwr8f$ubbJ(V=<<t1~GT59s!RQXtarf<OG};11w&z
zbQiMT=5KNw2NaP(+4|pl%lDhucl3ja=Z`oybr|vrysWn$)-{Z|U#{l0FveBSgzjh1
zMO>m@2XKK;Q8f~axyJ2E0<wOEuQx|d3YC>J1@WUC6Pf2K$1sX_qW-+aBM1AHZb?V&
zHhoi71aEYaa*cvykELQXR#G~BF~l}`n+&3HCWFd)w$-d5J_gO`Wmqa)&nl&0)Xii7
zf{R%TxOft!ER{pZ1~ZG4ioD7#`&7QV2@k0F@G6g!1Cd&LuaH%(_iN^pK+}mg<^-*d
zoVhgfl>oN#A?2IT<_+9^p)|qn=QrI|6gJvr+uDZjSLQyM>(py}uJATGGF-{RK||9u
zhQtZ$5I=;$Q<I=KW?J-%n=>Q#t<bvgFRDMXn-wZ?DYaSjSF=op-NDR6)uhhsT_fCP
zYKWv;E~)qUQk*TfZof*~)Cr>Z-A6k*bfbX4+iz0A8*WXE9ckaHG5V*WiY6g${1F~L
zAyay>p)VwzfT%(%|8?1)YSJGQy#`Xs!;JjiJYW#GvV$z>Ap#N0Imn^-?&l;=<kJ(U
z{(|B~2!Uk>_2ZFAJ9k!?pPg#Vhc!<{w((4MRWZEY?8m%y5?urS^h7P{S$U?O{H*<A
z{CVO`g(NQX5243MDLo7%sDjXWjFHTm&p5Z<Va{F?ht*6`fvWrJdU043YgGkVJWX>s
zZ`#7Fbv1PCiQ%0_+p8y;uyqV<dqjVgy#Ornvw_HCLTbpe9D=Z{WHO_3z0Sd-z>JMX
z^GbF&jx3wpX^T}81y99*qBzm1+tXnc_3rZN&bjs;hgt{4RcYmTL`$YzGa+?eF7alL
zzuwi<*7^{>Q;idQv}_h$qft(Gl)iuboFGXZdvDPwibiq{<K|-d1GzTwWAk7S>D*v#
z9aAujPK0~Zpxg$kiW3Mic!3E_e_6EM@SX%9Jf@=$bw7_(RY!fx^oZ5D{w_CGj`BwC
z6SFU^bGGq}JjWXLo{C*zepir3Rud7*63~)1g@~W&JPLKb)0e1wgP*@j^BVUE(hN1~
zy}}&U$rv+9yikdkES+n}BA9IWvBL1oWj}eP9u4-?B<E!a+!plDd_?a+HQUK+<9<&e
z#Utp^+3+SM6$HT=oH5hH|I8)_J_B2Ja5%_=8_hu0UvKC88(XLq{OEFy7b^Da=S<w_
z`_c~bKYk-7^oBOv^14!B9pox$NR@J2*=&A*^C~xA!6%1_DM*MDaaz~=ZucT|r~5_G
zieeWljZ?vu-0NSKK35|GYpgGfD*O6l1}IK~mUZx`@MT<^!_i}-iW-~z^oepKOD!H&
z*X8qp^brcn4lR1FQ%6?QWROTC88nU9ku<1uP(?>SB+j6U`5zBK7`%4d6iW{}-pcWc
ze*t<EM#XH*T6|aLDYAA<{2p`ny6_h~#hY6-+s?W?b|$9tPvbzBulPL#;G&ODvS-C`
z_w&5*ijdbFtW$)-Du+y&g63*xCu0>gSW3A>U942)Sp^#ELwJXV(3+S^T?XoBMgtKN
zyj#_Oqb3C~b2otmv~klIX+FDIG8{-j1D;QZ3tvwCbi2g7%qNSreQSz_x_T$6O{ICY
z(vxxtL-a;lx0jh`(Ut*g`KcNr|6JD~%2C^0LaFt@C!>G1xKQuY+7vRhY(uJQ#a8Ex
zsHxVm+(|i1LZIf-By-N<JuPT{-KfRz`cM76O8=`{EC-f}t0WjP$!e>_pBda9Wkkqg
zd{z3h^zQKk(twrW*8Zv2=HI_;3#i*YOR$P+lon&oNkB?exmvmVg&wQ<F(dK%9SDx~
zYhLW`m$%)v&#AlXI*I#LRIEGge-JVHVdj)5B|{@1tM~oE6>ubqzO-a7j(@dqPUX%)
zhvBq^1yOcP)}LFeX2?cYdp`1pq>UO!#-e3Oh3e94Y3cbFhw5FXA_omb{zGs^`W1t!
zG<`kr`GP~_ihr<nAG?kUjoc1RJ1hu>m2kzzqTDgMrE2Zr9i$7SzOj#0Z!9G~#p%ZG
z=xRaUF@2~t<CB*D)bfor&YYlx+Zt%CM|KR*wP4=5Wt|!*QFq(C{G3|RFTOY5Jfi*j
z<{29YznK=z;p5$Jj#ek{bCGueREZU82i6iHKAGVe6mXs3ish?K3kVkOSa@ll;zLyE
zKO4(c>TLT?V2X!<=i1Fn@qe&y$16n=_i?Z<H{2ORj&^i$*dw&jPymO8g~R88kNVGX
z>*czi8k#C(q4f*e1}Ua3HcONx>hGLUSagW8cahg_hn`PGRFg3+1t#Mz`in{0%t@w1
zwh70?(5xCmmVo71Th)LL;b4x(J7wRlFI7R{^;_fH*~itq+Ml{7SIHns3G}j5OU`DI
zIFRsh*rrv(#)a2w)a3DQ{*a$~V<KI`#|>vP``XJENZT(jZCe|da$LSplKF6f`l>42
zin9%6PvCln=fpI%t$5T!B9u`KX7&n4A&*bZybUrQ(N6Q4x4YEQ5YDuH>&^zxW1*wH
z6P^<Zr=!RzjROEU##y1dlHGh|ypeI<NdX9AGEOb)>?kukz{_q9?jg=ud1Y<L?^<pR
zr#LB2xxYE|I399Ab<a|Xt)<b=MU)c7ai57gxTU)J0ZQ}y3TL-`T-7d*!{R#&ed$#P
z;nD0Hm|bj~$I@$MUt$f$M3)SLsB9cd*0YXDK?v?M-beg*{nQDmFx-UZsJoS})|h08
zgz?yu#IJXjwy;<iYL6lDyCp)iLkqRqUlQ(<c$~Mxs-v1bbs~2T)KIw1wo0mkHGSvZ
zi`0QSL1s2B7ou<8P@cm>1pLeh>dH>;SG6!(h2Gg^;d9?Nex7)AS96sTn;hY)q8<*x
zI=odS!iV2q&mehAT=mUyjFCUBWYnqDBd1$D0>tpTR;y{@!IH$=&iNMm(6eMvrBhy{
zj^GY)Fj36?_Qe}*r_EQkXApIWZzTW5eqnaoiPU*MSljM#T+Mn04jlVj**_mWyNdgv
z#-38(o2QMuQq*Yy4G+@fF(u~Z2<sc3`AQRcm^t0zY23_6umOxjo;%~pE^|C~V=-HC
zQlSNJ+hi+qRy&<83)R1AW4mKNZ#+A6`gN@)jWaS-ng5g8F}I)b&~rbDqE%Kjy!DaW
z%sFiT@OE#SZ$45zD$!Fsf3WVGrp#UIwWN)#CFBONdhpTw<^}dmZ(+Dd9wI-=im^}(
z5$)%Fd}EaZPt=XmA;9ZcF*8kDZ$3M0VfHZl8$YyK&;i$B$3ySxAn8j>%N<fQvySzM
zIDRu+Ky?{WeY8{pj-i;?bSo^dk|1@l1g1uY%I@oy%O-O1sjRhaGro~8**tL`&-8(e
z6s9$0ACtH>)&ZN&e>Ll|k)He!bj8m5=W)!<L<a%WEpzE_MT^F_7aQoyKz7t&Ej-Eh
zZymOEf*Jg@T<|qg=i0LM+~@Q;A*O^Zi_wp99PvL)%S1!FY+m42zeBK;bD3eaU>1T*
zr+m$nQGe$fForF>BS1XsLB&`juQ&EabkZ8#xQRr8Mel`*kNOI>MZ6|`^wHPslF<{m
z3gT9(uT!Cw>UM_?De?r-9DPjHF)Q{Nrxhxf#q3yyUrp0{GjX8L&H0#3+v=-9x6&iL
zWmp#eoltP=b}>P>osTD}JuPk3gK15(f974OT^$QD!By<hG%Z5ZW5cg@sj+n(Ln<r@
zXls_h%~n44#9G~1Rv)wOy5BrR-$;CyTN$Y&pN1LiI{wHa^NvmEk5HGYm$N(sNTJs=
zEJw!aIE-!#zh*VyhPhLkug0*K$gg8OLh~pdiACS@t7xP0o_>#ilw2TmWU*U$2w+fS
z%tn@WvgVST_eswts!Tf2AV2<jC#kMkdhV6P-gY008B~Px^F}If`-8(UjmPdKI$nP?
zX(i+&D8HRSUhzu^oDS$3-V>{iksl{f*T&MmO|i8o_Lf*Y*LG3a@JhXca<SMSUKhC~
zvOPsN<xb!b9PJ|j%K6`ccDjIi1>~c22f-WsFu>j)k5py^z_`}H$v=+aCmdA|P7N>v
zNZSU_J;=#fC^jz4L%~`MuUY9@XwT2QcPCSPbk6GKM5pxayp9Vd!zIvtgF-=X`y#dv
z$cv#XZHQ^}pVF50oCfVNf{C?{sFQZ39Crq2hzc<3)q|Q8`KgfdnKe$Ir1X$(7R&p)
zH_jXEguMi<N84srY3aCn2Yh-texVt<RY?ZPddc`%hwI>zGb>$0atgeBb!+f`^`SYA
zEHh{OlNaj`Yh7c5?j-z-R_*xmX&VLIlwWS0y_|F}tF1n^!3I;@6!49g$~w+iW9LpV
zZ)dw2Pq5H4Sl@Y?HowNB_uxwBc%;5WM`p1<T~vwCI1+R|Pw&26$|8>=Rp5gwb3qh0
ze|Q&X@(?b$XXpLk&aA}7v47XIoqFkTG_5&~w?Lkf*Psh{eDTz-@UBp&=9a}}zbjJ6
z{N@8ln(WIsMF)pI(t2?KC#rYMh5*4>ZHMjAvl><5W1D>4ir_VzrcG%(5_;y+s;1_*
z*^AH0*Ivtxrp(h;c2(rx6p~UEc2H#=-S7iLy^99q8s!x|BQK)Hu4dk_O+S$y6QpGl
zXQk`9`r*0(T5ihAbKg%<q&U~g+Ob&_&t4YMa4e`lPWQ5{pgg*($?CD0-D4i<ur!@B
zbT3}Inl<yDyftZ4Xy*2+MzGB~WOBV*8u|$hsj;xI8q_`A*Y!M6oATZlF5dKuyl7vc
z>9Nw14V09=hMNuhmMlq<1fTG~4{bs5`{`k=_*>0ua^`$hIQ8(3vgXQJ)A?Bas*_Be
zd5aK^IW3+N{qTpEE;!%1e;V8dp5nJo!pH*Xo~m@;O{fk|T~$2O*Y^v@0etF3$QdT}
zb_>cExm1@xOm!*>U$!TE&2R(3L^_gq9iDwZ=cAeC_z~~?t?@Lwp^(l^uCR;4ObCHs
zp~O=R9`ljD0bSz|TwF~QN5ecVl*{cY?xgk0_}T;qS9}Mp!eoUq;^uyZ(^DE2+2F`J
zhs(k{M7VKdN%Kjn@ZJt_@9-sYn_x1IvaiI9;qyNh+NY#g^WjD&r)oZ`7BLsp*F{cZ
z^};TechZ03(}cefQ?d|s-fo4(a)UecEG&ngBiwY?Pehcw7t`D7TOCti7p7aWcIuZ}
zBHAt)L>G3?SUmKj7}=`YW;4@wD%NFuEm~7w`#jl8B<TE;Fm1C(>~_b>_s<UW$&5sf
z<6v!8kk<Ttiy3Hp^X10+%k}eSJ0y(<cZxKq4b(I4Yv1bc#J;$8GULy=cYy90Uwe$W
zvORG`s@_pndw_q6+!u#MUM@o4&NU@e_$qBGY~GX5OsuF5Zcx8JXsE#{TpxJ;*#+8W
zA$;DtS6ip5P^R9U?KW<StprTEgzA7{%D2m9i;aOlek5`AvdOgYF;Z(>;o~hHQkKBQ
z?oYIluHyoVvwv86K_4_Wrc7o~U9C9?%}cxb^0to9o0r#vLij##LpHV_=lRK$<J28W
zbql7g)jp|gw9C!(w|hc8Y?JpNNA=V_7n_T}-YJk7E3r|!UYc~A2*0V|f6k{#2~{UQ
z*5r05eqkUiK)D~QA2~70Na`NmG@Qb&&CI7-J{rq0b&Z&9!Ky9jMCT^xh^T$MNa?Xe
z0MFkq#WK<poY5^L7Axb;Mipq8_4L{l@x%1(>+X@lwO)C?5#Ku`6+&p7%{nza!{@Fd
z`D90kw?8?~ii~BCNIq*Esfxm|PEGb-QIyGft-cwWc&liq;?r{wWtlYQ<ek9q&r*&7
z^|8i>m{qr#D=I9y{ow9h#Ph+#g#M*l>+xn^y~ZxUo?q)14bW3S&O5+%Yi3mKk-gp)
zJ?H>b`23{fOP}RUDYqI|$AAmCv~_b~L};m>mAkuVVU!Nb@I%#=@#}2q`9^{9c|l3F
zoR;)62W=?U0l-u#FvV>o>ABj-m1QAJyJcj&9b~%M5-5Gd?5RRGmlD2wp|3xA|MOkd
zjEz3BV(jWwIVpM#j2i5@s_E4nR_g0pcRXJy#q^oJXmtNp(1s4H&rHSZ8sX*kvs|3{
zczF5RwLc&yB_@A-?eIFakanN(R;1Ur)0x}y>Cgix>;#DyfSjv2^52gEa4kT_|KLb0
z`#&rqIP;g7uw!!9QTbkRJ>yYS+!0`m0EMEqUc0>sZeCK5-l63t=GpQah7A|8>o)Rl
z1(K+mTuI8qKQs7%)Rm%LQ&VQn)nh*KM~`+T_OCzr;ttwuQTe7ckz;Am;Yr)3tRpK;
zkGK%YW$MG^Gm!<gW*H&M`7b++Uz@u9^5)J1$@6f=`=@VGovcy&En^XfT-A`HBLJRo
zM=sgAx71H|#63n6+9r!rE^sZeb&~k8qco=7%)II#P&s3z{b8Z-ip=KltM>urKpn&)
z;&?ynNK)~DSAoW`VrFC$SQS@n*uObuGLE?G7W?K$b=SX2NHHadI-I8&eL+*|!N*g<
zX`a|GG6Zuvp5%%Jb<Lb;!skR@Nh4+m1V03J2<{vAXW{I)`OnB*aM-<BF+4n?(V{BK
z?1!0I`S}@c8&UjJ!Ft-yp40!8BF;bWeGvd8+~VBGpwD@!q!bM+de_;%ocMQZMA;K^
z2>N#i-JM(?ey-k;s%{8rK&AG1?B;3Pgj!I?*T#xXoC6)YjwZUV{99>gJZz>U@ac0h
zvdAaDI6ym*MmS0fXL!MzS-ZRYQVJWh9sw|gUUb9KH*du4>{#R)Pq9kSU+YpTq?AST
z9Qi~?l+{!a?@AnM4meU$Rc+!VFZXF0O)-qrws8uv$m65muMcTi1R7|tf&O1^6upFB
zvZ*-$ymd!-g@2zB`A;Bkkede?v^FrF_o<W&8p|rx{f8mx%$D$Ce)EP8WY8xd%UFgw
zzA`2rw`a3%;>N$n$-g)8XB`e`<P+>Fwi)Rqs)t7!oHzp5mC(^fac|qc!NP{%M6Lh=
z-};Xe#iMfrr2Z!Hh7FS&Ku;1`z3i)x9tZ}F5M})Rddy{@%nOZeG^E4ztTOS+k)H&&
z2BV-;t+BfNja_x36U3jzj~-AP?|+_b7YGsxx~6lMJCsXQXwD~5f8KCNg9?D&2mV*u
zjb9*rWnp#Oe+%&|3;Qp1{$C;Nf8^eqa4Nt^PW=G?{_p498Dh~VW=Esd?5Ly4y+V6X
zcBI%p(kTCp1mwApO8gzS{xu_ai`#uye^X)(O}HO$!%q@Gjbo^BsqHHDstpSJ`q9ff
z&qr7s+WP7Db+cya<`}QO=-*`sUv>Q2h$~BmWBp~?Wu^|(I@6t892{JiZtP&+w-1*M
zdSrB-sk@kfjJ92Xu)y+wm?6q=Ueb7Sm!`8<7X~DL>9!h$o#v!y2wIzPMrA8+VMO>J
zjJzQcyfq6oZ`EkJMbqo=wdcKKJ+6~?20LuCnG%wWDu-nOb4j$#)A-0@W#9VjxSUeV
zqeG^~M6E)LVCZwLj8_+4FCUma)xY-qh_w?!=O>4!)bxkz9sA0^^UX<G?&Uuc{vK`F
zwR7zlD@KAf$`f;Kms!nEvWT=;Mke`k-@F;woihPrP=;b}`WSr3sn;{+LfN#5MW5=@
z6qF1oFt<3locd3E`w=Zu13&M48ZFdho^kcthBZLZ#K|Ro#sC(A;Jj#9qds=~!fJ{S
zDl*-ATq5qZ6lK&2gObH6CP+yscz5#fGK=)v$j7}p2N`d%UOD@*_=!-zdlf4lH31R?
zr#wfW`w^aH9ncF)HZPsx`-L%$!7T7R!A)dEZCo<nlN9&x$ws4NWddtw+Sq<Ukwa?3
z9%bpL$w??@O^RZ|15SmM_`Y~6mdH=DFI5+eG<yP1!B_+O&<u@M2)*JpbGdnaxy71~
za-}tu{!VrhW2n@HZ~F=NJ}VbpdJVIccs<H;t3yoHglD6)Ku-d0C0Xz0)W8b%M$Q83
z!QzRT;DA$rG*S0$Z<(G#r~t07`FayBMBZRERo|CGN~Ef$!69In9e<^xnVEZ0;d6qQ
zfyurVAp?6HM@3{FBB~Wm6N2E(UKS6=ocC^Kr*GaH-?qO*w=J{fgp_KEo4urc_Nxk_
zl?<9hYT;6<KY0aPMCxf-CAFKERofRA&Fb3FOgTUE%CpdNp|m{n^t-L;0o`YY+uc9n
z_FZs5?aEM<w(N}N#QmZC6zG}F>S;pe41nu110_6T-1j(t`bQE6yJjCn^q#>klnio0
zF8x64Ogq`mZ;Hagw$Ogn$90C*CNt}pBmGUg`35u<A}a6mAa|P{jyEo<nqHMwJS=nH
zd!aFE19#)?p?Lk7SZw0w!?5wI=Z@fjR*#w$#ivF}xv2zv+aU>89J~2UdGVXz#;Tke
z1<flQfmFp|aJbgO=~JA}7Uj5!FDOcamX4FuJ=_NKDk57A(^Z+NBPrmoD0o@d!;C9f
zY<IC+Ii4UCcwHfo#NcItX<9f)lWWeA*!V1gnA(eZvim6W2F2H~%aVD?*>N{i&Djla
z3E7?-xet?oTlmjz%{0L>O!~Iz<%v1)rd|J&P3Ngdu|lSQaJi#sA^c62*iBB~u)Jn;
zp|++G8M!t2V=K&!AtpE>-b*<Uehx4`Ls;7&KcyTXgj1XV$o>}38)JUr#Y}FX3dI))
z|9%le0cd$Z%S@cmPMMG>Gyt&u;ORwnKrNgR4MVM152JABqyV5k;kL3IV0-NYP?bOf
zI9`^bP6Gh1w4TKvjyS6jM9Qc?5&I-2eR}v{QyzSd4MNldIUAV5aY0^CQs|~Q83fAy
z6Sx1Hpx+$&EknQGvfp~>ze6#e0{?2dvRCpp8kd(`>x^$Z|2?7eqtT-mAS%GU169LF
zB`1#Nk^f}LR{f)n76toXqXCA1dt<Cx?RlFFG5`RSR`7kJ5A=%vMX%zo#{11f@cfbf
z!SvBYu%9*|IoD5N(8=B3k+v{9b6LOgS^3b@v1f_U?r5<N0i|T4Fr9_+E3dgV)#2Vr
zBycVn<c})#(rTCD_EP(D#E@0!)$nMzN}PAx<H>>XDBm}Q6wEt;Xtsw9r)-~!QZOkc
zvMEUwh9t#YiIp7C3)pS%Gkmj}UsgS06|m{o9~H+mZMJ4C55~n+A6woVOfNRB(w=zU
z&-F5~uYyT_>k|7nhST#MZa!`)(mD?A`ArRsRqC1Ug@$J&KSU$SPoN?^W&vXbQM%rj
zs*w(tk{>5TKTf<;nzU555#J?z{{6z-d!tQCo^)Sv*?Kc$NOfC#pBZXg)b(oC5xWm@
z*kx>XIY~c1OM0L^F3K3f=-yA+GuJvadoO=4Q-dDnJbYDLgq9NSd)1%f(D?ohVLL+q
zd;_qXK>+jTG1>1<1+}#kKb<b8teVAZcotfi_k0njsuk*cv`i5w#2oLfAIoulN=Kqm
z{=ph@_S&(0WZLry1)EI^b1nZ45u-4%`Kt@rixqPxTi#OBCRoW=IL|qB$TyVAsJv*0
zD!+k8dG`yNNp~KLLaLWLERT^HuaqimF4inPKr0!#<((1J>~}uiL-oSc<4R)`2a}Su
za%+evgi=}fuzMdt=0OIDYeixA?gXc(7BXY;U6c$ezQiofP-h`+<?0swVn$O-n69g=
z?xirNhIt`MqngAzu9g<T*t2E2txm|f*~k|R%3z8XASA=rYuOoUrQ7MVSB2&A)sxo&
z36JXrmNYTAG<>k?Cr}I~R&V&ZU^$>?B};zdVRC&OyXP%cTem*ldVFa>Vnxa2UP82k
z3(r(zUGm%0W)ErlJ$`TriXOjTAe~=UJ%XWE&xdKx>aH7Hdzw7T?9S0A8+mpP31Y|(
zRC)5(2+02g$LC(cOJJpJI;8UEst43S{O|w6_{KlzDE~*!H5J91nvWN$M15a?x##4z
zLtNK~)K0a#=uu11mMtIm(*<c?DAxQU>r7R6Aq=!f#j6#apE!MIB0o$s(sa|ASLlwK
zIQ5e=f*(&eUxQ6Ks1CW)DGo$$tM<Fd<tBb#sju!+eQqfG@=EuGt1;C>Z<2-V^kDZZ
z^V6zFwp@+dTASXd>p3}b{X7|HG2Mp!?48aK)oZ99`K~N8qjh08oWYhs$cy$egD!pE
zi~E}5*&}}Z#jS%93&qk}S9*pkPrFuyQPFZUWH&#-p*s52Iu&Jo%961XPb_W<r7%59
zT$<HLyv%)>IVTBPES^(HNAb1fE!;lJH}3UA0+;f$Pg(40u>FFOz6sEw7tN)w-@X2L
z!K3<vd*x_hUq|rNs{$US2Ep|v7$zkS%gC%&Hfuf`W6on-5xF)_TAE1|aq9b^B@n2y
z_PzVd@+;##Dauu!&QQA?>(%Uay~>nv`-`cr)V|&<+B`xFZ@h+HDKFkeL}E@p$sQxv
zptBdM@v+tZWh*Y;6Z`T8vjalVWEy*Q)yKVvp7c1;Siqfz^R6E-X<dQVf7W5U#$wqQ
z6C~4<ct*9|)+95rdg9?=13|wmU_gqv3+wr;%#T@5m!h!uxu4_MEvsnDfxNbS<1tHd
z*`uQ%X?i~(=5ZGB8or`Z2*BoSv;M@brxC(Q!BbhJ(>g%$m=VR#C_mX@)5O6w>dd;)
zGN9W3>zjB#>7mlCOBv6%rnGI%9`-fWy3>_#-5*}Kz{Dt6?FUp+{HJ|BpczL!OC@<D
z7quoAfl36YLz8aEufrPrwF{6xbl)4koF4GVKNE7egiiJYtlHnB{Og1I&9PtP_P31v
zeqVlTnBUs+|IGD8@1nS1(6yKLxk)EB{`R^5i`$`L_%32prU0-Rlv36S$Jk+E&YI!s
zwKaJY?(KF)dVPk<;s*1TS=Pj@S&gKJ^=UWN5e46P++JSlY5rjoy$sF^05W+AuI-S$
zWmf>THPvkRfq{Rnrxysh-mBDux<A{`cxKuVE57a^k-w8ST4o!Qc9Y%8Q?MBH<-#>Y
zqN<{P2|-~$6K~QU!^ufMN2qFuYG6%1)@U3$G?d)nZDdm9Z_2{79E*3&0%f2f{90i^
zwHMrJSJYG)rO?}Hlzy{?w7qW0Nt9JMxdq<wB!j9a8iK#dTy14;$6TJh+yrA^9Ls6x
zd^YI}F*JQ{rK^*dm-`d#_HKtUeA7>D&UYX}@#j(KOnAgz+$kUid|P&ZM^<{e*VUT`
zW6IQXCjos^r0F3xH3m`Te+N39#XI}$$&l=1_Mp?=bSaNPq^5@Orq&xAnu|e~hCxd_
z!Fr4gqNgW=9xxJtO%pfW;B9K~Q6R1%3O}~;VGP;L$hk8<JsXzRdpwf|OZZY-Dbnx~
z;c;+ahiW)MZ;`<8G_T^Wv+j5sv+%B2DZE19PvOB5XFo7|#*Z4;ds}RMoi|>weyiXO
zla4_zY(orI^_LtxoJBI{WX1f}KN0I7i~KVAV^dv)JSH@PRC(UuzhTVKFQmCc*79AU
zzKp~)Ui3j@HyLEU14RQU{DsPm)SW}fk}o^^JPJ36-U9(UMW9I&T7T2iC>d1bJwPM?
z|9k`*JE_JNy-Nlz6b8y91^%-;he(CX;N5v3aR)@@h4|N>>XAWUAaiX7G|_ScYo%Xk
z{Dpn|YngzEse*Ux;XAi~kNCIG<2P6T*OeS=%(1Ei3fENS4;1e8k=l-_*J)Cbn>+_V
z(dp&5@miO63^u95`-C*h@~ML;V!G02?|rMcM*HU-8;57d6N$rOL)QI3;(MJ8dPqak
zf$#T#F&Xf+KY;gAiZp3Wv7ITS;j2KDL7d#<KNJQkY{Y<FgWSoW!eHP9L>=3Ky!zQU
z(711?y=(A=R5Iv!4vA}ml+Q>6wiexdLoqLFQVDNUB!fm<fZawRClzsoXON?};2m9H
z6UzrQ(f`@txA=Y!{l7EA!{+w`m@333tQ~A`XwR3QibI%PqpADA$e-N=Cvh$9lHks8
z9ETJMtc|Mt86|5lP^<@Cx~A@Zb+FK8(C+CCqCM4jAdhZ}HW&LRTk6Bv&40!!8CZ}(
z#bEQ4e;9Hj@NYx;MJD`fUE^hcvm0E22HOLQ|Hei`KOc9<sjLCO@IApt62*&Gj*wma
zJCa^eq{c1yPYaWtM-)pae(-{RC17s-1YlY(AI{@*ylgO>ZC_2gG9_{8I?sHt(`|#>
zkq`GL!<NC&aQf*jBo)GOUD*m(X-%lh@5&+g+z@ICGVs7+4Y&^-OTv@Tl8we3ACE5*
zt;b|%r1b0c<Sq216(Wbh)t$=K-rSm8v`HyXm6s8D0j2TtpP73d9<+}exTGiPSEsF|
ztpz;HoGahyE1GB<dTAJ>Q;Kjq&{IiO6KzX-UQWE(H2QQDfyf_I=7-C4dAlrW(bAjC
zshWb_uJfb|ExwEM`3NlKK|bJM0Mf^F`=l7kP?)2|WhRV271H$}>E+e(yfl7A76lpl
zj@szpAWbDx_*4Fvnj^(kgYNv&pzq!JuLFk8l0jODkCpO0p1lJ-Ul#mu16$^Zszi1<
zLDiG)Hl}JPPT9{#dK;A{E@1>aJJK*VEIq}cUpxdC`GT5OmcE<2*I^=qqLN@HKwd*p
zyx0lRhlf|-WpcV7o>dU4^03&+J@B_i&RC3b5DoHL@U;6u!4F3^=i}O~x$(TY<{6E*
zs63xY3Yu;OZVfGH0_NqR)5*Lge~2p6%B;fphbV)Z(YbWfF;S$Rrk-_Y>v>>njj#DU
z-j%{jjI-Y#EQ?VHik4KKqz5Rj{<*zhNYB(-D<jKczsvs0RyK7V_4Uihr)PwiqnUh+
z!@_Z*^UL8Hh2=?iSf=hBa3DFE%?9KuD^p=t%mkiAa|mr3o;JmImuWAZcVtUeJYVa;
zmQpdI<)1Q932`qeIT8+RAAet9aaP67KYOd*N&z~pwdy-l<BbecW%E)!>T2DedNYOI
zFGJcDTFLd743=;J;5^^xC-Tu%>Em=ZK6CFEmMQA>Tl#4_==0a=ZD!3@DpyPt8&ecx
z3!T`eG4=AZruA;ILdH*G>XKe1ba(ptLMBt^L*Lii0(EyetDh|`%(JTQ_|-1&DlabH
zoLTDb?&hiOJk6<llBRlN1vPnR*h*S<<YLPBlt?CLTzB&1Ld;NnNm1X%-P*?kAJ`|A
z(Tv`A&wrgkdu;9#8*WI6D|zq2ZO7~NPe}sL?j#nulT|v=+eQ*5fbZjyx89wk!(r~c
zn;$YQOx=~7Zf`6w^gQt{{nq)_=*6uIH>hp7=ZjY%tR%pCJx_e>P5-TsFG=wB8DyBa
z_eW&d`4%A|7tP4m(mJiLx<w01;Djgq5oFNrYCaH-eF+qD`g{+#*!xjOvrI(N8bL&<
z=<~3~Kb0w3)S?&WRg|3C*essVQ?ju6@(JM&teKYxfN`su1|!i({kfedof;lY@|Q&u
z86dc_xV@+=z<yt^-=o*IpulcqJId!G572)dj~~|#W+j7`M*ySu3*bDT1OH#_y=PdH
z-L@`_BBBDKRH;!wP^u!mMMau4>7A%_Fd#iZAOZ@~1q2iXL_|bNqy(fT(gmb92|e_l
zPy-3^JZtUqeb;{7@7a5ub*=CH&b5z!`0+5G%w&u)$DDJFd)#AKw2%_U2m1ZeZ$bPk
zEdo6Qcx5<{O~}Ue90CY30K%S5!2D{C!3-o(vLZpK!A7=OL8a<FYGVES=ZTL4i_eDa
zgqLL5edk`fpA0nnM$q$k((2Hi%KPbMNrlG9<4S{sh6UjgzU0_PyH0%-rm}mx`$u%?
z*+*m)K%Fy)@PHtCs|$dRM4b?dG9(UEmLc_@kGaY9d%)jh^;-`AW;6c^IY^rW=5GRC
z?M57(QRtyeq5+Vgw-M0zEfqr!EVOaOzKl&8TT+OXJs-yU23zyO^F`-kY&@-JFF@Q&
z7M)^1G~POjJjw%LhWhIXYmEHg(kNgIaBi!B7moqR`ZnaItdj?R(ZAD@!cLIMnDsch
z%*)s39x|G{(*6M9;5}av@`$O<gKPahhgZ?r3&(GKV>KeVHDh*YKtx(vApT)AGlZ&2
zzY*}xM(d11`#iP<%C=yX_$4Ft8*SG0n_PzfeNx@Ogc!F{f~|f%C{yk9JOORV+agAQ
zTj|ob)jFVfzst;)7)j9ALDy`r`RF<h_cuz2vA0L<j9I+)q7r+3*{_o+<dF{*qXH51
z$zoYWv^M>4gHA5Jy;sYVuQjjkTeqaXHT8jMg@Sk~;0~WpiV$ccC_+T>9!&zz5o&ih
z^V3p~zmlV(Ids&}XOQn6$NdwI{9zLWJuowRnCy<xPRLa_d6XmNKo9|J>7iTXEnRbG
z1{_A1ajowo#|enjPb#a$#cn={8|_>B<k}?<EVDX<36-!X_{E21DB{Va$ktm*7iZ`T
zwLQcxIs}g6+g8e!Ky;$N>=W(sWpV^lRRo-8XV$VV@?M70t`UyP*CDR#5tsZI*Gcl#
zypBEmh|BQr02>*-&%&TUe_L}LH(WWnnYy|7?NJuEGPrR4HojS$e1vh5<^Mi9=P`$T
z9et0aG-<Z8@9~qWYOK#=*kP|as5UTY_<$_C8{{+j`6%BfX2Nno2%i@HXot%tP@ka(
z=a%8-Fwnp2u3Ls=ejP1UYphwyLi6=u{$t*Dh!9*JAG{d-$hVg99BlD!KluXOyTFZK
z)r}^a;>}ikxvQETsxnnxD`4n(a>1zi3Fo05g&D$#=nc{VPB`BH%C%2!)4H4aa~Laf
zkufGps!|tOjXJP89CM6r1<uMNO;bC(17N%)m?E4&)P*bmq`I|!1E84Y{~L+isnK_(
zm>FsO{t<0O51BsU-{|fB`0-S~$NV>xPmxuK(t}UghmlysQfE5kSIa2mn`*(blP_Do
z%(on9*0*zwxm1(j^5AmAMef~WY1_Og$@<JE2kd(;#4a2%W|B@<wFS(oa@Arrd6nHy
z1QbE0rRh$8p2olpSr*CFy>M`I+!B(OIri9XGUQ^&VC57&l&l+209PP}0u(m%Wo-qp
zbY)2_myTzR`LGKv6*dCC(Znoa&1@Lb6+(^QU)#pFC|v>&6Roe2>x1H(=+Rf}n33=O
zk46_T6BDtEZ=9&}U<s`fKc+N^#f7lLdCFCoo&&*}^2i@8*lRZ|t;})lSXk>k(%^nV
z{PC7APMiMk$8#@yEJT|iti%pCgH{(HHH(6K3kct-WjALZe*~%*@81TF_CKUlD^E3I
z7>Qu5ks@f(+WEN&uMX;*Qvwf3y${A*7loWpB<KwMq;ma9byj1#hXv@@G|5c>hgL{1
zVpW61DAi8PW43Bn4j77CD#}$A*@SCZ_fOtGg;X3sIPri1P6~rh%V2h&F0jF{(xneD
zOOE&q7q7~NE>`$nybG_q>CN|2I=RQ<L2?g5^C>M9nz{kBz$h`iEd&(S)0z%NX9}iG
z*Q@Z*=~Ot|*lVX$lI}F*Z=ctC`1Sl3Xp>P2?u{L?#e;lCH+VHIxUx4oaH}KG2Jj0+
zdS-G!y)7EOI1R#TXcCZdKAWefh;Gfc^dl%60k=RK<nv2a$0-9BZ?x`hf6qx1sgXY;
zM}KzQti-Gna~x>K0sLM#KOrysxrdSweqq6iM{x}xPC%Yqt`X?hgt<kPuxU9siM%kC
zi8`xiJkkdS05U-U+{1}-QUk!{m=4U7R1g_eGaf{)D+w}nlaM6LEP|+5pDx1ntA7Lk
z(Loh;$~YH!{A#ctfAMuseh>OB0^z?}1mf*~15O$U9kw3!-DCz*eSZtFzN-R2&mkzW
zKdrm{N6LNGel7g_6DrPsF?RGU*B`^GCc7T-;5lmo`%wTuJik<leEE@jK!pnNZ9V{q
z#WDhI`S^?LRdI_69(4}0aUzYACwrxXWgFA3jp=(Az_Q##GSY%H-4oKz{OEWi+MvFy
zuDR%3r2p>^q$@26Ynv`O4<LNRyo+IJL#jR~a`i*nv1Q|eRz<8jeFlni5HeWoJU?Pm
zgYUes!;?0V=(O)+RZja8R=eU#80RJUY0_<Hid-EN9bzqxp-w2DSWeP6{9<1B#XW2K
zQKBX`g9fd=d2wThL)kQ(;P!P=xZlXQ4`{-$Zl@De@!6`A!SM8qTiQi|OMTxYuU|aT
zt*1^CAlvNmEQG&`pb@>PBs2pR-vqKUPwi9(d^)e@pkV%%E9^||+}AW4rBmf>IGF{1
zG-uiBv6i;DU{2N9m1q#s%n0{JDq4of%0NiWVP*1qh6ej%TMMlK1;se2Qz?9jZy%eN
z@)N&ctFAU&)m>)w->N+V_ZWD|tYx8BNP`~4%BfsSpMCu(5H}o*pEB$dLrUy4u@0Bi
zIOZ!I#y%ELDt7YpDc66qP6H*<n_b#>isjiQ=1o`PT?jCOUl?5Qq+Xr0Tu4*gnD^qw
z=>C`0k?4bReY|WmTmwHc(~}^C7lCCoE9A#U#_iU#aTFVxPX{T!Do)XDj^a|fj!J)G
z4)1&uuQ@fSL_0e_0A#h8&0$(a623@wwc-yk9!D9zMJ}eR`Nf6$k4DZuN+P_Ey}`8c
zPF{d3GybZU+@vHCNO-yUF4mzao8!r`<JNwzl$wsoG%c6KDTGe~E+PCf-i}zWWRnAa
z+UpJt@V%lN7B}YQH1QeUe70Us<TB-&i!84sK0f7xwo{=qg|FY2!IH<Xf0uW{uF~y6
zitw45&(+=2Ul}sD+j`LKL@<F?yS$#jn>$l%1gWbeMzKdnBos}!XcxGeVnz+H9)e??
z6&ABE)6Q71yiz$3xIReCBdGuZMXnMhem<ffIMOuH+^yWk+{$`Pmlm5}U6DGeNSx(z
zvcL)j%*NIvap=UKwXCS|pm2?2y1T~oK27-W69j;!IF{rj1XwIs!VowTDF?Mp;S<OT
zy-GK5W5;qV-7sUa6rrT<)9=0jgkk3n*v$BDQR1!qrcXVH9d3K3BJHxz%R>;`kL}K!
z4$7u?rPUd>+%!@z4Krdf%+>H&z2J<M$6YMrcX%#*s)YJ35nKq@`&+E^v{Rgqu75+j
z(9??5lNG&n7Tyo}>f#p&`7yrFzB?Mz3FySmUhr(J+~qh&zD}45rGPqfT+->&yIPZR
z2OiK};Qmm)fvO$jxa5`gxqkHk;vU;kxh~c~lN+s0m1=NrlWhb^NljYcLonQ06_WwZ
zIhwVbdlg%OyoJS-J_w_aG2XhVMZq-J0k^JC61t22JSa9_m1@9Ms!0BNX8uHgZW}zY
zM8$R02g3vNuo{t&vQcbSFrkz?)UI0w7q1yw->P^9B^0t*GadUjod}U2-k6bSRu)`Q
z_PE#U+eu0#CS_gn9dxfqET~Ujde@%+!b-}ow%<%KMPB%6!nJ8sb{O~sJZL6?2@W1m
z(w2?cSXa8Re^#0<W1T@Nz~XJ{x#3lss<FK76YtBcFW|je<{WIWYGT>D8~Gl|&;asl
z$v&4c!a@=`X{WxYXz5*a=(GQw&T<dLn*9KBH>Z>^HQo%@jkeEB)=6%D)nhT&3t-{4
z6^>WLrD~mH&W>fxIqiP@IQMp3cFFkBTy>?22i{}7<s&gkp9&TYT;pCWBN*WRxYJ6C
z?^Yq$1?`z@PMQwsBj3(MiL`cH7dbz;*E4K#K$?~q9mA!O`9RcYW>*SBhVB4`huBH-
zBehJqqk_0<%`c}rE4{SPX_=FJCfpbIr8{tT=S-!m5_%O~GBQ!9-nCx5*%!z`8prM(
z*lzWyd>RHyXWqX37vt`azsX+ZT2*8EdvxWP^53iAT}2SPPOFB!|7<Ss#4nje)ZN)5
z^TVMD+9S2LBfuw(nm&;NwD(Upocyy!l{0mo0McghIcBwc5s2=h;CT>J;w(k2=R9I0
zvxgGndT{rNGUIeeHd`L9^UH^TKDFe}GW$%oN4NPw0T;Nv2johUn<4Bfyf8>JXLnqH
z|LoOGDaSj5_5(g$frr@M3his#;uoWEkcpb>op?sMKW>MsicN>K3DQ%sAg&yITXY48
zaoKCt)WBLiXXe0Tql$;`8+)`#T@>>?|FL(EPqlwLaD_-?qeo?ld53AN>;&H8oRzF|
z1i{jy*LiOKBl}LeRm1Gxh9o#;%QN4bJ6g!D-EX~mJxrg2|AZqDxk0nT+*+`A4oz$A
zWKenC#+`N_Cwr8bSnqdM^K`tCEToS=MT>&Z0Gt6j#7iZxLXFQg!_bOgS|I8Zu`^gR
zP?B3!V{T?p_q6?TgYM@K{gxIfx9)87_?3mpBO@u|1feJ~6){VqQFFLKcvFpIB2J8x
zhptT)EfnVVG<|vD31nP7Nuj&4khxL2wOX!0fB_^np1g0eyJZ-7ggxA27fnOf?xR4%
z5PyHnh#6Kq?L(&nOG0sJZN?fQIBV<B#Pmm3+LdPYFRI(<p>~sQi(3HD|5RoIc^|-0
z;oSfWX%ktPI!j!&B*3$d6DH8rg&MiUk+!nnu;vNP7jeR3l6!LJ)X&P?9mY+^&-Z2-
z+a(ea#jvraEE>Ex68M?tU@^^TVFDPSYBCNAb~4I35*mc|cNN>p**|<f{k?PR%C@BJ
zhDUlzx^Aa7VoTigz#@NfKmNRtr$NyA7|xZFZW92yk7yLK1GuAi!DP+}1Ob4NFqZ-a
z1c%ua26qf`9zK9iYvH|6L)@MYlAs~0kJLe2{AcP7gWZSsn`JFuiufhiq4MI*9ID=U
zq+KiNk?=o6^?=QW?uuW96#t|;IkN11q4Tx%-@-n6ZIqxD1W~#MNEf{PW!BMmZ_BwT
z=M_N*n?MR-dUzFRiw0!dei4^rG}V9oK~Jt@{`eC>_^2O0sd|F-fv-~qA$wIizoqve
z-dRH|qVs=Jg<Byh@)~45Kk)XH(NC&|;*p<J+{NivsDFQ*f4Ov0lA$v(G6ct5_RNp#
zRwAY+T+J?(*?*xGTd<F(Ju&Sa|Doov704zfb{Ydv2|E2F4)2@1iUlFP00EU$Hk;Z^
zP$81H4R<v+$yus&W|OI6yET$gz*($4@fbzAgEUftC3@)Bp(1edDBpNv3l{B<lj_sw
z6RSCzuI`>Vu5ccrK)(5t>SY}K>QAcD8^9cW@4g^ek<ZCCB%_et)KXY(>v1qWOksHY
zcArmN;2svM*e+%sMN=m!{EhVDf#$_j`4h7jNP3o@FRNUEXe*T$c~IM6q2Nk!ngm6+
z^FSW-@E5!OWi^F(lI4%yD_P)>=i$tKb(dzLsx7>d<wTF79w03JvGcLeP(N1Hz3O>S
zlpvFk*4ddhoiJaFBP5zvk{}U>K39Y|SDa5!8?T~nskyu#;|sOiB3#qyR7u@a(R+o*
zOkA4Pli7HDt0a%EGwqO=$8{J~6{72I*R>%(b~qmkBtrr_yA&%z!vFJc`KZq#)$YF<
za`VE}FY7Adk88?G>fTA&U)ELbFH79Ci{)6{uge@js{O%Nq?D^s{r)=t&z|vHI{$Nf
z_?zwM>F>tvzKa46%MM3VW&L#=M?L_l**_8``A5e*o|O9qPme5w|9aHB$s37RQEXST
z@Dp_kn}n3gi+NJt&4Yg<e)V0HTLeD~LX5&yfmDhvkV>H`Lu{$RdcHCqa3jI14M3+l
zd*4ONk+2F0R<bK`Sb}mQA*v?B)$j3`TinUS7(Sm2<$^^UYYn9!^UVn+8CeIfapP*~
zAhBeGxAw01b;F5llcTg|@D2-*wy_cdAf%?Uu}VTX4>lMOWY&yniFqrH3vjyEUokCi
zXC2o-U#1YIZl>r(&GcU8)HVa1IWg^RTMk*a9|b+Be;x*DH06w_E+U3a1`Q1gB4k*j
z#mX8x=qx)d{iovEU$qN<u6sbzfz>N1&M>mMSo$VJkS`Z5v84~Dr<T8ixTd0g@_i}?
ztSvsvT;+TOc1gX;?{Pzp>y(0mtJh=8!m&=|^nq>m0`k0<<(P@n0r;pIp*1$K17=vs
z;_7=Ol_|4-7d5aZHjAnZ^;!1hcFr+TcUPpM(JpX4+$+eRu;?tZXyIwrW!8c^DM^03
zf9uNfV-Bt?4$X^O3~<6Mm<tAAWM|c$B}8oOcrcDO5zh2C9=cg^SnArc<M*+MejA}%
z;wR_0LaFU(K{ZWNL_30AYnB2Y`E&<KV*%4>EVqIwG&P%*+PEZ3Cz4d}CR$oHaT5+s
zsq_fii(R~T?~%K!{lVKsaCW!S2no5kU20vII=_}ZC2m5bI0a2-hjA`GCEr8y4b~xo
zXT=`Uu?>ip_>!))msdJzBuTL?Tx`M0#b{j&U`)lxS|1AiVioYkl8^yjx#0@Msq;kB
z&ZfhVxO_6WlrYp%QVvyaX-0cs)3xuw=Htvn++{jUJ_r1mK|}5kg?uE?f^diGGj{Bp
z!`TXjzF4esD&xW_GJ{K}$QLvaQRy!xqMa5U567)54}e`3GW`Z29=nrJ2GqWWr8`mR
z0L<rHv`Qxqba=)kn<$nH0E<NyAdxC^gg5M!Asl!e3J3)`=Qb%lm~Ak;E?8^cQc}Vy
zqruY0M(^l*@pBtJV<!r)<Uv5TsHotJ<4b3T`jz;LT@~2FeYH@UYxAi~M@aLet719@
zQaZgN^iPFS0$%>4D(d+DaIw?9lI(Gqzt4WfDg*KiwXdB<P^l(K07bf2KOm+n^(mUU
zYiJ;s^Y+7Q075egKH#hKXhLwnU4Vl}4VVs=<5N$kSA(uCFaKQIi!7e2iI;MW2!d*V
zPwJm}Z@kTb=1*jO<q+%SkM6IGw)s98*RpdR{fA$A<qxwJr(Zo5w+A@~bPhEFa;6QB
zvdM`Wzv$^vAF#zb5EzeWq<Y9~?EtWwCHRj^@Nb*v?+^JG$b>Z+3M~Q3fGj?Seq6>u
zN-b0E^4lwruRWj8W;f9K%P;bW#s?lYR{=g6V*;>P2Io^^tMi2Hf1Le;R)4i2G|f>Y
z<?I-7U`Et3Zm|LUZ1Q?VS~F*~E)teBhj3wGcjwyI(0L`(@Z#<F8=?C_fm4wP=c(t1
z^}y9oajFWgh~Iol;aMqn9u_HX@PBmK!Jte#Chv8sv`zcgdr^DvmC+Z2)oH@X6kCtf
zY;uyC9NM3iJdRNW+97;9MEqjVzpO)IH)1TP4G8cF9P0k)CF`<Ig-b)uEg!z*Wg6PK
zVz|;h@nN)b&uH%pm-~rC*3~h-k1)_)8Id$AW;QeYinb{zD1j4Zy4Z4E$0t+O;pO8E
zK0!#Oqv@lI->$HGgEE}y@7`6wtc*{QpQ1OV1ga}3Xb?pY9I!X_i<N`^VBxF;|7Qq~
z9v1SQm>Ku956+cX-cp?Y*>PC7=^#ITArE(@cvvwCnkx6GDAQ5#{qv*^(!Fb->(ZlL
zByG6!pb9%o(W;UN33#_+W?4L(iQ_6vIO{6g+9(&f;hr;aENsR6gTOiOS6?qiN_$zS
zVKhA2QYhfX((CZyDicIg#@2es*ix|af3c<>0|82p7OU)7$eFdeTZjS}!esDS^Kg|B
z5$c|cVp5-jUbo@Xr3<a{aX)Psm>6!Gt7>{;!|soZ*Wi~)tfUM%elO-f%INR6gJ0zI
z+e`jKPM>A&|E>ayB?R@xd^y2ktjR%{#3jUQQ>%V~QBozu&mf^qX{5+`@0znhWp%+1
zqP9^O$?A)C;Xz0rbwL07t7j@18t9(m)Eb|tSnmpCUn3p^$d*$Uk6>U*Y<yXWC|qxZ
ze1kyU{Me16Fi2rrN-bn0^VdA-eOTS+v3;f@2{KW~^!)3qvVezHj@Tz*{wJB{0FprD
z>iqniq3-xj0uNcS7cF2rX6u&4IFi0~%JPS`%OU26K<deh6-=o5aYu74j_X?0r`L_)
zX86d)EGf4K1u%9?Z5Z@nsZ5|>@gA2|jC1?-2BmX0@4cIAU!9say~h!8t(hANo?U!i
zgub+53mSq8c@|oj6OD9+L&j#Dhj+y{HXKt0HNn9PQlj6E-%t;|<E0c^fQnDqT)&aG
zR0DZ_c|?gH7sqJc2hw9#VE8UmZc_uX4$i`AV(FqT`MOYtov4jy?bzwc>sSXZPax6z
z+rvj%s|fuTL5%cvZ{HAw6XtJ?$vRwYQ0TDj`qWTd-%|2HY>xfKSaF?~h=xy|M$$XU
zm}d^67kJv_fo<~yoadwSA#W<24ryuWTnQdc%^p1*v((yC0r%x4akT`@$Z^CQuILMR
z+iK1BU?I&_S$x@A!yXdo)6S+Ll4X4xzN^bpp`xb!OItP-hV7zh_e8TECoG&{w|f^M
zR?Cytgo!`DR0i8PTx?^8>*G%*c;utAanCQJQmrzU=aOVUM{CRFhRrn5J(js5JaWu@
zpDD?63rHPlnUZ<ZCVmtMaY4#RCH-XI0VpXe@*|Me+{-_Uq@O&7HkY<JuO#nJ&wj0m
z%jMhU)9#<rsMM*l?wz6f;a*2?KmavnpCx9S*yznR+Fl;m1(mnBT%L?Xw&!(v(g)Rj
zS#Y!Z@x7xq<GAjRanHd4cd+w1mnI$vZzECb)Ykan!kplQb<30Q#X^P~Hz0z2NtzqH
z<3(l1Lu0>QHJvm%G2js*9e_ndikH9@XBin*WYgEujV-zNhWV_jQ@ewnslVvJSDY=K
zw-56`pVxjOK$YCrzLK_kNKtucR+Wkg)8C5TDzk~p6ltgtjP!l0pATc&YuVwtx^pBA
zU8uTK6=uhga+@x$_)8Oq=+`%4-tFDgaAv%nvskV~(~+$R^#===V-JSl)lAra$MhjX
zwc4|B-0|H4ri^dk<MdvLGXTm+(|41443^unSA$M0kbSl@VY{dift^^l*(h-eV|L{a
zxa=aBc>HOS4%NdM^Ljt_sfrJYIYfjFdPM0uPM6W7&v5|g(0J6$whdXU|8maiZq}VG
zHNG3EAc1F*sb5Q`0iMp5Eg*Ikhb3)HO)YS?Yz2nIg*>eRHzF9EZAB{yJjihPZC??K
z%ju)yzO7$Sm8tg{Egqth2s!Z@XG!nx(8JpV$m>}@sYrn&%Cjzg3dIuW(!27F=uHqF
z!-*<%Fo2Bkdj+|+N8$JSeM1?sVsqCuB$FPbnvYt_+ub+mvF1K^bM1<MKTxK&%WlN=
z$BzM5Au&tVKumvg9tlu)6{ic|Z2bjzoUp#SdB`peCLI7t-L}HNg|UB+`2SoUWn#O9
zJ;b}aLX@wA+R`bXUNxtVy*w>04@S41)mP1axybGGxx^2LcR^nqDH+v{2^Pm1=AuWc
z<&K0ME?=nDwl{t_H)y)Hh~^~9suOR`gwAZ9A_^Jx`*?av_g0eZjY9U!B>3Hv<}*?S
zX61NF&9m`&WELL;w~HZMwmw^rfEt@}+Z$o`>_{vBOois0J8+seRH}b_?o`79+6vFU
z&P>pzoM#N3V)I3C=oSEX&GSkQN6&(1K%cRBad|~sUG%)#MYr!0UWK!{oL!sc@jj$6
z6AoEjgiN@hUP2_?p!4~OXF~RYd`ew*usw1=e-Zh5dmK@g-()h(W%>nA8!fG@R#6+r
zV!eVEu0i{hr3*%zvv*^HKN<Xhf8m$EpfqwT2Ywc>5rHJ%2>L))7+f!b%jUMfT{c9x
zoFpm7W`~GM_c3N~s#x`olo@EPVs<e==!ddD5f@a8utFq}((GZvl<ydZY5d}|3T;Tl
zVRxq6jPGlypg2XVq~X1;%Z@Gv?zno3j#9_V+KD}PONG`J@pxiTlY%s+DGs7ji60!W
zc9G>Y{XA@zTBdXAb<3b^l)THagHHL0R_{>Tc{Q1_p|E`)jV!V(-Y*6p(?}o&HLB2{
z`*vQYEM6DDYi`|Lln_W#Zc>VGD^z59ERU<n{&lh4F0;BYH8wJ}BkqdW-vZJ*2?j^?
zKxx<Q1t7tJ5JM1mfvfg+k3YQj@t5`a_ngu15&v%~#cz51v9tt`$A2f9!v9#i5hp}O
zXi-QApgLQ{|359gFf%E(oI<#{Z7qRWrQD!5EFTEegw#?EKFbKuW~NhTHwe&dPz`Vw
zIM#XX>qFG)prNO}=ZE2s2ieBN^KWn2r#nTU1zZlr<FnWY;8J26p57v2aZS3^QZb&q
zOK&s@U2Id6{H&(&WpG)N4AD&$!mL1S`+BIjnzZEA?lz-*oJX)jZ~&RU`uM4hrbQn8
zxpH}aj=2XKLxUy-(&Fw_0HpW3jeMXB(L$n~mo@-Drd2~M7mCqn=3$i@CNNY*fy?Qy
zIBw$xK>GIEl)kvw*m$u23zG=V2URu1*B&zoQMPAmdgloy@fkOCX(ozyWbc)aSvN@P
zpUBZ_f+k5_*1vaZR(i?Uc%7AXfI>U(;6BU+twj$XxKSSApTpVrzQIjCxgnU~03HWp
zTY(zLT9*0}k(%na$zNP+q3tcApy`<L{d>bJaEIMj>BwntD1{4elNAx(#>*XA@so-b
z9EELDEst-~N$`DHneB)|kyMW|!9fem_4_K@p|$gedeI8fQ{>Hqx3#Ex1gl4<zIbG3
z2-i{8j<QITmy2a&^Mvq<rj)LP#(;-t@oZI)`T)|gw&KhEfo}qj`_(VNbLz^U<b19@
zjBMi4XT4Pnf#u>%!#~2ZjlcBmPq_@oxVg}jWCn`QMkxAMr-b@TZuq_2ntpc7FFVKi
z(}KSqNs+u6bhy7c$mIn^WAcg*B))HFpRxk{Hd>U>3S0(GJH7_M52s~vXn=UyqI$OD
z*4lLCXl=k$HoxEF>$MkVg#h5dABsdOoeP%I+$3sm0;&D~yf9?+T_}*3`~$2TfDxQV
zMq-wffk>}Cf^72lxa#jAEq|F@Os^1RBTL{~g{M(*^2JUA>p7w$2J48dS+7Oq$4y{^
zRhvwfbVE}+-FzRo%k<Q!hCY5giweIEK=L!WjEQd3`scK8V(e-pJ`K2|=qA09P4L^o
z_xk+Tn(ykXj)5!?jczlWhm1f@fdd4r4=X7X8|>7Gn*YR7R2fEgw-%A%C`%SZ^8(@h
zJs`Xvng|^JjsL?-K=|z!L-`wR=f5ux__tj3A8hIO5TI=0w_N_svHbsO4F6pc;oe(_
zww>(Fs@($DtN)WReej>=djDrR;{TcV{@Xwierw?0VUPczhFAY;FvxEW{MNuf+gIuL
zKJZUU?w|4bfA-Grec;ce@>f3lt%1K%+J83JZw>sJRQ}3mzcuhzO8d{|`mKRKlgeND
z?6(H~N@@SuT)#E&XHxkqpZ(UrUn%WBo9nj*{!A)=<+I-!_$#ITXLJ45z@JIwuYC4f
z1AnEo|7@<`8u&A*{FTpsYv8Yx_Mgr5TLXV4mA~@YpK3s_`;bHYL%CVkcg>`l`}h*`
zt37&mW#@MaU;eXS?LRizc9?{y@(jv6B37sH8td?yoQaTMZtx{jPz7&)QhogLld1$m
z67D^j`M<+-^bd^XCTA0lgh=Z^-|(JcY3*R1{F)jkbJKg-wQ=tsexEU>xu<pc)mhp=
z00??IC6Y{9!4QRNf#P0c0CWBTC<t@_{-_)xDW^98ddeXX{<S%x2VDr1hz4OO(l`I!
z0AA1(w7vNW;4I171R&~v1J(Tgpx-k3ZAZVC$Zt3Df1rot{5uC6D^2E(LS~U{YR`g|
zHD4$5bX-T%#px|M<)ycbnHVUN$(6!*luZ$L<2+j!^v-p-hWE|z{-JNvSkGzMq%gxf
zmtIaMHZ19y;o07#L!M1F`%`F%s1NlK;WF<=C>M#5(p@Q?bPw+yc&_Zz)6J#WF@Ad-
zVtTw@5tr5&%s-nNH>*$WHX>V@u2-68-sCCsW-0ljRZJuoN8d~Ck1P4HBmO}%T#B0~
zdS(Iy-neYLtbeG8?p%2O-PYpbEZ672S&nt{cHP@ZiRS*Q35g-(V4vK!k>RK;f}BQv
zO#p>soJWO-d3d`P{Q#u1%84&TXrbHwA^qX}@W<=aH#}126-(?*R3klWQT_-9#B@##
zC?2jcZ6#8C<91aA;Z0IQep4>-6OV>*>C5B2A9C#ZRNWn@<B;ORm#jFt0@<GdwD_A;
zM90~NuDv*jNhQ7~VxvxkCM#D;f?*=1!9%u(jw$uwh1ddv_@R9K*xti3;)8&hodL>4
z*v5X-B#22cZKklx^!kJ2*De{5KN?!}9cvWb(qyNOr}1&|L*`A?b1zUxa$`N-C$fdI
zvJdZ=O?=O!m@3Kc1>g}e7}tc*{cPqEiZHRZiC1=HSFxa?T57`3V{M&DPTW@REoWfO
zEd^-8B2$55&m?X7&QGdXP(}80H0>mhtYJT#Ur8qLuJunUC^5e^6(~yo@x|n5&d=|i
z{G8B5fGy%jwH!@pvN)2&Ow_+O0M~kJK`d-J1cfQ7!Lbnq(?^nrCk8$w{*d}q)GI~M
z;HR(f)QR?j9WNm!?<wP3^Ko5P&78DRN?Oy}6ISYxx6lwq<<}O_OZwtcM&8gHcWyuM
zW){53v6$YiXI-GGFSfpcMpusb*IJs7P3*-Rj+pJs3whK~v>ob47-xVcazPfvWSmFD
zq0kc*?y*&NfOK>`i*H*Rc@E)pIa))cNdQ`geMt2A>cG9heJn|C-D&ST0h1uT8D<Dm
z)vm(#4%OKN%^+EvE#v0AuGLf;ELa7+8zo$zg{8=p${D`3j{$fBASFbh*^wqqC9V=!
z+D$y@Ssj=Edq!sLgiBBGV!O}Qa~r{DL*b`Sk6-+X#a!d7R~g&8j8{pUL7Z5&R4>IT
z^X@kWHe+``R5J`RL#!t|PNlnrD@kfxg2$4>FI@PFL#hv^j!%UZ9C(S>)GOE$N#pSn
zAvWX*<q-Jhd`TB~&oi*Q2T_ZB1K(r@0KbKRwuBil^lW36{$3@(({-S?ycc^w{k{A9
zTH`DjHcXR1X5rjo43|YPOK(Q*S&E0NR8DgD0UUNulfS3#?PwbLv=kdEI#X$J<T*W5
zD&OJ`3ld}BJA;QrlC@@_>OC!4k~ShuoO9?){dEfamBsmE6L(2d{A(1q0`TOE7pEII
z<R{CIV(VyK@Fpr88O5$p;VIK#uQp@1_qyth2OnHaVGf!@CW)^r=k;l6eeX!v&jVeZ
zDj`3qJlsAM<`flwU$4?9XHAhD4wu#!FjPQ^WS7XSSwfw`bbUI<@Q$Lh$Fs#}<XgN|
zA8f7KTbDN2<X$O&JC^ziPK2GNm0skPtRd*HA0z3)P4R3Avpi0XSf}th5KnR1o_MbG
zYszpvL{jlJy<Q#}rFKm3W~Yz9NH5}S2%GLSkGAQc1l$AH!wGTalXg+k2<B>W|9Fj!
z3&t0HVSDJ6Ms<Rz^j8<~8DP0%#`tXrlNk@I(2Rua$Qjg7zE-YADW>A$?eDP-nF7yK
zXsFM1cAsQ0ewTlxW!=Sg;=ln49Mj#bZmH`D8(K}u034yYh><yXwb1(YS;iQE6amxN
z^Grb<Q?Lx;sJn$@B&zv#t}0eVQyP5sP3x*9)%{Oif8H+772ZB_vfdqnZ5)`f$!>{S
zs{@s|wlz))SKxPBhI4UNwPW6BaSulqCRK;yq1(?54lYh7r|;^Y4r!xM;UVj+IL6pR
zxc_2H;GjNt<+e7L#n+8<TB2I>krto_?cy5UL~3avvLrgjC1_Uhh^vKi4L@0c%M1aX
zx1L%$eEC?5rdhi}J-p)z{`1Vc*Y>8Hz4sSh)-UF!mQ`)ra!QC?Ro)K?Gl+H-nX#16
z);3%8XI<v~z6G=z3mHyanL2?Z*cQyh1ieab8MYJv<0IiWx1l8`;*VR4&A+mci6swj
zt3Ta&5IsonUoS3Lh7x4W@%v4=IM4~Q{2;j8cf!`j<<Ul#aT`{p>rwCLmv>HeAJ|Qh
z<%>suQa!*>PdO{B$LACilkW}sPMx<TW*85D_)@7|SXk&M#+0JT#uDpL#&lUoInbCQ
zeiRMo!POhJN~+;NtT2QsA*J6I7Soh2y>sn{t2gzw;>NQ)RY4Vm@STat+{yAyA|#b-
z-^>Wx17+%ycANM~b$4s5yITgGu{(s%QaY+bL1~+LWfGbo3~vdbNPt*+48&}^G7@r$
zM2~dyWzvj(5uC2`g&rnO+u?QVB2SkDZn!X?GBQbD8E%ZO#e-Wwv;eL0s7qSX4Y9cS
zvMAZs3p3P>(~|k7RN0oRzCAw{zhdXxnuXXcTeBJCd_?rmPGU!U&a}9=K|yKkqfK}2
zR*oOH{xbWvNdGLQ=dgw{Fv&>}n<x!$05QV(c|&K~O9bOKBuX6amFqPyb}(H&<u5${
zD&<1)19#tG6sgz<LC*J34&>0zP8{8V1TRlcAhPFzL24|<=`??GSLEI+080Ho0M!8Z
zNj*m#ZZ>-1ly|cfpqg^6(7N(}{lA->|7yVJ;}BpmCHMfuGI?o~R5fdfbrHhw;xuEp
zkM6^sCIT#J({$v~&K0N-M#i9U{D$zi{!@8!C+MF%ogVR>^&sx+lg?NmmhDYu5tTiS
zJVCQ&O_<Y5foG%GFnb8L<+Q_RZ?<kHW?Y%5+iEmDGwU)RAiz!KYO~3LNtfMOFe6JK
zZ!NpXjvnp#k@Xr$ZBysrH*!`EG0}DMK%IJ1jYC|aPE|P(H)j<kd*!%`K>zG{jWD2(
z-t`m5Di-<D!hcdZJJ;zFFfAZPr~Y;5rQSNQLkYV2<a@!Y*X}3pe!PzSmYXQJ`N|vR
zmU@(faD<+LjP~I4BPgtJW&B<OP8XrWR@(5++|s4XuTjlG<HZSU##eD&p|(P^J)>MW
zNP<L&s0G$aBc}}qip289Jou;`#M|z^t^-vrKQ&%cc*41p6|0IkSKtN5im?MKHQI7j
z^w%yaR#Oz*NBGy%9LNWE?F_7By6r{ajo|VGWHkITPMNk%Y$jNE*62%FpYL-aa81<Q
zsdWy7gkPiNtwV0J7jQr&VWBW@T;y8QtVDE;X;hUZKC>cuC5t)v9<RNQq+RnQ<a$kb
zloI+=6NizlryEHDeurRhiOwm2L0a|#hq9Y8Rh!tvkZxJL=6L`Gv*y~v82w5kfxAS9
z)BroI=hoIxH(BQ>5`AeRq%(veyFn&<6vQS57h!xx*2)nZKd$_+{WB-~3rv!I6xug<
z;uwdVfr%->Xny2QT5{bGc`t56d&;<eQLq2;!hjh<`_>pCB`eLfdqSl>d@qiO_&|g-
zDP+ee+07tnCRfvK&;`61#cqCkTf*9$X!h^+;>YC1sk#+S^P|v;aA~}Cr2#Aw2YM!T
z)dQNkH6*R<jveozW@0lwJL|>GeeIi^(nv>$AW@qTqs?T#JV}=#wVFV^e0Zw{@52kS
zxxKe8W-1RwDQ4a&y_QQa@uEm(mR7n2e);G{h~Uy;PpG((#`IFUGXIQe-pr^a?ywud
zZ<w^O{4UD#>Rg*Jd3EbaGaGmGr+%&OdzON0^(<4foh;e=X(?N|{ui#+Tbk{{fgQ#@
z*)3mWyJ{kYjo30R!M)0~tinaWL@M3-)B?)99G_NbHQpW7_T7N{v#o|)8(;R!R7}=M
zLIlP)$5}-RzgJa^b(iII&-~Qj>3wzpRn;1A@far0%}GpX1kM>#&Dy%4Y4uEKPScE4
zF{mNrQlIa-A$Db4COaU2x=3`=I3rKRE_H%;jXO!*+Y)me?3n%BQl%IV(Ix;#3llf?
zrcuX4>8sN9CzXYNtP4*$SOuxqXA1Rnad&ZWbjsgnaD7*OU`45ZkK7knAIS*W?h@B?
z3jj#b3dS~}ki1H#Nx-q#JO87cfq&IFFACt+6D0uR)X%{fN{r<dJeV96NP83x6lm>O
z_TA)WZ_Q?P4LRShr0SV#oe4HV+c0W|-#b;jcf<dq)+48;<L_GqMX8TV=3b2c7T?n}
zbs8-UgJ_}%Rq;?1&u!C)g{fv$(_yJD$Ti)IYBkxusG{jshojE?Qyq2}OQ!A&*IoSp
zTf9e54=)<aD3%1at%?ZPuzK8Qhw6)k1@y6}PDsAEMHP0cB(QBS2tw4XlP_}1KBwe>
zYs0UPV0SugEOCAJpM96S?O(*!#}TDv?*6hi_HmO~wfge%6heFonYrG_fbl&9WAb-v
zmaD(PWsfS9_z7Gswp{&1n`C*>kjqK1ov~oWW&@iQPnM_!NTtOmYL99@_jr=|r^?ea
zO8g{CYg-sXkH8*7z65$k27iK?ZnyZ!vekS>XTC+5>kv9Y$rsd{m)awtO`DHAQh^X7
z4V<wOFPLW!czrflx;~>zC*LxfJ#~e3U@fJA%gkO|>eJg7QA%aPo)rv*(TDzPi!+Km
z%#GMP>%kwF(IEWpV#LGBx|A~$7UM-CS2A3$ikeQ`6sdrBwdu%~S3#AbV&u)vIcs_O
zqa);z73K6I7P4Oa*45JK`Q%A_H>v=bte$X(&6b0pWBJN?Jw<zLT;UGJ>E&x?kVwJ7
zHz6m*5T&DcK%Hh}&DjJ*v9m2bo={&Q>rS~$7_{`Qb;xk2uSJzB4XmB%rgz~?J7y1O
zQjkM95gqX(&!u?m+EgS6XI~dO4>)U`&GmO@>lik^GHKXyTr`R1D^vI77j~QQu9}`G
zG!1q&i$BtOpw0xxNaqR~^>iNbask<SRASY;+qqt^&tDA}#YhRt92_W#pGijohZYFn
z;@yJD$YK7ng(-k#9cWjcg<Yyxp3^5;O}St<@^BNkvE2hRMd;JLyJM#$q)*L!iC>^B
znCEWmr=X~1hWu_ws+ZCptNLu4)U@}t52a5}#BS;)s1n2rW}ZJP5rojWH5m`2i@dg!
zPTa_i^l>|{#pyAg%<Wc3+bt5vIBM2ddFY;<FJ4%}MF&1>PIPN0-MQGw`X<6kF~XB=
z_e=C`8%r4z0dmI=DxVreqhhBGyD6%c7&$9uWQ2j)_!h?0!b-%Ha%-;B-)rD0i(Pzt
zt?JcTfyuYWyyk1?u^W?bx~CNVm)+l4eIR2_y{G6-K|O#@LSY!@d1CqWfp4|gj<HE?
z5>}q9SkhAT!}`-RKFyNz=VligEcn$;4NtY6o>r1288fKjh9jqgt@Q9iDSIey<syZU
zG~8;ihS}#NIf3jWg)V-&Q&*Xf_jO&PH+G_x2H5^X*hb;fsNuSLeS;B{vikuFxK0eg
zLp^=3$MMsoo_ZH|Q?u{p9fMXqEdc_B`RKI|=G02o`BVvZSo*#eKBZ~gIgqu}LM$_D
zHK2IUQPDp29)IQ&$#3f~&d#QeyR0)Wd$eK?Hn&1<lbAM(F<c>84j!>n6#y&b8M5aG
zq8|^JW@gE{UA5LOz-%raD`7G3+=YtKqEl0Dd^T<u@Ni~!ZsS`EF^A*`)1OI*ChNYR
zoPzmXnn`)OVtSbNJg01ML4xVKX-D6SjHJ?<U7zvhx9unEsd3p2<|?8ZzAnKy%=u+Y
z@f=rClrooI`Z#u~kt0fj`4;~jO-Dn#l-?~4JCtUm4Uyfc02WO$?yob#>7Q0BH~?d7
zLPEXiudSREFz<VfWlv|@>a&yZ5YRC(6S{aV*G>Isip-d_HH2>I7#XPA-JWW9ss2t@
zKgvOEp2ZS34e{CGqGXeIFu~rQ!ikC1h(?X%wL;%{tBX}HwJNXjIXb}zm^0jA!jv46
zYH*}3!C{2VPXe5+Zx@&P=-xQ6L3prfeD6uyIzT{4bsJlUB^#0|ij?kTSa~Ccsf2vB
z#O=_r5C`RimFzGmtYAfyZP%feS2#^mB?+a+1?;#>MTD=wV;7AKkJ6=(=O?v;eUZdS
zorHVYI)f-4&enxvF3RhPPnvG9MZND9d#A4ULjCrl`<n6<OUmBl=He#2DBZ@_taAC$
zyCcNjSfD%5mc+JAQDQxBnh1Y<cy$miyJt^0h3~6U;eYA|ak&xWePy81shyeIeOvYs
z%^nU_2PUbzV)J7}p-3{%$&8tpO{SF))~cb_(uHy9p`2<yRgbwRr3zSZ*Mht0tc$h$
z3N(av%q2@W&1e8_frE-@WEespBItiisPa4=c{Dcn;&aUUF`MqaD97{Hg|4VxlN=>8
z^gdF5!L(`c-~%juyXD?S5;%Qjab4x1Z2NcBb+7yM*Ra)g;VGbJkRZes_^6yzu_SLb
zS(S-iwj#`7dWN=D*pJ@sLbV%Y(clp{`%w&jlf*e|TQ)Q^WVo<gY-h8)7xOS`Ua6G4
zR9e>*#_?kIIvI3SpadpZwLT_3q}!yQaP?QPTIrRDfNQaRE^STVRPyIKuk`xX>r;@|
zR*Vq`0i$^2+Uh%)oqm1a`;j`eV%S|Px;XK%#>12q9+i%MJJL4;yExukRdc49^Y+$H
zg->@+Jqcm!nhw4>-ZKkff#vX5vYDpV3=N*7)4WA{=7lQzt100G84pRF<Fe$g^aKf+
zTC2sqs%@g*fW~$dbYl~-1@w9n+#Z|nfX&Ws<j-Vkx5{!tp3V-##in0GHU}oJIMTgp
z4;2@%vht)WT9C8XfNrVnf3AHl{TlCYOpu6DQG`P<!{#c<c@twUSVODT{TsKq8JwVU
z$G<nfD3nY7E;;|^M#cS?_qqm%J$u1@WS8PqSueNrh1VZop?10m)q#c3N4rm+^Yi)<
z(~ZBomX=0n#t1(dZSh5O<9V}Zf`eKtQSYl5JaLL&km9*1FW-CcRG4<~AG<|Y#-fzm
zt~5*KP@J*aG18;o6cpgR=~ag*$sRbcY<sedK&celO<QTN(|v_tIc+nhgTNdRbUrSC
z(-~?#PW#M0>9oL?H)5s2{j7}1)xDv$b*030<ZHN*!$+d{aF1F4POX@G2UaZb2Fsqm
zWEW?BbAKDBRfQw13oE>r1s->`F@w`-t!V}-)EJl+nS&YT?7WAMRiRJSNaw#;9&1gL
zs2RiSf2i~N;Tb$ef>uFA{3sjX=aNTRAU&k7e8BYYtl#(^ripnh>TbJ$wP)a>=IbEE
z<G+m=v$WnC&dW%X$X=A7j3B^9T3+6^DiD&xPpUm2pfU}YAU$-!PNvN=>;<xwe_8^M
z7BCSdniE%LWXmDl!+X7AJt^n4l!lrL)ny7rj*qrj4Fh-Xb^V#|lM+?L$Qc`nretRY
z$IOqHTOyx_47qrDi04&C>6DFqeQ1(&y?)~H+tLQQ<e+9q`E)sOyNi=PEe$#c*Cg;(
z+1Wkn{ItTB%9LH&BQU(~zPH!cM*p$D{HicZnIKNG%X+E_6Pi8}C*8xdl6{W3<+M|=
z3)cNLz*=<r@=@qe%DPb2j62l;rY~nyp!7ns)8%)Lp2;149(vjH1KmHV#45%+u(Hb@
z^|pslw{#g$<;l2DFsIg$z(;|gC|`YgRR>Q=qLnS@hc|uJ9eqU!zTAqr=Fyr1A#<;0
zZKhQ`Z5A6b(dh{-5|Nhi7zewVW#qe5m}qw#?#^IEZL1m$hX;R*SAme5$5RDXkHiQ%
z{d_Wi9jedQ%?k58L~=-4asSzf<(Jy3Xxu=DV(z&X`txH~y8Wr?0-oS>SE$(0oMf>w
zqMJ)`K;Vbml9j-TZC8cV;!g+1aVvJ5>Zi^5qAwrgS37oIqcb^4OW5ymx)mzFBpZ*=
zp1S3;O9pq4tZXR=c}h11q@*@86`5K_EZ349-lH%lv?%~&ZzXjG-iYfF54@-%PTxOf
zonb(FY9B>i*{|7j=vY9y<3?fV_bnhln2res-HU8ur(Eq{7e(dzR6WM>NmXvkS0#LV
z6Hsb@CFNRkB=n0i`%{b)_%u;pJ57fn_>;>3?BQ7BjWJiIT!G<hufI){vTQ!^zWDIe
z1!^iUU4U2Iv7T>LYk18U!vafd?)m6!rotJvX`J<=E8kr2tv1U^xpQoa{DN9f<8|Cp
zmBLLxQrY5I;8X<|))>4fyGP#}`+>p*B)#XVfG0nx7){z&Ar)<LN?=eF!(x+-GU!(C
zJAz3+GAar*>aqCvRK>ztaF+gfY)jiA<@F(3|BB+6xjUagi=^mt&9)QvX2)<6p*8GL
zh?y%SlgV5VSwMv&p)t)?CowLyU@ulFKs5@{Z$CE=gQSIL?X~tvkwk*bJ1@+{*%?k~
zQd6stjo>w7heMQ9vW{;B0#pM=%67Dh|F7dU|5+^Oo<Y7E%e{ZL1JIwnM@#$fA~ME5
zsW@TD_wbcTY3csYP#)}h4(Czvabj=0`Ce$f$S9c6&82$WQr1Kj`=W;n(N=@_Vc7YT
zN}I3vARzAR-1M+7S1gxRPH_(i-?J{_gd{6g66#07lnma&{C0PCO1}*1Wp%+M#QDPn
z?}Tlr-zyfr5fFOH=<CDlwk;5Gt!ZV5!BHfcW+teZ(NG~NKa#I*fbF#5%4OZ#eILf7
zsXsKEIV!Q9(Q3V6Jrak$K#XbJ3%TA$2F(z65{{3;7CI=pJaBoLoCo4>u*NE$G5Qgj
z=?j<qOeU{brD`aeN0=@p-8GC}THmT;r;JaKYgjvqumG(V%|O>$aH1Lvv)1%(rN#`P
z=G!mR5tWd3aT<6u)?@?=p>u|Rd|6!oa-Uh3MVIf}wAOrP_uNIp_e5k9=xI$P5WIq`
zp!@1*K7J?<!q1w9Ln^DbD@Yln7VR5QsJGWOf_An%zvHD`uCV*!C%yK84#)tLc(oP2
z>ig9;tRe(BT0(Y@^ea9hdsZRTu3K8O75KLB?u_><5Qg1Viy|@@zi))c>W8RdMscsK
zYq?fqrE&z$7sw~ml-=ITQapO&69c1MZhKD#k4NFRCik_?rS=~m;z@q){L`BlE+7gX
zctciX8$Lc%l#Cc@0UC}kPp%H~^ST9KD7Qkq4lE8qj(rejxD`GjL0X>|_HnZ4-KI&{
zSD2e&6`pmW3gHyvoYI2Y@qf{j|1Isf1!H!p*8P@O4rY@G=ghN<X85*erhpaH5QLr;
z`@V7{s3hAt?n?yOUwr)-`r?K@XLqk=@26^q+t+?H=vs&cjuAHJN2+l4dA`2DvVFF`
zJ?gJ0@qCI&EQA}o)M_JPFuaM1Q<JJ?P}DY$9q8_B7rY^w*!ot3C01MRf;D~8v4UBT
z*eq^5PuEc9PMR!d+{Xc1HxY9!pZLS=HNP$QhaWZ4xh`J6f9{RfjqSYbhA}YH$|QX?
zFjYufgXz?H&dpo>ySns+E|wPk0!*@_^EZcID6c&jmNjk)aB{wUGkGtbXvS7ch-o%^
zOcdXutXlQg-8k-w^-GY7#qsz^<4}WIq)Kz82*pe-%?6J1<oD7(13>a4Iogjru)Zbj
z;x?By)s6c}#em>R_<RK!M~33tnr+$33GoTjfIDh?4`b_WlQG+JUY%)2(Vg~L=+D`X
zwS-IF8M#hZRzkQvUM7!imaO_5@?Al^Fb^i%ED<Eo*tl()8Z6UQCoLN}T^ZRCb$4u3
z-&3mZPQPgUi7bopjoUQBI#c5>Ng|Z<eq_jads<EZB+!Bb(H^|oLFrS2+u#%STze@!
zc*iL#qF*u1Vz`G{Gr?8C?d6wRBYS8PSq|tag}5r%-Iafz3#QB12qEaem>@B>G$Z5(
zAM!&*MSW^-<XEp5xg3^psPrw|NvyT9R^Yl{b@rS@>ere#3;7#@`zQ(mV6@@S|Fl{!
zx2s=;>nDqkZ9)98CWhUIZW*;p_5&MY{8fpDB2np*X*b`M!WwI@Q6)vfj8f!i9=FX+
zvB71DF?Mf>AYG^e$!Lh|JCh+nOYAW~A+8tM$L1=J_fxl@)B32ffMAI@Prm({p!iiq
zv<gk=>#Yd)n5iI?q<Q<hsf4N&K>a^xylXf;=D?DuJg%kC_6qmV`}|BLi&&vn;Q?Z<
z9n#;WrCc0>y%RjYg_#Kwuc{v^4rWWwhLqD7>fh=LkVzn^yVO*L+7U8{X3c&;KdG!%
z9`L;yEAY;4ZH>0``*^8p7gm4Gc=(VycKo;(Dv_i>)LYPx3=t_Cc|h_DFLb?el%`dS
zFRI~5x2(I~CSc&Q&)mg#b3*`{^z?F;{)?+8$eYuM@W7r370@!9B_b4Ak(S&Pa#~hq
z53AZ}xgqAKRnqzC?e~vQwFE=%=X>Ax);nLX#$G;8cAuJXV=0=RAEmHs?_yShw*loc
zhQqx9>i-Cw@*PywbFNmWdsRb4Wa<EKA1nmx-x_H}dT5!48I4TBmu;GizY}^$7aW(j
z@?w>(Sn0GUr)b%2wV{gC&1&x4b7jJ%hQ{lUpbsJpjl&!j8uT#-LWg5R%wmGGW8|r&
zlDgOK(O2I8*-r9TJfoGEMjSBhFp_A2jeQCAE`b{GW@&E`KU9Fwidr$tu?0v4V{Dzt
zv@)0R5S%v^T9}c|XvrK5PRcUIW;Cv53qtBZ4-~yVt$p3N$Eb{toYjchskiXiWO4z!
z^!g@ntsY)3sg)@kuvV2aV^7vNSLOVr%6uIw?Q!j#F{7wBXT4%8nyQ?bDe)W*4h{<V
z&>{`k9M3|Q9dm;eySF|NZR5`o5paIo@pbmZdG$2BkZS+>j$8{R-wUfBC81=*lB~tH
z?7pGa^9oscO$oD8{rJ<d3$}V^HEnDGvepBHi0t;ZLwmk{8bI%VAX-eUcnoGhJlZ@(
z)+BgD_+%&o9iz2DQSZ~mWTx~8EN1C%F7pMLt*5#^8Qajz$VCqe7l(~Q;f_bSLPWFK
zIN7GpOF#t9-;Jk3A4v@Lc@Wqmq7E&x)U>?1!6H7()yd8X67ti~m`(_jnLIWmYV|+b
zyUwVlx^#^)ir*kr0RbsOC{n}*A_l@ms!^hR(gFcdfnb1$7>Tr4P<k(76i|vFp?s7O
znuH#jBK?CvfFRPOgc_0%?)lceKW1h*Yu1{3*UbFj=Q(HXz0P^}d!GI7_j&hzU<~*0
z>=iB^gt*YShA^teBvdLQhebcD<$=^-8WFFh4b>`fEs$rw1=_{T_p)1p6?@AOM*a5p
z{B_c_tBB-~bD}ZI&rHMOx^gCvZ=ao|+>1x0tn^uuSG-7R;k~QmPA*^PCn6B!Nd))E
zs;qEp7)OV(2Zye&ke-ZO7$DkXCJ!{1Nb445>ul04^F6e%wEh-t*%JzitHOzpy`1?t
z`V5IMPnCYK-fvicU8R2c{#p~Ulw93}qZ#5ho*|<mN5dXtSO3VQ6kKNO(U@;2<42w0
zgTzJYhLN2epzP6Qj-*5IAPj#<L-7{mc6!jIvl7lKCMa}VY?B{C7YFi}^>a*u^5>jR
zGE_n`2f!JNUJ%2omxnOCq8iAqMSdv%V`NDRDM49{kQjF;b?42da}bft5%3lq58d{f
zsbhIsJ4eaN4{~t%bF$(qcZ20uxZ!>tck<^y6LQKIG#F-QbiST<b+e=6RUWK12j3P!
zJmH%mM>vaZEC6{Vhi;FY@#5zXkzm)nd2t=!3~$d<%Wi$%DQ9V(XV5nry6%UAiEF7Y
z#wW>hn$U^p<e}J8wbxwfI8h`t59Wo*_mD!&n|OS@=AB;x&Te(POjI-#fl9UIi8UR5
z9{Vx4OICb9V?+8edgBFj0VnKWZiRfxUvRZ_OsorG9d~H$?ct^I!Gu{(#<-V(Ew$+L
zLPgT(uZ}wdI~vPt)OFounc$IhQKlaO1rHjV34I3>q*5*NM&{F%B8G0h>Uyf8?tX6;
zK3niWOR3Ly+SFWIi_#rMQ>|VRY9ZkEBYzvUXSy3#L4gl7DBrvwKBVQc8KQ%XBjbB0
z8kY9ip$DE=nd>fVqn}jx`40@f&%u(Kxe}4rp_qY!b!eVTtkvnv)>y@IO%2#|=KFxh
z3b*smV*g~#3};`Qa~1|74={`l6ZoB(+4lp#l#%J^kgXSUV%5g%Y!fJqewvV)#`~x1
zll3hZ;TAbFJN{_z8pkh@-;$13!vP7#@qrp~jvwQa&59flgUTkqR|&`9M4g+m0iodc
zo5mJWy&jxl(CW8KX7Mi@Q<_A1g~h7Shk|=Xz7q-9AnoD<!?SqPoFZrJZHM29WxgQ2
z<owasIwqSBg*=1X<?Zd`S50s8;@5WqCTVZzzXYC8YThyrJQG|Uosl#TZWkXx-3N8N
zEovWSlz$WSiM#)UdvI0<GQ989hPae-crn;oa8B`35uk+gT`}-&v%!!uPg5WvMfc*z
zOC8rR=%qfc(1fv7v??V(cFH5THbea_B?KTkHJzJ(#oDkI8pT<DFTrlTW|L`VY`b>w
zl6(SP#E&oTKxK+5A}Kpjr2f_~Ew_PXQnMbVX?Xee%0P1gSUMjY3NH_}4X`gCtmwLi
zR0+Qj4tGsIqwquh+a*Cmu@UU$&A=X@8UnfHgYy5Lm4npMzzaSN(hzXXDi7-^3B3zd
zr(Xn_tYs(OuHuhTQ^mQF55Y$`*I$kik3l9~C>n4=Vf`KN1T%$d>Om{T1K-m<g`2uw
zcVB+`X2Gif5Rf@F=FLJT5Qm96Nqqr`prN6a9!!ll3v|CuZ-)q0T3NOkEN9nr3mK-%
zkSY>y5C_0ev7(P!ulHNpd>Qyk?6Ldi1pP465fz7W-^2qCeIbpX)tjOOlZ&y*nj4bx
zLrV|CKZL*?ElV^88Rwr$cbvLX`nEN^qpih1+F?R7jYh=ec{ph@ATQ|sqGK48zUch1
zJN;3rC0T6%5**n^<~h;KQ}d(J_sO_Nf`4(DQs3n{JVyKMdo*3HpG?#=7JLULyW<5)
zT`g@rm#__OPCjvt8@4f9<f-jPoU1O_*6wZicI9v|BqtVMtHm~hV{K?Nt<_1S3Ij=?
zK8Uwz2p|@63)_8<OGwxb#eGcp?%NJ(mi3#t=k2-J5+Kxo8xb;QwiIHYo0O1OcDOm2
zEnYt7X~%|+*p^^(*Y<m*-8gi=NB6Gk=|3Nj8@1_!xRUt*Hik>Nphx_i_w_swj%IZ}
zhy?iz!$6yT1cj4F3QCto4IDEAv-tF3FAb#UWo2G6ioikiw1Df?=`;&%0V}9c=kaS9
zN+)8*<f9$EZC%668|t{$f1Vw62qKbzm@{{}BB#GB@%}1DA(4~OjfkN|s1`K;jQ#T+
z$>qf0&32wIL)KjjF?vF*$XFJDf~AEI{_YbVj~<hw|7@NA&)V&4U{bmIecZ+rXO_L#
zL+x(tUtXU7S(E>?d;2<>a$Mt2yxayLC&$u6X4UWVtSZgQZ3+PykRFr?4=m@8MIT|8
z=XZndT^>_}?PbT!SAP_f04vd3Y{-rz1|jfI>%;Ah;{T6B{=-q!u~<HzafCKT$sU^*
z-x7bsP#>vPL@9o;69Gl|n}0e-E5tS40O%K^FThCSG$G6#S7sj_{i05W!jFxFllr7w
zQt7vPuT1HVitO`ruRCM*jH`@wrh|MTp_o+db3$iL-)F@i^ht=}y`#^F=%oHSvl-a=
zhPc6w(*JY{=lZ3p2I0J<4VhG}-Q{`j4hTz;0U4*@SmqTi$gKSUZh0EmmLgZEr~a$&
zIj(4AH%lDj%-#kPu9R5-%QDxvFVbuQDFYmX)tC9?+&}C_m3Oa~pPBbxXP47i>1#F&
z{p3;nHChH5?e23#zpv5nT<rlvkv2-U*N_Y$r-U3J7H<rdcOOgm_K5c@%i3R+*NpuD
zw!Ib0<z{e|V;9>YGc*XBg=6BsWIj2MBY}2#GO&R7#M~|q9myR{<OWRvjs$p4c)|Q$
z+dW(Tec#Y(i63>EV=orszu&E`s9aTaJ>2n1q^(|D?HmzibrQ(W{}nEk&%Qeh$Z7l~
zHf8-aEG<|@@O1qRH2p0s?e+3MkiBOY9?G7-{O=sU_a6Zh^Byw%Uyz}5rSR<t`sMim
zXwvGE=8wq;<F-ro*Y}^*1;_N;<Q(0%LXI?$WxkmTZH<*{xZ{UX&i#OBE3q5v=xNt1
zg-&ZF=_EH6H9jz-wS7T(dpW~aLKj6QaEGx5{McJxi@eqYsBEm5Yrm48b@gMr!@o(&
zA8fPAj!_@<=86H4hqT5+xo{`RK@4m^3g4O`(C0XufJiX!e=;d!xtM)Yt^W7}0c0R(
z=k+@JrQQ_k6+DTiFxa}RHjfRl5b<2P^(OU|e0*ieJhnKA+g(IJl9ES@d$*_oTk8fY
zY%O5|yM$)zONNe{d}6Z{B{n<Ci2Y!RiAdDAAh1c>!9y!*J?GWs)rrK#4n&zaHPk$2
zDK-8X3qM(_I8~?)2qtq?n3N{kF<@>&T}^Z-?eVOWrBdTqk3L*|)YVIcYeY(4>1|S`
zqEA~$nJJ`r_&8f&r|R`^v57_qBUGwsEWrV#0}$!1h(5#S%)04|MlktXED=Xa6vQ3i
zrLa^uL-~D>&&T>Y`4SSbj#ySD-0Vh53?j}8*GyG~O=v_ky)gW*kxEu-K)%I>7pG9D
zJGt1wAe35j#V06%GtKqbCKOfQcamUk(9u6|QWwIO)^~KM(2{5Koh_{K5DZ#ij-XN0
zld4ov7qnV^+En%?xpvJc_(xQQQ#kuNHbPUPwSqL5g0J}W0pTHzH|b{`%$jT~%I?o)
zgB2X0y=w{-RX2AVR}8!ImWjXDOnvJRSTk>yCeN6O;FxH34Zac|NnRK>G04y?&qRWr
z$`fN`8Vw7s#@8*r)493trtPNtSMS1FhaF-uHnbTY8$MAM!@k*n$^{xWUG5&>F|m3(
z`b4~Be{`6{nUB<j>Edf>AhTXnH-jp!8e8xff-P`mFgxO*Z?Tb9%(O|ZD5<fH{`BJv
zxv=u(&@EE+-Reha`q>gDXn8kMfLmqq^T1R=<t9nKjvC8l#EMzVNdpBeYn5uOQ^vf~
zZCqLxFc#&SGQa$=#I!&fY5mmtcQMd5V<JlUv6k2@UggJBE;DlKLPV_@eO%gUPD;^k
z%;KSESL{7^GydZuCg?-Ac6pljCmdVSe9@Zm<Nd@X2royv&s+0d%oxXT5vbK@>(w~(
z<tWuqvPwbPl@x*-QaiDdM=s5nI!SIHZ!aP#6+b3!;yH=}!5e7pQz`lp0EJ`sZ$5bV
zKP`#!llQp~heRxbZ+v9RU4LVC<8hwyJFje13gqtUKr)kuPeY=T{^U~70W?NNnLpZy
zCZopKLYJYElPGJ#2Cm00^2mA8BW8jnnTfE(20fIWhsfL}xpvso@iS>FIwLK{(h^qT
KW5R>m9sX~}(<E5{

literal 0
HcmV?d00001

diff --git a/_static/bonsai-predictcentroids.jpg b/_static/bonsai-predictcentroids.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..e284f2338ae73e228178ef7866f24bf4e114d118
GIT binary patch
literal 76547
zcmeFZcU)87wl5l_3yAa%A|NVFk)l!pB29jXpi+g1NR1HbK?1SSMFa#C6r>9XkrG0W
zbOGrl5+EU=Nl&N&LUOb3d*|$Z?|b*&bN6}w-1pflnV-*^pS9*%W6U|`n4^5>qW`AP
z0#4sDF*gA)F);z4j30nL1uz0IpE&WiKjUCw{8>3zSy@<EPqMSKad4mH=H@!d#l>@q
z|1{4jzEfOWr-e`R2?z=a332m^hzbje@(T(H{_P=5%#1cHtemW@oPs=DJc9q*2mJ?t
zkK;rF3!a(jEZ_tm6EhzZy#pWv05GvKsQnwk|M$Ukf<Yr2I|t`UF2)VW(|{99%*-cP
znEysK<L(H?bpQ(=E5GbzLpA|xSN601g6c1lb2;R0RJID)43p%qJPLTpc~V$JR7_mq
zoZ|Tl7d14sw69*fZgkVw#MI3E*4=yeZSCwG9Niwfdw6<z`ve9BhlE1Ip1pb<6&({B
z7oU>)J}o`t!^h0L{DQ)w;;$v8Rn;}fT2x(qLtA@CXIJ;np5BqsvGEDa?@26yI5)qr
zxb$avg}k-Bv%9xXIXL_qUrYez|BJ1E=j^}Zi;uzA2^JP+7WTjK#dIQ+QJDEySY<D>
z@f%vRyZQ^9Re!-Dcq2KtvXxWriVaEVQNZv?VR;RL0{L&O{a4QZV~oB0r#Sm}#{SM1
z7QoHS#27qgJ^&DK<YrGH9H*Ay?dgC=2Wa<fGuR+<E@VRP>SUMyJvxB!1-QaY2c+g4
zS2QINi>gC&DX$#ss)u%YPu~i@@E*H7=$CRHw-k1zPkh!*QY%a{xLcOANo0oI943rE
zB5joCEn`l1`-A;UkQJ(YU559g!$cpt>9p-iQ-o1l%zw}nFW$rFbin))9RSX!i8Ir_
z01vLw0oc`_N^+epib&GicbXDE7t?uboThwV82XO>^)7e*EwdTt(igpF*7NL*@OA_E
z7&{%H2t0C@Aq~mwGSLA&gEWkb03k;JOQ{%GSsb~R{}#Kq*xJ#{q;9BmbN*q@`zp!M
zE_L=>m@n6Bp3wnR4fbOL_*NhtfE=a+co<K(BSZZ}2doFr3*KQz&;kFWcNiK2{?+`S
z?z#xgS)eGq&9}fZV$Z5FEU-o&l(p!2viRB$+WxY$Zb)^%xLKL>1XJ{brJS}V%&aLj
zUOQ)(N(UUOUf)fSp=_b(fC?}*3UK#`XRrXYKS>7=z5ad*)>`d8^iAsEf^b+#bvPyt
zq-i0z*Iun{a@%xEmP%a171{TQc1Z?4FpUO!Be<cBgpC;D5s!}Jk#uEwd#;mxzfWAq
zxpY@?v!9QDB)JNB`s+IkOjB!2amWTbAV-asR{RwhtaS*_&|WB62D;YoB+!n4R7;tE
zKl8?1BMwU0LxDD==Cn0!{yw6V)zP`}FZVB|9A|s)o2QF9f4};(r^n@T3h$YXBk;bD
zS48yEh}zK_9YEIj`>AF-^)ZE9KoxmAwT@M-Mp$V^?Uha(YCIK=9_Z)WI2Wi<sVZnL
ziWRLE7QH*5$KOTP<WdU%+O8bfRrS^H%Ow_1e5P>e>LKkLZBeTnynB2S{_iyY1GU2j
zZ{p-qMDD*kugoc7pZ#G^x+cxdh7JI+A0em{WkXEHf_MJe!Y)H*CbCn{pdZo!21lXT
zrojW>h4m@aaq6dyW2JX=Kr}-OeC4IxhVM04VL8^t`RD*d(s40J=?`8~1!)O19ruGN
z5qV?`#|xr$d|md}eA$*iC_AMq#UwLdIp1TzBzHB}XANDpCZ!s2_rICMLI0>C!dL~F
zR_TC$P!y*BF#9p@tPP~m|Krxs+o?pb^D_t~!&Gyj)Y9Kq6Z_SSkSV5%N_=&CKAoHC
zzS5SoiQV-b_T+#5CI2}R|JC~6R)GJVf&Evn{j1mhZUp?V)WHAQ(t?<?I*`!ZJUq^!
z157lbQg2_BBi07Lc>CMqy1yE^K0;+aD>DutDJN%rFt{p<l|4IGGN6NLp>)$s)&pti
zVme?r(S9d%8%2O`OAY4Clcq=iv|OeGUaPh5AF91BPO{`GFm*kbxv!w;f!<P?ar5f?
zasNsUMDES**9w^S(lYLhJ|AgjFPUJD(2(;j;!$-{!SLbBru(JJ&er(IZK1S&NVw|G
zW*!R1gH3Y4z}3Djk=^6S=3Ra75vtM?Y}%m%)Z#vhyiA5+p03{Y4(v*1Ed5>?Ax<c3
z(YP--T_ofadRnL2aRllIp_nVkkdp8@v8Ek)3%bU8oVt}o$6vMZwZ6a49;v!rg7iG0
z)?Y(>IbSFff&&$+(E)N4v=aaRr<@Qx9==T#`AiidBt@$R4aTRepsq`oKnQ(Z%?xfs
za769Ng370UV#M>hFP<7g_oYRYwl=3+c8vvng```XJ~c)OS{tzb2aEJ3NseYLF?dLW
zAB#vm>rA4CIne<aP1=p7od2=q165p)Ajt6hVAY~`lh}VdgTUlL|1UZYHy)>RS>9r+
zO^M{VSZT2r_F;#MPw_=^S^@RqSda`h+;T)YNbCzG-2Ei93P-z2221BkCGFW@UtVGk
zb8-hP<?054EQW7JV1hAQe_E_?yPV;@INKu}`q}V-%$fe#to%9eZt7X_N=gS#+vyM0
z{Ww+oorV?aTNb<~<lqLA>j%r?thyF>Xlg^1X3~omxiJkLc2Z{=<`vAf6lOh9;RpLG
z(=c-p_bbufX<C047#!D~{wsv}s(TMSYu)jJ@%c{ELxwT+@&GlTwhavLN9~u;9vnVX
z(}D@jL0(;gX0=^+Y9&Sc0z0k)U%;FSW@3hMD|bs-QCkJ^%fs2TgYp^bovOsGu#udW
z$INAvkF((keVdX+C9}<7C(4TmcaI!i<do$(Z_KT;z4G@hI&Wl#n}(cAF_P<UwvQle
zB~^VbTZF(@Z8sYAe3x_e&X$w#5uvq7E%+<M&Dd-%$}7TPY-WE~5@E*3X1Gc0Yfh|q
z=w3vl@7P9#ZtTXBFOr+@zmEU@^2cvh&kRx6!bUOKYz3Z#=@j{xm3-4++qc1T^sl$q
z>?@{NsrM*8+xSx`(s#;g)e@+rb`%`kys)x)97>(eIc_?%9~?ttT4@Y}LbG)DAMCLa
zQEOvCz&}}-=IDSL#pxkhhRo6J79ENOsb~Wd)4~nnTWvV*&<fc@fk;WXZeJFy?QY~8
zH$TgVwTB#b@3!(oz-u5ZGwmU8kIZ<&>d*6zI){-WG8QsGnG;+<-WeAh)Q`B>yVNCu
zxxV`?#7d#__<8nf&##9fnP;|)71Np7oI99PKSB7s8iK@c$7aUFY^Jt~NNeyfV@390
zZN@d?LAmc8S+|9+T<pc&{<W^S(Eg|;vsAq3mD{&z?`oXkl11@@<BD-q)j_}kM0$f`
zuutxd<l6{?JpEOWT6+c^kW2-o#p!!q2o9x&p%2vR)pDM%WQ$DLwPhd(GiNQ8TKkD+
z6~y6+9dA!&E8*P7z{y_nPd9B~)0W8uO(mzzeT2#rjPQ<>M2_t&YdW6<iD(Y~eoUe^
z1b-98Km5ocT-7Jn&6M%U0;M8*@U$Ix)B3hr28pUd2mC~IfaWoO(;7)1HLRFcGHr`X
zCn+GqEW@!8ppd-w<>f&zAl-Wwti`{6@8KQQ36|#qGHnLvSrE@kARPQPo({08(U;C1
z*uvZ-6=0Q<bbL?V%n(sltb1Tyd(#&5v5NW?Xku;dm9m9eqQR4I+l6jJHoa*viIIa1
zI$ohTIaKYiY~{$_RM@lE;yeSdRSTHfnLRIe?MeiyUyPY%JB8CrkH1)`>KSsw_T>4z
z0@we9MgPxw0{^pi?BkJ_C<%m|=x66L!im6InRO-n+L2k!`ngq`y1GvVDJ*8bc_~I)
z#&Vv1fV=&cvubhD{FL|DH;%G(BAJ0_ao&eTsewf}r%jfwK}eurK<RgOuu!}2RYkTY
z{9ZI_!mmP~yOVHqB&B8#5vK$4XI5<0!XaX`{;>cT!A#*u^{dHG&2;D3vnkD_qmFD|
z5+n`=Dx>sl=(;0Tq$P&k$6!}{4s~s7&C71)U9}s?H1-g=yXldJVq*qy^J}Mnp2--@
zV0fq%DEZWQXB+sz6;R62U?Hs?LI>!|(E%;29?%@yVVDG=*$y<p9rEmU23Tc%p*scC
z8?B+bjx=<OvU$p+D0z}HHz)G82{em{bx|YDv@{n{BIm54yYQ_xl$h@8<;bz+Juf;S
zOz3q;2D@G0)5i|lH7Sy;XC7a%I=TZ*y;V0tyJ}B0AV(JyW&-dx(t#X*w4ROkzt`T|
zEMiMBThDiQBKnC3!7s~iO=$i!>~ZSO)Q4jb!K=T6h=n$nGMplT6Li3fJ?Yf1-kGBh
z5}dp_P>*d>z!zjq@!~&=w1Q3;W9O7#KMyE~wqZtS@(5UsovCc7|6bdiE#@a%n|ae*
zK3=5-T$-9v(yYz*16SW=3`-?K#8<RX*>W`#P4~<G=(8krW$L{}mS9(|IMM+g-u+*k
zdIcZIvDHLv{F0qF`75k^o2LVFnOxR(2z(1C@FujjXkvmVgW>EHv>9o!xC|Vsw09J(
zMdX4`P71FHU13Rn9epoRUQI2%Le*!`5`A<VGA`e+x&Y1J+`0gaQ{2Z>a|lb2QkVp}
zk+7eH%yPbB+>#Q8aPxkmSU2KVULGuHU7tBZv`<(1BC_~Pl@3tTyN0-N#Dk`il4um6
zxomL?frKFn7tYQ|7{p^b_0Pn9B#wST#Z@$E^*ovF3lm6Ih-<yHmD`iyWy9sxP6v46
z<7nDs#U!ejXX$FDytHC+#b81Zh+?rTWqJo-bm4J^Iau~oRu$xY87JS##Q2f0S3x~r
zaWeu3Q>fM71M2(Na9Q3{eVD`1z_oH4U!|L)e!Eg87CJooBj>V-8_Vfea}y0cLPLJs
z2k@I;1>7S~Ia-RFOWx^`c;4Il<t6V8-oi*Y&T&;Fr<(RgwaJ>;90?bP$zhOEwt7a+
zjdj}<<%DXewEXsEKIy=Gjvw3iV^hQAv-@sNsKE(5lwu+;Ck5RhUXfKe{m0QcZ6>=h
zu*BG{z{0(-dAGkoOx{6(Ph*QS*3(9`gf%PVQ*A&(IDG3iFnHp);?N#vy|93CqDi6`
zX>)6qw&6VDgX`MDDdsXCC7pbq%-2N>ti*?ye48?KP45zP{}9CzpM4RtBg3-Hd~YIK
zM@gn;3LDm9WPUWcXm<9&#nf%{v#S0tK79gM!b&3M;8k8Nmq0NtTH`QN5)E%d%KW{W
z52_nKfwL?t2%Q+h_sFN*ik6XkCMK$u+4o%Z>k{oECDmj+f1F6RdlPK0uh=<vtdGOD
z<xJUKZp-umrxV99x9zqvg;9Jxd+7-Z(@B>k&)=IHJYTVaSL&t%QjU<jgI?QVQzXbv
z_1ezHU|q}KXSIwgBNmm-h*{uIw3w41Zx$+#%Vb9i$ggS2O=FtAx3#NBsP{t@N7vhK
z^$6#F8Hg4=6MOco^Nrz}#04$UTcHu*w{UCjGpAhEro&H1MB`~tQ#t^4QvdI7urfA_
z{fFv5ys$X$`I&!{QNh2r&5j4WTGOHfw%dS5v(|eO@B@O(G3ny7^Cx(}PtgH`9CX0u
z3)A1ebimSU`2LTQKVLZhzj;3M$2#M;TMPGC?@*bBo+^QFM@d*A6~eYM(BHp>BPcXR
z@T?+d$QGAH2b@_29nPbF-eEy|CLN(m^4AzPMd+@kf*_O3<Q#2t#p_5HK67lTFg~Vr
zNn>BY6ms%-S>a*nwQl(dt+csKZZPJ?lqH#=q^70cVe!UFv|AlRY?yy+YHkcL-GkBr
zjfa`~gEIT@yEfJaLY||(8z(=WPO0ybiUH;BeWe2`eSvj!!2T{B(1oxz5QP6jJ!#F!
zY6EzD3@4Y3-C>pa_gemu@aRzz9RLSWeRc(zO#fc>qIK5L81@w&LkEaA8w0HWu@3x)
zNB`?p|L+~>ow}@5*)CUl$=$|uQ*`feBGEXQ4v?;f*Wr7g>pIN|*Sis<<nEDGXh{%+
z(H9EDc04*KutvF`<Hg>_#VK<@#d@*WpJ#T_{YS=wgw3cHcFJcW{3TAXiX;-}QE6H~
z5{#||{WcALA(`UTQ`}a;bWv2k;0^hc9TFBo!sXAYu~HBQ!|SxkGM{YTiM4O*`=w#k
zX7BP;7iCQxp7)&L-U3Hac#mW$U{i_}X+TZtPeV~I5yCs^c>B<!lBhFd($!15`a?yk
z>_KyBn!?L9%Y4ey3!{{*^q%gGrg}6>v&P0XDpy8m|AA=+8G>=?E^vaD&5kd{t5P)~
z-=M;9K5STy*YI3J75FFYyQ3xPRddL5i^ory`t3<B)uarfku%~oqDh39Uek?bOVBJN
z#OP@sxYsVRBMM>tUn6^Hx@2H`4z~El+X>!ND>OZ_gv%e))`PCdDX${!dm(WjlVzU7
zA%Ap#=&0cQ70G-2iFSt4=n{VHWJ!icX@&!taL`ACc9`9AuAZpH*sgT4Q6V-?AzG^=
zK|aN#*ObFt^V8!QeGlZqZJeLV)Mkw@diL|x3NYxlN6Rg9&|(%`EpsO)7E}XjgPf!q
zkwjiK>ABAiiu>T?$tV7VrXPCbdW|kSu&ej*%T>1XG~ONv#{}k4+6f(zoONViyaEN1
zw|ocR_b6CG<q>wJ)xpUp|3Q!7#VYc)Y0`Ug`YK470-E*gw5pE$g%i%2B@g-1B*E6x
zt1hyZAj|MDHzQ<W6u;Y*IN*fVg6#m?=$J79IO$2ucpW4YEj}~&8i+baN=gd;?6cn{
zXI;Y^3J&^V7%b7$#w>5IEHq=)KNjeS9a(@_G{X><ijYq^Jv)x>A-l*L$1rR@l4cFk
z#Dns6gUPG2@aITK%tWYBJNAwA)dsMyn)PU8yGmZv+(62WSq>B|D<kxm{{yWo^5I-p
zqy&Zvo`pQO%~qpC6yTfRhfZLgR+N>>+k^%7yFaQOP`EE!8~ss?{9{lkXKwR#<Q1|Y
zo(*PMKozefm3m35G>#+TC!ygs-;WGl+Q&!a{{C>r`@@@tvKDpykR^Vq90}8fpDZS@
z<&qlP@TbbDcO!~@v}H(dq1j9Kk(L+AJwF%nhKYSty+5}ts%|R5*3E;<)84N~wPZgt
zrj-Y^G&*+n6*e&vUa9!4Bpj-FFaV?LKuT!SdZvrJP*2KuwK)F8DZvTIFX61e&<+SW
z2o&=D@twu(49OdJi5%LR6oay@#a*0e;v|%1eQg<roT?0V_u4kO&HUxv%f5FerjmC3
zIRa=;Iv}25UP?9v&xx*UqJ_M1>ZIn1G=JV#8#`cs!Q?i%?2&upa~Zxhi2C*$7ChYn
zm8~TjypzCAahVQCOaLkxyWf7u-mWwQ=r8dJL2WtW%{JUn+gdjs%SwJpcD+B<6>)EV
zEGz}%tS2~w-nxgtWQ}ZG9#`4{B`MH0JcEyC_EF!#Inw$ZlyD*{rinGAMMea>m~uE?
z`!%ooNY8gn*m*01MdXgdr2_Srf$_^K`f@N%^biEu2NIy_)DJ&-Z05InKA4NG78N?s
znPM<?ul5wvnNN~-Kc&TvKj^DMLG8ek#@#PBf+d=@2^;kUXA>;uZl$_|^Mj@q!$5_$
zGE*exoK^r}8QaK6y?W$GQ$h*wwdPRv*SQmrNQp*E1XNY0&kGx-@a96Mwl`~IpTLHH
zeV0XwM1>!A3$<4cR}E8|>b;<k+X@#mu1y68hlhPNp6cE51tXL4G&cpPdRDBzDOv3?
zgo*M(XS}hebUwVy32SfWT$}cmaQ35%OP@5`T-QQH_}*(uA1P}OPJ+j(?Q&4X0*Mh?
ztq9*HmEdr1@9^;Bv77Fm2lcHSGuOorFdd*ORLACNifK3YGvXCtSQTbR7P&<hLcJ!A
zsaX7WbgC4vR4R+r?LJy?^Uw`;y9+a+u#%ljhv437+{<|Gq@)dfd7q@O-ueo!Y`;f6
zjQ-0v+CNbJRuMq|PU*eWRu|-(`p?&#Onvpj9<EP?S%5as>7<3F2U8%@VhVaXJ?p4%
zX$O}>R+`<p&K{*owFsMqbF8$S8TAd8h>*MAR2>cbE57xPIr5Wx?Yx?N(7H>CE*Bw1
zU01IlnGSHVhLbf(OG#)?nb?rM&=|%f3DzKM<q~`aEbE$*?hOa?@ZQVDam+z?b+~pF
zSadk-P}|zvrwxfMYa}G{I43$NA>#g%(Yx^jq8k}jHnnXJpI?nV&0#r-<yeki&Qa4?
zA6{M#1;bJ>zZPF*C<~&zwO8HYoz1!}#&C+Et|lFDV=s*kc+Q|h1K!3-KpcwHAgCu2
z=zzHLUZpE{d1<Guc;G|Lh45KH5F?dHdI<&7qSlq@06-oiy9o}62Qa*f*8=UJ9W`$d
zg;j(O5HO{l$!7Snx#73yfFI}4l+5Y8=a(3%dlnMrm<wx1V=|@#I)>>04#jBnAD0m{
z>7*L%9P0cRZs7ltqVvrEBItiUX8-D;|CLn#tB3y2?;(G*sIFq}a?Yt;aF(8gRZYE>
z;*%lAyz7gP*;%;Xa>^t=3x0~1fX{ze1?@3hWz%j0DLH7>k$G~t8>zPZYgdd?u(!&W
z#8$Co`9D8in$3q4FqGJlHA891f#?AAgbX#!62q8iYggCN{!0g67{>mzn{@X?6`G(=
z2NcUt6X}58Lu_cOA~xxmlcDU6Y2O*3p;910e+>S|jNxZ(?#9po3bv+HMf?uK^-4w`
zi8C(#tL^`f`|3~4XQQLLR`E-HTTQ>H^#hm8Rw;K_tteJVnL~btFdZGcGN54h;B65M
z2Y)D;Vb$VSd1<%Gea<s0(gfB`X|EIEG&;Z@zI9)Ql;Z>3kq0wE0EdkAHKoik9?<ju
z5K)G3{kP&bk@s93{@;lpL#R3T@)_&A2?Iw0{{}p`i2VA6JLf+_mHNaAn1Wp)_WE8o
z|A3nG6BEBF_G8-mV}QXR_cM$|5hKBH9Yd>E+Y(Cfr+tr50$b7n?}8D>0igNwC;DL#
zM^7Vhmm>$N4QJsjzuSpc@s98HpM;D^O<*6G7GMCUE%<tiOl7W|z9g}{87qBtlqOBN
zYcsJqe?!H$FjMIf!G$o+#WAYq5DSO^lW+sw9oXXdCet}JDaU^HH&pX*mEGC;8|C8F
zDPN!O{1q|nUvqv<DTmMQcl(F^s=vw?e|N_C4bPwzrhX-LbIqwGKFjyP3sU{dC=I()
z&}wZqrSrV|kD{J_1x|B~k?jRwBBZSITTk4p-%EZQn(k?~`a0banRdu@q(f_W6zSZ2
zRokLGPBs4Oup(akdGGw%O^@>@%*cy(dM#3Z-ey*1V!nn*)+|@q(R)M(+)f*DP~MP7
zdCOh1J>O~0J{nK<vEmafzTg+xoH4s~eWdJlu8DfkOd)h;Z|0yV2+qcU8`1%9WU89i
ztIf-CR0~R|KvJz&=C2~BMz6}36Wg^8A#X#D2hit_+IOc2GM!00u&0#pe9CB>T^k*6
zbNuY~*++mBElFJ))!P^oR{Ljf=h)D~X51WdXPY-^oZ;l_@MvHSFt&xiPiv5eQZ>9=
zesOvy{!oF<*vep7{`HZOR4C!HYW{roQis1O??@l95^C(F?a`>87$#E|WNrM+bII^O
zlbx}4hIfC%Z)k6o_$VPLyIIFww0lhRK!GNAI^g6+_kFV6Z$|zO_?@a*<^y`X4MPBX
zPMcRpPE38S8#r}kXu>3!4DY%Qk86@GqKvj?>L<GsOR_Om%|kFl)gl{$?|dDnMu3CU
z`@51huwD@BQv-HQwQeh(75&R3NPLjgS^WC5iV^~kPwxo+S@B1WxRYl$9tj2fVlxML
zrrJ@tND}S(r!)Q<x7B;J|DYPV^`#C2GlZ9AoqNnB`@*j{&GnIbVLl*j?kX{*VKJly
zU(T3Og?{!^q5X6~?bLQ;!%Fa>=Ub{~IXI(1^)pfGB4Z*~&g_-zPwZV#J%crXsB1pz
zduecunFaH~2aoi8<w5y8yU7B*+u<TW)-SH8LFg36Jv8(_$?;am%|YC}fd7@G{T+hj
zCcrL}DnyZ;!*f$&eq6VUQy@po6u)v@z}%^C85%WB%j9(B@%!!+(Ez%1%R-^u^i{f=
z+{KO5zmBtE?*61vLU;7Yur4snjkM`R+T9%)<a3fZ`noouyY|RQMY5pEPv*k49<lK7
zsY1%K;?|_<VVY={jlUl*0zEL5WAq0ttjkBV1Uj^g=(-b*ZTgI-gDtXDY&i43V>uyu
zX+IN~>_7khad-Ki2W6Rr#p>8)5qYB?SrprsPwtjIJQpaR%z6W~c+Sk-fSqhISNcvD
z*a_l+d6FdzG2-|i{LqU|GBKHmD3zT_#FMF`+lz#=rwh)BIosl{aQS+_@%`9BOdKyK
zbIoSgS7>)cSCWpK`b%J{G40|G_!H@kY!qa}LUiZx0$Py+y0mZ=VhjhhXKh#QEUr#?
zS3tWCwJBg$N}%>q@7@G%+>?B_^;vXI1qe$)8`f72ea;`Ra2=b4*6cmlaEj)Mjcxr2
za}&xXfW*9{uMfLT=xSDYx5$pjd)z~hRLs@C_;_2CQ%=F0{nellLH{yEyA%hOKI=tQ
z3k=WeCs^_Ro|^h|_(oX!#76JoqT60Tw^#b<;p?1Q9UUynOix1K)#w<iQywL-Bazsr
zli;w4^j3W`P;x3e^V`seJS+8s8aiOd^F!4o?vFacdfnIBQx@J}L%nsvCa1`_nTD3+
zVf{Voqnz29oXq9QCcF$yq`Y~zm6A;0V|`<78F~TZ8#2|cs$2Z-;KTihzM4;LKf3~*
zC0jH}JEvzX&NUaA3YEWH$>7(VcXh3P*y;CR<KU6L6VYhvbK)RdlL&$L$Bbtinv{pG
zfn5W8Gn}*T*TE{lyX^*e_vV~6C^J~9liK*E>T$$cPF6Ng<SFW3Bz#-n=TrFBK=Ac+
zB&l?&jr;pbh<{N|#0i|cr{{Zj(Fb|?#)~|?Mo&PQaS9S;c=$161xf}{__tcfuOIJ*
zI7WH{>z%VFzebL^2$AT3BtPBD6&8@fKE>E_s!q`~a>vPUKI8V=om)m}*jN4bEza|e
z-N)2ZTQXu{c%H#{{oya4yvyNBMU{=Wi0VfzQw#8txsOdPA}S~()ELv2BevXf%{fHv
zl;u|^qqzue+b!RU_16fm78zmw-aPYru++=FDWav9@8HjC1>q-;?!Vf*e_f-tp^9=u
zJAI*sset#i)t*d^W`38z8WCRQH*K&Ta<s8|%l1wh%J$$01%J!HZ7N%8iuHcpX|UV!
zZ9MMnuGQ|eVs~1dc*crJfC1TKD?S+srv#ftj>;Elq&_mS*O5w*d#D=H@abXvlN=FU
z9^P0@6;`S{VHXqhQ+nmd;(CFM4>C&dN{_PB{@JKA=??)c;$$AD*^D?pWtyxd+bg~?
zq?I5AWL@>a5|Cm6Rv3jX;Tx3}KBzt|c=>e7aQX@`h9*gAH>WJP>t_*W?pl&+p0yN~
zfTTQ@{z8wun^x`^EMjC9tfIpnCtH+7#Tos|m-F^i$giFl++47XR6%I9tj$q1wOcYH
z;h3Fcu6B6Memf8RhqM?usyi`Ah1s0U?cyFgfpl{6mWtsOQBPBsJ#e`HmY1yoaT<1y
z1bV&c`)zkn<j-}@8BE%|(~9_~)8=ZK_QsXYFMJoZxssZ!D?f=Jykb~2y!sdZNVAiQ
z;$&VUXT&W)T&-W6T^Qj7zGbz-Op!DPzoM2f!z%Y6=!Am$iLLbe@)vT9kF#~1dE3@K
z|7s$QseJ1rfvresRZV-p@L22au*E6%nJA`gX0$9P)(F<5r?;$<I@%wOpGT8a8dtv5
zA3&za{-Bq>bih>O-r=jelqiC;NT)bF8p+H5XHQyg%q}85DHtvonqe_`x$T!!cklF9
zJ>EhySTLcI2&b3~#s=jQmHlWxkDh?QT{Dwgp^z?oR`G3cXAPQ_5!wQtZnBhGYEOWR
zd(+ONM}nEkon5dh&CgM&m**^^-hH%aBWVnWS(|a0sSkhpK1ZO)%>97{`xm%OLm6wh
zVOCdhC!fnC$m^*$UzcC!h=^xxgi-%r9KXQtpXT0-O^mX(_-NdGNX0CK&adZJH!Lmg
zUt6!pgk@1joOcgo{-U&ZyNFZ53ssAWh||k1Vxx}MCvjkrO2-?I&%Apv|4{ohuh<~l
zM+D4}go^TA^PFh1DDXZF+V)mBrruHM2GG{3d8ntMgVqn&J2!b(C-D4DymP*aZK+Tk
z#-wKKo8pLys!Co%T*XaR)kjBvC3ihN1D#T5u63)?KHD_fopTZv>Phsy;zFXOmjA)}
zVqUNBh&5fFl}P07F20pwA$Rk3v7emQ{4J3nS*Ocyy4wAwH`d{+o(GdYwdmlAT4y?-
ztdMdJ|2gRk)tqe9+Hz{_%y$gcDc36$!zRm9QCMOGQI5XRclv$^%jV^uB2u@suDp8C
ziCDit!dab$qRpDfAxZhuP{%_(eVQdB7WNXTS>8bhTu}F9x)~qJDx3S>ZT?(S!oB0q
zR72h<s#}rMA3XMo1u1>B&GGSm#`g!73iFTlgQWfKtNW&Cq;By_?9)pIwP?;JSG6uB
zr^Xx}3;lO3mq%eYvH8<e8fQg!*sI|0`xX!4o|aKX+siIo^=%BOh_u~dY}8<w2L(km
z@OWkbzL$ZM5^aP{EgS%iH_@ajU3u8U76Iw=!*IeyucqB0hSei`<ewygiq(cpy$yWx
z=<5TR4QVrmsvTAV&F_em7}6+<xDn0_KUXa?c|XNA(R$&XNr`q}bi(JiVgbK<$YJ^-
zl-7be(o<6P`;2Zdad<8VITi8AK$Eo*d-vCoDfXt(+LtgFmm&R#CL1p}57m9viqDIt
z4B?!HehgX<?^KsM)wMC~_xf4XewJ@IXQmf%TJFZT^0VI^rkKx^4CrrJE;AtLA_uiC
zb`ejqGgdtz8O?~<oPim{8lIDO&Ix8m8s9*9a$7qZm4TMO_ptt~dmB_+pTeYG2XO5O
z{~^}3_IUT3^Oe1qq~f@i^UKmaq=CwO%IAu2yIcu<&5HNqXMdaf?5!@CC2zEAV|i@~
z!aQGPUjs5xmB@OSwbeO1C*Hwyb_U*I7bzQV?^b7?GUPHnt;^G*db#$#B7koYfRQEg
zM$@iR3QLx2{h_Bo(R&m6tP|8r4e$3SY?SS^FL=F+wZ%Ee{FJ_%_DHDi+Rdv!cfy|_
zRt}u-RC8ba{IdL#$4)>-&jtAw#_;j#AbiLo-8m<q;Xks^lYO9B7UMez9&GbO?r{d3
zWh2W_Nd|vXKA`cwec~^JavflT3s;LJM{)d)afb6trO)c@V-NXVOckfSGT@s`x+m&c
zb`Y80A~_iV80{FCop3RjNGL%(-<9FESBul+VfA&-3^J&6>MnLF(=g1lk*|YJ8g^%F
zUv_L!rd%e=bckfD#7;u7Qi@%B=}PCa3Rk3MUP{QQjwpr6Hm4Jn;_|V>HSv2MuG=Ys
zR~0JyGBN81J&|z-zU8A~II<N(-<@YzsI~fK<NC4`J7cm*4x!w%1FIJkyDNlu&z}Kx
zO3n0c|9V=@?!Se)$ZNFxptwCJDN>Yz{Q7}<xpY~(c8GyX(&oSxm~8jge3VlW<lE%G
z+1kFTT6mwJWA$;1Tr!J4O@R<I-ta?Wo<f&yk!?!~SNOoF(DVe>6sJ}eV?SFf+-BUY
zcivXsJCXP9`PW?AWM1g&V1j+aEogmNhA+KzAz-$8h4IFlw@ZRqCIZfO%iUs|A)wAy
zeVBTC@6+h&g5tKh2HHvw<7QT<bZ$MQWt9<Ljt7nEza5TLhRSthMpj1@K>OMc3)2(o
zRx-343j|DEYxr^DO3A_}Ds?KK7mrj>K;7A5WHbY9K#(E}r!qWn&tJZqz^cu>`Xi0}
zZvju-Dx2T(z?O-K=$0###4oFfkXcS$A(f=|e!I5x!zY+p-{a!Ir^y^P0^Aj45V5G*
znl7zhcz=_XT^S~-{ut~sIox{uu0?#iVQQPWewu{AaA`yy4%|jwyPI68ki;(W^f&uU
z)=4m8H!RXy1=xffeK2~A&2Yg}*m4f8V2%)$>lvi|y)p)3qXsPueaQ%v>kACQ<nqD%
zUVDc%G=&#$Hr_6`+_`VVcdmb!_wujUP*}6l2$DB8TW|zk*)l_n5Fb1OL-stenjHB@
zA&bT!IhCVdHI+=LYwBlcvN>$SkrI@?AK7O|bnS8cb!7Er?0)>?>r#m`c{3bHu1ziM
zmVm}^hrLOZgQM}o>RL9#yZNNXSxhYT%A8uv_4L9*r&e&W6D6Anb<Nbays*plM^$D0
z?x)({)DGq6Uw^k>RZ5f4zsJ}*6zPjF4u|9hpmbBP8@P-Rct%$9pdVTLElhhBT}219
z84-EZ43F*#p=hJc@%RBbG0x^_qWbk`@g;L=@zOk$&^<NEEr$Asx~+e+#T<FsM2`+K
zZd$xF9VVq`3`KXV#YkVPqm*mM!*sOc5kcjq?LB$sZUl!Q<MeC%-Ua7eOK<kg|5Uj1
zmnrn})?2Oh9x{77PzDOJBu&J8S$3TJ-AYi><F~xq4#}PNF|lvz1c$Gd-%{XrEUgl{
zWLQ`EG5$^Ir-K|tbw@me%0*3Y*sR5bJeWo?l;KaB9XxbqM<h;Pj$+kX<!!x6J3kzB
zP19SqTrXJVhD0Eg3i}@8x+wDz`rw$b`d}c5JldWC=hrozEx&6$vZLo_HI_G8a|G_c
zrCB#_G1(f>&#i)t1J$rzY;4&GxblM7!bx$M1IAnjM)5GV`(1@WF~CzbgPbnj0U!tE
zXA%AA4<7Ney;F`@&^4o)vmVFYhqV&mFJ#kJ_-VW)S$$=&eQ!0qjMCWdn4Kek<%Zb$
z6FQ(8Heo9-#CM>5)fRjE)bKX<PC)0qh30WABP=z7o(Y{qSR{)?$hhxM0{f9$=#+gG
zc;{ZE#CUn`nA%Q+e(SdiM8&x=O$Aiu)!D5+vr2UW@y>jYY%m}7lN24Gs!df(iq3XI
zQ<Z@ypI8hRRl|L!T>S?8{ch6~t^yTtbA@61>nf!74)kd?sU4{+>wjQ;L|++w7$Zbg
z=-lm7OQx>rAH-W_Z6u;Ei-~oT@eH4YvE3^w`8b2nc=x!F)WFN_9qD8!y^1^v@x>@>
zo!?eDxO)2&wOC9odh>7slxIQ1S@o>{%u#KmiFd-+`wxu|5l=9*i%_+;?6Z|anrMby
zs^1uaTFZ3FEoo&P@0vrx>OQ~Hw_*m&kLo&+-G0!R##GjHt>)BtDFKIz=d!t+o$kGy
zaBzFD5$qr-Z2c<bs;K6+RX08wC<t@PgEGbjs}{2hk#iZDi^f@<N@P>AG%<9Se4=R%
z(RS9zyZPAmIY5rjwUy-?P-ec76#5$O6)V#|fm23dc3y`e&m$$yXas1CMeBg;3f}!y
z{7lT1&-C0LBRtn{)R1_%dK~s_eL44&i0+r#MN}VTd#Z2au=oynKUyQvvHcdB0X1?j
z^<i=x=p`PWU##a2DM{q=d-^wS#t!!yj}yc4+T2I;#iKSS{o&zzp^zE#b7!0rmn@E&
z4#Gh1b^H!4McPtckhF<x`4!NSfk~<}{0n6S+o@ZA;>f7B2E~UphRBBoxJ>l~p0;_j
zvX^JoVs`(jML4<>$i708oro0aFLpw%IjY5W?otsKFT7<c#u+V2Cf6OLnq0~UYWY8Z
zy0lSN9!x_M4tK7+uFr>Wptg}1Vo;>W-@+e~J|e>Ul4Oq<1a8V$s$&o9Qpocw*<@BQ
z<ldwE_NNVXCEwV60^^aWmqAJ551=7>H#!SPb#G(}BZGrW47nYlzWU(yr?(dmM0&g!
z&Y0S%DsJkhS(pIj8d2s%56NKqkH7b;b{g-PMzd#6y3F*<Uwf)wLoRP?Yuvr;KCRBF
zDDP2$B;|A*W^{+C6_VPWp)kr~dq-bMk6xMKylHOp$K%P={%wB3ow>e0y}h?3d;<!t
zO%A~nOSLX-IO48nZ@%{YI>R37rit!fpeaCaVA_Qv<+C~+<@KeSuqwzfO;OvCh{>s{
zM^h44ueL6+)t2vC=rn8TnQ+ye8=a56o#FqpT)@O+meAVzAn?<i&{ldb$&Pp~#y=OF
zX+~yE+Ovw)wU!-)De1q1BqEQ=nK-D)7ldzO!uEy7PZh3Sn@(JbWTt>Z3nB-uC%q1K
zzPNp_e#|#m&Od(z`|YfE<hcBZ25%sk#lOR;UvA&LR`0h>tC1+~ltm=IIA;eXl=3w<
zLUo*)iezBmjy`WjM_zzl>B8#K$_aQ8uj_jm#pxm$SP#ai2zRIw<%1rTM9t5?y;k-N
z__j%#sKUSGFw$gH==03Db`yA~x;fqnT3fL=bGGt++q>IVfW;O@z&lEttk#KRp~Mu4
z?tYecR(%)rk$NY{t$aGmBI&xBG_mJ#)a>ul<cKe3t=X2dxvkltSNghSyC`Zz=>se@
zbJbxi$3&@!LS*a>;$M09+gTz`rPur2H$G0mCPSR5O5-WMdmi(7S>}7knxk8eWnw!R
z`08Q(&Z$@Ae8PVGx5b8GSSaDM%Y*=bk6v^o<hpD+=;E1Q6$WaGKkJYBx87tcc`!hB
z+0rBHyG{u+p?_7QMx5Qo-0PYWh5J9=PxDZy6ubDx;K}3b*~AtbatH%79G4>)DT~Z`
zp{lD_!Nb+Y0DJ2lFXT+A*^JsVbcwEd)5W##4d<542-@l=LYAR91iOwT5oc&>r`DlK
zCBdSWTpkm<9@Ho?qwZ`%oKQKZV{-3G*hF4%s<N=n-OG$XNXV_ID#+n}%CXad$m36`
z3<T>%TdHs1pQFJz+^JEjjxS;cT5VK2<mZsiW8-^E5c$5e@sfju?<3)BoUOBR(p|Y@
zG-Zm6G3j$oe6WldG_t*2yU-Tn*f4(PZEwINmfz#)o7<yrq@PA8wZYU#cRNOV2$oj&
zMm5Xoort_tdF-3W^UWiR4bcjt=W5Mol=I%sNc=jTqbhfSB}OHs)v17bnVj?-X848j
z5hkD(|8QhCN9aYagNs99$DI`sTrhG>Kd3}lx8`E+V_UALkv5?xpTS+?rWR)qBN%W{
zDl!!`QAMq{qp>t0cFr}E+1qeJsLC8ivTAaHA-umJW1nn*l4L%r`lU@WoYOP*Udil?
zj`L96l~@IH-Ie#}+$gjkg}p5i-+094FnRCP!gI;YgM@#=|HN%~$#7}*!6$PJNYP1P
zt&(qE{(q`eK6g?BEU#8I)P5w2a3{;v+l$?wbhCXT@Lk}Qh4VhDcU^m5X&FDVx_-W*
zInrX+7uK?AfoYM0TD47F|Ga@DCW@F<k1SY(A0-^$5S6P*9(Ulk^*EvB9&|>Fle-{T
z$i^z}86T|={R%`{Ymsaktq2CLn@q?#4;d|_Y8aeIxxt*WwRtB;5r<DBMOY9v-)OET
ziBVGMfL3X_Q8ft6&a-@(@6Yv#%*@V|qEsQor<V@m!_$}2wd3tBASd6qh=gb%7)r#X
znY45p*)*hrMxqHk<j4;s!#yQi5s^V=hk`yXWocOao5#?Yv_;4@z)J>ft(FlLQDfLS
zR@+aod~b?ba7l0#IARrEv;$=9X&`H6mj&ljijBw1gEG8!#nys3JS@)Lu~7c??dD^E
zInNs=Yl=28>6N}*7@ibIX}m!OrX<Lte`DWkhYMDp_t={J&S?V3UNLoZVJ15f&J{AU
z9h@p_Aq_6cU*K(;gsl4(t92?*HqTX#6=hn8$P-wSYo12C)*R#pm>36vICjg>f{~;6
z)4LLglU9TaL{e7*MDO!L=QjqPdpNS!<47aLM6^Wogo)0HRMeT13lAo&(1eJ{2oxyr
z+Qz{HbUS5h23A4a$dazb3+al;yoq`3)OmQ4@_kD~pyY>IQeWe-$vOk;oBjFPH&7f<
zV3Vpygm6>r+7DHVP(=7?g|WSND~=g^5qZW><kYKwRo*&6hl}1f`}8R#<hdnO`)#Bc
z)sn0jM;Of`(aOOEQ&8fd)9GcUFH?U69J@=puhsibj*0*2;QAS9Fv8wt$kbdmXBV|d
zcJmH#%xOPXBbvVY=&s%Hpo9tia^81K@%gk_vM+PL?ZJ5K<BL|mloUd5tB0ddJwP3w
z^{wSmQ?CK*8MSRNjrU;pF;HAnCzOQDuAW_O7O)*}GNVb>qw4B@S*v5M0#tn$ezX2k
zmSwtqn%Z&6_68D6wCX?zKq1}w!ldFTm}bH1VK&M$0yY8`A8UAW$@Iy@ih<X<AJ-L$
z<gOvM(Lv4mOH>VV_iGI_H|RIi(|Ka2T(Mix;XA2pWdd{k?Z+Z1^>8<_?8QGOc=WrU
zasro(GXU?jDv~;rxLv4s$T^Hap8zyr)})v?BVt)SzV6p_t&9$Eb+&?NAUqy^C_ojR
z;D~;B@~3?$vt|TZhROsyRw9qgmj6N^bMXCYx8owkeV+B>lY;30HtJOj)pG8$fcc1D
z(9eQQ9VuS>j5`)lC$-L}I5`E#hw7Yot(!KV`l8qJ=mw7AVkAPu+X=Dkv4runLlx9(
z{-uI%Ck~J-RZ!f5s~gg}J(|xw)thVk<btMJw05=z2_FL!BY>h&covv4cFU)sX(HL{
zT|>mjn?t&H7$y)`^|V5r!VvX)t;xyEb~0k^Xs*tw3ny7SrkNx`3GEEclt!tRhlP2%
zhWUm=zB;L3U-_g@895pWly&94GN5j{|IE-RM`DlFDry@Rn-2?p*A#pu4yZFtwZqH9
zjqAZ{E<%)Cq8nL!s+B=aZC0+2-QknN!l;3mZ#^Yb7LxqNl09<fV87wID;=i2y@tR`
zazZ03AfYCKd}y>D^BwZr7j6|jvQZk1pKUlc91Rj=4FxVenv#+b9hkT~6kjrnenk@{
zohs8WD{B2ZvodP5=k|3qvvaXL=-tNjBlC*Iw7ECmxvkA?4CMJ)!`PGbIaVXjLK2j-
zrBF=}J}9c4f%B95VL6^wychb(De!5@{4_P`zlNRFO>zDSqm6U`DusG>bL(N+6jgt7
zTd#DGh<X_;!#+)^8toEi#YB}SaPIyLzAD{ywt$W2s&4`#Bc|)|O|8S~B=A*If?dbq
z73dw~rO!d2jwi71PeFq-)qc<c|4cHkAnYRfvHn%EOq{O<^EZ1(3bDH*9jhFANzNyo
z{D<s#Eh)TIxw8MLlfUZ3&?%Gqr8PNRR9E>xXg7_|^At|UN|+%J+?Wu&mzRi@aFoSC
zj^~k8rv7-6cHeaV(oDa?rc)6<NrjPvDQ3Xo4jTFQVLNEX7d2*)wEgs!tl69-oCvv?
zdJZ$Rg1zY?N}5>{P9O@7pH1mqc1OO1U%GWEp4n>Qc(Lcf(kC9Na3U_tP4NZD6bw4O
z%)p=CP8K>&BsN_pVi(=cRz>qaoD#%xEL3M-f-$>L%*i61BGQUv%c7`Kiq?jeqysb^
ze`^$>QLiac=C$*6UBWw~zeXTV{m5P6idg&cq*7>`3<t~^q8+uHP{w=OGKi)J=Ucf<
ziKhepO4&LOtSUzGH#nFIHCzF{*R6Yl4i9^h&ShPV+(9w;+j61>Oo1ahzNOYSSZvmm
zQwmqZpB?KM-yhjj7}J<?Dp~W>=vj<=9IjN9Sok8g$MkYzuItdc$Ah<=lH;XL9Ov`m
z4Fs)?0smHn+F`y7go0Z+d`jk6Mk-YEzpVrOLC7WQJD`Mr4y9<<>iC=w$qE3|egkc6
zb?lC!3Gzn0^UgCzIvaUpJqf-lJ~`~W=U{muL%k*-az_=fc4p%$qP#_cq6(CzJaZOa
z8C=Kuu78`zco$ZrM>c!m5X#2#m|vaaY*jAFB_9Hj<eOZN*Wf5yENOf3Tz|jN;`WID
z3}fM>1Jnx+@<sR1bC4}~?$%<)krA|*#s`c+iC0Y)mCbQoK)Pp`kh=0qy4@lCh944}
zo062@4qnptu)g=y@T$tFH(Z$NJf{TbSQdB5f)JAURxht;3rtfI%p3TOc|P^<1=WPU
zwDH(>3$_hD|1GfYZG_h5p8trN?&Ul$X5+e=??XeTMLn*igWzDV+z_J9heC(<LbHsG
zd6!l*Ee0piCQYP`w7kSq;X0NVuof3;j%~meU&Xc;)u%fz_Ge6Y4p8eAL%=4Z5KA>m
z&{559ZDXp+4h@a?anRTb=i6wJ`0ebn^2{Wo6z70SWEirs58bmOYQMa3d(q=286bS{
zf@fbKCnFn03@IeLbQ2sJn=mkKw{We(JPi@7=I0Sn>eL?_rw}9m48!^>XWoUyJ~k46
ztQjsGbEH9mw|~dByAhynFR?m*+@hV*)Ace^lrwJiZWwWj3BLO=Co0MkU+8&sf#TC`
zI2IJU!-c9rJ3$x1cFR8c@tuhv$=_{SG_%n@NOp5S8DPLN^tm2ZPSK*-+0qctAh?E2
z#r<ae1vq*bAxBDju6u>d)!K4_41a6OB_nVSQEI9fSCA~%Ff3Qwkbd3~>#ZG;W^+Ct
z#(hD)+HX1#+?7*fg2<!fcV734Da5Zbx5GV?)<RlD#*ku~-qIXPva0K;Jt&38_a)i`
zw@(Yk@0MHbxeWy;x;UGx-lz+!kiMomYv~AGM|>8#@3J!E!S~=hM}AQ0(t_CXT}u8Q
zWSHvyO~1JGPikMFdftO>NU0@+btSh~<=5p~mD5Z^ocFs<UH3Ql!nhxUXzWA_oE2j)
zz6h%CuyUTUgQ0UyMY~g4AseT{2UGd7dU!io;oDwn)=}&x|7akm?oe`#Z*aCXiV4MG
zviO7IG>4Cs>zI(1X_%<_d`sduu%mj~dB%HRP1$zB@wIjABewauvvypmv6zS0sZ$KF
z8l;8gw+{$QoHN>HD~|w|xG(}K-?#l#s5!C}cm`yWKzMp;z_90+xZ>JD5j3=;3cM$-
zJhTE)nQoo3PB$B_aMZ55AaKiCyT+{J;){1ywoj5ra7m2K6AZxpxnq6e)a?tw4b;qI
zs|zGZ=b(^J%#Tc5x%3;F1i2$oI>*%=auMe}S)0Bnr%`0<u=VcUlv|S9rK2C>zbIg`
zM5q3>AzhUMsO?m{z6c{t?pBWbhJ8}yxL)UNqXYJG%LHC9eRmKrAaYSF8$8KC(|4Pa
zx|)I$b#A09A=N*~rP`!BfsJJKT~Vm8zibz(0vFCMDe76?Z#d4N0t1Ww)G`~^A!~=%
zS)&lYk>2Y$Yxy#NwweN!Ect%?3P`^@ea}2lQJM}|P(#}`!>_d@wd7=@qjM4mKf&b4
zcRTUCk=#^`BY7=x&zQUI!gc~ahi?<ZzdH74o6Dwv?fT<OSt0uNGrS*%o?P1-Oc3d|
z8PjDY{BTRPBD2+$g#^|=e#sT)R~sJV<;GrV^J8ApwjZ;WLrCaXvK-}&smL9DqD&l3
z8NfKyzZ5IaSL}->@IOUy)IU*f?GlS2ORiIn2&zb;g|GOM6*5}I$#~3l?1`iIOVczS
z$=J>S1L)(ltLzdlTDz3Odezg`_YhkQu{woU%}Yn2Jol%lhm63J%oXTh+dHTqP>|B@
zNIFI&)a0ylmEVNjBn=dcVrrl^kN)zJcj2FMeN$lb-Bp75l*I+?Sjz=;J3<s@klA}d
zRaJ9P#QQKvgwGpwJ@&PfIL=LNay=$Xtw#V*^46Z;fUS9An=t_p5WfxiVMVh|Ri5$j
z*VhY%?WcRqJi>f!thu9N`}~aRwk9<_?T>SI`a|^GAsSkq;}t}7L#EPFJgCN4+*oqh
zEQKv<T+TpTy}(V+Tv0T<(xMmt25#*4x3FK}E$7Ys9XLtTnco5~mox{8Rc#Uo4Ivxm
z`tX5)lj(=iD8kHbCOBWr06%v5P2QsCQO|sg%*y;zip1?x8RX3_+S$=3JIAV0Qo&2)
z=cGRTN`vO{Yc<YM?DpW1;6Qc9T-5IYBj2A7C586z6psiI8eiwINfNYnHT*bFl`%d}
zjIXTuUi3VAn6)nLdgE^B$Ir!nA>;?eadg1rt7@`C$~9olb!9oSC??q31EZR_(9R|1
z`#LUgu_n*O_25;>0q4N8G2ebf25ky{+{j=mU^%$C@Zc8y$nUs?rqqE0Vtj*~Z|&P3
zv1BWdfblh}VroG@GjX^*+}ad7mt>00IoGB$xuFtsHk?<xf8zugAc;r}jg>LKGfAuX
z$&T*AKZnw3MTHeH(&fU45@hGxU~stdU-EByl-~?0iP@YTloP1E)HLtA59h^NxaCi^
zUVoAn>K?#rtB9J2B28{GK3~@v>Ag&Qyi03Duaa7#69ui!T>m>T7i)OZ?yF@dM+o&*
zIeeAjZOYoCWLDWnYHRQZ?2e2te<;vyJO?shjI)(8+%$P87T=1B6#G7gL~u{V32${<
zCN4m(rfvQ1s#i(M`7-x(XKtPjh@r_~nq)|D(_!j;!HI_d!rpraG}UG6!%<XJKosc+
zs30g+r70~c(nPw{08vp|43XZ6bfgLh2nbOD0Vxq7AT`njM5Ie7fzW$G4Uoj|yfb%-
z&fF>9xpRN_e*Td_4m*c)_St*wwby!{XI0{dEzC_voRqTD1aAxEP04BS2J)#3lk>O0
zjkWwGB-=1I?c;hXXF{`K;_3Bybr(K{K<JC>6bHt|TY@mi03=Jb7BREwW=H;Pva?Kj
z?;9gG2DQ6$7j1KId$>&qx*5iZ-KR#rQ^tlUC2tnW85=unuDN4s7bxq>FUWvuN-TFy
zJ(9)Blg_aBfv+UUMz&ZBv%;Sh*n3oG?DCl#lTmvi5gXq?gTk94rG2hkiY6p}Z7GRN
zN7CGm;#gjcoPpvtz1}4M6C^X!04H&YV{?#PI6flOqt(4(sobDj8hs}a$0RahAWa7X
zu@oU2yeRj8rx@7t4YWTAQ1w&wdOEKc{k&*ti~rTcd9OatzEEDn*RQv<^B^U@l~{0G
z1MLfRpPMA0Ja)8bou_7i+cgEHQ!B8Su-8+#D|z|t24C>;&$M=@PM2LTU-p>Rvf)ez
z77cIOtl1~Xn%)GdD2On#Bt9V-naU_TD%C6IxLCkjWSzFCS5Kt8zsq~;XqFpS-G_Cy
zd4A#zWH3jDH2E<B9b=xJ^<<VfR0b%BTt2w=?!&N7#iH!uKBp8x-dp|eP0yLQgw@9?
zJicILcaPLyZSMBEyQeUHyUa4V1C{esU)9HUBH)LywCA7s{8xd24RVxNZYVVf4PJ*A
zy1IS-?xX9^ij4f#@0myyQZ5iZ9{XzMkRwfq8DVe8LApbmO6I%Tp|E^-j_6cdo7s6C
zwS&`@4rdo0t^zt0kN*56A{2XvuTVBO$#@~J<1j?iyZCGhLCr>Eko5jdNyTBEJ?;W|
z0D$s4;T<-NxNN6c&+-Eq22;rbM{BzrOhuF?Yi<K$LBYY>>if9ciO5G3DdLJ2T&lj7
zx7%0hb!n(gN!Q}t(%=dZ`#J7_gn-XdO^#h^AO7Hn|06&Dp0Dms0x%fRZ22_)9n6kl
zF6dRn*`Imst!de*TQ4k>qh0OCZnPqw6hMBoEmzoKGg#ch<^>{Z$$EUq&_^+YpOlnT
zNR1;vWImNmKHqkhP)|AL=stPW`Bv0(>LIK91Sn3b&Weq6q~sOg(${Nx_+?&{c!kpq
z#=zAWjP2rY^ZCHc1nAr2IHQLTa5YJu&g}4MdAY4bV%W5#zN|ess);C|fGbF}I39SU
z1u-zK1i)x|z0jXN*JAlGb9keON4JC=jkfShP$gfU=(x7)-oUXbIGa1{H+|4O#v~!i
z2J7M4q0?gyP`j9v=Xr)0-Y5C^q<Ju(D1OUWRB%b;>b_$GG7c5@gR(mc9{6eViFDZa
z=tV<{qO<V$ff<+8*fA94xB@}FY0047klZ6}b%^L|AXYdImmj~6t)YAJ?qhJl4bNpr
zpzO;A=3XR&CsdQD8zwzMceQ7%yA&?JzpV6KnRamlchr75N<weHAF5HCUXR3{{Ss$k
zVd>xvU-x!f)HD6O0GCdxpV}95+#xSKZ&c@l*{w1H>(YE`cE;n^BlGgn!|dFCS=vfX
zBfXI)^qOCkC{!I~yRt7{2gKIJ+5FsZ<k%Ydxtai%3GS}ol<+V{Rec|6MCHVnkM9~z
zhiMAjZfjG!y3cLjtwf&_Y5$C%)L#fs{fR@Jom+&R=tpevZ0pf8;#U>zk()=V0a>py
zOxM7r??3;SL?8btyY$c4us?5WA=ixG7Lp&J76W%c8gTjk^A~!=?<Q9s@Ymm_+IObp
z&&|r;`(1PxS-^h-*Mpqo1!Acd$=ipU09>D`aw6tWxO~<<Hns3CrOe^-GfgYeM6zL1
zEgx|?O$cMlx{yb9hO5nZbu9*XPC~xw`x~6+mYcjc$%|P}^s=cS={9+zBHc#l#{t%X
zG@b<<hPxyK8*zao+z4<Kd@Ka<JZwHJT_g}M>)aMz>j3)X5p1kox={%uL8lNYSrE;y
zB<KlGbt|%uA0hS#nZbxe*Et3aO@)zNbb|DrT_3oneT`pP%d=MN@l7XmTUTX?2ueKS
z#R=udBmtuY&E5eGQql8Kkbti8X)JN3QQV}-g!_S{7q^8DE~Kf{rr|sn-K#b(RxTrJ
z+$wWXCV6mNfmgI4neS{L+?PHN*3w?m<!j`@6teb-6OeGOqLG4*g5Av^&WU1Nnq9rh
zqGVwWl2Uo-3dQ<`nP<d2@?Bk)g~Jnd53wPA3MRO0)9;<y(3Q`;x@v45j<23#239=a
z_{u&}FK@^>%%!r@F5kp>Y(iJnHU<bY@_I5iy6+$w5l68Nh~xN$9i^E@1ygY}k10yu
z@+y5uwiAy<tANZWs|OBOZJnN4@ISrV(X<kx=(MXq{)`(SNJap7H@>Y8=1*yY#Ffpd
zsn$?3Bs9(ue)WFOJ<Tn(x9Xb@$3BJ#8r4&U$0czt<2jMlas{wf!V<q$w`<|umFw>-
zdpAVwtk;&>mrqgZtIP!KO$3i~jy#gG<c)3C1<hwjD6<xmT0&P84fY@wFUk0L)O^t1
zUZqQJ?gZeA-o{d~O2P~)Y-orx00mIO`$!tTytV0eO17{yyHYmm8%R7E?ypI7gbPiK
zs|VlvvT-bbPFSm)ed`q!UEOyq0Hu_U({a9Nnjt@6ZKV#_+jt_nfj<OEe7caj71CpZ
zQhnc7I&zx-Xj6%?=gW(n1=DWE>6*Bg{a^w;;!;tjHo0Z2y#t{^+-Qj?6_w<0FfJZZ
z8E*w(v!iXK?VtzygQH{vojG15ShAg2645OB21*AI3)B`d`^xiCP(&MkI;K0~Y(e|q
zn==3D-y;@iN<bFS37e?(M{nc>vqGN?jF-Qe`{uw_(&l+g2;q>#ebe6Zlm|}n>X+zO
z3nj@q6u=)=Ig$t{6BQ##MtKYcFafFkMqFZmErrrpm?UX|#>g!yxVh;3(^@v?GnZNj
z?&pBD%LCxqa=M}^OpySDDU)^_3s7tu0E+DcD8<MEm~cT_-#{-CQB(k@OUjom-yCRb
z#t}px)!fwK;ZKy)aWZkb8?B)^D19IKH&5lLZ<fl5B&L&Lv`}8rZy>t60A&~89tE!e
zblvlFP%;ld*JXn|)jPmX5%%W>T-7iDQ}y91mJATt6acm^qr3_2K+`u+lkpVm9)Qnw
z2oErJ?*NQlSxK5m?2or!p$Y=>y?^J5uKc>7Up@3|4E>sA|99=sdwdk1*q04(m^TJU
z%I1s=V-w0aCE`YkFCKq%L>GjruaS=cJwVP6%;~&tp%wZ+od*F)m7jW{f8<??vQJQN
z?-=+0nU>Z5RFVyp69{^?L?fiJbn{lqU5dJXeCP+R3hkW2#qZhQ4z)8bcsPO1j$V^j
zd@wWPcF*z+6la%d>xRRN^G_#iGq<7_b}F|~0-jsr&?eXlheIlL(f?#+T9XxTS@#AH
zf7llHNlaP1?E4Xff&)qFnfoxK7QNW@ta|K8If+^jxkQ(Uh!rc84)Dmwdj{2)%X%1$
zUX>m&-aIx3la7`;xb;AGOMT|Lm<rp~YPt<Lo4#)}rXL?%kvt4wv6Zgc^;U`cL(}8d
zghTz0lix8xR7-1#!;fLvzJ`5@HwWS^Ac|6^!%ex0hRyr3SJ6&~+3uS##qZItOw0D<
zXvSL;@c}|arKM_fa8%Yn7Aei@D<<l#Uztx0>BaNm?e2nFL(yv>N#W0GoIoDT7U6t;
zj>rG-IWQ%0zxtBJ^U4wAn2z>0(3~!D4cmw(X43>bO~#UH#)HXTauG5JR^Q7d#4qO^
zWDiza4zq7vC^cb{ZhzKpalkmNgC}U2p+U|FMnpI7m^6wjD+x4ke@?S@DzLAN$=S@W
z5zr_0btaq;74K0$zw1{s7Nhh={6m!13EheScd$v9`5M?^egh7!C`-;B9Zh#*(;byC
znIUJTQIyRKO96N0#omasKlgf72@tN#u)~NcgI*ITX+dr`nvzQwMWUw!*ziWZ4O(V=
za}donpe*Dd4~<=vmj1Glvgla4q_%14PG8FH_Pl>JI1~T};*X!4oOu3jk@1yr&|vjd
zKtatOxlqc#BLi!8>wGNFmqrrSCN4ESk1m?$_M;UdPR|-@fnwT>q7zrHevN62x_30?
z%2)eSEIYltxsu+aTeGq0Eo(kVH|WR;42Eqlhor;N;WCJ&s#%0<?y2q<^zD^EzQKMH
zKF{g}wK1yO7jLZAtPmB~NA?SX#mNWK;XZop<FDbn$lSZJfmVPN3Y!mX)QGejb|GX8
zcU#5oR+ae#--@H-5?D1$rl*@83A{4)MEwQXM(Ri*j0?n85~pEd9UA$Cc~UtMZ#??h
z*rO{&AAxJ0$Du!c=<z>iG3;xed_CjP@vGHq1+KfV1G=nMzaDsoILlV<n5f3N=0G`8
zs9Z%DG4I?ACfbJHsTzFcz{lN>VMv3DGp8tW$U9uy4Ih>*{L}z?<gt3>z-1V(P}?q{
zqc%EcXimL>#+|t0xXg32)U<HLHsL%3a*WSFf3-Ba%4u7ky~v}*3fwCV;B@lG(P5dy
z3;-xU%HEf$ZFNSd!#c_5&>4>eM4R=lz&X8xBTfzslBw>VMsEyrdvbWpAj=DD3)6ll
z&fR*InkR~fVj)ONF&{)Mc9Py6Kx6nxeFJHY<&O$hXMFxFRGuzAl7+jMz;osBCl}O~
zL+tm>a>#>x@QY_Nr?R|`G6p=alenzB@!8H<&U`CEC?&cG?O43LPYuA|All<-1v}tN
zWAD6uaWvMxjTA*q-q$aYyi?Wv&YuGGuNtdA_BzFZdLCuH&|M-qoq8WsN4MWb#POwO
zamC{gLQX?;#0ff(_P7}xdHcH4!P~q0>tfEEOz|uQ>}Cag6eexGU{?Lf5Y^9wwf@oX
z?D>za{l6oMX7`f|$6s4;ME=0bHWL3+Xy(!7iuJC1GhIh$`$%c%$FZO%UJD=DJ`MYx
z3!R@G)^c)KWgk{+IjH08oZ;A$Fc9PH(*GWdi)>69Jwnz*qDQmd4bJIo^$3EvxGxn>
zyB|ITDr21xWZYs~Evq5<7!W^)<t}h1Hz3&{_Bg-lWDhB?9m8!Wwe0k=!a3o?hd%7v
zF;SJ@IM<RSseAn!NGhp}J=&wgK+bl!+gBk&EhQ#D=)(T8>JN-Z{q=;lz1>@ohUPoq
zQ0UF~YuQy=VWyO<q18uOfi6jcsHZL&eXCa|QrDP60m_w0h_bMqvT)xz$>$Dyyu}S`
zc}-f~$Vlwo9?Jj&wzrz}Q;J5qT*mqix>}aAWo(MPXEkqJ=%<d(%<h98p~xI~g$wrQ
z>&nE!g4_qR><>QutaCHFIDK-hst{J2ye~P<vQ$r|SEDPd#c{*<<h!7+EcvJx=+F2N
zFmEY7DuchQpK|%df`R=iR4J<4qOjEF>zjP86JFbAJwURK4`!xXWaf-I+S{{ZvbM`?
z2b6`s<n)(*6(5Ud^;{`f){MW&RBX{CdPJ)A)=8g<9P_GNyZ%ZHHqH+Y^E#m7WpX%P
zLR}@O$=5}dPPqht;3CI6cR+UEKx77HvQdp(eWh`%g7%%)_M1`{7cW9oPdqR=IdSsH
zMfxe|$*wE>w%fwHt=~X?5sC>^Acs@i<a<XI<cFF>tq4;$&#vljCD3-+NFl6yHj(kj
zYSh-(%vSKI?(U*VGf<ppkfupnMdI(q3If4Z@Jcb1Q~%c)fA!=4^P|TSRfpd`i&Lip
zIM^=9EBspOzw^w0@1|${Y~c7GN~(%={k&}cUu3jy{lS@gWB#Y87RYyS*VZn%R*0HG
z(aoHDOU=>m6?cS-zZ}h*);g0lA6XGjCw8Uvr1wCxztR^O=rK=3laoE3x%Q+)alZXP
z|Ch)W)~jdY+npHWqWfnI?ds`-L2B2=;B<RyH_1!iK&RnoDiY9HdNqZnf3Xiy>7TgW
zH;xYkv$(=M;jVs?rls#l8%@n|=q8u@r>TLgRr_?D4>Z|5Zcx5@;EFP_WkC+$KZop?
zM3*u|E-XZT%#N9R$>2E4VR!xVl<s9~<CjE4{ZK3?IQSK*sWd}GRRYKD(EplzmP@cM
znoO7RG?Rby9Aaw3jdEaU5;jAFKHiQcmX9OCeWl#S6l93~#qj&D$-8=KQf4<4^L5X#
zUp9*%hH)P8ytDwp9;xQ?8OKIgalh}ToGWlBk<If&xG^8K^N{z?%xP>4^}_5+Ufg=)
zd?RmEqxn2+gO0{ZABLS-GX!*^fQ6n^2_C=V=|e-xfOmKKzkzsOx_S)I_L)qyFW?$6
znOf*=z?3evOk+P_57xa}NU{fr4138*D+F_xhl{0WDvy2Uqx=&EXpahk_`T2ck4k<7
zSE1j)@Bl^eyVJhGCD(xu*;T^(U*<Y&C9$&?OK)+Q$enLty^t&RdGX`kTu{8`hrMrT
z^Oo4QCSL;tb{oH;C={`!wGv2U?p{xeHdr&SEbS2=hozuxD??7joU!Z>xUBv1TA<eP
zN0;BELhDyR3*OC{S#zU8Fsh-pX~t%AKW$WV$=GP*SR$L&NjIM4hdf~d($yi|Gm-jM
zb+!OgVZ7|6*O_c|lt4vQ47D^TkXL7L1hJQksaUUqCD_=Se-wA9c|QUO3O*ukQD_dx
ztxYgZPW_n|-yT)(fiMaWvAijDiT3y<nJ{G5?@B{RFM*9Q8H6Ev`cKh&?#;wpDN6pa
zyXA$4<-ZaSi;gZU(~)huM>S`9-45P~)~b%zWqAXPi1_Cqg?>q3ai++TgREs%LaMDY
z&Xk1eia??T49hB``S+BSsAbUw9$~(FmaZii?OH;X9HU-a-suMj5*3xjad)ahd2%h6
zM&c}uRS!MA8&j~S?o`06DgYA+*)zVfP3Z;1y%v!a$482LBesAwRgVuq8KF9KX*6I4
zG|&F;Z?>R}2-sourlyBi_4?)&=A=Ap<s17*d_>>c!y-3%mag$E8u{a0#{Sq1bRT}T
z`u%|T)f0c18ozqt|NWj2Eg>ra!i78<fUFgM=gdOS56fF2n2EyQRV`i!aY#y*>$bRB
z<<N1ORpDvg$Aiq*eE!|p?VpW){k8A?^@vzgtpf668jxX!lD>gvD`0gCqsl)VnWY<{
z9rVkjv27M^-t@~~GT4SxB33hpP%UhAzJN;*%s}Hx9xtby0RxnSAy)uy=$G8Ft4k5A
z_1><kMveMqF8SN0pf<NCY55p3zid8#@N!dhdU_)i?rct2BHSLe@V$L~*!si_vGA)u
z@7yNs0)o8>Sbgh{kqe(@!tzR?Yy$hZ07lt9r~dHSmFXIz^z|iwDz|QSns@UkHWs;1
zIqw;bYS9f(O^%XuTw3?LFTD%p;b*GmkdtStjS){;gk;>SfJ``dp=7Rb+T1%Fm33S>
zdIHUF1*<fo;$tKQdoXBTSnV<HTjqZ8>h|!1y-d{~r&*>wJKivsu-+ff6c+P)jR@P6
z<GUEy)J>+Zzep|;T%8!O$g5SYvwY<Ot7c@mkP%wcGG1|NSGn^bK-}&05z0I=NSYea
z#C}RgcD+xm9=td)t!#ad(wvjD-pda|&WW!3JM3<pYd4b_K!~3~RA7L@Yfxbd11Tq3
zk%L%Pw^*&2JHKNa;?FinPQqV{@9yL%H}I<#GthW@NHP0p<KCAh%H~sP68r$1p%KOH
zEB^&kT27*5$fWL$Pv;Y={L~<#GV=F3Dpe<|PjkkZkE%$)&4Qz!`R_9IEsb&!wE-?x
zsxGr9PkrU_PgQ5&7n&;KJgT4FO==MEjh6#AE_x>zxk5wDbwk&wHZ2-tm$4CvcJpBw
zK$c0v8asRtp=8{9$eAl(<awYUOtU>h$8v?d=5NssUEZGE&Rk+$dIi&^&{I&f{Zr^5
zUj#`v+}DXXFp*M4Y8++lsg@j{c#wKicGKS~OF|m+tiyTjdN({?(_HXqwM8S|lX&@=
zuRe1fG&<ulIh7<esO^FnA5lH(A}SkV?PQHIoN4@mu9>Ov0~2(CON>GnO1c4nOe>}V
zN%4~~h)=Yp0YT%=Z56es&8#;^HAF6*OvTCtH#K@(Xdmljype2jbmTKy;(E!oqbj07
z@HNG|XE%$dd;z_S2<a-A)*N{OA&Z_&!!<i~Ux>oILn5mX70uMaHQ~BlKj>Izv;;-L
z*8T19JpSl|v4JepKH8wzD9^5C`~3jX$|;Yjhn~$&akl^m^0u!@FO6v~0T|7e9t1e7
z8PDYAXC<nC2M*s<_$pAgiu>BYU2EnDIW_ohjP<Ls`A#84dez|om2zslycU*SE2ph|
zUygw}zQl1?*0+vhxxwww>dXm)IT}xbChe2#Z<YO0x|eX>4U22fXsrzJ5mwW(?W5cF
zi57Ile*1-b6ST$SSIL*%mwdo+r9v;ywqt7HvMXy}rlgz67O9AOn;!PoUQvE}iRkGE
z0*kTJ2PvrJwp9<uHLyo|#dg`i9wMIL5xJ5!t{c8WF~t!8Xf;8!laWX3UYG`hz2eN`
zVqC#sz_{tXwXB3QpZ7jLyn$w>DD+T-(GJ|BVZY5{Zdpt&qm_fa|5phKug(=gExpc8
z#e@#cEYBl74K%@CfMHEOUhnOQ%_I4}8GIWtzw(^6X@p^<Jo3NmeXF;68)ZU9rtfas
zxu_`?>n|351JO^_lr8(F0QHRkfr3V8$vpp2&&^#UT^?VFeoX$e9-ZO<xKLTdU|>TT
z<5Hrq>yyjYt#+T(K^{BcMoG>&amhfFrY-$?+PQ9u4E<w?f<;<a&XCqCRrecj>fff^
zOWU%ulGU@}Klg!cxpOSdOL`Y}fG8JYUf<tlDa&B|@NPu6-YbY3wB*j+hv8iGZP$vf
ze#mj)V+=FyakEt3s^ep??%9B2akAws^buPQB9hNThgj|$V|Aq8!|J-XblL<i&Oavv
zP_dHR5a-SbR^p)B7d63&MZRsdiX>q`X3O{L+pZ0`Nwdcex={(NQqlh)`KIO7TeEL3
zGfZ#G-i4FI!18aVV%wm0N)|{KnjkEx@<5nSp5M4=6`D^jY?L-@&rz9DOlwb_N^&jp
zT6*=~jDs#<wcx==m5b%YIEqwvfv=q3_z)79F6Bh(>Kn_s3nJo$^>5ysjXwp$Ji4XI
zMZf3D?0)ku$}z<FlB8CTE1J&J@MceGUi4>_tqtMqgDHUg9K!zf;Md0|9QBUyhPm93
zHCOWIuT%A{1wa=8ek5VPDsWAn84bW??Afi>0%(%Qs^Jq1D~CKqCR$v`-ULgP#{L1T
zv*qvJ?ZUr-Ht&xGa`v4GS?U@&(JFL$^4gI5UMI;Y<qMu&6tf&pk89O^idrz{;M)Vp
z@zX?u+2X6YZL69*It}_2U&VW`M~9n*uX#O`wf!oC58XjTNUL{YP+_&qed}RWKASx&
zft`*65}!IRL>Jyj;Yx4yy_Y1cnCcNx%R4tJ|E_0|qDkDXqn#@4`6`oMn_o7qa2Es=
zrg6A-0(7ZmZ>DiFvQrH`l|N{o<B<t>(afHkTCu{rC(pUqj>m?(NwQNe6Kl&0Am%s+
zTxbfy&4Ver(Sv<Xw7GzAPQXL`W&$Sh*|8X1^SOMM%Nyj(TiU`}*Y!h;dhdbud_`vV
z)b<hhclOVr$QJ=C5x1a;IVYPKMR<NULDsGSf$<EZ1$f)i7y+2lq_^^SP3I5)y%w`{
zHWAIcCgI3gj@QVsHeH%}=91MZ0DNPW@(mPp4!q6xNEXJuKvNf&q*Y4*Bl13m-+{zT
z-MijN;nD<jU^Wt5TmU>$y;;uj9D4$^*mA#qu4XAZ1re_sS4WTFoDq(Txgx(K%Jai%
zW52&{(BP}$?{9PVyQ_M^l=|xw|M7%=^~Ik?>#x4}+kNr6<bD6W3e~E&#i980ZWZs<
zOA}G&Laxs6Hi7JZ554$(fcNKe=)f-G@YatUnPNw9KXW$R9=m!31Bj|<Q9?jw|H>rg
zAEDB>Fl+-9?+)4oQUpXo#Nbzj0PyAj1-T<(PW0Z2{ow#;3mM1(y+ihUGrVZ;n^tk}
zYYpU>+C#9&q(3uR(i{*0+N$GcQID>^YD*Y#<WX{-wZQ8tr)_$bwVw&^q4bsCI#%)c
z-l6Hd!&A?K(h<RYieCT+H7!lsV0W>nX6Q!V<i~A;aH&KC`JinRu<ba1Xn(S?drLFz
z<a=Ut;?BH#Eps~C`bS3=sQks8MZDeD_&Z6*FFd}Hdng2StlMOL0E0$C=3am9_m-6_
zfG?~q%^~`Cvs`0$ug){%YQAAUT<_1<LfA9@dH{=I9#Zbxd}(u^Jwpdy5}7cyvbIHp
zv^Tj}?IG6?v|wJ<N?CIYK%A9*-d~s`1yL__=r)HyOb6$6xg;+|co;!t4|cK3&eONO
z(Req?OLnB4fvu2lV(MPh7RYQ8bCSR%IxhrSsF^e+3T69)*3|YWK&(u=e{rG#dL%P+
z)KZ{%XDiGtAFgZAyR=cG%nEy^O__$}x_7_?#gAXG5N&Ssvb#DgXAl!PI*QLi-lAD%
zgeRujBV`71usfG1REdp3tlg4|c`Y$ICVG5d#;%dJ#r#9OgTMNw^<4RdtFA%!E_!uT
zst-oLEPL%9dPb`DpMOMYG>&32bevZgATe-HZJ}P^=GnWr;?;MxtU6c+q;zhY*>h`@
z-(%x)pWvk9?yLth6-pE=#YR)EI*s#)ZuBns^2>|PJsB3Pb-k}USld-B*dQik&622h
z{4gjNC%fW?hFek{3v(c5vtP3d9#qr(YGH*n#(BH~v@hO)^DGWQ^>0yO#pFkx_L!9=
zmCa6!U#?k-YEDA=B4KjdheD;Wuaa)9gK?skebxH5Fv-Bef^{Ci37Yq3d)J$0j+ag@
zL)nXTfPyANuR2V)ueP=Da6Oz+s#YM5P;Lz1i7Bw4tA3`irEy&`q)qx2jWfII6rb(w
zwMWVAYt;Z*Qnsz;ehx*dnAWlih_<5^vB#}Ei+H=_I^0|~Q5Sf&6dRQXgg^~U`AL^9
z>kYafq0AE`nI2eCM0(AF*U&eRP&sW7nO;><0Cv!p_hZ)WD=M`Y7e7Vp*m|sZMY)@I
zJ-~q=T-IDS{n5tki*pc_w-e)%3SBB3r)u3K@qD0h^=~N2)o<0F4Qy{vJ{xJs#QQGK
z$0tl9d2u!2wDpu5`0~ER8E6BImkb_XSP3^dHfmwnGLd7Qj~gjiVNPlr9}nNhXYpW7
z_Y{b_J$ZR@j}9+=ouMbJk^2amq|!t?)guGvgC9E8Yj$nWye)*M&9qSe-V+H>l9N~v
zzt1$h-8g;MUwLcF6RjYZy(rU`g$B$Kv&Y(2jYNFyMD_h$Jy-^PZ?)3v`3zPfEj@QW
zNgsdEO_!a4b9rCF#WR?vtGEIaj;m~z1l#a)aUc(RH@E|md*~|3Hewwv+_2{0e<q=^
zER`uFd%EQG(1@;33fqK+mO!LN{L!gkjqS4#dwV7uUsa%XF-brhF0w%2e=OZ`V7gU$
zNnQQzby-|U<;;|mOhNRqQz>R+cFzJY`WuO)HU|(}Zl#%51vTyzCFad@W6H`Axu(t6
zE8)ZJaX7Wj@v(a?30wy63L-|W1J6H{ag-avj&$%-%V=mHJ6oRB1*{}Z8?lCehnpH&
z!HGgJp)EIJ>)h8o2Pry=q8?f4w(zXun4r-X`Qc}BntP(uT&6@Hb*!?ygnX$zPPvjy
z&<zIUCtnxE<__Vl?34R14Ew!2^}g(V{Gn@F7`c~rt;Ql*ffx4>s@pZY{JG}}Gl#Va
zg%HEO2X50O!JX_$+s<A|?R?zwbHT#_*Ckj5H97grZ!;eCM$?gVK3MJPK`fS1oe6em
zH-!;nwTeKsJK`M;ou{LVW1mU~pPdJB39jk$I1jGm;O+2J{)h&6L}Eb&vhe<nGfiRH
zJDaT;Sx+E_Q2}b`syTvoI+ko<brf(`u{14hV4r7jY_QA1mLef}5cvfP#PD7NXz#N?
zpuTCGd4~b;j!W?rRgk-@3D-9eLu5VwI<OCEwa1bpz~fGyDzw0FplxwEU;<v@P5(Rp
zNRQrD)d@v>Uk!N5SCN#Yv~QsBOd1DAFnCrHkb>wzl34frBbVY94MZdal9GWsSRBZm
zi@0Ls_g7+6NIPZcJrF6D{k9@w<IcMZv;8&Cm)p!W4pUxA#Xen&^1u5?p-q&B#?sEe
zJ)Q?pQ{OCn14VZMK=u2v6e5tQAsPYyq4{@SJ`jg{*Fy7jK$1lO=-Y%Vvfj)ZsDUg4
zrm$HeNtM5O;Mj37AfuPTjiMbc!jeb$iAknFT~a1pfWB!01rk2L?Iy@tV#t|LKqDkD
z9}T3kC})zuYYjjyu+C@zNdEoKANzGVzk27_*!f#COWT$p3R!p>cnO_2bV<s7>zFW|
zV#!Im9G5HH7yf+=)c=0$98emzobtE~yYUUgPm4^d=Bt_b;lQI~F@3_gYQqoE6Eav9
z3P-P^ieJuGbmZ;(fjm497ZB1t!7#-GcZW>1k0R1?{XNmJI~5w6Xl^31jtLMaLj?(C
z?0g!uue3Pokn|<_j+2LDQQ2~e{%WkUfo5HsYRd_&nk;NlA<ftd&elye*+A%E1rQBj
zZt00JFoQ#_U~+3827XatyRbB$Co=Kzjp$Xj=zDgAeG@)U=nGv_XBtt=v*5dC=rfh^
zKq-U|sqp6m1$QQnAKQD`beq5Wfn?3Y3@)-Q0l#N%M`e!xYh^SLsoG!m*pGMPQ?cst
zYxnM@3Yk8C|Ecp?+v|te^V4X1FcTm@PlJn7q=eGmY<si44s)TRrIU8|S2!oM6bp`^
zWY##^j&K~gU}R__0aZXw;;Cr-f;}cZIe88UY%Z_Y#s(`(nmX|9Cw@j3YL@8tNq^kn
zg0Bx<UfySSi`_Z5sFhDg?s5x#hA0O?y;*dHv9%!m;*5GsJzl39al037log;lvhZQ`
zvYDPVj60O`{<7BHidouT*p5~D_@MYykX|p`mBEjF^&sFj&YcepuR5f^6gp>e@%-wD
z`dv<!^s1FW@4zOuGL$T8e-<D*W6kO_y+=E7O;<<`aXb4UL$Y<oA(}{e+sqWQer3IU
z)OcXq-w8__p(66eSHh#+vhWTUfeN(9aQG=ezhrHlS61DW&n$7T2FvPQt>wcxQ-^gX
z-tQIFZNRdgOw)S*zF#-lg<70z-`|h(E9hw#fedSM-r=TqzV<@w!|mfkCymP+5sV|A
zRSg0<CGng&yKae}=UN%ccyeDmeNaq{3TQEqcSB$5t39~DB9rlQU)L}m=irU0$Eg&K
z&e@A`BrDr>G`(G8>lL~0{w_oPc0GdXhbLZZPyP(J4k!v8v<1w4hB#d^E?OwT;Ac`_
z-oKTcTG$=72DrL*dTl`hRLG5+5e*M-Ifgv+K-Zpu%+k*HEy*xMno9GC^L6TJVHiKJ
zD6TcK-cH|MojClaLhLKoN4B^z;Ouf%yuH4n3Wf2IxlQlq9%ZEAVGGBkWxM;^R?aIY
z+mxr>_a!{Zyu|6wzn`Kz?x)fO{VKnB%)V#PY;ZQNRUoVXZIhE=w|Qv*&jX+PcHOG`
z<21&e?kh~JL0aL3YW+3Xp8KzKz_x(R8*y=IV1grfK&#j7(8i&~Q#~2`s$%temf^Nx
zpC86asffW&8e2aaawmF?5fs4DR>x|qBl;B<J}&c3iH#~mi!2{UNF^T2+9mK^b#5>D
zG=vp!gay7O3lrp$8#y?b?Co#~vmGan!v=6A!h@3LUM@;)2ezL^zX5?JZ~IO3!tji4
z+J-ciZy*sL8QRwY=?)#H<%tKz_<L(~l<Et+3S*2hCW0;znwHQVGO~kOzw8Dl(O<Q^
zHy28xb!`UN-QShQ+OB;hNP-c`)2=*nw!)m9Y+WXgRf&>%<)m9#;pkFLw~|7?S2K<B
zshq8-&ef*C^8qRMcHeQ(y*%4^Vi$gcZ0G}U`=Eu5G@&Fs5juDKKMP8rZ<6-_Zdu|C
z)e}sMimCowBQ-`(i7`QK0MBgaqC8NN><0ye|CCqxFKEZ}Z)HpWwY>1p9{VrklZR*r
z$;cahiozaBv@}~Uk<rS>5+^k|&&Eha#q}g?_p61GU>_<caWVrGQR=mpdfZgE*_^US
zWq0BJkj3~5rf>5FEUv#hU$6M3xxnn9|NTx)BIkRe@(XL#c-^vA3|CpG@-5GOV`KgY
zs>7~)*<d=dR=+~r(jv74_+G;$zL}>TtkueoJpzHGgbnEnCCQdWvx8+aVVXy!9@IZi
zd?nwj?_*EKwz?s0-JfM=@jw>HNQk|Ew4>iSpl}d03_D8)#A;6<DcS;_QVnfc>GU^0
z0Gp23OF*E!7D=`q_Y=@2SLC%5_(9~*8G^NEtU(2zcuc)(xx7mXQ_+%&bMq#Fj`3QL
zPh#BF)zFcJX)Kek)SDJS(YHB2$MXc)cxLz7lCLUnP7si5{it6cQLDA&Ot4H%^{NLy
z7dY@Riqp61W!Z!qrJMrh>ns_5T?iO6K!HU8x^`osrt?iLUwN?d8BhLeB;%Cvb`5Fj
z-Oma`p394O%#3(dK;s^uhdwMy3qT?Cs8U^K@+RWo9QY6h%sCH~0Fcjam@V76rLV&H
zuH(e*w5r@gcfV-X2L;4<q0^3kpk3^yoFb09b@0eDTJ0<887uBnq(c(S9$?PcGtP@I
z6`l6cOko6RpI)Ml%AJCy^(*ocZ|pXrSPG!cCRXR;wIo4yNxB8*^8-5qBbHfPq4(1Z
zipr{Au-X`krD;v_rwMzhURHRJ2KA8JmaI8q9D8xJ)n5WkNwCV>R2ottjz;F&vnMB8
z9VzQ75{gnx_1h@S*LBn-UJl8Mila5W-IDR49mO&uztG<Vi~;aG$f?`d&8Nht^2w3X
z)QiXiR|{||uhfsA3ShOYeTw&@&wH`G)OeX<+<jZ|wVd;Nd&Q>xu}?Lor16JR*0Et*
zP<9F(0n5|q&u{OZ`r!OrvPrqIeA}{|PegjZha>k8hk-Q2_w}bYw^GA*w|rm-K-zQ5
zea*@wMVP+=xrCkZ*;Xo8+7OC`jivh-4@kE&N#4xZz5ZTHi;OMl^G&g0O))tGde9;w
zdn<0uI`<?1!>H#NU)ccx$sl?FS#caHL{1woECX0ZEj^uMyw@fgl1IgQWS~cZu#~}P
zH$UOsvud|W?dpkVwpPxR8w!;1M)2Zg77!uUO~=T#z{W0EV#%MYgT%8`31|;7iD<kk
z>28izbi7yUc3^a5xw`&z94oW*-b_zcmC>nf<Wl)JP&bOUcR<}%TyhRsuQOUHS>s`{
zlkj7wg2x$%TU3y1CxtHdw^GU>dT-dSfEYmv2HnR17wNk>!cPn#`_LZA+tT&`0K&%p
z4%R*Y_nX`sCaYhAvfF=&=FmMRlh{*ww_hncIBAUKLhk{@r=0<)=>T-~vORKR+~XT4
zQ#=<0JSxcs{Q90AF!2bOCKOA)Mp60(y22|3RHB~vtrE3O1gQB}qhDR{YZUzNm<X>+
zhydx)BD?8zf;H@6%8;1s$CwJvODFF=)S$q0{=Ee0e}SL+M<4i4tcyR_5@G+5CV}?+
z@yH8i1X<wqu$8CIvIYj$zK^4b(S6~N;qLowg$Zj^%ZXdEU_NDldYxFCexC#^blxG2
z`TJy_z;IgaVo~mIME1Ad)d1~@*M|85#xhQJjfTa41F5#$az$>-uK?=DUsJM>bB>uh
z@~y(<+d@F(78_`=glNj%8D8gK<XPF_+yL)x0V`~?L;AXaR%)AQC#R*kmY>{Uq4o2@
zObe}IRC^SszYQ*CNKJ<=_6n@-Hh_qXTvim#$6JtRAha)izKA?eAI%<DIKOVv)ZMa*
z>O=eUZ|1Yy=&DOfo&A`0=Y!r-^%t@aDZfPyz=34WC*y(*a1>PplLZofHHY^a47_M2
zsaTb8&kFyEC2{-3(CVweJ;RK`KDDi2ED$A*H~9t%t$J?3qnjf~)9L`=7gZj+OY1h-
zb~uSNw_FE4iR!m+F^_ZJS61u_I;n66ep>E0c=*L$O<~aCR3+|ZyKW7Nc5W-rqqerB
ztfSsXeKQhQjZ;Xm1gj8Yjwb005IcU|zrlQrBHiz+5h<Hy%j)g6ktN&7_#`--oxW3{
z-(cSCs$S4!5z3UDII#<NX{ucJzmTdV+Au0aTux?xROy!!=fbpT{_s6Z+ue&Za*nh5
zPx5UQ1`bz=Akq?c$Q8&7ZA_5gHrCkAorU$-O{Og>qc@7W3#4WyWI7y1&l9JE2$7OZ
z^T^{%r7Opvx6S7`JSu%nJh<kj>bt@o?9y4CBH$leNyC{9N$J?ebFTCl(l8%~?xDiO
zNTBAML$P_HO}xP63{x%biu*YBtC&k85M^LH7c50%ooR|C0|D*3ofJX&7Rs6T_|0gP
z1M(eW73)*Ah`qO3<~!tD^PEr;TV1h({1CZYGgVYM2Y1vmu3YekqJWgzM<C4oWefY=
zV6mgs)dIpXu|NnM>yD+^0RDX(Q1HWDoCLM(QM8l{TcYtcIQA=_i0?im(AQwr@{oHe
zJ36NkEpR16t64?OrQ+RwB)hK~X(~3-*QKzIP!%c8-{3F$Mf#bd+$Mtx;}@%z9PxVX
zb5U&?a*iMMd-HA23>+>x*LFj+5X&A<8OYJo#4hYC`^bbvNb?#tnfP<PBPTM=JQk;u
zX^usB+}?T;P=03o*ovN<gflqYJ7O057J;3${*A>kg8l~LLdVqddk!_B4vdkGy&aHY
zX1fAy!8O=dhG1Uos`f=w?h1|UJ63C29rfV^%j8**NURyrB)nQblVV5Yi}WBVBY5Wg
zPZK+6B~G?7>A?>10)ye%0WU&MT)EyIa>U~KiK=4_5!3qRM#y@=-Gn3i7Uv~2WoCe}
z_NWwl%g6S+As~&sjYocI0Wul*Dkx0~B7G~~`ujub-?f)1Q3d^PY+xoQFFr=Hd5#*|
zd2I8p^0W$X+0<Sw-+<lcw<q1Ww$(0t_TFg*vNUn+THiou1}v2eeue(KJO7KG(mNpc
zs{t@#e)YreXU?yF_{n~dk;zDKgE_~y7D6pw&7A-AWS`@O&tISJ5`B!*SwUaFf%?$h
z>DsYWbLzR?zhIy0|6@$nJyt)e)EJBZAQ}3CLHf^%46yzee`ot6R4VCDVYw@OKppbB
ztqj>Or;3Y{4g#c88}WyqIPBCp$M~G<tA_yQVK{PT4TzLr-vJrotw;nRZ<g*6899zP
zO6E{YZCKp7zmoE>XB!2NSG;fJ^c&^fW1}$ZU2kUJ>%QX+IPQ6uYf0T!g$su=1a0m`
zqZYSC3YM=WJhBX*hh@KdbV(cX!9C+7kvbak#*?|;JL$2<!vPO&_|PHv>y+moyQ|?2
zM$aoGOrw$nH*(hjIU!^y99~$845J<Gh8H@jP9FWb6mt5x>f~cR@!Z1%$&uv~OWQ6n
zqS#llr-sK-b+W7WoBd2<h{KqW!l#uF*bBy**s@+;FPq2O)HjYFy~zAD#3Crbuyx#8
z+QSJORDF>&s>{^ZRyufG`$(sdI={_!iehfr??L}j?zar9@H@-&pZU&G(<qt~l8<2o
zB@43bbgLVWpUP)TUV#C#lZis19hfswHeWTleP8oVnb<$sK<o7ZAmW{^2r*x+JPMc?
znn@64VS0`BD1%h_x~@*c3>ISV0+$@WWe_p<Hlwh-cU%U&*PNWtXl{ee{|)#4x+?n{
zs3kPdyB?OW<FcD)-uup_vrQH9o;aM0a$KVbj91V`oR~)NDVGtYy6l}IW>WAqc`tpq
zb$qS4HIJIY;j=@x*DiTg3%GD=38jr=nP+h+T4(`-lEmwq-4<`XCa?8pIeP7_x_5N(
z#ao9p{$OOu=NjTli*S=kf9;`on&{=lv9Nla$)?~%u2JC(!j#edwIcqH;fVdfe9%pc
zeVO}8d12ZLh1w44GXsc10~=Kr{uY?GF!e0m-@yY9zIyb(^b~no%;ER?;RvAG<a*sC
zq=P=0`chZ{C;QnqW)!-xWGMqgw>HrLYUPV0FtO^MqA0QIxM~$3lv4wQat;_ESCt-g
zUE#!khq7Mx%Wnj)GXSxl`@%qK;&&mkUnlw1C%?wYzhv$hnhnD)TD%(hYTAJKJa?+b
z>x5G?YyFG07((gniv$0caMJZ<G?9!KzrFpm?(!c4J^#v}(SLHT|7*$5!$agZ-hscp
z{gm3_UlXnR?>zQ<Zot1sqwzl(0Y5d`{%Q!4Q50}r=NW$UVgFGfv;WC<|F=Zc&KF`M
za&pCix1VO3?EP6B!cTnfm(Bma#K%v~!k?Vczij^h=-Tkh=KoyHmw&}t@yq7_gsuKp
ztqZ?w{x6&V&pRi7il+X#jPx%*pXv;q{Ht2*Kj{_xRS)EspZ~}5WdBgfvHyO;$1gws
zcXcBFM>z1mVmp7M4A{TYvG`^4|F}c>$LeqVviaZdDSr9+|H|q3PoC?Spa09x|K;cZ
zx7S?yW%Ga8{C|F${gbx*AC3JmvynWSW0@)7>MppVE`jRZVno`%`4W+3?=2g;6IH$!
zXJdRThCvfGr5(Jy(7Nc9LtG(vgei+S;QABBS9oC^1+%X&<j$FAJds`)x^sG24Jdqk
zY4UJGCcJI~oq@HB^t-#T`GlHs`io?NYlkJKlWm*BJTYMu((KjhX<(N^{zy=*JGmCc
zPU`{0IVk$X<Z<0*eAGa#QpOrVDEpP!T-m@VZ>b+QuYAA4vzU*!bj7Nk#cP#jtzP*+
zD4EgEUa~E(A;YQm_|}Eij!X3lZTP5EWZ-V+TkJS57Qi2)i~y9DF^Y0GmaGn7PiAfV
zkSl$FN(U!8#9wT|N{DPSu@b_^UrmzDLlB#H8{mFU`JOo~xS}Vip>NwmH8<>%?4N|!
ze|+M<d|JzT7n9;b(rrmXG0+Z3M)?Z>4B~oBU$v$mqKG`BT3LgMD=_509Ru4wS(ain
zTq!*tA$X7zEL3?yX<1jn%hv*%kzZME$!J)8tI`~hRgsiMiSuuj_Br_ZAV?++$+(us
zuKE1j<YYj+Qx16`CEt9mlEG3Frr2J2{<^Vt;)m8a=Sa#;(J$a1ZVoF3s0Z_^zQFk+
z`uur3d(7T<7by#SI``u(!b~fY2=gORk*8(a!6_mq*xIEI=v@{#P}8GDehl{^tr6g;
zCfx{sS)yE|n4%@ZejLo~>AyK9-kcV@ZWF0{Jyjt)4Rs+^xcR^hV^=E8sBefAEm!W-
z>Dazv;9K1ZAk+_O`}K!XaVt5!v{U5Baq!_;f5Cnk-H5Is`%xYz_TuyJ-@4p6)qFb9
zn5F&H$2|v!H9>T#tz;xArw*X%RbR&)2qDp1Sb8i*<_e|c+kFEKCQCu&Gzk&&pJkK4
zn_rIgWWdHGc|5Je<p^@k3gJ+_UiHvgkvV07eqUw8vP6-iv}x#Ek$>Cg4iSMH!L6m^
z52Z@?Ky5q|HbaS$jR+Q~3K4w+Lyjf!XcKE_FS9+<G>gk?Fx`{q%M;;ex^7#0p&YhF
zTWKdTM-yk?pe-fu)AmbtV0vuIWu8JyH(}av51KNx-3rU%Elypa9RpOLcDOd6yBqcZ
zD*{l`33mSmqJ{dI(5U7}QUqm}nyP%3+}@ZWMOG`OsFbBOaX4GVvd<=uiuY~6+8gVn
z#C9O2+wu^gxP@z2^rfMSx&sw}tfapLDWn!fRB4jqKdNY(hbNa1WM7U`4e<IT-pS2)
z)8+V6EcM+FuD$dS4S0Xyqaq`$ZM&C!1>iqK`7gA>YS1sAX1CMJmhEmP41DhL>&LAi
zzI+2QjhK<!>Q;ba&&Od&`kyveTwb@|apG<ZI}#GlLU(Q~=EX)K^vB83&L0q6ev524
z#;eerA=t%ZK#oNZkz+NxDEeM7=Q~l3AwsvFDc;$X>sT$Jv?$a%tSAdnH2O<<Vq?B|
zh7shV6vat$SNiUjl6}Uj#BG-LPo*xgNAu>|K-T1xVISY|Wk>X%Pfe6Dti+-k4lK>A
z<Vy+<S)MRY3tB-?QZ<2^QXc-WT9p4#>zRRYwDffn$7AEwe%1`ld)@ltW{?+JNu@_W
zj=qU8_1OE)4Z&%*!{4aH@L`}<zV9|krvk1tin_CT3DOyHECuq3y@`Xa<h*bhYyuJf
z4aDJpah7(D{1qooD;O2+sXbw5LYgIU05X~t<`;aVWe3l`6G?%bL^AuDoXaAo)ybDH
zT_rZpeBfu@=w!jGk1^f#p*pM?R+0AYxV4*AiRClCzIo%ISU)KyH}m$#&*>{YseshQ
zN@;l;ab#=v)m8)yEsma9TtO=6A|$cv$R9?JXbQ>XaVd7+f)!7<4C=&uQe{r~3R7sJ
zP10K}@lSOOZx1JZGBUa<wiKct6&D8;^lt$lFDg=&@kBOsPFNZeuBV2DPF7U*yXta`
zEN~ket(HJdM^&cn2gIZHY3>cIb3ZbgG9eVdMmwxNrs2j(Y3Jbw;fLpL=(xCU@d)9V
z_f)^2@2TOr!GM*c-2m#*x5Fz(NA}Vd@uov7j_3S_2iD4{I&XG2Gs$;ek65&Tn=FkU
z6r!;uw*4j=lG!S5&%Jd0%hU!TN>PDT9g@xRvL&SBCPZ;m%!Ve#Fl(?O*UB{Vl6vAk
z>^%OH>=2J~L}SdDNrQKCq@n<EWegoYL3WwAEDpcjo1;ksQM9`rY;Y}0WK4kX#S9|5
zjVwdg8`|EtpWtmmP_C0qVoSl1D1HonuU(J5^5xj^vLI=0%Ytk64;bXV7P!Y8?J_S~
zGxQx`T`$<~)<+^~d(PSSxYgh*H_D=C$F_$#d?YDHE(}c4sA;{_--x5%P1_%BYAqjm
z4@nZzTkS0P#n)d3&`47s%Ui}2@?R&}<Pkcq4~raU8`k!^TWmJM-bPKy(rb(zNw40<
z2YdtBF<{8qL{2iK**o(a2(f^=+vhq0B`9@<%ZNsrU1Z)2ecmOGdHN_W;P~F0M>QI_
z8+~lb=#7Nf5w``$<jM=qrM&9z29_g|@#yRw+$SnP3i<w8XB34A;&LtAVc}w~y1Wi1
z$(8%p>~ZPEW#utiX7!=Cw#^S?`FlC<n>rste`el8vFM#9gEJjy{A3KF^Qf<9T1)8X
zf%<N8#xhygLBh<O%XKsE31oZE^N+1{CoauF%_s`iz#>_vrYwS+dt6q(TZg+pnm=W2
zJ7jz<jsv)*HeudNo%DrgB^=c$t<-rDq+3yqA4A@5`I&xMfC@bfR`OGF-2wkF*#E**
z4!?Q=lzol?_-J1`-vH#6d|lcFb&5baAgC)fkKLLFoPwzTC^3C-=^Kd3xPzwdlmeV6
zR%E>b5-72)oc|5vUkD~QIqLsbK>F{Wsl!gn(l6f(Q(2Sr4YZgX$?Bq0oWFElMd(PA
z{#JUG2?7vD`zSv^tB_n}YqZ0!a+Bn!!N<RWG{qOndytd701wV0dHXP^M3q(VXNy1m
z#rOVLTRhL<52f~Q?P+k}z@av$V$ySTVelD9r;}i^@siw5+2bpF?%lNQP@G(61AnCe
zJeaF+Q9y{y8PI-Hy@Fp=v`20psRn9Qm0^g<OH;eoJh5ZoB_CprQIDISn=h~R>lQz>
z-6n^WUp&Oxo!Bn;CGTsaXf9*J%7d9pS*dQPw5d4=Z5}GUZR@#0FQX;=twJ)&+b^~e
zyUq?2_%@OJ1`^cY1Ot*KUm!qH@Hz`DB@X($ABm~O5Om3B1!<=uo8(J;S$o9XGzwq7
zt&LkBT7Uc2%Td_WsO+Y1vlu_UfDQ8rzjI^wAqC?;iCRu8fT`a~oC&50>cC5x4Db+0
zdq`Y>-TuDv_zb<c17rG=49`y(#@2AQZ)1(Scb8rXX{#CN=FBD_-8Pg%cKIJ8XT^aM
z=|xbQM&_bFFjZcZB?3BQ70B(=wH(w}SePUaMTd%lHY?M0Nhperd5|oNQtS!O6jjd@
z<vDR%9vq>ypAdMUQ<reS@9X?~Ld|i2ZBNukFnfX<{gulwJ+2CcQgw0@I=uqD)%t(}
z0#u{0%q8cmqWaQaWiNZ{(W?v@gg(nTV5Q|iN<MZp)Yfb%pYak!P8RGW{BS|B#pGmK
zAkKt#cnZ+ni=Y5QI*TUb`LbGdR|QIjHOnuBCeS}}6*p0XD5=>Q_vK`qwh1wM6g|kZ
zKY3qJAZ%1|e~wV{xCe<BkP4@X+oSKi=*R^;uyEy59WM8s>%Yd=T6otq?A<<N+OT?p
zKk-rSbgA%VyG$R-^JE&XPQqU8!ay3L5_pD=>RtTSBuVQ2QZ$9F9T5Cu{r(Y~G}a!W
zW(uMR$#p0;rRka%WR{8XvO)p_^?;xA!Y$5#l=8OK^nr^ac@6KyaaEn6NTOaka|`VZ
zQ7GLd$4=HgLpBKlzbuVnQ^}ooXL)o}G}r9llc5yQ1gstNwlJ!%ZBBW9$NPpp#lkXt
z+IV{IHFk48c`Pe7r?!Ov-C?kYl0|=GcAow20i!$6gUp++$RT6UFsW9ONuwODJh0jh
zmyQingycfo6>&B>p!>UmYv8l*mPWWn9^IPAf3S&qgfGnj7Mxv4%IRK=XKD29c^tg4
zd{lxQFyZMMiIi~^JajDnda76UN#<?aQpWpL^f{6;>sv_Hks6~~591Lw(dL{IpI!X6
zC@|t<+d4~>tG~OfzZiGA-Pp00$+hQ;$+jH9v=c)M@>@}ubggSJO^U`_98TAEio>-6
z;<c~q;tkGAs+UgCrIrAK|C`(WzWPpMLp8ARU~lFyMNLAAB428h@@YuLE9}jw>e!Ok
zvmSfpl8wjGpEVowk0OcS?K4|1HFJw>slaxE5>-3xr~AXyRR7UL_@)6CuI{Gdx!iX9
zH-#~wwZ|(5nnQzVEaB?mYz*5l?;ZW{V+btvxy8AW61Gh@1!W_NSV{sVm1pbT*cNnp
zU^*r2@BI%)eGNwN_d{LzyQeDZ906+hn_XCb_QkJL{D&F))fay<>A(8o{~vu(Pbw=X
zKo5|uwoG>&6TLR!s0vzVqMV}}v}WeH-{Sg`WyeE24|4L^0(28tuhm<waBlH)RLc=@
zc-<aMQeSqjj8=y+L`#WZrM6Ef?o`zs(Wj-M#zux>zHjNK*#^k`1R(0bqnW1vMPjld
z@DQ$_0J!FGNAcGd>=AZC*W-mjdARu_oF{Ha5Ifn;$f>&%hz3)NXuiF?M9wCdo&9~k
zbcNf&n@B`STX<&0qflXfSqdZH5zkDL%T<zk9daKM)&%G^C1(hc?#Ydw@8?@aFb}H0
zHw{?`p2l7emg>qeDnc`ND#|Y%F}r2;xE(#y*;Z>@C86I1>7yhdUzKklau=Qx(T(B^
z&A8BV+gHP#o;xx<W>(spmnX`P=ZTP7q+Y}rs~Hps*nIu}+I#PyCfn{^lvfl~L<9r`
zq&Jl&9i@o~2na}*8WE(2SSSGkQ4x@?AfO;1y#}O4ASBX7dX0n<5_(Ul0Yco*_nm!a
zzW3W_mviQv-<;X=`-5S^K=N=u&%N$zt?OFX(&J@}IfUq70Vu%dp?RUIf$VKSn6@GE
z{<e33>UZ3Z7(Y?>w2{|9?&s-3q<;l1_ln;AO}LklclZp&<&sOe$Q#0PiBjW4s%-Pd
zb3n_zf}6WyWj8*~Xcu3uj<jF=qJDna_iD-q^%W7`o~pxA+!{~<`4CJkfT&<mzkXe>
z08XEDD&aujg&F8Ng~goA7zeW_C17o_)VqFd-bC-o7s#Cnbcr_@?$zjXa93GJ=CLaz
zL#Qotg6*a{7BCrK<-dlc$XzU%gY!-o#Zq}gO4jsgY_ap+(X+~$IwGPuUniH)55yHD
z_?^REYPn}%Ns9o25+H0B@K&xJ?QVwV>B3lum>8Z)sw6yeMtmIQ|6C?V<*nlJ>zo#1
zMl-!-Q<GFjg;Wab?3hII`AD4QQ}eP+bfA_Y41VlZ=1(dzjItV6RC9Fcko1P1q=kxo
z+5X|THo?Q0B!1y9SBDM3H-DL5kz_6~?wCBD!o*`{fm4*-0a4;-eDZ6HikbBa5p$zt
zv<C-<Q*V`e<C?oB6cIAx_LEKVaWqT0oIr3X)72Zvug<B?+WeZKPqrDXgPivACT6pH
zDfEz;>_{0pCEkH5H#IX(t=_$oCGjHZb_O(K;MoWps^O~5U!_#rj1J71UGN1;;g<Cn
zBJ46WQ{Bj2Y(5DU{MLFuwweUwC<15gK&hsgfyu9UzXZlU-+)t7dbSN|kVUXb%4+FL
zptAD0YbO?&(#$nUqKxRFz9-BK?`-VUwc2&_eKdUr&WFNw+22~s)ANrr%M9&&?RV53
z%&?4PHXm<EnjLt4H7T#4Hva5c%bP)WKkBJ<a(v=!12cNN%&}aUVz~f=p{j<c+959G
zt1~Ji%zbgVy7nIFJc^sXb%<LAgMh}BXLOo|Uj^^wgvjEU+WFN<@nza2(CS-Ca<j5r
zP8oo!oN0mfVjIV2<okPA1y<1zy!Y16spm-X3hRWj0aBt8Mu$hv|EY;UZ}RaZsOa;a
zt|fY0O0l6l2i1(MO+A9t*mbFw#y^+ogv;DC?VU86^yqG_T4sjjRc2bWvdm#vCaKzE
zyG`6E*{^G|WzZ-qRD*p;bfvFLgSfwo*cS!S8wupg_vyJWi{s`MLtow}>m`#$%tz}o
z4g4~BiJo<dHlVAHUX899yAv_7j=Izm86kzsaY2*s0#h&bmRn+_GnytUPp<3IWKAkd
z@0}#T0WXzp6MvfNp5}4|dzS#it}H)vu0>O}F?GAGU6ZXbHH4{$OV?8*RX*AN7dNk;
ziQM~}KZ=E-ZXMbaiaI{m$;Jy@WWPt2ZeK%qaRuP|QuF4z_Z)3JzBPW*56uq}#+rP1
z8et;nDwuO8pR;|20jmxsV4`Se#}!#vp~>xGH^^FT(k_-sd^YCUX0~;_&~@Ax@y*?z
z2^V_5>+-wlo%HQkhsXpDfk?^J-0|l9s?x#*6(h|2bloxP#5c5A7ZxQEuXro~WE!Jk
zm!N<L8NbFTp>`N=P8%GTshq1}xs900UC8^qQM7#lAu5;Te754-^-kK-z9TuZfY|r4
zV}r9P$6J#lQccb+T_C0sX3W>z&31?<-~3+xPI=){GF#K9UrBcl!&Q39^Cpjun>5k-
zx09Y(Y@LLj{Gqd0zJ2zmFxnNck&Tx$r__~pR6=*!FTqSqSyuzP31>##z2WP-@=SO3
zdF@m;(*L@UyVY}d_$I&Sbk+C7S8#T9Zz8bZw+Af8cOSuwMI5s^mNBF>FP?<2bdh(v
z#XpWe;Zt*+F|%e9PT}Rq-(h&s_-rw>k^-U$Ag3IfXCM;5V7Vhtd{5C3=q&9r#bw`<
z;{H6>^9b<^8rC1uY#stQwhKD5sToVE#}t>wCU;-`t9#1^m@g|rI(S7jo|C=2lH>qd
z)4bVibfWXH>4U4AWt`~EPOCetdJVW+T5_Uauw6E(xMuthjVq4Oj1Ifq8F`9vh1c!;
zz@vpJn*SI|Ig!2-?It49@8^7l|49`aL^S#La%0^VRp<}hWu>k26eMU&aaw*(XGCgq
zL#ZcIPf}XkUPErHLVEP)nrN<gmiF}vt2!KlHNM!il)BIqw8Ip}pIEn1X~biqjLd*+
z)?*G8G~36N?)!z9*7m2d=`ZSYPTG6Q2*|9)1n8AdYCETPUVi#1;&e>db+TK^Y_q6W
za3A*fde{)ebMesk4_$MXYou7~n=nsN-Ts>jKzOT6kwt;`jd?qINIRYBMek7bfcKA;
zPhCr;X-Nbs!}|UfI2WB{z;Ai;kazHB3vAqhuEU#=nD(wV)SAyu_A&)y8AWt;()fl9
zZ^cZiu(qea@D3o3VNGnd(>al>-TSGDVpEAXuS(zW&N6eHM%yF;hV}>EYgt2!JIGo5
ziu=TJXp8P;%UA!^ap14M|9%mNs0cY^hGx)O&;6mxPu-hb`41h<Ldc-$$IH@Mu-Qk-
z<KMMnK-Ej(G+8KaseMK;+rqmnk=YSyZS+Vi9^6;HZi{D-pYc#H^ZoJT&^QpV0Tk&3
zvg5~V11G%RPG;n`uJoly^$;+)1ey#YvApX1RMSP2Wm$f{Dr`R|85Ox|Aic>sBO34W
zxbLcST>SA6lH<#vpZOZRLy@-Uzat}KpxkJ`V||>N{c92Q7XCOpQj-~|Ir6B1HoVKU
z&aE^iq%mjClhtj=AVw}}zteY`SVjK&_LZLg%9S!2(@88<WcI|qWu%X&#X(6lB)KLT
z6Z#l!(UC@ca+9jAtgT7+AF^HQ-3IeX>KLDVl~L4S{1KpH%?>jmOGG1NWqdZNbl(h~
z1~E5vUO3*Ah!>tp)@$dw!5bW@C9@gRGq)dK?iqZc(d0w9S7T?D*p%7wrLNO9-!$?L
z1D`;YEu+;xpU(^1IbFRmbylgVdFuptFeTuoBBIGg15^a1U<YaO%RiAKfq7FWeTJ&w
zJ!g8=W9A%ff;Ju7njU)c^;_gz7**!jEv;vI$4CCsxt}eIO&S*arbF57=xbvzjX4kx
zloQ{<<q0oe3A-}tT#syOI1TxvaChYHHXVH}o1XnzpD!qEi09@oRYN8=$9v}bSX~#&
zD3r-C@(d~tr<E4C_<1o#blbHZ{P~vK!;BrL5c1^yn=wBP27oxabPVE<Y4&&d+M9}S
zc467~Q#pE@O7ln{brP56J@$<nIO0SL@k(#Y%5;On3}?KjjgU?A9+lz?ClHES_u=X>
zFlK>O%LRUh`d~~w1YWY5CHMx!?J*wA|5dsoSNNKAzD<0t{Jp+Kw!D|OU^_q5u@z1f
zz5e4C1&Wh_1P=@s;)xKhKFn3%O1-bzxz7K+GsGqxDlunO9wsp2Ba+w=c9YCEo>%g_
zAZ6Fi$1XPGdasR4r+DOVeO{l3Ki&iu&4`X6?DszoZuo7n80CA|Ek*Ac`+!2Tbw@bI
z)!LO`oUIs(XeSFn^B+gBPJG&a{Yt#>p+g04NtjC0i9M!RKjX7fBW9)?vp;rlMUxA7
zn&=4`iHs^ZrBLh1Hcj3}QaNXE+#pINz~&4JuP$Y1`|gQ+fgX$DFy-a(%?X%7<Dl>e
zO%S@d)7E(EXS#c2rP8C0w`XS~^tW>i-0v}-t?3qYxTwleM9w49%4P-LHo{|fSuQ%Q
zB5l%e4)i7s5rXO0HY;+X-+t(m5)+*iy2Avkmfz&nVeNs5oaD$V8M_&?v>#d5@*j5u
zk4G1+RZrUMyP#)?3%NDF?u*KNy8{ne`1s^4H`Bb;M0dBA+R@AwIV-U#U@R%gax@k2
zDPQQ}gC(J3jd%ps-ATv#Cz*RJ;Q(Iq^yX(rg+#|4#Z_s{%l8PqH{vb#DqUW>5%BzV
zX4|b_0D0$(KT9#h{}>=(E<=mjj=I8k8+JpOy>Ld?E1bq{?W}D%Lu`eoxF;`u$fJ9n
zzN^JV^&f&+6fc|U7Otk+xxaIIw5#E^BE;F4C#?P{))gVVnQ%)Kq{RLrr|a=!GP3It
zN@A`XCFDF?$pLM$K@JngC1|a0Mmsoax}_f13_QPtZ{$2=u&q{(Y<uFA3OTCwCdF?p
zpd-|pkV_Px5YKJDR9}=$;o4^>X}$6{J-up;`$HGwFNr_5W*6D)t(+I=yYcd6@Q3`F
zu-{xSXzQ#<&+c0i^^wWP9261#LD=>Dwo>Tp4U`ITLLG4GcDUnY)aB-K_0wDWiw@MU
z3@bY}Pm5vx^~naq{-zfwIvxC~BmR8K8jilZMjj%OlJ2KV+#A!+xc|5nBfl-fVCm~{
z$k?8UYqR$*>m(JX1*5rd>qq)eGQW_#pJo59a!)t%J~i!q_wmmcG3fixdW1yX<Q*&)
z&5SHJmNzkvWT}jI9fFe4_(gzgKw_w8H`V2iMVGVAq7qmxLh+q60aAW&jrTbMpMqrq
z-o(~6R?#_?^@Wg|Se!8UZE~k-rC$Ya3ovt++?h<g<xJYFDp<2S(`M(i%#%0~LF^SW
z_@E;?aBe24(+V|>Q4xFgd=Y<9%gcL!9d?V9P!Ku`4zVPg6y)kxc-c9<<|wh3w05&<
z(0n?`?+}+@t`ztwLQ7oxaz=Gcu%uS9o%f1Bl4Fc#{rn#~1l*biK5!qjv`L~mQYN(`
zG%tEOsgZm?%}9k?HVvnaIkUbNtDEkX+iI_Fdt>e(V;|)wr5q`eLD4_H>*ST*n=TN&
zi}27mh}lE1n_{2V*C%v^Hh1*sx_^9s-|_J)+D^~4j<fPyKTFL5Jy*PGoPX$)c1E6u
zMrbBdGl1IQ{^-G~WFJ+QjMXKZ<mKpcW|Z1`W(ftk4DMy*zqMP9d$;{WI;l3cQFZ<U
zY}EKy6yNJ1GNi)@bb<1ESIwB+nk-(%oVcGkR`LFuZ(Y6A&!k)Fk6Na>1+DJ(!s-v9
zCBb$C_Yk-vMnl6-Em*R1FA-PS(r8kW@7WD5^HNDaDTv>Sa;d4;IcetTIm3W`St;kL
zm}@HAbN33hNbynFNTcp1;w3x9s6#oJxuRJja*3?eS698V5{wj|u!-*N;BhP!eH3$Q
zR?g!4`2{d;8oXmH8M+~B^Ta7;tA{)mb`F!QkrcB%ev+g3lq~rVuRQN4ZR1BC%tXu$
zmi3`x?$aRmKXmp>OZ~#K3f<cW)TXthag682T{s8S#gw%4+P?`Qv(u@<&xszBJ(GU7
z@04S7{C5>^ajT8+O;?=DMF!A$am_shmd1EknQU5OdT$bN5eAa#*yC&${BorVsjCaj
z)xi@PaW&Yt<Y@sVV!@cjw&}iZMX?#b=`=Vc-*Ko*G<Gr9*HU$zg{6cxh^2`Gijr$o
zrCjeucmz-vVCB^&GX0d+>?p&<N_NX#+Hr!mK=b_FW29Tc%(WNtCyg+BC;se1%17d~
z1^D36shC3~5#PQB=A9c`ic$hB5b>lEMOB!}(_%Z_{dj)-Mj)HRbbe9jNo~QmD&5vg
z3_)x5T1;pSuNn_D!!S%cZ*@kH%<^twG(PhrB44>%DWU*(54lrbttVnDl+OL_)#mAS
zhmUvjwgC3Z4n#$7eVT6oNp4fE{?N&GnRKCsoQ!h9h8k_&Rg_y}+nj_FO~*OsXsjF7
zL^ansVE*ZhIl$myW+IYQTveFPH<k!~=mR*g|NHx?qI?qIZWy;~Al0MNCveA^*rt(}
z%Qqwe$bgg8c>@+x*UEv0cQJ+L>k=8^7j5$~Pe%$=V+i90BJTLQ{f<jX77O(vb7_>O
z{m3DdNSW98R?ZqOD=kOV+#9zLkJ54-O1;M>n<bhe{zT;IhqrP+%^q>mD_MKPS8-v%
z!OO;2)3wU^kzMbGRX>|AR5fxyNhRfgSdkqRAS31H=rlfQ5~mm59lbkg``eYyUfysc
zRh0UBnXd8+ngixfRyv5Jaj&scjyr>~@6>eyizY#yXmQfV;D9%+fo59C!HR4CG9{>X
zPNMPp<ul2Od0UE2+d%1L65g_I+5G#~E{L>*^RxLhMGYIm8r3w`4(-L?4Z&XmQQXG+
zPrrVex;^9@yS;Pc^h<X^YqzV_zUR+L?j-)e;4HfA3aeboq1$kdwy(00K6?{F%IVjT
zjSc8555b7jQX_{_D?E1^*tTTczy%(OsC5j27Zj}X(ZK>zbhu3m7VtH50tPt^@#shT
zfdiIoNVI0>(&gcmm6th+0oyaxGgC|<p6orJbNxM45=4J)Pl~8AGR`fvCvq*<!!P*w
zF-x@HRvL(@`@X(+PLW*D6XJZOK0P4(MA$mKPh8XXUP=9}XOnRuQAIOp0Kd#qjK~X3
zhQGj2RbEalSAL<gAM4!nKj-hg0kXo7qNqZ4;+3BKIi5bq11_7uz=tw~Q>8(}A73k9
zXN~MUHJ`rE!U(%pPj*9;F7J2-N4~CDdETbcpH><qZWRfpS?1&<-Jf{Waf37;&U=cH
zyNDAb&@#DuE(B3DjE^zhJMOQJ*@<VG^G}ObfBHOkg~m-5>r6XCJ+gpiO3i^)Hr|IX
zw%>o>n`7&E0m0VMY#1f_$_r|=|8@xFi11V%T25cg%wI#?@VI@lds8~~+fd%mR=%X#
zl23CVu>A=dt8Xiy#f?Z6W55(~<a<3hfQGr)oM@hlem;)zF$x)oygl%)mW6@x2DgX0
z9z+B)9>18usVZHWwcI(&^z!@J^hZ5qK2o3MxMs6z9-ZV)w$cP{_g9gQJ6S!gl`rAF
z?eYWaF>WIqxwZng8t~6p?WMrh&u-d?eSB>lh}^^O8EX#Q{o(*#g16!q6QAUw0_`Y&
z=n9~w?RFins>v$R5PZAx5p#WpJJ{>)R8Jv-D0;dU-gxbt`{(M4jijnLgyaCurLF3>
zE#-{|%E09Uk7jeR<8roJom70G0h^<)O0Q{FoZ;65Q)%R_Cwi@1MnJ5683sz<RN@j(
zT0{2lAFprB3{-9G+I*Ox4dWR1;qPSp!7T0kS2OO{)u8RAO~yr@hn|^==P3K}!SGQ2
zSLDsas<unqHcB<K0&&rEkSks^E)*&49ZIZXB|!NoZoH9b_kOT{JO8OD^84p2i;Vre
z34bjKDyIhkvob(wR)hVk*j*|Zf?l284Cdq5&U$@p2E|&F`(ybA-qdwVXRUQ88-HE7
zIiY^=rga;4O>E{s>|tl(4fKkOHEe_V)6Uxaut7WVDwli*O5pu*m^m&omh0UG9@l45
zsrBYdJxOMYRdH3ebh=XQoz&+>>m|Ez68s*%d*3ZVTDeIXy#yKXT`D$n;EF{9N}yKw
zc`!{#G!VBHwVCwG$SmLf5XmL!FJ;#15+9SCEPvI%6Dkq2xTWP8FR44^-;)Vx4U-tG
z0}JhvWnKp|m(BF>ZTvvir^?<S5dAz=uAeL}{Qgm_dS;?<#BK(!p}031QEKcn;GBy>
zTAzGX>|<8C1%atmw0Y;WoTdpiFnO)yd={~N5XkE;HX-64c&z&IjNtrYgWRzxXvImB
z9QL||B4CG37->(P;Ri5nwQhz%h-jk+>M(Ztl|1Wj1)qmW_jpem)EaK~9F58Ll?5)V
zdBU^zR<KA5d~v1h7+T$a`S!XpE-hT$#wDkxR5uTx*`)=IWb2HsMJjFg;yteoSAYF@
z$3-oi@miquWq7SycHxwE;iF|Bb16iG7JGlN$KfU)MeFZ};@WQe%D3@!7Tyt&36w9s
z@Zc@O#}S7fFq}z+qIjajq=Bix%nyKbQD_z%)4g#izu4@HD?9@+_78`ikpBvQkZ>kx
z{FlSw3q$c#xl~EPvjv|%;S;Xo61)1uJr$ua+1xd!+u0VwxmXUr%SZwDZ)w7k(agPr
zakrT3F4so6IIAj}P8p-K3joIbWT1wtau0b}*&L`}Wah;56)?M<X*d?3Kss+fc;W4&
zJgVr;Tas@ingp6rgnN9J#y&wsvl|Gv%Qd{e#;B^E1alePKrxdwciKRgGdc!xq2|Zt
z_HZ`O=Z~P_2IvG8kIx&r<>w!(YOM*Kmj5g5!rjr6?q4iWJV{1w@wri%(bmYGW>GFG
zWdAj}vxdO(${jSKWZx%Y5w!a$`{77=(&nOK);{if+{0S)vIc{zHx7MX=w1F|DZog*
zgT$Sz8O#ktI|)VdDu38OdCwB^?DSM*w>Vu2?0lb(Kt4gG0=~F1+jeWFfnGFx0V#?o
zjqb-t{s?1)Jpv*E%J~^uw9(TK!sQuH1fA2{C%1#whh2Cl>j$|rekJ|#L7TY0cRx?J
zogD#C8g<5jt)Pqkj*@e)0m{D5Y0bpg<_C0GoV82w9rA65r&rp|vm)-vAq`9mgq?IY
zaPdZnYz8=tS+R51JQVW_$VdQn<>;|_%&ev`^>H`>5KPo+oRYn;HgkQvF2lR2$w=d(
z+6zPbEBBgaf^zQ>{Y75VJ<|*1uwH^m0vp@Dy-FuUGXvql!#5M5^={oC@#JhC5>@?5
z&zFzmsi$SvlK8z3g|ja*x|CEJF?;{Kx4jiM1&vv4x<S4X$9Fxjg$Kv*dsy|0g2QfD
zBIHuXsaS-jG#cC*#MV<QPC3vQ?`GKhWh4BDu2zGZeOX;qNrT;yq#thI7Tk{vIHGc9
z_~s&&!F;Wso~GV!6|GL@J%sGUcfY@RG;v7d!qAeirJ<kY4iLUgK^w>!^xm|<9Wy5(
zGSl3AJs?Ygd5a}{E`#cHaMIX-IbwNm?>3)KC?QJo#pxugLq0<9iwMY!jAh*1?^?BJ
zSw%<9pbS{`Was8@dMQgb=t1d1FusV&K~WspVV24cp>e{GUsMIeTr2s%w05ZAor6%5
z`Mha>Bj_?^t>N#lm+2ubTH9xUnj07X$VS?fkj?4obhD<haaab8J%iZj%vLz`&2WEX
z9pl<YXM)w^zIxS_?N;SkN=BuvXJgS@DXKlu9>YjgZEV8wnvWssg8bg9l=x|OpGH=x
zL(a?B_0WkrZt#>0OlwN#QkKj}m^Uy*TtJa8hQ*g^K}4d))pdzY0~g!3<8^7BGD=sz
zCRs`$<jx=8i_3fIO4sf`27u9mN{ATNWI-u644$9896F&~<qRpcTdAIz>?oer7aiY+
z+=wo`|0|mN(wFKIzixW^;H6@KK|&XlL*I9kMmmzbB}*vNC`sZnN=2-qGI$($z$7YN
z)LJI5z29)bb|D8SZhsU#Y%OEfio&qNz@;lQQfV!?#l&*FE#=1k!tS&}M|xC@OlNs%
z<V3!Gtf-LzGN=?BcxpaQ|6Mozkml+ix+s+ufeyc>y9V16RJ*Y2p}3Wqfr_XJ0FpW9
z;6lsHn~@_Mu%CpTih@VUu!Itx%_j_t0ao-SQgrh}g{$vZ2YaN>J&V0Me5Z90cmGQD
zEj=MuE2;eOx=yP|ZiCZ26l9kI>^52AN2s+UJrNR(vMIF1LwyfyzN5%u^Mj8t`*Fga
zPRr=WX+1K-_J=vn+jGw^-($GFpQW0x&CI6`o(na-NAXsPR=48lN1+9?S{@9po>4sZ
zR63o)e+^Zs{*~>T3!ALSd9QJ>2C(KALH(Q4)U3DS1;&<+c2o;^VKZvH==WIcoGdpW
z2xp<H1Xg-u;m*Sw0q4XXWm!vY7QZWw>EG=CRa(j<_n1x;+pEb#o1vtX!)^}w53oZE
zC-qPxR12w>k5<!o!2Jd8XZA`xM|#Qf_w9fBJ)3<qS#7|k>WT6gp_-q(qg`Ef62^?m
zDDjq1<up2`3#j`q`b6+Jz0EDA9MLGZdD5|ibIN^#TcTL&G`xhs(6Z+hgzQOhkml0Z
zZQmTuvLf6}YmKN1tJF{AO?9@2$7fltzINozvU(**x})TVz7yM#{F_qVA`rWmM}EP=
zU6{!dzA#=P8ec01tBO*-u)xpm{5gJ34-{sQECLQ*T3=$GUQC+{_1dx=^9SFZ>$p9Z
zE|XVcs(48UQtXf~r^7CH|D!Co?&()wlQ`~{mFV_?F$6%CiIhzbvx!w*K4`}5R<=zC
z>^TMO-<Bi-mf`~cU_-P7DDnUsQtJV*A=3i?U_(+@{~I=BhxL4X%y3ZnEjV+S)RY&9
zNAs#oIPun|NnNqyq&Fu|cA|`u<HOg^{_<V${Gd}<<M&4A$@JIFBTl`4L|?iF1U~|%
z^1a68Y#_;Kp#p6QZxU-!#zYOZ7TVP@r46UK^>phrZ|RNc&QAD^D}zn-V`maL41S#c
z^6bW5U3*|E*^~(qV|g+DjeNc<e`p!eh=tu_M(`vH1(zO3r0SN#46}cv_o8o*78a=3
zn~$asr9^GPU*sPt`q~QL+rAh5eIwb|1{cO;=Lj_-jyDnB?IGN)hFwV&iRD4T{dv-J
z`u<Yb55I;!UpX}Fxd3lU^zpJ;uv)RsYlM?D$4+dT@@OdrFFK5{J8wX0J)6|fO5-u1
zv7FRLYy{3*`L84TYTwJbzKFZ~v((mr(L1i}f_&0hC13Pi(wY~Zl_u<Qn*l+rWY;M$
zE`iPuT+FC&v{ml`mri#1L_W%NWn!y3%Mr=gb6|9RfD%Z|e5p+uXajO{M`&6nPv25B
z`l=TWUGWXCG`~G%FP5w$mRZ~MHO8OqXIcHttKW9$v%D2g`r(o^Hu$kOm02U><+H>9
z^peXuPLcpR1+yfxLwi7g<h}w&dA&%IsVcFXn1JrDb<lhZ=qH(6WO2ew-ZytXT@iYL
zXnId)?w3QZYp#ry)*?Iv$UP@=e?9FCwo+oG(*GD?E>hZA?}qP<wDF6aRr{_yE7SLT
zdr(EjCT1}-<G2BH$aqkx_LQSfj04!c+iJjwT~PQoW|G5vz%ue^^9ZN--BxP;QGq8S
zKl;i?8wQo9Ou~dnTP8YfbDit-ivkl-U5#BaXdzt)gviHkKFogCeXQIE=2A>aAgT?>
z3-R{)N)e5?>2}5t2PL<NTGL5_<iZjJG@Q7fL%@<{nw<+AVE3y>vNTjI(<dVGpdS7E
z%gZf!dha1>XN3}Gb(|x<Yd+-%Dkw>bYj9JIGYo|VGToR)bv)dZ!Z>=ARpg$#g0zdw
z2|N=KQEsf!F*X`|^<n4N8nv2B9CuD9nb2`$=TO@8$>Ig&zdh5+0{!y_`07?mZ?YPQ
z%JsjJqq`~B!~4`gKw06vE^MP%KZFSR-A3F`4uzL`RTj#|C^p$gR~ChI3)OvZH1Z1y
z4@$aJm7CO_M$g<Np&9tK%ezo#PKyU>(GlEEUYWT;%B;v1%fhJd+dMP|=eNjx3X1SO
z=#rFKiF^B)<HifaHx?O*Mtb=D9Rtu74N;MlD3b{h+le(IH)$^BqGh6p?(qwhj^_oA
z-nYe*i3Q}(^-;zsazM_0ckPFlbsUPS>oelf8+y9^>kQ(0AI`tWc#L+<yeV!31eWkh
z{DRv~T9e8}{33SwU;VUS+Nn?@Q9w@Vd>W8+8wGFd;tv@Cogr3?cK3TcFHOe@Ah^H1
zfv<=)2OL4D;RyIz(jPjZJDmUD+i9q}0rHbMAPxNHzeJ0#E61=xwK`;p4gUBGlm^9t
zTAaS;^MW;Hc`wde@tLU}E(QdT0<~Wa7p&eC+DN94LJs*AF~^Kf@a0Ln<z@QC*nu~+
zYb4Wl)5!X;vtur`>Emtffx(dtr5@FwM;RGCO}FOtV{YYchc_7x6@R=WLQ_}vQ#U*2
z_v9BGq^gR2A5t2)w4xFmzn<oFleBJE7cr0A3Bu~GQ-9XeNc7wWm%m;SYw<Z@(dA>?
zpn+*k7q^b6C;q`(C4_x=O(sCJ?PfFd%nn>yrty#kqMV!SkMD(XNBH%8@|_mV()u>p
z&G17sM*MEItIuI?#HY_x<Kx@Ylu}yA8qm(t7Z~!~`Rl}h|MQ}*MZ=?jTr5KJ?8GH*
zBOhGpDgQ(CMu;(op|O%o34ZBBqn>TZ7H(H`m=D_^VtyYyAWI`Yj+hcoKsHt5{_!T0
zG`33SLPXJ(I^JaU_uXD2LT)<Wq{@V^mZ-LIlpSrYw_@?<cR^<!O;Hqo+eYISdu&Kb
zh|V<8>MHncCRNw)pC8UWj^q38g6V~4HLu1b*7S@V(VG3ID%=+O53pqXjyk4Y;LOue
zG~)`0WgVrX7-;7Z44J4yn6N_!q`wUI@%FyCC&T@f_m#h@%Nx~F9@8I;)pexxz}vfl
zrUjI?hh*0RA4Ac_Whb|(>mS#d_6M3%2hs+8=djCVFJAbXmNikex<lI$B0mA#`wBZh
zvD4KG`h9SZVm!wzVfHP&RcoU0dfGah9YuMZ*|V+eMr`K28pRB&#J;P~njS;k4#PoY
zt#DLODQyTx%OQZ>@NLTMP*dZYG_%oKiBX}?5z~jR5pRcX^u8WKy9oZ(IWHJ<5OV{B
zN=p4lmf^i8U=TD(@}x!e&)hh9@3UtlaM3e3lokd5x_l(c!{dooCXui%@lRx9GyD}g
zuBu2~je4IKUaj@<OpB{lwefUT3ZYpZ-+2j<mGRYsM-2NppEGHG#iq>va@7Y_)}U;Z
zwRvUyFYUwoki2W9=Fee0A^4>VOf0RO8Hp*HhHUGL5VYvvt+K?*t7}@2V;idI(OZ=n
z{p6bU8OJZ2lPbPyJM+H%3@R>HT@T&g!rky3RE=T3VwBR{iDs8Y{ZKAxVQji9D$8S2
z>TwXy*tk|*MbD@9;Y7t65#L3jSB!&-mWt`fB#>x1hOVg&gHTfMYR{(M*rc6Z$-Y?2
zY;-KlC!{!3gm<z(sr0rry)eY>Vw2c=VNhn4``pwsU3X6{I{shSwyLSMp=r&uIcA-X
zq?D>_$r(988P5tWzR6B%?%QjEw3`w^xh3_JkP1xNq3m~=(V*M#AKu$OhjRqk^^jbc
zI5{%8__sDeV14>>&;B`S7_T=&EnzH0w?Ef5-MF-CF+tVwn@SC=r7lb<tfvvw0pV#Q
zexD`fH+Jp6Sg}atqF74@vxqs)>qqcEiZd^}|A$U_(fTxlPA%)ErFk&!<jnK|eup}6
zeoV{l*5;lxmsvI0b3_&;>+bGNFs)vRW=5yFbm1~L2vB0$#trp)t4CuZe?7c@n0Sk{
zG&>NBGAmjO;1?Rl8j(E7{RQPf1g(EZ!uLNJ%WpsmZ!Avk2fD_X2u=NhCNrQv<!0hl
zR0fv(whnS^Uu++RovR!r2cBdrF^q~1cg5rgu(*=x$dVpf6Yd+?UdInna!eRC$VA`I
z9pxb2L&?P_`N{jC4N(t<2-W1vtkWxK1vJ^q$iMLOlMlXG%iTUh(I3ftbH&MlGgeRX
zpL$PRD>3|q$C}b~LpzkGHxDp#07Tm@Iye$ggIzyK@RJ4FVu!cNH>iR*>%=7{u6h!<
z{C4l_u@AcA|6JU8-vrgn|6_6eSSBXhd8D^q`M8q%n!CbLEXARf`ood2^&2^yBa>as
z9YbHDwufPoqCgsU^GP6!UoOI+EGR7d6A}XiJwEPkTME;9Hb)(VC$tGZx$CVVvaV^O
z9K=4=vj~~)->RMS9g6?l-B0g75VeM&8j`TF{JX*)M*%rS8SEDcs*T=m)0n^><ofT<
zk<1glTWiZD>rEhXGjKQ6ni~A!xklUf$#4eR04^;GqAmhpi)v~6R?Q0^=7E_^{GT2-
zy!#}IjNPzdaI<YUd$8Ih99so$w^sRov@OiN(+}1#S;r^qAnep8h7UYNR{cZg9&g%m
zz@tOGO@2>GpQMCg+f>|+4UKgd>+9Wb8neplRZUf9sPwN%xkG#PulCHDPW+*>By$|Z
zBK+`Ng(-wY9%0#N#lu>R=IRqcBTxSES;JU*g9{$<jo0;TZd^{^55AEBo<5|>Y%PRk
zoEn<e@T0OBk@Wdy@lghVK9LPNX&FN@7I?Sy6=nZ4>Tw;bZHpJXS_1Wuw<e-<-6bwt
zo|L%0Mm$kzp|Oxa@6?0xR<kaWM#mM}AO#d(BgwVYniMglt*>bFXXUMUrj2ML7rF{>
zd~uF)dME<G4jTuI6AQ~w|FreBvI$DlgC>6Y2$WJGv~;H<pwV<+Pv!#FZ2DYo>QaB(
zwUByONuEm#<1_c2V+Gc|3#Os`rWaO^h^wjbgiwjLAL#T5SP#=@Uy<(~)6<T&^A88V
z36{Qpv^lrjcwJiL7YcLcDFtpL{mZyP1#`F)7M!gKh3^~J55r{3S6U(GJdo+BIR#D%
z6s?wT9^<8{HqWQt%9U42#Vwpq4f<>&SjzHArIfPsLx7awg=TI(hW`c{#DT7*`XEhr
zoYs)9G!?bU3!j?{*bG~g>(_VGWjjcdoP_sKKQK)j%OI(K?)@KVosEvz*Ym)9@d}v|
zj~T9Em^#x+u}$wcpu<`%GJbk&EK8S}>D2j>GYOk&px5|Q2sBF%!jTDn&fdu@%uq3m
zh4h7oxm47-D#a$%_DY&)TFS@AV5>u9a2<z3|Hi?IW2@6K3R5PE5lxgNCAMm@LKWJv
z(JpiMu|JSJ1&bAa^!Y-K@>2jeFCE}2xer&?e}AL3`qZ5r!#5W};c)k9qc&s~Th%Dc
zmJ>k`U{LEOOiL@ao$0;TB8CEaeOHk07mjhanz-y3>nq%x8xt$R@00;gn;!Y5VIR6h
z**e(Ugi7J7QGxt_kuZ^BWQ8lWAO^3nOWtp*+Dv4Hk#SRhE%HTZ4rFF42c8fTf<Vy%
zRS9|Iy93tY4rGZ4|66s&nrix(G4-nu-`WYzX=i*PDlOkm&+y!NaKOwk-=Q;d)$qn)
zwu%afB25FQ4M;0Z0`*UkpV_kvTLSP_qm$-DN~tu}U>{WVhi*VxbY;e5O92h*qs(n?
zXazTHxa?cd#=ngrFobK{XuKz}%n?6vIB`;Q+qj7zqG{6W5`Q?*J=}a016qKjBvF&j
zcDJRat>f^LfLhrQ9z=k1YBF^x0ira$13Sbn{&%o(N?k?tlVm4t&FmntQ5&<epQrQH
zV>++8zWaG5_Ag5<cZ#|V5frPbKrVVG(TJ+bzCYp@Yt)l?VTuwY1}$hz)#k_?scX%R
zp6mH}-_32ZbvBeyS4kq^B<!GxhLPl+c(Gk6+hB)3G6X6%qh!o!+Q2~=8|7V=a9aEO
znctIBYe>%e>xG{Rv`ePio(p%&ercg1qhASe-I!%f5&Q0Fv@bloGz=518Xy?EJQ{*&
zFh)r<k>TOQalzCpM)JlNWaY1jF^lXpkG`Ro1fLGo(A6#f7!n3Z!{)PRKnmCa8{i;Q
zH=0Xqn_SbHSZr$z<BZgAtU?cdt@q|r)|xKA9iQWg9yl-MRa)HtE2fyk^gQFaV%mTm
zCYJrtD9nPm2G(gm`q+58ZZRXD*GN^`@0bvDRc^9&TXX59pu4C2R@^nF+GBqXSU+_d
z0?*1`IPx(Zz9H!f+aZhWYYhSXB#Rwc{UPNZ8a1eyPZJ{fSJZ^C>+*Co_6L2DQp?jl
z<C^evO63f<;I2a0K$<l9A*E23)xNC&-m2|Q<QR_p1RG12mfPLA^63GWvz|WvuU{$7
zNms!8b~ju$C)0|kEZa`6_L`5%VD*|lWP$bI8By}^8um6e%2&OTwS{;^Npg!RHZxdO
zhAb2IQ;DnL@G-BQm}PW8j@i!R%}sRy5(|yhl*bcLt?X~%(B038+_Zsx4M0&<CoA00
zxpi~y*LMAqfaW^&dq?o}<)no~;Ym;jFwf_7%*QVo5a6uQ+Hvo_owlx9Fpy6%aLH~>
z@2twD1hZT;4Fu{gj2-3LQa4jw)2}$W-ZbRVUkWKX&0vi?yEQrMR6=Ohd4ML#Mz6yZ
zp7Te83CDX4gSlC}u^!wwZlOr7H>%<mQon`lq%_rY=_&z0bc{7c;PD|e$WfK-Odxy=
z0qG9y`b-I#oyJDe7%4A*C_71xN<ntjjP6VC9iP7RbKAaJzNRAy!TS+MW{=c+`tzp<
zOGh@{9^~E7+JWYEvc}t^j-^O{mJ5#@Ja4O~zQ^f6+5403&n#MhuDN`Vd6~|l(?q}4
z>|RFli;yPTjRDf&^W<1#KjiEthP?x(O^PB)#AymlQQRHK*vPA0GTG0Irafuw-^068
zhgW$c6hv0dgGy)ao0qzXJXdS7>I)$ZbjWf;L5`&TCNg82+;6XOd0Tay#M@1<%o8?e
z$CUZe0*8+>c?&<Kc(r6$D$R5bDzZVNW(RC$!;}09&yRuD0$nWXdIg264ad@U@=|-0
za~|)}D8;I=P!&x~AAj<iNishw$pmZZbM^57=p2>t58ZhTAv;oCWPfWzYu!<kR{#k3
zHknm-w<a3um-ytv>c-7<r_(nj#j1X;h0)P|Q$7<axk54Bw9b(_X07!wlP$FkW1q=`
zP2PZ$jKYh7dsmkF*Z<HxEn_65g~M)@u9+5lHgdhKV>vhVZUa_?O4yJdDzs=u<>pS!
z!lp2wwCoqTX-066_dil?cuv0iwc3&2WT4#Q-lx+60t>LhtXN`(P+Eb6GVLNyca!Z6
z-?vYKp7xFerCkt;T>?onjJ)ggipQ=mR8V>>BdGAX5D1&+drF8oX|QHw*Yulu>cN{c
zLHX|oWZX_&w7r`ndMP?QJ0Ld-F|tF2z)!B9%<E4fWsOllLL>^CAYguv_vnxgVij9d
z!u=-v6E99f17?ofG#TrLm+oaMb8ouZaxcB7%P&=w%X;w{XM?9|72#I`D#1ZPkigR+
zY$HGGJ?vCd70pG(^)7a_)z4mXP6=2rQZC?d=mTm<m3vRB4P@{LO!t|~X683@|L*nk
zPr~5uguwrVAb{BX!qi&Vox~D0E(THE-3iSaGJ8d7CZp<RAq40aenY}Vzr>zCI`6Q4
z^83@|-X}H(L06IOvay=N$nPW=u8v4~*~T@)REzrdotzdggHm{Q6q~m5HVJ3fI5kyz
zmgRkmes71%t;a<Z|I`oDfIz@hhd^|W@1B^SKn}7!hd^RpS&GMc4TKPuI89Xysh|d`
z+<9P%8umBpW8<^8aeMFh5hs{xYsvJaK1m!7d&#PPk=~SoErGhvHgEJd&pDc7&Bnr*
z?HoO7fcd2|vPw$!bBgtcYFPwrq`T|ksE^wZx=5_7{6^6<u!KLgnuGE!5^E=qF^4L$
z3w@P&+nT$%XkkhTS*H`{9*;717H`V;;7OYF1lp9}^^{RC4@VZV{c9h%h8rVmKKsl2
z+FALlWxUF;bj~Ow*X=ifiCgbz68sX_Sv1gO8n3S8>{8THHYJmJS><{`L9g5In+dw_
z9SlAQe_?O9TlK_KZ`+K9G};N?R|vN->T@`ni8)LrWXC8XfO={fXY<^^C2jI8b^Ve_
ziML#4nE(!~8T=_t{~PQvSMPCJJ{BX*bpJ<CQ9vgU|J?D>VTxo4Am^W0O&^bCiFOUi
zBldC8*rs*NFm$>?uE~bpau+kJA=k=+-WBL|c&fL2j>15c;A<n5knb-Jf+~k5H^E`b
zBbYSPn&nDB@Z-r6lLS*Eb9ESnVA!WU^tK3?48ro6(GT2pr@?oA?Z;fSbG7j8ci~OB
ztMHgkUxT>C>)-+Bt7XXU<l<=_8_&zv1{0>WI#hzgN_-M88wv7Q<k@>Em)kBR&2G3d
z3$lFD<5NA>OIj-;VpmQy_$(7^nvW!Dtng!h^n;_d|D{^6HtAp=*`D~$j`MXXrLukh
z$N`s@09)Sed$HKJU=OyQQFE2Qw@P?oEp+PJkHfH)r;6$d<cqTG9w3&@up7`V8&c<o
zpxIE(1Ec1e@uJ~`k}`$+GA~6hMiTw4>Lq;xBX=Ghf+y8z8>T?f;VnkM|75Ly?Amri
zkTPc57evtly5%@_0;AL+{@%yT6pQxY{>sWJhrS$7L{KSi=3S+IFEnK#b<xU&%{e;N
z$M-(9%bZ)A8zAOQ6J$`@WM1OM50DDmRT^*2NNrP(*a+&WKe5SUA!Uor&FHu|>8@$=
z6}dOd5m3RJzsh2*Bv=*{g0)SLmN6Dsf1tJJN3aP-Gua<u#5+07#1yaXw^ld;-ZVHb
zm`b?^&P#iFV?Bue)DIU4j`3xh)NtK~y~Bu<4C;7;jUX*c+@6MgcDk}?)lH}DlB)bG
zNz=#wG|NX61o0bcV?-_?1;$$A%`(g|No%uP@jlWOf9UFEvI6AAQ1@apkq-hpSGPN&
zbxwC$20GDGnrGpssXi1;J|(-8Ux3tK6I!vd1u68x)iu|(=hS9!@Ul*E9lm~OBlzL9
znJM1z7})z-fe>jB0cj&GCN>1mcET^!1gJ4v>_?C^iyOy3@KAmoI%h2zHZ~5Tgepuv
z^zkZMNCjH&7X*qh@Xm#u{i$y4pBhU5h5w<u1XBnh4hTTQ+q15X=Y}%2Z4rgoKjpc_
z&qs&#f!X5M-_uJSx;gV~^*pBF+Y+xr^>x0JGwUml4J+Gveo~)S)MoL5joL`>v`32s
zQ<`d%o}+-{g#7$<%orsxA38F(6^j;{p@=*fRLYbvsts>DQbv?`@BD5q%%d+@PqSX(
z`_dyLaV(tV_?PmhmA-4W4H651D`4=!c2O8c!b)VzB-?Em69(3jaci8^!K*uy+gj0#
zWWao*F=~Hfv07>l6XM9;7Mxn@@_KrRude`A`q?9&PL6RwX@sKRfzwZ*s<~Gzs|>B)
zw!)=G5^vh5J}EDF#9(cG#_-#j`+~3ZBo`rc(5?;)w<cgW&}uP@aM2~9u(VUdqNCFO
zCl^Bva=tjNUKgd)5A-q1O?j%<zn-!zH;_Z=Y1x5souq^b*?HP^%%|6(V-a2h<@&Kf
zed|3My~*OeLD1B@mh>rq=^1br$xe4P_*PPsG1{G)!=*+1sHVdXt-cLP)}$`XWv-9W
zoU{&S8y@MLQR<8)ZQ?uXBTd<6*pzhk*ZWOBAFxc!^?mQafTsOA24b8v8WOF@3!No%
z%=X%p*Pfo<vTY})x-8Ri8pi4U5*8du-;K5pz=|)VqQwEP5dB`yu9Uh-DsH5yF*pF*
zXe(_Af^WxS>ytoghbntf(J1{Qs8(BgVRPA3XtLq(#?$7e@4p8Yba;SPuCG<CC8~7E
z31^(@zb9<ZTql6ZhOdXd=~>2w57fb2jvqjiAABFCDL<kdcQmF&`S*lgyhq9k_e40@
zwuSm0d3cH@;3TZj;)Ejo9X4&APF3yh0<sR#CT<sWg_8V$h8sfNf^!sRp;^E-9k|F$
z(O=_LY$VUr{g*O@VS}zVIt|~?)#-7+s{jk6&B42R3iCpAE0Fk}H0nULBAnCHZXZ$M
znabr9x*z>4HH{qPi+)gf;pL6_SG@m~_PG6qp&5#71-p%yQX0s?_=l*vnq~}HERK_8
zaMMD)?Ox4?B*EV@ck}A$`o|=$Ff<pY)W&Wm9}hnU7zV+An3mr{{C61y2fcWJN2p|n
z|JK&pO`2&6RiSB*WA?h?@Bbr)e_mC9|0Dx6=I`4(f9UjZSbc}NfBV550Neck`|Izm
z(f>`WG;i3sKXmV<@cWm55#7TL-ofZ6Sk3Fu?6xeyalZ{Ui(`#1w<EA=0uw|ZH+Pqw
z>ly}ImKZ_+WrP;4btnmdYSudfmerf)EO+G`{?N7IPd}n6k(46{j`H@=yC?kpYLY=R
z6L!}39)4}GdT7G<n6VDF0pKcRXZ9{|%yd7}qQq$Z?d|liJ11I%A4jDzhc+}WNW9jv
z0)Usl!O4Kn&jz{ZgaN*7%|a7?rCFMb2K2!ZlXIp=7l0(UT!5-8Kx%>kP)IwQngYb#
zcVAgbiz`#$e|vj<stle7KIjg#iaq{AN4!SAT1r5Lk~-%)4@zEo&cQE3?`OExe2>-n
zj@+(zBBAQBDU@P%DAM;%^pooJD>(xG?;nx(_e=a@iPSMhBo<I*2?({8EZ*T-W`6uX
zS#FM$0yJ^P!L0b6LzU?4U){G5xBrL#UkOqou;Kc(GLh9%$psL_5v>Tg*$mJO0h_#b
zvwvGao&WdN;(z*CbpL@j=6~U0xb%Rh7aNUP7tB5~fH<OoY+Y~-bTp6oX_E-!z{<k(
z2`8pzqi)B#^EiNc^|9&~iq=0*x6J6Z8(|*9@$|Fh1>iwoT)V@^!7%{)L$`y290OK7
zbNT~+=rjreQTYH(@vp%lhl784{V%_EhmZL`{|s27|Gk!E)CNReQ#p>Oma#1Tq01V9
z<AHBHdegPEeU~mBNmlAIJu_w=It;jm@df>%%L9AoRbHh`6M^1x6&&+ovj4tOTEe0@
zTpD!afoR-Kk<!FpsJXufiof=(IJPVQEf)P-G@}c((uI}%+g|^Bu)HH+Wd^zdV5)zK
zjER2_77rQ6EP>4Ve;Bs@^J@O()cCiM{-0~}pKBxhpL^>6#CrdG#+$!`WGB9V3%Yp-
lsySKf-EKqehQ7tI?w3i}Puc+6(uFH8N?9JN4Kn_j_%D@$`o;hN

literal 0
HcmV?d00001

diff --git a/_static/bonsai-predictposeidentities.jpg b/_static/bonsai-predictposeidentities.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..8582fd70727fbfff46555a0c463506627fa05ab4
GIT binary patch
literal 85845
zcmeFZXIN9+w=Wt5ML?v3fPm5k=}iHNh;#t~r9+e=orts`fhb595D*ZMCLJkKBSj*;
ziAaY~LPGBeB?JiJ=H2_8{r>m4XWzZ=zW2lV@K3UyXRWL?%b1yC%-<N}H*-F9z5rm<
z)6&%fP*6|+AmlgTd<LKapuBM5uMhb}MgCm8eDNX`)kRur>Pwfe&|bMhM@vUXf0db$
z{wmW|Iyy#nMkW?kHa4~^3>??kS+6m(va$YE2?Zs&4An)Nix+8F>FMZM|6jk(e*%~;
zUwBWoOi95HxWGg~$wYD93E%?&C@zwF`&Wnm$BW_uxsR8qFVoP{ksqjM1YDq?q`W{y
z`B$&WPY09l1E`oTGT)R@zr<qvoSNU8RrYOS-erLY)opAh!&pH%`&XefwCo($IJty`
zMMQ6j$tx%--MOo*@laDsTSr&V^vP2*a|=r=2S=wD&MvNQKE8hb0g%AgVei5tBBP>X
zl9E5Bq^5mJ&-jvGP*_x4Qd(A1TUXxz|K8Zt-qG3B-P7CGKQcNtJ~25pjYMM>mzGyn
z*VZ?1d;156_#?vc$zO6&04V=etbdd2f02ucEY}4pDoQHqzvQC05I{aCnW!$_l)1#L
zZcP2$n}uKY?Pb;niFws+Gy-xaST_4t!?f&z@@OI4U!whsWdAk6LjPNm{hMHa%Y_77
zp`;-Hc$7>4AmDTtL*2<hP!DM%y~Xd=2j=X}49vj6ry_UfPRwtf1L!Hv0gw5JpU(l?
zz9D~r`{YY_!dWp$YKmy_qZMZgffA@Y8;*v%*g6MqH-XbwWmUfP>&-K3IfSsXAHRMj
zD`6W}I8*{e=K?n<&jHCfXO#&A7lxBJ=YXZvbAWr*jW2KR!AM+`q+H<f-E#nP?Hmw6
zzFERRGKArqjt4XeS%sTT%MB@DFeHtGc|%5ACyJ5#)(Gb-4VKW~w7A~i-Xl#L$|AEt
zD!hn)-(OhAeb^ZFIY0z>Y8!Zw_4464;2)dL|A$sLdqr~@K>x1kf9!HgXbj+=rT^sz
z`@fA^05+kUfuIM~a4kFsI0mN;p-xeCyEFBiO&*-<@jo_41mbnGQl$$@sCMj?SEo+Q
zfAi4*m-1Xe_{npC=DlEvqH{o1&e<5r^Zu1n`k?~Q(eyb0<MQ{ZTVyBj`V>?SYgmoR
zGT|;gm=mM)@ojM7)r?~L@l)^ut5d=~*28s+eWLxJ1p>keyVL<ij?aJeb8NCuucP^q
zVlV48l+k|oy5y{9T(rO&!v^}4b3kS6p|%m8J4XkxTYC=Z9RB-s0^EV%ogzNMmaE+u
z^{m|Xc=7#v=*`zL2{$4F-t}InrH^;GU3^!pUx)hbh@7X1(S14HdN%({;3=<Q{&PU<
z&^drb`ybA2Cem<W>f>tk+Bo^b<>H7LEl=-M$lLGnbM~^IzsOx^Q(s^idhOhfh&4m7
zu~=Ii@Hlk9o9_KmoHVoQh3{+VgW1{{pJhC~(Id<E<HM`Vn5vYyo-))Pw}BuEG)m&(
z`Nzio@!ZPyNmPH*1jHeb(*c=V-SHVmFcsH2C8<^dH#1=ZmSX^8P1gVJL1APu2CVUo
z-Q1NLJiX0qu)NCMP<m&`1JhFbk7_kcbT|j_$2Xn>?Bvb?YYkP=M0FgOfk>ALB&4Gq
zIT=;;r8fTL#i-pN^PQzYCUv>lhq}Td3zW#W2aN{_M0x77`>;Kqb3py@zf*0$1rig^
zHZ4wG)K8y6-dm-QrQ>C$P=v*7L1NPqIwj?8n?O9x^@+iI?_Qq+{^MGfSWPI&lkh*Z
zO?2Z>eb+)%M=@LK>eKHH*&7qxY}`BSJelpajBi#RQU-}0H!UzLFvap$o4k9U;^7$N
zt^J9&?ca6P_kXcM6r1M&>>s{EvLWpoB28-lL+>U(7~lU-Y}h|b|8vy-herH!w)~Au
z{;6aCf^hy5>%eCU(eC8Kek8meeR2w%hMt`RnotO@NvF-X+3D-^F+Bv`bAaHH-Z`KF
zaiXx6dsPEZMq-~B&RSsG%Ex6LMomU>Tyy)@j~Yx|*rf~m-XDAN&)TNmpQ8ix>ugXm
zSe=;vLxDY7<yn|h&kw|<oHZyJghi~J1Ky>op93zR9TVO`^(0L@LC=!=xxmOU&(jxk
zp;9#~a+ipyLj%bfg4kIUHmMhg=0gA(6QVWyA+cM`ff@#<k3A6{Gv74vHr>xqNz=!#
zd_eb&9|q3^hI~d46EMUI*`q2lsH7&I11_$@GQ5Z~&V14Gx-Hh|d#2`p^2b{A_S!QP
zx(=QV1hd>9(cqGCbh?-(rvCH!sDL3KU7veC!J5R3phvld&_o0qM?_38yU}~p7;<&|
zu2&b3El?SgoY)(X^o~~x`x5>|AAWd*-+GcG2OdPUl96Na07)1NjH)-fvSv1+tXRu-
zSZDQaQpFaXWI(@jttizI8Sn^q{G8=x{#2mN{y|a9#9+P!x5|r}+C70i&umJ%hnZSq
zG9PFer4>KOynot%tAY(twRzaQ(f7Lf8~~bJcN)IPYOM2Ill3vhf8n5hj!h=%v6GRG
zC+xs<as&xJIX(x>8Lf=(8?_r9i&~kV)t>`~jX?UYh)wdK`2jN@Q`w1Qmy6%4N|n1P
ztqE3--(IQt{($1rUB>(?d@mGmMd@gUD5TNP&77zjh^a{nA?M4c8D>(K{TJMv9iIxD
zmZ+u?>^|M8OrvttxyGwr^<A`d+3qN-YA*m(cHqf22q*W3?M3VB;NIh$9vI!h5cN9n
zavhL`;I`WhyV9akU{LI)Y+ISASQ&9_57)PjF8`v->qH@W|K>G;{&Mf52;iF#HaF$s
z*j6TBjH*n4U!kjFt8nwTFf+L8IiLm;bgC==y-4CivLJ0UTdSjt_r7-GVK|#<%S`ed
z3WvZ{6+7iI*_2A6VBnsrfYM_e8WzSYINEw)Q&}or^=2t#^vPr^_aH~)N{(id)}O-A
znE<@0^@sU~ar|v!`fo;cIj?AF(@qA6wD+~B@I{qpwsGWOE1%>1lYPEIf(%Bk_XHOe
zg;PBesE-4V@arRP_p!Rnv=npeUNoBi__Pr&B8evo4W0ui5w?etCM!+BfrvfFG_<G6
zYr%wA1u!)P-qFhT0|)BJp>b0*`v3{u=xIk@$?#1bwQ^0vGa{Sm%xEX@n<1sI>>X_r
zy9etoh$GHEk+?SJnQ_^r#qO^?$6FH@^eNGxs~|f8SdmrV%SmC2sn5z$7vBk0XJK{Z
zx`J<MKDswx{RKYTXuhy~4tT59KPnYyeKLOz$U<P*=H}^k347nk@Ys_W=WG^tL=3cn
zO)8SKYD@pSNb#?76bXxb$E<!sC%7Ck=C6k`eI)=bCrEN`0jaN7>e0=>E$%xZ<3%$w
zvW6JU+D8d*s(dj{Lk{Xy&JFQh*tK&&Z+<0g^8GpBMB?jmhBE48OJ!{mwr_Nxn!1Yi
z&nYIH{!IO_p5(^$BGb9dCiCl<5-mK*z*B5MfGI(yVE?-B5_7KSwUrQFf|vC@_jbM)
zSW;dQp=m^yJnXTtoc#LZ(s~t5%O55F1im+XYl){I-suXaP+xj>P@3eW`gQ2=9B?ma
znhv&kat?@8vnTN7uGhMP=p~_2=z%#Un@}0W#`3{peJY<Bg}N{Yn>%zbD+C@qWy<kF
zEzFpQuo40~C$hz`8yn9n>NDkD`lVM8#l5{IE#ul)Hl$wKH=4bQZ?&rKuj+quJ(x0g
z%}njuut!~%EJLj;-uO17#*Z9((Vv|5x{G7uCyl3r$0%|y-Y_3koxzNT{v3;)q91V?
zb?nj@aD*0;rZd}`>dW`k^IPB#Nct`hADEQb+JvV(X_Yxr=LGB!HaqM@GCP|_^x~83
zG+df^X4TlfkL^^vdCMB*m4vyo&GX`F&mfXU1<v&I#6AX!M4ikbCU?Y4R){ceSmFY&
zG<LR{XkR-nabS7)!_448@5<4G9BHkxn7tV{Z^d4ITFwlb0Iz3@Ue}#bYhPkd%LEU<
z8*8%NU^UhR{GW6L^N`J>?#{oONB3cEvPl$a`d5=!r3nalZR4siZ2oxj+h^(C4RPx&
z#dE-ZJMeVD%rMp4`{s4NTZm6p@vs*tt{Jxm&?PU`yEuVPUPWAYK_f0ZJs86sr|ao4
z`2M#YW_-5CwZX69NwS;D&gqDk;E=4I!2x^`wm12WG>}G~R}9mB&NDgB;pc#CrgH%5
zLc+J1bHLE$bHMIf5ZyGPqjR2xpoCR1ZJk_h!X8D{HP-VnZe~m4R`T-igMYDOrumF@
zR$LgXKAQ>cjcC~<*$~e>;c>f_L>_Eos|aLaEgLY#VIDNZ?5cdYlSQ>BxF+8#G&U_E
zO#4h01NvYiulEYwz3s%U#QPgHIaeI4@~(oz`3z<o#P<jS>(1uJJ?%*;I8OWC_@g#3
z+3=Ib^@-#_C8m}L>6rFMZ?46iS`;l5#08seYX9*dO_mxs6<ZxAcC5}RKg@QFukOyw
zr1Vh{dpnj2c;${h2LK50&euwj-q>@%lks}af;vJ%aAAk|fH=ygtMq}VV3HDN#(kUZ
z`GTy~(;lT3Kc$WYW+LTyA^G_(2381MzZQWV->LQv#JErAeuHg`J{<Wm4V@WSPIHy$
z%zM7K_(7A2qF*1qzl`{80Mxf6@%q2Aseyk^BVE`g%!R1wp3(c9>O<vfTgPsE&om32
zF=|aSXdW-XRu(lmwczyZ6YU5y!C6loFY2c&hAK~%(}}+TMDYupwSh#SQUqtrP%Ixk
zF%Z2*bpJ9%RGuK3V#M`h-GgZ@4|OF#Cf_X|1`I;%2SQ)2%zfi-JT8NGCvYf6VqnpX
zc90GW>~4JBC(lWAPo#&RWkvp2&$NK!UU}lgz573gQ*W_9Ot#NG4O@|`(cN_<0~o?)
zJs#otBn=_j<A~fFo_0TO4tX8yhv$YJqxsITqNnr-A~u0UVB3@GOyHU&1mKfKFHXk3
z7z2Cx#YrMj?I%VtI|ti(eLVb@O!BE(4PzD8Y6pbcXnTw9^GUyV=)eJc<e*O9N4)AQ
z7<6#<gs~a*B0EBd@u?&7eiJe#gVo=NPE#niw-NQ`m!m6VDV=wMTnq`kA9D_vNt@jc
zmMq@U?86%kJwdG#s#Jml&{;5bTuRpXd^~R68aY3^J_`KM*MONfi*W1^6>-m()TZX7
zig&^OM&uIue{bw+Lv|&$!Ct=WIlYcE-e7SJ+2<cw>W9+=!rq5!E9}{&S53RRJ2_*>
z@f16V61rkCmgtB~t#B_l$j=7}m!TX(-KVmZ!a1#$-!O~%$30oy$=24ytrVg`l!WJX
z+a_}%JoAvRT~&eQgygQY*6YzLT(M{GZQNwk1xswL{Ao`gzR&kzzED?T0xmOJEjnvS
zvsFpEQ_`>Nn+*DEmF}(T(b&!TTOBRAQW3f90X~-fxIq|gn_w#YONGG^ZTTY6Wfyxh
zZ_bw;zr^dPUg=!!gf^pr1RatH2+BW9)L!IyV&SH&6zx@z*50&zRyN<kHkUCiXLIa7
z`p6+(J-VSPJ#^YJ5q@l!iP?m=U{_<oP~*}W$fzCG{Oy$fS-m$4$ZL|yJp36m)${5P
zYU4lmr*Nx^o9_dQuO0cH&Zd!&yEDIZMGOW5LuB<2)$lcw2RtSx;?~p&kDcA4eoZ`<
z$K(1aqgG;c6CZi^R{9@%wMCl)c;ez<e+vh?4B@bxfU=ZBGPMc8U){JOqqc`S^{ck-
z9tOo?z>7Z`6xAGMdkL@N%X1TQe!p1i{q-97?i|3Rs(}F$bTFf;-ucs0Jqw$S3fyrb
zU0gY_yYkx_ON?{6tP{qumJYvWDm1>{UFX~B<LsG_N!sK-2dsn0n7zqvTy(F0^3)LM
zn{);~ItFbG{oN_;DbI7j>O0uc&#FwA!w`pX{a47OUYI4AS=~O>5LuMDsGq>Pwk=+s
zmr9qJvY+?q9lcQNF%zYhI(K|X`q$5aJnxemssF8#{C_j!PY#}33H!Gz<Nv8sVy{In
zF#PYDFI~&MtkwZO2dD~=BYYS6{x;XeU*6pKpPI-T9DyQ1w9f%h+Uyv>|Gs`oZYy{U
zU+t&GCqlg{_xajg<t-+3cLJ^uONq$B$<}T7DH)5a%MUV~1rME)zk8#gHpF)XGu576
zB<*gv>4{zd*7lsC#)k2bb3pS+M$$J0sVH8K;a&Qyb3n9wEO%>;b7O1FUFF0heabN=
z!sAP*e<|s5stTJsGZ6amOX)+wAHC%EMR7q7PNP3Dsc!`S^%kFV!YQJ(R5bUpV_l5S
zG^cF2w3d-x)mg0TS<g8DQAJXC9RtY!`%)fekA!)X*C}#9E6(oA8_{#XKYYPA)sQm_
zJ0&BPMe~weTHxQ8+IWfLBR&!dK^Wf<y^!#4SHpi-{m+*E9<KkV(W49_MAKRmOgoLc
zFKdpjcIyK7dXE$qP}o|j7H!-NhM@yYP2|UV6^n%BRc9<do~TcLj(_WSIpu*iZ;smK
zrR(3;7@!PL0vQGI4@CpF_<*Enh5e%NQ`rq=a>B<Mm(xWOt)G-X2heV&Ha&W+xf6>C
zs!-bYWP_*O@vUu)>oo%~$U7v%OA9Jjy;5Nzsp5vZO0793CJ3|-2o2h%$3}Le=qrdu
zm>DS=qI@%HxM_H&ZkKUxy0bLUqwPiftsMJb`LDA2sjtXhe9f~OOQKtE>dq>#-3+0u
zANLc*q;_|^Hw$L33vot&d7t(Bdhkf@YJ+<pNg!mP+7-rxaCV19lCIV^V1c2BnKRp3
zqDxUiB@nGXos^^gmB&}njWrZ%^-F4Z&QeH1#Zj$FE+gx_BDjD^wVUo5CNTyK=6|+t
zFNc3(i`d@XUTF%i!0pvhbYIwhD!(Y{l~dHO*6?G-dL|XQlO~lzq-%U_*C$>f6|_FN
z$rn8@Nie|18YF>PZ3uD?THIZ|ByQISZr@&No}!t#mVB}1<sIr|8||MtvAkS_>A$eg
z208E>1rfhb4uj<DeIG;ANGj+Qj0@~7lQr4L6YsQ-IhGyqGALMd-L!M9BC-);v8#(5
zoggL)$R%*h2{;DI?VRU23Ui+pJc0hPj<Vy78|}REs=?7Y`l^u%#hpqybmUe7&ait3
zZI=Tp%%SfqBHo#VD#1!64w!kAPc|&yU)Q|lj2?$;f4HY!u*-^^BwZ*i9_;4YE`w*x
z8(VvZG&jN0<XczwPQGknoZBuRBrv0<H+iWis3SdvevEC`4SeNXfyb-v%;&XgB*IsI
z&fXt{484Q7B``o8apV|?ssa{{j*L)Po*>H33!Z|a6m$;u-Po6NgIrDKM>d&5Sr-Ge
zR<A9GP!ol4@hRBahgjiyEKuu@LNHx>8d8=pvT0JCs+IjPNm_!6U)TSUg-SZtbuX`L
zSm-G}@GO}acn+WwN4Lx?V~Y@ITZRr0LuLouya^UP(Qhf0dbs>Jq1Y~PkV#i=|Cv)D
z=Y>-3qe`WgpzUC>kTPX(+1jDhY4Z#SQ4Z9W@E>hOsY3+_^$WY)#hv0H4uVn#>B2aX
zy5ua<nuDe7-aEH})>|X%U!qu@R#zX2|B2Zp0xO7uIIp+Cm1_Ycx(Q!>#_nB{(UrPv
zB#KhxS0c@mpG!;e;@v=(s%d5#(iP+yjFt%WI#O&EufrJ@7ucb6_8?K+O9O0Ozjg)%
zTGi*|m_Bj}lwevkPgMxb81tA`MM#es;c){OFuFy>DIZDT_Y9f%z5gVeCdu^G5&OKa
zbklY_`-{HE*UFNko81l00Z&<X*Oq<L44c@%fuI-U(3iONGLP3m5T7AI+lruzyU=}d
z)P6*w!wF?25QAr3S*&ce!EL}6mp+=GBA;K13rk&x3Qe|%V{9GTO$k!no_Ut~f!^RR
zvsN}K)7*E=Wp<~VT{#7?3j6d&TSGV)h9fjY7V^?cKi3tB4|i}X*szl#^{d@z%I0K3
za>rIkGYK_nLfX>PBXHeHW~jF$3doTqP6WTwh_2#TN2aDN=p(Ko2dmue_?UD?I*o6C
zi8QP?y{$Lpl;rAw^xpfN5tD}Zc{w~iGcj^zGCDM}lU2p*wLwZtIQ{``n4j;#Zeotw
zcUhoJ*nmzL`!qxwy`2*+@ow09)7-4OJ6`t;Imq9AeZl#ST1q-re-Ykc6oa_pH^f7Z
z@Pw;!pwE=-nubkCCp*&>&KBtp9*H={-A{gbGY~J_VoKJYBb+vxw-Kw<HQ7o5<CKaZ
zasVU!80cr*{Pleb=jBIq@ln4<Z(rrKbP(e6zxPr+5nHvDj7u;g12w-%?BQ-?J(z)T
zWQMy!VjDiZC$lp#zirPr=qX5Y%G9ONYVW_KerL+tRze%ukGhJG3KrVjnZwskx_{Cl
z{pLXFAD_n6FYj8ZwKk|B8*%1s&A`j*n1MKlpWh%>7$aexSNfSkHA#jcnxC`6TTOr7
zT{tnP$BE;k2=v&=pZ;pnfkZis!PWhzvD)#A9aYbLsdhtk!7}D<x2_PEpRBSiv<i|n
z$OmH-`8qNgeb4ETCjeuXg^|#IgcGhU7`ok1&0IUz$5S9Z^YC|pS2UN&6!l#}V;m4~
zMMy8RA-ZI0V@tl*kGBNoL1MLQBTh!-g4{j6e|PE7D~tQ=ee)5f+Wgg|Cgg0g7<MLR
z+sd)(o-#6Odk(;#!q#PWaZ&<EwHqPBAevT|b>)aEjJ4DyJJ;w6>ot5>hjMt~S>Jxz
z%cPZC%xrKZLPa!CI!DnRLH7HiR4<llV+J<YNW2|+Yo=nYdBq#4M-Q&WTw)0PWbA28
z#n?RiEUc;36o=Rv#Tj)X`ML&9!IzBKCK_=Jx;VqE!oiZM*x1-R57M<Bt_URta$DWM
zdX?syFpxK-AIPQ(%Ol$SXc!h578!4W6+_@gBa#YqGK$5nR|M}hKHI;abm`gsPx4Or
zoEHO2i)QTh8c^b80f%t;Q>esiQ#*V%Ohq~M;72}UUt|+6`L-M?jB}3M_1k7XYRl*j
zcWB<W%FdA}z)zfNS^r9*c}lhr?`Z=<{ea;>=G>;tc8-v%xF?F`%oqb^ef_TRsNSm%
z5_KO`CVv^7-i2g#Tfw?*!JXL`aOP183@onfRsneP7^rc+rH#7;<=3wXTLVVIH2gE~
zl60>(T&OAAyGMz)B?RDD+H+V{L8XKQ&Bkxo!rz(B?hZkil7nbfZEyK!7iw+RbIZSp
z&T{;=KL-TxY-*O5pTcsEC9`zXnz_j_;<RH#^8|_<sBfo1SDkRwRS_@5lCfnziP4B2
z_UA_-Y=IR-H1IOkdBI1#k<E1u$O|H;41S6t2pK>$6?XEB4rxxJ&?G0-Cze?$BFO=N
zkqE?^-3S6pD@77cAgHS<oC6-<$)MwnF$|~Bs0n!fpSE@X8Vnl>aVLdui<4K<FMkmP
z$jc~82k1b`6-2nmLG}@?KR&+n&kp@_g#N?%`p-G^pK0{}c_VbaDog@XKkO87ZRDB*
z;>W{y1=_MN8;=2$xff!R+hKe3WM;{?xFb@^fF{6o)-8=oFM9i+xxxm%gXC6<c6_WA
z5OO%I(@R+ySm}<Qt-ucb%wTwTfS`tC2dik+#&^$K55(Oqa@3@yOZooTn({N?-q&*X
zgc_vH5Ii6?9KXu>Q-o~@)*gHgFa#cxV;0r<?#HAY>=r$YU{ud{!c4Ygr)OmI)%S$_
zP{}#qPhu50h`Y%^(ko9O7$Mflp0fZ9FaM7<<o>hlKga6-pShthIz1a)HbR5n-e&@N
z`MiIRc|&Gdq1DD<YdU0J)9xX8@f9*Y2hia2$t$iFnH~u%2Of*nm*if_`OoHWg_|g)
z2#IA|oAfRYMmh&r!1kW<VRPJo2ZHYBfTNR+b3oEps~A8Zh{<jfPIG;pXRI#j%f(-l
zBi6FjYpj=xe6n?yD7rk>V4FLzztYZEP#egiKj(nSSu*V)XRw6Q{<~2eimOHH6bMxs
zhk9a8hiY6E={mg^27PXwe&{TS``wc}R4NG2*>Oy?kIS0k7afqh4xsh{eMDeQq*&eg
zI73uixNfuW%|R8gaaBnAS%KjCiF`>Vujg&ZW>(5|Q9j6+q26bk9E~cPh-^8gEBEd#
zji@#+a%|h;67a<>-&6NNEd-whPMb`lt1hP3iZdj;G0!7DB!-3UEjixUw7bX92BQ&#
zq!}fx7#3N2InElDc2Jy{j{<9kel~*=P|O8?{OI*;<qYd;f=(mhGpjmfYnd!BZ(Q%a
zd*cnoVxeouL{9BE*{xWVknPbKmsF6DGO1T$j2$e(@K@BQ#JsuYd*hADd+%hlVa4l2
zJiKNP=+d@4ElkH)GkH<D!fwwTf%<Zw{hLJDh#UI(L)Q-!Zg+RQMdoGPiejeViaT8b
z-Z8wU;?T*%zikC!2qPP!nT%2Vr}`FbRg4EecCw3}mq}<RBa#vHP@2B0RrgLG_lIQo
zs}B~uK*i0RaiVb<*`b9+gxrF7MAegB>-iq<HIX-2H4h0J<>gPvDfMkDw71F1<IE=K
z^iCUv2#Y{_toQOqKUa@QTV;jT$lZFj6f}^<f-u^SREqL#RNa;{tBD{@80~h6QEL#s
z(vDYo5TT-9VgUbN4s@#!thW|pbV^O;R(>%E{GEsM>4DG7;<{7NU`iLR)=aeiCr{_b
zu@)nRl)-@?1xnfj;T}$g?9cLj^96nC%=4jSfSTysnS*B_=bs;D?78zz&yRMZSV*Go
zA-59jB04O_w(YtlfSpGZX7JSYVykwNXMKIY0BP?o<8f&!<@HK2kesLF&~6ia!zM6f
zc-{rZ4dp9^aJ09IjVUzq(N95=`o(819}ax4aP;#vM^b*9{VDx>_V9qej0hr}Ex<xU
zc*g4)m?6p9HLW$@l1M!Ko2H@jJNq;O{N!NA?U8F)=9EbVS0h^B<a|dvh{F}ZBrpNx
z=8L9|6vqBi<1yfCi`00BJT~cXvsJj0p#XPEF+UP{xz}ynJy!>A1rRcN+UMm)w(1(C
zoN~|-#qa9YE5QrdmS6$vg+H7zT{(x}La;qtit0bxE)=_#>T!X=#0*r@>6zrZrccQr
z3Qe>zJ*^~1sJKsFuZF296CA9ERfSJhqoD!?n<Sw(M#bx9Y61~kwx#wQTRQ_RMx)ej
z&lK<fmVM2mjmZf;2)5hRL*PQfoXRZ2l@yh@J~Itnei8koiG8+F?&l?KVV549B$T5Y
zMN4!%fpwzjtqAJ%i;Jf!7M@TI{J@h)ZO59J8PDE8cH3`#+(w*lg$8s&)|cjS#ce>w
z%~Tj$E~zWQg!cvx6e;!0bt&vj@%NbP7xZWLpL9z$(74c-{$bypwOyao48t<@vCRRS
zp7LYuDYNE^_;Y~N5Xi8&Y8}Si%Hy8wD&sn@hHcH0tP*c47OJfE6tL@kN3HXqai8;*
zEQg?T#&9hm4Nb4z&7sCphck=z!+d=rYn0bf(tb273P^=ef{i0glT;F{a3BsS7mofA
z)umDa)QH`-LcNFHEt`6j)SjSFkvIBi#OpTgmbA@tYpR89>X*MlxCsH6Grq-&@a|0X
zoT&k}&Z@hx6vgvm^$}v^&+l3mwT#6xHO}QEP031jAqLs^afK=J1!;;oq+g&F<e}NY
zgx;_xa1%t}%Sqf05hmm<pe}AOPC+8On%KjFWui1aOE(_R=-*7ta-gmJnYD8F!eXfw
zWnQr>f<=|D!phy;3Knyqa%#(jlrVGm)Z%1r8)F_&5ocloT#x9McJFPKh2ET0eU6ce
z4Ztf4D)m?sUaH<J{I&Oh%ev-h>(7N>w3I6|NrT~cuM>+|iB8zsu2dvET1pj&j#I|Z
z5pSW-)Fwt&R4nZSx1}nD(gh;4n2v|L`0FHu1(*Y2ij)rb0>P6cJ|bezslL8cQzU<z
z%rYYr?)6{<%HyC*DY&qoR=W{9P|YF&sv{#Ia=Xw%WW2)9J<!uWG5x7=p}APO<wt;_
z?lYAMrFW#ekf6@V`CC#YuBld%$)y(S_}2q&#`O&aUKJ8MH(gNS$AkS}-VC<9Q9rH*
zsa3VXqO!$WarnHiF7r29xkd@b&&b|mQ!t~QUaD1jvaNV!P6J~!%O&nuIm-9Rd-3_^
z1Od#p_~LF9No@VZqiJzw6JKLN2*K=7TpUkcpJ;$9nW3YYY`c3;t+Z_WxVsgj19<q;
zlzx!2__ZYP$l#{Evi<;W7)am^(f(81W5>-aLT=E55UIz@-@Ut)(9nf%@tn&M@h5s7
zH~FqK8U!sCM|wb?qUR{J$K(-AZd~z=XVJkI;|sY5ij)!qB##|zJ9#EvY1cRGaevNZ
z!$=bua0hD2-$$gpRG-6g2Nz8-QB@vQb$8^W7M1MR>1t=x@>YtdUMTe6dHm&}R{J5B
zR2$hxo~jD3Er@llCuEirVs>;^BE_BVjJ;E$UOA3t`Kh?q@v@iOkYBl5HpB|Hpd2E#
zg9y57=~Fm@KP`4XDmlYXR()R&QO3?UR@%8V1w{FNJ$o?_fUMiJo+|hr=1M!WXm<O%
zAZ213e;EXZNeEN5EODFQ5{a#ocyX#TgpX^;Db-MIun^?CYaE?$(E>WP7A(HX7mA>Q
z2BUKrA=LV~>)Db&>ah+sO{h84TH@NiocisXx$OAv6Ogois&yFeXsVQe*iPC{1U1pp
zQxX2f1|nwyqN;kgE3b0^V|BCgc+d6Vtw-JVgFoE$bz+rtwou~32ccKR`nFDhtwTFA
zo_0T!qs0+Vk2gsHQzyW-9Hy3AINslw-S!ZX;TB_eut>7}iDVudQd3y2Iz#apH`R4s
zcg4*`^Ns-?y<NlaxYY>HL%KUf*((#a_H$05cPx{feWg;KI6B#Dv6fO_Q}`M3o5X>O
zkHko^%|RkM%<D#kL;F!inO*LMx)nP1%)6GtqtUuI?)M7pTNY6DC5Orlx$;pSmhiPF
zc%x@L@d#Ew7|XnzDpjm5tWGsHc^~=<I^#i+21>i(dFNHks<Dv)kUec<uBBonZOU^8
zam;=Wkk8-PT?*<>SWLJI3FVQr74PJ-D`fvlm1Y$*J>MfX{~-Kf(!-|-pJHq&oO-#_
z?r8=_#v$BxFA;$gl4NFm)on6lyq;h<0<NeZ2!6mGv$=hgY<<JvSBWu|u-VJirTP2E
zcQFSHk#@>hzSue`Ms@V9`jQGiFnha)r_{DXlBge#mP)L3sAHme{|u8PYdQy{)8iF2
zcOmg2A-oWJZ!|Z#LrN{*0S&y2y!W1CqHwus^w3qZvSK*nsX@E!_ZLoDw=eB{|1|v_
z{-p_#zk--eIz}WG-?lHxY_^6i=z>-%&lL96;Ml6D<jxGgE~HVBjW*9*S5sdhD7f;r
zw)l<NQx1<DsUhC0-LuV^oOeU65t<i3-0sY-YCO2ANWU|aJDMF+>#xf><jS?Fbd0Zm
zGi?m<a7f&=5org#^O;&dQcjD?o$$>v<73*i$|}riZT7xGPd|7t`p`La9>#BPJchYX
z_<g5J@GKK5Q%UGSSN*goC$fmVJE)@TRtO|&N_6F<h16IK(OA2n6AL8-X2ttf1vQSG
zcR!o#&ROFSAaTzA`NJk9unK5@Wa#(SNvt*@4sFCpNH)Rlq+o)%O>vjf&jEbMr`dU|
zaI==hqiS(>LmfV89@lIx0-1Kn<<^v)h9<Ko*D`hr3mXPb?<pBaVBEF(Z1kc!^LhnV
zC_lKjw{jC@aN)5~QS9>3yZ9IcLlGI_b`A!5{Bh^55P1{peLYUmIWW}~y0+M+5x08j
zo2ocgyC=Y`RS4UanqEfG?x=x~^M3cgWp={3oc7lznYS(MM=Q5|d_6g?SZ`bSH1M-Z
z-%(a;Fj{p)^`Bb945W;8cecxpZ8tZaD)2<3VpWC3Dtf|jWn?5#qtLx*Jam>`5igz~
z_<2FCTSd2RbImOqfirCjVa2h<zpWr3)Q1mewh@jahxCS*B`cM9zAkreaqdB`Q46=Y
zF917ou1!JJzfBO08Q$L@_59(}ifhpb$_q}1_tCL7#2UcOA(0<w`&7F7FJAy=#q6B}
z>fjx)=n(k?j8`Ptl$xC~M0qq=4xl^#<kpT^Z$<Me=z@KG9k0HRdG?sf{>F7iy|E;d
z7;E6Z`!d<XA9{6*w0ATL9^I`cwrUR_!B~hq#e{4I`Z-9gMgxv@KR8>0-&2pRImtFw
zxa(cR-V6^n!1-!%9@_x%$ysj8E$r`yqt+I~hYW1>D>+`OihdSfT4iSEI>0N(Ah~U0
zva&UY!8tj5_p4Zl95^0g4)achuV^;ex55QP)T_|Fy-g?B(BNROpX2ROm9{jCwg`PW
zqdqH_j7f+f-AU*;=?SfXQVTbLReh+V$x186fp_)5_n5_OrJ8><z(;)niG$PX$~D{%
z^ll`7WZ9`&-vNzcU<8I)(5%VEiT^NSSSqLQOenB2IIZ*KMas(h0Dd4Sn%G$7ZwpVg
z=Zv9`{H4AE-EH(|xz+(*j2Z>HNYO$vG4Y8Ds4JV~q!e;v=<)ZK<`F-OO8R%HRbH)6
zqItLEf^!eWv{NB>7pWH!?Fc5Q4(@5xaL$(E9+_=!CASgr)Y#p}zaf@FJZp(hOe2Mw
zJBV4X4HAY%#H-zMd+qtzjxr7XqdwuSAL5E+;l0C;v226=R)YL&@Ko(R>W|A7;a*pb
zugcM^hxDzJ`KmK}2^FL>RtLz}T|_Esg|yqeW}t*|4m->OK0%BU&GS$e>pD(jk`?vW
z9AgFn_Es;<_@ysY4>;`P=q*E^li6R2Sn#`DsHRfvfv5JPYI0UH7Cu_9;u<pKE&uhJ
zF3krjv1`3QKRnRd4r^!;V+chsLN(Daj-o<`MH&SQ;>A+Pi*`Q74dn~gM)J5^3(rxb
zw{)6KLNh`g-A@=*5$Q1pIn1qB2w(KKU#xbJxH#Bj^B^Xox3`KL-p!Kv%_BC07pN0i
zWjpF_y($R+aEHZ7unASBf&AvkoUfySiu$PO?rpM%y#?DdIMF<w8QulDR$U;#x-vkK
zP<0HL!F3Xuu^~4zt5MM=RHor7{XW{!hf~*D<mkDDxTfrmnw?RRB)nMhn{NuDUO5-e
zq1E}Q74ojySoqh@lm%1swcal15ZRK1k_0k4Pjn}Vx4aeKmfAvokD5k%RajP3_~}or
zgz8df?F(pgUl5q@w|w<OXN|nGkB7tN2X?25?4GcxD!pFb?8>u7X|;9Ym0xB0ev}r8
zcIGfma7`^zbOi<3p6wc?HE&ilJXaAu4yq@topBkB8rTlKP*cWE;ebQ|=T7CvICdqN
zV{n61{i4^_u?l<<JUCjckR#GHM5%$H*H?#eLW2F~WT~54M}C)5BP*V3UfI)UKkLoY
z5&8t6xYX(&xd2PZ!;N<*26->iom~QR<*;f7IR)pn6=-udbLiM+PaiTUNDIhR8n=ft
z^DA=Xk&{Vei{c8rUS2#3RVj0AWlu04s~2A^pIuKoG2ssXlzuSDw~igUopj6ZWwmJF
zM@LTy<^CMS!Q!K71wup1_G-Srp{~30(bpfX%j+;9m<A!}^dUq8qZF}4G<K+5^UvmL
z9<2bISpTqNYks?ujwbs+{o1FUOcDFmzL}$d;=#s3q8Guh(x9NL7;@Tv@~)|+>eA=b
zrop<nAbv)ko3yLHoO*k060_Zx=P2<ZxaHNip1#4RZMzikKdK^M2El;=&irIIyk2nf
z(S=_>dzP4P%!Rf!Y|o|eCS!KP>~6L;6N>$c*Sv$jTH_KT6|8BJER0v`>y%wH47%Rh
zM$o+%fA*OtcZU->>@(_4x;|`vnhBFK?ibq@nb`xCZVi%cvoWC;qS*q<t%~XG{S#jd
zXa!2{9&*Fa=3kR>>QE$*m%PTf8_N(Jo{oS$^P0!Xy_o8C;#M<VWwoqC1?B~#!y@Nx
z^}e?I=g2LSNBlZbj*y8lV(BQ>M4wqr>NeoIvKwxWn5)D82-hc=#{0aQsVz|$j^g-0
zd|SZqK9o|GNQ0Yi{_2hd;cbvtSkxDTEEM`AE0Fwt`X$1(iR^c|9xLwBH?n=~o_$q-
zzBJ2P_@l5mL$~TjL&fr*c&aKfZDbuZFppTu`Q0;&l<G>Lf+QPFfa?-CAuzIg7{Ft)
zGVmnSWlJ&CddTS&|4%a!e`(f}rzkqYeutXUDU2x4%0J>q*1zMpAFFuV19RYukACq6
z@BNjk+p4lupJFGwVN7*A-W5a&k<h|wLSPxe1Z9wcA4Jsccro9keegVG+;dU7Vd7Kw
zbmoreoL&9`FcO&%p(f>~T;}R+;UWvs&>qQm;ZiWFvmS)l-HUmpC|JF)=*11d-X9sT
zg?$~^*L6KxuFSyX?2I*${nwvl{!L#tV(vBA0}+X+2ZceM3JJBH9Hg7r>@YMBw?(ox
ze7qQU`--z~fWcm9*@tLFuc*LyM~&v?b-R=j=tFE#@>v{ymt}1NHe(H|RU0nR<eP!Z
ziXC%;=bZ|p=o|YIsc}~bim_N!w;dIZJrUE5P3ehFGn{NOz4>R$Yny#dvjh~_F!>?9
zD4xGvUaWsa{9@KOS=ubVaMYz1ZFCMN^zK-_gxEZttum_fh^6DoaFeT12}s==?Dl!N
zp=af$<3+7*&NiEp>Am6^Q1ia$Yw?+2x|GJH)Jgw?=2~LQ+)hwHbc+ufaq%pT#LUO&
z`kDdxsRDt|+A<?~U+Xp4QesLR9eZlT34M$uZ_llUwVPw+=sNg3?Hdp?${!W81F!WO
z$wVynu5XBUk?un8;0AlTT**28>6WKKg>GuR<oqfh?ZB7Y)xF6EH|6L}gh9gUVxjgb
zZ3!Q};a)~jYMcbJLV_de?osPa@^&-OSgs|xWpf8#T{MyHS`oKK*}55vCXGM+P#Q|#
z_5Rlrrzfkwj_)Q-oWd4BOZAh%4b3$x%LpTq_FnU8>uzp~E{?A~gg#j>0ff_v&FlNB
zyX?AI>c%*e*|WI(PK805IMBYj%V4$Q3yaUb-b_fu057^_#HZzLa)lJKXBE%bxfxe<
zm)ZTMOLckfS*~}Md+X`kkx$}h1~#lw`w%Db?l4glOycZV>~19Z6+!g&^^;Ee97cCM
zi5>Fhm&}k8Bgg$bc`Qmh!0~T0w$w3dbcBc#aCDkk2p`;f4%9WzHFj`(GlOx+-T5qY
zrKb8uNkU&^d&9^}|AbG@`Ke;X5<8WXB*PTaX6xW<o2cfsQ>oTIl4n&4F|E;^Bvp%@
zukgH@2g_f_(>wVEdn@LZO0*33P^VgHMyD|_M+)oGl_V@J>>hg~&`_xS9(i#?KPQ;l
z+r=qlo_fvD`TNAa3c%fJJ61d55g$Uhf+Ki^2yY>~x$8M#jfDgjm*ag*!kO5P%2Rm4
z>hq&tCQy{G=W$f;Z>f0FrA_5D;_ZYj1U)imTPpky#6f=#dHKk;YNpn!@p-j&I=HXh
zcXjzrADQ|zQFF!&!taF`c|AP#Lt!V#*<&y8h}3w{i~k}HO~;^Od8x5i^T5wB<RuB#
zkv1tKkG0xB7O>Y0aIN~3^)A%Bzn=KBttul#jVMFN$IP+xka&loz^{Q^__XU}6y)n7
zW924GC%X8&Oh2$`y^vXezYaigE%xeVq2~m0fF2drB5?ZfK8d#h=hTWQ5Ev}gESaXr
zPb?~snX&B>40TfZGjKWPu2LO2Fc{*-GX|anZ4OL=4X!LSE?<Nzv|7Wm=RpDE%8>~z
zG6iIZ`ze_;agZrt`j~dY-<)GuTcOvX{*jBdPv~E`x_Gi4wdFp@&LWw0ZA>a;Aodhh
z+Z?i?D|2BBtPGl49q1?K%>?v??ZG=4-Xa(kczn83{f97{XVl%P6StOKIFLN`h2P^x
z!~+GIlkoHUpF%%6o7Xc3?>+b$YCT7~MF5)PuGV5(J(cIG?1!ORQdP@=#*bdotL^bg
zFU9rz%8&`?Jg$9D842{Wb+0-k?|+!g`z$m_PObSNvMLcQ`B+5fE;VHF@%L&n5nBFU
z2p%cABhoN(DC*@HqXfvW$!9M68A`LTWZ$xAcUIBDQUDoQ?5N&gmyk@yu>Pt4`n5vu
zvi)d>>%JX#Xn<00tNz1)V~F1e=cQH3g`Bn_MgqH0=%rPS*hjp4*pR%!O|?hgNGlu8
z)qWb)3F(br(qkpjA9U{uDEL*qBZ-W)s`Ld*MR09)toA7K3fmg$Zd)<%-=p6e+WD0Y
z{**==$qUjlGkH1Vo+1{bGzw~W4UuU+d!4EFloKHWwpIxNVtX2maN_Okt=DRCg7ad-
zH?i9*!=7LJO4=KecYe<$vrkR8bu-*is0Wu>&aGE(NZYn4<_JFUZI0J;$&gi|zIv5*
z@$>4ns*m+#B+4YO5=Un<q8)V`Rzt>u(N{CkmIq-_&#K+`6Z`E77lS-JK<WAdZ*R5z
zUILq9`Pxr^2;jZPpaNy<<{5>*^W96i<M_<RpR?%WONA42gEB1y;`!%Y`I-!>ciK~F
zuXSTzS>w-C&jE$s%TqMz*8QuNvZ~GkW_O`-IQ_P4#o^sU@Kr*y`efY?)o(4WqY9#4
z&l+w9FbU+dJ(3a;^d7VxCi<ZfOfG54zf&+g;-(`K@W7MstSyt7uKEDQb+(vF*3efF
zYK46U;w?mUh$L2|^PY51`8xl)Z&<+)CM=)FYYAx+ICB4pU0+b@QrDB%?F(90_P`gf
zUfCmZ;#Ol?q!zbLRz^D~t$r7r$k;0NaGXe|J&9;vo7MF)eH8vIN~aKWb?>g{)XSKB
zB_UR%5q!NeCu!VbsHmB|hjM%<*OxDjkD=wMv%*B|BI3#h{@bTi-7SUKz~L)3$HteO
zUQq<9X%5YUKRV*9l}Pefd>XiH)^4-ODXTSUlQ&<YYssa-qrd#Cj6_tQ`2$Lan`1|8
zdug$CIV{jegnS%*Ig!R2z5UX~j(yC2cz5?UY5H|)Mi<$sMW>+6ox+Ee^t5RjKd`{{
zJIS2dRzXP9W@q*-_=rLRtt~vxLsi?oaB6-%X+w#<+EoTAo0s{!o~1nfSxr)ULgZ=9
z6NzNlWJ}rxs<&BiDRYLL(lCK9y60%mLk+S1IppwB$YxgtPea(NU*Fu4-5wW~7e1X1
z?YDMktnM{rpty!txLS;J{&^zxZi{GILKx`4^K$+WYiJsuD5df@P5u2RS?-;^|L14>
zVe57*Xjo{raJCVx+v}qA3fq1ZBMdD(=77`ONmCe;kxW;pF<f_I^<SHR_U_o5Q%CTw
zM=I@{2s+_CloCTv-8rz34mG`$R!%6Vr|sVxl;n?m+`3;;nYUl4Q@xyLX3GD~I&3fb
zyXol3fgwt9J0!K)x*)4OgH)fk+>gKtEyMB@cB9nTu@yhwpG-#k#_`jUJOvSmn2X;R
z_~ufdBSIO-8v~7=Ox%{&mkB6ASzB183q%?DiFIc#`RPphFv|Y!&keW@SepE0r%Etg
zYzDHpW*kKxWP!Cty;?-Txr68gz^u!Qd&*CeEV=moRJ=Li{EwUSXs6%5P~`SM%9bWr
zR$7yOolw7%53;N`aL`E5O!U~1T1;~7N*S*kV*KtH&C~z0zjA=9_F2}O&s<i!KU&RQ
z5se^2QX0rr({FPR^woX&=(XJs9Cv#Nd-d3^UlprMlV*0Bm^Eq^w87ZMQxIR)n^Y^!
zR4blG@lD4`prkAbs{bbeZ%ZUIsv@1R%PUzvrRM-Q#cHmdP`=ygUjVF0jMvlBUcWn5
z!xO8*Hb4x}=T$b4q0ThYWgKsMP!Oj19#|(qeWGq`;&zJsrwZs2bMHhv`dxL4NuxpK
zuR?5dbv-N;!K|<wsq9y=^49aLg3QeV#~$>AGE`fA``!xS8Wgs$e73*z#6h4_wMs6q
zmx?(!^AJv&4uUOM!7^X`dGB;b_#o)m{S0P(rj$q2!KOu*A6i%Nbty|$xEC~Qmkc?S
zXvkX1hFZRGvKP2<LC516Z9*G@70QX@{!Na68+9Q$qKO|Zg-G3FdCgOW=^9F@gKzYt
zsg6ukxUhUNC>jNE5J4C0P(N}IM@eF;x`HeN*|xbm2P(uD%8ytyid+_q%YP?FUpz>c
zBrE=fpj}~0$j4_BB1`5VCNLT!jtF_{z^z?L38y9$J{E7PG+mG+-KF!BS!Rkgs}W!>
zwo*fm^w4eYZ#SK$4+ZoullXd4H<tsBe5Vc@2&D_(iH@rL^HP7zyCM^FvkWPqA-1Ru
zCGcs>!TO**QJ0VdBd0qd&h4#y!=4HV7Q$P};<`waF+}Y)v-oq)lx(LM^SRTh_Ogdd
zOUZQ?+}m1wutOMdw5rk7ERM)ScU$3^9`@H$k-D~*Q1>$??it=49Ps*`JE?WUPGj%V
z2Q}Ht^xGhEN?a+HTfb(E#J3J(R=qxO>&M8l4n2`W0Gz0L3G*=4h1U)KyOQpvNyu|c
z@h0!&xgV{%$1~>OG{X!Ptyl5uuuKpZHuUu-VYqY#5ng_nLv6d3(NKZW#x4*3=<2qQ
zP8|tK^Z`r$lzjgDGVROAm7Ae8Y+pglAwrPMj*z=pc(|%Kx<UjrcG#@4xsq%#Q8@ZM
zROJQSgi_o6rRAsDc*Qqlz+H7S5?xH*HC;%sCv|iv@4yRP)$S?}MuvmT8XCuZ2c_4O
zXRiG^QhK0xMvPFumL7WhJJs8I-;mo7`y))`dlDZZvlDbd0m#1A<e6RN?uQ-(oziXu
z1r`R$QFRIZkY;=+2QYik@tyPQHW?)?(EPPwC_e3?ms4;#iRAj)S06iSI^jhOsv;<%
zyAc@h^|cJcWS=5U!QrX3*PTtFHloJW50j8Hi2{vNtT&gc1V3Lj0VCOuNUZLV3BD`$
zRkg2{vWEwt2jU(V_`?TFgp#|D#OZj-cCJgu-p<qkeKNmQ3YBTW6-N+N7KUhv3atbI
z^oz;lr{gPiu@A#XxB7+OVj{>0$Tk}-(Iy$W99|>+WP&da=8WQLpI5{#H{_K=P#vw}
zEq@Lh{oDrzROklWXM!J0dyR2FX+Uz3Q@(ZzSGcHGgCrj|7J61rt_S<oPb%z=i-Vtw
zyP$?vwwD_c4C7GlRrc7bE(R99sEPS2xZ?1-IXV->sPp)3N#A=f3=EML3rV6Yx?K9!
z{&8*D{Ay+<G{<kRRi>E`Di*T1J7w0mBH#Pc@%Zarxm7s3@j?652$d!6<@7N2skNzY
z|C;P~PmL}T`LI22{45V`73sBh%`*6Ed9-aNq!io>BK;Sc4Fcq=t~<T4k9m+papQw{
zuWBGx+Zx9LwRFU8|5${?cD3>#>Rov+c$HTGYg;fgnc+slt4mOp2;^0ClG^@zyDSuW
zYaN;cmP}QNLLb4>x6NQ|Ei5GjwbicFZ0SFeEfO9UMwJ?+Kk+hbH>6!VJZzXL`W6+Q
zhwwv;iPxc4ts=vj?V(btn_ox*4$EPoq>$}QpKhP`lTna-w0Z1JTu&5H-!+%VzJ0QX
zT}`lI*rD+)_Y-)6cFgdvpgCiwc<{Vv$_Ms0uVgEp77##-Gy<93Ky2(Ce;g75Co#Tg
zZaCGr-vp@>1N}M$BB8_h^1xQRWj?hXSE<)?io`(jwy#1Uk$P|{t$CIF7gq-rz>8H!
zE%gT|Dq;v05eGyownmzb@MC)-cAwRO7>i0-Q_b}y%2|#>jo&@yf=TTv>aD6SxNM@>
z4`w)3X-`zF2$6<3vl`#1sx2@ZI)Qysj+5=9bn>OCqCeUapK)OeTu1FjQLUoialjZ)
zc?t72_$vrzptIq`EGw|EuVl_zyHML@=25F>?q}V!nOnC${U&d#vu<T1Xka!2TcpZu
zEg+%p#kfsEf^kBp`H5p%qWW3Ny^@rX{yP=J>BPL{PN&nE0mpAvaamP`jb!nXQ2VJf
z);mIuS@k2J;6iYhl=aNyI#R4MEl@JA2svlu;9X=SuP-odEAM&xik9BpxH-D3k?yif
zFLhQbOs2Hu(7@wMWLnRZ)+sroyZj}AtgO9?tE6@%)QPzS>zgt1|6_b0?^ZZ_2|J^j
zY(v7|`GQ%X)`ag;A@mGUufxcToO<~+-(OZNMONEJiWuJvRhDL%e22T7UgEp@QFe<)
zEip2U^TJPgWOZ(sXRgEOP?}?haSw(b?ST~nM^^Agbzb|9I!Qtv>^EN~=C;4NGh||5
zBo=^pfnXzPcH1=b-9r-gyMeTxj>CSjSwQm|Ea<JLUxmE`>!_WZ5`|>q$gO~q$5zwU
zL0UbWy#r3R`60u~U6ncI8(a;sC2m&136WlYlg+X1$m}>)*#~7PCn^6zC&_ZXrbJRj
zw4G+q7J`O((FQTTGMt}5B6Z(WTwngEtnU#gaN8riVNp(ZM#TT>eYS9P&NM;``~FCF
z%Ejl`C^D0|DYaDLq|3w5phaJ`l0RzU#hvz2WXnsLSINoALCt5RqvM{P!<N%(w*gIX
zBYB72NAgJOLw9j1=D3hZ4Bb6&mv;NfQQ88C8Kho3<6857vG?9#O?7R$IH+_G=^dnr
zNRuvIMVg3!bOHoLdWq6|5TrK&0i{LhJ<^qu&=C-kE}hUj2{k~7XTNjKocZ4O%*;3E
z{eHi>X3q8gvG)aJC#<Z!*4odypZmU_icjrkat&FEd&jR8FLm_KNbbi?6s%iGVH#os
zElbU@!D3e@*Oj?GS64UHAS0l)mbLX5U7@a9%0C%+v+tM0=xo1syX!8W37!;aA4dWl
z>lpvtrNNO(sj%Y#x6P&g>a3e-b1D+v`+3|<<eB%ER;E+T&C|ChK5owOg;{(-pV#5Q
z_QF-0v^|g)KiFGGTC1aq9KMs+mUekplD2kP$&11+YM@FSNRA=-&uU6zd)T09l`Tk$
ztLt8~!4mwcK!TzeS2t1O6U=Mlv*+G2<<^>eJ4qndi;48Q7xzP2ORI}Xh!Tsib}BT1
zf<x+=`Rnz;j49omH)YO<ql1kDEXil4Sh&6=kdP(27QLi=vNL-tWN8)dGNfWZAlZT7
zPXT`#vWqnEb6dSS6<4hu?>^MhrkU|1x+Pbg+mJHUzsBCA3r5<LAQe`)B@eOy`Nc{|
znW&R}OnPjrSr9+A^uk}_Ht%?@^7gala{S@^3Swp8K+eAx4wFV~<#3k<nqmfEkzP%p
zAo<~Q@DH!$VJ{`!N5i6G6~nU?PZlgSp0c;wtF**Ol0W=b5NjLKWj}DP2Cj@<h~mu|
z-!q!vMW(Gi8N$hxI=$!=d9FDb+`$?fHB%e?<vDNPtv+pz{BxR%%)>;i2G5XzW^?G2
zLSYN2Ea=&y&|a-Xs$=}(H&M#C{1O{({>Q_bi<z?9Bf0&k7BN*F+STEaM@g$SpR31;
z>YF39$Rs{DCfLF}!{2=oOYfb{fZEM^;>v#!aJYy_HNdtYBWGjS$0Zo24KFw;cFXJs
z+(pzANrBB>4N_hlc9y%rxIY=P#ooskZ)@67{4({bWR_j_BAn3|-ujE+TC>XHJhM*T
zCdSlP1w3X``=fld;W?@Ba|i9_>eChiT7vh>+YE=q+2<;b@`nEAAYXTG&P+!y8@FT&
z7N*`W&z(b1S*H=|@!FER&)hu@*jIg_$Cl`mh`I-(u2N*HjS1gfm1Rby9ruhC%)B8D
z4|42{VwWw{35?oWS>%-D1t2gLz{Uu%hsfjKJrHTYsc4_g?_#h>2QXa+YLIrWK1pfI
z%-8C`BJ^yj-M=u!9^yex402>$C%hdrNyR4EsnhzES71@26C5@F^DHM?6rQ``-w8nt
z7|#KdulkFTrs$Cv_fA$&zT}QWw*><&uf+DF<$%iUa}jL7qv|nN-XE2fB1a<1I=7y;
zjJvrtH_0*w_)^x0hf0X=t9=SA>**8ETcn=cdWgA;zz_XcPK65tCWFaBr`*t6bH*72
z<;j=glEtkzRqff);wBco1!KWf$bGOF=o?7CX8sq!<OMzAX#do{3w%vci)v+Mr#6@?
z>_N>)pxX(83*Y~Caued{9*yv%3-^BW#<UvNB8xjb4S>=H5BYZkd$Q0TwI6$8x{=T7
zR11D$O$<ia^pkd$vhxyFXCKMT3zx?5QGM-gUu^NwzexLy(o~Q=7`e+zBzrnpHp8na
zrX}mNT|?~kCM)&XU2aW+VxGY$6}xQeF4m})Q7OSGRr(i$42TlNYgQUuXLVm9<ct;G
z<b~WYSxR`DyPX`Qvap4nnr4tGC<PB->mVK8enm0CqP}GSos@n@YPz`GQ;b5_&)yXx
zQ<7)Z2+r^O_|Az&E@DcbmwB9JiKve{3&OmPvkkf^!i2R=QFZV*)~7MeT~GhEK63v@
z$JL<p9a<HoF<t$n@l9AwPD;6Nj*|bL#+T$H(Pz!KE6e1~dS5c@vk#k>=aTVRMn^su
zmJw6A*#&Ih`NxmMn6NgT#xg%DcGE^ck8ixYnfBd_)9I4r0Z9TA0ot3U+g(6|84m8(
z2`qr?3c09mts@sMwp3Ta)L7$rRC5m9Sl~U>76>l+q?Qb0(C(k*Zm}7!_>tH7DCt2w
zWw)BKk>3xJD-7i24b4gzx3$(6H^OY`dp*j1VVx%;ew<+e<)`{>#h-<aoptAE2PANv
z;d)oyH5m@Crk~5MpDE4ekNfjs?%B&uiFoeqXl1ld-PJOV`O#+`SHf30%BIUp9ItT<
z89EYCa+z2ATu%mlS)$?bv~<wXi=lIOP@(4~ec3G%dzaErPp*oG5J~4Y!12>~%ZlL+
zw_gPQQ1|rHygqoav~R*zTQGaeg9dbE1(bhL2U6<mVSsE&wNCG+o!)p!x=ZPPshZK7
z)70~doJ-swCC&x2)~SXTNekl)%bVL0tO!Ev8IQQl+ME-OdZU)=B~*FRg|%YTDKkHr
zrgFC=9~|2cxg7@a>gaGD=tgSMR%A_C)p*!=Fu#d-%NBJ*%R;kNa@VFM#IBd6(!lWI
z%|d2p&NWN<_8`d)d6D|EMm7%j&2hAa<=~a+Vc!B;>uQ~*ld!#4p?5Ad$d*A<fDCc&
z55|5ZP6bD0RU}}Gq1GRjR*B5Gn2{MjYo6mBRpcn+e=kJU8fTFxal6NRUS^2HaOJH+
z+`iVifW~iOO8UhM-3!2Np{SJw*gBDRJuR4a#IvI(U|!C)lzARjn(@p@)qd^Ta>816
zeiD9H8k#mfwIE>5^kL5V56k5jLJ0GUjxHcR2Ld~e7f!&6A(r$k=4Srw+?bmLz(JS-
zEpy6L1<)uD{3cYU7!Yk5kI6*h5o`Zsfy%D2BmEKB8Nmo~u7y}-`>;+QR+rxALs_!#
zYnMo>zEh*-kXP{pI5F(21#q;<6g;}t(%{|@cWqOwQMzGtuJ!;%zWgC5+O$tE^h+CC
zO&9dO*Qq^_T7HN%fC74A>d0V8bG{K@uz7<elLcvczvBz!W})({(VH9NbhBOdPjHE;
zPChG}DP6-x1^t}1KGO!@5F{IOcgBMGHjL+~Z#33pgsNy}Kr%!0K^*x?SFAbRNtJGE
zU218cKL%srn3ScRxl|18{=RALkvd{xSs^#T0w&^`a4Gv2febxi?XEE46)r}s*7<$k
zhJ~H%2^{x?w@jGcG+%>`E#>~h4}p0Bf(ij4SLJTBT*k8vE(u&vGQzOJ0JCmUXsx6#
zImQ4o^6|W0KEeJ;ZK?H=jsTmW!h!!zgYCh-lIrpBMbhCOkpm+_5-pk~;#tOQQ$Hq!
zb@#p%QsOSzJ4iTafl4%%Myqs2Ca4f@Md&sin_SLWwhsObFNQ<iAhC_9S;y(-WAOtO
zov!qs`U^nSnsK-LT8plSstc`&P^0i4&ou+B(DUs<L}2d%?{xXGvBnh_Vf`@8ouAh3
zgqzbO81nHS6;`6j4ITqt;(m{HjpedNGj!Anr+v%?JnL~S)IY_3&QX*>CRI?EtYkKj
zD25T!mvP(;_ruWl^U!dc(nhOv=Uhi`f3G+SS6h2>P#dKU1-rvXBaKf4<aOY?!OZ`t
zjna#GTI4by%Do6%0}SA9G?bAnn&4uU`2ZzNAmFgYH-7XVa%$r6Rg5Xg-;j~N`rf&4
zBhC2C8|KU$GdDcUA9GWWlIVfnr4R+en1+Ad_Q`lm3P7uL8b|8bG{p|jKW9+Hf{MpA
z3*%oIYp7q5c%yke?wZ4h=d=!+Zr!Jmou^e*b<x%hb@g!?^y0;Y`t$@0cL@mY5Ey0r
z)13NkUj3U#r9B91v#>{tP0__L!cj`pMi^pP!?u6jaVF)2@$_Sp{?xbC)>&~h1D>P$
z<EZ67cCVohXS<TQtG;l>gh2*<ee%&WeaX(aZ%H*vFOBTo>9a0vYZw@>=Z28q{Fg(>
z$=o%krfBA;6XPea=}h;KuCoVovY>thi<#-Q)&x@N^o3spgjjpjT-ua<c%VW))?UYv
za=<?CXO|<HJAv>M>B|C_v+QGLbK^t)J?8vspujJJ>(j7v(s}5HMXmlH_F*BiJII)#
zViDj3=%sG{se@PtUs3A%&vqmqGcQ1fQ(R}~uoHvRIszACf%s;Ik2rO7P8kYDyYb2C
zjIrc0$a2zxGCwLfEA%R<^}Q>XNUyU_SDGv`b?#h71;dJFxAx~tH<?1)hdCs*8%Ny)
zvby)LXu9SmtM04u+Pe`Lq}&~b*c{L;$>Qit7Sr3Z)pyTPSB7x+(KedH8mRNragK+*
zy8QY-OCORHUx)3|B?kw#)MF#_EGU#f3ww#>V!gMn%d@Y&WD#(@DNA|Do#eX7&U^@7
z&^KUF_c+v{dcMc2sF8mx0bXFak^W+jMZ-CuI4|9XU>fW;eyf$%C}hhyqmVIccDYVD
zmZ;M#ff!bu_XK0fk=-JXOY8ziLaP&Zs4?C93PG~1HMl!N&s}SrgGGp^G$+1*yg3~O
z>aQihB-W(QY|n4td+;<DDb2h}IdArzm2ZrOBB~Y@x;jrP<yu*kEv%F8U-Dm0Q=EYN
z#&(;lyxy(Vb~WDHu5mW$c=QP@IRE{#xYkYGw;V{mC+zhk{c+17kPZ1P`SJJ8P`7lN
zHV@XmSRYsZ+7C2uI0YFj*~ZN7adbI8ek)k?HBp7*-QodD$ztKkbndB0V=z&pPzuJU
z3rvpXQXiKF>@Za!d4mvV=v5ef`owJ=1@%u4tY2AK?%KsVh4)@9VwNk-A;tEh5~EnV
zx>E{e;P$0$1dTHC$6?TXT|R~S_q^I)c^nO8E=v%To2lBkuNm#R_$AJo1aP#_>)!=@
zD*yDsF53w|Wen8%MPLnRkq@;hHrsyhPtx+IKl&eO*8lB$*#`l0q-8e1y80>G-i1(F
zjN<QRsvhx01IF-(2sj=7X=BA+<XD}~N5DND5J3X=oraRXi!44+#bCsy%;(J;#}L=l
zP+R9zVx|xl>YCY2Z?OOf3%-9;5<#EA7%Xuo>4m1^5($vQ#PlSnpeS3=1<0R>$&03P
zjr`pMZ$Gw=X}XeZkGK6yr+enU`z|*=i9d%UHs=FwiKPO!$s2s8-E`^(C=nn8*vxk2
zYZ!r8FBl6OjM*8Wy&T$)DoWblNMxvsajFtarG3BpMlz^$DS$=OdchUA#gI5fgGCrI
z&Je>G&J^w9+2Wx+>gpOO;JnBcW!`S3`|MfdQyp%RO5$=p@6S|1wbi@A1$pA4^949t
zfO}=qp>nNRd?Awxq%n@uEp^u4kZn{$R4`3@jyjta(e)C{U@wN_ZiA0LtbI^YhGBeK
z8|!+au2*4-A|dwZ{iZE((B>!2sb_^dtIH(cxV@Eqme_PZ7Lf&2jLokw6%|4Eq-Snq
zCi2eW8?;oKApAqT%UaF%^0vVA!HkHK#fygri!nIWW=oq66V|cL=%<%Jn?fWw?fz5a
z-W5{vhgBtSWOnu;d^x0;<IGYlq=Pol{5Lif+1)wPVYs4#7X&`@Q@eOh{-N7SVt?SK
z?8AOH<w2f-lav#kRUlr{#Eu;2W1lX-NH=qM)4R<@%ZfJq<Lgs{F49ZQza2NwJ>nq~
z*t>iGi@kN{wWO~vypx{z)*Q#66GL!n`Hud_Re7CWE$$rDI?hVQEbLnRA3jsNCm8gS
z-1;E+CP|6}Z@zHTk%?#wp0c8^<=rv@9^p8Ji;1<1>dYgKz3b<BR})zs>!*TSFvs&c
z<94x~CkSD**KD)@*jQILub*d2gUIvYqTV%#!6%Iq%BOFyQ*3k5DUx9>tD@&)@YgpH
zg4Gj>;5@9;iSCd|2kVtI3WpRacZ)~7$*UVHnb$OGC;`N<Tg0lUI1r46hX<;N)-3KZ
z_<}mjUyN2t{oE6k{u*AR#h#EWDI9wHMyD%6^q~Y$31U}PdM}of6xg2t-*F<YStW1N
zzk7arTH)de=?%n8n@O+6i|75<h<~_i(raiIYXrmY6gc~f1kAXnOqqkAsVNg<ez?IT
zoXL_HRhe_zJOx60&}(6J*i~SxF5njdvy*&Xc*VrDp4yehHyMt#Lob=8XjU|IZ{AWS
zTj&VF2$@4(n+rF(Rn)X;FK)$hfXrL4Nxo;G>7cj0q+y*&)0uSupnQvk;~s6^-RGo+
zT#XJN9WsDz`;;!)YI8lUKVZofauFzlJA!zh?Le+yq<ig{baGMcib_)%=#IPOjTS{}
zH;+hzDy$f5^qT0wGxsRwTwS-;LfS$Inm*ZBA=T)oTM*zkYG%%>K(N9S`X`?3SJ{3a
z^Nk=;3V?G1rMwBAxC<~_G#&@m6>g?8_&$UAl1Wv~n;xZ|W{ImMe5FPEH%i6dX6IHu
z%Fl%h*eKxXkbo;h2jB|f0>%Lf(QK?UK7s?_*<1$PB^FE9aD1%biz@)Ow08!LR~CUh
z&QVJcIyd992{(GLrE)I=X{#TUenPY3h^$5IN8$v6e>0Sb{LN6}ClXt#@`q-vFeH^k
z0)p0gz)|8U;3y#@hQF1FeT<X-MWA{$71;b$K`}=E%TdAyaIawYgyX1U0T?sy_w97k
z|4r-u`?eMX3CcadQR4On7|RAYN-+FSM~T1JqJOfu_^Uzx&OQ69hyEHv|9eIFYnJ`(
zdimdW9U>80=#Q$bHM5I&kijChXUP{ICM4MSHk5ReG`EoqIWA34a1L24*`xqY>;LB7
z{NLNR1F!uZtc;DJHP8@^C|cApK;fpm8m4%MaavRtOMWp3Lkh@?aQrjL1I@6@0y$Z%
zd+YgQGhjQ|%*X}FP~`@^1$3mW-i*~*YU*vKKoOx0kH_&`n6$hKOWz@gOMLSui>_kV
z96q|B5B-^+J7b1QD2tr$oR+1G{35W2MVo-HQilV|c3Tf0KuJOBl);g_nUNluH*B8<
zVyxHP#`n54Jn7!4ik!-z`e+365Ea@rzZp|&%e6Yc+pzp3#UO4RGrhNZI=dGXIQSMX
z7D$H)E-2or@%8D5eQAwgL$|b46qFzTcwaPf<kOfjcYF7vscJ|;Ot&`SXKO%!5shHM
zeuqRQ`&LeOFqwFlVJLJk1T_h}LvE}Cp8VCR-5a5Ha_=IZEWC*`@#${@GP!Rv0_S`H
z8u?qKRC>|e&E9TN3VTj)`mMxe#K984Bm=mG&9aNmVhoN|*g}$61B{f2<HG5oI?`0k
zOSkUc9oCGxFY!xiGYrW_%3w-aqJc&%^%9a|-Fpfe87R4EGzLG6a;X%3KIW-ppw)cI
z&o+U|Zfcd(J73x3@&%0Zky~H6w+7e=DAq7%ghQe|4}G4GUM#Q3Hp8KrNR+F}pnfrV
zZdlw>c8lPszNWUCHE{Qt|73Y0!ti3~S2cVYF5I`ME0J~Hm%1InJMOu|r3c8hQ(KO<
zSsy14iHvPm^_3<GJtXaBPpLCQ!+4sWvP~UPpOPZSpl{&x*e^@*Vkm%`tG3ADG^L#_
zrV(*&_*qkG6`7hQ__B`sNi$4xuja~KmaNhU+rt-?KQdeDM==K!=g~r2^%G-`Q1P{{
zspX7EvaH4l4ar~oSm#X|#dR)AiGQ_8s^I<{S6$PXTm<R6J(xA2>*vEU7LHYWD|K_>
zJ&)&Uh+>Jg<kw9Jq3E$@ebl;?rF_7ww${+LQRf@wayspq3R-P^?y8LMjNb<dj`AXf
zPr>?QYK6T;Q(KQQouM9lF*cLQSC>-irPgv*62EK)WyqYk>{Tsk^{N$CAp3*F#-}Zi
znAjfkqA<?@{f}n-#ZLQ7jpvGbT7u5L^jf!1X{wr*qi9JaubS+UbFnFNTLNj|8gN1&
z1O>tnt-!<K(m|qF^<)>XI=^Qr&Y?wD1UoJ%-75@qPv+okr@G%aKDF0*GK5GhGH27|
zPs6O*60+p0KfM&c$CE(5oGfl3u%pk7^*E=&;zx9c)6GqBN<KvvJs+p}1ZmRbcrv^q
zq^C;Q6D)h!^q)V|vXK4WS0MK!j+af#?R*|r+c0pXTrX9s{<ti4@z}Bg4;}9rlBs-N
zAN_9SJ)r=)5(-R6T!<TVTc`ve3<Y&*0UDG}CQc!=a#;DSqpUgoS$LL@fDVGlO(ZP7
zFS+Eh%d3aGR^Igs6oe^UKoaVd;5QzFiNP;|;q#a^DY;qIPV*%-u1RTU)91Ex<{<aY
z4<*x|OBdXY5_IoPD@1s$MJbxt6j;qF!^r|1@-rP9CciN8JvHY}iM!+xcV<^<-usf{
zGmqG99a>nXv(RPwyDN`kFS<7s(Eef3i;gl_dwmKBgdOo{I=><2&a1f24#!LC0?E`x
zmlJMYo-F<coiqMD?C{UNRwjGaog4k1w8H-pjQvM{e>zXO`ABKvlzdeh9{F;MRxaS(
z107yXhF3EM38us*2BCgEpAH#DoVfOBprybL2n?G5)43rVjw-~o(SV)l-gf5e+AmF2
z$?bOn$3q2=;n`knS$vq~oo+4m#rh&+Pk8@mdmXvgSg8=qu><{Wdb;O|Eao1o9CViD
zS`Fv+gGoq_5=nu1WUE9`R<+YZ%;U~Jr1%}Vua?8N^#`BEVCUO{rNGRUm5KRItQ@PU
z{=#nMN4$hd&u^>Ux!xfhWthM2>#}zJRm@tmC1#`pL59snAuny}gGEc=qFbPMOR6sU
zHMh#DUbLoJPzr7eeYA0&Ra3H>65E%<=S6BEESWNY5vWM+T{yQj%0{i$Wf~8A&OQNI
z&f6y+ldd`ddc>#J-fNfdY`uEYfy$PdSC@Yv!+3i`T_$#0ygT-4_jeQPB@%%+ANU}U
zYzaZ{q<8aS9XSf~0v2WJWaSz(n_op`=(E_pxwc|x6BpQ@EZ6hV(Imx*HIPY`k%{&B
zu;^+{qr;SG;)!YZE+BaJiaM!eaP|<}I;)zmSyBE83MwTK54l5dyBtojZm{0?yndvY
zqG@i<B8>e#o9&l0Ble{?m4qjQR(1Q*^sQ%OI3m;rNA{rz_xl8G>GB6DTf@NDBmYh3
zcStY<n3*tgq2Vb&Wycy^%EdC^+}*{gh2MwHKSSKk4N)H4*F8zWpVRvZocI1B2#UeB
z#}J&W^s3D1o^R_f!}bI&j)j)UZ(iO8*aW`_{8aX;G?Jmsu#=zgxW7K~ufF{6G>F<i
zTh`oOxVRdiZ;GG+bOYKz{uxdIy`Zmsa@l<vTk0SB+v!i6G(G>lwO1+2?}2GO;;w%n
zA@Hvu@66_rxYyuK<X4kbw<^?tx(c)Jmt7q95|#1y6C>>{p+PF&P$hZUxSwWs+uG;v
zZny2s1a97CFFB15e*Z6PJDVN}Av#fW)N^UEKl6f1Z-@su_AFy~+NddEb-jZVW)DGV
z(K`Bw`*NkZ_pm5tputleD6S|x56*0)j^pZxfz~0Esz1m!$lh&$YnmJ8f8T-<@-*w_
z^)-^5$Q>A=Z}5DVvT)V#?wsk)+JUST9|@a{9Wx)z=+&`xWpOMNJ&l@!aRA^H?kQJ<
z%X6rRXkGl{so@)t-H4>S0#gMQhAA~5|H*>WRHi)RmR|$|DnD=)=%NlT_W3}yA)L=5
z_0`QVD~xD%#jMlAv?o#9y$!w<AFhY3B$M=&`*5_nt2qy#7grD*I7LhF-H8HZ6}s~W
zhnOC^rMsd?Dneb^K#ih&d_sVi78Ps}IP(l?SRPE`Yhq9nx<6=$XsZ1jZyT{l5tSoH
z6JeAZY8^^?JqCuSCy7L^9RTqWvg#KBrqMkis|lp5?E&2fzY_O9usXSL#cQl8m!N_k
zpBrPHp-b}wn-g=C4d~JFLDV#}s5JP~{7PtTlx%NWtW%jBKUv%LQ0I-r$ok{)9I_!)
z&G3;3?U2Z*YSHEai(ADA_KCWLraSQ+(HFPxbd!v7p+6`p0S~YY*u_O6P<8-V<H;{~
z%paTr!QT8&puF!aD0L60GiMCYD?VqsE+Zcxb(_)TLbJkn3apA=6QZteyt2}*{JE(c
zDq@o<z~{M%QPO=|=Dae37n;`oG-m$kA;GNc6Zfa}Fo5Eh3lgin@8+fLFu#-h!MEa*
zA(!+}YMM%OYbYUceqPTy+<6pFkEK%^lNylK$MB`GCKT^+glCT((hPWiZOVS8uJ`;_
z^P3L~?oQuR(BO8~<M+<R8yN$0c}278{0&XDjS;4W_a#_3^dheKkkAvNJWVQmYub<$
z&6i)=l0@TPLQuVXnYQ`n^%c%e6I5Bx#IF-21W8)n#9l@XONe<%sBD!(ac&9e)V!@X
z;m}_M;e@EYQ`{(kwQd7{|FQ+Ygp;Vk*doI#1MR7h(Yy3g^1_W3rE_0N?TDTVHHR`J
zUe1L*R0Ik%bLX}UpD!T84!@M=Hjh_U@FOy}$Eyat$fY}*Sa6@)bl=zK)eohF+x%b!
zm>7TPa{^x?hCSeE`}gk6f6wd2p;5sk>*kMNmCUR96xY^|4}Xi<S2wKwc%5C~Q49fw
zTnNDSQFna<@K9-$^K-6W1Z;VKoFxRn82sz1-`0!2dg7l~&cAx%|B;^94CDpsr=$UC
zDJ#H@>rBw0f`UU?=|=Z>H|+pJtsuGN8`IKaf-Ilb+z?R^ltEYTx6_{{F;M=FfukyW
zR$o%9a&)4yTXf(=@bm8gs{gk@>}3i)fD##dVRA~2okH}Esx0#WH7D@*5i%dH4leBB
z-n#PQr=jcVm3@O>f$|^qV}Pk2qJIq#stecy!07yx_^Ru<&fsxk3gB{tSy{Eo@-tgv
z=)m1_j=}{M)aC2zygZ++8sLtn-i66omus+WJ5M&-F1%CbaqKO^!dvkS7^T;cRsPEA
z#a({L9#Qhokf05x4Cg_NPB2os2$Z<}`J`cC<bs-M0>O6=Rnkm`g1*l1Zt_6IyKjL$
zJ2p$WgtvWe$YSAckS~v>erucNb1F+!6};}!OZfxmhv8b3-s^~G!0F0{7L5rjs^0kM
zQP@*3KcjOj)Z|bB^S{WQ<(uX{V4%2Zm-HfEP}dPd^2#;P4D~rsp%O6v^dF2w{p204
zBc_`wY<oRTH<ik!@#w08Q4`nI{z8u1lUVJJv`Fm+X9a4Wg`&ypu9>;9jPpEPqKc&p
z9I@*+>*K{#dJ8sOr6$}^K?Zo_!WoQn8mNz6SrUxACX6%lD!_)Bqjkd=pT~S3)^?cM
zGfgIYua6QLDzK)=Vu2Nt)r>a#EbXyigQpi0iXcn<WinW2SC{%JquJCHs-XBRgVTxE
zciB$%74C=rY`@HFO}C$Dk%CG~k!p0Q4wOSX_ZGIM)r<DMo@!4Pj%L@`M)C;d-Z5&W
zEY-Mf6z|-chTy|6zK3t(ub^{EDmS1QRH@acjLnfO{HAf8aMr<olDV8ft|jPDXEtbR
zqaLB@?%3Nmi@cHsm|pkdR53PP6P!{LIO(gkQ$%%Z@qmp9bz5mr21$x$L%MpxPAEiO
zJ$3b)({OnV>(jRO3SW34f~$WCN#mQ|Zs3~%rrD$m)M*H%az|Y^8{So+P*CNKqT!`g
zyTyH~@Ao)lVv@D!wSh&|xbP7d5f-dhGj0YlXlP9TmVhlS70fpO8BJ=a##R#&BTr~l
z3y!SkWQ&;M;055p*HXn5_Ml?k>S^*WiO#2H*F=h^7I|za<;cAc7Yk*q9`~b{p;@2<
zWLm|;c`SuhMTroy#c^e=XyCSQo${SPJ*+C`1_;VuJL^t$i4=slgtKY2z@FUO<=^L4
zj(6!8Paat!>71ttRBFUfS}DLDWvXE<pDvMchiO;8ga+$uycG0({4?8i%*^JeFOiwD
z>%4v&%EKwGmpZIjnigcGRoq%NP^;#3;}?O1S64$nuV1xo&$qzt-qZMd)S<3B<#yGP
zt#b@9ea=L~1q~XPWNR$XUEizKQ>uPTp@F@HxXDSK)_U(j`L_3p-0TCcO2qg$ovHav
z0y*`7gvjhb(dJ<<>L?Xfw`TK3cPE3E{PG5nc05__I;Hg*jQ6`Na(nMU^1HLle5Y>t
zra2@WN3{eRjKq@;H8VKE9Zl@U<?N8*Q-)xv`7Yil?gZx8IKHAmi$)Ag7o*V;FWW=w
zxVNYhiJ-<@-sohzF?MMt^O-zz|5SE1$!yQV%r3;0I?Vo0cr=a-r9u-+7Pxh<>^1;D
zX9Z$T(~4|f$T2G#YI)qqD@;32(L(Hj&gM^R2dw%tO#lwFEaXONi^w!Mr*1F~MuXEw
zWgB#kEAywfb4`LN4u27_X7fxuu1}MLXIUw9?<l%05?{NO`UD=IFMn<NJxa;B(~1du
zx=3-ZOA9FM=lC!?@u|n}m0I4%e}r5OD$zHyHX}6FOnMCXw6U%UTV`&6fDt=Tsug^K
zphNF0Z8F;EHao|LuC9(kI%(~u*Ah>X10#9Y+~lNLv8Dx!b7Eo55)IDo4Y><C=WQ=O
zol=PI)!6&So;+=Qed>`Q^w{>>aw>A652z(es|KEJNC7JwrksyJCD$kbk6VPq=L?vk
zV}<XO>%+B9D?aBztQzCocLirAR#*J`2O-+;<~$GFY?E~}6~!*qDO2D)u#rnZ;tO9=
zykXwS3W?!dx9TagSbu5S%E<DZ!urh&+l23&w7QmlCghzdx0O=q1Tgb$79DXKuIR0{
zdOUXyv`)M1L2-loBZl;Ej%t<r=a$`z9uZGxMjOE$TO8C~viKV#&4efpl}`T<o&(c5
zA&!-=cfKau)?Gbq-PFCOhi6PnNA_*)1;d}%?zv8ibs^KvT>|$|;CSHXi2P=^6)ZFs
zs4WmzS+^JI506gB^eURZN7S;C<vqxhy38YILifqb-X-bnPi6CMMSToK7Z-Uj%~r;Y
z3QfwQaRq-`iWfcch;GDYm*_4dQ)6QP?Ve-z@_}nv)>1kMJByuTSCtNTNAM)<89Ffn
zK<TP?=7u}A58j^hQ8jq#Kbu}wT`4qrakcA^L8Ajo?V`w5xGC@bYJr;6!9O_v+ev+f
z+q7@x`-pLgT=S6S1Tt&)Tlpg`=||o>CsJ@kPFji7R4||(5>i3#tp<}|N4m}?Vw=rA
zVv~D}8)`;&wkut^I(^I)HlG$8Jct(dcljR6$*Ug!Uax+GWT4h&QYaD)3H<<=^Yj+^
z8URl+bT^DI!oGGW%lyIjyF@V>E+4W8ze{#OF3LJp4k<=d(6$1=x;Z4^AsyuAGGsdM
zgrUB;bU27(0XTMF4}*Sp)#%&5cf}9l0fzWv_%)y;f!L#81Z{6&xI~SqW&roO+amdk
zfcOb?Ux~nFTsB^1S!V*d)xPvwU1qM|{m-M6#jR|wW`PQ_+8rex83U5%qkMK0EA^Vh
zgeeJ=T;>E@>*bI--$?u`e|tP-K0tT$R~}pBIwRVEVI@)klxh94T}=16NfdOQsSvu*
z2bd$}?O1XDBKTkhK)+#vDyN(wG7f)z#9w{!*SPqH=Ex=7{1$7B-3E){x0vKqqA}*`
zh@R`T#OeVHK{r2;Rs8=f0r-<y`hSb{{4d%2Kv(ed&HtT<1Hf8)v=9EFOwum`SLC7g
zgykZI8RAx;_jof`u~Xx2Dr)}fy0LuZq^Hi4rAN-wv0o}p0#l|Nb?T*S33xs~dH$^J
z)8|DzgJnTtV(ph@Gc%@vuMq;Nr_nv+A5tbh!L>2Sa8D%T)_4X!y8_$PhUaOhcNW{x
zq0o7D07{!*(Pro>|MB(m^Ad-CLAxpS)xOPaWSPpu?yOT1*tP%x%Yf{efYIdXoHsDK
z9d~e&3p$Ej7^ZId(7Bmui!CPmM=?suWOS=vvU#)L|ImdnuEf1(TjdV6v9KyQzuS%%
zMyJQ@f%JwvbyPa$Ze>~yr3IlxJ<nq`3!Z!kPfbd*j=TFcFup0OGDG(@g0XhUlfsqc
z2?i2=<jSaB;)=UhxUt8781H$$`wi((mnImv)=|IyoyD+nHjla<tGaO0wKmhtRr<VN
zn?u%ccqMeRd!=jNNkW5vIRz%*p}EK6Qldwh*z<^W?qmXt7!Vr?g23OXrfPYJsOTQ-
z47nJMsW`+#+ZrF&sl+MLK=L-PVF7w)m%L>`TCH2!F53h3v^jej%qd->jBq$@A@iG^
z1`-AB=5ag_E;s*BG(0PMeK^b3aV%fvsLX;P8u<JrV?W?SJparkvgiZ_!;7gmnFVj?
zt{XWuakW9IW*y~c6B-t|*nQt0QIW58ys+F%;~OY4ef&Y2o><|*lOC7ny|<Hk<J~)-
zbL`P$g)}GXL}ZPJ>js&9RjsP3@!u^$4@OLTrR!Bm3lbK`O`j9Kx@E4^XI7s&GYR(z
z0H38IN!X4Z>%3D@;3T1~2_V6T+Yv)rO6gnpW_GDiE};isNgV*CH<>pn0z6zj$}O`c
zZXRO!bTd{x6TI#3DK&%AB%VXwg`dXTz@`*H6F1QH*(k^r)e`5aPV?3;HDl6Awv4cs
zg)?eIow__7f-kJp$@{WUWO~eR>j2Mb3Uh}>qWZ8B--h<H`dI$adi?bf&y(ac^~RtR
zJ8-ReKRv<&HSK9nS!&`o6wgJ8Q$R<)J`$sxHx-L&z91gaua6pd>i&zs(Rm{`W7RJ@
zQiNBFaXU0mbjF52dv&i88I+AMsoSeK`dK_FU37#G&H$+rt4!4oH4BI~<2lmqxCQb%
zFUE7u%qGg}`+@qL6<yjjQ<NIrh+duF=ae^wbm~qkbbzS<k1;ASlP~Cj?4F3l23M=T
z^UfiV6E)wP{O@TpgPd;ld6i`1ZZ5?v#C6jP&=CG%WfsSSo<&d;o~zll;Onso`2_hZ
z@`CR_36d1Qen@v9mu+KNzwOzqfj-e1Z`Lo~lJ|)Ye2LO8dj~SD`cc6bA)=LM^uXSy
z&hH~c#E4?LmyW@x_v;mhQ@`E%rgL~v&i6qdPnhwmX^=PIk%X(PH=msQnzsm^5;7ce
zQ|8rmMlpL)QMB!wgE;OwxpA(7qFilF9Fk|hd1DItrK(qA_&}k`n&yD@%n4UqE$ot0
zdWLr<&t{Hbie|(?J6_~GIP3nT#lZ0^shc_e&tz1u3>8kC#^9oKb!1;3>WjsR3VZ1?
zZZP3*AR0#PT0T{0^VelCbB9x!HQX|BZDiDX5Ho4rS~6`K%!SoJS<)`{oh?1V*wvJ)
zF48Rq#6LNIBqY5vSsQSd-_(^i?s_<yQR=ZFt(M3a_3?=SF{Cyg8D8fv>jqX(x!^C@
zZ$VnhX0{<~84DJL2Fbuq%5<?>=)}0$CG>7_#`~PrC)bk~#p*!sjI3$9D5|&9tlH{R
z^a9x~KGr2hn<u``8xQ25jwlBLR-H;mI@Pj^dUaECqBOab-g_*7vOxWQ`FXlRsBFJ3
zI=vlA>8oKr>}E5j2rk}`r+avI^;MAdj%{9T6s9WDwzLnMphXk7!S}QB{!+{jG_+%$
z+?T(Z+~qljuc+-zsXO}ValE|EYKLsnmx9jAqRQXi3DvsKu*cEV%^m}kY>}OD9OC{6
zc<aonac|jwg&`gwaN8#0&_JRBZRGWfpnDzIlvnrQeVw4#Ta7@)XfK}M;zF#pFmx@V
z2q34G!SS*hIQ}%1gAO3^)Mf-sShf#T8A5D$+J3H66p;K`DBb#>>7YM+@NW_5A3od@
z+HPh8RPvTq!7*4u0YrK-w+e|82C`W>tS~%zB@#Q!g-$aC{J^Pnfjy@Q80Yp+Vd~#)
zPB*3q7u|Sr;cEe1uh08K)5djUn|1A9qLiHYBlRhd4Q`4R&g3q4cch7jkbJrG;VR-k
zubuzJy6vy#{5Klof8DHq`({xbK+By3bKGZc^?eL&j3E4xXC1ene~G$^B{$?oJ^ohC
zhu==W(+U5rX87;@%<?;^KLT>(YT)p{YO9SBR;khJ5&nO85Ho)KXZ;VRE-2-n3AKT)
zcR65f^QSkbG%o4^@wk{(c4?qRYzLTL(IDSBolzbVAqA%TxuqL)^C?fS^uWyHa*t5)
zD*=Y0i{CWl2>cTN_Pv(iF0ohsJ&8coAxpbKy{Z}}^T)A_u(iEFE620AiP9T)Z$EV(
z5^NLL!?gAurvi_XkY5C43`N3C8xR~B82v>o66z?zYk_^W*d)hkDb_y4_{yeKOSvlF
zYNMe<YJy3>x>ojPf??FNpI7{zbzYhz@REgbD$*2SJ&Y!#smxtPJ051tv8{}HZk2&z
z?LH^jgf!EtSBHDQxV$s<Vg?1`)L-HXQC7sZ_$}$Ec1f%&v%>(oT&#}r^!+h0fqj80
zWX85%`XwPhnI?Ivq1VrQdv3Ahjr=08L(*6{D03&TPUsAK4FGVW>dQ6X22;%Cpp5nl
z(RWI-CyI%(&mGCbg?$v<`i8Y190CmNhx5A|!=1?3U4VIN1U%oi#<DBn=TCgC=C<0!
zH;fzWiyPOX5+g+!Y6GwWBX*2~%66}IaGiH(u2s?ra=+n->aI}~IXA^Rop&*sU;XHz
zG*Ejcnp&c^7~o2apKIUM-k2kO%iFz3B+T2PCAg|?u`1i4sD@VUk_^1xj8T7oBhjO&
zJ1v+eVtjYk)CY7ULF&}#{nMaoVmmKAQtJ<lmYr~tmuTfZRAh(3GA5_)!^RvNIMO(n
zc~lzOpv5)`25Bj%N|FC~b=^_az|D+V{w8_X=LCmt#r+M)1<wv(ASp1f;bFoc=1py!
z+G248kkNg95lGfv<RKk1Vd2Lz6$@AwujvcfPVY0Hs0z`FwHp+=p9X^Ngug9)kBbTu
z6D71}@na~#&@B~KeZalII-<C!ryLumh`(^-r$z;??%z$Gp3$<qKGkYFkiVQpDf40D
zjWcyCl1NdlGEfQY7_caAr+9acJF_e2X2g>JUO<`l=P4pKKZmYw;!o@2t2o^WZevM~
z?J;8EE`gkl5j91dK2riQL<ckJZ|Dj@8z+o1dL@D>w|e*<owo1#Y1(u<>Az^UEz?~-
z)T%s2R2}+k+kPwJbU2DV-xFZkdY*D*EWZVvifv_8j?U{X$L51AE!U6FM$`S>!h$~q
zB^Y~x>f`td=QxHC<|dtikf)z%&j;DDTRqCRF?N{?Y>8Aq9p@M`6ofM4-M$C(itU+S
z`Rs5_O_Aq{>FR60{N6P0)7wZ{5+I>58QUeV+YuDdaR6T534o%%PqZazraZJ^N{an4
zEH1tVAkJEel!R_<D+~734cGtx>x-0MixO?7B8l)4QlYuXS>5MwIu6jNLTEee<oHhA
z;XfEkX4ZY%&ifVuTW$X`DF^yKUVkzYBxC8>WHP*3z4CZy`mh$PD>IQB^f=KFyON8Q
z)%r0QSFu;ibq)M8cDjQ0`}ZJoXCeYJbGzDvyKnMn3+<-4I5ZwYVtx@g3`XWqAQMIm
zC#DC2lsUjb@!<o@S~f!R1tYzbZqdXxE`6Q1a0OA6U)H>+BYqLkD9SEsHmiPc0%}Pe
zWvh17Hxk2NE^2{wzkO2t-jGy&O|!_uLc4LEChT=*7>VKuPLdOzk5%c!bBsu9E4&j^
zW;Ms&=v&;$k#T-!jCoXDY&^+w{=j}kSe@DCRq9v+7CJo*9+pR3?7lV7YLYF4C5r5A
z_yx^2G=0GTI0@##9#mrW=BFJRIp3G7ew?13!-Dl4kF+DIWkpuz(lvS3SMF)dB<)=(
zF9yo2T@=(h!`Xd#JMc_tI-1VP*Mn|t?k{OIYs>LygGwx>W~+ti<a;j(?$6yNUQqXA
zd4V51JHzgC*zQTqSHWY2kH*z9?g>mT{z%B#RDR@_<2Xb7vL-8Re~jxYjt(1M*n@89
zkkvuKunz(+7e_xE{m}gRJUGuK%|~X4G3+ZDWyqt<D&bdO?TEx{hqc@*b|5)1_>6g-
zz=~&}M#Mzo%#MhX_XJ|C!U4>zD6ptwCnBjknxDMrAu-j@{}uzXVy;=G@SDA2kV+W6
z|NgVF5}|bB0&+?LHQ#w@O&vq%OPqDqIBWHDW)jrz5dI)VLkg3tS(T=D&V2eDb(|yA
zhqxBpFW=KKPlJ8TanE({5JL7oKB7z|_Nb}1=mraST;+S-SZ#GOA>v!M7H;IxO)Foh
zx0@=Q08?fvl&gkORlSZ-YX<!F5k;AL^8Fi431f_=hNXLI`|R0W)VIWYJC!o`ajs%x
z{1Bx9p4>G@1wd)qarTSAGEo=3w?nibM(=P|Sz=$IO;q?E!ZAB|oR;lnBan3Cfq@1)
zhPR(R{9#@<i$y4XN|$Fa9cDLn$*v_)J7>XSQi38=VM}nk^64Czx`X|LH$nv7RTdl-
zp)X#pDBr+ZVwe^u+OuuibHp<{{3+Xn*&ai#&dnj;zjT&=Y@2NTNl}njjWueKD39&k
z!yX4q@GBM4Bfw$Cif6*BiNcvm+c0Ko?1>7)<{VFY-|Md(L1U)|@gvHkn(1yJ;d2D2
z79u*UG6-1P!)>mt^kvSUA#uus5Xql_03fveZ}q)Z&5i)szztvnS}}2*iImy>;Sa^Y
zj_1!AeO&qz?<21PLY|wM7dHBLssEpeaERPZxag$Hz|rFfqBNn}u}aUS61A5r@X1M^
z>eGM*`)}<EHf}N{a!U;36Hz$5A6}l=@DgmE&M-u5*HcOI5uyg<YZl+H_R?EaWa&<a
z=8uP4wI1hPNAz$zcmwWIg>T>V;rY<I6&Ln>fp2Dy6(71H%LEMhVa(W=qd9taRZr7&
z+S{~})Qy;$eZmYZKZ9lJ^nS4wAay?J2A5KjSkgb2sLo7xRbLD^_nnVJrt?<*@QBc#
zF{($;95~F0E0g~aJJcl`<ji2wN@&-S-enQ6L+*>r+Bo#L8cEhdo_70>nfD4H!3T%@
zxX06ukPL7_QLXA?fEmur2ep#$G{D)G>inyrpPihQq25)GCUF_xZqnsrMWe+4FzLo+
z{pJKm>vOf1`k(ZSvr@g}oSF+#Avyk+Ckg=jpkiNh?Gl`;KnIe177NjZX^!wF$PV#;
zidnN=9(0d(n=_z&#>a3<u@-hfT(TXgP&Vu90S-`b%xQPkcyN2;SegDapv_uB#IL3)
z_!an(O*RAo!Y8MDE;8&!Q@V{fAQ#5@i@;|&(P$ViXz84OBJPMF!%*#{w&11$Hnl`^
z-auju=;Ii6UuUWc;wj(UVQ2!-hS%bN`_GSEVTcF>lGl{MCE0(|ronf#>ed<|Zq%0J
zWp<k`Lua5E7IH<>sS%i4$45CI0#!>adM-Pi`~X$6eR<@1{PbpY>f5jrQkS)?31$QX
zM%&8b7lEyrmFcDqS8RB}=$l$+_8QCXaiHoXQ?RU{z#9n(BDL+*b6G6-Kx*Z&EBLW<
zFmyQp9;0{<f>NQ{Y@D6a7OBj16j?4IAXeA>(L{75<S}8l=Gi#C3jHAnPWaQee#Sz#
z&`KL}sk`cxl)JWkTh(Y%U9!g1k{lbTgc*KqBv7)#0&CyZ>u4`^9sBtBdysrjJCppw
z8A}OjVFm3MtE<+kAsC@c9Ee~;&h$l%u%lUGHh~|W{qY(a6LnOgI#1f?B!<l_)jTe<
z)#>9XlTy=vl*;-@=f<pRwXlIR!7*DKGv80N&hYxUO8nY{>Rd7W9nK(ke%AL)sGo~6
zlgf^Z4pf^i?(2QFc)An(^Hl+B)l1s;A=*c`n0aX55k8g{LC{UjfgNQVa+v)fqBzG=
zs{4Mf7R?y}ncaqO04SgHdthEIekb;FSkt@s6$ViN5>gpN-_lOQgI5I30(XOG)B?v!
zd_h*jhZVF$BTu*}9H$U<bdD!E?4QRzMzgM8#D~8XO&E$gd2LHs%KS2v{G6%nLK(1K
z!$J$&iU(c_PZxSJ8QTpkMHhI-E2ap?#S|MEne`KqlU}*TI%e@cpmGlRYH~=8{@*lO
zI_t+@LsCP3kN|nq);j>nI70<Rfj+q)@EpWl23W|^6&W`#v;B3=U;Xpf*!ZtLFD|90
za8#Gy?ygc}eXTik*GTU+!DYY}E%3JlDnZbHf+f@6n8f`(z~?{m`t7=ZRxHPtZD*_E
z?~5IQmxY#`2-X#eQ0Gv3Y;})zlb=lk?Ay$bH|(zQ@2uZw=qZ?hmUMU*);fY7iYRm;
zX%6QpHt8M+Buto3xNNWZ_TFOWd!=PjM)@i(Y{L{u0$t8?<T?f*iU%#uTT_v7+fXMM
z4YZy6gM<85LzpMLw(?8GKw(tRv=>zqrLhm&!FP8iW)6?`&2;$k{JK%gq;?$=s9Eck
zVc(`g=4>8RhtDpX1OJ@q<`3N*UuEG*CH-n^d!LzR#t4Sh;{=}WumN2>z6Zz)Y<;Wr
z6r%~ej)uL)QT;}S1YdJ9*qr*-2zgaIP5dQ2=xfe{hI9i+jVgq&(XgZPl<<Dj?bbz{
z95#DtFBXihsGY=F7W+B?*~%;@6Gj<xkJ%Xx6}^Zls+DA4hQ1oV2;q;y(=TLFADhdh
zP2lJt`_7}r(}Q>_#9^X=3V9$Y;sz(XGT$5D@CxNQDOkh3;~H0_efEm${L1dB&l1a$
zTxmKNHJ;6E(>x+uC5|^tZ)Bix&RnZxTDZR3o6L>x{uf>GD8m`z#k(x<6g*uAAe}hg
zL%}<>TOI~<q2aHZ6+}P!c~%J~YBx2GUjL%Hl1qJyhVRL~WURGmRj*QIY`N1M1qL3c
z=uz}_3KA6vwn14MZ)!wacFJZPqWK<Gk`lRm)A#&3rP;RHpm(ttepYvYfZa%$X!QQP
z#%aPoosL;S&UzKuB^DxoFHxv;;v&Y0e8z-fABwSqv=;N9?0hV&aDs%sN`AyVA0oG5
zVHJs?SY!h1d(<n$Vpm38jvyRc2>Qf($?f9`4iyXdA2R!7jN*c+=Ht>cU`K<{rs_2V
zz||0U9gO~lq(Y(u0W<yJ>7t9WZ3JWrisC<*7?@qgAM9{BA1HqNMbLm;o;B;4tJ&B+
zJD+W;D=KpYgL3nOGG7LiojFWRtipO_e;k7UFd%?L(eE`b&3(EtoML98iq8}cQ9PJo
ztfnv9cG|rCZD8*y0q#H7n0lnKbN3WsHLR@DJRY371m&7{Hn-gwOxT0g91XH>A+7;M
z59GUOzfW-&v-vG~zsfF(n<}J8i%*Y;PcX~NoLK{pCwZS9Twkf%8>&0;0Pil_`=MZf
zum)3b{LvP*$I?+bx3N>dKmTaV)s3%S;r&aV+~i32PiyQY9A#!e*)?Syb5{}RvcgeR
zxajk0Q1PQ)x|&;>y)`e4pTA&^qm3g`1xp9AVbd1}BUFfLTVw`1B`kk-EWoKOr*%uF
zKz{xP^FX;zofm<{OCs$0;nwU8XU$b|&-_)`d}}(nh=RFs`lBv5edyAxQ-3?2qv720
z(A?kBT%^0mhmDbjznSI1+ph>~wezNFHyziT`~B%-fdEY#oc>PB{fp>0&JU^;bU_{T
z$Bs?>L(gyZNW){nbW%tvY+p}s$p<SlE3CH?5NCA`<N~Gff@<_N1<YKxZ}{T!ILRe%
ziO=7l?j!vYgrV!6dR!5-z!LcK#SLZoI`k&CeYfwcRhzTS+lIFL!a{C#jw&zWp7PRC
zmROUeG9H4*tH9^(7evPvO_sRi7W@*>*58lZ|3h4J0A&q)gn>qSaS{4zXiiA$a%q+H
zmWhnFpzhTA!dC(`)IfGj?iIzU4-!|m^*URpZSH**D-9&Y*nL@oMED06E#Bb`Gd|d@
znD75mm-RKtLW|vP$65GIG&xVYp*zD#arbRn6BF>n(S?IRVH}W;{Db`IkD|1RLVZnY
zfMfUjk@|mh=m;Pp{x5J<4NeuUa(~#9>hT#{i&{Z_ecE%AWQl@QG6IzRbCXKr-!lXW
zxPHpcABt~6Pj-PMY~0XA7%*~FIe|_I-hpBVT!!v4&;NDBU;Xjd`1sG94fM_R(MGK5
z?^Ewe+#!7TYIK1xWt$)u_M1^h2$w6)Na@~hr~i<v^S@@WA@R4l#xcs@i@l34Kl>BF
z;lBqqQi}Xb9D6SCHv*&(`T202mt2Xp^|djAl_~-Ca{1yPFGkjc8KRby;sg53ob)$G
zoT5=XR9<o)wH?Kbk6sVbN=m2?9tWlx`*3TuigtQQqhNm<&_V)VjNkx*HQ7$x1wMUA
zCjfQm|18%WOJNcq)TF&9{VnDDo53q1-8Mj7U8xX;z^<>-q81fUOO=96<-{{Y8J=s1
z^VO2dWl*3-RcL<~f(Ts{Bm5!dd(6D~OCNSfsO+s?msS}v7WRWD>v$Sn>Wiq2U<x!M
z^Ozx9B)CT`_1Nou$2Z<-7nx`dqfF+Ry8?VBnZDp9nG&vWMT%Ts^?MlIPXBN6uX~G~
ztLr0p>ucU=2sU-n%3JXEQW6CwHQrs0LaR9j>Ro)O14lHvCHWLPSh1nk8>>KNiC7(+
zt#kgZZyt@q#1CNK-@oi_*NCu?QwSZ+C<@qt9)R<}A?gRyXZT7ld1!}{kk+SB7nfiT
zUzxss{qPF7YsDCE!sZQo?*t>p&t4+JaYof-Nr@QNp=NZ|XrW&vjsulhVs(ed@kVBV
z#69jbBED(Qg!`5x=&G-6tIhWScG`f7)_5Hx$|qgcc#f_S%@tYatR#xZJdF^9L)@0`
zFIOc<Xy1p6UBC5w`DQshTBWuwy>l=;SbP|#QZi-b?5~Cou}eYi#Tak6iR>QT%oV>E
zz|tVz*Tx^88W*5iG8^%=(oC(mpw#<tA|1WI&W03l4SE7Cov%pzn1`Nua{=oIbO9sV
zC)G}RJ6VahHg)i}?Y{`}UJCt-phhnekl;g>7%?tjY8*FL%;ZCv#!NNLliT-T+f8T<
z4(CKPI2%whv?KT@wgDkbi|8?Zym3#Dl`Lp|)pF)*wN{B`k6U!|ux#J7Mql+c#!re0
z=Qd`}tSWI2H_ooMl&`ByR&_9iDe5p*Jq?h??Lb@!8pR*jqzqWxalji{cLcvQ81MHw
z+bCMWQW$^989c*HZrZn|RV;BYE-mPi!5E^NlPY()-1adI_ON5i^OxO{aYmnUQW!r%
zMQ&6rBk<*J@4z@MM>BD~6BlgJpn(*R#CNChDF>_vVr+m`DKBqxo@!@TQ7j+pZmHS@
zo6N|q{482gxue{nc>W=M8)2t>|1lP2jV5Bzke~t+<A0c^N5K{OgNm__MFiTFC4CY*
z8grfO0^VzeYDXc)+@2bRXw!@K{}1-wJF2PeTN@6?f(i(NbOoh~G(n0ACMwcJK<O<<
z>BNBa0D*YuML<A6KtMW3iAagG<j@5KL`tM3BuE!Xr~yL!_V14GecyY$-|x0@@3{BA
zpZ`F{9y{54&$ZTEbI&=S=Xq}#bbP`<(vfFhp2kk44}WJj%*6GX?n9){XXl2^4#RT@
z!>x9_b(o;D!vdEaPsF=q)my)iefL<%o&;VbOnJE_Ml$%9?AWRsYCeO{BUCkN8p6KU
zOfJ1mqxPM)Y_2z{V@g-NK7r`Nl$C`rBeh0DWRa?0)NRWaRS4%3%_G&C)UGKXK`=D(
zUwlzb<ujYIo)e7xTyi0~oM_r}4a1M@_D8MkF9ytqQ3mxW^<fe*`ScqU_oM>dr|BLg
z7ME|k4;M%m7Ax1Cvb8)DW`E&r(r43t`_^hiXuEm!sQMWaT(8f<C*?tX^9%j+o0s2~
z->ZCbH(bVr>B)>pLR(YdWji3V3>^)x{kplTPr_zYBD5ykGNalDnV|@~kqcpW8(73&
zT7&1m3XJ5>)-!Xr6iqB=3T44dmH_cUpF{ANawLnjH9q!Ik`}iy&(txH59o2@4Jjf@
z!EZwW`&S^vUqBR<nNqC(JglPL<qhjUi_BPYiUs|L|3pPDugyv&>86W8iR;pt_pj6>
z@<gvdW#8Yr!us0p<31j0K>#3tNvd{#;~q&o<QW0|71T2H6I{hw$>_N1Pmq<L5GpmA
zAp2iG@>kT(FPi|+reE{qKW6P5ll`JtmccE4ndPm5BU{2PG1lk1f9@L_{$Ez|J@3(0
zURzDwNmq}P7TIgy=-2>A*+?dwfjitw1QE(WdO4yB%DH4U7^f~(Wn-mn@9$ga`h3No
z^~ow{4*(3Qo?}gZG^+9Fql^}46BUNrFs@BJfd2sH+co7*@OBDwG&AS7q7nL*l>VyR
zofe2;a<|oIcX@GFIgiKniZKN(;m3>>>)XPUqvi#%rlaCbeg^=Y^_g;2Wbi_pFpFHq
z8_m4s9C=DHI=?4CMqn+aq@g*rYBHFZDOr~+rM{%n<OJq7DgMNyUyhU>s;aY=)xSqI
zPgG*NBP(&9b8T&`zp7*e{@B^Ua;~X1(_lR9fkS-Kz-IE(4=ddrI`m_nB=l*-tPN&_
z4y{rmItZK&^FL>PDp{+dO*|-t^G&Agzy=&tll@lkiQT!wH|lKU{=h+<VP#nQ2i)fI
zTyM`J+sSb^@v}|A$;&!vg0C;vgv=2y7bY+%>Gk_t{Q#xxxvkZY##xlNKmdZ8a<s<t
zVa9R9>F)NWf)sLGmIYm{7~!s(HZY@q&Wy9U9&>8xx1<{%4=}+~`SgO<i>mF}S07Lg
zcQVMrJc!LsMO(sGySlX!2h&-V^pl>P+Glx^{*1L#Xs)<DC1gR;>+v`+MjX)|QVLMc
zJsY=D-XcV_N?tek>@mh(-8&HHHx&MIvb|BYPmh)u|6TPrKjqwn9BztVQH5p|EmD<Z
znX<UElXgRMDy8UCn!ZmOfClCij5)7*<JAazhJCB&rmLi!vm{U?7V;Qa-`WtC#!iOl
z%C|~~I1-~9{lchM+f;LeM)1}c01}enF6-0K@vfh>OzD9f&GMwBs8)NqR2LwaZO#nj
zksh`OhJ&TxsE&~9RFyokpqilWw)6_S2n<!-FEjT3mTU!Chj3kd>hzbBnjax5YXnpx
zfjO$JQ9GLJVRO^_(c&~h??kt5a|13@Cr;+0%xp~65iR4eFg~f;wN7tdh%hf%9H3>W
zg+}=AJ5Dodl~3lHj8&|u$MMM&>B#X-I`NC6z!tM7u?oAB{=G*uv!@>5sX)hm2~_?9
zHZ2GRef!lCEU<)0yP-3JgwC#7n-BZe;vbE9Jc%1zkdbG8DZ=x3T0}u-(d=4^THHc_
z{Jno8m|#gJ1b|;0qhO2tsW;SY!7#P$>N?!e>bF$o7t6(x^)*_VzDXGw>WSH#{1@K8
zaK%GwC7*pqUVT(GSYaccy^}w2_X^5!I%^Pff?)W*t~72?x4x@G>m%RXt)uTlvx|Hi
zQE&C#Gp!9q_Z5Rtg;k7IA`F`6O<JvrqTv7B=Z)~`99tOP3cB{a@>|oeoPgO~qc;r}
zCR~4fWm>gm{|Rw`K2A$Nu$ZuNV{p|2dN<sWiEmvqi>rglY=G8VaRGOs%Dpl%IIEq$
z)-uESmDtnA8*T5yE;|n)9S^h+>?C|nCZpP?cmU|+Ta!_OO@5HcbCf|prD%#wu7Of5
zE%p8SZL&h>jks>;hNG_1m*D%uKVb-le#T9#m`#w?`Vb`%n9j@0TCMyK6xqR};*!1v
z;!2}K|1!#=M)Tq;;!jYA#J|EE{vihY$qxSn+R%TN>0fg(ko`}{3&>wl7ypnv{I}l@
z5OyMR1^^|;5&hHo{NIa>pau<^J@HX8_8@jIPupO7#4zLTbP%Ii4zgu!#PqWb{|WZw
zZ}HQA%8mXvv)ixR0kYQr04@3-Fv))n);}7oWAfMCxgCvKZ8boP#_zNO9xyR~eLnPm
zRZQxDYU3rX?}Gp;1!jOzj_Z~X(_g^P?~qQEn?Ui+@JPI%K^lQ2@9k3h<XT3miZ!&=
z^FnW#TI}6FJ)b8fGmE1Bg~p@;r(A#*f9+-hszc~FZZzZyLZkC(zk6m#Vv$?g)voUZ
zZ(XmQc>dd%qfG95J?G6qn~8jsZh&ordLAo|;O!8S7?G7HwqrJ@h;C@|>V3*`5WtMl
zk+hU}%F42SDJizQ>Bv*I?*1(Z{bC4E+<$W#Ab)t~34CAar)Sjb9ovAd!x`x9arTT4
zA9le=Ao2FSlbkI{!nTn1&|7{jIGQtT=3}95kwV<Z&P7+gSDSOF!tO{AhHZm#^|KZI
z2=%BihwOG&^RPi-h+GZ6_oTxnc<^+%&yGz1g5bFKY(T~PMTpf7gm#AxnD5qkpa2H~
z+TqDA07y{2ow&sW@Nk@lFeGXhkOQ{RA0YAmy@Ur|E&Ih-S}Cqobe_oNvKcZMSRX-d
z(e4o9t{zFKA#1SOiH|rB*EUr83U)T0!V~Vsc{~lh_O2r8(;>O++7#DJ?jJzf3SoYW
zmYUaH&_hmoON)h8FNXvd<rEbYm)Wk3%k_wbU)4BT{n0+<(tY!V-a|=%JqsFzTRZ{~
zXPiUP4TLEXv>?W_8_o=-ZeZ9)G(VL4nc3;1iTKgj2lFlaO*mWB;D&mW4b7j*+mY#>
zgm$};7UxMm6z&kU9j?799WO_af8XQw1Uqtp6Z^o`VRM30<6TFP*5!l@Kn-&jeBley
z;O&&F%6<nySx9PB=1EDKeh|W3FS+$%s8-B9bfQ$;d+pb_BOqbGDt|Fm;?K5p?c8WW
zuK>7&Z|1p#b-A_O<;*>>2Ls5m>sl{V2JJke-Y+YLrBR`a%e!M><@)E%U!DnCH0Z+h
zAIrtQz58n5w95L5jcRwL*DMK}O9hy!@7IvBTCBSQ@BO|cP&C>h&}0kFKA+iLGI;{f
z^ZO|B{I}!_F@0+TDtJ8cuXX|=roWg`F7~FZM_LszNO>nQ!|*MfK54+voaPz*UK_*`
zpfIm-)U*DE!=>Xu)$?DPtfo@eRV7Cj!c9c-8wHjA<?|m{)@TA@#a|Y+_CL7@(|^m8
ze$Lzd4}FaPvG)IU<Nqy}#jjxr@z{>nyxNm6ZpLSCksp6Y%;nAHQa%vVXC_UM!T;R4
zv((7+|675U8u@!PHUFKCGnxHu9Ju}l#bMeZc0L^80sa7E@b7~J9|3p=|B)iOQU5>I
z|9z(N@0AVtd*5sO-=w112LI>dzndQQ|BEgDKkFE`(+`lvi1GjQHj7`={nvE=*DR4=
z)BWs!h1dV(-T(b!;{Tdy{rArKHQj$r_rDjk{JrD9f|>u-Kl)$QGx{r-`4!CkJ&*p6
zkN@)Szr6e3^WuN+_%HAN%e!m-o0iS5*!@@R{%>dF`QHGTU%||;VCEkOEr0LxuVCii
z<dy%UvN*rI`+vQ6zs`ORktWZ2(qo&=ZT8ORt;5T^Cb-7>2yw>*_Dg9K_FlKlyd(I*
zhtmDExQQUrd?{Aw74jC<F(NA3Id)QSWZUD^)K?2T-3t=R3CqDZFN*)U7ZMU=R<f!F
zkizGZnGd6LE&xQQ3g&T0;lmvIL+Z+oIqiBrt)y$;M(DTDZoN@Kogw;7F<$)1Q_`JB
z)4gmzei-=j=|$ePSr>9PB+xL+VRlzl`mo34`oTK`_dV=Zfk)?ZFhG5IV)X~;kaZt@
zxL*rUUY|v^si6QUHMksrQddG~ok&4&169P9d8dXN6@k<s_rtpE_^IdiJ)*XVt3l;)
z`YawP11?XR2c@#@leKI=PU=Y+otQTaLvc03g2^mPSRoqmKGibrPk(cNA>p{rMogS|
zma$Gu-3O<J_jhYDirSQ`Q&o?&Il`{>=;cCoO|Z0v5Iq7`x9&cBCXwKIP*|{sYw&*$
zZl{(!S3*TKmPGmxIS?zR!|O=>Qn~aZ&7Erw2L?Syqc_qzBti6nTf3DWY(0{(6+#<*
z9d_cOVi_~iWv){85_8wYv~T7sK(&s*PTv;6sVG4O5zbv;@xzq0LRwrEt%mG77@Plr
zDwkRg^H+sD&K1)W=>OX3_CTZBl>4L&MO!qe9mlKg7(iJkmD6gv?L@yYgosG3u?nOy
z8L=G{?Ji47Fx--MZays*dK})X;|KyNs;%^nGzjgw9V+>ZuCJ)B)E~_CkHQ`3JEc48
zF7SojkZ_o4n%#NquuL5^n#(mkv#^-ZM%b|CD{Okf`ZC6f!{PVqZQsbvq4=E!+M@tc
z9aRi<M%_mJiqi&rv=}8~D$os=F-NMJyw6d}zL9JL3e4PFnLUa<*^E2Pq-#0?Vn$$b
ziYewfS{5KMkfLgp`MQWE_D#K&v;2LNLowB|$LT3Q5)X>4KDdAsq~><Be+|U9>Ca{Z
z#E~b)s?e)zZ!naV!LmIF{6%k^4?f3;7<zTb%S6>n)#T2`LqD1oJs;oz99+O__X(<w
z5f1<y{#Nd0^{y{A&foW_w+6XI&&9!LNzB!$w^!@-(c2$I0EM1#zaJn*WUx8o&<aJ3
zrFoFMqh~0pT^{FX8r@J1qyc5N6?~ooe;oockhlCW{9qpdyuCjj5j9c8scCZN%9F&I
zn}!OrhB4q9DWu)DX>&4fM?<rGW&cchR=-BuL{0;WbQ0}o$tOO=yF=ys*7Q|G*zx6+
zKKG5TS-l6biM0vD5*+Oyvnt(u#gVUfS&;yc2z){9$!F)L{s7$p40Rs&1C*C*3MfOG
z<<cKFKzV669g0v3DVW!mI#zC9M$4pDD7^xAv*x;Hn|m7cJh|gJU{kt_jzc<9Yr4!Q
za{t_`D_?bP7iQfyJlW$d#!#f%;v*fJM~~b%+B_-)-%a0071xyF>}m3soGW(SUx+~h
zNLALX0J0%x>+H$I1m$Dz#`0BV+@EyZy<?{z2lLN()(g;wh@Il&O0m-QYC;77tyzC;
zKyppL5IxxrUK6K107u~kl}mIERv)YbizC{#X%K*(Q%Bu)w$fbp3wW5VWCFNhVoPe9
zR+D|b(_A!v7c<h_{AUlqy?2vpPgX|9gMaIn?=Tcv!BLtDsZ&OF2*q{thFBJz<VASA
z&u4~SNA<BsQuopz$Q<@!<O5nA^*AI(nEy%h?b%mI7IGY$+qLEk7>3Nq)-c-(&mC9U
z+Boa$<dHG=#HJulownZvjK`#~xYRdWYllBT2EuTkmsYReG`eoiXX5*&Hv_BoLja1_
zpPZ4m6(**(QfS5OeM)y`>RNf>r1<n&V$tLz(M{pb!igUshtTo)i;4Fx4!Y4`vwkhJ
zFaRDA@;NYqn5b@9K%}Yt*?7O~k^P&M-Fg)csP3uS_b0ehF1*i89eLXAwNS!Ez1qbK
z7NVAP{s8e}Vzy<EbY*t8Rq~n)S3Mf5l9gy^ND_6Om_AzY)acx*(Otyj-!@?UhG}xP
zFLB3=(T6I4Oj~kdi?|RS7Z<@;cfkqgF4ir?ZN))WPvA&5M%}k%!2@asEdpHwf|U>W
z#Jaj+e3h9q&NtPYw`ap_Wvdc{0$^+%ArfO|*Nq)a(<UZVLC@nqj*(vOhccZ%f&P>m
zJ3-H3rtWq@IW)=H)*4h_VDby>>NPP44WaYPeirokD8;agMio!rgzrV4@`AmmrBm&@
z0O$y3-FZkC7+|#vH*eRU2?o!$t2U;``wFvZtFJ$a_x4M+n(W^>2!ECIt#6V~Ys$cF
z>t+5vl;X?F7@up|nRpa1G5r7)2=u=8>@7BKAD;Yq2}4Lie_1iuleD_oK^{vQj8wPq
z9DK#v3W^CWZf*t|;p*$4%*ExG>;a|iJ@~`mA*c+C?rfK}e|ltpuxKG|(vm9q{kkvN
znk6`p%?o;b+Rz;<=6fS16<&B^qfz9z(`CMEnp3ChkGoy&8IAy&`2b|e&K&TwxciOM
zy!XLhd`Os^I6+kIR-OEG$md-4O7*ME(SqcJB8d#&w+024b=cP_`Y3;{C&_9S$)9_?
zBagbr*cv`|l|{K2M(qM*tj^Kh9n*GfkTwKkI&xi?XSdP)KJP$na=)|93(o`WQ_s)%
z{t&@o#yMA1jmP(zV6>Z%DCOQxS|nVOd_1y##WA0t%j0Cd@LI_&1#CVaAI5#*{qxLu
ztQ2D90hPUtIN_|Dwhj=}et)8_3Tq-4#}%|Y%5D+dnCv}euE&xJ9=`6nWy4{@72!*)
zHhnPKojKhyirO5S^mQUEr?209UH<@h+@3=G9NgS`S{$<|mtwEdP&B@*Vka}|RMQ+G
zcOm6toh`L7@?n_!iI1Y8SIz#3L&K?uTtMEZx~EWoTi2pOVnrz);c)uN;vSOjI?}Ne
z@t8TO(*EjH$5`4X%lK^i+=ke9(22=-Kk#*0dD&N2-0|;76gk^sB`_-TTT{cRU>C35
z3Cif~k;EI;5q0`qjndWO$Ih~ikN3lkwb&+M0B4sTJ($=%Hr}@J+<H0XU<AsTUMbkX
zP2wmCL_ivD@GYR5UT(-xuvth@*DR*141f1%)}O7^*L~lDjk+H>Zw>QH989h6s_J`k
z=eWBn_$()rs0dEAS(Hk&raFg(oREATav3%uA0a^{V2a!XM$s{AaDku`1ram4>$l^-
z-5SL7rvRFy^pcW17~x`iSMGcCsx{0T?Phm{1}&P#&@b-qVEn~rKlO5Fi!`Rj4W9Eq
z6`NpN^A)4l*_Xyyx)Y1*oAMZ%u=02t9RK)b5g*jwBYPe_j~152_ieE-&H&H2&%X^l
z*mea7!!G^2KKf~p;L_#XlSuA>ij%x%ABJR#nz?|j^7@-Q2OrL@Nh;&bMlXI}xCBn<
zQd(HO$iHiT<(~{%B263c@r%x*sP$1`CGu7FN?+AT#fBLTZ5y0D@3pfpSY97$;ckyP
zW&Sa>wfAn}`Hw9tm7jlr@KZlPjk*2huc_Mf@MfUZ_z@4V-RZ|JZvE#8_9jt_i{uiE
zmv<>|mB#Ad-;cW>t}o;G?3(EF*vcIUCotyLdVqwA4UV2<Km#^ik--?qcVP%4M)?~G
z;~%4pIuX)?I^!G9Y}zuU470wLN%#S}{~;)VBtgEQrxULDn)BMtlNHN`<E{N$N+rUs
z%BL65!HFHXEuJqpin%<n7aIR`8R*N9n}-;67@7TRRvUqRX#3EwefC)VVJAA;U8FgU
z&5Y37$%^(AUE{3Ub}^~pu;^CY@P^tT!G{%eLtoM~q+RGN#UAkkWJAG(`m}g=WTv6f
zg1<TB42fIJIz>KIA87Eh$X_UTs!ujvXre(06-3qI#D(H{1$Ub=+2K89alyc8bn|}7
z_YlBky{QKtI<x_`Cz4oAOJ{^mm@`fR%lca^g-v*8)G$UMv3**QgbvH?0C0<OK2qvh
z@a-1u14ZKf8eIV%#ZT`BEI)nx$jy;(^~DMF85-|_3z4o!v*bifEF7435+taM>SD`8
zE<|)|$<n*FA^xNIje@<Woz0gcp0A0)jWe%9A)-SoyJeRq*F$1I`R|&fVdxDu=w11a
z1I#~vgS7Nn?bTpNr(vkSDwt=)x;^9q&>x>zo6Y+mw_RX&VpZ;OvP~4GIb1UPXju-w
z38lBMQ<s7>!mE2zVY9ETuu)WKciO(;b9b2v!8XSAV8yXEQ0(2`Bt`)xF`bEUkY~>5
zYwHvkej|(?fS*!zz*TKR5A4BX2f{I^McE?MOcj)&Q}EMGVwDaHuyXn&S*}@hxVo}%
ztU9EM25l7x-V)ft%3&K)?<R{xSmbC~%p1QLE>92fS6HzXsyy&pVjOFZYo)ZmA;JRL
zBPx7t%~m=I(usEuOukHtm3z22#A<k}NxNUa^yb~M4aC_3>G4&Dx8d41RMA#bO`{p3
zWpFTC*yLZBc{wN^h^t{ti27jxSO76R)3R0+oCkS^B3*8z7>s_4;VEMEzgqN3=#Gb0
z!{kOe=)S$6?u8=&BJ4EEO*%1}Seu@JG?<C&*Y679OC(L_l-ZL|t0r|3a*WSBnx0B=
zteohJ(+y09gc(81rfWsY#@swzSX6(InxW=aXP8DnPT|rOLvq!E<9=Eb+F$AVKR~zP
z(tyAapyRcniEIE|%(34>hdZotqRL$Cct?Css4J8rBByAOuA~*C7}<-i4mAzF^@$%}
zyyf6y=4a-w005*6f)E%8M#7%J!7Q#TLFConP@JyO3`N0Xh1MAUtnW}!jsmw*<hsr3
zUTk}EY!d@3&(MN;>OwvKp+CQ1_v`-PPXqeXve<JXGEQA3O~VXmUX6Yk1xq(@akVjZ
zWe}1gRlsH6FPqh?=k!?9m7aN>FCEKot2d6%*0&7rc+;pwAw#*Wt909<36-3XVdePg
zj1LJj8lGHmUtN1O2aR^=1b&NitmL|H{;>KVAW3K1qPF6Dz%`qq`X0XjXXhN|Qt74k
zf8mTbtdPI1Z*Vsmy3gVYrAep%03}O%#e#nQ#4lTXi~Th({?8ux|7%;YvoG`B$jSLa
z;LDtobC(@|b0?Z$Gst&QUxEL`X~n=HBZML2HU_eD1`7mn4K8m^2}w~UvS$vBRwi1y
z$Ca|>?~Zu5uc_Cg<oG70r|U~)zM4vcI6eP-3+ATn&aCsn#%?GVHjd?%5(|19d74Pm
z+Pb`t%)L)otr5giL+PDH$skF-`38>(7(*ADi2JSy<VLpcnMy`5cxbo|UcxE7x+}Sb
z*AhdSIS5HFaS3G2e5xw%Mme6N>%OKRk>GlTLwl3s{?lV(42c~xx(7p3*Gd8A-%s<J
z+{AC(%x@UvMBe$_ELpX9L$1L4+Nf`joivM@_4PG}6sPI6HZc!5B|=tVaA{Mdg@bvo
zFo+?6z<1cGQ+#vEDi`0;GPRn((nBbc+=%k|TTHAG$p%st9XI@Xn)55fG%u@2>&~-o
z3JFo>-}^wrbFO~zg`P!3k-3TO@&4xZL4bw}bSh^3i}aCG(E|4T4+g{d4Bm4TwR4;Z
za@(0c{1S+oQA_39>}kE}2WZnU>cM;YSlJPEPY-93yg(0Sn0&l}oUH<h3hZPwpggOl
zpgqNIL*4v!k3^dU)gq_83>pI-A>aAQ4Vi_AE$%10#Dc92`4PR-$g7lIV#f`HA*Wp;
z?Nx>{KU+u1r%%mY1$~bnU*^{1zI$roebXlWI{F8wuqm*zu68)CjPJwtp$!j?CV(p-
z>K~?PB2&ngz7Q50#DNBex>H_gIYqO-Ggdh2yZT^*tQM4kc7O0~uXpZyUWSN?bJXKo
zCzI~L(He&-NHeSig*RF~tGx2gSf2%;G&M`MRM&L8vQc2aBYVe0*E#&^nS=gL8=bsp
zqa!Z^?5{)K=gt7Y%JYqTA~z%n#OEgIGuxYMLD}3?|29A{K<E;Nrf|A!79}`c5=R#d
zDqGQJaK?Ru#muOHHAKbo-b^lF?@3&h6nYfmomh^vNVlK`QKp-3lda2x{We?E%$HQq
zvnz?GET^Zy+s+u6P7Z9cIX^w7{Q}#WQ>;-}AMUym3xH4<x5vC&;NdjAilLvyJzPa`
zArkLXP*KVTLFOM0BvOke`hxsDt8;rJmE>0IZrevZc*-Ps?o<oh^ZZ=%!?sCs>JdQr
ztc;&I*Nx_MU^~VTMM_b5s>y69i{>22A8CP2Rz>u8RtA8uYH7uAp9jCiHz&~3UadQ?
z9nV&6(uic{NX9w3M*WEVngp%H<8+o^J)pPVUEKXfeJC%d)l5Ztt_#m1)25o?V9+i7
zt;qG#hMsW{dK0#Vd->4AUbR+WWxrJU2dK!pe_YtfOe>MxFKstk*RR;NGKd|L<NU&q
zm&Ny_k`f$kUL0|{ZvUiAbiUi@7HtP2gb0Juf@ht+ARqls5Mqo{MvICD!;y|CbW6ps
zo2Rx9@54-j%z8}0adH3gP4`n`tu?dp*48xdGUVN;_KjCt*N`5C<14BYRXAVAZ%eax
z>f5K>^im8tGt7<!3|P;ry*lp11-JbHs{9?h9<;wqZb*R@0bw6%<<r5PSlSTXv0_E|
znEU#A$VD3J-e=y<peG)pRKHTCm8^^Kw3?I-5dyz4@^B3!-J)w6=6$ZxiI&wJWkeIE
z9GHKgR!khvtX__WBlTyqrX>%|9a=>)+58B~t~;0PPwDfkp80tG<x5$O_*7y#8JO=!
z^!F!?o(E*=R|0DAKR|u^f$|s^bS^^-k#0JIbA+-kr&&csBX5>2KS1~jER%hsOk$jI
zHM;6}O_q7ZrsI3J)S{(B&6+o@sz;E5rHGvsL#lI}81w{<*uEp6*|HwJ?*3RmsHQmU
z)jX%3K`(B?5p>_->xqq>$K^=}+6G&G^vAfJIa-Qv;3@sG9j+^;Vq53%qJ?xLN^YDw
z#I@A5F?M;!Dk8h?o8M0Tm;R?a7&-e$v-%T`7UN;Refr~;w#+Rv(S>+n@}~_%O&ftL
zy;C)jY{8IRxw0(q&`!r)bohDT>$I{P&eBIN{E=_o_D)mkMKfYBwCW8dC!55kH*7ni
zlpSP5(GIn*i@oVYmao<F#a*7W5BqegSJY)RwCe1=@$SsY{MV+|lC`hR8OMN^cXUQY
z!S|`u?L;78?fU^@UITxlU8h88(az6QlMGK!L3Q&q{5`#0VL>9gJ4$I97ehtfmh_rA
zO$d5}o?1|CJCUOYnr;H?fG#TsR+ScWI*bJ5FZc>oZ=^@2&E9r-0&WkS7mVp2DGcoq
zT4IW3b!XCHZAQ~2u6B~FM=;coTH<jWNjY+u!F7CeO~`a;AV(}I<$?320dA7YI}x5?
z>h#+^*6>`2tq*ix57<C{vt4(dLl(XSFUTKCd)S#5XQ9=CQJTnr3PgTTZp41ySGood
z_Gm5Dt21u}kY0^OKdIafR=1RHX<B(&V{rD)T+z6XE#{S(%!eoE%Gcurwoq)0vv!AN
z6?87&!*Zd#w={IO`3(n7gR`wNyD%ebez6GrfQ#*@`kX~y^%D=6#b#j0<^8=a3<{8h
z9;A5CaV0^)SE)y4s$!_Z-62d=ry}3d+od$Qji70;wB?ycsdI^akFGd8{_}K}-saS3
zVjN)RE%|r`Pgr}1co63E7=y4s-#ejsw`Idd=UsoQSciqw?@hWB*H>(y?TOqgl?PV{
zu>G+D?{cWdqtYpB0?o0osWEL_$le`JSSE<7=U)`4NaPu})S&n7L6j!w79*U7=UYi0
zyoLMLbX9zvhC~XLbF-6EGYEv)a0-x!*L=Awftqf;2z_;q?X4coX(kPy-VcDJc8FL)
z(H*%GX)I&CcjpIa61|)I=2V`^<!2y1kYB|sq-Bu};{AHN@wEHcs!ZO9MBT;|Jk&h-
z8qrVXb*&=nktECuCjLE-jvI+F*z^6faLKz=$ryRK7-b>KH?FT~a<5LXNjZms@D(ga
zUBRS<^b3MN{&`pGrRkC_FR)Me&a4bhh<2f8JkCxdnlyyCznHjjt)vU|2IRs<SY-UY
zbw#i7^D}*w`ZyDM{B_QD3yOuoYX!z?ncF1j*FqV``u6qL*Ak0e@?$?0<o?|p)q=V)
zAO$JU`UCV;6StL2ZoJ8U7(94TMxXeP2ZNEY;;4@`WWxjIL4Z8Vz=7^WGb}`U(HyI2
zf_5KTYQOkKrwPW?F5Jz&)A-@*!1FhA+NSDGb$|VqJ<?^Y*$}{=M>cl;0EJcL96%F$
z&S-PvCd70m94w1;F!Gty4hP>v$khur=Vm(gY^>5c+m>~^y?c}szjBch>d1-bXrY$M
zrM?i}Mg2E+J!~BpZlBOeDhlU|wzRpEm4P|^;fMs<E~FEDHi6{s!_JLpYQ?9}LWu2F
z*#lt#aKYvJr&`r_oi}BtPW~oGRJ<q%dV1n5;TUZiUeG>!-V`e`MFV8ZutJ{~M(fVY
zTMdg2H|swVvwRHht>1amwphiSA;7T=_G9(inP6Y#FMtBPiIe-x(dsOxmbg~NHW3LY
zM6H|n?qn3-X8rR1C8w6v^*cI864R@c0?RSOlwI8YU^QsonQ7~mbaWHLc*}3uI~`h`
z8~S8Toi~6aYY6Oe<p{%j!Muhe&1$~vuXo%nGx;Mr9*R9{y85kZ_0xSApx^W^@ejK6
z%%PypaI+@)V_iJ2%?Ik~Y;ryJk2)BvtgV9P&78J$;TFF`e3*ZKpmJ_^Hm7nE9CdHB
z`NF%!VZ<MhdeP{;9gfw_;@d})XPs!fqHq*Jz7NlbhYCH$Qjq(_c_g$T;v(mY9!JY)
zL`jzS`ROU64&&yYyopbf4)L*{qC388U2<z#^MH)@ZO$8v!+Z`Dq%GxgDYBcp4(SE@
zZ!aUFc8mdjrJ0AMTu#JbSBNNERV$f!Wc#L`l>0r(2Zwmx9*$0S;bn3R^K8E_Q0N1n
zTFgBO@cz4HYL4Z31N;b%+fCaoN&KSOFQqI(^IyC=5-AhB(#7BaXi$MfVXW_=b~gqq
z`B`Rcw&TI%5@u}Qe8;vjEz9$qmqL)Ycuy+#OO>TlH33<-jUro2Auep31H(oi@LPaV
zTt5D8Zih~0p;3*^*IQ@%Wghc8%&B(23`QSx?gf&$ljm(s8D`MJ8z}h{Ib(Mlv{^RR
zwQjEm(6Y93hETb7v?-{LmLRfF<y$waZ@l1ZL+~wGO}J#lgG(2me*7F66Wbt~+!^(*
z*e&<HQ#j#biXDX*N4K6aa9Yt0171bculqkqENwk?f)o%wwo#gT_L1XrQ4r#$bQ8^$
zdL1ivJ3y3X4y6pfsMZ(WX|q=MH`XaEe9;eA8oD+Unzt)2_|n0-&4+y=uB#D~;A63d
z<gU_vvTG;x9?q;A*|X5h!<#mATW+`Ea->p#>8`#&&4SC5?v(WLz7)9Np4&8_<0R=j
zzjmg%Wt%{Zz%yP{&{GplfHIU5@w5w$r}-xJjIwRWE_4W~1{6X+K_+;AfZTLf>FpQO
zoRjXozh{|X(Ua&#s`>%?T{yzrPMnreE_S2nHHt$9+xd2>KK)E5jD4N+q0A}KsRW1T
z*ZUN|<)$-pFJ1}xY$Z``UkLDE=!_vBOhm@<MzabyJe5hSf3VP(Y^+rMyxJ0^DbvwK
z_p}O;qm2S?;v`)PLkmwL!mz@P`ckr5Ngnf4HkBjQ+9wlKF5kcR(G$-<lXK~%+_o*A
zadaBrPn^$vBt0$H2%#U74(=sN`>R*2><Cikt1GSD0v8h1UD5ikq?JL|@pfN{Y*v+e
z|0_zICuT!?ubh*1P#C=6?5AMM8xE9Aq0BUh8AYYQu?&_x2q<T;^^<~*<skI>VDf|j
zPmWJH5qjtYiN^Divvqab*cR@EFuG|WB1}7yKsG#WW2NDVQdd3Qb2N#?wJ1I*`@F6q
zlTer96C(r9Tn?0*tbRLzZ*(Ag=6v8_=Y%OEk$dcq3KoeMRz$lKoQ}!T?JVw-igFU-
z4ySDIIDE#d%>@z->8DnHfIL$7m)9;WVGJ2##Ql44AUO~0`0WgK3OHq^v{{<$!HsaX
zrevpfWjT^%!Sf?TLsuoV^yJh@SpvV2S{kS`_1cB`mMqeG=tI(a1YGIvSRMKdAIsI{
zkc+sWzMk<{)y_%pZsc&^dHWpn#ypuWdYC7#40khN@7Oo1n};IMRzpgMe%3IT@BWih
zHnnB}x<@0%POz`5zkhC&cm=EJhRE6p%W8%2s++%oS+Pxp@D9--kIFjBCJP_vdU|_%
zE1mr9QdSWEdHV$AjaI^Y_ltU5?hjJ44uE|ADluVad3$UrzB_GY&xalWxv5_NW$b2g
zEE!;q0kfZT6R1k{&RZt%P>-j*=<_dt-8phYL*)lZ!f0oskE`I-H&8`*=ySd!cI<CL
zM(mVFo4xb9y=N-D^qm{21}QE#?x`5woO#z=AA8cjrN=~}pcma}F`BEU;iTSe9iY0)
zCNYW+3SnbJjlMKbZ#}r}lG&M%<V~6uwWR913CP%8xZ;}h=X<Mh{aiWSyI;B#I~|zL
z)^G_@JADGvJ0Yi+fV?!E{uBcbARROX1!Y}_7l^M#v}$5>MrPKM3gd5Gn~U`9S9$Fb
zl1nk42RK(cwlUkobUY!UK6azhc_(_PS)mwEDBX#ua;b5FJ4;kLd!BDj{%SeZBm%u2
z>#X5kBu4o7>7%HZ2~%bdbfQ3eBrz)e7R5hK{XRn^_}1b|v9(ocOK3k?SI%_)L;M3?
z#W2B-<_La<7{S^aWewdp3{n8-#Br5P1rv+t%aB^F#DM%oa{-k<(6D_%wd@oJ&AnF|
zu}%)s-dO7(Acxw8lWAHKHk`d2^`uu$hlNU6c_f!wc{W5;M`Kj<BU~~I3twe^+IU4#
zf?(+yTL)DJel8g<>Up#PWZ2i}u;<Tcj0u`VLGkuuivT;H`N2I~S|4dJ!WYP^GcV&q
zq(+o3ZTD;68ydsIE1RxfHazI7c#@QC5PiPz;%SpJvCG@SiJAAlxKL<N&9P=*a7mG7
z)#XJgcktjNxx40^f*E;yUV?5?uBW@&p1O`4POlGd;r1X6JC9Zl;SAW=O5@ENv3mt3
zv{thCVGLbrwhw*wJhRP+Y^^Ge%Ax`1$LMGExvWV@%fL@PzAkpqQx`ZJ8`cgvN21_k
z>L#aE$&TG!iA=6wJ}h22SZ~~CGDB&6bwVZKXl57RBY7pK$08~7hhwXM3Y5uN>Tw3C
zVCVd4XY;?MSP|)84h>66(Gtyvc_J8VK6o<=xnJKb^Ur1c4TC~6t@*Z&k3Hu*CaX*p
zB)X&ouJ+z}VSh~L+$Lirt)8JiI<(c4mCK|-g3CjPx0`fk{acbZ0v<;y8LEAfcp#!P
z(ouLV#S@fwd4P_?LpNak#hAMinec!RS^2{Th*nN_i?N*vvG#s^i3?b)8DD!osNJe6
zG;PqAbU^z(bN{y2yqcl(_f7q?IX=(2njA=fuwl4wc6VC<Q4Cw<PSFhU?2^?;W3np`
zR-myiY!#nwDEzAYXq-ZQ(!EEoKT3YhJEtaP#^iz7M$vqCt4JX@JgRYGYHL}#qs}lM
zFR)WKbD|cp-VE`EJiq@3nN9d%D3|wfrOB+^i@mim1Cs$qI=*pe@r7!7LwUg$-RdDi
zXmwS|c%&~Wo|D+_XBDo(I7!Ke$!zuj{TE8z2P)JT@W?A~zB98u-_*|sDiD2HRBPVe
zFCOm=*VE;R!GndRXnvYBwV~Pk4NY#)${SUQ2*sN=*J0}tl;c+`){-v_mwHSO^eBlJ
z7)PKKbPSqcZD8q;Q9Ct!U348g(-Kx@SY1^mB9CaLT#1eC(i%a>H>4SRU%lc8o$e@q
zCh@@WxQs$M*pVA=IJj+a)LhA?QGT>&Ki7T8?qNRNYgl&nJ*MeBzGV3EhzNzwki_a6
zURV|%<2oD4f5c;R@a5n+JM&jM%pS_sB;mN2*(?Z;%@UXoFLeI5wRJ;F37+bBz4-~t
z$7yO$1`iiMxFT!8kLgp_Or^Ettc%cd4A{DDkNHZ8VGiyaW8v$1Df)Tyv4cB^b`zQ*
ztbADjA5Ef$<ncx+>;66}Q=9jQt&&kC`^xGt+j6H-ru<;XndKcFow?7ARdXpTSJE~Z
zg;eue2pmDgIY1uq&Kt9*B&rSo)!Vyj^vW#S^z%%tc93ri){R)MkH6jHML2}?{s5(6
zGjozOJ%!JJFHh1TBT8^NHfBqz8LoV9(6wOEl_@(v!lYkXeWR08#N#<~;H2UuBi)K%
z@7d|JEV*2GG0m_OEHdJ*)G_8>J1KbSO~Y`-*HlY6(Z)2+3QGrL&b1e>h$lQPH>H8(
zi!DkUo$|f&FeEoiXKP=&lWuTeig3RDY3zn%w7NmDHfhC!+r};MW~~H)c<)v3vGxq7
z_=F)Ag?N*XxA#onO+ErD`!2Fho1M95l*W-KtK8}*c&cH=u17RO3A*xW)|aYS)UvXZ
z5;Jl?y=wbINL=!hxLd{OFmUDL>GP{AA?@|)jC#Wj*FA}*g2_^}NDXbZh!(o8feC`1
zp}|an4u)5>z)kHO!snjzpR3arHJ`Q?X`G3c`D68)q~~+@E~?ETZuQX{L{_W$K6*z!
zu36UJlLQ;&Txxu|llB<aWbN}%&017nWUo5oA<K=+Dz^oPGjBlmC;n#S2rW4SO&|_d
z!6^l;a?fIfYq4=9wB!Ztly-(#&hY&ar9){WJY&#Zrqa1Nfw^7|IrE)ws;i_gzFuK3
zw9HJJ%syeS_t~RL4`JA`U;GEWqWz5lS)WhVAQpPF*P4&A%c5{n<`1jeXKtB3d#ok3
z(NsiaFz=yT{IB8~i-72_osqca{B8&e2^of-Yko-S+J_6b@89i|Qx`<9Xjq&4!RYpO
zW`(mHOEyk@DEX;avL{)_;iST67f5a7iwU_(iIxIb<MF)0af|7?sv?gQSw;z~6Vk31
zEap(#HKJOahO(p4r^_A+2dek&kD`GKn#&n`>SA#e$R40+zJ+@?WRn|)6|qLtn2q`b
z)V65c!uxz6<t*rrs;69T2e$glMxT4co>V+eMR?tg|KrGW%8qdDT2fHCO$URMYLF8K
zL+g?n=my^KCyjx2l|A*QWN!tnbhj8Dk!K%RP1cykiRoy%B7MTHs-uM~WgOMc&|_rg
z^k@cX=B{!V5<5$q!GcsKj+OT`_f9J{#~XF@2mR-&Cyzkx@~q9xZQwn;G%n8~N_OI!
z>PG3jj(IewQA4f~)sjMJa`cgn=z7&Qp#Pd{=S4(pz=KE)kyq`BSW~^tEjq9tt=o1D
z#}6-|+IH!=QwKsM%>dpgluOtw`?19k(a+d(h)L5p>xaq_g=Ra(J+dF}+N<e`e`)8O
z+lffKo77u<*HN$>e6W*F842{uB^%O8^J)xu?MBv=Qfo0s&#UMZG$s7L;GZj)cK1xr
zT)j=&qj#UT8iOI4sA0QFg;I?8**X+{dpBoNGe9o)K-cA}<;t*IBhb3puvHFI>!=xp
zlembt{Y5ohC^B9+#oov($!C3^Fq>8iCMF6<q!iO);ob(LNJ#0TE=+tNG2q;T=^IlG
z3HaU(u_Q6Y-b#UIdsBtyn+4$4fQ}SAU0Ak<=&^i&%+;=5QZi3KUZ<6ih53<Bsd7<e
zi&otcJ}bhfm+;jt`(G?G%v{|)&dhOI>npx<h*e0uGW%%HBqdjk%G`-Qy9AaW%flms
z<pFIzj_wM5c-|uT!cej>o103gQT;*Pkt0|2m5{k|b9z_*_?F#d8C=?p+nZe9I2rWT
zunxzKY8r&m6U!FSYCtn>zquVPf@*jV*SPt;*7ot(Sev{oZ%_Qzq-BN|NH#+}>HRoF
z<pzj3g62=Qt0Ws(cQTvNGN{mdAL^)b<s}-<q(mlhMlLG{q3`VFbe8-9!4u^lGG2wf
zp859BZd1R}a2>Pv5-$^w8x*{R`@Bz%rN}4bayJ{0pgbDDOVsG3T;j|DAJ8lQF=X&u
z<b0PU9ewGloN*c-T5`|tbvxexUsw{=np~8VMg6jHYm9M?J)9wI$oh5J>@r5+=7*LX
z>#W+OFyf>h^yFEaiy-Hl%O3f`lTa@r#c$nxB6dP`TynP-XF}A!x}+!F4J7H!M=Eg6
zS?lL%J?`@9x|k8P&^+PuZIuFhxER5pw{6nY%0*NcM7mGWV&eORJFHNemb+<LMQ57T
z_bzfBTJeX8*S%73^_YJQWJ8Lc<~F7%2yZ36WGsZ4!k^&K%Y)yt7TfyE_Zb2qaCwzc
zjCRA2ojDoEd1+9J%ip#i##b)yYek)Vu+q4beU{C!P$pXJp~{|_dB>CYS0b<LmH+m&
z-y7&d#ksJcParhNvbgZto8IQ5bhUgN8=I9t2Nkqyp^czfCcBd1<=PV0gge#Ra@^@p
z`ODltKD8Sv%voqOwlgn9BEK}>goWS({mM(n7gg#v5n&cJ3Bon5kveG)9N9!ZevKc~
zIwQJjCam`vm;}=gJA^qxq+Ics$Mfx)W3q4`T89hvo?-FCr@mBBy+i7xjP73bXC^DJ
zrVl%~nJ<U;X60PU<Qj8gG|!rx{a*mfpZ|M<xqg6Nzy}Y*J#br9R+lKb^e}hScTpf(
zDe@Y+#j!jQZe<u#=~1Ff5QIzYJ6l@^NWOgAQhiZ0-g5fgHPd4;Grz&?JHV2|NNG~M
ze#ecvcFQFJ-@<hK!`l1zjJhJGV?h=#-ekXVer~TkFI=lq5|B^7NzKVCnI3$Eg3Yw}
z1qGE`3S57bWFL__^fI&pP*3(cfj$b5!5b9WAs&*qOA^%`%Uv=*fzL^kWChsV=KGZe
z{lARnnAFFQdSp}y_@;c)J?a*=miTG_yWNOtJs_sTvhJMhJ$YD+pYCHgPnzM+5Ue}Z
zWwBC5H&jmaCcOyQDFgPKSYgg)=X_dsN2c@6SUjnZR;jd2339{AjVLWinAJt6huy=}
zWcXg)2@`2N-~Xq<+bjLUw21>74As8_#g)iK3uRenrwzB8@!SZ;qMv5^7)=$Yr7L9p
z-YRB!tmCy#$<Z~V3XALp53nh5TYnp0J!r7m3H3_>zsl++2!&oRhMOI6bWiv2E|stv
zz<>Y5Gh_IaDVFDsj<l%Mp74=xPdv0m3<PuqE0R92R+L-_`y93e&7LSK<e*uT<-?U)
zye+bY2FXej3CWq;Q(sQbHez5<1=h}&ufASk7VH?fzpfk?n3I8aJ435!hq57+$-(wU
zKwsDMAE5Vjhw?UImYLMbtrbK~sV4WZ&U$s`*fhd^;LX?4h+?4sKSRE2cTb$>kW1-p
zsv33xwI{8Mob=lU|Cx_P9KKbDRnfFNz{ia}OI!)84ZI~L*wJQVZ&H(HgK6CNGTFms
zMfKk!Ux%xij^C1f!QI^66>?>)kfuCiVBUc{qm;#$;cI@ThrIf*eQdu_<ab$to3?Or
zvTlAH(*vE@sdd+<tr{jJz9`5FgPE8w5dW=UYX1l50o5tBXj<P@7N3ne)hd>z;@Q&Q
ztF2Ug>6?o^D_M~gxzTh>TkKhdRtr*;?7>l)n;_Gy&x0o4Wa-ExhSPZ{)_}lkwcS8s
zW%ODqN!wyP9L|^G@H_kUfzeG1Rb%@Sx6FZZ(&3%it`K=k-@%y7^(eX7s19D^Hl-1`
zZW`g*Xj^xfhmiO!@!M@5-&_*dxY;<3Zq$Qh;TJZ|Y^s8}aa#%gG>H~?{leGL^a@&m
z{%4Es0Kfe2lLi6jpI;uf9XxySUgg;ntjVi{XW?eoefj!8A-BiC7bKEqf(3Q(bloyd
zm<L(@%u+pOVj-bQ{kpxFU;66VGX%Gx@@e(9tI?koHh<G|m@+!EAb^fT%QrKVp!_hf
z)QBun15Sbpe{~If9wO4nR9ydo#r2$*lYFg?h}K1df><bb(jP9ZHYttsz?=R=JvLYt
zD7t4Y7^RzgWWs|dZlaEC*<C+6R8abR#Syk%lg(xJ3kL$eNN=iqH!qJr;ac#+uE2q~
z*#kjoRT?U9V!7s4W0{KQjrW%2;Ufb2HA-G5ISpb9fCWeZq)VGT$m`LBzxB&zb@1RB
z!qk$;iYGo_l>_kiBL!c6`7$r+&ZjN%QBLfm?-Q;Dzv{<6#x3^^SIp9hRpohqTO#0<
z{jH6)U}>^pv=wjo_B6BRh`X~uRfBZrBWIZN%}>csCR@JVji2kv8um%7<9*u48oRhM
z51lZ)J7PYo8mYX_7O#imwV+9NU<6Dnu?1c)dVJZxemVvJa{F9Hz-iN$F<1VmsaB!v
zM<vL9r;E=5LIlWLc=d<H-IR3;9A17+c`|NKso(0_tV)HLzrFL*$tqopp^SXm>COph
zae2S>VAN@#2ZBDa2@ARJ+9ti~N&{-G7-|2S4ebdN^C-Fr>I~3h`fAOQ626~MP~OQ_
zkv#1t5D#$S-ou>9>bW=PnB>>7dF4{&+#eefXUv+>?WnWzxqm{ql&_MlVe8CR-No8R
zu`JZ=vBoiPT?PHf8!glMGM;A@`2DW7G|D8X*E=eE@7mxu0F@|zg1y0Jmm<UdyWU5<
zx@8f<eN(e;qzc;)t_eEKeVk*B$%dqgH7UM{*t_^*?j6|1S?mZE5*2a@w)%!{Rr?uf
z<VabKFNs<UnIsMJ+0roId4>stJVl<z4nK+5uE4nzOl7M1`lQ%isO5540xg4nz^k?o
zk(svH0lq{T!9vck8TdBE)kiVJc_Ve;z7i6a9GmHu*-6=>wx=Srd)Lx4=~f46jS#y=
zcEi9!E|gC|I-Q%g4UVJkMr^gf4p^*52z2&SL>0`@K%~Ld9+64IlMgvXo5LK|?r5p<
zusq+odC%))fwobr#c{!xWeX!hw|A5`ebZx<vxLWa$<1YzZV#4)grR4@y6zY5ChG>B
z^Q)QecJJ~c*z3eb9}QpIyDgRji=_HA&V46xY;RH1W1pb663-%>?++K$j+G=#ObC|x
z6gCJuPtWN#+vuPItGE;9PIbnJG!*p-{|%C(g!}l#*tKxi1_=89w4%$6?>KUA@WwXZ
z$FXk_g=t;~9pDR4c2}%o&AYm$5u~l><20&4VY`=o%jle<&EzD1#=FJI-Da25W70P3
zp&{aK6Y0;CEqo1m3|waf4TZalm#yXjv64#NCrXjSBCnr}iRo0a%KD1Axu0Z>lVaL=
z)@HxHIu6051KowTJ0a#VB}cP1QC=QDKybS8&f$4!QdVMgW?E_<t)I?IZW`pIohEjh
zsiZ9LFT}3fhvN0&5>kQ!Z*+?zr>0&W_nm(6ExkQm=<whQEybLw8iv&H4=P^@9C+DS
zTUKb}dZTl!n$?ed38N*yCQE|75m!tSSqG`~f6HTngg*R^rfL2y+ZndenBGXvh0-8%
zG_+daTFP6EO4l^Z&XVUj-ZL@A?yH6MFMm@!(kW7&_XFhk^32o&H-Y}h6!F*on6Vu%
zyZ)UZlN?j69NXs%o8NOY>I>?JQX=d>3(%|$#*n&XkBG*|2LD+ow<anNDKy-y?E?4h
zdu3Y3;d}hxUaL*OC*(`k_wMGx!PYF8DU?g}*7Tj(89PiBvA=k-o%}(-`qlRJir>vr
z3>&|mD~v7i$rlr>$)3B;saR)b*AJ2O=<MD?MfK@+C)MstFc!;Dt3AOC^PTmRv~qoF
zccjlChf6vX{zX~Zr?K|K1_7d|!f!S)bLHdrTgsKdMk<xk31JW6qpM}L2IkbfT?t%o
zXR~6+%I)u4mpftW6F!aa@qEAM$$eGvv9j$<{6W@y)h0kqy;e;|M;lvD(xUE;)P{e~
z*oFxT^DTo#KO}5^iq_f`jJuZHbxVC47QB2Pk+|y+f6iPjbP3hkgIYze+xf-DJk;F<
zCL;qCwt*W01Mc-jdL-(*2ytK6(^e>ofn7KpVZdz`Vj<teNx26nFpOjyb8W!*=mwyk
z3&}FO^B>Ny`BO<%GND<)m%M(u?~#2K6yZhF=kCN7lwgIfizYc3^&*m0r0ZtY6|b^B
z|0dso-s3^)0tJ{opb<^Cav8L@IV;&Fd<F(Uq1L0fv$cpFSjogztFD;BaT-b|?eiIL
z4(EsbIsTbih)1WGK);i8uM3z7*D#7^gSaF+63?`wMQL6u?e+D6b}xN~Jm7vwagpgh
z!~BvWe1ATwK0X}XsvZ=7wj#4;(&O;*4-o3AbOu4ux*T^XxiP0bfIWGbpk&DYl@{6l
zAg)cy`R$_j;&H>1G{s@fCKm}NJIe(&X<a$cL5@v8k3tH@!@!dQ?XnZ6_%sT~g<PCm
zQFE^?tF#-uZLKqFE&ZX2<wiy}!4>(Yf;7r}>@lBOOL8pM*;n4Mbvi`sQW4DOZ=<L~
zsXt8JI=lmWRR?!nukHItH$2u~;fJ+W5A6KpeZ`zGdGGeAu&@uWNuFWWB#-aayzOe0
zdTW3fiGV3R(<IXtC?Q55*&U0fipda^=SA0h8_4ohsGR5~7yq8$!78`(^pGE>54!27
zcmMu+6G~+bp!)xaWu_m4oY)@ZTHdf-gn=8rNiQ+DKPSW$ch7kou8%kGVeZdsU~Stk
za{O-pN}cNfGELJbiy&g5oEky)9^O?`R$q%VAEzEK@n-9Y6T^BlZxAFv;Wp*Zv`o@=
z^e&}0uN_!_+hX6u4UEw<HcuBFoN@JD0&`j{P>QQ7T(Cl6?J3hVTq_mEq0cIwArZ-&
z!O9t#v`1LJ&6HXnswSKOVfU!`0dme_KS{IV7h3K*xDz^F;q&ze$nV{<N<{I2`laf(
zZwu5eUw&h%e(78g{On#f!<D{{8XyJrWv>G}6GMl3({KlcN1B!(s@6NeN|CdjjjNTZ
z&v0LAOWs$9PK^2GTr<E-2<dWaC*Qu}%JA`e^QKB$`!GINYnTpn&q3QJ*ci=9{{K&V
z*B%Y^yY6+Bgj{kBMRJQtxfG&8nlz+D3?@>O2BU<*FhyuwL+&xTCHGurWH1wElv_lu
z<2o+IxD8=2rt|&nz5m(!>~;3pYn`>uS?m1%n7`h4&Ajt`-}m`G@8|hEpC`uf5bL5V
z(YtEPdve579;()(NpiYx)1LLA?(`SlU%ev9wbg6=tGu<gzqz28czuAW!>yjq9Yz`2
zvWMidntyXqOPK>HeLh67)MtB?2Si2#*#tBx#FvMuIDUwH7!5CAxYCbwPmsEc-Pn2M
ztMK5tR_4;F#e1*5i0n)J${LhHaqTzvV*I)_zl}*GgI0&7fNlW*qu+ikoy^|z3}E2>
zum-|;ES@aF%5lUq1+Z&S1cE!=yeF-(oA-{Sb<DX5zjt@dI#thVhM$Pqx93O)N^u=i
zyPQlhW{QwbgsnG^{&4t+FNbJSRrt&Sa*!w(JtHt(cYW7^8XK;t0=4<qR<sqw)P`7c
z24@=CGO!tbbLVAM%`Y6!DKFNoJAIlkflCuDi)`393;$9fOsGA<h_^kA$LK11$@|!|
zR@x<*qDjzmJF#d3p0Xxr`N3IByP*c*l29V0%;P6hgJ5;X2N`W<eE9g@QZB7o&D%Qd
zX@L$LU*zh&+L<3f!jO+++G!^+pS2S=5$$Ul@kg&znOKUjT~lFMaGf+iR%8o(KZ?*+
z0e`LEkZNa>*jS(^$fIT}CcB^QHBQFD_^Ud3P!S3}Re^?%5lxo8Rm#Ycs_pX%A5Pz~
zAkfcD-*&i%(croBOoes_m#>t~+0a?TlLb~e@Tc*A^ki2Ugi_rx@OslG?XCSe6Btob
z_@tV)!!<-S)ViC~j|kk!I`^^GRxKw<sPMkA3(v8^4|~Xpx{iD-Zn2OxpoU_ut)wNu
zE0@NQW9AWdl`Y~`Q#=%L$IHcqj0djkw2s54Th<~}+<T^19(Atn`11L3fUXPeo6fL-
zpiu~@De}Tbi1!ZE^jl9;BZDf#6KnNs(0H+L;piFr5a~v&xaRX#1HYF}7k`DZ#0pW!
z8$BQ(R}&z6XxKCxiVE8Cc&d~iMjzVm6L!@pbUcRVb({79>qvoErDv-BKl6!|oN55A
z6&Ur!^o>oCCR^NtcBf;WkwjUKyeE$XKy(F}{>bgq*`0_Pp5I)(+?&3eQb9~wt^Oo8
zft5VI#JYpyN8g%&;eq4&o6E!1*A6+vF4JjQ&3?Es|79e4u=&d4?>1nQ^E!H~-tJ@@
z@|2a!cu&mFl`<jp@qVBZP7V^#z){05j}|=6Y08>|1m>pyTs#(<`!E-4%oKlTq&e{5
zjJoVTsogahM(dV%O9VWJ6<XCH*WW~R#3GL-T}Hfz_mtYrv@SeYJ@Vv)v&)sJrB4m!
zUQ16ceS;7T@(-Y*QqO0$;<&2%Z2~S7;3DRhePmuRTlf3UcF1R?tx{#w!<L$LD_mRL
zi)JNl2Gy#5bA7Pa>d1;20>8V59n9Fc<{R?-gLBoI#fwaJrc7l1Fv36=U&G8yC@L+_
zo%0zw9BxthnD?U7HYn32UNHvj8}?D%?(*%feU=Q(n|AIy80*b7*)<k)rHi=$8t0?<
zC&iV_Wy(X0D3+{zj!=gAqO@U4*y5%97Ux$+f{9TY#qKgX`r%4X%;zIXDy(~x=M9t{
zSe<~Gx~gpIzb#Znre8od^*p-x^5n-v7t@&aF^{0GOM7)jj`kN?ni|&vZbFUQXDNsn
zq(7EPUKS9n+{w@a@t3_QW$R4}!b=}66z)`PTXy%&g7)WmR*SjCor-&j>$rID1o7=e
zyg5EPm(#*+wmh1eCo3uuofJK>Am6F$x!+>G{!*EKI}d8r*Nw&&{|e0u6fikgEt<B7
z8^Nq%P!59tm*NMeeRF7R{d^Z-MXVPprhBY$e`BCKe$#QIflzJNm65lad2UTYeBouw
zmz|clUB_MQm0?g8a$JCA1EBBuK;{8p2=sYNL)%x-^|(jGcnnA?#>X$+_wgI}o^z$d
zi1dKW&b*S99`-9kKf+cm8#rOw=+$>YErx@m(mhpV)-`r#`Oi_AZuHpvJd)G)U2?Vk
zbNc7>+6Tkz6kyS3{@vAOKP;L-507|^byST_1w%Tt)XCY^Q+_`vxG-WTwhjysJVa1X
ziuI@@Q#+QgHx2UMHU9=avv0EP%;yuTnrg3>FV8hz3eG?NoO+KT%Z<J*Q;sZU1$-_C
z0*%Ql;l_R5d{7fJa`k87tIV%)O{aJd&GkS+!mb+}g72>Ph={T3SEaF+ILIHR6FWi5
z99JOWGvDBZm>3g$XlIRANXjTc35|VPXJ~x@nhCyi*!84c8r^K@I_+7P^<!pIj|!{g
z&nq24J)@zD%sHLC-n`O~s(KsafLd@~Fx72T^<&?hhL^`r?6Jrxnw+Yr=k9*8<!!}w
zS|A2Us%7TPVEVIRuWi=6evF0VD$qfDwuMcKW<`s*Bgqr0+I9lr{gCo2x=z4F`cviU
z=qJ}}vG=|oO(5pg?j#2FE^jnBiUANn4su;5c+KJfz)dnV@TJLB$1X_ek0rz7GE()b
z11Dkv*CV9dp<CVh<Rsj(sl+v*n=uSFjZM?{y0McPH1dvB(Z7v>udmnSAoMs?oc#Rm
zFMg<11LWCxSU@xheudWg3f0qOpjgA)_Zh(1`S^mQg&56A#J1VwOu-aWA<RQ+fR`>~
zx6vbatgXu(AMD2I>#?;kCrb#ppxL3Avk2pufXq8@8hn<Utsd^5^8<YbJ+H=y<yR-Q
zQBME7u6Y4=@uTo;^E`MT+cM{j^TTyrO^pwVp7NfbzBrlE1?JoG*w>JUPnNweCZv^y
zyn9F=nUB96W2RbAdM`Dx#r8<137@d~!GgUR!vQ8ocukOuvQ2V(Dcprz(!-{cCR>62
z(+3GGLv{R9HMWAgU%{r*H<n?e=R@uHabeB39c^8l?c*J-EuZ6UW=4LD@dq7ZRa3Qf
zN5Wrf;k()5Xs-f<ypg*go&1Jc1orYUp+TzW7h6Q$T&j7^Jhq>doL#t-YF7B0OA`>m
zP+MHt99u@L>4T`S3Q{V)lZza_##^s<>ti@ohon*wc+b>~rjSPWn_7Y6eJS5we0+ND
z>4|5bvi1o6#8mR4kWCz!5b$@jR@NL5hBbXj+ZJf_oZUhIjnBbuU24T=iUdZk4a9?U
zJqJ!yJPR3BDj-Dr2Z>fsk&6;fPSFC9EFiyvIlJA7TZ<ihfanZ26hN!V)MU@ap)4$P
zpAtyQ%iE9q&9e_S@9nnbO$kp61I0x}iGYtajKtyIaXxFzl6Dc6DF!F2gOxOq!mbDU
z(#|v-<a6|Xedcn-f=P%+*RzSR^Vgk8ud)jWw7?}yMh|=N+7t6L7Vb7{tNO5i>uI^5
zkn4~S-DpKRL>H_N%AT{!CD@2=IPxq<6R`W<^u=QrrdOKUP82KLejs$s>UpSDoAA;v
zOJASu-&`3>jG%WugFe;c@vvaiRJ0)rMu%|IPgYMFim*yQQ*<-)^Xyz<@Sdk$`7~&*
z(C}4>Bl)Rcm#U72d3`x?_`vOUrfGwrAPUzMB)1&eA5C2f#`k{&YAq}vhzxU8zu_tZ
z(_wQiYyZrhw#I3(hi&fr>w_ae)yeF0YxP^RELG5_6+OdHN1f%*6af)LkT^3ueEt{g
zu;b9Ey(_UgTpgq?zfy9SppwjkIlP&~_2gSTuNSd^<vU?PS=vLuBPPf)EOTi&pa%iD
zmmYq5Ik2~ba&m44JfxkdbuMu|^}t@aDgML-k(&#%$RylUc=w1VTi9&bjrj$*?34Gh
zBJG&!HB9a%jx;Nz3}KV$OT$ABiXrs8v2qbZvS-^}Ob$Sg*T5glJ?d{DKle2}=&c%1
zR3;3j3%gIyKzxlob_a4-)ZjHJ2U>s!Qhh>R|4>RKM)Vs)^Oi+2a;~MUg<NN0)x*JC
zteP#F-An)P2PDGu&S&K1*O*3E;65}~iD?9nf2orPQaS0I&dLSYtSn<cd#7me@rJ>M
zzJISsEpS0ifEB5kTjGhBG|U^DkvUW>mJD~dxcvvd3F&6TWD)^__azYn`5qv9W#4W+
zqLv$AOKBy{Ty-{fXLf8nsRe6}y~I?0t){DDq-RND%*3+LDv*?IidF#l7mUyvzmw>W
zrn%~~`YwDk6(SPB8L2m4%+xnVxQ}5NZ9Q$y$G2;r{Gx4bqC4N{yKhVUB&jeHIkiHl
zD>l8|z23@oM&mRG=QT<3c<FVo_U*ePM-hqf=aW5$MO=RL>VuZ)tBnSynFi%fg+u9v
zhabHfvmXTC^)(T(7PKq2B)0C(Sk6svjjO6~|9aH-=`FzPw{Tv96{-<{B;3u@*%3-y
z|KzhL_3zU}rU2`3N(xAcIY}+;s3t{l1z^4~0Z?>6cl9^dm7#u9M=z$j?F9aNX8-Nv
zOGFso9Lc(fFUR3RI^aAOQZB!(3gwr2I5l}>RW>Z>1S+O~0!dY3OBA-y5j&?{K_`py
zP@_u}|C`hk+v-ESHCX|Xt}k|tUhV%XD4^EmG7-97SF%Z)inAvwj?e&9)R1AeLp4D=
z<*{!tjmZKA@{e?7v6{58Lr{Y_iMr){>{8v$Eqe>E>YJH*+mUxRj^K~8t6UE$@V=@I
z^U^}fxcZypCkG8=&{z7=mnQhMn;QF<fP5H0L@aTCm~ZzArtf>x?6}5GhCrH8MW~JS
zF=b2Zjn@GjNynwO&i-=MQAinx3b1S*+nj}u>jhwB$56ycOY(QnOchiLib#r`-$I|a
zj=O7Rf*jSA*Sq=^Y{FIe48QU@nsgT7UxZ><FB9rTi^dx!qqX}OH*pRhdy3Vnn!k<5
zrGm0ztO?bK*i)(54*Bhhp8WQZELI4`;9M74?7mLq04u<X6}B$g?<mOBTKkcjAFVQD
zO5pUwi=U{mIDZs-%V+e6i@;%(w{OxAfrcyhWxFxn+quIfM~aA*7~iwoJ7aBM$L@WD
ziRVWNjdlY4fMe@?OHlz6)0x%uj5C4$5Py23Exd}P;kv$|8$oO}^jZ(rxpt`JVA!bX
zS+pH1q3EIRM)=!wO5Er1)H=m|%?YI_GPRl{l(zn>F4)uy>2iHzw}!_U{HX=$YIFq+
zo@#5B1JEGX=6}&T^k@$zU9W-O;hxq!T_l?)8+^&GlcFnP>nw6PZT|b~3+KxAleyRV
z`pOGkvjc7mMz-DZQ$7^ea`XyF>y7To@9Gj4_u8be$aH5^v+T>Qm<*;|zXkBZIQ_v4
zTD#-FYSz8DfGHpMsdg|QSggcO-&5(3pvk|7rFYB>yDMJ-w0<_JFkmrk@q#kfBkI4o
zro9*~_^Wq&)!Zc14{ZpxsOUnkc0CI<;(o1ce6P;Q_PLvOmk&X$ziF$CG6M9JQ^yA6
zAsppCM6vcXCw$gwCTzq*6e?*Y^{o#3qJ~w1o29~UOp9xrceXI2s~J(q0{5AOxNKAc
zrJ4|r@@gVdr$?sQL!l#S!>~^N9B<@|s79RoUrRG;j`-htTv@$VmsFpay|iPyGITJA
z91_8vBFn1o_?n={whdV!P5ooi`uFous-GqEqv~V@!ZmAO+A%0Py1m6VATR%l6pxef
z5Xy@vZRrm-NGYG#Q_LUK)MC=9PeUNAx*C3fjfCe756qGD-aJx$wDsC&MUEA84pKH~
zr?@lx)%tQwyr}w>2g<FU;j0)NXl?3xnRqL#4cQC?bo4;q$yjWr49izKcD=XvQw_<-
z%#6>Vna)|6dOys$jVJ1tiaeoOh3<t*6sFlUgSZ(!x(Xvn6_wQDIZ&=0C{Cqioe~3<
z8MRW8lsMv)@T65F_~W?~dz#n*l6wtKEoaQ+)1#piU(1n5e<Ct!Ty)YR#4o&~sz&GF
z^309*t)*{q#|gTxF0Dx6UcQ<w7%vHEmaGmd`Gt_hh8j+s$4980wdkXBs5O>8)4BJD
zkM6u7trR;+J>IySzoci#7&Tj4`Gv4-v32mlynw=z*=G%sm`l_pKDPGvsuDe0cRL3o
zs~A`JQ@;ZFx2@8QUo<N`NqW`VT6jM;i3{CW{1YRrc_E;2EBWP(N5lsS`8<+y5MPt}
z?kN*v{-tRLBdLmghA3t3y?b6=1h7sJ<u}(}R3yE>u|BA2@=7W|(H0;jmR-zL>DF>k
z8;`d<iA#U1_TB3uc+QG%47ArP740>NKuV!6bl35WgRr773GXst2Etvdx39#xJnywi
zwT@?pb_SwCX-J7(oMKz~R5$pv>PJ`+`vL>jJyahyU%Z7$T;Hx+yFi3RkMhM%gpw)8
z3n4n8>iStndPz91Q(wEUe~{z#j_RmuZP0cbmI~g_2OcL!Wr{Hu@dmN)ItuL`EjO&^
z<TKI~785*|7Z2XQx;TN&y+~2Mzjyb=6F_0_s5Y5t5sEfs?wR1=^Jitmdy#dQvQTAB
zOj}1t;ee_1*UE><DiU-8e9-5eP1Iic!bPV(Qo?dsBT&SPX>1@canWeVJH{3Xmv50b
z+x0U|b(YX&q8|r7OKkD-TsnT*co%;mm%OW@mStsCCU8mFy5_T5g;Q00NGrrlmk=Eg
zwSe!xo=q}h;dUx@um&m&PgE0C(t9*=62w27=Ay0J&a@k9dXnaBzsp0aH39YU72U+M
z*jLH=lo`I`!A8S|%&BE+e}oo8Ae;z`ahqRrc))-X<Cn~P9(2onypPyCan?lGq3+TB
z)m_o|?@E}Cq-?hstOIoKO|6YH(^4n5PjaTl*cxlSk9C)ONJ@P0i`{jW#Se8qhc8W-
zrFRIjQx`lADW6dmYVK+;5&3eBcs1Nt{t0uHhU%L|i=B?84wbVE^xcvSms5bEh4vo4
z!vow8Bx9B|8`hVGcgi}3593DfdX%bigt?a&VkuHfT!`pm))}PCY!{G3DxgEZ$U7Kr
zOYU}9yOCi2YP0R&&RdU7i?M^$#^}7Fp_VP}3B-D}&L}RiucPcFXVLuvOtq)Ok2crI
z`!h&}rIfU;<3F19o*|>&2Mx+Eq(d{V?Ub~kIoA;)056qGMQ+z&@s=N5jL9(`RJs-O
z7z-Kj9wBL6o|8{05ChkA<I_u=4=jX5qwF*=kCD6%yFh8iA*fvhq{0DNdsqLGq?&g_
zbGq)+B!K)usichzwN_9y-OiN8<c8fma4aO4&lTyz!6~$+!J?~rAP0OP&_Z%&8kl2J
zsFeNI3Ftb;ZAg#Z<UzqXjcNd%lYHdcPYh!bGh#8*`xTYLx$GzbKsb?@Sp|UCF1xjj
z+qkuf=W?_B_ms8&m4@rCCiJ!6Tyby;Ze0NPo69(Dv{)7kTRFu_;4}cy&d&ts_kZ;j
z{};x0wX^<Cbs(n^v%*&c*e`m6lh|g$b%8vE>IS;xO8WsJnmZz=6m#pF9cOE3Mfx(s
zye9a(oPe90USg~0-m6i%&Gj0Vn0t>5f1RDfZOrum9FrT|X7&qZ&*#;Gj?EcpfIay0
z1K&YiZyL9gu14^{6D6*U&?4;(z|N<|xSp69mg8jsr7erUDQ$6@zheMREW^KQV&#-H
z<p2y<;Iyad+Nq}f!{D}*<sY(+R)3RqeD*&zrIl$D-cNOut?y31@m;g5>SS+ycPXU4
z{nP=KpOaa8`46aEG1|?;ZTJtrA3N`E3aDV-`l|{iH1?hj01`ER(dR2dNOpWB7o*B3
zx)qoB`pQc1A5)t6fmq{+8OjMOc)db4=``5ql!f>L<6r#6l$fN58S;N+vA6vHuV(O|
zfi3V)t4M2Hme~t|a4+8y*w{|I(J9aN4=Nr?EeRD0hTgP#w*nPPk*;vVz<domn#)_n
z=Il(%Z>|dufUVTY(e$gpblQ@afCC-P_4NCnV}IJ^@At%m|MU`whFU=2*XznNrdCR`
zP|}GN6Smjlc{<Nsn2TN42JSdU1_gf;VCH&SWuvtd{r)d9iI~5rB%1y?g+=|-E`M%@
zKM%>Do`IwBzvm7W{{hLY$(cpu3`E{fgW^WPsu1mw#cn-4;i_F5DZh3{aQz<q5Aln)
ASO5S3

literal 0
HcmV?d00001

diff --git a/_static/bonsai-predictposes.jpg b/_static/bonsai-predictposes.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..2e4f04a22020342a8f45bde98ab45eb6de607a1f
GIT binary patch
literal 79334
zcmeFZ2UJsQ*De|aMSAZdDov$HS1D1ECL0lyUZT>85J5_)Q52*jARr*1bfiRjCqk$q
zz1NV?q$iXRAmwJCbMF6td!Kv0^WSsE9rup=e~YY)u?8#ensdJMo#mPHU6d)xB7psY
zj=m0nhK2?JrTzgZvj8ms?U^%wey9f>^>g<8*|T(XXBp|~&z)yxWM*b!WMX1r<z#1J
z<zQuEV&`S&;Ns@t;bCUu<LBk(=j7(${&Nr-TIz4;&N7@m%fQXT#KQf5`=I;)aGXCA
zL5HEGxdJ%DK|{+yL+JpB001;+snq^S@P9jK&QNJQM}MAyk%@Xk4Ljfr4K3{%I@&*}
zrrsS){Xc+?<1FW8`Mc-1jGxh8@#I!`lkoMt=)H<I9+P3LnBsG<PzFX`K7Ii~aS6$*
z*RCrmtEk?*bzAGcwvMiz{)5L)OwG(KEUoMv9G#q9T-{!I`}q1n{R7^<3k!cA5gC=3
z^f5Ul^;24UPHtX)L19txx5}#On%cVhhQ@YeM`u@ePw&r>(XsJ~$*JiXG-hdeWp(ZM
z`UY-q|KJdRL^wYAlP?+o?SEzK-#GglUmR4v&d|}(($W9P7tI+z>PgE%clNUUInKMr
z^v^uGt|+`Y&wVf9YegG_sG<p$=egG~Bd?eeS{(N$YyX9_{}^MT|4W?x8)N^@*9?G}
zmWDccv>X5+;FK04*<r-~+fIH=`h^eMl?gdHhmb$qW4iM8o@C`s-l71QAf4F<JQP5$
z3xW`pi<=OB0OQ+?%(I=D+Mw^<D{QF&e@Z>_T=NEtIo9`}yBO%R-gmYN5aK%n#><Pp
zL6gjf9;Vv+2RjT;yL!!hSJ`h40Y5l00pwkN-UJ<7SW0R+VVwaDbT?4|?Xx&%(plJ^
zl^pp>Ji$6dngaL&CmU!}03GH@1Q^CIocPc~JDHe3%3`Zgoh_57nx`k?jc#%MiYBK^
zm?sP>N0tyBMm*GoFu_K5G7HDNTO@)Xo>vwUigwI2QE^P5iwd`h2B>yD*wggTme?l2
zY(MEPxXL0XH{xGW0AY2Q!4pObV7VGZ)E2?k5-ll!yC7mpGzAa=&sIXkr8{Ovs%z&H
zLZ3M)9dD=Haxe7|i@Ta%6K;FN`K|JP)ACD!pUiSWT()KnI5lL5@C2|5vb!9$a!+Yc
zIB%rw#RW9;{uO>S%j_@5{>Nq&7k*A^IFor%(ZJtQ6hPuI495-oM-eoCZIyt!`n=7{
zG!QJrzaG`>TsBw>{?fhrrpm+j7zRu`y1_7St09{z)w2aeyA!I&p@?G{J_>+Cmneha
zg$<cg0OC0m07fc4=d~K!Uj=IXw~1Z){e>0zPl%~Mxk~%}pUbFRGYasZjF5RU8==Vz
zXM{)K^HDR+@sfR{8HxGX{i8dBcPB?$R3?7ew2LY9<fZW2*oW#C)<7q>C;$pTe)rUm
z0+>nOZiEp1;irQXz*j+lGv0~<81#qXP}tm48Q5Agg0M?4zl`JRPJbU7WQiqDyKgTK
z9g?gJublYv{Q96bG1BMp^X?B8F*Vjj{{+t{Tp!y^Qb%KUC|MMQEL>+Ou|ioFzT?RG
zj0^<(L-8t{4^_`}u7>M~Gsec$cd2p&&ISY>j7~mLvxcpM;rZh&MtO+JNpfHx1rSG_
zX_09#1)#A)0d&LK>vy?`URdYQ<~vFnV6zmKd&-gJ%WsVuvL-t8JOvu-<5Z%?<WE~)
z73f}RfJcF<HPg_r?N~|V2+ReVtvjxiA#-nG>0u+-+M%hbDkl8%_0Cm(4`j}9niyC3
z`9+SSyqr1@R4l?+#>kA^2Rh!BBdd|QVUBR(Q7Q!>OaU-rccVqB4-1k-!b_W3UXk+s
zOQFfl>L(T0o!oC0v(+zq#X2<y*dP9Ur2Fn8AZSE?Pn;=tF?$BA1qlnGn{2j!IWcn-
zp=41L7%jv0pt*%z=ZSxk(#yOuX2$M|4tZC9TSxI$c>PBFqyRq5c>2O7%0gUor-1(h
zX*4iVC6q-s`T`D-Ro6Z<V>v%%X>hw!^lReL&r^vm)5{}9x7GF19zN8YnOgT2;dQl>
zXq<4bo`u&K<St}t9&t35`khhw@Z~w+g>-RZ-fMk<o=a*PuyWLD>)~9+?`+uS0qo^z
z38?beL2tLA_QTYH<ad10#zOJYB?_QC`&7#GffnP{oY()W=gsVB%`0Q-is(LYnaa}E
zXv<7Te+B5e@xYBAabL`n{%VL@M3xZ5C>7WtG5bXJ0Ri@i5W55xVIUNCiFjal@tV=F
z;PWGSqmue-Q444cqio1%wI2GWrm;{2e{R`(XD|@~+|x|^=avI5mWqiF--i1H8bq(l
ztPNu<YiPx*bEMpR;=YY+Pd7($G4iLZ0Twam2yJM+AYjp^`R$kq-rL>@y3o1_Ro>Ab
z4o@pmL-i)^@P9}0v@Ndzh8|2zYy|TC)ec_$8x{VtX*eyURc%HwNH<uf`nHQ%Qyh69
zP(R1)^j+t5sSoqqVMvZQN&22R1Eg7i2ENcIP&vOQ=0m4fs@daCbwjCl(Kjj?Zv6T|
zTSuP-y-DBNUG`Vm{oRR${ssz4s~cjXGbMN$M>PAtX+0Adt1hXZIr>$lD#U$Tcex%J
zOhR4vzW6+SLkSUK#adIWYwsGct5k58F5X@G<sA1ZM!@o(K@8WP{36$o;L`&QgMVVQ
zo3oKXE_Jmf7X2-yriPUUi;$5lpO08E><-gv``(q(-b788KDJM<ddhR6Z)^-t{l)$K
zop=%LOaF2r|8lGU3N-#bD*R6v@%;ZQ2GnfLiMmR=CK3(XdsG8Db)*1t5}S39=<2;7
zO&loPhbb@q)MM?5r$({NO1F<vY2gRWr<KfcX><;8;B{aGh=r(+VdHn>I|=zTsgigc
zC}#P{nWIRZw`W;!4TgS5=qA5TgrTM=0Kzw0;!hK|iQl`&6hPX3*(4IkPFUI&?UXBU
zedaCH!bqEIB-RD0K93xjU5a;7&+pKJE)!S3OtcCSs=87H6DD94FXSF0Q)hUh){hi=
zFATmcVjf9}{KJh|KVnQ%m8*Wyd<2Ls46>&H?&t$|w_rq5k-uI73BcIhW5l}D)m8)R
z0@V9sx%H~mSOaujOaVm5RVDYHpk$QEO^4R>9Z2<z<Wx@WhX^`r2cRg=#pk|aS;>%&
zY-V*W1wxhXh{LGZ5Ybdn*WWK#Df*Q?IWA?uK1Qcc#YFiMUp$y=%zQZeuaEv4C2ACa
z-95YQ#PvEfS$Cihi?E5ud=9x~>5UBJDo;r%OJq&tOLW*f3eNpSl~&W6D)kh&S@|HD
zdp8F5n?aI9il+dga{qdnKZ2S(oE(p$eji<M<e8>A;!IpTiAIb9KvYlwxF@i`T{ACT
z{hti^g;gPn8m|9!8i9Yk1pI51|Ek`9m&+#5nF2VDhs`3$V{rFT*it`i0Y*AMnH(g7
zjf10gCiqONRNPgQzAKh+#UB4QDuSE-+f76COt@3@P8m)S9a5MF_1&8_#N9W!3(G~>
z@mZAZUP^Y1{QO0l&&&NL9Xg&q-;<j1R+)BaDiQAb<hk+IE?mGS$M_1ieEzoJeqwzI
zeC8c&@K&1_{}RjGWXtgV;O>6mo$P)I_z`fUM|@W|dE2S=;Y<kqYvMhE)RM;YO^jLW
zm)<Xk@DS1Ku@5N#lZG@o?;Lm0s*gy<_H{32W2vjIm#3CVXk)@A-gAuj>}!zW@XTuG
zR*)%1n9n%Op&^am%1Sn?Dle|2!Qq*!?0g<+xli_?W@t`WeyhQ9x@%Br)T-+|a1*{9
z4;UiT5HgS>P~hF#mXT#F<So&NRGj0V_)``$YSpzNQh9spdOvp^W6^A@%j@~@KU$sN
zw9*G@5e#w(UPi1`EtfBIL$<4E@{K<mhm>Ecd$XPH6t_A*v7nRu%JY&-n*DqY>)~o^
zVos7xZ|5OuP*<dPx8-n7Znh7bs&JZZ*y#RsZ=mXwZ?Fi~el15ALkDBDfSM{98j7)2
z&D&X94MH4WM*sX))Kc%I_+hv$W_xLxe2&0pK?uaIc0a_4AGQ`^q9Y5(e4=L+RvF-u
z$@w3<o6k)#xj(woUbJWq)3#1>ucKl+Dii=*4}V|7JtTVtvN?X_DD!ydE(Kr<#G8Ff
z2Ae*uXt#K+K2S)S_y%-|CI?884zBCyYwO}85NkOw)V=s>Dhi~*MLyxC5$6$^xk3SC
zMN$C4jGDuI!~>(#-;d}%K7swWA6`AZ5859F?tZ<)X>3f&kvsmKeKeH2%>f&{PXV-s
z0sfN>d(j_<B5Ml)Psm0jKJlQi#q7|Nb76t761S|2LlAW>l?wl&&TnNOAumg0kiC4n
zmbBrfju$S<zt)3)Z&jHY3`S=#1zLzV)+dtgY0){(S8(So(j@(m#KHwBfOo(SkpsXu
z40{elmWa(n`t*@}C;-Lo8hF@ATxpidH9s(OV{M83xAm>rG}G<e+L<a@dKU2pdVynW
z4QRO$#{7Mdw#fTK6qnnuYf5_m;lz7K$TP=R;+@T4)|{Bqy=u=YUe^rZr5OON@&SZ+
znM_TAznm+f0A4Flt-2DdT?E5+0&qhRXyqsXu6rX3W<;42Db}?xXf#?Afjc+Z;g`NO
zaa*&b$l7AoYG&Q<RUh}g33NF}R&a%$LDL7{>mLE^`m4p$M1k?ukwd?9CN-9a_)UXn
ztKgIH(x43Mn#{LGuLea!#w^Cp_}e$FoY0fmDFAlZ5UnW%@IwGbNC%?XaP%aAGKfe+
zMJpl;`TD-sp-#x9(EK%|u-6(!sj(;+cq)82rJP9txO}jA=Dk+!f<KR6^PONpd@!P=
zy@_HAIk%z7F`>l2^G?+&B-BAAbszfnaf82p?9x5y%no+p^1CxO_Loilf)Mxq#<K2z
zCzBMa6LdUcHSB|J!uHk8(QsrYtUcs<;)bzfu8V;SHdD@ET(-QdTtYVQkCgpO2Fj0w
zI65qX2x8lGCz({t=dU9D@5i5Q2^+7v5;gg^*^eh^5veQs4vRSj0H8)v7+0DE?f<}V
z|GMk{ROWw4;$M>ZyMg<cPW)%r3FA2^$hsy}zhKzs>a$pb=0wW}-@cvG7lQ*B0Vm~A
zSHG~rFwPxnh%f}JoDT`8|K1uQ7#17b65Efij?$O)_)=@K7qNWxn-jC1&P}`Iv|OUo
zq!LI-goUV!B}bzrQ*|3BWeyGHK*9GqtgY1zJlA@a%PQU#W(qkJn=l-|;V)`cgR&VD
zQnBIt$Qj4xNmV|<fh6fjcVBmLcI_goKE=GR!wD%@=iivIBp-*oFl7@V56oz!w#P@{
zTX{wgW$ZiCQQ?q-XzNTr+orLdj9+ist@Aq%_V{yM856%IGXS#EiVeGD4MFCGq(&-&
zngr|@?2=E4v96h!{JDA1@LMtHDOR{P*}k}c3BuZ0fYO&%NsTYDj_?qRjnYeK2vQS0
zQ%WPo@wVRz7vB-jK-7r(X~Yf9)yB@C^p~eIL65?RY<fvO#4e%Pgde@DtuR6{HlsVm
z?q=m=vnmC^@JFL*_q%PPnr!Ogv65Kgf?i&7&quce1%Fm?Qwgqm{x0LL`9>=?WIQ+V
z%0x9&ACP{5a570|&tPMW&?$D>Ur(H^E?u>BKE2&bm-B>8%rjMD-YD@d;1=B125#$4
z<;Z%sa`JS+AcO9}5D$u%Tk+fM8#xJ(Y)8e`X0ta7<rC6RPety0Z^CZFBb-BRpXz%$
zUTA&hl*a7B%PxBJQ!Rh#Teq!Vr<&70f>cJHVck14Wrq*Ja=aU5x$CJj9)<48%}J#m
zYYGfEo^V?X%y$P>lxw;jZ)%e=h{1eF14D%+WC{vgh1Csv0W!$K)3U5gWELCwyn1SB
z0Jx}7bm8^89?ci$x{IJm_i#p0<;2UFKv0+o!E3d<D`CTL#>;Mf&7`r8JdiasZ^ND1
zJ6Cp{CC6XrXAH;wb9tZA&kbti3loU>Bdc9quN8-t^+LEf1#nOXWa%P+@E>uRN1@F^
zOR+FCIl6!#*HJZfSGX$`;|hmww@zQzrqzFeXD=37zB4QST=9bt&sG^uwB2Co?&3Zs
zlN(yyxfV;e!A@%Kj-4gOE;!A-`fSg$L{0bxx7u={s%XS(p-vUS&hE&AKp|I&(HamZ
z%1uQjHRA1B)cB5AjgHPhSL}`KwkTvO8WLk=5Py!1@m3t@B0Rb~jyws*dwM9JZJv4j
z>r{8<zSh<ke>=M&&0vJglhSWq-mPTwD5A#N8X>13^&<=(iNYGJ9i8@PckI%;*(o$)
zV+*5|)ZjcbUlW(^y*CNl{rZST8hdUbm$^{P?_!O!0d6har$6B6{%DYP-iF|K%5IJ8
zMBT7~=eZG{x~@{omwNP<o+j9xMN<IbtqLO{Y8L(Zgg~>AAgPP+;tjvSQk(UF@Ki1v
zv5q-cD&TSXyoY$AwzDo1=lirf&+7bcVRaWF9c)a5Xf!z)V;S@;5WHB>;4>zB$ik{2
z=;R{)PPx9WR(0<<)WlMjIq;!Hb4*F&Z8-xX1t1jQ(H^py+zOgJgeb1Vn1dc-{bFm}
zjUG^|Y)Tri$0M_jq?FWRtyjvq9#WCU{FN6%HJ{$1GdBb&kp~g1LC+RZ5fivR^t`ne
z)-LU9Z{c0~$a1BoV;M1p4*R-M*)qgP`q1@TkeUzOFT}2@RUFhjNz_+5q5!Cija4!)
zFQw;dPSpRjdUZqD=$JvDjG+LmhKFIx7qWMO<Z!Us6KaX>WA={RVVN)=`ED<Hp91*p
z^jgkXn;T$kIdc@Ll+8FLv<-ezIU@aWTUGUlf$r&v$Mdr<@f1LHYI{@0U}a&42pfZ!
zxm=M>ardc17|DNIzq(Gf@Lri|z)N?3i9!9GlJ%3=f-+~8mDpdZwMnwScVw11b(Y30
zE^r)_aQ`|(wYmShC!5BkMR^&z@^<tw8~K5a`P$$+3INQ;vw^A{Bi~HLhq;Exc6%z!
zKzXDOEc-X}r{k__b_#!KBw=XwrD8nhKQof(C;5CU(D~~b$Mk=WaQRf8$V^owF9NUt
z%6AGObBa{h*bGwHt_;qAWh(Z?%LOh8;f@-*pz9B=Ef}^&E!0i+dHim5$SJL9JT0XF
ztS{6ipW3Wb0NBsAP&sofETYV(pG~WJf@oNzlTuvZGHh_>=eAg-jr2PS3!W!VYC?=t
z%HTFo%~4phV=+|anmQ0;ORJB|{Ub|oQW*^PsTvo&VeNMOsyTu0ktH3jyPowHrZjsx
zvHcW-=!EdPH6l^2+3ZSc!Hk=_adRCLI9;y$Tav7Pyyn36reH&V3ZM-&Wj2a7iU^W=
zIg|a)heZ(?vkehPyq1<Zj2;-u$h@1bBO@ZBuFq8=AT#kzVRhocBn5EG^~h^>e1W9s
zxe}YR3n6R6w&x~~t!6*J-xfIabmJ3YlvN&lH{%!a^?P(jZKDEVacWvWzdv$ZQg6ai
z(CqR4rKFzOrd)63byZN!<lCTgD6U;)cZ1G=DrjKG2N#CCK}QEW>uGcOnf1Q&I?2{a
zX|sOxs{nKQ5?C9sSz8k6NN~<2T1G!4K+d%)_t~m!BJZu9+z{UnYqAY4roEITeLhB2
zJ@KXnFdFzQ5XLRnAr(*fW3_U;SqrlX5^z-=T3@69sE;V0LS9bgV+x(7ced+r`sn?D
z<hKj2M5Gdlbw&uQ?5;WA^;xo=_24lz^`+G#tA|WwiDT)Jddl1MXn2HonG=w6+9fpr
z7G1p@dH<CvD*S!}hhcrtx8VmHW#lRsxWq8bHQ3q)6!!`W$G^)!*EXQx9B#gvBG}o^
zuL9~3sW-mXB~Iv<TWCH$^5X1#$)G;jLewLr1v%ni=(4r<L|dG`vhdT*lz{QD<%<`h
z7P^pII|9$F9W(lUJpB0Toz_FYum@yMFD+`I+Gv+%W3@US5u#2z7ewoE+<VrpT|cji
zeY)y0%si6#rT9|H0IQnAujeR@`D)+OV)C~@Vx3e*YI1(C`QjQP6?wAtGK+>(E>}im
z!*(0sAa4)n?J<*)2U;aIP!O{ALs`e8%a`VU)^Xkckh@KHEoi0owWrm4{4xRNgU&v`
z5nvnMQnGPTa?Ha!Lxxex#U1{<sSYrASH2<T>T~a(v1PIr-=`bvG_QsXre{ZK-f#hO
z2MJ@Qh~i7ass<VN({>N4iY@1-oA}NTMSi{hbNRa??D_d}IYUzPj=Cy#P!C&>hS`=g
zV=9^74A=^}fz2$hDBLtxrIqe~VW;xe9`N{QvS*L`@wbnkXOqi(4=8}TYJ<Mikh<A>
zqC7%6N0lh@{QU90`^fB(nxPT?+mI5Bg`l~&+fJ6L#)(AB<&9Qn)7!ZggG&vSoR8A<
z^YoIrpT&gRetN2PVZyHX)Q!NWh4m;da|;M*DcrEKLx_aSso#3#&Ir0&r2orP(I4S_
zzHgr)X2DpbJ4<R1hMo_&LjW(9oh2Hx7MBss%^SyG>?pTzs%?vv4WMohtq)3^ZRPNV
zXEgBhvh@238f_&6e<3PQ1gRAe-nR!h<(9Y&gXT%^(K{Ke2^y6zrSrdwC?&8P4UDGU
zy=89v!4R|Cntcj^KlGpg-Zzi8ps3Kt9vHUW(s%;5A%Jj-WL}R{m^YaRC;85~vJq@4
zP1JYzuKo!Sc?G(lR}k|S*|o*WraWUFP40(PMTHFdl4WK~XHkfV8o6!W(SuCYQ^g0P
zWSjHbGU~lSUtaA8t=+Wy1a~PM4v=e|A&cpppqhvWM5>w2d-Pvi!=LWq|HacxgxXTG
ztcznKy)6eGJT_Z-M%X3ahf$HFsoT_OMq3#G%89yqrfdWn=E#!gmhPV4B-hV_|9r%P
zkQ7rGyhy{p_;wVo%ttgs9S9ErvARGCK*9u?VUley@;rUP=-RI)19$ou`5lcCBoTOO
z34PTuvi^L~bo^`GxR_okq=t<JWjQ@Z1!7B{G^9^oc}|6SXoKdf?u}2=8yskKlcgXh
ztiMGFqS@4FmD(o}HA{ZmNKZb$My2}a3bhC#Jq}yjq(&iy1Zw9PgA)~i1A>lOfbFku
z#}mpBztz6L7UN*#KXU&vkkmjKSzczfA~1h;gKG|CX_p^1_Hum(!;nSu^6Qjt&6BR5
zeo1xGd}<LqB}LG^UH$J3t*v9|%F5c<f~?i0zPOvYj||l0`Aq?w`inHW25|#;Y74^u
z_;H;+@R^MJLq4%ae`G(J@NgbwWSXwtWk%V}TJJ_W+U_G?;MnFHEBaotK9p)04GQ)t
zU7x7K-~R09Yi=x!wYgW+EB1oM%!K|@=y@gY`ndn6q>@TS&Z?;BVK$r?o{fH@42tCX
zU1_1~Uvz8&vMI~2l%|={e^BL4Lk)Ep6*-|a`eHP9{4)`MnS1Ii@YIp$z1fHygVf&4
z<fTUQZPhPetMS@2R}6!dK|}h?$ey8sODRNLsOkcrJA0_pvGq#qRA%AOq!BgbjAFxN
zSZ8F)JwAP+a`bBIC(3!c*2Tdk{||+jZSQ9ZZC!Uo(^bc5<|%+qY65e2o|XcjqxMMd
zIf~${VMBg*sJ{6E1wc%*K@i2$;Y1Q_S&2$5@E`Q#5&hq2{0RHoP@Zx)L5zl*jQF8&
z`nw_JUoWdKL1HaU(;m)U;QQmXm#S<))IphwbN%SH<A)u;NXt0+#5)r4H>kibrk08V
z@4TS^LT^#4ObUNd!LpxMY1#gwhJV-Rxt!NG{y#E?LfPf9aXUk+y$SsK?E9)3Fm(J5
zzmS#gt$~Mre(RmyR&vC*NhInj9O`$RqX1akTo8M#Mn}|Ds-;8|&@6>7A{D%#0I0P#
zW+cz&mncG&@ngKuHHK+HR^~mLfgvdErHaLu48rl<a`AR-q^ux5$;f8MB2xGdm7eh@
z)roaSrn+Jz1@PYJl3$k;a7MK&y{<!;>xtpxd;W=C{7>`}Ew00RxpIrHTeBVum>4V`
z^7Ka(U-zENZ8-R>5sF;G;qoJsb6XH;cygm?8m_y8jf1E^DK<_tESU%M?S9sOdCU>|
zs)jE|?({mKJfL-k6iVd7R&^|};4YA_R}T9aad!I}mmTKE5=)F^7@J)sX{8<d;&cE@
zF~k-eNPAdli9VtlekRvVP|?lkT$*tCdh-u!8qqF;*p9Hxgt!LynfzN`-(nZmwm+!{
z3p{sQyf*6jBynOQzVBd1ADg=LKAOz98Ce7cKS?}q5MDo;|9qnEj&Gv+A}*o+Svd3x
z!c6lV(Pg;xHk5U;!%9zOZ>_hb!nbc`03iMD=t8Jaz3A5-k=Pgc7A0r+FR%who#c`k
zP)nQHY(XyBNbR&zbuH&BpWZAQGFBH+)sexROuubwaeFUjT;;%d5NBqj%N3TT@Y&r&
z5Y@N|GaNk@$*(rn4(vvwwYeYB{BNA}S(~NiX4q+AZQ^@spRcUcibxo>SMt!%uN}y$
zY%8_&21fzJ`vjJ$Xqf+ij!1BHD@D}3>vVNXy>VXDIVrG#X$8HXeuA*Eo#_rwEr2YG
z9AEQ%M3lse5U*3yh!B2k{4AoH0;upNya<_IKDP7iWI{Nt&VKgpJsB1jTR!c^Qhj?!
z<VQ07ML01=g|MAdP&zEE`h?tJ`_zR6qw#_C4wQCZ^PxIQjpWI`yH^WY@>Kx;B#V=W
zvo7P#6{1`C+BAi6zNyasOV^r}12Iw-)t?(``Rsyx`+Q!`rPR*1pl4U!1CBpgjn<Lk
zyk&qGn}sXYb=db)l9NK9<lP`4uumDo4CvNUym3O@RLtl2xiof8UhVee%@-knvb+FB
z%~}aVxh|{t=m#0<3}KVAz~NSE_8T&NbUc`Pq_cB=TR`IgK{i07cJ5UkibH~R2T<PI
z`v^%4Nt&RGe>dE0UFynP@jHDrJ0Dzj-Q_mv;D$`f_u<^0cHzY8;!jHnlLK)#Hr|y(
zOIalG<=-a{5KOk~FgBtqOf^b+$A{(0&zmpIz!^~9x!!9szMRSwfF8g<CO^hKzUAVn
z9SbSjM~-y?s?%+ZbFiv=7sS6S`QE#Fl$lSSi!KEh{B7}9-_0n2OmhcgUzz%-mPL`*
z2<%&`<F2AsnRaJkXeR981oir22xwM)w~~BBdbWorrS1@?QDgQW@xI5wBd7slNiD)t
z2n5<BpIh+TTZnsqY6dm87^MJeJd|PswTB9@q`}BiAKM2L;!)(E^Y-BINKUhlN+E7R
zl{y(!9~YV@9kGL$S3!E=d%D>27IggO=P8P%;#HkK+))qonI}dKx!;B|zMpIMFiyx)
zL&B<By2HKW0ai26VoW18HC!%8jb}L^fH%32_lFIgucK}Ftfn<R3nEdzA5OL3G)C<*
zuR2beQv(DC%b9%K)!G|5hP|hnsW4#6scN>$`dp7+U%kT>X69FQ4SodJ;zLxBJK=Cz
zUo1Hbfkk0K$39;rw~g~D05?(`s!da_Im*^)A6c9FeP-rZH9fiN55XCYNt)JYbTQ&S
z7OzE=BD824oN*WpF!HuXVGdNJGyRY>-#`DkQ8~Hg+n}RzX_wHu`gH4eHTNp70sNN{
z;;OIpN(zW)$y6`<4n~%2b;cXT9J2Ajy0g?#HNjOR5t4IX-F7zY^D)fZW9bu4ZVRqb
zQE(T0u_0$eUbl2pu(Z)RDoAstAnB#tPV5g(E2d9x`xuT*OqKy7L2v!|H-Dzu0*%HI
ziFV~HeZP)<n@(FcLl{MF0kuch5!E5F)Nom6P#;U(gCMuttOpsE8Cf_#YXp{x*xuC~
zaY(y&Y@+w-d+T;=mFHEVbOQaUfh$>_V4DxxRIL-tUrk2W*MPVQQCCF`gczEWlXE^6
zgHYTOMzUh=QRhe*Ln`s@py?c7m={#z>%{iX73{(7t!oh1{IWb*Wucy(JNM0{I6GIi
z7-{G@jy$7&Vqsfv@}X+&wPeBS30@8AW(lMM6w)!FGQq?Mz7<pI7e$+W1?|_HdVExZ
zt~3>Jv(m9?Y|UytYz49Pmo!OS-P@L>HsMkLYA~d*&*U~oo~xh^Jh;q75dRK;n`JQa
zSygM?Tz|0rnYU8yqG@9ayD{7!a%rzD`*h!O#+?Kj2@^t(Y-+|1K+LlQdxJ3<Z|feJ
zesn_vqPgx5|A)t}%}p<h2D?_~IO^WKM-NU-^l4#_qLtcI-+E8`#Iw#reHTuVn)Edz
zYxgr%1a=%O1=c<Q8lJe=163ZYJvfB1Bb;pIFpmjul!LV$tH-^C6?J)c9TH1TrsfS_
zUHEk&Il*$}(XhheH+KE`PR%Df2zWG(5CkvJl(^kK2qRFVifyu0z~%<3S|SqN&(~od
zu$oZf(wGu1-QN}MkB3y(H>LL0ICy7%nHBr`LGi(vWcL+7zc%dj?##L-PvzL5$T>o`
z;W(rQwq;<%W+(nx)lXGI<)gn>O;3%teC~}8M%CNbSt)>&@9I3KiglqxMNEc)KoA!-
zW5Bd=!P?ap4Qzpb$m{6K8-Z8VQ<*L~nX}wHHrgRpr6@N8Pi5x_5Og^l9c^gd{;RQ|
zbS85k3HZ*$i1mehDa^@GzFhHwzjo@%XD(Tb#Jv2AI|#Xc5VjTJwIgK-Y9dw>Vd#^r
zWMHR(ArP%nxU_9>f1)~CL)EBZ%;O;A)HI4L{>G+PAf|N$)#RRcOQ+{@)Ccez1gX?}
zURJ}hoJg}|{VLyec+(FoUh_%b6|(xAKmp{x%x+AQ7deRkNM3tubt3YEXBg&eXpMfp
z`IK5LUvCVH922dcW~pf0_qtTO&z8y@;xQHtV~lEQ8f~`3@{u{?35GCCARa={R~O)W
z2U{A1P2+=0NTQ^tsO*~H0nN3AwJ@ny4`#iIFl#v-lBu?KQL$#@iJjD*S4eSyhQ{Ni
z?)TO8K?YK@Zd|%w+)g}VUk-H(aP|kt{2_PI#*u#Z%G|^A$fp;W37@m8*CG$3w(lV=
zD#vz}9;v(fyn?$vvJC###iD@i281~#n5N(Cp%0?&i{SL5xh6ar33Mt0nU5=nRp3f*
zhrBIyv(l^xwwH)svr+DUQ^Uv!xpl8&-I-r%C5735zPrbTkc&5JnyB9F1IzIt9VUx0
z(Sh%?^%dDfEmQj)@5pk2+xsr^C-R>=Hh+1&`XQN`go&HM2t24j9>cC-Q`KhcgDS+5
zeGLw@ExZV$heDpT2)09$j7L~L`psO7RN}KLm3gJ~sB_g?Dm1UTw~Idb%V*jPz-qHt
zZ~-(5!v6cuhK?EyP9e+HVIRGxi96682YGA$kU^!2-qoLF>@<59o`sHsW>ni7TZ^%G
zm24S2WR>*}HJG)<Yl<_IevXw{7Ra-Hi;CGEI8)pUaQe0oGDvm4mv9%qVs?3TVVHgs
z8BHN~1Kkla(!G!rjxH|~RP#%5a^h<6laulgxjJ=i9S*M9s|=7mIUH(N#e*|o1Zwsn
z+?cK^LN$fC1Xzdp4E{|y6l^-w+Kn%xHQ~($Kl77Ia_JTGWAiYxq`A}e^qXnj$2R4!
zY~4(b*zwlfBB)VEx@L*bH<pW|$fF_avX%K>txN2Uzb5uB$0#00tNh^D`%%o3%+>~B
zU(mvoJ+{W#r4~aGcGWebS?lVW=TcI9cSXyZEDh94*Dt3(lW<~gBR{@KQ+^#dh0xDH
zH@9AFt;Zhrj}YFIISthAV@KH7X8yRm7xZpq!TF-rWv?D_X4}5-$1&nnh}2a$0d!!y
zYY<Fb!l`?!+hqtg53FvybkHksF~3ZC_3XARu2fn0Ht|x{$jeBH?yCdu<$Hy~b)IU)
zz4~wq!A$fU4l*us@$4?xNU;J%2tFA)EUA8HwGpo5&-RhiEWuCrLZV!HWW%GS?}|Zu
z1T9x=Qv1=2mDU%r)FX(Wr${&vcrNxi)AcX+>tZT?n2IPR>MbYj+q=O(6R(lt4nSd=
zvLiJtnsVn0GIRY=)BBbLl@S*Ax{@oTLSyOGdvD_VFL)|Gg*XYl6%GN#))0Ls#}6g#
z%onI2j}bvU(;ojV*Z{2>qYgqt=m?<}-V^}aJUCr5AI3t3)T4A*g@VlQ7igNh6v+Sa
zNM%g=Jnt6FRZ8QQqiVWHv@(mwzD3n=(w(@4&a9a|cNvHof0{Vc%18QFy;APJkWV@a
z4b(2O%KLRzsWH1XM(i#)8$vX4vQ;E~3E3)SD{Bp}el55|EZwT}V~S4>?fBCC$+-UP
z+z`-N?No%_0-RnL<qoeoAEXV&2y|wBoWSV~t<xLsC0qG3o*T?a?2m7Hd-N?;X6sar
zBn2X-+P2)zwe;HTBW(r`uMj<4pt|^!ks!;`-?ftM3lhWVab>1V9=XG-GAZYMq<qo}
zDua`%QjH;UyBt=W=4b9fOW4Un^0gI(7PbVo&Xaqa=b#4$Wz=mc++(Xz0+M@Xik@qC
zwlA0{L>Hk&WEtqA;*ALOc%C78{0;@M1sB-Rq;>#agv_)~ljD!bw%IREiEDkQxqcT^
zLlO2}SVV^i12k*X3cI~M`QD=-(?<=)bl#`M*lr>)2;?U7O7tPM*Uem-gT~>sIkrKq
z*LM_G_j@O`ppLu2TvZAIQM+Zm`#|QT$vTgkwj<p1e0FJPp}C#-#@2*l)v$wkyEWbL
znfYzoopR?;E;mtsaaW0iGw%}r06a%qwD|I&stXdn5@5Eilj`yqLpIZ4r_rS*EUANc
zAGFm{&u(oXlAxZT7qjxd_wuuqdZ~JVzX(f6PvyFuP||}jb-Owf>)+{i@dXQ(t}y37
zol&?`Q+55*Q9ZGjBRc#+Uz@99zhawC-ftb1K)<oLopxT7t{GD|`ex&Dx?8<{xy=1i
z%_v-Z?q#>f5$AW=Roj7)EF1POvZJL9j^gfIFfYWK`-$(?HS5e-cc3^d1-}Vv-U<@N
z*{Zwd5K;nRESpv#b?D=Tk$q$x1>k5seB+JeeKY+T+EuZNWQZ!Hf1jtVJC4Xk%2RGa
zh1-d=Odg2vL;E_@Yd@=N1v@*s=fm+nG4I{<-blp$aW?U;AaCMIt>%-ZcnCK8H4x2*
zGmrDht|JwaIfL&B?TYqJ;~}yGOzs(V!pD26&P=gW-2$3mG-|pVM1bWILgeV7n8^p@
zGU{qzblHQw)BCE%3wJ#nTneI;>tm1#X$43{z4{>T7mU5NNnqbploN_<=(t|zjgpgT
zo?L^^!69mPYtCPYUc=3nOYTz08P<$%V0kw|ad?DImHPz)dzQ@Z&zJoa^eP==#N3}a
z9gi*1{H*KD7FL2VxZOT5Sx+w`XQIbr59jUrFRRR0cyopvD3yr1N2ciPC16V#)}kQD
z>`)aE<{({`o#bHA_1Xpg(I7>Lv(nR^Vpnlwq}81eCm1lH+ADAEGb<BDcPluDu!*@?
zgMog<Mh5$T3DzfSEpFO6;U%J!qkTT_-?Z@4iDK-%&{yxlsBm|}rg>u)eB2Zw+?aqG
z4tYQU#1i9{Fl^mm*n2OxI=&DVoY`0O_C@jRCpezu(lb9t6y8k#NO4H<d2*?uF7DAS
z5hxEQD#)|qZMV6m(0A-$r1I}Ar<%I4!`SP+;?<-1^`<U~j~uveP2rmOa}w+=c|I5Q
zUIgQ+@EX)m6j=;STOKFg!ppdgyQ6~x+mWWN!x3+uiA?X;{1mTr<h_@3#jonakMFY!
z@OBYSmj%sn<#-n8w2~p4O=ETIgwo^)CkCVcCF-Wd?7+y`FvlqV1Z3wmbd-WT-5)ZA
zs)w5^;O(;Q`uvqy+$f~Gj|$>qv+<!#V>Rp5Y@F+jzheC{{Sd9u<~tJ;xGBZiIh$&g
zx8O)7CMc`O&#MpE9pn;hB|F&S1XZb9xi<TXVGtA*?s1)YvRi0j)g}#EPaT=LkMmB>
zS76ID;9)HO?jg;jgnEsjHo#-4RpXAMU*uj~7S;a}_PaCi@LE_d1yFkIORC!4&!*-h
zb0QqAXQ(M93{z%qMV>#UpDn94fQfE}I1O%*^mnV8ehKWW9z%c3#-dc`b5xGF_KbGS
zVZr3xw9;e$f~Bf)*dmM=mW^J1R7)!IUMezlE`o}7rYl-(Of;&%Wwy+8WUgbD+hv+p
z<JLQ(C1<X*z3rd6;GZD{wjkgjXi|ujt!`WE^~#~V@hiir9iuaMQgk--@5>j(7Vv2c
znmyr+P_L@L@x0LI36AdvXq&o`!jOvG1unzd4KTuzc2tBz1LbteHihsZe}|y(axGmb
z#a#EbFr3I%z=9ZVWOMR1)jX=hIodG%%2K;bOql_4vN+k09%a6U<EKJa;!h%dzZ&m9
z-ML))wZAs?pc+E0%vcXYCLB^lJp&G?F(b?>Gdlw48Y^|Twc#2X$hFaUwzi7(jcTOy
zr5crMK}XkK*dynpJH#r_y3%RWn9)S?m!<p%v5>YPFglxSju5(pKBB&};PS9+vn#l?
zHPs(2@Xk{bA7fspSpHp11>q#_nDA^ya;{B5bK8&V0xJf>3W)|89cC$H_EWl(CGcL$
z?pt?{ny+dGYwDgG3i7iYNeQP(SI*q)sW-`&QWh|9M$EJArv@rJ`=-BoJiltZu(ObL
z*czaV+bR1k!W|;x3Sw3+@vqUZc^}2bwsEqYVilya%Z!~aWx4j808?TB#7JGw97N<K
zK+uz2LUIuS1BFz-@pRMf!dPjUd(GL}2~4<ZLv_8vc@I&JJ_()}lQZy2ZK-hW(!rQ)
zIvCpBMhP(snzn9ME+qs)rh|5=k-r|_!09F#-ZOy0%#V8xEWE)^p0Z(-*P;;BLDX8k
z3-4`_Y4zBqkbaqtEDut*TI+CZ{gU(Jtn)JqM~9_|Q&0=B+-%u~TGD8XU;jlm7$)da
zfiO2Y!HtDm!}bkWvpVsrd1ygzwXlQ`U)KH>U&`=Z*`tQ*)L6;%$Uv}a&E-XRw(Wa-
z=&*9M@Rc=Wqh=K*WlYJU`Oec((Kik?m8qIGiIMp+j!`{lE;~!8E=BcWji~!U{Lpd{
zk(g^i`YhYAH3Lna;OFVCQKo{Kzf?J;6OVt1^g%zJ;qWS_5iqe18$&!7mUWkGH@HCN
zo=}eELk>Z`0ug)!O&KSKXYZ=ShT8Wphdvqss?F<NQm7reUqJe7T2qODx4>-uL-0x4
zCd(d%e!d5*)GCbc@x?)mQH<RM>~8y1P>4j536?a*TQ%MC#We7%7cRdMp5F7#Hb2Hm
zN85J#`4>*bX9V<%R>8V4A2tR;lJ2-i&E4D?XvhhZPjgelRfjWtmgx)k<-=j}PE%(z
zEqJa}@3vm31~G0Eg>vmIT{qpGSw4E%tn=<$=|ylWc};(bQ{;|ftoA&HGT-A~jK30(
zmm6^+K;+)T%hferdADAsrVfb%!|F<Jk@%$ZLD`p;>e9MD^w?A~Xz_WWKQ0Ms+PPYT
zO9E^|sMQbIW8Fn_LJc*ebg%e52@<Kcwmk>V&D-SMOxReWmKNSQ-9+Dh>vm3F;Ri5q
zd7puESK;`Xf&udd6YSdd<vehDM}RgeoT!Z^*$~3im1W|JS?Zx4>22`0irQ)=x+}7(
zxhfVhXEfc`s<WrP34!gm9Rx2HQ>z{<ghGQ!1o~2C^N_mKx$OLI`LMDXiHfK@Blm}z
zqYMTd`W*BCrWe-2$&7?i8>%0dQ)XlDcE-f2cKHV2Gs<mTWR*5+1(j3=O5E3^lrB4_
z$k%$c<qBm+JbmKc4^LfVn%zxBgJ-t$15bd6g*x&qf!LxW5}&U5#eMsrOvqwFIWX41
zWc2BXyV<;C(iYbz52s&AV$a2c>!la7<uz?SGn2INABS_}k&(Fsa7WhFF`p?PhNq+6
z8G)Wka`w`mww?0nF2lkX?={zs7>8Df(|THUZ-Jf5TpEq=ak*chsU5dHN=i25=Lqn2
zoSRipNs&$3N}K^uS<~^dN>ActAtV0hHM3AYd`K-;C4Vux$$Im=IWpb1*a4sJ6n!q+
zU2UO{t`yhFIu`b*?RCA=<u}e$hXU?fJK00D;}OShYURPasW6}UR^Pe}71)fuR>NDZ
zR3up#Cl_-rKL?2s;JQ~Cr!W0NTufKGSgh#RFS9#W-sxYH*JNLGe%wJtE0sV@3*cyp
z)Dn<umWW>7qz~94hxg<Wv?F)GJi+mKg+fpJ&AD&i2i47ms4-GqM+nU*Q9X;5@P;=U
z21fURm^b`V1qJ7WzX|#JhC|tZ(J2<EzYtJgM#gl_`sd@g@M^=&?n~RCj*7x}D@isg
z+r|ADqjNK0)gGKq0;DR-jDf~w*sD@o!9yrC=7V=o^NhRS@%mof00rQc>M&nBg7QKv
z!KAF$wid48CbH4OTJ3m0&pAWMh#KHliB6rVqa`&XKK-1+Z~Nh9Pj9cCeBw>v$N2PN
z_mDnF0Tm6qcx8MCerSDODfXsi{x8Jz=rNxOJlj)H;k{IiRvQ1!%Qs(mBe+9^?4UE<
zNE4Uzmex0=&{t35pNy_GepmHvnh<z;O<GFqj~@F>T<!f!VYe=^2Q8VBQhhYE(3^zk
z5DtPL$RO7I@wjYvw7TfUd{?nlKe?L;m$|xz`#F3~Jtam2D%095Y(rrA2+*d<q)aX>
zn`{Up)uI%?;DFl<UkR#@#^m5%aQ7>axxbIPb!~578je1lmiam><eoi<>PV<)|0X*=
zUoT`FGMuhI3~{Q*PDZ>SYJ*E9UaLz<`}P<02Vl!OJiqGCoDt;R_e%>s%hHhQ%Jata
zRJt_9ccK^5RBXl~ikx|rjz<lq8*Me=oA+&rZ4LSJRs2E$M3iqykQu>VGHj;?A!NDW
zR=II>QL<{X)M><9a%>=Fy|C=<@kB+Bcw<;D=Q#_{OZz(XJyzv5kuBa$rwzG~w!Yxf
z7QuN<{ODu?Y<oUlkKl-VK8tZ!5UmwFSA%ZRMIIFwqNc!}?Ui|DbgyCy1RQj26XCx~
z@80UtVdxzutIcj)ygsZH{Ac8W9TPklnf-lIP~}SZXrU)Ty;D6ut^AVx1(O?m9NO&(
zZ9mkh=~!<bAx?+z9s}aYs7>k!*s!%S^vi2a+}9FLu0ec~UbdG$Zzsu^m?SZ7(y`P3
zDukXCpV6Mpp)N|@>hFxKtFseoN<sTb{gV`}SS~<))~n>C8oq{@pIC14dN61j7zE#|
zA(K;7LT(cFZLo-V_p%x}PB57U%RkBs{nPxhYTVZka(s!M*B5VTpY2G9m{Qc`w-*gL
zaPPFhns;P#Y%C}c%<W%cH26T@Ck}+2@EfPduXcULE)rGNu1n9EE>A0_Xg~j&8E$bs
zeTS&MxXX=*TA2(|#2k!jj=Boxccqq2^%hw;srXucjg<_``M7(1S>B8>&88gu<~`nf
zQeDc2x<J~pWJkUG&2<c)&g+-`_5!u}OXR=e=||>y+WvTU#$sTc`G?fg<(Ra>o!N}R
zy<`M0f<;Z$S&E7tyOUwDhYxy?gc=-^X$9fP58a>+UV=0KP9Y0bj$N+98ijFLx=JRl
zZi~6b<K34o%`!i`JH;v_aH~AJV?mDko(^t;vp#k(O8GRK7g}O0XqQn=NYaiQee`}c
z(XudQc>Ag8-j{IquU`Nd_3-&5uYEMA14{G_0_8*3+9w)27b3yA8`u2NYH8amE|T@h
zj$wQzIuRn(?562PjjTJK564;o*%9O`!!<@x`vge$#EfZdXR{!XDEeHAt=izae8XO0
z*3ueBS4B)x?7&^2?>ulG`gzpI_Q^(Rqjg{I$Cjh3c(cP5gHc`IBShK{=;%^Pu4+3l
zETFt_$r$@H*q{S6Wp{17t>Uh;I`3rkEB?UxyfB)ET7bNrj(qI+bPx@0<tQr17DKw-
z*R_U$y_I&~${!_K^t$-X^*HhI=74YJ)@AfDFgbE79M4)y<|+ZFA%@ua=(1h)TSZpC
zHF;}rVww`Q*#4AH3Dz?MO2?(EdQl4R^woQoe;6evoE)ctJj>*osc}>yk*fu<m?WDJ
z*9+SoA-aQGjJT5UMI+Q=QAjsrF`I8eTeiv{J!>%K?jDrU|6sa+#r@*uYt)TU9nQwm
zJ8bFqFTWm*#>b589vJcl%ON!{j~|lER!>HNOf&i4HrRWeuSM<lyUOy;n@Q3b_|IB2
zzL2RmaYfL;s1+@=$e&fMG*Z%%hE^vEw-&Fu4Wd@IbtwSLz3o$H)S4z}UZ{THEk1BO
zNW%wQO0e(rZS;$h@iBDUlj_&f1DEL5j+M8L%2a=s>3A_~`Fv*c`u;^dBu|kwRne^E
z#;N_O$_pD-yGaPNe@K+vgwYDMm54rP-X%+<ap=Gn(lJw8KdBE!lQgf2wwl$zWVa0P
zk<X)m5`T#DemK){c0M7?8V%~+oHs{;NhJkfK~!Xo9Q)?PgrIP<%J=zBSLZ-6Xp(nA
zM~>Kyug{k&9NWBJS(zcKsNFN+?&`N-Xb3mbeu+i=P%_(Wyn3+K8bk%SAIi$n)#Cb@
z08URFX0J|vr)_pt&ni9@wAvJ8jZ@Yzx`We=IH0~PYRqJ}GlVzVcy!J7-r}HwzVm9~
zuTm~>UWCqQ7k?QO;;?bHujO!|W4#XoY#%{kXN~H&4p7k}Py{9wIUCHmW6`WpvTn7N
zI-&R@`uhCzoRwC^0E1cNg7;->*IU3FjNmD7SF_xyM&Du*S}2pvY((jFmv!B$EXH`2
z5FiES%fefT?`zjIbrcIK@$a?GwY9pcp6TC;&dl|sTS`24eZDktpE!?TV#?u-x>^LF
zHZ%R(Sqc=U0D9svBB%5=u-^q`Kan^TF@j8u7{B{;lcU@K2x=MpoMaqu85B=$q*~iD
z&;=qJ!4m_!K;$C0L*USi2QY@&S=X750Up^o`Yf}b1SG<5$=Yq)cVvyQesSHJU+5Lc
z-EBq}KaqXmA|$EiXlJ&mf`3B;EJS7(ISHXY9Y{*%II<4Ork0+5`5pNiMbwl9b_7Vi
z@evL|nYJ{A?slh#g;#i%K%;hkPV^auZ{VIA-<Hm{kGL6Tr8jhey|nGlHD@DSZG}G;
z&s6rf1CJ%}LU@DI_pOI~W}7-IN65b{;P|``?b`y4j0I?hsY;&xb0g;4Mua=C^-~Jq
z5(R*#Zb7GZ(HNnQ>8U7Y&TT0TU*auTA1=gC;y7<+<cR2x;DNRITchD;HhoiCf!SmZ
zHg?cPfbd#}6xgRJ$t3-^-92@6Yx{>={px|kdQQ&tZXX`fd^2_v)QxF9OBfJ}7bGN^
z6W)5Fg&#mTpg|kvvMF^@?<(~Hn4Y}XT?vIB#qa$n{c<6gs*&f}NI=-K2>ECHKN4tQ
zQGaYrEh^Lz`SxZeWffLFJETkyg<xff5)r{5k&$L?42yO5<|0JkckO~u%F@FKn+4Z>
zFdr9fo@Km(8oahh(&L7$!?Hulx*5^Zo*Hn6pL8Xm**FMIrr7wmN)TpiY6ugd?C-tW
z(xQ_fJ})LClC6+qo&0*oY@dJ21NHPDMU?vwlgFb$svg}ll-z5Xw7+^<VZ04#gAYs`
zXmSuTmuz7MLqw5MXl_T61-J~|Bv6Q*6C@41DQa04&r2Lo(g`O!-_!zM;koziCg&?Z
z%u$CC?>d6L+4{MUp&WAZXDtF7O-&pLS@?Ey7DjoO>din)^GBNg3ASF{uRj@whEVx<
zM^Da1?6FbNLS5n|-~ox+gbq8sKKyfM-3PKuuG`YgmJ$(w?sq2rhFvVBBTFPbYOEhu
zB3#1aGcalyLZ_w<u@0uA0?d399aTO1A4<R3oI7_p-UN4l03{M6%#B_c(V#W0uZ@Iq
z+C<85FNvLLSA9vNA%I{D{Ksco3ZmMYY`so(TT)HZnObDmwpI4;a)z%6TvX}3rDJv}
z{^?!1Gm59zHQ<Dj5TRqljbKnBQJLtgd1uHcP~+TayP)6O=094mj11)3pC>qD0;T+Y
zG$eqx{9cH@7-}i{m6qM#oqY73#bV}tbWJwXv_(-?JF?yHq7?4#7(^M!G3UbD_p<RV
zDO0Q^^qK(8x#!&5=GSR%(O6R<>wnavzs@4t5@hcX@n!cUgpmrXqgR6rN(uN?V0i0U
z&FeVkryFcV>Vofy5`1<dBhHZ$XXc&9S_g%HB}RPm4}2uu=I~)h$f{wNm2e^qen_wj
z3dYes!s6fDz$uxlJ|ZLrUiAnWgmJ4B%lzJRcv;-uGb?<$x5b8nVP8`KqetYsRC`jw
z^oSrz?NIO2!AUoM-h|~_*;*n6yILQMBkv7Z_=xdy@r#x?T)Z*a8L~|_3?kiq(F7*|
zy(Y1mc0)DpR=b7-EkBw5!<9B%EGa*19&7LEmv^j+xh5tP{&Yx8t|i?1e0Ng%>07F=
z`e2w!dTB(WGiUR``i>lRq}-l6d6zgh)h4DgT&>j?(PJL;+F7mVQG?{05AqFF3)8ct
zl#%nQkiFr@zVe@$oub#_hm|C`jcE-C$pf_sxgS}S6ZL=jvad@GWWlk2-xndfMgTw7
zrV`)<Yo-8B1ir9+LM(m-RXV{>sXL-at5IwqZP&7ORoIKVxRr<~`<hhn#X0#ky8nZ{
z_Y7-lP1}Z}fQW!dS3sgv6#)xH5eVu=L<9t-3sC_nAw;Bw8bkqUA|N0jM5Rd=A%vPp
z2NCJLgoNG`YJd>m<(}vL-tC^UXZGx8j+y!V5f5FidaZTe_j#4`yjHisEf99v;nJO?
zA_6zk$OIp_U8flLJdLIhd<0u$tQd3SOO|<GQ`Du=T~+7hxAUQFz63&H%zB!quLRYt
zJ<l)TUeCfC{SMG$7UyTPJWLKQdo$8W=TBdAjJ~0o5op1C|JKOy!A&n`7yMdfgxpR7
z>Yn-KG5#RO3JsE0Hk!k>@MCk@rY#$mi*xz20e<2+<kHSlA#^Il8r(N2>%S2vEK0dN
zXzzPsMJ%wjQT$C&sJMIP5o>@4<elZzh4FPrFMU^^mHp2A#Ji6M@qP@Qxw7l)89>RR
zAby)p7_(4O>dNe2wy$XIF4#1*d}eUGLO)EhUBh~S*4vwp(&os+BV_w?kVXd4B*h?U
zh3SpuF+*&yuNf>*`8sGiN*MPHH30G;%=4$*o<$d7gBI_ECrc|0EYA{ye!A>P*^L@~
z_?iZ*Mq%VH%$p^amq*Aw;q=HEY??lu3|qynAKgiE#Boqnr=S$3{C?Wayc3Jq=icjX
zH8H`Jha*1M6!Qq*>gl?d-8&WX!8F;;&eTLf`=o?OzVi4h%d|vYZC5WDTu|pNpkyqA
zbo8Gn7yDLf#*ChWEDMlZcb))POH)QFEe~b_d@;R&sCi(0+R}6~mj?0ZssCI5c!S39
zo1C=E-=dEZJ3QThrrCwBph$2nWPN|hS5UjQ8vpd$5D<oA3zN|aKn%m8EfByq`dvT&
zUyZ4pax^hl`<JXEePN_>i0O*%L2ug<@>0Zo^0UZMFgLaJ7Pj+Fw+#5FIPH*QM-xm^
z!cLc0_{B;yzB+b<&ACPRa$%tlo`_fQz}KPr6p2%{W~FO~kLix?=--ZopT0Fv&b^Vq
z@Nw(tMGFHkjeF1OW3rT^3ChdaZ~UCrwlC5R)+K8Q!gd6%3MqKrOyrPRooF*-Ph3d0
zN?NVH+|7%}^qX<#IolTQ>l>Z;yjL<c&x4qG&FF=z3Jr4L{!4q`FEQ?j3y>Cthyzsl
zZ7@xhde(z4!^TDdo3G)Yket;lTH7~X>&K*P>0fhh*n)5yK2<w-)f<HZayXLuU6@1_
zg;s+|_!Fr^{-nvPyV%9DKOMAk0}q2`){nvpByGi3d2?=AvNova6H}C5>4@M?WG9R0
z=&!7n(GC_FMcbo_?#GDvvY}Y02q@K0+2pQs=1j_EGcD(JTehR)DC})B_#{bLR=W35
z$Sh9CTeFB7ZkhSPt}O6z){J8K-OTBM_dyGWHyKTMZ`X)Lp0XW(rq(37Tj=55Bev0~
z-Xy$=ijJM6AT)n%>#z@x-qnl$#Hjf(V*as_^V=iF=m{Xqfe-f9+t!>bi%lf8a#KaM
zg?T!q>*&EYmSRQ^bBDL(95!O3^TziFu@pTzBK<1YtCGi6cWa=|+3Q1umoGb{ZL_?V
zP`H=`A0q?_NXLdsVA5txC7;R|>Gj-Ibjr(=uM49LeVMrC(W;nuYV*n=<G0~{9DB%!
z8ZfwqSRSgq@5|UVN*2<0nNZ7ojm$Lp*gbRT<+b<aJ&v+rMk{jV@sfU5wF(}zrxY4Z
z8}PWmdRL-%9JW)7dC+G<7N(|9X6rsO+vvS5KvAA%83h}>NIyh7)l(rw^vJ0Y75Iqb
zDeDkTK+3O3*PP^17vy^;owI~}yD0Xktdf<8engW62o#nkNIWsQG}Bu@(WL833dv5+
zF6?Z7W9O!}Li(i|CtLW)!Mcoi2}0!cBlXvOOw=W!g9^R8MsO>-tMufQNjJA;nn()v
z6CG!y0mM`p=m%hy7pWYnGpr;;!1A=OP|t#VXp8J6|9(=%S!2g%7d`AwYP49paWU;o
z#a=4~O|{`9y}|7?K}(`@-b_MoZ9c0xGQ{U*-B*x9$Cm=9cc0gF*(+B11}_&Ik6Ass
zq`s%CUvPd?D<bV(_CvS~kXuA78_kTDy^fxwp`a_Yt+S)D+}@`<m*h$oA;kS%^A%!c
zfjM*m@_ISBXQNi&x?$9&C}Q1e&xIsa=QCPj`n=(W=gl8zf?xpmR5oNy_JAEfPJ&E3
zby4w2QJ;&K8J=<0LLu%-7pK+sb+_Id=2v>$cMWi_$2(0rgjW!T58$2K^{ktkI(5yt
zp}u8!X>5g5=6UJ0)$WW9JM0(+Cssg3W!j+x<qlsPs~Lz3uDCN08`~S_D@L~d1XbhO
zn^ICzL}5x@f-}8WTNXSoto=nIG&M_Vq3l@pB#<sFN5yTTt~Fir4=r0jw9>wUGTeZL
zQS<Kam`Z>tq}|*n`$`$+Lo9$r<HAoq|Ja_ra7tVF%DUc0l2<pHdNm*5uo`!xmKFfq
z0?uWMm+mmJf)~^W^jCt$6M(q%lMiPb-&cqh?DWwQW;aum5~>b%<?i#;2HX}F-mjn?
zR)PGT9iUcwY!y|lj`|9^RG<A7gn~n;J798*<PO6UWSRdfi2r6EHAJvE6?;k-Av$G~
zz|W@E1eOwhz%td5XKY`fT@opzeO&IvQwUK64=55_0t^*=F`DKZ@#`D>B=vFzD0Td;
z-Bfh-vF)pa<=ailiCV_V9;pLn_9tR3{aDU{R<q@H1^wpDJL;(&M3EzbI7wSU)>jZ=
zP?rwTpotj)-13w5_>Fa-iaDZZg6GGXe%R^9SFrPAwEUP0sDJS*Fu{p;$!lBIl;u8A
zJ=7Ho%htSeCCW_gP5eH8r}JYHWwXb2$MEQ>@xA%kU9y7A_Mp|6fDOOrci{H-EH(K!
zXPSKGTOb5`9s&qJvb8bI{u9dOe@XHDT#F@L%@e)j*$i+36DYb*4RO+lgVg(QUYDAb
zs_a8+C78pU`4bhL^k&Rmf3=!VA`Zn$<Jsd`MK`xC3V_h8iY~p<8dzhrkr}nJywu0)
zl!!n$P-owu*WoB{$%k?B%DHXaWRtNI_pq7KDU+)pQB+d$h{W!4E#{0XBULL<-=Kg9
zd~rZ7Y6;%!Y(K*8*lLJLFu$p9n>*h<-Weil)_mlN|Dn3<^P7Q7vG<(&B3`VU9<CXN
z+Hp03Uo1#Pp%uYp(c7OgR(6-jcSVf>6dAqFyp&`ZNCv|KDeO7;$gl8hHXa=|;q*Sw
zGrUxOi2CpX($U2V69WL%ZLzQ;H@8cb9Co3OTI4J=U2sE1PU#_%6J|1CJ4v)J7wUnw
z0>RS0VguA?&Dve*?Q)^hMRxF`A4)xBj>NOx)`@(&f7NJtud5G!dZFus0U?~QtnUj1
zs{6*AzJk)F4GSy05PbO*`)-Lg*IVr_K4md$-FeZ<Em_0!23YJf^-FA5_e>pPEjgwl
zP>8aVWu{%HNY=^v$0f6rbdI&H4*LRCMde{dvGxv&_Y^tr?rjS6da`SSLvGN!{HOTM
zEzHh`5KO57KL5fntu-rv>B_P`UI7=eMdP?ljsFS~Bh;bRS;_&=Ue?rCTu&nAj1X}=
zTt{iN@ok~6q{32=>dABUF4;Kd9VgE=%9W-0coWMg^QhX1d*Mm%KV+UWU`*aMx^GHO
zx%CEl4Oeg|DT-VtlEqOtTp$%A?w<ALgj=byO;mD-dL@Haac70W>!!U4$44fwj%$K$
z+#38`<=2-w{w&p%xl^vEBp$BMx&OVP{5}Oc_x5N%eyXvX-VC9uYd~>esp9VIMz}oF
zGv+(Xx_6bYWgU6ByH(<$=7Tdj9<Mp{ewhyWH~IU2^pM}VZ1BEaZacI^A1YBVzguIu
zd1rFp9%f0gkTC$4DpH*X4NCQ<vt2;66TNllP056Y4)D(5!@k}~4N`q6Nn5fe!kwZ>
z?1|2s<1=)Zdef3tTJ}uqY1B_ok9NhxN4$LP{kr%fLqZR)jqYgl73Yk7XrY^`K+ylm
z7_oPUDGx7|mE1~_teInZ(l#5l`TCCB^P%%#?T;j!E)_=7LLa0lk&E%3k?I~<Gltks
zG0@W<&AUc-UPoP;cy>m!!P!EEJhh%sN?~q52vfOhIcJLbOl6a|d_JF;5esJZC0{jZ
zv@D$5R3n-u)^^909<NL!yd8=OO5^?V8mr})Z@K!I&d=-*I1K<5n#uzBliV6K84f03
z4g=;fG8Z7liI>4~(0fgTYT;6K0?CE$DM+YJL3hslyd?b0<aA?n#Usl|eTU5_B7S50
z_G{{&%ZF7^DWh-*6fW%J5L$#YKRsb~Aqb;Y*s-UC^JT1#@5}=Y&t`LeBf2L2U9Yw`
zeiYCNP;Kc|6{GUk3P`+38-h8DSFTmWcE}liVBR-->-ja#SHh#+W0asO8}tHD^Yu{^
zkjkXS8B>==&6#VXr;WB|t566z!wvRzpb~Aw1k#;a_~i7Jmgz6-?nN#oTYX#3zx9wf
z%eQ{2uYe-ns!{wF8=k&f3n+MP2O8Tvu0pLILeUt3%LeDaoPN!b0U<s>%F-VLrt*vW
zlWyoJ8M?Cva4mL;0|lk9Nu^^9Gp*Fy!mPj|0xTq>VtAT1m|UU^_)Qo8@7-@c3jzJO
zs~@)d@m>ADiRQ-`{5G%s*Ub!F>rh^0(RU*k?GV|@aomjsSJ2ik9PgiF5Pxu9Wp3ir
z02xYqX3DXH8U$9My4M_;K#oPtG-$`PbLKgqF=OiUyOT&4zxNk<%v;Gl>7f8jAO8yS
z03ttU)%byw=<mws{IxvLGCCBptScXp3fWRpqJ~!cGEpCmF72BhV2TR^s*96jtl7h4
zA%Tu;;X+qAwK(5_%=Vr2+vBxw+nv-(i*B);;h;ZzhC{q&gY2f-&=2G@BE0}*vu&Go
zog~EFb+l~riub!1#f7H9q|$nwvt7NPAqVi5vImfOJ&HnZ#_YiTm^^z6k%VPd>SX+F
zyH53qX10JETn*c=GMNI2CB4(=k3RgmJrJxa;V#T;VpCAHeP0t+^}f?{r&k=Gl1Q>4
zBh2KT5$rK4$|n<vP)o{PL-Ux-#G?DRyk74y9~?JH<D8nZ6?uM?jfKgW<-^039T_}(
z+?U~N<yyKk{rM~O$jxn!ZE;v^HYrhCWKN^F6M@MUS+ACiH8?#uaNN*n_Zf!BFv}w=
zQBtWC$tn(k5$eEK&@D1;OZLe`dG}a_kF!Qs{_6yob=B!1<2zN#8P9K*Fh67Rjn@ER
z0(*;qrf_hYrl8-aW;;$s;U_{aK$x?dpSZLInn6>U!2o<Q%Cf8Mr2zn@IKQDZDJH=T
zXjU=V`UT8JJ55=t!5>1FLD_CoANp|ACiJdX#F-W(<>k5qy|BVA3j{<LcUa$c-&6Sk
zmq>2O_xHYQ<La7f=V?8}hs6qbx9%io79+)N)32Y|?N)ZGadhlRWD5-SKbH@p4*`I2
zpP9N0A*P7NisDB2Am)5f+~c!SvvAoOrNi_Xgu63Q%3Q|Il1n9KIJaDTm#DZ<y*Vq#
zx=GL*`3Y4P>KM79!9t0g5#T^OE}i~{gkaAIo|mZTl<wgW@BjEH-ux+=vvFHZ3Ue@U
zjAy~!ddHS9jvhKJR3^KEootnTTbZa%4IhPaduviM9<3snv@+*ZxfP&}$fE6^{B?_7
z-It6~AD#f!Ump!mm090}bwU8)hm9SOODl~b0suH^?OG{(Ee%<!+HWiGw~~4v$t>n9
z>iQw1Y2seKW!Cv&lAUnw>j}h(Wy^iagdzuzQ`x7V{8aZMsETo`m3Guu7(hJx0Z+>h
zaQ7&3j(Pa}Io)X8cPfYDm*Ko?p-9=9tkfjt>mR4+%7Cp%upp`tiG!VH+Md#rqbjVM
zI&E}Xx*nZrAL}M{n08&KH(WGk{DvBPu-^WIPOto=FQ>Kcd5e2q99^{mJ>Uvv&Ippe
zl3|%{X8++@&#6s`;}gkZ1`w|YF!#n2d-jbct4-s-f4r2ms%e0Akt@4-c6>l;PL_I!
z2Jp7j4xa^7#=qT<Er7~pYY3~mIw9o!5W0jPU<cfXUu}%|v!fXPO-}Z~HGl84p1tYr
z^w5mAO`eUj#RY!N>vmr_gu?uMU*BN<>|)ddB?nNw=zdBz^j!^qLb!XOLHKEgbkwL2
zilz<nPE_{+fd6+J!QUU##@tl&<3vAf^y54Ff8*wVafo0rqxemsBJdW#CImuPq0sC9
zcg6Xi1y47qT8X!)y=8xBM%_t=!(v*-HBt=EFEaSEgG{~&0d{>0!e0M|r>6hEI@*6Q
zaI~BC`)ZXBi>N@19CVXea=R5s?<IgLe@8a+TN%+``!T=x)-gBABgY4*yJ>IlZCP6l
z`bF@?o1iA#vaJLZ;8uMFIerCg8<w^*QL~MxU?R{@gDOY8P@615InuLsVVGyoZly9X
zE6)Al@J|+-qDc6qO}bh9r{|XUcpu%K^JadZyuIz*)RkAXgS+DUrQq=9iqg`RnT2PV
z0sE0t<jKdRP+9g_6!${W5-0SI=?uSnnYWSqp_z&5_8^CChD&&PAf8w=GnFxwai1=z
znm~swc5RlAp&rwmGQD*v*a6Yw3hz!f1gr^ycaOk&W3L?vH#W8B9Ij%2oc1vW5WNz;
zS#th9W7a%!ka)OdDikM*EJ0e#Zq&un<;X(=s(G<z<eD8b)6%+P8<-uef0{e4voD2N
zOcJb%x<f#<HDm(0%yKCG?5%mCK)~=?Zhn|do@0)cu_N(GfB9g!NLfzE(NCu}3_+vM
zKdu77%E)X$NZ}m5zQ<>g-yI(VEYNu8iHIizCYM2ISb};5ngdn1AeUrQx6vxy5Hunv
zo}hi{i@DOBN%@eQ<5@5JwcyL|_u7)npnH}1oieD2#J&&Y!2GwgpE@T?AP1Dy-P5!Y
z!s&rMu09(%{ue0X%@@Tk3hPQ6y=4-?y`fj(fUhOuCXftv!_)4U=9<+D3$=OcSraRD
zNj}9%B&m5;bCJaj-H^pG73C7sQw!)jFjWiND{(P+$V^_QOqJy43CV8k1|ZqTI=!W$
zo2bk&ij>2%)0A|97Cs!92p-$_NKxIw4wo9@=9=vMu!-=^2V#%(!i&CyJMekzKJm1V
zU&r|oQ8vg|mfRZPEk*J|?VYnRC)+%#IO_wu({qM3coJKJH*4l$biQEPD&jE9mKoWY
zj{)k-5%_7#uCJh_;ry?lLn+&Pw&{UDw0{Rcuat!Drs9EiEDBxGJ0_bHmoyj^C*NCV
zwWBhu>WvTe$z4{>CFj>EdB-hC1iUk!RSHZZUYOg=%@vCj-mK72nV8Mq)%rsGWDWJk
z@pHVnM8acdJNl6}Wpy|)v;|U&#RD~Jy!7|Q4_RY6GU%QCm5q(X_fIK36Fu2jGwFbS
zhyR=RmL%N_m)}k_^)xYS%kwbVaO-WkSPu~6cXauvkTG)@aifO7K$90{p|DSd78%3k
zE~GeHDhnX`iGw_GH8++MhM(658iYX-YO{E6zIw2Nl$oaX0Qhe2vmD=3is7;Akb^)e
zv-Ws6_rj-bRFT`}rCpXUwKEb+a47favOjz2nW-Y$rf_NHyDkH-PNjX0z2@+_iM#N`
z)u!>qeV?B(aQf@{P1~)UzqKCKV}vK6hZpB&X}QrM6BDMgRCd$?!G8t+xlFd|U`kJs
zwmpimNg_AX`wE2}m+$JFAs3+#ey)omH27lj%cVPaGBzB(<m`b;J_8y|x!3ygQ|)_9
zWgok;bGMKs1FvLQTi)wAYr)f%XVsP4<riP-^^h2Mq4nkK<&V9<OYKu6te&1lZ7lYF
z1%<WmJoW)(5`_V0!*>BiZ8?TBu%4hw<j&*{$iEsIN}K~^GgnX_S2Y?E8vG`4WmL3c
zRqL)<Vb%_riVJUQc=VC?d`_?~N$HhFZJo;n&(D9pb1w&7Zl`z$oOHVhT(g6yg>qED
znR3I2w)j%AsP^YpeMCRTjqqXE`7*eNb=h+^&ghnZlA656C0BdxvO<q2aK(ASrLBW&
zknIiFS5Up}uf|q_8|)L|ab+b$R26Fh_19YMos5W~X^|_C_s43><b3zy;)p8f2#5T9
zW%*P)eNDJ?hC=iKXFxC3w|zH3?x7}+?dXT|16S?iNQ!|n`1d%WA*xX|))ZW$bzc6W
z(`#yIRWI&{u~=m1d{mNzU_g?{M|AUT-_dKY>%e3mjbGUd{GBzs=3t(_EYe;JuKl!J
z?UCC+pQT|yna9_{eGG4-Eo#_CKQP^s&{8*1*XR{2!ao4B>riKC6}`;gKj4udcD3Jt
zb6n=#yMt1@Q#!H*zaNN@?}tKG^rZXQkPiig>XzS6mm2Hr8#`=3rf(Y<+GFzKf^vq6
z-))lnet)&!-a$S*AJurX2E9fH>R+`~zJKH%9n9MiiL}=Y`Li;nYkv2i4IcAtIp0{T
z@?h_jDC0&I=&%oO_}?^I+n$ymd;DTZ{;<dQU)T?O{1Nt0J<PX!JjJbd?NE#{w_t+<
zub;~Pqs+{GkJb2W|8=F{f3Ujvt_cVqYCIVRzKU9Pab5B8IZCmiYfcud0%da=8c@YH
z=0e{u_?1)N8f<pwUH%9m2coK9Q40W&1)c|B_>&-G;18@Iy}fWjcU_xqwB+s+H=k_S
znuv%id}Sn^PH*!>eMFyK*wX%zxsx4^wu!uH6Z2|H6?cc04L01#Q{{(^94ny6WRR14
z^(ef-dpeEEF__pz!eQmoc-B0Xr};J<QIiMD>@oXfZbaFC1<7yamiH61y}hrv#t|i^
zjdD`X6)h~IQc6d>uTtYk<$ToqyqRQ~xSDmK3kVSl1Mk<gV~m_7=8ToFyLHrel8ho>
z>xz#ZO*uc*&!99{SN0{EBGusO81oJiC{9s~kgJp!Gg*Ect9@Qj4I4MkC8aT0Ui&`N
z)tdd-rY>kjpm*UBxfO6BI^}nuL@TBA^0cp@hQLfO1YEv5<NQunPFi%bul@=6&GpYz
z_dMe{KF^36@vf6=v4&j>i*JjgE|9X`rMp^-T;k=tGCwNv%+KgL(o0_FE66Brl{A8L
zTQ$#l2x#Y+XeSsyolThQw#$g6`{-oNk(i&!<mV_KhL4%0MsYlakVBhD?@l@l%r19s
zi)hue<+AwcM3K%k9k^Yq3av<=s-cY5Q57k-y{{@b5XjE~%aLQz=ml|K0H|6KCVV&C
z`+b`x95#zLz0TJ;3D_*ekmuAp<-!$psrO<9+#)2C(g#fI;6jG_aBIT86ORPpk@dG2
zg6<A=!_C%9sOI^|w+*#r=xQ%>jaox7L2Zl0o*uJuZ<pmsE18mYk${@gEKMKth5HK=
zVAVY7!Sc%0p0rvMEldHYVXylLG&L4yX6Cp;(x}Dr_F?OM+^<tya9}LeoxR+c`cUht
zScDO~&Eb!*;#dVVS4H@457|TPhrUcU^3qHwf(@~%GL(8ObI9V!RiJ@Ep<2glGNDhJ
z?d(M_;a6A|x%9Kgj|`U|)zHs<6?gjtW$W*JXTuE(xub%uSRnqtUjQHoTy}S#6fgIA
zi!AogZ;IJWQhrxl&;*q}9iFmZNF|~Enl6V7IsvhyxhXX*UTXp<O2tv=cqtF}3$5%l
zLyBP<#Zxwr%YI3<^vxXqR&4$Q={H=rZT-cgnw$#l<Gjj|*msOOu?`0w!d_8jDXk=M
zxV&!AR(eVr?0_BanAaOG;aul8b(SDw>*vz`mmeU^$B^BJ#IL|{ckcus-+&i-N2gtb
z*KBDl&cfMF`^;LrbN41+eO@T$KpisBD&42E>wd#c<1OfdD;a;80_Yh97cB+w4WOQ`
zfz&4pU~t&siN<q{QPyfpulA>|@`7g<;mwG<QUe5){FRQKq=6xQ(#f=uzNY}3<e(~M
z#6<V?bOJ3oERCtM<R$ZBN^D_{y+BW-0ybMFh%ab=?!L_%ds`~_olbI@zOgwozF9sI
z0SmaWC%JI@GBVtyIrs9_-g{-8dEQ`R;<nGVIc+m!Zd#W-OKwJ&u!z=)oV!-b{ryJG
zi;CP0FapM?8jzDk>9k?;AFm7Gg3`z4P2oE+sp~x@f*LX0M{j)`n_vj(Lkz+`hHdO8
z>`zK2wYY4)%sCX>?GAs0v#=B-?Df^FEE(@<4O-E{uiKe!Y{Js$Fvs4?1y~%XwY1&S
zh&3^w?5Fq3pv4ZyGu5NWxsMO1gz6q)1++P?W0Kqzn=IDlaB^UBXjUMue-fK;?p_Fg
zt?gkaf%hh*=LfoHAxuEmC2eOCIM8+2ZLGGHSZOb%Wv;)z_h?D%AqRNgrN;LfLRQ(=
zXZO7w-<5n(a;lPly_jl3%-;`iitNBFaF<-x1y=W+IBOV;J^xi+-iOU=kKBWPy3^Nq
zBI6w|)@Iusd0Dp{SB4?WcVoI&L*<9sDvoXJfA=m|bXVM+r&%mbyVRzczD*t>cEkz5
z^MoPFT9+hu?tOwI5a7A3(zRjU<sh1}U=5JYIz9`favlbg$p{Xhd?O<hHTsh)WOfm7
z57d_Vx9Vb0Z9nN|Pv^gY5JyLVOvxKveV~-$=`<j?_!x<z2dHR$1??71*)eZHO$%S7
z7g$4AIk{)Iq$W71NJU_&n7-MCCyOZk+0jPM&@JZYs2mfc9mb!wa3PzQ>B8(l9A=AW
zO_sLzS;Y58cMtW_AB}_O_zZw*2u{`fx40X>x`FSHUUQcRk_sNozx)2{d-3DPiLF5R
z47DSkjI;axYI&hV!h*I9CA)q=2aw6}^tv$U7P~6H){*0o&HCncRsr)y{IsW`{D<oe
z9NOL4>hE&<F+xi+wYP3%8tUj;`3S08`efHaR#}0RfwkmUkZAjQ{sF-HveyAEsTXKk
zgvFu7Z@$Ao#bw|%R05fXJ^H9+6|g<fzGL@lEA5abP|moz3vf{W_Gx~0Kz^L$JJbHK
z$)A62|39$FWv6Ud@qE_S{a5C=Sq;?Lf(8qwQn_7tb!YUT>Bi13K;?I74G2q%bge6;
z7rMGVZ%XgKq`I<L9)DZX&A^u5jk#ni<Y%D>rgy0BT-Eq*{LZ$&7K#3??Tq}88q<Wo
z*;?Zd=<&Au7w6zOPAvphwH4znU`k0z$xob{;rK<-!6wd6*?m3mE9fQjP6cW~X;#^q
z+^b1(iy*gdpQ!KEij{(^wweO<h+##}rs5yVq8VbV&8;IO@1%R*0xAwsPf`6Sa`-gX
zw4skiDDp|VxV$kx_$qnZY&i`}#y7B<nO_?LGwSlVZ;1);yjDX7igD(I#*wrWUtq6a
zr7|e{)<usUt2t3~jC7Qx2CX56jUI*;!o4FJ3cU}TO!j3cD^KazpP+<#9h>i%?wt$)
z^HzX~t(?fHjoR4)s$%tn86}|Z+or>1HQ}=S=L;#}A_Re`OBc@agI$s}>zp>gd*Ln~
zYl$tE7J&L{-(?#)rGS_?w0vKASd$|cUnEGJfp2KcvHHxAO=VCO`rtuLpcJ6Spu4EQ
zXj?@IuCREB@r?8broc;<@m40F))S+LKc9XaVo#MF;~9u8P)H$<+Hd)03LR)o9afvM
z4+XJ0iY}*VayRQ09oVcoKOAi_EL5Mg*XtbVja5t~RYhfFYsK4aCTj&7VI~_Z;@e@Y
zIlQ!}4YL{4fd(PGP)P_*8dpS}DRw|^uk50nkD^+KnDd+IbkgFi$tWYFzq@Pqpm|a#
z*{eV#mnutCwm8r$Tjb-e6S!qdyG+C%U&`@sgf9vkT3J1~#;2b}NuIs<<*cEsvMlBC
zEqnz~e@OJHoQ!w3fmtg&>*>(2Hyn2R*gWz1gXpo=6JtvUuIxWioH&XPGgU}&b-ha-
zxgdwW@xDtvw-O!4rfX|l8<KrwQ)jHC;N-BDRn&v6<rv9p#$|3r@THoRP|uN#4s+@I
z;o^_%EjrF!d({IQMK3qmCV0YoHjLI9dKKrA<vUzhp`28S8H9WvoJ(4?x<Fk%@BTo}
z#Qa?@5%NC7(Xs<qw6!FwHkEBMkcOQ)7=#kxdRM+HOHr<zw(CQYF7aTuO99>uVs8UZ
zf)0N)ZSBDAqnWhR?%I&9dUJU8_F$Xdi9J+*85Vlv`ECXqM~ytEkuoh5GtDL-Cx#A|
zZ%M1n?S4`0Y*rAm$TDR&uIY3Hf$L}57m4v`Y(d;S6qt7Kq{4}WVmy!HMAQsy5xW^F
zB<<Pxj&jg67H&3-uKA2M8`N4@K1z<LxgZ;;G=hbMx*T8g5g4ofV9z-)+q0)*SzOWS
z;?d<HZ1kD%3FFhm7gLNO)#X$hI1$fG#R&$(?_&Ti-qmQW-tkBAreB<$f?s6Cf&3J+
z{4XE9%~g`?ffMvLG?Bk}ms~9^((@IhMQ?(c2^!wI8kTwdPK(rqMib^1(eqUeYNru3
zbROyyaM}Ut!@;VJhbAVo4`z#_bU_~QiQv$zH^UJp3r^nhKcjiD>GO^!x}M@P=%~0s
zD$-V6kSKO*A6~MOeb=S8uEnA!mS)o=nH`R`Q%LM~HHBwBcHLHjM5YW4j?Bfzx({@x
zKkJ0ygnc0H1FTF5AZx_=JsnrYIiwnt)vd`IHJ(orY=i2t4sip`NM8kdZPoUj<;^Pe
zj>Z!6mj?jM{JcIN5k!|Ud-){^@aL}~7QD<I4YAKAuOyRpPbd~)_q!!6#OSLOv{>2!
z&FIhAq5FM}{H74|3$7Ix`KX%154<a~eOF%WbdXPZC~uvcbktn2Jc3`C^(kA#@+>zt
z5UT(`W)O1$uiYz6cU!y%v`>NcD(j`q+}t$9CS|qC@pTx_aG9G_TRyKVn;bp0e5*0%
z3@<swDgS7Rzm~~(puzEz4tiZ{j6$5yizZYs>9gCZ!C5e?G^QRCYnOAsYeD$(ed~LR
zS=lF2Ln~rgPc-h5e5QlEvXkg5S;*zhpXuDn`Hqg6COX^ib4G?4D3p1%FVeP6H|7ea
z#?B<<oi@HL{nOUSnfWr)!I2&>#KhzaBOpFdBTv%>K(#E8-NNzHfKrk`7!?ihCoO=a
zfwwHRH*Kc?_)yxeL&)d8g5oDomXpv`b_i`ZblfA2cLqgnL2&~ZTkBhjDc~ze076Wl
z19-A6E)&S`vC`#G4%YZpZVuXh$RxOLGj4|gxELsIQ`k=36G}S^1(>^ATtG5_h^mjE
z-p2<6?@1iUUQ`2#9`N8H#113iATUI12Cb`F130u30RG~1jRZXlpdOj6mja<bsnn$d
zz4TJ(hw@3N9Tp%mPR5UEtk;_XDE$<$_Za=|PDDJaO%4b#u%IIUu19YF<rBaB!yedT
zdmDg(O>RYBLHVp`Kp#c$`)9~E-HG$%N0t-Ob<HZ!THN6-mr9k~j7QeC0+%|@SRK37
z%pG9)ahe@@e%<C#Zmdy;PKCHc(KStc*;M88iI=nIllH|N(>bkcNb1<0_vr;v2h}E)
zUqQ<95E`#Jic$tNCIdoa$pUElzC1j27zF{hNdU2VfP!irK_dh2k8I!yWMlxp>dA)&
zI?^6RJy8ZE8C<@B{J(uFkVCsMfgB4&+gjOy*UMxCr8!1WwN9d@0)danKRchU_%ptv
ziN``#YJd-|$^ght38v7wn!%ey;6oYB0v!49pF_bvZs&(}{wuy6gYXan^u5@b!5V+1
zu!?y4rFXv?L>bPcmv7>yFh#gmjiEx%TQbb+_cA`eyjPRaYJEx$G#C*qER4Y>lMp<1
z&#M>}<G7k7-U#G4nKX)2L<T!B`CsBPRK8NXr;3Cf4hzpH=W1CG=64z&YkWSo$~)D|
z=POi(O&&o64LG#vvQJbLXr>Q+XyP6BH*4$@ROU@q6E;x3`jM|`vdjlkTfMbDEcv0d
z$iR6b)y>_J^}~zi3)PR**gPo&<r{l)|8t3v${F94C+DZCo&=^~<sQ}Nz?wBZH@O8`
zNA&m7?{jRse+3O->ShUtaFTXt4w^fKf7_@kqXu`V8)IhOCebdFQdZ{wW%256RsxoX
zt5NhI#75M47dBx5VgucL30WWj@@XC3r*`6L;zV;@1InIqpze2{y{c>pg)c3qDqkj^
z4I7!d@vwz?Y2@?~m+)?+GN<3w$p8wpc@3J@Ty|rnY&%zquAoTM+P;JYR}H*aSNf9T
z0h3RtA9&I>W8|)oaDUB6&%>2Xm%U)%6QR;|(^&BmyED)O6J;N<&s(0}g9#+uqU<k7
zCCscVu6w#g`^dQgHx)Z{+>V5eQW-J-)R_6I@k-rh;j^>rK~xrsAgL?)Y-f%ut63%3
z<b%T2ZLF|N*XoOoLPt^LctO$AI?voAHqvrQVp2s}e#6Y6xIM_YTOGb);i^6oTEkwo
z0t%(+b<R#F^&;=Z3Z}&ONSt1RDej7a?s_qH-t=h4&B6S%o}~54?ZJztEm8#C2HjR@
zNd`G?JswZtA@J*lot!V4L+!N~8Jg)Z(Ga^or4{J)l;P@!(-L>ZQ{NOkTG3bz>-0T9
z?+}R17&TLYkz9Ksh;Y*pX<wh4*E(0X<NfQ>hjJb^Mg{rc^dl>;6+ACFUUZ!K0{sxG
z*4=1;eW)xp?!NBRHCUzRUaMI3@(#x}4s7m+Xo1O#CvO}d0u8KyJPw6g?7Z~8nPrzS
zQDL;DRJj%r4o{Jld%dSeM&C@oFZ69zq{P|B;pJPMrqb>cH2uz=Ey>BG+|su<Vzd-(
zR{%}&{KQ2}(X#!Nd%kaH@<9aioIp@O8r;UDIHCSir*4;Yp37QE^449Z=)i;{rX!jh
zThld}w;rYygp80!GAXPzTjK>tA`Oy>^vSWRfO~0R;Y!j<S3ga%q$E!oMw%RYc|A>S
zx$lh3#XLwop>QN~jN^d)mS2~HLsHRUuEFaiRvzKFv<!<@ccBj&0qlGmPw$+%rx+3U
z*whpI$XCSMpxs;ZEwy9i3K<uWX=~(p>OFa)a~5nH>&miTDxNUH7tw59byxf268y^p
zCaxkx6Al+!oP<G`U&wn^uUAxhtH@$ud$=DT)$NTo6ExLb;2f2<)a+S1=jxU1r2^K8
zHZFzjoLFv8sM~eFEI=aIlqAHl_B`n3^Ic+L6fids_|@f@ChT|hv8-Aiwohomj3n6z
z8L5K1Qg87`KP8o&%Fz;Q<K;S|Y$cZCFX!|5br(gk3i=enwVGs0jiYlWN!|>tO2`Os
zdAFL_G#NN)R{K=`^L*QtyK3;U+7se~rO(!n7%V*ngJ>5?xxQOl9V^<;j;|8fO|n`}
zK6*SekUgTKBdRRtu-g;Mcx*Bu1TL{!wq1&z;whK+-C{iaHg#hGDEmges`T(hq_xqN
zraf)^=P`#-wNiP|+xGmX4<#NGlP+Bdp7Phww2aUd_GvopaJ0{gjc@ZJuP(85fioQ$
zUDc91cjk;r4?b(g6P_NIZm#R>#Gcz#AmtO!J2@m8em%%`pYG9zOm}8bj5M3kq<b~B
zTbymKaJN=N3J<<NA$+x`xX{Gjy?R17`Tf=Q7H|xo6E{QL?N$4UJcu>nU{l7lX=O_{
zfUbjAcg!rx6LmY4wX$saum|jo?u|`U6x}={=Owy%O$qy%L;X&^57b~oMfEGlB5v+(
zQB22-<ch@xjiFzDX@LKu#o-Tai;k|ri~MH(q$!;s?M|7cjFPK|_`KB_b%-fv*r#uu
zAujK!ukpUPt?sQtZpn*q_Q;QyUNhU3ZPsFG>5L4%=SEtzWOI}r-3(f??t9EUIXV_Y
zbt3cYlXbA8f=kpmv&$)VTjQPqm6dA??TU&U3M&)hWp?5(7<dxLBfAO@SMexxPtjP&
zx+J6O3$pB5#CSls#y*rN1!kMIG%PK>3Kz$CSX_9i@RKQkN3X`YNRK-Xpet^_QjyI^
z(Y0tPc>_DmI$CQ0vcxUOEG)0BEqIObKk-(~(!7hxQqtU~F5XhB;M@wc1xnzvVWCgk
zfnK-chYlg9GJs~<pCQ~R>q<nOg(AggEu$afBUul&fnSvMHp`v4F>uDL0dY)0L4R+)
zUdJi7Cb8>hZk67)oGC{dj_f7fUI@Pw7NrAfxY<OrlX^(VU)zCZs=#+p(74avRJQK^
zx-isuV1pK*eEus4y7U#4TG-QK2_r1?kf(M~czRST0Ez#FLUa26#U-_cWUzYYTEs#t
zw$F5A)Avei(WMXY*7#ReLN*M`sokp5NPex7ZpVn4g-cV_3EJ}sZYiW@pA%g(>xB)H
zamTFAu*O~`oahjf7Ff7vw(n}?cNMx7g!+WTJ7LL^UCO+vK%C=SLm6CYRHs%4-|R-s
z7N+X?rXu%!iGrVA!9C12^z)|JssgV;1#}TqDf<2ax+rRYB0T}{M7w8JY`3lhRFYIH
zThqJXwW)8UA0T!09c96L^BvL>NWlbb<`a;Fk4hWC-s-OJW*$<_%sO%;6Qo#g1s)aA
z@J!Qh(RVG4y5%8yCS}Ya(I0%CG8O1*wiJ5GUdiaumuQj}O0d5I`yQ!ytD;-HK;XsQ
zMdqh$Z<(-O4~+*;lr;#-0UfnX{U=|D(g=Z>zWlwI<SmRW@gX4nNh_u&MQ@=$E+Zig
z1Mr<iG_l$E9?<L0C6jNmp4U<uQ&T0%Ub?Au<yvwo=W`C;018L#2$(DzT*SfAPm?RB
z_!j1x4Vzzty&1G?ny7QU%XJphMLsr(B<~ViY2Y2T5GNzT;L=6eiQ&gk)lwWGG$YkF
z*4ap<m0Bl$^o{!;O5sW)#uG97Jj%sy%;N4i+<h#25r9&U%kqOUh$nm_;wWCtjQ$x5
zk;wcUz<CW4FEVjl5a=1Ex4Ly)@;_pD&gx}x<BKzamw45kfno)yH9j`mu|BHE{+f!c
zY*kUE?i*#vlo?-<cFW2xWstJ$!_S{whD}H6rquQ`aKlfwkFFhj=%Y4c_N&tP-Y33o
zBkTp8F0D+{aH5^E?rXE$KBb6Yxp1@k_XAolKf~moJOqMBqfNyg#+#SUypA}_{MlP-
zbns;b1G&Xl5-Dj)f%wbK<#SzjeG{?Yd{~TIehzj2(}2y=#DUzh&yl-@eiDnAWMEg{
z%v<jv^d?T16_;^!hJ*rB2Q#^Ct4t8eIE4JLRHHZJX$f<Sy)nye3WS^A4Kn77IpG!(
zbxaTK+0{7M2nqvS>F=3kP%-KlF)=?B0|Z4lkPPOX6I~U&Z`9JY&P#Mzbg*9>)GLVE
zG*y!kX#Erc`)sPld#jib3wI|D32X!^!(pQchFMQ=P57maYRVuLwY_gli)*MME1S^D
z8`>ZgDf#^F2;DSrx&OnZ*bG=voDqVS(7U>W=+v`ofz4!)qJ4|u>+SBU?|+MrWWOGr
z+BdCao}QR)Vm342`bvulla#8(bLz@#rk~V7C;UJlHqeP-jz%`bpHZ29b?#!@)IExS
z#_ReO|G-S{#(&XH^3l0SjZ2~WZM<^kDtuZ!8U~oKJ<vj(2gq?NhPB;gTi@rC{WyNx
zhOhq~L;u*x_J;|6n4tLIzyx+=s>Y?_g!6COU8rf{k5hwT2d(b_lCTdh=Gy;9%Uk_+
z6lV?Xx6VyKIeI9?^q{egt^GZscvTzo>0y<Eo+1q58Jf-Iu;J46F8T$eyG~cmgsqIj
z;@a?u<8LB$4jn#p(U9@L#Y2p@poRIg&){`n#l_eY!XV>Qs)ta0mH9xZXjtLtoV@S`
zjT`m4r@==!++EI2%vYkzD*U6hwU+)~mA$IE*K)Hi$z9OJ*jt<`>s;_|B(CIzztY5M
z+yZt!;%H-iLn-Rxer;syQ`IBL*oIp0L7;1xlSitZDPLxWZGu5KE&}c*0yx!shF)it
zf9}v}I^>~Tbe1pD{1y}7*~)ugZj+>Z3G(qFv3i9I&sBGrX_Fe|+`3|9*DfO-gzvmE
z9jb57XntYB?*(2Ic(5^`l`l)zhO0Pq-XW3l0P-;2$Z{4jFV)^e5l&~`amIY(4;X3s
z!3M1JQqqytOM0D)DO`iY1A^!8#qVvja(xoW=+4YLn#@NH8pBZTKh7plad3@m-QGUA
z)Em>%cVUMs4Brc``(G5P7eDg`qdo+M9k0^vcdi}qzC%qU$p&F9j)pck&qUkeS#y-{
zik-4^(~(_js+19zWf|noEz3UX#R1BeY$<t`otTc6%8+exp?SjQ(>@csQOmc{=usb!
zUl9wOElyJ0gS;7>teJ!9l4jJ0UX%THe7Rm31DG0{gJGHnPdd*4;%6af?g~pE;!>E!
z(N1L_Nx#I81Gk-x?Vk7dYx52^Wu$#-ec&kKeqy=tUf7lPuv+Je3D`TTu2Z@Na&AnF
zej2W_X6IYKEZ=wWXxo>hU6+6e+eu}B!|1Z<zL_+0p&`EVyuq3q%@vChcZZ*}6x3AD
z$6$$c*wmZ(u}hVrqO}R};I%K`*f$+OG2Jh@ZWf#fl4Osov|=@V4<&aB$ZRp%j_fPy
zEy^y02^B)C7f#3~AV`t52O8Ma>$z@D<(VT}!Fk?R)Cf`p+ibE#0p%p_0R0mcnnCTF
z;VW}3DroU0QAQN)#@r8_dUPu*#v6XLEpXVdJ2wjWgb@h*@|idMijHNz%Eezn9qp^q
zy&-afU_LXd=@jXf0$sA?M%(m<;XdhkzrZ_&+``Ly4VCwY_(;9|6(t8}Ku*SxWLat6
z6sDbwU{}FV1Q&9>{?qpL<dM(yZP*~yu6H@_Px@txUClrZhdam0v=k5_!NO;|#Hf;_
z(EO(#ba2YuBhr_vuY9V5=UQ&_cI7g%9pwkri0P#Uj6XC+fi7HqyCAr@und>LtD+yB
zp<82R6`P^UHXYFK2spo{GdZ;D@A;ebm!Bct-{Qyod+q>WIa_RSrY&f0D>Sz9&Y_gz
zPvAnGho?Di?>^sHr9#$#{wlK2w(D0z<$F#+PYchkzs=J5vBwXKeNPqqVUK^*9)FdC
zW7>6t=>(woDSerqYq#3?t#;4vFR6I`J==b(<q2!o9~@Kkm<m^hIFZn-gc-L89(V31
z1D#w4;E~4Te!6O2MlVp+eiRNG-44C6u1i<6SVGX>@0_Rud7x|ZcT<BtLI%L)owD@a
ziLapMp~!{7a{9|xQ|>e7$~~qr^_j3tE#ks{K-m2JQEY`*&)G{+yk}0F$9@Vt77rSo
zzrD)a(XJ-sx)f-pkLi)lqk)Op*>D^Br3WpltGG=*2;nh2SXJF%vvB#{VNi|_D^WFo
z=*8mTf(Z0w={T8gYpMl1nBuQ^s?qvMjM)tqO%~VI;d@uBCMzXE_!3cu;K+ssVn%2q
z3z38UHMTNZFUWzqyqHRu)WJ^=T^n&oeC6Hn*#Z5s+t!;K0k@QYp}L2hV|de8CPupK
z9A{<1(xAg9>G$`9tzL%FQoNj+@+L!eFQMk}Q<*Cf^}78HwP=1!Y4RNXKx2_t=;^!k
z_>{Ds><qO}`nIdvOVaEmk(ytq6IJ(&7(kvj!8k!XoPv@a1zfYlObu8a5|r#N9+odS
zSQXY}?<UNA-6j9h+km&&&l)Tw!fsRQrh`!}oso=wTK5F!`nvn2xlgi=Dw5WGv!FyM
zO!x$l(+HaE7Q=fHSg3B4HP+J^Bl2}s&d&CVB$l)Am4@vm)yt79SfPZucdr%t#*C`#
z5KJWMg&HY0tdyq=BcXzymE-w`z-<Nhr+7EPce-)sNEM+q&_Yx&P)8GDsCudgLS)A!
zRIkG?T?m)e&{ovPmmt#uz&FaTHkKY2>aeuZW3N3u6SAcmVpp^oOYDq$6aK6Ea*baV
zg*&ugWkX<)LG(wSHgvXBAhYJF+D!s)F#-BWXQF}jCUf^gK=s>3Pf^qSfauF7UAjif
zg2hpw#@-?xhz=fyYzqJEgZ(?6A`Qjy6p>fdk1t%U-9B7}Myx*x^1T<OjCwb=b05Xw
zIr9z(A^bbu_<zkC`0*hBg~RF(c&XlsN^{x=6LmT}L_S`Scq4l1_OjY3wpVVATRj+_
zUDtr>b=Dg{Yx$i4ufcR*|FWF@ANl)UneTOHn!jsYX1jascPfqlHFa{K3H(;gYBsM0
zVdw6a!5^y>_}3H!|Jm(nK=$-Mli0BT;h)T$|3Bd8-@Sj>xBL6gO$VR<la-CIUqQyc
zG53BqlKy+p_8+y+aex`W%l_LGlK!&~0Q#Lw`CUxERsuC~|DMJJvH<e@zgI5nFCEA5
z>VFKKJqR$rA=e6iS4)JyHc|a=ZRdISpR0QC1xV9=GqeGK;(Nep#=m>j`A1f$e|Fz}
zw}46IKUMs8O8QTh&RYKpn&arm`K$77zeTG1SN{KhZb-p7_D^Ou@ItKx%^3cn)74*^
z!v2VD|H-W)&Nu&T=M$AbgF>F)yZHPokNW|=|5-Twj~xF4djBgS`w!^-N5Sg9=R7~4
z_a7Wv_$|o(1A6~L^WdL_)ql_NKcM%&5_W&P)(`0YM?mVo_Z&Z<_rDaf|A5~A2z35u
z_x%CAe?aeF)5iY7znTAIquwi14ND=hC}wG5H;2C*{?IJIg$m12mWT7DJt)rPzHNF%
zZvZNCf8dbEF4R->y=HJotg0ZjHG`&1jU@M0C-WrUYF~5>hnRF=#JWz#OE24KCYWit
zJ;`Dp42cn~w4!$tm@70WDH?4q`291T_K^OsAkNj?n|j+STm8*UFWhV;V#HQBA0Kmy
zkj9Qrc<>G*Q`W7>d2vLv;H<Au`w|8!WS=Zl(Y}C@b~Y>TT3lDw;bKf)_h7qfUb-vB
zqbX<T)$vbm<vOOQ!4Sgc+0EH-k6UR6C1Gjqg9Lyy-eff10)sq7?_N8+XZv;$MX{zi
zm#g|Jsl4kA{=FB>-Fx0dj2**zXOxSjst(>i;w{tVEh-tZAW07B4M@2Xxz=dWJs-nE
zeECIHZP%=Gop*e;Z!?6;!IzExdFpEC;`~~E&`-@dNH6vB&W@EYF(AbQDvePsli@Xr
zOJ@W>JDgJP2J`qf;SVTZB%^Dw38<ho%Ul9=y|peb`5@&=X?l}HDQ}608tKzs_z{m_
z!($fe5~s8n%-Ln^q~;e;QgzVQD<es~<%0)@>(9Vky}VNw##gV=eU!G+@C&kJLM@bA
z*^Y={wF7ftd}M2Ihl2o>t0!4w=EXtR+&y9y{OwZw4W7u1YE5iCDzcHEGi?aLaUwNc
zY`t|esjM{vf!VvYVQ>#Gn}en7k6yARvSmj5Re3ay+VCzz>Nxzqg7{QrkTqlYr_#Q>
zw~_jUTHM~^wxZArm1f#@`7)Lz!C$_DB;2lk>Au6NaO&}7)5p0xRMD}eAmBh{O_CHk
zfDg@I6v2apTy5{vt>|*N7kWKm;MiC~kJH2PgNI)<nu<k7Zm>)iMgaY-a<x?_z*>3v
zdNh6PS{kIgiH3$NaXiBpGgl)XV^IfI%^-Vce2-F~k8|L%>7DVp(r?PSTZHZ(8{e(5
zBFYuydiufXxdZ}^nyL7R<^)L6Ql_dQ`|M1i&t$$N-OZw+3<V<nr|X5!O@^2WK4b_B
z|75)Td_LDs2I@XEMEE3yxz37=uUV@spdg+UsYvOXi|T&}>G<TO^)fY<VM>eji9y&C
z>j*|T(Su=*W3O+E>fSb26bJR$m~OW`l-qWF;$$8`b7sG6q`)@KcX{`)?ulBR(=R}y
z`)xB)&u3(7&kRefuW6bP@-6bbOjQs7=d>!nwA7`i#jb<qD?wF}4dZwcZ9u+2&4e~9
z2@IJej`n>ah+bHE<`g2UtW9>;rabZGF7Ni{fRW<OEw%T<LgjF7cVqN+U1$=0e!9gQ
z4%<FP(}zCc1MjsW%a_}`;S}j3h~N+QhB&Qi8Rey-fE9Maz&`qtV-=k)l|||O2zmOR
zrq!FL%tBrt(-DXZBle`joo}?Sp;$+b9BFzdceY`D&@6hVk7q%2;^FzXN+r<T{FU+E
zPeZLfKEv2Tba|^|qQ{&p-s&r8VPO%o6>3Idt_HJF69#J%>eEJ(q-qt-7zMOf*B=Zw
zR}d9yYcDveoI0OQ&8|f=SFq%m&Zvq5k3(EAoZf>i=yp80VDpl@h(~kpeMZX1Ij9J7
zY~6?&n=C>dPSHw5qU%P4r~$d`*N)4nKra^+#`4`U6vX1O3A23IhEHwG2U#(<dX^f%
z{e%X|T8m@oDq*G`SxmUDddKQYj%sd-y{SSmStNs|?SR=E^P-+V=#eqb9TYT?Sw0??
zJBu^IHqH1T`Z9Q1#F9h4f)b22GrU|D*I=`V;G0NpBD7`)Rj)J|kk*iDTdZ#}ba4LO
z$xorrvradku6{HWX(7to@J`;`i->>XLU>%ZkGfGWz)TYhPq!}%GmJIlOMNAEdnabZ
z<iYvyqhij<du9iAFv?&OnA5jL^=t>@1E3&$a5}b^|AF1HpV)eqcnVXmUGc8CLOuB9
z^lHu$zwHX`UZ?NjSw%2Q#d(S|Q6_`fXJrs_!MdktFrM>}Rq$d)2t+KM=jo@$y^_yC
zvUb`QGx8gzo+GVpadZZ_>_h&_y()>i<Oe%~+gk&5x;(yZhRJ)W_voS&@JIYSyW?P;
z!pOnW4tbtu=fuvQ({qQpj&nM3D4$YqV)U-m#lU=zP%jhF+-68;<E{$sG-Tl1oyVHE
z<S&Ay?yn^JUIxJ%(wXI#buim!X%^kM{V)L`0rgeS>&Yse@`(LZja$9<D$CkkO2j^w
zFIvUCm<r@P`2k+POYBwESVTXRWQq(Oa|@S=B%%0`&LqAf`D3YVYd-rsz8IT4GIkuH
zkHFrU62MYwLU&}kZ|cZw=g-9^C&9C0%VrZ+-Fi`T37)IC5GA`5-QGFWA!^}hFF=O`
zO6g&$OytV>_RWGpn317jl#-2LjMa&&jk70Yxsim?IDgYxWGIxp-ygBiv)E?u9;-mG
zeX<&>GOze%zDA^ipVUy@8p?;0p)lbd5xue(_N-L*Y~5DE28gslF!ClHU};jfBSxOB
z-G2|;=~WdTiiIpWy4qKoDoq(1-TyeB2C1MZh7=)!04Z(TSR*$Ml|sRp;>Hi!N|zqw
zWG^?pI$!@lU2Jh-vUeyQNNr(_=!}w>g?30~6)-m-ASGKHBuJesFqLw&)+Fgb*r!dG
z9(&)-mUxB}>+}x3gNe0<+(*UiYZ5Grkr9HAt_=1<h>Yak9o7LO`Yv^V2;Dn{ujP15
zm!!IHgQrpZF`<G!XX`DLsS)fwk;K8{>b0hqT3jTT86-a$iiL-snYc2xmfs=Yc!Q;B
z(Z1t$_$H827JrOQdX<3JH``d~AZ)wD(Cs8Gx#<UV*->qkX{b>i=(irk{zw=-U_lB{
z4tx8p{occns4xiiqOv?7Q??!h=r`!0C~~N37GnMue^EYdcPOBgko)Nv1IKTk2LzBX
zo^aPdS7q&>R3`qPyGS2N76A0nya37UkH_%TgdspZ0FW#E?z!6m+O}jMe%hh~Pu?d>
zm%tu-y<bDC!@&i0(7vtD#H_EnhtJWnIM`Bu#JcedNNAMfQ+Xvn{bCaxkPMl=Ujk?s
zM2_78n)A4#>EUI0HGnOjKSWgkT4;w}|1VVVK7FhBf7pBPpr*U1T@)2XL_p~s1q7u^
z?;!F50%9oAOH^tIAs{VCAOg}m2q*}sRHaD`B_v7}klv(*gbo4;H9&|r@11k!e&^or
zn>lCZ{Q1rB4>KhEva{FPYd`B*&qLKi`I3X@@|Jmj=S;Dq{nzK2z)LT0_%=U1+_pvE
zK@k8EJN2#58?XuwYqcV*+Zb*fWun-+V4SMa6ghZ3Da)+kWw1a)c?DJV7u=zJbOUiN
z(o`L*y;#1$I7JREgR_soC7;5g9d*s~J2~$@{x;il|C(=ypbW#TFZV4!+dfYw<>`9#
zk;98M|5oK=i2@h!OiVagq7tO0g~U&c96}t)ZE952x6Uyg&8=m|Z{7RRsVi@v>#W37
z`U?`Ge<Oyv%c(Ta1PnzTn9fRen8J8MX8?<cuGv;wg?ZK|M2-a9qly)-eVNS51Ak&K
z+>&{>czy3_iH{EZ&Jm^|s7sWj^PKn!z&Be}8nA@;!$EQt%>^$Pi7xe?-G@p#o`00t
zRfcNsyl4@olL;VOBQta36HfL^S0=kz@~Wld#MVL#8}^nx&dMD|$?CWyp<T<#fP^+N
z<NCyYztGTXhYs4iYIUuT*&iH=mevp!Qoh>EE=+q;y{+;(!3xSZz(D}MvhVBx&!w-W
z>d^Y=REPhi@eoW9kR-JcdX9>DkSRTp^(j{tj6d4Kj20$`AP2^;@8lceL2(~Gh`rrN
z^lfdMmsvXQoNe<ZGvLnh+&fO)qrwE%kB?Y95<Y!t$+5pXH}BY3POYSs`fhsba6tx3
z+1~;c98ei2Dt*9oU|7j$g87OQ==6)e_6N+*Gi&Q+0x;&#a`X_u_Ls(O(*z`SI=ukP
zf|V$ZB=+9de<vtKqh7$pKk(8ns2du=)BOqZ{5dhC1Bsk6uefh-Pg*yWtd+fB#^$i{
zE}iC5x(Pgqc&VFbHw7&W*CsN@eH`09+}2&h#=4;#MRz31b;Yu8$*|QJ3dt!cGgfsl
zCK8>_&zLrVH>QqYBr`vmcnvuUJEaTxyhJQ|hvTg8cVP7ldcnd+r9YTU<?QoU`o@zM
z#GXY8WjeIweY3}hSAS{t?q^fxNdicBJTOKr<tS)c!w(#AUKlJ|;JQqzGB~pwM{M0n
zU?uO#_&a>tk@!jtb)^nUm?DtO=SX$Gnjs&Kpzpy7@RQAFYTz)5aZJOZiq_pNE7^e~
z*r(3>Uu?#_JPX6SeKY2ACI~Xyw#3qIC&3zGC%_LPiN3F9*%Q+sQNV2LVKx(&sQ*jj
z_~zzC0`ohLvw`o4U#^rS9Q6jz?szhW$)DJV9a<d27p%|5a>iCJZC5k*FJd`>R97dl
zxWT&MBEXTOJ{CX}+3$5#dbGFt^xZqGI9B2rAygd7)PVzV+08k6Xx+}WZ$vhQ7?zBo
z^n88N9>*}Vvb}q~`RdDZQ1u`%T1_xfi=XUFIwy{O-wv(-;4p%lN&-TJA{Vv^g_o;_
zcDIpQ@W=43wbY_l<$uBk5Zq!7BnhfIU%MiHz^~)j={_l~($KX8hIrV3o(S;xnG6uW
zCnNZ$u~XuT`&;G4`yM6Z<3MY+nf;8UHjCA<0maG~vk+C4gH=z8nK@?Q<*9Ay#;FW7
zFG6C-WKRr9rIk8hSdVDueg2n5o2Qz>B>k6$v>oOTOQ0&v_cF)84RB}LoK7IM?p&gb
z5#g8GWO2!c#U{Ob(DfC3o>a$-d|@|6!RejIO9|sHtYDZDjKPG|QAvH^v6hK#iU6zw
zFuYwMqoAp1Z59!1=5nUnb1~It@e;?cXKSjw4xuOd5|tPM0L50Y^F7)I*;Y-W4)X1>
zkeP>pBwa7<wIxfc3}F{lwTr)ya$}bHgfxL5_)_n4Jz!h`N7&e38ZJc)3*571cn7M$
zx)|!}zby|4_Gbv??b8=TjW3Uj#ji^)lB?<UXp)L!h0p}TWT9UL%+CmeLK+*lmOm=Z
zSx~jIjGMkB7h`VuS|sUvs&A@^N|ndGQ|&fK+P3MJrxYH!YVz$y!F?iiXD;G*|I#G?
zd#?P)^=oL>yZsLFKc_!~|IceCSJ*C4kMRHLb^pJ0_cnXL_22LFU+>HRUZ4NoKmVh=
z`0x1o5Bd5(i3?jTNy2>Ndz_=&-psCL^6Ek#=mlaMT-cDMbZfCU^5?E&=5x|6<sZZi
zsu`i$(q9@qMH=Ae_`8?-;b0=C4T0$sdsX-+o2^q!EXqP4jaDtI<cNb<@aKnW+R7s_
z@2ur#w|k%etOQ;n+l&V=v%m^E^5sXoG^8rlIjm;#Ivdr>fuLQG>^sQ_{G(^C;*gc=
z)jUfy5eC8;q+-J0E<UjNqg@yKK4>N=sSq>nHWr}jzZEOholM*{{G6_8uzM%Ym)HAt
z3UAQU0yb@T`(Ujbq=J745>9ro81wwT9Y+Jx@f9?>?M$)rWvtWY!QiiJg`XvJ3(d(@
zqu-1E<Wl_<8Ke-}jtS}3@s8QpUsPr6u<Hl2h`0r&R>B|3t;nt%zM1PR6)c-+o2CsF
zFXhF*Dtu~;zB5eo#L$($E4U{T;kkR6gzeT+YWPMJe3SA<L*lwr0A3dFAsF%~4Nc_v
zqK5>^KYw2=h_W(4xuF&Sky?t>`Fs4?xxo{n<7#PzrvRTWWJzNC#*k)P*5{1ARVg)*
zaCT?VeLL%5n(m32Pb~m;xCHS9X}>f7UgI!R0|>Gd!hF+year8nRyl+tg}*R3WZujr
z`Gt#)#4RbBWmmh(I|Ef%${^-M9wBvHzm5}|01_stW;<&k0t3E8=N}wl>~%!No<x*3
zUG%OYM@wf;5rezxwO<b(1D&5c`~6r?4B;vFCD!5QOFjiW*G_@UG9DILI5^n#R@hg3
z>tU69yi>gQH8HUQaP4ViA`kIg8K)c{K*h8cgj)z}N0x1Coj{V2=!jSoGFYMCbyTE&
z|IY7i?#wlY+cxhBu5@D8LH;S>r{l~xgC@du;~6ks4HLJ^oqtiU2s0y~*L#9?(}Fb8
z^jG(%Dt)+nHi^zg_K`8!@`o!MBfTP34V2Q8y<+N}SLy&$vx(u>xbO+){)8x6yp+PY
zLE#@yo|SUivPQ;_t}lMjlUFfV6l?+L+Mi}q)tz`BIa-@eNLKha%7%yf`J{b^cUz{c
zz2IDRGgV#wcBvBHQ3)3&hGkQA!L-IYoJ7=?dDh~JYg!3a#l^vdnFVZpFJI&F_7O;G
zRfku}FzwpiYs#_$;Y-kHv@p@&&lC<r@3emEIN8Rqu05mXoyTYT`|Iq&Oz+E@!W$8a
zkCxk$uR2`#$u?>pAL7`vod1_5$Pr@-&hk`7Q}b0VLUVA-m^4q=a>zAIctSjmI;0N>
zc4AbO3djZV9Q|`k?IHWnyZ*G)W?_wqNz#|aCbEl$(1OgcbyrN<$Gj&O;9-u~O=ajA
z>R@94Hk5#7ghik3;`1g8K6M>cLc6v2CcH!qpBv;%4sVozlsPU3a!KE>jPMDJY;sR0
z4s~e9p+$%qYW_}w6P5nCP(ItO7yNhpc|$LPZ4On|l%-nLA|1KHgPDKwJRtqNB^T3n
zSFD(dIhRo!hW^rA*r<$?(CO9Q0+JX`Yj$#udjL$oN#z|aS5nGAVM61I{C;OFrwsf?
z%Gr;hLOYIo-L0|e70W4L$g;BPLy5u`(t8{O|Me|vE%6bl1K*cN2Bz6&Q4q>-;!k35
zRO7CdfqC;!aUDLM6KrQe9U>Oxx;P-yQP+d`Xy8i}w3|F+I2SnL^8Wm-i6=HWzy}~*
z_r2(P6LX?8J$*S`o|G$)L71e>KAG^B7`m)m=3((0#(v<RzQcK=h;uRxTW(+El}a8_
zP5x(}`)R6HfMAT{0ofcgcW{FBs2MksuMeCeysbPgDsU~`7hbQl#|yrxFFPn6r^<28
zRMpdrS#46-IM3q6V}q1;UygPbdcSnhA?s>bw-3cTT^XY=k~nSN!a{J3nU%i{0-k?v
zo8$PVi=y+u;}yL)duIxF-T9_C3CKk>28UpQ%eJA$s!h<DrLd%3kqKy}w72ArOjTt`
z5`Dzo!u|`&H+pJRAN*3P3to6nP{s#E)@xs$`ym(~Ssh+MRzwuRtkxcaAXP%aj-5T`
zNlW*e-d-G*_3G(snr*-YXlsuZkh9R9{VnlNlo6AS<OXEc+C;t>={;4IjdqjAH6|8i
z{n<nLG}r4}!-OIiq`1{9s?+EX!-v{L>0sf>f5Dir8>3`L5Y8CO?vWLC7iD@hbEmuT
zn$uvZ2;1w%BS8s&{*+?)Eb-D}L81bJVV%eQ4OzPi3-%$|^*jg}+SZgX9h1(wcm_H2
zfI27kUU&U{(pW>!r%o<=7&%PUHXBM@CSe$lW8j)!M3iSHbIUw~EQU%PNr5ZR9+~a9
zL@b@lF%&&&>^XXll<n*~fqQ-uaOFJ&uY?NDS5gcEg#R=L1J_Clrl4s%`7)#=oOaB&
z9_4sApb+MH?`DyqV$wqiJy$C|w=4SG7m5;Df43y#zr7&3x6j*mAy4OC4ZV?}%E4_E
zS1Yo&G=$8n{fK$@INcp?dNXD@O4Bnt5ycQv4fiK-#zyC|FJu(1pAMj=%(;3aLZa(#
zUba%N?dR(dSD}f}#}(W_o_`=e!(JuR<8mQ9v&QHK5~zygO;+qQX5*alwY!?u$J!*{
zBf;_DJG;n-v=1C}LqmI1v3?;8K?VbRDEIcod`1Ihe=badwAG<?Gor^S<H;yo7tb%Z
z+OBSsV-q!Bq!zUt81_nO(qDgW?)9Y;J-_icBwylghag9TFlW@EiRjS`Z<|p&j`wNo
zLU3b-i0kJFP5h&GDxoXmUr}P9eJvpO6osGskQDxPv!w`qrFEFaM7=%B#}4bMWd3;D
zxuR`g1S;;X$8d8&%el!}P{|ZT5?dUKTP3}XtMaf*&i$wpgx{5aT0dbvU6NvkX>4xv
z5J7%$&^MBZ@JXzQsWG1ssqUv}w@;oOQqt1v;-Miu#02P0R!lRg6+ab@x)97?8b@}g
z=36GmpJf|Ys=w-hg(b4CKK!|u+Qr9go|GJ20V~HFyg@Mt6o|#aEk2W8X+8vq4>UMF
z<TZOqhS$93f@&mlYU;r_zPQRnAVB;t`gI8$r9|4EMH97d<y?!1kOcpb0bQYxgp<w`
z)<wG8MTpEW{Us&>wjIL-za8^%A1)du>i)gai8V)^x7DNT13{tWA=Wcpa@Oe1gDz$5
z9nNF-*K#k9`O<gyNthG%>SI*HORJ0&Z(x@mruF^l>=57;QM+@9jwEI}CaIx^Z)4TL
zjL*78RedSzfJn`{NDEm%bD-j1e!T(>&Q`>!=mUQ5e*JN>ZZT-4Z3kPsaU4{m=c{G!
zG0l75t>5Q9vxap)MS;Y>ly~U|nrBpjQP^@41%To-wL7NNii;QC`xXKfcrkuo>^v{|
zBJ~5G9c_-u{ajr9x{62;qz$_T7LX%FLZLG8hnjTV{T|m~FZchcuYG-|K5);#|G05)
z{+~<2k1D-HZmG7%&^8Xrnj0NYRPP<MOrH6R6_c|!&h10BjJNFWWf9*!OX8l<x4h;!
zX#6Zdg9&jh)a}mKaVkyX%D;_%KWa7*?xrgKKikqi=M8X-rv83w`7aIC)$l;xfr{z;
zBz{HwR->K{`t#P$sbKF<RqX}D`Ssz;n^XaX{}sOH|LWgp{H3v_qOMcXh(EWexT~SB
z9S`!&fo90l(L*VK{@5YEfl*rV*sk=pxoyRKHdsOW`@3E~mLeCcu)4oAx2vyHi(4J)
zL#N|Wvqx_TQd7w&J={NyH3V#u^VmmJQ!@#H2k#1DpX%?Wc{iTCEU%{%r%jV{L!e@4
z`$G`Gslzd|Dx~I>DI9aSB*cFU4805sT)>*Nsbwb0f=tEj<})&UZ_vMaTd?Gy+$KV`
zj-lgBWKOX7p~VMamp_Sda1@;ba&#*PpDC~LxhpPe@sZ-bk_<Xdp-~ip4dD)+6>Ku^
z1N1PQ+Bt(X4hL=QIk7dIDSTAmHCGjWKssL%+ab{-tHVYGgb5}|M=_PNwhOSYQ(dMc
zkIfwEn9yL)EveTS`>1bG%S|tY`na<6PolTa{1O@mg6-G(scmlq3b#ka5x5XKqKe3%
z$_=<z!0HwO9Y8@ab2+R4w-yCA_}U@~2`%@?nzG0QwiGWv{LLQ?`+_nd*yDDSDi3~Z
z=q+kr-x}UXO$gyS)D&1)fe6uD;L{oY<8YyU(MO`MGE58BV>j%-t<|CLDBt<2G~OV_
zOEKjp3}HIvRUGptDR3&HZ?Q-9f^gPt!FS4!-ZLjl<enFy0861_KdkMn3I(IKQMNJ>
z8W#Qwx4zZp8^lR<Rg**9o~F+Y@V$4ziuTwHZqOn$TDv@`$orlXw12JIf2307%%~>n
z@D;pjYPDc~HRbaznxAB0jzYGg2;E6R0Z^Ay7x|&m&RzDxq-yU9YavlSHo<%RdZ6`J
z0=c+6p)sS8gEL1l<GMs$h_my>e_>&9q1RvOK5|v3lf{UJXj+mt)pSJ-5lz-Cbwaco
zlzPZIk}TST^u<8_vDGfkS=|%MIYHO>X{>vB4OOfZ=LU#lJ*P1pt@8%Pgtk+sX>f;0
zfmI73u;f$4<BLF~UeQmYW8bCP@6HAy_uVhQg;d_*Rmy4&Q$m~t;m_2a1z*WRjmyau
ze`!v5;pu-~ofovH(1NhjSQkQ2%r>H~no6rc<w?_?NsSk5i6zkQJ({2Q>1i&rU&Ou^
z>=y2-T!b<r<2&<xrk<ohk9Os`uxo#?t$gmP)v`ojCnTf`TZ2hNS(a>_x@d{SJQ&BG
zSr{N&qE(8+U9X$pvs2w(y>RvSn!kAf7P-2Qp(oAuAeiB1OWVP_3XH}kKo-)=ep`c@
zJZ~9+&5p;7pj%LlSS4;t0ws{5Z^0YKh4EOd_rxZc0ZVavtxorjF*qp2cwei07%un`
z<^*{H;nW@Tk83c_LSfR|M^$)s0?$VKo-)f01$t#{IUE#Rr0|;Hd~v0uP@FYkZN_Bn
zNY|sWdu&Mf3Q){#U>U(7zw2b^d}a!Mg-oovJyy>MFde%^)FlNU%mV2)N1Ni+t84pV
z0gAdF7w>}}yZFREZSu~;b^*9y_#wP|1d4yLnXLURpq1a~XxDVAqO^67W4iL;iLai5
zeXla@4aX2Za;qK7YNFjVB_VeYxzfV;bk2UliPBPfO5FzXP!rN6LWh10haI9L<?8S0
zgr@4Qw#Fzh9|9D6KR&daOW?&7NcBVz-_4vcG|6l+?+Ml#i6R7s2VJRoncec%V#7xb
z`{WH<)-wwhwbZL_y<A+1OH)8W6d+*n{C4`;gJH7HdEr-3(3yg!fn!#%q_;_HuY1Xi
z+%KZ+h4(Xnd2hYj$1bq7&Y^@b1ybeWT}AzIsrn@8t?vt}$nk@yZ>?_}`&~95Yr!x3
zk{I|T#Ep41?28}Tpe_BprvM{<i0$$rh=O>Y@;!ZK<5u|27_cDgbTc#rFB1u5CO^kt
zRoo8wzQ8+q*dH3f4=@{*+L3nt{W90pUs5SwpP9eBrdp7sz64|+z;7G>KubV)R_#p~
zOeVdm6hc7HZSYOt*hH4#bnSOPA=DAQs_#EA6D6`vU_u>Uu(Q8B5_c7@^LYUlX-O1}
z+Yz2@2$fq636YTCCDKPdJ9`E)3{Dk_Gd1Aq7W9xR`n)7&5{zk{rXUYCT?!P~u<2*`
zWeoGsr$bAsDzO|X6EDLV_LE8;($IC)(BdaPd0(;>_|nn#B!w@C1E3*i?A^D%)ztng
zsr9a{Un_`i5GA$*<tJ+r4N@A`P{WE+ig06Gzm%&$cR&*=q?}Q1x+omAck8#Sb;rr<
z3HM`fafW<vt%3x@FxS2HYE0W-n)I11BANMHS#_=?!jlndMa=9}3k?6V$>Sn$7s0HR
z3tl|8v%YB%aAXT*Dk+1|=ri6+OulK9%{W#^!V<r3;fCTh#Bc*z+@re=@p~K9p9w3J
z*5xkf%Eix<{l{<Wo)oA)`^vT^Qubv*Hhqg?P0ssE!#MV^Zq;WKGf+K%cyx44VkLz)
zY(hYZkV4(b_$!2lOjN#b{?>yOz|`%g(b5=mRyn(c76-O~nl#g&o2G_b{wKL&JZKS8
z8_~j`iS(sxwtw+NK_<qx<IV&s2ZEhxaBT7C47!#q9r*9Kuur0u(^oEd@%veX%t-Y<
zVB}-`%RuUQtMPuziTsAR#2S{H{E`@oLfnAs;Yj`}D5*s*=WMgFLt~KB+0%g!*6Fcs
zwwY2jY(Kx+sUn-8Evi)(OH^0pIGzJ7^~jE#ofKT2a0f+{aFh$VJ?346F4&d>)u!*=
z8P_r8&9ZucQDbR-V$;B}TnO1F)n|}R3wK2(hbVL>g_|=HtCh+}1yc<`PLs=7QC~c*
zI4Gh}rvB1tlCpIgs%a;zUvcdy{Y!(T)&E%b`7uHKaU#l<$ojn{x`y_YJtg7~8CvSG
zRov?=zbVJ-7`Q<9VeaGytAAR}O)mP`NQ@Xw{F+bWD46I=m9xn{$cx!&EFPI%t138=
zpyXgC>a7O~>a!Axwy=C?PJ6I6BLWASVH9-q{T!MLC~&45BPz{qd`?*D<z{SMU369~
zz?~s@;e4Y9^U579N*Rmg2p-aOWqmdC=19U%e+G^*m(;TON!KKY0nfc)Gb87V=9exS
zvNFF~KolrTVSs7oNtx%sUQz6FsDU1wg5^cMKtmJYZP!UWGVCeW_{ILzAv|TzU}FQr
z1%tn}!-B4Aq#H3GzE|ucOOj%Ua+`~!hk`dekomVJ;0E}R*uzTT<L3N@zP@hq!{;_y
zzj@4^T&{k8db3+LW`^M=s*zF;*h@!5pVF7I2aQm7gVFhmsihifwC%6v5?i|$)S&dR
zkdEPX9<g{udd2$j21fIlkXbHF?M~JxP)H<;BV!=zAJ<<Pp`F1!38)Pb+ZhVzDj>vr
z6t=fWw-%AYX6+~4#&0)a8uD$6aukq_5_ej~-Db?)d}eiT?ez|bjywN4I%nUZ>SwPV
zmuJ1&Xrw&Zu-|pBHv?*uxrp@$7PNdn{%56Aw<Z)f8D7R<4e6FqS~CsJ{if`!GN{@O
z{vMEWaR5Tt#bXi|Nub3BupkR5dIg1m2G?~?-V-R)I*Vof1AZ@LI!iY*b=yE#-E+>t
zwO>ZN-J#HTWN)j&nsQkHN}|J#TSt*HmhxLa&ysb(_7E*I+%D_7-JM($wl}?(*T=|H
z8BB6Z+vyNt4_@nAi)$-+ynqR~AW<A3@UI01kVP~zXbF15>7RU_w~mheB@6$=E`aBI
z<8JUb?pUG+qsF20O-2j5dTzVxZ!u!0>|ClpjMN%OjVF{a!Uvx5C((j4(5`PJV3+mw
z=Q&_Q%o`|57h2{(QnnfKf_QH)F*VjUEMD|Ei(}@!-3i!B`ko{>V}$KGJ^s@VV%#bJ
zB>m-!XN!hi^1OIZUCu(NKJA`Hb;A{jxSI}jX+jl|p23l>D!mG*(X&}#4kCZxCa-a}
zBR>@zDA+_J<kx2;miT4xm#bd~3<!3Lb+l;Zx#RdHdiu|^ak<RjW*;|NHzvG|<@`N*
zwYEnrOSCOy5!;QuL0amXoY5~W+T>KSwzKf0vlL*Os5cH#&0G?6b8mC3%eG~Z_yjzZ
z$~=jtZLaK}0RFqbgs`}_Dnk{SkXE3=C?bcFq1hh4cUJ>S1)(|MrbMMYTrpuD6Qdzj
zHUmO7w$xi+**h<akccaacM1JQo9sVRc^&O6eDNm}Zm^fSL$8g&>h(TRJC5|633O3U
zQXy`XaqW#IfHKfyKad&AzhID-9(nHeMBI7zW$`37DgA5*<CvF$qIMQ5bIJb}#WlMJ
z)&r-2?xEP6=(ZR8$mnOo4N#6Pts8am5)3gd5ysD7J!?%H_--w@9`1DG>ysdQPIb-%
z0&_P;&;vnB5uiue7H;zVTHr=bNQ*!jzyd<!RYBSGYdWBNFjo4=0O6*z4!pu_q(-m7
z`|VPEE;*ADdZ6Gx<E?3Z(nJAO_dozOcdOq@V?a_bca0Ld?8Kts+ThB~d{#n!A@72L
zuwkCDdHNM(R)F{b+w<vE2(N8ktm>>IACMjO&w}izhB7{ZnQ_Z*Fng)#z-oNEep)Ej
zKr}&W69DDQcr^U$Qm`JfNgT+94W}NjVctpA%~atog@@KARaAo8iH!1cD7PM*Dh4_e
z(%k}$JOHqhM*DTpV~W+Zo&r+Hnqi201lYaKILIU4@J{L@5!=11dV4>A?9&Z6Vx!=G
zL@@xV5+}n1YusvYiGGil;`mGRl<_m@m}B40_>x^X(}}c*tu?3Eakxbw-Oo0x#Olf+
zS+$x<d&F#fOkY73Y=9}?)5k}B5=%1NPNXwa(pnIRGuD6h6PT#fht-ch9mIVa>wtNK
zxH8Xl7CyPqcIR82Txwj|zn_;_(7=QNyW{c%_GVPzh9QV5TcD>MAUZ(G?Kh*`y1jgt
ztLMOdmh2=n=xU-1OJ~&!wJzf}j0m+m@Rug@u%VAhN@mLmO3|Y<Qe|MDm^o=+xD}_h
z-Kts053+T#9}%x)6nh34J}>|6Fe6!O7l&$o%u4%S=&gS3Wv~0TG=uu!6rx!7P@Ev6
zf}=i=HHE<wc^r0q0uCrn675^PS}9hWET*)7lF^*UJM{Pw+U$hOi<?$?r?CLA2i*fu
z_&+;yDzhOokn`00pF*gx9()!Lm|^Hb^G;QuWPF$Vt3@LPS_6WLND?*^+bi+|azDJ%
zX4LD|{}E9bm1TNjNhgjfh}BYpKN=%b)p2t=XE#ipNY?G6T{xK=HuINC=dhdU_C|Nk
zPhb2qf6Hy}ZR79oOa1CtiI--M+xt_8CSfxA)#3xdk@Fqxq71U~<I~gllb9ugZN?>;
zH#OiXhOe*?TLhy`kLFQsG0DOj4Q{4okc<Go<Nq{@^uEIZ)xK&r{ZoZAU?_pAdrJzh
z@W^K$CCis=s-+}(9f*y~NV>ix_I^z0(7>zS7!YTB>hzRN?{PJjRS<!r!c#I+E0bNe
zl?Nl3bZbP=(&*Z#g7vSLVdHA&l?BoMA^jdW;+<_uvuo%VkQVoNL+&$${!d;*?3v{X
z-;JD<PKtzNQ;k`g*A)tqoD_)@$cTCqVWUPuN_@cY2IJg>YBrClh`Y~f@>L$kx$Ht^
z<xfT)#qSi~Ob)u)-*A|WnWRR3S<wf_;#D;G^gwej{$vWrro10UWisEaJQTb+O|pKP
z?LhoTjug#t|HpHA;JnO?wO_>7#Gpj9zSi-pHHskVML$lvv-*n77LaKhEx0+g8mfi`
z61+?KPNA`)9M!i|Os$etXCBsQeBBdH-AGZ`xs5kp!F&#Dal&TjqGQdEq07tYHSk{=
zI|<LBuVvNQ2hbD!KPY#KR9g=vRnv>iPEyI|o;=`xPPqaXBcx<Nngghan$t<nTfNW1
zcQMnZFQ>(^ZG7^fJ2AEfS$>fSX0qTOU$&&hE}|W~8p^B(w^-CpcojsJVHCsb#5hI4
zT`T5DFE_y{%$~B;>*cS0fiyWM4WW6Y;YDY!7%?-r|C0P;l09^3mzk{JIE-eHTu^Bv
zI$<~ph_V?*NDg@_A!<;#>Fhzlc1Q2`vv28v5N?Z{*t0&FmZ^TrQt)&MWmmROqq5`(
z6uN15P7H&9mqW+&f>XK$xnYGrdKWe2J7kI*zu@_O{XyP`Wg<CuO{eF6xze+C+eSVp
z^<Lz1ec3#RqKgbuhCj%)B6KbYQ3*lb5J6L-e(~bjJB3z;*d3k_J(LiUf9+Z9>X4wj
z_%hv4PS`d-8MBDUIV0y~=EP?>%A8Qa0y%_?BBIcvnuZ|Lz02j9MA+?Ao3rC3S>S0?
z1aeD3lXX&*EXIl{pZ6tlEpu)v+zujUUT%W0OAbj>|CW%~&K8xE0ze0M;WAaQOz0nl
z5MJ)6?GMmmZ2b%SNO_T8rsN;bQ^0%|fA<E~)z{Y<o;au}Pb+Uxqq(o=-@*|=j{pt$
z;oZLARI_!)OGn_3MctCMuX;|#74|HnzV%8M%7x~vJn2e&cf>gQ#r(V90;QMw+z`R*
zm_HXle`!8<X<55rjl-156};*K)p~4CWB;>JCJE*q#GWuIhVo-d@bq5!%Wq^VMMg+F
z7V(FlTGM{cc`Y+SRW=%$2IoMdv%6Zds>|#)>;ljqqrv^tSzOLnCVyWX?7Yx>DPy4m
z2!BkVw#jbebFu8}f%#k_7b*c|?OC=`^~@@bLv4G9TV280-PIRJ+1-s20l1JhBi`*x
zU6ZHvS*N<wl`Ea#(hAq-%CSe37w;{{jV`7a1V5SNYbb4(Vd+>@A28b2C(f_Mdc&>Y
zIt-%7&j!`DkEFDINjpEnCp?+!yH<aFr`5%MZfNf^e{gj%<_x%x2&ybMF2+)rPh>E=
zhcawoqfT_g4G_*pjYQF?fHLrm;!Z=c!yNdAZs`DH%a^%hnwqCRzjQyV%w8WLb7?9L
zbq&#Ov~)Ell)?Ht8twl45jK%Cv7=JuS1O7vxg%wNub;~|yj`NT{<R=V(k3dKlc1mN
z2XK~c6RQMdf<LJ85AT!u5D@ei)Z<t@-QYr6r%h(%51?Z2C6J^V;s*KZ)M2O~{xJa8
z8j_pj*NNsvMgB^k%gt3+9Q$V_4%(8hH`MoAjIo}-L+x(K33Tc&jW-g<s~(kExQ6%_
zJg?Op#p2h(y-|YN3Ac!)F(rfwCg*I26wB}m=VuSZ$6|h86}iqjKFQR4rbpHs5hge+
zb*Vimbj5E!-mtSgOfG~<p=x92<MuUvv0tVQHtB(<RcU$dr*OkHiGu~zuqDe84GpMp
z8!<Gi9CRlgwC^c9E1kwzpRII}*m-k~KEuteXNNp@DRb!F;?NuHje~@;!vWonnT2AX
zFp<*Od`+V*L!P&dfX@j^Ic8oX_E!vTcAl<zvS@TfOgq+N-vY7y0VK>5pgoI`_%YR#
z_MSb_cyevHk{H>ZBNnZpSq>}g3gZ~oua7%2WNK!$bA0qb9a5TVu$HWU)%4M=MXQuJ
zt-CPgA2V+_sC?c8e<`NZ#m?Ie;|+HqE2%$iUtFvDp0}!cap{YGw-0zNB?fby3?U$O
z37yCJG}+x21B>c1409WS$Bjm(uAKl0BN6$>vBSn|w*upeyuM?EuRli$))mBip?L9a
zLXKKkr}11Id%+v824*FhgQSOwsC^;b+t<sRw|~0NF!veueE><^Gp>cM5fOV5z};Z5
z(Y=B0QxK_Xp4?nOxr9DC+)W1ZE9^!};EW9g4KnPAN@XpsQH>21qVMVE+=HFW)9>re
zl<yU~uEfC>7z0kVJE7NLk)5xyW9o-k*ITk<6@KKNiw~K`+m}Cipyg%Mm61;Oi_UI|
zavLV&K)l@3Ze5QA+5a<D*fw6YZ>n&UQEi9iFOBTU(0fT|hKzl8F50ONaAlG+{)~fU
zhNB|9^o`L4%&SdV`<<&h)40h?c7O|KYDGy(?!j3gjrFA;x)D8BK09q@Tje5iYRTy1
zP$jvxai^(xgnB=qH`sd!K|m&<)Rz+ORck0$gTijrN9(X5c^sF6WDDO~U;CB#K%JkL
zuI}Eto_q&dgD{lHL0H!k8V#s%d(f;Fa9}N<quPC*7<fLpwH_+bPi7@#J4PO9){^{0
zv5cfFAU!(uPgRAKcM;M>(oXI>ZdT-r>T>)(_0~ZJS<lCFDp$)(z*Da>QzEN1w4Fgq
z<$2c}MK^8#33<kG3Tdftoc-ibzsq}B<SHdl!s6Ifr78iSDY%)9J>GB)sM_6`2unK0
z1`RJTlZqFIV$B!2a6nEEC(-oIaQ_^wV#aBt*e!ZmufPLM-{C14e98=K`zu5^R0Dvk
z1nUtqi?hBs7tjBtarGST<!_yH?C^&73NNsI%nch$-0tE2Nr<IfCqE{I;~{ns{i)8J
z$I4C@C*jJMhvI<`lhxJBvl{l&X~!<9L>@zbCH}eSZ~k&VCYH>M7mNd5UmJBHX@ha<
zO+|#m!s3aQH+4?-Gn+RCG<5sifwf&cKYxm(%6t2+-0NN~tnfag-it1RDA$hmv3l^a
z3E=6d_Uhr@tfS^}J6e8NBcL$JzQb-c7tb%1S7^U2zNh|<LFPKz#idGTQ@+d3hwMf0
zE+|^BgT-_^-K37nsYtEFvrXaBxf5sKg#DAU@;<Ha(ai|PS>8*LmSKD_?Wcs@WVD3S
zylsM`y8UG3uK$tsLs6ZQ8C~IRi%cz-Crpd<v-xj$2-?h)LCSq335e%cI_@ydb6dh)
zu8xwj+>x42{)q9GvyWMLz@2r`{x-9q?2NCb%yI$6XcqX&cnxWQdd}^r48d?@QVi!$
z{k8JP!bD-(cJ-50^;9R+^`N5`;d;A@$(cmCJCS8~vI4DyglJS56Q7OX3gXbr#FChx
zN>WHyjzC+X?atYg!V=H>bM_q#K}sh#<E!Yz#KH$)$V4;;tajFlcxNq!EL-X{i;2?E
zE}fZi=mARtG>Z>A>qLBhwri`Y6j?<Tb-n<|tbX?jKK$3QoM&E#sN1P^m#4V6suYYH
zs_R|a+U8iWKdK!U&9c9j`{W|6A$wt*ahBL1O%1nYx#z?ImBRr(IY#1tg$*f}<bP3z
zSG$q(6H3ROxJhk`KpxwLc~VF>nt!U&>(iOZ_^Dg@<#rtoo8aW1%i=47t15!RuXv^i
zQwjAl+#B1WxN|UI$}97iiC@XQfiIWO1#w$??jLjlc(frA8Ur^Z_TxFbS~bNwj<x8f
zVq)M*CdCsI3G;mfPi1QXhsv8x;dYOfW<Dh*)PM4=&#VTLE-hkMKxlDdP`rjwzI6PS
zWklZq0w>)%vao6yIjBzU8n8#v?8isTsefDidf6e(+KyUpQh)r}+NPj@F0oSzouzUo
zRS4Y8c<2T!*NFg)ffj`%Tq;A?(8$fsIw(9B0MWW|Ej$yD{>R+NTI54+0V`S#W{ne!
z_(%lC2B_Q`5o@V$h+pv*vP_*bN-Hc^|C)K>SE3H?=5KfQW%Nao44zRN9~fB<=D98g
zs6`X=e&m3jW%P5NzLB&0b|LOpYMk2jG7+UwANLGgN`lv48fWYk(qoU>c~ZGv%)|6h
zA&GYy-AX}OW5z6xMz1D#MMbQ%XM{H~f%NBUpl8Sf2zG}$2O{rQI5?FH^s6On<AxrP
zT!~v9+BX{*w(kUJiuQB{o+cN9IARr5id*v;?7u)LqcuKff_?Mtsq#rZ+*Pf{)b{P+
zz{`ETF_dmGa7)P^JyA;ao4eD=`^XB*Q{(#M7GE+w`yz#hq>eY1#i>&Zw2}o~qJ%Ax
zf4`!U;%{YrudKpPULv)#;%M_Jhu5>4Q7Yd_!UTgJG$&E4<6OS84u;+t`;p?aWZH}#
z9y7?yp?^N_{mnkK1$f`c_|A_MI!Pye%Q<{TZ77CUl&Zu&unoxykm8hf3<#tSxfEqD
zW}H+WBWy?X*oFR|_Ni*ue?6j7rHNREO=$n&cmGRM=w1G81hLFlu&Ygp{y#5dD}bQ(
zt^VJMlVT#GqwP|Uyp4nBWCriN24-53fw#G$YP;jL$kTi}Lw1MDbDzG=QbDxmiUlW(
zFn3BjGZhZVW*;4<OCmow&r`s#+Q8%u$ArO}VvmWprxaCrpSC8SE-3}wf6uY)FgUiG
zC8?Tv|0xb_Dp3Xu&!-2__2SZs4A1lKOdUlILaP|}9V4!`DCfJ+Xb@l1F1QhBc)v-Y
zKBsK${iP9MwkmLic45w&^Iu_U$joAq*bYFkgW<E|4h!&mMfJvUt>0I>hGr8np#9gT
zp8@hdW=M4y6)}kpBQSV;+mQO0U7G?JSNNs5-C-!g*=4f!{&ogD-N?+gI8e@AEYMJF
zc)jqHEOi3f_$U)lf;3?M0%s?9SGrf&JAfFGmMsydt2R)vc-N)~E?l%nu>mt}9G_X4
z$9n+xprPnj@@0ZU7`G#nI{!m1j>85)q;4fpw8;DYHHB`{V`<~J>I~d3EmpK;*w>$$
zoreLJ(aXGc^GwzzBp@zH3`eHWH;#_O-0%w_wtrS~3?FYOEy<?qI}6ukza0E!o51qq
z30SUU5g2iB&T?jGLT$vTZ?#ICRX2JDS+fjietd2JXtY;C_1=+(cEalPzPp&_Hrvk0
zhZfA3C25oHk}pB#hqV)R(CST=iS*K$?l!orp=HzB_}MWy2`S}r=S=<aRk{J$?7qkk
zuff9jm>+j3cc~W&ZzRIkmnr4*#%U%BGd)wekoQWoDMXi7PpVEh>LUFm_>%hTj*GsV
z_2x7ER5pP`n9c-xt-w?80{mu~M~i5-!25tA`}J~h8{<!#$XBkN68ElH?r#?-Td~NM
ztGKEEzT)+5z_j4VjpXZ#JLA8xZqZXR-BZ@Zc7GF@mcf7GhB63vrw}J0`=aU*?L8WH
zbIphGS8}b8x~u?#Q&KT09`b5tsBgmUTcmMFisB&TQ?S6@qpaDz`cb|mu^^E}=M40D
zKAExX&M-vVv@w0jJlC3&Lv8-B;^^4IeJ~{W#e1hY!*oZXyH$pg^rCG7@d7QFEqltV
z(~OmKj|vQ<rh@9vFn-d@+3)EjIVn*tvaD25mK<o)M!4hFbD^3+yuBaCiy)THjEo#A
z0L;1Aru`3k)PjoI51#JyQMR^o@+La#03NqEMipv}28}{<nDHl#@af*65IqKi0L+Oe
zdadlJ){*JT67`F%LOrKeg_1drJx4J!Q)b|R@v5vJJ5#*gc0Y}e0^i+W&T*E>q*JAt
z#3YCDdI;Y9SZm_Bur*kd+7AjT1<rwv1ApmH?(Ik|CqKQe@~q+#6BD@fb8@`qjezF0
z8p;rbCV4=LImS6_LLnjVx~^vDXz8@HU}$^lpiZ#Om0-XwUwCtpZw}|5a@;P-1pfRh
zEQ0<}Q^pKc8^X)HU49g89}>draS`~y<>vfNk?S3UEXy|xdxhF2<2jP4y~AVj?lBrS
zik$-N0bi<Uv%wh&z(kbZIkFriHw4>aap++@|HXmq=l=C#b4HbeH?{4}PZA%K_+%>q
z(H<sJLsTrLych1mtCz_Hh(_<hT(V#sccP#A`;5pjz<iv=bsoH0o^sWbklETjzu>Wx
zLN#EHqF;=}#~$vNIP#7+8!ug$@E&-Z{fNf?y_5!4$grF3WPlr92AQUr2=zM(Akf7r
zo-~B&>NjQK!M_7krIsdSgzCUQ^Z2Drz8c-$kJNvk;kJ5-HuaHKhMu{%l;Rk*jV053
zt^bsQ!Y;Thz$}c$H2@{B=n$a1%@|}yN=h2uBj8~1F0s6xztlHc`jH4^iCa~1=o-jL
z1U2H_AOWl&Ds-ZPE;g8VTgyW$h8*O37tn@9TwW=;DnDt#X2!^Dvi4fp;7v3gNeblx
zH}(I5sUH}|P=nON;rz|~h47bN<e|Fv-E_OCwiBA20gM05&1_}WfT`5(4ekGi02Z8J
z+VT+<@XG|P2HZTeUWP$d-;l1f|Av3w!`1f9S(uOsQfW^Fi<%1}|D{n?h4kdOHJ@pd
zSb;Soe^y3ZzadAdr+Xd5+F)GTA1|dsB@m^~_Ht(lTzQdN%i7DAbD(y&mV(A{sR6?M
z`57D`&QHc()S@A3Si|`Xu1SzvARex)`+;{YB;S3fqvNN!l#oG<dW6!aP(Q`w8qK!D
zgO;SpOyxMmdkW2KSr(FsM^ZkJcM5N6bOq+}GdG`F%wtb?Pd2r(HYi5@J9e)Gg7E;V
z62<p?DWb^C&X+=Ctxp<9NDIkVQR7N`Ltg8X`zZQbam4`f`yzLNo^J-RS@n`=CXzZ}
zoa0D}MJew|5h7F}y%6DHROb|}|Dr;~-&}|G`g0$8_URGh;ubJz_xO+>d$@v}3$DsO
zQzg+EjvXf4J;?3Pg;<+p>*pylzBt<Lt4V#nUHNNFwlOvlp87{noVZt|_Lt@cOh%um
zAI_BUW%|7IaWoh-%B(oX3>7z_pR3?qfu$b)=w{LpYFz)?N}-aBf7wIMAr#{s18!GB
z(2tq0kH{L2wEE2e;j@jROy0AHmN^G<^6bxrKc`>4<;?K<Q5lLFS?xfk07FH7mTn@H
z2m8mjzsxf`2<wKN<Nb-`WNC_!j5?pQw_b75GE2tC8=GLNH+OgGcA@71DkdR|&<K>t
zk_b`1un^VK1QW^bZ)=L5*;s@OYj)K>y4?gDyrxjr6<Bq3=DzaxzciPTi7D|+naW4+
zTR2w*w;2k)<$k;`ZBp?5$WP__`EfSp7XNhBPCQ2z60mgu^vMDEr6WlIoW0pIq0z<!
zez&DgYFA-{B`<sIWWj$9qmCX^?t58?pn2w;9dq=Y6aoghLcq@}Y<{?3s85na8Gf1F
zMJ^0VEl>Ms{AO$2jye{id-nBXnp3%+giv$dE@JbTJR4TAKZ|&+b+KWrQc`F*(9X^V
z)Z!wW&`dFgXd$!)HBR1ZIm844q(%~$B+Q|2DC|u{hOKmveS`jlg`NEQXS3YSdL^**
z{2A=^+u5WxVk_0B<twD`c<XfQr1+SD!b?D|lzwWL`GH{s*-2-B3PDQWj6ba(t4Sv#
zyTQ3Isa2bf9uYU4wcC3yemXZr-MZgftw54L)7%ocNy>J`4}bxq^_HUt!3VkOQp%lb
z$qlQ)qYr=e9mhvcvfZG`8j051=h|6xU-!Ig^2HAe+{aLh8_L;knel=ZViLQYEDDX(
zy4ggmgEg8cIaetz+r_;F&c%j}QDsC#79M2uE!<8kl&TYf3<g%+n$WsTlCc`&=sK7!
z*Gn;LnRwAM)?nPK-sJ7=0+JR`f9Gmg$RgjkVtq}g`#w$8QM{wn*eUh@Da~R}v1IJR
zkUBBD-QGSWT;MJ#1ZNxqk=E4x2R-IBi8qj~sOu8i%^X~N;q-8D&5y<AL&gAri82^Q
zKT^+V3@CWqJkD!5nrnd0Hl3)JaXAyN5<x~77Q08VemL=EbGq>Tmh;$pVL7y#EV%!i
z+<2yp9Hf3Y!EWGX@|~rg+Ki{QZSnZQx#=FiVrq5ZzwPb&P9%NCX68{G&55_I-=MtZ
zA@p_r<;m&_?;QpGxQa?C$y?EbNBEgJ9kAM*y|6A@0&i@wX@9e!!#6(pOY3YFtxvYv
zD6h7*%Fm{ifK&bC?70*rBJ$7nA|N*Vv|+Nj)RPtArMJ9k)tIw#UR+GL8T%uDfqV3x
z%`Ho&bsO3zVV2!v@H>P;Fm~2}dNGMBI4XKIRB#lBSlUwu)D*P_EyXeoG|kQ>%}oLI
z=x3o@r?ukSyBm8dQWn$`{177R1vF`?w>jjkw34V)f7)8t8oj!zpN?B<j((aWE3EEq
z=~Kk#Omoe6H(cu)nlPC3F-aQV-ER<-)gG#B-cpM~22tk;g06qj+EAAD!^p81?+gTu
zn{Pn3Ox_#@h5V(-U%Q7xM1`>uppo7vv0{%bBFTHB%4x}@Ny_&+)*~>uBhJ{{TKdGI
z=a;^0mnC8C2bKY5Z&8n_ZN%i@`;_8`&J3(k=;Q(0z%*LFzAfhdy_Ji3Mb*0T+a8~`
zLJk>^HLCT_-)mgPoilCyQy<eQ^=WQnA7l<vc9NW^@`|b`bB2x5OXn!?8q$1?z_ja`
zI8(EuE`ZnkxZ6T^xvjael{kp>y^xgp$$~9A@eRZ|F<C{J_L5Vc9`nN7zn=g&86#5;
zmnC#JFLIGtLyCH2qLz@+lbV{2!v=hJ9^ZDfk$Zdh&HY;uAJ-HMaRu}nPE2(z^`lZ@
zy->ek-!db#b8k!V3djv!BjO$89R7Wf=2HCB=UhnR!k@S4MiMSATls9GD8t~-^PW~1
zP=S6MWw-_*JJnmzmo7yzzXZ>Cbbne<$v5@tPbA3#XAk&8EnAJJj~J<$N(uXKqT2sA
zQ6c|BqN@CyKKAc7eHP#9IrM+W%-ILA=rB5x%n!;P%KT{kmyc-u%*v%~)8?GXpM8dJ
zXJ!qBb-jRMzEP~Iw+<BzOwRM*7m7%^fT>D<@Ei0PR0K1IhnN$$CEWtjT)-i2Z#%9&
zm)Nlxv5f1beQi)Q^6(hq{bm}x>Z2t{VqGj?AiY<dS2u2o-s8+M_7EAbDz2*yet8EB
zHhBBP`xke(tWY{P2IbefxJ$E+zA#p)aDI&;K~rnAk5Ox3q}n})>r+DD_b<Y=(I<lr
zjK3~Mi+L`6qPktxQZVHyn0Nn`SAvxwwXRz>ATpmvR%sttSgri$V`o`V&UL5DkzKEP
z;uPD|J5C>~y#?Q)ln3$GU$vW@yDO)9jb7}MX#Gx@2+U}xLn%zSiTEUeE~Yy0JliV6
zGzua%NAH-P^?RCTiS7f>J|G^=*+!Hu2$9Y!+G@%AR7^^i*aDGd%k1nnt;`uUg;^g+
z6QTDy6O;HlrKh<PYhS|`pl|-tuvNm<31d#(HP6dE+->5IjS28q?C9d!9aor}OLO9H
zg;|T^Zn@X>Rn8(VDvgk1vJ~Z2Fx%+Z>Fzng?py-T9E)ZnWJlwOOwIUEUj1Rw=@rkG
z=ud1MwNP%8+SbeJF~W5D<L8_`3tU8*-+yUpH4KV2F<!Pavwy5dOx9v;&0pl&?A)vq
zZ`9g&#v~8ahPmTn618{?2t&Yx_ETg+hJlHP)4JT|EB~C#imCmIEuWMG`-e{@(`s*<
zMg%^UyO8mV%sWiBaQ3A#?>RsM*OW&#PInXQTL(Bz(2a?|djq}-3wvi4>c=s*mcQHU
zxb;<A#eAgIA3&I$JoT&}H*^wr3MBKMD`^#s|BK5Tjw!SmoprUa=jVNuK4WN2GGdG0
z_F;PQ@^|X9$V{=a!p;IV4c^7=hYXB}t`?4%P@fk!W;X5KD((Grk*v9F{p`s;U%2mh
zp|gtOUP%T)9gm1S5WMWr#Tjp9Nl+|^ZeSE_53*x5hWN1>?bJ3F^L7toaxQpqH9B)!
z^9CNAmj2nrUxmvo@ylWLDI+<+H;wr$V`8Yb2s}ten6vfcetcKH6ZSY-Wg>YjJmN}{
z2sQ(u(2P-+LbNScd&>H)pmV^Rcnm{p_0Db^I)0prvs>#C^$T#7x}7bWF|mj2Sl7aN
zU>E<&=>9LCIy&Dyq&TzghPfa=%nMdPxmkzjzY9y?#nfuea?TKuf$+RpB64+OEQbh@
z`4J{c80wT^H3%l)vE4a{c$tH|2ViR|Q3m=<YI`v*9p@ryPCzj#I6XDLIC{O_T~dWa
z4fj()eg_aIQs6#j*p2vw_>Xm?^$Z`+Kfbv&{!!E01*ZEJ7v*_Y`0kns2|s^ozG?;y
z!zi9Oo~V2L+4;75th!y51J5OoIv$<~J{_vCvD*xs@tT?(@#Hk?Rp3ZBDwnM3e)?$r
z%}S<m$s;b_DNQ}Uoa}Cy3m#jaa{tosl3sR#KiL~9t2LkyGc>6eko_qNEPYpg9SSt`
z!OchD#$ckw_XVk8l&W&vw*MKUg>iF3Lu&;6b8tz6$0}5HnCtEA{m|kM-i~GDrFLmF
zJzQsm!a-7hGD^NvE|$pJuB67lWrqbRzxrp0RXsHKX(MWIT^PjuCM$gTxp1;ll~G`3
zb1SvQ30!X2@bk8W93r>Qsf|viPy+-}UWc&VT?^O4Gc&DcS#cVB);WN0mS0h7U#!@~
z(aL!(;FOh|%|suF&<DuGFVK&Wg&}zAdvSSiXmPa6E`x*AKHQw9)+~WfY=qtExHRx6
zXLFJ=9%77NJL;NDxICFn<$YZ2aoSHzeS4<iB&by4leiK+oBEZLH~hlu@AEdN){$f#
z!cZ55bE0&Z)1etTNw&RH5CsX~y<6=Pu}1>1uY4U+@NN7V&2L;T?vO8^dC#lrEX<H1
z)v)3o)@WX9SS{0&yH`uSQDn^kH%#cd)MVo8)4qD-v_;pHaQJgJa^BNWPtEjgqj&g1
zD*(H-TBjjM)csL#Q?xCi9Zg+!6R9+Y0D+V*x`D$ik1p!)^0zTN_$w}H(|A$YU}y5A
zE`>Z+{WG`9ucY@C3c1@z9&mm3l#Wn_2ZP8RwAJ?xY2P1TwXeb%mQJ)ZJkSzEj`#Qw
zh-goT@{F052!#GJYSh8DMGqp7o3&@uC8{y?Ldkh$&?dg^iS_3ON}xCP6>?~K)i_h#
zn~I|%$%1LvEg(1C(07RnH{~-CT&v?U+TGYD&=Zvm5?@5JF6TMdztO#wi9$Ycg*~nL
z?i^vDi-X34@hJiT-1x{&{ULza28=gee9;7KF!ss+Czbt|cY*twWbljoO3NqMu9O%#
zSbygXYT-7?JYS+Y@d6IZ26H=N)HbV+?Uv^|W^w4y@T!B}YPU6#BYz6Nk`jG@iN!F$
zL!G~RqFJW2?l>4P<_YVgClz=r)wJcP0#S{5dEPsV&aH#`*Dw5EJLm5=W#!xVz@cvM
zuil?_%WNxq3p`%Q{Y(CzseSwWyW#~*&&yRDcl<EDX^-iO8$a?saZBcxt`~`k@HUwg
z<y<_q%hs+yg!9moeV)Mwmi;ODvE@E*oy-mIbroKhkL+q&F1>E?!`IdOtr&A<_)R~z
zRXEPFc<go{<9cOEkZ`?R=IR>Vjk|tq{Ky{vG5^g&^&_p<^>yBD`Rf%O&wtm(VBP#&
z;X}vXeU$Y&Ei;pg>pRPH(<{I=Y)e1NAMx)nyLEi#x*DYq|C%5E*wQQRcRAR7R*<^l
zyNi!2);;VMaL$(C?EjkI6w)=fQs{MR{jPWad|$lX`sH@i_mK6M?mvCK>VDj9sqM3_
z$a9s>=lG%jkp1v`j@kJ_QSsSa8}IW99S!2&%=Gxpwc?*Wj~*J7^$D+hwA;S5UP$jx
zrF_SmueFjUvh{~|O}z9~+q-&mdG51HL%;Pr&n!+ab~j0X=91va>R!+DAstvUch60J
z_^$EMjj!{6xc?E(UiaeLu6NTPNt(Rw-t9K+%*%6c(@lb&9F|sI#`LaXl0m@yMSo-;
z$nrn@)@?R>!A;XG#$MlrPn#Z@ZZ2+Yz_@f)W#p1+P7JJ1(sDE;4~Q1mY%Gd;@%Daw
z+26hQT1vWJU;ntjZvXWfw;%r*KJ3kXzmES%KSzg7Qdjm)zuno}?pYQ-wP`<dX!X0;
zB@fKpjAvc{D1Vfn?T6K6?_JmbL@#_%wre%(<!7LI<x3wXa0nW25EA0B*Q@FRZtk{c
zE}hQ>4E(Nt7e72*yZn!`oZyu$fBkuxYo~py-}r!cO}Y4|cSa{Ce&1A=SNKo<kLkz#
zhyL;(`YwKtarUvvXCG+GbKS|D5Fg2YS-o{>rkJzIiQasTsmwEGx&1x#r^m3V?p4{f
z>p%b0&({JL47*-d7e+6?{V}tS>;2Qq#V^)N)VO_Ee)z|lB_HiOx772k{W8yY;nl4i
zYh91bztevvapz7Mo_X0bKg1u;=gq#)16=<N92&@~3U6WW%t_j_9k?LZV)8VHqk{L-
zXSh8)!Mx%@_$=U1@cN!Q{`(bq4}RqNck=Axxco<3JM!*rmz;IFz*)YHJ1ait>d$DL
zX@2r|>zUkRz)e=}Kg2&;AFPx6;PS!oQO8{y?(~PV|4JVUw-WZuF68z&2d<hE3N?a&
zJ5G-otuuRDw54+Gx?MFp|1(&-$4$TVH_a?NUjO*C{|u$~?>=9CWtyGc@ARX&(Fc=`
zXKc8b^R<3^<;~AFKi?&N`j-1R|K{f1h3EbxB&)Am&+#Mv!{hiLiI1hd+b+KUYHM`=
zU==&?ESCM-cg*l*J$hkw@%D*(Tqe7{d1P{r@kZ-DukTewU45PR>(Ac)u+e;X?OOT2
zs_`Fp1iFecgC<1==Y+_*X*<`Z|8tI$1fC_<3>?Q$JzBDbp7*qnWfzeRcVpm`>L;O(
z=6_ql|1)^%zFq*l(o1XE?8P@SHk{@IPP8P7Z1yZxDZ8kX_36aZClgwkb<P}}xoeWA
z%`D)pjU~X%hM*0HAS~MYSe~&W{gK+VUHM&mtbrXJzetwV>1!g6mU9$v@CdmoE33~b
zU(^--j&0rSWw$bZ)o$%=U1Mfub`f;Ym78*(N?>bf07C!+1LJ|!B7vd%)|dfP!~}d7
zA;<vp67p8Y&?Wby4<~i19`SfMU+Wy>?_*sWO!vP{UH<2<^M8hV_8%AjE(Nw#>;wKY
t+&}owB+!9<RBkj}(9(pA+h)|Ar@Nvw2LIh~=qm<+3I8wfMiTk|HvtpCChq_M

literal 0
HcmV?d00001

diff --git a/_static/bonsai-workflow.jpg b/_static/bonsai-workflow.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..0481c3dcfdc1105ea5676c4cc55f907580168188
GIT binary patch
literal 33481
zcmeFY2UwHMx;7f5cah#jK#(p~YNU$@2r9j(^e!b3YEYy%0RaI8=`{rDy%%ZHLocC9
zC)5BToP7H~d$09<XYaN4+1Eb%I@f>wZ!$?{-pR~7GxN?o&;8tUGkvoRxcyX3Lk)m~
zg989#Ux1rAfHDC0)~!F@*b5K)#wWtZ$HT)XAtWRqA|oLqBPAgvC8xM^o1B7*f|T?&
z-EAsr8d_RfGD>;|IvR#MG_*8-6oP|`{SO{KF+M&q4LK<}&HwA`rVT(vgky0l3>W7f
z;1(4QE)~vA7l0iAz`@71_Kybt-xtm;Y#Rv(iHJ!^u^Fmw18(8q;@-l;{iD^`+yU78
z06Z%EI~)%c38;175Z-g65q%$@MZ~G}qn%c7^pH#Jt@{UJ5;}SYMken2JiHJ1#3dx9
z9!bk6tEj4}Khb#l;^ixS14AQYODk&|TRVFP4^NPnH`vEF_+v<DSa?KaLgMG7<diR|
zY1uisdHDr}Ma31BRn;}Mb@dG$on75My+8Z<$Hpfn;ZxHyvxt?|we^k7-&@<rqvMm)
zvvbtN<sWk40C4}8SpO#3U*w{~%5@754;PQ{54mt|d1C`E6&^mvLxMYsx`c1ssPBos
zC!$e`&-&3$%qgaKNc+}(l!T5;9KntJL$rU9>^~>ihyRjf|0dYq<eCML;o@Ki50?r6
z1YC||=v!KE01CdY*O!SmfT*0m#9<_;)D@^ly8=@){Nx+64qUM6%IH9tvuPr?@?CsB
zTD^u?IRO{MA5EbR{*i;qKAY-%;m9T&rG7Zs^!llF*em<xC*chC+{p||EMG10JU74U
zE}SusE!x)H06Na6ZUCFZS~q}JJqXeh^5-~cjkAY|6zq+;g9F{_Rh-#?I0TddEB1a#
z`P9!vBfrAU^N@EsXb-C{O_OIALYoT%BKZeo^T3~ds5LkEc{zIby_Tt_U<?l$U4(%<
z%JZ!VznvSv@n-%FV5ta*Y5@Kvey9`1PAASDKC7M1U6WIw3N@Wk_fd$VJ@GJ|<L7;3
z&EQ1dnUGTmErGCr-H-JqJW+Crc|(x=BELDq15117mmvF>lFupiZxx&GR8n$IL;h0!
zU+X(c2j1Q?PU;ddcJT6kG2~PmuTJ)RkBjVGMIqs5o*WAK?d!Yp?0*@K|Htq=$`1NJ
znf|A{Kq2GMAY4dJ6~V8ix5cTkeuvIbk@TpMw?{)BaA2l>=G1z;06dtv0j!jKL>b%w
zxaPEP0A8-wOs89(IGNY!!sv6It1}3y7+7h21IXCA0R#@;0BFs<R}SLn7D*=+IE^f}
zR5TFsB;<i)X1p?=^&%#971=p`YWCP=>#gOHubOy}`#QA3osZrgqM~$>ubVL=Raan-
z+<<rn)3uM+X9ICU6cK~(>#lN*_pV(r+Sd)i<LcLTnQN1cC-PEPk`@>HndiGV0OU)E
z&A<&nejIb1iQ%ghheWX5K@vtt04ePQc}GFQx9z_!ZRz`d6kRa(h+FI`HIlyT-v%l(
zA75Px1r4+PmLYgwY_GsEe#*?MbQEBzHF+<CU$UAmX1!&oQy?YsLO=xYh~xdODOZcB
z6UK!bz~5CCw>tybY=A6P-T=A{-h+?Nh7U3Ckpu&s=8O;uR5jC*W@CL_Sl18r2eFq(
z!Idj<-Cp60S7tU&6*$a2in85)FZ`bzCZK+TRhx9PVUcJU@QP^r_I2Yx&mKl?@5=3$
zTeR+(PlF@Mg(n|l5vdhO&IFmMxB(1%V%Hnn8$e{v-*v+){obBMwuxV90;hYTyW9G(
z0LwLVwHts#tybek6!~#a<G4UQuJ7?rC^`SDiYd<lEp@G?+6nQix@14UMk9U0{Fu;J
zqZ6O{74dGrQ_RPIz-f2?VI`rgvK1MKXK6MeujK|1H+};!ow}%2eR3<3_xpcmpbS&E
z|NA_H1Le8`71jycqd%X0HCf$c^5GzU#roC2Z2}FJ1bj=>L1)9a#YBW`i7amMK-N-c
zfWqerwHWsjviLovwfiUbbgJ|1?!ocxu$MQ0kILq(o?7ZR035Cx00|I5H3j)iT6hCk
z=E_8<Erl&h71zR6t!@CEn`ehQ;bvAy`ZH<*1zs}<hALQ&%|6_i)Sqr?-DmCw0ErW`
zd3;ehq-nWq7Pw<r4!6*7mXw9jeRI%~*bHy}%sw;hV$}abZ|^=s_)OS=Oy-xVX6)X;
zS#|>$q1U|uv}pi4;!rA}rOR8}Kol_#O0OpQs1}3WtbS(6yF!3HHvn{^f*Z<pIg?5~
z8~tds49$h1b0w{>YntGytM>dB)AU1Bi6eeG#a{X0yWn*UXJj=3LIh6Ksy)<BmkPD2
zbtO#QNPii-F<(}!)%3$T$S1u$b<F@GZ3&1i2)0my{DZc<NOW($ya8BDDefO!LyRvk
zGAIv)5sR?HP2I_q=VOA1D8oou?DjhQe4wCZ;|37%4!`eVchhHHN%YveCfjL83k(&j
zHzued4s%NbL-Uua(!DXQu$=%^_zhs{^xDZv*Z2l-7{1ebf<FN{yrXlN^QS)t{mibu
z0sLvC)Y1234z6)P>}b41G%xPGNbH9DQQhnY@Lee91`w1031R>H^1hQr^8dpI1&Q)P
zA%8y>*zqBXZM*?IHv8A90{Gvh6_lQTS-L@+KyoP050u{<{6SHZ!4k>e>q@0!objIm
z!eMIPDTz@#hhgT(;cnruhWt2}gx^ClGpH|BKhvEMb)lC8v-7P7P=9F2FCk_*tB<HR
zhjCpB%qZmQkgeVa27f^&O>9qgce$7JrC;dp==Qrmj|v}WY`c3XGu1fy^nt1o#P0@h
zf9dl0jj$g?xJ4Fdo)IAwHV7dPq(^mhw@{;8u+NNERzc@X8Fv#>v9SgW3W6ujWmfV%
z8MpCHwMiU~m7{qi@Xj%=^*YgCcU3qHNZ>E#&5Qyy$s?Grv@|T!hy`2bhu$Jh%+zwP
zvlrDvmG9*%<8!2w56^sl(b$-D3i+1SiqX38-xhVl&P!;3PO_ZE;nE6am!Q&)Y5k%1
z8Ta~GBI1e7c*^sd_}3W&rU@H+RPbk_TwF4>%t=AGfzM465#G^EnsEeiB=s8p;1(e^
znJY8cOn2P`-7&u0a8*FAib6kO()u3m=3IAxw5ncGnqyAuDr2CV%D|#2Y~V1CXU+e`
zit9|{%4Y-2-Nuo7AEZ6gXR0hnZvc5JsiUU^4URM6zo`4A%2y;AU3Y&TQXn!b#O2xM
zIva#C^NiH)RsN`rg%eEx=)|8F1K5H{npK|#Odr8^E}dJA6cEC`rU#d;XuOnc&8*ab
z8vuI&rXF_e5c#Lo`!PO^?tg2zvz@^grWmJLsma{@LboX<jg+5TYqyZ5kuE--hJ4`s
zi&Z(Ugm-5KXAM#!cU+O?Y;0~2eoj}bQDMmaoLC<aKk|Wy>$W0Z<VKUVrLE6(B#2yC
zFJXowAgv{it7H3@5<LkW)dsEG*Z8D<K<6jN&u{_%7k&r{{p20pG3aR?$k{;KFDi8x
z&+z)R!pHS4HOpSx+hSrumb}ethm_Ov#kLy&tiE+;|Hp2XL1bBqH~R-0&6tO$@-nN2
zi^)g7y+!6`yVotj<8b|C!=`o&v<Cgl#16oegE$~Ib$D(xz4&`bD7J;~IAQAQzKN`^
zlN0lXi9IR9is&689_EOjjNN($UMFa03?xV%wu7=jx3ae`GLsN@+k#!qYLRNeZ%T}F
z<LAsfDtjscEc|3akGX7-F!{rpcx(HWxta-o*&>Y<kD`Sq5}vhg>@(@!N$Jrcb|0;&
z`{uhu*m)RfFbpVpv;=550x@>KJS~6%e5&z`taL0GC&q>=0C5~b{uA)Nt($kJHDzsS
z>-B!@@l)jDw-WGx@!_EEyn5jlcK56*#B(nxbY1wQA?q#VP9Q~5)k2Ace_IWM6MSRs
zC`<J{6P5c(YXt)AKch~PUj7=uKP;!2yEAJM>94%P%B)qNaSfua2^Y}1nzpG36<uCU
z(nyee`{<VBNnzfh|E$zm%-1yQg@Z7`(<or|T(8R!bK}eMSh<gY)Xz_ER}?z!$+1wW
zfSz8bEbd+KTpPL{ouVZ#DrZZio=rhd%{28ZLb!}P`x1y#+eijSxq`|lIVogTzfPsF
zQn>C)@#id35<pyuZva13b0Wy&KGDQgT6(9P#bx%*eYIWRq*d1$r`Me1o)kZz*p>+l
zaa}QV8Cw(-ed$Jc<`e$&6aG@|;G2#zHSNdk?X2EMXdlSXb;rr^HDmxOAGl*)TZJUZ
z%$0fK;`$~HJo9*(j+5cFwnyfo(4$>n{UmTmM$`&(#kl<$eFNxZ6dXR+=SHgbHo-A`
z_Ev9p@-~H^m1hB1Q@J_uN<}~_S$|VCd3Jm0>ETI(NfWddPj0+C4b-^@!Vk1eUb8as
zhh5wNQdZ=Jk84*_(A0B|UniPC1zRP)&!nWtT|JYZBqPHdqe7oAzO<D4MAY{P_iN_o
z)z2uYe91%~XeUJyEi`mm(!4u&I5#Q`Cm7e6qi0!iLzRU29n0QIX!;v(ip0Eadq+1P
zk8E%<#_!4mv>+bEInVJL7#)(zj+>B!`<29*YcBw}4BN)Uz42^4t%OWne1pQe1ZA%U
zCHVcI6#J9QM$W?VOpose#EU!>E8zzYeCowxS*flL@Lf`E82kQ>%XY%0rH?B57A?9q
zQ7}Q|A>&rYlQP^D{2Amyi10Xus-zJrG`8Jiw|LZ{E52s>Al<sMhpVGE{;U7o7l69y
z5PSQP(yZBJO_G}>$Ze&uZ==CPsiKtGrAvD*<s<Wy3!+wdd}CeE`v7CHfmyy1yu@f>
z!b%)xyVzJqPRaGxx%~W?eT1wGgM?nu{2(F5eZ=~+G)B0)AvBLmv55m})P9eio_fIQ
zxqt5ak99VE%M>#4U&+$eB;vv@;VFnE_QA`uJSgi^IO2Qd5p+OXD_f0}Xjy-$o(P8s
z4}Uqq;XK%+^Xz=qcTPAO|E%pjhzqV&C;QHuxc4)msmrok?1W(8I%=LR!v6-K+&Cto
z8Q}4(rnUO*i1z!bcb$~a)!$MmvW82z)E$I4tahFsd=Si=JXot?ZOZYuPqh4+Izc|h
z>!nYaCpS&i^IN)_s(?TJAmo<1Ophzg9Go&(FE}?iD${d4m0~vnrk2$NG5<?S&VbH7
zY+XD1Pf_KDp1jl#g=>{8IL|^in4>TBavObjU%OpA=fD_4KoCB;D$F+K<u)(Fk|405
zHHqzC5<xULWXWE~l3RCUhM!g7KEn)E-2gHTAqUe~rMXfSp#P@?V5Ku~6V85h2J9#u
zeN5Z>w`kn%oX*8*80I?eutfYhzU$wjSR-W|6YvIWvRI<|h7=`En;UnN!8w7AW^LlT
ziyh6i%-8AmSOv2ji)#G9Dv&m_kl|9``7R{HDEm|Y(o9X}@ebwn6+1%x%((!4*@-nZ
zdr_~_cyHDIU0nC?%b>N11VH~j;m?Hz@&8>N?0+j?-|jofe^{O4&mTp$uZwfIzbumg
zg8xn=-amCv0&}5D<=PO-A3Z;fE;jaaVcz`*UHIfbuJC_YzW>jwQ#HyCdej4NGI1Cp
zV`7VP!?P-N?<*Oi7R5bER&;1%vAp~6atzX;x+;80?<tIC!(wVbLD&gs`Ij^-$o7{!
zrbAO*2ZV;G7IqR-RHbq)@>!d>F-yZ|e&6Nj59M@=?$XP;@b<CwCDfB9ELfqmA94c-
z?rXiEi2G|ojoo)L*S{wI&*v(hvq0tlI@;f6C;M$LF~>is*6bE${rvmt?iWG;+52Av
zz9heY|Cd{uF4Vlrm<^umoV}}$uf5<4lDnj`k3XfYRb~F&*7jr4-ynac-QF4TjeyyX
z$}(1B>V8Kd0|1hB`Euzn9z1{vS-|S5go%(NR;SU&41Y~y|Nnm;MiW~T>a)BUA!E_8
zQUN1essx%(#DEN9+~54=WyV+gLzpehtBDaDb46tBYLx{A76f~dTk7q55>X#@%vzig
z66PhJf|P%MBIw->pOvdH?#74b6bhKT&In}L8&Y`f-+iOV89mkQYY_%=4bP;Zem1e@
zsP#gx(W^e;rF2{ga9|gw?d(?lj<&9bdJ)2Ax_IJ2&o}%7t=a3=<x2XljE8WvndcN!
zPk~3FwCsoe&9^szhE|Yz`K?N)10nINj~Q?u!2;!^1bW`i->Ly-?7Q)ZyFm&CP%+Dd
zN+rVebpukQ$(eKQAv84812H$3UR&Q3Ij}o5m@3FI`%#31aOnNdgQ>`sB&MiE+dln%
zx2lyLE%QZh%fLIBv9#DkKZ4Yi#?T4z86wrz#MJNS9%#p%%rAZ?u)OLx8hc;zJm}=<
zin1%{o+?>WMWvLsIzzAY)_V&Uv@S|os$+pvHRJ5UrJ8i}lO_x2xh_7*W4xo@jh)`H
zW)lS4_drg6^2rna)+Mtojsya^Z&}{m!#~Jc70twLxrJ@T7W;p^+Jb`x^v@bvuJ_K_
zl}1flJ6Y|~ip%nBNBT<-!uq}Tjfj<Wrd%LQ+x+F1ZP}5$fImX%4dAvTT!DVH?9@yf
z;q2u6yiXuz+LuyK2jI;`G6GA1IK?qNt+d0?z3=aOv3GtKLe=-A7ZLX8oGod~dRi*B
zt0%PY;I~hG?8DdZ#$b-swIz>$5jTLUf(SWRZ;6&9H_71aO=9xKy)?xg&R8GJ??;;~
zr!bBL^)xka8*j^G8)>e~w3FeP%{CdSE**AOSyL~`c;n`dxzZ5wP6FLZ&aZEx-gz8d
zJDl%vA8sn5JXQvn1MsRu7EFA`jH>dLiOC+{{lP=X+V|V>)KuP-jq-NIrAQStpz5M|
zKk!v}UUbi!#W%*@F?%ho+|{wbb_h+A?lSY2wb4_1M2L=L!b32NFT*f<c>9UO8Kk{6
z+8ELuL%XJ8B5yu{<89seXjon6CGd+JTNj$x4Awr(IIwIH-k^{htG|$M;+MBOQctpE
z`mS?9)nnsAOTd{?o2pv1H~^hG`q^gk-K!kp{*%QRuVeM!S~R*ApV(m`|7JnzdxuvS
zHc5YXFfLscnV!>dUj}i|)Bv9rcJ~!RXG9Ys-TgTG6=X3Z&(RWC^hhB#ei0n1GT(g>
zJF279*+})%LR;r`pbAndrh19dth%SRc)>xGx=vt>H3G98Fv-?Nl2_(v5O@zR4AR_^
zn;Ym$vp2cUmwt4@7ChGSbrfRTio1xW4Kqqpmx*q$olAdprHmheZtnK7b(sV{z_Shh
zDRBwqGn5NWlG2m<`mHPpi{YyN4t%9wCZ<sGVy1s2vN^q?cKcc<Mc|;!bLV4SRb4XN
zQ`+c_dvvRkH0?nvA;dvYN4=$XtjXftY-|b#Raq8~n8{{i=BkA$x^v<Lmhl_rP5&V)
zCOE`yX}I{rHsoW^RJ)dTG<1FRGUsHUF=2%(na79|L{=8Y@qX>LUt|EdU^iz~+F#>v
z>trrP)YR@c;G5y8NcjT$`myk+s-Ib)U{vP-vAHLAlhKtnH5H2)kwNU3qAHC62b@M%
z74VP3yUg3Uc1^e|25P)J`KC$MkM4%2$D3P}In2t3bvyHD)S}+jGMU(?eU5XbjZI4L
zjrHS>T#>YY|MFFWVs+%h<CHa7?P2T5s-GYz-?>ixyz@T7&&TU4YgwK)rDd==#0Nne
z*K4dA>}Z^ul4xIB*&8@I<MSz7(*bt-2B7wf<!AJ}$Ag-mz&%xK*$2GysrS#kq~r*_
zoKMDWo%08s40M}g`NN-S;^{^;elMm6zot=(3TizXefWu`;Fp5Pcq@V#<%(oB9nI{7
z+)<#fsV>}3+XyLAbSv1oZO+~2_yy6Re5}b*6#gnZrf0e>h1s|TIwEe^B%jx7NmlWz
z_~^;lprW_8m$zjV&cH)C0Kd=`BJ-n!&K>m?=3x@M6koJcY+C81r?6+~muZze?^|^Y
z(yO$QV-6PZEllR&Q{SFpEOblIYJnmQK|e)huFTmh2!O-1kyaTjVTcpSBO|bbrmT0I
z{(?Pgo0Qs<S*NE*;zS(apLV~zS>8)Uw@84a)GZ=7B);tOjLTUfhB&o97=B+Oqj$`&
zIt(({6yT{_TMyM9lZ^I@tv=l8$lyZKrD82HhS6#*Z!CZoN3M6S6X{r@?QvCPq{$<P
z&VA`y+X*DxS-flJ_u?==tZ*la#Z9QdyFgj+r^=@%k=)jTAs_=>IT1W}zMTC&shZ4S
zt?RQ|>j3f5D4m`0qp7K)<6X$@CXmk?qLkhg8CNt{CY39ipZs82pyv0P*YT3iTfoSt
z@qsgN)`CcUpmm2!pEHoj{}~bz24cqnt3R2N&|%jo>MIe1HzOWZfnn=@>Z&F-^zP8Y
zB$c?0-RE6xbC!yA!j0Q(w+?^75LeeESDk5E`(Y6NVefoX!*1^Ct~j1A2VAONjbh){
zK54k<NW9QPnJhOuc1SMLOvrJ3Ujmy?+znQ#ALmHu7tHi<BFm#^Eq_#G%<mhh5$J&N
zo`P;`N!JTjqvtGwHTxG_^MP$~bdDL4DBFVVK+X@3BwQ7?7toFAMuQQwUGY8p@QOa7
zPuUMZ>Kqon(U!O)d}S$lMs~i=HJSM{b6b}IrH*8DH9rJ&RU$=-!@uJLJOOW^ulg1i
z7<*N)kCq1ZL3@t#PXa#kL8H`F!G*yK)9Uku^y_hSNzr!-k?m11kKxq)$-8xQ-id+}
zz8@tIM3>dVEdzp*_@1wFGRS+R9e$(lF?Njx+7Ai8EWafy&1E+_{kr~$Q4y_k3@ciT
zv5twKkP+$RC;LF@fO1^v?VMW*Srkp+{rF>AgCkSBL0@{7;_<@u`Xj<t<#_`MpGUdy
z-eFtX;*k*P9;wI#*R}nY<9*-dz=swlTeIq6n++MVV9u#s*0rP$#l?e+5j^F9%<r&x
zEX?kV8pnjjyto0}N=G~Ln<GQ(S3)8{0mjFAh8ho{gM$c-?Z@|jBy&qx#Ywh6wDY8u
zZhPB~uvf43x#gntpM0~5S^JhWu_nULMkL>DaQ*q^G}ybxH+qIHNJZ4wKZ0NC_kcOR
zy(APS7#iksPG7PS?`XwPPJtq{+xtp(&A<H{!@iwCgpzkg#N3-KtCm+&)!bU;baK8g
z*CT%xbjy~NK#S%;4<;?`pBFdC_E4=)xCj0Nvb+=~bRPvN7`lcQ8kIshOGU)rJYMVs
zic<d)n`-u2I*gu&jI@_Pmfm4~gAL8!x2Wux1Y^GuJ~Zz5fxGkm1B7~b)nKRT>IEr=
zDn0I!LBjQFpB*Emj^>nNKM`?QI-dukU23t0fG|qU1^D~*Wa~j41hv5a=W-XagG5Wk
zHA-Fd^(RkBs{z5ZoodgKOGK1A7Mz$VxO+QNY1W&3U8g`rxAEOR5&LFyK)Si~)iZy=
z%oW0qS*_!5#2lG4?C)@2jjKFf4DxotVOJu&-;9o<B@>bnE!@q>grKS0hD9p7Ut2Rd
z>^Yu8+7w=(ES3d5RtO`VcdH~5W}k#7n4EI!PkNjWP+)@Gbyn2N?JO#O)ff6dt+${y
z*YL|SNlLB8<q)-aa8=aDt-~(q>-9?!Mls?KS`j9m8*t9qk!mp2Q@G%)`P{?!Bl;^y
zqUZ)dn%E0Gk#tx>(cA!5Epp%f-k`PRQF4nmuK9(j1xj`34{RwZEZ+H?*s{HtjA>c@
zEj(+o3~D|L&EPsbO;7~ocl?;2HOU|4ZXhFgF8iJR{O1k8%Ah~ZIkzD+%h3^V-)G%0
zhjgOyHUW^@Oas-snvBOzyWItu8i&OA((kq&)@*j03&wSwk5p{QSY=pgciwV*-~WD_
z{(${mc736rQx3i`bVx(p1(SBM>p68M-6F~UGB_&vy4mw+{_;w1eCc_FuCR;K(;Ce=
zTD&XPesJl*p9a!p)Z$YnjKSj1;|WAC*4+TIg7t3zEeNdSphi`|8%T$k8y4UpfZhNC
zu&_hVZ>;OdW)%}kZi2ZAy8*<&_h@jiwj|L|?==}y?e#h34S=~65@~FD<j;+i`0AnM
z{b=0wo`#5({{tug?-aqT{MGq=Ohe3^g`RJ!lR^~<s~w*fR1`6NOM|#283)3@7M*Ja
zY;Lgn*`!>_gSWAdXz(_b>Ad<zT^f*5JUZX-@_<i!<eErNEcT+U%w*-fGacKaz(<}R
zHP|w_@>jJQSkT|s$Zh~W6gX%~H>0J)dN~iQ05(_w-s@l_uH=oTdh`km@Ofu;%36hw
zw*wzOr4a!+==Jm%7~2;N4-aIW0y~yw#?=p7KO6{v=D^kDFZ6>4niC5nw25NH=J0Uq
zK(8s^SUf=Mpu7VP2f9;a<TtFs1CM4m8K}E-`j^bRyzWl4{-s)_t1bl|R8+v5H$DLY
zWecftjfOTJ{4P#pLcX@RM`Z#Ol&m2znd!)4-Q=sSjYh8%S-0sy3cioLh^<Hcvd%CV
z<X9o(I&zO&3)r2a!hTP~Is3(&djx4<Z1qXCn*||;ou^ahmf;p&iMgb06G)m|45yKB
z1VH1*TLy2H7<x$6z3QMJI<$49e@!`7(Ii*?h0oKr!5EEI1fOR|^7qSQKmV5z@gIES
z#vnuU6k-spuFq>YT$}}LGwyZX73~=ml|91uX{lU=k#Dh6m5Q5WM(U}k#$41lB)%HY
zi5;`cuU2#wFjrf8cwUcR-hcNKPN=`HZKQb0RFhbVeSAYV_g*f%e{WPX>$y&?#AeE@
zWIa#ZiDUiiaRcM1XDyMl^A43ht}aKjj9)S<PW3h4DZDwK3!Vh!FK4!tJ*#i&n?@OQ
zeHwB-W8TwetX65NuYQ@ds<Hku1Hs8n&r)IX6a1Vqd5!ErZwz<`;q?u4r3zNomN}1c
zKH2kl2Crt`k=0>#lss%2)wEv!{A7Rf?W^N;98QutC(HM7?+ydwJ+gk=JA+5l^D|af
zbhxALiVhy_kv-;NMt&t*{a$>T@_tuBht<wpAlW2j!qZQ7fEgVBS*g!`L_DP3G?`%G
z;RixNtRAR4=Ae^z1K_!w9+ek(=izTX^#)1UB`>RU7*f5FvOM)+WXW`=&*fZ9!`1nb
z0C%*2+db_b4JM^FHObuQIfJmV8-P<`H7j9v#q1UvJ5mUSxHth122vtl`sQc4OOOs%
z^ptqLiOwx-Pt5_h^$e$z$8WRULs5SRrw#PfoJ$#hwCy$y$uIH{&|%v#Nw?dB6Mhu8
z4t(KAz4u8bnecAWoMAj>NXB~NauQy<azG7E)SHY=Jd4^>+#T)MnsNrY6n-*MXZo<O
z{6^YBZGfJ9N)}^+VOfG9oE4Ae7BE;A*oW6UOI7<t2q4)`qGKT4*h#aW+SJMj41Mjc
zv+JlRSZt;al0a20&~UR-dnk5A7lUO>B9XN{Ly#9pgN_>j^)?-R*8BbPh(so}!>8YY
z!U-y0XPR(*<x5S3?-8}<#4PArNPmoAj3zi2)LfpYc2KxCags@c3hOkksV!12*7D06
zB6yo1`(#f^z(bi!hcQ;<X@6X(Z0Hz0Ba`s%!16VZ&e(3g<F(%6Sf~E@KHb?9WT8+V
zm;0i|BNg4|GzGcl$gD5;JtGtmrxuw!!+VZ_X-4#55B|NW%ac)IlNMA%!5$S1C8Yr`
zK!$w~JPA-@C$m5mXjWCgW5yOb?ZUDCyp2WRJ-3bSPuVX!@37V^Z=f>IJ|p8o4GfMI
zSnbDaA`szdlVWEi2Ptj3hNoQN@QfH2#;n{qm(Rt<LIhtwFatOP(`Xn2HL|*PvvDKE
zwwY*$Lzf#Td1}P$;nan9x^ZIh^m?7F`b6UUV%)o|SX*zSDMp&GIx{hFw(OJvyIOJ_
z4hGUR99cy>RqIr}S?y=h#3b_Y@OPk_SYqMhSbmaAEoKg)zP<7dN-KYPec0S$dUoN<
z8_44Sa?;;umE5v3?I7&#;q__RH`~(LfOI{nY)9FYZ`6a=6K|!HpmXjI;?>k<nA1+u
z)=%QXd|jeeH8ZK(<;R!i?dj6^)j5m8m0h{UH~}=QYrig}du}aLcuc%VY8gH)%M6L0
zyu1i}irq9YSoQMd=YM3v{c%sX<*E|W07(K#3|Iurv@QVq8`1#$JY9>lV`%H#f&A?=
zdmnoqS(6XYna{B=uOi_v(dz3QIp#vgDJHGY<_{)>1W$>@>s<u=nM+NLpk3}?axJR4
zi^52@`z-=TgA?DUS<p8;P<=|7Y8FVC4o*K)N#CR72|KMD6bLkwpt%b}lj96uAMx{R
zUB}C*a<|c6!c(0IeEeKRZLmA!(}H7I_G{!>gjBebC#QTDQ}xlJhO?uGwVWWIaUY9E
zoluZ7xqzR^qj&9}DvpiQHtGw&w(8B3)qXnxVwx2pWkRmP+=f}95{$OqZ*LXia;3G8
z6d)zCRx~-=`@Po!Q*1d7mzI0OFify-w#?=ZPL1Z6O?>1$@9&(g;UbB}DdOeKm<(zp
zrhHaA)vajOAw#hta=>&{xGJ^m4IV)<USlYy=mTrA(S6cpDXM0Vg(;=Dd)1}B6Gbm_
z*ljh8?q(M`Fu}j&YFW}=Uqe6Fjz^iO)TLTFMnqPIui~K(s#$BIzw)#7l#Mw3ZJ-QW
zM3Bi*%j}f%JArp6?v0C2OO|Nf*YJKz%c=6*Sc5hERSb@;4axTeeEvDNgCCm^>Q<Xp
z3V)@B6y0jCk!l}`Eg``XaraDvI8L7V+W{Wt(x59QRJF|;*HXKR&Y8RfD@zp7@G&e0
zO^)>F1X2mZ1UyquNdvUUB+{IvjK_q<OWLXL(CNB*)98t-2J?P;px56ijYrYnj8^cD
z?PymLYuEyLv~{h0VR<>fNvcpYa?#ih`7xCAZP1&IXY$wgVt1#~mpw6<lduiWX+Xfv
z?*J*Ss?pNBKd_8~v0qq1wMgMBRn{O{he$;CD)Q0BH=o4?nSfp|YMjuvlIJ0?DcB;7
zt`QqA(mZLonaja<Xli;lW~ljl%7FLv?lIRxAsXsqjeA0F@4tQpX!%?_`4Tk$8%m`6
zm;u9g7@N52%w{HRToHTrOWUgz-=vbC0y}vxhPy3(e7ynK`N&DbRNh!7)lQ6j$XI-f
zn|ZZn;0C&Su4wC6wmp1!&_mcE!<Oa!rVyNE@=Wp_snINR1R7Rzo!b>1+v~&~S;V39
z^z2KXMveyOYh)c+X~Es#k;qdKpJe`9FeNXY0-gF@pD**3q|r$~Zl5Qo%xpC@?E1&$
z2lfPVCohvLn~p(W4+}eP3@)7y@KM5>R}l~_ae%^K&_bbKvU$?Tt0fEc0kwbIQ}JZ&
z{C+b<-4ZTZs|xj=d(PfhJ6B*&LQwr_T~podQa2r2x3x!ipPEK6NGlbCl^fsiQB0Sv
zGRK(r=bkJMXvtwtEJOOr!BTtnon~dHQwEn}71=-h%~o02+kBZFF87wEJ=Zf9r-#<%
zGMmz@Lb*F`1=@3|{t^*O3?haWBXz8Jvm*ijEUbgBjXYUN^(Gd@_>4?6+((PrUq_k%
zu2$-0cVa++ckO;?=G&2$)TH>s55H;uZhXvLMJ?jyTn-ol@U{Jz&wE0$e8MJOnf@G0
zj}vsxu`lBp-@g|r!6-CD4Ir$)-xO?gh?c4C(j*Uvi_)7EfMwSiuEAG7X^F$jGW6$t
z+*TO9@6noGKpxCCg7UvFwe0?(v+S|4P3FR@mZ~&w%{@yTcIMLHh*$3B@0v4w|LN%+
zV%!gpXqQLOKFHAKRrWiEb}1PYOc_x{!&aFTr*wV+-p^|zwKDBB7CoPI6_dpWyZlsq
zY|Y3W;MX>*n5d<tv?c7kpWEC}XK0mdXW(dRT0!^mleAGEM+(mx&u6WcH%Qs#spvpn
zu&v=?a;q!C;<nL(=`+s*y2a!$y|K3mN~2Z!11>*pgKihy8)93D?6v>pX2G;~1Hgy0
zFX^biUFvk*p2hO{<}8wa4H%;~5fB20v_Y90K=|zorY0FVI(=i+`4p-J4XavBC;uND
zb@AgKMKO;rlI7^xB?UM4ShFA<>R9%ovt9=~u7j^aWs!wLTC}@yp9B1bao4Mljby6E
z_OHfN<{2L^f1vPRx>db_%J5eaF17lBRrLF(x*jDM9M3Kal5Oqg5}jRyzG!5etNj{$
zG{pf8>`hS+pY()J9qhfWE}xLoDw>_2Hhn52DCAmAR90%M%1b-TE>Ow8t{l?l_A}3!
z87ox{?7F_up=PSyE3WbA4`P(DJR6n|xn)LaRD-08vo1D@3aM%eu?;sf(R3DqGcso<
zy&74*hCf$#Gbo{M=j_r%$P4-#U4Qj|QqH&3lJLvf>hn8u`E3S*p1oYw`W_SJ*ac}L
zjl3VxE@O4Ns<hnyS>6$iYs>D^e&)KFmiyF_m+QICQCWTv$;I56&!IxK*Wx|+3Mynp
zQ}CgjGwjFVJl*A+@^=k+ZG2|Cv#Vrk+V=^$mfyd(P#M)0PFlKL0FF2>vSip__rWGm
z-d!1d$P#lihYZ$+rimsqYeyuVyx^Lkhd`iyi9<Y>LwnzJ<gL5<SH)Q-A<}6I^~I=8
zm5H2@Q$gXL461SN(bHxf&BODUY8~czy4s1o{G_Ix^p_8T@hmZCs_y%5neX*EUGJzq
z>g#U50SLC7%uO{k&h=l-PU@MDT?@@!NcjmZz-|C|7jr9Nd$h~Kv|x3!l(vS?fpSQK
zq-=}rhYsfC$%|)Z)VD=1VjgLzR>dV0x0w8#_E&3XXP2JCTJ|HJt>F!lsmLt<SBNV=
zSJEY5IM$ljrq0h8-%(H<%Ntra-A#|?P{vDo!NTxfu<=B{eu)jt*1N(J))QAa;D%q`
zT-mbZ&=1=%G3|C^7<j;}+3_9umHGoEhg!c|ik9$Z36@s5x89N+_4{h+2+P=OW<-WX
z2+2tlgO}Q;(kx;?W_}2pL?rGr8Ay1UoRJU6nYOC|G+h@WJnL=~{j?^CF21ABP2h3R
z5UZCi+B^eD2bS%iTqO-fnnGwnPe803AVSB&6S&z^6T{8NGlopS$|?tx9<5|tN*wHW
zVP!Q8OBTJvk})|@7Y_5yi!U2YZU9&{Zw+&L9terM?btxs*%~}`xfxl2eJEVJ0$H3j
zvHF`ClK<u58ufOgV3+-M7Rl4mc&g9Px`$k0e({`HoV)Bn4K=l8{+x#k=i#o-r-J%+
zj2TB$za9>=f9c>M7Z+*V5DD{%z`b5)He@>Z4K;-YNI^=xj;nW@5>#{+Gl8eqpny3i
z?2;Qqa%2&gNKSEu*o=fF>mN>qfzF!YT(MH0OqAdTBHRPTAN=dqk|k|8iR$CSiJqJm
z^G>OSjTTHwoR>f>B9O2i%D8fHls?iACDjv1fpq>DNQA`i4y{J`5wkxZS&M~BQDIrz
zy&x`=g`Q)kTo*jIAA~q@^j8{=iT-kNMv-$XEqTjxLzEpFQ2Xbo#4eey$By0=0om$T
zG;Zc|7iBOe0;85mCW09KMc#AbKumW25Wniq=-JY#_l1LT+SvA)>t>(@MirL1bo9}8
zDN49S#J-h6yBZo0*G=zbF0G~Ll27iK-%WB|(Z@}mr0h*wH%$_gGWgwDk3aK}FxGVs
z^Ju|v8r6(;g_+Lp;P#KBO~6x<+eHHL>P29$4Ip37qtrXJ8FXsiN7@UOBCsM)SlHRD
zcU*NG{cw@Rwr|pPmHu;8@Y^HL9Z&9PeFUc-ZvEo!`b=cuLN%@WD+7K_#Ml;8#=Oty
zLY4pAI{?%pWi0RP`mNU4r)DXyk(~i$utFHN7f8K53y(vx_n32_ve-Wg9muJEpWP8N
z8?yK1RDPHSH{6WVOnabNtQM~t-$aym)Mf!3^*250Tcsp3^4A*2@YsnAEbnOud!}FQ
z>^S<ROpQrX*)+8@dpQ@eDLpL|0eV5}GDo)gR)vfs^QyGZwsKq+pPtobIX=m2O<w|`
zs?q`**{Q&UEA-JWEuzM=+iV=8&?bjP-ny!ZD#tkHZI8uo;-*h%3B!_ypa`AH@Kl$i
zpt}Mw5AP@()Y@fN#(vVp)L3}U`J6YppQfikP{2(Tcf}vzvj`T4z}<)oYXwN?Fn}RF
zbmsSwkZAEGaxBeqify%nA3vu0+d^(y^e)b)*#KC7D$!egw|dh(jwiHzXj7DQ@eX|}
zBj{SV(*b0QWJ|0#TUATA)NE!HI60_kCwq!)9u!`e8Nkzf@I`GDJ@iolYP;$h$$mQr
z-lV)7kMP5uoM;GH%(Ttf(KooWup{Uw*2y1GCi_uSBH_uJ%6-PeEGwt|qp70A)z%OJ
ziGcvhmucrZaH99ZmP`TN-TCQq#icdp1|vOZ<xjo8lN^2U?om_D->5O$Z4MkT_njSb
zDI2dIvX&Z_Rwy$#nzGQ8&onVJ?;U8M3^HDv9SiX2lwvw>|9x7w)^b=De+WNUAD?XL
zlPN9A))Y7wQ$GuRpLvcW*H#hY47<0E@<3!pr*tHCE)NqsVwmFh&|vWro0>$e<XPDm
zwb4CJgMe&;1a4P;ZnxUIF}ZQYi0fPo`vN%9jAH3!OWg!#Lu2_hJj#DWexdO5utR0I
zT+KHXQIQbCMqaYw_Rn8|p*MhP+Yo=}Wnmo0*^ruoYzK3Z>V%a6H8Z};ZHt8_`-k|>
z?GLlkM6LOXs#1PU+7JVMs^NnExlo9F%ihn(cjJ{8TSEn)#@4i!ut>BT%JFcu8$!^e
z>9`F1aM_b;i{BIC?Pm-8DCG2}{s=_0=*If-N6EY?ck8znyJgFJj}*DSOe+kHrxA?r
zX_cyxTHg7RZTF0u8ke*@=mjqO13>H-+I3OPxH-Y3c!0T<lv%?#SFlFoq%%m%qSH6h
za2Sv3z%QI}HNw^M`Kz<JpPE_yUvE#wAv4di#uV7LYz9=mZnxamQr4)%VtKm_{y@iu
z+L#2B&ldL?qs8bAb3X%z7-y9M6R?Wysl}74*5!SEI~W#|DLA?Iwezc{-Mpmijl&10
zckdveK)uEk-2@FpZJHL}K0kjw2bUIW^J<sirswR8VeVT=AUnV12%j&{yxHHGRsX&_
z)1!8(+`h(u-dmK&8M<n39pAuL*vIo5r%dSByfVF)NxN78ds0AXO1r6=%Z}bpOrWMz
z)XRnNoColLaHG$Z*BKVRbfn|2a|3wY2)m}4+|0OJkA(L2$W!Z9wTz9y1vy_uGDnDO
zti>*H)s<UE>Int;FsPA~^FE}><QVRQK6=n7_12n)Ek)i@b5gK6imS283v84c2%eQB
zD}9xe6vU_e0Pi-@-KkV?)LG=19MBuuFDJXgs@ds_x93`l%Kk+JtUNh?Ox9RCSoxH5
z_}7T)T_uJ)B;~9+@3l^cqEo0$O?uzXe!CJUcNuzc19;Z35Vi>Ddq{Qk70>P0;fp?h
z`W2|fG9(PGyga^qa0eXK6>Ktb;Q7M2;MW;bvDv}^h{|s*27#qnT)5csDaM5B<9kG+
zTA^N;((0WG>+AH3>;7SjW??~{tep7!F<de5C-VgQFFq;$Ry-*%z=HCOP0VgZsGJq)
zNjc^c$3WmjcI^!y1X4C_pJ=9@7w^ba-0fSa*OoV(rop5ZHK<Rx!OjaqGB2m2qD)a)
zum!N*&T}LO)=LPilrVJWriK>AX~rIDs#~Pq{$6LqCMsIk{);*?u9>TG4rw2-eR&DO
zYmmBzv~L|NWFReK1?1R^z;FAHm?Nx_1O?Yg4kLU?8iB*&SGfaZLenf(R%<YZKBn`j
zabqp3czdk($W1DV>y!8!=M+n+)cfap27L5Hj1@aat8VPf;A4bMCSt*6M?D3BDUj;a
z6po-vSydhJ<d1-cq=WrS$JAY2J~j~`!sB(HR*Dso83vDlE!r+aTB;Exag^W}dLvE5
zSl7pG2PVmcOQ+<_*FP{_84s)_^OTKM&efEtCZ#{ImNel?nbw{_d_yxY#7-+OUat0^
zVDvGLndp?f%hSuTN`I{tdD(Uh>!ie<`RZ<(eu|^u-a=0gM2KOev1SrKmR6$;@9R7M
z)f+%4kQhW67I+605Rd~V)zz{&NAy4mTAf`t?a!soBa>}rXE)Z~tfVNos|_lvN#V7b
zQV95qG$V14&9Pz2j$m=E7l@5oyoSJ}!vo#JU1|qKXxb<by-Ua0y3QBN^tWHb?j~qY
zN%YSfsPde91X89zB-Z_&n~pY)R<#&dY@&SUPDo}~WCy?nA+_Cnglg@KdZLsHlidCq
zE0nmiX7=XyBiX|mCKmac#|3gw&V?pmmkrucTBmH4819ESc~#z;;t9mhWRt=c*?G@h
zKQ7{!0aw+#Vbt5c_wr$&!kM{QM`~kQduVH=z{hODH<;1^Mlzb^s@E)^8AiK3s%i|-
z@Qj8XEMyq${=wC&b>Pzc8>KZCi5KnoC#T9tr2f?!sCWEV|LcG0AOjnx<8o%C1tkX@
zEx!ywsluvQpJEQ4S4^G`?1a0Vcr~PMBq<EAHs;OtawW*v*oo#@IyoORe?ao=ro)!N
zm0eFORaXS*yHeg+SU51HM1!E>-FG7y@{h$!=wjgq-ib}DhXL2Qw}WaYJzqqv+e6PT
z><gBw7fzjCm~Itd^;<uz^TAjsf{oJgkh4qj5T%QMmQnB&6;gyMTP_{*hkNrppJ76O
z?Z;p9`-}iXzAW2{At1ZQD-a5-6N;j$*>_b#6GDn|G&r0;+N#wXJSTn!|G+pQ$4iq`
z>;6+?_G!>BQT0*Jm+mFvXApFp@cKo%pOMEg))BPahjF(ZklAb#j%ay=g5;xsCOX7h
zH7%@%@T{DzhS<zvGp+kMZju>iB7Ndo4w*J_R?)j9T$^4aVzu`-qg!sFRhq^nhDm=r
zBO?5zpy1$VNSe#!tN;(%^-4vF^dXmW?nbWnEqzI!f7UHk6)+EvztnF!a`*6SMxSn=
z$9%o^Jn(55dPr62LF?HT3AIC+3Si3R%xHfxq`SN`=a|1(7U|sF021qkqjC^6-cfL`
zdp;}PcV;3b)F(DX6T6G7<$8*`q=@9Bi`U)DGBa)b8yD%{Kh?sZ*Ol0V{29gNLQ)~}
z^R@2>*~sMsInwXFt1<WCKuSB!{DA^<P4WBv)3W^)v*FVZ%+~OI6#0*01{xEpm-N|+
ze9u~V$M1+IATr?obKQJJoWE*b+=r(s5-}wB*WR@v&}-)gOji7;8F#j<sj7{H&p!MZ
zN<$PW4`3k~Na<x`3Dzv1SVIz|u%$!ZC3~gB_%VcTcn3t#xHeMkwiQ0t;Cf9q2?f5a
zZ;GdLW;<&)l1ixq6TPSrtJ;l^b>D5eB}RJuAq4Xf-bC%UDohP-M3ATa#85O$$r%(_
z>wfg6w@cXiNM`FW;O7@Bmi~^<C&4Qq{X)><*OcREQvd9DgGJ!bg?Zm)roA$BYf7dW
zo<x3K+K`v4JM8v0@j(W|ngFNwV{g76F>PG0cbz;Xyjq?ewsvdA=pc0BFyeN1+Ff}f
zt7D|J%O}d-VqJBS64i-=tDNqg;wcq~a$wWLv{7Vl7ds2;4m&eSf1u~wcwsSBD%5<c
ztF}f`ZGU>3DP?A}HtAmIBZlyC-xZN|o3+I15_P)VY<Stg#$G``*3s!z8s6e;a0+b&
zZqH_F&g{@Jg)E#xp?l3klhIM+^CJCaXZt_PO6~=Jej8hg)6Vh!x2hG^C_c<DS}AEI
zn*nF4H6gjb+tiKjof2gn-IiQy;;1cTPLXEsl2`R_&ILDivyoqC%e_L>OE`WL?wDHS
z9d`y%GMlM8a(6Fmp2^20+C0A-N&W7pver8OK&e}Z4{c)fp&1-@Skgt0;jEpgrli}J
z@wA|VLH9sJ&=mg>2fAL286rH9^RCS=AFjK)A;czXXX~!}kr0@Py=~>QnHSE5&^gTm
zy&C|(*F}C@vmGqCNpTex&PM)m)?Pt2v$xh@kt(yh?_qfbIj-@L!{W<(yzyGigYE-N
z7wK!v?iE&dc){3HHArAw%RQ6d4S@x<T1fLoiY!FQ7w8*TRJ=)bhxzNIi=^3~QeWKL
zK@2tVAsmIZ+*kxBw;rMaeA0I!qNmL^UVz5C0laKB-Vd<<&p00bL+8UkLmU4)?st_&
z71f|BwW5Z4c9`>zng(6El2>Kr+P}K&tk+0c`uo-N!3=%#^9V9mVnj?AdYxI4D+`Tm
ziCAYHtrV;JrSAMH271lc;f#CD0HpXn%iRCwT(aO$K5g7xY}U;}Pv1upu@NiUBkqms
z!I#Psq)l+v`$BgM!!-u7zaz`4*MFu#J)}}wr$Ksudg&eDNOFkOcI+jvOYCSf)}=`i
z%UJ5dD#xenY$%64kVcs}jK%I;)B@$+(>X-2DEm$3fl`QJn^A^iVBYkzN7fow+w$1u
zHc$rh8bfyjC_6tHKd3Ff3?1&2QOL}yiE_-Hp0gWFP@DCa{F%INny_t&>%vfZ=i#r_
zjOvtZwD6&8Xp`*q`&_W5(pZ&nmEh8uS#mm=WQ%)4!@xRU$RaCMu;tz2jPC~YE0l4;
zN)}A<*`)D^*4y?>bgCPp9%lzX0<ZR2(-r2<1DzKwbQ|M<C~ssz4;y2}lyLPh<qg0d
zB?@HSmf^>`j8^4O%w3?|RcvNi5$``pf^#;ktpR({#nBT6V){F-*ELI*dX@=QH0Zta
zd@P~hH>+~HIol-kRMT&d@L5yvtN8m9_v$KR;72M%g-0Li;y<f1;8#3W?N!vNJPZvk
zuu%80&$frx##eXdbxFqOMov+1)o}#<6m2L0o`>VDMINRP&lYqp64%sJ)(7w?w@g%D
zdyPGivslTox1sAnW+y9yM5WDk`_|jSl|uRIUP1#n&%P%1UP^tW>^z{HMrE2H0TWh7
zE)sdQ<+ZUry7+^c4zUlSBdx)3(I1}>ova@cChxaGf&78S9}}b80+tXU3?yqigBInA
z5Wa==JW!IOOjHjKC#;TLAr)09=avaHddx|;>Bz_FRKii~&^7GDYizbtY23*~v&NC1
zpr$HT%-&_rTbT}UG_w1wc$qb-_g2$DJC<qZR)!g2$`3BFr@NF2<2+1GFD6%ToL+CN
zYE1L{{BF@y)i9AtIU@CO1f#pv>y`oSA?)#VD6_@p5%p-@W#9wgq(4>-%Kc{qv_J0u
zyCk$*MccE@$sq+p0flC&wzfnIT;r0H#<k;#7Bv3vP<bFEbN8J&lV`okm7jX;$*lH1
z2M0em6`UIh&~_`*ePJNi)XB<1E@T7asgU>XTB@#zDVUd&xu57(g8%4B<tQLSY(Iyr
zD?kaZpMLea#1y~vu<r@+`d-<&2o^T{%!uv=`$NdlCie(+NFB9HMHxLW!hxtg8wU*A
zq(AW9{i1wx`!aO#;$tM;B>}EP$Ih#HRNux-r2pQGul)tPCyIW>HVkubwyHsX<06MK
zdF)MvT}#Q<J~7+rq`_wF`%;Gt+6h4ksHBFoO{mVoJg=-K#nX(g*GKZ?0p4y%W?iII
zhaZs6j^dMP+W_}o*+?gmn}Gij!8?yBS9rs3_*t6#aAE&3u&h8kyJ~7B?nQQyT|=(S
zMxt$@N#^{5N(HQ(^U-BwqZv=j-#mZjCRa-SFCn4k6c*(TF^;-6AI7~E3~@J8SkKlh
z7eYMW93FSSieYjzVnIc#h*aen=Zs=E4K(Pmwo`*?_Hz9IBqDZ3D#E6i5fX7%c*P?&
zXew?Ya5m9r=nO*qJ*+m}Uz0tMbEPF<qzMQGADh$pIyYUG6}cO$F;@)wR*F4rJLfvj
zL)r#|GI6)Vw5J5P#t$UJsOrfWwcqvkKY8=Xhwuk%Tx5pAOl_h>!DserZeL^XJTrAg
zY`3R+d(`w(bCto>oXK@BoQ?Dv+VpUR`O98)lo481>?^UAn>gT-9zb=<ANRt1m27mr
zBcb!&08Jz#ml8?2+y|m<{^L6>;nztmLZ({{6Xp~Xo1*qpek~niEMKg5%Gvj8<sg0x
z+m8K1b-v%2v|Fh2p{@n-+v{@5deB$O^iwKlQ-hbYk-V@V1@zx`h{S^XK@WDAhT_6b
zW=_eUVzB~VwJo0155;lEvrWw17B!L)1*UQ`ba8H_f1Y(y(P)2Ey$6!N`-RDzwtCpE
zmEGPwwF~RJ%Yy-$UMCEv(i?l2JE(@8j+hn0D%r5nS7iCiw7=RXJ8a#|uY!7{VIzmq
z6ii9HlY?v(Gv)*1jkcR#cCV=QON*>`p6u^=@62${pA64w87~XvAoM;_4<^^9r$Uwm
zg4hQQYSBPc5n?#P%ULMOuf?us%V~p|Cu+O7w~otiYBRtd<<Vsv_VKjD%s{i=Pjk+e
zBH;nLowfb9;x_+i&%4DtrR<YaT+JpHvP*|@yX<$~ZfOO)Iqn_hIjU}@J}HY_Z(&4j
zQ-&{+j#vMDWtm_-C&%Zp{<KNvTI}c9y@oV&OP;iDdVz#ZmF+gowptbK6WM}d65i}e
zaqLIXKMD{3=y?7C;TwQ$^Zlbvga50&^9*aMPuD$)ihzhT=|m}_bg5FKC|y8$2}J~H
zp$JHaL`8~J0f83~kS;ZJB!nPMs(^Gt2>~fm6Ql$Z@T_;vHM3{tynANPK6~%;<$T}@
zS!;cP)t={m?)&%umQT3Q;)0{ISqzQ=y50`aW()n*rX-vkN>s_KXJ9!=-k5Ha`Q{yj
zQC@r_JzuYq5lPk=c9U?q@MBi6?Z{@m@D|tz_3io5f{y%lc?RXV;^(5*j@<K9qZcJv
z75(i>TXHrh3x6Dkf3FS5@8W29DcdObhO<452}64o9i04}jpgMHq1r3`oVv@)##J6#
z2MV{`q0a$}+K~mMKBa+YeaMqY2h%;-X<EpkKfSQdfq3oit8BSyL4+@{Ov+T4jx0UQ
z7=-k525(H?klppviE)t=Ofq+W#(vMVFQT&1g2(_ZbvflEfn$|SiUPv!L0&fyq~+cY
zebh{2lcN^LTu-`cCKTzJdX{ape~OuAZ~Yb;JGqZ-HBff=e&`PmZfYmOaoh<JWG9zJ
z9bKqp(mFRIH%)-JSn3bY>)~^YsV)NQ>4m1Lw!djodMbuJwnyeg`@<(II);wFL3RNJ
z*YHn~6hOuON1tco{%jontI!eEb(xCE*nLsQ>3!h}#fP=Yn=svU<0RADWl~aCG})d}
zu>Hv1|G63o17QLL50XZtjlKL`xGJP!j?|z}F03jGO9XaAfeeD?tB;otykOm*L#0sK
z&U*18q0iV(C%=T9TkaAUjD^q;vy`RZ-pX=qa(`+WIVMt><+11!xz@9ky}*=E=ws?-
z)wmsU;BFWvr!^jr>bL2YffMbwkw1sxES0kx^VVb=Cw_RH%h-y2SaCA>j_*>ux=Ga1
zMAHt@b?7_*s1XHR#b?q07*{1|5wUf>kQk;rdL6%C*kOgic7!;fV}&T0JWM`6(KEla
zdxwWCKOQMOHp=?j<7HKyEHE#2%f1NX!a@IfbJIPqu0C38;*Q<jsC?>S8t_w#_!68x
zyYo**cevk-zW#hSZ{rU(a#<L$ivUbUio!l1rV6UY+`hk?)yVpm)%MH)d7#A5-D<f9
zdL`=)NUo>X3EaaH{U5hUcRb`M3i-e>wJ=JV*S0dyZu{<z!8@^7%p?D}*2@r>lRw_S
zOUo1gi$d=eL|4`Eb<ZnJk?)C`^PP6@>|>Ykm7}MO8}MVQbu3W~@L|3w1=H;L52McH
zTlO?Rl{gxte`J-`cI-ikudHkys+VZRi2(xJ2N=+TzxvTi%mDx0cfe8nA2~*{TuJS6
zu6LDaQ7)|nzf9Nr#A3>3hL^Siuqlprjkb)NTkM&=^`lEuEsau5_C#_%9#K=aEIuRL
z?xR-WjOLG*g}UUDE)7s!T-PZIosxoMS@y{CH@hVg<czo0LxvdCz5AE@zm*;HSm4bT
zUy*MT!dz>nIX5h}FcZ$h;m-<F8gD7L-D}(U>IRG&!hH!#Uoh6_xB^{<IF){=$COKf
zqg%DxGpfeN)Pf_CZCTLIr>qm;1|<vb<^8YClrB!^k*Ir-D)vbo{5)Z7+@ZHOkxtzS
zuaV#wak7?Efvd#&;YUW7)N%O213KdQPqP|Up*{xzy}?|NPQwmuAA_1SU=-Nldjw()
zOK4~pQMP($PbK_mA9yzF=Op6b%xDHEh{n?`cg#ZUar0{hYgPI*)ETr&a0xi7ZRx}J
zJ!@@iPV?&4m(jXVC7*opM}2LP>%u>P34Tr;RflKPJWKO8&xP?62hQC0me-x1f_%G&
zX7<m;@BiHC5U18zvIRLGKM)$}7VWDW$lq=+7MsW31jh^4Qbt!8(cZGp?OO_-vyj1(
z2Os)fjZ-qVf0-khY+Bo_-*^Pd<wz~cu+rh{@VVoNxr({)F2S3bi)Ij*VIr}oFOM5f
zX@Syu5$YFBb~t$lxB~xXv-9752M(k7&m034Uo`!6&`LwrD6xF=;hoDU8@Xq|n~z24
z@@P++eC3y#5Q|*#H)QcLm&-Hf_sbhdtw;+^D!iaf3$?n8-L1`x)_FS{$&_1PIO~9P
zUdH~OdJ4%zt(Yttb5DUkp$sw=ry}2hnC_vjwu*21JUII}am){|RW~_WA6YdPHr0D>
zh4Y1x&x8ARXsg)7A$CNr^v)aG3vMC6L*wH64AT;yC*QS1cHCaM+@L)1SU95l5l^vH
z%yPW#LStgD4;Voz_I|AXqM$4}0-oS?g+@s=#f9}7%RnAAcg5xmk#i9o1>+o-DZBXE
zZ{j0tu!b_T_slZWn$*z4g`mXM8b0;S_(E<ynyNOc;mIka)0P2#M}2f_)6K135u*5<
zq3e=g6btin%2h-TOg12<-A0^nZ-OkA=C|XY+S2&xwL02~yk~WrG}rxcC}`+R+?R<?
zy*n9V67b8JfneK#QP8V4gN}A{*FGcLtIbCU{hkBQsYb(8YfzW$`8N^N3t47WE-d4%
znv=VTf{k{#gFf}wJmTd;(@$~vp_U8M_ada2AD1>RcKxEbWY%o9YB83t)Ls@bNjB6v
z6<li)ak_*~%03{^K=Huq9(8}_AHDQ{-}k2%|G!mT|J*V3*T?3E9s*WB;I%BY4o&^D
z8&R6V>^0esNF^9k$xon<G`LW1c5-9jZQd%^1;YUP_@;%EAW}UG)2obHd9OQ`0Z5M&
zliZ`PB`*FM%HLDd4r(pt)Sux8!?|G}#w|}+#-)|ztYS4vTn>=0q2I(lSllJEaTFH^
z2zDTqKWBH`8x<*{;?)Vhvt&mMuLPGiuMH)oUGC4ED=E(1&1tutCYcf{zO!PwP#Bw6
z1&Y!i3K}U%AL*}wc6j8MWcF2R7ee{0M(R#vgS}I)N~+^$+PfZ|W1}(4wem{P*T&`{
zOGb?+B1aaUF;__}M^tQg{`kCL_)X0<==T}ypZYJr@`sPlTFSXR4>Soj?uC5|DY(Q)
zq4i)1rC${LWv<bm^o+k3Q#nSxw>K1HG<!&7TJ&$DXCO<VQYuBWe^F>s9R_6~nb^4-
z6Bbmu6I7u?!h3r!KEIL_&O#EsaKh6W&;;7G+JV`0fl-_z_>CnACC8#IVbE#RyxrPE
ze5g<J#b~8ttY-5#WN4_BE<oR$5RuoQAeN<)?~U!C`F?ZXaV>xLt?(B8OYpgCM6;q^
z{~*E~F=FfNS4r=xVp~!AqPzTu*MiK7&0QKYOg=T%)JF`S>4~0~hz3*KHih;M{9K<R
z2H(_3<qmE&n&_A0xI;~{K)dX1wwwiO_B53be|<$9kq%y+WGxd?2D16^BiANq0A!H`
zBeT`gi*@}ogh?fk{p9yKBkh)No&~jIqBm-qV{y1ovH6HGjuWX;pbRYBAY80hmK`;C
z6C^N8dlM8!X=|P4#bU)Rqh1%d*s$F?aKzOe>0&RWF!r=pu+`~QV<j$FybU|uZ%!EB
zZ#k{^lugDhU!N{COqxyD&da_Fo>%hu0^Mkl`;6xJ=5g*A9$8mL-PV(Y>AN`v6Uu9Z
z*0Ogol~7%hmWQm*xAG-Wwd0FrlpY=>tu-9IHJ_UKk!pJ2_qMR50tIC@P?z{vD)Bo|
zDLYyim+3y2<DI141N`~-|DW7teSmW`v^-$6o1+C?NDy3f&ze;c8LL}+{#CfmKr2t=
zimeW_fec}U{}I~qroWv0xdsb;{nD}|A~04angyv8+a2s|(Kl=;#3NP$RZNHKt^2)}
zm%Ijtq&&8zVi%m4mkKILu(DHX!`gzyMkJ|1e4;yP7B*~p42qN9Xm0U0NN8(lm={Si
zRY)ngpf9D5ZU}$ygn8CLZlL<*Ph`!`EO{a=7VKQ1KG{(_g$Xb-z(>Z53j!LrG=PDu
zsEhjH@6S>S3_0?wv=LnHVwO_8ntm@-sZ2|c5t}JR#OV3mezQpj<ftbfLByzyB_p3@
ze22HGM`WPsbtmhJso9_=rH~Iwki_YtV3S#X?<2fjpkQvU&6A~m2Z-Waup|aS0`0QI
zc$=KW6?=vHnIE!xckJTw-0<r2i>dvv9aD^rUrj4np)}5#q_RDq_N^W>C$~3!cohSV
zQ(^h4QeQ$`G9`xLqTGry4Qt2hnjOdN0*33{pQK)`4$O>Sb}lfpe<!q2pQ@+W0dJ&w
z8KRGCWmuocfE6oYm@|TaX4rQ=vs4cVCp29hW`=V1E)J!?wRgFC_k!Q!N<WkaUhWH)
zu};}un=siKC06cHSC*wPKhu89_v6`4x-9#45~M7lE(PSLQ~;OSHWkngT<wCtXuZOz
z53nnDl#J{b!Hll9__!ZeQ_&ziDz(bI3tHa&%xPD5Z0nkVaPzCr=VRZ0T$`wk4ptP`
zwOwjBw%EzNt#yoDK7iBOx_h7o%Q^w)!i_SgMKF~!68=J_ncI@o^~l+BBDp*?W=r<1
zj+0EMk2+zV?Kg1R+rHgP_Z8l931uIE6Kp_9$}C;kg>jHBRv2U0p#izkZ{@q@)=vTh
ziT^h~(Vt?*4oh!aPTrI6RX|NugvLXhif4(7Seu&hxJ?eGuHvJs6F+APOD!{t67NCE
zBIlj#S5&;}&K1Ku<Tk+g-&|OAmW0j8!Ue>d)vksN1?TQ9QRQ~4H~Fb+vse08tQD`D
zzx{ZmcY1LM0s)F;dhcHpB>2#>-%;T6d2GTC*hW8XAm{E$$tKyNV9m@D)Um_n+K61v
z9jluYx?Qa1<MpJBhlJ`6NbKf7mGJQU`q{rzziZQC+s-30it@rWbNgte6jn7GKI#eQ
zwlFWtzqM>(VWFu^lW0SSfPF#wmEb^;rWb5!p&6}Vd+<%Q%r7>~x{mGZXZiY1Qi|j9
z-brhljMz{rRE4}kkK-plPhjan=n9A)qnl{mST*?YbE3$#*#Qewa{!Bdod50im+$Qb
zGDUntkTb0~f6w;GpRKBb1X45H7Vo|cU4Oo3pmz6s;N%w&^M<8;fLRBO3zeWf5n@u(
zkjYaf`YC7gwg769>0x3u&4}CUQjgi(Ig6BW?Hskr`-^5BKbN+o^kp*&GOo51*ofs9
ztXHchhh{7rjIPFgN|Mv!j~f-zRy^<ErcgZkU1svz$-(F`nWsCggP=|S8pca_8ec(h
z?O0T92+N&=a-_xt`xJ}6DYrRip&36va+3iU5bwpyC+w~`?Qn*d_>$y8w8gn8JVy)T
zyQ7rA+{?{?=|oJr`5`3YngTWA(QLH=s>6L{;N@E-=fUU+B?u5=8(p-w)m;Kb1-TE@
ztO2U1yJud2dwGQ&WGYoQD^;hdAdvav(pQsfXX?1BZKL6}kx?k;5ED>jh^HNH2F?=T
zDo*G61%NXuyzXa)DwNG=iapt_5d(Q6dkS8tk0$BVb7Hn&ihw28c6DSVtbT7CJ33d7
zgi9TTo;=K{(~mUqm4?t+>X1MT7@`wN`AL+%a##C(e#tMFUa&osqV{LJXdF^$<}L-;
zHS2DqFX?)$(28FoMuX|qyn>z4LP^3-qn1$hr&*x>tl}3D@pmr~)Ttj*raqeW3&W#a
zZR^ixH!%n0x!5q__pX{5+L&v(yM&f(B`F4eaC#(g8>0KXl8|a}5H;!LaNQl~GHS?4
zWK)&f6Z`HL_S>52f(waHrRSiJ;+-uS=uY;G+?(F;dA+`(ks5t<{ry-9n(nz@Yf2LG
zr%P|3mfs{>!(Lj-FACYT!jb)lYbZI=eM~^i!%Q<CDnemSBvv=gbkx^>WMom)aE{{f
zaJMIat@?I(dEyY4-U7K)yNbt0ZY!oBL>q5YE%?eQS$pz%7UW*^GSA_|vB+tJ%pLcc
z!&HB{N3VKUubVzi^yYN#tM`3$Is_9&5q&;}fG1#l71*#}6jepQ74BBY|7!C=Pj#2*
z_aTv(XN0_$QA;mDK0AgNX>+_tQUUp$^StBxLoCD{9W=V7@tY_wfYV+wpPA_@cD41O
z;G65~?9o!y!TG$s$mzP&zhzHugdFI$gv=#D-F4!7`{AOEX4t~YN$bm_nkOIX;_PiA
z=9Ot7ksXZ+F>41fk#y0BOy8g+Y4vjkw?9!0Xz^Xfm%H_LWc5JCpsV0Zr(cf6!(n_H
z2ulc2oFvO_qwC0?*(QrXys=SF>vuPpt7S_ZWV<VHktJj{Zgs715Pa4XPNdy$-7iZc
z)B%aE$P-cGF)!doumeg#s*Yk4g_VS*BWY-B3`b;F6iUAW;;jD^95th+wKR4^=Yf4Q
zN#gEH`^%Mn=S?q3CEBzt4PlaYaC57%Ge-E+l*^I0EY*Is&?mnrtgpH#@6TSXQMI<b
zU@kf~n<HuZ_MLHG!Rsxa3Ou$02@2oKG9A4))>^l<GT}=*lYWWN_XL@}qE#GolfpVl
zv%ik~nSxp{W2J~_gadKS1Yn@t9xIF1J{Hjj!6%ZtH<BNYl`ln{laMT-=GS!%PrCby
zLc;f%@w;4QUZe$ltFf`~#BHt;D#&j9zPSmqGGhj~sG?UXad6C$k}p3Dgs+bbM;R@Z
zh>v`A-veKz{mNv2Pqgjb`YkuBGhAXcq3>JMU&a*=R4U@Sx2XIV=A;b9l|1drY}c<I
z&P$^LDtt>kOy?Dqwb>>H@65OX#O1sP+f4n&pD%L1ILJi25PVeInEDg2Od^!@p+Vwq
zWCM|dC)?j%>oRD5xY|Cz2VAjqO>@mYjD4{vfQl;xjCW|)E`S)ILD{)Q{U;SJJ}S*0
z)j*gWGi3M*Jukn5rlaPBa=51W%jCwhBksQ_d=Z)26Gw<q;|DfxbXrbLlO}CtYApkQ
zYh_1^hS=kC+M7=_1EkdOgN;1-`2zKGjZxwI1&LO+U6-nAm6MP<5j5B)2Wqjs#4K4e
z(4wyGgK-s+5J)}2WT*=oXWDl#!Yw)4#;S57Soo>C(1^Y$U4*DfY~>bdxL|vB%E+d=
zGK6+ij4b7k!(Jup5=^UAnnbfIa{|k3E!3o$EPFh1A?Egt@l@}m)m4X><2TqZS(s4{
zqxcp$!Bp#~{9_9!R7(rt{Fq!?jrDK8C?+(ZbX7tO%oK*_pMX&iHZ}^a;K_|(l~QuE
zN}R>9t<GE;yLUxBSpmiYCUWxV6p$^6IfcZaQFTBsxUWM1$Al>2%IGw@yj?k_NA)p<
zZ~E`MRbd8`dVk{=6y<pOR``k@W*&4>nMZKy{3_m&;gOJk|LI4W+{Q>V0+%1h++0f+
za<I`al~I*(W$ykJOU1b9=toy-+?7`o(C0grtl<XZ<stTqwp9A1Q#Ue8yFb+Dq<vJ)
zJfI?})pfBysJR&s>mKnvtE{0Qr?x2lMqyDdLH`Y;JtK6v@T4_tC41goLB67ajRy5r
zW<*HI;8UES@ocOA`Iq%#uJ1=pbx$jS@wp||r(sn(Z!FO68@Ly;8a#jOYaP$p(b_N>
z0T5-POflD*@pVDI+x0hILeIf&MeD-1bfi>y+{ODY+_8M%CXt>4RUe3EZwM8%amgjq
zM|Pn3ScBlPAqG&~a%-|-Li@AXHl8f5T`A4)3hiPK9X0%SAFha0bIRg03db}S%&@0t
zPds<4^IH`guz(s^i1z<SEx~_B(`}%}Oz|)NoY|cZuY?+8+S0BalTVp=1c+}Cm|3Xl
zvzdM$#!e3>230Y>q-(ouFZzMk>&=i$xU_y4V5IwU9NfLoN{0=|F^d3V+oEGm6J%Z_
z$GX`_Tj`NK2=lO^`Sz{AbD9vZgfpzyPD(1s^7x!Lb+mJ|=6G~m=~`NCW?6+SNk0PW
zvLcS^A=IQeiU0Jt_WtKYt*`hi_~reCOC7j=KpYNuLh2Zv<@{8anZ+}&qq!<SP{yG8
za&YD7#n3<~RKe?+fqsE;==12-)an*w+ASZSz~25*3zt=b$G)`!NOB#^rgAeNx&2gh
zBDo8JV<$^3Ff8g<xk$LA_4)>Le<@^V?2mc9JLk@a<uZI~Zs%ZKfz;*fL|?<VM#1EW
z#kvIULyRXGg1e)JIyk6cHki{WEXchZ>3>{;x!#mXQwlq^I<cC+fS<w4LgXU%8XB5L
zC$+Mfo?h*BdnW;L<a{>#lJ6z^k@~?>vdMQM6vtjK9H**Jvxb0saiff^WwWSLAH3|C
zDU+Q|<SeZ@3p4x>%K%X;G|<}mMKNQ2@<YHQCw-u9wEk3E_@^Pvf9%K;=N}cKgP-gm
z#DT5GWNJ&v<^=AnuKKx_YL~?aW6Qf~{lH^R6y>40>Zgj+!{p;^<8!Vel=uoZlB|b{
zdKX(RFK*NRqTsR(@SA*4+e`vyvQDY=Dq2gcs$Pi+PfyX7x?&O*>JG&pwl^pg9H%r!
zJP!_455{iY^5m_0<CwJ5?~EQEWW8Z4IxUyn3}txFC={QLhkp(dE3R0#^nVqiX!mjB
z4S!I@TxGH+7xGn)5AWdmi$1m_O>b{~nIv7K?;Cw-h54Rc8_U!21x>3%+%ldMKfTFW
zDAMLLP;G0BlFmoK9S__=F`g*n<G~!3tm#p*ty7`sT%Aw(3&~j*_k+MZ$MAy^wG{2!
zZe>Nv4Ncf{zh7S+)(Y<6b{@8GC}BHkbVvj8Mh?q+L#-Fh)AlCkt0;O`Da_O$T1DHj
zPu`pS2IP#+|D&7{+q-{4Q2MvI@Be8{{=-=AUuz`)UB7q!lVQ@oJZL(j#_}D+2~$+V
z@0<!KXm|S~G&6r|y^SeZGxxWjhBHmEzkA0rJrvbx)IRDf)XX2W@$cb9%~prrDA`vK
zkv}SB;`CM@dHbJF83ZIuU)XxW<MWQw^Hwg0)XL?%5WPrYeldf{;pnB~2_)b0(@w9I
zLl>8gUlbH4+<9nYq7F&hRY7-?!zp^3^U<io^{c0Pj$v~{1VkmwZf69YI-z7xxF(z0
zXg2wVu4U{Ev*d2yz??jnJ=bc{d0(TCrNa4Z_<(iuDIcRB$n`kOn|XDL?SC<p3WGa2
z_7lSQpe<%bpCAxUV@Vx8Kc29$<j>r_XG?E;yWTc)OV?79*P7;2{g81yKX;&`WYB6|
zxxfK<q~Tk)7WFyjzIql~>3&#iYQOJ6a(o*E49noi%ktgI&>n}xmBg&+tf}SncrBG8
z^bB45c<DXu7Cke}Bx&Eagm5SOR=%Ex^sS#_hmCXjVEvnjLF{?5AwlDXhkdffcyR=a
z5=OIY!E4Q43C0I#)4;+T7fx@x`~oQ7W{C`L`@xUS&!uiIDRV17I(>`}W?FW%;0{5n
zjF5-7kwN+6{XTu{uUoR3oTm5poN~XZvHXnu_BXD`XJ7`jN{Ad$y&cR>T)l^%{q6yA
zS3SM6Y8U(cf!@#V58HrGPK{&&!|>cDmJn!7iMn`2Qx2ZAg5K`|3EifvVo44Wjdgsv
z=dM`mE%y1Pn9R;I;c}+SJ|A--0zdxPT8y)Cl(bq`*vL#5^!6U;!-{DQxg96l9M~o;
z`dlPw23KzbzM-}Osd8CcF)h-_W~<R&SqG@W|8pywE5p<eTb2-Unc29SJC0+F7yOjZ
z9))|uQZ}o5lGSU}`_?{7*1Ug3uSy5I4+Wor@jx4j=9k)zAKrvV4^F$khkwq;z%3CA
z;HVr%xZTWDeZoc8w0UWw8Lg8M-lIg`$Av-O#oZQkoh$o-VwGLb`j=-1!qikIGW?A|
zsqfVvw#X$W=9;W0x)qB4d2l&;8O9o-i$DCLPvMM~#8>RDD_liTWY61mT-J2(B8!!?
zTLeocTAPG4?f*$6{<B#8?|fbm_^S}~KYB3z(Y`+nLQ2$QSs~85y-;RW%lHq~fvpZT
z;f0UoXFk0QKNl}$%kl%pJ&xuLd88|4lvp97tG$I;(R#F^0uU0#tEecaXpV{hmzb(b
z*KdZ)&Fditp6Xei4y8I<0v?C1tG@q94&`ytCs3@}#Iau@?r6I>e*WS<i7HUwacODo
zg(q5@6oP9vvu=8~&~`ojO-5Dsp19PCcr=*m1`(3$Tsw~yoyZN@rM7<!Tfdg8Ns(<%
zo(niOD+}Qpqwn;5iA0C{lQvrTCGU18*NG9E4z8!I34t)X4xdIsg!%Pq)OYpz*1g$S
z63+1~mYwzqxhW|wDXY81+~s9gR(X(;x%agEag(rjw|!5CQ9P^cf%NmZ)`OW-r+V`3
ziJVRz%5OUY?z_K*6niYceaC;|w@RqYL)HFke5xIlvAMje*1w|<0gT(fBM(vfMgKnT
z`)m5px4$C~eW1e}`5*KU3d&>?aL!!iYART7DgxKLSQwj6g4YjQ^(p6M^V#x>vUOJ1
zGZ%y-Vsn?!S3nbJjT1!qa@b;c-8N@mj$Ty0A8z&vcqr)DgS(g}{dFMn20)NgQZszM
zd*r?E0>1zLBs+vMKjZ;{Fk0Vj@4`|xD~jxSG%N9xQ&Rp5TO$=%d;$R*vAz5p6saOo
z-8igspKiP{OCVF*LE>%aSz5+tUbEQ|P2%CFXD3yz1r*Kq-**u$h<YQ?YG~qa1`es(
zS)<}V<qy6PP7>M3i0mv&A;^4BL$WkkWImE)64=c_zE3hH{B$aI`_4g%HRa7_<=8Ok
zv=jnU?N1VpK=k1qV_lC^ILw24`<G6UFp$(L%@*u(z@Y6}vf8(Oe(#1Y$(cEE$M7E0
z<E03mbzv;vgY<kkafsmvNbZ=>IzN#~)9vHpk@+BXO@jUI_6s<c?<CPy$fY4^og=)W
zoFaWS<w>l#;lN!svaX+9^>ZmV{(=U7UB5eoJK9HjVz+~fGD3$&ErqfRi_y7^c{a0C
zDQ;;ELevL}?+5^p>2Kg<WaE>M|Al8CkC+Kw5*YfNFzw$B6CN#whT4gArDh3*u5CN{
zhV$D#pSL|dV*@0~{{~ovNNQau=b)R#R@;;lw3OFfGAhEyJCi6Z@_(XQbDSwT3-#vT
zo!RYfqd(Jy-@#QT6kR;9>1CX=l_w2(UAuIoOlVJd%5o?cLsBO494?eA42;N+Jsnk<
z#UL<jBKphQ!nKNiB}bZkN<8-v$#tYPUmaR#zds{n_Cc-sQob<7J#t)d_J06)g_yd%
zhf|N^c)mg6&(Ik<CN<|IEC}h<yN(MzI;beZ0_!{+DCs3VuweKdpx0)~elCcE1jTu7
z{>UfVw)v+jY@sY8O3Lmn&-ac!C_bmbzV+(NtU?u472@5gAXyPje7e}5#q(400XWdz
zCzg6qaM=^_MDYqVpDNRiFkY;hIiVO>xpDkc;>*)Jv2<)^AM1jTR0aTr`dlaBX?Mp?
z43<TOx?dzNU;WKu9L-gEb6&p(S*$aXpKq_n1}EF*n;eCaSRrU2ws!ss5cth~8mWoD
zuMyC5d1S}yjIsUt{RqmF1lJQ4MDI_UXZbN@dYreU_CwGG8+Nyhv$utJfu<AnbNC;%
zp1<$=?`S`#`7G=md@sjJ^mSYd!=O2c)%b%$U>%f#=<6bWjG9iPl_~LlVQ=N&fqG|l
zh7VZ|l^uwg?MU0$y7cFBHeba>`B$OOOqwMV`*OU@$BVaE)4w~KpN}s}@=Kb|0gTA)
zGlb$3I*#Jg^K!%q)0QPVpDtR)_2c@Lf~q8YNP2pE^4JUAV8okN2QBpQZU7t-+upq#
zW}pfTAU_Xt0oZunnmX(>Q>UTT8F+<D>abH<!_OO@r0=D0g@xzqn?o@rX)C%U`5UJ}
zFsC?6-7;vuUG7Xv*ZNU`<@5y+O7t0wq7x-p`xpt1JT88vLJtYB*Q|$D=o=&&9tZTy
zaB3yoP6+)#k=}hT*;>2<9>xPAx}tsGy9bukYz33vvvxnr&r2v)DVPkle(rJyDmqls
zx7%wcZd!|)&yLFdV*=q?9x@^CFDH{Qgt5oD0SIFy0lL#n>P+(y9nD1X`3&9*dz8)3
zE*@Wzm)U%7X3EN>B;Y(*jFk5>cYd<g7p|Lr)cS$A(P_ifm^z~|;raYiLFlfA`>j|h
zd-gl)P@_{;!Fc$Z+juMf7J#oqR}!;GG5~m(Wegct)(~FS<FDoA@iPD2F)EN<Wg%=M
zgV2wdn(Qi*q7AiIOXc}0lFSGCtPanDW!{(TBN(ifzI!^cvENX<;=3C=N-L(V+ovo7
zMDy1IbN@w-Uvtj+({5Q)|J$>G&Y<wBq_?Kph(nGv0{SiC@tH$J?%tg2sCw0_FJI;r
zb?z|pKWfi4VfLdiOH5x4a5hZ2bauj0Hl^;wV&iGLn2@i%EzewjsV>%bxE>(nuCA{A
zr7E*Z>q10eY`=&{z;596YMY+V`L#pX9q!V0wLGg43z&_IVBgL|Vs26r8W*-}D*&&P
z%uc&F$M#5w5AGH!)Aa3v&CIR2OT)>BFLVp+iGJB%6LMQ|-T|wp5nll&!Py)#^amJ|
zd~gS^v1L$zQK>zMqwNX;tU7OOqoq*-%SD$L3Vu=aYaTjmDN_q*C+UV5EYeSV9+VO_
zo=r4ZnuZf%)Dn1wHxxUoTc%WgBnmiiwxkV!w>U_sW~C0&&2Qab$jDok$ASAQOt*Uo
zJQF5Wj5Vp1at&x}`UgN*vv(=ym{Z1+OamuWQZC7Ly8sIznN6VVE3I8`TVA+YW%U)U
z<zRPODG~hs((#$(&*e~1`0sJ%*aC>P_(E4sgj`nKM<{ZDj#>Gh`^)j6YfG+;25sE1
zS0_2}g-kDJ6|pM9&X-n34FafZ!fBv(#4ffnu^4^t(#Oy2nxC(wm<$P+zZjp69(pFt
z7sADqqtWlsYLq~mXur`n)mwkLEMx05qMOX5u^`)sKg409RCtMlU&~Q0gbP1>sokGi
z$rL6<kdz+q7-wyF`gUC76M}A*0-aHzpuD`+T<xBDm#WZhQ%#_yPI~p1KJJf~ox7y{
zPdu=HxzqmOxSQBN6SkHSf98%*u&kK2)&Y)dIv{CR6Vs&--i6M+ktW)e5at^bBl9WV
zh&kq<pk4xJZK1P+k??*dWs_;E2pdwNny8%Vw4AlQg3@|gBp6qJpu+hI`%v~WtMRAv
zIxU%2c2T{ZpTEDqB5|wR!eYyF9G-4yTCmcHv<*r=p?3=F#_-QyuO5y)yy809A1F{?
z@%0_Pt^~^OHMoF0Qzu=$<%a9HpxXb{6`HY<h0EeOG#%fm*!wEhclA>f3YGFCf^F<+
z)?5_Lhf7J&#a70EvM&{ZzNG<v>pu+%zA1=K^S#Em^#j#MmKCKB63eIrpWM9o60%(y
z?!mm`e7sV&M8{;9aDS|>l_NmoVVcIozOs17yey#bKeNOdHDs9ntaN|WdHTRxZknaP
zDatKIphnHdTNpz^$C#DvYQc3(O!ziuHXJ7_vdU4|{cqq(6ceGg9R{A@)EX6&QPef+
zku$J>W(XW4%M4t5G4eU>S*N3#NN$F5n}5-wTxBrwe+-P$v?D-=z^)#Of@W3CCI_*{
z+DFa3Aqm5A_bo^3-jqLSZLP0KJ@$e+r!k4Rd)+qm;(DBub&%iiEjiqNY8*pU>`dd~
z0RAxx8(+o0AS+w&wOYiNMcm*{GaeBXvLSlWT4B>lHEl~?N?ibQmsS(!@G14$UB~$5
zo2TOXpco}?92FtFBcKu-4a5)P-y*~6$AX|@g`(v?AKc36K+_nRkXtmF8vS2F`sV!`
zPZNJp^qzFx!6;4sfS<?%OJjs?9-l^&e^KyUn*=ItI28hzw&??<9X5)52a{S-N^D*<
zEy48&j{8BPMoFs7Dfzs7w;<1+Xk9+!)lk)#0-y19_p~Bv;F@=?%+IV_mJdrQFC*UD
zXM=CQ*5h!t^B=Ig{F9N2)F)ddQx=E#zI7<us)G;)^9pY3Cm)~IPu3z8k8kFEI9nZJ
zw<OB?Sq0QILB97<t;nUy-8A)K(srkiden!yB*iOb9A^gySnl!DU=RYtBZ=W1G*3Y3
z(Nj;l0DoKP+!M{E!1HyYPkqi>AwLTC$7|U;K9Pt?XM6Gb^>Ii9@)+EIe8QxVKpA3%
zAL`WE=;yu>s){wLi*wA>ov4fXDujQ~8U7A3*To#i@VJkk#<-?x@E&CiA`T`I5*pTN
z*%*UuLrdZPW8C=g8c|{*G6>o!&Etqje!uEZ2_#d+65jc3x!=b5p7F&+-MJLdiqWyk
z#C;DMFvYU$i(i;`hB4|wwIe&TuaB)}LRyly7{OEd(FfSCz5&CMUk^SGOS?7@G(Ktj
z2=F-NZ4gh6#dk%Gyz8^K5Y1`u_lbx5EdL{jm12IJuh^Rw8KQe|kRrmom5;4g>jrX$
z0I;fr!5Jz;<q6CI2w?D=B%|!OvFWG2GC!=d+<bZvn8cPwLw!_`$a?4VpdQ5yv?d<W
zNl%X<3y#-UfEK~AsQUy}@I~wSfS7yr6OPRpBYWx*_A=$wLp*8HhH1ueyfJ(M3_8m%
z(C|EB({l0`9<~}{W-(cB*e<ONlVtNl?r?Hq;!qq>;LZCrFf?d$#!%^X%A?1yEr^$b
zx?CBD?)J#&`24)7-q3d|Uu|opz2L&e;CkQTlgoK%UOauLINMx^Lw)_2(*1l$z&Ogl
z%caD5l1FZAwMD`C(_0~WdeKZsdyW%><eV_YT?>vU{GnKn39rKs#aMFk3B1s8`G?$P
zEFRv!yYW2wX5P!+H8CC|FOq%;xPDB$q7XTmM9X<3J=5}t2r)SwXv~&>C`m1j{(9j1
z;4O>GrA*dysT7Va#$3DdEW0@`VC=-v#mJYiYXQC}yF3>d<%E*iqOkA`8(CQ-BeyYB
z59jm%C6{ej8~6EkgL^KLQOX<;@WfEMg>xZKChFXwW+va1)Cyc+53;YJYN|%-(nac4
zTOMC$ym(>4dy<w3b-f^whWo3&j_{7d>8Z(~jpa@u30j^@>iwdq)*CvAl_gm~J&r?T
ztR^E)d4accQxZ4t;^JXHvZXr7xcdElu7ji~r?~Go8O-HB_j2O9<j(Vlgnl63O<l)~
zU~LFnsX2J}`cF92q>}v7iN104aHiJmeeo!PXzHtSpRPQsWi0P!QYeq%^qHq3z@wgn
zsIc&M<V1bkT}_p1b3eA2dV^(VG)gc>Oh=wMb=oC-@5I=Dt0v9F8xm)2H5(84_e13X
zxB;(hhUr{0jBy5C{o_Z|;Y4+X%RH!B4q13ond!(M&*v2=*~N3jZ74I_d8xMMBX>WA
zyv*vU06?!~0f1~93Zhstf0{o=ig#VKfT&|S_PE~h&5bjw4fRP9Ax#WXbD!<rn|-xq
zh2zYLF$+#_T5<HmtVIy(+JRX%sAX(fMHD?ID&ker$)Ap}RCjjRZLIBBJs+#Vg*JAe
zQc1*dGj#)XNERVIgkcGMc@4vSToLQ2t8p*K^i#55oc)_8S)O_NNHfl*<isR%u1Cos
z^T3FF#%u)}igYzLwroCaI00ejJoh5fOE6GN_sZPjx%<I({)|JOJ|4KGRM|5r<B{hz
zuf0-OsW9udQE0^Lx3}DXgg1*(xen#I_>C-wqosvkD2Ltr^evn4O;D>q?Ui}z!Gglw
zOKu?0jVsib$)fa&7aFO$b}qF>p{&Kq7RzR~Zpf46>=*mLikppQ=mexLQhv~4sBfvi
zH1hE9F5g9rYPltgy^Ff9{I$K7w5%DO?Z%cG>m3K5K0O9hv)my+cD3kmZ;TWjO@sho
zNtBad=W4`ZC&ngc2~O0k+^VOOAa)p|$ISFB3TR4j1vLQ-YM<VW;4X90Yzb3?{GCXA
z#1;fO)TbbUm7eSCNwgDsd=JiHz2WC6IS^fAJvCQ!a%1K0rYYtYjh&+*_<2NL&W@-D
zw(xYrzZTtFq4N<~jBD=~g(BPrGQAe+4k)F+BaZ%PW<kmLPb$5$|6u<=#zXj^mGH_5
zo%-}~AcNFJTDxGUQ1p0;?Lb3Q^y2MQL)o<OdrNl&&+%xesc}k^=oDRlFFNo2ovKSE
zTo?jw7)1fzOmrxKb2S=6Y~|Iw67FD4E{5caJZYrnc*{L_=p;5^qQ3R9RH5`{ex7w$
nStzjnB#|fMk0joIW!wLsUH$Ksnf~si{H5XhFWgS`Yx;iyo?w)7

literal 0
HcmV?d00001

diff --git a/_static/css/tabs.css b/_static/css/tabs.css
new file mode 100644
index 000000000..95765dff6
--- /dev/null
+++ b/_static/css/tabs.css
@@ -0,0 +1,91 @@
+.sphinx-tabs {
+  margin-bottom: 1rem;
+}
+
+[role="tablist"] {
+  border-bottom: 1px solid #a0b3bf;
+}
+
+.sphinx-tabs-tab {
+  position: relative;
+  font-family: Lato,'Helvetica Neue',Arial,Helvetica,sans-serif;
+  color: var(--pst-color-link);
+  line-height: 24px;
+  margin: 3px;
+  font-size: 16px;
+  font-weight: 400;
+  background-color: rgb(241 244 249);
+  border-radius: 5px 5px 0 0;
+  border: 0;
+  padding: 1rem 1.5rem;
+  margin-bottom: 0;
+}
+
+.sphinx-tabs-tab[aria-selected="true"] {
+  font-weight: 700;
+  border: 1px solid #a0b3bf;
+  border-bottom: 1px solid rgb(241 244 249);
+  margin: -1px;
+  background-color: rgb(242 247 255);
+}
+
+.admonition .sphinx-tabs-tab[aria-selected="true"]:last-child {
+  margin-bottom: -1px;
+}
+
+.sphinx-tabs-tab:focus {
+  z-index: 1;
+  outline-offset: 1px;
+}
+
+.sphinx-tabs-panel {
+  position: relative;
+  padding: 1rem;
+  border: 1px solid #a0b3bf;
+  margin: 0px -1px -1px -1px;
+  border-radius: 0 0 5px 5px;
+  border-top: 0;
+  background: rgb(242 247 255);
+}
+
+.sphinx-tabs-panel.code-tab {
+  padding: 0.4rem;
+}
+
+.sphinx-tab img {
+    margin-bottom: 24px;
+}
+
+/* Dark theme preference styling */
+
+html[data-theme="dark"] .sphinx-tabs-panel {
+  color: white;
+  background-color: rgb(50, 50, 50);
+}
+
+html[data-theme="dark"] .sphinx-tabs-tab {
+  color: var(--pst-color-link);
+  background-color: rgba(255, 255, 255, 0.05);
+}
+
+html[data-theme="dark"] .sphinx-tabs-tab[aria-selected="true"] {
+  border-bottom: 2px solid rgb(50, 50, 50);
+  background-color: rgb(50, 50, 50);
+}
+
+/* Light theme preference styling */
+
+html[data-theme="light"] .sphinx-tabs-panel {
+  color: black;
+  background-color: white;
+}
+
+html[data-theme="light"] .sphinx-tabs-tab {
+  color: var(--pst-color-link);
+  background-color: rgba(0, 0, 0, 0.05);
+}
+
+html[data-theme="light"] .sphinx-tabs-tab[aria-selected="true"] {
+  border-bottom: 2px solid white;
+  background-color: white;
+}
\ No newline at end of file
diff --git a/_static/doctools.js b/_static/doctools.js
index e1bfd708b..527b876ca 100644
--- a/_static/doctools.js
+++ b/_static/doctools.js
@@ -2,357 +2,155 @@
  * doctools.js
  * ~~~~~~~~~~~
  *
- * Sphinx JavaScript utilities for all documentation.
+ * Base JavaScript utilities for all Sphinx HTML documentation.
  *
  * :copyright: Copyright 2007-2022 by the Sphinx team, see AUTHORS.
  * :license: BSD, see LICENSE for details.
  *
  */
-
-/**
- * select a different prefix for underscore
- */
-$u = _.noConflict();
-
-/**
- * make the code below compatible with browsers without
- * an installed firebug like debugger
-if (!window.console || !console.firebug) {
-  var names = ["log", "debug", "info", "warn", "error", "assert", "dir",
-    "dirxml", "group", "groupEnd", "time", "timeEnd", "count", "trace",
-    "profile", "profileEnd"];
-  window.console = {};
-  for (var i = 0; i < names.length; ++i)
-    window.console[names[i]] = function() {};
-}
- */
-
-/**
- * small helper function to urldecode strings
- *
- * See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/decodeURIComponent#Decoding_query_parameters_from_a_URL
- */
-jQuery.urldecode = function(x) {
-  if (!x) {
-    return x
-  }
-  return decodeURIComponent(x.replace(/\+/g, ' '));
-};
-
-/**
- * small helper function to urlencode strings
- */
-jQuery.urlencode = encodeURIComponent;
-
-/**
- * This function returns the parsed url parameters of the
- * current request. Multiple values per key are supported,
- * it will always return arrays of strings for the value parts.
- */
-jQuery.getQueryParameters = function(s) {
-  if (typeof s === 'undefined')
-    s = document.location.search;
-  var parts = s.substr(s.indexOf('?') + 1).split('&');
-  var result = {};
-  for (var i = 0; i < parts.length; i++) {
-    var tmp = parts[i].split('=', 2);
-    var key = jQuery.urldecode(tmp[0]);
-    var value = jQuery.urldecode(tmp[1]);
-    if (key in result)
-      result[key].push(value);
-    else
-      result[key] = [value];
-  }
-  return result;
-};
-
-/**
- * highlight a given string on a jquery object by wrapping it in
- * span elements with the given class name.
- */
-jQuery.fn.highlightText = function(text, className) {
-  function highlight(node, addItems) {
-    if (node.nodeType === 3) {
-      var val = node.nodeValue;
-      var pos = val.toLowerCase().indexOf(text);
-      if (pos >= 0 &&
-          !jQuery(node.parentNode).hasClass(className) &&
-          !jQuery(node.parentNode).hasClass("nohighlight")) {
-        var span;
-        var isInSVG = jQuery(node).closest("body, svg, foreignObject").is("svg");
-        if (isInSVG) {
-          span = document.createElementNS("http://www.w3.org/2000/svg", "tspan");
-        } else {
-          span = document.createElement("span");
-          span.className = className;
-        }
-        span.appendChild(document.createTextNode(val.substr(pos, text.length)));
-        node.parentNode.insertBefore(span, node.parentNode.insertBefore(
-          document.createTextNode(val.substr(pos + text.length)),
-          node.nextSibling));
-        node.nodeValue = val.substr(0, pos);
-        if (isInSVG) {
-          var rect = document.createElementNS("http://www.w3.org/2000/svg", "rect");
-          var bbox = node.parentElement.getBBox();
-          rect.x.baseVal.value = bbox.x;
-          rect.y.baseVal.value = bbox.y;
-          rect.width.baseVal.value = bbox.width;
-          rect.height.baseVal.value = bbox.height;
-          rect.setAttribute('class', className);
-          addItems.push({
-              "parent": node.parentNode,
-              "target": rect});
-        }
-      }
-    }
-    else if (!jQuery(node).is("button, select, textarea")) {
-      jQuery.each(node.childNodes, function() {
-        highlight(this, addItems);
-      });
-    }
-  }
-  var addItems = [];
-  var result = this.each(function() {
-    highlight(this, addItems);
-  });
-  for (var i = 0; i < addItems.length; ++i) {
-    jQuery(addItems[i].parent).before(addItems[i].target);
+"use strict";
+
+const BLACKLISTED_KEY_CONTROL_ELEMENTS = new Set([
+  "TEXTAREA",
+  "INPUT",
+  "SELECT",
+  "BUTTON",
+]);
+
+const _ready = (callback) => {
+  if (document.readyState !== "loading") {
+    callback();
+  } else {
+    document.addEventListener("DOMContentLoaded", callback);
   }
-  return result;
 };
 
-/*
- * backward compatibility for jQuery.browser
- * This will be supported until firefox bug is fixed.
- */
-if (!jQuery.browser) {
-  jQuery.uaMatch = function(ua) {
-    ua = ua.toLowerCase();
-
-    var match = /(chrome)[ \/]([\w.]+)/.exec(ua) ||
-      /(webkit)[ \/]([\w.]+)/.exec(ua) ||
-      /(opera)(?:.*version|)[ \/]([\w.]+)/.exec(ua) ||
-      /(msie) ([\w.]+)/.exec(ua) ||
-      ua.indexOf("compatible") < 0 && /(mozilla)(?:.*? rv:([\w.]+)|)/.exec(ua) ||
-      [];
-
-    return {
-      browser: match[ 1 ] || "",
-      version: match[ 2 ] || "0"
-    };
-  };
-  jQuery.browser = {};
-  jQuery.browser[jQuery.uaMatch(navigator.userAgent).browser] = true;
-}
-
 /**
  * Small JavaScript module for the documentation.
  */
-var Documentation = {
-
-  init : function() {
-    this.fixFirefoxAnchorBug();
-    this.highlightSearchWords();
-    this.initIndexTable();
-    this.initOnKeyListeners();
+const Documentation = {
+  init: () => {
+    Documentation.initDomainIndexTable();
+    Documentation.initOnKeyListeners();
   },
 
   /**
    * i18n support
    */
-  TRANSLATIONS : {},
-  PLURAL_EXPR : function(n) { return n === 1 ? 0 : 1; },
-  LOCALE : 'unknown',
+  TRANSLATIONS: {},
+  PLURAL_EXPR: (n) => (n === 1 ? 0 : 1),
+  LOCALE: "unknown",
 
   // gettext and ngettext don't access this so that the functions
   // can safely bound to a different name (_ = Documentation.gettext)
-  gettext : function(string) {
-    var translated = Documentation.TRANSLATIONS[string];
-    if (typeof translated === 'undefined')
-      return string;
-    return (typeof translated === 'string') ? translated : translated[0];
-  },
-
-  ngettext : function(singular, plural, n) {
-    var translated = Documentation.TRANSLATIONS[singular];
-    if (typeof translated === 'undefined')
-      return (n == 1) ? singular : plural;
-    return translated[Documentation.PLURALEXPR(n)];
-  },
-
-  addTranslations : function(catalog) {
-    for (var key in catalog.messages)
-      this.TRANSLATIONS[key] = catalog.messages[key];
-    this.PLURAL_EXPR = new Function('n', 'return +(' + catalog.plural_expr + ')');
-    this.LOCALE = catalog.locale;
-  },
-
-  /**
-   * add context elements like header anchor links
-   */
-  addContextElements : function() {
-    $('div[id] > :header:first').each(function() {
-      $('<a class="headerlink">\u00B6</a>').
-      attr('href', '#' + this.id).
-      attr('title', _('Permalink to this headline')).
-      appendTo(this);
-    });
-    $('dt[id]').each(function() {
-      $('<a class="headerlink">\u00B6</a>').
-      attr('href', '#' + this.id).
-      attr('title', _('Permalink to this definition')).
-      appendTo(this);
-    });
-  },
-
-  /**
-   * workaround a firefox stupidity
-   * see: https://bugzilla.mozilla.org/show_bug.cgi?id=645075
-   */
-  fixFirefoxAnchorBug : function() {
-    if (document.location.hash && $.browser.mozilla)
-      window.setTimeout(function() {
-        document.location.href += '';
-      }, 10);
-  },
-
-  /**
-   * highlight the search words provided in the url in the text
-   */
-  highlightSearchWords : function() {
-    var params = $.getQueryParameters();
-    var terms = (params.highlight) ? params.highlight[0].split(/\s+/) : [];
-    if (terms.length) {
-      var body = $('div.body');
-      if (!body.length) {
-        body = $('body');
-      }
-      window.setTimeout(function() {
-        $.each(terms, function() {
-          body.highlightText(this.toLowerCase(), 'highlighted');
-        });
-      }, 10);
-      $('<p class="highlight-link"><a href="javascript:Documentation.' +
-        'hideSearchWords()">' + _('Hide Search Matches') + '</a></p>')
-          .appendTo($('#searchbox'));
+  gettext: (string) => {
+    const translated = Documentation.TRANSLATIONS[string];
+    switch (typeof translated) {
+      case "undefined":
+        return string; // no translation
+      case "string":
+        return translated; // translation exists
+      default:
+        return translated[0]; // (singular, plural) translation tuple exists
     }
   },
 
-  /**
-   * init the domain index toggle buttons
-   */
-  initIndexTable : function() {
-    var togglers = $('img.toggler').click(function() {
-      var src = $(this).attr('src');
-      var idnum = $(this).attr('id').substr(7);
-      $('tr.cg-' + idnum).toggle();
-      if (src.substr(-9) === 'minus.png')
-        $(this).attr('src', src.substr(0, src.length-9) + 'plus.png');
-      else
-        $(this).attr('src', src.substr(0, src.length-8) + 'minus.png');
-    }).css('display', '');
-    if (DOCUMENTATION_OPTIONS.COLLAPSE_INDEX) {
-        togglers.click();
-    }
+  ngettext: (singular, plural, n) => {
+    const translated = Documentation.TRANSLATIONS[singular];
+    if (typeof translated !== "undefined")
+      return translated[Documentation.PLURAL_EXPR(n)];
+    return n === 1 ? singular : plural;
   },
 
-  /**
-   * helper function to hide the search marks again
-   */
-  hideSearchWords : function() {
-    $('#searchbox .highlight-link').fadeOut(300);
-    $('span.highlighted').removeClass('highlighted');
-    var url = new URL(window.location);
-    url.searchParams.delete('highlight');
-    window.history.replaceState({}, '', url);
+  addTranslations: (catalog) => {
+    Object.assign(Documentation.TRANSLATIONS, catalog.messages);
+    Documentation.PLURAL_EXPR = new Function(
+      "n",
+      `return (${catalog.plural_expr})`
+    );
+    Documentation.LOCALE = catalog.locale;
   },
 
-   /**
+  /**
    * helper function to focus on search bar
    */
-  focusSearchBar : function() {
-    $('input[name=q]').first().focus();
+  focusSearchBar: () => {
+    document.querySelectorAll("input[name=q]")[0]?.focus();
   },
 
   /**
-   * make the url absolute
+   * Initialise the domain index toggle buttons
    */
-  makeURL : function(relativeURL) {
-    return DOCUMENTATION_OPTIONS.URL_ROOT + '/' + relativeURL;
-  },
+  initDomainIndexTable: () => {
+    const toggler = (el) => {
+      const idNumber = el.id.substr(7);
+      const toggledRows = document.querySelectorAll(`tr.cg-${idNumber}`);
+      if (el.src.substr(-9) === "minus.png") {
+        el.src = `${el.src.substr(0, el.src.length - 9)}plus.png`;
+        toggledRows.forEach((el) => (el.style.display = "none"));
+      } else {
+        el.src = `${el.src.substr(0, el.src.length - 8)}minus.png`;
+        toggledRows.forEach((el) => (el.style.display = ""));
+      }
+    };
 
-  /**
-   * get the current relative url
-   */
-  getCurrentURL : function() {
-    var path = document.location.pathname;
-    var parts = path.split(/\//);
-    $.each(DOCUMENTATION_OPTIONS.URL_ROOT.split(/\//), function() {
-      if (this === '..')
-        parts.pop();
-    });
-    var url = parts.join('/');
-    return path.substring(url.lastIndexOf('/') + 1, path.length - 1);
+    const togglerElements = document.querySelectorAll("img.toggler");
+    togglerElements.forEach((el) =>
+      el.addEventListener("click", (event) => toggler(event.currentTarget))
+    );
+    togglerElements.forEach((el) => (el.style.display = ""));
+    if (DOCUMENTATION_OPTIONS.COLLAPSE_INDEX) togglerElements.forEach(toggler);
   },
 
-  initOnKeyListeners: function() {
+  initOnKeyListeners: () => {
     // only install a listener if it is really needed
-    if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS &&
-        !DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS)
-        return;
-
-    $(document).keydown(function(event) {
-      var activeElementType = document.activeElement.tagName;
-      // don't navigate when in search box, textarea, dropdown or button
-      if (activeElementType !== 'TEXTAREA' && activeElementType !== 'INPUT' && activeElementType !== 'SELECT'
-          && activeElementType !== 'BUTTON') {
-        if (event.altKey || event.ctrlKey || event.metaKey)
-          return;
-
-          if (!event.shiftKey) {
-            switch (event.key) {
-              case 'ArrowLeft':
-                if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS)
-                  break;
-                var prevHref = $('link[rel="prev"]').prop('href');
-                if (prevHref) {
-                  window.location.href = prevHref;
-                  return false;
-                }
-                break;
-              case 'ArrowRight':
-                if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS)
-                  break;
-                var nextHref = $('link[rel="next"]').prop('href');
-                if (nextHref) {
-                  window.location.href = nextHref;
-                  return false;
-                }
-                break;
-              case 'Escape':
-                if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS)
-                  break;
-                Documentation.hideSearchWords();
-                return false;
-          }
-        }
-
-        // some keyboard layouts may need Shift to get /
+    if (
+      !DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS &&
+      !DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS
+    )
+      return;
+
+    document.addEventListener("keydown", (event) => {
+      // bail for input elements
+      if (BLACKLISTED_KEY_CONTROL_ELEMENTS.has(document.activeElement.tagName)) return;
+      // bail with special keys
+      if (event.altKey || event.ctrlKey || event.metaKey) return;
+
+      if (!event.shiftKey) {
         switch (event.key) {
-          case '/':
-            if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS)
-              break;
-            Documentation.focusSearchBar();
-            return false;
+          case "ArrowLeft":
+            if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS) break;
+
+            const prevLink = document.querySelector('link[rel="prev"]');
+            if (prevLink && prevLink.href) {
+              window.location.href = prevLink.href;
+              event.preventDefault();
+            }
+            break;
+          case "ArrowRight":
+            if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS) break;
+
+            const nextLink = document.querySelector('link[rel="next"]');
+            if (nextLink && nextLink.href) {
+              window.location.href = nextLink.href;
+              event.preventDefault();
+            }
+            break;
         }
       }
+
+      // some keyboard layouts may need Shift to get /
+      switch (event.key) {
+        case "/":
+          if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS) break;
+          Documentation.focusSearchBar();
+          event.preventDefault();
+      }
     });
-  }
+  },
 };
 
 // quick alias for translations
-_ = Documentation.gettext;
+const _ = Documentation.gettext;
 
-$(document).ready(function() {
-  Documentation.init();
-});
+_ready(Documentation.init);
diff --git a/_static/documentation_options.js b/_static/documentation_options.js
index 96beb0ebf..82426b5b2 100644
--- a/_static/documentation_options.js
+++ b/_static/documentation_options.js
@@ -1,7 +1,7 @@
 var DOCUMENTATION_OPTIONS = {
     URL_ROOT: document.getElementById("documentation_options").getAttribute('data-url_root'),
-    VERSION: 'v1.3.3',
-    LANGUAGE: 'None',
+    VERSION: 'v1.4.1',
+    LANGUAGE: 'en',
     COLLAPSE_INDEX: false,
     BUILDER: 'html',
     FILE_SUFFIX: '.html',
diff --git a/_static/jquery-3.6.0.js b/_static/jquery-3.6.0.js
new file mode 100644
index 000000000..fc6c299b7
--- /dev/null
+++ b/_static/jquery-3.6.0.js
@@ -0,0 +1,10881 @@
+/*!
+ * jQuery JavaScript Library v3.6.0
+ * https://jquery.com/
+ *
+ * Includes Sizzle.js
+ * https://sizzlejs.com/
+ *
+ * Copyright OpenJS Foundation and other contributors
+ * Released under the MIT license
+ * https://jquery.org/license
+ *
+ * Date: 2021-03-02T17:08Z
+ */
+( function( global, factory ) {
+
+	"use strict";
+
+	if ( typeof module === "object" && typeof module.exports === "object" ) {
+
+		// For CommonJS and CommonJS-like environments where a proper `window`
+		// is present, execute the factory and get jQuery.
+		// For environments that do not have a `window` with a `document`
+		// (such as Node.js), expose a factory as module.exports.
+		// This accentuates the need for the creation of a real `window`.
+		// e.g. var jQuery = require("jquery")(window);
+		// See ticket #14549 for more info.
+		module.exports = global.document ?
+			factory( global, true ) :
+			function( w ) {
+				if ( !w.document ) {
+					throw new Error( "jQuery requires a window with a document" );
+				}
+				return factory( w );
+			};
+	} else {
+		factory( global );
+	}
+
+// Pass this if window is not defined yet
+} )( typeof window !== "undefined" ? window : this, function( window, noGlobal ) {
+
+// Edge <= 12 - 13+, Firefox <=18 - 45+, IE 10 - 11, Safari 5.1 - 9+, iOS 6 - 9.1
+// throw exceptions when non-strict code (e.g., ASP.NET 4.5) accesses strict mode
+// arguments.callee.caller (trac-13335). But as of jQuery 3.0 (2016), strict mode should be common
+// enough that all such attempts are guarded in a try block.
+"use strict";
+
+var arr = [];
+
+var getProto = Object.getPrototypeOf;
+
+var slice = arr.slice;
+
+var flat = arr.flat ? function( array ) {
+	return arr.flat.call( array );
+} : function( array ) {
+	return arr.concat.apply( [], array );
+};
+
+
+var push = arr.push;
+
+var indexOf = arr.indexOf;
+
+var class2type = {};
+
+var toString = class2type.toString;
+
+var hasOwn = class2type.hasOwnProperty;
+
+var fnToString = hasOwn.toString;
+
+var ObjectFunctionString = fnToString.call( Object );
+
+var support = {};
+
+var isFunction = function isFunction( obj ) {
+
+		// Support: Chrome <=57, Firefox <=52
+		// In some browsers, typeof returns "function" for HTML <object> elements
+		// (i.e., `typeof document.createElement( "object" ) === "function"`).
+		// We don't want to classify *any* DOM node as a function.
+		// Support: QtWeb <=3.8.5, WebKit <=534.34, wkhtmltopdf tool <=0.12.5
+		// Plus for old WebKit, typeof returns "function" for HTML collections
+		// (e.g., `typeof document.getElementsByTagName("div") === "function"`). (gh-4756)
+		return typeof obj === "function" && typeof obj.nodeType !== "number" &&
+			typeof obj.item !== "function";
+	};
+
+
+var isWindow = function isWindow( obj ) {
+		return obj != null && obj === obj.window;
+	};
+
+
+var document = window.document;
+
+
+
+	var preservedScriptAttributes = {
+		type: true,
+		src: true,
+		nonce: true,
+		noModule: true
+	};
+
+	function DOMEval( code, node, doc ) {
+		doc = doc || document;
+
+		var i, val,
+			script = doc.createElement( "script" );
+
+		script.text = code;
+		if ( node ) {
+			for ( i in preservedScriptAttributes ) {
+
+				// Support: Firefox 64+, Edge 18+
+				// Some browsers don't support the "nonce" property on scripts.
+				// On the other hand, just using `getAttribute` is not enough as
+				// the `nonce` attribute is reset to an empty string whenever it
+				// becomes browsing-context connected.
+				// See https://github.com/whatwg/html/issues/2369
+				// See https://html.spec.whatwg.org/#nonce-attributes
+				// The `node.getAttribute` check was added for the sake of
+				// `jQuery.globalEval` so that it can fake a nonce-containing node
+				// via an object.
+				val = node[ i ] || node.getAttribute && node.getAttribute( i );
+				if ( val ) {
+					script.setAttribute( i, val );
+				}
+			}
+		}
+		doc.head.appendChild( script ).parentNode.removeChild( script );
+	}
+
+
+function toType( obj ) {
+	if ( obj == null ) {
+		return obj + "";
+	}
+
+	// Support: Android <=2.3 only (functionish RegExp)
+	return typeof obj === "object" || typeof obj === "function" ?
+		class2type[ toString.call( obj ) ] || "object" :
+		typeof obj;
+}
+/* global Symbol */
+// Defining this global in .eslintrc.json would create a danger of using the global
+// unguarded in another place, it seems safer to define global only for this module
+
+
+
+var
+	version = "3.6.0",
+
+	// Define a local copy of jQuery
+	jQuery = function( selector, context ) {
+
+		// The jQuery object is actually just the init constructor 'enhanced'
+		// Need init if jQuery is called (just allow error to be thrown if not included)
+		return new jQuery.fn.init( selector, context );
+	};
+
+jQuery.fn = jQuery.prototype = {
+
+	// The current version of jQuery being used
+	jquery: version,
+
+	constructor: jQuery,
+
+	// The default length of a jQuery object is 0
+	length: 0,
+
+	toArray: function() {
+		return slice.call( this );
+	},
+
+	// Get the Nth element in the matched element set OR
+	// Get the whole matched element set as a clean array
+	get: function( num ) {
+
+		// Return all the elements in a clean array
+		if ( num == null ) {
+			return slice.call( this );
+		}
+
+		// Return just the one element from the set
+		return num < 0 ? this[ num + this.length ] : this[ num ];
+	},
+
+	// Take an array of elements and push it onto the stack
+	// (returning the new matched element set)
+	pushStack: function( elems ) {
+
+		// Build a new jQuery matched element set
+		var ret = jQuery.merge( this.constructor(), elems );
+
+		// Add the old object onto the stack (as a reference)
+		ret.prevObject = this;
+
+		// Return the newly-formed element set
+		return ret;
+	},
+
+	// Execute a callback for every element in the matched set.
+	each: function( callback ) {
+		return jQuery.each( this, callback );
+	},
+
+	map: function( callback ) {
+		return this.pushStack( jQuery.map( this, function( elem, i ) {
+			return callback.call( elem, i, elem );
+		} ) );
+	},
+
+	slice: function() {
+		return this.pushStack( slice.apply( this, arguments ) );
+	},
+
+	first: function() {
+		return this.eq( 0 );
+	},
+
+	last: function() {
+		return this.eq( -1 );
+	},
+
+	even: function() {
+		return this.pushStack( jQuery.grep( this, function( _elem, i ) {
+			return ( i + 1 ) % 2;
+		} ) );
+	},
+
+	odd: function() {
+		return this.pushStack( jQuery.grep( this, function( _elem, i ) {
+			return i % 2;
+		} ) );
+	},
+
+	eq: function( i ) {
+		var len = this.length,
+			j = +i + ( i < 0 ? len : 0 );
+		return this.pushStack( j >= 0 && j < len ? [ this[ j ] ] : [] );
+	},
+
+	end: function() {
+		return this.prevObject || this.constructor();
+	},
+
+	// For internal use only.
+	// Behaves like an Array's method, not like a jQuery method.
+	push: push,
+	sort: arr.sort,
+	splice: arr.splice
+};
+
+jQuery.extend = jQuery.fn.extend = function() {
+	var options, name, src, copy, copyIsArray, clone,
+		target = arguments[ 0 ] || {},
+		i = 1,
+		length = arguments.length,
+		deep = false;
+
+	// Handle a deep copy situation
+	if ( typeof target === "boolean" ) {
+		deep = target;
+
+		// Skip the boolean and the target
+		target = arguments[ i ] || {};
+		i++;
+	}
+
+	// Handle case when target is a string or something (possible in deep copy)
+	if ( typeof target !== "object" && !isFunction( target ) ) {
+		target = {};
+	}
+
+	// Extend jQuery itself if only one argument is passed
+	if ( i === length ) {
+		target = this;
+		i--;
+	}
+
+	for ( ; i < length; i++ ) {
+
+		// Only deal with non-null/undefined values
+		if ( ( options = arguments[ i ] ) != null ) {
+
+			// Extend the base object
+			for ( name in options ) {
+				copy = options[ name ];
+
+				// Prevent Object.prototype pollution
+				// Prevent never-ending loop
+				if ( name === "__proto__" || target === copy ) {
+					continue;
+				}
+
+				// Recurse if we're merging plain objects or arrays
+				if ( deep && copy && ( jQuery.isPlainObject( copy ) ||
+					( copyIsArray = Array.isArray( copy ) ) ) ) {
+					src = target[ name ];
+
+					// Ensure proper type for the source value
+					if ( copyIsArray && !Array.isArray( src ) ) {
+						clone = [];
+					} else if ( !copyIsArray && !jQuery.isPlainObject( src ) ) {
+						clone = {};
+					} else {
+						clone = src;
+					}
+					copyIsArray = false;
+
+					// Never move original objects, clone them
+					target[ name ] = jQuery.extend( deep, clone, copy );
+
+				// Don't bring in undefined values
+				} else if ( copy !== undefined ) {
+					target[ name ] = copy;
+				}
+			}
+		}
+	}
+
+	// Return the modified object
+	return target;
+};
+
+jQuery.extend( {
+
+	// Unique for each copy of jQuery on the page
+	expando: "jQuery" + ( version + Math.random() ).replace( /\D/g, "" ),
+
+	// Assume jQuery is ready without the ready module
+	isReady: true,
+
+	error: function( msg ) {
+		throw new Error( msg );
+	},
+
+	noop: function() {},
+
+	isPlainObject: function( obj ) {
+		var proto, Ctor;
+
+		// Detect obvious negatives
+		// Use toString instead of jQuery.type to catch host objects
+		if ( !obj || toString.call( obj ) !== "[object Object]" ) {
+			return false;
+		}
+
+		proto = getProto( obj );
+
+		// Objects with no prototype (e.g., `Object.create( null )`) are plain
+		if ( !proto ) {
+			return true;
+		}
+
+		// Objects with prototype are plain iff they were constructed by a global Object function
+		Ctor = hasOwn.call( proto, "constructor" ) && proto.constructor;
+		return typeof Ctor === "function" && fnToString.call( Ctor ) === ObjectFunctionString;
+	},
+
+	isEmptyObject: function( obj ) {
+		var name;
+
+		for ( name in obj ) {
+			return false;
+		}
+		return true;
+	},
+
+	// Evaluates a script in a provided context; falls back to the global one
+	// if not specified.
+	globalEval: function( code, options, doc ) {
+		DOMEval( code, { nonce: options && options.nonce }, doc );
+	},
+
+	each: function( obj, callback ) {
+		var length, i = 0;
+
+		if ( isArrayLike( obj ) ) {
+			length = obj.length;
+			for ( ; i < length; i++ ) {
+				if ( callback.call( obj[ i ], i, obj[ i ] ) === false ) {
+					break;
+				}
+			}
+		} else {
+			for ( i in obj ) {
+				if ( callback.call( obj[ i ], i, obj[ i ] ) === false ) {
+					break;
+				}
+			}
+		}
+
+		return obj;
+	},
+
+	// results is for internal usage only
+	makeArray: function( arr, results ) {
+		var ret = results || [];
+
+		if ( arr != null ) {
+			if ( isArrayLike( Object( arr ) ) ) {
+				jQuery.merge( ret,
+					typeof arr === "string" ?
+						[ arr ] : arr
+				);
+			} else {
+				push.call( ret, arr );
+			}
+		}
+
+		return ret;
+	},
+
+	inArray: function( elem, arr, i ) {
+		return arr == null ? -1 : indexOf.call( arr, elem, i );
+	},
+
+	// Support: Android <=4.0 only, PhantomJS 1 only
+	// push.apply(_, arraylike) throws on ancient WebKit
+	merge: function( first, second ) {
+		var len = +second.length,
+			j = 0,
+			i = first.length;
+
+		for ( ; j < len; j++ ) {
+			first[ i++ ] = second[ j ];
+		}
+
+		first.length = i;
+
+		return first;
+	},
+
+	grep: function( elems, callback, invert ) {
+		var callbackInverse,
+			matches = [],
+			i = 0,
+			length = elems.length,
+			callbackExpect = !invert;
+
+		// Go through the array, only saving the items
+		// that pass the validator function
+		for ( ; i < length; i++ ) {
+			callbackInverse = !callback( elems[ i ], i );
+			if ( callbackInverse !== callbackExpect ) {
+				matches.push( elems[ i ] );
+			}
+		}
+
+		return matches;
+	},
+
+	// arg is for internal usage only
+	map: function( elems, callback, arg ) {
+		var length, value,
+			i = 0,
+			ret = [];
+
+		// Go through the array, translating each of the items to their new values
+		if ( isArrayLike( elems ) ) {
+			length = elems.length;
+			for ( ; i < length; i++ ) {
+				value = callback( elems[ i ], i, arg );
+
+				if ( value != null ) {
+					ret.push( value );
+				}
+			}
+
+		// Go through every key on the object,
+		} else {
+			for ( i in elems ) {
+				value = callback( elems[ i ], i, arg );
+
+				if ( value != null ) {
+					ret.push( value );
+				}
+			}
+		}
+
+		// Flatten any nested arrays
+		return flat( ret );
+	},
+
+	// A global GUID counter for objects
+	guid: 1,
+
+	// jQuery.support is not used in Core but other projects attach their
+	// properties to it so it needs to exist.
+	support: support
+} );
+
+if ( typeof Symbol === "function" ) {
+	jQuery.fn[ Symbol.iterator ] = arr[ Symbol.iterator ];
+}
+
+// Populate the class2type map
+jQuery.each( "Boolean Number String Function Array Date RegExp Object Error Symbol".split( " " ),
+	function( _i, name ) {
+		class2type[ "[object " + name + "]" ] = name.toLowerCase();
+	} );
+
+function isArrayLike( obj ) {
+
+	// Support: real iOS 8.2 only (not reproducible in simulator)
+	// `in` check used to prevent JIT error (gh-2145)
+	// hasOwn isn't used here due to false negatives
+	// regarding Nodelist length in IE
+	var length = !!obj && "length" in obj && obj.length,
+		type = toType( obj );
+
+	if ( isFunction( obj ) || isWindow( obj ) ) {
+		return false;
+	}
+
+	return type === "array" || length === 0 ||
+		typeof length === "number" && length > 0 && ( length - 1 ) in obj;
+}
+var Sizzle =
+/*!
+ * Sizzle CSS Selector Engine v2.3.6
+ * https://sizzlejs.com/
+ *
+ * Copyright JS Foundation and other contributors
+ * Released under the MIT license
+ * https://js.foundation/
+ *
+ * Date: 2021-02-16
+ */
+( function( window ) {
+var i,
+	support,
+	Expr,
+	getText,
+	isXML,
+	tokenize,
+	compile,
+	select,
+	outermostContext,
+	sortInput,
+	hasDuplicate,
+
+	// Local document vars
+	setDocument,
+	document,
+	docElem,
+	documentIsHTML,
+	rbuggyQSA,
+	rbuggyMatches,
+	matches,
+	contains,
+
+	// Instance-specific data
+	expando = "sizzle" + 1 * new Date(),
+	preferredDoc = window.document,
+	dirruns = 0,
+	done = 0,
+	classCache = createCache(),
+	tokenCache = createCache(),
+	compilerCache = createCache(),
+	nonnativeSelectorCache = createCache(),
+	sortOrder = function( a, b ) {
+		if ( a === b ) {
+			hasDuplicate = true;
+		}
+		return 0;
+	},
+
+	// Instance methods
+	hasOwn = ( {} ).hasOwnProperty,
+	arr = [],
+	pop = arr.pop,
+	pushNative = arr.push,
+	push = arr.push,
+	slice = arr.slice,
+
+	// Use a stripped-down indexOf as it's faster than native
+	// https://jsperf.com/thor-indexof-vs-for/5
+	indexOf = function( list, elem ) {
+		var i = 0,
+			len = list.length;
+		for ( ; i < len; i++ ) {
+			if ( list[ i ] === elem ) {
+				return i;
+			}
+		}
+		return -1;
+	},
+
+	booleans = "checked|selected|async|autofocus|autoplay|controls|defer|disabled|hidden|" +
+		"ismap|loop|multiple|open|readonly|required|scoped",
+
+	// Regular expressions
+
+	// http://www.w3.org/TR/css3-selectors/#whitespace
+	whitespace = "[\\x20\\t\\r\\n\\f]",
+
+	// https://www.w3.org/TR/css-syntax-3/#ident-token-diagram
+	identifier = "(?:\\\\[\\da-fA-F]{1,6}" + whitespace +
+		"?|\\\\[^\\r\\n\\f]|[\\w-]|[^\0-\\x7f])+",
+
+	// Attribute selectors: http://www.w3.org/TR/selectors/#attribute-selectors
+	attributes = "\\[" + whitespace + "*(" + identifier + ")(?:" + whitespace +
+
+		// Operator (capture 2)
+		"*([*^$|!~]?=)" + whitespace +
+
+		// "Attribute values must be CSS identifiers [capture 5]
+		// or strings [capture 3 or capture 4]"
+		"*(?:'((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\"|(" + identifier + "))|)" +
+		whitespace + "*\\]",
+
+	pseudos = ":(" + identifier + ")(?:\\((" +
+
+		// To reduce the number of selectors needing tokenize in the preFilter, prefer arguments:
+		// 1. quoted (capture 3; capture 4 or capture 5)
+		"('((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\")|" +
+
+		// 2. simple (capture 6)
+		"((?:\\\\.|[^\\\\()[\\]]|" + attributes + ")*)|" +
+
+		// 3. anything else (capture 2)
+		".*" +
+		")\\)|)",
+
+	// Leading and non-escaped trailing whitespace, capturing some non-whitespace characters preceding the latter
+	rwhitespace = new RegExp( whitespace + "+", "g" ),
+	rtrim = new RegExp( "^" + whitespace + "+|((?:^|[^\\\\])(?:\\\\.)*)" +
+		whitespace + "+$", "g" ),
+
+	rcomma = new RegExp( "^" + whitespace + "*," + whitespace + "*" ),
+	rcombinators = new RegExp( "^" + whitespace + "*([>+~]|" + whitespace + ")" + whitespace +
+		"*" ),
+	rdescend = new RegExp( whitespace + "|>" ),
+
+	rpseudo = new RegExp( pseudos ),
+	ridentifier = new RegExp( "^" + identifier + "$" ),
+
+	matchExpr = {
+		"ID": new RegExp( "^#(" + identifier + ")" ),
+		"CLASS": new RegExp( "^\\.(" + identifier + ")" ),
+		"TAG": new RegExp( "^(" + identifier + "|[*])" ),
+		"ATTR": new RegExp( "^" + attributes ),
+		"PSEUDO": new RegExp( "^" + pseudos ),
+		"CHILD": new RegExp( "^:(only|first|last|nth|nth-last)-(child|of-type)(?:\\(" +
+			whitespace + "*(even|odd|(([+-]|)(\\d*)n|)" + whitespace + "*(?:([+-]|)" +
+			whitespace + "*(\\d+)|))" + whitespace + "*\\)|)", "i" ),
+		"bool": new RegExp( "^(?:" + booleans + ")$", "i" ),
+
+		// For use in libraries implementing .is()
+		// We use this for POS matching in `select`
+		"needsContext": new RegExp( "^" + whitespace +
+			"*[>+~]|:(even|odd|eq|gt|lt|nth|first|last)(?:\\(" + whitespace +
+			"*((?:-\\d)?\\d*)" + whitespace + "*\\)|)(?=[^-]|$)", "i" )
+	},
+
+	rhtml = /HTML$/i,
+	rinputs = /^(?:input|select|textarea|button)$/i,
+	rheader = /^h\d$/i,
+
+	rnative = /^[^{]+\{\s*\[native \w/,
+
+	// Easily-parseable/retrievable ID or TAG or CLASS selectors
+	rquickExpr = /^(?:#([\w-]+)|(\w+)|\.([\w-]+))$/,
+
+	rsibling = /[+~]/,
+
+	// CSS escapes
+	// http://www.w3.org/TR/CSS21/syndata.html#escaped-characters
+	runescape = new RegExp( "\\\\[\\da-fA-F]{1,6}" + whitespace + "?|\\\\([^\\r\\n\\f])", "g" ),
+	funescape = function( escape, nonHex ) {
+		var high = "0x" + escape.slice( 1 ) - 0x10000;
+
+		return nonHex ?
+
+			// Strip the backslash prefix from a non-hex escape sequence
+			nonHex :
+
+			// Replace a hexadecimal escape sequence with the encoded Unicode code point
+			// Support: IE <=11+
+			// For values outside the Basic Multilingual Plane (BMP), manually construct a
+			// surrogate pair
+			high < 0 ?
+				String.fromCharCode( high + 0x10000 ) :
+				String.fromCharCode( high >> 10 | 0xD800, high & 0x3FF | 0xDC00 );
+	},
+
+	// CSS string/identifier serialization
+	// https://drafts.csswg.org/cssom/#common-serializing-idioms
+	rcssescape = /([\0-\x1f\x7f]|^-?\d)|^-$|[^\0-\x1f\x7f-\uFFFF\w-]/g,
+	fcssescape = function( ch, asCodePoint ) {
+		if ( asCodePoint ) {
+
+			// U+0000 NULL becomes U+FFFD REPLACEMENT CHARACTER
+			if ( ch === "\0" ) {
+				return "\uFFFD";
+			}
+
+			// Control characters and (dependent upon position) numbers get escaped as code points
+			return ch.slice( 0, -1 ) + "\\" +
+				ch.charCodeAt( ch.length - 1 ).toString( 16 ) + " ";
+		}
+
+		// Other potentially-special ASCII characters get backslash-escaped
+		return "\\" + ch;
+	},
+
+	// Used for iframes
+	// See setDocument()
+	// Removing the function wrapper causes a "Permission Denied"
+	// error in IE
+	unloadHandler = function() {
+		setDocument();
+	},
+
+	inDisabledFieldset = addCombinator(
+		function( elem ) {
+			return elem.disabled === true && elem.nodeName.toLowerCase() === "fieldset";
+		},
+		{ dir: "parentNode", next: "legend" }
+	);
+
+// Optimize for push.apply( _, NodeList )
+try {
+	push.apply(
+		( arr = slice.call( preferredDoc.childNodes ) ),
+		preferredDoc.childNodes
+	);
+
+	// Support: Android<4.0
+	// Detect silently failing push.apply
+	// eslint-disable-next-line no-unused-expressions
+	arr[ preferredDoc.childNodes.length ].nodeType;
+} catch ( e ) {
+	push = { apply: arr.length ?
+
+		// Leverage slice if possible
+		function( target, els ) {
+			pushNative.apply( target, slice.call( els ) );
+		} :
+
+		// Support: IE<9
+		// Otherwise append directly
+		function( target, els ) {
+			var j = target.length,
+				i = 0;
+
+			// Can't trust NodeList.length
+			while ( ( target[ j++ ] = els[ i++ ] ) ) {}
+			target.length = j - 1;
+		}
+	};
+}
+
+function Sizzle( selector, context, results, seed ) {
+	var m, i, elem, nid, match, groups, newSelector,
+		newContext = context && context.ownerDocument,
+
+		// nodeType defaults to 9, since context defaults to document
+		nodeType = context ? context.nodeType : 9;
+
+	results = results || [];
+
+	// Return early from calls with invalid selector or context
+	if ( typeof selector !== "string" || !selector ||
+		nodeType !== 1 && nodeType !== 9 && nodeType !== 11 ) {
+
+		return results;
+	}
+
+	// Try to shortcut find operations (as opposed to filters) in HTML documents
+	if ( !seed ) {
+		setDocument( context );
+		context = context || document;
+
+		if ( documentIsHTML ) {
+
+			// If the selector is sufficiently simple, try using a "get*By*" DOM method
+			// (excepting DocumentFragment context, where the methods don't exist)
+			if ( nodeType !== 11 && ( match = rquickExpr.exec( selector ) ) ) {
+
+				// ID selector
+				if ( ( m = match[ 1 ] ) ) {
+
+					// Document context
+					if ( nodeType === 9 ) {
+						if ( ( elem = context.getElementById( m ) ) ) {
+
+							// Support: IE, Opera, Webkit
+							// TODO: identify versions
+							// getElementById can match elements by name instead of ID
+							if ( elem.id === m ) {
+								results.push( elem );
+								return results;
+							}
+						} else {
+							return results;
+						}
+
+					// Element context
+					} else {
+
+						// Support: IE, Opera, Webkit
+						// TODO: identify versions
+						// getElementById can match elements by name instead of ID
+						if ( newContext && ( elem = newContext.getElementById( m ) ) &&
+							contains( context, elem ) &&
+							elem.id === m ) {
+
+							results.push( elem );
+							return results;
+						}
+					}
+
+				// Type selector
+				} else if ( match[ 2 ] ) {
+					push.apply( results, context.getElementsByTagName( selector ) );
+					return results;
+
+				// Class selector
+				} else if ( ( m = match[ 3 ] ) && support.getElementsByClassName &&
+					context.getElementsByClassName ) {
+
+					push.apply( results, context.getElementsByClassName( m ) );
+					return results;
+				}
+			}
+
+			// Take advantage of querySelectorAll
+			if ( support.qsa &&
+				!nonnativeSelectorCache[ selector + " " ] &&
+				( !rbuggyQSA || !rbuggyQSA.test( selector ) ) &&
+
+				// Support: IE 8 only
+				// Exclude object elements
+				( nodeType !== 1 || context.nodeName.toLowerCase() !== "object" ) ) {
+
+				newSelector = selector;
+				newContext = context;
+
+				// qSA considers elements outside a scoping root when evaluating child or
+				// descendant combinators, which is not what we want.
+				// In such cases, we work around the behavior by prefixing every selector in the
+				// list with an ID selector referencing the scope context.
+				// The technique has to be used as well when a leading combinator is used
+				// as such selectors are not recognized by querySelectorAll.
+				// Thanks to Andrew Dupont for this technique.
+				if ( nodeType === 1 &&
+					( rdescend.test( selector ) || rcombinators.test( selector ) ) ) {
+
+					// Expand context for sibling selectors
+					newContext = rsibling.test( selector ) && testContext( context.parentNode ) ||
+						context;
+
+					// We can use :scope instead of the ID hack if the browser
+					// supports it & if we're not changing the context.
+					if ( newContext !== context || !support.scope ) {
+
+						// Capture the context ID, setting it first if necessary
+						if ( ( nid = context.getAttribute( "id" ) ) ) {
+							nid = nid.replace( rcssescape, fcssescape );
+						} else {
+							context.setAttribute( "id", ( nid = expando ) );
+						}
+					}
+
+					// Prefix every selector in the list
+					groups = tokenize( selector );
+					i = groups.length;
+					while ( i-- ) {
+						groups[ i ] = ( nid ? "#" + nid : ":scope" ) + " " +
+							toSelector( groups[ i ] );
+					}
+					newSelector = groups.join( "," );
+				}
+
+				try {
+					push.apply( results,
+						newContext.querySelectorAll( newSelector )
+					);
+					return results;
+				} catch ( qsaError ) {
+					nonnativeSelectorCache( selector, true );
+				} finally {
+					if ( nid === expando ) {
+						context.removeAttribute( "id" );
+					}
+				}
+			}
+		}
+	}
+
+	// All others
+	return select( selector.replace( rtrim, "$1" ), context, results, seed );
+}
+
+/**
+ * Create key-value caches of limited size
+ * @returns {function(string, object)} Returns the Object data after storing it on itself with
+ *	property name the (space-suffixed) string and (if the cache is larger than Expr.cacheLength)
+ *	deleting the oldest entry
+ */
+function createCache() {
+	var keys = [];
+
+	function cache( key, value ) {
+
+		// Use (key + " ") to avoid collision with native prototype properties (see Issue #157)
+		if ( keys.push( key + " " ) > Expr.cacheLength ) {
+
+			// Only keep the most recent entries
+			delete cache[ keys.shift() ];
+		}
+		return ( cache[ key + " " ] = value );
+	}
+	return cache;
+}
+
+/**
+ * Mark a function for special use by Sizzle
+ * @param {Function} fn The function to mark
+ */
+function markFunction( fn ) {
+	fn[ expando ] = true;
+	return fn;
+}
+
+/**
+ * Support testing using an element
+ * @param {Function} fn Passed the created element and returns a boolean result
+ */
+function assert( fn ) {
+	var el = document.createElement( "fieldset" );
+
+	try {
+		return !!fn( el );
+	} catch ( e ) {
+		return false;
+	} finally {
+
+		// Remove from its parent by default
+		if ( el.parentNode ) {
+			el.parentNode.removeChild( el );
+		}
+
+		// release memory in IE
+		el = null;
+	}
+}
+
+/**
+ * Adds the same handler for all of the specified attrs
+ * @param {String} attrs Pipe-separated list of attributes
+ * @param {Function} handler The method that will be applied
+ */
+function addHandle( attrs, handler ) {
+	var arr = attrs.split( "|" ),
+		i = arr.length;
+
+	while ( i-- ) {
+		Expr.attrHandle[ arr[ i ] ] = handler;
+	}
+}
+
+/**
+ * Checks document order of two siblings
+ * @param {Element} a
+ * @param {Element} b
+ * @returns {Number} Returns less than 0 if a precedes b, greater than 0 if a follows b
+ */
+function siblingCheck( a, b ) {
+	var cur = b && a,
+		diff = cur && a.nodeType === 1 && b.nodeType === 1 &&
+			a.sourceIndex - b.sourceIndex;
+
+	// Use IE sourceIndex if available on both nodes
+	if ( diff ) {
+		return diff;
+	}
+
+	// Check if b follows a
+	if ( cur ) {
+		while ( ( cur = cur.nextSibling ) ) {
+			if ( cur === b ) {
+				return -1;
+			}
+		}
+	}
+
+	return a ? 1 : -1;
+}
+
+/**
+ * Returns a function to use in pseudos for input types
+ * @param {String} type
+ */
+function createInputPseudo( type ) {
+	return function( elem ) {
+		var name = elem.nodeName.toLowerCase();
+		return name === "input" && elem.type === type;
+	};
+}
+
+/**
+ * Returns a function to use in pseudos for buttons
+ * @param {String} type
+ */
+function createButtonPseudo( type ) {
+	return function( elem ) {
+		var name = elem.nodeName.toLowerCase();
+		return ( name === "input" || name === "button" ) && elem.type === type;
+	};
+}
+
+/**
+ * Returns a function to use in pseudos for :enabled/:disabled
+ * @param {Boolean} disabled true for :disabled; false for :enabled
+ */
+function createDisabledPseudo( disabled ) {
+
+	// Known :disabled false positives: fieldset[disabled] > legend:nth-of-type(n+2) :can-disable
+	return function( elem ) {
+
+		// Only certain elements can match :enabled or :disabled
+		// https://html.spec.whatwg.org/multipage/scripting.html#selector-enabled
+		// https://html.spec.whatwg.org/multipage/scripting.html#selector-disabled
+		if ( "form" in elem ) {
+
+			// Check for inherited disabledness on relevant non-disabled elements:
+			// * listed form-associated elements in a disabled fieldset
+			//   https://html.spec.whatwg.org/multipage/forms.html#category-listed
+			//   https://html.spec.whatwg.org/multipage/forms.html#concept-fe-disabled
+			// * option elements in a disabled optgroup
+			//   https://html.spec.whatwg.org/multipage/forms.html#concept-option-disabled
+			// All such elements have a "form" property.
+			if ( elem.parentNode && elem.disabled === false ) {
+
+				// Option elements defer to a parent optgroup if present
+				if ( "label" in elem ) {
+					if ( "label" in elem.parentNode ) {
+						return elem.parentNode.disabled === disabled;
+					} else {
+						return elem.disabled === disabled;
+					}
+				}
+
+				// Support: IE 6 - 11
+				// Use the isDisabled shortcut property to check for disabled fieldset ancestors
+				return elem.isDisabled === disabled ||
+
+					// Where there is no isDisabled, check manually
+					/* jshint -W018 */
+					elem.isDisabled !== !disabled &&
+					inDisabledFieldset( elem ) === disabled;
+			}
+
+			return elem.disabled === disabled;
+
+		// Try to winnow out elements that can't be disabled before trusting the disabled property.
+		// Some victims get caught in our net (label, legend, menu, track), but it shouldn't
+		// even exist on them, let alone have a boolean value.
+		} else if ( "label" in elem ) {
+			return elem.disabled === disabled;
+		}
+
+		// Remaining elements are neither :enabled nor :disabled
+		return false;
+	};
+}
+
+/**
+ * Returns a function to use in pseudos for positionals
+ * @param {Function} fn
+ */
+function createPositionalPseudo( fn ) {
+	return markFunction( function( argument ) {
+		argument = +argument;
+		return markFunction( function( seed, matches ) {
+			var j,
+				matchIndexes = fn( [], seed.length, argument ),
+				i = matchIndexes.length;
+
+			// Match elements found at the specified indexes
+			while ( i-- ) {
+				if ( seed[ ( j = matchIndexes[ i ] ) ] ) {
+					seed[ j ] = !( matches[ j ] = seed[ j ] );
+				}
+			}
+		} );
+	} );
+}
+
+/**
+ * Checks a node for validity as a Sizzle context
+ * @param {Element|Object=} context
+ * @returns {Element|Object|Boolean} The input node if acceptable, otherwise a falsy value
+ */
+function testContext( context ) {
+	return context && typeof context.getElementsByTagName !== "undefined" && context;
+}
+
+// Expose support vars for convenience
+support = Sizzle.support = {};
+
+/**
+ * Detects XML nodes
+ * @param {Element|Object} elem An element or a document
+ * @returns {Boolean} True iff elem is a non-HTML XML node
+ */
+isXML = Sizzle.isXML = function( elem ) {
+	var namespace = elem && elem.namespaceURI,
+		docElem = elem && ( elem.ownerDocument || elem ).documentElement;
+
+	// Support: IE <=8
+	// Assume HTML when documentElement doesn't yet exist, such as inside loading iframes
+	// https://bugs.jquery.com/ticket/4833
+	return !rhtml.test( namespace || docElem && docElem.nodeName || "HTML" );
+};
+
+/**
+ * Sets document-related variables once based on the current document
+ * @param {Element|Object} [doc] An element or document object to use to set the document
+ * @returns {Object} Returns the current document
+ */
+setDocument = Sizzle.setDocument = function( node ) {
+	var hasCompare, subWindow,
+		doc = node ? node.ownerDocument || node : preferredDoc;
+
+	// Return early if doc is invalid or already selected
+	// Support: IE 11+, Edge 17 - 18+
+	// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+	// two documents; shallow comparisons work.
+	// eslint-disable-next-line eqeqeq
+	if ( doc == document || doc.nodeType !== 9 || !doc.documentElement ) {
+		return document;
+	}
+
+	// Update global variables
+	document = doc;
+	docElem = document.documentElement;
+	documentIsHTML = !isXML( document );
+
+	// Support: IE 9 - 11+, Edge 12 - 18+
+	// Accessing iframe documents after unload throws "permission denied" errors (jQuery #13936)
+	// Support: IE 11+, Edge 17 - 18+
+	// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+	// two documents; shallow comparisons work.
+	// eslint-disable-next-line eqeqeq
+	if ( preferredDoc != document &&
+		( subWindow = document.defaultView ) && subWindow.top !== subWindow ) {
+
+		// Support: IE 11, Edge
+		if ( subWindow.addEventListener ) {
+			subWindow.addEventListener( "unload", unloadHandler, false );
+
+		// Support: IE 9 - 10 only
+		} else if ( subWindow.attachEvent ) {
+			subWindow.attachEvent( "onunload", unloadHandler );
+		}
+	}
+
+	// Support: IE 8 - 11+, Edge 12 - 18+, Chrome <=16 - 25 only, Firefox <=3.6 - 31 only,
+	// Safari 4 - 5 only, Opera <=11.6 - 12.x only
+	// IE/Edge & older browsers don't support the :scope pseudo-class.
+	// Support: Safari 6.0 only
+	// Safari 6.0 supports :scope but it's an alias of :root there.
+	support.scope = assert( function( el ) {
+		docElem.appendChild( el ).appendChild( document.createElement( "div" ) );
+		return typeof el.querySelectorAll !== "undefined" &&
+			!el.querySelectorAll( ":scope fieldset div" ).length;
+	} );
+
+	/* Attributes
+	---------------------------------------------------------------------- */
+
+	// Support: IE<8
+	// Verify that getAttribute really returns attributes and not properties
+	// (excepting IE8 booleans)
+	support.attributes = assert( function( el ) {
+		el.className = "i";
+		return !el.getAttribute( "className" );
+	} );
+
+	/* getElement(s)By*
+	---------------------------------------------------------------------- */
+
+	// Check if getElementsByTagName("*") returns only elements
+	support.getElementsByTagName = assert( function( el ) {
+		el.appendChild( document.createComment( "" ) );
+		return !el.getElementsByTagName( "*" ).length;
+	} );
+
+	// Support: IE<9
+	support.getElementsByClassName = rnative.test( document.getElementsByClassName );
+
+	// Support: IE<10
+	// Check if getElementById returns elements by name
+	// The broken getElementById methods don't pick up programmatically-set names,
+	// so use a roundabout getElementsByName test
+	support.getById = assert( function( el ) {
+		docElem.appendChild( el ).id = expando;
+		return !document.getElementsByName || !document.getElementsByName( expando ).length;
+	} );
+
+	// ID filter and find
+	if ( support.getById ) {
+		Expr.filter[ "ID" ] = function( id ) {
+			var attrId = id.replace( runescape, funescape );
+			return function( elem ) {
+				return elem.getAttribute( "id" ) === attrId;
+			};
+		};
+		Expr.find[ "ID" ] = function( id, context ) {
+			if ( typeof context.getElementById !== "undefined" && documentIsHTML ) {
+				var elem = context.getElementById( id );
+				return elem ? [ elem ] : [];
+			}
+		};
+	} else {
+		Expr.filter[ "ID" ] =  function( id ) {
+			var attrId = id.replace( runescape, funescape );
+			return function( elem ) {
+				var node = typeof elem.getAttributeNode !== "undefined" &&
+					elem.getAttributeNode( "id" );
+				return node && node.value === attrId;
+			};
+		};
+
+		// Support: IE 6 - 7 only
+		// getElementById is not reliable as a find shortcut
+		Expr.find[ "ID" ] = function( id, context ) {
+			if ( typeof context.getElementById !== "undefined" && documentIsHTML ) {
+				var node, i, elems,
+					elem = context.getElementById( id );
+
+				if ( elem ) {
+
+					// Verify the id attribute
+					node = elem.getAttributeNode( "id" );
+					if ( node && node.value === id ) {
+						return [ elem ];
+					}
+
+					// Fall back on getElementsByName
+					elems = context.getElementsByName( id );
+					i = 0;
+					while ( ( elem = elems[ i++ ] ) ) {
+						node = elem.getAttributeNode( "id" );
+						if ( node && node.value === id ) {
+							return [ elem ];
+						}
+					}
+				}
+
+				return [];
+			}
+		};
+	}
+
+	// Tag
+	Expr.find[ "TAG" ] = support.getElementsByTagName ?
+		function( tag, context ) {
+			if ( typeof context.getElementsByTagName !== "undefined" ) {
+				return context.getElementsByTagName( tag );
+
+			// DocumentFragment nodes don't have gEBTN
+			} else if ( support.qsa ) {
+				return context.querySelectorAll( tag );
+			}
+		} :
+
+		function( tag, context ) {
+			var elem,
+				tmp = [],
+				i = 0,
+
+				// By happy coincidence, a (broken) gEBTN appears on DocumentFragment nodes too
+				results = context.getElementsByTagName( tag );
+
+			// Filter out possible comments
+			if ( tag === "*" ) {
+				while ( ( elem = results[ i++ ] ) ) {
+					if ( elem.nodeType === 1 ) {
+						tmp.push( elem );
+					}
+				}
+
+				return tmp;
+			}
+			return results;
+		};
+
+	// Class
+	Expr.find[ "CLASS" ] = support.getElementsByClassName && function( className, context ) {
+		if ( typeof context.getElementsByClassName !== "undefined" && documentIsHTML ) {
+			return context.getElementsByClassName( className );
+		}
+	};
+
+	/* QSA/matchesSelector
+	---------------------------------------------------------------------- */
+
+	// QSA and matchesSelector support
+
+	// matchesSelector(:active) reports false when true (IE9/Opera 11.5)
+	rbuggyMatches = [];
+
+	// qSa(:focus) reports false when true (Chrome 21)
+	// We allow this because of a bug in IE8/9 that throws an error
+	// whenever `document.activeElement` is accessed on an iframe
+	// So, we allow :focus to pass through QSA all the time to avoid the IE error
+	// See https://bugs.jquery.com/ticket/13378
+	rbuggyQSA = [];
+
+	if ( ( support.qsa = rnative.test( document.querySelectorAll ) ) ) {
+
+		// Build QSA regex
+		// Regex strategy adopted from Diego Perini
+		assert( function( el ) {
+
+			var input;
+
+			// Select is set to empty string on purpose
+			// This is to test IE's treatment of not explicitly
+			// setting a boolean content attribute,
+			// since its presence should be enough
+			// https://bugs.jquery.com/ticket/12359
+			docElem.appendChild( el ).innerHTML = "<a id='" + expando + "'></a>" +
+				"<select id='" + expando + "-\r\\' msallowcapture=''>" +
+				"<option selected=''></option></select>";
+
+			// Support: IE8, Opera 11-12.16
+			// Nothing should be selected when empty strings follow ^= or $= or *=
+			// The test attribute must be unknown in Opera but "safe" for WinRT
+			// https://msdn.microsoft.com/en-us/library/ie/hh465388.aspx#attribute_section
+			if ( el.querySelectorAll( "[msallowcapture^='']" ).length ) {
+				rbuggyQSA.push( "[*^$]=" + whitespace + "*(?:''|\"\")" );
+			}
+
+			// Support: IE8
+			// Boolean attributes and "value" are not treated correctly
+			if ( !el.querySelectorAll( "[selected]" ).length ) {
+				rbuggyQSA.push( "\\[" + whitespace + "*(?:value|" + booleans + ")" );
+			}
+
+			// Support: Chrome<29, Android<4.4, Safari<7.0+, iOS<7.0+, PhantomJS<1.9.8+
+			if ( !el.querySelectorAll( "[id~=" + expando + "-]" ).length ) {
+				rbuggyQSA.push( "~=" );
+			}
+
+			// Support: IE 11+, Edge 15 - 18+
+			// IE 11/Edge don't find elements on a `[name='']` query in some cases.
+			// Adding a temporary attribute to the document before the selection works
+			// around the issue.
+			// Interestingly, IE 10 & older don't seem to have the issue.
+			input = document.createElement( "input" );
+			input.setAttribute( "name", "" );
+			el.appendChild( input );
+			if ( !el.querySelectorAll( "[name='']" ).length ) {
+				rbuggyQSA.push( "\\[" + whitespace + "*name" + whitespace + "*=" +
+					whitespace + "*(?:''|\"\")" );
+			}
+
+			// Webkit/Opera - :checked should return selected option elements
+			// http://www.w3.org/TR/2011/REC-css3-selectors-20110929/#checked
+			// IE8 throws error here and will not see later tests
+			if ( !el.querySelectorAll( ":checked" ).length ) {
+				rbuggyQSA.push( ":checked" );
+			}
+
+			// Support: Safari 8+, iOS 8+
+			// https://bugs.webkit.org/show_bug.cgi?id=136851
+			// In-page `selector#id sibling-combinator selector` fails
+			if ( !el.querySelectorAll( "a#" + expando + "+*" ).length ) {
+				rbuggyQSA.push( ".#.+[+~]" );
+			}
+
+			// Support: Firefox <=3.6 - 5 only
+			// Old Firefox doesn't throw on a badly-escaped identifier.
+			el.querySelectorAll( "\\\f" );
+			rbuggyQSA.push( "[\\r\\n\\f]" );
+		} );
+
+		assert( function( el ) {
+			el.innerHTML = "<a href='' disabled='disabled'></a>" +
+				"<select disabled='disabled'><option/></select>";
+
+			// Support: Windows 8 Native Apps
+			// The type and name attributes are restricted during .innerHTML assignment
+			var input = document.createElement( "input" );
+			input.setAttribute( "type", "hidden" );
+			el.appendChild( input ).setAttribute( "name", "D" );
+
+			// Support: IE8
+			// Enforce case-sensitivity of name attribute
+			if ( el.querySelectorAll( "[name=d]" ).length ) {
+				rbuggyQSA.push( "name" + whitespace + "*[*^$|!~]?=" );
+			}
+
+			// FF 3.5 - :enabled/:disabled and hidden elements (hidden elements are still enabled)
+			// IE8 throws error here and will not see later tests
+			if ( el.querySelectorAll( ":enabled" ).length !== 2 ) {
+				rbuggyQSA.push( ":enabled", ":disabled" );
+			}
+
+			// Support: IE9-11+
+			// IE's :disabled selector does not pick up the children of disabled fieldsets
+			docElem.appendChild( el ).disabled = true;
+			if ( el.querySelectorAll( ":disabled" ).length !== 2 ) {
+				rbuggyQSA.push( ":enabled", ":disabled" );
+			}
+
+			// Support: Opera 10 - 11 only
+			// Opera 10-11 does not throw on post-comma invalid pseudos
+			el.querySelectorAll( "*,:x" );
+			rbuggyQSA.push( ",.*:" );
+		} );
+	}
+
+	if ( ( support.matchesSelector = rnative.test( ( matches = docElem.matches ||
+		docElem.webkitMatchesSelector ||
+		docElem.mozMatchesSelector ||
+		docElem.oMatchesSelector ||
+		docElem.msMatchesSelector ) ) ) ) {
+
+		assert( function( el ) {
+
+			// Check to see if it's possible to do matchesSelector
+			// on a disconnected node (IE 9)
+			support.disconnectedMatch = matches.call( el, "*" );
+
+			// This should fail with an exception
+			// Gecko does not error, returns false instead
+			matches.call( el, "[s!='']:x" );
+			rbuggyMatches.push( "!=", pseudos );
+		} );
+	}
+
+	rbuggyQSA = rbuggyQSA.length && new RegExp( rbuggyQSA.join( "|" ) );
+	rbuggyMatches = rbuggyMatches.length && new RegExp( rbuggyMatches.join( "|" ) );
+
+	/* Contains
+	---------------------------------------------------------------------- */
+	hasCompare = rnative.test( docElem.compareDocumentPosition );
+
+	// Element contains another
+	// Purposefully self-exclusive
+	// As in, an element does not contain itself
+	contains = hasCompare || rnative.test( docElem.contains ) ?
+		function( a, b ) {
+			var adown = a.nodeType === 9 ? a.documentElement : a,
+				bup = b && b.parentNode;
+			return a === bup || !!( bup && bup.nodeType === 1 && (
+				adown.contains ?
+					adown.contains( bup ) :
+					a.compareDocumentPosition && a.compareDocumentPosition( bup ) & 16
+			) );
+		} :
+		function( a, b ) {
+			if ( b ) {
+				while ( ( b = b.parentNode ) ) {
+					if ( b === a ) {
+						return true;
+					}
+				}
+			}
+			return false;
+		};
+
+	/* Sorting
+	---------------------------------------------------------------------- */
+
+	// Document order sorting
+	sortOrder = hasCompare ?
+	function( a, b ) {
+
+		// Flag for duplicate removal
+		if ( a === b ) {
+			hasDuplicate = true;
+			return 0;
+		}
+
+		// Sort on method existence if only one input has compareDocumentPosition
+		var compare = !a.compareDocumentPosition - !b.compareDocumentPosition;
+		if ( compare ) {
+			return compare;
+		}
+
+		// Calculate position if both inputs belong to the same document
+		// Support: IE 11+, Edge 17 - 18+
+		// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+		// two documents; shallow comparisons work.
+		// eslint-disable-next-line eqeqeq
+		compare = ( a.ownerDocument || a ) == ( b.ownerDocument || b ) ?
+			a.compareDocumentPosition( b ) :
+
+			// Otherwise we know they are disconnected
+			1;
+
+		// Disconnected nodes
+		if ( compare & 1 ||
+			( !support.sortDetached && b.compareDocumentPosition( a ) === compare ) ) {
+
+			// Choose the first element that is related to our preferred document
+			// Support: IE 11+, Edge 17 - 18+
+			// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+			// two documents; shallow comparisons work.
+			// eslint-disable-next-line eqeqeq
+			if ( a == document || a.ownerDocument == preferredDoc &&
+				contains( preferredDoc, a ) ) {
+				return -1;
+			}
+
+			// Support: IE 11+, Edge 17 - 18+
+			// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+			// two documents; shallow comparisons work.
+			// eslint-disable-next-line eqeqeq
+			if ( b == document || b.ownerDocument == preferredDoc &&
+				contains( preferredDoc, b ) ) {
+				return 1;
+			}
+
+			// Maintain original order
+			return sortInput ?
+				( indexOf( sortInput, a ) - indexOf( sortInput, b ) ) :
+				0;
+		}
+
+		return compare & 4 ? -1 : 1;
+	} :
+	function( a, b ) {
+
+		// Exit early if the nodes are identical
+		if ( a === b ) {
+			hasDuplicate = true;
+			return 0;
+		}
+
+		var cur,
+			i = 0,
+			aup = a.parentNode,
+			bup = b.parentNode,
+			ap = [ a ],
+			bp = [ b ];
+
+		// Parentless nodes are either documents or disconnected
+		if ( !aup || !bup ) {
+
+			// Support: IE 11+, Edge 17 - 18+
+			// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+			// two documents; shallow comparisons work.
+			/* eslint-disable eqeqeq */
+			return a == document ? -1 :
+				b == document ? 1 :
+				/* eslint-enable eqeqeq */
+				aup ? -1 :
+				bup ? 1 :
+				sortInput ?
+				( indexOf( sortInput, a ) - indexOf( sortInput, b ) ) :
+				0;
+
+		// If the nodes are siblings, we can do a quick check
+		} else if ( aup === bup ) {
+			return siblingCheck( a, b );
+		}
+
+		// Otherwise we need full lists of their ancestors for comparison
+		cur = a;
+		while ( ( cur = cur.parentNode ) ) {
+			ap.unshift( cur );
+		}
+		cur = b;
+		while ( ( cur = cur.parentNode ) ) {
+			bp.unshift( cur );
+		}
+
+		// Walk down the tree looking for a discrepancy
+		while ( ap[ i ] === bp[ i ] ) {
+			i++;
+		}
+
+		return i ?
+
+			// Do a sibling check if the nodes have a common ancestor
+			siblingCheck( ap[ i ], bp[ i ] ) :
+
+			// Otherwise nodes in our document sort first
+			// Support: IE 11+, Edge 17 - 18+
+			// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+			// two documents; shallow comparisons work.
+			/* eslint-disable eqeqeq */
+			ap[ i ] == preferredDoc ? -1 :
+			bp[ i ] == preferredDoc ? 1 :
+			/* eslint-enable eqeqeq */
+			0;
+	};
+
+	return document;
+};
+
+Sizzle.matches = function( expr, elements ) {
+	return Sizzle( expr, null, null, elements );
+};
+
+Sizzle.matchesSelector = function( elem, expr ) {
+	setDocument( elem );
+
+	if ( support.matchesSelector && documentIsHTML &&
+		!nonnativeSelectorCache[ expr + " " ] &&
+		( !rbuggyMatches || !rbuggyMatches.test( expr ) ) &&
+		( !rbuggyQSA     || !rbuggyQSA.test( expr ) ) ) {
+
+		try {
+			var ret = matches.call( elem, expr );
+
+			// IE 9's matchesSelector returns false on disconnected nodes
+			if ( ret || support.disconnectedMatch ||
+
+				// As well, disconnected nodes are said to be in a document
+				// fragment in IE 9
+				elem.document && elem.document.nodeType !== 11 ) {
+				return ret;
+			}
+		} catch ( e ) {
+			nonnativeSelectorCache( expr, true );
+		}
+	}
+
+	return Sizzle( expr, document, null, [ elem ] ).length > 0;
+};
+
+Sizzle.contains = function( context, elem ) {
+
+	// Set document vars if needed
+	// Support: IE 11+, Edge 17 - 18+
+	// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+	// two documents; shallow comparisons work.
+	// eslint-disable-next-line eqeqeq
+	if ( ( context.ownerDocument || context ) != document ) {
+		setDocument( context );
+	}
+	return contains( context, elem );
+};
+
+Sizzle.attr = function( elem, name ) {
+
+	// Set document vars if needed
+	// Support: IE 11+, Edge 17 - 18+
+	// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+	// two documents; shallow comparisons work.
+	// eslint-disable-next-line eqeqeq
+	if ( ( elem.ownerDocument || elem ) != document ) {
+		setDocument( elem );
+	}
+
+	var fn = Expr.attrHandle[ name.toLowerCase() ],
+
+		// Don't get fooled by Object.prototype properties (jQuery #13807)
+		val = fn && hasOwn.call( Expr.attrHandle, name.toLowerCase() ) ?
+			fn( elem, name, !documentIsHTML ) :
+			undefined;
+
+	return val !== undefined ?
+		val :
+		support.attributes || !documentIsHTML ?
+			elem.getAttribute( name ) :
+			( val = elem.getAttributeNode( name ) ) && val.specified ?
+				val.value :
+				null;
+};
+
+Sizzle.escape = function( sel ) {
+	return ( sel + "" ).replace( rcssescape, fcssescape );
+};
+
+Sizzle.error = function( msg ) {
+	throw new Error( "Syntax error, unrecognized expression: " + msg );
+};
+
+/**
+ * Document sorting and removing duplicates
+ * @param {ArrayLike} results
+ */
+Sizzle.uniqueSort = function( results ) {
+	var elem,
+		duplicates = [],
+		j = 0,
+		i = 0;
+
+	// Unless we *know* we can detect duplicates, assume their presence
+	hasDuplicate = !support.detectDuplicates;
+	sortInput = !support.sortStable && results.slice( 0 );
+	results.sort( sortOrder );
+
+	if ( hasDuplicate ) {
+		while ( ( elem = results[ i++ ] ) ) {
+			if ( elem === results[ i ] ) {
+				j = duplicates.push( i );
+			}
+		}
+		while ( j-- ) {
+			results.splice( duplicates[ j ], 1 );
+		}
+	}
+
+	// Clear input after sorting to release objects
+	// See https://github.com/jquery/sizzle/pull/225
+	sortInput = null;
+
+	return results;
+};
+
+/**
+ * Utility function for retrieving the text value of an array of DOM nodes
+ * @param {Array|Element} elem
+ */
+getText = Sizzle.getText = function( elem ) {
+	var node,
+		ret = "",
+		i = 0,
+		nodeType = elem.nodeType;
+
+	if ( !nodeType ) {
+
+		// If no nodeType, this is expected to be an array
+		while ( ( node = elem[ i++ ] ) ) {
+
+			// Do not traverse comment nodes
+			ret += getText( node );
+		}
+	} else if ( nodeType === 1 || nodeType === 9 || nodeType === 11 ) {
+
+		// Use textContent for elements
+		// innerText usage removed for consistency of new lines (jQuery #11153)
+		if ( typeof elem.textContent === "string" ) {
+			return elem.textContent;
+		} else {
+
+			// Traverse its children
+			for ( elem = elem.firstChild; elem; elem = elem.nextSibling ) {
+				ret += getText( elem );
+			}
+		}
+	} else if ( nodeType === 3 || nodeType === 4 ) {
+		return elem.nodeValue;
+	}
+
+	// Do not include comment or processing instruction nodes
+
+	return ret;
+};
+
+Expr = Sizzle.selectors = {
+
+	// Can be adjusted by the user
+	cacheLength: 50,
+
+	createPseudo: markFunction,
+
+	match: matchExpr,
+
+	attrHandle: {},
+
+	find: {},
+
+	relative: {
+		">": { dir: "parentNode", first: true },
+		" ": { dir: "parentNode" },
+		"+": { dir: "previousSibling", first: true },
+		"~": { dir: "previousSibling" }
+	},
+
+	preFilter: {
+		"ATTR": function( match ) {
+			match[ 1 ] = match[ 1 ].replace( runescape, funescape );
+
+			// Move the given value to match[3] whether quoted or unquoted
+			match[ 3 ] = ( match[ 3 ] || match[ 4 ] ||
+				match[ 5 ] || "" ).replace( runescape, funescape );
+
+			if ( match[ 2 ] === "~=" ) {
+				match[ 3 ] = " " + match[ 3 ] + " ";
+			}
+
+			return match.slice( 0, 4 );
+		},
+
+		"CHILD": function( match ) {
+
+			/* matches from matchExpr["CHILD"]
+				1 type (only|nth|...)
+				2 what (child|of-type)
+				3 argument (even|odd|\d*|\d*n([+-]\d+)?|...)
+				4 xn-component of xn+y argument ([+-]?\d*n|)
+				5 sign of xn-component
+				6 x of xn-component
+				7 sign of y-component
+				8 y of y-component
+			*/
+			match[ 1 ] = match[ 1 ].toLowerCase();
+
+			if ( match[ 1 ].slice( 0, 3 ) === "nth" ) {
+
+				// nth-* requires argument
+				if ( !match[ 3 ] ) {
+					Sizzle.error( match[ 0 ] );
+				}
+
+				// numeric x and y parameters for Expr.filter.CHILD
+				// remember that false/true cast respectively to 0/1
+				match[ 4 ] = +( match[ 4 ] ?
+					match[ 5 ] + ( match[ 6 ] || 1 ) :
+					2 * ( match[ 3 ] === "even" || match[ 3 ] === "odd" ) );
+				match[ 5 ] = +( ( match[ 7 ] + match[ 8 ] ) || match[ 3 ] === "odd" );
+
+				// other types prohibit arguments
+			} else if ( match[ 3 ] ) {
+				Sizzle.error( match[ 0 ] );
+			}
+
+			return match;
+		},
+
+		"PSEUDO": function( match ) {
+			var excess,
+				unquoted = !match[ 6 ] && match[ 2 ];
+
+			if ( matchExpr[ "CHILD" ].test( match[ 0 ] ) ) {
+				return null;
+			}
+
+			// Accept quoted arguments as-is
+			if ( match[ 3 ] ) {
+				match[ 2 ] = match[ 4 ] || match[ 5 ] || "";
+
+			// Strip excess characters from unquoted arguments
+			} else if ( unquoted && rpseudo.test( unquoted ) &&
+
+				// Get excess from tokenize (recursively)
+				( excess = tokenize( unquoted, true ) ) &&
+
+				// advance to the next closing parenthesis
+				( excess = unquoted.indexOf( ")", unquoted.length - excess ) - unquoted.length ) ) {
+
+				// excess is a negative index
+				match[ 0 ] = match[ 0 ].slice( 0, excess );
+				match[ 2 ] = unquoted.slice( 0, excess );
+			}
+
+			// Return only captures needed by the pseudo filter method (type and argument)
+			return match.slice( 0, 3 );
+		}
+	},
+
+	filter: {
+
+		"TAG": function( nodeNameSelector ) {
+			var nodeName = nodeNameSelector.replace( runescape, funescape ).toLowerCase();
+			return nodeNameSelector === "*" ?
+				function() {
+					return true;
+				} :
+				function( elem ) {
+					return elem.nodeName && elem.nodeName.toLowerCase() === nodeName;
+				};
+		},
+
+		"CLASS": function( className ) {
+			var pattern = classCache[ className + " " ];
+
+			return pattern ||
+				( pattern = new RegExp( "(^|" + whitespace +
+					")" + className + "(" + whitespace + "|$)" ) ) && classCache(
+						className, function( elem ) {
+							return pattern.test(
+								typeof elem.className === "string" && elem.className ||
+								typeof elem.getAttribute !== "undefined" &&
+									elem.getAttribute( "class" ) ||
+								""
+							);
+				} );
+		},
+
+		"ATTR": function( name, operator, check ) {
+			return function( elem ) {
+				var result = Sizzle.attr( elem, name );
+
+				if ( result == null ) {
+					return operator === "!=";
+				}
+				if ( !operator ) {
+					return true;
+				}
+
+				result += "";
+
+				/* eslint-disable max-len */
+
+				return operator === "=" ? result === check :
+					operator === "!=" ? result !== check :
+					operator === "^=" ? check && result.indexOf( check ) === 0 :
+					operator === "*=" ? check && result.indexOf( check ) > -1 :
+					operator === "$=" ? check && result.slice( -check.length ) === check :
+					operator === "~=" ? ( " " + result.replace( rwhitespace, " " ) + " " ).indexOf( check ) > -1 :
+					operator === "|=" ? result === check || result.slice( 0, check.length + 1 ) === check + "-" :
+					false;
+				/* eslint-enable max-len */
+
+			};
+		},
+
+		"CHILD": function( type, what, _argument, first, last ) {
+			var simple = type.slice( 0, 3 ) !== "nth",
+				forward = type.slice( -4 ) !== "last",
+				ofType = what === "of-type";
+
+			return first === 1 && last === 0 ?
+
+				// Shortcut for :nth-*(n)
+				function( elem ) {
+					return !!elem.parentNode;
+				} :
+
+				function( elem, _context, xml ) {
+					var cache, uniqueCache, outerCache, node, nodeIndex, start,
+						dir = simple !== forward ? "nextSibling" : "previousSibling",
+						parent = elem.parentNode,
+						name = ofType && elem.nodeName.toLowerCase(),
+						useCache = !xml && !ofType,
+						diff = false;
+
+					if ( parent ) {
+
+						// :(first|last|only)-(child|of-type)
+						if ( simple ) {
+							while ( dir ) {
+								node = elem;
+								while ( ( node = node[ dir ] ) ) {
+									if ( ofType ?
+										node.nodeName.toLowerCase() === name :
+										node.nodeType === 1 ) {
+
+										return false;
+									}
+								}
+
+								// Reverse direction for :only-* (if we haven't yet done so)
+								start = dir = type === "only" && !start && "nextSibling";
+							}
+							return true;
+						}
+
+						start = [ forward ? parent.firstChild : parent.lastChild ];
+
+						// non-xml :nth-child(...) stores cache data on `parent`
+						if ( forward && useCache ) {
+
+							// Seek `elem` from a previously-cached index
+
+							// ...in a gzip-friendly way
+							node = parent;
+							outerCache = node[ expando ] || ( node[ expando ] = {} );
+
+							// Support: IE <9 only
+							// Defend against cloned attroperties (jQuery gh-1709)
+							uniqueCache = outerCache[ node.uniqueID ] ||
+								( outerCache[ node.uniqueID ] = {} );
+
+							cache = uniqueCache[ type ] || [];
+							nodeIndex = cache[ 0 ] === dirruns && cache[ 1 ];
+							diff = nodeIndex && cache[ 2 ];
+							node = nodeIndex && parent.childNodes[ nodeIndex ];
+
+							while ( ( node = ++nodeIndex && node && node[ dir ] ||
+
+								// Fallback to seeking `elem` from the start
+								( diff = nodeIndex = 0 ) || start.pop() ) ) {
+
+								// When found, cache indexes on `parent` and break
+								if ( node.nodeType === 1 && ++diff && node === elem ) {
+									uniqueCache[ type ] = [ dirruns, nodeIndex, diff ];
+									break;
+								}
+							}
+
+						} else {
+
+							// Use previously-cached element index if available
+							if ( useCache ) {
+
+								// ...in a gzip-friendly way
+								node = elem;
+								outerCache = node[ expando ] || ( node[ expando ] = {} );
+
+								// Support: IE <9 only
+								// Defend against cloned attroperties (jQuery gh-1709)
+								uniqueCache = outerCache[ node.uniqueID ] ||
+									( outerCache[ node.uniqueID ] = {} );
+
+								cache = uniqueCache[ type ] || [];
+								nodeIndex = cache[ 0 ] === dirruns && cache[ 1 ];
+								diff = nodeIndex;
+							}
+
+							// xml :nth-child(...)
+							// or :nth-last-child(...) or :nth(-last)?-of-type(...)
+							if ( diff === false ) {
+
+								// Use the same loop as above to seek `elem` from the start
+								while ( ( node = ++nodeIndex && node && node[ dir ] ||
+									( diff = nodeIndex = 0 ) || start.pop() ) ) {
+
+									if ( ( ofType ?
+										node.nodeName.toLowerCase() === name :
+										node.nodeType === 1 ) &&
+										++diff ) {
+
+										// Cache the index of each encountered element
+										if ( useCache ) {
+											outerCache = node[ expando ] ||
+												( node[ expando ] = {} );
+
+											// Support: IE <9 only
+											// Defend against cloned attroperties (jQuery gh-1709)
+											uniqueCache = outerCache[ node.uniqueID ] ||
+												( outerCache[ node.uniqueID ] = {} );
+
+											uniqueCache[ type ] = [ dirruns, diff ];
+										}
+
+										if ( node === elem ) {
+											break;
+										}
+									}
+								}
+							}
+						}
+
+						// Incorporate the offset, then check against cycle size
+						diff -= last;
+						return diff === first || ( diff % first === 0 && diff / first >= 0 );
+					}
+				};
+		},
+
+		"PSEUDO": function( pseudo, argument ) {
+
+			// pseudo-class names are case-insensitive
+			// http://www.w3.org/TR/selectors/#pseudo-classes
+			// Prioritize by case sensitivity in case custom pseudos are added with uppercase letters
+			// Remember that setFilters inherits from pseudos
+			var args,
+				fn = Expr.pseudos[ pseudo ] || Expr.setFilters[ pseudo.toLowerCase() ] ||
+					Sizzle.error( "unsupported pseudo: " + pseudo );
+
+			// The user may use createPseudo to indicate that
+			// arguments are needed to create the filter function
+			// just as Sizzle does
+			if ( fn[ expando ] ) {
+				return fn( argument );
+			}
+
+			// But maintain support for old signatures
+			if ( fn.length > 1 ) {
+				args = [ pseudo, pseudo, "", argument ];
+				return Expr.setFilters.hasOwnProperty( pseudo.toLowerCase() ) ?
+					markFunction( function( seed, matches ) {
+						var idx,
+							matched = fn( seed, argument ),
+							i = matched.length;
+						while ( i-- ) {
+							idx = indexOf( seed, matched[ i ] );
+							seed[ idx ] = !( matches[ idx ] = matched[ i ] );
+						}
+					} ) :
+					function( elem ) {
+						return fn( elem, 0, args );
+					};
+			}
+
+			return fn;
+		}
+	},
+
+	pseudos: {
+
+		// Potentially complex pseudos
+		"not": markFunction( function( selector ) {
+
+			// Trim the selector passed to compile
+			// to avoid treating leading and trailing
+			// spaces as combinators
+			var input = [],
+				results = [],
+				matcher = compile( selector.replace( rtrim, "$1" ) );
+
+			return matcher[ expando ] ?
+				markFunction( function( seed, matches, _context, xml ) {
+					var elem,
+						unmatched = matcher( seed, null, xml, [] ),
+						i = seed.length;
+
+					// Match elements unmatched by `matcher`
+					while ( i-- ) {
+						if ( ( elem = unmatched[ i ] ) ) {
+							seed[ i ] = !( matches[ i ] = elem );
+						}
+					}
+				} ) :
+				function( elem, _context, xml ) {
+					input[ 0 ] = elem;
+					matcher( input, null, xml, results );
+
+					// Don't keep the element (issue #299)
+					input[ 0 ] = null;
+					return !results.pop();
+				};
+		} ),
+
+		"has": markFunction( function( selector ) {
+			return function( elem ) {
+				return Sizzle( selector, elem ).length > 0;
+			};
+		} ),
+
+		"contains": markFunction( function( text ) {
+			text = text.replace( runescape, funescape );
+			return function( elem ) {
+				return ( elem.textContent || getText( elem ) ).indexOf( text ) > -1;
+			};
+		} ),
+
+		// "Whether an element is represented by a :lang() selector
+		// is based solely on the element's language value
+		// being equal to the identifier C,
+		// or beginning with the identifier C immediately followed by "-".
+		// The matching of C against the element's language value is performed case-insensitively.
+		// The identifier C does not have to be a valid language name."
+		// http://www.w3.org/TR/selectors/#lang-pseudo
+		"lang": markFunction( function( lang ) {
+
+			// lang value must be a valid identifier
+			if ( !ridentifier.test( lang || "" ) ) {
+				Sizzle.error( "unsupported lang: " + lang );
+			}
+			lang = lang.replace( runescape, funescape ).toLowerCase();
+			return function( elem ) {
+				var elemLang;
+				do {
+					if ( ( elemLang = documentIsHTML ?
+						elem.lang :
+						elem.getAttribute( "xml:lang" ) || elem.getAttribute( "lang" ) ) ) {
+
+						elemLang = elemLang.toLowerCase();
+						return elemLang === lang || elemLang.indexOf( lang + "-" ) === 0;
+					}
+				} while ( ( elem = elem.parentNode ) && elem.nodeType === 1 );
+				return false;
+			};
+		} ),
+
+		// Miscellaneous
+		"target": function( elem ) {
+			var hash = window.location && window.location.hash;
+			return hash && hash.slice( 1 ) === elem.id;
+		},
+
+		"root": function( elem ) {
+			return elem === docElem;
+		},
+
+		"focus": function( elem ) {
+			return elem === document.activeElement &&
+				( !document.hasFocus || document.hasFocus() ) &&
+				!!( elem.type || elem.href || ~elem.tabIndex );
+		},
+
+		// Boolean properties
+		"enabled": createDisabledPseudo( false ),
+		"disabled": createDisabledPseudo( true ),
+
+		"checked": function( elem ) {
+
+			// In CSS3, :checked should return both checked and selected elements
+			// http://www.w3.org/TR/2011/REC-css3-selectors-20110929/#checked
+			var nodeName = elem.nodeName.toLowerCase();
+			return ( nodeName === "input" && !!elem.checked ) ||
+				( nodeName === "option" && !!elem.selected );
+		},
+
+		"selected": function( elem ) {
+
+			// Accessing this property makes selected-by-default
+			// options in Safari work properly
+			if ( elem.parentNode ) {
+				// eslint-disable-next-line no-unused-expressions
+				elem.parentNode.selectedIndex;
+			}
+
+			return elem.selected === true;
+		},
+
+		// Contents
+		"empty": function( elem ) {
+
+			// http://www.w3.org/TR/selectors/#empty-pseudo
+			// :empty is negated by element (1) or content nodes (text: 3; cdata: 4; entity ref: 5),
+			//   but not by others (comment: 8; processing instruction: 7; etc.)
+			// nodeType < 6 works because attributes (2) do not appear as children
+			for ( elem = elem.firstChild; elem; elem = elem.nextSibling ) {
+				if ( elem.nodeType < 6 ) {
+					return false;
+				}
+			}
+			return true;
+		},
+
+		"parent": function( elem ) {
+			return !Expr.pseudos[ "empty" ]( elem );
+		},
+
+		// Element/input types
+		"header": function( elem ) {
+			return rheader.test( elem.nodeName );
+		},
+
+		"input": function( elem ) {
+			return rinputs.test( elem.nodeName );
+		},
+
+		"button": function( elem ) {
+			var name = elem.nodeName.toLowerCase();
+			return name === "input" && elem.type === "button" || name === "button";
+		},
+
+		"text": function( elem ) {
+			var attr;
+			return elem.nodeName.toLowerCase() === "input" &&
+				elem.type === "text" &&
+
+				// Support: IE<8
+				// New HTML5 attribute values (e.g., "search") appear with elem.type === "text"
+				( ( attr = elem.getAttribute( "type" ) ) == null ||
+					attr.toLowerCase() === "text" );
+		},
+
+		// Position-in-collection
+		"first": createPositionalPseudo( function() {
+			return [ 0 ];
+		} ),
+
+		"last": createPositionalPseudo( function( _matchIndexes, length ) {
+			return [ length - 1 ];
+		} ),
+
+		"eq": createPositionalPseudo( function( _matchIndexes, length, argument ) {
+			return [ argument < 0 ? argument + length : argument ];
+		} ),
+
+		"even": createPositionalPseudo( function( matchIndexes, length ) {
+			var i = 0;
+			for ( ; i < length; i += 2 ) {
+				matchIndexes.push( i );
+			}
+			return matchIndexes;
+		} ),
+
+		"odd": createPositionalPseudo( function( matchIndexes, length ) {
+			var i = 1;
+			for ( ; i < length; i += 2 ) {
+				matchIndexes.push( i );
+			}
+			return matchIndexes;
+		} ),
+
+		"lt": createPositionalPseudo( function( matchIndexes, length, argument ) {
+			var i = argument < 0 ?
+				argument + length :
+				argument > length ?
+					length :
+					argument;
+			for ( ; --i >= 0; ) {
+				matchIndexes.push( i );
+			}
+			return matchIndexes;
+		} ),
+
+		"gt": createPositionalPseudo( function( matchIndexes, length, argument ) {
+			var i = argument < 0 ? argument + length : argument;
+			for ( ; ++i < length; ) {
+				matchIndexes.push( i );
+			}
+			return matchIndexes;
+		} )
+	}
+};
+
+Expr.pseudos[ "nth" ] = Expr.pseudos[ "eq" ];
+
+// Add button/input type pseudos
+for ( i in { radio: true, checkbox: true, file: true, password: true, image: true } ) {
+	Expr.pseudos[ i ] = createInputPseudo( i );
+}
+for ( i in { submit: true, reset: true } ) {
+	Expr.pseudos[ i ] = createButtonPseudo( i );
+}
+
+// Easy API for creating new setFilters
+function setFilters() {}
+setFilters.prototype = Expr.filters = Expr.pseudos;
+Expr.setFilters = new setFilters();
+
+tokenize = Sizzle.tokenize = function( selector, parseOnly ) {
+	var matched, match, tokens, type,
+		soFar, groups, preFilters,
+		cached = tokenCache[ selector + " " ];
+
+	if ( cached ) {
+		return parseOnly ? 0 : cached.slice( 0 );
+	}
+
+	soFar = selector;
+	groups = [];
+	preFilters = Expr.preFilter;
+
+	while ( soFar ) {
+
+		// Comma and first run
+		if ( !matched || ( match = rcomma.exec( soFar ) ) ) {
+			if ( match ) {
+
+				// Don't consume trailing commas as valid
+				soFar = soFar.slice( match[ 0 ].length ) || soFar;
+			}
+			groups.push( ( tokens = [] ) );
+		}
+
+		matched = false;
+
+		// Combinators
+		if ( ( match = rcombinators.exec( soFar ) ) ) {
+			matched = match.shift();
+			tokens.push( {
+				value: matched,
+
+				// Cast descendant combinators to space
+				type: match[ 0 ].replace( rtrim, " " )
+			} );
+			soFar = soFar.slice( matched.length );
+		}
+
+		// Filters
+		for ( type in Expr.filter ) {
+			if ( ( match = matchExpr[ type ].exec( soFar ) ) && ( !preFilters[ type ] ||
+				( match = preFilters[ type ]( match ) ) ) ) {
+				matched = match.shift();
+				tokens.push( {
+					value: matched,
+					type: type,
+					matches: match
+				} );
+				soFar = soFar.slice( matched.length );
+			}
+		}
+
+		if ( !matched ) {
+			break;
+		}
+	}
+
+	// Return the length of the invalid excess
+	// if we're just parsing
+	// Otherwise, throw an error or return tokens
+	return parseOnly ?
+		soFar.length :
+		soFar ?
+			Sizzle.error( selector ) :
+
+			// Cache the tokens
+			tokenCache( selector, groups ).slice( 0 );
+};
+
+function toSelector( tokens ) {
+	var i = 0,
+		len = tokens.length,
+		selector = "";
+	for ( ; i < len; i++ ) {
+		selector += tokens[ i ].value;
+	}
+	return selector;
+}
+
+function addCombinator( matcher, combinator, base ) {
+	var dir = combinator.dir,
+		skip = combinator.next,
+		key = skip || dir,
+		checkNonElements = base && key === "parentNode",
+		doneName = done++;
+
+	return combinator.first ?
+
+		// Check against closest ancestor/preceding element
+		function( elem, context, xml ) {
+			while ( ( elem = elem[ dir ] ) ) {
+				if ( elem.nodeType === 1 || checkNonElements ) {
+					return matcher( elem, context, xml );
+				}
+			}
+			return false;
+		} :
+
+		// Check against all ancestor/preceding elements
+		function( elem, context, xml ) {
+			var oldCache, uniqueCache, outerCache,
+				newCache = [ dirruns, doneName ];
+
+			// We can't set arbitrary data on XML nodes, so they don't benefit from combinator caching
+			if ( xml ) {
+				while ( ( elem = elem[ dir ] ) ) {
+					if ( elem.nodeType === 1 || checkNonElements ) {
+						if ( matcher( elem, context, xml ) ) {
+							return true;
+						}
+					}
+				}
+			} else {
+				while ( ( elem = elem[ dir ] ) ) {
+					if ( elem.nodeType === 1 || checkNonElements ) {
+						outerCache = elem[ expando ] || ( elem[ expando ] = {} );
+
+						// Support: IE <9 only
+						// Defend against cloned attroperties (jQuery gh-1709)
+						uniqueCache = outerCache[ elem.uniqueID ] ||
+							( outerCache[ elem.uniqueID ] = {} );
+
+						if ( skip && skip === elem.nodeName.toLowerCase() ) {
+							elem = elem[ dir ] || elem;
+						} else if ( ( oldCache = uniqueCache[ key ] ) &&
+							oldCache[ 0 ] === dirruns && oldCache[ 1 ] === doneName ) {
+
+							// Assign to newCache so results back-propagate to previous elements
+							return ( newCache[ 2 ] = oldCache[ 2 ] );
+						} else {
+
+							// Reuse newcache so results back-propagate to previous elements
+							uniqueCache[ key ] = newCache;
+
+							// A match means we're done; a fail means we have to keep checking
+							if ( ( newCache[ 2 ] = matcher( elem, context, xml ) ) ) {
+								return true;
+							}
+						}
+					}
+				}
+			}
+			return false;
+		};
+}
+
+function elementMatcher( matchers ) {
+	return matchers.length > 1 ?
+		function( elem, context, xml ) {
+			var i = matchers.length;
+			while ( i-- ) {
+				if ( !matchers[ i ]( elem, context, xml ) ) {
+					return false;
+				}
+			}
+			return true;
+		} :
+		matchers[ 0 ];
+}
+
+function multipleContexts( selector, contexts, results ) {
+	var i = 0,
+		len = contexts.length;
+	for ( ; i < len; i++ ) {
+		Sizzle( selector, contexts[ i ], results );
+	}
+	return results;
+}
+
+function condense( unmatched, map, filter, context, xml ) {
+	var elem,
+		newUnmatched = [],
+		i = 0,
+		len = unmatched.length,
+		mapped = map != null;
+
+	for ( ; i < len; i++ ) {
+		if ( ( elem = unmatched[ i ] ) ) {
+			if ( !filter || filter( elem, context, xml ) ) {
+				newUnmatched.push( elem );
+				if ( mapped ) {
+					map.push( i );
+				}
+			}
+		}
+	}
+
+	return newUnmatched;
+}
+
+function setMatcher( preFilter, selector, matcher, postFilter, postFinder, postSelector ) {
+	if ( postFilter && !postFilter[ expando ] ) {
+		postFilter = setMatcher( postFilter );
+	}
+	if ( postFinder && !postFinder[ expando ] ) {
+		postFinder = setMatcher( postFinder, postSelector );
+	}
+	return markFunction( function( seed, results, context, xml ) {
+		var temp, i, elem,
+			preMap = [],
+			postMap = [],
+			preexisting = results.length,
+
+			// Get initial elements from seed or context
+			elems = seed || multipleContexts(
+				selector || "*",
+				context.nodeType ? [ context ] : context,
+				[]
+			),
+
+			// Prefilter to get matcher input, preserving a map for seed-results synchronization
+			matcherIn = preFilter && ( seed || !selector ) ?
+				condense( elems, preMap, preFilter, context, xml ) :
+				elems,
+
+			matcherOut = matcher ?
+
+				// If we have a postFinder, or filtered seed, or non-seed postFilter or preexisting results,
+				postFinder || ( seed ? preFilter : preexisting || postFilter ) ?
+
+					// ...intermediate processing is necessary
+					[] :
+
+					// ...otherwise use results directly
+					results :
+				matcherIn;
+
+		// Find primary matches
+		if ( matcher ) {
+			matcher( matcherIn, matcherOut, context, xml );
+		}
+
+		// Apply postFilter
+		if ( postFilter ) {
+			temp = condense( matcherOut, postMap );
+			postFilter( temp, [], context, xml );
+
+			// Un-match failing elements by moving them back to matcherIn
+			i = temp.length;
+			while ( i-- ) {
+				if ( ( elem = temp[ i ] ) ) {
+					matcherOut[ postMap[ i ] ] = !( matcherIn[ postMap[ i ] ] = elem );
+				}
+			}
+		}
+
+		if ( seed ) {
+			if ( postFinder || preFilter ) {
+				if ( postFinder ) {
+
+					// Get the final matcherOut by condensing this intermediate into postFinder contexts
+					temp = [];
+					i = matcherOut.length;
+					while ( i-- ) {
+						if ( ( elem = matcherOut[ i ] ) ) {
+
+							// Restore matcherIn since elem is not yet a final match
+							temp.push( ( matcherIn[ i ] = elem ) );
+						}
+					}
+					postFinder( null, ( matcherOut = [] ), temp, xml );
+				}
+
+				// Move matched elements from seed to results to keep them synchronized
+				i = matcherOut.length;
+				while ( i-- ) {
+					if ( ( elem = matcherOut[ i ] ) &&
+						( temp = postFinder ? indexOf( seed, elem ) : preMap[ i ] ) > -1 ) {
+
+						seed[ temp ] = !( results[ temp ] = elem );
+					}
+				}
+			}
+
+		// Add elements to results, through postFinder if defined
+		} else {
+			matcherOut = condense(
+				matcherOut === results ?
+					matcherOut.splice( preexisting, matcherOut.length ) :
+					matcherOut
+			);
+			if ( postFinder ) {
+				postFinder( null, results, matcherOut, xml );
+			} else {
+				push.apply( results, matcherOut );
+			}
+		}
+	} );
+}
+
+function matcherFromTokens( tokens ) {
+	var checkContext, matcher, j,
+		len = tokens.length,
+		leadingRelative = Expr.relative[ tokens[ 0 ].type ],
+		implicitRelative = leadingRelative || Expr.relative[ " " ],
+		i = leadingRelative ? 1 : 0,
+
+		// The foundational matcher ensures that elements are reachable from top-level context(s)
+		matchContext = addCombinator( function( elem ) {
+			return elem === checkContext;
+		}, implicitRelative, true ),
+		matchAnyContext = addCombinator( function( elem ) {
+			return indexOf( checkContext, elem ) > -1;
+		}, implicitRelative, true ),
+		matchers = [ function( elem, context, xml ) {
+			var ret = ( !leadingRelative && ( xml || context !== outermostContext ) ) || (
+				( checkContext = context ).nodeType ?
+					matchContext( elem, context, xml ) :
+					matchAnyContext( elem, context, xml ) );
+
+			// Avoid hanging onto element (issue #299)
+			checkContext = null;
+			return ret;
+		} ];
+
+	for ( ; i < len; i++ ) {
+		if ( ( matcher = Expr.relative[ tokens[ i ].type ] ) ) {
+			matchers = [ addCombinator( elementMatcher( matchers ), matcher ) ];
+		} else {
+			matcher = Expr.filter[ tokens[ i ].type ].apply( null, tokens[ i ].matches );
+
+			// Return special upon seeing a positional matcher
+			if ( matcher[ expando ] ) {
+
+				// Find the next relative operator (if any) for proper handling
+				j = ++i;
+				for ( ; j < len; j++ ) {
+					if ( Expr.relative[ tokens[ j ].type ] ) {
+						break;
+					}
+				}
+				return setMatcher(
+					i > 1 && elementMatcher( matchers ),
+					i > 1 && toSelector(
+
+					// If the preceding token was a descendant combinator, insert an implicit any-element `*`
+					tokens
+						.slice( 0, i - 1 )
+						.concat( { value: tokens[ i - 2 ].type === " " ? "*" : "" } )
+					).replace( rtrim, "$1" ),
+					matcher,
+					i < j && matcherFromTokens( tokens.slice( i, j ) ),
+					j < len && matcherFromTokens( ( tokens = tokens.slice( j ) ) ),
+					j < len && toSelector( tokens )
+				);
+			}
+			matchers.push( matcher );
+		}
+	}
+
+	return elementMatcher( matchers );
+}
+
+function matcherFromGroupMatchers( elementMatchers, setMatchers ) {
+	var bySet = setMatchers.length > 0,
+		byElement = elementMatchers.length > 0,
+		superMatcher = function( seed, context, xml, results, outermost ) {
+			var elem, j, matcher,
+				matchedCount = 0,
+				i = "0",
+				unmatched = seed && [],
+				setMatched = [],
+				contextBackup = outermostContext,
+
+				// We must always have either seed elements or outermost context
+				elems = seed || byElement && Expr.find[ "TAG" ]( "*", outermost ),
+
+				// Use integer dirruns iff this is the outermost matcher
+				dirrunsUnique = ( dirruns += contextBackup == null ? 1 : Math.random() || 0.1 ),
+				len = elems.length;
+
+			if ( outermost ) {
+
+				// Support: IE 11+, Edge 17 - 18+
+				// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+				// two documents; shallow comparisons work.
+				// eslint-disable-next-line eqeqeq
+				outermostContext = context == document || context || outermost;
+			}
+
+			// Add elements passing elementMatchers directly to results
+			// Support: IE<9, Safari
+			// Tolerate NodeList properties (IE: "length"; Safari: <number>) matching elements by id
+			for ( ; i !== len && ( elem = elems[ i ] ) != null; i++ ) {
+				if ( byElement && elem ) {
+					j = 0;
+
+					// Support: IE 11+, Edge 17 - 18+
+					// IE/Edge sometimes throw a "Permission denied" error when strict-comparing
+					// two documents; shallow comparisons work.
+					// eslint-disable-next-line eqeqeq
+					if ( !context && elem.ownerDocument != document ) {
+						setDocument( elem );
+						xml = !documentIsHTML;
+					}
+					while ( ( matcher = elementMatchers[ j++ ] ) ) {
+						if ( matcher( elem, context || document, xml ) ) {
+							results.push( elem );
+							break;
+						}
+					}
+					if ( outermost ) {
+						dirruns = dirrunsUnique;
+					}
+				}
+
+				// Track unmatched elements for set filters
+				if ( bySet ) {
+
+					// They will have gone through all possible matchers
+					if ( ( elem = !matcher && elem ) ) {
+						matchedCount--;
+					}
+
+					// Lengthen the array for every element, matched or not
+					if ( seed ) {
+						unmatched.push( elem );
+					}
+				}
+			}
+
+			// `i` is now the count of elements visited above, and adding it to `matchedCount`
+			// makes the latter nonnegative.
+			matchedCount += i;
+
+			// Apply set filters to unmatched elements
+			// NOTE: This can be skipped if there are no unmatched elements (i.e., `matchedCount`
+			// equals `i`), unless we didn't visit _any_ elements in the above loop because we have
+			// no element matchers and no seed.
+			// Incrementing an initially-string "0" `i` allows `i` to remain a string only in that
+			// case, which will result in a "00" `matchedCount` that differs from `i` but is also
+			// numerically zero.
+			if ( bySet && i !== matchedCount ) {
+				j = 0;
+				while ( ( matcher = setMatchers[ j++ ] ) ) {
+					matcher( unmatched, setMatched, context, xml );
+				}
+
+				if ( seed ) {
+
+					// Reintegrate element matches to eliminate the need for sorting
+					if ( matchedCount > 0 ) {
+						while ( i-- ) {
+							if ( !( unmatched[ i ] || setMatched[ i ] ) ) {
+								setMatched[ i ] = pop.call( results );
+							}
+						}
+					}
+
+					// Discard index placeholder values to get only actual matches
+					setMatched = condense( setMatched );
+				}
+
+				// Add matches to results
+				push.apply( results, setMatched );
+
+				// Seedless set matches succeeding multiple successful matchers stipulate sorting
+				if ( outermost && !seed && setMatched.length > 0 &&
+					( matchedCount + setMatchers.length ) > 1 ) {
+
+					Sizzle.uniqueSort( results );
+				}
+			}
+
+			// Override manipulation of globals by nested matchers
+			if ( outermost ) {
+				dirruns = dirrunsUnique;
+				outermostContext = contextBackup;
+			}
+
+			return unmatched;
+		};
+
+	return bySet ?
+		markFunction( superMatcher ) :
+		superMatcher;
+}
+
+compile = Sizzle.compile = function( selector, match /* Internal Use Only */ ) {
+	var i,
+		setMatchers = [],
+		elementMatchers = [],
+		cached = compilerCache[ selector + " " ];
+
+	if ( !cached ) {
+
+		// Generate a function of recursive functions that can be used to check each element
+		if ( !match ) {
+			match = tokenize( selector );
+		}
+		i = match.length;
+		while ( i-- ) {
+			cached = matcherFromTokens( match[ i ] );
+			if ( cached[ expando ] ) {
+				setMatchers.push( cached );
+			} else {
+				elementMatchers.push( cached );
+			}
+		}
+
+		// Cache the compiled function
+		cached = compilerCache(
+			selector,
+			matcherFromGroupMatchers( elementMatchers, setMatchers )
+		);
+
+		// Save selector and tokenization
+		cached.selector = selector;
+	}
+	return cached;
+};
+
+/**
+ * A low-level selection function that works with Sizzle's compiled
+ *  selector functions
+ * @param {String|Function} selector A selector or a pre-compiled
+ *  selector function built with Sizzle.compile
+ * @param {Element} context
+ * @param {Array} [results]
+ * @param {Array} [seed] A set of elements to match against
+ */
+select = Sizzle.select = function( selector, context, results, seed ) {
+	var i, tokens, token, type, find,
+		compiled = typeof selector === "function" && selector,
+		match = !seed && tokenize( ( selector = compiled.selector || selector ) );
+
+	results = results || [];
+
+	// Try to minimize operations if there is only one selector in the list and no seed
+	// (the latter of which guarantees us context)
+	if ( match.length === 1 ) {
+
+		// Reduce context if the leading compound selector is an ID
+		tokens = match[ 0 ] = match[ 0 ].slice( 0 );
+		if ( tokens.length > 2 && ( token = tokens[ 0 ] ).type === "ID" &&
+			context.nodeType === 9 && documentIsHTML && Expr.relative[ tokens[ 1 ].type ] ) {
+
+			context = ( Expr.find[ "ID" ]( token.matches[ 0 ]
+				.replace( runescape, funescape ), context ) || [] )[ 0 ];
+			if ( !context ) {
+				return results;
+
+			// Precompiled matchers will still verify ancestry, so step up a level
+			} else if ( compiled ) {
+				context = context.parentNode;
+			}
+
+			selector = selector.slice( tokens.shift().value.length );
+		}
+
+		// Fetch a seed set for right-to-left matching
+		i = matchExpr[ "needsContext" ].test( selector ) ? 0 : tokens.length;
+		while ( i-- ) {
+			token = tokens[ i ];
+
+			// Abort if we hit a combinator
+			if ( Expr.relative[ ( type = token.type ) ] ) {
+				break;
+			}
+			if ( ( find = Expr.find[ type ] ) ) {
+
+				// Search, expanding context for leading sibling combinators
+				if ( ( seed = find(
+					token.matches[ 0 ].replace( runescape, funescape ),
+					rsibling.test( tokens[ 0 ].type ) && testContext( context.parentNode ) ||
+						context
+				) ) ) {
+
+					// If seed is empty or no tokens remain, we can return early
+					tokens.splice( i, 1 );
+					selector = seed.length && toSelector( tokens );
+					if ( !selector ) {
+						push.apply( results, seed );
+						return results;
+					}
+
+					break;
+				}
+			}
+		}
+	}
+
+	// Compile and execute a filtering function if one is not provided
+	// Provide `match` to avoid retokenization if we modified the selector above
+	( compiled || compile( selector, match ) )(
+		seed,
+		context,
+		!documentIsHTML,
+		results,
+		!context || rsibling.test( selector ) && testContext( context.parentNode ) || context
+	);
+	return results;
+};
+
+// One-time assignments
+
+// Sort stability
+support.sortStable = expando.split( "" ).sort( sortOrder ).join( "" ) === expando;
+
+// Support: Chrome 14-35+
+// Always assume duplicates if they aren't passed to the comparison function
+support.detectDuplicates = !!hasDuplicate;
+
+// Initialize against the default document
+setDocument();
+
+// Support: Webkit<537.32 - Safari 6.0.3/Chrome 25 (fixed in Chrome 27)
+// Detached nodes confoundingly follow *each other*
+support.sortDetached = assert( function( el ) {
+
+	// Should return 1, but returns 4 (following)
+	return el.compareDocumentPosition( document.createElement( "fieldset" ) ) & 1;
+} );
+
+// Support: IE<8
+// Prevent attribute/property "interpolation"
+// https://msdn.microsoft.com/en-us/library/ms536429%28VS.85%29.aspx
+if ( !assert( function( el ) {
+	el.innerHTML = "<a href='#'></a>";
+	return el.firstChild.getAttribute( "href" ) === "#";
+} ) ) {
+	addHandle( "type|href|height|width", function( elem, name, isXML ) {
+		if ( !isXML ) {
+			return elem.getAttribute( name, name.toLowerCase() === "type" ? 1 : 2 );
+		}
+	} );
+}
+
+// Support: IE<9
+// Use defaultValue in place of getAttribute("value")
+if ( !support.attributes || !assert( function( el ) {
+	el.innerHTML = "<input/>";
+	el.firstChild.setAttribute( "value", "" );
+	return el.firstChild.getAttribute( "value" ) === "";
+} ) ) {
+	addHandle( "value", function( elem, _name, isXML ) {
+		if ( !isXML && elem.nodeName.toLowerCase() === "input" ) {
+			return elem.defaultValue;
+		}
+	} );
+}
+
+// Support: IE<9
+// Use getAttributeNode to fetch booleans when getAttribute lies
+if ( !assert( function( el ) {
+	return el.getAttribute( "disabled" ) == null;
+} ) ) {
+	addHandle( booleans, function( elem, name, isXML ) {
+		var val;
+		if ( !isXML ) {
+			return elem[ name ] === true ? name.toLowerCase() :
+				( val = elem.getAttributeNode( name ) ) && val.specified ?
+					val.value :
+					null;
+		}
+	} );
+}
+
+return Sizzle;
+
+} )( window );
+
+
+
+jQuery.find = Sizzle;
+jQuery.expr = Sizzle.selectors;
+
+// Deprecated
+jQuery.expr[ ":" ] = jQuery.expr.pseudos;
+jQuery.uniqueSort = jQuery.unique = Sizzle.uniqueSort;
+jQuery.text = Sizzle.getText;
+jQuery.isXMLDoc = Sizzle.isXML;
+jQuery.contains = Sizzle.contains;
+jQuery.escapeSelector = Sizzle.escape;
+
+
+
+
+var dir = function( elem, dir, until ) {
+	var matched = [],
+		truncate = until !== undefined;
+
+	while ( ( elem = elem[ dir ] ) && elem.nodeType !== 9 ) {
+		if ( elem.nodeType === 1 ) {
+			if ( truncate && jQuery( elem ).is( until ) ) {
+				break;
+			}
+			matched.push( elem );
+		}
+	}
+	return matched;
+};
+
+
+var siblings = function( n, elem ) {
+	var matched = [];
+
+	for ( ; n; n = n.nextSibling ) {
+		if ( n.nodeType === 1 && n !== elem ) {
+			matched.push( n );
+		}
+	}
+
+	return matched;
+};
+
+
+var rneedsContext = jQuery.expr.match.needsContext;
+
+
+
+function nodeName( elem, name ) {
+
+	return elem.nodeName && elem.nodeName.toLowerCase() === name.toLowerCase();
+
+}
+var rsingleTag = ( /^<([a-z][^\/\0>:\x20\t\r\n\f]*)[\x20\t\r\n\f]*\/?>(?:<\/\1>|)$/i );
+
+
+
+// Implement the identical functionality for filter and not
+function winnow( elements, qualifier, not ) {
+	if ( isFunction( qualifier ) ) {
+		return jQuery.grep( elements, function( elem, i ) {
+			return !!qualifier.call( elem, i, elem ) !== not;
+		} );
+	}
+
+	// Single element
+	if ( qualifier.nodeType ) {
+		return jQuery.grep( elements, function( elem ) {
+			return ( elem === qualifier ) !== not;
+		} );
+	}
+
+	// Arraylike of elements (jQuery, arguments, Array)
+	if ( typeof qualifier !== "string" ) {
+		return jQuery.grep( elements, function( elem ) {
+			return ( indexOf.call( qualifier, elem ) > -1 ) !== not;
+		} );
+	}
+
+	// Filtered directly for both simple and complex selectors
+	return jQuery.filter( qualifier, elements, not );
+}
+
+jQuery.filter = function( expr, elems, not ) {
+	var elem = elems[ 0 ];
+
+	if ( not ) {
+		expr = ":not(" + expr + ")";
+	}
+
+	if ( elems.length === 1 && elem.nodeType === 1 ) {
+		return jQuery.find.matchesSelector( elem, expr ) ? [ elem ] : [];
+	}
+
+	return jQuery.find.matches( expr, jQuery.grep( elems, function( elem ) {
+		return elem.nodeType === 1;
+	} ) );
+};
+
+jQuery.fn.extend( {
+	find: function( selector ) {
+		var i, ret,
+			len = this.length,
+			self = this;
+
+		if ( typeof selector !== "string" ) {
+			return this.pushStack( jQuery( selector ).filter( function() {
+				for ( i = 0; i < len; i++ ) {
+					if ( jQuery.contains( self[ i ], this ) ) {
+						return true;
+					}
+				}
+			} ) );
+		}
+
+		ret = this.pushStack( [] );
+
+		for ( i = 0; i < len; i++ ) {
+			jQuery.find( selector, self[ i ], ret );
+		}
+
+		return len > 1 ? jQuery.uniqueSort( ret ) : ret;
+	},
+	filter: function( selector ) {
+		return this.pushStack( winnow( this, selector || [], false ) );
+	},
+	not: function( selector ) {
+		return this.pushStack( winnow( this, selector || [], true ) );
+	},
+	is: function( selector ) {
+		return !!winnow(
+			this,
+
+			// If this is a positional/relative selector, check membership in the returned set
+			// so $("p:first").is("p:last") won't return true for a doc with two "p".
+			typeof selector === "string" && rneedsContext.test( selector ) ?
+				jQuery( selector ) :
+				selector || [],
+			false
+		).length;
+	}
+} );
+
+
+// Initialize a jQuery object
+
+
+// A central reference to the root jQuery(document)
+var rootjQuery,
+
+	// A simple way to check for HTML strings
+	// Prioritize #id over <tag> to avoid XSS via location.hash (#9521)
+	// Strict HTML recognition (#11290: must start with <)
+	// Shortcut simple #id case for speed
+	rquickExpr = /^(?:\s*(<[\w\W]+>)[^>]*|#([\w-]+))$/,
+
+	init = jQuery.fn.init = function( selector, context, root ) {
+		var match, elem;
+
+		// HANDLE: $(""), $(null), $(undefined), $(false)
+		if ( !selector ) {
+			return this;
+		}
+
+		// Method init() accepts an alternate rootjQuery
+		// so migrate can support jQuery.sub (gh-2101)
+		root = root || rootjQuery;
+
+		// Handle HTML strings
+		if ( typeof selector === "string" ) {
+			if ( selector[ 0 ] === "<" &&
+				selector[ selector.length - 1 ] === ">" &&
+				selector.length >= 3 ) {
+
+				// Assume that strings that start and end with <> are HTML and skip the regex check
+				match = [ null, selector, null ];
+
+			} else {
+				match = rquickExpr.exec( selector );
+			}
+
+			// Match html or make sure no context is specified for #id
+			if ( match && ( match[ 1 ] || !context ) ) {
+
+				// HANDLE: $(html) -> $(array)
+				if ( match[ 1 ] ) {
+					context = context instanceof jQuery ? context[ 0 ] : context;
+
+					// Option to run scripts is true for back-compat
+					// Intentionally let the error be thrown if parseHTML is not present
+					jQuery.merge( this, jQuery.parseHTML(
+						match[ 1 ],
+						context && context.nodeType ? context.ownerDocument || context : document,
+						true
+					) );
+
+					// HANDLE: $(html, props)
+					if ( rsingleTag.test( match[ 1 ] ) && jQuery.isPlainObject( context ) ) {
+						for ( match in context ) {
+
+							// Properties of context are called as methods if possible
+							if ( isFunction( this[ match ] ) ) {
+								this[ match ]( context[ match ] );
+
+							// ...and otherwise set as attributes
+							} else {
+								this.attr( match, context[ match ] );
+							}
+						}
+					}
+
+					return this;
+
+				// HANDLE: $(#id)
+				} else {
+					elem = document.getElementById( match[ 2 ] );
+
+					if ( elem ) {
+
+						// Inject the element directly into the jQuery object
+						this[ 0 ] = elem;
+						this.length = 1;
+					}
+					return this;
+				}
+
+			// HANDLE: $(expr, $(...))
+			} else if ( !context || context.jquery ) {
+				return ( context || root ).find( selector );
+
+			// HANDLE: $(expr, context)
+			// (which is just equivalent to: $(context).find(expr)
+			} else {
+				return this.constructor( context ).find( selector );
+			}
+
+		// HANDLE: $(DOMElement)
+		} else if ( selector.nodeType ) {
+			this[ 0 ] = selector;
+			this.length = 1;
+			return this;
+
+		// HANDLE: $(function)
+		// Shortcut for document ready
+		} else if ( isFunction( selector ) ) {
+			return root.ready !== undefined ?
+				root.ready( selector ) :
+
+				// Execute immediately if ready is not present
+				selector( jQuery );
+		}
+
+		return jQuery.makeArray( selector, this );
+	};
+
+// Give the init function the jQuery prototype for later instantiation
+init.prototype = jQuery.fn;
+
+// Initialize central reference
+rootjQuery = jQuery( document );
+
+
+var rparentsprev = /^(?:parents|prev(?:Until|All))/,
+
+	// Methods guaranteed to produce a unique set when starting from a unique set
+	guaranteedUnique = {
+		children: true,
+		contents: true,
+		next: true,
+		prev: true
+	};
+
+jQuery.fn.extend( {
+	has: function( target ) {
+		var targets = jQuery( target, this ),
+			l = targets.length;
+
+		return this.filter( function() {
+			var i = 0;
+			for ( ; i < l; i++ ) {
+				if ( jQuery.contains( this, targets[ i ] ) ) {
+					return true;
+				}
+			}
+		} );
+	},
+
+	closest: function( selectors, context ) {
+		var cur,
+			i = 0,
+			l = this.length,
+			matched = [],
+			targets = typeof selectors !== "string" && jQuery( selectors );
+
+		// Positional selectors never match, since there's no _selection_ context
+		if ( !rneedsContext.test( selectors ) ) {
+			for ( ; i < l; i++ ) {
+				for ( cur = this[ i ]; cur && cur !== context; cur = cur.parentNode ) {
+
+					// Always skip document fragments
+					if ( cur.nodeType < 11 && ( targets ?
+						targets.index( cur ) > -1 :
+
+						// Don't pass non-elements to Sizzle
+						cur.nodeType === 1 &&
+							jQuery.find.matchesSelector( cur, selectors ) ) ) {
+
+						matched.push( cur );
+						break;
+					}
+				}
+			}
+		}
+
+		return this.pushStack( matched.length > 1 ? jQuery.uniqueSort( matched ) : matched );
+	},
+
+	// Determine the position of an element within the set
+	index: function( elem ) {
+
+		// No argument, return index in parent
+		if ( !elem ) {
+			return ( this[ 0 ] && this[ 0 ].parentNode ) ? this.first().prevAll().length : -1;
+		}
+
+		// Index in selector
+		if ( typeof elem === "string" ) {
+			return indexOf.call( jQuery( elem ), this[ 0 ] );
+		}
+
+		// Locate the position of the desired element
+		return indexOf.call( this,
+
+			// If it receives a jQuery object, the first element is used
+			elem.jquery ? elem[ 0 ] : elem
+		);
+	},
+
+	add: function( selector, context ) {
+		return this.pushStack(
+			jQuery.uniqueSort(
+				jQuery.merge( this.get(), jQuery( selector, context ) )
+			)
+		);
+	},
+
+	addBack: function( selector ) {
+		return this.add( selector == null ?
+			this.prevObject : this.prevObject.filter( selector )
+		);
+	}
+} );
+
+function sibling( cur, dir ) {
+	while ( ( cur = cur[ dir ] ) && cur.nodeType !== 1 ) {}
+	return cur;
+}
+
+jQuery.each( {
+	parent: function( elem ) {
+		var parent = elem.parentNode;
+		return parent && parent.nodeType !== 11 ? parent : null;
+	},
+	parents: function( elem ) {
+		return dir( elem, "parentNode" );
+	},
+	parentsUntil: function( elem, _i, until ) {
+		return dir( elem, "parentNode", until );
+	},
+	next: function( elem ) {
+		return sibling( elem, "nextSibling" );
+	},
+	prev: function( elem ) {
+		return sibling( elem, "previousSibling" );
+	},
+	nextAll: function( elem ) {
+		return dir( elem, "nextSibling" );
+	},
+	prevAll: function( elem ) {
+		return dir( elem, "previousSibling" );
+	},
+	nextUntil: function( elem, _i, until ) {
+		return dir( elem, "nextSibling", until );
+	},
+	prevUntil: function( elem, _i, until ) {
+		return dir( elem, "previousSibling", until );
+	},
+	siblings: function( elem ) {
+		return siblings( ( elem.parentNode || {} ).firstChild, elem );
+	},
+	children: function( elem ) {
+		return siblings( elem.firstChild );
+	},
+	contents: function( elem ) {
+		if ( elem.contentDocument != null &&
+
+			// Support: IE 11+
+			// <object> elements with no `data` attribute has an object
+			// `contentDocument` with a `null` prototype.
+			getProto( elem.contentDocument ) ) {
+
+			return elem.contentDocument;
+		}
+
+		// Support: IE 9 - 11 only, iOS 7 only, Android Browser <=4.3 only
+		// Treat the template element as a regular one in browsers that
+		// don't support it.
+		if ( nodeName( elem, "template" ) ) {
+			elem = elem.content || elem;
+		}
+
+		return jQuery.merge( [], elem.childNodes );
+	}
+}, function( name, fn ) {
+	jQuery.fn[ name ] = function( until, selector ) {
+		var matched = jQuery.map( this, fn, until );
+
+		if ( name.slice( -5 ) !== "Until" ) {
+			selector = until;
+		}
+
+		if ( selector && typeof selector === "string" ) {
+			matched = jQuery.filter( selector, matched );
+		}
+
+		if ( this.length > 1 ) {
+
+			// Remove duplicates
+			if ( !guaranteedUnique[ name ] ) {
+				jQuery.uniqueSort( matched );
+			}
+
+			// Reverse order for parents* and prev-derivatives
+			if ( rparentsprev.test( name ) ) {
+				matched.reverse();
+			}
+		}
+
+		return this.pushStack( matched );
+	};
+} );
+var rnothtmlwhite = ( /[^\x20\t\r\n\f]+/g );
+
+
+
+// Convert String-formatted options into Object-formatted ones
+function createOptions( options ) {
+	var object = {};
+	jQuery.each( options.match( rnothtmlwhite ) || [], function( _, flag ) {
+		object[ flag ] = true;
+	} );
+	return object;
+}
+
+/*
+ * Create a callback list using the following parameters:
+ *
+ *	options: an optional list of space-separated options that will change how
+ *			the callback list behaves or a more traditional option object
+ *
+ * By default a callback list will act like an event callback list and can be
+ * "fired" multiple times.
+ *
+ * Possible options:
+ *
+ *	once:			will ensure the callback list can only be fired once (like a Deferred)
+ *
+ *	memory:			will keep track of previous values and will call any callback added
+ *					after the list has been fired right away with the latest "memorized"
+ *					values (like a Deferred)
+ *
+ *	unique:			will ensure a callback can only be added once (no duplicate in the list)
+ *
+ *	stopOnFalse:	interrupt callings when a callback returns false
+ *
+ */
+jQuery.Callbacks = function( options ) {
+
+	// Convert options from String-formatted to Object-formatted if needed
+	// (we check in cache first)
+	options = typeof options === "string" ?
+		createOptions( options ) :
+		jQuery.extend( {}, options );
+
+	var // Flag to know if list is currently firing
+		firing,
+
+		// Last fire value for non-forgettable lists
+		memory,
+
+		// Flag to know if list was already fired
+		fired,
+
+		// Flag to prevent firing
+		locked,
+
+		// Actual callback list
+		list = [],
+
+		// Queue of execution data for repeatable lists
+		queue = [],
+
+		// Index of currently firing callback (modified by add/remove as needed)
+		firingIndex = -1,
+
+		// Fire callbacks
+		fire = function() {
+
+			// Enforce single-firing
+			locked = locked || options.once;
+
+			// Execute callbacks for all pending executions,
+			// respecting firingIndex overrides and runtime changes
+			fired = firing = true;
+			for ( ; queue.length; firingIndex = -1 ) {
+				memory = queue.shift();
+				while ( ++firingIndex < list.length ) {
+
+					// Run callback and check for early termination
+					if ( list[ firingIndex ].apply( memory[ 0 ], memory[ 1 ] ) === false &&
+						options.stopOnFalse ) {
+
+						// Jump to end and forget the data so .add doesn't re-fire
+						firingIndex = list.length;
+						memory = false;
+					}
+				}
+			}
+
+			// Forget the data if we're done with it
+			if ( !options.memory ) {
+				memory = false;
+			}
+
+			firing = false;
+
+			// Clean up if we're done firing for good
+			if ( locked ) {
+
+				// Keep an empty list if we have data for future add calls
+				if ( memory ) {
+					list = [];
+
+				// Otherwise, this object is spent
+				} else {
+					list = "";
+				}
+			}
+		},
+
+		// Actual Callbacks object
+		self = {
+
+			// Add a callback or a collection of callbacks to the list
+			add: function() {
+				if ( list ) {
+
+					// If we have memory from a past run, we should fire after adding
+					if ( memory && !firing ) {
+						firingIndex = list.length - 1;
+						queue.push( memory );
+					}
+
+					( function add( args ) {
+						jQuery.each( args, function( _, arg ) {
+							if ( isFunction( arg ) ) {
+								if ( !options.unique || !self.has( arg ) ) {
+									list.push( arg );
+								}
+							} else if ( arg && arg.length && toType( arg ) !== "string" ) {
+
+								// Inspect recursively
+								add( arg );
+							}
+						} );
+					} )( arguments );
+
+					if ( memory && !firing ) {
+						fire();
+					}
+				}
+				return this;
+			},
+
+			// Remove a callback from the list
+			remove: function() {
+				jQuery.each( arguments, function( _, arg ) {
+					var index;
+					while ( ( index = jQuery.inArray( arg, list, index ) ) > -1 ) {
+						list.splice( index, 1 );
+
+						// Handle firing indexes
+						if ( index <= firingIndex ) {
+							firingIndex--;
+						}
+					}
+				} );
+				return this;
+			},
+
+			// Check if a given callback is in the list.
+			// If no argument is given, return whether or not list has callbacks attached.
+			has: function( fn ) {
+				return fn ?
+					jQuery.inArray( fn, list ) > -1 :
+					list.length > 0;
+			},
+
+			// Remove all callbacks from the list
+			empty: function() {
+				if ( list ) {
+					list = [];
+				}
+				return this;
+			},
+
+			// Disable .fire and .add
+			// Abort any current/pending executions
+			// Clear all callbacks and values
+			disable: function() {
+				locked = queue = [];
+				list = memory = "";
+				return this;
+			},
+			disabled: function() {
+				return !list;
+			},
+
+			// Disable .fire
+			// Also disable .add unless we have memory (since it would have no effect)
+			// Abort any pending executions
+			lock: function() {
+				locked = queue = [];
+				if ( !memory && !firing ) {
+					list = memory = "";
+				}
+				return this;
+			},
+			locked: function() {
+				return !!locked;
+			},
+
+			// Call all callbacks with the given context and arguments
+			fireWith: function( context, args ) {
+				if ( !locked ) {
+					args = args || [];
+					args = [ context, args.slice ? args.slice() : args ];
+					queue.push( args );
+					if ( !firing ) {
+						fire();
+					}
+				}
+				return this;
+			},
+
+			// Call all the callbacks with the given arguments
+			fire: function() {
+				self.fireWith( this, arguments );
+				return this;
+			},
+
+			// To know if the callbacks have already been called at least once
+			fired: function() {
+				return !!fired;
+			}
+		};
+
+	return self;
+};
+
+
+function Identity( v ) {
+	return v;
+}
+function Thrower( ex ) {
+	throw ex;
+}
+
+function adoptValue( value, resolve, reject, noValue ) {
+	var method;
+
+	try {
+
+		// Check for promise aspect first to privilege synchronous behavior
+		if ( value && isFunction( ( method = value.promise ) ) ) {
+			method.call( value ).done( resolve ).fail( reject );
+
+		// Other thenables
+		} else if ( value && isFunction( ( method = value.then ) ) ) {
+			method.call( value, resolve, reject );
+
+		// Other non-thenables
+		} else {
+
+			// Control `resolve` arguments by letting Array#slice cast boolean `noValue` to integer:
+			// * false: [ value ].slice( 0 ) => resolve( value )
+			// * true: [ value ].slice( 1 ) => resolve()
+			resolve.apply( undefined, [ value ].slice( noValue ) );
+		}
+
+	// For Promises/A+, convert exceptions into rejections
+	// Since jQuery.when doesn't unwrap thenables, we can skip the extra checks appearing in
+	// Deferred#then to conditionally suppress rejection.
+	} catch ( value ) {
+
+		// Support: Android 4.0 only
+		// Strict mode functions invoked without .call/.apply get global-object context
+		reject.apply( undefined, [ value ] );
+	}
+}
+
+jQuery.extend( {
+
+	Deferred: function( func ) {
+		var tuples = [
+
+				// action, add listener, callbacks,
+				// ... .then handlers, argument index, [final state]
+				[ "notify", "progress", jQuery.Callbacks( "memory" ),
+					jQuery.Callbacks( "memory" ), 2 ],
+				[ "resolve", "done", jQuery.Callbacks( "once memory" ),
+					jQuery.Callbacks( "once memory" ), 0, "resolved" ],
+				[ "reject", "fail", jQuery.Callbacks( "once memory" ),
+					jQuery.Callbacks( "once memory" ), 1, "rejected" ]
+			],
+			state = "pending",
+			promise = {
+				state: function() {
+					return state;
+				},
+				always: function() {
+					deferred.done( arguments ).fail( arguments );
+					return this;
+				},
+				"catch": function( fn ) {
+					return promise.then( null, fn );
+				},
+
+				// Keep pipe for back-compat
+				pipe: function( /* fnDone, fnFail, fnProgress */ ) {
+					var fns = arguments;
+
+					return jQuery.Deferred( function( newDefer ) {
+						jQuery.each( tuples, function( _i, tuple ) {
+
+							// Map tuples (progress, done, fail) to arguments (done, fail, progress)
+							var fn = isFunction( fns[ tuple[ 4 ] ] ) && fns[ tuple[ 4 ] ];
+
+							// deferred.progress(function() { bind to newDefer or newDefer.notify })
+							// deferred.done(function() { bind to newDefer or newDefer.resolve })
+							// deferred.fail(function() { bind to newDefer or newDefer.reject })
+							deferred[ tuple[ 1 ] ]( function() {
+								var returned = fn && fn.apply( this, arguments );
+								if ( returned && isFunction( returned.promise ) ) {
+									returned.promise()
+										.progress( newDefer.notify )
+										.done( newDefer.resolve )
+										.fail( newDefer.reject );
+								} else {
+									newDefer[ tuple[ 0 ] + "With" ](
+										this,
+										fn ? [ returned ] : arguments
+									);
+								}
+							} );
+						} );
+						fns = null;
+					} ).promise();
+				},
+				then: function( onFulfilled, onRejected, onProgress ) {
+					var maxDepth = 0;
+					function resolve( depth, deferred, handler, special ) {
+						return function() {
+							var that = this,
+								args = arguments,
+								mightThrow = function() {
+									var returned, then;
+
+									// Support: Promises/A+ section 2.3.3.3.3
+									// https://promisesaplus.com/#point-59
+									// Ignore double-resolution attempts
+									if ( depth < maxDepth ) {
+										return;
+									}
+
+									returned = handler.apply( that, args );
+
+									// Support: Promises/A+ section 2.3.1
+									// https://promisesaplus.com/#point-48
+									if ( returned === deferred.promise() ) {
+										throw new TypeError( "Thenable self-resolution" );
+									}
+
+									// Support: Promises/A+ sections 2.3.3.1, 3.5
+									// https://promisesaplus.com/#point-54
+									// https://promisesaplus.com/#point-75
+									// Retrieve `then` only once
+									then = returned &&
+
+										// Support: Promises/A+ section 2.3.4
+										// https://promisesaplus.com/#point-64
+										// Only check objects and functions for thenability
+										( typeof returned === "object" ||
+											typeof returned === "function" ) &&
+										returned.then;
+
+									// Handle a returned thenable
+									if ( isFunction( then ) ) {
+
+										// Special processors (notify) just wait for resolution
+										if ( special ) {
+											then.call(
+												returned,
+												resolve( maxDepth, deferred, Identity, special ),
+												resolve( maxDepth, deferred, Thrower, special )
+											);
+
+										// Normal processors (resolve) also hook into progress
+										} else {
+
+											// ...and disregard older resolution values
+											maxDepth++;
+
+											then.call(
+												returned,
+												resolve( maxDepth, deferred, Identity, special ),
+												resolve( maxDepth, deferred, Thrower, special ),
+												resolve( maxDepth, deferred, Identity,
+													deferred.notifyWith )
+											);
+										}
+
+									// Handle all other returned values
+									} else {
+
+										// Only substitute handlers pass on context
+										// and multiple values (non-spec behavior)
+										if ( handler !== Identity ) {
+											that = undefined;
+											args = [ returned ];
+										}
+
+										// Process the value(s)
+										// Default process is resolve
+										( special || deferred.resolveWith )( that, args );
+									}
+								},
+
+								// Only normal processors (resolve) catch and reject exceptions
+								process = special ?
+									mightThrow :
+									function() {
+										try {
+											mightThrow();
+										} catch ( e ) {
+
+											if ( jQuery.Deferred.exceptionHook ) {
+												jQuery.Deferred.exceptionHook( e,
+													process.stackTrace );
+											}
+
+											// Support: Promises/A+ section 2.3.3.3.4.1
+											// https://promisesaplus.com/#point-61
+											// Ignore post-resolution exceptions
+											if ( depth + 1 >= maxDepth ) {
+
+												// Only substitute handlers pass on context
+												// and multiple values (non-spec behavior)
+												if ( handler !== Thrower ) {
+													that = undefined;
+													args = [ e ];
+												}
+
+												deferred.rejectWith( that, args );
+											}
+										}
+									};
+
+							// Support: Promises/A+ section 2.3.3.3.1
+							// https://promisesaplus.com/#point-57
+							// Re-resolve promises immediately to dodge false rejection from
+							// subsequent errors
+							if ( depth ) {
+								process();
+							} else {
+
+								// Call an optional hook to record the stack, in case of exception
+								// since it's otherwise lost when execution goes async
+								if ( jQuery.Deferred.getStackHook ) {
+									process.stackTrace = jQuery.Deferred.getStackHook();
+								}
+								window.setTimeout( process );
+							}
+						};
+					}
+
+					return jQuery.Deferred( function( newDefer ) {
+
+						// progress_handlers.add( ... )
+						tuples[ 0 ][ 3 ].add(
+							resolve(
+								0,
+								newDefer,
+								isFunction( onProgress ) ?
+									onProgress :
+									Identity,
+								newDefer.notifyWith
+							)
+						);
+
+						// fulfilled_handlers.add( ... )
+						tuples[ 1 ][ 3 ].add(
+							resolve(
+								0,
+								newDefer,
+								isFunction( onFulfilled ) ?
+									onFulfilled :
+									Identity
+							)
+						);
+
+						// rejected_handlers.add( ... )
+						tuples[ 2 ][ 3 ].add(
+							resolve(
+								0,
+								newDefer,
+								isFunction( onRejected ) ?
+									onRejected :
+									Thrower
+							)
+						);
+					} ).promise();
+				},
+
+				// Get a promise for this deferred
+				// If obj is provided, the promise aspect is added to the object
+				promise: function( obj ) {
+					return obj != null ? jQuery.extend( obj, promise ) : promise;
+				}
+			},
+			deferred = {};
+
+		// Add list-specific methods
+		jQuery.each( tuples, function( i, tuple ) {
+			var list = tuple[ 2 ],
+				stateString = tuple[ 5 ];
+
+			// promise.progress = list.add
+			// promise.done = list.add
+			// promise.fail = list.add
+			promise[ tuple[ 1 ] ] = list.add;
+
+			// Handle state
+			if ( stateString ) {
+				list.add(
+					function() {
+
+						// state = "resolved" (i.e., fulfilled)
+						// state = "rejected"
+						state = stateString;
+					},
+
+					// rejected_callbacks.disable
+					// fulfilled_callbacks.disable
+					tuples[ 3 - i ][ 2 ].disable,
+
+					// rejected_handlers.disable
+					// fulfilled_handlers.disable
+					tuples[ 3 - i ][ 3 ].disable,
+
+					// progress_callbacks.lock
+					tuples[ 0 ][ 2 ].lock,
+
+					// progress_handlers.lock
+					tuples[ 0 ][ 3 ].lock
+				);
+			}
+
+			// progress_handlers.fire
+			// fulfilled_handlers.fire
+			// rejected_handlers.fire
+			list.add( tuple[ 3 ].fire );
+
+			// deferred.notify = function() { deferred.notifyWith(...) }
+			// deferred.resolve = function() { deferred.resolveWith(...) }
+			// deferred.reject = function() { deferred.rejectWith(...) }
+			deferred[ tuple[ 0 ] ] = function() {
+				deferred[ tuple[ 0 ] + "With" ]( this === deferred ? undefined : this, arguments );
+				return this;
+			};
+
+			// deferred.notifyWith = list.fireWith
+			// deferred.resolveWith = list.fireWith
+			// deferred.rejectWith = list.fireWith
+			deferred[ tuple[ 0 ] + "With" ] = list.fireWith;
+		} );
+
+		// Make the deferred a promise
+		promise.promise( deferred );
+
+		// Call given func if any
+		if ( func ) {
+			func.call( deferred, deferred );
+		}
+
+		// All done!
+		return deferred;
+	},
+
+	// Deferred helper
+	when: function( singleValue ) {
+		var
+
+			// count of uncompleted subordinates
+			remaining = arguments.length,
+
+			// count of unprocessed arguments
+			i = remaining,
+
+			// subordinate fulfillment data
+			resolveContexts = Array( i ),
+			resolveValues = slice.call( arguments ),
+
+			// the primary Deferred
+			primary = jQuery.Deferred(),
+
+			// subordinate callback factory
+			updateFunc = function( i ) {
+				return function( value ) {
+					resolveContexts[ i ] = this;
+					resolveValues[ i ] = arguments.length > 1 ? slice.call( arguments ) : value;
+					if ( !( --remaining ) ) {
+						primary.resolveWith( resolveContexts, resolveValues );
+					}
+				};
+			};
+
+		// Single- and empty arguments are adopted like Promise.resolve
+		if ( remaining <= 1 ) {
+			adoptValue( singleValue, primary.done( updateFunc( i ) ).resolve, primary.reject,
+				!remaining );
+
+			// Use .then() to unwrap secondary thenables (cf. gh-3000)
+			if ( primary.state() === "pending" ||
+				isFunction( resolveValues[ i ] && resolveValues[ i ].then ) ) {
+
+				return primary.then();
+			}
+		}
+
+		// Multiple arguments are aggregated like Promise.all array elements
+		while ( i-- ) {
+			adoptValue( resolveValues[ i ], updateFunc( i ), primary.reject );
+		}
+
+		return primary.promise();
+	}
+} );
+
+
+// These usually indicate a programmer mistake during development,
+// warn about them ASAP rather than swallowing them by default.
+var rerrorNames = /^(Eval|Internal|Range|Reference|Syntax|Type|URI)Error$/;
+
+jQuery.Deferred.exceptionHook = function( error, stack ) {
+
+	// Support: IE 8 - 9 only
+	// Console exists when dev tools are open, which can happen at any time
+	if ( window.console && window.console.warn && error && rerrorNames.test( error.name ) ) {
+		window.console.warn( "jQuery.Deferred exception: " + error.message, error.stack, stack );
+	}
+};
+
+
+
+
+jQuery.readyException = function( error ) {
+	window.setTimeout( function() {
+		throw error;
+	} );
+};
+
+
+
+
+// The deferred used on DOM ready
+var readyList = jQuery.Deferred();
+
+jQuery.fn.ready = function( fn ) {
+
+	readyList
+		.then( fn )
+
+		// Wrap jQuery.readyException in a function so that the lookup
+		// happens at the time of error handling instead of callback
+		// registration.
+		.catch( function( error ) {
+			jQuery.readyException( error );
+		} );
+
+	return this;
+};
+
+jQuery.extend( {
+
+	// Is the DOM ready to be used? Set to true once it occurs.
+	isReady: false,
+
+	// A counter to track how many items to wait for before
+	// the ready event fires. See #6781
+	readyWait: 1,
+
+	// Handle when the DOM is ready
+	ready: function( wait ) {
+
+		// Abort if there are pending holds or we're already ready
+		if ( wait === true ? --jQuery.readyWait : jQuery.isReady ) {
+			return;
+		}
+
+		// Remember that the DOM is ready
+		jQuery.isReady = true;
+
+		// If a normal DOM Ready event fired, decrement, and wait if need be
+		if ( wait !== true && --jQuery.readyWait > 0 ) {
+			return;
+		}
+
+		// If there are functions bound, to execute
+		readyList.resolveWith( document, [ jQuery ] );
+	}
+} );
+
+jQuery.ready.then = readyList.then;
+
+// The ready event handler and self cleanup method
+function completed() {
+	document.removeEventListener( "DOMContentLoaded", completed );
+	window.removeEventListener( "load", completed );
+	jQuery.ready();
+}
+
+// Catch cases where $(document).ready() is called
+// after the browser event has already occurred.
+// Support: IE <=9 - 10 only
+// Older IE sometimes signals "interactive" too soon
+if ( document.readyState === "complete" ||
+	( document.readyState !== "loading" && !document.documentElement.doScroll ) ) {
+
+	// Handle it asynchronously to allow scripts the opportunity to delay ready
+	window.setTimeout( jQuery.ready );
+
+} else {
+
+	// Use the handy event callback
+	document.addEventListener( "DOMContentLoaded", completed );
+
+	// A fallback to window.onload, that will always work
+	window.addEventListener( "load", completed );
+}
+
+
+
+
+// Multifunctional method to get and set values of a collection
+// The value/s can optionally be executed if it's a function
+var access = function( elems, fn, key, value, chainable, emptyGet, raw ) {
+	var i = 0,
+		len = elems.length,
+		bulk = key == null;
+
+	// Sets many values
+	if ( toType( key ) === "object" ) {
+		chainable = true;
+		for ( i in key ) {
+			access( elems, fn, i, key[ i ], true, emptyGet, raw );
+		}
+
+	// Sets one value
+	} else if ( value !== undefined ) {
+		chainable = true;
+
+		if ( !isFunction( value ) ) {
+			raw = true;
+		}
+
+		if ( bulk ) {
+
+			// Bulk operations run against the entire set
+			if ( raw ) {
+				fn.call( elems, value );
+				fn = null;
+
+			// ...except when executing function values
+			} else {
+				bulk = fn;
+				fn = function( elem, _key, value ) {
+					return bulk.call( jQuery( elem ), value );
+				};
+			}
+		}
+
+		if ( fn ) {
+			for ( ; i < len; i++ ) {
+				fn(
+					elems[ i ], key, raw ?
+						value :
+						value.call( elems[ i ], i, fn( elems[ i ], key ) )
+				);
+			}
+		}
+	}
+
+	if ( chainable ) {
+		return elems;
+	}
+
+	// Gets
+	if ( bulk ) {
+		return fn.call( elems );
+	}
+
+	return len ? fn( elems[ 0 ], key ) : emptyGet;
+};
+
+
+// Matches dashed string for camelizing
+var rmsPrefix = /^-ms-/,
+	rdashAlpha = /-([a-z])/g;
+
+// Used by camelCase as callback to replace()
+function fcamelCase( _all, letter ) {
+	return letter.toUpperCase();
+}
+
+// Convert dashed to camelCase; used by the css and data modules
+// Support: IE <=9 - 11, Edge 12 - 15
+// Microsoft forgot to hump their vendor prefix (#9572)
+function camelCase( string ) {
+	return string.replace( rmsPrefix, "ms-" ).replace( rdashAlpha, fcamelCase );
+}
+var acceptData = function( owner ) {
+
+	// Accepts only:
+	//  - Node
+	//    - Node.ELEMENT_NODE
+	//    - Node.DOCUMENT_NODE
+	//  - Object
+	//    - Any
+	return owner.nodeType === 1 || owner.nodeType === 9 || !( +owner.nodeType );
+};
+
+
+
+
+function Data() {
+	this.expando = jQuery.expando + Data.uid++;
+}
+
+Data.uid = 1;
+
+Data.prototype = {
+
+	cache: function( owner ) {
+
+		// Check if the owner object already has a cache
+		var value = owner[ this.expando ];
+
+		// If not, create one
+		if ( !value ) {
+			value = {};
+
+			// We can accept data for non-element nodes in modern browsers,
+			// but we should not, see #8335.
+			// Always return an empty object.
+			if ( acceptData( owner ) ) {
+
+				// If it is a node unlikely to be stringify-ed or looped over
+				// use plain assignment
+				if ( owner.nodeType ) {
+					owner[ this.expando ] = value;
+
+				// Otherwise secure it in a non-enumerable property
+				// configurable must be true to allow the property to be
+				// deleted when data is removed
+				} else {
+					Object.defineProperty( owner, this.expando, {
+						value: value,
+						configurable: true
+					} );
+				}
+			}
+		}
+
+		return value;
+	},
+	set: function( owner, data, value ) {
+		var prop,
+			cache = this.cache( owner );
+
+		// Handle: [ owner, key, value ] args
+		// Always use camelCase key (gh-2257)
+		if ( typeof data === "string" ) {
+			cache[ camelCase( data ) ] = value;
+
+		// Handle: [ owner, { properties } ] args
+		} else {
+
+			// Copy the properties one-by-one to the cache object
+			for ( prop in data ) {
+				cache[ camelCase( prop ) ] = data[ prop ];
+			}
+		}
+		return cache;
+	},
+	get: function( owner, key ) {
+		return key === undefined ?
+			this.cache( owner ) :
+
+			// Always use camelCase key (gh-2257)
+			owner[ this.expando ] && owner[ this.expando ][ camelCase( key ) ];
+	},
+	access: function( owner, key, value ) {
+
+		// In cases where either:
+		//
+		//   1. No key was specified
+		//   2. A string key was specified, but no value provided
+		//
+		// Take the "read" path and allow the get method to determine
+		// which value to return, respectively either:
+		//
+		//   1. The entire cache object
+		//   2. The data stored at the key
+		//
+		if ( key === undefined ||
+				( ( key && typeof key === "string" ) && value === undefined ) ) {
+
+			return this.get( owner, key );
+		}
+
+		// When the key is not a string, or both a key and value
+		// are specified, set or extend (existing objects) with either:
+		//
+		//   1. An object of properties
+		//   2. A key and value
+		//
+		this.set( owner, key, value );
+
+		// Since the "set" path can have two possible entry points
+		// return the expected data based on which path was taken[*]
+		return value !== undefined ? value : key;
+	},
+	remove: function( owner, key ) {
+		var i,
+			cache = owner[ this.expando ];
+
+		if ( cache === undefined ) {
+			return;
+		}
+
+		if ( key !== undefined ) {
+
+			// Support array or space separated string of keys
+			if ( Array.isArray( key ) ) {
+
+				// If key is an array of keys...
+				// We always set camelCase keys, so remove that.
+				key = key.map( camelCase );
+			} else {
+				key = camelCase( key );
+
+				// If a key with the spaces exists, use it.
+				// Otherwise, create an array by matching non-whitespace
+				key = key in cache ?
+					[ key ] :
+					( key.match( rnothtmlwhite ) || [] );
+			}
+
+			i = key.length;
+
+			while ( i-- ) {
+				delete cache[ key[ i ] ];
+			}
+		}
+
+		// Remove the expando if there's no more data
+		if ( key === undefined || jQuery.isEmptyObject( cache ) ) {
+
+			// Support: Chrome <=35 - 45
+			// Webkit & Blink performance suffers when deleting properties
+			// from DOM nodes, so set to undefined instead
+			// https://bugs.chromium.org/p/chromium/issues/detail?id=378607 (bug restricted)
+			if ( owner.nodeType ) {
+				owner[ this.expando ] = undefined;
+			} else {
+				delete owner[ this.expando ];
+			}
+		}
+	},
+	hasData: function( owner ) {
+		var cache = owner[ this.expando ];
+		return cache !== undefined && !jQuery.isEmptyObject( cache );
+	}
+};
+var dataPriv = new Data();
+
+var dataUser = new Data();
+
+
+
+//	Implementation Summary
+//
+//	1. Enforce API surface and semantic compatibility with 1.9.x branch
+//	2. Improve the module's maintainability by reducing the storage
+//		paths to a single mechanism.
+//	3. Use the same single mechanism to support "private" and "user" data.
+//	4. _Never_ expose "private" data to user code (TODO: Drop _data, _removeData)
+//	5. Avoid exposing implementation details on user objects (eg. expando properties)
+//	6. Provide a clear path for implementation upgrade to WeakMap in 2014
+
+var rbrace = /^(?:\{[\w\W]*\}|\[[\w\W]*\])$/,
+	rmultiDash = /[A-Z]/g;
+
+function getData( data ) {
+	if ( data === "true" ) {
+		return true;
+	}
+
+	if ( data === "false" ) {
+		return false;
+	}
+
+	if ( data === "null" ) {
+		return null;
+	}
+
+	// Only convert to a number if it doesn't change the string
+	if ( data === +data + "" ) {
+		return +data;
+	}
+
+	if ( rbrace.test( data ) ) {
+		return JSON.parse( data );
+	}
+
+	return data;
+}
+
+function dataAttr( elem, key, data ) {
+	var name;
+
+	// If nothing was found internally, try to fetch any
+	// data from the HTML5 data-* attribute
+	if ( data === undefined && elem.nodeType === 1 ) {
+		name = "data-" + key.replace( rmultiDash, "-$&" ).toLowerCase();
+		data = elem.getAttribute( name );
+
+		if ( typeof data === "string" ) {
+			try {
+				data = getData( data );
+			} catch ( e ) {}
+
+			// Make sure we set the data so it isn't changed later
+			dataUser.set( elem, key, data );
+		} else {
+			data = undefined;
+		}
+	}
+	return data;
+}
+
+jQuery.extend( {
+	hasData: function( elem ) {
+		return dataUser.hasData( elem ) || dataPriv.hasData( elem );
+	},
+
+	data: function( elem, name, data ) {
+		return dataUser.access( elem, name, data );
+	},
+
+	removeData: function( elem, name ) {
+		dataUser.remove( elem, name );
+	},
+
+	// TODO: Now that all calls to _data and _removeData have been replaced
+	// with direct calls to dataPriv methods, these can be deprecated.
+	_data: function( elem, name, data ) {
+		return dataPriv.access( elem, name, data );
+	},
+
+	_removeData: function( elem, name ) {
+		dataPriv.remove( elem, name );
+	}
+} );
+
+jQuery.fn.extend( {
+	data: function( key, value ) {
+		var i, name, data,
+			elem = this[ 0 ],
+			attrs = elem && elem.attributes;
+
+		// Gets all values
+		if ( key === undefined ) {
+			if ( this.length ) {
+				data = dataUser.get( elem );
+
+				if ( elem.nodeType === 1 && !dataPriv.get( elem, "hasDataAttrs" ) ) {
+					i = attrs.length;
+					while ( i-- ) {
+
+						// Support: IE 11 only
+						// The attrs elements can be null (#14894)
+						if ( attrs[ i ] ) {
+							name = attrs[ i ].name;
+							if ( name.indexOf( "data-" ) === 0 ) {
+								name = camelCase( name.slice( 5 ) );
+								dataAttr( elem, name, data[ name ] );
+							}
+						}
+					}
+					dataPriv.set( elem, "hasDataAttrs", true );
+				}
+			}
+
+			return data;
+		}
+
+		// Sets multiple values
+		if ( typeof key === "object" ) {
+			return this.each( function() {
+				dataUser.set( this, key );
+			} );
+		}
+
+		return access( this, function( value ) {
+			var data;
+
+			// The calling jQuery object (element matches) is not empty
+			// (and therefore has an element appears at this[ 0 ]) and the
+			// `value` parameter was not undefined. An empty jQuery object
+			// will result in `undefined` for elem = this[ 0 ] which will
+			// throw an exception if an attempt to read a data cache is made.
+			if ( elem && value === undefined ) {
+
+				// Attempt to get data from the cache
+				// The key will always be camelCased in Data
+				data = dataUser.get( elem, key );
+				if ( data !== undefined ) {
+					return data;
+				}
+
+				// Attempt to "discover" the data in
+				// HTML5 custom data-* attrs
+				data = dataAttr( elem, key );
+				if ( data !== undefined ) {
+					return data;
+				}
+
+				// We tried really hard, but the data doesn't exist.
+				return;
+			}
+
+			// Set the data...
+			this.each( function() {
+
+				// We always store the camelCased key
+				dataUser.set( this, key, value );
+			} );
+		}, null, value, arguments.length > 1, null, true );
+	},
+
+	removeData: function( key ) {
+		return this.each( function() {
+			dataUser.remove( this, key );
+		} );
+	}
+} );
+
+
+jQuery.extend( {
+	queue: function( elem, type, data ) {
+		var queue;
+
+		if ( elem ) {
+			type = ( type || "fx" ) + "queue";
+			queue = dataPriv.get( elem, type );
+
+			// Speed up dequeue by getting out quickly if this is just a lookup
+			if ( data ) {
+				if ( !queue || Array.isArray( data ) ) {
+					queue = dataPriv.access( elem, type, jQuery.makeArray( data ) );
+				} else {
+					queue.push( data );
+				}
+			}
+			return queue || [];
+		}
+	},
+
+	dequeue: function( elem, type ) {
+		type = type || "fx";
+
+		var queue = jQuery.queue( elem, type ),
+			startLength = queue.length,
+			fn = queue.shift(),
+			hooks = jQuery._queueHooks( elem, type ),
+			next = function() {
+				jQuery.dequeue( elem, type );
+			};
+
+		// If the fx queue is dequeued, always remove the progress sentinel
+		if ( fn === "inprogress" ) {
+			fn = queue.shift();
+			startLength--;
+		}
+
+		if ( fn ) {
+
+			// Add a progress sentinel to prevent the fx queue from being
+			// automatically dequeued
+			if ( type === "fx" ) {
+				queue.unshift( "inprogress" );
+			}
+
+			// Clear up the last queue stop function
+			delete hooks.stop;
+			fn.call( elem, next, hooks );
+		}
+
+		if ( !startLength && hooks ) {
+			hooks.empty.fire();
+		}
+	},
+
+	// Not public - generate a queueHooks object, or return the current one
+	_queueHooks: function( elem, type ) {
+		var key = type + "queueHooks";
+		return dataPriv.get( elem, key ) || dataPriv.access( elem, key, {
+			empty: jQuery.Callbacks( "once memory" ).add( function() {
+				dataPriv.remove( elem, [ type + "queue", key ] );
+			} )
+		} );
+	}
+} );
+
+jQuery.fn.extend( {
+	queue: function( type, data ) {
+		var setter = 2;
+
+		if ( typeof type !== "string" ) {
+			data = type;
+			type = "fx";
+			setter--;
+		}
+
+		if ( arguments.length < setter ) {
+			return jQuery.queue( this[ 0 ], type );
+		}
+
+		return data === undefined ?
+			this :
+			this.each( function() {
+				var queue = jQuery.queue( this, type, data );
+
+				// Ensure a hooks for this queue
+				jQuery._queueHooks( this, type );
+
+				if ( type === "fx" && queue[ 0 ] !== "inprogress" ) {
+					jQuery.dequeue( this, type );
+				}
+			} );
+	},
+	dequeue: function( type ) {
+		return this.each( function() {
+			jQuery.dequeue( this, type );
+		} );
+	},
+	clearQueue: function( type ) {
+		return this.queue( type || "fx", [] );
+	},
+
+	// Get a promise resolved when queues of a certain type
+	// are emptied (fx is the type by default)
+	promise: function( type, obj ) {
+		var tmp,
+			count = 1,
+			defer = jQuery.Deferred(),
+			elements = this,
+			i = this.length,
+			resolve = function() {
+				if ( !( --count ) ) {
+					defer.resolveWith( elements, [ elements ] );
+				}
+			};
+
+		if ( typeof type !== "string" ) {
+			obj = type;
+			type = undefined;
+		}
+		type = type || "fx";
+
+		while ( i-- ) {
+			tmp = dataPriv.get( elements[ i ], type + "queueHooks" );
+			if ( tmp && tmp.empty ) {
+				count++;
+				tmp.empty.add( resolve );
+			}
+		}
+		resolve();
+		return defer.promise( obj );
+	}
+} );
+var pnum = ( /[+-]?(?:\d*\.|)\d+(?:[eE][+-]?\d+|)/ ).source;
+
+var rcssNum = new RegExp( "^(?:([+-])=|)(" + pnum + ")([a-z%]*)$", "i" );
+
+
+var cssExpand = [ "Top", "Right", "Bottom", "Left" ];
+
+var documentElement = document.documentElement;
+
+
+
+	var isAttached = function( elem ) {
+			return jQuery.contains( elem.ownerDocument, elem );
+		},
+		composed = { composed: true };
+
+	// Support: IE 9 - 11+, Edge 12 - 18+, iOS 10.0 - 10.2 only
+	// Check attachment across shadow DOM boundaries when possible (gh-3504)
+	// Support: iOS 10.0-10.2 only
+	// Early iOS 10 versions support `attachShadow` but not `getRootNode`,
+	// leading to errors. We need to check for `getRootNode`.
+	if ( documentElement.getRootNode ) {
+		isAttached = function( elem ) {
+			return jQuery.contains( elem.ownerDocument, elem ) ||
+				elem.getRootNode( composed ) === elem.ownerDocument;
+		};
+	}
+var isHiddenWithinTree = function( elem, el ) {
+
+		// isHiddenWithinTree might be called from jQuery#filter function;
+		// in that case, element will be second argument
+		elem = el || elem;
+
+		// Inline style trumps all
+		return elem.style.display === "none" ||
+			elem.style.display === "" &&
+
+			// Otherwise, check computed style
+			// Support: Firefox <=43 - 45
+			// Disconnected elements can have computed display: none, so first confirm that elem is
+			// in the document.
+			isAttached( elem ) &&
+
+			jQuery.css( elem, "display" ) === "none";
+	};
+
+
+
+function adjustCSS( elem, prop, valueParts, tween ) {
+	var adjusted, scale,
+		maxIterations = 20,
+		currentValue = tween ?
+			function() {
+				return tween.cur();
+			} :
+			function() {
+				return jQuery.css( elem, prop, "" );
+			},
+		initial = currentValue(),
+		unit = valueParts && valueParts[ 3 ] || ( jQuery.cssNumber[ prop ] ? "" : "px" ),
+
+		// Starting value computation is required for potential unit mismatches
+		initialInUnit = elem.nodeType &&
+			( jQuery.cssNumber[ prop ] || unit !== "px" && +initial ) &&
+			rcssNum.exec( jQuery.css( elem, prop ) );
+
+	if ( initialInUnit && initialInUnit[ 3 ] !== unit ) {
+
+		// Support: Firefox <=54
+		// Halve the iteration target value to prevent interference from CSS upper bounds (gh-2144)
+		initial = initial / 2;
+
+		// Trust units reported by jQuery.css
+		unit = unit || initialInUnit[ 3 ];
+
+		// Iteratively approximate from a nonzero starting point
+		initialInUnit = +initial || 1;
+
+		while ( maxIterations-- ) {
+
+			// Evaluate and update our best guess (doubling guesses that zero out).
+			// Finish if the scale equals or crosses 1 (making the old*new product non-positive).
+			jQuery.style( elem, prop, initialInUnit + unit );
+			if ( ( 1 - scale ) * ( 1 - ( scale = currentValue() / initial || 0.5 ) ) <= 0 ) {
+				maxIterations = 0;
+			}
+			initialInUnit = initialInUnit / scale;
+
+		}
+
+		initialInUnit = initialInUnit * 2;
+		jQuery.style( elem, prop, initialInUnit + unit );
+
+		// Make sure we update the tween properties later on
+		valueParts = valueParts || [];
+	}
+
+	if ( valueParts ) {
+		initialInUnit = +initialInUnit || +initial || 0;
+
+		// Apply relative offset (+=/-=) if specified
+		adjusted = valueParts[ 1 ] ?
+			initialInUnit + ( valueParts[ 1 ] + 1 ) * valueParts[ 2 ] :
+			+valueParts[ 2 ];
+		if ( tween ) {
+			tween.unit = unit;
+			tween.start = initialInUnit;
+			tween.end = adjusted;
+		}
+	}
+	return adjusted;
+}
+
+
+var defaultDisplayMap = {};
+
+function getDefaultDisplay( elem ) {
+	var temp,
+		doc = elem.ownerDocument,
+		nodeName = elem.nodeName,
+		display = defaultDisplayMap[ nodeName ];
+
+	if ( display ) {
+		return display;
+	}
+
+	temp = doc.body.appendChild( doc.createElement( nodeName ) );
+	display = jQuery.css( temp, "display" );
+
+	temp.parentNode.removeChild( temp );
+
+	if ( display === "none" ) {
+		display = "block";
+	}
+	defaultDisplayMap[ nodeName ] = display;
+
+	return display;
+}
+
+function showHide( elements, show ) {
+	var display, elem,
+		values = [],
+		index = 0,
+		length = elements.length;
+
+	// Determine new display value for elements that need to change
+	for ( ; index < length; index++ ) {
+		elem = elements[ index ];
+		if ( !elem.style ) {
+			continue;
+		}
+
+		display = elem.style.display;
+		if ( show ) {
+
+			// Since we force visibility upon cascade-hidden elements, an immediate (and slow)
+			// check is required in this first loop unless we have a nonempty display value (either
+			// inline or about-to-be-restored)
+			if ( display === "none" ) {
+				values[ index ] = dataPriv.get( elem, "display" ) || null;
+				if ( !values[ index ] ) {
+					elem.style.display = "";
+				}
+			}
+			if ( elem.style.display === "" && isHiddenWithinTree( elem ) ) {
+				values[ index ] = getDefaultDisplay( elem );
+			}
+		} else {
+			if ( display !== "none" ) {
+				values[ index ] = "none";
+
+				// Remember what we're overwriting
+				dataPriv.set( elem, "display", display );
+			}
+		}
+	}
+
+	// Set the display of the elements in a second loop to avoid constant reflow
+	for ( index = 0; index < length; index++ ) {
+		if ( values[ index ] != null ) {
+			elements[ index ].style.display = values[ index ];
+		}
+	}
+
+	return elements;
+}
+
+jQuery.fn.extend( {
+	show: function() {
+		return showHide( this, true );
+	},
+	hide: function() {
+		return showHide( this );
+	},
+	toggle: function( state ) {
+		if ( typeof state === "boolean" ) {
+			return state ? this.show() : this.hide();
+		}
+
+		return this.each( function() {
+			if ( isHiddenWithinTree( this ) ) {
+				jQuery( this ).show();
+			} else {
+				jQuery( this ).hide();
+			}
+		} );
+	}
+} );
+var rcheckableType = ( /^(?:checkbox|radio)$/i );
+
+var rtagName = ( /<([a-z][^\/\0>\x20\t\r\n\f]*)/i );
+
+var rscriptType = ( /^$|^module$|\/(?:java|ecma)script/i );
+
+
+
+( function() {
+	var fragment = document.createDocumentFragment(),
+		div = fragment.appendChild( document.createElement( "div" ) ),
+		input = document.createElement( "input" );
+
+	// Support: Android 4.0 - 4.3 only
+	// Check state lost if the name is set (#11217)
+	// Support: Windows Web Apps (WWA)
+	// `name` and `type` must use .setAttribute for WWA (#14901)
+	input.setAttribute( "type", "radio" );
+	input.setAttribute( "checked", "checked" );
+	input.setAttribute( "name", "t" );
+
+	div.appendChild( input );
+
+	// Support: Android <=4.1 only
+	// Older WebKit doesn't clone checked state correctly in fragments
+	support.checkClone = div.cloneNode( true ).cloneNode( true ).lastChild.checked;
+
+	// Support: IE <=11 only
+	// Make sure textarea (and checkbox) defaultValue is properly cloned
+	div.innerHTML = "<textarea>x</textarea>";
+	support.noCloneChecked = !!div.cloneNode( true ).lastChild.defaultValue;
+
+	// Support: IE <=9 only
+	// IE <=9 replaces <option> tags with their contents when inserted outside of
+	// the select element.
+	div.innerHTML = "<option></option>";
+	support.option = !!div.lastChild;
+} )();
+
+
+// We have to close these tags to support XHTML (#13200)
+var wrapMap = {
+
+	// XHTML parsers do not magically insert elements in the
+	// same way that tag soup parsers do. So we cannot shorten
+	// this by omitting <tbody> or other required elements.
+	thead: [ 1, "<table>", "</table>" ],
+	col: [ 2, "<table><colgroup>", "</colgroup></table>" ],
+	tr: [ 2, "<table><tbody>", "</tbody></table>" ],
+	td: [ 3, "<table><tbody><tr>", "</tr></tbody></table>" ],
+
+	_default: [ 0, "", "" ]
+};
+
+wrapMap.tbody = wrapMap.tfoot = wrapMap.colgroup = wrapMap.caption = wrapMap.thead;
+wrapMap.th = wrapMap.td;
+
+// Support: IE <=9 only
+if ( !support.option ) {
+	wrapMap.optgroup = wrapMap.option = [ 1, "<select multiple='multiple'>", "</select>" ];
+}
+
+
+function getAll( context, tag ) {
+
+	// Support: IE <=9 - 11 only
+	// Use typeof to avoid zero-argument method invocation on host objects (#15151)
+	var ret;
+
+	if ( typeof context.getElementsByTagName !== "undefined" ) {
+		ret = context.getElementsByTagName( tag || "*" );
+
+	} else if ( typeof context.querySelectorAll !== "undefined" ) {
+		ret = context.querySelectorAll( tag || "*" );
+
+	} else {
+		ret = [];
+	}
+
+	if ( tag === undefined || tag && nodeName( context, tag ) ) {
+		return jQuery.merge( [ context ], ret );
+	}
+
+	return ret;
+}
+
+
+// Mark scripts as having already been evaluated
+function setGlobalEval( elems, refElements ) {
+	var i = 0,
+		l = elems.length;
+
+	for ( ; i < l; i++ ) {
+		dataPriv.set(
+			elems[ i ],
+			"globalEval",
+			!refElements || dataPriv.get( refElements[ i ], "globalEval" )
+		);
+	}
+}
+
+
+var rhtml = /<|&#?\w+;/;
+
+function buildFragment( elems, context, scripts, selection, ignored ) {
+	var elem, tmp, tag, wrap, attached, j,
+		fragment = context.createDocumentFragment(),
+		nodes = [],
+		i = 0,
+		l = elems.length;
+
+	for ( ; i < l; i++ ) {
+		elem = elems[ i ];
+
+		if ( elem || elem === 0 ) {
+
+			// Add nodes directly
+			if ( toType( elem ) === "object" ) {
+
+				// Support: Android <=4.0 only, PhantomJS 1 only
+				// push.apply(_, arraylike) throws on ancient WebKit
+				jQuery.merge( nodes, elem.nodeType ? [ elem ] : elem );
+
+			// Convert non-html into a text node
+			} else if ( !rhtml.test( elem ) ) {
+				nodes.push( context.createTextNode( elem ) );
+
+			// Convert html into DOM nodes
+			} else {
+				tmp = tmp || fragment.appendChild( context.createElement( "div" ) );
+
+				// Deserialize a standard representation
+				tag = ( rtagName.exec( elem ) || [ "", "" ] )[ 1 ].toLowerCase();
+				wrap = wrapMap[ tag ] || wrapMap._default;
+				tmp.innerHTML = wrap[ 1 ] + jQuery.htmlPrefilter( elem ) + wrap[ 2 ];
+
+				// Descend through wrappers to the right content
+				j = wrap[ 0 ];
+				while ( j-- ) {
+					tmp = tmp.lastChild;
+				}
+
+				// Support: Android <=4.0 only, PhantomJS 1 only
+				// push.apply(_, arraylike) throws on ancient WebKit
+				jQuery.merge( nodes, tmp.childNodes );
+
+				// Remember the top-level container
+				tmp = fragment.firstChild;
+
+				// Ensure the created nodes are orphaned (#12392)
+				tmp.textContent = "";
+			}
+		}
+	}
+
+	// Remove wrapper from fragment
+	fragment.textContent = "";
+
+	i = 0;
+	while ( ( elem = nodes[ i++ ] ) ) {
+
+		// Skip elements already in the context collection (trac-4087)
+		if ( selection && jQuery.inArray( elem, selection ) > -1 ) {
+			if ( ignored ) {
+				ignored.push( elem );
+			}
+			continue;
+		}
+
+		attached = isAttached( elem );
+
+		// Append to fragment
+		tmp = getAll( fragment.appendChild( elem ), "script" );
+
+		// Preserve script evaluation history
+		if ( attached ) {
+			setGlobalEval( tmp );
+		}
+
+		// Capture executables
+		if ( scripts ) {
+			j = 0;
+			while ( ( elem = tmp[ j++ ] ) ) {
+				if ( rscriptType.test( elem.type || "" ) ) {
+					scripts.push( elem );
+				}
+			}
+		}
+	}
+
+	return fragment;
+}
+
+
+var rtypenamespace = /^([^.]*)(?:\.(.+)|)/;
+
+function returnTrue() {
+	return true;
+}
+
+function returnFalse() {
+	return false;
+}
+
+// Support: IE <=9 - 11+
+// focus() and blur() are asynchronous, except when they are no-op.
+// So expect focus to be synchronous when the element is already active,
+// and blur to be synchronous when the element is not already active.
+// (focus and blur are always synchronous in other supported browsers,
+// this just defines when we can count on it).
+function expectSync( elem, type ) {
+	return ( elem === safeActiveElement() ) === ( type === "focus" );
+}
+
+// Support: IE <=9 only
+// Accessing document.activeElement can throw unexpectedly
+// https://bugs.jquery.com/ticket/13393
+function safeActiveElement() {
+	try {
+		return document.activeElement;
+	} catch ( err ) { }
+}
+
+function on( elem, types, selector, data, fn, one ) {
+	var origFn, type;
+
+	// Types can be a map of types/handlers
+	if ( typeof types === "object" ) {
+
+		// ( types-Object, selector, data )
+		if ( typeof selector !== "string" ) {
+
+			// ( types-Object, data )
+			data = data || selector;
+			selector = undefined;
+		}
+		for ( type in types ) {
+			on( elem, type, selector, data, types[ type ], one );
+		}
+		return elem;
+	}
+
+	if ( data == null && fn == null ) {
+
+		// ( types, fn )
+		fn = selector;
+		data = selector = undefined;
+	} else if ( fn == null ) {
+		if ( typeof selector === "string" ) {
+
+			// ( types, selector, fn )
+			fn = data;
+			data = undefined;
+		} else {
+
+			// ( types, data, fn )
+			fn = data;
+			data = selector;
+			selector = undefined;
+		}
+	}
+	if ( fn === false ) {
+		fn = returnFalse;
+	} else if ( !fn ) {
+		return elem;
+	}
+
+	if ( one === 1 ) {
+		origFn = fn;
+		fn = function( event ) {
+
+			// Can use an empty set, since event contains the info
+			jQuery().off( event );
+			return origFn.apply( this, arguments );
+		};
+
+		// Use same guid so caller can remove using origFn
+		fn.guid = origFn.guid || ( origFn.guid = jQuery.guid++ );
+	}
+	return elem.each( function() {
+		jQuery.event.add( this, types, fn, data, selector );
+	} );
+}
+
+/*
+ * Helper functions for managing events -- not part of the public interface.
+ * Props to Dean Edwards' addEvent library for many of the ideas.
+ */
+jQuery.event = {
+
+	global: {},
+
+	add: function( elem, types, handler, data, selector ) {
+
+		var handleObjIn, eventHandle, tmp,
+			events, t, handleObj,
+			special, handlers, type, namespaces, origType,
+			elemData = dataPriv.get( elem );
+
+		// Only attach events to objects that accept data
+		if ( !acceptData( elem ) ) {
+			return;
+		}
+
+		// Caller can pass in an object of custom data in lieu of the handler
+		if ( handler.handler ) {
+			handleObjIn = handler;
+			handler = handleObjIn.handler;
+			selector = handleObjIn.selector;
+		}
+
+		// Ensure that invalid selectors throw exceptions at attach time
+		// Evaluate against documentElement in case elem is a non-element node (e.g., document)
+		if ( selector ) {
+			jQuery.find.matchesSelector( documentElement, selector );
+		}
+
+		// Make sure that the handler has a unique ID, used to find/remove it later
+		if ( !handler.guid ) {
+			handler.guid = jQuery.guid++;
+		}
+
+		// Init the element's event structure and main handler, if this is the first
+		if ( !( events = elemData.events ) ) {
+			events = elemData.events = Object.create( null );
+		}
+		if ( !( eventHandle = elemData.handle ) ) {
+			eventHandle = elemData.handle = function( e ) {
+
+				// Discard the second event of a jQuery.event.trigger() and
+				// when an event is called after a page has unloaded
+				return typeof jQuery !== "undefined" && jQuery.event.triggered !== e.type ?
+					jQuery.event.dispatch.apply( elem, arguments ) : undefined;
+			};
+		}
+
+		// Handle multiple events separated by a space
+		types = ( types || "" ).match( rnothtmlwhite ) || [ "" ];
+		t = types.length;
+		while ( t-- ) {
+			tmp = rtypenamespace.exec( types[ t ] ) || [];
+			type = origType = tmp[ 1 ];
+			namespaces = ( tmp[ 2 ] || "" ).split( "." ).sort();
+
+			// There *must* be a type, no attaching namespace-only handlers
+			if ( !type ) {
+				continue;
+			}
+
+			// If event changes its type, use the special event handlers for the changed type
+			special = jQuery.event.special[ type ] || {};
+
+			// If selector defined, determine special event api type, otherwise given type
+			type = ( selector ? special.delegateType : special.bindType ) || type;
+
+			// Update special based on newly reset type
+			special = jQuery.event.special[ type ] || {};
+
+			// handleObj is passed to all event handlers
+			handleObj = jQuery.extend( {
+				type: type,
+				origType: origType,
+				data: data,
+				handler: handler,
+				guid: handler.guid,
+				selector: selector,
+				needsContext: selector && jQuery.expr.match.needsContext.test( selector ),
+				namespace: namespaces.join( "." )
+			}, handleObjIn );
+
+			// Init the event handler queue if we're the first
+			if ( !( handlers = events[ type ] ) ) {
+				handlers = events[ type ] = [];
+				handlers.delegateCount = 0;
+
+				// Only use addEventListener if the special events handler returns false
+				if ( !special.setup ||
+					special.setup.call( elem, data, namespaces, eventHandle ) === false ) {
+
+					if ( elem.addEventListener ) {
+						elem.addEventListener( type, eventHandle );
+					}
+				}
+			}
+
+			if ( special.add ) {
+				special.add.call( elem, handleObj );
+
+				if ( !handleObj.handler.guid ) {
+					handleObj.handler.guid = handler.guid;
+				}
+			}
+
+			// Add to the element's handler list, delegates in front
+			if ( selector ) {
+				handlers.splice( handlers.delegateCount++, 0, handleObj );
+			} else {
+				handlers.push( handleObj );
+			}
+
+			// Keep track of which events have ever been used, for event optimization
+			jQuery.event.global[ type ] = true;
+		}
+
+	},
+
+	// Detach an event or set of events from an element
+	remove: function( elem, types, handler, selector, mappedTypes ) {
+
+		var j, origCount, tmp,
+			events, t, handleObj,
+			special, handlers, type, namespaces, origType,
+			elemData = dataPriv.hasData( elem ) && dataPriv.get( elem );
+
+		if ( !elemData || !( events = elemData.events ) ) {
+			return;
+		}
+
+		// Once for each type.namespace in types; type may be omitted
+		types = ( types || "" ).match( rnothtmlwhite ) || [ "" ];
+		t = types.length;
+		while ( t-- ) {
+			tmp = rtypenamespace.exec( types[ t ] ) || [];
+			type = origType = tmp[ 1 ];
+			namespaces = ( tmp[ 2 ] || "" ).split( "." ).sort();
+
+			// Unbind all events (on this namespace, if provided) for the element
+			if ( !type ) {
+				for ( type in events ) {
+					jQuery.event.remove( elem, type + types[ t ], handler, selector, true );
+				}
+				continue;
+			}
+
+			special = jQuery.event.special[ type ] || {};
+			type = ( selector ? special.delegateType : special.bindType ) || type;
+			handlers = events[ type ] || [];
+			tmp = tmp[ 2 ] &&
+				new RegExp( "(^|\\.)" + namespaces.join( "\\.(?:.*\\.|)" ) + "(\\.|$)" );
+
+			// Remove matching events
+			origCount = j = handlers.length;
+			while ( j-- ) {
+				handleObj = handlers[ j ];
+
+				if ( ( mappedTypes || origType === handleObj.origType ) &&
+					( !handler || handler.guid === handleObj.guid ) &&
+					( !tmp || tmp.test( handleObj.namespace ) ) &&
+					( !selector || selector === handleObj.selector ||
+						selector === "**" && handleObj.selector ) ) {
+					handlers.splice( j, 1 );
+
+					if ( handleObj.selector ) {
+						handlers.delegateCount--;
+					}
+					if ( special.remove ) {
+						special.remove.call( elem, handleObj );
+					}
+				}
+			}
+
+			// Remove generic event handler if we removed something and no more handlers exist
+			// (avoids potential for endless recursion during removal of special event handlers)
+			if ( origCount && !handlers.length ) {
+				if ( !special.teardown ||
+					special.teardown.call( elem, namespaces, elemData.handle ) === false ) {
+
+					jQuery.removeEvent( elem, type, elemData.handle );
+				}
+
+				delete events[ type ];
+			}
+		}
+
+		// Remove data and the expando if it's no longer used
+		if ( jQuery.isEmptyObject( events ) ) {
+			dataPriv.remove( elem, "handle events" );
+		}
+	},
+
+	dispatch: function( nativeEvent ) {
+
+		var i, j, ret, matched, handleObj, handlerQueue,
+			args = new Array( arguments.length ),
+
+			// Make a writable jQuery.Event from the native event object
+			event = jQuery.event.fix( nativeEvent ),
+
+			handlers = (
+				dataPriv.get( this, "events" ) || Object.create( null )
+			)[ event.type ] || [],
+			special = jQuery.event.special[ event.type ] || {};
+
+		// Use the fix-ed jQuery.Event rather than the (read-only) native event
+		args[ 0 ] = event;
+
+		for ( i = 1; i < arguments.length; i++ ) {
+			args[ i ] = arguments[ i ];
+		}
+
+		event.delegateTarget = this;
+
+		// Call the preDispatch hook for the mapped type, and let it bail if desired
+		if ( special.preDispatch && special.preDispatch.call( this, event ) === false ) {
+			return;
+		}
+
+		// Determine handlers
+		handlerQueue = jQuery.event.handlers.call( this, event, handlers );
+
+		// Run delegates first; they may want to stop propagation beneath us
+		i = 0;
+		while ( ( matched = handlerQueue[ i++ ] ) && !event.isPropagationStopped() ) {
+			event.currentTarget = matched.elem;
+
+			j = 0;
+			while ( ( handleObj = matched.handlers[ j++ ] ) &&
+				!event.isImmediatePropagationStopped() ) {
+
+				// If the event is namespaced, then each handler is only invoked if it is
+				// specially universal or its namespaces are a superset of the event's.
+				if ( !event.rnamespace || handleObj.namespace === false ||
+					event.rnamespace.test( handleObj.namespace ) ) {
+
+					event.handleObj = handleObj;
+					event.data = handleObj.data;
+
+					ret = ( ( jQuery.event.special[ handleObj.origType ] || {} ).handle ||
+						handleObj.handler ).apply( matched.elem, args );
+
+					if ( ret !== undefined ) {
+						if ( ( event.result = ret ) === false ) {
+							event.preventDefault();
+							event.stopPropagation();
+						}
+					}
+				}
+			}
+		}
+
+		// Call the postDispatch hook for the mapped type
+		if ( special.postDispatch ) {
+			special.postDispatch.call( this, event );
+		}
+
+		return event.result;
+	},
+
+	handlers: function( event, handlers ) {
+		var i, handleObj, sel, matchedHandlers, matchedSelectors,
+			handlerQueue = [],
+			delegateCount = handlers.delegateCount,
+			cur = event.target;
+
+		// Find delegate handlers
+		if ( delegateCount &&
+
+			// Support: IE <=9
+			// Black-hole SVG <use> instance trees (trac-13180)
+			cur.nodeType &&
+
+			// Support: Firefox <=42
+			// Suppress spec-violating clicks indicating a non-primary pointer button (trac-3861)
+			// https://www.w3.org/TR/DOM-Level-3-Events/#event-type-click
+			// Support: IE 11 only
+			// ...but not arrow key "clicks" of radio inputs, which can have `button` -1 (gh-2343)
+			!( event.type === "click" && event.button >= 1 ) ) {
+
+			for ( ; cur !== this; cur = cur.parentNode || this ) {
+
+				// Don't check non-elements (#13208)
+				// Don't process clicks on disabled elements (#6911, #8165, #11382, #11764)
+				if ( cur.nodeType === 1 && !( event.type === "click" && cur.disabled === true ) ) {
+					matchedHandlers = [];
+					matchedSelectors = {};
+					for ( i = 0; i < delegateCount; i++ ) {
+						handleObj = handlers[ i ];
+
+						// Don't conflict with Object.prototype properties (#13203)
+						sel = handleObj.selector + " ";
+
+						if ( matchedSelectors[ sel ] === undefined ) {
+							matchedSelectors[ sel ] = handleObj.needsContext ?
+								jQuery( sel, this ).index( cur ) > -1 :
+								jQuery.find( sel, this, null, [ cur ] ).length;
+						}
+						if ( matchedSelectors[ sel ] ) {
+							matchedHandlers.push( handleObj );
+						}
+					}
+					if ( matchedHandlers.length ) {
+						handlerQueue.push( { elem: cur, handlers: matchedHandlers } );
+					}
+				}
+			}
+		}
+
+		// Add the remaining (directly-bound) handlers
+		cur = this;
+		if ( delegateCount < handlers.length ) {
+			handlerQueue.push( { elem: cur, handlers: handlers.slice( delegateCount ) } );
+		}
+
+		return handlerQueue;
+	},
+
+	addProp: function( name, hook ) {
+		Object.defineProperty( jQuery.Event.prototype, name, {
+			enumerable: true,
+			configurable: true,
+
+			get: isFunction( hook ) ?
+				function() {
+					if ( this.originalEvent ) {
+						return hook( this.originalEvent );
+					}
+				} :
+				function() {
+					if ( this.originalEvent ) {
+						return this.originalEvent[ name ];
+					}
+				},
+
+			set: function( value ) {
+				Object.defineProperty( this, name, {
+					enumerable: true,
+					configurable: true,
+					writable: true,
+					value: value
+				} );
+			}
+		} );
+	},
+
+	fix: function( originalEvent ) {
+		return originalEvent[ jQuery.expando ] ?
+			originalEvent :
+			new jQuery.Event( originalEvent );
+	},
+
+	special: {
+		load: {
+
+			// Prevent triggered image.load events from bubbling to window.load
+			noBubble: true
+		},
+		click: {
+
+			// Utilize native event to ensure correct state for checkable inputs
+			setup: function( data ) {
+
+				// For mutual compressibility with _default, replace `this` access with a local var.
+				// `|| data` is dead code meant only to preserve the variable through minification.
+				var el = this || data;
+
+				// Claim the first handler
+				if ( rcheckableType.test( el.type ) &&
+					el.click && nodeName( el, "input" ) ) {
+
+					// dataPriv.set( el, "click", ... )
+					leverageNative( el, "click", returnTrue );
+				}
+
+				// Return false to allow normal processing in the caller
+				return false;
+			},
+			trigger: function( data ) {
+
+				// For mutual compressibility with _default, replace `this` access with a local var.
+				// `|| data` is dead code meant only to preserve the variable through minification.
+				var el = this || data;
+
+				// Force setup before triggering a click
+				if ( rcheckableType.test( el.type ) &&
+					el.click && nodeName( el, "input" ) ) {
+
+					leverageNative( el, "click" );
+				}
+
+				// Return non-false to allow normal event-path propagation
+				return true;
+			},
+
+			// For cross-browser consistency, suppress native .click() on links
+			// Also prevent it if we're currently inside a leveraged native-event stack
+			_default: function( event ) {
+				var target = event.target;
+				return rcheckableType.test( target.type ) &&
+					target.click && nodeName( target, "input" ) &&
+					dataPriv.get( target, "click" ) ||
+					nodeName( target, "a" );
+			}
+		},
+
+		beforeunload: {
+			postDispatch: function( event ) {
+
+				// Support: Firefox 20+
+				// Firefox doesn't alert if the returnValue field is not set.
+				if ( event.result !== undefined && event.originalEvent ) {
+					event.originalEvent.returnValue = event.result;
+				}
+			}
+		}
+	}
+};
+
+// Ensure the presence of an event listener that handles manually-triggered
+// synthetic events by interrupting progress until reinvoked in response to
+// *native* events that it fires directly, ensuring that state changes have
+// already occurred before other listeners are invoked.
+function leverageNative( el, type, expectSync ) {
+
+	// Missing expectSync indicates a trigger call, which must force setup through jQuery.event.add
+	if ( !expectSync ) {
+		if ( dataPriv.get( el, type ) === undefined ) {
+			jQuery.event.add( el, type, returnTrue );
+		}
+		return;
+	}
+
+	// Register the controller as a special universal handler for all event namespaces
+	dataPriv.set( el, type, false );
+	jQuery.event.add( el, type, {
+		namespace: false,
+		handler: function( event ) {
+			var notAsync, result,
+				saved = dataPriv.get( this, type );
+
+			if ( ( event.isTrigger & 1 ) && this[ type ] ) {
+
+				// Interrupt processing of the outer synthetic .trigger()ed event
+				// Saved data should be false in such cases, but might be a leftover capture object
+				// from an async native handler (gh-4350)
+				if ( !saved.length ) {
+
+					// Store arguments for use when handling the inner native event
+					// There will always be at least one argument (an event object), so this array
+					// will not be confused with a leftover capture object.
+					saved = slice.call( arguments );
+					dataPriv.set( this, type, saved );
+
+					// Trigger the native event and capture its result
+					// Support: IE <=9 - 11+
+					// focus() and blur() are asynchronous
+					notAsync = expectSync( this, type );
+					this[ type ]();
+					result = dataPriv.get( this, type );
+					if ( saved !== result || notAsync ) {
+						dataPriv.set( this, type, false );
+					} else {
+						result = {};
+					}
+					if ( saved !== result ) {
+
+						// Cancel the outer synthetic event
+						event.stopImmediatePropagation();
+						event.preventDefault();
+
+						// Support: Chrome 86+
+						// In Chrome, if an element having a focusout handler is blurred by
+						// clicking outside of it, it invokes the handler synchronously. If
+						// that handler calls `.remove()` on the element, the data is cleared,
+						// leaving `result` undefined. We need to guard against this.
+						return result && result.value;
+					}
+
+				// If this is an inner synthetic event for an event with a bubbling surrogate
+				// (focus or blur), assume that the surrogate already propagated from triggering the
+				// native event and prevent that from happening again here.
+				// This technically gets the ordering wrong w.r.t. to `.trigger()` (in which the
+				// bubbling surrogate propagates *after* the non-bubbling base), but that seems
+				// less bad than duplication.
+				} else if ( ( jQuery.event.special[ type ] || {} ).delegateType ) {
+					event.stopPropagation();
+				}
+
+			// If this is a native event triggered above, everything is now in order
+			// Fire an inner synthetic event with the original arguments
+			} else if ( saved.length ) {
+
+				// ...and capture the result
+				dataPriv.set( this, type, {
+					value: jQuery.event.trigger(
+
+						// Support: IE <=9 - 11+
+						// Extend with the prototype to reset the above stopImmediatePropagation()
+						jQuery.extend( saved[ 0 ], jQuery.Event.prototype ),
+						saved.slice( 1 ),
+						this
+					)
+				} );
+
+				// Abort handling of the native event
+				event.stopImmediatePropagation();
+			}
+		}
+	} );
+}
+
+jQuery.removeEvent = function( elem, type, handle ) {
+
+	// This "if" is needed for plain objects
+	if ( elem.removeEventListener ) {
+		elem.removeEventListener( type, handle );
+	}
+};
+
+jQuery.Event = function( src, props ) {
+
+	// Allow instantiation without the 'new' keyword
+	if ( !( this instanceof jQuery.Event ) ) {
+		return new jQuery.Event( src, props );
+	}
+
+	// Event object
+	if ( src && src.type ) {
+		this.originalEvent = src;
+		this.type = src.type;
+
+		// Events bubbling up the document may have been marked as prevented
+		// by a handler lower down the tree; reflect the correct value.
+		this.isDefaultPrevented = src.defaultPrevented ||
+				src.defaultPrevented === undefined &&
+
+				// Support: Android <=2.3 only
+				src.returnValue === false ?
+			returnTrue :
+			returnFalse;
+
+		// Create target properties
+		// Support: Safari <=6 - 7 only
+		// Target should not be a text node (#504, #13143)
+		this.target = ( src.target && src.target.nodeType === 3 ) ?
+			src.target.parentNode :
+			src.target;
+
+		this.currentTarget = src.currentTarget;
+		this.relatedTarget = src.relatedTarget;
+
+	// Event type
+	} else {
+		this.type = src;
+	}
+
+	// Put explicitly provided properties onto the event object
+	if ( props ) {
+		jQuery.extend( this, props );
+	}
+
+	// Create a timestamp if incoming event doesn't have one
+	this.timeStamp = src && src.timeStamp || Date.now();
+
+	// Mark it as fixed
+	this[ jQuery.expando ] = true;
+};
+
+// jQuery.Event is based on DOM3 Events as specified by the ECMAScript Language Binding
+// https://www.w3.org/TR/2003/WD-DOM-Level-3-Events-20030331/ecma-script-binding.html
+jQuery.Event.prototype = {
+	constructor: jQuery.Event,
+	isDefaultPrevented: returnFalse,
+	isPropagationStopped: returnFalse,
+	isImmediatePropagationStopped: returnFalse,
+	isSimulated: false,
+
+	preventDefault: function() {
+		var e = this.originalEvent;
+
+		this.isDefaultPrevented = returnTrue;
+
+		if ( e && !this.isSimulated ) {
+			e.preventDefault();
+		}
+	},
+	stopPropagation: function() {
+		var e = this.originalEvent;
+
+		this.isPropagationStopped = returnTrue;
+
+		if ( e && !this.isSimulated ) {
+			e.stopPropagation();
+		}
+	},
+	stopImmediatePropagation: function() {
+		var e = this.originalEvent;
+
+		this.isImmediatePropagationStopped = returnTrue;
+
+		if ( e && !this.isSimulated ) {
+			e.stopImmediatePropagation();
+		}
+
+		this.stopPropagation();
+	}
+};
+
+// Includes all common event props including KeyEvent and MouseEvent specific props
+jQuery.each( {
+	altKey: true,
+	bubbles: true,
+	cancelable: true,
+	changedTouches: true,
+	ctrlKey: true,
+	detail: true,
+	eventPhase: true,
+	metaKey: true,
+	pageX: true,
+	pageY: true,
+	shiftKey: true,
+	view: true,
+	"char": true,
+	code: true,
+	charCode: true,
+	key: true,
+	keyCode: true,
+	button: true,
+	buttons: true,
+	clientX: true,
+	clientY: true,
+	offsetX: true,
+	offsetY: true,
+	pointerId: true,
+	pointerType: true,
+	screenX: true,
+	screenY: true,
+	targetTouches: true,
+	toElement: true,
+	touches: true,
+	which: true
+}, jQuery.event.addProp );
+
+jQuery.each( { focus: "focusin", blur: "focusout" }, function( type, delegateType ) {
+	jQuery.event.special[ type ] = {
+
+		// Utilize native event if possible so blur/focus sequence is correct
+		setup: function() {
+
+			// Claim the first handler
+			// dataPriv.set( this, "focus", ... )
+			// dataPriv.set( this, "blur", ... )
+			leverageNative( this, type, expectSync );
+
+			// Return false to allow normal processing in the caller
+			return false;
+		},
+		trigger: function() {
+
+			// Force setup before trigger
+			leverageNative( this, type );
+
+			// Return non-false to allow normal event-path propagation
+			return true;
+		},
+
+		// Suppress native focus or blur as it's already being fired
+		// in leverageNative.
+		_default: function() {
+			return true;
+		},
+
+		delegateType: delegateType
+	};
+} );
+
+// Create mouseenter/leave events using mouseover/out and event-time checks
+// so that event delegation works in jQuery.
+// Do the same for pointerenter/pointerleave and pointerover/pointerout
+//
+// Support: Safari 7 only
+// Safari sends mouseenter too often; see:
+// https://bugs.chromium.org/p/chromium/issues/detail?id=470258
+// for the description of the bug (it existed in older Chrome versions as well).
+jQuery.each( {
+	mouseenter: "mouseover",
+	mouseleave: "mouseout",
+	pointerenter: "pointerover",
+	pointerleave: "pointerout"
+}, function( orig, fix ) {
+	jQuery.event.special[ orig ] = {
+		delegateType: fix,
+		bindType: fix,
+
+		handle: function( event ) {
+			var ret,
+				target = this,
+				related = event.relatedTarget,
+				handleObj = event.handleObj;
+
+			// For mouseenter/leave call the handler if related is outside the target.
+			// NB: No relatedTarget if the mouse left/entered the browser window
+			if ( !related || ( related !== target && !jQuery.contains( target, related ) ) ) {
+				event.type = handleObj.origType;
+				ret = handleObj.handler.apply( this, arguments );
+				event.type = fix;
+			}
+			return ret;
+		}
+	};
+} );
+
+jQuery.fn.extend( {
+
+	on: function( types, selector, data, fn ) {
+		return on( this, types, selector, data, fn );
+	},
+	one: function( types, selector, data, fn ) {
+		return on( this, types, selector, data, fn, 1 );
+	},
+	off: function( types, selector, fn ) {
+		var handleObj, type;
+		if ( types && types.preventDefault && types.handleObj ) {
+
+			// ( event )  dispatched jQuery.Event
+			handleObj = types.handleObj;
+			jQuery( types.delegateTarget ).off(
+				handleObj.namespace ?
+					handleObj.origType + "." + handleObj.namespace :
+					handleObj.origType,
+				handleObj.selector,
+				handleObj.handler
+			);
+			return this;
+		}
+		if ( typeof types === "object" ) {
+
+			// ( types-object [, selector] )
+			for ( type in types ) {
+				this.off( type, selector, types[ type ] );
+			}
+			return this;
+		}
+		if ( selector === false || typeof selector === "function" ) {
+
+			// ( types [, fn] )
+			fn = selector;
+			selector = undefined;
+		}
+		if ( fn === false ) {
+			fn = returnFalse;
+		}
+		return this.each( function() {
+			jQuery.event.remove( this, types, fn, selector );
+		} );
+	}
+} );
+
+
+var
+
+	// Support: IE <=10 - 11, Edge 12 - 13 only
+	// In IE/Edge using regex groups here causes severe slowdowns.
+	// See https://connect.microsoft.com/IE/feedback/details/1736512/
+	rnoInnerhtml = /<script|<style|<link/i,
+
+	// checked="checked" or checked
+	rchecked = /checked\s*(?:[^=]|=\s*.checked.)/i,
+	rcleanScript = /^\s*<!(?:\[CDATA\[|--)|(?:\]\]|--)>\s*$/g;
+
+// Prefer a tbody over its parent table for containing new rows
+function manipulationTarget( elem, content ) {
+	if ( nodeName( elem, "table" ) &&
+		nodeName( content.nodeType !== 11 ? content : content.firstChild, "tr" ) ) {
+
+		return jQuery( elem ).children( "tbody" )[ 0 ] || elem;
+	}
+
+	return elem;
+}
+
+// Replace/restore the type attribute of script elements for safe DOM manipulation
+function disableScript( elem ) {
+	elem.type = ( elem.getAttribute( "type" ) !== null ) + "/" + elem.type;
+	return elem;
+}
+function restoreScript( elem ) {
+	if ( ( elem.type || "" ).slice( 0, 5 ) === "true/" ) {
+		elem.type = elem.type.slice( 5 );
+	} else {
+		elem.removeAttribute( "type" );
+	}
+
+	return elem;
+}
+
+function cloneCopyEvent( src, dest ) {
+	var i, l, type, pdataOld, udataOld, udataCur, events;
+
+	if ( dest.nodeType !== 1 ) {
+		return;
+	}
+
+	// 1. Copy private data: events, handlers, etc.
+	if ( dataPriv.hasData( src ) ) {
+		pdataOld = dataPriv.get( src );
+		events = pdataOld.events;
+
+		if ( events ) {
+			dataPriv.remove( dest, "handle events" );
+
+			for ( type in events ) {
+				for ( i = 0, l = events[ type ].length; i < l; i++ ) {
+					jQuery.event.add( dest, type, events[ type ][ i ] );
+				}
+			}
+		}
+	}
+
+	// 2. Copy user data
+	if ( dataUser.hasData( src ) ) {
+		udataOld = dataUser.access( src );
+		udataCur = jQuery.extend( {}, udataOld );
+
+		dataUser.set( dest, udataCur );
+	}
+}
+
+// Fix IE bugs, see support tests
+function fixInput( src, dest ) {
+	var nodeName = dest.nodeName.toLowerCase();
+
+	// Fails to persist the checked state of a cloned checkbox or radio button.
+	if ( nodeName === "input" && rcheckableType.test( src.type ) ) {
+		dest.checked = src.checked;
+
+	// Fails to return the selected option to the default selected state when cloning options
+	} else if ( nodeName === "input" || nodeName === "textarea" ) {
+		dest.defaultValue = src.defaultValue;
+	}
+}
+
+function domManip( collection, args, callback, ignored ) {
+
+	// Flatten any nested arrays
+	args = flat( args );
+
+	var fragment, first, scripts, hasScripts, node, doc,
+		i = 0,
+		l = collection.length,
+		iNoClone = l - 1,
+		value = args[ 0 ],
+		valueIsFunction = isFunction( value );
+
+	// We can't cloneNode fragments that contain checked, in WebKit
+	if ( valueIsFunction ||
+			( l > 1 && typeof value === "string" &&
+				!support.checkClone && rchecked.test( value ) ) ) {
+		return collection.each( function( index ) {
+			var self = collection.eq( index );
+			if ( valueIsFunction ) {
+				args[ 0 ] = value.call( this, index, self.html() );
+			}
+			domManip( self, args, callback, ignored );
+		} );
+	}
+
+	if ( l ) {
+		fragment = buildFragment( args, collection[ 0 ].ownerDocument, false, collection, ignored );
+		first = fragment.firstChild;
+
+		if ( fragment.childNodes.length === 1 ) {
+			fragment = first;
+		}
+
+		// Require either new content or an interest in ignored elements to invoke the callback
+		if ( first || ignored ) {
+			scripts = jQuery.map( getAll( fragment, "script" ), disableScript );
+			hasScripts = scripts.length;
+
+			// Use the original fragment for the last item
+			// instead of the first because it can end up
+			// being emptied incorrectly in certain situations (#8070).
+			for ( ; i < l; i++ ) {
+				node = fragment;
+
+				if ( i !== iNoClone ) {
+					node = jQuery.clone( node, true, true );
+
+					// Keep references to cloned scripts for later restoration
+					if ( hasScripts ) {
+
+						// Support: Android <=4.0 only, PhantomJS 1 only
+						// push.apply(_, arraylike) throws on ancient WebKit
+						jQuery.merge( scripts, getAll( node, "script" ) );
+					}
+				}
+
+				callback.call( collection[ i ], node, i );
+			}
+
+			if ( hasScripts ) {
+				doc = scripts[ scripts.length - 1 ].ownerDocument;
+
+				// Reenable scripts
+				jQuery.map( scripts, restoreScript );
+
+				// Evaluate executable scripts on first document insertion
+				for ( i = 0; i < hasScripts; i++ ) {
+					node = scripts[ i ];
+					if ( rscriptType.test( node.type || "" ) &&
+						!dataPriv.access( node, "globalEval" ) &&
+						jQuery.contains( doc, node ) ) {
+
+						if ( node.src && ( node.type || "" ).toLowerCase()  !== "module" ) {
+
+							// Optional AJAX dependency, but won't run scripts if not present
+							if ( jQuery._evalUrl && !node.noModule ) {
+								jQuery._evalUrl( node.src, {
+									nonce: node.nonce || node.getAttribute( "nonce" )
+								}, doc );
+							}
+						} else {
+							DOMEval( node.textContent.replace( rcleanScript, "" ), node, doc );
+						}
+					}
+				}
+			}
+		}
+	}
+
+	return collection;
+}
+
+function remove( elem, selector, keepData ) {
+	var node,
+		nodes = selector ? jQuery.filter( selector, elem ) : elem,
+		i = 0;
+
+	for ( ; ( node = nodes[ i ] ) != null; i++ ) {
+		if ( !keepData && node.nodeType === 1 ) {
+			jQuery.cleanData( getAll( node ) );
+		}
+
+		if ( node.parentNode ) {
+			if ( keepData && isAttached( node ) ) {
+				setGlobalEval( getAll( node, "script" ) );
+			}
+			node.parentNode.removeChild( node );
+		}
+	}
+
+	return elem;
+}
+
+jQuery.extend( {
+	htmlPrefilter: function( html ) {
+		return html;
+	},
+
+	clone: function( elem, dataAndEvents, deepDataAndEvents ) {
+		var i, l, srcElements, destElements,
+			clone = elem.cloneNode( true ),
+			inPage = isAttached( elem );
+
+		// Fix IE cloning issues
+		if ( !support.noCloneChecked && ( elem.nodeType === 1 || elem.nodeType === 11 ) &&
+				!jQuery.isXMLDoc( elem ) ) {
+
+			// We eschew Sizzle here for performance reasons: https://jsperf.com/getall-vs-sizzle/2
+			destElements = getAll( clone );
+			srcElements = getAll( elem );
+
+			for ( i = 0, l = srcElements.length; i < l; i++ ) {
+				fixInput( srcElements[ i ], destElements[ i ] );
+			}
+		}
+
+		// Copy the events from the original to the clone
+		if ( dataAndEvents ) {
+			if ( deepDataAndEvents ) {
+				srcElements = srcElements || getAll( elem );
+				destElements = destElements || getAll( clone );
+
+				for ( i = 0, l = srcElements.length; i < l; i++ ) {
+					cloneCopyEvent( srcElements[ i ], destElements[ i ] );
+				}
+			} else {
+				cloneCopyEvent( elem, clone );
+			}
+		}
+
+		// Preserve script evaluation history
+		destElements = getAll( clone, "script" );
+		if ( destElements.length > 0 ) {
+			setGlobalEval( destElements, !inPage && getAll( elem, "script" ) );
+		}
+
+		// Return the cloned set
+		return clone;
+	},
+
+	cleanData: function( elems ) {
+		var data, elem, type,
+			special = jQuery.event.special,
+			i = 0;
+
+		for ( ; ( elem = elems[ i ] ) !== undefined; i++ ) {
+			if ( acceptData( elem ) ) {
+				if ( ( data = elem[ dataPriv.expando ] ) ) {
+					if ( data.events ) {
+						for ( type in data.events ) {
+							if ( special[ type ] ) {
+								jQuery.event.remove( elem, type );
+
+							// This is a shortcut to avoid jQuery.event.remove's overhead
+							} else {
+								jQuery.removeEvent( elem, type, data.handle );
+							}
+						}
+					}
+
+					// Support: Chrome <=35 - 45+
+					// Assign undefined instead of using delete, see Data#remove
+					elem[ dataPriv.expando ] = undefined;
+				}
+				if ( elem[ dataUser.expando ] ) {
+
+					// Support: Chrome <=35 - 45+
+					// Assign undefined instead of using delete, see Data#remove
+					elem[ dataUser.expando ] = undefined;
+				}
+			}
+		}
+	}
+} );
+
+jQuery.fn.extend( {
+	detach: function( selector ) {
+		return remove( this, selector, true );
+	},
+
+	remove: function( selector ) {
+		return remove( this, selector );
+	},
+
+	text: function( value ) {
+		return access( this, function( value ) {
+			return value === undefined ?
+				jQuery.text( this ) :
+				this.empty().each( function() {
+					if ( this.nodeType === 1 || this.nodeType === 11 || this.nodeType === 9 ) {
+						this.textContent = value;
+					}
+				} );
+		}, null, value, arguments.length );
+	},
+
+	append: function() {
+		return domManip( this, arguments, function( elem ) {
+			if ( this.nodeType === 1 || this.nodeType === 11 || this.nodeType === 9 ) {
+				var target = manipulationTarget( this, elem );
+				target.appendChild( elem );
+			}
+		} );
+	},
+
+	prepend: function() {
+		return domManip( this, arguments, function( elem ) {
+			if ( this.nodeType === 1 || this.nodeType === 11 || this.nodeType === 9 ) {
+				var target = manipulationTarget( this, elem );
+				target.insertBefore( elem, target.firstChild );
+			}
+		} );
+	},
+
+	before: function() {
+		return domManip( this, arguments, function( elem ) {
+			if ( this.parentNode ) {
+				this.parentNode.insertBefore( elem, this );
+			}
+		} );
+	},
+
+	after: function() {
+		return domManip( this, arguments, function( elem ) {
+			if ( this.parentNode ) {
+				this.parentNode.insertBefore( elem, this.nextSibling );
+			}
+		} );
+	},
+
+	empty: function() {
+		var elem,
+			i = 0;
+
+		for ( ; ( elem = this[ i ] ) != null; i++ ) {
+			if ( elem.nodeType === 1 ) {
+
+				// Prevent memory leaks
+				jQuery.cleanData( getAll( elem, false ) );
+
+				// Remove any remaining nodes
+				elem.textContent = "";
+			}
+		}
+
+		return this;
+	},
+
+	clone: function( dataAndEvents, deepDataAndEvents ) {
+		dataAndEvents = dataAndEvents == null ? false : dataAndEvents;
+		deepDataAndEvents = deepDataAndEvents == null ? dataAndEvents : deepDataAndEvents;
+
+		return this.map( function() {
+			return jQuery.clone( this, dataAndEvents, deepDataAndEvents );
+		} );
+	},
+
+	html: function( value ) {
+		return access( this, function( value ) {
+			var elem = this[ 0 ] || {},
+				i = 0,
+				l = this.length;
+
+			if ( value === undefined && elem.nodeType === 1 ) {
+				return elem.innerHTML;
+			}
+
+			// See if we can take a shortcut and just use innerHTML
+			if ( typeof value === "string" && !rnoInnerhtml.test( value ) &&
+				!wrapMap[ ( rtagName.exec( value ) || [ "", "" ] )[ 1 ].toLowerCase() ] ) {
+
+				value = jQuery.htmlPrefilter( value );
+
+				try {
+					for ( ; i < l; i++ ) {
+						elem = this[ i ] || {};
+
+						// Remove element nodes and prevent memory leaks
+						if ( elem.nodeType === 1 ) {
+							jQuery.cleanData( getAll( elem, false ) );
+							elem.innerHTML = value;
+						}
+					}
+
+					elem = 0;
+
+				// If using innerHTML throws an exception, use the fallback method
+				} catch ( e ) {}
+			}
+
+			if ( elem ) {
+				this.empty().append( value );
+			}
+		}, null, value, arguments.length );
+	},
+
+	replaceWith: function() {
+		var ignored = [];
+
+		// Make the changes, replacing each non-ignored context element with the new content
+		return domManip( this, arguments, function( elem ) {
+			var parent = this.parentNode;
+
+			if ( jQuery.inArray( this, ignored ) < 0 ) {
+				jQuery.cleanData( getAll( this ) );
+				if ( parent ) {
+					parent.replaceChild( elem, this );
+				}
+			}
+
+		// Force callback invocation
+		}, ignored );
+	}
+} );
+
+jQuery.each( {
+	appendTo: "append",
+	prependTo: "prepend",
+	insertBefore: "before",
+	insertAfter: "after",
+	replaceAll: "replaceWith"
+}, function( name, original ) {
+	jQuery.fn[ name ] = function( selector ) {
+		var elems,
+			ret = [],
+			insert = jQuery( selector ),
+			last = insert.length - 1,
+			i = 0;
+
+		for ( ; i <= last; i++ ) {
+			elems = i === last ? this : this.clone( true );
+			jQuery( insert[ i ] )[ original ]( elems );
+
+			// Support: Android <=4.0 only, PhantomJS 1 only
+			// .get() because push.apply(_, arraylike) throws on ancient WebKit
+			push.apply( ret, elems.get() );
+		}
+
+		return this.pushStack( ret );
+	};
+} );
+var rnumnonpx = new RegExp( "^(" + pnum + ")(?!px)[a-z%]+$", "i" );
+
+var getStyles = function( elem ) {
+
+		// Support: IE <=11 only, Firefox <=30 (#15098, #14150)
+		// IE throws on elements created in popups
+		// FF meanwhile throws on frame elements through "defaultView.getComputedStyle"
+		var view = elem.ownerDocument.defaultView;
+
+		if ( !view || !view.opener ) {
+			view = window;
+		}
+
+		return view.getComputedStyle( elem );
+	};
+
+var swap = function( elem, options, callback ) {
+	var ret, name,
+		old = {};
+
+	// Remember the old values, and insert the new ones
+	for ( name in options ) {
+		old[ name ] = elem.style[ name ];
+		elem.style[ name ] = options[ name ];
+	}
+
+	ret = callback.call( elem );
+
+	// Revert the old values
+	for ( name in options ) {
+		elem.style[ name ] = old[ name ];
+	}
+
+	return ret;
+};
+
+
+var rboxStyle = new RegExp( cssExpand.join( "|" ), "i" );
+
+
+
+( function() {
+
+	// Executing both pixelPosition & boxSizingReliable tests require only one layout
+	// so they're executed at the same time to save the second computation.
+	function computeStyleTests() {
+
+		// This is a singleton, we need to execute it only once
+		if ( !div ) {
+			return;
+		}
+
+		container.style.cssText = "position:absolute;left:-11111px;width:60px;" +
+			"margin-top:1px;padding:0;border:0";
+		div.style.cssText =
+			"position:relative;display:block;box-sizing:border-box;overflow:scroll;" +
+			"margin:auto;border:1px;padding:1px;" +
+			"width:60%;top:1%";
+		documentElement.appendChild( container ).appendChild( div );
+
+		var divStyle = window.getComputedStyle( div );
+		pixelPositionVal = divStyle.top !== "1%";
+
+		// Support: Android 4.0 - 4.3 only, Firefox <=3 - 44
+		reliableMarginLeftVal = roundPixelMeasures( divStyle.marginLeft ) === 12;
+
+		// Support: Android 4.0 - 4.3 only, Safari <=9.1 - 10.1, iOS <=7.0 - 9.3
+		// Some styles come back with percentage values, even though they shouldn't
+		div.style.right = "60%";
+		pixelBoxStylesVal = roundPixelMeasures( divStyle.right ) === 36;
+
+		// Support: IE 9 - 11 only
+		// Detect misreporting of content dimensions for box-sizing:border-box elements
+		boxSizingReliableVal = roundPixelMeasures( divStyle.width ) === 36;
+
+		// Support: IE 9 only
+		// Detect overflow:scroll screwiness (gh-3699)
+		// Support: Chrome <=64
+		// Don't get tricked when zoom affects offsetWidth (gh-4029)
+		div.style.position = "absolute";
+		scrollboxSizeVal = roundPixelMeasures( div.offsetWidth / 3 ) === 12;
+
+		documentElement.removeChild( container );
+
+		// Nullify the div so it wouldn't be stored in the memory and
+		// it will also be a sign that checks already performed
+		div = null;
+	}
+
+	function roundPixelMeasures( measure ) {
+		return Math.round( parseFloat( measure ) );
+	}
+
+	var pixelPositionVal, boxSizingReliableVal, scrollboxSizeVal, pixelBoxStylesVal,
+		reliableTrDimensionsVal, reliableMarginLeftVal,
+		container = document.createElement( "div" ),
+		div = document.createElement( "div" );
+
+	// Finish early in limited (non-browser) environments
+	if ( !div.style ) {
+		return;
+	}
+
+	// Support: IE <=9 - 11 only
+	// Style of cloned element affects source element cloned (#8908)
+	div.style.backgroundClip = "content-box";
+	div.cloneNode( true ).style.backgroundClip = "";
+	support.clearCloneStyle = div.style.backgroundClip === "content-box";
+
+	jQuery.extend( support, {
+		boxSizingReliable: function() {
+			computeStyleTests();
+			return boxSizingReliableVal;
+		},
+		pixelBoxStyles: function() {
+			computeStyleTests();
+			return pixelBoxStylesVal;
+		},
+		pixelPosition: function() {
+			computeStyleTests();
+			return pixelPositionVal;
+		},
+		reliableMarginLeft: function() {
+			computeStyleTests();
+			return reliableMarginLeftVal;
+		},
+		scrollboxSize: function() {
+			computeStyleTests();
+			return scrollboxSizeVal;
+		},
+
+		// Support: IE 9 - 11+, Edge 15 - 18+
+		// IE/Edge misreport `getComputedStyle` of table rows with width/height
+		// set in CSS while `offset*` properties report correct values.
+		// Behavior in IE 9 is more subtle than in newer versions & it passes
+		// some versions of this test; make sure not to make it pass there!
+		//
+		// Support: Firefox 70+
+		// Only Firefox includes border widths
+		// in computed dimensions. (gh-4529)
+		reliableTrDimensions: function() {
+			var table, tr, trChild, trStyle;
+			if ( reliableTrDimensionsVal == null ) {
+				table = document.createElement( "table" );
+				tr = document.createElement( "tr" );
+				trChild = document.createElement( "div" );
+
+				table.style.cssText = "position:absolute;left:-11111px;border-collapse:separate";
+				tr.style.cssText = "border:1px solid";
+
+				// Support: Chrome 86+
+				// Height set through cssText does not get applied.
+				// Computed height then comes back as 0.
+				tr.style.height = "1px";
+				trChild.style.height = "9px";
+
+				// Support: Android 8 Chrome 86+
+				// In our bodyBackground.html iframe,
+				// display for all div elements is set to "inline",
+				// which causes a problem only in Android 8 Chrome 86.
+				// Ensuring the div is display: block
+				// gets around this issue.
+				trChild.style.display = "block";
+
+				documentElement
+					.appendChild( table )
+					.appendChild( tr )
+					.appendChild( trChild );
+
+				trStyle = window.getComputedStyle( tr );
+				reliableTrDimensionsVal = ( parseInt( trStyle.height, 10 ) +
+					parseInt( trStyle.borderTopWidth, 10 ) +
+					parseInt( trStyle.borderBottomWidth, 10 ) ) === tr.offsetHeight;
+
+				documentElement.removeChild( table );
+			}
+			return reliableTrDimensionsVal;
+		}
+	} );
+} )();
+
+
+function curCSS( elem, name, computed ) {
+	var width, minWidth, maxWidth, ret,
+
+		// Support: Firefox 51+
+		// Retrieving style before computed somehow
+		// fixes an issue with getting wrong values
+		// on detached elements
+		style = elem.style;
+
+	computed = computed || getStyles( elem );
+
+	// getPropertyValue is needed for:
+	//   .css('filter') (IE 9 only, #12537)
+	//   .css('--customProperty) (#3144)
+	if ( computed ) {
+		ret = computed.getPropertyValue( name ) || computed[ name ];
+
+		if ( ret === "" && !isAttached( elem ) ) {
+			ret = jQuery.style( elem, name );
+		}
+
+		// A tribute to the "awesome hack by Dean Edwards"
+		// Android Browser returns percentage for some values,
+		// but width seems to be reliably pixels.
+		// This is against the CSSOM draft spec:
+		// https://drafts.csswg.org/cssom/#resolved-values
+		if ( !support.pixelBoxStyles() && rnumnonpx.test( ret ) && rboxStyle.test( name ) ) {
+
+			// Remember the original values
+			width = style.width;
+			minWidth = style.minWidth;
+			maxWidth = style.maxWidth;
+
+			// Put in the new values to get a computed value out
+			style.minWidth = style.maxWidth = style.width = ret;
+			ret = computed.width;
+
+			// Revert the changed values
+			style.width = width;
+			style.minWidth = minWidth;
+			style.maxWidth = maxWidth;
+		}
+	}
+
+	return ret !== undefined ?
+
+		// Support: IE <=9 - 11 only
+		// IE returns zIndex value as an integer.
+		ret + "" :
+		ret;
+}
+
+
+function addGetHookIf( conditionFn, hookFn ) {
+
+	// Define the hook, we'll check on the first run if it's really needed.
+	return {
+		get: function() {
+			if ( conditionFn() ) {
+
+				// Hook not needed (or it's not possible to use it due
+				// to missing dependency), remove it.
+				delete this.get;
+				return;
+			}
+
+			// Hook needed; redefine it so that the support test is not executed again.
+			return ( this.get = hookFn ).apply( this, arguments );
+		}
+	};
+}
+
+
+var cssPrefixes = [ "Webkit", "Moz", "ms" ],
+	emptyStyle = document.createElement( "div" ).style,
+	vendorProps = {};
+
+// Return a vendor-prefixed property or undefined
+function vendorPropName( name ) {
+
+	// Check for vendor prefixed names
+	var capName = name[ 0 ].toUpperCase() + name.slice( 1 ),
+		i = cssPrefixes.length;
+
+	while ( i-- ) {
+		name = cssPrefixes[ i ] + capName;
+		if ( name in emptyStyle ) {
+			return name;
+		}
+	}
+}
+
+// Return a potentially-mapped jQuery.cssProps or vendor prefixed property
+function finalPropName( name ) {
+	var final = jQuery.cssProps[ name ] || vendorProps[ name ];
+
+	if ( final ) {
+		return final;
+	}
+	if ( name in emptyStyle ) {
+		return name;
+	}
+	return vendorProps[ name ] = vendorPropName( name ) || name;
+}
+
+
+var
+
+	// Swappable if display is none or starts with table
+	// except "table", "table-cell", or "table-caption"
+	// See here for display values: https://developer.mozilla.org/en-US/docs/CSS/display
+	rdisplayswap = /^(none|table(?!-c[ea]).+)/,
+	rcustomProp = /^--/,
+	cssShow = { position: "absolute", visibility: "hidden", display: "block" },
+	cssNormalTransform = {
+		letterSpacing: "0",
+		fontWeight: "400"
+	};
+
+function setPositiveNumber( _elem, value, subtract ) {
+
+	// Any relative (+/-) values have already been
+	// normalized at this point
+	var matches = rcssNum.exec( value );
+	return matches ?
+
+		// Guard against undefined "subtract", e.g., when used as in cssHooks
+		Math.max( 0, matches[ 2 ] - ( subtract || 0 ) ) + ( matches[ 3 ] || "px" ) :
+		value;
+}
+
+function boxModelAdjustment( elem, dimension, box, isBorderBox, styles, computedVal ) {
+	var i = dimension === "width" ? 1 : 0,
+		extra = 0,
+		delta = 0;
+
+	// Adjustment may not be necessary
+	if ( box === ( isBorderBox ? "border" : "content" ) ) {
+		return 0;
+	}
+
+	for ( ; i < 4; i += 2 ) {
+
+		// Both box models exclude margin
+		if ( box === "margin" ) {
+			delta += jQuery.css( elem, box + cssExpand[ i ], true, styles );
+		}
+
+		// If we get here with a content-box, we're seeking "padding" or "border" or "margin"
+		if ( !isBorderBox ) {
+
+			// Add padding
+			delta += jQuery.css( elem, "padding" + cssExpand[ i ], true, styles );
+
+			// For "border" or "margin", add border
+			if ( box !== "padding" ) {
+				delta += jQuery.css( elem, "border" + cssExpand[ i ] + "Width", true, styles );
+
+			// But still keep track of it otherwise
+			} else {
+				extra += jQuery.css( elem, "border" + cssExpand[ i ] + "Width", true, styles );
+			}
+
+		// If we get here with a border-box (content + padding + border), we're seeking "content" or
+		// "padding" or "margin"
+		} else {
+
+			// For "content", subtract padding
+			if ( box === "content" ) {
+				delta -= jQuery.css( elem, "padding" + cssExpand[ i ], true, styles );
+			}
+
+			// For "content" or "padding", subtract border
+			if ( box !== "margin" ) {
+				delta -= jQuery.css( elem, "border" + cssExpand[ i ] + "Width", true, styles );
+			}
+		}
+	}
+
+	// Account for positive content-box scroll gutter when requested by providing computedVal
+	if ( !isBorderBox && computedVal >= 0 ) {
+
+		// offsetWidth/offsetHeight is a rounded sum of content, padding, scroll gutter, and border
+		// Assuming integer scroll gutter, subtract the rest and round down
+		delta += Math.max( 0, Math.ceil(
+			elem[ "offset" + dimension[ 0 ].toUpperCase() + dimension.slice( 1 ) ] -
+			computedVal -
+			delta -
+			extra -
+			0.5
+
+		// If offsetWidth/offsetHeight is unknown, then we can't determine content-box scroll gutter
+		// Use an explicit zero to avoid NaN (gh-3964)
+		) ) || 0;
+	}
+
+	return delta;
+}
+
+function getWidthOrHeight( elem, dimension, extra ) {
+
+	// Start with computed style
+	var styles = getStyles( elem ),
+
+		// To avoid forcing a reflow, only fetch boxSizing if we need it (gh-4322).
+		// Fake content-box until we know it's needed to know the true value.
+		boxSizingNeeded = !support.boxSizingReliable() || extra,
+		isBorderBox = boxSizingNeeded &&
+			jQuery.css( elem, "boxSizing", false, styles ) === "border-box",
+		valueIsBorderBox = isBorderBox,
+
+		val = curCSS( elem, dimension, styles ),
+		offsetProp = "offset" + dimension[ 0 ].toUpperCase() + dimension.slice( 1 );
+
+	// Support: Firefox <=54
+	// Return a confounding non-pixel value or feign ignorance, as appropriate.
+	if ( rnumnonpx.test( val ) ) {
+		if ( !extra ) {
+			return val;
+		}
+		val = "auto";
+	}
+
+
+	// Support: IE 9 - 11 only
+	// Use offsetWidth/offsetHeight for when box sizing is unreliable.
+	// In those cases, the computed value can be trusted to be border-box.
+	if ( ( !support.boxSizingReliable() && isBorderBox ||
+
+		// Support: IE 10 - 11+, Edge 15 - 18+
+		// IE/Edge misreport `getComputedStyle` of table rows with width/height
+		// set in CSS while `offset*` properties report correct values.
+		// Interestingly, in some cases IE 9 doesn't suffer from this issue.
+		!support.reliableTrDimensions() && nodeName( elem, "tr" ) ||
+
+		// Fall back to offsetWidth/offsetHeight when value is "auto"
+		// This happens for inline elements with no explicit setting (gh-3571)
+		val === "auto" ||
+
+		// Support: Android <=4.1 - 4.3 only
+		// Also use offsetWidth/offsetHeight for misreported inline dimensions (gh-3602)
+		!parseFloat( val ) && jQuery.css( elem, "display", false, styles ) === "inline" ) &&
+
+		// Make sure the element is visible & connected
+		elem.getClientRects().length ) {
+
+		isBorderBox = jQuery.css( elem, "boxSizing", false, styles ) === "border-box";
+
+		// Where available, offsetWidth/offsetHeight approximate border box dimensions.
+		// Where not available (e.g., SVG), assume unreliable box-sizing and interpret the
+		// retrieved value as a content box dimension.
+		valueIsBorderBox = offsetProp in elem;
+		if ( valueIsBorderBox ) {
+			val = elem[ offsetProp ];
+		}
+	}
+
+	// Normalize "" and auto
+	val = parseFloat( val ) || 0;
+
+	// Adjust for the element's box model
+	return ( val +
+		boxModelAdjustment(
+			elem,
+			dimension,
+			extra || ( isBorderBox ? "border" : "content" ),
+			valueIsBorderBox,
+			styles,
+
+			// Provide the current computed size to request scroll gutter calculation (gh-3589)
+			val
+		)
+	) + "px";
+}
+
+jQuery.extend( {
+
+	// Add in style property hooks for overriding the default
+	// behavior of getting and setting a style property
+	cssHooks: {
+		opacity: {
+			get: function( elem, computed ) {
+				if ( computed ) {
+
+					// We should always get a number back from opacity
+					var ret = curCSS( elem, "opacity" );
+					return ret === "" ? "1" : ret;
+				}
+			}
+		}
+	},
+
+	// Don't automatically add "px" to these possibly-unitless properties
+	cssNumber: {
+		"animationIterationCount": true,
+		"columnCount": true,
+		"fillOpacity": true,
+		"flexGrow": true,
+		"flexShrink": true,
+		"fontWeight": true,
+		"gridArea": true,
+		"gridColumn": true,
+		"gridColumnEnd": true,
+		"gridColumnStart": true,
+		"gridRow": true,
+		"gridRowEnd": true,
+		"gridRowStart": true,
+		"lineHeight": true,
+		"opacity": true,
+		"order": true,
+		"orphans": true,
+		"widows": true,
+		"zIndex": true,
+		"zoom": true
+	},
+
+	// Add in properties whose names you wish to fix before
+	// setting or getting the value
+	cssProps: {},
+
+	// Get and set the style property on a DOM Node
+	style: function( elem, name, value, extra ) {
+
+		// Don't set styles on text and comment nodes
+		if ( !elem || elem.nodeType === 3 || elem.nodeType === 8 || !elem.style ) {
+			return;
+		}
+
+		// Make sure that we're working with the right name
+		var ret, type, hooks,
+			origName = camelCase( name ),
+			isCustomProp = rcustomProp.test( name ),
+			style = elem.style;
+
+		// Make sure that we're working with the right name. We don't
+		// want to query the value if it is a CSS custom property
+		// since they are user-defined.
+		if ( !isCustomProp ) {
+			name = finalPropName( origName );
+		}
+
+		// Gets hook for the prefixed version, then unprefixed version
+		hooks = jQuery.cssHooks[ name ] || jQuery.cssHooks[ origName ];
+
+		// Check if we're setting a value
+		if ( value !== undefined ) {
+			type = typeof value;
+
+			// Convert "+=" or "-=" to relative numbers (#7345)
+			if ( type === "string" && ( ret = rcssNum.exec( value ) ) && ret[ 1 ] ) {
+				value = adjustCSS( elem, name, ret );
+
+				// Fixes bug #9237
+				type = "number";
+			}
+
+			// Make sure that null and NaN values aren't set (#7116)
+			if ( value == null || value !== value ) {
+				return;
+			}
+
+			// If a number was passed in, add the unit (except for certain CSS properties)
+			// The isCustomProp check can be removed in jQuery 4.0 when we only auto-append
+			// "px" to a few hardcoded values.
+			if ( type === "number" && !isCustomProp ) {
+				value += ret && ret[ 3 ] || ( jQuery.cssNumber[ origName ] ? "" : "px" );
+			}
+
+			// background-* props affect original clone's values
+			if ( !support.clearCloneStyle && value === "" && name.indexOf( "background" ) === 0 ) {
+				style[ name ] = "inherit";
+			}
+
+			// If a hook was provided, use that value, otherwise just set the specified value
+			if ( !hooks || !( "set" in hooks ) ||
+				( value = hooks.set( elem, value, extra ) ) !== undefined ) {
+
+				if ( isCustomProp ) {
+					style.setProperty( name, value );
+				} else {
+					style[ name ] = value;
+				}
+			}
+
+		} else {
+
+			// If a hook was provided get the non-computed value from there
+			if ( hooks && "get" in hooks &&
+				( ret = hooks.get( elem, false, extra ) ) !== undefined ) {
+
+				return ret;
+			}
+
+			// Otherwise just get the value from the style object
+			return style[ name ];
+		}
+	},
+
+	css: function( elem, name, extra, styles ) {
+		var val, num, hooks,
+			origName = camelCase( name ),
+			isCustomProp = rcustomProp.test( name );
+
+		// Make sure that we're working with the right name. We don't
+		// want to modify the value if it is a CSS custom property
+		// since they are user-defined.
+		if ( !isCustomProp ) {
+			name = finalPropName( origName );
+		}
+
+		// Try prefixed name followed by the unprefixed name
+		hooks = jQuery.cssHooks[ name ] || jQuery.cssHooks[ origName ];
+
+		// If a hook was provided get the computed value from there
+		if ( hooks && "get" in hooks ) {
+			val = hooks.get( elem, true, extra );
+		}
+
+		// Otherwise, if a way to get the computed value exists, use that
+		if ( val === undefined ) {
+			val = curCSS( elem, name, styles );
+		}
+
+		// Convert "normal" to computed value
+		if ( val === "normal" && name in cssNormalTransform ) {
+			val = cssNormalTransform[ name ];
+		}
+
+		// Make numeric if forced or a qualifier was provided and val looks numeric
+		if ( extra === "" || extra ) {
+			num = parseFloat( val );
+			return extra === true || isFinite( num ) ? num || 0 : val;
+		}
+
+		return val;
+	}
+} );
+
+jQuery.each( [ "height", "width" ], function( _i, dimension ) {
+	jQuery.cssHooks[ dimension ] = {
+		get: function( elem, computed, extra ) {
+			if ( computed ) {
+
+				// Certain elements can have dimension info if we invisibly show them
+				// but it must have a current display style that would benefit
+				return rdisplayswap.test( jQuery.css( elem, "display" ) ) &&
+
+					// Support: Safari 8+
+					// Table columns in Safari have non-zero offsetWidth & zero
+					// getBoundingClientRect().width unless display is changed.
+					// Support: IE <=11 only
+					// Running getBoundingClientRect on a disconnected node
+					// in IE throws an error.
+					( !elem.getClientRects().length || !elem.getBoundingClientRect().width ) ?
+					swap( elem, cssShow, function() {
+						return getWidthOrHeight( elem, dimension, extra );
+					} ) :
+					getWidthOrHeight( elem, dimension, extra );
+			}
+		},
+
+		set: function( elem, value, extra ) {
+			var matches,
+				styles = getStyles( elem ),
+
+				// Only read styles.position if the test has a chance to fail
+				// to avoid forcing a reflow.
+				scrollboxSizeBuggy = !support.scrollboxSize() &&
+					styles.position === "absolute",
+
+				// To avoid forcing a reflow, only fetch boxSizing if we need it (gh-3991)
+				boxSizingNeeded = scrollboxSizeBuggy || extra,
+				isBorderBox = boxSizingNeeded &&
+					jQuery.css( elem, "boxSizing", false, styles ) === "border-box",
+				subtract = extra ?
+					boxModelAdjustment(
+						elem,
+						dimension,
+						extra,
+						isBorderBox,
+						styles
+					) :
+					0;
+
+			// Account for unreliable border-box dimensions by comparing offset* to computed and
+			// faking a content-box to get border and padding (gh-3699)
+			if ( isBorderBox && scrollboxSizeBuggy ) {
+				subtract -= Math.ceil(
+					elem[ "offset" + dimension[ 0 ].toUpperCase() + dimension.slice( 1 ) ] -
+					parseFloat( styles[ dimension ] ) -
+					boxModelAdjustment( elem, dimension, "border", false, styles ) -
+					0.5
+				);
+			}
+
+			// Convert to pixels if value adjustment is needed
+			if ( subtract && ( matches = rcssNum.exec( value ) ) &&
+				( matches[ 3 ] || "px" ) !== "px" ) {
+
+				elem.style[ dimension ] = value;
+				value = jQuery.css( elem, dimension );
+			}
+
+			return setPositiveNumber( elem, value, subtract );
+		}
+	};
+} );
+
+jQuery.cssHooks.marginLeft = addGetHookIf( support.reliableMarginLeft,
+	function( elem, computed ) {
+		if ( computed ) {
+			return ( parseFloat( curCSS( elem, "marginLeft" ) ) ||
+				elem.getBoundingClientRect().left -
+					swap( elem, { marginLeft: 0 }, function() {
+						return elem.getBoundingClientRect().left;
+					} )
+			) + "px";
+		}
+	}
+);
+
+// These hooks are used by animate to expand properties
+jQuery.each( {
+	margin: "",
+	padding: "",
+	border: "Width"
+}, function( prefix, suffix ) {
+	jQuery.cssHooks[ prefix + suffix ] = {
+		expand: function( value ) {
+			var i = 0,
+				expanded = {},
+
+				// Assumes a single number if not a string
+				parts = typeof value === "string" ? value.split( " " ) : [ value ];
+
+			for ( ; i < 4; i++ ) {
+				expanded[ prefix + cssExpand[ i ] + suffix ] =
+					parts[ i ] || parts[ i - 2 ] || parts[ 0 ];
+			}
+
+			return expanded;
+		}
+	};
+
+	if ( prefix !== "margin" ) {
+		jQuery.cssHooks[ prefix + suffix ].set = setPositiveNumber;
+	}
+} );
+
+jQuery.fn.extend( {
+	css: function( name, value ) {
+		return access( this, function( elem, name, value ) {
+			var styles, len,
+				map = {},
+				i = 0;
+
+			if ( Array.isArray( name ) ) {
+				styles = getStyles( elem );
+				len = name.length;
+
+				for ( ; i < len; i++ ) {
+					map[ name[ i ] ] = jQuery.css( elem, name[ i ], false, styles );
+				}
+
+				return map;
+			}
+
+			return value !== undefined ?
+				jQuery.style( elem, name, value ) :
+				jQuery.css( elem, name );
+		}, name, value, arguments.length > 1 );
+	}
+} );
+
+
+function Tween( elem, options, prop, end, easing ) {
+	return new Tween.prototype.init( elem, options, prop, end, easing );
+}
+jQuery.Tween = Tween;
+
+Tween.prototype = {
+	constructor: Tween,
+	init: function( elem, options, prop, end, easing, unit ) {
+		this.elem = elem;
+		this.prop = prop;
+		this.easing = easing || jQuery.easing._default;
+		this.options = options;
+		this.start = this.now = this.cur();
+		this.end = end;
+		this.unit = unit || ( jQuery.cssNumber[ prop ] ? "" : "px" );
+	},
+	cur: function() {
+		var hooks = Tween.propHooks[ this.prop ];
+
+		return hooks && hooks.get ?
+			hooks.get( this ) :
+			Tween.propHooks._default.get( this );
+	},
+	run: function( percent ) {
+		var eased,
+			hooks = Tween.propHooks[ this.prop ];
+
+		if ( this.options.duration ) {
+			this.pos = eased = jQuery.easing[ this.easing ](
+				percent, this.options.duration * percent, 0, 1, this.options.duration
+			);
+		} else {
+			this.pos = eased = percent;
+		}
+		this.now = ( this.end - this.start ) * eased + this.start;
+
+		if ( this.options.step ) {
+			this.options.step.call( this.elem, this.now, this );
+		}
+
+		if ( hooks && hooks.set ) {
+			hooks.set( this );
+		} else {
+			Tween.propHooks._default.set( this );
+		}
+		return this;
+	}
+};
+
+Tween.prototype.init.prototype = Tween.prototype;
+
+Tween.propHooks = {
+	_default: {
+		get: function( tween ) {
+			var result;
+
+			// Use a property on the element directly when it is not a DOM element,
+			// or when there is no matching style property that exists.
+			if ( tween.elem.nodeType !== 1 ||
+				tween.elem[ tween.prop ] != null && tween.elem.style[ tween.prop ] == null ) {
+				return tween.elem[ tween.prop ];
+			}
+
+			// Passing an empty string as a 3rd parameter to .css will automatically
+			// attempt a parseFloat and fallback to a string if the parse fails.
+			// Simple values such as "10px" are parsed to Float;
+			// complex values such as "rotate(1rad)" are returned as-is.
+			result = jQuery.css( tween.elem, tween.prop, "" );
+
+			// Empty strings, null, undefined and "auto" are converted to 0.
+			return !result || result === "auto" ? 0 : result;
+		},
+		set: function( tween ) {
+
+			// Use step hook for back compat.
+			// Use cssHook if its there.
+			// Use .style if available and use plain properties where available.
+			if ( jQuery.fx.step[ tween.prop ] ) {
+				jQuery.fx.step[ tween.prop ]( tween );
+			} else if ( tween.elem.nodeType === 1 && (
+				jQuery.cssHooks[ tween.prop ] ||
+					tween.elem.style[ finalPropName( tween.prop ) ] != null ) ) {
+				jQuery.style( tween.elem, tween.prop, tween.now + tween.unit );
+			} else {
+				tween.elem[ tween.prop ] = tween.now;
+			}
+		}
+	}
+};
+
+// Support: IE <=9 only
+// Panic based approach to setting things on disconnected nodes
+Tween.propHooks.scrollTop = Tween.propHooks.scrollLeft = {
+	set: function( tween ) {
+		if ( tween.elem.nodeType && tween.elem.parentNode ) {
+			tween.elem[ tween.prop ] = tween.now;
+		}
+	}
+};
+
+jQuery.easing = {
+	linear: function( p ) {
+		return p;
+	},
+	swing: function( p ) {
+		return 0.5 - Math.cos( p * Math.PI ) / 2;
+	},
+	_default: "swing"
+};
+
+jQuery.fx = Tween.prototype.init;
+
+// Back compat <1.8 extension point
+jQuery.fx.step = {};
+
+
+
+
+var
+	fxNow, inProgress,
+	rfxtypes = /^(?:toggle|show|hide)$/,
+	rrun = /queueHooks$/;
+
+function schedule() {
+	if ( inProgress ) {
+		if ( document.hidden === false && window.requestAnimationFrame ) {
+			window.requestAnimationFrame( schedule );
+		} else {
+			window.setTimeout( schedule, jQuery.fx.interval );
+		}
+
+		jQuery.fx.tick();
+	}
+}
+
+// Animations created synchronously will run synchronously
+function createFxNow() {
+	window.setTimeout( function() {
+		fxNow = undefined;
+	} );
+	return ( fxNow = Date.now() );
+}
+
+// Generate parameters to create a standard animation
+function genFx( type, includeWidth ) {
+	var which,
+		i = 0,
+		attrs = { height: type };
+
+	// If we include width, step value is 1 to do all cssExpand values,
+	// otherwise step value is 2 to skip over Left and Right
+	includeWidth = includeWidth ? 1 : 0;
+	for ( ; i < 4; i += 2 - includeWidth ) {
+		which = cssExpand[ i ];
+		attrs[ "margin" + which ] = attrs[ "padding" + which ] = type;
+	}
+
+	if ( includeWidth ) {
+		attrs.opacity = attrs.width = type;
+	}
+
+	return attrs;
+}
+
+function createTween( value, prop, animation ) {
+	var tween,
+		collection = ( Animation.tweeners[ prop ] || [] ).concat( Animation.tweeners[ "*" ] ),
+		index = 0,
+		length = collection.length;
+	for ( ; index < length; index++ ) {
+		if ( ( tween = collection[ index ].call( animation, prop, value ) ) ) {
+
+			// We're done with this property
+			return tween;
+		}
+	}
+}
+
+function defaultPrefilter( elem, props, opts ) {
+	var prop, value, toggle, hooks, oldfire, propTween, restoreDisplay, display,
+		isBox = "width" in props || "height" in props,
+		anim = this,
+		orig = {},
+		style = elem.style,
+		hidden = elem.nodeType && isHiddenWithinTree( elem ),
+		dataShow = dataPriv.get( elem, "fxshow" );
+
+	// Queue-skipping animations hijack the fx hooks
+	if ( !opts.queue ) {
+		hooks = jQuery._queueHooks( elem, "fx" );
+		if ( hooks.unqueued == null ) {
+			hooks.unqueued = 0;
+			oldfire = hooks.empty.fire;
+			hooks.empty.fire = function() {
+				if ( !hooks.unqueued ) {
+					oldfire();
+				}
+			};
+		}
+		hooks.unqueued++;
+
+		anim.always( function() {
+
+			// Ensure the complete handler is called before this completes
+			anim.always( function() {
+				hooks.unqueued--;
+				if ( !jQuery.queue( elem, "fx" ).length ) {
+					hooks.empty.fire();
+				}
+			} );
+		} );
+	}
+
+	// Detect show/hide animations
+	for ( prop in props ) {
+		value = props[ prop ];
+		if ( rfxtypes.test( value ) ) {
+			delete props[ prop ];
+			toggle = toggle || value === "toggle";
+			if ( value === ( hidden ? "hide" : "show" ) ) {
+
+				// Pretend to be hidden if this is a "show" and
+				// there is still data from a stopped show/hide
+				if ( value === "show" && dataShow && dataShow[ prop ] !== undefined ) {
+					hidden = true;
+
+				// Ignore all other no-op show/hide data
+				} else {
+					continue;
+				}
+			}
+			orig[ prop ] = dataShow && dataShow[ prop ] || jQuery.style( elem, prop );
+		}
+	}
+
+	// Bail out if this is a no-op like .hide().hide()
+	propTween = !jQuery.isEmptyObject( props );
+	if ( !propTween && jQuery.isEmptyObject( orig ) ) {
+		return;
+	}
+
+	// Restrict "overflow" and "display" styles during box animations
+	if ( isBox && elem.nodeType === 1 ) {
+
+		// Support: IE <=9 - 11, Edge 12 - 15
+		// Record all 3 overflow attributes because IE does not infer the shorthand
+		// from identically-valued overflowX and overflowY and Edge just mirrors
+		// the overflowX value there.
+		opts.overflow = [ style.overflow, style.overflowX, style.overflowY ];
+
+		// Identify a display type, preferring old show/hide data over the CSS cascade
+		restoreDisplay = dataShow && dataShow.display;
+		if ( restoreDisplay == null ) {
+			restoreDisplay = dataPriv.get( elem, "display" );
+		}
+		display = jQuery.css( elem, "display" );
+		if ( display === "none" ) {
+			if ( restoreDisplay ) {
+				display = restoreDisplay;
+			} else {
+
+				// Get nonempty value(s) by temporarily forcing visibility
+				showHide( [ elem ], true );
+				restoreDisplay = elem.style.display || restoreDisplay;
+				display = jQuery.css( elem, "display" );
+				showHide( [ elem ] );
+			}
+		}
+
+		// Animate inline elements as inline-block
+		if ( display === "inline" || display === "inline-block" && restoreDisplay != null ) {
+			if ( jQuery.css( elem, "float" ) === "none" ) {
+
+				// Restore the original display value at the end of pure show/hide animations
+				if ( !propTween ) {
+					anim.done( function() {
+						style.display = restoreDisplay;
+					} );
+					if ( restoreDisplay == null ) {
+						display = style.display;
+						restoreDisplay = display === "none" ? "" : display;
+					}
+				}
+				style.display = "inline-block";
+			}
+		}
+	}
+
+	if ( opts.overflow ) {
+		style.overflow = "hidden";
+		anim.always( function() {
+			style.overflow = opts.overflow[ 0 ];
+			style.overflowX = opts.overflow[ 1 ];
+			style.overflowY = opts.overflow[ 2 ];
+		} );
+	}
+
+	// Implement show/hide animations
+	propTween = false;
+	for ( prop in orig ) {
+
+		// General show/hide setup for this element animation
+		if ( !propTween ) {
+			if ( dataShow ) {
+				if ( "hidden" in dataShow ) {
+					hidden = dataShow.hidden;
+				}
+			} else {
+				dataShow = dataPriv.access( elem, "fxshow", { display: restoreDisplay } );
+			}
+
+			// Store hidden/visible for toggle so `.stop().toggle()` "reverses"
+			if ( toggle ) {
+				dataShow.hidden = !hidden;
+			}
+
+			// Show elements before animating them
+			if ( hidden ) {
+				showHide( [ elem ], true );
+			}
+
+			/* eslint-disable no-loop-func */
+
+			anim.done( function() {
+
+				/* eslint-enable no-loop-func */
+
+				// The final step of a "hide" animation is actually hiding the element
+				if ( !hidden ) {
+					showHide( [ elem ] );
+				}
+				dataPriv.remove( elem, "fxshow" );
+				for ( prop in orig ) {
+					jQuery.style( elem, prop, orig[ prop ] );
+				}
+			} );
+		}
+
+		// Per-property setup
+		propTween = createTween( hidden ? dataShow[ prop ] : 0, prop, anim );
+		if ( !( prop in dataShow ) ) {
+			dataShow[ prop ] = propTween.start;
+			if ( hidden ) {
+				propTween.end = propTween.start;
+				propTween.start = 0;
+			}
+		}
+	}
+}
+
+function propFilter( props, specialEasing ) {
+	var index, name, easing, value, hooks;
+
+	// camelCase, specialEasing and expand cssHook pass
+	for ( index in props ) {
+		name = camelCase( index );
+		easing = specialEasing[ name ];
+		value = props[ index ];
+		if ( Array.isArray( value ) ) {
+			easing = value[ 1 ];
+			value = props[ index ] = value[ 0 ];
+		}
+
+		if ( index !== name ) {
+			props[ name ] = value;
+			delete props[ index ];
+		}
+
+		hooks = jQuery.cssHooks[ name ];
+		if ( hooks && "expand" in hooks ) {
+			value = hooks.expand( value );
+			delete props[ name ];
+
+			// Not quite $.extend, this won't overwrite existing keys.
+			// Reusing 'index' because we have the correct "name"
+			for ( index in value ) {
+				if ( !( index in props ) ) {
+					props[ index ] = value[ index ];
+					specialEasing[ index ] = easing;
+				}
+			}
+		} else {
+			specialEasing[ name ] = easing;
+		}
+	}
+}
+
+function Animation( elem, properties, options ) {
+	var result,
+		stopped,
+		index = 0,
+		length = Animation.prefilters.length,
+		deferred = jQuery.Deferred().always( function() {
+
+			// Don't match elem in the :animated selector
+			delete tick.elem;
+		} ),
+		tick = function() {
+			if ( stopped ) {
+				return false;
+			}
+			var currentTime = fxNow || createFxNow(),
+				remaining = Math.max( 0, animation.startTime + animation.duration - currentTime ),
+
+				// Support: Android 2.3 only
+				// Archaic crash bug won't allow us to use `1 - ( 0.5 || 0 )` (#12497)
+				temp = remaining / animation.duration || 0,
+				percent = 1 - temp,
+				index = 0,
+				length = animation.tweens.length;
+
+			for ( ; index < length; index++ ) {
+				animation.tweens[ index ].run( percent );
+			}
+
+			deferred.notifyWith( elem, [ animation, percent, remaining ] );
+
+			// If there's more to do, yield
+			if ( percent < 1 && length ) {
+				return remaining;
+			}
+
+			// If this was an empty animation, synthesize a final progress notification
+			if ( !length ) {
+				deferred.notifyWith( elem, [ animation, 1, 0 ] );
+			}
+
+			// Resolve the animation and report its conclusion
+			deferred.resolveWith( elem, [ animation ] );
+			return false;
+		},
+		animation = deferred.promise( {
+			elem: elem,
+			props: jQuery.extend( {}, properties ),
+			opts: jQuery.extend( true, {
+				specialEasing: {},
+				easing: jQuery.easing._default
+			}, options ),
+			originalProperties: properties,
+			originalOptions: options,
+			startTime: fxNow || createFxNow(),
+			duration: options.duration,
+			tweens: [],
+			createTween: function( prop, end ) {
+				var tween = jQuery.Tween( elem, animation.opts, prop, end,
+					animation.opts.specialEasing[ prop ] || animation.opts.easing );
+				animation.tweens.push( tween );
+				return tween;
+			},
+			stop: function( gotoEnd ) {
+				var index = 0,
+
+					// If we are going to the end, we want to run all the tweens
+					// otherwise we skip this part
+					length = gotoEnd ? animation.tweens.length : 0;
+				if ( stopped ) {
+					return this;
+				}
+				stopped = true;
+				for ( ; index < length; index++ ) {
+					animation.tweens[ index ].run( 1 );
+				}
+
+				// Resolve when we played the last frame; otherwise, reject
+				if ( gotoEnd ) {
+					deferred.notifyWith( elem, [ animation, 1, 0 ] );
+					deferred.resolveWith( elem, [ animation, gotoEnd ] );
+				} else {
+					deferred.rejectWith( elem, [ animation, gotoEnd ] );
+				}
+				return this;
+			}
+		} ),
+		props = animation.props;
+
+	propFilter( props, animation.opts.specialEasing );
+
+	for ( ; index < length; index++ ) {
+		result = Animation.prefilters[ index ].call( animation, elem, props, animation.opts );
+		if ( result ) {
+			if ( isFunction( result.stop ) ) {
+				jQuery._queueHooks( animation.elem, animation.opts.queue ).stop =
+					result.stop.bind( result );
+			}
+			return result;
+		}
+	}
+
+	jQuery.map( props, createTween, animation );
+
+	if ( isFunction( animation.opts.start ) ) {
+		animation.opts.start.call( elem, animation );
+	}
+
+	// Attach callbacks from options
+	animation
+		.progress( animation.opts.progress )
+		.done( animation.opts.done, animation.opts.complete )
+		.fail( animation.opts.fail )
+		.always( animation.opts.always );
+
+	jQuery.fx.timer(
+		jQuery.extend( tick, {
+			elem: elem,
+			anim: animation,
+			queue: animation.opts.queue
+		} )
+	);
+
+	return animation;
+}
+
+jQuery.Animation = jQuery.extend( Animation, {
+
+	tweeners: {
+		"*": [ function( prop, value ) {
+			var tween = this.createTween( prop, value );
+			adjustCSS( tween.elem, prop, rcssNum.exec( value ), tween );
+			return tween;
+		} ]
+	},
+
+	tweener: function( props, callback ) {
+		if ( isFunction( props ) ) {
+			callback = props;
+			props = [ "*" ];
+		} else {
+			props = props.match( rnothtmlwhite );
+		}
+
+		var prop,
+			index = 0,
+			length = props.length;
+
+		for ( ; index < length; index++ ) {
+			prop = props[ index ];
+			Animation.tweeners[ prop ] = Animation.tweeners[ prop ] || [];
+			Animation.tweeners[ prop ].unshift( callback );
+		}
+	},
+
+	prefilters: [ defaultPrefilter ],
+
+	prefilter: function( callback, prepend ) {
+		if ( prepend ) {
+			Animation.prefilters.unshift( callback );
+		} else {
+			Animation.prefilters.push( callback );
+		}
+	}
+} );
+
+jQuery.speed = function( speed, easing, fn ) {
+	var opt = speed && typeof speed === "object" ? jQuery.extend( {}, speed ) : {
+		complete: fn || !fn && easing ||
+			isFunction( speed ) && speed,
+		duration: speed,
+		easing: fn && easing || easing && !isFunction( easing ) && easing
+	};
+
+	// Go to the end state if fx are off
+	if ( jQuery.fx.off ) {
+		opt.duration = 0;
+
+	} else {
+		if ( typeof opt.duration !== "number" ) {
+			if ( opt.duration in jQuery.fx.speeds ) {
+				opt.duration = jQuery.fx.speeds[ opt.duration ];
+
+			} else {
+				opt.duration = jQuery.fx.speeds._default;
+			}
+		}
+	}
+
+	// Normalize opt.queue - true/undefined/null -> "fx"
+	if ( opt.queue == null || opt.queue === true ) {
+		opt.queue = "fx";
+	}
+
+	// Queueing
+	opt.old = opt.complete;
+
+	opt.complete = function() {
+		if ( isFunction( opt.old ) ) {
+			opt.old.call( this );
+		}
+
+		if ( opt.queue ) {
+			jQuery.dequeue( this, opt.queue );
+		}
+	};
+
+	return opt;
+};
+
+jQuery.fn.extend( {
+	fadeTo: function( speed, to, easing, callback ) {
+
+		// Show any hidden elements after setting opacity to 0
+		return this.filter( isHiddenWithinTree ).css( "opacity", 0 ).show()
+
+			// Animate to the value specified
+			.end().animate( { opacity: to }, speed, easing, callback );
+	},
+	animate: function( prop, speed, easing, callback ) {
+		var empty = jQuery.isEmptyObject( prop ),
+			optall = jQuery.speed( speed, easing, callback ),
+			doAnimation = function() {
+
+				// Operate on a copy of prop so per-property easing won't be lost
+				var anim = Animation( this, jQuery.extend( {}, prop ), optall );
+
+				// Empty animations, or finishing resolves immediately
+				if ( empty || dataPriv.get( this, "finish" ) ) {
+					anim.stop( true );
+				}
+			};
+
+		doAnimation.finish = doAnimation;
+
+		return empty || optall.queue === false ?
+			this.each( doAnimation ) :
+			this.queue( optall.queue, doAnimation );
+	},
+	stop: function( type, clearQueue, gotoEnd ) {
+		var stopQueue = function( hooks ) {
+			var stop = hooks.stop;
+			delete hooks.stop;
+			stop( gotoEnd );
+		};
+
+		if ( typeof type !== "string" ) {
+			gotoEnd = clearQueue;
+			clearQueue = type;
+			type = undefined;
+		}
+		if ( clearQueue ) {
+			this.queue( type || "fx", [] );
+		}
+
+		return this.each( function() {
+			var dequeue = true,
+				index = type != null && type + "queueHooks",
+				timers = jQuery.timers,
+				data = dataPriv.get( this );
+
+			if ( index ) {
+				if ( data[ index ] && data[ index ].stop ) {
+					stopQueue( data[ index ] );
+				}
+			} else {
+				for ( index in data ) {
+					if ( data[ index ] && data[ index ].stop && rrun.test( index ) ) {
+						stopQueue( data[ index ] );
+					}
+				}
+			}
+
+			for ( index = timers.length; index--; ) {
+				if ( timers[ index ].elem === this &&
+					( type == null || timers[ index ].queue === type ) ) {
+
+					timers[ index ].anim.stop( gotoEnd );
+					dequeue = false;
+					timers.splice( index, 1 );
+				}
+			}
+
+			// Start the next in the queue if the last step wasn't forced.
+			// Timers currently will call their complete callbacks, which
+			// will dequeue but only if they were gotoEnd.
+			if ( dequeue || !gotoEnd ) {
+				jQuery.dequeue( this, type );
+			}
+		} );
+	},
+	finish: function( type ) {
+		if ( type !== false ) {
+			type = type || "fx";
+		}
+		return this.each( function() {
+			var index,
+				data = dataPriv.get( this ),
+				queue = data[ type + "queue" ],
+				hooks = data[ type + "queueHooks" ],
+				timers = jQuery.timers,
+				length = queue ? queue.length : 0;
+
+			// Enable finishing flag on private data
+			data.finish = true;
+
+			// Empty the queue first
+			jQuery.queue( this, type, [] );
+
+			if ( hooks && hooks.stop ) {
+				hooks.stop.call( this, true );
+			}
+
+			// Look for any active animations, and finish them
+			for ( index = timers.length; index--; ) {
+				if ( timers[ index ].elem === this && timers[ index ].queue === type ) {
+					timers[ index ].anim.stop( true );
+					timers.splice( index, 1 );
+				}
+			}
+
+			// Look for any animations in the old queue and finish them
+			for ( index = 0; index < length; index++ ) {
+				if ( queue[ index ] && queue[ index ].finish ) {
+					queue[ index ].finish.call( this );
+				}
+			}
+
+			// Turn off finishing flag
+			delete data.finish;
+		} );
+	}
+} );
+
+jQuery.each( [ "toggle", "show", "hide" ], function( _i, name ) {
+	var cssFn = jQuery.fn[ name ];
+	jQuery.fn[ name ] = function( speed, easing, callback ) {
+		return speed == null || typeof speed === "boolean" ?
+			cssFn.apply( this, arguments ) :
+			this.animate( genFx( name, true ), speed, easing, callback );
+	};
+} );
+
+// Generate shortcuts for custom animations
+jQuery.each( {
+	slideDown: genFx( "show" ),
+	slideUp: genFx( "hide" ),
+	slideToggle: genFx( "toggle" ),
+	fadeIn: { opacity: "show" },
+	fadeOut: { opacity: "hide" },
+	fadeToggle: { opacity: "toggle" }
+}, function( name, props ) {
+	jQuery.fn[ name ] = function( speed, easing, callback ) {
+		return this.animate( props, speed, easing, callback );
+	};
+} );
+
+jQuery.timers = [];
+jQuery.fx.tick = function() {
+	var timer,
+		i = 0,
+		timers = jQuery.timers;
+
+	fxNow = Date.now();
+
+	for ( ; i < timers.length; i++ ) {
+		timer = timers[ i ];
+
+		// Run the timer and safely remove it when done (allowing for external removal)
+		if ( !timer() && timers[ i ] === timer ) {
+			timers.splice( i--, 1 );
+		}
+	}
+
+	if ( !timers.length ) {
+		jQuery.fx.stop();
+	}
+	fxNow = undefined;
+};
+
+jQuery.fx.timer = function( timer ) {
+	jQuery.timers.push( timer );
+	jQuery.fx.start();
+};
+
+jQuery.fx.interval = 13;
+jQuery.fx.start = function() {
+	if ( inProgress ) {
+		return;
+	}
+
+	inProgress = true;
+	schedule();
+};
+
+jQuery.fx.stop = function() {
+	inProgress = null;
+};
+
+jQuery.fx.speeds = {
+	slow: 600,
+	fast: 200,
+
+	// Default speed
+	_default: 400
+};
+
+
+// Based off of the plugin by Clint Helfers, with permission.
+// https://web.archive.org/web/20100324014747/http://blindsignals.com/index.php/2009/07/jquery-delay/
+jQuery.fn.delay = function( time, type ) {
+	time = jQuery.fx ? jQuery.fx.speeds[ time ] || time : time;
+	type = type || "fx";
+
+	return this.queue( type, function( next, hooks ) {
+		var timeout = window.setTimeout( next, time );
+		hooks.stop = function() {
+			window.clearTimeout( timeout );
+		};
+	} );
+};
+
+
+( function() {
+	var input = document.createElement( "input" ),
+		select = document.createElement( "select" ),
+		opt = select.appendChild( document.createElement( "option" ) );
+
+	input.type = "checkbox";
+
+	// Support: Android <=4.3 only
+	// Default value for a checkbox should be "on"
+	support.checkOn = input.value !== "";
+
+	// Support: IE <=11 only
+	// Must access selectedIndex to make default options select
+	support.optSelected = opt.selected;
+
+	// Support: IE <=11 only
+	// An input loses its value after becoming a radio
+	input = document.createElement( "input" );
+	input.value = "t";
+	input.type = "radio";
+	support.radioValue = input.value === "t";
+} )();
+
+
+var boolHook,
+	attrHandle = jQuery.expr.attrHandle;
+
+jQuery.fn.extend( {
+	attr: function( name, value ) {
+		return access( this, jQuery.attr, name, value, arguments.length > 1 );
+	},
+
+	removeAttr: function( name ) {
+		return this.each( function() {
+			jQuery.removeAttr( this, name );
+		} );
+	}
+} );
+
+jQuery.extend( {
+	attr: function( elem, name, value ) {
+		var ret, hooks,
+			nType = elem.nodeType;
+
+		// Don't get/set attributes on text, comment and attribute nodes
+		if ( nType === 3 || nType === 8 || nType === 2 ) {
+			return;
+		}
+
+		// Fallback to prop when attributes are not supported
+		if ( typeof elem.getAttribute === "undefined" ) {
+			return jQuery.prop( elem, name, value );
+		}
+
+		// Attribute hooks are determined by the lowercase version
+		// Grab necessary hook if one is defined
+		if ( nType !== 1 || !jQuery.isXMLDoc( elem ) ) {
+			hooks = jQuery.attrHooks[ name.toLowerCase() ] ||
+				( jQuery.expr.match.bool.test( name ) ? boolHook : undefined );
+		}
+
+		if ( value !== undefined ) {
+			if ( value === null ) {
+				jQuery.removeAttr( elem, name );
+				return;
+			}
+
+			if ( hooks && "set" in hooks &&
+				( ret = hooks.set( elem, value, name ) ) !== undefined ) {
+				return ret;
+			}
+
+			elem.setAttribute( name, value + "" );
+			return value;
+		}
+
+		if ( hooks && "get" in hooks && ( ret = hooks.get( elem, name ) ) !== null ) {
+			return ret;
+		}
+
+		ret = jQuery.find.attr( elem, name );
+
+		// Non-existent attributes return null, we normalize to undefined
+		return ret == null ? undefined : ret;
+	},
+
+	attrHooks: {
+		type: {
+			set: function( elem, value ) {
+				if ( !support.radioValue && value === "radio" &&
+					nodeName( elem, "input" ) ) {
+					var val = elem.value;
+					elem.setAttribute( "type", value );
+					if ( val ) {
+						elem.value = val;
+					}
+					return value;
+				}
+			}
+		}
+	},
+
+	removeAttr: function( elem, value ) {
+		var name,
+			i = 0,
+
+			// Attribute names can contain non-HTML whitespace characters
+			// https://html.spec.whatwg.org/multipage/syntax.html#attributes-2
+			attrNames = value && value.match( rnothtmlwhite );
+
+		if ( attrNames && elem.nodeType === 1 ) {
+			while ( ( name = attrNames[ i++ ] ) ) {
+				elem.removeAttribute( name );
+			}
+		}
+	}
+} );
+
+// Hooks for boolean attributes
+boolHook = {
+	set: function( elem, value, name ) {
+		if ( value === false ) {
+
+			// Remove boolean attributes when set to false
+			jQuery.removeAttr( elem, name );
+		} else {
+			elem.setAttribute( name, name );
+		}
+		return name;
+	}
+};
+
+jQuery.each( jQuery.expr.match.bool.source.match( /\w+/g ), function( _i, name ) {
+	var getter = attrHandle[ name ] || jQuery.find.attr;
+
+	attrHandle[ name ] = function( elem, name, isXML ) {
+		var ret, handle,
+			lowercaseName = name.toLowerCase();
+
+		if ( !isXML ) {
+
+			// Avoid an infinite loop by temporarily removing this function from the getter
+			handle = attrHandle[ lowercaseName ];
+			attrHandle[ lowercaseName ] = ret;
+			ret = getter( elem, name, isXML ) != null ?
+				lowercaseName :
+				null;
+			attrHandle[ lowercaseName ] = handle;
+		}
+		return ret;
+	};
+} );
+
+
+
+
+var rfocusable = /^(?:input|select|textarea|button)$/i,
+	rclickable = /^(?:a|area)$/i;
+
+jQuery.fn.extend( {
+	prop: function( name, value ) {
+		return access( this, jQuery.prop, name, value, arguments.length > 1 );
+	},
+
+	removeProp: function( name ) {
+		return this.each( function() {
+			delete this[ jQuery.propFix[ name ] || name ];
+		} );
+	}
+} );
+
+jQuery.extend( {
+	prop: function( elem, name, value ) {
+		var ret, hooks,
+			nType = elem.nodeType;
+
+		// Don't get/set properties on text, comment and attribute nodes
+		if ( nType === 3 || nType === 8 || nType === 2 ) {
+			return;
+		}
+
+		if ( nType !== 1 || !jQuery.isXMLDoc( elem ) ) {
+
+			// Fix name and attach hooks
+			name = jQuery.propFix[ name ] || name;
+			hooks = jQuery.propHooks[ name ];
+		}
+
+		if ( value !== undefined ) {
+			if ( hooks && "set" in hooks &&
+				( ret = hooks.set( elem, value, name ) ) !== undefined ) {
+				return ret;
+			}
+
+			return ( elem[ name ] = value );
+		}
+
+		if ( hooks && "get" in hooks && ( ret = hooks.get( elem, name ) ) !== null ) {
+			return ret;
+		}
+
+		return elem[ name ];
+	},
+
+	propHooks: {
+		tabIndex: {
+			get: function( elem ) {
+
+				// Support: IE <=9 - 11 only
+				// elem.tabIndex doesn't always return the
+				// correct value when it hasn't been explicitly set
+				// https://web.archive.org/web/20141116233347/http://fluidproject.org/blog/2008/01/09/getting-setting-and-removing-tabindex-values-with-javascript/
+				// Use proper attribute retrieval(#12072)
+				var tabindex = jQuery.find.attr( elem, "tabindex" );
+
+				if ( tabindex ) {
+					return parseInt( tabindex, 10 );
+				}
+
+				if (
+					rfocusable.test( elem.nodeName ) ||
+					rclickable.test( elem.nodeName ) &&
+					elem.href
+				) {
+					return 0;
+				}
+
+				return -1;
+			}
+		}
+	},
+
+	propFix: {
+		"for": "htmlFor",
+		"class": "className"
+	}
+} );
+
+// Support: IE <=11 only
+// Accessing the selectedIndex property
+// forces the browser to respect setting selected
+// on the option
+// The getter ensures a default option is selected
+// when in an optgroup
+// eslint rule "no-unused-expressions" is disabled for this code
+// since it considers such accessions noop
+if ( !support.optSelected ) {
+	jQuery.propHooks.selected = {
+		get: function( elem ) {
+
+			/* eslint no-unused-expressions: "off" */
+
+			var parent = elem.parentNode;
+			if ( parent && parent.parentNode ) {
+				parent.parentNode.selectedIndex;
+			}
+			return null;
+		},
+		set: function( elem ) {
+
+			/* eslint no-unused-expressions: "off" */
+
+			var parent = elem.parentNode;
+			if ( parent ) {
+				parent.selectedIndex;
+
+				if ( parent.parentNode ) {
+					parent.parentNode.selectedIndex;
+				}
+			}
+		}
+	};
+}
+
+jQuery.each( [
+	"tabIndex",
+	"readOnly",
+	"maxLength",
+	"cellSpacing",
+	"cellPadding",
+	"rowSpan",
+	"colSpan",
+	"useMap",
+	"frameBorder",
+	"contentEditable"
+], function() {
+	jQuery.propFix[ this.toLowerCase() ] = this;
+} );
+
+
+
+
+	// Strip and collapse whitespace according to HTML spec
+	// https://infra.spec.whatwg.org/#strip-and-collapse-ascii-whitespace
+	function stripAndCollapse( value ) {
+		var tokens = value.match( rnothtmlwhite ) || [];
+		return tokens.join( " " );
+	}
+
+
+function getClass( elem ) {
+	return elem.getAttribute && elem.getAttribute( "class" ) || "";
+}
+
+function classesToArray( value ) {
+	if ( Array.isArray( value ) ) {
+		return value;
+	}
+	if ( typeof value === "string" ) {
+		return value.match( rnothtmlwhite ) || [];
+	}
+	return [];
+}
+
+jQuery.fn.extend( {
+	addClass: function( value ) {
+		var classes, elem, cur, curValue, clazz, j, finalValue,
+			i = 0;
+
+		if ( isFunction( value ) ) {
+			return this.each( function( j ) {
+				jQuery( this ).addClass( value.call( this, j, getClass( this ) ) );
+			} );
+		}
+
+		classes = classesToArray( value );
+
+		if ( classes.length ) {
+			while ( ( elem = this[ i++ ] ) ) {
+				curValue = getClass( elem );
+				cur = elem.nodeType === 1 && ( " " + stripAndCollapse( curValue ) + " " );
+
+				if ( cur ) {
+					j = 0;
+					while ( ( clazz = classes[ j++ ] ) ) {
+						if ( cur.indexOf( " " + clazz + " " ) < 0 ) {
+							cur += clazz + " ";
+						}
+					}
+
+					// Only assign if different to avoid unneeded rendering.
+					finalValue = stripAndCollapse( cur );
+					if ( curValue !== finalValue ) {
+						elem.setAttribute( "class", finalValue );
+					}
+				}
+			}
+		}
+
+		return this;
+	},
+
+	removeClass: function( value ) {
+		var classes, elem, cur, curValue, clazz, j, finalValue,
+			i = 0;
+
+		if ( isFunction( value ) ) {
+			return this.each( function( j ) {
+				jQuery( this ).removeClass( value.call( this, j, getClass( this ) ) );
+			} );
+		}
+
+		if ( !arguments.length ) {
+			return this.attr( "class", "" );
+		}
+
+		classes = classesToArray( value );
+
+		if ( classes.length ) {
+			while ( ( elem = this[ i++ ] ) ) {
+				curValue = getClass( elem );
+
+				// This expression is here for better compressibility (see addClass)
+				cur = elem.nodeType === 1 && ( " " + stripAndCollapse( curValue ) + " " );
+
+				if ( cur ) {
+					j = 0;
+					while ( ( clazz = classes[ j++ ] ) ) {
+
+						// Remove *all* instances
+						while ( cur.indexOf( " " + clazz + " " ) > -1 ) {
+							cur = cur.replace( " " + clazz + " ", " " );
+						}
+					}
+
+					// Only assign if different to avoid unneeded rendering.
+					finalValue = stripAndCollapse( cur );
+					if ( curValue !== finalValue ) {
+						elem.setAttribute( "class", finalValue );
+					}
+				}
+			}
+		}
+
+		return this;
+	},
+
+	toggleClass: function( value, stateVal ) {
+		var type = typeof value,
+			isValidValue = type === "string" || Array.isArray( value );
+
+		if ( typeof stateVal === "boolean" && isValidValue ) {
+			return stateVal ? this.addClass( value ) : this.removeClass( value );
+		}
+
+		if ( isFunction( value ) ) {
+			return this.each( function( i ) {
+				jQuery( this ).toggleClass(
+					value.call( this, i, getClass( this ), stateVal ),
+					stateVal
+				);
+			} );
+		}
+
+		return this.each( function() {
+			var className, i, self, classNames;
+
+			if ( isValidValue ) {
+
+				// Toggle individual class names
+				i = 0;
+				self = jQuery( this );
+				classNames = classesToArray( value );
+
+				while ( ( className = classNames[ i++ ] ) ) {
+
+					// Check each className given, space separated list
+					if ( self.hasClass( className ) ) {
+						self.removeClass( className );
+					} else {
+						self.addClass( className );
+					}
+				}
+
+			// Toggle whole class name
+			} else if ( value === undefined || type === "boolean" ) {
+				className = getClass( this );
+				if ( className ) {
+
+					// Store className if set
+					dataPriv.set( this, "__className__", className );
+				}
+
+				// If the element has a class name or if we're passed `false`,
+				// then remove the whole classname (if there was one, the above saved it).
+				// Otherwise bring back whatever was previously saved (if anything),
+				// falling back to the empty string if nothing was stored.
+				if ( this.setAttribute ) {
+					this.setAttribute( "class",
+						className || value === false ?
+							"" :
+							dataPriv.get( this, "__className__" ) || ""
+					);
+				}
+			}
+		} );
+	},
+
+	hasClass: function( selector ) {
+		var className, elem,
+			i = 0;
+
+		className = " " + selector + " ";
+		while ( ( elem = this[ i++ ] ) ) {
+			if ( elem.nodeType === 1 &&
+				( " " + stripAndCollapse( getClass( elem ) ) + " " ).indexOf( className ) > -1 ) {
+				return true;
+			}
+		}
+
+		return false;
+	}
+} );
+
+
+
+
+var rreturn = /\r/g;
+
+jQuery.fn.extend( {
+	val: function( value ) {
+		var hooks, ret, valueIsFunction,
+			elem = this[ 0 ];
+
+		if ( !arguments.length ) {
+			if ( elem ) {
+				hooks = jQuery.valHooks[ elem.type ] ||
+					jQuery.valHooks[ elem.nodeName.toLowerCase() ];
+
+				if ( hooks &&
+					"get" in hooks &&
+					( ret = hooks.get( elem, "value" ) ) !== undefined
+				) {
+					return ret;
+				}
+
+				ret = elem.value;
+
+				// Handle most common string cases
+				if ( typeof ret === "string" ) {
+					return ret.replace( rreturn, "" );
+				}
+
+				// Handle cases where value is null/undef or number
+				return ret == null ? "" : ret;
+			}
+
+			return;
+		}
+
+		valueIsFunction = isFunction( value );
+
+		return this.each( function( i ) {
+			var val;
+
+			if ( this.nodeType !== 1 ) {
+				return;
+			}
+
+			if ( valueIsFunction ) {
+				val = value.call( this, i, jQuery( this ).val() );
+			} else {
+				val = value;
+			}
+
+			// Treat null/undefined as ""; convert numbers to string
+			if ( val == null ) {
+				val = "";
+
+			} else if ( typeof val === "number" ) {
+				val += "";
+
+			} else if ( Array.isArray( val ) ) {
+				val = jQuery.map( val, function( value ) {
+					return value == null ? "" : value + "";
+				} );
+			}
+
+			hooks = jQuery.valHooks[ this.type ] || jQuery.valHooks[ this.nodeName.toLowerCase() ];
+
+			// If set returns undefined, fall back to normal setting
+			if ( !hooks || !( "set" in hooks ) || hooks.set( this, val, "value" ) === undefined ) {
+				this.value = val;
+			}
+		} );
+	}
+} );
+
+jQuery.extend( {
+	valHooks: {
+		option: {
+			get: function( elem ) {
+
+				var val = jQuery.find.attr( elem, "value" );
+				return val != null ?
+					val :
+
+					// Support: IE <=10 - 11 only
+					// option.text throws exceptions (#14686, #14858)
+					// Strip and collapse whitespace
+					// https://html.spec.whatwg.org/#strip-and-collapse-whitespace
+					stripAndCollapse( jQuery.text( elem ) );
+			}
+		},
+		select: {
+			get: function( elem ) {
+				var value, option, i,
+					options = elem.options,
+					index = elem.selectedIndex,
+					one = elem.type === "select-one",
+					values = one ? null : [],
+					max = one ? index + 1 : options.length;
+
+				if ( index < 0 ) {
+					i = max;
+
+				} else {
+					i = one ? index : 0;
+				}
+
+				// Loop through all the selected options
+				for ( ; i < max; i++ ) {
+					option = options[ i ];
+
+					// Support: IE <=9 only
+					// IE8-9 doesn't update selected after form reset (#2551)
+					if ( ( option.selected || i === index ) &&
+
+							// Don't return options that are disabled or in a disabled optgroup
+							!option.disabled &&
+							( !option.parentNode.disabled ||
+								!nodeName( option.parentNode, "optgroup" ) ) ) {
+
+						// Get the specific value for the option
+						value = jQuery( option ).val();
+
+						// We don't need an array for one selects
+						if ( one ) {
+							return value;
+						}
+
+						// Multi-Selects return an array
+						values.push( value );
+					}
+				}
+
+				return values;
+			},
+
+			set: function( elem, value ) {
+				var optionSet, option,
+					options = elem.options,
+					values = jQuery.makeArray( value ),
+					i = options.length;
+
+				while ( i-- ) {
+					option = options[ i ];
+
+					/* eslint-disable no-cond-assign */
+
+					if ( option.selected =
+						jQuery.inArray( jQuery.valHooks.option.get( option ), values ) > -1
+					) {
+						optionSet = true;
+					}
+
+					/* eslint-enable no-cond-assign */
+				}
+
+				// Force browsers to behave consistently when non-matching value is set
+				if ( !optionSet ) {
+					elem.selectedIndex = -1;
+				}
+				return values;
+			}
+		}
+	}
+} );
+
+// Radios and checkboxes getter/setter
+jQuery.each( [ "radio", "checkbox" ], function() {
+	jQuery.valHooks[ this ] = {
+		set: function( elem, value ) {
+			if ( Array.isArray( value ) ) {
+				return ( elem.checked = jQuery.inArray( jQuery( elem ).val(), value ) > -1 );
+			}
+		}
+	};
+	if ( !support.checkOn ) {
+		jQuery.valHooks[ this ].get = function( elem ) {
+			return elem.getAttribute( "value" ) === null ? "on" : elem.value;
+		};
+	}
+} );
+
+
+
+
+// Return jQuery for attributes-only inclusion
+
+
+support.focusin = "onfocusin" in window;
+
+
+var rfocusMorph = /^(?:focusinfocus|focusoutblur)$/,
+	stopPropagationCallback = function( e ) {
+		e.stopPropagation();
+	};
+
+jQuery.extend( jQuery.event, {
+
+	trigger: function( event, data, elem, onlyHandlers ) {
+
+		var i, cur, tmp, bubbleType, ontype, handle, special, lastElement,
+			eventPath = [ elem || document ],
+			type = hasOwn.call( event, "type" ) ? event.type : event,
+			namespaces = hasOwn.call( event, "namespace" ) ? event.namespace.split( "." ) : [];
+
+		cur = lastElement = tmp = elem = elem || document;
+
+		// Don't do events on text and comment nodes
+		if ( elem.nodeType === 3 || elem.nodeType === 8 ) {
+			return;
+		}
+
+		// focus/blur morphs to focusin/out; ensure we're not firing them right now
+		if ( rfocusMorph.test( type + jQuery.event.triggered ) ) {
+			return;
+		}
+
+		if ( type.indexOf( "." ) > -1 ) {
+
+			// Namespaced trigger; create a regexp to match event type in handle()
+			namespaces = type.split( "." );
+			type = namespaces.shift();
+			namespaces.sort();
+		}
+		ontype = type.indexOf( ":" ) < 0 && "on" + type;
+
+		// Caller can pass in a jQuery.Event object, Object, or just an event type string
+		event = event[ jQuery.expando ] ?
+			event :
+			new jQuery.Event( type, typeof event === "object" && event );
+
+		// Trigger bitmask: & 1 for native handlers; & 2 for jQuery (always true)
+		event.isTrigger = onlyHandlers ? 2 : 3;
+		event.namespace = namespaces.join( "." );
+		event.rnamespace = event.namespace ?
+			new RegExp( "(^|\\.)" + namespaces.join( "\\.(?:.*\\.|)" ) + "(\\.|$)" ) :
+			null;
+
+		// Clean up the event in case it is being reused
+		event.result = undefined;
+		if ( !event.target ) {
+			event.target = elem;
+		}
+
+		// Clone any incoming data and prepend the event, creating the handler arg list
+		data = data == null ?
+			[ event ] :
+			jQuery.makeArray( data, [ event ] );
+
+		// Allow special events to draw outside the lines
+		special = jQuery.event.special[ type ] || {};
+		if ( !onlyHandlers && special.trigger && special.trigger.apply( elem, data ) === false ) {
+			return;
+		}
+
+		// Determine event propagation path in advance, per W3C events spec (#9951)
+		// Bubble up to document, then to window; watch for a global ownerDocument var (#9724)
+		if ( !onlyHandlers && !special.noBubble && !isWindow( elem ) ) {
+
+			bubbleType = special.delegateType || type;
+			if ( !rfocusMorph.test( bubbleType + type ) ) {
+				cur = cur.parentNode;
+			}
+			for ( ; cur; cur = cur.parentNode ) {
+				eventPath.push( cur );
+				tmp = cur;
+			}
+
+			// Only add window if we got to document (e.g., not plain obj or detached DOM)
+			if ( tmp === ( elem.ownerDocument || document ) ) {
+				eventPath.push( tmp.defaultView || tmp.parentWindow || window );
+			}
+		}
+
+		// Fire handlers on the event path
+		i = 0;
+		while ( ( cur = eventPath[ i++ ] ) && !event.isPropagationStopped() ) {
+			lastElement = cur;
+			event.type = i > 1 ?
+				bubbleType :
+				special.bindType || type;
+
+			// jQuery handler
+			handle = ( dataPriv.get( cur, "events" ) || Object.create( null ) )[ event.type ] &&
+				dataPriv.get( cur, "handle" );
+			if ( handle ) {
+				handle.apply( cur, data );
+			}
+
+			// Native handler
+			handle = ontype && cur[ ontype ];
+			if ( handle && handle.apply && acceptData( cur ) ) {
+				event.result = handle.apply( cur, data );
+				if ( event.result === false ) {
+					event.preventDefault();
+				}
+			}
+		}
+		event.type = type;
+
+		// If nobody prevented the default action, do it now
+		if ( !onlyHandlers && !event.isDefaultPrevented() ) {
+
+			if ( ( !special._default ||
+				special._default.apply( eventPath.pop(), data ) === false ) &&
+				acceptData( elem ) ) {
+
+				// Call a native DOM method on the target with the same name as the event.
+				// Don't do default actions on window, that's where global variables be (#6170)
+				if ( ontype && isFunction( elem[ type ] ) && !isWindow( elem ) ) {
+
+					// Don't re-trigger an onFOO event when we call its FOO() method
+					tmp = elem[ ontype ];
+
+					if ( tmp ) {
+						elem[ ontype ] = null;
+					}
+
+					// Prevent re-triggering of the same event, since we already bubbled it above
+					jQuery.event.triggered = type;
+
+					if ( event.isPropagationStopped() ) {
+						lastElement.addEventListener( type, stopPropagationCallback );
+					}
+
+					elem[ type ]();
+
+					if ( event.isPropagationStopped() ) {
+						lastElement.removeEventListener( type, stopPropagationCallback );
+					}
+
+					jQuery.event.triggered = undefined;
+
+					if ( tmp ) {
+						elem[ ontype ] = tmp;
+					}
+				}
+			}
+		}
+
+		return event.result;
+	},
+
+	// Piggyback on a donor event to simulate a different one
+	// Used only for `focus(in | out)` events
+	simulate: function( type, elem, event ) {
+		var e = jQuery.extend(
+			new jQuery.Event(),
+			event,
+			{
+				type: type,
+				isSimulated: true
+			}
+		);
+
+		jQuery.event.trigger( e, null, elem );
+	}
+
+} );
+
+jQuery.fn.extend( {
+
+	trigger: function( type, data ) {
+		return this.each( function() {
+			jQuery.event.trigger( type, data, this );
+		} );
+	},
+	triggerHandler: function( type, data ) {
+		var elem = this[ 0 ];
+		if ( elem ) {
+			return jQuery.event.trigger( type, data, elem, true );
+		}
+	}
+} );
+
+
+// Support: Firefox <=44
+// Firefox doesn't have focus(in | out) events
+// Related ticket - https://bugzilla.mozilla.org/show_bug.cgi?id=687787
+//
+// Support: Chrome <=48 - 49, Safari <=9.0 - 9.1
+// focus(in | out) events fire after focus & blur events,
+// which is spec violation - http://www.w3.org/TR/DOM-Level-3-Events/#events-focusevent-event-order
+// Related ticket - https://bugs.chromium.org/p/chromium/issues/detail?id=449857
+if ( !support.focusin ) {
+	jQuery.each( { focus: "focusin", blur: "focusout" }, function( orig, fix ) {
+
+		// Attach a single capturing handler on the document while someone wants focusin/focusout
+		var handler = function( event ) {
+			jQuery.event.simulate( fix, event.target, jQuery.event.fix( event ) );
+		};
+
+		jQuery.event.special[ fix ] = {
+			setup: function() {
+
+				// Handle: regular nodes (via `this.ownerDocument`), window
+				// (via `this.document`) & document (via `this`).
+				var doc = this.ownerDocument || this.document || this,
+					attaches = dataPriv.access( doc, fix );
+
+				if ( !attaches ) {
+					doc.addEventListener( orig, handler, true );
+				}
+				dataPriv.access( doc, fix, ( attaches || 0 ) + 1 );
+			},
+			teardown: function() {
+				var doc = this.ownerDocument || this.document || this,
+					attaches = dataPriv.access( doc, fix ) - 1;
+
+				if ( !attaches ) {
+					doc.removeEventListener( orig, handler, true );
+					dataPriv.remove( doc, fix );
+
+				} else {
+					dataPriv.access( doc, fix, attaches );
+				}
+			}
+		};
+	} );
+}
+var location = window.location;
+
+var nonce = { guid: Date.now() };
+
+var rquery = ( /\?/ );
+
+
+
+// Cross-browser xml parsing
+jQuery.parseXML = function( data ) {
+	var xml, parserErrorElem;
+	if ( !data || typeof data !== "string" ) {
+		return null;
+	}
+
+	// Support: IE 9 - 11 only
+	// IE throws on parseFromString with invalid input.
+	try {
+		xml = ( new window.DOMParser() ).parseFromString( data, "text/xml" );
+	} catch ( e ) {}
+
+	parserErrorElem = xml && xml.getElementsByTagName( "parsererror" )[ 0 ];
+	if ( !xml || parserErrorElem ) {
+		jQuery.error( "Invalid XML: " + (
+			parserErrorElem ?
+				jQuery.map( parserErrorElem.childNodes, function( el ) {
+					return el.textContent;
+				} ).join( "\n" ) :
+				data
+		) );
+	}
+	return xml;
+};
+
+
+var
+	rbracket = /\[\]$/,
+	rCRLF = /\r?\n/g,
+	rsubmitterTypes = /^(?:submit|button|image|reset|file)$/i,
+	rsubmittable = /^(?:input|select|textarea|keygen)/i;
+
+function buildParams( prefix, obj, traditional, add ) {
+	var name;
+
+	if ( Array.isArray( obj ) ) {
+
+		// Serialize array item.
+		jQuery.each( obj, function( i, v ) {
+			if ( traditional || rbracket.test( prefix ) ) {
+
+				// Treat each array item as a scalar.
+				add( prefix, v );
+
+			} else {
+
+				// Item is non-scalar (array or object), encode its numeric index.
+				buildParams(
+					prefix + "[" + ( typeof v === "object" && v != null ? i : "" ) + "]",
+					v,
+					traditional,
+					add
+				);
+			}
+		} );
+
+	} else if ( !traditional && toType( obj ) === "object" ) {
+
+		// Serialize object item.
+		for ( name in obj ) {
+			buildParams( prefix + "[" + name + "]", obj[ name ], traditional, add );
+		}
+
+	} else {
+
+		// Serialize scalar item.
+		add( prefix, obj );
+	}
+}
+
+// Serialize an array of form elements or a set of
+// key/values into a query string
+jQuery.param = function( a, traditional ) {
+	var prefix,
+		s = [],
+		add = function( key, valueOrFunction ) {
+
+			// If value is a function, invoke it and use its return value
+			var value = isFunction( valueOrFunction ) ?
+				valueOrFunction() :
+				valueOrFunction;
+
+			s[ s.length ] = encodeURIComponent( key ) + "=" +
+				encodeURIComponent( value == null ? "" : value );
+		};
+
+	if ( a == null ) {
+		return "";
+	}
+
+	// If an array was passed in, assume that it is an array of form elements.
+	if ( Array.isArray( a ) || ( a.jquery && !jQuery.isPlainObject( a ) ) ) {
+
+		// Serialize the form elements
+		jQuery.each( a, function() {
+			add( this.name, this.value );
+		} );
+
+	} else {
+
+		// If traditional, encode the "old" way (the way 1.3.2 or older
+		// did it), otherwise encode params recursively.
+		for ( prefix in a ) {
+			buildParams( prefix, a[ prefix ], traditional, add );
+		}
+	}
+
+	// Return the resulting serialization
+	return s.join( "&" );
+};
+
+jQuery.fn.extend( {
+	serialize: function() {
+		return jQuery.param( this.serializeArray() );
+	},
+	serializeArray: function() {
+		return this.map( function() {
+
+			// Can add propHook for "elements" to filter or add form elements
+			var elements = jQuery.prop( this, "elements" );
+			return elements ? jQuery.makeArray( elements ) : this;
+		} ).filter( function() {
+			var type = this.type;
+
+			// Use .is( ":disabled" ) so that fieldset[disabled] works
+			return this.name && !jQuery( this ).is( ":disabled" ) &&
+				rsubmittable.test( this.nodeName ) && !rsubmitterTypes.test( type ) &&
+				( this.checked || !rcheckableType.test( type ) );
+		} ).map( function( _i, elem ) {
+			var val = jQuery( this ).val();
+
+			if ( val == null ) {
+				return null;
+			}
+
+			if ( Array.isArray( val ) ) {
+				return jQuery.map( val, function( val ) {
+					return { name: elem.name, value: val.replace( rCRLF, "\r\n" ) };
+				} );
+			}
+
+			return { name: elem.name, value: val.replace( rCRLF, "\r\n" ) };
+		} ).get();
+	}
+} );
+
+
+var
+	r20 = /%20/g,
+	rhash = /#.*$/,
+	rantiCache = /([?&])_=[^&]*/,
+	rheaders = /^(.*?):[ \t]*([^\r\n]*)$/mg,
+
+	// #7653, #8125, #8152: local protocol detection
+	rlocalProtocol = /^(?:about|app|app-storage|.+-extension|file|res|widget):$/,
+	rnoContent = /^(?:GET|HEAD)$/,
+	rprotocol = /^\/\//,
+
+	/* Prefilters
+	 * 1) They are useful to introduce custom dataTypes (see ajax/jsonp.js for an example)
+	 * 2) These are called:
+	 *    - BEFORE asking for a transport
+	 *    - AFTER param serialization (s.data is a string if s.processData is true)
+	 * 3) key is the dataType
+	 * 4) the catchall symbol "*" can be used
+	 * 5) execution will start with transport dataType and THEN continue down to "*" if needed
+	 */
+	prefilters = {},
+
+	/* Transports bindings
+	 * 1) key is the dataType
+	 * 2) the catchall symbol "*" can be used
+	 * 3) selection will start with transport dataType and THEN go to "*" if needed
+	 */
+	transports = {},
+
+	// Avoid comment-prolog char sequence (#10098); must appease lint and evade compression
+	allTypes = "*/".concat( "*" ),
+
+	// Anchor tag for parsing the document origin
+	originAnchor = document.createElement( "a" );
+
+originAnchor.href = location.href;
+
+// Base "constructor" for jQuery.ajaxPrefilter and jQuery.ajaxTransport
+function addToPrefiltersOrTransports( structure ) {
+
+	// dataTypeExpression is optional and defaults to "*"
+	return function( dataTypeExpression, func ) {
+
+		if ( typeof dataTypeExpression !== "string" ) {
+			func = dataTypeExpression;
+			dataTypeExpression = "*";
+		}
+
+		var dataType,
+			i = 0,
+			dataTypes = dataTypeExpression.toLowerCase().match( rnothtmlwhite ) || [];
+
+		if ( isFunction( func ) ) {
+
+			// For each dataType in the dataTypeExpression
+			while ( ( dataType = dataTypes[ i++ ] ) ) {
+
+				// Prepend if requested
+				if ( dataType[ 0 ] === "+" ) {
+					dataType = dataType.slice( 1 ) || "*";
+					( structure[ dataType ] = structure[ dataType ] || [] ).unshift( func );
+
+				// Otherwise append
+				} else {
+					( structure[ dataType ] = structure[ dataType ] || [] ).push( func );
+				}
+			}
+		}
+	};
+}
+
+// Base inspection function for prefilters and transports
+function inspectPrefiltersOrTransports( structure, options, originalOptions, jqXHR ) {
+
+	var inspected = {},
+		seekingTransport = ( structure === transports );
+
+	function inspect( dataType ) {
+		var selected;
+		inspected[ dataType ] = true;
+		jQuery.each( structure[ dataType ] || [], function( _, prefilterOrFactory ) {
+			var dataTypeOrTransport = prefilterOrFactory( options, originalOptions, jqXHR );
+			if ( typeof dataTypeOrTransport === "string" &&
+				!seekingTransport && !inspected[ dataTypeOrTransport ] ) {
+
+				options.dataTypes.unshift( dataTypeOrTransport );
+				inspect( dataTypeOrTransport );
+				return false;
+			} else if ( seekingTransport ) {
+				return !( selected = dataTypeOrTransport );
+			}
+		} );
+		return selected;
+	}
+
+	return inspect( options.dataTypes[ 0 ] ) || !inspected[ "*" ] && inspect( "*" );
+}
+
+// A special extend for ajax options
+// that takes "flat" options (not to be deep extended)
+// Fixes #9887
+function ajaxExtend( target, src ) {
+	var key, deep,
+		flatOptions = jQuery.ajaxSettings.flatOptions || {};
+
+	for ( key in src ) {
+		if ( src[ key ] !== undefined ) {
+			( flatOptions[ key ] ? target : ( deep || ( deep = {} ) ) )[ key ] = src[ key ];
+		}
+	}
+	if ( deep ) {
+		jQuery.extend( true, target, deep );
+	}
+
+	return target;
+}
+
+/* Handles responses to an ajax request:
+ * - finds the right dataType (mediates between content-type and expected dataType)
+ * - returns the corresponding response
+ */
+function ajaxHandleResponses( s, jqXHR, responses ) {
+
+	var ct, type, finalDataType, firstDataType,
+		contents = s.contents,
+		dataTypes = s.dataTypes;
+
+	// Remove auto dataType and get content-type in the process
+	while ( dataTypes[ 0 ] === "*" ) {
+		dataTypes.shift();
+		if ( ct === undefined ) {
+			ct = s.mimeType || jqXHR.getResponseHeader( "Content-Type" );
+		}
+	}
+
+	// Check if we're dealing with a known content-type
+	if ( ct ) {
+		for ( type in contents ) {
+			if ( contents[ type ] && contents[ type ].test( ct ) ) {
+				dataTypes.unshift( type );
+				break;
+			}
+		}
+	}
+
+	// Check to see if we have a response for the expected dataType
+	if ( dataTypes[ 0 ] in responses ) {
+		finalDataType = dataTypes[ 0 ];
+	} else {
+
+		// Try convertible dataTypes
+		for ( type in responses ) {
+			if ( !dataTypes[ 0 ] || s.converters[ type + " " + dataTypes[ 0 ] ] ) {
+				finalDataType = type;
+				break;
+			}
+			if ( !firstDataType ) {
+				firstDataType = type;
+			}
+		}
+
+		// Or just use first one
+		finalDataType = finalDataType || firstDataType;
+	}
+
+	// If we found a dataType
+	// We add the dataType to the list if needed
+	// and return the corresponding response
+	if ( finalDataType ) {
+		if ( finalDataType !== dataTypes[ 0 ] ) {
+			dataTypes.unshift( finalDataType );
+		}
+		return responses[ finalDataType ];
+	}
+}
+
+/* Chain conversions given the request and the original response
+ * Also sets the responseXXX fields on the jqXHR instance
+ */
+function ajaxConvert( s, response, jqXHR, isSuccess ) {
+	var conv2, current, conv, tmp, prev,
+		converters = {},
+
+		// Work with a copy of dataTypes in case we need to modify it for conversion
+		dataTypes = s.dataTypes.slice();
+
+	// Create converters map with lowercased keys
+	if ( dataTypes[ 1 ] ) {
+		for ( conv in s.converters ) {
+			converters[ conv.toLowerCase() ] = s.converters[ conv ];
+		}
+	}
+
+	current = dataTypes.shift();
+
+	// Convert to each sequential dataType
+	while ( current ) {
+
+		if ( s.responseFields[ current ] ) {
+			jqXHR[ s.responseFields[ current ] ] = response;
+		}
+
+		// Apply the dataFilter if provided
+		if ( !prev && isSuccess && s.dataFilter ) {
+			response = s.dataFilter( response, s.dataType );
+		}
+
+		prev = current;
+		current = dataTypes.shift();
+
+		if ( current ) {
+
+			// There's only work to do if current dataType is non-auto
+			if ( current === "*" ) {
+
+				current = prev;
+
+			// Convert response if prev dataType is non-auto and differs from current
+			} else if ( prev !== "*" && prev !== current ) {
+
+				// Seek a direct converter
+				conv = converters[ prev + " " + current ] || converters[ "* " + current ];
+
+				// If none found, seek a pair
+				if ( !conv ) {
+					for ( conv2 in converters ) {
+
+						// If conv2 outputs current
+						tmp = conv2.split( " " );
+						if ( tmp[ 1 ] === current ) {
+
+							// If prev can be converted to accepted input
+							conv = converters[ prev + " " + tmp[ 0 ] ] ||
+								converters[ "* " + tmp[ 0 ] ];
+							if ( conv ) {
+
+								// Condense equivalence converters
+								if ( conv === true ) {
+									conv = converters[ conv2 ];
+
+								// Otherwise, insert the intermediate dataType
+								} else if ( converters[ conv2 ] !== true ) {
+									current = tmp[ 0 ];
+									dataTypes.unshift( tmp[ 1 ] );
+								}
+								break;
+							}
+						}
+					}
+				}
+
+				// Apply converter (if not an equivalence)
+				if ( conv !== true ) {
+
+					// Unless errors are allowed to bubble, catch and return them
+					if ( conv && s.throws ) {
+						response = conv( response );
+					} else {
+						try {
+							response = conv( response );
+						} catch ( e ) {
+							return {
+								state: "parsererror",
+								error: conv ? e : "No conversion from " + prev + " to " + current
+							};
+						}
+					}
+				}
+			}
+		}
+	}
+
+	return { state: "success", data: response };
+}
+
+jQuery.extend( {
+
+	// Counter for holding the number of active queries
+	active: 0,
+
+	// Last-Modified header cache for next request
+	lastModified: {},
+	etag: {},
+
+	ajaxSettings: {
+		url: location.href,
+		type: "GET",
+		isLocal: rlocalProtocol.test( location.protocol ),
+		global: true,
+		processData: true,
+		async: true,
+		contentType: "application/x-www-form-urlencoded; charset=UTF-8",
+
+		/*
+		timeout: 0,
+		data: null,
+		dataType: null,
+		username: null,
+		password: null,
+		cache: null,
+		throws: false,
+		traditional: false,
+		headers: {},
+		*/
+
+		accepts: {
+			"*": allTypes,
+			text: "text/plain",
+			html: "text/html",
+			xml: "application/xml, text/xml",
+			json: "application/json, text/javascript"
+		},
+
+		contents: {
+			xml: /\bxml\b/,
+			html: /\bhtml/,
+			json: /\bjson\b/
+		},
+
+		responseFields: {
+			xml: "responseXML",
+			text: "responseText",
+			json: "responseJSON"
+		},
+
+		// Data converters
+		// Keys separate source (or catchall "*") and destination types with a single space
+		converters: {
+
+			// Convert anything to text
+			"* text": String,
+
+			// Text to html (true = no transformation)
+			"text html": true,
+
+			// Evaluate text as a json expression
+			"text json": JSON.parse,
+
+			// Parse text as xml
+			"text xml": jQuery.parseXML
+		},
+
+		// For options that shouldn't be deep extended:
+		// you can add your own custom options here if
+		// and when you create one that shouldn't be
+		// deep extended (see ajaxExtend)
+		flatOptions: {
+			url: true,
+			context: true
+		}
+	},
+
+	// Creates a full fledged settings object into target
+	// with both ajaxSettings and settings fields.
+	// If target is omitted, writes into ajaxSettings.
+	ajaxSetup: function( target, settings ) {
+		return settings ?
+
+			// Building a settings object
+			ajaxExtend( ajaxExtend( target, jQuery.ajaxSettings ), settings ) :
+
+			// Extending ajaxSettings
+			ajaxExtend( jQuery.ajaxSettings, target );
+	},
+
+	ajaxPrefilter: addToPrefiltersOrTransports( prefilters ),
+	ajaxTransport: addToPrefiltersOrTransports( transports ),
+
+	// Main method
+	ajax: function( url, options ) {
+
+		// If url is an object, simulate pre-1.5 signature
+		if ( typeof url === "object" ) {
+			options = url;
+			url = undefined;
+		}
+
+		// Force options to be an object
+		options = options || {};
+
+		var transport,
+
+			// URL without anti-cache param
+			cacheURL,
+
+			// Response headers
+			responseHeadersString,
+			responseHeaders,
+
+			// timeout handle
+			timeoutTimer,
+
+			// Url cleanup var
+			urlAnchor,
+
+			// Request state (becomes false upon send and true upon completion)
+			completed,
+
+			// To know if global events are to be dispatched
+			fireGlobals,
+
+			// Loop variable
+			i,
+
+			// uncached part of the url
+			uncached,
+
+			// Create the final options object
+			s = jQuery.ajaxSetup( {}, options ),
+
+			// Callbacks context
+			callbackContext = s.context || s,
+
+			// Context for global events is callbackContext if it is a DOM node or jQuery collection
+			globalEventContext = s.context &&
+				( callbackContext.nodeType || callbackContext.jquery ) ?
+				jQuery( callbackContext ) :
+				jQuery.event,
+
+			// Deferreds
+			deferred = jQuery.Deferred(),
+			completeDeferred = jQuery.Callbacks( "once memory" ),
+
+			// Status-dependent callbacks
+			statusCode = s.statusCode || {},
+
+			// Headers (they are sent all at once)
+			requestHeaders = {},
+			requestHeadersNames = {},
+
+			// Default abort message
+			strAbort = "canceled",
+
+			// Fake xhr
+			jqXHR = {
+				readyState: 0,
+
+				// Builds headers hashtable if needed
+				getResponseHeader: function( key ) {
+					var match;
+					if ( completed ) {
+						if ( !responseHeaders ) {
+							responseHeaders = {};
+							while ( ( match = rheaders.exec( responseHeadersString ) ) ) {
+								responseHeaders[ match[ 1 ].toLowerCase() + " " ] =
+									( responseHeaders[ match[ 1 ].toLowerCase() + " " ] || [] )
+										.concat( match[ 2 ] );
+							}
+						}
+						match = responseHeaders[ key.toLowerCase() + " " ];
+					}
+					return match == null ? null : match.join( ", " );
+				},
+
+				// Raw string
+				getAllResponseHeaders: function() {
+					return completed ? responseHeadersString : null;
+				},
+
+				// Caches the header
+				setRequestHeader: function( name, value ) {
+					if ( completed == null ) {
+						name = requestHeadersNames[ name.toLowerCase() ] =
+							requestHeadersNames[ name.toLowerCase() ] || name;
+						requestHeaders[ name ] = value;
+					}
+					return this;
+				},
+
+				// Overrides response content-type header
+				overrideMimeType: function( type ) {
+					if ( completed == null ) {
+						s.mimeType = type;
+					}
+					return this;
+				},
+
+				// Status-dependent callbacks
+				statusCode: function( map ) {
+					var code;
+					if ( map ) {
+						if ( completed ) {
+
+							// Execute the appropriate callbacks
+							jqXHR.always( map[ jqXHR.status ] );
+						} else {
+
+							// Lazy-add the new callbacks in a way that preserves old ones
+							for ( code in map ) {
+								statusCode[ code ] = [ statusCode[ code ], map[ code ] ];
+							}
+						}
+					}
+					return this;
+				},
+
+				// Cancel the request
+				abort: function( statusText ) {
+					var finalText = statusText || strAbort;
+					if ( transport ) {
+						transport.abort( finalText );
+					}
+					done( 0, finalText );
+					return this;
+				}
+			};
+
+		// Attach deferreds
+		deferred.promise( jqXHR );
+
+		// Add protocol if not provided (prefilters might expect it)
+		// Handle falsy url in the settings object (#10093: consistency with old signature)
+		// We also use the url parameter if available
+		s.url = ( ( url || s.url || location.href ) + "" )
+			.replace( rprotocol, location.protocol + "//" );
+
+		// Alias method option to type as per ticket #12004
+		s.type = options.method || options.type || s.method || s.type;
+
+		// Extract dataTypes list
+		s.dataTypes = ( s.dataType || "*" ).toLowerCase().match( rnothtmlwhite ) || [ "" ];
+
+		// A cross-domain request is in order when the origin doesn't match the current origin.
+		if ( s.crossDomain == null ) {
+			urlAnchor = document.createElement( "a" );
+
+			// Support: IE <=8 - 11, Edge 12 - 15
+			// IE throws exception on accessing the href property if url is malformed,
+			// e.g. http://example.com:80x/
+			try {
+				urlAnchor.href = s.url;
+
+				// Support: IE <=8 - 11 only
+				// Anchor's host property isn't correctly set when s.url is relative
+				urlAnchor.href = urlAnchor.href;
+				s.crossDomain = originAnchor.protocol + "//" + originAnchor.host !==
+					urlAnchor.protocol + "//" + urlAnchor.host;
+			} catch ( e ) {
+
+				// If there is an error parsing the URL, assume it is crossDomain,
+				// it can be rejected by the transport if it is invalid
+				s.crossDomain = true;
+			}
+		}
+
+		// Convert data if not already a string
+		if ( s.data && s.processData && typeof s.data !== "string" ) {
+			s.data = jQuery.param( s.data, s.traditional );
+		}
+
+		// Apply prefilters
+		inspectPrefiltersOrTransports( prefilters, s, options, jqXHR );
+
+		// If request was aborted inside a prefilter, stop there
+		if ( completed ) {
+			return jqXHR;
+		}
+
+		// We can fire global events as of now if asked to
+		// Don't fire events if jQuery.event is undefined in an AMD-usage scenario (#15118)
+		fireGlobals = jQuery.event && s.global;
+
+		// Watch for a new set of requests
+		if ( fireGlobals && jQuery.active++ === 0 ) {
+			jQuery.event.trigger( "ajaxStart" );
+		}
+
+		// Uppercase the type
+		s.type = s.type.toUpperCase();
+
+		// Determine if request has content
+		s.hasContent = !rnoContent.test( s.type );
+
+		// Save the URL in case we're toying with the If-Modified-Since
+		// and/or If-None-Match header later on
+		// Remove hash to simplify url manipulation
+		cacheURL = s.url.replace( rhash, "" );
+
+		// More options handling for requests with no content
+		if ( !s.hasContent ) {
+
+			// Remember the hash so we can put it back
+			uncached = s.url.slice( cacheURL.length );
+
+			// If data is available and should be processed, append data to url
+			if ( s.data && ( s.processData || typeof s.data === "string" ) ) {
+				cacheURL += ( rquery.test( cacheURL ) ? "&" : "?" ) + s.data;
+
+				// #9682: remove data so that it's not used in an eventual retry
+				delete s.data;
+			}
+
+			// Add or update anti-cache param if needed
+			if ( s.cache === false ) {
+				cacheURL = cacheURL.replace( rantiCache, "$1" );
+				uncached = ( rquery.test( cacheURL ) ? "&" : "?" ) + "_=" + ( nonce.guid++ ) +
+					uncached;
+			}
+
+			// Put hash and anti-cache on the URL that will be requested (gh-1732)
+			s.url = cacheURL + uncached;
+
+		// Change '%20' to '+' if this is encoded form body content (gh-2658)
+		} else if ( s.data && s.processData &&
+			( s.contentType || "" ).indexOf( "application/x-www-form-urlencoded" ) === 0 ) {
+			s.data = s.data.replace( r20, "+" );
+		}
+
+		// Set the If-Modified-Since and/or If-None-Match header, if in ifModified mode.
+		if ( s.ifModified ) {
+			if ( jQuery.lastModified[ cacheURL ] ) {
+				jqXHR.setRequestHeader( "If-Modified-Since", jQuery.lastModified[ cacheURL ] );
+			}
+			if ( jQuery.etag[ cacheURL ] ) {
+				jqXHR.setRequestHeader( "If-None-Match", jQuery.etag[ cacheURL ] );
+			}
+		}
+
+		// Set the correct header, if data is being sent
+		if ( s.data && s.hasContent && s.contentType !== false || options.contentType ) {
+			jqXHR.setRequestHeader( "Content-Type", s.contentType );
+		}
+
+		// Set the Accepts header for the server, depending on the dataType
+		jqXHR.setRequestHeader(
+			"Accept",
+			s.dataTypes[ 0 ] && s.accepts[ s.dataTypes[ 0 ] ] ?
+				s.accepts[ s.dataTypes[ 0 ] ] +
+					( s.dataTypes[ 0 ] !== "*" ? ", " + allTypes + "; q=0.01" : "" ) :
+				s.accepts[ "*" ]
+		);
+
+		// Check for headers option
+		for ( i in s.headers ) {
+			jqXHR.setRequestHeader( i, s.headers[ i ] );
+		}
+
+		// Allow custom headers/mimetypes and early abort
+		if ( s.beforeSend &&
+			( s.beforeSend.call( callbackContext, jqXHR, s ) === false || completed ) ) {
+
+			// Abort if not done already and return
+			return jqXHR.abort();
+		}
+
+		// Aborting is no longer a cancellation
+		strAbort = "abort";
+
+		// Install callbacks on deferreds
+		completeDeferred.add( s.complete );
+		jqXHR.done( s.success );
+		jqXHR.fail( s.error );
+
+		// Get transport
+		transport = inspectPrefiltersOrTransports( transports, s, options, jqXHR );
+
+		// If no transport, we auto-abort
+		if ( !transport ) {
+			done( -1, "No Transport" );
+		} else {
+			jqXHR.readyState = 1;
+
+			// Send global event
+			if ( fireGlobals ) {
+				globalEventContext.trigger( "ajaxSend", [ jqXHR, s ] );
+			}
+
+			// If request was aborted inside ajaxSend, stop there
+			if ( completed ) {
+				return jqXHR;
+			}
+
+			// Timeout
+			if ( s.async && s.timeout > 0 ) {
+				timeoutTimer = window.setTimeout( function() {
+					jqXHR.abort( "timeout" );
+				}, s.timeout );
+			}
+
+			try {
+				completed = false;
+				transport.send( requestHeaders, done );
+			} catch ( e ) {
+
+				// Rethrow post-completion exceptions
+				if ( completed ) {
+					throw e;
+				}
+
+				// Propagate others as results
+				done( -1, e );
+			}
+		}
+
+		// Callback for when everything is done
+		function done( status, nativeStatusText, responses, headers ) {
+			var isSuccess, success, error, response, modified,
+				statusText = nativeStatusText;
+
+			// Ignore repeat invocations
+			if ( completed ) {
+				return;
+			}
+
+			completed = true;
+
+			// Clear timeout if it exists
+			if ( timeoutTimer ) {
+				window.clearTimeout( timeoutTimer );
+			}
+
+			// Dereference transport for early garbage collection
+			// (no matter how long the jqXHR object will be used)
+			transport = undefined;
+
+			// Cache response headers
+			responseHeadersString = headers || "";
+
+			// Set readyState
+			jqXHR.readyState = status > 0 ? 4 : 0;
+
+			// Determine if successful
+			isSuccess = status >= 200 && status < 300 || status === 304;
+
+			// Get response data
+			if ( responses ) {
+				response = ajaxHandleResponses( s, jqXHR, responses );
+			}
+
+			// Use a noop converter for missing script but not if jsonp
+			if ( !isSuccess &&
+				jQuery.inArray( "script", s.dataTypes ) > -1 &&
+				jQuery.inArray( "json", s.dataTypes ) < 0 ) {
+				s.converters[ "text script" ] = function() {};
+			}
+
+			// Convert no matter what (that way responseXXX fields are always set)
+			response = ajaxConvert( s, response, jqXHR, isSuccess );
+
+			// If successful, handle type chaining
+			if ( isSuccess ) {
+
+				// Set the If-Modified-Since and/or If-None-Match header, if in ifModified mode.
+				if ( s.ifModified ) {
+					modified = jqXHR.getResponseHeader( "Last-Modified" );
+					if ( modified ) {
+						jQuery.lastModified[ cacheURL ] = modified;
+					}
+					modified = jqXHR.getResponseHeader( "etag" );
+					if ( modified ) {
+						jQuery.etag[ cacheURL ] = modified;
+					}
+				}
+
+				// if no content
+				if ( status === 204 || s.type === "HEAD" ) {
+					statusText = "nocontent";
+
+				// if not modified
+				} else if ( status === 304 ) {
+					statusText = "notmodified";
+
+				// If we have data, let's convert it
+				} else {
+					statusText = response.state;
+					success = response.data;
+					error = response.error;
+					isSuccess = !error;
+				}
+			} else {
+
+				// Extract error from statusText and normalize for non-aborts
+				error = statusText;
+				if ( status || !statusText ) {
+					statusText = "error";
+					if ( status < 0 ) {
+						status = 0;
+					}
+				}
+			}
+
+			// Set data for the fake xhr object
+			jqXHR.status = status;
+			jqXHR.statusText = ( nativeStatusText || statusText ) + "";
+
+			// Success/Error
+			if ( isSuccess ) {
+				deferred.resolveWith( callbackContext, [ success, statusText, jqXHR ] );
+			} else {
+				deferred.rejectWith( callbackContext, [ jqXHR, statusText, error ] );
+			}
+
+			// Status-dependent callbacks
+			jqXHR.statusCode( statusCode );
+			statusCode = undefined;
+
+			if ( fireGlobals ) {
+				globalEventContext.trigger( isSuccess ? "ajaxSuccess" : "ajaxError",
+					[ jqXHR, s, isSuccess ? success : error ] );
+			}
+
+			// Complete
+			completeDeferred.fireWith( callbackContext, [ jqXHR, statusText ] );
+
+			if ( fireGlobals ) {
+				globalEventContext.trigger( "ajaxComplete", [ jqXHR, s ] );
+
+				// Handle the global AJAX counter
+				if ( !( --jQuery.active ) ) {
+					jQuery.event.trigger( "ajaxStop" );
+				}
+			}
+		}
+
+		return jqXHR;
+	},
+
+	getJSON: function( url, data, callback ) {
+		return jQuery.get( url, data, callback, "json" );
+	},
+
+	getScript: function( url, callback ) {
+		return jQuery.get( url, undefined, callback, "script" );
+	}
+} );
+
+jQuery.each( [ "get", "post" ], function( _i, method ) {
+	jQuery[ method ] = function( url, data, callback, type ) {
+
+		// Shift arguments if data argument was omitted
+		if ( isFunction( data ) ) {
+			type = type || callback;
+			callback = data;
+			data = undefined;
+		}
+
+		// The url can be an options object (which then must have .url)
+		return jQuery.ajax( jQuery.extend( {
+			url: url,
+			type: method,
+			dataType: type,
+			data: data,
+			success: callback
+		}, jQuery.isPlainObject( url ) && url ) );
+	};
+} );
+
+jQuery.ajaxPrefilter( function( s ) {
+	var i;
+	for ( i in s.headers ) {
+		if ( i.toLowerCase() === "content-type" ) {
+			s.contentType = s.headers[ i ] || "";
+		}
+	}
+} );
+
+
+jQuery._evalUrl = function( url, options, doc ) {
+	return jQuery.ajax( {
+		url: url,
+
+		// Make this explicit, since user can override this through ajaxSetup (#11264)
+		type: "GET",
+		dataType: "script",
+		cache: true,
+		async: false,
+		global: false,
+
+		// Only evaluate the response if it is successful (gh-4126)
+		// dataFilter is not invoked for failure responses, so using it instead
+		// of the default converter is kludgy but it works.
+		converters: {
+			"text script": function() {}
+		},
+		dataFilter: function( response ) {
+			jQuery.globalEval( response, options, doc );
+		}
+	} );
+};
+
+
+jQuery.fn.extend( {
+	wrapAll: function( html ) {
+		var wrap;
+
+		if ( this[ 0 ] ) {
+			if ( isFunction( html ) ) {
+				html = html.call( this[ 0 ] );
+			}
+
+			// The elements to wrap the target around
+			wrap = jQuery( html, this[ 0 ].ownerDocument ).eq( 0 ).clone( true );
+
+			if ( this[ 0 ].parentNode ) {
+				wrap.insertBefore( this[ 0 ] );
+			}
+
+			wrap.map( function() {
+				var elem = this;
+
+				while ( elem.firstElementChild ) {
+					elem = elem.firstElementChild;
+				}
+
+				return elem;
+			} ).append( this );
+		}
+
+		return this;
+	},
+
+	wrapInner: function( html ) {
+		if ( isFunction( html ) ) {
+			return this.each( function( i ) {
+				jQuery( this ).wrapInner( html.call( this, i ) );
+			} );
+		}
+
+		return this.each( function() {
+			var self = jQuery( this ),
+				contents = self.contents();
+
+			if ( contents.length ) {
+				contents.wrapAll( html );
+
+			} else {
+				self.append( html );
+			}
+		} );
+	},
+
+	wrap: function( html ) {
+		var htmlIsFunction = isFunction( html );
+
+		return this.each( function( i ) {
+			jQuery( this ).wrapAll( htmlIsFunction ? html.call( this, i ) : html );
+		} );
+	},
+
+	unwrap: function( selector ) {
+		this.parent( selector ).not( "body" ).each( function() {
+			jQuery( this ).replaceWith( this.childNodes );
+		} );
+		return this;
+	}
+} );
+
+
+jQuery.expr.pseudos.hidden = function( elem ) {
+	return !jQuery.expr.pseudos.visible( elem );
+};
+jQuery.expr.pseudos.visible = function( elem ) {
+	return !!( elem.offsetWidth || elem.offsetHeight || elem.getClientRects().length );
+};
+
+
+
+
+jQuery.ajaxSettings.xhr = function() {
+	try {
+		return new window.XMLHttpRequest();
+	} catch ( e ) {}
+};
+
+var xhrSuccessStatus = {
+
+		// File protocol always yields status code 0, assume 200
+		0: 200,
+
+		// Support: IE <=9 only
+		// #1450: sometimes IE returns 1223 when it should be 204
+		1223: 204
+	},
+	xhrSupported = jQuery.ajaxSettings.xhr();
+
+support.cors = !!xhrSupported && ( "withCredentials" in xhrSupported );
+support.ajax = xhrSupported = !!xhrSupported;
+
+jQuery.ajaxTransport( function( options ) {
+	var callback, errorCallback;
+
+	// Cross domain only allowed if supported through XMLHttpRequest
+	if ( support.cors || xhrSupported && !options.crossDomain ) {
+		return {
+			send: function( headers, complete ) {
+				var i,
+					xhr = options.xhr();
+
+				xhr.open(
+					options.type,
+					options.url,
+					options.async,
+					options.username,
+					options.password
+				);
+
+				// Apply custom fields if provided
+				if ( options.xhrFields ) {
+					for ( i in options.xhrFields ) {
+						xhr[ i ] = options.xhrFields[ i ];
+					}
+				}
+
+				// Override mime type if needed
+				if ( options.mimeType && xhr.overrideMimeType ) {
+					xhr.overrideMimeType( options.mimeType );
+				}
+
+				// X-Requested-With header
+				// For cross-domain requests, seeing as conditions for a preflight are
+				// akin to a jigsaw puzzle, we simply never set it to be sure.
+				// (it can always be set on a per-request basis or even using ajaxSetup)
+				// For same-domain requests, won't change header if already provided.
+				if ( !options.crossDomain && !headers[ "X-Requested-With" ] ) {
+					headers[ "X-Requested-With" ] = "XMLHttpRequest";
+				}
+
+				// Set headers
+				for ( i in headers ) {
+					xhr.setRequestHeader( i, headers[ i ] );
+				}
+
+				// Callback
+				callback = function( type ) {
+					return function() {
+						if ( callback ) {
+							callback = errorCallback = xhr.onload =
+								xhr.onerror = xhr.onabort = xhr.ontimeout =
+									xhr.onreadystatechange = null;
+
+							if ( type === "abort" ) {
+								xhr.abort();
+							} else if ( type === "error" ) {
+
+								// Support: IE <=9 only
+								// On a manual native abort, IE9 throws
+								// errors on any property access that is not readyState
+								if ( typeof xhr.status !== "number" ) {
+									complete( 0, "error" );
+								} else {
+									complete(
+
+										// File: protocol always yields status 0; see #8605, #14207
+										xhr.status,
+										xhr.statusText
+									);
+								}
+							} else {
+								complete(
+									xhrSuccessStatus[ xhr.status ] || xhr.status,
+									xhr.statusText,
+
+									// Support: IE <=9 only
+									// IE9 has no XHR2 but throws on binary (trac-11426)
+									// For XHR2 non-text, let the caller handle it (gh-2498)
+									( xhr.responseType || "text" ) !== "text"  ||
+									typeof xhr.responseText !== "string" ?
+										{ binary: xhr.response } :
+										{ text: xhr.responseText },
+									xhr.getAllResponseHeaders()
+								);
+							}
+						}
+					};
+				};
+
+				// Listen to events
+				xhr.onload = callback();
+				errorCallback = xhr.onerror = xhr.ontimeout = callback( "error" );
+
+				// Support: IE 9 only
+				// Use onreadystatechange to replace onabort
+				// to handle uncaught aborts
+				if ( xhr.onabort !== undefined ) {
+					xhr.onabort = errorCallback;
+				} else {
+					xhr.onreadystatechange = function() {
+
+						// Check readyState before timeout as it changes
+						if ( xhr.readyState === 4 ) {
+
+							// Allow onerror to be called first,
+							// but that will not handle a native abort
+							// Also, save errorCallback to a variable
+							// as xhr.onerror cannot be accessed
+							window.setTimeout( function() {
+								if ( callback ) {
+									errorCallback();
+								}
+							} );
+						}
+					};
+				}
+
+				// Create the abort callback
+				callback = callback( "abort" );
+
+				try {
+
+					// Do send the request (this may raise an exception)
+					xhr.send( options.hasContent && options.data || null );
+				} catch ( e ) {
+
+					// #14683: Only rethrow if this hasn't been notified as an error yet
+					if ( callback ) {
+						throw e;
+					}
+				}
+			},
+
+			abort: function() {
+				if ( callback ) {
+					callback();
+				}
+			}
+		};
+	}
+} );
+
+
+
+
+// Prevent auto-execution of scripts when no explicit dataType was provided (See gh-2432)
+jQuery.ajaxPrefilter( function( s ) {
+	if ( s.crossDomain ) {
+		s.contents.script = false;
+	}
+} );
+
+// Install script dataType
+jQuery.ajaxSetup( {
+	accepts: {
+		script: "text/javascript, application/javascript, " +
+			"application/ecmascript, application/x-ecmascript"
+	},
+	contents: {
+		script: /\b(?:java|ecma)script\b/
+	},
+	converters: {
+		"text script": function( text ) {
+			jQuery.globalEval( text );
+			return text;
+		}
+	}
+} );
+
+// Handle cache's special case and crossDomain
+jQuery.ajaxPrefilter( "script", function( s ) {
+	if ( s.cache === undefined ) {
+		s.cache = false;
+	}
+	if ( s.crossDomain ) {
+		s.type = "GET";
+	}
+} );
+
+// Bind script tag hack transport
+jQuery.ajaxTransport( "script", function( s ) {
+
+	// This transport only deals with cross domain or forced-by-attrs requests
+	if ( s.crossDomain || s.scriptAttrs ) {
+		var script, callback;
+		return {
+			send: function( _, complete ) {
+				script = jQuery( "<script>" )
+					.attr( s.scriptAttrs || {} )
+					.prop( { charset: s.scriptCharset, src: s.url } )
+					.on( "load error", callback = function( evt ) {
+						script.remove();
+						callback = null;
+						if ( evt ) {
+							complete( evt.type === "error" ? 404 : 200, evt.type );
+						}
+					} );
+
+				// Use native DOM manipulation to avoid our domManip AJAX trickery
+				document.head.appendChild( script[ 0 ] );
+			},
+			abort: function() {
+				if ( callback ) {
+					callback();
+				}
+			}
+		};
+	}
+} );
+
+
+
+
+var oldCallbacks = [],
+	rjsonp = /(=)\?(?=&|$)|\?\?/;
+
+// Default jsonp settings
+jQuery.ajaxSetup( {
+	jsonp: "callback",
+	jsonpCallback: function() {
+		var callback = oldCallbacks.pop() || ( jQuery.expando + "_" + ( nonce.guid++ ) );
+		this[ callback ] = true;
+		return callback;
+	}
+} );
+
+// Detect, normalize options and install callbacks for jsonp requests
+jQuery.ajaxPrefilter( "json jsonp", function( s, originalSettings, jqXHR ) {
+
+	var callbackName, overwritten, responseContainer,
+		jsonProp = s.jsonp !== false && ( rjsonp.test( s.url ) ?
+			"url" :
+			typeof s.data === "string" &&
+				( s.contentType || "" )
+					.indexOf( "application/x-www-form-urlencoded" ) === 0 &&
+				rjsonp.test( s.data ) && "data"
+		);
+
+	// Handle iff the expected data type is "jsonp" or we have a parameter to set
+	if ( jsonProp || s.dataTypes[ 0 ] === "jsonp" ) {
+
+		// Get callback name, remembering preexisting value associated with it
+		callbackName = s.jsonpCallback = isFunction( s.jsonpCallback ) ?
+			s.jsonpCallback() :
+			s.jsonpCallback;
+
+		// Insert callback into url or form data
+		if ( jsonProp ) {
+			s[ jsonProp ] = s[ jsonProp ].replace( rjsonp, "$1" + callbackName );
+		} else if ( s.jsonp !== false ) {
+			s.url += ( rquery.test( s.url ) ? "&" : "?" ) + s.jsonp + "=" + callbackName;
+		}
+
+		// Use data converter to retrieve json after script execution
+		s.converters[ "script json" ] = function() {
+			if ( !responseContainer ) {
+				jQuery.error( callbackName + " was not called" );
+			}
+			return responseContainer[ 0 ];
+		};
+
+		// Force json dataType
+		s.dataTypes[ 0 ] = "json";
+
+		// Install callback
+		overwritten = window[ callbackName ];
+		window[ callbackName ] = function() {
+			responseContainer = arguments;
+		};
+
+		// Clean-up function (fires after converters)
+		jqXHR.always( function() {
+
+			// If previous value didn't exist - remove it
+			if ( overwritten === undefined ) {
+				jQuery( window ).removeProp( callbackName );
+
+			// Otherwise restore preexisting value
+			} else {
+				window[ callbackName ] = overwritten;
+			}
+
+			// Save back as free
+			if ( s[ callbackName ] ) {
+
+				// Make sure that re-using the options doesn't screw things around
+				s.jsonpCallback = originalSettings.jsonpCallback;
+
+				// Save the callback name for future use
+				oldCallbacks.push( callbackName );
+			}
+
+			// Call if it was a function and we have a response
+			if ( responseContainer && isFunction( overwritten ) ) {
+				overwritten( responseContainer[ 0 ] );
+			}
+
+			responseContainer = overwritten = undefined;
+		} );
+
+		// Delegate to script
+		return "script";
+	}
+} );
+
+
+
+
+// Support: Safari 8 only
+// In Safari 8 documents created via document.implementation.createHTMLDocument
+// collapse sibling forms: the second one becomes a child of the first one.
+// Because of that, this security measure has to be disabled in Safari 8.
+// https://bugs.webkit.org/show_bug.cgi?id=137337
+support.createHTMLDocument = ( function() {
+	var body = document.implementation.createHTMLDocument( "" ).body;
+	body.innerHTML = "<form></form><form></form>";
+	return body.childNodes.length === 2;
+} )();
+
+
+// Argument "data" should be string of html
+// context (optional): If specified, the fragment will be created in this context,
+// defaults to document
+// keepScripts (optional): If true, will include scripts passed in the html string
+jQuery.parseHTML = function( data, context, keepScripts ) {
+	if ( typeof data !== "string" ) {
+		return [];
+	}
+	if ( typeof context === "boolean" ) {
+		keepScripts = context;
+		context = false;
+	}
+
+	var base, parsed, scripts;
+
+	if ( !context ) {
+
+		// Stop scripts or inline event handlers from being executed immediately
+		// by using document.implementation
+		if ( support.createHTMLDocument ) {
+			context = document.implementation.createHTMLDocument( "" );
+
+			// Set the base href for the created document
+			// so any parsed elements with URLs
+			// are based on the document's URL (gh-2965)
+			base = context.createElement( "base" );
+			base.href = document.location.href;
+			context.head.appendChild( base );
+		} else {
+			context = document;
+		}
+	}
+
+	parsed = rsingleTag.exec( data );
+	scripts = !keepScripts && [];
+
+	// Single tag
+	if ( parsed ) {
+		return [ context.createElement( parsed[ 1 ] ) ];
+	}
+
+	parsed = buildFragment( [ data ], context, scripts );
+
+	if ( scripts && scripts.length ) {
+		jQuery( scripts ).remove();
+	}
+
+	return jQuery.merge( [], parsed.childNodes );
+};
+
+
+/**
+ * Load a url into a page
+ */
+jQuery.fn.load = function( url, params, callback ) {
+	var selector, type, response,
+		self = this,
+		off = url.indexOf( " " );
+
+	if ( off > -1 ) {
+		selector = stripAndCollapse( url.slice( off ) );
+		url = url.slice( 0, off );
+	}
+
+	// If it's a function
+	if ( isFunction( params ) ) {
+
+		// We assume that it's the callback
+		callback = params;
+		params = undefined;
+
+	// Otherwise, build a param string
+	} else if ( params && typeof params === "object" ) {
+		type = "POST";
+	}
+
+	// If we have elements to modify, make the request
+	if ( self.length > 0 ) {
+		jQuery.ajax( {
+			url: url,
+
+			// If "type" variable is undefined, then "GET" method will be used.
+			// Make value of this field explicit since
+			// user can override it through ajaxSetup method
+			type: type || "GET",
+			dataType: "html",
+			data: params
+		} ).done( function( responseText ) {
+
+			// Save response for use in complete callback
+			response = arguments;
+
+			self.html( selector ?
+
+				// If a selector was specified, locate the right elements in a dummy div
+				// Exclude scripts to avoid IE 'Permission Denied' errors
+				jQuery( "<div>" ).append( jQuery.parseHTML( responseText ) ).find( selector ) :
+
+				// Otherwise use the full result
+				responseText );
+
+		// If the request succeeds, this function gets "data", "status", "jqXHR"
+		// but they are ignored because response was set above.
+		// If it fails, this function gets "jqXHR", "status", "error"
+		} ).always( callback && function( jqXHR, status ) {
+			self.each( function() {
+				callback.apply( this, response || [ jqXHR.responseText, status, jqXHR ] );
+			} );
+		} );
+	}
+
+	return this;
+};
+
+
+
+
+jQuery.expr.pseudos.animated = function( elem ) {
+	return jQuery.grep( jQuery.timers, function( fn ) {
+		return elem === fn.elem;
+	} ).length;
+};
+
+
+
+
+jQuery.offset = {
+	setOffset: function( elem, options, i ) {
+		var curPosition, curLeft, curCSSTop, curTop, curOffset, curCSSLeft, calculatePosition,
+			position = jQuery.css( elem, "position" ),
+			curElem = jQuery( elem ),
+			props = {};
+
+		// Set position first, in-case top/left are set even on static elem
+		if ( position === "static" ) {
+			elem.style.position = "relative";
+		}
+
+		curOffset = curElem.offset();
+		curCSSTop = jQuery.css( elem, "top" );
+		curCSSLeft = jQuery.css( elem, "left" );
+		calculatePosition = ( position === "absolute" || position === "fixed" ) &&
+			( curCSSTop + curCSSLeft ).indexOf( "auto" ) > -1;
+
+		// Need to be able to calculate position if either
+		// top or left is auto and position is either absolute or fixed
+		if ( calculatePosition ) {
+			curPosition = curElem.position();
+			curTop = curPosition.top;
+			curLeft = curPosition.left;
+
+		} else {
+			curTop = parseFloat( curCSSTop ) || 0;
+			curLeft = parseFloat( curCSSLeft ) || 0;
+		}
+
+		if ( isFunction( options ) ) {
+
+			// Use jQuery.extend here to allow modification of coordinates argument (gh-1848)
+			options = options.call( elem, i, jQuery.extend( {}, curOffset ) );
+		}
+
+		if ( options.top != null ) {
+			props.top = ( options.top - curOffset.top ) + curTop;
+		}
+		if ( options.left != null ) {
+			props.left = ( options.left - curOffset.left ) + curLeft;
+		}
+
+		if ( "using" in options ) {
+			options.using.call( elem, props );
+
+		} else {
+			curElem.css( props );
+		}
+	}
+};
+
+jQuery.fn.extend( {
+
+	// offset() relates an element's border box to the document origin
+	offset: function( options ) {
+
+		// Preserve chaining for setter
+		if ( arguments.length ) {
+			return options === undefined ?
+				this :
+				this.each( function( i ) {
+					jQuery.offset.setOffset( this, options, i );
+				} );
+		}
+
+		var rect, win,
+			elem = this[ 0 ];
+
+		if ( !elem ) {
+			return;
+		}
+
+		// Return zeros for disconnected and hidden (display: none) elements (gh-2310)
+		// Support: IE <=11 only
+		// Running getBoundingClientRect on a
+		// disconnected node in IE throws an error
+		if ( !elem.getClientRects().length ) {
+			return { top: 0, left: 0 };
+		}
+
+		// Get document-relative position by adding viewport scroll to viewport-relative gBCR
+		rect = elem.getBoundingClientRect();
+		win = elem.ownerDocument.defaultView;
+		return {
+			top: rect.top + win.pageYOffset,
+			left: rect.left + win.pageXOffset
+		};
+	},
+
+	// position() relates an element's margin box to its offset parent's padding box
+	// This corresponds to the behavior of CSS absolute positioning
+	position: function() {
+		if ( !this[ 0 ] ) {
+			return;
+		}
+
+		var offsetParent, offset, doc,
+			elem = this[ 0 ],
+			parentOffset = { top: 0, left: 0 };
+
+		// position:fixed elements are offset from the viewport, which itself always has zero offset
+		if ( jQuery.css( elem, "position" ) === "fixed" ) {
+
+			// Assume position:fixed implies availability of getBoundingClientRect
+			offset = elem.getBoundingClientRect();
+
+		} else {
+			offset = this.offset();
+
+			// Account for the *real* offset parent, which can be the document or its root element
+			// when a statically positioned element is identified
+			doc = elem.ownerDocument;
+			offsetParent = elem.offsetParent || doc.documentElement;
+			while ( offsetParent &&
+				( offsetParent === doc.body || offsetParent === doc.documentElement ) &&
+				jQuery.css( offsetParent, "position" ) === "static" ) {
+
+				offsetParent = offsetParent.parentNode;
+			}
+			if ( offsetParent && offsetParent !== elem && offsetParent.nodeType === 1 ) {
+
+				// Incorporate borders into its offset, since they are outside its content origin
+				parentOffset = jQuery( offsetParent ).offset();
+				parentOffset.top += jQuery.css( offsetParent, "borderTopWidth", true );
+				parentOffset.left += jQuery.css( offsetParent, "borderLeftWidth", true );
+			}
+		}
+
+		// Subtract parent offsets and element margins
+		return {
+			top: offset.top - parentOffset.top - jQuery.css( elem, "marginTop", true ),
+			left: offset.left - parentOffset.left - jQuery.css( elem, "marginLeft", true )
+		};
+	},
+
+	// This method will return documentElement in the following cases:
+	// 1) For the element inside the iframe without offsetParent, this method will return
+	//    documentElement of the parent window
+	// 2) For the hidden or detached element
+	// 3) For body or html element, i.e. in case of the html node - it will return itself
+	//
+	// but those exceptions were never presented as a real life use-cases
+	// and might be considered as more preferable results.
+	//
+	// This logic, however, is not guaranteed and can change at any point in the future
+	offsetParent: function() {
+		return this.map( function() {
+			var offsetParent = this.offsetParent;
+
+			while ( offsetParent && jQuery.css( offsetParent, "position" ) === "static" ) {
+				offsetParent = offsetParent.offsetParent;
+			}
+
+			return offsetParent || documentElement;
+		} );
+	}
+} );
+
+// Create scrollLeft and scrollTop methods
+jQuery.each( { scrollLeft: "pageXOffset", scrollTop: "pageYOffset" }, function( method, prop ) {
+	var top = "pageYOffset" === prop;
+
+	jQuery.fn[ method ] = function( val ) {
+		return access( this, function( elem, method, val ) {
+
+			// Coalesce documents and windows
+			var win;
+			if ( isWindow( elem ) ) {
+				win = elem;
+			} else if ( elem.nodeType === 9 ) {
+				win = elem.defaultView;
+			}
+
+			if ( val === undefined ) {
+				return win ? win[ prop ] : elem[ method ];
+			}
+
+			if ( win ) {
+				win.scrollTo(
+					!top ? val : win.pageXOffset,
+					top ? val : win.pageYOffset
+				);
+
+			} else {
+				elem[ method ] = val;
+			}
+		}, method, val, arguments.length );
+	};
+} );
+
+// Support: Safari <=7 - 9.1, Chrome <=37 - 49
+// Add the top/left cssHooks using jQuery.fn.position
+// Webkit bug: https://bugs.webkit.org/show_bug.cgi?id=29084
+// Blink bug: https://bugs.chromium.org/p/chromium/issues/detail?id=589347
+// getComputedStyle returns percent when specified for top/left/bottom/right;
+// rather than make the css module depend on the offset module, just check for it here
+jQuery.each( [ "top", "left" ], function( _i, prop ) {
+	jQuery.cssHooks[ prop ] = addGetHookIf( support.pixelPosition,
+		function( elem, computed ) {
+			if ( computed ) {
+				computed = curCSS( elem, prop );
+
+				// If curCSS returns percentage, fallback to offset
+				return rnumnonpx.test( computed ) ?
+					jQuery( elem ).position()[ prop ] + "px" :
+					computed;
+			}
+		}
+	);
+} );
+
+
+// Create innerHeight, innerWidth, height, width, outerHeight and outerWidth methods
+jQuery.each( { Height: "height", Width: "width" }, function( name, type ) {
+	jQuery.each( {
+		padding: "inner" + name,
+		content: type,
+		"": "outer" + name
+	}, function( defaultExtra, funcName ) {
+
+		// Margin is only for outerHeight, outerWidth
+		jQuery.fn[ funcName ] = function( margin, value ) {
+			var chainable = arguments.length && ( defaultExtra || typeof margin !== "boolean" ),
+				extra = defaultExtra || ( margin === true || value === true ? "margin" : "border" );
+
+			return access( this, function( elem, type, value ) {
+				var doc;
+
+				if ( isWindow( elem ) ) {
+
+					// $( window ).outerWidth/Height return w/h including scrollbars (gh-1729)
+					return funcName.indexOf( "outer" ) === 0 ?
+						elem[ "inner" + name ] :
+						elem.document.documentElement[ "client" + name ];
+				}
+
+				// Get document width or height
+				if ( elem.nodeType === 9 ) {
+					doc = elem.documentElement;
+
+					// Either scroll[Width/Height] or offset[Width/Height] or client[Width/Height],
+					// whichever is greatest
+					return Math.max(
+						elem.body[ "scroll" + name ], doc[ "scroll" + name ],
+						elem.body[ "offset" + name ], doc[ "offset" + name ],
+						doc[ "client" + name ]
+					);
+				}
+
+				return value === undefined ?
+
+					// Get width or height on the element, requesting but not forcing parseFloat
+					jQuery.css( elem, type, extra ) :
+
+					// Set width or height on the element
+					jQuery.style( elem, type, value, extra );
+			}, type, chainable ? margin : undefined, chainable );
+		};
+	} );
+} );
+
+
+jQuery.each( [
+	"ajaxStart",
+	"ajaxStop",
+	"ajaxComplete",
+	"ajaxError",
+	"ajaxSuccess",
+	"ajaxSend"
+], function( _i, type ) {
+	jQuery.fn[ type ] = function( fn ) {
+		return this.on( type, fn );
+	};
+} );
+
+
+
+
+jQuery.fn.extend( {
+
+	bind: function( types, data, fn ) {
+		return this.on( types, null, data, fn );
+	},
+	unbind: function( types, fn ) {
+		return this.off( types, null, fn );
+	},
+
+	delegate: function( selector, types, data, fn ) {
+		return this.on( types, selector, data, fn );
+	},
+	undelegate: function( selector, types, fn ) {
+
+		// ( namespace ) or ( selector, types [, fn] )
+		return arguments.length === 1 ?
+			this.off( selector, "**" ) :
+			this.off( types, selector || "**", fn );
+	},
+
+	hover: function( fnOver, fnOut ) {
+		return this.mouseenter( fnOver ).mouseleave( fnOut || fnOver );
+	}
+} );
+
+jQuery.each(
+	( "blur focus focusin focusout resize scroll click dblclick " +
+	"mousedown mouseup mousemove mouseover mouseout mouseenter mouseleave " +
+	"change select submit keydown keypress keyup contextmenu" ).split( " " ),
+	function( _i, name ) {
+
+		// Handle event binding
+		jQuery.fn[ name ] = function( data, fn ) {
+			return arguments.length > 0 ?
+				this.on( name, null, data, fn ) :
+				this.trigger( name );
+		};
+	}
+);
+
+
+
+
+// Support: Android <=4.0 only
+// Make sure we trim BOM and NBSP
+var rtrim = /^[\s\uFEFF\xA0]+|[\s\uFEFF\xA0]+$/g;
+
+// Bind a function to a context, optionally partially applying any
+// arguments.
+// jQuery.proxy is deprecated to promote standards (specifically Function#bind)
+// However, it is not slated for removal any time soon
+jQuery.proxy = function( fn, context ) {
+	var tmp, args, proxy;
+
+	if ( typeof context === "string" ) {
+		tmp = fn[ context ];
+		context = fn;
+		fn = tmp;
+	}
+
+	// Quick check to determine if target is callable, in the spec
+	// this throws a TypeError, but we will just return undefined.
+	if ( !isFunction( fn ) ) {
+		return undefined;
+	}
+
+	// Simulated bind
+	args = slice.call( arguments, 2 );
+	proxy = function() {
+		return fn.apply( context || this, args.concat( slice.call( arguments ) ) );
+	};
+
+	// Set the guid of unique handler to the same of original handler, so it can be removed
+	proxy.guid = fn.guid = fn.guid || jQuery.guid++;
+
+	return proxy;
+};
+
+jQuery.holdReady = function( hold ) {
+	if ( hold ) {
+		jQuery.readyWait++;
+	} else {
+		jQuery.ready( true );
+	}
+};
+jQuery.isArray = Array.isArray;
+jQuery.parseJSON = JSON.parse;
+jQuery.nodeName = nodeName;
+jQuery.isFunction = isFunction;
+jQuery.isWindow = isWindow;
+jQuery.camelCase = camelCase;
+jQuery.type = toType;
+
+jQuery.now = Date.now;
+
+jQuery.isNumeric = function( obj ) {
+
+	// As of jQuery 3.0, isNumeric is limited to
+	// strings and numbers (primitives or objects)
+	// that can be coerced to finite numbers (gh-2662)
+	var type = jQuery.type( obj );
+	return ( type === "number" || type === "string" ) &&
+
+		// parseFloat NaNs numeric-cast false positives ("")
+		// ...but misinterprets leading-number strings, particularly hex literals ("0x...")
+		// subtraction forces infinities to NaN
+		!isNaN( obj - parseFloat( obj ) );
+};
+
+jQuery.trim = function( text ) {
+	return text == null ?
+		"" :
+		( text + "" ).replace( rtrim, "" );
+};
+
+
+
+// Register as a named AMD module, since jQuery can be concatenated with other
+// files that may use define, but not via a proper concatenation script that
+// understands anonymous AMD modules. A named AMD is safest and most robust
+// way to register. Lowercase jquery is used because AMD module names are
+// derived from file names, and jQuery is normally delivered in a lowercase
+// file name. Do this after creating the global so that if an AMD module wants
+// to call noConflict to hide this version of jQuery, it will work.
+
+// Note that for maximum portability, libraries that are not jQuery should
+// declare themselves as anonymous modules, and avoid setting a global if an
+// AMD loader is present. jQuery is a special case. For more information, see
+// https://github.com/jrburke/requirejs/wiki/Updating-existing-libraries#wiki-anon
+
+if ( typeof define === "function" && define.amd ) {
+	define( "jquery", [], function() {
+		return jQuery;
+	} );
+}
+
+
+
+
+var
+
+	// Map over jQuery in case of overwrite
+	_jQuery = window.jQuery,
+
+	// Map over the $ in case of overwrite
+	_$ = window.$;
+
+jQuery.noConflict = function( deep ) {
+	if ( window.$ === jQuery ) {
+		window.$ = _$;
+	}
+
+	if ( deep && window.jQuery === jQuery ) {
+		window.jQuery = _jQuery;
+	}
+
+	return jQuery;
+};
+
+// Expose jQuery and $ identifiers, even in AMD
+// (#7102#comment:10, https://github.com/jquery/jquery/pull/557)
+// and CommonJS for browser emulators (#13566)
+if ( typeof noGlobal === "undefined" ) {
+	window.jQuery = window.$ = jQuery;
+}
+
+
+
+
+return jQuery;
+} );
diff --git a/_static/jquery.js b/_static/jquery.js
index b0614034a..c4c6022f2 100644
--- a/_static/jquery.js
+++ b/_static/jquery.js
@@ -1,2 +1,2 @@
-/*! jQuery v3.5.1 | (c) JS Foundation and other contributors | jquery.org/license */
-!function(e,t){"use strict";"object"==typeof module&&"object"==typeof module.exports?module.exports=e.document?t(e,!0):function(e){if(!e.document)throw new Error("jQuery requires a window with a document");return t(e)}:t(e)}("undefined"!=typeof window?window:this,function(C,e){"use strict";var t=[],r=Object.getPrototypeOf,s=t.slice,g=t.flat?function(e){return t.flat.call(e)}:function(e){return t.concat.apply([],e)},u=t.push,i=t.indexOf,n={},o=n.toString,v=n.hasOwnProperty,a=v.toString,l=a.call(Object),y={},m=function(e){return"function"==typeof e&&"number"!=typeof e.nodeType},x=function(e){return null!=e&&e===e.window},E=C.document,c={type:!0,src:!0,nonce:!0,noModule:!0};function b(e,t,n){var r,i,o=(n=n||E).createElement("script");if(o.text=e,t)for(r in c)(i=t[r]||t.getAttribute&&t.getAttribute(r))&&o.setAttribute(r,i);n.head.appendChild(o).parentNode.removeChild(o)}function w(e){return null==e?e+"":"object"==typeof e||"function"==typeof e?n[o.call(e)]||"object":typeof e}var f="3.5.1",S=function(e,t){return new S.fn.init(e,t)};function p(e){var t=!!e&&"length"in e&&e.length,n=w(e);return!m(e)&&!x(e)&&("array"===n||0===t||"number"==typeof t&&0<t&&t-1 in e)}S.fn=S.prototype={jquery:f,constructor:S,length:0,toArray:function(){return s.call(this)},get:function(e){return null==e?s.call(this):e<0?this[e+this.length]:this[e]},pushStack:function(e){var t=S.merge(this.constructor(),e);return t.prevObject=this,t},each:function(e){return S.each(this,e)},map:function(n){return this.pushStack(S.map(this,function(e,t){return n.call(e,t,e)}))},slice:function(){return this.pushStack(s.apply(this,arguments))},first:function(){return this.eq(0)},last:function(){return this.eq(-1)},even:function(){return this.pushStack(S.grep(this,function(e,t){return(t+1)%2}))},odd:function(){return this.pushStack(S.grep(this,function(e,t){return t%2}))},eq:function(e){var t=this.length,n=+e+(e<0?t:0);return this.pushStack(0<=n&&n<t?[this[n]]:[])},end:function(){return this.prevObject||this.constructor()},push:u,sort:t.sort,splice:t.splice},S.extend=S.fn.extend=function(){var e,t,n,r,i,o,a=arguments[0]||{},s=1,u=arguments.length,l=!1;for("boolean"==typeof a&&(l=a,a=arguments[s]||{},s++),"object"==typeof a||m(a)||(a={}),s===u&&(a=this,s--);s<u;s++)if(null!=(e=arguments[s]))for(t in e)r=e[t],"__proto__"!==t&&a!==r&&(l&&r&&(S.isPlainObject(r)||(i=Array.isArray(r)))?(n=a[t],o=i&&!Array.isArray(n)?[]:i||S.isPlainObject(n)?n:{},i=!1,a[t]=S.extend(l,o,r)):void 0!==r&&(a[t]=r));return a},S.extend({expando:"jQuery"+(f+Math.random()).replace(/\D/g,""),isReady:!0,error:function(e){throw new Error(e)},noop:function(){},isPlainObject:function(e){var t,n;return!(!e||"[object Object]"!==o.call(e))&&(!(t=r(e))||"function"==typeof(n=v.call(t,"constructor")&&t.constructor)&&a.call(n)===l)},isEmptyObject:function(e){var t;for(t in e)return!1;return!0},globalEval:function(e,t,n){b(e,{nonce:t&&t.nonce},n)},each:function(e,t){var n,r=0;if(p(e)){for(n=e.length;r<n;r++)if(!1===t.call(e[r],r,e[r]))break}else for(r in e)if(!1===t.call(e[r],r,e[r]))break;return e},makeArray:function(e,t){var n=t||[];return null!=e&&(p(Object(e))?S.merge(n,"string"==typeof e?[e]:e):u.call(n,e)),n},inArray:function(e,t,n){return null==t?-1:i.call(t,e,n)},merge:function(e,t){for(var n=+t.length,r=0,i=e.length;r<n;r++)e[i++]=t[r];return e.length=i,e},grep:function(e,t,n){for(var r=[],i=0,o=e.length,a=!n;i<o;i++)!t(e[i],i)!==a&&r.push(e[i]);return r},map:function(e,t,n){var r,i,o=0,a=[];if(p(e))for(r=e.length;o<r;o++)null!=(i=t(e[o],o,n))&&a.push(i);else for(o in e)null!=(i=t(e[o],o,n))&&a.push(i);return g(a)},guid:1,support:y}),"function"==typeof Symbol&&(S.fn[Symbol.iterator]=t[Symbol.iterator]),S.each("Boolean Number String Function Array Date RegExp Object Error Symbol".split(" "),function(e,t){n["[object "+t+"]"]=t.toLowerCase()});var d=function(n){var e,d,b,o,i,h,f,g,w,u,l,T,C,a,E,v,s,c,y,S="sizzle"+1*new Date,p=n.document,k=0,r=0,m=ue(),x=ue(),A=ue(),N=ue(),D=function(e,t){return e===t&&(l=!0),0},j={}.hasOwnProperty,t=[],q=t.pop,L=t.push,H=t.push,O=t.slice,P=function(e,t){for(var n=0,r=e.length;n<r;n++)if(e[n]===t)return n;return-1},R="checked|selected|async|autofocus|autoplay|controls|defer|disabled|hidden|ismap|loop|multiple|open|readonly|required|scoped",M="[\\x20\\t\\r\\n\\f]",I="(?:\\\\[\\da-fA-F]{1,6}"+M+"?|\\\\[^\\r\\n\\f]|[\\w-]|[^\0-\\x7f])+",W="\\["+M+"*("+I+")(?:"+M+"*([*^$|!~]?=)"+M+"*(?:'((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\"|("+I+"))|)"+M+"*\\]",F=":("+I+")(?:\\((('((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\")|((?:\\\\.|[^\\\\()[\\]]|"+W+")*)|.*)\\)|)",B=new RegExp(M+"+","g"),$=new RegExp("^"+M+"+|((?:^|[^\\\\])(?:\\\\.)*)"+M+"+$","g"),_=new RegExp("^"+M+"*,"+M+"*"),z=new RegExp("^"+M+"*([>+~]|"+M+")"+M+"*"),U=new RegExp(M+"|>"),X=new RegExp(F),V=new RegExp("^"+I+"$"),G={ID:new RegExp("^#("+I+")"),CLASS:new RegExp("^\\.("+I+")"),TAG:new RegExp("^("+I+"|[*])"),ATTR:new RegExp("^"+W),PSEUDO:new RegExp("^"+F),CHILD:new RegExp("^:(only|first|last|nth|nth-last)-(child|of-type)(?:\\("+M+"*(even|odd|(([+-]|)(\\d*)n|)"+M+"*(?:([+-]|)"+M+"*(\\d+)|))"+M+"*\\)|)","i"),bool:new RegExp("^(?:"+R+")$","i"),needsContext:new RegExp("^"+M+"*[>+~]|:(even|odd|eq|gt|lt|nth|first|last)(?:\\("+M+"*((?:-\\d)?\\d*)"+M+"*\\)|)(?=[^-]|$)","i")},Y=/HTML$/i,Q=/^(?:input|select|textarea|button)$/i,J=/^h\d$/i,K=/^[^{]+\{\s*\[native \w/,Z=/^(?:#([\w-]+)|(\w+)|\.([\w-]+))$/,ee=/[+~]/,te=new RegExp("\\\\[\\da-fA-F]{1,6}"+M+"?|\\\\([^\\r\\n\\f])","g"),ne=function(e,t){var n="0x"+e.slice(1)-65536;return t||(n<0?String.fromCharCode(n+65536):String.fromCharCode(n>>10|55296,1023&n|56320))},re=/([\0-\x1f\x7f]|^-?\d)|^-$|[^\0-\x1f\x7f-\uFFFF\w-]/g,ie=function(e,t){return t?"\0"===e?"\ufffd":e.slice(0,-1)+"\\"+e.charCodeAt(e.length-1).toString(16)+" ":"\\"+e},oe=function(){T()},ae=be(function(e){return!0===e.disabled&&"fieldset"===e.nodeName.toLowerCase()},{dir:"parentNode",next:"legend"});try{H.apply(t=O.call(p.childNodes),p.childNodes),t[p.childNodes.length].nodeType}catch(e){H={apply:t.length?function(e,t){L.apply(e,O.call(t))}:function(e,t){var n=e.length,r=0;while(e[n++]=t[r++]);e.length=n-1}}}function se(t,e,n,r){var i,o,a,s,u,l,c,f=e&&e.ownerDocument,p=e?e.nodeType:9;if(n=n||[],"string"!=typeof t||!t||1!==p&&9!==p&&11!==p)return n;if(!r&&(T(e),e=e||C,E)){if(11!==p&&(u=Z.exec(t)))if(i=u[1]){if(9===p){if(!(a=e.getElementById(i)))return n;if(a.id===i)return n.push(a),n}else if(f&&(a=f.getElementById(i))&&y(e,a)&&a.id===i)return n.push(a),n}else{if(u[2])return H.apply(n,e.getElementsByTagName(t)),n;if((i=u[3])&&d.getElementsByClassName&&e.getElementsByClassName)return H.apply(n,e.getElementsByClassName(i)),n}if(d.qsa&&!N[t+" "]&&(!v||!v.test(t))&&(1!==p||"object"!==e.nodeName.toLowerCase())){if(c=t,f=e,1===p&&(U.test(t)||z.test(t))){(f=ee.test(t)&&ye(e.parentNode)||e)===e&&d.scope||((s=e.getAttribute("id"))?s=s.replace(re,ie):e.setAttribute("id",s=S)),o=(l=h(t)).length;while(o--)l[o]=(s?"#"+s:":scope")+" "+xe(l[o]);c=l.join(",")}try{return H.apply(n,f.querySelectorAll(c)),n}catch(e){N(t,!0)}finally{s===S&&e.removeAttribute("id")}}}return g(t.replace($,"$1"),e,n,r)}function ue(){var r=[];return function e(t,n){return r.push(t+" ")>b.cacheLength&&delete e[r.shift()],e[t+" "]=n}}function le(e){return e[S]=!0,e}function ce(e){var t=C.createElement("fieldset");try{return!!e(t)}catch(e){return!1}finally{t.parentNode&&t.parentNode.removeChild(t),t=null}}function fe(e,t){var n=e.split("|"),r=n.length;while(r--)b.attrHandle[n[r]]=t}function pe(e,t){var n=t&&e,r=n&&1===e.nodeType&&1===t.nodeType&&e.sourceIndex-t.sourceIndex;if(r)return r;if(n)while(n=n.nextSibling)if(n===t)return-1;return e?1:-1}function de(t){return function(e){return"input"===e.nodeName.toLowerCase()&&e.type===t}}function he(n){return function(e){var t=e.nodeName.toLowerCase();return("input"===t||"button"===t)&&e.type===n}}function ge(t){return function(e){return"form"in e?e.parentNode&&!1===e.disabled?"label"in e?"label"in e.parentNode?e.parentNode.disabled===t:e.disabled===t:e.isDisabled===t||e.isDisabled!==!t&&ae(e)===t:e.disabled===t:"label"in e&&e.disabled===t}}function ve(a){return le(function(o){return o=+o,le(function(e,t){var n,r=a([],e.length,o),i=r.length;while(i--)e[n=r[i]]&&(e[n]=!(t[n]=e[n]))})})}function ye(e){return e&&"undefined"!=typeof e.getElementsByTagName&&e}for(e in d=se.support={},i=se.isXML=function(e){var t=e.namespaceURI,n=(e.ownerDocument||e).documentElement;return!Y.test(t||n&&n.nodeName||"HTML")},T=se.setDocument=function(e){var t,n,r=e?e.ownerDocument||e:p;return r!=C&&9===r.nodeType&&r.documentElement&&(a=(C=r).documentElement,E=!i(C),p!=C&&(n=C.defaultView)&&n.top!==n&&(n.addEventListener?n.addEventListener("unload",oe,!1):n.attachEvent&&n.attachEvent("onunload",oe)),d.scope=ce(function(e){return a.appendChild(e).appendChild(C.createElement("div")),"undefined"!=typeof e.querySelectorAll&&!e.querySelectorAll(":scope fieldset div").length}),d.attributes=ce(function(e){return e.className="i",!e.getAttribute("className")}),d.getElementsByTagName=ce(function(e){return e.appendChild(C.createComment("")),!e.getElementsByTagName("*").length}),d.getElementsByClassName=K.test(C.getElementsByClassName),d.getById=ce(function(e){return a.appendChild(e).id=S,!C.getElementsByName||!C.getElementsByName(S).length}),d.getById?(b.filter.ID=function(e){var t=e.replace(te,ne);return function(e){return e.getAttribute("id")===t}},b.find.ID=function(e,t){if("undefined"!=typeof t.getElementById&&E){var n=t.getElementById(e);return n?[n]:[]}}):(b.filter.ID=function(e){var n=e.replace(te,ne);return function(e){var t="undefined"!=typeof e.getAttributeNode&&e.getAttributeNode("id");return t&&t.value===n}},b.find.ID=function(e,t){if("undefined"!=typeof t.getElementById&&E){var n,r,i,o=t.getElementById(e);if(o){if((n=o.getAttributeNode("id"))&&n.value===e)return[o];i=t.getElementsByName(e),r=0;while(o=i[r++])if((n=o.getAttributeNode("id"))&&n.value===e)return[o]}return[]}}),b.find.TAG=d.getElementsByTagName?function(e,t){return"undefined"!=typeof t.getElementsByTagName?t.getElementsByTagName(e):d.qsa?t.querySelectorAll(e):void 0}:function(e,t){var n,r=[],i=0,o=t.getElementsByTagName(e);if("*"===e){while(n=o[i++])1===n.nodeType&&r.push(n);return r}return o},b.find.CLASS=d.getElementsByClassName&&function(e,t){if("undefined"!=typeof t.getElementsByClassName&&E)return t.getElementsByClassName(e)},s=[],v=[],(d.qsa=K.test(C.querySelectorAll))&&(ce(function(e){var t;a.appendChild(e).innerHTML="<a id='"+S+"'></a><select id='"+S+"-\r\\' msallowcapture=''><option selected=''></option></select>",e.querySelectorAll("[msallowcapture^='']").length&&v.push("[*^$]="+M+"*(?:''|\"\")"),e.querySelectorAll("[selected]").length||v.push("\\["+M+"*(?:value|"+R+")"),e.querySelectorAll("[id~="+S+"-]").length||v.push("~="),(t=C.createElement("input")).setAttribute("name",""),e.appendChild(t),e.querySelectorAll("[name='']").length||v.push("\\["+M+"*name"+M+"*="+M+"*(?:''|\"\")"),e.querySelectorAll(":checked").length||v.push(":checked"),e.querySelectorAll("a#"+S+"+*").length||v.push(".#.+[+~]"),e.querySelectorAll("\\\f"),v.push("[\\r\\n\\f]")}),ce(function(e){e.innerHTML="<a href='' disabled='disabled'></a><select disabled='disabled'><option/></select>";var t=C.createElement("input");t.setAttribute("type","hidden"),e.appendChild(t).setAttribute("name","D"),e.querySelectorAll("[name=d]").length&&v.push("name"+M+"*[*^$|!~]?="),2!==e.querySelectorAll(":enabled").length&&v.push(":enabled",":disabled"),a.appendChild(e).disabled=!0,2!==e.querySelectorAll(":disabled").length&&v.push(":enabled",":disabled"),e.querySelectorAll("*,:x"),v.push(",.*:")})),(d.matchesSelector=K.test(c=a.matches||a.webkitMatchesSelector||a.mozMatchesSelector||a.oMatchesSelector||a.msMatchesSelector))&&ce(function(e){d.disconnectedMatch=c.call(e,"*"),c.call(e,"[s!='']:x"),s.push("!=",F)}),v=v.length&&new RegExp(v.join("|")),s=s.length&&new RegExp(s.join("|")),t=K.test(a.compareDocumentPosition),y=t||K.test(a.contains)?function(e,t){var n=9===e.nodeType?e.documentElement:e,r=t&&t.parentNode;return e===r||!(!r||1!==r.nodeType||!(n.contains?n.contains(r):e.compareDocumentPosition&&16&e.compareDocumentPosition(r)))}:function(e,t){if(t)while(t=t.parentNode)if(t===e)return!0;return!1},D=t?function(e,t){if(e===t)return l=!0,0;var n=!e.compareDocumentPosition-!t.compareDocumentPosition;return n||(1&(n=(e.ownerDocument||e)==(t.ownerDocument||t)?e.compareDocumentPosition(t):1)||!d.sortDetached&&t.compareDocumentPosition(e)===n?e==C||e.ownerDocument==p&&y(p,e)?-1:t==C||t.ownerDocument==p&&y(p,t)?1:u?P(u,e)-P(u,t):0:4&n?-1:1)}:function(e,t){if(e===t)return l=!0,0;var n,r=0,i=e.parentNode,o=t.parentNode,a=[e],s=[t];if(!i||!o)return e==C?-1:t==C?1:i?-1:o?1:u?P(u,e)-P(u,t):0;if(i===o)return pe(e,t);n=e;while(n=n.parentNode)a.unshift(n);n=t;while(n=n.parentNode)s.unshift(n);while(a[r]===s[r])r++;return r?pe(a[r],s[r]):a[r]==p?-1:s[r]==p?1:0}),C},se.matches=function(e,t){return se(e,null,null,t)},se.matchesSelector=function(e,t){if(T(e),d.matchesSelector&&E&&!N[t+" "]&&(!s||!s.test(t))&&(!v||!v.test(t)))try{var n=c.call(e,t);if(n||d.disconnectedMatch||e.document&&11!==e.document.nodeType)return n}catch(e){N(t,!0)}return 0<se(t,C,null,[e]).length},se.contains=function(e,t){return(e.ownerDocument||e)!=C&&T(e),y(e,t)},se.attr=function(e,t){(e.ownerDocument||e)!=C&&T(e);var n=b.attrHandle[t.toLowerCase()],r=n&&j.call(b.attrHandle,t.toLowerCase())?n(e,t,!E):void 0;return void 0!==r?r:d.attributes||!E?e.getAttribute(t):(r=e.getAttributeNode(t))&&r.specified?r.value:null},se.escape=function(e){return(e+"").replace(re,ie)},se.error=function(e){throw new Error("Syntax error, unrecognized expression: "+e)},se.uniqueSort=function(e){var t,n=[],r=0,i=0;if(l=!d.detectDuplicates,u=!d.sortStable&&e.slice(0),e.sort(D),l){while(t=e[i++])t===e[i]&&(r=n.push(i));while(r--)e.splice(n[r],1)}return u=null,e},o=se.getText=function(e){var t,n="",r=0,i=e.nodeType;if(i){if(1===i||9===i||11===i){if("string"==typeof e.textContent)return e.textContent;for(e=e.firstChild;e;e=e.nextSibling)n+=o(e)}else if(3===i||4===i)return e.nodeValue}else while(t=e[r++])n+=o(t);return n},(b=se.selectors={cacheLength:50,createPseudo:le,match:G,attrHandle:{},find:{},relative:{">":{dir:"parentNode",first:!0}," ":{dir:"parentNode"},"+":{dir:"previousSibling",first:!0},"~":{dir:"previousSibling"}},preFilter:{ATTR:function(e){return e[1]=e[1].replace(te,ne),e[3]=(e[3]||e[4]||e[5]||"").replace(te,ne),"~="===e[2]&&(e[3]=" "+e[3]+" "),e.slice(0,4)},CHILD:function(e){return e[1]=e[1].toLowerCase(),"nth"===e[1].slice(0,3)?(e[3]||se.error(e[0]),e[4]=+(e[4]?e[5]+(e[6]||1):2*("even"===e[3]||"odd"===e[3])),e[5]=+(e[7]+e[8]||"odd"===e[3])):e[3]&&se.error(e[0]),e},PSEUDO:function(e){var t,n=!e[6]&&e[2];return G.CHILD.test(e[0])?null:(e[3]?e[2]=e[4]||e[5]||"":n&&X.test(n)&&(t=h(n,!0))&&(t=n.indexOf(")",n.length-t)-n.length)&&(e[0]=e[0].slice(0,t),e[2]=n.slice(0,t)),e.slice(0,3))}},filter:{TAG:function(e){var t=e.replace(te,ne).toLowerCase();return"*"===e?function(){return!0}:function(e){return e.nodeName&&e.nodeName.toLowerCase()===t}},CLASS:function(e){var t=m[e+" "];return t||(t=new RegExp("(^|"+M+")"+e+"("+M+"|$)"))&&m(e,function(e){return t.test("string"==typeof e.className&&e.className||"undefined"!=typeof e.getAttribute&&e.getAttribute("class")||"")})},ATTR:function(n,r,i){return function(e){var t=se.attr(e,n);return null==t?"!="===r:!r||(t+="","="===r?t===i:"!="===r?t!==i:"^="===r?i&&0===t.indexOf(i):"*="===r?i&&-1<t.indexOf(i):"$="===r?i&&t.slice(-i.length)===i:"~="===r?-1<(" "+t.replace(B," ")+" ").indexOf(i):"|="===r&&(t===i||t.slice(0,i.length+1)===i+"-"))}},CHILD:function(h,e,t,g,v){var y="nth"!==h.slice(0,3),m="last"!==h.slice(-4),x="of-type"===e;return 1===g&&0===v?function(e){return!!e.parentNode}:function(e,t,n){var r,i,o,a,s,u,l=y!==m?"nextSibling":"previousSibling",c=e.parentNode,f=x&&e.nodeName.toLowerCase(),p=!n&&!x,d=!1;if(c){if(y){while(l){a=e;while(a=a[l])if(x?a.nodeName.toLowerCase()===f:1===a.nodeType)return!1;u=l="only"===h&&!u&&"nextSibling"}return!0}if(u=[m?c.firstChild:c.lastChild],m&&p){d=(s=(r=(i=(o=(a=c)[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]||[])[0]===k&&r[1])&&r[2],a=s&&c.childNodes[s];while(a=++s&&a&&a[l]||(d=s=0)||u.pop())if(1===a.nodeType&&++d&&a===e){i[h]=[k,s,d];break}}else if(p&&(d=s=(r=(i=(o=(a=e)[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]||[])[0]===k&&r[1]),!1===d)while(a=++s&&a&&a[l]||(d=s=0)||u.pop())if((x?a.nodeName.toLowerCase()===f:1===a.nodeType)&&++d&&(p&&((i=(o=a[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]=[k,d]),a===e))break;return(d-=v)===g||d%g==0&&0<=d/g}}},PSEUDO:function(e,o){var t,a=b.pseudos[e]||b.setFilters[e.toLowerCase()]||se.error("unsupported pseudo: "+e);return a[S]?a(o):1<a.length?(t=[e,e,"",o],b.setFilters.hasOwnProperty(e.toLowerCase())?le(function(e,t){var n,r=a(e,o),i=r.length;while(i--)e[n=P(e,r[i])]=!(t[n]=r[i])}):function(e){return a(e,0,t)}):a}},pseudos:{not:le(function(e){var r=[],i=[],s=f(e.replace($,"$1"));return s[S]?le(function(e,t,n,r){var i,o=s(e,null,r,[]),a=e.length;while(a--)(i=o[a])&&(e[a]=!(t[a]=i))}):function(e,t,n){return r[0]=e,s(r,null,n,i),r[0]=null,!i.pop()}}),has:le(function(t){return function(e){return 0<se(t,e).length}}),contains:le(function(t){return t=t.replace(te,ne),function(e){return-1<(e.textContent||o(e)).indexOf(t)}}),lang:le(function(n){return V.test(n||"")||se.error("unsupported lang: "+n),n=n.replace(te,ne).toLowerCase(),function(e){var t;do{if(t=E?e.lang:e.getAttribute("xml:lang")||e.getAttribute("lang"))return(t=t.toLowerCase())===n||0===t.indexOf(n+"-")}while((e=e.parentNode)&&1===e.nodeType);return!1}}),target:function(e){var t=n.location&&n.location.hash;return t&&t.slice(1)===e.id},root:function(e){return e===a},focus:function(e){return e===C.activeElement&&(!C.hasFocus||C.hasFocus())&&!!(e.type||e.href||~e.tabIndex)},enabled:ge(!1),disabled:ge(!0),checked:function(e){var t=e.nodeName.toLowerCase();return"input"===t&&!!e.checked||"option"===t&&!!e.selected},selected:function(e){return e.parentNode&&e.parentNode.selectedIndex,!0===e.selected},empty:function(e){for(e=e.firstChild;e;e=e.nextSibling)if(e.nodeType<6)return!1;return!0},parent:function(e){return!b.pseudos.empty(e)},header:function(e){return J.test(e.nodeName)},input:function(e){return Q.test(e.nodeName)},button:function(e){var t=e.nodeName.toLowerCase();return"input"===t&&"button"===e.type||"button"===t},text:function(e){var t;return"input"===e.nodeName.toLowerCase()&&"text"===e.type&&(null==(t=e.getAttribute("type"))||"text"===t.toLowerCase())},first:ve(function(){return[0]}),last:ve(function(e,t){return[t-1]}),eq:ve(function(e,t,n){return[n<0?n+t:n]}),even:ve(function(e,t){for(var n=0;n<t;n+=2)e.push(n);return e}),odd:ve(function(e,t){for(var n=1;n<t;n+=2)e.push(n);return e}),lt:ve(function(e,t,n){for(var r=n<0?n+t:t<n?t:n;0<=--r;)e.push(r);return e}),gt:ve(function(e,t,n){for(var r=n<0?n+t:n;++r<t;)e.push(r);return e})}}).pseudos.nth=b.pseudos.eq,{radio:!0,checkbox:!0,file:!0,password:!0,image:!0})b.pseudos[e]=de(e);for(e in{submit:!0,reset:!0})b.pseudos[e]=he(e);function me(){}function xe(e){for(var t=0,n=e.length,r="";t<n;t++)r+=e[t].value;return r}function be(s,e,t){var u=e.dir,l=e.next,c=l||u,f=t&&"parentNode"===c,p=r++;return e.first?function(e,t,n){while(e=e[u])if(1===e.nodeType||f)return s(e,t,n);return!1}:function(e,t,n){var r,i,o,a=[k,p];if(n){while(e=e[u])if((1===e.nodeType||f)&&s(e,t,n))return!0}else while(e=e[u])if(1===e.nodeType||f)if(i=(o=e[S]||(e[S]={}))[e.uniqueID]||(o[e.uniqueID]={}),l&&l===e.nodeName.toLowerCase())e=e[u]||e;else{if((r=i[c])&&r[0]===k&&r[1]===p)return a[2]=r[2];if((i[c]=a)[2]=s(e,t,n))return!0}return!1}}function we(i){return 1<i.length?function(e,t,n){var r=i.length;while(r--)if(!i[r](e,t,n))return!1;return!0}:i[0]}function Te(e,t,n,r,i){for(var o,a=[],s=0,u=e.length,l=null!=t;s<u;s++)(o=e[s])&&(n&&!n(o,r,i)||(a.push(o),l&&t.push(s)));return a}function Ce(d,h,g,v,y,e){return v&&!v[S]&&(v=Ce(v)),y&&!y[S]&&(y=Ce(y,e)),le(function(e,t,n,r){var i,o,a,s=[],u=[],l=t.length,c=e||function(e,t,n){for(var r=0,i=t.length;r<i;r++)se(e,t[r],n);return n}(h||"*",n.nodeType?[n]:n,[]),f=!d||!e&&h?c:Te(c,s,d,n,r),p=g?y||(e?d:l||v)?[]:t:f;if(g&&g(f,p,n,r),v){i=Te(p,u),v(i,[],n,r),o=i.length;while(o--)(a=i[o])&&(p[u[o]]=!(f[u[o]]=a))}if(e){if(y||d){if(y){i=[],o=p.length;while(o--)(a=p[o])&&i.push(f[o]=a);y(null,p=[],i,r)}o=p.length;while(o--)(a=p[o])&&-1<(i=y?P(e,a):s[o])&&(e[i]=!(t[i]=a))}}else p=Te(p===t?p.splice(l,p.length):p),y?y(null,t,p,r):H.apply(t,p)})}function Ee(e){for(var i,t,n,r=e.length,o=b.relative[e[0].type],a=o||b.relative[" "],s=o?1:0,u=be(function(e){return e===i},a,!0),l=be(function(e){return-1<P(i,e)},a,!0),c=[function(e,t,n){var r=!o&&(n||t!==w)||((i=t).nodeType?u(e,t,n):l(e,t,n));return i=null,r}];s<r;s++)if(t=b.relative[e[s].type])c=[be(we(c),t)];else{if((t=b.filter[e[s].type].apply(null,e[s].matches))[S]){for(n=++s;n<r;n++)if(b.relative[e[n].type])break;return Ce(1<s&&we(c),1<s&&xe(e.slice(0,s-1).concat({value:" "===e[s-2].type?"*":""})).replace($,"$1"),t,s<n&&Ee(e.slice(s,n)),n<r&&Ee(e=e.slice(n)),n<r&&xe(e))}c.push(t)}return we(c)}return me.prototype=b.filters=b.pseudos,b.setFilters=new me,h=se.tokenize=function(e,t){var n,r,i,o,a,s,u,l=x[e+" "];if(l)return t?0:l.slice(0);a=e,s=[],u=b.preFilter;while(a){for(o in n&&!(r=_.exec(a))||(r&&(a=a.slice(r[0].length)||a),s.push(i=[])),n=!1,(r=z.exec(a))&&(n=r.shift(),i.push({value:n,type:r[0].replace($," ")}),a=a.slice(n.length)),b.filter)!(r=G[o].exec(a))||u[o]&&!(r=u[o](r))||(n=r.shift(),i.push({value:n,type:o,matches:r}),a=a.slice(n.length));if(!n)break}return t?a.length:a?se.error(e):x(e,s).slice(0)},f=se.compile=function(e,t){var n,v,y,m,x,r,i=[],o=[],a=A[e+" "];if(!a){t||(t=h(e)),n=t.length;while(n--)(a=Ee(t[n]))[S]?i.push(a):o.push(a);(a=A(e,(v=o,m=0<(y=i).length,x=0<v.length,r=function(e,t,n,r,i){var o,a,s,u=0,l="0",c=e&&[],f=[],p=w,d=e||x&&b.find.TAG("*",i),h=k+=null==p?1:Math.random()||.1,g=d.length;for(i&&(w=t==C||t||i);l!==g&&null!=(o=d[l]);l++){if(x&&o){a=0,t||o.ownerDocument==C||(T(o),n=!E);while(s=v[a++])if(s(o,t||C,n)){r.push(o);break}i&&(k=h)}m&&((o=!s&&o)&&u--,e&&c.push(o))}if(u+=l,m&&l!==u){a=0;while(s=y[a++])s(c,f,t,n);if(e){if(0<u)while(l--)c[l]||f[l]||(f[l]=q.call(r));f=Te(f)}H.apply(r,f),i&&!e&&0<f.length&&1<u+y.length&&se.uniqueSort(r)}return i&&(k=h,w=p),c},m?le(r):r))).selector=e}return a},g=se.select=function(e,t,n,r){var i,o,a,s,u,l="function"==typeof e&&e,c=!r&&h(e=l.selector||e);if(n=n||[],1===c.length){if(2<(o=c[0]=c[0].slice(0)).length&&"ID"===(a=o[0]).type&&9===t.nodeType&&E&&b.relative[o[1].type]){if(!(t=(b.find.ID(a.matches[0].replace(te,ne),t)||[])[0]))return n;l&&(t=t.parentNode),e=e.slice(o.shift().value.length)}i=G.needsContext.test(e)?0:o.length;while(i--){if(a=o[i],b.relative[s=a.type])break;if((u=b.find[s])&&(r=u(a.matches[0].replace(te,ne),ee.test(o[0].type)&&ye(t.parentNode)||t))){if(o.splice(i,1),!(e=r.length&&xe(o)))return H.apply(n,r),n;break}}}return(l||f(e,c))(r,t,!E,n,!t||ee.test(e)&&ye(t.parentNode)||t),n},d.sortStable=S.split("").sort(D).join("")===S,d.detectDuplicates=!!l,T(),d.sortDetached=ce(function(e){return 1&e.compareDocumentPosition(C.createElement("fieldset"))}),ce(function(e){return e.innerHTML="<a href='#'></a>","#"===e.firstChild.getAttribute("href")})||fe("type|href|height|width",function(e,t,n){if(!n)return e.getAttribute(t,"type"===t.toLowerCase()?1:2)}),d.attributes&&ce(function(e){return e.innerHTML="<input/>",e.firstChild.setAttribute("value",""),""===e.firstChild.getAttribute("value")})||fe("value",function(e,t,n){if(!n&&"input"===e.nodeName.toLowerCase())return e.defaultValue}),ce(function(e){return null==e.getAttribute("disabled")})||fe(R,function(e,t,n){var r;if(!n)return!0===e[t]?t.toLowerCase():(r=e.getAttributeNode(t))&&r.specified?r.value:null}),se}(C);S.find=d,S.expr=d.selectors,S.expr[":"]=S.expr.pseudos,S.uniqueSort=S.unique=d.uniqueSort,S.text=d.getText,S.isXMLDoc=d.isXML,S.contains=d.contains,S.escapeSelector=d.escape;var h=function(e,t,n){var r=[],i=void 0!==n;while((e=e[t])&&9!==e.nodeType)if(1===e.nodeType){if(i&&S(e).is(n))break;r.push(e)}return r},T=function(e,t){for(var n=[];e;e=e.nextSibling)1===e.nodeType&&e!==t&&n.push(e);return n},k=S.expr.match.needsContext;function A(e,t){return e.nodeName&&e.nodeName.toLowerCase()===t.toLowerCase()}var N=/^<([a-z][^\/\0>:\x20\t\r\n\f]*)[\x20\t\r\n\f]*\/?>(?:<\/\1>|)$/i;function D(e,n,r){return m(n)?S.grep(e,function(e,t){return!!n.call(e,t,e)!==r}):n.nodeType?S.grep(e,function(e){return e===n!==r}):"string"!=typeof n?S.grep(e,function(e){return-1<i.call(n,e)!==r}):S.filter(n,e,r)}S.filter=function(e,t,n){var r=t[0];return n&&(e=":not("+e+")"),1===t.length&&1===r.nodeType?S.find.matchesSelector(r,e)?[r]:[]:S.find.matches(e,S.grep(t,function(e){return 1===e.nodeType}))},S.fn.extend({find:function(e){var t,n,r=this.length,i=this;if("string"!=typeof e)return this.pushStack(S(e).filter(function(){for(t=0;t<r;t++)if(S.contains(i[t],this))return!0}));for(n=this.pushStack([]),t=0;t<r;t++)S.find(e,i[t],n);return 1<r?S.uniqueSort(n):n},filter:function(e){return this.pushStack(D(this,e||[],!1))},not:function(e){return this.pushStack(D(this,e||[],!0))},is:function(e){return!!D(this,"string"==typeof e&&k.test(e)?S(e):e||[],!1).length}});var j,q=/^(?:\s*(<[\w\W]+>)[^>]*|#([\w-]+))$/;(S.fn.init=function(e,t,n){var r,i;if(!e)return this;if(n=n||j,"string"==typeof e){if(!(r="<"===e[0]&&">"===e[e.length-1]&&3<=e.length?[null,e,null]:q.exec(e))||!r[1]&&t)return!t||t.jquery?(t||n).find(e):this.constructor(t).find(e);if(r[1]){if(t=t instanceof S?t[0]:t,S.merge(this,S.parseHTML(r[1],t&&t.nodeType?t.ownerDocument||t:E,!0)),N.test(r[1])&&S.isPlainObject(t))for(r in t)m(this[r])?this[r](t[r]):this.attr(r,t[r]);return this}return(i=E.getElementById(r[2]))&&(this[0]=i,this.length=1),this}return e.nodeType?(this[0]=e,this.length=1,this):m(e)?void 0!==n.ready?n.ready(e):e(S):S.makeArray(e,this)}).prototype=S.fn,j=S(E);var L=/^(?:parents|prev(?:Until|All))/,H={children:!0,contents:!0,next:!0,prev:!0};function O(e,t){while((e=e[t])&&1!==e.nodeType);return e}S.fn.extend({has:function(e){var t=S(e,this),n=t.length;return this.filter(function(){for(var e=0;e<n;e++)if(S.contains(this,t[e]))return!0})},closest:function(e,t){var n,r=0,i=this.length,o=[],a="string"!=typeof e&&S(e);if(!k.test(e))for(;r<i;r++)for(n=this[r];n&&n!==t;n=n.parentNode)if(n.nodeType<11&&(a?-1<a.index(n):1===n.nodeType&&S.find.matchesSelector(n,e))){o.push(n);break}return this.pushStack(1<o.length?S.uniqueSort(o):o)},index:function(e){return e?"string"==typeof e?i.call(S(e),this[0]):i.call(this,e.jquery?e[0]:e):this[0]&&this[0].parentNode?this.first().prevAll().length:-1},add:function(e,t){return this.pushStack(S.uniqueSort(S.merge(this.get(),S(e,t))))},addBack:function(e){return this.add(null==e?this.prevObject:this.prevObject.filter(e))}}),S.each({parent:function(e){var t=e.parentNode;return t&&11!==t.nodeType?t:null},parents:function(e){return h(e,"parentNode")},parentsUntil:function(e,t,n){return h(e,"parentNode",n)},next:function(e){return O(e,"nextSibling")},prev:function(e){return O(e,"previousSibling")},nextAll:function(e){return h(e,"nextSibling")},prevAll:function(e){return h(e,"previousSibling")},nextUntil:function(e,t,n){return h(e,"nextSibling",n)},prevUntil:function(e,t,n){return h(e,"previousSibling",n)},siblings:function(e){return T((e.parentNode||{}).firstChild,e)},children:function(e){return T(e.firstChild)},contents:function(e){return null!=e.contentDocument&&r(e.contentDocument)?e.contentDocument:(A(e,"template")&&(e=e.content||e),S.merge([],e.childNodes))}},function(r,i){S.fn[r]=function(e,t){var n=S.map(this,i,e);return"Until"!==r.slice(-5)&&(t=e),t&&"string"==typeof t&&(n=S.filter(t,n)),1<this.length&&(H[r]||S.uniqueSort(n),L.test(r)&&n.reverse()),this.pushStack(n)}});var P=/[^\x20\t\r\n\f]+/g;function R(e){return e}function M(e){throw e}function I(e,t,n,r){var i;try{e&&m(i=e.promise)?i.call(e).done(t).fail(n):e&&m(i=e.then)?i.call(e,t,n):t.apply(void 0,[e].slice(r))}catch(e){n.apply(void 0,[e])}}S.Callbacks=function(r){var e,n;r="string"==typeof r?(e=r,n={},S.each(e.match(P)||[],function(e,t){n[t]=!0}),n):S.extend({},r);var i,t,o,a,s=[],u=[],l=-1,c=function(){for(a=a||r.once,o=i=!0;u.length;l=-1){t=u.shift();while(++l<s.length)!1===s[l].apply(t[0],t[1])&&r.stopOnFalse&&(l=s.length,t=!1)}r.memory||(t=!1),i=!1,a&&(s=t?[]:"")},f={add:function(){return s&&(t&&!i&&(l=s.length-1,u.push(t)),function n(e){S.each(e,function(e,t){m(t)?r.unique&&f.has(t)||s.push(t):t&&t.length&&"string"!==w(t)&&n(t)})}(arguments),t&&!i&&c()),this},remove:function(){return S.each(arguments,function(e,t){var n;while(-1<(n=S.inArray(t,s,n)))s.splice(n,1),n<=l&&l--}),this},has:function(e){return e?-1<S.inArray(e,s):0<s.length},empty:function(){return s&&(s=[]),this},disable:function(){return a=u=[],s=t="",this},disabled:function(){return!s},lock:function(){return a=u=[],t||i||(s=t=""),this},locked:function(){return!!a},fireWith:function(e,t){return a||(t=[e,(t=t||[]).slice?t.slice():t],u.push(t),i||c()),this},fire:function(){return f.fireWith(this,arguments),this},fired:function(){return!!o}};return f},S.extend({Deferred:function(e){var o=[["notify","progress",S.Callbacks("memory"),S.Callbacks("memory"),2],["resolve","done",S.Callbacks("once memory"),S.Callbacks("once memory"),0,"resolved"],["reject","fail",S.Callbacks("once memory"),S.Callbacks("once memory"),1,"rejected"]],i="pending",a={state:function(){return i},always:function(){return s.done(arguments).fail(arguments),this},"catch":function(e){return a.then(null,e)},pipe:function(){var i=arguments;return S.Deferred(function(r){S.each(o,function(e,t){var n=m(i[t[4]])&&i[t[4]];s[t[1]](function(){var e=n&&n.apply(this,arguments);e&&m(e.promise)?e.promise().progress(r.notify).done(r.resolve).fail(r.reject):r[t[0]+"With"](this,n?[e]:arguments)})}),i=null}).promise()},then:function(t,n,r){var u=0;function l(i,o,a,s){return function(){var n=this,r=arguments,e=function(){var e,t;if(!(i<u)){if((e=a.apply(n,r))===o.promise())throw new TypeError("Thenable self-resolution");t=e&&("object"==typeof e||"function"==typeof e)&&e.then,m(t)?s?t.call(e,l(u,o,R,s),l(u,o,M,s)):(u++,t.call(e,l(u,o,R,s),l(u,o,M,s),l(u,o,R,o.notifyWith))):(a!==R&&(n=void 0,r=[e]),(s||o.resolveWith)(n,r))}},t=s?e:function(){try{e()}catch(e){S.Deferred.exceptionHook&&S.Deferred.exceptionHook(e,t.stackTrace),u<=i+1&&(a!==M&&(n=void 0,r=[e]),o.rejectWith(n,r))}};i?t():(S.Deferred.getStackHook&&(t.stackTrace=S.Deferred.getStackHook()),C.setTimeout(t))}}return S.Deferred(function(e){o[0][3].add(l(0,e,m(r)?r:R,e.notifyWith)),o[1][3].add(l(0,e,m(t)?t:R)),o[2][3].add(l(0,e,m(n)?n:M))}).promise()},promise:function(e){return null!=e?S.extend(e,a):a}},s={};return S.each(o,function(e,t){var n=t[2],r=t[5];a[t[1]]=n.add,r&&n.add(function(){i=r},o[3-e][2].disable,o[3-e][3].disable,o[0][2].lock,o[0][3].lock),n.add(t[3].fire),s[t[0]]=function(){return s[t[0]+"With"](this===s?void 0:this,arguments),this},s[t[0]+"With"]=n.fireWith}),a.promise(s),e&&e.call(s,s),s},when:function(e){var n=arguments.length,t=n,r=Array(t),i=s.call(arguments),o=S.Deferred(),a=function(t){return function(e){r[t]=this,i[t]=1<arguments.length?s.call(arguments):e,--n||o.resolveWith(r,i)}};if(n<=1&&(I(e,o.done(a(t)).resolve,o.reject,!n),"pending"===o.state()||m(i[t]&&i[t].then)))return o.then();while(t--)I(i[t],a(t),o.reject);return o.promise()}});var W=/^(Eval|Internal|Range|Reference|Syntax|Type|URI)Error$/;S.Deferred.exceptionHook=function(e,t){C.console&&C.console.warn&&e&&W.test(e.name)&&C.console.warn("jQuery.Deferred exception: "+e.message,e.stack,t)},S.readyException=function(e){C.setTimeout(function(){throw e})};var F=S.Deferred();function B(){E.removeEventListener("DOMContentLoaded",B),C.removeEventListener("load",B),S.ready()}S.fn.ready=function(e){return F.then(e)["catch"](function(e){S.readyException(e)}),this},S.extend({isReady:!1,readyWait:1,ready:function(e){(!0===e?--S.readyWait:S.isReady)||(S.isReady=!0)!==e&&0<--S.readyWait||F.resolveWith(E,[S])}}),S.ready.then=F.then,"complete"===E.readyState||"loading"!==E.readyState&&!E.documentElement.doScroll?C.setTimeout(S.ready):(E.addEventListener("DOMContentLoaded",B),C.addEventListener("load",B));var $=function(e,t,n,r,i,o,a){var s=0,u=e.length,l=null==n;if("object"===w(n))for(s in i=!0,n)$(e,t,s,n[s],!0,o,a);else if(void 0!==r&&(i=!0,m(r)||(a=!0),l&&(a?(t.call(e,r),t=null):(l=t,t=function(e,t,n){return l.call(S(e),n)})),t))for(;s<u;s++)t(e[s],n,a?r:r.call(e[s],s,t(e[s],n)));return i?e:l?t.call(e):u?t(e[0],n):o},_=/^-ms-/,z=/-([a-z])/g;function U(e,t){return t.toUpperCase()}function X(e){return e.replace(_,"ms-").replace(z,U)}var V=function(e){return 1===e.nodeType||9===e.nodeType||!+e.nodeType};function G(){this.expando=S.expando+G.uid++}G.uid=1,G.prototype={cache:function(e){var t=e[this.expando];return t||(t={},V(e)&&(e.nodeType?e[this.expando]=t:Object.defineProperty(e,this.expando,{value:t,configurable:!0}))),t},set:function(e,t,n){var r,i=this.cache(e);if("string"==typeof t)i[X(t)]=n;else for(r in t)i[X(r)]=t[r];return i},get:function(e,t){return void 0===t?this.cache(e):e[this.expando]&&e[this.expando][X(t)]},access:function(e,t,n){return void 0===t||t&&"string"==typeof t&&void 0===n?this.get(e,t):(this.set(e,t,n),void 0!==n?n:t)},remove:function(e,t){var n,r=e[this.expando];if(void 0!==r){if(void 0!==t){n=(t=Array.isArray(t)?t.map(X):(t=X(t))in r?[t]:t.match(P)||[]).length;while(n--)delete r[t[n]]}(void 0===t||S.isEmptyObject(r))&&(e.nodeType?e[this.expando]=void 0:delete e[this.expando])}},hasData:function(e){var t=e[this.expando];return void 0!==t&&!S.isEmptyObject(t)}};var Y=new G,Q=new G,J=/^(?:\{[\w\W]*\}|\[[\w\W]*\])$/,K=/[A-Z]/g;function Z(e,t,n){var r,i;if(void 0===n&&1===e.nodeType)if(r="data-"+t.replace(K,"-$&").toLowerCase(),"string"==typeof(n=e.getAttribute(r))){try{n="true"===(i=n)||"false"!==i&&("null"===i?null:i===+i+""?+i:J.test(i)?JSON.parse(i):i)}catch(e){}Q.set(e,t,n)}else n=void 0;return n}S.extend({hasData:function(e){return Q.hasData(e)||Y.hasData(e)},data:function(e,t,n){return Q.access(e,t,n)},removeData:function(e,t){Q.remove(e,t)},_data:function(e,t,n){return Y.access(e,t,n)},_removeData:function(e,t){Y.remove(e,t)}}),S.fn.extend({data:function(n,e){var t,r,i,o=this[0],a=o&&o.attributes;if(void 0===n){if(this.length&&(i=Q.get(o),1===o.nodeType&&!Y.get(o,"hasDataAttrs"))){t=a.length;while(t--)a[t]&&0===(r=a[t].name).indexOf("data-")&&(r=X(r.slice(5)),Z(o,r,i[r]));Y.set(o,"hasDataAttrs",!0)}return i}return"object"==typeof n?this.each(function(){Q.set(this,n)}):$(this,function(e){var t;if(o&&void 0===e)return void 0!==(t=Q.get(o,n))?t:void 0!==(t=Z(o,n))?t:void 0;this.each(function(){Q.set(this,n,e)})},null,e,1<arguments.length,null,!0)},removeData:function(e){return this.each(function(){Q.remove(this,e)})}}),S.extend({queue:function(e,t,n){var r;if(e)return t=(t||"fx")+"queue",r=Y.get(e,t),n&&(!r||Array.isArray(n)?r=Y.access(e,t,S.makeArray(n)):r.push(n)),r||[]},dequeue:function(e,t){t=t||"fx";var n=S.queue(e,t),r=n.length,i=n.shift(),o=S._queueHooks(e,t);"inprogress"===i&&(i=n.shift(),r--),i&&("fx"===t&&n.unshift("inprogress"),delete o.stop,i.call(e,function(){S.dequeue(e,t)},o)),!r&&o&&o.empty.fire()},_queueHooks:function(e,t){var n=t+"queueHooks";return Y.get(e,n)||Y.access(e,n,{empty:S.Callbacks("once memory").add(function(){Y.remove(e,[t+"queue",n])})})}}),S.fn.extend({queue:function(t,n){var e=2;return"string"!=typeof t&&(n=t,t="fx",e--),arguments.length<e?S.queue(this[0],t):void 0===n?this:this.each(function(){var e=S.queue(this,t,n);S._queueHooks(this,t),"fx"===t&&"inprogress"!==e[0]&&S.dequeue(this,t)})},dequeue:function(e){return this.each(function(){S.dequeue(this,e)})},clearQueue:function(e){return this.queue(e||"fx",[])},promise:function(e,t){var n,r=1,i=S.Deferred(),o=this,a=this.length,s=function(){--r||i.resolveWith(o,[o])};"string"!=typeof e&&(t=e,e=void 0),e=e||"fx";while(a--)(n=Y.get(o[a],e+"queueHooks"))&&n.empty&&(r++,n.empty.add(s));return s(),i.promise(t)}});var ee=/[+-]?(?:\d*\.|)\d+(?:[eE][+-]?\d+|)/.source,te=new RegExp("^(?:([+-])=|)("+ee+")([a-z%]*)$","i"),ne=["Top","Right","Bottom","Left"],re=E.documentElement,ie=function(e){return S.contains(e.ownerDocument,e)},oe={composed:!0};re.getRootNode&&(ie=function(e){return S.contains(e.ownerDocument,e)||e.getRootNode(oe)===e.ownerDocument});var ae=function(e,t){return"none"===(e=t||e).style.display||""===e.style.display&&ie(e)&&"none"===S.css(e,"display")};function se(e,t,n,r){var i,o,a=20,s=r?function(){return r.cur()}:function(){return S.css(e,t,"")},u=s(),l=n&&n[3]||(S.cssNumber[t]?"":"px"),c=e.nodeType&&(S.cssNumber[t]||"px"!==l&&+u)&&te.exec(S.css(e,t));if(c&&c[3]!==l){u/=2,l=l||c[3],c=+u||1;while(a--)S.style(e,t,c+l),(1-o)*(1-(o=s()/u||.5))<=0&&(a=0),c/=o;c*=2,S.style(e,t,c+l),n=n||[]}return n&&(c=+c||+u||0,i=n[1]?c+(n[1]+1)*n[2]:+n[2],r&&(r.unit=l,r.start=c,r.end=i)),i}var ue={};function le(e,t){for(var n,r,i,o,a,s,u,l=[],c=0,f=e.length;c<f;c++)(r=e[c]).style&&(n=r.style.display,t?("none"===n&&(l[c]=Y.get(r,"display")||null,l[c]||(r.style.display="")),""===r.style.display&&ae(r)&&(l[c]=(u=a=o=void 0,a=(i=r).ownerDocument,s=i.nodeName,(u=ue[s])||(o=a.body.appendChild(a.createElement(s)),u=S.css(o,"display"),o.parentNode.removeChild(o),"none"===u&&(u="block"),ue[s]=u)))):"none"!==n&&(l[c]="none",Y.set(r,"display",n)));for(c=0;c<f;c++)null!=l[c]&&(e[c].style.display=l[c]);return e}S.fn.extend({show:function(){return le(this,!0)},hide:function(){return le(this)},toggle:function(e){return"boolean"==typeof e?e?this.show():this.hide():this.each(function(){ae(this)?S(this).show():S(this).hide()})}});var ce,fe,pe=/^(?:checkbox|radio)$/i,de=/<([a-z][^\/\0>\x20\t\r\n\f]*)/i,he=/^$|^module$|\/(?:java|ecma)script/i;ce=E.createDocumentFragment().appendChild(E.createElement("div")),(fe=E.createElement("input")).setAttribute("type","radio"),fe.setAttribute("checked","checked"),fe.setAttribute("name","t"),ce.appendChild(fe),y.checkClone=ce.cloneNode(!0).cloneNode(!0).lastChild.checked,ce.innerHTML="<textarea>x</textarea>",y.noCloneChecked=!!ce.cloneNode(!0).lastChild.defaultValue,ce.innerHTML="<option></option>",y.option=!!ce.lastChild;var ge={thead:[1,"<table>","</table>"],col:[2,"<table><colgroup>","</colgroup></table>"],tr:[2,"<table><tbody>","</tbody></table>"],td:[3,"<table><tbody><tr>","</tr></tbody></table>"],_default:[0,"",""]};function ve(e,t){var n;return n="undefined"!=typeof e.getElementsByTagName?e.getElementsByTagName(t||"*"):"undefined"!=typeof e.querySelectorAll?e.querySelectorAll(t||"*"):[],void 0===t||t&&A(e,t)?S.merge([e],n):n}function ye(e,t){for(var n=0,r=e.length;n<r;n++)Y.set(e[n],"globalEval",!t||Y.get(t[n],"globalEval"))}ge.tbody=ge.tfoot=ge.colgroup=ge.caption=ge.thead,ge.th=ge.td,y.option||(ge.optgroup=ge.option=[1,"<select multiple='multiple'>","</select>"]);var me=/<|&#?\w+;/;function xe(e,t,n,r,i){for(var o,a,s,u,l,c,f=t.createDocumentFragment(),p=[],d=0,h=e.length;d<h;d++)if((o=e[d])||0===o)if("object"===w(o))S.merge(p,o.nodeType?[o]:o);else if(me.test(o)){a=a||f.appendChild(t.createElement("div")),s=(de.exec(o)||["",""])[1].toLowerCase(),u=ge[s]||ge._default,a.innerHTML=u[1]+S.htmlPrefilter(o)+u[2],c=u[0];while(c--)a=a.lastChild;S.merge(p,a.childNodes),(a=f.firstChild).textContent=""}else p.push(t.createTextNode(o));f.textContent="",d=0;while(o=p[d++])if(r&&-1<S.inArray(o,r))i&&i.push(o);else if(l=ie(o),a=ve(f.appendChild(o),"script"),l&&ye(a),n){c=0;while(o=a[c++])he.test(o.type||"")&&n.push(o)}return f}var be=/^key/,we=/^(?:mouse|pointer|contextmenu|drag|drop)|click/,Te=/^([^.]*)(?:\.(.+)|)/;function Ce(){return!0}function Ee(){return!1}function Se(e,t){return e===function(){try{return E.activeElement}catch(e){}}()==("focus"===t)}function ke(e,t,n,r,i,o){var a,s;if("object"==typeof t){for(s in"string"!=typeof n&&(r=r||n,n=void 0),t)ke(e,s,n,r,t[s],o);return e}if(null==r&&null==i?(i=n,r=n=void 0):null==i&&("string"==typeof n?(i=r,r=void 0):(i=r,r=n,n=void 0)),!1===i)i=Ee;else if(!i)return e;return 1===o&&(a=i,(i=function(e){return S().off(e),a.apply(this,arguments)}).guid=a.guid||(a.guid=S.guid++)),e.each(function(){S.event.add(this,t,i,r,n)})}function Ae(e,i,o){o?(Y.set(e,i,!1),S.event.add(e,i,{namespace:!1,handler:function(e){var t,n,r=Y.get(this,i);if(1&e.isTrigger&&this[i]){if(r.length)(S.event.special[i]||{}).delegateType&&e.stopPropagation();else if(r=s.call(arguments),Y.set(this,i,r),t=o(this,i),this[i](),r!==(n=Y.get(this,i))||t?Y.set(this,i,!1):n={},r!==n)return e.stopImmediatePropagation(),e.preventDefault(),n.value}else r.length&&(Y.set(this,i,{value:S.event.trigger(S.extend(r[0],S.Event.prototype),r.slice(1),this)}),e.stopImmediatePropagation())}})):void 0===Y.get(e,i)&&S.event.add(e,i,Ce)}S.event={global:{},add:function(t,e,n,r,i){var o,a,s,u,l,c,f,p,d,h,g,v=Y.get(t);if(V(t)){n.handler&&(n=(o=n).handler,i=o.selector),i&&S.find.matchesSelector(re,i),n.guid||(n.guid=S.guid++),(u=v.events)||(u=v.events=Object.create(null)),(a=v.handle)||(a=v.handle=function(e){return"undefined"!=typeof S&&S.event.triggered!==e.type?S.event.dispatch.apply(t,arguments):void 0}),l=(e=(e||"").match(P)||[""]).length;while(l--)d=g=(s=Te.exec(e[l])||[])[1],h=(s[2]||"").split(".").sort(),d&&(f=S.event.special[d]||{},d=(i?f.delegateType:f.bindType)||d,f=S.event.special[d]||{},c=S.extend({type:d,origType:g,data:r,handler:n,guid:n.guid,selector:i,needsContext:i&&S.expr.match.needsContext.test(i),namespace:h.join(".")},o),(p=u[d])||((p=u[d]=[]).delegateCount=0,f.setup&&!1!==f.setup.call(t,r,h,a)||t.addEventListener&&t.addEventListener(d,a)),f.add&&(f.add.call(t,c),c.handler.guid||(c.handler.guid=n.guid)),i?p.splice(p.delegateCount++,0,c):p.push(c),S.event.global[d]=!0)}},remove:function(e,t,n,r,i){var o,a,s,u,l,c,f,p,d,h,g,v=Y.hasData(e)&&Y.get(e);if(v&&(u=v.events)){l=(t=(t||"").match(P)||[""]).length;while(l--)if(d=g=(s=Te.exec(t[l])||[])[1],h=(s[2]||"").split(".").sort(),d){f=S.event.special[d]||{},p=u[d=(r?f.delegateType:f.bindType)||d]||[],s=s[2]&&new RegExp("(^|\\.)"+h.join("\\.(?:.*\\.|)")+"(\\.|$)"),a=o=p.length;while(o--)c=p[o],!i&&g!==c.origType||n&&n.guid!==c.guid||s&&!s.test(c.namespace)||r&&r!==c.selector&&("**"!==r||!c.selector)||(p.splice(o,1),c.selector&&p.delegateCount--,f.remove&&f.remove.call(e,c));a&&!p.length&&(f.teardown&&!1!==f.teardown.call(e,h,v.handle)||S.removeEvent(e,d,v.handle),delete u[d])}else for(d in u)S.event.remove(e,d+t[l],n,r,!0);S.isEmptyObject(u)&&Y.remove(e,"handle events")}},dispatch:function(e){var t,n,r,i,o,a,s=new Array(arguments.length),u=S.event.fix(e),l=(Y.get(this,"events")||Object.create(null))[u.type]||[],c=S.event.special[u.type]||{};for(s[0]=u,t=1;t<arguments.length;t++)s[t]=arguments[t];if(u.delegateTarget=this,!c.preDispatch||!1!==c.preDispatch.call(this,u)){a=S.event.handlers.call(this,u,l),t=0;while((i=a[t++])&&!u.isPropagationStopped()){u.currentTarget=i.elem,n=0;while((o=i.handlers[n++])&&!u.isImmediatePropagationStopped())u.rnamespace&&!1!==o.namespace&&!u.rnamespace.test(o.namespace)||(u.handleObj=o,u.data=o.data,void 0!==(r=((S.event.special[o.origType]||{}).handle||o.handler).apply(i.elem,s))&&!1===(u.result=r)&&(u.preventDefault(),u.stopPropagation()))}return c.postDispatch&&c.postDispatch.call(this,u),u.result}},handlers:function(e,t){var n,r,i,o,a,s=[],u=t.delegateCount,l=e.target;if(u&&l.nodeType&&!("click"===e.type&&1<=e.button))for(;l!==this;l=l.parentNode||this)if(1===l.nodeType&&("click"!==e.type||!0!==l.disabled)){for(o=[],a={},n=0;n<u;n++)void 0===a[i=(r=t[n]).selector+" "]&&(a[i]=r.needsContext?-1<S(i,this).index(l):S.find(i,this,null,[l]).length),a[i]&&o.push(r);o.length&&s.push({elem:l,handlers:o})}return l=this,u<t.length&&s.push({elem:l,handlers:t.slice(u)}),s},addProp:function(t,e){Object.defineProperty(S.Event.prototype,t,{enumerable:!0,configurable:!0,get:m(e)?function(){if(this.originalEvent)return e(this.originalEvent)}:function(){if(this.originalEvent)return this.originalEvent[t]},set:function(e){Object.defineProperty(this,t,{enumerable:!0,configurable:!0,writable:!0,value:e})}})},fix:function(e){return e[S.expando]?e:new S.Event(e)},special:{load:{noBubble:!0},click:{setup:function(e){var t=this||e;return pe.test(t.type)&&t.click&&A(t,"input")&&Ae(t,"click",Ce),!1},trigger:function(e){var t=this||e;return pe.test(t.type)&&t.click&&A(t,"input")&&Ae(t,"click"),!0},_default:function(e){var t=e.target;return pe.test(t.type)&&t.click&&A(t,"input")&&Y.get(t,"click")||A(t,"a")}},beforeunload:{postDispatch:function(e){void 0!==e.result&&e.originalEvent&&(e.originalEvent.returnValue=e.result)}}}},S.removeEvent=function(e,t,n){e.removeEventListener&&e.removeEventListener(t,n)},S.Event=function(e,t){if(!(this instanceof S.Event))return new S.Event(e,t);e&&e.type?(this.originalEvent=e,this.type=e.type,this.isDefaultPrevented=e.defaultPrevented||void 0===e.defaultPrevented&&!1===e.returnValue?Ce:Ee,this.target=e.target&&3===e.target.nodeType?e.target.parentNode:e.target,this.currentTarget=e.currentTarget,this.relatedTarget=e.relatedTarget):this.type=e,t&&S.extend(this,t),this.timeStamp=e&&e.timeStamp||Date.now(),this[S.expando]=!0},S.Event.prototype={constructor:S.Event,isDefaultPrevented:Ee,isPropagationStopped:Ee,isImmediatePropagationStopped:Ee,isSimulated:!1,preventDefault:function(){var e=this.originalEvent;this.isDefaultPrevented=Ce,e&&!this.isSimulated&&e.preventDefault()},stopPropagation:function(){var e=this.originalEvent;this.isPropagationStopped=Ce,e&&!this.isSimulated&&e.stopPropagation()},stopImmediatePropagation:function(){var e=this.originalEvent;this.isImmediatePropagationStopped=Ce,e&&!this.isSimulated&&e.stopImmediatePropagation(),this.stopPropagation()}},S.each({altKey:!0,bubbles:!0,cancelable:!0,changedTouches:!0,ctrlKey:!0,detail:!0,eventPhase:!0,metaKey:!0,pageX:!0,pageY:!0,shiftKey:!0,view:!0,"char":!0,code:!0,charCode:!0,key:!0,keyCode:!0,button:!0,buttons:!0,clientX:!0,clientY:!0,offsetX:!0,offsetY:!0,pointerId:!0,pointerType:!0,screenX:!0,screenY:!0,targetTouches:!0,toElement:!0,touches:!0,which:function(e){var t=e.button;return null==e.which&&be.test(e.type)?null!=e.charCode?e.charCode:e.keyCode:!e.which&&void 0!==t&&we.test(e.type)?1&t?1:2&t?3:4&t?2:0:e.which}},S.event.addProp),S.each({focus:"focusin",blur:"focusout"},function(e,t){S.event.special[e]={setup:function(){return Ae(this,e,Se),!1},trigger:function(){return Ae(this,e),!0},delegateType:t}}),S.each({mouseenter:"mouseover",mouseleave:"mouseout",pointerenter:"pointerover",pointerleave:"pointerout"},function(e,i){S.event.special[e]={delegateType:i,bindType:i,handle:function(e){var t,n=e.relatedTarget,r=e.handleObj;return n&&(n===this||S.contains(this,n))||(e.type=r.origType,t=r.handler.apply(this,arguments),e.type=i),t}}}),S.fn.extend({on:function(e,t,n,r){return ke(this,e,t,n,r)},one:function(e,t,n,r){return ke(this,e,t,n,r,1)},off:function(e,t,n){var r,i;if(e&&e.preventDefault&&e.handleObj)return r=e.handleObj,S(e.delegateTarget).off(r.namespace?r.origType+"."+r.namespace:r.origType,r.selector,r.handler),this;if("object"==typeof e){for(i in e)this.off(i,t,e[i]);return this}return!1!==t&&"function"!=typeof t||(n=t,t=void 0),!1===n&&(n=Ee),this.each(function(){S.event.remove(this,e,n,t)})}});var Ne=/<script|<style|<link/i,De=/checked\s*(?:[^=]|=\s*.checked.)/i,je=/^\s*<!(?:\[CDATA\[|--)|(?:\]\]|--)>\s*$/g;function qe(e,t){return A(e,"table")&&A(11!==t.nodeType?t:t.firstChild,"tr")&&S(e).children("tbody")[0]||e}function Le(e){return e.type=(null!==e.getAttribute("type"))+"/"+e.type,e}function He(e){return"true/"===(e.type||"").slice(0,5)?e.type=e.type.slice(5):e.removeAttribute("type"),e}function Oe(e,t){var n,r,i,o,a,s;if(1===t.nodeType){if(Y.hasData(e)&&(s=Y.get(e).events))for(i in Y.remove(t,"handle events"),s)for(n=0,r=s[i].length;n<r;n++)S.event.add(t,i,s[i][n]);Q.hasData(e)&&(o=Q.access(e),a=S.extend({},o),Q.set(t,a))}}function Pe(n,r,i,o){r=g(r);var e,t,a,s,u,l,c=0,f=n.length,p=f-1,d=r[0],h=m(d);if(h||1<f&&"string"==typeof d&&!y.checkClone&&De.test(d))return n.each(function(e){var t=n.eq(e);h&&(r[0]=d.call(this,e,t.html())),Pe(t,r,i,o)});if(f&&(t=(e=xe(r,n[0].ownerDocument,!1,n,o)).firstChild,1===e.childNodes.length&&(e=t),t||o)){for(s=(a=S.map(ve(e,"script"),Le)).length;c<f;c++)u=e,c!==p&&(u=S.clone(u,!0,!0),s&&S.merge(a,ve(u,"script"))),i.call(n[c],u,c);if(s)for(l=a[a.length-1].ownerDocument,S.map(a,He),c=0;c<s;c++)u=a[c],he.test(u.type||"")&&!Y.access(u,"globalEval")&&S.contains(l,u)&&(u.src&&"module"!==(u.type||"").toLowerCase()?S._evalUrl&&!u.noModule&&S._evalUrl(u.src,{nonce:u.nonce||u.getAttribute("nonce")},l):b(u.textContent.replace(je,""),u,l))}return n}function Re(e,t,n){for(var r,i=t?S.filter(t,e):e,o=0;null!=(r=i[o]);o++)n||1!==r.nodeType||S.cleanData(ve(r)),r.parentNode&&(n&&ie(r)&&ye(ve(r,"script")),r.parentNode.removeChild(r));return e}S.extend({htmlPrefilter:function(e){return e},clone:function(e,t,n){var r,i,o,a,s,u,l,c=e.cloneNode(!0),f=ie(e);if(!(y.noCloneChecked||1!==e.nodeType&&11!==e.nodeType||S.isXMLDoc(e)))for(a=ve(c),r=0,i=(o=ve(e)).length;r<i;r++)s=o[r],u=a[r],void 0,"input"===(l=u.nodeName.toLowerCase())&&pe.test(s.type)?u.checked=s.checked:"input"!==l&&"textarea"!==l||(u.defaultValue=s.defaultValue);if(t)if(n)for(o=o||ve(e),a=a||ve(c),r=0,i=o.length;r<i;r++)Oe(o[r],a[r]);else Oe(e,c);return 0<(a=ve(c,"script")).length&&ye(a,!f&&ve(e,"script")),c},cleanData:function(e){for(var t,n,r,i=S.event.special,o=0;void 0!==(n=e[o]);o++)if(V(n)){if(t=n[Y.expando]){if(t.events)for(r in t.events)i[r]?S.event.remove(n,r):S.removeEvent(n,r,t.handle);n[Y.expando]=void 0}n[Q.expando]&&(n[Q.expando]=void 0)}}}),S.fn.extend({detach:function(e){return Re(this,e,!0)},remove:function(e){return Re(this,e)},text:function(e){return $(this,function(e){return void 0===e?S.text(this):this.empty().each(function(){1!==this.nodeType&&11!==this.nodeType&&9!==this.nodeType||(this.textContent=e)})},null,e,arguments.length)},append:function(){return Pe(this,arguments,function(e){1!==this.nodeType&&11!==this.nodeType&&9!==this.nodeType||qe(this,e).appendChild(e)})},prepend:function(){return Pe(this,arguments,function(e){if(1===this.nodeType||11===this.nodeType||9===this.nodeType){var t=qe(this,e);t.insertBefore(e,t.firstChild)}})},before:function(){return Pe(this,arguments,function(e){this.parentNode&&this.parentNode.insertBefore(e,this)})},after:function(){return Pe(this,arguments,function(e){this.parentNode&&this.parentNode.insertBefore(e,this.nextSibling)})},empty:function(){for(var e,t=0;null!=(e=this[t]);t++)1===e.nodeType&&(S.cleanData(ve(e,!1)),e.textContent="");return this},clone:function(e,t){return e=null!=e&&e,t=null==t?e:t,this.map(function(){return S.clone(this,e,t)})},html:function(e){return $(this,function(e){var t=this[0]||{},n=0,r=this.length;if(void 0===e&&1===t.nodeType)return t.innerHTML;if("string"==typeof e&&!Ne.test(e)&&!ge[(de.exec(e)||["",""])[1].toLowerCase()]){e=S.htmlPrefilter(e);try{for(;n<r;n++)1===(t=this[n]||{}).nodeType&&(S.cleanData(ve(t,!1)),t.innerHTML=e);t=0}catch(e){}}t&&this.empty().append(e)},null,e,arguments.length)},replaceWith:function(){var n=[];return Pe(this,arguments,function(e){var t=this.parentNode;S.inArray(this,n)<0&&(S.cleanData(ve(this)),t&&t.replaceChild(e,this))},n)}}),S.each({appendTo:"append",prependTo:"prepend",insertBefore:"before",insertAfter:"after",replaceAll:"replaceWith"},function(e,a){S.fn[e]=function(e){for(var t,n=[],r=S(e),i=r.length-1,o=0;o<=i;o++)t=o===i?this:this.clone(!0),S(r[o])[a](t),u.apply(n,t.get());return this.pushStack(n)}});var Me=new RegExp("^("+ee+")(?!px)[a-z%]+$","i"),Ie=function(e){var t=e.ownerDocument.defaultView;return t&&t.opener||(t=C),t.getComputedStyle(e)},We=function(e,t,n){var r,i,o={};for(i in t)o[i]=e.style[i],e.style[i]=t[i];for(i in r=n.call(e),t)e.style[i]=o[i];return r},Fe=new RegExp(ne.join("|"),"i");function Be(e,t,n){var r,i,o,a,s=e.style;return(n=n||Ie(e))&&(""!==(a=n.getPropertyValue(t)||n[t])||ie(e)||(a=S.style(e,t)),!y.pixelBoxStyles()&&Me.test(a)&&Fe.test(t)&&(r=s.width,i=s.minWidth,o=s.maxWidth,s.minWidth=s.maxWidth=s.width=a,a=n.width,s.width=r,s.minWidth=i,s.maxWidth=o)),void 0!==a?a+"":a}function $e(e,t){return{get:function(){if(!e())return(this.get=t).apply(this,arguments);delete this.get}}}!function(){function e(){if(l){u.style.cssText="position:absolute;left:-11111px;width:60px;margin-top:1px;padding:0;border:0",l.style.cssText="position:relative;display:block;box-sizing:border-box;overflow:scroll;margin:auto;border:1px;padding:1px;width:60%;top:1%",re.appendChild(u).appendChild(l);var e=C.getComputedStyle(l);n="1%"!==e.top,s=12===t(e.marginLeft),l.style.right="60%",o=36===t(e.right),r=36===t(e.width),l.style.position="absolute",i=12===t(l.offsetWidth/3),re.removeChild(u),l=null}}function t(e){return Math.round(parseFloat(e))}var n,r,i,o,a,s,u=E.createElement("div"),l=E.createElement("div");l.style&&(l.style.backgroundClip="content-box",l.cloneNode(!0).style.backgroundClip="",y.clearCloneStyle="content-box"===l.style.backgroundClip,S.extend(y,{boxSizingReliable:function(){return e(),r},pixelBoxStyles:function(){return e(),o},pixelPosition:function(){return e(),n},reliableMarginLeft:function(){return e(),s},scrollboxSize:function(){return e(),i},reliableTrDimensions:function(){var e,t,n,r;return null==a&&(e=E.createElement("table"),t=E.createElement("tr"),n=E.createElement("div"),e.style.cssText="position:absolute;left:-11111px",t.style.height="1px",n.style.height="9px",re.appendChild(e).appendChild(t).appendChild(n),r=C.getComputedStyle(t),a=3<parseInt(r.height),re.removeChild(e)),a}}))}();var _e=["Webkit","Moz","ms"],ze=E.createElement("div").style,Ue={};function Xe(e){var t=S.cssProps[e]||Ue[e];return t||(e in ze?e:Ue[e]=function(e){var t=e[0].toUpperCase()+e.slice(1),n=_e.length;while(n--)if((e=_e[n]+t)in ze)return e}(e)||e)}var Ve=/^(none|table(?!-c[ea]).+)/,Ge=/^--/,Ye={position:"absolute",visibility:"hidden",display:"block"},Qe={letterSpacing:"0",fontWeight:"400"};function Je(e,t,n){var r=te.exec(t);return r?Math.max(0,r[2]-(n||0))+(r[3]||"px"):t}function Ke(e,t,n,r,i,o){var a="width"===t?1:0,s=0,u=0;if(n===(r?"border":"content"))return 0;for(;a<4;a+=2)"margin"===n&&(u+=S.css(e,n+ne[a],!0,i)),r?("content"===n&&(u-=S.css(e,"padding"+ne[a],!0,i)),"margin"!==n&&(u-=S.css(e,"border"+ne[a]+"Width",!0,i))):(u+=S.css(e,"padding"+ne[a],!0,i),"padding"!==n?u+=S.css(e,"border"+ne[a]+"Width",!0,i):s+=S.css(e,"border"+ne[a]+"Width",!0,i));return!r&&0<=o&&(u+=Math.max(0,Math.ceil(e["offset"+t[0].toUpperCase()+t.slice(1)]-o-u-s-.5))||0),u}function Ze(e,t,n){var r=Ie(e),i=(!y.boxSizingReliable()||n)&&"border-box"===S.css(e,"boxSizing",!1,r),o=i,a=Be(e,t,r),s="offset"+t[0].toUpperCase()+t.slice(1);if(Me.test(a)){if(!n)return a;a="auto"}return(!y.boxSizingReliable()&&i||!y.reliableTrDimensions()&&A(e,"tr")||"auto"===a||!parseFloat(a)&&"inline"===S.css(e,"display",!1,r))&&e.getClientRects().length&&(i="border-box"===S.css(e,"boxSizing",!1,r),(o=s in e)&&(a=e[s])),(a=parseFloat(a)||0)+Ke(e,t,n||(i?"border":"content"),o,r,a)+"px"}function et(e,t,n,r,i){return new et.prototype.init(e,t,n,r,i)}S.extend({cssHooks:{opacity:{get:function(e,t){if(t){var n=Be(e,"opacity");return""===n?"1":n}}}},cssNumber:{animationIterationCount:!0,columnCount:!0,fillOpacity:!0,flexGrow:!0,flexShrink:!0,fontWeight:!0,gridArea:!0,gridColumn:!0,gridColumnEnd:!0,gridColumnStart:!0,gridRow:!0,gridRowEnd:!0,gridRowStart:!0,lineHeight:!0,opacity:!0,order:!0,orphans:!0,widows:!0,zIndex:!0,zoom:!0},cssProps:{},style:function(e,t,n,r){if(e&&3!==e.nodeType&&8!==e.nodeType&&e.style){var i,o,a,s=X(t),u=Ge.test(t),l=e.style;if(u||(t=Xe(s)),a=S.cssHooks[t]||S.cssHooks[s],void 0===n)return a&&"get"in a&&void 0!==(i=a.get(e,!1,r))?i:l[t];"string"===(o=typeof n)&&(i=te.exec(n))&&i[1]&&(n=se(e,t,i),o="number"),null!=n&&n==n&&("number"!==o||u||(n+=i&&i[3]||(S.cssNumber[s]?"":"px")),y.clearCloneStyle||""!==n||0!==t.indexOf("background")||(l[t]="inherit"),a&&"set"in a&&void 0===(n=a.set(e,n,r))||(u?l.setProperty(t,n):l[t]=n))}},css:function(e,t,n,r){var i,o,a,s=X(t);return Ge.test(t)||(t=Xe(s)),(a=S.cssHooks[t]||S.cssHooks[s])&&"get"in a&&(i=a.get(e,!0,n)),void 0===i&&(i=Be(e,t,r)),"normal"===i&&t in Qe&&(i=Qe[t]),""===n||n?(o=parseFloat(i),!0===n||isFinite(o)?o||0:i):i}}),S.each(["height","width"],function(e,u){S.cssHooks[u]={get:function(e,t,n){if(t)return!Ve.test(S.css(e,"display"))||e.getClientRects().length&&e.getBoundingClientRect().width?Ze(e,u,n):We(e,Ye,function(){return Ze(e,u,n)})},set:function(e,t,n){var r,i=Ie(e),o=!y.scrollboxSize()&&"absolute"===i.position,a=(o||n)&&"border-box"===S.css(e,"boxSizing",!1,i),s=n?Ke(e,u,n,a,i):0;return a&&o&&(s-=Math.ceil(e["offset"+u[0].toUpperCase()+u.slice(1)]-parseFloat(i[u])-Ke(e,u,"border",!1,i)-.5)),s&&(r=te.exec(t))&&"px"!==(r[3]||"px")&&(e.style[u]=t,t=S.css(e,u)),Je(0,t,s)}}}),S.cssHooks.marginLeft=$e(y.reliableMarginLeft,function(e,t){if(t)return(parseFloat(Be(e,"marginLeft"))||e.getBoundingClientRect().left-We(e,{marginLeft:0},function(){return e.getBoundingClientRect().left}))+"px"}),S.each({margin:"",padding:"",border:"Width"},function(i,o){S.cssHooks[i+o]={expand:function(e){for(var t=0,n={},r="string"==typeof e?e.split(" "):[e];t<4;t++)n[i+ne[t]+o]=r[t]||r[t-2]||r[0];return n}},"margin"!==i&&(S.cssHooks[i+o].set=Je)}),S.fn.extend({css:function(e,t){return $(this,function(e,t,n){var r,i,o={},a=0;if(Array.isArray(t)){for(r=Ie(e),i=t.length;a<i;a++)o[t[a]]=S.css(e,t[a],!1,r);return o}return void 0!==n?S.style(e,t,n):S.css(e,t)},e,t,1<arguments.length)}}),((S.Tween=et).prototype={constructor:et,init:function(e,t,n,r,i,o){this.elem=e,this.prop=n,this.easing=i||S.easing._default,this.options=t,this.start=this.now=this.cur(),this.end=r,this.unit=o||(S.cssNumber[n]?"":"px")},cur:function(){var e=et.propHooks[this.prop];return e&&e.get?e.get(this):et.propHooks._default.get(this)},run:function(e){var t,n=et.propHooks[this.prop];return this.options.duration?this.pos=t=S.easing[this.easing](e,this.options.duration*e,0,1,this.options.duration):this.pos=t=e,this.now=(this.end-this.start)*t+this.start,this.options.step&&this.options.step.call(this.elem,this.now,this),n&&n.set?n.set(this):et.propHooks._default.set(this),this}}).init.prototype=et.prototype,(et.propHooks={_default:{get:function(e){var t;return 1!==e.elem.nodeType||null!=e.elem[e.prop]&&null==e.elem.style[e.prop]?e.elem[e.prop]:(t=S.css(e.elem,e.prop,""))&&"auto"!==t?t:0},set:function(e){S.fx.step[e.prop]?S.fx.step[e.prop](e):1!==e.elem.nodeType||!S.cssHooks[e.prop]&&null==e.elem.style[Xe(e.prop)]?e.elem[e.prop]=e.now:S.style(e.elem,e.prop,e.now+e.unit)}}}).scrollTop=et.propHooks.scrollLeft={set:function(e){e.elem.nodeType&&e.elem.parentNode&&(e.elem[e.prop]=e.now)}},S.easing={linear:function(e){return e},swing:function(e){return.5-Math.cos(e*Math.PI)/2},_default:"swing"},S.fx=et.prototype.init,S.fx.step={};var tt,nt,rt,it,ot=/^(?:toggle|show|hide)$/,at=/queueHooks$/;function st(){nt&&(!1===E.hidden&&C.requestAnimationFrame?C.requestAnimationFrame(st):C.setTimeout(st,S.fx.interval),S.fx.tick())}function ut(){return C.setTimeout(function(){tt=void 0}),tt=Date.now()}function lt(e,t){var n,r=0,i={height:e};for(t=t?1:0;r<4;r+=2-t)i["margin"+(n=ne[r])]=i["padding"+n]=e;return t&&(i.opacity=i.width=e),i}function ct(e,t,n){for(var r,i=(ft.tweeners[t]||[]).concat(ft.tweeners["*"]),o=0,a=i.length;o<a;o++)if(r=i[o].call(n,t,e))return r}function ft(o,e,t){var n,a,r=0,i=ft.prefilters.length,s=S.Deferred().always(function(){delete u.elem}),u=function(){if(a)return!1;for(var e=tt||ut(),t=Math.max(0,l.startTime+l.duration-e),n=1-(t/l.duration||0),r=0,i=l.tweens.length;r<i;r++)l.tweens[r].run(n);return s.notifyWith(o,[l,n,t]),n<1&&i?t:(i||s.notifyWith(o,[l,1,0]),s.resolveWith(o,[l]),!1)},l=s.promise({elem:o,props:S.extend({},e),opts:S.extend(!0,{specialEasing:{},easing:S.easing._default},t),originalProperties:e,originalOptions:t,startTime:tt||ut(),duration:t.duration,tweens:[],createTween:function(e,t){var n=S.Tween(o,l.opts,e,t,l.opts.specialEasing[e]||l.opts.easing);return l.tweens.push(n),n},stop:function(e){var t=0,n=e?l.tweens.length:0;if(a)return this;for(a=!0;t<n;t++)l.tweens[t].run(1);return e?(s.notifyWith(o,[l,1,0]),s.resolveWith(o,[l,e])):s.rejectWith(o,[l,e]),this}}),c=l.props;for(!function(e,t){var n,r,i,o,a;for(n in e)if(i=t[r=X(n)],o=e[n],Array.isArray(o)&&(i=o[1],o=e[n]=o[0]),n!==r&&(e[r]=o,delete e[n]),(a=S.cssHooks[r])&&"expand"in a)for(n in o=a.expand(o),delete e[r],o)n in e||(e[n]=o[n],t[n]=i);else t[r]=i}(c,l.opts.specialEasing);r<i;r++)if(n=ft.prefilters[r].call(l,o,c,l.opts))return m(n.stop)&&(S._queueHooks(l.elem,l.opts.queue).stop=n.stop.bind(n)),n;return S.map(c,ct,l),m(l.opts.start)&&l.opts.start.call(o,l),l.progress(l.opts.progress).done(l.opts.done,l.opts.complete).fail(l.opts.fail).always(l.opts.always),S.fx.timer(S.extend(u,{elem:o,anim:l,queue:l.opts.queue})),l}S.Animation=S.extend(ft,{tweeners:{"*":[function(e,t){var n=this.createTween(e,t);return se(n.elem,e,te.exec(t),n),n}]},tweener:function(e,t){m(e)?(t=e,e=["*"]):e=e.match(P);for(var n,r=0,i=e.length;r<i;r++)n=e[r],ft.tweeners[n]=ft.tweeners[n]||[],ft.tweeners[n].unshift(t)},prefilters:[function(e,t,n){var r,i,o,a,s,u,l,c,f="width"in t||"height"in t,p=this,d={},h=e.style,g=e.nodeType&&ae(e),v=Y.get(e,"fxshow");for(r in n.queue||(null==(a=S._queueHooks(e,"fx")).unqueued&&(a.unqueued=0,s=a.empty.fire,a.empty.fire=function(){a.unqueued||s()}),a.unqueued++,p.always(function(){p.always(function(){a.unqueued--,S.queue(e,"fx").length||a.empty.fire()})})),t)if(i=t[r],ot.test(i)){if(delete t[r],o=o||"toggle"===i,i===(g?"hide":"show")){if("show"!==i||!v||void 0===v[r])continue;g=!0}d[r]=v&&v[r]||S.style(e,r)}if((u=!S.isEmptyObject(t))||!S.isEmptyObject(d))for(r in f&&1===e.nodeType&&(n.overflow=[h.overflow,h.overflowX,h.overflowY],null==(l=v&&v.display)&&(l=Y.get(e,"display")),"none"===(c=S.css(e,"display"))&&(l?c=l:(le([e],!0),l=e.style.display||l,c=S.css(e,"display"),le([e]))),("inline"===c||"inline-block"===c&&null!=l)&&"none"===S.css(e,"float")&&(u||(p.done(function(){h.display=l}),null==l&&(c=h.display,l="none"===c?"":c)),h.display="inline-block")),n.overflow&&(h.overflow="hidden",p.always(function(){h.overflow=n.overflow[0],h.overflowX=n.overflow[1],h.overflowY=n.overflow[2]})),u=!1,d)u||(v?"hidden"in v&&(g=v.hidden):v=Y.access(e,"fxshow",{display:l}),o&&(v.hidden=!g),g&&le([e],!0),p.done(function(){for(r in g||le([e]),Y.remove(e,"fxshow"),d)S.style(e,r,d[r])})),u=ct(g?v[r]:0,r,p),r in v||(v[r]=u.start,g&&(u.end=u.start,u.start=0))}],prefilter:function(e,t){t?ft.prefilters.unshift(e):ft.prefilters.push(e)}}),S.speed=function(e,t,n){var r=e&&"object"==typeof e?S.extend({},e):{complete:n||!n&&t||m(e)&&e,duration:e,easing:n&&t||t&&!m(t)&&t};return S.fx.off?r.duration=0:"number"!=typeof r.duration&&(r.duration in S.fx.speeds?r.duration=S.fx.speeds[r.duration]:r.duration=S.fx.speeds._default),null!=r.queue&&!0!==r.queue||(r.queue="fx"),r.old=r.complete,r.complete=function(){m(r.old)&&r.old.call(this),r.queue&&S.dequeue(this,r.queue)},r},S.fn.extend({fadeTo:function(e,t,n,r){return this.filter(ae).css("opacity",0).show().end().animate({opacity:t},e,n,r)},animate:function(t,e,n,r){var i=S.isEmptyObject(t),o=S.speed(e,n,r),a=function(){var e=ft(this,S.extend({},t),o);(i||Y.get(this,"finish"))&&e.stop(!0)};return a.finish=a,i||!1===o.queue?this.each(a):this.queue(o.queue,a)},stop:function(i,e,o){var a=function(e){var t=e.stop;delete e.stop,t(o)};return"string"!=typeof i&&(o=e,e=i,i=void 0),e&&this.queue(i||"fx",[]),this.each(function(){var e=!0,t=null!=i&&i+"queueHooks",n=S.timers,r=Y.get(this);if(t)r[t]&&r[t].stop&&a(r[t]);else for(t in r)r[t]&&r[t].stop&&at.test(t)&&a(r[t]);for(t=n.length;t--;)n[t].elem!==this||null!=i&&n[t].queue!==i||(n[t].anim.stop(o),e=!1,n.splice(t,1));!e&&o||S.dequeue(this,i)})},finish:function(a){return!1!==a&&(a=a||"fx"),this.each(function(){var e,t=Y.get(this),n=t[a+"queue"],r=t[a+"queueHooks"],i=S.timers,o=n?n.length:0;for(t.finish=!0,S.queue(this,a,[]),r&&r.stop&&r.stop.call(this,!0),e=i.length;e--;)i[e].elem===this&&i[e].queue===a&&(i[e].anim.stop(!0),i.splice(e,1));for(e=0;e<o;e++)n[e]&&n[e].finish&&n[e].finish.call(this);delete t.finish})}}),S.each(["toggle","show","hide"],function(e,r){var i=S.fn[r];S.fn[r]=function(e,t,n){return null==e||"boolean"==typeof e?i.apply(this,arguments):this.animate(lt(r,!0),e,t,n)}}),S.each({slideDown:lt("show"),slideUp:lt("hide"),slideToggle:lt("toggle"),fadeIn:{opacity:"show"},fadeOut:{opacity:"hide"},fadeToggle:{opacity:"toggle"}},function(e,r){S.fn[e]=function(e,t,n){return this.animate(r,e,t,n)}}),S.timers=[],S.fx.tick=function(){var e,t=0,n=S.timers;for(tt=Date.now();t<n.length;t++)(e=n[t])()||n[t]!==e||n.splice(t--,1);n.length||S.fx.stop(),tt=void 0},S.fx.timer=function(e){S.timers.push(e),S.fx.start()},S.fx.interval=13,S.fx.start=function(){nt||(nt=!0,st())},S.fx.stop=function(){nt=null},S.fx.speeds={slow:600,fast:200,_default:400},S.fn.delay=function(r,e){return r=S.fx&&S.fx.speeds[r]||r,e=e||"fx",this.queue(e,function(e,t){var n=C.setTimeout(e,r);t.stop=function(){C.clearTimeout(n)}})},rt=E.createElement("input"),it=E.createElement("select").appendChild(E.createElement("option")),rt.type="checkbox",y.checkOn=""!==rt.value,y.optSelected=it.selected,(rt=E.createElement("input")).value="t",rt.type="radio",y.radioValue="t"===rt.value;var pt,dt=S.expr.attrHandle;S.fn.extend({attr:function(e,t){return $(this,S.attr,e,t,1<arguments.length)},removeAttr:function(e){return this.each(function(){S.removeAttr(this,e)})}}),S.extend({attr:function(e,t,n){var r,i,o=e.nodeType;if(3!==o&&8!==o&&2!==o)return"undefined"==typeof e.getAttribute?S.prop(e,t,n):(1===o&&S.isXMLDoc(e)||(i=S.attrHooks[t.toLowerCase()]||(S.expr.match.bool.test(t)?pt:void 0)),void 0!==n?null===n?void S.removeAttr(e,t):i&&"set"in i&&void 0!==(r=i.set(e,n,t))?r:(e.setAttribute(t,n+""),n):i&&"get"in i&&null!==(r=i.get(e,t))?r:null==(r=S.find.attr(e,t))?void 0:r)},attrHooks:{type:{set:function(e,t){if(!y.radioValue&&"radio"===t&&A(e,"input")){var n=e.value;return e.setAttribute("type",t),n&&(e.value=n),t}}}},removeAttr:function(e,t){var n,r=0,i=t&&t.match(P);if(i&&1===e.nodeType)while(n=i[r++])e.removeAttribute(n)}}),pt={set:function(e,t,n){return!1===t?S.removeAttr(e,n):e.setAttribute(n,n),n}},S.each(S.expr.match.bool.source.match(/\w+/g),function(e,t){var a=dt[t]||S.find.attr;dt[t]=function(e,t,n){var r,i,o=t.toLowerCase();return n||(i=dt[o],dt[o]=r,r=null!=a(e,t,n)?o:null,dt[o]=i),r}});var ht=/^(?:input|select|textarea|button)$/i,gt=/^(?:a|area)$/i;function vt(e){return(e.match(P)||[]).join(" ")}function yt(e){return e.getAttribute&&e.getAttribute("class")||""}function mt(e){return Array.isArray(e)?e:"string"==typeof e&&e.match(P)||[]}S.fn.extend({prop:function(e,t){return $(this,S.prop,e,t,1<arguments.length)},removeProp:function(e){return this.each(function(){delete this[S.propFix[e]||e]})}}),S.extend({prop:function(e,t,n){var r,i,o=e.nodeType;if(3!==o&&8!==o&&2!==o)return 1===o&&S.isXMLDoc(e)||(t=S.propFix[t]||t,i=S.propHooks[t]),void 0!==n?i&&"set"in i&&void 0!==(r=i.set(e,n,t))?r:e[t]=n:i&&"get"in i&&null!==(r=i.get(e,t))?r:e[t]},propHooks:{tabIndex:{get:function(e){var t=S.find.attr(e,"tabindex");return t?parseInt(t,10):ht.test(e.nodeName)||gt.test(e.nodeName)&&e.href?0:-1}}},propFix:{"for":"htmlFor","class":"className"}}),y.optSelected||(S.propHooks.selected={get:function(e){var t=e.parentNode;return t&&t.parentNode&&t.parentNode.selectedIndex,null},set:function(e){var t=e.parentNode;t&&(t.selectedIndex,t.parentNode&&t.parentNode.selectedIndex)}}),S.each(["tabIndex","readOnly","maxLength","cellSpacing","cellPadding","rowSpan","colSpan","useMap","frameBorder","contentEditable"],function(){S.propFix[this.toLowerCase()]=this}),S.fn.extend({addClass:function(t){var e,n,r,i,o,a,s,u=0;if(m(t))return this.each(function(e){S(this).addClass(t.call(this,e,yt(this)))});if((e=mt(t)).length)while(n=this[u++])if(i=yt(n),r=1===n.nodeType&&" "+vt(i)+" "){a=0;while(o=e[a++])r.indexOf(" "+o+" ")<0&&(r+=o+" ");i!==(s=vt(r))&&n.setAttribute("class",s)}return this},removeClass:function(t){var e,n,r,i,o,a,s,u=0;if(m(t))return this.each(function(e){S(this).removeClass(t.call(this,e,yt(this)))});if(!arguments.length)return this.attr("class","");if((e=mt(t)).length)while(n=this[u++])if(i=yt(n),r=1===n.nodeType&&" "+vt(i)+" "){a=0;while(o=e[a++])while(-1<r.indexOf(" "+o+" "))r=r.replace(" "+o+" "," ");i!==(s=vt(r))&&n.setAttribute("class",s)}return this},toggleClass:function(i,t){var o=typeof i,a="string"===o||Array.isArray(i);return"boolean"==typeof t&&a?t?this.addClass(i):this.removeClass(i):m(i)?this.each(function(e){S(this).toggleClass(i.call(this,e,yt(this),t),t)}):this.each(function(){var e,t,n,r;if(a){t=0,n=S(this),r=mt(i);while(e=r[t++])n.hasClass(e)?n.removeClass(e):n.addClass(e)}else void 0!==i&&"boolean"!==o||((e=yt(this))&&Y.set(this,"__className__",e),this.setAttribute&&this.setAttribute("class",e||!1===i?"":Y.get(this,"__className__")||""))})},hasClass:function(e){var t,n,r=0;t=" "+e+" ";while(n=this[r++])if(1===n.nodeType&&-1<(" "+vt(yt(n))+" ").indexOf(t))return!0;return!1}});var xt=/\r/g;S.fn.extend({val:function(n){var r,e,i,t=this[0];return arguments.length?(i=m(n),this.each(function(e){var t;1===this.nodeType&&(null==(t=i?n.call(this,e,S(this).val()):n)?t="":"number"==typeof t?t+="":Array.isArray(t)&&(t=S.map(t,function(e){return null==e?"":e+""})),(r=S.valHooks[this.type]||S.valHooks[this.nodeName.toLowerCase()])&&"set"in r&&void 0!==r.set(this,t,"value")||(this.value=t))})):t?(r=S.valHooks[t.type]||S.valHooks[t.nodeName.toLowerCase()])&&"get"in r&&void 0!==(e=r.get(t,"value"))?e:"string"==typeof(e=t.value)?e.replace(xt,""):null==e?"":e:void 0}}),S.extend({valHooks:{option:{get:function(e){var t=S.find.attr(e,"value");return null!=t?t:vt(S.text(e))}},select:{get:function(e){var t,n,r,i=e.options,o=e.selectedIndex,a="select-one"===e.type,s=a?null:[],u=a?o+1:i.length;for(r=o<0?u:a?o:0;r<u;r++)if(((n=i[r]).selected||r===o)&&!n.disabled&&(!n.parentNode.disabled||!A(n.parentNode,"optgroup"))){if(t=S(n).val(),a)return t;s.push(t)}return s},set:function(e,t){var n,r,i=e.options,o=S.makeArray(t),a=i.length;while(a--)((r=i[a]).selected=-1<S.inArray(S.valHooks.option.get(r),o))&&(n=!0);return n||(e.selectedIndex=-1),o}}}}),S.each(["radio","checkbox"],function(){S.valHooks[this]={set:function(e,t){if(Array.isArray(t))return e.checked=-1<S.inArray(S(e).val(),t)}},y.checkOn||(S.valHooks[this].get=function(e){return null===e.getAttribute("value")?"on":e.value})}),y.focusin="onfocusin"in C;var bt=/^(?:focusinfocus|focusoutblur)$/,wt=function(e){e.stopPropagation()};S.extend(S.event,{trigger:function(e,t,n,r){var i,o,a,s,u,l,c,f,p=[n||E],d=v.call(e,"type")?e.type:e,h=v.call(e,"namespace")?e.namespace.split("."):[];if(o=f=a=n=n||E,3!==n.nodeType&&8!==n.nodeType&&!bt.test(d+S.event.triggered)&&(-1<d.indexOf(".")&&(d=(h=d.split(".")).shift(),h.sort()),u=d.indexOf(":")<0&&"on"+d,(e=e[S.expando]?e:new S.Event(d,"object"==typeof e&&e)).isTrigger=r?2:3,e.namespace=h.join("."),e.rnamespace=e.namespace?new RegExp("(^|\\.)"+h.join("\\.(?:.*\\.|)")+"(\\.|$)"):null,e.result=void 0,e.target||(e.target=n),t=null==t?[e]:S.makeArray(t,[e]),c=S.event.special[d]||{},r||!c.trigger||!1!==c.trigger.apply(n,t))){if(!r&&!c.noBubble&&!x(n)){for(s=c.delegateType||d,bt.test(s+d)||(o=o.parentNode);o;o=o.parentNode)p.push(o),a=o;a===(n.ownerDocument||E)&&p.push(a.defaultView||a.parentWindow||C)}i=0;while((o=p[i++])&&!e.isPropagationStopped())f=o,e.type=1<i?s:c.bindType||d,(l=(Y.get(o,"events")||Object.create(null))[e.type]&&Y.get(o,"handle"))&&l.apply(o,t),(l=u&&o[u])&&l.apply&&V(o)&&(e.result=l.apply(o,t),!1===e.result&&e.preventDefault());return e.type=d,r||e.isDefaultPrevented()||c._default&&!1!==c._default.apply(p.pop(),t)||!V(n)||u&&m(n[d])&&!x(n)&&((a=n[u])&&(n[u]=null),S.event.triggered=d,e.isPropagationStopped()&&f.addEventListener(d,wt),n[d](),e.isPropagationStopped()&&f.removeEventListener(d,wt),S.event.triggered=void 0,a&&(n[u]=a)),e.result}},simulate:function(e,t,n){var r=S.extend(new S.Event,n,{type:e,isSimulated:!0});S.event.trigger(r,null,t)}}),S.fn.extend({trigger:function(e,t){return this.each(function(){S.event.trigger(e,t,this)})},triggerHandler:function(e,t){var n=this[0];if(n)return S.event.trigger(e,t,n,!0)}}),y.focusin||S.each({focus:"focusin",blur:"focusout"},function(n,r){var i=function(e){S.event.simulate(r,e.target,S.event.fix(e))};S.event.special[r]={setup:function(){var e=this.ownerDocument||this.document||this,t=Y.access(e,r);t||e.addEventListener(n,i,!0),Y.access(e,r,(t||0)+1)},teardown:function(){var e=this.ownerDocument||this.document||this,t=Y.access(e,r)-1;t?Y.access(e,r,t):(e.removeEventListener(n,i,!0),Y.remove(e,r))}}});var Tt=C.location,Ct={guid:Date.now()},Et=/\?/;S.parseXML=function(e){var t;if(!e||"string"!=typeof e)return null;try{t=(new C.DOMParser).parseFromString(e,"text/xml")}catch(e){t=void 0}return t&&!t.getElementsByTagName("parsererror").length||S.error("Invalid XML: "+e),t};var St=/\[\]$/,kt=/\r?\n/g,At=/^(?:submit|button|image|reset|file)$/i,Nt=/^(?:input|select|textarea|keygen)/i;function Dt(n,e,r,i){var t;if(Array.isArray(e))S.each(e,function(e,t){r||St.test(n)?i(n,t):Dt(n+"["+("object"==typeof t&&null!=t?e:"")+"]",t,r,i)});else if(r||"object"!==w(e))i(n,e);else for(t in e)Dt(n+"["+t+"]",e[t],r,i)}S.param=function(e,t){var n,r=[],i=function(e,t){var n=m(t)?t():t;r[r.length]=encodeURIComponent(e)+"="+encodeURIComponent(null==n?"":n)};if(null==e)return"";if(Array.isArray(e)||e.jquery&&!S.isPlainObject(e))S.each(e,function(){i(this.name,this.value)});else for(n in e)Dt(n,e[n],t,i);return r.join("&")},S.fn.extend({serialize:function(){return S.param(this.serializeArray())},serializeArray:function(){return this.map(function(){var e=S.prop(this,"elements");return e?S.makeArray(e):this}).filter(function(){var e=this.type;return this.name&&!S(this).is(":disabled")&&Nt.test(this.nodeName)&&!At.test(e)&&(this.checked||!pe.test(e))}).map(function(e,t){var n=S(this).val();return null==n?null:Array.isArray(n)?S.map(n,function(e){return{name:t.name,value:e.replace(kt,"\r\n")}}):{name:t.name,value:n.replace(kt,"\r\n")}}).get()}});var jt=/%20/g,qt=/#.*$/,Lt=/([?&])_=[^&]*/,Ht=/^(.*?):[ \t]*([^\r\n]*)$/gm,Ot=/^(?:GET|HEAD)$/,Pt=/^\/\//,Rt={},Mt={},It="*/".concat("*"),Wt=E.createElement("a");function Ft(o){return function(e,t){"string"!=typeof e&&(t=e,e="*");var n,r=0,i=e.toLowerCase().match(P)||[];if(m(t))while(n=i[r++])"+"===n[0]?(n=n.slice(1)||"*",(o[n]=o[n]||[]).unshift(t)):(o[n]=o[n]||[]).push(t)}}function Bt(t,i,o,a){var s={},u=t===Mt;function l(e){var r;return s[e]=!0,S.each(t[e]||[],function(e,t){var n=t(i,o,a);return"string"!=typeof n||u||s[n]?u?!(r=n):void 0:(i.dataTypes.unshift(n),l(n),!1)}),r}return l(i.dataTypes[0])||!s["*"]&&l("*")}function $t(e,t){var n,r,i=S.ajaxSettings.flatOptions||{};for(n in t)void 0!==t[n]&&((i[n]?e:r||(r={}))[n]=t[n]);return r&&S.extend(!0,e,r),e}Wt.href=Tt.href,S.extend({active:0,lastModified:{},etag:{},ajaxSettings:{url:Tt.href,type:"GET",isLocal:/^(?:about|app|app-storage|.+-extension|file|res|widget):$/.test(Tt.protocol),global:!0,processData:!0,async:!0,contentType:"application/x-www-form-urlencoded; charset=UTF-8",accepts:{"*":It,text:"text/plain",html:"text/html",xml:"application/xml, text/xml",json:"application/json, text/javascript"},contents:{xml:/\bxml\b/,html:/\bhtml/,json:/\bjson\b/},responseFields:{xml:"responseXML",text:"responseText",json:"responseJSON"},converters:{"* text":String,"text html":!0,"text json":JSON.parse,"text xml":S.parseXML},flatOptions:{url:!0,context:!0}},ajaxSetup:function(e,t){return t?$t($t(e,S.ajaxSettings),t):$t(S.ajaxSettings,e)},ajaxPrefilter:Ft(Rt),ajaxTransport:Ft(Mt),ajax:function(e,t){"object"==typeof e&&(t=e,e=void 0),t=t||{};var c,f,p,n,d,r,h,g,i,o,v=S.ajaxSetup({},t),y=v.context||v,m=v.context&&(y.nodeType||y.jquery)?S(y):S.event,x=S.Deferred(),b=S.Callbacks("once memory"),w=v.statusCode||{},a={},s={},u="canceled",T={readyState:0,getResponseHeader:function(e){var t;if(h){if(!n){n={};while(t=Ht.exec(p))n[t[1].toLowerCase()+" "]=(n[t[1].toLowerCase()+" "]||[]).concat(t[2])}t=n[e.toLowerCase()+" "]}return null==t?null:t.join(", ")},getAllResponseHeaders:function(){return h?p:null},setRequestHeader:function(e,t){return null==h&&(e=s[e.toLowerCase()]=s[e.toLowerCase()]||e,a[e]=t),this},overrideMimeType:function(e){return null==h&&(v.mimeType=e),this},statusCode:function(e){var t;if(e)if(h)T.always(e[T.status]);else for(t in e)w[t]=[w[t],e[t]];return this},abort:function(e){var t=e||u;return c&&c.abort(t),l(0,t),this}};if(x.promise(T),v.url=((e||v.url||Tt.href)+"").replace(Pt,Tt.protocol+"//"),v.type=t.method||t.type||v.method||v.type,v.dataTypes=(v.dataType||"*").toLowerCase().match(P)||[""],null==v.crossDomain){r=E.createElement("a");try{r.href=v.url,r.href=r.href,v.crossDomain=Wt.protocol+"//"+Wt.host!=r.protocol+"//"+r.host}catch(e){v.crossDomain=!0}}if(v.data&&v.processData&&"string"!=typeof v.data&&(v.data=S.param(v.data,v.traditional)),Bt(Rt,v,t,T),h)return T;for(i in(g=S.event&&v.global)&&0==S.active++&&S.event.trigger("ajaxStart"),v.type=v.type.toUpperCase(),v.hasContent=!Ot.test(v.type),f=v.url.replace(qt,""),v.hasContent?v.data&&v.processData&&0===(v.contentType||"").indexOf("application/x-www-form-urlencoded")&&(v.data=v.data.replace(jt,"+")):(o=v.url.slice(f.length),v.data&&(v.processData||"string"==typeof v.data)&&(f+=(Et.test(f)?"&":"?")+v.data,delete v.data),!1===v.cache&&(f=f.replace(Lt,"$1"),o=(Et.test(f)?"&":"?")+"_="+Ct.guid+++o),v.url=f+o),v.ifModified&&(S.lastModified[f]&&T.setRequestHeader("If-Modified-Since",S.lastModified[f]),S.etag[f]&&T.setRequestHeader("If-None-Match",S.etag[f])),(v.data&&v.hasContent&&!1!==v.contentType||t.contentType)&&T.setRequestHeader("Content-Type",v.contentType),T.setRequestHeader("Accept",v.dataTypes[0]&&v.accepts[v.dataTypes[0]]?v.accepts[v.dataTypes[0]]+("*"!==v.dataTypes[0]?", "+It+"; q=0.01":""):v.accepts["*"]),v.headers)T.setRequestHeader(i,v.headers[i]);if(v.beforeSend&&(!1===v.beforeSend.call(y,T,v)||h))return T.abort();if(u="abort",b.add(v.complete),T.done(v.success),T.fail(v.error),c=Bt(Mt,v,t,T)){if(T.readyState=1,g&&m.trigger("ajaxSend",[T,v]),h)return T;v.async&&0<v.timeout&&(d=C.setTimeout(function(){T.abort("timeout")},v.timeout));try{h=!1,c.send(a,l)}catch(e){if(h)throw e;l(-1,e)}}else l(-1,"No Transport");function l(e,t,n,r){var i,o,a,s,u,l=t;h||(h=!0,d&&C.clearTimeout(d),c=void 0,p=r||"",T.readyState=0<e?4:0,i=200<=e&&e<300||304===e,n&&(s=function(e,t,n){var r,i,o,a,s=e.contents,u=e.dataTypes;while("*"===u[0])u.shift(),void 0===r&&(r=e.mimeType||t.getResponseHeader("Content-Type"));if(r)for(i in s)if(s[i]&&s[i].test(r)){u.unshift(i);break}if(u[0]in n)o=u[0];else{for(i in n){if(!u[0]||e.converters[i+" "+u[0]]){o=i;break}a||(a=i)}o=o||a}if(o)return o!==u[0]&&u.unshift(o),n[o]}(v,T,n)),!i&&-1<S.inArray("script",v.dataTypes)&&(v.converters["text script"]=function(){}),s=function(e,t,n,r){var i,o,a,s,u,l={},c=e.dataTypes.slice();if(c[1])for(a in e.converters)l[a.toLowerCase()]=e.converters[a];o=c.shift();while(o)if(e.responseFields[o]&&(n[e.responseFields[o]]=t),!u&&r&&e.dataFilter&&(t=e.dataFilter(t,e.dataType)),u=o,o=c.shift())if("*"===o)o=u;else if("*"!==u&&u!==o){if(!(a=l[u+" "+o]||l["* "+o]))for(i in l)if((s=i.split(" "))[1]===o&&(a=l[u+" "+s[0]]||l["* "+s[0]])){!0===a?a=l[i]:!0!==l[i]&&(o=s[0],c.unshift(s[1]));break}if(!0!==a)if(a&&e["throws"])t=a(t);else try{t=a(t)}catch(e){return{state:"parsererror",error:a?e:"No conversion from "+u+" to "+o}}}return{state:"success",data:t}}(v,s,T,i),i?(v.ifModified&&((u=T.getResponseHeader("Last-Modified"))&&(S.lastModified[f]=u),(u=T.getResponseHeader("etag"))&&(S.etag[f]=u)),204===e||"HEAD"===v.type?l="nocontent":304===e?l="notmodified":(l=s.state,o=s.data,i=!(a=s.error))):(a=l,!e&&l||(l="error",e<0&&(e=0))),T.status=e,T.statusText=(t||l)+"",i?x.resolveWith(y,[o,l,T]):x.rejectWith(y,[T,l,a]),T.statusCode(w),w=void 0,g&&m.trigger(i?"ajaxSuccess":"ajaxError",[T,v,i?o:a]),b.fireWith(y,[T,l]),g&&(m.trigger("ajaxComplete",[T,v]),--S.active||S.event.trigger("ajaxStop")))}return T},getJSON:function(e,t,n){return S.get(e,t,n,"json")},getScript:function(e,t){return S.get(e,void 0,t,"script")}}),S.each(["get","post"],function(e,i){S[i]=function(e,t,n,r){return m(t)&&(r=r||n,n=t,t=void 0),S.ajax(S.extend({url:e,type:i,dataType:r,data:t,success:n},S.isPlainObject(e)&&e))}}),S.ajaxPrefilter(function(e){var t;for(t in e.headers)"content-type"===t.toLowerCase()&&(e.contentType=e.headers[t]||"")}),S._evalUrl=function(e,t,n){return S.ajax({url:e,type:"GET",dataType:"script",cache:!0,async:!1,global:!1,converters:{"text script":function(){}},dataFilter:function(e){S.globalEval(e,t,n)}})},S.fn.extend({wrapAll:function(e){var t;return this[0]&&(m(e)&&(e=e.call(this[0])),t=S(e,this[0].ownerDocument).eq(0).clone(!0),this[0].parentNode&&t.insertBefore(this[0]),t.map(function(){var e=this;while(e.firstElementChild)e=e.firstElementChild;return e}).append(this)),this},wrapInner:function(n){return m(n)?this.each(function(e){S(this).wrapInner(n.call(this,e))}):this.each(function(){var e=S(this),t=e.contents();t.length?t.wrapAll(n):e.append(n)})},wrap:function(t){var n=m(t);return this.each(function(e){S(this).wrapAll(n?t.call(this,e):t)})},unwrap:function(e){return this.parent(e).not("body").each(function(){S(this).replaceWith(this.childNodes)}),this}}),S.expr.pseudos.hidden=function(e){return!S.expr.pseudos.visible(e)},S.expr.pseudos.visible=function(e){return!!(e.offsetWidth||e.offsetHeight||e.getClientRects().length)},S.ajaxSettings.xhr=function(){try{return new C.XMLHttpRequest}catch(e){}};var _t={0:200,1223:204},zt=S.ajaxSettings.xhr();y.cors=!!zt&&"withCredentials"in zt,y.ajax=zt=!!zt,S.ajaxTransport(function(i){var o,a;if(y.cors||zt&&!i.crossDomain)return{send:function(e,t){var n,r=i.xhr();if(r.open(i.type,i.url,i.async,i.username,i.password),i.xhrFields)for(n in i.xhrFields)r[n]=i.xhrFields[n];for(n in i.mimeType&&r.overrideMimeType&&r.overrideMimeType(i.mimeType),i.crossDomain||e["X-Requested-With"]||(e["X-Requested-With"]="XMLHttpRequest"),e)r.setRequestHeader(n,e[n]);o=function(e){return function(){o&&(o=a=r.onload=r.onerror=r.onabort=r.ontimeout=r.onreadystatechange=null,"abort"===e?r.abort():"error"===e?"number"!=typeof r.status?t(0,"error"):t(r.status,r.statusText):t(_t[r.status]||r.status,r.statusText,"text"!==(r.responseType||"text")||"string"!=typeof r.responseText?{binary:r.response}:{text:r.responseText},r.getAllResponseHeaders()))}},r.onload=o(),a=r.onerror=r.ontimeout=o("error"),void 0!==r.onabort?r.onabort=a:r.onreadystatechange=function(){4===r.readyState&&C.setTimeout(function(){o&&a()})},o=o("abort");try{r.send(i.hasContent&&i.data||null)}catch(e){if(o)throw e}},abort:function(){o&&o()}}}),S.ajaxPrefilter(function(e){e.crossDomain&&(e.contents.script=!1)}),S.ajaxSetup({accepts:{script:"text/javascript, application/javascript, application/ecmascript, application/x-ecmascript"},contents:{script:/\b(?:java|ecma)script\b/},converters:{"text script":function(e){return S.globalEval(e),e}}}),S.ajaxPrefilter("script",function(e){void 0===e.cache&&(e.cache=!1),e.crossDomain&&(e.type="GET")}),S.ajaxTransport("script",function(n){var r,i;if(n.crossDomain||n.scriptAttrs)return{send:function(e,t){r=S("<script>").attr(n.scriptAttrs||{}).prop({charset:n.scriptCharset,src:n.url}).on("load error",i=function(e){r.remove(),i=null,e&&t("error"===e.type?404:200,e.type)}),E.head.appendChild(r[0])},abort:function(){i&&i()}}});var Ut,Xt=[],Vt=/(=)\?(?=&|$)|\?\?/;S.ajaxSetup({jsonp:"callback",jsonpCallback:function(){var e=Xt.pop()||S.expando+"_"+Ct.guid++;return this[e]=!0,e}}),S.ajaxPrefilter("json jsonp",function(e,t,n){var r,i,o,a=!1!==e.jsonp&&(Vt.test(e.url)?"url":"string"==typeof e.data&&0===(e.contentType||"").indexOf("application/x-www-form-urlencoded")&&Vt.test(e.data)&&"data");if(a||"jsonp"===e.dataTypes[0])return r=e.jsonpCallback=m(e.jsonpCallback)?e.jsonpCallback():e.jsonpCallback,a?e[a]=e[a].replace(Vt,"$1"+r):!1!==e.jsonp&&(e.url+=(Et.test(e.url)?"&":"?")+e.jsonp+"="+r),e.converters["script json"]=function(){return o||S.error(r+" was not called"),o[0]},e.dataTypes[0]="json",i=C[r],C[r]=function(){o=arguments},n.always(function(){void 0===i?S(C).removeProp(r):C[r]=i,e[r]&&(e.jsonpCallback=t.jsonpCallback,Xt.push(r)),o&&m(i)&&i(o[0]),o=i=void 0}),"script"}),y.createHTMLDocument=((Ut=E.implementation.createHTMLDocument("").body).innerHTML="<form></form><form></form>",2===Ut.childNodes.length),S.parseHTML=function(e,t,n){return"string"!=typeof e?[]:("boolean"==typeof t&&(n=t,t=!1),t||(y.createHTMLDocument?((r=(t=E.implementation.createHTMLDocument("")).createElement("base")).href=E.location.href,t.head.appendChild(r)):t=E),o=!n&&[],(i=N.exec(e))?[t.createElement(i[1])]:(i=xe([e],t,o),o&&o.length&&S(o).remove(),S.merge([],i.childNodes)));var r,i,o},S.fn.load=function(e,t,n){var r,i,o,a=this,s=e.indexOf(" ");return-1<s&&(r=vt(e.slice(s)),e=e.slice(0,s)),m(t)?(n=t,t=void 0):t&&"object"==typeof t&&(i="POST"),0<a.length&&S.ajax({url:e,type:i||"GET",dataType:"html",data:t}).done(function(e){o=arguments,a.html(r?S("<div>").append(S.parseHTML(e)).find(r):e)}).always(n&&function(e,t){a.each(function(){n.apply(this,o||[e.responseText,t,e])})}),this},S.expr.pseudos.animated=function(t){return S.grep(S.timers,function(e){return t===e.elem}).length},S.offset={setOffset:function(e,t,n){var r,i,o,a,s,u,l=S.css(e,"position"),c=S(e),f={};"static"===l&&(e.style.position="relative"),s=c.offset(),o=S.css(e,"top"),u=S.css(e,"left"),("absolute"===l||"fixed"===l)&&-1<(o+u).indexOf("auto")?(a=(r=c.position()).top,i=r.left):(a=parseFloat(o)||0,i=parseFloat(u)||0),m(t)&&(t=t.call(e,n,S.extend({},s))),null!=t.top&&(f.top=t.top-s.top+a),null!=t.left&&(f.left=t.left-s.left+i),"using"in t?t.using.call(e,f):("number"==typeof f.top&&(f.top+="px"),"number"==typeof f.left&&(f.left+="px"),c.css(f))}},S.fn.extend({offset:function(t){if(arguments.length)return void 0===t?this:this.each(function(e){S.offset.setOffset(this,t,e)});var e,n,r=this[0];return r?r.getClientRects().length?(e=r.getBoundingClientRect(),n=r.ownerDocument.defaultView,{top:e.top+n.pageYOffset,left:e.left+n.pageXOffset}):{top:0,left:0}:void 0},position:function(){if(this[0]){var e,t,n,r=this[0],i={top:0,left:0};if("fixed"===S.css(r,"position"))t=r.getBoundingClientRect();else{t=this.offset(),n=r.ownerDocument,e=r.offsetParent||n.documentElement;while(e&&(e===n.body||e===n.documentElement)&&"static"===S.css(e,"position"))e=e.parentNode;e&&e!==r&&1===e.nodeType&&((i=S(e).offset()).top+=S.css(e,"borderTopWidth",!0),i.left+=S.css(e,"borderLeftWidth",!0))}return{top:t.top-i.top-S.css(r,"marginTop",!0),left:t.left-i.left-S.css(r,"marginLeft",!0)}}},offsetParent:function(){return this.map(function(){var e=this.offsetParent;while(e&&"static"===S.css(e,"position"))e=e.offsetParent;return e||re})}}),S.each({scrollLeft:"pageXOffset",scrollTop:"pageYOffset"},function(t,i){var o="pageYOffset"===i;S.fn[t]=function(e){return $(this,function(e,t,n){var r;if(x(e)?r=e:9===e.nodeType&&(r=e.defaultView),void 0===n)return r?r[i]:e[t];r?r.scrollTo(o?r.pageXOffset:n,o?n:r.pageYOffset):e[t]=n},t,e,arguments.length)}}),S.each(["top","left"],function(e,n){S.cssHooks[n]=$e(y.pixelPosition,function(e,t){if(t)return t=Be(e,n),Me.test(t)?S(e).position()[n]+"px":t})}),S.each({Height:"height",Width:"width"},function(a,s){S.each({padding:"inner"+a,content:s,"":"outer"+a},function(r,o){S.fn[o]=function(e,t){var n=arguments.length&&(r||"boolean"!=typeof e),i=r||(!0===e||!0===t?"margin":"border");return $(this,function(e,t,n){var r;return x(e)?0===o.indexOf("outer")?e["inner"+a]:e.document.documentElement["client"+a]:9===e.nodeType?(r=e.documentElement,Math.max(e.body["scroll"+a],r["scroll"+a],e.body["offset"+a],r["offset"+a],r["client"+a])):void 0===n?S.css(e,t,i):S.style(e,t,n,i)},s,n?e:void 0,n)}})}),S.each(["ajaxStart","ajaxStop","ajaxComplete","ajaxError","ajaxSuccess","ajaxSend"],function(e,t){S.fn[t]=function(e){return this.on(t,e)}}),S.fn.extend({bind:function(e,t,n){return this.on(e,null,t,n)},unbind:function(e,t){return this.off(e,null,t)},delegate:function(e,t,n,r){return this.on(t,e,n,r)},undelegate:function(e,t,n){return 1===arguments.length?this.off(e,"**"):this.off(t,e||"**",n)},hover:function(e,t){return this.mouseenter(e).mouseleave(t||e)}}),S.each("blur focus focusin focusout resize scroll click dblclick mousedown mouseup mousemove mouseover mouseout mouseenter mouseleave change select submit keydown keypress keyup contextmenu".split(" "),function(e,n){S.fn[n]=function(e,t){return 0<arguments.length?this.on(n,null,e,t):this.trigger(n)}});var Gt=/^[\s\uFEFF\xA0]+|[\s\uFEFF\xA0]+$/g;S.proxy=function(e,t){var n,r,i;if("string"==typeof t&&(n=e[t],t=e,e=n),m(e))return r=s.call(arguments,2),(i=function(){return e.apply(t||this,r.concat(s.call(arguments)))}).guid=e.guid=e.guid||S.guid++,i},S.holdReady=function(e){e?S.readyWait++:S.ready(!0)},S.isArray=Array.isArray,S.parseJSON=JSON.parse,S.nodeName=A,S.isFunction=m,S.isWindow=x,S.camelCase=X,S.type=w,S.now=Date.now,S.isNumeric=function(e){var t=S.type(e);return("number"===t||"string"===t)&&!isNaN(e-parseFloat(e))},S.trim=function(e){return null==e?"":(e+"").replace(Gt,"")},"function"==typeof define&&define.amd&&define("jquery",[],function(){return S});var Yt=C.jQuery,Qt=C.$;return S.noConflict=function(e){return C.$===S&&(C.$=Qt),e&&C.jQuery===S&&(C.jQuery=Yt),S},"undefined"==typeof e&&(C.jQuery=C.$=S),S});
+/*! jQuery v3.6.0 | (c) OpenJS Foundation and other contributors | jquery.org/license */
+!function(e,t){"use strict";"object"==typeof module&&"object"==typeof module.exports?module.exports=e.document?t(e,!0):function(e){if(!e.document)throw new Error("jQuery requires a window with a document");return t(e)}:t(e)}("undefined"!=typeof window?window:this,function(C,e){"use strict";var t=[],r=Object.getPrototypeOf,s=t.slice,g=t.flat?function(e){return t.flat.call(e)}:function(e){return t.concat.apply([],e)},u=t.push,i=t.indexOf,n={},o=n.toString,v=n.hasOwnProperty,a=v.toString,l=a.call(Object),y={},m=function(e){return"function"==typeof e&&"number"!=typeof e.nodeType&&"function"!=typeof e.item},x=function(e){return null!=e&&e===e.window},E=C.document,c={type:!0,src:!0,nonce:!0,noModule:!0};function b(e,t,n){var r,i,o=(n=n||E).createElement("script");if(o.text=e,t)for(r in c)(i=t[r]||t.getAttribute&&t.getAttribute(r))&&o.setAttribute(r,i);n.head.appendChild(o).parentNode.removeChild(o)}function w(e){return null==e?e+"":"object"==typeof e||"function"==typeof e?n[o.call(e)]||"object":typeof e}var f="3.6.0",S=function(e,t){return new S.fn.init(e,t)};function p(e){var t=!!e&&"length"in e&&e.length,n=w(e);return!m(e)&&!x(e)&&("array"===n||0===t||"number"==typeof t&&0<t&&t-1 in e)}S.fn=S.prototype={jquery:f,constructor:S,length:0,toArray:function(){return s.call(this)},get:function(e){return null==e?s.call(this):e<0?this[e+this.length]:this[e]},pushStack:function(e){var t=S.merge(this.constructor(),e);return t.prevObject=this,t},each:function(e){return S.each(this,e)},map:function(n){return this.pushStack(S.map(this,function(e,t){return n.call(e,t,e)}))},slice:function(){return this.pushStack(s.apply(this,arguments))},first:function(){return this.eq(0)},last:function(){return this.eq(-1)},even:function(){return this.pushStack(S.grep(this,function(e,t){return(t+1)%2}))},odd:function(){return this.pushStack(S.grep(this,function(e,t){return t%2}))},eq:function(e){var t=this.length,n=+e+(e<0?t:0);return this.pushStack(0<=n&&n<t?[this[n]]:[])},end:function(){return this.prevObject||this.constructor()},push:u,sort:t.sort,splice:t.splice},S.extend=S.fn.extend=function(){var e,t,n,r,i,o,a=arguments[0]||{},s=1,u=arguments.length,l=!1;for("boolean"==typeof a&&(l=a,a=arguments[s]||{},s++),"object"==typeof a||m(a)||(a={}),s===u&&(a=this,s--);s<u;s++)if(null!=(e=arguments[s]))for(t in e)r=e[t],"__proto__"!==t&&a!==r&&(l&&r&&(S.isPlainObject(r)||(i=Array.isArray(r)))?(n=a[t],o=i&&!Array.isArray(n)?[]:i||S.isPlainObject(n)?n:{},i=!1,a[t]=S.extend(l,o,r)):void 0!==r&&(a[t]=r));return a},S.extend({expando:"jQuery"+(f+Math.random()).replace(/\D/g,""),isReady:!0,error:function(e){throw new Error(e)},noop:function(){},isPlainObject:function(e){var t,n;return!(!e||"[object Object]"!==o.call(e))&&(!(t=r(e))||"function"==typeof(n=v.call(t,"constructor")&&t.constructor)&&a.call(n)===l)},isEmptyObject:function(e){var t;for(t in e)return!1;return!0},globalEval:function(e,t,n){b(e,{nonce:t&&t.nonce},n)},each:function(e,t){var n,r=0;if(p(e)){for(n=e.length;r<n;r++)if(!1===t.call(e[r],r,e[r]))break}else for(r in e)if(!1===t.call(e[r],r,e[r]))break;return e},makeArray:function(e,t){var n=t||[];return null!=e&&(p(Object(e))?S.merge(n,"string"==typeof e?[e]:e):u.call(n,e)),n},inArray:function(e,t,n){return null==t?-1:i.call(t,e,n)},merge:function(e,t){for(var n=+t.length,r=0,i=e.length;r<n;r++)e[i++]=t[r];return e.length=i,e},grep:function(e,t,n){for(var r=[],i=0,o=e.length,a=!n;i<o;i++)!t(e[i],i)!==a&&r.push(e[i]);return r},map:function(e,t,n){var r,i,o=0,a=[];if(p(e))for(r=e.length;o<r;o++)null!=(i=t(e[o],o,n))&&a.push(i);else for(o in e)null!=(i=t(e[o],o,n))&&a.push(i);return g(a)},guid:1,support:y}),"function"==typeof Symbol&&(S.fn[Symbol.iterator]=t[Symbol.iterator]),S.each("Boolean Number String Function Array Date RegExp Object Error Symbol".split(" "),function(e,t){n["[object "+t+"]"]=t.toLowerCase()});var d=function(n){var e,d,b,o,i,h,f,g,w,u,l,T,C,a,E,v,s,c,y,S="sizzle"+1*new Date,p=n.document,k=0,r=0,m=ue(),x=ue(),A=ue(),N=ue(),j=function(e,t){return e===t&&(l=!0),0},D={}.hasOwnProperty,t=[],q=t.pop,L=t.push,H=t.push,O=t.slice,P=function(e,t){for(var n=0,r=e.length;n<r;n++)if(e[n]===t)return n;return-1},R="checked|selected|async|autofocus|autoplay|controls|defer|disabled|hidden|ismap|loop|multiple|open|readonly|required|scoped",M="[\\x20\\t\\r\\n\\f]",I="(?:\\\\[\\da-fA-F]{1,6}"+M+"?|\\\\[^\\r\\n\\f]|[\\w-]|[^\0-\\x7f])+",W="\\["+M+"*("+I+")(?:"+M+"*([*^$|!~]?=)"+M+"*(?:'((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\"|("+I+"))|)"+M+"*\\]",F=":("+I+")(?:\\((('((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\")|((?:\\\\.|[^\\\\()[\\]]|"+W+")*)|.*)\\)|)",B=new RegExp(M+"+","g"),$=new RegExp("^"+M+"+|((?:^|[^\\\\])(?:\\\\.)*)"+M+"+$","g"),_=new RegExp("^"+M+"*,"+M+"*"),z=new RegExp("^"+M+"*([>+~]|"+M+")"+M+"*"),U=new RegExp(M+"|>"),X=new RegExp(F),V=new RegExp("^"+I+"$"),G={ID:new RegExp("^#("+I+")"),CLASS:new RegExp("^\\.("+I+")"),TAG:new RegExp("^("+I+"|[*])"),ATTR:new RegExp("^"+W),PSEUDO:new RegExp("^"+F),CHILD:new RegExp("^:(only|first|last|nth|nth-last)-(child|of-type)(?:\\("+M+"*(even|odd|(([+-]|)(\\d*)n|)"+M+"*(?:([+-]|)"+M+"*(\\d+)|))"+M+"*\\)|)","i"),bool:new RegExp("^(?:"+R+")$","i"),needsContext:new RegExp("^"+M+"*[>+~]|:(even|odd|eq|gt|lt|nth|first|last)(?:\\("+M+"*((?:-\\d)?\\d*)"+M+"*\\)|)(?=[^-]|$)","i")},Y=/HTML$/i,Q=/^(?:input|select|textarea|button)$/i,J=/^h\d$/i,K=/^[^{]+\{\s*\[native \w/,Z=/^(?:#([\w-]+)|(\w+)|\.([\w-]+))$/,ee=/[+~]/,te=new RegExp("\\\\[\\da-fA-F]{1,6}"+M+"?|\\\\([^\\r\\n\\f])","g"),ne=function(e,t){var n="0x"+e.slice(1)-65536;return t||(n<0?String.fromCharCode(n+65536):String.fromCharCode(n>>10|55296,1023&n|56320))},re=/([\0-\x1f\x7f]|^-?\d)|^-$|[^\0-\x1f\x7f-\uFFFF\w-]/g,ie=function(e,t){return t?"\0"===e?"\ufffd":e.slice(0,-1)+"\\"+e.charCodeAt(e.length-1).toString(16)+" ":"\\"+e},oe=function(){T()},ae=be(function(e){return!0===e.disabled&&"fieldset"===e.nodeName.toLowerCase()},{dir:"parentNode",next:"legend"});try{H.apply(t=O.call(p.childNodes),p.childNodes),t[p.childNodes.length].nodeType}catch(e){H={apply:t.length?function(e,t){L.apply(e,O.call(t))}:function(e,t){var n=e.length,r=0;while(e[n++]=t[r++]);e.length=n-1}}}function se(t,e,n,r){var i,o,a,s,u,l,c,f=e&&e.ownerDocument,p=e?e.nodeType:9;if(n=n||[],"string"!=typeof t||!t||1!==p&&9!==p&&11!==p)return n;if(!r&&(T(e),e=e||C,E)){if(11!==p&&(u=Z.exec(t)))if(i=u[1]){if(9===p){if(!(a=e.getElementById(i)))return n;if(a.id===i)return n.push(a),n}else if(f&&(a=f.getElementById(i))&&y(e,a)&&a.id===i)return n.push(a),n}else{if(u[2])return H.apply(n,e.getElementsByTagName(t)),n;if((i=u[3])&&d.getElementsByClassName&&e.getElementsByClassName)return H.apply(n,e.getElementsByClassName(i)),n}if(d.qsa&&!N[t+" "]&&(!v||!v.test(t))&&(1!==p||"object"!==e.nodeName.toLowerCase())){if(c=t,f=e,1===p&&(U.test(t)||z.test(t))){(f=ee.test(t)&&ye(e.parentNode)||e)===e&&d.scope||((s=e.getAttribute("id"))?s=s.replace(re,ie):e.setAttribute("id",s=S)),o=(l=h(t)).length;while(o--)l[o]=(s?"#"+s:":scope")+" "+xe(l[o]);c=l.join(",")}try{return H.apply(n,f.querySelectorAll(c)),n}catch(e){N(t,!0)}finally{s===S&&e.removeAttribute("id")}}}return g(t.replace($,"$1"),e,n,r)}function ue(){var r=[];return function e(t,n){return r.push(t+" ")>b.cacheLength&&delete e[r.shift()],e[t+" "]=n}}function le(e){return e[S]=!0,e}function ce(e){var t=C.createElement("fieldset");try{return!!e(t)}catch(e){return!1}finally{t.parentNode&&t.parentNode.removeChild(t),t=null}}function fe(e,t){var n=e.split("|"),r=n.length;while(r--)b.attrHandle[n[r]]=t}function pe(e,t){var n=t&&e,r=n&&1===e.nodeType&&1===t.nodeType&&e.sourceIndex-t.sourceIndex;if(r)return r;if(n)while(n=n.nextSibling)if(n===t)return-1;return e?1:-1}function de(t){return function(e){return"input"===e.nodeName.toLowerCase()&&e.type===t}}function he(n){return function(e){var t=e.nodeName.toLowerCase();return("input"===t||"button"===t)&&e.type===n}}function ge(t){return function(e){return"form"in e?e.parentNode&&!1===e.disabled?"label"in e?"label"in e.parentNode?e.parentNode.disabled===t:e.disabled===t:e.isDisabled===t||e.isDisabled!==!t&&ae(e)===t:e.disabled===t:"label"in e&&e.disabled===t}}function ve(a){return le(function(o){return o=+o,le(function(e,t){var n,r=a([],e.length,o),i=r.length;while(i--)e[n=r[i]]&&(e[n]=!(t[n]=e[n]))})})}function ye(e){return e&&"undefined"!=typeof e.getElementsByTagName&&e}for(e in d=se.support={},i=se.isXML=function(e){var t=e&&e.namespaceURI,n=e&&(e.ownerDocument||e).documentElement;return!Y.test(t||n&&n.nodeName||"HTML")},T=se.setDocument=function(e){var t,n,r=e?e.ownerDocument||e:p;return r!=C&&9===r.nodeType&&r.documentElement&&(a=(C=r).documentElement,E=!i(C),p!=C&&(n=C.defaultView)&&n.top!==n&&(n.addEventListener?n.addEventListener("unload",oe,!1):n.attachEvent&&n.attachEvent("onunload",oe)),d.scope=ce(function(e){return a.appendChild(e).appendChild(C.createElement("div")),"undefined"!=typeof e.querySelectorAll&&!e.querySelectorAll(":scope fieldset div").length}),d.attributes=ce(function(e){return e.className="i",!e.getAttribute("className")}),d.getElementsByTagName=ce(function(e){return e.appendChild(C.createComment("")),!e.getElementsByTagName("*").length}),d.getElementsByClassName=K.test(C.getElementsByClassName),d.getById=ce(function(e){return a.appendChild(e).id=S,!C.getElementsByName||!C.getElementsByName(S).length}),d.getById?(b.filter.ID=function(e){var t=e.replace(te,ne);return function(e){return e.getAttribute("id")===t}},b.find.ID=function(e,t){if("undefined"!=typeof t.getElementById&&E){var n=t.getElementById(e);return n?[n]:[]}}):(b.filter.ID=function(e){var n=e.replace(te,ne);return function(e){var t="undefined"!=typeof e.getAttributeNode&&e.getAttributeNode("id");return t&&t.value===n}},b.find.ID=function(e,t){if("undefined"!=typeof t.getElementById&&E){var n,r,i,o=t.getElementById(e);if(o){if((n=o.getAttributeNode("id"))&&n.value===e)return[o];i=t.getElementsByName(e),r=0;while(o=i[r++])if((n=o.getAttributeNode("id"))&&n.value===e)return[o]}return[]}}),b.find.TAG=d.getElementsByTagName?function(e,t){return"undefined"!=typeof t.getElementsByTagName?t.getElementsByTagName(e):d.qsa?t.querySelectorAll(e):void 0}:function(e,t){var n,r=[],i=0,o=t.getElementsByTagName(e);if("*"===e){while(n=o[i++])1===n.nodeType&&r.push(n);return r}return o},b.find.CLASS=d.getElementsByClassName&&function(e,t){if("undefined"!=typeof t.getElementsByClassName&&E)return t.getElementsByClassName(e)},s=[],v=[],(d.qsa=K.test(C.querySelectorAll))&&(ce(function(e){var t;a.appendChild(e).innerHTML="<a id='"+S+"'></a><select id='"+S+"-\r\\' msallowcapture=''><option selected=''></option></select>",e.querySelectorAll("[msallowcapture^='']").length&&v.push("[*^$]="+M+"*(?:''|\"\")"),e.querySelectorAll("[selected]").length||v.push("\\["+M+"*(?:value|"+R+")"),e.querySelectorAll("[id~="+S+"-]").length||v.push("~="),(t=C.createElement("input")).setAttribute("name",""),e.appendChild(t),e.querySelectorAll("[name='']").length||v.push("\\["+M+"*name"+M+"*="+M+"*(?:''|\"\")"),e.querySelectorAll(":checked").length||v.push(":checked"),e.querySelectorAll("a#"+S+"+*").length||v.push(".#.+[+~]"),e.querySelectorAll("\\\f"),v.push("[\\r\\n\\f]")}),ce(function(e){e.innerHTML="<a href='' disabled='disabled'></a><select disabled='disabled'><option/></select>";var t=C.createElement("input");t.setAttribute("type","hidden"),e.appendChild(t).setAttribute("name","D"),e.querySelectorAll("[name=d]").length&&v.push("name"+M+"*[*^$|!~]?="),2!==e.querySelectorAll(":enabled").length&&v.push(":enabled",":disabled"),a.appendChild(e).disabled=!0,2!==e.querySelectorAll(":disabled").length&&v.push(":enabled",":disabled"),e.querySelectorAll("*,:x"),v.push(",.*:")})),(d.matchesSelector=K.test(c=a.matches||a.webkitMatchesSelector||a.mozMatchesSelector||a.oMatchesSelector||a.msMatchesSelector))&&ce(function(e){d.disconnectedMatch=c.call(e,"*"),c.call(e,"[s!='']:x"),s.push("!=",F)}),v=v.length&&new RegExp(v.join("|")),s=s.length&&new RegExp(s.join("|")),t=K.test(a.compareDocumentPosition),y=t||K.test(a.contains)?function(e,t){var n=9===e.nodeType?e.documentElement:e,r=t&&t.parentNode;return e===r||!(!r||1!==r.nodeType||!(n.contains?n.contains(r):e.compareDocumentPosition&&16&e.compareDocumentPosition(r)))}:function(e,t){if(t)while(t=t.parentNode)if(t===e)return!0;return!1},j=t?function(e,t){if(e===t)return l=!0,0;var n=!e.compareDocumentPosition-!t.compareDocumentPosition;return n||(1&(n=(e.ownerDocument||e)==(t.ownerDocument||t)?e.compareDocumentPosition(t):1)||!d.sortDetached&&t.compareDocumentPosition(e)===n?e==C||e.ownerDocument==p&&y(p,e)?-1:t==C||t.ownerDocument==p&&y(p,t)?1:u?P(u,e)-P(u,t):0:4&n?-1:1)}:function(e,t){if(e===t)return l=!0,0;var n,r=0,i=e.parentNode,o=t.parentNode,a=[e],s=[t];if(!i||!o)return e==C?-1:t==C?1:i?-1:o?1:u?P(u,e)-P(u,t):0;if(i===o)return pe(e,t);n=e;while(n=n.parentNode)a.unshift(n);n=t;while(n=n.parentNode)s.unshift(n);while(a[r]===s[r])r++;return r?pe(a[r],s[r]):a[r]==p?-1:s[r]==p?1:0}),C},se.matches=function(e,t){return se(e,null,null,t)},se.matchesSelector=function(e,t){if(T(e),d.matchesSelector&&E&&!N[t+" "]&&(!s||!s.test(t))&&(!v||!v.test(t)))try{var n=c.call(e,t);if(n||d.disconnectedMatch||e.document&&11!==e.document.nodeType)return n}catch(e){N(t,!0)}return 0<se(t,C,null,[e]).length},se.contains=function(e,t){return(e.ownerDocument||e)!=C&&T(e),y(e,t)},se.attr=function(e,t){(e.ownerDocument||e)!=C&&T(e);var n=b.attrHandle[t.toLowerCase()],r=n&&D.call(b.attrHandle,t.toLowerCase())?n(e,t,!E):void 0;return void 0!==r?r:d.attributes||!E?e.getAttribute(t):(r=e.getAttributeNode(t))&&r.specified?r.value:null},se.escape=function(e){return(e+"").replace(re,ie)},se.error=function(e){throw new Error("Syntax error, unrecognized expression: "+e)},se.uniqueSort=function(e){var t,n=[],r=0,i=0;if(l=!d.detectDuplicates,u=!d.sortStable&&e.slice(0),e.sort(j),l){while(t=e[i++])t===e[i]&&(r=n.push(i));while(r--)e.splice(n[r],1)}return u=null,e},o=se.getText=function(e){var t,n="",r=0,i=e.nodeType;if(i){if(1===i||9===i||11===i){if("string"==typeof e.textContent)return e.textContent;for(e=e.firstChild;e;e=e.nextSibling)n+=o(e)}else if(3===i||4===i)return e.nodeValue}else while(t=e[r++])n+=o(t);return n},(b=se.selectors={cacheLength:50,createPseudo:le,match:G,attrHandle:{},find:{},relative:{">":{dir:"parentNode",first:!0}," ":{dir:"parentNode"},"+":{dir:"previousSibling",first:!0},"~":{dir:"previousSibling"}},preFilter:{ATTR:function(e){return e[1]=e[1].replace(te,ne),e[3]=(e[3]||e[4]||e[5]||"").replace(te,ne),"~="===e[2]&&(e[3]=" "+e[3]+" "),e.slice(0,4)},CHILD:function(e){return e[1]=e[1].toLowerCase(),"nth"===e[1].slice(0,3)?(e[3]||se.error(e[0]),e[4]=+(e[4]?e[5]+(e[6]||1):2*("even"===e[3]||"odd"===e[3])),e[5]=+(e[7]+e[8]||"odd"===e[3])):e[3]&&se.error(e[0]),e},PSEUDO:function(e){var t,n=!e[6]&&e[2];return G.CHILD.test(e[0])?null:(e[3]?e[2]=e[4]||e[5]||"":n&&X.test(n)&&(t=h(n,!0))&&(t=n.indexOf(")",n.length-t)-n.length)&&(e[0]=e[0].slice(0,t),e[2]=n.slice(0,t)),e.slice(0,3))}},filter:{TAG:function(e){var t=e.replace(te,ne).toLowerCase();return"*"===e?function(){return!0}:function(e){return e.nodeName&&e.nodeName.toLowerCase()===t}},CLASS:function(e){var t=m[e+" "];return t||(t=new RegExp("(^|"+M+")"+e+"("+M+"|$)"))&&m(e,function(e){return t.test("string"==typeof e.className&&e.className||"undefined"!=typeof e.getAttribute&&e.getAttribute("class")||"")})},ATTR:function(n,r,i){return function(e){var t=se.attr(e,n);return null==t?"!="===r:!r||(t+="","="===r?t===i:"!="===r?t!==i:"^="===r?i&&0===t.indexOf(i):"*="===r?i&&-1<t.indexOf(i):"$="===r?i&&t.slice(-i.length)===i:"~="===r?-1<(" "+t.replace(B," ")+" ").indexOf(i):"|="===r&&(t===i||t.slice(0,i.length+1)===i+"-"))}},CHILD:function(h,e,t,g,v){var y="nth"!==h.slice(0,3),m="last"!==h.slice(-4),x="of-type"===e;return 1===g&&0===v?function(e){return!!e.parentNode}:function(e,t,n){var r,i,o,a,s,u,l=y!==m?"nextSibling":"previousSibling",c=e.parentNode,f=x&&e.nodeName.toLowerCase(),p=!n&&!x,d=!1;if(c){if(y){while(l){a=e;while(a=a[l])if(x?a.nodeName.toLowerCase()===f:1===a.nodeType)return!1;u=l="only"===h&&!u&&"nextSibling"}return!0}if(u=[m?c.firstChild:c.lastChild],m&&p){d=(s=(r=(i=(o=(a=c)[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]||[])[0]===k&&r[1])&&r[2],a=s&&c.childNodes[s];while(a=++s&&a&&a[l]||(d=s=0)||u.pop())if(1===a.nodeType&&++d&&a===e){i[h]=[k,s,d];break}}else if(p&&(d=s=(r=(i=(o=(a=e)[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]||[])[0]===k&&r[1]),!1===d)while(a=++s&&a&&a[l]||(d=s=0)||u.pop())if((x?a.nodeName.toLowerCase()===f:1===a.nodeType)&&++d&&(p&&((i=(o=a[S]||(a[S]={}))[a.uniqueID]||(o[a.uniqueID]={}))[h]=[k,d]),a===e))break;return(d-=v)===g||d%g==0&&0<=d/g}}},PSEUDO:function(e,o){var t,a=b.pseudos[e]||b.setFilters[e.toLowerCase()]||se.error("unsupported pseudo: "+e);return a[S]?a(o):1<a.length?(t=[e,e,"",o],b.setFilters.hasOwnProperty(e.toLowerCase())?le(function(e,t){var n,r=a(e,o),i=r.length;while(i--)e[n=P(e,r[i])]=!(t[n]=r[i])}):function(e){return a(e,0,t)}):a}},pseudos:{not:le(function(e){var r=[],i=[],s=f(e.replace($,"$1"));return s[S]?le(function(e,t,n,r){var i,o=s(e,null,r,[]),a=e.length;while(a--)(i=o[a])&&(e[a]=!(t[a]=i))}):function(e,t,n){return r[0]=e,s(r,null,n,i),r[0]=null,!i.pop()}}),has:le(function(t){return function(e){return 0<se(t,e).length}}),contains:le(function(t){return t=t.replace(te,ne),function(e){return-1<(e.textContent||o(e)).indexOf(t)}}),lang:le(function(n){return V.test(n||"")||se.error("unsupported lang: "+n),n=n.replace(te,ne).toLowerCase(),function(e){var t;do{if(t=E?e.lang:e.getAttribute("xml:lang")||e.getAttribute("lang"))return(t=t.toLowerCase())===n||0===t.indexOf(n+"-")}while((e=e.parentNode)&&1===e.nodeType);return!1}}),target:function(e){var t=n.location&&n.location.hash;return t&&t.slice(1)===e.id},root:function(e){return e===a},focus:function(e){return e===C.activeElement&&(!C.hasFocus||C.hasFocus())&&!!(e.type||e.href||~e.tabIndex)},enabled:ge(!1),disabled:ge(!0),checked:function(e){var t=e.nodeName.toLowerCase();return"input"===t&&!!e.checked||"option"===t&&!!e.selected},selected:function(e){return e.parentNode&&e.parentNode.selectedIndex,!0===e.selected},empty:function(e){for(e=e.firstChild;e;e=e.nextSibling)if(e.nodeType<6)return!1;return!0},parent:function(e){return!b.pseudos.empty(e)},header:function(e){return J.test(e.nodeName)},input:function(e){return Q.test(e.nodeName)},button:function(e){var t=e.nodeName.toLowerCase();return"input"===t&&"button"===e.type||"button"===t},text:function(e){var t;return"input"===e.nodeName.toLowerCase()&&"text"===e.type&&(null==(t=e.getAttribute("type"))||"text"===t.toLowerCase())},first:ve(function(){return[0]}),last:ve(function(e,t){return[t-1]}),eq:ve(function(e,t,n){return[n<0?n+t:n]}),even:ve(function(e,t){for(var n=0;n<t;n+=2)e.push(n);return e}),odd:ve(function(e,t){for(var n=1;n<t;n+=2)e.push(n);return e}),lt:ve(function(e,t,n){for(var r=n<0?n+t:t<n?t:n;0<=--r;)e.push(r);return e}),gt:ve(function(e,t,n){for(var r=n<0?n+t:n;++r<t;)e.push(r);return e})}}).pseudos.nth=b.pseudos.eq,{radio:!0,checkbox:!0,file:!0,password:!0,image:!0})b.pseudos[e]=de(e);for(e in{submit:!0,reset:!0})b.pseudos[e]=he(e);function me(){}function xe(e){for(var t=0,n=e.length,r="";t<n;t++)r+=e[t].value;return r}function be(s,e,t){var u=e.dir,l=e.next,c=l||u,f=t&&"parentNode"===c,p=r++;return e.first?function(e,t,n){while(e=e[u])if(1===e.nodeType||f)return s(e,t,n);return!1}:function(e,t,n){var r,i,o,a=[k,p];if(n){while(e=e[u])if((1===e.nodeType||f)&&s(e,t,n))return!0}else while(e=e[u])if(1===e.nodeType||f)if(i=(o=e[S]||(e[S]={}))[e.uniqueID]||(o[e.uniqueID]={}),l&&l===e.nodeName.toLowerCase())e=e[u]||e;else{if((r=i[c])&&r[0]===k&&r[1]===p)return a[2]=r[2];if((i[c]=a)[2]=s(e,t,n))return!0}return!1}}function we(i){return 1<i.length?function(e,t,n){var r=i.length;while(r--)if(!i[r](e,t,n))return!1;return!0}:i[0]}function Te(e,t,n,r,i){for(var o,a=[],s=0,u=e.length,l=null!=t;s<u;s++)(o=e[s])&&(n&&!n(o,r,i)||(a.push(o),l&&t.push(s)));return a}function Ce(d,h,g,v,y,e){return v&&!v[S]&&(v=Ce(v)),y&&!y[S]&&(y=Ce(y,e)),le(function(e,t,n,r){var i,o,a,s=[],u=[],l=t.length,c=e||function(e,t,n){for(var r=0,i=t.length;r<i;r++)se(e,t[r],n);return n}(h||"*",n.nodeType?[n]:n,[]),f=!d||!e&&h?c:Te(c,s,d,n,r),p=g?y||(e?d:l||v)?[]:t:f;if(g&&g(f,p,n,r),v){i=Te(p,u),v(i,[],n,r),o=i.length;while(o--)(a=i[o])&&(p[u[o]]=!(f[u[o]]=a))}if(e){if(y||d){if(y){i=[],o=p.length;while(o--)(a=p[o])&&i.push(f[o]=a);y(null,p=[],i,r)}o=p.length;while(o--)(a=p[o])&&-1<(i=y?P(e,a):s[o])&&(e[i]=!(t[i]=a))}}else p=Te(p===t?p.splice(l,p.length):p),y?y(null,t,p,r):H.apply(t,p)})}function Ee(e){for(var i,t,n,r=e.length,o=b.relative[e[0].type],a=o||b.relative[" "],s=o?1:0,u=be(function(e){return e===i},a,!0),l=be(function(e){return-1<P(i,e)},a,!0),c=[function(e,t,n){var r=!o&&(n||t!==w)||((i=t).nodeType?u(e,t,n):l(e,t,n));return i=null,r}];s<r;s++)if(t=b.relative[e[s].type])c=[be(we(c),t)];else{if((t=b.filter[e[s].type].apply(null,e[s].matches))[S]){for(n=++s;n<r;n++)if(b.relative[e[n].type])break;return Ce(1<s&&we(c),1<s&&xe(e.slice(0,s-1).concat({value:" "===e[s-2].type?"*":""})).replace($,"$1"),t,s<n&&Ee(e.slice(s,n)),n<r&&Ee(e=e.slice(n)),n<r&&xe(e))}c.push(t)}return we(c)}return me.prototype=b.filters=b.pseudos,b.setFilters=new me,h=se.tokenize=function(e,t){var n,r,i,o,a,s,u,l=x[e+" "];if(l)return t?0:l.slice(0);a=e,s=[],u=b.preFilter;while(a){for(o in n&&!(r=_.exec(a))||(r&&(a=a.slice(r[0].length)||a),s.push(i=[])),n=!1,(r=z.exec(a))&&(n=r.shift(),i.push({value:n,type:r[0].replace($," ")}),a=a.slice(n.length)),b.filter)!(r=G[o].exec(a))||u[o]&&!(r=u[o](r))||(n=r.shift(),i.push({value:n,type:o,matches:r}),a=a.slice(n.length));if(!n)break}return t?a.length:a?se.error(e):x(e,s).slice(0)},f=se.compile=function(e,t){var n,v,y,m,x,r,i=[],o=[],a=A[e+" "];if(!a){t||(t=h(e)),n=t.length;while(n--)(a=Ee(t[n]))[S]?i.push(a):o.push(a);(a=A(e,(v=o,m=0<(y=i).length,x=0<v.length,r=function(e,t,n,r,i){var o,a,s,u=0,l="0",c=e&&[],f=[],p=w,d=e||x&&b.find.TAG("*",i),h=k+=null==p?1:Math.random()||.1,g=d.length;for(i&&(w=t==C||t||i);l!==g&&null!=(o=d[l]);l++){if(x&&o){a=0,t||o.ownerDocument==C||(T(o),n=!E);while(s=v[a++])if(s(o,t||C,n)){r.push(o);break}i&&(k=h)}m&&((o=!s&&o)&&u--,e&&c.push(o))}if(u+=l,m&&l!==u){a=0;while(s=y[a++])s(c,f,t,n);if(e){if(0<u)while(l--)c[l]||f[l]||(f[l]=q.call(r));f=Te(f)}H.apply(r,f),i&&!e&&0<f.length&&1<u+y.length&&se.uniqueSort(r)}return i&&(k=h,w=p),c},m?le(r):r))).selector=e}return a},g=se.select=function(e,t,n,r){var i,o,a,s,u,l="function"==typeof e&&e,c=!r&&h(e=l.selector||e);if(n=n||[],1===c.length){if(2<(o=c[0]=c[0].slice(0)).length&&"ID"===(a=o[0]).type&&9===t.nodeType&&E&&b.relative[o[1].type]){if(!(t=(b.find.ID(a.matches[0].replace(te,ne),t)||[])[0]))return n;l&&(t=t.parentNode),e=e.slice(o.shift().value.length)}i=G.needsContext.test(e)?0:o.length;while(i--){if(a=o[i],b.relative[s=a.type])break;if((u=b.find[s])&&(r=u(a.matches[0].replace(te,ne),ee.test(o[0].type)&&ye(t.parentNode)||t))){if(o.splice(i,1),!(e=r.length&&xe(o)))return H.apply(n,r),n;break}}}return(l||f(e,c))(r,t,!E,n,!t||ee.test(e)&&ye(t.parentNode)||t),n},d.sortStable=S.split("").sort(j).join("")===S,d.detectDuplicates=!!l,T(),d.sortDetached=ce(function(e){return 1&e.compareDocumentPosition(C.createElement("fieldset"))}),ce(function(e){return e.innerHTML="<a href='#'></a>","#"===e.firstChild.getAttribute("href")})||fe("type|href|height|width",function(e,t,n){if(!n)return e.getAttribute(t,"type"===t.toLowerCase()?1:2)}),d.attributes&&ce(function(e){return e.innerHTML="<input/>",e.firstChild.setAttribute("value",""),""===e.firstChild.getAttribute("value")})||fe("value",function(e,t,n){if(!n&&"input"===e.nodeName.toLowerCase())return e.defaultValue}),ce(function(e){return null==e.getAttribute("disabled")})||fe(R,function(e,t,n){var r;if(!n)return!0===e[t]?t.toLowerCase():(r=e.getAttributeNode(t))&&r.specified?r.value:null}),se}(C);S.find=d,S.expr=d.selectors,S.expr[":"]=S.expr.pseudos,S.uniqueSort=S.unique=d.uniqueSort,S.text=d.getText,S.isXMLDoc=d.isXML,S.contains=d.contains,S.escapeSelector=d.escape;var h=function(e,t,n){var r=[],i=void 0!==n;while((e=e[t])&&9!==e.nodeType)if(1===e.nodeType){if(i&&S(e).is(n))break;r.push(e)}return r},T=function(e,t){for(var n=[];e;e=e.nextSibling)1===e.nodeType&&e!==t&&n.push(e);return n},k=S.expr.match.needsContext;function A(e,t){return e.nodeName&&e.nodeName.toLowerCase()===t.toLowerCase()}var N=/^<([a-z][^\/\0>:\x20\t\r\n\f]*)[\x20\t\r\n\f]*\/?>(?:<\/\1>|)$/i;function j(e,n,r){return m(n)?S.grep(e,function(e,t){return!!n.call(e,t,e)!==r}):n.nodeType?S.grep(e,function(e){return e===n!==r}):"string"!=typeof n?S.grep(e,function(e){return-1<i.call(n,e)!==r}):S.filter(n,e,r)}S.filter=function(e,t,n){var r=t[0];return n&&(e=":not("+e+")"),1===t.length&&1===r.nodeType?S.find.matchesSelector(r,e)?[r]:[]:S.find.matches(e,S.grep(t,function(e){return 1===e.nodeType}))},S.fn.extend({find:function(e){var t,n,r=this.length,i=this;if("string"!=typeof e)return this.pushStack(S(e).filter(function(){for(t=0;t<r;t++)if(S.contains(i[t],this))return!0}));for(n=this.pushStack([]),t=0;t<r;t++)S.find(e,i[t],n);return 1<r?S.uniqueSort(n):n},filter:function(e){return this.pushStack(j(this,e||[],!1))},not:function(e){return this.pushStack(j(this,e||[],!0))},is:function(e){return!!j(this,"string"==typeof e&&k.test(e)?S(e):e||[],!1).length}});var D,q=/^(?:\s*(<[\w\W]+>)[^>]*|#([\w-]+))$/;(S.fn.init=function(e,t,n){var r,i;if(!e)return this;if(n=n||D,"string"==typeof e){if(!(r="<"===e[0]&&">"===e[e.length-1]&&3<=e.length?[null,e,null]:q.exec(e))||!r[1]&&t)return!t||t.jquery?(t||n).find(e):this.constructor(t).find(e);if(r[1]){if(t=t instanceof S?t[0]:t,S.merge(this,S.parseHTML(r[1],t&&t.nodeType?t.ownerDocument||t:E,!0)),N.test(r[1])&&S.isPlainObject(t))for(r in t)m(this[r])?this[r](t[r]):this.attr(r,t[r]);return this}return(i=E.getElementById(r[2]))&&(this[0]=i,this.length=1),this}return e.nodeType?(this[0]=e,this.length=1,this):m(e)?void 0!==n.ready?n.ready(e):e(S):S.makeArray(e,this)}).prototype=S.fn,D=S(E);var L=/^(?:parents|prev(?:Until|All))/,H={children:!0,contents:!0,next:!0,prev:!0};function O(e,t){while((e=e[t])&&1!==e.nodeType);return e}S.fn.extend({has:function(e){var t=S(e,this),n=t.length;return this.filter(function(){for(var e=0;e<n;e++)if(S.contains(this,t[e]))return!0})},closest:function(e,t){var n,r=0,i=this.length,o=[],a="string"!=typeof e&&S(e);if(!k.test(e))for(;r<i;r++)for(n=this[r];n&&n!==t;n=n.parentNode)if(n.nodeType<11&&(a?-1<a.index(n):1===n.nodeType&&S.find.matchesSelector(n,e))){o.push(n);break}return this.pushStack(1<o.length?S.uniqueSort(o):o)},index:function(e){return e?"string"==typeof e?i.call(S(e),this[0]):i.call(this,e.jquery?e[0]:e):this[0]&&this[0].parentNode?this.first().prevAll().length:-1},add:function(e,t){return this.pushStack(S.uniqueSort(S.merge(this.get(),S(e,t))))},addBack:function(e){return this.add(null==e?this.prevObject:this.prevObject.filter(e))}}),S.each({parent:function(e){var t=e.parentNode;return t&&11!==t.nodeType?t:null},parents:function(e){return h(e,"parentNode")},parentsUntil:function(e,t,n){return h(e,"parentNode",n)},next:function(e){return O(e,"nextSibling")},prev:function(e){return O(e,"previousSibling")},nextAll:function(e){return h(e,"nextSibling")},prevAll:function(e){return h(e,"previousSibling")},nextUntil:function(e,t,n){return h(e,"nextSibling",n)},prevUntil:function(e,t,n){return h(e,"previousSibling",n)},siblings:function(e){return T((e.parentNode||{}).firstChild,e)},children:function(e){return T(e.firstChild)},contents:function(e){return null!=e.contentDocument&&r(e.contentDocument)?e.contentDocument:(A(e,"template")&&(e=e.content||e),S.merge([],e.childNodes))}},function(r,i){S.fn[r]=function(e,t){var n=S.map(this,i,e);return"Until"!==r.slice(-5)&&(t=e),t&&"string"==typeof t&&(n=S.filter(t,n)),1<this.length&&(H[r]||S.uniqueSort(n),L.test(r)&&n.reverse()),this.pushStack(n)}});var P=/[^\x20\t\r\n\f]+/g;function R(e){return e}function M(e){throw e}function I(e,t,n,r){var i;try{e&&m(i=e.promise)?i.call(e).done(t).fail(n):e&&m(i=e.then)?i.call(e,t,n):t.apply(void 0,[e].slice(r))}catch(e){n.apply(void 0,[e])}}S.Callbacks=function(r){var e,n;r="string"==typeof r?(e=r,n={},S.each(e.match(P)||[],function(e,t){n[t]=!0}),n):S.extend({},r);var i,t,o,a,s=[],u=[],l=-1,c=function(){for(a=a||r.once,o=i=!0;u.length;l=-1){t=u.shift();while(++l<s.length)!1===s[l].apply(t[0],t[1])&&r.stopOnFalse&&(l=s.length,t=!1)}r.memory||(t=!1),i=!1,a&&(s=t?[]:"")},f={add:function(){return s&&(t&&!i&&(l=s.length-1,u.push(t)),function n(e){S.each(e,function(e,t){m(t)?r.unique&&f.has(t)||s.push(t):t&&t.length&&"string"!==w(t)&&n(t)})}(arguments),t&&!i&&c()),this},remove:function(){return S.each(arguments,function(e,t){var n;while(-1<(n=S.inArray(t,s,n)))s.splice(n,1),n<=l&&l--}),this},has:function(e){return e?-1<S.inArray(e,s):0<s.length},empty:function(){return s&&(s=[]),this},disable:function(){return a=u=[],s=t="",this},disabled:function(){return!s},lock:function(){return a=u=[],t||i||(s=t=""),this},locked:function(){return!!a},fireWith:function(e,t){return a||(t=[e,(t=t||[]).slice?t.slice():t],u.push(t),i||c()),this},fire:function(){return f.fireWith(this,arguments),this},fired:function(){return!!o}};return f},S.extend({Deferred:function(e){var o=[["notify","progress",S.Callbacks("memory"),S.Callbacks("memory"),2],["resolve","done",S.Callbacks("once memory"),S.Callbacks("once memory"),0,"resolved"],["reject","fail",S.Callbacks("once memory"),S.Callbacks("once memory"),1,"rejected"]],i="pending",a={state:function(){return i},always:function(){return s.done(arguments).fail(arguments),this},"catch":function(e){return a.then(null,e)},pipe:function(){var i=arguments;return S.Deferred(function(r){S.each(o,function(e,t){var n=m(i[t[4]])&&i[t[4]];s[t[1]](function(){var e=n&&n.apply(this,arguments);e&&m(e.promise)?e.promise().progress(r.notify).done(r.resolve).fail(r.reject):r[t[0]+"With"](this,n?[e]:arguments)})}),i=null}).promise()},then:function(t,n,r){var u=0;function l(i,o,a,s){return function(){var n=this,r=arguments,e=function(){var e,t;if(!(i<u)){if((e=a.apply(n,r))===o.promise())throw new TypeError("Thenable self-resolution");t=e&&("object"==typeof e||"function"==typeof e)&&e.then,m(t)?s?t.call(e,l(u,o,R,s),l(u,o,M,s)):(u++,t.call(e,l(u,o,R,s),l(u,o,M,s),l(u,o,R,o.notifyWith))):(a!==R&&(n=void 0,r=[e]),(s||o.resolveWith)(n,r))}},t=s?e:function(){try{e()}catch(e){S.Deferred.exceptionHook&&S.Deferred.exceptionHook(e,t.stackTrace),u<=i+1&&(a!==M&&(n=void 0,r=[e]),o.rejectWith(n,r))}};i?t():(S.Deferred.getStackHook&&(t.stackTrace=S.Deferred.getStackHook()),C.setTimeout(t))}}return S.Deferred(function(e){o[0][3].add(l(0,e,m(r)?r:R,e.notifyWith)),o[1][3].add(l(0,e,m(t)?t:R)),o[2][3].add(l(0,e,m(n)?n:M))}).promise()},promise:function(e){return null!=e?S.extend(e,a):a}},s={};return S.each(o,function(e,t){var n=t[2],r=t[5];a[t[1]]=n.add,r&&n.add(function(){i=r},o[3-e][2].disable,o[3-e][3].disable,o[0][2].lock,o[0][3].lock),n.add(t[3].fire),s[t[0]]=function(){return s[t[0]+"With"](this===s?void 0:this,arguments),this},s[t[0]+"With"]=n.fireWith}),a.promise(s),e&&e.call(s,s),s},when:function(e){var n=arguments.length,t=n,r=Array(t),i=s.call(arguments),o=S.Deferred(),a=function(t){return function(e){r[t]=this,i[t]=1<arguments.length?s.call(arguments):e,--n||o.resolveWith(r,i)}};if(n<=1&&(I(e,o.done(a(t)).resolve,o.reject,!n),"pending"===o.state()||m(i[t]&&i[t].then)))return o.then();while(t--)I(i[t],a(t),o.reject);return o.promise()}});var W=/^(Eval|Internal|Range|Reference|Syntax|Type|URI)Error$/;S.Deferred.exceptionHook=function(e,t){C.console&&C.console.warn&&e&&W.test(e.name)&&C.console.warn("jQuery.Deferred exception: "+e.message,e.stack,t)},S.readyException=function(e){C.setTimeout(function(){throw e})};var F=S.Deferred();function B(){E.removeEventListener("DOMContentLoaded",B),C.removeEventListener("load",B),S.ready()}S.fn.ready=function(e){return F.then(e)["catch"](function(e){S.readyException(e)}),this},S.extend({isReady:!1,readyWait:1,ready:function(e){(!0===e?--S.readyWait:S.isReady)||(S.isReady=!0)!==e&&0<--S.readyWait||F.resolveWith(E,[S])}}),S.ready.then=F.then,"complete"===E.readyState||"loading"!==E.readyState&&!E.documentElement.doScroll?C.setTimeout(S.ready):(E.addEventListener("DOMContentLoaded",B),C.addEventListener("load",B));var $=function(e,t,n,r,i,o,a){var s=0,u=e.length,l=null==n;if("object"===w(n))for(s in i=!0,n)$(e,t,s,n[s],!0,o,a);else if(void 0!==r&&(i=!0,m(r)||(a=!0),l&&(a?(t.call(e,r),t=null):(l=t,t=function(e,t,n){return l.call(S(e),n)})),t))for(;s<u;s++)t(e[s],n,a?r:r.call(e[s],s,t(e[s],n)));return i?e:l?t.call(e):u?t(e[0],n):o},_=/^-ms-/,z=/-([a-z])/g;function U(e,t){return t.toUpperCase()}function X(e){return e.replace(_,"ms-").replace(z,U)}var V=function(e){return 1===e.nodeType||9===e.nodeType||!+e.nodeType};function G(){this.expando=S.expando+G.uid++}G.uid=1,G.prototype={cache:function(e){var t=e[this.expando];return t||(t={},V(e)&&(e.nodeType?e[this.expando]=t:Object.defineProperty(e,this.expando,{value:t,configurable:!0}))),t},set:function(e,t,n){var r,i=this.cache(e);if("string"==typeof t)i[X(t)]=n;else for(r in t)i[X(r)]=t[r];return i},get:function(e,t){return void 0===t?this.cache(e):e[this.expando]&&e[this.expando][X(t)]},access:function(e,t,n){return void 0===t||t&&"string"==typeof t&&void 0===n?this.get(e,t):(this.set(e,t,n),void 0!==n?n:t)},remove:function(e,t){var n,r=e[this.expando];if(void 0!==r){if(void 0!==t){n=(t=Array.isArray(t)?t.map(X):(t=X(t))in r?[t]:t.match(P)||[]).length;while(n--)delete r[t[n]]}(void 0===t||S.isEmptyObject(r))&&(e.nodeType?e[this.expando]=void 0:delete e[this.expando])}},hasData:function(e){var t=e[this.expando];return void 0!==t&&!S.isEmptyObject(t)}};var Y=new G,Q=new G,J=/^(?:\{[\w\W]*\}|\[[\w\W]*\])$/,K=/[A-Z]/g;function Z(e,t,n){var r,i;if(void 0===n&&1===e.nodeType)if(r="data-"+t.replace(K,"-$&").toLowerCase(),"string"==typeof(n=e.getAttribute(r))){try{n="true"===(i=n)||"false"!==i&&("null"===i?null:i===+i+""?+i:J.test(i)?JSON.parse(i):i)}catch(e){}Q.set(e,t,n)}else n=void 0;return n}S.extend({hasData:function(e){return Q.hasData(e)||Y.hasData(e)},data:function(e,t,n){return Q.access(e,t,n)},removeData:function(e,t){Q.remove(e,t)},_data:function(e,t,n){return Y.access(e,t,n)},_removeData:function(e,t){Y.remove(e,t)}}),S.fn.extend({data:function(n,e){var t,r,i,o=this[0],a=o&&o.attributes;if(void 0===n){if(this.length&&(i=Q.get(o),1===o.nodeType&&!Y.get(o,"hasDataAttrs"))){t=a.length;while(t--)a[t]&&0===(r=a[t].name).indexOf("data-")&&(r=X(r.slice(5)),Z(o,r,i[r]));Y.set(o,"hasDataAttrs",!0)}return i}return"object"==typeof n?this.each(function(){Q.set(this,n)}):$(this,function(e){var t;if(o&&void 0===e)return void 0!==(t=Q.get(o,n))?t:void 0!==(t=Z(o,n))?t:void 0;this.each(function(){Q.set(this,n,e)})},null,e,1<arguments.length,null,!0)},removeData:function(e){return this.each(function(){Q.remove(this,e)})}}),S.extend({queue:function(e,t,n){var r;if(e)return t=(t||"fx")+"queue",r=Y.get(e,t),n&&(!r||Array.isArray(n)?r=Y.access(e,t,S.makeArray(n)):r.push(n)),r||[]},dequeue:function(e,t){t=t||"fx";var n=S.queue(e,t),r=n.length,i=n.shift(),o=S._queueHooks(e,t);"inprogress"===i&&(i=n.shift(),r--),i&&("fx"===t&&n.unshift("inprogress"),delete o.stop,i.call(e,function(){S.dequeue(e,t)},o)),!r&&o&&o.empty.fire()},_queueHooks:function(e,t){var n=t+"queueHooks";return Y.get(e,n)||Y.access(e,n,{empty:S.Callbacks("once memory").add(function(){Y.remove(e,[t+"queue",n])})})}}),S.fn.extend({queue:function(t,n){var e=2;return"string"!=typeof t&&(n=t,t="fx",e--),arguments.length<e?S.queue(this[0],t):void 0===n?this:this.each(function(){var e=S.queue(this,t,n);S._queueHooks(this,t),"fx"===t&&"inprogress"!==e[0]&&S.dequeue(this,t)})},dequeue:function(e){return this.each(function(){S.dequeue(this,e)})},clearQueue:function(e){return this.queue(e||"fx",[])},promise:function(e,t){var n,r=1,i=S.Deferred(),o=this,a=this.length,s=function(){--r||i.resolveWith(o,[o])};"string"!=typeof e&&(t=e,e=void 0),e=e||"fx";while(a--)(n=Y.get(o[a],e+"queueHooks"))&&n.empty&&(r++,n.empty.add(s));return s(),i.promise(t)}});var ee=/[+-]?(?:\d*\.|)\d+(?:[eE][+-]?\d+|)/.source,te=new RegExp("^(?:([+-])=|)("+ee+")([a-z%]*)$","i"),ne=["Top","Right","Bottom","Left"],re=E.documentElement,ie=function(e){return S.contains(e.ownerDocument,e)},oe={composed:!0};re.getRootNode&&(ie=function(e){return S.contains(e.ownerDocument,e)||e.getRootNode(oe)===e.ownerDocument});var ae=function(e,t){return"none"===(e=t||e).style.display||""===e.style.display&&ie(e)&&"none"===S.css(e,"display")};function se(e,t,n,r){var i,o,a=20,s=r?function(){return r.cur()}:function(){return S.css(e,t,"")},u=s(),l=n&&n[3]||(S.cssNumber[t]?"":"px"),c=e.nodeType&&(S.cssNumber[t]||"px"!==l&&+u)&&te.exec(S.css(e,t));if(c&&c[3]!==l){u/=2,l=l||c[3],c=+u||1;while(a--)S.style(e,t,c+l),(1-o)*(1-(o=s()/u||.5))<=0&&(a=0),c/=o;c*=2,S.style(e,t,c+l),n=n||[]}return n&&(c=+c||+u||0,i=n[1]?c+(n[1]+1)*n[2]:+n[2],r&&(r.unit=l,r.start=c,r.end=i)),i}var ue={};function le(e,t){for(var n,r,i,o,a,s,u,l=[],c=0,f=e.length;c<f;c++)(r=e[c]).style&&(n=r.style.display,t?("none"===n&&(l[c]=Y.get(r,"display")||null,l[c]||(r.style.display="")),""===r.style.display&&ae(r)&&(l[c]=(u=a=o=void 0,a=(i=r).ownerDocument,s=i.nodeName,(u=ue[s])||(o=a.body.appendChild(a.createElement(s)),u=S.css(o,"display"),o.parentNode.removeChild(o),"none"===u&&(u="block"),ue[s]=u)))):"none"!==n&&(l[c]="none",Y.set(r,"display",n)));for(c=0;c<f;c++)null!=l[c]&&(e[c].style.display=l[c]);return e}S.fn.extend({show:function(){return le(this,!0)},hide:function(){return le(this)},toggle:function(e){return"boolean"==typeof e?e?this.show():this.hide():this.each(function(){ae(this)?S(this).show():S(this).hide()})}});var ce,fe,pe=/^(?:checkbox|radio)$/i,de=/<([a-z][^\/\0>\x20\t\r\n\f]*)/i,he=/^$|^module$|\/(?:java|ecma)script/i;ce=E.createDocumentFragment().appendChild(E.createElement("div")),(fe=E.createElement("input")).setAttribute("type","radio"),fe.setAttribute("checked","checked"),fe.setAttribute("name","t"),ce.appendChild(fe),y.checkClone=ce.cloneNode(!0).cloneNode(!0).lastChild.checked,ce.innerHTML="<textarea>x</textarea>",y.noCloneChecked=!!ce.cloneNode(!0).lastChild.defaultValue,ce.innerHTML="<option></option>",y.option=!!ce.lastChild;var ge={thead:[1,"<table>","</table>"],col:[2,"<table><colgroup>","</colgroup></table>"],tr:[2,"<table><tbody>","</tbody></table>"],td:[3,"<table><tbody><tr>","</tr></tbody></table>"],_default:[0,"",""]};function ve(e,t){var n;return n="undefined"!=typeof e.getElementsByTagName?e.getElementsByTagName(t||"*"):"undefined"!=typeof e.querySelectorAll?e.querySelectorAll(t||"*"):[],void 0===t||t&&A(e,t)?S.merge([e],n):n}function ye(e,t){for(var n=0,r=e.length;n<r;n++)Y.set(e[n],"globalEval",!t||Y.get(t[n],"globalEval"))}ge.tbody=ge.tfoot=ge.colgroup=ge.caption=ge.thead,ge.th=ge.td,y.option||(ge.optgroup=ge.option=[1,"<select multiple='multiple'>","</select>"]);var me=/<|&#?\w+;/;function xe(e,t,n,r,i){for(var o,a,s,u,l,c,f=t.createDocumentFragment(),p=[],d=0,h=e.length;d<h;d++)if((o=e[d])||0===o)if("object"===w(o))S.merge(p,o.nodeType?[o]:o);else if(me.test(o)){a=a||f.appendChild(t.createElement("div")),s=(de.exec(o)||["",""])[1].toLowerCase(),u=ge[s]||ge._default,a.innerHTML=u[1]+S.htmlPrefilter(o)+u[2],c=u[0];while(c--)a=a.lastChild;S.merge(p,a.childNodes),(a=f.firstChild).textContent=""}else p.push(t.createTextNode(o));f.textContent="",d=0;while(o=p[d++])if(r&&-1<S.inArray(o,r))i&&i.push(o);else if(l=ie(o),a=ve(f.appendChild(o),"script"),l&&ye(a),n){c=0;while(o=a[c++])he.test(o.type||"")&&n.push(o)}return f}var be=/^([^.]*)(?:\.(.+)|)/;function we(){return!0}function Te(){return!1}function Ce(e,t){return e===function(){try{return E.activeElement}catch(e){}}()==("focus"===t)}function Ee(e,t,n,r,i,o){var a,s;if("object"==typeof t){for(s in"string"!=typeof n&&(r=r||n,n=void 0),t)Ee(e,s,n,r,t[s],o);return e}if(null==r&&null==i?(i=n,r=n=void 0):null==i&&("string"==typeof n?(i=r,r=void 0):(i=r,r=n,n=void 0)),!1===i)i=Te;else if(!i)return e;return 1===o&&(a=i,(i=function(e){return S().off(e),a.apply(this,arguments)}).guid=a.guid||(a.guid=S.guid++)),e.each(function(){S.event.add(this,t,i,r,n)})}function Se(e,i,o){o?(Y.set(e,i,!1),S.event.add(e,i,{namespace:!1,handler:function(e){var t,n,r=Y.get(this,i);if(1&e.isTrigger&&this[i]){if(r.length)(S.event.special[i]||{}).delegateType&&e.stopPropagation();else if(r=s.call(arguments),Y.set(this,i,r),t=o(this,i),this[i](),r!==(n=Y.get(this,i))||t?Y.set(this,i,!1):n={},r!==n)return e.stopImmediatePropagation(),e.preventDefault(),n&&n.value}else r.length&&(Y.set(this,i,{value:S.event.trigger(S.extend(r[0],S.Event.prototype),r.slice(1),this)}),e.stopImmediatePropagation())}})):void 0===Y.get(e,i)&&S.event.add(e,i,we)}S.event={global:{},add:function(t,e,n,r,i){var o,a,s,u,l,c,f,p,d,h,g,v=Y.get(t);if(V(t)){n.handler&&(n=(o=n).handler,i=o.selector),i&&S.find.matchesSelector(re,i),n.guid||(n.guid=S.guid++),(u=v.events)||(u=v.events=Object.create(null)),(a=v.handle)||(a=v.handle=function(e){return"undefined"!=typeof S&&S.event.triggered!==e.type?S.event.dispatch.apply(t,arguments):void 0}),l=(e=(e||"").match(P)||[""]).length;while(l--)d=g=(s=be.exec(e[l])||[])[1],h=(s[2]||"").split(".").sort(),d&&(f=S.event.special[d]||{},d=(i?f.delegateType:f.bindType)||d,f=S.event.special[d]||{},c=S.extend({type:d,origType:g,data:r,handler:n,guid:n.guid,selector:i,needsContext:i&&S.expr.match.needsContext.test(i),namespace:h.join(".")},o),(p=u[d])||((p=u[d]=[]).delegateCount=0,f.setup&&!1!==f.setup.call(t,r,h,a)||t.addEventListener&&t.addEventListener(d,a)),f.add&&(f.add.call(t,c),c.handler.guid||(c.handler.guid=n.guid)),i?p.splice(p.delegateCount++,0,c):p.push(c),S.event.global[d]=!0)}},remove:function(e,t,n,r,i){var o,a,s,u,l,c,f,p,d,h,g,v=Y.hasData(e)&&Y.get(e);if(v&&(u=v.events)){l=(t=(t||"").match(P)||[""]).length;while(l--)if(d=g=(s=be.exec(t[l])||[])[1],h=(s[2]||"").split(".").sort(),d){f=S.event.special[d]||{},p=u[d=(r?f.delegateType:f.bindType)||d]||[],s=s[2]&&new RegExp("(^|\\.)"+h.join("\\.(?:.*\\.|)")+"(\\.|$)"),a=o=p.length;while(o--)c=p[o],!i&&g!==c.origType||n&&n.guid!==c.guid||s&&!s.test(c.namespace)||r&&r!==c.selector&&("**"!==r||!c.selector)||(p.splice(o,1),c.selector&&p.delegateCount--,f.remove&&f.remove.call(e,c));a&&!p.length&&(f.teardown&&!1!==f.teardown.call(e,h,v.handle)||S.removeEvent(e,d,v.handle),delete u[d])}else for(d in u)S.event.remove(e,d+t[l],n,r,!0);S.isEmptyObject(u)&&Y.remove(e,"handle events")}},dispatch:function(e){var t,n,r,i,o,a,s=new Array(arguments.length),u=S.event.fix(e),l=(Y.get(this,"events")||Object.create(null))[u.type]||[],c=S.event.special[u.type]||{};for(s[0]=u,t=1;t<arguments.length;t++)s[t]=arguments[t];if(u.delegateTarget=this,!c.preDispatch||!1!==c.preDispatch.call(this,u)){a=S.event.handlers.call(this,u,l),t=0;while((i=a[t++])&&!u.isPropagationStopped()){u.currentTarget=i.elem,n=0;while((o=i.handlers[n++])&&!u.isImmediatePropagationStopped())u.rnamespace&&!1!==o.namespace&&!u.rnamespace.test(o.namespace)||(u.handleObj=o,u.data=o.data,void 0!==(r=((S.event.special[o.origType]||{}).handle||o.handler).apply(i.elem,s))&&!1===(u.result=r)&&(u.preventDefault(),u.stopPropagation()))}return c.postDispatch&&c.postDispatch.call(this,u),u.result}},handlers:function(e,t){var n,r,i,o,a,s=[],u=t.delegateCount,l=e.target;if(u&&l.nodeType&&!("click"===e.type&&1<=e.button))for(;l!==this;l=l.parentNode||this)if(1===l.nodeType&&("click"!==e.type||!0!==l.disabled)){for(o=[],a={},n=0;n<u;n++)void 0===a[i=(r=t[n]).selector+" "]&&(a[i]=r.needsContext?-1<S(i,this).index(l):S.find(i,this,null,[l]).length),a[i]&&o.push(r);o.length&&s.push({elem:l,handlers:o})}return l=this,u<t.length&&s.push({elem:l,handlers:t.slice(u)}),s},addProp:function(t,e){Object.defineProperty(S.Event.prototype,t,{enumerable:!0,configurable:!0,get:m(e)?function(){if(this.originalEvent)return e(this.originalEvent)}:function(){if(this.originalEvent)return this.originalEvent[t]},set:function(e){Object.defineProperty(this,t,{enumerable:!0,configurable:!0,writable:!0,value:e})}})},fix:function(e){return e[S.expando]?e:new S.Event(e)},special:{load:{noBubble:!0},click:{setup:function(e){var t=this||e;return pe.test(t.type)&&t.click&&A(t,"input")&&Se(t,"click",we),!1},trigger:function(e){var t=this||e;return pe.test(t.type)&&t.click&&A(t,"input")&&Se(t,"click"),!0},_default:function(e){var t=e.target;return pe.test(t.type)&&t.click&&A(t,"input")&&Y.get(t,"click")||A(t,"a")}},beforeunload:{postDispatch:function(e){void 0!==e.result&&e.originalEvent&&(e.originalEvent.returnValue=e.result)}}}},S.removeEvent=function(e,t,n){e.removeEventListener&&e.removeEventListener(t,n)},S.Event=function(e,t){if(!(this instanceof S.Event))return new S.Event(e,t);e&&e.type?(this.originalEvent=e,this.type=e.type,this.isDefaultPrevented=e.defaultPrevented||void 0===e.defaultPrevented&&!1===e.returnValue?we:Te,this.target=e.target&&3===e.target.nodeType?e.target.parentNode:e.target,this.currentTarget=e.currentTarget,this.relatedTarget=e.relatedTarget):this.type=e,t&&S.extend(this,t),this.timeStamp=e&&e.timeStamp||Date.now(),this[S.expando]=!0},S.Event.prototype={constructor:S.Event,isDefaultPrevented:Te,isPropagationStopped:Te,isImmediatePropagationStopped:Te,isSimulated:!1,preventDefault:function(){var e=this.originalEvent;this.isDefaultPrevented=we,e&&!this.isSimulated&&e.preventDefault()},stopPropagation:function(){var e=this.originalEvent;this.isPropagationStopped=we,e&&!this.isSimulated&&e.stopPropagation()},stopImmediatePropagation:function(){var e=this.originalEvent;this.isImmediatePropagationStopped=we,e&&!this.isSimulated&&e.stopImmediatePropagation(),this.stopPropagation()}},S.each({altKey:!0,bubbles:!0,cancelable:!0,changedTouches:!0,ctrlKey:!0,detail:!0,eventPhase:!0,metaKey:!0,pageX:!0,pageY:!0,shiftKey:!0,view:!0,"char":!0,code:!0,charCode:!0,key:!0,keyCode:!0,button:!0,buttons:!0,clientX:!0,clientY:!0,offsetX:!0,offsetY:!0,pointerId:!0,pointerType:!0,screenX:!0,screenY:!0,targetTouches:!0,toElement:!0,touches:!0,which:!0},S.event.addProp),S.each({focus:"focusin",blur:"focusout"},function(e,t){S.event.special[e]={setup:function(){return Se(this,e,Ce),!1},trigger:function(){return Se(this,e),!0},_default:function(){return!0},delegateType:t}}),S.each({mouseenter:"mouseover",mouseleave:"mouseout",pointerenter:"pointerover",pointerleave:"pointerout"},function(e,i){S.event.special[e]={delegateType:i,bindType:i,handle:function(e){var t,n=e.relatedTarget,r=e.handleObj;return n&&(n===this||S.contains(this,n))||(e.type=r.origType,t=r.handler.apply(this,arguments),e.type=i),t}}}),S.fn.extend({on:function(e,t,n,r){return Ee(this,e,t,n,r)},one:function(e,t,n,r){return Ee(this,e,t,n,r,1)},off:function(e,t,n){var r,i;if(e&&e.preventDefault&&e.handleObj)return r=e.handleObj,S(e.delegateTarget).off(r.namespace?r.origType+"."+r.namespace:r.origType,r.selector,r.handler),this;if("object"==typeof e){for(i in e)this.off(i,t,e[i]);return this}return!1!==t&&"function"!=typeof t||(n=t,t=void 0),!1===n&&(n=Te),this.each(function(){S.event.remove(this,e,n,t)})}});var ke=/<script|<style|<link/i,Ae=/checked\s*(?:[^=]|=\s*.checked.)/i,Ne=/^\s*<!(?:\[CDATA\[|--)|(?:\]\]|--)>\s*$/g;function je(e,t){return A(e,"table")&&A(11!==t.nodeType?t:t.firstChild,"tr")&&S(e).children("tbody")[0]||e}function De(e){return e.type=(null!==e.getAttribute("type"))+"/"+e.type,e}function qe(e){return"true/"===(e.type||"").slice(0,5)?e.type=e.type.slice(5):e.removeAttribute("type"),e}function Le(e,t){var n,r,i,o,a,s;if(1===t.nodeType){if(Y.hasData(e)&&(s=Y.get(e).events))for(i in Y.remove(t,"handle events"),s)for(n=0,r=s[i].length;n<r;n++)S.event.add(t,i,s[i][n]);Q.hasData(e)&&(o=Q.access(e),a=S.extend({},o),Q.set(t,a))}}function He(n,r,i,o){r=g(r);var e,t,a,s,u,l,c=0,f=n.length,p=f-1,d=r[0],h=m(d);if(h||1<f&&"string"==typeof d&&!y.checkClone&&Ae.test(d))return n.each(function(e){var t=n.eq(e);h&&(r[0]=d.call(this,e,t.html())),He(t,r,i,o)});if(f&&(t=(e=xe(r,n[0].ownerDocument,!1,n,o)).firstChild,1===e.childNodes.length&&(e=t),t||o)){for(s=(a=S.map(ve(e,"script"),De)).length;c<f;c++)u=e,c!==p&&(u=S.clone(u,!0,!0),s&&S.merge(a,ve(u,"script"))),i.call(n[c],u,c);if(s)for(l=a[a.length-1].ownerDocument,S.map(a,qe),c=0;c<s;c++)u=a[c],he.test(u.type||"")&&!Y.access(u,"globalEval")&&S.contains(l,u)&&(u.src&&"module"!==(u.type||"").toLowerCase()?S._evalUrl&&!u.noModule&&S._evalUrl(u.src,{nonce:u.nonce||u.getAttribute("nonce")},l):b(u.textContent.replace(Ne,""),u,l))}return n}function Oe(e,t,n){for(var r,i=t?S.filter(t,e):e,o=0;null!=(r=i[o]);o++)n||1!==r.nodeType||S.cleanData(ve(r)),r.parentNode&&(n&&ie(r)&&ye(ve(r,"script")),r.parentNode.removeChild(r));return e}S.extend({htmlPrefilter:function(e){return e},clone:function(e,t,n){var r,i,o,a,s,u,l,c=e.cloneNode(!0),f=ie(e);if(!(y.noCloneChecked||1!==e.nodeType&&11!==e.nodeType||S.isXMLDoc(e)))for(a=ve(c),r=0,i=(o=ve(e)).length;r<i;r++)s=o[r],u=a[r],void 0,"input"===(l=u.nodeName.toLowerCase())&&pe.test(s.type)?u.checked=s.checked:"input"!==l&&"textarea"!==l||(u.defaultValue=s.defaultValue);if(t)if(n)for(o=o||ve(e),a=a||ve(c),r=0,i=o.length;r<i;r++)Le(o[r],a[r]);else Le(e,c);return 0<(a=ve(c,"script")).length&&ye(a,!f&&ve(e,"script")),c},cleanData:function(e){for(var t,n,r,i=S.event.special,o=0;void 0!==(n=e[o]);o++)if(V(n)){if(t=n[Y.expando]){if(t.events)for(r in t.events)i[r]?S.event.remove(n,r):S.removeEvent(n,r,t.handle);n[Y.expando]=void 0}n[Q.expando]&&(n[Q.expando]=void 0)}}}),S.fn.extend({detach:function(e){return Oe(this,e,!0)},remove:function(e){return Oe(this,e)},text:function(e){return $(this,function(e){return void 0===e?S.text(this):this.empty().each(function(){1!==this.nodeType&&11!==this.nodeType&&9!==this.nodeType||(this.textContent=e)})},null,e,arguments.length)},append:function(){return He(this,arguments,function(e){1!==this.nodeType&&11!==this.nodeType&&9!==this.nodeType||je(this,e).appendChild(e)})},prepend:function(){return He(this,arguments,function(e){if(1===this.nodeType||11===this.nodeType||9===this.nodeType){var t=je(this,e);t.insertBefore(e,t.firstChild)}})},before:function(){return He(this,arguments,function(e){this.parentNode&&this.parentNode.insertBefore(e,this)})},after:function(){return He(this,arguments,function(e){this.parentNode&&this.parentNode.insertBefore(e,this.nextSibling)})},empty:function(){for(var e,t=0;null!=(e=this[t]);t++)1===e.nodeType&&(S.cleanData(ve(e,!1)),e.textContent="");return this},clone:function(e,t){return e=null!=e&&e,t=null==t?e:t,this.map(function(){return S.clone(this,e,t)})},html:function(e){return $(this,function(e){var t=this[0]||{},n=0,r=this.length;if(void 0===e&&1===t.nodeType)return t.innerHTML;if("string"==typeof e&&!ke.test(e)&&!ge[(de.exec(e)||["",""])[1].toLowerCase()]){e=S.htmlPrefilter(e);try{for(;n<r;n++)1===(t=this[n]||{}).nodeType&&(S.cleanData(ve(t,!1)),t.innerHTML=e);t=0}catch(e){}}t&&this.empty().append(e)},null,e,arguments.length)},replaceWith:function(){var n=[];return He(this,arguments,function(e){var t=this.parentNode;S.inArray(this,n)<0&&(S.cleanData(ve(this)),t&&t.replaceChild(e,this))},n)}}),S.each({appendTo:"append",prependTo:"prepend",insertBefore:"before",insertAfter:"after",replaceAll:"replaceWith"},function(e,a){S.fn[e]=function(e){for(var t,n=[],r=S(e),i=r.length-1,o=0;o<=i;o++)t=o===i?this:this.clone(!0),S(r[o])[a](t),u.apply(n,t.get());return this.pushStack(n)}});var Pe=new RegExp("^("+ee+")(?!px)[a-z%]+$","i"),Re=function(e){var t=e.ownerDocument.defaultView;return t&&t.opener||(t=C),t.getComputedStyle(e)},Me=function(e,t,n){var r,i,o={};for(i in t)o[i]=e.style[i],e.style[i]=t[i];for(i in r=n.call(e),t)e.style[i]=o[i];return r},Ie=new RegExp(ne.join("|"),"i");function We(e,t,n){var r,i,o,a,s=e.style;return(n=n||Re(e))&&(""!==(a=n.getPropertyValue(t)||n[t])||ie(e)||(a=S.style(e,t)),!y.pixelBoxStyles()&&Pe.test(a)&&Ie.test(t)&&(r=s.width,i=s.minWidth,o=s.maxWidth,s.minWidth=s.maxWidth=s.width=a,a=n.width,s.width=r,s.minWidth=i,s.maxWidth=o)),void 0!==a?a+"":a}function Fe(e,t){return{get:function(){if(!e())return(this.get=t).apply(this,arguments);delete this.get}}}!function(){function e(){if(l){u.style.cssText="position:absolute;left:-11111px;width:60px;margin-top:1px;padding:0;border:0",l.style.cssText="position:relative;display:block;box-sizing:border-box;overflow:scroll;margin:auto;border:1px;padding:1px;width:60%;top:1%",re.appendChild(u).appendChild(l);var e=C.getComputedStyle(l);n="1%"!==e.top,s=12===t(e.marginLeft),l.style.right="60%",o=36===t(e.right),r=36===t(e.width),l.style.position="absolute",i=12===t(l.offsetWidth/3),re.removeChild(u),l=null}}function t(e){return Math.round(parseFloat(e))}var n,r,i,o,a,s,u=E.createElement("div"),l=E.createElement("div");l.style&&(l.style.backgroundClip="content-box",l.cloneNode(!0).style.backgroundClip="",y.clearCloneStyle="content-box"===l.style.backgroundClip,S.extend(y,{boxSizingReliable:function(){return e(),r},pixelBoxStyles:function(){return e(),o},pixelPosition:function(){return e(),n},reliableMarginLeft:function(){return e(),s},scrollboxSize:function(){return e(),i},reliableTrDimensions:function(){var e,t,n,r;return null==a&&(e=E.createElement("table"),t=E.createElement("tr"),n=E.createElement("div"),e.style.cssText="position:absolute;left:-11111px;border-collapse:separate",t.style.cssText="border:1px solid",t.style.height="1px",n.style.height="9px",n.style.display="block",re.appendChild(e).appendChild(t).appendChild(n),r=C.getComputedStyle(t),a=parseInt(r.height,10)+parseInt(r.borderTopWidth,10)+parseInt(r.borderBottomWidth,10)===t.offsetHeight,re.removeChild(e)),a}}))}();var Be=["Webkit","Moz","ms"],$e=E.createElement("div").style,_e={};function ze(e){var t=S.cssProps[e]||_e[e];return t||(e in $e?e:_e[e]=function(e){var t=e[0].toUpperCase()+e.slice(1),n=Be.length;while(n--)if((e=Be[n]+t)in $e)return e}(e)||e)}var Ue=/^(none|table(?!-c[ea]).+)/,Xe=/^--/,Ve={position:"absolute",visibility:"hidden",display:"block"},Ge={letterSpacing:"0",fontWeight:"400"};function Ye(e,t,n){var r=te.exec(t);return r?Math.max(0,r[2]-(n||0))+(r[3]||"px"):t}function Qe(e,t,n,r,i,o){var a="width"===t?1:0,s=0,u=0;if(n===(r?"border":"content"))return 0;for(;a<4;a+=2)"margin"===n&&(u+=S.css(e,n+ne[a],!0,i)),r?("content"===n&&(u-=S.css(e,"padding"+ne[a],!0,i)),"margin"!==n&&(u-=S.css(e,"border"+ne[a]+"Width",!0,i))):(u+=S.css(e,"padding"+ne[a],!0,i),"padding"!==n?u+=S.css(e,"border"+ne[a]+"Width",!0,i):s+=S.css(e,"border"+ne[a]+"Width",!0,i));return!r&&0<=o&&(u+=Math.max(0,Math.ceil(e["offset"+t[0].toUpperCase()+t.slice(1)]-o-u-s-.5))||0),u}function Je(e,t,n){var r=Re(e),i=(!y.boxSizingReliable()||n)&&"border-box"===S.css(e,"boxSizing",!1,r),o=i,a=We(e,t,r),s="offset"+t[0].toUpperCase()+t.slice(1);if(Pe.test(a)){if(!n)return a;a="auto"}return(!y.boxSizingReliable()&&i||!y.reliableTrDimensions()&&A(e,"tr")||"auto"===a||!parseFloat(a)&&"inline"===S.css(e,"display",!1,r))&&e.getClientRects().length&&(i="border-box"===S.css(e,"boxSizing",!1,r),(o=s in e)&&(a=e[s])),(a=parseFloat(a)||0)+Qe(e,t,n||(i?"border":"content"),o,r,a)+"px"}function Ke(e,t,n,r,i){return new Ke.prototype.init(e,t,n,r,i)}S.extend({cssHooks:{opacity:{get:function(e,t){if(t){var n=We(e,"opacity");return""===n?"1":n}}}},cssNumber:{animationIterationCount:!0,columnCount:!0,fillOpacity:!0,flexGrow:!0,flexShrink:!0,fontWeight:!0,gridArea:!0,gridColumn:!0,gridColumnEnd:!0,gridColumnStart:!0,gridRow:!0,gridRowEnd:!0,gridRowStart:!0,lineHeight:!0,opacity:!0,order:!0,orphans:!0,widows:!0,zIndex:!0,zoom:!0},cssProps:{},style:function(e,t,n,r){if(e&&3!==e.nodeType&&8!==e.nodeType&&e.style){var i,o,a,s=X(t),u=Xe.test(t),l=e.style;if(u||(t=ze(s)),a=S.cssHooks[t]||S.cssHooks[s],void 0===n)return a&&"get"in a&&void 0!==(i=a.get(e,!1,r))?i:l[t];"string"===(o=typeof n)&&(i=te.exec(n))&&i[1]&&(n=se(e,t,i),o="number"),null!=n&&n==n&&("number"!==o||u||(n+=i&&i[3]||(S.cssNumber[s]?"":"px")),y.clearCloneStyle||""!==n||0!==t.indexOf("background")||(l[t]="inherit"),a&&"set"in a&&void 0===(n=a.set(e,n,r))||(u?l.setProperty(t,n):l[t]=n))}},css:function(e,t,n,r){var i,o,a,s=X(t);return Xe.test(t)||(t=ze(s)),(a=S.cssHooks[t]||S.cssHooks[s])&&"get"in a&&(i=a.get(e,!0,n)),void 0===i&&(i=We(e,t,r)),"normal"===i&&t in Ge&&(i=Ge[t]),""===n||n?(o=parseFloat(i),!0===n||isFinite(o)?o||0:i):i}}),S.each(["height","width"],function(e,u){S.cssHooks[u]={get:function(e,t,n){if(t)return!Ue.test(S.css(e,"display"))||e.getClientRects().length&&e.getBoundingClientRect().width?Je(e,u,n):Me(e,Ve,function(){return Je(e,u,n)})},set:function(e,t,n){var r,i=Re(e),o=!y.scrollboxSize()&&"absolute"===i.position,a=(o||n)&&"border-box"===S.css(e,"boxSizing",!1,i),s=n?Qe(e,u,n,a,i):0;return a&&o&&(s-=Math.ceil(e["offset"+u[0].toUpperCase()+u.slice(1)]-parseFloat(i[u])-Qe(e,u,"border",!1,i)-.5)),s&&(r=te.exec(t))&&"px"!==(r[3]||"px")&&(e.style[u]=t,t=S.css(e,u)),Ye(0,t,s)}}}),S.cssHooks.marginLeft=Fe(y.reliableMarginLeft,function(e,t){if(t)return(parseFloat(We(e,"marginLeft"))||e.getBoundingClientRect().left-Me(e,{marginLeft:0},function(){return e.getBoundingClientRect().left}))+"px"}),S.each({margin:"",padding:"",border:"Width"},function(i,o){S.cssHooks[i+o]={expand:function(e){for(var t=0,n={},r="string"==typeof e?e.split(" "):[e];t<4;t++)n[i+ne[t]+o]=r[t]||r[t-2]||r[0];return n}},"margin"!==i&&(S.cssHooks[i+o].set=Ye)}),S.fn.extend({css:function(e,t){return $(this,function(e,t,n){var r,i,o={},a=0;if(Array.isArray(t)){for(r=Re(e),i=t.length;a<i;a++)o[t[a]]=S.css(e,t[a],!1,r);return o}return void 0!==n?S.style(e,t,n):S.css(e,t)},e,t,1<arguments.length)}}),((S.Tween=Ke).prototype={constructor:Ke,init:function(e,t,n,r,i,o){this.elem=e,this.prop=n,this.easing=i||S.easing._default,this.options=t,this.start=this.now=this.cur(),this.end=r,this.unit=o||(S.cssNumber[n]?"":"px")},cur:function(){var e=Ke.propHooks[this.prop];return e&&e.get?e.get(this):Ke.propHooks._default.get(this)},run:function(e){var t,n=Ke.propHooks[this.prop];return this.options.duration?this.pos=t=S.easing[this.easing](e,this.options.duration*e,0,1,this.options.duration):this.pos=t=e,this.now=(this.end-this.start)*t+this.start,this.options.step&&this.options.step.call(this.elem,this.now,this),n&&n.set?n.set(this):Ke.propHooks._default.set(this),this}}).init.prototype=Ke.prototype,(Ke.propHooks={_default:{get:function(e){var t;return 1!==e.elem.nodeType||null!=e.elem[e.prop]&&null==e.elem.style[e.prop]?e.elem[e.prop]:(t=S.css(e.elem,e.prop,""))&&"auto"!==t?t:0},set:function(e){S.fx.step[e.prop]?S.fx.step[e.prop](e):1!==e.elem.nodeType||!S.cssHooks[e.prop]&&null==e.elem.style[ze(e.prop)]?e.elem[e.prop]=e.now:S.style(e.elem,e.prop,e.now+e.unit)}}}).scrollTop=Ke.propHooks.scrollLeft={set:function(e){e.elem.nodeType&&e.elem.parentNode&&(e.elem[e.prop]=e.now)}},S.easing={linear:function(e){return e},swing:function(e){return.5-Math.cos(e*Math.PI)/2},_default:"swing"},S.fx=Ke.prototype.init,S.fx.step={};var Ze,et,tt,nt,rt=/^(?:toggle|show|hide)$/,it=/queueHooks$/;function ot(){et&&(!1===E.hidden&&C.requestAnimationFrame?C.requestAnimationFrame(ot):C.setTimeout(ot,S.fx.interval),S.fx.tick())}function at(){return C.setTimeout(function(){Ze=void 0}),Ze=Date.now()}function st(e,t){var n,r=0,i={height:e};for(t=t?1:0;r<4;r+=2-t)i["margin"+(n=ne[r])]=i["padding"+n]=e;return t&&(i.opacity=i.width=e),i}function ut(e,t,n){for(var r,i=(lt.tweeners[t]||[]).concat(lt.tweeners["*"]),o=0,a=i.length;o<a;o++)if(r=i[o].call(n,t,e))return r}function lt(o,e,t){var n,a,r=0,i=lt.prefilters.length,s=S.Deferred().always(function(){delete u.elem}),u=function(){if(a)return!1;for(var e=Ze||at(),t=Math.max(0,l.startTime+l.duration-e),n=1-(t/l.duration||0),r=0,i=l.tweens.length;r<i;r++)l.tweens[r].run(n);return s.notifyWith(o,[l,n,t]),n<1&&i?t:(i||s.notifyWith(o,[l,1,0]),s.resolveWith(o,[l]),!1)},l=s.promise({elem:o,props:S.extend({},e),opts:S.extend(!0,{specialEasing:{},easing:S.easing._default},t),originalProperties:e,originalOptions:t,startTime:Ze||at(),duration:t.duration,tweens:[],createTween:function(e,t){var n=S.Tween(o,l.opts,e,t,l.opts.specialEasing[e]||l.opts.easing);return l.tweens.push(n),n},stop:function(e){var t=0,n=e?l.tweens.length:0;if(a)return this;for(a=!0;t<n;t++)l.tweens[t].run(1);return e?(s.notifyWith(o,[l,1,0]),s.resolveWith(o,[l,e])):s.rejectWith(o,[l,e]),this}}),c=l.props;for(!function(e,t){var n,r,i,o,a;for(n in e)if(i=t[r=X(n)],o=e[n],Array.isArray(o)&&(i=o[1],o=e[n]=o[0]),n!==r&&(e[r]=o,delete e[n]),(a=S.cssHooks[r])&&"expand"in a)for(n in o=a.expand(o),delete e[r],o)n in e||(e[n]=o[n],t[n]=i);else t[r]=i}(c,l.opts.specialEasing);r<i;r++)if(n=lt.prefilters[r].call(l,o,c,l.opts))return m(n.stop)&&(S._queueHooks(l.elem,l.opts.queue).stop=n.stop.bind(n)),n;return S.map(c,ut,l),m(l.opts.start)&&l.opts.start.call(o,l),l.progress(l.opts.progress).done(l.opts.done,l.opts.complete).fail(l.opts.fail).always(l.opts.always),S.fx.timer(S.extend(u,{elem:o,anim:l,queue:l.opts.queue})),l}S.Animation=S.extend(lt,{tweeners:{"*":[function(e,t){var n=this.createTween(e,t);return se(n.elem,e,te.exec(t),n),n}]},tweener:function(e,t){m(e)?(t=e,e=["*"]):e=e.match(P);for(var n,r=0,i=e.length;r<i;r++)n=e[r],lt.tweeners[n]=lt.tweeners[n]||[],lt.tweeners[n].unshift(t)},prefilters:[function(e,t,n){var r,i,o,a,s,u,l,c,f="width"in t||"height"in t,p=this,d={},h=e.style,g=e.nodeType&&ae(e),v=Y.get(e,"fxshow");for(r in n.queue||(null==(a=S._queueHooks(e,"fx")).unqueued&&(a.unqueued=0,s=a.empty.fire,a.empty.fire=function(){a.unqueued||s()}),a.unqueued++,p.always(function(){p.always(function(){a.unqueued--,S.queue(e,"fx").length||a.empty.fire()})})),t)if(i=t[r],rt.test(i)){if(delete t[r],o=o||"toggle"===i,i===(g?"hide":"show")){if("show"!==i||!v||void 0===v[r])continue;g=!0}d[r]=v&&v[r]||S.style(e,r)}if((u=!S.isEmptyObject(t))||!S.isEmptyObject(d))for(r in f&&1===e.nodeType&&(n.overflow=[h.overflow,h.overflowX,h.overflowY],null==(l=v&&v.display)&&(l=Y.get(e,"display")),"none"===(c=S.css(e,"display"))&&(l?c=l:(le([e],!0),l=e.style.display||l,c=S.css(e,"display"),le([e]))),("inline"===c||"inline-block"===c&&null!=l)&&"none"===S.css(e,"float")&&(u||(p.done(function(){h.display=l}),null==l&&(c=h.display,l="none"===c?"":c)),h.display="inline-block")),n.overflow&&(h.overflow="hidden",p.always(function(){h.overflow=n.overflow[0],h.overflowX=n.overflow[1],h.overflowY=n.overflow[2]})),u=!1,d)u||(v?"hidden"in v&&(g=v.hidden):v=Y.access(e,"fxshow",{display:l}),o&&(v.hidden=!g),g&&le([e],!0),p.done(function(){for(r in g||le([e]),Y.remove(e,"fxshow"),d)S.style(e,r,d[r])})),u=ut(g?v[r]:0,r,p),r in v||(v[r]=u.start,g&&(u.end=u.start,u.start=0))}],prefilter:function(e,t){t?lt.prefilters.unshift(e):lt.prefilters.push(e)}}),S.speed=function(e,t,n){var r=e&&"object"==typeof e?S.extend({},e):{complete:n||!n&&t||m(e)&&e,duration:e,easing:n&&t||t&&!m(t)&&t};return S.fx.off?r.duration=0:"number"!=typeof r.duration&&(r.duration in S.fx.speeds?r.duration=S.fx.speeds[r.duration]:r.duration=S.fx.speeds._default),null!=r.queue&&!0!==r.queue||(r.queue="fx"),r.old=r.complete,r.complete=function(){m(r.old)&&r.old.call(this),r.queue&&S.dequeue(this,r.queue)},r},S.fn.extend({fadeTo:function(e,t,n,r){return this.filter(ae).css("opacity",0).show().end().animate({opacity:t},e,n,r)},animate:function(t,e,n,r){var i=S.isEmptyObject(t),o=S.speed(e,n,r),a=function(){var e=lt(this,S.extend({},t),o);(i||Y.get(this,"finish"))&&e.stop(!0)};return a.finish=a,i||!1===o.queue?this.each(a):this.queue(o.queue,a)},stop:function(i,e,o){var a=function(e){var t=e.stop;delete e.stop,t(o)};return"string"!=typeof i&&(o=e,e=i,i=void 0),e&&this.queue(i||"fx",[]),this.each(function(){var e=!0,t=null!=i&&i+"queueHooks",n=S.timers,r=Y.get(this);if(t)r[t]&&r[t].stop&&a(r[t]);else for(t in r)r[t]&&r[t].stop&&it.test(t)&&a(r[t]);for(t=n.length;t--;)n[t].elem!==this||null!=i&&n[t].queue!==i||(n[t].anim.stop(o),e=!1,n.splice(t,1));!e&&o||S.dequeue(this,i)})},finish:function(a){return!1!==a&&(a=a||"fx"),this.each(function(){var e,t=Y.get(this),n=t[a+"queue"],r=t[a+"queueHooks"],i=S.timers,o=n?n.length:0;for(t.finish=!0,S.queue(this,a,[]),r&&r.stop&&r.stop.call(this,!0),e=i.length;e--;)i[e].elem===this&&i[e].queue===a&&(i[e].anim.stop(!0),i.splice(e,1));for(e=0;e<o;e++)n[e]&&n[e].finish&&n[e].finish.call(this);delete t.finish})}}),S.each(["toggle","show","hide"],function(e,r){var i=S.fn[r];S.fn[r]=function(e,t,n){return null==e||"boolean"==typeof e?i.apply(this,arguments):this.animate(st(r,!0),e,t,n)}}),S.each({slideDown:st("show"),slideUp:st("hide"),slideToggle:st("toggle"),fadeIn:{opacity:"show"},fadeOut:{opacity:"hide"},fadeToggle:{opacity:"toggle"}},function(e,r){S.fn[e]=function(e,t,n){return this.animate(r,e,t,n)}}),S.timers=[],S.fx.tick=function(){var e,t=0,n=S.timers;for(Ze=Date.now();t<n.length;t++)(e=n[t])()||n[t]!==e||n.splice(t--,1);n.length||S.fx.stop(),Ze=void 0},S.fx.timer=function(e){S.timers.push(e),S.fx.start()},S.fx.interval=13,S.fx.start=function(){et||(et=!0,ot())},S.fx.stop=function(){et=null},S.fx.speeds={slow:600,fast:200,_default:400},S.fn.delay=function(r,e){return r=S.fx&&S.fx.speeds[r]||r,e=e||"fx",this.queue(e,function(e,t){var n=C.setTimeout(e,r);t.stop=function(){C.clearTimeout(n)}})},tt=E.createElement("input"),nt=E.createElement("select").appendChild(E.createElement("option")),tt.type="checkbox",y.checkOn=""!==tt.value,y.optSelected=nt.selected,(tt=E.createElement("input")).value="t",tt.type="radio",y.radioValue="t"===tt.value;var ct,ft=S.expr.attrHandle;S.fn.extend({attr:function(e,t){return $(this,S.attr,e,t,1<arguments.length)},removeAttr:function(e){return this.each(function(){S.removeAttr(this,e)})}}),S.extend({attr:function(e,t,n){var r,i,o=e.nodeType;if(3!==o&&8!==o&&2!==o)return"undefined"==typeof e.getAttribute?S.prop(e,t,n):(1===o&&S.isXMLDoc(e)||(i=S.attrHooks[t.toLowerCase()]||(S.expr.match.bool.test(t)?ct:void 0)),void 0!==n?null===n?void S.removeAttr(e,t):i&&"set"in i&&void 0!==(r=i.set(e,n,t))?r:(e.setAttribute(t,n+""),n):i&&"get"in i&&null!==(r=i.get(e,t))?r:null==(r=S.find.attr(e,t))?void 0:r)},attrHooks:{type:{set:function(e,t){if(!y.radioValue&&"radio"===t&&A(e,"input")){var n=e.value;return e.setAttribute("type",t),n&&(e.value=n),t}}}},removeAttr:function(e,t){var n,r=0,i=t&&t.match(P);if(i&&1===e.nodeType)while(n=i[r++])e.removeAttribute(n)}}),ct={set:function(e,t,n){return!1===t?S.removeAttr(e,n):e.setAttribute(n,n),n}},S.each(S.expr.match.bool.source.match(/\w+/g),function(e,t){var a=ft[t]||S.find.attr;ft[t]=function(e,t,n){var r,i,o=t.toLowerCase();return n||(i=ft[o],ft[o]=r,r=null!=a(e,t,n)?o:null,ft[o]=i),r}});var pt=/^(?:input|select|textarea|button)$/i,dt=/^(?:a|area)$/i;function ht(e){return(e.match(P)||[]).join(" ")}function gt(e){return e.getAttribute&&e.getAttribute("class")||""}function vt(e){return Array.isArray(e)?e:"string"==typeof e&&e.match(P)||[]}S.fn.extend({prop:function(e,t){return $(this,S.prop,e,t,1<arguments.length)},removeProp:function(e){return this.each(function(){delete this[S.propFix[e]||e]})}}),S.extend({prop:function(e,t,n){var r,i,o=e.nodeType;if(3!==o&&8!==o&&2!==o)return 1===o&&S.isXMLDoc(e)||(t=S.propFix[t]||t,i=S.propHooks[t]),void 0!==n?i&&"set"in i&&void 0!==(r=i.set(e,n,t))?r:e[t]=n:i&&"get"in i&&null!==(r=i.get(e,t))?r:e[t]},propHooks:{tabIndex:{get:function(e){var t=S.find.attr(e,"tabindex");return t?parseInt(t,10):pt.test(e.nodeName)||dt.test(e.nodeName)&&e.href?0:-1}}},propFix:{"for":"htmlFor","class":"className"}}),y.optSelected||(S.propHooks.selected={get:function(e){var t=e.parentNode;return t&&t.parentNode&&t.parentNode.selectedIndex,null},set:function(e){var t=e.parentNode;t&&(t.selectedIndex,t.parentNode&&t.parentNode.selectedIndex)}}),S.each(["tabIndex","readOnly","maxLength","cellSpacing","cellPadding","rowSpan","colSpan","useMap","frameBorder","contentEditable"],function(){S.propFix[this.toLowerCase()]=this}),S.fn.extend({addClass:function(t){var e,n,r,i,o,a,s,u=0;if(m(t))return this.each(function(e){S(this).addClass(t.call(this,e,gt(this)))});if((e=vt(t)).length)while(n=this[u++])if(i=gt(n),r=1===n.nodeType&&" "+ht(i)+" "){a=0;while(o=e[a++])r.indexOf(" "+o+" ")<0&&(r+=o+" ");i!==(s=ht(r))&&n.setAttribute("class",s)}return this},removeClass:function(t){var e,n,r,i,o,a,s,u=0;if(m(t))return this.each(function(e){S(this).removeClass(t.call(this,e,gt(this)))});if(!arguments.length)return this.attr("class","");if((e=vt(t)).length)while(n=this[u++])if(i=gt(n),r=1===n.nodeType&&" "+ht(i)+" "){a=0;while(o=e[a++])while(-1<r.indexOf(" "+o+" "))r=r.replace(" "+o+" "," ");i!==(s=ht(r))&&n.setAttribute("class",s)}return this},toggleClass:function(i,t){var o=typeof i,a="string"===o||Array.isArray(i);return"boolean"==typeof t&&a?t?this.addClass(i):this.removeClass(i):m(i)?this.each(function(e){S(this).toggleClass(i.call(this,e,gt(this),t),t)}):this.each(function(){var e,t,n,r;if(a){t=0,n=S(this),r=vt(i);while(e=r[t++])n.hasClass(e)?n.removeClass(e):n.addClass(e)}else void 0!==i&&"boolean"!==o||((e=gt(this))&&Y.set(this,"__className__",e),this.setAttribute&&this.setAttribute("class",e||!1===i?"":Y.get(this,"__className__")||""))})},hasClass:function(e){var t,n,r=0;t=" "+e+" ";while(n=this[r++])if(1===n.nodeType&&-1<(" "+ht(gt(n))+" ").indexOf(t))return!0;return!1}});var yt=/\r/g;S.fn.extend({val:function(n){var r,e,i,t=this[0];return arguments.length?(i=m(n),this.each(function(e){var t;1===this.nodeType&&(null==(t=i?n.call(this,e,S(this).val()):n)?t="":"number"==typeof t?t+="":Array.isArray(t)&&(t=S.map(t,function(e){return null==e?"":e+""})),(r=S.valHooks[this.type]||S.valHooks[this.nodeName.toLowerCase()])&&"set"in r&&void 0!==r.set(this,t,"value")||(this.value=t))})):t?(r=S.valHooks[t.type]||S.valHooks[t.nodeName.toLowerCase()])&&"get"in r&&void 0!==(e=r.get(t,"value"))?e:"string"==typeof(e=t.value)?e.replace(yt,""):null==e?"":e:void 0}}),S.extend({valHooks:{option:{get:function(e){var t=S.find.attr(e,"value");return null!=t?t:ht(S.text(e))}},select:{get:function(e){var t,n,r,i=e.options,o=e.selectedIndex,a="select-one"===e.type,s=a?null:[],u=a?o+1:i.length;for(r=o<0?u:a?o:0;r<u;r++)if(((n=i[r]).selected||r===o)&&!n.disabled&&(!n.parentNode.disabled||!A(n.parentNode,"optgroup"))){if(t=S(n).val(),a)return t;s.push(t)}return s},set:function(e,t){var n,r,i=e.options,o=S.makeArray(t),a=i.length;while(a--)((r=i[a]).selected=-1<S.inArray(S.valHooks.option.get(r),o))&&(n=!0);return n||(e.selectedIndex=-1),o}}}}),S.each(["radio","checkbox"],function(){S.valHooks[this]={set:function(e,t){if(Array.isArray(t))return e.checked=-1<S.inArray(S(e).val(),t)}},y.checkOn||(S.valHooks[this].get=function(e){return null===e.getAttribute("value")?"on":e.value})}),y.focusin="onfocusin"in C;var mt=/^(?:focusinfocus|focusoutblur)$/,xt=function(e){e.stopPropagation()};S.extend(S.event,{trigger:function(e,t,n,r){var i,o,a,s,u,l,c,f,p=[n||E],d=v.call(e,"type")?e.type:e,h=v.call(e,"namespace")?e.namespace.split("."):[];if(o=f=a=n=n||E,3!==n.nodeType&&8!==n.nodeType&&!mt.test(d+S.event.triggered)&&(-1<d.indexOf(".")&&(d=(h=d.split(".")).shift(),h.sort()),u=d.indexOf(":")<0&&"on"+d,(e=e[S.expando]?e:new S.Event(d,"object"==typeof e&&e)).isTrigger=r?2:3,e.namespace=h.join("."),e.rnamespace=e.namespace?new RegExp("(^|\\.)"+h.join("\\.(?:.*\\.|)")+"(\\.|$)"):null,e.result=void 0,e.target||(e.target=n),t=null==t?[e]:S.makeArray(t,[e]),c=S.event.special[d]||{},r||!c.trigger||!1!==c.trigger.apply(n,t))){if(!r&&!c.noBubble&&!x(n)){for(s=c.delegateType||d,mt.test(s+d)||(o=o.parentNode);o;o=o.parentNode)p.push(o),a=o;a===(n.ownerDocument||E)&&p.push(a.defaultView||a.parentWindow||C)}i=0;while((o=p[i++])&&!e.isPropagationStopped())f=o,e.type=1<i?s:c.bindType||d,(l=(Y.get(o,"events")||Object.create(null))[e.type]&&Y.get(o,"handle"))&&l.apply(o,t),(l=u&&o[u])&&l.apply&&V(o)&&(e.result=l.apply(o,t),!1===e.result&&e.preventDefault());return e.type=d,r||e.isDefaultPrevented()||c._default&&!1!==c._default.apply(p.pop(),t)||!V(n)||u&&m(n[d])&&!x(n)&&((a=n[u])&&(n[u]=null),S.event.triggered=d,e.isPropagationStopped()&&f.addEventListener(d,xt),n[d](),e.isPropagationStopped()&&f.removeEventListener(d,xt),S.event.triggered=void 0,a&&(n[u]=a)),e.result}},simulate:function(e,t,n){var r=S.extend(new S.Event,n,{type:e,isSimulated:!0});S.event.trigger(r,null,t)}}),S.fn.extend({trigger:function(e,t){return this.each(function(){S.event.trigger(e,t,this)})},triggerHandler:function(e,t){var n=this[0];if(n)return S.event.trigger(e,t,n,!0)}}),y.focusin||S.each({focus:"focusin",blur:"focusout"},function(n,r){var i=function(e){S.event.simulate(r,e.target,S.event.fix(e))};S.event.special[r]={setup:function(){var e=this.ownerDocument||this.document||this,t=Y.access(e,r);t||e.addEventListener(n,i,!0),Y.access(e,r,(t||0)+1)},teardown:function(){var e=this.ownerDocument||this.document||this,t=Y.access(e,r)-1;t?Y.access(e,r,t):(e.removeEventListener(n,i,!0),Y.remove(e,r))}}});var bt=C.location,wt={guid:Date.now()},Tt=/\?/;S.parseXML=function(e){var t,n;if(!e||"string"!=typeof e)return null;try{t=(new C.DOMParser).parseFromString(e,"text/xml")}catch(e){}return n=t&&t.getElementsByTagName("parsererror")[0],t&&!n||S.error("Invalid XML: "+(n?S.map(n.childNodes,function(e){return e.textContent}).join("\n"):e)),t};var Ct=/\[\]$/,Et=/\r?\n/g,St=/^(?:submit|button|image|reset|file)$/i,kt=/^(?:input|select|textarea|keygen)/i;function At(n,e,r,i){var t;if(Array.isArray(e))S.each(e,function(e,t){r||Ct.test(n)?i(n,t):At(n+"["+("object"==typeof t&&null!=t?e:"")+"]",t,r,i)});else if(r||"object"!==w(e))i(n,e);else for(t in e)At(n+"["+t+"]",e[t],r,i)}S.param=function(e,t){var n,r=[],i=function(e,t){var n=m(t)?t():t;r[r.length]=encodeURIComponent(e)+"="+encodeURIComponent(null==n?"":n)};if(null==e)return"";if(Array.isArray(e)||e.jquery&&!S.isPlainObject(e))S.each(e,function(){i(this.name,this.value)});else for(n in e)At(n,e[n],t,i);return r.join("&")},S.fn.extend({serialize:function(){return S.param(this.serializeArray())},serializeArray:function(){return this.map(function(){var e=S.prop(this,"elements");return e?S.makeArray(e):this}).filter(function(){var e=this.type;return this.name&&!S(this).is(":disabled")&&kt.test(this.nodeName)&&!St.test(e)&&(this.checked||!pe.test(e))}).map(function(e,t){var n=S(this).val();return null==n?null:Array.isArray(n)?S.map(n,function(e){return{name:t.name,value:e.replace(Et,"\r\n")}}):{name:t.name,value:n.replace(Et,"\r\n")}}).get()}});var Nt=/%20/g,jt=/#.*$/,Dt=/([?&])_=[^&]*/,qt=/^(.*?):[ \t]*([^\r\n]*)$/gm,Lt=/^(?:GET|HEAD)$/,Ht=/^\/\//,Ot={},Pt={},Rt="*/".concat("*"),Mt=E.createElement("a");function It(o){return function(e,t){"string"!=typeof e&&(t=e,e="*");var n,r=0,i=e.toLowerCase().match(P)||[];if(m(t))while(n=i[r++])"+"===n[0]?(n=n.slice(1)||"*",(o[n]=o[n]||[]).unshift(t)):(o[n]=o[n]||[]).push(t)}}function Wt(t,i,o,a){var s={},u=t===Pt;function l(e){var r;return s[e]=!0,S.each(t[e]||[],function(e,t){var n=t(i,o,a);return"string"!=typeof n||u||s[n]?u?!(r=n):void 0:(i.dataTypes.unshift(n),l(n),!1)}),r}return l(i.dataTypes[0])||!s["*"]&&l("*")}function Ft(e,t){var n,r,i=S.ajaxSettings.flatOptions||{};for(n in t)void 0!==t[n]&&((i[n]?e:r||(r={}))[n]=t[n]);return r&&S.extend(!0,e,r),e}Mt.href=bt.href,S.extend({active:0,lastModified:{},etag:{},ajaxSettings:{url:bt.href,type:"GET",isLocal:/^(?:about|app|app-storage|.+-extension|file|res|widget):$/.test(bt.protocol),global:!0,processData:!0,async:!0,contentType:"application/x-www-form-urlencoded; charset=UTF-8",accepts:{"*":Rt,text:"text/plain",html:"text/html",xml:"application/xml, text/xml",json:"application/json, text/javascript"},contents:{xml:/\bxml\b/,html:/\bhtml/,json:/\bjson\b/},responseFields:{xml:"responseXML",text:"responseText",json:"responseJSON"},converters:{"* text":String,"text html":!0,"text json":JSON.parse,"text xml":S.parseXML},flatOptions:{url:!0,context:!0}},ajaxSetup:function(e,t){return t?Ft(Ft(e,S.ajaxSettings),t):Ft(S.ajaxSettings,e)},ajaxPrefilter:It(Ot),ajaxTransport:It(Pt),ajax:function(e,t){"object"==typeof e&&(t=e,e=void 0),t=t||{};var c,f,p,n,d,r,h,g,i,o,v=S.ajaxSetup({},t),y=v.context||v,m=v.context&&(y.nodeType||y.jquery)?S(y):S.event,x=S.Deferred(),b=S.Callbacks("once memory"),w=v.statusCode||{},a={},s={},u="canceled",T={readyState:0,getResponseHeader:function(e){var t;if(h){if(!n){n={};while(t=qt.exec(p))n[t[1].toLowerCase()+" "]=(n[t[1].toLowerCase()+" "]||[]).concat(t[2])}t=n[e.toLowerCase()+" "]}return null==t?null:t.join(", ")},getAllResponseHeaders:function(){return h?p:null},setRequestHeader:function(e,t){return null==h&&(e=s[e.toLowerCase()]=s[e.toLowerCase()]||e,a[e]=t),this},overrideMimeType:function(e){return null==h&&(v.mimeType=e),this},statusCode:function(e){var t;if(e)if(h)T.always(e[T.status]);else for(t in e)w[t]=[w[t],e[t]];return this},abort:function(e){var t=e||u;return c&&c.abort(t),l(0,t),this}};if(x.promise(T),v.url=((e||v.url||bt.href)+"").replace(Ht,bt.protocol+"//"),v.type=t.method||t.type||v.method||v.type,v.dataTypes=(v.dataType||"*").toLowerCase().match(P)||[""],null==v.crossDomain){r=E.createElement("a");try{r.href=v.url,r.href=r.href,v.crossDomain=Mt.protocol+"//"+Mt.host!=r.protocol+"//"+r.host}catch(e){v.crossDomain=!0}}if(v.data&&v.processData&&"string"!=typeof v.data&&(v.data=S.param(v.data,v.traditional)),Wt(Ot,v,t,T),h)return T;for(i in(g=S.event&&v.global)&&0==S.active++&&S.event.trigger("ajaxStart"),v.type=v.type.toUpperCase(),v.hasContent=!Lt.test(v.type),f=v.url.replace(jt,""),v.hasContent?v.data&&v.processData&&0===(v.contentType||"").indexOf("application/x-www-form-urlencoded")&&(v.data=v.data.replace(Nt,"+")):(o=v.url.slice(f.length),v.data&&(v.processData||"string"==typeof v.data)&&(f+=(Tt.test(f)?"&":"?")+v.data,delete v.data),!1===v.cache&&(f=f.replace(Dt,"$1"),o=(Tt.test(f)?"&":"?")+"_="+wt.guid+++o),v.url=f+o),v.ifModified&&(S.lastModified[f]&&T.setRequestHeader("If-Modified-Since",S.lastModified[f]),S.etag[f]&&T.setRequestHeader("If-None-Match",S.etag[f])),(v.data&&v.hasContent&&!1!==v.contentType||t.contentType)&&T.setRequestHeader("Content-Type",v.contentType),T.setRequestHeader("Accept",v.dataTypes[0]&&v.accepts[v.dataTypes[0]]?v.accepts[v.dataTypes[0]]+("*"!==v.dataTypes[0]?", "+Rt+"; q=0.01":""):v.accepts["*"]),v.headers)T.setRequestHeader(i,v.headers[i]);if(v.beforeSend&&(!1===v.beforeSend.call(y,T,v)||h))return T.abort();if(u="abort",b.add(v.complete),T.done(v.success),T.fail(v.error),c=Wt(Pt,v,t,T)){if(T.readyState=1,g&&m.trigger("ajaxSend",[T,v]),h)return T;v.async&&0<v.timeout&&(d=C.setTimeout(function(){T.abort("timeout")},v.timeout));try{h=!1,c.send(a,l)}catch(e){if(h)throw e;l(-1,e)}}else l(-1,"No Transport");function l(e,t,n,r){var i,o,a,s,u,l=t;h||(h=!0,d&&C.clearTimeout(d),c=void 0,p=r||"",T.readyState=0<e?4:0,i=200<=e&&e<300||304===e,n&&(s=function(e,t,n){var r,i,o,a,s=e.contents,u=e.dataTypes;while("*"===u[0])u.shift(),void 0===r&&(r=e.mimeType||t.getResponseHeader("Content-Type"));if(r)for(i in s)if(s[i]&&s[i].test(r)){u.unshift(i);break}if(u[0]in n)o=u[0];else{for(i in n){if(!u[0]||e.converters[i+" "+u[0]]){o=i;break}a||(a=i)}o=o||a}if(o)return o!==u[0]&&u.unshift(o),n[o]}(v,T,n)),!i&&-1<S.inArray("script",v.dataTypes)&&S.inArray("json",v.dataTypes)<0&&(v.converters["text script"]=function(){}),s=function(e,t,n,r){var i,o,a,s,u,l={},c=e.dataTypes.slice();if(c[1])for(a in e.converters)l[a.toLowerCase()]=e.converters[a];o=c.shift();while(o)if(e.responseFields[o]&&(n[e.responseFields[o]]=t),!u&&r&&e.dataFilter&&(t=e.dataFilter(t,e.dataType)),u=o,o=c.shift())if("*"===o)o=u;else if("*"!==u&&u!==o){if(!(a=l[u+" "+o]||l["* "+o]))for(i in l)if((s=i.split(" "))[1]===o&&(a=l[u+" "+s[0]]||l["* "+s[0]])){!0===a?a=l[i]:!0!==l[i]&&(o=s[0],c.unshift(s[1]));break}if(!0!==a)if(a&&e["throws"])t=a(t);else try{t=a(t)}catch(e){return{state:"parsererror",error:a?e:"No conversion from "+u+" to "+o}}}return{state:"success",data:t}}(v,s,T,i),i?(v.ifModified&&((u=T.getResponseHeader("Last-Modified"))&&(S.lastModified[f]=u),(u=T.getResponseHeader("etag"))&&(S.etag[f]=u)),204===e||"HEAD"===v.type?l="nocontent":304===e?l="notmodified":(l=s.state,o=s.data,i=!(a=s.error))):(a=l,!e&&l||(l="error",e<0&&(e=0))),T.status=e,T.statusText=(t||l)+"",i?x.resolveWith(y,[o,l,T]):x.rejectWith(y,[T,l,a]),T.statusCode(w),w=void 0,g&&m.trigger(i?"ajaxSuccess":"ajaxError",[T,v,i?o:a]),b.fireWith(y,[T,l]),g&&(m.trigger("ajaxComplete",[T,v]),--S.active||S.event.trigger("ajaxStop")))}return T},getJSON:function(e,t,n){return S.get(e,t,n,"json")},getScript:function(e,t){return S.get(e,void 0,t,"script")}}),S.each(["get","post"],function(e,i){S[i]=function(e,t,n,r){return m(t)&&(r=r||n,n=t,t=void 0),S.ajax(S.extend({url:e,type:i,dataType:r,data:t,success:n},S.isPlainObject(e)&&e))}}),S.ajaxPrefilter(function(e){var t;for(t in e.headers)"content-type"===t.toLowerCase()&&(e.contentType=e.headers[t]||"")}),S._evalUrl=function(e,t,n){return S.ajax({url:e,type:"GET",dataType:"script",cache:!0,async:!1,global:!1,converters:{"text script":function(){}},dataFilter:function(e){S.globalEval(e,t,n)}})},S.fn.extend({wrapAll:function(e){var t;return this[0]&&(m(e)&&(e=e.call(this[0])),t=S(e,this[0].ownerDocument).eq(0).clone(!0),this[0].parentNode&&t.insertBefore(this[0]),t.map(function(){var e=this;while(e.firstElementChild)e=e.firstElementChild;return e}).append(this)),this},wrapInner:function(n){return m(n)?this.each(function(e){S(this).wrapInner(n.call(this,e))}):this.each(function(){var e=S(this),t=e.contents();t.length?t.wrapAll(n):e.append(n)})},wrap:function(t){var n=m(t);return this.each(function(e){S(this).wrapAll(n?t.call(this,e):t)})},unwrap:function(e){return this.parent(e).not("body").each(function(){S(this).replaceWith(this.childNodes)}),this}}),S.expr.pseudos.hidden=function(e){return!S.expr.pseudos.visible(e)},S.expr.pseudos.visible=function(e){return!!(e.offsetWidth||e.offsetHeight||e.getClientRects().length)},S.ajaxSettings.xhr=function(){try{return new C.XMLHttpRequest}catch(e){}};var Bt={0:200,1223:204},$t=S.ajaxSettings.xhr();y.cors=!!$t&&"withCredentials"in $t,y.ajax=$t=!!$t,S.ajaxTransport(function(i){var o,a;if(y.cors||$t&&!i.crossDomain)return{send:function(e,t){var n,r=i.xhr();if(r.open(i.type,i.url,i.async,i.username,i.password),i.xhrFields)for(n in i.xhrFields)r[n]=i.xhrFields[n];for(n in i.mimeType&&r.overrideMimeType&&r.overrideMimeType(i.mimeType),i.crossDomain||e["X-Requested-With"]||(e["X-Requested-With"]="XMLHttpRequest"),e)r.setRequestHeader(n,e[n]);o=function(e){return function(){o&&(o=a=r.onload=r.onerror=r.onabort=r.ontimeout=r.onreadystatechange=null,"abort"===e?r.abort():"error"===e?"number"!=typeof r.status?t(0,"error"):t(r.status,r.statusText):t(Bt[r.status]||r.status,r.statusText,"text"!==(r.responseType||"text")||"string"!=typeof r.responseText?{binary:r.response}:{text:r.responseText},r.getAllResponseHeaders()))}},r.onload=o(),a=r.onerror=r.ontimeout=o("error"),void 0!==r.onabort?r.onabort=a:r.onreadystatechange=function(){4===r.readyState&&C.setTimeout(function(){o&&a()})},o=o("abort");try{r.send(i.hasContent&&i.data||null)}catch(e){if(o)throw e}},abort:function(){o&&o()}}}),S.ajaxPrefilter(function(e){e.crossDomain&&(e.contents.script=!1)}),S.ajaxSetup({accepts:{script:"text/javascript, application/javascript, application/ecmascript, application/x-ecmascript"},contents:{script:/\b(?:java|ecma)script\b/},converters:{"text script":function(e){return S.globalEval(e),e}}}),S.ajaxPrefilter("script",function(e){void 0===e.cache&&(e.cache=!1),e.crossDomain&&(e.type="GET")}),S.ajaxTransport("script",function(n){var r,i;if(n.crossDomain||n.scriptAttrs)return{send:function(e,t){r=S("<script>").attr(n.scriptAttrs||{}).prop({charset:n.scriptCharset,src:n.url}).on("load error",i=function(e){r.remove(),i=null,e&&t("error"===e.type?404:200,e.type)}),E.head.appendChild(r[0])},abort:function(){i&&i()}}});var _t,zt=[],Ut=/(=)\?(?=&|$)|\?\?/;S.ajaxSetup({jsonp:"callback",jsonpCallback:function(){var e=zt.pop()||S.expando+"_"+wt.guid++;return this[e]=!0,e}}),S.ajaxPrefilter("json jsonp",function(e,t,n){var r,i,o,a=!1!==e.jsonp&&(Ut.test(e.url)?"url":"string"==typeof e.data&&0===(e.contentType||"").indexOf("application/x-www-form-urlencoded")&&Ut.test(e.data)&&"data");if(a||"jsonp"===e.dataTypes[0])return r=e.jsonpCallback=m(e.jsonpCallback)?e.jsonpCallback():e.jsonpCallback,a?e[a]=e[a].replace(Ut,"$1"+r):!1!==e.jsonp&&(e.url+=(Tt.test(e.url)?"&":"?")+e.jsonp+"="+r),e.converters["script json"]=function(){return o||S.error(r+" was not called"),o[0]},e.dataTypes[0]="json",i=C[r],C[r]=function(){o=arguments},n.always(function(){void 0===i?S(C).removeProp(r):C[r]=i,e[r]&&(e.jsonpCallback=t.jsonpCallback,zt.push(r)),o&&m(i)&&i(o[0]),o=i=void 0}),"script"}),y.createHTMLDocument=((_t=E.implementation.createHTMLDocument("").body).innerHTML="<form></form><form></form>",2===_t.childNodes.length),S.parseHTML=function(e,t,n){return"string"!=typeof e?[]:("boolean"==typeof t&&(n=t,t=!1),t||(y.createHTMLDocument?((r=(t=E.implementation.createHTMLDocument("")).createElement("base")).href=E.location.href,t.head.appendChild(r)):t=E),o=!n&&[],(i=N.exec(e))?[t.createElement(i[1])]:(i=xe([e],t,o),o&&o.length&&S(o).remove(),S.merge([],i.childNodes)));var r,i,o},S.fn.load=function(e,t,n){var r,i,o,a=this,s=e.indexOf(" ");return-1<s&&(r=ht(e.slice(s)),e=e.slice(0,s)),m(t)?(n=t,t=void 0):t&&"object"==typeof t&&(i="POST"),0<a.length&&S.ajax({url:e,type:i||"GET",dataType:"html",data:t}).done(function(e){o=arguments,a.html(r?S("<div>").append(S.parseHTML(e)).find(r):e)}).always(n&&function(e,t){a.each(function(){n.apply(this,o||[e.responseText,t,e])})}),this},S.expr.pseudos.animated=function(t){return S.grep(S.timers,function(e){return t===e.elem}).length},S.offset={setOffset:function(e,t,n){var r,i,o,a,s,u,l=S.css(e,"position"),c=S(e),f={};"static"===l&&(e.style.position="relative"),s=c.offset(),o=S.css(e,"top"),u=S.css(e,"left"),("absolute"===l||"fixed"===l)&&-1<(o+u).indexOf("auto")?(a=(r=c.position()).top,i=r.left):(a=parseFloat(o)||0,i=parseFloat(u)||0),m(t)&&(t=t.call(e,n,S.extend({},s))),null!=t.top&&(f.top=t.top-s.top+a),null!=t.left&&(f.left=t.left-s.left+i),"using"in t?t.using.call(e,f):c.css(f)}},S.fn.extend({offset:function(t){if(arguments.length)return void 0===t?this:this.each(function(e){S.offset.setOffset(this,t,e)});var e,n,r=this[0];return r?r.getClientRects().length?(e=r.getBoundingClientRect(),n=r.ownerDocument.defaultView,{top:e.top+n.pageYOffset,left:e.left+n.pageXOffset}):{top:0,left:0}:void 0},position:function(){if(this[0]){var e,t,n,r=this[0],i={top:0,left:0};if("fixed"===S.css(r,"position"))t=r.getBoundingClientRect();else{t=this.offset(),n=r.ownerDocument,e=r.offsetParent||n.documentElement;while(e&&(e===n.body||e===n.documentElement)&&"static"===S.css(e,"position"))e=e.parentNode;e&&e!==r&&1===e.nodeType&&((i=S(e).offset()).top+=S.css(e,"borderTopWidth",!0),i.left+=S.css(e,"borderLeftWidth",!0))}return{top:t.top-i.top-S.css(r,"marginTop",!0),left:t.left-i.left-S.css(r,"marginLeft",!0)}}},offsetParent:function(){return this.map(function(){var e=this.offsetParent;while(e&&"static"===S.css(e,"position"))e=e.offsetParent;return e||re})}}),S.each({scrollLeft:"pageXOffset",scrollTop:"pageYOffset"},function(t,i){var o="pageYOffset"===i;S.fn[t]=function(e){return $(this,function(e,t,n){var r;if(x(e)?r=e:9===e.nodeType&&(r=e.defaultView),void 0===n)return r?r[i]:e[t];r?r.scrollTo(o?r.pageXOffset:n,o?n:r.pageYOffset):e[t]=n},t,e,arguments.length)}}),S.each(["top","left"],function(e,n){S.cssHooks[n]=Fe(y.pixelPosition,function(e,t){if(t)return t=We(e,n),Pe.test(t)?S(e).position()[n]+"px":t})}),S.each({Height:"height",Width:"width"},function(a,s){S.each({padding:"inner"+a,content:s,"":"outer"+a},function(r,o){S.fn[o]=function(e,t){var n=arguments.length&&(r||"boolean"!=typeof e),i=r||(!0===e||!0===t?"margin":"border");return $(this,function(e,t,n){var r;return x(e)?0===o.indexOf("outer")?e["inner"+a]:e.document.documentElement["client"+a]:9===e.nodeType?(r=e.documentElement,Math.max(e.body["scroll"+a],r["scroll"+a],e.body["offset"+a],r["offset"+a],r["client"+a])):void 0===n?S.css(e,t,i):S.style(e,t,n,i)},s,n?e:void 0,n)}})}),S.each(["ajaxStart","ajaxStop","ajaxComplete","ajaxError","ajaxSuccess","ajaxSend"],function(e,t){S.fn[t]=function(e){return this.on(t,e)}}),S.fn.extend({bind:function(e,t,n){return this.on(e,null,t,n)},unbind:function(e,t){return this.off(e,null,t)},delegate:function(e,t,n,r){return this.on(t,e,n,r)},undelegate:function(e,t,n){return 1===arguments.length?this.off(e,"**"):this.off(t,e||"**",n)},hover:function(e,t){return this.mouseenter(e).mouseleave(t||e)}}),S.each("blur focus focusin focusout resize scroll click dblclick mousedown mouseup mousemove mouseover mouseout mouseenter mouseleave change select submit keydown keypress keyup contextmenu".split(" "),function(e,n){S.fn[n]=function(e,t){return 0<arguments.length?this.on(n,null,e,t):this.trigger(n)}});var Xt=/^[\s\uFEFF\xA0]+|[\s\uFEFF\xA0]+$/g;S.proxy=function(e,t){var n,r,i;if("string"==typeof t&&(n=e[t],t=e,e=n),m(e))return r=s.call(arguments,2),(i=function(){return e.apply(t||this,r.concat(s.call(arguments)))}).guid=e.guid=e.guid||S.guid++,i},S.holdReady=function(e){e?S.readyWait++:S.ready(!0)},S.isArray=Array.isArray,S.parseJSON=JSON.parse,S.nodeName=A,S.isFunction=m,S.isWindow=x,S.camelCase=X,S.type=w,S.now=Date.now,S.isNumeric=function(e){var t=S.type(e);return("number"===t||"string"===t)&&!isNaN(e-parseFloat(e))},S.trim=function(e){return null==e?"":(e+"").replace(Xt,"")},"function"==typeof define&&define.amd&&define("jquery",[],function(){return S});var Vt=C.jQuery,Gt=C.$;return S.noConflict=function(e){return C.$===S&&(C.$=Gt),e&&C.jQuery===S&&(C.jQuery=Vt),S},"undefined"==typeof e&&(C.jQuery=C.$=S),S});
diff --git a/_static/language_data.js b/_static/language_data.js
index ebe2f03bf..2e22b06ab 100644
--- a/_static/language_data.js
+++ b/_static/language_data.js
@@ -10,7 +10,7 @@
  *
  */
 
-var stopwords = ["a","and","are","as","at","be","but","by","for","if","in","into","is","it","near","no","not","of","on","or","such","that","the","their","then","there","these","they","this","to","was","will","with"];
+var stopwords = ["a", "and", "are", "as", "at", "be", "but", "by", "for", "if", "in", "into", "is", "it", "near", "no", "not", "of", "on", "or", "such", "that", "the", "their", "then", "there", "these", "they", "this", "to", "was", "will", "with"];
 
 
 /* Non-minified version is copied as a separate JS file, is available */
@@ -197,101 +197,3 @@ var Stemmer = function() {
   }
 }
 
-
-
-
-var splitChars = (function() {
-    var result = {};
-    var singles = [96, 180, 187, 191, 215, 247, 749, 885, 903, 907, 909, 930, 1014, 1648,
-         1748, 1809, 2416, 2473, 2481, 2526, 2601, 2609, 2612, 2615, 2653, 2702,
-         2706, 2729, 2737, 2740, 2857, 2865, 2868, 2910, 2928, 2948, 2961, 2971,
-         2973, 3085, 3089, 3113, 3124, 3213, 3217, 3241, 3252, 3295, 3341, 3345,
-         3369, 3506, 3516, 3633, 3715, 3721, 3736, 3744, 3748, 3750, 3756, 3761,
-         3781, 3912, 4239, 4347, 4681, 4695, 4697, 4745, 4785, 4799, 4801, 4823,
-         4881, 5760, 5901, 5997, 6313, 7405, 8024, 8026, 8028, 8030, 8117, 8125,
-         8133, 8181, 8468, 8485, 8487, 8489, 8494, 8527, 11311, 11359, 11687, 11695,
-         11703, 11711, 11719, 11727, 11735, 12448, 12539, 43010, 43014, 43019, 43587,
-         43696, 43713, 64286, 64297, 64311, 64317, 64319, 64322, 64325, 65141];
-    var i, j, start, end;
-    for (i = 0; i < singles.length; i++) {
-        result[singles[i]] = true;
-    }
-    var ranges = [[0, 47], [58, 64], [91, 94], [123, 169], [171, 177], [182, 184], [706, 709],
-         [722, 735], [741, 747], [751, 879], [888, 889], [894, 901], [1154, 1161],
-         [1318, 1328], [1367, 1368], [1370, 1376], [1416, 1487], [1515, 1519], [1523, 1568],
-         [1611, 1631], [1642, 1645], [1750, 1764], [1767, 1773], [1789, 1790], [1792, 1807],
-         [1840, 1868], [1958, 1968], [1970, 1983], [2027, 2035], [2038, 2041], [2043, 2047],
-         [2070, 2073], [2075, 2083], [2085, 2087], [2089, 2307], [2362, 2364], [2366, 2383],
-         [2385, 2391], [2402, 2405], [2419, 2424], [2432, 2436], [2445, 2446], [2449, 2450],
-         [2483, 2485], [2490, 2492], [2494, 2509], [2511, 2523], [2530, 2533], [2546, 2547],
-         [2554, 2564], [2571, 2574], [2577, 2578], [2618, 2648], [2655, 2661], [2672, 2673],
-         [2677, 2692], [2746, 2748], [2750, 2767], [2769, 2783], [2786, 2789], [2800, 2820],
-         [2829, 2830], [2833, 2834], [2874, 2876], [2878, 2907], [2914, 2917], [2930, 2946],
-         [2955, 2957], [2966, 2968], [2976, 2978], [2981, 2983], [2987, 2989], [3002, 3023],
-         [3025, 3045], [3059, 3076], [3130, 3132], [3134, 3159], [3162, 3167], [3170, 3173],
-         [3184, 3191], [3199, 3204], [3258, 3260], [3262, 3293], [3298, 3301], [3312, 3332],
-         [3386, 3388], [3390, 3423], [3426, 3429], [3446, 3449], [3456, 3460], [3479, 3481],
-         [3518, 3519], [3527, 3584], [3636, 3647], [3655, 3663], [3674, 3712], [3717, 3718],
-         [3723, 3724], [3726, 3731], [3752, 3753], [3764, 3772], [3774, 3775], [3783, 3791],
-         [3802, 3803], [3806, 3839], [3841, 3871], [3892, 3903], [3949, 3975], [3980, 4095],
-         [4139, 4158], [4170, 4175], [4182, 4185], [4190, 4192], [4194, 4196], [4199, 4205],
-         [4209, 4212], [4226, 4237], [4250, 4255], [4294, 4303], [4349, 4351], [4686, 4687],
-         [4702, 4703], [4750, 4751], [4790, 4791], [4806, 4807], [4886, 4887], [4955, 4968],
-         [4989, 4991], [5008, 5023], [5109, 5120], [5741, 5742], [5787, 5791], [5867, 5869],
-         [5873, 5887], [5906, 5919], [5938, 5951], [5970, 5983], [6001, 6015], [6068, 6102],
-         [6104, 6107], [6109, 6111], [6122, 6127], [6138, 6159], [6170, 6175], [6264, 6271],
-         [6315, 6319], [6390, 6399], [6429, 6469], [6510, 6511], [6517, 6527], [6572, 6592],
-         [6600, 6607], [6619, 6655], [6679, 6687], [6741, 6783], [6794, 6799], [6810, 6822],
-         [6824, 6916], [6964, 6980], [6988, 6991], [7002, 7042], [7073, 7085], [7098, 7167],
-         [7204, 7231], [7242, 7244], [7294, 7400], [7410, 7423], [7616, 7679], [7958, 7959],
-         [7966, 7967], [8006, 8007], [8014, 8015], [8062, 8063], [8127, 8129], [8141, 8143],
-         [8148, 8149], [8156, 8159], [8173, 8177], [8189, 8303], [8306, 8307], [8314, 8318],
-         [8330, 8335], [8341, 8449], [8451, 8454], [8456, 8457], [8470, 8472], [8478, 8483],
-         [8506, 8507], [8512, 8516], [8522, 8525], [8586, 9311], [9372, 9449], [9472, 10101],
-         [10132, 11263], [11493, 11498], [11503, 11516], [11518, 11519], [11558, 11567],
-         [11622, 11630], [11632, 11647], [11671, 11679], [11743, 11822], [11824, 12292],
-         [12296, 12320], [12330, 12336], [12342, 12343], [12349, 12352], [12439, 12444],
-         [12544, 12548], [12590, 12592], [12687, 12689], [12694, 12703], [12728, 12783],
-         [12800, 12831], [12842, 12880], [12896, 12927], [12938, 12976], [12992, 13311],
-         [19894, 19967], [40908, 40959], [42125, 42191], [42238, 42239], [42509, 42511],
-         [42540, 42559], [42592, 42593], [42607, 42622], [42648, 42655], [42736, 42774],
-         [42784, 42785], [42889, 42890], [42893, 43002], [43043, 43055], [43062, 43071],
-         [43124, 43137], [43188, 43215], [43226, 43249], [43256, 43258], [43260, 43263],
-         [43302, 43311], [43335, 43359], [43389, 43395], [43443, 43470], [43482, 43519],
-         [43561, 43583], [43596, 43599], [43610, 43615], [43639, 43641], [43643, 43647],
-         [43698, 43700], [43703, 43704], [43710, 43711], [43715, 43738], [43742, 43967],
-         [44003, 44015], [44026, 44031], [55204, 55215], [55239, 55242], [55292, 55295],
-         [57344, 63743], [64046, 64047], [64110, 64111], [64218, 64255], [64263, 64274],
-         [64280, 64284], [64434, 64466], [64830, 64847], [64912, 64913], [64968, 65007],
-         [65020, 65135], [65277, 65295], [65306, 65312], [65339, 65344], [65371, 65381],
-         [65471, 65473], [65480, 65481], [65488, 65489], [65496, 65497]];
-    for (i = 0; i < ranges.length; i++) {
-        start = ranges[i][0];
-        end = ranges[i][1];
-        for (j = start; j <= end; j++) {
-            result[j] = true;
-        }
-    }
-    return result;
-})();
-
-function splitQuery(query) {
-    var result = [];
-    var start = -1;
-    for (var i = 0; i < query.length; i++) {
-        if (splitChars[query.charCodeAt(i)]) {
-            if (start !== -1) {
-                result.push(query.slice(start, i));
-                start = -1;
-            }
-        } else if (start === -1) {
-            start = i;
-        }
-    }
-    if (start !== -1) {
-        result.push(query.slice(start));
-    }
-    return result;
-}
-
-
diff --git a/_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css b/_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css
new file mode 100644
index 000000000..335663106
--- /dev/null
+++ b/_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css
@@ -0,0 +1,2342 @@
+/* Variables */
+:root {
+  --mystnb-source-bg-color: #f7f7f7;
+  --mystnb-stdout-bg-color: #fcfcfc;
+  --mystnb-stderr-bg-color: #fdd;
+  --mystnb-traceback-bg-color: #fcfcfc;
+  --mystnb-source-border-color: #ccc;
+  --mystnb-source-margin-color: green;
+  --mystnb-stdout-border-color: #f7f7f7;
+  --mystnb-stderr-border-color: #f7f7f7;
+  --mystnb-traceback-border-color: #ffd6d6;
+  --mystnb-hide-prompt-opacity: 70%;
+  --mystnb-source-border-radius: .4em;
+  --mystnb-source-border-width: 1px;
+}
+
+/* Whole cell */
+div.container.cell {
+  padding-left: 0;
+  margin-bottom: 1em;
+}
+
+/* Removing all background formatting so we can control at the div level */
+.cell_input div.highlight,
+.cell_output pre,
+.cell_input pre,
+.cell_output .output {
+  border: none;
+  box-shadow: none;
+}
+
+.cell_output .output pre,
+.cell_input pre {
+  margin: 0px;
+}
+
+/* Input cells */
+div.cell div.cell_input,
+div.cell details.above-input>summary {
+  padding-left: 0em;
+  padding-right: 0em;
+  border: var(--mystnb-source-border-width) var(--mystnb-source-border-color) solid;
+  background-color: var(--mystnb-source-bg-color);
+  border-left-color: var(--mystnb-source-margin-color);
+  border-left-width: medium;
+  border-radius: var(--mystnb-source-border-radius);
+}
+
+div.cell_input>div,
+div.cell_output div.output>div.highlight {
+  margin: 0em !important;
+  border: none !important;
+}
+
+/* All cell outputs */
+.cell_output {
+  padding-left: 1em;
+  padding-right: 0em;
+  margin-top: 1em;
+}
+
+/* Text outputs from cells */
+.cell_output .output.text_plain,
+.cell_output .output.traceback,
+.cell_output .output.stream,
+.cell_output .output.stderr {
+  margin-top: 1em;
+  margin-bottom: 0em;
+  box-shadow: none;
+}
+
+.cell_output .output.text_plain,
+.cell_output .output.stream {
+  background: var(--mystnb-stdout-bg-color);
+  border: 1px solid var(--mystnb-stdout-border-color);
+}
+
+.cell_output .output.stderr {
+  background: var(--mystnb-stderr-bg-color);
+  border: 1px solid var(--mystnb-stderr-border-color);
+}
+
+.cell_output .output.traceback {
+  background: var(--mystnb-traceback-bg-color);
+  border: 1px solid var(--mystnb-traceback-border-color);
+}
+
+/* Collapsible cell content */
+div.cell details.above-input div.cell_input {
+  border-top-left-radius: 0;
+  border-top-right-radius: 0;
+  border-top: var(--mystnb-source-border-width) var(--mystnb-source-border-color) dashed;
+}
+
+div.cell div.cell_input.above-output-prompt {
+  border-bottom-left-radius: 0;
+  border-bottom-right-radius: 0;
+}
+
+div.cell details.above-input>summary {
+  border-bottom-left-radius: 0;
+  border-bottom-right-radius: 0;
+  border-bottom: var(--mystnb-source-border-width) var(--mystnb-source-border-color) dashed;
+  padding-left: 1em;
+  margin-bottom: 0;
+}
+
+div.cell details.above-output>summary {
+  background-color: var(--mystnb-source-bg-color);
+  padding-left: 1em;
+  padding-right: 0em;
+  border: var(--mystnb-source-border-width) var(--mystnb-source-border-color) solid;
+  border-radius: var(--mystnb-source-border-radius);
+  border-left-color: var(--mystnb-source-margin-color);
+  border-left-width: medium;
+}
+
+div.cell details.below-input>summary {
+  background-color: var(--mystnb-source-bg-color);
+  padding-left: 1em;
+  padding-right: 0em;
+  border: var(--mystnb-source-border-width) var(--mystnb-source-border-color) solid;
+  border-top: none;
+  border-bottom-left-radius: var(--mystnb-source-border-radius);
+  border-bottom-right-radius: var(--mystnb-source-border-radius);
+  border-left-color: var(--mystnb-source-margin-color);
+  border-left-width: medium;
+}
+
+div.cell details.hide>summary>span {
+  opacity: var(--mystnb-hide-prompt-opacity);
+}
+
+div.cell details.hide[open]>summary>span.collapsed {
+  display: none;
+}
+
+div.cell details.hide:not([open])>summary>span.expanded {
+  display: none;
+}
+
+@keyframes collapsed-fade-in {
+  0% {
+    opacity: 0;
+  }
+
+  100% {
+    opacity: 1;
+  }
+}
+div.cell details.hide[open]>summary~* {
+  -moz-animation: collapsed-fade-in 0.3s ease-in-out;
+  -webkit-animation: collapsed-fade-in 0.3s ease-in-out;
+  animation: collapsed-fade-in 0.3s ease-in-out;
+}
+
+/* Math align to the left */
+.cell_output .MathJax_Display {
+  text-align: left !important;
+}
+
+/* Pandas tables. Pulled from the Jupyter / nbsphinx CSS */
+div.cell_output table {
+  border: none;
+  border-collapse: collapse;
+  border-spacing: 0;
+  color: black;
+  font-size: 1em;
+  table-layout: fixed;
+}
+
+div.cell_output thead {
+  border-bottom: 1px solid black;
+  vertical-align: bottom;
+}
+
+div.cell_output tr,
+div.cell_output th,
+div.cell_output td {
+  text-align: right;
+  vertical-align: middle;
+  padding: 0.5em 0.5em;
+  line-height: normal;
+  white-space: normal;
+  max-width: none;
+  border: none;
+}
+
+div.cell_output th {
+  font-weight: bold;
+}
+
+div.cell_output tbody tr:nth-child(odd) {
+  background: #f5f5f5;
+}
+
+div.cell_output tbody tr:hover {
+  background: rgba(66, 165, 245, 0.2);
+}
+
+/** source code line numbers **/
+span.linenos {
+  opacity: 0.5;
+}
+
+/* Inline text from `paste` operation */
+
+span.pasted-text {
+  font-weight: bold;
+}
+
+span.pasted-inline img {
+  max-height: 2em;
+}
+
+tbody span.pasted-inline img {
+  max-height: none;
+}
+
+/* Font colors for translated ANSI escape sequences
+Color values are copied from Jupyter Notebook
+https://github.com/jupyter/notebook/blob/52581f8eda9b319eb0390ac77fe5903c38f81e3e/notebook/static/notebook/less/ansicolors.less#L14-L21
+Background colors from
+https://nbsphinx.readthedocs.io/en/latest/code-cells.html#ANSI-Colors
+*/
+div.highlight .-Color-Bold {
+  font-weight: bold;
+}
+
+div.highlight .-Color[class*=-Black] {
+  color: #3E424D
+}
+
+div.highlight .-Color[class*=-Red] {
+  color: #E75C58
+}
+
+div.highlight .-Color[class*=-Green] {
+  color: #00A250
+}
+
+div.highlight .-Color[class*=-Yellow] {
+  color: #DDB62B
+}
+
+div.highlight .-Color[class*=-Blue] {
+  color: #208FFB
+}
+
+div.highlight .-Color[class*=-Magenta] {
+  color: #D160C4
+}
+
+div.highlight .-Color[class*=-Cyan] {
+  color: #60C6C8
+}
+
+div.highlight .-Color[class*=-White] {
+  color: #C5C1B4
+}
+
+div.highlight .-Color[class*=-BGBlack] {
+  background-color: #3E424D
+}
+
+div.highlight .-Color[class*=-BGRed] {
+  background-color: #E75C58
+}
+
+div.highlight .-Color[class*=-BGGreen] {
+  background-color: #00A250
+}
+
+div.highlight .-Color[class*=-BGYellow] {
+  background-color: #DDB62B
+}
+
+div.highlight .-Color[class*=-BGBlue] {
+  background-color: #208FFB
+}
+
+div.highlight .-Color[class*=-BGMagenta] {
+  background-color: #D160C4
+}
+
+div.highlight .-Color[class*=-BGCyan] {
+  background-color: #60C6C8
+}
+
+div.highlight .-Color[class*=-BGWhite] {
+  background-color: #C5C1B4
+}
+
+/* Font colors for 8-bit ANSI */
+
+div.highlight .-Color[class*=-C0] {
+  color: #000000
+}
+
+div.highlight .-Color[class*=-BGC0] {
+  background-color: #000000
+}
+
+div.highlight .-Color[class*=-C1] {
+  color: #800000
+}
+
+div.highlight .-Color[class*=-BGC1] {
+  background-color: #800000
+}
+
+div.highlight .-Color[class*=-C2] {
+  color: #008000
+}
+
+div.highlight .-Color[class*=-BGC2] {
+  background-color: #008000
+}
+
+div.highlight .-Color[class*=-C3] {
+  color: #808000
+}
+
+div.highlight .-Color[class*=-BGC3] {
+  background-color: #808000
+}
+
+div.highlight .-Color[class*=-C4] {
+  color: #000080
+}
+
+div.highlight .-Color[class*=-BGC4] {
+  background-color: #000080
+}
+
+div.highlight .-Color[class*=-C5] {
+  color: #800080
+}
+
+div.highlight .-Color[class*=-BGC5] {
+  background-color: #800080
+}
+
+div.highlight .-Color[class*=-C6] {
+  color: #008080
+}
+
+div.highlight .-Color[class*=-BGC6] {
+  background-color: #008080
+}
+
+div.highlight .-Color[class*=-C7] {
+  color: #C0C0C0
+}
+
+div.highlight .-Color[class*=-BGC7] {
+  background-color: #C0C0C0
+}
+
+div.highlight .-Color[class*=-C8] {
+  color: #808080
+}
+
+div.highlight .-Color[class*=-BGC8] {
+  background-color: #808080
+}
+
+div.highlight .-Color[class*=-C9] {
+  color: #FF0000
+}
+
+div.highlight .-Color[class*=-BGC9] {
+  background-color: #FF0000
+}
+
+div.highlight .-Color[class*=-C10] {
+  color: #00FF00
+}
+
+div.highlight .-Color[class*=-BGC10] {
+  background-color: #00FF00
+}
+
+div.highlight .-Color[class*=-C11] {
+  color: #FFFF00
+}
+
+div.highlight .-Color[class*=-BGC11] {
+  background-color: #FFFF00
+}
+
+div.highlight .-Color[class*=-C12] {
+  color: #0000FF
+}
+
+div.highlight .-Color[class*=-BGC12] {
+  background-color: #0000FF
+}
+
+div.highlight .-Color[class*=-C13] {
+  color: #FF00FF
+}
+
+div.highlight .-Color[class*=-BGC13] {
+  background-color: #FF00FF
+}
+
+div.highlight .-Color[class*=-C14] {
+  color: #00FFFF
+}
+
+div.highlight .-Color[class*=-BGC14] {
+  background-color: #00FFFF
+}
+
+div.highlight .-Color[class*=-C15] {
+  color: #FFFFFF
+}
+
+div.highlight .-Color[class*=-BGC15] {
+  background-color: #FFFFFF
+}
+
+div.highlight .-Color[class*=-C16] {
+  color: #000000
+}
+
+div.highlight .-Color[class*=-BGC16] {
+  background-color: #000000
+}
+
+div.highlight .-Color[class*=-C17] {
+  color: #00005F
+}
+
+div.highlight .-Color[class*=-BGC17] {
+  background-color: #00005F
+}
+
+div.highlight .-Color[class*=-C18] {
+  color: #000087
+}
+
+div.highlight .-Color[class*=-BGC18] {
+  background-color: #000087
+}
+
+div.highlight .-Color[class*=-C19] {
+  color: #0000AF
+}
+
+div.highlight .-Color[class*=-BGC19] {
+  background-color: #0000AF
+}
+
+div.highlight .-Color[class*=-C20] {
+  color: #0000D7
+}
+
+div.highlight .-Color[class*=-BGC20] {
+  background-color: #0000D7
+}
+
+div.highlight .-Color[class*=-C21] {
+  color: #0000FF
+}
+
+div.highlight .-Color[class*=-BGC21] {
+  background-color: #0000FF
+}
+
+div.highlight .-Color[class*=-C22] {
+  color: #005F00
+}
+
+div.highlight .-Color[class*=-BGC22] {
+  background-color: #005F00
+}
+
+div.highlight .-Color[class*=-C23] {
+  color: #005F5F
+}
+
+div.highlight .-Color[class*=-BGC23] {
+  background-color: #005F5F
+}
+
+div.highlight .-Color[class*=-C24] {
+  color: #005F87
+}
+
+div.highlight .-Color[class*=-BGC24] {
+  background-color: #005F87
+}
+
+div.highlight .-Color[class*=-C25] {
+  color: #005FAF
+}
+
+div.highlight .-Color[class*=-BGC25] {
+  background-color: #005FAF
+}
+
+div.highlight .-Color[class*=-C26] {
+  color: #005FD7
+}
+
+div.highlight .-Color[class*=-BGC26] {
+  background-color: #005FD7
+}
+
+div.highlight .-Color[class*=-C27] {
+  color: #005FFF
+}
+
+div.highlight .-Color[class*=-BGC27] {
+  background-color: #005FFF
+}
+
+div.highlight .-Color[class*=-C28] {
+  color: #008700
+}
+
+div.highlight .-Color[class*=-BGC28] {
+  background-color: #008700
+}
+
+div.highlight .-Color[class*=-C29] {
+  color: #00875F
+}
+
+div.highlight .-Color[class*=-BGC29] {
+  background-color: #00875F
+}
+
+div.highlight .-Color[class*=-C30] {
+  color: #008787
+}
+
+div.highlight .-Color[class*=-BGC30] {
+  background-color: #008787
+}
+
+div.highlight .-Color[class*=-C31] {
+  color: #0087AF
+}
+
+div.highlight .-Color[class*=-BGC31] {
+  background-color: #0087AF
+}
+
+div.highlight .-Color[class*=-C32] {
+  color: #0087D7
+}
+
+div.highlight .-Color[class*=-BGC32] {
+  background-color: #0087D7
+}
+
+div.highlight .-Color[class*=-C33] {
+  color: #0087FF
+}
+
+div.highlight .-Color[class*=-BGC33] {
+  background-color: #0087FF
+}
+
+div.highlight .-Color[class*=-C34] {
+  color: #00AF00
+}
+
+div.highlight .-Color[class*=-BGC34] {
+  background-color: #00AF00
+}
+
+div.highlight .-Color[class*=-C35] {
+  color: #00AF5F
+}
+
+div.highlight .-Color[class*=-BGC35] {
+  background-color: #00AF5F
+}
+
+div.highlight .-Color[class*=-C36] {
+  color: #00AF87
+}
+
+div.highlight .-Color[class*=-BGC36] {
+  background-color: #00AF87
+}
+
+div.highlight .-Color[class*=-C37] {
+  color: #00AFAF
+}
+
+div.highlight .-Color[class*=-BGC37] {
+  background-color: #00AFAF
+}
+
+div.highlight .-Color[class*=-C38] {
+  color: #00AFD7
+}
+
+div.highlight .-Color[class*=-BGC38] {
+  background-color: #00AFD7
+}
+
+div.highlight .-Color[class*=-C39] {
+  color: #00AFFF
+}
+
+div.highlight .-Color[class*=-BGC39] {
+  background-color: #00AFFF
+}
+
+div.highlight .-Color[class*=-C40] {
+  color: #00D700
+}
+
+div.highlight .-Color[class*=-BGC40] {
+  background-color: #00D700
+}
+
+div.highlight .-Color[class*=-C41] {
+  color: #00D75F
+}
+
+div.highlight .-Color[class*=-BGC41] {
+  background-color: #00D75F
+}
+
+div.highlight .-Color[class*=-C42] {
+  color: #00D787
+}
+
+div.highlight .-Color[class*=-BGC42] {
+  background-color: #00D787
+}
+
+div.highlight .-Color[class*=-C43] {
+  color: #00D7AF
+}
+
+div.highlight .-Color[class*=-BGC43] {
+  background-color: #00D7AF
+}
+
+div.highlight .-Color[class*=-C44] {
+  color: #00D7D7
+}
+
+div.highlight .-Color[class*=-BGC44] {
+  background-color: #00D7D7
+}
+
+div.highlight .-Color[class*=-C45] {
+  color: #00D7FF
+}
+
+div.highlight .-Color[class*=-BGC45] {
+  background-color: #00D7FF
+}
+
+div.highlight .-Color[class*=-C46] {
+  color: #00FF00
+}
+
+div.highlight .-Color[class*=-BGC46] {
+  background-color: #00FF00
+}
+
+div.highlight .-Color[class*=-C47] {
+  color: #00FF5F
+}
+
+div.highlight .-Color[class*=-BGC47] {
+  background-color: #00FF5F
+}
+
+div.highlight .-Color[class*=-C48] {
+  color: #00FF87
+}
+
+div.highlight .-Color[class*=-BGC48] {
+  background-color: #00FF87
+}
+
+div.highlight .-Color[class*=-C49] {
+  color: #00FFAF
+}
+
+div.highlight .-Color[class*=-BGC49] {
+  background-color: #00FFAF
+}
+
+div.highlight .-Color[class*=-C50] {
+  color: #00FFD7
+}
+
+div.highlight .-Color[class*=-BGC50] {
+  background-color: #00FFD7
+}
+
+div.highlight .-Color[class*=-C51] {
+  color: #00FFFF
+}
+
+div.highlight .-Color[class*=-BGC51] {
+  background-color: #00FFFF
+}
+
+div.highlight .-Color[class*=-C52] {
+  color: #5F0000
+}
+
+div.highlight .-Color[class*=-BGC52] {
+  background-color: #5F0000
+}
+
+div.highlight .-Color[class*=-C53] {
+  color: #5F005F
+}
+
+div.highlight .-Color[class*=-BGC53] {
+  background-color: #5F005F
+}
+
+div.highlight .-Color[class*=-C54] {
+  color: #5F0087
+}
+
+div.highlight .-Color[class*=-BGC54] {
+  background-color: #5F0087
+}
+
+div.highlight .-Color[class*=-C55] {
+  color: #5F00AF
+}
+
+div.highlight .-Color[class*=-BGC55] {
+  background-color: #5F00AF
+}
+
+div.highlight .-Color[class*=-C56] {
+  color: #5F00D7
+}
+
+div.highlight .-Color[class*=-BGC56] {
+  background-color: #5F00D7
+}
+
+div.highlight .-Color[class*=-C57] {
+  color: #5F00FF
+}
+
+div.highlight .-Color[class*=-BGC57] {
+  background-color: #5F00FF
+}
+
+div.highlight .-Color[class*=-C58] {
+  color: #5F5F00
+}
+
+div.highlight .-Color[class*=-BGC58] {
+  background-color: #5F5F00
+}
+
+div.highlight .-Color[class*=-C59] {
+  color: #5F5F5F
+}
+
+div.highlight .-Color[class*=-BGC59] {
+  background-color: #5F5F5F
+}
+
+div.highlight .-Color[class*=-C60] {
+  color: #5F5F87
+}
+
+div.highlight .-Color[class*=-BGC60] {
+  background-color: #5F5F87
+}
+
+div.highlight .-Color[class*=-C61] {
+  color: #5F5FAF
+}
+
+div.highlight .-Color[class*=-BGC61] {
+  background-color: #5F5FAF
+}
+
+div.highlight .-Color[class*=-C62] {
+  color: #5F5FD7
+}
+
+div.highlight .-Color[class*=-BGC62] {
+  background-color: #5F5FD7
+}
+
+div.highlight .-Color[class*=-C63] {
+  color: #5F5FFF
+}
+
+div.highlight .-Color[class*=-BGC63] {
+  background-color: #5F5FFF
+}
+
+div.highlight .-Color[class*=-C64] {
+  color: #5F8700
+}
+
+div.highlight .-Color[class*=-BGC64] {
+  background-color: #5F8700
+}
+
+div.highlight .-Color[class*=-C65] {
+  color: #5F875F
+}
+
+div.highlight .-Color[class*=-BGC65] {
+  background-color: #5F875F
+}
+
+div.highlight .-Color[class*=-C66] {
+  color: #5F8787
+}
+
+div.highlight .-Color[class*=-BGC66] {
+  background-color: #5F8787
+}
+
+div.highlight .-Color[class*=-C67] {
+  color: #5F87AF
+}
+
+div.highlight .-Color[class*=-BGC67] {
+  background-color: #5F87AF
+}
+
+div.highlight .-Color[class*=-C68] {
+  color: #5F87D7
+}
+
+div.highlight .-Color[class*=-BGC68] {
+  background-color: #5F87D7
+}
+
+div.highlight .-Color[class*=-C69] {
+  color: #5F87FF
+}
+
+div.highlight .-Color[class*=-BGC69] {
+  background-color: #5F87FF
+}
+
+div.highlight .-Color[class*=-C70] {
+  color: #5FAF00
+}
+
+div.highlight .-Color[class*=-BGC70] {
+  background-color: #5FAF00
+}
+
+div.highlight .-Color[class*=-C71] {
+  color: #5FAF5F
+}
+
+div.highlight .-Color[class*=-BGC71] {
+  background-color: #5FAF5F
+}
+
+div.highlight .-Color[class*=-C72] {
+  color: #5FAF87
+}
+
+div.highlight .-Color[class*=-BGC72] {
+  background-color: #5FAF87
+}
+
+div.highlight .-Color[class*=-C73] {
+  color: #5FAFAF
+}
+
+div.highlight .-Color[class*=-BGC73] {
+  background-color: #5FAFAF
+}
+
+div.highlight .-Color[class*=-C74] {
+  color: #5FAFD7
+}
+
+div.highlight .-Color[class*=-BGC74] {
+  background-color: #5FAFD7
+}
+
+div.highlight .-Color[class*=-C75] {
+  color: #5FAFFF
+}
+
+div.highlight .-Color[class*=-BGC75] {
+  background-color: #5FAFFF
+}
+
+div.highlight .-Color[class*=-C76] {
+  color: #5FD700
+}
+
+div.highlight .-Color[class*=-BGC76] {
+  background-color: #5FD700
+}
+
+div.highlight .-Color[class*=-C77] {
+  color: #5FD75F
+}
+
+div.highlight .-Color[class*=-BGC77] {
+  background-color: #5FD75F
+}
+
+div.highlight .-Color[class*=-C78] {
+  color: #5FD787
+}
+
+div.highlight .-Color[class*=-BGC78] {
+  background-color: #5FD787
+}
+
+div.highlight .-Color[class*=-C79] {
+  color: #5FD7AF
+}
+
+div.highlight .-Color[class*=-BGC79] {
+  background-color: #5FD7AF
+}
+
+div.highlight .-Color[class*=-C80] {
+  color: #5FD7D7
+}
+
+div.highlight .-Color[class*=-BGC80] {
+  background-color: #5FD7D7
+}
+
+div.highlight .-Color[class*=-C81] {
+  color: #5FD7FF
+}
+
+div.highlight .-Color[class*=-BGC81] {
+  background-color: #5FD7FF
+}
+
+div.highlight .-Color[class*=-C82] {
+  color: #5FFF00
+}
+
+div.highlight .-Color[class*=-BGC82] {
+  background-color: #5FFF00
+}
+
+div.highlight .-Color[class*=-C83] {
+  color: #5FFF5F
+}
+
+div.highlight .-Color[class*=-BGC83] {
+  background-color: #5FFF5F
+}
+
+div.highlight .-Color[class*=-C84] {
+  color: #5FFF87
+}
+
+div.highlight .-Color[class*=-BGC84] {
+  background-color: #5FFF87
+}
+
+div.highlight .-Color[class*=-C85] {
+  color: #5FFFAF
+}
+
+div.highlight .-Color[class*=-BGC85] {
+  background-color: #5FFFAF
+}
+
+div.highlight .-Color[class*=-C86] {
+  color: #5FFFD7
+}
+
+div.highlight .-Color[class*=-BGC86] {
+  background-color: #5FFFD7
+}
+
+div.highlight .-Color[class*=-C87] {
+  color: #5FFFFF
+}
+
+div.highlight .-Color[class*=-BGC87] {
+  background-color: #5FFFFF
+}
+
+div.highlight .-Color[class*=-C88] {
+  color: #870000
+}
+
+div.highlight .-Color[class*=-BGC88] {
+  background-color: #870000
+}
+
+div.highlight .-Color[class*=-C89] {
+  color: #87005F
+}
+
+div.highlight .-Color[class*=-BGC89] {
+  background-color: #87005F
+}
+
+div.highlight .-Color[class*=-C90] {
+  color: #870087
+}
+
+div.highlight .-Color[class*=-BGC90] {
+  background-color: #870087
+}
+
+div.highlight .-Color[class*=-C91] {
+  color: #8700AF
+}
+
+div.highlight .-Color[class*=-BGC91] {
+  background-color: #8700AF
+}
+
+div.highlight .-Color[class*=-C92] {
+  color: #8700D7
+}
+
+div.highlight .-Color[class*=-BGC92] {
+  background-color: #8700D7
+}
+
+div.highlight .-Color[class*=-C93] {
+  color: #8700FF
+}
+
+div.highlight .-Color[class*=-BGC93] {
+  background-color: #8700FF
+}
+
+div.highlight .-Color[class*=-C94] {
+  color: #875F00
+}
+
+div.highlight .-Color[class*=-BGC94] {
+  background-color: #875F00
+}
+
+div.highlight .-Color[class*=-C95] {
+  color: #875F5F
+}
+
+div.highlight .-Color[class*=-BGC95] {
+  background-color: #875F5F
+}
+
+div.highlight .-Color[class*=-C96] {
+  color: #875F87
+}
+
+div.highlight .-Color[class*=-BGC96] {
+  background-color: #875F87
+}
+
+div.highlight .-Color[class*=-C97] {
+  color: #875FAF
+}
+
+div.highlight .-Color[class*=-BGC97] {
+  background-color: #875FAF
+}
+
+div.highlight .-Color[class*=-C98] {
+  color: #875FD7
+}
+
+div.highlight .-Color[class*=-BGC98] {
+  background-color: #875FD7
+}
+
+div.highlight .-Color[class*=-C99] {
+  color: #875FFF
+}
+
+div.highlight .-Color[class*=-BGC99] {
+  background-color: #875FFF
+}
+
+div.highlight .-Color[class*=-C100] {
+  color: #878700
+}
+
+div.highlight .-Color[class*=-BGC100] {
+  background-color: #878700
+}
+
+div.highlight .-Color[class*=-C101] {
+  color: #87875F
+}
+
+div.highlight .-Color[class*=-BGC101] {
+  background-color: #87875F
+}
+
+div.highlight .-Color[class*=-C102] {
+  color: #878787
+}
+
+div.highlight .-Color[class*=-BGC102] {
+  background-color: #878787
+}
+
+div.highlight .-Color[class*=-C103] {
+  color: #8787AF
+}
+
+div.highlight .-Color[class*=-BGC103] {
+  background-color: #8787AF
+}
+
+div.highlight .-Color[class*=-C104] {
+  color: #8787D7
+}
+
+div.highlight .-Color[class*=-BGC104] {
+  background-color: #8787D7
+}
+
+div.highlight .-Color[class*=-C105] {
+  color: #8787FF
+}
+
+div.highlight .-Color[class*=-BGC105] {
+  background-color: #8787FF
+}
+
+div.highlight .-Color[class*=-C106] {
+  color: #87AF00
+}
+
+div.highlight .-Color[class*=-BGC106] {
+  background-color: #87AF00
+}
+
+div.highlight .-Color[class*=-C107] {
+  color: #87AF5F
+}
+
+div.highlight .-Color[class*=-BGC107] {
+  background-color: #87AF5F
+}
+
+div.highlight .-Color[class*=-C108] {
+  color: #87AF87
+}
+
+div.highlight .-Color[class*=-BGC108] {
+  background-color: #87AF87
+}
+
+div.highlight .-Color[class*=-C109] {
+  color: #87AFAF
+}
+
+div.highlight .-Color[class*=-BGC109] {
+  background-color: #87AFAF
+}
+
+div.highlight .-Color[class*=-C110] {
+  color: #87AFD7
+}
+
+div.highlight .-Color[class*=-BGC110] {
+  background-color: #87AFD7
+}
+
+div.highlight .-Color[class*=-C111] {
+  color: #87AFFF
+}
+
+div.highlight .-Color[class*=-BGC111] {
+  background-color: #87AFFF
+}
+
+div.highlight .-Color[class*=-C112] {
+  color: #87D700
+}
+
+div.highlight .-Color[class*=-BGC112] {
+  background-color: #87D700
+}
+
+div.highlight .-Color[class*=-C113] {
+  color: #87D75F
+}
+
+div.highlight .-Color[class*=-BGC113] {
+  background-color: #87D75F
+}
+
+div.highlight .-Color[class*=-C114] {
+  color: #87D787
+}
+
+div.highlight .-Color[class*=-BGC114] {
+  background-color: #87D787
+}
+
+div.highlight .-Color[class*=-C115] {
+  color: #87D7AF
+}
+
+div.highlight .-Color[class*=-BGC115] {
+  background-color: #87D7AF
+}
+
+div.highlight .-Color[class*=-C116] {
+  color: #87D7D7
+}
+
+div.highlight .-Color[class*=-BGC116] {
+  background-color: #87D7D7
+}
+
+div.highlight .-Color[class*=-C117] {
+  color: #87D7FF
+}
+
+div.highlight .-Color[class*=-BGC117] {
+  background-color: #87D7FF
+}
+
+div.highlight .-Color[class*=-C118] {
+  color: #87FF00
+}
+
+div.highlight .-Color[class*=-BGC118] {
+  background-color: #87FF00
+}
+
+div.highlight .-Color[class*=-C119] {
+  color: #87FF5F
+}
+
+div.highlight .-Color[class*=-BGC119] {
+  background-color: #87FF5F
+}
+
+div.highlight .-Color[class*=-C120] {
+  color: #87FF87
+}
+
+div.highlight .-Color[class*=-BGC120] {
+  background-color: #87FF87
+}
+
+div.highlight .-Color[class*=-C121] {
+  color: #87FFAF
+}
+
+div.highlight .-Color[class*=-BGC121] {
+  background-color: #87FFAF
+}
+
+div.highlight .-Color[class*=-C122] {
+  color: #87FFD7
+}
+
+div.highlight .-Color[class*=-BGC122] {
+  background-color: #87FFD7
+}
+
+div.highlight .-Color[class*=-C123] {
+  color: #87FFFF
+}
+
+div.highlight .-Color[class*=-BGC123] {
+  background-color: #87FFFF
+}
+
+div.highlight .-Color[class*=-C124] {
+  color: #AF0000
+}
+
+div.highlight .-Color[class*=-BGC124] {
+  background-color: #AF0000
+}
+
+div.highlight .-Color[class*=-C125] {
+  color: #AF005F
+}
+
+div.highlight .-Color[class*=-BGC125] {
+  background-color: #AF005F
+}
+
+div.highlight .-Color[class*=-C126] {
+  color: #AF0087
+}
+
+div.highlight .-Color[class*=-BGC126] {
+  background-color: #AF0087
+}
+
+div.highlight .-Color[class*=-C127] {
+  color: #AF00AF
+}
+
+div.highlight .-Color[class*=-BGC127] {
+  background-color: #AF00AF
+}
+
+div.highlight .-Color[class*=-C128] {
+  color: #AF00D7
+}
+
+div.highlight .-Color[class*=-BGC128] {
+  background-color: #AF00D7
+}
+
+div.highlight .-Color[class*=-C129] {
+  color: #AF00FF
+}
+
+div.highlight .-Color[class*=-BGC129] {
+  background-color: #AF00FF
+}
+
+div.highlight .-Color[class*=-C130] {
+  color: #AF5F00
+}
+
+div.highlight .-Color[class*=-BGC130] {
+  background-color: #AF5F00
+}
+
+div.highlight .-Color[class*=-C131] {
+  color: #AF5F5F
+}
+
+div.highlight .-Color[class*=-BGC131] {
+  background-color: #AF5F5F
+}
+
+div.highlight .-Color[class*=-C132] {
+  color: #AF5F87
+}
+
+div.highlight .-Color[class*=-BGC132] {
+  background-color: #AF5F87
+}
+
+div.highlight .-Color[class*=-C133] {
+  color: #AF5FAF
+}
+
+div.highlight .-Color[class*=-BGC133] {
+  background-color: #AF5FAF
+}
+
+div.highlight .-Color[class*=-C134] {
+  color: #AF5FD7
+}
+
+div.highlight .-Color[class*=-BGC134] {
+  background-color: #AF5FD7
+}
+
+div.highlight .-Color[class*=-C135] {
+  color: #AF5FFF
+}
+
+div.highlight .-Color[class*=-BGC135] {
+  background-color: #AF5FFF
+}
+
+div.highlight .-Color[class*=-C136] {
+  color: #AF8700
+}
+
+div.highlight .-Color[class*=-BGC136] {
+  background-color: #AF8700
+}
+
+div.highlight .-Color[class*=-C137] {
+  color: #AF875F
+}
+
+div.highlight .-Color[class*=-BGC137] {
+  background-color: #AF875F
+}
+
+div.highlight .-Color[class*=-C138] {
+  color: #AF8787
+}
+
+div.highlight .-Color[class*=-BGC138] {
+  background-color: #AF8787
+}
+
+div.highlight .-Color[class*=-C139] {
+  color: #AF87AF
+}
+
+div.highlight .-Color[class*=-BGC139] {
+  background-color: #AF87AF
+}
+
+div.highlight .-Color[class*=-C140] {
+  color: #AF87D7
+}
+
+div.highlight .-Color[class*=-BGC140] {
+  background-color: #AF87D7
+}
+
+div.highlight .-Color[class*=-C141] {
+  color: #AF87FF
+}
+
+div.highlight .-Color[class*=-BGC141] {
+  background-color: #AF87FF
+}
+
+div.highlight .-Color[class*=-C142] {
+  color: #AFAF00
+}
+
+div.highlight .-Color[class*=-BGC142] {
+  background-color: #AFAF00
+}
+
+div.highlight .-Color[class*=-C143] {
+  color: #AFAF5F
+}
+
+div.highlight .-Color[class*=-BGC143] {
+  background-color: #AFAF5F
+}
+
+div.highlight .-Color[class*=-C144] {
+  color: #AFAF87
+}
+
+div.highlight .-Color[class*=-BGC144] {
+  background-color: #AFAF87
+}
+
+div.highlight .-Color[class*=-C145] {
+  color: #AFAFAF
+}
+
+div.highlight .-Color[class*=-BGC145] {
+  background-color: #AFAFAF
+}
+
+div.highlight .-Color[class*=-C146] {
+  color: #AFAFD7
+}
+
+div.highlight .-Color[class*=-BGC146] {
+  background-color: #AFAFD7
+}
+
+div.highlight .-Color[class*=-C147] {
+  color: #AFAFFF
+}
+
+div.highlight .-Color[class*=-BGC147] {
+  background-color: #AFAFFF
+}
+
+div.highlight .-Color[class*=-C148] {
+  color: #AFD700
+}
+
+div.highlight .-Color[class*=-BGC148] {
+  background-color: #AFD700
+}
+
+div.highlight .-Color[class*=-C149] {
+  color: #AFD75F
+}
+
+div.highlight .-Color[class*=-BGC149] {
+  background-color: #AFD75F
+}
+
+div.highlight .-Color[class*=-C150] {
+  color: #AFD787
+}
+
+div.highlight .-Color[class*=-BGC150] {
+  background-color: #AFD787
+}
+
+div.highlight .-Color[class*=-C151] {
+  color: #AFD7AF
+}
+
+div.highlight .-Color[class*=-BGC151] {
+  background-color: #AFD7AF
+}
+
+div.highlight .-Color[class*=-C152] {
+  color: #AFD7D7
+}
+
+div.highlight .-Color[class*=-BGC152] {
+  background-color: #AFD7D7
+}
+
+div.highlight .-Color[class*=-C153] {
+  color: #AFD7FF
+}
+
+div.highlight .-Color[class*=-BGC153] {
+  background-color: #AFD7FF
+}
+
+div.highlight .-Color[class*=-C154] {
+  color: #AFFF00
+}
+
+div.highlight .-Color[class*=-BGC154] {
+  background-color: #AFFF00
+}
+
+div.highlight .-Color[class*=-C155] {
+  color: #AFFF5F
+}
+
+div.highlight .-Color[class*=-BGC155] {
+  background-color: #AFFF5F
+}
+
+div.highlight .-Color[class*=-C156] {
+  color: #AFFF87
+}
+
+div.highlight .-Color[class*=-BGC156] {
+  background-color: #AFFF87
+}
+
+div.highlight .-Color[class*=-C157] {
+  color: #AFFFAF
+}
+
+div.highlight .-Color[class*=-BGC157] {
+  background-color: #AFFFAF
+}
+
+div.highlight .-Color[class*=-C158] {
+  color: #AFFFD7
+}
+
+div.highlight .-Color[class*=-BGC158] {
+  background-color: #AFFFD7
+}
+
+div.highlight .-Color[class*=-C159] {
+  color: #AFFFFF
+}
+
+div.highlight .-Color[class*=-BGC159] {
+  background-color: #AFFFFF
+}
+
+div.highlight .-Color[class*=-C160] {
+  color: #D70000
+}
+
+div.highlight .-Color[class*=-BGC160] {
+  background-color: #D70000
+}
+
+div.highlight .-Color[class*=-C161] {
+  color: #D7005F
+}
+
+div.highlight .-Color[class*=-BGC161] {
+  background-color: #D7005F
+}
+
+div.highlight .-Color[class*=-C162] {
+  color: #D70087
+}
+
+div.highlight .-Color[class*=-BGC162] {
+  background-color: #D70087
+}
+
+div.highlight .-Color[class*=-C163] {
+  color: #D700AF
+}
+
+div.highlight .-Color[class*=-BGC163] {
+  background-color: #D700AF
+}
+
+div.highlight .-Color[class*=-C164] {
+  color: #D700D7
+}
+
+div.highlight .-Color[class*=-BGC164] {
+  background-color: #D700D7
+}
+
+div.highlight .-Color[class*=-C165] {
+  color: #D700FF
+}
+
+div.highlight .-Color[class*=-BGC165] {
+  background-color: #D700FF
+}
+
+div.highlight .-Color[class*=-C166] {
+  color: #D75F00
+}
+
+div.highlight .-Color[class*=-BGC166] {
+  background-color: #D75F00
+}
+
+div.highlight .-Color[class*=-C167] {
+  color: #D75F5F
+}
+
+div.highlight .-Color[class*=-BGC167] {
+  background-color: #D75F5F
+}
+
+div.highlight .-Color[class*=-C168] {
+  color: #D75F87
+}
+
+div.highlight .-Color[class*=-BGC168] {
+  background-color: #D75F87
+}
+
+div.highlight .-Color[class*=-C169] {
+  color: #D75FAF
+}
+
+div.highlight .-Color[class*=-BGC169] {
+  background-color: #D75FAF
+}
+
+div.highlight .-Color[class*=-C170] {
+  color: #D75FD7
+}
+
+div.highlight .-Color[class*=-BGC170] {
+  background-color: #D75FD7
+}
+
+div.highlight .-Color[class*=-C171] {
+  color: #D75FFF
+}
+
+div.highlight .-Color[class*=-BGC171] {
+  background-color: #D75FFF
+}
+
+div.highlight .-Color[class*=-C172] {
+  color: #D78700
+}
+
+div.highlight .-Color[class*=-BGC172] {
+  background-color: #D78700
+}
+
+div.highlight .-Color[class*=-C173] {
+  color: #D7875F
+}
+
+div.highlight .-Color[class*=-BGC173] {
+  background-color: #D7875F
+}
+
+div.highlight .-Color[class*=-C174] {
+  color: #D78787
+}
+
+div.highlight .-Color[class*=-BGC174] {
+  background-color: #D78787
+}
+
+div.highlight .-Color[class*=-C175] {
+  color: #D787AF
+}
+
+div.highlight .-Color[class*=-BGC175] {
+  background-color: #D787AF
+}
+
+div.highlight .-Color[class*=-C176] {
+  color: #D787D7
+}
+
+div.highlight .-Color[class*=-BGC176] {
+  background-color: #D787D7
+}
+
+div.highlight .-Color[class*=-C177] {
+  color: #D787FF
+}
+
+div.highlight .-Color[class*=-BGC177] {
+  background-color: #D787FF
+}
+
+div.highlight .-Color[class*=-C178] {
+  color: #D7AF00
+}
+
+div.highlight .-Color[class*=-BGC178] {
+  background-color: #D7AF00
+}
+
+div.highlight .-Color[class*=-C179] {
+  color: #D7AF5F
+}
+
+div.highlight .-Color[class*=-BGC179] {
+  background-color: #D7AF5F
+}
+
+div.highlight .-Color[class*=-C180] {
+  color: #D7AF87
+}
+
+div.highlight .-Color[class*=-BGC180] {
+  background-color: #D7AF87
+}
+
+div.highlight .-Color[class*=-C181] {
+  color: #D7AFAF
+}
+
+div.highlight .-Color[class*=-BGC181] {
+  background-color: #D7AFAF
+}
+
+div.highlight .-Color[class*=-C182] {
+  color: #D7AFD7
+}
+
+div.highlight .-Color[class*=-BGC182] {
+  background-color: #D7AFD7
+}
+
+div.highlight .-Color[class*=-C183] {
+  color: #D7AFFF
+}
+
+div.highlight .-Color[class*=-BGC183] {
+  background-color: #D7AFFF
+}
+
+div.highlight .-Color[class*=-C184] {
+  color: #D7D700
+}
+
+div.highlight .-Color[class*=-BGC184] {
+  background-color: #D7D700
+}
+
+div.highlight .-Color[class*=-C185] {
+  color: #D7D75F
+}
+
+div.highlight .-Color[class*=-BGC185] {
+  background-color: #D7D75F
+}
+
+div.highlight .-Color[class*=-C186] {
+  color: #D7D787
+}
+
+div.highlight .-Color[class*=-BGC186] {
+  background-color: #D7D787
+}
+
+div.highlight .-Color[class*=-C187] {
+  color: #D7D7AF
+}
+
+div.highlight .-Color[class*=-BGC187] {
+  background-color: #D7D7AF
+}
+
+div.highlight .-Color[class*=-C188] {
+  color: #D7D7D7
+}
+
+div.highlight .-Color[class*=-BGC188] {
+  background-color: #D7D7D7
+}
+
+div.highlight .-Color[class*=-C189] {
+  color: #D7D7FF
+}
+
+div.highlight .-Color[class*=-BGC189] {
+  background-color: #D7D7FF
+}
+
+div.highlight .-Color[class*=-C190] {
+  color: #D7FF00
+}
+
+div.highlight .-Color[class*=-BGC190] {
+  background-color: #D7FF00
+}
+
+div.highlight .-Color[class*=-C191] {
+  color: #D7FF5F
+}
+
+div.highlight .-Color[class*=-BGC191] {
+  background-color: #D7FF5F
+}
+
+div.highlight .-Color[class*=-C192] {
+  color: #D7FF87
+}
+
+div.highlight .-Color[class*=-BGC192] {
+  background-color: #D7FF87
+}
+
+div.highlight .-Color[class*=-C193] {
+  color: #D7FFAF
+}
+
+div.highlight .-Color[class*=-BGC193] {
+  background-color: #D7FFAF
+}
+
+div.highlight .-Color[class*=-C194] {
+  color: #D7FFD7
+}
+
+div.highlight .-Color[class*=-BGC194] {
+  background-color: #D7FFD7
+}
+
+div.highlight .-Color[class*=-C195] {
+  color: #D7FFFF
+}
+
+div.highlight .-Color[class*=-BGC195] {
+  background-color: #D7FFFF
+}
+
+div.highlight .-Color[class*=-C196] {
+  color: #FF0000
+}
+
+div.highlight .-Color[class*=-BGC196] {
+  background-color: #FF0000
+}
+
+div.highlight .-Color[class*=-C197] {
+  color: #FF005F
+}
+
+div.highlight .-Color[class*=-BGC197] {
+  background-color: #FF005F
+}
+
+div.highlight .-Color[class*=-C198] {
+  color: #FF0087
+}
+
+div.highlight .-Color[class*=-BGC198] {
+  background-color: #FF0087
+}
+
+div.highlight .-Color[class*=-C199] {
+  color: #FF00AF
+}
+
+div.highlight .-Color[class*=-BGC199] {
+  background-color: #FF00AF
+}
+
+div.highlight .-Color[class*=-C200] {
+  color: #FF00D7
+}
+
+div.highlight .-Color[class*=-BGC200] {
+  background-color: #FF00D7
+}
+
+div.highlight .-Color[class*=-C201] {
+  color: #FF00FF
+}
+
+div.highlight .-Color[class*=-BGC201] {
+  background-color: #FF00FF
+}
+
+div.highlight .-Color[class*=-C202] {
+  color: #FF5F00
+}
+
+div.highlight .-Color[class*=-BGC202] {
+  background-color: #FF5F00
+}
+
+div.highlight .-Color[class*=-C203] {
+  color: #FF5F5F
+}
+
+div.highlight .-Color[class*=-BGC203] {
+  background-color: #FF5F5F
+}
+
+div.highlight .-Color[class*=-C204] {
+  color: #FF5F87
+}
+
+div.highlight .-Color[class*=-BGC204] {
+  background-color: #FF5F87
+}
+
+div.highlight .-Color[class*=-C205] {
+  color: #FF5FAF
+}
+
+div.highlight .-Color[class*=-BGC205] {
+  background-color: #FF5FAF
+}
+
+div.highlight .-Color[class*=-C206] {
+  color: #FF5FD7
+}
+
+div.highlight .-Color[class*=-BGC206] {
+  background-color: #FF5FD7
+}
+
+div.highlight .-Color[class*=-C207] {
+  color: #FF5FFF
+}
+
+div.highlight .-Color[class*=-BGC207] {
+  background-color: #FF5FFF
+}
+
+div.highlight .-Color[class*=-C208] {
+  color: #FF8700
+}
+
+div.highlight .-Color[class*=-BGC208] {
+  background-color: #FF8700
+}
+
+div.highlight .-Color[class*=-C209] {
+  color: #FF875F
+}
+
+div.highlight .-Color[class*=-BGC209] {
+  background-color: #FF875F
+}
+
+div.highlight .-Color[class*=-C210] {
+  color: #FF8787
+}
+
+div.highlight .-Color[class*=-BGC210] {
+  background-color: #FF8787
+}
+
+div.highlight .-Color[class*=-C211] {
+  color: #FF87AF
+}
+
+div.highlight .-Color[class*=-BGC211] {
+  background-color: #FF87AF
+}
+
+div.highlight .-Color[class*=-C212] {
+  color: #FF87D7
+}
+
+div.highlight .-Color[class*=-BGC212] {
+  background-color: #FF87D7
+}
+
+div.highlight .-Color[class*=-C213] {
+  color: #FF87FF
+}
+
+div.highlight .-Color[class*=-BGC213] {
+  background-color: #FF87FF
+}
+
+div.highlight .-Color[class*=-C214] {
+  color: #FFAF00
+}
+
+div.highlight .-Color[class*=-BGC214] {
+  background-color: #FFAF00
+}
+
+div.highlight .-Color[class*=-C215] {
+  color: #FFAF5F
+}
+
+div.highlight .-Color[class*=-BGC215] {
+  background-color: #FFAF5F
+}
+
+div.highlight .-Color[class*=-C216] {
+  color: #FFAF87
+}
+
+div.highlight .-Color[class*=-BGC216] {
+  background-color: #FFAF87
+}
+
+div.highlight .-Color[class*=-C217] {
+  color: #FFAFAF
+}
+
+div.highlight .-Color[class*=-BGC217] {
+  background-color: #FFAFAF
+}
+
+div.highlight .-Color[class*=-C218] {
+  color: #FFAFD7
+}
+
+div.highlight .-Color[class*=-BGC218] {
+  background-color: #FFAFD7
+}
+
+div.highlight .-Color[class*=-C219] {
+  color: #FFAFFF
+}
+
+div.highlight .-Color[class*=-BGC219] {
+  background-color: #FFAFFF
+}
+
+div.highlight .-Color[class*=-C220] {
+  color: #FFD700
+}
+
+div.highlight .-Color[class*=-BGC220] {
+  background-color: #FFD700
+}
+
+div.highlight .-Color[class*=-C221] {
+  color: #FFD75F
+}
+
+div.highlight .-Color[class*=-BGC221] {
+  background-color: #FFD75F
+}
+
+div.highlight .-Color[class*=-C222] {
+  color: #FFD787
+}
+
+div.highlight .-Color[class*=-BGC222] {
+  background-color: #FFD787
+}
+
+div.highlight .-Color[class*=-C223] {
+  color: #FFD7AF
+}
+
+div.highlight .-Color[class*=-BGC223] {
+  background-color: #FFD7AF
+}
+
+div.highlight .-Color[class*=-C224] {
+  color: #FFD7D7
+}
+
+div.highlight .-Color[class*=-BGC224] {
+  background-color: #FFD7D7
+}
+
+div.highlight .-Color[class*=-C225] {
+  color: #FFD7FF
+}
+
+div.highlight .-Color[class*=-BGC225] {
+  background-color: #FFD7FF
+}
+
+div.highlight .-Color[class*=-C226] {
+  color: #FFFF00
+}
+
+div.highlight .-Color[class*=-BGC226] {
+  background-color: #FFFF00
+}
+
+div.highlight .-Color[class*=-C227] {
+  color: #FFFF5F
+}
+
+div.highlight .-Color[class*=-BGC227] {
+  background-color: #FFFF5F
+}
+
+div.highlight .-Color[class*=-C228] {
+  color: #FFFF87
+}
+
+div.highlight .-Color[class*=-BGC228] {
+  background-color: #FFFF87
+}
+
+div.highlight .-Color[class*=-C229] {
+  color: #FFFFAF
+}
+
+div.highlight .-Color[class*=-BGC229] {
+  background-color: #FFFFAF
+}
+
+div.highlight .-Color[class*=-C230] {
+  color: #FFFFD7
+}
+
+div.highlight .-Color[class*=-BGC230] {
+  background-color: #FFFFD7
+}
+
+div.highlight .-Color[class*=-C231] {
+  color: #FFFFFF
+}
+
+div.highlight .-Color[class*=-BGC231] {
+  background-color: #FFFFFF
+}
+
+div.highlight .-Color[class*=-C232] {
+  color: #080808
+}
+
+div.highlight .-Color[class*=-BGC232] {
+  background-color: #080808
+}
+
+div.highlight .-Color[class*=-C233] {
+  color: #121212
+}
+
+div.highlight .-Color[class*=-BGC233] {
+  background-color: #121212
+}
+
+div.highlight .-Color[class*=-C234] {
+  color: #1C1C1C
+}
+
+div.highlight .-Color[class*=-BGC234] {
+  background-color: #1C1C1C
+}
+
+div.highlight .-Color[class*=-C235] {
+  color: #262626
+}
+
+div.highlight .-Color[class*=-BGC235] {
+  background-color: #262626
+}
+
+div.highlight .-Color[class*=-C236] {
+  color: #303030
+}
+
+div.highlight .-Color[class*=-BGC236] {
+  background-color: #303030
+}
+
+div.highlight .-Color[class*=-C237] {
+  color: #3A3A3A
+}
+
+div.highlight .-Color[class*=-BGC237] {
+  background-color: #3A3A3A
+}
+
+div.highlight .-Color[class*=-C238] {
+  color: #444444
+}
+
+div.highlight .-Color[class*=-BGC238] {
+  background-color: #444444
+}
+
+div.highlight .-Color[class*=-C239] {
+  color: #4E4E4E
+}
+
+div.highlight .-Color[class*=-BGC239] {
+  background-color: #4E4E4E
+}
+
+div.highlight .-Color[class*=-C240] {
+  color: #585858
+}
+
+div.highlight .-Color[class*=-BGC240] {
+  background-color: #585858
+}
+
+div.highlight .-Color[class*=-C241] {
+  color: #626262
+}
+
+div.highlight .-Color[class*=-BGC241] {
+  background-color: #626262
+}
+
+div.highlight .-Color[class*=-C242] {
+  color: #6C6C6C
+}
+
+div.highlight .-Color[class*=-BGC242] {
+  background-color: #6C6C6C
+}
+
+div.highlight .-Color[class*=-C243] {
+  color: #767676
+}
+
+div.highlight .-Color[class*=-BGC243] {
+  background-color: #767676
+}
+
+div.highlight .-Color[class*=-C244] {
+  color: #808080
+}
+
+div.highlight .-Color[class*=-BGC244] {
+  background-color: #808080
+}
+
+div.highlight .-Color[class*=-C245] {
+  color: #8A8A8A
+}
+
+div.highlight .-Color[class*=-BGC245] {
+  background-color: #8A8A8A
+}
+
+div.highlight .-Color[class*=-C246] {
+  color: #949494
+}
+
+div.highlight .-Color[class*=-BGC246] {
+  background-color: #949494
+}
+
+div.highlight .-Color[class*=-C247] {
+  color: #9E9E9E
+}
+
+div.highlight .-Color[class*=-BGC247] {
+  background-color: #9E9E9E
+}
+
+div.highlight .-Color[class*=-C248] {
+  color: #A8A8A8
+}
+
+div.highlight .-Color[class*=-BGC248] {
+  background-color: #A8A8A8
+}
+
+div.highlight .-Color[class*=-C249] {
+  color: #B2B2B2
+}
+
+div.highlight .-Color[class*=-BGC249] {
+  background-color: #B2B2B2
+}
+
+div.highlight .-Color[class*=-C250] {
+  color: #BCBCBC
+}
+
+div.highlight .-Color[class*=-BGC250] {
+  background-color: #BCBCBC
+}
+
+div.highlight .-Color[class*=-C251] {
+  color: #C6C6C6
+}
+
+div.highlight .-Color[class*=-BGC251] {
+  background-color: #C6C6C6
+}
+
+div.highlight .-Color[class*=-C252] {
+  color: #D0D0D0
+}
+
+div.highlight .-Color[class*=-BGC252] {
+  background-color: #D0D0D0
+}
+
+div.highlight .-Color[class*=-C253] {
+  color: #DADADA
+}
+
+div.highlight .-Color[class*=-BGC253] {
+  background-color: #DADADA
+}
+
+div.highlight .-Color[class*=-C254] {
+  color: #E4E4E4
+}
+
+div.highlight .-Color[class*=-BGC254] {
+  background-color: #E4E4E4
+}
+
+div.highlight .-Color[class*=-C255] {
+  color: #EEEEEE
+}
+
+div.highlight .-Color[class*=-BGC255] {
+  background-color: #EEEEEE
+}
diff --git a/_static/searchtools.js b/_static/searchtools.js
index 0a44e8582..e89e34d4e 100644
--- a/_static/searchtools.js
+++ b/_static/searchtools.js
@@ -8,18 +8,20 @@
  * :license: BSD, see LICENSE for details.
  *
  */
+"use strict";
 
-if (!Scorer) {
-  /**
-   * Simple result scoring code.
-   */
+/**
+ * Simple result scoring code.
+ */
+if (typeof Scorer === "undefined") {
   var Scorer = {
     // Implement the following function to further tweak the score for each result
-    // The function takes a result array [filename, title, anchor, descr, score]
+    // The function takes a result array [docname, title, anchor, descr, score, filename]
     // and returns the new score.
     /*
-    score: function(result) {
-      return result[4];
+    score: result => {
+      const [docname, title, anchor, descr, score, filename] = result
+      return score
     },
     */
 
@@ -28,9 +30,11 @@ if (!Scorer) {
     // or matches in the last dotted part of the object name
     objPartialMatch: 6,
     // Additive scores depending on the priority of the object
-    objPrio: {0:  15,   // used to be importantResults
-              1:  5,   // used to be objectResults
-              2: -5},  // used to be unimportantResults
+    objPrio: {
+      0: 15, // used to be importantResults
+      1: 5, // used to be objectResults
+      2: -5, // used to be unimportantResults
+    },
     //  Used when the priority is not in the mapping.
     objPrioDefault: 0,
 
@@ -39,452 +43,495 @@ if (!Scorer) {
     partialTitle: 7,
     // query found in terms
     term: 5,
-    partialTerm: 2
+    partialTerm: 2,
   };
 }
 
-if (!splitQuery) {
-  function splitQuery(query) {
-    return query.split(/\s+/);
+const _removeChildren = (element) => {
+  while (element && element.lastChild) element.removeChild(element.lastChild);
+};
+
+/**
+ * See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions#escaping
+ */
+const _escapeRegExp = (string) =>
+  string.replace(/[.*+\-?^${}()|[\]\\]/g, "\\$&"); // $& means the whole matched string
+
+const _displayItem = (item, searchTerms) => {
+  const docBuilder = DOCUMENTATION_OPTIONS.BUILDER;
+  const docUrlRoot = DOCUMENTATION_OPTIONS.URL_ROOT;
+  const docFileSuffix = DOCUMENTATION_OPTIONS.FILE_SUFFIX;
+  const docLinkSuffix = DOCUMENTATION_OPTIONS.LINK_SUFFIX;
+  const showSearchSummary = DOCUMENTATION_OPTIONS.SHOW_SEARCH_SUMMARY;
+
+  const [docName, title, anchor, descr, score, _filename] = item;
+
+  let listItem = document.createElement("li");
+  let requestUrl;
+  let linkUrl;
+  if (docBuilder === "dirhtml") {
+    // dirhtml builder
+    let dirname = docName + "/";
+    if (dirname.match(/\/index\/$/))
+      dirname = dirname.substring(0, dirname.length - 6);
+    else if (dirname === "index/") dirname = "";
+    requestUrl = docUrlRoot + dirname;
+    linkUrl = requestUrl;
+  } else {
+    // normal html builders
+    requestUrl = docUrlRoot + docName + docFileSuffix;
+    linkUrl = docName + docLinkSuffix;
   }
+  let linkEl = listItem.appendChild(document.createElement("a"));
+  linkEl.href = linkUrl + anchor;
+  linkEl.dataset.score = score;
+  linkEl.innerHTML = title;
+  if (descr)
+    listItem.appendChild(document.createElement("span")).innerHTML =
+      " (" + descr + ")";
+  else if (showSearchSummary)
+    fetch(requestUrl)
+      .then((responseData) => responseData.text())
+      .then((data) => {
+        if (data)
+          listItem.appendChild(
+            Search.makeSearchSummary(data, searchTerms)
+          );
+      });
+  Search.output.appendChild(listItem);
+};
+const _finishSearch = (resultCount) => {
+  Search.stopPulse();
+  Search.title.innerText = _("Search Results");
+  if (!resultCount)
+    Search.status.innerText = Documentation.gettext(
+      "Your search did not match any documents. Please make sure that all words are spelled correctly and that you've selected enough categories."
+    );
+  else
+    Search.status.innerText = _(
+      `Search finished, found ${resultCount} page(s) matching the search query.`
+    );
+};
+const _displayNextItem = (
+  results,
+  resultCount,
+  searchTerms
+) => {
+  // results left, load the summary and display it
+  // this is intended to be dynamic (don't sub resultsCount)
+  if (results.length) {
+    _displayItem(results.pop(), searchTerms);
+    setTimeout(
+      () => _displayNextItem(results, resultCount, searchTerms),
+      5
+    );
+  }
+  // search finished, update title and status message
+  else _finishSearch(resultCount);
+};
+
+/**
+ * Default splitQuery function. Can be overridden in ``sphinx.search`` with a
+ * custom function per language.
+ *
+ * The regular expression works by splitting the string on consecutive characters
+ * that are not Unicode letters, numbers, underscores, or emoji characters.
+ * This is the same as ``\W+`` in Python, preserving the surrogate pair area.
+ */
+if (typeof splitQuery === "undefined") {
+  var splitQuery = (query) => query
+      .split(/[^\p{Letter}\p{Number}_\p{Emoji_Presentation}]+/gu)
+      .filter(term => term)  // remove remaining empty strings
 }
 
 /**
  * Search Module
  */
-var Search = {
-
-  _index : null,
-  _queued_query : null,
-  _pulse_status : -1,
-
-  htmlToText : function(htmlString) {
-      var virtualDocument = document.implementation.createHTMLDocument('virtual');
-      var htmlElement = $(htmlString, virtualDocument);
-      htmlElement.find('.headerlink').remove();
-      docContent = htmlElement.find('[role=main]')[0];
-      if(docContent === undefined) {
-          console.warn("Content block not found. Sphinx search tries to obtain it " +
-                       "via '[role=main]'. Could you check your theme or template.");
-          return "";
-      }
-      return docContent.textContent || docContent.innerText;
+const Search = {
+  _index: null,
+  _queued_query: null,
+  _pulse_status: -1,
+
+  htmlToText: (htmlString) => {
+    const htmlElement = new DOMParser().parseFromString(htmlString, 'text/html');
+    htmlElement.querySelectorAll(".headerlink").forEach((el) => { el.remove() });
+    const docContent = htmlElement.querySelector('[role="main"]');
+    if (docContent !== undefined) return docContent.textContent;
+    console.warn(
+      "Content block not found. Sphinx search tries to obtain it via '[role=main]'. Could you check your theme or template."
+    );
+    return "";
   },
 
-  init : function() {
-      var params = $.getQueryParameters();
-      if (params.q) {
-          var query = params.q[0];
-          $('input[name="q"]')[0].value = query;
-          this.performSearch(query);
-      }
+  init: () => {
+    const query = new URLSearchParams(window.location.search).get("q");
+    document
+      .querySelectorAll('input[name="q"]')
+      .forEach((el) => (el.value = query));
+    if (query) Search.performSearch(query);
   },
 
-  loadIndex : function(url) {
-    $.ajax({type: "GET", url: url, data: null,
-            dataType: "script", cache: true,
-            complete: function(jqxhr, textstatus) {
-              if (textstatus != "success") {
-                document.getElementById("searchindexloader").src = url;
-              }
-            }});
-  },
+  loadIndex: (url) =>
+    (document.body.appendChild(document.createElement("script")).src = url),
 
-  setIndex : function(index) {
-    var q;
-    this._index = index;
-    if ((q = this._queued_query) !== null) {
-      this._queued_query = null;
-      Search.query(q);
+  setIndex: (index) => {
+    Search._index = index;
+    if (Search._queued_query !== null) {
+      const query = Search._queued_query;
+      Search._queued_query = null;
+      Search.query(query);
     }
   },
 
-  hasIndex : function() {
-      return this._index !== null;
-  },
+  hasIndex: () => Search._index !== null,
 
-  deferQuery : function(query) {
-      this._queued_query = query;
-  },
+  deferQuery: (query) => (Search._queued_query = query),
 
-  stopPulse : function() {
-      this._pulse_status = 0;
-  },
+  stopPulse: () => (Search._pulse_status = -1),
 
-  startPulse : function() {
-    if (this._pulse_status >= 0)
-        return;
-    function pulse() {
-      var i;
+  startPulse: () => {
+    if (Search._pulse_status >= 0) return;
+
+    const pulse = () => {
       Search._pulse_status = (Search._pulse_status + 1) % 4;
-      var dotString = '';
-      for (i = 0; i < Search._pulse_status; i++)
-        dotString += '.';
-      Search.dots.text(dotString);
-      if (Search._pulse_status > -1)
-        window.setTimeout(pulse, 500);
-    }
+      Search.dots.innerText = ".".repeat(Search._pulse_status);
+      if (Search._pulse_status >= 0) window.setTimeout(pulse, 500);
+    };
     pulse();
   },
 
   /**
    * perform a search for something (or wait until index is loaded)
    */
-  performSearch : function(query) {
+  performSearch: (query) => {
     // create the required interface elements
-    this.out = $('#search-results');
-    this.title = $('<h2>' + _('Searching') + '</h2>').appendTo(this.out);
-    this.dots = $('<span></span>').appendTo(this.title);
-    this.status = $('<p class="search-summary">&nbsp;</p>').appendTo(this.out);
-    this.output = $('<ul class="search"/>').appendTo(this.out);
-
-    $('#search-progress').text(_('Preparing search...'));
-    this.startPulse();
+    const searchText = document.createElement("h2");
+    searchText.textContent = _("Searching");
+    const searchSummary = document.createElement("p");
+    searchSummary.classList.add("search-summary");
+    searchSummary.innerText = "";
+    const searchList = document.createElement("ul");
+    searchList.classList.add("search");
+
+    const out = document.getElementById("search-results");
+    Search.title = out.appendChild(searchText);
+    Search.dots = Search.title.appendChild(document.createElement("span"));
+    Search.status = out.appendChild(searchSummary);
+    Search.output = out.appendChild(searchList);
+
+    const searchProgress = document.getElementById("search-progress");
+    // Some themes don't use the search progress node
+    if (searchProgress) {
+      searchProgress.innerText = _("Preparing search...");
+    }
+    Search.startPulse();
 
     // index already loaded, the browser was quick!
-    if (this.hasIndex())
-      this.query(query);
-    else
-      this.deferQuery(query);
+    if (Search.hasIndex()) Search.query(query);
+    else Search.deferQuery(query);
   },
 
   /**
    * execute search (requires search index to be loaded)
    */
-  query : function(query) {
-    var i;
-
-    // stem the searchterms and add them to the correct list
-    var stemmer = new Stemmer();
-    var searchterms = [];
-    var excluded = [];
-    var hlterms = [];
-    var tmp = splitQuery(query);
-    var objectterms = [];
-    for (i = 0; i < tmp.length; i++) {
-      if (tmp[i] !== "") {
-          objectterms.push(tmp[i].toLowerCase());
-      }
+  query: (query) => {
+    const filenames = Search._index.filenames;
+    const docNames = Search._index.docnames;
+    const titles = Search._index.titles;
+    const allTitles = Search._index.alltitles;
+    const indexEntries = Search._index.indexentries;
+
+    // stem the search terms and add them to the correct list
+    const stemmer = new Stemmer();
+    const searchTerms = new Set();
+    const excludedTerms = new Set();
+    const highlightTerms = new Set();
+    const objectTerms = new Set(splitQuery(query.toLowerCase().trim()));
+    splitQuery(query.trim()).forEach((queryTerm) => {
+      const queryTermLower = queryTerm.toLowerCase();
+
+      // maybe skip this "word"
+      // stopwords array is from language_data.js
+      if (
+        stopwords.indexOf(queryTermLower) !== -1 ||
+        queryTerm.match(/^\d+$/)
+      )
+        return;
 
-      if ($u.indexOf(stopwords, tmp[i].toLowerCase()) != -1 || tmp[i] === "") {
-        // skip this "word"
-        continue;
-      }
       // stem the word
-      var word = stemmer.stemWord(tmp[i].toLowerCase());
-      var toAppend;
+      let word = stemmer.stemWord(queryTermLower);
       // select the correct list
-      if (word[0] == '-') {
-        toAppend = excluded;
-        word = word.substr(1);
-      }
+      if (word[0] === "-") excludedTerms.add(word.substr(1));
       else {
-        toAppend = searchterms;
-        hlterms.push(tmp[i].toLowerCase());
+        searchTerms.add(word);
+        highlightTerms.add(queryTermLower);
       }
-      // only add if not already in the list
-      if (!$u.contains(toAppend, word))
-        toAppend.push(word);
-    }
-    var highlightstring = '?highlight=' + $.urlencode(hlterms.join(" "));
+    });
 
-    // console.debug('SEARCH: searching for:');
-    // console.info('required: ', searchterms);
-    // console.info('excluded: ', excluded);
+    if (SPHINX_HIGHLIGHT_ENABLED) {  // set in sphinx_highlight.js
+      localStorage.setItem("sphinx_highlight_terms", [...highlightTerms].join(" "))
+    }
 
-    // prepare search
-    var terms = this._index.terms;
-    var titleterms = this._index.titleterms;
+    // console.debug("SEARCH: searching for:");
+    // console.info("required: ", [...searchTerms]);
+    // console.info("excluded: ", [...excludedTerms]);
+
+    // array of [docname, title, anchor, descr, score, filename]
+    let results = [];
+    _removeChildren(document.getElementById("search-progress"));
+
+    const queryLower = query.toLowerCase();
+    for (const [title, foundTitles] of Object.entries(allTitles)) {
+      if (title.toLowerCase().includes(queryLower) && (queryLower.length >= title.length/2)) {
+        for (const [file, id] of foundTitles) {
+          let score = Math.round(100 * queryLower.length / title.length)
+          results.push([
+            docNames[file],
+            titles[file] !== title ? `${titles[file]} > ${title}` : title,
+            id !== null ? "#" + id : "",
+            null,
+            score,
+            filenames[file],
+          ]);
+        }
+      }
+    }
 
-    // array of [filename, title, anchor, descr, score]
-    var results = [];
-    $('#search-progress').empty();
+    // search for explicit entries in index directives
+    for (const [entry, foundEntries] of Object.entries(indexEntries)) {
+      if (entry.includes(queryLower) && (queryLower.length >= entry.length/2)) {
+        for (const [file, id] of foundEntries) {
+          let score = Math.round(100 * queryLower.length / entry.length)
+          results.push([
+            docNames[file],
+            titles[file],
+            id ? "#" + id : "",
+            null,
+            score,
+            filenames[file],
+          ]);
+        }
+      }
+    }
 
     // lookup as object
-    for (i = 0; i < objectterms.length; i++) {
-      var others = [].concat(objectterms.slice(0, i),
-                             objectterms.slice(i+1, objectterms.length));
-      results = results.concat(this.performObjectSearch(objectterms[i], others));
-    }
+    objectTerms.forEach((term) =>
+      results.push(...Search.performObjectSearch(term, objectTerms))
+    );
 
     // lookup as search terms in fulltext
-    results = results.concat(this.performTermsSearch(searchterms, excluded, terms, titleterms));
+    results.push(...Search.performTermsSearch(searchTerms, excludedTerms));
 
     // let the scorer override scores with a custom scoring function
-    if (Scorer.score) {
-      for (i = 0; i < results.length; i++)
-        results[i][4] = Scorer.score(results[i]);
-    }
+    if (Scorer.score) results.forEach((item) => (item[4] = Scorer.score(item)));
 
     // now sort the results by score (in opposite order of appearance, since the
     // display function below uses pop() to retrieve items) and then
     // alphabetically
-    results.sort(function(a, b) {
-      var left = a[4];
-      var right = b[4];
-      if (left > right) {
-        return 1;
-      } else if (left < right) {
-        return -1;
-      } else {
+    results.sort((a, b) => {
+      const leftScore = a[4];
+      const rightScore = b[4];
+      if (leftScore === rightScore) {
         // same score: sort alphabetically
-        left = a[1].toLowerCase();
-        right = b[1].toLowerCase();
-        return (left > right) ? -1 : ((left < right) ? 1 : 0);
+        const leftTitle = a[1].toLowerCase();
+        const rightTitle = b[1].toLowerCase();
+        if (leftTitle === rightTitle) return 0;
+        return leftTitle > rightTitle ? -1 : 1; // inverted is intentional
       }
+      return leftScore > rightScore ? 1 : -1;
     });
 
+    // remove duplicate search results
+    // note the reversing of results, so that in the case of duplicates, the highest-scoring entry is kept
+    let seen = new Set();
+    results = results.reverse().reduce((acc, result) => {
+      let resultStr = result.slice(0, 4).concat([result[5]]).map(v => String(v)).join(',');
+      if (!seen.has(resultStr)) {
+        acc.push(result);
+        seen.add(resultStr);
+      }
+      return acc;
+    }, []);
+
+    results = results.reverse();
+
     // for debugging
     //Search.lastresults = results.slice();  // a copy
-    //console.info('search results:', Search.lastresults);
+    // console.info("search results:", Search.lastresults);
 
     // print the results
-    var resultCount = results.length;
-    function displayNextItem() {
-      // results left, load the summary and display it
-      if (results.length) {
-        var item = results.pop();
-        var listItem = $('<li></li>');
-        var requestUrl = "";
-        var linkUrl = "";
-        if (DOCUMENTATION_OPTIONS.BUILDER === 'dirhtml') {
-          // dirhtml builder
-          var dirname = item[0] + '/';
-          if (dirname.match(/\/index\/$/)) {
-            dirname = dirname.substring(0, dirname.length-6);
-          } else if (dirname == 'index/') {
-            dirname = '';
-          }
-          requestUrl = DOCUMENTATION_OPTIONS.URL_ROOT + dirname;
-          linkUrl = requestUrl;
-
-        } else {
-          // normal html builders
-          requestUrl = DOCUMENTATION_OPTIONS.URL_ROOT + item[0] + DOCUMENTATION_OPTIONS.FILE_SUFFIX;
-          linkUrl = item[0] + DOCUMENTATION_OPTIONS.LINK_SUFFIX;
-        }
-        listItem.append($('<a/>').attr('href',
-            linkUrl +
-            highlightstring + item[2]).html(item[1]));
-        if (item[3]) {
-          listItem.append($('<span> (' + item[3] + ')</span>'));
-          Search.output.append(listItem);
-          setTimeout(function() {
-            displayNextItem();
-          }, 5);
-        } else if (DOCUMENTATION_OPTIONS.SHOW_SEARCH_SUMMARY) {
-          $.ajax({url: requestUrl,
-                  dataType: "text",
-                  complete: function(jqxhr, textstatus) {
-                    var data = jqxhr.responseText;
-                    if (data !== '' && data !== undefined) {
-                      var summary = Search.makeSearchSummary(data, searchterms, hlterms);
-                      if (summary) {
-                        listItem.append(summary);
-                      }
-                    }
-                    Search.output.append(listItem);
-                    setTimeout(function() {
-                      displayNextItem();
-                    }, 5);
-                  }});
-        } else {
-          // just display title
-          Search.output.append(listItem);
-          setTimeout(function() {
-            displayNextItem();
-          }, 5);
-        }
-      }
-      // search finished, update title and status message
-      else {
-        Search.stopPulse();
-        Search.title.text(_('Search Results'));
-        if (!resultCount)
-          Search.status.text(_('Your search did not match any documents. Please make sure that all words are spelled correctly and that you\'ve selected enough categories.'));
-        else
-            Search.status.text(_('Search finished, found %s page(s) matching the search query.').replace('%s', resultCount));
-        Search.status.fadeIn(500);
-      }
-    }
-    displayNextItem();
+    _displayNextItem(results, results.length, searchTerms);
   },
 
   /**
    * search for object names
    */
-  performObjectSearch : function(object, otherterms) {
-    var filenames = this._index.filenames;
-    var docnames = this._index.docnames;
-    var objects = this._index.objects;
-    var objnames = this._index.objnames;
-    var titles = this._index.titles;
-
-    var i;
-    var results = [];
-
-    for (var prefix in objects) {
-      for (var iMatch = 0; iMatch != objects[prefix].length; ++iMatch) {
-        var match = objects[prefix][iMatch];
-        var name = match[4];
-        var fullname = (prefix ? prefix + '.' : '') + name;
-        var fullnameLower = fullname.toLowerCase()
-        if (fullnameLower.indexOf(object) > -1) {
-          var score = 0;
-          var parts = fullnameLower.split('.');
-          // check for different match types: exact matches of full name or
-          // "last name" (i.e. last dotted part)
-          if (fullnameLower == object || parts[parts.length - 1] == object) {
-            score += Scorer.objNameMatch;
-          // matches in last name
-          } else if (parts[parts.length - 1].indexOf(object) > -1) {
-            score += Scorer.objPartialMatch;
-          }
-          var objname = objnames[match[1]][2];
-          var title = titles[match[0]];
-          // If more than one term searched for, we require other words to be
-          // found in the name/title/description
-          if (otherterms.length > 0) {
-            var haystack = (prefix + ' ' + name + ' ' +
-                            objname + ' ' + title).toLowerCase();
-            var allfound = true;
-            for (i = 0; i < otherterms.length; i++) {
-              if (haystack.indexOf(otherterms[i]) == -1) {
-                allfound = false;
-                break;
-              }
-            }
-            if (!allfound) {
-              continue;
-            }
-          }
-          var descr = objname + _(', in ') + title;
-
-          var anchor = match[3];
-          if (anchor === '')
-            anchor = fullname;
-          else if (anchor == '-')
-            anchor = objnames[match[1]][1] + '-' + fullname;
-          // add custom score for some objects according to scorer
-          if (Scorer.objPrio.hasOwnProperty(match[2])) {
-            score += Scorer.objPrio[match[2]];
-          } else {
-            score += Scorer.objPrioDefault;
-          }
-          results.push([docnames[match[0]], fullname, '#'+anchor, descr, score, filenames[match[0]]]);
-        }
+  performObjectSearch: (object, objectTerms) => {
+    const filenames = Search._index.filenames;
+    const docNames = Search._index.docnames;
+    const objects = Search._index.objects;
+    const objNames = Search._index.objnames;
+    const titles = Search._index.titles;
+
+    const results = [];
+
+    const objectSearchCallback = (prefix, match) => {
+      const name = match[4]
+      const fullname = (prefix ? prefix + "." : "") + name;
+      const fullnameLower = fullname.toLowerCase();
+      if (fullnameLower.indexOf(object) < 0) return;
+
+      let score = 0;
+      const parts = fullnameLower.split(".");
+
+      // check for different match types: exact matches of full name or
+      // "last name" (i.e. last dotted part)
+      if (fullnameLower === object || parts.slice(-1)[0] === object)
+        score += Scorer.objNameMatch;
+      else if (parts.slice(-1)[0].indexOf(object) > -1)
+        score += Scorer.objPartialMatch; // matches in last name
+
+      const objName = objNames[match[1]][2];
+      const title = titles[match[0]];
+
+      // If more than one term searched for, we require other words to be
+      // found in the name/title/description
+      const otherTerms = new Set(objectTerms);
+      otherTerms.delete(object);
+      if (otherTerms.size > 0) {
+        const haystack = `${prefix} ${name} ${objName} ${title}`.toLowerCase();
+        if (
+          [...otherTerms].some((otherTerm) => haystack.indexOf(otherTerm) < 0)
+        )
+          return;
       }
-    }
 
+      let anchor = match[3];
+      if (anchor === "") anchor = fullname;
+      else if (anchor === "-") anchor = objNames[match[1]][1] + "-" + fullname;
+
+      const descr = objName + _(", in ") + title;
+
+      // add custom score for some objects according to scorer
+      if (Scorer.objPrio.hasOwnProperty(match[2]))
+        score += Scorer.objPrio[match[2]];
+      else score += Scorer.objPrioDefault;
+
+      results.push([
+        docNames[match[0]],
+        fullname,
+        "#" + anchor,
+        descr,
+        score,
+        filenames[match[0]],
+      ]);
+    };
+    Object.keys(objects).forEach((prefix) =>
+      objects[prefix].forEach((array) =>
+        objectSearchCallback(prefix, array)
+      )
+    );
     return results;
   },
 
-  /**
-   * See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions
-   */
-  escapeRegExp : function(string) {
-    return string.replace(/[.*+\-?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
-  },
-
   /**
    * search for full-text terms in the index
    */
-  performTermsSearch : function(searchterms, excluded, terms, titleterms) {
-    var docnames = this._index.docnames;
-    var filenames = this._index.filenames;
-    var titles = this._index.titles;
+  performTermsSearch: (searchTerms, excludedTerms) => {
+    // prepare search
+    const terms = Search._index.terms;
+    const titleTerms = Search._index.titleterms;
+    const filenames = Search._index.filenames;
+    const docNames = Search._index.docnames;
+    const titles = Search._index.titles;
 
-    var i, j, file;
-    var fileMap = {};
-    var scoreMap = {};
-    var results = [];
+    const scoreMap = new Map();
+    const fileMap = new Map();
 
     // perform the search on the required terms
-    for (i = 0; i < searchterms.length; i++) {
-      var word = searchterms[i];
-      var files = [];
-      var _o = [
-        {files: terms[word], score: Scorer.term},
-        {files: titleterms[word], score: Scorer.title}
+    searchTerms.forEach((word) => {
+      const files = [];
+      const arr = [
+        { files: terms[word], score: Scorer.term },
+        { files: titleTerms[word], score: Scorer.title },
       ];
       // add support for partial matches
       if (word.length > 2) {
-        var word_regex = this.escapeRegExp(word);
-        for (var w in terms) {
-          if (w.match(word_regex) && !terms[word]) {
-            _o.push({files: terms[w], score: Scorer.partialTerm})
-          }
-        }
-        for (var w in titleterms) {
-          if (w.match(word_regex) && !titleterms[word]) {
-              _o.push({files: titleterms[w], score: Scorer.partialTitle})
-          }
-        }
+        const escapedWord = _escapeRegExp(word);
+        Object.keys(terms).forEach((term) => {
+          if (term.match(escapedWord) && !terms[word])
+            arr.push({ files: terms[term], score: Scorer.partialTerm });
+        });
+        Object.keys(titleTerms).forEach((term) => {
+          if (term.match(escapedWord) && !titleTerms[word])
+            arr.push({ files: titleTerms[word], score: Scorer.partialTitle });
+        });
       }
 
       // no match but word was a required one
-      if ($u.every(_o, function(o){return o.files === undefined;})) {
-        break;
-      }
+      if (arr.every((record) => record.files === undefined)) return;
+
       // found search word in contents
-      $u.each(_o, function(o) {
-        var _files = o.files;
-        if (_files === undefined)
-          return
-
-        if (_files.length === undefined)
-          _files = [_files];
-        files = files.concat(_files);
-
-        // set score for the word in each file to Scorer.term
-        for (j = 0; j < _files.length; j++) {
-          file = _files[j];
-          if (!(file in scoreMap))
-            scoreMap[file] = {};
-          scoreMap[file][word] = o.score;
-        }
+      arr.forEach((record) => {
+        if (record.files === undefined) return;
+
+        let recordFiles = record.files;
+        if (recordFiles.length === undefined) recordFiles = [recordFiles];
+        files.push(...recordFiles);
+
+        // set score for the word in each file
+        recordFiles.forEach((file) => {
+          if (!scoreMap.has(file)) scoreMap.set(file, {});
+          scoreMap.get(file)[word] = record.score;
+        });
       });
 
       // create the mapping
-      for (j = 0; j < files.length; j++) {
-        file = files[j];
-        if (file in fileMap && fileMap[file].indexOf(word) === -1)
-          fileMap[file].push(word);
-        else
-          fileMap[file] = [word];
-      }
-    }
+      files.forEach((file) => {
+        if (fileMap.has(file) && fileMap.get(file).indexOf(word) === -1)
+          fileMap.get(file).push(word);
+        else fileMap.set(file, [word]);
+      });
+    });
 
     // now check if the files don't contain excluded terms
-    for (file in fileMap) {
-      var valid = true;
-
+    const results = [];
+    for (const [file, wordList] of fileMap) {
       // check if all requirements are matched
-      var filteredTermCount = // as search terms with length < 3 are discarded: ignore
-        searchterms.filter(function(term){return term.length > 2}).length
+
+      // as search terms with length < 3 are discarded
+      const filteredTermCount = [...searchTerms].filter(
+        (term) => term.length > 2
+      ).length;
       if (
-        fileMap[file].length != searchterms.length &&
-        fileMap[file].length != filteredTermCount
-      ) continue;
+        wordList.length !== searchTerms.size &&
+        wordList.length !== filteredTermCount
+      )
+        continue;
 
       // ensure that none of the excluded terms is in the search result
-      for (i = 0; i < excluded.length; i++) {
-        if (terms[excluded[i]] == file ||
-            titleterms[excluded[i]] == file ||
-            $u.contains(terms[excluded[i]] || [], file) ||
-            $u.contains(titleterms[excluded[i]] || [], file)) {
-          valid = false;
-          break;
-        }
-      }
+      if (
+        [...excludedTerms].some(
+          (term) =>
+            terms[term] === file ||
+            titleTerms[term] === file ||
+            (terms[term] || []).includes(file) ||
+            (titleTerms[term] || []).includes(file)
+        )
+      )
+        break;
 
-      // if we have still a valid result we can add it to the result list
-      if (valid) {
-        // select one (max) score for the file.
-        // for better ranking, we should calculate ranking by using words statistics like basic tf-idf...
-        var score = $u.max($u.map(fileMap[file], function(w){return scoreMap[file][w]}));
-        results.push([docnames[file], titles[file], '', null, score, filenames[file]]);
-      }
+      // select one (max) score for the file.
+      const score = Math.max(...wordList.map((w) => scoreMap.get(file)[w]));
+      // add result to the result list
+      results.push([
+        docNames[file],
+        titles[file],
+        "",
+        null,
+        score,
+        filenames[file],
+      ]);
     }
     return results;
   },
@@ -492,34 +539,28 @@ var Search = {
   /**
    * helper function to return a node containing the
    * search summary for a given text. keywords is a list
-   * of stemmed words, hlwords is the list of normal, unstemmed
-   * words. the first one is used to find the occurrence, the
-   * latter for highlighting it.
+   * of stemmed words.
    */
-  makeSearchSummary : function(htmlText, keywords, hlwords) {
-    var text = Search.htmlToText(htmlText);
-    if (text == "") {
-      return null;
-    }
-    var textLower = text.toLowerCase();
-    var start = 0;
-    $.each(keywords, function() {
-      var i = textLower.indexOf(this.toLowerCase());
-      if (i > -1)
-        start = i;
-    });
-    start = Math.max(start - 120, 0);
-    var excerpt = ((start > 0) ? '...' : '') +
-      $.trim(text.substr(start, 240)) +
-      ((start + 240 - text.length) ? '...' : '');
-    var rv = $('<p class="context"></p>').text(excerpt);
-    $.each(hlwords, function() {
-      rv = rv.highlightText(this, 'highlighted');
-    });
-    return rv;
-  }
+  makeSearchSummary: (htmlText, keywords) => {
+    const text = Search.htmlToText(htmlText);
+    if (text === "") return null;
+
+    const textLower = text.toLowerCase();
+    const actualStartPosition = [...keywords]
+      .map((k) => textLower.indexOf(k.toLowerCase()))
+      .filter((i) => i > -1)
+      .slice(-1)[0];
+    const startWithContext = Math.max(actualStartPosition - 120, 0);
+
+    const top = startWithContext === 0 ? "" : "...";
+    const tail = startWithContext + 240 < text.length ? "..." : "";
+
+    let summary = document.createElement("p");
+    summary.classList.add("context");
+    summary.textContent = top + text.substr(startWithContext, 240).trim() + tail;
+
+    return summary;
+  },
 };
 
-$(document).ready(function() {
-  Search.init();
-});
+_ready(Search.init);
diff --git a/_static/sphinx_highlight.js b/_static/sphinx_highlight.js
new file mode 100644
index 000000000..aae669d7e
--- /dev/null
+++ b/_static/sphinx_highlight.js
@@ -0,0 +1,144 @@
+/* Highlighting utilities for Sphinx HTML documentation. */
+"use strict";
+
+const SPHINX_HIGHLIGHT_ENABLED = true
+
+/**
+ * highlight a given string on a node by wrapping it in
+ * span elements with the given class name.
+ */
+const _highlight = (node, addItems, text, className) => {
+  if (node.nodeType === Node.TEXT_NODE) {
+    const val = node.nodeValue;
+    const parent = node.parentNode;
+    const pos = val.toLowerCase().indexOf(text);
+    if (
+      pos >= 0 &&
+      !parent.classList.contains(className) &&
+      !parent.classList.contains("nohighlight")
+    ) {
+      let span;
+
+      const closestNode = parent.closest("body, svg, foreignObject");
+      const isInSVG = closestNode && closestNode.matches("svg");
+      if (isInSVG) {
+        span = document.createElementNS("http://www.w3.org/2000/svg", "tspan");
+      } else {
+        span = document.createElement("span");
+        span.classList.add(className);
+      }
+
+      span.appendChild(document.createTextNode(val.substr(pos, text.length)));
+      parent.insertBefore(
+        span,
+        parent.insertBefore(
+          document.createTextNode(val.substr(pos + text.length)),
+          node.nextSibling
+        )
+      );
+      node.nodeValue = val.substr(0, pos);
+
+      if (isInSVG) {
+        const rect = document.createElementNS(
+          "http://www.w3.org/2000/svg",
+          "rect"
+        );
+        const bbox = parent.getBBox();
+        rect.x.baseVal.value = bbox.x;
+        rect.y.baseVal.value = bbox.y;
+        rect.width.baseVal.value = bbox.width;
+        rect.height.baseVal.value = bbox.height;
+        rect.setAttribute("class", className);
+        addItems.push({ parent: parent, target: rect });
+      }
+    }
+  } else if (node.matches && !node.matches("button, select, textarea")) {
+    node.childNodes.forEach((el) => _highlight(el, addItems, text, className));
+  }
+};
+const _highlightText = (thisNode, text, className) => {
+  let addItems = [];
+  _highlight(thisNode, addItems, text, className);
+  addItems.forEach((obj) =>
+    obj.parent.insertAdjacentElement("beforebegin", obj.target)
+  );
+};
+
+/**
+ * Small JavaScript module for the documentation.
+ */
+const SphinxHighlight = {
+
+  /**
+   * highlight the search words provided in localstorage in the text
+   */
+  highlightSearchWords: () => {
+    if (!SPHINX_HIGHLIGHT_ENABLED) return;  // bail if no highlight
+
+    // get and clear terms from localstorage
+    const url = new URL(window.location);
+    const highlight =
+        localStorage.getItem("sphinx_highlight_terms")
+        || url.searchParams.get("highlight")
+        || "";
+    localStorage.removeItem("sphinx_highlight_terms")
+    url.searchParams.delete("highlight");
+    window.history.replaceState({}, "", url);
+
+    // get individual terms from highlight string
+    const terms = highlight.toLowerCase().split(/\s+/).filter(x => x);
+    if (terms.length === 0) return; // nothing to do
+
+    // There should never be more than one element matching "div.body"
+    const divBody = document.querySelectorAll("div.body");
+    const body = divBody.length ? divBody[0] : document.querySelector("body");
+    window.setTimeout(() => {
+      terms.forEach((term) => _highlightText(body, term, "highlighted"));
+    }, 10);
+
+    const searchBox = document.getElementById("searchbox");
+    if (searchBox === null) return;
+    searchBox.appendChild(
+      document
+        .createRange()
+        .createContextualFragment(
+          '<p class="highlight-link">' +
+            '<a href="javascript:SphinxHighlight.hideSearchWords()">' +
+            _("Hide Search Matches") +
+            "</a></p>"
+        )
+    );
+  },
+
+  /**
+   * helper function to hide the search marks again
+   */
+  hideSearchWords: () => {
+    document
+      .querySelectorAll("#searchbox .highlight-link")
+      .forEach((el) => el.remove());
+    document
+      .querySelectorAll("span.highlighted")
+      .forEach((el) => el.classList.remove("highlighted"));
+    localStorage.removeItem("sphinx_highlight_terms")
+  },
+
+  initEscapeListener: () => {
+    // only install a listener if it is really needed
+    if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS) return;
+
+    document.addEventListener("keydown", (event) => {
+      // bail for input elements
+      if (BLACKLISTED_KEY_CONTROL_ELEMENTS.has(document.activeElement.tagName)) return;
+      // bail with special keys
+      if (event.shiftKey || event.altKey || event.ctrlKey || event.metaKey) return;
+      if (DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS && (event.key === "Escape")) {
+        SphinxHighlight.hideSearchWords();
+        event.preventDefault();
+      }
+    });
+  },
+};
+
+_ready(SphinxHighlight.highlightSearchWords);
+_ready(SphinxHighlight.initEscapeListener);
diff --git a/_static/tabs.css b/_static/tabs.css
new file mode 100644
index 000000000..957ba60d6
--- /dev/null
+++ b/_static/tabs.css
@@ -0,0 +1,89 @@
+.sphinx-tabs {
+  margin-bottom: 1rem;
+}
+
+[role="tablist"] {
+  border-bottom: 1px solid #a0b3bf;
+}
+
+.sphinx-tabs-tab {
+  position: relative;
+  font-family: Lato,'Helvetica Neue',Arial,Helvetica,sans-serif;
+  color: #1D5C87;
+  line-height: 24px;
+  margin: 0;
+  font-size: 16px;
+  font-weight: 400;
+  background-color: rgba(255, 255, 255, 0);
+  border-radius: 5px 5px 0 0;
+  border: 0;
+  padding: 1rem 1.5rem;
+  margin-bottom: 0;
+}
+
+.sphinx-tabs-tab[aria-selected="true"] {
+  font-weight: 700;
+  border: 1px solid #a0b3bf;
+  border-bottom: 1px solid white;
+  margin: -1px;
+  background-color: white;
+}
+
+.sphinx-tabs-tab:focus {
+  z-index: 1;
+  outline-offset: 1px;
+}
+
+.sphinx-tabs-panel {
+  position: relative;
+  padding: 1rem;
+  border: 1px solid #a0b3bf;
+  margin: 0px -1px -1px -1px;
+  border-radius: 0 0 5px 5px;
+  border-top: 0;
+  background: white;
+}
+
+.sphinx-tabs-panel.code-tab {
+  padding: 0.4rem;
+}
+
+.sphinx-tab img {
+	margin-bottom: 24 px;
+}
+
+/* Dark theme preference styling */
+
+@media (prefers-color-scheme: dark) {
+  body[data-theme="auto"] .sphinx-tabs-panel {
+    color: white;
+    background-color: rgb(50, 50, 50);
+  }
+
+  body[data-theme="auto"] .sphinx-tabs-tab {
+    color: white;
+    background-color: rgba(255, 255, 255, 0.05);
+  }
+
+  body[data-theme="auto"] .sphinx-tabs-tab[aria-selected="true"] {
+    border-bottom: 1px solid rgb(50, 50, 50);
+    background-color: rgb(50, 50, 50);
+  }
+}
+
+/* Explicit dark theme styling */
+
+body[data-theme="dark"] .sphinx-tabs-panel {
+  color: white;
+  background-color: rgb(50, 50, 50);
+}
+
+body[data-theme="dark"] .sphinx-tabs-tab {
+  color: white;
+  background-color: rgba(255, 255, 255, 0.05);
+}
+
+body[data-theme="dark"] .sphinx-tabs-tab[aria-selected="true"] {
+  border-bottom: 2px solid rgb(50, 50, 50);
+  background-color: rgb(50, 50, 50);
+}
diff --git a/_static/tabs.js b/_static/tabs.js
new file mode 100644
index 000000000..48dc303c8
--- /dev/null
+++ b/_static/tabs.js
@@ -0,0 +1,145 @@
+try {
+  var session = window.sessionStorage || {};
+} catch (e) {
+  var session = {};
+}
+
+window.addEventListener("DOMContentLoaded", () => {
+  const allTabs = document.querySelectorAll('.sphinx-tabs-tab');
+  const tabLists = document.querySelectorAll('[role="tablist"]');
+
+  allTabs.forEach(tab => {
+    tab.addEventListener("click", changeTabs);
+  });
+
+  tabLists.forEach(tabList => {
+    tabList.addEventListener("keydown", keyTabs);
+  });
+
+  // Restore group tab selection from session
+  const lastSelected = session.getItem('sphinx-tabs-last-selected');
+  if (lastSelected != null) selectNamedTabs(lastSelected);
+});
+
+/**
+ * Key focus left and right between sibling elements using arrows
+ * @param  {Node} e the element in focus when key was pressed
+ */
+function keyTabs(e) {
+    const tab = e.target;
+    let nextTab = null;
+    if (e.keyCode === 39 || e.keyCode === 37) {
+      tab.setAttribute("tabindex", -1);
+      // Move right
+      if (e.keyCode === 39) {
+        nextTab = tab.nextElementSibling;
+        if (nextTab === null) {
+          nextTab = tab.parentNode.firstElementChild;
+        }
+      // Move left
+      } else if (e.keyCode === 37) {
+        nextTab = tab.previousElementSibling;
+        if (nextTab === null) {
+          nextTab = tab.parentNode.lastElementChild;
+        }
+      }
+    }
+
+    if (nextTab !== null) {
+      nextTab.setAttribute("tabindex", 0);
+      nextTab.focus();
+    }
+}
+
+/**
+ * Select or deselect clicked tab. If a group tab
+ * is selected, also select tab in other tabLists.
+ * @param  {Node} e the element that was clicked
+ */
+function changeTabs(e) {
+  // Use this instead of the element that was clicked, in case it's a child
+  const notSelected = this.getAttribute("aria-selected") === "false";
+  const positionBefore = this.parentNode.getBoundingClientRect().top;
+  const notClosable = !this.parentNode.classList.contains("closeable");
+
+  deselectTabList(this);
+
+  if (notSelected || notClosable) {
+    selectTab(this);
+    const name = this.getAttribute("name");
+    selectNamedTabs(name, this.id);
+
+    if (this.classList.contains("group-tab")) {
+      // Persist during session
+      session.setItem('sphinx-tabs-last-selected', name);
+    }
+  }
+
+  const positionAfter = this.parentNode.getBoundingClientRect().top;
+  const positionDelta = positionAfter - positionBefore;
+  // Scroll to offset content resizing
+  window.scrollTo(0, window.scrollY + positionDelta);
+}
+
+/**
+ * Select tab and show associated panel.
+ * @param  {Node} tab tab to select
+ */
+function selectTab(tab) {
+  tab.setAttribute("aria-selected", true);
+
+  // Show the associated panel
+  document
+    .getElementById(tab.getAttribute("aria-controls"))
+    .removeAttribute("hidden");
+}
+
+/**
+ * Hide the panels associated with all tabs within the
+ * tablist containing this tab.
+ * @param  {Node} tab a tab within the tablist to deselect
+ */
+function deselectTabList(tab) {
+  const parent = tab.parentNode;
+  const grandparent = parent.parentNode;
+
+  Array.from(parent.children)
+  .forEach(t => t.setAttribute("aria-selected", false));
+
+  Array.from(grandparent.children)
+    .slice(1)  // Skip tablist
+    .forEach(panel => panel.setAttribute("hidden", true));
+}
+
+/**
+ * Select grouped tabs with the same name, but no the tab
+ * with the given id.
+ * @param  {Node} name name of grouped tab to be selected
+ * @param  {Node} clickedId id of clicked tab
+ */
+function selectNamedTabs(name, clickedId=null) {
+  const groupedTabs = document.querySelectorAll(`.sphinx-tabs-tab[name="${name}"]`);
+  const tabLists = Array.from(groupedTabs).map(tab => tab.parentNode);
+
+  tabLists
+    .forEach(tabList => {
+      // Don't want to change the tabList containing the clicked tab
+      const clickedTab = tabList.querySelector(`[id="${clickedId}"]`);
+      if (clickedTab === null ) {
+        // Select first tab with matching name
+        const tab = tabList.querySelector(`.sphinx-tabs-tab[name="${name}"]`);
+        deselectTabList(tab);
+        selectTab(tab);
+      }
+    })
+}
+
+if (typeof exports === 'undefined') {
+  exports = {};
+}
+
+exports.keyTabs = keyTabs;
+exports.changeTabs = changeTabs;
+exports.selectTab = selectTab;
+exports.deselectTabList = deselectTabList;
+exports.selectNamedTabs = selectNamedTabs;
diff --git a/api.html b/api.html
index aa0c52f6a..b31aa6b48 100644
--- a/api.html
+++ b/api.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Developer API &#8212; SLEAP (v1.3.3)</title>
+    <title>Developer API &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api';</script>
     <link rel="canonical" href="/develop/api.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Datasets" href="datasets.html" />
     <link rel="prev" title="Interactive and realtime inference" href="notebooks/Interactive_and_realtime_inference.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Developer API</h1>
                 <article class="bd-article" role="main">
                   
   <section id="developer-api">
-<span id="api"></span><h1>Developer API<a class="headerlink" href="#developer-api" title="Permalink to this headline">#</a></h1>
+<span id="api"></span><h1>Developer API<a class="headerlink" href="#developer-api" title="Permalink to this heading">#</a></h1>
 <div class="admonition hint">
 <p class="admonition-title">Hint</p>
 <p>Head to our <a class="reference external" href="https://github.com/talmolab/sleap/discussions">discussions page</a> to get help from the SLEAP community.</p>
@@ -328,10 +326,6 @@ <h1>Developer API</h1>
 <div class="toctree-wrapper compound">
 </div>
 <table class="autosummary longtable table autosummary">
-<colgroup>
-<col style="width: 10%" />
-<col style="width: 90%" />
-</colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.instance.html#module-sleap.instance" title="sleap.instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.instance</span></code></a></p></td>
 <td><p>Data structures for all labeled data contained with a SLEAP project.</p></td>
@@ -459,6 +453,24 @@ <h1>Developer API</h1>
 <tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.data.utils.html#module-sleap.nn.data.utils" title="sleap.nn.data.utils"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.utils</span></code></a></p></td>
 <td><p>Miscellaneous utility functions for data processing.</p></td>
 </tr>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.data.html#module-sleap.nn.config.data" title="sleap.nn.config.data"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.data</span></code></a></p></td>
+<td><p></p></td>
+</tr>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.model.html#module-sleap.nn.config.model" title="sleap.nn.config.model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.model</span></code></a></p></td>
+<td><p></p></td>
+</tr>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.optimization.html#module-sleap.nn.config.optimization" title="sleap.nn.config.optimization"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.optimization</span></code></a></p></td>
+<td><p></p></td>
+</tr>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.outputs.html#module-sleap.nn.config.outputs" title="sleap.nn.config.outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.outputs</span></code></a></p></td>
+<td><p></p></td>
+</tr>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.training_job.html#module-sleap.nn.config.training_job" title="sleap.nn.config.training_job"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.training_job</span></code></a></p></td>
+<td><p>Serializable configuration classes for specifying all training job parameters.</p></td>
+</tr>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.utils.html#module-sleap.nn.config.utils" title="sleap.nn.config.utils"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.utils</span></code></a></p></td>
+<td><p>Utilities for config building and validation.</p></td>
+</tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.architectures.common.html#module-sleap.nn.architectures.common" title="sleap.nn.architectures.common"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.architectures.common</span></code></a></p></td>
 <td><p>Common utilities for architecture and model building.</p></td>
 </tr>
@@ -486,103 +498,82 @@ <h1>Developer API</h1>
 <tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.architectures.upsampling.html#module-sleap.nn.architectures.upsampling" title="sleap.nn.architectures.upsampling"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.architectures.upsampling</span></code></a></p></td>
 <td><p>This module defines common upsampling layer stack configurations.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.data.html#module-sleap.nn.config.data" title="sleap.nn.config.data"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.data</span></code></a></p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.model.html#module-sleap.nn.config.model" title="sleap.nn.config.model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.model</span></code></a></p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.optimization.html#module-sleap.nn.config.optimization" title="sleap.nn.config.optimization"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.optimization</span></code></a></p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.outputs.html#module-sleap.nn.config.outputs" title="sleap.nn.config.outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.outputs</span></code></a></p></td>
-<td><p></p></td>
-</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.config.training_job.html#module-sleap.nn.config.training_job" title="sleap.nn.config.training_job"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.training_job</span></code></a></p></td>
-<td><p>Serializable configuration classes for specifying all training job parameters.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.config.utils.html#module-sleap.nn.config.utils" title="sleap.nn.config.utils"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.utils</span></code></a></p></td>
-<td><p>Utilities for config building and validation.</p></td>
-</tr>
 <tr class="row-even"><td><p><a class="reference internal" href="api/sleap.nn.tracker.components.html#module-sleap.nn.tracker.components" title="sleap.nn.tracker.components"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.tracker.components</span></code></a></p></td>
 <td><p>Functions/classes used by multiple trackers.</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.nn.tracker.kalman.html#module-sleap.nn.tracker.kalman" title="sleap.nn.tracker.kalman"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.tracker.kalman</span></code></a></p></td>
 <td><p>Module to use Kalman filters for tracking instance identities.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.asyncvideo.html#module-sleap.io.asyncvideo" title="sleap.io.asyncvideo"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.asyncvideo</span></code></a></p></td>
-<td><p>Support for loading video frames (by chunk) in background process.</p></td>
-</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.convert.html#module-sleap.io.convert" title="sleap.io.convert"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.convert</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.convert.html#module-sleap.io.convert" title="sleap.io.convert"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.convert</span></code></a></p></td>
 <td><p>Command line utility for converting between various dataset formats.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.dataset.html#module-sleap.io.dataset" title="sleap.io.dataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.dataset</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.dataset.html#module-sleap.io.dataset" title="sleap.io.dataset"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.dataset</span></code></a></p></td>
 <td><p>A SLEAP dataset collects labeled video frames, together with required metadata.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.legacy.html#module-sleap.io.legacy" title="sleap.io.legacy"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.legacy</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.legacy.html#module-sleap.io.legacy" title="sleap.io.legacy"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.legacy</span></code></a></p></td>
 <td><p>Module for legacy LEAP dataset.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.pathutils.html#module-sleap.io.pathutils" title="sleap.io.pathutils"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.pathutils</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.pathutils.html#module-sleap.io.pathutils" title="sleap.io.pathutils"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.pathutils</span></code></a></p></td>
 <td><p>Utilities for working with file paths.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.video.html#module-sleap.io.video" title="sleap.io.video"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.video</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.video.html#module-sleap.io.video" title="sleap.io.video"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.video</span></code></a></p></td>
 <td><p>Video reading and writing interfaces for different formats.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.videowriter.html#module-sleap.io.videowriter" title="sleap.io.videowriter"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.videowriter</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.videowriter.html#module-sleap.io.videowriter" title="sleap.io.videowriter"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.videowriter</span></code></a></p></td>
 <td><p>Module for writing avi/mp4 videos.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.visuals.html#module-sleap.io.visuals" title="sleap.io.visuals"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.visuals</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.visuals.html#module-sleap.io.visuals" title="sleap.io.visuals"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.visuals</span></code></a></p></td>
 <td><p>Module for generating videos with visual annotation overlays.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.adaptor.html#module-sleap.io.format.adaptor" title="sleap.io.format.adaptor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.adaptor</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.adaptor.html#module-sleap.io.format.adaptor" title="sleap.io.format.adaptor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.adaptor</span></code></a></p></td>
 <td><p>File format adaptor base class.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.alphatracker.html#module-sleap.io.format.alphatracker" title="sleap.io.format.alphatracker"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.alphatracker</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.alphatracker.html#module-sleap.io.format.alphatracker" title="sleap.io.format.alphatracker"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.alphatracker</span></code></a></p></td>
 <td><p>Adaptor for reading AlphaTracker datasets.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.coco.html#module-sleap.io.format.coco" title="sleap.io.format.coco"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.coco</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.coco.html#module-sleap.io.format.coco" title="sleap.io.format.coco"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.coco</span></code></a></p></td>
 <td><p>Adaptor for reading COCO keypoint detection datasets.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.csv.html#module-sleap.io.format.csv" title="sleap.io.format.csv"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.csv</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.csv.html#module-sleap.io.format.csv" title="sleap.io.format.csv"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.csv</span></code></a></p></td>
 <td><p>Adaptor for writing SLEAP analysis as csv.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.deeplabcut.html#module-sleap.io.format.deeplabcut" title="sleap.io.format.deeplabcut"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.deeplabcut</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.deeplabcut.html#module-sleap.io.format.deeplabcut" title="sleap.io.format.deeplabcut"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.deeplabcut</span></code></a></p></td>
 <td><p>Adaptor for reading DeepLabCut datasets.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.deepposekit.html#module-sleap.io.format.deepposekit" title="sleap.io.format.deepposekit"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.deepposekit</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.deepposekit.html#module-sleap.io.format.deepposekit" title="sleap.io.format.deepposekit"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.deepposekit</span></code></a></p></td>
 <td><p>Adaptor for reading DeepPoseKit datasets (HDF5).</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.dispatch.html#module-sleap.io.format.dispatch" title="sleap.io.format.dispatch"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.dispatch</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.dispatch.html#module-sleap.io.format.dispatch" title="sleap.io.format.dispatch"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.dispatch</span></code></a></p></td>
 <td><p>Dispatcher for dynamically supporting multiple dataset file formats.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.filehandle.html#module-sleap.io.format.filehandle" title="sleap.io.format.filehandle"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.filehandle</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.filehandle.html#module-sleap.io.format.filehandle" title="sleap.io.format.filehandle"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.filehandle</span></code></a></p></td>
 <td><p>File object which can be passed to adaptors.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.genericjson.html#module-sleap.io.format.genericjson" title="sleap.io.format.genericjson"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.genericjson</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.genericjson.html#module-sleap.io.format.genericjson" title="sleap.io.format.genericjson"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.genericjson</span></code></a></p></td>
 <td><p>Adaptor for reading and writing any generic JSON file.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.hdf5.html#module-sleap.io.format.hdf5" title="sleap.io.format.hdf5"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.hdf5</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.hdf5.html#module-sleap.io.format.hdf5" title="sleap.io.format.hdf5"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.hdf5</span></code></a></p></td>
 <td><p>Adaptor for reading/writing SLEAP datasets as HDF5 (including <code class="xref py py-obj docutils literal notranslate"><span class="pre">slp</span></code>).</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.labels_json.html#module-sleap.io.format.labels_json" title="sleap.io.format.labels_json"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.labels_json</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.labels_json.html#module-sleap.io.format.labels_json" title="sleap.io.format.labels_json"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.labels_json</span></code></a></p></td>
 <td><p>Adaptor for reading/writing old, JSON dataset format (kind of deprecated).</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.leap_matlab.html#module-sleap.io.format.leap_matlab" title="sleap.io.format.leap_matlab"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.leap_matlab</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.leap_matlab.html#module-sleap.io.format.leap_matlab" title="sleap.io.format.leap_matlab"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.leap_matlab</span></code></a></p></td>
 <td><p>Adaptor to read (not write) LEAP MATLAB data files.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.main.html#module-sleap.io.format.main" title="sleap.io.format.main"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.main</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.main.html#module-sleap.io.format.main" title="sleap.io.format.main"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.main</span></code></a></p></td>
 <td><p>Read/write for multiple dataset formats.</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.ndx_pose.html#module-sleap.io.format.ndx_pose" title="sleap.io.format.ndx_pose"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.ndx_pose</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.ndx_pose.html#module-sleap.io.format.ndx_pose" title="sleap.io.format.ndx_pose"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.ndx_pose</span></code></a></p></td>
 <td><p>Adaptor to read and write ndx-pose files.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.nix.html#module-sleap.io.format.nix" title="sleap.io.format.nix"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.nix</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.nix.html#module-sleap.io.format.nix" title="sleap.io.format.nix"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.nix</span></code></a></p></td>
 <td><p></p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.sleap_analysis.html#module-sleap.io.format.sleap_analysis" title="sleap.io.format.sleap_analysis"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.sleap_analysis</span></code></a></p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.sleap_analysis.html#module-sleap.io.format.sleap_analysis" title="sleap.io.format.sleap_analysis"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.sleap_analysis</span></code></a></p></td>
 <td><p>Adaptor to read and write analysis HDF5 files.</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="api/sleap.io.format.text.html#module-sleap.io.format.text" title="sleap.io.format.text"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.text</span></code></a></p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="api/sleap.io.format.text.html#module-sleap.io.format.text" title="sleap.io.format.text"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format.text</span></code></a></p></td>
 <td><p>Adaptor for reading and writing any generic text file.</p></td>
 </tr>
 </tbody>
@@ -647,7 +638,7 @@ <h1>Developer API</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.align.html b/api/sleap.info.align.html
index 0e742b74b..62bce7dba 100644
--- a/api/sleap.info.align.html
+++ b/api/sleap.info.align.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.align &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.align &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.align';</script>
     <link rel="canonical" href="/develop/api/sleap.info.align.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,21 @@ <h1>sleap.info.align</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instance_points"><code class="docutils literal notranslate"><span class="pre">align_instance_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instances"><code class="docutils literal notranslate"><span class="pre">align_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instances_on_most_stable"><code class="docutils literal notranslate"><span class="pre">align_instances_on_most_stable()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_instances_points"><code class="docutils literal notranslate"><span class="pre">get_instances_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_mean_and_std_for_points"><code class="docutils literal notranslate"><span class="pre">get_mean_and_std_for_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_most_stable_node_pair"><code class="docutils literal notranslate"><span class="pre">get_most_stable_node_pair()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_stable_node_pairs"><code class="docutils literal notranslate"><span class="pre">get_stable_node_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_template_points_array"><code class="docutils literal notranslate"><span class="pre">get_template_points_array()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +333,7 @@ <h1>sleap.info.align</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.align">
-<span id="sleap-info-align"></span><h1>sleap.info.align<a class="headerlink" href="#module-sleap.info.align" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-align"></span><h1>sleap.info.align<a class="headerlink" href="#module-sleap.info.align" title="Permalink to this heading">#</a></h1>
 <p>Functions to align instances.</p>
 <p>Usually you’ll want to</p>
 <ol class="arabic simple">
@@ -343,49 +358,49 @@ <h1>sleap.info.align</h1>
 which doesn’t yet have all points).</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.align_instance_points">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instance_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source_points_array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_points_array</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L177-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instance_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instance_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source_points_array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_points_array</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L177-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instance_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Transforms source for best fit on to target.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.align_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotate_on_node_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L91-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotate_on_node_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L91-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Rotates every instance so that line from node_a to node_b aligns.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.align_instances_on_most_stable">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instances_on_most_stable</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_stable_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L131-L141"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instances_on_most_stable" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">align_instances_on_most_stable</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_stable_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L131-L141"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.align_instances_on_most_stable" title="Permalink to this definition">#</a></dt>
 <dd><p>Gets most stable pair of nodes and aligned instances along these nodes.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.get_instances_points">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_instances_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L212-L214"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_instances_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_instances_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L212-L214"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_instances_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns single (instance, node, 2) matrix with points for all instances.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.get_mean_and_std_for_points">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_mean_and_std_for_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">aligned_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L144-L153"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_mean_and_std_for_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_mean_and_std_for_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">aligned_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L144-L153"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_mean_and_std_for_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns mean and standard deviation for every node given aligned points.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.get_most_stable_node_pair">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_most_stable_node_pair</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L83-L88"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_most_stable_node_pair" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_most_stable_node_pair</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L83-L88"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_most_stable_node_pair" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns pair of nodes which are at stable distance (over min threshold).</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.get_stable_node_pairs">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_stable_node_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_names</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L36-L80"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_stable_node_pairs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_stable_node_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_points_arrays</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_names</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L36-L80"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_stable_node_pairs" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns sorted list of node pairs with mean and standard dev distance.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.align.get_template_points_array">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_template_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/align.py#L217-L225"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_template_points_array" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.align.</span></span><span class="sig-name descname"><span class="pre">get_template_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/align.py#L217-L225"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.align.get_template_points_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns mean of aligned points for instances.</p>
 </dd></dl>
 
@@ -413,6 +428,27 @@ <h1>sleap.info.align</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instance_points"><code class="docutils literal notranslate"><span class="pre">align_instance_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instances"><code class="docutils literal notranslate"><span class="pre">align_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.align_instances_on_most_stable"><code class="docutils literal notranslate"><span class="pre">align_instances_on_most_stable()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_instances_points"><code class="docutils literal notranslate"><span class="pre">get_instances_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_mean_and_std_for_points"><code class="docutils literal notranslate"><span class="pre">get_mean_and_std_for_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_most_stable_node_pair"><code class="docutils literal notranslate"><span class="pre">get_most_stable_node_pair()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_stable_node_pairs"><code class="docutils literal notranslate"><span class="pre">get_stable_node_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.align.get_template_points_array"><code class="docutils literal notranslate"><span class="pre">get_template_points_array()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -431,7 +467,7 @@ <h1>sleap.info.align</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.feature_suggestions.html b/api/sleap.info.feature_suggestions.html
index 795c466b0..66319ecae 100644
--- a/api/sleap.info.feature_suggestions.html
+++ b/api/sleap.info.feature_suggestions.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.feature_suggestions &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.feature_suggestions &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.feature_suggestions';</script>
     <link rel="canonical" href="/develop/api/sleap.info.feature_suggestions.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,55 @@ <h1>sleap.info.feature_suggestions</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.method"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.item_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.item_group</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.group_data"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.group_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.groupset_data"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.groupset_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.all_items"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.all_items</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.append_to_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.append_to_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.extend_group_items"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.extend_group_items()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.get_item_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.get_item_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.groups"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.groups</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.sample"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.sample()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameItem"><code class="docutils literal notranslate"><span class="pre">FrameItem</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack"><code class="docutils literal notranslate"><span class="pre">ItemStack</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.items"><code class="docutils literal notranslate"><span class="pre">ItemStack.items</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.data"><code class="docutils literal notranslate"><span class="pre">ItemStack.data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.ownership"><code class="docutils literal notranslate"><span class="pre">ItemStack.ownership</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.meta"><code class="docutils literal notranslate"><span class="pre">ItemStack.meta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.group_sets"><code class="docutils literal notranslate"><span class="pre">ItemStack.group_sets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features"><code class="docutils literal notranslate"><span class="pre">ItemStack.brisk_bag_of_features()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.current_groupset"><code class="docutils literal notranslate"><span class="pre">ItemStack.current_groupset</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.extend_ownership"><code class="docutils literal notranslate"><span class="pre">ItemStack.extend_ownership()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.flatten"><code class="docutils literal notranslate"><span class="pre">ItemStack.flatten()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_all_items_from_group"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_all_items_from_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_item_data"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_item_data()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_item_data_idxs"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_item_data_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_raw_images"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_raw_images()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.hog_bag_of_features"><code class="docutils literal notranslate"><span class="pre">ItemStack.hog_bag_of_features()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.kmeans"><code class="docutils literal notranslate"><span class="pre">ItemStack.kmeans()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.make_sample_group"><code class="docutils literal notranslate"><span class="pre">ItemStack.make_sample_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.pca"><code class="docutils literal notranslate"><span class="pre">ItemStack.pca()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.sample_groups"><code class="docutils literal notranslate"><span class="pre">ItemStack.sample_groups()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.get"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.get()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.make"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.make()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.run"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.run()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.tuples_to_suggestions()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +367,11 @@ <h1>sleap.info.feature_suggestions</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.feature_suggestions">
-<span id="sleap-info-feature-suggestions"></span><h1>sleap.info.feature_suggestions<a class="headerlink" href="#module-sleap.info.feature_suggestions" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-feature-suggestions"></span><h1>sleap.info.feature_suggestions<a class="headerlink" href="#module-sleap.info.feature_suggestions" title="Permalink to this heading">#</a></h1>
 <p>Module for generating lists of frames using frame features, pca, kmeans, etc.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameGroupSet">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">FrameGroupSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">sleap.info.feature_suggestions.FrameItem</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">groupset_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L252-L340"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">FrameGroupSet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">FrameItem</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">groupset_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L252-L340"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for a set of groups of FrameItem objects.</p>
 <p>Each item can have at most one group; each group is represented as an int.</p>
 <dl class="py attribute">
@@ -330,7 +379,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">method</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.method" title="Permalink to this definition">#</a></dt>
 <dd><p>Label for the method used to generate group set.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -341,7 +390,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">item_group</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.item_group" title="Permalink to this definition">#</a></dt>
 <dd><p>Dictionary which maps each item to its group.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dict[<a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem">sleap.info.feature_suggestions.FrameItem</a>, int]</p>
 </dd>
 </dl>
@@ -353,7 +402,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <dd><p>Dictionary of any extra data for each group;
 keys are group ids, values are dictionaries of data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dict[int, dict]</p>
 </dd>
 </dl>
@@ -364,7 +413,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">groupset_data</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.groupset_data" title="Permalink to this definition">#</a></dt>
 <dd><p>Dictionary for any data about the entire set of groups.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dict</p>
 </dd>
 </dl>
@@ -378,19 +427,19 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameGroupSet.append_to_group">
-<span class="sig-name descname"><span class="pre">append_to_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">sleap.info.feature_suggestions.FrameItem</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L271-L275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.append_to_group" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">append_to_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">FrameItem</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L271-L275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.append_to_group" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds item to group.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameGroupSet.extend_group_items">
-<span class="sig-name descname"><span class="pre">extend_group_items</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">sleap.info.feature_suggestions.FrameItem</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L277-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.extend_group_items" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">extend_group_items</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">item_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">FrameItem</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L277-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.extend_group_items" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds all items in list to group.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameGroupSet.get_item_group">
-<span class="sig-name descname"><span class="pre">get_item_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">sleap.info.feature_suggestions.FrameItem</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L282-L284"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.get_item_group" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_item_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.info.feature_suggestions.FrameItem" title="sleap.info.feature_suggestions.FrameItem"><span class="pre">FrameItem</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L282-L284"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.get_item_group" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns group that contain item.</p>
 </dd></dl>
 
@@ -402,19 +451,19 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameGroupSet.sample">
-<span class="sig-name descname"><span class="pre">sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">per_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unique_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L302-L340"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.sample" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">per_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unique_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L302-L340"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameGroupSet.sample" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns new FrameGroupSet with groups sampled from current groups.</p>
 <p>Note that the order of items in the new groups will not match order of
 items in the groups from which samples are drawn.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>per_group</strong> – The number of samples to take from each group.</p></li>
 <li><p><strong>unique_samples</strong> – Whether to ensure that there are no shared items
 in the resulting groups.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>New FrameGroupSet.</p>
 </dd>
 </dl>
@@ -424,20 +473,20 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.FrameItem">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">FrameItem</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L231-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameItem" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">FrameItem</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L231-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.FrameItem" title="Permalink to this definition">#</a></dt>
 <dd><p>Just a simple wrapper for (video, frame_idx), plus method to get image.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">ItemStack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ownership</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">meta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_sets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameGroupSet" title="sleap.info.feature_suggestions.FrameGroupSet"><span class="pre">sleap.info.feature_suggestions.FrameGroupSet</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L344-L546"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">ItemStack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ownership</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tuple</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">meta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">group_sets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.info.feature_suggestions.FrameGroupSet" title="sleap.info.feature_suggestions.FrameGroupSet"><span class="pre">FrameGroupSet</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L344-L546"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack" title="Permalink to this definition">#</a></dt>
 <dd><p>Container for items, each item can “own” one or more rows of data.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.items">
 <span class="sig-name descname"><span class="pre">items</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.items" title="Permalink to this definition">#</a></dt>
 <dd><p>The list of items</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List</p>
 </dd>
 </dl>
@@ -448,7 +497,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">data</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.data" title="Permalink to this definition">#</a></dt>
 <dd><p>An ndarray with rows of data corresponding to items.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[numpy.ndarray]</p>
 </dd>
 </dl>
@@ -460,7 +509,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <dd><p>List which specifies which rows of data correspond to which
 items.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[tuple]]</p>
 </dd>
 </dl>
@@ -471,7 +520,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">meta</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.meta" title="Permalink to this definition">#</a></dt>
 <dd><p>List which stores metadata about each operation on stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List</p>
 </dd>
 </dl>
@@ -482,7 +531,7 @@ <h1>sleap.info.feature_suggestions</h1>
 <span class="sig-name descname"><span class="pre">group_sets</span></span><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.group_sets" title="Permalink to this definition">#</a></dt>
 <dd><p>List of GroupSets of items.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="#sleap.info.feature_suggestions.FrameGroupSet" title="sleap.info.feature_suggestions.FrameGroupSet">sleap.info.feature_suggestions.FrameGroupSet</a>]</p>
 </dd>
 </dl>
@@ -490,7 +539,7 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features">
-<span class="sig-name descname"><span class="pre">brisk_bag_of_features</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">brisk_threshold</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">40</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vocab_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L443-L446"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">brisk_bag_of_features</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">brisk_threshold</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">40</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vocab_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L443-L446"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features" title="Permalink to this definition">#</a></dt>
 <dd><p>Transform data using bag of features based on brisk features.</p>
 </dd></dl>
 
@@ -502,67 +551,67 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.extend_ownership">
-<span class="sig-name descname"><span class="pre">extend_ownership</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ownership</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">row_count</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L396-L404"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.extend_ownership" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">extend_ownership</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ownership</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">row_count</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L396-L404"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.extend_ownership" title="Permalink to this definition">#</a></dt>
 <dd><p>Extends an ownership list with number of rows owned by next item.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.flatten">
-<span class="sig-name descname"><span class="pre">flatten</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L434-L441"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.flatten" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">flatten</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L434-L441"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.flatten" title="Permalink to this definition">#</a></dt>
 <dd><p>Flattens each row of data to 1-d array.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.get_all_items_from_group">
-<span class="sig-name descname"><span class="pre">get_all_items_from_group</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L511-L515"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_all_items_from_group" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_all_items_from_group</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L511-L515"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_all_items_from_group" title="Permalink to this definition">#</a></dt>
 <dd><p>Sets items for Stack to all items from current GroupSet.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.get_item_data">
-<span class="sig-name descname"><span class="pre">get_item_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L380-L383"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_item_data" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_item_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L380-L383"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_item_data" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns rows of data which belong to item.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.get_item_data_idxs">
-<span class="sig-name descname"><span class="pre">get_item_data_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L371-L378"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_item_data_idxs" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_item_data_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">item</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L371-L378"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_item_data_idxs" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns indexes of rows in data which belong to item.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.get_raw_images">
-<span class="sig-name descname"><span class="pre">get_raw_images</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L406-L432"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_raw_images" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_raw_images</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L406-L432"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.get_raw_images" title="Permalink to this definition">#</a></dt>
 <dd><p>Sets data to raw image for each FrameItem.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.hog_bag_of_features">
-<span class="sig-name descname"><span class="pre">hog_bag_of_features</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">brisk_threshold</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">40</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vocab_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L448-L451"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.hog_bag_of_features" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">hog_bag_of_features</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">brisk_threshold</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">40</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vocab_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">20</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L448-L451"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.hog_bag_of_features" title="Permalink to this definition">#</a></dt>
 <dd><p>Transforms data into bag of features vector of hog descriptors.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.kmeans">
-<span class="sig-name descname"><span class="pre">kmeans</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n_clusters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L466-L483"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.kmeans" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">kmeans</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n_clusters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L466-L483"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.kmeans" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds GroupSet using k-means clustering on data.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.make_sample_group">
-<span class="sig-name descname"><span class="pre">make_sample_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">samples_per_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stride'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L485-L509"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.make_sample_group" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_sample_group</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">samples_per_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stride'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L485-L509"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.make_sample_group" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds GroupSet by sampling frames from each video.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.pca">
-<span class="sig-name descname"><span class="pre">pca</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n_components</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L453-L464"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.pca" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">pca</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n_components</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L453-L464"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.pca" title="Permalink to this definition">#</a></dt>
 <dd><p>Transforms data by applying PCA.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ItemStack.sample_groups">
-<span class="sig-name descname"><span class="pre">sample_groups</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">samples_per_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L517-L523"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.sample_groups" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">sample_groups</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">samples_per_group</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L517-L523"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ItemStack.sample_groups" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds GroupSet by sampling items from current GroupSet.</p>
 </dd></dl>
 
@@ -570,7 +619,7 @@ <h1>sleap.info.feature_suggestions</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ParallelFeaturePipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">ParallelFeaturePipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.info.feature_suggestions.FeatureSuggestionPipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos_as_dicts</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L630-L697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.feature_suggestions.</span></span><span class="sig-name descname"><span class="pre">ParallelFeaturePipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FeatureSuggestionPipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos_as_dicts</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L630-L697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Enables easy per-video pipeline parallelization for feature suggestions.</p>
 <p>Create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">FeatureSuggestionPipeline</span></code> with the desired parameters, and
 then call <a class="reference internal" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.run" title="sleap.info.feature_suggestions.ParallelFeaturePipeline.run"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.run()</span></code></a> with the pipeline and the list
@@ -579,25 +628,25 @@ <h1>sleap.info.feature_suggestions</h1>
 the results back into a single list of <code class="xref py py-obj docutils literal notranslate"><span class="pre">SuggestionFrame</span></code> objects.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ParallelFeaturePipeline.get">
-<span class="sig-name descname"><span class="pre">get</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video_idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L644-L659"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.get" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video_idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L644-L659"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.get" title="Permalink to this definition">#</a></dt>
 <dd><p>Apply pipeline to single video by idx. Can be called in process.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ParallelFeaturePipeline.make">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L661-L665"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.make" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L661-L665"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.make" title="Permalink to this definition">#</a></dt>
 <dd><p>Make class object from pipeline and list of videos.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ParallelFeaturePipeline.run">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parallel</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L678-L697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.run" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipeline</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parallel</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L678-L697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.run" title="Permalink to this definition">#</a></dt>
 <dd><p>Runs pipeline on all videos in parallel and returns suggestions.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tuples_to_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tuples</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/feature_suggestions.py#L667-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tuples_to_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tuples</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/feature_suggestions.py#L667-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts serialized data from processes back into SuggestionFrames.</p>
 </dd></dl>
 
@@ -627,6 +676,61 @@ <h1>sleap.info.feature_suggestions</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.method"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.item_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.item_group</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.group_data"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.group_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.groupset_data"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.groupset_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.all_items"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.all_items</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.append_to_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.append_to_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.extend_group_items"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.extend_group_items()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.get_item_group"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.get_item_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.groups"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.groups</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameGroupSet.sample"><code class="docutils literal notranslate"><span class="pre">FrameGroupSet.sample()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.FrameItem"><code class="docutils literal notranslate"><span class="pre">FrameItem</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack"><code class="docutils literal notranslate"><span class="pre">ItemStack</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.items"><code class="docutils literal notranslate"><span class="pre">ItemStack.items</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.data"><code class="docutils literal notranslate"><span class="pre">ItemStack.data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.ownership"><code class="docutils literal notranslate"><span class="pre">ItemStack.ownership</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.meta"><code class="docutils literal notranslate"><span class="pre">ItemStack.meta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.group_sets"><code class="docutils literal notranslate"><span class="pre">ItemStack.group_sets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features"><code class="docutils literal notranslate"><span class="pre">ItemStack.brisk_bag_of_features()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.current_groupset"><code class="docutils literal notranslate"><span class="pre">ItemStack.current_groupset</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.extend_ownership"><code class="docutils literal notranslate"><span class="pre">ItemStack.extend_ownership()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.flatten"><code class="docutils literal notranslate"><span class="pre">ItemStack.flatten()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_all_items_from_group"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_all_items_from_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_item_data"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_item_data()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_item_data_idxs"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_item_data_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.get_raw_images"><code class="docutils literal notranslate"><span class="pre">ItemStack.get_raw_images()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.hog_bag_of_features"><code class="docutils literal notranslate"><span class="pre">ItemStack.hog_bag_of_features()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.kmeans"><code class="docutils literal notranslate"><span class="pre">ItemStack.kmeans()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.make_sample_group"><code class="docutils literal notranslate"><span class="pre">ItemStack.make_sample_group()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.pca"><code class="docutils literal notranslate"><span class="pre">ItemStack.pca()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ItemStack.sample_groups"><code class="docutils literal notranslate"><span class="pre">ItemStack.sample_groups()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.get"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.get()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.make"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.make()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.run"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.run()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions"><code class="docutils literal notranslate"><span class="pre">ParallelFeaturePipeline.tuples_to_suggestions()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -645,7 +749,7 @@ <h1>sleap.info.feature_suggestions</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.labels.html b/api/sleap.info.labels.html
index 494f096ba..b99048a97 100644
--- a/api/sleap.info.labels.html
+++ b/api/sleap.info.labels.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.labels &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.labels &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.labels';</script>
     <link rel="canonical" href="/develop/api/sleap.info.labels.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -318,7 +316,7 @@ <h1>sleap.info.labels</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.labels">
-<span id="sleap-info-labels"></span><h1>sleap.info.labels<a class="headerlink" href="#module-sleap.info.labels" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-labels"></span><h1>sleap.info.labels<a class="headerlink" href="#module-sleap.info.labels" title="Permalink to this heading">#</a></h1>
 <p>Command line utility which prints data about labels file.</p>
 </section>
 
@@ -362,7 +360,7 @@ <h1>sleap.info.labels</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.metrics.html b/api/sleap.info.metrics.html
index 8d2be41db..94dd8f63a 100644
--- a/api/sleap.info.metrics.html
+++ b/api/sleap.info.metrics.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.metrics &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.metrics &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.metrics';</script>
     <link rel="canonical" href="/develop/api/sleap.info.metrics.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,23 @@ <h1>sleap.info.metrics</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.calculate_pairwise_cost"><code class="docutils literal notranslate"><span class="pre">calculate_pairwise_cost()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.compare_instance_lists"><code class="docutils literal notranslate"><span class="pre">compare_instance_lists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.list_points_array"><code class="docutils literal notranslate"><span class="pre">list_points_array()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.match_instance_lists"><code class="docutils literal notranslate"><span class="pre">match_instance_lists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.match_instance_lists_nodewise"><code class="docutils literal notranslate"><span class="pre">match_instance_lists_nodewise()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.matched_instance_distances"><code class="docutils literal notranslate"><span class="pre">matched_instance_distances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.nodeless_point_dist"><code class="docutils literal notranslate"><span class="pre">nodeless_point_dist()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_dist"><code class="docutils literal notranslate"><span class="pre">point_dist()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_match_count"><code class="docutils literal notranslate"><span class="pre">point_match_count()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_nonmatch_count"><code class="docutils literal notranslate"><span class="pre">point_nonmatch_count()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,47 +335,47 @@ <h1>sleap.info.metrics</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.metrics">
-<span id="sleap-info-metrics"></span><h1>sleap.info.metrics<a class="headerlink" href="#module-sleap.info.metrics" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-metrics"></span><h1>sleap.info.metrics<a class="headerlink" href="#module-sleap.info.metrics" title="Permalink to this heading">#</a></h1>
 <p>Module for producing prediction metrics for SLEAP datasets.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.calculate_pairwise_cost">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">calculate_pairwise_cost</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L102-L124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.calculate_pairwise_cost" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">calculate_pairwise_cost</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L33-L55"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.calculate_pairwise_cost" title="Permalink to this definition">#</a></dt>
 <dd><p>Calculate (a * b) matrix of pairwise costs using cost function.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.compare_instance_lists">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">compare_instance_lists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L211-L222"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.compare_instance_lists" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">compare_instance_lists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L211-L222"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.compare_instance_lists" title="Permalink to this definition">#</a></dt>
 <dd><p>Given two lists of corresponding Instances, returns
 (instances * nodes) matrix of distances between corresponding nodes.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.list_points_array">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">list_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L225-L230"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.list_points_array" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">list_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L225-L230"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.list_points_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Given list of Instances, returns (instances * nodes * 2) matrix.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.match_instance_lists">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">match_instance_lists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L82-L99"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.match_instance_lists" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">match_instance_lists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L13-L30"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.match_instance_lists" title="Permalink to this definition">#</a></dt>
 <dd><p>Sorts two lists of Instances to find best overall correspondence
 for a given cost function (e.g., total distance between points).</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.match_instance_lists_nodewise">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">match_instance_lists_nodewise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L127-L165"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.match_instance_lists_nodewise" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">match_instance_lists_nodewise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L58-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.match_instance_lists_nodewise" title="Permalink to this definition">#</a></dt>
 <dd><p>For each node for each instance in the first list, pairs it with the
 closest corresponding node from <em>any</em> instance in the second list.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.matched_instance_distances">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">matched_instance_distances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_lists_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">range</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L13-L79"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.matched_instance_distances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">matched_instance_distances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt:</span> <span class="pre">~sleap.io.dataset.Labels</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr:</span> <span class="pre">~sleap.io.dataset.Labels</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_lists_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">match_instance_lists_nodewise&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range:</span> <span class="pre">~typing.Optional[range]</span> <span class="pre">=</span> <span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L99-L165"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.matched_instance_distances" title="Permalink to this definition">#</a></dt>
 <dd><p>Distances between ground truth and predicted nodes over a set of frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels_gt</strong> – the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object with ground truth data</p></li>
 <li><p><strong>labels_pr</strong> – the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object with predicted data</p></li>
@@ -369,7 +386,7 @@ <h1>sleap.info.metrics</h1>
 frame in labels_pr.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><p>frame indices map: instance idx (for other matrices) -&gt; frame idx</p></li>
 <li><p>distance matrix: (instances * nodes)</p></li>
@@ -378,7 +395,7 @@ <h1>sleap.info.metrics</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Tuple</p>
 </dd>
 </dl>
@@ -386,26 +403,26 @@ <h1>sleap.info.metrics</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.nodeless_point_dist">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">nodeless_point_dist</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inst_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inst_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L180-L208"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.nodeless_point_dist" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">nodeless_point_dist</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inst_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inst_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L180-L208"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.nodeless_point_dist" title="Permalink to this definition">#</a></dt>
 <dd><p>Given two instances, returns array of distances for closest points
 ignoring node identities.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.point_dist">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_dist</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inst_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inst_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L168-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_dist" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_dist</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inst_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inst_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L168-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_dist" title="Permalink to this definition">#</a></dt>
 <dd><p>Given two instances, returns array of distances for corresponding nodes.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.point_match_count">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_match_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_array</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L233-L235"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_match_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_match_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_array</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L233-L235"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_match_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Given an array of distances, returns number which are &lt;= threshold.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.metrics.point_nonmatch_count">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_nonmatch_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_array</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/metrics.py#L238-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_nonmatch_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.metrics.</span></span><span class="sig-name descname"><span class="pre">point_nonmatch_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_array</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/metrics.py#L238-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.metrics.point_nonmatch_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Given an array of distances, returns number which are not &lt;= threshold.</p>
 </dd></dl>
 
@@ -433,6 +450,29 @@ <h1>sleap.info.metrics</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.calculate_pairwise_cost"><code class="docutils literal notranslate"><span class="pre">calculate_pairwise_cost()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.compare_instance_lists"><code class="docutils literal notranslate"><span class="pre">compare_instance_lists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.list_points_array"><code class="docutils literal notranslate"><span class="pre">list_points_array()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.match_instance_lists"><code class="docutils literal notranslate"><span class="pre">match_instance_lists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.match_instance_lists_nodewise"><code class="docutils literal notranslate"><span class="pre">match_instance_lists_nodewise()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.matched_instance_distances"><code class="docutils literal notranslate"><span class="pre">matched_instance_distances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.nodeless_point_dist"><code class="docutils literal notranslate"><span class="pre">nodeless_point_dist()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_dist"><code class="docutils literal notranslate"><span class="pre">point_dist()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_match_count"><code class="docutils literal notranslate"><span class="pre">point_match_count()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.metrics.point_nonmatch_count"><code class="docutils literal notranslate"><span class="pre">point_nonmatch_count()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -451,7 +491,7 @@ <h1>sleap.info.metrics</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.summary.html b/api/sleap.info.summary.html
index 61a0b3de3..babcd9ad5 100644
--- a/api/sleap.info.summary.html
+++ b/api/sleap.info.summary.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.summary &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.summary &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.summary';</script>
     <link rel="canonical" href="/develop/api/sleap.info.summary.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,22 @@ <h1>sleap.info.summary</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries"><code class="docutils literal notranslate"><span class="pre">StatisticSeries</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_instance_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_instance_score_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_count_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_count_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_displacement_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_displacement_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_score_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_primary_point_displacement_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_primary_point_displacement_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_tracking_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_tracking_score_series()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,34 +334,34 @@ <h1>sleap.info.summary</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.summary">
-<span id="sleap-info-summary"></span><h1>sleap.info.summary<a class="headerlink" href="#module-sleap.info.summary" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-summary"></span><h1>sleap.info.summary<a class="headerlink" href="#module-sleap.info.summary" title="Permalink to this heading">#</a></h1>
 <p>Module for getting a series which gives some statistic based on labeling
 data for each frame of some labeled video.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.summary.</span></span><span class="sig-name descname"><span class="pre">StatisticSeries</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L16-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.info.summary.</span></span><span class="sig-name descname"><span class="pre">StatisticSeries</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L16-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries" title="Permalink to this definition">#</a></dt>
 <dd><p>Class to calculate various statistical series for labeled frames.</p>
 <p>Each method returns a series which is a dictionary in which keys
 are frame index and value are some numerical value for the frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>labels</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> for which to calculate series.</p>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><p><strong>labels</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> for which to calculate series.</p>
 </dd>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_instance_score_series">
-<span class="sig-name descname"><span class="pre">get_instance_score_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L66-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_instance_score_series" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_instance_score_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L66-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_instance_score_series" title="Permalink to this definition">#</a></dt>
 <dd><p>Get series with statistic of instance scores in each frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
-<li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
+<li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
 <li><p><strong>reduction</strong> – name of function applied to scores:
 * sum
 * min</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The series dictionary (see class docs for details)</p>
 </dd>
 </dl>
@@ -353,28 +369,28 @@ <h1>sleap.info.summary</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_point_count_series">
-<span class="sig-name descname"><span class="pre">get_point_count_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L29-L36"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_count_series" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_point_count_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L29-L36"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_count_series" title="Permalink to this definition">#</a></dt>
 <dd><p>Get series with total number of labeled points in each frame.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_point_displacement_series">
-<span class="sig-name descname"><span class="pre">get_point_displacement_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L87-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_displacement_series" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_point_displacement_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L87-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_displacement_series" title="Permalink to this definition">#</a></dt>
 <dd><p>Get series with statistic of point displacement in each frame.</p>
 <p>Point displacement is the distance between the point location in
 frame and the location of the corresponding point (same node,
 same track) from the closest earlier labeled frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
-<li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
+<li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
 <li><p><strong>reduction</strong> – name of function applied to point scores:
 * sum
 * mean
 * max</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The series dictionary (see class docs for details)</p>
 </dd>
 </dl>
@@ -382,18 +398,18 @@ <h1>sleap.info.summary</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_point_score_series">
-<span class="sig-name descname"><span class="pre">get_point_score_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L38-L64"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_score_series" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_point_score_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'sum'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L38-L64"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_point_score_series" title="Permalink to this definition">#</a></dt>
 <dd><p>Get series with statistic of point scores in each frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
-<li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
+<li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
 <li><p><strong>reduction</strong> – name of function applied to scores:
 * sum
 * min</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The series dictionary (see class docs for details)</p>
 </dd>
 </dl>
@@ -401,12 +417,12 @@ <h1>sleap.info.summary</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_primary_point_displacement_series">
-<span class="sig-name descname"><span class="pre">get_primary_point_displacement_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">primary_node</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/summary.py#L117-L196"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_primary_point_displacement_series" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_primary_point_displacement_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'sum'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">primary_node</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L117-L196"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_primary_point_displacement_series" title="Permalink to this definition">#</a></dt>
 <dd><p>Get sum of displacement for single node of each instance per frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
-<li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
+<li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
 <li><p><strong>reduction</strong> – name of function applied to point scores:
 * sum
 * mean
@@ -416,7 +432,26 @@ <h1>sleap.info.summary</h1>
 then defaults to first node.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>The series dictionary (see class docs for details)</p>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.info.summary.StatisticSeries.get_tracking_score_series">
+<span class="sig-name descname"><span class="pre">get_tracking_score_series</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'min'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/summary.py#L250-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.summary.StatisticSeries.get_tracking_score_series" title="Permalink to this definition">#</a></dt>
+<dd><p>Get series with statistic of tracking scores in each frame.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> for which to calculate statistic.</p></li>
+<li><p><strong>reduction</strong> – name of function applied to scores:
+* mean
+* min</p></li>
+</ul>
+</dd>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The series dictionary (see class docs for details)</p>
 </dd>
 </dl>
@@ -448,6 +483,28 @@ <h1>sleap.info.summary</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries"><code class="docutils literal notranslate"><span class="pre">StatisticSeries</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_instance_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_instance_score_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_count_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_count_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_displacement_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_displacement_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_point_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_point_score_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_primary_point_displacement_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_primary_point_displacement_series()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.summary.StatisticSeries.get_tracking_score_series"><code class="docutils literal notranslate"><span class="pre">StatisticSeries.get_tracking_score_series()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -466,7 +523,7 @@ <h1>sleap.info.summary</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.trackcleaner.html b/api/sleap.info.trackcleaner.html
index 45e0fbd84..28f00cbe8 100644
--- a/api/sleap.info.trackcleaner.html
+++ b/api/sleap.info.trackcleaner.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.trackcleaner &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.trackcleaner &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.trackcleaner';</script>
     <link rel="canonical" href="/develop/api/sleap.info.trackcleaner.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,14 @@ <h1>sleap.info.trackcleaner</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.trackcleaner.fit_tracks"><code class="docutils literal notranslate"><span class="pre">fit_tracks()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +326,7 @@ <h1>sleap.info.trackcleaner</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.trackcleaner">
-<span id="sleap-info-trackcleaner"></span><h1>sleap.info.trackcleaner<a class="headerlink" href="#module-sleap.info.trackcleaner" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-trackcleaner"></span><h1>sleap.info.trackcleaner<a class="headerlink" href="#module-sleap.info.trackcleaner" title="Permalink to this heading">#</a></h1>
 <p>CLI for TrackCleaner (mostly deprecated).</p>
 <p>This runs the track cleaner to (i) cull instances and (ii) connect track breaks
 in predictions which have already been tracked.</p>
@@ -328,7 +336,7 @@ <h1>sleap.info.trackcleaner</h1>
 it will be better to use the <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap-track</span></code> CLI.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.trackcleaner.fit_tracks">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.trackcleaner.</span></span><span class="sig-name descname"><span class="pre">fit_tracks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/trackcleaner.py#L20-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.trackcleaner.fit_tracks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.trackcleaner.</span></span><span class="sig-name descname"><span class="pre">fit_tracks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/trackcleaner.py#L20-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.trackcleaner.fit_tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>Wraps <code class="xref py py-obj docutils literal notranslate"><span class="pre">TrackCleaner</span></code> for easier cli api.</p>
 </dd></dl>
 
@@ -356,6 +364,20 @@ <h1>sleap.info.trackcleaner</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.trackcleaner.fit_tracks"><code class="docutils literal notranslate"><span class="pre">fit_tracks()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -374,7 +396,7 @@ <h1>sleap.info.trackcleaner</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.info.write_tracking_h5.html b/api/sleap.info.write_tracking_h5.html
index c6672c934..5ee0e6e07 100644
--- a/api/sleap.info.write_tracking_h5.html
+++ b/api/sleap.info.write_tracking_h5.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.info.write_tracking_h5 &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.info.write_tracking_h5 &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.info.write_tracking_h5';</script>
     <link rel="canonical" href="/develop/api/sleap.info.write_tracking_h5.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,21 @@ <h1>sleap.info.write_tracking_h5</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_edges_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_edges_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_nodes_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_nodes_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_occupancy_and_points_matrices"><code class="docutils literal notranslate"><span class="pre">get_occupancy_and_points_matrices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_tracks_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_tracks_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices"><code class="docutils literal notranslate"><span class="pre">remove_empty_tracks_from_matrices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.write_csv_file"><code class="docutils literal notranslate"><span class="pre">write_csv_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.write_occupancy_file"><code class="docutils literal notranslate"><span class="pre">write_occupancy_file()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +333,7 @@ <h1>sleap.info.write_tracking_h5</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.info.write_tracking_h5">
-<span id="sleap-info-write-tracking-h5"></span><h1>sleap.info.write_tracking_h5<a class="headerlink" href="#module-sleap.info.write_tracking_h5" title="Permalink to this headline">#</a></h1>
+<span id="sleap-info-write-tracking-h5"></span><h1>sleap.info.write_tracking_h5<a class="headerlink" href="#module-sleap.info.write_tracking_h5" title="Permalink to this heading">#</a></h1>
 <p>Generate an HDF5 or CSV file with track occupancy and point location data.</p>
 <p>Ignores tracks that are entirely empty. By default will also ignore
 empty frames from the beginning and end of video, although
@@ -343,24 +358,24 @@ <h1>sleap.info.write_tracking_h5</h1>
 <p>Note: the datasets are stored column-major as expected by MATLAB.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.get_edges_as_np_strings">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_edges_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.bytes_</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.bytes_</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L51-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_edges_as_np_strings" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_edges_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">bytes_</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bytes_</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L51-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_edges_as_np_strings" title="Permalink to this definition">#</a></dt>
 <dd><p>Get list of edge names as <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.string_</span></code>.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.get_nodes_as_np_strings">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_nodes_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.bytes_</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L46-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_nodes_as_np_strings" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_nodes_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">bytes_</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L46-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_nodes_as_np_strings" title="Permalink to this definition">#</a></dt>
 <dd><p>Get list of node names as <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.string_</span></code>.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.get_occupancy_and_points_matrices">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_occupancy_and_points_matrices</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L59-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_occupancy_and_points_matrices" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_occupancy_and_points_matrices</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L59-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_occupancy_and_points_matrices" title="Permalink to this definition">#</a></dt>
 <dd><p>Builds numpy matrices with track occupancy and point location data.</p>
 <p>Note: This function assumes either all instances have tracks or no instances have
 tracks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> from which to get data.</p></li>
 <li><p><strong>all_frames</strong> – If True, then includes zeros so that frame index
@@ -372,7 +387,7 @@ <h1>sleap.info.write_tracking_h5</h1>
 are no labeled frames in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">video</span></code>, then None will be returned.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><p>occupancy matrix with shape (tracks, frames)</p></li>
 <li><p>point location array with shape (frames, nodes, 2, tracks)</p></li>
@@ -382,7 +397,7 @@ <h1>sleap.info.write_tracking_h5</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>tuple of arrays</p>
 </dd>
 </dl>
@@ -390,16 +405,16 @@ <h1>sleap.info.write_tracking_h5</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.get_tracks_as_np_strings">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_tracks_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.bytes_</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L41-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_tracks_as_np_strings" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">get_tracks_as_np_strings</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">bytes_</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L41-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.get_tracks_as_np_strings" title="Permalink to this definition">#</a></dt>
 <dd><p>Get list of track names as <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.string_</span></code>.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.main">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">csv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L354-L440"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.main" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">csv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L354-L440"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.main" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes HDF5 file with matrices of track occupancy and coordinates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> from which to get data.</p></li>
 <li><p><strong>output_path</strong> – Path of HDF5 file to create.</p></li>
@@ -414,7 +429,7 @@ <h1>sleap.info.write_tracking_h5</h1>
 <li><p><strong>csv</strong> – Bool to save the analysis as a csv file if set to True</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -422,10 +437,10 @@ <h1>sleap.info.write_tracking_h5</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">remove_empty_tracks_from_matrices</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">occupancy_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">locations_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L194-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">remove_empty_tracks_from_matrices</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">occupancy_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">locations_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L194-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices" title="Permalink to this definition">#</a></dt>
 <dd><p>Removes matrix rows/columns for unoccupied tracks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>track_names</strong> – List of track names</p></li>
 <li><p><strong>occupancy_matrix</strong> – 2d numpy matrix, rows correspond to tracks</p></li>
@@ -435,7 +450,7 @@ <h1>sleap.info.write_tracking_h5</h1>
 <li><p><strong>tracking_scores</strong> – 2d numpy matrix, last index is track</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>track_names, occupancy_matrix, locations_matrix, point_scores, instance_scores
 tracking_scores but without the rows/columns corresponding to unoccupied tracks.</p>
 </dd>
@@ -444,17 +459,17 @@ <h1>sleap.info.write_tracking_h5</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.write_csv_file">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">write_csv_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L290-L351"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.write_csv_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">write_csv_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L290-L351"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.write_csv_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Write CSV file with data from given dictionary.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>output_path</strong> – Path of HDF5 file.</p></li>
 <li><p><strong>data_dict</strong> – Dictionary with data to save. Keys are dataset names,
 values are the data.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -462,10 +477,10 @@ <h1>sleap.info.write_tracking_h5</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.info.write_tracking_h5.write_occupancy_file">
-<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">write_occupancy_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transpose</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/info/write_tracking_h5.py#L243-L287"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.write_occupancy_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.info.write_tracking_h5.</span></span><span class="sig-name descname"><span class="pre">write_occupancy_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">data_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transpose</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/info/write_tracking_h5.py#L243-L287"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.info.write_tracking_h5.write_occupancy_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Write HDF5 file with data from given dictionary.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>output_path</strong> – Path of HDF5 file.</p></li>
 <li><p><strong>data_dict</strong> – Dictionary with data to save. Keys are dataset names,
@@ -476,7 +491,7 @@ <h1>sleap.info.write_tracking_h5</h1>
 column-major format.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -506,6 +521,27 @@ <h1>sleap.info.write_tracking_h5</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_edges_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_edges_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_nodes_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_nodes_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_occupancy_and_points_matrices"><code class="docutils literal notranslate"><span class="pre">get_occupancy_and_points_matrices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.get_tracks_as_np_strings"><code class="docutils literal notranslate"><span class="pre">get_tracks_as_np_strings()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices"><code class="docutils literal notranslate"><span class="pre">remove_empty_tracks_from_matrices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.write_csv_file"><code class="docutils literal notranslate"><span class="pre">write_csv_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.info.write_tracking_h5.write_occupancy_file"><code class="docutils literal notranslate"><span class="pre">write_occupancy_file()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -524,7 +560,7 @@ <h1>sleap.info.write_tracking_h5</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.instance.html b/api/sleap.instance.html
index f3140d7b8..6280ce004 100644
--- a/api/sleap.instance.html
+++ b/api/sleap.instance.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.instance &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.instance &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.instance';</script>
     <link rel="canonical" href="/develop/api/sleap.instance.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,106 @@ <h1>sleap.instance</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance"><code class="docutils literal notranslate"><span class="pre">Instance</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.bounding_box"><code class="docutils literal notranslate"><span class="pre">Instance.bounding_box</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.centroid"><code class="docutils literal notranslate"><span class="pre">Instance.centroid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.fill_missing"><code class="docutils literal notranslate"><span class="pre">Instance.fill_missing()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.frame_idx"><code class="docutils literal notranslate"><span class="pre">Instance.frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.from_numpy"><code class="docutils literal notranslate"><span class="pre">Instance.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.from_pointsarray"><code class="docutils literal notranslate"><span class="pre">Instance.from_pointsarray()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.get_points_array"><code class="docutils literal notranslate"><span class="pre">Instance.get_points_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.matches"><code class="docutils literal notranslate"><span class="pre">Instance.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.midpoint"><code class="docutils literal notranslate"><span class="pre">Instance.midpoint</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.n_visible_points"><code class="docutils literal notranslate"><span class="pre">Instance.n_visible_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.nodes"><code class="docutils literal notranslate"><span class="pre">Instance.nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.nodes_points"><code class="docutils literal notranslate"><span class="pre">Instance.nodes_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.numpy"><code class="docutils literal notranslate"><span class="pre">Instance.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.points"><code class="docutils literal notranslate"><span class="pre">Instance.points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.points_array"><code class="docutils literal notranslate"><span class="pre">Instance.points_array</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.transform_points"><code class="docutils literal notranslate"><span class="pre">Instance.transform_points()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.video"><code class="docutils literal notranslate"><span class="pre">Instance.video</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList"><code class="docutils literal notranslate"><span class="pre">InstancesList</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.append"><code class="docutils literal notranslate"><span class="pre">InstancesList.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.clear"><code class="docutils literal notranslate"><span class="pre">InstancesList.clear()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.copy"><code class="docutils literal notranslate"><span class="pre">InstancesList.copy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.extend"><code class="docutils literal notranslate"><span class="pre">InstancesList.extend()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.insert"><code class="docutils literal notranslate"><span class="pre">InstancesList.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.labeled_frame"><code class="docutils literal notranslate"><span class="pre">InstancesList.labeled_frame</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.pop"><code class="docutils literal notranslate"><span class="pre">InstancesList.pop()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.remove"><code class="docutils literal notranslate"><span class="pre">InstancesList.remove()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame"><code class="docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.complex_frame_merge"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.complex_frame_merge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.complex_merge_between"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.complex_merge_between()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.find"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.find()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.image"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.index"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.insert"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.instances_to_show"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.instances_to_show</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.merge_frames"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.merge_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.numpy"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.plot"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.plot()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.plot_predicted"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.plot_predicted()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.remove_empty_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.remove_empty_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.remove_untracked"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.remove_untracked()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.training_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.training_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.unused_predictions"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.unused_predictions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.user_instances</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point"><code class="docutils literal notranslate"><span class="pre">Point</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.dtype"><code class="docutils literal notranslate"><span class="pre">Point.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.isnan"><code class="docutils literal notranslate"><span class="pre">Point.isnan()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.numpy"><code class="docutils literal notranslate"><span class="pre">Point.numpy()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray"><code class="docutils literal notranslate"><span class="pre">PointArray</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray.from_array"><code class="docutils literal notranslate"><span class="pre">PointArray.from_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray.make_default"><code class="docutils literal notranslate"><span class="pre">PointArray.make_default()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance"><code class="docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_arrays"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_arrays()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_instance"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_numpy"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_pointsarray"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_pointsarray()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.points_and_scores_array"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.points_and_scores_array</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.scores"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.scores</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint"><code class="docutils literal notranslate"><span class="pre">PredictedPoint</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint.dtype"><code class="docutils literal notranslate"><span class="pre">PredictedPoint.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint.from_point"><code class="docutils literal notranslate"><span class="pre">PredictedPoint.from_point()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPointArray"><code class="docutils literal notranslate"><span class="pre">PredictedPointArray</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPointArray.to_array"><code class="docutils literal notranslate"><span class="pre">PredictedPointArray.to_array()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Track"><code class="docutils literal notranslate"><span class="pre">Track</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Track.matches"><code class="docutils literal notranslate"><span class="pre">Track.matches()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.make_instance_cattr"><code class="docutils literal notranslate"><span class="pre">make_instance_cattr()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +418,7 @@ <h1>sleap.instance</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.instance">
-<span id="sleap-instance"></span><h1>sleap.instance<a class="headerlink" href="#module-sleap.instance" title="Permalink to this headline">#</a></h1>
+<span id="sleap-instance"></span><h1>sleap.instance<a class="headerlink" href="#module-sleap.instance" title="Permalink to this heading">#</a></h1>
 <p>Data structures for all labeled data contained with a SLEAP project.</p>
 <p>The relationships between objects in this module:</p>
 <ul class="simple">
@@ -334,10 +434,10 @@ <h1>sleap.instance</h1>
 </ul>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.Instance">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_predicted</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L344-L991"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_predicted</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L344-L991"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance" title="Permalink to this definition">#</a></dt>
 <dd><p>This class represents a labeled instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>skeleton</strong> – The skeleton that this instance is associated with.</p></li>
 <li><p><strong>points</strong> – A dictionary where keys are skeleton node names and
@@ -356,13 +456,13 @@ <h1>sleap.instance</h1>
 </dl>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.bounding_box">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">bounding_box</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.bounding_box" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">bounding_box</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.bounding_box" title="Permalink to this definition">#</a></dt>
 <dd><p>Return bounding box containing all points in <code class="xref py py-obj docutils literal notranslate"><span class="pre">[y1,</span> <span class="pre">x1,</span> <span class="pre">y2,</span> <span class="pre">x2]</span></code> format.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.centroid">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">centroid</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.centroid" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">centroid</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.centroid" title="Permalink to this definition">#</a></dt>
 <dd><p>Return instance centroid as an array of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates</p>
 <div class="admonition-notes admonition">
 <p class="admonition-title">Notes</p>
@@ -372,11 +472,11 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.fill_missing">
-<span class="sig-name descname"><span class="pre">fill_missing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L789-L819"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.fill_missing" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">fill_missing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L789-L819"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.fill_missing" title="Permalink to this definition">#</a></dt>
 <dd><p>Add points for skeleton nodes that are missing in the instance.</p>
 <p>This is useful when modifying the skeleton so the nodes appears in the GUI.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>max_x</strong> – If specified, make sure points are not added outside of valid range.</p></li>
 <li><p><strong>max_y</strong> – If specified, make sure points are not added outside of valid range.</p></li>
@@ -393,10 +493,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.from_numpy">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L949-L969"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.from_numpy" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L949-L969"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.from_numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance from a numpy array.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A numpy array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> that
 contains the points in (x, y) coordinates of each node. Missing nodes
@@ -406,7 +506,7 @@ <h1>sleap.instance</h1>
 <li><p><strong>track</strong> – Optional <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Track</span></code> object to associate with the instance.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> object.</p>
 </dd>
 </dl>
@@ -418,10 +518,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.from_pointsarray">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pointsarray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L923-L947"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.from_pointsarray" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pointsarray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L923-L947"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.from_pointsarray" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance from an array of points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A numpy array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> that
 contains the points in (x, y) coordinates of each node. Missing nodes
@@ -431,7 +531,7 @@ <h1>sleap.instance</h1>
 <li><p><strong>track</strong> – Optional <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Track</span></code> object to associate with the instance.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> object.</p>
 </dd>
 </dl>
@@ -439,10 +539,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.get_points_array">
-<span class="sig-name descname"><span class="pre">get_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invisible_as_nan</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.recarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L745-L787"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.get_points_array" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_points_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invisible_as_nan</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">recarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L745-L787"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.get_points_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the instance’s points in array form.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>copy</strong> – If True, the return a copy of the points array as an ndarray.
 If False, return a view of the underlying recarray.</p></li>
@@ -453,7 +553,7 @@ <h1>sleap.instance</h1>
 the x and y coordinates.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Either a recarray (if copy is False) or an ndarray (if copy True).</p>
 <p>The order of the rows corresponds to the ordering of the skeleton
 nodes. Any skeleton node not defined will have NaNs present.</p>
@@ -467,14 +567,14 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L669-L699"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L669-L699"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Whether two instances match by value.</p>
 <p>Checks the types, points, track, and frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code></a>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if match, False otherwise.</p>
 </dd>
 </dl>
@@ -482,7 +582,7 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.midpoint">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">midpoint</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.midpoint" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">midpoint</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.midpoint" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the center of the bounding box of the instance points.</p>
 </dd></dl>
 
@@ -494,23 +594,23 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.nodes">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="p"><span class="pre">...</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.nodes" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="p"><span class="pre">...</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Return nodes that have been labelled for this instance.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.nodes_points">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes_points</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">sleap.instance.Point</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.nodes_points" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes_points</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">Point</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.nodes_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of (node, point) tuples for all labeled points.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.numpy">
-<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L833-L843"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.numpy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L833-L843"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the instance node coordinates as a numpy array.</p>
 <p>Alias for <a class="reference internal" href="#sleap.instance.Instance.points_array" title="sleap.instance.Instance.points_array"><code class="xref py py-obj docutils literal notranslate"><span class="pre">points_array</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> of dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> containing the coordinates
 of the instance’s nodes. Missing/not visible nodes will be replaced with
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">NaN</span></code>.</p>
@@ -520,18 +620,18 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.points">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">sleap.instance.Point</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="p"><span class="pre">...</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.points" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">Point</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="p"><span class="pre">...</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.points" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a tuple of labelled points, in the order they were labelled.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.points_array">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points_array</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.points_array" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points_array</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.Instance.points_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Return array of x and y coordinates for visible points.</p>
 <p>Row in array corresponds to order of points in skeleton. Invisible points will
 be denoted by NaNs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A numpy array of of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> point coordinates.</p>
 </dd>
 </dl>
@@ -539,10 +639,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Instance.transform_points">
-<span class="sig-name descname"><span class="pre">transform_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">transformation_matrix</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L845-L864"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.transform_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">transformation_matrix</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L845-L864"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Instance.transform_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Apply affine transformation matrix to points in the instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>transformation_matrix</strong> – Affine transformation matrix as a numpy array of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(3,</span> <span class="pre">3)</span></code>.</p>
 </dd>
@@ -551,18 +651,88 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.Instance.video">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.video" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.Instance.video" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the video of the labeled frame this instance is associated with.</p>
 </dd></dl>
 
 </dd></dl>
 
+<dl class="py class">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">InstancesList</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labeled_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1307-L1425"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList" title="Permalink to this definition">#</a></dt>
+<dd><p>A list of <code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance`s</span> <span class="pre">associated</span> <span class="pre">with</span> <span class="pre">a</span> <span class="pre">`LabeledFrame</span></code>.</p>
+<p>This class should only be used for the <a class="reference internal" href="#sleap.instance.LabeledFrame.instances" title="sleap.instance.LabeledFrame.instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame.instances</span></code></a> attribute.</p>
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.append">
+<span class="sig-name descname"><span class="pre">append</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1348-L1361"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.append" title="Permalink to this definition">#</a></dt>
+<dd><p>Append an <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> or <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> to the list, setting the frame.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><p><strong>item</strong> – The <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> or <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> to append to the list.</p>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.clear">
+<span class="sig-name descname"><span class="pre">clear</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">None</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1414-L1418"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.clear" title="Permalink to this definition">#</a></dt>
+<dd><p>Remove all instances from list, setting instance.frame to None.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.copy">
+<span class="sig-name descname"><span class="pre">copy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">list</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1420-L1425"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.copy" title="Permalink to this definition">#</a></dt>
+<dd><p>Return a shallow copy of the list of instances as a list.</p>
+<p>Note: This will not return an <a class="reference internal" href="#sleap.instance.InstancesList" title="sleap.instance.InstancesList"><code class="xref py py-obj docutils literal notranslate"><span class="pre">InstancesList</span></code></a> object, but a normal list.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.extend">
+<span class="sig-name descname"><span class="pre">extend</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1363-L1374"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.extend" title="Permalink to this definition">#</a></dt>
+<dd><p>Extend the list with a list of <a href="#id1"><span class="problematic" id="id2">`</span></a>Instance`s or <a href="#id3"><span class="problematic" id="id4">`</span></a>PredictedInstance`s.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><p><strong>instances</strong> – A list of <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> or <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> objects to add to the
+list.</p>
+</dd>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>None</p>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.insert">
+<span class="sig-name descname"><span class="pre">insert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">None</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1385-L1387"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.insert" title="Permalink to this definition">#</a></dt>
+<dd><p>Insert object before index.</p>
+</dd></dl>
+
+<dl class="py property">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.labeled_frame">
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">labeled_frame</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></em><a class="headerlink" href="#sleap.instance.InstancesList.labeled_frame" title="Permalink to this definition">#</a></dt>
+<dd><p>Return the <a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a> associated with this list of instances.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.pop">
+<span class="sig-name descname"><span class="pre">pop</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1402-L1407"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.pop" title="Permalink to this definition">#</a></dt>
+<dd><p>Remove and return instance at index, setting instance.frame to None.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.instance.InstancesList.remove">
+<span class="sig-name descname"><span class="pre">remove</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">None</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1409-L1412"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.InstancesList.remove" title="Permalink to this definition">#</a></dt>
+<dd><p>Remove instance from list, setting instance.frame to None.</p>
+</dd></dl>
+
+</dd></dl>
+
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">LabeledFrame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1282-L1808"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">LabeledFrame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.InstancesList" title="sleap.instance.InstancesList"><span class="pre">InstancesList</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1429-L1941"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame" title="Permalink to this definition">#</a></dt>
 <dd><p>Holds labeled data for a single frame of a video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> associated with this frame.</p></li>
 <li><p><strong>frame_idx</strong> – The index of frame in video.</p></li>
@@ -572,7 +742,7 @@ <h1>sleap.instance</h1>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.complex_frame_merge">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_frame_merge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1666-L1760"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.complex_frame_merge" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_frame_merge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1801-L1893"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.complex_frame_merge" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge two frames, return conflicts if any.</p>
 <p>A conflict occurs when
 * each frame has Instances which don’t perfectly match those</p>
@@ -584,13 +754,13 @@ <h1>sleap.instance</h1>
 those in the other frame.</p></li>
 </ul>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>base_frame</strong> – The <a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a> into which we want to merge.</p></li>
 <li><p><strong>new_frame</strong> – The <a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a> from which we want to merge.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><p>list of instances that were merged</p></li>
 <li><p>list of conflicting instances from base</p></li>
@@ -598,7 +768,7 @@ <h1>sleap.instance</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>tuple of three items</p>
 </dd>
 </dl>
@@ -606,23 +776,23 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.complex_merge_between">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_merge_between</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1613-L1664"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.complex_merge_between" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_merge_between</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1748-L1799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.complex_merge_between" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge data from new frames into a <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> object.</p>
 <p>Everything that can be merged cleanly is merged, any conflicts
 are returned.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>base_labels</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> into which we are merging.</p></li>
 <li><p><strong>new_frames</strong> – The list of <a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a> objects from
 which we are merging.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><dl class="simple">
 <dt>Dictionary, keys are <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code>, values are</dt><dd><p>dictionary in which keys are frame index (int)
-and value is list of :class:<a href="#id1"><span class="problematic" id="id2">`</span></a>Instance`s</p>
+and value is list of :class:<a href="#id5"><span class="problematic" id="id6">`</span></a>Instance`s</p>
 </dd>
 </dl>
 </li>
@@ -631,7 +801,7 @@ <h1>sleap.instance</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>tuple of three items</p>
 </dd>
 </dl>
@@ -639,18 +809,18 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.find">
-<span class="sig-name descname"><span class="pre">find</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1369-L1388"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.find" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1502-L1521"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.find" title="Permalink to this definition">#</a></dt>
 <dd><p>Retrieve instances (if any) matching specifications.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>track</strong> – The <a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><code class="xref py py-class docutils literal notranslate"><span class="pre">Track</span></code></a> to match. Note that None will only
-match instances where <a href="#id3"><span class="problematic" id="id4">:attribute:`Instance.track`</span></a> is
+match instances where <a href="#id7"><span class="problematic" id="id8">:attribute:`Instance.track`</span></a> is
 None. If track is -1, then we’ll match any track.</p></li>
 <li><p><strong>user</strong> – Whether to only match user (non-predicted) instances.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of instances.</p>
 </dd>
 </dl>
@@ -676,29 +846,29 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.image">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">image</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.image" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">image</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.image" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the image for this frame of shape (height, width, channels).</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.index">
-<span class="sig-name descname"><span class="pre">index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1316-L1318"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.index" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1460-L1462"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.index" title="Permalink to this definition">#</a></dt>
 <dd><p>Return index of given <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code></a>.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.insert">
-<span class="sig-name descname"><span class="pre">insert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1338-L1352"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.insert" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">insert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1477-L1488"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.insert" title="Permalink to this definition">#</a></dt>
 <dd><p>Add instance to frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>index</strong> – The index in list of frame instances where we should
 insert the new instance.</p></li>
 <li><p><strong>value</strong> – The instance to associate with frame.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None.</p>
 </dd>
 </dl>
@@ -706,18 +876,18 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of all instances associated with this frame.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.instances_to_show">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">instances_to_show</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.instances_to_show" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">instances_to_show</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.instances_to_show" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of instances to show in GUI for this frame.</p>
 <p>This list will not include any predicted instances for which
 there’s a corresponding regular instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List of instances to show in GUI.</p>
 </dd>
 </dl>
@@ -725,10 +895,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.merge_frames">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">merge_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labeled_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">remove_redundant</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1565-L1611"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.merge_frames" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">merge_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labeled_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">remove_redundant</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1700-L1746"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.merge_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Return merged LabeledFrames for same video and frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labeled_frames</strong> – List of <a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a> objects to merge.</p></li>
 <li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which to merge.
@@ -738,8 +908,8 @@ <h1>sleap.instance</h1>
 where there’s a perfect match.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p>The merged list of :class:<a href="#id5"><span class="problematic" id="id6">`</span></a>LabeledFrame`s.</p>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>The merged list of :class:<a href="#id9"><span class="problematic" id="id10">`</span></a>LabeledFrame`s.</p>
 </dd>
 </dl>
 </dd></dl>
@@ -764,16 +934,16 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.numpy">
-<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1767-L1772"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.numpy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1900-L1905"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the instances as an array of shape (instances, nodes, 2).</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.plot">
-<span class="sig-name descname"><span class="pre">plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1774-L1788"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.plot" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">plot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1907-L1921"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.plot" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot the frame with all instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – If False, only the instances will be plotted without loading the
 original image.</p></li>
@@ -790,10 +960,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.plot_predicted">
-<span class="sig-name descname"><span class="pre">plot_predicted</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1790-L1808"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.plot_predicted" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">plot_predicted</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1923-L1941"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.plot_predicted" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot the frame with all predicted instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – If False, only the instances will be plotted without loading the
 original image.</p></li>
@@ -810,37 +980,37 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.predicted_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">predicted_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.predicted_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">predicted_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.predicted_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of predicted instances associated with frame.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.remove_empty_instances">
-<span class="sig-name descname"><span class="pre">remove_empty_instances</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1500-L1502"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.remove_empty_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_empty_instances</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1635-L1637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.remove_empty_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove instances with no visible nodes from the labeled frame.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.remove_untracked">
-<span class="sig-name descname"><span class="pre">remove_untracked</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1445-L1447"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.remove_untracked" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_untracked</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1580-L1582"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.LabeledFrame.remove_untracked" title="Permalink to this definition">#</a></dt>
 <dd><p>Removes any instances without a track assignment.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.tracked_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tracked_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.tracked_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tracked_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.tracked_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of predicted instances with tracks associated with frame.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.training_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">training_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.training_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">training_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.training_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of user instances with points for training.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.unused_predictions">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unused_predictions</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.unused_predictions" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unused_predictions</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.unused_predictions" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of “unused” <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-class docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> objects in frame.</p>
 <p>This is all the <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-class docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> objects which do not have
 a corresponding <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code></a> in the same track in frame.</p>
@@ -848,7 +1018,7 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.LabeledFrame.user_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.user_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.instance.LabeledFrame.user_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of user instances associated with this frame.</p>
 </dd></dl>
 
@@ -856,10 +1026,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.Point">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visible</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">complete</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L37-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visible</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">complete</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L37-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point" title="Permalink to this definition">#</a></dt>
 <dd><p>A labelled point and any metadata associated with it.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – The horizontal pixel location of point within image frame.</p></li>
 <li><p><strong>y</strong> – The vertical pixel location of point within image frame.</p></li>
@@ -875,10 +1045,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Point.isnan">
-<span class="sig-name descname"><span class="pre">isnan</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L79-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point.isnan" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">isnan</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L79-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point.isnan" title="Permalink to this definition">#</a></dt>
 <dd><p>Are either of the coordinates a NaN value.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>True if x or y is NaN, False otherwise.</p>
 </dd>
 </dl>
@@ -886,7 +1056,7 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Point.numpy">
-<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L88-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point.numpy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L88-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Point.numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the point as a numpy array.</p>
 </dd></dl>
 
@@ -894,21 +1064,21 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.PointArray">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PointArray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">strides</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">formats</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">titles</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">byteorder</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">aligned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'C'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L166-L279"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PointArray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">strides</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">formats</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">titles</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">byteorder</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">aligned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'C'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L166-L279"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray" title="Permalink to this definition">#</a></dt>
 <dd><p>PointArray is a sub-class of numpy recarray which stores
 Point objects as records.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PointArray.from_array">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L257-L279"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray.from_array" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L257-L279"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray.from_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts a <a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><code class="xref py py-class docutils literal notranslate"><span class="pre">PointArray</span></code></a> (or child) to a new instance.</p>
 <p>This will convert an object to the same type as itself,
 so a <a class="reference internal" href="#sleap.instance.PredictedPointArray" title="sleap.instance.PredictedPointArray"><code class="xref py py-class docutils literal notranslate"><span class="pre">PredictedPointArray</span></code></a> will result in the same.</p>
 <p>Uses the default attribute values for new array.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>a</strong> – The array to convert.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><code class="xref py py-class docutils literal notranslate"><span class="pre">PointArray</span></code></a> or <a class="reference internal" href="#sleap.instance.PredictedPointArray" title="sleap.instance.PredictedPointArray"><code class="xref py py-class docutils literal notranslate"><span class="pre">PredictedPointArray</span></code></a> with
 the same points as a.</p>
 </dd>
@@ -917,16 +1087,16 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PointArray.make_default">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_default</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L220-L237"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray.make_default" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_default</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L220-L237"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PointArray.make_default" title="Permalink to this definition">#</a></dt>
 <dd><p>Construct a point array where points are all set to default.</p>
 <p>The constructed <a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><code class="xref py py-class docutils literal notranslate"><span class="pre">PointArray</span></code></a> will have specified size
 and each value in the array is assigned the default values for
 a <code class="xref py py-class docutils literal notranslate"><span class="pre">Point`</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>size</strong> – The number of points to allocate.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A point array with all elements set to Point()</p>
 </dd>
 </dl>
@@ -936,10 +1106,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedInstance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_predicted</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L995-L1177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedInstance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_predicted</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L995-L1198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance" title="Permalink to this definition">#</a></dt>
 <dd><p>A predicted instance is an output of the inference procedure.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>score</strong> – The instance-level grouping prediction score.</p></li>
 <li><p><strong>tracking_score</strong> – The instance-level track matching score.</p></li>
@@ -948,10 +1118,10 @@ <h1>sleap.instance</h1>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.from_arrays">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_arrays</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1075-L1117"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_arrays" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_arrays</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1079-L1124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_arrays" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a predicted instance from data arrays.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A numpy array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> that
 contains the points in <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates of each node. Missing nodes
@@ -963,9 +1133,10 @@ <h1>sleap.instance</h1>
 <li><p><strong>skeleton</strong> – A sleap.Skeleton instance with n_nodes nodes to associate with the
 predicted instance.</p></li>
 <li><p><strong>track</strong> – Optional <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Track</span></code> to associate with the instance.</p></li>
+<li><p><strong>tracking_score</strong> – Optional float representing the track matching score.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a>.</p>
 </dd>
 </dl>
@@ -973,19 +1144,20 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.from_instance">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1051-L1073"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_instance" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1051-L1077"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> from an <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a>.</p>
 <p>The fields are copied in a shallow manner with the exception of points. For each
 point in the instance a <a class="reference internal" href="#sleap.instance.PredictedPoint" title="sleap.instance.PredictedPoint"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedPoint</span></code></a> is created with score set to default
 value.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instance</strong> – The <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> object to shallow copy data from.</p></li>
 <li><p><strong>score</strong> – The score for this instance.</p></li>
+<li><p><strong>tracking_score</strong> – The tracking score for this instance.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> for the given <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a>.</p>
 </dd>
 </dl>
@@ -993,10 +1165,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.from_numpy">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1149-L1177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_numpy" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1163-L1198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a predicted instance from data arrays.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A numpy array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> that
 contains the points in <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates of each node. Missing nodes
@@ -1008,9 +1180,10 @@ <h1>sleap.instance</h1>
 <li><p><strong>skeleton</strong> – A sleap.Skeleton instance with n_nodes nodes to associate with the
 predicted instance.</p></li>
 <li><p><strong>track</strong> – Optional <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Track</span></code> to associate with the instance.</p></li>
+<li><p><strong>tracking_score</strong> – Optional float representing the track matching score.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a>.</p>
 </dd>
 </dl>
@@ -1018,10 +1191,10 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.from_pointsarray">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pointsarray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1119-L1147"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_pointsarray" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pointsarray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">point_confidences</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracking_score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1126-L1161"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedInstance.from_pointsarray" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a predicted instance from data arrays.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A numpy array of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> that
 contains the points in <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates of each node. Missing nodes
@@ -1033,9 +1206,10 @@ <h1>sleap.instance</h1>
 <li><p><strong>skeleton</strong> – A sleap.Skeleton instance with n_nodes nodes to associate with the
 predicted instance.</p></li>
 <li><p><strong>track</strong> – Optional <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Track</span></code> to associate with the instance.</p></li>
+<li><p><strong>tracking_score</strong> – Optional float representing the track matching score.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a>.</p>
 </dd>
 </dl>
@@ -1043,7 +1217,7 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.points_and_scores_array">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points_and_scores_array</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.PredictedInstance.points_and_scores_array" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">points_and_scores_array</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.PredictedInstance.points_and_scores_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the instance points and scores as an array.</p>
 <p>This will be a <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">3)</span></code> array of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y,</span> <span class="pre">score)</span></code> for each predicted point.</p>
 <p>Rows in the array correspond to the order of points in skeleton. Invisible
@@ -1052,7 +1226,7 @@ <h1>sleap.instance</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.instance.PredictedInstance.scores">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">scores</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.instance.PredictedInstance.scores" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">scores</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.instance.PredictedInstance.scores" title="Permalink to this definition">#</a></dt>
 <dd><p>Return point scores for each predicted node.</p>
 </dd></dl>
 
@@ -1060,11 +1234,11 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.PredictedPoint">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedPoint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visible</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">complete</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L98-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPoint" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedPoint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visible</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">complete</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L98-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPoint" title="Permalink to this definition">#</a></dt>
 <dd><p>A predicted point is an output of the inference procedure.</p>
 <p>It has all the properties of a labeled point, plus a score.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – The horizontal pixel location of point within image frame.</p></li>
 <li><p><strong>y</strong> – The vertical pixel location of point within image frame.</p></li>
@@ -1081,16 +1255,16 @@ <h1>sleap.instance</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedPoint.from_point">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">sleap.instance.Point</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedPoint" title="sleap.instance.PredictedPoint"><span class="pre">sleap.instance.PredictedPoint</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L144-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPoint.from_point" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Point" title="sleap.instance.Point"><span class="pre">Point</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PredictedPoint" title="sleap.instance.PredictedPoint"><span class="pre">PredictedPoint</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L144-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPoint.from_point" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a PredictedPoint from a Point</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>point</strong> – The point to copy all data from.</p></li>
 <li><p><strong>score</strong> – The score for this predicted point.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A scored point based on the point passed in.</p>
 </dd>
 </dl>
@@ -1100,18 +1274,18 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.PredictedPointArray">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedPointArray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">strides</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">formats</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">titles</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">byteorder</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">aligned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'C'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L282-L306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPointArray" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">PredictedPointArray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">strides</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">formats</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">titles</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">byteorder</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">aligned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'C'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L282-L306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPointArray" title="Permalink to this definition">#</a></dt>
 <dd><p>PredictedPointArray is analogous to PointArray except for predicted
 points.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.PredictedPointArray.to_array">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">to_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PredictedPointArray" title="sleap.instance.PredictedPointArray"><span class="pre">sleap.instance.PredictedPointArray</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">sleap.instance.PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L290-L306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPointArray.to_array" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">to_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.PredictedPointArray" title="sleap.instance.PredictedPointArray"><span class="pre">PredictedPointArray</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.instance.PointArray" title="sleap.instance.PointArray"><span class="pre">PointArray</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L290-L306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.PredictedPointArray.to_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert a PredictedPointArray to a normal PointArray.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>a</strong> – The array to convert.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The converted array.</p>
 </dd>
 </dl>
@@ -1121,12 +1295,12 @@ <h1>sleap.instance</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.instance.Track">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">spawned_on</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L310-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Track" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">Track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">spawned_on</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L310-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Track" title="Permalink to this definition">#</a></dt>
 <dd><p>A track object is associated with a set of animal/object instances
 across multiple frames of video. This allows tracking of unique
 entities in the video over time and space.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>spawned_on</strong> – The video frame that this track was spawned on.</p></li>
 <li><p><strong>name</strong> – A name given to this track for identifying purposes.</p></li>
@@ -1135,13 +1309,13 @@ <h1>sleap.instance</h1>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.instance.Track.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L324-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Track.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L324-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.Track.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if two tracks match by value.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other track to check</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if they match, False otherwise.</p>
 </dd>
 </dl>
@@ -1151,12 +1325,12 @@ <h1>sleap.instance</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.instance.make_instance_cattr">
-<span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">make_instance_cattr</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">cattr.converters.Converter</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/instance.py#L1180-L1278"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.make_instance_cattr" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.instance.</span></span><span class="sig-name descname"><span class="pre">make_instance_cattr</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Converter</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/instance.py#L1201-L1304"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.instance.make_instance_cattr" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a cattr converter for Lists of Instances/PredictedInstances.</p>
 <p>This is required because cattrs doesn’t automatically detect the class when the
 attributes of one class are a subset of another.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A cattr converter with hooks registered for structuring and unstructuring
 <a class="reference internal" href="#sleap.instance.Instance" title="sleap.instance.Instance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Instance</span></code></a> and <a class="reference internal" href="#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a> objects.</p>
 </dd>
@@ -1187,6 +1361,112 @@ <h1>sleap.instance</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance"><code class="docutils literal notranslate"><span class="pre">Instance</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.bounding_box"><code class="docutils literal notranslate"><span class="pre">Instance.bounding_box</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.centroid"><code class="docutils literal notranslate"><span class="pre">Instance.centroid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.fill_missing"><code class="docutils literal notranslate"><span class="pre">Instance.fill_missing()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.frame_idx"><code class="docutils literal notranslate"><span class="pre">Instance.frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.from_numpy"><code class="docutils literal notranslate"><span class="pre">Instance.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.from_pointsarray"><code class="docutils literal notranslate"><span class="pre">Instance.from_pointsarray()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.get_points_array"><code class="docutils literal notranslate"><span class="pre">Instance.get_points_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.matches"><code class="docutils literal notranslate"><span class="pre">Instance.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.midpoint"><code class="docutils literal notranslate"><span class="pre">Instance.midpoint</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.n_visible_points"><code class="docutils literal notranslate"><span class="pre">Instance.n_visible_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.nodes"><code class="docutils literal notranslate"><span class="pre">Instance.nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.nodes_points"><code class="docutils literal notranslate"><span class="pre">Instance.nodes_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.numpy"><code class="docutils literal notranslate"><span class="pre">Instance.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.points"><code class="docutils literal notranslate"><span class="pre">Instance.points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.points_array"><code class="docutils literal notranslate"><span class="pre">Instance.points_array</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.transform_points"><code class="docutils literal notranslate"><span class="pre">Instance.transform_points()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Instance.video"><code class="docutils literal notranslate"><span class="pre">Instance.video</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList"><code class="docutils literal notranslate"><span class="pre">InstancesList</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.append"><code class="docutils literal notranslate"><span class="pre">InstancesList.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.clear"><code class="docutils literal notranslate"><span class="pre">InstancesList.clear()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.copy"><code class="docutils literal notranslate"><span class="pre">InstancesList.copy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.extend"><code class="docutils literal notranslate"><span class="pre">InstancesList.extend()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.insert"><code class="docutils literal notranslate"><span class="pre">InstancesList.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.labeled_frame"><code class="docutils literal notranslate"><span class="pre">InstancesList.labeled_frame</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.pop"><code class="docutils literal notranslate"><span class="pre">InstancesList.pop()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.InstancesList.remove"><code class="docutils literal notranslate"><span class="pre">InstancesList.remove()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame"><code class="docutils literal notranslate"><span class="pre">LabeledFrame</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.complex_frame_merge"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.complex_frame_merge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.complex_merge_between"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.complex_merge_between()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.find"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.find()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.has_user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.has_user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.image"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.index"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.insert"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.instances_to_show"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.instances_to_show</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.merge_frames"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.merge_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.n_user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.n_user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.numpy"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.plot"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.plot()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.plot_predicted"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.plot_predicted()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.predicted_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.remove_empty_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.remove_empty_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.remove_untracked"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.remove_untracked()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.tracked_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.tracked_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.training_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.training_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.unused_predictions"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.unused_predictions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.LabeledFrame.user_instances"><code class="docutils literal notranslate"><span class="pre">LabeledFrame.user_instances</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point"><code class="docutils literal notranslate"><span class="pre">Point</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.dtype"><code class="docutils literal notranslate"><span class="pre">Point.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.isnan"><code class="docutils literal notranslate"><span class="pre">Point.isnan()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Point.numpy"><code class="docutils literal notranslate"><span class="pre">Point.numpy()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray"><code class="docutils literal notranslate"><span class="pre">PointArray</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray.from_array"><code class="docutils literal notranslate"><span class="pre">PointArray.from_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PointArray.make_default"><code class="docutils literal notranslate"><span class="pre">PointArray.make_default()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance"><code class="docutils literal notranslate"><span class="pre">PredictedInstance</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_arrays"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_arrays()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_instance"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_numpy"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.from_pointsarray"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.from_pointsarray()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.points_and_scores_array"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.points_and_scores_array</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedInstance.scores"><code class="docutils literal notranslate"><span class="pre">PredictedInstance.scores</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint"><code class="docutils literal notranslate"><span class="pre">PredictedPoint</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint.dtype"><code class="docutils literal notranslate"><span class="pre">PredictedPoint.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPoint.from_point"><code class="docutils literal notranslate"><span class="pre">PredictedPoint.from_point()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPointArray"><code class="docutils literal notranslate"><span class="pre">PredictedPointArray</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.PredictedPointArray.to_array"><code class="docutils literal notranslate"><span class="pre">PredictedPointArray.to_array()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Track"><code class="docutils literal notranslate"><span class="pre">Track</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.Track.matches"><code class="docutils literal notranslate"><span class="pre">Track.matches()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.instance.make_instance_cattr"><code class="docutils literal notranslate"><span class="pre">make_instance_cattr()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1205,7 +1485,7 @@ <h1>sleap.instance</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.convert.html b/api/sleap.io.convert.html
index f36cd14b5..70511b8e7 100644
--- a/api/sleap.io.convert.html
+++ b/api/sleap.io.convert.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.convert &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.convert &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.convert';</script>
     <link rel="canonical" href="/develop/api/sleap.io.convert.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,14 @@ <h1>sleap.io.convert</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.convert.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +326,7 @@ <h1>sleap.io.convert</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.convert">
-<span id="sleap-io-convert"></span><h1>sleap.io.convert<a class="headerlink" href="#module-sleap.io.convert" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-convert"></span><h1>sleap.io.convert<a class="headerlink" href="#module-sleap.io.convert" title="Permalink to this heading">#</a></h1>
 <p>Command line utility for converting between various dataset formats.</p>
 <p>Reads:
 * SLEAP dataset in .slp, .h5, .json, or .json.zip file
@@ -354,10 +362,10 @@ <h1>sleap.io.convert</h1>
 first transpose the datasets so they matche the shapes described above.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.convert.main">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.convert.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/convert.py#L100-L184"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.convert.main" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.convert.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/convert.py#L101-L204"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.convert.main" title="Permalink to this definition">#</a></dt>
 <dd><p>Entrypoint for <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap-convert</span></code> CLI for converting .slp to different formats.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>args</strong> – A list of arguments to be passed into sleap-convert.</p>
 </dd>
 </dl>
@@ -387,6 +395,20 @@ <h1>sleap.io.convert</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.convert.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -405,7 +427,7 @@ <h1>sleap.io.convert</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.dataset.html b/api/sleap.io.dataset.html
index 0a655e46c..da964b458 100644
--- a/api/sleap.io.dataset.html
+++ b/api/sleap.io.dataset.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.dataset &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.dataset &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.dataset';</script>
     <link rel="canonical" href="/develop/api/sleap.io.dataset.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,120 @@ <h1>sleap.io.dataset</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels"><code class="docutils literal notranslate"><span class="pre">Labels</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.labeled_frames"><code class="docutils literal notranslate"><span class="pre">Labels.labeled_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.videos"><code class="docutils literal notranslate"><span class="pre">Labels.videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.skeletons"><code class="docutils literal notranslate"><span class="pre">Labels.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.tracks"><code class="docutils literal notranslate"><span class="pre">Labels.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.suggestions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.negative_anchors"><code class="docutils literal notranslate"><span class="pre">Labels.negative_anchors</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.provenance"><code class="docutils literal notranslate"><span class="pre">Labels.provenance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_instance"><code class="docutils literal notranslate"><span class="pre">Labels.add_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.add_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_track"><code class="docutils literal notranslate"><span class="pre">Labels.add_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_video"><code class="docutils literal notranslate"><span class="pre">Labels.add_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.all_instances"><code class="docutils literal notranslate"><span class="pre">Labels.all_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.append"><code class="docutils literal notranslate"><span class="pre">Labels.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.append_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.append_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.clear_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.clear_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.complex_merge_between"><code class="docutils literal notranslate"><span class="pre">Labels.complex_merge_between()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.copy"><code class="docutils literal notranslate"><span class="pre">Labels.copy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.delete_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.delete_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.describe"><code class="docutils literal notranslate"><span class="pre">Labels.describe()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export"><code class="docutils literal notranslate"><span class="pre">Labels.export()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export_csv"><code class="docutils literal notranslate"><span class="pre">Labels.export_csv()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export_nwb"><code class="docutils literal notranslate"><span class="pre">Labels.export_nwb()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.extend_from"><code class="docutils literal notranslate"><span class="pre">Labels.extend_from()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.extract"><code class="docutils literal notranslate"><span class="pre">Labels.extract()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find"><code class="docutils literal notranslate"><span class="pre">Labels.find()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_first"><code class="docutils literal notranslate"><span class="pre">Labels.find_first()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_last"><code class="docutils literal notranslate"><span class="pre">Labels.find_last()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.find_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_track_occupancy"><code class="docutils literal notranslate"><span class="pre">Labels.find_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.finish_complex_merge"><code class="docutils literal notranslate"><span class="pre">Labels.finish_complex_merge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.frames"><code class="docutils literal notranslate"><span class="pre">Labels.frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get"><code class="docutils literal notranslate"><span class="pre">Labels.get()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_next_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.get_next_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.get_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_track_count"><code class="docutils literal notranslate"><span class="pre">Labels.get_track_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_track_occupancy"><code class="docutils literal notranslate"><span class="pre">Labels.get_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_unlabeled_suggestion_inds"><code class="docutils literal notranslate"><span class="pre">Labels.get_unlabeled_suggestion_inds()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_video_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.get_video_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.has_frame"><code class="docutils literal notranslate"><span class="pre">Labels.has_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.has_missing_videos"><code class="docutils literal notranslate"><span class="pre">Labels.has_missing_videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.index"><code class="docutils literal notranslate"><span class="pre">Labels.index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.insert"><code class="docutils literal notranslate"><span class="pre">Labels.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.instance_count"><code class="docutils literal notranslate"><span class="pre">Labels.instance_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.instances"><code class="docutils literal notranslate"><span class="pre">Labels.instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.is_multi_instance"><code class="docutils literal notranslate"><span class="pre">Labels.is_multi_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.labels"><code class="docutils literal notranslate"><span class="pre">Labels.labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.load_file"><code class="docutils literal notranslate"><span class="pre">Labels.load_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.make_video_callback"><code class="docutils literal notranslate"><span class="pre">Labels.make_video_callback()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_container_dicts"><code class="docutils literal notranslate"><span class="pre">Labels.merge_container_dicts()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_matching_frames"><code class="docutils literal notranslate"><span class="pre">Labels.merge_matching_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_nodes"><code class="docutils literal notranslate"><span class="pre">Labels.merge_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.numpy"><code class="docutils literal notranslate"><span class="pre">Labels.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.predicted_instances"><code class="docutils literal notranslate"><span class="pre">Labels.predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove"><code class="docutils literal notranslate"><span class="pre">Labels.remove()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_all_tracks"><code class="docutils literal notranslate"><span class="pre">Labels.remove_all_tracks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_empty_frames"><code class="docutils literal notranslate"><span class="pre">Labels.remove_empty_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_empty_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_empty_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_frame"><code class="docutils literal notranslate"><span class="pre">Labels.remove_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_frames"><code class="docutils literal notranslate"><span class="pre">Labels.remove_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_instance"><code class="docutils literal notranslate"><span class="pre">Labels.remove_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_predictions"><code class="docutils literal notranslate"><span class="pre">Labels.remove_predictions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.remove_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_track"><code class="docutils literal notranslate"><span class="pre">Labels.remove_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_untracked_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_untracked_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_unused_tracks"><code class="docutils literal notranslate"><span class="pre">Labels.remove_unused_tracks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_user_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_user_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_video"><code class="docutils literal notranslate"><span class="pre">Labels.remove_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save"><code class="docutils literal notranslate"><span class="pre">Labels.save()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_file"><code class="docutils literal notranslate"><span class="pre">Labels.save_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_frame_data_hdf5"><code class="docutils literal notranslate"><span class="pre">Labels.save_frame_data_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_frame_data_imgstore"><code class="docutils literal notranslate"><span class="pre">Labels.save_frame_data_imgstore()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.set_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.set_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.skeleton"><code class="docutils literal notranslate"><span class="pre">Labels.skeleton</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.split"><code class="docutils literal notranslate"><span class="pre">Labels.split()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_dict"><code class="docutils literal notranslate"><span class="pre">Labels.to_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_json"><code class="docutils literal notranslate"><span class="pre">Labels.to_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_pipeline"><code class="docutils literal notranslate"><span class="pre">Labels.to_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.track_set_instance"><code class="docutils literal notranslate"><span class="pre">Labels.track_set_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.track_swap"><code class="docutils literal notranslate"><span class="pre">Labels.track_swap()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.unlabeled_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.unlabeled_suggestions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_instances"><code class="docutils literal notranslate"><span class="pre">Labels.user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_labeled_frame_inds"><code class="docutils literal notranslate"><span class="pre">Labels.user_labeled_frame_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_labeled_frames"><code class="docutils literal notranslate"><span class="pre">Labels.user_labeled_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.video"><code class="docutils literal notranslate"><span class="pre">Labels.video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.with_user_labels_only"><code class="docutils literal notranslate"><span class="pre">Labels.with_user_labels_only()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.add_instance"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.add_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.add_track"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.add_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.find_fancy_frame_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.find_frames"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.find_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_filtered_frame_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_frame_count"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_frame_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_track_occupancy"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_video_track_occupancy"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_video_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_frame"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_instance"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_video"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.track_swap"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.track_swap()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.update"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.update()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.update_counts_for_frame"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.update_counts_for_frame()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.find_path_using_paths"><code class="docutils literal notranslate"><span class="pre">find_path_using_paths()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.load_file"><code class="docutils literal notranslate"><span class="pre">load_file()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +432,7 @@ <h1>sleap.io.dataset</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.dataset">
-<span id="sleap-io-dataset"></span><h1>sleap.io.dataset<a class="headerlink" href="#module-sleap.io.dataset" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-dataset"></span><h1>sleap.io.dataset<a class="headerlink" href="#module-sleap.io.dataset" title="Permalink to this heading">#</a></h1>
 <p>A SLEAP dataset collects labeled video frames, together with required metadata.</p>
 <p>This contains labeled frame data (user annotations and/or predictions),
 together with all the other data that is saved for a SLEAP project
@@ -355,7 +469,7 @@ <h1>sleap.io.dataset</h1>
 default extension to use if none is provided in the filename.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">Labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labeled_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.suggestions.SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">negative_anchors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">provenance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L389-L2697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">Labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labeled_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">negative_anchors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">provenance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L389-L2710"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels" title="Permalink to this definition">#</a></dt>
 <dd><p>The <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code></a> class collects the data for a SLEAP project.</p>
 <p>This class is front-end for all interactions with loading, writing,
 and modifying these labels. The actual storage backend for the data
@@ -365,7 +479,7 @@ <h1>sleap.io.dataset</h1>
 <span class="sig-name descname"><span class="pre">labeled_frames</span></span><a class="headerlink" href="#sleap.io.dataset.Labels.labeled_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>A list of <code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame">sleap.instance.LabeledFrame</a>]</p>
 </dd>
 </dl>
@@ -379,7 +493,7 @@ <h1>sleap.io.dataset</h1>
 stored in <a class="reference internal" href="#sleap.io.dataset.Labels.videos" title="sleap.io.dataset.Labels.videos"><code class="xref py py-obj docutils literal notranslate"><span class="pre">videos</span></code></a> attribute, but some videos in
 this list may not have any associated labeled frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video">sleap.io.video.Video</a>]</p>
 </dd>
 </dl>
@@ -391,7 +505,7 @@ <h1>sleap.io.dataset</h1>
 <dd><p>A list of <code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code> objects (again, that may or may
 not be referenced by an <code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code> in labeled frame).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton">sleap.skeleton.Skeleton</a>]</p>
 </dd>
 </dl>
@@ -402,7 +516,7 @@ <h1>sleap.io.dataset</h1>
 <span class="sig-name descname"><span class="pre">tracks</span></span><a class="headerlink" href="#sleap.io.dataset.Labels.tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>A list of <code class="xref py py-class docutils literal notranslate"><span class="pre">Track</span></code> that instances can belong to.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track">sleap.instance.Track</a>]</p>
 </dd>
 </dl>
@@ -415,7 +529,7 @@ <h1>sleap.io.dataset</h1>
 videos in project. These can be suggested frames for user
 to label or suggested frames for user to review.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[sleap.gui.suggestions.SuggestionFrame]</p>
 </dd>
 </dl>
@@ -429,7 +543,7 @@ <h1>sleap.io.dataset</h1>
 Dictionary key is <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code>, value is list of
 (frame index, x, y) tuples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dict[<a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video">sleap.io.video.Video</a>, list]</p>
 </dd>
 </dl>
@@ -440,7 +554,7 @@ <h1>sleap.io.dataset</h1>
 <span class="sig-name descname"><span class="pre">provenance</span></span><a class="headerlink" href="#sleap.io.dataset.Labels.provenance" title="Permalink to this definition">#</a></dt>
 <dd><p>Dictionary that denotes the origin of the <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dict[str, Union[str, int, float, bool]]</p>
 </dd>
 </dl>
@@ -448,16 +562,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.add_instance">
-<span class="sig-name descname"><span class="pre">add_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1326-L1343"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1326-L1343"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Add instance to frame, updating track occupancy.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.add_suggestion">
-<span class="sig-name descname"><span class="pre">add_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1383-L1393"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_suggestion" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1383-L1393"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_suggestion" title="Permalink to this definition">#</a></dt>
 <dd><p>Add a suggested frame to the labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> instance of the suggestion.</p></li>
 <li><p><strong>frame_idx</strong> – Index of the frame of the suggestion.</p></li>
@@ -468,19 +582,19 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.add_track">
-<span class="sig-name descname"><span class="pre">add_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1229-L1232"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_track" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1229-L1232"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_track" title="Permalink to this definition">#</a></dt>
 <dd><p>Add track to labels, updating occupancy.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.add_video">
-<span class="sig-name descname"><span class="pre">add_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1550-L1562"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_video" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1550-L1562"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.add_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Add a video to the labels if it is not already in it.</p>
 <p>Video instances are added automatically when adding labeled frames,
 but this function allows for adding videos to the labels before any
 labeled frames are added.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>video</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> instance</p>
 </dd>
 </dl>
@@ -488,31 +602,31 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.all_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.all_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.all_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of all instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.append">
-<span class="sig-name descname"><span class="pre">append</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L875-L877"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.append" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">append</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L875-L877"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.append" title="Permalink to this definition">#</a></dt>
 <dd><p>Add labeled frame to list of labeled frames.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.append_suggestions">
-<span class="sig-name descname"><span class="pre">append_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.suggestions.SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1497-L1499"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.append_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">append_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1497-L1499"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.append_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Append the suggested frames.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.clear_suggestions">
-<span class="sig-name descname"><span class="pre">clear_suggestions</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1509-L1511"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.clear_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">clear_suggestions</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1509-L1511"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.clear_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete all suggestions.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.complex_merge_between">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_merge_between</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">unify</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tuple</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1766-L1827"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.complex_merge_between" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">complex_merge_between</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">unify</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tuple</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1766-L1827"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.complex_merge_between" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge frames and other data from one dataset into another.</p>
 <p>Anything that can be merged cleanly is merged into base_labels.</p>
 <p>Frames conflict just in case each labels object has a matching
@@ -527,7 +641,7 @@ <h1>sleap.io.dataset</h1>
 instances not in the other.</p></li>
 </ul>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>base_labels</strong> – the <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> that we’re merging into</p></li>
 <li><p><strong>new_labels</strong> – the <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> that we’re merging from</p></li>
@@ -535,7 +649,7 @@ <h1>sleap.io.dataset</h1>
 new_labels with <em>matching</em> objects from base</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><dl class="simple">
 <dt>Dictionary, keys are <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code>, values are</dt><dd><p>dictionary in which keys are frame index (int)
@@ -548,7 +662,7 @@ <h1>sleap.io.dataset</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>tuple of three items</p>
 </dd>
 </dl>
@@ -556,7 +670,7 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.copy">
-<span class="sig-name descname"><span class="pre">copy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L811-L818"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.copy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">copy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L811-L818"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.copy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a full deep copy of the labels.
 .. admonition:: Notes</p>
 <blockquote>
@@ -568,23 +682,23 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.delete_suggestions">
-<span class="sig-name descname"><span class="pre">delete_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1505-L1507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.delete_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">delete_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1505-L1507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.delete_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete suggestions for specified video.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.describe">
-<span class="sig-name descname"><span class="pre">describe</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1135-L1154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.describe" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">describe</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1135-L1154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.describe" title="Permalink to this definition">#</a></dt>
 <dd><p>Print basic statistics about the labels dataset.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.export">
-<span class="sig-name descname"><span class="pre">export</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2035-L2056"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.export" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2035-L2056"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.export" title="Permalink to this definition">#</a></dt>
 <dd><p>Export labels to analysis HDF5 format.</p>
 <p>This expects the labels to contain data for a single video (e.g., predictions).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>filename</strong> – Path to output HDF5 file.</p>
 </dd>
 </dl>
@@ -608,9 +722,24 @@ <h1>sleap.io.dataset</h1>
 </div>
 </dd></dl>
 
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.io.dataset.Labels.export_csv">
+<span class="sig-name descname"><span class="pre">export_csv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2058-L2069"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.export_csv" title="Permalink to this definition">#</a></dt>
+<dd><p>Export labels to CSV format.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><p><strong>filename</strong> – Output path for the CSV format file.</p>
+</dd>
+</dl>
+<div class="admonition-notes admonition">
+<p class="admonition-title">Notes</p>
+<p>This will write the contents of the labels out as a CSV file.</p>
+</div>
+</dd></dl>
+
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.export_nwb">
-<span class="sig-name descname"><span class="pre">export_nwb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">overwrite</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'Processed</span> <span class="pre">SLEAP</span> <span class="pre">pose</span> <span class="pre">data'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">identifier</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_start_time</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">datetime.datetime</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2058-L2126"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.export_nwb" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_nwb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">overwrite</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'Processed</span> <span class="pre">SLEAP</span> <span class="pre">pose</span> <span class="pre">data'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">identifier</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_start_time</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">datetime</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2071-L2139"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.export_nwb" title="Permalink to this definition">#</a></dt>
 <dd><p>Export all <code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code> objects in a <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object to an NWB file.</p>
 <blockquote>
 <div><p>Use <a class="reference internal" href="#sleap.io.dataset.Labels.numpy" title="sleap.io.dataset.Labels.numpy"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels.numpy</span></code></a> to create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.NWBFile</span></code> with a separate
@@ -643,7 +772,7 @@ <h1>sleap.io.dataset</h1>
 </div></blockquote>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Output path for the NWB format file.</p></li>
 <li><p><strong>labels</strong> – The <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object to covert to a NWB format file.</p></li>
@@ -661,7 +790,7 @@ <h1>sleap.io.dataset</h1>
 not be used.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.NWBFile</span></code> with a separate <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.ProcessingModule</span></code> for each
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> in the <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object.</p>
 </dd>
@@ -670,7 +799,7 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.extend_from">
-<span class="sig-name descname"><span class="pre">extend_from</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unify</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1593-L1634"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.extend_from" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">extend_from</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unify</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1593-L1634"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.extend_from" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge data from another <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object or <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> list.</p>
 <dl>
 <dt>Arg:</dt><dd><p>new_frames: the object from which to copy data
@@ -681,7 +810,7 @@ <h1>sleap.io.dataset</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool, True if we added frames, False otherwise</p>
 </dd>
 </dl>
@@ -689,20 +818,20 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.extract">
-<span class="sig-name descname"><span class="pre">extract</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L779-L809"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.extract" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">extract</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L779-L809"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.extract" title="Permalink to this definition">#</a></dt>
 <dd><p>Extract labeled frames from indices and return a new <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object.
 :param inds: Any valid indexing keys, e.g., a range, slice, list of label indices,</p>
 <blockquote>
 <div><p>numpy array, <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code>, etc. See <code class="xref py py-obj docutils literal notranslate"><span class="pre">__getitem__</span></code> for full list.</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>copy</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, create a new copy of all of the extracted labeled frames
 and associated labels. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code> (the default), a shallow copy with
 references to the original labeled frames and other objects will be
 returned.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A new <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> object with the specified labeled frames.
 This will preserve the other data structures even if they are not found in
 the extracted labels, including:</p>
@@ -722,10 +851,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.find">
-<span class="sig-name descname"><span class="pre">find</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_new</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L933-L961"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_new</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L933-L961"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find" title="Permalink to this definition">#</a></dt>
 <dd><p>Search for labeled frames given video and/or frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – A <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> that is associated with the project.</p></li>
 <li><p><strong>frame_idx</strong> – The frame index (or indices) which we want to
@@ -736,7 +865,7 @@ <h1>sleap.io.dataset</h1>
 <code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code> if none is found in project.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects that match the criteria.
 Empty if no matches found, unless return_new is True,
 in which case it contains a new <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> with
@@ -747,11 +876,11 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.find_first">
-<span class="sig-name descname"><span class="pre">find_first</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L981-L1010"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_first" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_first</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L981-L1010"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_first" title="Permalink to this definition">#</a></dt>
 <dd><p>Find the first occurrence of a matching labeled frame.</p>
 <p>Matches on frames for the given video and/or frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> instance that is associated with the
 labeled frames</p></li>
@@ -762,7 +891,7 @@ <h1>sleap.io.dataset</h1>
 True, use the labels data cache, else loop through all labels to search.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>First <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> that match the criteria
 or None if none were found.</p>
 </dd>
@@ -771,11 +900,11 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.find_last">
-<span class="sig-name descname"><span class="pre">find_last</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1012-L1034"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_last" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_last</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1012-L1034"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_last" title="Permalink to this definition">#</a></dt>
 <dd><p>Find the last occurrence of a matching labeled frame.</p>
 <p>Matches on frames for the given video and/or frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> instance that is associated with the
 labeled frames</p></li>
@@ -783,7 +912,7 @@ <h1>sleap.io.dataset</h1>
 the video</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Last <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> that match the criteria
 or None if none were found.</p>
 </dd>
@@ -792,16 +921,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.find_suggestion">
-<span class="sig-name descname"><span class="pre">find_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1436-L1447"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_suggestion" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1436-L1447"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_suggestion" title="Permalink to this definition">#</a></dt>
 <dd><p>Find SuggestionFrame by video and frame index.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.find_track_occupancy">
-<span class="sig-name descname"><span class="pre">find_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1345-L1381"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_track_occupancy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1345-L1381"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.find_track_occupancy" title="Permalink to this definition">#</a></dt>
 <dd><p>Get instances for a given video, track, and range of frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code></p></li>
 <li><p><strong>track</strong> – the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Track</span></code> or int (“pseudo-track” index to instance list)</p></li>
@@ -809,7 +938,7 @@ <h1>sleap.io.dataset</h1>
 If None, return all instances for given track.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of <code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code> objects.</p>
 </dd>
 </dl>
@@ -817,10 +946,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.finish_complex_merge">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">finish_complex_merge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">resolved_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1829-L1849"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.finish_complex_merge" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">finish_complex_merge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">resolved_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1829-L1849"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.finish_complex_merge" title="Permalink to this definition">#</a></dt>
 <dd><p>Finish conflicted merge from complex_merge_between.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>base_labels</strong> – the <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> that we’re merging into</p></li>
 <li><p><strong>resolved_frames</strong> – the list of frames to add into base_labels</p></li>
@@ -831,10 +960,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.frames">
-<span class="sig-name descname"><span class="pre">frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reverse</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L963-L979"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reverse</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L963-L979"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Return an iterator over all labeled frames in a video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – A <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> that is associated with the project.</p></li>
 <li><p><strong>from_frame_idx</strong> – The frame index from which we want to start.
@@ -842,7 +971,7 @@ <h1>sleap.io.dataset</h1>
 <li><p><strong>reverse</strong> – Whether to iterate over frames in reverse order.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Yields</dt>
+<dt class="field-even">Yields<span class="colon">:</span></dt>
 <dd class="field-even"><p><code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code></p>
 </dd>
 </dl>
@@ -850,12 +979,12 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get">
-<span class="sig-name descname"><span class="pre">get</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">slice</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">secondary_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">slice</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_errors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L676-L777"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">slice</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">secondary_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">slice</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">integer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">list</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">range</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_errors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L676-L777"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get" title="Permalink to this definition">#</a></dt>
 <dd><p>Return labeled frames matching key or return <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> if not found.</p>
 <p>This is a safe version of <code class="xref py py-obj docutils literal notranslate"><span class="pre">labels[...]</span></code> that will not raise an exception if the
 item is not found.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>key</strong> – Indexing argument to match against. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">key</span></code> is a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> or tuple of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(Video,</span> <span class="pre">frame_index)</span></code>, frames that match the criteria will be searched
@@ -871,10 +1000,10 @@ <h1>sleap.io.dataset</h1>
 of raising KeyError.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>KeyError</strong> – If the specified key could not be found.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list with the matching <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame`s,</span> <span class="pre">or</span> <span class="pre">a</span> <span class="pre">single</span> <span class="pre">`LabeledFrame</span></code> if a
 scalar key was provided, or <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> if not found.</p>
 </dd>
@@ -883,36 +1012,36 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_next_suggestion">
-<span class="sig-name descname"><span class="pre">get_next_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">seek_direction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1449-L1495"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_next_suggestion" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_next_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">seek_direction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1449-L1495"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_next_suggestion" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a (video, frame_idx) tuple seeking from given frame.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_suggestions">
-<span class="sig-name descname"><span class="pre">get_suggestions</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.suggestions.SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1432-L1434"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_suggestions</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1432-L1434"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Return all suggestions as a list of SuggestionFrame items.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_track_count">
-<span class="sig-name descname"><span class="pre">get_track_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1221-L1223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_track_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_track_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1221-L1223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_track_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the number of occupied tracks for a given video.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_track_occupancy">
-<span class="sig-name descname"><span class="pre">get_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1225-L1227"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_track_occupancy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1225-L1227"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_track_occupancy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return track occupancy list for given video.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_unlabeled_suggestion_inds">
-<span class="sig-name descname"><span class="pre">get_unlabeled_suggestion_inds</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1523-L1548"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_unlabeled_suggestion_inds" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_unlabeled_suggestion_inds</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1523-L1548"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_unlabeled_suggestion_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>Find labeled frames for unlabeled suggestions and return their indices.</p>
 <p>This is useful for generating a list of example indices for inference on
 unlabeled suggestions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p><p>List of indices of the labeled frames that correspond to the suggestions
 that do not have user instances.</p>
 <p>If a labeled frame corresponding to a suggestion does not exist, an empty
@@ -925,10 +1054,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.get_video_suggestions">
-<span class="sig-name descname"><span class="pre">get_video_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1407-L1430"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_video_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_video_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1407-L1430"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.get_video_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of suggested frame indices.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – Video to get suggestions for.</p></li>
 <li><p><strong>user_labeled</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), return frame indices for suggestions
@@ -936,7 +1065,7 @@ <h1>sleap.io.dataset</h1>
 labeled instances will be returned.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Indices of the suggested frames for for the specified video.</p>
 </dd>
 </dl>
@@ -944,10 +1073,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.has_frame">
-<span class="sig-name descname"><span class="pre">has_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1636-L1680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.has_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">has_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1636-L1680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.has_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if the labels contain a specified frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>lf</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> to search for. If not provided, the <a class="reference internal" href="#sleap.io.dataset.Labels.video" title="sleap.io.dataset.Labels.video"><code class="xref py py-obj docutils literal notranslate"><span class="pre">video</span></code></a> and
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">frame_idx</span></code> must not be <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p></li>
@@ -957,7 +1086,7 @@ <h1>sleap.io.dataset</h1>
 searching. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, check every frame without the cache.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">bool</span></code> indicating whether the specified <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> is contained in the
 labels.</p>
 <p>This will return <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> if there is a matching frame with the same video and
@@ -981,34 +1110,34 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.index">
-<span class="sig-name descname"><span class="pre">index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L590-L592"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.index" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L590-L592"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.index" title="Permalink to this definition">#</a></dt>
 <dd><p>Return index of labeled frame in list of labeled frames.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.insert">
-<span class="sig-name descname"><span class="pre">insert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L867-L873"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.insert" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">insert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">index</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L867-L873"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.insert" title="Permalink to this definition">#</a></dt>
 <dd><p>Insert labeled frame at given index.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.instance_count">
-<span class="sig-name descname"><span class="pre">instance_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1080-L1088"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.instance_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">instance_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1080-L1088"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.instance_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Return number of instances matching video/frame index.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.instances">
-<span class="sig-name descname"><span class="pre">instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1156-L1172"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1156-L1172"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Iterate over instances in the labels, optionally with filters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – Only iterate through instances in this video</p></li>
 <li><p><strong>skeleton</strong> – Only iterate through instances with this skeleton</p></li>
 </ul>
 </dd>
-<dt class="field-even">Yields</dt>
+<dt class="field-even">Yields<span class="colon">:</span></dt>
 <dd class="field-even"><p><em>Instance</em> – The next labeled instance</p>
 </dd>
 </dl>
@@ -1028,20 +1157,20 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.load_file">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1957-L1970"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.load_file" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1957-L1970"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.load_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Load file, detecting format from filename.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.make_video_callback">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_video_callback</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2587-L2697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.make_video_callback" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_video_callback</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2600-L2710"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.make_video_callback" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a callback for finding missing videos.</p>
 <p>The callback can be used while loading a saved project and
 allows the user to find videos which have been moved (or have
 paths from a different system).</p>
 <p>The callback function returns True to signal “abort”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>search_paths</strong> – If specified, this is a list of paths where
 we’ll automatically try to find the missing videos.</p></li>
@@ -1049,7 +1178,7 @@ <h1>sleap.io.dataset</h1>
 value. Used externally to determine if any filenames were updated.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The callback function.</p>
 </dd>
 </dl>
@@ -1057,16 +1186,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.merge_container_dicts">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">merge_container_dicts</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dict_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dict_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1851-L1859"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_container_dicts" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">merge_container_dicts</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dict_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dict_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1851-L1859"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_container_dicts" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge data from dict_b into dict_a.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.merge_matching_frames">
-<span class="sig-name descname"><span class="pre">merge_matching_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1861-L1873"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_matching_frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">merge_matching_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1861-L1873"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_matching_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects that are for the same video frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>video</strong> – combine for this video; if None, do all videos</p>
 </dd>
 </dl>
@@ -1074,10 +1203,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.merge_nodes">
-<span class="sig-name descname"><span class="pre">merge_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">merge_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2548-L2575"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_nodes" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">merge_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">base_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">merge_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2561-L2588"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.merge_nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Merge two nodes and update data accordingly.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>base_node</strong> – Name of skeleton node that will remain after merging.</p></li>
 <li><p><strong>merge_node</strong> – Name of skeleton node that will be merged into the base node.</p></li>
@@ -1098,10 +1227,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.numpy">
-<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">untracked</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_confidence</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2429-L2546"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.numpy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">untracked</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_confidence</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2442-L2559"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Construct a numpy array from instance points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – Video or video index to convert to numpy arrays. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> (the
 default), uses the first video.</p></li>
@@ -1115,7 +1244,7 @@ <h1>sleap.io.dataset</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, return the points and scores of nodes.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An array of tracks of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_frames,</span> <span class="pre">n_tracks,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code> if
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">return_confidence</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>. Otherwise returned shape is
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_frames,</span> <span class="pre">n_tracks,</span> <span class="pre">n_nodes,</span> <span class="pre">3)</span></code> if <code class="xref py py-obj docutils literal notranslate"><span class="pre">return_confidence</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>.</p>
@@ -1134,34 +1263,34 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.predicted_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">predicted_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.predicted_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">predicted_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.predicted_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of all predicted instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove">
-<span class="sig-name descname"><span class="pre">remove</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L883-L885"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L883-L885"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove given labeled frame.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_all_tracks">
-<span class="sig-name descname"><span class="pre">remove_all_tracks</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1241-L1245"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_all_tracks" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_all_tracks</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1241-L1245"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_all_tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove all tracks from labels, updating (but not removing) instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_empty_frames">
-<span class="sig-name descname"><span class="pre">remove_empty_frames</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L926-L931"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_empty_frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_empty_frames</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L926-L931"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_empty_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove frames with no instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_empty_instances">
-<span class="sig-name descname"><span class="pre">remove_empty_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keep_empty_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L909-L924"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_empty_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_empty_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keep_empty_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L909-L924"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_empty_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove instances with no visible points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>keep_empty_frames</strong> – If True (the default), frames with no remaining instances
 will not be removed.</p>
 </dd>
@@ -1175,10 +1304,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_frame">
-<span class="sig-name descname"><span class="pre">remove_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L887-L897"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L887-L897"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a given labeled frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>lf</strong> – Labeled frame instance to remove.</p></li>
 <li><p><strong>update_cache</strong> – If True, update the internal frame cache. If False, cache
@@ -1190,10 +1319,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_frames">
-<span class="sig-name descname"><span class="pre">remove_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lfs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L899-L907"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lfs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L899-L907"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a list of frames from the labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>lfs</strong> – A sequence of labeled frames to remove.</p>
 </dd>
 </dl>
@@ -1201,18 +1330,18 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_instance">
-<span class="sig-name descname"><span class="pre">remove_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">in_transaction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1318-L1324"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">in_transaction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1318-L1324"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove instance from frame, updating track occupancy.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_predictions">
-<span class="sig-name descname"><span class="pre">remove_predictions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1718-L1752"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_predictions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_predictions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1718-L1752"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_predictions" title="Permalink to this definition">#</a></dt>
 <dd><p>Clear predicted instances from the labels.</p>
 <p>Useful prior to merging operations to prevent overlapping instances from new
 predictions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>new_labels</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, only predicted instances in frames that also
 contain predictions in the new labels will be removed. If not provided
 (the default), all predicted instances will be removed.</p>
@@ -1230,10 +1359,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_suggestion">
-<span class="sig-name descname"><span class="pre">remove_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1395-L1405"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_suggestion" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_suggestion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1395-L1405"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_suggestion" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a suggestion from the list by video and frame index.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> instance of the suggestion.</p></li>
 <li><p><strong>frame_idx</strong> – Index of the frame of the suggestion.</p></li>
@@ -1244,16 +1373,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_track">
-<span class="sig-name descname"><span class="pre">remove_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1234-L1239"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_track" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1234-L1239"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_track" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a track from the labels, updating (but not removing) instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_untracked_instances">
-<span class="sig-name descname"><span class="pre">remove_untracked_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">remove_empty_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1754-L1764"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_untracked_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_untracked_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">remove_empty_frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1754-L1764"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_untracked_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove instances that do not have a track assignment.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>remove_empty_frames</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), removes frames that do not
 contain any instances after removing untracked ones.</p>
 </dd>
@@ -1262,18 +1391,18 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_unused_tracks">
-<span class="sig-name descname"><span class="pre">remove_unused_tracks</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1247-L1261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_unused_tracks" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_unused_tracks</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1247-L1261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_unused_tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove tracks that are not used by any instances.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_user_instances">
-<span class="sig-name descname"><span class="pre">remove_user_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1682-L1716"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_user_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_user_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1682-L1716"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_user_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Clear user instances from the labels.</p>
 <p>Useful prior to merging operations to prevent overlapping instances from new
 labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>new_labels</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, only user instances in frames that also contain
 user instances in the new labels will be removed. If not provided
 (the default), all user instances will be removed.</p>
@@ -1291,10 +1420,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.remove_video">
-<span class="sig-name descname"><span class="pre">remove_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1564-L1585"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_video" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1564-L1585"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.remove_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a video from the labels and all associated labeled frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>video</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> instance to be removed.</p>
 </dd>
 </dl>
@@ -1302,10 +1431,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.save">
-<span class="sig-name descname"><span class="pre">save</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">embed_all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">embed_suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1999-L2033"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">embed_all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">embed_suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1999-L2033"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save" title="Permalink to this definition">#</a></dt>
 <dd><p>Save the labels to a file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to save the labels to ending in <code class="xref py py-obj docutils literal notranslate"><span class="pre">slp</span></code>. If the filename does
 not end in <code class="xref py py-obj docutils literal notranslate"><span class="pre">slp</span></code>, the extension will be automatically appended.</p></li>
@@ -1330,10 +1459,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.save_file">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">save_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1972-L1997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_file" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">save_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1972-L1997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Save file, detecting format from filename.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – The dataset to save.</p></li>
 <li><p><strong>filename</strong> – Path where we’ll save it. We attempt to detect format
@@ -1343,7 +1472,7 @@ <h1>sleap.io.dataset</h1>
 format). Doesn’t need to have “.” before file extension.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If cannot detect valid filetype.</p>
 </dd>
 </dl>
@@ -1351,11 +1480,11 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.save_frame_data_hdf5">
-<span class="sig-name descname"><span class="pre">save_frame_data_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">sleap.io.video.HDF5Video</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2302-L2386"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_frame_data_hdf5" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save_frame_data_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">HDF5Video</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2315-L2399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_frame_data_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Write images for labeled frames from all videos to hdf5 file.</p>
 <p>Note that this will make an HDF5 video, not an HDF5 labels dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>output_path</strong> – Path to HDF5 file.</p></li>
 <li><p><strong>format</strong> – The image format to use for the data. Defaults to png.</p></li>
@@ -1375,7 +1504,7 @@ <h1>sleap.io.dataset</h1>
 deleted.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of <code class="xref py py-class docutils literal notranslate"><span class="pre">HDF5Video</span></code> objects with the stored frames.</p>
 </dd>
 </dl>
@@ -1383,12 +1512,12 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.save_frame_data_imgstore">
-<span class="sig-name descname"><span class="pre">save_frame_data_imgstore</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'./'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.ImgStoreVideo" title="sleap.io.video.ImgStoreVideo"><span class="pre">sleap.io.video.ImgStoreVideo</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2226-L2300"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_frame_data_imgstore" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save_frame_data_imgstore</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'./'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.ImgStoreVideo" title="sleap.io.video.ImgStoreVideo"><span class="pre">ImgStoreVideo</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2239-L2313"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.save_frame_data_imgstore" title="Permalink to this definition">#</a></dt>
 <dd><p>Write images for labeled frames from all videos to imgstore datasets.</p>
 <p>This only writes frames that have been labeled. Videos without
 any labeled frames will be included as empty imgstores.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>output_dir</strong> – Path to directory which will contain imgstores.</p></li>
 <li><p><strong>format</strong> – The image format to use for the data.
@@ -1408,7 +1537,7 @@ <h1>sleap.io.dataset</h1>
 deleted.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of <code class="xref py py-class docutils literal notranslate"><span class="pre">ImgStoreVideo</span></code> objects with the stored
 frames.</p>
 </dd>
@@ -1417,28 +1546,28 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.set_suggestions">
-<span class="sig-name descname"><span class="pre">set_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.suggestions.SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1501-L1503"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.set_suggestions" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">set_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">suggestions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">SuggestionFrame</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1501-L1503"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.set_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Set the suggested frames.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.skeleton">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">skeleton</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></em><a class="headerlink" href="#sleap.io.dataset.Labels.skeleton" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">skeleton</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></em><a class="headerlink" href="#sleap.io.dataset.Labels.skeleton" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the skeleton if there is only a single skeleton in the labels.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.split">
-<span class="sig-name descname"><span class="pre">split</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L820-L859"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.split" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">split</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L820-L859"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.split" title="Permalink to this definition">#</a></dt>
 <dd><p>Split labels randomly.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>n</strong> – Number or fraction of elements in the first split.</p></li>
 <li><p><strong>copy</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), return copies of the labels.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(labels_a,</span> <span class="pre">labels_b)</span></code> where both are <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instances
 subsampled from these labels.</p>
 </dd>
@@ -1460,16 +1589,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.to_dict">
-<span class="sig-name descname"><span class="pre">to_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skip_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1875-L1946"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_dict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skip_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1875-L1946"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Serialize all labels to dicts.</p>
 <p>Serializes the labels in the underling list of LabeledFrames to a dict
 structure. This function returns a nested dict structure composed entirely of
 primitive python types. It is used to create JSON and HDF5 serialized datasets.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>skip_labels</strong> – If True, skip labels serialization and just do the metadata.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><p>version - The version of the dict/json serialization format.</p></li>
 <li><p>skeletons - The skeletons associated with these underlying
@@ -1483,7 +1612,7 @@ <h1>sleap.io.dataset</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dict containing the followings top level keys</p>
 </dd>
 </dl>
@@ -1491,10 +1620,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.to_json">
-<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1948-L1955"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_json" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1948-L1955"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Serialize all labels in the underling list of LabeledFrame(s) to JSON.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The JSON representation of the string.</p>
 </dd>
 </dl>
@@ -1502,10 +1631,10 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.to_pipeline">
-<span class="sig-name descname"><span class="pre">to_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">sleap.pipelines.Pipeline</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2388-L2427"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labeled_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">sleap.pipelines.Pipeline</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2401-L2440"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.to_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a pipeline for reading the dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>batch_size</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the video frames will be batched into rank-4
 tensors. Otherwise, single rank-3 images will be returned.</p></li>
@@ -1517,7 +1646,7 @@ <h1>sleap.io.dataset</h1>
 labeled instances.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> that builds <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> for high
 throughput I/O during inference.</p>
 </dd>
@@ -1527,19 +1656,19 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.track_set_instance">
-<span class="sig-name descname"><span class="pre">track_set_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1263-L1275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.track_set_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">track_set_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1263-L1275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.track_set_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Set track on given instance, updating occupancy.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.track_swap">
-<span class="sig-name descname"><span class="pre">track_swap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tuple</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1277-L1316"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.track_swap" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">track_swap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tuple</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1277-L1316"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.track_swap" title="Permalink to this definition">#</a></dt>
 <dd><p>Swap track assignment for instances in two tracks.</p>
 <p>If you need to change the track to or from None, you’ll need
 to use <a class="reference internal" href="#sleap.io.dataset.Labels.track_set_instance" title="sleap.io.dataset.Labels.track_set_instance"><code class="xref py py-meth docutils literal notranslate"><span class="pre">track_set_instance()</span></code></a> for each specific
 instance you want to modify.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>video</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> for which we want to swap tracks.</p></li>
 <li><p><strong>new_track</strong> – A <code class="xref py py-class docutils literal notranslate"><span class="pre">Track</span></code> for which we want to swap
@@ -1555,13 +1684,13 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.unlabeled_suggestions">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unlabeled_suggestions</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.suggestions.SuggestionFrame</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.unlabeled_suggestions" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unlabeled_suggestions</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">SuggestionFrame</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.unlabeled_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Return suggestions without user labels.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.user_instances">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.user_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_instances</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.user_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of all user (non-predicted) instances.</p>
 </dd></dl>
 
@@ -1573,23 +1702,23 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.user_labeled_frames">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_labeled_frames</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.user_labeled_frames" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">user_labeled_frames</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.io.dataset.Labels.user_labeled_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Return all labeled frames with user (non-predicted) instances.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.video">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></em><a class="headerlink" href="#sleap.io.dataset.Labels.video" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></em><a class="headerlink" href="#sleap.io.dataset.Labels.video" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the video if there is only a single video in the labels.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.Labels.with_user_labels_only">
-<span class="sig-name descname"><span class="pre">with_user_labels_only</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">user_instances_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L1046-L1075"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.with_user_labels_only" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">with_user_labels_only</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">user_instances_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L1046-L1075"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.Labels.with_user_labels_only" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a new <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> containing only user labels.</p>
 <p>This is useful as a preprocessing step to train on only user-labeled data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>user_instances_only</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), predicted instances will be
 removed from frames that also have user instances.</p></li>
@@ -1601,7 +1730,7 @@ <h1>sleap.io.dataset</h1>
 instances will be removed from the original labels as well!</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> with only the specified subset of frames and instances.</p>
 </dd>
 </dl>
@@ -1611,89 +1740,89 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">LabelsDataCache</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L97-L385"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">LabelsDataCache</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L97-L385"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for maintaining cache of data in labels dataset.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.add_instance">
-<span class="sig-name descname"><span class="pre">add_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L254-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.add_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L254-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.add_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Add an instance to the labels.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.add_track">
-<span class="sig-name descname"><span class="pre">add_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L250-L252"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.add_track" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L250-L252"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.add_track" title="Permalink to this definition">#</a></dt>
 <dd><p>Add a track to the labels.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs">
-<span class="sig-name descname"><span class="pre">find_fancy_frame_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reverse</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L158-L180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_fancy_frame_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">from_frame_idx</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reverse</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L158-L180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of frame idxs, with optional start position/order.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.find_frames">
-<span class="sig-name descname"><span class="pre">find_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L134-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.find_frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L134-L156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.find_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of LabeledFrames matching video/frame_idx, or None.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs">
-<span class="sig-name descname"><span class="pre">get_filtered_frame_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Set</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L295-L331"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_filtered_frame_idxs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Set</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L295-L331"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs" title="Permalink to this definition">#</a></dt>
 <dd><p>Return list of (video_idx, frame_idx) tuples matching video/filter.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.get_frame_count">
-<span class="sig-name descname"><span class="pre">get_frame_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L279-L293"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_frame_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame_count</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L279-L293"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_frame_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Return (possibly cached) count of frames matching video/filter.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.get_track_occupancy">
-<span class="sig-name descname"><span class="pre">get_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">sleap.rangelist.RangeList</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L196-L200"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_track_occupancy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">RangeList</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L196-L200"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_track_occupancy" title="Permalink to this definition">#</a></dt>
 <dd><p>Access track occupancy cache that adds video/track as needed.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.get_video_track_occupancy">
-<span class="sig-name descname"><span class="pre">get_video_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">sleap.rangelist.RangeList</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L202-L207"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_video_track_occupancy" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_video_track_occupancy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RangeList</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L202-L207"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.get_video_track_occupancy" title="Permalink to this definition">#</a></dt>
 <dd><p>Return track occupancy information for specified video.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.remove_frame">
-<span class="sig-name descname"><span class="pre">remove_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L209-L216"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L209-L216"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove frame and update cache as needed.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.remove_instance">
-<span class="sig-name descname"><span class="pre">remove_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L266-L277"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L266-L277"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove an instance and update the cache as needed.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.remove_video">
-<span class="sig-name descname"><span class="pre">remove_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L218-L223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_video" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">remove_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L218-L223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.remove_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove video and update cache as needed.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.track_swap">
-<span class="sig-name descname"><span class="pre">track_swap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tuple</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L225-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.track_swap" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">track_swap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tuple</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L225-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.track_swap" title="Permalink to this definition">#</a></dt>
 <dd><p>Swap tracks and update cache as needed.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.update">
-<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L105-L132"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.update" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">new_frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L105-L132"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.update" title="Permalink to this definition">#</a></dt>
 <dd><p>Build (or rebuilds) various caches.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.dataset.LabelsDataCache.update_counts_for_frame">
-<span class="sig-name descname"><span class="pre">update_counts_for_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L333-L361"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.update_counts_for_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">update_counts_for_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L333-L361"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.LabelsDataCache.update_counts_for_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Updated the cached count. Should be called after frame is modified.</p>
 </dd></dl>
 
@@ -1701,16 +1830,16 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.dataset.find_path_using_paths">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">find_path_using_paths</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">missing_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2700-L2731"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.find_path_using_paths" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">find_path_using_paths</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">missing_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2713-L2744"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.find_path_using_paths" title="Permalink to this definition">#</a></dt>
 <dd><p>Find a path to a missing file given a set of paths to search in.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>missing_path</strong> – Path to the missing filename.</p></li>
 <li><p><strong>search_paths</strong> – List of paths to search in.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The corrected path if it was found, or the original missing path if it was not.</p>
 </dd>
 </dl>
@@ -1718,14 +1847,14 @@ <h1>sleap.io.dataset</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.dataset.load_file">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">load_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">detect_videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/dataset.py#L2734-L2776"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.load_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.dataset.</span></span><span class="sig-name descname"><span class="pre">load_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">detect_videos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/dataset.py#L2747-L2789"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.dataset.load_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a SLEAP labels file.</p>
 <p>SLEAP labels files (<code class="xref py py-obj docutils literal notranslate"><span class="pre">slp</span></code>) contain all the metadata for a labeling project or the
 predicted labels from a video. This includes the skeleton, videos, labeled frames,
 user-labeled and predicted instances, suggestions and tracks.</p>
 <p>See <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.dataset.Labels</span></code></a> for more detailed information.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a SLEAP labels (.slp) file.</p></li>
 <li><p><strong>detect_videos</strong> – If True, will attempt to detect missing videos by searching for
@@ -1740,7 +1869,7 @@ <h1>sleap.io.dataset</h1>
 contents across sets of labels.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The loaded <a class="reference internal" href="#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code></a> instance.</p>
 </dd>
 </dl>
@@ -1777,6 +1906,126 @@ <h1>sleap.io.dataset</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels"><code class="docutils literal notranslate"><span class="pre">Labels</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.labeled_frames"><code class="docutils literal notranslate"><span class="pre">Labels.labeled_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.videos"><code class="docutils literal notranslate"><span class="pre">Labels.videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.skeletons"><code class="docutils literal notranslate"><span class="pre">Labels.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.tracks"><code class="docutils literal notranslate"><span class="pre">Labels.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.suggestions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.negative_anchors"><code class="docutils literal notranslate"><span class="pre">Labels.negative_anchors</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.provenance"><code class="docutils literal notranslate"><span class="pre">Labels.provenance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_instance"><code class="docutils literal notranslate"><span class="pre">Labels.add_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.add_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_track"><code class="docutils literal notranslate"><span class="pre">Labels.add_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.add_video"><code class="docutils literal notranslate"><span class="pre">Labels.add_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.all_instances"><code class="docutils literal notranslate"><span class="pre">Labels.all_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.append"><code class="docutils literal notranslate"><span class="pre">Labels.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.append_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.append_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.clear_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.clear_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.complex_merge_between"><code class="docutils literal notranslate"><span class="pre">Labels.complex_merge_between()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.copy"><code class="docutils literal notranslate"><span class="pre">Labels.copy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.delete_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.delete_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.describe"><code class="docutils literal notranslate"><span class="pre">Labels.describe()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export"><code class="docutils literal notranslate"><span class="pre">Labels.export()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export_csv"><code class="docutils literal notranslate"><span class="pre">Labels.export_csv()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.export_nwb"><code class="docutils literal notranslate"><span class="pre">Labels.export_nwb()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.extend_from"><code class="docutils literal notranslate"><span class="pre">Labels.extend_from()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.extract"><code class="docutils literal notranslate"><span class="pre">Labels.extract()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find"><code class="docutils literal notranslate"><span class="pre">Labels.find()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_first"><code class="docutils literal notranslate"><span class="pre">Labels.find_first()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_last"><code class="docutils literal notranslate"><span class="pre">Labels.find_last()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.find_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.find_track_occupancy"><code class="docutils literal notranslate"><span class="pre">Labels.find_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.finish_complex_merge"><code class="docutils literal notranslate"><span class="pre">Labels.finish_complex_merge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.frames"><code class="docutils literal notranslate"><span class="pre">Labels.frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get"><code class="docutils literal notranslate"><span class="pre">Labels.get()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_next_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.get_next_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.get_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_track_count"><code class="docutils literal notranslate"><span class="pre">Labels.get_track_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_track_occupancy"><code class="docutils literal notranslate"><span class="pre">Labels.get_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_unlabeled_suggestion_inds"><code class="docutils literal notranslate"><span class="pre">Labels.get_unlabeled_suggestion_inds()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.get_video_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.get_video_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.has_frame"><code class="docutils literal notranslate"><span class="pre">Labels.has_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.has_missing_videos"><code class="docutils literal notranslate"><span class="pre">Labels.has_missing_videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.index"><code class="docutils literal notranslate"><span class="pre">Labels.index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.insert"><code class="docutils literal notranslate"><span class="pre">Labels.insert()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.instance_count"><code class="docutils literal notranslate"><span class="pre">Labels.instance_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.instances"><code class="docutils literal notranslate"><span class="pre">Labels.instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.is_multi_instance"><code class="docutils literal notranslate"><span class="pre">Labels.is_multi_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.labels"><code class="docutils literal notranslate"><span class="pre">Labels.labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.load_file"><code class="docutils literal notranslate"><span class="pre">Labels.load_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.make_video_callback"><code class="docutils literal notranslate"><span class="pre">Labels.make_video_callback()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_container_dicts"><code class="docutils literal notranslate"><span class="pre">Labels.merge_container_dicts()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_matching_frames"><code class="docutils literal notranslate"><span class="pre">Labels.merge_matching_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.merge_nodes"><code class="docutils literal notranslate"><span class="pre">Labels.merge_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.numpy"><code class="docutils literal notranslate"><span class="pre">Labels.numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.predicted_instances"><code class="docutils literal notranslate"><span class="pre">Labels.predicted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove"><code class="docutils literal notranslate"><span class="pre">Labels.remove()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_all_tracks"><code class="docutils literal notranslate"><span class="pre">Labels.remove_all_tracks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_empty_frames"><code class="docutils literal notranslate"><span class="pre">Labels.remove_empty_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_empty_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_empty_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_frame"><code class="docutils literal notranslate"><span class="pre">Labels.remove_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_frames"><code class="docutils literal notranslate"><span class="pre">Labels.remove_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_instance"><code class="docutils literal notranslate"><span class="pre">Labels.remove_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_predictions"><code class="docutils literal notranslate"><span class="pre">Labels.remove_predictions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_suggestion"><code class="docutils literal notranslate"><span class="pre">Labels.remove_suggestion()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_track"><code class="docutils literal notranslate"><span class="pre">Labels.remove_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_untracked_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_untracked_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_unused_tracks"><code class="docutils literal notranslate"><span class="pre">Labels.remove_unused_tracks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_user_instances"><code class="docutils literal notranslate"><span class="pre">Labels.remove_user_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.remove_video"><code class="docutils literal notranslate"><span class="pre">Labels.remove_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save"><code class="docutils literal notranslate"><span class="pre">Labels.save()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_file"><code class="docutils literal notranslate"><span class="pre">Labels.save_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_frame_data_hdf5"><code class="docutils literal notranslate"><span class="pre">Labels.save_frame_data_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.save_frame_data_imgstore"><code class="docutils literal notranslate"><span class="pre">Labels.save_frame_data_imgstore()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.set_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.set_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.skeleton"><code class="docutils literal notranslate"><span class="pre">Labels.skeleton</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.split"><code class="docutils literal notranslate"><span class="pre">Labels.split()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_dict"><code class="docutils literal notranslate"><span class="pre">Labels.to_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_json"><code class="docutils literal notranslate"><span class="pre">Labels.to_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.to_pipeline"><code class="docutils literal notranslate"><span class="pre">Labels.to_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.track_set_instance"><code class="docutils literal notranslate"><span class="pre">Labels.track_set_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.track_swap"><code class="docutils literal notranslate"><span class="pre">Labels.track_swap()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.unlabeled_suggestions"><code class="docutils literal notranslate"><span class="pre">Labels.unlabeled_suggestions</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_instances"><code class="docutils literal notranslate"><span class="pre">Labels.user_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_labeled_frame_inds"><code class="docutils literal notranslate"><span class="pre">Labels.user_labeled_frame_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.user_labeled_frames"><code class="docutils literal notranslate"><span class="pre">Labels.user_labeled_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.video"><code class="docutils literal notranslate"><span class="pre">Labels.video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.Labels.with_user_labels_only"><code class="docutils literal notranslate"><span class="pre">Labels.with_user_labels_only()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.add_instance"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.add_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.add_track"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.add_track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.find_fancy_frame_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.find_frames"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.find_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_filtered_frame_idxs()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_frame_count"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_frame_count()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_track_occupancy"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.get_video_track_occupancy"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.get_video_track_occupancy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_frame"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_instance"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_instance()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.remove_video"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.remove_video()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.track_swap"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.track_swap()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.update"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.update()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.LabelsDataCache.update_counts_for_frame"><code class="docutils literal notranslate"><span class="pre">LabelsDataCache.update_counts_for_frame()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.find_path_using_paths"><code class="docutils literal notranslate"><span class="pre">find_path_using_paths()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.dataset.load_file"><code class="docutils literal notranslate"><span class="pre">load_file()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1795,7 +2044,7 @@ <h1>sleap.io.dataset</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.adaptor.html b/api/sleap.io.format.adaptor.html
index cd71538b5..c6d7a6cd1 100644
--- a/api/sleap.io.format.adaptor.html
+++ b/api/sleap.io.format.adaptor.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.adaptor &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.adaptor &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.adaptor';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.adaptor.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,29 @@ <h1>sleap.io.format.adaptor</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor"><code class="docutils literal notranslate"><span class="pre">Adaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">Adaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">Adaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">Adaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">Adaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_match_ext"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_match_ext()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_read"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_write"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.formatted_ext_options"><code class="docutils literal notranslate"><span class="pre">Adaptor.formatted_ext_options</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.handles"><code class="docutils literal notranslate"><span class="pre">Adaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.name"><code class="docutils literal notranslate"><span class="pre">Adaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.read"><code class="docutils literal notranslate"><span class="pre">Adaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.write"><code class="docutils literal notranslate"><span class="pre">Adaptor.write()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.SleapObjectType"><code class="docutils literal notranslate"><span class="pre">SleapObjectType</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +341,11 @@ <h1>sleap.io.format.adaptor</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.adaptor">
-<span id="sleap-io-format-adaptor"></span><h1>sleap.io.format.adaptor<a class="headerlink" href="#module-sleap.io.format.adaptor" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-adaptor"></span><h1>sleap.io.format.adaptor<a class="headerlink" href="#module-sleap.io.format.adaptor" title="Permalink to this heading">#</a></h1>
 <p>File format adaptor base class.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.adaptor.</span></span><span class="sig-name descname"><span class="pre">Adaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L21-L92"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.adaptor.</span></span><span class="sig-name descname"><span class="pre">Adaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L21-L92"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor" title="Permalink to this definition">#</a></dt>
 <dd><p>File format adaptor base class.</p>
 <p>An adaptor handles reading and/or writing a specific file format. To add
 support for a new file format, you’ll create a new class which inherits from
@@ -335,13 +358,13 @@ <h1>sleap.io.format.adaptor</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L55-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L55-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L59-L61"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L59-L61"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -353,19 +376,19 @@ <h1>sleap.io.format.adaptor</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.does_match_ext">
-<span class="sig-name descname"><span class="pre">does_match_ext</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L81-L87"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_match_ext" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_match_ext</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L81-L87"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_match_ext" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L63-L65"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L63-L65"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L67-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L67-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -377,7 +400,7 @@ <h1>sleap.io.format.adaptor</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.handles">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">sleap.io.format.adaptor.SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.handles" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.handles" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the type of object that can be read/written.</p>
 <p>The Dispatch class calls this method on all registered adaptors to
 determine which to use for reading/writing.</p>
@@ -391,13 +414,13 @@ <h1>sleap.io.format.adaptor</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.read">
-<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L71-L73"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L71-L73"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.Adaptor.write">
-<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L75-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L75-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.Adaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to a file.</p>
 </dd></dl>
 
@@ -405,7 +428,7 @@ <h1>sleap.io.format.adaptor</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.adaptor.SleapObjectType">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.adaptor.</span></span><span class="sig-name descname"><span class="pre">SleapObjectType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/adaptor.py#L13-L17"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.SleapObjectType" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.adaptor.</span></span><span class="sig-name descname"><span class="pre">SleapObjectType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/adaptor.py#L13-L17"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.adaptor.SleapObjectType" title="Permalink to this definition">#</a></dt>
 <dd><p>Types of files that an adaptor could read/write.</p>
 </dd></dl>
 
@@ -433,6 +456,35 @@ <h1>sleap.io.format.adaptor</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor"><code class="docutils literal notranslate"><span class="pre">Adaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">Adaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">Adaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">Adaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">Adaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_match_ext"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_match_ext()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_read"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.does_write"><code class="docutils literal notranslate"><span class="pre">Adaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.formatted_ext_options"><code class="docutils literal notranslate"><span class="pre">Adaptor.formatted_ext_options</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.handles"><code class="docutils literal notranslate"><span class="pre">Adaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.name"><code class="docutils literal notranslate"><span class="pre">Adaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.read"><code class="docutils literal notranslate"><span class="pre">Adaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.Adaptor.write"><code class="docutils literal notranslate"><span class="pre">Adaptor.write()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.adaptor.SleapObjectType"><code class="docutils literal notranslate"><span class="pre">SleapObjectType</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -451,7 +503,7 @@ <h1>sleap.io.format.adaptor</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.alphatracker.html b/api/sleap.io.format.alphatracker.html
index 2854bab1b..a14074c0c 100644
--- a/api/sleap.io.format.alphatracker.html
+++ b/api/sleap.io.format.alphatracker.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.alphatracker &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.alphatracker &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.alphatracker';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.alphatracker.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,32 @@ <h1>sleap.io.format.alphatracker</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_match_ext()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.formatted_ext_options"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.formatted_ext_options</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_frame_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_instance_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_point_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.make_video_for_image_list()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.name"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.read"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.write"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +344,7 @@ <h1>sleap.io.format.alphatracker</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.alphatracker">
-<span id="sleap-io-format-alphatracker"></span><h1>sleap.io.format.alphatracker<a class="headerlink" href="#module-sleap.io.format.alphatracker" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-alphatracker"></span><h1>sleap.io.format.alphatracker<a class="headerlink" href="#module-sleap.io.format.alphatracker" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading AlphaTracker datasets.</p>
 <p>This can read a JSON file with labeled frames for a single video,
 or multiple videos.</p>
@@ -329,7 +355,7 @@ <h1>sleap.io.format.alphatracker</h1>
 create a video object which wraps the individual frame images.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.alphatracker.</span></span><span class="sig-name descname"><span class="pre">AlphaTrackerAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L28-L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.alphatracker.</span></span><span class="sig-name descname"><span class="pre">AlphaTrackerAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L28-L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads AlphaTracker JSON file with annotations for both single and multiple animals.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.all_exts">
@@ -339,7 +365,7 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L58-L138"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L58-L138"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 <p>Checks the format of the file at three different levels:
 - First, the upper-level format of file.json must be a list of dictionaries.
@@ -364,7 +390,7 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L140-L142"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L140-L142"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -376,19 +402,19 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext">
-<span class="sig-name descname"><span class="pre">does_match_ext</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L431-L437"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_match_ext</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L431-L437"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L144-L146"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L144-L146"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L148-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L148-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -400,16 +426,16 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict">
-<span class="sig-name descname"><span class="pre">get_alpha_tracker_frame_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L296-L309"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_alpha_tracker_frame_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L296-L309"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns a deep copy of the dictionary used for frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>annotations</strong> – List of AlphaTracker annotations for current frame.</p></li>
 <li><p><strong>filename</strong> – The filename of the image to use for the current frame.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary containing the annotations in the frame, the AlphaTracker
 class of the dictionary (“image”), and the filename.</p>
 </dd>
@@ -418,10 +444,10 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict">
-<span class="sig-name descname"><span class="pre">get_alpha_tracker_instance_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L311-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_alpha_tracker_instance_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L311-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns a deep copy of the dictionary used for instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>height</strong> – The height of the bounding box.</p></li>
 <li><p><strong>width</strong> – The width of the bounding box.</p></li>
@@ -429,7 +455,7 @@ <h1>sleap.io.format.alphatracker</h1>
 <li><p><strong>y</strong> – The y-coor of lower left hand corner of the bounding box.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary containing the AlphaTracker class of the dictionary
 (“Face”) along with the height, width, and x,y-coordinates of the
 bounding box around the instance.</p>
@@ -439,16 +465,16 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict">
-<span class="sig-name descname"><span class="pre">get_alpha_tracker_point_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L340-L351"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_alpha_tracker_point_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L340-L351"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns a deep copy of the dictionary used for nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – The x-coordinate of the node location.</p></li>
 <li><p><strong>y</strong> – The y-coordinate of the node location.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary containing the AlphaTracker class of the dictionary
 (“point”) and the x,y-coordinates of the node location.</p>
 </dd>
@@ -457,7 +483,7 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.format.adaptor.html#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">sleap.io.format.adaptor.SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.format.adaptor.html#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the type of object that can be read/written.</p>
 <p>The Dispatch class calls this method on all registered adaptors to
 determine which to use for reading/writing.</p>
@@ -465,16 +491,16 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list">
-<span class="sig-name descname"><span class="pre">make_video_for_image_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L273-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_video_for_image_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L273-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list" title="Permalink to this definition">#</a></dt>
 <dd><p>Creates a Video object from frame images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image_dir</strong> – The directory that contains the AlphaTracker file.</p></li>
 <li><p><strong>filenames</strong> – A list of filenames for the frame images.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> object.</p>
 </dd>
 </dl>
@@ -488,10 +514,10 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.read">
-<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L152-L271"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L152-L271"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>file</strong> – The file to read.</p></li>
 <li><p><strong>skeleton</strong> – The skeleton to use for <code class="xref py py-class docutils literal notranslate"><span class="pre">Instance</span></code> objects. If no skeleton
@@ -499,7 +525,7 @@ <h1>sleap.io.format.alphatracker</h1>
 <li><p><strong>full_video</strong> – The video to use for the <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> object.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> object containing all the AlphaTracker annotations.</p>
 </dd>
 </dl>
@@ -507,21 +533,21 @@ <h1>sleap.io.format.alphatracker</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.alphatracker.AlphaTrackerAdaptor.write">
-<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/alphatracker.py#L355-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/alphatracker.py#L355-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to an AlphaTracker JSON file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The name of the file being written to.</p></li>
 <li><p><strong>souce_object</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> object which contains the relevant
 information to write to an AlphaTracker JSON file.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>NotImplementedError</strong> – The code for the write functionality is not complete,
     see TODOs.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list of annotated frames in the same format seen in the AlphaTracker JSON.</p>
 </dd>
 </dl>
@@ -553,6 +579,38 @@ <h1>sleap.io.format.alphatracker</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_match_ext()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.formatted_ext_options"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.formatted_ext_options</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_frame_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_instance_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.get_alpha_tracker_point_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.make_video_for_image_list()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.name"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.read"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.alphatracker.AlphaTrackerAdaptor.write"><code class="docutils literal notranslate"><span class="pre">AlphaTrackerAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -571,7 +629,7 @@ <h1>sleap.io.format.alphatracker</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.coco.html b/api/sleap.io.format.coco.html
index a61858105..b781db41d 100644
--- a/api/sleap.io.format.coco.html
+++ b/api/sleap.io.format.coco.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.coco &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.coco &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.coco';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.coco.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.io.format.coco</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +337,12 @@ <h1>sleap.io.format.coco</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.coco">
-<span id="sleap-io-format-coco"></span><h1>sleap.io.format.coco<a class="headerlink" href="#module-sleap.io.format.coco" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-coco"></span><h1>sleap.io.format.coco<a class="headerlink" href="#module-sleap.io.format.coco" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading COCO keypoint detection datasets.</p>
 <p>See <a class="reference external" href="http://cocodataset.org/#format-data">http://cocodataset.org/#format-data</a> for details about this format.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.coco.</span></span><span class="sig-name descname"><span class="pre">LabelsCocoAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L19-L203"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.coco.</span></span><span class="sig-name descname"><span class="pre">LabelsCocoAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L19-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -332,13 +351,13 @@ <h1>sleap.io.format.coco</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L36-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L36-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L47-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L47-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -350,13 +369,13 @@ <h1>sleap.io.format.coco</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L50-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L50-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L53-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L53-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -376,7 +395,7 @@ <h1>sleap.io.format.coco</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.coco.LabelsCocoAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_missing_gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/coco.py#L56-L203"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_missing_gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/coco.py#L56-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.coco.LabelsCocoAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
@@ -406,6 +425,31 @@ <h1>sleap.io.format.coco</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.coco.LabelsCocoAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsCocoAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -424,7 +468,7 @@ <h1>sleap.io.format.coco</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.csv.html b/api/sleap.io.format.csv.html
index ffd1cd937..893965477 100644
--- a/api/sleap.io.format.csv.html
+++ b/api/sleap.io.format.csv.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.csv &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.csv &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.csv';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.csv.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.io.format.csv</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.name"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.write"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +337,11 @@ <h1>sleap.io.format.csv</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.csv">
-<span id="sleap-io-format-csv"></span><h1>sleap.io.format.csv<a class="headerlink" href="#module-sleap.io.format.csv" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-csv"></span><h1>sleap.io.format.csv<a class="headerlink" href="#module-sleap.io.format.csv" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for writing SLEAP analysis as csv.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.csv.</span></span><span class="sig-name descname"><span class="pre">CSVAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L9-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.csv.</span></span><span class="sig-name descname"><span class="pre">CSVAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L9-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -331,13 +350,13 @@ <h1>sleap.io.format.csv</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L30-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L30-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L33-L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L33-L34"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -349,13 +368,13 @@ <h1>sleap.io.format.csv</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L36-L37"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L36-L37"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L39-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L39-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -375,10 +394,10 @@ <h1>sleap.io.format.csv</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.csv.CSVAdaptor.write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/csv.py#L42-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/csv.py#L42-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.csv.CSVAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes csv file for <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> <code class="xref py py-obj docutils literal notranslate"><span class="pre">source_object</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The filename for the output file.</p></li>
 <li><p><strong>source_object</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> from which to get data from.</p></li>
@@ -418,6 +437,31 @@ <h1>sleap.io.format.csv</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.name"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.csv.CSVAdaptor.write"><code class="docutils literal notranslate"><span class="pre">CSVAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -436,7 +480,7 @@ <h1>sleap.io.format.csv</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.deeplabcut.html b/api/sleap.io.format.deeplabcut.html
index f6c3a3079..8137ed394 100644
--- a/api/sleap.io.format.deeplabcut.html
+++ b/api/sleap.io.format.deeplabcut.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.deeplabcut &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.deeplabcut &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.deeplabcut';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.deeplabcut.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,38 @@ <h1>sleap.io.format.deeplabcut</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.make_video_for_image_list()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +350,7 @@ <h1>sleap.io.format.deeplabcut</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.deeplabcut">
-<span id="sleap-io-format-deeplabcut"></span><h1>sleap.io.format.deeplabcut<a class="headerlink" href="#module-sleap.io.format.deeplabcut" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-deeplabcut"></span><h1>sleap.io.format.deeplabcut<a class="headerlink" href="#module-sleap.io.format.deeplabcut" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading DeepLabCut datasets.</p>
 <p>This can either read a CSV file with labeled frames for a single video,
 or a YAML file which potentially contains multiple videos.</p>
@@ -329,7 +361,7 @@ <h1>sleap.io.format.deeplabcut</h1>
 create a video object which wraps the individual frame images.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deeplabcut.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepLabCutCsvAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L32-L221"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deeplabcut.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepLabCutCsvAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L32-L232"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads DeepLabCut csv file with labeled frames for single video.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.all_exts">
@@ -339,13 +371,13 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L53-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L53-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L59-L60"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L59-L60"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -357,13 +389,13 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L62-L63"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L62-L63"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L65-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L65-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -377,7 +409,7 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_video_for_image_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_dir</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L82-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_video_for_image_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_dir</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L82-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list" title="Permalink to this definition">#</a></dt>
 <dd><p>Creates a Video object from frame images.</p>
 </dd></dl>
 
@@ -389,7 +421,7 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L68-L80"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L68-L80"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
@@ -397,7 +429,7 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deeplabcut.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepLabCutYamlAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L224-L323"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deeplabcut.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepLabCutYamlAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L235-L347"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -406,13 +438,13 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L241-L246"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L252-L257"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L248-L249"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L259-L260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -424,13 +456,13 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L251-L252"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L262-L263"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L254-L255"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L265-L266"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -450,7 +482,7 @@ <h1>sleap.io.format.deeplabcut</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deeplabcut.py#L257-L323"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deeplabcut.py#L268-L347"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
@@ -480,6 +512,44 @@ <h1>sleap.io.format.deeplabcut</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.make_video_for_image_list()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutCsvAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepLabCutYamlAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -498,7 +568,7 @@ <h1>sleap.io.format.deeplabcut</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.deepposekit.html b/api/sleap.io.format.deepposekit.html
index 4c39e852a..c63ff9c87 100644
--- a/api/sleap.io.format.deepposekit.html
+++ b/api/sleap.io.format.deepposekit.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.deepposekit &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.deepposekit &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.deepposekit';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.deepposekit.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.io.format.deepposekit</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +337,11 @@ <h1>sleap.io.format.deepposekit</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.deepposekit">
-<span id="sleap-io-format-deepposekit"></span><h1>sleap.io.format.deepposekit<a class="headerlink" href="#module-sleap.io.format.deepposekit" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-deepposekit"></span><h1>sleap.io.format.deepposekit<a class="headerlink" href="#module-sleap.io.format.deepposekit" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading DeepPoseKit datasets (HDF5).</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deepposekit.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepPoseKitAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L16-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.deepposekit.</span></span><span class="sig-name descname"><span class="pre">LabelsDeepPoseKitAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L16-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -331,13 +350,13 @@ <h1>sleap.io.format.deepposekit</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L33-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L33-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L42-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L42-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -349,13 +368,13 @@ <h1>sleap.io.format.deepposekit</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L45-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L45-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L48-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L48-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -375,7 +394,7 @@ <h1>sleap.io.format.deepposekit</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/deepposekit.py#L51-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/deepposekit.py#L51-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
@@ -405,6 +424,31 @@ <h1>sleap.io.format.deepposekit</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsDeepPoseKitAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -423,7 +467,7 @@ <h1>sleap.io.format.deepposekit</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.dispatch.html b/api/sleap.io.format.dispatch.html
index 68aad9a54..1ed8a7aaf 100644
--- a/api/sleap.io.format.dispatch.html
+++ b/api/sleap.io.format.dispatch.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.dispatch &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.dispatch &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.dispatch';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.dispatch.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -318,7 +316,7 @@ <h1>sleap.io.format.dispatch</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.dispatch">
-<span id="sleap-io-format-dispatch"></span><h1>sleap.io.format.dispatch<a class="headerlink" href="#module-sleap.io.format.dispatch" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-dispatch"></span><h1>sleap.io.format.dispatch<a class="headerlink" href="#module-sleap.io.format.dispatch" title="Permalink to this heading">#</a></h1>
 <p>Dispatcher for dynamically supporting multiple dataset file formats.</p>
 <p>See <code class="xref py py-obj docutils literal notranslate"><span class="pre">read()</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">write()</span></code> in <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.io.format</span></code> for usage.</p>
 </section>
@@ -363,7 +361,7 @@ <h1>sleap.io.format.dispatch</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.filehandle.html b/api/sleap.io.format.filehandle.html
index 68d84b545..cf3855ee0 100644
--- a/api/sleap.io.format.filehandle.html
+++ b/api/sleap.io.format.filehandle.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.filehandle &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.filehandle &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.filehandle';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.filehandle.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,24 @@ <h1>sleap.io.format.filehandle</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle"><code class="docutils literal notranslate"><span class="pre">FileHandle</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.close"><code class="docutils literal notranslate"><span class="pre">FileHandle.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.file"><code class="docutils literal notranslate"><span class="pre">FileHandle.file</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.format_id"><code class="docutils literal notranslate"><span class="pre">FileHandle.format_id</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.is_hdf5"><code class="docutils literal notranslate"><span class="pre">FileHandle.is_hdf5</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.is_json"><code class="docutils literal notranslate"><span class="pre">FileHandle.is_json</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.json"><code class="docutils literal notranslate"><span class="pre">FileHandle.json</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.open"><code class="docutils literal notranslate"><span class="pre">FileHandle.open()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.text"><code class="docutils literal notranslate"><span class="pre">FileHandle.text</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,18 +336,18 @@ <h1>sleap.io.format.filehandle</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.filehandle">
-<span id="sleap-io-format-filehandle"></span><h1>sleap.io.format.filehandle<a class="headerlink" href="#module-sleap.io.format.filehandle" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-filehandle"></span><h1>sleap.io.format.filehandle<a class="headerlink" href="#module-sleap.io.format.filehandle" title="Permalink to this heading">#</a></h1>
 <p>File object which can be passed to adaptors.</p>
 <p>We use this since multiple file adaptors may need to open/read the file while
 dispatch is determining which adaptor to use, and the <a class="reference internal" href="#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><code class="xref py py-obj docutils literal notranslate"><span class="pre">FileHandle</span></code></a> allows us
 to keep any results from previous reads.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.filehandle.FileHandle">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.filehandle.</span></span><span class="sig-name descname"><span class="pre">FileHandle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_hdf5</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_open</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">text</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/filehandle.py#L18-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.filehandle.</span></span><span class="sig-name descname"><span class="pre">FileHandle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_hdf5</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_open</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">text</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/filehandle.py#L18-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle" title="Permalink to this definition">#</a></dt>
 <dd><p>Reference to a file; can hold loaded data so it needn’t be read twice.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.filehandle.FileHandle.close">
-<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/filehandle.py#L53-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle.close" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/filehandle.py#L53-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle.close" title="Permalink to this definition">#</a></dt>
 <dd><p>Closes the file.</p>
 </dd></dl>
 
@@ -369,7 +387,7 @@ <h1>sleap.io.format.filehandle</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.filehandle.FileHandle.open">
-<span class="sig-name descname"><span class="pre">open</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/filehandle.py#L36-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle.open" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">open</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/filehandle.py#L36-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.filehandle.FileHandle.open" title="Permalink to this definition">#</a></dt>
 <dd><p>Opens the file (if it’s not already open).</p>
 </dd></dl>
 
@@ -405,6 +423,30 @@ <h1>sleap.io.format.filehandle</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle"><code class="docutils literal notranslate"><span class="pre">FileHandle</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.close"><code class="docutils literal notranslate"><span class="pre">FileHandle.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.file"><code class="docutils literal notranslate"><span class="pre">FileHandle.file</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.format_id"><code class="docutils literal notranslate"><span class="pre">FileHandle.format_id</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.is_hdf5"><code class="docutils literal notranslate"><span class="pre">FileHandle.is_hdf5</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.is_json"><code class="docutils literal notranslate"><span class="pre">FileHandle.is_json</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.json"><code class="docutils literal notranslate"><span class="pre">FileHandle.json</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.open"><code class="docutils literal notranslate"><span class="pre">FileHandle.open()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.filehandle.FileHandle.text"><code class="docutils literal notranslate"><span class="pre">FileHandle.text</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -423,7 +465,7 @@ <h1>sleap.io.format.filehandle</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.genericjson.html b/api/sleap.io.format.genericjson.html
index b824f1b7b..2cd9621f2 100644
--- a/api/sleap.io.format.genericjson.html
+++ b/api/sleap.io.format.genericjson.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.genericjson &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.genericjson &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.genericjson';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.genericjson.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.genericjson</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.name"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.read"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.write"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +338,12 @@ <h1>sleap.io.format.genericjson</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.genericjson">
-<span id="sleap-io-format-genericjson"></span><h1>sleap.io.format.genericjson<a class="headerlink" href="#module-sleap.io.format.genericjson" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-genericjson"></span><h1>sleap.io.format.genericjson<a class="headerlink" href="#module-sleap.io.format.genericjson" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading and writing any generic JSON file.</p>
 <p>This is a good example of a very simple adaptor class.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.genericjson.</span></span><span class="sig-name descname"><span class="pre">GenericJsonAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L13-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.genericjson.</span></span><span class="sig-name descname"><span class="pre">GenericJsonAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L13-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -332,13 +352,13 @@ <h1>sleap.io.format.genericjson</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L30-L33"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L30-L33"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L35-L36"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L35-L36"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -350,13 +370,13 @@ <h1>sleap.io.format.genericjson</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L38-L39"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L38-L39"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L41-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L41-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -376,13 +396,13 @@ <h1>sleap.io.format.genericjson</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.read">
-<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L44-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L44-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.genericjson.GenericJsonAdaptor.write">
-<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/genericjson.py#L47-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/genericjson.py#L47-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.genericjson.GenericJsonAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to a file.</p>
 </dd></dl>
 
@@ -412,6 +432,32 @@ <h1>sleap.io.format.genericjson</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.name"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.read"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.genericjson.GenericJsonAdaptor.write"><code class="docutils literal notranslate"><span class="pre">GenericJsonAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -430,7 +476,7 @@ <h1>sleap.io.format.genericjson</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.hdf5.html b/api/sleap.io.format.hdf5.html
index 373682548..c95921735 100644
--- a/api/sleap.io.format.hdf5.html
+++ b/api/sleap.io.format.hdf5.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.hdf5 &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.hdf5 &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.hdf5';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.hdf5.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.hdf5</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.write"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,13 +338,13 @@ <h1>sleap.io.format.hdf5</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.hdf5">
-<span id="sleap-io-format-hdf5"></span><h1>sleap.io.format.hdf5<a class="headerlink" href="#module-sleap.io.format.hdf5" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-hdf5"></span><h1>sleap.io.format.hdf5<a class="headerlink" href="#module-sleap.io.format.hdf5" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading/writing SLEAP datasets as HDF5 (including <code class="xref py py-obj docutils literal notranslate"><span class="pre">slp</span></code>).</p>
 <p>Note that this is not the adaptor for reading/writing the “analysis” HDF5
 format.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.hdf5.</span></span><span class="sig-name descname"><span class="pre">LabelsV1Adaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L30-L532"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.hdf5.</span></span><span class="sig-name descname"><span class="pre">LabelsV1Adaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L30-L575"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -333,13 +353,13 @@ <h1>sleap.io.format.hdf5</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L53-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L53-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L64-L65"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L64-L65"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -351,13 +371,13 @@ <h1>sleap.io.format.hdf5</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L67-L68"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L67-L68"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L70-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L70-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -377,13 +397,13 @@ <h1>sleap.io.format.hdf5</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L128-L220"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L131-L262"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.hdf5.LabelsV1Adaptor.write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">append</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_frame_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_data_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/hdf5.py#L222-L532"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">append</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_frame_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_data_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/hdf5.py#L264-L575"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.hdf5.LabelsV1Adaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to a file.</p>
 </dd></dl>
 
@@ -413,6 +433,32 @@ <h1>sleap.io.format.hdf5</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.hdf5.LabelsV1Adaptor.write"><code class="docutils literal notranslate"><span class="pre">LabelsV1Adaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -431,7 +477,7 @@ <h1>sleap.io.format.hdf5</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.labels_json.html b/api/sleap.io.format.labels_json.html
index 0e4e97d4f..a7282eb37 100644
--- a/api/sleap.io.format.labels_json.html
+++ b/api/sleap.io.format.labels_json.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.labels_json &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.labels_json &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.labels_json';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.labels_json.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,27 @@ <h1>sleap.io.format.labels_json</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.from_json_data()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.write"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +339,7 @@ <h1>sleap.io.format.labels_json</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.labels_json">
-<span id="sleap-io-format-labels-json"></span><h1>sleap.io.format.labels_json<a class="headerlink" href="#module-sleap.io.format.labels_json" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-labels-json"></span><h1>sleap.io.format.labels_json<a class="headerlink" href="#module-sleap.io.format.labels_json" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading/writing old, JSON dataset format (kind of deprecated).</p>
 <p>This supports reading and writing both <a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle.json" title="sleap.io.format.filehandle.FileHandle.json"><code class="xref py py-obj docutils literal notranslate"><span class="pre">json</span></code></a> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">json.zip</span></code>. The zip allows
 you to include image files, in imgstore videos. A better option now is to save
@@ -326,7 +347,7 @@ <h1>sleap.io.format.labels_json</h1>
 also the videos/frames as HDF5 datasets.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.labels_json.</span></span><span class="sig-name descname"><span class="pre">LabelsJsonAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L34-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.labels_json.</span></span><span class="sig-name descname"><span class="pre">LabelsJsonAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L34-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -335,13 +356,13 @@ <h1>sleap.io.format.labels_json</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L53-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L53-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L68-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L68-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -353,23 +374,23 @@ <h1>sleap.io.format.labels_json</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L71-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L71-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L74-L75"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L74-L75"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L359-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json_data</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L359-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data" title="Permalink to this definition">#</a></dt>
 <dd><p>Create instance of class from data in dictionary.</p>
 <p>Method is used by other methods that load from JSON.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – Dictionary, deserialized from JSON.</p></li>
 <li><p><strong>match_to</strong> – If given, we’ll replace particular objects in the
@@ -379,7 +400,7 @@ <h1>sleap.io.format.labels_json</h1>
 duplicate matching objects (e.g., <code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code> objects ).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> object.</p>
 </dd>
 </dl>
@@ -401,16 +422,16 @@ <h1>sleap.io.format.labels_json</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L77-L234"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_to</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L77-L234"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.labels_json.LabelsJsonAdaptor.write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">compress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_frame_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_data_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/labels_json.py#L236-L357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">compress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_frame_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_data_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_labeled</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suggested</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">None</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/labels_json.py#L236-L357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Save a Labels instance to a JSON format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The filename to save the data to.</p></li>
 <li><p><strong>source_object</strong> – The labels dataset to save.</p></li>
@@ -452,7 +473,7 @@ <h1>sleap.io.format.labels_json</h1>
 <li><p><strong>progress_callback</strong> – A function that will be called with the current progress.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -484,6 +505,33 @@ <h1>sleap.io.format.labels_json</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.from_json_data()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.labels_json.LabelsJsonAdaptor.write"><code class="docutils literal notranslate"><span class="pre">LabelsJsonAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -502,7 +550,7 @@ <h1>sleap.io.format.labels_json</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.leap_matlab.html b/api/sleap.io.format.leap_matlab.html
index be4b38528..f7013b16b 100644
--- a/api/sleap.io.format.leap_matlab.html
+++ b/api/sleap.io.format.leap_matlab.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.leap_matlab &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.leap_matlab &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.leap_matlab';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.leap_matlab.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.io.format.leap_matlab</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,13 +337,13 @@ <h1>sleap.io.format.leap_matlab</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.leap_matlab">
-<span id="sleap-io-format-leap-matlab"></span><h1>sleap.io.format.leap_matlab<a class="headerlink" href="#module-sleap.io.format.leap_matlab" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-leap-matlab"></span><h1>sleap.io.format.leap_matlab<a class="headerlink" href="#module-sleap.io.format.leap_matlab" title="Permalink to this heading">#</a></h1>
 <p>Adaptor to read (not write) LEAP MATLAB data files.</p>
 <p>This attempts to find videos. If they cannot automatically be found and the
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">gui</span></code> param is True, then the user will be prompted to find the videos.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.leap_matlab.</span></span><span class="sig-name descname"><span class="pre">LabelsLeapMatlabAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L22-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.leap_matlab.</span></span><span class="sig-name descname"><span class="pre">LabelsLeapMatlabAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L22-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -333,13 +352,13 @@ <h1>sleap.io.format.leap_matlab</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L39-L44"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L39-L44"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L46-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L46-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -351,13 +370,13 @@ <h1>sleap.io.format.leap_matlab</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L49-L50"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L49-L50"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L52-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L52-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -377,7 +396,7 @@ <h1>sleap.io.format.leap_matlab</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/leap_matlab.py#L55-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">gui</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/leap_matlab.py#L55-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
@@ -407,6 +426,31 @@ <h1>sleap.io.format.leap_matlab</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.name"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read"><code class="docutils literal notranslate"><span class="pre">LabelsLeapMatlabAdaptor.read()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -425,7 +469,7 @@ <h1>sleap.io.format.leap_matlab</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.main.html b/api/sleap.io.format.main.html
index bb2a6f4fe..2d287000c 100644
--- a/api/sleap.io.format.main.html
+++ b/api/sleap.io.format.main.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.main &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.main &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.main';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.main.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,15 @@ <h1>sleap.io.format.main</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.main.read"><code class="docutils literal notranslate"><span class="pre">read()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.main.write"><code class="docutils literal notranslate"><span class="pre">write()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +327,7 @@ <h1>sleap.io.format.main</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.main">
-<span id="sleap-io-format-main"></span><h1>sleap.io.format.main<a class="headerlink" href="#module-sleap.io.format.main" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-main"></span><h1>sleap.io.format.main<a class="headerlink" href="#module-sleap.io.format.main" title="Permalink to this heading">#</a></h1>
 <p>Read/write for multiple dataset formats.</p>
 <p>File adaptors provide a common API for code that reads and/or writes data files.
 Usually these are for reading or writing SLEAP datasets or something with
@@ -348,10 +357,10 @@ <h1>sleap.io.format.main</h1>
 also a non-default file extension for the <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsV1Adaptor</span></code> adaptor.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.format.main.read">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.format.main.</span></span><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">for_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">as_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/main.py#L71-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.main.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.format.main.</span></span><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">for_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">object</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">as_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/main.py#L71-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.main.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads file using the appropriate file format adaptor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Full filename of the file to read.</p></li>
 <li><p><strong>for_object</strong> – The type of object we’re trying to read; can be given as
@@ -373,10 +382,10 @@ <h1>sleap.io.format.main</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.format.main.write">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.format.main.</span></span><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">as_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/main.py#L118-L166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.main.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.format.main.</span></span><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">as_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/main.py#L118-L166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.main.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes SLEAP dataset file using the appropriate file format adaptor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Full filename of the file to write.
 All directories should exist.</p></li>
@@ -420,6 +429,21 @@ <h1>sleap.io.format.main</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.main.read"><code class="docutils literal notranslate"><span class="pre">read()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.main.write"><code class="docutils literal notranslate"><span class="pre">write()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -438,7 +462,7 @@ <h1>sleap.io.format.main</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.ndx_pose.html b/api/sleap.io.format.ndx_pose.html
index 154f7f9ea..196445fd3 100644
--- a/api/sleap.io.format.ndx_pose.html
+++ b/api/sleap.io.format.ndx_pose.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.ndx_pose &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.ndx_pose &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.ndx_pose';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.ndx_pose.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.ndx_pose</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.name"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.read"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.write"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +338,11 @@ <h1>sleap.io.format.ndx_pose</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.ndx_pose">
-<span id="sleap-io-format-ndx-pose"></span><h1>sleap.io.format.ndx_pose<a class="headerlink" href="#module-sleap.io.format.ndx_pose" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-ndx-pose"></span><h1>sleap.io.format.ndx_pose<a class="headerlink" href="#module-sleap.io.format.ndx_pose" title="Permalink to this heading">#</a></h1>
 <p>Adaptor to read and write ndx-pose files.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.ndx_pose.</span></span><span class="sig-name descname"><span class="pre">NDXPoseAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L23-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.ndx_pose.</span></span><span class="sig-name descname"><span class="pre">NDXPoseAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L23-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><p>Adaptor to read and write ndx-pose files.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.all_exts">
@@ -332,13 +352,13 @@ <h1>sleap.io.format.ndx_pose</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L42-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L42-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L45-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L45-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -350,19 +370,19 @@ <h1>sleap.io.format.ndx_pose</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L48-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L48-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L51-L52"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L51-L52"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.handles">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.format.adaptor.html#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">sleap.io.format.adaptor.SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.handles" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">handles</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.format.adaptor.html#sleap.io.format.adaptor.SleapObjectType" title="sleap.io.format.adaptor.SleapObjectType"><span class="pre">SleapObjectType</span></a></em><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.handles" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the type of object that can be read/written.</p>
 <p>The Dispatch class calls this method on all registered adaptors to
 determine which to use for reading/writing.</p>
@@ -376,13 +396,13 @@ <h1>sleap.io.format.ndx_pose</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.read">
-<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L54-L168"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L54-L168"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Read the NWB file and returns the appropriate deserialized <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>file</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">FileHandle</span></code> object for the NWB file to read.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object.</p>
 </dd>
 </dl>
@@ -390,7 +410,7 @@ <h1>sleap.io.format.ndx_pose</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.ndx_pose.NDXPoseAdaptor.write">
-<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">overwrite</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'Processed</span> <span class="pre">SLEAP</span> <span class="pre">pose</span> <span class="pre">data'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">identifier</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_start_time</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">datetime.datetime</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/ndx_pose.py#L170-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">overwrite</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'Processed</span> <span class="pre">SLEAP</span> <span class="pre">pose</span> <span class="pre">data'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">identifier</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">session_start_time</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">datetime</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/ndx_pose.py#L170-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Write all <code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedInstance</span></code> objects in a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object to an NWB file.</p>
 <blockquote>
 <div><p>Use <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels.numpy</span></code> to create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.NWBFile</span></code> with a separate
@@ -423,7 +443,7 @@ <h1>sleap.io.format.ndx_pose</h1>
 </div></blockquote>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Output path for the NWB format file.</p></li>
 <li><p><strong>labels</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object to covert to a NWB format file.</p></li>
@@ -441,7 +461,7 @@ <h1>sleap.io.format.ndx_pose</h1>
 not be used.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.NWBFile</span></code> with a separate <code class="xref py py-obj docutils literal notranslate"><span class="pre">pynwb.ProcessingModule</span></code> for each
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> object.</p>
 </dd>
@@ -474,6 +494,32 @@ <h1>sleap.io.format.ndx_pose</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.name"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.read"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.ndx_pose.NDXPoseAdaptor.write"><code class="docutils literal notranslate"><span class="pre">NDXPoseAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -492,7 +538,7 @@ <h1>sleap.io.format.ndx_pose</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.nix.html b/api/sleap.io.format.nix.html
index 6299ea651..b46e080ff 100644
--- a/api/sleap.io.format.nix.html
+++ b/api/sleap.io.format.nix.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.nix &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.nix &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.nix';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.nix.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.nix</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor"><code class="docutils literal notranslate"><span class="pre">NixAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.name"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.read"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.write"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,10 +338,10 @@ <h1>sleap.io.format.nix</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.nix">
-<span id="sleap-io-format-nix"></span><h1>sleap.io.format.nix<a class="headerlink" href="#module-sleap.io.format.nix" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-nix"></span><h1>sleap.io.format.nix<a class="headerlink" href="#module-sleap.io.format.nix" title="Permalink to this heading">#</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.nix.</span></span><span class="sig-name descname"><span class="pre">NixAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L15-L463"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.nix.</span></span><span class="sig-name descname"><span class="pre">NixAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L15-L463"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><p>Adaptor class for export of tracking analysis results to the generic
 [NIX](<a class="github reference external" href="https://github.com/g-node/nix">g-node/nix</a>) format.
 NIX defines a generic data model for scientific data that combines data and data
@@ -346,13 +366,13 @@ <h1>sleap.io.format.nix</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.can_read_file">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L55-L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L55-L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L60-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L60-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -364,13 +384,13 @@ <h1>sleap.io.format.nix</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.does_read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L64-L67"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L64-L67"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.does_write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L69-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L69-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -390,13 +410,13 @@ <h1>sleap.io.format.nix</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L74-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">object</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L74-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.nix.NixAdaptor.write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/nix.py#L98-L463"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">object</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/nix.py#L98-L463"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.nix.NixAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to a file.</p>
 </dd></dl>
 
@@ -426,6 +446,32 @@ <h1>sleap.io.format.nix</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor"><code class="docutils literal notranslate"><span class="pre">NixAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.name"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.read"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.nix.NixAdaptor.write"><code class="docutils literal notranslate"><span class="pre">NixAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -444,7 +490,7 @@ <h1>sleap.io.format.nix</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.sleap_analysis.html b/api/sleap.io.format.sleap_analysis.html
index dd65568cc..935a0cfb4 100644
--- a/api/sleap.io.format.sleap_analysis.html
+++ b/api/sleap.io.format.sleap_analysis.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.sleap_analysis &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.sleap_analysis &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.sleap_analysis';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.sleap_analysis.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.sleap_analysis</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.name"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +338,7 @@ <h1>sleap.io.format.sleap_analysis</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.sleap_analysis">
-<span id="sleap-io-format-sleap-analysis"></span><h1>sleap.io.format.sleap_analysis<a class="headerlink" href="#module-sleap.io.format.sleap_analysis" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-sleap-analysis"></span><h1>sleap.io.format.sleap_analysis<a class="headerlink" href="#module-sleap.io.format.sleap_analysis" title="Permalink to this heading">#</a></h1>
 <p>Adaptor to read and write analysis HDF5 files.</p>
 <p>These contain location and track data, but lack other metadata included in a
 full SLEAP dataset file.</p>
@@ -328,7 +348,7 @@ <h1>sleap.io.format.sleap_analysis</h1>
 with a <code class="xref py py-obj docutils literal notranslate"><span class="pre">track_occupancy</span></code> dataset.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.sleap_analysis.</span></span><span class="sig-name descname"><span class="pre">SleapAnalysisAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L24-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.sleap_analysis.</span></span><span class="sig-name descname"><span class="pre">SleapAnalysisAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L24-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -337,13 +357,13 @@ <h1>sleap.io.format.sleap_analysis</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L41-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L41-L48"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L50-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L50-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -355,13 +375,13 @@ <h1>sleap.io.format.sleap_analysis</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L53-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L53-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L56-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L56-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -381,16 +401,16 @@ <h1>sleap.io.format.sleap_analysis</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L59-L124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L59-L124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/sleap_analysis.py#L126-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/sleap_analysis.py#L126-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes analysis file for <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> <code class="xref py py-obj docutils literal notranslate"><span class="pre">source_object</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The filename for the output file.</p></li>
 <li><p><strong>source_object</strong> – The <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> from which to get data from.</p></li>
@@ -429,6 +449,32 @@ <h1>sleap.io.format.sleap_analysis</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.name"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write"><code class="docutils literal notranslate"><span class="pre">SleapAnalysisAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -447,7 +493,7 @@ <h1>sleap.io.format.sleap_analysis</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.format.text.html b/api/sleap.io.format.text.html
index bcd868291..8158269b2 100644
--- a/api/sleap.io.format.text.html
+++ b/api/sleap.io.format.text.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.format.text &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.format.text &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.format.text';</script>
     <link rel="canonical" href="/develop/api/sleap.io.format.text.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.io.format.text</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor"><code class="docutils literal notranslate"><span class="pre">TextAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.name"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.read"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.write"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +338,12 @@ <h1>sleap.io.format.text</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.format.text">
-<span id="sleap-io-format-text"></span><h1>sleap.io.format.text<a class="headerlink" href="#module-sleap.io.format.text" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-format-text"></span><h1>sleap.io.format.text<a class="headerlink" href="#module-sleap.io.format.text" title="Permalink to this heading">#</a></h1>
 <p>Adaptor for reading and writing any generic text file.</p>
 <p>This is a good example of a very simple adaptor class.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.text.</span></span><span class="sig-name descname"><span class="pre">TextAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L11-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.format.text.</span></span><span class="sig-name descname"><span class="pre">TextAdaptor</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L11-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.all_exts">
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">all_exts</span></span><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.all_exts" title="Permalink to this definition">#</a></dt>
@@ -332,13 +352,13 @@ <h1>sleap.io.format.text</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.can_read_file">
-<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L28-L29"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_read_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L28-L29"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.can_read_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can read this file.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.can_write_filename">
-<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L31-L32"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">can_write_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L31-L32"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.can_write_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor can write format of this filename.</p>
 </dd></dl>
 
@@ -350,13 +370,13 @@ <h1>sleap.io.format.text</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.does_read">
-<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L34-L35"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.does_read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_read</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L34-L35"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.does_read" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports reading.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.does_write">
-<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L37-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.does_write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">does_write</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L37-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.does_write" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns whether this adaptor supports writing.</p>
 </dd></dl>
 
@@ -376,13 +396,13 @@ <h1>sleap.io.format.text</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.read">
-<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">sleap.io.format.filehandle.FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L40-L41"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.read" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">read</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle" title="sleap.io.format.filehandle.FileHandle"><span class="pre">FileHandle</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L40-L41"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.read" title="Permalink to this definition">#</a></dt>
 <dd><p>Reads the file and returns the appropriate deserialized object.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.format.text.TextAdaptor.write">
-<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/format/text.py#L43-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.write" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">write</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">source_object</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/format/text.py#L43-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.format.text.TextAdaptor.write" title="Permalink to this definition">#</a></dt>
 <dd><p>Writes the object to a file.</p>
 </dd></dl>
 
@@ -412,6 +432,32 @@ <h1>sleap.io.format.text</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor"><code class="docutils literal notranslate"><span class="pre">TextAdaptor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.all_exts"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.all_exts</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.can_read_file"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.can_read_file()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.can_write_filename"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.can_write_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.default_ext"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.default_ext</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.does_read"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.does_read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.does_write"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.does_write()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.handles"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.handles</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.name"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.read"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.read()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.format.text.TextAdaptor.write"><code class="docutils literal notranslate"><span class="pre">TextAdaptor.write()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -430,7 +476,7 @@ <h1>sleap.io.format.text</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.legacy.html b/api/sleap.io.legacy.html
index 197dff1f8..c72f65d14 100644
--- a/api/sleap.io.legacy.html
+++ b/api/sleap.io.legacy.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.legacy &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.legacy &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.legacy';</script>
     <link rel="canonical" href="/develop/api/sleap.io.legacy.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,15 @@ <h1>sleap.io.legacy</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.legacy.load_labels_json_old"><code class="docutils literal notranslate"><span class="pre">load_labels_json_old()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.legacy.load_predicted_labels_json_old"><code class="docutils literal notranslate"><span class="pre">load_predicted_labels_json_old()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +327,14 @@ <h1>sleap.io.legacy</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.legacy">
-<span id="sleap-io-legacy"></span><h1>sleap.io.legacy<a class="headerlink" href="#module-sleap.io.legacy" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-legacy"></span><h1>sleap.io.legacy<a class="headerlink" href="#module-sleap.io.legacy" title="Permalink to this heading">#</a></h1>
 <p>Module for legacy LEAP dataset.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.legacy.load_labels_json_old">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.legacy.</span></span><span class="sig-name descname"><span class="pre">load_labels_json_old</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parsed_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adjust_matlab_indexing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_rel_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/legacy.py#L174-L292"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.legacy.load_labels_json_old" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.legacy.</span></span><span class="sig-name descname"><span class="pre">load_labels_json_old</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parsed_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adjust_matlab_indexing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_rel_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/legacy.py#L174-L292"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.legacy.load_labels_json_old" title="Permalink to this definition">#</a></dt>
 <dd><p>Load predicted instances from Talmo’s old JSON format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_path</strong> – The path to the JSON file.</p></li>
 <li><p><strong>parsed_json</strong> – The parsed json if already loaded, so we can save
@@ -334,7 +343,7 @@ <h1>sleap.io.legacy</h1>
 <li><p><strong>fix_rel_paths</strong> – Whether to fix paths to videos to absolute paths.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A newly constructed Labels object.</p>
 </dd>
 </dl>
@@ -342,10 +351,10 @@ <h1>sleap.io.legacy</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.legacy.load_predicted_labels_json_old">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.legacy.</span></span><span class="sig-name descname"><span class="pre">load_predicted_labels_json_old</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parsed_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adjust_matlab_indexing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_rel_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/legacy.py#L25-L171"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.legacy.load_predicted_labels_json_old" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.legacy.</span></span><span class="sig-name descname"><span class="pre">load_predicted_labels_json_old</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parsed_json</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adjust_matlab_indexing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_rel_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/legacy.py#L25-L171"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.legacy.load_predicted_labels_json_old" title="Permalink to this definition">#</a></dt>
 <dd><p>Load predicted instances from Talmo’s old JSON format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_path</strong> – The path to the JSON file.</p></li>
 <li><p><strong>parsed_json</strong> – The parsed json if already loaded, so we can save
@@ -354,7 +363,7 @@ <h1>sleap.io.legacy</h1>
 <li><p><strong>fix_rel_paths</strong> – Whether to fix paths to videos to absolute paths.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of <code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects.</p>
 </dd>
 </dl>
@@ -384,6 +393,21 @@ <h1>sleap.io.legacy</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.legacy.load_labels_json_old"><code class="docutils literal notranslate"><span class="pre">load_labels_json_old()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.legacy.load_predicted_labels_json_old"><code class="docutils literal notranslate"><span class="pre">load_predicted_labels_json_old()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -402,7 +426,7 @@ <h1>sleap.io.legacy</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.pathutils.html b/api/sleap.io.pathutils.html
index d9c4a8a18..e67a07338 100644
--- a/api/sleap.io.pathutils.html
+++ b/api/sleap.io.pathutils.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.pathutils &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.pathutils &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.pathutils';</script>
     <link rel="canonical" href="/develop/api/sleap.io.pathutils.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,16 @@ <h1>sleap.io.pathutils</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.filenames_prefix_change"><code class="docutils literal notranslate"><span class="pre">filenames_prefix_change()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.find_changed_subpath"><code class="docutils literal notranslate"><span class="pre">find_changed_subpath()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.list_file_missing"><code class="docutils literal notranslate"><span class="pre">list_file_missing()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +328,14 @@ <h1>sleap.io.pathutils</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.pathutils">
-<span id="sleap-io-pathutils"></span><h1>sleap.io.pathutils<a class="headerlink" href="#module-sleap.io.pathutils" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-pathutils"></span><h1>sleap.io.pathutils<a class="headerlink" href="#module-sleap.io.pathutils" title="Permalink to this heading">#</a></h1>
 <p>Utilities for working with file paths.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.pathutils.filenames_prefix_change">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">filenames_prefix_change</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_prefix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_prefix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">missing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confirm_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/pathutils.py#L16-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.filenames_prefix_change" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">filenames_prefix_change</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">old_prefix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_prefix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">missing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confirm_callback</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/pathutils.py#L16-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.filenames_prefix_change" title="Permalink to this definition">#</a></dt>
 <dd><p>Finds missing files by changing the initial part of paths.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filenames</strong> – The list of filenames, needn’t all be missing.</p></li>
 <li><p><strong>old_prefix</strong> – Initial part of path to replace.</p></li>
@@ -336,7 +346,7 @@ <h1>sleap.io.pathutils</h1>
 change to confirm that user wants to apply the change.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None; <code class="xref py py-obj docutils literal notranslate"><span class="pre">filenames</span></code> (and <code class="xref py py-obj docutils literal notranslate"><span class="pre">missing</span></code>, if given) have new data.</p>
 </dd>
 </dl>
@@ -344,16 +354,16 @@ <h1>sleap.io.pathutils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.pathutils.find_changed_subpath">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">find_changed_subpath</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/pathutils.py#L78-L107"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.find_changed_subpath" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">find_changed_subpath</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/pathutils.py#L78-L107"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.find_changed_subpath" title="Permalink to this definition">#</a></dt>
 <dd><p>Finds the smallest initial section of path that was changed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>old_path</strong> – Old path</p></li>
 <li><p><strong>new_path</strong> – New path</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>(initial part of old path), (corresponding replacement in new path)</p>
 </dd>
 </dl>
@@ -361,7 +371,7 @@ <h1>sleap.io.pathutils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.pathutils.list_file_missing">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">list_file_missing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/pathutils.py#L11-L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.list_file_missing" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.pathutils.</span></span><span class="sig-name descname"><span class="pre">list_file_missing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/pathutils.py#L11-L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.pathutils.list_file_missing" title="Permalink to this definition">#</a></dt>
 <dd><p>Given a list of filenames, returns list of whether file exists.</p>
 </dd></dl>
 
@@ -389,6 +399,22 @@ <h1>sleap.io.pathutils</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.filenames_prefix_change"><code class="docutils literal notranslate"><span class="pre">filenames_prefix_change()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.find_changed_subpath"><code class="docutils literal notranslate"><span class="pre">find_changed_subpath()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.pathutils.list_file_missing"><code class="docutils literal notranslate"><span class="pre">list_file_missing()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -407,7 +433,7 @@ <h1>sleap.io.pathutils</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.video.html b/api/sleap.io.video.html
index 596ac43cc..30a07e71c 100644
--- a/api/sleap.io.video.html
+++ b/api/sleap.io.video.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.video &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.video &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.video';</script>
     <link rel="canonical" href="/develop/api/sleap.io.video.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,106 @@ <h1>sleap.io.video</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.DummyVideo"><code class="docutils literal notranslate"><span class="pre">DummyVideo</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video"><code class="docutils literal notranslate"><span class="pre">HDF5Video</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.channels"><code class="docutils literal notranslate"><span class="pre">HDF5Video.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.check"><code class="docutils literal notranslate"><span class="pre">HDF5Video.check()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.close"><code class="docutils literal notranslate"><span class="pre">HDF5Video.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.dtype"><code class="docutils literal notranslate"><span class="pre">HDF5Video.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.embedded_frame_inds"><code class="docutils literal notranslate"><span class="pre">HDF5Video.embedded_frame_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.enable_source_video"><code class="docutils literal notranslate"><span class="pre">HDF5Video.enable_source_video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.frames"><code class="docutils literal notranslate"><span class="pre">HDF5Video.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.get_frame"><code class="docutils literal notranslate"><span class="pre">HDF5Video.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.has_embedded_images"><code class="docutils literal notranslate"><span class="pre">HDF5Video.has_embedded_images</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.height"><code class="docutils literal notranslate"><span class="pre">HDF5Video.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">HDF5Video.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.matches"><code class="docutils literal notranslate"><span class="pre">HDF5Video.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.reset"><code class="docutils literal notranslate"><span class="pre">HDF5Video.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.source_video"><code class="docutils literal notranslate"><span class="pre">HDF5Video.source_video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.source_video_available"><code class="docutils literal notranslate"><span class="pre">HDF5Video.source_video_available</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.width"><code class="docutils literal notranslate"><span class="pre">HDF5Video.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.channels"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.close"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.dtype"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.frames"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.height"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.imgstore"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.imgstore</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.matches"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.open"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.open()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.reset"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.width"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo"><code class="docutils literal notranslate"><span class="pre">MediaVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.channels"><code class="docutils literal notranslate"><span class="pre">MediaVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.dtype"><code class="docutils literal notranslate"><span class="pre">MediaVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.fps"><code class="docutils literal notranslate"><span class="pre">MediaVideo.fps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.frames"><code class="docutils literal notranslate"><span class="pre">MediaVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">MediaVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.height"><code class="docutils literal notranslate"><span class="pre">MediaVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.matches"><code class="docutils literal notranslate"><span class="pre">MediaVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.reset"><code class="docutils literal notranslate"><span class="pre">MediaVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.width"><code class="docutils literal notranslate"><span class="pre">MediaVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo"><code class="docutils literal notranslate"><span class="pre">NumpyVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.channels"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.dtype"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.frames"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.height"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.is_missing"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.is_missing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.matches"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.reset"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.width"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.channels"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.dtype"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.frames"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.height"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.matches"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.reset"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.width"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video"><code class="docutils literal notranslate"><span class="pre">Video</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.cattr"><code class="docutils literal notranslate"><span class="pre">Video.cattr()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.fixup_path"><code class="docutils literal notranslate"><span class="pre">Video.fixup_path()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_filename"><code class="docutils literal notranslate"><span class="pre">Video.from_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_hdf5"><code class="docutils literal notranslate"><span class="pre">Video.from_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_image_filenames"><code class="docutils literal notranslate"><span class="pre">Video.from_image_filenames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_media"><code class="docutils literal notranslate"><span class="pre">Video.from_media()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_numpy"><code class="docutils literal notranslate"><span class="pre">Video.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frame"><code class="docutils literal notranslate"><span class="pre">Video.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frames"><code class="docutils literal notranslate"><span class="pre">Video.get_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frames_safely"><code class="docutils literal notranslate"><span class="pre">Video.get_frames_safely()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.imgstore_from_filenames"><code class="docutils literal notranslate"><span class="pre">Video.imgstore_from_filenames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.is_missing"><code class="docutils literal notranslate"><span class="pre">Video.is_missing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">Video.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.num_frames"><code class="docutils literal notranslate"><span class="pre">Video.num_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.shape"><code class="docutils literal notranslate"><span class="pre">Video.shape</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_hdf5"><code class="docutils literal notranslate"><span class="pre">Video.to_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_imgstore"><code class="docutils literal notranslate"><span class="pre">Video.to_imgstore()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_pipeline"><code class="docutils literal notranslate"><span class="pre">Video.to_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.available_video_exts"><code class="docutils literal notranslate"><span class="pre">available_video_exts()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.load_video"><code class="docutils literal notranslate"><span class="pre">load_video()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +418,11 @@ <h1>sleap.io.video</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.video">
-<span id="sleap-io-video"></span><h1>sleap.io.video<a class="headerlink" href="#module-sleap.io.video" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-video"></span><h1>sleap.io.video<a class="headerlink" href="#module-sleap.io.video" title="Permalink to this heading">#</a></h1>
 <p>Video reading and writing interfaces for different formats.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.DummyVideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">DummyVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dummy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L23-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.DummyVideo" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">DummyVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dummy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L23-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.DummyVideo" title="Permalink to this definition">#</a></dt>
 <dd><p>Fake video backend,returns frames with all zeros.</p>
 <p>This can be useful when you want to look at labels for a dataset but don’t
 have access to the real video.</p>
@@ -330,10 +430,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">HDF5Video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'channels_last'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convert_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L47-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">HDF5Video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'channels_last'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convert_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L47-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video" title="Permalink to this definition">#</a></dt>
 <dd><p>Video data stored as 4D datasets in HDF5 files.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The name of the HDF5 file where the dataset with video data
 is stored.</p></li>
@@ -362,13 +462,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.check">
-<span class="sig-name descname"><span class="pre">check</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">attribute</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L82-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.check" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">check</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">attribute</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L82-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.check" title="Permalink to this definition">#</a></dt>
 <dd><p>Called by attrs to validates input format.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.close">
-<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L227-L233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.close" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L227-L233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.close" title="Permalink to this definition">#</a></dt>
 <dd><p>Close the HDF5 file object (if it’s open).</p>
 </dd></dl>
 
@@ -399,13 +499,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L301-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L301-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a frame from the underlying HDF5 video data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>idx</strong> – The index of the frame to get.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The numpy.ndarray representing the video frame data.</p>
 </dd>
 </dl>
@@ -434,13 +534,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">sleap.io.video.HDF5Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L210-L225"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">HDF5Video</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L210-L225"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if attributes match those of another video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other video to compare with.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if attributes match, False otherwise.</p>
 </dd>
 </dl>
@@ -448,13 +548,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L296-L299"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.reset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L296-L299"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.HDF5Video.reset" title="Permalink to this definition">#</a></dt>
 <dd><p>Reloads the video.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.io.video.HDF5Video.source_video">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">source_video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></em><a class="headerlink" href="#sleap.io.video.HDF5Video.source_video" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">source_video</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></em><a class="headerlink" href="#sleap.io.video.HDF5Video.source_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the source video if available, otherwise return <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 </dd></dl>
 
@@ -474,13 +574,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">ImgStoreVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L611-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">ImgStoreVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L611-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo" title="Permalink to this definition">#</a></dt>
 <dd><p>Video data stored as an ImgStore dataset.</p>
 <p>See: <a class="github reference external" href="https://github.com/loopbio/imgstore">loopbio/imgstore</a>
 This class is just a lightweight wrapper for reading such datasets as
 video sources for SLEAP.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The name of the file or directory to the imgstore.</p></li>
 <li><p><strong>index_by_original</strong> – ImgStores are great for storing a collection of
@@ -502,10 +602,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo.close">
-<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L789-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.close" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">close</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L789-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.close" title="Permalink to this definition">#</a></dt>
 <dd><p>Close the imgstore if it isn’t already closed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -525,17 +625,17 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_number</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L732-L763"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_number</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L732-L763"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a frame from the underlying ImgStore video data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>frame_number</strong> – The number of the frame to get. If
 index_by_original is set to True, then this number should
 actually be a frame index within the imgstore. That is,
 if there are 4 frames in the imgstore, this number should be
 be from 0 to 3.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The numpy.ndarray representing the video frame data.</p>
 </dd>
 </dl>
@@ -552,7 +652,7 @@ <h1>sleap.io.video</h1>
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">imgstore</span></span><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.imgstore" title="Permalink to this definition">#</a></dt>
 <dd><p>Get the underlying ImgStore object for this Video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The imgstore that is backing this video object.</p>
 </dd>
 </dl>
@@ -569,13 +669,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L655-L668"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L655-L668"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if attributes match.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The instance to comapare with.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if attributes match, False otherwise</p>
 </dd>
 </dl>
@@ -583,10 +683,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo.open">
-<span class="sig-name descname"><span class="pre">open</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L775-L787"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.open" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">open</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L775-L787"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.open" title="Permalink to this definition">#</a></dt>
 <dd><p>Open the image store if it isn’t already open.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -594,7 +694,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.ImgStoreVideo.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L727-L730"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.reset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L727-L730"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.ImgStoreVideo.reset" title="Permalink to this definition">#</a></dt>
 <dd><p>Reloads the video.</p>
 </dd></dl>
 
@@ -608,12 +708,12 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.MediaVideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">MediaVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bgr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L340-L507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">MediaVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bgr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L340-L507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo" title="Permalink to this definition">#</a></dt>
 <dd><p>Video data stored in traditional media formats readable by FFMPEG</p>
 <p>This class provides bare minimum read only interface on top of
 OpenCV’s VideoCapture class.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The name of the file (.mp4, .avi, etc)</p></li>
 <li><p><strong>grayscale</strong> – Whether the video is grayscale or not. “auto” means detect
@@ -648,7 +748,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.MediaVideo.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L486-L507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L486-L507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>See <a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code></a>.</p>
 </dd></dl>
 
@@ -660,13 +760,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.MediaVideo.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><span class="pre">sleap.io.video.MediaVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L416-L430"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><span class="pre">MediaVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L416-L430"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if attributes match those of another video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other video to compare with.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if attributes match, False otherwise.</p>
 </dd>
 </dl>
@@ -674,7 +774,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.MediaVideo.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bgr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L468-L484"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.reset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bgr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L468-L484"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.MediaVideo.reset" title="Permalink to this definition">#</a></dt>
 <dd><p>Reloads the video.</p>
 </dd></dl>
 
@@ -688,10 +788,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.NumpyVideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">NumpyVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L511-L607"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">NumpyVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L511-L607"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo" title="Permalink to this definition">#</a></dt>
 <dd><p>Video data stored as Numpy array.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Either a file to load or a numpy array of the data.</p></li>
 <li><p><strong>shape</strong> (<em>* numpy data</em>) – (frames, height, width, channels)</p></li>
@@ -718,7 +818,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.NumpyVideo.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L598-L600"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L598-L600"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>See <a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code></a>.</p>
 </dd></dl>
 
@@ -736,13 +836,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.NumpyVideo.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><span class="pre">sleap.io.video.NumpyVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L556-L566"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><span class="pre">NumpyVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L556-L566"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if attributes match those of another video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other video to compare with.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if attributes match, False otherwise.</p>
 </dd>
 </dl>
@@ -750,7 +850,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.NumpyVideo.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L593-L596"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.reset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L593-L596"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.NumpyVideo.reset" title="Permalink to this definition">#</a></dt>
 <dd><p>Reload the video.</p>
 </dd></dl>
 
@@ -764,10 +864,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.SingleImageVideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">SingleImageVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L803-L997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">SingleImageVideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L803-L997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo" title="Permalink to this definition">#</a></dt>
 <dd><p>Video wrapper for individual image files.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>filenames</strong> – Files to load as video.</p>
 </dd>
 </dl>
@@ -791,7 +891,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.SingleImageVideo.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L981-L997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L981-L997"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>See <a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><code class="xref py py-class docutils literal notranslate"><span class="pre">Video</span></code></a>.</p>
 </dd></dl>
 
@@ -803,13 +903,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.SingleImageVideo.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.SingleImageVideo" title="sleap.io.video.SingleImageVideo"><span class="pre">sleap.io.video.SingleImageVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L891-L901"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.io.video.SingleImageVideo" title="sleap.io.video.SingleImageVideo"><span class="pre">SingleImageVideo</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L891-L901"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if attributes match those of another video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other video to compare with.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if attributes match, False otherwise.</p>
 </dd>
 </dl>
@@ -817,7 +917,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.SingleImageVideo.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L947-L979"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.reset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L947-L979"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.SingleImageVideo.reset" title="Permalink to this definition">#</a></dt>
 <dd><p>Reloads the video.</p>
 </dd></dl>
 
@@ -831,7 +931,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.video.Video">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">Video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backend</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">sleap.io.video.HDF5Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><span class="pre">sleap.io.video.NumpyVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><span class="pre">sleap.io.video.MediaVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.ImgStoreVideo" title="sleap.io.video.ImgStoreVideo"><span class="pre">sleap.io.video.ImgStoreVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.SingleImageVideo" title="sleap.io.video.SingleImageVideo"><span class="pre">sleap.io.video.SingleImageVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.DummyVideo" title="sleap.io.video.DummyVideo"><span class="pre">sleap.io.video.DummyVideo</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1001-L1612"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">Video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backend</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.io.video.HDF5Video" title="sleap.io.video.HDF5Video"><span class="pre">HDF5Video</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><span class="pre">NumpyVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><span class="pre">MediaVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.ImgStoreVideo" title="sleap.io.video.ImgStoreVideo"><span class="pre">ImgStoreVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.SingleImageVideo" title="sleap.io.video.SingleImageVideo"><span class="pre">SingleImageVideo</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.io.video.DummyVideo" title="sleap.io.video.DummyVideo"><span class="pre">DummyVideo</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1001-L1620"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video" title="Permalink to this definition">#</a></dt>
 <dd><p>The top-level interface to any Video data used by SLEAP.</p>
 <p>This class provides a common interface for various supported video data
 backends. It provides the bare minimum of properties and methods that
@@ -849,7 +949,7 @@ <h1>sleap.io.video</h1>
 </pre></div>
 </div>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>backend</strong> – A backend is an object that implements the following basic
 required methods and properties</p></li>
@@ -872,10 +972,10 @@ <h1>sleap.io.video</h1>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.cattr">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">cattr</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1527-L1552"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.cattr" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">cattr</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1540-L1560"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.cattr" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a cattr converter for serialiazing/deserializing Video objects.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A cattr converter.</p>
 </dd>
 </dl>
@@ -883,7 +983,7 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.fixup_path">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">fixup_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_error</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_warning</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1554-L1612"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.fixup_path" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">fixup_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_error</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">raise_warning</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1562-L1620"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.fixup_path" title="Permalink to this definition">#</a></dt>
 <dd><p>Try to locate video if the given path doesn’t work.</p>
 <p>Given a path to a video try to find it. This is attempt to make the
 paths serialized for different video objects portable across multiple
@@ -896,18 +996,18 @@ <h1>sleap.io.video</h1>
 saved <code class="xref py py-class docutils literal notranslate"><span class="pre">Labels</span></code> dataset, it’s usually preferable to fix video
 paths using a <code class="xref py py-obj docutils literal notranslate"><span class="pre">video_search</span></code> callback or path list.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>path</strong> – The path the video asset.</p></li>
 <li><p><strong>raise_error</strong> – Whether to raise error if we cannot find video.</p></li>
 <li><p><strong>raise_warning</strong> – Whether to raise warning if we cannot find video.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>FileNotFoundError</strong> – If file still cannot be found and raise_error
     is True.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The fixed up path</p>
 </dd>
 </dl>
@@ -915,10 +1015,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.from_filename">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1242-L1282"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_filename" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1243-L1283"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of a video object, auto-detecting the backend.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – <p>The path to the video filename.
 Currently supported types are:</p>
@@ -937,7 +1037,7 @@ <h1>sleap.io.video</h1>
 <li><p><strong>kwargs</strong> – Arguments to pass to <a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><code class="xref py py-class docutils literal notranslate"><span class="pre">NumpyVideo</span></code></a></p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A Video object with the detected backend.</p>
 </dd>
 </dl>
@@ -945,11 +1045,11 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.from_hdf5">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.dataset.Dataset</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.files.File</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'channels_last'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convert_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1157-L1188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_hdf5" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dataset</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">File</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'channels_last'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convert_range</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1158-L1189"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of a video object from an HDF5 file and dataset.</p>
 <p>This is a helper method that invokes the HDF5Video backend.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>dataset</strong> – The name of the dataset or and h5.Dataset object. If
 filename is h5.File, dataset must be a str of the dataset name.</p></li>
@@ -959,7 +1059,7 @@ <h1>sleap.io.video</h1>
 <li><p><strong>convert_range</strong> – Whether we should convert data to [0, 255]-range</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A Video object with HDF5Video backend.</p>
 </dd>
 </dl>
@@ -967,24 +1067,24 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.from_image_filenames">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_image_filenames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1225-L1240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_image_filenames" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_image_filenames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1226-L1241"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_image_filenames" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of a SingleImageVideo from individual image file(s).</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.from_media">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_media</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1207-L1223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_media" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_media</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1208-L1224"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_media" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of a video object from a typical media file.</p>
 <p>For example, mp4, avi, or other types readable by FFMPEG.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The name of the file</p></li>
 <li><p><strong>args</strong> – Arguments to pass to <a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><code class="xref py py-class docutils literal notranslate"><span class="pre">MediaVideo</span></code></a></p></li>
 <li><p><strong>kwargs</strong> – Arguments to pass to <a class="reference internal" href="#sleap.io.video.MediaVideo" title="sleap.io.video.MediaVideo"><code class="xref py py-class docutils literal notranslate"><span class="pre">MediaVideo</span></code></a></p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A Video object with a MediaVideo backend</p>
 </dd>
 </dl>
@@ -992,17 +1092,17 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.from_numpy">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1190-L1205"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_numpy" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1191-L1206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.from_numpy" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of a video object from a numpy array.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The numpy array or the name of the file</p></li>
 <li><p><strong>args</strong> – Arguments to pass to <a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><code class="xref py py-class docutils literal notranslate"><span class="pre">NumpyVideo</span></code></a></p></li>
 <li><p><strong>kwargs</strong> – Arguments to pass to <a class="reference internal" href="#sleap.io.video.NumpyVideo" title="sleap.io.video.NumpyVideo"><code class="xref py py-class docutils literal notranslate"><span class="pre">NumpyVideo</span></code></a></p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A Video object with a NumpyVideo backend</p>
 </dd>
 </dl>
@@ -1010,13 +1110,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.get_frame">
-<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1094-L1104"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frame" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frame</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1094-L1104"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frame" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a single frame of video from the underlying video data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>idx</strong> – The index of the video frame</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The video frame with shape (height, width, channels)</p>
 </dd>
 </dl>
@@ -1024,13 +1124,13 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.get_frames">
-<span class="sig-name descname"><span class="pre">get_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idxs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1106-L1117"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frames" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idxs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1106-L1117"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a collection of video frames from the underlying video data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>idxs</strong> – An iterable object that contains the indices of frames.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The requested video frames with shape (len(idxs), height, width, channels).</p>
 </dd>
 </dl>
@@ -1038,9 +1138,9 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.get_frames_safely">
-<span class="sig-name descname"><span class="pre">get_frames_safely</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idxs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1119-L1149"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frames_safely" title="Permalink to this definition">#</a></dt>
-<dd><p>Return list of frame indices and frames which were successfully loaded.</p>
-<p>idxs: An iterable object that contains the indices of frames.</p>
+<span class="sig-name descname"><span class="pre">get_frames_safely</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idxs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1119-L1150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.get_frames_safely" title="Permalink to this definition">#</a></dt>
+<dd><p>Return list of frame indices and frames which were successfully loaded.
+:param idxs: An iterable object that contains the indices of frames.</p>
 <dl>
 <dt>Returns: A tuple of (frame indices, frames), where</dt><dd><ul class="simple">
 <li><p>frame indices is a subset of the specified idxs, and</p></li>
@@ -1053,16 +1153,16 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.imgstore_from_filenames">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">imgstore_from_filenames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">list</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1284-L1315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.imgstore_from_filenames" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">imgstore_from_filenames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filenames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">list</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1285-L1316"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.imgstore_from_filenames" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an imgstore from a list of image files.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filenames</strong> – List of filenames for the image files.</p></li>
 <li><p><strong>output_filename</strong> – Filename for the imgstore to create.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code></a> object for the new imgstore.</p>
 </dd>
 </dl>
@@ -1094,11 +1194,11 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.to_hdf5">
-<span class="sig-name descname"><span class="pre">to_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_numbers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1398-L1480"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_hdf5" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_numbers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1399-L1493"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert frames from arbitrary video backend to HDF5Video.</p>
 <p>Used for building an HDF5 that holds all data needed for training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>path</strong> – Filename to HDF5 (which could already exist).</p></li>
 <li><p><strong>dataset</strong> – The HDF5 dataset in which to store video frames.</p></li>
@@ -1115,7 +1215,7 @@ <h1>sleap.io.video</h1>
 all the frame indices in the dataset.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new Video object that references the HDF5 dataset.</p>
 </dd>
 </dl>
@@ -1123,11 +1223,11 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.to_imgstore">
-<span class="sig-name descname"><span class="pre">to_imgstore</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_numbers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1317-L1396"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_imgstore" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_imgstore</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_numbers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">format</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'png'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index_by_original</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1318-L1397"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_imgstore" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert frames from arbitrary video backend to ImgStoreVideo.</p>
 <p>This should facilitate conversion of any video to a loopbio imgstore.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>path</strong> – Filename or directory name to store imgstore.</p></li>
 <li><p><strong>frame_numbers</strong> – A list of frame numbers from the video to save.
@@ -1145,7 +1245,7 @@ <h1>sleap.io.video</h1>
 all the frame indices on <code class="xref py py-class docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects in the dataset.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new Video object that references the imgstore.</p>
 </dd>
 </dl>
@@ -1153,10 +1253,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.video.Video.to_pipeline">
-<span class="sig-name descname"><span class="pre">to_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">sleap.pipelines.Pipeline</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1482-L1514"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">sleap.pipelines.Pipeline</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1495-L1527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.Video.to_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a pipeline for reading the video.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>batch_size</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the video frames will be batched into rank-4
 tensors. Otherwise, single rank-3 images will be returned.</p></li>
@@ -1165,7 +1265,7 @@ <h1>sleap.io.video</h1>
 specified (default), pipeline will read the entire video.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> that builds <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> for high
 throughput I/O during inference.</p>
 </dd>
@@ -1177,10 +1277,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.video.available_video_exts">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">available_video_exts</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1615-L1627"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.available_video_exts" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">available_video_exts</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1623-L1635"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.available_video_exts" title="Permalink to this definition">#</a></dt>
 <dd><p>Return tuple of supported video extensions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Tuple of supported video extensions.</p>
 </dd>
 </dl>
@@ -1188,10 +1288,10 @@ <h1>sleap.io.video</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.video.load_video">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">load_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename:</span> <span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale:</span> <span class="pre">typing.Optional[bool]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset=&lt;class</span> <span class="pre">'NoneType'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_first:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">**kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/video.py#L1630-L1676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.load_video" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.video.</span></span><span class="sig-name descname"><span class="pre">load_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename:</span> <span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grayscale:</span> <span class="pre">~typing.Optional[bool]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset=&lt;class</span> <span class="pre">'NoneType'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">channels_first:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">**kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/video.py#L1638-L1684"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.video.load_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Open a video from disk.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a video file. The video reader backend will be determined by
 the file extension. Support extensions include: <code class="xref py py-obj docutils literal notranslate"><span class="pre">mp4</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">avi</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">h5</span></code>,
@@ -1208,7 +1308,7 @@ <h1>sleap.io.video</h1>
 no effect for non-HDF5 inputs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> instance with the appropriate backend for its format.</p>
 <p>This enables numpy-like access to video data.</p>
 </p>
@@ -1253,6 +1353,112 @@ <h1>sleap.io.video</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.DummyVideo"><code class="docutils literal notranslate"><span class="pre">DummyVideo</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video"><code class="docutils literal notranslate"><span class="pre">HDF5Video</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.channels"><code class="docutils literal notranslate"><span class="pre">HDF5Video.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.check"><code class="docutils literal notranslate"><span class="pre">HDF5Video.check()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.close"><code class="docutils literal notranslate"><span class="pre">HDF5Video.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.dtype"><code class="docutils literal notranslate"><span class="pre">HDF5Video.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.embedded_frame_inds"><code class="docutils literal notranslate"><span class="pre">HDF5Video.embedded_frame_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.enable_source_video"><code class="docutils literal notranslate"><span class="pre">HDF5Video.enable_source_video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.frames"><code class="docutils literal notranslate"><span class="pre">HDF5Video.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.get_frame"><code class="docutils literal notranslate"><span class="pre">HDF5Video.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.has_embedded_images"><code class="docutils literal notranslate"><span class="pre">HDF5Video.has_embedded_images</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.height"><code class="docutils literal notranslate"><span class="pre">HDF5Video.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">HDF5Video.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.matches"><code class="docutils literal notranslate"><span class="pre">HDF5Video.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.reset"><code class="docutils literal notranslate"><span class="pre">HDF5Video.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.source_video"><code class="docutils literal notranslate"><span class="pre">HDF5Video.source_video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.source_video_available"><code class="docutils literal notranslate"><span class="pre">HDF5Video.source_video_available</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.HDF5Video.width"><code class="docutils literal notranslate"><span class="pre">HDF5Video.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.channels"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.close"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.close()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.dtype"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.frames"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.height"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.imgstore"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.imgstore</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.matches"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.open"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.open()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.reset"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.ImgStoreVideo.width"><code class="docutils literal notranslate"><span class="pre">ImgStoreVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo"><code class="docutils literal notranslate"><span class="pre">MediaVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.channels"><code class="docutils literal notranslate"><span class="pre">MediaVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.dtype"><code class="docutils literal notranslate"><span class="pre">MediaVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.fps"><code class="docutils literal notranslate"><span class="pre">MediaVideo.fps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.frames"><code class="docutils literal notranslate"><span class="pre">MediaVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">MediaVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.height"><code class="docutils literal notranslate"><span class="pre">MediaVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.matches"><code class="docutils literal notranslate"><span class="pre">MediaVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.reset"><code class="docutils literal notranslate"><span class="pre">MediaVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.MediaVideo.width"><code class="docutils literal notranslate"><span class="pre">MediaVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo"><code class="docutils literal notranslate"><span class="pre">NumpyVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.channels"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.dtype"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.frames"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.height"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.is_missing"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.is_missing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.matches"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.reset"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.NumpyVideo.width"><code class="docutils literal notranslate"><span class="pre">NumpyVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.channels"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.dtype"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.dtype</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.frames"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.get_frame"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.height"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.matches"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.reset"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.reset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.SingleImageVideo.width"><code class="docutils literal notranslate"><span class="pre">SingleImageVideo.width</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video"><code class="docutils literal notranslate"><span class="pre">Video</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.cattr"><code class="docutils literal notranslate"><span class="pre">Video.cattr()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.fixup_path"><code class="docutils literal notranslate"><span class="pre">Video.fixup_path()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_filename"><code class="docutils literal notranslate"><span class="pre">Video.from_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_hdf5"><code class="docutils literal notranslate"><span class="pre">Video.from_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_image_filenames"><code class="docutils literal notranslate"><span class="pre">Video.from_image_filenames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_media"><code class="docutils literal notranslate"><span class="pre">Video.from_media()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.from_numpy"><code class="docutils literal notranslate"><span class="pre">Video.from_numpy()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frame"><code class="docutils literal notranslate"><span class="pre">Video.get_frame()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frames"><code class="docutils literal notranslate"><span class="pre">Video.get_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.get_frames_safely"><code class="docutils literal notranslate"><span class="pre">Video.get_frames_safely()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.imgstore_from_filenames"><code class="docutils literal notranslate"><span class="pre">Video.imgstore_from_filenames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.is_missing"><code class="docutils literal notranslate"><span class="pre">Video.is_missing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.last_frame_idx"><code class="docutils literal notranslate"><span class="pre">Video.last_frame_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.num_frames"><code class="docutils literal notranslate"><span class="pre">Video.num_frames</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.shape"><code class="docutils literal notranslate"><span class="pre">Video.shape</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_hdf5"><code class="docutils literal notranslate"><span class="pre">Video.to_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_imgstore"><code class="docutils literal notranslate"><span class="pre">Video.to_imgstore()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.Video.to_pipeline"><code class="docutils literal notranslate"><span class="pre">Video.to_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.available_video_exts"><code class="docutils literal notranslate"><span class="pre">available_video_exts()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.video.load_video"><code class="docutils literal notranslate"><span class="pre">load_video()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1271,7 +1477,7 @@ <h1>sleap.io.video</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.videowriter.html b/api/sleap.io.videowriter.html
index 76a376f7b..3e769b305 100644
--- a/api/sleap.io.videowriter.html
+++ b/api/sleap.io.videowriter.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.videowriter &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.videowriter &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.videowriter';</script>
     <link rel="canonical" href="/develop/api/sleap.io.videowriter.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,28 @@ <h1>sleap.io.videowriter</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter"><code class="docutils literal notranslate"><span class="pre">VideoWriter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter.can_use_ffmpeg"><code class="docutils literal notranslate"><span class="pre">VideoWriter.can_use_ffmpeg()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter.safe_builder"><code class="docutils literal notranslate"><span class="pre">VideoWriter.safe_builder()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.filename"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.filename</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.height"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.width"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.fps"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.fps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.crf"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.crf</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.preset"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.preset</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterOpenCV"><code class="docutils literal notranslate"><span class="pre">VideoWriterOpenCV</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +340,7 @@ <h1>sleap.io.videowriter</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.videowriter">
-<span id="sleap-io-videowriter"></span><h1>sleap.io.videowriter<a class="headerlink" href="#module-sleap.io.videowriter" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-videowriter"></span><h1>sleap.io.videowriter<a class="headerlink" href="#module-sleap.io.videowriter" title="Permalink to this heading">#</a></h1>
 <p>Module for writing avi/mp4 videos.</p>
 <p>Usage:</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="o">&gt;</span> <span class="n">writer</span> <span class="o">=</span> <span class="n">VideoWriter</span><span class="o">.</span><span class="n">safe_builder</span><span class="p">(</span><span class="n">filename</span><span class="p">,</span> <span class="n">height</span><span class="p">,</span> <span class="n">width</span><span class="p">,</span> <span class="n">fps</span><span class="p">)</span>
@@ -328,66 +350,72 @@ <h1>sleap.io.videowriter</h1>
 </div>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriter">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/videowriter.py#L17-L52"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriter" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/videowriter.py#L18-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriter" title="Permalink to this definition">#</a></dt>
 <dd><p>Abstract base class for writing avi/mp4 videos.</p>
 <dl class="py method">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriter.safe_builder">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">safe_builder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/videowriter.py#L32-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriter.safe_builder" title="Permalink to this definition">#</a></dt>
-<dd><p>Builds VideoWriter based on available dependencies.</p>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriter.can_use_ffmpeg">
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">can_use_ffmpeg</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/videowriter.py#L41-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriter.can_use_ffmpeg" title="Permalink to this definition">#</a></dt>
+<dd><p>Check if ffmpeg is available for writing videos.</p>
 </dd></dl>
 
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriter.safe_builder">
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">safe_builder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/videowriter.py#L33-L39"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriter.safe_builder" title="Permalink to this definition">#</a></dt>
+<dd><p>Builds VideoWriter based on available dependencies.</p>
 </dd></dl>
 
-<dl class="py class">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterOpenCV">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriterOpenCV</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/videowriter.py#L55-L68"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriterOpenCV" title="Permalink to this definition">#</a></dt>
-<dd><p>Writes video using OpenCV as wrapper for ffmpeg.</p>
 </dd></dl>
 
 <dl class="py class">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriterSkvideo</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'superfast'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/videowriter.py#L71-L112"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo" title="Permalink to this definition">#</a></dt>
-<dd><p>Writes video using scikit-video as wrapper for ffmpeg.</p>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriterImageio</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'superfast'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/videowriter.py#L76-L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio" title="Permalink to this definition">#</a></dt>
+<dd><p>Writes video using imageio as a wrapper for ffmpeg.</p>
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.filename">
-<span class="sig-name descname"><span class="pre">filename</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.filename" title="Permalink to this definition">#</a></dt>
-<dd><p>Path to mp4 file to save to.</p>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.filename">
+<span class="sig-name descname"><span class="pre">filename</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.filename" title="Permalink to this definition">#</a></dt>
+<dd><p>Path to video file to save to.</p>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.height">
-<span class="sig-name descname"><span class="pre">height</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.height" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.height">
+<span class="sig-name descname"><span class="pre">height</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.height" title="Permalink to this definition">#</a></dt>
 <dd><p>Height of movie frames.</p>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.width">
-<span class="sig-name descname"><span class="pre">width</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.width" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.width">
+<span class="sig-name descname"><span class="pre">width</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.width" title="Permalink to this definition">#</a></dt>
 <dd><p>Width of movie frames.</p>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.fps">
-<span class="sig-name descname"><span class="pre">fps</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.fps" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.fps">
+<span class="sig-name descname"><span class="pre">fps</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.fps" title="Permalink to this definition">#</a></dt>
 <dd><p>Playback framerate to save at.</p>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.crf">
-<span class="sig-name descname"><span class="pre">crf</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.crf" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.crf">
+<span class="sig-name descname"><span class="pre">crf</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.crf" title="Permalink to this definition">#</a></dt>
 <dd><p>Compression rate factor to control lossiness of video. Values go from
 2 to 32, with numbers in the 18 to 30 range being most common. Lower values
 mean less compressed/higher quality.</p>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterSkvideo.preset">
-<span class="sig-name descname"><span class="pre">preset</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterSkvideo.preset" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterImageio.preset">
+<span class="sig-name descname"><span class="pre">preset</span></span><a class="headerlink" href="#sleap.io.videowriter.VideoWriterImageio.preset" title="Permalink to this definition">#</a></dt>
 <dd><p>Name of the libx264 preset to use (default: “superfast”).</p>
 </dd></dl>
 
 </dd></dl>
 
+<dl class="py class">
+<dt class="sig sig-object py" id="sleap.io.videowriter.VideoWriterOpenCV">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.videowriter.</span></span><span class="sig-name descname"><span class="pre">VideoWriterOpenCV</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">height</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/videowriter.py#L60-L73"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.videowriter.VideoWriterOpenCV" title="Permalink to this definition">#</a></dt>
+<dd><p>Writes video using OpenCV as wrapper for ffmpeg.</p>
+</dd></dl>
+
 </section>
 
 
@@ -412,6 +440,34 @@ <h1>sleap.io.videowriter</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter"><code class="docutils literal notranslate"><span class="pre">VideoWriter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter.can_use_ffmpeg"><code class="docutils literal notranslate"><span class="pre">VideoWriter.can_use_ffmpeg()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriter.safe_builder"><code class="docutils literal notranslate"><span class="pre">VideoWriter.safe_builder()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.filename"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.filename</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.height"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.width"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.fps"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.fps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.crf"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.crf</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterImageio.preset"><code class="docutils literal notranslate"><span class="pre">VideoWriterImageio.preset</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.videowriter.VideoWriterOpenCV"><code class="docutils literal notranslate"><span class="pre">VideoWriterOpenCV</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -430,7 +486,7 @@ <h1>sleap.io.videowriter</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.io.visuals.html b/api/sleap.io.visuals.html
index 42937df51..7414e2289 100644
--- a/api/sleap.io.visuals.html
+++ b/api/sleap.io.visuals.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.io.visuals &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.io.visuals &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.io.visuals';</script>
     <link rel="canonical" href="/develop/api/sleap.io.visuals.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,22 @@ <h1>sleap.io.visuals</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.VideoMarkerThread"><code class="docutils literal notranslate"><span class="pre">VideoMarkerThread</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.VideoMarkerThread.run"><code class="docutils literal notranslate"><span class="pre">VideoMarkerThread.run()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.img_to_cv"><code class="docutils literal notranslate"><span class="pre">img_to_cv()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.reader"><code class="docutils literal notranslate"><span class="pre">reader()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.resize_image"><code class="docutils literal notranslate"><span class="pre">resize_image()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.save_labeled_video"><code class="docutils literal notranslate"><span class="pre">save_labeled_video()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.writer"><code class="docutils literal notranslate"><span class="pre">writer()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +334,14 @@ <h1>sleap.io.visuals</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.io.visuals">
-<span id="sleap-io-visuals"></span><h1>sleap.io.visuals<a class="headerlink" href="#module-sleap.io.visuals" title="Permalink to this headline">#</a></h1>
+<span id="sleap-io-visuals"></span><h1>sleap.io.visuals<a class="headerlink" href="#module-sleap.io.visuals" title="Permalink to this heading">#</a></h1>
 <p>Module for generating videos with visual annotation overlays.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.io.visuals.VideoMarkerThread">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">VideoMarkerThread</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">in_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">queue.Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">out_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">queue.Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">show_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_is_wedge</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">marker_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_manager</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.color.ColorManager</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">palette</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'standard'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">distinctly_color</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L156-L489"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.VideoMarkerThread" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">VideoMarkerThread</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">in_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">out_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">show_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_is_wedge</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">marker_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_manager</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ColorManager</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">palette</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'standard'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">distinctly_color</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L174-L507"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.VideoMarkerThread" title="Permalink to this definition">#</a></dt>
 <dd><p>Annotate frame images (draw instances).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>in_q</strong> – Queue with (list of frame indexes, ndarray of frame images).</p></li>
 <li><p><strong>out_q</strong> – Queue to send annotated images as
@@ -341,7 +357,7 @@ <h1>sleap.io.visuals</h1>
 </dl>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.io.visuals.VideoMarkerThread.run">
-<span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L226-L229"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.VideoMarkerThread.run" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L244-L247"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.VideoMarkerThread.run" title="Permalink to this definition">#</a></dt>
 <dd><p>Method representing the thread’s activity.</p>
 <p>You may override this method in a subclass. The standard run() method
 invokes the callable object passed to the object’s constructor as the
@@ -353,25 +369,26 @@ <h1>sleap.io.visuals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.visuals.img_to_cv">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">img_to_cv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L600-L608"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.img_to_cv" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">img_to_cv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L620-L628"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.img_to_cv" title="Permalink to this definition">#</a></dt>
 <dd><p>Prepares frame image as needed for opencv.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.visuals.reader">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">reader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">out_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">queue.Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L30-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.reader" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">reader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">out_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">background</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'original'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L30-L104"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.reader" title="Permalink to this definition">#</a></dt>
 <dd><p>Read frame images from video and send them into queue.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>out_q</strong> – Queue to send (list of frame indexes, ndarray of frame images)
 for chunks of video.</p></li>
 <li><p><strong>video</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Video</span></code> object to read.</p></li>
 <li><p><strong>frames</strong> – Full list frame indexes we want to read.</p></li>
 <li><p><strong>scale</strong> – Output scale for frame images.</p></li>
+<li><p><strong>background</strong> – output video background. Either original, black, white, grey</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None.</p>
 </dd>
 </dl>
@@ -379,16 +396,16 @@ <h1>sleap.io.visuals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.visuals.resize_image">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">resize_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L611-L624"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.resize_image" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">resize_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L631-L644"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.resize_image" title="Permalink to this definition">#</a></dt>
 <dd><p>Resizes single image with shape (height, width, channels).</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.visuals.save_labeled_video">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">save_labeled_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">show_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_is_wedge</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">marker_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_manager</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.gui.color.ColorManager</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">palette</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'standard'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">distinctly_color</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gui_progress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L492-L593"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.save_labeled_video" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">save_labeled_video</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">background</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'original'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">show_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_is_wedge</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">marker_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_manager</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ColorManager</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">palette</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'standard'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">distinctly_color</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gui_progress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L510-L613"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.save_labeled_video" title="Permalink to this definition">#</a></dt>
 <dd><p>Function to generate and save video with annotations.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Output filename.</p></li>
 <li><p><strong>labels</strong> – The dataset from which to get data.</p></li>
@@ -397,6 +414,7 @@ <h1>sleap.io.visuals</h1>
 <li><p><strong>fps</strong> – Frames per second for output video.</p></li>
 <li><p><strong>scale</strong> – scale of image (so we can scale point locations to match)</p></li>
 <li><p><strong>crop_size_xy</strong> – size of crop around instances, or None for full images</p></li>
+<li><p><strong>background</strong> – output video background. Either original, black, white, grey</p></li>
 <li><p><strong>show_edges</strong> – whether to draw lines between nodes</p></li>
 <li><p><strong>edge_is_wedge</strong> – whether to draw edges as wedges (draw as line if False)</p></li>
 <li><p><strong>marker_size</strong> – Size of marker in pixels before scaling by <code class="xref py py-obj docutils literal notranslate"><span class="pre">scale</span></code></p></li>
@@ -409,7 +427,7 @@ <h1>sleap.io.visuals</h1>
 <li><p><strong>gui_progress</strong> – Whether to show Qt GUI progress dialog.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None.</p>
 </dd>
 </dl>
@@ -417,11 +435,11 @@ <h1>sleap.io.visuals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.io.visuals.writer">
-<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">writer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">in_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">queue.Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_queue</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">queue.Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/io/visuals.py#L89-L153"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.writer" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.io.visuals.</span></span><span class="sig-name descname"><span class="pre">writer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">in_q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_queue</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Queue</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/io/visuals.py#L107-L171"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.io.visuals.writer" title="Permalink to this definition">#</a></dt>
 <dd><p>Write annotated images to video.</p>
 <p>Image size is determined by the first image received in queue.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>in_q</strong> – Queue with annotated images as (images, h, w, channels) ndarray</p></li>
 <li><p><strong>progress_queue</strong> – Queue to send progress as
@@ -431,7 +449,7 @@ <h1>sleap.io.visuals</h1>
 <li><p><strong>fps</strong> – frames per second for output video</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None.</p>
 </dd>
 </dl>
@@ -461,6 +479,28 @@ <h1>sleap.io.visuals</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.VideoMarkerThread"><code class="docutils literal notranslate"><span class="pre">VideoMarkerThread</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.VideoMarkerThread.run"><code class="docutils literal notranslate"><span class="pre">VideoMarkerThread.run()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.img_to_cv"><code class="docutils literal notranslate"><span class="pre">img_to_cv()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.reader"><code class="docutils literal notranslate"><span class="pre">reader()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.resize_image"><code class="docutils literal notranslate"><span class="pre">resize_image()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.save_labeled_video"><code class="docutils literal notranslate"><span class="pre">save_labeled_video()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.io.visuals.writer"><code class="docutils literal notranslate"><span class="pre">writer()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -479,7 +519,7 @@ <h1>sleap.io.visuals</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.message.html b/api/sleap.message.html
index bfdb2211d..f3de7fba4 100644
--- a/api/sleap.message.html
+++ b/api/sleap.message.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.message &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.message &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.message';</script>
     <link rel="canonical" href="/develop/api/sleap.message.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.message</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.BaseMessageParticipant"><code class="docutils literal notranslate"><span class="pre">BaseMessageParticipant</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver"><code class="docutils literal notranslate"><span class="pre">Receiver</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.check_message"><code class="docutils literal notranslate"><span class="pre">Receiver.check_message()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.check_messages"><code class="docutils literal notranslate"><span class="pre">Receiver.check_messages()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.push_back_message"><code class="docutils literal notranslate"><span class="pre">Receiver.push_back_message()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender"><code class="docutils literal notranslate"><span class="pre">Sender</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender.send_array"><code class="docutils literal notranslate"><span class="pre">Sender.send_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender.send_dict"><code class="docutils literal notranslate"><span class="pre">Sender.send_dict()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +337,7 @@ <h1>sleap.message</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.message">
-<span id="sleap-message"></span><h1>sleap.message<a class="headerlink" href="#module-sleap.message" title="Permalink to this headline">#</a></h1>
+<span id="sleap-message"></span><h1>sleap.message<a class="headerlink" href="#module-sleap.message" title="Permalink to this heading">#</a></h1>
 <p>Module with classes for sending and receiving messages between processes.</p>
 <p>These use ZMQ pub/sub sockets.</p>
 <p>Most of the time you’ll want the PairedSender and PairedReceiver.
@@ -327,23 +346,23 @@ <h1>sleap.message</h1>
 <p>Each message is either dictionary or dictionary + numpy ndarray.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.message.BaseMessageParticipant">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">BaseMessageParticipant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.context.Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.socket.Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L22-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.BaseMessageParticipant" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">BaseMessageParticipant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L22-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.BaseMessageParticipant" title="Permalink to this definition">#</a></dt>
 <dd><p>Base class for simple Sender and Receiver.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.message.Receiver">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">Receiver</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.context.Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.socket.Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">message_queue</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L42-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">Receiver</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">message_queue</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L42-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver" title="Permalink to this definition">#</a></dt>
 <dd><p>Receives messages from corresponding Sender.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.message.Receiver.check_message">
-<span class="sig-name descname"><span class="pre">check_message</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Any</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L75-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.check_message" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">check_message</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Any</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L75-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.check_message" title="Permalink to this definition">#</a></dt>
 <dd><p>Attempt to receive a single message.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.message.Receiver.check_messages">
-<span class="sig-name descname"><span class="pre">check_messages</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">times_to_check</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L88-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.check_messages" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">check_messages</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">times_to_check</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L88-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.check_messages" title="Permalink to this definition">#</a></dt>
 <dd><p>Attempt to receive multiple messages.</p>
 <p>This method allows us to keep up with the messages by getting
 multiple messages that have been sent since the last check.
@@ -353,7 +372,7 @@ <h1>sleap.message</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.message.Receiver.push_back_message">
-<span class="sig-name descname"><span class="pre">push_back_message</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">message</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L58-L60"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.push_back_message" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">push_back_message</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">message</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L58-L60"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Receiver.push_back_message" title="Permalink to this definition">#</a></dt>
 <dd><p>Act like we didn’t receive this message yet.</p>
 </dd></dl>
 
@@ -361,17 +380,17 @@ <h1>sleap.message</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.message.Sender">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">Sender</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.context.Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">zmq.sugar.socket.Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L119-L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.message.</span></span><span class="sig-name descname"><span class="pre">Sender</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">context</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Context</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">socket</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Socket</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L119-L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender" title="Permalink to this definition">#</a></dt>
 <dd><p>Publishes messages to corresponding Receiver.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.message.Sender.send_array">
-<span class="sig-name descname"><span class="pre">send_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">header_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">A</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flags</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L137-L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender.send_array" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">send_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">header_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">A</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flags</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L137-L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender.send_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Sends dictionary + numpy ndarray.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.message.Sender.send_dict">
-<span class="sig-name descname"><span class="pre">send_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/message.py#L131-L135"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender.send_dict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">send_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/message.py#L131-L135"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.message.Sender.send_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Sends dictionary.</p>
 </dd></dl>
 
@@ -401,6 +420,31 @@ <h1>sleap.message</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.BaseMessageParticipant"><code class="docutils literal notranslate"><span class="pre">BaseMessageParticipant</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver"><code class="docutils literal notranslate"><span class="pre">Receiver</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.check_message"><code class="docutils literal notranslate"><span class="pre">Receiver.check_message()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.check_messages"><code class="docutils literal notranslate"><span class="pre">Receiver.check_messages()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Receiver.push_back_message"><code class="docutils literal notranslate"><span class="pre">Receiver.push_back_message()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender"><code class="docutils literal notranslate"><span class="pre">Sender</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender.send_array"><code class="docutils literal notranslate"><span class="pre">Sender.send_array()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.message.Sender.send_dict"><code class="docutils literal notranslate"><span class="pre">Sender.send_dict()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -419,7 +463,7 @@ <h1>sleap.message</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.common.html b/api/sleap.nn.architectures.common.html
index d400696c6..2d339924b 100644
--- a/api/sleap.nn.architectures.common.html
+++ b/api/sleap.nn.architectures.common.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.common &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.common &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.common';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.common.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,19 @@ <h1>sleap.nn.architectures.common</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.tensor"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.tensor</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.stride"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.scale"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.scale</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +331,11 @@ <h1>sleap.nn.architectures.common</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.common">
-<span id="sleap-nn-architectures-common"></span><h1>sleap.nn.architectures.common<a class="headerlink" href="#module-sleap.nn.architectures.common" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-common"></span><h1>sleap.nn.architectures.common<a class="headerlink" href="#module-sleap.nn.architectures.common" title="Permalink to this heading">#</a></h1>
 <p>Common utilities for architecture and model building.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.common.IntermediateFeature">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.common.</span></span><span class="sig-name descname"><span class="pre">IntermediateFeature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tensor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/common.py#L8-L27"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.common.IntermediateFeature" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.common.</span></span><span class="sig-name descname"><span class="pre">IntermediateFeature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tensor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/common.py#L8-L27"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.common.IntermediateFeature" title="Permalink to this definition">#</a></dt>
 <dd><p>Intermediate feature tensor for use in skip connections.</p>
 <p>This class is effectively a named tuple to store the stride (resolution) metadata.</p>
 <dl class="py attribute">
@@ -330,7 +343,7 @@ <h1>sleap.nn.architectures.common</h1>
 <span class="sig-name descname"><span class="pre">tensor</span></span><a class="headerlink" href="#sleap.nn.architectures.common.IntermediateFeature.tensor" title="Permalink to this definition">#</a></dt>
 <dd><p>The tensor output from an intermediate layer.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>tensorflow.python.framework.ops.Tensor</p>
 </dd>
 </dl>
@@ -341,7 +354,7 @@ <h1>sleap.nn.architectures.common</h1>
 <span class="sig-name descname"><span class="pre">stride</span></span><a class="headerlink" href="#sleap.nn.architectures.common.IntermediateFeature.stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Stride of the tensor relative to the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -380,6 +393,25 @@ <h1>sleap.nn.architectures.common</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.tensor"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.tensor</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.stride"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.common.IntermediateFeature.scale"><code class="docutils literal notranslate"><span class="pre">IntermediateFeature.scale</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -398,7 +430,7 @@ <h1>sleap.nn.architectures.common</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.encoder_decoder.html b/api/sleap.nn.architectures.encoder_decoder.html
index 84c79e5b6..0d4a3f307 100644
--- a/api/sleap.nn.architectures.encoder_decoder.html
+++ b/api/sleap.nn.architectures.encoder_decoder.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.encoder_decoder &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.encoder_decoder &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.encoder_decoder';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.encoder_decoder.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,78 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock"><code class="docutils literal notranslate"><span class="pre">DecoderBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">DecoderBlock.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block"><code class="docutils literal notranslate"><span class="pre">DecoderBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock"><code class="docutils literal notranslate"><span class="pre">EncoderBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.pool"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stacks"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.decoder_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.encoder_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_decoder()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_encoder()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_stem()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.maximum_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.output_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stem_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stem_stack</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool_before_convs"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pool_before_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.num_convs"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.num_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.filters"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.activation"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.block_prefix"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.block_prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.interp_method"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_connection"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.skip_connection</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_add"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.skip_add</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_first_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_first_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +390,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.encoder_decoder">
-<span id="sleap-nn-architectures-encoder-decoder"></span><h1>sleap.nn.architectures.encoder_decoder<a class="headerlink" href="#module-sleap.nn.architectures.encoder_decoder" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-encoder-decoder"></span><h1>sleap.nn.architectures.encoder_decoder<a class="headerlink" href="#module-sleap.nn.architectures.encoder_decoder" title="Permalink to this heading">#</a></h1>
 <p>Generic encoder-decoder fully convolutional backbones.</p>
 <p>This module contains building blocks for creating encoder-decoder architectures of
 general form.</p>
@@ -346,7 +418,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <p>See the <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder" title="sleap.nn.architectures.encoder_decoder.EncoderDecoder"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EncoderDecoder</span></code></a> base class for requirements for creating new architectures.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.DecoderBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">DecoderBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L148-L182"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">DecoderBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L148-L182"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Base class for decoder blocks.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.DecoderBlock.upsampling_stride">
@@ -355,7 +427,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 to 2, such that the input tensor doubles in size after the block, but can be
 set higher to upsample in fewer steps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -363,10 +435,10 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L159-L182"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L159-L182"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the decoder block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>current_stride</strong> – The stride of input tensor.</p></li>
@@ -377,7 +449,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -387,14 +459,14 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">EncoderBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L37-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">EncoderBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L37-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Base class for encoder blocks.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderBlock.pool">
 <span class="sig-name descname"><span class="pre">pool</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.pool" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, applies max pooling at the end of the block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -406,7 +478,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Stride of the max pooling operation. If 1, the output of this
 block will be at the same stride (== 1/scale) as the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -414,7 +486,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L49-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L49-L53"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the encoder block from an input tensor.</p>
 </dd></dl>
 
@@ -422,7 +494,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">EncoderDecoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L403-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">EncoderDecoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L403-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder" title="Permalink to this definition">#</a></dt>
 <dd><p>General encoder-decoder base class.</p>
 <p>New architectures that follow the encoder-decoder pattern can be defined by
 inheriting from this class and implementing the <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack" title="sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack"><code class="xref py py-obj docutils literal notranslate"><span class="pre">encoder_stack</span></code></a> and <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack" title="sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack"><code class="xref py py-obj docutils literal notranslate"><span class="pre">decoder_stack</span></code></a>
@@ -432,7 +504,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">stacks</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stacks" title="Permalink to this definition">#</a></dt>
 <dd><p>If greater than 1, the encoder-decoder architecture will be repeated.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -448,7 +520,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="sleap.nn.architectures.encoder_decoder.DecoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.DecoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="sleap.nn.architectures.encoder_decoder.DecoderBlock"><span class="pre">DecoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of decoder blocks that define the decoder.</p>
 </dd></dl>
 
@@ -462,16 +534,16 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of encoder blocks that define the encoder.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone">
-<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L607-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L607-L676"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the entire encoder-decoder backbone.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – The input tensor.</p></li>
 <li><p><strong>current_stride</strong> – The stride of <code class="xref py py-obj docutils literal notranslate"><span class="pre">x_in</span></code> relative to the original input. This is
@@ -480,7 +552,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 during decoder construction.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of the final output tensor of the decoder and a list of
 <a href="#id7"><span class="problematic" id="id8">`</span></a>IntermediateFeature`s.</p>
 <p>The intermediate features contain the output tensors from every block except
@@ -495,10 +567,10 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder">
-<span class="sig-name descname"><span class="pre">make_decoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source_features</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'dec'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L546-L605"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_decoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source_features</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'dec'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L546-L605"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the encoder layers defined by the decoder stack configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – The input tensor.</p></li>
 <li><p><strong>current_stride</strong> – The stride of <code class="xref py py-obj docutils literal notranslate"><span class="pre">x_in</span></code> relative to the original input. This is
@@ -512,7 +584,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <li><p><strong>prefix</strong> – String prefix for naming decoder layers.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of the final output tensor of the decoder and a list of
 <a href="#id11"><span class="problematic" id="id12">`</span></a>IntermediateFeature`s.</p>
 <p>The intermediate features contain the output tensors from every block except
@@ -525,10 +597,10 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder">
-<span class="sig-name descname"><span class="pre">make_encoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'enc'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L508-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_encoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'enc'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L508-L544"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the encoder layers defined by the encoder stack configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – The input tensor.</p></li>
 <li><p><strong>current_stride</strong> – The stride of <code class="xref py py-obj docutils literal notranslate"><span class="pre">x_in</span></code> relative to the original input. If any
@@ -538,7 +610,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <li><p><strong>prefix</strong> – String prefix for naming encoder layers.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of the final output tensor of the encoder and a list of
 <a href="#id13"><span class="problematic" id="id14">`</span></a>IntermediateFeature`s.</p>
 <p>The intermediate features contain the output tensors from every block except
@@ -550,11 +622,11 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem">
-<span class="sig-name descname"><span class="pre">make_stem</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stem'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L484-L506"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_stem</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stem'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L484-L506"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the stem layers defined by the stem block configuration.</p>
 <p>Unlike in the encoder, the stem layers do not get repeated in stacked models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – The input tensor.</p></li>
 <li><p><strong>current_stride</strong> – The stride of <code class="xref py py-obj docutils literal notranslate"><span class="pre">x_in</span></code> relative to the original input. If any
@@ -563,7 +635,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <li><p><strong>prefix</strong> – String prefix for naming stem layers.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The final output tensor of the stem.</p>
 </dd>
 </dl>
@@ -591,7 +663,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.EncoderBlock</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">EncoderBlock</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of encoder blocks that define the stem.</p>
 </dd></dl>
 
@@ -599,14 +671,14 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleConvBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">SimpleConvBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pool_before_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">32</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L57-L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">SimpleConvBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pool_before_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">32</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L57-L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Flexible block of convolutions and max pooling.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool">
 <span class="sig-name descname"><span class="pre">pool</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, applies max pooling at the end of the block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -618,7 +690,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Stride of the max pooling operation. If 1, the output of this
 block will be at the same stride (== 1/scale) as the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -629,7 +701,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">pool_before_convs</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool_before_convs" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, max pooling is performed before convolutions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -641,7 +713,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Number of convolution layers with activation. All attributes below
 are the same for all convolution layers within the block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -652,7 +724,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of convolutional kernel filters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -663,7 +735,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">kernel_size</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.kernel_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Size of convolutional kernels (== height == width).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -674,7 +746,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">use_bias</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.use_bias" title="Permalink to this definition">#</a></dt>
 <dd><p>If False, convolution layers will not have a bias term.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -685,7 +757,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">batch_norm</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, applies batch normalization after each convolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -705,7 +777,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -716,7 +788,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">activation</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.activation" title="Permalink to this definition">#</a></dt>
 <dd><p>Name of activation function (typically “relu” or “linear”).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -727,7 +799,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">block_prefix</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.block_prefix" title="Permalink to this definition">#</a></dt>
 <dd><p>String to append to the prefix provided at block creation time.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -739,10 +811,10 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv_block'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L94-L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv_block'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L94-L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>prefix</strong> – String that will be added to the name of every layer in the block.
@@ -750,7 +822,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -760,7 +832,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">SimpleUpsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'bilinear'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connection</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_add</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_first_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L186-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.encoder_decoder.</span></span><span class="sig-name descname"><span class="pre">SimpleUpsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'bilinear'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connection</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_add</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_first_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batch_norm_before_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'relu'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L186-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Standard block of upsampling with optional refinement and skip connections.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.upsampling_stride">
@@ -769,7 +841,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 to 2, such that the input tensor doubles in size after the block, but can be
 set higher to upsample in fewer steps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -782,7 +854,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 learnable upsampling. If False, interpolated upsampling will be used (see
 <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.interp_method" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.interp_method"><code class="xref py py-obj docutils literal notranslate"><span class="pre">interp_method</span></code></a>) and <code class="xref py py-obj docutils literal notranslate"><span class="pre">transposed_conv_*</span></code> attributes will have no effect.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -794,7 +866,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Integer that specifies the number of filters in the
 transposed convolution layer.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -805,7 +877,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">transposed_conv_kernel_size</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_kernel_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Size of the kernel for the transposed convolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -817,7 +889,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>If False, transposed convolution layers will not have
 a bias term.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -829,7 +901,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>If True, applies batch normalization after the
 transposed convolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -849,7 +921,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -861,7 +933,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Name of activation function (typically “relu” or
 “linear”).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -874,7 +946,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv"><code class="xref py py-obj docutils literal notranslate"><span class="pre">transposed_conv</span></code></a> is set to False. This can be <code class="xref py py-obj docutils literal notranslate"><span class="pre">bilinear</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">nearest</span></code>. See
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.layers.UpSampling2D</span></code> for more details on the implementation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -888,7 +960,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 False, no skip connection will be formed even if a source feature is
 available.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -903,7 +975,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 first (similar to residual shortcut connections). If False, the two tensors
 will be concatenated channelwise instead.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -921,7 +993,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 attributes will have no effect. If greater than 1, all layers will be
 identical with respect to these attributes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -933,7 +1005,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>If not None, the first refinement conv layer will
 have this many filters, otherwise <a class="reference internal" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_filters" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_filters"><code class="xref py py-obj docutils literal notranslate"><span class="pre">refine_convs_filters</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -945,7 +1017,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Specifies the number of filters to use for the refinement
 convolutions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -956,7 +1028,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <span class="sig-name descname"><span class="pre">refine_convs_kernel_size</span></span><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_kernel_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Size of the kernel for the refinement convolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -968,7 +1040,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>If False, refinement convolution layers will not have a
 bias term.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -980,7 +1052,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>If True, applies batch normalization after each
 refinement convolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1000,7 +1072,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1012,7 +1084,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 <dd><p>Name of activation function (typically “relu” or
 “linear”).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1024,10 +1096,10 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/encoder_decoder.py#L275-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/encoder_decoder.py#L275-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the decoder block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>current_stride</strong> – The stride of input tensor. Not required but if provided,
@@ -1039,7 +1111,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -1071,6 +1143,84 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock"><code class="docutils literal notranslate"><span class="pre">DecoderBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">DecoderBlock.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block"><code class="docutils literal notranslate"><span class="pre">DecoderBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock"><code class="docutils literal notranslate"><span class="pre">EncoderBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.pool"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block"><code class="docutils literal notranslate"><span class="pre">EncoderBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stacks"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.decoder_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.encoder_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_decoder()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_encoder()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.make_stem()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.maximum_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.output_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_features_stride"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stem_features_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack"><code class="docutils literal notranslate"><span class="pre">EncoderDecoder.stem_stack</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool_before_convs"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.pool_before_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.num_convs"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.num_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.filters"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.activation"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.block_prefix"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.block_prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block"><code class="docutils literal notranslate"><span class="pre">SimpleConvBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.transposed_conv_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.interp_method"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_connection"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.skip_connection</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_add"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.skip_add</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_first_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_first_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_filters"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_kernel_size"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_use_bias"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_use_bias</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_batch_norm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm_before_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_batch_norm_before_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_activation"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.refine_convs_activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">SimpleUpsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1089,7 +1239,7 @@ <h1>sleap.nn.architectures.encoder_decoder</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.hourglass.html b/api/sleap.nn.architectures.hourglass.html
index a9662f871..cd6c342c2 100644
--- a/api/sleap.nn.architectures.hourglass.html
+++ b/api/sleap.nn.architectures.hourglass.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.hourglass &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.hourglass &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.hourglass';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.hourglass.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,49 @@ <h1>sleap.nn.architectures.hourglass</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.filters"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass"><code class="docutils literal notranslate"><span class="pre">Hourglass</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.down_blocks"><code class="docutils literal notranslate"><span class="pre">Hourglass.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.up_blocks"><code class="docutils literal notranslate"><span class="pre">Hourglass.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_filters"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stride"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.filters"><code class="docutils literal notranslate"><span class="pre">Hourglass.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.filter_increase"><code class="docutils literal notranslate"><span class="pre">Hourglass.filter_increase</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.interp_method"><code class="docutils literal notranslate"><span class="pre">Hourglass.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stacks"><code class="docutils literal notranslate"><span class="pre">Hourglass.stacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.decoder_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.encoder_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.from_config"><code class="docutils literal notranslate"><span class="pre">Hourglass.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_stack</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock"><code class="docutils literal notranslate"><span class="pre">StemBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.pool"><code class="docutils literal notranslate"><span class="pre">StemBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">StemBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.filters"><code class="docutils literal notranslate"><span class="pre">StemBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.output_filters"><code class="docutils literal notranslate"><span class="pre">StemBlock.output_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.first_conv_stride"><code class="docutils literal notranslate"><span class="pre">StemBlock.first_conv_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.make_block"><code class="docutils literal notranslate"><span class="pre">StemBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.interp_method"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.conv"><code class="docutils literal notranslate"><span class="pre">conv()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +361,12 @@ <h1>sleap.nn.architectures.hourglass</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.hourglass">
-<span id="sleap-nn-architectures-hourglass"></span><h1>sleap.nn.architectures.hourglass<a class="headerlink" href="#module-sleap.nn.architectures.hourglass" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-hourglass"></span><h1>sleap.nn.architectures.hourglass<a class="headerlink" href="#module-sleap.nn.architectures.hourglass" title="Permalink to this heading">#</a></h1>
 <p>This module provides a generalized implementation of (stacked) hourglass.</p>
 <p>See the <a class="reference internal" href="#sleap.nn.architectures.hourglass.Hourglass" title="sleap.nn.architectures.hourglass.Hourglass"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Hourglass</span></code></a> class docstring for more information.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.DownsamplingBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">DownsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L104-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.DownsamplingBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">DownsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L104-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.DownsamplingBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Convolutional downsampling block of the hourglass.</p>
 <p>This block is the simplified convolution-only block described in the <a class="reference external" href="https://arxiv.org/abs/1611.05424">Associative
 Embedding paper</a>, not the original residual
@@ -338,7 +381,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of filters in the convolutional layer of the block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -346,10 +389,10 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.DownsamplingBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'downsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L121-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'downsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L121-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>prefix</strong> – String that will be added to the name of every layer in the block.
@@ -357,7 +400,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -367,7 +410,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.Hourglass">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">Hourglass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter_increase</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'nearest'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L195-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">Hourglass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter_increase</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'nearest'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L195-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass" title="Permalink to this definition">#</a></dt>
 <dd><p>Encoder-decoder definition of the (stacked) hourglass network backbone.</p>
 <p>This implements the architecture of the <a class="reference external" href="https://arxiv.org/abs/1611.05424">Associative Embedding paper</a>, which improves upon the architecture in the
 <a class="reference external" href="https://arxiv.org/abs/1603.06937">original hourglass paper</a>. The primary changes
@@ -383,7 +426,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <dd><p>Number of downsampling blocks. The original implementation has 4
 downsampling blocks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -396,7 +439,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 and has 4 upsampling blocks. If specifying more than 1 stack, this should be
 equal to down_blocks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -410,7 +453,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 convenient way to reduce the input image size while extracting fine-scale
 image features. In the original implementation this is 128.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -423,7 +466,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 increases the spatial receptive field at the cost of losing fine details at
 higher resolution. In the original implementation this is 4.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -436,7 +479,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 block, where subsequent blocks will have an linearly increasing number of
 filters (see <a class="reference internal" href="#sleap.nn.architectures.hourglass.Hourglass.filter_increase" title="sleap.nn.architectures.hourglass.Hourglass.filter_increase"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filter_increase</span></code></a>). In the original implementation this is 256.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -449,7 +492,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 block by. This number is added, not multiplied, at each block. In the
 original implementation this is 128.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -462,7 +505,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 original implementation this is nearest neighbor interpolation. Valid values
 are “nearest” or “bilinear”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -475,7 +518,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 stacks. Intermediate outputs are returned which can be used to apply
 intermediate supervision.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -483,25 +526,25 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.Hourglass.decoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="sleap.nn.architectures.encoder_decoder.DecoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.DecoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.decoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.DecoderBlock" title="sleap.nn.architectures.encoder_decoder.DecoderBlock"><span class="pre">DecoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.decoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define decoder stack configuration.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.Hourglass.encoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.encoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.encoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define encoder stack configuration.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.Hourglass.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.HourglassConfig" title="sleap.nn.config.model.HourglassConfig"><span class="pre">sleap.nn.config.model.HourglassConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.hourglass.Hourglass" title="sleap.nn.architectures.hourglass.Hourglass"><span class="pre">sleap.nn.architectures.hourglass.Hourglass</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L283-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.HourglassConfig" title="sleap.nn.config.model.HourglassConfig"><span class="pre">HourglassConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.hourglass.Hourglass" title="sleap.nn.architectures.hourglass.Hourglass"><span class="pre">Hourglass</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L283-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a model from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">HourglassConfig</span></code> instance with the desired parameters.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class with the specified configuration.</p>
 </dd>
 </dl>
@@ -509,7 +552,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.Hourglass.stem_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.EncoderBlock" title="sleap.nn.architectures.encoder_decoder.EncoderBlock"><span class="pre">EncoderBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define stem stack configuration.</p>
 </dd></dl>
 
@@ -517,7 +560,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.StemBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">StemBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L49-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">StemBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L49-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Stem layers of the hourglass. These are not repeated with multiple stacks.</p>
 <dl class="simple">
 <dt>The default structure of this block is:</dt><dd><p>Conv(7 x 7 x filters, stride 2) -&gt; Conv(3 x 3 x 2*filters) -&gt; MaxPool(stride 2)
@@ -529,7 +572,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <span class="sig-name descname"><span class="pre">pool</span></span><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock.pool" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, pooling is applied. See <a class="reference internal" href="#sleap.nn.architectures.hourglass.StemBlock.pooling_stride" title="sleap.nn.architectures.hourglass.StemBlock.pooling_stride"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pooling_stride</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -543,7 +586,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 have a stride of 2. If set to 4, the first convolution and the max pooling
 layer will both have a stride of 2.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -554,7 +597,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Base number of convolutional filters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -565,7 +608,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <span class="sig-name descname"><span class="pre">output_filters</span></span><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock.output_filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of filters to output at the end of this block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -580,10 +623,10 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.StemBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stem'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L73-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'stem'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L73-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.StemBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>prefix</strong> – String that will be added to the name of every layer in the block.
@@ -591,7 +634,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -601,7 +644,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.UpsamplingBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">UpsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'bilinear'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L141-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.UpsamplingBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">UpsamplingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">interp_method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'bilinear'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L141-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.UpsamplingBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Upsampling block that integrates skip connections with refinement.</p>
 <p>This block implements both the intermediate block after the skip connection from the
 downsampling path, as well as the upsampling block from the main network backbone
@@ -617,7 +660,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of filters in the output tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -630,7 +673,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 implementation, nearest neighbor interpolation was used. Valid values are
 “nearest” or “bilinear”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -638,10 +681,10 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.UpsamplingBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L163-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'upsample'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L163-L191"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the upsampling block from an input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input tensor to the block.</p></li>
 <li><p><strong>current_stride</strong> – The stride of input tensor.</p></li>
@@ -652,7 +695,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 name conflicts if existing layers have the same name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying all operations in the block.</p>
 </dd>
 </dl>
@@ -662,10 +705,10 @@ <h1>sleap.nn.architectures.hourglass</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hourglass.conv">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">conv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hourglass.py#L17-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.conv" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hourglass.</span></span><span class="sig-name descname"><span class="pre">conv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hourglass.py#L17-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hourglass.conv" title="Permalink to this definition">#</a></dt>
 <dd><p>Apply basic convolution with ReLU and batch normalization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – Input tensor.</p></li>
 <li><p><strong>filters</strong> – Number of convolutional filters (output channels).</p></li>
@@ -674,7 +717,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
 <li><p><strong>prefix</strong> – String to prepend to the sublayers of this convolution.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The output tensor after applying convolution and batch normalization.</p>
 </dd>
 </dl>
@@ -704,6 +747,55 @@ <h1>sleap.nn.architectures.hourglass</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.filters"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.DownsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">DownsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass"><code class="docutils literal notranslate"><span class="pre">Hourglass</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.down_blocks"><code class="docutils literal notranslate"><span class="pre">Hourglass.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.up_blocks"><code class="docutils literal notranslate"><span class="pre">Hourglass.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_filters"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stride"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.filters"><code class="docutils literal notranslate"><span class="pre">Hourglass.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.filter_increase"><code class="docutils literal notranslate"><span class="pre">Hourglass.filter_increase</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.interp_method"><code class="docutils literal notranslate"><span class="pre">Hourglass.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stacks"><code class="docutils literal notranslate"><span class="pre">Hourglass.stacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.decoder_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.encoder_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.from_config"><code class="docutils literal notranslate"><span class="pre">Hourglass.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.Hourglass.stem_stack"><code class="docutils literal notranslate"><span class="pre">Hourglass.stem_stack</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock"><code class="docutils literal notranslate"><span class="pre">StemBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.pool"><code class="docutils literal notranslate"><span class="pre">StemBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">StemBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.filters"><code class="docutils literal notranslate"><span class="pre">StemBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.output_filters"><code class="docutils literal notranslate"><span class="pre">StemBlock.output_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.first_conv_stride"><code class="docutils literal notranslate"><span class="pre">StemBlock.first_conv_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.StemBlock.make_block"><code class="docutils literal notranslate"><span class="pre">StemBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.interp_method"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.interp_method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.UpsamplingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">UpsamplingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hourglass.conv"><code class="docutils literal notranslate"><span class="pre">conv()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -722,7 +814,7 @@ <h1>sleap.nn.architectures.hourglass</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.hrnet.html b/api/sleap.nn.architectures.hrnet.html
index 24c2cf0ba..eda0a33c8 100644
--- a/api/sleap.nn.architectures.hrnet.html
+++ b/api/sleap.nn.architectures.hrnet.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.hrnet &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.hrnet &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.hrnet';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.hrnet.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,28 @@ <h1>sleap.nn.architectures.hrnet</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet"><code class="docutils literal notranslate"><span class="pre">HigherHRNet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.C"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.C</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.initial_downsampling_steps"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.initial_downsampling_steps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.n_deconv_modules"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.n_deconv_modules</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.bottleneck"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.bottleneck</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.deconv_filters"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.deconv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.bilinear_upsampling"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.bilinear_upsampling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.down_blocks"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.output"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.output()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.output_scale"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.output_scale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.adjust_prefix"><code class="docutils literal notranslate"><span class="pre">adjust_prefix()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.bottleneck_block"><code class="docutils literal notranslate"><span class="pre">bottleneck_block()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.simple_block"><code class="docutils literal notranslate"><span class="pre">simple_block()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +340,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.hrnet">
-<span id="sleap-nn-architectures-hrnet"></span><h1>sleap.nn.architectures.hrnet<a class="headerlink" href="#module-sleap.nn.architectures.hrnet" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-hrnet"></span><h1>sleap.nn.architectures.hrnet<a class="headerlink" href="#module-sleap.nn.architectures.hrnet" title="Permalink to this heading">#</a></h1>
 <p>(Higher)HRNet backbone.</p>
 <p>This implementation is based on the PyTorch implementation of HRNet, modified to
 implement HigherHRNet’s configuration and new deconvolution heads.</p>
@@ -327,7 +349,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 <a class="reference external" href="https://arxiv.org/pdf/1908.10357.pdf">https://arxiv.org/pdf/1908.10357.pdf</a></p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.HigherHRNet">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">HigherHRNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">18</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_downsampling_steps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_deconv_modules</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bottleneck</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">deconv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bilinear_upsampling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hrnet.py#L576-L645"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.HigherHRNet" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">HigherHRNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">C</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">18</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_downsampling_steps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_deconv_modules</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bottleneck</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">deconv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bilinear_upsampling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hrnet.py#L576-L645"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.HigherHRNet" title="Permalink to this definition">#</a></dt>
 <dd><p>HigherHRNet backbone.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.HigherHRNet.C">
@@ -336,7 +358,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 params. This number is effectively the number of filters at the highest
 resolution output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -349,7 +371,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 Decrease if this introduces too much loss of resolution from the initial
 images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -362,7 +384,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 equal to initial_downsampling_steps, the output will be at the same scale as
 the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -373,7 +395,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 <span class="sig-name descname"><span class="pre">bottleneck</span></span><a class="headerlink" href="#sleap.nn.architectures.hrnet.HigherHRNet.bottleneck" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, uses bottleneck blocks instead of simple residual blocks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -385,7 +407,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 <dd><p>Number of filters to use in deconv blocks if using transposed
 convolutions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -397,7 +419,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
 <dd><p>Use bilinear upsampling instead of transposed convolutions
 at the output heads.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -411,23 +433,23 @@ <h1>sleap.nn.architectures.hrnet</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.HigherHRNet.output">
-<span class="sig-name descname"><span class="pre">output</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_output_channels</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hrnet.py#L604-L633"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.HigherHRNet.output" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">output</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_output_channels</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hrnet.py#L604-L633"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.HigherHRNet.output" title="Permalink to this definition">#</a></dt>
 <dd><p>Builds the layers for this backbone and return the output tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – Input 4D tf.Tensor.</p></li>
 <li><p><strong>n_output_channels</strong> – The number of final output channels.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt>A tf.keras.model whose outputs are a list of tf.Tensors</dt><dd><p>at each scale of the deconv_modules.</p>
 </dd>
 </dl>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>higher_hrnet_model</p>
 </dd>
 </dl>
@@ -443,19 +465,19 @@ <h1>sleap.nn.architectures.hrnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.adjust_prefix">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">adjust_prefix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hrnet.py#L16-L25"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.adjust_prefix" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">adjust_prefix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hrnet.py#L16-L25"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.adjust_prefix" title="Permalink to this definition">#</a></dt>
 <dd><p>Adds a delimiter if the prefix is not empty.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.bottleneck_block">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">bottleneck_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">expansion_rate</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hrnet.py#L99-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.bottleneck_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">bottleneck_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">expansion_rate</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hrnet.py#L99-L137"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.bottleneck_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Creates a convolutional block with bottleneck.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.hrnet.simple_block">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">simple_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">downsampling_layer</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/hrnet.py#L61-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.simple_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.hrnet.</span></span><span class="sig-name descname"><span class="pre">simple_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">downsampling_layer</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name_prefix</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/hrnet.py#L61-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.hrnet.simple_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Creates a basic residual convolutional block.</p>
 </dd></dl>
 
@@ -483,6 +505,34 @@ <h1>sleap.nn.architectures.hrnet</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet"><code class="docutils literal notranslate"><span class="pre">HigherHRNet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.C"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.C</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.initial_downsampling_steps"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.initial_downsampling_steps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.n_deconv_modules"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.n_deconv_modules</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.bottleneck"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.bottleneck</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.deconv_filters"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.deconv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.bilinear_upsampling"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.bilinear_upsampling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.down_blocks"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.output"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.output()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.HigherHRNet.output_scale"><code class="docutils literal notranslate"><span class="pre">HigherHRNet.output_scale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.adjust_prefix"><code class="docutils literal notranslate"><span class="pre">adjust_prefix()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.bottleneck_block"><code class="docutils literal notranslate"><span class="pre">bottleneck_block()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.hrnet.simple_block"><code class="docutils literal notranslate"><span class="pre">simple_block()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -501,7 +551,7 @@ <h1>sleap.nn.architectures.hrnet</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.leap.html b/api/sleap.nn.architectures.leap.html
index e77b46f44..47172da01 100644
--- a/api/sleap.nn.architectures.leap.html
+++ b/api/sleap.nn.architectures.leap.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.leap &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.leap &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.leap';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.leap.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.nn.architectures.leap</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN"><code class="docutils literal notranslate"><span class="pre">LeapCNN</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.filters"><code class="docutils literal notranslate"><span class="pre">LeapCNN.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.filters_rate"><code class="docutils literal notranslate"><span class="pre">LeapCNN.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.down_blocks"><code class="docutils literal notranslate"><span class="pre">LeapCNN.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.down_convs_per_block"><code class="docutils literal notranslate"><span class="pre">LeapCNN.down_convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_blocks"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_interpolate"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_convs_per_block"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.decoder_stack"><code class="docutils literal notranslate"><span class="pre">LeapCNN.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.encoder_stack"><code class="docutils literal notranslate"><span class="pre">LeapCNN.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.from_config"><code class="docutils literal notranslate"><span class="pre">LeapCNN.from_config()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +338,12 @@ <h1>sleap.nn.architectures.leap</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.leap">
-<span id="sleap-nn-architectures-leap"></span><h1>sleap.nn.architectures.leap<a class="headerlink" href="#module-sleap.nn.architectures.leap" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-leap"></span><h1>sleap.nn.architectures.leap<a class="headerlink" href="#module-sleap.nn.architectures.leap" title="Permalink to this heading">#</a></h1>
 <p>This module provides a generalized implementation of the LEAP CNN.</p>
 <p>See the <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN" title="sleap.nn.architectures.leap.LeapCNN"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LeapCNN</span></code></a> class docstring for more information.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.leap.LeapCNN">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.leap.</span></span><span class="sig-name descname"><span class="pre">LeapCNN</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/leap.py#L15-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.leap.</span></span><span class="sig-name descname"><span class="pre">LeapCNN</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/leap.py#L15-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN" title="Permalink to this definition">#</a></dt>
 <dd><p>LEAP CNN from “Fast animal pose estimation using deep neural networks” (2019).</p>
 <p>This is a simple encoder-decoder style architecture without skip connections.</p>
 <p>This implementation is generalized from original paper (<a class="reference external" href="https://www.nature.com/articles/s41592-018-0234-5">Pereira et al., 2019</a>) and <a class="reference external" href="https://github.com/talmo/leap">code</a>.</p>
@@ -335,7 +355,7 @@ <h1>sleap.nn.architectures.leap</h1>
 increase the representational capacity of the network at the cost of memory
 and runtime.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -346,7 +366,7 @@ <h1>sleap.nn.architectures.leap</h1>
 <span class="sig-name descname"><span class="pre">filters_rate</span></span><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.filters_rate" title="Permalink to this definition">#</a></dt>
 <dd><p>Factor to increase the number of filters by in each block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -359,7 +379,7 @@ <h1>sleap.nn.architectures.leap</h1>
 increase the effective maximum receptive field, but may incur loss of
 spatial precision.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -372,7 +392,7 @@ <h1>sleap.nn.architectures.leap</h1>
 convolutions per block will increase the representational capacity of the
 network at the cost of memory and runtime.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -385,7 +405,7 @@ <h1>sleap.nn.architectures.leap</h1>
 <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN.down_blocks" title="sleap.nn.architectures.leap.LeapCNN.down_blocks"><code class="xref py py-obj docutils literal notranslate"><span class="pre">down_blocks</span></code></a>, the output of this network will be at the same stride (scale)
 as the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -401,7 +421,7 @@ <h1>sleap.nn.architectures.leap</h1>
 number of filters are determined by <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN.filters" title="sleap.nn.architectures.leap.LeapCNN.filters"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters</span></code></a> and <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN.filters_rate" title="sleap.nn.architectures.leap.LeapCNN.filters_rate"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters_rate</span></code></a> to
 progressively decrease the number of filters at each step.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -414,7 +434,7 @@ <h1>sleap.nn.architectures.leap</h1>
 operation. These will use the <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN.filters" title="sleap.nn.architectures.leap.LeapCNN.filters"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters</span></code></a> and <a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN.filters_rate" title="sleap.nn.architectures.leap.LeapCNN.filters_rate"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters_rate</span></code></a> to progressively
 decrease the number of filters at each step.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -422,25 +442,25 @@ <h1>sleap.nn.architectures.leap</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.leap.LeapCNN.decoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.decoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><span class="pre">SimpleUpsamplingBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.decoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the decoder block configuration.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.leap.LeapCNN.encoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.encoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.encoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the encoder block configuration.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.leap.LeapCNN.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.LEAPConfig" title="sleap.nn.config.model.LEAPConfig"><span class="pre">sleap.nn.config.model.LEAPConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN" title="sleap.nn.architectures.leap.LeapCNN"><span class="pre">sleap.nn.architectures.leap.LeapCNN</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/leap.py#L109-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.LEAPConfig" title="sleap.nn.config.model.LEAPConfig"><span class="pre">LEAPConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.leap.LeapCNN" title="sleap.nn.architectures.leap.LeapCNN"><span class="pre">LeapCNN</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/leap.py#L109-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.leap.LeapCNN.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a model from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">LEAPConfig</span></code> instance with the desired parameters.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class with the specified configuration.</p>
 </dd>
 </dl>
@@ -472,6 +492,32 @@ <h1>sleap.nn.architectures.leap</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN"><code class="docutils literal notranslate"><span class="pre">LeapCNN</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.filters"><code class="docutils literal notranslate"><span class="pre">LeapCNN.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.filters_rate"><code class="docutils literal notranslate"><span class="pre">LeapCNN.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.down_blocks"><code class="docutils literal notranslate"><span class="pre">LeapCNN.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.down_convs_per_block"><code class="docutils literal notranslate"><span class="pre">LeapCNN.down_convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_blocks"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_interpolate"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.up_convs_per_block"><code class="docutils literal notranslate"><span class="pre">LeapCNN.up_convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.decoder_stack"><code class="docutils literal notranslate"><span class="pre">LeapCNN.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.encoder_stack"><code class="docutils literal notranslate"><span class="pre">LeapCNN.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.leap.LeapCNN.from_config"><code class="docutils literal notranslate"><span class="pre">LeapCNN.from_config()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -490,7 +536,7 @@ <h1>sleap.nn.architectures.leap</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.pretrained_encoders.html b/api/sleap.nn.architectures.pretrained_encoders.html
index 913512bbe..d17f4aada 100644
--- a/api/sleap.nn.architectures.pretrained_encoders.html
+++ b/api/sleap.nn.architectures.pretrained_encoders.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.pretrained_encoders &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.pretrained_encoders &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.pretrained_encoders';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.pretrained_encoders.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,25 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.encoder"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.encoder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.decoder_filters"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.decoder_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.pretrained"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.down_blocks"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.maximum_stride"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.output_stride"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.up_blocks"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.up_blocks</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +337,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.pretrained_encoders">
-<span id="sleap-nn-architectures-pretrained-encoders"></span><h1>sleap.nn.architectures.pretrained_encoders<a class="headerlink" href="#module-sleap.nn.architectures.pretrained_encoders" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-pretrained-encoders"></span><h1>sleap.nn.architectures.pretrained_encoders<a class="headerlink" href="#module-sleap.nn.architectures.pretrained_encoders" title="Permalink to this heading">#</a></h1>
 <p>Encoder-decoder backbones with pretrained encoder models.</p>
 <p>This module enables encoder-decoder style architectures using a wide range of
 state-of-the-art architectures as the encoder, with a UNet-like upsampling stack that
@@ -354,7 +373,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 THE SOFTWARE.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.pretrained_encoders.</span></span><span class="sig-name descname"><span class="pre">UnetPretrainedEncoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">encoder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'efficientnetb0'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">(256,</span> <span class="pre">256,</span> <span class="pre">128,</span> <span class="pre">128)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/pretrained_encoders.py#L98-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.pretrained_encoders.</span></span><span class="sig-name descname"><span class="pre">UnetPretrainedEncoder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">encoder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'efficientnetb0'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">(256,</span> <span class="pre">256,</span> <span class="pre">128,</span> <span class="pre">128)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/pretrained_encoders.py#L98-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder" title="Permalink to this definition">#</a></dt>
 <dd><p>UNet with an (optionally) pretrained encoder model.</p>
 <p>This backbone enables the use of a variety of popular neural network architectures
 for feature extraction in the backbone. These can be used with ImageNet-pretrained
@@ -385,7 +404,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 </ul>
 <p>Defaults to <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;mobilenetv2&quot;</span></code></p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -400,7 +419,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 steps and therefore the output stride of the backbone. Specify 5 filter
 numbers to get an output stride of 1 (same size as the input). Defaults to</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Tuple[int]</p>
 </dd>
 </dl>
@@ -413,7 +432,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, the same model architecture will be used for the encoder but the
 weights will be randomly initialized.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -427,14 +446,14 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.PretrainedEncoderConfig" title="sleap.nn.config.model.PretrainedEncoderConfig"><span class="pre">sleap.nn.config.model.PretrainedEncoderConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder" title="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"><span class="pre">sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/pretrained_encoders.py#L142-L163"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.PretrainedEncoderConfig" title="sleap.nn.config.model.PretrainedEncoderConfig"><span class="pre">PretrainedEncoderConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder" title="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"><span class="pre">UnetPretrainedEncoder</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/pretrained_encoders.py#L142-L163"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the backbone from a configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">PretrainedEncoderConfig</span></code> instance specifying the
 configuration of the backbone.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instantiated <a class="reference internal" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder" title="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"><code class="xref py py-obj docutils literal notranslate"><span class="pre">UnetPretrainedEncoder</span></code></a>.</p>
 </dd>
 </dl>
@@ -442,15 +461,15 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone">
-<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/pretrained_encoders.py#L185-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/pretrained_encoders.py#L185-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the backbone and return the output tensors for building a model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>x_in</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> representing the input to this backbone. This is
 typically an instance of <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.layers.Input()</span></code> but can also be any
 rank-4 tensor. Can be grayscale or RGB.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (<code class="xref py py-obj docutils literal notranslate"><span class="pre">x_main</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">intermediate_activations</span></code>).</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">x_main</span></code> is the output tensor from the last upsampling block.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">intermediate_activations</span></code> is a list of <a href="#id1"><span class="problematic" id="id2">`</span></a>IntermediateActivation`s containing
@@ -505,6 +524,31 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.encoder"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.encoder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.decoder_filters"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.decoder_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.pretrained"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.down_blocks"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.maximum_stride"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.output_stride"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.up_blocks"><code class="docutils literal notranslate"><span class="pre">UnetPretrainedEncoder.up_blocks</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -523,7 +567,7 @@ <h1>sleap.nn.architectures.pretrained_encoders</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.resnet.html b/api/sleap.nn.architectures.resnet.html
index cada3408a..45fa140e4 100644
--- a/api/sleap.nn.architectures.resnet.html
+++ b/api/sleap.nn.architectures.resnet.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.resnet &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.resnet &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.resnet';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.resnet.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,59 @@ <h1>sleap.nn.architectures.resnet</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101"><code class="docutils literal notranslate"><span class="pre">ResNet101</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet101.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet101.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet101.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet101.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet101.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152"><code class="docutils literal notranslate"><span class="pre">ResNet152</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet152.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet152.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet152.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet152.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet152.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50"><code class="docutils literal notranslate"><span class="pre">ResNet50</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet50.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet50.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet50.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet50.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet50.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1"><code class="docutils literal notranslate"><span class="pre">ResNetv1</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.model_name"><code class="docutils literal notranslate"><span class="pre">ResNetv1.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.stack_configs"><code class="docutils literal notranslate"><span class="pre">ResNetv1.stack_configs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNetv1.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNetv1.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.frozen"><code class="docutils literal notranslate"><span class="pre">ResNetv1.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNetv1.skip_connections</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.down_blocks"><code class="docutils literal notranslate"><span class="pre">ResNetv1.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.from_config"><code class="docutils literal notranslate"><span class="pre">ResNetv1.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.make_backbone"><code class="docutils literal notranslate"><span class="pre">ResNetv1.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.maximum_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.output_scale"><code class="docutils literal notranslate"><span class="pre">ResNetv1.output_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.output_stride</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.block_v1"><code class="docutils literal notranslate"><span class="pre">block_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.imagenet_preproc_v1"><code class="docutils literal notranslate"><span class="pre">imagenet_preproc_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.make_backbone_fn"><code class="docutils literal notranslate"><span class="pre">make_backbone_fn()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.make_resnet_model"><code class="docutils literal notranslate"><span class="pre">make_resnet_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.stack_v1"><code class="docutils literal notranslate"><span class="pre">stack_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.tile_channels"><code class="docutils literal notranslate"><span class="pre">tile_channels()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +371,7 @@ <h1>sleap.nn.architectures.resnet</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.resnet">
-<span id="sleap-nn-architectures-resnet"></span><h1>sleap.nn.architectures.resnet<a class="headerlink" href="#module-sleap.nn.architectures.resnet" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-resnet"></span><h1>sleap.nn.architectures.resnet<a class="headerlink" href="#module-sleap.nn.architectures.resnet" title="Permalink to this heading">#</a></h1>
 <p>ResNet-based backbones.</p>
 <p>This module primarily generalizes the ResNet architectures for configurable output
 stride based on atrous convolutions, DeepLabv2-style (<a class="reference external" href="https://arxiv.org/abs/1606.00915">https://arxiv.org/abs/1606.00915</a>).</p>
@@ -327,7 +380,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <a class="github reference external" href="https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/keras/applications/resnet.py">tensorflow/tensorflow</a></p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNet101">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet101</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">sleap.nn.architectures.upsampling.UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L598-L648"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet101" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet101</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L598-L648"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet101" title="Permalink to this definition">#</a></dt>
 <dd><p>ResNet101 backbone.</p>
 <p>This model has a stack of 3, 4, 23 and 3 residual blocks.</p>
 <dl class="py attribute">
@@ -338,7 +391,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 the <code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingStack</span></code> documentation for more.  If not provided, the
 activations from the last backbone block will be the output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack">sleap.nn.architectures.upsampling.UpsamplingStack</a>]</p>
 </dd>
 </dl>
@@ -355,7 +408,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 convolution kernel expansion, in the same style as DeepLabv2. Valid values
 are 1, 2, 4, 8, 16 or 32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -367,7 +420,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, initialize with weights pretrained on ImageNet. If False,
 random weights will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -380,7 +433,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 for fast fine-tuning of ResNet features, but relies on having an upsampling
 stack with sufficient representational capacity to adapt the fixed features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -392,7 +445,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, form skip connections between outputs of each block
 in the ResNet backbone and the upsampling stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -406,7 +459,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNet152">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet152</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">sleap.nn.architectures.upsampling.UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L652-L702"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet152" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet152</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L652-L702"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet152" title="Permalink to this definition">#</a></dt>
 <dd><p>ResNet152 backbone.</p>
 <p>This model has a stack of 3, 4, 23 and 3 residual blocks.</p>
 <dl class="py attribute">
@@ -417,7 +470,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 the <code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingStack</span></code> documentation for more. If not provided, the
 activations from the last backbone block will be the output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack">sleap.nn.architectures.upsampling.UpsamplingStack</a>]</p>
 </dd>
 </dl>
@@ -434,7 +487,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 convolution kernel expansion, in the same style as DeepLabv2. Valid values
 are 1, 2, 4, 8, 16 or 32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -446,7 +499,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, initialize with weights pretrained on ImageNet. If False,
 random weights will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -459,7 +512,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 for fast fine-tuning of ResNet features, but relies on having an upsampling
 stack with sufficient representational capacity to adapt the fixed features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -471,7 +524,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, form skip connections between outputs of each block
 in the ResNet backbone and the upsampling stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -485,7 +538,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNet50">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet50</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">sleap.nn.architectures.upsampling.UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L544-L594"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet50" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNet50</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L544-L594"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNet50" title="Permalink to this definition">#</a></dt>
 <dd><p>ResNet50 backbone.</p>
 <p>This model has a stack of 3, 4, 6 and 3 residual blocks.</p>
 <dl class="py attribute">
@@ -496,7 +549,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 the <code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingStack</span></code> documentation for more.  If not provided, the
 activations from the last backbone block will be the output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack">sleap.nn.architectures.upsampling.UpsamplingStack</a>]</p>
 </dd>
 </dl>
@@ -513,7 +566,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 convolution kernel expansion, in the same style as DeepLabv2. Valid values
 are 1, 2, 4, 8, 16 or 32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -525,7 +578,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, initialize with weights pretrained on ImageNet. If False,
 random weights will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -538,7 +591,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 for fast fine-tuning of ResNet features, but relies on having an upsampling
 stack with sufficient representational capacity to adapt the fixed features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -550,7 +603,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, form skip connections between outputs of each block
 in the ResNet backbone and the upsampling stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -564,7 +617,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNetv1">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNetv1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Mapping</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">sleap.nn.architectures.upsampling.UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L364-L540"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">ResNetv1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Mapping</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling_stack</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">UpsamplingStack</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">features_output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">frozen</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L364-L540"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1" title="Permalink to this definition">#</a></dt>
 <dd><p>ResNetv1 backbone with configurable output stride and pretrained weights.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNetv1.model_name">
@@ -572,7 +625,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>Backbone name. Must be one of “resnet50”, “resnet101”, or
 “resnet152” if using pretrained weights.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -588,7 +641,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 may be adjusted to achieve the desired target output stride by converting
 strided convs into dilated convs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Sequence[Mapping[str, Any]]</p>
 </dd>
 </dl>
@@ -602,7 +655,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 the <code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingStack</span></code> documentation for more.  If not provided, the
 activations from the last backbone block will be the output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack">sleap.nn.architectures.upsampling.UpsamplingStack</a>]</p>
 </dd>
 </dl>
@@ -619,7 +672,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 convolution kernel expansion, in the same style as DeepLabv2. Valid values
 are 1, 2, 4, 8, 16 or 32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -631,7 +684,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, initialize with weights pretrained on ImageNet. If False,
 random weights will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -644,7 +697,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 for fast fine-tuning of ResNet features, but relies on having an upsampling
 stack with sufficient representational capacity to adapt the fixed features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -656,7 +709,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <dd><p>If True, form skip connections between outputs of each block
 in the ResNet backbone and the upsampling stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -674,13 +727,13 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNetv1.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ResNetConfig" title="sleap.nn.config.model.ResNetConfig"><span class="pre">sleap.nn.config.model.ResNetConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.resnet.ResNetv1" title="sleap.nn.architectures.resnet.ResNetv1"><span class="pre">sleap.nn.architectures.resnet.ResNetv1</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L407-L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ResNetConfig" title="sleap.nn.config.model.ResNetConfig"><span class="pre">ResNetConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.resnet.ResNetv1" title="sleap.nn.architectures.resnet.ResNetv1"><span class="pre">ResNetv1</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L407-L442"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a model from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">ResNetConfig</span></code> instance with the desired parameters.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class with the specified configuration.</p>
 </dd>
 </dl>
@@ -688,13 +741,13 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.ResNetv1.make_backbone">
-<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L467-L540"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1.make_backbone" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_backbone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L467-L540"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.ResNetv1.make_backbone" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the full backbone starting with the specified input tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>x_in</strong> – Input tensor of shape (samples, height, width, channels).</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of the final output tensor at the stride specified by the
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">upsampling_stack.features_output_stride</span></code> class attribute, and a list of
 intermediate tensors after each upsampling step.</p>
@@ -727,10 +780,10 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.block_v1">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">block_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dilation_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">conv_shortcut</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L168-L224"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.block_v1" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">block_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dilation_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">conv_shortcut</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L168-L224"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.block_v1" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a ResNetv1 residual block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – input tensor.</p></li>
 <li><p><strong>filters</strong> – integer, filters of the bottleneck layer.</p></li>
@@ -742,7 +795,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <li><p><strong>name</strong> – string, block label.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Output tensor for the residual block.</p>
 </dd>
 </dl>
@@ -750,14 +803,14 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.imagenet_preproc_v1">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">imagenet_preproc_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L341-L360"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.imagenet_preproc_v1" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">imagenet_preproc_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L341-L360"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.imagenet_preproc_v1" title="Permalink to this definition">#</a></dt>
 <dd><p>Preprocess images according to ImageNet/caffe/channels_last.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>X</strong> – Tensor of shape (samples, height, width, 3) of dtype float32 with values in
 the range [0, 1]. The channels axis is in RGB ordering.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Tensor of the same shape and dtype with channels reversed to BGR ordering and
 values scaled to [0, 255] and subtracted by the ImageNet/caffe pretrained model
 channel means (103.939, 116.779, 123.68) for BGR respectively. The effective
@@ -768,10 +821,10 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.make_backbone_fn">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">make_backbone_fn</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stack_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Mapping</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L256-L323"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.make_backbone_fn" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">make_backbone_fn</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stack_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stack_configs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Mapping</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L256-L323"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.make_backbone_fn" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a function that creates a block stack with output stride adjustments.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>stack_fn</strong> – Function that takes a tensor as the first positional argument,
 followed by any number of keyword arguments. This function will construct
@@ -787,7 +840,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 input stride.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Function that creates the backbone stacks based on the stack_configs.</p>
 <dl class="simple">
 <dt>This function will have the signature:</dt><dd><p>x_out, intermediate_feats = backbone_fn(x_in, current_stride)</p>
@@ -796,7 +849,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <p>The current stride describes the stride of the x_in input tensor.</p>
 </p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – If the desired output stride cannot be achieved.</p>
 </dd>
 </dl>
@@ -804,11 +857,11 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.make_resnet_model">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">make_resnet_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preact</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'resnet'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'imagenet'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_tensor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.engine.training.Model</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L45-L165"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.make_resnet_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">make_resnet_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preact</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'resnet'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'imagenet'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_tensor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_stride2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Model</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L45-L165"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.make_resnet_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the ResNet, ResNetV2 (TODO), and ResNeXt (TODO) architecture.</p>
 <p>Optionally loads weights pre-trained on ImageNet.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>backbone_fn</strong> – a function that returns output tensor for the
 stacked residual blocks.</p></li>
@@ -832,11 +885,11 @@ <h1>sleap.nn.architectures.resnet</h1>
 It should have exactly 3 inputs channels.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tuple of the <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> mapping input to final feature outputs, and a
 list of <a href="#id1"><span class="problematic" id="id2">`</span></a>IntermediateFeature`s from every block in the backbone.</p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – in case of invalid argument for <code class="xref py py-obj docutils literal notranslate"><span class="pre">weights</span></code>.</p>
 </dd>
 </dl>
@@ -844,10 +897,10 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.stack_v1">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">stack_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dilation_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L227-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.stack_v1" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">stack_v1</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dilation_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L227-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.stack_v1" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a set of stacked ResNetv1 residual blocks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – input tensor.</p></li>
 <li><p><strong>filters</strong> – integer, filters of the bottleneck layer in a block.</p></li>
@@ -857,7 +910,7 @@ <h1>sleap.nn.architectures.resnet</h1>
 <li><p><strong>name</strong> – string, stack label.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Output tensor for the stacked blocks.</p>
 </dd>
 </dl>
@@ -865,15 +918,15 @@ <h1>sleap.nn.architectures.resnet</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.architectures.resnet.tile_channels">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">tile_channels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/resnet.py#L326-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.tile_channels" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.resnet.</span></span><span class="sig-name descname"><span class="pre">tile_channels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/resnet.py#L326-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.resnet.tile_channels" title="Permalink to this definition">#</a></dt>
 <dd><p>Tile single channel to 3 channel tensor.</p>
 <p>This functon is useful to replicate grayscale single-channel images into 3-channel
 monochrome RGB images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>X</strong> – Tensor of shape (samples, height, width, 1).</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Tensor of shape (samples, height, width, 3) where the channels are identical.</p>
 </dd>
 </dl>
@@ -903,6 +956,65 @@ <h1>sleap.nn.architectures.resnet</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101"><code class="docutils literal notranslate"><span class="pre">ResNet101</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet101.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet101.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet101.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet101.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet101.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet101.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152"><code class="docutils literal notranslate"><span class="pre">ResNet152</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet152.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet152.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet152.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet152.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet152.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet152.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50"><code class="docutils literal notranslate"><span class="pre">ResNet50</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNet50.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNet50.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNet50.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.frozen"><code class="docutils literal notranslate"><span class="pre">ResNet50.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNet50.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNet50.skip_connections</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1"><code class="docutils literal notranslate"><span class="pre">ResNetv1</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.model_name"><code class="docutils literal notranslate"><span class="pre">ResNetv1.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.stack_configs"><code class="docutils literal notranslate"><span class="pre">ResNetv1.stack_configs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.upsampling_stack"><code class="docutils literal notranslate"><span class="pre">ResNetv1.upsampling_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.features_output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.features_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.pretrained"><code class="docutils literal notranslate"><span class="pre">ResNetv1.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.frozen"><code class="docutils literal notranslate"><span class="pre">ResNetv1.frozen</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.skip_connections"><code class="docutils literal notranslate"><span class="pre">ResNetv1.skip_connections</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.down_blocks"><code class="docutils literal notranslate"><span class="pre">ResNetv1.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.from_config"><code class="docutils literal notranslate"><span class="pre">ResNetv1.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.make_backbone"><code class="docutils literal notranslate"><span class="pre">ResNetv1.make_backbone()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.maximum_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.maximum_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.output_scale"><code class="docutils literal notranslate"><span class="pre">ResNetv1.output_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.ResNetv1.output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetv1.output_stride</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.block_v1"><code class="docutils literal notranslate"><span class="pre">block_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.imagenet_preproc_v1"><code class="docutils literal notranslate"><span class="pre">imagenet_preproc_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.make_backbone_fn"><code class="docutils literal notranslate"><span class="pre">make_backbone_fn()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.make_resnet_model"><code class="docutils literal notranslate"><span class="pre">make_resnet_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.stack_v1"><code class="docutils literal notranslate"><span class="pre">stack_v1()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.resnet.tile_channels"><code class="docutils literal notranslate"><span class="pre">tile_channels()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -921,7 +1033,7 @@ <h1>sleap.nn.architectures.resnet</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.unet.html b/api/sleap.nn.architectures.unet.html
index 4d99fe715..f8ef040de 100644
--- a/api/sleap.nn.architectures.unet.html
+++ b/api/sleap.nn.architectures.unet.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.unet &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.unet &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.unet';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.unet.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,37 @@ <h1>sleap.nn.architectures.unet</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock"><code class="docutils literal notranslate"><span class="pre">PoolingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.pool"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet"><code class="docutils literal notranslate"><span class="pre">UNet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.filters"><code class="docutils literal notranslate"><span class="pre">UNet.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.filters_rate"><code class="docutils literal notranslate"><span class="pre">UNet.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.kernel_size"><code class="docutils literal notranslate"><span class="pre">UNet.kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_kernel_size"><code class="docutils literal notranslate"><span class="pre">UNet.stem_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.stem_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.down_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.convs_per_block"><code class="docutils literal notranslate"><span class="pre">UNet.convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.up_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.middle_block"><code class="docutils literal notranslate"><span class="pre">UNet.middle_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.up_interpolate"><code class="docutils literal notranslate"><span class="pre">UNet.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.block_contraction"><code class="docutils literal notranslate"><span class="pre">UNet.block_contraction</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.decoder_stack"><code class="docutils literal notranslate"><span class="pre">UNet.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.encoder_stack"><code class="docutils literal notranslate"><span class="pre">UNet.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.from_config"><code class="docutils literal notranslate"><span class="pre">UNet.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_stack"><code class="docutils literal notranslate"><span class="pre">UNet.stem_stack</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +349,12 @@ <h1>sleap.nn.architectures.unet</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.unet">
-<span id="sleap-nn-architectures-unet"></span><h1>sleap.nn.architectures.unet<a class="headerlink" href="#module-sleap.nn.architectures.unet" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-unet"></span><h1>sleap.nn.architectures.unet<a class="headerlink" href="#module-sleap.nn.architectures.unet" title="Permalink to this heading">#</a></h1>
 <p>This module provides a generalized implementation of UNet.</p>
 <p>See the <a class="reference internal" href="#sleap.nn.architectures.unet.UNet" title="sleap.nn.architectures.unet.UNet"><code class="xref py py-obj docutils literal notranslate"><span class="pre">UNet</span></code></a> class docstring for more information.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.PoolingBlock">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.unet.</span></span><span class="sig-name descname"><span class="pre">PoolingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/unet.py#L15-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.PoolingBlock" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.unet.</span></span><span class="sig-name descname"><span class="pre">PoolingBlock</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pooling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/unet.py#L15-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.PoolingBlock" title="Permalink to this definition">#</a></dt>
 <dd><p>Pooling-only encoder block.</p>
 <p>Used to compensate for UNet having a skip source before the pooling, so the blocks
 need to end with a conv, not the pooling layer. This is added to the end of the
@@ -334,7 +365,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">pool</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.PoolingBlock.pool" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, applies max pooling at the end of the block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -346,7 +377,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <dd><p>Stride of the max pooling operation. If 1, the output of this
 block will be at the same stride (== 1/scale) as the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -354,7 +385,7 @@ <h1>sleap.nn.architectures.unet</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.PoolingBlock.make_block">
-<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv_block'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/unet.py#L32-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.PoolingBlock.make_block" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_block</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'conv_block'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/unet.py#L32-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.PoolingBlock.make_block" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate the encoder block from an input tensor.</p>
 </dd></dl>
 
@@ -362,7 +393,7 @@ <h1>sleap.nn.architectures.unet</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.UNet">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.unet.</span></span><span class="sig-name descname"><span class="pre">UNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">middle_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_contraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/unet.py#L46-L278"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.UNet" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.unet.</span></span><span class="sig-name descname"><span class="pre">UNet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">convs_per_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">down_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">middle_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_contraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/unet.py#L46-L278"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.UNet" title="Permalink to this definition">#</a></dt>
 <dd><p>UNet encoder-decoder architecture for fully convolutional networks.</p>
 <p>This is the canonical architecture described in <a class="reference external" href="https://arxiv.org/abs/1505.04597">Ronneberger et al., 2015</a>.</p>
 <p>The default configuration with 4 down/up blocks and 64 base filters has ~34.5M
@@ -374,7 +405,7 @@ <h1>sleap.nn.architectures.unet</h1>
 increase the representational capacity of the network at the cost of memory
 and runtime.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -385,7 +416,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">filters_rate</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.filters_rate" title="Permalink to this definition">#</a></dt>
 <dd><p>Factor to increase the number of filters by in each block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -396,7 +427,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">kernel_size</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.kernel_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Size of convolutional kernels (== height == width).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -407,7 +438,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">stem_kernel_size</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.stem_kernel_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Size of convolutional kernels in stem blocks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -419,7 +450,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <dd><p>If &gt;0, will create additional “down” blocks for initial
 downsampling. These will be configured identically to the down blocks below.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -430,7 +461,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">down_blocks</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.down_blocks" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of blocks with pooling in the encoder. More down blocks will</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -444,7 +475,7 @@ <h1>sleap.nn.architectures.unet</h1>
 of memory and runtime.
 increase the effective maximum receptive field.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -457,7 +488,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <a class="reference internal" href="#sleap.nn.architectures.unet.UNet.down_blocks" title="sleap.nn.architectures.unet.UNet.down_blocks"><code class="xref py py-obj docutils literal notranslate"><span class="pre">down_blocks</span></code></a>, the output of this network will be at the same stride (scale)
 as the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -468,7 +499,7 @@ <h1>sleap.nn.architectures.unet</h1>
 <span class="sig-name descname"><span class="pre">middle_block</span></span><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.middle_block" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, add an additional block at the end of the encoder.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -484,7 +515,7 @@ <h1>sleap.nn.architectures.unet</h1>
 number of filters are determined by <a class="reference internal" href="#sleap.nn.architectures.unet.UNet.filters" title="sleap.nn.architectures.unet.UNet.filters"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters</span></code></a> and <a class="reference internal" href="#sleap.nn.architectures.unet.UNet.filters_rate" title="sleap.nn.architectures.unet.UNet.filters_rate"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters_rate</span></code></a> to
 progressively decrease the number of filters at each step.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -498,7 +529,7 @@ <h1>sleap.nn.architectures.unet</h1>
 bottleneck before each upsampling step. The original implementation does not
 do this, but the CARE implementation does.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -514,25 +545,25 @@ <h1>sleap.nn.architectures.unet</h1>
 </div>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.UNet.decoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.decoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock" title="sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"><span class="pre">SimpleUpsamplingBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.decoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define the decoder stack.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.UNet.encoder_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.encoder_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encoder_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.encoder_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define the encoder stack.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.UNet.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.UNetConfig" title="sleap.nn.config.model.UNetConfig"><span class="pre">sleap.nn.config.model.UNetConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.unet.UNet" title="sleap.nn.architectures.unet.UNet"><span class="pre">sleap.nn.architectures.unet.UNet</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/unet.py#L250-L278"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.UNetConfig" title="sleap.nn.config.model.UNetConfig"><span class="pre">UNetConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.unet.UNet" title="sleap.nn.architectures.unet.UNet"><span class="pre">UNet</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/unet.py#L250-L278"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a model from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">UNetConfig</span></code> instance with the desired parameters.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class with the specified configuration.</p>
 </dd>
 </dl>
@@ -540,7 +571,7 @@ <h1>sleap.nn.architectures.unet</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.architectures.unet.UNet.stem_stack">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">sleap.nn.architectures.encoder_decoder.SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.stem_stack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">stem_stack</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.encoder_decoder.html#sleap.nn.architectures.encoder_decoder.SimpleConvBlock" title="sleap.nn.architectures.encoder_decoder.SimpleConvBlock"><span class="pre">SimpleConvBlock</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.architectures.unet.UNet.stem_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Define the downsampling stem.</p>
 </dd></dl>
 
@@ -570,6 +601,43 @@ <h1>sleap.nn.architectures.unet</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock"><code class="docutils literal notranslate"><span class="pre">PoolingBlock</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.pool"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.pool</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.pooling_stride"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.pooling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.PoolingBlock.make_block"><code class="docutils literal notranslate"><span class="pre">PoolingBlock.make_block()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet"><code class="docutils literal notranslate"><span class="pre">UNet</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.filters"><code class="docutils literal notranslate"><span class="pre">UNet.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.filters_rate"><code class="docutils literal notranslate"><span class="pre">UNet.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.kernel_size"><code class="docutils literal notranslate"><span class="pre">UNet.kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_kernel_size"><code class="docutils literal notranslate"><span class="pre">UNet.stem_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.stem_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.down_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.down_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.convs_per_block"><code class="docutils literal notranslate"><span class="pre">UNet.convs_per_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.up_blocks"><code class="docutils literal notranslate"><span class="pre">UNet.up_blocks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.middle_block"><code class="docutils literal notranslate"><span class="pre">UNet.middle_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.up_interpolate"><code class="docutils literal notranslate"><span class="pre">UNet.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.block_contraction"><code class="docutils literal notranslate"><span class="pre">UNet.block_contraction</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.decoder_stack"><code class="docutils literal notranslate"><span class="pre">UNet.decoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.encoder_stack"><code class="docutils literal notranslate"><span class="pre">UNet.encoder_stack</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.from_config"><code class="docutils literal notranslate"><span class="pre">UNet.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.unet.UNet.stem_stack"><code class="docutils literal notranslate"><span class="pre">UNet.stem_stack</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -588,7 +656,7 @@ <h1>sleap.nn.architectures.unet</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.architectures.upsampling.html b/api/sleap.nn.architectures.upsampling.html
index 39661e2c6..899de3614 100644
--- a/api/sleap.nn.architectures.upsampling.html
+++ b/api/sleap.nn.architectures.upsampling.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.architectures.upsampling &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.architectures.upsampling &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.architectures.upsampling';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.architectures.upsampling.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,31 @@ <h1>sleap.nn.architectures.upsampling</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.output_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_skip_connection"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.make_skip_connection</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.skip_add"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.skip_add</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.from_config"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_stack"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.make_stack()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +343,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.architectures.upsampling">
-<span id="sleap-nn-architectures-upsampling"></span><h1>sleap.nn.architectures.upsampling<a class="headerlink" href="#module-sleap.nn.architectures.upsampling" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-architectures-upsampling"></span><h1>sleap.nn.architectures.upsampling<a class="headerlink" href="#module-sleap.nn.architectures.upsampling" title="Permalink to this heading">#</a></h1>
 <p>This module defines common upsampling layer stack configurations.</p>
 <dl class="simple">
 <dt>The generic upsampling stack consists of:</dt><dd><ul class="simple">
@@ -333,14 +358,14 @@ <h1>sleap.nn.architectures.upsampling</h1>
 simpler patterns like shallow or direct upsampling (e.g., DLC).</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.architectures.upsampling.UpsamplingStack">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.upsampling.</span></span><span class="sig-name descname"><span class="pre">UpsamplingStack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">make_skip_connection</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_add</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/upsampling.py#L23-L259"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.architectures.upsampling.</span></span><span class="sig-name descname"><span class="pre">UpsamplingStack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">make_skip_connection</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_add</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/upsampling.py#L23-L259"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack" title="Permalink to this definition">#</a></dt>
 <dd><p>Standard stack of upsampling layers with refinement and skip connections.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.architectures.upsampling.UpsamplingStack.output_stride">
 <span class="sig-name descname"><span class="pre">output_stride</span></span><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack.output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>The desired final stride of the output tensor of the stack.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -355,7 +380,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.output_stride" title="sleap.nn.architectures.upsampling.UpsamplingStack.output_stride"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_stride</span></code></a> directly in fewer steps. See the notes in the <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_stack" title="sleap.nn.architectures.upsampling.UpsamplingStack.make_stack"><code class="xref py py-obj docutils literal notranslate"><span class="pre">make_stack</span></code></a>
 method for examples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -367,7 +392,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <dd><p>If True, use a strided transposed convolution to perform
 learnable upsampling. If False, bilinear upsampling will be used instead.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -381,7 +406,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters_rate" title="sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters_rate"><code class="xref py py-obj docutils literal notranslate"><span class="pre">transposed_conv_filters_rate</span></code></a> at every upsampling step. No effect if
 bilinear upsampling is used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -394,7 +419,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 transposed convolution layer after each upsampling step. If set to 1, the
 number of filters won’t change. No effect if bilinear upsampling is used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -406,7 +431,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <dd><p>Size of the kernel for the transposed convolution.
 No effect if bilinear upsampling is used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -419,7 +444,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 the transposed convolution (and before the ReLU activation). No effect if
 bilinear upsampling is used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -431,7 +456,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <dd><p>If True, incoming feature tensors form skip connection
 with upsampled features. If False, no skip connection will be formed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -445,7 +470,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 a 1x1 linear conv is applied if the channel counts do no match up. If False,
 the skip connection is formed via channel-wise concatenation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -461,7 +486,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 aliasing and checkerboard effects. If 0, no additional convolutions will be
 applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -474,7 +499,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 of filters to use for the refinement convolutions in each upsampling step.
 No effect if <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs" title="sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">refine_convs</span></code></a> is 0.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -488,7 +513,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 for all convs within the same upsampling step. If set to 1, the number of
 filters won’t change. No effect if <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs" title="sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">refine_convs</span></code></a> is 0.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -501,7 +526,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 each 3x3 convolution and before the ReLU activation. No effect if
 <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs" title="sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">refine_convs</span></code></a> is 0.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -509,16 +534,16 @@ <h1>sleap.nn.architectures.upsampling</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.upsampling.UpsamplingStack.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig"><span class="pre">sleap.nn.config.model.UpsamplingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">sleap.nn.architectures.upsampling.UpsamplingStack</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/upsampling.py#L88-L116"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig"><span class="pre">UpsamplingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack" title="sleap.nn.architectures.upsampling.UpsamplingStack"><span class="pre">UpsamplingStack</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/upsampling.py#L88-L116"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a model from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingConfig</span></code> instance with the desired parameters.</p></li>
 <li><p><strong>output_stride</strong> – The desired final stride of the output tensor of the stack.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class with the specified configuration.</p>
 </dd>
 </dl>
@@ -526,10 +551,10 @@ <h1>sleap.nn.architectures.upsampling</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.architectures.upsampling.UpsamplingStack.make_stack">
-<span class="sig-name descname"><span class="pre">make_stack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_sources</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">sleap.nn.architectures.common.IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/architectures/upsampling.py#L118-L259"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_stack" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_stack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">current_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_sources</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.architectures.common.html#sleap.nn.architectures.common.IntermediateFeature" title="sleap.nn.architectures.common.IntermediateFeature"><span class="pre">IntermediateFeature</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/architectures/upsampling.py#L118-L259"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_stack" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the stack of upsampling layers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – Feature tensor at low resolution, typically from encoder/backbone.</p></li>
 <li><p><strong>current_stride</strong> – Stride of the feature tensor relative to the network input.</p></li>
@@ -540,7 +565,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
 <span class="pre">concatenation</span> <span class="pre">or</span> <span class="pre">addition,</span> <span class="pre">depending</span> <span class="pre">on</span> <span class="pre">the</span> <span class="pre">`skip_add</span></code> class attribute.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of the resulting upsampled tensor at the stride specified by the
 <a class="reference internal" href="#sleap.nn.architectures.upsampling.UpsamplingStack.output_stride" title="sleap.nn.architectures.upsampling.UpsamplingStack.output_stride"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_stride</span></code></a> class attribute, and a list of intermediate tensors after
 each upsampling step.</p>
@@ -595,6 +620,37 @@ <h1>sleap.nn.architectures.upsampling</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.output_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.upsampling_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.upsampling_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_kernel_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.transposed_conv_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_skip_connection"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.make_skip_connection</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.skip_add"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.skip_add</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.refine_convs_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.from_config"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.architectures.upsampling.UpsamplingStack.make_stack"><code class="docutils literal notranslate"><span class="pre">UpsamplingStack.make_stack()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -613,7 +669,7 @@ <h1>sleap.nn.architectures.upsampling</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.callbacks.html b/api/sleap.nn.callbacks.html
index 832dff611..8395894f4 100644
--- a/api/sleap.nn.callbacks.html
+++ b/api/sleap.nn.callbacks.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.callbacks &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.callbacks &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.callbacks';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.callbacks.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,49 @@ <h1>sleap.nn.callbacks</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.plot_fn"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.plot_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.save_folder"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.save_folder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.prefix"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.on_epoch_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.filepath"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.filepath</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.event"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.event</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_epoch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_train_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_train_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_batch_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_batch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_epoch_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_epoch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_train_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_train_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.logdir"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.logdir</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.plot_fn"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.plot_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.tag"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.tag</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.on_epoch_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ.on_batch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ.set_lr"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ.set_lr()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +361,11 @@ <h1>sleap.nn.callbacks</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.callbacks">
-<span id="sleap-nn-callbacks"></span><h1>sleap.nn.callbacks<a class="headerlink" href="#module-sleap.nn.callbacks" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-callbacks"></span><h1>sleap.nn.callbacks<a class="headerlink" href="#module-sleap.nn.callbacks" title="Permalink to this heading">#</a></h1>
 <p>Training-related tf.keras callbacks.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.MatplotlibSaver">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">MatplotlibSaver</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_folder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">matplotlib.figure.Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L243-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.MatplotlibSaver" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">MatplotlibSaver</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_folder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L243-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.MatplotlibSaver" title="Permalink to this definition">#</a></dt>
 <dd><p>Callback for saving images rendered with matplotlib during training.</p>
 <p>This is useful for saving visualizations of the training to disk. It will be called
 at the end of each epoch.</p>
@@ -359,7 +402,7 @@ <h1>sleap.nn.callbacks</h1>
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.MatplotlibSaver.on_epoch_end">
-<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L277-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.MatplotlibSaver.on_epoch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L277-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.MatplotlibSaver.on_epoch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Save figure at the end of each epoch.</p>
 </dd></dl>
 
@@ -367,7 +410,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ModelCheckpointOnEvent">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">ModelCheckpointOnEvent</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filepath</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">event</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'train_end'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L163-L194"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">ModelCheckpointOnEvent</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filepath</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">event</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'train_end'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L163-L194"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent" title="Permalink to this definition">#</a></dt>
 <dd><p>Callback for model checkpointing on a fixed event.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ModelCheckpointOnEvent.filepath">
@@ -383,19 +426,19 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end">
-<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L183-L189"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L183-L189"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of each epoch.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin">
-<span class="sig-name descname"><span class="pre">on_train_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L178-L181"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_train_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L178-L181"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the start of training.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end">
-<span class="sig-name descname"><span class="pre">on_train_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L191-L194"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_train_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L191-L194"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of training.</p>
 </dd></dl>
 
@@ -403,22 +446,22 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">ProgressReporterZMQ</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">what</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'not_set'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L69-L160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">ProgressReporterZMQ</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">what</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'not_set'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L69-L160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin">
-<span class="sig-name descname"><span class="pre">on_batch_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L103-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_batch_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L103-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin" title="Permalink to this definition">#</a></dt>
 <dd><p>A backwards compatibility alias for <code class="xref py py-obj docutils literal notranslate"><span class="pre">on_train_batch_begin</span></code>.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end">
-<span class="sig-name descname"><span class="pre">on_batch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L112-L118"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_batch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L112-L118"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>A backwards compatibility alias for <code class="xref py py-obj docutils literal notranslate"><span class="pre">on_train_batch_end</span></code>.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin">
-<span class="sig-name descname"><span class="pre">on_epoch_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L120-L133"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_epoch_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L120-L133"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the start of an epoch.
 Subclasses should override for any actions to run. This function should only
 be called during train mode.
@@ -434,7 +477,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end">
-<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L135-L149"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L135-L149"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of an epoch.
 Subclasses should override for any actions to run. This function should only
 be called during train mode.
@@ -451,7 +494,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin">
-<span class="sig-name descname"><span class="pre">on_train_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L92-L101"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_train_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L92-L101"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the beginning of training.
 Subclasses should override for any actions to run.
 # Arguments</p>
@@ -465,7 +508,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.ProgressReporterZMQ.on_train_end">
-<span class="sig-name descname"><span class="pre">on_train_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L151-L160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_train_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L151-L160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of training.
 Subclasses should override for any actions to run.
 # Arguments</p>
@@ -481,7 +524,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TensorBoardMatplotlibWriter">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">TensorBoardMatplotlibWriter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">log_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">matplotlib.figure.Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tag</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'viz'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L197-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">TensorBoardMatplotlibWriter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">log_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plot_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tag</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'viz'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L197-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter" title="Permalink to this definition">#</a></dt>
 <dd><p>Callback for writing image summaries with visualizations during training.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TensorBoardMatplotlibWriter.logdir">
@@ -503,7 +546,7 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end">
-<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L221-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_epoch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">epoch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L221-L240"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of each epoch.</p>
 </dd></dl>
 
@@ -511,16 +554,16 @@ <h1>sleap.nn.callbacks</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TrainingControllerZMQ">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">TrainingControllerZMQ</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9000'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">topic</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">poll_timeout</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L18-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.callbacks.</span></span><span class="sig-name descname"><span class="pre">TrainingControllerZMQ</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">address</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9000'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">topic</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">poll_timeout</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L18-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ" title="Permalink to this definition">#</a></dt>
 <dd><dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end">
-<span class="sig-name descname"><span class="pre">on_batch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L43-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">on_batch_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">logs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L43-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end" title="Permalink to this definition">#</a></dt>
 <dd><p>Called at the end of a training batch.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.callbacks.TrainingControllerZMQ.set_lr">
-<span class="sig-name descname"><span class="pre">set_lr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lr</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/callbacks.py#L58-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ.set_lr" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">set_lr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lr</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/callbacks.py#L58-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.callbacks.TrainingControllerZMQ.set_lr" title="Permalink to this definition">#</a></dt>
 <dd><p>Adjust the model learning rate.</p>
 <p>This is the based off of the implementation used in the native learning rate
 scheduling callbacks.</p>
@@ -552,6 +595,55 @@ <h1>sleap.nn.callbacks</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.plot_fn"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.plot_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.save_folder"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.save_folder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.prefix"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.MatplotlibSaver.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">MatplotlibSaver.on_epoch_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.filepath"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.filepath</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.event"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.event</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_epoch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_train_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end"><code class="docutils literal notranslate"><span class="pre">ModelCheckpointOnEvent.on_train_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_batch_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_batch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_epoch_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_epoch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_train_begin()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.ProgressReporterZMQ.on_train_end"><code class="docutils literal notranslate"><span class="pre">ProgressReporterZMQ.on_train_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.logdir"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.logdir</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.plot_fn"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.plot_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.tag"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.tag</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end"><code class="docutils literal notranslate"><span class="pre">TensorBoardMatplotlibWriter.on_epoch_end()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ.on_batch_end()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.callbacks.TrainingControllerZMQ.set_lr"><code class="docutils literal notranslate"><span class="pre">TrainingControllerZMQ.set_lr()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -570,7 +662,7 @@ <h1>sleap.nn.callbacks</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.data.html b/api/sleap.nn.config.data.html
index 25d84fd1f..18761550e 100644
--- a/api/sleap.nn.config.data.html
+++ b/api/sleap.nn.config.data.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.data &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.data &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.data';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.data.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,44 @@ <h1>sleap.nn.config.data</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.DataConfig"><code class="docutils literal notranslate"><span class="pre">DataConfig</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.center_on_part"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.center_on_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.crop_size"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.crop_size_detection_padding"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.crop_size_detection_padding</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig"><code class="docutils literal notranslate"><span class="pre">LabelsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.training_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.training_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_fraction"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_fraction</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.test_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.test_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.split_by_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.split_by_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.training_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.training_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.test_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.test_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.search_path_hints"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.search_path_hints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.skeletons"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.skeletons</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.ensure_rgb</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.imagenet_mode"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.imagenet_mode</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.input_scale"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.resize_and_pad_to_target"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.resize_and_pad_to_target</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.target_height"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.target_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.target_width"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.target_width</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,10 +356,10 @@ <h1>sleap.nn.config.data</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.data">
-<span id="sleap-nn-config-data"></span><h1>sleap.nn.config.data<a class="headerlink" href="#module-sleap.nn.config.data" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-data"></span><h1>sleap.nn.config.data<a class="headerlink" href="#module-sleap.nn.config.data" title="Permalink to this heading">#</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.data.DataConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">DataConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">sleap.nn.config.data.LabelsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preprocessing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">sleap.nn.config.data.PreprocessingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_cropping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">sleap.nn.config.data.InstanceCroppingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/data.py#L147-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.DataConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">DataConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">LabelsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preprocessing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">PreprocessingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_cropping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">InstanceCroppingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/data.py#L147-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.DataConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Data configuration.</p>
 <p>labels: Configuration options related to user labels for training or testing.
 preprocessing: Configuration options related to data preprocessing.
@@ -333,7 +371,7 @@ <h1>sleap.nn.config.data</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.data.InstanceCroppingConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">InstanceCroppingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">center_on_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_detection_padding</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/data.py#L118-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.InstanceCroppingConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">InstanceCroppingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">center_on_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size_detection_padding</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/data.py#L118-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.InstanceCroppingConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Instance cropping configuration.</p>
 <p>These are only used in topdown or centroid models.</p>
 <dl class="py attribute">
@@ -344,7 +382,7 @@ <h1>sleap.nn.config.data</h1>
 box. This value will be used for both topdown and centroid models. It must
 match the name of a node on the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -361,7 +399,7 @@ <h1>sleap.nn.config.data</h1>
 layer during inference. This must be divisible by the model’s max stride
 (typically 32).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -375,7 +413,7 @@ <h1>sleap.nn.config.data</h1>
 appropriate crop size from the data. No effect if the <a class="reference internal" href="#sleap.nn.config.data.InstanceCroppingConfig.crop_size" title="sleap.nn.config.data.InstanceCroppingConfig.crop_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">crop_size</span></code></a> is
 already specified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -385,7 +423,7 @@ <h1>sleap.nn.config.data</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.data.LabelsConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">LabelsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_fraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_by_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">training_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_path_hints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/data.py#L7-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.LabelsConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">LabelsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_fraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_by_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">training_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_path_hints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/data.py#L7-L57"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.LabelsConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Labels configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.data.LabelsConfig.training_labels">
@@ -393,7 +431,7 @@ <h1>sleap.nn.config.data</h1>
 <dd><p>A filepath to a saved labels file containing user labeled
 frames to use for generating the training set.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -408,7 +446,7 @@ <h1>sleap.nn.config.data</h1>
 early stopping. If not specified, the validation set will be sampled from
 the training labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -422,7 +460,7 @@ <h1>sleap.nn.config.data</h1>
 labeled frames will be left in the training set. If the <a class="reference internal" href="#sleap.nn.config.data.LabelsConfig.validation_labels" title="sleap.nn.config.data.LabelsConfig.validation_labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">validation_labels</span></code></a>
 are already specified, this has no effect.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -439,7 +477,7 @@ <h1>sleap.nn.config.data</h1>
 useful to explicitly keep track of the test set that should be used when
 multiple splits are created for training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -455,7 +493,7 @@ <h1>sleap.nn.config.data</h1>
 labels file. If splits are generated automatically (using
 <a class="reference internal" href="#sleap.nn.config.data.LabelsConfig.validation_fraction" title="sleap.nn.config.data.LabelsConfig.validation_fraction"><code class="xref py py-obj docutils literal notranslate"><span class="pre">validation_fraction</span></code></a>), the selected indices are stored below for reference.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -466,7 +504,7 @@ <h1>sleap.nn.config.data</h1>
 <span class="sig-name descname"><span class="pre">training_inds</span></span><a class="headerlink" href="#sleap.nn.config.data.LabelsConfig.training_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>List of indices of the training split labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[int]]</p>
 </dd>
 </dl>
@@ -477,7 +515,7 @@ <h1>sleap.nn.config.data</h1>
 <span class="sig-name descname"><span class="pre">validation_inds</span></span><a class="headerlink" href="#sleap.nn.config.data.LabelsConfig.validation_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>List of indices of the validation split labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[int]]</p>
 </dd>
 </dl>
@@ -488,7 +526,7 @@ <h1>sleap.nn.config.data</h1>
 <span class="sig-name descname"><span class="pre">test_inds</span></span><a class="headerlink" href="#sleap.nn.config.data.LabelsConfig.test_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>List of indices of the test split labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[int]]</p>
 </dd>
 </dl>
@@ -503,7 +541,7 @@ <h1>sleap.nn.config.data</h1>
 in the labels. This has no effect if the labels were exported as a package
 with the user labeled data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -516,7 +554,7 @@ <h1>sleap.nn.config.data</h1>
 not specified, these will be pulled out of the labels during training, but
 must be specified for inference in order to generate predicted instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton">sleap.skeleton.Skeleton</a>]</p>
 </dd>
 </dl>
@@ -526,14 +564,14 @@ <h1>sleap.nn.config.data</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.data.PreprocessingConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">PreprocessingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ensure_rgb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">imagenet_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_scaling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_and_pad_to_target</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/data.py#L61-L114"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.PreprocessingConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.data.</span></span><span class="sig-name descname"><span class="pre">PreprocessingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ensure_rgb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">imagenet_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_scaling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_and_pad_to_target</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/data.py#L61-L114"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.data.PreprocessingConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Preprocessing configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.data.PreprocessingConfig.ensure_rgb">
 <span class="sig-name descname"><span class="pre">ensure_rgb</span></span><a class="headerlink" href="#sleap.nn.config.data.PreprocessingConfig.ensure_rgb" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, converts the image to RGB if not already.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -544,7 +582,7 @@ <h1>sleap.nn.config.data</h1>
 <span class="sig-name descname"><span class="pre">ensure_grayscale</span></span><a class="headerlink" href="#sleap.nn.config.data.PreprocessingConfig.ensure_grayscale" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, converts the image to grayscale if not already.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -571,7 +609,7 @@ <h1>sleap.nn.config.data</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -598,7 +636,7 @@ <h1>sleap.nn.config.data</h1>
 cropping inputs since the crop size should already be divisible by the
 model’s max stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -612,7 +650,7 @@ <h1>sleap.nn.config.data</h1>
 mixed image dimensions (from different video resolutions). Aspect ratio is
 preserved, and padding applied (if needed) to bottom or right of image only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -624,7 +662,7 @@ <h1>sleap.nn.config.data</h1>
 <dd><p>Target image height for ‘resize_and_pad_to_target’. When not
 explicitly provided, inferred as the max image height from the dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -636,7 +674,7 @@ <h1>sleap.nn.config.data</h1>
 <dd><p>Target image width for ‘resize_and_pad_to_target’. When not
 explicitly provided, inferred as the max image width from the dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -668,6 +706,50 @@ <h1>sleap.nn.config.data</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.DataConfig"><code class="docutils literal notranslate"><span class="pre">DataConfig</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.center_on_part"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.center_on_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.crop_size"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.InstanceCroppingConfig.crop_size_detection_padding"><code class="docutils literal notranslate"><span class="pre">InstanceCroppingConfig.crop_size_detection_padding</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig"><code class="docutils literal notranslate"><span class="pre">LabelsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.training_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.training_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_fraction"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_fraction</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.test_labels"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.test_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.split_by_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.split_by_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.training_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.training_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.validation_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.validation_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.test_inds"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.test_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.search_path_hints"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.search_path_hints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.LabelsConfig.skeletons"><code class="docutils literal notranslate"><span class="pre">LabelsConfig.skeletons</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.ensure_rgb</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.imagenet_mode"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.imagenet_mode</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.input_scale"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.resize_and_pad_to_target"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.resize_and_pad_to_target</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.target_height"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.target_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.data.PreprocessingConfig.target_width"><code class="docutils literal notranslate"><span class="pre">PreprocessingConfig.target_width</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -686,7 +768,7 @@ <h1>sleap.nn.config.data</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.model.html b/api/sleap.nn.config.model.html
index 77bc450cb..e63b1a365 100644
--- a/api/sleap.nn.config.model.html
+++ b/api/sleap.nn.config.model.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.model &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.model &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.model';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.model.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,162 @@ <h1>sleap.nn.config.model</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig"><code class="docutils literal notranslate"><span class="pre">BackboneConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.leap"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.leap</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.unet"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.unet</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.hourglass"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.hourglass</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.resnet"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.resnet</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.anchor_part"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.anchor_part"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.classes"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.classes"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_layers"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.num_fc_layers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_units"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.num_fc_units</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig"><code class="docutils literal notranslate"><span class="pre">HeadsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.single_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.single_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.centroid"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.centroid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.centered_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.centered_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_class_bottomup"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_class_bottomup</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_class_topdown"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_class_topdown</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig"><code class="docutils literal notranslate"><span class="pre">HourglassConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.stem_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.filters"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.filters_increase"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.filters_increase</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.stacks"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig"><code class="docutils literal notranslate"><span class="pre">LEAPConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.filters"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.up_interpolate"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.stacks"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig"><code class="docutils literal notranslate"><span class="pre">ModelConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.backbone"><code class="docutils literal notranslate"><span class="pre">ModelConfig.backbone</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.heads"><code class="docutils literal notranslate"><span class="pre">ModelConfig.heads</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.base_checkpoint"><code class="docutils literal notranslate"><span class="pre">ModelConfig.base_checkpoint</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig.class_maps"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig.class_maps</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig.class_vectors"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig.class_vectors</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig.pafs"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig.pafs</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.edges"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.encoder"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.encoder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.pretrained"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters_rate"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_batchnorm"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_batchnorm</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig"><code class="docutils literal notranslate"><span class="pre">ResNetConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.version"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.version</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.weights"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.weights</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.upsampling"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.upsampling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.output_stride</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig"><code class="docutils literal notranslate"><span class="pre">UNetConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.stem_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.filters"><code class="docutils literal notranslate"><span class="pre">UNetConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">UNetConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.middle_block"><code class="docutils literal notranslate"><span class="pre">UNetConfig.middle_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.up_interpolate"><code class="docutils literal notranslate"><span class="pre">UNetConfig.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.stacks"><code class="docutils literal notranslate"><span class="pre">UNetConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.method"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.skip_connections"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.skip_connections</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.block_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.block_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.refine_convs"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.conv_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.conv_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.transposed_conv_kernel_size</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,10 +474,10 @@ <h1>sleap.nn.config.model</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.model">
-<span id="sleap-nn-config-model"></span><h1>sleap.nn.config.model<a class="headerlink" href="#module-sleap.nn.config.model" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-model"></span><h1>sleap.nn.config.model<a class="headerlink" href="#module-sleap.nn.config.model" title="Permalink to this heading">#</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.BackboneConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">BackboneConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L630-L646"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">BackboneConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L630-L646"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations related to the model backbone.</p>
 <p>Only one field can be set and will determine which backbone architecture to use.</p>
 <dl class="py attribute">
@@ -329,7 +485,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">leap</span></span><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig.leap" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.config.model.LEAPConfig" title="sleap.nn.config.model.LEAPConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LEAPConfig</span></code></a> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.LEAPConfig" title="sleap.nn.config.model.LEAPConfig">sleap.nn.config.model.LEAPConfig</a>]</p>
 </dd>
 </dl>
@@ -340,7 +496,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">unet</span></span><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig.unet" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.config.model.UNetConfig" title="sleap.nn.config.model.UNetConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">UNetConfig</span></code></a> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.UNetConfig" title="sleap.nn.config.model.UNetConfig">sleap.nn.config.model.UNetConfig</a>]</p>
 </dd>
 </dl>
@@ -351,7 +507,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">hourglass</span></span><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig.hourglass" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.config.model.HourglassConfig" title="sleap.nn.config.model.HourglassConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">HourglassConfig</span></code></a> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.HourglassConfig" title="sleap.nn.config.model.HourglassConfig">sleap.nn.config.model.HourglassConfig</a>]</p>
 </dd>
 </dl>
@@ -362,7 +518,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">resnet</span></span><a class="headerlink" href="#sleap.nn.config.model.BackboneConfig.resnet" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.config.model.ResNetConfig" title="sleap.nn.config.model.ResNetConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ResNetConfig</span></code></a> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.ResNetConfig" title="sleap.nn.config.model.ResNetConfig">sleap.nn.config.model.ResNetConfig</a>]</p>
 </dd>
 </dl>
@@ -372,7 +528,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L99-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L99-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for centered instance confidence map heads.</p>
 <p>These heads are used in topdown multi-instance models that make the assumption that
 there is an instance reliably centered in the cropped input image. These heads are
@@ -397,7 +553,7 @@ <h1>sleap.nn.config.model</h1>
 benefit from a consistent geometry of the body parts relative to the center
 of the image.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -410,7 +566,7 @@ <h1>sleap.nn.config.model</h1>
 to produce. The number of parts determines the number of channels in the
 output. If not specified, all body parts in the skeleton will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -426,7 +582,7 @@ <h1>sleap.nn.config.model</h1>
 is in units of pixels of the model input image, i.e., the image resolution
 after any input scaling is applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -441,7 +597,7 @@ <h1>sleap.nn.config.model</h1>
 this value can considerably speed up model performance and decrease memory
 requirements, at the cost of decreased spatial resolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -454,7 +610,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -471,7 +627,7 @@ <h1>sleap.nn.config.model</h1>
 does not require additional resources or training, but may not achieve the
 same accuracy as learned refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -481,7 +637,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.CentroidsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">CentroidsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L51-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.CentroidsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">CentroidsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L51-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.CentroidsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for centroid confidence map heads.</p>
 <p>These heads are used in topdown models that rely on centroid detection to detect
 instances for cropping before predicting the remaining body parts.</p>
@@ -498,7 +654,7 @@ <h1>sleap.nn.config.model</h1>
 benefit from a consistent geometry of the body parts relative to the center
 of the image.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -514,7 +670,7 @@ <h1>sleap.nn.config.model</h1>
 is in units of pixels of the model input image, i.e., the image resolution
 after any input scaling is applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -529,7 +685,7 @@ <h1>sleap.nn.config.model</h1>
 this value can considerably speed up model performance and decrease memory
 requirements, at the cost of decreased spatial resolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -542,7 +698,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -559,7 +715,7 @@ <h1>sleap.nn.config.model</h1>
 does not require additional resources or training, but may not achieve the
 same accuracy as learned refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -569,7 +725,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ClassMapsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ClassMapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L290-L316"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ClassMapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L290-L316"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for class map heads.</p>
 <p>These heads are used in bottom-up multi-instance models that classify detected
 points using a fixed set of learned classes (e.g., animal identities).</p>
@@ -581,7 +737,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">classes</span></span><a class="headerlink" href="#sleap.nn.config.model.ClassMapsHeadConfig.classes" title="Permalink to this definition">#</a></dt>
 <dd><p>List of string names of the classes that this head will predict.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -593,7 +749,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Spread of the Gaussian distribution that determines the neighborhood
 that the class maps will be nonzero around each landmark.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -607,7 +763,7 @@ <h1>sleap.nn.config.model</h1>
 results in maps that are 0.5x the size of the input. This should be the same
 size as the confidence maps they are associated with.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -620,7 +776,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -630,7 +786,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ClassVectorsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ClassVectorsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_layers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_units</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">global_pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L346-L377"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ClassVectorsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_layers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_units</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">global_pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L346-L377"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for class vectors heads.</p>
 <p>These heads are used in top-down multi-instance models that classify detected
 points using a fixed set of learned classes (e.g., animal identities).</p>
@@ -641,7 +797,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">classes</span></span><a class="headerlink" href="#sleap.nn.config.model.ClassVectorsHeadConfig.classes" title="Permalink to this definition">#</a></dt>
 <dd><p>List of string names of the classes that this head will predict.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -654,7 +810,7 @@ <h1>sleap.nn.config.model</h1>
 layer. These can help in transforming general image features into
 classification-specific features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -667,7 +823,7 @@ <h1>sleap.nn.config.model</h1>
 classification. Increasing this can improve the representational capacity in
 the pre-classification layers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -681,7 +837,7 @@ <h1>sleap.nn.config.model</h1>
 results in maps that are 0.5x the size of the input. This should be the same
 size as the confidence maps they are associated with.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -694,7 +850,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -704,7 +860,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.HeadsConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">HeadsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L410-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">HeadsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L410-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations related to the model output head type.</p>
 <p>Only one attribute of this class can be set, which defines the model output type.</p>
 <dl class="py attribute">
@@ -712,7 +868,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">single_instance</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.single_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig">sleap.nn.config.model.SingleInstanceConfmapsHeadConfig</a>]</p>
 </dd>
 </dl>
@@ -723,7 +879,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">centroid</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.centroid" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CentroidsHeadConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig">sleap.nn.config.model.CentroidsHeadConfig</a>]</p>
 </dd>
 </dl>
@@ -734,7 +890,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">centered_instance</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.centered_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig">sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig</a>]</p>
 </dd>
 </dl>
@@ -745,7 +901,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">multi_instance</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.multi_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfig" title="sleap.nn.config.model.MultiInstanceConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiInstanceConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfig" title="sleap.nn.config.model.MultiInstanceConfig">sleap.nn.config.model.MultiInstanceConfig</a>]</p>
 </dd>
 </dl>
@@ -756,7 +912,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">multi_class_bottomup</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.multi_class_bottomup" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.MultiClassBottomUpConfig" title="sleap.nn.config.model.MultiClassBottomUpConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.MultiClassBottomUpConfig" title="sleap.nn.config.model.MultiClassBottomUpConfig">sleap.nn.config.model.MultiClassBottomUpConfig</a>]</p>
 </dd>
 </dl>
@@ -767,7 +923,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">multi_class_topdown</span></span><a class="headerlink" href="#sleap.nn.config.model.HeadsConfig.multi_class_topdown" title="Permalink to this definition">#</a></dt>
 <dd><p>An instance of <a class="reference internal" href="#sleap.nn.config.model.MultiClassTopDownConfig" title="sleap.nn.config.model.MultiClassTopDownConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiClassTopDownConfig</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.MultiClassTopDownConfig" title="sleap.nn.config.model.MultiClassTopDownConfig">sleap.nn.config.model.MultiClassTopDownConfig</a>]</p>
 </dd>
 </dl>
@@ -777,7 +933,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.HourglassConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">HourglassConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter_increase</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L489-L510"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.HourglassConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">HourglassConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stem_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filter_increase</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L489-L510"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.HourglassConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Hourglass backbone configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.HourglassConfig.stem_stride">
@@ -786,7 +942,7 @@ <h1>sleap.nn.config.model</h1>
 These are useful for learned downsampling that is able to retain spatial
 information while reducing large input image sizes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -798,7 +954,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Determines the number of downsampling blocks in the network,
 increasing receptive field size at the cost of network size.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -809,7 +965,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">output_stride</span></span><a class="headerlink" href="#sleap.nn.config.model.HourglassConfig.output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Determines the number of upsampling blocks in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -820,7 +976,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.config.model.HourglassConfig.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Base number of filters in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -837,7 +993,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">stacks</span></span><a class="headerlink" href="#sleap.nn.config.model.HourglassConfig.stacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of repeated stacks of the network (excluding the stem).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -847,7 +1003,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.LEAPConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">LEAPConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L433-L453"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">LEAPConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L433-L453"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>LEAP backbone configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.LEAPConfig.max_stride">
@@ -855,7 +1011,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Determines the number of downsampling blocks in the network,
 increasing receptive field size at the cost of network size.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -866,7 +1022,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">output_stride</span></span><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig.output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Determines the number of upsampling blocks in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -877,7 +1033,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Base number of filters in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -888,7 +1044,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">filters_rate</span></span><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig.filters_rate" title="Permalink to this definition">#</a></dt>
 <dd><p>Factor to scale the number of filters by at each block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -901,7 +1057,7 @@ <h1>sleap.nn.config.model</h1>
 convolutions for upsampling. This can save computations but may lower
 overall accuracy.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -912,7 +1068,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">stacks</span></span><a class="headerlink" href="#sleap.nn.config.model.LEAPConfig.stacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of repeated stacks of the network (excluding the stem).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -922,14 +1078,14 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ModelConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ModelConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.BackboneConfig" title="sleap.nn.config.model.BackboneConfig"><span class="pre">sleap.nn.config.model.BackboneConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">heads</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.HeadsConfig" title="sleap.nn.config.model.HeadsConfig"><span class="pre">sleap.nn.config.model.HeadsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">base_checkpoint</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L650-L661"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ModelConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ModelConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.BackboneConfig" title="sleap.nn.config.model.BackboneConfig"><span class="pre">BackboneConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">heads</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.HeadsConfig" title="sleap.nn.config.model.HeadsConfig"><span class="pre">HeadsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">base_checkpoint</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L650-L661"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ModelConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations related to model architecture.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ModelConfig.backbone">
 <span class="sig-name descname"><span class="pre">backbone</span></span><a class="headerlink" href="#sleap.nn.config.model.ModelConfig.backbone" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations related to the main network architecture.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.BackboneConfig" title="sleap.nn.config.model.BackboneConfig">sleap.nn.config.model.BackboneConfig</a></p>
 </dd>
 </dl>
@@ -940,7 +1096,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">heads</span></span><a class="headerlink" href="#sleap.nn.config.model.ModelConfig.heads" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations related to the output heads.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.HeadsConfig" title="sleap.nn.config.model.HeadsConfig">sleap.nn.config.model.HeadsConfig</a></p>
 </dd>
 </dl>
@@ -951,7 +1107,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">base_checkpoint</span></span><a class="headerlink" href="#sleap.nn.config.model.ModelConfig.base_checkpoint" title="Permalink to this definition">#</a></dt>
 <dd><p>Path to model folder for loading a checkpoint. Should contain the .h5 file</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -961,7 +1117,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.MultiClassBottomUpConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiClassBottomUpConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig"><span class="pre">sleap.nn.config.model.ClassMapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L320-L342"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiClassBottomUpConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiClassBottomUpConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">MultiInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig"><span class="pre">ClassMapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L320-L342"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiClassBottomUpConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for multi-instance confidence map and class map models.</p>
 <p>This configuration specifies a multi-head model that outputs both multi-instance
 confidence maps and class maps, which together enable multi-instance pose tracking
@@ -977,7 +1133,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Part confidence map configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</a></p>
 </dd>
 </dl>
@@ -989,7 +1145,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Class map configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ClassMapsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig">sleap.nn.config.model.ClassMapsHeadConfig</a></p>
 </dd>
 </dl>
@@ -999,7 +1155,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.MultiClassTopDownConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiClassTopDownConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_vectors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig"><span class="pre">sleap.nn.config.model.ClassVectorsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L381-L405"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiClassTopDownConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiClassTopDownConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_vectors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig"><span class="pre">ClassVectorsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L381-L405"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiClassTopDownConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for centered-instance confidence map and class map models.</p>
 <p>This configuration specifies a multi-head model that outputs both centered-instance
 confidence maps and class vectors, which together enable multi-instance pose
@@ -1016,7 +1172,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Part confidence map configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig">sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig</a></p>
 </dd>
 </dl>
@@ -1028,7 +1184,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Class map configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig">sleap.nn.config.model.ClassVectorsHeadConfig</a></p>
 </dd>
 </dl>
@@ -1038,7 +1194,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.MultiInstanceConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig"><span class="pre">sleap.nn.config.model.PartAffinityFieldsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L268-L286"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiInstanceConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">MultiInstanceConfmapsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig"><span class="pre">PartAffinityFieldsHeadConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L268-L286"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiInstanceConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for combined multi-instance confidence map and PAF model heads.</p>
 <p>This configuration specifies a multi-head model that outputs both multi-instance
 confidence maps and part affinity fields, which together enable multi-instance pose
@@ -1050,7 +1206,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Part confidence map configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</a></p>
 </dd>
 </dl>
@@ -1062,7 +1218,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Part affinity fields configuration (see the description in
 <a class="reference internal" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig">sleap.nn.config.model.PartAffinityFieldsHeadConfig</a></p>
 </dd>
 </dl>
@@ -1072,7 +1228,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L161-L213"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L161-L213"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for multi-instance confidence map heads.</p>
 <p>These heads are used in bottom-up multi-instance models that do not make any
 assumption about the connectivity of the body parts. These heads will generate
@@ -1095,7 +1251,7 @@ <h1>sleap.nn.config.model</h1>
 to produce. The number of parts determines the number of channels in the
 output. If not specified, all body parts in the skeleton will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -1111,7 +1267,7 @@ <h1>sleap.nn.config.model</h1>
 is in units of pixels of the model input image, i.e., the image resolution
 after any input scaling is applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1126,7 +1282,7 @@ <h1>sleap.nn.config.model</h1>
 this value can considerably speed up model performance and decrease memory
 requirements, at the cost of decreased spatial resolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1139,7 +1295,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1156,7 +1312,7 @@ <h1>sleap.nn.config.model</h1>
 does not require additional resources or training, but may not achieve the
 same accuracy as learned refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1166,7 +1322,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.PartAffinityFieldsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L217-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L217-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for multi-instance part affinity field heads.</p>
 <p>These heads are used in bottom-up multi-instance models that require information
 about body part connectivity in order to group multiple detections of each body part
@@ -1190,7 +1346,7 @@ <h1>sleap.nn.config.model</h1>
 define pairs of text names of the directed edges of the graph. If not set,
 all edges in the skeleton will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[Sequence[Tuple[str, str]]]</p>
 </dd>
 </dl>
@@ -1208,7 +1364,7 @@ <h1>sleap.nn.config.model</h1>
 spread is in units of pixels of the model input image, i.e., the image
 resolution after any input scaling is applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1223,7 +1379,7 @@ <h1>sleap.nn.config.model</h1>
 this value can considerably speed up model performance and decrease memory
 requirements, at the cost of decreased spatial resolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1236,7 +1392,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1246,7 +1402,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.PretrainedEncoderConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">PretrainedEncoderConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">encoder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'efficientnetb0'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L593-L625"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.PretrainedEncoderConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">PretrainedEncoderConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">encoder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'efficientnetb0'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pretrained</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">decoder_batchnorm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L593-L625"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.PretrainedEncoderConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for UNet backbone with pretrained encoder.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.PretrainedEncoderConfig.encoder">
@@ -1269,7 +1425,7 @@ <h1>sleap.nn.config.model</h1>
 </ul>
 <p>Defaults to <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;efficientnetb0&quot;</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1280,7 +1436,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">pretrained</span></span><a class="headerlink" href="#sleap.nn.config.model.PretrainedEncoderConfig.pretrained" title="Permalink to this definition">#</a></dt>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, use initialized with weights pretrained on ImageNet.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1292,7 +1448,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Base number of filters for the upsampling blocks in the
 decoder.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1304,7 +1460,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Factor to scale the number of filters by at each
 consecutive upsampling block in the decoder.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1315,7 +1471,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">output_stride</span></span><a class="headerlink" href="#sleap.nn.config.model.PretrainedEncoderConfig.output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Stride of the final output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1327,7 +1483,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), use batch normalization in the
 decoder layers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1337,7 +1493,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ResNetConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ResNetConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">version</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'ResNet50'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'frozen'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig"><span class="pre">sleap.nn.config.model.UpsamplingConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">32</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L563-L589"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ResNetConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">ResNetConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">version</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'ResNet50'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'frozen'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">upsampling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig"><span class="pre">UpsamplingConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">32</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L563-L589"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.ResNetConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>ResNet backbone configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.ResNetConfig.version">
@@ -1345,7 +1501,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Name of the ResNetV1 variant. Can be one of: “ResNet50”, “ResNet101”,
 or “ResNet152”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1359,7 +1515,7 @@ <h1>sleap.nn.config.model</h1>
 and keeps them fixed. If “tunable”, the network uses pretrained weights and
 allows them to be trainable.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1370,7 +1526,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">upsampling</span></span><a class="headerlink" href="#sleap.nn.config.model.ResNetConfig.upsampling" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">UpsamplingConfig</span></code></a> that defines an upsampling branch if not None.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.config.model.UpsamplingConfig" title="sleap.nn.config.model.UpsamplingConfig">sleap.nn.config.model.UpsamplingConfig</a>]</p>
 </dd>
 </dl>
@@ -1381,7 +1537,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">max_stride</span></span><a class="headerlink" href="#sleap.nn.config.model.ResNetConfig.max_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Stride of the backbone feature activations. These should be &lt;= 32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1394,7 +1550,7 @@ <h1>sleap.nn.config.model</h1>
 defined, the output stride is controlled via dilated convolutions or reduced
 pooling in the backbone.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1404,7 +1560,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L7-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsHeadConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L7-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configurations for single instance confidence map heads.</p>
 <p>These heads are used in single instance models that make the assumption that only
 one of each body part is present in the image. These heads produce confidence maps
@@ -1417,7 +1573,7 @@ <h1>sleap.nn.config.model</h1>
 to produce. The number of parts determines the number of channels in the
 output. If not specified, all body parts in the skeleton will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -1433,7 +1589,7 @@ <h1>sleap.nn.config.model</h1>
 is in units of pixels of the model input image, i.e., the image resolution
 after any input scaling is applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1448,7 +1604,7 @@ <h1>sleap.nn.config.model</h1>
 this value can considerably speed up model performance and decrease memory
 requirements, at the cost of decreased spatial resolution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1461,7 +1617,7 @@ <h1>sleap.nn.config.model</h1>
 training. Increase this to encourage the optimization to focus on improving
 this specific output in multi-head models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1478,7 +1634,7 @@ <h1>sleap.nn.config.model</h1>
 does not require additional resources or training, but may not achieve the
 same accuracy as learned refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1488,7 +1644,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.UNetConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">UNetConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">middle_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L457-L485"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.UNetConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">UNetConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stem_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">16</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">middle_block</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_interpolate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stacks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L457-L485"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.UNetConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>UNet backbone configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.UNetConfig.stem_stride">
@@ -1497,7 +1653,7 @@ <h1>sleap.nn.config.model</h1>
 downsampling. These are useful for learned downsampling that is able to
 retain spatial information while reducing large input image sizes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -1509,7 +1665,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Determines the number of downsampling blocks in the network,
 increasing receptive field size at the cost of network size.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1520,7 +1676,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">output_stride</span></span><a class="headerlink" href="#sleap.nn.config.model.UNetConfig.output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Determines the number of upsampling blocks in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1531,7 +1687,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">filters</span></span><a class="headerlink" href="#sleap.nn.config.model.UNetConfig.filters" title="Permalink to this definition">#</a></dt>
 <dd><p>Base number of filters in the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1542,7 +1698,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">filters_rate</span></span><a class="headerlink" href="#sleap.nn.config.model.UNetConfig.filters_rate" title="Permalink to this definition">#</a></dt>
 <dd><p>Factor to scale the number of filters by at each block.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1555,7 +1711,7 @@ <h1>sleap.nn.config.model</h1>
 upsampling branch for additional processing for features at the largest
 receptive field size. This will not introduce an extra pooling step.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1568,7 +1724,7 @@ <h1>sleap.nn.config.model</h1>
 convolutions for upsampling. This can save computations but may lower
 overall accuracy.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1579,7 +1735,7 @@ <h1>sleap.nn.config.model</h1>
 <span class="sig-name descname"><span class="pre">stacks</span></span><a class="headerlink" href="#sleap.nn.config.model.UNetConfig.stacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of repeated stacks of the network (excluding the stem).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1589,7 +1745,7 @@ <h1>sleap.nn.config.model</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.model.UpsamplingConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">UpsamplingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'interpolation'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/model.py#L514-L559"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.UpsamplingConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.model.</span></span><span class="sig-name descname"><span class="pre">UpsamplingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'interpolation'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skip_connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filters_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refine_convs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_norm</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transposed_conv_kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/model.py#L514-L559"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.model.UpsamplingConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Upsampling stack configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.model.UpsamplingConfig.method">
@@ -1598,7 +1754,7 @@ <h1>sleap.nn.config.model</h1>
 learnable upsampling. If “interpolation”, bilinear upsampling will be used
 instead.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1613,7 +1769,7 @@ <h1>sleap.nn.config.model</h1>
 up. If “concatenate”, the skip connection is formed via channel-wise
 concatenation. If None, skip connections will not be formed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -1627,7 +1783,7 @@ <h1>sleap.nn.config.model</h1>
 upsampling step, but can be set higher to upsample to the desired
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">output_stride</span></code> directly in fewer steps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1639,7 +1795,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Integer that specifies the base number of filters in each convolution
 layer. This will be scaled by the <a class="reference internal" href="#sleap.nn.config.model.UpsamplingConfig.filters_rate" title="sleap.nn.config.model.UpsamplingConfig.filters_rate"><code class="xref py py-obj docutils literal notranslate"><span class="pre">filters_rate</span></code></a> at every upsampling step.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1651,7 +1807,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Factor to scale the number of filters in the convolution layers
 after each upsampling step. If set to 1, the number of filters won’t change.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1667,7 +1823,7 @@ <h1>sleap.nn.config.model</h1>
 aliasing and checkerboard effects. If 0, no additional convolutions will be
 applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1686,7 +1842,7 @@ <h1>sleap.nn.config.model</h1>
 <dd><p>Size of the kernel for the transposed convolution.
 No effect if bilinear upsampling is used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1718,6 +1874,168 @@ <h1>sleap.nn.config.model</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig"><code class="docutils literal notranslate"><span class="pre">BackboneConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.leap"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.leap</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.unet"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.unet</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.hourglass"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.hourglass</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.BackboneConfig.resnet"><code class="docutils literal notranslate"><span class="pre">BackboneConfig.resnet</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.anchor_part"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.anchor_part"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.CentroidsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">CentroidsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.classes"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassMapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassMapsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.classes"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_layers"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.num_fc_layers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_units"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.num_fc_units</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ClassVectorsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig"><code class="docutils literal notranslate"><span class="pre">HeadsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.single_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.single_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.centroid"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.centroid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.centered_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.centered_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_instance"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_class_bottomup"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_class_bottomup</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HeadsConfig.multi_class_topdown"><code class="docutils literal notranslate"><span class="pre">HeadsConfig.multi_class_topdown</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig"><code class="docutils literal notranslate"><span class="pre">HourglassConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.stem_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.filters"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.filters_increase"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.filters_increase</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.HourglassConfig.stacks"><code class="docutils literal notranslate"><span class="pre">HourglassConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig"><code class="docutils literal notranslate"><span class="pre">LEAPConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.filters"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.up_interpolate"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.LEAPConfig.stacks"><code class="docutils literal notranslate"><span class="pre">LEAPConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig"><code class="docutils literal notranslate"><span class="pre">ModelConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.backbone"><code class="docutils literal notranslate"><span class="pre">ModelConfig.backbone</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.heads"><code class="docutils literal notranslate"><span class="pre">ModelConfig.heads</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ModelConfig.base_checkpoint"><code class="docutils literal notranslate"><span class="pre">ModelConfig.base_checkpoint</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassBottomUpConfig.class_maps"><code class="docutils literal notranslate"><span class="pre">MultiClassBottomUpConfig.class_maps</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiClassTopDownConfig.class_vectors"><code class="docutils literal notranslate"><span class="pre">MultiClassTopDownConfig.class_vectors</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig.confmaps"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig.confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfig.pafs"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfig.pafs</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.edges"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PartAffinityFieldsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHeadConfig.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.encoder"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.encoder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.pretrained"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.pretrained</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters_rate"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.PretrainedEncoderConfig.decoder_batchnorm"><code class="docutils literal notranslate"><span class="pre">PretrainedEncoderConfig.decoder_batchnorm</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig"><code class="docutils literal notranslate"><span class="pre">ResNetConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.version"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.version</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.weights"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.weights</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.upsampling"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.upsampling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.ResNetConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">ResNetConfig.output_stride</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.part_names"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.loss_weight"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.offset_refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHeadConfig.offset_refinement</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig"><code class="docutils literal notranslate"><span class="pre">UNetConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.stem_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.stem_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.max_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.max_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.output_stride"><code class="docutils literal notranslate"><span class="pre">UNetConfig.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.filters"><code class="docutils literal notranslate"><span class="pre">UNetConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">UNetConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.middle_block"><code class="docutils literal notranslate"><span class="pre">UNetConfig.middle_block</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.up_interpolate"><code class="docutils literal notranslate"><span class="pre">UNetConfig.up_interpolate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UNetConfig.stacks"><code class="docutils literal notranslate"><span class="pre">UNetConfig.stacks</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.method"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.method</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.skip_connections"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.skip_connections</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.block_stride"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.block_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.filters"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.filters</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.filters_rate"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.filters_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.refine_convs"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.refine_convs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.conv_batchnorm"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.conv_batchnorm</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.model.UpsamplingConfig.transposed_conv_kernel_size"><code class="docutils literal notranslate"><span class="pre">UpsamplingConfig.transposed_conv_kernel_size</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1736,7 +2054,7 @@ <h1>sleap.nn.config.model</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.optimization.html b/api/sleap.nn.config.optimization.html
index d1afb9d96..71b441879 100644
--- a/api/sleap.nn.config.optimization.html
+++ b/api/sleap.nn.config.optimization.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.optimization &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.optimization &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.optimization';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.optimization.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,84 @@ <h1>sleap.nn.config.optimization</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotate"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_min_angle"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotation_min_angle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_max_angle"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotation_max_angle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate_min"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate_min</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate_max"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate_max</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale_min"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale_min</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale_max"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale_max</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_min_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise_min_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_max_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise_max_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_mean"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise_mean</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_stddev"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise_stddev</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast_min_gamma"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast_min_gamma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast_max_gamma"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast_max_gamma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_min_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness_min_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_max_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness_max_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_width"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_height"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_flip"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_flip</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.flip_horizontal"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.flip_horizontal</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.stop_training_on_plateau"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.stop_training_on_plateau</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.plateau_min_delta"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.plateau_min_delta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.plateau_patience"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.plateau_patience</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.online_mining"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.online_mining</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.hard_to_easy_ratio"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.hard_to_easy_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.min_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.min_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.max_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.max_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.loss_scale"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.loss_scale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.reduce_on_plateau"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.reduce_on_plateau</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.reduction_factor"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.reduction_factor</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_min_delta"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_min_delta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_patience"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_patience</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_cooldown"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_cooldown</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.min_learning_rate"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.min_learning_rate</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.preload_data"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.preload_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.augmentation_config"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.augmentation_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.online_shuffling"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.online_shuffling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.shuffle_buffer_size"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.shuffle_buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.prefetch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.prefetch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.batch_size"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.min_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.min_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.val_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.val_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.min_val_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.min_val_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.epochs"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.epochs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.optimizer"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.optimizer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.initial_learning_rate"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.initial_learning_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.learning_rate_schedule"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.learning_rate_schedule</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.hard_keypoint_mining"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.hard_keypoint_mining</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.early_stopping"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.early_stopping</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,10 +396,10 @@ <h1>sleap.nn.config.optimization</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.optimization">
-<span id="sleap-nn-config-optimization"></span><h1>sleap.nn.config.optimization<a class="headerlink" href="#module-sleap.nn.config.optimization" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-optimization"></span><h1>sleap.nn.config.optimization<a class="headerlink" href="#module-sleap.nn.config.optimization" title="Permalink to this heading">#</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.AugmentationConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">AugmentationConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">rotate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotation_min_angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">180</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotation_max_angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">180</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate_min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate_max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.9</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise_min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise_max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise_mean</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise_stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast_min_gamma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast_max_gamma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness_min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness_max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_flip</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flip_horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/optimization.py#L6-L87"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">AugmentationConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">rotate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotation_min_angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-180</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rotation_max_angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">180</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate_min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translate_max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.9</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise_min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uniform_noise_max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise_mean</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gaussian_noise_stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast_min_gamma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contrast_max_gamma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness_min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">brightness_max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">random_flip</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flip_horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/optimization.py#L6-L87"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Parameters for configuring an augmentation stack.</p>
 <p>The augmentations will be applied in the the order of the attributes.</p>
 <dl class="py attribute">
@@ -330,7 +408,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>If True, rotational augmentation will be applied. Rotation is relative
 to the center of the image. See <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.geometric.Affine</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -341,7 +419,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">rotation_min_angle</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_min_angle" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum rotation angle in degrees in [-180, 180].</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -352,7 +430,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">rotation_max_angle</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_max_angle" title="Permalink to this definition">#</a></dt>
 <dd><p>Maximum rotation angle in degrees in [-180, 180].</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -365,7 +443,7 @@ <h1>sleap.nn.config.optimization</h1>
 sampled independently for x and y coordinates. See
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.geometric.Affine</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -376,7 +454,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">translate_min</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.translate_min" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum translation in integer pixel units.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -387,7 +465,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">translate_max</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.translate_max" title="Permalink to this definition">#</a></dt>
 <dd><p>Maximum translation in integer pixel units.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -399,7 +477,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>If True, scaling augmentation will be applied. See
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.geometric.Affine</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -410,7 +488,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">scale_min</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.scale_min" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum scaling factor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -421,7 +499,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">scale_max</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.scale_max" title="Permalink to this definition">#</a></dt>
 <dd><p>Maximum scaling factor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -434,7 +512,7 @@ <h1>sleap.nn.config.optimization</h1>
 This is effectively adding a different random value to each pixel to
 simulate shot noise. See <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.arithmetic.AddElementwise</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -445,7 +523,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">uniform_noise_min_val</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_min_val" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum value to add.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -456,7 +534,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">uniform_noise_max_val</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_max_val" title="Permalink to this definition">#</a></dt>
 <dd><p>Maximum value to add.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -470,7 +548,7 @@ <h1>sleap.nn.config.optimization</h1>
 mean noise magnitude. This is applied independently to each pixel.
 See <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.arithmetic.AdditiveGaussianNoise</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -481,7 +559,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">gaussian_noise_mean</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_mean" title="Permalink to this definition">#</a></dt>
 <dd><p>Mean of the distribution to sample from.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -492,7 +570,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">gaussian_noise_stddev</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_stddev" title="Permalink to this definition">#</a></dt>
 <dd><p>Standard deviation of the distribution to sample from.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -506,7 +584,7 @@ <h1>sleap.nn.config.optimization</h1>
 the [0, 1] range. Values in [0, 255] are first scaled to [0, 1]. See
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.contrast.GammaContrast</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -518,7 +596,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Minimum gamma to use for augmentation. Reasonable values are
 in [0.5, 2.0].</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -530,7 +608,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Maximum gamma to use for augmentation. Reasonable values are
 in [0.5, 2.0].</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -543,7 +621,7 @@ <h1>sleap.nn.config.optimization</h1>
 simply adds the same value to all pixels in the image to simulate broadfield
 illumination change. See <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.arithmetic.Add</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -554,7 +632,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">brightness_min_val</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_min_val" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum value to add to all pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -565,7 +643,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">brightness_max_val</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_max_val" title="Permalink to this definition">#</a></dt>
 <dd><p>Maximum value to add to all pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -579,7 +657,7 @@ <h1>sleap.nn.config.optimization</h1>
 global structure beyond the crop size. Random cropping will be applied after
 the augmentations above.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -590,7 +668,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">random_crop_width</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_width" title="Permalink to this definition">#</a></dt>
 <dd><p>Width of random crops.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -601,7 +679,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">random_crop_height</span></span><a class="headerlink" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_height" title="Permalink to this definition">#</a></dt>
 <dd><p>Height of random crops.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -614,7 +692,7 @@ <h1>sleap.nn.config.optimization</h1>
 the instances will be adjusted accordingly. Body parts that are left/right
 symmetric must be marked on the skeleton in order to be swapped correctly.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -626,7 +704,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, flip images left/right when randomly reflecting
 them. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, flipping is down up/down instead.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -636,7 +714,7 @@ <h1>sleap.nn.config.optimization</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.EarlyStoppingConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">EarlyStoppingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stop_training_on_plateau</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_min_delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_patience</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/optimization.py#L151-L167"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.EarlyStoppingConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">EarlyStoppingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stop_training_on_plateau</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_min_delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_patience</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/optimization.py#L151-L167"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.EarlyStoppingConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for early stopping.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.EarlyStoppingConfig.stop_training_on_plateau">
@@ -646,7 +724,7 @@ <h1>sleap.nn.config.optimization</h1>
 resources when there are minimal improvements to be gained from further
 training, as well as to prevent training into the overfitting regime.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -658,7 +736,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Minimum absolute decrease in the loss in order to consider an
 epoch as not in a plateau.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -670,7 +748,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Number of epochs without an improvement of at least
 <a class="reference internal" href="#sleap.nn.config.optimization.EarlyStoppingConfig.plateau_min_delta" title="sleap.nn.config.optimization.EarlyStoppingConfig.plateau_min_delta"><code class="xref py py-obj docutils literal notranslate"><span class="pre">plateau_min_delta</span></code></a> in order for a plateau to be detected.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -680,7 +758,7 @@ <h1>sleap.nn.config.optimization</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.HardKeypointMiningConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">HardKeypointMiningConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">online_mining</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/optimization.py#L91-L118"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.HardKeypointMiningConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">HardKeypointMiningConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">online_mining</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/optimization.py#L91-L118"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.HardKeypointMiningConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for online hard keypoint mining.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.HardKeypointMiningConfig.online_mining">
@@ -693,7 +771,7 @@ <h1>sleap.nn.config.optimization</h1>
 If False, no mining will be performed and all keypoints will be weighted
 equally in the loss.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -706,7 +784,7 @@ <h1>sleap.nn.config.optimization</h1>
 respect to the lowest keypoint loss in order to be considered as “hard”.
 This helps to switch focus on across groups of keypoints during training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -718,7 +796,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>The minimum number of keypoints that will be considered as
 “hard”, even if they are not below the <a class="reference internal" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.hard_to_easy_ratio" title="sleap.nn.config.optimization.HardKeypointMiningConfig.hard_to_easy_ratio"><code class="xref py py-obj docutils literal notranslate"><span class="pre">hard_to_easy_ratio</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -732,7 +810,7 @@ <h1>sleap.nn.config.optimization</h1>
 ratio and result in loss scaling being applied to most keypoints, which can
 reduce the impact of hard mining altogether.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -743,7 +821,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">loss_scale</span></span><a class="headerlink" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.loss_scale" title="Permalink to this definition">#</a></dt>
 <dd><p>Factor to scale the hard keypoint losses by.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -753,7 +831,7 @@ <h1>sleap.nn.config.optimization</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.LearningRateScheduleConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">LearningRateScheduleConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">reduce_on_plateau</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction_factor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_min_delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_patience</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_cooldown</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_learning_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-08</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/optimization.py#L122-L147"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.LearningRateScheduleConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">LearningRateScheduleConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">reduce_on_plateau</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reduction_factor</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_min_delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-06</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_patience</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">plateau_cooldown</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_learning_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1e-08</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/optimization.py#L122-L147"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.LearningRateScheduleConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration for learning rate scheduling.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.LearningRateScheduleConfig.reduce_on_plateau">
@@ -763,7 +841,7 @@ <h1>sleap.nn.config.optimization</h1>
 updates are required for fine-tuning the optimization, balancing out an
 initially high learning rate required for practical initial optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -775,7 +853,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Factor by which the learning rate will be scaled when a
 plateau is detected.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -787,7 +865,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Minimum absolute decrease in the loss in order to consider an
 epoch as not in a plateau.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -799,7 +877,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Number of epochs without an improvement of at least
 <a class="reference internal" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_min_delta" title="sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_min_delta"><code class="xref py py-obj docutils literal notranslate"><span class="pre">plateau_min_delta</span></code></a> in order for a plateau to be detected.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -811,7 +889,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Number of epochs after a reduction step before epochs without
 improvement will begin to be counted again.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -823,7 +901,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>The minimum absolute value that the learning rate can be
 reduced to.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -833,7 +911,7 @@ <h1>sleap.nn.config.optimization</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.OptimizationConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">OptimizationConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preload_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">augmentation_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.AugmentationConfig" title="sleap.nn.config.optimization.AugmentationConfig"><span class="pre">sleap.nn.config.optimization.AugmentationConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">online_shuffling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shuffle_buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">val_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_val_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">epochs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimizer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'adam'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_learning_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">learning_rate_schedule</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.LearningRateScheduleConfig" title="sleap.nn.config.optimization.LearningRateScheduleConfig"><span class="pre">sleap.nn.config.optimization.LearningRateScheduleConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_keypoint_mining</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.HardKeypointMiningConfig" title="sleap.nn.config.optimization.HardKeypointMiningConfig"><span class="pre">sleap.nn.config.optimization.HardKeypointMiningConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">early_stopping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.EarlyStoppingConfig" title="sleap.nn.config.optimization.EarlyStoppingConfig"><span class="pre">sleap.nn.config.optimization.EarlyStoppingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/optimization.py#L171-L254"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.OptimizationConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.optimization.</span></span><span class="sig-name descname"><span class="pre">OptimizationConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preload_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">augmentation_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.AugmentationConfig" title="sleap.nn.config.optimization.AugmentationConfig"><span class="pre">AugmentationConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">online_shuffling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shuffle_buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">128</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">200</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">val_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_val_batches_per_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">epochs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimizer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'adam'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_learning_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.0001</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">learning_rate_schedule</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.LearningRateScheduleConfig" title="sleap.nn.config.optimization.LearningRateScheduleConfig"><span class="pre">LearningRateScheduleConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_keypoint_mining</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.HardKeypointMiningConfig" title="sleap.nn.config.optimization.HardKeypointMiningConfig"><span class="pre">HardKeypointMiningConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">early_stopping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.optimization.EarlyStoppingConfig" title="sleap.nn.config.optimization.EarlyStoppingConfig"><span class="pre">EarlyStoppingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/optimization.py#L171-L254"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.optimization.OptimizationConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization configuration.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.optimization.OptimizationConfig.preload_data">
@@ -845,7 +923,7 @@ <h1>sleap.nn.config.optimization</h1>
 be loaded into memory. This is especially beneficial for datasets with few
 examples or when the raw data is on (slow) network storage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -856,7 +934,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">augmentation_config</span></span><a class="headerlink" href="#sleap.nn.config.optimization.OptimizationConfig.augmentation_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to data augmentation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.optimization.AugmentationConfig" title="sleap.nn.config.optimization.AugmentationConfig">sleap.nn.config.optimization.AugmentationConfig</a></p>
 </dd>
 </dl>
@@ -871,7 +949,7 @@ <h1>sleap.nn.config.optimization</h1>
 composition, which in turn can promote generalization. Note that the data
 are shuffled at the start of training regardless.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -886,7 +964,7 @@ <h1>sleap.nn.config.optimization</h1>
 the entire dataset will be used, which results in perfect randomization at
 the cost of increased memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -898,7 +976,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>If True, data will generated in parallel to training to minimize the
 bottleneck of the preprocessing pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -914,7 +992,7 @@ <h1>sleap.nn.config.optimization</h1>
 numbers may lead to overfitting, but may be beneficial to the optimization
 process when few but varied examples are available.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -935,7 +1013,7 @@ <h1>sleap.nn.config.optimization</h1>
 too high, the training procedure may take longer, especially since several
 hyperparameter tuning heuristics only consider epoch-to-epoch performance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -949,7 +1027,7 @@ <h1>sleap.nn.config.optimization</h1>
 explicitly specified. This should be set to 200-400 to compensate for short
 loops through the data when there are few examples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -960,7 +1038,7 @@ <h1>sleap.nn.config.optimization</h1>
 <span class="sig-name descname"><span class="pre">val_batches_per_epoch</span></span><a class="headerlink" href="#sleap.nn.config.optimization.OptimizationConfig.val_batches_per_epoch" title="Permalink to this definition">#</a></dt>
 <dd><p>Same as <a class="reference internal" href="#sleap.nn.config.optimization.OptimizationConfig.batches_per_epoch" title="sleap.nn.config.optimization.OptimizationConfig.batches_per_epoch"><code class="xref py py-obj docutils literal notranslate"><span class="pre">batches_per_epoch</span></code></a>, but for the validation set.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -972,7 +1050,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Same as <a class="reference internal" href="#sleap.nn.config.optimization.OptimizationConfig.min_batches_per_epoch" title="sleap.nn.config.optimization.OptimizationConfig.min_batches_per_epoch"><code class="xref py py-obj docutils literal notranslate"><span class="pre">min_batches_per_epoch</span></code></a>, but for the
 validation set.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -984,7 +1062,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Maximum number of epochs to train for. Training can be stopped manually
 or automatically if early stopping is enabled and a plateau is detected.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -997,7 +1075,7 @@ <h1>sleap.nn.config.optimization</h1>
 but the name of any class from <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.optimizers</span></code> may be used. If “adam”
 is specified, the Adam optimizer will have AMSGrad enabled.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1011,7 +1089,7 @@ <h1>sleap.nn.config.optimization</h1>
 learning rate reduction on plateau is enabled. If this is too high or too
 low, the training may fail to find good initial local minima to descend.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1023,7 +1101,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Configuration options related to learning rate
 scheduling.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.optimization.LearningRateScheduleConfig" title="sleap.nn.config.optimization.LearningRateScheduleConfig">sleap.nn.config.optimization.LearningRateScheduleConfig</a></p>
 </dd>
 </dl>
@@ -1035,7 +1113,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Configuration options related to online hard keypoint
 mining.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.optimization.HardKeypointMiningConfig" title="sleap.nn.config.optimization.HardKeypointMiningConfig">sleap.nn.config.optimization.HardKeypointMiningConfig</a></p>
 </dd>
 </dl>
@@ -1047,7 +1125,7 @@ <h1>sleap.nn.config.optimization</h1>
 <dd><p>Configuration options related to early stopping of training on
 plateau/convergence is detected.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.optimization.EarlyStoppingConfig" title="sleap.nn.config.optimization.EarlyStoppingConfig">sleap.nn.config.optimization.EarlyStoppingConfig</a></p>
 </dd>
 </dl>
@@ -1079,6 +1157,90 @@ <h1>sleap.nn.config.optimization</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotate"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_min_angle"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotation_min_angle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.rotation_max_angle"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.rotation_max_angle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate_min"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate_min</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.translate_max"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.translate_max</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale_min"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale_min</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.scale_max"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.scale_max</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_min_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise_min_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.uniform_noise_max_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.uniform_noise_max_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_mean"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise_mean</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_stddev"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.gaussian_noise_stddev</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast_min_gamma"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast_min_gamma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.contrast_max_gamma"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.contrast_max_gamma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_min_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness_min_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.brightness_max_val"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.brightness_max_val</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_width"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_crop_height"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.random_flip"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.random_flip</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.AugmentationConfig.flip_horizontal"><code class="docutils literal notranslate"><span class="pre">AugmentationConfig.flip_horizontal</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.stop_training_on_plateau"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.stop_training_on_plateau</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.plateau_min_delta"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.plateau_min_delta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.EarlyStoppingConfig.plateau_patience"><code class="docutils literal notranslate"><span class="pre">EarlyStoppingConfig.plateau_patience</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.online_mining"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.online_mining</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.hard_to_easy_ratio"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.hard_to_easy_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.min_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.min_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.max_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.max_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.HardKeypointMiningConfig.loss_scale"><code class="docutils literal notranslate"><span class="pre">HardKeypointMiningConfig.loss_scale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.reduce_on_plateau"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.reduce_on_plateau</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.reduction_factor"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.reduction_factor</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_min_delta"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_min_delta</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_patience"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_patience</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_cooldown"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.plateau_cooldown</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.LearningRateScheduleConfig.min_learning_rate"><code class="docutils literal notranslate"><span class="pre">LearningRateScheduleConfig.min_learning_rate</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.preload_data"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.preload_data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.augmentation_config"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.augmentation_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.online_shuffling"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.online_shuffling</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.shuffle_buffer_size"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.shuffle_buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.prefetch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.prefetch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.batch_size"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.min_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.min_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.val_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.val_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.min_val_batches_per_epoch"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.min_val_batches_per_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.epochs"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.epochs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.optimizer"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.optimizer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.initial_learning_rate"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.initial_learning_rate</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.learning_rate_schedule"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.learning_rate_schedule</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.hard_keypoint_mining"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.hard_keypoint_mining</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.optimization.OptimizationConfig.early_stopping"><code class="docutils literal notranslate"><span class="pre">OptimizationConfig.early_stopping</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1097,7 +1259,7 @@ <h1>sleap.nn.config.optimization</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.outputs.html b/api/sleap.nn.config.outputs.html
index 905bbcd42..b2959e859 100644
--- a/api/sleap.nn.config.outputs.html
+++ b/api/sleap.nn.config.outputs.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.outputs &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.outputs &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.outputs';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.outputs.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,54 @@ <h1>sleap.nn.config.outputs</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.initial_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.initial_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.best_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.best_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.every_epoch"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.every_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.latest_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.latest_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.final_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.final_model</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig"><code class="docutils literal notranslate"><span class="pre">OutputsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.save_outputs"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.save_outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name_prefix"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name_prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name_suffix"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name_suffix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.runs_folder"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.runs_folder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.tags"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.tags</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.save_visualizations"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.save_visualizations</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.keep_viz_images"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.keep_viz_images</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.zip_outputs"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.zip_outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.log_to_csv"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.log_to_csv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.checkpointing"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.checkpointing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.tensorboard"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.tensorboard</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.zmq"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.zmq</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_path"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_path</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.write_logs"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.write_logs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.loss_frequency"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.loss_frequency</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.architecture_graph"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.architecture_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.profile_graph"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.profile_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.visualizations"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.visualizations</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig"><code class="docutils literal notranslate"><span class="pre">ZMQConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.subscribe_to_controller"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.subscribe_to_controller</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.controller_address"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.controller_address</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.controller_polling_timeout"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.controller_polling_timeout</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.publish_updates"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.publish_updates</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.publish_address"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.publish_address</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,10 +366,10 @@ <h1>sleap.nn.config.outputs</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.outputs">
-<span id="sleap-nn-config-outputs"></span><h1>sleap.nn.config.outputs<a class="headerlink" href="#module-sleap.nn.config.outputs" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-outputs"></span><h1>sleap.nn.config.outputs<a class="headerlink" href="#module-sleap.nn.config.outputs" title="Permalink to this heading">#</a></h1>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.CheckpointingConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">CheckpointingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">initial_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">best_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">every_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">latest_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">final_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/outputs.py#L7-L37"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.CheckpointingConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">CheckpointingConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">initial_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">best_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">every_epoch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">latest_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">final_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/outputs.py#L7-L37"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.CheckpointingConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration of model checkpointing.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.CheckpointingConfig.initial_model">
@@ -334,7 +382,7 @@ <h1>sleap.nn.config.outputs</h1>
 <div><p>“{run_folder}/initial_model.h5”</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -349,7 +397,7 @@ <h1>sleap.nn.config.outputs</h1>
 <div><p>“{run_folder}/best_model.h5”</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -365,7 +413,7 @@ <h1>sleap.nn.config.outputs</h1>
 <div><p>“{run_folder}/model.epoch{04d}.h5”</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -381,7 +429,7 @@ <h1>sleap.nn.config.outputs</h1>
 <div><p>“{run_folder}/latest_model.h5”</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -397,7 +445,7 @@ <h1>sleap.nn.config.outputs</h1>
 <div><p>“{run_folder}/final_model.h5”</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -407,7 +455,7 @@ <h1>sleap.nn.config.outputs</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.OutputsConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">OutputsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name_prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name_suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">runs_folder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'models'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tags</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_visualizations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">delete_viz_images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">zip_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">log_to_csv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">checkpointing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.CheckpointingConfig" title="sleap.nn.config.outputs.CheckpointingConfig"><span class="pre">sleap.nn.config.outputs.CheckpointingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensorboard</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.TensorBoardConfig" title="sleap.nn.config.outputs.TensorBoardConfig"><span class="pre">sleap.nn.config.outputs.TensorBoardConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">zmq</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.ZMQConfig" title="sleap.nn.config.outputs.ZMQConfig"><span class="pre">sleap.nn.config.outputs.ZMQConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/outputs.py#L113-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">OutputsConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name_prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_name_suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">runs_folder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'models'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tags</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_visualizations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_viz_images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">zip_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">log_to_csv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">checkpointing</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.CheckpointingConfig" title="sleap.nn.config.outputs.CheckpointingConfig"><span class="pre">CheckpointingConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensorboard</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.TensorBoardConfig" title="sleap.nn.config.outputs.TensorBoardConfig"><span class="pre">TensorBoardConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">zmq</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.config.outputs.ZMQConfig" title="sleap.nn.config.outputs.ZMQConfig"><span class="pre">ZMQConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/outputs.py#L113-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration of training outputs.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.OutputsConfig.save_outputs">
@@ -416,7 +464,7 @@ <h1>sleap.nn.config.outputs</h1>
 nothing will be written to disk, which may be useful for interactive
 training where no outputs are desired.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -433,7 +481,7 @@ <h1>sleap.nn.config.outputs</h1>
 runs can end up overwriting each other if <a class="reference internal" href="#sleap.nn.config.outputs.OutputsConfig.run_name_prefix" title="sleap.nn.config.outputs.OutputsConfig.run_name_prefix"><code class="xref py py-obj docutils literal notranslate"><span class="pre">run_name_prefix</span></code></a> or
 <a class="reference internal" href="#sleap.nn.config.outputs.OutputsConfig.run_name_suffix" title="sleap.nn.config.outputs.OutputsConfig.run_name_suffix"><code class="xref py py-obj docutils literal notranslate"><span class="pre">run_name_suffix</span></code></a> are not specified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -446,7 +494,7 @@ <h1>sleap.nn.config.outputs</h1>
 multiple runs started at the same exact time to be mapped to the same
 folder, or when a fixed run name is specified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -468,7 +516,7 @@ <h1>sleap.nn.config.outputs</h1>
 directory listing across clients.</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -488,7 +536,7 @@ <h1>sleap.nn.config.outputs</h1>
 folders will be created if they do not already exist. Defaults to the
 “models” subdirectory of the current working directory.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -501,7 +549,7 @@ <h1>sleap.nn.config.outputs</h1>
 runs. These are not used for anything during training or inference, so they
 can be used to store arbitrary user-specified metadata.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -514,20 +562,20 @@ <h1>sleap.nn.config.outputs</h1>
 predictions as PNGs to “{run_folder}/viz/{split}.{epoch:04d}.png”, where the
 split is one of “train”, “validation”, “test”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.nn.config.outputs.OutputsConfig.delete_viz_images">
-<span class="sig-name descname"><span class="pre">delete_viz_images</span></span><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig.delete_viz_images" title="Permalink to this definition">#</a></dt>
-<dd><p>If True, delete the saved visualizations after training
-completes. This is useful to reduce the model folder size if you do not need
+<dt class="sig sig-object py" id="sleap.nn.config.outputs.OutputsConfig.keep_viz_images">
+<span class="sig-name descname"><span class="pre">keep_viz_images</span></span><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig.keep_viz_images" title="Permalink to this definition">#</a></dt>
+<dd><p>If True, keep the saved visualization images after training
+completes. This is useful unchecked to reduce the model folder size if you do not need
 to keep the visualization images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -539,7 +587,7 @@ <h1>sleap.nn.config.outputs</h1>
 <dd><p>If True, compress the run folder to a zip file. This will be named
 “{run_folder}.zip”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -551,7 +599,7 @@ <h1>sleap.nn.config.outputs</h1>
 <dd><p>If True, loss and metrics will be saved to a simple CSV after each
 epoch to “{run_folder}/training_log.csv”</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -562,7 +610,7 @@ <h1>sleap.nn.config.outputs</h1>
 <span class="sig-name descname"><span class="pre">checkpointing</span></span><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig.checkpointing" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to model checkpointing.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.outputs.CheckpointingConfig" title="sleap.nn.config.outputs.CheckpointingConfig">sleap.nn.config.outputs.CheckpointingConfig</a></p>
 </dd>
 </dl>
@@ -573,7 +621,7 @@ <h1>sleap.nn.config.outputs</h1>
 <span class="sig-name descname"><span class="pre">tensorboard</span></span><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig.tensorboard" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to TensorBoard logging.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.outputs.TensorBoardConfig" title="sleap.nn.config.outputs.TensorBoardConfig">sleap.nn.config.outputs.TensorBoardConfig</a></p>
 </dd>
 </dl>
@@ -584,7 +632,7 @@ <h1>sleap.nn.config.outputs</h1>
 <span class="sig-name descname"><span class="pre">zmq</span></span><a class="headerlink" href="#sleap.nn.config.outputs.OutputsConfig.zmq" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to ZeroMQ-based control and monitoring.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.config.outputs.ZMQConfig" title="sleap.nn.config.outputs.ZMQConfig">sleap.nn.config.outputs.ZMQConfig</a></p>
 </dd>
 </dl>
@@ -600,7 +648,7 @@ <h1>sleap.nn.config.outputs</h1>
 </dl>
 <p>If <a class="reference internal" href="#sleap.nn.config.outputs.OutputsConfig.run_name_suffix" title="sleap.nn.config.outputs.OutputsConfig.run_name_suffix"><code class="xref py py-obj docutils literal notranslate"><span class="pre">run_name_suffix</span></code></a> is set to None, it will be ignored.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – If <a class="reference internal" href="#sleap.nn.config.outputs.OutputsConfig.run_name" title="sleap.nn.config.outputs.OutputsConfig.run_name"><code class="xref py py-obj docutils literal notranslate"><span class="pre">run_name</span></code></a> is not set.</p>
 </dd>
 </dl>
@@ -617,7 +665,7 @@ <h1>sleap.nn.config.outputs</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.TensorBoardConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">TensorBoardConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">write_logs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_frequency</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'epoch'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">architecture_graph</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">profile_graph</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visualizations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/outputs.py#L41-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.TensorBoardConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">TensorBoardConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">write_logs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_frequency</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'epoch'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">architecture_graph</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">profile_graph</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">visualizations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/outputs.py#L41-L69"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.TensorBoardConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration of TensorBoard-based monitoring of the training.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.TensorBoardConfig.write_logs">
@@ -627,7 +675,7 @@ <h1>sleap.nn.config.outputs</h1>
 folder that may contain multiple models/runs. Both will be displayed
 correctly in the dashboard.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -643,7 +691,7 @@ <h1>sleap.nn.config.outputs</h1>
 “epoch” if training interactively. This value only affects the monitored
 losses and metrics, not other summaries like visualizations.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -657,7 +705,7 @@ <h1>sleap.nn.config.outputs</h1>
 beginning of training, but can consume a lot of disk space for large models,
 as well as potentially freezing the browser tab when rendered.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -669,7 +717,7 @@ <h1>sleap.nn.config.outputs</h1>
 <dd><p>If True, profiles the second batch of examples to collect compute
 statistics.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -681,7 +729,7 @@ <h1>sleap.nn.config.outputs</h1>
 <dd><p>If True, visualizations of the model predictions are rendered
 and logged for display in TensorBoard -&gt; Images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -691,7 +739,7 @@ <h1>sleap.nn.config.outputs</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.ZMQConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">ZMQConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">subscribe_to_controller</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">controller_address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9000'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">controller_polling_timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">publish_updates</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">publish_address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/outputs.py#L73-L109"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.ZMQConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.outputs.</span></span><span class="sig-name descname"><span class="pre">ZMQConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">subscribe_to_controller</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">controller_address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9000'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">controller_polling_timeout</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">publish_updates</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">publish_address</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'tcp://127.0.0.1:9001'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/outputs.py#L73-L109"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.outputs.ZMQConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration of ZeroMQ-based monitoring of the training.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.outputs.ZMQConfig.subscribe_to_controller">
@@ -711,7 +759,7 @@ <h1>sleap.nn.config.outputs</h1>
 </div></blockquote>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -725,7 +773,7 @@ <h1>sleap.nn.config.outputs</h1>
 the form of “<a class="reference external" href="tcp:/">tcp:/</a>/{ip_address}:{port_number}”. Defaults to
 “<a class="reference external" href="tcp://127.0.0.1:9000">tcp://127.0.0.1:9000</a>”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -738,7 +786,7 @@ <h1>sleap.nn.config.outputs</h1>
 integer. This controls how long the poller should wait to receive a response
 and should be set to a small value to minimize the impact on training speed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -754,7 +802,7 @@ <h1>sleap.nn.config.outputs</h1>
 All data will be broadcast as JSON serialized strings.
 TODO: Describe published message keys.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -767,7 +815,7 @@ <h1>sleap.nn.config.outputs</h1>
 updates to. For TCP-based endpoints, this must be in the form of
 “<a class="reference external" href="tcp:/">tcp:/</a>/{ip_address}:{port_number}”. Defaults to “<a class="reference external" href="tcp://127.0.0.1:9001">tcp://127.0.0.1:9001</a>”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -799,6 +847,60 @@ <h1>sleap.nn.config.outputs</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.initial_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.initial_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.best_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.best_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.every_epoch"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.every_epoch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.latest_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.latest_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.CheckpointingConfig.final_model"><code class="docutils literal notranslate"><span class="pre">CheckpointingConfig.final_model</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig"><code class="docutils literal notranslate"><span class="pre">OutputsConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.save_outputs"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.save_outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name_prefix"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name_prefix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_name_suffix"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_name_suffix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.runs_folder"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.runs_folder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.tags"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.tags</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.save_visualizations"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.save_visualizations</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.keep_viz_images"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.keep_viz_images</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.zip_outputs"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.zip_outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.log_to_csv"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.log_to_csv</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.checkpointing"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.checkpointing</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.tensorboard"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.tensorboard</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.zmq"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.zmq</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.OutputsConfig.run_path"><code class="docutils literal notranslate"><span class="pre">OutputsConfig.run_path</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.write_logs"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.write_logs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.loss_frequency"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.loss_frequency</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.architecture_graph"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.architecture_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.profile_graph"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.profile_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.TensorBoardConfig.visualizations"><code class="docutils literal notranslate"><span class="pre">TensorBoardConfig.visualizations</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig"><code class="docutils literal notranslate"><span class="pre">ZMQConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.subscribe_to_controller"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.subscribe_to_controller</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.controller_address"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.controller_address</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.controller_polling_timeout"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.controller_polling_timeout</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.publish_updates"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.publish_updates</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.outputs.ZMQConfig.publish_address"><code class="docutils literal notranslate"><span class="pre">ZMQConfig.publish_address</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -817,7 +919,7 @@ <h1>sleap.nn.config.outputs</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.training_job.html b/api/sleap.nn.config.training_job.html
index c23b0c5c5..76903a602 100644
--- a/api/sleap.nn.config.training_job.html
+++ b/api/sleap.nn.config.training_job.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.training_job &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.training_job &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.training_job';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.training_job.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,30 @@ <h1>sleap.nn.config.training_job</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.data"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.model"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.optimization"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.optimization</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.outputs"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.name"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.description"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.description</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.sleap_version"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.sleap_version</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.filename"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.filename</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.from_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.from_json_dicts()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.load_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.load_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.save_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.save_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.to_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.to_json()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.load_config"><code class="docutils literal notranslate"><span class="pre">load_config()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +342,7 @@ <h1>sleap.nn.config.training_job</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.training_job">
-<span id="sleap-nn-config-training-job"></span><h1>sleap.nn.config.training_job<a class="headerlink" href="#module-sleap.nn.config.training_job" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-training-job"></span><h1>sleap.nn.config.training_job<a class="headerlink" href="#module-sleap.nn.config.training_job" title="Permalink to this heading">#</a></h1>
 <p>Serializable configuration classes for specifying all training job parameters.</p>
 <p>These configuration classes are intended to specify all the parameters required to run
 a training job or perform inference from a serialized one.</p>
@@ -349,14 +373,14 @@ <h1>sleap.nn.config.training_job</h1>
 parameters are aggregated and documented for end users (as opposed to developers).</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.training_job.</span></span><span class="sig-name descname"><span class="pre">TrainingJobConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ModelConfig" title="sleap.nn.config.model.ModelConfig"><span class="pre">sleap.nn.config.model.ModelConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">sleap.nn.config.outputs.OutputsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sleap_version</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'1.3.3'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L41-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.config.training_job.</span></span><span class="sig-name descname"><span class="pre">TrainingJobConfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ModelConfig" title="sleap.nn.config.model.ModelConfig"><span class="pre">ModelConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">OutputsConfig</span></a></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sleap_version</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'1.4.1'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L41-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration of a training job.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.data">
 <span class="sig-name descname"><span class="pre">data</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.data" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to the training data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -367,7 +391,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">model</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.model" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to the model architecture.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ModelConfig" title="sleap.nn.config.model.ModelConfig">sleap.nn.config.model.ModelConfig</a></p>
 </dd>
 </dl>
@@ -378,7 +402,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">optimization</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.optimization" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to the training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -389,7 +413,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">outputs</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.outputs" title="Permalink to this definition">#</a></dt>
 <dd><p>Configuration options related to outputs during training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig">sleap.nn.config.outputs.OutputsConfig</a></p>
 </dd>
 </dl>
@@ -400,7 +424,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">name</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.name" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional name for this configuration profile.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -411,7 +435,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">description</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.description" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional description of the configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -422,7 +446,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">sleap_version</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.sleap_version" title="Permalink to this definition">#</a></dt>
 <dd><p>Version of SLEAP that generated this configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -433,7 +457,7 @@ <h1>sleap.nn.config.training_job</h1>
 <span class="sig-name descname"><span class="pre">filename</span></span><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Path to this config file if it was loaded from disk.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -441,13 +465,13 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.from_json">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L78-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L78-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Create training job configuration from JSON text data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>json_data</strong> – JSON-formatted string that specifies the configurations.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A TrainingJobConfig instance parsed from the JSON text.</p>
 </dd>
 </dl>
@@ -455,14 +479,14 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json_dicts</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_data_dicts</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L64-L76"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json_dicts</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_data_dicts</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L64-L76"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts" title="Permalink to this definition">#</a></dt>
 <dd><p>Create training job configuration from dictionaries decoded from JSON.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>json_data_dicts</strong> – Dictionaries that specify the configurations. These are
 typically generated by structuring raw JSON formatted text.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A TrainingJobConfig instance parsed from the JSON dicts.</p>
 </dd>
 </dl>
@@ -470,10 +494,10 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.load_json">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">load_training_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L92-L124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.load_json" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">load_training_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L92-L124"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.load_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a training job configuration from a file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a training job configuration JSON file or a directory
 containing <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;training_job.json&quot;</span></code>.</p></li>
@@ -481,7 +505,7 @@ <h1>sleap.nn.config.training_job</h1>
 over <code class="xref py py-obj docutils literal notranslate"><span class="pre">initial_config.json</span></code> if it is present in the same folder.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A TrainingJobConfig instance parsed from the file.</p>
 </dd>
 </dl>
@@ -489,10 +513,10 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.save_json">
-<span class="sig-name descname"><span class="pre">save_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L135-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.save_json" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L135-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.save_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Save the configuration to a JSON file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>filename</strong> – Path to save the training job file to.</p>
 </dd>
 </dl>
@@ -500,10 +524,10 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.TrainingJobConfig.to_json">
-<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L126-L133"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.to_json" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L126-L133"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.TrainingJobConfig.to_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Serialize the configuration into JSON-encoded string format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The JSON encoded string representation of the configuration.</p>
 </dd>
 </dl>
@@ -513,17 +537,17 @@ <h1>sleap.nn.config.training_job</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.config.training_job.load_config">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.config.training_job.</span></span><span class="sig-name descname"><span class="pre">load_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">load_training_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/training_job.py#L146-L159"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.load_config" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.config.training_job.</span></span><span class="sig-name descname"><span class="pre">load_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">load_training_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/training_job.py#L146-L159"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.training_job.load_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a training job configuration for a model run.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a JSON file or directory containing <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_job.json</span></code>.</p></li>
 <li><p><strong>load_training_config</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), prefer <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_job.json</span></code> over
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">initial_config.json</span></code> if it is present in the same folder.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The parsed <a class="reference internal" href="#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><code class="xref py py-obj docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code></a>.</p>
 </dd>
 </dl>
@@ -553,6 +577,36 @@ <h1>sleap.nn.config.training_job</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.data"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.data</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.model"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.optimization"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.optimization</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.outputs"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.outputs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.name"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.description"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.description</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.sleap_version"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.sleap_version</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.filename"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.filename</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.from_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.from_json_dicts()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.load_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.load_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.save_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.save_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.TrainingJobConfig.to_json"><code class="docutils literal notranslate"><span class="pre">TrainingJobConfig.to_json()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.training_job.load_config"><code class="docutils literal notranslate"><span class="pre">load_config()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -571,7 +625,7 @@ <h1>sleap.nn.config.training_job</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.config.utils.html b/api/sleap.nn.config.utils.html
index b9c934e65..5b7feb56c 100644
--- a/api/sleap.nn.config.utils.html
+++ b/api/sleap.nn.config.utils.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.config.utils &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.config.utils &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.config.utils';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.config.utils.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,14 @@ <h1>sleap.nn.config.utils</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.utils.oneof"><code class="docutils literal notranslate"><span class="pre">oneof()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,22 +326,22 @@ <h1>sleap.nn.config.utils</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.config.utils">
-<span id="sleap-nn-config-utils"></span><h1>sleap.nn.config.utils<a class="headerlink" href="#module-sleap.nn.config.utils" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-config-utils"></span><h1>sleap.nn.config.utils<a class="headerlink" href="#module-sleap.nn.config.utils" title="Permalink to this heading">#</a></h1>
 <p>Utilities for config building and validation.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.config.utils.oneof">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.config.utils.</span></span><span class="sig-name descname"><span class="pre">oneof</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">attrs_cls</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">must_be_set</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/config/utils.py#L4-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.utils.oneof" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.config.utils.</span></span><span class="sig-name descname"><span class="pre">oneof</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">attrs_cls</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">must_be_set</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/config/utils.py#L4-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.config.utils.oneof" title="Permalink to this definition">#</a></dt>
 <dd><p>Ensure that the decorated attrs class only has a single attribute set.</p>
 <p>This decorator is inspired by the <a class="reference internal" href="#sleap.nn.config.utils.oneof" title="sleap.nn.config.utils.oneof"><code class="xref py py-obj docutils literal notranslate"><span class="pre">oneof</span></code></a> protobuffer field behavior.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>attrs_cls</strong> – An attrs decorated class.</p></li>
 <li><p><strong>must_be_set</strong> – If True, raise an error if none of the attributes are set. If not,
 error will only be raised if more than one attribute is set.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">attrs_cls</span></code> with an <code class="xref py py-obj docutils literal notranslate"><span class="pre">__init__</span></code> method that checks for the number of
 attributes that are set.</p>
 </dd>
@@ -364,6 +372,20 @@ <h1>sleap.nn.config.utils</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.config.utils.oneof"><code class="docutils literal notranslate"><span class="pre">oneof()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -382,7 +404,7 @@ <h1>sleap.nn.config.utils</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.augmentation.html b/api/sleap.nn.data.augmentation.html
index fa46f0874..3038a20c8 100644
--- a/api/sleap.nn.data.augmentation.html
+++ b/api/sleap.nn.data.augmentation.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.augmentation &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.augmentation &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.augmentation';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.augmentation.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,39 @@ <h1>sleap.nn.data.augmentation</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.augmenter</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper"><code class="docutils literal notranslate"><span class="pre">RandomCropper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.crop_height"><code class="docutils literal notranslate"><span class="pre">RandomCropper.crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.crop_width"><code class="docutils literal notranslate"><span class="pre">RandomCropper.crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">RandomCropper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper"><code class="docutils literal notranslate"><span class="pre">RandomFlipper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.symmetric_inds"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.symmetric_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.horizontal"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.horizontal</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.probability"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.probability</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.from_skeleton"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.from_skeleton()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.flip_instances_lr"><code class="docutils literal notranslate"><span class="pre">flip_instances_lr()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.flip_instances_ud"><code class="docutils literal notranslate"><span class="pre">flip_instances_ud()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,57 +351,57 @@ <h1>sleap.nn.data.augmentation</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.augmentation">
-<span id="sleap-nn-data-augmentation"></span><h1>sleap.nn.data.augmentation<a class="headerlink" href="#module-sleap.nn.data.augmentation" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-augmentation"></span><h1>sleap.nn.data.augmentation<a class="headerlink" href="#module-sleap.nn.data.augmentation" title="Permalink to this heading">#</a></h1>
 <p>Transformers for applying data augmentation.</p>
 <dl class="py class">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">ImgaugAugmenter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">augmenter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">imgaug.augmenters.meta.Sequential</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L114-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter" title="Permalink to this definition">#</a></dt>
-<dd><p>Data transformer based on the <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug</span></code> library.</p>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">AlbumentationsAugmenter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">augmenter</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Compose</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L106-L261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter" title="Permalink to this definition">#</a></dt>
+<dd><p>Data transformer based on the <code class="xref py py-obj docutils literal notranslate"><span class="pre">albumentations</span></code> library.</p>
 <p>This class can generate a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> from an existing one that generates
 image and instance data. Element of the output dataset will have a set of
 augmentation transformations applied.</p>
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.augmenter">
-<span class="sig-name descname"><span class="pre">augmenter</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.augmenter" title="Permalink to this definition">#</a></dt>
-<dd><p>An instance of <code class="xref py py-obj docutils literal notranslate"><span class="pre">imgaug.augmenters.Sequential</span></code> that will be applied to
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter">
+<span class="sig-name descname"><span class="pre">augmenter</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter" title="Permalink to this definition">#</a></dt>
+<dd><p>An instance of <code class="xref py py-obj docutils literal notranslate"><span class="pre">albumentations.Compose</span></code> that will be applied to
 each element of the input dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
-<dd class="field-odd"><p>imgaug.augmenters.meta.Sequential</p>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
+<dd class="field-odd"><p>albumentations.core.composition.Compose</p>
 </dd>
 </dl>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.image_key">
-<span class="sig-name descname"><span class="pre">image_key</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.image_key" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key">
+<span class="sig-name descname"><span class="pre">image_key</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>Name of the example key where the image is stored. Defaults to
 “image”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
 </dd></dl>
 
 <dl class="py attribute">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.instances_key">
-<span class="sig-name descname"><span class="pre">instances_key</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.instances_key" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key">
+<span class="sig-name descname"><span class="pre">instances_key</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key" title="Permalink to this definition">#</a></dt>
 <dd><p>Name of the example key where the instance points are stored.
 Defaults to “instances”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
 </dd></dl>
 
 <dl class="py method">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.AugmentationConfig" title="sleap.nn.config.optimization.AugmentationConfig"><span class="pre">sleap.nn.config.optimization.AugmentationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.augmentation.ImgaugAugmenter" title="sleap.nn.data.augmentation.ImgaugAugmenter"><span class="pre">sleap.nn.data.augmentation.ImgaugAugmenter</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L134-L199"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.from_config" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config">
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.AugmentationConfig" title="sleap.nn.config.optimization.AugmentationConfig"><span class="pre">AugmentationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter" title="sleap.nn.data.augmentation.AlbumentationsAugmenter"><span class="pre">AlbumentationsAugmenter</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L126-L203"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an augmenter from a set of configuration parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">AugmentationConfig</span></code> instance with the desired parameters.</p></li>
 <li><p><strong>image_key</strong> – Name of the example key where the image is stored. Defaults to
@@ -377,35 +410,35 @@ <h1>sleap.nn.data.augmentation</h1>
 Defaults to “instances”.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.data.augmentation.ImgaugAugmenter" title="sleap.nn.data.augmentation.ImgaugAugmenter"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ImgaugAugmenter</span></code></a> with the specified augmentation
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter" title="sleap.nn.data.augmentation.AlbumentationsAugmenter"><code class="xref py py-obj docutils literal notranslate"><span class="pre">AlbumentationsAugmenter</span></code></a> with the specified augmentation
 configuration.</p>
 </dd>
 </dl>
 </dd></dl>
 
 <dl class="py property">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.input_keys">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">input_keys</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.input_keys" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys">
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">input_keys</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the keys that incoming elements are expected to have.</p>
 </dd></dl>
 
 <dl class="py property">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.output_keys">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">output_keys</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.output_keys" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys">
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">output_keys</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the keys that outgoing elements will have.</p>
 </dd></dl>
 
 <dl class="py method">
-<dt class="sig sig-object py" id="sleap.nn.data.augmentation.ImgaugAugmenter.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L211-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.ImgaugAugmenter.transform_dataset" title="Permalink to this definition">#</a></dt>
+<dt class="sig sig-object py" id="sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset">
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L215-L261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing augmented data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys “image” and
 “instances”. This is typically raw data from a data provider.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, but with images and
 instance points updated with the applied augmentations.</p>
 </dd>
@@ -421,7 +454,7 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.RandomCropper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">RandomCropper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L268-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">RandomCropper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">256</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L265-L333"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer for applying random crops to input images.</p>
 <p>This class can generate a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> from an existing one that generates
 image and instance data. Element of the output dataset will have random crops
@@ -431,7 +464,7 @@ <h1>sleap.nn.data.augmentation</h1>
 <span class="sig-name descname"><span class="pre">crop_height</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper.crop_height" title="Permalink to this definition">#</a></dt>
 <dd><p>The height of the cropped region in pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -442,7 +475,7 @@ <h1>sleap.nn.data.augmentation</h1>
 <span class="sig-name descname"><span class="pre">crop_width</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper.crop_width" title="Permalink to this definition">#</a></dt>
 <dd><p>The width of the cropped region in pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -450,14 +483,14 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.RandomCropper.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L291-L336"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L288-L333"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomCropper.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing augmented data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys “image” and
 “instances”. This is typically raw data from a data provider.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, but with images and
 instance points updated with the applied random crop.</p>
 <p>Additionally, the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;crop_bbox&quot;</span></code> key will contain the bounding box of the
@@ -471,7 +504,7 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.RandomFlipper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">RandomFlipper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">probability</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L340-L427"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">RandomFlipper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">probability</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L337-L424"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer for applying random flipping to input images.</p>
 <p>This class can generate a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> from an existing one that generates
 image and instance data. Elements of the output dataset will have random horizontal
@@ -486,7 +519,7 @@ <h1>sleap.nn.data.augmentation</h1>
 of nodes will be swapped to account for the reflection if this is not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>
 (the default).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[numpy.ndarray]</p>
 </dd>
 </dl>
@@ -498,7 +531,7 @@ <h1>sleap.nn.data.augmentation</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), flips are applied horizontally instead of
 vertically.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -509,7 +542,7 @@ <h1>sleap.nn.data.augmentation</h1>
 <span class="sig-name descname"><span class="pre">probability</span></span><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper.probability" title="Permalink to this definition">#</a></dt>
 <dd><p>The probability that the augmentation should be applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -517,10 +550,10 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.RandomFlipper.from_skeleton">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">probability</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.augmentation.RandomFlipper" title="sleap.nn.data.augmentation.RandomFlipper"><span class="pre">sleap.nn.data.augmentation.RandomFlipper</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L363-L382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper.from_skeleton" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">horizontal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">probability</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.augmentation.RandomFlipper" title="sleap.nn.data.augmentation.RandomFlipper"><span class="pre">RandomFlipper</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L360-L379"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper.from_skeleton" title="Permalink to this definition">#</a></dt>
 <dd><p>Create an instance of <a class="reference internal" href="#sleap.nn.data.augmentation.RandomFlipper" title="sleap.nn.data.augmentation.RandomFlipper"><code class="xref py py-obj docutils literal notranslate"><span class="pre">RandomFlipper</span></code></a> from a skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>skeleton</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Skeleton</span></code> that may define symmetric nodes.</p></li>
 <li><p><strong>horizontal</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> (the default), flips are applied horizontally instead
@@ -528,7 +561,7 @@ <h1>sleap.nn.data.augmentation</h1>
 <li><p><strong>probability</strong> – The probability that the augmentation should be applied.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.data.augmentation.RandomFlipper" title="sleap.nn.data.augmentation.RandomFlipper"><code class="xref py py-obj docutils literal notranslate"><span class="pre">RandomFlipper</span></code></a>.</p>
 </dd>
 </dl>
@@ -536,14 +569,14 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.RandomFlipper.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L392-L427"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L389-L424"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.RandomFlipper.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing augmented data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code> and
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instances&quot;</span></code>. This is typically raw data from a data provider.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, but with images and
 instance points updated with the applied random flip.</p>
 </dd>
@@ -554,10 +587,10 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.flip_instances_lr">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">flip_instances_lr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L21-L64"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.flip_instances_lr" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">flip_instances_lr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L13-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.flip_instances_lr" title="Permalink to this definition">#</a></dt>
 <dd><p>Flip a set of instance points horizontally with symmetric node adjustment.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – Instance points as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>
 and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p></li>
@@ -570,7 +603,7 @@ <h1>sleap.nn.data.augmentation</h1>
 reflection if this is not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> (the default).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The instance points with x-coordinates flipped horizontally.</p>
 </dd>
 </dl>
@@ -578,10 +611,10 @@ <h1>sleap.nn.data.augmentation</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.augmentation.flip_instances_ud">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">flip_instances_ud</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/augmentation.py#L67-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.flip_instances_ud" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.augmentation.</span></span><span class="sig-name descname"><span class="pre">flip_instances_ud</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">symmetric_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/augmentation.py#L59-L102"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.augmentation.flip_instances_ud" title="Permalink to this definition">#</a></dt>
 <dd><p>Flip a set of instance points vertically with symmetric node adjustment.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – Instance points as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>
 and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p></li>
@@ -594,7 +627,7 @@ <h1>sleap.nn.data.augmentation</h1>
 reflection if this is not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> (the default).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The instance points with y-coordinates flipped horizontally.</p>
 </dd>
 </dl>
@@ -624,6 +657,45 @@ <h1>sleap.nn.data.augmentation</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.augmenter</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">AlbumentationsAugmenter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper"><code class="docutils literal notranslate"><span class="pre">RandomCropper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.crop_height"><code class="docutils literal notranslate"><span class="pre">RandomCropper.crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.crop_width"><code class="docutils literal notranslate"><span class="pre">RandomCropper.crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomCropper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">RandomCropper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper"><code class="docutils literal notranslate"><span class="pre">RandomFlipper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.symmetric_inds"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.symmetric_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.horizontal"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.horizontal</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.probability"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.probability</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.from_skeleton"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.from_skeleton()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.RandomFlipper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">RandomFlipper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.flip_instances_lr"><code class="docutils literal notranslate"><span class="pre">flip_instances_lr()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.augmentation.flip_instances_ud"><code class="docutils literal notranslate"><span class="pre">flip_instances_ud()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -642,7 +714,7 @@ <h1>sleap.nn.data.augmentation</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.confidence_maps.html b/api/sleap.nn.data.confidence_maps.html
index f8446b57a..f7c8fae55 100644
--- a/api/sleap.nn.data.confidence_maps.html
+++ b/api/sleap.nn.data.confidence_maps.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.confidence_maps &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.confidence_maps &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.confidence_maps';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.confidence_maps.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,51 @@ <h1>sleap.nn.data.confidence_maps</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.all_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.centroids"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.centroids</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_confmaps"><code class="docutils literal notranslate"><span class="pre">make_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_multi_confmaps"><code class="docutils literal notranslate"><span class="pre">make_multi_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets"><code class="docutils literal notranslate"><span class="pre">make_multi_confmaps_with_offsets()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +363,11 @@ <h1>sleap.nn.data.confidence_maps</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.confidence_maps">
-<span id="sleap-nn-data-confidence-maps"></span><h1>sleap.nn.data.confidence_maps<a class="headerlink" href="#module-sleap.nn.data.confidence_maps" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-confidence-maps"></span><h1>sleap.nn.data.confidence_maps<a class="headerlink" href="#module-sleap.nn.data.confidence_maps" title="Permalink to this heading">#</a></h1>
 <p>Transformers for confidence map generation.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">InstanceConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L319-L456"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">InstanceConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">all_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L319-L456"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate instance-centered confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.sigma">
@@ -331,7 +376,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 confidence maps. This defines the spread in units of the input image’s grid,
 i.e., it does not take scaling in previous steps into account.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -344,7 +389,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 effectively the reciprocal of the output scale, i.e., increase this to
 generate confidence maps that are smaller than the input images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -355,7 +400,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <span class="sig-name descname"><span class="pre">all_instances</span></span><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, will also generate the multi-instance confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -366,7 +411,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <span class="sig-name descname"><span class="pre">with_offsets</span></span><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, also return offsets for refining the peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -378,7 +423,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <dd><p>Minimum confidence map value below which offsets will be
 replaced with zeros.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -391,7 +436,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 flattened to produce rank-3 tensors. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, the generated offset maps
 will be rank-4 with shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(height,</span> <span class="pre">width,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -411,15 +456,15 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L365-L456"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L365-L456"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_image&quot;</span></code>,
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;center_instance&quot;</span></code> and, if the attribute <a class="reference internal" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances" title="sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">all_instances</span></code></a> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>,
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;all_instances&quot;</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_confidence_maps&quot;</span></code> and, if the attribute <a class="reference internal" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances" title="sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">all_instances</span></code></a> is
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;all_instance_confidence_maps&quot;</span></code> keys containing the generated
@@ -443,7 +488,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">MultiConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L196-L315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">MultiConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L196-L315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate multi-instance confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.sigma">
@@ -452,7 +497,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 confidence maps. This defines the spread in units of the input image’s grid,
 i.e., it does not take scaling in previous steps into account.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -465,7 +510,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 effectively the reciprocal of the output scale, i.e., increase this to
 generate confidence maps that are smaller than the input images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -477,7 +522,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, generate confidence maps for centroids rather than
 instance points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -488,7 +533,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <span class="sig-name descname"><span class="pre">with_offsets</span></span><a class="headerlink" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, also return offsets for refining the peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -500,7 +545,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <dd><p>Minimum confidence map value below which offsets will be
 replaced with zeros.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -513,7 +558,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 flattened to produce rank-3 tensors. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, the generated offset maps
 will be rank-4 with shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(height,</span> <span class="pre">width,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -533,15 +578,15 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L243-L315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L243-L315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;scale&quot;</span></code>
 and either “instances” or “centroids” depending on whether the
 <a class="reference internal" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.centroids" title="sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.centroids"><code class="xref py py-obj docutils literal notranslate"><span class="pre">centroids</span></code></a> attribute is set to <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;confidence_maps&quot;</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroid_confidence_maps&quot;</span></code> key containing the
 generated confidence maps.</p>
@@ -564,7 +609,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L460-L556"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfidenceMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L460-L556"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate single-instance confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.sigma">
@@ -573,7 +618,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 confidence maps. This defines the spread in units of the input image’s grid,
 i.e., it does not take scaling in previous steps into account.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -586,7 +631,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 effectively the reciprocal of the output scale, i.e., increase this to
 generate confidence maps that are smaller than the input images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -597,7 +642,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <span class="sig-name descname"><span class="pre">with_offsets</span></span><a class="headerlink" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, also return offsets for refining the peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -609,7 +654,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 <dd><p>Minimum confidence map value below which offsets will be
 replaced with zeros.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -622,7 +667,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 flattened to produce rank-3 tensors. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, the generated offset maps
 will be rank-4 with shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(height,</span> <span class="pre">width,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -642,14 +687,14 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L497-L556"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L497-L556"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instances&quot;</span></code> and
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;confidence_maps&quot;</span></code> containing the generated confidence maps.</p>
 </dd>
@@ -668,10 +713,10 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.make_confmaps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L10-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_confmaps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L10-L54"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_confmaps" title="Permalink to this definition">#</a></dt>
 <dd><p>Make confidence maps from a set of points from a single instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – A tensor of points of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> where
 the last axis corresponds to (x, y) pixel coordinates on the image. These
@@ -686,7 +731,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 confidence maps.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Confidence maps as a tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_nodes)</span></code> of
 dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p>
 <p>Each channel of the confidence maps will contain the unnormalized PDF of a 2D
@@ -705,10 +750,10 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.make_multi_confmaps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L57-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_multi_confmaps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L57-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_multi_confmaps" title="Permalink to this definition">#</a></dt>
 <dd><p>Make confidence maps for multiple instances through reduction.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – A tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>
 containing instance points where the last axis corresponds to (x, y) pixel
@@ -724,7 +769,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 confidence maps.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Confidence maps as a tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_nodes)</span></code> of
 dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p>
 <p>Each channel will contain the elementwise maximum of the confidence maps
@@ -744,10 +789,10 @@ <h1>sleap.nn.data.confidence_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_confmaps_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/confidence_maps.py#L113-L192"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.confidence_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_confmaps_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/confidence_maps.py#L113-L192"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>Make confidence maps and offsets for multiple instances through reduction.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – A tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>
 containing instance points where the last axis corresponds to (x, y) pixel
@@ -771,7 +816,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
 will be rank-4 with shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(height,</span> <span class="pre">width,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(confmaps,</span> <span class="pre">offsets)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">confmaps</span></code> are confidence maps as a tensor of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_nodes)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p>
@@ -817,6 +862,57 @@ <h1>sleap.nn.data.confidence_maps</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.all_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.centroids"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.centroids</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">MultiConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.with_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.with_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.offsets_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.offsets_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.flatten_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.flatten_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfidenceMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_confmaps"><code class="docutils literal notranslate"><span class="pre">make_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_multi_confmaps"><code class="docutils literal notranslate"><span class="pre">make_multi_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets"><code class="docutils literal notranslate"><span class="pre">make_multi_confmaps_with_offsets()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -835,7 +931,7 @@ <h1>sleap.nn.data.confidence_maps</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.dataset_ops.html b/api/sleap.nn.data.dataset_ops.html
index c56bd199f..de8fd234a 100644
--- a/api/sleap.nn.data.dataset_ops.html
+++ b/api/sleap.nn.data.dataset_ops.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.dataset_ops &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.dataset_ops &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.dataset_ops';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.dataset_ops.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,67 @@ <h1>sleap.nn.data.dataset_ops</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher"><code class="docutils literal notranslate"><span class="pre">Batcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.batch_size"><code class="docutils literal notranslate"><span class="pre">Batcher.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.drop_remainder"><code class="docutils literal notranslate"><span class="pre">Batcher.drop_remainder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.unrag"><code class="docutils literal notranslate"><span class="pre">Batcher.unrag</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Batcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Batcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Batcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter"><code class="docutils literal notranslate"><span class="pre">LambdaFilter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.filter_fn"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.filter_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.input_keys"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.output_keys"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher"><code class="docutils literal notranslate"><span class="pre">Prefetcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.prefetch"><code class="docutils literal notranslate"><span class="pre">Prefetcher.prefetch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.buffer_size"><code class="docutils literal notranslate"><span class="pre">Prefetcher.buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Prefetcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Prefetcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Prefetcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader"><code class="docutils literal notranslate"><span class="pre">Preloader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.examples"><code class="docutils literal notranslate"><span class="pre">Preloader.examples</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.input_keys"><code class="docutils literal notranslate"><span class="pre">Preloader.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.output_keys"><code class="docutils literal notranslate"><span class="pre">Preloader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Preloader.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater"><code class="docutils literal notranslate"><span class="pre">Repeater</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.repeat"><code class="docutils literal notranslate"><span class="pre">Repeater.repeat</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.epochs"><code class="docutils literal notranslate"><span class="pre">Repeater.epochs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.input_keys"><code class="docutils literal notranslate"><span class="pre">Repeater.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.output_keys"><code class="docutils literal notranslate"><span class="pre">Repeater.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Repeater.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler"><code class="docutils literal notranslate"><span class="pre">Shuffler</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.shuffle"><code class="docutils literal notranslate"><span class="pre">Shuffler.shuffle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.buffer_size"><code class="docutils literal notranslate"><span class="pre">Shuffler.buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.reshuffle_each_iteration"><code class="docutils literal notranslate"><span class="pre">Shuffler.reshuffle_each_iteration</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.input_keys"><code class="docutils literal notranslate"><span class="pre">Shuffler.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.output_keys"><code class="docutils literal notranslate"><span class="pre">Shuffler.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Shuffler.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher"><code class="docutils literal notranslate"><span class="pre">Unbatcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Unbatcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Unbatcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Unbatcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,12 +379,12 @@ <h1>sleap.nn.data.dataset_ops</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.dataset_ops">
-<span id="sleap-nn-data-dataset-ops"></span><h1>sleap.nn.data.dataset_ops<a class="headerlink" href="#module-sleap.nn.data.dataset_ops" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-dataset-ops"></span><h1>sleap.nn.data.dataset_ops<a class="headerlink" href="#module-sleap.nn.data.dataset_ops" title="Permalink to this heading">#</a></h1>
 <p>Transformers for dataset (multi-example) operations, e.g., shuffling and batching.</p>
 <p>These are mostly wrappers for standard tf.data.Dataset ops.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Batcher">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Batcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">drop_remainder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L76-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Batcher" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Batcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">drop_remainder</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L76-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Batcher" title="Permalink to this definition">#</a></dt>
 <dd><p>Batching transformer for use in pipelines.</p>
 <p>This class enables variable-length example keys to be batched by converting them to
 ragged tensors prior to concatenation, then converting them back to dense tensors.</p>
@@ -339,7 +400,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <dd><p>Number of elements within a batch. Every key will be stacked within
 their first axis (with expansion) such that it has <a class="reference internal" href="#sleap.nn.data.dataset_ops.Batcher.batch_size" title="sleap.nn.data.dataset_ops.Batcher.batch_size"><code class="xref py py-obj docutils literal notranslate"><span class="pre">batch_size</span></code></a> length.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -352,7 +413,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 will be dropped once the end of the input dataset iteration is reached. This
 should be True for training and False for inference.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -366,7 +427,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 with NaN padding. Leaving tensors as ragged is useful when downstream ops
 will continue to use them in variable length form.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -386,14 +447,14 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Batcher.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L114-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Batcher.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L114-L158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Batcher.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with batched elements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any dataset that produces dictionaries keyed by strings and values
 with any rank tensors.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same keys, but with each
 tensor promoted to 1 rank higher (except for scalars with rank 0 will be
 promoted to rank 2).</p>
@@ -411,7 +472,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.LambdaFilter">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">LambdaFilter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filter_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L324-L362"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.LambdaFilter" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">LambdaFilter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filter_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bool</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L324-L362"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.LambdaFilter" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for filtering examples out of a dataset.</p>
 <p>This class is useful for eliminating examples that fail to meet some criteria, e.g.,
 when no peaks are found.</p>
@@ -421,7 +482,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <dd><p>Callable that takes an example dictionary as input and returns True
 if the the element should be kept.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Callable[[Dict[str, tensorflow.python.framework.ops.Tensor]], bool]</p>
 </dd>
 </dl>
@@ -441,14 +502,14 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L347-L362"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L347-L362"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with filtering applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any dataset that produces dictionaries keyed by strings and values
 with any rank tensors.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same keys, but with
 potentially fewer elements.</p>
 </dd>
@@ -459,7 +520,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Prefetcher">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Prefetcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L232-L272"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Prefetcher" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Prefetcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">prefetch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L232-L272"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Prefetcher" title="Permalink to this definition">#</a></dt>
 <dd><p>Prefetching transformer for use in pipelines.</p>
 <p>Prefetches elements from the input dataset to minimize the processing bottleneck
 as elements are requested since prefetching can occur in parallel.</p>
@@ -468,7 +529,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <span class="sig-name descname"><span class="pre">prefetch</span></span><a class="headerlink" href="#sleap.nn.data.dataset_ops.Prefetcher.prefetch" title="Permalink to this definition">#</a></dt>
 <dd><p>If False, returns the input dataset unmodified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -481,7 +542,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 (<code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.experimental.AUTOTUNE</span></code>), this value will be optimized
 automatically to decrease latency.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -501,13 +562,13 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Prefetcher.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L258-L272"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Prefetcher.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L258-L272"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Prefetcher.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with prefetching to maintain a buffer during iteration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any dataset.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with identical elements. Processing that occurs with the
 elements that are produced can be done in parallel (e.g., training on the
 GPU) while new elements are generated from the pipeline.</p>
@@ -519,7 +580,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Preloader">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Preloader</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L276-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Preloader" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Preloader</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L276-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Preloader" title="Permalink to this definition">#</a></dt>
 <dd><p>Preload elements of the underlying dataset to generate in-memory examples.</p>
 <p>This transformer can lead to considerable performance improvements at the cost of
 memory consumption.</p>
@@ -530,7 +591,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <span class="sig-name descname"><span class="pre">examples</span></span><a class="headerlink" href="#sleap.nn.data.dataset_ops.Preloader.examples" title="Permalink to this definition">#</a></dt>
 <dd><p>Stored list of preloaded elements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[Any]</p>
 </dd>
 </dl>
@@ -550,14 +611,14 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Preloader.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L301-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Preloader.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L301-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Preloader.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that generates preloaded elements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> that generates examples as a dictionary of
 tensors. Should not be repeating infinitely.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A dataset that generates the same examples.</p>
 <p>This is similar to prefetching, except that examples are yielded through a
 generator and loaded when this method is called rather than during pipeline
@@ -571,7 +632,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Repeater">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Repeater</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">repeat</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">epochs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L181-L228"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Repeater" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Repeater</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">repeat</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">epochs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L181-L228"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Repeater" title="Permalink to this definition">#</a></dt>
 <dd><p>Repeating transformer for use in pipelines.</p>
 <p>Repeats the underlying elements indefinitely or for a number of “iterations” or
 “epochs”.</p>
@@ -588,7 +649,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <span class="sig-name descname"><span class="pre">repeat</span></span><a class="headerlink" href="#sleap.nn.data.dataset_ops.Repeater.repeat" title="Permalink to this definition">#</a></dt>
 <dd><p>If False, returns the input dataset unmodified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -600,7 +661,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <dd><p>If -1, repeats the input dataset elements infinitely. Otherwise, loops
 through the elements of the input dataset this number of times.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -620,13 +681,13 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Repeater.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L215-L228"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Repeater.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L215-L228"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Repeater.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with repeated loops over the input elements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any dataset.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same keys, but repeated for
 <a class="reference internal" href="#sleap.nn.data.dataset_ops.Repeater.epochs" title="sleap.nn.data.dataset_ops.Repeater.epochs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">epochs</span></code></a> iterations.</p>
 </dd>
@@ -637,7 +698,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Shuffler">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Shuffler</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shuffle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reshuffle_each_iteration</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L14-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Shuffler" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Shuffler</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shuffle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">buffer_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reshuffle_each_iteration</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L14-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Shuffler" title="Permalink to this definition">#</a></dt>
 <dd><p>Shuffling transformer for use in pipelines.</p>
 <p>The input to this transformer should not be repeated or batched (though the latter
 would technically work). Repeating prevents the shuffling from going through “epoch”
@@ -656,7 +717,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <span class="sig-name descname"><span class="pre">shuffle</span></span><a class="headerlink" href="#sleap.nn.data.dataset_ops.Shuffler.shuffle" title="Permalink to this definition">#</a></dt>
 <dd><p>If False, returns the input dataset unmodified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -669,7 +730,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 set too high, it may take a long time to fill the initial buffer, especially
 if it resets every epoch.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -681,7 +742,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 <dd><p>If True, resets the sampling buffer every iteration
 through the underlying dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -701,13 +762,13 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Shuffler.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L53-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Shuffler.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L53-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Shuffler.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with shuffled element order.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any dataset.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same keys, but in a
 shuffled order, if enabled.</p>
 <p>If the input dataset is repeated, this doesn’t really respect epoch
@@ -721,7 +782,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Unbatcher">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Unbatcher</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L162-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Unbatcher" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.dataset_ops.</span></span><span class="sig-name descname"><span class="pre">Unbatcher</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L162-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Unbatcher" title="Permalink to this definition">#</a></dt>
 <dd><p>Unbatching transformer for use in pipelines.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Unbatcher.input_keys">
@@ -737,7 +798,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.dataset_ops.Unbatcher.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/dataset_ops.py#L175-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Unbatcher.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/dataset_ops.py#L175-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.dataset_ops.Unbatcher.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with unbatched elements.</p>
 </dd></dl>
 
@@ -767,6 +828,73 @@ <h1>sleap.nn.data.dataset_ops</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher"><code class="docutils literal notranslate"><span class="pre">Batcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.batch_size"><code class="docutils literal notranslate"><span class="pre">Batcher.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.drop_remainder"><code class="docutils literal notranslate"><span class="pre">Batcher.drop_remainder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.unrag"><code class="docutils literal notranslate"><span class="pre">Batcher.unrag</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Batcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Batcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Batcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Batcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter"><code class="docutils literal notranslate"><span class="pre">LambdaFilter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.filter_fn"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.filter_fn</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.input_keys"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.output_keys"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">LambdaFilter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher"><code class="docutils literal notranslate"><span class="pre">Prefetcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.prefetch"><code class="docutils literal notranslate"><span class="pre">Prefetcher.prefetch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.buffer_size"><code class="docutils literal notranslate"><span class="pre">Prefetcher.buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Prefetcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Prefetcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Prefetcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Prefetcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader"><code class="docutils literal notranslate"><span class="pre">Preloader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.examples"><code class="docutils literal notranslate"><span class="pre">Preloader.examples</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.input_keys"><code class="docutils literal notranslate"><span class="pre">Preloader.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.output_keys"><code class="docutils literal notranslate"><span class="pre">Preloader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Preloader.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Preloader.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater"><code class="docutils literal notranslate"><span class="pre">Repeater</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.repeat"><code class="docutils literal notranslate"><span class="pre">Repeater.repeat</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.epochs"><code class="docutils literal notranslate"><span class="pre">Repeater.epochs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.input_keys"><code class="docutils literal notranslate"><span class="pre">Repeater.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.output_keys"><code class="docutils literal notranslate"><span class="pre">Repeater.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Repeater.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Repeater.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler"><code class="docutils literal notranslate"><span class="pre">Shuffler</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.shuffle"><code class="docutils literal notranslate"><span class="pre">Shuffler.shuffle</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.buffer_size"><code class="docutils literal notranslate"><span class="pre">Shuffler.buffer_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.reshuffle_each_iteration"><code class="docutils literal notranslate"><span class="pre">Shuffler.reshuffle_each_iteration</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.input_keys"><code class="docutils literal notranslate"><span class="pre">Shuffler.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.output_keys"><code class="docutils literal notranslate"><span class="pre">Shuffler.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Shuffler.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Shuffler.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher"><code class="docutils literal notranslate"><span class="pre">Unbatcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.input_keys"><code class="docutils literal notranslate"><span class="pre">Unbatcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.output_keys"><code class="docutils literal notranslate"><span class="pre">Unbatcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.dataset_ops.Unbatcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Unbatcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -785,7 +913,7 @@ <h1>sleap.nn.data.dataset_ops</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.edge_maps.html b/api/sleap.nn.data.edge_maps.html
index 4b9c307fa..eb0942389 100644
--- a/api/sleap.nn.data.edge_maps.html
+++ b/api/sleap.nn.data.edge_maps.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.edge_maps &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.edge_maps &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.edge_maps';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.edge_maps.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,28 @@ <h1>sleap.nn.data.edge_maps</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.skeletons"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.flatten_channels"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.flatten_channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.distance_to_edge"><code class="docutils literal notranslate"><span class="pre">distance_to_edge()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.get_edge_points"><code class="docutils literal notranslate"><span class="pre">get_edge_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_edge_maps"><code class="docutils literal notranslate"><span class="pre">make_edge_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_multi_pafs"><code class="docutils literal notranslate"><span class="pre">make_multi_pafs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_pafs"><code class="docutils literal notranslate"><span class="pre">make_pafs()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +340,11 @@ <h1>sleap.nn.data.edge_maps</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.edge_maps">
-<span id="sleap-nn-data-edge-maps"></span><h1>sleap.nn.data.edge_maps<a class="headerlink" href="#module-sleap.nn.data.edge_maps" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-edge-maps"></span><h1>sleap.nn.data.edge_maps<a class="headerlink" href="#module-sleap.nn.data.edge_maps" title="Permalink to this heading">#</a></h1>
 <p>Transformers for generating edge confidence maps and part affinity fields.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.PartAffinityFieldsGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L241-L359"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flatten_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L241-L359"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate part affinity fields.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.sigma">
@@ -332,7 +354,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 in units of the input image’s grid, i.e., it does not take scaling in
 previous steps into account.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -345,7 +367,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 effectively the reciprocal of the output scale, i.e., increase this to
 generate confidence maps that are smaller than the input images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -357,7 +379,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 <dd><p>List of <a href="#id1"><span class="problematic" id="id2">`</span></a>sleap.Skeleton`s to use for looking up the index of the
 edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[<a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton">sleap.skeleton.Skeleton</a>]]</p>
 </dd>
 </dl>
@@ -370,7 +392,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 [height, width, n_edges, 2]. If True, generated tensors are of shape
 [height, width, n_edges * 2] by flattening the last 2 axes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -390,14 +412,14 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L276-L359"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L276-L359"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys “image”,
 “instances” and “skeleton_inds”.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as
 “part_affinity_fields”.</p>
 <p>The “part_affinity_fields” key will be a tensor of shape
@@ -425,10 +447,10 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.distance_to_edge">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">distance_to_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L16-L79"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.distance_to_edge" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">distance_to_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L16-L79"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.distance_to_edge" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute pairwise distance between points and undirected edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – Tensor of dtype tf.float32 of shape (d_0, …, d_n, 2) where the last
 axis corresponds to x- and y-coordinates. Distances will be broadcast across
@@ -440,7 +462,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 edge.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of dtype tf.float32 of shape (d_0, …, d_n, n_edges) where the first
 axes correspond to the initial dimensions of <code class="xref py py-obj docutils literal notranslate"><span class="pre">points</span></code>, and the last indicates
 the distance of each point to each edge.</p>
@@ -450,10 +472,10 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.get_edge_points">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">get_edge_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L214-L237"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.get_edge_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">get_edge_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L214-L237"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.get_edge_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the points in each instance that form a directed graph.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – A tensor of shape (n_instances, n_nodes, 2) and dtype tf.float32
 containing instance points where the last axis corresponds to (x, y) pixel
@@ -464,7 +486,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 source and destination node indices.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Tuple of (edge_sources, edge_destinations) containing the edge and destination
 points respectively. Both will be tensors of shape (n_instances, n_edges, 2),
 where the last axis corresponds to (x, y) pixel coordinates on the image.</p>
@@ -474,10 +496,10 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.make_edge_maps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_edge_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L82-L116"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_edge_maps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_edge_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L82-L116"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_edge_maps" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate confidence maps for a set of undirected edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>xv</strong> – Sampling grid vector for x-coordinates of shape (grid_width,) and dtype
 tf.float32. This can be generated by
@@ -494,7 +516,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 confidence maps.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A set of confidence maps corresponding to the probability of each point on a
 sampling grid being on each edge. These will be in a tensor of shape
 (grid_height, grid_width, n_edges) of dtype tf.float32.</p>
@@ -504,10 +526,10 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.make_multi_pafs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_sources</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destinations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L165-L211"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_multi_pafs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_multi_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_sources</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destinations</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L165-L211"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_multi_pafs" title="Permalink to this definition">#</a></dt>
 <dd><p>Make multiple instance PAFs with max reduction.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>xv</strong> – Sampling grid vector for x-coordinates of shape (grid_width,) and dtype
 tf.float32. This can be generated by
@@ -525,7 +547,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 the edge maps for masking the PAFs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A set of part affinity fields generated for each instance. These will be in a
 tensor of shape (grid_height, grid_width, n_edges, 2). If multiple instance
 PAFs are defined on the same pixel, they will be summed.</p>
@@ -535,10 +557,10 @@ <h1>sleap.nn.data.edge_maps</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.edge_maps.make_pafs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/edge_maps.py#L119-L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_pafs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.edge_maps.</span></span><span class="sig-name descname"><span class="pre">make_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/edge_maps.py#L119-L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.edge_maps.make_pafs" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate part affinity fields for a set of directed edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>xv</strong> – Sampling grid vector for x-coordinates of shape (grid_width,) and dtype
 tf.float32. This can be generated by
@@ -555,7 +577,7 @@ <h1>sleap.nn.data.edge_maps</h1>
 the edge maps for masking the PAFs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A set of part affinity fields corresponding to the unit vector pointing along
 the direction of each edge weighted by the probability of each point on a
 sampling grid being on each edge. These will be in a tensor of shape
@@ -589,6 +611,34 @@ <h1>sleap.nn.data.edge_maps</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.skeletons"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.flatten_channels"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.flatten_channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.distance_to_edge"><code class="docutils literal notranslate"><span class="pre">distance_to_edge()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.get_edge_points"><code class="docutils literal notranslate"><span class="pre">get_edge_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_edge_maps"><code class="docutils literal notranslate"><span class="pre">make_edge_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_multi_pafs"><code class="docutils literal notranslate"><span class="pre">make_multi_pafs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.edge_maps.make_pafs"><code class="docutils literal notranslate"><span class="pre">make_pafs()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -607,7 +657,7 @@ <h1>sleap.nn.data.edge_maps</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.general.html b/api/sleap.nn.data.general.html
index 5f4df54a9..6e0dfc7cc 100644
--- a/api/sleap.nn.data.general.html
+++ b/api/sleap.nn.data.general.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.general &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.general &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.general';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.general.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,40 @@ <h1>sleap.nn.data.general</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter"><code class="docutils literal notranslate"><span class="pre">KeyFilter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyFilter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyFilter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyFilter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer"><code class="docutils literal notranslate"><span class="pre">KeyRenamer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap"><code class="docutils literal notranslate"><span class="pre">LambdaMap</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.func"><code class="docutils literal notranslate"><span class="pre">LambdaMap.func</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.input_key_names"><code class="docutils literal notranslate"><span class="pre">LambdaMap.input_key_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.output_key_names"><code class="docutils literal notranslate"><span class="pre">LambdaMap.output_key_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.input_keys"><code class="docutils literal notranslate"><span class="pre">LambdaMap.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.output_keys"><code class="docutils literal notranslate"><span class="pre">LambdaMap.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.transform_dataset"><code class="docutils literal notranslate"><span class="pre">LambdaMap.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +352,11 @@ <h1>sleap.nn.data.general</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.general">
-<span id="sleap-nn-data-general"></span><h1>sleap.nn.data.general<a class="headerlink" href="#module-sleap.nn.data.general" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-general"></span><h1>sleap.nn.data.general<a class="headerlink" href="#module-sleap.nn.data.general" title="Permalink to this heading">#</a></h1>
 <p>General purpose transformers for common pipeline processing tasks.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyDeviceMover">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyDeviceMover</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'/cpu:0'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L81-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyDeviceMover" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyDeviceMover</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'/cpu:0'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L81-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyDeviceMover" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for moving example keys to a device.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyDeviceMover.input_keys">
@@ -338,7 +372,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyDeviceMover.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L97-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyDeviceMover.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L97-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyDeviceMover.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains data but moved to cpu.</p>
 </dd></dl>
 
@@ -346,7 +380,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyFilter">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyFilter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keep_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L50-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyFilter" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyFilter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keep_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L50-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyFilter" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for filtering example keys.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyFilter.input_keys">
@@ -362,7 +396,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyFilter.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L65-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyFilter.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L65-L77"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyFilter.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains filtered data.</p>
 </dd></dl>
 
@@ -370,7 +404,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyRenamer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyRenamer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">drop_old</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L9-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyRenamer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">KeyRenamer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">drop_old</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L9-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyRenamer" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for renaming example keys.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyRenamer.input_keys">
@@ -386,7 +420,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.general.KeyRenamer.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L29-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyRenamer.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L29-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.KeyRenamer.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains filtered data.</p>
 </dd></dl>
 
@@ -394,7 +428,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.general.LambdaMap">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">LambdaMap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">func</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L117-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.LambdaMap" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.general.</span></span><span class="sig-name descname"><span class="pre">LambdaMap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">func</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_key_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L117-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.LambdaMap" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for mapping an arbitrary function to the dataset.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.general.LambdaMap.func">
@@ -402,7 +436,7 @@ <h1>sleap.nn.data.general</h1>
 <dd><p>A callable of the form <code class="xref py py-obj docutils literal notranslate"><span class="pre">func(example)</span> <span class="pre">-&gt;</span> <span class="pre">example</span></code>, where the input and
 output are each a dictionary of tensors.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Callable[[Dict[str, tensorflow.python.framework.ops.Tensor]], Dict[str, tensorflow.python.framework.ops.Tensor]]</p>
 </dd>
 </dl>
@@ -414,7 +448,7 @@ <h1>sleap.nn.data.general</h1>
 <dd><p>List of input key names that the function expects to find in
 the example.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -425,7 +459,7 @@ <h1>sleap.nn.data.general</h1>
 <span class="sig-name descname"><span class="pre">output_key_names</span></span><a class="headerlink" href="#sleap.nn.data.general.LambdaMap.output_key_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of output key names that the function will return.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -445,7 +479,7 @@ <h1>sleap.nn.data.general</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.general.LambdaMap.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/general.py#L142-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.LambdaMap.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/general.py#L142-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.general.LambdaMap.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains transformed data.</p>
 </dd></dl>
 
@@ -475,6 +509,46 @@ <h1>sleap.nn.data.general</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyDeviceMover.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyDeviceMover.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter"><code class="docutils literal notranslate"><span class="pre">KeyFilter</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyFilter.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyFilter.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyFilter.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyFilter.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer"><code class="docutils literal notranslate"><span class="pre">KeyRenamer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.KeyRenamer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyRenamer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap"><code class="docutils literal notranslate"><span class="pre">LambdaMap</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.func"><code class="docutils literal notranslate"><span class="pre">LambdaMap.func</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.input_key_names"><code class="docutils literal notranslate"><span class="pre">LambdaMap.input_key_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.output_key_names"><code class="docutils literal notranslate"><span class="pre">LambdaMap.output_key_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.input_keys"><code class="docutils literal notranslate"><span class="pre">LambdaMap.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.output_keys"><code class="docutils literal notranslate"><span class="pre">LambdaMap.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.general.LambdaMap.transform_dataset"><code class="docutils literal notranslate"><span class="pre">LambdaMap.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -493,7 +567,7 @@ <h1>sleap.nn.data.general</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.grouping.html b/api/sleap.nn.data.grouping.html
index 4679c5fcf..807ccb5e8 100644
--- a/api/sleap.nn.data.grouping.html
+++ b/api/sleap.nn.data.grouping.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.grouping &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.grouping &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.grouping';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.grouping.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,15 @@ <h1>sleap.nn.data.grouping</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.grouping.group_examples"><code class="docutils literal notranslate"><span class="pre">group_examples()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.grouping.group_examples_iter"><code class="docutils literal notranslate"><span class="pre">group_examples_iter()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,18 +327,18 @@ <h1>sleap.nn.data.grouping</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.grouping">
-<span id="sleap-nn-data-grouping"></span><h1>sleap.nn.data.grouping<a class="headerlink" href="#module-sleap.nn.data.grouping" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-grouping"></span><h1>sleap.nn.data.grouping<a class="headerlink" href="#module-sleap.nn.data.grouping" title="Permalink to this heading">#</a></h1>
 <p>Group inference results (“examples”) by frame.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.grouping.group_examples">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.grouping.</span></span><span class="sig-name descname"><span class="pre">group_examples</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">examples</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/grouping.py#L8-L19"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.grouping.group_examples" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.grouping.</span></span><span class="sig-name descname"><span class="pre">group_examples</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">examples</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/grouping.py#L8-L19"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.grouping.group_examples" title="Permalink to this definition">#</a></dt>
 <dd><p>Group examples into dictionary.</p>
 <p>Key is (video_ind, frame_ind), val is list of examples matching key.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.grouping.group_examples_iter">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.grouping.</span></span><span class="sig-name descname"><span class="pre">group_examples_iter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">examples</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/grouping.py#L22-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.grouping.group_examples_iter" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.grouping.</span></span><span class="sig-name descname"><span class="pre">group_examples_iter</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">examples</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/grouping.py#L22-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.grouping.group_examples_iter" title="Permalink to this definition">#</a></dt>
 <dd><p>Iterator which groups examples.</p>
 <p>Yields ((video_ind, frame_ind), list of examples matching vid/frame).</p>
 </dd></dl>
@@ -358,6 +367,21 @@ <h1>sleap.nn.data.grouping</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.grouping.group_examples"><code class="docutils literal notranslate"><span class="pre">group_examples()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.grouping.group_examples_iter"><code class="docutils literal notranslate"><span class="pre">group_examples_iter()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -376,7 +400,7 @@ <h1>sleap.nn.data.grouping</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.identity.html b/api/sleap.nn.data.identity.html
index 21bd719ad..497a800f0 100644
--- a/api/sleap.nn.data.identity.html
+++ b/api/sleap.nn.data.identity.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.identity &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.identity &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.identity';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.identity.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,31 @@ <h1>sleap.nn.data.identity</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.centroids"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.centroids</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.class_map_threshold"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.class_map_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.make_class_maps"><code class="docutils literal notranslate"><span class="pre">make_class_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.make_class_vectors"><code class="docutils literal notranslate"><span class="pre">make_class_vectors()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +343,11 @@ <h1>sleap.nn.data.identity</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.identity">
-<span id="sleap-nn-data-identity"></span><h1>sleap.nn.data.identity<a class="headerlink" href="#module-sleap.nn.data.identity" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-identity"></span><h1>sleap.nn.data.identity<a class="headerlink" href="#module-sleap.nn.data.identity" title="Permalink to this heading">#</a></h1>
 <p>Utilities for generating data for track identity models.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassMapGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">ClassMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_map_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L115-L202"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassMapGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">ClassMapGenerator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_map_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L115-L202"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassMapGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate class maps from track indices.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassMapGenerator.sigma">
@@ -332,7 +357,7 @@ <h1>sleap.nn.data.identity</h1>
 units of the input image’s grid, i.e., it does not take scaling in previous
 steps into account.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -345,7 +370,7 @@ <h1>sleap.nn.data.identity</h1>
 reciprocal of the output scale, i.e., increase this to generate maps that
 are smaller than the input images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -357,7 +382,7 @@ <h1>sleap.nn.data.identity</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, generate masking confidence maps for centroids rather than
 instance points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -369,7 +394,7 @@ <h1>sleap.nn.data.identity</h1>
 <dd><p>Minimum confidence map value below which map values will be
 replaced with zeros.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -389,15 +414,15 @@ <h1>sleap.nn.data.identity</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassMapGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L150-L202"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L150-L202"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassMapGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated class identity maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code>,
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;track_inds&quot;</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;n_tracks&quot;</span></code> and either <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instances&quot;</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;</span></code>
 depending on whether the <a class="reference internal" href="#sleap.nn.data.identity.ClassMapGenerator.centroids" title="sleap.nn.data.identity.ClassMapGenerator.centroids"><code class="xref py py-obj docutils literal notranslate"><span class="pre">centroids</span></code></a> attribute is set to <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;class_maps&quot;</span></code> key containing the generated class maps.</p>
 </dd>
@@ -408,7 +433,7 @@ <h1>sleap.nn.data.identity</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassVectorGenerator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">ClassVectorGenerator</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L74-L111"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassVectorGenerator" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">ClassVectorGenerator</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L74-L111"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassVectorGenerator" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to generate class probability vectors from track indices.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassVectorGenerator.input_keys">
@@ -424,14 +449,14 @@ <h1>sleap.nn.data.identity</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.ClassVectorGenerator.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L87-L111"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassVectorGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L87-L111"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.ClassVectorGenerator.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains the generated class identity vectors.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_ds</strong> – A dataset with elements that contain the keys`”track_inds”` and
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;n_tracks&quot;</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with the same keys as the input, as well as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;class&quot;</span></code>
 key containing the generated class vectors.</p>
 </dd>
@@ -442,12 +467,12 @@ <h1>sleap.nn.data.identity</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.make_class_maps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">make_class_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L26-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.make_class_maps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">make_class_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L26-L70"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.make_class_maps" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate identity class maps using instance-wise confidence maps.</p>
 <p>This is useful for making class maps defined on local neighborhoods around the
 peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>confmaps</strong> – Confidence maps for the same points as the offset maps as a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_instances)</span></code> and dtype
@@ -459,7 +484,7 @@ <h1>sleap.nn.data.identity</h1>
 with zeros.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The class maps with shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_classes)</span></code> and dtype
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> where each channel will be a binary mask with 1 where the instance
 confidence maps were higher than the threshold.</p>
@@ -475,17 +500,17 @@ <h1>sleap.nn.data.identity</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.identity.make_class_vectors">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">make_class_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">class_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/identity.py#L9-L23"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.make_class_vectors" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.identity.</span></span><span class="sig-name descname"><span class="pre">make_class_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">class_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/identity.py#L9-L23"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.identity.make_class_vectors" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a binary class vectors from class indices.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>class_inds</strong> – Class indices as <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code> and shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,)</span></code>. Indices of <code class="xref py py-obj docutils literal notranslate"><span class="pre">-1</span></code> will be interpreted as having no class.</p></li>
 <li><p><strong>n_classes</strong> – Integer number of maximum classes.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor with binary class vectors of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_classes)</span></code> of dtype
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>. Instances with no class will have all zeros in their row.</p>
 </dd>
@@ -517,6 +542,37 @@ <h1>sleap.nn.data.identity</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.centroids"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.centroids</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.class_map_threshold"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.class_map_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassMapGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">ClassMapGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.input_keys"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.output_keys"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.ClassVectorGenerator.transform_dataset"><code class="docutils literal notranslate"><span class="pre">ClassVectorGenerator.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.make_class_maps"><code class="docutils literal notranslate"><span class="pre">make_class_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.identity.make_class_vectors"><code class="docutils literal notranslate"><span class="pre">make_class_vectors()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -535,7 +591,7 @@ <h1>sleap.nn.data.identity</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.inference.html b/api/sleap.nn.data.inference.html
index fb8ab65b6..0d3efb0e5 100644
--- a/api/sleap.nn.data.inference.html
+++ b/api/sleap.nn.data.inference.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.inference &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.inference &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.inference';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.inference.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,23 @@ <h1>sleap.nn.data.inference</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.GlobalPeakFinder"><code class="docutils literal notranslate"><span class="pre">GlobalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.KerasModelPredictor"><code class="docutils literal notranslate"><span class="pre">KerasModelPredictor</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.LocalPeakFinder"><code class="docutils literal notranslate"><span class="pre">LocalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.MockGlobalPeakFinder"><code class="docutils literal notranslate"><span class="pre">MockGlobalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.input_keys"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.output_keys"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,35 +335,35 @@ <h1>sleap.nn.data.inference</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.inference">
-<span id="sleap-nn-data-inference"></span><h1>sleap.nn.data.inference<a class="headerlink" href="#module-sleap.nn.data.inference" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-inference"></span><h1>sleap.nn.data.inference<a class="headerlink" href="#module-sleap.nn.data.inference" title="Permalink to this heading">#</a></h1>
 <p>Transformers for performing inference.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.GlobalPeakFinder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">GlobalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L75-L141"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.GlobalPeakFinder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">GlobalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L75-L141"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.GlobalPeakFinder" title="Permalink to this definition">#</a></dt>
 <dd><p>Global peak finding transformer.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.KerasModelPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">KerasModelPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_input_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instance_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_output_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L18-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.KerasModelPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">KerasModelPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_input_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instance_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_output_keys</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L18-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.KerasModelPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for performing tf.keras model inference.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.LocalPeakFinder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">LocalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroids'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_sample_inds'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_channel_inds'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L204-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.LocalPeakFinder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">LocalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroids'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_sample_inds'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_channel_inds'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L204-L280"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.LocalPeakFinder" title="Permalink to this definition">#</a></dt>
 <dd><p>Local peak finding transformer.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.MockGlobalPeakFinder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">MockGlobalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_peaks_in_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_out_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_in_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_out_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L145-L200"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.MockGlobalPeakFinder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">MockGlobalPeakFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">all_peaks_in_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_out_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_in_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instance_confidence_maps'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_out_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidence_maps'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L145-L200"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.MockGlobalPeakFinder" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer that mimics <a class="reference internal" href="#sleap.nn.data.inference.GlobalPeakFinder" title="sleap.nn.data.inference.GlobalPeakFinder"><code class="xref py py-obj docutils literal notranslate"><span class="pre">GlobalPeakFinder</span></code></a> but passes ground truth data.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.PredictedCenterInstanceNormalizer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">PredictedCenterInstanceNormalizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_confidence_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid_confidence'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_confidences_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_centroid_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_centroid_confidence_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_confidence'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_peak_confidences_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidences'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L284-L348"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.inference.</span></span><span class="sig-name descname"><span class="pre">PredictedCenterInstanceNormalizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_confidence_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroid_confidence'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_points'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_confidences_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_center_instance_confidences'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_centroid_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_centroid_confidence_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_centroid_confidence'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_peaks_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_peak_confidences_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instance_confidences'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L284-L348"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer for adjusting centered instance coordinates.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.PredictedCenterInstanceNormalizer.input_keys">
@@ -362,7 +379,7 @@ <h1>sleap.nn.data.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/inference.py#L320-L348"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/inference.py#L320-L348"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains instance cropped data.</p>
 </dd></dl>
 
@@ -392,6 +409,29 @@ <h1>sleap.nn.data.inference</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.GlobalPeakFinder"><code class="docutils literal notranslate"><span class="pre">GlobalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.KerasModelPredictor"><code class="docutils literal notranslate"><span class="pre">KerasModelPredictor</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.LocalPeakFinder"><code class="docutils literal notranslate"><span class="pre">LocalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.MockGlobalPeakFinder"><code class="docutils literal notranslate"><span class="pre">MockGlobalPeakFinder</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.input_keys"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.output_keys"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PredictedCenterInstanceNormalizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -410,7 +450,7 @@ <h1>sleap.nn.data.inference</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.instance_centroids.html b/api/sleap.nn.data.instance_centroids.html
index d38ca6795..365945a8a 100644
--- a/api/sleap.nn.data.instance_centroids.html
+++ b/api/sleap.nn.data.instance_centroids.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.instance_centroids &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.instance_centroids &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.instance_centroids';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.instance_centroids.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.nn.data.instance_centroids</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.center_on_anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.anchor_part_names"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.anchor_part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.skeletons"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.instances_key"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.find_points_bbox_midpoint"><code class="docutils literal notranslate"><span class="pre">find_points_bbox_midpoint()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.get_instance_anchors"><code class="docutils literal notranslate"><span class="pre">get_instance_anchors()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +338,11 @@ <h1>sleap.nn.data.instance_centroids</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.instance_centroids">
-<span id="sleap-nn-data-instance-centroids"></span><h1>sleap.nn.data.instance_centroids<a class="headerlink" href="#module-sleap.nn.data.instance_centroids" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-instance-centroids"></span><h1>sleap.nn.data.instance_centroids<a class="headerlink" href="#module-sleap.nn.data.instance_centroids" title="Permalink to this heading">#</a></h1>
 <p>Transformers for finding instance centroids.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_centroids.InstanceCentroidFinder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">InstanceCentroidFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">center_on_anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_centroids.py#L55-L195"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">InstanceCentroidFinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">center_on_anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_centroids.py#L55-L195"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer to add centroid information to instances.</p>
 <p>This is useful as a transformation to data streams that will be used in centroid
 networks or for instance cropping.</p>
@@ -333,7 +353,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 to a body part rather than the midpoint of the instance bounding box. If
 False, the midpoint of the bounding box of all points will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -346,7 +366,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 skeleton to use as anchors for centering. If <a class="reference internal" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part" title="sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part"><code class="xref py py-obj docutils literal notranslate"><span class="pre">center_on_anchor_part</span></code></a> is
 False, this has no effect and does not need to be specified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[str]]</p>
 </dd>
 </dl>
@@ -359,7 +379,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 <span class="pre">anchor</span> <span class="pre">body</span> <span class="pre">parts.</span> <span class="pre">If</span> <span class="pre">`center_on_anchor_part</span></code> is False, this has no effect
 and does not need to be specified.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[<a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton">sleap.skeleton.Skeleton</a>]]</p>
 </dd>
 </dl>
@@ -371,7 +391,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 <dd><p>Name of the example key where the instance points are stored.
 Defaults to “instances”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -379,23 +399,23 @@ <h1>sleap.nn.data.instance_centroids</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">sleap.nn.config.data.InstanceCroppingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder" title="sleap.nn.data.instance_centroids.InstanceCentroidFinder"><span class="pre">sleap.nn.data.instance_centroids.InstanceCentroidFinder</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_centroids.py#L84-L122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">InstanceCroppingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder" title="sleap.nn.data.instance_centroids.InstanceCentroidFinder"><span class="pre">InstanceCentroidFinder</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_centroids.py#L84-L122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Build an instance of this class from its configuration options.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">InstanceCroppingConfig</span></code> instance with the desired parameters.</p></li>
 <li><p><strong>skeletons</strong> – List of skeletons to use. This must be provided if doing instance
 cropping centered on an anchor part.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An instance of this class.</p>
 <p>This will assume that <a class="reference internal" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part" title="sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part"><code class="xref py py-obj docutils literal notranslate"><span class="pre">center_on_anchor_part</span></code></a> is False when the
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">config.center_on_part</span></code> attribute is not a string.</p>
 </p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – If the skeletons are not provided in the arguments and the
     config specifies an anchor part name.</p>
 </dd>
@@ -416,10 +436,10 @@ <h1>sleap.nn.data.instance_centroids</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_centroids.py#L137-L195"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_centroids.py#L137-L195"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains centroids computed from the inputs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – A dataset with “instances” key containing instance points in a
 tf.float32 tensor of shape (n_instances, n_nodes, 2). If centering on
 anchor parts, a “skeleton_inds” key of dtype tf.int32 and shape
@@ -427,7 +447,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 associated with each instance. These must match the order in the
 <a class="reference internal" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.skeletons" title="sleap.nn.data.instance_centroids.InstanceCentroidFinder.skeletons"><code class="xref py py-obj docutils literal notranslate"><span class="pre">skeletons</span></code></a> attribute of this class.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing a “centroids” key containing
 a tf.float32 tensor of shape (n_instances, 2) with the computed centroids.</p>
 </dd>
@@ -438,14 +458,14 @@ <h1>sleap.nn.data.instance_centroids</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_centroids.find_points_bbox_midpoint">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">find_points_bbox_midpoint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_centroids.py#L12-L33"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.find_points_bbox_midpoint" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">find_points_bbox_midpoint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_centroids.py#L12-L33"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.find_points_bbox_midpoint" title="Permalink to this definition">#</a></dt>
 <dd><p>Find the midpoint of the bounding box of a set of points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>instances</strong> – A tf.Tensor of dtype tf.float32 and of shape (…, n_points, 2),
 i.e., rank &gt;= 2.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The midpoints between the bounds of each set of points. The output will be of
 shape (…, 2), reducing the rank of the input by 1. NaNs will be ignored in the
 calculation.</p>
@@ -467,10 +487,10 @@ <h1>sleap.nn.data.instance_centroids</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_centroids.get_instance_anchors">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">get_instance_anchors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_centroids.py#L36-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.get_instance_anchors" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_centroids.</span></span><span class="sig-name descname"><span class="pre">get_instance_anchors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_centroids.py#L36-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_centroids.get_instance_anchors" title="Permalink to this definition">#</a></dt>
 <dd><p>Gather the anchor points of a set of instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances</strong> – A tensor of shape (n_instances, n_nodes, 2) containing instance
 points. This must be rank-3 even if a single instance is present.</p></li>
@@ -478,7 +498,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
 the index of the anchor node for each instance.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of shape (n_instances, 2) containing the anchor points for each
 each instance. This is basically a slice along the nodes axis, where each
 instance may potentially have a different node to use as an anchor.</p>
@@ -510,6 +530,32 @@ <h1>sleap.nn.data.instance_centroids</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.center_on_anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.anchor_part_names"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.anchor_part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.skeletons"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.skeletons</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.instances_key"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceCentroidFinder.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.find_points_bbox_midpoint"><code class="docutils literal notranslate"><span class="pre">find_points_bbox_midpoint()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_centroids.get_instance_anchors"><code class="docutils literal notranslate"><span class="pre">get_instance_anchors()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -528,7 +574,7 @@ <h1>sleap.nn.data.instance_centroids</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.instance_cropping.html b/api/sleap.nn.data.instance_cropping.html
index 2e1d7fd6b..770c297e1 100644
--- a/api/sleap.nn.data.instance_cropping.html
+++ b/api/sleap.nn.data.instance_cropping.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.instance_cropping &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.instance_cropping &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.instance_cropping';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.instance_cropping.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,33 @@ <h1>sleap.nn.data.instance_cropping</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper"><code class="docutils literal notranslate"><span class="pre">InstanceCropper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_width"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_height"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.keep_full_image"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.mock_centroid_confidence"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.mock_centroid_confidence</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.unbatch"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.unbatch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.image_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.instances_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.centroids_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.centroids_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.from_config"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.crop_bboxes"><code class="docutils literal notranslate"><span class="pre">crop_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.find_instance_crop_size"><code class="docutils literal notranslate"><span class="pre">find_instance_crop_size()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.make_centered_bboxes"><code class="docutils literal notranslate"><span class="pre">make_centered_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.normalize_bboxes"><code class="docutils literal notranslate"><span class="pre">normalize_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.unnormalize_bboxes"><code class="docutils literal notranslate"><span class="pre">unnormalize_bboxes()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +345,11 @@ <h1>sleap.nn.data.instance_cropping</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.instance_cropping">
-<span id="sleap-nn-data-instance-cropping"></span><h1>sleap.nn.data.instance_cropping<a class="headerlink" href="#module-sleap.nn.data.instance_cropping" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-instance-cropping"></span><h1>sleap.nn.data.instance_cropping<a class="headerlink" href="#module-sleap.nn.data.instance_cropping" title="Permalink to this heading">#</a></h1>
 <p>Transformers for cropping instances for topdown processing.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.InstanceCropper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">InstanceCropper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mock_centroid_confidence</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unbatch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroids'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L225-L458"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">InstanceCropper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">crop_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mock_centroid_confidence</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unbatch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroids_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'centroids'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L225-L458"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer to crop and generate individual examples for instances.</p>
 <p>This generates datasets that are instance cropped for topdown processing.</p>
 <dl class="py attribute">
@@ -330,7 +357,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <span class="sig-name descname"><span class="pre">crop_width</span></span><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_width" title="Permalink to this definition">#</a></dt>
 <dd><p>Width of the crops in pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -341,7 +368,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <span class="sig-name descname"><span class="pre">crop_height</span></span><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_height" title="Permalink to this definition">#</a></dt>
 <dd><p>Height of the crops in pixels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -356,7 +383,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 requirements usage. Setting this to False can substantially improve
 performance of large pipelines if the full images are no longer required.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -368,7 +395,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <dd><p>If True, add confidence keys for compatibility with
 predicted instance cropping.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -382,7 +409,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 the same example. Use this when building pipelines that require knowledge
 about all instances within a single example.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -394,7 +421,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <dd><p>Name of the example key where the image is stored. Defaults to
 “image”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -406,7 +433,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <dd><p>Name of the example key where the instance points are stored.
 Defaults to “instances”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -418,7 +445,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <dd><p>Name of the example key where the instance centroids are stored.
 Defaults to “centroids”.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -426,10 +453,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.InstanceCropper.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">sleap.nn.config.data.InstanceCroppingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.instance_cropping.InstanceCropper" title="sleap.nn.data.instance_cropping.InstanceCropper"><span class="pre">sleap.nn.data.instance_cropping.InstanceCropper</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L261-L288"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.InstanceCroppingConfig" title="sleap.nn.config.data.InstanceCroppingConfig"><span class="pre">InstanceCroppingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.instance_cropping.InstanceCropper" title="sleap.nn.data.instance_cropping.InstanceCropper"><span class="pre">InstanceCropper</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L261-L288"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Build an instance of this class from its configuration options.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">InstanceCroppingConfig</span></code> instance with the desired parameters.</p></li>
 <li><p><strong>crop_size</strong> – Integer specifying the crop height and width. This is only
@@ -437,10 +464,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 not specify an explicit integer crop size (e.g., it is set to None).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class.</p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – If the <code class="xref py py-obj docutils literal notranslate"><span class="pre">crop_size</span></code> is not specified in either the config
     attribute or function arguments.</p>
 </dd>
@@ -461,10 +488,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L315-L458"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L315-L458"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains instance cropped data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – <p>A dataset with examples containing the following keys:
 “image”: The full image in a tensor of shape (height, width, channels).
 “instances”: Instance points in a tf.float32 tensor of shape</p>
@@ -479,7 +506,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 Any additional keys present will be replicated in each output.</p>
 </p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing instance cropped data. Each
 instance will generate an example, so the total number of elements may
 change relative to the input dataset.</p>
@@ -534,13 +561,13 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.crop_bboxes">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">crop_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L169-L221"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.crop_bboxes" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">crop_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L169-L221"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.crop_bboxes" title="Permalink to this definition">#</a></dt>
 <dd><p>Crop bounding boxes from an image.</p>
 <p>This method serves as a convenience method for specifying the arguments of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.image.crop_and_resize</span></code>, becoming especially useful in the case of multiple
 bounding boxes with a single image and no resizing.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – Tensor of shape (height, width, channels) of a single image.</p></li>
 <li><p><strong>bboxes</strong> – Tensor of shape (n_bboxes, 4) and dtype tf.float32, where the last axis
@@ -548,7 +575,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 be generated from centroids using <a class="reference internal" href="#sleap.nn.data.instance_cropping.make_centered_bboxes" title="sleap.nn.data.instance_cropping.make_centered_bboxes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">make_centered_bboxes</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of shape (n_bboxes, crop_height, crop_width, channels) of the same
 dtype as the input image. The crop size is inferred from the bounding box
 coordinates.</p>
@@ -568,10 +595,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.find_instance_crop_size">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">find_instance_crop_size</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">padding</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">maximum_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_scaling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L11-L55"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.find_instance_crop_size" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">find_instance_crop_size</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">padding</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">maximum_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_scaling</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L11-L55"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.find_instance_crop_size" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the size of the largest instance bounding box from labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> containing user-labeled instances.</p></li>
 <li><p><strong>padding</strong> – Integer number of pixels to add to the bounds as margin padding.</p></li>
@@ -583,7 +610,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <li><p><strong>min_crop_size</strong> – The (optional) crop size set by the user. None if not set.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An integer crop size denoting the length of the side of the bounding boxes that
 will contain the instances when cropped. The returned crop size will be larger
 or equal to the input <code class="xref py py-obj docutils literal notranslate"><span class="pre">crop_size</span></code>.</p>
@@ -595,10 +622,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.make_centered_bboxes">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">make_centered_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L124-L166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.make_centered_bboxes" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">make_centered_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroids</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">box_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L124-L166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.make_centered_bboxes" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate bounding boxes centered on a set of centroid coordinates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>centroids</strong> – A tensor of shape (n_centroids, 2) and dtype tf.float32, where the
 last axis corresponds to the (x, y) coordinates of each centroid.</p></li>
@@ -606,7 +633,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <li><p><strong>box_width</strong> – Scalar integer indicating the width of the bounding boxes.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Tensor of shape (n_centroids, 4) and dtype tf.float32, where the last axis
 corresponds to (y1, x1, y2, x2) coordinates of the bounding boxes in absolute
 image coordinates.</p>
@@ -641,12 +668,12 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.normalize_bboxes">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">normalize_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L58-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.normalize_bboxes" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">normalize_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L58-L90"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.normalize_bboxes" title="Permalink to this definition">#</a></dt>
 <dd><p>Normalize bounding box coordinates to the range [0, 1].</p>
 <p>This is useful for transforming points for TensorFlow operations that require
 normalized image coordinates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>bboxes</strong> – Tensor of shape (n_bboxes, 4) and dtype tf.float32, where the last axis
 corresponds to (y1, x1, y2, x2) coordinates of the bounding boxes.</p></li>
@@ -654,7 +681,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <li><p><strong>image_width</strong> – Scalar integer indicating the width of the image.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Tensor of the normalized points of the same shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">bboxes</span></code>.</p>
 <p>The normalization applied to each point is <code class="xref py py-obj docutils literal notranslate"><span class="pre">x</span> <span class="pre">/</span> <span class="pre">(image_width</span> <span class="pre">-</span> <span class="pre">1)</span></code> and
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">y</span> <span class="pre">/</span> <span class="pre">(image_width</span> <span class="pre">-</span> <span class="pre">1)</span></code>.</p>
@@ -666,10 +693,10 @@ <h1>sleap.nn.data.instance_cropping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.instance_cropping.unnormalize_bboxes">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">unnormalize_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">normalized_bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/instance_cropping.py#L93-L121"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.unnormalize_bboxes" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.instance_cropping.</span></span><span class="sig-name descname"><span class="pre">unnormalize_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">normalized_bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/instance_cropping.py#L93-L121"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.instance_cropping.unnormalize_bboxes" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert bounding boxes coordinates in the range [0, 1] to absolute coordinates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>normalized_bboxes</strong> – Tensor of shape (n_bboxes, 4) and dtype tf.float32, where the
 last axis corresponds to (y1, x1, y2, x2) normalized coordinates of the
@@ -678,7 +705,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
 <li><p><strong>image_width</strong> – Scalar integer indicating the width of the image.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Tensor of the same shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">bboxes</span></code> mapped back to absolute image coordinates
 by multiplying (x, y) coordinates by <code class="xref py py-obj docutils literal notranslate"><span class="pre">(image_width</span> <span class="pre">-</span> <span class="pre">1,</span> <span class="pre">image_height</span> <span class="pre">-</span> <span class="pre">1)</span></code>.</p>
 </dd>
@@ -710,6 +737,39 @@ <h1>sleap.nn.data.instance_cropping</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper"><code class="docutils literal notranslate"><span class="pre">InstanceCropper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_width"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.crop_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.crop_height"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.crop_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.keep_full_image"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.mock_centroid_confidence"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.mock_centroid_confidence</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.unbatch"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.unbatch</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.image_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.instances_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.instances_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.centroids_key"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.centroids_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.from_config"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.input_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.output_keys"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">InstanceCropper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.crop_bboxes"><code class="docutils literal notranslate"><span class="pre">crop_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.find_instance_crop_size"><code class="docutils literal notranslate"><span class="pre">find_instance_crop_size()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.make_centered_bboxes"><code class="docutils literal notranslate"><span class="pre">make_centered_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.normalize_bboxes"><code class="docutils literal notranslate"><span class="pre">normalize_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.instance_cropping.unnormalize_bboxes"><code class="docutils literal notranslate"><span class="pre">unnormalize_bboxes()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -728,7 +788,7 @@ <h1>sleap.nn.data.instance_cropping</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.normalization.html b/api/sleap.nn.data.normalization.html
index ba8ba9431..accd8c8a7 100644
--- a/api/sleap.nn.data.normalization.html
+++ b/api/sleap.nn.data.normalization.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.normalization &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.normalization &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.normalization';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.normalization.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,35 @@ <h1>sleap.nn.data.normalization</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer"><code class="docutils literal notranslate"><span class="pre">Normalizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.image_key"><code class="docutils literal notranslate"><span class="pre">Normalizer.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_float"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_rgb</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.imagenet_mode"><code class="docutils literal notranslate"><span class="pre">Normalizer.imagenet_mode</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.from_config"><code class="docutils literal notranslate"><span class="pre">Normalizer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.input_keys"><code class="docutils literal notranslate"><span class="pre">Normalizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.output_keys"><code class="docutils literal notranslate"><span class="pre">Normalizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Normalizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.convert_rgb_to_bgr"><code class="docutils literal notranslate"><span class="pre">convert_rgb_to_bgr()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_float"><code class="docutils literal notranslate"><span class="pre">ensure_float()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">ensure_grayscale()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_int"><code class="docutils literal notranslate"><span class="pre">ensure_int()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_min_image_rank"><code class="docutils literal notranslate"><span class="pre">ensure_min_image_rank()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">ensure_rgb()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_image_range"><code class="docutils literal notranslate"><span class="pre">scale_image_range()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_caffe_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_caffe_mode()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_tf_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_tf_mode()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_torch_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_torch_mode()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +347,11 @@ <h1>sleap.nn.data.normalization</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.normalization">
-<span id="sleap-nn-data-normalization"></span><h1>sleap.nn.data.normalization<a class="headerlink" href="#module-sleap.nn.data.normalization" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-normalization"></span><h1>sleap.nn.data.normalization<a class="headerlink" href="#module-sleap.nn.data.normalization" title="Permalink to this heading">#</a></h1>
 <p>Transformers for normalizing data formats.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.Normalizer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">Normalizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_float</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_rgb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">imagenet_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L269-L382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">Normalizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_float</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_rgb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_grayscale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">imagenet_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L269-L382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer to normalize images.</p>
 <p>This is useful as a transformation to data streams that require specific data ranges
 such as for pretrained models with specific preprocessing constraints.</p>
@@ -331,7 +360,7 @@ <h1>sleap.nn.data.normalization</h1>
 <span class="sig-name descname"><span class="pre">image_key</span></span><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the images to normalize.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -342,7 +371,7 @@ <h1>sleap.nn.data.normalization</h1>
 <span class="sig-name descname"><span class="pre">ensure_float</span></span><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.ensure_float" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, converts the image to a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> if not already.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -353,7 +382,7 @@ <h1>sleap.nn.data.normalization</h1>
 <span class="sig-name descname"><span class="pre">ensure_rgb</span></span><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.ensure_rgb" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, converts the image to RGB if not already.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -364,7 +393,7 @@ <h1>sleap.nn.data.normalization</h1>
 <span class="sig-name descname"><span class="pre">ensure_grayscale</span></span><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.ensure_grayscale" title="Permalink to this definition">#</a></dt>
 <dd><p>If True, converts the image to grayscale if not already.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -386,7 +415,7 @@ <h1>sleap.nn.data.normalization</h1>
 </dd>
 </dl>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -394,16 +423,16 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.Normalizer.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">sleap.nn.config.data.PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.normalization.Normalizer" title="sleap.nn.data.normalization.Normalizer"><span class="pre">sleap.nn.data.normalization.Normalizer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L301-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.normalization.Normalizer" title="sleap.nn.data.normalization.Normalizer"><span class="pre">Normalizer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L301-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Build an instance of this class from its configuration options.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">PreprocessingConfig</span></code> instance with the desired parameters.</p></li>
 <li><p><strong>image_key</strong> – String name of the key containing the images to normalize.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class.</p>
 </dd>
 </dl>
@@ -423,13 +452,13 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.Normalizer.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L332-L382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L332-L382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.Normalizer.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains centroids computed from the inputs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – A dataset with image key specified in the <a class="reference internal" href="#sleap.nn.data.normalization.Normalizer.image_key" title="sleap.nn.data.normalization.Normalizer.image_key"><code class="xref py py-obj docutils literal notranslate"><span class="pre">image_key</span></code></a> attribute.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same images with
 normalization applied.</p>
 </dd>
@@ -440,14 +469,14 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.convert_rgb_to_bgr">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">convert_rgb_to_bgr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L117-L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.convert_rgb_to_bgr" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">convert_rgb_to_bgr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L117-L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.convert_rgb_to_bgr" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert an RGB image to BGR format by reversing the channel order.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any dtype with shape (…, 3) in RGB format. If grayscale, the
 image will be converted to RGB first.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The input image with the channels axis reversed.</p>
 </dd>
 </dl>
@@ -455,13 +484,13 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.ensure_float">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_float</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L34-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_float" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_float</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L34-L49"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_float" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert the image to a tf.float32.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any dtype.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tensor of the same shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">image</span></code> but with dtype tf.float32. If the image
 was already of tf.float32 dtype, it will not be changed.</p>
 <p>If the input was of an integer type, it will be scaled to the range [0, 1]
@@ -474,14 +503,14 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.ensure_grayscale">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_grayscale</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L81-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_grayscale" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_grayscale</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L81-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_grayscale" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert image to grayscale if in RGB format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any dtype of shape (height, width, channels). Channels are
 expected to be 1 or 3.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A grayscale image of shape (height, width, 1) of the same dtype as the input.</p>
 </dd>
 </dl>
@@ -490,15 +519,15 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.ensure_int">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_int</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L52-L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_int" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_int</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L52-L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_int" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert the image to a tf.uint8.</p>
 <p>If the image is a floating dtype, then converts and scales data from [0, 1]
 to [0, 255] as needed. Otherwise, returns image as is.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any dtype.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tensor of the same shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">image</span></code> but with dtype tf.uint8.
 If the image was not a floating dtype, then it will not be changed.</p>
 <p>If the input was float with range [0, 1], it will be scaled to [0, 255].</p>
@@ -509,13 +538,13 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.ensure_min_image_rank">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_min_image_rank</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L11-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_min_image_rank" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_min_image_rank</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L11-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_min_image_rank" title="Permalink to this definition">#</a></dt>
 <dd><p>Expand the image to a minimum rank of 3 by adding single dimensions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any rank and dtype.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The expanded image to a minimum rank of 3.</p>
 <p>If the input was rank-2, it is assumed be of shape (height, width), so a
 singleton channels axis is appended to produce a tensor of shape
@@ -529,14 +558,14 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.ensure_rgb">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_rgb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L99-L114"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_rgb" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">ensure_rgb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L99-L114"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.ensure_rgb" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert image to RGB if in grayscale format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Tensor of any dtype of shape (height, width, channels). Channels are
 expected to be 1 or 3.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A grayscale image of shape (height, width, 1) of the same dtype as the input.</p>
 </dd>
 </dl>
@@ -545,17 +574,17 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.scale_image_range">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_image_range</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L130-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_image_range" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_image_range</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L130-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_image_range" title="Permalink to this definition">#</a></dt>
 <dd><p>Scale the range of image values.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – Tensor of any shape of dtype tf.float32 with values in the range [0, 1].</p></li>
 <li><p><strong>min_val</strong> – The minimum number that values will be scaled to.</p></li>
 <li><p><strong>max_val</strong> – The maximum number that values will be scaled to.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The scaled image of the same shape and dtype tf.float32. Values in the input
 that were 0 will now be scaled to <code class="xref py py-obj docutils literal notranslate"><span class="pre">min_val</span></code>, and values that were 1.0 will be
 scaled to <code class="xref py py-obj docutils literal notranslate"><span class="pre">max_val</span></code>.</p>
@@ -565,16 +594,16 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.scale_to_imagenet_caffe_mode">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_caffe_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L180-L223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_caffe_mode" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_caffe_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L180-L223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_caffe_mode" title="Permalink to this definition">#</a></dt>
 <dd><p>Scale images according to the “caffe” preprocessing mode.</p>
 <p>This applies the preprocessing operations implemented in <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.applications</span></code> for
 models pretrained on ImageNet.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Any image tensor of rank &gt;= 2. If rank &gt;=3, the last axis is assumed to
 be of size 3 corresponding to RGB-ordered channels.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The preprocessed image of dtype tf.float32 and shape (…, height, width, 3)
 with BGR channel ordering.</p>
 <p>Values will be in the approximate range of [-127.5, 127.5].</p>
@@ -603,15 +632,15 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.scale_to_imagenet_tf_mode">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_tf_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L146-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_tf_mode" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_tf_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L146-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_tf_mode" title="Permalink to this definition">#</a></dt>
 <dd><p>Scale images according to the “tf” preprocessing mode.</p>
 <p>This applies the preprocessing operations implemented in <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.applications</span></code> for
 models pretrained on ImageNet.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Any image tensor of rank &gt;= 2.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The preprocessed image of dtype tf.float32 and shape (…, height, width, 3)
 with RGB channel ordering.</p>
 <p>Values will be in the range [-1.0, 1.0].</p>
@@ -638,16 +667,16 @@ <h1>sleap.nn.data.normalization</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.normalization.scale_to_imagenet_torch_mode">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_torch_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/normalization.py#L226-L265"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_torch_mode" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.normalization.</span></span><span class="sig-name descname"><span class="pre">scale_to_imagenet_torch_mode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/normalization.py#L226-L265"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.normalization.scale_to_imagenet_torch_mode" title="Permalink to this definition">#</a></dt>
 <dd><p>Scale images according to the “torch” preprocessing mode.</p>
 <p>This applies the preprocessing operations implemented in <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.applications</span></code> for
 models pretrained on ImageNet.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>image</strong> – Any image tensor of rank &gt;= 2. If rank &gt;=3, the last axis is assumed to
 be of size 3 corresponding to RGB-ordered channels.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The preprocessed image of dtype tf.float32 and shape (…, height, width, 3)
 with RGB channel ordering.</p>
 <p>Values will be in the approximate range of [-0.5, 0.5].</p>
@@ -698,6 +727,41 @@ <h1>sleap.nn.data.normalization</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer"><code class="docutils literal notranslate"><span class="pre">Normalizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.image_key"><code class="docutils literal notranslate"><span class="pre">Normalizer.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_float"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_rgb</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">Normalizer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.imagenet_mode"><code class="docutils literal notranslate"><span class="pre">Normalizer.imagenet_mode</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.from_config"><code class="docutils literal notranslate"><span class="pre">Normalizer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.input_keys"><code class="docutils literal notranslate"><span class="pre">Normalizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.output_keys"><code class="docutils literal notranslate"><span class="pre">Normalizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.Normalizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Normalizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.convert_rgb_to_bgr"><code class="docutils literal notranslate"><span class="pre">convert_rgb_to_bgr()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_float"><code class="docutils literal notranslate"><span class="pre">ensure_float()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">ensure_grayscale()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_int"><code class="docutils literal notranslate"><span class="pre">ensure_int()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_min_image_rank"><code class="docutils literal notranslate"><span class="pre">ensure_min_image_rank()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.ensure_rgb"><code class="docutils literal notranslate"><span class="pre">ensure_rgb()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_image_range"><code class="docutils literal notranslate"><span class="pre">scale_image_range()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_caffe_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_caffe_mode()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_tf_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_tf_mode()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.normalization.scale_to_imagenet_torch_mode"><code class="docutils literal notranslate"><span class="pre">scale_to_imagenet_torch_mode()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -716,7 +780,7 @@ <h1>sleap.nn.data.normalization</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.offset_regression.html b/api/sleap.nn.data.offset_regression.html
index 973d72d1f..04ff138b5 100644
--- a/api/sleap.nn.data.offset_regression.html
+++ b/api/sleap.nn.data.offset_regression.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.offset_regression &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.offset_regression &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.offset_regression';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.offset_regression.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,15 @@ <h1>sleap.nn.data.offset_regression</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.offset_regression.make_offsets"><code class="docutils literal notranslate"><span class="pre">make_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.offset_regression.mask_offsets"><code class="docutils literal notranslate"><span class="pre">mask_offsets()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +327,14 @@ <h1>sleap.nn.data.offset_regression</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.offset_regression">
-<span id="sleap-nn-data-offset-regression"></span><h1>sleap.nn.data.offset_regression<a class="headerlink" href="#module-sleap.nn.data.offset_regression" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-offset-regression"></span><h1>sleap.nn.data.offset_regression<a class="headerlink" href="#module-sleap.nn.data.offset_regression" title="Permalink to this heading">#</a></h1>
 <p>Utilities for creating offset regression maps.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.offset_regression.make_offsets">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.offset_regression.</span></span><span class="sig-name descname"><span class="pre">make_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/offset_regression.py#L6-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.offset_regression.make_offsets" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.offset_regression.</span></span><span class="sig-name descname"><span class="pre">make_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/offset_regression.py#L6-L56"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.offset_regression.make_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>Make point offset maps on a grid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points</strong> – Point locations as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_points,</span> <span class="pre">2)</span></code> and dtype
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> where each row specifies the x- and y-coordinates of the map
@@ -341,7 +350,7 @@ <h1>sleap.nn.data.offset_regression</h1>
 grids so that the offsets point to the smaller grid coordinates.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The offset maps as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_points,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. The last axis
 corresponds to the x- and y-offsets at each grid point for each input point.</p>
@@ -355,11 +364,11 @@ <h1>sleap.nn.data.offset_regression</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.offset_regression.mask_offsets">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.offset_regression.</span></span><span class="sig-name descname"><span class="pre">mask_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/offset_regression.py#L59-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.offset_regression.mask_offsets" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.offset_regression.</span></span><span class="sig-name descname"><span class="pre">mask_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/offset_regression.py#L59-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.offset_regression.mask_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>Mask offset maps using a confidence map threshold.</p>
 <p>This is useful for restricting offset maps to local neighborhoods around the peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>offsets</strong> – A set of offset maps as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(grid_height,</span> <span class="pre">grid_width,</span> <span class="pre">n_points,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. This can be
@@ -372,7 +381,7 @@ <h1>sleap.nn.data.offset_regression</h1>
 with zeros.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The offset maps with the same shape as the inputs but with zeros where the
 confidence maps are below the specified threshold.</p>
 </dd>
@@ -404,6 +413,21 @@ <h1>sleap.nn.data.offset_regression</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.offset_regression.make_offsets"><code class="docutils literal notranslate"><span class="pre">make_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.offset_regression.mask_offsets"><code class="docutils literal notranslate"><span class="pre">mask_offsets()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -422,7 +446,7 @@ <h1>sleap.nn.data.offset_regression</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.pipelines.html b/api/sleap.nn.data.pipelines.html
index f7767deb4..86777444e 100644
--- a/api/sleap.nn.data.pipelines.html
+++ b/api/sleap.nn.data.pipelines.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.pipelines &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.pipelines &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.pipelines';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.pipelines.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,90 @@ <h1>sleap.nn.data.pipelines</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.class_maps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.class_maps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.pafs_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.pafs_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.centroid_confmap_head"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.centroid_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline"><code class="docutils literal notranslate"><span class="pre">Pipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.providers"><code class="docutils literal notranslate"><span class="pre">Pipeline.providers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.transformers"><code class="docutils literal notranslate"><span class="pre">Pipeline.transformers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.append"><code class="docutils literal notranslate"><span class="pre">Pipeline.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.describe"><code class="docutils literal notranslate"><span class="pre">Pipeline.describe()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.from_blocks"><code class="docutils literal notranslate"><span class="pre">Pipeline.from_blocks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.from_pipelines"><code class="docutils literal notranslate"><span class="pre">Pipeline.from_pipelines()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.make_dataset"><code class="docutils literal notranslate"><span class="pre">Pipeline.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.output_keys"><code class="docutils literal notranslate"><span class="pre">Pipeline.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.peek"><code class="docutils literal notranslate"><span class="pre">Pipeline.peek()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.run"><code class="docutils literal notranslate"><span class="pre">Pipeline.run()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.validate_pipeline"><code class="docutils literal notranslate"><span class="pre">Pipeline.validate_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.single_instance_confmap_head"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.single_instance_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.class_vectors_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.class_vectors_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.instance_confmap_head"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.instance_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +402,7 @@ <h1>sleap.nn.data.pipelines</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.pipelines">
-<span id="sleap-nn-data-pipelines"></span><h1>sleap.nn.data.pipelines<a class="headerlink" href="#module-sleap.nn.data.pipelines" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-pipelines"></span><h1>sleap.nn.data.pipelines<a class="headerlink" href="#module-sleap.nn.data.pipelines" title="Permalink to this heading">#</a></h1>
 <p>This module defines high level pipeline configurations from providers/transformers.</p>
 <p>The <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> class has the capability to create sequences of data I/O and processing
 operations wrapped in a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data</span></code>-based pipeline.</p>
@@ -326,14 +410,14 @@ <h1>sleap.nn.data.pipelines</h1>
 well as to define training vs inference versions based on the same configurations.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpMultiClassPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">sleap.nn.heads.MultiInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_maps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassMapsHead" title="sleap.nn.heads.ClassMapsHead"><span class="pre">sleap.nn.heads.ClassMapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L953-L1113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">MultiInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_maps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassMapsHead" title="sleap.nn.heads.ClassMapsHead"><span class="pre">ClassMapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L953-L1113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for confidence maps and class maps models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpMultiClassPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -344,7 +428,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -355,7 +439,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">confmaps_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.confmaps_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiated head describing the output confidence maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead">sleap.nn.heads.MultiInstanceConfmapsHead</a></p>
 </dd>
 </dl>
@@ -366,7 +450,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">class_maps_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.class_maps_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiated head describing the output class maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassMapsHead" title="sleap.nn.heads.ClassMapsHead">sleap.nn.heads.ClassMapsHead</a></p>
 </dd>
 </dl>
@@ -377,7 +461,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -385,14 +469,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L970-L993"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L970-L993"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -400,14 +484,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L995-L1077"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L995-L1077"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -421,17 +505,17 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L1079-L1113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L1079-L1113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p></li>
 <li><p><strong>keras_model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> that can be used for inference.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and run inference to generate
 predictions useful for visualization during training.</p>
 </dd>
@@ -442,14 +526,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">BottomUpPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">sleap.nn.heads.MultiInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.PartAffinityFieldsHead" title="sleap.nn.heads.PartAffinityFieldsHead"><span class="pre">sleap.nn.heads.PartAffinityFieldsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L789-L949"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">BottomUpPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmaps_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">MultiInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.PartAffinityFieldsHead" title="sleap.nn.heads.PartAffinityFieldsHead"><span class="pre">PartAffinityFieldsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L789-L949"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for confidence maps + part affinity fields models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -460,7 +544,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -471,7 +555,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">confmaps_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.confmaps_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiated head describing the output confidence maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead">sleap.nn.heads.MultiInstanceConfmapsHead</a></p>
 </dd>
 </dl>
@@ -482,7 +566,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">pafs_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.pafs_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiated head describing the output PAFs tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.PartAffinityFieldsHead" title="sleap.nn.heads.PartAffinityFieldsHead">sleap.nn.heads.PartAffinityFieldsHead</a></p>
 </dd>
 </dl>
@@ -493,7 +577,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -501,14 +585,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L806-L829"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L806-L829"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -516,14 +600,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L831-L912"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L831-L912"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -537,17 +621,17 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L914-L949"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L914-L949"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p></li>
 <li><p><strong>keras_model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> that can be used for inference.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and run inference to generate
 predictions useful for visualization during training.</p>
 </dd>
@@ -558,14 +642,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.CentroidConfmapsPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CentroidConfmapsHead" title="sleap.nn.heads.CentroidConfmapsHead"><span class="pre">sleap.nn.heads.CentroidConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L474-L637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CentroidConfmapsHead" title="sleap.nn.heads.CentroidConfmapsHead"><span class="pre">CentroidConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L474-L637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for centroid confidence map models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.CentroidConfmapsPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -576,7 +660,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -588,7 +672,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <dd><p>Instantiated head describing the output centroid
 confidence maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CentroidConfmapsHead" title="sleap.nn.heads.CentroidConfmapsHead">sleap.nn.heads.CentroidConfmapsHead</a></p>
 </dd>
 </dl>
@@ -599,7 +683,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -607,14 +691,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L490-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L490-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -622,14 +706,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L519-L599"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L519-L599"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -643,17 +727,17 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L601-L637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L601-L637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p></li>
 <li><p><strong>keras_model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> that can be used for inference.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and run inference to generate
 predictions useful for visualization during training.</p>
 </dd>
@@ -664,14 +748,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">Pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">providers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transformers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L113-L330"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">Pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">providers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">transformers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L113-L330"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline composed of providers and transformers.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.providers">
 <span class="sig-name descname"><span class="pre">providers</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.providers" title="Permalink to this definition">#</a></dt>
 <dd><p>A single or a list of data providers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[sleap.nn.data.pipelines.Provider]</p>
 </dd>
 </dl>
@@ -682,7 +766,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">transformers</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.transformers" title="Permalink to this definition">#</a></dt>
 <dd><p>A single or a list of transformers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[sleap.nn.data.pipelines.Transformer]</p>
 </dd>
 </dl>
@@ -690,14 +774,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.append">
-<span class="sig-name descname"><span class="pre">append</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">sleap.nn.data.pipelines.Transformer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Transformer</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L180-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.append" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">append</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Transformer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Transformer</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L180-L206"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.append" title="Permalink to this definition">#</a></dt>
 <dd><p>Append one or more blocks to this pipeline instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – A single <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">Transformer</span></code> or list of <a href="#id1"><span class="problematic" id="id2">`</span></a>Transformer`s to
 append to the end of this pipeline.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If blocks provided are not a <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">Transformer</span></code> or list
     of <a href="#id3"><span class="problematic" id="id4">`</span></a>Transformer`s.</p>
 </dd>
@@ -706,14 +790,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.describe">
-<span class="sig-name descname"><span class="pre">describe</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">return_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L319-L330"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.describe" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">describe</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">return_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L319-L330"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.describe" title="Permalink to this definition">#</a></dt>
 <dd><p>Prints the keys in the examples generated by the pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>return_description</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, returns the string description instead of
 printing it.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>String description if <code class="xref py py-obj docutils literal notranslate"><span class="pre">return_description</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, otherwise <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 </dd>
 </dl>
@@ -721,13 +805,13 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.from_blocks">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_blocks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">sleap.nn.data.pipelines.Transformer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">sleap.nn.data.pipelines.Transformer</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L124-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.from_blocks" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_blocks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">blocks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Transformer</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Transformer</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L124-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.from_blocks" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a pipeline from a sequence of providers and transformers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>blocks</strong> – List or tuple of providers and transformer instances.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instantiated pipeline with all blocks chained.</p>
 </dd>
 </dl>
@@ -735,13 +819,13 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.from_pipelines">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pipelines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipelines</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L154-L170"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.from_pipelines" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_pipelines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pipelines</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L154-L170"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.from_pipelines" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a new pipeline instance by chaining together multiple pipelines.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>pipelines</strong> – A sequence of <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instances.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance formed by concatenating the individual pipelines.</p>
 </dd>
 </dl>
@@ -749,10 +833,10 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.make_dataset">
-<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L268-L290"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.make_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L268-L290"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.make_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset instance that generates examples from the pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> pipeline that generates examples with the
 keys in the <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline.output_keys" title="sleap.nn.data.pipelines.Pipeline.output_keys"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_keys</span></code></a> attribute.</p>
 </dd>
@@ -767,14 +851,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.peek">
-<span class="sig-name descname"><span class="pre">peek</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L300-L317"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.peek" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">peek</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L300-L317"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.peek" title="Permalink to this definition">#</a></dt>
 <dd><p>Build and return the first <code class="xref py py-obj docutils literal notranslate"><span class="pre">n</span></code> examples from the pipeline.</p>
 <p>This function is useful for quickly inspecting the output of a pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>n</strong> – Number of examples to get from the pipeline.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary with tensors if <code class="xref py py-obj docutils literal notranslate"><span class="pre">n</span> <span class="pre">=</span> <span class="pre">1</span></code>, or a list of dictionaries if <code class="xref py py-obj docutils literal notranslate"><span class="pre">n</span> <span class="pre">&gt;</span> <span class="pre">1</span></code>.</p>
 </dd>
 </dl>
@@ -782,10 +866,10 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.run">
-<span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L292-L298"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.run" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">run</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L292-L298"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.run" title="Permalink to this definition">#</a></dt>
 <dd><p>Build and evaluate the pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List of example dictionaries after processing the pipeline.</p>
 </dd>
 </dl>
@@ -793,13 +877,13 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.Pipeline.validate_pipeline">
-<span class="sig-name descname"><span class="pre">validate_pipeline</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L218-L261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.validate_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">validate_pipeline</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L218-L261"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.Pipeline.validate_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Check that all pipeline blocks meet the data requirements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The final keys that will be present in each example.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If keys required for a block are dropped at some point in the
     pipeline.</p>
 </dd>
@@ -810,14 +894,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">single_instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.SingleInstanceConfmapsHead" title="sleap.nn.heads.SingleInstanceConfmapsHead"><span class="pre">sleap.nn.heads.SingleInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L334-L470"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">single_instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.SingleInstanceConfmapsHead" title="sleap.nn.heads.SingleInstanceConfmapsHead"><span class="pre">SingleInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L334-L470"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for single-instance confidence map models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -828,7 +912,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -840,7 +924,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <dd><p>Instantiated head describing the output confidence
 maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.SingleInstanceConfmapsHead" title="sleap.nn.heads.SingleInstanceConfmapsHead">sleap.nn.heads.SingleInstanceConfmapsHead</a></p>
 </dd>
 </dl>
@@ -851,7 +935,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -859,14 +943,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L350-L373"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L350-L373"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -874,14 +958,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L375-L446"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L375-L446"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -895,17 +979,17 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L448-L470"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L448-L470"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p></li>
 <li><p><strong>keras_model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> that can be used for inference.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and run inference to generate
 predictions useful for visualization during training.</p>
 </dd>
@@ -916,14 +1000,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopDownMultiClassPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">sleap.nn.heads.CenteredInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_vectors_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassVectorsHead" title="sleap.nn.heads.ClassVectorsHead"><span class="pre">sleap.nn.heads.ClassVectorsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L1117-L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">CenteredInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">class_vectors_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassVectorsHead" title="sleap.nn.heads.ClassVectorsHead"><span class="pre">ClassVectorsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L1117-L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for confidence maps and class maps models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopDownMultiClassPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -934,7 +1018,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -952,7 +1036,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <dd><p>Instantiated head describing the output class vectors
 tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.ClassVectorsHead" title="sleap.nn.heads.ClassVectorsHead">sleap.nn.heads.ClassVectorsHead</a></p>
 </dd>
 </dl>
@@ -963,7 +1047,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -971,14 +1055,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L1135-L1158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L1135-L1158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -986,14 +1070,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L1160-L1233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L1160-L1233"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -1007,14 +1091,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L1235-L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L1235-L1260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and run inference to generate
 predictions useful for visualization during training.</p>
 </dd>
@@ -1025,14 +1109,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopdownConfmapsPipeline">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">TopdownConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">sleap.nn.config.data.DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">sleap.nn.heads.CenteredInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L641-L785"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.pipelines.</span></span><span class="sig-name descname"><span class="pre">TopdownConfmapsPipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig"><span class="pre">DataConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">optimization_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_confmap_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">CenteredInstanceConfmapsHead</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets_head</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L641-L785"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Pipeline builder for instance-centered confidence map models.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopdownConfmapsPipeline.data_config">
 <span class="sig-name descname"><span class="pre">data_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.data_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Data-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.DataConfig" title="sleap.nn.config.data.DataConfig">sleap.nn.config.data.DataConfig</a></p>
 </dd>
 </dl>
@@ -1043,7 +1127,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">optimization_config</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.optimization_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Optimization-related configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig">sleap.nn.config.optimization.OptimizationConfig</a></p>
 </dd>
 </dl>
@@ -1055,7 +1139,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <dd><p>Instantiated head describing the output centered
 confidence maps tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead">sleap.nn.heads.CenteredInstanceConfmapsHead</a></p>
 </dd>
 </dl>
@@ -1066,7 +1150,7 @@ <h1>sleap.nn.data.pipelines</h1>
 <span class="sig-name descname"><span class="pre">offsets_head</span></span><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.offsets_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Optional head describing the offset refinement maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.heads.html#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead">sleap.nn.heads.OffsetRefinementHead</a>]</p>
 </dd>
 </dl>
@@ -1074,14 +1158,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline">
-<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L657-L680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_base_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L657-L680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create base pipeline with input data only.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce input examples.</p>
 </dd>
 </dl>
@@ -1089,14 +1173,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline">
-<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L682-L758"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_training_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L682-L758"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create full training pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to produce all data keys required for
 training.</p>
 </dd>
@@ -1110,14 +1194,14 @@ <h1>sleap.nn.data.pipelines</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline">
-<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/pipelines.py#L760-L785"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_viz_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/pipelines.py#L760-L785"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Create visualization pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Provider</span></code> that generates data examples, typically a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code></a> instance configured to fetch data and for running inference to
 generate predictions useful for visualization during training.</p>
 </dd>
@@ -1150,6 +1234,96 @@ <h1>sleap.nn.data.pipelines</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.class_maps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.class_maps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.pafs_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.pafs_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.centroid_confmap_head"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.centroid_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline"><code class="docutils literal notranslate"><span class="pre">Pipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.providers"><code class="docutils literal notranslate"><span class="pre">Pipeline.providers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.transformers"><code class="docutils literal notranslate"><span class="pre">Pipeline.transformers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.append"><code class="docutils literal notranslate"><span class="pre">Pipeline.append()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.describe"><code class="docutils literal notranslate"><span class="pre">Pipeline.describe()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.from_blocks"><code class="docutils literal notranslate"><span class="pre">Pipeline.from_blocks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.from_pipelines"><code class="docutils literal notranslate"><span class="pre">Pipeline.from_pipelines()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.make_dataset"><code class="docutils literal notranslate"><span class="pre">Pipeline.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.output_keys"><code class="docutils literal notranslate"><span class="pre">Pipeline.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.peek"><code class="docutils literal notranslate"><span class="pre">Pipeline.peek()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.run"><code class="docutils literal notranslate"><span class="pre">Pipeline.run()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.Pipeline.validate_pipeline"><code class="docutils literal notranslate"><span class="pre">Pipeline.validate_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.single_instance_confmap_head"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.single_instance_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.confmaps_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.confmaps_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.class_vectors_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.class_vectors_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.data_config"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.data_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.optimization_config"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.optimization_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.instance_confmap_head"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.instance_confmap_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.offsets_head"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.offsets_head</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_base_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_training_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsPipeline.make_viz_pipeline()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1168,7 +1342,7 @@ <h1>sleap.nn.data.pipelines</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.providers.html b/api/sleap.nn.data.providers.html
index 0cfb51abc..e24dc7017 100644
--- a/api/sleap.nn.data.providers.html
+++ b/api/sleap.nn.data.providers.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.providers &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.providers &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.providers';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.providers.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,41 @@ <h1>sleap.nn.data.providers</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader"><code class="docutils literal notranslate"><span class="pre">LabelsReader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.labels"><code class="docutils literal notranslate"><span class="pre">LabelsReader.labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.example_indices"><code class="docutils literal notranslate"><span class="pre">LabelsReader.example_indices</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.user_instances_only"><code class="docutils literal notranslate"><span class="pre">LabelsReader.user_instances_only</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.with_track_only"><code class="docutils literal notranslate"><span class="pre">LabelsReader.with_track_only</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_filename"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_unlabeled_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_user_instances"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_user_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_user_labeled_frames"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_user_labeled_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.is_from_multi_size_videos"><code class="docutils literal notranslate"><span class="pre">LabelsReader.is_from_multi_size_videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.make_dataset"><code class="docutils literal notranslate"><span class="pre">LabelsReader.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.max_height_and_width"><code class="docutils literal notranslate"><span class="pre">LabelsReader.max_height_and_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.output_keys"><code class="docutils literal notranslate"><span class="pre">LabelsReader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.tracks"><code class="docutils literal notranslate"><span class="pre">LabelsReader.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.videos"><code class="docutils literal notranslate"><span class="pre">LabelsReader.videos</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader"><code class="docutils literal notranslate"><span class="pre">VideoReader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.video"><code class="docutils literal notranslate"><span class="pre">VideoReader.video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.example_indices"><code class="docutils literal notranslate"><span class="pre">VideoReader.example_indices</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.video_ind"><code class="docutils literal notranslate"><span class="pre">VideoReader.video_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.from_filepath"><code class="docutils literal notranslate"><span class="pre">VideoReader.from_filepath()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.make_dataset"><code class="docutils literal notranslate"><span class="pre">VideoReader.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.max_height_and_width"><code class="docutils literal notranslate"><span class="pre">VideoReader.max_height_and_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.output_keys"><code class="docutils literal notranslate"><span class="pre">VideoReader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.videos"><code class="docutils literal notranslate"><span class="pre">VideoReader.videos</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +353,11 @@ <h1>sleap.nn.data.providers</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.providers">
-<span id="sleap-nn-data-providers"></span><h1>sleap.nn.data.providers<a class="headerlink" href="#module-sleap.nn.data.providers" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-providers"></span><h1>sleap.nn.data.providers<a class="headerlink" href="#module-sleap.nn.data.providers" title="Permalink to this heading">#</a></h1>
 <p>Data providers for pipeline I/O.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.providers.</span></span><span class="sig-name descname"><span class="pre">LabelsReader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_instances_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L11-L303"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.providers.</span></span><span class="sig-name descname"><span class="pre">LabelsReader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_instances_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L11-L303"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader" title="Permalink to this definition">#</a></dt>
 <dd><p>Data provider from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance.</p>
 <p>This class can generate <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset`s</span> <span class="pre">from</span> <span class="pre">a</span> <span class="pre">set</span> <span class="pre">of</span> <span class="pre">labels</span> <span class="pre">for</span> <span class="pre">use</span> <span class="pre">in</span> <span class="pre">data</span>
 <span class="pre">pipelines.</span> <span class="pre">Each</span> <span class="pre">element</span> <span class="pre">in</span> <span class="pre">the</span> <span class="pre">dataset</span> <span class="pre">will</span> <span class="pre">contain</span> <span class="pre">the</span> <span class="pre">data</span> <span class="pre">contained</span> <span class="pre">in</span> <span class="pre">a</span> <span class="pre">single</span>
@@ -332,7 +367,7 @@ <h1>sleap.nn.data.providers</h1>
 <span class="sig-name descname"><span class="pre">labels</span></span><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.labels" title="Permalink to this definition">#</a></dt>
 <dd><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance to generate data from.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels">sleap.io.dataset.Labels</a></p>
 </dd>
 </dl>
@@ -348,7 +383,7 @@ <h1>sleap.nn.data.providers</h1>
 the labeled frames in <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader.labels" title="sleap.nn.data.providers.LabelsReader.labels"><code class="xref py py-obj docutils literal notranslate"><span class="pre">labels</span></code></a> attribute, which may have changed in ordering
 or filtered.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[Union[Sequence[int], numpy.ndarray]]</p>
 </dd>
 </dl>
@@ -360,7 +395,7 @@ <h1>sleap.nn.data.providers</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, load only user labeled instances. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>,
 all instances will be loaded.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -372,7 +407,7 @@ <h1>sleap.nn.data.providers</h1>
 <dd><p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, load only instances that have a track assigned.
 Useful when training supervised ID models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -380,16 +415,16 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.from_filename">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L85-L102"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_filename" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filename</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L85-L102"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_filename" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> from a saved labels file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a saved labels file.</p></li>
 <li><p><strong>user_instances</strong> – If True, will use only labeled frames with user instances.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> instance that can create a dataset for pipelining.</p>
 </dd>
 </dl>
@@ -397,11 +432,11 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_unlabeled_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L74-L83"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_unlabeled_suggestions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L74-L83"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> using the unlabeled suggestions in a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> set.
 :param labels: A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance containing unlabeled suggestions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> instance that can create a dataset for pipelining.</p>
 </dd>
 </dl>
@@ -409,7 +444,7 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.from_user_instances">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_user_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L37-L59"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_user_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_user_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L37-L59"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_user_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> using the user instances in a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> set.
 :param labels: A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance containing user instances.
 :param with_track_only: If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, load only instances that have a track assigned.</p>
@@ -417,7 +452,7 @@ <h1>sleap.nn.data.providers</h1>
 <div><p>Useful when training supervised ID models.</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> instance that can create a dataset for pipelining.</p>
 </dd>
 </dl>
@@ -431,11 +466,11 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.from_user_labeled_frames">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_user_labeled_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L61-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_user_labeled_frames" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_user_labeled_frames</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L61-L72"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.from_user_labeled_frames" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> using the user labeled frames in a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> set.
 :param labels: A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance containing user labeled frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> instance that can create a dataset for pipelining.
 Note that this constructor will load ALL instances in frames that have user
 instances. To load only user labeled instances, use
@@ -452,7 +487,7 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.make_dataset">
-<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L152-L303"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.make_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">DatasetV2</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L152-L303"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.make_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> whose elements are data from labeled frames.
 :returns: A dataset whose elements are dictionaries with the loaded data associated</p>
 <blockquote>
@@ -504,13 +539,13 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.tracks">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tracks</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.tracks" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">tracks</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the list of tracks that <code class="xref py py-obj docutils literal notranslate"><span class="pre">track_inds</span></code> in examples match up with.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.LabelsReader.videos">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">videos</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.videos" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">videos</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.LabelsReader.videos" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the list of videos that <code class="xref py py-obj docutils literal notranslate"><span class="pre">video_ind</span></code> in examples match up with.</p>
 </dd></dl>
 
@@ -518,7 +553,7 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.VideoReader">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.providers.</span></span><span class="sig-name descname"><span class="pre">VideoReader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L307-L441"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.providers.</span></span><span class="sig-name descname"><span class="pre">VideoReader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L307-L439"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader" title="Permalink to this definition">#</a></dt>
 <dd><p>Data provider from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> instance.</p>
 <p>This class can generate <a href="#id1"><span class="problematic" id="id2">`</span></a>tf.data.Dataset`s from a video for use in data pipelines.
 Each element in the dataset will contain the image data from a single frame.</p>
@@ -527,7 +562,7 @@ <h1>sleap.nn.data.providers</h1>
 <span class="sig-name descname"><span class="pre">video</span></span><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.video" title="Permalink to this definition">#</a></dt>
 <dd><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> instance to generate data from.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video">sleap.io.video.Video</a></p>
 </dd>
 </dl>
@@ -540,7 +575,7 @@ <h1>sleap.nn.data.providers</h1>
 iterating over the video. Use this to specify subsets of the video to read.
 If not provided, the entire video will be read.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[Union[Sequence[int], numpy.ndarray]]</p>
 </dd>
 </dl>
@@ -555,10 +590,10 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.VideoReader.from_filepath">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filepath</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.VideoReader" title="sleap.nn.data.providers.VideoReader"><span class="pre">sleap.nn.data.providers.VideoReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L325-L345"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.from_filepath" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_filepath</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">example_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.providers.VideoReader" title="sleap.nn.data.providers.VideoReader"><span class="pre">VideoReader</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L325-L345"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.from_filepath" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a <a class="reference internal" href="#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code></a> from a saved labels file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – Path to a video file.</p></li>
 <li><p><strong>example_indices</strong> – List or numpy array of ints with the frame indices to use
@@ -567,7 +602,7 @@ <h1>sleap.nn.data.providers</h1>
 <li><p><strong>**kwargs</strong> – Any other video keyword argument (e.g., grayscale, dataset).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <a class="reference internal" href="#sleap.nn.data.providers.VideoReader" title="sleap.nn.data.providers.VideoReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">VideoReader</span></code></a> instance that can create a dataset for pipelining.</p>
 </dd>
 </dl>
@@ -575,10 +610,10 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.VideoReader.make_dataset">
-<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/providers.py#L371-L441"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.make_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_dataset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/providers.py#L371-L439"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.make_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> whose elements are data from video frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p><p>A dataset whose elements are dictionaries with the loaded data associated
 with a single video frame. Items will be converted to tensors. These are:</p>
 <blockquote>
@@ -622,7 +657,7 @@ <h1>sleap.nn.data.providers</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.providers.VideoReader.videos">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">videos</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.videos" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">videos</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.data.providers.VideoReader.videos" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the list of videos that <a class="reference internal" href="#sleap.nn.data.providers.VideoReader.video_ind" title="sleap.nn.data.providers.VideoReader.video_ind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">video_ind</span></code></a> in examples match up with.</p>
 </dd></dl>
 
@@ -652,6 +687,47 @@ <h1>sleap.nn.data.providers</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader"><code class="docutils literal notranslate"><span class="pre">LabelsReader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.labels"><code class="docutils literal notranslate"><span class="pre">LabelsReader.labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.example_indices"><code class="docutils literal notranslate"><span class="pre">LabelsReader.example_indices</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.user_instances_only"><code class="docutils literal notranslate"><span class="pre">LabelsReader.user_instances_only</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.with_track_only"><code class="docutils literal notranslate"><span class="pre">LabelsReader.with_track_only</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_filename"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_filename()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_unlabeled_suggestions()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_user_instances"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_user_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.from_user_labeled_frames"><code class="docutils literal notranslate"><span class="pre">LabelsReader.from_user_labeled_frames()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.is_from_multi_size_videos"><code class="docutils literal notranslate"><span class="pre">LabelsReader.is_from_multi_size_videos</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.make_dataset"><code class="docutils literal notranslate"><span class="pre">LabelsReader.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.max_height_and_width"><code class="docutils literal notranslate"><span class="pre">LabelsReader.max_height_and_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.output_keys"><code class="docutils literal notranslate"><span class="pre">LabelsReader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.tracks"><code class="docutils literal notranslate"><span class="pre">LabelsReader.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.LabelsReader.videos"><code class="docutils literal notranslate"><span class="pre">LabelsReader.videos</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader"><code class="docutils literal notranslate"><span class="pre">VideoReader</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.video"><code class="docutils literal notranslate"><span class="pre">VideoReader.video</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.example_indices"><code class="docutils literal notranslate"><span class="pre">VideoReader.example_indices</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.video_ind"><code class="docutils literal notranslate"><span class="pre">VideoReader.video_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.from_filepath"><code class="docutils literal notranslate"><span class="pre">VideoReader.from_filepath()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.make_dataset"><code class="docutils literal notranslate"><span class="pre">VideoReader.make_dataset()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.max_height_and_width"><code class="docutils literal notranslate"><span class="pre">VideoReader.max_height_and_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.output_keys"><code class="docutils literal notranslate"><span class="pre">VideoReader.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.providers.VideoReader.videos"><code class="docutils literal notranslate"><span class="pre">VideoReader.videos</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -670,7 +746,7 @@ <h1>sleap.nn.data.providers</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.resizing.html b/api/sleap.nn.data.resizing.html
index a13f96268..bb47fba85 100644
--- a/api/sleap.nn.data.resizing.html
+++ b/api/sleap.nn.data.resizing.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.resizing &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.resizing &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.resizing';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.resizing.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,51 @@ <h1>sleap.nn.data.resizing</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler"><code class="docutils literal notranslate"><span class="pre">PointsRescaler</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.input_keys"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.output_keys"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer"><code class="docutils literal notranslate"><span class="pre">Resizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.image_key"><code class="docutils literal notranslate"><span class="pre">Resizer.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.scale_key"><code class="docutils literal notranslate"><span class="pre">Resizer.scale_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.points_key"><code class="docutils literal notranslate"><span class="pre">Resizer.points_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.scale"><code class="docutils literal notranslate"><span class="pre">Resizer.scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">Resizer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.keep_full_image"><code class="docutils literal notranslate"><span class="pre">Resizer.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.full_image_key"><code class="docutils literal notranslate"><span class="pre">Resizer.full_image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.from_config"><code class="docutils literal notranslate"><span class="pre">Resizer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.input_keys"><code class="docutils literal notranslate"><span class="pre">Resizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.output_keys"><code class="docutils literal notranslate"><span class="pre">Resizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Resizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher"><code class="docutils literal notranslate"><span class="pre">SizeMatcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.image_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.scale_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.scale_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.points_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.points_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.keep_full_image"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.full_image_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.full_image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.max_image_height"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.max_image_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.max_image_width"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.max_image_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.center_pad"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.center_pad</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.from_config"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.input_keys"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.output_keys"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.find_padding_for_stride"><code class="docutils literal notranslate"><span class="pre">find_padding_for_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">pad_to_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.resize_image"><code class="docutils literal notranslate"><span class="pre">resize_image()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +363,11 @@ <h1>sleap.nn.data.resizing</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.resizing">
-<span id="sleap-nn-data-resizing"></span><h1>sleap.nn.data.resizing<a class="headerlink" href="#module-sleap.nn.data.resizing" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-resizing"></span><h1>sleap.nn.data.resizing<a class="headerlink" href="#module-sleap.nn.data.resizing" title="Permalink to this heading">#</a></h1>
 <p>Transformers for image resizing and padding.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.PointsRescaler">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">PointsRescaler</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invert</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L473-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.PointsRescaler" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">PointsRescaler</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'predicted_instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invert</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L473-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.PointsRescaler" title="Permalink to this definition">#</a></dt>
 <dd><p>Transformer to apply or invert scaling operations on points.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.PointsRescaler.input_keys">
@@ -338,7 +383,7 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.PointsRescaler.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L490-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.PointsRescaler.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L490-L517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.PointsRescaler.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains instance cropped data.</p>
 </dd></dl>
 
@@ -346,7 +391,7 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.Resizer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">Resizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L109-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">Resizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L109-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer to resize or pad images.</p>
 <p>This is useful as a transformation to data streams that require resizing or padding
 in order to be downsampled or meet divisibility criteria.</p>
@@ -355,7 +400,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">image_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the images to resize.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -366,7 +411,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">scale_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.scale_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the scale of the images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -378,7 +423,7 @@ <h1>sleap.nn.data.resizing</h1>
 <dd><p>String name of the key containing points to adjust for the resizing
 operation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -389,7 +434,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">scale</span></span><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.scale" title="Permalink to this definition">#</a></dt>
 <dd><p>Scalar float specifying scaling factor to resize images by.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -403,7 +448,7 @@ <h1>sleap.nn.data.resizing</h1>
 this divisibility criteria. Padding is applied after the scaling specified
 in the <a class="reference internal" href="#sleap.nn.data.resizing.Resizer.scale" title="sleap.nn.data.resizing.Resizer.scale"><code class="xref py py-obj docutils literal notranslate"><span class="pre">scale</span></code></a> attribute.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -415,7 +460,7 @@ <h1>sleap.nn.data.resizing</h1>
 <dd><p>If True, keeps the (original size) full image in the examples.
 This is useful for multi-scale inference.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -426,7 +471,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">full_image_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.full_image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the full images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -434,10 +479,10 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.Resizer.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">sleap.nn.config.data.PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.resizing.Resizer" title="sleap.nn.data.resizing.Resizer"><span class="pre">sleap.nn.data.resizing.Resizer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L138-L188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pad_to_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.resizing.Resizer" title="sleap.nn.data.resizing.Resizer"><span class="pre">Resizer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L138-L188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Build an instance of this class from its configuration options.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">PreprocessingConfig</span></code> instance with the desired parameters. If
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">config.pad_to_stride</span></code> is not an explicit integer, the <a class="reference internal" href="#sleap.nn.data.resizing.pad_to_stride" title="sleap.nn.data.resizing.pad_to_stride"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pad_to_stride</span></code></a>
@@ -453,10 +498,10 @@ <h1>sleap.nn.data.resizing</h1>
 resizing operation.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class.</p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ValueError</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">config.pad_to_stride</span></code> is not set to an integer and the
     <a class="reference internal" href="#sleap.nn.data.resizing.pad_to_stride" title="sleap.nn.data.resizing.pad_to_stride"><code class="xref py py-obj docutils literal notranslate"><span class="pre">pad_to_stride</span></code></a> argument is not provided.</p>
 </dd>
@@ -477,15 +522,15 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.Resizer.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L206-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L206-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.Resizer.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset that contains centroids computed from the inputs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – A dataset with the image specified in the <a class="reference internal" href="#sleap.nn.data.resizing.Resizer.image_key" title="sleap.nn.data.resizing.Resizer.image_key"><code class="xref py py-obj docutils literal notranslate"><span class="pre">image_key</span></code></a> attribute,
 points specified in the <a class="reference internal" href="#sleap.nn.data.resizing.Resizer.points_key" title="sleap.nn.data.resizing.Resizer.points_key"><code class="xref py py-obj docutils literal notranslate"><span class="pre">points_key</span></code></a> attribute, and the “scale” key for
 tracking scaling transformations.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same images and points with
 resizing applied.</p>
 <p>The “scale” key of the example will be multipled by the <a class="reference internal" href="#sleap.nn.data.resizing.Resizer.scale" title="sleap.nn.data.resizing.Resizer.scale"><code class="xref py py-obj docutils literal notranslate"><span class="pre">scale</span></code></a> attribute of
@@ -502,14 +547,14 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.SizeMatcher">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">SizeMatcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">center_pad</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L257-L469"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">SizeMatcher</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">center_pad</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L257-L469"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher" title="Permalink to this definition">#</a></dt>
 <dd><p>Data transformer that ensures output images have uniform shape by resizing/padding smaller images.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.SizeMatcher.image_key">
 <span class="sig-name descname"><span class="pre">image_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the images to resize.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -520,7 +565,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">scale_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.scale_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the scale of the images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -532,7 +577,7 @@ <h1>sleap.nn.data.resizing</h1>
 <dd><p>String name of the key containing points to adjust for the resizing
 operation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -544,7 +589,7 @@ <h1>sleap.nn.data.resizing</h1>
 <dd><p>If True, keeps the (original size) full image in the examples.
 This is useful for multi-scale inference.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -555,7 +600,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">full_image_key</span></span><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.full_image_key" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the key containing the full images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -566,7 +611,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">max_image_height</span></span><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.max_image_height" title="Permalink to this definition">#</a></dt>
 <dd><p>int The target height to which all smaller images will be resized/padded to.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -577,7 +622,7 @@ <h1>sleap.nn.data.resizing</h1>
 <span class="sig-name descname"><span class="pre">max_image_width</span></span><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.max_image_width" title="Permalink to this definition">#</a></dt>
 <dd><p>int The target width to which all smaller images will be resized/padded to.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -589,7 +634,7 @@ <h1>sleap.nn.data.resizing</h1>
 <dd><p>If True, pad on the left/top to center the non-zero pixels rather than aligning
 them to the top-left. The offsets will be stored in the “offset_x” and “offset_y” keys.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -597,10 +642,10 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.SizeMatcher.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">sleap.nn.config.data.PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.resizing.SizeMatcher" title="sleap.nn.data.resizing.SizeMatcher"><span class="pre">SizeMatcher</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L283-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig" title="sleap.nn.config.data.PreprocessingConfig"><span class="pre">PreprocessingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'scale'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keep_full_image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">full_image_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'full_image'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_key</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'instances'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.data.resizing.SizeMatcher" title="sleap.nn.data.resizing.SizeMatcher"><span class="pre">SizeMatcher</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L283-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Build an instance of this class from configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – An <code class="xref py py-obj docutils literal notranslate"><span class="pre">PreprocessingConfig</span></code> instance with the desired parameters. If
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">config.resize_and_pad_to_target</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> and ‘target_height’ /
@@ -620,7 +665,7 @@ <h1>sleap.nn.data.resizing</h1>
 resizing operation.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of this class.</p>
 </dd>
 </dl>
@@ -640,15 +685,15 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.SizeMatcher.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L357-L469"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L357-L469"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.SizeMatcher.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Transform a dataset with variable size images into one with fixed sizes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – A dataset with the image specified in the <a class="reference internal" href="#sleap.nn.data.resizing.SizeMatcher.image_key" title="sleap.nn.data.resizing.SizeMatcher.image_key"><code class="xref py py-obj docutils literal notranslate"><span class="pre">image_key</span></code></a> attribute,
 points specified in the <a class="reference internal" href="#sleap.nn.data.resizing.SizeMatcher.points_key" title="sleap.nn.data.resizing.SizeMatcher.points_key"><code class="xref py py-obj docutils literal notranslate"><span class="pre">points_key</span></code></a> attribute, and the “scale” key for
 tracking scaling transformations.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> with elements containing the same images and points of
 equal size.</p>
 <p>If the <a class="reference internal" href="#sleap.nn.data.resizing.SizeMatcher.keep_full_image" title="sleap.nn.data.resizing.SizeMatcher.keep_full_image"><code class="xref py py-obj docutils literal notranslate"><span class="pre">keep_full_image</span></code></a> attribute is True, a key specified by
@@ -663,12 +708,12 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.find_padding_for_stride">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">find_padding_for_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L10-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.find_padding_for_stride" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">find_padding_for_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L10-L31"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.find_padding_for_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute padding required to ensure image is divisible by a stride.</p>
 <p>This function is useful for determining how to pad images such that they will not
 have issues with divisibility after repeated pooling steps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image_height</strong> – Scalar integer specifying the image height (rows).</p></li>
 <li><p><strong>image_width</strong> – Scalar integer specifying the image height (columns).</p></li>
@@ -676,7 +721,7 @@ <h1>sleap.nn.data.resizing</h1>
 divisible by.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tuple of (pad_bottom, pad_right), integers with the number of pixels that the
 image would need to be padded by to meet the divisibility requirement.</p>
 </dd>
@@ -685,12 +730,12 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.pad_to_stride">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">pad_to_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="headerlink" href="#sleap.nn.data.resizing.pad_to_stride" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">pad_to_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="headerlink" href="#sleap.nn.data.resizing.pad_to_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Pad an image to meet a max stride constraint.</p>
 <p>This is useful for ensuring there is no size mismatch between an image and the
 output tensors after multiple downsampling and upsampling steps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – Single image tensor of shape (height, width, channels).</p></li>
 <li><p><strong>max_stride</strong> – Scalar integer specifying the maximum stride that the image must be
@@ -700,7 +745,7 @@ <h1>sleap.nn.data.resizing</h1>
 reduction layers in the model.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The input image with 0-padding applied to the bottom and/or right such that the
 new shape’s height and width are both divisible by <code class="xref py py-obj docutils literal notranslate"><span class="pre">max_stride</span></code>.</p>
 </dd>
@@ -709,12 +754,12 @@ <h1>sleap.nn.data.resizing</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.resizing.resize_image">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">resize_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/resizing.py#L71-L105"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.resize_image" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.resizing.</span></span><span class="sig-name descname"><span class="pre">resize_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/resizing.py#L71-L105"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.resizing.resize_image" title="Permalink to this definition">#</a></dt>
 <dd><p>Rescale an image by a scale factor.</p>
 <p>This function is primarily a convenience wrapper for <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.image.resize</span></code> that
 calculates the new shape from the scale factor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image</strong> – Single image tensor of shape (height, width, channels).</p></li>
 <li><p><strong>scale</strong> – Factor to resize the image dimensions by, specified as either a float
@@ -722,7 +767,7 @@ <h1>sleap.nn.data.resizing</h1>
 dimensions are resized by the same factor.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The resized image tensor of the same dtype but scaled height and width.</p>
 </dd>
 </dl>
@@ -753,6 +798,57 @@ <h1>sleap.nn.data.resizing</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler"><code class="docutils literal notranslate"><span class="pre">PointsRescaler</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.input_keys"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.output_keys"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.PointsRescaler.transform_dataset"><code class="docutils literal notranslate"><span class="pre">PointsRescaler.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer"><code class="docutils literal notranslate"><span class="pre">Resizer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.image_key"><code class="docutils literal notranslate"><span class="pre">Resizer.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.scale_key"><code class="docutils literal notranslate"><span class="pre">Resizer.scale_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.points_key"><code class="docutils literal notranslate"><span class="pre">Resizer.points_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.scale"><code class="docutils literal notranslate"><span class="pre">Resizer.scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">Resizer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.keep_full_image"><code class="docutils literal notranslate"><span class="pre">Resizer.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.full_image_key"><code class="docutils literal notranslate"><span class="pre">Resizer.full_image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.from_config"><code class="docutils literal notranslate"><span class="pre">Resizer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.input_keys"><code class="docutils literal notranslate"><span class="pre">Resizer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.output_keys"><code class="docutils literal notranslate"><span class="pre">Resizer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.Resizer.transform_dataset"><code class="docutils literal notranslate"><span class="pre">Resizer.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher"><code class="docutils literal notranslate"><span class="pre">SizeMatcher</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.image_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.scale_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.scale_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.points_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.points_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.keep_full_image"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.keep_full_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.full_image_key"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.full_image_key</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.max_image_height"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.max_image_height</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.max_image_width"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.max_image_width</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.center_pad"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.center_pad</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.from_config"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.input_keys"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.output_keys"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.SizeMatcher.transform_dataset"><code class="docutils literal notranslate"><span class="pre">SizeMatcher.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.find_padding_for_stride"><code class="docutils literal notranslate"><span class="pre">find_padding_for_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">pad_to_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.resizing.resize_image"><code class="docutils literal notranslate"><span class="pre">resize_image()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -771,7 +867,7 @@ <h1>sleap.nn.data.resizing</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.training.html b/api/sleap.nn.data.training.html
index 7e063e604..5ebb87f1a 100644
--- a/api/sleap.nn.data.training.html
+++ b/api/sleap.nn.data.training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.training &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.training &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.training';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,23 @@ <h1>sleap.nn.data.training</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper"><code class="docutils literal notranslate"><span class="pre">KeyMapper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.key_maps"><code class="docutils literal notranslate"><span class="pre">KeyMapper.key_maps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyMapper.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyMapper.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyMapper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels"><code class="docutils literal notranslate"><span class="pre">split_labels()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels_reader"><code class="docutils literal notranslate"><span class="pre">split_labels_reader()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels_train_val"><code class="docutils literal notranslate"><span class="pre">split_labels_train_val()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +335,11 @@ <h1>sleap.nn.data.training</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.training">
-<span id="sleap-nn-data-training"></span><h1>sleap.nn.data.training<a class="headerlink" href="#module-sleap.nn.data.training" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-training"></span><h1>sleap.nn.data.training<a class="headerlink" href="#module-sleap.nn.data.training" title="Permalink to this heading">#</a></h1>
 <p>Transformers and utilities for training-related operations.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.data.training.KeyMapper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">KeyMapper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">key_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/training.py#L194-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.KeyMapper" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">KeyMapper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">key_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/training.py#L194-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.KeyMapper" title="Permalink to this definition">#</a></dt>
 <dd><p>Maps example keys to specified outputs.</p>
 <p>This is useful for transforming examples into tuples that map onto specific layer
 names for training.</p>
@@ -333,7 +350,7 @@ <h1>sleap.nn.data.training</h1>
 the form: {input_key: output_key}. If a list, the examples will be in tuples
 in the same order.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[Dict[str, str]]</p>
 </dd>
 </dl>
@@ -353,14 +370,14 @@ <h1>sleap.nn.data.training</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.data.training.KeyMapper.transform_dataset">
-<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/training.py#L226-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.KeyMapper.transform_dataset" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">transform_dataset</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds_input</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">DatasetV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/training.py#L226-L248"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.KeyMapper.transform_dataset" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a dataset with input keys mapped to new key names.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>ds_input</strong> – Any <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.data.Dataset</span></code> that generates examples as a dictionary of
 tensors with the keys in <a class="reference internal" href="#sleap.nn.data.training.KeyMapper.input_keys" title="sleap.nn.data.training.KeyMapper.input_keys"><code class="xref py py-obj docutils literal notranslate"><span class="pre">input_keys</span></code></a>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dataset that generates examples with the tensors in <a class="reference internal" href="#sleap.nn.data.training.KeyMapper.input_keys" title="sleap.nn.data.training.KeyMapper.input_keys"><code class="xref py py-obj docutils literal notranslate"><span class="pre">input_keys</span></code></a> mapped to
 keys in <a class="reference internal" href="#sleap.nn.data.training.KeyMapper.output_keys" title="sleap.nn.data.training.KeyMapper.output_keys"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_keys</span></code></a> according to the structure in <a class="reference internal" href="#sleap.nn.data.training.KeyMapper.key_maps" title="sleap.nn.data.training.KeyMapper.key_maps"><code class="xref py py-obj docutils literal notranslate"><span class="pre">key_maps</span></code></a>.</p>
 </dd>
@@ -371,10 +388,10 @@ <h1>sleap.nn.data.training</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.training.split_labels">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_fractions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/training.py#L65-L119"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_fractions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/training.py#L65-L119"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels" title="Permalink to this definition">#</a></dt>
 <dd><p>Split a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> into multiple new ones with random subsets of the data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – An instance of <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code>.</p></li>
 <li><p><strong>split_fractions</strong> – One or more floats between 0 and 1 that specify the fraction of
@@ -384,10 +401,10 @@ <h1>sleap.nn.data.training</h1>
 that it should contain all elements left over from the other splits.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tuple of new <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> instances of the same length as <code class="xref py py-obj docutils literal notranslate"><span class="pre">split_fractions</span></code>.</p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ValueError</strong> – If more than one split fraction is specified as -1.</p></li>
 <li><p><strong>ValueError</strong> – If the splits add up to more than the total available examples.</p></li>
@@ -402,10 +419,10 @@ <h1>sleap.nn.data.training</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.training.split_labels_reader">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels_reader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_fractions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/training.py#L122-L190"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels_reader" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels_reader</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_fractions</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/training.py#L122-L190"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels_reader" title="Permalink to this definition">#</a></dt>
 <dd><p>Split a <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> into multiple new ones with random subsets of the data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels_reader</strong> – An instance of <a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.providers.LabelsReader</span></code></a>. This is a
 provider that generates datasets that contain elements read from a
@@ -417,7 +434,7 @@ <h1>sleap.nn.data.training</h1>
 that it should contain all elements left over from the other splits.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instances of the same length as <code class="xref py py-obj docutils literal notranslate"><span class="pre">split_fractions</span></code>. The
 indices will be stored in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">example_indices</span></code> in each <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> instance.</p>
 <p>The actual <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> instance will be the same for each instance, only the
@@ -426,7 +443,7 @@ <h1>sleap.nn.data.training</h1>
 will be sampled to generate the splits.</p>
 </p>
 </dd>
-<dt class="field-odd">Raises</dt>
+<dt class="field-odd">Raises<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ValueError</strong> – If more than one split fraction is specified as -1.</p></li>
 <li><p><strong>ValueError</strong> – If the splits add up to more than the total available examples.</p></li>
@@ -441,16 +458,16 @@ <h1>sleap.nn.data.training</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.training.split_labels_train_val">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels_train_val</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_fraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/training.py#L14-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels_train_val" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.training.</span></span><span class="sig-name descname"><span class="pre">split_labels_train_val</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_fraction</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/training.py#L14-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.training.split_labels_train_val" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a train/validation split from a labels dataset.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> dataset with labeled frames.</p></li>
 <li><p><strong>validation_fraction</strong> – Fraction of frames to use for validation.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(labels_train,</span> <span class="pre">idx_train,</span> <span class="pre">labels_val,</span> <span class="pre">idx_val)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">labels_train</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">labels_val</span></code> are <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Label</span></code> objects containing the
 selected frames for each split. Their <code class="xref py py-obj docutils literal notranslate"><span class="pre">videos</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">tracks</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">provenance</span></code>
@@ -494,6 +511,29 @@ <h1>sleap.nn.data.training</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper"><code class="docutils literal notranslate"><span class="pre">KeyMapper</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.key_maps"><code class="docutils literal notranslate"><span class="pre">KeyMapper.key_maps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.input_keys"><code class="docutils literal notranslate"><span class="pre">KeyMapper.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.output_keys"><code class="docutils literal notranslate"><span class="pre">KeyMapper.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.KeyMapper.transform_dataset"><code class="docutils literal notranslate"><span class="pre">KeyMapper.transform_dataset()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels"><code class="docutils literal notranslate"><span class="pre">split_labels()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels_reader"><code class="docutils literal notranslate"><span class="pre">split_labels_reader()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.training.split_labels_train_val"><code class="docutils literal notranslate"><span class="pre">split_labels_train_val()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -512,7 +552,7 @@ <h1>sleap.nn.data.training</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.data.utils.html b/api/sleap.nn.data.utils.html
index a9f66307c..e5d3532ed 100644
--- a/api/sleap.nn.data.utils.html
+++ b/api/sleap.nn.data.utils.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.data.utils &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.data.utils &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.data.utils';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.data.utils.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,20 @@ <h1>sleap.nn.data.utils</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.describe_tensors"><code class="docutils literal notranslate"><span class="pre">describe_tensors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.ensure_list"><code class="docutils literal notranslate"><span class="pre">ensure_list()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.expand_to_rank"><code class="docutils literal notranslate"><span class="pre">expand_to_rank()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.gaussian_pdf"><code class="docutils literal notranslate"><span class="pre">gaussian_pdf()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.make_grid_vectors"><code class="docutils literal notranslate"><span class="pre">make_grid_vectors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.unrag_example"><code class="docutils literal notranslate"><span class="pre">unrag_example()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.unrag_tensor"><code class="docutils literal notranslate"><span class="pre">unrag_tensor()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,21 +332,21 @@ <h1>sleap.nn.data.utils</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.data.utils">
-<span id="sleap-nn-data-utils"></span><h1>sleap.nn.data.utils<a class="headerlink" href="#module-sleap.nn.data.utils" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-data-utils"></span><h1>sleap.nn.data.utils<a class="headerlink" href="#module-sleap.nn.data.utils" title="Permalink to this heading">#</a></h1>
 <p>Miscellaneous utility functions for data processing.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.describe_tensors">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">describe_tensors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L87-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.describe_tensors" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">describe_tensors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_description</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L87-L115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.describe_tensors" title="Permalink to this definition">#</a></dt>
 <dd><p>Print the keys in a example.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>example</strong> – Dictionary keyed by strings with tensors as values.</p></li>
 <li><p><strong>return_description</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, returns the string description instead of
 printing it.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>String description if <code class="xref py py-obj docutils literal notranslate"><span class="pre">return_description</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, otherwise <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 </dd>
 </dl>
@@ -340,16 +354,16 @@ <h1>sleap.nn.data.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.ensure_list">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">ensure_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L9-L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.ensure_list" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">ensure_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L9-L13"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.ensure_list" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert the input into a list if it is not already.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.expand_to_rank">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">expand_to_rank</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_rank</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prepend</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L16-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.expand_to_rank" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">expand_to_rank</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_rank</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prepend</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L16-L38"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.expand_to_rank" title="Permalink to this definition">#</a></dt>
 <dd><p>Expand a tensor to a target rank by adding singleton dimensions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – Any <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> with rank &lt;= <code class="xref py py-obj docutils literal notranslate"><span class="pre">target_rank</span></code>. If the rank is higher than
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">target_rank</span></code>, the tensor will be returned with the same shape.</p></li>
@@ -358,7 +372,7 @@ <h1>sleap.nn.data.utils</h1>
 data. If False, singleton dimensions are added after the last axis.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The expanded tensor of the same dtype as the input, but with rank <code class="xref py py-obj docutils literal notranslate"><span class="pre">target_rank</span></code>.</p>
 <p>The output has the same exact data as the input tensor and will be identical if
 they are both flattened.</p>
@@ -369,13 +383,13 @@ <h1>sleap.nn.data.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.gaussian_pdf">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">gaussian_pdf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L74-L84"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.gaussian_pdf" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">gaussian_pdf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L74-L84"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.gaussian_pdf" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the PDF of an unnormalized 0-centered Gaussian distribution.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>x</strong> – Any tensor of dtype tf.float32 with values to compute the PDF for.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of the same shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">x</span></code>, but with values of a PDF of an unnormalized
 Gaussian distribution. Values of 0 have an unnormalized PDF value of 1.0.</p>
 </dd>
@@ -384,13 +398,13 @@ <h1>sleap.nn.data.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.make_grid_vectors">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">make_grid_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L41-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.make_grid_vectors" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">make_grid_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">image_height</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">image_width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L41-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.make_grid_vectors" title="Permalink to this definition">#</a></dt>
 <dd><p>Make sampling grid vectors from image dimensions.</p>
 <p>This is a useful function for creating the x- and y-vectors that define a sampling
 grid over an image space. These vectors can be used to generate a full meshgrid or
 for equivalent broadcasting operations.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>image_height</strong> – Height of the image grid that will be sampled, specified as a
 scalar integer.</p></li>
@@ -403,7 +417,7 @@ <h1>sleap.nn.data.utils</h1>
 values greater than 1.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Tuple of grid vectors (xv, yv). These are tensors of dtype tf.float32 with
 shapes (grid_width,) and (grid_height,) respectively.</p>
 <dl class="simple">
@@ -418,16 +432,16 @@ <h1>sleap.nn.data.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.unrag_example">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">unrag_example</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L118-L146"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.unrag_example" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">unrag_example</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L118-L146"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.unrag_example" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert ragged tensors in an example into normal tensors with NaN padding.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>example</strong> – Dictionary keyed by strings with tensors as values.</p></li>
 <li><p><strong>numpy</strong> – If <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, convert values to numpy arrays or Python primitives.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The same dictionary, but values of type <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> will be converted to
 tensors of type <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> with NaN padding if the ragged dimensions are of
 variable length.</p>
@@ -442,12 +456,12 @@ <h1>sleap.nn.data.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.data.utils.unrag_tensor">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">unrag_tensor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/data/utils.py#L149-L176"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.unrag_tensor" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.data.utils.</span></span><span class="sig-name descname"><span class="pre">unrag_tensor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/data/utils.py#L149-L176"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.data.utils.unrag_tensor" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts a ragged tensor to a full tensor by padding to a maximum size.</p>
 <p>This function is useful for converting ragged tensors to a fixed size when one or
 more of the dimensions are of variable length.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x</strong> – Ragged tensor to convert.</p></li>
 <li><p><strong>max_size</strong> – Maximum size of the axis to pad.</p></li>
@@ -456,7 +470,7 @@ <h1>sleap.nn.data.utils</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">axis</span></code>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A padded version of <code class="xref py py-obj docutils literal notranslate"><span class="pre">x</span></code>. Padding will use the equivalent of NaNs in the tensor’s
 native dtype.</p>
 <p>This will replace the shape of the specified <code class="xref py py-obj docutils literal notranslate"><span class="pre">axis</span></code> with <code class="xref py py-obj docutils literal notranslate"><span class="pre">max_size</span></code>, leaving the
@@ -490,6 +504,26 @@ <h1>sleap.nn.data.utils</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.describe_tensors"><code class="docutils literal notranslate"><span class="pre">describe_tensors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.ensure_list"><code class="docutils literal notranslate"><span class="pre">ensure_list()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.expand_to_rank"><code class="docutils literal notranslate"><span class="pre">expand_to_rank()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.gaussian_pdf"><code class="docutils literal notranslate"><span class="pre">gaussian_pdf()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.make_grid_vectors"><code class="docutils literal notranslate"><span class="pre">make_grid_vectors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.unrag_example"><code class="docutils literal notranslate"><span class="pre">unrag_example()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.data.utils.unrag_tensor"><code class="docutils literal notranslate"><span class="pre">unrag_tensor()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -508,7 +542,7 @@ <h1>sleap.nn.data.utils</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.evals.html b/api/sleap.nn.evals.html
index d6691b349..ef59c102d 100644
--- a/api/sleap.nn.evals.html
+++ b/api/sleap.nn.evals.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.evals &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.evals &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.evals';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.evals.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,27 @@ <h1>sleap.nn.evals</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_dist_metrics"><code class="docutils literal notranslate"><span class="pre">compute_dist_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_dists"><code class="docutils literal notranslate"><span class="pre">compute_dists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_generalized_voc_metrics"><code class="docutils literal notranslate"><span class="pre">compute_generalized_voc_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_instance_area"><code class="docutils literal notranslate"><span class="pre">compute_instance_area()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_oks"><code class="docutils literal notranslate"><span class="pre">compute_oks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_pck_metrics"><code class="docutils literal notranslate"><span class="pre">compute_pck_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_visibility_conf"><code class="docutils literal notranslate"><span class="pre">compute_visibility_conf()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.evaluate"><code class="docutils literal notranslate"><span class="pre">evaluate()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.evaluate_model"><code class="docutils literal notranslate"><span class="pre">evaluate_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.find_frame_pairs"><code class="docutils literal notranslate"><span class="pre">find_frame_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.load_metrics"><code class="docutils literal notranslate"><span class="pre">load_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.match_frame_pairs"><code class="docutils literal notranslate"><span class="pre">match_frame_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.match_instances"><code class="docutils literal notranslate"><span class="pre">match_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.replace_path"><code class="docutils literal notranslate"><span class="pre">replace_path()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +339,7 @@ <h1>sleap.nn.evals</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.evals">
-<span id="sleap-nn-evals"></span><h1>sleap.nn.evals<a class="headerlink" href="#module-sleap.nn.evals" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-evals"></span><h1>sleap.nn.evals<a class="headerlink" href="#module-sleap.nn.evals" title="Permalink to this heading">#</a></h1>
 <p>Evaluation utilities for measuring pose estimation accuracy.</p>
 <p>To generate metrics, you’ll need two <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> datasets, one with ground truth
 data and one with predicted data. The video paths in the datasets must match.
@@ -342,13 +363,13 @@ <h1>sleap.nn.evals</h1>
 </div></blockquote>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_dist_metrics">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_dist_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dists_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L523-L553"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_dist_metrics" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_dist_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dists_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L523-L553"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_dist_metrics" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the Euclidean distance error at different percentiles.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>dists</strong> – An array of pairwise distances of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_positive_pairs,</span> <span class="pre">n_nodes)</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary of distance metrics.</p>
 </dd>
 </dl>
@@ -356,19 +377,19 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_dists">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_dists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L485-L520"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_dists" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_dists</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L485-L520"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_dists" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute Euclidean distances between matched pairs of instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>positive_pairs</strong> – A list of tuples of the form <code class="xref py py-obj docutils literal notranslate"><span class="pre">(instance_gt,</span> <span class="pre">instance_pr,</span> <span class="pre">_)</span></code>
 containing the matched pair of instances.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>dists: An array of pairwise distances of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_positive_pairs,</span> <span class="pre">n_nodes)</span></code>
 frame_idxs: A list of frame indices corresponding to the <code class="xref py py-obj docutils literal notranslate"><span class="pre">dists</span></code>
 video_paths: A list of video paths corresponding to the <code class="xref py py-obj docutils literal notranslate"><span class="pre">dists</span></code></p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dictionary with the following keys</p>
 </dd>
 </dl>
@@ -376,10 +397,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_generalized_voc_metrics">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_generalized_voc_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">false_negatives</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_score_thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([0.5,</span> <span class="pre">0.55,</span> <span class="pre">0.6,</span> <span class="pre">0.65,</span> <span class="pre">0.7,</span> <span class="pre">0.75,</span> <span class="pre">0.8,</span> <span class="pre">0.85,</span> <span class="pre">0.9,</span> <span class="pre">0.95])</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">recall_thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([0.0,</span> <span class="pre">0.01,</span> <span class="pre">0.02,</span> <span class="pre">0.03,</span> <span class="pre">0.04,</span> <span class="pre">0.05,</span> <span class="pre">0.06,</span> <span class="pre">0.07,</span> <span class="pre">0.08,</span> <span class="pre">0.09,</span> <span class="pre">0.1,</span> <span class="pre">0.11,</span> <span class="pre">0.12,</span> <span class="pre">0.13,</span> <span class="pre">0.14,</span> <span class="pre">0.15,</span> <span class="pre">0.16,</span> <span class="pre">0.17,</span> <span class="pre">0.18,</span> <span class="pre">0.19,</span> <span class="pre">0.2,</span> <span class="pre">0.21,</span> <span class="pre">0.22,</span> <span class="pre">0.23,</span> <span class="pre">0.24,</span> <span class="pre">0.25,</span> <span class="pre">0.26,</span> <span class="pre">0.27,</span> <span class="pre">0.28,</span> <span class="pre">0.29,</span> <span class="pre">0.3,</span> <span class="pre">0.31,</span> <span class="pre">0.32,</span> <span class="pre">0.33,</span> <span class="pre">0.34,</span> <span class="pre">0.35,</span> <span class="pre">0.36,</span> <span class="pre">0.37,</span> <span class="pre">0.38,</span> <span class="pre">0.39,</span> <span class="pre">0.4,</span> <span class="pre">0.41,</span> <span class="pre">0.42,</span> <span class="pre">0.43,</span> <span class="pre">0.44,</span> <span class="pre">0.45,</span> <span class="pre">0.46,</span> <span class="pre">0.47,</span> <span class="pre">0.48,</span> <span class="pre">0.49,</span> <span class="pre">0.5,</span> <span class="pre">0.51,</span> <span class="pre">0.52,</span> <span class="pre">0.53,</span> <span class="pre">0.54,</span> <span class="pre">0.55,</span> <span class="pre">0.56,</span> <span class="pre">0.57,</span> <span class="pre">0.58,</span> <span class="pre">0.59,</span> <span class="pre">0.6,</span> <span class="pre">0.61,</span> <span class="pre">0.62,</span> <span class="pre">0.63,</span> <span class="pre">0.64,</span> <span class="pre">0.65,</span> <span class="pre">0.66,</span> <span class="pre">0.67,</span> <span class="pre">0.68,</span> <span class="pre">0.69,</span> <span class="pre">0.7,</span> <span class="pre">0.71,</span> <span class="pre">0.72,</span> <span class="pre">0.73,</span> <span class="pre">0.74,</span> <span class="pre">0.75,</span> <span class="pre">0.76,</span> <span class="pre">0.77,</span> <span class="pre">0.78,</span> <span class="pre">0.79,</span> <span class="pre">0.8,</span> <span class="pre">0.81,</span> <span class="pre">0.82,</span> <span class="pre">0.83,</span> <span class="pre">0.84,</span> <span class="pre">0.85,</span> <span class="pre">0.86,</span> <span class="pre">0.87,</span> <span class="pre">0.88,</span> <span class="pre">0.89,</span> <span class="pre">0.9,</span> <span class="pre">0.91,</span> <span class="pre">0.92,</span> <span class="pre">0.93,</span> <span class="pre">0.94,</span> <span class="pre">0.95,</span> <span class="pre">0.96,</span> <span class="pre">0.97,</span> <span class="pre">0.98,</span> <span class="pre">0.99,</span> <span class="pre">1.0])</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'gvoc'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L402-L482"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_generalized_voc_metrics" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_generalized_voc_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">false_negatives</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_score_thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([0.5,</span> <span class="pre">0.55,</span> <span class="pre">0.6,</span> <span class="pre">0.65,</span> <span class="pre">0.7,</span> <span class="pre">0.75,</span> <span class="pre">0.8,</span> <span class="pre">0.85,</span> <span class="pre">0.9,</span> <span class="pre">0.95])</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">recall_thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([0.0,</span> <span class="pre">0.01,</span> <span class="pre">0.02,</span> <span class="pre">0.03,</span> <span class="pre">0.04,</span> <span class="pre">0.05,</span> <span class="pre">0.06,</span> <span class="pre">0.07,</span> <span class="pre">0.08,</span> <span class="pre">0.09,</span> <span class="pre">0.1,</span> <span class="pre">0.11,</span> <span class="pre">0.12,</span> <span class="pre">0.13,</span> <span class="pre">0.14,</span> <span class="pre">0.15,</span> <span class="pre">0.16,</span> <span class="pre">0.17,</span> <span class="pre">0.18,</span> <span class="pre">0.19,</span> <span class="pre">0.2,</span> <span class="pre">0.21,</span> <span class="pre">0.22,</span> <span class="pre">0.23,</span> <span class="pre">0.24,</span> <span class="pre">0.25,</span> <span class="pre">0.26,</span> <span class="pre">0.27,</span> <span class="pre">0.28,</span> <span class="pre">0.29,</span> <span class="pre">0.3,</span> <span class="pre">0.31,</span> <span class="pre">0.32,</span> <span class="pre">0.33,</span> <span class="pre">0.34,</span> <span class="pre">0.35,</span> <span class="pre">0.36,</span> <span class="pre">0.37,</span> <span class="pre">0.38,</span> <span class="pre">0.39,</span> <span class="pre">0.4,</span> <span class="pre">0.41,</span> <span class="pre">0.42,</span> <span class="pre">0.43,</span> <span class="pre">0.44,</span> <span class="pre">0.45,</span> <span class="pre">0.46,</span> <span class="pre">0.47,</span> <span class="pre">0.48,</span> <span class="pre">0.49,</span> <span class="pre">0.5,</span> <span class="pre">0.51,</span> <span class="pre">0.52,</span> <span class="pre">0.53,</span> <span class="pre">0.54,</span> <span class="pre">0.55,</span> <span class="pre">0.56,</span> <span class="pre">0.57,</span> <span class="pre">0.58,</span> <span class="pre">0.59,</span> <span class="pre">0.6,</span> <span class="pre">0.61,</span> <span class="pre">0.62,</span> <span class="pre">0.63,</span> <span class="pre">0.64,</span> <span class="pre">0.65,</span> <span class="pre">0.66,</span> <span class="pre">0.67,</span> <span class="pre">0.68,</span> <span class="pre">0.69,</span> <span class="pre">0.7,</span> <span class="pre">0.71,</span> <span class="pre">0.72,</span> <span class="pre">0.73,</span> <span class="pre">0.74,</span> <span class="pre">0.75,</span> <span class="pre">0.76,</span> <span class="pre">0.77,</span> <span class="pre">0.78,</span> <span class="pre">0.79,</span> <span class="pre">0.8,</span> <span class="pre">0.81,</span> <span class="pre">0.82,</span> <span class="pre">0.83,</span> <span class="pre">0.84,</span> <span class="pre">0.85,</span> <span class="pre">0.86,</span> <span class="pre">0.87,</span> <span class="pre">0.88,</span> <span class="pre">0.89,</span> <span class="pre">0.9,</span> <span class="pre">0.91,</span> <span class="pre">0.92,</span> <span class="pre">0.93,</span> <span class="pre">0.94,</span> <span class="pre">0.95,</span> <span class="pre">0.96,</span> <span class="pre">0.97,</span> <span class="pre">0.98,</span> <span class="pre">0.99,</span> <span class="pre">1.0])</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'gvoc'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L402-L482"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_generalized_voc_metrics" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute VOC metrics given matched pairs of instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>positive_pairs</strong> – A list of tuples of the form <code class="xref py py-obj docutils literal notranslate"><span class="pre">(instance_gt,</span> <span class="pre">instance_pr,</span> <span class="pre">_)</span></code>
 containing the matched pair of instances.</p></li>
@@ -392,7 +413,7 @@ <h1>sleap.nn.evals</h1>
 <li><p><strong>name</strong> – Name to use to prefix returned metric keys.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary of VOC metrics.</p>
 </dd>
 </dl>
@@ -400,13 +421,13 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_instance_area">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_instance_area</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L116-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_instance_area" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_instance_area</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L116-L131"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_instance_area" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the area of the bounding box of a set of keypoints.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>points</strong> – A numpy array of coordinates.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The area of the bounding box of the points.</p>
 </dd>
 </dl>
@@ -414,10 +435,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_oks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_oks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cocoeval</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L134-L250"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_oks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_oks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_cocoeval</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L134-L250"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_oks" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the object keypoints similarity between sets of points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points_gt</strong> – Ground truth instances of shape (n_gt, n_nodes, n_ed),
 where n_nodes is the number of body parts/keypoint types, and n_ed
@@ -441,7 +462,7 @@ <h1>sleap.nn.evals</h1>
 to reflect the smaller spread expected in localizing it.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The object keypoints similarity between every pair of ground truth and
 predicted instance, a numpy array of of shape (n_gt, n_pr) in the range
 of [0, 1.0], with 1.0 denoting a perfect match.</p>
@@ -464,16 +485,16 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_pck_metrics">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_pck_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dists</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([1.0,</span> <span class="pre">2.0,</span> <span class="pre">3.0,</span> <span class="pre">4.0,</span> <span class="pre">5.0,</span> <span class="pre">6.0,</span> <span class="pre">7.0,</span> <span class="pre">8.0,</span> <span class="pre">9.0,</span> <span class="pre">10.0])</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L556-L579"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_pck_metrics" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_pck_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dists</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresholds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">array([1.0,</span> <span class="pre">2.0,</span> <span class="pre">3.0,</span> <span class="pre">4.0,</span> <span class="pre">5.0,</span> <span class="pre">6.0,</span> <span class="pre">7.0,</span> <span class="pre">8.0,</span> <span class="pre">9.0,</span> <span class="pre">10.0])</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L556-L579"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_pck_metrics" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute PCK across a range of thresholds.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>dists</strong> – An array of pairwise distances of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_positive_pairs,</span> <span class="pre">n_nodes)</span></code>.</p></li>
 <li><p><strong>thresholds</strong> – A list of distance thresholds in pixels.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary of PCK metrics evaluated at each threshold.</p>
 </dd>
 </dl>
@@ -481,14 +502,14 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.compute_visibility_conf">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_visibility_conf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L582-L615"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_visibility_conf" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">compute_visibility_conf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">positive_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L582-L615"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.compute_visibility_conf" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute node visibility metrics.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>positive_pairs</strong> – A list of tuples of the form <code class="xref py py-obj docutils literal notranslate"><span class="pre">(instance_gt,</span> <span class="pre">instance_pr,</span> <span class="pre">_)</span></code>
 containing the matched pair of instances.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary of visibility metrics, including the confusion matrix.</p>
 </dd>
 </dl>
@@ -496,10 +517,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.evaluate">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">evaluate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">oks_stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">oks_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L618-L680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.evaluate" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">evaluate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">oks_stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">oks_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L618-L680"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.evaluate" title="Permalink to this definition">#</a></dt>
 <dd><p>Calculate all metrics from ground truth and predicted labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels_gt</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> dataset object with ground truth labels.</p></li>
 <li><p><strong>labels_pr</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> dataset object with predicted labels.</p></li>
@@ -513,7 +534,7 @@ <h1>sleap.nn.evals</h1>
 considered for matching.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Dict, keys are strings, values are metrics (floats or ndarrays).</p>
 </dd>
 </dl>
@@ -521,10 +542,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.evaluate_model">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">evaluate_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cfg</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">save</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'test'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L683-L773"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.evaluate_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">evaluate_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cfg</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">save</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'test'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L683-L773"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.evaluate_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Evaluate a trained model and save metrics and predictions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cfg</strong> – The <code class="xref py py-obj docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code> associated with the model.</p></li>
 <li><p><strong>labels_gt</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> pipeline generator that reads the ground truth
@@ -534,7 +555,7 @@ <h1>sleap.nn.evals</h1>
 <li><p><strong>split_name</strong> – String name to append to the saved filenames.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(labels_pr,</span> <span class="pre">metrics)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">labels_pr</span></code> will contain the predicted labels.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">metrics</span></code> will contain the evaluated metrics given the predictions, or None if
@@ -546,10 +567,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.find_frame_pairs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">find_frame_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L61-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.find_frame_pairs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">find_frame_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L61-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.find_frame_pairs" title="Permalink to this definition">#</a></dt>
 <dd><p>Find corresponding frames across two sets of labels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>labels_gt</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance with ground truth instances.</p></li>
 <li><p><strong>labels_pr</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> instance with predicted instances.</p></li>
@@ -557,7 +578,7 @@ <h1>sleap.nn.evals</h1>
 also be considered for matching.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of pairs of <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.LabeledFrame`s</span> <span class="pre">in</span> <span class="pre">the</span> <span class="pre">form</span> <span class="pre">`(frame_gt,</span> <span class="pre">frame_pr)</span></code>.</p>
 </dd>
 </dl>
@@ -565,10 +586,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.load_metrics">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">load_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'val'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L776-L815"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.load_metrics" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">load_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">split</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'val'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L776-L815"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.load_metrics" title="Permalink to this definition">#</a></dt>
 <dd><p>Load metrics for a model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to a model folder or metrics file (.npz).</p></li>
 <li><p><strong>split</strong> – Name of the split to load the metrics for. Must be <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;train&quot;</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;val&quot;</span></code> or
@@ -576,7 +597,7 @@ <h1>sleap.nn.evals</h1>
 provided.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><ul class="simple">
 <li><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;vis.tp&quot;</span></code>: Visibility - True Positives</p></li>
 <li><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;vis.fp&quot;</span></code>: Visibility - False Positives</p></li>
@@ -602,7 +623,7 @@ <h1>sleap.nn.evals</h1>
 </ul>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The loaded metrics as a dictionary with keys</p>
 </dd>
 </dl>
@@ -610,11 +631,11 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.match_frame_pairs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">match_frame_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L353-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.match_frame_pairs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">match_frame_pairs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_pairs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L353-L399"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.match_frame_pairs" title="Permalink to this definition">#</a></dt>
 <dd><p>Match all ground truth and predicted instances within each pair of frames.</p>
 <p>This is a wrapper for <a class="reference internal" href="#sleap.nn.evals.match_instances" title="sleap.nn.evals.match_instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">match_instances()</span></code></a> but operates on lists of frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>frame_pairs</strong> – A list of pairs of <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.LabeledFrame`s</span> <span class="pre">in</span> <span class="pre">the</span> <span class="pre">form</span>
 <span class="pre">`(frame_gt,</span> <span class="pre">frame_pr)</span></code>. These can be obtained with <a class="reference internal" href="#sleap.nn.evals.find_frame_pairs" title="sleap.nn.evals.find_frame_pairs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">find_frame_pairs()</span></code></a>.</p></li>
@@ -627,7 +648,7 @@ <h1>sleap.nn.evals</h1>
 considered for matching.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (<code class="xref py py-obj docutils literal notranslate"><span class="pre">positive_pairs</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">false_negatives</span></code>).</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">positive_pairs</span></code> is a list of 3-tuples of the form
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(instance_gt,</span> <span class="pre">instance_pr,</span> <span class="pre">oks)</span></code> containing the matched pair of instances and
@@ -641,10 +662,10 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.match_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">match_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">sleap.instance.PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L253-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.match_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">match_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">frame_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">stddev</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.025</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">user_labels_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.instance.html#sleap.instance.PredictedInstance" title="sleap.instance.PredictedInstance"><span class="pre">PredictedInstance</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L253-L350"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.match_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Match pairs of instances between ground truth and predictions in a frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>frame_gt</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.LabeledFrame</span></code> with ground truth instances.</p></li>
 <li><p><strong>frame_pr</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.LabeledFrame</span></code> with predicted instances.</p></li>
@@ -657,7 +678,7 @@ <h1>sleap.nn.evals</h1>
 considered for matching.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (<code class="xref py py-obj docutils literal notranslate"><span class="pre">positive_pairs</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">false_negatives</span></code>).</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">positive_pairs</span></code> is a list of 3-tuples of the form
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(instance_gt,</span> <span class="pre">instance_pr,</span> <span class="pre">oks)</span></code> containing the matched pair of instances and
@@ -679,7 +700,7 @@ <h1>sleap.nn.evals</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.evals.replace_path">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">replace_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/evals.py#L52-L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.replace_path" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.evals.</span></span><span class="sig-name descname"><span class="pre">replace_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">video_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">dict</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/evals.py#L52-L58"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.evals.replace_path" title="Permalink to this definition">#</a></dt>
 <dd><p>Replace video paths in unstructured video objects.</p>
 </dd></dl>
 
@@ -707,6 +728,33 @@ <h1>sleap.nn.evals</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_dist_metrics"><code class="docutils literal notranslate"><span class="pre">compute_dist_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_dists"><code class="docutils literal notranslate"><span class="pre">compute_dists()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_generalized_voc_metrics"><code class="docutils literal notranslate"><span class="pre">compute_generalized_voc_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_instance_area"><code class="docutils literal notranslate"><span class="pre">compute_instance_area()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_oks"><code class="docutils literal notranslate"><span class="pre">compute_oks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_pck_metrics"><code class="docutils literal notranslate"><span class="pre">compute_pck_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.compute_visibility_conf"><code class="docutils literal notranslate"><span class="pre">compute_visibility_conf()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.evaluate"><code class="docutils literal notranslate"><span class="pre">evaluate()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.evaluate_model"><code class="docutils literal notranslate"><span class="pre">evaluate_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.find_frame_pairs"><code class="docutils literal notranslate"><span class="pre">find_frame_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.load_metrics"><code class="docutils literal notranslate"><span class="pre">load_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.match_frame_pairs"><code class="docutils literal notranslate"><span class="pre">match_frame_pairs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.match_instances"><code class="docutils literal notranslate"><span class="pre">match_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.evals.replace_path"><code class="docutils literal notranslate"><span class="pre">replace_path()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -725,7 +773,7 @@ <h1>sleap.nn.evals</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.heads.html b/api/sleap.nn.heads.html
index 1e01c430d..b0199e23a 100644
--- a/api/sleap.nn.heads.html
+++ b/api/sleap.nn.heads.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.heads &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.heads &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.heads';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.heads.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,115 @@ <h1>sleap.nn.heads</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.anchor_part"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id0"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.anchor_part"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.config</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.classes"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.activation"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.channels"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.config"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id1"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.classes</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.classes"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.num_fc_layers"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.num_fc_layers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.num_fc_units"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.num_fc_units</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.activation"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.channels"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.from_config"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.config"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id2"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.loss_function"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.loss_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.make_head"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.make_head()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head"><code class="docutils literal notranslate"><span class="pre">Head</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.activation"><code class="docutils literal notranslate"><span class="pre">Head.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.channels"><code class="docutils literal notranslate"><span class="pre">Head.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.loss_function"><code class="docutils literal notranslate"><span class="pre">Head.loss_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.make_head"><code class="docutils literal notranslate"><span class="pre">Head.make_head()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id3"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.part_names"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.sigma_threshold"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.sigma_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.output_stride"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.channels"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.from_config"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.config"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id4"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id5"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.sigma_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id6"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.edges"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.channels"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.from_config"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.config"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id7"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.edges</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id8"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,18 +427,18 @@ <h1>sleap.nn.heads</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.heads">
-<span id="sleap-nn-heads"></span><h1>sleap.nn.heads<a class="headerlink" href="#module-sleap.nn.heads" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-heads"></span><h1>sleap.nn.heads<a class="headerlink" href="#module-sleap.nn.heads" title="Permalink to this heading">#</a></h1>
 <p>Model head definitions for defining model output types.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.CenteredInstanceConfmapsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">CenteredInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L158-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">CenteredInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L158-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying centered instance confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.CenteredInstanceConfmapsHead.part_names">
 <span class="sig-name descname"><span class="pre">part_names</span></span><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.part_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of strings specifying the part names associated with channels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -341,7 +450,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Name of the part to use as an anchor node. If not specified, the
 bounding box centroid will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -352,7 +461,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -364,7 +473,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -375,7 +484,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -389,7 +498,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.CenteredInstanceConfmapsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">sleap.nn.heads.CenteredInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L182-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.CenteredInstanceConfmapsHead" title="sleap.nn.heads.CenteredInstanceConfmapsHead"><span class="pre">CenteredInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L182-L209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.CenteredInstanceConfmapsHead.config">
@@ -408,7 +517,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -418,7 +527,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.CentroidConfmapsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L117-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">anchor_part</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L117-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying instance centroid confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.CentroidConfmapsHead.anchor_part">
@@ -426,7 +535,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Name of the part to use as an anchor node. If not specified, the
 bounding box centroid will be used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -437,7 +546,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -449,7 +558,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -460,7 +569,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -474,7 +583,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.CentroidConfmapsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig"><span class="pre">sleap.nn.config.model.CentroidsHeadConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.CentroidConfmapsHead" title="sleap.nn.heads.CentroidConfmapsHead"><span class="pre">sleap.nn.heads.CentroidConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L139-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig"><span class="pre">CentroidsHeadConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.CentroidConfmapsHead" title="sleap.nn.heads.CentroidConfmapsHead"><span class="pre">CentroidConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L139-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.CentroidConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.CentroidConfmapsHead.config">
@@ -483,7 +592,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -493,14 +602,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassMapsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">ClassMapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L314-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">ClassMapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L314-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying class identity maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassMapsHead.classes">
 <span class="sig-name descname"><span class="pre">classes</span></span><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead.classes" title="Permalink to this definition">#</a></dt>
 <dd><p>List of string names of the classes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -511,7 +620,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the class maps around each node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -523,7 +632,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -534,7 +643,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -554,7 +663,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassMapsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig"><span class="pre">sleap.nn.config.model.ClassMapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.ClassMapsHead" title="sleap.nn.heads.ClassMapsHead"><span class="pre">sleap.nn.heads.ClassMapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L340-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ClassMapsHeadConfig" title="sleap.nn.config.model.ClassMapsHeadConfig"><span class="pre">ClassMapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.ClassMapsHead" title="sleap.nn.heads.ClassMapsHead"><span class="pre">ClassMapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L340-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassMapsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassMapsHead.config">
@@ -571,7 +680,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -581,14 +690,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassVectorsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">ClassVectorsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_layers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_units</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">global_pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L368-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">ClassVectorsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_layers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_fc_units</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">64</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">global_pool</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L368-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying classification heads.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassVectorsHead.classes">
 <span class="sig-name descname"><span class="pre">classes</span></span><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.classes" title="Permalink to this definition">#</a></dt>
 <dd><p>List of string names of the classes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -599,7 +708,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">num_fc_layers</span></span><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.num_fc_layers" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of fully connected layers after flattening input features.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -611,7 +720,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Number of units (dimensions) in fully connected layers prior to
 classification output.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -623,7 +732,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -634,7 +743,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -654,7 +763,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassVectorsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig"><span class="pre">sleap.nn.config.model.ClassVectorsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.ClassVectorsHead" title="sleap.nn.heads.ClassVectorsHead"><span class="pre">sleap.nn.heads.ClassVectorsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L403-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ClassVectorsHeadConfig" title="sleap.nn.config.model.ClassVectorsHeadConfig"><span class="pre">ClassVectorsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">classes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.ClassVectorsHead" title="sleap.nn.heads.ClassVectorsHead"><span class="pre">ClassVectorsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L403-L429"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassVectorsHead.config">
@@ -671,7 +780,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -685,17 +794,17 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.ClassVectorsHead.make_head">
-<span class="sig-name descname"><span class="pre">make_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L431-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.make_head" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L431-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.ClassVectorsHead.make_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Make head output tensor from input feature tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – An input <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code>.</p></li>
 <li><p><strong>name</strong> – If provided, specifies the name of the output layer. If not (the
 default), uses the name of the head as the layer name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> with the correct shape for the head.</p>
 </dd>
 </dl>
@@ -705,7 +814,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.Head">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">Head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L20-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.Head" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">Head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L20-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.Head" title="Permalink to this definition">#</a></dt>
 <dd><p>Base class for model output heads.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.heads.Head.activation">
@@ -727,17 +836,17 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.Head.make_head">
-<span class="sig-name descname"><span class="pre">make_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L42-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.Head.make_head" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x_in</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L42-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.Head.make_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Make head output tensor from input feature tensor.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>x_in</strong> – An input <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code>.</p></li>
 <li><p><strong>name</strong> – If provided, specifies the name of the output layer. If not (the
 default), uses the name of the head as the layer name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> with the correct shape for the head.</p>
 </dd>
 </dl>
@@ -747,14 +856,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.MultiInstanceConfmapsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L213-L260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">MultiInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L213-L260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying multi-instance confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.MultiInstanceConfmapsHead.part_names">
 <span class="sig-name descname"><span class="pre">part_names</span></span><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead.part_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of strings specifying the part names associated with channels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -765,7 +874,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -777,7 +886,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -788,7 +897,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -802,7 +911,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.MultiInstanceConfmapsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">sleap.nn.heads.MultiInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L234-L260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">MultiInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.MultiInstanceConfmapsHead" title="sleap.nn.heads.MultiInstanceConfmapsHead"><span class="pre">MultiInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L234-L260"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.MultiInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.MultiInstanceConfmapsHead.config">
@@ -821,7 +930,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -831,14 +940,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.OffsetRefinementHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">OffsetRefinementHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L472-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">OffsetRefinementHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L472-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying offset refinement maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.OffsetRefinementHead.part_names">
 <span class="sig-name descname"><span class="pre">part_names</span></span><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead.part_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of strings specifying the part names associated with channels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -850,7 +959,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Threshold of confidence map values to use for defining the
 boundary of the offset maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -862,7 +971,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -873,7 +982,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -887,7 +996,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.OffsetRefinementHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig"><span class="pre">sleap.nn.config.model.CentroidsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.SingleInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.MultiInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">sleap.nn.heads.OffsetRefinementHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L494-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CentroidsHeadConfig" title="sleap.nn.config.model.CentroidsHeadConfig"><span class="pre">CentroidsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><span class="pre">SingleInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig" title="sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"><span class="pre">CenteredInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfmapsHeadConfig" title="sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"><span class="pre">MultiInstanceConfmapsHeadConfig</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.OffsetRefinementHead" title="sleap.nn.heads.OffsetRefinementHead"><span class="pre">OffsetRefinementHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L494-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.OffsetRefinementHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.OffsetRefinementHead.config">
@@ -918,7 +1027,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -928,14 +1037,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.PartAffinityFieldsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L264-L310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">PartAffinityFieldsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">15.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L264-L310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying multi-instance part affinity fields.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.PartAffinityFieldsHead.edges">
 <span class="sig-name descname"><span class="pre">edges</span></span><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead.edges" title="Permalink to this definition">#</a></dt>
 <dd><p>List of tuples of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(source,</span> <span class="pre">destination)</span></code> node names.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Sequence[Tuple[str, str]]</p>
 </dd>
 </dl>
@@ -946,7 +1055,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the part affinity fields.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -958,7 +1067,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -969,7 +1078,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -983,7 +1092,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.PartAffinityFieldsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig"><span class="pre">sleap.nn.config.model.PartAffinityFieldsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.PartAffinityFieldsHead" title="sleap.nn.heads.PartAffinityFieldsHead"><span class="pre">sleap.nn.heads.PartAffinityFieldsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L285-L310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.PartAffinityFieldsHeadConfig" title="sleap.nn.config.model.PartAffinityFieldsHeadConfig"><span class="pre">PartAffinityFieldsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Sequence</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.PartAffinityFieldsHead" title="sleap.nn.heads.PartAffinityFieldsHead"><span class="pre">PartAffinityFieldsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L285-L310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.PartAffinityFieldsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.PartAffinityFieldsHead.config">
@@ -1001,7 +1110,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -1011,14 +1120,14 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.heads.SingleInstanceConfmapsHead">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L66-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.heads.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceConfmapsHead</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L66-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead" title="Permalink to this definition">#</a></dt>
 <dd><p>Head for specifying single instance confidence maps.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.SingleInstanceConfmapsHead.part_names">
 <span class="sig-name descname"><span class="pre">part_names</span></span><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead.part_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of strings specifying the part names associated with channels.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -1029,7 +1138,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">sigma</span></span><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead.sigma" title="Permalink to this definition">#</a></dt>
 <dd><p>Spread of the confidence maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1041,7 +1150,7 @@ <h1>sleap.nn.heads</h1>
 <dd><p>Stride of the output head tensor. The input tensor is expected to
 be at the same stride.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1052,7 +1161,7 @@ <h1>sleap.nn.heads</h1>
 <span class="sig-name descname"><span class="pre">loss_weight</span></span><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead.loss_weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the loss term for this head during optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1066,7 +1175,7 @@ <h1>sleap.nn.heads</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.heads.SingleInstanceConfmapsHead.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><span class="pre">sleap.nn.config.model.SingleInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.SingleInstanceConfmapsHead" title="sleap.nn.heads.SingleInstanceConfmapsHead"><span class="pre">sleap.nn.heads.SingleInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/heads.py#L87-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.SingleInstanceConfmapsHeadConfig" title="sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"><span class="pre">SingleInstanceConfmapsHeadConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.heads.SingleInstanceConfmapsHead" title="sleap.nn.heads.SingleInstanceConfmapsHead"><span class="pre">SingleInstanceConfmapsHead</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/heads.py#L87-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.heads.SingleInstanceConfmapsHead.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create this head from a set of configurations.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.heads.SingleInstanceConfmapsHead.config">
@@ -1085,7 +1194,7 @@ <h1>sleap.nn.heads</h1>
 </dd></dl>
 
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The instantiated head with the specified configuration options.</p>
 </dd>
 </dl>
@@ -1117,6 +1226,121 @@ <h1>sleap.nn.heads</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.anchor_part"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CenteredInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id0"><code class="docutils literal notranslate"><span class="pre">CenteredInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.anchor_part"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.anchor_part</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.CentroidConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsHead.config</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.classes"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.activation"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.channels"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassMapsHead.config"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id1"><code class="docutils literal notranslate"><span class="pre">ClassMapsHead.classes</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.classes"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.num_fc_layers"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.num_fc_layers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.num_fc_units"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.num_fc_units</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.activation"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.channels"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.from_config"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.config"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id2"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.classes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.loss_function"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.loss_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.ClassVectorsHead.make_head"><code class="docutils literal notranslate"><span class="pre">ClassVectorsHead.make_head()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head"><code class="docutils literal notranslate"><span class="pre">Head</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.activation"><code class="docutils literal notranslate"><span class="pre">Head.activation</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.channels"><code class="docutils literal notranslate"><span class="pre">Head.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.loss_function"><code class="docutils literal notranslate"><span class="pre">Head.loss_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.Head.make_head"><code class="docutils literal notranslate"><span class="pre">Head.make_head()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.MultiInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id3"><code class="docutils literal notranslate"><span class="pre">MultiInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.part_names"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.sigma_threshold"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.sigma_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.output_stride"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.channels"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.from_config"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.OffsetRefinementHead.config"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id4"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id5"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.sigma_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id6"><code class="docutils literal notranslate"><span class="pre">OffsetRefinementHead.loss_weight</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.edges"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.sigma"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.channels"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.from_config"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.PartAffinityFieldsHead.config"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id7"><code class="docutils literal notranslate"><span class="pre">PartAffinityFieldsHead.edges</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.part_names"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.sigma"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.sigma</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.loss_weight"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.loss_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.channels"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.channels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.from_config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.heads.SingleInstanceConfmapsHead.config"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#id8"><code class="docutils literal notranslate"><span class="pre">SingleInstanceConfmapsHead.part_names</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1135,7 +1359,7 @@ <h1>sleap.nn.heads</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.identity.html b/api/sleap.nn.identity.html
index 1c9d2e3be..2badaa3a3 100644
--- a/api/sleap.nn.identity.html
+++ b/api/sleap.nn.identity.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.identity &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.identity &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.identity';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.identity.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,16 @@ <h1>sleap.nn.identity</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.classify_peaks_from_maps"><code class="docutils literal notranslate"><span class="pre">classify_peaks_from_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.classify_peaks_from_vectors"><code class="docutils literal notranslate"><span class="pre">classify_peaks_from_vectors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.group_class_peaks"><code class="docutils literal notranslate"><span class="pre">group_class_peaks()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,16 +328,16 @@ <h1>sleap.nn.identity</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.identity">
-<span id="sleap-nn-identity"></span><h1>sleap.nn.identity<a class="headerlink" href="#module-sleap.nn.identity" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-identity"></span><h1>sleap.nn.identity<a class="headerlink" href="#module-sleap.nn.identity" title="Permalink to this heading">#</a></h1>
 <p>Utilities for models that learn identity.</p>
 <p>These functions implement the inference logic for classifying peaks using class maps or
 classification vectors.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.identity.classify_peaks_from_maps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">classify_peaks_from_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">class_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/identity.py#L97-L179"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.classify_peaks_from_maps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">classify_peaks_from_maps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">class_maps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/identity.py#L97-L179"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.classify_peaks_from_maps" title="Permalink to this definition">#</a></dt>
 <dd><p>Classify and group local peaks by their class map probability.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>class_maps</strong> – Class maps for a batch as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> and
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">n_classes)</span></code>.</p></li>
@@ -342,7 +352,7 @@ <h1>sleap.nn.identity</h1>
 <li><p><strong>n_channels</strong> – Integer number of channels (nodes) the instances should have.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(points,</span> <span class="pre">point_vals,</span> <span class="pre">class_probs)</span></code> containing the grouped peaks.</p>
 <dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">points</span></code>: Class-grouped peaks as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> and shape</dt><dd><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">n_classes,</span> <span class="pre">n_channels,</span> <span class="pre">2)</span></code>. Missing points will be denoted by
@@ -361,11 +371,11 @@ <h1>sleap.nn.identity</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.identity.classify_peaks_from_vectors">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">classify_peaks_from_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_class_probs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/identity.py#L182-L250"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.classify_peaks_from_vectors" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">classify_peaks_from_vectors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_class_probs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/identity.py#L182-L250"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.classify_peaks_from_vectors" title="Permalink to this definition">#</a></dt>
 <dd><p>Group peaks by classification probabilities.</p>
 <p>This is used in top-down classification models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peak_points</strong> – </p></li>
 <li><p><strong>peak_vals</strong> – </p></li>
@@ -374,7 +384,7 @@ <h1>sleap.nn.identity</h1>
 <li><p><strong>n_samples</strong> – Number of samples in the batch.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(points,</span> <span class="pre">point_vals,</span> <span class="pre">class_probs)</span></code>.</p>
 <dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">points</span></code>: Class-grouped peaks as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> and shape</dt><dd><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">n_classes,</span> <span class="pre">n_channels,</span> <span class="pre">2)</span></code>. Missing points will be denoted by
@@ -392,12 +402,12 @@ <h1>sleap.nn.identity</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.identity.group_class_peaks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">group_class_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_class_probs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/identity.py#L13-L94"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.group_class_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.identity.</span></span><span class="sig-name descname"><span class="pre">group_class_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_class_probs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_samples</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_channels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/identity.py#L13-L94"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.identity.group_class_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Group local peaks using class probabilities.</p>
 <p>This is useful for matching peaks that span multiple samples and channels into
 classes (e.g., instance identities) by their class probability.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peak_class_probs</strong> – Class probabilities for each peak as <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of dtype
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> and shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_peaks,</span> <span class="pre">n_classes)</span></code>.</p></li>
@@ -409,7 +419,7 @@ <h1>sleap.nn.identity</h1>
 <li><p><strong>n_channels</strong> – Integer number of channels (nodes) the instances should have.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(peak_inds,</span> <span class="pre">class_inds)</span></code>.</p>
 <dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">peak_inds</span></code>: Indices of class-grouped peaks within <code class="xref py py-obj docutils literal notranslate"><span class="pre">[0,</span> <span class="pre">n_peaks)</span></code>. Will be at</dt><dd><p>most <code class="xref py py-obj docutils literal notranslate"><span class="pre">n_classes</span></code> long.</p>
@@ -453,6 +463,22 @@ <h1>sleap.nn.identity</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.classify_peaks_from_maps"><code class="docutils literal notranslate"><span class="pre">classify_peaks_from_maps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.classify_peaks_from_vectors"><code class="docutils literal notranslate"><span class="pre">classify_peaks_from_vectors()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.identity.group_class_peaks"><code class="docutils literal notranslate"><span class="pre">group_class_peaks()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -471,7 +497,7 @@ <h1>sleap.nn.identity</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.inference.html b/api/sleap.nn.inference.html
index 24cda7059..5681a012d 100644
--- a/api/sleap.nn.inference.html
+++ b/api/sleap.nn.inference.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.inference &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.inference &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.inference';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.inference.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,306 @@ <h1>sleap.nn.inference</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.paf_scorer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.paf_scorer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.cm_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.cm_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.paf_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.paf_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_pafs"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_pafs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_paf_graph"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_paf_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.pafs_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.pafs_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.find_peaks"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.find_peaks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.forward_pass"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.forward_pass()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel.bottomup_layer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel.bottomup_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.cm_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.cm_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.class_maps_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_class_maps"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.return_class_maps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.class_maps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.find_peaks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.forward_pass()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.inference_layer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel.inference_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.tracks"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.bottomup_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.bottomup_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.bottomup_model"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.bottomup_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.max_edge_length_ratio"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.max_edge_length_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.dist_penalty_weight"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.dist_penalty_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.paf_line_points"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.paf_line_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.min_line_scores"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.min_line_scores</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.max_instances"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop"><code class="docutils literal notranslate"><span class="pre">CentroidCrop</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.keras_model"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.crop_size"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.input_scale"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.peak_threshold"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.refinement"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.return_confmaps"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.offsets_ind"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.return_crops"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.return_crops</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.max_instances"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.call"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth.crop_size"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth.call"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.keras_model"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.input_scale"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.output_stride"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.peak_threshold"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.refinement"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.return_confmaps"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.offsets_ind"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.call"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaksGroundTruth"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaksGroundTruth</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaksGroundTruth.call"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaksGroundTruth.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer"><code class="docutils literal notranslate"><span class="pre">InferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.ensure_float"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.preprocess"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.preprocess()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel"><code class="docutils literal notranslate"><span class="pre">InferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.export_model"><code class="docutils literal notranslate"><span class="pre">InferenceModel.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.predict"><code class="docutils literal notranslate"><span class="pre">InferenceModel.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.predict_on_batch"><code class="docutils literal notranslate"><span class="pre">InferenceModel.predict_on_batch()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.model_name"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.ensure_float"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel.inference_layer"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel.inference_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.model_name"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor"><code class="docutils literal notranslate"><span class="pre">Predictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.export_model"><code class="docutils literal notranslate"><span class="pre">Predictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.from_model_paths"><code class="docutils literal notranslate"><span class="pre">Predictor.from_model_paths()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">Predictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">Predictor.make_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.predict"><code class="docutils literal notranslate"><span class="pre">Predictor.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.report_period"><code class="docutils literal notranslate"><span class="pre">Predictor.report_period</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.RateColumn"><code class="docutils literal notranslate"><span class="pre">RateColumn</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.RateColumn.render"><code class="docutils literal notranslate"><span class="pre">RateColumn.render()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel.single_instance_layer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel.single_instance_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.export_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.instance_peaks"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.keras_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.input_scale"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.output_stride"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.refinement"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.return_confmaps"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.return_class_vectors"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.return_class_vectors</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.offsets_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.class_vectors_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.class_vectors_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.optimal_grouping"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.optimal_grouping</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.call"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.instance_peaks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.export_model()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.centroid_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.centroid_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.centroid_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.centroid_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.tracks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.make_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.centroid_config"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.centroid_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.centroid_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.centroid_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.max_instances"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.make_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor"><code class="docutils literal notranslate"><span class="pre">VisualPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.make_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.predict"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.safely_generate"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.safely_generate()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.export_cli"><code class="docutils literal notranslate"><span class="pre">export_cli()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.export_model"><code class="docutils literal notranslate"><span class="pre">export_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.find_head"><code class="docutils literal notranslate"><span class="pre">find_head()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.get_keras_model_path"><code class="docutils literal notranslate"><span class="pre">get_keras_model_path()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.get_model_output_stride"><code class="docutils literal notranslate"><span class="pre">get_model_output_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.load_model"><code class="docutils literal notranslate"><span class="pre">load_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.make_model_movenet"><code class="docutils literal notranslate"><span class="pre">make_model_movenet()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +618,7 @@ <h1>sleap.nn.inference</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.inference">
-<span id="sleap-nn-inference"></span><h1>sleap.nn.inference<a class="headerlink" href="#module-sleap.nn.inference" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-inference"></span><h1>sleap.nn.inference<a class="headerlink" href="#module-sleap.nn.inference" title="Permalink to this heading">#</a></h1>
 <p>Inference pipelines and utilities.</p>
 <p>This module contains the classes and high level APIs for predicting instances on new
 data using trained models.</p>
@@ -338,7 +638,7 @@ <h1>sleap.nn.inference</h1>
 function which provides a simplified interface for creating <a href="#id3"><span class="problematic" id="id4">`</span></a>Predictor`s.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceLayer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2674-L2940"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2695-L2961"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that predicts instances from images using a trained model.</p>
 <p>This layer encapsulates all of the inference operations required for generating
 predictions from a centered instance confidence map model. This includes
@@ -461,17 +761,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceLayer.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2875-L2940"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2896-L2961"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – This may be either a single batch of images as a 4-D tensor of shape</p></li>
 <li><p><strong>`</strong> (<em>batch_size</em><em>, </em><em>height</em><em>, </em><em>width</em><em>, </em><em>channels</em>) – </p></li>
 <li><p><strong>key.</strong> (<em>image batch in the &quot;images&quot;</em>) – </p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>: Instance skeleton
 points.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peak_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes)</span></code>: Confidence
@@ -487,7 +787,7 @@ <h1>sleap.nn.inference</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;edge_inds&quot;</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;edge_peak_inds&quot;</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;line_scores&quot;</span></code> keys.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -495,13 +795,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceLayer.find_peaks">
-<span class="sig-name descname"><span class="pre">find_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2829-L2873"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.find_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2850-L2894"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.find_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Run peak finding on predicted confidence maps.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceLayer.forward_pass">
-<span class="sig-name descname"><span class="pre">forward_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2801-L2827"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.forward_pass" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">forward_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2822-L2848"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceLayer.forward_pass" title="Permalink to this definition">#</a></dt>
 <dd><p>Run preprocessing and model inference on a batch.</p>
 </dd></dl>
 
@@ -509,7 +809,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2943-L2989"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2964-L3010"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Bottom-up instance prediction model.</p>
 <p>This model encapsulates the bottom-up approach where points are first detected by
 local peak detection and then grouped into instances by connectivity scoring using
@@ -523,15 +823,15 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2963-L2989"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2984-L3010"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>: Instance skeleton</dt><dd><p>points.</p>
 </dd>
@@ -546,7 +846,7 @@ <h1>sleap.nn.inference</h1>
 predicted PAFs will be returned in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;part_affinity_fields&quot;</span></code> key.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -556,7 +856,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceLayer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3278-L3516"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3309-L3547"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that predicts instances from images using a trained model.</p>
 <p>This layer encapsulates all of the inference operations required for generating
 predictions from a centered instance confidence map model. This includes
@@ -662,17 +962,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceLayer.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3452-L3516"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3483-L3547"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – This may be either a single batch of images as a 4-D tensor of shape</p></li>
 <li><p><strong>`</strong> (<em>batch_size</em><em>, </em><em>height</em><em>, </em><em>width</em><em>, </em><em>channels</em>) – </p></li>
 <li><p><strong>key.</strong> (<em>image batch in the &quot;images&quot;</em>) – </p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>: Instance skeleton
 points.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peak_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes)</span></code>: Confidence
@@ -685,7 +985,7 @@ <h1>sleap.nn.inference</h1>
 will be returned in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;class_maps&quot;</span></code> key.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -693,13 +993,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks">
-<span class="sig-name descname"><span class="pre">find_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3420-L3450"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3451-L3481"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Run peak finding on predicted confidence maps.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass">
-<span class="sig-name descname"><span class="pre">forward_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3394-L3418"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">forward_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3425-L3449"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass" title="Permalink to this definition">#</a></dt>
 <dd><p>Run preprocessing and model inference on a batch.</p>
 </dd></dl>
 
@@ -707,7 +1007,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3519-L3561"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3550-L3592"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Bottom-up multi-class instance prediction model.</p>
 <p>This model encapsulates the bottom-up multi-class approach where points are first
 detected by local peak finding and then grouped into instances by their identity
@@ -721,15 +1021,15 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3535-L3561"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3566-L3592"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>: Instance skeleton</dt><dd><p>points.</p>
 </dd>
@@ -744,7 +1044,7 @@ <h1>sleap.nn.inference</h1>
 will be returned in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;class_maps&quot;</span></code> key.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -754,7 +1054,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel" title="sleap.nn.inference.BottomUpMultiClassInferenceModel"><span class="pre">sleap.nn.inference.BottomUpMultiClassInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3565-L3777"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel" title="sleap.nn.inference.BottomUpMultiClassInferenceModel"><span class="pre">BottomUpMultiClassInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3596-L3818"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Bottom-up multi-instance predictor.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained bottom-up multi-instance SLEAP model.</p>
@@ -766,7 +1066,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.TrainingJobConfig</span></code> containing the metadata for the
 trained model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a></p>
 </dd>
 </dl>
@@ -779,7 +1079,7 @@ <h1>sleap.nn.inference</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be used if available from the data
 source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a></p>
 </dd>
 </dl>
@@ -792,7 +1092,7 @@ <h1>sleap.nn.inference</h1>
 wraps a trained <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing, peak finding
 and classification.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel" title="sleap.nn.inference.BottomUpMultiClassInferenceModel">sleap.nn.inference.BottomUpMultiClassInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -804,7 +1104,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -816,7 +1116,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -827,7 +1127,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassPredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a local peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -841,7 +1141,7 @@ <h1>sleap.nn.inference</h1>
 gradient offset. This has no effect if the model has an offset regression
 head.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -853,7 +1153,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Size of patches to crop around each rough peak for integral
 refinement as an integer scalar.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -865,7 +1165,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>If provided, instances will be created using these track instances. If
 not, instances will be assigned tracks from the provider if possible.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[<a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track">sleap.instance.Track</a>]]</p>
 </dd>
 </dl>
@@ -873,10 +1173,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.BottomUpMultiClassPredictor" title="sleap.nn.inference.BottomUpMultiClassPredictor"><span class="pre">sleap.nn.inference.BottomUpMultiClassPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3623-L3674"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.BottomUpMultiClassPredictor" title="sleap.nn.inference.BottomUpMultiClassPredictor"><span class="pre">BottomUpMultiClassPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3654-L3705"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create predictor from a saved model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to a model folder or training job JSON file inside a model
 folder. This folder should contain <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_config.json</span></code> and
@@ -896,7 +1196,7 @@ <h1>sleap.nn.inference</h1>
 resized to (None, None, None, num_color_channels).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.inference.BottomUpPredictor" title="sleap.nn.inference.BottomUpPredictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">BottomUpPredictor</span></code></a> with the loaded model.</p>
 </dd>
 </dl>
@@ -912,7 +1212,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bottomup_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bottomup_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.BottomUpInferenceModel" title="sleap.nn.inference.BottomUpInferenceModel"><span class="pre">sleap.nn.inference.BottomUpInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paf_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2993-L3275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">BottomUpPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bottomup_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bottomup_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.BottomUpInferenceModel" title="sleap.nn.inference.BottomUpInferenceModel"><span class="pre">BottomUpInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paf_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3014-L3306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Bottom-up multi-instance predictor.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained bottom-up multi-instance SLEAP model.</p>
@@ -924,7 +1224,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.TrainingJobConfig</span></code> containing the metadata
 for the trained bottomup model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a></p>
 </dd>
 </dl>
@@ -937,7 +1237,7 @@ <h1>sleap.nn.inference</h1>
 bottomup model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be used if available
 from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a></p>
 </dd>
 </dl>
@@ -950,7 +1250,7 @@ <h1>sleap.nn.inference</h1>
 trained <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing, centroid detection,
 cropping and peak finding.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.BottomUpInferenceModel" title="sleap.nn.inference.BottomUpInferenceModel">sleap.nn.inference.BottomUpInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -962,7 +1262,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -975,7 +1275,7 @@ <h1>sleap.nn.inference</h1>
 detections over time. Predicted instances will not be assigned to tracks if
 if this is <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.tracking.html#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker">sleap.nn.tracking.Tracker</a>]</p>
 </dd>
 </dl>
@@ -987,7 +1287,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -998,7 +1298,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a local peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1012,7 +1312,7 @@ <h1>sleap.nn.inference</h1>
 gradient offset. This has no effect if the model has an offset regression
 head.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -1024,7 +1324,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Size of patches to crop around each rough peak for integral
 refinement as an integer scalar.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1037,7 +1337,7 @@ <h1>sleap.nn.inference</h1>
 as a fraction of the image size. Candidate connections longer than this
 length will be penalized during matching.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1050,7 +1350,7 @@ <h1>sleap.nn.inference</h1>
 a scalar float. Set to values greater than 1.0 to enforce the distance
 penalty more strictly.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1061,7 +1361,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">paf_line_points</span></span><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor.paf_line_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of points to sample along the line integral.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1074,7 +1374,7 @@ <h1>sleap.nn.inference</h1>
 between candidate point pairs. Useful for rejecting spurious detections when
 there are no better ones.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1088,7 +1388,7 @@ <h1>sleap.nn.inference</h1>
 This is useful for preventing extraneous instances from being created when
 tracking is not being applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -1096,10 +1396,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.BottomUpPredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paf_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.BottomUpPredictor" title="sleap.nn.inference.BottomUpPredictor"><span class="pre">sleap.nn.inference.BottomUpPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3088-L3165"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paf_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.BottomUpPredictor" title="sleap.nn.inference.BottomUpPredictor"><span class="pre">BottomUpPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3109-L3186"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.BottomUpPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create predictor from a saved model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to a bottom-up model folder or training job JSON file
 inside a model folder. This folder should contain <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_config.json</span></code>
@@ -1133,7 +1433,7 @@ <h1>sleap.nn.inference</h1>
 created when tracking is not being applied.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.inference.BottomUpPredictor" title="sleap.nn.inference.BottomUpPredictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">BottomUpPredictor</span></code></a> with the loaded model.</p>
 </dd>
 </dl>
@@ -1149,7 +1449,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.CentroidCrop">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidCrop</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1614-L1930"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCrop" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidCrop</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1625-L1941"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCrop" title="Permalink to this definition">#</a></dt>
 <dd><p>Inference layer for applying centroid crop-based models.</p>
 <p>This layer encapsulates all of the inference operations requires for generating
 predictions from a centroid confidence map model. This includes preprocessing,
@@ -1265,12 +1565,12 @@ <h1>sleap.nn.inference</h1>
 <p>This layer can be chained with a <a class="reference internal" href="#sleap.nn.inference.FindInstancePeaks" title="sleap.nn.inference.FindInstancePeaks"><code class="xref py py-obj docutils literal notranslate"><span class="pre">FindInstancePeaks</span></code></a> layer to create a top-down
 inference function from full images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>inputs</strong> – Full frame images as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> or a dictionary with key:
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code>: Full frame images in the same format as above.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;</span></code>: The predicted centroids of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">?,</span> <span class="pre">2)</span></code>.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroid_vals&quot;:</span> <span class="pre">The</span> <span class="pre">centroid</span> <span class="pre">confidence</span> <span class="pre">values</span> <span class="pre">of</span> <span class="pre">shape</span> <span class="pre">`(samples,</span> <span class="pre">?)</span></code>.</p>
 <p>If the <a class="reference internal" href="#sleap.nn.inference.CentroidCrop.return_confmaps" title="sleap.nn.inference.CentroidCrop.return_confmaps"><code class="xref py py-obj docutils literal notranslate"><span class="pre">return_confmaps</span></code></a> attribute is set to <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>, the output will also
@@ -1286,7 +1586,7 @@ <h1>sleap.nn.inference</h1>
 coordinates.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dictionary of outputs grouped by sample with keys</p>
 </dd>
 </dl>
@@ -1296,7 +1596,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.CentroidCropGroundTruth">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidCropGroundTruth</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L722-L798"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCropGroundTruth" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidCropGroundTruth</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L723-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCropGroundTruth" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that simulates a centroid cropping model using ground truth.</p>
 <p>This layer is useful for testing and evaluating centered instance models.</p>
 <dl class="py attribute">
@@ -1307,10 +1607,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.CentroidCropGroundTruth.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L735-L798"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCropGroundTruth.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L736-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidCropGroundTruth.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the ground truth instance crops.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example_gt</strong> – <p>Dictionary generated from a labels pipeline with the keys:
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_centroids,</span> <span class="pre">2)</span></code>: The input centroids.</p>
@@ -1320,7 +1620,7 @@ <h1>sleap.nn.inference</h1>
 <p>These can be generated by the <code class="xref py py-obj docutils literal notranslate"><span class="pre">InstanceCentroidFinder</span></code> transformer.</p>
 </p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;crops&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_centroids,</span> <span class="pre">crop_size,</span> <span class="pre">crop_size,</span> <span class="pre">channels)</span></code>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;crop_offsets&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_centroids,</span> <span class="pre">crop_size,</span> <span class="pre">crop_size,</span> <span class="pre">channels)</span></code></p>
 <blockquote>
@@ -1333,7 +1633,7 @@ <h1>sleap.nn.inference</h1>
 filled with ones.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Dictionary containing the output of the instance cropping layer with keys</p>
 </dd>
 </dl>
@@ -1343,7 +1643,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.CentroidInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2158-L2198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">CentroidInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2169-L2209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Centroid only instance prediction model.</p>
 <p>This model encapsulates the first step in a top-down approach where instances are detected by
 local peak detection of an anchor point and then cropped.</p>
@@ -1357,17 +1657,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.CentroidInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2174-L2198"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2185-L2209"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.CentroidInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key. If using a ground
 truth model for either centroid cropping or instance peaks, the full
 example from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code> is required for providing the metadata.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">2)</span></code>: Instance centroids.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroid_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances)</span></code>: Instance centroid confidence</p>
 <blockquote>
@@ -1375,7 +1675,7 @@ <h1>sleap.nn.inference</h1>
 </div></blockquote>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -1385,7 +1685,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.FindInstancePeaks">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">FindInstancePeaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1933-L2155"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaks" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">FindInstancePeaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1944-L2166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that predicts instance peaks from images using a trained model.</p>
 <p>This layer encapsulates all of the inference operations required for generating
 predictions from a centered instance confidence map model. This includes
@@ -1466,12 +1766,12 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.FindInstancePeaks.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2016-L2155"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaks.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2027-L2166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaks.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict confidence maps and infer peak coordinates.</p>
 <p>This layer can be chained with a <a class="reference internal" href="#sleap.nn.inference.CentroidCrop" title="sleap.nn.inference.CentroidCrop"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CentroidCrop</span></code></a> layer to create a top-down
 inference function from full images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>inputs</strong> – <p>Instance-centered images as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">?,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> where images are grouped by
@@ -1492,7 +1792,7 @@ <h1>sleap.nn.inference</h1>
 </dl>
 </p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;</span></code>: The predicted peaks for each instance in the batch as a</dt><dd><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">?,</span> <span class="pre">nodes,</span> <span class="pre">2)</span></code>.</p>
 </dd>
@@ -1509,7 +1809,7 @@ <h1>sleap.nn.inference</h1>
 confidence maps predicted by the model.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dictionary of outputs with keys</p>
 </dd>
 </dl>
@@ -1519,15 +1819,15 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.FindInstancePeaksGroundTruth">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">FindInstancePeaksGroundTruth</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L801-L883"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaksGroundTruth" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">FindInstancePeaksGroundTruth</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L802-L884"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaksGroundTruth" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that simulates a centered instance peaks model.</p>
 <p>This layer is useful for testing and evaluating centroid models.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.FindInstancePeaksGroundTruth.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_output</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L809-L883"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaksGroundTruth.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_output</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L810-L884"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.FindInstancePeaksGroundTruth.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the ground truth instance peaks given a set of crops.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>example_gt</strong> – <p>Dictionary generated from a labels pipeline with the key:
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instances&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances_gt,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code></p>
@@ -1545,7 +1845,7 @@ <h1>sleap.nn.inference</h1>
 </p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A dictionary with the instance peaks for each frame. The peaks are just the
 ground truth instances matched to the crop output centroids via greedy
 matching of the closest node point to each centroid.</p>
@@ -1572,7 +1872,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceLayer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">InferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L886-L966"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">InferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L887-L967"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer" title="Permalink to this definition">#</a></dt>
 <dd><p>Base layer for wrapping a Keras model into a layer with preprocessing.</p>
 <p>This layer is useful for wrapping input preprocessing operations that would
 otherwise be handled by a separate pipeline.</p>
@@ -1617,13 +1917,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceLayer.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L957-L966"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L958-L967"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Call the model with preprocessed data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data</strong> – Inputs to the model.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Output of the model after being called with preprocessing.</p>
 </dd>
 </dl>
@@ -1631,13 +1931,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceLayer.preprocess">
-<span class="sig-name descname"><span class="pre">preprocess</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">imgs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L929-L955"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer.preprocess" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">preprocess</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">imgs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L930-L956"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceLayer.preprocess" title="Permalink to this definition">#</a></dt>
 <dd><p>Apply all preprocessing operations configured for this layer.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>imgs</strong> – A batch of images as a tensor.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The input tensor after applying preprocessing operations. The tensor will
 always be a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>, which will be adjusted to the range <code class="xref py py-obj docutils literal notranslate"><span class="pre">[0,</span> <span class="pre">1]</span></code> if it
 was previously an integer.</p>
@@ -1649,17 +1949,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">InferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L969-L1156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">InferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L970-L1158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>SLEAP inference model base class.</p>
 <p>This class wraps the <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> class to provide SLEAP-specific inference
 utilities such as handling different input data types, preprocessing and variable
 output shapes.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceModel.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1080-L1156"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1081-L1158"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Save the frozen graph of a model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -1685,10 +1985,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceModel.predict">
-<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L977-L1033"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.predict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">DatasetV2</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L978-L1034"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.predict" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances in the data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – <p>Input data in any form. Possible types:
 - <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.ndarray</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code>: Images of shape</p>
@@ -1712,7 +2012,7 @@ <h1>sleap.nn.inference</h1>
 pipeline as input since those are expected to generate batches.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The model outputs as a dictionary of (potentially ragged) tensors or numpy
 arrays if <code class="xref py py-obj docutils literal notranslate"><span class="pre">numpy</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>.</p>
 <p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">numpy</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, values of the dictionary may be <a href="#id7"><span class="problematic" id="id8">`</span></a>tf.RaggedTensor`s
@@ -1729,10 +2029,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.InferenceModel.predict_on_batch">
-<span class="sig-name descname"><span class="pre">predict_on_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1035-L1078"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.predict_on_batch" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">predict_on_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">numpy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1036-L1079"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.InferenceModel.predict_on_batch" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict a single batch of samples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – <p>Input data in any form. Possible types:
 - <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.ndarray</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code>: Images of shape</p>
@@ -1747,7 +2047,7 @@ <h1>sleap.nn.inference</h1>
 <a href="#id9"><span class="problematic" id="id10">`</span></a>np.ndarray`s or Python primitives if scalars.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The model outputs as a dictionary of (potentially ragged) tensors or numpy
 arrays if <code class="xref py py-obj docutils literal notranslate"><span class="pre">numpy</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>.</p>
 <p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">numpy</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">False</span></code>, values of the dictionary may be <a href="#id11"><span class="problematic" id="id12">`</span></a>tf.RaggedTensor`s
@@ -1766,7 +2066,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetInferenceLayer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4533-L4578"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceLayer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4593-L4638"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceLayer" title="Permalink to this definition">#</a></dt>
 <dd><p>Inference layer for applying single instance models.</p>
 <p>This layer encapsulates all of the inference operations requires for generating
 predictions from a single instance confidence map model. This includes
@@ -1816,13 +2116,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetInferenceLayer.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ex</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4568-L4578"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceLayer.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ex</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4628-L4638"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceLayer.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Call the model with preprocessed data.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data</strong> – Inputs to the model.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Output of the model after being called with preprocessing.</p>
 </dd>
 </dl>
@@ -1832,7 +2132,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4581-L4605"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4641-L4665"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>MoveNet prediction model.</p>
 <p>This model encapsulates the basic MoveNet approach. The images are passed to a model
 which is trained to detect all body parts (17 joints in total).</p>
@@ -1849,7 +2149,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4604-L4605"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4664-L4665"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Calls the model on new inputs and returns the outputs as tensors.</p>
 <p>In this case <a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel.call" title="sleap.nn.inference.MoveNetInferenceModel.call"><code class="xref py py-obj docutils literal notranslate"><span class="pre">call()</span></code></a> just reapplies
 all ops in the graph to the new inputs
@@ -1859,7 +2159,7 @@ <h1>sleap.nn.inference</h1>
 To call a model on an input, always use the <code class="xref py py-obj docutils literal notranslate"><span class="pre">__call__()</span></code> method,
 i.e. <code class="xref py py-obj docutils literal notranslate"><span class="pre">model(inputs)</span></code>, which relies on the underlying <a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel.call" title="sleap.nn.inference.MoveNetInferenceModel.call"><code class="xref py py-obj docutils literal notranslate"><span class="pre">call()</span></code></a> method.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>inputs</strong> – Input tensor, or dict/list/tuple of input tensors.</p></li>
 <li><p><strong>training</strong> – Boolean or boolean scalar tensor, indicating whether to run
@@ -1872,7 +2172,7 @@ <h1>sleap.nn.inference</h1>
 </p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor if there is a single output, or
 a list of tensors if there are more than one outputs.</p>
 </dd>
@@ -1883,7 +2183,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel" title="sleap.nn.inference.MoveNetInferenceModel"><span class="pre">sleap.nn.inference.MoveNetInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'lightning'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4609-L4741"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">MoveNetPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel" title="sleap.nn.inference.MoveNetInferenceModel"><span class="pre">MoveNetInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'lightning'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4669-L4810"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>MoveNet predictor.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained MoveNet model.
@@ -1895,7 +2195,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel" title="sleap.nn.inference.MoveNetInferenceModel"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.inference.MoveNetInferenceModel</span></code></a> that wraps
 a trained <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing and peak finding.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.MoveNetInferenceModel" title="sleap.nn.inference.MoveNetInferenceModel">sleap.nn.inference.MoveNetInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -1907,7 +2207,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -1918,7 +2218,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.MoveNetPredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a global peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -1930,7 +2230,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -1942,7 +2242,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Variable indicating which model of MoveNet to use, either “lightning”
 or “thunder.”</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -1950,10 +2250,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.MoveNetPredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.MoveNetPredictor" title="sleap.nn.inference.MoveNetPredictor"><span class="pre">sleap.nn.inference.MoveNetPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4661-L4683"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.MoveNetPredictor" title="sleap.nn.inference.MoveNetPredictor"><span class="pre">MoveNetPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4721-L4743"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.MoveNetPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the predictor from a saved model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_name</strong> – Variable indicating which model of MoveNet to use, either “lightning”
 or “thunder.”</p></li>
@@ -1961,7 +2261,7 @@ <h1>sleap.nn.inference</h1>
 valid.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of`MoveNetPredictor` with the models loaded.</p>
 </dd>
 </dl>
@@ -1977,16 +2277,16 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.Predictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">Predictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L158-L588"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">Predictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L159-L589"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Base interface class for predictors.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.Predictor.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L532-L588"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L533-L589"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Export a trained SLEAP model as a frozen graph. Initializes model,
 creates a dummy tracing batch and passes it through the model. The
 frozen graph is saved along with training meta info.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -2010,10 +2310,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.Predictor.from_model_paths">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_model_paths</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><span class="pre">sleap.nn.inference.Predictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L174-L310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.from_model_paths" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_model_paths</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><span class="pre">Predictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L175-L311"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.from_model_paths" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the appropriate <a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Predictor</span></code></a> subclass from a list of model paths.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_paths</strong> – A single or list of trained model paths. Special cases of
 non-SLEAP models include “movenet-thunder” and “movenet-lightning”.</p></li>
@@ -2035,7 +2335,7 @@ <h1>sleap.nn.inference</h1>
 created when tracking is not being applied.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A subclass of <a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Predictor</span></code></a>.</p>
 </dd>
 </dl>
@@ -2054,14 +2354,14 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.Predictor.make_pipeline">
-<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L328-L370"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.make_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L329-L371"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.make_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a data loading pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the pipeline will be created with an instance
 of a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The created <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> with batching and prefetching.</p>
 </dd>
 </dl>
@@ -2074,10 +2374,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.Predictor.predict">
-<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">sleap.io.video.Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">make_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L495-L530"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.predict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.video.html#sleap.io.video.Video" title="sleap.io.video.Video"><span class="pre">Video</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">make_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L496-L531"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.Predictor.predict" title="Permalink to this definition">#</a></dt>
 <dd><p>Run inference on a data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> to
 run inference over.</p></li>
@@ -2086,7 +2386,7 @@ <h1>sleap.nn.inference</h1>
 dictionaries containing the raw arrays returned by the inference model.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> with <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.PredictedInstance`s</span> <span class="pre">if</span> <span class="pre">`make_labels</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>,
 otherwise a list of dictionaries containing batches of numpy arrays with the
 raw results.</p>
@@ -2104,11 +2404,11 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.RateColumn">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">RateColumn</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">table_column</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">rich.table.Column</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L146-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.RateColumn" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">RateColumn</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">table_column</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Column</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L147-L155"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.RateColumn" title="Permalink to this definition">#</a></dt>
 <dd><p>Renders the progress rate.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.RateColumn.render">
-<span class="sig-name descname"><span class="pre">render</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Task</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">rich.text.Text</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L149-L154"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.RateColumn.render" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">render</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Task</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Text</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L150-L155"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.RateColumn.render" title="Permalink to this definition">#</a></dt>
 <dd><p>Show progress rate.</p>
 </dd></dl>
 
@@ -2116,7 +2416,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstanceInferenceLayer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1214-L1365"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceLayer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceInferenceLayer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1216-L1367"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceLayer" title="Permalink to this definition">#</a></dt>
 <dd><p>Inference layer for applying single instance models.</p>
 <p>This layer encapsulates all of the inference operations requires for generating
 predictions from a single instance confidence map model. This includes
@@ -2205,15 +2505,15 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstanceInferenceLayer.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1304-L1365"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceLayer.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1306-L1367"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceLayer.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instance confidence maps and find peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>inputs</strong> – Full frame images as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> or a dictionary with key:
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;image&quot;</span></code>: Full frame images in the same format as above.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;</span></code>: The predicted peaks of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">1,</span> <span class="pre">nodes,</span> <span class="pre">2)</span></code>.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peak_vals&quot;:</span> <span class="pre">The</span> <span class="pre">peak</span> <span class="pre">confidence</span> <span class="pre">values</span> <span class="pre">of</span> <span class="pre">shape</span>
 <span class="pre">`(samples,</span> <span class="pre">1,</span> <span class="pre">nodes)</span></code>.</p>
@@ -2223,7 +2523,7 @@ <h1>sleap.nn.inference</h1>
 predicted by the model.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dictionary of outputs grouped by sample with keys</p>
 </dd>
 </dl>
@@ -2233,7 +2533,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstanceInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1368-L1400"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1370-L1402"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Single instance prediction model.</p>
 <p>This model encapsulates the basic single instance approach where it is assumed that
 there is only one instance in the frame. The images are passed to a peak detector
@@ -2248,15 +2548,15 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstanceInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1385-L1400"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1387-L1402"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstanceInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">1,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code>: Instance skeleton points.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peak_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">1,</span> <span class="pre">n_nodes)</span></code>: Confidence</p>
 <blockquote>
@@ -2264,7 +2564,7 @@ <h1>sleap.nn.inference</h1>
 </div></blockquote>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -2274,7 +2574,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstancePredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstancePredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="sleap.nn.inference.SingleInstanceInferenceModel"><span class="pre">sleap.nn.inference.SingleInstanceInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1404-L1611"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">SingleInstancePredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="sleap.nn.inference.SingleInstanceInferenceModel"><span class="pre">SingleInstanceInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1406-L1622"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Single instance predictor.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained single instance SLEAP model.</p>
@@ -2286,7 +2586,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.config.TrainingJobConfig</span></code> containing the metadata
 for the trained model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a></p>
 </dd>
 </dl>
@@ -2297,7 +2597,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">confmap_model</span></span><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.confmap_model" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.model.Model</span></code></a> instance created from the trained model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a></p>
 </dd>
 </dl>
@@ -2309,7 +2609,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <a class="reference internal" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="sleap.nn.inference.SingleInstanceInferenceModel"><code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.inference.SingleInstanceInferenceModel</span></code></a> that wraps
 a trained <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing and peak finding.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.SingleInstanceInferenceModel" title="sleap.nn.inference.SingleInstanceInferenceModel">sleap.nn.inference.SingleInstanceInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -2321,7 +2621,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -2332,7 +2632,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a global peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -2346,7 +2646,7 @@ <h1>sleap.nn.inference</h1>
 gradient offset. This has no effect if the model has an offset regression
 head.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -2358,7 +2658,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Size of patches to crop around each rough peak for integral
 refinement as an integer scalar.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -2370,7 +2670,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -2378,12 +2678,12 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstancePredictor.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1591-L1611"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1602-L1622"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Export a trained SLEAP model as a frozen graph. Initializes model,
 creates a dummy tracing batch and passes it through the model. The
 frozen graph is saved along with training meta info.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -2407,10 +2707,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.SingleInstancePredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.SingleInstancePredictor" title="sleap.nn.inference.SingleInstancePredictor"><span class="pre">sleap.nn.inference.SingleInstancePredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1464-L1517"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.SingleInstancePredictor" title="sleap.nn.inference.SingleInstancePredictor"><span class="pre">SingleInstancePredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1466-L1519"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.SingleInstancePredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the predictor from a saved model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to a model folder or training job JSON file inside a model
 folder. This folder should contain <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_config.json</span></code> and
@@ -2430,7 +2730,7 @@ <h1>sleap.nn.inference</h1>
 resized to (None, None, None, num_color_channels).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of`SingleInstancePredictor` with the models loaded.</p>
 </dd>
 </dl>
@@ -2446,7 +2746,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2201-L2266"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2212-L2277"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Top-down instance prediction model.</p>
 <p>This model encapsulates the top-down approach where instances are first detected by
 local peak detection of an anchor point and then cropped. These instance-centered
@@ -2471,17 +2771,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2228-L2266"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2239-L2277"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key. If using a ground
 truth model for either centroid cropping or instance peaks, the full
 example from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code> is required for providing the metadata.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">2)</span></code>: Instance centroids.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroid_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances)</span></code>: Instance centroid confidence</p>
 <blockquote>
@@ -2495,7 +2795,7 @@ <h1>sleap.nn.inference</h1>
 </dl>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -2505,7 +2805,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassFindPeaks">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassFindPeaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3780-L4046"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassFindPeaks" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassFindPeaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3821-L4087"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassFindPeaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras layer that predicts and classifies peaks from images using a trained model.</p>
 <p>This layer encapsulates all of the inference operations required for generating
 predictions from a centered instance confidence map and multi-class model. This
@@ -2613,12 +2913,12 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassFindPeaks.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L3888-L4046"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">inputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L3929-L4087"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict confidence maps and infer peak coordinates.</p>
 <p>This layer can be chained with a <a class="reference internal" href="#sleap.nn.inference.CentroidCrop" title="sleap.nn.inference.CentroidCrop"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CentroidCrop</span></code></a> layer to create a top-down
 inference function from full images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>inputs</strong> – <p>Instance-centered images as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">?,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code> where images are grouped by
@@ -2639,7 +2939,7 @@ <h1>sleap.nn.inference</h1>
 </dl>
 </p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;instance_peaks&quot;</span></code>: The predicted peaks for each instance in the batch as a</dt><dd><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">n_classes,</span> <span class="pre">nodes,</span> <span class="pre">2)</span></code>. Instances will
 be ordered by class and will be filled with <code class="xref py py-obj docutils literal notranslate"><span class="pre">NaN</span></code> where not found.</p>
@@ -2664,7 +2964,7 @@ <h1>sleap.nn.inference</h1>
 this will change the output dict keys and shapes.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A dictionary of outputs with keys</p>
 </dd>
 </dl>
@@ -2674,7 +2974,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassInferenceModel">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4049-L4119"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassInferenceModel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4090-L4160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel" title="Permalink to this definition">#</a></dt>
 <dd><p>Top-down instance prediction model.</p>
 <p>This model encapsulates the top-down approach where instances are first detected by
 local peak detection of an anchor point and then cropped. These instance-centered
@@ -2699,17 +2999,17 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassInferenceModel.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4076-L4104"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">example</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4117-L4145"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Predict instances for one batch of images.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>example</strong> – This may be either a single batch of images as a 4-D tensor of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(batch_size,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>, or a dictionary
 containing the image batch in the <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;images&quot;</span></code> key. If using a ground
 truth model for either centroid cropping or instance peaks, the full
 example from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Pipeline</span></code> is required for providing the metadata.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroids&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances,</span> <span class="pre">2)</span></code>: Instance centroids.
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">&quot;centroid_vals&quot;:</span> <span class="pre">(batch_size,</span> <span class="pre">n_instances)</span></code>: Instance centroid confidence</p>
 <blockquote>
@@ -2723,7 +3023,7 @@ <h1>sleap.nn.inference</h1>
 </dl>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The predicted instances as a dictionary of tensors with keys</p>
 </dd>
 </dl>
@@ -2731,10 +3031,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassInferenceModel.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4106-L4119"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4147-L4160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Save the frozen graph of a model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -2762,7 +3062,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassInferenceModel" title="sleap.nn.inference.TopDownMultiClassInferenceModel"><span class="pre">sleap.nn.inference.TopDownMultiClassInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4123-L4493"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassInferenceModel" title="sleap.nn.inference.TopDownMultiClassInferenceModel"><span class="pre">TopDownMultiClassInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4164-L4553"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Top-down multi-instance predictor with classification.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained top-down multi-instance classification SLEAP model.</p>
@@ -2775,7 +3075,7 @@ <h1>sleap.nn.inference</h1>
 for the trained centroid model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be
 used if available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a>]</p>
 </dd>
 </dl>
@@ -2788,7 +3088,7 @@ <h1>sleap.nn.inference</h1>
 centroid model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be used if available
 from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a>]</p>
 </dd>
 </dl>
@@ -2801,7 +3101,7 @@ <h1>sleap.nn.inference</h1>
 for the trained centered instance model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth instances
 will be used if available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a>]</p>
 </dd>
 </dl>
@@ -2814,7 +3114,7 @@ <h1>sleap.nn.inference</h1>
 centered-instance model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth instances will be used if
 available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a>]</p>
 </dd>
 </dl>
@@ -2827,7 +3127,7 @@ <h1>sleap.nn.inference</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing, centroid detection, cropping,
 peak finding and classification.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassInferenceModel" title="sleap.nn.inference.TopDownMultiClassInferenceModel">sleap.nn.inference.TopDownMultiClassInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -2839,7 +3139,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -2852,7 +3152,7 @@ <h1>sleap.nn.inference</h1>
 detections over time. Predicted instances will not be assigned to tracks if
 if this is <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.tracking.html#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker">sleap.nn.tracking.Tracker</a>]</p>
 </dd>
 </dl>
@@ -2864,7 +3164,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -2875,7 +3175,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a local peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -2889,7 +3189,7 @@ <h1>sleap.nn.inference</h1>
 gradient offset. This has no effect if the model has an offset regression
 head.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -2901,7 +3201,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Size of patches to crop around each rough peak for integral
 refinement as an integer scalar.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -2913,7 +3213,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>If provided, instances will be created using these track instances. If
 not, instances will be assigned tracks from the provider if possible.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[List[<a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track">sleap.instance.Track</a>]]</p>
 </dd>
 </dl>
@@ -2921,12 +3221,12 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassPredictor.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4466-L4493"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4526-L4553"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Export a trained SLEAP model as a frozen graph. Initializes model,
 creates a dummy tracing batch and passes it through the model. The
 frozen graph is saved along with training meta info.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -2950,10 +3250,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassPredictor" title="sleap.nn.inference.TopDownMultiClassPredictor"><span class="pre">sleap.nn.inference.TopDownMultiClassPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4222-L4310"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassPredictor" title="sleap.nn.inference.TopDownMultiClassPredictor"><span class="pre">TopDownMultiClassPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4263-L4351"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create predictor from saved models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>centroid_model_path</strong> – Path to a centroid model folder or training job JSON
 file inside a model folder. This folder should contain
@@ -2976,7 +3276,7 @@ <h1>sleap.nn.inference</h1>
 resized to (None, None, None, num_color_channels).</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An instance of <a class="reference internal" href="#sleap.nn.inference.TopDownMultiClassPredictor" title="sleap.nn.inference.TopDownMultiClassPredictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor</span></code></a> with the loaded models.</p>
 <p>One of the two models can be left as <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> to perform inference with ground
 truth data. This will only work with <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> as the provider.</p>
@@ -2993,14 +3293,14 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline">
-<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4330-L4364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4371-L4405"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a data loading pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the pipeline will be created with an instance
 of a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The created <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> with batching and prefetching.</p>
 </dd>
 </dl>
@@ -3015,7 +3315,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.TopDownInferenceModel" title="sleap.nn.inference.TopDownInferenceModel"><span class="pre">sleap.nn.inference.TopDownInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2270-L2671"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">TopDownPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">centroid_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inference_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.inference.TopDownInferenceModel" title="sleap.nn.inference.TopDownInferenceModel"><span class="pre">TopDownInferenceModel</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2281-L2692"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Top-down multi-instance predictor.</p>
 <p>This high-level class handles initialization, preprocessing and tracking using a
 trained top-down multi-instance SLEAP model.</p>
@@ -3028,7 +3328,7 @@ <h1>sleap.nn.inference</h1>
 for the trained centroid model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be
 used if available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a>]</p>
 </dd>
 </dl>
@@ -3041,7 +3341,7 @@ <h1>sleap.nn.inference</h1>
 centroid model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth centroids will be used if available
 from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a>]</p>
 </dd>
 </dl>
@@ -3054,7 +3354,7 @@ <h1>sleap.nn.inference</h1>
 for the trained centered instance model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth instances
 will be used if available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a>]</p>
 </dd>
 </dl>
@@ -3067,7 +3367,7 @@ <h1>sleap.nn.inference</h1>
 centered-instance model. If <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, ground truth instances will be used if
 available from the data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a>]</p>
 </dd>
 </dl>
@@ -3080,7 +3380,7 @@ <h1>sleap.nn.inference</h1>
 trained <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to implement preprocessing, centroid detection,
 cropping and peak finding.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="#sleap.nn.inference.TopDownInferenceModel" title="sleap.nn.inference.TopDownInferenceModel">sleap.nn.inference.TopDownInferenceModel</a>]</p>
 </dd>
 </dl>
@@ -3092,7 +3392,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.nn.data.Pipeline</span></code> that loads the data and batches input data.
 This will be updated dynamically if new data sources are used.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a>]</p>
 </dd>
 </dl>
@@ -3105,7 +3405,7 @@ <h1>sleap.nn.inference</h1>
 detections over time. Predicted instances will not be assigned to tracks if
 if this is <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.tracking.html#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker">sleap.nn.tracking.Tracker</a>]</p>
 </dd>
 </dl>
@@ -3117,7 +3417,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>The default batch size to use when loading data for inference.
 Higher values increase inference speed at the cost of higher memory usage.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -3128,7 +3428,7 @@ <h1>sleap.nn.inference</h1>
 <span class="sig-name descname"><span class="pre">peak_threshold</span></span><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.peak_threshold" title="Permalink to this definition">#</a></dt>
 <dd><p>Minimum confidence map value to consider a local peak as valid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -3142,7 +3442,7 @@ <h1>sleap.nn.inference</h1>
 gradient offset. This has no effect if the model has an offset regression
 head.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -3154,7 +3454,7 @@ <h1>sleap.nn.inference</h1>
 <dd><p>Size of patches to crop around each rough peak for integral
 refinement as an integer scalar.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -3168,7 +3468,7 @@ <h1>sleap.nn.inference</h1>
 This is useful for preventing extraneous instances from being created when
 tracking is not being applied.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[int]</p>
 </dd>
 </dl>
@@ -3176,12 +3476,12 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownPredictor.export_model">
-<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2644-L2671"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2665-L2692"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Export a trained SLEAP model as a frozen graph. Initializes model,
 creates a dummy tracing batch and passes it through the model. The
 frozen graph is saved along with training meta info.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>save_path</strong> – Path to output directory to store the frozen graph</p></li>
 <li><p><strong>signatures</strong> – String defining the input and output types for
@@ -3205,10 +3505,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownPredictor.from_trained_models">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.TopDownPredictor" title="sleap.nn.inference.TopDownPredictor"><span class="pre">sleap.nn.inference.TopDownPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2381-L2477"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_trained_models</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centroid_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">confmap_model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.TopDownPredictor" title="sleap.nn.inference.TopDownPredictor"><span class="pre">TopDownPredictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2392-L2488"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.from_trained_models" title="Permalink to this definition">#</a></dt>
 <dd><p>Create predictor from saved models.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>centroid_model_path</strong> – Path to a centroid model folder or training job JSON
 file inside a model folder. This folder should contain
@@ -3235,7 +3535,7 @@ <h1>sleap.nn.inference</h1>
 created when tracking is not being applied.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An instance of <a class="reference internal" href="#sleap.nn.inference.TopDownPredictor" title="sleap.nn.inference.TopDownPredictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">TopDownPredictor</span></code></a> with the loaded models.</p>
 <p>One of the two models can be left as <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> to perform inference with ground
 truth data. This will only work with <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabelsReader</span></code> as the provider.</p>
@@ -3252,14 +3552,14 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.TopDownPredictor.make_pipeline">
-<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.data.pipelines.Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">sleap.nn.data.pipelines.Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L2489-L2535"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Provider</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline"><span class="pre">Pipeline</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L2500-L2546"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.TopDownPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a data loading pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the pipeline will be created with an instance
 of a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The created <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> with batching and prefetching.</p>
 </dd>
 </dl>
@@ -3274,18 +3574,18 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.inference.VisualPredictor">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">VisualPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L593-L719"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">VisualPredictor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="o"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbosity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">report_rate</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L594-L720"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor" title="Permalink to this definition">#</a></dt>
 <dd><p>Predictor class for generating the visual output of model.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.VisualPredictor.make_pipeline">
-<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L649-L667"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_pipeline</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L650-L668"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.make_pipeline" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a data loading pipeline.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>data_provider</strong> – If not <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code>, the pipeline will be created with an instance
 of a <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The created <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Pipeline</span></code> with batching and prefetching.</p>
 </dd>
 </dl>
@@ -3298,10 +3598,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.VisualPredictor.predict">
-<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.data.pipelines.Provider</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L715-L719"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.predict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_provider</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Provider</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L716-L720"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.predict" title="Permalink to this definition">#</a></dt>
 <dd><p>Run inference on a data source.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>data</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.pipelines.Provider</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> or <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Video</span></code> to
 run inference over.</p></li>
@@ -3310,7 +3610,7 @@ <h1>sleap.nn.inference</h1>
 dictionaries containing the raw arrays returned by the inference model.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Labels</span></code> with <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.PredictedInstance`s</span> <span class="pre">if</span> <span class="pre">`make_labels</span></code> is <code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code>,
 otherwise a list of dictionaries containing batches of numpy arrays with the
 raw results.</p>
@@ -3320,7 +3620,7 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.inference.VisualPredictor.safely_generate">
-<span class="sig-name descname"><span class="pre">safely_generate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.data.ops.dataset_ops.DatasetV2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L669-L703"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.safely_generate" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">safely_generate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">DatasetV2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L670-L704"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.VisualPredictor.safely_generate" title="Permalink to this definition">#</a></dt>
 <dd><p>Yields examples from dataset, catching and logging exceptions.</p>
 </dd></dl>
 
@@ -3328,16 +3628,16 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.export_cli">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">export_cli</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4929-L4944"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.export_cli" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">export_cli</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4998-L5013"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.export_cli" title="Permalink to this definition">#</a></dt>
 <dd><p>CLI for sleap-export.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.export_model">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'exported_model'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4886-L4926"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.export_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">export_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'exported_model'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">signatures</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'serving_default'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_traces</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tensors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unrag_outputs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4955-L4995"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.export_model" title="Permalink to this definition">#</a></dt>
 <dd><p>High level export of a trained SLEAP model as a frozen graph.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to model or list of path to models that were trained by SLEAP.
 These should be the directories that contain <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_job.json</span></code> and
@@ -3362,16 +3662,16 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.find_head">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">find_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1189-L1211"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.find_head" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">find_head</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1191-L1213"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.find_head" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the index of a head in a model’s outputs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> trained by SLEAP.</p></li>
 <li><p><strong>name</strong> – A string that is contained in the model output tensor name.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The index of the first output with a matched name or <code class="xref py py-obj docutils literal notranslate"><span class="pre">None</span></code> if none were found.</p>
 </dd>
 </dl>
@@ -3389,13 +3689,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.get_keras_model_path">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">get_keras_model_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L131-L143"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.get_keras_model_path" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">get_keras_model_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L132-L144"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.get_keras_model_path" title="Permalink to this definition">#</a></dt>
 <dd><p>Utility method for finding the path to a saved Keras model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>path</strong> – Path to a model run folder or job file.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Path to <code class="xref py py-obj docutils literal notranslate"><span class="pre">best_model.h5</span></code> in the run folder.</p>
 </dd>
 </dl>
@@ -3403,10 +3703,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.get_model_output_stride">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">get_model_output_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L1159-L1186"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.get_model_output_stride" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">get_model_output_stride</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">input_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L1161-L1188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.get_model_output_stride" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the stride (1/scale) of the model outputs relative to the input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code>.</p></li>
 <li><p><strong>input_ind</strong> – The index of the input to use as reference. Defaults to 0, indicating
@@ -3415,7 +3715,7 @@ <h1>sleap.nn.inference</h1>
 indicating the last output for multi-output models.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The output stride of the model computed as the integer ratio of the input’s
 height relative to the output’s height, e.g., for a single input/output model:</p>
 <blockquote>
@@ -3429,10 +3729,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.load_model">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">load_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'integral'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker_max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">disable_gpu_preallocation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_reporting</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><span class="pre">sleap.nn.inference.Predictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4744-L4883"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.load_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">load_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">batch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'integral'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker_max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">disable_gpu_preallocation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">progress_reporting</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'rich'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">resize_input_layer</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><span class="pre">Predictor</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4813-L4952"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.load_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a trained SLEAP model.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>model_path</strong> – Path to model or list of path to models that were trained by SLEAP.
 These should be the directories that contain <code class="xref py py-obj docutils literal notranslate"><span class="pre">training_job.json</span></code> and
@@ -3470,7 +3770,7 @@ <h1>sleap.nn.inference</h1>
 tracking is not being applied.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>An instance of a <a class="reference internal" href="#sleap.nn.inference.Predictor" title="sleap.nn.inference.Predictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Predictor</span></code></a> based on which model type was detected.</p>
 <p>If this is a top-down model, paths to the centroids model as well as the
 centered instance model must be provided. A <a class="reference internal" href="#sleap.nn.inference.TopDownPredictor" title="sleap.nn.inference.TopDownPredictor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">TopDownPredictor</span></code></a> instance will be
@@ -3488,10 +3788,10 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.main">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L5351-L5486"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.main" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">list</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L5503-L5719"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.main" title="Permalink to this definition">#</a></dt>
 <dd><p>Entrypoint for <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap-track</span></code> CLI for running inference.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>args</strong> – A list of arguments to be passed into sleap-track.</p>
 </dd>
 </dl>
@@ -3499,13 +3799,13 @@ <h1>sleap.nn.inference</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.inference.make_model_movenet">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">make_model_movenet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">keras.engine.training.Model</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/inference.py#L4496-L4530"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.make_model_movenet" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.inference.</span></span><span class="sig-name descname"><span class="pre">make_model_movenet</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Model</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/inference.py#L4556-L4590"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.inference.make_model_movenet" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a MoveNet model by name.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>model_name</strong> – Name of the model (“lightning” or “thunder”)</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tf.keras.Model ready for inference.</p>
 </dd>
 </dl>
@@ -3535,6 +3835,312 @@ <h1>sleap.nn.inference</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.paf_scorer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.paf_scorer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.cm_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.cm_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.paf_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.paf_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_pafs"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_pafs</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.return_paf_graph"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.return_paf_graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.pafs_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.pafs_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.find_peaks"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.find_peaks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceLayer.forward_pass"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceLayer.forward_pass()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel.bottomup_layer"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel.bottomup_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">BottomUpInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.cm_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.cm_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_output_stride"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.class_maps_output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_class_maps"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.return_class_maps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.class_maps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.find_peaks()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceLayer.forward_pass()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.inference_layer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel.inference_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.config"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.tracks"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpMultiClassPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.bottomup_config"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.bottomup_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.bottomup_model"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.bottomup_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.max_edge_length_ratio"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.max_edge_length_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.dist_penalty_weight"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.dist_penalty_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.paf_line_points"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.paf_line_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.min_line_scores"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.min_line_scores</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.max_instances"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.BottomUpPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">BottomUpPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop"><code class="docutils literal notranslate"><span class="pre">CentroidCrop</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.keras_model"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.crop_size"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.input_scale"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.output_stride"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.peak_threshold"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.refinement"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.return_confmaps"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.offsets_ind"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.return_crops"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.return_crops</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.max_instances"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCrop.call"><code class="docutils literal notranslate"><span class="pre">CentroidCrop.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth.crop_size"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth.crop_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidCropGroundTruth.call"><code class="docutils literal notranslate"><span class="pre">CentroidCropGroundTruth.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.CentroidInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">CentroidInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.keras_model"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.input_scale"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.output_stride"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.peak_threshold"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.refinement"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.return_confmaps"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.offsets_ind"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaks.call"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaks.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaksGroundTruth"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaksGroundTruth</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.FindInstancePeaksGroundTruth.call"><code class="docutils literal notranslate"><span class="pre">FindInstancePeaksGroundTruth.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer"><code class="docutils literal notranslate"><span class="pre">InferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.ensure_float"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceLayer.preprocess"><code class="docutils literal notranslate"><span class="pre">InferenceLayer.preprocess()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel"><code class="docutils literal notranslate"><span class="pre">InferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.export_model"><code class="docutils literal notranslate"><span class="pre">InferenceModel.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.predict"><code class="docutils literal notranslate"><span class="pre">InferenceModel.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.InferenceModel.predict_on_batch"><code class="docutils literal notranslate"><span class="pre">InferenceModel.predict_on_batch()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.model_name"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.ensure_grayscale"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.ensure_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.ensure_float"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.ensure_float</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceLayer.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel.inference_layer"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel.inference_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">MoveNetInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.model_name"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.model_name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.MoveNetPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">MoveNetPredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor"><code class="docutils literal notranslate"><span class="pre">Predictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.export_model"><code class="docutils literal notranslate"><span class="pre">Predictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.from_model_paths"><code class="docutils literal notranslate"><span class="pre">Predictor.from_model_paths()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">Predictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">Predictor.make_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.predict"><code class="docutils literal notranslate"><span class="pre">Predictor.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.Predictor.report_period"><code class="docutils literal notranslate"><span class="pre">Predictor.report_period</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.RateColumn"><code class="docutils literal notranslate"><span class="pre">RateColumn</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.RateColumn.render"><code class="docutils literal notranslate"><span class="pre">RateColumn.render()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.keras_model"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.input_scale"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.pad_to_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.pad_to_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.output_stride"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.peak_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.return_confmaps"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.offsets_ind"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceLayer.call"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceLayer.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel.single_instance_layer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel.single_instance_layer</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstanceInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">SingleInstanceInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.export_model"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.SingleInstancePredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">SingleInstancePredictor.is_grayscale</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.instance_peaks"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">TopDownInferenceModel.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.keras_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.input_scale"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.input_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.output_stride"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.output_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.refinement"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.return_confmaps"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.return_confmaps</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.return_class_vectors"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.return_class_vectors</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.confmaps_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.confmaps_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.offsets_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.offsets_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.class_vectors_ind"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.class_vectors_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.optimal_grouping"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.optimal_grouping</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassFindPeaks.call"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassFindPeaks.call()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.centroid_crop"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.centroid_crop</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.instance_peaks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.call"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassInferenceModel.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassInferenceModel.export_model()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.centroid_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.centroid_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.centroid_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.centroid_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.tracks"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassPredictor.make_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.centroid_config"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.centroid_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.centroid_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.centroid_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.confmap_config"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.confmap_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.confmap_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.confmap_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.inference_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.inference_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.tracker"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.tracker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.batch_size"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.batch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.peak_threshold"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.peak_threshold</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.integral_refinement"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.integral_refinement</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.integral_patch_size"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.integral_patch_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.max_instances"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.max_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.export_model"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.export_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.from_trained_models"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.from_trained_models()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.is_grayscale"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.is_grayscale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.TopDownPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">TopDownPredictor.make_pipeline()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor"><code class="docutils literal notranslate"><span class="pre">VisualPredictor</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.make_pipeline"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.make_pipeline()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.predict"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.VisualPredictor.safely_generate"><code class="docutils literal notranslate"><span class="pre">VisualPredictor.safely_generate()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.export_cli"><code class="docutils literal notranslate"><span class="pre">export_cli()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.export_model"><code class="docutils literal notranslate"><span class="pre">export_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.find_head"><code class="docutils literal notranslate"><span class="pre">find_head()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.get_keras_model_path"><code class="docutils literal notranslate"><span class="pre">get_keras_model_path()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.get_model_output_stride"><code class="docutils literal notranslate"><span class="pre">get_model_output_stride()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.load_model"><code class="docutils literal notranslate"><span class="pre">load_model()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.inference.make_model_movenet"><code class="docutils literal notranslate"><span class="pre">make_model_movenet()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -3553,7 +4159,7 @@ <h1>sleap.nn.inference</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.losses.html b/api/sleap.nn.losses.html
index 781747871..46bc84ad3 100644
--- a/api/sleap.nn.losses.html
+++ b/api/sleap.nn.losses.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.losses &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.losses &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.losses';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.losses.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,30 @@ <h1>sleap.nn.losses</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss"><code class="docutils literal notranslate"><span class="pre">OHKMLoss</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.hard_to_easy_ratio"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.hard_to_easy_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.min_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.min_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.max_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.max_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.loss_scale"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.loss_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.call"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.from_config"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.from_config()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss"><code class="docutils literal notranslate"><span class="pre">PartLoss</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.channel_ind"><code class="docutils literal notranslate"><span class="pre">PartLoss.channel_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.name"><code class="docutils literal notranslate"><span class="pre">PartLoss.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.result"><code class="docutils literal notranslate"><span class="pre">PartLoss.result()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.update_state"><code class="docutils literal notranslate"><span class="pre">PartLoss.update_state()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.compute_ohkm_loss"><code class="docutils literal notranslate"><span class="pre">compute_ohkm_loss()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +342,11 @@ <h1>sleap.nn.losses</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.losses">
-<span id="sleap-nn-losses"></span><h1>sleap.nn.losses<a class="headerlink" href="#module-sleap.nn.losses" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-losses"></span><h1>sleap.nn.losses<a class="headerlink" href="#module-sleap.nn.losses" title="Permalink to this heading">#</a></h1>
 <p>Custom loss functions and metrics.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.losses.OHKMLoss">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">OHKMLoss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'ohkm'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L54-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">OHKMLoss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'ohkm'</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L54-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss" title="Permalink to this definition">#</a></dt>
 <dd><p>Online hard keypoint mining loss.</p>
 <p>This loss serves to dynamically reweight the MSE of the top-K worst channels in each
 batch. This is useful when fine tuning a model to improve performance on a hard
@@ -360,10 +384,10 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.losses.OHKMLoss.call">
-<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_weight</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L102-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss.call" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">call</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_weight</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L102-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss.call" title="Permalink to this definition">#</a></dt>
 <dd><p>Invokes the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Loss</span></code> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>y_true</strong> – Ground truth values. shape = <code class="xref py py-obj docutils literal notranslate"><span class="pre">[batch_size,</span> <span class="pre">d0,</span> <span class="pre">..</span> <span class="pre">dN]</span></code>, except
 sparse loss functions such as sparse categorical crossentropy where
@@ -371,7 +395,7 @@ <h1>sleap.nn.losses</h1>
 <li><p><strong>y_pred</strong> – The predicted values. shape = <code class="xref py py-obj docutils literal notranslate"><span class="pre">[batch_size,</span> <span class="pre">d0,</span> <span class="pre">..</span> <span class="pre">dN]</span></code></p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Loss values with the shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">[batch_size,</span> <span class="pre">d0,</span> <span class="pre">..</span> <span class="pre">dN-1]</span></code>.</p>
 </dd>
 </dl>
@@ -379,13 +403,13 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.losses.OHKMLoss.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.HardKeypointMiningConfig" title="sleap.nn.config.optimization.HardKeypointMiningConfig"><span class="pre">sleap.nn.config.optimization.HardKeypointMiningConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.losses.OHKMLoss" title="sleap.nn.losses.OHKMLoss"><span class="pre">sleap.nn.losses.OHKMLoss</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L91-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.HardKeypointMiningConfig" title="sleap.nn.config.optimization.HardKeypointMiningConfig"><span class="pre">HardKeypointMiningConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.losses.OHKMLoss" title="sleap.nn.losses.OHKMLoss"><span class="pre">OHKMLoss</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L91-L100"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.OHKMLoss.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiates a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Loss</span></code> from its config (output of <code class="xref py py-obj docutils literal notranslate"><span class="pre">get_config()</span></code>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>config</strong> – Output of <code class="xref py py-obj docutils literal notranslate"><span class="pre">get_config()</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Loss</span></code> instance.</p>
 </dd>
 </dl>
@@ -395,7 +419,7 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.losses.PartLoss">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">PartLoss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L113-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">PartLoss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">args</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L113-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute channelwise loss.</p>
 <p>Useful for monitoring the MSE for specific body parts (channels).</p>
 <dl class="py attribute">
@@ -412,7 +436,7 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.losses.PartLoss.result">
-<span class="sig-name descname"><span class="pre">result</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L154-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss.result" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">result</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L154-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss.result" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes and returns the metric value tensor.</p>
 <p>Result computation is an idempotent operation that simply calculates the
 metric value using the state variables.</p>
@@ -420,7 +444,7 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.losses.PartLoss.update_state">
-<span class="sig-name descname"><span class="pre">update_state</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_weight</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L139-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss.update_state" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">update_state</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_weight</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L139-L152"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.PartLoss.update_state" title="Permalink to this definition">#</a></dt>
 <dd><p>Accumulates statistics for the metric.</p>
 <p>Note: This function is executed as a graph function in graph mode.
 This means:</p>
@@ -436,7 +460,7 @@ <h1>sleap.nn.losses</h1>
 eager execution.</p>
 </div></blockquote>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>*args</strong> – </p></li>
 <li><p><strong>**kwargs</strong> – A mini-batch of inputs to the Metric.</p></li>
@@ -449,7 +473,7 @@ <h1>sleap.nn.losses</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.losses.compute_ohkm_loss">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">compute_ohkm_loss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-</span> <span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/losses.py#L7-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.compute_ohkm_loss" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.losses.</span></span><span class="sig-name descname"><span class="pre">compute_ohkm_loss</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hard_to_easy_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_hard_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">loss_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/losses.py#L7-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.losses.compute_ohkm_loss" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the online hard keypoint mining loss.</p>
 </dd></dl>
 
@@ -477,6 +501,36 @@ <h1>sleap.nn.losses</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss"><code class="docutils literal notranslate"><span class="pre">OHKMLoss</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.hard_to_easy_ratio"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.hard_to_easy_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.min_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.min_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.max_hard_keypoints"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.max_hard_keypoints</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.loss_scale"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.loss_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.call"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.call()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.OHKMLoss.from_config"><code class="docutils literal notranslate"><span class="pre">OHKMLoss.from_config()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss"><code class="docutils literal notranslate"><span class="pre">PartLoss</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.channel_ind"><code class="docutils literal notranslate"><span class="pre">PartLoss.channel_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.name"><code class="docutils literal notranslate"><span class="pre">PartLoss.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.result"><code class="docutils literal notranslate"><span class="pre">PartLoss.result()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.PartLoss.update_state"><code class="docutils literal notranslate"><span class="pre">PartLoss.update_state()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.losses.compute_ohkm_loss"><code class="docutils literal notranslate"><span class="pre">compute_ohkm_loss()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -495,7 +549,7 @@ <h1>sleap.nn.losses</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.model.html b/api/sleap.nn.model.html
index 2e8a88324..8d2f42c0c 100644
--- a/api/sleap.nn.model.html
+++ b/api/sleap.nn.model.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.model &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.model &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.model';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.model.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,22 @@ <h1>sleap.nn.model</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model"><code class="docutils literal notranslate"><span class="pre">Model</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.backbone"><code class="docutils literal notranslate"><span class="pre">Model.backbone</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.heads"><code class="docutils literal notranslate"><span class="pre">Model.heads</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.keras_model"><code class="docutils literal notranslate"><span class="pre">Model.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.from_config"><code class="docutils literal notranslate"><span class="pre">Model.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.make_model"><code class="docutils literal notranslate"><span class="pre">Model.make_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.maximum_stride"><code class="docutils literal notranslate"><span class="pre">Model.maximum_stride</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +334,14 @@ <h1>sleap.nn.model</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.model">
-<span id="sleap-nn-model"></span><h1>sleap.nn.model<a class="headerlink" href="#module-sleap.nn.model" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-model"></span><h1>sleap.nn.model<a class="headerlink" href="#module-sleap.nn.model" title="Permalink to this heading">#</a></h1>
 <p>This module defines the main SLEAP model class for defining a trainable model.</p>
 <p>This is a higher level wrapper around <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> that holds all the configuration
 parameters required to construct the actual model. This allows for easy querying of the
 model configuration without actually instantiating the model itself.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.model.Model">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.model.</span></span><span class="sig-name descname"><span class="pre">Model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.model.Architecture</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">heads</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.engine.training.Model</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/model.py#L90-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.model.</span></span><span class="sig-name descname"><span class="pre">Model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backbone</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Architecture</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">heads</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Model</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/model.py#L90-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model" title="Permalink to this definition">#</a></dt>
 <dd><p>SLEAP model that describes an architecture and output types.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.model.Model.backbone">
@@ -333,7 +349,7 @@ <h1>sleap.nn.model</h1>
 <dd><p>An <code class="xref py py-obj docutils literal notranslate"><span class="pre">Architecture</span></code> class that provides methods for building a
 tf.keras.Model given an input.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>sleap.nn.model.Architecture</p>
 </dd>
 </dl>
@@ -344,7 +360,7 @@ <h1>sleap.nn.model</h1>
 <span class="sig-name descname"><span class="pre">heads</span></span><a class="headerlink" href="#sleap.nn.model.Model.heads" title="Permalink to this definition">#</a></dt>
 <dd><p>List of <a href="#id1"><span class="problematic" id="id2">`</span></a>Head`s that define the outputs of the network.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[sleap.nn.model.Head]</p>
 </dd>
 </dl>
@@ -355,7 +371,7 @@ <h1>sleap.nn.model</h1>
 <span class="sig-name descname"><span class="pre">keras_model</span></span><a class="headerlink" href="#sleap.nn.model.Model.keras_model" title="Permalink to this definition">#</a></dt>
 <dd><p>The current <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> instance if one has been created.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[keras.engine.training.Model]</p>
 </dd>
 </dl>
@@ -363,10 +379,10 @@ <h1>sleap.nn.model</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.model.Model.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ModelConfig" title="sleap.nn.config.model.ModelConfig"><span class="pre">sleap.nn.config.model.ModelConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/model.py#L104-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.ModelConfig" title="sleap.nn.config.model.ModelConfig"><span class="pre">ModelConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.skeleton.html#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/model.py#L104-L305"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a SLEAP model from configurations.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – The configurations as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">ModelConfig</span></code> instance.</p></li>
 <li><p><strong>skeleton</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Skeleton</span></code> to use if not provided in the config.</p></li>
@@ -374,7 +390,7 @@ <h1>sleap.nn.model</h1>
 values inferred from other fields.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of <a class="reference internal" href="#sleap.nn.model.Model" title="sleap.nn.model.Model"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Model</span></code></a> built with the specified configurations.</p>
 </dd>
 </dl>
@@ -382,14 +398,14 @@ <h1>sleap.nn.model</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.model.Model.make_model">
-<span class="sig-name descname"><span class="pre">make_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">keras.engine.training.Model</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/model.py#L312-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model.make_model" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">make_model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">input_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Model</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/model.py#L312-L364"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.model.Model.make_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a trainable model by connecting the backbone with the heads.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>input_shape</strong> – Tuple of (height, width, channels) specifying the shape of the
 inputs before preprocessing.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instantiated <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code>.</p>
 </dd>
 </dl>
@@ -427,6 +443,28 @@ <h1>sleap.nn.model</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model"><code class="docutils literal notranslate"><span class="pre">Model</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.backbone"><code class="docutils literal notranslate"><span class="pre">Model.backbone</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.heads"><code class="docutils literal notranslate"><span class="pre">Model.heads</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.keras_model"><code class="docutils literal notranslate"><span class="pre">Model.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.from_config"><code class="docutils literal notranslate"><span class="pre">Model.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.make_model"><code class="docutils literal notranslate"><span class="pre">Model.make_model()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.model.Model.maximum_stride"><code class="docutils literal notranslate"><span class="pre">Model.maximum_stride</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -445,7 +483,7 @@ <h1>sleap.nn.model</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.paf_grouping.html b/api/sleap.nn.paf_grouping.html
index 9bd3b9e0f..6eaf5929c 100644
--- a/api/sleap.nn.paf_grouping.html
+++ b/api/sleap.nn.paf_grouping.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.paf_grouping &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.paf_grouping &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.paf_grouping';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.paf_grouping.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,63 @@ <h1>sleap.nn.paf_grouping</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection"><code class="docutils literal notranslate"><span class="pre">EdgeConnection</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.src_peak_ind"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.src_peak_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.dst_peak_ind"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.dst_peak_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.score"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.score</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType"><code class="docutils literal notranslate"><span class="pre">EdgeType</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType.src_node_ind"><code class="docutils literal notranslate"><span class="pre">EdgeType.src_node_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType.dst_node_ind"><code class="docutils literal notranslate"><span class="pre">EdgeType.dst_node_ind</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer"><code class="docutils literal notranslate"><span class="pre">PAFScorer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.part_names"><code class="docutils literal notranslate"><span class="pre">PAFScorer.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edges"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.pafs_stride"><code class="docutils literal notranslate"><span class="pre">PAFScorer.pafs_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.max_edge_length_ratio"><code class="docutils literal notranslate"><span class="pre">PAFScorer.max_edge_length_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.dist_penalty_weight"><code class="docutils literal notranslate"><span class="pre">PAFScorer.dist_penalty_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_points"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.min_instance_peaks"><code class="docutils literal notranslate"><span class="pre">PAFScorer.min_instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.min_line_scores"><code class="docutils literal notranslate"><span class="pre">PAFScorer.min_line_scores</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edge_inds"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edge_types"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edge_types</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_nodes"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_edges"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.sorted_edge_inds"><code class="docutils literal notranslate"><span class="pre">PAFScorer.sorted_edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.from_config"><code class="docutils literal notranslate"><span class="pre">PAFScorer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.group_instances"><code class="docutils literal notranslate"><span class="pre">PAFScorer.group_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.match_candidates"><code class="docutils literal notranslate"><span class="pre">PAFScorer.match_candidates()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.predict"><code class="docutils literal notranslate"><span class="pre">PAFScorer.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.score_paf_lines"><code class="docutils literal notranslate"><span class="pre">PAFScorer.score_paf_lines()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID"><code class="docutils literal notranslate"><span class="pre">PeakID</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID.node_ind"><code class="docutils literal notranslate"><span class="pre">PeakID.node_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID.peak_ind"><code class="docutils literal notranslate"><span class="pre">PeakID.peak_ind</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.assign_connections_to_instances"><code class="docutils literal notranslate"><span class="pre">assign_connections_to_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.compute_distance_penalty"><code class="docutils literal notranslate"><span class="pre">compute_distance_penalty()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.get_connection_candidates"><code class="docutils literal notranslate"><span class="pre">get_connection_candidates()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.get_paf_lines"><code class="docutils literal notranslate"><span class="pre">get_paf_lines()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.group_instances_batch"><code class="docutils literal notranslate"><span class="pre">group_instances_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.group_instances_sample"><code class="docutils literal notranslate"><span class="pre">group_instances_sample()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.make_line_subs"><code class="docutils literal notranslate"><span class="pre">make_line_subs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.make_predicted_instances"><code class="docutils literal notranslate"><span class="pre">make_predicted_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.match_candidates_batch"><code class="docutils literal notranslate"><span class="pre">match_candidates_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.match_candidates_sample"><code class="docutils literal notranslate"><span class="pre">match_candidates_sample()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.score_paf_lines"><code class="docutils literal notranslate"><span class="pre">score_paf_lines()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.score_paf_lines_batch"><code class="docutils literal notranslate"><span class="pre">score_paf_lines_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.toposort_edges"><code class="docutils literal notranslate"><span class="pre">toposort_edges()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +375,7 @@ <h1>sleap.nn.paf_grouping</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.paf_grouping">
-<span id="sleap-nn-paf-grouping"></span><h1>sleap.nn.paf_grouping<a class="headerlink" href="#module-sleap.nn.paf_grouping" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-paf-grouping"></span><h1>sleap.nn.paf_grouping<a class="headerlink" href="#module-sleap.nn.paf_grouping" title="Permalink to this heading">#</a></h1>
 <p>This module provides a set of utilities for grouping peaks based on PAFs.</p>
 <p>Part affinity fields (PAFs) are a representation used to resolve the peak grouping
 problem for multi-instance pose estimation [1].</p>
@@ -334,15 +391,16 @@ <h1>sleap.nn.paf_grouping</h1>
 network-based prediction from unlabeled images.</p>
 <p>A high-level API for grouping based on PAFs is provided through the <a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer" title="sleap.nn.paf_grouping.PAFScorer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PAFScorer</span></code></a> class.</p>
 <p class="rubric">References</p>
-<dl class="footnote brackets">
-<dt class="label" id="id1"><span class="brackets">1</span></dt>
-<dd><p>Zhe Cao, Tomas Simon, Shih-En Wei, Yaser Sheikh. Realtime Multi-Person 2D
+<aside class="footnote-list brackets">
+<aside class="footnote brackets" id="id1" role="note">
+<span class="label"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></span>
+<p>Zhe Cao, Tomas Simon, Shih-En Wei, Yaser Sheikh. Realtime Multi-Person 2D
 Pose Estimation using Part Affinity Fields. In _CVPR_, 2017.</p>
-</dd>
-</dl>
+</aside>
+</aside>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.EdgeConnection">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">EdgeConnection</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src_peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dst_peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L66-L79"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeConnection" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">EdgeConnection</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src_peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dst_peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L66-L79"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeConnection" title="Permalink to this definition">#</a></dt>
 <dd><p>Indices to specify a matched connection between two peaks.</p>
 <p>This is a convenience named tuple for use in the matching pipeline.</p>
 <dl class="py attribute">
@@ -350,7 +408,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">src_peak_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeConnection.src_peak_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the source peak within all peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -361,7 +419,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">dst_peak_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeConnection.dst_peak_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the destination peak within all peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -372,7 +430,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">score</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeConnection.score" title="Permalink to this definition">#</a></dt>
 <dd><p>Score of the match.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -382,7 +440,7 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.EdgeType">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">EdgeType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src_node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dst_node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L51-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeType" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">EdgeType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src_node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dst_node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L51-L62"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeType" title="Permalink to this definition">#</a></dt>
 <dd><p>Indices to uniquely identify a single edge type.</p>
 <p>This is a convenience named tuple for use in the matching pipeline.</p>
 <dl class="py attribute">
@@ -390,7 +448,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">src_node_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeType.src_node_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the source node type within the skeleton edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -401,7 +459,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">dst_node_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.EdgeType.dst_node_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the destination node type within the skeleton edges.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -411,7 +469,7 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">PAFScorer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1319-L1705"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">PAFScorer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1319-L1705"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer" title="Permalink to this definition">#</a></dt>
 <dd><p>Scoring pipeline based on part affinity fields.</p>
 <p>This class facilitates grouping of predicted peaks based on PAFs. It holds a set of
 common parameters that are used across different steps of the pipeline.</p>
@@ -420,7 +478,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">part_names</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.part_names" title="Permalink to this definition">#</a></dt>
 <dd><p>List of string node names in the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[str]</p>
 </dd>
 </dl>
@@ -431,7 +489,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">edges</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.edges" title="Permalink to this definition">#</a></dt>
 <dd><p>List of (src_node, dst_node) names in the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[Tuple[str, str]]</p>
 </dd>
 </dl>
@@ -443,7 +501,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <dd><p>Output stride of the part affinity fields. This will be used to
 adjust the peak coordinates from full image to PAF subscripts.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -456,7 +514,7 @@ <h1>sleap.nn.paf_grouping</h1>
 as a fraction of the image size. Candidate connections longer than this
 length will be penalized during matching.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -469,7 +527,7 @@ <h1>sleap.nn.paf_grouping</h1>
 a scalar float. Set to values greater than 1.0 to enforce the distance
 penalty more strictly.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -480,7 +538,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">n_points</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.n_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Number of points to sample along the line integral.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -493,7 +551,7 @@ <h1>sleap.nn.paf_grouping</h1>
 considered a real instance. Instances with fewer peaks than this will be
 discarded (useful for filtering spurious detections).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Union[int, float]</p>
 </dd>
 </dl>
@@ -506,7 +564,7 @@ <h1>sleap.nn.paf_grouping</h1>
 between candidate point pairs. Useful for rejecting spurious detections when
 there are no better ones.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -518,7 +576,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <dd><p>The edges of the skeleton defined as a list of (source, destination)
 tuples of node indices. This is created automatically on initialization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[Tuple[int, int]]</p>
 </dd>
 </dl>
@@ -530,7 +588,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <dd><p>A list of <a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EdgeType</span></code></a> instances representing the edges of the
 skeleton. This is created automatically on initialization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType">sleap.nn.paf_grouping.EdgeType</a>]</p>
 </dd>
 </dl>
@@ -542,7 +600,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <dd><p>The number of nodes in the skeleton as a scalar <code class="xref py py-obj docutils literal notranslate"><span class="pre">int</span></code>. This is created
 automatically on initialization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -554,7 +612,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <dd><p>The number of edges in the skeleton as a scalar <code class="xref py py-obj docutils literal notranslate"><span class="pre">int</span></code>. This is created
 automatically on initialization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -567,7 +625,7 @@ <h1>sleap.nn.paf_grouping</h1>
 edge types should be accessed in during instance assembly
 (<a class="reference internal" href="#sleap.nn.paf_grouping.assign_connections_to_instances" title="sleap.nn.paf_grouping.assign_connections_to_instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">assign_connections_to_instances</span></code></a>).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Tuple[int]</p>
 </dd>
 </dl>
@@ -595,10 +653,10 @@ <h1>sleap.nn.paf_grouping</h1>
 </div>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfig" title="sleap.nn.config.model.MultiInstanceConfig"><span class="pre">sleap.nn.config.model.MultiInstanceConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer" title="sleap.nn.paf_grouping.PAFScorer"><span class="pre">sleap.nn.paf_grouping.PAFScorer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1406-L1447"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.model.html#sleap.nn.config.model.MultiInstanceConfig" title="sleap.nn.config.model.MultiInstanceConfig"><span class="pre">MultiInstanceConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer" title="sleap.nn.paf_grouping.PAFScorer"><span class="pre">PAFScorer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1406-L1447"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Initialize the PAF scorer from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiInstanceConfig</span></code> head config.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">MultiInstanceConfig</span></code> from <code class="xref py py-obj docutils literal notranslate"><span class="pre">cfg.model.heads.multi_instance</span></code>.</p></li>
 <li><p><strong>max_edge_length_ratio</strong> – The maximum expected length of a connected pair of
@@ -617,7 +675,7 @@ <h1>sleap.nn.paf_grouping</h1>
 detections when there are no better ones.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The initialized instance of <a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer" title="sleap.nn.paf_grouping.PAFScorer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PAFScorer</span></code></a>.</p>
 </dd>
 </dl>
@@ -625,10 +683,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer.group_instances">
-<span class="sig-name descname"><span class="pre">group_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1552-L1627"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.group_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">group_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1552-L1627"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.group_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Group matched connections into full instances for a batch.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peaks</strong> – The sample-grouped detected peaks in a batch as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code>
 of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">(n_peaks),</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. These
@@ -658,7 +716,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. This can be generated by <a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer.match_candidates" title="sleap.nn.paf_grouping.PAFScorer.match_candidates"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PAFScorer.match_candidates()</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of arrays with the grouped instances for the whole batch grouped by
 sample:</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">predicted_instances</span></code>: The sample- and instance-grouped coordinates for each
@@ -683,10 +741,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer.match_candidates">
-<span class="sig-name descname"><span class="pre">match_candidates</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1498-L1550"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.match_candidates" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">match_candidates</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1498-L1550"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.match_candidates" title="Permalink to this definition">#</a></dt>
 <dd><p>Match candidate connections for a batch based on PAF scores.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>edge_inds</strong> – Sample-grouped edge indices as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">(n_candidates))</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code> indicating the
@@ -701,7 +759,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. Can be generated using <a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer.score_paf_lines" title="sleap.nn.paf_grouping.PAFScorer.score_paf_lines"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PAFScorer.score_paf_lines()</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The connection peaks for each edge matched based on score as tuple of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(match_edge_inds,</span> <span class="pre">match_src_peak_inds,</span> <span class="pre">match_dst_peak_inds,</span> <span class="pre">match_line_scores)</span></code></p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">match_edge_inds</span></code>: Sample-grouped indices of the skeleton edge for each
@@ -732,10 +790,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer.predict">
-<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1629-L1705"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.predict" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">predict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1629-L1705"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.predict" title="Permalink to this definition">#</a></dt>
 <dd><p>Group a batch of predicted peaks into full instance predictions using PAFs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>pafs</strong> – The batch of part affinity fields as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">n_edges)</span></code> and type <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p></li>
@@ -749,7 +807,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of arrays with the grouped instances for the whole batch grouped by
 sample:</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">predicted_instances</span></code>: The sample- and instance-grouped coordinates for each
@@ -779,10 +837,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PAFScorer.score_paf_lines">
-<span class="sig-name descname"><span class="pre">score_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1449-L1496"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.score_paf_lines" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">score_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1449-L1496"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PAFScorer.score_paf_lines" title="Permalink to this definition">#</a></dt>
 <dd><p>Create and score PAF lines formed between connection candidates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>pafs</strong> – The batch of part affinity fields as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">n_edges)</span></code> and type <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p></li>
@@ -793,7 +851,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(edge_inds,</span> <span class="pre">edge_peak_inds,</span> <span class="pre">line_scores)</span></code> with the connections
 and their scores based on the PAFs.</p>
 <p><a class="reference internal" href="#sleap.nn.paf_grouping.PAFScorer.edge_inds" title="sleap.nn.paf_grouping.PAFScorer.edge_inds"><code class="xref py py-obj docutils literal notranslate"><span class="pre">edge_inds</span></code></a>: Sample-grouped indices of the edge in the skeleton that each
@@ -821,7 +879,7 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.PeakID">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">PeakID</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L36-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PeakID" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">PeakID</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L36-L47"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.PeakID" title="Permalink to this definition">#</a></dt>
 <dd><p>Indices to uniquely identify a single peak.</p>
 <p>This is a convenience named tuple for use in the matching pipeline.</p>
 <dl class="py attribute">
@@ -829,7 +887,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">node_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.PeakID.node_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the node type (channel) of the peak.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -840,7 +898,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <span class="sig-name descname"><span class="pre">peak_ind</span></span><a class="headerlink" href="#sleap.nn.paf_grouping.PeakID.peak_ind" title="Permalink to this definition">#</a></dt>
 <dd><p>Index of the peak within its node type.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -850,10 +908,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.assign_connections_to_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">assign_connections_to_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">sleap.nn.paf_grouping.EdgeType</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeConnection" title="sleap.nn.paf_grouping.EdgeConnection"><span class="pre">sleap.nn.paf_grouping.EdgeConnection</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.PeakID" title="sleap.nn.paf_grouping.PeakID"><span class="pre">sleap.nn.paf_grouping.PeakID</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L799-L914"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.assign_connections_to_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">assign_connections_to_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">EdgeType</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeConnection" title="sleap.nn.paf_grouping.EdgeConnection"><span class="pre">EdgeConnection</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.PeakID" title="sleap.nn.paf_grouping.PeakID"><span class="pre">PeakID</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L799-L914"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.assign_connections_to_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Assigns connected edges to instances via greedy graph partitioning.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>connections</strong> – A dict that maps EdgeType to a list of EdgeConnections found
 through connection scoring. This can be generated by the
@@ -868,7 +926,7 @@ <h1>sleap.nn.paf_grouping</h1>
 count is inferred from the unique node inds in connections.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A dict mapping PeakID to a unique instance ID specified
 as an integer.</p>
 <p>A PeakID is a tuple of (node_type_ind, peak_ind), where the peak_ind is the
@@ -876,7 +934,7 @@ <h1>sleap.nn.paf_grouping</h1>
 dst_peak_ind.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>instance_assignments</p>
 </dd>
 </dl>
@@ -890,10 +948,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.compute_distance_penalty">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">compute_distance_penalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">spatial_vec_lengths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L278-L322"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.compute_distance_penalty" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">compute_distance_penalty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">spatial_vec_lengths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L278-L322"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.compute_distance_penalty" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the distance penalty component of the PAF line integral score.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>spatial_vec_lengths</strong> – Euclidean distance between candidate source and
 destination points as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> tensor of any shape (typically
@@ -907,7 +965,7 @@ <h1>sleap.nn.paf_grouping</h1>
 penalty more strictly.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The distance penalty for each candidate as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code> tensor of the same
 shape as <code class="xref py py-obj docutils literal notranslate"><span class="pre">spatial_vec_lengths</span></code>.</p>
 <p>The penalty will be 0 (when below the threshold) and -1 as the distance
@@ -936,10 +994,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.get_connection_candidates">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">get_connection_candidates</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L82-L142"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.get_connection_candidates" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">get_connection_candidates</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L82-L142"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.get_connection_candidates" title="Permalink to this definition">#</a></dt>
 <dd><p>Find the indices of all the possible connections formed by the detected peaks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peak_channel_inds_sample</strong> – The channel indices of the peaks found in a sample.
 This is a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_peaks,)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code> that is
@@ -950,7 +1008,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <li><p><strong>n_nodes</strong> – The total number of nodes in the skeleton as a scalar integer.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(edge_inds,</span> <span class="pre">edge_peak_inds)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">edge_inds</span></code> is a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_candidates,)</span></code> indicating the indices
 of the edge that each of the candidate connections belongs to.</p>
@@ -964,10 +1022,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.get_paf_lines">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">get_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L225-L275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.get_paf_lines" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">get_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L225-L275"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.get_paf_lines" title="Permalink to this definition">#</a></dt>
 <dd><p>Gets the PAF values at the lines formed between all detected peaks in a sample.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>pafs_sample</strong> – The PAFs for the sample as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">n_edges)</span></code>.</p></li>
@@ -989,7 +1047,7 @@ <h1>sleap.nn.paf_grouping</h1>
 divided by this value to adjust the indexing into the PAFs tensor.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The PAF vectors at all of the line points as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_candidates,</span> <span class="pre">n_line_points,</span> <span class="pre">2,</span> <span class="pre">3)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>. These subscripts can
 be used directly with <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.gather_nd</span></code> to pull out the PAF values at the lines.</p>
@@ -1010,10 +1068,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.group_instances_batch">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">group_instances_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sorted_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">sleap.nn.paf_grouping.EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1115-L1290"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.group_instances_batch" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">group_instances_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_vals</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sorted_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1115-L1290"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.group_instances_batch" title="Permalink to this definition">#</a></dt>
 <dd><p>Group matched connections into full instances for a batch.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peaks</strong> – The sample-grouped detected peaks in a batch as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">(n_peaks),</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. These should be
@@ -1054,7 +1112,7 @@ <h1>sleap.nn.paf_grouping</h1>
 between candidate point pairs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of arrays with the grouped instances for the whole batch grouped by
 sample:</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">predicted_instances</span></code>: The sample- and instance-grouped coordinates for each
@@ -1074,10 +1132,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.group_instances_sample">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">group_instances_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sorted_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">sleap.nn.paf_grouping.EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L984-L1112"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.group_instances_sample" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">group_instances_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_edge_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_src_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_dst_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_line_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sorted_edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_instance_peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L984-L1112"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.group_instances_sample" title="Permalink to this definition">#</a></dt>
 <dd><p>Group matched connections into full instances for a single sample.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peaks_sample</strong> – The detected peaks in a sample as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_peaks,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. These should be <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates
@@ -1117,7 +1175,7 @@ <h1>sleap.nn.paf_grouping</h1>
 between candidate point pairs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">predicted_instances</span></code>: The grouped coordinates for each instance as an array of
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,</span> <span class="pre">n_nodes,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code>. Missing peaks are
 represented by <a href="#id18"><span class="problematic" id="id19">`</span></a>np.NaN`s.</p>
@@ -1127,7 +1185,7 @@ <h1>sleap.nn.paf_grouping</h1>
 shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_instances,)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code>.</p>
 </p>
 </dd>
-<dt class="field-odd">Return type</dt>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A tuple of arrays with the grouped instances</p>
 </dd>
 </dl>
@@ -1140,10 +1198,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.make_line_subs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">make_line_subs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L145-L222"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.make_line_subs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">make_line_subs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L145-L222"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.make_line_subs" title="Permalink to this definition">#</a></dt>
 <dd><p>Create the lines between candidate connections for evaluating the PAFs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peaks_sample</strong> – The detected peaks in a sample as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_peaks,</span> <span class="pre">2)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. These should be <code class="xref py py-obj docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> coordinates
@@ -1163,7 +1221,7 @@ <h1>sleap.nn.paf_grouping</h1>
 divided by this value to adjust the indexing into the PAFs tensor.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The line subscripts as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_candidates,</span> <span class="pre">n_line_points,</span> <span class="pre">2,</span> <span class="pre">3)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>. These subscripts can
 be used directly with <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.gather_nd</span></code> to pull out the PAF values at the lines.</p>
@@ -1184,10 +1242,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.make_predicted_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">make_predicted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeConnection" title="sleap.nn.paf_grouping.EdgeConnection"><span class="pre">sleap.nn.paf_grouping.EdgeConnection</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_assignments</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.PeakID" title="sleap.nn.paf_grouping.PeakID"><span class="pre">sleap.nn.paf_grouping.PeakID</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.array</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.array</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.array</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L917-L981"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.make_predicted_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">make_predicted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">connections</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeConnection" title="sleap.nn.paf_grouping.EdgeConnection"><span class="pre">EdgeConnection</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_assignments</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.PeakID" title="sleap.nn.paf_grouping.PeakID"><span class="pre">PeakID</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">array</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">array</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">array</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L917-L981"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.make_predicted_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Group peaks by assignments and accumulate scores.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>peaks</strong> – Node-grouped peaks</p></li>
 <li><p><strong>peak_scores</strong> – Node-grouped peak scores</p></li>
@@ -1195,7 +1253,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <li><p><strong>instance_assignments</strong> – <a class="reference internal" href="#sleap.nn.paf_grouping.PeakID" title="sleap.nn.paf_grouping.PeakID"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PeakID</span></code></a> to instance ID mapping</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>Tuple of (predicted_instances, predicted_peak_scores, predicted_instance_scores)</p>
 <p>predicted_instances: (n_instances, n_nodes, 2) array
 predicted_peak_scores: (n_instances, n_nodes) array
@@ -1207,10 +1265,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.match_candidates_batch">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">match_candidates_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L673-L796"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.match_candidates_batch" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">match_candidates_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L673-L796"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.match_candidates_batch" title="Permalink to this definition">#</a></dt>
 <dd><p>Match candidate connections for a batch based on PAF scores.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>edge_inds</strong> – Sample-grouped edge indices as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.RaggedTensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">(n_candidates))</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code> indicating the indices
@@ -1226,7 +1284,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <li><p><strong>n_edges</strong> – A scalar <code class="xref py py-obj docutils literal notranslate"><span class="pre">int</span></code> denoting the number of edges in the skeleton.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The connection peaks for each edge matched based on score as tuple of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(match_edge_inds,</span> <span class="pre">match_src_peak_inds,</span> <span class="pre">match_dst_peak_inds,</span> <span class="pre">match_line_scores)</span></code></p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">match_edge_inds</span></code>: Sample-grouped indices of the skeleton edge for each
@@ -1257,10 +1315,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.match_candidates_sample">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">match_candidates_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L553-L670"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.match_candidates_sample" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">match_candidates_sample</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">line_scores_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L553-L670"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.match_candidates_sample" title="Permalink to this definition">#</a></dt>
 <dd><p>Match candidate connections for a sample based on PAF scores.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>edge_inds_sample</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_candidates,)</span></code> and dtype <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code>
 indicating the indices of the edge that each of the candidate connections
@@ -1276,7 +1334,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <li><p><strong>n_edges</strong> – A scalar <code class="xref py py-obj docutils literal notranslate"><span class="pre">int</span></code> denoting the number of edges in the skeleton.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The connection peaks for each edge matched based on score as tuple of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(match_edge_inds,</span> <span class="pre">match_src_peak_inds,</span> <span class="pre">match_dst_peak_inds,</span> <span class="pre">match_line_scores)</span></code></p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">match_edge_inds</span></code>: Indices of the skeleton edge that each connection corresponds
@@ -1305,10 +1363,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.score_paf_lines">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">score_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">paf_lines_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L325-L403"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.score_paf_lines" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">score_paf_lines</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">paf_lines_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_peak_inds_sample</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L325-L403"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.score_paf_lines" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute the connectivity score for each PAF line in a sample.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>paf_lines_sample</strong> – The PAF vectors evaluated at the lines formed between
 candidate conncetions as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
@@ -1330,7 +1388,7 @@ <h1>sleap.nn.paf_grouping</h1>
 penalty more strictly.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>The line scores as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_candidates,)</span></code> and dtype
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>. Each score value is the average dot product between the PAFs and
 the normalized displacement vector between source and destination peaks.</p>
@@ -1350,10 +1408,10 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.score_paf_lines_batch">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">score_paf_lines_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.ops.ragged.ragged_tensor.RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L406-L550"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.score_paf_lines_batch" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">score_paf_lines_batch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peaks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">peak_channel_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">RaggedTensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton_edges</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_line_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pafs_stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_edge_length_ratio</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_penalty_weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_nodes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">RaggedTensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L406-L550"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.score_paf_lines_batch" title="Permalink to this definition">#</a></dt>
 <dd><p>Create and score PAF lines formed between connection candidates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>pafs</strong> – The batch of part affinity fields as a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.Tensor</span></code> of shape
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">n_edges)</span></code> and type <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.float32</span></code>.</p></li>
@@ -1380,7 +1438,7 @@ <h1>sleap.nn.paf_grouping</h1>
 <li><p><strong>n_nodes</strong> – The total number of nodes in the skeleton as a scalar integer.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(edge_inds,</span> <span class="pre">edge_peak_inds,</span> <span class="pre">line_scores)</span></code> with the connections and
 their scores based on the PAFs.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">edge_inds</span></code>: Sample-grouped indices of the edge in the skeleton that each
@@ -1409,13 +1467,13 @@ <h1>sleap.nn.paf_grouping</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.paf_grouping.toposort_edges">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">toposort_edges</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">sleap.nn.paf_grouping.EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/paf_grouping.py#L1293-L1315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.toposort_edges" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.paf_grouping.</span></span><span class="sig-name descname"><span class="pre">toposort_edges</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><span class="pre">EdgeType</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/paf_grouping.py#L1293-L1315"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.paf_grouping.toposort_edges" title="Permalink to this definition">#</a></dt>
 <dd><p>Find a topological ordering for a list of edge types.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>edge_types</strong> – A list of <a class="reference internal" href="#sleap.nn.paf_grouping.EdgeType" title="sleap.nn.paf_grouping.EdgeType"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EdgeType</span></code></a> instances describing a skeleton.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of indices specifying the topological order that the edge types should
 be accessed in during instance assembly (<a class="reference internal" href="#sleap.nn.paf_grouping.assign_connections_to_instances" title="sleap.nn.paf_grouping.assign_connections_to_instances"><code class="xref py py-obj docutils literal notranslate"><span class="pre">assign_connections_to_instances</span></code></a>).</p>
 <p>This is important to ensure that instances are assembled starting at the root
@@ -1450,6 +1508,69 @@ <h1>sleap.nn.paf_grouping</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection"><code class="docutils literal notranslate"><span class="pre">EdgeConnection</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.src_peak_ind"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.src_peak_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.dst_peak_ind"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.dst_peak_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeConnection.score"><code class="docutils literal notranslate"><span class="pre">EdgeConnection.score</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType"><code class="docutils literal notranslate"><span class="pre">EdgeType</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType.src_node_ind"><code class="docutils literal notranslate"><span class="pre">EdgeType.src_node_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.EdgeType.dst_node_ind"><code class="docutils literal notranslate"><span class="pre">EdgeType.dst_node_ind</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer"><code class="docutils literal notranslate"><span class="pre">PAFScorer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.part_names"><code class="docutils literal notranslate"><span class="pre">PAFScorer.part_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edges"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.pafs_stride"><code class="docutils literal notranslate"><span class="pre">PAFScorer.pafs_stride</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.max_edge_length_ratio"><code class="docutils literal notranslate"><span class="pre">PAFScorer.max_edge_length_ratio</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.dist_penalty_weight"><code class="docutils literal notranslate"><span class="pre">PAFScorer.dist_penalty_weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_points"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.min_instance_peaks"><code class="docutils literal notranslate"><span class="pre">PAFScorer.min_instance_peaks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.min_line_scores"><code class="docutils literal notranslate"><span class="pre">PAFScorer.min_line_scores</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edge_inds"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.edge_types"><code class="docutils literal notranslate"><span class="pre">PAFScorer.edge_types</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_nodes"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.n_edges"><code class="docutils literal notranslate"><span class="pre">PAFScorer.n_edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.sorted_edge_inds"><code class="docutils literal notranslate"><span class="pre">PAFScorer.sorted_edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.from_config"><code class="docutils literal notranslate"><span class="pre">PAFScorer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.group_instances"><code class="docutils literal notranslate"><span class="pre">PAFScorer.group_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.match_candidates"><code class="docutils literal notranslate"><span class="pre">PAFScorer.match_candidates()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.predict"><code class="docutils literal notranslate"><span class="pre">PAFScorer.predict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PAFScorer.score_paf_lines"><code class="docutils literal notranslate"><span class="pre">PAFScorer.score_paf_lines()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID"><code class="docutils literal notranslate"><span class="pre">PeakID</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID.node_ind"><code class="docutils literal notranslate"><span class="pre">PeakID.node_ind</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.PeakID.peak_ind"><code class="docutils literal notranslate"><span class="pre">PeakID.peak_ind</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.assign_connections_to_instances"><code class="docutils literal notranslate"><span class="pre">assign_connections_to_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.compute_distance_penalty"><code class="docutils literal notranslate"><span class="pre">compute_distance_penalty()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.get_connection_candidates"><code class="docutils literal notranslate"><span class="pre">get_connection_candidates()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.get_paf_lines"><code class="docutils literal notranslate"><span class="pre">get_paf_lines()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.group_instances_batch"><code class="docutils literal notranslate"><span class="pre">group_instances_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.group_instances_sample"><code class="docutils literal notranslate"><span class="pre">group_instances_sample()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.make_line_subs"><code class="docutils literal notranslate"><span class="pre">make_line_subs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.make_predicted_instances"><code class="docutils literal notranslate"><span class="pre">make_predicted_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.match_candidates_batch"><code class="docutils literal notranslate"><span class="pre">match_candidates_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.match_candidates_sample"><code class="docutils literal notranslate"><span class="pre">match_candidates_sample()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.score_paf_lines"><code class="docutils literal notranslate"><span class="pre">score_paf_lines()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.score_paf_lines_batch"><code class="docutils literal notranslate"><span class="pre">score_paf_lines_batch()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.paf_grouping.toposort_edges"><code class="docutils literal notranslate"><span class="pre">toposort_edges()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1468,7 +1589,7 @@ <h1>sleap.nn.paf_grouping</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.peak_finding.html b/api/sleap.nn.peak_finding.html
index 97256f59b..6d9450c71 100644
--- a/api/sleap.nn.peak_finding.html
+++ b/api/sleap.nn.peak_finding.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.peak_finding &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.peak_finding &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.peak_finding';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.peak_finding.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,26 @@ <h1>sleap.nn.peak_finding</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.crop_bboxes"><code class="docutils literal notranslate"><span class="pre">crop_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks"><code class="docutils literal notranslate"><span class="pre">find_global_peaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_integral"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_integral()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_rough"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_rough()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_with_offsets"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_with_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks"><code class="docutils literal notranslate"><span class="pre">find_local_peaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_integral"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_integral()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_rough"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_rough()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_with_offsets"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_with_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_offsets_local_direction"><code class="docutils literal notranslate"><span class="pre">find_offsets_local_direction()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.integral_regression"><code class="docutils literal notranslate"><span class="pre">integral_regression()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.make_gaussian_kernel"><code class="docutils literal notranslate"><span class="pre">make_gaussian_kernel()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.smooth_imgs"><code class="docutils literal notranslate"><span class="pre">smooth_imgs()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +338,7 @@ <h1>sleap.nn.peak_finding</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.peak_finding">
-<span id="sleap-nn-peak-finding"></span><h1>sleap.nn.peak_finding<a class="headerlink" href="#module-sleap.nn.peak_finding" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-peak-finding"></span><h1>sleap.nn.peak_finding<a class="headerlink" href="#module-sleap.nn.peak_finding" title="Permalink to this heading">#</a></h1>
 <p>This module contains TensorFlow-based peak finding methods.</p>
 <p>In general, the inputs to the functions provided here operate on confidence maps
 (sometimes referred to as heatmaps), which are image-based representations of the
@@ -329,12 +349,12 @@ <h1>sleap.nn.peak_finding</h1>
 <p>Peak finding entails finding either the global or local maxima of these confidence maps.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.crop_bboxes">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">crop_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L135-L190"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.crop_bboxes" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">crop_bboxes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">images</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bboxes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sample_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L135-L190"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.crop_bboxes" title="Permalink to this definition">#</a></dt>
 <dd><p>Crop bounding boxes from a batch of images.</p>
 <p>This method serves as a convenience method for specifying the arguments of
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.image.crop_and_resize</span></code>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>images</strong> – Tensor of shape (samples, height, width, channels) of a batch of images.</p></li>
 <li><p><strong>bboxes</strong> – Tensor of shape (n_bboxes, 4) and dtype tf.float32, where the last axis
@@ -344,7 +364,7 @@ <h1>sleap.nn.peak_finding</h1>
 box should be cropped from.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of shape (n_bboxes, crop_height, crop_width, channels) of the same
 dtype as the input image. The crop size is inferred from the bounding box
 coordinates.</p>
@@ -364,10 +384,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_global_peaks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Find global peaks with optional refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps. Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>threshold</strong> – Minimum confidence threshold. Peaks with values below this will
@@ -379,7 +399,7 @@ <h1>sleap.nn.peak_finding</h1>
 integer scalar.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals).</p>
 <p>peak_points: float32 tensor of shape (samples, channels, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -392,12 +412,12 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_global_peaks_integral">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_integral</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L423-L448"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_integral" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_integral</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L423-L448"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_integral" title="Permalink to this definition">#</a></dt>
 <dd><p>Find local peaks with integral refinement.</p>
 <p>Integral regression refinement will be computed by taking the weighted average of
 the local neighborhood around each rough peak.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps. Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>crop_size</strong> – Size of patches to crop around each rough peak as an integer scalar.</p></li>
@@ -405,7 +425,7 @@ <h1>sleap.nn.peak_finding</h1>
 set to NaNs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals).</p>
 <p>peak_points: float32 tensor of shape (samples, channels, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -418,17 +438,17 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_global_peaks_rough">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_rough</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L193-L246"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_rough" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_rough</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.1</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L193-L246"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_rough" title="Permalink to this definition">#</a></dt>
 <dd><p>Find the global maximum for each sample and channel.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>threshold</strong> – Scalar float specifying the minimum confidence value for peaks. Peaks
 with values below this threshold will be replaced with NaNs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals).</p>
 <p>peak_points: float32 tensor of shape (samples, channels, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -441,10 +461,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_global_peaks_with_offsets">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_with_offsets" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_global_peaks_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_global_peaks_with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>Find global peaks and refine with learned offset maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>.</p></li>
 <li><p><strong>offsets</strong> – Offset maps tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">channels)</span></code>.</p></li>
@@ -452,7 +472,7 @@ <h1>sleap.nn.peak_finding</h1>
 with values below this threshold will be replaced with NaNs.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(peak_points,</span> <span class="pre">peak_vals)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">peak_points</span></code>: <code class="xref py py-obj docutils literal notranslate"><span class="pre">float32</span></code> tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">channels,</span> <span class="pre">2)</span></code>, where the
 last axis indicates peak locations in xy order.</p>
@@ -465,10 +485,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_local_peaks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">refinement</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">integral_patch_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Find local peaks with optional refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps. Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>threshold</strong> – Minimum confidence threshold. Peaks with values below this will
@@ -480,7 +500,7 @@ <h1>sleap.nn.peak_finding</h1>
 integer scalar.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals, peak_sample_inds, peak_channel_inds).</p>
 <p>peak_points: float32 tensor of shape (n_peaks, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -497,10 +517,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_local_peaks_integral">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_integral</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L535-L563"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_integral" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_integral</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">crop_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L535-L563"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_integral" title="Permalink to this definition">#</a></dt>
 <dd><p>Find local peaks with integral refinement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps. Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>crop_size</strong> – </p></li>
@@ -508,7 +528,7 @@ <h1>sleap.nn.peak_finding</h1>
 ignored.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals, peak_sample_inds, peak_channel_inds).</p>
 <p>peak_points: float32 tensor of shape (n_peaks, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -525,17 +545,17 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_local_peaks_rough">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_rough</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L249-L308"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_rough" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_rough</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L249-L308"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_rough" title="Permalink to this definition">#</a></dt>
 <dd><p>Find local maxima via non-maximum suppresion.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>threshold</strong> – Scalar float specifying the minimum confidence value for peaks. Peaks
 with values below this threshold will not be returned.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (peak_points, peak_vals, peak_sample_inds, peak_channel_inds).
 peak_points: float32 tensor of shape (n_peaks, 2), where the last axis
 indicates peak locations in xy order.</p>
@@ -552,10 +572,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_local_peaks_with_offsets">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_with_offsets" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_local_peaks_with_offsets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offsets</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.2</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#sleap.nn.peak_finding.find_local_peaks_with_offsets" title="Permalink to this definition">#</a></dt>
 <dd><p>Find local peaks and refine with learned offset maps.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps. Tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">channels)</span></code>.</p></li>
 <li><p><strong>offsets</strong> – Offset maps tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(samples,</span> <span class="pre">height,</span> <span class="pre">width,</span> <span class="pre">2</span> <span class="pre">*</span> <span class="pre">channels)</span></code>.</p></li>
@@ -563,7 +583,7 @@ <h1>sleap.nn.peak_finding</h1>
 ignored.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(peak_points,</span> <span class="pre">peak_vals,</span> <span class="pre">peak_sample_inds,</span> <span class="pre">peak_channel_inds)</span></code>.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">peak_points</span></code>: float32 tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_peaks,</span> <span class="pre">2)</span></code>, where the last axis
 indicates peak locations in xy order.</p>
@@ -580,12 +600,12 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.find_offsets_local_direction">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_offsets_local_direction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centered_patches</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L78-L132"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_offsets_local_direction" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">find_offsets_local_direction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">centered_patches</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">delta</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.25</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L78-L132"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.find_offsets_local_direction" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes subpixel offsets from the direction of the pixels around the peak.</p>
 <p>This function finds the delta-offset from the center pixel of peak-centered patches
 by finding the direction of the gradient around each center.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>centered_patches</strong> – A rank-4 tensor of shape (samples, 3, 3, 1) corresponding
 to the centered crops around the grid-anchored peaks. For multi-channel
@@ -594,7 +614,7 @@ <h1>sleap.nn.peak_finding</h1>
 <li><p><strong>delta</strong> – Scalar float that will scaled by the gradient direction.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>offsets, a float32 tensor of shape (samples, 2) where the columns correspond to
 the offsets relative to the center pixel for the x and y directions
 respectively, i.e., for the i-th sample:</p>
@@ -628,27 +648,28 @@ <h1>sleap.nn.peak_finding</h1>
 </pre></div>
 </div>
 <p class="rubric">References</p>
-<dl class="footnote brackets">
-<dt class="label" id="id1"><span class="brackets">1</span></dt>
-<dd><p>Alejandro Newell, Kaiyu Yang, and Jia Deng. Stacked Hourglass Networks
+<aside class="footnote-list brackets">
+<aside class="footnote brackets" id="id1" role="note">
+<span class="label"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></span>
+<p>Alejandro Newell, Kaiyu Yang, and Jia Deng. Stacked Hourglass Networks
 for Human Pose Estimation. In _European conference on computer <a href="#id2"><span class="problematic" id="id3">vision_</span></a>, 2016.</p>
-</dd>
-</dl>
+</aside>
+</aside>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.integral_regression">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">integral_regression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L311-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.integral_regression" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">integral_regression</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cms</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">yv</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L311-L334"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.integral_regression" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute regression by integrating over the confidence maps on a grid.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cms</strong> – Confidence maps with shape (samples, height, width, channels).</p></li>
 <li><p><strong>xv</strong> – X grid vector tf.float32 of grid coordinates to sample.</p></li>
 <li><p><strong>yv</strong> – Y grid vector tf.float32 of grid coordinates to sample.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of (x_hat, y_hat) with the regressed x- and y-coordinates for each
 channel of the confidence maps.</p>
 <p>x_hat and y_hat are of shape (samples, channels)</p>
@@ -659,16 +680,16 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.make_gaussian_kernel">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">make_gaussian_kernel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L20-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.make_gaussian_kernel" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">make_gaussian_kernel</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L20-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.make_gaussian_kernel" title="Permalink to this definition">#</a></dt>
 <dd><p>Generates a square unnormalized 2D symmetric Gaussian kernel.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>size</strong> – Length of kernel. This should be an odd integer.</p></li>
 <li><p><strong>sigma</strong> – Standard deviation of the Gaussian specified as a scalar float.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>kernel, a float32 tensor of shape (size, size) with values corresponding to the
 unnormalized probability density of a 2D Gaussian distribution with symmetric
 covariance along the x and y directions.</p>
@@ -683,10 +704,10 @@ <h1>sleap.nn.peak_finding</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.peak_finding.smooth_imgs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">smooth_imgs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">imgs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/peak_finding.py#L49-L75"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.smooth_imgs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.peak_finding.</span></span><span class="sig-name descname"><span class="pre">smooth_imgs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">imgs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/peak_finding.py#L49-L75"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.peak_finding.smooth_imgs" title="Permalink to this definition">#</a></dt>
 <dd><p>Smooths the input image by convolving it with a Gaussian kernel.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>imgs</strong> – A rank-4 tensor of shape (samples, height, width, channels).</p></li>
 <li><p><strong>kernel_size</strong> – An odd-valued scalar integer specifying the width and height of
@@ -694,7 +715,7 @@ <h1>sleap.nn.peak_finding</h1>
 <li><p><strong>sigma</strong> – Standard deviation of the Gaussian specified as a scalar float.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tensor of the same shape as imgs after convolving with a Gaussian sample and
 channelwise.</p>
 </dd>
@@ -725,6 +746,32 @@ <h1>sleap.nn.peak_finding</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.crop_bboxes"><code class="docutils literal notranslate"><span class="pre">crop_bboxes()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks"><code class="docutils literal notranslate"><span class="pre">find_global_peaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_integral"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_integral()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_rough"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_rough()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_global_peaks_with_offsets"><code class="docutils literal notranslate"><span class="pre">find_global_peaks_with_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks"><code class="docutils literal notranslate"><span class="pre">find_local_peaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_integral"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_integral()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_rough"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_rough()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_local_peaks_with_offsets"><code class="docutils literal notranslate"><span class="pre">find_local_peaks_with_offsets()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.find_offsets_local_direction"><code class="docutils literal notranslate"><span class="pre">find_offsets_local_direction()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.integral_regression"><code class="docutils literal notranslate"><span class="pre">integral_regression()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.make_gaussian_kernel"><code class="docutils literal notranslate"><span class="pre">make_gaussian_kernel()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.peak_finding.smooth_imgs"><code class="docutils literal notranslate"><span class="pre">smooth_imgs()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -743,7 +790,7 @@ <h1>sleap.nn.peak_finding</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.system.html b/api/sleap.nn.system.html
index 83d0c6c55..03de6c159 100644
--- a/api/sleap.nn.system.html
+++ b/api/sleap.nn.system.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.system &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.system &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.system';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.system.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,28 @@ <h1>sleap.nn.system</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.best_logical_device_name"><code class="docutils literal notranslate"><span class="pre">best_logical_device_name()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.disable_preallocation"><code class="docutils literal notranslate"><span class="pre">disable_preallocation()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.enable_preallocation"><code class="docutils literal notranslate"><span class="pre">enable_preallocation()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_all_gpus"><code class="docutils literal notranslate"><span class="pre">get_all_gpus()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_available_gpus"><code class="docutils literal notranslate"><span class="pre">get_available_gpus()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_current_gpu"><code class="docutils literal notranslate"><span class="pre">get_current_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_gpu_memory"><code class="docutils literal notranslate"><span class="pre">get_gpu_memory()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.initialize_devices"><code class="docutils literal notranslate"><span class="pre">initialize_devices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.is_gpu_system"><code class="docutils literal notranslate"><span class="pre">is_gpu_system()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.is_initialized"><code class="docutils literal notranslate"><span class="pre">is_initialized()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.summary"><code class="docutils literal notranslate"><span class="pre">summary()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_cpu_only"><code class="docutils literal notranslate"><span class="pre">use_cpu_only()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_first_gpu"><code class="docutils literal notranslate"><span class="pre">use_first_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_gpu"><code class="docutils literal notranslate"><span class="pre">use_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_last_gpu"><code class="docutils literal notranslate"><span class="pre">use_last_gpu()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,17 +340,17 @@ <h1>sleap.nn.system</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.system">
-<span id="sleap-nn-system"></span><h1>sleap.nn.system<a class="headerlink" href="#module-sleap.nn.system" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-system"></span><h1>sleap.nn.system<a class="headerlink" href="#module-sleap.nn.system" title="Permalink to this heading">#</a></h1>
 <p>Utilities for working with the physical system (e.g., GPUs).</p>
 <p>This module mostly provides convenience functions for changing the state of the runtime
 environment by wrapping <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.config</span></code> module functions.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.best_logical_device_name">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">best_logical_device_name</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L171-L188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.best_logical_device_name" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">best_logical_device_name</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L191-L208"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.best_logical_device_name" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the name of the best logical device for performance.</p>
 <p>This is particularly useful to use with <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.device()</span></code> for explicit tensor placement.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The name of the first logical GPU device if available, or alternatively the CPU.</p>
 </dd>
 </dl>
@@ -340,7 +362,7 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.disable_preallocation">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">disable_preallocation</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L115-L127"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.disable_preallocation" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">disable_preallocation</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L135-L147"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.disable_preallocation" title="Permalink to this definition">#</a></dt>
 <dd><p>Disable preallocation of full GPU memory on all available GPUs.</p>
 <p>This enables memory growth policy so that TensorFlow will not pre-allocate all
 available GPU memory.</p>
@@ -351,7 +373,7 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.enable_preallocation">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">enable_preallocation</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L130-L142"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.enable_preallocation" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">enable_preallocation</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L150-L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.enable_preallocation" title="Permalink to this definition">#</a></dt>
 <dd><p>Enable preallocation of full GPU memory on all available GPUs.</p>
 <p>This disables memory growth policy so that TensorFlow will pre-allocate all
 available GPU memory.</p>
@@ -362,27 +384,27 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.get_all_gpus">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_all_gpus</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.eager.context.PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L14-L16"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_all_gpus" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_all_gpus</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L14-L16"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_all_gpus" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of GPUs including unavailable devices.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.get_available_gpus">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_available_gpus</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.eager.context.PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L24-L26"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_available_gpus" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_available_gpus</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L24-L26"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_available_gpus" title="Permalink to this definition">#</a></dt>
 <dd><p>Return a list of available GPUs.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.get_current_gpu">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_current_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.eager.context.PhysicalDevice</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L29-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_current_gpu" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_current_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">PhysicalDevice</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L29-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_current_gpu" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the current (single) GPU device.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p><p>The tf.config.PhysicalDevice for the available GPU.</p>
 <p>If no GPUs are available, returns None.</p>
 </p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If multiple GPUs are available.</p>
 </dd>
 </dl>
@@ -390,10 +412,10 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.get_gpu_memory">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_gpu_memory</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L191-L229"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_gpu_memory" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">get_gpu_memory</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L211-L249"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.get_gpu_memory" title="Permalink to this definition">#</a></dt>
 <dd><p>Get the available memory on each GPU.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list of the available memory on each GPU in MiB.</p>
 </dd>
 </dl>
@@ -401,26 +423,26 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.initialize_devices">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">initialize_devices</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L145-L150"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.initialize_devices" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">initialize_devices</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L165-L170"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.initialize_devices" title="Permalink to this definition">#</a></dt>
 <dd><p>Initialize available physical devices as logical devices.</p>
 <p>If preallocation was enabled on the GPUs, this will trigger memory allocation.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.is_gpu_system">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">is_gpu_system</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L19-L21"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.is_gpu_system" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">is_gpu_system</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L19-L21"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.is_gpu_system" title="Permalink to this definition">#</a></dt>
 <dd><p>Return True if the system has discoverable GPUs.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.is_initialized">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">is_initialized</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">gpu</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.eager.context.PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L74-L112"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.is_initialized" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">is_initialized</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">gpu</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">PhysicalDevice</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L94-L132"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.is_initialized" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if a physical GPU has been initialized without triggering initialization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>gpu</strong> – The GPU to check for initialization. If None, defaults to the current GPU.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if the GPU is initialized.</p>
 </dd>
 </dl>
@@ -437,28 +459,28 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.summary">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">summary</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L153-L168"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.summary" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">summary</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L173-L188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.summary" title="Permalink to this definition">#</a></dt>
 <dd><p>Print a summary of the state of the system.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.use_cpu_only">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_cpu_only</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L49-L51"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_cpu_only" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_cpu_only</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L49-L61"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_cpu_only" title="Permalink to this definition">#</a></dt>
 <dd><p>Hide GPUs from TensorFlow to ensure only the CPU is available.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.use_first_gpu">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_first_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L64-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_first_gpu" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_first_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L84-L86"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_first_gpu" title="Permalink to this definition">#</a></dt>
 <dd><p>Make only the first GPU available to TensorFlow.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.use_gpu">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_gpu</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L54-L61"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_gpu" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_gpu</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device_ind</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L64-L81"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_gpu" title="Permalink to this definition">#</a></dt>
 <dd><p>Make a single GPU available to TensorFlow.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>device_ind</strong> – Index of the GPU within the list of system GPUs.</p>
 </dd>
 </dl>
@@ -466,7 +488,7 @@ <h1>sleap.nn.system</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.system.use_last_gpu">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_last_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/system.py#L69-L71"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_last_gpu" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.system.</span></span><span class="sig-name descname"><span class="pre">use_last_gpu</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/system.py#L89-L91"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.system.use_last_gpu" title="Permalink to this definition">#</a></dt>
 <dd><p>Make only the last GPU available to TensorFlow.</p>
 </dd></dl>
 
@@ -494,6 +516,34 @@ <h1>sleap.nn.system</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.best_logical_device_name"><code class="docutils literal notranslate"><span class="pre">best_logical_device_name()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.disable_preallocation"><code class="docutils literal notranslate"><span class="pre">disable_preallocation()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.enable_preallocation"><code class="docutils literal notranslate"><span class="pre">enable_preallocation()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_all_gpus"><code class="docutils literal notranslate"><span class="pre">get_all_gpus()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_available_gpus"><code class="docutils literal notranslate"><span class="pre">get_available_gpus()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_current_gpu"><code class="docutils literal notranslate"><span class="pre">get_current_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.get_gpu_memory"><code class="docutils literal notranslate"><span class="pre">get_gpu_memory()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.initialize_devices"><code class="docutils literal notranslate"><span class="pre">initialize_devices()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.is_gpu_system"><code class="docutils literal notranslate"><span class="pre">is_gpu_system()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.is_initialized"><code class="docutils literal notranslate"><span class="pre">is_initialized()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.summary"><code class="docutils literal notranslate"><span class="pre">summary()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_cpu_only"><code class="docutils literal notranslate"><span class="pre">use_cpu_only()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_first_gpu"><code class="docutils literal notranslate"><span class="pre">use_first_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_gpu"><code class="docutils literal notranslate"><span class="pre">use_gpu()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.system.use_last_gpu"><code class="docutils literal notranslate"><span class="pre">use_last_gpu()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -512,7 +562,7 @@ <h1>sleap.nn.system</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.tracker.components.html b/api/sleap.nn.tracker.components.html
index fabb0ce56..616e241da 100644
--- a/api/sleap.nn.tracker.components.html
+++ b/api/sleap.nn.tracker.components.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.tracker.components &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.tracker.components &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.tracker.components';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.tracker.components.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,34 @@ <h1>sleap.nn.tracker.components</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches"><code class="docutils literal notranslate"><span class="pre">FrameMatches</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.matches"><code class="docutils literal notranslate"><span class="pre">FrameMatches.matches</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.cost_matrix"><code class="docutils literal notranslate"><span class="pre">FrameMatches.cost_matrix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.unmatched_instances"><code class="docutils literal notranslate"><span class="pre">FrameMatches.unmatched_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.from_candidate_instances"><code class="docutils literal notranslate"><span class="pre">FrameMatches.from_candidate_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.has_only_first_choice_matches"><code class="docutils literal notranslate"><span class="pre">FrameMatches.has_only_first_choice_matches</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.Match"><code class="docutils literal notranslate"><span class="pre">Match</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.centroid_distance"><code class="docutils literal notranslate"><span class="pre">centroid_distance()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.connect_single_track_breaks"><code class="docutils literal notranslate"><span class="pre">connect_single_track_breaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.cull_frame_instances"><code class="docutils literal notranslate"><span class="pre">cull_frame_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.cull_instances"><code class="docutils literal notranslate"><span class="pre">cull_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.factory_object_keypoint_similarity"><code class="docutils literal notranslate"><span class="pre">factory_object_keypoint_similarity()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.first_choice_matching"><code class="docutils literal notranslate"><span class="pre">first_choice_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.greedy_matching"><code class="docutils literal notranslate"><span class="pre">greedy_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.hungarian_matching"><code class="docutils literal notranslate"><span class="pre">hungarian_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.instance_iou"><code class="docutils literal notranslate"><span class="pre">instance_iou()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.instance_similarity"><code class="docutils literal notranslate"><span class="pre">instance_similarity()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.nms_fast"><code class="docutils literal notranslate"><span class="pre">nms_fast()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.normalized_instance_similarity"><code class="docutils literal notranslate"><span class="pre">normalized_instance_similarity()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +346,7 @@ <h1>sleap.nn.tracker.components</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.tracker.components">
-<span id="sleap-nn-tracker-components"></span><h1>sleap.nn.tracker.components<a class="headerlink" href="#module-sleap.nn.tracker.components" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-tracker-components"></span><h1>sleap.nn.tracker.components<a class="headerlink" href="#module-sleap.nn.tracker.components" title="Permalink to this heading">#</a></h1>
 <p>Functions/classes used by multiple trackers.</p>
 <p>Main types of functions:</p>
 <ol class="arabic simple">
@@ -329,7 +357,7 @@ <h1>sleap.nn.tracker.components</h1>
 </ol>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.FrameMatches">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">FrameMatches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">matches</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match"><span class="pre">sleap.nn.tracker.components.Match</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unmatched_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L371-L532"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">FrameMatches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">matches</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match"><span class="pre">Match</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">unmatched_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L479-L640"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches" title="Permalink to this definition">#</a></dt>
 <dd><p>Calculates (and stores) matches for a frame.</p>
 <p>This class encapsulates the logic to generate matches (using a custom
 matching function) from a cost matrix. One key feature is that it retains
@@ -344,7 +372,7 @@ <h1>sleap.nn.tracker.components</h1>
 <span class="sig-name descname"><span class="pre">matches</span></span><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>the list of <a class="reference internal" href="#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Match</span></code></a> objects.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[<a class="reference internal" href="#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match">sleap.nn.tracker.components.Match</a>]</p>
 </dd>
 </dl>
@@ -356,7 +384,7 @@ <h1>sleap.nn.tracker.components</h1>
 <dd><p>the cost matrix, shape is
 (number of untracked instances, number of candidate tracks).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>numpy.ndarray</p>
 </dd>
 </dl>
@@ -367,7 +395,7 @@ <h1>sleap.nn.tracker.components</h1>
 <span class="sig-name descname"><span class="pre">unmatched_instances</span></span><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches.unmatched_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>the instances for which we are finding matches.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[sleap.nn.tracker.components.InstanceType]</p>
 </dd>
 </dl>
@@ -375,10 +403,10 @@ <h1>sleap.nn.tracker.components</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.FrameMatches.from_candidate_instances">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_candidate_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L406-L481"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches.from_candidate_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_candidate_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L514-L589"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.FrameMatches.from_candidate_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Calculates (and stores) matches for a frame from candidate instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>untracked_instances</strong> – list of untracked instances in the frame.</p></li>
 <li><p><strong>candidate_instances</strong> – list of instances use as match.</p></li>
@@ -407,13 +435,13 @@ <h1>sleap.nn.tracker.components</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.Match">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">Match</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_first_choice</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L361-L367"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.Match" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">Match</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">score</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_first_choice</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L469-L475"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.Match" title="Permalink to this definition">#</a></dt>
 <dd><p>Stores a match between a specific instance and specific track.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.centroid_distance">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">centroid_distance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">{}</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L43-L61"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.centroid_distance" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">centroid_distance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">{}</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L151-L169"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.centroid_distance" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the negative distance between the centroids of two instances.</p>
 <p>Uses <code class="xref py py-obj docutils literal notranslate"><span class="pre">cache</span></code> dictionary (created with function so it persists between calls)
 since without cache this method is significantly slower than others.</p>
@@ -421,16 +449,16 @@ <h1>sleap.nn.tracker.components</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.connect_single_track_breaks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">connect_single_track_breaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L309-L357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.connect_single_track_breaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">connect_single_track_breaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L417-L465"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.connect_single_track_breaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Merges breaks in tracks by connecting single lost with single new track.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>frames</strong> – The list of <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects with predictions.</p></li>
 <li><p><strong>instance_count</strong> – The maximum number of instances we want per frame.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Updated list of frames, also modifies frames in place.</p>
 </dd>
 </dl>
@@ -438,10 +466,10 @@ <h1>sleap.nn.tracker.components</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.cull_frame_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">cull_frame_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L258-L306"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.cull_frame_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">cull_frame_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L366-L414"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.cull_frame_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Removes instances (for single frame) over instance per frame threshold.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instances_list</strong> – The list of instances for a single frame.</p></li>
 <li><p><strong>instance_count</strong> – The maximum number of instances we want per frame.</p></li>
@@ -450,7 +478,7 @@ <h1>sleap.nn.tracker.components</h1>
 only use score to determine which instances to remove.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Updated list of frames, also modifies frames in place.</p>
 </dd>
 </dl>
@@ -458,10 +486,10 @@ <h1>sleap.nn.tracker.components</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.cull_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">cull_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L202-L255"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.cull_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">cull_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L310-L363"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.cull_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Removes instances from frames over instance per frame threshold.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>frames</strong> – The list of <code class="xref py py-obj docutils literal notranslate"><span class="pre">LabeledFrame</span></code> objects with predictions.</p></li>
 <li><p><strong>instance_count</strong> – The maximum number of instances we want per frame.</p></li>
@@ -470,49 +498,84 @@ <h1>sleap.nn.tracker.components</h1>
 only use score to determine which instances to remove.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None; modifies frames in place.</p>
 </dd>
 </dl>
 </dd></dl>
 
+<dl class="py function">
+<dt class="sig sig-object py" id="sleap.nn.tracker.components.factory_object_keypoint_similarity">
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">factory_object_keypoint_similarity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keypoint_errors</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">score_weighting</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">normalization_keypoints</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">'all'</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L62-L148"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.factory_object_keypoint_similarity" title="Permalink to this definition">#</a></dt>
+<dd><p>Factory for similarity function based on object keypoints.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>keypoint_errors</strong> – The standard error of the distance between the predicted
+keypoint and the true value, in pixels.
+If None or empty list, defaults to 1.
+If a scalar or singleton list, every keypoint has the same error.
+If a list, defines the error for each keypoint, the length should be equal
+to the number of keypoints in the skeleton.</p></li>
+<li><p><strong>score_weighting</strong> – If True, use <code class="xref py py-obj docutils literal notranslate"><span class="pre">score</span></code> of <code class="xref py py-obj docutils literal notranslate"><span class="pre">PredictedPoint</span></code> to weigh
+<code class="xref py py-obj docutils literal notranslate"><span class="pre">keypoint_errors</span></code>. If False, do not add a weight to <code class="xref py py-obj docutils literal notranslate"><span class="pre">keypoint_errors</span></code>.</p></li>
+<li><p><strong>normalization_keypoints</strong> – Determine how to normalize similarity score. One of
+[“all”, “ref”, “union”]. If “all”, similarity score is normalized by number
+of reference points. If “ref”, similarity score is normalized by number of
+visible reference points. If “union”, similarity score is normalized by
+number of points both visible in query and reference instance.
+Default is “all”.</p></li>
+</ul>
+</dd>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>Callable that returns object keypoint similarity between two <a href="#id1"><span class="problematic" id="id2">`</span></a>Instance`s.</p>
+</dd>
+</dl>
+</dd></dl>
+
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.first_choice_matching">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">first_choice_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L535-L545"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.first_choice_matching" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">first_choice_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L643-L653"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.first_choice_matching" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns match indices where each row gets matched to best column.</p>
 <p>The means that multiple rows might be matched to the same column.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.greedy_matching">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">greedy_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L88-L109"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.greedy_matching" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">greedy_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L196-L217"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.greedy_matching" title="Permalink to this definition">#</a></dt>
 <dd><p>Performs greedy bipartite matching.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.hungarian_matching">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">hungarian_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L81-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.hungarian_matching" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">hungarian_matching</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L189-L193"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.hungarian_matching" title="Permalink to this definition">#</a></dt>
 <dd><p>Wrapper for Hungarian matching algorithm in scipy.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.instance_iou">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">instance_iou</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">{}</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L64-L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.instance_iou" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">instance_iou</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cache</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">{}</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L172-L186"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.instance_iou" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes IOU between bounding boxes of instances.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.instance_similarity">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">instance_similarity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L29-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.instance_similarity" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">instance_similarity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L48-L59"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.instance_similarity" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes similarity between instances.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.components.nms_fast">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">nms_fast</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">boxes</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scores</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_count</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L125-L199"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.nms_fast" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">nms_fast</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">boxes</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scores</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_count</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L233-L307"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.nms_fast" title="Permalink to this definition">#</a></dt>
 <dd><p><a class="reference external" href="https://www.pyimagesearch.com/2015/02/16/faster-non-maximum-suppression-python/">https://www.pyimagesearch.com/2015/02/16/faster-non-maximum-suppression-python/</a></p>
 </dd></dl>
 
+<dl class="py function">
+<dt class="sig sig-object py" id="sleap.nn.tracker.components.normalized_instance_similarity">
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.components.</span></span><span class="sig-name descname"><span class="pre">normalized_instance_similarity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_hw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L33-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.components.normalized_instance_similarity" title="Permalink to this definition">#</a></dt>
+<dd><p>Computes similarity between instances with normalized keypoints.</p>
+</dd></dl>
+
 </section>
 
 
@@ -537,6 +600,40 @@ <h1>sleap.nn.tracker.components</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches"><code class="docutils literal notranslate"><span class="pre">FrameMatches</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.matches"><code class="docutils literal notranslate"><span class="pre">FrameMatches.matches</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.cost_matrix"><code class="docutils literal notranslate"><span class="pre">FrameMatches.cost_matrix</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.unmatched_instances"><code class="docutils literal notranslate"><span class="pre">FrameMatches.unmatched_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.from_candidate_instances"><code class="docutils literal notranslate"><span class="pre">FrameMatches.from_candidate_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.FrameMatches.has_only_first_choice_matches"><code class="docutils literal notranslate"><span class="pre">FrameMatches.has_only_first_choice_matches</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.Match"><code class="docutils literal notranslate"><span class="pre">Match</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.centroid_distance"><code class="docutils literal notranslate"><span class="pre">centroid_distance()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.connect_single_track_breaks"><code class="docutils literal notranslate"><span class="pre">connect_single_track_breaks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.cull_frame_instances"><code class="docutils literal notranslate"><span class="pre">cull_frame_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.cull_instances"><code class="docutils literal notranslate"><span class="pre">cull_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.factory_object_keypoint_similarity"><code class="docutils literal notranslate"><span class="pre">factory_object_keypoint_similarity()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.first_choice_matching"><code class="docutils literal notranslate"><span class="pre">first_choice_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.greedy_matching"><code class="docutils literal notranslate"><span class="pre">greedy_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.hungarian_matching"><code class="docutils literal notranslate"><span class="pre">hungarian_matching()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.instance_iou"><code class="docutils literal notranslate"><span class="pre">instance_iou()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.instance_similarity"><code class="docutils literal notranslate"><span class="pre">instance_similarity()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.nms_fast"><code class="docutils literal notranslate"><span class="pre">nms_fast()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.components.normalized_instance_similarity"><code class="docutils literal notranslate"><span class="pre">normalized_instance_similarity()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -555,7 +652,7 @@ <h1>sleap.nn.tracker.components</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.tracker.kalman.html b/api/sleap.nn.tracker.kalman.html
index b5bb035e6..2a873f95e 100644
--- a/api/sleap.nn.tracker.kalman.html
+++ b/api/sleap.nn.tracker.kalman.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.tracker.kalman &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.tracker.kalman &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.tracker.kalman';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.tracker.kalman.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,16 @@ <h1>sleap.nn.tracker.kalman</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.get_track_instance_matches"><code class="docutils literal notranslate"><span class="pre">get_track_instance_matches()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.match_dict_from_match_function"><code class="docutils literal notranslate"><span class="pre">match_dict_from_match_function()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix"><code class="docutils literal notranslate"><span class="pre">remove_second_bests_from_cost_matrix()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,7 +328,7 @@ <h1>sleap.nn.tracker.kalman</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.tracker.kalman">
-<span id="sleap-nn-tracker-kalman"></span><h1>sleap.nn.tracker.kalman<a class="headerlink" href="#module-sleap.nn.tracker.kalman" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-tracker-kalman"></span><h1>sleap.nn.tracker.kalman<a class="headerlink" href="#module-sleap.nn.tracker.kalman" title="Permalink to this heading">#</a></h1>
 <p>Module to use Kalman filters for tracking instance identities.</p>
 <p>The Kalman filters needs a small number of frames already tracked in order
 to initialize the filters. Then you can use the module for tracking on the
@@ -328,7 +338,7 @@ <h1>sleap.nn.tracker.kalman</h1>
 up the filters.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.kalman.get_track_instance_matches">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">get_track_instance_matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">are_too_close_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.tracker.components.html#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match"><span class="pre">sleap.nn.tracker.components.Match</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/kalman.py#L447-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.get_track_instance_matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">get_track_instance_matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">are_too_close_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.tracker.components.html#sleap.nn.tracker.components.Match" title="sleap.nn.tracker.components.Match"><span class="pre">Match</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/kalman.py#L447-L527"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.get_track_instance_matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Matches track identities (from filters) to instances in frame.</p>
 <p>Algorithm is modified greedy matching.</p>
 <p>Standard greedy matching:
@@ -360,7 +370,7 @@ <h1>sleap.nn.tracker.kalman</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.kalman.match_dict_from_match_function">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">match_dict_from_match_function</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">row_items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">column_items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">key_by_column</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/kalman.py#L530-L560"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.match_dict_from_match_function" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">match_dict_from_match_function</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">row_items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">column_items</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">match_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">key_by_column</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/kalman.py#L530-L560"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.match_dict_from_match_function" title="Permalink to this definition">#</a></dt>
 <dd><p>Dict keys are from column (tracks), values are from row (instances).</p>
 <p>If multiple rows (instances) match on the same column (track), then
 dict will just contain the best match.</p>
@@ -368,7 +378,7 @@ <h1>sleap.nn.tracker.kalman</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">remove_second_bests_from_cost_matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invalid_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">numpy.ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/kalman.py#L585-L667"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracker.kalman.</span></span><span class="sig-name descname"><span class="pre">remove_second_bests_from_cost_matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thresh</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invalid_val</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">nan</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">ndarray</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/kalman.py#L585-L667"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix" title="Permalink to this definition">#</a></dt>
 <dd><p>Removes unclear matches from cost matrix.</p>
 <p>If the best match for a given track is too close to the second best match,
 then this will clear all the matches for that track (and ensure that any
@@ -377,7 +387,7 @@ <h1>sleap.nn.tracker.kalman</h1>
 <p>It removes the matches by setting the appropriate rows/columns to the
 specified invalid_val (usually nan or inf).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>cost_matrix</strong> – Cost matrix for matching, lower means better match.</p></li>
 <li><p><strong>thresh</strong> – Best match must be better than second best + threshold
@@ -385,7 +395,7 @@ <h1>sleap.nn.tracker.kalman</h1>
 <li><p><strong>invalid_val</strong> – Value to set invalid rows/columns to.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>cost matrix with invalid matches set to specified invalid value.</p>
 </dd>
 </dl>
@@ -415,6 +425,22 @@ <h1>sleap.nn.tracker.kalman</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.get_track_instance_matches"><code class="docutils literal notranslate"><span class="pre">get_track_instance_matches()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.match_dict_from_match_function"><code class="docutils literal notranslate"><span class="pre">match_dict_from_match_function()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix"><code class="docutils literal notranslate"><span class="pre">remove_second_bests_from_cost_matrix()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -433,7 +459,7 @@ <h1>sleap.nn.tracker.kalman</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.tracking.html b/api/sleap.nn.tracking.html
index 8095f5470..8f63b9c08 100644
--- a/api/sleap.nn.tracking.html
+++ b/api/sleap.nn.tracking.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.tracking &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.tracking &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.tracking';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.tracking.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,68 @@ <h1>sleap.nn.tracking</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.BaseTracker"><code class="docutils literal notranslate"><span class="pre">BaseTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.min_points"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.min_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.img_scale"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.img_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.of_window_size"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.of_window_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.of_max_levels"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.of_max_levels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.save_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.save_shifted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.track_window"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.track_window</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.flow_shift_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.get_shifted_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.get_shifted_instances_from_earlier_time()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.prune_shifted_instances()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker.matching_function"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker.matching_function()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker.similarity_function"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker.similarity_function()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.max_tracks"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker.max_tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker.get_ref_instances()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowTracker"><code class="docutils literal notranslate"><span class="pre">FlowTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker"><code class="docutils literal notranslate"><span class="pre">KalmanTracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.is_valid"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.is_valid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.make_tracker"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.make_tracker()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.track"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.track()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleCandidateMaker"><code class="docutils literal notranslate"><span class="pre">SimpleCandidateMaker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleMaxTracker"><code class="docutils literal notranslate"><span class="pre">SimpleMaxTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleMaxTracksCandidateMaker"><code class="docutils literal notranslate"><span class="pre">SimpleMaxTracksCandidateMaker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleTracker"><code class="docutils literal notranslate"><span class="pre">SimpleTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner"><code class="docutils literal notranslate"><span class="pre">TrackCleaner</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner.instance_count"><code class="docutils literal notranslate"><span class="pre">TrackCleaner.instance_count</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner.iou_threshold"><code class="docutils literal notranslate"><span class="pre">TrackCleaner.iou_threshold</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker"><code class="docutils literal notranslate"><span class="pre">Tracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.track_window"><code class="docutils literal notranslate"><span class="pre">Tracker.track_window</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.similarity_function"><code class="docutils literal notranslate"><span class="pre">Tracker.similarity_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.matching_function"><code class="docutils literal notranslate"><span class="pre">Tracker.matching_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.candidate_maker"><code class="docutils literal notranslate"><span class="pre">Tracker.candidate_maker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.cleaner"><code class="docutils literal notranslate"><span class="pre">Tracker.cleaner</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.min_new_track_points"><code class="docutils literal notranslate"><span class="pre">Tracker.min_new_track_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.robust_best_instance"><code class="docutils literal notranslate"><span class="pre">Tracker.robust_best_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.max_tracking"><code class="docutils literal notranslate"><span class="pre">Tracker.max_tracking</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.final_pass"><code class="docutils literal notranslate"><span class="pre">Tracker.final_pass()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.track"><code class="docutils literal notranslate"><span class="pre">Tracker.track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.unique_tracks_in_queue"><code class="docutils literal notranslate"><span class="pre">Tracker.unique_tracks_in_queue</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.run_tracker"><code class="docutils literal notranslate"><span class="pre">run_tracker()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,17 +380,17 @@ <h1>sleap.nn.tracking</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.tracking">
-<span id="sleap-nn-tracking"></span><h1>sleap.nn.tracking<a class="headerlink" href="#module-sleap.nn.tracking" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-tracking"></span><h1>sleap.nn.tracking<a class="headerlink" href="#module-sleap.nn.tracking" title="Permalink to this heading">#</a></h1>
 <p>Tracking tools for linking grouped instances over time.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.BaseTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">BaseTracker</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L504-L531"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.BaseTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">BaseTracker</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L511-L538"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.BaseTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Abstract base class for tracker.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracking.ShiftedInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L107-L354"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">ShiftedInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L110-L357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for producing optical flow shift matching candidates.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker.min_points">
@@ -336,7 +398,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>Minimum number of points that must be detected in the new frame in
 order to generate a new shifted instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -349,7 +411,7 @@ <h1>sleap.nn.tracking</h1>
 this to increase performance at the cost of finer accuracy. Sometimes
 decreasing the image scale can improve performance with fast movements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -361,7 +423,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>Optical flow window size to consider at each pyramid scale
 level.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -373,7 +435,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>Number of pyramid scale levels to consider. This is different
 from the scale parameter, which determines the initial image scaling.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -385,7 +447,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>If True, save the shifted instances between elapsed
 frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -397,7 +459,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>How many frames back to look for candidate instances to match
 instances in the current frame against.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -405,10 +467,10 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">flow_shift_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_shifted_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracking.ShiftedInstance</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L252-L354"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">flow_shift_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_shifted_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">ShiftedInstance</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L255-L357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Generates instances in a new frame by applying optical flow displacements.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ref_instances</strong> – Reference instances in the previous frame.</p></li>
 <li><p><strong>ref_img</strong> – Previous frame image as a numpy array.</p></li>
@@ -424,7 +486,7 @@ <h1>sleap.nn.tracking</h1>
 from the scale parameter, which determines the initial image scaling.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of ShiftedInstances with the optical flow displacements applied to
 the reference instance points. Points that are not found will be represented
 as NaNs in the points array for each shifted instance.</p>
@@ -438,10 +500,10 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances">
-<span class="sig-name descname"><span class="pre">get_shifted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracking.ShiftedInstance</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L163-L200"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_shifted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">ShiftedInstance</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L166-L203"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns a list of shifted instances and save shifted instances if needed.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ref_instances</strong> – Reference instances in the previous frame.</p></li>
 <li><p><strong>ref_img</strong> – Previous frame image as a numpy array.</p></li>
@@ -455,10 +517,10 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time">
-<span class="sig-name descname"><span class="pre">get_shifted_instances_from_earlier_time</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">ref_t:</span> <span class="pre">int,</span> <span class="pre">ref_img:</span> <span class="pre">numpy.ndarray,</span> <span class="pre">ref_instances:</span> <span class="pre">typing.List[sleap.nn.tracker.components.InstanceType],</span> <span class="pre">t:</span> <span class="pre">int)</span> <span class="pre">-&gt;</span> <span class="pre">(&lt;class</span> <span class="pre">'numpy.ndarray'&gt;,</span> <span class="pre">typing.List[~InstanceType]</span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L142-L161"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_shifted_instances_from_earlier_time</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">ref_t:</span> <span class="pre">int,</span> <span class="pre">ref_img:</span> <span class="pre">~numpy.ndarray,</span> <span class="pre">ref_instances:</span> <span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType],</span> <span class="pre">t:</span> <span class="pre">int)</span> <span class="pre">-&gt;</span> <span class="pre">(&lt;class</span> <span class="pre">'numpy.ndarray'&gt;,</span> <span class="pre">typing.List[~InstanceType]</span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L145-L164"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate shifted instances and corresponding image from earlier time.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ref_instances</strong> – Reference instances in the previous frame.</p></li>
 <li><p><strong>ref_img</strong> – Previous frame image as a numpy array.</p></li>
@@ -471,7 +533,7 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances">
-<span class="sig-name descname"><span class="pre">prune_shifted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L233-L250"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">prune_shifted_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L236-L253"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Prune the shifted instances older than <code class="xref py py-obj docutils literal notranslate"><span class="pre">self.track_window</span></code>.</p>
 <p>If <code class="xref py py-obj docutils literal notranslate"><span class="pre">self.save_shifted_instances</span></code> is False, do nothing.</p>
 <dl class="simple">
@@ -487,17 +549,17 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowMaxTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">typing.Dict[sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">typing.List[sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.List[sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">typing.Optional[sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1091-L1098"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowMaxTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">~typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">~typing.Dict[~sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">~typing.List[~sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">~typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">~typing.Optional[~sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1172-L1179"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Pre-configured tracker to use optical flow shifted candidates with max tracks.</p>
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracker.matching_function">
-<span class="sig-name descname"><span class="pre">matching_function</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L88-L109"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker.matching_function" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matching_function</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L196-L217"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker.matching_function" title="Permalink to this definition">#</a></dt>
 <dd><p>Performs greedy bipartite matching.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracker.similarity_function">
-<span class="sig-name descname"><span class="pre">similarity_function</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.components.InstanceType</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracker/components.py#L29-L40"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker.similarity_function" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">similarity_function</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">query_instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">InstanceType</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracker/components.py#L48-L59"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracker.similarity_function" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes similarity between instances.</p>
 </dd></dl>
 
@@ -505,14 +567,14 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracksCandidateMaker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowMaxTracksCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracking.ShiftedInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L358-L434"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowMaxTracksCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_window_size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">21</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">of_max_levels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">3</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shifted_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">ShiftedInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L361-L438"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for producing optical flow shift matching candidates with maximum tracks.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracksCandidateMaker.max_tracks">
 <span class="sig-name descname"><span class="pre">max_tracks</span></span><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.max_tracks" title="Permalink to this definition">#</a></dt>
 <dd><p>The maximum number of tracks to avoid redundant tracks.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -520,10 +582,10 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">get_ref_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Deque</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracking.MatchedFrameInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L368-L389"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">get_ref_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ref_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ref_img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Deque</span><span class="p"><span class="pre">[</span></span><span class="pre">MatchedFrameInstance</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L371-L392"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Generates a list of instances based on the reference time and image.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>ref_t</strong> – Previous frame time instance.</p></li>
 <li><p><strong>ref_img</strong> – Previous frame image as a numpy array.</p></li>
@@ -538,18 +600,18 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.FlowTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">typing.Dict[sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">typing.List[sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.List[sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">typing.Optional[sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1080-L1085"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">FlowTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">~typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">~typing.Dict[~sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">~typing.List[~sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">~typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">~typing.Optional[~sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1161-L1166"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.FlowTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>A Tracker pre-configured to use optical flow shifted candidates.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.KalmanTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">KalmanTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">init_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker"><span class="pre">sleap.nn.tracking.Tracker</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_set</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracking.KalmanInitSet</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kalman_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">sleap.nn.tracker.kalman.BareKalmanTracker</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cull_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_frame_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">re_init_cooldown</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">re_init_after</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_done</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_tracked</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_init_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1198-L1391"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">KalmanTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">init_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker"><span class="pre">Tracker</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_set</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">KalmanInitSet</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kalman_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">BareKalmanTracker</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cull_function</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Callable</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_frame_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">re_init_cooldown</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">100</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">re_init_after</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">20</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_done</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_tracked</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_init_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1279-L1472"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for Kalman filter-based tracking pipeline.</p>
 <p>Kalman filters need to be initialized with a certain number of already
 tracked instances.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>init_tracker</strong> – The regular Tracker we can use to track data needed
 for initializing Kalman filters. If not specified, then you can
@@ -585,10 +647,10 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.KalmanTracker.make_tracker">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">init_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker"><span class="pre">sleap.nn.tracking.Tracker</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_frame_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1250-L1301"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker.make_tracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">init_tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.nn.tracking.Tracker" title="sleap.nn.tracking.Tracker"><span class="pre">Tracker</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_indices</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">instance_iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0.8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">init_frame_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">10</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1331-L1382"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker.make_tracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Creates KalmanTracker object.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>init_tracker</strong> – The Kalman filters need to be initialized with data
 that’s already been tracked. This is a regular Tracker which
@@ -610,7 +672,7 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.KalmanTracker.track">
-<span class="sig-name descname"><span class="pre">track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1303-L1381"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker.track" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1384-L1462"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.KalmanTracker.track" title="Permalink to this definition">#</a></dt>
 <dd><p>Tracks individual frame, using Kalman filters if possible.</p>
 </dd></dl>
 
@@ -618,31 +680,31 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.SimpleCandidateMaker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L438-L457"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleCandidateMaker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L442-L461"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleCandidateMaker" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for producing list of matching candidates from prior frames.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.SimpleMaxTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleMaxTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">typing.Dict[sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">typing.List[sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.List[sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">typing.Optional[sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_iou&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">hungarian_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1111-L1118"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleMaxTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleMaxTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">~typing.Dict[~sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">~typing.List[~sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">~typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">~typing.Optional[~sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_iou&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">hungarian_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">*</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">int</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1192-L1199"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleMaxTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Pre-configured tracker to use simple, non-image-based candidates with max tracks.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.SimpleMaxTracksCandidateMaker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleMaxTracksCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L461-L481"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleMaxTracksCandidateMaker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleMaxTracksCandidateMaker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min_points</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracks</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L465-L486"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleMaxTracksCandidateMaker" title="Permalink to this definition">#</a></dt>
 <dd><p>Class to generate instances with maximum number of tracks from prior frames.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.SimpleTracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">typing.Dict[sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">typing.List[sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.List[sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">typing.Optional[sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_iou&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">hungarian_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1102-L1107"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleTracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">SimpleTracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">~typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">~typing.Dict[~sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">~typing.List[~sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">~typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">~typing.Optional[~sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_iou&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">hungarian_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1183-L1188"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.SimpleTracker" title="Permalink to this definition">#</a></dt>
 <dd><p>A Tracker pre-configured to use simple, non-image-based candidates.</p>
 </dd></dl>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.TrackCleaner">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">TrackCleaner</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1395-L1420"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.TrackCleaner" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">TrackCleaner</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">iou_threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1476-L1501"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.TrackCleaner" title="Permalink to this definition">#</a></dt>
 <dd><p>Class for merging breaks in the predicted tracks.</p>
 <p>Method:
 1. You specify how many instances there should be in each frame.
@@ -658,7 +720,7 @@ <h1>sleap.nn.tracking</h1>
 <span class="sig-name descname"><span class="pre">instance_count</span></span><a class="headerlink" href="#sleap.nn.tracking.TrackCleaner.instance_count" title="Permalink to this definition">#</a></dt>
 <dd><p>The maximum number of instances we want per frame.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -671,7 +733,7 @@ <h1>sleap.nn.tracking</h1>
 removing overlapping instances over target count; if None, then
 only use score to determine which instances to remove.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[float]</p>
 </dd>
 </dl>
@@ -681,7 +743,7 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.tracking.Tracker">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">Tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">typing.Optional[typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">typing.Dict[sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.Deque[sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">typing.List[sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">typing.List[sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">typing.Optional[sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L535-L1076"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">Tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">max_tracks:</span> <span class="pre">~typing.Optional[int]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_window:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">similarity_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">instance_similarity&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">matching_function:</span> <span class="pre">~typing.Callable</span> <span class="pre">=</span> <span class="pre">&lt;function</span> <span class="pre">greedy_matching&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">candidate_maker:</span> <span class="pre">object</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_tracking:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cleaner:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_instance_count:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pre_cull_function:</span> <span class="pre">~typing.Optional[~typing.Callable]</span> <span class="pre">=</span> <span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">post_connect_single_breaks:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">robust_best_instance:</span> <span class="pre">float</span> <span class="pre">=</span> <span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_new_track_points:</span> <span class="pre">int</span> <span class="pre">=</span> <span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue:</span> <span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstances]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">track_matching_queue_dict:</span> <span class="pre">~typing.Dict[~sleap.instance.Track</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.Deque[~sleap.nn.tracking.MatchedFrameInstance]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spawned_tracks:</span> <span class="pre">~typing.List[~sleap.instance.Track]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">save_tracked_instances:</span> <span class="pre">bool</span> <span class="pre">=</span> <span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracked_instances:</span> <span class="pre">~typing.Dict[int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">~typing.List[~sleap.nn.tracker.components.InstanceType]]</span> <span class="pre">=</span> <span class="pre">_Nothing.NOTHING</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">last_matches:</span> <span class="pre">~typing.Optional[~sleap.nn.tracker.components.FrameMatches]</span> <span class="pre">=</span> <span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L542-L1157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Instance pose tracker.</p>
 <p>Use by instantiated with the desired parameters and then calling the
 <a class="reference internal" href="#sleap.nn.tracking.Tracker.track" title="sleap.nn.tracking.Tracker.track"><code class="xref py py-obj docutils literal notranslate"><span class="pre">track</span></code></a> method for each frame.</p>
@@ -691,7 +753,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>How many frames back to look for candidate instances to
 match instances in the current frame against.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -703,7 +765,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>A function that returns a numeric pairwise
 instance similarity value.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[Callable]</p>
 </dd>
 </dl>
@@ -715,7 +777,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>A function that takes a matrix of pairwise similarities
 and determines the matches to use.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Callable</p>
 </dd>
 </dl>
@@ -728,7 +790,7 @@ <h1>sleap.nn.tracking</h1>
 which returns a list of Instances-like objects  which we can match
 the predicted instances in a frame against.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>object</p>
 </dd>
 </dl>
@@ -740,7 +802,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>A class with a <code class="xref py py-obj docutils literal notranslate"><span class="pre">run</span></code> method which attempts to clean tracks
 after the other tracking has run for all frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[Callable]</p>
 </dd>
 </dl>
@@ -752,7 +814,7 @@ <h1>sleap.nn.tracking</h1>
 <dd><p>We won’t spawn a new track for an instance with
 fewer than this many points.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>int</p>
 </dd>
 </dl>
@@ -766,7 +828,7 @@ <h1>sleap.nn.tracking</h1>
 use the max similarity (non-robust). For selecting a robust score,
 0.95 is a good value.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -777,7 +839,7 @@ <h1>sleap.nn.tracking</h1>
 <span class="sig-name descname"><span class="pre">max_tracking</span></span><a class="headerlink" href="#sleap.nn.tracking.Tracker.max_tracking" title="Permalink to this definition">#</a></dt>
 <dd><p>Max tracking is incorporated when this is set to true.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -785,23 +847,24 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.Tracker.final_pass">
-<span class="sig-name descname"><span class="pre">final_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L790-L802"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker.final_pass" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">final_pass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L816-L828"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker.final_pass" title="Permalink to this definition">#</a></dt>
 <dd><p>Called after tracking has run on all frames to do any post-processing.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.tracking.Tracker.track">
-<span class="sig-name descname"><span class="pre">track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">sleap.nn.tracker.components.InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L627-L748"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker.track" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">track</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">untracked_instances</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img_hw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">InstanceType</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L642-L773"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.Tracker.track" title="Permalink to this definition">#</a></dt>
 <dd><p>Performs a single step of tracking.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>untracked_instances</strong> – List of instances to assign to tracks.</p></li>
+<li><p><strong>img_hw</strong> – (height, width) of the image used to normalize the keypoints.</p></li>
 <li><p><strong>img</strong> – Image data of the current frame for flow shifting.</p></li>
 <li><p><strong>t</strong> – Current timestep. If not provided, increments from the internal queue.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of the instances that were tracked.</p>
 </dd>
 </dl>
@@ -809,7 +872,7 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.tracking.Tracker.unique_tracks_in_queue">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unique_tracks_in_queue</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">sleap.instance.Track</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.tracking.Tracker.unique_tracks_in_queue" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">unique_tracks_in_queue</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.Track" title="sleap.instance.Track"><span class="pre">Track</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.tracking.Tracker.unique_tracks_in_queue" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the unique tracks in the matching queue.</p>
 </dd></dl>
 
@@ -817,16 +880,16 @@ <h1>sleap.nn.tracking</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.tracking.run_tracker">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">run_tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.tracking.BaseTracker" title="sleap.nn.tracking.BaseTracker"><span class="pre">sleap.nn.tracking.BaseTracker</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">sleap.instance.LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/tracking.py#L1423-L1460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.run_tracker" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.tracking.</span></span><span class="sig-name descname"><span class="pre">run_tracker</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frames</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracker</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.tracking.BaseTracker" title="sleap.nn.tracking.BaseTracker"><span class="pre">BaseTracker</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.instance.html#sleap.instance.LabeledFrame" title="sleap.instance.LabeledFrame"><span class="pre">LabeledFrame</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/tracking.py#L1504-L1542"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.tracking.run_tracker" title="Permalink to this definition">#</a></dt>
 <dd><p>Run a tracker on a set of labeled frames.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>frames</strong> – A list of labeled frames with instances.</p></li>
 <li><p><strong>tracker</strong> – An initialized Tracker.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The input frames with the new tracks assigned. If the frames already had tracks,
 they will be cleared if the tracker has been re-initialized.</p>
 </dd>
@@ -857,6 +920,74 @@ <h1>sleap.nn.tracking</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.BaseTracker"><code class="docutils literal notranslate"><span class="pre">BaseTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.min_points"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.min_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.img_scale"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.img_scale</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.of_window_size"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.of_window_size</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.of_max_levels"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.of_max_levels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.save_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.save_shifted_instances</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.track_window"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.track_window</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.flow_shift_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.get_shifted_instances()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.get_shifted_instances_from_earlier_time()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances"><code class="docutils literal notranslate"><span class="pre">FlowCandidateMaker.prune_shifted_instances()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker.matching_function"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker.matching_function()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracker.similarity_function"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracker.similarity_function()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.max_tracks"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker.max_tracks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances"><code class="docutils literal notranslate"><span class="pre">FlowMaxTracksCandidateMaker.get_ref_instances()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.FlowTracker"><code class="docutils literal notranslate"><span class="pre">FlowTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker"><code class="docutils literal notranslate"><span class="pre">KalmanTracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.is_valid"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.is_valid</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.make_tracker"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.make_tracker()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.KalmanTracker.track"><code class="docutils literal notranslate"><span class="pre">KalmanTracker.track()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleCandidateMaker"><code class="docutils literal notranslate"><span class="pre">SimpleCandidateMaker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleMaxTracker"><code class="docutils literal notranslate"><span class="pre">SimpleMaxTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleMaxTracksCandidateMaker"><code class="docutils literal notranslate"><span class="pre">SimpleMaxTracksCandidateMaker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.SimpleTracker"><code class="docutils literal notranslate"><span class="pre">SimpleTracker</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner"><code class="docutils literal notranslate"><span class="pre">TrackCleaner</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner.instance_count"><code class="docutils literal notranslate"><span class="pre">TrackCleaner.instance_count</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.TrackCleaner.iou_threshold"><code class="docutils literal notranslate"><span class="pre">TrackCleaner.iou_threshold</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker"><code class="docutils literal notranslate"><span class="pre">Tracker</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.track_window"><code class="docutils literal notranslate"><span class="pre">Tracker.track_window</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.similarity_function"><code class="docutils literal notranslate"><span class="pre">Tracker.similarity_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.matching_function"><code class="docutils literal notranslate"><span class="pre">Tracker.matching_function</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.candidate_maker"><code class="docutils literal notranslate"><span class="pre">Tracker.candidate_maker</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.cleaner"><code class="docutils literal notranslate"><span class="pre">Tracker.cleaner</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.min_new_track_points"><code class="docutils literal notranslate"><span class="pre">Tracker.min_new_track_points</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.robust_best_instance"><code class="docutils literal notranslate"><span class="pre">Tracker.robust_best_instance</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.max_tracking"><code class="docutils literal notranslate"><span class="pre">Tracker.max_tracking</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.final_pass"><code class="docutils literal notranslate"><span class="pre">Tracker.final_pass()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.track"><code class="docutils literal notranslate"><span class="pre">Tracker.track()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.Tracker.unique_tracks_in_queue"><code class="docutils literal notranslate"><span class="pre">Tracker.unique_tracks_in_queue</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.tracking.run_tracker"><code class="docutils literal notranslate"><span class="pre">run_tracker()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -875,7 +1006,7 @@ <h1>sleap.nn.tracking</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.training.html b/api/sleap.nn.training.html
index d7b6167b8..f74a1c9c0 100644
--- a/api/sleap.nn.training.html
+++ b/api/sleap.nn.training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.training &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.training &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.training';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,100 @@ <h1>sleap.nn.training</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders"><code class="docutils literal notranslate"><span class="pre">DataReaders</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.training_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.training_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.validation_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.validation_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.test_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.test_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.from_config"><code class="docutils literal notranslate"><span class="pre">DataReaders.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.from_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.from_labels()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.test_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.test_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.training_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.training_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.validation_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.validation_labels</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer"><code class="docutils literal notranslate"><span class="pre">Trainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.data_readers"><code class="docutils literal notranslate"><span class="pre">Trainer.data_readers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.model"><code class="docutils literal notranslate"><span class="pre">Trainer.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.config"><code class="docutils literal notranslate"><span class="pre">Trainer.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.initial_config"><code class="docutils literal notranslate"><span class="pre">Trainer.initial_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.pipeline_builder"><code class="docutils literal notranslate"><span class="pre">Trainer.pipeline_builder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.training_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.training_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.validation_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.validation_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.training_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.training_viz_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.validation_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.validation_viz_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.optimization_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.optimization_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.output_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.output_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.visualization_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.visualization_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.run_path"><code class="docutils literal notranslate"><span class="pre">Trainer.run_path</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.cleanup"><code class="docutils literal notranslate"><span class="pre">Trainer.cleanup()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.evaluate"><code class="docutils literal notranslate"><span class="pre">Trainer.evaluate()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.from_config"><code class="docutils literal notranslate"><span class="pre">Trainer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.input_keys"><code class="docutils literal notranslate"><span class="pre">Trainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.keras_model"><code class="docutils literal notranslate"><span class="pre">Trainer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.output_keys"><code class="docutils literal notranslate"><span class="pre">Trainer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.package"><code class="docutils literal notranslate"><span class="pre">Trainer.package()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.setup"><code class="docutils literal notranslate"><span class="pre">Trainer.setup()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.train"><code class="docutils literal notranslate"><span class="pre">Trainer.train()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.create_trainer_using_cli"><code class="docutils literal notranslate"><span class="pre">create_trainer_using_cli()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.get_timestamp"><code class="docutils literal notranslate"><span class="pre">get_timestamp()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.sanitize_scope_name"><code class="docutils literal notranslate"><span class="pre">sanitize_scope_name()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_checkpointing"><code class="docutils literal notranslate"><span class="pre">setup_checkpointing()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_losses"><code class="docutils literal notranslate"><span class="pre">setup_losses()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_metrics"><code class="docutils literal notranslate"><span class="pre">setup_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_new_run_folder"><code class="docutils literal notranslate"><span class="pre">setup_new_run_folder()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_optimization_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_optimization_callbacks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_optimizer"><code class="docutils literal notranslate"><span class="pre">setup_optimizer()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_output_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_output_callbacks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_tensorboard"><code class="docutils literal notranslate"><span class="pre">setup_tensorboard()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_visualization"><code class="docutils literal notranslate"><span class="pre">setup_visualization()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_zmq_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_zmq_callbacks()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,11 +412,11 @@ <h1>sleap.nn.training</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.training">
-<span id="sleap-nn-training"></span><h1>sleap.nn.training<a class="headerlink" href="#module-sleap.nn.training" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-training"></span><h1>sleap.nn.training<a class="headerlink" href="#module-sleap.nn.training" title="Permalink to this heading">#</a></h1>
 <p>Training functionality and high level APIs.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.BottomUpModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">BottomUpModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1361-L1498"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.BottomUpModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">BottomUpModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1361-L1498"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.BottomUpModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output multi-instance confidence maps and PAFs.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.BottomUpModelTrainer.has_offsets">
@@ -346,7 +440,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.BottomUpMultiClassModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1502-L1637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.BottomUpMultiClassModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">BottomUpMultiClassModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1502-L1637"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.BottomUpMultiClassModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output multi-instance confidence maps and class maps.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.BottomUpMultiClassModelTrainer.has_offsets">
@@ -370,7 +464,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.CentroidConfmapsModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1126-L1226"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.CentroidConfmapsModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">CentroidConfmapsModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1126-L1226"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.CentroidConfmapsModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output centroid confidence maps.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.CentroidConfmapsModelTrainer.has_offsets">
@@ -394,7 +488,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">DataReaders</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">sleap.nn.data.providers.LabelsReader</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L86-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">DataReaders</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels_reader</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader"><span class="pre">LabelsReader</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L86-L264"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders" title="Permalink to this definition">#</a></dt>
 <dd><p>Container class for SLEAP labels that serve as training data sources.</p>
 <dl class="py attribute">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.training_labels_reader">
@@ -402,7 +496,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>LabelsReader pipeline provider for a training data from
 a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader">sleap.nn.data.providers.LabelsReader</a></p>
 </dd>
 </dl>
@@ -414,7 +508,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>LabelsReader pipeline provider for a validation data
 from a <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader">sleap.nn.data.providers.LabelsReader</a></p>
 </dd>
 </dl>
@@ -426,7 +520,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>LabelsReader pipeline provider for a test set data from a
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> instance. This is not necessary for training.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader" title="sleap.nn.data.providers.LabelsReader">sleap.nn.data.providers.LabelsReader</a>]</p>
 </dd>
 </dl>
@@ -434,31 +528,31 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">sleap.nn.config.data.LabelsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">training</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L102-L151"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">labels_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">LabelsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">training</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L102-L151"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Create data readers from a (possibly incomplete) configuration.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.from_labels">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">sleap.nn.config.data.LabelsConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L153-L247"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders.from_labels" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_labels</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">training</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">labels_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.data.html#sleap.nn.config.data.LabelsConfig" title="sleap.nn.config.data.LabelsConfig"><span class="pre">LabelsConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_track_only</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L153-L247"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.DataReaders.from_labels" title="Permalink to this definition">#</a></dt>
 <dd><p>Create data readers from <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> datasets as data providers.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.test_labels">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">test_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.test_labels" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">test_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.test_labels" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> underlying the test data reader.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.training_labels">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">training_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.training_labels" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">training_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.training_labels" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> underlying the training data reader.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.DataReaders.validation_labels">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">validation_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.validation_labels" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">validation_labels</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a></em><a class="headerlink" href="#sleap.nn.training.DataReaders.validation_labels" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the <code class="xref py py-obj docutils literal notranslate"><span class="pre">Labels</span></code> underlying the validation data reader.</p>
 </dd></dl>
 
@@ -466,7 +560,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.SingleInstanceModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1009-L1122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.SingleInstanceModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">SingleInstanceModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1009-L1122"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.SingleInstanceModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output single-instance confidence maps.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.SingleInstanceModelTrainer.has_offsets">
@@ -490,7 +584,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.TopDownMultiClassModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1641-L1797"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.TopDownMultiClassModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">TopDownMultiClassModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1641-L1797"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.TopDownMultiClassModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output multi-instance confidence maps and class maps.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.TopDownMultiClassModelTrainer.has_offsets">
@@ -514,7 +608,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.TopdownConfmapsModelTrainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">TopdownConfmapsModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1230-L1357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.TopdownConfmapsModelTrainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">TopdownConfmapsModelTrainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1230-L1357"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.TopdownConfmapsModelTrainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Trainer for models that output instance centered confidence maps.</p>
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.TopdownConfmapsModelTrainer.has_offsets">
@@ -538,7 +632,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">Trainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">sleap.nn.training.DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">sleap.nn.model.Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L560-L1005"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">Trainer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data_readers</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><span class="pre">DataReaders</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model"><span class="pre">Model</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">initial_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L560-L1005"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer" title="Permalink to this definition">#</a></dt>
 <dd><p>Base trainer class that provides general model training functionality.</p>
 <p>This class is intended to be instantiated using the <a class="reference internal" href="#sleap.nn.training.Trainer.from_config" title="sleap.nn.training.Trainer.from_config"><code class="xref py py-obj docutils literal notranslate"><span class="pre">from_config()</span></code></a> class method,
 which will return the appropriate subclass based on the input configuration.</p>
@@ -549,7 +643,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">data_readers</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.data_readers" title="Permalink to this definition">#</a></dt>
 <dd><p>A <a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders"><code class="xref py py-obj docutils literal notranslate"><span class="pre">DataReaders</span></code></a> instance that contains training data providers.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="#sleap.nn.training.DataReaders" title="sleap.nn.training.DataReaders">sleap.nn.training.DataReaders</a></p>
 </dd>
 </dl>
@@ -560,7 +654,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">model</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.model" title="Permalink to this definition">#</a></dt>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">Model</span></code> instance describing the SLEAP model to train.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.model.html#sleap.nn.model.Model" title="sleap.nn.model.Model">sleap.nn.model.Model</a></p>
 </dd>
 </dl>
@@ -571,7 +665,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">config</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.config" title="Permalink to this definition">#</a></dt>
 <dd><p>A <code class="xref py py-obj docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code> that describes the training parameters.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a></p>
 </dd>
 </dl>
@@ -583,7 +677,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>This attribute will contain a copy of the input configuration
 before any attributes are updated in <a class="reference internal" href="#sleap.nn.training.Trainer.config" title="sleap.nn.training.Trainer.config"><code class="xref py py-obj docutils literal notranslate"><span class="pre">config</span></code></a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[<a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig">sleap.nn.config.training_job.TrainingJobConfig</a>]</p>
 </dd>
 </dl>
@@ -596,7 +690,7 @@ <h1>sleap.nn.training</h1>
 pipelines that generate data used for training. This must be specified in
 subclasses.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>sleap.nn.training.PipelineBuilder</p>
 </dd>
 </dl>
@@ -608,7 +702,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>The data pipeline that generates examples from the training
 set for optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a></p>
 </dd>
 </dl>
@@ -620,7 +714,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>The data pipeline that generates examples from the
 validation set for optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a></p>
 </dd>
 </dl>
@@ -632,7 +726,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>The data pipeline that generates examples from the
 training set for visualization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a></p>
 </dd>
 </dl>
@@ -644,7 +738,7 @@ <h1>sleap.nn.training</h1>
 <dd><p>The data pipeline that generates examples from the
 validation set for visualization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p><a class="reference internal" href="sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline" title="sleap.nn.data.pipelines.Pipeline">sleap.nn.data.pipelines.Pipeline</a></p>
 </dd>
 </dl>
@@ -655,7 +749,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">optimization_callbacks</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.optimization_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras callbacks related to optimization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[keras.callbacks.Callback]</p>
 </dd>
 </dl>
@@ -666,7 +760,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">output_callbacks</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.output_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras callbacks related to outputs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[keras.callbacks.Callback]</p>
 </dd>
 </dl>
@@ -677,7 +771,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">visualization_callbacks</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.visualization_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Keras callbacks related to visualization.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>List[keras.callbacks.Callback]</p>
 </dd>
 </dl>
@@ -688,7 +782,7 @@ <h1>sleap.nn.training</h1>
 <span class="sig-name descname"><span class="pre">run_path</span></span><a class="headerlink" href="#sleap.nn.training.Trainer.run_path" title="Permalink to this definition">#</a></dt>
 <dd><p>The path to the run folder that will contain training results, if any.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -696,28 +790,28 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.callbacks">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">callbacks</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.training.Trainer.callbacks" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">callbacks</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.nn.training.Trainer.callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Return all callbacks currently configured.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.cleanup">
-<span class="sig-name descname"><span class="pre">cleanup</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L984-L996"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.cleanup" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">cleanup</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L984-L996"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.cleanup" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete visualization images subdirectory.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.evaluate">
-<span class="sig-name descname"><span class="pre">evaluate</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L958-L982"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.evaluate" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">evaluate</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L958-L982"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.evaluate" title="Permalink to this definition">#</a></dt>
 <dd><p>Compute evaluation metrics on data splits and save them.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.from_config">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">sleap.nn.config.training_job.TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">training_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">sleap.io.dataset.Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.Trainer" title="sleap.nn.training.Trainer"><span class="pre">sleap.nn.training.Trainer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L615-L690"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.from_config" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_config</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig" title="sleap.nn.config.training_job.TrainingJobConfig"><span class="pre">TrainingJobConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">training_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">validation_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">float</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">test_labels</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="sleap.io.dataset.html#sleap.io.dataset.Labels" title="sleap.io.dataset.Labels"><span class="pre">Labels</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">video_search_paths</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.nn.training.Trainer" title="sleap.nn.training.Trainer"><span class="pre">Trainer</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L615-L690"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.from_config" title="Permalink to this definition">#</a></dt>
 <dd><p>Initialize the trainer from a training job configuration.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>config</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code> instance.</p></li>
 <li><p><strong>training_labels</strong> – Training labels to use instead of the ones in the config,
@@ -740,7 +834,7 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.keras_model">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">keras_model</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">keras.engine.training.Model</span></em><a class="headerlink" href="#sleap.nn.training.Trainer.keras_model" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">keras_model</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Model</span></em><a class="headerlink" href="#sleap.nn.training.Trainer.keras_model" title="Permalink to this definition">#</a></dt>
 <dd><p>Alias for <code class="xref py py-obj docutils literal notranslate"><span class="pre">self.model.keras_model</span></code>.</p>
 </dd></dl>
 
@@ -752,19 +846,19 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.package">
-<span class="sig-name descname"><span class="pre">package</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L998-L1005"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.package" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">package</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L998-L1005"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.package" title="Permalink to this definition">#</a></dt>
 <dd><p>Package model folder into a zip file for portability.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.setup">
-<span class="sig-name descname"><span class="pre">setup</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L902-L919"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.setup" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">setup</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L902-L919"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.setup" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up data pipeline and model for training.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.nn.training.Trainer.train">
-<span class="sig-name descname"><span class="pre">train</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L921-L956"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.train" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">train</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L921-L956"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.Trainer.train" title="Permalink to this definition">#</a></dt>
 <dd><p>Execute the optimization loop to train the model.</p>
 </dd></dl>
 
@@ -772,85 +866,85 @@ <h1>sleap.nn.training</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.create_trainer_using_cli">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">create_trainer_using_cli</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L1800-L2008"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.create_trainer_using_cli" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">create_trainer_using_cli</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L1800-L2033"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.create_trainer_using_cli" title="Permalink to this definition">#</a></dt>
 <dd><p>Create CLI for training.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.get_timestamp">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">get_timestamp</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L358-L360"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.get_timestamp" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">get_timestamp</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L358-L360"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.get_timestamp" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the date and time as a string.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.main">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L2011-L2014"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.main" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">main</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">args</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L2036-L2039"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.main" title="Permalink to this definition">#</a></dt>
 <dd><p>Create CLI for training and run.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.sanitize_scope_name">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">sanitize_scope_name</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L539-L545"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.sanitize_scope_name" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">sanitize_scope_name</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L539-L545"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.sanitize_scope_name" title="Permalink to this definition">#</a></dt>
 <dd><p>Sanitizes string which will be used as TensorFlow scope name.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_checkpointing">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_checkpointing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.CheckpointingConfig" title="sleap.nn.config.outputs.CheckpointingConfig"><span class="pre">sleap.nn.config.outputs.CheckpointingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L415-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_checkpointing" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_checkpointing</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.CheckpointingConfig" title="sleap.nn.config.outputs.CheckpointingConfig"><span class="pre">CheckpointingConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L415-L460"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_checkpointing" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up model checkpointing callbacks from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_losses">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_losses</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L288-L304"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_losses" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_losses</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L288-L304"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_losses" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up model loss function from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_metrics">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.losses.Loss</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">keras.metrics.Metric</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L307-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_metrics" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_metrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">part_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Loss</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Metric</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L307-L320"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_metrics" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up training metrics from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_new_run_folder">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_new_run_folder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">sleap.nn.config.outputs.OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">base_run_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L363-L393"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_new_run_folder" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_new_run_folder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">base_run_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L363-L393"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_new_run_folder" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a new run folder from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_optimization_callbacks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_optimization_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L323-L355"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_optimization_callbacks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_optimization_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L323-L355"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_optimization_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up optimization callbacks from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_optimizer">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_optimizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">sleap.nn.config.optimization.OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">keras.optimizer_v2.optimizer_v2.OptimizerV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L267-L285"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_optimizer" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_optimizer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.optimization.html#sleap.nn.config.optimization.OptimizationConfig" title="sleap.nn.config.optimization.OptimizationConfig"><span class="pre">OptimizationConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">OptimizerV2</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L267-L285"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_optimizer" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up model optimizer from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_output_callbacks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_output_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">sleap.nn.config.outputs.OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L484-L498"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_output_callbacks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_output_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L484-L498"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_output_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up training outputs callbacks from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_tensorboard">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_tensorboard</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.TensorBoardConfig" title="sleap.nn.config.outputs.TensorBoardConfig"><span class="pre">sleap.nn.config.outputs.TensorBoardConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L463-L481"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_tensorboard" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_tensorboard</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.TensorBoardConfig" title="sleap.nn.config.outputs.TensorBoardConfig"><span class="pre">TensorBoardConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L463-L481"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_tensorboard" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up TensorBoard callbacks from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_visualization">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_visualization</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">sleap.nn.config.outputs.OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">viz_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">matplotlib.figure.Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L501-L536"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_visualization" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_visualization</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig" title="sleap.nn.config.outputs.OutputsConfig"><span class="pre">OutputsConfig</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">run_path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">viz_fn</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">[</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Figure</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L501-L536"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_visualization" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up visualization callbacks from config.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.training.setup_zmq_callbacks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_zmq_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">zmq_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.ZMQConfig" title="sleap.nn.config.outputs.ZMQConfig"><span class="pre">sleap.nn.config.outputs.ZMQConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">keras.callbacks.Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/training.py#L396-L412"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_zmq_callbacks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.training.</span></span><span class="sig-name descname"><span class="pre">setup_zmq_callbacks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">zmq_config</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.nn.config.outputs.html#sleap.nn.config.outputs.ZMQConfig" title="sleap.nn.config.outputs.ZMQConfig"><span class="pre">ZMQConfig</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Callback</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/training.py#L396-L412"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.training.setup_zmq_callbacks" title="Permalink to this definition">#</a></dt>
 <dd><p>Set up ZeroMQ callbacks from config.</p>
 </dd></dl>
 
@@ -878,6 +972,106 @@ <h1>sleap.nn.training</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.BottomUpMultiClassModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">BottomUpMultiClassModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.CentroidConfmapsModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">CentroidConfmapsModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders"><code class="docutils literal notranslate"><span class="pre">DataReaders</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.training_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.training_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.validation_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.validation_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.test_labels_reader"><code class="docutils literal notranslate"><span class="pre">DataReaders.test_labels_reader</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.from_config"><code class="docutils literal notranslate"><span class="pre">DataReaders.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.from_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.from_labels()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.test_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.test_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.training_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.training_labels</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.DataReaders.validation_labels"><code class="docutils literal notranslate"><span class="pre">DataReaders.validation_labels</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.SingleInstanceModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">SingleInstanceModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopDownMultiClassModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">TopDownMultiClassModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.has_offsets"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.has_offsets</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.input_keys"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.TopdownConfmapsModelTrainer.output_keys"><code class="docutils literal notranslate"><span class="pre">TopdownConfmapsModelTrainer.output_keys</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer"><code class="docutils literal notranslate"><span class="pre">Trainer</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.data_readers"><code class="docutils literal notranslate"><span class="pre">Trainer.data_readers</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.model"><code class="docutils literal notranslate"><span class="pre">Trainer.model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.config"><code class="docutils literal notranslate"><span class="pre">Trainer.config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.initial_config"><code class="docutils literal notranslate"><span class="pre">Trainer.initial_config</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.pipeline_builder"><code class="docutils literal notranslate"><span class="pre">Trainer.pipeline_builder</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.training_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.training_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.validation_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.validation_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.training_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.training_viz_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.validation_viz_pipeline"><code class="docutils literal notranslate"><span class="pre">Trainer.validation_viz_pipeline</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.optimization_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.optimization_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.output_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.output_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.visualization_callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.visualization_callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.run_path"><code class="docutils literal notranslate"><span class="pre">Trainer.run_path</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.callbacks"><code class="docutils literal notranslate"><span class="pre">Trainer.callbacks</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.cleanup"><code class="docutils literal notranslate"><span class="pre">Trainer.cleanup()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.evaluate"><code class="docutils literal notranslate"><span class="pre">Trainer.evaluate()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.from_config"><code class="docutils literal notranslate"><span class="pre">Trainer.from_config()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.input_keys"><code class="docutils literal notranslate"><span class="pre">Trainer.input_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.keras_model"><code class="docutils literal notranslate"><span class="pre">Trainer.keras_model</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.output_keys"><code class="docutils literal notranslate"><span class="pre">Trainer.output_keys</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.package"><code class="docutils literal notranslate"><span class="pre">Trainer.package()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.setup"><code class="docutils literal notranslate"><span class="pre">Trainer.setup()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.Trainer.train"><code class="docutils literal notranslate"><span class="pre">Trainer.train()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.create_trainer_using_cli"><code class="docutils literal notranslate"><span class="pre">create_trainer_using_cli()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.get_timestamp"><code class="docutils literal notranslate"><span class="pre">get_timestamp()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.main"><code class="docutils literal notranslate"><span class="pre">main()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.sanitize_scope_name"><code class="docutils literal notranslate"><span class="pre">sanitize_scope_name()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_checkpointing"><code class="docutils literal notranslate"><span class="pre">setup_checkpointing()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_losses"><code class="docutils literal notranslate"><span class="pre">setup_losses()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_metrics"><code class="docutils literal notranslate"><span class="pre">setup_metrics()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_new_run_folder"><code class="docutils literal notranslate"><span class="pre">setup_new_run_folder()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_optimization_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_optimization_callbacks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_optimizer"><code class="docutils literal notranslate"><span class="pre">setup_optimizer()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_output_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_output_callbacks()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_tensorboard"><code class="docutils literal notranslate"><span class="pre">setup_tensorboard()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_visualization"><code class="docutils literal notranslate"><span class="pre">setup_visualization()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.training.setup_zmq_callbacks"><code class="docutils literal notranslate"><span class="pre">setup_zmq_callbacks()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -896,7 +1090,7 @@ <h1>sleap.nn.training</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.utils.html b/api/sleap.nn.utils.html
index d29f1e540..9dbd2b1b2 100644
--- a/api/sleap.nn.utils.html
+++ b/api/sleap.nn.utils.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.utils &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.utils &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.utils';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.utils.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,18 @@ <h1>sleap.nn.utils</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.compute_iou"><code class="docutils literal notranslate"><span class="pre">compute_iou()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.group_array"><code class="docutils literal notranslate"><span class="pre">group_array()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.match_points"><code class="docutils literal notranslate"><span class="pre">match_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.reset_input_layer"><code class="docutils literal notranslate"><span class="pre">reset_input_layer()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.tf_linear_sum_assignment"><code class="docutils literal notranslate"><span class="pre">tf_linear_sum_assignment()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,20 +330,20 @@ <h1>sleap.nn.utils</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.utils">
-<span id="sleap-nn-utils"></span><h1>sleap.nn.utils<a class="headerlink" href="#module-sleap.nn.utils" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-utils"></span><h1>sleap.nn.utils<a class="headerlink" href="#module-sleap.nn.utils" title="Permalink to this heading">#</a></h1>
 <p>This module contains generic utilities used for training and inference.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.utils.compute_iou">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">compute_iou</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bbox1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bbox2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/utils.py#L45-L76"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.compute_iou" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">compute_iou</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bbox1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bbox2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">float</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/utils.py#L45-L76"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.compute_iou" title="Permalink to this definition">#</a></dt>
 <dd><p>Computes the intersection over union for a pair of bounding boxes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>bbox1</strong> – Bounding box specified by corner coordinates [y1, x1, y2, x2].</p></li>
 <li><p><strong>bbox2</strong> – Bounding box specified by corner coordinates [y1, x1, y2, x2].</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A float scalar calculated as the ratio between the areas of the intersection
 and the union of the two bounding boxes.</p>
 </dd>
@@ -340,10 +352,10 @@ <h1>sleap.nn.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.utils.group_array">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">group_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">groups</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/utils.py#L10-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.group_array" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">group_array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">X</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">groups</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/utils.py#L10-L42"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.group_array" title="Permalink to this definition">#</a></dt>
 <dd><p>Groups an array into a dictionary keyed by a grouping vector.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>X</strong> – Numpy array with length n along the specified axis.</p></li>
 <li><p><strong>groups</strong> – Vector of n values denoting the group that each slice of X should be
@@ -353,7 +365,7 @@ <h1>sleap.nn.utils</h1>
 to the length of groups.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary with keys mapping each unique value in groups to a subset of X.</p>
 </dd>
 </dl>
@@ -369,16 +381,16 @@ <h1>sleap.nn.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.utils.match_points">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">match_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">tensorflow.python.framework.ops.Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/utils.py#L101-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.match_points" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">match_points</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Tensor</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tensor</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/utils.py#L101-L128"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.match_points" title="Permalink to this definition">#</a></dt>
 <dd><p>Match closest points across two sets.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>points1</strong> – A set of reference points of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">D)</span></code>.</p></li>
 <li><p><strong>points2</strong> – A set of unmatched points of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(M,</span> <span class="pre">D)</span></code>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(inds1,</span> <span class="pre">inds2)</span></code> with the indices of matching points across the two
 lists of points.</p>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">inds1</span></code> is a <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.int32</span></code> tensor of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n,)</span></code> where <code class="xref py py-obj docutils literal notranslate"><span class="pre">n</span> <span class="pre">&lt;=</span> <span class="pre">N</span></code> that indexes into
@@ -397,17 +409,17 @@ <h1>sleap.nn.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.utils.reset_input_layer">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">reset_input_layer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">keras.engine.training.Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/utils.py#L131-L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.reset_input_layer" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">reset_input_layer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">keras_model</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/utils.py#L131-L162"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.utils.reset_input_layer" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns a copy of <code class="xref py py-obj docutils literal notranslate"><span class="pre">keras_model</span></code> with input shape reset to <code class="xref py py-obj docutils literal notranslate"><span class="pre">new_shape</span></code>.</p>
 <p>This method was modified from <a class="reference external" href="https://stackoverflow.com/a/58485055">https://stackoverflow.com/a/58485055</a>.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>keras_model</strong> – <code class="xref py py-obj docutils literal notranslate"><span class="pre">tf.keras.Model</span></code> to return a copy of (with input shape reset).</p></li>
 <li><p><strong>new_shape</strong> – Shape of the returned model’s input layer.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A copy of <code class="xref py py-obj docutils literal notranslate"><span class="pre">keras_model</span></code> with input shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">new_shape</span></code>.</p>
 </dd>
 </dl>
@@ -415,14 +427,14 @@ <h1>sleap.nn.utils</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.utils.tf_linear_sum_assignment">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">tf_linear_sum_assignment</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">tensorflow.python.framework.ops.Tensor</span></span></span><a class="headerlink" href="#sleap.nn.utils.tf_linear_sum_assignment" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.utils.</span></span><span class="sig-name descname"><span class="pre">tf_linear_sum_assignment</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cost_matrix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tensor</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tensor</span></span></span><a class="headerlink" href="#sleap.nn.utils.tf_linear_sum_assignment" title="Permalink to this definition">#</a></dt>
 <dd><p>Run <code class="xref py py-obj docutils literal notranslate"><span class="pre">linear_sum_assignment</span></code> as a TensorFlow function.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>cost_matrix</strong> – Cost matrix of shape <code class="xref py py-obj docutils literal notranslate"><span class="pre">(n_src,</span> <span class="pre">n_dst)</span></code>. Make sure to replace <code class="xref py py-obj docutils literal notranslate"><span class="pre">NaN`s</span>
 <span class="pre">with</span> <span class="pre">`np.inf</span></code>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A tuple of <code class="xref py py-obj docutils literal notranslate"><span class="pre">(row,</span> <span class="pre">col)</span></code> with the indices of the optimal assignments.</p>
 </dd>
 </dl>
@@ -452,6 +464,24 @@ <h1>sleap.nn.utils</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.compute_iou"><code class="docutils literal notranslate"><span class="pre">compute_iou()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.group_array"><code class="docutils literal notranslate"><span class="pre">group_array()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.match_points"><code class="docutils literal notranslate"><span class="pre">match_points()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.reset_input_layer"><code class="docutils literal notranslate"><span class="pre">reset_input_layer()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.utils.tf_linear_sum_assignment"><code class="docutils literal notranslate"><span class="pre">tf_linear_sum_assignment()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -470,7 +500,7 @@ <h1>sleap.nn.utils</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.nn.viz.html b/api/sleap.nn.viz.html
index df6356361..eb83c299b 100644
--- a/api/sleap.nn.viz.html
+++ b/api/sleap.nn.viz.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.nn.viz &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.nn.viz &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.nn.viz';</script>
     <link rel="canonical" href="/develop/api/sleap.nn.viz.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,21 @@ <h1>sleap.nn.viz</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.generate_skeleton_preview_image"><code class="docutils literal notranslate"><span class="pre">generate_skeleton_preview_image()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.imgfig"><code class="docutils literal notranslate"><span class="pre">imgfig()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_confmaps"><code class="docutils literal notranslate"><span class="pre">plot_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_img"><code class="docutils literal notranslate"><span class="pre">plot_img()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_instance"><code class="docutils literal notranslate"><span class="pre">plot_instance()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_instances"><code class="docutils literal notranslate"><span class="pre">plot_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_pafs"><code class="docutils literal notranslate"><span class="pre">plot_pafs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_peaks"><code class="docutils literal notranslate"><span class="pre">plot_peaks()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,21 +333,21 @@ <h1>sleap.nn.viz</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.nn.viz">
-<span id="sleap-nn-viz"></span><h1>sleap.nn.viz<a class="headerlink" href="#module-sleap.nn.viz" title="Permalink to this headline">#</a></h1>
+<span id="sleap-nn-viz"></span><h1>sleap.nn.viz<a class="headerlink" href="#module-sleap.nn.viz" title="Permalink to this heading">#</a></h1>
 <p>Visualization and plotting utilities.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.generate_skeleton_preview_image">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">generate_skeleton_preview_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">sleap.instance.Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">square_bb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thumbnail_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(128,</span> <span class="pre">128)</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bytes</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L305-L386"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.generate_skeleton_preview_image" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">generate_skeleton_preview_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="sleap.instance.html#sleap.instance.Instance" title="sleap.instance.Instance"><span class="pre">Instance</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">square_bb</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thumbnail_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(128,</span> <span class="pre">128)</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bytes</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L305-L386"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.generate_skeleton_preview_image" title="Permalink to this definition">#</a></dt>
 <dd><p>Generate preview image for skeleton based on given instance.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>instance</strong> – A <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap.Instance</span></code> object for which to generate the preview image from.</p></li>
 <li><p><strong>square_bb</strong> – A boolean flag for whether or not the preview image should be a square image</p></li>
 <li><p><strong>thumbnail_size</strong> – A tuple of (w,h) for what the size of the thumbnail image should be</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A byte string encoding of the preview image.</p>
 </dd>
 </dl>
@@ -340,10 +355,10 @@ <h1>sleap.nn.viz</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.imgfig">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">imgfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">6</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dpi</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">72</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">matplotlib.figure.Figure</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L14-L39"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.imgfig" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">imgfig</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">size</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Tuple</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">6</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dpi</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">72</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Figure</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L14-L39"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.imgfig" title="Permalink to this definition">#</a></dt>
 <dd><p>Create a tight figure for image plotting.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>size</strong> – Scalar or 2-tuple specifying the (width, height) of the figure in inches.
 If scalar, will assume equal width and height.</p></li>
@@ -352,7 +367,7 @@ <h1>sleap.nn.viz</h1>
 increasing the size of the plot at the same DPI.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A matplotlib.figure.Figure to use for plotting.</p>
 </dd>
 </dl>
@@ -360,37 +375,37 @@ <h1>sleap.nn.viz</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_confmaps">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L81-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_confmaps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_confmaps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">confmaps</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L81-L96"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_confmaps" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot confidence maps reduced over channels.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_img">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_img</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dpi</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">72</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">matplotlib.figure.Figure</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L42-L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_img" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_img</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dpi</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">72</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Figure</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L42-L78"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_img" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot an image in a tight figure.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_instance">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_by_node</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">lw</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ms</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bbox</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L192-L255"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_instance" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_instance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instance</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_by_node</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">lw</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ms</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bbox</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L192-L255"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_instance" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot a single instance with edge coloring.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_instances">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_by_track</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L258-L295"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_instances" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_instances</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">instances</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color_by_track</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tracks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L258-L295"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_instances" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot a list of instances with identity coloring.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_pafs">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L143-L189"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_pafs" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_pafs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pafs</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stride</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">4.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cmap</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L143-L189"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_pafs" title="Permalink to this definition">#</a></dt>
 <dd><p>Quiver plot for a single frame of pafs.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.nn.viz.plot_peaks">
-<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pts_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">numpy.ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pts_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">numpy.ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paired</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/nn/viz.py#L99-L140"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_peaks" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.nn.viz.</span></span><span class="sig-name descname"><span class="pre">plot_peaks</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pts_gt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pts_pr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">ndarray</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">paired</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/nn/viz.py#L99-L140"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.nn.viz.plot_peaks" title="Permalink to this definition">#</a></dt>
 <dd><p>Plot ground truth and detected peaks.</p>
 </dd></dl>
 
@@ -418,6 +433,27 @@ <h1>sleap.nn.viz</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.generate_skeleton_preview_image"><code class="docutils literal notranslate"><span class="pre">generate_skeleton_preview_image()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.imgfig"><code class="docutils literal notranslate"><span class="pre">imgfig()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_confmaps"><code class="docutils literal notranslate"><span class="pre">plot_confmaps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_img"><code class="docutils literal notranslate"><span class="pre">plot_img()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_instance"><code class="docutils literal notranslate"><span class="pre">plot_instance()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_instances"><code class="docutils literal notranslate"><span class="pre">plot_instances()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_pafs"><code class="docutils literal notranslate"><span class="pre">plot_pafs()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.nn.viz.plot_peaks"><code class="docutils literal notranslate"><span class="pre">plot_peaks()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -436,7 +472,7 @@ <h1>sleap.nn.viz</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.skeleton.html b/api/sleap.skeleton.html
index 77e84d6f8..3f53bd4dd 100644
--- a/api/sleap.skeleton.html
+++ b/api/sleap.skeleton.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.skeleton &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.skeleton &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.skeleton';</script>
     <link rel="canonical" href="/develop/api/sleap.skeleton.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,88 @@ <h1>sleap.skeleton</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.EdgeType"><code class="docutils literal notranslate"><span class="pre">EdgeType</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node"><code class="docutils literal notranslate"><span class="pre">Node</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.name"><code class="docutils literal notranslate"><span class="pre">Node.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.weight"><code class="docutils literal notranslate"><span class="pre">Node.weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.as_node"><code class="docutils literal notranslate"><span class="pre">Node.as_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.from_names"><code class="docutils literal notranslate"><span class="pre">Node.from_names()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.matches"><code class="docutils literal notranslate"><span class="pre">Node.matches()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton"><code class="docutils literal notranslate"><span class="pre">Skeleton</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton._skeleton_idx"><code class="docutils literal notranslate"><span class="pre">Skeleton._skeleton_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.preview_image"><code class="docutils literal notranslate"><span class="pre">Skeleton.preview_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.description"><code class="docutils literal notranslate"><span class="pre">Skeleton.description</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton._is_template"><code class="docutils literal notranslate"><span class="pre">Skeleton._is_template</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.clear_edges"><code class="docutils literal notranslate"><span class="pre">Skeleton.clear_edges()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_to_index"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_to_index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edges"><code class="docutils literal notranslate"><span class="pre">Skeleton.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edges_full"><code class="docutils literal notranslate"><span class="pre">Skeleton.edges_full</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_neighbors"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_neighbors()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_unique_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_unique_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_dict"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_names_and_edge_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_names_and_edge_inds()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.get_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.get_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.get_symmetry_name"><code class="docutils literal notranslate"><span class="pre">Skeleton.get_symmetry_name()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.graph"><code class="docutils literal notranslate"><span class="pre">Skeleton.graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.graph_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.graph_symmetry</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.is_arborescence"><code class="docutils literal notranslate"><span class="pre">Skeleton.is_arborescence</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.is_template"><code class="docutils literal notranslate"><span class="pre">Skeleton.is_template</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_all_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_all_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_mat"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_mat()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.make_cattr"><code class="docutils literal notranslate"><span class="pre">Skeleton.make_cattr()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.matches"><code class="docutils literal notranslate"><span class="pre">Skeleton.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.name"><code class="docutils literal notranslate"><span class="pre">Skeleton.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.node_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.node_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.node_to_index"><code class="docutils literal notranslate"><span class="pre">Skeleton.node_to_index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.relabel_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.relabel_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.relabel_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.relabel_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.rename_skeleton"><code class="docutils literal notranslate"><span class="pre">Skeleton.rename_skeleton()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_all_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_all_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetric_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetric_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetries"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetries</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetries_full"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetries_full</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetry_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetry_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.to_dict"><code class="docutils literal notranslate"><span class="pre">Skeleton.to_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.to_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.to_json()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder.decode"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder.decode()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder.decode_preview_image"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder.decode_preview_image()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonEncoder"><code class="docutils literal notranslate"><span class="pre">SkeletonEncoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonEncoder.encode"><code class="docutils literal notranslate"><span class="pre">SkeletonEncoder.encode()</span></code></a></li>
+</ul>
+</li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,14 +400,14 @@ <h1>sleap.skeleton</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.skeleton">
-<span id="sleap-skeleton"></span><h1>sleap.skeleton<a class="headerlink" href="#module-sleap.skeleton" title="Permalink to this headline">#</a></h1>
+<span id="sleap-skeleton"></span><h1>sleap.skeleton<a class="headerlink" href="#module-sleap.skeleton" title="Permalink to this heading">#</a></h1>
 <p>Implementation of skeleton data structure and API.</p>
 <p>This module implements and API for creating animal skeletons. The goal
 is to provide a common interface for defining the parts of the animal,
 their connection to each other, and needed meta-data.</p>
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.skeleton.EdgeType">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">EdgeType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L31-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.EdgeType" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">EdgeType</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L31-L45"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.EdgeType" title="Permalink to this definition">#</a></dt>
 <dd><p>Type of edge in the skeleton graph.</p>
 <p>The skeleton graph can store different types of edges to represent
 different things. All edges must specify one or more of the
@@ -342,7 +424,7 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.skeleton.Node">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">Node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L49-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">Node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">weight</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">float</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">1.0</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L49-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node" title="Permalink to this definition">#</a></dt>
 <dd><p>This class represents node in the skeleton graph, i.e., a body part.</p>
 <p>Note: Nodes can exist without being part of a skeleton.</p>
 <dl class="py attribute">
@@ -350,7 +432,7 @@ <h1>sleap.skeleton</h1>
 <span class="sig-name descname"><span class="pre">name</span></span><a class="headerlink" href="#sleap.skeleton.Node.name" title="Permalink to this definition">#</a></dt>
 <dd><p>String name of the node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>str</p>
 </dd>
 </dl>
@@ -361,7 +443,7 @@ <h1>sleap.skeleton</h1>
 <span class="sig-name descname"><span class="pre">weight</span></span><a class="headerlink" href="#sleap.skeleton.Node.weight" title="Permalink to this definition">#</a></dt>
 <dd><p>Weight of the node (not currently used).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>float</p>
 </dd>
 </dl>
@@ -369,25 +451,25 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Node.as_node">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">as_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L70-L73"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.as_node" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">as_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L70-L73"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.as_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert given <code class="xref py py-obj docutils literal notranslate"><span class="pre">node</span></code> to <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> object (if not already).</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Node.from_names">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_names</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L62-L68"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.from_names" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_names</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L62-L68"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.from_names" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert list of node names to list of nodes objects.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Node.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L75-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L75-L85"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Node.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Check whether all attributes match between two nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> to compare to this one.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if all attributes match, False otherwise.</p>
 </dd>
 </dl>
@@ -397,7 +479,7 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">Skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L88-L1279"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">Skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L584-L1775"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton" title="Permalink to this definition">#</a></dt>
 <dd><p>The main object for representing animal skeletons.</p>
 <p>The skeleton represents the constituent parts of the animal whose
 pose is being estimated.</p>
@@ -414,7 +496,7 @@ <h1>sleap.skeleton</h1>
 <dd><p>A byte string containing an encoded preview image for the
 skeleton. Used only for templates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[bytes]</p>
 </dd>
 </dl>
@@ -425,7 +507,7 @@ <h1>sleap.skeleton</h1>
 <span class="sig-name descname"><span class="pre">description</span></span><a class="headerlink" href="#sleap.skeleton.Skeleton.description" title="Permalink to this definition">#</a></dt>
 <dd><p>A text description of the skeleton. Used only for templates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>Optional[str]</p>
 </dd>
 </dl>
@@ -436,7 +518,7 @@ <h1>sleap.skeleton</h1>
 <span class="sig-name descname"><span class="pre">_is_template</span></span><a class="headerlink" href="#sleap.skeleton.Skeleton._is_template" title="Permalink to this definition">#</a></dt>
 <dd><p>Whether this skeleton is a template. Used only for templates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Type</dt>
+<dt class="field-odd">Type<span class="colon">:</span></dt>
 <dd class="field-odd"><p>bool</p>
 </dd>
 </dl>
@@ -444,20 +526,20 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.add_edge">
-<span class="sig-name descname"><span class="pre">add_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L589-L639"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_edge" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1085-L1135"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_edge" title="Permalink to this definition">#</a></dt>
 <dd><p>Add an edge between two nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>source</strong> – The name of the source node.</p></li>
 <li><p><strong>destination</strong> – The name of the destination node.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If source or destination nodes cannot be found,
     or if edge already exists between those nodes.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None.</p>
 </dd>
 </dl>
@@ -465,14 +547,14 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.add_node">
-<span class="sig-name descname"><span class="pre">add_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L505-L521"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_node" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1001-L1017"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Add a node representing an animal part to the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>name</strong> – The name of the node to add to the skeleton.
 This name must be unique within the skeleton.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If name is not unique.</p>
 </dd>
 </dl>
@@ -480,10 +562,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.add_nodes">
-<span class="sig-name descname"><span class="pre">add_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L523-L530"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_nodes" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name_list</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1019-L1026"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Add a list of nodes representing animal parts to the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>name_list</strong> – List of strings representing the nodes.</p>
 </dd>
 </dl>
@@ -491,22 +573,22 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.add_symmetry">
-<span class="sig-name descname"><span class="pre">add_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L691-L729"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_symmetry" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">add_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1187-L1225"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.add_symmetry" title="Permalink to this definition">#</a></dt>
 <dd><p>Specify that two parts (nodes) in skeleton are symmetrical.</p>
 <p>Certain parts of an animal body can be related as symmetrical
 parts in a pair. For example, left and right hands of a person.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>node1</strong> – The name of the first part in the symmetric pair</p></li>
 <li><p><strong>node2</strong> – The name of the second part in the symmetric pair</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If node1 and node2 match, or if there is already
     a symmetry between them.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -514,26 +596,26 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.clear_edges">
-<span class="sig-name descname"><span class="pre">clear_edges</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L686-L689"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.clear_edges" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">clear_edges</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1182-L1185"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.clear_edges" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete all edges in skeleton.</p>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.delete_edge">
-<span class="sig-name descname"><span class="pre">delete_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L641-L684"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_edge" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">delete_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1137-L1180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_edge" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete an edge between two nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>source</strong> – The name of the source node.</p></li>
 <li><p><strong>destination</strong> – The name of the destination node.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If skeleton does not have either source node,
     destination node, or edge between them.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -541,18 +623,18 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.delete_node">
-<span class="sig-name descname"><span class="pre">delete_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L532-L553"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_node" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">delete_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1028-L1049"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Remove a node from the skeleton.</p>
 <p>The method removes a node from the skeleton and any edge that is
 connected to it.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>name</strong> – The name of the node to remove</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If node cannot be found.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -560,19 +642,19 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.delete_symmetry">
-<span class="sig-name descname"><span class="pre">delete_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L731-L760"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_symmetry" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">delete_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1227-L1256"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.delete_symmetry" title="Permalink to this definition">#</a></dt>
 <dd><p>Delete a previously established symmetry between two nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>node1</strong> – One node (by <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> object or name) in symmetric pair.</p></li>
 <li><p><strong>node2</strong> – Other node in symmetric pair.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If there’s no symmetry between node1 and node2.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -583,7 +665,7 @@ <h1>sleap.skeleton</h1>
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edge_inds</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edge_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of edges as node indices.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list of (src_node_ind, dst_node_ind), where indices are subscripts into
 the Skeleton.nodes list.</p>
 </dd>
@@ -595,7 +677,7 @@ <h1>sleap.skeleton</h1>
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edge_names</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edge_names" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of edge name tuples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>list of (src_node.name, dst_node.name)</p>
 </dd>
 </dl>
@@ -603,16 +685,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.edge_to_index">
-<span class="sig-name descname"><span class="pre">edge_to_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L495-L503"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.edge_to_index" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">edge_to_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">destination</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L991-L999"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.edge_to_index" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the index of edge from source to destination.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.edges">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edges</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edges" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edges</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edges" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of edge tuples.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>list of (src_node, dst_node)</p>
 </dd>
 </dl>
@@ -620,10 +702,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.edges_full">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edges_full</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edges_full" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">edges_full</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.edges_full" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of edge tuples with keys and attributes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>list of (src_node, dst_node, key, attributes)</p>
 </dd>
 </dl>
@@ -631,13 +713,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.find_neighbors">
-<span class="sig-name descname"><span class="pre">find_neighbors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L577-L587"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_neighbors" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_neighbors</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1073-L1083"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_neighbors" title="Permalink to this definition">#</a></dt>
 <dd><p>Find nodes that are predecessors or successors from a node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>node</strong> – Name or <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> instance.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> objects that are neighbors to the node.</p>
 </dd>
 </dl>
@@ -645,13 +727,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.find_node">
-<span class="sig-name descname"><span class="pre">find_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L555-L575"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_node" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">find_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1051-L1071"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Find node in skeleton by name of node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>name</strong> – The name of the <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-class docutils literal notranslate"><span class="pre">Node</span></code></a> (or a <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-class docutils literal notranslate"><span class="pre">Node</span></code></a>)</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a>, or None if no match found</p>
 </dd>
 </dl>
@@ -659,13 +741,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.find_unique_nodes">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">find_unique_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L244-L254"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_unique_nodes" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">find_unique_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L740-L750"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.find_unique_nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Find all unique nodes from a list of skeletons.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>skeletons</strong> – The list of skeletons.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A list of unique <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> objects.</p>
 </dd>
 </dl>
@@ -673,10 +755,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.from_dict">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L942-L962"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_dict" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1438-L1458"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Create skeleton from dict; used for loading from JSON.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>d</strong> – the dict from which to deserialize</p></li>
 <li><p><strong>node_to_idx</strong> – optional dict which maps <code class="xref py py-class docutils literal notranslate"><span class="pre">Node`sto</span> <span class="pre">index</span>
@@ -689,7 +771,7 @@ <h1>sleap.skeleton</h1>
 the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p>
 </dd>
 </dl>
@@ -697,10 +779,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.from_json">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1056-L1087"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_json" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1552-L1583"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Instantiate <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a> from JSON string.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>json_str</strong> – The JSON encoded Skeleton.</p></li>
 <li><p><strong>idx_to_node</strong> – optional dict which maps an int (indexing a
@@ -711,7 +793,7 @@ <h1>sleap.skeleton</h1>
 left in the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a> when it was saved.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Skeleton</span></code></a> object decoded from the JSON.</p>
 </dd>
 </dl>
@@ -719,17 +801,17 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.from_names_and_edge_inds">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_names_and_edge_inds</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L964-L984"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_names_and_edge_inds" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">from_names_and_edge_inds</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_inds</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1460-L1480"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.from_names_and_edge_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>Create skeleton from a list of node names and edge indices.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>node_names</strong> – List of strings defining the nodes.</p></li>
 <li><p><strong>edge_inds</strong> – List of tuples in the form (src_node_ind, dst_node_ind). If not
 specified, the resulting skeleton will have no edges.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The instantiated skeleton.</p>
 </dd>
 </dl>
@@ -737,16 +819,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.get_symmetry">
-<span class="sig-name descname"><span class="pre">get_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L762-L787"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.get_symmetry" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_symmetry</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1258-L1283"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.get_symmetry" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the node symmetric with the specified node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>node</strong> – Node (by <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> object or name) to query.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If node has more than one symmetry.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The symmetric <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-class docutils literal notranslate"><span class="pre">Node</span></code></a>, None if no symmetry.</p>
 </dd>
 </dl>
@@ -754,13 +836,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.get_symmetry_name">
-<span class="sig-name descname"><span class="pre">get_symmetry_name</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L789-L799"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.get_symmetry_name" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">get_symmetry_name</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1285-L1295"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.get_symmetry_name" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the name of the node symmetric with the specified node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>node</strong> – Node (by <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> object or name) to query.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Name of symmetric node, None if no symmetry.</p>
 </dd>
 </dl>
@@ -780,16 +862,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.has_edge">
-<span class="sig-name descname"><span class="pre">has_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dest_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L901-L917"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_edge" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">has_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dest_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1397-L1413"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_edge" title="Permalink to this definition">#</a></dt>
 <dd><p>Check whether the skeleton has an edge.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>source_name</strong> – The name of the source node for the edge.</p></li>
 <li><p><strong>dest_name</strong> – The name of the destination node for the edge.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True is yes, False if no.</p>
 </dd>
 </dl>
@@ -797,13 +879,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.has_node">
-<span class="sig-name descname"><span class="pre">has_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L870-L881"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_node" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">has_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1366-L1377"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Check whether the skeleton has a node.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>name</strong> – The name of the node to check for.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True for yes, False for no.</p>
 </dd>
 </dl>
@@ -811,13 +893,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.has_nodes">
-<span class="sig-name descname"><span class="pre">has_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L883-L899"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_nodes" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">has_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">names</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1379-L1395"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.has_nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Check whether the skeleton has a list of nodes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>names</strong> – The list names of the nodes to check for.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True for yes, False for no.</p>
 </dd>
 </dl>
@@ -841,10 +923,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.load_all_hdf5">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_all_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.files.File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1137-L1164"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_all_hdf5" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_all_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1633-L1660"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_all_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Load all skeletons found in the HDF5 file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>file</strong> – The file name or open h5py.File</p></li>
 <li><p><strong>return_dict</strong> – Whether the the return value should be a dict
@@ -853,7 +935,7 @@ <h1>sleap.skeleton</h1>
 return just a list of the skeletons.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The skeleton instances stored in the HDF5 file.
 Either in List or Dict form.</p>
 </dd>
@@ -862,16 +944,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.load_hdf5">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.files.File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1117-L1135"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_hdf5" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1613-L1631"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a specific skeleton (by name) from the HDF5 file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>file</strong> – The file name or open h5py.File</p></li>
 <li><p><strong>name</strong> – The name of the skeleton.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The specified <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Skeleton</span></code></a> instance stored in the HDF5 file.</p>
 </dd>
 </dl>
@@ -879,12 +961,12 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.load_json">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1089-L1115"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_json" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1585-L1611"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Load a skeleton from a JSON file.</p>
 <p>This method will load the Skeleton from JSON file saved
 with; <a class="reference internal" href="#sleap.skeleton.Skeleton.save_json" title="sleap.skeleton.Skeleton.save_json"><code class="xref py py-meth docutils literal notranslate"><span class="pre">save_json()</span></code></a></p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The file that contains the JSON.</p></li>
 <li><p><strong>idx_to_node</strong> – optional dict which maps an int (indexing a
@@ -895,7 +977,7 @@ <h1>sleap.skeleton</h1>
 left in the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a> when it was saved.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Skeleton</span></code></a> object stored in the JSON filename.</p>
 </dd>
 </dl>
@@ -903,15 +985,15 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.load_mat">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_mat</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1241-L1273"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_mat" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_mat</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1737-L1769"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.load_mat" title="Permalink to this definition">#</a></dt>
 <dd><p>Load the skeleton from a Matlab MAT file.</p>
 <p>This is to support backwards compatibility with old LEAP
 MATLAB code and datasets.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>filename</strong> – The name of the skeleton file</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>An instance of the skeleton.</p>
 </dd>
 </dl>
@@ -919,15 +1001,15 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.make_cattr">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_cattr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">cattr.converters.Converter</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L256-L283"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.make_cattr" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">make_cattr</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">idx_to_node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Converter</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L752-L779"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.make_cattr" title="Permalink to this definition">#</a></dt>
 <dd><p>Make cattr.Convert() for <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p>
 <p>Make a cattr.Converter() that registers structure/unstructure
 hooks for Skeleton objects to handle serialization of skeletons.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>idx_to_node</strong> – A dict that maps node index to Node objects.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A cattr.Converter() instance for skeleton serialization
 and deserialization.</p>
 </dd>
@@ -936,13 +1018,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.matches">
-<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L153-L179"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.matches" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">matches</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">other</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L649-L675"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.matches" title="Permalink to this definition">#</a></dt>
 <dd><p>Compare this <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Skeleton</span></code></a> to another, ignoring name and node identities.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>other</strong> – The other skeleton.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">True</span></code> if the skeleton graphs are isomorphic and node names.</p>
 </dd>
 </dl>
@@ -953,7 +1035,7 @@ <h1>sleap.skeleton</h1>
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">name</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">str</span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.name" title="Permalink to this definition">#</a></dt>
 <dd><p>Get the name of the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A string representing the name of the skeleton.</p>
 </dd>
 </dl>
@@ -964,7 +1046,7 @@ <h1>sleap.skeleton</h1>
 <em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">node_names</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.node_names" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of node names.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list of node names.</p>
 </dd>
 </dl>
@@ -972,16 +1054,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.node_to_index">
-<span class="sig-name descname"><span class="pre">node_to_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L477-L493"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.node_to_index" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">node_to_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L973-L989"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.node_to_index" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the index of the node, accepts either <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> or name.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>node</strong> – The name of the node or the Node object.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError if node cannot be found in skeleton.</strong> – </p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The index of the node in the graph.</p>
 </dd>
 </dl>
@@ -989,10 +1071,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.nodes">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.nodes" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">nodes</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of :class:<a href="#id1"><span class="problematic" id="id2">`</span></a>Node`s.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>A list of :class:<a href="#id3"><span class="problematic" id="id4">`</span></a>Node`s</p>
 </dd>
 </dl>
@@ -1000,16 +1082,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.relabel_node">
-<span class="sig-name descname"><span class="pre">relabel_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L835-L845"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.relabel_node" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">relabel_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">old_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">new_name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1331-L1341"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.relabel_node" title="Permalink to this definition">#</a></dt>
 <dd><p>Relabel a single node to a new name.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>old_name</strong> – The old name of the node.</p></li>
 <li><p><strong>new_name</strong> – The new name of the node.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -1017,17 +1099,17 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.relabel_nodes">
-<span class="sig-name descname"><span class="pre">relabel_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mapping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L847-L866"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.relabel_nodes" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">relabel_nodes</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mapping</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1343-L1362"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.relabel_nodes" title="Permalink to this definition">#</a></dt>
 <dd><p>Relabel the nodes of the skeleton.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>mapping</strong> – A dictionary with the old labels as keys and new
 labels as values. A partial mapping is allowed.</p>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If node already present with one of the new names.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -1035,7 +1117,7 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.rename_skeleton">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">rename_skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L320-L340"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.rename_skeleton" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">rename_skeleton</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">skeleton</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L816-L836"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.rename_skeleton" title="Permalink to this definition">#</a></dt>
 <dd><p>Make copy of skeleton with new name.</p>
 <p>This property is immutable because it is used to hash skeletons.
 If you want to rename a Skeleton you must use this class method.</p>
@@ -1044,13 +1126,13 @@ <h1>sleap.skeleton</h1>
 </pre></div>
 </div>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>skeleton</strong> – The skeleton to copy.</p></li>
 <li><p><strong>name</strong> – The name of the new skeleton.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A new deep copied skeleton with the changed name.</p>
 </dd>
 </dl>
@@ -1058,21 +1140,21 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.save_all_hdf5">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">save_all_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.files.File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1175-L1201"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_all_hdf5" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">save_all_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">File</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">skeletons</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1671-L1697"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_all_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Convenience method to save a list of skeletons to HDF5 file.</p>
 <p>Skeletons are saved as attributes of a /skeleton group in the
 file.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>file</strong> – The filename or the open h5py.File object.</p></li>
 <li><p><strong>skeletons</strong> – The list of skeletons to save.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If multiple skeletons have the same name.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -1080,13 +1162,13 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.save_hdf5">
-<span class="sig-name descname"><span class="pre">save_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">h5py._hl.files.File</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1203-L1218"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_hdf5" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">File</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1699-L1714"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Wrapper for HDF5 saving which takes either filename or h5py.File.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>file</strong> – can be filename (string) or <code class="xref py py-obj docutils literal notranslate"><span class="pre">h5py.File</span></code> object</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -1094,11 +1176,11 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.save_json">
-<span class="sig-name descname"><span class="pre">save_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L1030-L1054"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_json" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">save_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1526-L1550"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.save_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Save the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a> as JSON file.</p>
 <p>Output the complete skeleton to a file in JSON format.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename</strong> – The filename to save the JSON to.</p></li>
 <li><p><strong>node_to_idx</strong> – optional dict which maps <code class="xref py py-class docutils literal notranslate"><span class="pre">Node`sto</span> <span class="pre">index</span>
@@ -1111,7 +1193,7 @@ <h1>sleap.skeleton</h1>
 the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -1119,16 +1201,16 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.symmetric_inds">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetric_inds</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">numpy.ndarray</span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetric_inds" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetric_inds</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">ndarray</span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetric_inds" title="Permalink to this definition">#</a></dt>
 <dd><p>Return the symmetric nodes as an array of indices.</p>
 </dd></dl>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.symmetries">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetries</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetries" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetries</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetries" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of all symmetries without duplicates.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>list of (node1, node2)</p>
 </dd>
 </dl>
@@ -1136,12 +1218,12 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py property">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.symmetries_full">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetries_full</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetries_full" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">symmetries_full</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#sleap.skeleton.Skeleton.symmetries_full" title="Permalink to this definition">#</a></dt>
 <dd><p>Get a list of all symmetries with keys and attributes.</p>
 <p>Note: The returned list will contain duplicates (node1, node2)
 and (node2, node1).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>list of (node1, node2, key, attr)</p>
 </dd>
 </dl>
@@ -1155,10 +1237,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.to_dict">
-<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">to_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">obj</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">sleap.skeleton.Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L919-L940"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.to_dict" title="Permalink to this definition">#</a></dt>
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">to_dict</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">obj</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><span class="pre">Skeleton</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1415-L1436"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.to_dict" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert skeleton to dict; used for saving as JSON.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>obj</strong> – the <a href="#id5"><span class="problematic" id="id6">:object:`Skeleton`</span></a> to convert</p></li>
 <li><p><strong>node_to_idx</strong> – optional dict which maps <code class="xref py py-class docutils literal notranslate"><span class="pre">Node`sto</span> <span class="pre">index</span>
@@ -1171,7 +1253,7 @@ <h1>sleap.skeleton</h1>
 the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>dict with data from skeleton</p>
 </dd>
 </dl>
@@ -1179,10 +1261,10 @@ <h1>sleap.skeleton</h1>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="sleap.skeleton.Skeleton.to_json">
-<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">sleap.skeleton.Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/skeleton.py#L986-L1028"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.to_json" title="Permalink to this definition">#</a></dt>
+<span class="sig-name descname"><span class="pre">to_json</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">node_to_idx</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><span class="pre">Node</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L1482-L1524"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.Skeleton.to_json" title="Permalink to this definition">#</a></dt>
 <dd><p>Convert the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a> to a JSON representation.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>node_to_idx</strong> – optional dict which maps <code class="xref py py-class docutils literal notranslate"><span class="pre">Node`sto</span> <span class="pre">index</span>
 <span class="pre">in</span> <span class="pre">some</span> <span class="pre">list.</span> <span class="pre">This</span> <span class="pre">is</span> <span class="pre">used</span> <span class="pre">when</span> <span class="pre">saving</span>
 <span class="pre">:class:`Labels`where</span> <span class="pre">we</span> <span class="pre">want</span> <span class="pre">to</span> <span class="pre">serialize</span> <span class="pre">the</span>
@@ -1192,7 +1274,7 @@ <h1>sleap.skeleton</h1>
 Otherwise, we convert <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-class docutils literal notranslate"><span class="pre">Node</span></code></a> objects with the rest of
 the <a class="reference internal" href="#sleap.skeleton.Skeleton" title="sleap.skeleton.Skeleton"><code class="xref py py-class docutils literal notranslate"><span class="pre">Skeleton</span></code></a>.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A string containing the JSON representation of the skeleton.</p>
 </dd>
 </dl>
@@ -1200,6 +1282,133 @@ <h1>sleap.skeleton</h1>
 
 </dd></dl>
 
+<dl class="py class">
+<dt class="sig sig-object py" id="sleap.skeleton.SkeletonDecoder">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">SkeletonDecoder</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L88-L375"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.SkeletonDecoder" title="Permalink to this definition">#</a></dt>
+<dd><p>Replace jsonpickle.decode with our own decoder.</p>
+<p>This function will decode the following from jsonpickle’s encoded format:</p>
+<dl>
+<dt><a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> objects from</dt><dd><dl class="simple">
+<dt>{</dt><dd><p>“py/object”: “sleap.skeleton.Node”,
+“py/state”: { “py/tuple”: [“thorax1”, 1.0] }</p>
+</dd>
+</dl>
+<p>}</p>
+</dd>
+</dl>
+<p>to <code class="xref py py-obj docutils literal notranslate"><span class="pre">Node(name=&quot;thorax1&quot;,</span> <span class="pre">weight=1.0)</span></code></p>
+<dl>
+<dt><a class="reference internal" href="#sleap.skeleton.EdgeType" title="sleap.skeleton.EdgeType"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EdgeType</span></code></a> objects from</dt><dd><dl>
+<dt>{</dt><dd><dl class="simple">
+<dt>“py/reduce”: [</dt><dd><p>{ “py/type”: “sleap.skeleton.EdgeType” },
+{ “py/tuple”: [1] }</p>
+</dd>
+</dl>
+<p>]</p>
+</dd>
+</dl>
+<p>}</p>
+</dd>
+</dl>
+<p>to <code class="xref py py-obj docutils literal notranslate"><span class="pre">EdgeType(1)</span></code></p>
+<dl>
+<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">bytes</span></code> from</dt><dd><dl class="simple">
+<dt>{</dt><dd><p>“py/b64”: “aVZC…”</p>
+</dd>
+</dl>
+<p>}</p>
+</dd>
+</dl>
+<p>to <code class="xref py py-obj docutils literal notranslate"><span class="pre">b&quot;iVBO...&quot;</span></code></p>
+<dl>
+<dt>and any repeated objects from</dt><dd><dl class="simple">
+<dt>{</dt><dd><p>“py/id”: 1</p>
+</dd>
+</dl>
+<p>}</p>
+</dd>
+</dl>
+<p>to the object with the same reconstruction id (from top to bottom).</p>
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.skeleton.SkeletonDecoder.decode">
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L367-L375"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.SkeletonDecoder.decode" title="Permalink to this definition">#</a></dt>
+<dd><p>Decode the given json string into a dictionary.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
+<dd class="field-odd"><p>A dict with <code class="xref py py-obj docutils literal notranslate"><span class="pre">Node`s,</span> <span class="pre">`EdgeType`s,</span> <span class="pre">and</span> <span class="pre">`bytes</span></code> decoded/reconstructed.</p>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.skeleton.SkeletonDecoder.decode_preview_image">
+<em class="property"><span class="pre">static</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">decode_preview_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_b64</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bytes</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">return_bytes</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">Image</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">bytes</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L318-L338"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.SkeletonDecoder.decode_preview_image" title="Permalink to this definition">#</a></dt>
+<dd><p>Decode a skeleton preview image byte string representation to a <code class="xref py py-obj docutils literal notranslate"><span class="pre">PIL.Image</span></code></p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>img_b64</strong> – a byte string representation of a skeleton preview image</p></li>
+<li><p><strong>return_bytes</strong> – whether to return the decoded image as bytes</p></li>
+</ul>
+</dd>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>Either a PIL.Image of the skeleton preview image or the decoded image as bytes
+(if <code class="xref py py-obj docutils literal notranslate"><span class="pre">return_bytes</span></code> is True).</p>
+</dd>
+</dl>
+</dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="sleap.skeleton.SkeletonEncoder">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">sleap.skeleton.</span></span><span class="sig-name descname"><span class="pre">SkeletonEncoder</span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L378-L581"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.SkeletonEncoder" title="Permalink to this definition">#</a></dt>
+<dd><p>Replace jsonpickle.encode with our own encoder.</p>
+<p>The input is a dictionary containing python objects that need to be encoded as
+JSON strings. The output is a JSON string that represents the input dictionary.</p>
+<dl>
+<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node(name='neck',</span> <span class="pre">weight=1.0)</span></code> =&gt;</dt><dd><dl class="simple">
+<dt>{</dt><dd><p>“py/object”: “sleap.Skeleton.Node”,
+“py/state”: {“py/tuple” [“neck”, 1.0]}</p>
+</dd>
+</dl>
+<p>}</p>
+</dd>
+<dt><code class="xref py py-obj docutils literal notranslate"><span class="pre">&lt;EdgeType.BODY:</span> <span class="pre">1&gt;</span></code> =&gt;</dt><dd><dl class="simple">
+<dt>{“py/reduce”: [</dt><dd><p>{“py/type”: “sleap.Skeleton.EdgeType”},
+{“py/tuple”: [1] }
+]</p>
+</dd>
+</dl>
+<p>}`</p>
+</dd>
+</dl>
+<p>Where <code class="xref py py-obj docutils literal notranslate"><span class="pre">name</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">weight</span></code> are the attributes of the <a class="reference internal" href="#sleap.skeleton.Node" title="sleap.skeleton.Node"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Node</span></code></a> class; weight is always 1.0.
+<a class="reference internal" href="#sleap.skeleton.EdgeType" title="sleap.skeleton.EdgeType"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EdgeType</span></code></a> is an enum with values <code class="xref py py-obj docutils literal notranslate"><span class="pre">BODY</span> <span class="pre">=</span> <span class="pre">1</span></code> and <code class="xref py py-obj docutils literal notranslate"><span class="pre">SYMMETRY</span> <span class="pre">=</span> <span class="pre">2</span></code>.</p>
+<p>See sleap.skeleton.Node and sleap.skeleton.EdgeType.</p>
+<p>If the object has been “seen” before, it will not be encoded as the full JSON string
+but referenced by its <code class="xref py py-obj docutils literal notranslate"><span class="pre">py/id</span></code>, which starts at 1 and indexes the objects in the
+order they are seen so that the second time the first object is used, it will be
+referenced as <code class="xref py py-obj docutils literal notranslate"><span class="pre">{&quot;py/id&quot;:</span> <span class="pre">1}</span></code>.</p>
+<dl class="py method">
+<dt class="sig sig-object py" id="sleap.skeleton.SkeletonEncoder.encode">
+<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">encode</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/skeleton.py#L413-L430"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.skeleton.SkeletonEncoder.encode" title="Permalink to this definition">#</a></dt>
+<dd><p>Encodes the input dictionary as a JSON string.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><p><strong>data</strong> – The data to encode.</p>
+</dd>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
+<dd class="field-even"><p>The JSON string representation of the data.</p>
+</dd>
+<dt class="field-odd">Return type<span class="colon">:</span></dt>
+<dd class="field-odd"><p>json_str</p>
+</dd>
+</dl>
+</dd></dl>
+
+</dd></dl>
+
 </section>
 
 
@@ -1224,6 +1433,94 @@ <h1>sleap.skeleton</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.EdgeType"><code class="docutils literal notranslate"><span class="pre">EdgeType</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node"><code class="docutils literal notranslate"><span class="pre">Node</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.name"><code class="docutils literal notranslate"><span class="pre">Node.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.weight"><code class="docutils literal notranslate"><span class="pre">Node.weight</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.as_node"><code class="docutils literal notranslate"><span class="pre">Node.as_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.from_names"><code class="docutils literal notranslate"><span class="pre">Node.from_names()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Node.matches"><code class="docutils literal notranslate"><span class="pre">Node.matches()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton"><code class="docutils literal notranslate"><span class="pre">Skeleton</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton._skeleton_idx"><code class="docutils literal notranslate"><span class="pre">Skeleton._skeleton_idx</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.preview_image"><code class="docutils literal notranslate"><span class="pre">Skeleton.preview_image</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.description"><code class="docutils literal notranslate"><span class="pre">Skeleton.description</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton._is_template"><code class="docutils literal notranslate"><span class="pre">Skeleton._is_template</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.add_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.add_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.clear_edges"><code class="docutils literal notranslate"><span class="pre">Skeleton.clear_edges()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.delete_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.delete_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edge_to_index"><code class="docutils literal notranslate"><span class="pre">Skeleton.edge_to_index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edges"><code class="docutils literal notranslate"><span class="pre">Skeleton.edges</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.edges_full"><code class="docutils literal notranslate"><span class="pre">Skeleton.edges_full</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_neighbors"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_neighbors()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.find_unique_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.find_unique_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_dict"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.from_names_and_edge_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.from_names_and_edge_inds()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.get_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.get_symmetry()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.get_symmetry_name"><code class="docutils literal notranslate"><span class="pre">Skeleton.get_symmetry_name()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.graph"><code class="docutils literal notranslate"><span class="pre">Skeleton.graph</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.graph_symmetry"><code class="docutils literal notranslate"><span class="pre">Skeleton.graph_symmetry</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_edge"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_edge()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.has_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.has_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.is_arborescence"><code class="docutils literal notranslate"><span class="pre">Skeleton.is_arborescence</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.is_template"><code class="docutils literal notranslate"><span class="pre">Skeleton.is_template</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_all_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_all_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.load_mat"><code class="docutils literal notranslate"><span class="pre">Skeleton.load_mat()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.make_cattr"><code class="docutils literal notranslate"><span class="pre">Skeleton.make_cattr()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.matches"><code class="docutils literal notranslate"><span class="pre">Skeleton.matches()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.name"><code class="docutils literal notranslate"><span class="pre">Skeleton.name</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.node_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.node_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.node_to_index"><code class="docutils literal notranslate"><span class="pre">Skeleton.node_to_index()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.nodes</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.relabel_node"><code class="docutils literal notranslate"><span class="pre">Skeleton.relabel_node()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.relabel_nodes"><code class="docutils literal notranslate"><span class="pre">Skeleton.relabel_nodes()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.rename_skeleton"><code class="docutils literal notranslate"><span class="pre">Skeleton.rename_skeleton()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_all_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_all_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_hdf5"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_hdf5()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.save_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.save_json()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetric_inds"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetric_inds</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetries"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetries</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetries_full"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetries_full</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.symmetry_names"><code class="docutils literal notranslate"><span class="pre">Skeleton.symmetry_names</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.to_dict"><code class="docutils literal notranslate"><span class="pre">Skeleton.to_dict()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.Skeleton.to_json"><code class="docutils literal notranslate"><span class="pre">Skeleton.to_json()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder.decode"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder.decode()</span></code></a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonDecoder.decode_preview_image"><code class="docutils literal notranslate"><span class="pre">SkeletonDecoder.decode_preview_image()</span></code></a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonEncoder"><code class="docutils literal notranslate"><span class="pre">SkeletonEncoder</span></code></a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.skeleton.SkeletonEncoder.encode"><code class="docutils literal notranslate"><span class="pre">SkeletonEncoder.encode()</span></code></a></li>
+</ul>
+</li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -1242,7 +1539,7 @@ <h1>sleap.skeleton</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/api/sleap.util.html b/api/sleap.util.html
index bb1ca8b15..0db3556ef 100644
--- a/api/sleap.util.html
+++ b/api/sleap.util.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>sleap.util &#8212; SLEAP (v1.3.3)</title>
+    <title>sleap.util &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'api/sleap.util';</script>
     <link rel="canonical" href="/develop/api/sleap.util.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -292,7 +290,9 @@
   </button>
 `);
 </script>
-
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
 </div></div>
       
     </div>
@@ -308,6 +308,27 @@ <h1>sleap.util</h1>
     <div id="print-main-content">
         <div id="jb-print-toc">
             
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.attr_to_dtype"><code class="docutils literal notranslate"><span class="pre">attr_to_dtype()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.dict_cut"><code class="docutils literal notranslate"><span class="pre">dict_cut()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.find_files_by_suffix"><code class="docutils literal notranslate"><span class="pre">find_files_by_suffix()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.frame_list"><code class="docutils literal notranslate"><span class="pre">frame_list()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.get_config_file"><code class="docutils literal notranslate"><span class="pre">get_config_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.get_package_file"><code class="docutils literal notranslate"><span class="pre">get_package_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.json_dumps"><code class="docutils literal notranslate"><span class="pre">json_dumps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.json_loads"><code class="docutils literal notranslate"><span class="pre">json_loads()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.make_scoped_dictionary"><code class="docutils literal notranslate"><span class="pre">make_scoped_dictionary()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.parse_uri_path"><code class="docutils literal notranslate"><span class="pre">parse_uri_path()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.save_dict_to_hdf5"><code class="docutils literal notranslate"><span class="pre">save_dict_to_hdf5()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.uniquify"><code class="docutils literal notranslate"><span class="pre">uniquify()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.usable_cpu_count"><code class="docutils literal notranslate"><span class="pre">usable_cpu_count()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.weak_filename_match"><code class="docutils literal notranslate"><span class="pre">weak_filename_match()</span></code></a></li>
+</ul>
+            </nav>
         </div>
     </div>
 </div>
@@ -318,51 +339,37 @@ <h1>sleap.util</h1>
                 <article class="bd-article" role="main">
                   
   <section id="module-sleap.util">
-<span id="sleap-util"></span><h1>sleap.util<a class="headerlink" href="#module-sleap.util" title="Permalink to this headline">#</a></h1>
+<span id="sleap-util"></span><h1>sleap.util<a class="headerlink" href="#module-sleap.util" title="Permalink to this heading">#</a></h1>
 <p>A miscellaneous set of utility functions.</p>
 <p>Try not to put things in here unless they really have no other place.</p>
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.attr_to_dtype">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">attr_to_dtype</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cls</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L69-L95"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.attr_to_dtype" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">attr_to_dtype</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cls</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L66-L92"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.attr_to_dtype" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts classes with basic types to numpy composite dtypes.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>cls</strong> – class to convert</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>numpy dtype.</p>
 </dd>
 </dl>
 </dd></dl>
 
-<dl class="py function">
-<dt class="sig sig-object py" id="sleap.util.decode_preview_image">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">decode_preview_image</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">img_b64</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bytes</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">&lt;module</span> <span class="pre">'PIL.Image'</span> <span class="pre">from</span> <span class="pre">'/usr/share/miniconda/envs/sleap_ci/lib/python3.7/site-packages/PIL/Image.py'&gt;</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L389-L401"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.decode_preview_image" title="Permalink to this definition">#</a></dt>
-<dd><p>Decode a skeleton preview image byte string representation to a <code class="xref py py-obj docutils literal notranslate"><span class="pre">PIL.Image</span></code></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
-<dd class="field-odd"><p><strong>img_b64</strong> – a byte string representation of a skeleton preview image</p>
-</dd>
-<dt class="field-even">Returns</dt>
-<dd class="field-even"><p>A PIL.Image of the skeleton preview</p>
-</dd>
-</dl>
-</dd></dl>
-
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.dict_cut">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">dict_cut</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L226-L239"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.dict_cut" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">dict_cut</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L223-L236"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.dict_cut" title="Permalink to this definition">#</a></dt>
 <dd><p>Helper function for creating subdictionary by numeric indexing of items.</p>
 <p>Assumes that <code class="xref py py-obj docutils literal notranslate"><span class="pre">dict.items()</span></code> will have a fixed order.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>d</strong> – The dictionary to “split”</p></li>
 <li><p><strong>a</strong> – Start index of range of items to include in result.</p></li>
 <li><p><strong>b</strong> – End index of range of items to include in result.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>A dictionary that contains a subset of the items in the original dict.</p>
 </dd>
 </dl>
@@ -370,10 +377,10 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.find_files_by_suffix">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">find_files_by_suffix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">root_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">depth</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">posix.DirEntry</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L348-L381"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.find_files_by_suffix" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">find_files_by_suffix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">root_dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">suffix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">prefix</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">''</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">depth</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">int</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">DirEntry</span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L335-L368"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.find_files_by_suffix" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns list of files matching suffix, optionally searching in subdirs.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>root_dir</strong> – Path to directory where we start searching</p></li>
 <li><p><strong>suffix</strong> – File suffix to match (e.g., ‘.json’)</p></li>
@@ -381,7 +388,7 @@ <h1>sleap.util</h1>
 <li><p><strong>depth</strong> – How many subdirectories deep to keep searching</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of os.DirEntry objects.</p>
 </dd>
 </dl>
@@ -389,13 +396,13 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.frame_list">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">frame_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L163-L180"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.frame_list" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">frame_list</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">frame_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">List</span><span class="p"><span class="pre">[</span></span><span class="pre">int</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L160-L177"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.frame_list" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts ‘n-m’ string to list of ints.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>frame_str</strong> – string representing range</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>List of ints, or None if string does not represent valid range.</p>
 </dd>
 </dl>
@@ -403,14 +410,14 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.get_config_file">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">get_config_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shortname</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_file_not_found</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">get_defaults</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L249-L307"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.get_config_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">get_config_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shortname</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_file_not_found</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">get_defaults</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L246-L294"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.get_config_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns the full path to the specified config file.</p>
 <p>The config file will be at ~/.sleap/&lt;version&gt;/&lt;shortname&gt;</p>
 <p>If that file doesn’t yet exist, we’ll look for a &lt;shortname&gt; file inside
 the package config directory (sleap/config) and copy the file into the
 user’s config directory (creating the directory if needed).</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>shortname</strong> – The short filename, e.g., shortcuts.yaml</p></li>
 <li><p><strong>ignore_file_not_found</strong> – If True, then return path for config file
@@ -418,10 +425,10 @@ <h1>sleap.util</h1>
 <li><p><strong>get_defaults</strong> – If True, then just return the path to default config file.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>FileNotFoundError</strong> – If the specified config file cannot be found.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The full path to the specified config file.</p>
 </dd>
 </dl>
@@ -429,22 +436,22 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.get_package_file">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">get_package_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L242-L246"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.get_package_file" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">get_package_file</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L239-L243"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.get_package_file" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns full path to specified file within sleap package.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.json_dumps">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">json_dumps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L49-L66"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.json_dumps" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">json_dumps</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Optional</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">]</span></span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L46-L63"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.json_dumps" title="Permalink to this definition">#</a></dt>
 <dd><p>A simple wrapper around the JSON encoder we are using.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>d</strong> – The dict to write.</p></li>
 <li><p><strong>filename</strong> – The filename to write to.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>None</p>
 </dd>
 </dl>
@@ -452,13 +459,13 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.json_loads">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">json_loads</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L34-L46"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.json_loads" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">json_loads</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">json_str</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L31-L43"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.json_loads" title="Permalink to this definition">#</a></dt>
 <dd><p>A simple wrapper around the JSON decoder we are using.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>json_str</strong> – JSON string to decode.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>Result of decoding JSON string.</p>
 </dd>
 </dl>
@@ -466,17 +473,17 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.make_scoped_dictionary">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">make_scoped_dictionary</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">flat_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">exclude_nones</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L323-L345"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.make_scoped_dictionary" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">make_scoped_dictionary</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">flat_dict</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">exclude_nones</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bool</span></span><span class="w"> </span><span class="o"><span class="pre">=</span></span><span class="w"> </span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L310-L332"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.make_scoped_dictionary" title="Permalink to this definition">#</a></dt>
 <dd><p>Converts dictionary with scoped keys to dictionary of dictionaries.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>flat_dict</strong> – The dictionary to convert. Keys should be strings with
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">scope.foo</span></code> format.</p></li>
 <li><p><strong>exclude_nodes</strong> – Whether to exclude items where value is None.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p><dl class="simple">
 <dt>Dictionary in which keys are <code class="xref py py-obj docutils literal notranslate"><span class="pre">scope</span></code> and values are dictionary with</dt><dd><p><code class="xref py py-obj docutils literal notranslate"><span class="pre">foo</span></code> (etc) as keys and original value of <code class="xref py py-obj docutils literal notranslate"><span class="pre">scope.foo</span></code> as value.</p>
 </dd>
@@ -488,19 +495,19 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.parse_uri_path">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">parse_uri_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">uri</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L384-L386"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.parse_uri_path" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">parse_uri_path</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">uri</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">str</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L371-L373"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.parse_uri_path" title="Permalink to this definition">#</a></dt>
 <dd><p>Parse a URI starting with ‘<a class="reference external" href="file:///">file:///</a>’ to a posix path.</p>
 </dd></dl>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.save_dict_to_hdf5">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">save_dict_to_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">h5file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">h5py._hl.files.File</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dic</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L116-L160"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.save_dict_to_hdf5" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">save_dict_to_hdf5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">h5file</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">File</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">path</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dic</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">dict</span></span></em><span class="sig-paren">)</span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L113-L157"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.save_dict_to_hdf5" title="Permalink to this definition">#</a></dt>
 <dd><p>Saves dictionary to an HDF5 file.</p>
 <p>Calls itself recursively if items in dictionary are not
 <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.ndarray</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.int64</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">np.float64</span></code>, <code class="xref py py-obj docutils literal notranslate"><span class="pre">str</span></code>, or bytes.
 Objects must be iterable.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>h5file</strong> – The HDF5 filename object to save the data to.
 Assume it is open.</p></li>
@@ -508,10 +515,10 @@ <h1>sleap.util</h1>
 <li><p><strong>dic</strong> – The dict to save.</p></li>
 </ul>
 </dd>
-<dt class="field-even">Raises</dt>
+<dt class="field-even">Raises<span class="colon">:</span></dt>
 <dd class="field-even"><p><strong>ValueError</strong> – If type for item in dict cannot be saved.</p>
 </dd>
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>None</p>
 </dd>
 </dl>
@@ -519,15 +526,15 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.uniquify">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">uniquify</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">seq</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">Hashable</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L183-L199"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.uniquify" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">uniquify</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">seq</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Iterable</span><span class="p"><span class="pre">[</span></span><span class="pre">Hashable</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">List</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L180-L196"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.uniquify" title="Permalink to this definition">#</a></dt>
 <dd><p>Returns unique elements from list, preserving order.</p>
 <p>Note: This will not work on Python 3.5 or lower since dicts don’t
 preserve order.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><p><strong>seq</strong> – The list to remove duplicates from.</p>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>The unique elements from the input list extracted in original
 order.</p>
 </dd>
@@ -536,11 +543,11 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.usable_cpu_count">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">usable_cpu_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L98-L113"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.usable_cpu_count" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">usable_cpu_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">int</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L95-L110"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.usable_cpu_count" title="Permalink to this definition">#</a></dt>
 <dd><p>Gets number of CPUs usable by the current process.</p>
 <p>Takes into consideration cpusets restrictions.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Returns</dt>
+<dt class="field-odd">Returns<span class="colon">:</span></dt>
 <dd class="field-odd"><p>The number of usable cpus</p>
 </dd>
 </dl>
@@ -548,17 +555,17 @@ <h1>sleap.util</h1>
 
 <dl class="py function">
 <dt class="sig sig-object py" id="sleap.util.weak_filename_match">
-<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">weak_filename_match</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.3.3/sleap/util.py#L202-L223"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.weak_filename_match" title="Permalink to this definition">#</a></dt>
+<span class="sig-prename descclassname"><span class="pre">sleap.util.</span></span><span class="sig-name descname"><span class="pre">weak_filename_match</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">filename_a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">filename_b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">str</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bool</span></span></span><a class="reference external" href="https://github.com/talmolab/sleap/blob/v1.4.1/sleap/util.py#L199-L220"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#sleap.util.weak_filename_match" title="Permalink to this definition">#</a></dt>
 <dd><p>Check if paths probably point to same file.</p>
 <p>Compares the filename and names of two directories up.</p>
 <dl class="field-list simple">
-<dt class="field-odd">Parameters</dt>
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
 <li><p><strong>filename_a</strong> – first path to check</p></li>
 <li><p><strong>filename_b</strong> – path to check against first path</p></li>
 </ul>
 </dd>
-<dt class="field-even">Returns</dt>
+<dt class="field-even">Returns<span class="colon">:</span></dt>
 <dd class="field-even"><p>True if the paths probably match.</p>
 </dd>
 </dl>
@@ -588,6 +595,33 @@ <h1>sleap.util</h1>
             
             
               
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.attr_to_dtype"><code class="docutils literal notranslate"><span class="pre">attr_to_dtype()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.dict_cut"><code class="docutils literal notranslate"><span class="pre">dict_cut()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.find_files_by_suffix"><code class="docutils literal notranslate"><span class="pre">find_files_by_suffix()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.frame_list"><code class="docutils literal notranslate"><span class="pre">frame_list()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.get_config_file"><code class="docutils literal notranslate"><span class="pre">get_config_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.get_package_file"><code class="docutils literal notranslate"><span class="pre">get_package_file()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.json_dumps"><code class="docutils literal notranslate"><span class="pre">json_dumps()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.json_loads"><code class="docutils literal notranslate"><span class="pre">json_loads()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.make_scoped_dictionary"><code class="docutils literal notranslate"><span class="pre">make_scoped_dictionary()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.parse_uri_path"><code class="docutils literal notranslate"><span class="pre">parse_uri_path()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.save_dict_to_hdf5"><code class="docutils literal notranslate"><span class="pre">save_dict_to_hdf5()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.uniquify"><code class="docutils literal notranslate"><span class="pre">uniquify()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.usable_cpu_count"><code class="docutils literal notranslate"><span class="pre">usable_cpu_count()</span></code></a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap.util.weak_filename_match"><code class="docutils literal notranslate"><span class="pre">weak_filename_match()</span></code></a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
             
           </div>
           <footer class="bd-footer-content">
@@ -606,7 +640,7 @@ <h1>sleap.util</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/datasets.html b/datasets.html
index 3864faa91..256e0e4db 100644
--- a/datasets.html
+++ b/datasets.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Datasets &#8212; SLEAP (v1.3.3)</title>
+    <title>Datasets &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'datasets';</script>
     <link rel="canonical" href="/develop/datasets.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Help" href="help.html" />
     <link rel="prev" title="Developer API" href="api.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -335,7 +333,7 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="datasets">
-<h1>Datasets<a class="headerlink" href="#datasets" title="Permalink to this headline">#</a></h1>
+<h1>Datasets<a class="headerlink" href="#datasets" title="Permalink to this heading">#</a></h1>
 <p>Here we provide a number of sample and benchmark datasets that can be used for
 experimenting with or extending SLEAP.</p>
 <p>These datasets are all saved as <code class="docutils literal notranslate"><span class="pre">.pkg.slp</span></code> files, which are self-contained SLEAP labels
@@ -349,12 +347,12 @@ <h1>Datasets<a class="headerlink" href="#datasets" title="Permalink to this head
 <p>Need a quick testing clip? <a class="reference external" href="https://storage.googleapis.com/sleap-data/sandbox/sleap-mice-demo/mice.mp4">Here’s a video of a pair of mice in a standard home cage setting.</a></p>
 </div>
 <section id="fly32">
-<h2><code class="docutils literal notranslate"><span class="pre">fly32</span></code><a class="headerlink" href="#fly32" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">fly32</span></code><a class="headerlink" href="#fly32" title="Permalink to this heading">#</a></h2>
 <p><img alt="fly32" src="_images/example.fly32.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -400,12 +398,12 @@ <h2><code class="docutils literal notranslate"><span class="pre">fly32</span></c
 </table>
 </section>
 <section id="flies13">
-<h2><code class="docutils literal notranslate"><span class="pre">flies13</span></code><a class="headerlink" href="#flies13" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">flies13</span></code><a class="headerlink" href="#flies13" title="Permalink to this heading">#</a></h2>
 <p><img alt="flies13" src="_images/example.flies13.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -451,12 +449,12 @@ <h2><code class="docutils literal notranslate"><span class="pre">flies13</span><
 </table>
 </section>
 <section id="mice-of">
-<h2><code class="docutils literal notranslate"><span class="pre">mice_of</span></code><a class="headerlink" href="#mice-of" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">mice_of</span></code><a class="headerlink" href="#mice-of" title="Permalink to this heading">#</a></h2>
 <p><img alt="mice_of" src="_images/example.mice_of.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -502,12 +500,12 @@ <h2><code class="docutils literal notranslate"><span class="pre">mice_of</span><
 </table>
 </section>
 <section id="mice-hc">
-<h2><code class="docutils literal notranslate"><span class="pre">mice_hc</span></code><a class="headerlink" href="#mice-hc" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">mice_hc</span></code><a class="headerlink" href="#mice-hc" title="Permalink to this heading">#</a></h2>
 <p><img alt="mice_hc" src="_images/example.mice_hc.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -553,12 +551,12 @@ <h2><code class="docutils literal notranslate"><span class="pre">mice_hc</span><
 </table>
 </section>
 <section id="bees">
-<h2><code class="docutils literal notranslate"><span class="pre">bees</span></code><a class="headerlink" href="#bees" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">bees</span></code><a class="headerlink" href="#bees" title="Permalink to this heading">#</a></h2>
 <p><img alt="bees" src="_images/example.bees.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -604,12 +602,12 @@ <h2><code class="docutils literal notranslate"><span class="pre">bees</span></co
 </table>
 </section>
 <section id="gerbils">
-<h2><code class="docutils literal notranslate"><span class="pre">gerbils</span></code><a class="headerlink" href="#gerbils" title="Permalink to this headline">#</a></h2>
+<h2><code class="docutils literal notranslate"><span class="pre">gerbils</span></code><a class="headerlink" href="#gerbils" title="Permalink to this heading">#</a></h2>
 <p><img alt="gerbils" src="_images/example.gerbils.jpg" /></p>
-<table class="colwidths-given table" style="width: 80%">
+<table class="table" style="width: 80%">
 <colgroup>
-<col style="width: 20%" />
-<col style="width: 80%" />
+<col style="width: 20.0%" />
+<col style="width: 80.0%" />
 </colgroup>
 <tbody>
 <tr class="row-odd"><th class="stub"><p>Name</p></th>
@@ -733,7 +731,7 @@ <h2><code class="docutils literal notranslate"><span class="pre">gerbils</span><
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/docs/_static/bonsai-connection.jpg b/docs/_static/bonsai-connection.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..32b725416e2ed32f12977cdf50c73b53ce9b3d9c
GIT binary patch
literal 113805
zcmeFZ2UJwgmp|A9MTwFXl$<0=R#J;dmZ;F=77@v`l5-<S&LAKlpkyR886>r2$r&Uk
z$*G&zK!@$`&QAC`vpchMcK^F)&ade@R8_xM@7`PWs&4pH;eO$k0k<?%)Kvfk1Oxyt
z{2u@}2T%eK5fc9O#eazL-)lFnT_YyGMoL0*{RTNHIXM|A85sp7%`FN_YDzM)TMW0T
zY3b<c>B*@WZ!^%{rlF&!`>PQGBK$qX*KS_Bc9V{RjDqg}_6PR^Kz)NSiugAX0SADP
znt+I!0M`j%1po-H;i>(L;Q!-8K!~UDI?0Wjq-6LDYHtAu35bXYiHZKA8h^Dv{&xT|
z^)(uHQN`=DdX^*{@94zdCw#rZ`J}RqUVjw9CH~4K=q4!x<Lx_4+&sK|_xU9xrKBG|
zl2Ll9tfH!>uJQbZfg#Aq*u?6!^&1;oJ9}3*cMnf5Z=c`~A)#U65s`^WACps3Kc%JT
z<mTlU6c!bKtE#T4t*dYN-q_yJ+11_iv$t<-d}4BH`q#|t%Iezh^^MJ~?H%ON@yY4g
zIqKr_FTMx>ME}Irzi{>kU(|TM2#JY_h)Mq9i-6D*pNOc5ud$0>r%}`+v3y6%A@=?T
z-IIi`m2EdU#q|;NuUtk+8Mq`?xRHOc_P=xXUt=ujzs1?VF!rx}%>u}Y2=JXpL=6A}
zt~6JNW8LB8sMy(Qt(GeZD%+~63$=?^2jVIEJ_KeUv=E#DIs(E0xN~rTRRPtr52g`v
z%1kr;bc{{~@5!q>U&pZeWoWWl3lelcw#?R>YIQ?MRjy^4cC)ZmOuENWX%%jzCd~HL
z`EVinU07xx1q;-!g(Iwqk{$Q=v#ym?xP3naBkIt`IDjG;of3rugy&)zmZE@L0yse8
zC=^Nm`%<LykOCEKg6*$_Cz_^+-7w$hi6u+S_EQvRi+$Db=xC-gIO}6Nu=9}8;j(R9
zP$Vy9ySa_5*G}TC*(WS2=#Aptxwo$C)m9|({AP(DqABQ#6RH{;1iuhs!~v*P|6p+g
zd=onoA<?LdWD+F2>hLi6sU@P__&r`-V1S$3iiSSoson8|9}E310wa>JNQ*louPPJQ
zbQ4!ZoUk)D$M<=faRAk|^Znnsdr=8j-R)PCy)K6s2=*Cd!4GyhhaTPiPuTSd>w=Qn
zxoT2$S2Fb_m<$vpsBO44_{U4Izqn|dTW7I2@L6r_dzu;cce?p-fROst;Y(5+;CBre
zt^DppSG?knI@Z4nx?CX=um}9V6h>T5pez;2G(Zps_yWUfD&qj1dN=?Xq$~S`9tY^P
zg`*<Zw!Tny{H@FApzU)=8_YH6k%<VFqv4W|=zkU=@~JD*_;-fZD|TaAzvb%LP1~_=
z7iyngZiy%B`p{Z5IXIlb9;DrqsvBWs2|Q;xUCt&1{&vQX3mAU9jV4Twn1JWAv5nBZ
zlE)-xCOE*bHx#*q$h{JRZZ^YFhbRy`GNUIwtOHE?Yj?`y>|E;I&KbjOeCEAq57+rT
zn!XGj@~l^_+A8!7FzJl5w3acnJ5pmwzqTT(J>c%xp@>`5uQ9O9j^>S3nZWo>hifSt
zT)Kg`^p4c!bfnG}zz309SZF{<QQeriW2@7Qh(&99)*&VZe!K`qd!qTU{osp35v(kh
ze$xrrq>T^>K?`T!5~iq+g47hjYRT@AXf(HKx={OF7_+OJFuS{`rrjqM4Vt-Lgf_wC
zoKO*?96IfCz&s_Z8DKIquMC~~*zS#;F*Rb^Uh=U)nhOWo+9vWeMj#)FaN2j@va)^u
zW*AoI;PrFN{&$}_Wc>xKg(d$#B4-uM2)7pkwt=Z6O&17;V5`2&nc_MPnocZNx;0hb
zn^Ppn?$;}&JayT)wKDVy{Be+F7fc1UErFsy+a+GB%lAyQJfDx2mDNy?UTQ`yRn7M^
zp!R5o8%SPTes-+}O5LBGhU=hMH{jYi@P@)s93TmrD<AhSy`uLoGzmmzoCTtw*Tae&
zL{=X{s|6*jR>`80gI*QCVtH!!@L9D$=gw$!V>}Tbmred?GbG0=*ra95ZFj{TYGdc*
zD`!W$ujVZM$<&)!-I6x*%K?D!F}b+x4!jKiMqvh^)9YA=L-T%5!&kGkY$+;iMIBam
z<K^Xdm|BXVJsHsE?8{r<BmjEKbpNj<&=BzNGX*TSWEslru*enHopl#+9tM*rliOiP
zt$%K|<KtKV>Bt1|?ORe;^hFO>(#ShiQ>b~j?;k72pVN~ULom`uUTWGPxGU)gL{Ks}
zN7XfVMn}+d+&|NX@FDGBf~lSL{gjLNK=p^^NmvHYP#j=|1KWA!mhE)h1xq_qZyn&}
z<e+_5-A8i-^n`mp?6%!^hL^zy{a1Y;7^tC5DGu;&r=hsN&tjk$6Tdfzn6rk<7cw6Q
zUGMGM!PqOlo4Pd`@>$vq1n9bS5NKLgiGGgWuJ62e%{{4rx#voC_S)Cfy)jJoFlTr(
zrWKqv;<)hnvz44~?c61{ycOYNt`JuIx4G}nhX(H~-K$}ichwjdw49KXZk%FP$+C6Z
zh4arE7}sMFPxShljrf@Z=Xz(wP3nMj5)V)AaU5(umx)?3CY#LA8W&dcdMhp(@-&Y8
zH4fl+8U2|_8icKb|CW{5$3P$hXTZNR`R7!9Bm`Zp4?&Bat)A;8Bzn<o>DbxHPl*25
zyeU}qG`doy^R8_2^P1zf9@FUr?csO#{?jE!|9{xb{*SfpZ`$5JR{uYh!~ba#*%_v=
z&?@qB)o5&hih1vIq^SJ3^X9SokG0~QuLw_cqzo&@s|^RpnTiA)QN`i_N78sjeL51l
z)ho~`C%)p4$=aDE3U5f{sh@&bet-Y!&EmJZigu;94?I;mUuNC71dT4yqe)<1C7mAU
zCWcwl@!UDx166DtB(BjkR8It=@d~Q!8tWyYIt~D}!W$UHp2y>tS7EN3U^28gk|F{T
zno!*Osgdf#Li*6rkT=ET0bAmY&V)~smhBpv^q%vNw<uT`Q4Xs@^3CjdJ39I0VJo~g
zRn0Z&Dvhn3G>W$q#hFSUs;fJ}cveXsJ8Z#F&s#<Dsw`Lb`56$)ANTiS#yKn?2is~M
z_o<l|vDRZjkgDk~EJ9_^1EiR-t!H&?&0-a{wRj`OX2k3haN<{J`A`n#e@|c~F1(es
z{Ij39aB4~8mf`HZ;j`u{(`!fMA-q}%y-hrz0afR>aG=ShIF{}(8oG6p7lVnz0U~q%
zUd*vLJ)Fi;UBI_FK4&<c*RMDPXkI#A1VG-;PS$RQ2=BopW(pDJ$+hKSXZgTwp}=~r
zw|;e^p~|H1$((@F@+5T_hw4A&S@dVcx6c_%gP5y?jdveGHaGyOXE*SKI0=jG!2!DP
z%f;WyFQV7@%7_0><sU?Up%jbU!>`QRs2*%!B6R5&4uJXw{Cg4bk5>L;xlEv_aR5&D
zO^g>D8y(vkG=}|+1DuZI0M;f*cnHh2X+MEgho~6Cw$fE!l^EkZr4-??+r(Lc(^Ym*
z7wF|S4p3DFutGmWm;tF<Sx`wGv=uRuP6`g9_lFN=MnVD}_0`?B`W%TUGd(j!^3)@$
zSL&i%%j`;4wWMD-4AU)&q>t<{oV(P_<~^un#sR!J37hJI=xpu@Y=7f3k-ZFH-JEU!
zdK+J0WdiQWvZ${?iM*ydR+Z1A(6kJ}#agAd6<euwS!~VeGfvHEHi-h^=?|@sfomI&
zN8;bt6)!$27s1|vg2w!qSK}z{(#w(uTMjz}bclY<FI0E$yOJ*6Pd6}dO31G14$$wk
z41IX#QX$fCv<Hp@IN=px;I@F>oj$P*(`}T?91frk9=92c#}3Tm2Q-?u2*ugK=wlXe
zGh25|vP9EiE^XjUIv(8i&X=iDRZN*LBu{^^LF0g`??*mHWH~OzkGj;zD_2p55(YE2
z$8w}ziR-s!O-=a<X`{@FkEd>$+|NDA8RB^9vYQnHexe5&INyb>SzmCqkSEqWVwi>x
z4F;55W{ZOZ`h14I+AO=J8dokc!2Eqzi$1DVEX*!|%k#Nw8<(TWKD~^F5Q(O#w>g6R
zuO9l{n+%{tfnQWkORBlYXl$Gaf*w@I*FAWeGNqUK!*BJ5l}M~~b;9S37UxJkzvBFQ
zwDd)QNE35&<{}P&zyUH(Ps|fQ`Zz$G6ux_aC+sim&kCN*RLW8G;{abt2d|!jk4J%r
zU)65%6=6pi(I+}rTY6{V@XZ|P(vv{0oZNrDKy{V_2gr)R0sKh;iIH*p)X?FlI6!L%
zt8En4hZl3gN2tLMtifCN^0e4XI{zGzi>>VQkuU#f_W##HK(Tf&k)7ZGEMDUs>7ljx
z+a^PC1+x}i+*yUr)f>hpgt0|^p!c-De^WPDJ(4_!>)zP%ravl!hnzMQai-s6)K9-d
zC!>DvV`*vL>swEbZseBr__DZ{c@c78$%B<@E5~XE$qcR>Oq6biN=EG+_tZ%!|Bx+s
z9(!GU7$&hqj(#%86Uh1u2l!xqx>BC0G@P{Gfe|$SBnK0LT(;E@9=q0~#4Q$ZfE&<o
z;1<C*9AKFoj@DvDgrS15flM$g06&U5$F+xW0AN2lKI_}kH2BzNNP!x$8IHEj+nH5O
z&7Va%dbOS%Y2pAiYSUr86N?Z&#q|#tSr6of#RGMIYMDeWTYQDA7OPW(Z4-BGOs<rN
z!`&rBlYrUlQN)%+V!h|Dg-M?1g+IEa#oodJZb3(g3~+!Scc7?r;0hIz1Ot!+qX|-R
zfQWonk3lya;5}CJvSq&Iz+rC=>W>2q;sAzn&$0Ab-)*e(GHk{-8!M648;w>e-khrS
zndK@A?NaQM&IH}_G<nW|gL2~lLWfz4?3D|l{7@`*TCslAd4CPbkek<DM(ts2uperD
zC*i%S!PX7y-r=$AzYAjYPeBRTL3SMilw6$Zm(Uk*@gH(}2#5)GaMTek$&gmK@XIdV
zZp1<Avg|aZD`ntB`B-7{){x<y&j3%t!G?sr(Go8E(ry(1VJATSbhQrpcW!h4ltw+z
zIcH^pvIE{|W?iMa1lYq-L?So<tty%w%ZAsOu9x5dfnqoSQj$}QuMCP(LM!3`k16nu
zVSqJWZ?naZ0$?ZW31AY6xDLkh1pecqe~928qu_u1i=YtqcPmln`U42b)>AjBTde5N
z8eJ05R~(=;dka2QTf(b^0~iBGp=A!b1j=;(t#n=ZI}X4a51p36zA7AnPMATb!I)%w
z3Rf%;KSz~2p}Tax-~dGp_&ofIf8=^5o!;YX^lfNe)c!J65W*B5BWF;!r9V$Nb(Z)&
zac7ybfbW(>q7`d<L9;h}Ifs=@VYJp!?ll5B{1Y)aVrEr7h9@66f2m6{UIR);3t;jT
znEmY1_Wd|lo8*<A$H?`%rFG0Mb+*iu<r{nicn!__PUrszoY9Asd2Mzq-b=`ts4ZcB
zlzJ!BctTBl{P)n|SdzzN^6ft7QP7)p`e8fPTg~FIXYs!rX8i{a<_zf@^f+0qd3le@
zLqCj966f;CsvC#U={<|jiR1(PKahx$<1kNw2O>wXxBzxPOC&H1D~A?GNGOORbwd3(
z{g?HM=T%pzD3wbVTgDzLJB4;1JkIdqVk06-$O>rx=X>uVhJP&W1EaDCP_>x)bsXT0
zm!uo(9<<74;Fz>`<6;igqt!&|vr#adB~nn2Uwy41-AiepbFRj12>K=%@jVnyFC@oY
zj25kgBZoW7>~$C&#*Jp~S#?Y?x&G{BcqXmeZ|NlX_N|eG?cJpF$Yl^cg6yol5gc;q
z-V^IANp;(69E=Pt{Icbvo#5bfZ(&wZn6QZFwgKl8GIdT7;@61|bFCtN6nM*_WNkPE
z%Z;7G;x&&CydZzj9{lxj>s+IJmc{<=`day(^jNb77{%+HD*YZUpFg<jAf#zJBBxs9
zFeU8+4_j*)|E<|`-N7{K4tHm9i0*Bf3jW)#!ogWba|TD4rL{Z^{bkE@jb4%M*}@f1
zs%^Wk$f&cQ={jNCR?7mWF+{8P&ZC{=$hcD@gsAgIA-r5^c~28^Ki=s6{W>9`{(dhY
z@&rbKvRUovfZxeKs9ak|g)CF&%kGHon3(DAAS^f%nZ{%V+>EUPLZz?uFl(|<(ZDOt
zlKc@cBgEhVEZ;7DVW^t2rEpi_7NRyKf+5wNV6ZZB@j>_Fw6}iST(9m0TL0{^EJRZw
z4=PuSPvLja<`W`8TjO?8>5ve|yivhbZ`~3L^*asOgAI53MA>ztzeScbbelC{SqVoy
z&SB?e<{!qKc=3Jc1jiCVzHh-wnLPKc3!v@l%L}}n76eH6`|+-G3Y%Q8S8Vsx?ZVH5
zPj&8kibx3mc*`9_RW6W<H){4TmiJwitC8eba%5f1a(^ZSi#LN&8t!BQik9@FF2P>0
z+w0Va%zkZmbhU3@Zx|+rw#hk5urm5RoGL^@!?=^$8q;@Nz{V)qpOdGB&$CbFHkuRi
ztaj*xz30a7LiRSV_)wdxAiP_m5T)5;GQUWNco87fI9ZMs0)}zbPaGu579>8L-7v1G
zXbMT<py+(dO#fnU)6VJAg6yz5Wg+E5-;Pi|4nSoGZ`_eR_sNP`8J_#LtBFkQG^M)h
z_2hC`ZPd@kZbhEopP!FO>-^3QVNdhVS<(#APNb^LRRWdB&HdixY{1|evu`F*7gO?d
z$mx{`rP<teUCNQVtEW4&oW*0iX_u87vUQ&m0)3{1*1*5?54)I_V$1jC{24MFR#|C{
zy&9(&M+U}>?HPXExQ89dF}ktz@kUG{9N4F%5B()#JjH_)!UxM0cod5rn9ieX;EZNw
zvnbEKQTFP^YmKUB1s6sS3HoWIj&6URdQrYc|GWl*1AM%IpYyIX@fRc}g*M@>YTkP6
z4C+K)1cWJ*Xb~}+IJxb2a}{IS>BI9;GSgReJh9#Q#KAQ&$m7Z==J#vnsfR2efmrb_
zA9x`8U9O)7!Z~g=n{ts1Lu**4*`yg0BHi7=$^V$`j4e9o1}qyZ9MI_l?WBq_Iqan1
zzv_6flYaWa&?lrm3e8Ru{zaYRae|rYP3!0mw#NjlVp0VB?<`Icn{84{lyYP_UiPD3
z6YojQygU}#ds(H<T=Fnzfrjr~l-Hh*NV3#%#T(DgHEaqtGlAPq7~~Ea4QO5p8_vs#
z<bTYU6AAcVmDbTzH~@#mdmP~96%@lzqqhvbpgEzsBEd^{t2QT4;vmViPa&qrXolr!
ziT|r}_Dd&D$I4{Rtj$MfMZncMylQ=bzd7g;4san>yN?4*W|zI$v3M}z=8)@D<ly5J
z(--(mTKZl0GD{e_;+GrnKW>RbmBF`Uzd)B`q1b07GCGkGPClw}&n8RV?{G3_3JL4$
z=T3z15>_?8LyB}PQLa^CXB!(BrH6X$EDL|lgBQ(`nX~)q>Q$55V~xJWe~Dg`fI0QF
zb`AgAW^=IukW0!fRupG8J}AnEfW*}3gfsC?x{;m2tBx&5WSJhZCsR&G8qJf$D$9vI
zB&X69a4B)>Z8>{@Hx2&I;@{{k>$^|hf%k;YVS^h<us1gG(1pG~Wm3=gg&g>QPK|&|
zB;X%S{8P{TS7i}p<PH;EpK>Q-p88>w(%ris<Svvs*4vOJ8*Q~%f5Ljo$a*^cmlrbs
z9k$fp{va5evxaTN0q}SQCii2{{dZUD_F!^gbu-m+Fd15Lnq8XARqlyeO_WfV$^Ho9
zz(}k+EU2%m^#vP-pxhyTOUgl_nCnQLEt1&_<d*o;hd<8Z=4K#)Bx1M=@}X5|1j~bJ
z>RIFsF>Y4&DzWWld*@(BLmafkwtVYLWshOP@S{P=5+3cJM}tH^>P)oOr;5D=FPFk0
zu~+*J&*O(Z-LNoD7u7FYvVe!>akgZE0G8xkEG-I>YqlljDP18saZX*DA07SW^=klK
z>}2sf1iYjAN$<Uk$G_ikJo_CNpicjN$ZvWTN?^yrAIq4<P}f=y5gXU7cBL3B&CH8b
zW={zF7Vof~J;QjGI4!4GX6E+*73O+&NIx?txfoaOc*DI!;>SC-+j{R<))%5l49NlP
z$c`;2rDcXv^Q4F{c@&e@x)ST%uh#vn-%eOFy*xW(oS}f896!a%`Nq`X)7$PJL_dpn
zFhld6>YTs+(D%Vd{#kNgbm{Q;@C{UZPj|XwYRVI}PFD#rC9(a4+bnBxuU}#%zgyU>
z=%)j_scspY#tYVJK0M`XKM{(*$*Mo(FfKs*<zFYjRs&1{90CaiUpsX;BwE_tcUrak
zDk>4*!4+ZGv_8M_KqLNUPs6YVY2=2TMR!;?Bo34wg6WP}DXFWi2&vP8I(N$L$$YRJ
z6`yzSF>^<BzZy_~#G9xkw&>;jHPp-6xU4`q@lg(Y@@OFY$8+La%Kp+wfetX0AeL71
zT|rkNR}j6FwBfqn1(*|1agibSo8|LKm)}53!Tlk9kM*s9Vr?VGZ*+G0!KpQ4yXCWR
z_<_P3*h`cma`CFYe#9>QbRf0y?gtg_kEw~{6>O(`5BtvU#(q*zt2>r#8mg3i99?vD
zP1Ydx?ae)HN`=sza(e?C6|+%!GeeqKsf<Dc@UcbM)+@Ctj>ZRWmhpWqH%<#d=>fwE
zT&bO43LyrULE&rxqg1Q&<R?`D8ar?V(@EpyPjzj6&0^Gok66cK@_h)yh@LY2oByuW
z@^9}W5n#Oxht7l7E$&k`!BEvEW6k$+QRNV#Z5h@ON6Ac6P*q)B%7GFvJ-&z8tcOc4
zC(15uCij&~AblI}>>3WBM|Bm2F6qxjIh5Hy@Os|0NQvZzg*K}+(zj-KV)FAdpY&b+
z;&U^8W&AZ>&Ur{L(V*wkawxH-J9t_ZuXay`_@^%E41y8x(_+;hP-el>W?sQ6uW40O
z4c8P%brjNk-_=2K<IDbyk4kSVt4w=jLef5!aNHhIxRqI3_6<9lu8>+dQ;+6cWNV%(
zmt0#Z7!k6pclafO9XQv#`(c2Cbd<j$UQPEVT?#<dLfRCQFdXgX{}u)7G=W6=DHeGJ
zEp;XgfAjLV@Am6}quNx5doMa{uwJwTQ1Ez--><#Dc^@TxW>bgAObr=hE+f&-RznJS
z#d`TkuY8ioxF1oSKB07%`Vk-Clhg$Zh+A!lG^Ezf&jNhqgae!%WVX7TpnQSw#d>pC
zKMGx6GdxSCHZ74!V=n$=>ae!O(09%avBO;jethQgsk}3P!{~)F&m@<IqHvSvn<(fk
zmGVR_kYxK(JF0A#F^lWU`RD$tkbt(cN2@`5W@+4y3wS>U(FXX{ku`jz=s>~F<j3XI
z9jqnBQBct2o@A!gbNpKOp`^pE#I9RN<@Yd`m`Tk~<M=%?{-0s|+7G%M$`P!1t<-7R
zT(gbQqkifEEKFiwyLaEMg^5CD(X%Dxrtw<jy<*a@!31vyln9567+z<kLZ{+%G-o`I
z>vEI7r^n7>radlZhflb|&fdq=#>liE{s<5mV>A-)$uYM#tg9#}l+f)3CRS#|$&lEt
zCVVtky>anYO%5cBfSsi`AyYfDNq0b4D!eZIyvSB>H$65jx2$~5)Z>jo_Iaw&3@2B{
zEm^fEcZ)dbBbH&mkj)}Ul6DK;#;ImmU-y-{;zD!nrS|rFkQboZtR=?TD6^mA@WbSO
z;>X4chn&u@vnx)Uwk_@&A2u_?>}M5Xvsf4{_z=&0@~-#sdeqylB>ujAK&RH0N2zx_
zphQZ%lj>QxE~4GsQnUYEf4!+bbraS%=ianaBQIiv$#-LR!0^v0j+`DcMeTFUfyKOh
zZgbfR^1d-De9$l9szKQ!%cg$!+PcaMmc#B=VJOMA?t<c%T^8=9hBT+;@fuV2L>m_}
z>vUDM99_EgbvLtR8&^IBWu)vG2zX0Q{HvGqk|6?95LIA@`1NFZk+wels4XejwP29!
zN8R%WJ?!`Ba_E!+p$Aw|^!<$0xDO}$981sZYotI43%A{)cqSZ0lC8W3q*GRnqLB$a
z`vr=}Iif=kob$_UJ<%dq3-Iq}o}hiy92)P8C@Ad6j_`m|m!5jWJ)5WzQdMn{@yO4&
zAMF+v;-&84(;t;fcqrP>d_897?SjjVyRaar9R*c)x)!#<P005tBEL0FxRNSD^2SSF
zvJy4yRY{vf8S}M^=Hiq`qS;C7FS$@82r=OMOVx<ImBul(X-V+DPBjXuti{te7*liW
z(@H&!8oet)fcm2V7Z)AQTx_q$u~$@=Ok9M!7Rv$aEn80$`W=*O$8PRm5qF#Ux%5G`
zTD>$tq;G}OK!p9~F@Uad?gDz`4903|VKL)jLM;b4fGqgn*8x8I5IZaE46UzOc;2=u
z#NrenUdHt>tNmW9TmSEV9x3mH1J&8J0dQ*hql~!OLai~awrM$vH7Y8V#%p+|yVLIp
zp}n((nOQ+6b{*mMzMoHv!`E+gpn?pkf^YyUb1J2Av{Iq{J1=cnA&^q`V=*tywy8$$
zIYmCwqv5w@0=>e3t1L6`UeWASXPfX=Qq)}M(ybZ_vhDyL1+Gz;+#T4Kn0=$d2S<ic
z8wS0w<aFa;Gc}@wKK7&7U)@X*CTDRr7h15agrlXF#Qf6SeF+d=@8QO)?mWkHo4Y3G
zJrGJSXB^-~O>5)@SH+m4`zu3kOR+^A?v`fil`3nyS93vx>(d*%(fh7vk)zEbIlJ{f
zQ^7F1#*`@GtW*<bCF1jZ5$9_yHI?CSycSHf*4AA*`XQ+gNfIW3)|&9UMyM2#d#LjM
zPMTdNsf@gDCbpUoCuc|J{Dx2+e&d2;0y{>ZN9=VKE2j2Yte7OHnryONnYS73CPEwe
zMRuTQEEU3$N%9+a^1LTIxr0DI=c6b2S#Jz}@&|8FSR6_=f)N|&eT-SMzsUN902=1&
zE)$2c6XKC)*fB<ntZG83+saf|R#sFZCu{FOEa)4a8X6C_juUw|4YU!KybcmI8AN`Y
z+w&pi4e?X+Q!17G8P=gcr*sYdY9VOoz&hFTXrzit)Y=*<O3Fx1@zN{x%)e=(S-A`q
zr;e1_wqVch^l0T87v>pn9i_;twA+4(=x0qYu>BpbnF*t2dD(X1;@Zj)Fi?fo9d8!#
zUUdR>Eb=2gLZ^M5(K;(ekdtDo@77|fXRMN>ACA_j>p#ENg^Txouf9jSTlgCWx$-$b
znj2d9UNgHC3(tqMrsm_L5Oi|D5<lQ~WLl$x5ap0S8pL9sv?Z<1M)PBW%G=Qz{rlZ2
zFP0O6{HlB4HS0qei1BXEOK5Pj@>f(&e+S418)H+ZWJ|H%Y)|lAw_flohaHDaq>B2J
zdg{c-<(nN##BTl~%VpCYzBScO{Fi+@L&sOFTEUthqGGU?kvCp&J%-2zAeR|z94mj<
z>dY*Gkts=KN^4Zia}F{e7H`85b*Y{<IT^||q5IU<4QUjeS$>;+QQXheLtavQzJ4}2
zpE!plhFgG>FLyF|K7LuRJw0i}Jc1q7+f?Ux^__QMsZb`ZToZm;m}(ffEH~q)+ft2|
zfXV#u`hqvv?^4TGT+Gl6f!l(Lek-=DwdL8XQT{YqA}dQV2if59>=&b!SCWxZ`NVI4
za$kKbDlY=;Uk_MOC9J23iEpv+qwH{in-W=skbRMAXuD8xK|VGYI$Jk0Z|Zo}VRMO@
zxoABQ9a4R?%o+$s1mgomc(h5x;T&{@3~@SzFS7)I=j0Bnu;&=dqccqE0eWU>!YY$r
z_IvHZ+<^z$u@x#V<MV{nbOt<ypFTpOX#M<x&O0u7YFFUbykOlj0mb>sr4cn5re2AQ
zi;p!9ebZJ{l7c_w*u=lIer(O1%37lm=Y2hlH&hO|O2yn-lBWa<cT;0mts7q-PL|8r
zuzlAMMku+!{_B@g<ng}J*&|;#B}!8z$rN$WVx6-kb$K^sQ3z3gR$C#;hy!TUgp?-;
zZCbLE4{-lf>wZw^VOYDo|NP7^tUV=wc|!8I$ONS_+>z>&92ZnB+2o`yBpv=l)k%0)
zHvV;d$#5gNW^faPS}(8A`Cu_68P2!e)MVNs)MKi=1R0XX0p^<Pw&6uv__(D<+^Rp-
zlL<E}N>s2}>WfhhB-t;MF2_^`L;cq($2WPE=exKFs52vC8dft(1I8LDYO%>qcwgpu
zjPhB66D&<mo}5O4fwo%k`6ui2NBJaHnekk)gh#pjR3B}ZDuZ?{2vFYc-;7aue&z)z
z?P0}<ugI{FWhAla#miPpu$1WCpGRuap|35YzqYT{qB5@pvcp;Kj5JF_O1uKbmB%Ge
zj_eMHH|GYuU-q_rAAb|FY<mBbReWru9w|ql62!LKN&X|8tp8;Fc&RlpJ!Q@m3iK?*
zUPF0IgU;fShmn2)?+}M-h)$*fo6amYQ?fx#Wi9#5Ao}|BSK%VymmWveiP40rBC4XH
z0by6#MqWwnt;~}(uxnoB*yGN~<E>f#o>!@<b)!AZ{ch3=f_j|=%KFv`w=0<qg6}(z
zcP~-)M2nzlT_hQ5S(qlfi|E(HYKMe`h0YiRrLs?A=DuHj>UnfmA@xdOr<Dc)>nM^F
zU2QUDrSp>3@MjWwt~+M3Gj=ws9IjC{Cy1=*SEMaA4jpSgAdc*oxzD?#qX|t1^DM5*
zDu8~sUiF{UuWXn#Ve)I<gmnnypzsXwPWUNa|0>j)QD8ckhp7En+2u~x?VY2x!1AQI
zZcH!mmB#gY#L`*lXaLVHlYC}MMEQ2g>_*CEcU;j<yb_&EH1($xu#ll)vN1XGw?;(6
z-GcRvuL#JGhs~@oFOZ?Ri{@M=&zv!1jAR7`)V<iO!&UBrBK$`pPrGF+h{&@)(GQ5i
zJFagC_=!RGhwGv3j4M}VWLSAQc%Ao>HRf<h!D6Wjvy>rw8SSc9KZ|V@!A>7Sx%1($
zp5c+>IEH{R3(-ncuItyfl^Au2eUTo9SsG<4io9z3siUoNb8jWLO@})lTu2l&O0A_I
za6Fa=KEr&eiE&jQ<sO$1a=tR1v}vi=q0BMmKB2O?AGxPtVIA^R-%0;z9?fpOhlAAd
zGR4AlSDKuOoI0sp*(x)H2KwfnL|QYdre}Oc4zDCI-??));E=6^JiY@Wz6$?!{tU@o
z_HkTi1p@|EOu5Hs&tYj`6a{64w`-Y`<l=kn$JyT~RS1@F+%10CL0=sABWXz!GYXDs
z3NST3bX|9Pcm}@mg<jcToNn)&WivF(yCa5mk#kxSOe~oY&v!oD;TKcg6OHrTTjIyv
zqiT7CiUMt%#K_}-u_+ly;3<UOe(?q}Aj!^>J5tq&yR2*=wck@riEOr|rL&^{xM4m=
zYj?q0cLDew%NPeh3+UXi1J(Ihx3`Ivoh9U&%qY^0iHmyeQ_JNWyjd|4WuqAQouYu(
zXJ&Fi2w=w&qnYivole>AWlz|Qpxm96l<}G?NcGO!CUa&{xy9EFZ@jq##S(;{gm5i2
z4Na>|Z^BmQ!}nk<cxUqt9`Pe{bq%&pk=s#-dVh)`F-N2n=c+nLGzyo|H<@dvirp?V
zp(P&c#yWnpA_`C<e_&B7i39XtXq86M>aOK>GTjlywxYAW?{dtP>82B-3PnvWES^@W
zY#1BlngPiRJw^vzE!Z$=;8m5F)lno-Cqn?+6k5^2%6jvc?XH8B6aDA?cg_q`#x`OB
z7c+S@YHe+^gs}oZMNDGp$uJ2j^%)}z<A|iWzNtWl=`b-$VNSWeI(^7ezB!cbs8{sZ
z!1zNFJ)7lI?^jlMhG06fo+B37(`d5A1rHe~(3+`GR<PuBsGQBpg2|eRznDCsgDHLb
zk)@S<U-wuIjOghuD~*yg%{4=7!pCxRXa1O+^77^zST^KY{;GGa_lZ2v)L-Fj{F$k)
zi13@BMOkWlx%Sryg#1J66;|^vQegU!4_Kb*0z}JUB-##AYHuPE4y^0><z{K)P`&3m
zN9L6)F8c*@5#OAsq3KLyAI}&TcP2S@&`iY}>hAx}T~h^7m$b4B*3xQq3K^IdotJPv
zC@($Yvo*S9peE?~cIxxsh5yvlq9@+QOUb_~mpZw+$i$lA-C~CFq|i7So}>a)q?4!C
zq#v+oOWlEcD9-z?{({W>_t@#aBd!E4vA!~<z>E33do2#kE5lvDTd4lEu!tIYohMUr
z0tLTNp4u%@s#A#rebt(-UU7Vhv&PB6Tzb*oVa!y0d7oFjAZJ<?V}87=S~BW#A~`!I
zd37L$bBmjZn7W8-pQjQd_ScU(`nd+U4c^dbym%k-qMvpK4Mc%gwMbTZ>eqmDsPdjp
znXJNm6FacYxSMO?oM|4-yxJZ5!nC2on!BN{Ke=w*Fvxet13*jzJTIKAWLwpR8=>18
z_>~$DQY6%w&DF$CUfqKS`GLLZ8V5JSoJ1^ZYmXq30&hM^*7%ka`L*A?t+7)HP_Vld
zN~A3^E3OG2l`uZRZ!HW1O`n=(?5d$f>{CjLl(;_^2*4%t8cf}dK3@~sbb7U2QMaOA
zfWC$u#+%PjEN`pLnNIX66(h7KOKz#wzZ%1evFWeh&xU@!fVw%a$xzK48X!(5%d1!7
z=x4VV`D!O6OMb#6DcH-)SmyK_ub$P7{ivs|cCwCgC@mTA<|H@#>IlQw!VM39*G&8K
zn;a&=SG)rvMW%Gt$83Qt&62BRnWPWbT5G~U)$qpHjU(>Zhg?+&R*ki(8}rpFT~!ZW
zgwKaQEQe}ga^+;-3QP6}uaXHol6zF(pu6pvQn#XEVKzun6wNN?p93;5kG=wk`cH|x
zkON=w&!DPTgfo}%@h$bkQpwwrnlUDu;2^r=gV#;o@ES(vw^0(FAnIt^3NhACUxrdU
zVKdn(s7&uw@IG0?+PJ$vmDU(X#TAX_y`~xSL07sDj|4avl3#T82NfxDNt3`%b8!HP
z)Le*flqPU@Uj$RPe>rmx9*pIf^QtzStnscXQ+hmI)3B1YsWE|066=t{0r+)Cdbu1+
zUnP6FD8{@eibi+XTs`&bgHocOP1Z6G)7#4Y_D~SaRvZ^DbMqV*0zG@jRQKxG#x5cK
zfnSg9O!EL?28A!|$KiX!C6e{2@(}okvO`HAHxrNqrA={tlIFp$zLI{yXw5{;zP<-i
z-_Ud4ovhBDW2Y)6%it?~N3BbCt0m@3GZ!<MS{bb-7KnoOfK|-JaPL@zpZ1&!ymdA}
zUW>0!I!<A}|DYP)cQ&$9OBL4$3Uh+Z9%@gzGs|C}T7=wkkb2i~=JcSKrtjI;JY#0+
z#-HtcOZF9)pl<@#b^8>FUsEr(zC+gUS?8cix<QwF>MM6xL7D~=1xTu3C-<lc+gECy
zGsEKh!EW!$4b?WyJKpLgWq)z&H|<w#(u~-DU3*OOE2qBh!6(ZbxmJmL_v%SxKeEmC
zHM!(xs2eZ9fV=4C`ib?BiC5XyQ1GtHDV?5i`nA8%2LDOA3&#h0ert{Z_gL}qD?F@T
z?(g9vT@qCYwiM&gjTJ<dm?Ea%D+nn>+*Rjj^A@bf0U(L9tHVs@turrg(n8*az9j1j
zDmivd*h|3yc7~@K4#Qa(>*ExWl=YT2-mYc&AM0<;%`Td&)YIlYjCuC<XH&$ls2(nW
z_*glOZekLsQp~J3N5I_IWD;$e+j5XrI1A2CUcF$L{gBl4J8jPs_-TZF@%j<1^GyTh
z&{a?*YTnHqNfInlyF|SMCPyn%Wd5o`x+nd(v@QWAY|iL#54_<y_Q>T;W2(Nv>>xl$
zRbq>!8at64_q0zHa$W!QAr+KlMz62P?$_h=o&S4}dG1)<H||5(G9V0oxu-vJ!J>25
zf!2#fy|%?-sK4#VYsN`(DcpiVDX-_-HvHAkW^q+^Th&<xW@BOf$AIW$YgTu`R_@-F
zet9yG&h15Ddp}k9tJ&es%h+PBl;dr!y#8#3BNYYzh89;QO@WFxFPI}p9d^_2cOBJC
zr>@`;06*c{HU;OhOWzz0H>Rf-E$S=`FKAa_;in?uOVns4E~GI0U0&}uud%a-9FxLj
ziN8(lr`kukN&5!r?0*@}t<P%2zepBE3n6Rjp_J}rsvV2cNXBsNLn^WzMw9vw=fgKm
zq0<I>33uya%kD*S&BXG>KjAh*S2@kY!K3=q-+wllV0?B_CFfO>My(KNX2~v8=+J5%
zIuq41+Y{N!t_*>!Qc&rSjz6)Pbaa%nD)Ne5l^qsiXC81CT-Ce-5HB%3)OYQKHgWmm
z{nL5*gHFH_{3s#9UJn(GipASlh+h17)HV7qY!xJ+@42izEor2>jEcM5{R>#Gk4MoM
zD%5%^;(gBVZ~$}YnZXy*R=)A_17<CFMTE2P$!@UNT+=>H2`48XZ{RP-Am3<DSbFQj
zfDsekxSA{cUV`lh+V$xcU#2c2=f4;Rc^#Fp5(iHlQOQ{agN;FL@H*8Fhqw<w+5%LG
zmGj^-j5_=!l3BHHd%BN-ulyI3T>nd~lyfBVE6^{!=-?U5iAo1y-s!18KQ_xt$*^#1
z`Q#;l^O(MxD~d_HiCFoLU=M)G<RWqit>v6Jr_;#DoSf@S)uJ&?Qx!k<ouo+6!Rpf)
z73FS<bSv3}5c_wP@3FE4w$-6+HZLC^JW$m@=tO3dp`F%5LS#;}!_cf{JJmbsZV9vA
zCVpSJY1m2I`a-&I-`%L4wIoa<yRcLh`uwUDiNf!O;exl}eb3kN#_;G(Jzgud?oFzk
zU4~gmr-SV9-XR`z`udk`<u7HI*=rYKG6~&snF}|6$4+j65!DjVi<4=5FMXbXqPiT_
zM<ySYaJpop1qWHN{&>v}%`3@@6;=JCY)%13`*vZsum%jp-;8a2oECG77TOB8&l2R$
z89nt&-zhN#4IJz}HUQA5KLqH{8Oooiu>R4;pKY3Let_CS3BBgb#%nUA^(NSxCON&`
zx+f(}_hl&54LmDr(l>6eeZNqEE{Ly-t*sMJ=NZtr{>ecADBu^kf-w>+?B_StY_mLb
z_-3rOn8xA5goKvLBes{{J}I8L!b9%TJUSd^E%T7+3N4o0GX&M0yUb{vIphTycm$b$
z!<q&L7|w3ZX;a>2SuGp8b8))D`T@&^@ceM1ML5@y#nBL~Q5o9Hlce^Yf-Q#dfYatq
z_q(}=Fs(+EsvYKip=3+<cTB|?RUKE>-3h;B70v@{NFr$I(4$A`w5FPiTCS?Ghly}t
z?=2mrI&Vdh#Dd1K$#|JNg~ATmE?H5EuWyMRae@{zz+m)aJboRaZoA%~kguVwpS&>q
z)%Za8GrM+XPVNlE`P}hXNNTQ4#`Aq1qGIgyLh#KqElw|2-D9pKtPkyP|K-P=<Akkm
zC*_D>#yg+~Gg~sJm~W4$&B~)Zf|U2>OdhD-i}8n)`k5l)!uHTwzTX_)+MTg5+pbMQ
zoD<gSgH#QyqcuQQKb5{5jZ_nH5Q^<k=n$hF5!sy-E>4WkzE6~%7L{RVFlpsm^^M9C
z@y@evA_W>=d0UY`NqyXx|B;4_eBqgp%#?YP2ewwU-R5>F^XDYE_!I_v3R4P45tYRe
zqofXcOvY-jWQ#=>-Q%tZFLmO3n1Yi(#C~%a!*A(%gZHK#Z^Q7pM@4uPT;#TG`6tEa
zk&#>8pCXSsCzI_Cd0o@-+F=@A2N^EF@7nmVfzNgG{rf6`e3(2E+B19&#oNhfa&I5+
zyR<)hTY8@2b3F`N(re8eq*t!|S?q!!klSytDJW5l;M;?0UuO6pZwTujWvvw;%a+0;
z>=d?R!Vm|z;z#_la8lLbw`9_+-ymY^5p=w5t@c(?GQuC(f$ykxyH*}E2Mh4~m3qAI
z!#~AEF8)e7yxluG?}S|&uQA-R>t6t+WCoabvHGg**yH_M19(XG<~SIibJfEE*6_y~
z{82;Cz1b;3`D?3@1;$fMyPZCNRyXTDK|0psa!!0h>LW?m{DYoY7q#-x%=DH^yW@S-
zSDvgzxbP<a<t7tX#R0Olpm>+tAU+q3f54wrhLf?Cmfy5|>67<c=IfsOr*E?*n}12Q
zho>yFCf7AejR%L4G<;*ZgG`BAyFNLs)_i3==6=(yX=y>m8zdSc!a8t$kHiARwi&G1
z--iPTNuEslZ}~#D@pJ$iivRud{sR*5Sb2Icm<B>S*tN_71b#Ix6nWF;Xi=+aK`-D1
z@#kv3WzV7$_#NBJ;fUH{s+mRa+1ALqeJEsI=7PihP1FXPfU|+sQBt&PB;ddHK#-;~
z;NMLwe|EeZ9$5Nc6k*J8q{JMM+SvB_yo5~Vo|ujkbLR@ri*7`fmE~JX4;zmW)7!)`
z8gd;!H{kzn!~_3b&9?0qN(U^%n>ZXm0YCP!wCVWVZTx{Hf7FzxL<0U-G#Glg+WOwo
zL2_`wWTB;oZ%k8Sk#&NeQ&-G5A{YR;$u$i!tM7tbaR0UGu|jea8}0+x-71H+DLkk{
z)`c<T+vmSZ`XV{<?;xxH5~hC9d=n3E0*1l&?m_VSYD#$zHiz|`@xnla5c)H2<?xEc
z6pwSt|3cBnmlH_(|8aT*CBiFpsb8SK6qaxR<uW=iR_xOjB5d0k4sa=i@4DC}jPd1k
zte-iOGBX$Td{zFSkR`up(WK2Sg~1lG*!f|6C0_Nd&@T>^VfOhk`kz9_mfG_FbW?_(
zHX?M10yT?0I3pPm5iD$&aC}|qZS9pZp~}x$inMZjha3sg8IU!*`ywg5kq2Zq=W-TN
zW~YZr((&ie^7}P!YStsclxOljzOULu(2?Uib9<b8)ed2xQPPKpzv1?W#CU>8_Kw%g
z!KR3zv@&zM#XA!)iT)U?XQ~aA9<^_lZYN8(U%cRDHln}t8<b-kKszDm{_9j;Rlly;
zt1PcBR@PK1kbqNVz^De6eX1F4RlIA0<v|=|u2E6$q*KBmXDtFRYE$|SmV9~B$iwxf
zUE(|V)qb)?BOLf6)Fo+s#Lrx#^GCBpznSp10m=O2Mz9Ph-}V67jIQ(vDQO){T-}VO
zk=e8HPfm-hV$bHjtcR+os1m>Qs>b_)r9;r}PNe2rjYBcK5l%HXB`_i#>7~0g{T_As
zzo%3$xm-CFHtKCrFRd0$$^Fk?AS_}^TLaHi0$gpSHGK<NC8vLWQsQE7uxvVa`PtwY
zm+|l?@w@Y->g>Q;A-A1+Cg(3lnHushjmy6W)RFll39ha<cCKDa8q{6?@Ul`++f&M8
zL%+082nJ1UDxA+=I(IFaO$>XHcyToiz0{e<`mu&!rM(W$yJqEPKU=Us$n3MkCYw1a
zGBZne{j%Rkhcc(672N+>GdmiN14Mbf%M2cFwnSF1ibN>f_Io~B+tk9_uE4HCvSSi$
zqSX^t%PcU*6Y-8;>Fx0+kr7UMMyVq<Rnf|rU?)~BB}8g_yijC8qSUWlX&E%>c;6GS
z&ZhU+(fq;HvnBSv7&PCHZi=|Q6+9GHCkzzgmeC*Xt|jKgU<7lIn)?{qtrsCwmSX(L
zr2Da69YyeOor7|*s(0(D4gsv_l-3MWt1@lTZv}OYP`3;mV9c=vG;l;g(z7UpvOyen
z!V$4ED5?GqR`OCy?D)9QiFfIofverUy^U$7UQ@;Op)ZWa=w9ih8m0wgVUG^Wwr4iA
z^NDS<%&bk=X(prc_`b;A0r7FKpWEt~7XJVtF{4W$gP1y>&OAFtwv&~hnQAr0UjDh>
z7*FD|D!I54!&g^|+8t$Vv2)p&a^w~-M4t7YfQ}yO`HH#e6=dHBoPnt=GO@=SCEGM{
zGCR|~M_end=I}RR>t;mN@h+|<iOJgDc<9D8@F8eC|Kz}B%q&6ApYZ}};W>fyjf-KW
zZGKeb^}+rarH$=rZ{-=QkykFQDJkNf$ZeJK@=`K?SZ5ouJhb*8{wodu0&7fj7oqPX
z%F8hIXDuRw*x3moN&NhD5}}&u?(qA9=IxU3Qnk`iIl;`&M~`;x8S&0#EFtDORBQp}
zMyIFD{-0Uz)Mn5UR+FZGv&UdkAnw#^7v%P@m^r}sw{!eyJ!LUn$@?|ZDW23d;->R#
zUh68it*zy5DY6B(4I06Ks3_#SW|CDcqK-3*ThnN((p|%IPhFEO$g9wYWdj^49!=lh
zt4dP`(j|JYMaulLlde1-j)lLx>E`<#)_Mz7;blR>Jsc|e)8bf5U;7<VD9s$~3fQbS
z`hxVh8<P&TEjqXc!-I6LXGHtA$f>?%A9_4AZUb-i@lcpNGt5W6f>L===^`y!mb0mC
zrA$0Be2W+7%+(9~<!(AQol#RvL=JNt+^_xQ7v`-L=$j#%*IFI4@t}<-=C$s^TNM?k
zOqo7=m+dTmmud*?B6p371pCA1FtQ*P5|t$xyQ_uwkY?RMMD{un!}kU~ers&0DBeh*
zoj-#WDKJ2LMG#QT`%-+jJORm;1&pMa1bi*|YtsO*i-m8bMtyM(grtmuIB{owOvqj;
z=*bioCRaL)vVGG=9xNunw$=K#CSEM!9p}@q+^x$Kfz(t@D0rjsOy<4}Q=z0;(${aU
zy4TD8*<4!7u3axp1rr~ntEKxhOdF2DB!9LakZ49?Z;E&l^eU378BwiJSoGl6ITj?e
z5tyv5X<E=If!*}?Yr3Gi#5P~CFkb1{PsyBj-~f1#5cht|X1FAbdTW~H?iJU!L(#~i
zJL_3ARD3BV;?=Q)v`N9kBt?E&E#p|)fpJv9&OHlo)b7`I6HHfIG%Ebnh~HctCKcGU
zz3wpwR`demeT0EnmYr0e><~ZNm0=2dP^A9GvFe!9K2N5mGy9^%wa{hCo4bSR?ud*Z
z=dFC-kwQJeen4ROfgcQLl&bl@eEle0I?Am}x~@Bt$1M9}_lK87UE}sNTKm56RwnOl
zX0R5zmbuv*T9((Ux!6)W=kIm_v>4rQWk$<l3YrzaqOv=gDD&-BjrQ?T%cV}~df+RS
zV6BAwn-W9se|TK<9n5?f`oYl18U<opB6p{{ZI?xkj0-<io(N#t_1QmKbnuk<kdi7&
z%mAu8lC7*0t4Re2%5p+@`GWg@o&zJGM1Jx}B26Sg-T28MW2euAA;R1%R7gs?Ifd^{
zta7X>v-$M-V}L19@NA<930eXX)=?QCC$M5p+rg<Duj!46D^#*OOTE*fzFxHGE%*6U
zGC4CRRjmj!`VPG>Vz4#&L__lKY|Hu}Y6*ezmxD2`Do7%^>thtSP*&|FtD3!~lB}V#
z<t3hW`8&s^HQ(NhceV6-aTyY0QG$WLgV{)~mDFJpYnX*d&YVy#b&QPlqUKvt($VL$
z(dUJ}b1G^oY4dpj_#+D8?Kw>%H_*HYU?3Nb7r3pJ6)_(?*s;xI#!z_P3eFg?Rlvif
z84kfh;+eTj3iJqYSo6(7^+J?`;<%i=CD*XaS);*ck4lNMm@)d8$AWLg7PJKdtnG~P
zBd=F$4i7%T|M!@F=Is^)@4)JTPS2tDPNvtt^9C)g)c=4{?>Z;ou}j_6I~KeRjTf(i
zjuib8dLKDQrXF=RWe^y#KkQ-oimY0d*IeUmGw0Pv=0ZxAe^)4x)2%rEI+oJ6ArQgZ
z*g2-|{24Lyf>M<wA_px!CLx8C5G|SCW?{=baIP`Ui`A4!ROVRDaU9~|sxD#B?+OQ0
zJ45k7F_id9wDen)Y`YBOhm|I4-|!Q8DQ51pkGJwWHJZYS*W|?<9&S6F*rE4o(9%0K
zRl+7qgAdU#A%FkFq^Z3NU|V|a0=x|lKtHKLMXbuZ_Apuq1*>TK#yETLJ5QcU2l+Bq
zQT?urogctj@us(`x$;b|4u_U2@cU`wk1$wm(!@xsHjnu^2&Zl=pssZ5BD<%^R%mAW
zbGg%M6E=d0C1PK@y2O@eH{CphlR|CcZup7Tj+NuJ-xDp(t|?J<Dl|86=2jP!L-x_!
z26_xYUuWgL96&z~M)M#Piq{07jEgAgj`J+R`nlh`(@bVm{tTlbbC0yv0Lw3t8mSGr
z0ofSXg~Af;@doc9PwHuFcH|)f#?AWaR3OGxZPXna&fCqx+b~&k7%7l%G9y*_utG~g
z%S2XP$Yeq~nS(Cn36rwT6Oa|3E=e)A-zT#yD`IV5<auKx@U{`&Xa9YAC9{1fN@xOR
zClPqOUi+o~XpUKfjp*RB&-@|<0@D>;r@v+5nBuZX@kG<JI=cJ(i6Y<U@&~Vv0r=g@
zzHdn$ZB)rwBV2y~soe$4Rl5R65K9rKWG{~P4^A_A77vVGh&mM=vGa?m|B%<VE%bUG
zEmsEp$hgU{KiRZ+HL)+Xl!*h}G4>y;@2^}2-fhO85<rIbT}{YY;GajEj+E74sMj33
zXUEj)qBxRwOcoCowW;Q-Tt^)$P)MtlLlW$eN#Xk%-;$?JBsybXwM2QIOGxGH@O|}&
zz8eC$ANcBlhLjA`*`Q8`k9Mtl!TgSlc`~Lu4LIxgrm<d~Wz4BG>$UQGzB%Jmm*ti4
zEqe+j{M$8o01+0c5t2GvJ-c0Ykl#=f`@gaG-ce2cd%7?R0!mX*kggO_X-XAoD$+zi
zL~4MjbRtBg1qlQ}dXp|71nDgxHS|bVdJVmV-bttdLVoW)Gw00A`JH=b=8t#Hy7#X0
zN3vG3l3l)g?{9fNPf<_1{jFZFhF>IavpXG6$9zwl)htpGSmJv6Y*el-YaOM|?(ggF
zXext_l=KLLQT|Hz@IlSve0K2rdBh<tESz7ABJ?_W4CRp%1v4j&jgmdJDq)}JnHbFA
zuZXGh+R!%l#1p3|&nzO=T)d}_aQNNvT0{|BvtW6mnwBJlzlWOEp6l2061TasJIh9c
zZOcRD;OwESk%|;<9c`H<=)1Jt@mn(JOSC2#B8(m9g6XUIg_1ikMD00m{dPmLcmp09
zw4-1GMgJZxma67C5zVBl`jFCo?q}0{LCRv~!3d`LB7BQe>1bp-Ie$lH^cCjtl#~)r
zN}hzkj(&2pv?w!?90^E6B&e~17yq^R8`zY<J#~1RUH-#v)Vg(xH+fLw8i@1PZQrd+
zl`9Ve%bm0AW&}0r^*nL3;*GN2fg3MNt1m0Z^0lJa-(JEIEgm=BIG`9E@c<xD&V0z|
z$LZP66~ZV7{1Y*^I~8LV^Ka-1n|YqprC4T2ERwRXhF=zD>6oRK(&aQ+s|tYt=%X`a
zV52smc8tFx|ACq&tMxB>E5mcCXe;{54tzyxYxum5&dyhp2x;nxy{ZPyVrzM)sq8)H
zWUa-;B2umE@y7gHOu!~t5+<%SD(&dfeBv1S$*bIScw1VoV&iy+Het$aO1@J?^;&%+
zNZxVpgG|u*B9g(dw6=IFZ$(3USfTmq$Mh<E=i|Jw#^0#%vbwkvTojFe|02f;NQDTR
z>!T;yfEiZFSNdI%j9ucXw_+`GR2VXU$IEv=`%Y*5YK8q+JsVsPT$Y22(n7WU`;&>o
zr0?({Q?=zC06Q^Qj#2VPw$#acy$unA)|~)80gK#D6@O3!o_YAZJFa_01o6X-2s@wk
zJ1i7~S`r#rDn=c*G6~kiG=?W8(6{!3D?NLD0qB=E#9mKn?w(jC6QHdOM)O837^a6>
z1?b@YzMw-3A*^)lf>#NheHFxo4yk;`AOHzLH}Stq;J;LGHV1-u`%bVXpTzg3Pq-C=
zrzTNZM`{=LtOK+pT6@CFDEYr+&9OoCjX4sVdCx=GIJl0;m+)KrC-e~3Si#<Ss)$^^
z1&7I)piWAMxmy%J_p^o;opmju&>g^Nqh1yL?1W#(x;NWV`7G>=RCe5XvVlO7c{hNu
z2d;^0kZlK!ZofL6=#fc`yzOfK)|+c)dMZhEdQ!T)eNe>0O!!AEM^eb$57l!d=>-fI
zY@uoPf`a%P6gpG{uFBZpsZ^2&tKE5&Vwy04N7X~TZK+}@m*jhvo?ez{8|J$T%^-`y
z=@L7zCl>-^ZSlL48Ut^}t?lQds2FTFpHp1IR@Z&Han9h;i=geu_xH^i6Y<JT{cy=4
zg^NY2RTg6X?|j{ZS`EL<(1jdHR_yZM>bsvh{reJy$a_G4l%gLSV`$_wgGPH7H!cln
zEL^W1mqoab?o}8q*pQ*cC7tbw^HEGBen2KDI9yZ-X0dO4P#&g}*7Yh@-|n@e`Y$`4
zU5(5b6aLB=n1$M-?rZ26t@|fxHJso8AD*EDCI&4_Tb`60*atc4`0ryaS6NwO9+t7k
z2s{qDQXecgK~=K5xLmt#bY!}RnM+(lAV1X3b{#-<1-%@J#(+#CI^aJP4S9wCplFf>
z3_ZPo>r$SOm=8BDBd@PM>{&#eWfqJ%9t+ikm-Gv?p`TZ~Tg3G20013kF#&L2`Z<a`
zxw(;F44OxiVrH-ZK{3X$_XovOMG6<*awsVl#GC!G(o=U6m|h5AJJu{v|4^#rDnhK^
z{DZ<75&se|L<%PJ0E4C*h8${I_gBfnpRpic*?-rV{LcyQ{=CosN`UuJ9&up1VSymf
zm=4o?R=R2N+tb$m`>J~)oy(QMw$0tvDoup|mG0=$*3VRKcb)gFn1%aA|3!ue>ucMg
zh+?+)k@DB;&#Xq*2YQN1|7PxKPipMFo%Y%$o}JxPwAHVkFSUisOwKb=skQSEUYC{x
zbuL<Kyj3wIEnmmEx{VSO)7ui6nQDIvGoo=VbE4@zY@az^=>3Djj+YG<)_gR>p#?f~
zX=%J%eX0+!;KR>^*JdB26m%TAUb$|2kuxE#B9Mk|$@iI4&pV=W(~F6Xgh>TnDDI4B
z=Hj<WPu6_SDXFU77^;A;PT+F?YTw7NI&`Gr+H;N;<|<Dy_{8AE772UzK@$cUh>A;g
zALZsxFRK_n9%WHDTHXA6s>bUv_TP~0tus$D#;2{Ldk#)Vcx`TH!IH4kahcdYYCw>x
z)k!KMI>{sU$FNw>&mTi2&Rq>#(%^b@bvu&(GyUqlgT9I8+1A;MB+UY%mnwlNppq_%
z+CfPWznzUT1<`J|)!!&r3lj-)g^HM%=&4Y~&(x2CEKe1`9yrRI){VwELdndT^9|4J
zM8aQ~Yo8$nL))&Z8Q#l_$dnVkDah!9pA7Lly?a)In++t1z;)55yiFf-KP!!#A(8T1
z(V&jQh%{(d`LDjj*gB7y>&t1S-oOH;3bAVlj)<_^Rm<|M0t%X+l+UWB`yLZP1WO&_
zcZTtSh1y{e*MrE5l}k@qWsM$Pk$NGsHF$mO&A3q){j)%k;=&oDRJb%zk*yhp(D$Z!
z5joU65m0fVUjEuw4aXh5CyL$loL7p1^BYC8oF2{j_0<hHEYD9boFRU0EkQP15|U#g
zvd<9a@Eu1H09vrI1Kue9J1^BgYwlhZE008>%>T{qz>ZR<UUz<_MVc5w=)(;(gn`4<
z$_FY-lulJf#4B2^mfS)vWlouDaxyeo-DLkUd9i7-qe|}yF?I{lCVdjm{$LZ1IaPzp
z48JK4dyd~uDux&W`~gN1D7QnTY}&!4Vowkm%hD&TDG(!PbcsDoW{}#sBN^6$>kC^X
zelOATc&Fc*C|mV`2r-SqjVzbujZYk3z8yJg1p4K;X<%j2(juE6F(~O7CI;HBy_+e-
ziaa35d7qrke3x?>e3cLaCZKV-2bL=sYU29QoKr|KQOf}D5h@+wCLN)3s!47c5GxUN
zv)kUeRqcCL{9uC4rh5i-mOVfAs~h2NFgZm2yY;^j#LTF<?^uw`^cRF1Ne+I7GzZ9M
z9T=V87q_>tJ3D9cZrkGne<`jOET%H(Cx7Co0;*X%Y^RyWhDPBe%PhDOzBqSY@#9j_
zzPAjEVmF~lU^7@gc)`$JVgH5ZJ_>{&9)+C^kP3feAOX=Bn_1pqbIY%5t^69<aWV%@
zH?Bmdjq~bh&P>qcj}dw2-J=MXv?>NSNh%JGPm+|;3etINH@16aIIH=XUFF8VIe)V<
zQpty~a){j5%CndpQ<WN+R6CL=a@hUV#2?PeFVg)afGg(FwFU51M>r%M&Pi~}ou3Wn
zFh`M7xjqKQ=fiB(>jvu!gHzzn+UyP0khp}$TsBuCaO3aW0u}UIS<B%dqIwQ2rTrwb
zE(eygGcOizpHK7|Z8fTjwqumOegkyX*gC>PM3uLfQfxQm#az3ps*Cm$#d!?=j^N3O
zsMvwg%cIxJr&@p#A`FpRII7L5G?~#TNr*#SA$by*<^v<r!INuu=He?W2AnFHf|C=z
zoVcoW%Egf9CxlI6j(PNNS!_)kKU%AJtL78i`eFl}ILj+Nt<uJ^_%SG8|I=BLea;4(
z;h_wgDWkCKZfj)0u}jTU-_LIvp45n%yvoZTa(gI<h%avRKJ`p)W;D2_RHB=PTsaax
z#Ee|iHy%MrG4P$f!X*AGE&H_XBsc-Gvdd2TaZZSW#a9x1mzPBg&)aG$mYYw^ZA)H`
zFMvNBKA85{N=}*4IFflg(AHL=FB?x4j%Hzq4xSgXF<R<znXt&7S7soVw>PpjdZFZ>
z50g389}Wz-F>zVIlGkMnUj=$<cOo+CM9K;UB!078olE_cT*@_^DeM&2AkO0jpqq@-
z?vo&G;a7{X6FSN+OQf2_1IU?Y4y@}r)@grLq;dSgun2W7^h2%p&nml%W;P$vOlxLn
zBfiS9%fG+K35aR~-GYsb2P?KbKn+6*g1pUjkjYsg{p3#h_gojnsG=WM=v?NVxkFDe
z7rNJ=K}^K;{rrQ1jr0OfM$K!{!qq);$5*jFcL|1>&a4qxT5eah5t0^)9nm7ub=|#|
zP(gajH^Mi(bFXF%yVrq_71B#44o_pfMJ#o6oVB;Ni0u;Y`<TVr%q`PB-RnLBjg)xa
z;7SHg(6BUMTOsviB4Fwx<9EdfzfMU`GYfUf$Cti*Fx9+zd-02K3h^S2my@X6Hp7P=
zEiT&HzqpCAsT#oGq=yysT&PZ$n3R#;P_EJrzJfgM?v(FuziYw?X#)Llmj!@u!3DpA
z^HOF~$$=y!J0TG$hoilaqL-js>M#_)FRF>Q`;k5FxHLAj-6NvL4VpVUnp<LN{NeX$
zZx?3z2>C*tPn$!NvnaZnz9{#Gf4~2lPDyNK@B!w$a{HN!+KBX);RbFymdb~tA3(H<
zFT{Nii((f!{Dl+xeXdHQ==&$@b6+c6IdoEJd`Td@%L{x-1W9$ikEoTKAgXn$+sej~
z+L&+kPOuuX7^-fUJ}L1?1tg+DHF-bqh18(#^hZSM)-mK{e8PK)YO&HDQv={Oit?Cv
za&$)H9lQ<70}2qml`o@|_+U|HgtpPJ*8?##3!pSFP^Gu_Tw2?v|FMWAT2lGz@8yyh
zuHTiga$I3lD3YaK6GbcrD)8C^J5orFuIyv9kJ7McMroxV@oCkE)!#ObR-Wli4E#x4
zQ<;Ui+s)d_GXkSvGQcE>1xKEkcTEHMD;V+zMKoe$>yQ;yIJWxGu{RR{c-JPpi)>7`
zx0j4FLTC&<-k$%ot_`B$=Z^+oBR#{166+2>5>MxKFT<i*N71YvCkNSCl&wbC!-T5-
zDLD;NaE|fRc_uN`E(!VGvABFbp`n0uail~+K7IY9r<pj~g11j&BM*;=RkBuMSr<VG
z%zgZrZghX=hw-+;Z-nNEC0gnF82gVMN#Yi1wG3-AX}>FGGd1hnK^Rfpg4+eAJPEZt
zVQ+Fp+j=I(Cp(<P#BbT>Bsv643xsRQb6J%N#Ar&p=a0+ty`9voe4P<`hyKrV!T<PK
zIDv=)=DO&1<wI(~dVhe);@lq;RI`N9wjEvorLX`%@UGLs^bOe<^kHP-8}wq@vIOQ)
zEbFucZv3N`EQSP73SvMDyr+~Fh}EEUM$1D-;0uW5eu-hpowF3gV)EFT@JQaDZFv68
zQm2TC!GR+u3O9YX8A`vdEse^0A+Ij_{blDTP4;ij(I5g~R1|d9MmAtL3jSVR>*Y*!
z#C%Kk`0@L!@K1a9R;YLsAsYJe>UP3Oy;=5!J!Af4W?*S>!YU0uJq3h19jgH0A`~b=
z2SgZ=$rA{|i8Ai4?g`BdVo{td7E5OS)p4sC&$x1Ad#ngOFrDo>zV7W1wdeB|U^A!!
ztb-@*g@EO*qcZV~7at8!9QsQO{-Bth)Fmr}?mbVu+#%vvD|huFI*<Rx&_<#o`$%IR
zfO`LVhkO{ohSCxs31F&_Bp>wM%Hm9@x8mkK%|<P`&n(4*GI9|Dl+k@%d|Z)o2h?A>
z`-mDPK)Kj>1hKP~3phDt{^^`H{?*Lkzu3*c7()DWLw|qm|8yz-*+&2M?fTC)`oFV{
z7`>VgCD+o}2Sv8^yK2zofp_$8Y*{~|{WZ5MuV-3@D1xG{o>Rn}lphoe&5Knh2Yn`4
z<{+$Q%#B+t6;WHO^_3NV+}Nf`*K!RRiWp(Gu_l*qTvyC1|MHKLi^LuU^b9vK={pD(
zbb9OXgfl)!+kT7B{ApM2V-EhL{x1RXi611qmygTNvC-yh-r5X_wU>$BH00q+STKmv
ztsmH_NC|lpGcR6QK`-#i%BU&SsEPXNMr6IKbRYaFF#>^$uvEeC1XXB|)aOxFy3&=Y
zZ98LI-_f7NQ=Ao=1gyfPqJ@5Qmf9Q9g-6HYyfro&MTc-fZD^hwUS9I~Jg07%zd4eM
zetkZ^fA_*0&PXG7cKlCY!;)5}iyJ4rQT5yl(@a{`g!q}LoEf|10KcgwsHZdZeZS^P
zt69gzI8BEnHM`-W@W-4_z9z&xJMplHoQ+8Q?mT`(#^|=LD##Nabl_5<F9|)LmCkuH
z&P}dMuCpNsV(r7$X>>Krf(gvrYP)$V)GyOR2UF&A{ZJ%v+j$c>IK2Zm*)wZ1GOO=v
zs#|G3X*F3T!zy=u4u2);5u{rIoK1Q{XiUWYLBYJP#PT_nrtQiGYa;_F_`*I`%TeH_
zO<cEQ@|tyAZ@LCQk14}x;CfRjQDxo{()v+07+Nn;>xf!u)tk?+kWboL^p==*O%i=t
zw&@<;Jyf9a^f^)D{KcEW&PS<fDTpo1@@%4L18eA+D|BVY9vw4UB0IbFP`Ry-k{J84
zY{(VC4nr6Ym!KHVvg0H3sfNrZ_fA&_%o3yg_k7i9l2;n($^AT6iNA5v91u}kHUHMp
zWUGhqqk-uoZ>#G|ja1)h_%l*WVeUU3cJA@q2_iGV!qk;mVfznI)|JF06fAHnBjFlf
zJI+lo^sNefGQrEa!rJ#WEvL>sB|6f5LL3hVP(y1_!A+&nAkKG=)<&Wfm%^hhBu)A#
z+nhwRqsOy6TYgy?cl`2n2~9aI^L8brAlwG`suCJPW0q=l4l#2RBuBT&Byio4Fe1W&
zv2<|X1pat|_o6-9$v4*BHu{ew36=XP@1uT2lgcG(G|%6ATxa?EJwzW&%%8(tf(3Rw
z!ji!0t0!BBvt{-l19Pd^WH}>8cIVycj}_=Cd*Hu;t292_hKJuHP^Wt1@^s3Tn}xxb
zCJLuObqG80$>#~Uqc54R2`lbTSp4gj9(Z)j1}a~Ln`FTLplF$p4LX4GTMRA)&ge2m
zPz!56d2G$~PFfQxYc7@X)XLD8g8#~!CaU5xEo8i-$B}vVvBk*_`sI|7bSxRg_6G%8
z8oJ%!<+Yu7`UuAK606{uiwWMJRv_M5mUpz7KRKcO?hn_%b$KH1o?n&#L;E;O@Ys=#
zwkRqSqG4a_D31i5#KYMjV7kHV@J007qVWBT_q`(mBKl}D?r%bxNh<ihc6T83t8&_j
z>5>`bqg!t1{wzzAL!x_Va+cnjuob>z`B|N9sNTx$I04(YZ=at{euOqdudX)=4%Dl=
z;rcdHy!wJM_&DthU1z6V`-HA6P5Wi7(g{i7URE{ZB)l~hOQ7`SiXvn@5S9QUe*o+b
z;rs0;V4#Q~uQ`bad!tRy1d7;I?oeI%qW;nJcGkWB!pHIaKQK)G7p_dMe+8GWwZTPz
zSt=MH6P!{4-1ODY#n|eS0OYO~i5GAvgD9U}jyv}szz|<)U{bp<BYvd5TTr|0p99MT
zum!i&heJk#CH(iz#%3RtH$CJjSp8W@xvF+|<Zo08HHi0{Y?o_ip-+vsjLO_^xKkas
z2+f{PIJq~r1Xsk<;DSC|N!WoI75FoL*@P~*Bfp#egpYl#cxlbgE5ZB>Z6FXGIyutF
zP1Luj!fqWX)7wo%G*(+aEHe5Cf0>hRX<*!qiwkiNcp<xT?h*IhmGR2H6WTREM0HLG
zAZKK5POO|!0bb1{ideuJU4PodZ(#bm=}r?OXy$hHpb!>*5tnM=%_9(`9din}G#6Pu
zxg;R?<=7#M{=(1e*ZQ?^5j<`qpOdx2{f*3%gCSPRj^=3=w-%24U(@Z`E-hqB{{0U0
z4jo;{eh(333Qb_%o%gX&dHvLH=oMgZG$2+inwz3Fv;=s`Yh4$N8DY;r$3l}X#YmJ^
zwRT1_w*8tj6bd9W5^0*MVoX<gXz<HH5CQ#1__bQ?hFYl)n|m(`Zd?F&MZI60WRv@e
zDk+B`_pb~Al7RENWJUEqop!amd1!J;3+57}N@Co>;`+b=o!q-b?U<SM<~x>N2CuuE
z4C4GR-?)r)z4FaaJ-~+=E{hiiqXQdxVMZ-I2q$c!W42utwtbledq<&yv;Rmx+MMlA
z`}otD&z_LcuWgC^IHW(BePU{YsN9^)8z5TzO;h0^Pbx$5v$@xj!SO1WAN)oYI`Xym
zl6OyLI%nu6r;!ddk%A^dV=_61l%f^yCHq|qFgBygtNg&Ue~t=C`WW?hr>b`P$zfFE
zM`=SteN`D6UDs&sRT!LM5QB?<awpoi>vH;Sp$8O~n(!sffN-46XxI(RKvaH&Kg`c(
zhC(Y)5_I<4A#Xc+TwnC<G5fsrltbzM-RA$?lD6-6ALs@smv?+tx<0XHgc$pHD|>$9
zdR7pb8Mzt!78A|=ZEMfRp5oHv{d&CerL__=3yk*_As!R91Lwq@_OXJtcHK52#Aq7B
z?3%Z}<q&+|--DsR|JgpX;!#d1G>H2fxCm~kX%!vwTWzjqk8`}b^xmZzbZH@_etm68
z{5_>1U*?iLha(C)LKk%D%t4LO*)ZW|AFC6#fC_JJN9LJrjkH@yi_BN8FN#d+@kLQT
z(R_QsL-{{xT5Wsu!Q}SZ7*8F*kel7vm>S6_YwRgT^sDOvebYr{;sNj9{o7aCuIN9#
z57QY;g@ylgpZxt3LTMP<c`5@~^*-iS2K>PVfG%d(JH-sJI*ddDEGK3JVV&atlB0n9
z8AUKf{#|R;U_!pE=i^7)mX{BEjS;0K9s-7u;fVLd7IqNV=?iN0e=oRM=%2~}rB><E
z6QjcaJ>lAnH_DKZHQW>DtR>Oyj%q73Z|0hqmx;OHDv|Na^6jU%Z{AC%E%kr(-YsV8
zK<}-UIk+@2>lnU-=NDUGh2B8Ejx4L%ggno<XKoPOHw~@z!n@b+kV^K{W0Me!yz`)~
zQG^&-Jq^aIODN)HkjhhdobR33c6q|IC2yUBF}W(*Y${5!$1Py)>jw^n>IllXU&e_?
zkkwl5k^XV(_tw=icr}K_^=8e<w@WEk1Uv=4ZT{@TmpoohV<s4)3rXUGB_xw{R3lmt
zl|NiDed~lX?{&S>n=b<BVz%p7tSK9t$Be%C5E|%O3azDP9U?xn4S`yRjDH`xNKs_G
zF>Cww<^DhtqSiU|Ab&F2^bZQjwy?y69Y?g6r@JU|Ve16>$yJ<jt=~+lb}n$xD>z0w
z;gD>xmb7S`^=hhcDPn>BiyOtrPC(<0k$M&U1Xlaw;CdgZDq?d-X#^rdZO?xUUu3>p
zs($^wGs$?)j~d)sxCYo51sF0}e(x!9*j1;02p6_G(o=a)xMq^`l;%i-j#-00_rqO9
zmPf-C7{8ARL32sVRRo6YtS{s1JU7Y{1Zx{;d3)lyWxuAWR0f-EO|e|pw@`FA(g~<?
zoJGx&=)7lahX&H)EWs&;D1be1+}O%n1|}tsHuIidttdGYh^*7Sg!tvs4nAblJ-bm+
zOn3Mg-xpTTc`5^=R3%gga{F{?`O#FM%|x^AaAe&}lQ2^YsgG&tmT;7J|1?9?C#$4M
z0Ob!ru6-ffnOjI2HFhi8+nK^d0h{{u4~nedvZs-+7#R^GG3x0nH=>Q*8(D@H1kI(X
zB2`)w=}qG1&LzNzy%Y~D6BKu~9oHqbAo=E>sI#tj-B2=2ap8+ze)gD&=iK$!S+ujs
zpt+f+<kPC!mW%Ah2}9zb?JBkDrK-Kv+m!uq7JjX{=?h~^$<0%-+m@yiMF{F9(Gvgt
z2R0GkljUizEJQym`w{V&isv)OSo-<U+yA4lOCSDaVW>$iZLD}fb@(re*)jzWhNaO@
zBo#)avtTwU*{vBg^+Z7YYTzx}PJa4h8me~}s^?ffnf~jq%i&<A2?Zfk+ma{QI(%cv
z)5nI?-omL+<6%406{{#CJAH%WAaX`*ke~1tW?gd4t)mohjbXpeiZ-EohbyhXB|T>U
zHT9tBYPmV2Yx&sJ=^Jr&X=hl#d~DC%_A;04JNAtz37tmM<Fsq1ueA$^YD{#*jz=cS
zRg#`KjTzC!LI-Est+4=_8|H2H!yX$OVkaQ*NCPv_i=%e$2<&(`p*IsyXQSi_MR_$M
zVGh7?CAA3_sU_LVpYk1?R}T}NC!fP7bCH{C*R@mHt!_B+I}KsveH?ZQ7lHa8>^98k
zFub44W-@735?Ik5)WN#h6s}}uNU42!dy$nTAcclTNa3(hHbN{Y!d4ld=+Lroo;dz4
z6Yp_YG^077G?ia>D(?~ZT1M{c^fAbP<tT|0qS`t2Hto#EISO}cXR&T|&oOIpe|Np{
z4uE7%=EG+ljfRs~!lX(VX6&kJ<44DHu~0<$rhb?EUhZyIqRp$z7y${cDw?kFMc?;I
zf^aK<TK~W?i=dXNh@9Z7u8x#9S6V*Qy>)>}q3Da+W@&p-cB-#lwbichQWT3Ru|*G{
zi-THByU{Z;bwm`(cx@1tqE}P>qe6qlIP#i?$F&GO$>!UlpAF>=+;@Sji=ut#b)wPp
z5hDYF_Ds1Y{;2)ELn(w~YrMu^?~`La(TUH5RU||(_I4hh)oxWuLiB}zBIw@nhdm!o
z<yUTZA(Lvc8@WBt^nm5rYF|GcBzQ-3K8<zuY92Qym9@q%h0|B@1WYTGT`OlAHtfK2
z4=csXy@r*s#chAGF3@&LPJ5SOeSf^L09-!JZW(Q*DRVcfTJApS2{Ca23=k%<?-uK^
zjL~!6kBAEfR@DjC)dSn`;GTTt-GR$p1rODGXG@;|DdEM2;m02l;7}t+*K*L<V$RZo
zJEEYq-7}D;2ga(e{B|+hoOe|@Q*#NE_B{J`?x=YB!-$nhA^WsDPTV>70-eHmla>m?
zFTA3lzqFkfX9fzJzM33UI|g9ue@9gLUvcICd;IvnEqIEBakiuwfJ^-su1X)kRS|Tr
z{D<{RI%B9#V1AF0z^cQy;d6!1zmc=wz#kK^iTI(W31o;{LgFUt?du&*Da*-fucKQy
z7M-G}qpy9C=)T|t@Lf9U!P`+w7LZd_BEK<Vv;{2Ga!8nkd^9Y~g;XU-JmX`Fvh}>y
z7u`^;e<L{E=I4b#4R$G{)@pIdVoBk#s2ZkZ_c)da>flj`(MK$RFRc17;YY^>YQ07X
z$`i%!{Ct0!w>oBqDu`|MaP1b9&9Q>NLCoN`M1(EwOvN18VZq`5bf}oz<dv^kIEku6
z_3>86JUM--o8+T*gVYiU>4RGnOlRjY%?OUwWbpt62(-u?=0K`UL*XDBymwWcI&ba?
zhFB&)*}N*kbW4|Cc+ZHAApi4ZH0VIwFS#MRBTPXKr_8j9mSJO<YhY%0cBEaO!0CJy
zWiOcB6?cb=z!u58YQ=uHcscB}N0GUzv3`~mQL#SST@}`d2@*e^wdg*fF(Wq2f$520
z9fDZt0+<nSaU8Dl4l`~Xc^%lh@gYI?uFd<X_zQu9AI?el5l0WBi7tSC#Tf5|M8V26
zDwRY8*3jblAOTUfxLuK@<i}r;cB*bubhV5J^Ub6BM6YzBSO>tWRU9G-6-+@M#vg?$
z<LE&<ms>nUn}0<MX1S}1WTd8fvS*dIUwd`zG+v<0uy5}16T#?n0Sr-MCy40T`3Pep
zzrT|TtZ66<bsNt7w1ln*CuMbAp1$Py=mGyle^Wtku-hLL3>%2%35Vd=vY~PSnmQ6!
zHfPE$Q4)g5>k?AYpT+7;zE8rc))an8u~*)HNjA7qwdh-Qc->HrG6VUni&#Ps0|&V$
zQ?YWqz*g>DUJuD!b9XU2OPDU^kQ40Q+jQZ2`@l+|K%SZ6)mz9in2{unfigoHuM(<i
z2uh+)jyjkj_V$@Xhe2<i1MS@@lW7JG``)+43&sYY6E*of=|j&Szdqw8jMNJ`&#hL>
za>Brd1hr3jQ^5jJ>I2)AcO0!~cLh#G^>Tv?Z_r6uevrS}dB6^(?P|bim6)3ztX7#i
z%BfcZ9xck+$XufuR3@zD)u6(id<gFql$7`dv9!%0GtdTro;fz>?<LP0J9g41P)X1`
zu0N(OvO`=YvCShGfeO&$Oc>_zNCj^N3VE0vnM;Lj^m>@lwp@5^#$@2hia>3+Rp@bG
zpEe=2J>t8=Xst(BJ>W&r!Xt04);ZKF!`xM3X0K(n8uQ!nu4s4j#EGuNeLl`04TI)i
zO>|&lK*l=eGVi>4WW6NL(b?5Lwnc(YIQ5nnw~9fE_Hgq0y4{wL#<ED&62EpQhYHUK
z)deAKv676T#neiKmu?2cdrM~ROxd$gT>EtL`#$;m0t1akb)DZSXs9T+T*Im+a(?^4
z#MPwk_vL;*di;fVV~>(F*!gO<R2Fhu9Qj*hc=#KM0XPs4+{j$d`mwUT(vds6J6OXt
z^5rRaT;ch10*-QP;0GkN%3-oN1bMAw6V5)blD=*V{=8W<`D#GSkyp>y4U||jNtIWY
z7Xe;o3?1VnltxbukPKV^SA)L8&_;p{2G`?af}cJLK<%gU>^<(^c7txN@mD8pqb4|w
z7h~=}@UmglxAaCr?%hpqsT}~o3;NlGj6W!3t=6)4Kp+rd2{Y|Z21~C`9JN8Ih(<35
z_5s(AhtyGsqAhZFC1!Ze546>K_xRE#eUeDJBaL7X3IfY)P2gd|FXt9J#CjujhER*O
z)v@e1uV`^Gu0+zfw81|uK+jodF6@LRDBZyj6SG!%Bm)PePQ__qhC})lbwlLVFW#Y=
zjK~kzGW8oN!UEB13UliptlsSNb~v(bDN${>S6wk%k6AmcIufO{(cGR#C&*ca2e5m%
zMkG>LhrM>o75*S+F^g?9ih5~qJZy*=tgPJ_A|3V)D@nn=S1M9APpE&|+JPhT>~oD9
z2X~}np>HxL*ltI-xc`v%jQp9=!6lxcDR(KQ7vt9`sh+1j&n?`ZAk#=o5DMxR%`@%E
zD>`(f>h6NNT<8+<v3hyUsw?XqH~X`wI5cm*!z{j}QOX?tEJ|ktt~rNfhxs)BTIb;!
z${AiPOBml4449aiwy+Jgs(lb6^%`}3y5fQK5|T$MWrqhQyFk+lDPdS=onRmbWD`BK
z2HhYo_8$3d45wNmeN=T87ov+5Doy-kE|=f+%Zh>LS7DOLh7E~x3AGhdRpWMth%Tu7
z#Lkk(xLv~OnE_5Wn9M?0!sa^SO0L>9@>LHm?TX_D-DsWC2A}^Jza0^x*7>95H7R$n
zzz)e)e{tSslueyb669vsAD?2#`ml(F8yGFx7a+*n3{k1XcI>|NUSF#y3y3(0+HiqO
z<H?mcb^zcCnkWex)dN`N5o|qmHFiF~?V0;H*o@&JDcvtD0W)x*W1-W;D<}!zEy@+M
zN87ZuS~-+`E46P8sS?CLZ&lPm1e0lp$JOxm*!iWSr=JYjZ|P23@w}?g(Yf~llr>#9
znX*^r6Jxs7Q}j0bc3Z=AgAzCSXfJZK7W@K4dW#p_1gaW+?WPPT#1GUpSk^ARK#!pX
z0~^tYnJm|v?;rb~=F}+D!*z(U^OGP&ry*Q?05V`_!5&4OJ+H!8vJ_`Y*JW3Qy85QK
z*=UP?H7cgADY8Cl-u9=XN;@4+nbwX&uD__Hk#&QID`x9mvOGl2%-B1bJ;`XX<sz+q
zAM<M-$hZ;@2@1Y~uv-0y^oPGmhvf&2@4Ry&DSI5sn_Rxa&+th7P?P^y(@-<ywGpNB
zgv>XQskZ>&5eML}RHPErc2c-COideFxH~8+-gYwXSm2s<mVwG8vf;qNtQ^AAv&2Yq
zrOv*`UP!UWUM>$GVYhCz__HL*R?Sd}IlqL%&c5FuUMYl0OEX;Z)>6ef*=XzP`@P4E
zvku%%sD&NI3OJa1%=rbgA74^^&%=Oc#n_*!XyNWu2zNn)$ebhM_4c_Afy`Bjva-J`
zMdNEF+oVq70uZ0CGdY_(HQEr92=ch4%`{f5V+F*dUL|L?qtJd?dW`=vjZ<c4iv9gk
z{*ON%tX%t+;;bJ!Ju_n}W>hz`)`*4F80sYoe-stRo@`7Sk56XSdPhBu?ZEj6o5Doj
z`J>u{#LC25rx;>8*b>X3a(l8))05X%q_xOw9W5bJv&`U0*iMTPmdyF&q^1%pnsrKj
z(M&g}^=VmI1aa{SSV9+%p0`Ga)O+JkwX23%6~%Ge*_kk{w&z7Zw5j%<zS~Nh!;8JS
zL}1o=*1?5rb;iOu@s2q)1@`FZb!C=fw^WwqBUAIn#Q9meiPb~X56^xSnu?U#A9l$x
zFgsq+60z(Q{dVn^z-;Mq;9>0EjQh~<+d1_|J9^-TrL<A4E;}$-W^puUx*@Rz%&j23
z(C5#_dxdnz16pkFY#-ymYa6(J3%01CfUQc(v^8Eb3$XruSNCH~%R=I3Q4&o$F{PD!
zb3lQYCp&Axp^p(&&_aujbbDFjdCA(N>gO0m^hTQ})R&~|>3Mp8HSHpp7FNapAzub!
zeaNHV%|Pmo9*(wGb)A+CUiR*YZ+B8_D%4ymois}fIDqg1IrP`aY_qVa=H!uSED~Qf
z*x!@8>cJ5)lX5>n+7$xh_Z)S{zU=&JUj0_qa7511;(geWQxVw!g!bA6gh(r&%^Kna
z1yQM^y~(3PNx)F<&6Wq(Ee~p&xiIKToEs@pFkl#IYs`0a<q(X#ZD`;+K^txMpoyMZ
zZbdq14l%*f^6;qnOt?Mm)^G)q3It9mQZ00dRtx=MrXVt7m2!^vZs|hAoEK7rs63~L
zP#saaVs1usKUFJ}B?}F=e<LtUd4Ya&K(y?umfpVMEZFS5_DS!X0k6L@<AL$(GvuTZ
zetQ3<orjV@b;?%4%R8kp>TyqZY7p)wW?Mq+3~jc;?b_l*M-1Q4Xzl3A%AS({5cxig
z&JgSB&DYUR=4u$JOo3L}C&-*s*2m614pe$~C1{%O4cGPywJSAq!viEk{PQ#!D78?y
z6b->I933>O@4HU9xF?66UiR)w+1_8thVNvPL~91f*MmjiLW-$JBr$62(-&*qF|-J@
z?#YXDA6PcW&gK@ncB%l3ksV;44Yiv-m&GcH1fFUU!|Z`Un)WajC*`kNHMB5H5%t3B
zSQHo@U>Y6}ujnYLzz_pL>1^rVxvDy?K~m%&6k<@M=ze2l_S)XTlJ4?QGRafPtjk;c
zfVGkk_S1rsz?H^3LW1IY#KDog6-x15g`gJKQpf4~*52!4{W{L!E{65m1o@^JI;>O!
zf2)#6|30^o!Ruu2$lQDtsNw?aB`S?g4wV|;X8+%m4e(C{H3U(OOhGch>xL*?L0P~G
zc3RQ3c@5^wHbdK3M^&532DV9`zS;+-SFa>dr8i%Bk(#~*SI08~vSXG;absAWF)<#J
zLG0RGv*h)M--i^Mtg|XxYA|HQ${;@+4hm%W2*zCh`Sszqr0Y-JmiE9t6XJ?s-LidM
zRh`b4x>kU#Q^}fD+L@8IZe9hLqw*-C5<@pT7NPv2EL48d&?G#>%VsI=SylmVWFUTF
z#mQ1W_=}OAg}$Yg&{?`v-n*sbd2t4!4oK@XL}$gewi+$7z7@_kX)F2dgY0=VZkP2D
z(+@cS<BEE8J_e!GDm<OLM9cyhX?s|l>eZLG3$xa-4-XS<^Vbp+P9jshE}2ctm_93O
zQPBAw^VCg1qU?(br5`Qd4<%v5Z&5<j*{2faeU;GwM3W-IR-A*}qkL{_A4p+}S<o9U
zv620aBk*b2rw1*e=D(abNec12H)FU9FFjgHoc8&T8|bIG|L^FlQ5?Akv#b5iCw_gH
z;!oyS=Q{uDaA72hf-U?E%b!y9&g<e{6nm4{73wA*_vP`OrY3Gk{@=pS{FENRpAJ`u
zM<|I7A#fu268{P>X}^W7H2w;E`0=(FG_lL*)a=r|sw7IPZ>3vC?42p~+)2!zZBT-)
z-u{65OGit|37xr*(JMqV6u;6~_@QMWVbz~#)Q+I9cg}~Uv?liiU>ksvEX!K|8rJCW
zV{Q=2^t#nvPJQT8fvM-55K~4KMy>M?ie(S8&eAKPzK(OJVRXWrU7|XKg+?~#x`)jv
z7k4{reaCFpY1i5y2L+>=J?mRF$;|^E6>L*c15&`s@9T(xo4bLYZ$|G63LqEQ9)Wug
z6DRwN186<4Kqp8amjkx>>2>(o4si-1Z%e6tZ9rhnp}lx<u91iF$`@KvMlprF#8lvE
zm-$Q>VY21Qs?1SxP{`|%?0QU1S%tZfTAij1EUGu;ox-?T`e`rq49Vt4!^kXhZV-`#
z28jzD<p5kgZ-Bx`=ih|bXGf5z5&?7R0#ih5$@w+3X=<zKRQE~1&zd;JqH6k^ks{*1
zhu%H&>=QVnX^(-s;iY`84k@Tf4~Qk^Pn|_}H$fE2`;;{9i(CqPIp-fg89h_WOG6xe
zN;nLyt@`AX91x~4z-C}&6|KY0eVtm3d&HN2Fyhh4ESZf4u$^C(04j|afpZ7KPsNx3
zoBDohpd!Z%q?$HbW_&{GAe!g!;4^CQ9~9>V|26tFplJ?sLXRPz*8&1<N4s5tyg($r
zS^+3$KL$?Yr{=$;&HjCh?r~o|x4-hI?MD&Kc<6vKNf%Hg69A%e;T@_J-g19xl*~UL
zCHVfpBi?@=)j!+nFHh<}@2mgApTRL^%Wk;q?~EATgNTESFDw8+S*Q%;EWnlR@M&na
z_?6=X$&&b`<CHy4`ggla(@0&h=lf%`u4|zG<_~=#112o6j9w;L*M}jyN895$W>xOT
zHSPTJm|>;+-<B0@d!jtzeEQQa)DlL6>u)Dul{g7SA6?lJ*#K3sw)Y}vJ$UA&D0WuN
zz}<K0QOvF9eMevBlAN$)Ovy(2XwufOQ4cRYng8+JVn0bTeHBgSfgrgW{R?4G^Hedl
zP+si>ys8zx=JC04mh`R*%(Ve#7oMAttQQb~w+;vVit~4E+?kMZWn%kBjoRUND>djy
z?5JLO0Qi??bXIhNRNs!?hwf>=)T^G!S_K&Lv+R?SZjmid`!nFeM%!YgjJnx)E8MWj
zu4i5r_?cLbyw#7gVaOM0z0{KIu7zQ+R7ZigrnhtXU4*HXMLdOUv>VtC(?^RcQW5$i
z^~(7UCqAh-FU;l2DWA(D``2$IUobpnk2$#1X=Nm6_5AhI)?9v;!Hr6?#Svh91xU&#
z8a{Ncfeg3yYPI1h;6dBLIHPVrdLa!CuSGO1czMr@Us@YI;wbjTo<s&%d$LFM4&R6b
z*@jOo^+TU`3WBBSgWpEo+*q?~DY6r=w!bGUQQWWlfm(-V-l4`+yD@Us4^groKJ@aL
zmhBk=9GRK--@w`bg1!F_V(prG|AwL)gD=C~vH^1r_uP4N>+I7Dq15*$bB=SOzY4r>
zui?z{uu?S(FF&VQ#~@tC&2Vjnhn3I~B@j1D9Bul@>rawC)D%lQfa@L>rrb!s$L7pw
zHy=e;=c2AqP*i;(afM@)mz~TzPDH#IwIu+Iy8)>U%nyt8mkx&;ju4lQZGkV1!?DZY
z=wRi03(1JlRngw|(8Ytku$c?^zJUE5<_#Wc{B}I9N6U7sGHlC`_T4u6j@53CE$#6O
zV@IJIEC2-|0%6Y^q|9^>MZI=18YUH{Ki{mA^5G=D?8$(qE?=)n7JJCyXCw2Ub?G}g
zy@wm<z8@voC(vC*iO}W$E{gn%*hFA5AbMok#n>earZp5oINBPc9U*xR;+E&@zegU$
z-%Rr5d`R2tqVHWl0hZZ1j-UMvUIYLO{`!pi+$Ic9(Y7*9sFbj23ifrwFwBT%pY#+7
zU>SKRU9NsH`}HMJpVsE)f*M$K_Z}rrkSP&|Uuk*2TJAtr#y7^ZqAyL^f3Vn)=|_>Y
zBg!6`bR_3~m@9rfrD(r{KJZqpD%!FuhPGi?6mHI4UyN-f3lAvBaXA<DPE=(w05O$^
z!m<4HcMx$rp9$Dh6iFSA420id3YA7&wW~D>%rTx3wK;te$SG@be<?oc>rIB}3jbcU
zJ7BGS#;S*Rh2ut)L}8pN{UVjq>g_(fKL*6pPYrUi-3lIaSGLMMURLLSVEFv^^{I=z
z4^Kju;Y_DGa0z0-JcEns0LeN-E8pHUX3g>3h^W?7XBWF+fZ-E!I?W}2U#-kyiM!}T
zaWK*P<h!iho?_I~feLN;mfClwKCy2gImS<oJ0P}~T9{eic0b9Zu>Km`ArL!L2rF!^
zHHxU<Ol32zJatO?@S=<6D`9(VePiz?)th)-no1Kic7LPc<=%5YGS=}~4}Z*ZuUvf2
zC#&hD@^rZulOU={_iwHx68Mg`K33SNMPp)%62p*KJ#X%YzFnee!uJxzs&tEs^(jej
z%RxqJs_#!#`Z(`BaStyx+(|qkS8@)xxRQ~pA4jfrIsjP(Oy7ykL(Lx(!|RoEi8*_(
ztNOPqHTW7+Ja5W8>g7gXcK+1+00k;xvon-A=|1AXrb4O1dh<oldIe)Xn_8#4I(%o7
z!lbcdW1b)>XaQr@Iz-QmguJ3~hbaVAFDDdI0VAonZgFvFfm_9v0Q=C+Qm2_#wa&$J
z%|D!ir1e`cto01@r#%)!PYB4MgE|x!Oa9|rhc9M|U#R?7?#kVtD`ZOHbv=1WS(s6?
zLq$>p%tld4UchP^hEyXMgfkyAPunGbHi}lib#APO^B#jABhv%ES`L)q93X7Hap^Z8
z5HWfNy!{|1Zej@XGT`x9hdo}hBB@W(W@_uSL$2MMU^=WfqP&4#N_RycO#Mx-GY4&?
zChEQd%I0A&%pv;qkz3FhAk~32(<N76{X>jscx;OG?Vl2F#iQd*-v7KZtEb8-X7Iyl
z`{=IyiDF>PsH3Gd)Mhj`B_#nJVk!Cj5w-fy<F5gF)q4Rv@lJn@%KisoJap}0Wn@G4
z`1(idrN+#hndU?pbV5wgk3N<>t7vXb`l77MKj-{ti<m-ZceUV_f1ywIon78P*X{GS
z#pB45^1n&z|3TaPq4aM8AA(RD6E!iB%%&SkQo(H@o(=g-<iAm@%FWX9BxfwIjf})B
zbJ=c*(B3#3=<!DXRc#LzBMy<(MZL(FZRLH-7c|K*m1d6R#Ca41JbHjqHmDyKDcaPN
zn!dl88Tx={sp;m**Yf){PIDBQp^J?VNlb|F+9SW7jxnqL{hH#)OsH!&kH%8a+vC@F
zIQ?HQ#<!oNk=_{Tgw=I)G@oGOvnK60N`j@ipWRsrWw?(@eEY<bv3sb8`5^w@IU&^Z
zjBEflD0FxQ0`We8F&)YPet!tu)%r$kMUR%r5L^?QH$AN>)$P*+y&id;-zG{h&UQ)u
z)p?2s*M`@W%X5P(fbRCAK~GSi+uD{}*Oqqu)|10dyBM2kPSA97%KUrh>UXbO!ozJE
z(vCV})_B<JM%{E798SzDQ&W4kx+N7eyO^8ns*qytFMB17?D?1ELkVfgSez3oBlX-x
zKALm2bKxU(4d4SbFlL8L^qmX><j}-L_#V?wFt183k<O@=Aeb1eV2*QqxxRIC#L33X
zqw3p*dz>*5!V(uat*rWl_&-_9&K(eL47?)d?^lgXsM%<93$hY&d|ol!oLXc=1@!u?
z7M=0V+eD#&jsWipw(~a`WR??PX=jxiiRP-%j;(v0aBu27UCCPp2bgW^VWV()ZB0d_
z-9lAE%@1b@-{)VR4ai*=S`rE3CWL<)48V^&Viqwa%3q~tbC4-#AaZmG0FV4xPWJbo
z8!&T-pb8YI*~0~FpWqRL+DMO?r;7GF21Lb?NkV%iD!|hR#Qg4`8&bwIDwA*dDrsjl
zmi_6=DB=?mV3zZUTM{7WE&OE#OSw29ydnM%U;pWU|J>L=+w5<5%>Qe*#`#U;2Iqz<
z^M4Qz<vR%f#X00<|2ebSFf#Wlk6d!|f?>tinYAVI?>yt_9%~OLu5)}@_=g$=b7D&j
zffF?}l-z<>&o#}$8SpmQxUc(nDkh^L+((BQ=BoMA>OSK4jJ9|kO;kHlD&oYKoo>6Y
zK(9qPpWcJZ#f(1lW8i-=q+FKMkk;uaWIiln`)Nne#3{6k3J`kMyHWX#aeeaDD0!6;
zwXuIjc|!VwqUUULS4-C~+L)w)->mwm69x2hkMZR1*-e&QQ39j-7Fv|g1w%3?%imEY
zO2}&rmcG8H<pcxGDf=n#3T}+8Nltqg^i*J@jE0>mZ!eTv?tz{rO8olK{9f|vn-^Pp
zQE`@-jL3uo5(rVqG<0&tbg%R**@wq|hI>Rxqr$7BZ20BGovdmsG^a*mYqlF5$r<#u
z>-^J&4+1&`L4uD2MzbfItMKZ~Y<D<5Q>%-OQF4_E#BS?oYpTkZnuz*^{@2Qfzw~|o
zQ|gBj>9Y*dl_J;%t|S(B3InoYg$;%(C$-oGuh4ar*N3kBnwh%u<>?E_*XLUV=#%#g
z|7r?R@JzZE;yPrJKkdW6UAK0$+DA`FY=a;GW$r~1B*PA1hM*ct-Ef7w*pKLdk|KCM
zL*4D5?gYU|d%7Plekdd<gqQS~iMQ@hZdmXRH(qi$Oym*4uY4vcXH1Qm$@=YUNDEKP
z>>4|^zO2kjNECaK<<e^LK<oP#tAe`Zld4cHO*zqN45Rh27QWhIO$ZqGq5$BM|AbeN
z&RiF_?LSw?%QOQ!(#5A0Iv1@Q&D;ha-22da6^1D^9pBT0=<%y%q3)iGHZgtI4fL1l
zZ1PC518%eDNOHt$-bFwFS}72WSYrAx;Kv}x@?qzl(U*}l2{VI|yRjlFu6LFw=E83=
zcn7JPx*L*<YP~1CC!nyHWs=X4x6ct`m{{I|VSrAyChr6&I2OZo!(KEHT9YSC?Z_Mz
zNc9GNIG{0CdCB4OHQ@pF+8e<^jWiWL-$rAlty2w{kr5EmHRmPDvGQtY+mc}YdDp>I
zvED`3nj=J3JLGM`J!&<r%-em^8x^lW(|%+l0BZG}m_2!RCSq3Ug6Pf$wD1;yzxwgi
z@k%@L{MP=RdfI4!*d#yM#|?r!+%74+U6b}Up*%zDcBoY7?S;upcQXVa0Sh9UCAZR*
zxZoPZS_Saj7NWIqel6Ym){Z8gbdn{D6}IN0)Q!ePU>RaB_KMLvBP$Z=U2cLi^YWMK
zt!XMsj-;>Us^Sab*p}rHz}0JRa`X?1^Q5#@;!YDG$`5l*iCtwun^GTp;LU;y=Xf^}
z7?!PDq;2^iRWDB6DdxEW5DavM_SokwB|Q#AdPqP7iUNj<2?V$`;9vIGWnMV@2R<Q+
zx-id<n!M9jW+F0!T@7F*TZt~Vn%+uuGN6k+1e5q`jYxt)tr!5a4jC^a_nsO76kx*6
zvS*6S3Uo%US8^v5i;~OQ5+Ybwr0oNRnLBXgf}&MI?@X<#`cnB&$sD`|e}r=^PF#rk
zJX-6G45{?yqalbDoCJkRRaR8{EiW5?Rru0DbqBh9KR%k?x$63oitWWh7+4=Gs-y_}
zs3@uTgTQvU=_A}P?o?zG+3tCxqBLfFa0eQL7!(bafX-07E0!zC-JtB?p&nG?hxs@%
znSp-0{{FT1y?j_qB)rh?lKf*nNwuF}erX!g(9`C@2kHVNu@<H|$O30DURK*=5@`<I
zXc*00Mj%U&<Zhdl*0hVmN;lS7OKcGfg-4^T7j-N+B|;ctcy;&fGCs&lMOip5$f1OY
zsgNSL1b!zHE=S;lUVaQIY&!Et*nVHfmQHzkNLq9WfZxP<f72jL@gJ{TE&yimMqs@O
z)5p0JEf5I3JGMWg<LbTiR}mCH4JdsnC@Ailj#Qtj-bP(3s*f*1x>?&!eMa)X^nmUm
zu&Gf|2ioE<Dv4SIBb<Bw+_cwA!cnNSxdlk^2{)r>ku-H}>O?zXXHXJK>nC|WS}H^U
zVSUah_(*brkLcfG!F`rde@}sb9&u$O?f%f<f@-Y#z)DS31C3@N)@+LJYeqEBjN$!X
zotu|vN5)5V$6Q~v?Ve$eg+W8s0RIhS*V7K)s6)tXlg;Zqlpt6ic48tZ=rBtAe5a#u
zYvYX}$0Aqnmv-GeE(~`r(=Oehzep!G=VRFSM3`;<DH30au>}HR=fKpk@)o0H>EZ;j
z{g8SYEK7E{?7fC|q{FDQcN&hAXa3}S0^0;7bFr+qx$3jp$*vtEbeECavH9$SiTmMR
zK2;cUOwoxq7T@Rl@mj!2go5V$SmjQz#f{<OB2lT0FbpTrUGa(uzh|0M<jprWgA(k#
zVg_*+DHx7?j~?DGfMzh8gAayw0bltq&Ho#^IMY|eTpFf4F#=(O^gTSGiB!;Z&RaJs
z=Ul`SqXjGaRd%rijdLR1t@U4_!k$XTr@LL-nE(GR)8KUry2o6VJAgA_HDERo0oGIh
zSEGERu0|;Ux9kFXqmt98v+Kj>!hm}bsE#tR!FBPTHBkoI4`+3T>nBkXhs%9k?hy0P
zp+~DN?~R%hnkHZm7O-%|+>S!XiFJ5ef^bUSclRfh7i=k_(}>cWv2ZrLcVj*zrwe7S
znm4U3fIWn|(I#7k=S4?fvi3`S{^g!*@yF1Ixo7X%Q=@_!@4;kt<zbr@VWPS((LsO_
zae;jOQCX6e7D7xWzN=3D@}$arA$!W~4`~naqan)#v5J_?ti&%9VnH5E0Z!8Lg|Ba@
z@^NMXv}<(}|I$5oKb@`C_sXh3K*%GQ7a1X~e0mCq3^_B&S80MUD|`BYEd=lc5o+xx
zW=9SGpm+hD)&iJ|oQP=r=FGW~18=YPlZcrAkG=PfYHHorMp0B$nsg}wN>i#zlNwwI
zh=_<v4G<OS2BZfmvCyl4fKru?lz@~-OOVhJ=|pNs5TpeXN(d0*oonyA_BqRS?m7G1
zF~0AfZ`}O{Fvc7&bH4MPZ~2wy_dFTw99GK|?4JC4PcCsq^(6|qfs1y#J0gyp>1FWc
zrI5N6#5II<SR6~^5|;@mb5r%glsj8I{o-dLF$f!76&34~FR1$xVs`QgJfMFOOFPfb
z4)0$8s?6Yk$3Eu+Z+%AEq0wF8I}{LI73dNp3AwAVyc0AbF-c25FzFXL6r`TfW=iGl
zZ4VJn{mF2v=4IOniMsbiU8&@f=9ty4fiB3y)7VQLU%s8Zcx%7~B}j!=EC~ydzD%&C
zc^AO)Zj&6}d%)r1^2y3;y&l!fQG!{^lN1lN+oxF=^g<&4+yR4M`!`cku|o?$?<%y2
z$(v`@0Y#!Y$AY*9w`Dx|BsyPD1m~-RuLihz;5Fg|_@(5-ZYG&L0f^&L{D1+ny6)e*
zQNMEoh*MK<3j=I@IejK{dM6;k0$7Y;hw$sWw2_|->#J~@`-eLV^rkhZ(;Gk<UaaLj
z0{&}*gkSEkd_(ha^gnUqXaAF%u|E^pkeq>I_kon|r<Gxueoa+;ug9@()G)3#i>La@
zMK*6G(jAe<BjpH*-~C}-42QkE7?`(}`<HOSblw$#u-dSQ&K%Ymob8Ru&rlD9%Vi_k
zSe<)%%&8BrGq4#d$=RmjKQ?N`M|aRaqW~h5gx_bpADZU)u_kaugahK5YHg7qa<^${
zVPE;6Idkkzx0Af$d`~IIr8>2VZ%T)-rk1#Gs<^F7%kZ5wz*2}EJa`%WlR;M)n7tuH
zIJJ-A@!-E&{A)b^zn&kq$p`kq&bfcfS3Z+M0{EejS~<PnWeQ!b8U6_2{&)XI&mCCq
zjZQNi=ye|W0(=AJC4bRY7&w;o2QjKAgc<ZbjS8Z-jKTNL%GCa~bcDg@j*s#!*Lo@4
zdkgPa|KRNV5IZAWn{JU<fMQv-3>}d&?SIGqOe@R|c!>lUdRnCClSzWGIDnN6;aO2f
zgqF>0T0LuJN_O5!!Z6>A(ImIzDD`t2i^rV)N_ckv$1w*99zUh}ti(}xtD+M?_EpPH
z<F|p6dq8;p>YohR6A^Tl3J`TZ=%6C#Mn-E6gKs~e$VmMhOqRK!s_vgy((6)W)BTwB
zyP>y0c}m%Xw|U7fad|F05vQdXQMCv8N>^lqY^5TV6i&hPcQA%f<*#o)fc4}?zgWoL
z86}QQScMi^@#NiOML#7<7x9?rV;>mIj0#5p0s6ii8f64Ta_|JIU4D1*9Q)~kt|HXS
zfGskYGQSr=R4t;qW9-t}wNH{QLOp7-B#r&|Rqq1sZDUo`s8lK+vLRJ$p3iu3Naoq{
zZ?E*A%eWem&>9FoECJ(xd73$k3Denz^>;h9uQ0A}l-mV9rU#Mh$$a1#`y-pnReHUj
z)Cl*@nd%e?)H|@lEugmi(4^RLKo<-OL?n+{h|s(esQQlq@MGQkpA0W$66k7nfn;Q7
z!+J!>i3DM4kc|tm;2^;UJ^bT^nnc<y3qu9b2b*D@?hb0YuevUNThZ-=b1Gk_1Fz=f
z((sxR!Z#2k5kQdh10TPi8epyav0H5OUYMw^zv$OPm)@z_L`xoI?bZOs#!>jKgoS2#
z16+0dS6qzIHsq<60i7cQv`Z|8&-MUjSiI>kqbKuK2^W8Aa*7&=*>evg^q+C%bm0|9
zeP4W)HQLb0!|R5wrw>b|jN+U5hjmpb&AMG~D3dKp-T4AuEl3BdtWae@fKc^rhTIEr
znn2g+=b#12^SecU-i=Ng-Wy}_tVYS@MC#<GpM*wdMqy5L15H_j{g<IJpoyu#Hz#ic
zM8GM)aV}J$U_+f02@iVUlJ7dhMpntmvvW0d;iX?lmAPA_HIk{{Xg)jjp(a#c-1#uC
z*urO)04)(h@WB_=rLvXu^6H$(kFbyqW9f;3oC}txiSwKGV%?&btBW_EgqSlWW<r!+
zTR=9WkqA^)V^GEXS|QG4BN-^qgY+*1GX2t%yg!efMYc^)P<<u!Z7x-MACHJBtJseE
zVPt%DVdznv^MlT}KD}o)Z@zlgy{UA6;hJsZ8#x7Ft~!^CwOKF3xMvGvS7{)GjiE~}
z?8AbLXW_E9v_@F0Y@a%lWx@FkjKCF>l=!Wwu@cbA9_{l|;FcEipP&AhFPBEuB|n{<
zlaHPlOT$ll-xxf@QLfBtB{yf(>d(ZHdvk*|w!qNWOnja>-mUqa`Zs`*<4+6a^)IV>
zH22wM^q*I@$S<?H*7+R^;=lNO<<{pJW&Vq<n@2PJ|DrcGxVApFiGE<NEW{gi_zohT
zNHnMGEieg{sxgsy*jv!?*{2S2Q9?^I$|UnV3Q>Ax#|v|w+0)?)Lvz#OiE-&Qetyvx
zWd(^T@xSVtEr1+xyOMRF8{4iMIpavWVR^z(v-Mvz$<!==GK}cb34jptu75|)hMF60
z8z8;@$&e?2%KZb>nHJ;%1kqBP0dsjrWbaZj0|`Qlw`M<p12~Aqc7z2vf|mIMqp;5h
zY%vFdh1evjUOt4^xsco9wOy;JZFGspr6(sCW}Q&%L(TmiKuy|Ke^iVk>LN8@cAY2Y
z^1?T_=N_7I-=?ekYN7QLe^bH@%YWSnAsdI}**VFO+1u>S0c@VjA$*nOGBo=ZQ4r&~
z((CWx<|c#d+Ps;tAQ+i{`1rN+Hu@ahSs}is`^9o;x5@DKm+#B6rQzWkLCp4q@${jg
z36t7>wO2O;iKe|V&O!OIe2;R_HVxz)lM!vsMfJM4yd^-%rHLaTf3dvrX62`7)fa7h
zpZRMi`F}F_oJ13Fhpnl*k1g8hM@Tu5?V%{&ccX7?gR^gG!)U#Ct3C<E3QG$GW4Y;Y
zDla;Ecpv53>mTSxQ}fsvUpXMwVB4BB$n2dWe42&@Z#Chn7%(q2pa7P;@U^40<2U)D
zz(m=wwcJ-_C&g|em;^(7G?0NsK>iR1)VmtY;<#K>$&;?usAOoMC-)7rBT4#Y?N`5F
zkP`L&ER<c_L47%EhKxmBu%Tkw=-i~Ds62Js>>QV(<t}->{8z@F3ORz^?%(VwaZTsN
zobG0_-{iA9%gAUn44_*pU+2&gz-m&#00;cX1A%>%q4Iiw6@2LunvjivU|pj-_AMge
zi+qK^31LDv$RpaF1t6JrZ-ImB0%%wGons$Q90hR-fQ2gD1h(kk?`AO31e|j^&5FQb
zcL%<q^*6)amCSx{c&ner53t5`Z3UOMfwq`|w48$!FtrAb{U3FcCpo-;YIP#b5g0w_
z4n0E*ap@pp3mxgUuR%LspVltS?{V@c!^6qSVLpsMbMX9Q4GSW~<H9E)ho)E|k=GWr
z6S^hKNwDmm6}=&ka`%d$(o(n54(BZEF`h@FExIK(PNKeIr@t=(YPNu+f=`UGi~(sT
zEPu(*GZI?%*nQ&k*{g%5yi?N87mr=g^QzaTp(9jx<Lx%0iSW)eT0&@0@V-Tm4uAoA
zPhJmi@+O$@I^#&Gg>NYt2zKW^kBEx*in|_e+$Lv7Vnn+(y}a~q9W^PS7IkFvkPV1u
zq@Hb;i|Jw4wIwab6Noa6V3)Ny*sc9QsfI5iu1~bDo~yj#exg0tp*W;}=EP`iNo~yf
z%N#a@uS?<nHRKNOMMI6=%yk#mY{K{t@rBxLg|I#CDyOmu2<-~VcDthX3G!(^-}xj7
zTJMfSmT?K^rxuLL_4b<&hb(5sXfD7WL+spH2?l%lX9O$259Hv)-nFFy)K#(xF?g>3
z80`$%p*gj=beXO=9OUMNZ(K<@wfH+5TEuKU%T=cr_V!O&I?NpGIUPDk#}wCcs&m>P
zk?Ni82tnB0c12r=vx<s2a*UJuVpCUrx+VeB?#Y&>D3yu}QDbrGm^FIhHtd5AZ?KEN
zYXG_vXf~Q)5!G?se?kd-JJHIxQ4rZ$_g?S{f7cHl)fY{S4y3Um0=8BAFexpp++D~}
zN&Ny?DW#>w(UN$NEB>;ZrMB(%@J|M)(fkLm=E<@CQcnqAj}?N)#?Yj%ZCZey{hwFC
z;4jPQj-um#vG|^;IG1_U!XoI;pE}W5fZ+A_r`n?gk-ei!In>L5)snc#_^(_1Yb^eQ
zdGVhnc=fpdLdbLIpSX6!LS(sftYQ(gEDu^i(0gic{OZ&`Q(DaZyEKMO^sidsw&j2J
zYR@y~{;pVO*8RJ99fRw?K`!Y^X>FCT5Eo4ew*tGBC~<uasOtjQ&E@VFh_bciTw<QR
z@Ie_z%VUS@eUHI^&>h^lt+-IGO9>{Y#jMeS5w0qz)1-<jg43XBO&l);s2S46_(=se
z#)Q{}<$v;#6g?uw%zeMn>H9Zo$-epu=u}fMA?K8}>zIH`{cw#~zR_jP4w(ZLdSIJk
zo5}*Of<P9cQE>0R6V5r$G1I4R=4!BbcGKi8k#4=SQa@J*Vr?=db9KldXxE2*SmkGi
zuQd-4Yl4~YWSGYLG&vi$RU6CI){Dq;<aV5zlQ$B2ruh*|(Ip4>V}LqoE7v_roG$@Q
ze3&gUV(M^dW|1hwI|S}*HsTka>FL*~WG)q&LyXnjO=vj4t=M_=p;413rt@=5HjvxX
z`P1WI(_v^WRAt4aG0dsDA?jT^AHPN1>(UFtp`0?G>72kWexBsi3OVC!CpK25jNo=l
z9dhw&h;_Uw@D1Ze0UvplGW|*Pvf(p>w@Poi+JhPOoX%g+Hw`qj`pICsgDhRU_lPhj
zL9xYd??49qw0+5tmK@#{i|k|3tr$!?bm6%};bc*~>^8X6762pmxI@Rpk(2&VQJa7}
zJ&9T0YaSK%8AE0vZEw%jlTbOHcwbT!c|`c4@g2g#9C4%7_saT0wVNk<YOF}#<i(Vi
zNAxbe_gdO2=Rs;vvWLspHyhFx(v}^QVztsIx`VqB-RNPkh?Hy2M?`dOg3#nLP`mqL
zn|U9^uS*ao*A;kr`d1h|MCSTd#@VrNDaDJ;7sorrg~GmNyItPnq_Jc5`?s1k=BeQ`
zaz7cYq-altXgS^o9$MP#`)0%~bI+Xs=rE*%{U^iYgt@5_I=LH#{*VN?hx1fFO~%jK
zN2%p*WU+Q@fW?KKX*m)0Vr~^OVUCh&;96=>PdZok;@i@DBi4Sg<jd=c-CehzKS3R#
z_v+$#i|~^tLIi^3U`h<+t8=A-x6_FuuCcEgF%%uK%oo{@a_4tWeLs&2+w+P(S8vu-
zz60GHMXOBup9sXFE#NC8yD00iD+w>`m{3$GDG#H4dSo7%{gdHl%_(DvJjH8YPN?1K
z$&U|wZrjX#_s<xr{~HeY2MNi)`^(slb^wi=v;y2rnAfBq5nRNH^0DsPtPt3XDOc}j
zIu9@VU2A+}a7tWMD%H;)uJa4OxUI1l?%yYt78&3CgC@MR2mTN5<S!w!dHbIXrtJGM
zw9<K;0K97+0mmM6(qmosqCwk$nfVFbgR&(T0S?H{0elL+*zu0>GkinM83gz(9)RD%
z+mJJ%XgN%zD+H+y`>x;w^zi%TidN}vprgHICl=CGt2mUR3Ubs^bh~?5^WY^s-Bf)C
zB-t@W=9qL^W!3Sp$-{b@=Z_?>9&?-VmC<6qM;;3!bnAa=hzDp<>^(SL(hl<)r9RJ>
zXHk=7ZLmMW1{NiUQIZ4Qc0nV>jlvT6W|MYn08jQJci%r+E=Uee=6E;wkj#yhDeedI
z@b**n@jv=Kw33Mz!~hMs*>hZjp2N=Q+IwkeZl#1HxTO$2<fv~{g<QE$GpOSt#I^3O
z!I9+W#nuMAqIaq;r$$?~R+)bFwu>Y5$}x9XM}&JQSlGmyurI?W>PnBjeZn)puQ>os
zzL;4tJ~ey)yTX+&?O;6Ug`F@dz6gKeAz3xvqbBG0lvYQ?yQ!tfI={g31?2df)g&vO
zct3_-u{s@%gueQUp{EKPH|PzJYX8X+8D-yQv|#taM8V?2px+Mi6g1699@g~Mp*5do
zySbL9Uy*Rnqx9|?!ALP@bJ3o6H95{!uP<ImsBR3ju*ZwUW?w)yRSpK3&8Ae2Ah!6o
z<Ak*|!0{BfQ6$HSvv0bEE`3(pB#7-Kh<q7OoA3UveTF7CRI9~K_`XMpA%IxWZ2=Yc
zGj6FQNPNwCv~&(?5wCX6A^<3~Y;n4c6-s^~lUGGm!$(1tk82v01EsG&8C(V{G(jxs
z^D(7X-)?=V);m>K8|u(1`RK{enV8MFXbZpQTGyWpD*3C(@lWOc)giim3(<=i`^}^9
ztbDqQ{DM>j`f;4Rv0Iz9Uo&5Lez1s6Gnu>dGfFzf#`$dRrKkt>5NR%!AAPnTR7)Sw
zJ&~z2FcmQeRha$y9sw)GVDR7l8p2-!KWP;QRUv(n1@2PO`j&)5amyXqNRB)iNnV5A
zLsf+|KGHfR#ugUS7NSDVhkx_W76KdF8(H253nsH?oHjB#uXfP-3>SFhd-f$QL@SiP
zNmV7L42C2yMcsF)Svs3+hV+6a_2A{t#&K#~OGt}lN=z*9#9V*!Y0&D*E|`?OwA5QT
zNy9r2)9Vt_s11I@NM^F?90N6O79Nfg`-CJ*KA&y&ZYLZ!2=_pIEiGOmw31WuV{RRW
z?HCCllVVqSxVx*6=iGv1$Yy0AF6G50Muj+K1e=RGOZAGnLQKW>A(--Ea&1ye1Ey?n
zH&L!j_D0K-w`xb?Td{cm(6S%{;Zi6lT5dL?O{$ILV)#IHT(sHE<=vZ(LY!AVDzIx=
z@R-RP=oG$2(_t2JocaSYqHKFZE&zOIKt4#h7z|$8xzrxur0B#M^26`Dq;c$N=w|*(
zW9&?NXa7%z>mSkZz+v!>2}`7la&QWItB!8Bo<R&LN~GpcT1Gu+s$?dE!MmS<TE#a@
z*<2&FD;H?0K)(IWYFwqBZ=;9<MB3qUscfLQ!TBDWL+PVrc4Ad7jEP7I5w9neXBXl3
zWek0LWfHgzj;<zkAHKEnz@1cT3pM0BvI~Bb!YDdMpaBAuo3<)wLccKVoiTo8o#mXX
zz@sPmycTvA>BV(p7vtu-Z>k;H@Xo&E8N#{r%7(uTP$^`;>Gr97ZVXht8p>LxbiOLT
z`tta(a1%Km=7~2a%<N`4azB0yfn=hHa)G`Rlt)@={bbV(teFgCyK3#?m*1rbZVIb$
zU;ely8vD|u%#_f3>e6>tI90mg_xvuJlnMlN`dLkYSg!?y=Kk&mWn?_J3XPYo0s`^^
zIrNhiXzH}cFRh+uD4DC$BvL_}n7e=rZw4gafV4GdugEUG1Sq}yumaSH{LMuNP+SfP
zG!l?UC7JpDJ+m76v6Bw_$MX+W>)FFwt$-a|QyE0#v4xWxK!1KVLqLBgyTCGt4PbwE
zAnV7k@RsI}+e}YfZ#k7z5Go<aRKdo^UJx`*%Dat4@zWp&Z^lG;smZnp1dDJFN3xLm
zLX~$QPkYaVfgh5iQ|4pq=WoRvVS=#j{!wJ#dsw3}IXkTE8)Rll9aCH0ho(TgaV)8?
zI+i_!?<MBw4(uFj>co4j;_;yH5P8xVS~ao>N~qVADKyHXdJWqQ_v+<O8b0e4NpTU>
zczpYLzT;Wg)@PIH;_no|F?CGeMT#`+=ksVS``N(*J*#)E=wOv#vH$X&e@|TgGUf?X
zBU$*5@FL)#C`Hg!bp-!6eu95>=RcdE{~DeDY;^t)&VH7LQpp4pPO1UkuKEM@M&G*f
z8y3+3m4pD<i#1)<{X!=;Z<4PKU9XB=JAaJrNnqn2LT4f!Nz_z!_CBf5a-Y&^Pw~i3
zXt_-<6e=AMvruXz<fZi39#{L~3Ttc$e@W^D98%smY1eMx{n@ua>Mrc=EL%O9vGpgz
zX$6#oi~f&W8_H@>rIojG&zqv;?ae;c#wsd#K`+%n&sEMHne`t-*MvtLO^C@S9}6R}
zAD^MR$Q9NPD;qf%G}K4WM_V^HOd4H$-FvRa^*M6gw%{J&lidGuuhwy|bTRwjad~+o
zYNlcNe1Q|-Ta=o}O&`eD;OuW76m9_#{cmnV)S}OJ(xf;_a(*|!>*BcVVlR&LyQH4K
z(Ms95(-xGw)GVO)<jnOu$IF|+l$WH`m^lRws`~=?z4tpRjGDeid?@0uf4<wtiurz5
z>3sd&SlPkM%?_;uj0bnnHF7j?2yrS+Ga4al1{HiXpd!DG6pq+uDF|?R$_^Vr-3@w=
zI9G*U=}qbsWtXizctwN!&d7)1wMI4P*}@R}mPJv%+*UXJ8#D}#)})fx?{pX-KURNT
zJ-pm`Ww~rVZ`i~Z?DEvgtD82KAzqdgxbr(hyg14gJH5}|$$n;e0<^kL5aC(mvRM>B
zQ|{PKbSvHN*2>c!QWZNPCf`_jb*lbMySLrt<WA6GC*aAE4rEHn+&atDGUe&5szPdX
z>u0HOjnYUW;ZtpD&kW8@Jn6kp$|JVWux}WtX6n*xun7UwD#8tO(sR4MZLI$V^VlQb
z`5o@lsWT~OQeewGsPOuX^`k*LR7CM&C2Xx1W5-YWno@lAgHgev_3OQwhOz<V+r!;T
zbz`l^=ThuneaH|0?VWp^E>nSrbGo%HWN44hH)+@4UI0lf@uNZM@N)|V`+ClM?)6Gy
zn27^yII&i5P@ZP%!Bamhhp;wff8Fv${<Q3u#F{8({j%#$f<lP1q#Z(MEKQdTB8)DS
zyrb6pfS7IK6mO9ZYSfKBKLN7Qb*~>JnP*Pbzp0cvr&FcAK7p^pQy~<Sn9nG6)T<yR
zDwvQZztX1Y<_SYu*D~7H7*?>hCBC6|5nROleB&WC24Nc;XsNzYSmP`_dfSAJW?4H_
z3kDmZd0_GVmNnPc*W`|jRe`^CD(7#up&p%k;HKn|#~%7xNZ7nk(3S!m$~1(G%dpYf
zw%>y15*!2RHzIDRJ~+tR$_NI5gFP_X=SXAac@UR6i*)&dGo@3-=q%B?EV~om!w61(
z-ByMWxz9GwSWV(Z_#H@JqvmX-bdnZ2+ZtHo(~4{7bOkZ<wlkX2e15BSqG@V23_Vw+
ziqe-rA4Aag1vI^2MSYUUvjuFLZqtwti1iwwbz=Qiec-b+w#wfgS=s2b^a@m(i#^!*
z96ap}7{zcAyC8#6-(5|`f*Dwb(coiQAEGveupk4fvYj#(P4a)CC{I3rzg7A3!>}jW
z_QN;&(X^v`xV@^JZypGf=8!av1ZLWOG8+=mhCsIX3p0<%)o4j}d|hC5ZZ7=N<9>PL
z;bNC5Erla2=85!v>x;sA_V^r5JPFqk$AsC#c9>A_X`lH-I_RT*+T6;3*G|iQxn8p1
z%DeG0)L-K+Q!E9cGYl{!8uQJ~)DvQ0Ta9QIw+6kqtX@NZS&O=4>)K)YuF#uFc?a*k
zZYHKoz=Fm+5`tvrdLgj_kT^||e5$w23%QfQ!h~U)u{EVJ^T5ag3;pP?vlrVQo%(IN
zc>mtM8=Q8`9Q}0$d@+Tn;juQk9!K0j6QU~Um#wJ|H74wfbs^`d)ur9}?H`N+a+wbk
z9XlWSCv^>h!8yt)lrqh_G3hrvT|wg1zQ8$!6G2KNwXAn-n?`+?d>bR7vSsC>3Af_`
zn8=eU*)9afiFx8TdkySPgkUw&e-;;|ZqjBy*UW58YX4eL;hfrFUZ1<HKHYQGNusG*
zQ_*@#E<>p|=Iu+dcHY&qEKSkxN$6RND4Ga+Yd<m0$Ozp?YFsfXF>w};a~Ie!@NvFf
z(mAu6k}<!?>??0wxQE+@%as8|f_hM8X!)MP*a|8~l@F6M;kR!^bqq2fAdak9JC=A>
z*es^or8(9m^_AX2T5uS!HWZ9JQx|9I3rN56*-%I|Rh4xgF#)+Y7^Fdf>b7YE?4{wg
zB5cZUN0ek1<>s3#-3vIt{rt#uK*KTvwkz&<OzMa`gRn~HEGT}fkMBA-KvUzOYIdC$
zLXE*}8U{RhFe=)ZwAqJjgoq28Zy@*?JDd)^)*jN<4QzuPX>uZS;@Y1sH2KJhwe9sR
zj;ZXq=L9Q{W|bGjksBl`1dh)bieCnZ2V`H%a|Fysp>RVVv12=E_nwI{9CjxQTO^kr
zyhow*pIk|^6CZ%q3idIZlBU}xqlSI&9;%N#Cd#)Vk*cl3!r-MNc14ZR`V*QHpx<n#
zPc#DH12~-{VKuH!#F0L<Z!)18{2n!pcIr*2CT;-Ei<}~Qz8wlPp>ry%P9{)+B;rRp
zj?$YWwJ-H$1Jw%zF?IqnceBm=sJs4alN}e<Kd_<^Z)^XkAw(!sr|0_YTPbutU~>)^
z9%JW$HJS<N=L7^FN}u_1+~HF;evSp@*meKa2Y12DcT)U_8wgEmA8FE*ypZ*NSzCAn
zx`p7QI^7mP#drsEq>dJD<cHz2?uB16<o<Y8T33D+qv4lQm^QIYPs8@FVe1Oy>t?>r
z)D4r@i@c+$Ss!3HiZ9h<PK56>!8VXk(3cgFaWu8osh~|xT&L}eaf8)@Z17eunz{Ou
zmKM6xULj`7<0R;L(Cut-bXCNiPgU(Ptm@+nNnElU<%0>w?WfFD{mx=ttETuAJM6fr
z$z*RlB)&<3a7LpQ#q(*@C#tN5GOQe!Wh&rrerUnJ7J2ENhW@C9l=ydhCi~BHW-^E-
z`-7Yl3Ej%l1}CsiE-6~Nqzz4%d3SnihL$B90l&(R>NZoGy7cTs0a;HA@GfUBBWH#-
z$0qpEVYO)5X~btpul_ba3DUW<hil1FZ3lQCgMTAKD=ORz;>Gn7?Nd!#i_N}{FxK9q
z&XXnA4Gz+2x^oEt3#qdO=v#O31jJbACgpu=rJogJXNvicCk}JXcL#toL`%?|%2hyR
z>a<pQPB2`}`?QhYRAsa%SD08Zsr1{U)4o>f&x(rs%&=(3EUDiPvXy}%ydcm3P3^sF
zLwp8L^%QtqGV3U@RXRb%o8|F1v`8*OOl#!|F4bnnoaO=0T|jTq*A=n=8lez0B1?pw
z=1lT`-o!zG^A&Xj0<hCP{WoJ()15`V`H<d+4zKmY?%qO%Cbe(Y#&Dim<)5u4ND&R;
zcz}%(#No>-(_+1y6GM}7I5X$mFkzK2b@StC_wM+0`<9z`4wp)$pS<{u6-vcmK%8`G
zDxcYaFOjh|TQFtqvAmq)aFBs_aap$U_q*4Hv_b{7uYVm<eUy17FTI~yLgKO|uho;S
z*ZOY{m!Bm=Udjb&Jsf_^pIaYo)TtmHMo^?gcHB$Gc((LyV?8!3Xm&ZEtL-w{dm%0%
zCy*Q4Xu}UdK554E{;CL|tI@R{-2L_kMgI75&80s7BCy9JE2+eaG2D+w_oI_F0Fxmr
zp-xK`;5(-o5<qOubiqMo9A|HZi{HeQ1@A>&ZWW5+b@J{o|9oswYWHU6ODR$5ODs&8
z)D?>NC_rVh0uccsYOHjAD$Wu$_+Y3O#7_M%|D*0BOd^Gl{&jn6_jI^b)+2r7P5g0>
zWY(n+M(SOEyhtSP9#H=D)I3EbOm^#1p>yJC(HF3ybc<8%HPV><0lrLf7dN&IF^yYG
zLERc$Z^Y&%R24aRn9so*Hi&io{>w_Hdgor$d;F;)eR~8eM+VakJ!0XjtN}k6a)qC%
zCHw;b(^T|tx|M(7+h;TXk0Z>8;=KPjB+c+Ur<GXhO8^)H#0T~+xnZ@X0NLDEK{rVr
zp_(Gai|T4)N(VRS3p+MLP9vTa$6B*%*DK1&IW<+{Rr`6-8=#{A7-Ak^vcn#r>9N+^
z{ZT-L)K)@_2$!Z{87xhAAjlBk$7eF~F5?zfj&ntEO1UKZTKu&~yVp2hw7pPeagt&t
z$`BFXiPUe!lO^3$_*`c+TYV7>elO-)GdiNv7%!)LZ}@$ZhTEI)w-W0aEKr%6%Q8pe
zQnGbUe(LV;omzy~fS%C|%D_u?0tMfbW@O7BcuFP&@>CSezxU_$XFndj7;64lXYEFb
znCRUgQ&Rk~WPCXy?6D7meG$cjh1X=5)P{lq9^Lr0Gp{-YpsXnTEp5|hoochYkgyO2
znm!~dn~}^Ht;}phAuYVu8l``r%a5$H6$<w(Oc@HlAxjiwxW7GJ7FemuKnv<~%Y%cG
zutPh3n*$PVwtkaeH~jYN&HBx;bPd>V*jAPFazqrg+`@doef!`t1*%P027SDGvRmTI
zt3vGg7uUOg>*~wOV;vMfc;Ce2L<l2|cGSzbUGR3Z2c;EEXTaR$#o~xVR-umPg9fD^
z&&+vBT$Md8e1?R>%XPMi9H+SjGh0@UD#=;*-XWo1tcQk96qNO8ZhY+e;C-sTe#2i9
zU)SfEX%YM(h>?7hk~|l9uW9&N`j5yOR=aOCA0M2pvb}N>v5z`^5O4eL(V6a~`YXlO
z#%j)4IWs>Ql==J%caEj(2C6AErjgL9-996$0A6ZBu|=4ld?a1_ed%;ToBzO)mklQ0
zuK!V5Qo}j0^3;=>!*_;6FQrY@cQZIPOQWR7b<LhdI6IoufHIubl^8I$4>ehp^L>(i
zHrBATJ@`ra{RjLnKr^68&DL<`(|31DiP3NNM_%s(@>`eS8St?Rhqb=WTKY;2Y9DTL
z?H9Jcf6xAR=y_q;x-X%iHtC9c+$@KolH)aleQ`mf^A56yI&>c7n2w=BND1u<vFSgE
z7YI3FNWij|o`eM&eH<+q$t4xc1{P#hKD?*?p5ql0n-RPbqPqy%;3kV4z>DZfpY0$(
zQet8zY~P%^PJp(LW$&>yzqefS8zXluEJ=hW&5yjhT~x(lQduDQAu(7PO=pyWe;)%*
z?h|k}Lysc*YXsHC7((fy8opl^zF5;p<{^0;zy1(g6rnTXh#a0$nf`byb>X7nEJLRs
zo1(22mdHe+blQz>1){mY;}92h*^qN6K2sx>+b(NQl9kGyUo;evC(exGt&yv~gy29t
zk|>Ce+Rl<3@`w`HO0P@t<WDtw0h<0;Hk(U)X#aV<iNU-7#&_lQZwUlYI}uD!qN|TI
z$>I0m958hZ#+(s@O;hb>83V|^4HYsxj!XA*blK_LXA<*_tXKXp)=iswgqxO;dAe6x
z6M!6ki%3(t1E;bp?kePrK!EI`v%=w^Ti9d2vyFo$)*q&4st6RO`4{$kI(9O9p0sJ_
zPdx|Q6k};tTWID*L8!kgqvW-sxs*UW)JEfh_Zhl_IFs6jTCyLyVp_A}aQA%7IIj%s
zM(pkb3AK*X_XYM|G)U&6L_mF$q^j6}+Jy~T3J~)R<2JQQ1-MK8Lf-NJ=Gxu=c`$i|
z*vpO>(hkLU(07wBd+aGsV!UFeQXIGQvwimmIj%!lYUY_PpKXy2%U|5g!cKaXn@(IZ
zM(dSoN&c?608ybSjF@yBtKkdLA}7H<u76E}`9KF<_MGo@fRvyUx(t_Ic5B>Xx5EO4
zUU@KC0i!yw+-CG)$wR>K$p^l#2CP+)Lhie#g`P6bG?$s1S}>pJYuo^5=$HEqx$YMN
zlUpkxyXD8orP>R_rpOw!OoRWtVa@86ixP1eX|C0SqiX>$O&_>)Qr<uC`b|G5k!17c
zS&=X#U3NfTK$Ua=P~{fuROX{yhjUY`DYBQ7P9Qh0%Y?I(z8q^3#$o&CNA^N920ad3
zHnpUq$vJ0e*GI*uc?VgA%Xi1AkgKGWACa}=kI%lans;IA6;J59!~L9f>gAOq9na>y
zq*rKeF}g!R8bvPFQqd@}6{LWyK+}K=`J2SDbpzhr{bthSWa;(K&#r&9N}))OTrpgN
zuLLL@)Y59d*+LJ3)2N#YsLAYVy@12?5fpO<CPW-7H+;9gYS2dz#PLwWVG~haaSe52
zvdvQMb&~>LD*kvfGfsI7W<ltPnlpC0HC$6rpM~%dgmZ|`-!?c?vub{7&Tn?T?zdw@
zAir_ZS%K;WPasePte`IRCyU{XBAV*&QJl@{+$QxEN;IkSd59|Y2PQ<F`W%mAaiP>%
z+jgWlRk*6u$p*FO^gp`H<8thd=7!GP={^f;@V=hqvNkK3Ju-g;5O7C+BKyun;k$xB
zlmMKm=}Q*gwX!O7&&NxNQ|IqF`X4gn%u)PEgtyB)ohB};<~FW-d~V#cau4j|!8iQq
z1UI{a*z^E;oe-|EbL6blunOmfHWi{bi&xvws+?dimrR~M;d;J%zB>!QUX}%=Xi%wM
z)9?x@Cwgs5fyLF!C7sfIH1lx3v2&06TjbY|MyD*;I*xxo|1ORB8bl>#Z%Ce=y*DD#
zu=<ms{89PJTJ?l;Vdr1;;0ym59M9j|1mQ>+!ZmGaY99eAV@{nK+nbMwQYd?j63*Yd
z%wPV{=G{gda`H5EYCp7o*E^HZR_a-h3;FOI<rUc@EXXPVFq!PHzMEWbaLmbTp*|OH
zZj8EF-MwifV{pWwvN+nZ*nF~iL`3~11ZoSP)x@O{)B*vP+-W~Jl}}+I(1F|Z=2e?t
zK+Vg4kss3$>iq92iv5{L$YKP*&Hr6^d}BQh^xX@KUZKzZWYAY;P;9-}=}qY4<<GNv
zm_*@*O<y&?d+x<?t*bmMIj3mGfLVQ*nxR8F$iWz)`Bv0Nn@2}@8_v{aX4{xAaENKx
znjDUIN?_nL(e$)4o<;8p)rbJqu)xY4jIoO&d$&yN#L(4))N-zmZLc#=T08FzO}Sa0
zDX&3+HB$x8`qd+|C2&<;g1REIv$F!=ZUsm;`U_y=4bs#`o6eHY_20Wl)l;%HF>F!~
zDad!NQ(5tLx^nH6%&xKc<CE88`Omx#Kdt440Nm<-Pk_^8K=5y6hM$BP+=wG?BZL+$
zLZlKb$PLEqlOD>EAz#SHk|vI~_6l6Db128h96j5j+odU4MWUQB!iQ8JAMlOiYB46v
zk7#PVH^HB%b#s;->%Ltu7Jay%Vo<+1egB?$+=Nu?`7RgSyxZPBk_L8bFezUxUk3xQ
zPRfdcEzf~P3O$DWD1?h-8#Yt)f$BPThOMLH$8FT;ogTjMp)2<knXc}SIoh@zA4HGV
zJ-HFyT;MEl>;qO-^Mael<1o#bl5#nUZJs^?buqt^q3VwKy~iajVk{xLK(26QMuDpL
z&c%bYVRSvY1$FTwa_IeV#kCV+y{=<dUd%R5DxoZsH2fb9&Ytf~dtT}=`&Q6sOhWkN
zC>jq|jrNJjGQtcl@|~((NKNNHMFevRbjzAfACGCNO1gHl@z%+Vc?;M5Ak<`ha}BoJ
z?w|wPkUld62-w}JZFVQ(j6?vaSt}oYeUAaa8!`d0V=kMu0;!i<XaHBmm4-<zFJ9sG
zPaE4Ph?V3qa(w=X&E#y<r<?aVFpuIDH|No<XzC3hGMSBHLq*{8aU7`0LRiR-aq98a
zE%d6(9Q^WVig9!GbikW%J<}*1nFwZ!CU*S6UA$d6{*2!bjKDV%GGZIR`Ib5rx_I04
zJ#g?0`Dmr-H&~KF1Rz|H3xbEQcL}O-32Lah{A5@y+6d}XNd6{(%-L3;8;}heCJkrz
z5s~T%FNZOyd3URZ<!Y@Y+)f82b8qx>O6N3xQf~at!+7+GiKv^aF#o`6ezhe{!tH5t
z-+Zq)Zyg?*X0MtdxshLSPyUpm19rZHpzTgAz&BgW@pQ&D89L#ceSMh8cw>RDGUeI}
zJ(MN6?U$bZ3ggqfu(@lQL94|fhHMubOTc)+MLI}Zcvw?_f;#q+AZwwr2SvTn&pMNd
z5=xZv)j?T?Rh&tw+5n_8Xl>D10fN9*e+-n5K4zA?oBmW{SAYE{gJgqtI4KM9qRA7V
zrPwAQnO11GB#_czYi$E}Dhqk^_;Pvez{Rdr9p!SM(g_lQUStG-?Lu&XT=*oK`P>kl
zn|j2WrZrL#WKaNmI`dgm^#k?fo|}upT(&+SSxKZ74a-~KO>6QJOAIQygFhjX%ezWE
z;&_dt>z=#L<cmx%McfMb)+OuqCI1~C{}uSgnv(*+27*TPgC7H&VSk<>2DGywK%w$G
zHULcj45+yvD_841I;Q&TRVDbJ{|&5k`+svIrYu)K6hA3GVJ^=9_5f5~R_U(9WIJo-
zP5M1Q&l^3@D=&vUJf_pV)At0&N|$lT;ZZun2pOOsfxV*ogRlRn8+AW7aOW&@Vw5eY
zhULA^_h(_1SMI<ndI=6;W2K}{0x>0}2Ah_qqWbEyhkMxflu6@QY3X>YC{YGAacl;?
z6Mk5(Ns6S3FMMCB5bJ~HvR*n>`!%f?mTVq>+~J$9*f*`tq<2mSR*{DmFWu!^H*QW-
znCpXgvq4y(jSDn2=x#OOfZ>@*I|mFt{WQh|fD0s#ehSSFA7BD%7Fw<VCFAL|O==)X
z5^|K@sh&do@tHZ{6s*^bd~fb<ZF5Z%vPlL0SVJ=Vw91q%Skw}?+Pi_C`TSY=hcJi=
znTs<rICxjne-g%Qk$`DHhHvG$+2e5$dub6q`;&{?f)5vFTD`^id9`*Qy!fqZf}uMb
z%9C1WvL2mAW5W(NBhqm59!8VE+?_|SxZ?u$08qf6rltrfj2cLPrtU;P)cHH3_5v;z
z^hi*{*n3Xa+N8B{w6cck;h+$4XVNIM{h;!!;KbCShz6ECamQQW{Z17CuGLRqD<TXR
z2kDb21qgtlmSKpf<cK1j*@=Es`H43%rYkRghQM3ZtT-R~(F?YVb(b-(bpxDJHec!|
zm+#p3ueDYJ0PI>FeB&ulDxZ^=^XCo?{?gwdR<hgdT8Sqf5#HIZ5~4&-tu2H}yWJIu
z(qvYRRO;<!=B(^LnlhDa<$8o0QX`Z2oL(*ZoB~-TG_=7F)3}7u_<GH=r3Qt1LK8x1
zUen5jxRl%7GEEWlC!ef-sh_xYwa6o;dlG9ULl^WxHOnJNyQOqylTOk$()O3G{4#W7
zeLNDUasPSm9}nvMs|hg8e*LQd8kfH>zyDxdxW5D9Z#e7oe{*@?ogn^X7|kg+G6hTm
zfXh1dFGY6%fy+O2?Ho|AY_47Fe-(&60>^_boMQIj7mR()p`t&K`E~~EFPYjD#)tI=
zSaO+~2FjZ1*`9oqHErOEpM32E@HYRVe4bLEo|JcG<Yo1(mC@5<1uS}Mro{+;$(+2w
z9C1L?kBq|rY61c*Pc5RMzQ)%lI*+8wee8C*oF9kodvn1-&~mQ#K~@_`$O1|KiazOq
z5`a1GK`oEh+M3P<$c?@XIRCuvf><0*gDc^-hDA;Hz^Gd3;-&HV$IuIRJWYSN^h`fa
zg2=I__QKrQ(=5V#)r=CAx%X`A=uc5I;pKlQs|yztf!zdrN5j6$csluVT~(65ToHTn
zW6r}9JQ9DjkqG*do1ba-W&g;O?cnB4tI7C&yD$sYXs(>O?dfQEEg{Zp!dM<_8rvlX
zJ`D}&y`RL{h&*!4@Z#}J$z=Z16@l%LkD0@;qTk48NvarJRGJJv;|$~NA=={tF=P_f
z(<sL1rHhx2hl(ETi&5ia=qIRjNN#S)G*WWgkH!TFT<{Oxp0Wm~$SKpT;hAPAV!{zJ
z(<%}E0alF<TgVWJs9|L`sEx+9J=M=}Cw;yQ?&X=z^oz`M&G$Fu3h(R|vrh&?s{211
z8{kinO-B`E5l5$&Cta8y<ulu!W5C%XgNk>PzocF1+6_PP9a}kgi((FTr9-EGGT1#f
zKLD|>kZd_HRHv4mJ%x6-5G*?YI)kyP8-7q1uVOE)-0)TClQaWUpUwU3h%1P5<R7cm
zfpe^E_=+;xC4%5SNXQme(Wy^nG1h(?d}Odj&clnvxT>jMDB^^u!98fjAszr~&%0CK
zFoBDPXWriP0>B>6u=A2I$q16b8dit2sv^uxlERcc4zhBquvrS>=pWBMZZB>4%KG!d
z7elRwI>-2=_%-DdW&ub9`cw@l7R5H0fvJR9w1vow%F1sC_P)(S1JZ39bM+kHNIQ)P
z=ib+jsiI|bXJ2e?mva;9AI%RI6(OXSjuz1_x)s#86y_AEuBbe5aU=6~T7m;F;UD0?
z`AWodT&P!tjKAm?nw#G@n5d$6rqHK-Q0#OZ-~oYe0Kba%Rq322)l!rM3g|5b1QG!y
zF9hu?X1f14<7O9Kib7*)AX6(Di2;DRLHD0op;r>ZP)ul&I<*F3<e@-8MtUeaYk&PR
z`QS~7+2djii_r7LHP?a!%(VU_<=L2~*RXY#<R(VNTIorIho&*>=$9|;e#CF)J`8JL
zug*R(d2f0{CM^l)VD;2%gRc9uL}9@naD`PDfY{rmqko3duax|Sy_dmT<R9F-0Ve<E
z6YKn{ODa?bUp@;|_`*~Ed5$KA{^K1()M`Dw#P^QPp(joq?ztJ$)g>l2#R1WwR$B~f
zGUlt465x>_QoM^RK{2@0#&V+N^p@?$H9je(SnM+@;AF@#(b2%ZxCMl0gB}cgttRWX
zR67ex4l9{Nst0Ems3h7~EPb@IvTJEQTECwlC|j~$cG>Wg49fI%rcv_{1PfVN7ueL8
z7FH(nra?<w>@GKqY^~&TUJY;@^TJ4H$5vFn!b}M?{z4fOl>YG3Xp`w|=_O~)6J;)N
z9)|?W7>u{vSVpleDnV)8^TelTuUR7ObHgy2ZSy@{Dw=%cv<@%Yb*hUL1skP|W2O!*
zESwE4U24$FPVqcdl|3(B8P3m2vvj^2Q$FV=l+|vmQR?jj*h(8_3LSgUK%a5R%)4kk
zczrkMKyuw;J{UJ7P*5_cG+Az7O7@)(&vD^0@$~hjTqlGpAALJLaf(OhNYX6^o?e(L
z!Qus7%$4&S%!ELBO9))d+k8|zSgWP7cJJn#r{!QI|2pI>Y_Q2<jy)WZ+w+g4`2tkx
z;Z>xfcI#M5xzmHAAJ=NCk4bP$G95RP;k2pyep)*J{hrA+l1GB3KZ2%#4Z530^}vh?
zb14*q;Sb&T1X89qyPYIHg}#-O3%wBKU&oRWeZGZJ8R|9NTeRB^CpPk3*(h*F(&Tn=
zArSzm>pl)X&<Q|-ag7kVqFJqs{-Dofx$bS?y~4uEjJ4WF`^UwiPml0A`8n&DF>o4Q
zVXz1dQl-8ptkhBO5^N3Hag(MnIv@lbbeU2&IJt<2N_~=?9#YQOVLypc$P#}m?^!cz
zA>~GnASi^<)aIPXd{MQE1WA&15g?6tz;&Z>NHXat3-T$nuaYx>p<7t2Dempdvr|td
zQ?g6e5L8Jb6>e@HYLlPlhvstwD4-y6X7ua^%T=%mpRi&*Qav=6`I3wkV&c*Hubi5g
z&AK+fmhRT%0fFXU=S?-RU!8&_x8ekAyJyK~4!&!jC9h>U1UqOo7j{Nc-G~~u%cOh!
z#f`FXeiCaoNMjO?;2*P|0Z1XP(<~<lK|1LKnSK^Qqel_fnO!u0tga@OjX%d*KHAJr
zmiT%&^Q6?8!Q0)C3mWjx`>PIv+RFgK0T5Me_%R^zpK>4t?eA;RbjV{xB|)+|ZFKpJ
zwgKv7tA%xQU)dOM-Qf6?&RC!|7mMpRhk7V}{Q3Qpt0Hp2s%Gaut^c5^pq1{L`&~j;
zdY~8Wf9JxQ2Es%9dHcuq>ET)?^n(e&15F>#eW$NVVcj`kd75*u56tU1CHS8{``^6J
z|K$ys&H%;uz|1<NEjn_RV@7^w9J&e5`^{hFLCfr1EXyIBI02?efV>P5bK^@l2@8=O
zhb7#rzO%#o)QpnjDSJcE#qnP3E5=ib!{Se|s`~(3#2Jp!4DzS<ZlZ7p$$|3$B5=_7
zpp!7Q55Ssl?km!HcUA#l%W*^^l76aOmxz1NNqc$_4%&W0{IRz3^<a8a4-}FYvq7^5
zBFc3ks!JsxKcU{fgRkmOQBcj*h8&+i50k;4h6r%%8SMYnVD4Yz@b@+KKV=;L2PWzM
zSuU&62<kLh;0GR#*^dL8)JE@H)i>4*)iegaxTX-`>v^V0_<>L)rl;(iDDiH*r)~Qm
z?m~AEnCi{fZo)XdX#8}-b~c@Pg;h_CGee^{;Er2B9VcU0s*=PB!DnJ(Y6TPfwvc8-
zc@q)&2D`m7M7O|o;e$Y$GyK%7+{LFm#KY<R(LXG*S3avIF7dAT#IZ!E)REhep1GAX
z1%*JN(MvuxPj91xf||HnD>ncL6@Zken_(YC3>yzOIcV}?ZZ@PFY~ZDHy}hb>B{?Ku
zyehJxDp+$hq&5>Zjm+0>cOtoC&na}PToL2%jWZ5>x1wrs8P`Qo<;^gk0AFuA88S2k
zuXtV2g4Xtrg|G*4kZHn4Nja_4Rx}kr1ZMhcwmtc1!F(qrg;*X-?t?O0)d<_%x_L>a
zOJ3+yh1sDeoO6z&l!kpo=~_BizPaI~xseIgsUj<Dt8ubm$89J?-O??Av|}~u!+v^2
z_*C<dtUMysSdy7v4<jzylN3|x5<Z0>K{d18UQ*#?K{FBvgFa%rlHMN^B11)YJoRlp
zD2zE;n8v*n<CBWdyKpA|QY-4=d{J_#{)hHVnLzdgC7{3xCfH^zm)Ch_EU(CByls<{
z!|wW9?ka{?m-U~p4>e`bxp6T~7erduPr+Qaj1AhgIY@ZY^xE4Gw#HD0;3HAch0|+y
zE(l9n0kRkN%;B~@za{Z|hbQHz^rrz8q0OkEl3iC$Z0a;OnlcFif^5jCrxuu;ND7Ek
zX>xbAazqJgaI-mB<Ia$><)D0-JC7L+xiLNWI6twW#(lA8Syx~K4Iq}ZsBaNL1`XBJ
zDg^cMk}#lgyw#^Zd2^_!GENNne7Mg_j%DLQfZ!9ZJdTz}nh95_#<Nf`<}$3vhRik2
z&Oyz)E09I3$PnvwbLN~8guRb9xaM-W^Dx)hYg1o3U1JY*0B{;=?JR-=HgksY2I<kB
zMI-=Y1qa)kj;+!5m6J}}ORV1Me+;<kb<A&ezOyckSWIGfK^tzETev>Q>2hpwPHPz*
zi+2gBTiO8suG{B4S-O5`(c}Zwebb73FuImkU<Bkm%dQq>x8i3Tv&FoXJUbbr%UC|X
z&>Xp0{fhk(*?{tl%oHW3EjSJnG$j8{@p3{E?U}!vCJ>~4d&Fl6`o7+!T*a<uAnN60
z66A5>o53D6#{?f?`GW)~y?<6?#h9x#hf`UBy}ekk#AQaK?q1F4BUFmJy!&YCiFP2n
z=WM(Znfcg1VICg=EePUiszy@pJ@#QgK~+&;SwUJi(_`@Ym&Wyz#%FXmai=;Lv{xc~
z6~zN9iO*X{JFra+pj5z262Ou++^sL1r=*~Mc%=J567m+P>-kibPSd6aI5#z=q6s1u
zF86q>XLLS#)nEg2JksFJ3Eha-sYkp0j$cMV;;rE^O3hc^&`f8W?|^v0RKFdgz(C%{
z@WtZSx@wZnv7FbsI<DK)jq|x(_<8|D0@=pzeYN*5dK^LErpw+JPnI0SryOQaK5%~|
z6g<VaX$DSsh7!%_RA2k3pfu<fJlChu&T%%?m@qceD(IvycxizzE|MYTGyi>9S@CqS
z_fH0|pA5TzOPXc07ra)ONoy|fpp|&^h`^^aJoa_no{lAu@4L~3$O<offb#$4-olBM
zr>;v$Q|BFOCf*AZd|i+>e7$<}yN~nMU)0XPs^{_nRn}wWc`n*)WP`XL19(H<m;`o=
zt9|86_QSllYiBR_*kg8|oDh<cyyu}a<sYSeg4%0VORdcX*M<7Pd0p!!#&X~itKFX5
z(RGI6oUNkLe7qH{dash}uU_pnFPmYYM$IY+z~*j~#O?S;YtW}{J4^Zw(7~QgDrqpe
z+&X`~=o5M)4x&+Q4(~)uS-)R7kC3;V3^ChGNg1B#*})p8HcTlsts@nISUY&Bd8;mc
zV!bxT`Ug?%{RG5n0v3qL7AH^bWQ)YK^I|Y{13YOae)q*$J(X?<&NKCjYq$%VMczlf
z#e=wAW>iS1M8cVK_TB}qb%_w`oA*j!`8tbF0{D+SaqFcE(wII`-O9)1vbI#8Do4C1
zYf@lI<^@MdzA&^Z@MW%%IdKRd^;Q4kA>|NWDi@&$sLKft;UUH4&#-@rxi+{Jb4a$&
z=gUUt1zVl(((^&jP$mq!Z)a4-x(&>J6cs69=}EbCxL4q`4{}~)w0CVo3l-f?%Atuv
z#*rc<fN;g(08LEOHldijoexgy^Qd0iRNK@n&meECPa|W^M|FHF|DO<77Jb0C))v<#
zAYx}EhiX0I`aCKuVM%)(z4&F5ytC<J5<TqaCxffPr~5w{<)o>#C~t*%iZBPU!)JZT
z&ZyKG2+`xzjO##f*4sMFr8^(SO~X&CF^WwMC6L`I^YGv~58z#8LeUWp&Dd#Rlc2H5
zv}h|)pL+uA@{O(8GEIZ%s@Va#ZCgc?!Py|4B)jRhaI5^?f|R@$m%rW9lf8FgvKeoy
z+djicotrH`JlBauDZgxcwamF}Ew>%_TGu%GJf>i@9{F7Sy18xTL;p1Gr>yPI<;Esy
zTBA^u)=IXTVcFM=0DPd9z-cSoSH4qIQ(wOHA_b3UF`TV_HN|#<0B9ab5#Y~#k7cxL
z^O2wwL}~8=HKKhCb)hkvP7EhJ7Y$ZE+?G+-$&|^dEOjKQJsQr!H4W_OH&W2dZm5}(
z5o+OF_54qUA$xeoo<FXCrZ_kH;mLW_zZbs6-u)F+!0_K^N+SclAwU)33x(fBg|C}f
z0*L3gJl|hkH22cEa_f`wD#ss?f{<sLs(Ms0!Wj{(O7y<|Th(!Qv7Fw`@3xbcDvI;u
zOpEu_G)xk8_?7@XQd11y=ht%Owe_;aF@osqDM7bMj>P2t`U+Gu!|ZnwYvu%Ne1z1q
zCJD)y2kN%gE@eH?@&$9*Bm<Fgzr14-FBy(YDH|A24d21wZJcfs5R3xjQp5LT<=$mK
zPXnIb8=pt>FP9fTY%>&<%X~##b{i_rD}6B3nd1s!-RiS99##U9sS(+1sNt<oGlI-U
zuRiLz2)>US5?9GT_mMOIJVQz6jV1dGyKtp=EwLN%N{7rWOkXklU+leiSX0}!HylI(
zMFf>1ARtOnY0{Mz5$Ph*J5icIgeWx#7)9w#KtMoxl`6f3pi}{oP9W3-=`EoK2=QC)
zbM`)GyWhL-zUO)G{qFNV=PwAxRpwfA%rVFK73!PRqx*)lj1*1w<fE1N-C#)(hDFPx
z_>=@3oi-U<X(n~r!ooFmva>2Z^)>w$ZH?S)($zi*DKxtl5jveLUpayiGAXiFi%0`#
zB~Pjbz0%em1%CD9GTnN5dF;VB7K}Yw0i;dA?m&l5QI7wxE=NpKt9ezaOGWKF$a}|P
z>Z;lj63_8`I#D8$0t4K6VtkNW;}N*F7J6Gj<Zq|n{HN2eEoAN=blji=H@i51M+4fM
z6xhpUEb2{-eN$YyM!B8pGeViCE^-Np6+HJ4Nh|tETWK`o#oMVr(=i+FW~BJf#L#|e
zqx_E?<ckSrd8eA~U3ku-OVJk2Dm+i!pHu${P=(X!*+-JE7eha#(ztO2%CycEVIcvi
zllZy!@N8ibiL|PX&g^O?JI#A)f)wp{UMyNBDATICu@{K5|D^sF`FruH(TwM!sNV}v
z{Z2ftDDg7ozYp;@A-#W(#Xm#^|HsDSN(Z?VRiVSb-?ra(^Y31>aTx$8>xYK(zYKd-
zenKwvbzFN%&X^$syS$YG{bKM?WRXQLV$d9ch8)J)Q{Mg6ui@t=90tvVFDI)b01m+l
z-9vz`aOSzpl^>E<r<a*o;?Uxsd|?vIUkTcRE(L;jUj-Uoz-qj9p}*s*qvS=E{*Z2{
zkT8zvVe4Y-t#OMwXKP~oSn()>s(CNf(69Sl71-H@oeiY!(y(Y}m_gLqJ*G>3G(A)$
zKy$Jk30aGr)xAdrSfnI!9Z~`5Ch0^Z$sUNf<w^i9<!kD9&<h<v{K2vBCMyIVXAfAB
z4*=Q^SHT289-^C`{89!%KBh{REKmjkHXA1CfC%J6Z6L|Z1GFIkfddKf{>eokax-xe
zhl>~~VuAgHT_9iro(u-Ca{xY&4bX-F7(WZ*xd6LJA2DNE@$XCde{fWaVYr7u6$R@7
zC=tIgz4Drb3g|>>>*KI?h8x#G`zO}^Z=gDjmR|kZ*Mj_k*K_38$@f3_^MuPU-f8MT
z;?2FI06S=@XE-~r=x@k?KXm1O?$Z57pZ^t?<8|uqpb@s7@1Wbf9+h)!%z^GW*y!e+
zz7^QF{t|cklgbXqNG4NGmoUdnsAz8JDo#fv+f&Z1hwM{_$X%K#;Ub=G>R-W^M67Lf
zHE7;*xaq`fp|V<JclFhrwtQ4zS90yV{T!;SPl6px3rA@6xh;H=d6o=iO0KYpsk+>I
z@nlNP34xE!V4He*$0@TxVG_r5GP~Xzhm5W}1Gvpq`cY?7FU>E*3mrTD4`^4FUft79
z*o6)$FcSIJKh_Dqhq*fBEb2@`DrZrVnX>Y6$*dulNqNB$asl>}dO^L;2{mgs<z;@u
z#&1kvTwC@_fKA<-zIERIb;)puW^Y8c^O$W}$i@!4!^5y0cXi_e52xJ9r?J0=$oLt0
zY5;kNZ_C84GZ?qOx8HxI!XD~y+@jO#sMzbykIy?hm9BvNXsBWq-QC@D-F=O^?Il(Z
zo#r>jH&PjWgz+LPi}?9glpG%H=O!f5%0?YnQfxJd>U`D*w|<e;I=*zfnma|(PE1z(
z+C=h{q$y#y1;WV1;Gn?TJ3k;Xx*bWq^t89Jz9Me=(-X?ks}IiDcigHLvvH>8rg5f`
zNd$ka3cS<MFQ+`S&H|rr-ZgcrGjm`g4L>uAx#;CoXXb5h7h83vK0B)Ci&@-WelFnQ
z1kgaqmrDt!GI8sS#76znK4&@kqlw!?wfd#8>DHVh7lN3nC=UAN+9o~Au~JuBzlw$L
z-kl<zjGw~+0CMMdnba&!DX6FITe?$q+lGyHv?XVjz#t^8FRsCH%5q3V0xFwYr34ie
z?rT6k%CP3?uPEE_)QmlO*F)%OxC`V2nWG(_`$@*rzo@c7&thO}WO%aE!u+G1-vpEH
zbtNUn2eC{yPh#HzFj-0W%05#Kcizh@UmbNG`-N32Dl;1LrT>a*`A>dQzabB#@DKTj
z;|p^ly@cbv3LW<?U%3l-j0wFlivKXCDfh5EU!&Em+(|wk*h!QFs>LWaqN~BsGY%q5
z5uX~09k2@!VMvzGsg1u%oAmyS+FYHWzNiDQpgA?qhwi7~FgdPFA|NEe78!_CwC2pt
zTsnQuStyIw^U61dP9=dOlmb^y0^-vb*uIfC@FBa+Vl>+#K`0SIKw?~x!>QyFO*?Dl
z3&L~JwDy4h?l<Q|b$Z?iv$?`tmaDF>++rG1JBJU}$gt8Zo6PGSOR*Ju>xUkja^BdI
z7W$Bxj|j>wn#!J2j)~|{IfdQL<+_|$cm7da$Yft{2iKUixrPt=6xqtz&ai3q>O^i6
zgmIK8(hQ8LyM+pe%G@LuyQ9WpmV>8E*thaFsu7{gXhT*x>|JX97l{ibN<94?JOlnQ
zgkiaEigeL1p6yxeVct$lJM`);KW0%y*2keoj@fl>j3p4;&8DX)M~EEmrPe|%jTlUJ
zTM-z;YW8S=!NlmqwM<ITw}ZrmceMo?Sssp%u(bYzn7!Q_Gc0V^UQS4U2W53Y)}gZ_
zo00xp06_@@#TJocix-~0Ds^N=OJQPj(y`qtO`UKylZzwK<1O#GFH_x;6?=!Tndv}0
z9DA?8=c60g#K}K|B~ZoS-`gNg^~lsZxwmYZDZfASIe!&;-szDEEv2{`>DfwZ_B!gk
zadq?9I`sk?6X?I<{2-AL*>A_?ShZ=AfZ>m|p$NA7bYVu$n(|!xw03~tJpL#~%pC5b
zee}vpO&PO`YZdk_E*K3CGoQOmtaS<tLobhrS#hrHH7FcD0`#B87E!AO5}|P)W`Vsj
z9k3^bxsTa;XZ8Vz5h|uGAH(mUXtNov14>K8LSS*uPrTHB!72G!-rrrb4EFz^k_eph
zfBbK`zvRugf3XGyY5&dfBxLH)WPh;YpP7UImES3t+V@Mh3+m|iLnHTrQ7tRd{Uhit
zjZWrQ#2gOAjOY--BMx5_0$`F`Mo>!?Ic@jp)X`u$Wvka+-dRLN*tDj<5W6~4-1bR|
zqwlYZ`<;>v0WFXW2uhz&v?xt324cBVFU%(BP8=pBn+_Sy%0qo(<+!dvz*8G8jhLGR
z5C0i01<iGOsHsQ2?fC$f+5zF2ns6J?8UNR>&r#_`)I3;*urC{-L<#&s3QQy=qW^SH
zfXrcJm@`0%ZJDxbsHyQAUj4FskdSpvAWzI^TKvdJhb|P|Xd_GOB%nzwwY7DkOHQtC
z5wbidwA#z_M5O?vZ*cx9;Gp*ylt!QJ1oYe|MRrv;-d~mdRf4Ds;m~t-3B^h@)Kf*q
zYa)jweMN4`#MV2-bqaKiIAy1Zj}Gx#6^&M<3_4;$CMx0!JRm`&$u3B};nPZ{8JX(l
z9{<#Gu9xoXL5Zq<ZWChb%^2nG*01TgWB!98nfjB9^6mRw0UpY{IS+a9E%S+?T7;gK
z5^7@Z$5Qwbgw~M_5?Y*kml;DB-i^d)p+{|N!|(L>T&RzHbVkKomE!C0n4i8x@k(QH
z*E~SK>R;9`CXr{MjX=u(d8bZb4t9#<LQEl3;M-fAoX46no6s1fjKy5AHT202mJ!|d
zRUmK8)c1KH{Ow08u<-iiJjksMkli`$h_*bBA2V|Zow-6XLS!uU6`t{8-Z4zE>3RHq
z&Y4P$E~nr4Ia*Ea(rh-DFJEI(BfS@-@?hd<!;t5p{X~NsV!&uu<JWiRuly9~{L3Rf
zWQrIC)bgh2d~W0~0aD@8X+-T-{9?HhOVHhv0rRfdW8H(nJgyeZSI<>M#|cwL8{1{*
ze#^aDVvA_t`3^ETyg0+@&_?SCYaQEqSkea(@{uIIAeh7Hr=PuGuEloinGzlbPLh2+
z_4|vn4&-#qFfy!{4(yhLwG**ZU$nk-j#y-yRS-3CN#J8d1kSl2uQ_jgta-*s-9jQ<
zcwXmS9Q@<hu?iQBX8w0qK7PZCn|8{@+G;h#ka*7l!EEY9m~Ho6Nfr#0pfv?(V#spN
zDEo{c91Fr-cN_@q;Sxn6dM{h{IfL^YPZ#Y5O0v_+6C?^ws|#D{zX=a9(HRN5ChP5q
zpa%mg_<^Xi7W5X{^u|jl(P~;({;(3OGs`TlWv6+{EYEj-YQ8pFe_!td==o)J`AN*P
z#@R!t<5S)r2u51iZb^G5xfGD;IA`<|ZNDz@_odrjYS9BJfN&#Ox^Glq^oIPa?fl@+
z@rI(a{cmQ22K(uMcaJw(_^Yb7|DE4E^6pPumH(4O=RYFu{g*_yKeq=6Bb^RJ`2!zW
zo-1*O(FnadV=ihutp2F#iWLR7;)8cbo{XT2js8%QP;f3I(i6&c%ZZRQV!--GO>FA1
zC5**>M^2abPKR|lJBW9{TJ+Y$8|L;4Mhzrcq8=9Za{uJM^eeZ<m2fHVIAk<eg{IV^
zi4LuE)X3pT@BQ;nwu|FiCNK9PFXdFF7cULV2tB_au&h(7cJbs>y0x1fTit5ar!B_X
z<<9W*eopUK8xgCEdacNz$gx0a^;^2%MF@Pkh2SDu>)>~T1g!@vd~aQ%Ypr-yJS*qN
zs%CrtrmZdao9Bv0DNTXSAz_{6AahUuO=4X+$7;S@voew8n*;i(cCef$AmK@Tifhu6
ztbm>Fu<LzmAtklzhq;pDlwdKSHXq$Cw5bcYa41&?Y&Wpq8$?2Iuyd;E0>#0$xZ_1x
zp;|($L3I-MRC)Tn?(_u<a~ff~H*Mdh-p$Bg+a;rXcb|r4H<m1G>1@V-LX7(PiU6wH
zfvuSNpgU}L)Arj*+~pRSt{l~;{Bdg8f~BnQ1Y|l(6TWC*)f6iz`hCQSPVTORuC}SO
z+b}wFxRv*Ty;}jTz}Q>6<e?*>p;Zj(JPz|TMfX5sRhakpvv3VL0O3DQ6QFA_!Me1y
zWCZBVrivP=5j_xykxR(szsvxo`fo~oOKxYx#?<+$dS=&gT2FI5fBNuw;7dPZpv5pi
z(K%C5Gq6>%s#?V-Yog1@P#Qwh^~#t*KnfB!-qR*CMMdOn|F~Qf%IeHZ9v%{wbRhS3
zj@iBS>#Wi7KX;Cre;lo(Y%mrtJ_Tz9ck^k*);@*peFr6py7e4eNL2M7YmVD@6H2es
zWFHx~v!CF#b6|QIQ)^srb3W|6Td}}Lzz!4gQQFjp&9jB;To2awMsPX4VYR!k>UI0q
z&&W;KPIb7LRKfVQru=t#txtU!<06*#S9jf3dtnCme+Xy)0hQ}6`BPSycFOu6`XeP*
zeiOvb`#J1T#{S)Wr(pZmuc9O+z5h`V_@Dgyu^ZrF>ERGy1^dBxIO6?_Df}Ri`~E0!
z{L7^mnx87Bj<G-Ig@CNHGUG3y-Y6L^ZCQ#h`FKceI>dh<m5>YfnA})$3L~5Aj0w~`
z-BWM!drZN9-Y5ed6?xJEKDJ|}eQE?LYPmGYU`E4JV^vxhJ^Dzu<;X3+=IcjiO4I01
z6AVLVz~K-Mwi$I|QXU+v-9IA}AFx$0MKh6kY~xZjGOBg(+NIY6*k=wJp)bSV&3o<`
z1h$}P2}Uj20W#<iJHp-g>Canl-D>Bmj7_i42bh>p3%06$P80ud@40Yas|VZe3k5!6
zvt_6bOC8s95h1tG0>F{{POYz<ax}Eyiu$!~rz1|I4i{Zo=5F4^WfP_GMyAAvI2ao|
zu*DMSo4@KYAxMuP4^mgMK8)_zL`&}JjNUaEe9W@u$@i9X0`?umNkG(rW4)+XpH~$f
zTqm|zjvW?%ZhypIy>XIkgK!LV*}H#ckCpu<&-^*b<(*%np&?5PJ>>{V4IjQ6=zhcb
zLK#vtfv3XOgsOvC(XX>P|0H2R@{D<(y~F7Ep11m-#CiZHA3-}<;r!xoBMpfz&ve8g
zZWsWL`ojcQw>cnRV!uJ<neMQN_k24VNfVWD_nOT{P*DK_gJ@$(J-O!2S6w<#yfww7
zp0+f3xdGpn%Mq7qU)3}VoA3YTR_9Lx&v?L-ZY?q=5g^)A@1sGqbs}7xQn4;vO5qrC
zGh1KfkTvZ4!ew8T)EUjp{V>G%t%S?xEIFrUoBPjI1+vIr7qLYGI@P@j7l>WA2PK89
zw48Sh1H5HJLleHe;Eu9z7P1NLf(b$F!yY-GwF3FYb%WzeBp?&cw%-a1w~jj}E;o{o
zpr`PSW~)WSBnTsXe|M(+g)jmuhacMYFZN+~&!1?|T8tMPow}&~5M^&{8QT>s&gc*6
zGIT#&TWyxcimWQ*rh=dG;62h$1$qLST#OnXBdXKkj>B44M{!;kYGVu%o^?FVb)?e2
z&4ps;yqSQ|2Ux@p8>pM>0{xsTgtQI&==}@%`ey=0r=w*L)Ge>@2F%(9Z?Iwwg}CCT
z<C;NFw2273ut-4lu+Ac<1VIsBm9kM8_-c*6(pRC+{MzY)&?hc^IXk20W6FdXDPG;I
zW{!<nM6~H5LL7u7%Ew%Ii!jTuLtc?jIg`6}lii$2%8dGqe^uebOMCSO&ey0Je4-O@
zg<SMZY%9rLP}Qw=DjQYy@}{dOZ}XPPDm18<8o(52AApIn*pfg=3nT1Qz!ukWB8o%G
z-imO100T{MNvo}k;)#G}v??)2>$ZM$(L9OHBsPEA2S<`b^N8Lp%%V%C3hAc<aF|<0
zE-8v9v&24J5G#A#dGmrFMX~m$DVh==@e7f<#89AMb<W|v<M8W6%vP11gzk|whYM99
z4yn#6i&?KXjAv1foljhHudAxdj*^OCsP(x6HWdx=u$1>^+JHp@8YbS3Vk6xo7RyoN
z!8b$AfhK4XVDr6>P^iy$Fw*r+xm>i$aWd+#+`T2Y`|<i5-<<6u;r=3w4=+J0)sF-w
zNE-}37<6{R5~{L_6IJ3}^S)}z-FNcC+~9o>^ObSh1<qO=`4J4->Y<>gK1MiI57dC0
zcpaYPE<!MokM4MsWnJdBW;`8F^WlQwqeh|H)G&7x9bP{21=@^RY+UoODBs+1;4nS5
zs$29}?BmQ6*RGw?t}PDkO039NwaX@U(uvf9Q-4T|XVS<IK=$v%l6O!Rkr(hi;Szwl
z(EKzkh;+l#p}-+Apdfc*OkKV`u#!V&l21(iyqJ>E7mbJo11_2sgZS5?FG8`YOco^X
zVKTMp(sj*nzD@`rx;b>ic$?N;N~dW`^S!Y7N%xAKk_^B&xIwl_Mvl!e2DYO~V}iNv
zSw3A~UM;9Tn-FQa-QgokyaYUP1rb7Z(3VGo_jeGprO+zQ;tuC!8sWC<jF(4QqfN2}
zV&W5(QXXHF(OB;w4Xrw3>Z=b2f#OQF1?uiNdOf3T``}>za>fh){|@pCB>$g48h?fD
z^WXiQ|Bao=WB8q(1EuNR03n}6A=2uJVXzc4W--Z~{mss$N=maO_a7d!&f9qUK$JLf
z9bim^x@-GMSqcugdlkFeuL{ap_}+Ny&~at2nTm(yXyluB0_$8-DhfC};?*jN9-Ac3
zIXSm<@LAUuiF(r|-xB7sq1Ij?eg>iVk*PVbRZzL#u!^i*ov>=hYF;+Waw8|Z&}mqn
z)F3_8c^Lj{?VeZlA8ViV>XW0MkaAcgHc>K;RRR}zK7`doGw;I7!P6F)Ejg{qhsAvm
zX&-Yy0h-It5Ot1_+KixEOrb&-`6y?Pbg={!4z~#z4lnJndleNw@HgOR=e4`Fbmvj}
zoA+Oz%p4ZT>6H-caP!THLX+9pF34#q|HDZ77u)s0Yq*f*m1-7ffl8xwDD*hb>9T8M
z%xt&qE=qGf9bQ9|0c2Q*NV}%ME1)h>O#-@~7fE%N`&{~ckjsds*(5BdKMwA8;3X{C
zc$%a>SW&=iadlna%;l1}x~wVYO5)?Frui<d>|j<Ki?0<(6M`NtH_V5^Zuse|XDp2$
z2&l-MaL3IiwyId>QkZ;e)v%^)QjlFZQx<&o)`npID{3u3uyf!L_YwjHE`)!UR2yG%
zw~Bx51|(OStl7P{(WqUszY($rY5xwY*ai6YjKmTZe0T7WF!DTiC9=92&sUo-T<hbM
zSrZM^zNMA#OMMQ?(UU#0XNyVe*rrPMaHmrkUQCM{%?Y--cDLRp@-a2^RLjGo-UdQO
zVq>Ho1eoZ3d|_o50B@4liB1VtA(ykR{gczo&&lskpZ9XH>mRiP0Zlx0g6BPg>Wd0G
zs1br?!mPN+HMa$ZQ-ngRJ8LPsgWp6OM7_}KeC?H0a^?Dt*5Oxy?;z*nrdJf!rd}e_
zw<?(kQ-KKk#?wTlPm^cy=0@nclAxSMbdg>Q;Jq<OQUQFIF-`}j?SO{zO)a?yHbyb=
zjBm2#%-oP7p#N3b7-XqH6wN0Fs@3#u+a$F}l?h^;hTjPrULHC#CkxF!dpkPM=s2Io
znm}!7O#@oO<4^?FlY<p;-{|rln?DrV8LLx>|Mx!Oe{P=;041Lz1co)*IZ><P70O|0
zP_(K1E2|xB!mXAE+SjNItO|`TTw~?<;Key_Zq(GX$4jOMEFC3N$w3;yt#^e*-WRb&
zgsTmPXyy94L!Qiofq%8-ItmxZ3Qitx97rD?KD?cnOzNnDCbT_XYJM{sQ%*g*%@gen
zp|F>k@^|A}@wL{saF*rUD6G74{NT>;G#v`wCf9n9{X(`W;5tvp%C`l_cSG+k59B&4
zL+#xR+l*RVy&YU#WtXSwR>!I)q6&TsZmv^jkkqySYcdh-@6KLO&n{#*KGx-G^|`$5
zIJFClk+~oBq^^8HI)G*Qo0WH7cY%efK*mCKe}3ta+wg0XGh#1at{ro#ZF#)U|LGi-
zz?@J-%xqKOA-?d<>U`KAfEs?HXf3L$63bm)T^*r*qLyOTgq1?}gfpOVmU`p?P$CQ|
z-Ud8}EHC#~m`&VS+I_}SmN>XLU|vog`@-9!CfMdl!%T>cir{yU$k&7VlGUW1eb5ru
z?#2}0E8TZk?gdzaaww6>0Puu>DN2OylM7MP6DELm=FrsS!8_sd19>9CsBhg#PiI(S
zdC^CaKX_F#^vaB8gUlXLghKSqSVb9Ql!b{!>`Tg!7(?O(Io0VKNZBIzSSwq@vZLXC
zg#vd?bhk)R@rAe>+@e_?UCvCVOnWDJ%><1_t)NxpQ2>FSb_xKgAcO!=$|}%6OWmRv
z1h3CFeg|>9*No+<Gn<8)P9&D0eOVqfSWKnPc$1PRlN@`%g9@O+l#8h0ibzf%HIz=Q
zxTabZ&Zo(%WvZ5WI(XJZ2-1=8`j~?dGWu4DR&e!@DkW3wPd5y%f`Jz(^AJ;A2(qlf
z>>&d38rf|sAaG!}2`GC69+H3uH&cOlGzn;u-1`n{j03PZ9<SLd42^cW=MO>d^ap#u
zyU=d`^Id=(`%ECeyOYe{ACjU#+4iTyM*i;6^imc5`PJV1f9NVK)kRkJE-p`UBR>$M
zX5`s5(A!jl?sW8bW3)q#bM&22RwsXU&Nx|`H*tPxxx-M;PoDkv!}-%t#>TF1>^bC4
zzsyD<z>DR9Xz)Xz|1q(MI`gu*x-Rwkx*R{PsXZn>(2`rSvOFW}wb;0-pl%zMA?hml
zOUQHpMS+<%QO^nX9<Hg|h;~nQHNa0JMInx2O*VDVs9?7HQCA5h0k(`h5XJLjZh!Wm
zD^2UC@;o*q?F;{W=G}DsSs%CZckkv8PwKmKGk_=cOpH3RN~N#FRj}l^!C~|E7&rL!
z`c2mVGSU3sx%MAWfT-(u*vkrfhC|f6LGRYYU946Aj7f>mrP!u$&X4}bUR>|s=bS^e
zm%LG6115cOCV4I4J1ECVp@Ed*;;&IG{Da!$96UXT55uafTeu=t%Y+e4rFhT76HmFw
zv=6?6Z~%zU-+Fp|pK%LJVg@h)Eoe<;?<wOV()#*C|Lv#X&!}&%5PYEz$9K@Jlt93R
zXaqq_^xMVmzNi?xv9>NJiCs3+gkwx{6Pu1kNa)1bS=E3*>ijtZg(X)HbK;cn84j$R
zT36|`@74^TW~_4ihV)<CrRY#(aG6>O=wZ!Exr#K{o0{BeH2PNkOu3o%mab-$Z`XdK
zl?cre<T%MaPx)ac(P~&vPDs9TsK=kbvZ5}`dd?4L<{oli@uXa}Q`hFnh#wk6zrTxN
zNUO+Mvvz!7OI<{4ep|Fwx}}8!+Qah#OKZJM2krZL%9E_x=_Tw0<;Gw^NpK)1egAri
zyPA)xwR-_%+>v|@)j!*aRGYqdt~g!YCg_+K?}6M!$%|regZsUN*ui(_Cx+@!$f+#I
z*NHA;nKK$1v}RitLW|?Okwim<Q{O@9+5j-}(<3lR%b*U=|GcocNT^s&9{X8Xgd?*m
zz)Zm1YvhVc-ig^D1G@_zM|Vm%lE`O7(}<9dn2LDdQHJov=rz1N5O1sYVE!;V_9c&b
zT{lFcHUW2*2Yy=APS<8u6D^Lt&VqA}!JjoiS0-!h6SjSITVaB7ik@se=dVrCyj4G|
zJXdNQBhF|T<ys$Ln{F+j>7bAdfegvZ!t&%e&gG1=ScU@Se`@EdErsYm3?M96P5?ju
zX`f@RylCMWdw)fN#3nzouJpyFaMuWR%Wu!Ab=4?iEZMPp<cp-6=t{^LGs_W&cvDy?
z&p=qbf_Q_UtR~#K<<_v)6(_CwiVxF%YZhS(SoAmA#YQe?T>myCATg)l>L;uj*O{!k
z!p-xFr<`&%9r)y#j&&(+N>hZK?bSJN>-9-ptDtgdyQ(pAY8Vqfou}d~d>nJgqPe(P
zl^TC*_a%<#FcX@T<SX68Fg9-1R#su!_-cN@Br2HVwfR^((#FN;`sc4=;)6gj<#+Er
z^zC7qj|2GCGfwjk>-WVma%e-tU71(Tep(7&p?T>kE6zvhg*bvd*!8DV-WF50zaICw
zHYBH*MTA1o#e!?d(*=&6I*x2Z!|A+l@OxGm+>P9*Z$XU-*P1>UirSn^my{G8df(d7
zT@NW(c-P1a-QDtV!%H}-^ByaS-hcouf(D22WTwcMjSD}-?D1wm%6L|lB)~<D=*<OE
zkLfBdG9Qw<MI6ver$@Lu;1CRu?jn-hkh?%fC7rHAskd|7oa|eJB&Ygk<U57jENi4%
z9W4T1Jr4YGm18crVEGx(;I;_J(G=P1xVtgYE_9ezWj`Om$7|J-%C~@{d39_Lo*ySw
zasExt;n<YXK<anUYsAyd(vx&-?~tXII~b8QK&t}sVvG8xiz;xmq{K6aKak%^$4;Fx
z9yLpkh9)Jk!)L-(4ehRiIOmXO`Vm1WN_<?wjF8k$^Sx@s@kz(iKxjE0yB@@@Ut>HI
z&0g~$1ta0FU(RfNBK=I>Xp^-zt{V6Tee;1du|OxVm3#q>*=(MIK|4q~UV@%*ttRX}
zt<uw>Hdzm>xSx+bPorlEGBW?V><dZlS!r0^8&#za^g%6ERP}iwQWI93rc)_}D>;GS
zNAoPv*``|hqGCEF?87zPlVqbfE}jRp&X;&j9&z0L>DG|(?N4_~@y+)L63^IDbq1*-
zr&%j$*=3g(*Q?D1H<+TridQeu8}huCnm_k`>);%c9jEQTn{uAuQH#?is%J;b0{wD2
zXQx{nx6muPcVK6vJt8>D3V9@G*{qyunlB$|8-^w&Zvz?*i8_$|yN&pr)c1Xrd(7xj
zw=f$()|X!I=ULeI9dx4o)6~|L1q4wDh`zxkM~E<~k_;2lX#yjepFmZT2RI@*f%UYM
z^b?^D=}^st%p+9RY!yyN@2ll=Ja_ss7F?2-j}0lhK9zj9UWSOd_5M}nU2|k8s;eH7
z_z-|y{r>owKfjX}LQIHXRs%bPQ6?#LP4%z*7e-&J>dnT6z@Y`8gmDnn{TJ38u3t{_
z;6GY%z*u9VEZJWkm?p<Q02>3z{yx=V0fg+qh_7YjhDhys^iR+74{_M7LT5W<>MMDN
z>;LilD2BCJp6Zp8-(gP{P#SRS3JBE?&nd5n;J@nTy`diqnz?Y^`@~k_42l=gUYtx`
z19Vz)oNfhL7QP?UO@hW*AnHEd%Vr&T;-{S`&$4-G_#YO$8>zqfBcNywKKxXf`FAVw
zPg;@foOFbU7ufacF~_>Oq#SbzQ}a-|+3_bg^fE_4A4REIk3Jkhk8HR40K4L72v8zD
zpagcj3$IF4kA*}q>$*gWIgX9x`=(m=v^}s53DdXj96+HL(C*^|E~3fgkA&j~6XVkx
zFv)-mUycN_HKiXEikFuW#QgnRKV-=*<9D*T_O^hvh8F-^rFibx!uIwX@hm{ekxL#}
z9lruuAzdOD`6RIV-~eKVe=?92vP*zTtiXX=jRvHCYXwS2AfIHAr0^-(ILK*>%81Zh
z6`tCfh)MNP{SP`pGlm{KAI=%@Cu&Qi#ZK@8%n%XsZE!hIV(S?KG{G9lYI~~61LWgy
zXXsu<5(Kp9vc%BP@GVYfqzNs0qpO|!o^*ky-fO5)iz&Kkip9Pla~V)t8Lo)E+Vvo>
z8!87XIdR)rgD@5$u)uYMWP@*P(K0<y2W&HnD!ngXs!P%u-hxi>E!<k^EGv|L+_KJU
zV>_8vtRkMc>VUAvuEMNV$v*}l@&L>*aah(!gaE?dKCXF6_=kpl5)*8Da>u>09K!3`
zwQHDl%`jZXnc%COh*u{n^QK6A!x8{L;UJRAa`A~dd(Db7XTRDRmoIbo0%z%K6RNFm
zSb3w-Z*&f`mB_we(s9HF_hAx*{H%m@3|LB4+jnDpE)2>U&<7W%ev($;353NK-gz{+
zA)iTGWxW}8ii)18)|){65fl4vCq9K2zde-Q;>6gig`jM|c&%CB$|DBF3op@<+q{ej
zhX=piN8pZ=_Y~&WQ=rblUQQmbY^cI7;uT&v`5c*O&OVG76$TxpxNwIG`_?rjcQ}&1
z`qTRHMWcPY03Z5Hhq~t%H+r-W&NGbr)vHCFrP#=dP@Kl0no6#t0Cx;JYaBHBxI3NF
zz54QcGxn}zk=rO#X;}>q{Z-rsjU^3-=M~#%<;&-oF@PIPe{(g=3P>A}je88_uPwX?
z$G^Q_oC01r?SRBk73Xcp3G=tBHl<3Et>vrLuP;ZRj%j(s<kE1eFRJy9oGqRJD-uzu
z#YTBOq%|`$_bs0K62LcBC-G=jf8|cz^eRLVV$i33pzFWBl1e_>xW^^E8+ssxnjAt*
zivaUR0gV4LMI*<OX9<COk<O+~St8oIzrKC6`YkniN|uk-MSZ?OHKXKH<0;|%IyJmh
zTwU3lLFL!wuPm=jL`jW%JzsNX2gCQ#BX9Hp4Fc>+e5o&rx0-y!XvgCc;?dvd3a|as
zU4SA@+wPnwm5Jrx6`DQ}6tI>rym;A@Wn=KlIjxqezBYAv$C5y2h0}1bE^5$?Rqkwt
zBg?3?r>U#SSl5#{L7(<JEZ<@}<ldpt+rfG80kwgxZcCzhM_zkY$G6E(S!Nx1=FNOV
zFu_kpkp4S$QXjKFe4L>M`g9cwmgJ5_L<I`*aGhnx$+saNeNZyH6PaiQ*a)ay7~qaH
zz-K*Pgv}YNR5O2aVu$iA1Uyr!9CfL|2CYH$$jqKV!ujY&!f6Yvq)d>z0?zkT9;<kr
zlc@M{@RB{ZzsCn)SnJWhzE)o#KZg`B_90xB9MBCMgbJua5e&9g8NY)nF9VZ2_qQ33
z_-(?6{(NFplHfX!>5l%$bUV4oS5e6sk*gZuO@IY9t@#}kMZeNHwYmNGr)u?cFaoG%
zw6Z`}BxR2viUCnBq$Jlq^FTL}w2vU3`40M~bRbquI)o5uf!`zFHvV+2`iS4hx`IW_
zW%6L=caXv_?-T?13($E58<CFy>I;|VME>*e>Jb9|_nrSe%KynTH=hbaTBpGSSEMk8
zwK+?kw9M}>DLwQ&U(9H~J6(E0y-kgEmuZLY_g&mYO^NI-Ubefdk_1mJgveXLvJwgN
z`~tqzrhQ;KBz^urlk!n+`waiiPJTx%{ZpXS&p6I{)1UD3_o$WmmGtKdqJc)PBWMVy
zLI16JfVAG}GhyF(Kp62Kz7bx*&894vj_PAR4hjp~`Me*p<NbGFdVW=9hQs8g0setX
zD6NH__B#rp>=gkCU63M0)C=mPl%JQ5@f8<RcRdp!JPbu~z>{##N^z_bhpxDQ$Z7A;
z#^YMq#+AwbtI?#*l`1d3n4So%)uv+XV0!gchWe09tUN!C^CfmwwPn6V>I&e(M>{BD
zX|KS&#L1ncm)fCz0aY2{t)@G8#g2dFODWEVT|s|r+jlq6tw`H&z+JOj?CZMTz@31$
z4R1HEqC25gN@t_qj3YU+Y#u{X@w=Z&KyZ~A!<t?s9kzkl*HkK0CtkZ>Xc{_br776T
zz)HdMP0BS!=X$N)W3vi$_Fkjx#9qNpRaCT|ZdJGRyfdULUYp-ZhD5+N2JWIcC5W2Y
z7R23l7DEL=y#H_}?p#s}ItW(mG~Ck8<6K5vzE^Diq0;^(DF0)zV#ze@wBw{y3>NZS
zP92x*L3d)EeCB$p-b<szu6slC10|>)2gltgu{&m;H~Cy5O1NJNY6Q}4`&gdue^58D
zX5GIsQ!|hTHM<fwK{<A7Oj?z-O@}_7)8O%`)-9&^LVyp22ttbU2TDRt*e9k2W+fCu
z)R6<}iGA4)C&Rge-=tfud^KYoS6_l=xZE|;)U1XX2_UAW2m0Kuy)wU_F<DhfhLk`y
z6<i`)WlEYCCAEy<NydZ7v(qA+ISIVo<V?wq_&aswem7=P46V>d`q!2{6RvnD(q4J=
zz8L7yvb1P3TKY)!AB{nBF`O#JIS5ILut4DM2}(`FGCtCVy<HM2vPPSxG_SnB=!&1G
zHrk)$q98g>HF4pCTd@;-U|QdI7{r6@uv=XnH@;Y^OeCGs?h12N$2C$AAGRkl^vP*r
zC-_@Tuh~AybsQG8Fbjax`P4ENt6E10D4@}>lwjHQp)11G4P&)A$r4ku4UJR-i%=7R
z^N3I30&Q&IreL>qhNZZtDfKY<Eva&hW0_;h9V?Tg3?1%*I>`RpqU^G1R(f;eW5M&^
zK{mdwYSn;y;ia>diZI{I(W1zpGq{`4m?dC`#r58`W|mjHyxjQ{Z?oHdC>jaIt_n0g
zmCT9CaEN;<6=RoTQ5leWMDggA!@+Qn0p~{}`S&DtJXPfkjgLtlk-gcIeARy@TF7Nt
ztvY&NDw5MoaK-<V`AG)&$5w;G`Oz`7=JF=0VWP4(B?YJ%MQX?2#Vo$xbS}-eghyie
zTOq-_3Zew&n3;7tM@(g9K+?nFapi|V!A+JoJZW(OcRXzAgXme6b$(3Eby0r>K`sVM
zQ*!S!0T?qeQv-!}kIl)<aKJn(m9H#{{qppYCw(2y{qjpS6d7;iY^b=E#LsSh&BPac
zGHnlY@-ah9A0F9l^?0F-s(+HB$h*buT5U7ySuFPCWt*16BwY2@;DurCx<0pqx9)E?
zwy)1sh7a#~(CmDrIh-Cjf-T<&b1U4QRd;Tbm)!Oy7fmF-SH-wWW>1YztS-cbb~Fg3
zZy;J}!(H_Uo`H$2au0C6GRsYd%ujRtt#hGg$4pJe`4X7An1jy=c3iqIp&Jku-g4Q_
z3JNdRCngX)+o#T2_lpR+smqCQWGsrC!}o#GHk*~YEl1Oo)<f7sL8B)XMf}o&eNoPR
zugq`2istdKz_lVtGm{D-DX4|1D0Ak?xF=<n7s6DSZ7-dnd(Ovx$(}OX7M_Wve<25G
z&ejD6A(^lf&%T4S?o~;f=ENTRQkBN>@%GnOX-6qHqCV&MlOge`CE7}o?lidtW_zo-
zIpO<_dhJM2C+Sv^5XkAXZK$h&16dFCgu~9TjAdZS_+nh|$KlssVK!=6ijT|hZeG}I
zPBZgOqnvY|MttDL)B6MYYBKcZu#Y0NKoPg>ttdbYA<uLfh2C+wr7nKib?(|!_!!Nd
z6$;fs6fe<ix`z##XjcgshcC#imF6%(>6CN_xNB-ExT{|>Mr6MZpa|%{!J0ck636?J
zE8jy|EE)TR-JLYft*)2M4Wv7)WvCe5j=NQ#te?Sf?)3>jnG+7n-ewcf^(UTt8%s#?
z?M`XKP3<);vsquRxVj~Ha%)jzt9DQ-%M1Y$CZ~vX$Gd$PTO6(f2Nk>JtvteX-T<;s
z8#Nw9sXfyt2`y1CoGLE2B-O%>x8oq;m?}S%AORdAp$BdTaGmv~z1!+r$(UDVU+6kE
zSXs|IfiG~Mn&vRR`eG|ERDlUD(2j{q%_F9M+NTcMrn^4Zxo$Xcp)PK4jqad2O8xFl
z@Rjq=p>7Lx-i|IyDouk+VjqQU>NkwK;Y!S><J&ZjeDRoPwelXmE7aDt?^70jXcir_
zX&Hx3GCUXqG=xH{vtuA&Gd#c%%?+*O;`BL2^u`HUTD#^inKC7JkADN%#-{78KL@ei
zkUO=#5UZ!!w<0H>jd4gc2(L=W!Ki!~8y#&bRXXW-EiWkT*gYALsNr%()|>e&yf&k6
zj0AN709pCvKV$6)#1Vl6m8ZH5yh8<as{80Iw0@s6El|Gw4KQBN`H%9=A9#SX)_=FF
z%s;&MD~x>WXO_(WAD=1iYHx>cX9WL(AP1%VLLIDRYTPg1esBL1=;!iZ{d`0qa@%^@
z$c<~8g=-rP05nXqf0pC8$>#QVQ#+<V6k>{pe^TUlL(!)Wh}GwROQzFHd@kwqD;uv6
z#sQyECOTo|HCs>@>!BPhcf-_O0iDU1viJ8hPq!7WsNcVquPs>)`xT1~Z(q$QlU`|_
zDWxrRWw*a25n`g`n<03Hf|YeOYMV){PAhF&7WQAObwIzZvk=Q$r%oZY2w-Ev$;4;1
z1SM6xeCyP!oc==7uT480&%Acb<YLt4+aH&DrFAad8;hl}XnnA_%f+N{9?ow}*oj5t
zEX9?<g};OJBgTJ3$(Gf^2|=9>nAfp})^|D|^G!m9N{v4!7(JvhEUxuJ76VldnD%V>
z%GPjqzD`VHpy3K?VI9~G`c70SbZ4NP`nN{psBtAP>mvQ-p5R27QdzVFP2+b`gM=kz
zeRaQ#Xg!8yhN#-7jfve<G80rgIwl%$?|LLnRZin^1$xc*#P-%ZQw;7&-dqe&-DR}m
z%nPTrEWb(SO;JbN=<D2KN@c1OUGWzJaD%lnlT)%|!&3rVm0K5uG~^D2x;KUn_vPhf
zWK&HR$%6fr+NYQH>59-v`HP~Tgk~G#W;#t?zKy_1Rbkl=`=3NSX@wd!PYM%`?S887
ziChNc7RQ&)jsjE<Da}KeiC3y}fEO#j*kot9+R-Op`n}46O|MM@Vt>Qy!4h1-;8P^p
znI}?6)cMjHHvSQ(M?iTA9!LJ_CCM0~`QnK;jIp#;&_*QAr1xzS$l15c*JwKfQKGrL
z8&HMbAD+S=IBckn#Llx(uhLG#n4mlRhLL%WZ1EUj%YHh;vD+wLK9l&zHd?opb_}^=
zV}iB6p73205SZcj+qR(Iu?DbfNG64H-y*srWO)V6qFXpJ*(8Pk`q?|{mu?8&3L|_p
z2Z0uSC~@nwmOhv9JMXcy1+=NBIvy~}Xqa{uoL$ulKJxUMK>5Ra`hH?GS(r!(00E5#
zz&LS9Q?o@Tm5kM%d!}QDhg;>_77a{<UH@%?Q4}3XidYVv0W&Odf(caR*{Z(xF{>?A
zXS-ZIq96`kx$#ncL6;oR--2x-SH==TsjZo_pA<W=36~I#5}R9g*{YHW%5DW_e8c?A
z$8$RhR2_H=OanHuZbg4qQY&20Sg}0;s$^ce3iBS_pQN2@*xr?ws(Y%M;DcHO#{<Qf
zg0U4xfth5D8S{P#3kdEwAuxDN(n%H<vS?VPRd`qMtTx~P)URn+NjIzJ&-*$c%A_}T
zq?3_?I-pqLs?S{lUp1~?^H5q9+3y1tj-Q(#9^9)6X&a`+o*C`R<V~oIntORvBErVi
zC89XB{y4Dl^g7qofHJy=ICx6(DW0pnflPu(d~4i6@0Q5=CK<cJ$5bs8hRKt2T83Ze
zwj36vBAtZxgOSZ=sW+H*uF90vS23MPo_!2$Q8?}J+Q+z#AQCF4hbv()gZuTg%0z@8
zMq1-E>)ExW&OL++t(p$A>hgJUZ_I7-C~-?u#6Ca0id8tXs6Gw$e?}9dIJEuwYc@e@
zu5)vhZFLrdRDTa?vu-J?xpymk+S;40r?xgjW@BpgdI{-x`&N8vw@D?~GwA?1wRbgs
z`RtEzo3PeL61bYQLmJPn`B>wG_}eGhhAN>@I;kFunvkBmas3{e>_E!ja+Y=Iz7=2i
zq;}=w$JQr&j1?{*D-Y91hLb`Bl~~{DWdSYK+XEGcR^CuEfwx;DYg}$>pX^j0Qz~DE
ze?ECDz>e2~CutQCgPz~>8Xd2szOo^-hVGUH$4*W73K#nND;x9|e@)LGoe*1fNLN3x
zZ>!k8Y?z<*&g@R7m|jTq62LvU==W#REI%>n|F1YamkiN-2MsEbao<5#8~<fJmj6hn
z<zM;x{|XbCpJ9-z&=4Zfu;6L@1UUFnl>=$X@tYisKbFSgAFE^@rQvT}%%URuM9-kD
zj;&SGCba&*JV?{Q`$RY6bB3EX_Z2^ZHe_BH{<fIri<>Re3A`;?(Y@I2*bq}n`GNl{
zs{)Y`w7h<5(KMwfXDL3>HhD5PCxI^f+s?wKX#&fm0RBr?YT>eKpQndKUOhVt=laN&
zP0q$fQV|i&Pa}YeBUd~?zc^Ni21~6^%GuMBT2j=%%~d|@E=KgKY6GfN$HByHE;}@M
z;Sx~GoIaXLdKytpe@UdN@r2LWX(x!zkdO5AA!YM^PL_HDs=wm1P;OlpvjQ|Dz^9ad
zjz1=`)^{3wa>?>6jMftFsa}WgQsH)D_CA;LIgQ)zt;Wj73+Js_nXFeZ^M^-)aw`c0
zfUP5ScAQ+vet=9CNF4z3yf-nOpdLQdFr{xs9ETfe)(l8MXzlA1lAeGn6TdcV2hiBO
za&CDPl*Y(uP0F&Khw172$l_riIqCb>wvwqoPO|q&)OT3d)RqoZgnGmIqs7>y-EN=w
zlCAHjqDaXh0Q=LEZ^LTk&#XA2xgG$AWHSoOCb-AHKuSNvby&>CSXTE7_3ur5I@R_#
zspGyHzudRED)OHW9hWAn2lxb%DTnxVrgi5TeXo#@iOV_d62+XG@f&_1*<9L&G7vR?
z8dIk9OCRtPYoRxF%t^fOs$brVo>-X?ZRmZi!7mWTD)*5KF*#20)hqWf$s7P^MgWL~
z=VuUvC57Leo!xW-WHU(z_}hk6=`NT)igTHyyQS$_9@e&dOY!o%;%hdWAYVWe@-g*|
zxpEj8cs=w9v5leYKJ3_o9%i%E4fo}6VQN&Uebef@9K!R**Wg8va%k$-=>ecJengur
zuR2=UvrhvQs)-c*Kn9lyHA{neZ;LwvLa}0-ZLi^NGfgrl$-*ygwU%_`+(HwcXWumQ
zL)~Xrw`5PhnEDXk^J(8X(>u=;u6Mh#v7tdrV`+&)hB3~&a-s}7B5bNt`(b{LJyH(n
zmHp{yof?Z9Q_{A_y9S>DTOMOVcu6RX#Fb`AEbhHrjdqs!%!7I*?6LB$c)RWHHF!sn
zY!A(X=_|GtX;a};*HdLFmbvGLOY5eDKXHzJ;zF{6%%z^`M7(MI-D@BD`zz3-FWOIh
z()PM7sx9=q{gM2<k03YDQ@sL#KYN?<P6N17-#>d|zaUfp<h;MYN`D&k-=Fc`4`k}^
zCQ#miR4pJN#{vEh0s#!vnZY0E*8dGx@w2Mz;7=Di)J?h*XrK*9ph-(txilY!d)Hp9
zaI-!LY&+LEGLLyK2DY2p<EsK|Y&-r_H~tSUVK<Ebv1sXE`fZo@FDn=RRodTQ`QHBm
zFM-DpdUtjUE}+dGAW~BNpscL2x5zGK4%?*42zAa*ev8is@x**QQjggLpCVeK+onKn
zC8ws;C8SKN10H-x(R4W)dcG?C-1`g9N36sA#7BCm-a$@R&{;P<EeZ0G&r`&W3CY|r
zK{s!<!#C8~VApbnF*X;98KZXNa~uPkAwl~_tvwWrmO_1J=H$h7P1tylH>2OHvq_z#
zxWpY5cr^0qruI=ES%PPi$QgJ{t4swt<-!0mTz78T#1U<T#(e2{LTh@XbVHXWI|9RX
zX;<!i9_TVidPfw`Sv1Y{f|+4CZa2{*H*a!$>CB>M>D%z2ha;s|IdwGJcrad6Drxi?
ztH145MS>J$`pIl?Sv`VqS^uG5E9JS>0iOZ2FK4b>n}_$ymR&4XF?n7z_N&57H&AJc
zq+FVmBf`GfUlW?E%yvxlsCTg15VYFRl&vXNM(uFl`+V^-YY^xI&x;t%%ALB#XjI9!
z%4(gGWfIB;#YBL$tZJGK+^xrNs`kl&;l@Zj{N02jv*v7hZYpf&cIb^0VHa!O2!M|b
zG8ZKw!Z<1G%ZEiS5UK<DD*DnQmi1~P4Xf=}9z<!X@_gYsoq3r@{652RrkQg1TapT~
z55H`TA5JD*Qb7f*S>psata?lbDn1vshIhV#&yGD1^utcG2GxsF#3tx}n3I5jmL^$<
zw0N;O$O%Ul8A!nPn<0`qcG*`x3whtFmKHo3w8;y1<_+)-Vbh-Q3;B|NWe1h=tOTL?
zR7BR1d8(iiz-`G}G%}JRvn0SF&u9tYKm_(e9}<$bgrUb6_8jQ8>-j+;^2s_@n+754
zf(z-c(H1IV%Jxf~*LvL&N*<twx0bIwvI_CeTN9oMFPe!Rz34>&zFL3jYRvOiHP#F0
zE<i=P!omngNm3)A1AK&vg~P?*c8a#HSv}Tyj^z$@F%}kD99fI!TjR#F9BP%+tTAcN
zvo=%DOA17CET(d7Z{wdKQEUzPp1?<s8mqQC+-CACp8Ag_N<_@+3{c&BrIF`F7akGT
z+<(f#;5&#H)-XARpD8Q%TqRZ_>h|}!%m7a&@GFv+`};9Cw3SUHs92)bCzR-xe2YD|
zB1TY3<C2w5S>d%2@RhCN5G4Jl&Gwmf45bo*B@ii81b0=$?*<506f8GDlxzCs_V`jJ
zidHX%*!OmR2;#F^I0>E0Gvtks-$-s4JODa+knmz;s>QfFbboAXY_q&hnNdgW!M~t+
za?FC?uoPWUFRy(ESv}}BHljN>ywPtE$BJ`58GY$JckCO#bLSZpnO+dg$m*~##0s4k
zAgaoV6d)g8KDYr#?>G?YVQKK0mUy;^Y;RGNHhu6JGp(awAsFNHGhe=Fb_8rx$IM1H
z3Zj;Oe}v~ZB7Z|kA#?ws;_wH_gHwjT2{!zDkFfw=5(Y1qkURg$sPd0{3_qfG9tOY^
zQj<h|eO>8dV~hQv61&o;&Rowla~E7`_rFKMlLBI-82Y6$mm5HOkVYKSjy`Cvn@lb-
zQRVhubL+}myJP}7nr_2G)qJJh{PePUY7-<hQM;|@xH?YjRF`f=KY2)~P|We3pvsJ$
zxZr(p+vB`0k6prO&ZW`c`K|GpmG@f+Er#H7BN12=qVNplfPN=F#?=9~gOKg)Np)pp
zuXcjf&#krH7>f{31T9}K)@<Wc08|?ES|@4x6q<KRi0ZKU(gW4x@|8)8EHwMhs_f>s
z)ixFBhjRBj(n*GR?GQ;o@K=bvO@^wd=<uy2!fs^gwU+C*^HvZc<Eoy7XOFC-dL+Ck
z28XpnkOJB8F~!<G0Ks0h7|(n1gI6TyYD96P7=sF{KHl3<0(=g6><P}hgsyc4aY`L0
zU}88qqo!p#KGA&XnzCx(vu&#{`4p_Y&)1EF$h`1do91<W_X&o<Ql$<@tEJg|o`wlM
zIVpN0!i4kMSgg(`L#u1&?oapk6>M>tad0rT6Jay_&*dIwF4Z*#l3cM3o*Mo4I`XC*
z>KYnanC%%$9g@!%T053@oo<T3RKw)-G8Q*^q)i+AE6PkYpG9@M1a&)cPKn(=J|{7m
zl$thKD;EOOYN;F<$fymc#KpgqGs4j}P~Rqayex#qy$%fx7uOz4PnkRSS(gK+BAD-)
z%VRyqnNj!Dss^x52>3rIBIbdu^ubb%Cw;)jmKS$ThjS60oXOb>JG{r*(mnY+Zu^hw
zzzsd@E?m7-I<3oQB(pEj?xRfTG6pn#CcIn0%v&7PO;3eWoy*O-Gl=mAHz(9L?Z#<C
zz3*r|mb$t|x9L*Dx*|lD@KLC%?&l9}w`7OA2LwyZTUd^V4#!ShG`{;CG|2-0W<dL{
z>>BHZA+u6x!0{W>yOLK7#zwXnqqu#p4<OCt%j|nCR0YeFc|W@QTSxWAa2uR!dU*X@
zw<<vz7nh2k!G4<*WRR+K?slSkXQ<0J_TJepg8eZMzhkD{`*?l|b-68Z#4;L&%&f>j
zrkWD7XCM?rX+yM}?Q}PDN0GDO{i?{@1jg#VyTZb{RO1^aAN$H+hHY&pPf4G8(kR_J
zMGt*OH|*X+MT}`m;bD8!XcujSo(X9a2+P~Ztqr!@NSZtUh@V&GM((&-ayJt;h3Bvk
zDS({{lZp4U8zKofQTb<wYg#l6A~op>cUoU%DIKd-nwQzT`I!SZQr7m&>7!8^NtD=t
zg@mTk7y3NJc3bvu-ma}S`4;7D<!V1^THks@)mEM1+GI9L@79p3+tjQ_3Gv!f%~+tA
zOaVr9+k5C6iO;}-F#OL+)ZVz|gO^V-$3DiW3|~)J3-dQThXr9LRowJJq*inaGzDK0
zs=)8~bTaAi*Nnr=r}x*4qkVqNH|pZV>%1m`;9KRS>_O&SC(~2$@|*zmnBwPTed*r{
z#DBR0_Fq#VE)mDC`>*;q;i)z7t6MQjwgqrI_T%5GFQ}DvIi3!QrlAYC{`r!8_*2^_
z5L)oJcp{kmaTu}BeVMY|=Mw&G>%Ma(?(??AGuwgfE+H$-jA5WfZOYn)SwaSLypW?+
z#N?#0Q07KC{rSs4qO5hm{wW_=MJTpX>7gl{gGkx@g&1>FtyAfC!#MLV8^<qA><$J~
z89z3b^xpcZyHg?$fFgwo;O*1ohTH|iDd*u!QoF-d-mzL=;&xAu(<oXovTRiOsewLp
zl-&BqYLU(BCRsb(56{;QhcKbePuN>>5QI~+iH^_)BYhNs`K4D@p%d@Bsy?%}Y8@53
zs6-DdtMQPFH}0EA18IpLifx&bgXno6rJMt|o7WhUm~dQjx(&#|h&9BgK0wTMTpxQo
zGz;HlZZppHF+xk|@MIr)ji0R~4M1K(3LdsdvD^a+7AMI^;4FH0sp<za^e2hcYN0wc
z4SlwL!;_1=?lpA*gh3_O`O>Iz%FittCjVc1-yPM|nzxGw6qP1Lst^?v1p%c>jg2M;
zBcODMij)uoA{_!z5Rk4Qa0DdMrABI~iF5%0X+c6w=slqZ2=VSS^UZuS=X_;m?z(ra
z``z<b*4lfsdG}j?<@r6&^FyzeE6^c(fWkfa)pn(rP*qJWj!NIFXT1J)$sO+Z#4qca
z!SQgf!XPWB;d?wM8%9#EU|ha2j?f_2srR(RwFwoPrej%g!&^Pv&T3hvv8SchSX4~f
zPs!eYr=pX8qxQlX^Wcg57ZrCI{fJ$gb<JA;$mwNCs;{))LeCH$kN0$+i4C{ZUTa%1
zXig3bs#7ie-iqg{z;kxFVVs846eetJUKwVWo*Q*jFLY2f$BjkKJ{P(4+1@9<c?W+&
z>AR;D<d{$H?cVa95XR}UTKrJCLA0t}-D`8~aOKyCmY~))pCih1cDS~Yt(kiS#S}x7
z9<FbVys_;1z#sk!dvZb~&(u2T=og8{rwB@CXYSmMg`Ha*03d-XJ96qvcaWvQNxA@x
z7)W_ZQ`fN_6Kr&}u<pJlK-EmqsvzZ1kRX>qIXOk&Ld2z_?o^6Wxe9xbbf!lCt%)jc
z<Fk&BjtcGd<u9n&6o+LGFH~Of<u`qrUB~SrQ>u73UtV95iBrGJvbd(qBsHt1lChDP
zYJEl!7cN)JI0dAJn+N`*{IK*Fl#?9Sl8oCVSa?3!u<ext#k8DQ)E#YyT`}`2wY0YE
znm%$`<0Dq=!3~{*kn5w@wCc5m=fCaemKg>twgu5`sUK)v6alhBJPE={%O~&eADPtU
z&n0z^$;>6ZscbrK)53duj&=2j=%LfEIRGn+m>UUpa-|ub>Gizev7jM<fWrdvVJiH_
zx{@O<>~Hwe6g`kE=ggVQ^IDHe66Z5vIyM+up5*Sr!a!1?wf_&MR48MFhSsD>k}%;F
zJ;RQ-DF?VR3{BE9hOnu@??njLm`Qv3sRNt19-=dvE5%)d{A}7CNdnPK(TCy<X~8P5
z80TT778b+M`5`r>oUCnXX&c|0bM*<OiqtdeDKiP%>8C5DiSr+g^lyW^;1eBynAQ)b
z&Z4D4==uZ|yNj{}@jKgc4<hv`J7mcaq-M5Hr#0=q^t{nRvUilnKFRykgbSo{q(OB0
z(aH+{;gGTm_QsEoj@%F01o-#sxQeZPM#vM}x?z~?l7TuUQh7*Wp-pCWRhUurc>evT
zt8e(e2uantzlhi?jp$uE@U_}w{8FRuO;>_Itl&%Q8i8QW#dxF_Eyb%Q;7civ5)%)D
zd-R?Rg(ccRkj`^6mooI)g(lVJgid~?aW$#UHFFKCH0~Wq$@js0_FGN*a5H<12xPKv
z!rWw<O&>{OJqKj01tZa0B}|-J7FUw@gd@VWZ}SEd^#cU}C@h+^fpc6DH*&#u0BC#!
zG`4fj5I*S|Wp|L~@gBtB;jn5MKi@kC*pD^}7Ro1HD82l4=ZRh1_|bc;m)~-oLGm0!
zJ|1}HlfvLvX~l%8wS|_ig$v6p3&f$h&6oK72=8;W9WJ@t6K|6pIrdn9i|MQO@GJJh
zj_k1BP1Ekox1(s&iLNZX1Q6~H!(EApQWbE4CgKxDY|vkT>HPK6AFO!4a}j^~ud-c#
zbMK#G#rv1Rp4E^@5O)!dk0!j)?5%9#taGtw<w5z21ok1`?$<X$zK3Yb-+o?A(t%uU
zSo~Y>D0?>2anc3?sH$_>&~mS*>LLQNhhdO>jS~H{nY_*sJyNnOUr(B^7+<@Zm#~t%
z#jveE9<!H#R3}gHeuvcn5vOJwP!o>xQ$I;S{BA$}#ts2@m)#!rIGHfg@tu!HPdc=F
z^=N?~YYT>%hSA3kdO(C|DD%EmIWxo19mwG6G$Gj(kq2hFQueBX*St@~NY56j+@*$v
zAmyfLI5Ak7^dJsoB^y_BOY3T{-0JG)u3;+x^(nQqd0Af_%7&79tUI><*;M{BA742@
zVI)C)`-4eD4k6OAS2Wm-P6s&WhBO{#v}9XD7x^J>islj0R*%x{V=jD&J9EBI;jL3U
z*4|f>2a7RSz!;2V`m6IgQi?{fNl7jk+p#AB_W87>CO<yE+C|&beVd!TlLqhgi!qO0
znK$lN>MPHRSymF7ln)Wz`m!jPC`FLy{OctxVZ!FxiTRC0^9l~BRrLbj<0YnS5|8s5
zSWbx@Ra!TTpmR)tczw8<CQi5@Gz`-LZO1XBN$~<v`Tfc661zv0Yw$;hvzvLs#h5C~
zG_*`=u)loC9)ypA9Z?+LVU>Z>gyjl|;1VyKm@ztL>(iqft6!CN4;vwOMJBOtAb;hS
zP0xu-c9Ieg5;L}p$2>Or=|tR+rAn`I`OGj7!U58N`@@S#7L2n<zY0L_^JZfpcffgr
zrGWcWHjsNUh`g+0)0c-8!^>BBN2PD6is-d>c>$z{NhB_oMWSgE01R<T=FBYsC*)cz
z`b8=C+iK0&e9HVJFZwm}IM*dszun3;5~u|fRpZ)X=saQFry+?Tcea&(DgkN$J<n$&
z5RY5cRe7M5jc32U+BsIkFXb@m%^YFh7NMbE0LSrCV-iRhUceqYvDGE5-@moSwq(mw
zouXI&LiXH70*2+`lbn#ODo@{YN6j8y(XT$59Jn5lxPqzCFR#B}xU(tcr)rfun`}5y
zh650zd7E-E3?RVX$VQH3hLtT$IDMij=bew+VFdP+bZx^NBBje$IUim*n7k>!)pKg?
zCEkfcr3EZRM@@2lQp%;DO(g7qh|$XPYbvK<gC<UGakl3kv9Y(u9Bnb&H}G*arm{91
zSOy#?kawEZBr5l|lU!kFHF%0@9O>+}L+$aMSEf#_wfXJVwm~_OerQ%CH^ymPbdWU)
z>3BXbn-<C!!WhiCS&UEa|2bd&a6xuW+*9s*sqAiS@5W@dpvv_<^yHoOPn-Q%=&g|?
zh5_7nbWa)pRFtEGW$V~T9H<t?X#nc>9HU|6ge3+h6~XT|UHVjgFbq-YbUbHq?2wOK
zj9=P>EiH`XdVJ9|*`<tZ1JWOGm5305t1r<$B>SFoFn`1S3%}#zrytl$Oqs7LO(q-G
z=Qf}Pk&r(6&B<tYn@?1x@~|68+LX<`Q~|#yrrcKNlh@c?D#@y*fwJsw<mpd#?~aes
zcqw>mK%u@)Rzo*eUpYM!snoaWJwPv$(0>-+Id>!8j{P%BH>Y01yz_K_;A#(0MDn(1
zvD(;w@jy=aBUWaxe76PKZ%QD_S$Q)=4$_eqJt>_{0Nz{&ljU>8vr|>>blISiUgkMF
z<8>C8vMVvu?KC^*ox^|zCo9TXf`;XKv0AiQf6AOzv1Umyl(ZEN9&Tgd{>n3b^seqx
zWEbYB!}=GHx+60XuC+(A($FN!;oN8>xWE~kCeh32$%Mmmx_dITg`<gv&DXhrQaOcV
zF@p)egq?}IJm%PhVx>j&G@;6Fab!amcKpmjTx2ko{<_fR-R}3trMizZR<J?Jm7Ld4
zs+3D=t{^l<5{8Q6Y$|GM%Dm65UC2dvulGX@13U=b3;xl$z&^MrrkK*OjgzMw^hBnQ
zN_;yrcpAP*tWEJaL~jV0p7A+Fh{i|L)jb_^Y??3&?v~C;N4gt**2Evzrn-JLy8J{v
zRu+=*EKp}E7%Jmqb{mhQDFJZ@bBaTK4*kYArD&-&BV0|dWK8R^6^wY=&Rv_Ej>Fjs
z`Q1&O30dy$p?80N;f^$<<hIEBpj(ysaiGD+eSxBjDq@!@brX&<V%dQ5@?lEE_taIN
zuio3!hZ@J8A3Bd2yVJE$TrX^0yfmm{KiC%7U-TZkGHAO1or8~L;yTur5n%IAgUKsT
zToG%+-R_8ImFtC}LeE)RX<%ul4zIMQO@yMMmvkknX*#8el!lfqnVjHvwwx__m`7{6
z8*JF$_4Rf4dSqzHlZ2KAfuF9D+q4X#qAQlfgs84`i6P=3c3T#cjdHNob)9qe!u{;+
zPl~ajoO(a6(e;uVrfWN#+p+DRjhP0vj*a;>*E{Ht=*0^eS${gE^We!dVWA#2H_ZlR
zGfVmv0hYGn;+(5`kF+nXwtc%GNhzpk0=Vr_Zd5=dBN{0Ra}&)Tug<dmeqCeudz`+A
zRN`kz7rpL7SG#VF7$H#TNgKt~zQi_75;!7Fl`7#ScDjzj`JS|P$R@8!g@}1oS8qB#
z^w7R4xo5eRkA<lvaKOT-^x6YEUK;Cm8qJO>(CBR|-j8%lZ^*H<>2Z9zV_B5HD<^VW
ze#ZVm{sK!}y5ww>ORGXCU6VGBJIiSQwgo;yYq~}?;VK*dK1K^%TlOkFeO5>KdF*nW
z?s{BdJkRIiBU9x*Yhg&0fI^zxXNEAIf-BoDVkUdetY&ho#y)+0iOW6DkfZT-`%EV{
z7n^JJyL&aEr&CQ!4c_UfD<{xX<M0j%dluF$DxF>D9O4naAKOQ22Ca${w*aX0BqSmc
zd_u#BY@=epzhmi=AR0v=R%8=<I5xd6`*zIVvvau;Boxe?CS6!DxQt0$DRLpz^2aGE
zM0)cGX5D2IKin<&la@)(**bg^AU<2Hb&>wiKFRpo*+Jf)4P+dMEThMXCt;@*bg1g$
zC|{fFlfiu-k=!(Wa!=UaP(%fXdz;pILQ%12WWg(s5n4I=JXKVj?P=?HpDO>2d(r8=
zK0**{jQ6=x{aX59(kP&PgYehZ99yAjqA=^Y*hWmGtqd&>o<WQ2Z>X_BHyt=ASCtK*
zp^gItp2A0QR#)INIphx~t-|cb<>du|!nC7-T0T&6YeJTUTICyia>1CUr%;`2a+b-2
z)h)fIay$6SOyHyE$R<KvSVO5LC~{n2K#R`@)g~rTI#}#jO$~if<fImwl%dN!QUV*4
z%`i7!TM^R(%6`FhW71k%hg%?F6C7J|>@?7}y#a&`AU{cVsCC!M&&w^<cPjf1_3Z(7
z{?hb?x+)V@a)(yLGYD-*t|9g?ia7n~g|6w8{dljfKEeiMsFoJJZvy+l#OVQH%^n*H
zLO0{21|FC7ujEkuuIIcm={iqZ6P^eh)yZ4`{MqnqTm0*`a+)I*z|l<b({GXyY)!Uj
z`YU9tnsyurs|4cSvG>pQzIO??F1McU^Y#A%Qd66XScm*XyJ=fSK`$?{@14(NuYtZ=
zo5RN#=$|YShl?8CpJ54HG0+!0{OOpdXp0w$<H;5a726uidAnvO6tgAsbkR$AK=|S1
z6z@-E;SND$^T(CG{I}d@DrXNLd)y-BH<40?Qg7{Cokaz<JLSVVrU(<UK*{rk>_sdN
z*90#ZhRTdsjx~A7NH{t2uS#6e`@pZmmU@{5nVY8Vo79Mo6qBJ9xn}sp{M*<%E2s4s
z2Lzx^%x{-O=s#)s80~#h=zg`A&6N0==Up^$N_-2NYYTJIS$Hl6X;@@sy(D0EVgXBq
zX&I8$bofcp^;yp@z04E2nL&|Hs}l&C2JliL4J5=(5+sc7MuN*&F8ZSdozw`yA7OqH
zPc_)P)UpQ*j@uk$wu9y!cT%z)$3Ik^DTv<7IuNlz6{!_@oTGHM<C7s90qUC@duhLV
z27h4TEE9hRi}**$S<n7nTPEKOwAe-{{>SoE|1cIVF7vN3!jD~~C5vodVXMjRmRE*1
zJWuU0(U`yAJ(6%T0oX0N(DdyAFvZ>tK!;rb5=ag`41Z@3O&)0I)Z0H%w>2s_0;%W<
zF1nC4irN7t2?3KTjTpvYF9y?Q{)DliL;==$58@_Jvug&TKZ~X50AfXRl>j0xSUQ%W
z`X`(XU0fNpF#?qIHYBwDU^<n_;7$SWu6zWfNT&dHzCQtN{%FnroWA*Oh64z2{@v6{
z0N_a9Rtg!yG@R3PouiHL&QDguS+)<{I?7(7xc}gI`;XI%+D!k9(raBtTPpz%#8VwL
zRayi-b$YsH(VWMZM;c<~pH+64D)xxH`;@8q-K=2ft%c9|A+qlBOb*R<-52p-uujR<
z$>l-YyF+k$HYvyGFW(=gEnZ+&YXg8y(e&#%IR#|b2(|5SvTFR-eNbpzpQg&4N8mx7
z`0L_#_$amxkD7T{9)t*k+3EV2#k)&>k`Ys33?7-Ec9u<Si8p$!7B81xdTrDEvhv$J
z$-?_YsjgJejbLlmW+A#em6^2X17xwq2#W9}eLKd$V4cd$*zn@hS7XQ#502S4v%kDA
zn!fi`<*+<r)C2f7by28cpC{Zx1MEq{u#J>IT~PLWYcF^1bak@ti-yse_$Y*6BSqd<
z)T<z0qecrqG#Ah<5FXc@e=1oCzt!KpKkwi57cS>HaCuT>h1?<JF>3BB#Q?~@h;je2
zu#+dp8~kwhHM#KnuP1o|?ZpSoGOd8^@=xNm*1B!3EY2t%trN7|CN%}3ZvLWasE7kD
zAu~8|QfK>~mr>i@!hmRxP6ux>UlP!LVw@C}APoq4lh%WxqV@WP3F=jwyB?9dPkt_a
z^(`Mppv0Nuyh<M35DnH4>udt{%8a4dh~<d}+=YIPo22=N)$vU`Ppy<@rWL~0C7!pa
zH=F9jXzxzn$QB9wtYD72OzCOHN8k<*dWia8qvVFp_@}T%T9i)!g(lLH@xFY0tXvO|
z^u#HYM-NM~I?a~gWRa{?*Qh$1sc86(!J<nhmc1S}T7rfLU8(U0F?;@zk-iQg-c5Ye
z+h;e#FMU?0zkCWC|G`8^cHIF~2{l>y3MXu|{4jlMGyO(r9=y{?3N!?CLT<&Q$o(_!
z66IrQ@*0GIC@Pi}5Arp6+`jEAdnUo=R)?~}<;M@Moi?2<qi&AEa58-<O+X<~5Yhln
zTIz!9F~);N1Srifh0@n&-r5T`TV4#(%eXa~n(CW?Sw@qsQ0(;kmA&P$hmcw+Ed;-x
zExi&*AEC=?NJ5$3x<>YcupXIeE1&3N-HlAwR>lAUvB1+v-JuE_c*_#0OW5C)ZBY!=
zTj*i6yE>H!?@IF=Y>u;P?u{M2R=`uR^7Un{=m$F+{Z?mC2ZLdTMJpxs%n4@<X;^R9
z>=)zNpsQ<Hl|>9ygB%}JBjG8{-G(|Bv8E`3c(m8nwBQtXo$Z4dGL-Ce_U=0C@rai1
z1uHkWUv0Xw`e1*i&V_iOxt;nex)buP&W)AFis-bQdhBA#RKOXc$8zHOv{^S!>fr5c
zN{gC6IM)lD82#sJDEs$$;n1Wh`-`G=iM0{dmewK$am>D@(yfndM(hwxFq2J4(54_=
zjwA}wkA?IrCX`jg@Y-d3K3)1ED7uryN5qa*T}SOvt<<w>!?s#%_QZngl6DW|ozAgs
zXaHaYi2Pu(tWUm>QKIyp+A-_si1Euqt98C4XqXu9z^}z2o$w;#eQWsRs<<{trIYc^
zpY-?3)Z5*7(mg${*S`ZDf)tqjLcd4z#w?~uQ-)%Y0`9qO<dq|;%ea<8h3v#^yQjTE
zE7OTejv?5ayjl6N<vGVM+lA!H&yHY5S7=0r8AuoLg9+2Oy;0JE)Tf3;;~mk*7vitG
zVUSiy_o$rcN9Xa_zR9YJ2pBnh>MO!gL}2um<3r-ge(P|*ZVMEY^9G4XcLG?}IdIja
zgYyQ!Fz79J6(>lWALM|pM9qDDbf>H25?(%){tFO{5vA#p&6;DwmPEJTlWaiEKw(?X
zB0(5|D{}XAlqZg9t4m-V2r4(vx85}!t9%-WoU{lgxxSK3r&TLWEmZCVDprKcqz94p
z?Xa(5?N3Of%~>Dc+!2($J-Na<BV;<aeUgV(fn_j@Sk0l-w{cj5q>=0RFoV4fbp}5c
zVkOFfZq^1OVhz!ft-@<)ADsl4w-+K4BCq>KeDubQSt4@kg*mi5EV=^!lg~|kT)(S}
zqNW(a_^Q$xx5Y49`zAu8O<c22=gUG<q0?gzkLbsa_Gk1@YPjty`q+_^8UuKn#H4!_
zMGY+a6{{{Gw%{wW(zE6E=FOqJr`gk^d?o=VqXnCjYcD;yLT!cmHz!`&a#B2_3%#Sy
zbw?^jXIcslgWOxU3F{&8U)S5e-nu;Y<*L#kG+`h<W~grrv#aR-ZT22HyS@VN9|%kH
zB%@y?CpE<$p@Rp@F=C5g-pzXdmb+HEA4iDat$R>#nYt0g6x`e{<VCX5*=u>_pm&W0
zN*-`(Cxq&dl*<c>&=APiC3m7Y#}`~(9}PVnJIV9PQKE}GNbj9sJ6qz&*%f=$HaQY_
zc&<4)U<HTqP~_ktFSWtROZBWe$-19PqAcpm3V8(W<U6Iwj&`T6mB+T~%=J~#*vQA~
z+>4jaZW!`o$2I!vjCh0|kINvI&s2ImEE9X^^ZofTYwkqWh;EVniLdC4gTbL>+aFBv
z75fwVPj>yZ=@V-*KbR^RHnEyift8a~(Df?X0{InPj;#L#$=?PnK8qwBMwbqM<;FpV
zm(?4O^}tOW&$Tp@55u#eY?}`0R3RjoGDEj1p0~`2_|#lF3Nt%0Zcz1%^b=aj@m1RO
z-N)jr#XcA?RpoQNt1ZQ`4I!p}pfx_#=51^Zi!1TcQ^+BXTu4Q*1g?*TkJQeb-kOBb
z^=MpK^Z=SK!k)_9m?i+_lQ?p9HM$gjEH`e^+^S3QiYhy|dkCaWNX|*yv)pwFfjW`Y
zN*saqO}ud3eo5Hs5^oX{#Znlxo}}6?v*u)n1ldeGj=Hio_Wqag52_-}cTQ?ra7rJm
z0144Sw0D#B(x%jaxmctmc_hyRqi$JR9(K1MlQ<54HTJkX^k^EZp2Ur4{w+kIgw*ZG
z32v?9MX}GzWk4iUC|x1_J*Zh(ZQAl0KstE;WscCxIN2p1iSJ#pb>c35gVA!Iaq`r*
z7$lhN3TjD*@sFSK&FR<B{4mz}QQW3oC~_rktdpH3Ax8hrL-T7Qok`E`-_3=q|6o$s
zueRG8CoC0wGp_$tT#%l;4*a;zq%Q%c)R{jdJY@c5i2grga{i{2Ie|Fz)|^HhBH@?k
za%dUbLM}y{Xb!8f6rU@zcT9(Gx4xTlSLSo9J<hbLuO^(U<XaKj031iOU_(Q*<_SK}
z72ald$#>gej&oUlE~(9^{OI#RI;j^Y>%O+Uk~rFMTrp6H77rarCu~AawsFxVEDx`J
z&#5mD$-oXgd~9*LHvAYOy|=M+^~~d#`z#?xo@*-qwOH(L9OSy4?I<8xKiLx`c9~Ev
zvfZ~8seX2Q=uOM1XtqS7s!I`H^m?cl#RgdO<_FAzeL5hg=*Q4aj5A4X#N60A1^0sC
z1+~@`vNRh@n&4`&I5vVzu#)x<D2%%u(s%#08jpY-`+F44W?M`w@}5%$RX0jWz2~I0
zlW<`8Ddmf*RT1J6uSZIp>`oNA8Jyn_S4q7*dZ#7bf5ex;38u%b58<A#R-g#cz-%tu
zmXhuQ>aOGOQpSTH=1Vq#g`GQP;-TZj?=6qc_Z#_THCmNhNRV)P>{%;Q`Kwb~kk}nK
z)m5jQvYk@sGjfAa2NJ8zh53mVvPrZQGi`hwk~^-WQ3*}dH&boQbGZl|H(EnA=rc5m
z$J0J+ngv254FH;Fv#m|d*$8ShcnV5C&$vu81BVs6$au*3IQK4E2F%o)Yc?UaGGD)Z
z`j_d5yNeB0Lad2`N|~eyqO=~BzdmC`m$VYt)3mqMM084%s)-nd)ueXBknGqL?z|ve
zyshym<0>Q}W^<G=B)4^-^?r{n8#whBJXZ#`oPQPH;WPU0p<yRyz$fb-V#a(n8u6SO
zl6mKY3ziaxEVH+(2<>j?V$e^%WrecD?>KnJk<Z;&oDcSw*ZL{VoeYW6K<sPK%zjR0
zi}ojv*5tnO*ymO{z$G1SDKvrbLB#CfSHI)P;_943uH(wx0H6tRZqB=NcVYd(CO)oq
zy%ayQn>p@;5ulp2Eo4DOoGj%+HEy~V`=IpW7WePQsVl&<Zxg_<u&;Ki8-np=PYmR~
z=7P_k+Nwlx3jmh=5k<5h92Y<mzMnldTp6v=E_N%{jV-~}Vk71-+xRTjAW&n19!Y08
z7`NKOlUz4+y|sP4SB@1eCF_Q2sJoI5yq220lQH3nT~anWts=gTF+#b=&NsO6m8?qn
zm7X%?b=Fd#fvE}&ex^m_eL!In(t<T!u=T32hjFIoLS^cPxNF*B?A#ihuk}KNWfn)w
z_3kCG1gs!eHcC<XyD_#)*aU_dYMP;o1)#4?|G(*@c5|>A66#=iKw-)P0n0E&!)6Zh
z563$^uI^T0j+zX-Q4R=3vQUN&*obC~L)rkEV89Yek4eTK1TO>H(Ds8Kiig60pq6AP
ze%`hZPdP5fI3G*9t^okAZ)^i*_j4l@RjFE==^>EiJx0Rp36x{#KSct^VkYARgah@h
zJ{vX32c;Xj7+n)YQG$HnjF1&+pzbW=&+O_id7Ge{axf_EFo40~wR(H^2#9_<O&DP8
z^uG-R$uHn($>6_l9$|*>j|TnU>Y-CFzIluZv?e=ld%cH0oL$%nbR2$QzC$*6;^-vQ
z3Ymptzo7>H%*zi?90X5$^nsT_046h22A=oVJ5J1Z{({D-%jj%S#`Az3WmFn>;yd9-
zibw_iH{FkDd||V?U}2)+qSM5_V`Pi5H6sDYNCqqANTg#PP0pGldx2sFk>L4WlNh69
zD$R(miVBWVSh*&4ol4jt>&9pS#D<Nk7$1;jl5zj)*?@<2G6BL?@XAJ2zKYD{#9Ido
zA;e=3Ii7c9f1ZFGMA0RSe4y8<N2|U8bURS4pD6)hx2fkQ(<Xmr$074sPAuyj;SNfD
zE2%WrzJJ#TdBub@!TZxy@cdJN%yIF>4<;Cf3rz21oCj^sZ~;}W7&Y??o%|)%-3yjP
zR1>`xZ3NNEgPc!;EDv4E4usdPrVY*D5@#z`(pOFY!V_!8O&w_vMmNX4)VfINj@o{+
zSd)+*I*iG7nw+TOu1^qs>oy<_va|L;Rp^FgMw>CXdw4}ZaM3n_#TaXXC*{+D9L<f1
za6E+v#L%=2UuR@Tt{p~RqhUVIfAU%F1k!wz?u3um^9{1+E#K~y*A#xkzWb>i5M^pS
zI+bAeIzKD)#s`<^o2-pgfd<jQsm4HYAbMfTLoHOpM_`mn?28tYltGRW+6Q01jkX}9
zo7_d1Zl0ZfobX1k2ugFtbD_Rr$*y#kUz=m}gNeh8w6@1jHv~3>*S~g>bwTcyH$Sj}
zB<yjnVYV`%`|cVW&G^!_Xr-k9u^wA30a_-ztx%ZGMxCdpn;U<;U(rQ>T9L4j@v5nX
zxQRqGd)Bo>@jKBmkxi(hf#>PU19fb9G+zzrB3$HTB$N%hP?MPH?hp?Vkrv8$FOw}5
zSw1PaK2kj#;|KchTNmI`J3)z~=g{e!=M7Q<2|t+nP`f8-JNBR>jGmhAaJd(fgP!W5
zHx@e(?susypIpQ3YX&;?PgmOH?l`Y@sGRW1kB>8SM;`*v3?-DXdz6`|2@Wm(dVE+@
z3u{av;szXRlrMi{jcb@G=Gk%IxX7f@o3Aqi1xkRD$^y#H0bAgCU~|($0hRQW!9X!!
zw{vZtpdI^lF6ETR$1Uw4c05FO64Gk0J8L^zn)wmbf1t~+`QJZD#{c$xnG9h})5N9>
z4FDb;I`gMu-<5l25kS=IKFyyIg6Bm|N`UEpvA^lGyQ~#Pw!l99+i?Z|X5w8F8`So}
z(3PNJ_<1X;3gagrf715xcc01hKcQ{aR1?xT9zd5i*{V`(y3wG>a>2RMeg03~iIxe`
zO}8Aeapr?*{A28Q;_|ine!0*PBd_620)<0)v_OU4MDGv}6c21~Hl@HER=HQvHJI87
z_vBlS8%r@CV^0yfZsj%Hu)Fa>;J@w6<J5{8b&RAx!*2w!IUrWM)w^3@IpO`ap0#n}
zgw0LrGk9A+VwDbq^H|+H+#J{Ddb&SuxT!i^-{@@WW$T0nB3H+0YTI%zDK-LnY!kxn
zgSqw1Q^xV<zR42oAf~(uliZtN$0PXjbNSFKQ$f{&J=xd)`-z~hEj{e&+PDoYrC*%;
z@CFQJ{B$Jr|FG9=qVl`hjr#j}oiEe2hVO0<)c13#;MHH6aZMPZk8F+D5kF#1{%)53
zZZT#l5DR|GCuSS)`_5>gVpqg7Xz>lnj5i3<8n>?dN7xR3i6Wtdg`HX&)O`Dnyv|b&
zCNFfED?Q!{l<imH>%F`Lm&t8PxiA}@`Bc?ZP-#)|w_IX}I@%Z+pdEo-GwF1gr85n5
zhv11Yxf^sPSJ&Iy*R9C9MVRY?dM!v%>f<y}k6i;SyMLo%*+79%?FF<jM)OGQ!CB{y
z@-ek%$@3{n*<3K=%K?_!9O_dl4IFlND?fb#3XMiamhwymWs2w?5{VXq&p$@UKKrWA
zbg2D&4I7w;t~!7!!?jPPSYcK~+7<WZBwL<=iN%>em@HV#(Q|M7GiJivZ1wj!7oaDG
zwtZ+u?s5?t@{E&=7VN_u$<S5){f_Tm7$R=m<muH;U%{*U892XQ+(jQWnNYp21`m`T
zz*lTt)w-}`dZlaIC%yS{o_YTh=gb$aZt<hZ<JozG+YlTsl^V+!!*1?oN=n%d0WkI5
zOMrV-`Gd&@d;?G7`1FGb9c>06(F9O*uM!QgTrkoipuoZ@O;W9E_6>O0QzQV~9N%;9
z?XY;~JL5-o$O2?haKk1h9LctCiiA;<!j_S4yrUxqa#5TKZlsmlb4>(GKCcWr?lLaT
zX20ik%(I7a8&Acht``fCV#21l!cc+?Fw%pfZVTt7xwh>^1C)jrg$^qjo~n*BeTC>j
z^MbDUPG1zuUoY~QUM@mGf`-<21}6+{zYvYwYAoB;{pj9Ubk)j^Cs?OWohwA+A>7UL
z11%KcLL4WpoKLw!J=xNzEq_YmXX8iAd5gJk#w$?2`AobT*a06EPo$p&m^VnmREKKP
zK1L9!N60Ge*8rzmPB_<;Kh~8@;7f|#wt93jSfN8?L{QNua=Vh&&J~8#CGCR+Y+-=M
zPnEk(`Km_9&RH;7Jm%=lK0|&l(?nFrY{e_u{vz@0VvFjt0_TCi-Dj|Mcc@&##;j%S
zf^R_Y+`(xEN<9!T^czJU0))`Q7zQUmg$C%sQ9{3?dRhyBw0R_+p$<jc%F_0!WuGlw
zTSZ+N{O<`ESIS1Ka9|kXp>O&^eoYi_N2DZC>?&Vy7?WV79joe>5j?qvh?+WrB}1ak
zEJeF!gOFT>#qV=qYqBp~Gkx`%c9K(*Us7!jr-Bkl7d~*8m$2Mx$<sDO!0AN-s2ZnQ
zB}p+XwxBL)jz6Fa@LUjKbh<O1cC{8b{K`?kX|5%Okm+9N<9-`ohcz&$EzrwN9#=KJ
zA3n@1A7*_2xkU8m*KV&3VfQ;^Wm-1R${PbjOTg#;<|V)RW~;aN-u?p*{p#wRsRV6M
z^8tJpdF@wuzu&aB;@|wL|FVbw>@)rGqW@|Z{?SK&^3VU6NB{e4ja6lGKis8GJi|~i
zR<Y0g^O{7I#k(kHeTglJULzWaajTx0(EwOMdooTRj&0(1MU_zdtN^lTzB7Lwp4JBx
z8hH!?HR~Wpz)6-ZNdK3MS~&gh0U+*vJwRzSV2%-0QJi}nAhH1fP&TOT`^8rO#mx^K
zY8Nv84E}ojEsOg0&CN5599j`fodCMU(Lm!M6tq7Hcyvc`Ky*jz@(-qopDHeVV8_$6
zH!-`^Fb0he)hL&Z>WKpHd@j0ufmx2h9U#p>0?Q9~7(|I&fNo529Z;l=|NKHI5XC_Y
zqSjvM($DUh(P==y5P5tc<GDO=Y6<k==UV^xXaCQ1<5_G)MBkNDR|=ifq@s=|M*NOk
z_OBX8g_vXCt(@_k9&qUqC)Z1>9v>g3jgy$$q!guugdpr$JV!N7!2e3{QK})AJ-meC
z@Q7A7Di8X>1gY(N7(^y4D#cyQbh=ThcKOWPpzACTg7p^*m8wh9=lxtmJt7UIb^BL5
z-ph7h;i!m?Vq-t?Vg6&pPanmeKl-%H;PVl`!|2|Z46Xyi58nHpPq&|5cq%#Uk_A|;
z8s?BVoz<mK%`Ts|&j45h#Yxws#*)Xj6DH-@VM#Sw)o`&CcgIYtmU9ufc&F}9r5~<@
z%IoP_Kdu#<t%?l-3u_os!X~jnBF)NsrUj%*ny)XJjm<`O+-dC$qbFwcmzjht`=l<5
zhK|;hqix+<&fcbTt{Eu5XLME9rFB`w(fTiYRzr}1y@xms+9CQQ4a@N(lo6jo{W%Vg
z1%4n^v2i7u5!kg~((0Z({0#~E;A4xpgAf43s=_M_66^4BjeS>zSH1T8I__~@>G66J
zzhrAjaRmzgk#6KAgt$qcCtHei(Zvep{8(0J;<@_N4r*+f!g%p;XJJUfH_TpQ2fay1
z>k@5js)(y~TrO@L%!O(cC!u+r_D&H>E$<BNxJ{gU?PzhU@hVZJc2q$mjM?QxbQcoy
z&IB-0+AlD?0I+4Z9MLBCz>1EZ_wzvi#1Jtgxsh^d0VFfZ9MN!Y`=(Q6Yk!aU5vZg?
z*EEV>Va*7JX9^T_BAo4n(A-EOB3Ham(y*1^itJ)nt{mZpRM;_pz1CH+S3DJ@U~530
zAl__F6Qs`6U9oW@wuVcFBg$F(%X>JTv1~k@K9$+igs{D{KS%8+k7bfx#QI)+^c|;!
zG#;q1(OeN}JOQi0M{Pk^Et-_k9vfB%?=kG&aaXGjQ7T^wo-}h)<;n3Xys;6H>c>c}
zq))kwn|kGGx6;SJ9fpkKW@~*rX50H~A!5e{R<`^Bl5r=2r*MIb{(0J3Nz6r?^Vjtq
z4qp+^V+wup=0)&bpUBA}U`G@53F7d#rKp5%)jykf5N0yq5$PZK%l4_9VH-A%+&95(
z&QyJrGe;nyr^!-trp}EPvvGIm62?<gx^XYD#5G!e*-D|!&bdc^H+#x8`9=x-a~V(d
zCPHLNn@m46$T+iv6CF@osuyczEKdDgU7FTv(USL;RWJ9t1YEvV;_gxc50iV46V0x>
z(h|U4K$D)%#3!AraG6kF*_rsV<t#k|jhZ-tlUM+AAni#j?k1GQ#hSJkN(g_*K)E{G
zFCot>^<HkTa`Y=k8BhCqOZ7hQjOQlZ!M^fv<wj~!hL-o!ec<<ck($nXbT49bWqg(a
z9OfxvZu9lar|d7pTsxk^K8G7|psM?CF@n|LbF^~CEG|H?V#cCH-_sa%AzQ8moMf_a
zTY3uhFRY%e*kF_h^5#^mfE7$QE&ikRuno-cqv$;!b<<v%<yo_+fOlzIUfnj2FBONN
z1}e>F1qYkP-E2|8aX*+wpqqoELuYG87$Wqc&3CJOX5C_1v@xo{O{!!}PA{2K>m(*f
z<$RMN^M0$Sy{3NO!|SHodwVvn#2x>#nf7{}(K8~+Lv&fpso`-bP1%cbFcgg<T?%V5
zt@-MQq%7pe!3sY5u03%P>b&&m)_U`l|KT%ld04bnC5Eg5HSqK}M!ZO*J}w-(YKZTo
z$EG`B4u?&}hHt70gHGsX-cDwXETSurWpuj7%|FCR7dWL67Z1ar?OKSzQVo~B2mI<a
z!gI1<r^pfY0M<5Pd{}u(rx0C*9LsI4@9NyR>#<}ZV>!rU+=cB4iPOJaCD}RoWY(pq
zt8WNcr@nX@)a3ltM&r03OZ*0W1Dd20pswwMYQRV1&QRMTi^`)R`UA30ipz5bIP*0`
z3!D<S-KAT#zV`UlI?9&vo87=KT1juP``QF83WHaS$c9j}G79FEtQ^u6{=w3Abl&wK
zJ!j<o;o`Seqe@M?K1|YwuL-3&ejYz*jS+Ah##eu2xNd!aMKpSw*$buMXEqk>mVr4S
z7lHyQfsX<dCy2=dF(O$#63PtQtHW~B4DJoRT)7x`mBjW4{8WKCUObL-^5h+1+$Rbs
zTo(Co@;$3fU7yP)Ti5MYk-W_U8c}t5+JpUDO=%u`R)Q*X?sd!C%Fim>N2u^kRKMA?
z9s3Rd-1i(k4UBxPHlOLE{}dPY7jx)e{8dFU16&75A5h=5CVnt!7xDT0KV}C0i{Jgv
z_UGTwE@n?ccDaAeU_4kt9q?6n{e|l*&xe<B3&n-{uJ&tB$}kyZ(IB_KZP|DmmsG6F
zd3L_68928gTDQQ%7h=Wox++G(`U3MzkFnt+7>czHlu4uo5=U;uE)zN5X+`HYDVE-T
zWARm-r?F%C){B={f}n1@e(UbTM!E(lE30*cALlZO&Ow1TLODJ4l|*@3sjk8k%Wj<?
zos+yL9I&GzU$6o^RY8|R4o0oaW^A>LBK=|;(Z?4jMo{oR`ov1njsE-iA5400eTPhO
zkT;SX>nr<;V5v#b)ZqBE$_Q_7{1$-X(G2k9+40io+uwGl@MjmlH=QyZ5HrcsEo4&~
zk|Q*@8r!E@%lwjht<*RF)Tb!^bK%{y-CKCz{}54*yOy>+pWrHDXMk9j-rHq7x_cvM
zrh07q9nK^{gvQJ)ir|y;l&C=gcFeh*GnHd&LkzB6eErT<8Iwt9&=%s#6R2TLx%rkz
zk*Rc%`aM9TE_8NQCyq2#9eOYMinzMV{ij){?3j+vBf*maY>f=Q*d|M_ckGtJw<sy9
z3Q}K+_2a+F>c6gbxYVuxs5wp_Nc0+(yDc1?@tj@sZH?d8@L|*G^DI&pJ-bNm2UK$P
zs;v%8*H`*l8aUu=UD8T*SHX*<3z<yMj|$GyA(XZ2)R35HSRU~!)oI5KYOYi<z1=s`
zC0r_f>5Uq8%-i@_ll0{W*U#46zF6;p8iMv1Sy)a0IvDj5F)9W$O<=niYT70cB~T9E
zJOHeqmQZ%wy@>9If{FZQk0jX7MDgs$b`Hh=g8XO3G=}`^KtXA!4V1sg4pvo0mjns%
zbX~z$ECjl46VhS%z)Z}#t+7-9mHZMPy6bh23`%&)*?qnJtMFVy&rzQUO;6YI9B%^V
z;Xv_R=&Fyy{=F-@d?qbYSA8Vzg)=!th#gzh*UaXt!xy8caWmh&{ZFaVBqyu2L>I@8
zptS|k0F~tHA4kS*3^j(L&#Sa6HB~3;bF!`pbw(Eie{N01b0DuM5#FfJZ!deiR|`1X
zAzs@a(f5$2lPl?cRN?7Y$j_6qN95i6dJQ`FhB56G^WPJbF^l^Lb6k_IK@1+6d%ph1
z`9zgSAHokNyBM3B_iAzbO1Q~!?W(GTX70ywt7IZxin<<UnL+}GxM<win@%=eF{-#n
zhyox3Zyl&I9Lb&7!12@Pl&4&gF0(a@7wf$l75F*KM>((UB_I~QD~3SrtI>`F3$|A=
zJ$wAS<vamAoQTL?5~x&&yaVo?PCgc=|7cP)PT``)l+y4#PM{yDX5{fjtb1Ht@5!UG
z7~jK%ms_fIN*ZcAqn<bO-#`5PuKw{VLdN@X9BOE>$=mw?WqMtk&-fn|v-oG9V@{%g
z>EV8?2f-^q#?rD^_XXgl8+-#&*{%TVAt8bZZSbqF4TP9K1)v#GHN~OTD{>57Lnz?s
zqxR$%YlMF%9`YYi(ELZ&<|-h&qKdKmtmCNpPRqr=(chT)Pwz1G|3<y-pZ(r{_W8g1
zb@{d|Ku3U@Cbl`T0ExMQncZK=bCX_xr*(nMW^fVEgT~*E%={yWr~kvxoWlQR95sLh
z>sedi;)8wE0e|^jmc746x+n#I1SrV|XDu+Bwd%tOUdLo*`gGXh+~U+8qq~79Z6g<{
zXO9Vnq8$bSV3lj4+9!dS$=_eprJrV;0{FL&B>f7F0W#r=hX#zS@xQx@`ujio<6Zwl
z`iplFy{-lWZFBvqJ(O6{=d|<Ms!gp^(G8oDs{TB1aY!?l=B?j%9#*JWtA6pFmypa#
zpRzQNV^Dj{)lu#ZvSgZI^Z}48^heYhP^TmeE-wPlErRxG;MPyxAQPRbMJHze{oDPF
z!1NzX_+UC9EBiN3H;e`tip1u9;8<IchV1TzF&JFb%)faW&g!<M=%SVLTm}E{?~VQU
zXA4E!270UZl9|T;?zMk3;*a6@V@^>2x;h@C7SH@(aziB;QzYrH7yy3flP8`LVTq6P
zBS`*ktN-Hw_~XF*!zlQRIrN9&^JnYvf107l`Wst&DCp?Q!x%w3^YrK&5^f*w-)ClG
NQe(RCf$PWc{{rTTu?heH

literal 0
HcmV?d00001

diff --git a/docs/_static/bonsai-filecapture.jpg b/docs/_static/bonsai-filecapture.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..7a809d67ac10f043046af7dc5960e00b5640a2d0
GIT binary patch
literal 68990
zcmeFZcU%))w?7(+qSCv7fYKEKr7A5dT|@+>mplq65FsMH2C34UfP#QhrABJ#kuFU-
z2nZpe_XGq2gyhb1&i$P8`kZp#-+Ruz_g74^|Cq_l-fOQ}v-T?AwaMSfb0DU>nmU>w
z3JMC47w`p<r$K5UDoV;ff53+t_?<d)>J&BgsdF?mr_a!xqoboeM@vi3z`{h&z|25P
z%f!LN%*w{j&Q8b3d7gvqJPR8;+n=3KPyzQ)pE`T$)LAxqT6(ts;Sc!>i1`d<Gxa<b
zg#d_>nSzR$g4_<`2Z1O~0o49U@c;dxpaf_<O>^e#Ia=U?Y9<gR1r-$~HPxR~16TV4
z|AVNRPqAFOe)}}5!DAW$7dGiv2_Mf0-YIWkHypwV$vk-(boLwv=Xow}VG&WW%U5LO
z<P~n*R8&*f(A3h_x%=>uk+F%XnYoRv-P33G4vwyF?jD|A-af%0p<&^#BO()%-Xy1_
zzI~ULos*lFU+}50sG_o}x&~QWSKr#!-qG3hwYz6{WOQtN;``(j20Op7xU{^ox`yA`
z{kgY)KsY@5b6yl6s(+r=zs&3}^I`_(MM+IfMNRYPyeKF=fkwqled^Nn(=4|QXdb(;
z3P`^?!*(a(V|mM2K^a3F`;(VL=QxC9F~ayir}j@X`}-yq^bgJKUncgi^O^$DQBeR7
zkBS)t1|1)FWU*8YoL)1Sv~KadTA?mN?;L63a8>E^@}gW68AOY|MFwT5=#oyf$e@l1
z*5f<YB3W3mtvoaH8*W2_SDBS7nE3A8$)@b{XTL4_;dG-^8yUpur-Y4h_*tqVSJM=$
zf5oU4xBt;OzU~&r*88GSbn?ZRcH3)R2AW%66)lg7Nu15E$RJgp)RUt)GAJtNFUN>8
zAKuetzHut1njwQQzAOB@%Y|_CCo+go5B~M=u(+*Ragb;xbU6>=UzlGWZg_i9`yDbl
z!v_9N>md(*)Fxh9+$SULhjN@J85Fwx0X~qKM+ODtkU`j$zn)lc70ab*AcI;D2Eoe%
zhGY=RfDFRB!GAq=$FM#4ebt|w`TvO%!>X)=<b1DQ!^EB*+S<HW!RyeF2MxCkCgZ@@
z%BqcZ6uezzLT?B$-CX}M5#o1-3`(fqB{f3GAWbdEZ3p;E_|YmER8a~t-jM_^0|K@U
z2n4rsRNUbK{CE}KnU2-xbl_u(yPss*$YHR?Ik;N$?Q0(&_F;{z@a^eZ+}SVPRLYS(
z)r3-v?f{2ZR@;Mh=4K=&-QIZ*3X$lVe3DF?!y#VXY>>#7_@V6959zWJ^j>zuaY}P=
zWfQ8@Ds}ef-r&T)KXwKN&gw8LiyKcytn5~gCl#4@_&O-5<GC&+sO3C=!Syot!Zm77
zwq_B$NXFbY{aXD#!ew7iRJS^1e{0?boJ?nLHSHMGbuM>~^8O8W!}}724v)3aNJ%8u
zf+c(a5N)(BN!FW_e|L)vIv6H{VvO-7ak5gUyzatSp$cf*p(oluzJ2Z%RCwEyP}CV|
zHS=!Fh+UWE3rHsy2ac14yq;nnm*t7rx~kDW9=D%<vs5_N-qwCG2K$bAEP3orn8DZV
z0i*AVm)-0T0`nw)H0cO}&-q#lXcOS=YfdFX%_DEwG`{R?$mG!Y$`kQ$biO<r;RwMm
zp|?dzAy;BRYz7+taC=IKvPR;*#kmw(2_@b>J=1&URP55Z>xndmN)zmV)oSCY4%DG&
zJ>sN-3>qqg@I(1I8|6N(O~q!sJSstQyBIb4sxTa_@w^-=3HBShGZdGQ9BUEi-EnEM
zDXs|hs}lN)PQvl<tos{w&hC8k+H#T<G<%#TLYF8=S4^oGbVg3acuLxHjBop|^bGy=
z#6*aJ_Ot(IWwhJB)MR6$oU#pm=g!@}4Rwx@?Mb|;?#fNsLVNnK{8e?t#mfRGAYT+w
z%Abnlw<mSH#}6oLtzQqVnKgc2@|!RJEE~Ut;rC1OFN)&-SYhyo=A*lc9BPo|r%lVA
z=p<NzT1+WrVu?eKlzGV@f!J^|$od+5r8+xMId~pLj2VaTVY7~;?;b9HQ3{W37Igw4
z_p56tZg7zXsu3TXmB^s4GtYo~E|WnyLWmnVs5B%N_0>}qPVmL*l<gtT@aQu^)&j0K
zBtFJAabrn3hF{MJKSe(`E&XKHt{rNm5KJ`d=V$cOkyxt6Fjjg6KNxMa@)T<8NK@(f
znV6UOTKMVA^FHHIp;n<gMd>Z1t4qy++4;Kf<j_YWUm+7je#lx^xAnOOTu0+951$C$
zY9j0jjLPBZq?5aY?0#*$u6wlX^HM=ZOq`Ise$V&w8+Lf1-P<1+Bls?jHj2eyilI?Y
zBaigt%H#wf1N4g(-ie=>5-7eZO)sd>48AeCzk|vkgZ8&~iJZwr`f2yAHVGEH$oT_e
zZ?RW#yI!H!Q#~!qv8rj25w~vT7oTt5F3@g2e>eWibX95c+lyG`>(iqrauW#}?0V<|
zW73X9DKT{pFb1?+BgH;Y{vC}`i@;0<3A!(zc%ewqvCTolqy;i)Z-fl8Ge10-(;~)9
zWbI+epd)F;BU;_#Br<41|0LO-D)=NvS-~9}7eo>X*S~XA9eufYbuI6ef^nko=tcHr
zVT?f+A|QDO8Ie1RP~}|#Gs%A6PT8J6)miYNDuD(~r&yIxmM-J+hS_3Tvu?V@b>t?Z
z@lzgK5^C=mUh3G}43bwLgS1W63xL}ZGte2tB-Lkxjv{+i9zqmf_&oUhnl%V#lp6LC
zMqiU&(x00*`$0Ha<Q87hKihO(*D9sY%2d0n*Ll`^UKhbeXlMzLufu14EU|Fh2cMGI
z6VKq4dRkr09fDiaQBJZDydo3-f!+O*_60j8#I&11@%?QRcvm_Zlt>13o7h1MgzxL*
z<!1$lFZYl^IMHQzh@TZh4i+3<PX>iI3qiTFz1AdPZ3>^iuPS!TWTu}ai{v@6G1`_Z
z`v$!e*5Q2>f3IjEP-1`6-X}i`bF3nxi-$zYG13rak~#vUxWmf%>Rj@j8{-~6enxi`
zq@Ac)lD}B@#ygHXNMCI~k3A=OJO6-~-FzYiU-I0MzB@Xm+7E6i_JQu0duOpYW{ML+
zW;t+888vlS7S)TwWKf7~hEYMX=qpY`8@o^QRJL7u3etx^IClhu5;c7JHxSruCgR<4
zJY!__QZv#HYDM98xCj?fiV~}_{+fUURsTpm14-<qbPW`r6)TKd)Kh_DwvT$p*AE{o
zZUgUQ8{AxTrsSx`XcRTykDAp#Ib+1#R0uE(-k5d7R3sZHmK_QDN83C}0${qz7<@C8
z6myrLq%-XIu*mNY;T>N@*{Y9T)SB^7AJXSu>ccHX{~EIkmYT&a=jmd;c7Me-;|WMI
zNa(<b461<~$%!v>8{a=zD?NlK=uF$RNmlWywQ?^Za=je4AJ)R@2qNECrr=XQG=#q@
zBoueE#3$CBX_SuFrr|oL$-W$cK@KY4CDGxZ7M7WbhRD%2-Cwi14=t!bJ<`7z{?rC`
zy>BwqgW@x94rOzE37XZOgCvfdEzY)TET394u9)n>#q%1fi*@KwQw6!i2fVqv)3A=1
z?<c{7b}FOpy1?vUZK=epSyBL3E+NR*)gyqH5cdEd^A$I?ae#IhwH-B-&`^K+op&*X
z%Pjg!Yr}2Y9y%)1dRWX}`T!lF-b%^3+}^2C*jJ@zbym1>%=wSTN2f1y<szfawZYzR
z+o&?{5a03ZO-+}URz0{REO;vrX-N^8SY(deJ%JD4tc|~=yBIXDp;b`5_&rpgwOCbf
zOXkfq3oDpWlu^TQ6&0%$)I?hM1;4St{J9<8-s(5(RE`m6DAeMeyXN&-b?__}TzkCI
zZB~4O*ShRnu(YTuoj*idLq}L9b8PKz?jrwSZOV9Y>*!9Tqp`yw@}P<g((hW}x9cj)
z3UMNXsBuz^UJ;wkH?^J|5$iWQ#!lAcno{a+q>NuI>q#}YjI@aqnkuU7d{p<{o8`js
zCYxb%ga^;K-oZJs?D$h`2DkoaYWxqqlt>!9jHJdt0FMZ&!N&kN&RS|@5aK-er}@w=
zPC0t1HW<{u*ee#TXQ4#vdUayzwMOijM=~<XF|FBwIzn?Vp)rF2m;6BZdlN49&5{J)
z;$8<Q;mvjr?xr`J^o=&s_b`^58W&QODMm-Lsu+psaUFf|!NNlLs?b;*!jYd%m7hq7
zN4&1q=k#(WIVrT7nADpMHOk4oshMD&mgy3#nm}H@$P=)gLZRX{SG+2+55d8ENoR|~
zWeV6E(;=QxWDuY>C^8cYyiC5NwI9Trn%bvUn$>r?^450<-s+1ID1LRIqooPjGk!Q2
znPiOf43g!QUw$z9v!DO|=NdJk0~3b}2I@J*GBuQOFQH2+7TECDL{=>9f_#L-Cu@jT
zT$j0gs*6i*)Fn@rZp^WnPWWbQ3RcT2<@!<vs&S5EioSQ7%PYJi(Mz83wfvLm)+zmV
zyYgljla)i_H^9Yv)|hhV@cGlk*_*~k+D0ae-5i@$9Lx8OXHrw=dy()6GY7aJPnnVE
zxH38<s}}<$4g}!*Q`@9o`)KVWLiNM^^zyc^`!{H%(AFwM+j(~0U9fe@+^Zvl&ds-?
ze@X#y4FUrSW9@{O>fmF+1$>;})Kwc78f^63cM_N|YN{MrY_?;yl$F=M^wh%j#gYy)
zSFDMf(?$DHD(qBN=XOLU|1e2-MFu*x6#8jp8xwK0TA$n8zS5Bcur|`BTm@D)oYJHu
zq(3GL@nnf^U3=Z^X}3sdfbBMjC+B&hiKxD#UJ`p(C~5@)EW9{gcxxQNgpY#^`Uux2
zgTA-}bRO!@hjvApcXhxS5vl@?Q$$6pgAh6O3Hto^sWm%4)+t45;<cV%YCSZ#|1KJQ
z@w%H3-@H0u2z%3_#W;x&b`v4T7?HLdUZ55ELq{d_(`aMue#rUh1&hTur{`>mT_W$x
ztl_Cr>r!8wmCIlHgAX~SI;1*8$7cc7{K6wGqIeuFd=OPY2F)n~fj~bb1{DoqL4P5G
zPGQpx@Zcyp2(Lq`V*lb&Fzn;>Oe<c^*>HTunV9|}yf0C6X6den{13@H{<Jv8j;ZZf
z)$^-I1rZlx=`A)njfcHG<tyK@h(`#-%#_-tFke=GD3BA6Xwtm?%2`rDzksN)TtKAP
zNv+?20l~&v9P)(r2u`9x5P+cCFGN}ivx8jWxH4s6ExAL=B7*|gNo?@Hy~9LD<M$PR
z!@)oK@lx6gqCMzNopgW1>3OF1JF^d=j}j6(Z>^Hp5!tO3YBz<*pwce>tx_3FGAQmV
zcs(6{{ox9DO%lA0AoWyHcy1J~jru36nCE8)UOK+wQ_XLuZ(R84!<m4d%u=n%hhs&K
zVwC(91;qyYkM<nXB0<G21zj%8G4DoM7#>CATvv@ypP!IHnY(bnFVJ%sy2X|C{W^TW
z8-+)cMB+Biqm~=d1Rwy!YVP*n^s{|B$nGqR2*k%Ca*j_+9d0X-0NbQ}JcEtm-AOPf
z@c^*j0EYu>Pil!G`Sz1Ru|VWiu0R9y6#hpC8KkmA26e#+Y2e>O{&z9fdm^ri0FyN?
ze=l3J3o*Yfuyw_Ze)O)U=Jj~}yErS(t^j&GJW}~_wwKA=c1+^~68jiLYt(rIv2iiP
z@*;J}v1T4?#awVk)A|aS2i{c=|8Cez2I&eQQsK`~lPA@P2o&EQnB9ypa$6V8{X@S%
z#}45*C8UtDB-j3_X16`*@=d;{Joz#mcN+?w-<J}%*^kN_#2}@^L_l!WJ)4zBmAP<T
zUji=T!y`ysgsi(mZhF!xnwb8*s&9+0>7Ao^cb`6z*kyf92Gws`2V=;fFzYjTk>LEb
z@hRnQC%6;usIJ71me%^y6oK*4bJ25Y1i*7kAy{BHEwD73*F};rR=y9b#+0e@zON=t
ze;;@um9JysW}rH<^cB<%O4)*D5mIr8PFLdHQD?trn1&`;YpONP$cCS?2#$9<5?;Ar
zVQcZ}aqJY9<8y;_P^-y$*S7|jyDUo$$b#-s<HMcK*(U4l)JKRD^Q^_b?Unu2J^l^}
z3rs;9Bb%3JYg?=NawHLgIE;Jz!OKT0W%1)Z@1zD@Wfu*WM2ctb3;6dJObK9bT-&^|
z4^a^oUhTwVjyqW$*^6N`_x{*ep3e0EW9Rlq@2<FDMZ(}j6an1ACqod$yJAG!;VcfM
ztN5A((N~Nidr(ZS%6MshpXjl=^bd23ZkJ0xYGb0EB+s*SrLB?_ypB5{T&iaYiDo3~
zb(5nDiYKKtSp{pmvu!T^_B_j|+7YKQNp801@a)9?iwmEoBZrNH{N16|O(dg(Qul4&
z+;``8p!pF8s0Pmi71>Rw{LoJ+;^izFGH6g3g_DMpB#WT@w2fy!tSZ!J?7ts0oxF7*
z)4N|BHpzYY>2n&PIjS#Qh<^SEKF*0oB@BE@6AGdGLGb;8Y-av!3uUB+5KOdcn{~fa
z*q2IX47_E2%h8j<rKI(6-I|L4wgZe72SLoUHVKP1YdTmsXA0kCy@$85?p2OU<_4Z5
z@=GqVZpxIPuD_)@FTR`BV;JzwFm_depbvj9vD_CU=xm!gfg~jWJ`5SOP;yN3C4iM+
zhgWJjGKut)bDl~GF1jOmb&I%O+xWd(P)W${@mFOgdKkR*9S(j%X+Z{2i2^!neYX<B
z7~+l&wZYV&MB5m-)Eu7S=8bTqJXk*AF75{z1nm~rX2&pv3Ac%e3ijG`l0nJksAE~u
zxBA#}Jk@@HJ|FQ$bM3I3J_s$*>J_Y+v@!09o??~Nk?u)1wu-AW>hz|g0gVqMWpts{
z2d*baw=JN%QkfTq4H{<A#PFGtnd6tJ)Y&Xtz8?_Yy_Eu3A;5(GOe5YSfJv`_b@dex
z3c$M`uFk=O;mkRs&#w~o9+x?gLF`z3!E`>}_#KY4wnH-7)!AjJhkdMuazvdQjK8L8
z!1V`xz=t82i!P8xIGSE?5bw+hcSgyhbdO{r0DD4a4u~~SBT1(L!>0qV{LVdVA%lLN
zg%Rn1Wjv~EDl+~?7U7{PBN_BT7d~f81`)0J|4QWMrw<IGB({NIGH8VbP2d9mQ{es=
z32JZ@qx!ps17YF6P8mpR{`#zHvIBpU^qWh6hv@#6rN8F||IMYpi%W|{79E^*puXs_
zDw+(^C}s2FC#g44ky`cv1u9|xfwsQ6FlXw!mBZ0ihD+gU{TnHmT%(UA64%ScV|~+5
z)}=alp6;5ahBROY^RbV~rpH`4&>54GZFtQ+1r?8@Ln}+S_g3fFUURzgDE?72lZ&|m
z=g?{m%a0zegfm}nCv{UDjhhff<@t9z=ShbU(v?>;Z>0A4!|Th^i&rO_S<`m)$K}Z&
z2XjlWJDb?1sVMHluTJG6%*>i|FPOD%U6!5R4)>7?83+Y}LuYkuFo`i-Lr#tS?z}0T
zH|=(ezl)p|N{Dqe=f}PO4j<r&J6?nniolqgp8*F!f&Tyq#Nr>peD^rUk!n2J;TU*0
zQP>k>CG}-JpYUYkM(xmZBBv|+o9Rw|Kk?_xNQSx#2W3hfaC#EQS^!l&4i!ce%cQSt
zXdEe%gjW#Hx?XVYg8HQ1;f%R>tEa=(+6y8V%b=~A43O3zPQ(L=_+}F_C^Cn{n$CDQ
z??MLoA%K<c{t<cv_uXOIhDN`f@GzI4xZP@OO&K=Onrd6!848<A4=I0}cvcEj^L_H7
zJ8!7!FdE|&{+h?S8h6D48jb%t^<MQ{oke0dk9xEn$8<$`!^N8z(^oA5H2`?$mx;n#
zT_2M{%Ff_zfFf+6du_ex*Ksh$%qyzZgTyiB$9__PiTWdmo@2KqRmM!6b?a1UX{6-r
z_X{4S{i0pgoP%!DR!aAV_pFrkt79G`tEzDQKMn3gUMqV28B|1b>+}^V!u{vq<(RIM
z9Td^%7nb`I+g#O19AUs>sd}_cVx0;ngU;vt^+X9^8~jH-+_Hza4i15r(N=)d?)!iY
zng?9*Ur(?DYx{r3=v&<UxGGuw!-EMpE^`!2==jUwGoa_c<xr6>jLXc*QsbP)3vR75
zv{&YZ=Enmi|2pcwB_j{;4uP(ICGLOM%j~+a7Wyofsg+?Q*;m*4eIKio>GDxV80h-X
zy1cjb8D9z}WzUm<6vsyKqcZSVT&dL{eAE&?0r@lib3Z2&t7K|zW+~(-Z939@?yT}T
zY@<<W*C)#;3x{FT$4mMWuTBkI<p0pMym>1^PU`%RdwnU67vnx81WNk227lR}0Ea%!
z@O<Fl*XAj?v0O&8I9wY8#mXmTLv$Q2)@GU#;Fl=Q`Acgr6<H8XhMvAN$Y}_<-gzi!
zm9riu5?H<_&v&f>`G;g+LtGt5JTCtR_#R`a`##bDVv&oxktoVV2Ocorq1SCU4@eba
z)_>9Lwo!aRXII8{?t+i+gc$~~v@6#5pSqIbTz4vx0(ma*cNH{i<x!OX5I?<eA@;rD
zL>f`b!A}7TWAV~7<JNW(;`FflrvHYRkmb}%NU~18_4g*&{-Y!u17jx|zm>Jm1^{V>
zzZNHq2PkPclsS%biw60b%+oq?!z$N&GQ(ldoi^=2GM&n+*hEjGixtzkx%{cye$^9R
z(<GG<Qfmt2$c){~sJXC^C_PAgI2ZLv**DiQ33_l;mRv>vx8*z}4Tf!P=_c#n+}x7D
zys*0to>KRktiATI{$mZo867i0<mEu*cQ`-~5%mL*CSlL~g27B4JYV`eBjalQQjACH
zk#P=;cdWGUk5@wXKPG>^C3rFR47F$J6@qUEH>^F2$($f=fuD(+o^Qa}Rk5@c`KdPG
zFVq}0A9AmsaQAoxbjcK}MMQ=iUgNsHYfqS;i!H?J33iz92@bl0IlSU@33>%>d26Xt
zzC0=80ty>^a$3n6Yvr$(Zt{+H@02s{J|EU=jzFo@HGz{oW*Dw7rAizJNQ&c@fY1-E
zb(XyWgoS8<3y+i2=kqE=7#nyNwfvzhQ)odrw`+l_?5MhnojOwb`gmkP_tn+#f$zDW
zpY*93Z7S>G_I0h)=1~!nL|UoWd(eK}do@8ePA?LzOi$wObBaib9#q~b<*ABS*Uwlu
zAG?*zsE&q`9L#~xkuNi&f8H!QZm+;~T2+39bPmtoW~Hx52xpv97B13tdhE%tI??Ct
zEEvl+bC!j!XDB-8dZsp4*|*M4Ug^%+Z<ERmP_J>ZB%w_s5Jgh=oF^g>jxp{n@`7VE
zD!o{gw$EIh<rsC6?H9%L?}yt7b`<LFF9++^I@o%L*0x7Sa;smFXS(g&pofFS@L<a=
za(yKnh=0h@>fIl~Y;GXolVE#Bwr8f$ubbJ(V=<<t1~GT59s!RQXtarf<OG};11w&z
zbQiMT=5KNw2NaP(+4|pl%lDhucl3ja=Z`oybr|vrysWn$)-{Z|U#{l0FveBSgzjh1
zMO>m@2XKK;Q8f~axyJ2E0<wOEuQx|d3YC>J1@WUC6Pf2K$1sX_qW-+aBM1AHZb?V&
zHhoi71aEYaa*cvykELQXR#G~BF~l}`n+&3HCWFd)w$-d5J_gO`Wmqa)&nl&0)Xii7
zf{R%TxOft!ER{pZ1~ZG4ioD7#`&7QV2@k0F@G6g!1Cd&LuaH%(_iN^pK+}mg<^-*d
zoVhgfl>oN#A?2IT<_+9^p)|qn=QrI|6gJvr+uDZjSLQyM>(py}uJATGGF-{RK||9u
zhQtZ$5I=;$Q<I=KW?J-%n=>Q#t<bvgFRDMXn-wZ?DYaSjSF=op-NDR6)uhhsT_fCP
zYKWv;E~)qUQk*TfZof*~)Cr>Z-A6k*bfbX4+iz0A8*WXE9ckaHG5V*WiY6g${1F~L
zAyay>p)VwzfT%(%|8?1)YSJGQy#`Xs!;JjiJYW#GvV$z>Ap#N0Imn^-?&l;=<kJ(U
z{(|B~2!Uk>_2ZFAJ9k!?pPg#Vhc!<{w((4MRWZEY?8m%y5?urS^h7P{S$U?O{H*<A
z{CVO`g(NQX5243MDLo7%sDjXWjFHTm&p5Z<Va{F?ht*6`fvWrJdU043YgGkVJWX>s
zZ`#7Fbv1PCiQ%0_+p8y;uyqV<dqjVgy#Ornvw_HCLTbpe9D=Z{WHO_3z0Sd-z>JMX
z^GbF&jx3wpX^T}81y99*qBzm1+tXnc_3rZN&bjs;hgt{4RcYmTL`$YzGa+?eF7alL
zzuwi<*7^{>Q;idQv}_h$qft(Gl)iuboFGXZdvDPwibiq{<K|-d1GzTwWAk7S>D*v#
z9aAujPK0~Zpxg$kiW3Mic!3E_e_6EM@SX%9Jf@=$bw7_(RY!fx^oZ5D{w_CGj`BwC
z6SFU^bGGq}JjWXLo{C*zepir3Rud7*63~)1g@~W&JPLKb)0e1wgP*@j^BVUE(hN1~
zy}}&U$rv+9yikdkES+n}BA9IWvBL1oWj}eP9u4-?B<E!a+!plDd_?a+HQUK+<9<&e
z#Utp^+3+SM6$HT=oH5hH|I8)_J_B2Ja5%_=8_hu0UvKC88(XLq{OEFy7b^Da=S<w_
z`_c~bKYk-7^oBOv^14!B9pox$NR@J2*=&A*^C~xA!6%1_DM*MDaaz~=ZucT|r~5_G
zieeWljZ?vu-0NSKK35|GYpgGfD*O6l1}IK~mUZx`@MT<^!_i}-iW-~z^oepKOD!H&
z*X8qp^brcn4lR1FQ%6?QWROTC88nU9ku<1uP(?>SB+j6U`5zBK7`%4d6iW{}-pcWc
ze*t<EM#XH*T6|aLDYAA<{2p`ny6_h~#hY6-+s?W?b|$9tPvbzBulPL#;G&ODvS-C`
z_w&5*ijdbFtW$)-Du+y&g63*xCu0>gSW3A>U942)Sp^#ELwJXV(3+S^T?XoBMgtKN
zyj#_Oqb3C~b2otmv~klIX+FDIG8{-j1D;QZ3tvwCbi2g7%qNSreQSz_x_T$6O{ICY
z(vxxtL-a;lx0jh`(Ut*g`KcNr|6JD~%2C^0LaFt@C!>G1xKQuY+7vRhY(uJQ#a8Ex
zsHxVm+(|i1LZIf-By-N<JuPT{-KfRz`cM76O8=`{EC-f}t0WjP$!e>_pBda9Wkkqg
zd{z3h^zQKk(twrW*8Zv2=HI_;3#i*YOR$P+lon&oNkB?exmvmVg&wQ<F(dK%9SDx~
zYhLW`m$%)v&#AlXI*I#LRIEGge-JVHVdj)5B|{@1tM~oE6>ubqzO-a7j(@dqPUX%)
zhvBq^1yOcP)}LFeX2?cYdp`1pq>UO!#-e3Oh3e94Y3cbFhw5FXA_omb{zGs^`W1t!
zG<`kr`GP~_ihr<nAG?kUjoc1RJ1hu>m2kzzqTDgMrE2Zr9i$7SzOj#0Z!9G~#p%ZG
z=xRaUF@2~t<CB*D)bfor&YYlx+Zt%CM|KR*wP4=5Wt|!*QFq(C{G3|RFTOY5Jfi*j
z<{29YznK=z;p5$Jj#ek{bCGueREZU82i6iHKAGVe6mXs3ish?K3kVkOSa@ll;zLyE
zKO4(c>TLT?V2X!<=i1Fn@qe&y$16n=_i?Z<H{2ORj&^i$*dw&jPymO8g~R88kNVGX
z>*czi8k#C(q4f*e1}Ua3HcONx>hGLUSagW8cahg_hn`PGRFg3+1t#Mz`in{0%t@w1
zwh70?(5xCmmVo71Th)LL;b4x(J7wRlFI7R{^;_fH*~itq+Ml{7SIHns3G}j5OU`DI
zIFRsh*rrv(#)a2w)a3DQ{*a$~V<KI`#|>vP``XJENZT(jZCe|da$LSplKF6f`l>42
zin9%6PvCln=fpI%t$5T!B9u`KX7&n4A&*bZybUrQ(N6Q4x4YEQ5YDuH>&^zxW1*wH
z6P^<Zr=!RzjROEU##y1dlHGh|ypeI<NdX9AGEOb)>?kukz{_q9?jg=ud1Y<L?^<pR
zr#LB2xxYE|I399Ab<a|Xt)<b=MU)c7ai57gxTU)J0ZQ}y3TL-`T-7d*!{R#&ed$#P
z;nD0Hm|bj~$I@$MUt$f$M3)SLsB9cd*0YXDK?v?M-beg*{nQDmFx-UZsJoS})|h08
zgz?yu#IJXjwy;<iYL6lDyCp)iLkqRqUlQ(<c$~Mxs-v1bbs~2T)KIw1wo0mkHGSvZ
zi`0QSL1s2B7ou<8P@cm>1pLeh>dH>;SG6!(h2Gg^;d9?Nex7)AS96sTn;hY)q8<*x
zI=odS!iV2q&mehAT=mUyjFCUBWYnqDBd1$D0>tpTR;y{@!IH$=&iNMm(6eMvrBhy{
zj^GY)Fj36?_Qe}*r_EQkXApIWZzTW5eqnaoiPU*MSljM#T+Mn04jlVj**_mWyNdgv
z#-38(o2QMuQq*Yy4G+@fF(u~Z2<sc3`AQRcm^t0zY23_6umOxjo;%~pE^|C~V=-HC
zQlSNJ+hi+qRy&<83)R1AW4mKNZ#+A6`gN@)jWaS-ng5g8F}I)b&~rbDqE%Kjy!DaW
z%sFiT@OE#SZ$45zD$!Fsf3WVGrp#UIwWN)#CFBONdhpTw<^}dmZ(+Dd9wI-=im^}(
z5$)%Fd}EaZPt=XmA;9ZcF*8kDZ$3M0VfHZl8$YyK&;i$B$3ySxAn8j>%N<fQvySzM
zIDRu+Ky?{WeY8{pj-i;?bSo^dk|1@l1g1uY%I@oy%O-O1sjRhaGro~8**tL`&-8(e
z6s9$0ACtH>)&ZN&e>Ll|k)He!bj8m5=W)!<L<a%WEpzE_MT^F_7aQoyKz7t&Ej-Eh
zZymOEf*Jg@T<|qg=i0LM+~@Q;A*O^Zi_wp99PvL)%S1!FY+m42zeBK;bD3eaU>1T*
zr+m$nQGe$fForF>BS1XsLB&`juQ&EabkZ8#xQRr8Mel`*kNOI>MZ6|`^wHPslF<{m
z3gT9(uT!Cw>UM_?De?r-9DPjHF)Q{Nrxhxf#q3yyUrp0{GjX8L&H0#3+v=-9x6&iL
zWmp#eoltP=b}>P>osTD}JuPk3gK15(f974OT^$QD!By<hG%Z5ZW5cg@sj+n(Ln<r@
zXls_h%~n44#9G~1Rv)wOy5BrR-$;CyTN$Y&pN1LiI{wHa^NvmEk5HGYm$N(sNTJs=
zEJw!aIE-!#zh*VyhPhLkug0*K$gg8OLh~pdiACS@t7xP0o_>#ilw2TmWU*U$2w+fS
z%tn@WvgVST_eswts!Tf2AV2<jC#kMkdhV6P-gY008B~Px^F}If`-8(UjmPdKI$nP?
zX(i+&D8HRSUhzu^oDS$3-V>{iksl{f*T&MmO|i8o_Lf*Y*LG3a@JhXca<SMSUKhC~
zvOPsN<xb!b9PJ|j%K6`ccDjIi1>~c22f-WsFu>j)k5py^z_`}H$v=+aCmdA|P7N>v
zNZSU_J;=#fC^jz4L%~`MuUY9@XwT2QcPCSPbk6GKM5pxayp9Vd!zIvtgF-=X`y#dv
z$cv#XZHQ^}pVF50oCfVNf{C?{sFQZ39Crq2hzc<3)q|Q8`KgfdnKe$Ir1X$(7R&p)
zH_jXEguMi<N84srY3aCn2Yh-texVt<RY?ZPddc`%hwI>zGb>$0atgeBb!+f`^`SYA
zEHh{OlNaj`Yh7c5?j-z-R_*xmX&VLIlwWS0y_|F}tF1n^!3I;@6!49g$~w+iW9LpV
zZ)dw2Pq5H4Sl@Y?HowNB_uxwBc%;5WM`p1<T~vwCI1+R|Pw&26$|8>=Rp5gwb3qh0
ze|Q&X@(?b$XXpLk&aA}7v47XIoqFkTG_5&~w?Lkf*Psh{eDTz-@UBp&=9a}}zbjJ6
z{N@8ln(WIsMF)pI(t2?KC#rYMh5*4>ZHMjAvl><5W1D>4ir_VzrcG%(5_;y+s;1_*
z*^AH0*Ivtxrp(h;c2(rx6p~UEc2H#=-S7iLy^99q8s!x|BQK)Hu4dk_O+S$y6QpGl
zXQk`9`r*0(T5ihAbKg%<q&U~g+Ob&_&t4YMa4e`lPWQ5{pgg*($?CD0-D4i<ur!@B
zbT3}Inl<yDyftZ4Xy*2+MzGB~WOBV*8u|$hsj;xI8q_`A*Y!M6oATZlF5dKuyl7vc
z>9Nw14V09=hMNuhmMlq<1fTG~4{bs5`{`k=_*>0ua^`$hIQ8(3vgXQJ)A?Bas*_Be
zd5aK^IW3+N{qTpEE;!%1e;V8dp5nJo!pH*Xo~m@;O{fk|T~$2O*Y^v@0etF3$QdT}
zb_>cExm1@xOm!*>U$!TE&2R(3L^_gq9iDwZ=cAeC_z~~?t?@Lwp^(l^uCR;4ObCHs
zp~O=R9`ljD0bSz|TwF~QN5ecVl*{cY?xgk0_}T;qS9}Mp!eoUq;^uyZ(^DE2+2F`J
zhs(k{M7VKdN%Kjn@ZJt_@9-sYn_x1IvaiI9;qyNh+NY#g^WjD&r)oZ`7BLsp*F{cZ
z^};TechZ03(}cefQ?d|s-fo4(a)UecEG&ngBiwY?Pehcw7t`D7TOCti7p7aWcIuZ}
zBHAt)L>G3?SUmKj7}=`YW;4@wD%NFuEm~7w`#jl8B<TE;Fm1C(>~_b>_s<UW$&5sf
z<6v!8kk<Ttiy3Hp^X10+%k}eSJ0y(<cZxKq4b(I4Yv1bc#J;$8GULy=cYy90Uwe$W
zvORG`s@_pndw_q6+!u#MUM@o4&NU@e_$qBGY~GX5OsuF5Zcx8JXsE#{TpxJ;*#+8W
zA$;DtS6ip5P^R9U?KW<StprTEgzA7{%D2m9i;aOlek5`AvdOgYF;Z(>;o~hHQkKBQ
z?oYIluHyoVvwv86K_4_Wrc7o~U9C9?%}cxb^0to9o0r#vLij##LpHV_=lRK$<J28W
zbql7g)jp|gw9C!(w|hc8Y?JpNNA=V_7n_T}-YJk7E3r|!UYc~A2*0V|f6k{#2~{UQ
z*5r05eqkUiK)D~QA2~70Na`NmG@Qb&&CI7-J{rq0b&Z&9!Ky9jMCT^xh^T$MNa?Xe
z0MFkq#WK<poY5^L7Axb;Mipq8_4L{l@x%1(>+X@lwO)C?5#Ku`6+&p7%{nza!{@Fd
z`D90kw?8?~ii~BCNIq*Esfxm|PEGb-QIyGft-cwWc&liq;?r{wWtlYQ<ek9q&r*&7
z^|8i>m{qr#D=I9y{ow9h#Ph+#g#M*l>+xn^y~ZxUo?q)14bW3S&O5+%Yi3mKk-gp)
zJ?H>b`23{fOP}RUDYqI|$AAmCv~_b~L};m>mAkuVVU!Nb@I%#=@#}2q`9^{9c|l3F
zoR;)62W=?U0l-u#FvV>o>ABj-m1QAJyJcj&9b~%M5-5Gd?5RRGmlD2wp|3xA|MOkd
zjEz3BV(jWwIVpM#j2i5@s_E4nR_g0pcRXJy#q^oJXmtNp(1s4H&rHSZ8sX*kvs|3{
zczF5RwLc&yB_@A-?eIFakanN(R;1Ur)0x}y>Cgix>;#DyfSjv2^52gEa4kT_|KLb0
z`#&rqIP;g7uw!!9QTbkRJ>yYS+!0`m0EMEqUc0>sZeCK5-l63t=GpQah7A|8>o)Rl
z1(K+mTuI8qKQs7%)Rm%LQ&VQn)nh*KM~`+T_OCzr;ttwuQTe7ckz;Am;Yr)3tRpK;
zkGK%YW$MG^Gm!<gW*H&M`7b++Uz@u9^5)J1$@6f=`=@VGovcy&En^XfT-A`HBLJRo
zM=sgAx71H|#63n6+9r!rE^sZeb&~k8qco=7%)II#P&s3z{b8Z-ip=KltM>urKpn&)
z;&?ynNK)~DSAoW`VrFC$SQS@n*uObuGLE?G7W?K$b=SX2NHHadI-I8&eL+*|!N*g<
zX`a|GG6Zuvp5%%Jb<Lb;!skR@Nh4+m1V03J2<{vAXW{I)`OnB*aM-<BF+4n?(V{BK
z?1!0I`S}@c8&UjJ!Ft-yp40!8BF;bWeGvd8+~VBGpwD@!q!bM+de_;%ocMQZMA;K^
z2>N#i-JM(?ey-k;s%{8rK&AG1?B;3Pgj!I?*T#xXoC6)YjwZUV{99>gJZz>U@ac0h
zvdAaDI6ym*MmS0fXL!MzS-ZRYQVJWh9sw|gUUb9KH*du4>{#R)Pq9kSU+YpTq?AST
z9Qi~?l+{!a?@AnM4meU$Rc+!VFZXF0O)-qrws8uv$m65muMcTi1R7|tf&O1^6upFB
zvZ*-$ymd!-g@2zB`A;Bkkede?v^FrF_o<W&8p|rx{f8mx%$D$Ce)EP8WY8xd%UFgw
zzA`2rw`a3%;>N$n$-g)8XB`e`<P+>Fwi)Rqs)t7!oHzp5mC(^fac|qc!NP{%M6Lh=
z-};Xe#iMfrr2Z!Hh7FS&Ku;1`z3i)x9tZ}F5M})Rddy{@%nOZeG^E4ztTOS+k)H&&
z2BV-;t+BfNja_x36U3jzj~-AP?|+_b7YGsxx~6lMJCsXQXwD~5f8KCNg9?D&2mV*u
zjb9*rWnp#Oe+%&|3;Qp1{$C;Nf8^eqa4Nt^PW=G?{_p498Dh~VW=Esd?5Ly4y+V6X
zcBI%p(kTCp1mwApO8gzS{xu_ai`#uye^X)(O}HO$!%q@Gjbo^BsqHHDstpSJ`q9ff
z&qr7s+WP7Db+cya<`}QO=-*`sUv>Q2h$~BmWBp~?Wu^|(I@6t892{JiZtP&+w-1*M
zdSrB-sk@kfjJ92Xu)y+wm?6q=Ueb7Sm!`8<7X~DL>9!h$o#v!y2wIzPMrA8+VMO>J
zjJzQcyfq6oZ`EkJMbqo=wdcKKJ+6~?20LuCnG%wWDu-nOb4j$#)A-0@W#9VjxSUeV
zqeG^~M6E)LVCZwLj8_+4FCUma)xY-qh_w?!=O>4!)bxkz9sA0^^UX<G?&Uuc{vK`F
zwR7zlD@KAf$`f;Kms!nEvWT=;Mke`k-@F;woihPrP=;b}`WSr3sn;{+LfN#5MW5=@
z6qF1oFt<3locd3E`w=Zu13&M48ZFdho^kcthBZLZ#K|Ro#sC(A;Jj#9qds=~!fJ{S
zDl*-ATq5qZ6lK&2gObH6CP+yscz5#fGK=)v$j7}p2N`d%UOD@*_=!-zdlf4lH31R?
zr#wfW`w^aH9ncF)HZPsx`-L%$!7T7R!A)dEZCo<nlN9&x$ws4NWddtw+Sq<Ukwa?3
z9%bpL$w??@O^RZ|15SmM_`Y~6mdH=DFI5+eG<yP1!B_+O&<u@M2)*JpbGdnaxy71~
za-}tu{!VrhW2n@HZ~F=NJ}VbpdJVIccs<H;t3yoHglD6)Ku-d0C0Xz0)W8b%M$Q83
z!QzRT;DA$rG*S0$Z<(G#r~t07`FayBMBZRERo|CGN~Ef$!69In9e<^xnVEZ0;d6qQ
zfyurVAp?6HM@3{FBB~Wm6N2E(UKS6=ocC^Kr*GaH-?qO*w=J{fgp_KEo4urc_Nxk_
zl?<9hYT;6<KY0aPMCxf-CAFKERofRA&Fb3FOgTUE%CpdNp|m{n^t-L;0o`YY+uc9n
z_FZs5?aEM<w(N}N#QmZC6zG}F>S;pe41nu110_6T-1j(t`bQE6yJjCn^q#>klnio0
zF8x64Ogq`mZ;Hagw$Ogn$90C*CNt}pBmGUg`35u<A}a6mAa|P{jyEo<nqHMwJS=nH
zd!aFE19#)?p?Lk7SZw0w!?5wI=Z@fjR*#w$#ivF}xv2zv+aU>89J~2UdGVXz#;Tke
z1<flQfmFp|aJbgO=~JA}7Uj5!FDOcamX4FuJ=_NKDk57A(^Z+NBPrmoD0o@d!;C9f
zY<IC+Ii4UCcwHfo#NcItX<9f)lWWeA*!V1gnA(eZvim6W2F2H~%aVD?*>N{i&Djla
z3E7?-xet?oTlmjz%{0L>O!~Iz<%v1)rd|J&P3Ngdu|lSQaJi#sA^c62*iBB~u)Jn;
zp|++G8M!t2V=K&!AtpE>-b*<Uehx4`Ls;7&KcyTXgj1XV$o>}38)JUr#Y}FX3dI))
z|9%le0cd$Z%S@cmPMMG>Gyt&u;ORwnKrNgR4MVM152JABqyV5k;kL3IV0-NYP?bOf
zI9`^bP6Gh1w4TKvjyS6jM9Qc?5&I-2eR}v{QyzSd4MNldIUAV5aY0^CQs|~Q83fAy
z6Sx1Hpx+$&EknQGvfp~>ze6#e0{?2dvRCpp8kd(`>x^$Z|2?7eqtT-mAS%GU169LF
zB`1#Nk^f}LR{f)n76toXqXCA1dt<Cx?RlFFG5`RSR`7kJ5A=%vMX%zo#{11f@cfbf
z!SvBYu%9*|IoD5N(8=B3k+v{9b6LOgS^3b@v1f_U?r5<N0i|T4Fr9_+E3dgV)#2Vr
zBycVn<c})#(rTCD_EP(D#E@0!)$nMzN}PAx<H>>XDBm}Q6wEt;Xtsw9r)-~!QZOkc
zvMEUwh9t#YiIp7C3)pS%Gkmj}UsgS06|m{o9~H+mZMJ4C55~n+A6woVOfNRB(w=zU
z&-F5~uYyT_>k|7nhST#MZa!`)(mD?A`ArRsRqC1Ug@$J&KSU$SPoN?^W&vXbQM%rj
zs*w(tk{>5TKTf<;nzU555#J?z{{6z-d!tQCo^)Sv*?Kc$NOfC#pBZXg)b(oC5xWm@
z*kx>XIY~c1OM0L^F3K3f=-yA+GuJvadoO=4Q-dDnJbYDLgq9NSd)1%f(D?ohVLL+q
zd;_qXK>+jTG1>1<1+}#kKb<b8teVAZcotfi_k0njsuk*cv`i5w#2oLfAIoulN=Kqm
z{=ph@_S&(0WZLry1)EI^b1nZ45u-4%`Kt@rixqPxTi#OBCRoW=IL|qB$TyVAsJv*0
zD!+k8dG`yNNp~KLLaLWLERT^HuaqimF4inPKr0!#<((1J>~}uiL-oSc<4R)`2a}Su
za%+evgi=}fuzMdt=0OIDYeixA?gXc(7BXY;U6c$ezQiofP-h`+<?0swVn$O-n69g=
z?xirNhIt`MqngAzu9g<T*t2E2txm|f*~k|R%3z8XASA=rYuOoUrQ7MVSB2&A)sxo&
z36JXrmNYTAG<>k?Cr}I~R&V&ZU^$>?B};zdVRC&OyXP%cTem*ldVFa>Vnxa2UP82k
z3(r(zUGm%0W)ErlJ$`TriXOjTAe~=UJ%XWE&xdKx>aH7Hdzw7T?9S0A8+mpP31Y|(
zRC)5(2+02g$LC(cOJJpJI;8UEst43S{O|w6_{KlzDE~*!H5J91nvWN$M15a?x##4z
zLtNK~)K0a#=uu11mMtIm(*<c?DAxQU>r7R6Aq=!f#j6#apE!MIB0o$s(sa|ASLlwK
zIQ5e=f*(&eUxQ6Ks1CW)DGo$$tM<Fd<tBb#sju!+eQqfG@=EuGt1;C>Z<2-V^kDZZ
z^V6zFwp@+dTASXd>p3}b{X7|HG2Mp!?48aK)oZ99`K~N8qjh08oWYhs$cy$egD!pE
zi~E}5*&}}Z#jS%93&qk}S9*pkPrFuyQPFZUWH&#-p*s52Iu&Jo%961XPb_W<r7%59
zT$<HLyv%)>IVTBPES^(HNAb1fE!;lJH}3UA0+;f$Pg(40u>FFOz6sEw7tN)w-@X2L
z!K3<vd*x_hUq|rNs{$US2Ep|v7$zkS%gC%&Hfuf`W6on-5xF)_TAE1|aq9b^B@n2y
z_PzVd@+;##Dauu!&QQA?>(%Uay~>nv`-`cr)V|&<+B`xFZ@h+HDKFkeL}E@p$sQxv
zptBdM@v+tZWh*Y;6Z`T8vjalVWEy*Q)yKVvp7c1;Siqfz^R6E-X<dQVf7W5U#$wqQ
z6C~4<ct*9|)+95rdg9?=13|wmU_gqv3+wr;%#T@5m!h!uxu4_MEvsnDfxNbS<1tHd
z*`uQ%X?i~(=5ZGB8or`Z2*BoSv;M@brxC(Q!BbhJ(>g%$m=VR#C_mX@)5O6w>dd;)
zGN9W3>zjB#>7mlCOBv6%rnGI%9`-fWy3>_#-5*}Kz{Dt6?FUp+{HJ|BpczL!OC@<D
z7quoAfl36YLz8aEufrPrwF{6xbl)4koF4GVKNE7egiiJYtlHnB{Og1I&9PtP_P31v
zeqVlTnBUs+|IGD8@1nS1(6yKLxk)EB{`R^5i`$`L_%32prU0-Rlv36S$Jk+E&YI!s
zwKaJY?(KF)dVPk<;s*1TS=Pj@S&gKJ^=UWN5e46P++JSlY5rjoy$sF^05W+AuI-S$
zWmf>THPvkRfq{Rnrxysh-mBDux<A{`cxKuVE57a^k-w8ST4o!Qc9Y%8Q?MBH<-#>Y
zqN<{P2|-~$6K~QU!^ufMN2qFuYG6%1)@U3$G?d)nZDdm9Z_2{79E*3&0%f2f{90i^
zwHMrJSJYG)rO?}Hlzy{?w7qW0Nt9JMxdq<wB!j9a8iK#dTy14;$6TJh+yrA^9Ls6x
zd^YI}F*JQ{rK^*dm-`d#_HKtUeA7>D&UYX}@#j(KOnAgz+$kUid|P&ZM^<{e*VUT`
zW6IQXCjos^r0F3xH3m`Te+N39#XI}$$&l=1_Mp?=bSaNPq^5@Orq&xAnu|e~hCxd_
z!Fr4gqNgW=9xxJtO%pfW;B9K~Q6R1%3O}~;VGP;L$hk8<JsXzRdpwf|OZZY-Dbnx~
z;c;+ahiW)MZ;`<8G_T^Wv+j5sv+%B2DZE19PvOB5XFo7|#*Z4;ds}RMoi|>weyiXO
zla4_zY(orI^_LtxoJBI{WX1f}KN0I7i~KVAV^dv)JSH@PRC(UuzhTVKFQmCc*79AU
zzKp~)Ui3j@HyLEU14RQU{DsPm)SW}fk}o^^JPJ36-U9(UMW9I&T7T2iC>d1bJwPM?
z|9k`*JE_JNy-Nlz6b8y91^%-;he(CX;N5v3aR)@@h4|N>>XAWUAaiX7G|_ScYo%Xk
z{Dpn|YngzEse*Ux;XAi~kNCIG<2P6T*OeS=%(1Ei3fENS4;1e8k=l-_*J)Cbn>+_V
z(dp&5@miO63^u95`-C*h@~ML;V!G02?|rMcM*HU-8;57d6N$rOL)QI3;(MJ8dPqak
zf$#T#F&Xf+KY;gAiZp3Wv7ITS;j2KDL7d#<KNJQkY{Y<FgWSoW!eHP9L>=3Ky!zQU
z(711?y=(A=R5Iv!4vA}ml+Q>6wiexdLoqLFQVDNUB!fm<fZawRClzsoXON?};2m9H
z6UzrQ(f`@txA=Y!{l7EA!{+w`m@333tQ~A`XwR3QibI%PqpADA$e-N=Cvh$9lHks8
z9ETJMtc|Mt86|5lP^<@Cx~A@Zb+FK8(C+CCqCM4jAdhZ}HW&LRTk6Bv&40!!8CZ}(
z#bEQ4e;9Hj@NYx;MJD`fUE^hcvm0E22HOLQ|Hei`KOc9<sjLCO@IApt62*&Gj*wma
zJCa^eq{c1yPYaWtM-)pae(-{RC17s-1YlY(AI{@*ylgO>ZC_2gG9_{8I?sHt(`|#>
zkq`GL!<NC&aQf*jBo)GOUD*m(X-%lh@5&+g+z@ICGVs7+4Y&^-OTv@Tl8we3ACE5*
zt;b|%r1b0c<Sq216(Wbh)t$=K-rSm8v`HyXm6s8D0j2TtpP73d9<+}exTGiPSEsF|
ztpz;HoGahyE1GB<dTAJ>Q;Kjq&{IiO6KzX-UQWE(H2QQDfyf_I=7-C4dAlrW(bAjC
zshWb_uJfb|ExwEM`3NlKK|bJM0Mf^F`=l7kP?)2|WhRV271H$}>E+e(yfl7A76lpl
zj@szpAWbDx_*4Fvnj^(kgYNv&pzq!JuLFk8l0jODkCpO0p1lJ-Ul#mu16$^Zszi1<
zLDiG)Hl}JPPT9{#dK;A{E@1>aJJK*VEIq}cUpxdC`GT5OmcE<2*I^=qqLN@HKwd*p
zyx0lRhlf|-WpcV7o>dU4^03&+J@B_i&RC3b5DoHL@U;6u!4F3^=i}O~x$(TY<{6E*
zs63xY3Yu;OZVfGH0_NqR)5*Lge~2p6%B;fphbV)Z(YbWfF;S$Rrk-_Y>v>>njj#DU
z-j%{jjI-Y#EQ?VHik4KKqz5Rj{<*zhNYB(-D<jKczsvs0RyK7V_4Uihr)PwiqnUh+
z!@_Z*^UL8Hh2=?iSf=hBa3DFE%?9KuD^p=t%mkiAa|mr3o;JmImuWAZcVtUeJYVa;
zmQpdI<)1Q932`qeIT8+RAAet9aaP67KYOd*N&z~pwdy-l<BbecW%E)!>T2DedNYOI
zFGJcDTFLd743=;J;5^^xC-Tu%>Em=ZK6CFEmMQA>Tl#4_==0a=ZD!3@DpyPt8&ecx
z3!T`eG4=AZruA;ILdH*G>XKe1ba(ptLMBt^L*Lii0(EyetDh|`%(JTQ_|-1&DlabH
zoLTDb?&hiOJk6<llBRlN1vPnR*h*S<<YLPBlt?CLTzB&1Ld;NnNm1X%-P*?kAJ`|A
z(Tv`A&wrgkdu;9#8*WI6D|zq2ZO7~NPe}sL?j#nulT|v=+eQ*5fbZjyx89wk!(r~c
zn;$YQOx=~7Zf`6w^gQt{{nq)_=*6uIH>hp7=ZjY%tR%pCJx_e>P5-TsFG=wB8DyBa
z_eW&d`4%A|7tP4m(mJiLx<w01;Djgq5oFNrYCaH-eF+qD`g{+#*!xjOvrI(N8bL&<
z=<~3~Kb0w3)S?&WRg|3C*essVQ?ju6@(JM&teKYxfN`su1|!i({kfedof;lY@|Q&u
z86dc_xV@+=z<yt^-=o*IpulcqJId!G572)dj~~|#W+j7`M*ySu3*bDT1OH#_y=PdH
z-L@`_BBBDKRH;!wP^u!mMMau4>7A%_Fd#iZAOZ@~1q2iXL_|bNqy(fT(gmb92|e_l
zPy-3^JZtUqeb;{7@7a5ub*=CH&b5z!`0+5G%w&u)$DDJFd)#AKw2%_U2m1ZeZ$bPk
zEdo6Qcx5<{O~}Ue90CY30K%S5!2D{C!3-o(vLZpK!A7=OL8a<FYGVES=ZTL4i_eDa
zgqLL5edk`fpA0nnM$q$k((2Hi%KPbMNrlG9<4S{sh6UjgzU0_PyH0%-rm}mx`$u%?
z*+*m)K%Fy)@PHtCs|$dRM4b?dG9(UEmLc_@kGaY9d%)jh^;-`AW;6c^IY^rW=5GRC
z?M57(QRtyeq5+Vgw-M0zEfqr!EVOaOzKl&8TT+OXJs-yU23zyO^F`-kY&@-JFF@Q&
z7M)^1G~POjJjw%LhWhIXYmEHg(kNgIaBi!B7moqR`ZnaItdj?R(ZAD@!cLIMnDsch
z%*)s39x|G{(*6M9;5}av@`$O<gKPahhgZ?r3&(GKV>KeVHDh*YKtx(vApT)AGlZ&2
zzY*}xM(d11`#iP<%C=yX_$4Ft8*SG0n_PzfeNx@Ogc!F{f~|f%C{yk9JOORV+agAQ
zTj|ob)jFVfzst;)7)j9ALDy`r`RF<h_cuz2vA0L<j9I+)q7r+3*{_o+<dF{*qXH51
z$zoYWv^M>4gHA5Jy;sYVuQjjkTeqaXHT8jMg@Sk~;0~WpiV$ccC_+T>9!&zz5o&ih
z^V3p~zmlV(Ids&}XOQn6$NdwI{9zLWJuowRnCy<xPRLa_d6XmNKo9|J>7iTXEnRbG
z1{_A1ajowo#|enjPb#a$#cn={8|_>B<k}?<EVDX<36-!X_{E21DB{Va$ktm*7iZ`T
zwLQcxIs}g6+g8e!Ky;$N>=W(sWpV^lRRo-8XV$VV@?M70t`UyP*CDR#5tsZI*Gcl#
zypBEmh|BQr02>*-&%&TUe_L}LH(WWnnYy|7?NJuEGPrR4HojS$e1vh5<^Mi9=P`$T
z9et0aG-<Z8@9~qWYOK#=*kP|as5UTY_<$_C8{{+j`6%BfX2Nno2%i@HXot%tP@ka(
z=a%8-Fwnp2u3Ls=ejP1UYphwyLi6=u{$t*Dh!9*JAG{d-$hVg99BlD!KluXOyTFZK
z)r}^a;>}ikxvQETsxnnxD`4n(a>1zi3Fo05g&D$#=nc{VPB`BH%C%2!)4H4aa~Laf
zkufGps!|tOjXJP89CM6r1<uMNO;bC(17N%)m?E4&)P*bmq`I|!1E84Y{~L+isnK_(
zm>FsO{t<0O51BsU-{|fB`0-S~$NV>xPmxuK(t}UghmlysQfE5kSIa2mn`*(blP_Do
z%(on9*0*zwxm1(j^5AmAMef~WY1_Og$@<JE2kd(;#4a2%W|B@<wFS(oa@Arrd6nHy
z1QbE0rRh$8p2olpSr*CFy>M`I+!B(OIri9XGUQ^&VC57&l&l+209PP}0u(m%Wo-qp
zbY)2_myTzR`LGKv6*dCC(Znoa&1@Lb6+(^QU)#pFC|v>&6Roe2>x1H(=+Rf}n33=O
zk46_T6BDtEZ=9&}U<s`fKc+N^#f7lLdCFCoo&&*}^2i@8*lRZ|t;})lSXk>k(%^nV
z{PC7APMiMk$8#@yEJT|iti%pCgH{(HHH(6K3kct-WjALZe*~%*@81TF_CKUlD^E3I
z7>Qu5ks@f(+WEN&uMX;*Qvwf3y${A*7loWpB<KwMq;ma9byj1#hXv@@G|5c>hgL{1
zVpW61DAi8PW43Bn4j77CD#}$A*@SCZ_fOtGg;X3sIPri1P6~rh%V2h&F0jF{(xneD
zOOE&q7q7~NE>`$nybG_q>CN|2I=RQ<L2?g5^C>M9nz{kBz$h`iEd&(S)0z%NX9}iG
z*Q@Z*=~Ot|*lVX$lI}F*Z=ctC`1Sl3Xp>P2?u{L?#e;lCH+VHIxUx4oaH}KG2Jj0+
zdS-G!y)7EOI1R#TXcCZdKAWefh;Gfc^dl%60k=RK<nv2a$0-9BZ?x`hf6qx1sgXY;
zM}KzQti-Gna~x>K0sLM#KOrysxrdSweqq6iM{x}xPC%Yqt`X?hgt<kPuxU9siM%kC
zi8`xiJkkdS05U-U+{1}-QUk!{m=4U7R1g_eGaf{)D+w}nlaM6LEP|+5pDx1ntA7Lk
z(Loh;$~YH!{A#ctfAMuseh>OB0^z?}1mf*~15O$U9kw3!-DCz*eSZtFzN-R2&mkzW
zKdrm{N6LNGel7g_6DrPsF?RGU*B`^GCc7T-;5lmo`%wTuJik<leEE@jK!pnNZ9V{q
z#WDhI`S^?LRdI_69(4}0aUzYACwrxXWgFA3jp=(Az_Q##GSY%H-4oKz{OEWi+MvFy
zuDR%3r2p>^q$@26Ynv`O4<LNRyo+IJL#jR~a`i*nv1Q|eRz<8jeFlni5HeWoJU?Pm
zgYUes!;?0V=(O)+RZja8R=eU#80RJUY0_<Hid-EN9bzqxp-w2DSWeP6{9<1B#XW2K
zQKBX`g9fd=d2wThL)kQ(;P!P=xZlXQ4`{-$Zl@De@!6`A!SM8qTiQi|OMTxYuU|aT
zt*1^CAlvNmEQG&`pb@>PBs2pR-vqKUPwi9(d^)e@pkV%%E9^||+}AW4rBmf>IGF{1
zG-uiBv6i;DU{2N9m1q#s%n0{JDq4of%0NiWVP*1qh6ej%TMMlK1;se2Qz?9jZy%eN
z@)N&ctFAU&)m>)w->N+V_ZWD|tYx8BNP`~4%BfsSpMCu(5H}o*pEB$dLrUy4u@0Bi
zIOZ!I#y%ELDt7YpDc66qP6H*<n_b#>isjiQ=1o`PT?jCOUl?5Qq+Xr0Tu4*gnD^qw
z=>C`0k?4bReY|WmTmwHc(~}^C7lCCoE9A#U#_iU#aTFVxPX{T!Do)XDj^a|fj!J)G
z4)1&uuQ@fSL_0e_0A#h8&0$(a623@wwc-yk9!D9zMJ}eR`Nf6$k4DZuN+P_Ey}`8c
zPF{d3GybZU+@vHCNO-yUF4mzao8!r`<JNwzl$wsoG%c6KDTGe~E+PCf-i}zWWRnAa
z+UpJt@V%lN7B}YQH1QeUe70Us<TB-&i!84sK0f7xwo{=qg|FY2!IH<Xf0uW{uF~y6
zitw45&(+=2Ul}sD+j`LKL@<F?yS$#jn>$l%1gWbeMzKdnBos}!XcxGeVnz+H9)e??
z6&ABE)6Q71yiz$3xIReCBdGuZMXnMhem<ffIMOuH+^yWk+{$`Pmlm5}U6DGeNSx(z
zvcL)j%*NIvap=UKwXCS|pm2?2y1T~oK27-W69j;!IF{rj1XwIs!VowTDF?Mp;S<OT
zy-GK5W5;qV-7sUa6rrT<)9=0jgkk3n*v$BDQR1!qrcXVH9d3K3BJHxz%R>;`kL}K!
z4$7u?rPUd>+%!@z4Krdf%+>H&z2J<M$6YMrcX%#*s)YJ35nKq@`&+E^v{Rgqu75+j
z(9??5lNG&n7Tyo}>f#p&`7yrFzB?Mz3FySmUhr(J+~qh&zD}45rGPqfT+->&yIPZR
z2OiK};Qmm)fvO$jxa5`gxqkHk;vU;kxh~c~lN+s0m1=NrlWhb^NljYcLonQ06_WwZ
zIhwVbdlg%OyoJS-J_w_aG2XhVMZq-J0k^JC61t22JSa9_m1@9Ms!0BNX8uHgZW}zY
zM8$R02g3vNuo{t&vQcbSFrkz?)UI0w7q1yw->P^9B^0t*GadUjod}U2-k6bSRu)`Q
z_PE#U+eu0#CS_gn9dxfqET~Ujde@%+!b-}ow%<%KMPB%6!nJ8sb{O~sJZL6?2@W1m
z(w2?cSXa8Re^#0<W1T@Nz~XJ{x#3lss<FK76YtBcFW|je<{WIWYGT>D8~Gl|&;asl
z$v&4c!a@=`X{WxYXz5*a=(GQw&T<dLn*9KBH>Z>^HQo%@jkeEB)=6%D)nhT&3t-{4
z6^>WLrD~mH&W>fxIqiP@IQMp3cFFkBTy>?22i{}7<s&gkp9&TYT;pCWBN*WRxYJ6C
z?^Yq$1?`z@PMQwsBj3(MiL`cH7dbz;*E4K#K$?~q9mA!O`9RcYW>*SBhVB4`huBH-
zBehJqqk_0<%`c}rE4{SPX_=FJCfpbIr8{tT=S-!m5_%O~GBQ!9-nCx5*%!z`8prM(
z*lzWyd>RHyXWqX37vt`azsX+ZT2*8EdvxWP^53iAT}2SPPOFB!|7<Ss#4nje)ZN)5
z^TVMD+9S2LBfuw(nm&;NwD(Upocyy!l{0mo0McghIcBwc5s2=h;CT>J;w(k2=R9I0
zvxgGndT{rNGUIeeHd`L9^UH^TKDFe}GW$%oN4NPw0T;Nv2johUn<4Bfyf8>JXLnqH
z|LoOGDaSj5_5(g$frr@M3his#;uoWEkcpb>op?sMKW>MsicN>K3DQ%sAg&yITXY48
zaoKCt)WBLiXXe0Tql$;`8+)`#T@>>?|FL(EPqlwLaD_-?qeo?ld53AN>;&H8oRzF|
z1i{jy*LiOKBl}LeRm1Gxh9o#;%QN4bJ6g!D-EX~mJxrg2|AZqDxk0nT+*+`A4oz$A
zWKenC#+`N_Cwr8bSnqdM^K`tCEToS=MT>&Z0Gt6j#7iZxLXFQg!_bOgS|I8Zu`^gR
zP?B3!V{T?p_q6?TgYM@K{gxIfx9)87_?3mpBO@u|1feJ~6){VqQFFLKcvFpIB2J8x
zhptT)EfnVVG<|vD31nP7Nuj&4khxL2wOX!0fB_^np1g0eyJZ-7ggxA27fnOf?xR4%
z5PyHnh#6Kq?L(&nOG0sJZN?fQIBV<B#Pmm3+LdPYFRI(<p>~sQi(3HD|5RoIc^|-0
z;oSfWX%ktPI!j!&B*3$d6DH8rg&MiUk+!nnu;vNP7jeR3l6!LJ)X&P?9mY+^&-Z2-
z+a(ea#jvraEE>Ex68M?tU@^^TVFDPSYBCNAb~4I35*mc|cNN>p**|<f{k?PR%C@BJ
zhDUlzx^Aa7VoTigz#@NfKmNRtr$NyA7|xZFZW92yk7yLK1GuAi!DP+}1Ob4NFqZ-a
z1c%ua26qf`9zK9iYvH|6L)@MYlAs~0kJLe2{AcP7gWZSsn`JFuiufhiq4MI*9ID=U
zq+KiNk?=o6^?=QW?uuW96#t|;IkN11q4Tx%-@-n6ZIqxD1W~#MNEf{PW!BMmZ_BwT
z=M_N*n?MR-dUzFRiw0!dei4^rG}V9oK~Jt@{`eC>_^2O0sd|F-fv-~qA$wIizoqve
z-dRH|qVs=Jg<Byh@)~45Kk)XH(NC&|;*p<J+{NivsDFQ*f4Ov0lA$v(G6ct5_RNp#
zRwAY+T+J?(*?*xGTd<F(Ju&Sa|Doov704zfb{Ydv2|E2F4)2@1iUlFP00EU$Hk;Z^
zP$81H4R<v+$yus&W|OI6yET$gz*($4@fbzAgEUftC3@)Bp(1edDBpNv3l{B<lj_sw
z6RSCzuI`>Vu5ccrK)(5t>SY}K>QAcD8^9cW@4g^ek<ZCCB%_et)KXY(>v1qWOksHY
zcArmN;2svM*e+%sMN=m!{EhVDf#$_j`4h7jNP3o@FRNUEXe*T$c~IM6q2Nk!ngm6+
z^FSW-@E5!OWi^F(lI4%yD_P)>=i$tKb(dzLsx7>d<wTF79w03JvGcLeP(N1Hz3O>S
zlpvFk*4ddhoiJaFBP5zvk{}U>K39Y|SDa5!8?T~nskyu#;|sOiB3#qyR7u@a(R+o*
zOkA4Pli7HDt0a%EGwqO=$8{J~6{72I*R>%(b~qmkBtrr_yA&%z!vFJc`KZq#)$YF<
za`VE}FY7Adk88?G>fTA&U)ELbFH79Ci{)6{uge@js{O%Nq?D^s{r)=t&z|vHI{$Nf
z_?zwM>F>tvzKa46%MM3VW&L#=M?L_l**_8``A5e*o|O9qPme5w|9aHB$s37RQEXST
z@Dp_kn}n3gi+NJt&4Yg<e)V0HTLeD~LX5&yfmDhvkV>H`Lu{$RdcHCqa3jI14M3+l
zd*4ONk+2F0R<bK`Sb}mQA*v?B)$j3`TinUS7(Sm2<$^^UYYn9!^UVn+8CeIfapP*~
zAhBeGxAw01b;F5llcTg|@D2-*wy_cdAf%?Uu}VTX4>lMOWY&yniFqrH3vjyEUokCi
zXC2o-U#1YIZl>r(&GcU8)HVa1IWg^RTMk*a9|b+Be;x*DH06w_E+U3a1`Q1gB4k*j
z#mX8x=qx)d{iovEU$qN<u6sbzfz>N1&M>mMSo$VJkS`Z5v84~Dr<T8ixTd0g@_i}?
ztSvsvT;+TOc1gX;?{Pzp>y(0mtJh=8!m&=|^nq>m0`k0<<(P@n0r;pIp*1$K17=vs
z;_7=Ol_|4-7d5aZHjAnZ^;!1hcFr+TcUPpM(JpX4+$+eRu;?tZXyIwrW!8c^DM^03
zf9uNfV-Bt?4$X^O3~<6Mm<tAAWM|c$B}8oOcrcDO5zh2C9=cg^SnArc<M*+MejA}%
z;wR_0LaFU(K{ZWNL_30AYnB2Y`E&<KV*%4>EVqIwG&P%*+PEZ3Cz4d}CR$oHaT5+s
zsq_fii(R~T?~%K!{lVKsaCW!S2no5kU20vII=_}ZC2m5bI0a2-hjA`GCEr8y4b~xo
zXT=`Uu?>ip_>!))msdJzBuTL?Tx`M0#b{j&U`)lxS|1AiVioYkl8^yjx#0@Msq;kB
z&ZfhVxO_6WlrYp%QVvyaX-0cs)3xuw=Htvn++{jUJ_r1mK|}5kg?uE?f^diGGj{Bp
z!`TXjzF4esD&xW_GJ{K}$QLvaQRy!xqMa5U567)54}e`3GW`Z29=nrJ2GqWWr8`mR
z0L<rHv`Qxqba=)kn<$nH0E<NyAdxC^gg5M!Asl!e3J3)`=Qb%lm~Ak;E?8^cQc}Vy
zqruY0M(^l*@pBtJV<!r)<Uv5TsHotJ<4b3T`jz;LT@~2FeYH@UYxAi~M@aLet719@
zQaZgN^iPFS0$%>4D(d+DaIw?9lI(Gqzt4WfDg*KiwXdB<P^l(K07bf2KOm+n^(mUU
zYiJ;s^Y+7Q075egKH#hKXhLwnU4Vl}4VVs=<5N$kSA(uCFaKQIi!7e2iI;MW2!d*V
zPwJm}Z@kTb=1*jO<q+%SkM6IGw)s98*RpdR{fA$A<qxwJr(Zo5w+A@~bPhEFa;6QB
zvdM`Wzv$^vAF#zb5EzeWq<Y9~?EtWwCHRj^@Nb*v?+^JG$b>Z+3M~Q3fGj?Seq6>u
zN-b0E^4lwruRWj8W;f9K%P;bW#s?lYR{=g6V*;>P2Io^^tMi2Hf1Le;R)4i2G|f>Y
z<?I-7U`Et3Zm|LUZ1Q?VS~F*~E)teBhj3wGcjwyI(0L`(@Z#<F8=?C_fm4wP=c(t1
z^}y9oajFWgh~Iol;aMqn9u_HX@PBmK!Jte#Chv8sv`zcgdr^DvmC+Z2)oH@X6kCtf
zY;uyC9NM3iJdRNW+97;9MEqjVzpO)IH)1TP4G8cF9P0k)CF`<Ig-b)uEg!z*Wg6PK
zVz|;h@nN)b&uH%pm-~rC*3~h-k1)_)8Id$AW;QeYinb{zD1j4Zy4Z4E$0t+O;pO8E
zK0!#Oqv@lI->$HGgEE}y@7`6wtc*{QpQ1OV1ga}3Xb?pY9I!X_i<N`^VBxF;|7Qq~
z9v1SQm>Ku956+cX-cp?Y*>PC7=^#ITArE(@cvvwCnkx6GDAQ5#{qv*^(!Fb->(ZlL
zByG6!pb9%o(W;UN33#_+W?4L(iQ_6vIO{6g+9(&f;hr;aENsR6gTOiOS6?qiN_$zS
zVKhA2QYhfX((CZyDicIg#@2es*ix|af3c<>0|82p7OU)7$eFdeTZjS}!esDS^Kg|B
z5$c|cVp5-jUbo@Xr3<a{aX)Psm>6!Gt7>{;!|soZ*Wi~)tfUM%elO-f%INR6gJ0zI
z+e`jKPM>A&|E>ayB?R@xd^y2ktjR%{#3jUQQ>%V~QBozu&mf^qX{5+`@0znhWp%+1
zqP9^O$?A)C;Xz0rbwL07t7j@18t9(m)Eb|tSnmpCUn3p^$d*$Uk6>U*Y<yXWC|qxZ
ze1kyU{Me16Fi2rrN-bn0^VdA-eOTS+v3;f@2{KW~^!)3qvVezHj@Tz*{wJB{0FprD
z>iqniq3-xj0uNcS7cF2rX6u&4IFi0~%JPS`%OU26K<deh6-=o5aYu74j_X?0r`L_)
zX86d)EGf4K1u%9?Z5Z@nsZ5|>@gA2|jC1?-2BmX0@4cIAU!9say~h!8t(hANo?U!i
zgub+53mSq8c@|oj6OD9+L&j#Dhj+y{HXKt0HNn9PQlj6E-%t;|<E0c^fQnDqT)&aG
zR0DZ_c|?gH7sqJc2hw9#VE8UmZc_uX4$i`AV(FqT`MOYtov4jy?bzwc>sSXZPax6z
z+rvj%s|fuTL5%cvZ{HAw6XtJ?$vRwYQ0TDj`qWTd-%|2HY>xfKSaF?~h=xy|M$$XU
zm}d^67kJv_fo<~yoadwSA#W<24ryuWTnQdc%^p1*v((yC0r%x4akT`@$Z^CQuILMR
z+iK1BU?I&_S$x@A!yXdo)6S+Ll4X4xzN^bpp`xb!OItP-hV7zh_e8TECoG&{w|f^M
zR?Cytgo!`DR0i8PTx?^8>*G%*c;utAanCQJQmrzU=aOVUM{CRFhRrn5J(js5JaWu@
zpDD?63rHPlnUZ<ZCVmtMaY4#RCH-XI0VpXe@*|Me+{-_Uq@O&7HkY<JuO#nJ&wj0m
z%jMhU)9#<rsMM*l?wz6f;a*2?KmavnpCx9S*yznR+Fl;m1(mnBT%L?Xw&!(v(g)Rj
zS#Y!Z@x7xq<GAjRanHd4cd+w1mnI$vZzECb)Ykan!kplQb<30Q#X^P~Hz0z2NtzqH
z<3(l1Lu0>QHJvm%G2js*9e_ndikH9@XBin*WYgEujV-zNhWV_jQ@ewnslVvJSDY=K
zw-56`pVxjOK$YCrzLK_kNKtucR+Wkg)8C5TDzk~p6ltgtjP!l0pATc&YuVwtx^pBA
zU8uTK6=uhga+@x$_)8Oq=+`%4-tFDgaAv%nvskV~(~+$R^#===V-JSl)lAra$MhjX
zwc4|B-0|H4ri^dk<MdvLGXTm+(|41443^unSA$M0kbSl@VY{dift^^l*(h-eV|L{a
zxa=aBc>HOS4%NdM^Ljt_sfrJYIYfjFdPM0uPM6W7&v5|g(0J6$whdXU|8maiZq}VG
zHNG3EAc1F*sb5Q`0iMp5Eg*Ikhb3)HO)YS?Yz2nIg*>eRHzF9EZAB{yJjihPZC??K
z%ju)yzO7$Sm8tg{Egqth2s!Z@XG!nx(8JpV$m>}@sYrn&%Cjzg3dIuW(!27F=uHqF
z!-*<%Fo2Bkdj+|+N8$JSeM1?sVsqCuB$FPbnvYt_+ub+mvF1K^bM1<MKTxK&%WlN=
z$BzM5Au&tVKumvg9tlu)6{ic|Z2bjzoUp#SdB`peCLI7t-L}HNg|UB+`2SoUWn#O9
zJ;b}aLX@wA+R`bXUNxtVy*w>04@S41)mP1axybGGxx^2LcR^nqDH+v{2^Pm1=AuWc
z<&K0ME?=nDwl{t_H)y)Hh~^~9suOR`gwAZ9A_^Jx`*?av_g0eZjY9U!B>3Hv<}*?S
zX61NF&9m`&WELL;w~HZMwmw^rfEt@}+Z$o`>_{vBOois0J8+seRH}b_?o`79+6vFU
z&P>pzoM#N3V)I3C=oSEX&GSkQN6&(1K%cRBad|~sUG%)#MYr!0UWK!{oL!sc@jj$6
z6AoEjgiN@hUP2_?p!4~OXF~RYd`ew*usw1=e-Zh5dmK@g-()h(W%>nA8!fG@R#6+r
zV!eVEu0i{hr3*%zvv*^HKN<Xhf8m$EpfqwT2Ywc>5rHJ%2>L))7+f!b%jUMfT{c9x
zoFpm7W`~GM_c3N~s#x`olo@EPVs<e==!ddD5f@a8utFq}((GZvl<ydZY5d}|3T;Tl
zVRxq6jPGlypg2XVq~X1;%Z@Gv?zno3j#9_V+KD}PONG`J@pxiTlY%s+DGs7ji60!W
zc9G>Y{XA@zTBdXAb<3b^l)THagHHL0R_{>Tc{Q1_p|E`)jV!V(-Y*6p(?}o&HLB2{
z`*vQYEM6DDYi`|Lln_W#Zc>VGD^z59ERU<n{&lh4F0;BYH8wJ}BkqdW-vZJ*2?j^?
zKxx<Q1t7tJ5JM1mfvfg+k3YQj@t5`a_ngu15&v%~#cz51v9tt`$A2f9!v9#i5hp}O
zXi-QApgLQ{|359gFf%E(oI<#{Z7qRWrQD!5EFTEegw#?EKFbKuW~NhTHwe&dPz`Vw
zIM#XX>qFG)prNO}=ZE2s2ieBN^KWn2r#nTU1zZlr<FnWY;8J26p57v2aZS3^QZb&q
zOK&s@U2Id6{H&(&WpG)N4AD&$!mL1S`+BIjnzZEA?lz-*oJX)jZ~&RU`uM4hrbQn8
zxpH}aj=2XKLxUy-(&Fw_0HpW3jeMXB(L$n~mo@-Drd2~M7mCqn=3$i@CNNY*fy?Qy
zIBw$xK>GIEl)kvw*m$u23zG=V2URu1*B&zoQMPAmdgloy@fkOCX(ozyWbc)aSvN@P
zpUBZ_f+k5_*1vaZR(i?Uc%7AXfI>U(;6BU+twj$XxKSSApTpVrzQIjCxgnU~03HWp
zTY(zLT9*0}k(%na$zNP+q3tcApy`<L{d>bJaEIMj>BwntD1{4elNAx(#>*XA@so-b
z9EELDEst-~N$`DHneB)|kyMW|!9fem_4_K@p|$gedeI8fQ{>Hqx3#Ex1gl4<zIbG3
z2-i{8j<QITmy2a&^Mvq<rj)LP#(;-t@oZI)`T)|gw&KhEfo}qj`_(VNbLz^U<b19@
zjBMi4XT4Pnf#u>%!#~2ZjlcBmPq_@oxVg}jWCn`QMkxAMr-b@TZuq_2ntpc7FFVKi
z(}KSqNs+u6bhy7c$mIn^WAcg*B))HFpRxk{Hd>U>3S0(GJH7_M52s~vXn=UyqI$OD
z*4lLCXl=k$HoxEF>$MkVg#h5dABsdOoeP%I+$3sm0;&D~yf9?+T_}*3`~$2TfDxQV
zMq-wffk>}Cf^72lxa#jAEq|F@Os^1RBTL{~g{M(*^2JUA>p7w$2J48dS+7Oq$4y{^
zRhvwfbVE}+-FzRo%k<Q!hCY5giweIEK=L!WjEQd3`scK8V(e-pJ`K2|=qA09P4L^o
z_xk+Tn(ykXj)5!?jczlWhm1f@fdd4r4=X7X8|>7Gn*YR7R2fEgw-%A%C`%SZ^8(@h
zJs`Xvng|^JjsL?-K=|z!L-`wR=f5ux__tj3A8hIO5TI=0w_N_svHbsO4F6pc;oe(_
zww>(Fs@($DtN)WReej>=djDrR;{TcV{@Xwierw?0VUPczhFAY;FvxEW{MNuf+gIuL
zKJZUU?w|4bfA-Grec;ce@>f3lt%1K%+J83JZw>sJRQ}3mzcuhzO8d{|`mKRKlgeND
z?6(H~N@@SuT)#E&XHxkqpZ(UrUn%WBo9nj*{!A)=<+I-!_$#ITXLJ45z@JIwuYC4f
z1AnEo|7@<`8u&A*{FTpsYv8Yx_Mgr5TLXV4mA~@YpK3s_`;bHYL%CVkcg>`l`}h*`
zt37&mW#@MaU;eXS?LRizc9?{y@(jv6B37sH8td?yoQaTMZtx{jPz7&)QhogLld1$m
z67D^j`M<+-^bd^XCTA0lgh=Z^-|(JcY3*R1{F)jkbJKg-wQ=tsexEU>xu<pc)mhp=
z00??IC6Y{9!4QRNf#P0c0CWBTC<t@_{-_)xDW^98ddeXX{<S%x2VDr1hz4OO(l`I!
z0AA1(w7vNW;4I171R&~v1J(Tgpx-k3ZAZVC$Zt3Df1rot{5uC6D^2E(LS~U{YR`g|
zHD4$5bX-T%#px|M<)ycbnHVUN$(6!*luZ$L<2+j!^v-p-hWE|z{-JNvSkGzMq%gxf
zmtIaMHZ19y;o07#L!M1F`%`F%s1NlK;WF<=C>M#5(p@Q?bPw+yc&_Zz)6J#WF@Ad-
zVtTw@5tr5&%s-nNH>*$WHX>V@u2-68-sCCsW-0ljRZJuoN8d~Ck1P4HBmO}%T#B0~
zdS(Iy-neYLtbeG8?p%2O-PYpbEZ672S&nt{cHP@ZiRS*Q35g-(V4vK!k>RK;f}BQv
zO#p>soJWO-d3d`P{Q#u1%84&TXrbHwA^qX}@W<=aH#}126-(?*R3klWQT_-9#B@##
zC?2jcZ6#8C<91aA;Z0IQep4>-6OV>*>C5B2A9C#ZRNWn@<B;ORm#jFt0@<GdwD_A;
zM90~NuDv*jNhQ7~VxvxkCM#D;f?*=1!9%u(jw$uwh1ddv_@R9K*xti3;)8&hodL>4
z*v5X-B#22cZKklx^!kJ2*De{5KN?!}9cvWb(qyNOr}1&|L*`A?b1zUxa$`N-C$fdI
zvJdZ=O?=O!m@3Kc1>g}e7}tc*{cPqEiZHRZiC1=HSFxa?T57`3V{M&DPTW@REoWfO
zEd^-8B2$55&m?X7&QGdXP(}80H0>mhtYJT#Ur8qLuJunUC^5e^6(~yo@x|n5&d=|i
z{G8B5fGy%jwH!@pvN)2&Ow_+O0M~kJK`d-J1cfQ7!Lbnq(?^nrCk8$w{*d}q)GI~M
z;HR(f)QR?j9WNm!?<wP3^Ko5P&78DRN?Oy}6ISYxx6lwq<<}O_OZwtcM&8gHcWyuM
zW){53v6$YiXI-GGFSfpcMpusb*IJs7P3*-Rj+pJs3whK~v>ob47-xVcazPfvWSmFD
zq0kc*?y*&NfOK>`i*H*Rc@E)pIa))cNdQ`geMt2A>cG9heJn|C-D&ST0h1uT8D<Dm
z)vm(#4%OKN%^+EvE#v0AuGLf;ELa7+8zo$zg{8=p${D`3j{$fBASFbh*^wqqC9V=!
z+D$y@Ssj=Edq!sLgiBBGV!O}Qa~r{DL*b`Sk6-+X#a!d7R~g&8j8{pUL7Z5&R4>IT
z^X@kWHe+``R5J`RL#!t|PNlnrD@kfxg2$4>FI@PFL#hv^j!%UZ9C(S>)GOE$N#pSn
zAvWX*<q-Jhd`TB~&oi*Q2T_ZB1K(r@0KbKRwuBil^lW36{$3@(({-S?ycc^w{k{A9
zTH`DjHcXR1X5rjo43|YPOK(Q*S&E0NR8DgD0UUNulfS3#?PwbLv=kdEI#X$J<T*W5
zD&OJ`3ld}BJA;QrlC@@_>OC!4k~ShuoO9?){dEfamBsmE6L(2d{A(1q0`TOE7pEII
z<R{CIV(VyK@Fpr88O5$p;VIK#uQp@1_qyth2OnHaVGf!@CW)^r=k;l6eeX!v&jVeZ
zDj`3qJlsAM<`flwU$4?9XHAhD4wu#!FjPQ^WS7XSSwfw`bbUI<@Q$Lh$Fs#}<XgN|
zA8f7KTbDN2<X$O&JC^ziPK2GNm0skPtRd*HA0z3)P4R3Avpi0XSf}th5KnR1o_MbG
zYszpvL{jlJy<Q#}rFKm3W~Yz9NH5}S2%GLSkGAQc1l$AH!wGTalXg+k2<B>W|9Fj!
z3&t0HVSDJ6Ms<Rz^j8<~8DP0%#`tXrlNk@I(2Rua$Qjg7zE-YADW>A$?eDP-nF7yK
zXsFM1cAsQ0ewTlxW!=Sg;=ln49Mj#bZmH`D8(K}u034yYh><yXwb1(YS;iQE6amxN
z^Grb<Q?Lx;sJn$@B&zv#t}0eVQyP5sP3x*9)%{Oif8H+772ZB_vfdqnZ5)`f$!>{S
zs{@s|wlz))SKxPBhI4UNwPW6BaSulqCRK;yq1(?54lYh7r|;^Y4r!xM;UVj+IL6pR
zxc_2H;GjNt<+e7L#n+8<TB2I>krto_?cy5UL~3avvLrgjC1_Uhh^vKi4L@0c%M1aX
zx1L%$eEC?5rdhi}J-p)z{`1Vc*Y>8Hz4sSh)-UF!mQ`)ra!QC?Ro)K?Gl+H-nX#16
z);3%8XI<v~z6G=z3mHyanL2?Z*cQyh1ieab8MYJv<0IiWx1l8`;*VR4&A+mci6swj
zt3Ta&5IsonUoS3Lh7x4W@%v4=IM4~Q{2;j8cf!`j<<Ul#aT`{p>rwCLmv>HeAJ|Qh
z<%>suQa!*>PdO{B$LACilkW}sPMx<TW*85D_)@7|SXk&M#+0JT#uDpL#&lUoInbCQ
zeiRMo!POhJN~+;NtT2QsA*J6I7Soh2y>sn{t2gzw;>NQ)RY4Vm@STat+{yAyA|#b-
z-^>Wx17+%ycANM~b$4s5yITgGu{(s%QaY+bL1~+LWfGbo3~vdbNPt*+48&}^G7@r$
zM2~dyWzvj(5uC2`g&rnO+u?QVB2SkDZn!X?GBQbD8E%ZO#e-Wwv;eL0s7qSX4Y9cS
zvMAZs3p3P>(~|k7RN0oRzCAw{zhdXxnuXXcTeBJCd_?rmPGU!U&a}9=K|yKkqfK}2
zR*oOH{xbWvNdGLQ=dgw{Fv&>}n<x!$05QV(c|&K~O9bOKBuX6amFqPyb}(H&<u5${
zD&<1)19#tG6sgz<LC*J34&>0zP8{8V1TRlcAhPFzL24|<=`??GSLEI+080Ho0M!8Z
zNj*m#ZZ>-1ly|cfpqg^6(7N(}{lA->|7yVJ;}BpmCHMfuGI?o~R5fdfbrHhw;xuEp
zkM6^sCIT#J({$v~&K0N-M#i9U{D$zi{!@8!C+MF%ogVR>^&sx+lg?NmmhDYu5tTiS
zJVCQ&O_<Y5foG%GFnb8L<+Q_RZ?<kHW?Y%5+iEmDGwU)RAiz!KYO~3LNtfMOFe6JK
zZ!NpXjvnp#k@Xr$ZBysrH*!`EG0}DMK%IJ1jYC|aPE|P(H)j<kd*!%`K>zG{jWD2(
z-t`m5Di-<D!hcdZJJ;zFFfAZPr~Y;5rQSNQLkYV2<a@!Y*X}3pe!PzSmYXQJ`N|vR
zmU@(faD<+LjP~I4BPgtJW&B<OP8XrWR@(5++|s4XuTjlG<HZSU##eD&p|(P^J)>MW
zNP<L&s0G$aBc}}qip289Jou;`#M|z^t^-vrKQ&%cc*41p6|0IkSKtN5im?MKHQI7j
z^w%yaR#Oz*NBGy%9LNWE?F_7By6r{ajo|VGWHkITPMNk%Y$jNE*62%FpYL-aa81<Q
zsdWy7gkPiNtwV0J7jQr&VWBW@T;y8QtVDE;X;hUZKC>cuC5t)v9<RNQq+RnQ<a$kb
zloI+=6NizlryEHDeurRhiOwm2L0a|#hq9Y8Rh!tvkZxJL=6L`Gv*y~v82w5kfxAS9
z)BroI=hoIxH(BQ>5`AeRq%(veyFn&<6vQS57h!xx*2)nZKd$_+{WB-~3rv!I6xug<
z;uwdVfr%->Xny2QT5{bGc`t56d&;<eQLq2;!hjh<`_>pCB`eLfdqSl>d@qiO_&|g-
zDP+ee+07tnCRfvK&;`61#cqCkTf*9$X!h^+;>YC1sk#+S^P|v;aA~}Cr2#Aw2YM!T
z)dQNkH6*R<jveozW@0lwJL|>GeeIi^(nv>$AW@qTqs?T#JV}=#wVFV^e0Zw{@52kS
zxxKe8W-1RwDQ4a&y_QQa@uEm(mR7n2e);G{h~Uy;PpG((#`IFUGXIQe-pr^a?ywud
zZ<w^O{4UD#>Rg*Jd3EbaGaGmGr+%&OdzON0^(<4foh;e=X(?N|{ui#+Tbk{{fgQ#@
z*)3mWyJ{kYjo30R!M)0~tinaWL@M3-)B?)99G_NbHQpW7_T7N{v#o|)8(;R!R7}=M
zLIlP)$5}-RzgJa^b(iII&-~Qj>3wzpRn;1A@far0%}GpX1kM>#&Dy%4Y4uEKPScE4
zF{mNrQlIa-A$Db4COaU2x=3`=I3rKRE_H%;jXO!*+Y)me?3n%BQl%IV(Ix;#3llf?
zrcuX4>8sN9CzXYNtP4*$SOuxqXA1Rnad&ZWbjsgnaD7*OU`45ZkK7knAIS*W?h@B?
z3jj#b3dS~}ki1H#Nx-q#JO87cfq&IFFACt+6D0uR)X%{fN{r<dJeV96NP83x6lm>O
z_TA)WZ_Q?P4LRShr0SV#oe4HV+c0W|-#b;jcf<dq)+48;<L_GqMX8TV=3b2c7T?n}
zbs8-UgJ_}%Rq;?1&u!C)g{fv$(_yJD$Ti)IYBkxusG{jshojE?Qyq2}OQ!A&*IoSp
zTf9e54=)<aD3%1at%?ZPuzK8Qhw6)k1@y6}PDsAEMHP0cB(QBS2tw4XlP_}1KBwe>
zYs0UPV0SugEOCAJpM96S?O(*!#}TDv?*6hi_HmO~wfge%6heFonYrG_fbl&9WAb-v
zmaD(PWsfS9_z7Gswp{&1n`C*>kjqK1ov~oWW&@iQPnM_!NTtOmYL99@_jr=|r^?ea
zO8g{CYg-sXkH8*7z65$k27iK?ZnyZ!vekS>XTC+5>kv9Y$rsd{m)awtO`DHAQh^X7
z4V<wOFPLW!czrflx;~>zC*LxfJ#~e3U@fJA%gkO|>eJg7QA%aPo)rv*(TDzPi!+Km
z%#GMP>%kwF(IEWpV#LGBx|A~$7UM-CS2A3$ikeQ`6sdrBwdu%~S3#AbV&u)vIcs_O
zqa);z73K6I7P4Oa*45JK`Q%A_H>v=bte$X(&6b0pWBJN?Jw<zLT;UGJ>E&x?kVwJ7
zHz6m*5T&DcK%Hh}&DjJ*v9m2bo={&Q>rS~$7_{`Qb;xk2uSJzB4XmB%rgz~?J7y1O
zQjkM95gqX(&!u?m+EgS6XI~dO4>)U`&GmO@>lik^GHKXyTr`R1D^vI77j~QQu9}`G
zG!1q&i$BtOpw0xxNaqR~^>iNbask<SRASY;+qqt^&tDA}#YhRt92_W#pGijohZYFn
z;@yJD$YK7ng(-k#9cWjcg<Yyxp3^5;O}St<@^BNkvE2hRMd;JLyJM#$q)*L!iC>^B
znCEWmr=X~1hWu_ws+ZCptNLu4)U@}t52a5}#BS;)s1n2rW}ZJP5rojWH5m`2i@dg!
zPTa_i^l>|{#pyAg%<Wc3+bt5vIBM2ddFY;<FJ4%}MF&1>PIPN0-MQGw`X<6kF~XB=
z_e=C`8%r4z0dmI=DxVreqhhBGyD6%c7&$9uWQ2j)_!h?0!b-%Ha%-;B-)rD0i(Pzt
zt?JcTfyuYWyyk1?u^W?bx~CNVm)+l4eIR2_y{G6-K|O#@LSY!@d1CqWfp4|gj<HE?
z5>}q9SkhAT!}`-RKFyNz=VligEcn$;4NtY6o>r1288fKjh9jqgt@Q9iDSIey<syZU
zG~8;ihS}#NIf3jWg)V-&Q&*Xf_jO&PH+G_x2H5^X*hb;fsNuSLeS;B{vikuFxK0eg
zLp^=3$MMsoo_ZH|Q?u{p9fMXqEdc_B`RKI|=G02o`BVvZSo*#eKBZ~gIgqu}LM$_D
zHK2IUQPDp29)IQ&$#3f~&d#QeyR0)Wd$eK?Hn&1<lbAM(F<c>84j!>n6#y&b8M5aG
zq8|^JW@gE{UA5LOz-%raD`7G3+=YtKqEl0Dd^T<u@Ni~!ZsS`EF^A*`)1OI*ChNYR
zoPzmXnn`)OVtSbNJg01ML4xVKX-D6SjHJ?<U7zvhx9unEsd3p2<|?8ZzAnKy%=u+Y
z@f=rClrooI`Z#u~kt0fj`4;~jO-Dn#l-?~4JCtUm4Uyfc02WO$?yob#>7Q0BH~?d7
zLPEXiudSREFz<VfWlv|@>a&yZ5YRC(6S{aV*G>Isip-d_HH2>I7#XPA-JWW9ss2t@
zKgvOEp2ZS34e{CGqGXeIFu~rQ!ikC1h(?X%wL;%{tBX}HwJNXjIXb}zm^0jA!jv46
zYH*}3!C{2VPXe5+Zx@&P=-xQ6L3prfeD6uyIzT{4bsJlUB^#0|ij?kTSa~Ccsf2vB
z#O=_r5C`RimFzGmtYAfyZP%feS2#^mB?+a+1?;#>MTD=wV;7AKkJ6=(=O?v;eUZdS
zorHVYI)f-4&enxvF3RhPPnvG9MZND9d#A4ULjCrl`<n6<OUmBl=He#2DBZ@_taAC$
zyCcNjSfD%5mc+JAQDQxBnh1Y<cy$miyJt^0h3~6U;eYA|ak&xWePy81shyeIeOvYs
z%^nU_2PUbzV)J7}p-3{%$&8tpO{SF))~cb_(uHy9p`2<yRgbwRr3zSZ*Mht0tc$h$
z3N(av%q2@W&1e8_frE-@WEespBItiisPa4=c{Dcn;&aUUF`MqaD97{Hg|4VxlN=>8
z^gdF5!L(`c-~%juyXD?S5;%Qjab4x1Z2NcBb+7yM*Ra)g;VGbJkRZes_^6yzu_SLb
zS(S-iwj#`7dWN=D*pJ@sLbV%Y(clp{`%w&jlf*e|TQ)Q^WVo<gY-h8)7xOS`Ua6G4
zR9e>*#_?kIIvI3SpadpZwLT_3q}!yQaP?QPTIrRDfNQaRE^STVRPyIKuk`xX>r;@|
zR*Vq`0i$^2+Uh%)oqm1a`;j`eV%S|Px;XK%#>12q9+i%MJJL4;yExukRdc49^Y+$H
zg->@+Jqcm!nhw4>-ZKkff#vX5vYDpV3=N*7)4WA{=7lQzt100G84pRF<Fe$g^aKf+
zTC2sqs%@g*fW~$dbYl~-1@w9n+#Z|nfX&Ws<j-Vkx5{!tp3V-##in0GHU}oJIMTgp
z4;2@%vht)WT9C8XfNrVnf3AHl{TlCYOpu6DQG`P<!{#c<c@twUSVODT{TsKq8JwVU
z$G<nfD3nY7E;;|^M#cS?_qqm%J$u1@WS8PqSueNrh1VZop?10m)q#c3N4rm+^Yi)<
z(~ZBomX=0n#t1(dZSh5O<9V}Zf`eKtQSYl5JaLL&km9*1FW-CcRG4<~AG<|Y#-fzm
zt~5*KP@J*aG18;o6cpgR=~ag*$sRbcY<sedK&celO<QTN(|v_tIc+nhgTNdRbUrSC
z(-~?#PW#M0>9oL?H)5s2{j7}1)xDv$b*030<ZHN*!$+d{aF1F4POX@G2UaZb2Fsqm
zWEW?BbAKDBRfQw13oE>r1s->`F@w`-t!V}-)EJl+nS&YT?7WAMRiRJSNaw#;9&1gL
zs2RiSf2i~N;Tb$ef>uFA{3sjX=aNTRAU&k7e8BYYtl#(^ripnh>TbJ$wP)a>=IbEE
z<G+m=v$WnC&dW%X$X=A7j3B^9T3+6^DiD&xPpUm2pfU}YAU$-!PNvN=>;<xwe_8^M
z7BCSdniE%LWXmDl!+X7AJt^n4l!lrL)ny7rj*qrj4Fh-Xb^V#|lM+?L$Qc`nretRY
z$IOqHTOyx_47qrDi04&C>6DFqeQ1(&y?)~H+tLQQ<e+9q`E)sOyNi=PEe$#c*Cg;(
z+1Wkn{ItTB%9LH&BQU(~zPH!cM*p$D{HicZnIKNG%X+E_6Pi8}C*8xdl6{W3<+M|=
z3)cNLz*=<r@=@qe%DPb2j62l;rY~nyp!7ns)8%)Lp2;149(vjH1KmHV#45%+u(Hb@
z^|pslw{#g$<;l2DFsIg$z(;|gC|`YgRR>Q=qLnS@hc|uJ9eqU!zTAqr=Fyr1A#<;0
zZKhQ`Z5A6b(dh{-5|Nhi7zewVW#qe5m}qw#?#^IEZL1m$hX;R*SAme5$5RDXkHiQ%
z{d_Wi9jedQ%?k58L~=-4asSzf<(Jy3Xxu=DV(z&X`txH~y8Wr?0-oS>SE$(0oMf>w
zqMJ)`K;Vbml9j-TZC8cV;!g+1aVvJ5>Zi^5qAwrgS37oIqcb^4OW5ymx)mzFBpZ*=
zp1S3;O9pq4tZXR=c}h11q@*@86`5K_EZ349-lH%lv?%~&ZzXjG-iYfF54@-%PTxOf
zonb(FY9B>i*{|7j=vY9y<3?fV_bnhln2res-HU8ur(Eq{7e(dzR6WM>NmXvkS0#LV
z6Hsb@CFNRkB=n0i`%{b)_%u;pJ57fn_>;>3?BQ7BjWJiIT!G<hufI){vTQ!^zWDIe
z1!^iUU4U2Iv7T>LYk18U!vafd?)m6!rotJvX`J<=E8kr2tv1U^xpQoa{DN9f<8|Cp
zmBLLxQrY5I;8X<|))>4fyGP#}`+>p*B)#XVfG0nx7){z&Ar)<LN?=eF!(x+-GU!(C
zJAz3+GAar*>aqCvRK>ztaF+gfY)jiA<@F(3|BB+6xjUagi=^mt&9)QvX2)<6p*8GL
zh?y%SlgV5VSwMv&p)t)?CowLyU@ulFKs5@{Z$CE=gQSIL?X~tvkwk*bJ1@+{*%?k~
zQd6stjo>w7heMQ9vW{;B0#pM=%67Dh|F7dU|5+^Oo<Y7E%e{ZL1JIwnM@#$fA~ME5
zsW@TD_wbcTY3csYP#)}h4(Czvabj=0`Ce$f$S9c6&82$WQr1Kj`=W;n(N=@_Vc7YT
zN}I3vARzAR-1M+7S1gxRPH_(i-?J{_gd{6g66#07lnma&{C0PCO1}*1Wp%+M#QDPn
z?}Tlr-zyfr5fFOH=<CDlwk;5Gt!ZV5!BHfcW+teZ(NG~NKa#I*fbF#5%4OZ#eILf7
zsXsKEIV!Q9(Q3V6Jrak$K#XbJ3%TA$2F(z65{{3;7CI=pJaBoLoCo4>u*NE$G5Qgj
z=?j<qOeU{brD`aeN0=@p-8GC}THmT;r;JaKYgjvqumG(V%|O>$aH1Lvv)1%(rN#`P
z=G!mR5tWd3aT<6u)?@?=p>u|Rd|6!oa-Uh3MVIf}wAOrP_uNIp_e5k9=xI$P5WIq`
zp!@1*K7J?<!q1w9Ln^DbD@Yln7VR5QsJGWOf_An%zvHD`uCV*!C%yK84#)tLc(oP2
z>ig9;tRe(BT0(Y@^ea9hdsZRTu3K8O75KLB?u_><5Qg1Viy|@@zi))c>W8RdMscsK
zYq?fqrE&z$7sw~ml-=ITQapO&69c1MZhKD#k4NFRCik_?rS=~m;z@q){L`BlE+7gX
zctciX8$Lc%l#Cc@0UC}kPp%H~^ST9KD7Qkq4lE8qj(rejxD`GjL0X>|_HnZ4-KI&{
zSD2e&6`pmW3gHyvoYI2Y@qf{j|1Isf1!H!p*8P@O4rY@G=ghN<X85*erhpaH5QLr;
z`@V7{s3hAt?n?yOUwr)-`r?K@XLqk=@26^q+t+?H=vs&cjuAHJN2+l4dA`2DvVFF`
zJ?gJ0@qCI&EQA}o)M_JPFuaM1Q<JJ?P}DY$9q8_B7rY^w*!ot3C01MRf;D~8v4UBT
z*eq^5PuEc9PMR!d+{Xc1HxY9!pZLS=HNP$QhaWZ4xh`J6f9{RfjqSYbhA}YH$|QX?
zFjYufgXz?H&dpo>ySns+E|wPk0!*@_^EZcID6c&jmNjk)aB{wUGkGtbXvS7ch-o%^
zOcdXutXlQg-8k-w^-GY7#qsz^<4}WIq)Kz82*pe-%?6J1<oD7(13>a4Iogjru)Zbj
z;x?By)s6c}#em>R_<RK!M~33tnr+$33GoTjfIDh?4`b_WlQG+JUY%)2(Vg~L=+D`X
zwS-IF8M#hZRzkQvUM7!imaO_5@?Al^Fb^i%ED<Eo*tl()8Z6UQCoLN}T^ZRCb$4u3
z-&3mZPQPgUi7bopjoUQBI#c5>Ng|Z<eq_jads<EZB+!Bb(H^|oLFrS2+u#%STze@!
zc*iL#qF*u1Vz`G{Gr?8C?d6wRBYS8PSq|tag}5r%-Iafz3#QB12qEaem>@B>G$Z5(
zAM!&*MSW^-<XEp5xg3^psPrw|NvyT9R^Yl{b@rS@>ere#3;7#@`zQ(mV6@@S|Fl{!
zx2s=;>nDqkZ9)98CWhUIZW*;p_5&MY{8fpDB2np*X*b`M!WwI@Q6)vfj8f!i9=FX+
zvB71DF?Mf>AYG^e$!Lh|JCh+nOYAW~A+8tM$L1=J_fxl@)B32ffMAI@Prm({p!iiq
zv<gk=>#Yd)n5iI?q<Q<hsf4N&K>a^xylXf;=D?DuJg%kC_6qmV`}|BLi&&vn;Q?Z<
z9n#;WrCc0>y%RjYg_#Kwuc{v^4rWWwhLqD7>fh=LkVzn^yVO*L+7U8{X3c&;KdG!%
z9`L;yEAY;4ZH>0``*^8p7gm4Gc=(VycKo;(Dv_i>)LYPx3=t_Cc|h_DFLb?el%`dS
zFRI~5x2(I~CSc&Q&)mg#b3*`{^z?F;{)?+8$eYuM@W7r370@!9B_b4Ak(S&Pa#~hq
z53AZ}xgqAKRnqzC?e~vQwFE=%=X>Ax);nLX#$G;8cAuJXV=0=RAEmHs?_yShw*loc
zhQqx9>i-Cw@*PywbFNmWdsRb4Wa<EKA1nmx-x_H}dT5!48I4TBmu;GizY}^$7aW(j
z@?w>(Sn0GUr)b%2wV{gC&1&x4b7jJ%hQ{lUpbsJpjl&!j8uT#-LWg5R%wmGGW8|r&
zlDgOK(O2I8*-r9TJfoGEMjSBhFp_A2jeQCAE`b{GW@&E`KU9Fwidr$tu?0v4V{Dzt
zv@)0R5S%v^T9}c|XvrK5PRcUIW;Cv53qtBZ4-~yVt$p3N$Eb{toYjchskiXiWO4z!
z^!g@ntsY)3sg)@kuvV2aV^7vNSLOVr%6uIw?Q!j#F{7wBXT4%8nyQ?bDe)W*4h{<V
z&>{`k9M3|Q9dm;eySF|NZR5`o5paIo@pbmZdG$2BkZS+>j$8{R-wUfBC81=*lB~tH
z?7pGa^9oscO$oD8{rJ<d3$}V^HEnDGvepBHi0t;ZLwmk{8bI%VAX-eUcnoGhJlZ@(
z)+BgD_+%&o9iz2DQSZ~mWTx~8EN1C%F7pMLt*5#^8Qajz$VCqe7l(~Q;f_bSLPWFK
zIN7GpOF#t9-;Jk3A4v@Lc@Wqmq7E&x)U>?1!6H7()yd8X67ti~m`(_jnLIWmYV|+b
zyUwVlx^#^)ir*kr0RbsOC{n}*A_l@ms!^hR(gFcdfnb1$7>Tr4P<k(76i|vFp?s7O
znuH#jBK?CvfFRPOgc_0%?)lceKW1h*Yu1{3*UbFj=Q(HXz0P^}d!GI7_j&hzU<~*0
z>=iB^gt*YShA^teBvdLQhebcD<$=^-8WFFh4b>`fEs$rw1=_{T_p)1p6?@AOM*a5p
z{B_c_tBB-~bD}ZI&rHMOx^gCvZ=ao|+>1x0tn^uuSG-7R;k~QmPA*^PCn6B!Nd))E
zs;qEp7)OV(2Zye&ke-ZO7$DkXCJ!{1Nb445>ul04^F6e%wEh-t*%JzitHOzpy`1?t
z`V5IMPnCYK-fvicU8R2c{#p~Ulw93}qZ#5ho*|<mN5dXtSO3VQ6kKNO(U@;2<42w0
zgTzJYhLN2epzP6Qj-*5IAPj#<L-7{mc6!jIvl7lKCMa}VY?B{C7YFi}^>a*u^5>jR
zGE_n`2f!JNUJ%2omxnOCq8iAqMSdv%V`NDRDM49{kQjF;b?42da}bft5%3lq58d{f
zsbhIsJ4eaN4{~t%bF$(qcZ20uxZ!>tck<^y6LQKIG#F-QbiST<b+e=6RUWK12j3P!
zJmH%mM>vaZEC6{Vhi;FY@#5zXkzm)nd2t=!3~$d<%Wi$%DQ9V(XV5nry6%UAiEF7Y
z#wW>hn$U^p<e}J8wbxwfI8h`t59Wo*_mD!&n|OS@=AB;x&Te(POjI-#fl9UIi8UR5
z9{Vx4OICb9V?+8edgBFj0VnKWZiRfxUvRZ_OsorG9d~H$?ct^I!Gu{(#<-V(Ew$+L
zLPgT(uZ}wdI~vPt)OFounc$IhQKlaO1rHjV34I3>q*5*NM&{F%B8G0h>Uyf8?tX6;
zK3niWOR3Ly+SFWIi_#rMQ>|VRY9ZkEBYzvUXSy3#L4gl7DBrvwKBVQc8KQ%XBjbB0
z8kY9ip$DE=nd>fVqn}jx`40@f&%u(Kxe}4rp_qY!b!eVTtkvnv)>y@IO%2#|=KFxh
z3b*smV*g~#3};`Qa~1|74={`l6ZoB(+4lp#l#%J^kgXSUV%5g%Y!fJqewvV)#`~x1
zll3hZ;TAbFJN{_z8pkh@-;$13!vP7#@qrp~jvwQa&59flgUTkqR|&`9M4g+m0iodc
zo5mJWy&jxl(CW8KX7Mi@Q<_A1g~h7Shk|=Xz7q-9AnoD<!?SqPoFZrJZHM29WxgQ2
z<owasIwqSBg*=1X<?Zd`S50s8;@5WqCTVZzzXYC8YThyrJQG|Uosl#TZWkXx-3N8N
zEovWSlz$WSiM#)UdvI0<GQ989hPae-crn;oa8B`35uk+gT`}-&v%!!uPg5WvMfc*z
zOC8rR=%qfc(1fv7v??V(cFH5THbea_B?KTkHJzJ(#oDkI8pT<DFTrlTW|L`VY`b>w
zl6(SP#E&oTKxK+5A}Kpjr2f_~Ew_PXQnMbVX?Xee%0P1gSUMjY3NH_}4X`gCtmwLi
zR0+Qj4tGsIqwquh+a*Cmu@UU$&A=X@8UnfHgYy5Lm4npMzzaSN(hzXXDi7-^3B3zd
zr(Xn_tYs(OuHuhTQ^mQF55Y$`*I$kik3l9~C>n4=Vf`KN1T%$d>Om{T1K-m<g`2uw
zcVB+`X2Gif5Rf@F=FLJT5Qm96Nqqr`prN6a9!!ll3v|CuZ-)q0T3NOkEN9nr3mK-%
zkSY>y5C_0ev7(P!ulHNpd>Qyk?6Ldi1pP465fz7W-^2qCeIbpX)tjOOlZ&y*nj4bx
zLrV|CKZL*?ElV^88Rwr$cbvLX`nEN^qpih1+F?R7jYh=ec{ph@ATQ|sqGK48zUch1
zJN;3rC0T6%5**n^<~h;KQ}d(J_sO_Nf`4(DQs3n{JVyKMdo*3HpG?#=7JLULyW<5)
zT`g@rm#__OPCjvt8@4f9<f-jPoU1O_*6wZicI9v|BqtVMtHm~hV{K?Nt<_1S3Ij=?
zK8Uwz2p|@63)_8<OGwxb#eGcp?%NJ(mi3#t=k2-J5+Kxo8xb;QwiIHYo0O1OcDOm2
zEnYt7X~%|+*p^^(*Y<m*-8gi=NB6Gk=|3Nj8@1_!xRUt*Hik>Nphx_i_w_swj%IZ}
zhy?iz!$6yT1cj4F3QCto4IDEAv-tF3FAb#UWo2G6ioikiw1Df?=`;&%0V}9c=kaS9
zN+)8*<f9$EZC%668|t{$f1Vw62qKbzm@{{}BB#GB@%}1DA(4~OjfkN|s1`K;jQ#T+
z$>qf0&32wIL)KjjF?vF*$XFJDf~AEI{_YbVj~<hw|7@NA&)V&4U{bmIecZ+rXO_L#
zL+x(tUtXU7S(E>?d;2<>a$Mt2yxayLC&$u6X4UWVtSZgQZ3+PykRFr?4=m@8MIT|8
z=XZndT^>_}?PbT!SAP_f04vd3Y{-rz1|jfI>%;Ah;{T6B{=-q!u~<HzafCKT$sU^*
z-x7bsP#>vPL@9o;69Gl|n}0e-E5tS40O%K^FThCSG$G6#S7sj_{i05W!jFxFllr7w
zQt7vPuT1HVitO`ruRCM*jH`@wrh|MTp_o+db3$iL-)F@i^ht=}y`#^F=%oHSvl-a=
zhPc6w(*JY{=lZ3p2I0J<4VhG}-Q{`j4hTz;0U4*@SmqTi$gKSUZh0EmmLgZEr~a$&
zIj(4AH%lDj%-#kPu9R5-%QDxvFVbuQDFYmX)tC9?+&}C_m3Oa~pPBbxXP47i>1#F&
z{p3;nHChH5?e23#zpv5nT<rlvkv2-U*N_Y$r-U3J7H<rdcOOgm_K5c@%i3R+*NpuD
zw!Ib0<z{e|V;9>YGc*XBg=6BsWIj2MBY}2#GO&R7#M~|q9myR{<OWRvjs$p4c)|Q$
z+dW(Tec#Y(i63>EV=orszu&E`s9aTaJ>2n1q^(|D?HmzibrQ(W{}nEk&%Qeh$Z7l~
zHf8-aEG<|@@O1qRH2p0s?e+3MkiBOY9?G7-{O=sU_a6Zh^Byw%Uyz}5rSR<t`sMim
zXwvGE=8wq;<F-ro*Y}^*1;_N;<Q(0%LXI?$WxkmTZH<*{xZ{UX&i#OBE3q5v=xNt1
zg-&ZF=_EH6H9jz-wS7T(dpW~aLKj6QaEGx5{McJxi@eqYsBEm5Yrm48b@gMr!@o(&
zA8fPAj!_@<=86H4hqT5+xo{`RK@4m^3g4O`(C0XufJiX!e=;d!xtM)Yt^W7}0c0R(
z=k+@JrQQ_k6+DTiFxa}RHjfRl5b<2P^(OU|e0*ieJhnKA+g(IJl9ES@d$*_oTk8fY
zY%O5|yM$)zONNe{d}6Z{B{n<Ci2Y!RiAdDAAh1c>!9y!*J?GWs)rrK#4n&zaHPk$2
zDK-8X3qM(_I8~?)2qtq?n3N{kF<@>&T}^Z-?eVOWrBdTqk3L*|)YVIcYeY(4>1|S`
zqEA~$nJJ`r_&8f&r|R`^v57_qBUGwsEWrV#0}$!1h(5#S%)04|MlktXED=Xa6vQ3i
zrLa^uL-~D>&&T>Y`4SSbj#ySD-0Vh53?j}8*GyG~O=v_ky)gW*kxEu-K)%I>7pG9D
zJGt1wAe35j#V06%GtKqbCKOfQcamUk(9u6|QWwIO)^~KM(2{5Koh_{K5DZ#ij-XN0
zld4ov7qnV^+En%?xpvJc_(xQQQ#kuNHbPUPwSqL5g0J}W0pTHzH|b{`%$jT~%I?o)
zgB2X0y=w{-RX2AVR}8!ImWjXDOnvJRSTk>yCeN6O;FxH34Zac|NnRK>G04y?&qRWr
z$`fN`8Vw7s#@8*r)493trtPNtSMS1FhaF-uHnbTY8$MAM!@k*n$^{xWUG5&>F|m3(
z`b4~Be{`6{nUB<j>Edf>AhTXnH-jp!8e8xff-P`mFgxO*Z?Tb9%(O|ZD5<fH{`BJv
zxv=u(&@EE+-Reha`q>gDXn8kMfLmqq^T1R=<t9nKjvC8l#EMzVNdpBeYn5uOQ^vf~
zZCqLxFc#&SGQa$=#I!&fY5mmtcQMd5V<JlUv6k2@UggJBE;DlKLPV_@eO%gUPD;^k
z%;KSESL{7^GydZuCg?-Ac6pljCmdVSe9@Zm<Nd@X2royv&s+0d%oxXT5vbK@>(w~(
z<tWuqvPwbPl@x*-QaiDdM=s5nI!SIHZ!aP#6+b3!;yH=}!5e7pQz`lp0EJ`sZ$5bV
zKP`#!llQp~heRxbZ+v9RU4LVC<8hwyJFje13gqtUKr)kuPeY=T{^U~70W?NNnLpZy
zCZopKLYJYElPGJ#2Cm00^2mA8BW8jnnTfE(20fIWhsfL}xpvso@iS>FIwLK{(h^qT
KW5R>m9sX~}(<E5{

literal 0
HcmV?d00001

diff --git a/docs/_static/bonsai-predictcentroids.jpg b/docs/_static/bonsai-predictcentroids.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..e284f2338ae73e228178ef7866f24bf4e114d118
GIT binary patch
literal 76547
zcmeFZcU)87wl5l_3yAa%A|NVFk)l!pB29jXpi+g1NR1HbK?1SSMFa#C6r>9XkrG0W
zbOGrl5+EU=Nl&N&LUOb3d*|$Z?|b*&bN6}w-1pflnV-*^pS9*%W6U|`n4^5>qW`AP
z0#4sDF*gA)F);z4j30nL1uz0IpE&WiKjUCw{8>3zSy@<EPqMSKad4mH=H@!d#l>@q
z|1{4jzEfOWr-e`R2?z=a332m^hzbje@(T(H{_P=5%#1cHtemW@oPs=DJc9q*2mJ?t
zkK;rF3!a(jEZ_tm6EhzZy#pWv05GvKsQnwk|M$Ukf<Yr2I|t`UF2)VW(|{99%*-cP
znEysK<L(H?bpQ(=E5GbzLpA|xSN601g6c1lb2;R0RJID)43p%qJPLTpc~V$JR7_mq
zoZ|Tl7d14sw69*fZgkVw#MI3E*4=yeZSCwG9Niwfdw6<z`ve9BhlE1Ip1pb<6&({B
z7oU>)J}o`t!^h0L{DQ)w;;$v8Rn;}fT2x(qLtA@CXIJ;np5BqsvGEDa?@26yI5)qr
zxb$avg}k-Bv%9xXIXL_qUrYez|BJ1E=j^}Zi;uzA2^JP+7WTjK#dIQ+QJDEySY<D>
z@f%vRyZQ^9Re!-Dcq2KtvXxWriVaEVQNZv?VR;RL0{L&O{a4QZV~oB0r#Sm}#{SM1
z7QoHS#27qgJ^&DK<YrGH9H*Ay?dgC=2Wa<fGuR+<E@VRP>SUMyJvxB!1-QaY2c+g4
zS2QINi>gC&DX$#ss)u%YPu~i@@E*H7=$CRHw-k1zPkh!*QY%a{xLcOANo0oI943rE
zB5joCEn`l1`-A;UkQJ(YU559g!$cpt>9p-iQ-o1l%zw}nFW$rFbin))9RSX!i8Ir_
z01vLw0oc`_N^+epib&GicbXDE7t?uboThwV82XO>^)7e*EwdTt(igpF*7NL*@OA_E
z7&{%H2t0C@Aq~mwGSLA&gEWkb03k;JOQ{%GSsb~R{}#Kq*xJ#{q;9BmbN*q@`zp!M
zE_L=>m@n6Bp3wnR4fbOL_*NhtfE=a+co<K(BSZZ}2doFr3*KQz&;kFWcNiK2{?+`S
z?z#xgS)eGq&9}fZV$Z5FEU-o&l(p!2viRB$+WxY$Zb)^%xLKL>1XJ{brJS}V%&aLj
zUOQ)(N(UUOUf)fSp=_b(fC?}*3UK#`XRrXYKS>7=z5ad*)>`d8^iAsEf^b+#bvPyt
zq-i0z*Iun{a@%xEmP%a171{TQc1Z?4FpUO!Be<cBgpC;D5s!}Jk#uEwd#;mxzfWAq
zxpY@?v!9QDB)JNB`s+IkOjB!2amWTbAV-asR{RwhtaS*_&|WB62D;YoB+!n4R7;tE
zKl8?1BMwU0LxDD==Cn0!{yw6V)zP`}FZVB|9A|s)o2QF9f4};(r^n@T3h$YXBk;bD
zS48yEh}zK_9YEIj`>AF-^)ZE9KoxmAwT@M-Mp$V^?Uha(YCIK=9_Z)WI2Wi<sVZnL
ziWRLE7QH*5$KOTP<WdU%+O8bfRrS^H%Ow_1e5P>e>LKkLZBeTnynB2S{_iyY1GU2j
zZ{p-qMDD*kugoc7pZ#G^x+cxdh7JI+A0em{WkXEHf_MJe!Y)H*CbCn{pdZo!21lXT
zrojW>h4m@aaq6dyW2JX=Kr}-OeC4IxhVM04VL8^t`RD*d(s40J=?`8~1!)O19ruGN
z5qV?`#|xr$d|md}eA$*iC_AMq#UwLdIp1TzBzHB}XANDpCZ!s2_rICMLI0>C!dL~F
zR_TC$P!y*BF#9p@tPP~m|Krxs+o?pb^D_t~!&Gyj)Y9Kq6Z_SSkSV5%N_=&CKAoHC
zzS5SoiQV-b_T+#5CI2}R|JC~6R)GJVf&Evn{j1mhZUp?V)WHAQ(t?<?I*`!ZJUq^!
z157lbQg2_BBi07Lc>CMqy1yE^K0;+aD>DutDJN%rFt{p<l|4IGGN6NLp>)$s)&pti
zVme?r(S9d%8%2O`OAY4Clcq=iv|OeGUaPh5AF91BPO{`GFm*kbxv!w;f!<P?ar5f?
zasNsUMDES**9w^S(lYLhJ|AgjFPUJD(2(;j;!$-{!SLbBru(JJ&er(IZK1S&NVw|G
zW*!R1gH3Y4z}3Djk=^6S=3Ra75vtM?Y}%m%)Z#vhyiA5+p03{Y4(v*1Ed5>?Ax<c3
z(YP--T_ofadRnL2aRllIp_nVkkdp8@v8Ek)3%bU8oVt}o$6vMZwZ6a49;v!rg7iG0
z)?Y(>IbSFff&&$+(E)N4v=aaRr<@Qx9==T#`AiidBt@$R4aTRepsq`oKnQ(Z%?xfs
za769Ng370UV#M>hFP<7g_oYRYwl=3+c8vvng```XJ~c)OS{tzb2aEJ3NseYLF?dLW
zAB#vm>rA4CIne<aP1=p7od2=q165p)Ajt6hVAY~`lh}VdgTUlL|1UZYHy)>RS>9r+
zO^M{VSZT2r_F;#MPw_=^S^@RqSda`h+;T)YNbCzG-2Ei93P-z2221BkCGFW@UtVGk
zb8-hP<?054EQW7JV1hAQe_E_?yPV;@INKu}`q}V-%$fe#to%9eZt7X_N=gS#+vyM0
z{Ww+oorV?aTNb<~<lqLA>j%r?thyF>Xlg^1X3~omxiJkLc2Z{=<`vAf6lOh9;RpLG
z(=c-p_bbufX<C047#!D~{wsv}s(TMSYu)jJ@%c{ELxwT+@&GlTwhavLN9~u;9vnVX
z(}D@jL0(;gX0=^+Y9&Sc0z0k)U%;FSW@3hMD|bs-QCkJ^%fs2TgYp^bovOsGu#udW
z$INAvkF((keVdX+C9}<7C(4TmcaI!i<do$(Z_KT;z4G@hI&Wl#n}(cAF_P<UwvQle
zB~^VbTZF(@Z8sYAe3x_e&X$w#5uvq7E%+<M&Dd-%$}7TPY-WE~5@E*3X1Gc0Yfh|q
z=w3vl@7P9#ZtTXBFOr+@zmEU@^2cvh&kRx6!bUOKYz3Z#=@j{xm3-4++qc1T^sl$q
z>?@{NsrM*8+xSx`(s#;g)e@+rb`%`kys)x)97>(eIc_?%9~?ttT4@Y}LbG)DAMCLa
zQEOvCz&}}-=IDSL#pxkhhRo6J79ENOsb~Wd)4~nnTWvV*&<fc@fk;WXZeJFy?QY~8
zH$TgVwTB#b@3!(oz-u5ZGwmU8kIZ<&>d*6zI){-WG8QsGnG;+<-WeAh)Q`B>yVNCu
zxxV`?#7d#__<8nf&##9fnP;|)71Np7oI99PKSB7s8iK@c$7aUFY^Jt~NNeyfV@390
zZN@d?LAmc8S+|9+T<pc&{<W^S(Eg|;vsAq3mD{&z?`oXkl11@@<BD-q)j_}kM0$f`
zuutxd<l6{?JpEOWT6+c^kW2-o#p!!q2o9x&p%2vR)pDM%WQ$DLwPhd(GiNQ8TKkD+
z6~y6+9dA!&E8*P7z{y_nPd9B~)0W8uO(mzzeT2#rjPQ<>M2_t&YdW6<iD(Y~eoUe^
z1b-98Km5ocT-7Jn&6M%U0;M8*@U$Ix)B3hr28pUd2mC~IfaWoO(;7)1HLRFcGHr`X
zCn+GqEW@!8ppd-w<>f&zAl-Wwti`{6@8KQQ36|#qGHnLvSrE@kARPQPo({08(U;C1
z*uvZ-6=0Q<bbL?V%n(sltb1Tyd(#&5v5NW?Xku;dm9m9eqQR4I+l6jJHoa*viIIa1
zI$ohTIaKYiY~{$_RM@lE;yeSdRSTHfnLRIe?MeiyUyPY%JB8CrkH1)`>KSsw_T>4z
z0@we9MgPxw0{^pi?BkJ_C<%m|=x66L!im6InRO-n+L2k!`ngq`y1GvVDJ*8bc_~I)
z#&Vv1fV=&cvubhD{FL|DH;%G(BAJ0_ao&eTsewf}r%jfwK}eurK<RgOuu!}2RYkTY
z{9ZI_!mmP~yOVHqB&B8#5vK$4XI5<0!XaX`{;>cT!A#*u^{dHG&2;D3vnkD_qmFD|
z5+n`=Dx>sl=(;0Tq$P&k$6!}{4s~s7&C71)U9}s?H1-g=yXldJVq*qy^J}Mnp2--@
zV0fq%DEZWQXB+sz6;R62U?Hs?LI>!|(E%;29?%@yVVDG=*$y<p9rEmU23Tc%p*scC
z8?B+bjx=<OvU$p+D0z}HHz)G82{em{bx|YDv@{n{BIm54yYQ_xl$h@8<;bz+Juf;S
zOz3q;2D@G0)5i|lH7Sy;XC7a%I=TZ*y;V0tyJ}B0AV(JyW&-dx(t#X*w4ROkzt`T|
zEMiMBThDiQBKnC3!7s~iO=$i!>~ZSO)Q4jb!K=T6h=n$nGMplT6Li3fJ?Yf1-kGBh
z5}dp_P>*d>z!zjq@!~&=w1Q3;W9O7#KMyE~wqZtS@(5UsovCc7|6bdiE#@a%n|ae*
zK3=5-T$-9v(yYz*16SW=3`-?K#8<RX*>W`#P4~<G=(8krW$L{}mS9(|IMM+g-u+*k
zdIcZIvDHLv{F0qF`75k^o2LVFnOxR(2z(1C@FujjXkvmVgW>EHv>9o!xC|Vsw09J(
zMdX4`P71FHU13Rn9epoRUQI2%Le*!`5`A<VGA`e+x&Y1J+`0gaQ{2Z>a|lb2QkVp}
zk+7eH%yPbB+>#Q8aPxkmSU2KVULGuHU7tBZv`<(1BC_~Pl@3tTyN0-N#Dk`il4um6
zxomL?frKFn7tYQ|7{p^b_0Pn9B#wST#Z@$E^*ovF3lm6Ih-<yHmD`iyWy9sxP6v46
z<7nDs#U!ejXX$FDytHC+#b81Zh+?rTWqJo-bm4J^Iau~oRu$xY87JS##Q2f0S3x~r
zaWeu3Q>fM71M2(Na9Q3{eVD`1z_oH4U!|L)e!Eg87CJooBj>V-8_Vfea}y0cLPLJs
z2k@I;1>7S~Ia-RFOWx^`c;4Il<t6V8-oi*Y&T&;Fr<(RgwaJ>;90?bP$zhOEwt7a+
zjdj}<<%DXewEXsEKIy=Gjvw3iV^hQAv-@sNsKE(5lwu+;Ck5RhUXfKe{m0QcZ6>=h
zu*BG{z{0(-dAGkoOx{6(Ph*QS*3(9`gf%PVQ*A&(IDG3iFnHp);?N#vy|93CqDi6`
zX>)6qw&6VDgX`MDDdsXCC7pbq%-2N>ti*?ye48?KP45zP{}9CzpM4RtBg3-Hd~YIK
zM@gn;3LDm9WPUWcXm<9&#nf%{v#S0tK79gM!b&3M;8k8Nmq0NtTH`QN5)E%d%KW{W
z52_nKfwL?t2%Q+h_sFN*ik6XkCMK$u+4o%Z>k{oECDmj+f1F6RdlPK0uh=<vtdGOD
z<xJUKZp-umrxV99x9zqvg;9Jxd+7-Z(@B>k&)=IHJYTVaSL&t%QjU<jgI?QVQzXbv
z_1ezHU|q}KXSIwgBNmm-h*{uIw3w41Zx$+#%Vb9i$ggS2O=FtAx3#NBsP{t@N7vhK
z^$6#F8Hg4=6MOco^Nrz}#04$UTcHu*w{UCjGpAhEro&H1MB`~tQ#t^4QvdI7urfA_
z{fFv5ys$X$`I&!{QNh2r&5j4WTGOHfw%dS5v(|eO@B@O(G3ny7^Cx(}PtgH`9CX0u
z3)A1ebimSU`2LTQKVLZhzj;3M$2#M;TMPGC?@*bBo+^QFM@d*A6~eYM(BHp>BPcXR
z@T?+d$QGAH2b@_29nPbF-eEy|CLN(m^4AzPMd+@kf*_O3<Q#2t#p_5HK67lTFg~Vr
zNn>BY6ms%-S>a*nwQl(dt+csKZZPJ?lqH#=q^70cVe!UFv|AlRY?yy+YHkcL-GkBr
zjfa`~gEIT@yEfJaLY||(8z(=WPO0ybiUH;BeWe2`eSvj!!2T{B(1oxz5QP6jJ!#F!
zY6EzD3@4Y3-C>pa_gemu@aRzz9RLSWeRc(zO#fc>qIK5L81@w&LkEaA8w0HWu@3x)
zNB`?p|L+~>ow}@5*)CUl$=$|uQ*`feBGEXQ4v?;f*Wr7g>pIN|*Sis<<nEDGXh{%+
z(H9EDc04*KutvF`<Hg>_#VK<@#d@*WpJ#T_{YS=wgw3cHcFJcW{3TAXiX;-}QE6H~
z5{#||{WcALA(`UTQ`}a;bWv2k;0^hc9TFBo!sXAYu~HBQ!|SxkGM{YTiM4O*`=w#k
zX7BP;7iCQxp7)&L-U3Hac#mW$U{i_}X+TZtPeV~I5yCs^c>B<!lBhFd($!15`a?yk
z>_KyBn!?L9%Y4ey3!{{*^q%gGrg}6>v&P0XDpy8m|AA=+8G>=?E^vaD&5kd{t5P)~
z-=M;9K5STy*YI3J75FFYyQ3xPRddL5i^ory`t3<B)uarfku%~oqDh39Uek?bOVBJN
z#OP@sxYsVRBMM>tUn6^Hx@2H`4z~El+X>!ND>OZ_gv%e))`PCdDX${!dm(WjlVzU7
zA%Ap#=&0cQ70G-2iFSt4=n{VHWJ!icX@&!taL`ACc9`9AuAZpH*sgT4Q6V-?AzG^=
zK|aN#*ObFt^V8!QeGlZqZJeLV)Mkw@diL|x3NYxlN6Rg9&|(%`EpsO)7E}XjgPf!q
zkwjiK>ABAiiu>T?$tV7VrXPCbdW|kSu&ej*%T>1XG~ONv#{}k4+6f(zoONViyaEN1
zw|ocR_b6CG<q>wJ)xpUp|3Q!7#VYc)Y0`Ug`YK470-E*gw5pE$g%i%2B@g-1B*E6x
zt1hyZAj|MDHzQ<W6u;Y*IN*fVg6#m?=$J79IO$2ucpW4YEj}~&8i+baN=gd;?6cn{
zXI;Y^3J&^V7%b7$#w>5IEHq=)KNjeS9a(@_G{X><ijYq^Jv)x>A-l*L$1rR@l4cFk
z#Dns6gUPG2@aITK%tWYBJNAwA)dsMyn)PU8yGmZv+(62WSq>B|D<kxm{{yWo^5I-p
zqy&Zvo`pQO%~qpC6yTfRhfZLgR+N>>+k^%7yFaQOP`EE!8~ss?{9{lkXKwR#<Q1|Y
zo(*PMKozefm3m35G>#+TC!ygs-;WGl+Q&!a{{C>r`@@@tvKDpykR^Vq90}8fpDZS@
z<&qlP@TbbDcO!~@v}H(dq1j9Kk(L+AJwF%nhKYSty+5}ts%|R5*3E;<)84N~wPZgt
zrj-Y^G&*+n6*e&vUa9!4Bpj-FFaV?LKuT!SdZvrJP*2KuwK)F8DZvTIFX61e&<+SW
z2o&=D@twu(49OdJi5%LR6oay@#a*0e;v|%1eQg<roT?0V_u4kO&HUxv%f5FerjmC3
zIRa=;Iv}25UP?9v&xx*UqJ_M1>ZIn1G=JV#8#`cs!Q?i%?2&upa~Zxhi2C*$7ChYn
zm8~TjypzCAahVQCOaLkxyWf7u-mWwQ=r8dJL2WtW%{JUn+gdjs%SwJpcD+B<6>)EV
zEGz}%tS2~w-nxgtWQ}ZG9#`4{B`MH0JcEyC_EF!#Inw$ZlyD*{rinGAMMea>m~uE?
z`!%ooNY8gn*m*01MdXgdr2_Srf$_^K`f@N%^biEu2NIy_)DJ&-Z05InKA4NG78N?s
znPM<?ul5wvnNN~-Kc&TvKj^DMLG8ek#@#PBf+d=@2^;kUXA>;uZl$_|^Mj@q!$5_$
zGE*exoK^r}8QaK6y?W$GQ$h*wwdPRv*SQmrNQp*E1XNY0&kGx-@a96Mwl`~IpTLHH
zeV0XwM1>!A3$<4cR}E8|>b;<k+X@#mu1y68hlhPNp6cE51tXL4G&cpPdRDBzDOv3?
zgo*M(XS}hebUwVy32SfWT$}cmaQ35%OP@5`T-QQH_}*(uA1P}OPJ+j(?Q&4X0*Mh?
ztq9*HmEdr1@9^;Bv77Fm2lcHSGuOorFdd*ORLACNifK3YGvXCtSQTbR7P&<hLcJ!A
zsaX7WbgC4vR4R+r?LJy?^Uw`;y9+a+u#%ljhv437+{<|Gq@)dfd7q@O-ueo!Y`;f6
zjQ-0v+CNbJRuMq|PU*eWRu|-(`p?&#Onvpj9<EP?S%5as>7<3F2U8%@VhVaXJ?p4%
zX$O}>R+`<p&K{*owFsMqbF8$S8TAd8h>*MAR2>cbE57xPIr5Wx?Yx?N(7H>CE*Bw1
zU01IlnGSHVhLbf(OG#)?nb?rM&=|%f3DzKM<q~`aEbE$*?hOa?@ZQVDam+z?b+~pF
zSadk-P}|zvrwxfMYa}G{I43$NA>#g%(Yx^jq8k}jHnnXJpI?nV&0#r-<yeki&Qa4?
zA6{M#1;bJ>zZPF*C<~&zwO8HYoz1!}#&C+Et|lFDV=s*kc+Q|h1K!3-KpcwHAgCu2
z=zzHLUZpE{d1<Guc;G|Lh45KH5F?dHdI<&7qSlq@06-oiy9o}62Qa*f*8=UJ9W`$d
zg;j(O5HO{l$!7Snx#73yfFI}4l+5Y8=a(3%dlnMrm<wx1V=|@#I)>>04#jBnAD0m{
z>7*L%9P0cRZs7ltqVvrEBItiUX8-D;|CLn#tB3y2?;(G*sIFq}a?Yt;aF(8gRZYE>
z;*%lAyz7gP*;%;Xa>^t=3x0~1fX{ze1?@3hWz%j0DLH7>k$G~t8>zPZYgdd?u(!&W
z#8$Co`9D8in$3q4FqGJlHA891f#?AAgbX#!62q8iYggCN{!0g67{>mzn{@X?6`G(=
z2NcUt6X}58Lu_cOA~xxmlcDU6Y2O*3p;910e+>S|jNxZ(?#9po3bv+HMf?uK^-4w`
zi8C(#tL^`f`|3~4XQQLLR`E-HTTQ>H^#hm8Rw;K_tteJVnL~btFdZGcGN54h;B65M
z2Y)D;Vb$VSd1<%Gea<s0(gfB`X|EIEG&;Z@zI9)Ql;Z>3kq0wE0EdkAHKoik9?<ju
z5K)G3{kP&bk@s93{@;lpL#R3T@)_&A2?Iw0{{}p`i2VA6JLf+_mHNaAn1Wp)_WE8o
z|A3nG6BEBF_G8-mV}QXR_cM$|5hKBH9Yd>E+Y(Cfr+tr50$b7n?}8D>0igNwC;DL#
zM^7Vhmm>$N4QJsjzuSpc@s98HpM;D^O<*6G7GMCUE%<tiOl7W|z9g}{87qBtlqOBN
zYcsJqe?!H$FjMIf!G$o+#WAYq5DSO^lW+sw9oXXdCet}JDaU^HH&pX*mEGC;8|C8F
zDPN!O{1q|nUvqv<DTmMQcl(F^s=vw?e|N_C4bPwzrhX-LbIqwGKFjyP3sU{dC=I()
z&}wZqrSrV|kD{J_1x|B~k?jRwBBZSITTk4p-%EZQn(k?~`a0banRdu@q(f_W6zSZ2
zRokLGPBs4Oup(akdGGw%O^@>@%*cy(dM#3Z-ey*1V!nn*)+|@q(R)M(+)f*DP~MP7
zdCOh1J>O~0J{nK<vEmafzTg+xoH4s~eWdJlu8DfkOd)h;Z|0yV2+qcU8`1%9WU89i
ztIf-CR0~R|KvJz&=C2~BMz6}36Wg^8A#X#D2hit_+IOc2GM!00u&0#pe9CB>T^k*6
zbNuY~*++mBElFJ))!P^oR{Ljf=h)D~X51WdXPY-^oZ;l_@MvHSFt&xiPiv5eQZ>9=
zesOvy{!oF<*vep7{`HZOR4C!HYW{roQis1O??@l95^C(F?a`>87$#E|WNrM+bII^O
zlbx}4hIfC%Z)k6o_$VPLyIIFww0lhRK!GNAI^g6+_kFV6Z$|zO_?@a*<^y`X4MPBX
zPMcRpPE38S8#r}kXu>3!4DY%Qk86@GqKvj?>L<GsOR_Om%|kFl)gl{$?|dDnMu3CU
z`@51huwD@BQv-HQwQeh(75&R3NPLjgS^WC5iV^~kPwxo+S@B1WxRYl$9tj2fVlxML
zrrJ@tND}S(r!)Q<x7B;J|DYPV^`#C2GlZ9AoqNnB`@*j{&GnIbVLl*j?kX{*VKJly
zU(T3Og?{!^q5X6~?bLQ;!%Fa>=Ub{~IXI(1^)pfGB4Z*~&g_-zPwZV#J%crXsB1pz
zduecunFaH~2aoi8<w5y8yU7B*+u<TW)-SH8LFg36Jv8(_$?;am%|YC}fd7@G{T+hj
zCcrL}DnyZ;!*f$&eq6VUQy@po6u)v@z}%^C85%WB%j9(B@%!!+(Ez%1%R-^u^i{f=
z+{KO5zmBtE?*61vLU;7Yur4snjkM`R+T9%)<a3fZ`noouyY|RQMY5pEPv*k49<lK7
zsY1%K;?|_<VVY={jlUl*0zEL5WAq0ttjkBV1Uj^g=(-b*ZTgI-gDtXDY&i43V>uyu
zX+IN~>_7khad-Ki2W6Rr#p>8)5qYB?SrprsPwtjIJQpaR%z6W~c+Sk-fSqhISNcvD
z*a_l+d6FdzG2-|i{LqU|GBKHmD3zT_#FMF`+lz#=rwh)BIosl{aQS+_@%`9BOdKyK
zbIoSgS7>)cSCWpK`b%J{G40|G_!H@kY!qa}LUiZx0$Py+y0mZ=VhjhhXKh#QEUr#?
zS3tWCwJBg$N}%>q@7@G%+>?B_^;vXI1qe$)8`f72ea;`Ra2=b4*6cmlaEj)Mjcxr2
za}&xXfW*9{uMfLT=xSDYx5$pjd)z~hRLs@C_;_2CQ%=F0{nellLH{yEyA%hOKI=tQ
z3k=WeCs^_Ro|^h|_(oX!#76JoqT60Tw^#b<;p?1Q9UUynOix1K)#w<iQywL-Bazsr
zli;w4^j3W`P;x3e^V`seJS+8s8aiOd^F!4o?vFacdfnIBQx@J}L%nsvCa1`_nTD3+
zVf{Voqnz29oXq9QCcF$yq`Y~zm6A;0V|`<78F~TZ8#2|cs$2Z-;KTihzM4;LKf3~*
zC0jH}JEvzX&NUaA3YEWH$>7(VcXh3P*y;CR<KU6L6VYhvbK)RdlL&$L$Bbtinv{pG
zfn5W8Gn}*T*TE{lyX^*e_vV~6C^J~9liK*E>T$$cPF6Ng<SFW3Bz#-n=TrFBK=Ac+
zB&l?&jr;pbh<{N|#0i|cr{{Zj(Fb|?#)~|?Mo&PQaS9S;c=$161xf}{__tcfuOIJ*
zI7WH{>z%VFzebL^2$AT3BtPBD6&8@fKE>E_s!q`~a>vPUKI8V=om)m}*jN4bEza|e
z-N)2ZTQXu{c%H#{{oya4yvyNBMU{=Wi0VfzQw#8txsOdPA}S~()ELv2BevXf%{fHv
zl;u|^qqzue+b!RU_16fm78zmw-aPYru++=FDWav9@8HjC1>q-;?!Vf*e_f-tp^9=u
zJAI*sset#i)t*d^W`38z8WCRQH*K&Ta<s8|%l1wh%J$$01%J!HZ7N%8iuHcpX|UV!
zZ9MMnuGQ|eVs~1dc*crJfC1TKD?S+srv#ftj>;Elq&_mS*O5w*d#D=H@abXvlN=FU
z9^P0@6;`S{VHXqhQ+nmd;(CFM4>C&dN{_PB{@JKA=??)c;$$AD*^D?pWtyxd+bg~?
zq?I5AWL@>a5|Cm6Rv3jX;Tx3}KBzt|c=>e7aQX@`h9*gAH>WJP>t_*W?pl&+p0yN~
zfTTQ@{z8wun^x`^EMjC9tfIpnCtH+7#Tos|m-F^i$giFl++47XR6%I9tj$q1wOcYH
z;h3Fcu6B6Memf8RhqM?usyi`Ah1s0U?cyFgfpl{6mWtsOQBPBsJ#e`HmY1yoaT<1y
z1bV&c`)zkn<j-}@8BE%|(~9_~)8=ZK_QsXYFMJoZxssZ!D?f=Jykb~2y!sdZNVAiQ
z;$&VUXT&W)T&-W6T^Qj7zGbz-Op!DPzoM2f!z%Y6=!Am$iLLbe@)vT9kF#~1dE3@K
z|7s$QseJ1rfvresRZV-p@L22au*E6%nJA`gX0$9P)(F<5r?;$<I@%wOpGT8a8dtv5
zA3&za{-Bq>bih>O-r=jelqiC;NT)bF8p+H5XHQyg%q}85DHtvonqe_`x$T!!cklF9
zJ>EhySTLcI2&b3~#s=jQmHlWxkDh?QT{Dwgp^z?oR`G3cXAPQ_5!wQtZnBhGYEOWR
zd(+ONM}nEkon5dh&CgM&m**^^-hH%aBWVnWS(|a0sSkhpK1ZO)%>97{`xm%OLm6wh
zVOCdhC!fnC$m^*$UzcC!h=^xxgi-%r9KXQtpXT0-O^mX(_-NdGNX0CK&adZJH!Lmg
zUt6!pgk@1joOcgo{-U&ZyNFZ53ssAWh||k1Vxx}MCvjkrO2-?I&%Apv|4{ohuh<~l
zM+D4}go^TA^PFh1DDXZF+V)mBrruHM2GG{3d8ntMgVqn&J2!b(C-D4DymP*aZK+Tk
z#-wKKo8pLys!Co%T*XaR)kjBvC3ihN1D#T5u63)?KHD_fopTZv>Phsy;zFXOmjA)}
zVqUNBh&5fFl}P07F20pwA$Rk3v7emQ{4J3nS*Ocyy4wAwH`d{+o(GdYwdmlAT4y?-
ztdMdJ|2gRk)tqe9+Hz{_%y$gcDc36$!zRm9QCMOGQI5XRclv$^%jV^uB2u@suDp8C
ziCDit!dab$qRpDfAxZhuP{%_(eVQdB7WNXTS>8bhTu}F9x)~qJDx3S>ZT?(S!oB0q
zR72h<s#}rMA3XMo1u1>B&GGSm#`g!73iFTlgQWfKtNW&Cq;By_?9)pIwP?;JSG6uB
zr^Xx}3;lO3mq%eYvH8<e8fQg!*sI|0`xX!4o|aKX+siIo^=%BOh_u~dY}8<w2L(km
z@OWkbzL$ZM5^aP{EgS%iH_@ajU3u8U76Iw=!*IeyucqB0hSei`<ewygiq(cpy$yWx
z=<5TR4QVrmsvTAV&F_em7}6+<xDn0_KUXa?c|XNA(R$&XNr`q}bi(JiVgbK<$YJ^-
zl-7be(o<6P`;2Zdad<8VITi8AK$Eo*d-vCoDfXt(+LtgFmm&R#CL1p}57m9viqDIt
z4B?!HehgX<?^KsM)wMC~_xf4XewJ@IXQmf%TJFZT^0VI^rkKx^4CrrJE;AtLA_uiC
zb`ejqGgdtz8O?~<oPim{8lIDO&Ix8m8s9*9a$7qZm4TMO_ptt~dmB_+pTeYG2XO5O
z{~^}3_IUT3^Oe1qq~f@i^UKmaq=CwO%IAu2yIcu<&5HNqXMdaf?5!@CC2zEAV|i@~
z!aQGPUjs5xmB@OSwbeO1C*Hwyb_U*I7bzQV?^b7?GUPHnt;^G*db#$#B7koYfRQEg
zM$@iR3QLx2{h_Bo(R&m6tP|8r4e$3SY?SS^FL=F+wZ%Ee{FJ_%_DHDi+Rdv!cfy|_
zRt}u-RC8ba{IdL#$4)>-&jtAw#_;j#AbiLo-8m<q;Xks^lYO9B7UMez9&GbO?r{d3
zWh2W_Nd|vXKA`cwec~^JavflT3s;LJM{)d)afb6trO)c@V-NXVOckfSGT@s`x+m&c
zb`Y80A~_iV80{FCop3RjNGL%(-<9FESBul+VfA&-3^J&6>MnLF(=g1lk*|YJ8g^%F
zUv_L!rd%e=bckfD#7;u7Qi@%B=}PCa3Rk3MUP{QQjwpr6Hm4Jn;_|V>HSv2MuG=Ys
zR~0JyGBN81J&|z-zU8A~II<N(-<@YzsI~fK<NC4`J7cm*4x!w%1FIJkyDNlu&z}Kx
zO3n0c|9V=@?!Se)$ZNFxptwCJDN>Yz{Q7}<xpY~(c8GyX(&oSxm~8jge3VlW<lE%G
z+1kFTT6mwJWA$;1Tr!J4O@R<I-ta?Wo<f&yk!?!~SNOoF(DVe>6sJ}eV?SFf+-BUY
zcivXsJCXP9`PW?AWM1g&V1j+aEogmNhA+KzAz-$8h4IFlw@ZRqCIZfO%iUs|A)wAy
zeVBTC@6+h&g5tKh2HHvw<7QT<bZ$MQWt9<Ljt7nEza5TLhRSthMpj1@K>OMc3)2(o
zRx-343j|DEYxr^DO3A_}Ds?KK7mrj>K;7A5WHbY9K#(E}r!qWn&tJZqz^cu>`Xi0}
zZvju-Dx2T(z?O-K=$0###4oFfkXcS$A(f=|e!I5x!zY+p-{a!Ir^y^P0^Aj45V5G*
znl7zhcz=_XT^S~-{ut~sIox{uu0?#iVQQPWewu{AaA`yy4%|jwyPI68ki;(W^f&uU
z)=4m8H!RXy1=xffeK2~A&2Yg}*m4f8V2%)$>lvi|y)p)3qXsPueaQ%v>kACQ<nqD%
zUVDc%G=&#$Hr_6`+_`VVcdmb!_wujUP*}6l2$DB8TW|zk*)l_n5Fb1OL-stenjHB@
zA&bT!IhCVdHI+=LYwBlcvN>$SkrI@?AK7O|bnS8cb!7Er?0)>?>r#m`c{3bHu1ziM
zmVm}^hrLOZgQM}o>RL9#yZNNXSxhYT%A8uv_4L9*r&e&W6D6Anb<Nbays*plM^$D0
z?x)({)DGq6Uw^k>RZ5f4zsJ}*6zPjF4u|9hpmbBP8@P-Rct%$9pdVTLElhhBT}219
z84-EZ43F*#p=hJc@%RBbG0x^_qWbk`@g;L=@zOk$&^<NEEr$Asx~+e+#T<FsM2`+K
zZd$xF9VVq`3`KXV#YkVPqm*mM!*sOc5kcjq?LB$sZUl!Q<MeC%-Ua7eOK<kg|5Uj1
zmnrn})?2Oh9x{77PzDOJBu&J8S$3TJ-AYi><F~xq4#}PNF|lvz1c$Gd-%{XrEUgl{
zWLQ`EG5$^Ir-K|tbw@me%0*3Y*sR5bJeWo?l;KaB9XxbqM<h;Pj$+kX<!!x6J3kzB
zP19SqTrXJVhD0Eg3i}@8x+wDz`rw$b`d}c5JldWC=hrozEx&6$vZLo_HI_G8a|G_c
zrCB#_G1(f>&#i)t1J$rzY;4&GxblM7!bx$M1IAnjM)5GV`(1@WF~CzbgPbnj0U!tE
zXA%AA4<7Ney;F`@&^4o)vmVFYhqV&mFJ#kJ_-VW)S$$=&eQ!0qjMCWdn4Kek<%Zb$
z6FQ(8Heo9-#CM>5)fRjE)bKX<PC)0qh30WABP=z7o(Y{qSR{)?$hhxM0{f9$=#+gG
zc;{ZE#CUn`nA%Q+e(SdiM8&x=O$Aiu)!D5+vr2UW@y>jYY%m}7lN24Gs!df(iq3XI
zQ<Z@ypI8hRRl|L!T>S?8{ch6~t^yTtbA@61>nf!74)kd?sU4{+>wjQ;L|++w7$Zbg
z=-lm7OQx>rAH-W_Z6u;Ei-~oT@eH4YvE3^w`8b2nc=x!F)WFN_9qD8!y^1^v@x>@>
zo!?eDxO)2&wOC9odh>7slxIQ1S@o>{%u#KmiFd-+`wxu|5l=9*i%_+;?6Z|anrMby
zs^1uaTFZ3FEoo&P@0vrx>OQ~Hw_*m&kLo&+-G0!R##GjHt>)BtDFKIz=d!t+o$kGy
zaBzFD5$qr-Z2c<bs;K6+RX08wC<t@PgEGbjs}{2hk#iZDi^f@<N@P>AG%<9Se4=R%
z(RS9zyZPAmIY5rjwUy-?P-ec76#5$O6)V#|fm23dc3y`e&m$$yXas1CMeBg;3f}!y
z{7lT1&-C0LBRtn{)R1_%dK~s_eL44&i0+r#MN}VTd#Z2au=oynKUyQvvHcdB0X1?j
z^<i=x=p`PWU##a2DM{q=d-^wS#t!!yj}yc4+T2I;#iKSS{o&zzp^zE#b7!0rmn@E&
z4#Gh1b^H!4McPtckhF<x`4!NSfk~<}{0n6S+o@ZA;>f7B2E~UphRBBoxJ>l~p0;_j
zvX^JoVs`(jML4<>$i708oro0aFLpw%IjY5W?otsKFT7<c#u+V2Cf6OLnq0~UYWY8Z
zy0lSN9!x_M4tK7+uFr>Wptg}1Vo;>W-@+e~J|e>Ul4Oq<1a8V$s$&o9Qpocw*<@BQ
z<ldwE_NNVXCEwV60^^aWmqAJ551=7>H#!SPb#G(}BZGrW47nYlzWU(yr?(dmM0&g!
z&Y0S%DsJkhS(pIj8d2s%56NKqkH7b;b{g-PMzd#6y3F*<Uwf)wLoRP?Yuvr;KCRBF
zDDP2$B;|A*W^{+C6_VPWp)kr~dq-bMk6xMKylHOp$K%P={%wB3ow>e0y}h?3d;<!t
zO%A~nOSLX-IO48nZ@%{YI>R37rit!fpeaCaVA_Qv<+C~+<@KeSuqwzfO;OvCh{>s{
zM^h44ueL6+)t2vC=rn8TnQ+ye8=a56o#FqpT)@O+meAVzAn?<i&{ldb$&Pp~#y=OF
zX+~yE+Ovw)wU!-)De1q1BqEQ=nK-D)7ldzO!uEy7PZh3Sn@(JbWTt>Z3nB-uC%q1K
zzPNp_e#|#m&Od(z`|YfE<hcBZ25%sk#lOR;UvA&LR`0h>tC1+~ltm=IIA;eXl=3w<
zLUo*)iezBmjy`WjM_zzl>B8#K$_aQ8uj_jm#pxm$SP#ai2zRIw<%1rTM9t5?y;k-N
z__j%#sKUSGFw$gH==03Db`yA~x;fqnT3fL=bGGt++q>IVfW;O@z&lEttk#KRp~Mu4
z?tYecR(%)rk$NY{t$aGmBI&xBG_mJ#)a>ul<cKe3t=X2dxvkltSNghSyC`Zz=>se@
zbJbxi$3&@!LS*a>;$M09+gTz`rPur2H$G0mCPSR5O5-WMdmi(7S>}7knxk8eWnw!R
z`08Q(&Z$@Ae8PVGx5b8GSSaDM%Y*=bk6v^o<hpD+=;E1Q6$WaGKkJYBx87tcc`!hB
z+0rBHyG{u+p?_7QMx5Qo-0PYWh5J9=PxDZy6ubDx;K}3b*~AtbatH%79G4>)DT~Z`
zp{lD_!Nb+Y0DJ2lFXT+A*^JsVbcwEd)5W##4d<542-@l=LYAR91iOwT5oc&>r`DlK
zCBdSWTpkm<9@Ho?qwZ`%oKQKZV{-3G*hF4%s<N=n-OG$XNXV_ID#+n}%CXad$m36`
z3<T>%TdHs1pQFJz+^JEjjxS;cT5VK2<mZsiW8-^E5c$5e@sfju?<3)BoUOBR(p|Y@
zG-Zm6G3j$oe6WldG_t*2yU-Tn*f4(PZEwINmfz#)o7<yrq@PA8wZYU#cRNOV2$oj&
zMm5Xoort_tdF-3W^UWiR4bcjt=W5Mol=I%sNc=jTqbhfSB}OHs)v17bnVj?-X848j
z5hkD(|8QhCN9aYagNs99$DI`sTrhG>Kd3}lx8`E+V_UALkv5?xpTS+?rWR)qBN%W{
zDl!!`QAMq{qp>t0cFr}E+1qeJsLC8ivTAaHA-umJW1nn*l4L%r`lU@WoYOP*Udil?
zj`L96l~@IH-Ie#}+$gjkg}p5i-+094FnRCP!gI;YgM@#=|HN%~$#7}*!6$PJNYP1P
zt&(qE{(q`eK6g?BEU#8I)P5w2a3{;v+l$?wbhCXT@Lk}Qh4VhDcU^m5X&FDVx_-W*
zInrX+7uK?AfoYM0TD47F|Ga@DCW@F<k1SY(A0-^$5S6P*9(Ulk^*EvB9&|>Fle-{T
z$i^z}86T|={R%`{Ymsaktq2CLn@q?#4;d|_Y8aeIxxt*WwRtB;5r<DBMOY9v-)OET
ziBVGMfL3X_Q8ft6&a-@(@6Yv#%*@V|qEsQor<V@m!_$}2wd3tBASd6qh=gb%7)r#X
znY45p*)*hrMxqHk<j4;s!#yQi5s^V=hk`yXWocOao5#?Yv_;4@z)J>ft(FlLQDfLS
zR@+aod~b?ba7l0#IARrEv;$=9X&`H6mj&ljijBw1gEG8!#nys3JS@)Lu~7c??dD^E
zInNs=Yl=28>6N}*7@ibIX}m!OrX<Lte`DWkhYMDp_t={J&S?V3UNLoZVJ15f&J{AU
z9h@p_Aq_6cU*K(;gsl4(t92?*HqTX#6=hn8$P-wSYo12C)*R#pm>36vICjg>f{~;6
z)4LLglU9TaL{e7*MDO!L=QjqPdpNS!<47aLM6^Wogo)0HRMeT13lAo&(1eJ{2oxyr
z+Qz{HbUS5h23A4a$dazb3+al;yoq`3)OmQ4@_kD~pyY>IQeWe-$vOk;oBjFPH&7f<
zV3Vpygm6>r+7DHVP(=7?g|WSND~=g^5qZW><kYKwRo*&6hl}1f`}8R#<hdnO`)#Bc
z)sn0jM;Of`(aOOEQ&8fd)9GcUFH?U69J@=puhsibj*0*2;QAS9Fv8wt$kbdmXBV|d
zcJmH#%xOPXBbvVY=&s%Hpo9tia^81K@%gk_vM+PL?ZJ5K<BL|mloUd5tB0ddJwP3w
z^{wSmQ?CK*8MSRNjrU;pF;HAnCzOQDuAW_O7O)*}GNVb>qw4B@S*v5M0#tn$ezX2k
zmSwtqn%Z&6_68D6wCX?zKq1}w!ldFTm}bH1VK&M$0yY8`A8UAW$@Iy@ih<X<AJ-L$
z<gOvM(Lv4mOH>VV_iGI_H|RIi(|Ka2T(Mix;XA2pWdd{k?Z+Z1^>8<_?8QGOc=WrU
zasro(GXU?jDv~;rxLv4s$T^Hap8zyr)})v?BVt)SzV6p_t&9$Eb+&?NAUqy^C_ojR
z;D~;B@~3?$vt|TZhROsyRw9qgmj6N^bMXCYx8owkeV+B>lY;30HtJOj)pG8$fcc1D
z(9eQQ9VuS>j5`)lC$-L}I5`E#hw7Yot(!KV`l8qJ=mw7AVkAPu+X=Dkv4runLlx9(
z{-uI%Ck~J-RZ!f5s~gg}J(|xw)thVk<btMJw05=z2_FL!BY>h&covv4cFU)sX(HL{
zT|>mjn?t&H7$y)`^|V5r!VvX)t;xyEb~0k^Xs*tw3ny7SrkNx`3GEEclt!tRhlP2%
zhWUm=zB;L3U-_g@895pWly&94GN5j{|IE-RM`DlFDry@Rn-2?p*A#pu4yZFtwZqH9
zjqAZ{E<%)Cq8nL!s+B=aZC0+2-QknN!l;3mZ#^Yb7LxqNl09<fV87wID;=i2y@tR`
zazZ03AfYCKd}y>D^BwZr7j6|jvQZk1pKUlc91Rj=4FxVenv#+b9hkT~6kjrnenk@{
zohs8WD{B2ZvodP5=k|3qvvaXL=-tNjBlC*Iw7ECmxvkA?4CMJ)!`PGbIaVXjLK2j-
zrBF=}J}9c4f%B95VL6^wychb(De!5@{4_P`zlNRFO>zDSqm6U`DusG>bL(N+6jgt7
zTd#DGh<X_;!#+)^8toEi#YB}SaPIyLzAD{ywt$W2s&4`#Bc|)|O|8S~B=A*If?dbq
z73dw~rO!d2jwi71PeFq-)qc<c|4cHkAnYRfvHn%EOq{O<^EZ1(3bDH*9jhFANzNyo
z{D<s#Eh)TIxw8MLlfUZ3&?%Gqr8PNRR9E>xXg7_|^At|UN|+%J+?Wu&mzRi@aFoSC
zj^~k8rv7-6cHeaV(oDa?rc)6<NrjPvDQ3Xo4jTFQVLNEX7d2*)wEgs!tl69-oCvv?
zdJZ$Rg1zY?N}5>{P9O@7pH1mqc1OO1U%GWEp4n>Qc(Lcf(kC9Na3U_tP4NZD6bw4O
z%)p=CP8K>&BsN_pVi(=cRz>qaoD#%xEL3M-f-$>L%*i61BGQUv%c7`Kiq?jeqysb^
ze`^$>QLiac=C$*6UBWw~zeXTV{m5P6idg&cq*7>`3<t~^q8+uHP{w=OGKi)J=Ucf<
ziKhepO4&LOtSUzGH#nFIHCzF{*R6Yl4i9^h&ShPV+(9w;+j61>Oo1ahzNOYSSZvmm
zQwmqZpB?KM-yhjj7}J<?Dp~W>=vj<=9IjN9Sok8g$MkYzuItdc$Ah<=lH;XL9Ov`m
z4Fs)?0smHn+F`y7go0Z+d`jk6Mk-YEzpVrOLC7WQJD`Mr4y9<<>iC=w$qE3|egkc6
zb?lC!3Gzn0^UgCzIvaUpJqf-lJ~`~W=U{muL%k*-az_=fc4p%$qP#_cq6(CzJaZOa
z8C=Kuu78`zco$ZrM>c!m5X#2#m|vaaY*jAFB_9Hj<eOZN*Wf5yENOf3Tz|jN;`WID
z3}fM>1Jnx+@<sR1bC4}~?$%<)krA|*#s`c+iC0Y)mCbQoK)Pp`kh=0qy4@lCh944}
zo062@4qnptu)g=y@T$tFH(Z$NJf{TbSQdB5f)JAURxht;3rtfI%p3TOc|P^<1=WPU
zwDH(>3$_hD|1GfYZG_h5p8trN?&Ul$X5+e=??XeTMLn*igWzDV+z_J9heC(<LbHsG
zd6!l*Ee0piCQYP`w7kSq;X0NVuof3;j%~meU&Xc;)u%fz_Ge6Y4p8eAL%=4Z5KA>m
z&{559ZDXp+4h@a?anRTb=i6wJ`0ebn^2{Wo6z70SWEirs58bmOYQMa3d(q=286bS{
zf@fbKCnFn03@IeLbQ2sJn=mkKw{We(JPi@7=I0Sn>eL?_rw}9m48!^>XWoUyJ~k46
ztQjsGbEH9mw|~dByAhynFR?m*+@hV*)Ace^lrwJiZWwWj3BLO=Co0MkU+8&sf#TC`
zI2IJU!-c9rJ3$x1cFR8c@tuhv$=_{SG_%n@NOp5S8DPLN^tm2ZPSK*-+0qctAh?E2
z#r<ae1vq*bAxBDju6u>d)!K4_41a6OB_nVSQEI9fSCA~%Ff3Qwkbd3~>#ZG;W^+Ct
z#(hD)+HX1#+?7*fg2<!fcV734Da5Zbx5GV?)<RlD#*ku~-qIXPva0K;Jt&38_a)i`
zw@(Yk@0MHbxeWy;x;UGx-lz+!kiMomYv~AGM|>8#@3J!E!S~=hM}AQ0(t_CXT}u8Q
zWSHvyO~1JGPikMFdftO>NU0@+btSh~<=5p~mD5Z^ocFs<UH3Ql!nhxUXzWA_oE2j)
zz6h%CuyUTUgQ0UyMY~g4AseT{2UGd7dU!io;oDwn)=}&x|7akm?oe`#Z*aCXiV4MG
zviO7IG>4Cs>zI(1X_%<_d`sduu%mj~dB%HRP1$zB@wIjABewauvvypmv6zS0sZ$KF
z8l;8gw+{$QoHN>HD~|w|xG(}K-?#l#s5!C}cm`yWKzMp;z_90+xZ>JD5j3=;3cM$-
zJhTE)nQoo3PB$B_aMZ55AaKiCyT+{J;){1ywoj5ra7m2K6AZxpxnq6e)a?tw4b;qI
zs|zGZ=b(^J%#Tc5x%3;F1i2$oI>*%=auMe}S)0Bnr%`0<u=VcUlv|S9rK2C>zbIg`
zM5q3>AzhUMsO?m{z6c{t?pBWbhJ8}yxL)UNqXYJG%LHC9eRmKrAaYSF8$8KC(|4Pa
zx|)I$b#A09A=N*~rP`!BfsJJKT~Vm8zibz(0vFCMDe76?Z#d4N0t1Ww)G`~^A!~=%
zS)&lYk>2Y$Yxy#NwweN!Ect%?3P`^@ea}2lQJM}|P(#}`!>_d@wd7=@qjM4mKf&b4
zcRTUCk=#^`BY7=x&zQUI!gc~ahi?<ZzdH74o6Dwv?fT<OSt0uNGrS*%o?P1-Oc3d|
z8PjDY{BTRPBD2+$g#^|=e#sT)R~sJV<;GrV^J8ApwjZ;WLrCaXvK-}&smL9DqD&l3
z8NfKyzZ5IaSL}->@IOUy)IU*f?GlS2ORiIn2&zb;g|GOM6*5}I$#~3l?1`iIOVczS
z$=J>S1L)(ltLzdlTDz3Odezg`_YhkQu{woU%}Yn2Jol%lhm63J%oXTh+dHTqP>|B@
zNIFI&)a0ylmEVNjBn=dcVrrl^kN)zJcj2FMeN$lb-Bp75l*I+?Sjz=;J3<s@klA}d
zRaJ9P#QQKvgwGpwJ@&PfIL=LNay=$Xtw#V*^46Z;fUS9An=t_p5WfxiVMVh|Ri5$j
z*VhY%?WcRqJi>f!thu9N`}~aRwk9<_?T>SI`a|^GAsSkq;}t}7L#EPFJgCN4+*oqh
zEQKv<T+TpTy}(V+Tv0T<(xMmt25#*4x3FK}E$7Ys9XLtTnco5~mox{8Rc#Uo4Ivxm
z`tX5)lj(=iD8kHbCOBWr06%v5P2QsCQO|sg%*y;zip1?x8RX3_+S$=3JIAV0Qo&2)
z=cGRTN`vO{Yc<YM?DpW1;6Qc9T-5IYBj2A7C586z6psiI8eiwINfNYnHT*bFl`%d}
zjIXTuUi3VAn6)nLdgE^B$Ir!nA>;?eadg1rt7@`C$~9olb!9oSC??q31EZR_(9R|1
z`#LUgu_n*O_25;>0q4N8G2ebf25ky{+{j=mU^%$C@Zc8y$nUs?rqqE0Vtj*~Z|&P3
zv1BWdfblh}VroG@GjX^*+}ad7mt>00IoGB$xuFtsHk?<xf8zugAc;r}jg>LKGfAuX
z$&T*AKZnw3MTHeH(&fU45@hGxU~stdU-EByl-~?0iP@YTloP1E)HLtA59h^NxaCi^
zUVoAn>K?#rtB9J2B28{GK3~@v>Ag&Qyi03Duaa7#69ui!T>m>T7i)OZ?yF@dM+o&*
zIeeAjZOYoCWLDWnYHRQZ?2e2te<;vyJO?shjI)(8+%$P87T=1B6#G7gL~u{V32${<
zCN4m(rfvQ1s#i(M`7-x(XKtPjh@r_~nq)|D(_!j;!HI_d!rpraG}UG6!%<XJKosc+
zs30g+r70~c(nPw{08vp|43XZ6bfgLh2nbOD0Vxq7AT`njM5Ie7fzW$G4Uoj|yfb%-
z&fF>9xpRN_e*Td_4m*c)_St*wwby!{XI0{dEzC_voRqTD1aAxEP04BS2J)#3lk>O0
zjkWwGB-=1I?c;hXXF{`K;_3Bybr(K{K<JC>6bHt|TY@mi03=Jb7BREwW=H;Pva?Kj
z?;9gG2DQ6$7j1KId$>&qx*5iZ-KR#rQ^tlUC2tnW85=unuDN4s7bxq>FUWvuN-TFy
zJ(9)Blg_aBfv+UUMz&ZBv%;Sh*n3oG?DCl#lTmvi5gXq?gTk94rG2hkiY6p}Z7GRN
zN7CGm;#gjcoPpvtz1}4M6C^X!04H&YV{?#PI6flOqt(4(sobDj8hs}a$0RahAWa7X
zu@oU2yeRj8rx@7t4YWTAQ1w&wdOEKc{k&*ti~rTcd9OatzEEDn*RQv<^B^U@l~{0G
z1MLfRpPMA0Ja)8bou_7i+cgEHQ!B8Su-8+#D|z|t24C>;&$M=@PM2LTU-p>Rvf)ez
z77cIOtl1~Xn%)GdD2On#Bt9V-naU_TD%C6IxLCkjWSzFCS5Kt8zsq~;XqFpS-G_Cy
zd4A#zWH3jDH2E<B9b=xJ^<<VfR0b%BTt2w=?!&N7#iH!uKBp8x-dp|eP0yLQgw@9?
zJicILcaPLyZSMBEyQeUHyUa4V1C{esU)9HUBH)LywCA7s{8xd24RVxNZYVVf4PJ*A
zy1IS-?xX9^ij4f#@0myyQZ5iZ9{XzMkRwfq8DVe8LApbmO6I%Tp|E^-j_6cdo7s6C
zwS&`@4rdo0t^zt0kN*56A{2XvuTVBO$#@~J<1j?iyZCGhLCr>Eko5jdNyTBEJ?;W|
z0D$s4;T<-NxNN6c&+-Eq22;rbM{BzrOhuF?Yi<K$LBYY>>if9ciO5G3DdLJ2T&lj7
zx7%0hb!n(gN!Q}t(%=dZ`#J7_gn-XdO^#h^AO7Hn|06&Dp0Dms0x%fRZ22_)9n6kl
zF6dRn*`Imst!de*TQ4k>qh0OCZnPqw6hMBoEmzoKGg#ch<^>{Z$$EUq&_^+YpOlnT
zNR1;vWImNmKHqkhP)|AL=stPW`Bv0(>LIK91Sn3b&Weq6q~sOg(${Nx_+?&{c!kpq
z#=zAWjP2rY^ZCHc1nAr2IHQLTa5YJu&g}4MdAY4bV%W5#zN|ess);C|fGbF}I39SU
z1u-zK1i)x|z0jXN*JAlGb9keON4JC=jkfShP$gfU=(x7)-oUXbIGa1{H+|4O#v~!i
z2J7M4q0?gyP`j9v=Xr)0-Y5C^q<Ju(D1OUWRB%b;>b_$GG7c5@gR(mc9{6eViFDZa
z=tV<{qO<V$ff<+8*fA94xB@}FY0047klZ6}b%^L|AXYdImmj~6t)YAJ?qhJl4bNpr
zpzO;A=3XR&CsdQD8zwzMceQ7%yA&?JzpV6KnRamlchr75N<weHAF5HCUXR3{{Ss$k
zVd>xvU-x!f)HD6O0GCdxpV}95+#xSKZ&c@l*{w1H>(YE`cE;n^BlGgn!|dFCS=vfX
zBfXI)^qOCkC{!I~yRt7{2gKIJ+5FsZ<k%Ydxtai%3GS}ol<+V{Rec|6MCHVnkM9~z
zhiMAjZfjG!y3cLjtwf&_Y5$C%)L#fs{fR@Jom+&R=tpevZ0pf8;#U>zk()=V0a>py
zOxM7r??3;SL?8btyY$c4us?5WA=ixG7Lp&J76W%c8gTjk^A~!=?<Q9s@Ymm_+IObp
z&&|r;`(1PxS-^h-*Mpqo1!Acd$=ipU09>D`aw6tWxO~<<Hns3CrOe^-GfgYeM6zL1
zEgx|?O$cMlx{yb9hO5nZbu9*XPC~xw`x~6+mYcjc$%|P}^s=cS={9+zBHc#l#{t%X
zG@b<<hPxyK8*zao+z4<Kd@Ka<JZwHJT_g}M>)aMz>j3)X5p1kox={%uL8lNYSrE;y
zB<KlGbt|%uA0hS#nZbxe*Et3aO@)zNbb|DrT_3oneT`pP%d=MN@l7XmTUTX?2ueKS
z#R=udBmtuY&E5eGQql8Kkbti8X)JN3QQV}-g!_S{7q^8DE~Kf{rr|sn-K#b(RxTrJ
z+$wWXCV6mNfmgI4neS{L+?PHN*3w?m<!j`@6teb-6OeGOqLG4*g5Av^&WU1Nnq9rh
zqGVwWl2Uo-3dQ<`nP<d2@?Bk)g~Jnd53wPA3MRO0)9;<y(3Q`;x@v45j<23#239=a
z_{u&}FK@^>%%!r@F5kp>Y(iJnHU<bY@_I5iy6+$w5l68Nh~xN$9i^E@1ygY}k10yu
z@+y5uwiAy<tANZWs|OBOZJnN4@ISrV(X<kx=(MXq{)`(SNJap7H@>Y8=1*yY#Ffpd
zsn$?3Bs9(ue)WFOJ<Tn(x9Xb@$3BJ#8r4&U$0czt<2jMlas{wf!V<q$w`<|umFw>-
zdpAVwtk;&>mrqgZtIP!KO$3i~jy#gG<c)3C1<hwjD6<xmT0&P84fY@wFUk0L)O^t1
zUZqQJ?gZeA-o{d~O2P~)Y-orx00mIO`$!tTytV0eO17{yyHYmm8%R7E?ypI7gbPiK
zs|VlvvT-bbPFSm)ed`q!UEOyq0Hu_U({a9Nnjt@6ZKV#_+jt_nfj<OEe7caj71CpZ
zQhnc7I&zx-Xj6%?=gW(n1=DWE>6*Bg{a^w;;!;tjHo0Z2y#t{^+-Qj?6_w<0FfJZZ
z8E*w(v!iXK?VtzygQH{vojG15ShAg2645OB21*AI3)B`d`^xiCP(&MkI;K0~Y(e|q
zn==3D-y;@iN<bFS37e?(M{nc>vqGN?jF-Qe`{uw_(&l+g2;q>#ebe6Zlm|}n>X+zO
z3nj@q6u=)=Ig$t{6BQ##MtKYcFafFkMqFZmErrrpm?UX|#>g!yxVh;3(^@v?GnZNj
z?&pBD%LCxqa=M}^OpySDDU)^_3s7tu0E+DcD8<MEm~cT_-#{-CQB(k@OUjom-yCRb
z#t}px)!fwK;ZKy)aWZkb8?B)^D19IKH&5lLZ<fl5B&L&Lv`}8rZy>t60A&~89tE!e
zblvlFP%;ld*JXn|)jPmX5%%W>T-7iDQ}y91mJATt6acm^qr3_2K+`u+lkpVm9)Qnw
z2oErJ?*NQlSxK5m?2or!p$Y=>y?^J5uKc>7Up@3|4E>sA|99=sdwdk1*q04(m^TJU
z%I1s=V-w0aCE`YkFCKq%L>GjruaS=cJwVP6%;~&tp%wZ+od*F)m7jW{f8<??vQJQN
z?-=+0nU>Z5RFVyp69{^?L?fiJbn{lqU5dJXeCP+R3hkW2#qZhQ4z)8bcsPO1j$V^j
zd@wWPcF*z+6la%d>xRRN^G_#iGq<7_b}F|~0-jsr&?eXlheIlL(f?#+T9XxTS@#AH
zf7llHNlaP1?E4Xff&)qFnfoxK7QNW@ta|K8If+^jxkQ(Uh!rc84)Dmwdj{2)%X%1$
zUX>m&-aIx3la7`;xb;AGOMT|Lm<rp~YPt<Lo4#)}rXL?%kvt4wv6Zgc^;U`cL(}8d
zghTz0lix8xR7-1#!;fLvzJ`5@HwWS^Ac|6^!%ex0hRyr3SJ6&~+3uS##qZItOw0D<
zXvSL;@c}|arKM_fa8%Yn7Aei@D<<l#Uztx0>BaNm?e2nFL(yv>N#W0GoIoDT7U6t;
zj>rG-IWQ%0zxtBJ^U4wAn2z>0(3~!D4cmw(X43>bO~#UH#)HXTauG5JR^Q7d#4qO^
zWDiza4zq7vC^cb{ZhzKpalkmNgC}U2p+U|FMnpI7m^6wjD+x4ke@?S@DzLAN$=S@W
z5zr_0btaq;74K0$zw1{s7Nhh={6m!13EheScd$v9`5M?^egh7!C`-;B9Zh#*(;byC
znIUJTQIyRKO96N0#omasKlgf72@tN#u)~NcgI*ITX+dr`nvzQwMWUw!*ziWZ4O(V=
za}donpe*Dd4~<=vmj1Glvgla4q_%14PG8FH_Pl>JI1~T};*X!4oOu3jk@1yr&|vjd
zKtatOxlqc#BLi!8>wGNFmqrrSCN4ESk1m?$_M;UdPR|-@fnwT>q7zrHevN62x_30?
z%2)eSEIYltxsu+aTeGq0Eo(kVH|WR;42Eqlhor;N;WCJ&s#%0<?y2q<^zD^EzQKMH
zKF{g}wK1yO7jLZAtPmB~NA?SX#mNWK;XZop<FDbn$lSZJfmVPN3Y!mX)QGejb|GX8
zcU#5oR+ae#--@H-5?D1$rl*@83A{4)MEwQXM(Ri*j0?n85~pEd9UA$Cc~UtMZ#??h
z*rO{&AAxJ0$Du!c=<z>iG3;xed_CjP@vGHq1+KfV1G=nMzaDsoILlV<n5f3N=0G`8
zs9Z%DG4I?ACfbJHsTzFcz{lN>VMv3DGp8tW$U9uy4Ih>*{L}z?<gt3>z-1V(P}?q{
zqc%EcXimL>#+|t0xXg32)U<HLHsL%3a*WSFf3-Ba%4u7ky~v}*3fwCV;B@lG(P5dy
z3;-xU%HEf$ZFNSd!#c_5&>4>eM4R=lz&X8xBTfzslBw>VMsEyrdvbWpAj=DD3)6ll
z&fR*InkR~fVj)ONF&{)Mc9Py6Kx6nxeFJHY<&O$hXMFxFRGuzAl7+jMz;osBCl}O~
zL+tm>a>#>x@QY_Nr?R|`G6p=alenzB@!8H<&U`CEC?&cG?O43LPYuA|All<-1v}tN
zWAD6uaWvMxjTA*q-q$aYyi?Wv&YuGGuNtdA_BzFZdLCuH&|M-qoq8WsN4MWb#POwO
zamC{gLQX?;#0ff(_P7}xdHcH4!P~q0>tfEEOz|uQ>}Cag6eexGU{?Lf5Y^9wwf@oX
z?D>za{l6oMX7`f|$6s4;ME=0bHWL3+Xy(!7iuJC1GhIh$`$%c%$FZO%UJD=DJ`MYx
z3!R@G)^c)KWgk{+IjH08oZ;A$Fc9PH(*GWdi)>69Jwnz*qDQmd4bJIo^$3EvxGxn>
zyB|ITDr21xWZYs~Evq5<7!W^)<t}h1Hz3&{_Bg-lWDhB?9m8!Wwe0k=!a3o?hd%7v
zF;SJ@IM<RSseAn!NGhp}J=&wgK+bl!+gBk&EhQ#D=)(T8>JN-Z{q=;lz1>@ohUPoq
zQ0UF~YuQy=VWyO<q18uOfi6jcsHZL&eXCa|QrDP60m_w0h_bMqvT)xz$>$Dyyu}S`
zc}-f~$Vlwo9?Jj&wzrz}Q;J5qT*mqix>}aAWo(MPXEkqJ=%<d(%<h98p~xI~g$wrQ
z>&nE!g4_qR><>QutaCHFIDK-hst{J2ye~P<vQ$r|SEDPd#c{*<<h!7+EcvJx=+F2N
zFmEY7DuchQpK|%df`R=iR4J<4qOjEF>zjP86JFbAJwURK4`!xXWaf-I+S{{ZvbM`?
z2b6`s<n)(*6(5Ud^;{`f){MW&RBX{CdPJ)A)=8g<9P_GNyZ%ZHHqH+Y^E#m7WpX%P
zLR}@O$=5}dPPqht;3CI6cR+UEKx77HvQdp(eWh`%g7%%)_M1`{7cW9oPdqR=IdSsH
zMfxe|$*wE>w%fwHt=~X?5sC>^Acs@i<a<XI<cFF>tq4;$&#vljCD3-+NFl6yHj(kj
zYSh-(%vSKI?(U*VGf<ppkfupnMdI(q3If4Z@Jcb1Q~%c)fA!=4^P|TSRfpd`i&Lip
zIM^=9EBspOzw^w0@1|${Y~c7GN~(%={k&}cUu3jy{lS@gWB#Y87RYyS*VZn%R*0HG
z(aoHDOU=>m6?cS-zZ}h*);g0lA6XGjCw8Uvr1wCxztR^O=rK=3laoE3x%Q+)alZXP
z|Ch)W)~jdY+npHWqWfnI?ds`-L2B2=;B<RyH_1!iK&RnoDiY9HdNqZnf3Xiy>7TgW
zH;xYkv$(=M;jVs?rls#l8%@n|=q8u@r>TLgRr_?D4>Z|5Zcx5@;EFP_WkC+$KZop?
zM3*u|E-XZT%#N9R$>2E4VR!xVl<s9~<CjE4{ZK3?IQSK*sWd}GRRYKD(EplzmP@cM
znoO7RG?Rby9Aaw3jdEaU5;jAFKHiQcmX9OCeWl#S6l93~#qj&D$-8=KQf4<4^L5X#
zUp9*%hH)P8ytDwp9;xQ?8OKIgalh}ToGWlBk<If&xG^8K^N{z?%xP>4^}_5+Ufg=)
zd?RmEqxn2+gO0{ZABLS-GX!*^fQ6n^2_C=V=|e-xfOmKKzkzsOx_S)I_L)qyFW?$6
znOf*=z?3evOk+P_57xa}NU{fr4138*D+F_xhl{0WDvy2Uqx=&EXpahk_`T2ck4k<7
zSE1j)@Bl^eyVJhGCD(xu*;T^(U*<Y&C9$&?OK)+Q$enLty^t&RdGX`kTu{8`hrMrT
z^Oo4QCSL;tb{oH;C={`!wGv2U?p{xeHdr&SEbS2=hozuxD??7joU!Z>xUBv1TA<eP
zN0;BELhDyR3*OC{S#zU8Fsh-pX~t%AKW$WV$=GP*SR$L&NjIM4hdf~d($yi|Gm-jM
zb+!OgVZ7|6*O_c|lt4vQ47D^TkXL7L1hJQksaUUqCD_=Se-wA9c|QUO3O*ukQD_dx
ztxYgZPW_n|-yT)(fiMaWvAijDiT3y<nJ{G5?@B{RFM*9Q8H6Ev`cKh&?#;wpDN6pa
zyXA$4<-ZaSi;gZU(~)huM>S`9-45P~)~b%zWqAXPi1_Cqg?>q3ai++TgREs%LaMDY
z&Xk1eia??T49hB``S+BSsAbUw9$~(FmaZii?OH;X9HU-a-suMj5*3xjad)ahd2%h6
zM&c}uRS!MA8&j~S?o`06DgYA+*)zVfP3Z;1y%v!a$482LBesAwRgVuq8KF9KX*6I4
zG|&F;Z?>R}2-sourlyBi_4?)&=A=Ap<s17*d_>>c!y-3%mag$E8u{a0#{Sq1bRT}T
z`u%|T)f0c18ozqt|NWj2Eg>ra!i78<fUFgM=gdOS56fF2n2EyQRV`i!aY#y*>$bRB
z<<N1ORpDvg$Aiq*eE!|p?VpW){k8A?^@vzgtpf668jxX!lD>gvD`0gCqsl)VnWY<{
z9rVkjv27M^-t@~~GT4SxB33hpP%UhAzJN;*%s}Hx9xtby0RxnSAy)uy=$G8Ft4k5A
z_1><kMveMqF8SN0pf<NCY55p3zid8#@N!dhdU_)i?rct2BHSLe@V$L~*!si_vGA)u
z@7yNs0)o8>Sbgh{kqe(@!tzR?Yy$hZ07lt9r~dHSmFXIz^z|iwDz|QSns@UkHWs;1
zIqw;bYS9f(O^%XuTw3?LFTD%p;b*GmkdtStjS){;gk;>SfJ``dp=7Rb+T1%Fm33S>
zdIHUF1*<fo;$tKQdoXBTSnV<HTjqZ8>h|!1y-d{~r&*>wJKivsu-+ff6c+P)jR@P6
z<GUEy)J>+Zzep|;T%8!O$g5SYvwY<Ot7c@mkP%wcGG1|NSGn^bK-}&05z0I=NSYea
z#C}RgcD+xm9=td)t!#ad(wvjD-pda|&WW!3JM3<pYd4b_K!~3~RA7L@Yfxbd11Tq3
zk%L%Pw^*&2JHKNa;?FinPQqV{@9yL%H}I<#GthW@NHP0p<KCAh%H~sP68r$1p%KOH
zEB^&kT27*5$fWL$Pv;Y={L~<#GV=F3Dpe<|PjkkZkE%$)&4Qz!`R_9IEsb&!wE-?x
zsxGr9PkrU_PgQ5&7n&;KJgT4FO==MEjh6#AE_x>zxk5wDbwk&wHZ2-tm$4CvcJpBw
zK$c0v8asRtp=8{9$eAl(<awYUOtU>h$8v?d=5NssUEZGE&Rk+$dIi&^&{I&f{Zr^5
zUj#`v+}DXXFp*M4Y8++lsg@j{c#wKicGKS~OF|m+tiyTjdN({?(_HXqwM8S|lX&@=
zuRe1fG&<ulIh7<esO^FnA5lH(A}SkV?PQHIoN4@mu9>Ov0~2(CON>GnO1c4nOe>}V
zN%4~~h)=Yp0YT%=Z56es&8#;^HAF6*OvTCtH#K@(Xdmljype2jbmTKy;(E!oqbj07
z@HNG|XE%$dd;z_S2<a-A)*N{OA&Z_&!!<i~Ux>oILn5mX70uMaHQ~BlKj>Izv;;-L
z*8T19JpSl|v4JepKH8wzD9^5C`~3jX$|;Yjhn~$&akl^m^0u!@FO6v~0T|7e9t1e7
z8PDYAXC<nC2M*s<_$pAgiu>BYU2EnDIW_ohjP<Ls`A#84dez|om2zslycU*SE2ph|
zUygw}zQl1?*0+vhxxwww>dXm)IT}xbChe2#Z<YO0x|eX>4U22fXsrzJ5mwW(?W5cF
zi57Ile*1-b6ST$SSIL*%mwdo+r9v;ywqt7HvMXy}rlgz67O9AOn;!PoUQvE}iRkGE
z0*kTJ2PvrJwp9<uHLyo|#dg`i9wMIL5xJ5!t{c8WF~t!8Xf;8!laWX3UYG`hz2eN`
zVqC#sz_{tXwXB3QpZ7jLyn$w>DD+T-(GJ|BVZY5{Zdpt&qm_fa|5phKug(=gExpc8
z#e@#cEYBl74K%@CfMHEOUhnOQ%_I4}8GIWtzw(^6X@p^<Jo3NmeXF;68)ZU9rtfas
zxu_`?>n|351JO^_lr8(F0QHRkfr3V8$vpp2&&^#UT^?VFeoX$e9-ZO<xKLTdU|>TT
z<5Hrq>yyjYt#+T(K^{BcMoG>&amhfFrY-$?+PQ9u4E<w?f<;<a&XCqCRrecj>fff^
zOWU%ulGU@}Klg!cxpOSdOL`Y}fG8JYUf<tlDa&B|@NPu6-YbY3wB*j+hv8iGZP$vf
ze#mj)V+=FyakEt3s^ep??%9B2akAws^buPQB9hNThgj|$V|Aq8!|J-XblL<i&Oavv
zP_dHR5a-SbR^p)B7d63&MZRsdiX>q`X3O{L+pZ0`Nwdcex={(NQqlh)`KIO7TeEL3
zGfZ#G-i4FI!18aVV%wm0N)|{KnjkEx@<5nSp5M4=6`D^jY?L-@&rz9DOlwb_N^&jp
zT6*=~jDs#<wcx==m5b%YIEqwvfv=q3_z)79F6Bh(>Kn_s3nJo$^>5ysjXwp$Ji4XI
zMZf3D?0)ku$}z<FlB8CTE1J&J@MceGUi4>_tqtMqgDHUg9K!zf;Md0|9QBUyhPm93
zHCOWIuT%A{1wa=8ek5VPDsWAn84bW??Afi>0%(%Qs^Jq1D~CKqCR$v`-ULgP#{L1T
zv*qvJ?ZUr-Ht&xGa`v4GS?U@&(JFL$^4gI5UMI;Y<qMu&6tf&pk89O^idrz{;M)Vp
z@zX?u+2X6YZL69*It}_2U&VW`M~9n*uX#O`wf!oC58XjTNUL{YP+_&qed}RWKASx&
zft`*65}!IRL>Jyj;Yx4yy_Y1cnCcNx%R4tJ|E_0|qDkDXqn#@4`6`oMn_o7qa2Es=
zrg6A-0(7ZmZ>DiFvQrH`l|N{o<B<t>(afHkTCu{rC(pUqj>m?(NwQNe6Kl&0Am%s+
zTxbfy&4Ver(Sv<Xw7GzAPQXL`W&$Sh*|8X1^SOMM%Nyj(TiU`}*Y!h;dhdbud_`vV
z)b<hhclOVr$QJ=C5x1a;IVYPKMR<NULDsGSf$<EZ1$f)i7y+2lq_^^SP3I5)y%w`{
zHWAIcCgI3gj@QVsHeH%}=91MZ0DNPW@(mPp4!q6xNEXJuKvNf&q*Y4*Bl13m-+{zT
z-MijN;nD<jU^Wt5TmU>$y;;uj9D4$^*mA#qu4XAZ1re_sS4WTFoDq(Txgx(K%Jai%
zW52&{(BP}$?{9PVyQ_M^l=|xw|M7%=^~Ik?>#x4}+kNr6<bD6W3e~E&#i980ZWZs<
zOA}G&Laxs6Hi7JZ554$(fcNKe=)f-G@YatUnPNw9KXW$R9=m!31Bj|<Q9?jw|H>rg
zAEDB>Fl+-9?+)4oQUpXo#Nbzj0PyAj1-T<(PW0Z2{ow#;3mM1(y+ihUGrVZ;n^tk}
zYYpU>+C#9&q(3uR(i{*0+N$GcQID>^YD*Y#<WX{-wZQ8tr)_$bwVw&^q4bsCI#%)c
z-l6Hd!&A?K(h<RYieCT+H7!lsV0W>nX6Q!V<i~A;aH&KC`JinRu<ba1Xn(S?drLFz
z<a=Ut;?BH#Eps~C`bS3=sQks8MZDeD_&Z6*FFd}Hdng2StlMOL0E0$C=3am9_m-6_
zfG?~q%^~`Cvs`0$ug){%YQAAUT<_1<LfA9@dH{=I9#Zbxd}(u^Jwpdy5}7cyvbIHp
zv^Tj}?IG6?v|wJ<N?CIYK%A9*-d~s`1yL__=r)HyOb6$6xg;+|co;!t4|cK3&eONO
z(Req?OLnB4fvu2lV(MPh7RYQ8bCSR%IxhrSsF^e+3T69)*3|YWK&(u=e{rG#dL%P+
z)KZ{%XDiGtAFgZAyR=cG%nEy^O__$}x_7_?#gAXG5N&Ssvb#DgXAl!PI*QLi-lAD%
zgeRujBV`71usfG1REdp3tlg4|c`Y$ICVG5d#;%dJ#r#9OgTMNw^<4RdtFA%!E_!uT
zst-oLEPL%9dPb`DpMOMYG>&32bevZgATe-HZJ}P^=GnWr;?;MxtU6c+q;zhY*>h`@
z-(%x)pWvk9?yLth6-pE=#YR)EI*s#)ZuBns^2>|PJsB3Pb-k}USld-B*dQik&622h
z{4gjNC%fW?hFek{3v(c5vtP3d9#qr(YGH*n#(BH~v@hO)^DGWQ^>0yO#pFkx_L!9=
zmCa6!U#?k-YEDA=B4KjdheD;Wuaa)9gK?skebxH5Fv-Bef^{Ci37Yq3d)J$0j+ag@
zL)nXTfPyANuR2V)ueP=Da6Oz+s#YM5P;Lz1i7Bw4tA3`irEy&`q)qx2jWfII6rb(w
zwMWVAYt;Z*Qnsz;ehx*dnAWlih_<5^vB#}Ei+H=_I^0|~Q5Sf&6dRQXgg^~U`AL^9
z>kYafq0AE`nI2eCM0(AF*U&eRP&sW7nO;><0Cv!p_hZ)WD=M`Y7e7Vp*m|sZMY)@I
zJ-~q=T-IDS{n5tki*pc_w-e)%3SBB3r)u3K@qD0h^=~N2)o<0F4Qy{vJ{xJs#QQGK
z$0tl9d2u!2wDpu5`0~ER8E6BImkb_XSP3^dHfmwnGLd7Qj~gjiVNPlr9}nNhXYpW7
z_Y{b_J$ZR@j}9+=ouMbJk^2amq|!t?)guGvgC9E8Yj$nWye)*M&9qSe-V+H>l9N~v
zzt1$h-8g;MUwLcF6RjYZy(rU`g$B$Kv&Y(2jYNFyMD_h$Jy-^PZ?)3v`3zPfEj@QW
zNgsdEO_!a4b9rCF#WR?vtGEIaj;m~z1l#a)aUc(RH@E|md*~|3Hewwv+_2{0e<q=^
zER`uFd%EQG(1@;33fqK+mO!LN{L!gkjqS4#dwV7uUsa%XF-brhF0w%2e=OZ`V7gU$
zNnQQzby-|U<;;|mOhNRqQz>R+cFzJY`WuO)HU|(}Zl#%51vTyzCFad@W6H`Axu(t6
zE8)ZJaX7Wj@v(a?30wy63L-|W1J6H{ag-avj&$%-%V=mHJ6oRB1*{}Z8?lCehnpH&
z!HGgJp)EIJ>)h8o2Pry=q8?f4w(zXun4r-X`Qc}BntP(uT&6@Hb*!?ygnX$zPPvjy
z&<zIUCtnxE<__Vl?34R14Ew!2^}g(V{Gn@F7`c~rt;Ql*ffx4>s@pZY{JG}}Gl#Va
zg%HEO2X50O!JX_$+s<A|?R?zwbHT#_*Ckj5H97grZ!;eCM$?gVK3MJPK`fS1oe6em
zH-!;nwTeKsJK`M;ou{LVW1mU~pPdJB39jk$I1jGm;O+2J{)h&6L}Eb&vhe<nGfiRH
zJDaT;Sx+E_Q2}b`syTvoI+ko<brf(`u{14hV4r7jY_QA1mLef}5cvfP#PD7NXz#N?
zpuTCGd4~b;j!W?rRgk-@3D-9eLu5VwI<OCEwa1bpz~fGyDzw0FplxwEU;<v@P5(Rp
zNRQrD)d@v>Uk!N5SCN#Yv~QsBOd1DAFnCrHkb>wzl34frBbVY94MZdal9GWsSRBZm
zi@0Ls_g7+6NIPZcJrF6D{k9@w<IcMZv;8&Cm)p!W4pUxA#Xen&^1u5?p-q&B#?sEe
zJ)Q?pQ{OCn14VZMK=u2v6e5tQAsPYyq4{@SJ`jg{*Fy7jK$1lO=-Y%Vvfj)ZsDUg4
zrm$HeNtM5O;Mj37AfuPTjiMbc!jeb$iAknFT~a1pfWB!01rk2L?Iy@tV#t|LKqDkD
z9}T3kC})zuYYjjyu+C@zNdEoKANzGVzk27_*!f#COWT$p3R!p>cnO_2bV<s7>zFW|
zV#!Im9G5HH7yf+=)c=0$98emzobtE~yYUUgPm4^d=Bt_b;lQI~F@3_gYQqoE6Eav9
z3P-P^ieJuGbmZ;(fjm497ZB1t!7#-GcZW>1k0R1?{XNmJI~5w6Xl^31jtLMaLj?(C
z?0g!uue3Pokn|<_j+2LDQQ2~e{%WkUfo5HsYRd_&nk;NlA<ftd&elye*+A%E1rQBj
zZt00JFoQ#_U~+3827XatyRbB$Co=Kzjp$Xj=zDgAeG@)U=nGv_XBtt=v*5dC=rfh^
zKq-U|sqp6m1$QQnAKQD`beq5Wfn?3Y3@)-Q0l#N%M`e!xYh^SLsoG!m*pGMPQ?cst
zYxnM@3Yk8C|Ecp?+v|te^V4X1FcTm@PlJn7q=eGmY<si44s)TRrIU8|S2!oM6bp`^
zWY##^j&K~gU}R__0aZXw;;Cr-f;}cZIe88UY%Z_Y#s(`(nmX|9Cw@j3YL@8tNq^kn
zg0Bx<UfySSi`_Z5sFhDg?s5x#hA0O?y;*dHv9%!m;*5GsJzl39al037log;lvhZQ`
zvYDPVj60O`{<7BHidouT*p5~D_@MYykX|p`mBEjF^&sFj&YcepuR5f^6gp>e@%-wD
z`dv<!^s1FW@4zOuGL$T8e-<D*W6kO_y+=E7O;<<`aXb4UL$Y<oA(}{e+sqWQer3IU
z)OcXq-w8__p(66eSHh#+vhWTUfeN(9aQG=ezhrHlS61DW&n$7T2FvPQt>wcxQ-^gX
z-tQIFZNRdgOw)S*zF#-lg<70z-`|h(E9hw#fedSM-r=TqzV<@w!|mfkCymP+5sV|A
zRSg0<CGng&yKae}=UN%ccyeDmeNaq{3TQEqcSB$5t39~DB9rlQU)L}m=irU0$Eg&K
z&e@A`BrDr>G`(G8>lL~0{w_oPc0GdXhbLZZPyP(J4k!v8v<1w4hB#d^E?OwT;Ac`_
z-oKTcTG$=72DrL*dTl`hRLG5+5e*M-Ifgv+K-Zpu%+k*HEy*xMno9GC^L6TJVHiKJ
zD6TcK-cH|MojClaLhLKoN4B^z;Ouf%yuH4n3Wf2IxlQlq9%ZEAVGGBkWxM;^R?aIY
z+mxr>_a!{Zyu|6wzn`Kz?x)fO{VKnB%)V#PY;ZQNRUoVXZIhE=w|Qv*&jX+PcHOG`
z<21&e?kh~JL0aL3YW+3Xp8KzKz_x(R8*y=IV1grfK&#j7(8i&~Q#~2`s$%temf^Nx
zpC86asffW&8e2aaawmF?5fs4DR>x|qBl;B<J}&c3iH#~mi!2{UNF^T2+9mK^b#5>D
zG=vp!gay7O3lrp$8#y?b?Co#~vmGan!v=6A!h@3LUM@;)2ezL^zX5?JZ~IO3!tji4
z+J-ciZy*sL8QRwY=?)#H<%tKz_<L(~l<Et+3S*2hCW0;znwHQVGO~kOzw8Dl(O<Q^
zHy28xb!`UN-QShQ+OB;hNP-c`)2=*nw!)m9Y+WXgRf&>%<)m9#;pkFLw~|7?S2K<B
zshq8-&ef*C^8qRMcHeQ(y*%4^Vi$gcZ0G}U`=Eu5G@&Fs5juDKKMP8rZ<6-_Zdu|C
z)e}sMimCowBQ-`(i7`QK0MBgaqC8NN><0ye|CCqxFKEZ}Z)HpWwY>1p9{VrklZR*r
z$;cahiozaBv@}~Uk<rS>5+^k|&&Eha#q}g?_p61GU>_<caWVrGQR=mpdfZgE*_^US
zWq0BJkj3~5rf>5FEUv#hU$6M3xxnn9|NTx)BIkRe@(XL#c-^vA3|CpG@-5GOV`KgY
zs>7~)*<d=dR=+~r(jv74_+G;$zL}>TtkueoJpzHGgbnEnCCQdWvx8+aVVXy!9@IZi
zd?nwj?_*EKwz?s0-JfM=@jw>HNQk|Ew4>iSpl}d03_D8)#A;6<DcS;_QVnfc>GU^0
z0Gp23OF*E!7D=`q_Y=@2SLC%5_(9~*8G^NEtU(2zcuc)(xx7mXQ_+%&bMq#Fj`3QL
zPh#BF)zFcJX)Kek)SDJS(YHB2$MXc)cxLz7lCLUnP7si5{it6cQLDA&Ot4H%^{NLy
z7dY@Riqp61W!Z!qrJMrh>ns_5T?iO6K!HU8x^`osrt?iLUwN?d8BhLeB;%Cvb`5Fj
z-Oma`p394O%#3(dK;s^uhdwMy3qT?Cs8U^K@+RWo9QY6h%sCH~0Fcjam@V76rLV&H
zuH(e*w5r@gcfV-X2L;4<q0^3kpk3^yoFb09b@0eDTJ0<887uBnq(c(S9$?PcGtP@I
z6`l6cOko6RpI)Ml%AJCy^(*ocZ|pXrSPG!cCRXR;wIo4yNxB8*^8-5qBbHfPq4(1Z
zipr{Au-X`krD;v_rwMzhURHRJ2KA8JmaI8q9D8xJ)n5WkNwCV>R2ottjz;F&vnMB8
z9VzQ75{gnx_1h@S*LBn-UJl8Mila5W-IDR49mO&uztG<Vi~;aG$f?`d&8Nht^2w3X
z)QiXiR|{||uhfsA3ShOYeTw&@&wH`G)OeX<+<jZ|wVd;Nd&Q>xu}?Lor16JR*0Et*
zP<9F(0n5|q&u{OZ`r!OrvPrqIeA}{|PegjZha>k8hk-Q2_w}bYw^GA*w|rm-K-zQ5
zea*@wMVP+=xrCkZ*;Xo8+7OC`jivh-4@kE&N#4xZz5ZTHi;OMl^G&g0O))tGde9;w
zdn<0uI`<?1!>H#NU)ccx$sl?FS#caHL{1woECX0ZEj^uMyw@fgl1IgQWS~cZu#~}P
zH$UOsvud|W?dpkVwpPxR8w!;1M)2Zg77!uUO~=T#z{W0EV#%MYgT%8`31|;7iD<kk
z>28izbi7yUc3^a5xw`&z94oW*-b_zcmC>nf<Wl)JP&bOUcR<}%TyhRsuQOUHS>s`{
zlkj7wg2x$%TU3y1CxtHdw^GU>dT-dSfEYmv2HnR17wNk>!cPn#`_LZA+tT&`0K&%p
z4%R*Y_nX`sCaYhAvfF=&=FmMRlh{*ww_hncIBAUKLhk{@r=0<)=>T-~vORKR+~XT4
zQ#=<0JSxcs{Q90AF!2bOCKOA)Mp60(y22|3RHB~vtrE3O1gQB}qhDR{YZUzNm<X>+
zhydx)BD?8zf;H@6%8;1s$CwJvODFF=)S$q0{=Ee0e}SL+M<4i4tcyR_5@G+5CV}?+
z@yH8i1X<wqu$8CIvIYj$zK^4b(S6~N;qLowg$Zj^%ZXdEU_NDldYxFCexC#^blxG2
z`TJy_z;IgaVo~mIME1Ad)d1~@*M|85#xhQJjfTa41F5#$az$>-uK?=DUsJM>bB>uh
z@~y(<+d@F(78_`=glNj%8D8gK<XPF_+yL)x0V`~?L;AXaR%)AQC#R*kmY>{Uq4o2@
zObe}IRC^SszYQ*CNKJ<=_6n@-Hh_qXTvim#$6JtRAha)izKA?eAI%<DIKOVv)ZMa*
z>O=eUZ|1Yy=&DOfo&A`0=Y!r-^%t@aDZfPyz=34WC*y(*a1>PplLZofHHY^a47_M2
zsaTb8&kFyEC2{-3(CVweJ;RK`KDDi2ED$A*H~9t%t$J?3qnjf~)9L`=7gZj+OY1h-
zb~uSNw_FE4iR!m+F^_ZJS61u_I;n66ep>E0c=*L$O<~aCR3+|ZyKW7Nc5W-rqqerB
ztfSsXeKQhQjZ;Xm1gj8Yjwb005IcU|zrlQrBHiz+5h<Hy%j)g6ktN&7_#`--oxW3{
z-(cSCs$S4!5z3UDII#<NX{ucJzmTdV+Au0aTux?xROy!!=fbpT{_s6Z+ue&Za*nh5
zPx5UQ1`bz=Akq?c$Q8&7ZA_5gHrCkAorU$-O{Og>qc@7W3#4WyWI7y1&l9JE2$7OZ
z^T^{%r7Opvx6S7`JSu%nJh<kj>bt@o?9y4CBH$leNyC{9N$J?ebFTCl(l8%~?xDiO
zNTBAML$P_HO}xP63{x%biu*YBtC&k85M^LH7c50%ooR|C0|D*3ofJX&7Rs6T_|0gP
z1M(eW73)*Ah`qO3<~!tD^PEr;TV1h({1CZYGgVYM2Y1vmu3YekqJWgzM<C4oWefY=
zV6mgs)dIpXu|NnM>yD+^0RDX(Q1HWDoCLM(QM8l{TcYtcIQA=_i0?im(AQwr@{oHe
zJ36NkEpR16t64?OrQ+RwB)hK~X(~3-*QKzIP!%c8-{3F$Mf#bd+$Mtx;}@%z9PxVX
zb5U&?a*iMMd-HA23>+>x*LFj+5X&A<8OYJo#4hYC`^bbvNb?#tnfP<PBPTM=JQk;u
zX^usB+}?T;P=03o*ovN<gflqYJ7O057J;3${*A>kg8l~LLdVqddk!_B4vdkGy&aHY
zX1fAy!8O=dhG1Uos`f=w?h1|UJ63C29rfV^%j8**NURyrB)nQblVV5Yi}WBVBY5Wg
zPZK+6B~G?7>A?>10)ye%0WU&MT)EyIa>U~KiK=4_5!3qRM#y@=-Gn3i7Uv~2WoCe}
z_NWwl%g6S+As~&sjYocI0Wul*Dkx0~B7G~~`ujub-?f)1Q3d^PY+xoQFFr=Hd5#*|
zd2I8p^0W$X+0<Sw-+<lcw<q1Ww$(0t_TFg*vNUn+THiou1}v2eeue(KJO7KG(mNpc
zs{t@#e)YreXU?yF_{n~dk;zDKgE_~y7D6pw&7A-AWS`@O&tISJ5`B!*SwUaFf%?$h
z>DsYWbLzR?zhIy0|6@$nJyt)e)EJBZAQ}3CLHf^%46yzee`ot6R4VCDVYw@OKppbB
ztqj>Or;3Y{4g#c88}WyqIPBCp$M~G<tA_yQVK{PT4TzLr-vJrotw;nRZ<g*6899zP
zO6E{YZCKp7zmoE>XB!2NSG;fJ^c&^fW1}$ZU2kUJ>%QX+IPQ6uYf0T!g$su=1a0m`
zqZYSC3YM=WJhBX*hh@KdbV(cX!9C+7kvbak#*?|;JL$2<!vPO&_|PHv>y+moyQ|?2
zM$aoGOrw$nH*(hjIU!^y99~$845J<Gh8H@jP9FWb6mt5x>f~cR@!Z1%$&uv~OWQ6n
zqS#llr-sK-b+W7WoBd2<h{KqW!l#uF*bBy**s@+;FPq2O)HjYFy~zAD#3Crbuyx#8
z+QSJORDF>&s>{^ZRyufG`$(sdI={_!iehfr??L}j?zar9@H@-&pZU&G(<qt~l8<2o
zB@43bbgLVWpUP)TUV#C#lZis19hfswHeWTleP8oVnb<$sK<o7ZAmW{^2r*x+JPMc?
znn@64VS0`BD1%h_x~@*c3>ISV0+$@WWe_p<Hlwh-cU%U&*PNWtXl{ee{|)#4x+?n{
zs3kPdyB?OW<FcD)-uup_vrQH9o;aM0a$KVbj91V`oR~)NDVGtYy6l}IW>WAqc`tpq
zb$qS4HIJIY;j=@x*DiTg3%GD=38jr=nP+h+T4(`-lEmwq-4<`XCa?8pIeP7_x_5N(
z#ao9p{$OOu=NjTli*S=kf9;`on&{=lv9Nla$)?~%u2JC(!j#edwIcqH;fVdfe9%pc
zeVO}8d12ZLh1w44GXsc10~=Kr{uY?GF!e0m-@yY9zIyb(^b~no%;ER?;RvAG<a*sC
zq=P=0`chZ{C;QnqW)!-xWGMqgw>HrLYUPV0FtO^MqA0QIxM~$3lv4wQat;_ESCt-g
zUE#!khq7Mx%Wnj)GXSxl`@%qK;&&mkUnlw1C%?wYzhv$hnhnD)TD%(hYTAJKJa?+b
z>x5G?YyFG07((gniv$0caMJZ<G?9!KzrFpm?(!c4J^#v}(SLHT|7*$5!$agZ-hscp
z{gm3_UlXnR?>zQ<Zot1sqwzl(0Y5d`{%Q!4Q50}r=NW$UVgFGfv;WC<|F=Zc&KF`M
za&pCix1VO3?EP6B!cTnfm(Bma#K%v~!k?Vczij^h=-Tkh=KoyHmw&}t@yq7_gsuKp
ztqZ?w{x6&V&pRi7il+X#jPx%*pXv;q{Ht2*Kj{_xRS)EspZ~}5WdBgfvHyO;$1gws
zcXcBFM>z1mVmp7M4A{TYvG`^4|F}c>$LeqVviaZdDSr9+|H|q3PoC?Spa09x|K;cZ
zx7S?yW%Ga8{C|F${gbx*AC3JmvynWSW0@)7>MppVE`jRZVno`%`4W+3?=2g;6IH$!
zXJdRThCvfGr5(Jy(7Nc9LtG(vgei+S;QABBS9oC^1+%X&<j$FAJds`)x^sG24Jdqk
zY4UJGCcJI~oq@HB^t-#T`GlHs`io?NYlkJKlWm*BJTYMu((KjhX<(N^{zy=*JGmCc
zPU`{0IVk$X<Z<0*eAGa#QpOrVDEpP!T-m@VZ>b+QuYAA4vzU*!bj7Nk#cP#jtzP*+
zD4EgEUa~E(A;YQm_|}Eij!X3lZTP5EWZ-V+TkJS57Qi2)i~y9DF^Y0GmaGn7PiAfV
zkSl$FN(U!8#9wT|N{DPSu@b_^UrmzDLlB#H8{mFU`JOo~xS}Vip>NwmH8<>%?4N|!
ze|+M<d|JzT7n9;b(rrmXG0+Z3M)?Z>4B~oBU$v$mqKG`BT3LgMD=_509Ru4wS(ain
zTq!*tA$X7zEL3?yX<1jn%hv*%kzZME$!J)8tI`~hRgsiMiSuuj_Br_ZAV?++$+(us
zuKE1j<YYj+Qx16`CEt9mlEG3Frr2J2{<^Vt;)m8a=Sa#;(J$a1ZVoF3s0Z_^zQFk+
z`uur3d(7T<7by#SI``u(!b~fY2=gORk*8(a!6_mq*xIEI=v@{#P}8GDehl{^tr6g;
zCfx{sS)yE|n4%@ZejLo~>AyK9-kcV@ZWF0{Jyjt)4Rs+^xcR^hV^=E8sBefAEm!W-
z>Dazv;9K1ZAk+_O`}K!XaVt5!v{U5Baq!_;f5Cnk-H5Is`%xYz_TuyJ-@4p6)qFb9
zn5F&H$2|v!H9>T#tz;xArw*X%RbR&)2qDp1Sb8i*<_e|c+kFEKCQCu&Gzk&&pJkK4
zn_rIgWWdHGc|5Je<p^@k3gJ+_UiHvgkvV07eqUw8vP6-iv}x#Ek$>Cg4iSMH!L6m^
z52Z@?Ky5q|HbaS$jR+Q~3K4w+Lyjf!XcKE_FS9+<G>gk?Fx`{q%M;;ex^7#0p&YhF
zTWKdTM-yk?pe-fu)AmbtV0vuIWu8JyH(}av51KNx-3rU%Elypa9RpOLcDOd6yBqcZ
zD*{l`33mSmqJ{dI(5U7}QUqm}nyP%3+}@ZWMOG`OsFbBOaX4GVvd<=uiuY~6+8gVn
z#C9O2+wu^gxP@z2^rfMSx&sw}tfapLDWn!fRB4jqKdNY(hbNa1WM7U`4e<IT-pS2)
z)8+V6EcM+FuD$dS4S0Xyqaq`$ZM&C!1>iqK`7gA>YS1sAX1CMJmhEmP41DhL>&LAi
zzI+2QjhK<!>Q;ba&&Od&`kyveTwb@|apG<ZI}#GlLU(Q~=EX)K^vB83&L0q6ev524
z#;eerA=t%ZK#oNZkz+NxDEeM7=Q~l3AwsvFDc;$X>sT$Jv?$a%tSAdnH2O<<Vq?B|
zh7shV6vat$SNiUjl6}Uj#BG-LPo*xgNAu>|K-T1xVISY|Wk>X%Pfe6Dti+-k4lK>A
z<Vy+<S)MRY3tB-?QZ<2^QXc-WT9p4#>zRRYwDffn$7AEwe%1`ld)@ltW{?+JNu@_W
zj=qU8_1OE)4Z&%*!{4aH@L`}<zV9|krvk1tin_CT3DOyHECuq3y@`Xa<h*bhYyuJf
z4aDJpah7(D{1qooD;O2+sXbw5LYgIU05X~t<`;aVWe3l`6G?%bL^AuDoXaAo)ybDH
zT_rZpeBfu@=w!jGk1^f#p*pM?R+0AYxV4*AiRClCzIo%ISU)KyH}m$#&*>{YseshQ
zN@;l;ab#=v)m8)yEsma9TtO=6A|$cv$R9?JXbQ>XaVd7+f)!7<4C=&uQe{r~3R7sJ
zP10K}@lSOOZx1JZGBUa<wiKct6&D8;^lt$lFDg=&@kBOsPFNZeuBV2DPF7U*yXta`
zEN~ket(HJdM^&cn2gIZHY3>cIb3ZbgG9eVdMmwxNrs2j(Y3Jbw;fLpL=(xCU@d)9V
z_f)^2@2TOr!GM*c-2m#*x5Fz(NA}Vd@uov7j_3S_2iD4{I&XG2Gs$;ek65&Tn=FkU
z6r!;uw*4j=lG!S5&%Jd0%hU!TN>PDT9g@xRvL&SBCPZ;m%!Ve#Fl(?O*UB{Vl6vAk
z>^%OH>=2J~L}SdDNrQKCq@n<EWegoYL3WwAEDpcjo1;ksQM9`rY;Y}0WK4kX#S9|5
zjVwdg8`|EtpWtmmP_C0qVoSl1D1HonuU(J5^5xj^vLI=0%Ytk64;bXV7P!Y8?J_S~
zGxQx`T`$<~)<+^~d(PSSxYgh*H_D=C$F_$#d?YDHE(}c4sA;{_--x5%P1_%BYAqjm
z4@nZzTkS0P#n)d3&`47s%Ui}2@?R&}<Pkcq4~raU8`k!^TWmJM-bPKy(rb(zNw40<
z2YdtBF<{8qL{2iK**o(a2(f^=+vhq0B`9@<%ZNsrU1Z)2ecmOGdHN_W;P~F0M>QI_
z8+~lb=#7Nf5w``$<jM=qrM&9z29_g|@#yRw+$SnP3i<w8XB34A;&LtAVc}w~y1Wi1
z$(8%p>~ZPEW#utiX7!=Cw#^S?`FlC<n>rste`el8vFM#9gEJjy{A3KF^Qf<9T1)8X
zf%<N8#xhygLBh<O%XKsE31oZE^N+1{CoauF%_s`iz#>_vrYwS+dt6q(TZg+pnm=W2
zJ7jz<jsv)*HeudNo%DrgB^=c$t<-rDq+3yqA4A@5`I&xMfC@bfR`OGF-2wkF*#E**
z4!?Q=lzol?_-J1`-vH#6d|lcFb&5baAgC)fkKLLFoPwzTC^3C-=^Kd3xPzwdlmeV6
zR%E>b5-72)oc|5vUkD~QIqLsbK>F{Wsl!gn(l6f(Q(2Sr4YZgX$?Bq0oWFElMd(PA
z{#JUG2?7vD`zSv^tB_n}YqZ0!a+Bn!!N<RWG{qOndytd701wV0dHXP^M3q(VXNy1m
z#rOVLTRhL<52f~Q?P+k}z@av$V$ySTVelD9r;}i^@siw5+2bpF?%lNQP@G(61AnCe
zJeaF+Q9y{y8PI-Hy@Fp=v`20psRn9Qm0^g<OH;eoJh5ZoB_CprQIDISn=h~R>lQz>
z-6n^WUp&Oxo!Bn;CGTsaXf9*J%7d9pS*dQPw5d4=Z5}GUZR@#0FQX;=twJ)&+b^~e
zyUq?2_%@OJ1`^cY1Ot*KUm!qH@Hz`DB@X($ABm~O5Om3B1!<=uo8(J;S$o9XGzwq7
zt&LkBT7Uc2%Td_WsO+Y1vlu_UfDQ8rzjI^wAqC?;iCRu8fT`a~oC&50>cC5x4Db+0
zdq`Y>-TuDv_zb<c17rG=49`y(#@2AQZ)1(Scb8rXX{#CN=FBD_-8Pg%cKIJ8XT^aM
z=|xbQM&_bFFjZcZB?3BQ70B(=wH(w}SePUaMTd%lHY?M0Nhperd5|oNQtS!O6jjd@
z<vDR%9vq>ypAdMUQ<reS@9X?~Ld|i2ZBNukFnfX<{gulwJ+2CcQgw0@I=uqD)%t(}
z0#u{0%q8cmqWaQaWiNZ{(W?v@gg(nTV5Q|iN<MZp)Yfb%pYak!P8RGW{BS|B#pGmK
zAkKt#cnZ+ni=Y5QI*TUb`LbGdR|QIjHOnuBCeS}}6*p0XD5=>Q_vK`qwh1wM6g|kZ
zKY3qJAZ%1|e~wV{xCe<BkP4@X+oSKi=*R^;uyEy59WM8s>%Yd=T6otq?A<<N+OT?p
zKk-rSbgA%VyG$R-^JE&XPQqU8!ay3L5_pD=>RtTSBuVQ2QZ$9F9T5Cu{r(Y~G}a!W
zW(uMR$#p0;rRka%WR{8XvO)p_^?;xA!Y$5#l=8OK^nr^ac@6KyaaEn6NTOaka|`VZ
zQ7GLd$4=HgLpBKlzbuVnQ^}ooXL)o}G}r9llc5yQ1gstNwlJ!%ZBBW9$NPpp#lkXt
z+IV{IHFk48c`Pe7r?!Ov-C?kYl0|=GcAow20i!$6gUp++$RT6UFsW9ONuwODJh0jh
zmyQingycfo6>&B>p!>UmYv8l*mPWWn9^IPAf3S&qgfGnj7Mxv4%IRK=XKD29c^tg4
zd{lxQFyZMMiIi~^JajDnda76UN#<?aQpWpL^f{6;>sv_Hks6~~591Lw(dL{IpI!X6
zC@|t<+d4~>tG~OfzZiGA-Pp00$+hQ;$+jH9v=c)M@>@}ubggSJO^U`_98TAEio>-6
z;<c~q;tkGAs+UgCrIrAK|C`(WzWPpMLp8ARU~lFyMNLAAB428h@@YuLE9}jw>e!Ok
zvmSfpl8wjGpEVowk0OcS?K4|1HFJw>slaxE5>-3xr~AXyRR7UL_@)6CuI{Gdx!iX9
zH-#~wwZ|(5nnQzVEaB?mYz*5l?;ZW{V+btvxy8AW61Gh@1!W_NSV{sVm1pbT*cNnp
zU^*r2@BI%)eGNwN_d{LzyQeDZ906+hn_XCb_QkJL{D&F))fay<>A(8o{~vu(Pbw=X
zKo5|uwoG>&6TLR!s0vzVqMV}}v}WeH-{Sg`WyeE24|4L^0(28tuhm<waBlH)RLc=@
zc-<aMQeSqjj8=y+L`#WZrM6Ef?o`zs(Wj-M#zux>zHjNK*#^k`1R(0bqnW1vMPjld
z@DQ$_0J!FGNAcGd>=AZC*W-mjdARu_oF{Ha5Ifn;$f>&%hz3)NXuiF?M9wCdo&9~k
zbcNf&n@B`STX<&0qflXfSqdZH5zkDL%T<zk9daKM)&%G^C1(hc?#Ydw@8?@aFb}H0
zHw{?`p2l7emg>qeDnc`ND#|Y%F}r2;xE(#y*;Z>@C86I1>7yhdUzKklau=Qx(T(B^
z&A8BV+gHP#o;xx<W>(spmnX`P=ZTP7q+Y}rs~Hps*nIu}+I#PyCfn{^lvfl~L<9r`
zq&Jl&9i@o~2na}*8WE(2SSSGkQ4x@?AfO;1y#}O4ASBX7dX0n<5_(Ul0Yco*_nm!a
zzW3W_mviQv-<;X=`-5S^K=N=u&%N$zt?OFX(&J@}IfUq70Vu%dp?RUIf$VKSn6@GE
z{<e33>UZ3Z7(Y?>w2{|9?&s-3q<;l1_ln;AO}LklclZp&<&sOe$Q#0PiBjW4s%-Pd
zb3n_zf}6WyWj8*~Xcu3uj<jF=qJDna_iD-q^%W7`o~pxA+!{~<`4CJkfT&<mzkXe>
z08XEDD&aujg&F8Ng~goA7zeW_C17o_)VqFd-bC-o7s#Cnbcr_@?$zjXa93GJ=CLaz
zL#Qotg6*a{7BCrK<-dlc$XzU%gY!-o#Zq}gO4jsgY_ap+(X+~$IwGPuUniH)55yHD
z_?^REYPn}%Ns9o25+H0B@K&xJ?QVwV>B3lum>8Z)sw6yeMtmIQ|6C?V<*nlJ>zo#1
zMl-!-Q<GFjg;Wab?3hII`AD4QQ}eP+bfA_Y41VlZ=1(dzjItV6RC9Fcko1P1q=kxo
z+5X|THo?Q0B!1y9SBDM3H-DL5kz_6~?wCBD!o*`{fm4*-0a4;-eDZ6HikbBa5p$zt
zv<C-<Q*V`e<C?oB6cIAx_LEKVaWqT0oIr3X)72Zvug<B?+WeZKPqrDXgPivACT6pH
zDfEz;>_{0pCEkH5H#IX(t=_$oCGjHZb_O(K;MoWps^O~5U!_#rj1J71UGN1;;g<Cn
zBJ46WQ{Bj2Y(5DU{MLFuwweUwC<15gK&hsgfyu9UzXZlU-+)t7dbSN|kVUXb%4+FL
zptAD0YbO?&(#$nUqKxRFz9-BK?`-VUwc2&_eKdUr&WFNw+22~s)ANrr%M9&&?RV53
z%&?4PHXm<EnjLt4H7T#4Hva5c%bP)WKkBJ<a(v=!12cNN%&}aUVz~f=p{j<c+959G
zt1~Ji%zbgVy7nIFJc^sXb%<LAgMh}BXLOo|Uj^^wgvjEU+WFN<@nza2(CS-Ca<j5r
zP8oo!oN0mfVjIV2<okPA1y<1zy!Y16spm-X3hRWj0aBt8Mu$hv|EY;UZ}RaZsOa;a
zt|fY0O0l6l2i1(MO+A9t*mbFw#y^+ogv;DC?VU86^yqG_T4sjjRc2bWvdm#vCaKzE
zyG`6E*{^G|WzZ-qRD*p;bfvFLgSfwo*cS!S8wupg_vyJWi{s`MLtow}>m`#$%tz}o
z4g4~BiJo<dHlVAHUX899yAv_7j=Izm86kzsaY2*s0#h&bmRn+_GnytUPp<3IWKAkd
z@0}#T0WXzp6MvfNp5}4|dzS#it}H)vu0>O}F?GAGU6ZXbHH4{$OV?8*RX*AN7dNk;
ziQM~}KZ=E-ZXMbaiaI{m$;Jy@WWPt2ZeK%qaRuP|QuF4z_Z)3JzBPW*56uq}#+rP1
z8et;nDwuO8pR;|20jmxsV4`Se#}!#vp~>xGH^^FT(k_-sd^YCUX0~;_&~@Ax@y*?z
z2^V_5>+-wlo%HQkhsXpDfk?^J-0|l9s?x#*6(h|2bloxP#5c5A7ZxQEuXro~WE!Jk
zm!N<L8NbFTp>`N=P8%GTshq1}xs900UC8^qQM7#lAu5;Te754-^-kK-z9TuZfY|r4
zV}r9P$6J#lQccb+T_C0sX3W>z&31?<-~3+xPI=){GF#K9UrBcl!&Q39^Cpjun>5k-
zx09Y(Y@LLj{Gqd0zJ2zmFxnNck&Tx$r__~pR6=*!FTqSqSyuzP31>##z2WP-@=SO3
zdF@m;(*L@UyVY}d_$I&Sbk+C7S8#T9Zz8bZw+Af8cOSuwMI5s^mNBF>FP?<2bdh(v
z#XpWe;Zt*+F|%e9PT}Rq-(h&s_-rw>k^-U$Ag3IfXCM;5V7Vhtd{5C3=q&9r#bw`<
z;{H6>^9b<^8rC1uY#stQwhKD5sToVE#}t>wCU;-`t9#1^m@g|rI(S7jo|C=2lH>qd
z)4bVibfWXH>4U4AWt`~EPOCetdJVW+T5_Uauw6E(xMuthjVq4Oj1Ifq8F`9vh1c!;
zz@vpJn*SI|Ig!2-?It49@8^7l|49`aL^S#La%0^VRp<}hWu>k26eMU&aaw*(XGCgq
zL#ZcIPf}XkUPErHLVEP)nrN<gmiF}vt2!KlHNM!il)BIqw8Ip}pIEn1X~biqjLd*+
z)?*G8G~36N?)!z9*7m2d=`ZSYPTG6Q2*|9)1n8AdYCETPUVi#1;&e>db+TK^Y_q6W
za3A*fde{)ebMesk4_$MXYou7~n=nsN-Ts>jKzOT6kwt;`jd?qINIRYBMek7bfcKA;
zPhCr;X-Nbs!}|UfI2WB{z;Ai;kazHB3vAqhuEU#=nD(wV)SAyu_A&)y8AWt;()fl9
zZ^cZiu(qea@D3o3VNGnd(>al>-TSGDVpEAXuS(zW&N6eHM%yF;hV}>EYgt2!JIGo5
ziu=TJXp8P;%UA!^ap14M|9%mNs0cY^hGx)O&;6mxPu-hb`41h<Ldc-$$IH@Mu-Qk-
z<KMMnK-Ej(G+8KaseMK;+rqmnk=YSyZS+Vi9^6;HZi{D-pYc#H^ZoJT&^QpV0Tk&3
zvg5~V11G%RPG;n`uJoly^$;+)1ey#YvApX1RMSP2Wm$f{Dr`R|85Ox|Aic>sBO34W
zxbLcST>SA6lH<#vpZOZRLy@-Uzat}KpxkJ`V||>N{c92Q7XCOpQj-~|Ir6B1HoVKU
z&aE^iq%mjClhtj=AVw}}zteY`SVjK&_LZLg%9S!2(@88<WcI|qWu%X&#X(6lB)KLT
z6Z#l!(UC@ca+9jAtgT7+AF^HQ-3IeX>KLDVl~L4S{1KpH%?>jmOGG1NWqdZNbl(h~
z1~E5vUO3*Ah!>tp)@$dw!5bW@C9@gRGq)dK?iqZc(d0w9S7T?D*p%7wrLNO9-!$?L
z1D`;YEu+;xpU(^1IbFRmbylgVdFuptFeTuoBBIGg15^a1U<YaO%RiAKfq7FWeTJ&w
zJ!g8=W9A%ff;Ju7njU)c^;_gz7**!jEv;vI$4CCsxt}eIO&S*arbF57=xbvzjX4kx
zloQ{<<q0oe3A-}tT#syOI1TxvaChYHHXVH}o1XnzpD!qEi09@oRYN8=$9v}bSX~#&
zD3r-C@(d~tr<E4C_<1o#blbHZ{P~vK!;BrL5c1^yn=wBP27oxabPVE<Y4&&d+M9}S
zc467~Q#pE@O7ln{brP56J@$<nIO0SL@k(#Y%5;On3}?KjjgU?A9+lz?ClHES_u=X>
zFlK>O%LRUh`d~~w1YWY5CHMx!?J*wA|5dsoSNNKAzD<0t{Jp+Kw!D|OU^_q5u@z1f
zz5e4C1&Wh_1P=@s;)xKhKFn3%O1-bzxz7K+GsGqxDlunO9wsp2Ba+w=c9YCEo>%g_
zAZ6Fi$1XPGdasR4r+DOVeO{l3Ki&iu&4`X6?DszoZuo7n80CA|Ek*Ac`+!2Tbw@bI
z)!LO`oUIs(XeSFn^B+gBPJG&a{Yt#>p+g04NtjC0i9M!RKjX7fBW9)?vp;rlMUxA7
zn&=4`iHs^ZrBLh1Hcj3}QaNXE+#pINz~&4JuP$Y1`|gQ+fgX$DFy-a(%?X%7<Dl>e
zO%S@d)7E(EXS#c2rP8C0w`XS~^tW>i-0v}-t?3qYxTwleM9w49%4P-LHo{|fSuQ%Q
zB5l%e4)i7s5rXO0HY;+X-+t(m5)+*iy2Avkmfz&nVeNs5oaD$V8M_&?v>#d5@*j5u
zk4G1+RZrUMyP#)?3%NDF?u*KNy8{ne`1s^4H`Bb;M0dBA+R@AwIV-U#U@R%gax@k2
zDPQQ}gC(J3jd%ps-ATv#Cz*RJ;Q(Iq^yX(rg+#|4#Z_s{%l8PqH{vb#DqUW>5%BzV
zX4|b_0D0$(KT9#h{}>=(E<=mjj=I8k8+JpOy>Ld?E1bq{?W}D%Lu`eoxF;`u$fJ9n
zzN^JV^&f&+6fc|U7Otk+xxaIIw5#E^BE;F4C#?P{))gVVnQ%)Kq{RLrr|a=!GP3It
zN@A`XCFDF?$pLM$K@JngC1|a0Mmsoax}_f13_QPtZ{$2=u&q{(Y<uFA3OTCwCdF?p
zpd-|pkV_Px5YKJDR9}=$;o4^>X}$6{J-up;`$HGwFNr_5W*6D)t(+I=yYcd6@Q3`F
zu-{xSXzQ#<&+c0i^^wWP9261#LD=>Dwo>Tp4U`ITLLG4GcDUnY)aB-K_0wDWiw@MU
z3@bY}Pm5vx^~naq{-zfwIvxC~BmR8K8jilZMjj%OlJ2KV+#A!+xc|5nBfl-fVCm~{
z$k?8UYqR$*>m(JX1*5rd>qq)eGQW_#pJo59a!)t%J~i!q_wmmcG3fixdW1yX<Q*&)
z&5SHJmNzkvWT}jI9fFe4_(gzgKw_w8H`V2iMVGVAq7qmxLh+q60aAW&jrTbMpMqrq
z-o(~6R?#_?^@Wg|Se!8UZE~k-rC$Ya3ovt++?h<g<xJYFDp<2S(`M(i%#%0~LF^SW
z_@E;?aBe24(+V|>Q4xFgd=Y<9%gcL!9d?V9P!Ku`4zVPg6y)kxc-c9<<|wh3w05&<
z(0n?`?+}+@t`ztwLQ7oxaz=Gcu%uS9o%f1Bl4Fc#{rn#~1l*biK5!qjv`L~mQYN(`
zG%tEOsgZm?%}9k?HVvnaIkUbNtDEkX+iI_Fdt>e(V;|)wr5q`eLD4_H>*ST*n=TN&
zi}27mh}lE1n_{2V*C%v^Hh1*sx_^9s-|_J)+D^~4j<fPyKTFL5Jy*PGoPX$)c1E6u
zMrbBdGl1IQ{^-G~WFJ+QjMXKZ<mKpcW|Z1`W(ftk4DMy*zqMP9d$;{WI;l3cQFZ<U
zY}EKy6yNJ1GNi)@bb<1ESIwB+nk-(%oVcGkR`LFuZ(Y6A&!k)Fk6Na>1+DJ(!s-v9
zCBb$C_Yk-vMnl6-Em*R1FA-PS(r8kW@7WD5^HNDaDTv>Sa;d4;IcetTIm3W`St;kL
zm}@HAbN33hNbynFNTcp1;w3x9s6#oJxuRJja*3?eS698V5{wj|u!-*N;BhP!eH3$Q
zR?g!4`2{d;8oXmH8M+~B^Ta7;tA{)mb`F!QkrcB%ev+g3lq~rVuRQN4ZR1BC%tXu$
zmi3`x?$aRmKXmp>OZ~#K3f<cW)TXthag682T{s8S#gw%4+P?`Qv(u@<&xszBJ(GU7
z@04S7{C5>^ajT8+O;?=DMF!A$am_shmd1EknQU5OdT$bN5eAa#*yC&${BorVsjCaj
z)xi@PaW&Yt<Y@sVV!@cjw&}iZMX?#b=`=Vc-*Ko*G<Gr9*HU$zg{6cxh^2`Gijr$o
zrCjeucmz-vVCB^&GX0d+>?p&<N_NX#+Hr!mK=b_FW29Tc%(WNtCyg+BC;se1%17d~
z1^D36shC3~5#PQB=A9c`ic$hB5b>lEMOB!}(_%Z_{dj)-Mj)HRbbe9jNo~QmD&5vg
z3_)x5T1;pSuNn_D!!S%cZ*@kH%<^twG(PhrB44>%DWU*(54lrbttVnDl+OL_)#mAS
zhmUvjwgC3Z4n#$7eVT6oNp4fE{?N&GnRKCsoQ!h9h8k_&Rg_y}+nj_FO~*OsXsjF7
zL^ansVE*ZhIl$myW+IYQTveFPH<k!~=mR*g|NHx?qI?qIZWy;~Al0MNCveA^*rt(}
z%Qqwe$bgg8c>@+x*UEv0cQJ+L>k=8^7j5$~Pe%$=V+i90BJTLQ{f<jX77O(vb7_>O
z{m3DdNSW98R?ZqOD=kOV+#9zLkJ54-O1;M>n<bhe{zT;IhqrP+%^q>mD_MKPS8-v%
z!OO;2)3wU^kzMbGRX>|AR5fxyNhRfgSdkqRAS31H=rlfQ5~mm59lbkg``eYyUfysc
zRh0UBnXd8+ngixfRyv5Jaj&scjyr>~@6>eyizY#yXmQfV;D9%+fo59C!HR4CG9{>X
zPNMPp<ul2Od0UE2+d%1L65g_I+5G#~E{L>*^RxLhMGYIm8r3w`4(-L?4Z&XmQQXG+
zPrrVex;^9@yS;Pc^h<X^YqzV_zUR+L?j-)e;4HfA3aeboq1$kdwy(00K6?{F%IVjT
zjSc8555b7jQX_{_D?E1^*tTTczy%(OsC5j27Zj}X(ZK>zbhu3m7VtH50tPt^@#shT
zfdiIoNVI0>(&gcmm6th+0oyaxGgC|<p6orJbNxM45=4J)Pl~8AGR`fvCvq*<!!P*w
zF-x@HRvL(@`@X(+PLW*D6XJZOK0P4(MA$mKPh8XXUP=9}XOnRuQAIOp0Kd#qjK~X3
zhQGj2RbEalSAL<gAM4!nKj-hg0kXo7qNqZ4;+3BKIi5bq11_7uz=tw~Q>8(}A73k9
zXN~MUHJ`rE!U(%pPj*9;F7J2-N4~CDdETbcpH><qZWRfpS?1&<-Jf{Waf37;&U=cH
zyNDAb&@#DuE(B3DjE^zhJMOQJ*@<VG^G}ObfBHOkg~m-5>r6XCJ+gpiO3i^)Hr|IX
zw%>o>n`7&E0m0VMY#1f_$_r|=|8@xFi11V%T25cg%wI#?@VI@lds8~~+fd%mR=%X#
zl23CVu>A=dt8Xiy#f?Z6W55(~<a<3hfQGr)oM@hlem;)zF$x)oygl%)mW6@x2DgX0
z9z+B)9>18usVZHWwcI(&^z!@J^hZ5qK2o3MxMs6z9-ZV)w$cP{_g9gQJ6S!gl`rAF
z?eYWaF>WIqxwZng8t~6p?WMrh&u-d?eSB>lh}^^O8EX#Q{o(*#g16!q6QAUw0_`Y&
z=n9~w?RFins>v$R5PZAx5p#WpJJ{>)R8Jv-D0;dU-gxbt`{(M4jijnLgyaCurLF3>
zE#-{|%E09Uk7jeR<8roJom70G0h^<)O0Q{FoZ;65Q)%R_Cwi@1MnJ5683sz<RN@j(
zT0{2lAFprB3{-9G+I*Ox4dWR1;qPSp!7T0kS2OO{)u8RAO~yr@hn|^==P3K}!SGQ2
zSLDsas<unqHcB<K0&&rEkSks^E)*&49ZIZXB|!NoZoH9b_kOT{JO8OD^84p2i;Vre
z34bjKDyIhkvob(wR)hVk*j*|Zf?l284Cdq5&U$@p2E|&F`(ybA-qdwVXRUQ88-HE7
zIiY^=rga;4O>E{s>|tl(4fKkOHEe_V)6Uxaut7WVDwli*O5pu*m^m&omh0UG9@l45
zsrBYdJxOMYRdH3ebh=XQoz&+>>m|Ez68s*%d*3ZVTDeIXy#yKXT`D$n;EF{9N}yKw
zc`!{#G!VBHwVCwG$SmLf5XmL!FJ;#15+9SCEPvI%6Dkq2xTWP8FR44^-;)Vx4U-tG
z0}JhvWnKp|m(BF>ZTvvir^?<S5dAz=uAeL}{Qgm_dS;?<#BK(!p}031QEKcn;GBy>
zTAzGX>|<8C1%atmw0Y;WoTdpiFnO)yd={~N5XkE;HX-64c&z&IjNtrYgWRzxXvImB
z9QL||B4CG37->(P;Ri5nwQhz%h-jk+>M(Ztl|1Wj1)qmW_jpem)EaK~9F58Ll?5)V
zdBU^zR<KA5d~v1h7+T$a`S!XpE-hT$#wDkxR5uTx*`)=IWb2HsMJjFg;yteoSAYF@
z$3-oi@miquWq7SycHxwE;iF|Bb16iG7JGlN$KfU)MeFZ};@WQe%D3@!7Tyt&36w9s
z@Zc@O#}S7fFq}z+qIjajq=Bix%nyKbQD_z%)4g#izu4@HD?9@+_78`ikpBvQkZ>kx
z{FlSw3q$c#xl~EPvjv|%;S;Xo61)1uJr$ua+1xd!+u0VwxmXUr%SZwDZ)w7k(agPr
zakrT3F4so6IIAj}P8p-K3joIbWT1wtau0b}*&L`}Wah;56)?M<X*d?3Kss+fc;W4&
zJgVr;Tas@ingp6rgnN9J#y&wsvl|Gv%Qd{e#;B^E1alePKrxdwciKRgGdc!xq2|Zt
z_HZ`O=Z~P_2IvG8kIx&r<>w!(YOM*Kmj5g5!rjr6?q4iWJV{1w@wri%(bmYGW>GFG
zWdAj}vxdO(${jSKWZx%Y5w!a$`{77=(&nOK);{if+{0S)vIc{zHx7MX=w1F|DZog*
zgT$Sz8O#ktI|)VdDu38OdCwB^?DSM*w>Vu2?0lb(Kt4gG0=~F1+jeWFfnGFx0V#?o
zjqb-t{s?1)Jpv*E%J~^uw9(TK!sQuH1fA2{C%1#whh2Cl>j$|rekJ|#L7TY0cRx?J
zogD#C8g<5jt)Pqkj*@e)0m{D5Y0bpg<_C0GoV82w9rA65r&rp|vm)-vAq`9mgq?IY
zaPdZnYz8=tS+R51JQVW_$VdQn<>;|_%&ev`^>H`>5KPo+oRYn;HgkQvF2lR2$w=d(
z+6zPbEBBgaf^zQ>{Y75VJ<|*1uwH^m0vp@Dy-FuUGXvql!#5M5^={oC@#JhC5>@?5
z&zFzmsi$SvlK8z3g|ja*x|CEJF?;{Kx4jiM1&vv4x<S4X$9Fxjg$Kv*dsy|0g2QfD
zBIHuXsaS-jG#cC*#MV<QPC3vQ?`GKhWh4BDu2zGZeOX;qNrT;yq#thI7Tk{vIHGc9
z_~s&&!F;Wso~GV!6|GL@J%sGUcfY@RG;v7d!qAeirJ<kY4iLUgK^w>!^xm|<9Wy5(
zGSl3AJs?Ygd5a}{E`#cHaMIX-IbwNm?>3)KC?QJo#pxugLq0<9iwMY!jAh*1?^?BJ
zSw%<9pbS{`Was8@dMQgb=t1d1FusV&K~WspVV24cp>e{GUsMIeTr2s%w05ZAor6%5
z`Mha>Bj_?^t>N#lm+2ubTH9xUnj07X$VS?fkj?4obhD<haaab8J%iZj%vLz`&2WEX
z9pl<YXM)w^zIxS_?N;SkN=BuvXJgS@DXKlu9>YjgZEV8wnvWssg8bg9l=x|OpGH=x
zL(a?B_0WkrZt#>0OlwN#QkKj}m^Uy*TtJa8hQ*g^K}4d))pdzY0~g!3<8^7BGD=sz
zCRs`$<jx=8i_3fIO4sf`27u9mN{ATNWI-u644$9896F&~<qRpcTdAIz>?oer7aiY+
z+=wo`|0|mN(wFKIzixW^;H6@KK|&XlL*I9kMmmzbB}*vNC`sZnN=2-qGI$($z$7YN
z)LJI5z29)bb|D8SZhsU#Y%OEfio&qNz@;lQQfV!?#l&*FE#=1k!tS&}M|xC@OlNs%
z<V3!Gtf-LzGN=?BcxpaQ|6Mozkml+ix+s+ufeyc>y9V16RJ*Y2p}3Wqfr_XJ0FpW9
z;6lsHn~@_Mu%CpTih@VUu!Itx%_j_t0ao-SQgrh}g{$vZ2YaN>J&V0Me5Z90cmGQD
zEj=MuE2;eOx=yP|ZiCZ26l9kI>^52AN2s+UJrNR(vMIF1LwyfyzN5%u^Mj8t`*Fga
zPRr=WX+1K-_J=vn+jGw^-($GFpQW0x&CI6`o(na-NAXsPR=48lN1+9?S{@9po>4sZ
zR63o)e+^Zs{*~>T3!ALSd9QJ>2C(KALH(Q4)U3DS1;&<+c2o;^VKZvH==WIcoGdpW
z2xp<H1Xg-u;m*Sw0q4XXWm!vY7QZWw>EG=CRa(j<_n1x;+pEb#o1vtX!)^}w53oZE
zC-qPxR12w>k5<!o!2Jd8XZA`xM|#Qf_w9fBJ)3<qS#7|k>WT6gp_-q(qg`Ef62^?m
zDDjq1<up2`3#j`q`b6+Jz0EDA9MLGZdD5|ibIN^#TcTL&G`xhs(6Z+hgzQOhkml0Z
zZQmTuvLf6}YmKN1tJF{AO?9@2$7fltzINozvU(**x})TVz7yM#{F_qVA`rWmM}EP=
zU6{!dzA#=P8ec01tBO*-u)xpm{5gJ34-{sQECLQ*T3=$GUQC+{_1dx=^9SFZ>$p9Z
zE|XVcs(48UQtXf~r^7CH|D!Co?&()wlQ`~{mFV_?F$6%CiIhzbvx!w*K4`}5R<=zC
z>^TMO-<Bi-mf`~cU_-P7DDnUsQtJV*A=3i?U_(+@{~I=BhxL4X%y3ZnEjV+S)RY&9
zNAs#oIPun|NnNqyq&Fu|cA|`u<HOg^{_<V${Gd}<<M&4A$@JIFBTl`4L|?iF1U~|%
z^1a68Y#_;Kp#p6QZxU-!#zYOZ7TVP@r46UK^>phrZ|RNc&QAD^D}zn-V`maL41S#c
z^6bW5U3*|E*^~(qV|g+DjeNc<e`p!eh=tu_M(`vH1(zO3r0SN#46}cv_o8o*78a=3
zn~$asr9^GPU*sPt`q~QL+rAh5eIwb|1{cO;=Lj_-jyDnB?IGN)hFwV&iRD4T{dv-J
z`u<Yb55I;!UpX}Fxd3lU^zpJ;uv)RsYlM?D$4+dT@@OdrFFK5{J8wX0J)6|fO5-u1
zv7FRLYy{3*`L84TYTwJbzKFZ~v((mr(L1i}f_&0hC13Pi(wY~Zl_u<Qn*l+rWY;M$
zE`iPuT+FC&v{ml`mri#1L_W%NWn!y3%Mr=gb6|9RfD%Z|e5p+uXajO{M`&6nPv25B
z`l=TWUGWXCG`~G%FP5w$mRZ~MHO8OqXIcHttKW9$v%D2g`r(o^Hu$kOm02U><+H>9
z^peXuPLcpR1+yfxLwi7g<h}w&dA&%IsVcFXn1JrDb<lhZ=qH(6WO2ew-ZytXT@iYL
zXnId)?w3QZYp#ry)*?Iv$UP@=e?9FCwo+oG(*GD?E>hZA?}qP<wDF6aRr{_yE7SLT
zdr(EjCT1}-<G2BH$aqkx_LQSfj04!c+iJjwT~PQoW|G5vz%ue^^9ZN--BxP;QGq8S
zKl;i?8wQo9Ou~dnTP8YfbDit-ivkl-U5#BaXdzt)gviHkKFogCeXQIE=2A>aAgT?>
z3-R{)N)e5?>2}5t2PL<NTGL5_<iZjJG@Q7fL%@<{nw<+AVE3y>vNTjI(<dVGpdS7E
z%gZf!dha1>XN3}Gb(|x<Yd+-%Dkw>bYj9JIGYo|VGToR)bv)dZ!Z>=ARpg$#g0zdw
z2|N=KQEsf!F*X`|^<n4N8nv2B9CuD9nb2`$=TO@8$>Ig&zdh5+0{!y_`07?mZ?YPQ
z%JsjJqq`~B!~4`gKw06vE^MP%KZFSR-A3F`4uzL`RTj#|C^p$gR~ChI3)OvZH1Z1y
z4@$aJm7CO_M$g<Np&9tK%ezo#PKyU>(GlEEUYWT;%B;v1%fhJd+dMP|=eNjx3X1SO
z=#rFKiF^B)<HifaHx?O*Mtb=D9Rtu74N;MlD3b{h+le(IH)$^BqGh6p?(qwhj^_oA
z-nYe*i3Q}(^-;zsazM_0ckPFlbsUPS>oelf8+y9^>kQ(0AI`tWc#L+<yeV!31eWkh
z{DRv~T9e8}{33SwU;VUS+Nn?@Q9w@Vd>W8+8wGFd;tv@Cogr3?cK3TcFHOe@Ah^H1
zfv<=)2OL4D;RyIz(jPjZJDmUD+i9q}0rHbMAPxNHzeJ0#E61=xwK`;p4gUBGlm^9t
zTAaS;^MW;Hc`wde@tLU}E(QdT0<~Wa7p&eC+DN94LJs*AF~^Kf@a0Ln<z@QC*nu~+
zYb4Wl)5!X;vtur`>Emtffx(dtr5@FwM;RGCO}FOtV{YYchc_7x6@R=WLQ_}vQ#U*2
z_v9BGq^gR2A5t2)w4xFmzn<oFleBJE7cr0A3Bu~GQ-9XeNc7wWm%m;SYw<Z@(dA>?
zpn+*k7q^b6C;q`(C4_x=O(sCJ?PfFd%nn>yrty#kqMV!SkMD(XNBH%8@|_mV()u>p
z&G17sM*MEItIuI?#HY_x<Kx@Ylu}yA8qm(t7Z~!~`Rl}h|MQ}*MZ=?jTr5KJ?8GH*
zBOhGpDgQ(CMu;(op|O%o34ZBBqn>TZ7H(H`m=D_^VtyYyAWI`Yj+hcoKsHt5{_!T0
zG`33SLPXJ(I^JaU_uXD2LT)<Wq{@V^mZ-LIlpSrYw_@?<cR^<!O;Hqo+eYISdu&Kb
zh|V<8>MHncCRNw)pC8UWj^q38g6V~4HLu1b*7S@V(VG3ID%=+O53pqXjyk4Y;LOue
zG~)`0WgVrX7-;7Z44J4yn6N_!q`wUI@%FyCC&T@f_m#h@%Nx~F9@8I;)pexxz}vfl
zrUjI?hh*0RA4Ac_Whb|(>mS#d_6M3%2hs+8=djCVFJAbXmNikex<lI$B0mA#`wBZh
zvD4KG`h9SZVm!wzVfHP&RcoU0dfGah9YuMZ*|V+eMr`K28pRB&#J;P~njS;k4#PoY
zt#DLODQyTx%OQZ>@NLTMP*dZYG_%oKiBX}?5z~jR5pRcX^u8WKy9oZ(IWHJ<5OV{B
zN=p4lmf^i8U=TD(@}x!e&)hh9@3UtlaM3e3lokd5x_l(c!{dooCXui%@lRx9GyD}g
zuBu2~je4IKUaj@<OpB{lwefUT3ZYpZ-+2j<mGRYsM-2NppEGHG#iq>va@7Y_)}U;Z
zwRvUyFYUwoki2W9=Fee0A^4>VOf0RO8Hp*HhHUGL5VYvvt+K?*t7}@2V;idI(OZ=n
z{p6bU8OJZ2lPbPyJM+H%3@R>HT@T&g!rky3RE=T3VwBR{iDs8Y{ZKAxVQji9D$8S2
z>TwXy*tk|*MbD@9;Y7t65#L3jSB!&-mWt`fB#>x1hOVg&gHTfMYR{(M*rc6Z$-Y?2
zY;-KlC!{!3gm<z(sr0rry)eY>Vw2c=VNhn4``pwsU3X6{I{shSwyLSMp=r&uIcA-X
zq?D>_$r(988P5tWzR6B%?%QjEw3`w^xh3_JkP1xNq3m~=(V*M#AKu$OhjRqk^^jbc
zI5{%8__sDeV14>>&;B`S7_T=&EnzH0w?Ef5-MF-CF+tVwn@SC=r7lb<tfvvw0pV#Q
zexD`fH+Jp6Sg}atqF74@vxqs)>qqcEiZd^}|A$U_(fTxlPA%)ErFk&!<jnK|eup}6
zeoV{l*5;lxmsvI0b3_&;>+bGNFs)vRW=5yFbm1~L2vB0$#trp)t4CuZe?7c@n0Sk{
zG&>NBGAmjO;1?Rl8j(E7{RQPf1g(EZ!uLNJ%WpsmZ!Avk2fD_X2u=NhCNrQv<!0hl
zR0fv(whnS^Uu++RovR!r2cBdrF^q~1cg5rgu(*=x$dVpf6Yd+?UdInna!eRC$VA`I
z9pxb2L&?P_`N{jC4N(t<2-W1vtkWxK1vJ^q$iMLOlMlXG%iTUh(I3ftbH&MlGgeRX
zpL$PRD>3|q$C}b~LpzkGHxDp#07Tm@Iye$ggIzyK@RJ4FVu!cNH>iR*>%=7{u6h!<
z{C4l_u@AcA|6JU8-vrgn|6_6eSSBXhd8D^q`M8q%n!CbLEXARf`ood2^&2^yBa>as
z9YbHDwufPoqCgsU^GP6!UoOI+EGR7d6A}XiJwEPkTME;9Hb)(VC$tGZx$CVVvaV^O
z9K=4=vj~~)->RMS9g6?l-B0g75VeM&8j`TF{JX*)M*%rS8SEDcs*T=m)0n^><ofT<
zk<1glTWiZD>rEhXGjKQ6ni~A!xklUf$#4eR04^;GqAmhpi)v~6R?Q0^=7E_^{GT2-
zy!#}IjNPzdaI<YUd$8Ih99so$w^sRov@OiN(+}1#S;r^qAnep8h7UYNR{cZg9&g%m
zz@tOGO@2>GpQMCg+f>|+4UKgd>+9Wb8neplRZUf9sPwN%xkG#PulCHDPW+*>By$|Z
zBK+`Ng(-wY9%0#N#lu>R=IRqcBTxSES;JU*g9{$<jo0;TZd^{^55AEBo<5|>Y%PRk
zoEn<e@T0OBk@Wdy@lghVK9LPNX&FN@7I?Sy6=nZ4>Tw;bZHpJXS_1Wuw<e-<-6bwt
zo|L%0Mm$kzp|Oxa@6?0xR<kaWM#mM}AO#d(BgwVYniMglt*>bFXXUMUrj2ML7rF{>
zd~uF)dME<G4jTuI6AQ~w|FreBvI$DlgC>6Y2$WJGv~;H<pwV<+Pv!#FZ2DYo>QaB(
zwUByONuEm#<1_c2V+Gc|3#Os`rWaO^h^wjbgiwjLAL#T5SP#=@Uy<(~)6<T&^A88V
z36{Qpv^lrjcwJiL7YcLcDFtpL{mZyP1#`F)7M!gKh3^~J55r{3S6U(GJdo+BIR#D%
z6s?wT9^<8{HqWQt%9U42#Vwpq4f<>&SjzHArIfPsLx7awg=TI(hW`c{#DT7*`XEhr
zoYs)9G!?bU3!j?{*bG~g>(_VGWjjcdoP_sKKQK)j%OI(K?)@KVosEvz*Ym)9@d}v|
zj~T9Em^#x+u}$wcpu<`%GJbk&EK8S}>D2j>GYOk&px5|Q2sBF%!jTDn&fdu@%uq3m
zh4h7oxm47-D#a$%_DY&)TFS@AV5>u9a2<z3|Hi?IW2@6K3R5PE5lxgNCAMm@LKWJv
z(JpiMu|JSJ1&bAa^!Y-K@>2jeFCE}2xer&?e}AL3`qZ5r!#5W};c)k9qc&s~Th%Dc
zmJ>k`U{LEOOiL@ao$0;TB8CEaeOHk07mjhanz-y3>nq%x8xt$R@00;gn;!Y5VIR6h
z**e(Ugi7J7QGxt_kuZ^BWQ8lWAO^3nOWtp*+Dv4Hk#SRhE%HTZ4rFF42c8fTf<Vy%
zRS9|Iy93tY4rGZ4|66s&nrix(G4-nu-`WYzX=i*PDlOkm&+y!NaKOwk-=Q;d)$qn)
zwu%afB25FQ4M;0Z0`*UkpV_kvTLSP_qm$-DN~tu}U>{WVhi*VxbY;e5O92h*qs(n?
zXazTHxa?cd#=ngrFobK{XuKz}%n?6vIB`;Q+qj7zqG{6W5`Q?*J=}a016qKjBvF&j
zcDJRat>f^LfLhrQ9z=k1YBF^x0ira$13Sbn{&%o(N?k?tlVm4t&FmntQ5&<epQrQH
zV>++8zWaG5_Ag5<cZ#|V5frPbKrVVG(TJ+bzCYp@Yt)l?VTuwY1}$hz)#k_?scX%R
zp6mH}-_32ZbvBeyS4kq^B<!GxhLPl+c(Gk6+hB)3G6X6%qh!o!+Q2~=8|7V=a9aEO
znctIBYe>%e>xG{Rv`ePio(p%&ercg1qhASe-I!%f5&Q0Fv@bloGz=518Xy?EJQ{*&
zFh)r<k>TOQalzCpM)JlNWaY1jF^lXpkG`Ro1fLGo(A6#f7!n3Z!{)PRKnmCa8{i;Q
zH=0Xqn_SbHSZr$z<BZgAtU?cdt@q|r)|xKA9iQWg9yl-MRa)HtE2fyk^gQFaV%mTm
zCYJrtD9nPm2G(gm`q+58ZZRXD*GN^`@0bvDRc^9&TXX59pu4C2R@^nF+GBqXSU+_d
z0?*1`IPx(Zz9H!f+aZhWYYhSXB#Rwc{UPNZ8a1eyPZJ{fSJZ^C>+*Co_6L2DQp?jl
z<C^evO63f<;I2a0K$<l9A*E23)xNC&-m2|Q<QR_p1RG12mfPLA^63GWvz|WvuU{$7
zNms!8b~ju$C)0|kEZa`6_L`5%VD*|lWP$bI8By}^8um6e%2&OTwS{;^Npg!RHZxdO
zhAb2IQ;DnL@G-BQm}PW8j@i!R%}sRy5(|yhl*bcLt?X~%(B038+_Zsx4M0&<CoA00
zxpi~y*LMAqfaW^&dq?o}<)no~;Ym;jFwf_7%*QVo5a6uQ+Hvo_owlx9Fpy6%aLH~>
z@2twD1hZT;4Fu{gj2-3LQa4jw)2}$W-ZbRVUkWKX&0vi?yEQrMR6=Ohd4ML#Mz6yZ
zp7Te83CDX4gSlC}u^!wwZlOr7H>%<mQon`lq%_rY=_&z0bc{7c;PD|e$WfK-Odxy=
z0qG9y`b-I#oyJDe7%4A*C_71xN<ntjjP6VC9iP7RbKAaJzNRAy!TS+MW{=c+`tzp<
zOGh@{9^~E7+JWYEvc}t^j-^O{mJ5#@Ja4O~zQ^f6+5403&n#MhuDN`Vd6~|l(?q}4
z>|RFli;yPTjRDf&^W<1#KjiEthP?x(O^PB)#AymlQQRHK*vPA0GTG0Irafuw-^068
zhgW$c6hv0dgGy)ao0qzXJXdS7>I)$ZbjWf;L5`&TCNg82+;6XOd0Tay#M@1<%o8?e
z$CUZe0*8+>c?&<Kc(r6$D$R5bDzZVNW(RC$!;}09&yRuD0$nWXdIg264ad@U@=|-0
za~|)}D8;I=P!&x~AAj<iNishw$pmZZbM^57=p2>t58ZhTAv;oCWPfWzYu!<kR{#k3
zHknm-w<a3um-ytv>c-7<r_(nj#j1X;h0)P|Q$7<axk54Bw9b(_X07!wlP$FkW1q=`
zP2PZ$jKYh7dsmkF*Z<HxEn_65g~M)@u9+5lHgdhKV>vhVZUa_?O4yJdDzs=u<>pS!
z!lp2wwCoqTX-066_dil?cuv0iwc3&2WT4#Q-lx+60t>LhtXN`(P+Eb6GVLNyca!Z6
z-?vYKp7xFerCkt;T>?onjJ)ggipQ=mR8V>>BdGAX5D1&+drF8oX|QHw*Yulu>cN{c
zLHX|oWZX_&w7r`ndMP?QJ0Ld-F|tF2z)!B9%<E4fWsOllLL>^CAYguv_vnxgVij9d
z!u=-v6E99f17?ofG#TrLm+oaMb8ouZaxcB7%P&=w%X;w{XM?9|72#I`D#1ZPkigR+
zY$HGGJ?vCd70pG(^)7a_)z4mXP6=2rQZC?d=mTm<m3vRB4P@{LO!t|~X683@|L*nk
zPr~5uguwrVAb{BX!qi&Vox~D0E(THE-3iSaGJ8d7CZp<RAq40aenY}Vzr>zCI`6Q4
z^83@|-X}H(L06IOvay=N$nPW=u8v4~*~T@)REzrdotzdggHm{Q6q~m5HVJ3fI5kyz
zmgRkmes71%t;a<Z|I`oDfIz@hhd^|W@1B^SKn}7!hd^RpS&GMc4TKPuI89Xysh|d`
z+<9P%8umBpW8<^8aeMFh5hs{xYsvJaK1m!7d&#PPk=~SoErGhvHgEJd&pDc7&Bnr*
z?HoO7fcd2|vPw$!bBgtcYFPwrq`T|ksE^wZx=5_7{6^6<u!KLgnuGE!5^E=qF^4L$
z3w@P&+nT$%XkkhTS*H`{9*;717H`V;;7OYF1lp9}^^{RC4@VZV{c9h%h8rVmKKsl2
z+FALlWxUF;bj~Ow*X=ifiCgbz68sX_Sv1gO8n3S8>{8THHYJmJS><{`L9g5In+dw_
z9SlAQe_?O9TlK_KZ`+K9G};N?R|vN->T@`ni8)LrWXC8XfO={fXY<^^C2jI8b^Ve_
ziML#4nE(!~8T=_t{~PQvSMPCJJ{BX*bpJ<CQ9vgU|J?D>VTxo4Am^W0O&^bCiFOUi
zBldC8*rs*NFm$>?uE~bpau+kJA=k=+-WBL|c&fL2j>15c;A<n5knb-Jf+~k5H^E`b
zBbYSPn&nDB@Z-r6lLS*Eb9ESnVA!WU^tK3?48ro6(GT2pr@?oA?Z;fSbG7j8ci~OB
ztMHgkUxT>C>)-+Bt7XXU<l<=_8_&zv1{0>WI#hzgN_-M88wv7Q<k@>Em)kBR&2G3d
z3$lFD<5NA>OIj-;VpmQy_$(7^nvW!Dtng!h^n;_d|D{^6HtAp=*`D~$j`MXXrLukh
z$N`s@09)Sed$HKJU=OyQQFE2Qw@P?oEp+PJkHfH)r;6$d<cqTG9w3&@up7`V8&c<o
zpxIE(1Ec1e@uJ~`k}`$+GA~6hMiTw4>Lq;xBX=Ghf+y8z8>T?f;VnkM|75Ly?Amri
zkTPc57evtly5%@_0;AL+{@%yT6pQxY{>sWJhrS$7L{KSi=3S+IFEnK#b<xU&%{e;N
z$M-(9%bZ)A8zAOQ6J$`@WM1OM50DDmRT^*2NNrP(*a+&WKe5SUA!Uor&FHu|>8@$=
z6}dOd5m3RJzsh2*Bv=*{g0)SLmN6Dsf1tJJN3aP-Gua<u#5+07#1yaXw^ld;-ZVHb
zm`b?^&P#iFV?Bue)DIU4j`3xh)NtK~y~Bu<4C;7;jUX*c+@6MgcDk}?)lH}DlB)bG
zNz=#wG|NX61o0bcV?-_?1;$$A%`(g|No%uP@jlWOf9UFEvI6AAQ1@apkq-hpSGPN&
zbxwC$20GDGnrGpssXi1;J|(-8Ux3tK6I!vd1u68x)iu|(=hS9!@Ul*E9lm~OBlzL9
znJM1z7})z-fe>jB0cj&GCN>1mcET^!1gJ4v>_?C^iyOy3@KAmoI%h2zHZ~5Tgepuv
z^zkZMNCjH&7X*qh@Xm#u{i$y4pBhU5h5w<u1XBnh4hTTQ+q15X=Y}%2Z4rgoKjpc_
z&qs&#f!X5M-_uJSx;gV~^*pBF+Y+xr^>x0JGwUml4J+Gveo~)S)MoL5joL`>v`32s
zQ<`d%o}+-{g#7$<%orsxA38F(6^j;{p@=*fRLYbvsts>DQbv?`@BD5q%%d+@PqSX(
z`_dyLaV(tV_?PmhmA-4W4H651D`4=!c2O8c!b)VzB-?Em69(3jaci8^!K*uy+gj0#
zWWao*F=~Hfv07>l6XM9;7Mxn@@_KrRude`A`q?9&PL6RwX@sKRfzwZ*s<~Gzs|>B)
zw!)=G5^vh5J}EDF#9(cG#_-#j`+~3ZBo`rc(5?;)w<cgW&}uP@aM2~9u(VUdqNCFO
zCl^Bva=tjNUKgd)5A-q1O?j%<zn-!zH;_Z=Y1x5souq^b*?HP^%%|6(V-a2h<@&Kf
zed|3My~*OeLD1B@mh>rq=^1br$xe4P_*PPsG1{G)!=*+1sHVdXt-cLP)}$`XWv-9W
zoU{&S8y@MLQR<8)ZQ?uXBTd<6*pzhk*ZWOBAFxc!^?mQafTsOA24b8v8WOF@3!No%
z%=X%p*Pfo<vTY})x-8Ri8pi4U5*8du-;K5pz=|)VqQwEP5dB`yu9Uh-DsH5yF*pF*
zXe(_Af^WxS>ytoghbntf(J1{Qs8(BgVRPA3XtLq(#?$7e@4p8Yba;SPuCG<CC8~7E
z31^(@zb9<ZTql6ZhOdXd=~>2w57fb2jvqjiAABFCDL<kdcQmF&`S*lgyhq9k_e40@
zwuSm0d3cH@;3TZj;)Ejo9X4&APF3yh0<sR#CT<sWg_8V$h8sfNf^!sRp;^E-9k|F$
z(O=_LY$VUr{g*O@VS}zVIt|~?)#-7+s{jk6&B42R3iCpAE0Fk}H0nULBAnCHZXZ$M
znabr9x*z>4HH{qPi+)gf;pL6_SG@m~_PG6qp&5#71-p%yQX0s?_=l*vnq~}HERK_8
zaMMD)?Ox4?B*EV@ck}A$`o|=$Ff<pY)W&Wm9}hnU7zV+An3mr{{C61y2fcWJN2p|n
z|JK&pO`2&6RiSB*WA?h?@Bbr)e_mC9|0Dx6=I`4(f9UjZSbc}NfBV550Neck`|Izm
z(f>`WG;i3sKXmV<@cWm55#7TL-ofZ6Sk3Fu?6xeyalZ{Ui(`#1w<EA=0uw|ZH+Pqw
z>ly}ImKZ_+WrP;4btnmdYSudfmerf)EO+G`{?N7IPd}n6k(46{j`H@=yC?kpYLY=R
z6L!}39)4}GdT7G<n6VDF0pKcRXZ9{|%yd7}qQq$Z?d|liJ11I%A4jDzhc+}WNW9jv
z0)Usl!O4Kn&jz{ZgaN*7%|a7?rCFMb2K2!ZlXIp=7l0(UT!5-8Kx%>kP)IwQngYb#
zcVAgbiz`#$e|vj<stle7KIjg#iaq{AN4!SAT1r5Lk~-%)4@zEo&cQE3?`OExe2>-n
zj@+(zBBAQBDU@P%DAM;%^pooJD>(xG?;nx(_e=a@iPSMhBo<I*2?({8EZ*T-W`6uX
zS#FM$0yJ^P!L0b6LzU?4U){G5xBrL#UkOqou;Kc(GLh9%$psL_5v>Tg*$mJO0h_#b
zvwvGao&WdN;(z*CbpL@j=6~U0xb%Rh7aNUP7tB5~fH<OoY+Y~-bTp6oX_E-!z{<k(
z2`8pzqi)B#^EiNc^|9&~iq=0*x6J6Z8(|*9@$|Fh1>iwoT)V@^!7%{)L$`y290OK7
zbNT~+=rjreQTYH(@vp%lhl784{V%_EhmZL`{|s27|Gk!E)CNReQ#p>Oma#1Tq01V9
z<AHBHdegPEeU~mBNmlAIJu_w=It;jm@df>%%L9AoRbHh`6M^1x6&&+ovj4tOTEe0@
zTpD!afoR-Kk<!FpsJXufiof=(IJPVQEf)P-G@}c((uI}%+g|^Bu)HH+Wd^zdV5)zK
zjER2_77rQ6EP>4Ve;Bs@^J@O()cCiM{-0~}pKBxhpL^>6#CrdG#+$!`WGB9V3%Yp-
lsySKf-EKqehQ7tI?w3i}Puc+6(uFH8N?9JN4Kn_j_%D@$`o;hN

literal 0
HcmV?d00001

diff --git a/docs/_static/bonsai-predictposeidentities.jpg b/docs/_static/bonsai-predictposeidentities.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..8582fd70727fbfff46555a0c463506627fa05ab4
GIT binary patch
literal 85845
zcmeFZXIN9+w=Wt5ML?v3fPm5k=}iHNh;#t~r9+e=orts`fhb595D*ZMCLJkKBSj*;
ziAaY~LPGBeB?JiJ=H2_8{r>m4XWzZ=zW2lV@K3UyXRWL?%b1yC%-<N}H*-F9z5rm<
z)6&%fP*6|+AmlgTd<LKapuBM5uMhb}MgCm8eDNX`)kRur>Pwfe&|bMhM@vUXf0db$
z{wmW|Iyy#nMkW?kHa4~^3>??kS+6m(va$YE2?Zs&4An)Nix+8F>FMZM|6jk(e*%~;
zUwBWoOi95HxWGg~$wYD93E%?&C@zwF`&Wnm$BW_uxsR8qFVoP{ksqjM1YDq?q`W{y
z`B$&WPY09l1E`oTGT)R@zr<qvoSNU8RrYOS-erLY)opAh!&pH%`&XefwCo($IJty`
zMMQ6j$tx%--MOo*@laDsTSr&V^vP2*a|=r=2S=wD&MvNQKE8hb0g%AgVei5tBBP>X
zl9E5Bq^5mJ&-jvGP*_x4Qd(A1TUXxz|K8Zt-qG3B-P7CGKQcNtJ~25pjYMM>mzGyn
z*VZ?1d;156_#?vc$zO6&04V=etbdd2f02ucEY}4pDoQHqzvQC05I{aCnW!$_l)1#L
zZcP2$n}uKY?Pb;niFws+Gy-xaST_4t!?f&z@@OI4U!whsWdAk6LjPNm{hMHa%Y_77
zp`;-Hc$7>4AmDTtL*2<hP!DM%y~Xd=2j=X}49vj6ry_UfPRwtf1L!Hv0gw5JpU(l?
zz9D~r`{YY_!dWp$YKmy_qZMZgffA@Y8;*v%*g6MqH-XbwWmUfP>&-K3IfSsXAHRMj
zD`6W}I8*{e=K?n<&jHCfXO#&A7lxBJ=YXZvbAWr*jW2KR!AM+`q+H<f-E#nP?Hmw6
zzFERRGKArqjt4XeS%sTT%MB@DFeHtGc|%5ACyJ5#)(Gb-4VKW~w7A~i-Xl#L$|AEt
zD!hn)-(OhAeb^ZFIY0z>Y8!Zw_4464;2)dL|A$sLdqr~@K>x1kf9!HgXbj+=rT^sz
z`@fA^05+kUfuIM~a4kFsI0mN;p-xeCyEFBiO&*-<@jo_41mbnGQl$$@sCMj?SEo+Q
zfAi4*m-1Xe_{npC=DlEvqH{o1&e<5r^Zu1n`k?~Q(eyb0<MQ{ZTVyBj`V>?SYgmoR
zGT|;gm=mM)@ojM7)r?~L@l)^ut5d=~*28s+eWLxJ1p>keyVL<ij?aJeb8NCuucP^q
zVlV48l+k|oy5y{9T(rO&!v^}4b3kS6p|%m8J4XkxTYC=Z9RB-s0^EV%ogzNMmaE+u
z^{m|Xc=7#v=*`zL2{$4F-t}InrH^;GU3^!pUx)hbh@7X1(S14HdN%({;3=<Q{&PU<
z&^drb`ybA2Cem<W>f>tk+Bo^b<>H7LEl=-M$lLGnbM~^IzsOx^Q(s^idhOhfh&4m7
zu~=Ii@Hlk9o9_KmoHVoQh3{+VgW1{{pJhC~(Id<E<HM`Vn5vYyo-))Pw}BuEG)m&(
z`Nzio@!ZPyNmPH*1jHeb(*c=V-SHVmFcsH2C8<^dH#1=ZmSX^8P1gVJL1APu2CVUo
z-Q1NLJiX0qu)NCMP<m&`1JhFbk7_kcbT|j_$2Xn>?Bvb?YYkP=M0FgOfk>ALB&4Gq
zIT=;;r8fTL#i-pN^PQzYCUv>lhq}Td3zW#W2aN{_M0x77`>;Kqb3py@zf*0$1rig^
zHZ4wG)K8y6-dm-QrQ>C$P=v*7L1NPqIwj?8n?O9x^@+iI?_Qq+{^MGfSWPI&lkh*Z
zO?2Z>eb+)%M=@LK>eKHH*&7qxY}`BSJelpajBi#RQU-}0H!UzLFvap$o4k9U;^7$N
zt^J9&?ca6P_kXcM6r1M&>>s{EvLWpoB28-lL+>U(7~lU-Y}h|b|8vy-herH!w)~Au
z{;6aCf^hy5>%eCU(eC8Kek8meeR2w%hMt`RnotO@NvF-X+3D-^F+Bv`bAaHH-Z`KF
zaiXx6dsPEZMq-~B&RSsG%Ex6LMomU>Tyy)@j~Yx|*rf~m-XDAN&)TNmpQ8ix>ugXm
zSe=;vLxDY7<yn|h&kw|<oHZyJghi~J1Ky>op93zR9TVO`^(0L@LC=!=xxmOU&(jxk
zp;9#~a+ipyLj%bfg4kIUHmMhg=0gA(6QVWyA+cM`ff@#<k3A6{Gv74vHr>xqNz=!#
zd_eb&9|q3^hI~d46EMUI*`q2lsH7&I11_$@GQ5Z~&V14Gx-Hh|d#2`p^2b{A_S!QP
zx(=QV1hd>9(cqGCbh?-(rvCH!sDL3KU7veC!J5R3phvld&_o0qM?_38yU}~p7;<&|
zu2&b3El?SgoY)(X^o~~x`x5>|AAWd*-+GcG2OdPUl96Na07)1NjH)-fvSv1+tXRu-
zSZDQaQpFaXWI(@jttizI8Sn^q{G8=x{#2mN{y|a9#9+P!x5|r}+C70i&umJ%hnZSq
zG9PFer4>KOynot%tAY(twRzaQ(f7Lf8~~bJcN)IPYOM2Ill3vhf8n5hj!h=%v6GRG
zC+xs<as&xJIX(x>8Lf=(8?_r9i&~kV)t>`~jX?UYh)wdK`2jN@Q`w1Qmy6%4N|n1P
ztqE3--(IQt{($1rUB>(?d@mGmMd@gUD5TNP&77zjh^a{nA?M4c8D>(K{TJMv9iIxD
zmZ+u?>^|M8OrvttxyGwr^<A`d+3qN-YA*m(cHqf22q*W3?M3VB;NIh$9vI!h5cN9n
zavhL`;I`WhyV9akU{LI)Y+ISASQ&9_57)PjF8`v->qH@W|K>G;{&Mf52;iF#HaF$s
z*j6TBjH*n4U!kjFt8nwTFf+L8IiLm;bgC==y-4CivLJ0UTdSjt_r7-GVK|#<%S`ed
z3WvZ{6+7iI*_2A6VBnsrfYM_e8WzSYINEw)Q&}or^=2t#^vPr^_aH~)N{(id)}O-A
znE<@0^@sU~ar|v!`fo;cIj?AF(@qA6wD+~B@I{qpwsGWOE1%>1lYPEIf(%Bk_XHOe
zg;PBesE-4V@arRP_p!Rnv=npeUNoBi__Pr&B8evo4W0ui5w?etCM!+BfrvfFG_<G6
zYr%wA1u!)P-qFhT0|)BJp>b0*`v3{u=xIk@$?#1bwQ^0vGa{Sm%xEX@n<1sI>>X_r
zy9etoh$GHEk+?SJnQ_^r#qO^?$6FH@^eNGxs~|f8SdmrV%SmC2sn5z$7vBk0XJK{Z
zx`J<MKDswx{RKYTXuhy~4tT59KPnYyeKLOz$U<P*=H}^k347nk@Ys_W=WG^tL=3cn
zO)8SKYD@pSNb#?76bXxb$E<!sC%7Ck=C6k`eI)=bCrEN`0jaN7>e0=>E$%xZ<3%$w
zvW6JU+D8d*s(dj{Lk{Xy&JFQh*tK&&Z+<0g^8GpBMB?jmhBE48OJ!{mwr_Nxn!1Yi
z&nYIH{!IO_p5(^$BGb9dCiCl<5-mK*z*B5MfGI(yVE?-B5_7KSwUrQFf|vC@_jbM)
zSW;dQp=m^yJnXTtoc#LZ(s~t5%O55F1im+XYl){I-suXaP+xj>P@3eW`gQ2=9B?ma
znhv&kat?@8vnTN7uGhMP=p~_2=z%#Un@}0W#`3{peJY<Bg}N{Yn>%zbD+C@qWy<kF
zEzFpQuo40~C$hz`8yn9n>NDkD`lVM8#l5{IE#ul)Hl$wKH=4bQZ?&rKuj+quJ(x0g
z%}njuut!~%EJLj;-uO17#*Z9((Vv|5x{G7uCyl3r$0%|y-Y_3koxzNT{v3;)q91V?
zb?nj@aD*0;rZd}`>dW`k^IPB#Nct`hADEQb+JvV(X_Yxr=LGB!HaqM@GCP|_^x~83
zG+df^X4TlfkL^^vdCMB*m4vyo&GX`F&mfXU1<v&I#6AX!M4ikbCU?Y4R){ceSmFY&
zG<LR{XkR-nabS7)!_448@5<4G9BHkxn7tV{Z^d4ITFwlb0Iz3@Ue}#bYhPkd%LEU<
z8*8%NU^UhR{GW6L^N`J>?#{oONB3cEvPl$a`d5=!r3nalZR4siZ2oxj+h^(C4RPx&
z#dE-ZJMeVD%rMp4`{s4NTZm6p@vs*tt{Jxm&?PU`yEuVPUPWAYK_f0ZJs86sr|ao4
z`2M#YW_-5CwZX69NwS;D&gqDk;E=4I!2x^`wm12WG>}G~R}9mB&NDgB;pc#CrgH%5
zLc+J1bHLE$bHMIf5ZyGPqjR2xpoCR1ZJk_h!X8D{HP-VnZe~m4R`T-igMYDOrumF@
zR$LgXKAQ>cjcC~<*$~e>;c>f_L>_Eos|aLaEgLY#VIDNZ?5cdYlSQ>BxF+8#G&U_E
zO#4h01NvYiulEYwz3s%U#QPgHIaeI4@~(oz`3z<o#P<jS>(1uJJ?%*;I8OWC_@g#3
z+3=Ib^@-#_C8m}L>6rFMZ?46iS`;l5#08seYX9*dO_mxs6<ZxAcC5}RKg@QFukOyw
zr1Vh{dpnj2c;${h2LK50&euwj-q>@%lks}af;vJ%aAAk|fH=ygtMq}VV3HDN#(kUZ
z`GTy~(;lT3Kc$WYW+LTyA^G_(2381MzZQWV->LQv#JErAeuHg`J{<Wm4V@WSPIHy$
z%zM7K_(7A2qF*1qzl`{80Mxf6@%q2Aseyk^BVE`g%!R1wp3(c9>O<vfTgPsE&om32
zF=|aSXdW-XRu(lmwczyZ6YU5y!C6loFY2c&hAK~%(}}+TMDYupwSh#SQUqtrP%Ixk
zF%Z2*bpJ9%RGuK3V#M`h-GgZ@4|OF#Cf_X|1`I;%2SQ)2%zfi-JT8NGCvYf6VqnpX
zc90GW>~4JBC(lWAPo#&RWkvp2&$NK!UU}lgz573gQ*W_9Ot#NG4O@|`(cN_<0~o?)
zJs#otBn=_j<A~fFo_0TO4tX8yhv$YJqxsITqNnr-A~u0UVB3@GOyHU&1mKfKFHXk3
z7z2Cx#YrMj?I%VtI|ti(eLVb@O!BE(4PzD8Y6pbcXnTw9^GUyV=)eJc<e*O9N4)AQ
z7<6#<gs~a*B0EBd@u?&7eiJe#gVo=NPE#niw-NQ`m!m6VDV=wMTnq`kA9D_vNt@jc
zmMq@U?86%kJwdG#s#Jml&{;5bTuRpXd^~R68aY3^J_`KM*MONfi*W1^6>-m()TZX7
zig&^OM&uIue{bw+Lv|&$!Ct=WIlYcE-e7SJ+2<cw>W9+=!rq5!E9}{&S53RRJ2_*>
z@f16V61rkCmgtB~t#B_l$j=7}m!TX(-KVmZ!a1#$-!O~%$30oy$=24ytrVg`l!WJX
z+a_}%JoAvRT~&eQgygQY*6YzLT(M{GZQNwk1xswL{Ao`gzR&kzzED?T0xmOJEjnvS
zvsFpEQ_`>Nn+*DEmF}(T(b&!TTOBRAQW3f90X~-fxIq|gn_w#YONGG^ZTTY6Wfyxh
zZ_bw;zr^dPUg=!!gf^pr1RatH2+BW9)L!IyV&SH&6zx@z*50&zRyN<kHkUCiXLIa7
z`p6+(J-VSPJ#^YJ5q@l!iP?m=U{_<oP~*}W$fzCG{Oy$fS-m$4$ZL|yJp36m)${5P
zYU4lmr*Nx^o9_dQuO0cH&Zd!&yEDIZMGOW5LuB<2)$lcw2RtSx;?~p&kDcA4eoZ`<
z$K(1aqgG;c6CZi^R{9@%wMCl)c;ez<e+vh?4B@bxfU=ZBGPMc8U){JOqqc`S^{ck-
z9tOo?z>7Z`6xAGMdkL@N%X1TQe!p1i{q-97?i|3Rs(}F$bTFf;-ucs0Jqw$S3fyrb
zU0gY_yYkx_ON?{6tP{qumJYvWDm1>{UFX~B<LsG_N!sK-2dsn0n7zqvTy(F0^3)LM
zn{);~ItFbG{oN_;DbI7j>O0uc&#FwA!w`pX{a47OUYI4AS=~O>5LuMDsGq>Pwk=+s
zmr9qJvY+?q9lcQNF%zYhI(K|X`q$5aJnxemssF8#{C_j!PY#}33H!Gz<Nv8sVy{In
zF#PYDFI~&MtkwZO2dD~=BYYS6{x;XeU*6pKpPI-T9DyQ1w9f%h+Uyv>|Gs`oZYy{U
zU+t&GCqlg{_xajg<t-+3cLJ^uONq$B$<}T7DH)5a%MUV~1rME)zk8#gHpF)XGu576
zB<*gv>4{zd*7lsC#)k2bb3pS+M$$J0sVH8K;a&Qyb3n9wEO%>;b7O1FUFF0heabN=
z!sAP*e<|s5stTJsGZ6amOX)+wAHC%EMR7q7PNP3Dsc!`S^%kFV!YQJ(R5bUpV_l5S
zG^cF2w3d-x)mg0TS<g8DQAJXC9RtY!`%)fekA!)X*C}#9E6(oA8_{#XKYYPA)sQm_
zJ0&BPMe~weTHxQ8+IWfLBR&!dK^Wf<y^!#4SHpi-{m+*E9<KkV(W49_MAKRmOgoLc
zFKdpjcIyK7dXE$qP}o|j7H!-NhM@yYP2|UV6^n%BRc9<do~TcLj(_WSIpu*iZ;smK
zrR(3;7@!PL0vQGI4@CpF_<*Enh5e%NQ`rq=a>B<Mm(xWOt)G-X2heV&Ha&W+xf6>C
zs!-bYWP_*O@vUu)>oo%~$U7v%OA9Jjy;5Nzsp5vZO0793CJ3|-2o2h%$3}Le=qrdu
zm>DS=qI@%HxM_H&ZkKUxy0bLUqwPiftsMJb`LDA2sjtXhe9f~OOQKtE>dq>#-3+0u
zANLc*q;_|^Hw$L33vot&d7t(Bdhkf@YJ+<pNg!mP+7-rxaCV19lCIV^V1c2BnKRp3
zqDxUiB@nGXos^^gmB&}njWrZ%^-F4Z&QeH1#Zj$FE+gx_BDjD^wVUo5CNTyK=6|+t
zFNc3(i`d@XUTF%i!0pvhbYIwhD!(Y{l~dHO*6?G-dL|XQlO~lzq-%U_*C$>f6|_FN
z$rn8@Nie|18YF>PZ3uD?THIZ|ByQISZr@&No}!t#mVB}1<sIr|8||MtvAkS_>A$eg
z208E>1rfhb4uj<DeIG;ANGj+Qj0@~7lQr4L6YsQ-IhGyqGALMd-L!M9BC-);v8#(5
zoggL)$R%*h2{;DI?VRU23Ui+pJc0hPj<Vy78|}REs=?7Y`l^u%#hpqybmUe7&ait3
zZI=Tp%%SfqBHo#VD#1!64w!kAPc|&yU)Q|lj2?$;f4HY!u*-^^BwZ*i9_;4YE`w*x
z8(VvZG&jN0<XczwPQGknoZBuRBrv0<H+iWis3SdvevEC`4SeNXfyb-v%;&XgB*IsI
z&fXt{484Q7B``o8apV|?ssa{{j*L)Po*>H33!Z|a6m$;u-Po6NgIrDKM>d&5Sr-Ge
zR<A9GP!ol4@hRBahgjiyEKuu@LNHx>8d8=pvT0JCs+IjPNm_!6U)TSUg-SZtbuX`L
zSm-G}@GO}acn+WwN4Lx?V~Y@ITZRr0LuLouya^UP(Qhf0dbs>Jq1Y~PkV#i=|Cv)D
z=Y>-3qe`WgpzUC>kTPX(+1jDhY4Z#SQ4Z9W@E>hOsY3+_^$WY)#hv0H4uVn#>B2aX
zy5ua<nuDe7-aEH})>|X%U!qu@R#zX2|B2Zp0xO7uIIp+Cm1_Ycx(Q!>#_nB{(UrPv
zB#KhxS0c@mpG!;e;@v=(s%d5#(iP+yjFt%WI#O&EufrJ@7ucb6_8?K+O9O0Ozjg)%
zTGi*|m_Bj}lwevkPgMxb81tA`MM#es;c){OFuFy>DIZDT_Y9f%z5gVeCdu^G5&OKa
zbklY_`-{HE*UFNko81l00Z&<X*Oq<L44c@%fuI-U(3iONGLP3m5T7AI+lruzyU=}d
z)P6*w!wF?25QAr3S*&ce!EL}6mp+=GBA;K13rk&x3Qe|%V{9GTO$k!no_Ut~f!^RR
zvsN}K)7*E=Wp<~VT{#7?3j6d&TSGV)h9fjY7V^?cKi3tB4|i}X*szl#^{d@z%I0K3
za>rIkGYK_nLfX>PBXHeHW~jF$3doTqP6WTwh_2#TN2aDN=p(Ko2dmue_?UD?I*o6C
zi8QP?y{$Lpl;rAw^xpfN5tD}Zc{w~iGcj^zGCDM}lU2p*wLwZtIQ{``n4j;#Zeotw
zcUhoJ*nmzL`!qxwy`2*+@ow09)7-4OJ6`t;Imq9AeZl#ST1q-re-Ykc6oa_pH^f7Z
z@Pw;!pwE=-nubkCCp*&>&KBtp9*H={-A{gbGY~J_VoKJYBb+vxw-Kw<HQ7o5<CKaZ
zasVU!80cr*{Pleb=jBIq@ln4<Z(rrKbP(e6zxPr+5nHvDj7u;g12w-%?BQ-?J(z)T
zWQMy!VjDiZC$lp#zirPr=qX5Y%G9ONYVW_KerL+tRze%ukGhJG3KrVjnZwskx_{Cl
z{pLXFAD_n6FYj8ZwKk|B8*%1s&A`j*n1MKlpWh%>7$aexSNfSkHA#jcnxC`6TTOr7
zT{tnP$BE;k2=v&=pZ;pnfkZis!PWhzvD)#A9aYbLsdhtk!7}D<x2_PEpRBSiv<i|n
z$OmH-`8qNgeb4ETCjeuXg^|#IgcGhU7`ok1&0IUz$5S9Z^YC|pS2UN&6!l#}V;m4~
zMMy8RA-ZI0V@tl*kGBNoL1MLQBTh!-g4{j6e|PE7D~tQ=ee)5f+Wgg|Cgg0g7<MLR
z+sd)(o-#6Odk(;#!q#PWaZ&<EwHqPBAevT|b>)aEjJ4DyJJ;w6>ot5>hjMt~S>Jxz
z%cPZC%xrKZLPa!CI!DnRLH7HiR4<llV+J<YNW2|+Yo=nYdBq#4M-Q&WTw)0PWbA28
z#n?RiEUc;36o=Rv#Tj)X`ML&9!IzBKCK_=Jx;VqE!oiZM*x1-R57M<Bt_URta$DWM
zdX?syFpxK-AIPQ(%Ol$SXc!h578!4W6+_@gBa#YqGK$5nR|M}hKHI;abm`gsPx4Or
zoEHO2i)QTh8c^b80f%t;Q>esiQ#*V%Ohq~M;72}UUt|+6`L-M?jB}3M_1k7XYRl*j
zcWB<W%FdA}z)zfNS^r9*c}lhr?`Z=<{ea;>=G>;tc8-v%xF?F`%oqb^ef_TRsNSm%
z5_KO`CVv^7-i2g#Tfw?*!JXL`aOP183@onfRsneP7^rc+rH#7;<=3wXTLVVIH2gE~
zl60>(T&OAAyGMz)B?RDD+H+V{L8XKQ&Bkxo!rz(B?hZkil7nbfZEyK!7iw+RbIZSp
z&T{;=KL-TxY-*O5pTcsEC9`zXnz_j_;<RH#^8|_<sBfo1SDkRwRS_@5lCfnziP4B2
z_UA_-Y=IR-H1IOkdBI1#k<E1u$O|H;41S6t2pK>$6?XEB4rxxJ&?G0-Cze?$BFO=N
zkqE?^-3S6pD@77cAgHS<oC6-<$)MwnF$|~Bs0n!fpSE@X8Vnl>aVLdui<4K<FMkmP
z$jc~82k1b`6-2nmLG}@?KR&+n&kp@_g#N?%`p-G^pK0{}c_VbaDog@XKkO87ZRDB*
z;>W{y1=_MN8;=2$xff!R+hKe3WM;{?xFb@^fF{6o)-8=oFM9i+xxxm%gXC6<c6_WA
z5OO%I(@R+ySm}<Qt-ucb%wTwTfS`tC2dik+#&^$K55(Oqa@3@yOZooTn({N?-q&*X
zgc_vH5Ii6?9KXu>Q-o~@)*gHgFa#cxV;0r<?#HAY>=r$YU{ud{!c4Ygr)OmI)%S$_
zP{}#qPhu50h`Y%^(ko9O7$Mflp0fZ9FaM7<<o>hlKga6-pShthIz1a)HbR5n-e&@N
z`MiIRc|&Gdq1DD<YdU0J)9xX8@f9*Y2hia2$t$iFnH~u%2Of*nm*if_`OoHWg_|g)
z2#IA|oAfRYMmh&r!1kW<VRPJo2ZHYBfTNR+b3oEps~A8Zh{<jfPIG;pXRI#j%f(-l
zBi6FjYpj=xe6n?yD7rk>V4FLzztYZEP#egiKj(nSSu*V)XRw6Q{<~2eimOHH6bMxs
zhk9a8hiY6E={mg^27PXwe&{TS``wc}R4NG2*>Oy?kIS0k7afqh4xsh{eMDeQq*&eg
zI73uixNfuW%|R8gaaBnAS%KjCiF`>Vujg&ZW>(5|Q9j6+q26bk9E~cPh-^8gEBEd#
zji@#+a%|h;67a<>-&6NNEd-whPMb`lt1hP3iZdj;G0!7DB!-3UEjixUw7bX92BQ&#
zq!}fx7#3N2InElDc2Jy{j{<9kel~*=P|O8?{OI*;<qYd;f=(mhGpjmfYnd!BZ(Q%a
zd*cnoVxeouL{9BE*{xWVknPbKmsF6DGO1T$j2$e(@K@BQ#JsuYd*hADd+%hlVa4l2
zJiKNP=+d@4ElkH)GkH<D!fwwTf%<Zw{hLJDh#UI(L)Q-!Zg+RQMdoGPiejeViaT8b
z-Z8wU;?T*%zikC!2qPP!nT%2Vr}`FbRg4EecCw3}mq}<RBa#vHP@2B0RrgLG_lIQo
zs}B~uK*i0RaiVb<*`b9+gxrF7MAegB>-iq<HIX-2H4h0J<>gPvDfMkDw71F1<IE=K
z^iCUv2#Y{_toQOqKUa@QTV;jT$lZFj6f}^<f-u^SREqL#RNa;{tBD{@80~h6QEL#s
z(vDYo5TT-9VgUbN4s@#!thW|pbV^O;R(>%E{GEsM>4DG7;<{7NU`iLR)=aeiCr{_b
zu@)nRl)-@?1xnfj;T}$g?9cLj^96nC%=4jSfSTysnS*B_=bs;D?78zz&yRMZSV*Go
zA-59jB04O_w(YtlfSpGZX7JSYVykwNXMKIY0BP?o<8f&!<@HK2kesLF&~6ia!zM6f
zc-{rZ4dp9^aJ09IjVUzq(N95=`o(819}ax4aP;#vM^b*9{VDx>_V9qej0hr}Ex<xU
zc*g4)m?6p9HLW$@l1M!Ko2H@jJNq;O{N!NA?U8F)=9EbVS0h^B<a|dvh{F}ZBrpNx
z=8L9|6vqBi<1yfCi`00BJT~cXvsJj0p#XPEF+UP{xz}ynJy!>A1rRcN+UMm)w(1(C
zoN~|-#qa9YE5QrdmS6$vg+H7zT{(x}La;qtit0bxE)=_#>T!X=#0*r@>6zrZrccQr
z3Qe>zJ*^~1sJKsFuZF296CA9ERfSJhqoD!?n<Sw(M#bx9Y61~kwx#wQTRQ_RMx)ej
z&lK<fmVM2mjmZf;2)5hRL*PQfoXRZ2l@yh@J~Itnei8koiG8+F?&l?KVV549B$T5Y
zMN4!%fpwzjtqAJ%i;Jf!7M@TI{J@h)ZO59J8PDE8cH3`#+(w*lg$8s&)|cjS#ce>w
z%~Tj$E~zWQg!cvx6e;!0bt&vj@%NbP7xZWLpL9z$(74c-{$bypwOyao48t<@vCRRS
zp7LYuDYNE^_;Y~N5Xi8&Y8}Si%Hy8wD&sn@hHcH0tP*c47OJfE6tL@kN3HXqai8;*
zEQg?T#&9hm4Nb4z&7sCphck=z!+d=rYn0bf(tb273P^=ef{i0glT;F{a3BsS7mofA
z)umDa)QH`-LcNFHEt`6j)SjSFkvIBi#OpTgmbA@tYpR89>X*MlxCsH6Grq-&@a|0X
zoT&k}&Z@hx6vgvm^$}v^&+l3mwT#6xHO}QEP031jAqLs^afK=J1!;;oq+g&F<e}NY
zgx;_xa1%t}%Sqf05hmm<pe}AOPC+8On%KjFWui1aOE(_R=-*7ta-gmJnYD8F!eXfw
zWnQr>f<=|D!phy;3Knyqa%#(jlrVGm)Z%1r8)F_&5ocloT#x9McJFPKh2ET0eU6ce
z4Ztf4D)m?sUaH<J{I&Oh%ev-h>(7N>w3I6|NrT~cuM>+|iB8zsu2dvET1pj&j#I|Z
z5pSW-)Fwt&R4nZSx1}nD(gh;4n2v|L`0FHu1(*Y2ij)rb0>P6cJ|bezslL8cQzU<z
z%rYYr?)6{<%HyC*DY&qoR=W{9P|YF&sv{#Ia=Xw%WW2)9J<!uWG5x7=p}APO<wt;_
z?lYAMrFW#ekf6@V`CC#YuBld%$)y(S_}2q&#`O&aUKJ8MH(gNS$AkS}-VC<9Q9rH*
zsa3VXqO!$WarnHiF7r29xkd@b&&b|mQ!t~QUaD1jvaNV!P6J~!%O&nuIm-9Rd-3_^
z1Od#p_~LF9No@VZqiJzw6JKLN2*K=7TpUkcpJ;$9nW3YYY`c3;t+Z_WxVsgj19<q;
zlzx!2__ZYP$l#{Evi<;W7)am^(f(81W5>-aLT=E55UIz@-@Ut)(9nf%@tn&M@h5s7
zH~FqK8U!sCM|wb?qUR{J$K(-AZd~z=XVJkI;|sY5ij)!qB##|zJ9#EvY1cRGaevNZ
z!$=bua0hD2-$$gpRG-6g2Nz8-QB@vQb$8^W7M1MR>1t=x@>YtdUMTe6dHm&}R{J5B
zR2$hxo~jD3Er@llCuEirVs>;^BE_BVjJ;E$UOA3t`Kh?q@v@iOkYBl5HpB|Hpd2E#
zg9y57=~Fm@KP`4XDmlYXR()R&QO3?UR@%8V1w{FNJ$o?_fUMiJo+|hr=1M!WXm<O%
zAZ213e;EXZNeEN5EODFQ5{a#ocyX#TgpX^;Db-MIun^?CYaE?$(E>WP7A(HX7mA>Q
z2BUKrA=LV~>)Db&>ah+sO{h84TH@NiocisXx$OAv6Ogois&yFeXsVQe*iPC{1U1pp
zQxX2f1|nwyqN;kgE3b0^V|BCgc+d6Vtw-JVgFoE$bz+rtwou~32ccKR`nFDhtwTFA
zo_0T!qs0+Vk2gsHQzyW-9Hy3AINslw-S!ZX;TB_eut>7}iDVudQd3y2Iz#apH`R4s
zcg4*`^Ns-?y<NlaxYY>HL%KUf*((#a_H$05cPx{feWg;KI6B#Dv6fO_Q}`M3o5X>O
zkHko^%|RkM%<D#kL;F!inO*LMx)nP1%)6GtqtUuI?)M7pTNY6DC5Orlx$;pSmhiPF
zc%x@L@d#Ew7|XnzDpjm5tWGsHc^~=<I^#i+21>i(dFNHks<Dv)kUec<uBBonZOU^8
zam;=Wkk8-PT?*<>SWLJI3FVQr74PJ-D`fvlm1Y$*J>MfX{~-Kf(!-|-pJHq&oO-#_
z?r8=_#v$BxFA;$gl4NFm)on6lyq;h<0<NeZ2!6mGv$=hgY<<JvSBWu|u-VJirTP2E
zcQFSHk#@>hzSue`Ms@V9`jQGiFnha)r_{DXlBge#mP)L3sAHme{|u8PYdQy{)8iF2
zcOmg2A-oWJZ!|Z#LrN{*0S&y2y!W1CqHwus^w3qZvSK*nsX@E!_ZLoDw=eB{|1|v_
z{-p_#zk--eIz}WG-?lHxY_^6i=z>-%&lL96;Ml6D<jxGgE~HVBjW*9*S5sdhD7f;r
zw)l<NQx1<DsUhC0-LuV^oOeU65t<i3-0sY-YCO2ANWU|aJDMF+>#xf><jS?Fbd0Zm
zGi?m<a7f&=5org#^O;&dQcjD?o$$>v<73*i$|}riZT7xGPd|7t`p`La9>#BPJchYX
z_<g5J@GKK5Q%UGSSN*goC$fmVJE)@TRtO|&N_6F<h16IK(OA2n6AL8-X2ttf1vQSG
zcR!o#&ROFSAaTzA`NJk9unK5@Wa#(SNvt*@4sFCpNH)Rlq+o)%O>vjf&jEbMr`dU|
zaI==hqiS(>LmfV89@lIx0-1Kn<<^v)h9<Ko*D`hr3mXPb?<pBaVBEF(Z1kc!^LhnV
zC_lKjw{jC@aN)5~QS9>3yZ9IcLlGI_b`A!5{Bh^55P1{peLYUmIWW}~y0+M+5x08j
zo2ocgyC=Y`RS4UanqEfG?x=x~^M3cgWp={3oc7lznYS(MM=Q5|d_6g?SZ`bSH1M-Z
z-%(a;Fj{p)^`Bb945W;8cecxpZ8tZaD)2<3VpWC3Dtf|jWn?5#qtLx*Jam>`5igz~
z_<2FCTSd2RbImOqfirCjVa2h<zpWr3)Q1mewh@jahxCS*B`cM9zAkreaqdB`Q46=Y
zF917ou1!JJzfBO08Q$L@_59(}ifhpb$_q}1_tCL7#2UcOA(0<w`&7F7FJAy=#q6B}
z>fjx)=n(k?j8`Ptl$xC~M0qq=4xl^#<kpT^Z$<Me=z@KG9k0HRdG?sf{>F7iy|E;d
z7;E6Z`!d<XA9{6*w0ATL9^I`cwrUR_!B~hq#e{4I`Z-9gMgxv@KR8>0-&2pRImtFw
zxa(cR-V6^n!1-!%9@_x%$ysj8E$r`yqt+I~hYW1>D>+`OihdSfT4iSEI>0N(Ah~U0
zva&UY!8tj5_p4Zl95^0g4)achuV^;ex55QP)T_|Fy-g?B(BNROpX2ROm9{jCwg`PW
zqdqH_j7f+f-AU*;=?SfXQVTbLReh+V$x186fp_)5_n5_OrJ8><z(;)niG$PX$~D{%
z^ll`7WZ9`&-vNzcU<8I)(5%VEiT^NSSSqLQOenB2IIZ*KMas(h0Dd4Sn%G$7ZwpVg
z=Zv9`{H4AE-EH(|xz+(*j2Z>HNYO$vG4Y8Ds4JV~q!e;v=<)ZK<`F-OO8R%HRbH)6
zqItLEf^!eWv{NB>7pWH!?Fc5Q4(@5xaL$(E9+_=!CASgr)Y#p}zaf@FJZp(hOe2Mw
zJBV4X4HAY%#H-zMd+qtzjxr7XqdwuSAL5E+;l0C;v226=R)YL&@Ko(R>W|A7;a*pb
zugcM^hxDzJ`KmK}2^FL>RtLz}T|_Esg|yqeW}t*|4m->OK0%BU&GS$e>pD(jk`?vW
z9AgFn_Es;<_@ysY4>;`P=q*E^li6R2Sn#`DsHRfvfv5JPYI0UH7Cu_9;u<pKE&uhJ
zF3krjv1`3QKRnRd4r^!;V+chsLN(Daj-o<`MH&SQ;>A+Pi*`Q74dn~gM)J5^3(rxb
zw{)6KLNh`g-A@=*5$Q1pIn1qB2w(KKU#xbJxH#Bj^B^Xox3`KL-p!Kv%_BC07pN0i
zWjpF_y($R+aEHZ7unASBf&AvkoUfySiu$PO?rpM%y#?DdIMF<w8QulDR$U;#x-vkK
zP<0HL!F3Xuu^~4zt5MM=RHor7{XW{!hf~*D<mkDDxTfrmnw?RRB)nMhn{NuDUO5-e
zq1E}Q74ojySoqh@lm%1swcal15ZRK1k_0k4Pjn}Vx4aeKmfAvokD5k%RajP3_~}or
zgz8df?F(pgUl5q@w|w<OXN|nGkB7tN2X?25?4GcxD!pFb?8>u7X|;9Ym0xB0ev}r8
zcIGfma7`^zbOi<3p6wc?HE&ilJXaAu4yq@topBkB8rTlKP*cWE;ebQ|=T7CvICdqN
zV{n61{i4^_u?l<<JUCjckR#GHM5%$H*H?#eLW2F~WT~54M}C)5BP*V3UfI)UKkLoY
z5&8t6xYX(&xd2PZ!;N<*26->iom~QR<*;f7IR)pn6=-udbLiM+PaiTUNDIhR8n=ft
z^DA=Xk&{Vei{c8rUS2#3RVj0AWlu04s~2A^pIuKoG2ssXlzuSDw~igUopj6ZWwmJF
zM@LTy<^CMS!Q!K71wup1_G-Srp{~30(bpfX%j+;9m<A!}^dUq8qZF}4G<K+5^UvmL
z9<2bISpTqNYks?ujwbs+{o1FUOcDFmzL}$d;=#s3q8Guh(x9NL7;@Tv@~)|+>eA=b
zrop<nAbv)ko3yLHoO*k060_Zx=P2<ZxaHNip1#4RZMzikKdK^M2El;=&irIIyk2nf
z(S=_>dzP4P%!Rf!Y|o|eCS!KP>~6L;6N>$c*Sv$jTH_KT6|8BJER0v`>y%wH47%Rh
zM$o+%fA*OtcZU->>@(_4x;|`vnhBFK?ibq@nb`xCZVi%cvoWC;qS*q<t%~XG{S#jd
zXa!2{9&*Fa=3kR>>QE$*m%PTf8_N(Jo{oS$^P0!Xy_o8C;#M<VWwoqC1?B~#!y@Nx
z^}e?I=g2LSNBlZbj*y8lV(BQ>M4wqr>NeoIvKwxWn5)D82-hc=#{0aQsVz|$j^g-0
zd|SZqK9o|GNQ0Yi{_2hd;cbvtSkxDTEEM`AE0Fwt`X$1(iR^c|9xLwBH?n=~o_$q-
zzBJ2P_@l5mL$~TjL&fr*c&aKfZDbuZFppTu`Q0;&l<G>Lf+QPFfa?-CAuzIg7{Ft)
zGVmnSWlJ&CddTS&|4%a!e`(f}rzkqYeutXUDU2x4%0J>q*1zMpAFFuV19RYukACq6
z@BNjk+p4lupJFGwVN7*A-W5a&k<h|wLSPxe1Z9wcA4Jsccro9keegVG+;dU7Vd7Kw
zbmoreoL&9`FcO&%p(f>~T;}R+;UWvs&>qQm;ZiWFvmS)l-HUmpC|JF)=*11d-X9sT
zg?$~^*L6KxuFSyX?2I*${nwvl{!L#tV(vBA0}+X+2ZceM3JJBH9Hg7r>@YMBw?(ox
ze7qQU`--z~fWcm9*@tLFuc*LyM~&v?b-R=j=tFE#@>v{ymt}1NHe(H|RU0nR<eP!Z
ziXC%;=bZ|p=o|YIsc}~bim_N!w;dIZJrUE5P3ehFGn{NOz4>R$Yny#dvjh~_F!>?9
zD4xGvUaWsa{9@KOS=ubVaMYz1ZFCMN^zK-_gxEZttum_fh^6DoaFeT12}s==?Dl!N
zp=af$<3+7*&NiEp>Am6^Q1ia$Yw?+2x|GJH)Jgw?=2~LQ+)hwHbc+ufaq%pT#LUO&
z`kDdxsRDt|+A<?~U+Xp4QesLR9eZlT34M$uZ_llUwVPw+=sNg3?Hdp?${!W81F!WO
z$wVynu5XBUk?un8;0AlTT**28>6WKKg>GuR<oqfh?ZB7Y)xF6EH|6L}gh9gUVxjgb
zZ3!Q};a)~jYMcbJLV_de?osPa@^&-OSgs|xWpf8#T{MyHS`oKK*}55vCXGM+P#Q|#
z_5Rlrrzfkwj_)Q-oWd4BOZAh%4b3$x%LpTq_FnU8>uzp~E{?A~gg#j>0ff_v&FlNB
zyX?AI>c%*e*|WI(PK805IMBYj%V4$Q3yaUb-b_fu057^_#HZzLa)lJKXBE%bxfxe<
zm)ZTMOLckfS*~}Md+X`kkx$}h1~#lw`w%Db?l4glOycZV>~19Z6+!g&^^;Ee97cCM
zi5>Fhm&}k8Bgg$bc`Qmh!0~T0w$w3dbcBc#aCDkk2p`;f4%9WzHFj`(GlOx+-T5qY
zrKb8uNkU&^d&9^}|AbG@`Ke;X5<8WXB*PTaX6xW<o2cfsQ>oTIl4n&4F|E;^Bvp%@
zukgH@2g_f_(>wVEdn@LZO0*33P^VgHMyD|_M+)oGl_V@J>>hg~&`_xS9(i#?KPQ;l
z+r=qlo_fvD`TNAa3c%fJJ61d55g$Uhf+Ki^2yY>~x$8M#jfDgjm*ag*!kO5P%2Rm4
z>hq&tCQy{G=W$f;Z>f0FrA_5D;_ZYj1U)imTPpky#6f=#dHKk;YNpn!@p-j&I=HXh
zcXjzrADQ|zQFF!&!taF`c|AP#Lt!V#*<&y8h}3w{i~k}HO~;^Od8x5i^T5wB<RuB#
zkv1tKkG0xB7O>Y0aIN~3^)A%Bzn=KBttul#jVMFN$IP+xka&loz^{Q^__XU}6y)n7
zW924GC%X8&Oh2$`y^vXezYaigE%xeVq2~m0fF2drB5?ZfK8d#h=hTWQ5Ev}gESaXr
zPb?~snX&B>40TfZGjKWPu2LO2Fc{*-GX|anZ4OL=4X!LSE?<Nzv|7Wm=RpDE%8>~z
zG6iIZ`ze_;agZrt`j~dY-<)GuTcOvX{*jBdPv~E`x_Gi4wdFp@&LWw0ZA>a;Aodhh
z+Z?i?D|2BBtPGl49q1?K%>?v??ZG=4-Xa(kczn83{f97{XVl%P6StOKIFLN`h2P^x
z!~+GIlkoHUpF%%6o7Xc3?>+b$YCT7~MF5)PuGV5(J(cIG?1!ORQdP@=#*bdotL^bg
zFU9rz%8&`?Jg$9D842{Wb+0-k?|+!g`z$m_PObSNvMLcQ`B+5fE;VHF@%L&n5nBFU
z2p%cABhoN(DC*@HqXfvW$!9M68A`LTWZ$xAcUIBDQUDoQ?5N&gmyk@yu>Pt4`n5vu
zvi)d>>%JX#Xn<00tNz1)V~F1e=cQH3g`Bn_MgqH0=%rPS*hjp4*pR%!O|?hgNGlu8
z)qWb)3F(br(qkpjA9U{uDEL*qBZ-W)s`Ld*MR09)toA7K3fmg$Zd)<%-=p6e+WD0Y
z{**==$qUjlGkH1Vo+1{bGzw~W4UuU+d!4EFloKHWwpIxNVtX2maN_Okt=DRCg7ad-
zH?i9*!=7LJO4=KecYe<$vrkR8bu-*is0Wu>&aGE(NZYn4<_JFUZI0J;$&gi|zIv5*
z@$>4ns*m+#B+4YO5=Un<q8)V`Rzt>u(N{CkmIq-_&#K+`6Z`E77lS-JK<WAdZ*R5z
zUILq9`Pxr^2;jZPpaNy<<{5>*^W96i<M_<RpR?%WONA42gEB1y;`!%Y`I-!>ciK~F
zuXSTzS>w-C&jE$s%TqMz*8QuNvZ~GkW_O`-IQ_P4#o^sU@Kr*y`efY?)o(4WqY9#4
z&l+w9FbU+dJ(3a;^d7VxCi<ZfOfG54zf&+g;-(`K@W7MstSyt7uKEDQb+(vF*3efF
zYK46U;w?mUh$L2|^PY51`8xl)Z&<+)CM=)FYYAx+ICB4pU0+b@QrDB%?F(90_P`gf
zUfCmZ;#Ol?q!zbLRz^D~t$r7r$k;0NaGXe|J&9;vo7MF)eH8vIN~aKWb?>g{)XSKB
zB_UR%5q!NeCu!VbsHmB|hjM%<*OxDjkD=wMv%*B|BI3#h{@bTi-7SUKz~L)3$HteO
zUQq<9X%5YUKRV*9l}Pefd>XiH)^4-ODXTSUlQ&<YYssa-qrd#Cj6_tQ`2$Lan`1|8
zdug$CIV{jegnS%*Ig!R2z5UX~j(yC2cz5?UY5H|)Mi<$sMW>+6ox+Ee^t5RjKd`{{
zJIS2dRzXP9W@q*-_=rLRtt~vxLsi?oaB6-%X+w#<+EoTAo0s{!o~1nfSxr)ULgZ=9
z6NzNlWJ}rxs<&BiDRYLL(lCK9y60%mLk+S1IppwB$YxgtPea(NU*Fu4-5wW~7e1X1
z?YDMktnM{rpty!txLS;J{&^zxZi{GILKx`4^K$+WYiJsuD5df@P5u2RS?-;^|L14>
zVe57*Xjo{raJCVx+v}qA3fq1ZBMdD(=77`ONmCe;kxW;pF<f_I^<SHR_U_o5Q%CTw
zM=I@{2s+_CloCTv-8rz34mG`$R!%6Vr|sVxl;n?m+`3;;nYUl4Q@xyLX3GD~I&3fb
zyXol3fgwt9J0!K)x*)4OgH)fk+>gKtEyMB@cB9nTu@yhwpG-#k#_`jUJOvSmn2X;R
z_~ufdBSIO-8v~7=Ox%{&mkB6ASzB183q%?DiFIc#`RPphFv|Y!&keW@SepE0r%Etg
zYzDHpW*kKxWP!Cty;?-Txr68gz^u!Qd&*CeEV=moRJ=Li{EwUSXs6%5P~`SM%9bWr
zR$7yOolw7%53;N`aL`E5O!U~1T1;~7N*S*kV*KtH&C~z0zjA=9_F2}O&s<i!KU&RQ
z5se^2QX0rr({FPR^woX&=(XJs9Cv#Nd-d3^UlprMlV*0Bm^Eq^w87ZMQxIR)n^Y^!
zR4blG@lD4`prkAbs{bbeZ%ZUIsv@1R%PUzvrRM-Q#cHmdP`=ygUjVF0jMvlBUcWn5
z!xO8*Hb4x}=T$b4q0ThYWgKsMP!Oj19#|(qeWGq`;&zJsrwZs2bMHhv`dxL4NuxpK
zuR?5dbv-N;!K|<wsq9y=^49aLg3QeV#~$>AGE`fA``!xS8Wgs$e73*z#6h4_wMs6q
zmx?(!^AJv&4uUOM!7^X`dGB;b_#o)m{S0P(rj$q2!KOu*A6i%Nbty|$xEC~Qmkc?S
zXvkX1hFZRGvKP2<LC516Z9*G@70QX@{!Na68+9Q$qKO|Zg-G3FdCgOW=^9F@gKzYt
zsg6ukxUhUNC>jNE5J4C0P(N}IM@eF;x`HeN*|xbm2P(uD%8ytyid+_q%YP?FUpz>c
zBrE=fpj}~0$j4_BB1`5VCNLT!jtF_{z^z?L38y9$J{E7PG+mG+-KF!BS!Rkgs}W!>
zwo*fm^w4eYZ#SK$4+ZoullXd4H<tsBe5Vc@2&D_(iH@rL^HP7zyCM^FvkWPqA-1Ru
zCGcs>!TO**QJ0VdBd0qd&h4#y!=4HV7Q$P};<`waF+}Y)v-oq)lx(LM^SRTh_Ogdd
zOUZQ?+}m1wutOMdw5rk7ERM)ScU$3^9`@H$k-D~*Q1>$??it=49Ps*`JE?WUPGj%V
z2Q}Ht^xGhEN?a+HTfb(E#J3J(R=qxO>&M8l4n2`W0Gz0L3G*=4h1U)KyOQpvNyu|c
z@h0!&xgV{%$1~>OG{X!Ptyl5uuuKpZHuUu-VYqY#5ng_nLv6d3(NKZW#x4*3=<2qQ
zP8|tK^Z`r$lzjgDGVROAm7Ae8Y+pglAwrPMj*z=pc(|%Kx<UjrcG#@4xsq%#Q8@ZM
zROJQSgi_o6rRAsDc*Qqlz+H7S5?xH*HC;%sCv|iv@4yRP)$S?}MuvmT8XCuZ2c_4O
zXRiG^QhK0xMvPFumL7WhJJs8I-;mo7`y))`dlDZZvlDbd0m#1A<e6RN?uQ-(oziXu
z1r`R$QFRIZkY;=+2QYik@tyPQHW?)?(EPPwC_e3?ms4;#iRAj)S06iSI^jhOsv;<%
zyAc@h^|cJcWS=5U!QrX3*PTtFHloJW50j8Hi2{vNtT&gc1V3Lj0VCOuNUZLV3BD`$
zRkg2{vWEwt2jU(V_`?TFgp#|D#OZj-cCJgu-p<qkeKNmQ3YBTW6-N+N7KUhv3atbI
z^oz;lr{gPiu@A#XxB7+OVj{>0$Tk}-(Iy$W99|>+WP&da=8WQLpI5{#H{_K=P#vw}
zEq@Lh{oDrzROklWXM!J0dyR2FX+Uz3Q@(ZzSGcHGgCrj|7J61rt_S<oPb%z=i-Vtw
zyP$?vwwD_c4C7GlRrc7bE(R99sEPS2xZ?1-IXV->sPp)3N#A=f3=EML3rV6Yx?K9!
z{&8*D{Ay+<G{<kRRi>E`Di*T1J7w0mBH#Pc@%Zarxm7s3@j?652$d!6<@7N2skNzY
z|C;P~PmL}T`LI22{45V`73sBh%`*6Ed9-aNq!io>BK;Sc4Fcq=t~<T4k9m+papQw{
zuWBGx+Zx9LwRFU8|5${?cD3>#>Rov+c$HTGYg;fgnc+slt4mOp2;^0ClG^@zyDSuW
zYaN;cmP}QNLLb4>x6NQ|Ei5GjwbicFZ0SFeEfO9UMwJ?+Kk+hbH>6!VJZzXL`W6+Q
zhwwv;iPxc4ts=vj?V(btn_ox*4$EPoq>$}QpKhP`lTna-w0Z1JTu&5H-!+%VzJ0QX
zT}`lI*rD+)_Y-)6cFgdvpgCiwc<{Vv$_Ms0uVgEp77##-Gy<93Ky2(Ce;g75Co#Tg
zZaCGr-vp@>1N}M$BB8_h^1xQRWj?hXSE<)?io`(jwy#1Uk$P|{t$CIF7gq-rz>8H!
zE%gT|Dq;v05eGyownmzb@MC)-cAwRO7>i0-Q_b}y%2|#>jo&@yf=TTv>aD6SxNM@>
z4`w)3X-`zF2$6<3vl`#1sx2@ZI)Qysj+5=9bn>OCqCeUapK)OeTu1FjQLUoialjZ)
zc?t72_$vrzptIq`EGw|EuVl_zyHML@=25F>?q}V!nOnC${U&d#vu<T1Xka!2TcpZu
zEg+%p#kfsEf^kBp`H5p%qWW3Ny^@rX{yP=J>BPL{PN&nE0mpAvaamP`jb!nXQ2VJf
z);mIuS@k2J;6iYhl=aNyI#R4MEl@JA2svlu;9X=SuP-odEAM&xik9BpxH-D3k?yif
zFLhQbOs2Hu(7@wMWLnRZ)+sroyZj}AtgO9?tE6@%)QPzS>zgt1|6_b0?^ZZ_2|J^j
zY(v7|`GQ%X)`ag;A@mGUufxcToO<~+-(OZNMONEJiWuJvRhDL%e22T7UgEp@QFe<)
zEip2U^TJPgWOZ(sXRgEOP?}?haSw(b?ST~nM^^Agbzb|9I!Qtv>^EN~=C;4NGh||5
zBo=^pfnXzPcH1=b-9r-gyMeTxj>CSjSwQm|Ea<JLUxmE`>!_WZ5`|>q$gO~q$5zwU
zL0UbWy#r3R`60u~U6ncI8(a;sC2m&136WlYlg+X1$m}>)*#~7PCn^6zC&_ZXrbJRj
zw4G+q7J`O((FQTTGMt}5B6Z(WTwngEtnU#gaN8riVNp(ZM#TT>eYS9P&NM;``~FCF
z%Ejl`C^D0|DYaDLq|3w5phaJ`l0RzU#hvz2WXnsLSINoALCt5RqvM{P!<N%(w*gIX
zBYB72NAgJOLw9j1=D3hZ4Bb6&mv;NfQQ88C8Kho3<6857vG?9#O?7R$IH+_G=^dnr
zNRuvIMVg3!bOHoLdWq6|5TrK&0i{LhJ<^qu&=C-kE}hUj2{k~7XTNjKocZ4O%*;3E
z{eHi>X3q8gvG)aJC#<Z!*4odypZmU_icjrkat&FEd&jR8FLm_KNbbi?6s%iGVH#os
zElbU@!D3e@*Oj?GS64UHAS0l)mbLX5U7@a9%0C%+v+tM0=xo1syX!8W37!;aA4dWl
z>lpvtrNNO(sj%Y#x6P&g>a3e-b1D+v`+3|<<eB%ER;E+T&C|ChK5owOg;{(-pV#5Q
z_QF-0v^|g)KiFGGTC1aq9KMs+mUekplD2kP$&11+YM@FSNRA=-&uU6zd)T09l`Tk$
ztLt8~!4mwcK!TzeS2t1O6U=Mlv*+G2<<^>eJ4qndi;48Q7xzP2ORI}Xh!Tsib}BT1
zf<x+=`Rnz;j49omH)YO<ql1kDEXil4Sh&6=kdP(27QLi=vNL-tWN8)dGNfWZAlZT7
zPXT`#vWqnEb6dSS6<4hu?>^MhrkU|1x+Pbg+mJHUzsBCA3r5<LAQe`)B@eOy`Nc{|
znW&R}OnPjrSr9+A^uk}_Ht%?@^7gala{S@^3Swp8K+eAx4wFV~<#3k<nqmfEkzP%p
zAo<~Q@DH!$VJ{`!N5i6G6~nU?PZlgSp0c;wtF**Ol0W=b5NjLKWj}DP2Cj@<h~mu|
z-!q!vMW(Gi8N$hxI=$!=d9FDb+`$?fHB%e?<vDNPtv+pz{BxR%%)>;i2G5XzW^?G2
zLSYN2Ea=&y&|a-Xs$=}(H&M#C{1O{({>Q_bi<z?9Bf0&k7BN*F+STEaM@g$SpR31;
z>YF39$Rs{DCfLF}!{2=oOYfb{fZEM^;>v#!aJYy_HNdtYBWGjS$0Zo24KFw;cFXJs
z+(pzANrBB>4N_hlc9y%rxIY=P#ooskZ)@67{4({bWR_j_BAn3|-ujE+TC>XHJhM*T
zCdSlP1w3X``=fld;W?@Ba|i9_>eChiT7vh>+YE=q+2<;b@`nEAAYXTG&P+!y8@FT&
z7N*`W&z(b1S*H=|@!FER&)hu@*jIg_$Cl`mh`I-(u2N*HjS1gfm1Rby9ruhC%)B8D
z4|42{VwWw{35?oWS>%-D1t2gLz{Uu%hsfjKJrHTYsc4_g?_#h>2QXa+YLIrWK1pfI
z%-8C`BJ^yj-M=u!9^yex402>$C%hdrNyR4EsnhzES71@26C5@F^DHM?6rQ``-w8nt
z7|#KdulkFTrs$Cv_fA$&zT}QWw*><&uf+DF<$%iUa}jL7qv|nN-XE2fB1a<1I=7y;
zjJvrtH_0*w_)^x0hf0X=t9=SA>**8ETcn=cdWgA;zz_XcPK65tCWFaBr`*t6bH*72
z<;j=glEtkzRqff);wBco1!KWf$bGOF=o?7CX8sq!<OMzAX#do{3w%vci)v+Mr#6@?
z>_N>)pxX(83*Y~Caued{9*yv%3-^BW#<UvNB8xjb4S>=H5BYZkd$Q0TwI6$8x{=T7
zR11D$O$<ia^pkd$vhxyFXCKMT3zx?5QGM-gUu^NwzexLy(o~Q=7`e+zBzrnpHp8na
zrX}mNT|?~kCM)&XU2aW+VxGY$6}xQeF4m})Q7OSGRr(i$42TlNYgQUuXLVm9<ct;G
z<b~WYSxR`DyPX`Qvap4nnr4tGC<PB->mVK8enm0CqP}GSos@n@YPz`GQ;b5_&)yXx
zQ<7)Z2+r^O_|Az&E@DcbmwB9JiKve{3&OmPvkkf^!i2R=QFZV*)~7MeT~GhEK63v@
z$JL<p9a<HoF<t$n@l9AwPD;6Nj*|bL#+T$H(Pz!KE6e1~dS5c@vk#k>=aTVRMn^su
zmJw6A*#&Ih`NxmMn6NgT#xg%DcGE^ck8ixYnfBd_)9I4r0Z9TA0ot3U+g(6|84m8(
z2`qr?3c09mts@sMwp3Ta)L7$rRC5m9Sl~U>76>l+q?Qb0(C(k*Zm}7!_>tH7DCt2w
zWw)BKk>3xJD-7i24b4gzx3$(6H^OY`dp*j1VVx%;ew<+e<)`{>#h-<aoptAE2PANv
z;d)oyH5m@Crk~5MpDE4ekNfjs?%B&uiFoeqXl1ld-PJOV`O#+`SHf30%BIUp9ItT<
z89EYCa+z2ATu%mlS)$?bv~<wXi=lIOP@(4~ec3G%dzaErPp*oG5J~4Y!12>~%ZlL+
zw_gPQQ1|rHygqoav~R*zTQGaeg9dbE1(bhL2U6<mVSsE&wNCG+o!)p!x=ZPPshZK7
z)70~doJ-swCC&x2)~SXTNekl)%bVL0tO!Ev8IQQl+ME-OdZU)=B~*FRg|%YTDKkHr
zrgFC=9~|2cxg7@a>gaGD=tgSMR%A_C)p*!=Fu#d-%NBJ*%R;kNa@VFM#IBd6(!lWI
z%|d2p&NWN<_8`d)d6D|EMm7%j&2hAa<=~a+Vc!B;>uQ~*ld!#4p?5Ad$d*A<fDCc&
z55|5ZP6bD0RU}}Gq1GRjR*B5Gn2{MjYo6mBRpcn+e=kJU8fTFxal6NRUS^2HaOJH+
z+`iVifW~iOO8UhM-3!2Np{SJw*gBDRJuR4a#IvI(U|!C)lzARjn(@p@)qd^Ta>816
zeiD9H8k#mfwIE>5^kL5V56k5jLJ0GUjxHcR2Ld~e7f!&6A(r$k=4Srw+?bmLz(JS-
zEpy6L1<)uD{3cYU7!Yk5kI6*h5o`Zsfy%D2BmEKB8Nmo~u7y}-`>;+QR+rxALs_!#
zYnMo>zEh*-kXP{pI5F(21#q;<6g;}t(%{|@cWqOwQMzGtuJ!;%zWgC5+O$tE^h+CC
zO&9dO*Qq^_T7HN%fC74A>d0V8bG{K@uz7<elLcvczvBz!W})({(VH9NbhBOdPjHE;
zPChG}DP6-x1^t}1KGO!@5F{IOcgBMGHjL+~Z#33pgsNy}Kr%!0K^*x?SFAbRNtJGE
zU218cKL%srn3ScRxl|18{=RALkvd{xSs^#T0w&^`a4Gv2febxi?XEE46)r}s*7<$k
zhJ~H%2^{x?w@jGcG+%>`E#>~h4}p0Bf(ij4SLJTBT*k8vE(u&vGQzOJ0JCmUXsx6#
zImQ4o^6|W0KEeJ;ZK?H=jsTmW!h!!zgYCh-lIrpBMbhCOkpm+_5-pk~;#tOQQ$Hq!
zb@#p%QsOSzJ4iTafl4%%Myqs2Ca4f@Md&sin_SLWwhsObFNQ<iAhC_9S;y(-WAOtO
zov!qs`U^nSnsK-LT8plSstc`&P^0i4&ou+B(DUs<L}2d%?{xXGvBnh_Vf`@8ouAh3
zgqzbO81nHS6;`6j4ITqt;(m{HjpedNGj!Anr+v%?JnL~S)IY_3&QX*>CRI?EtYkKj
zD25T!mvP(;_ruWl^U!dc(nhOv=Uhi`f3G+SS6h2>P#dKU1-rvXBaKf4<aOY?!OZ`t
zjna#GTI4by%Do6%0}SA9G?bAnn&4uU`2ZzNAmFgYH-7XVa%$r6Rg5Xg-;j~N`rf&4
zBhC2C8|KU$GdDcUA9GWWlIVfnr4R+en1+Ad_Q`lm3P7uL8b|8bG{p|jKW9+Hf{MpA
z3*%oIYp7q5c%yke?wZ4h=d=!+Zr!Jmou^e*b<x%hb@g!?^y0;Y`t$@0cL@mY5Ey0r
z)13NkUj3U#r9B91v#>{tP0__L!cj`pMi^pP!?u6jaVF)2@$_Sp{?xbC)>&~h1D>P$
z<EZ67cCVohXS<TQtG;l>gh2*<ee%&WeaX(aZ%H*vFOBTo>9a0vYZw@>=Z28q{Fg(>
z$=o%krfBA;6XPea=}h;KuCoVovY>thi<#-Q)&x@N^o3spgjjpjT-ua<c%VW))?UYv
za=<?CXO|<HJAv>M>B|C_v+QGLbK^t)J?8vspujJJ>(j7v(s}5HMXmlH_F*BiJII)#
zViDj3=%sG{se@PtUs3A%&vqmqGcQ1fQ(R}~uoHvRIszACf%s;Ik2rO7P8kYDyYb2C
zjIrc0$a2zxGCwLfEA%R<^}Q>XNUyU_SDGv`b?#h71;dJFxAx~tH<?1)hdCs*8%Ny)
zvby)LXu9SmtM04u+Pe`Lq}&~b*c{L;$>Qit7Sr3Z)pyTPSB7x+(KedH8mRNragK+*
zy8QY-OCORHUx)3|B?kw#)MF#_EGU#f3ww#>V!gMn%d@Y&WD#(@DNA|Do#eX7&U^@7
z&^KUF_c+v{dcMc2sF8mx0bXFak^W+jMZ-CuI4|9XU>fW;eyf$%C}hhyqmVIccDYVD
zmZ;M#ff!bu_XK0fk=-JXOY8ziLaP&Zs4?C93PG~1HMl!N&s}SrgGGp^G$+1*yg3~O
z>aQihB-W(QY|n4td+;<DDb2h}IdArzm2ZrOBB~Y@x;jrP<yu*kEv%F8U-Dm0Q=EYN
z#&(;lyxy(Vb~WDHu5mW$c=QP@IRE{#xYkYGw;V{mC+zhk{c+17kPZ1P`SJJ8P`7lN
zHV@XmSRYsZ+7C2uI0YFj*~ZN7adbI8ek)k?HBp7*-QodD$ztKkbndB0V=z&pPzuJU
z3rvpXQXiKF>@Za!d4mvV=v5ef`owJ=1@%u4tY2AK?%KsVh4)@9VwNk-A;tEh5~EnV
zx>E{e;P$0$1dTHC$6?TXT|R~S_q^I)c^nO8E=v%To2lBkuNm#R_$AJo1aP#_>)!=@
zD*yDsF53w|Wen8%MPLnRkq@;hHrsyhPtx+IKl&eO*8lB$*#`l0q-8e1y80>G-i1(F
zjN<QRsvhx01IF-(2sj=7X=BA+<XD}~N5DND5J3X=oraRXi!44+#bCsy%;(J;#}L=l
zP+R9zVx|xl>YCY2Z?OOf3%-9;5<#EA7%Xuo>4m1^5($vQ#PlSnpeS3=1<0R>$&03P
zjr`pMZ$Gw=X}XeZkGK6yr+enU`z|*=i9d%UHs=FwiKPO!$s2s8-E`^(C=nn8*vxk2
zYZ!r8FBl6OjM*8Wy&T$)DoWblNMxvsajFtarG3BpMlz^$DS$=OdchUA#gI5fgGCrI
z&Je>G&J^w9+2Wx+>gpOO;JnBcW!`S3`|MfdQyp%RO5$=p@6S|1wbi@A1$pA4^949t
zfO}=qp>nNRd?Awxq%n@uEp^u4kZn{$R4`3@jyjta(e)C{U@wN_ZiA0LtbI^YhGBeK
z8|!+au2*4-A|dwZ{iZE((B>!2sb_^dtIH(cxV@Eqme_PZ7Lf&2jLokw6%|4Eq-Snq
zCi2eW8?;oKApAqT%UaF%^0vVA!HkHK#fygri!nIWW=oq66V|cL=%<%Jn?fWw?fz5a
z-W5{vhgBtSWOnu;d^x0;<IGYlq=Pol{5Lif+1)wPVYs4#7X&`@Q@eOh{-N7SVt?SK
z?8AOH<w2f-lav#kRUlr{#Eu;2W1lX-NH=qM)4R<@%ZfJq<Lgs{F49ZQza2NwJ>nq~
z*t>iGi@kN{wWO~vypx{z)*Q#66GL!n`Hud_Re7CWE$$rDI?hVQEbLnRA3jsNCm8gS
z-1;E+CP|6}Z@zHTk%?#wp0c8^<=rv@9^p8Ji;1<1>dYgKz3b<BR})zs>!*TSFvs&c
z<94x~CkSD**KD)@*jQILub*d2gUIvYqTV%#!6%Iq%BOFyQ*3k5DUx9>tD@&)@YgpH
zg4Gj>;5@9;iSCd|2kVtI3WpRacZ)~7$*UVHnb$OGC;`N<Tg0lUI1r46hX<;N)-3KZ
z_<}mjUyN2t{oE6k{u*AR#h#EWDI9wHMyD%6^q~Y$31U}PdM}of6xg2t-*F<YStW1N
zzk7arTH)de=?%n8n@O+6i|75<h<~_i(raiIYXrmY6gc~f1kAXnOqqkAsVNg<ez?IT
zoXL_HRhe_zJOx60&}(6J*i~SxF5njdvy*&Xc*VrDp4yehHyMt#Lob=8XjU|IZ{AWS
zTj&VF2$@4(n+rF(Rn)X;FK)$hfXrL4Nxo;G>7cj0q+y*&)0uSupnQvk;~s6^-RGo+
zT#XJN9WsDz`;;!)YI8lUKVZofauFzlJA!zh?Le+yq<ig{baGMcib_)%=#IPOjTS{}
zH;+hzDy$f5^qT0wGxsRwTwS-;LfS$Inm*ZBA=T)oTM*zkYG%%>K(N9S`X`?3SJ{3a
z^Nk=;3V?G1rMwBAxC<~_G#&@m6>g?8_&$UAl1Wv~n;xZ|W{ImMe5FPEH%i6dX6IHu
z%Fl%h*eKxXkbo;h2jB|f0>%Lf(QK?UK7s?_*<1$PB^FE9aD1%biz@)Ow08!LR~CUh
z&QVJcIyd992{(GLrE)I=X{#TUenPY3h^$5IN8$v6e>0Sb{LN6}ClXt#@`q-vFeH^k
z0)p0gz)|8U;3y#@hQF1FeT<X-MWA{$71;b$K`}=E%TdAyaIawYgyX1U0T?sy_w97k
z|4r-u`?eMX3CcadQR4On7|RAYN-+FSM~T1JqJOfu_^Uzx&OQ69hyEHv|9eIFYnJ`(
zdimdW9U>80=#Q$bHM5I&kijChXUP{ICM4MSHk5ReG`EoqIWA34a1L24*`xqY>;LB7
z{NLNR1F!uZtc;DJHP8@^C|cApK;fpm8m4%MaavRtOMWp3Lkh@?aQrjL1I@6@0y$Z%
zd+YgQGhjQ|%*X}FP~`@^1$3mW-i*~*YU*vKKoOx0kH_&`n6$hKOWz@gOMLSui>_kV
z96q|B5B-^+J7b1QD2tr$oR+1G{35W2MVo-HQilV|c3Tf0KuJOBl);g_nUNluH*B8<
zVyxHP#`n54Jn7!4ik!-z`e+365Ea@rzZp|&%e6Yc+pzp3#UO4RGrhNZI=dGXIQSMX
z7D$H)E-2or@%8D5eQAwgL$|b46qFzTcwaPf<kOfjcYF7vscJ|;Ot&`SXKO%!5shHM
zeuqRQ`&LeOFqwFlVJLJk1T_h}LvE}Cp8VCR-5a5Ha_=IZEWC*`@#${@GP!Rv0_S`H
z8u?qKRC>|e&E9TN3VTj)`mMxe#K984Bm=mG&9aNmVhoN|*g}$61B{f2<HG5oI?`0k
zOSkUc9oCGxFY!xiGYrW_%3w-aqJc&%^%9a|-Fpfe87R4EGzLG6a;X%3KIW-ppw)cI
z&o+U|Zfcd(J73x3@&%0Zky~H6w+7e=DAq7%ghQe|4}G4GUM#Q3Hp8KrNR+F}pnfrV
zZdlw>c8lPszNWUCHE{Qt|73Y0!ti3~S2cVYF5I`ME0J~Hm%1InJMOu|r3c8hQ(KO<
zSsy14iHvPm^_3<GJtXaBPpLCQ!+4sWvP~UPpOPZSpl{&x*e^@*Vkm%`tG3ADG^L#_
zrV(*&_*qkG6`7hQ__B`sNi$4xuja~KmaNhU+rt-?KQdeDM==K!=g~r2^%G-`Q1P{{
zspX7EvaH4l4ar~oSm#X|#dR)AiGQ_8s^I<{S6$PXTm<R6J(xA2>*vEU7LHYWD|K_>
zJ&)&Uh+>Jg<kw9Jq3E$@ebl;?rF_7ww${+LQRf@wayspq3R-P^?y8LMjNb<dj`AXf
zPr>?QYK6T;Q(KQQouM9lF*cLQSC>-irPgv*62EK)WyqYk>{Tsk^{N$CAp3*F#-}Zi
znAjfkqA<?@{f}n-#ZLQ7jpvGbT7u5L^jf!1X{wr*qi9JaubS+UbFnFNTLNj|8gN1&
z1O>tnt-!<K(m|qF^<)>XI=^Qr&Y?wD1UoJ%-75@qPv+okr@G%aKDF0*GK5GhGH27|
zPs6O*60+p0KfM&c$CE(5oGfl3u%pk7^*E=&;zx9c)6GqBN<KvvJs+p}1ZmRbcrv^q
zq^C;Q6D)h!^q)V|vXK4WS0MK!j+af#?R*|r+c0pXTrX9s{<ti4@z}Bg4;}9rlBs-N
zAN_9SJ)r=)5(-R6T!<TVTc`ve3<Y&*0UDG}CQc!=a#;DSqpUgoS$LL@fDVGlO(ZP7
zFS+Eh%d3aGR^Igs6oe^UKoaVd;5QzFiNP;|;q#a^DY;qIPV*%-u1RTU)91Ex<{<aY
z4<*x|OBdXY5_IoPD@1s$MJbxt6j;qF!^r|1@-rP9CciN8JvHY}iM!+xcV<^<-usf{
zGmqG99a>nXv(RPwyDN`kFS<7s(Eef3i;gl_dwmKBgdOo{I=><2&a1f24#!LC0?E`x
zmlJMYo-F<coiqMD?C{UNRwjGaog4k1w8H-pjQvM{e>zXO`ABKvlzdeh9{F;MRxaS(
z107yXhF3EM38us*2BCgEpAH#DoVfOBprybL2n?G5)43rVjw-~o(SV)l-gf5e+AmF2
z$?bOn$3q2=;n`knS$vq~oo+4m#rh&+Pk8@mdmXvgSg8=qu><{Wdb;O|Eao1o9CViD
zS`Fv+gGoq_5=nu1WUE9`R<+YZ%;U~Jr1%}Vua?8N^#`BEVCUO{rNGRUm5KRItQ@PU
z{=#nMN4$hd&u^>Ux!xfhWthM2>#}zJRm@tmC1#`pL59snAuny}gGEc=qFbPMOR6sU
zHMh#DUbLoJPzr7eeYA0&Ra3H>65E%<=S6BEESWNY5vWM+T{yQj%0{i$Wf~8A&OQNI
z&f6y+ldd`ddc>#J-fNfdY`uEYfy$PdSC@Yv!+3i`T_$#0ygT-4_jeQPB@%%+ANU}U
zYzaZ{q<8aS9XSf~0v2WJWaSz(n_op`=(E_pxwc|x6BpQ@EZ6hV(Imx*HIPY`k%{&B
zu;^+{qr;SG;)!YZE+BaJiaM!eaP|<}I;)zmSyBE83MwTK54l5dyBtojZm{0?yndvY
zqG@i<B8>e#o9&l0Ble{?m4qjQR(1Q*^sQ%OI3m;rNA{rz_xl8G>GB6DTf@NDBmYh3
zcStY<n3*tgq2Vb&Wycy^%EdC^+}*{gh2MwHKSSKk4N)H4*F8zWpVRvZocI1B2#UeB
z#}J&W^s3D1o^R_f!}bI&j)j)UZ(iO8*aW`_{8aX;G?Jmsu#=zgxW7K~ufF{6G>F<i
zTh`oOxVRdiZ;GG+bOYKz{uxdIy`Zmsa@l<vTk0SB+v!i6G(G>lwO1+2?}2GO;;w%n
zA@Hvu@66_rxYyuK<X4kbw<^?tx(c)Jmt7q95|#1y6C>>{p+PF&P$hZUxSwWs+uG;v
zZny2s1a97CFFB15e*Z6PJDVN}Av#fW)N^UEKl6f1Z-@su_AFy~+NddEb-jZVW)DGV
z(K`Bw`*NkZ_pm5tputleD6S|x56*0)j^pZxfz~0Esz1m!$lh&$YnmJ8f8T-<@-*w_
z^)-^5$Q>A=Z}5DVvT)V#?wsk)+JUST9|@a{9Wx)z=+&`xWpOMNJ&l@!aRA^H?kQJ<
z%X6rRXkGl{so@)t-H4>S0#gMQhAA~5|H*>WRHi)RmR|$|DnD=)=%NlT_W3}yA)L=5
z_0`QVD~xD%#jMlAv?o#9y$!w<AFhY3B$M=&`*5_nt2qy#7grD*I7LhF-H8HZ6}s~W
zhnOC^rMsd?Dneb^K#ih&d_sVi78Ps}IP(l?SRPE`Yhq9nx<6=$XsZ1jZyT{l5tSoH
z6JeAZY8^^?JqCuSCy7L^9RTqWvg#KBrqMkis|lp5?E&2fzY_O9usXSL#cQl8m!N_k
zpBrPHp-b}wn-g=C4d~JFLDV#}s5JP~{7PtTlx%NWtW%jBKUv%LQ0I-r$ok{)9I_!)
z&G3;3?U2Z*YSHEai(ADA_KCWLraSQ+(HFPxbd!v7p+6`p0S~YY*u_O6P<8-V<H;{~
z%paTr!QT8&puF!aD0L60GiMCYD?VqsE+Zcxb(_)TLbJkn3apA=6QZteyt2}*{JE(c
zDq@o<z~{M%QPO=|=Dae37n;`oG-m$kA;GNc6Zfa}Fo5Eh3lgin@8+fLFu#-h!MEa*
zA(!+}YMM%OYbYUceqPTy+<6pFkEK%^lNylK$MB`GCKT^+glCT((hPWiZOVS8uJ`;_
z^P3L~?oQuR(BO8~<M+<R8yN$0c}278{0&XDjS;4W_a#_3^dheKkkAvNJWVQmYub<$
z&6i)=l0@TPLQuVXnYQ`n^%c%e6I5Bx#IF-21W8)n#9l@XONe<%sBD!(ac&9e)V!@X
z;m}_M;e@EYQ`{(kwQd7{|FQ+Ygp;Vk*doI#1MR7h(Yy3g^1_W3rE_0N?TDTVHHR`J
zUe1L*R0Ik%bLX}UpD!T84!@M=Hjh_U@FOy}$Eyat$fY}*Sa6@)bl=zK)eohF+x%b!
zm>7TPa{^x?hCSeE`}gk6f6wd2p;5sk>*kMNmCUR96xY^|4}Xi<S2wKwc%5C~Q49fw
zTnNDSQFna<@K9-$^K-6W1Z;VKoFxRn82sz1-`0!2dg7l~&cAx%|B;^94CDpsr=$UC
zDJ#H@>rBw0f`UU?=|=Z>H|+pJtsuGN8`IKaf-Ilb+z?R^ltEYTx6_{{F;M=FfukyW
zR$o%9a&)4yTXf(=@bm8gs{gk@>}3i)fD##dVRA~2okH}Esx0#WH7D@*5i%dH4leBB
z-n#PQr=jcVm3@O>f$|^qV}Pk2qJIq#stecy!07yx_^Ru<&fsxk3gB{tSy{Eo@-tgv
z=)m1_j=}{M)aC2zygZ++8sLtn-i66omus+WJ5M&-F1%CbaqKO^!dvkS7^T;cRsPEA
z#a({L9#Qhokf05x4Cg_NPB2os2$Z<}`J`cC<bs-M0>O6=Rnkm`g1*l1Zt_6IyKjL$
zJ2p$WgtvWe$YSAckS~v>erucNb1F+!6};}!OZfxmhv8b3-s^~G!0F0{7L5rjs^0kM
zQP@*3KcjOj)Z|bB^S{WQ<(uX{V4%2Zm-HfEP}dPd^2#;P4D~rsp%O6v^dF2w{p204
zBc_`wY<oRTH<ik!@#w08Q4`nI{z8u1lUVJJv`Fm+X9a4Wg`&ypu9>;9jPpEPqKc&p
z9I@*+>*K{#dJ8sOr6$}^K?Zo_!WoQn8mNz6SrUxACX6%lD!_)Bqjkd=pT~S3)^?cM
zGfgIYua6QLDzK)=Vu2Nt)r>a#EbXyigQpi0iXcn<WinW2SC{%JquJCHs-XBRgVTxE
zciB$%74C=rY`@HFO}C$Dk%CG~k!p0Q4wOSX_ZGIM)r<DMo@!4Pj%L@`M)C;d-Z5&W
zEY-Mf6z|-chTy|6zK3t(ub^{EDmS1QRH@acjLnfO{HAf8aMr<olDV8ft|jPDXEtbR
zqaLB@?%3Nmi@cHsm|pkdR53PP6P!{LIO(gkQ$%%Z@qmp9bz5mr21$x$L%MpxPAEiO
zJ$3b)({OnV>(jRO3SW34f~$WCN#mQ|Zs3~%rrD$m)M*H%az|Y^8{So+P*CNKqT!`g
zyTyH~@Ao)lVv@D!wSh&|xbP7d5f-dhGj0YlXlP9TmVhlS70fpO8BJ=a##R#&BTr~l
z3y!SkWQ&;M;055p*HXn5_Ml?k>S^*WiO#2H*F=h^7I|za<;cAc7Yk*q9`~b{p;@2<
zWLm|;c`SuhMTroy#c^e=XyCSQo${SPJ*+C`1_;VuJL^t$i4=slgtKY2z@FUO<=^L4
zj(6!8Paat!>71ttRBFUfS}DLDWvXE<pDvMchiO;8ga+$uycG0({4?8i%*^JeFOiwD
z>%4v&%EKwGmpZIjnigcGRoq%NP^;#3;}?O1S64$nuV1xo&$qzt-qZMd)S<3B<#yGP
zt#b@9ea=L~1q~XPWNR$XUEizKQ>uPTp@F@HxXDSK)_U(j`L_3p-0TCcO2qg$ovHav
z0y*`7gvjhb(dJ<<>L?Xfw`TK3cPE3E{PG5nc05__I;Hg*jQ6`Na(nMU^1HLle5Y>t
zra2@WN3{eRjKq@;H8VKE9Zl@U<?N8*Q-)xv`7Yil?gZx8IKHAmi$)Ag7o*V;FWW=w
zxVNYhiJ-<@-sohzF?MMt^O-zz|5SE1$!yQV%r3;0I?Vo0cr=a-r9u-+7Pxh<>^1;D
zX9Z$T(~4|f$T2G#YI)qqD@;32(L(Hj&gM^R2dw%tO#lwFEaXONi^w!Mr*1F~MuXEw
zWgB#kEAywfb4`LN4u27_X7fxuu1}MLXIUw9?<l%05?{NO`UD=IFMn<NJxa;B(~1du
zx=3-ZOA9FM=lC!?@u|n}m0I4%e}r5OD$zHyHX}6FOnMCXw6U%UTV`&6fDt=Tsug^K
zphNF0Z8F;EHao|LuC9(kI%(~u*Ah>X10#9Y+~lNLv8Dx!b7Eo55)IDo4Y><C=WQ=O
zol=PI)!6&So;+=Qed>`Q^w{>>aw>A652z(es|KEJNC7JwrksyJCD$kbk6VPq=L?vk
zV}<XO>%+B9D?aBztQzCocLirAR#*J`2O-+;<~$GFY?E~}6~!*qDO2D)u#rnZ;tO9=
zykXwS3W?!dx9TagSbu5S%E<DZ!urh&+l23&w7QmlCghzdx0O=q1Tgb$79DXKuIR0{
zdOUXyv`)M1L2-loBZl;Ej%t<r=a$`z9uZGxMjOE$TO8C~viKV#&4efpl}`T<o&(c5
zA&!-=cfKau)?Gbq-PFCOhi6PnNA_*)1;d}%?zv8ibs^KvT>|$|;CSHXi2P=^6)ZFs
zs4WmzS+^JI506gB^eURZN7S;C<vqxhy38YILifqb-X-bnPi6CMMSToK7Z-Uj%~r;Y
z3QfwQaRq-`iWfcch;GDYm*_4dQ)6QP?Ve-z@_}nv)>1kMJByuTSCtNTNAM)<89Ffn
zK<TP?=7u}A58j^hQ8jq#Kbu}wT`4qrakcA^L8Ajo?V`w5xGC@bYJr;6!9O_v+ev+f
z+q7@x`-pLgT=S6S1Tt&)Tlpg`=||o>CsJ@kPFji7R4||(5>i3#tp<}|N4m}?Vw=rA
zVv~D}8)`;&wkut^I(^I)HlG$8Jct(dcljR6$*Ug!Uax+GWT4h&QYaD)3H<<=^Yj+^
z8URl+bT^DI!oGGW%lyIjyF@V>E+4W8ze{#OF3LJp4k<=d(6$1=x;Z4^AsyuAGGsdM
zgrUB;bU27(0XTMF4}*Sp)#%&5cf}9l0fzWv_%)y;f!L#81Z{6&xI~SqW&roO+amdk
zfcOb?Ux~nFTsB^1S!V*d)xPvwU1qM|{m-M6#jR|wW`PQ_+8rex83U5%qkMK0EA^Vh
zgeeJ=T;>E@>*bI--$?u`e|tP-K0tT$R~}pBIwRVEVI@)klxh94T}=16NfdOQsSvu*
z2bd$}?O1XDBKTkhK)+#vDyN(wG7f)z#9w{!*SPqH=Ex=7{1$7B-3E){x0vKqqA}*`
zh@R`T#OeVHK{r2;Rs8=f0r-<y`hSb{{4d%2Kv(ed&HtT<1Hf8)v=9EFOwum`SLC7g
zgykZI8RAx;_jof`u~Xx2Dr)}fy0LuZq^Hi4rAN-wv0o}p0#l|Nb?T*S33xs~dH$^J
z)8|DzgJnTtV(ph@Gc%@vuMq;Nr_nv+A5tbh!L>2Sa8D%T)_4X!y8_$PhUaOhcNW{x
zq0o7D07{!*(Pro>|MB(m^Ad-CLAxpS)xOPaWSPpu?yOT1*tP%x%Yf{efYIdXoHsDK
z9d~e&3p$Ej7^ZId(7Bmui!CPmM=?suWOS=vvU#)L|ImdnuEf1(TjdV6v9KyQzuS%%
zMyJQ@f%JwvbyPa$Ze>~yr3IlxJ<nq`3!Z!kPfbd*j=TFcFup0OGDG(@g0XhUlfsqc
z2?i2=<jSaB;)=UhxUt8781H$$`wi((mnImv)=|IyoyD+nHjla<tGaO0wKmhtRr<VN
zn?u%ccqMeRd!=jNNkW5vIRz%*p}EK6Qldwh*z<^W?qmXt7!Vr?g23OXrfPYJsOTQ-
z47nJMsW`+#+ZrF&sl+MLK=L-PVF7w)m%L>`TCH2!F53h3v^jej%qd->jBq$@A@iG^
z1`-AB=5ag_E;s*BG(0PMeK^b3aV%fvsLX;P8u<JrV?W?SJparkvgiZ_!;7gmnFVj?
zt{XWuakW9IW*y~c6B-t|*nQt0QIW58ys+F%;~OY4ef&Y2o><|*lOC7ny|<Hk<J~)-
zbL`P$g)}GXL}ZPJ>js&9RjsP3@!u^$4@OLTrR!Bm3lbK`O`j9Kx@E4^XI7s&GYR(z
z0H38IN!X4Z>%3D@;3T1~2_V6T+Yv)rO6gnpW_GDiE};isNgV*CH<>pn0z6zj$}O`c
zZXRO!bTd{x6TI#3DK&%AB%VXwg`dXTz@`*H6F1QH*(k^r)e`5aPV?3;HDl6Awv4cs
zg)?eIow__7f-kJp$@{WUWO~eR>j2Mb3Uh}>qWZ8B--h<H`dI$adi?bf&y(ac^~RtR
zJ8-ReKRv<&HSK9nS!&`o6wgJ8Q$R<)J`$sxHx-L&z91gaua6pd>i&zs(Rm{`W7RJ@
zQiNBFaXU0mbjF52dv&i88I+AMsoSeK`dK_FU37#G&H$+rt4!4oH4BI~<2lmqxCQb%
zFUE7u%qGg}`+@qL6<yjjQ<NIrh+duF=ae^wbm~qkbbzS<k1;ASlP~Cj?4F3l23M=T
z^UfiV6E)wP{O@TpgPd;ld6i`1ZZ5?v#C6jP&=CG%WfsSSo<&d;o~zll;Onso`2_hZ
z@`CR_36d1Qen@v9mu+KNzwOzqfj-e1Z`Lo~lJ|)Ye2LO8dj~SD`cc6bA)=LM^uXSy
z&hH~c#E4?LmyW@x_v;mhQ@`E%rgL~v&i6qdPnhwmX^=PIk%X(PH=msQnzsm^5;7ce
zQ|8rmMlpL)QMB!wgE;OwxpA(7qFilF9Fk|hd1DItrK(qA_&}k`n&yD@%n4UqE$ot0
zdWLr<&t{Hbie|(?J6_~GIP3nT#lZ0^shc_e&tz1u3>8kC#^9oKb!1;3>WjsR3VZ1?
zZZP3*AR0#PT0T{0^VelCbB9x!HQX|BZDiDX5Ho4rS~6`K%!SoJS<)`{oh?1V*wvJ)
zF48Rq#6LNIBqY5vSsQSd-_(^i?s_<yQR=ZFt(M3a_3?=SF{Cyg8D8fv>jqX(x!^C@
zZ$VnhX0{<~84DJL2Fbuq%5<?>=)}0$CG>7_#`~PrC)bk~#p*!sjI3$9D5|&9tlH{R
z^a9x~KGr2hn<u``8xQ25jwlBLR-H;mI@Pj^dUaECqBOab-g_*7vOxWQ`FXlRsBFJ3
zI=vlA>8oKr>}E5j2rk}`r+avI^;MAdj%{9T6s9WDwzLnMphXk7!S}QB{!+{jG_+%$
z+?T(Z+~qljuc+-zsXO}ValE|EYKLsnmx9jAqRQXi3DvsKu*cEV%^m}kY>}OD9OC{6
zc<aonac|jwg&`gwaN8#0&_JRBZRGWfpnDzIlvnrQeVw4#Ta7@)XfK}M;zF#pFmx@V
z2q34G!SS*hIQ}%1gAO3^)Mf-sShf#T8A5D$+J3H66p;K`DBb#>>7YM+@NW_5A3od@
z+HPh8RPvTq!7*4u0YrK-w+e|82C`W>tS~%zB@#Q!g-$aC{J^Pnfjy@Q80Yp+Vd~#)
zPB*3q7u|Sr;cEe1uh08K)5djUn|1A9qLiHYBlRhd4Q`4R&g3q4cch7jkbJrG;VR-k
zubuzJy6vy#{5Klof8DHq`({xbK+By3bKGZc^?eL&j3E4xXC1ene~G$^B{$?oJ^ohC
zhu==W(+U5rX87;@%<?;^KLT>(YT)p{YO9SBR;khJ5&nO85Ho)KXZ;VRE-2-n3AKT)
zcR65f^QSkbG%o4^@wk{(c4?qRYzLTL(IDSBolzbVAqA%TxuqL)^C?fS^uWyHa*t5)
zD*=Y0i{CWl2>cTN_Pv(iF0ohsJ&8coAxpbKy{Z}}^T)A_u(iEFE620AiP9T)Z$EV(
z5^NLL!?gAurvi_XkY5C43`N3C8xR~B82v>o66z?zYk_^W*d)hkDb_y4_{yeKOSvlF
zYNMe<YJy3>x>ojPf??FNpI7{zbzYhz@REgbD$*2SJ&Y!#smxtPJ051tv8{}HZk2&z
z?LH^jgf!EtSBHDQxV$s<Vg?1`)L-HXQC7sZ_$}$Ec1f%&v%>(oT&#}r^!+h0fqj80
zWX85%`XwPhnI?Ivq1VrQdv3Ahjr=08L(*6{D03&TPUsAK4FGVW>dQ6X22;%Cpp5nl
z(RWI-CyI%(&mGCbg?$v<`i8Y190CmNhx5A|!=1?3U4VIN1U%oi#<DBn=TCgC=C<0!
zH;fzWiyPOX5+g+!Y6GwWBX*2~%66}IaGiH(u2s?ra=+n->aI}~IXA^Rop&*sU;XHz
zG*Ejcnp&c^7~o2apKIUM-k2kO%iFz3B+T2PCAg|?u`1i4sD@VUk_^1xj8T7oBhjO&
zJ1v+eVtjYk)CY7ULF&}#{nMaoVmmKAQtJ<lmYr~tmuTfZRAh(3GA5_)!^RvNIMO(n
zc~lzOpv5)`25Bj%N|FC~b=^_az|D+V{w8_X=LCmt#r+M)1<wv(ASp1f;bFoc=1py!
z+G248kkNg95lGfv<RKk1Vd2Lz6$@AwujvcfPVY0Hs0z`FwHp+=p9X^Ngug9)kBbTu
z6D71}@na~#&@B~KeZalII-<C!ryLumh`(^-r$z;??%z$Gp3$<qKGkYFkiVQpDf40D
zjWcyCl1NdlGEfQY7_caAr+9acJF_e2X2g>JUO<`l=P4pKKZmYw;!o@2t2o^WZevM~
z?J;8EE`gkl5j91dK2riQL<ckJZ|Dj@8z+o1dL@D>w|e*<owo1#Y1(u<>Az^UEz?~-
z)T%s2R2}+k+kPwJbU2DV-xFZkdY*D*EWZVvifv_8j?U{X$L51AE!U6FM$`S>!h$~q
zB^Y~x>f`td=QxHC<|dtikf)z%&j;DDTRqCRF?N{?Y>8Aq9p@M`6ofM4-M$C(itU+S
z`Rs5_O_Aq{>FR60{N6P0)7wZ{5+I>58QUeV+YuDdaR6T534o%%PqZazraZJ^N{an4
zEH1tVAkJEel!R_<D+~734cGtx>x-0MixO?7B8l)4QlYuXS>5MwIu6jNLTEee<oHhA
z;XfEkX4ZY%&ifVuTW$X`DF^yKUVkzYBxC8>WHP*3z4CZy`mh$PD>IQB^f=KFyON8Q
z)%r0QSFu;ibq)M8cDjQ0`}ZJoXCeYJbGzDvyKnMn3+<-4I5ZwYVtx@g3`XWqAQMIm
zC#DC2lsUjb@!<o@S~f!R1tYzbZqdXxE`6Q1a0OA6U)H>+BYqLkD9SEsHmiPc0%}Pe
zWvh17Hxk2NE^2{wzkO2t-jGy&O|!_uLc4LEChT=*7>VKuPLdOzk5%c!bBsu9E4&j^
zW;Ms&=v&;$k#T-!jCoXDY&^+w{=j}kSe@DCRq9v+7CJo*9+pR3?7lV7YLYF4C5r5A
z_yx^2G=0GTI0@##9#mrW=BFJRIp3G7ew?13!-Dl4kF+DIWkpuz(lvS3SMF)dB<)=(
zF9yo2T@=(h!`Xd#JMc_tI-1VP*Mn|t?k{OIYs>LygGwx>W~+ti<a;j(?$6yNUQqXA
zd4V51JHzgC*zQTqSHWY2kH*z9?g>mT{z%B#RDR@_<2Xb7vL-8Re~jxYjt(1M*n@89
zkkvuKunz(+7e_xE{m}gRJUGuK%|~X4G3+ZDWyqt<D&bdO?TEx{hqc@*b|5)1_>6g-
zz=~&}M#Mzo%#MhX_XJ|C!U4>zD6ptwCnBjknxDMrAu-j@{}uzXVy;=G@SDA2kV+W6
z|NgVF5}|bB0&+?LHQ#w@O&vq%OPqDqIBWHDW)jrz5dI)VLkg3tS(T=D&V2eDb(|yA
zhqxBpFW=KKPlJ8TanE({5JL7oKB7z|_Nb}1=mraST;+S-SZ#GOA>v!M7H;IxO)Foh
zx0@=Q08?fvl&gkORlSZ-YX<!F5k;AL^8Fi431f_=hNXLI`|R0W)VIWYJC!o`ajs%x
z{1Bx9p4>G@1wd)qarTSAGEo=3w?nibM(=P|Sz=$IO;q?E!ZAB|oR;lnBan3Cfq@1)
zhPR(R{9#@<i$y4XN|$Fa9cDLn$*v_)J7>XSQi38=VM}nk^64Czx`X|LH$nv7RTdl-
zp)X#pDBr+ZVwe^u+OuuibHp<{{3+Xn*&ai#&dnj;zjT&=Y@2NTNl}njjWueKD39&k
z!yX4q@GBM4Bfw$Cif6*BiNcvm+c0Ko?1>7)<{VFY-|Md(L1U)|@gvHkn(1yJ;d2D2
z79u*UG6-1P!)>mt^kvSUA#uus5Xql_03fveZ}q)Z&5i)szztvnS}}2*iImy>;Sa^Y
zj_1!AeO&qz?<21PLY|wM7dHBLssEpeaERPZxag$Hz|rFfqBNn}u}aUS61A5r@X1M^
z>eGM*`)}<EHf}N{a!U;36Hz$5A6}l=@DgmE&M-u5*HcOI5uyg<YZl+H_R?EaWa&<a
z=8uP4wI1hPNAz$zcmwWIg>T>V;rY<I6&Ln>fp2Dy6(71H%LEMhVa(W=qd9taRZr7&
z+S{~})Qy;$eZmYZKZ9lJ^nS4wAay?J2A5KjSkgb2sLo7xRbLD^_nnVJrt?<*@QBc#
zF{($;95~F0E0g~aJJcl`<ji2wN@&-S-enQ6L+*>r+Bo#L8cEhdo_70>nfD4H!3T%@
zxX06ukPL7_QLXA?fEmur2ep#$G{D)G>inyrpPihQq25)GCUF_xZqnsrMWe+4FzLo+
z{pJKm>vOf1`k(ZSvr@g}oSF+#Avyk+Ckg=jpkiNh?Gl`;KnIe177NjZX^!wF$PV#;
zidnN=9(0d(n=_z&#>a3<u@-hfT(TXgP&Vu90S-`b%xQPkcyN2;SegDapv_uB#IL3)
z_!an(O*RAo!Y8MDE;8&!Q@V{fAQ#5@i@;|&(P$ViXz84OBJPMF!%*#{w&11$Hnl`^
z-auju=;Ii6UuUWc;wj(UVQ2!-hS%bN`_GSEVTcF>lGl{MCE0(|ronf#>ed<|Zq%0J
zWp<k`Lua5E7IH<>sS%i4$45CI0#!>adM-Pi`~X$6eR<@1{PbpY>f5jrQkS)?31$QX
zM%&8b7lEyrmFcDqS8RB}=$l$+_8QCXaiHoXQ?RU{z#9n(BDL+*b6G6-Kx*Z&EBLW<
zFmyQp9;0{<f>NQ{Y@D6a7OBj16j?4IAXeA>(L{75<S}8l=Gi#C3jHAnPWaQee#Sz#
z&`KL}sk`cxl)JWkTh(Y%U9!g1k{lbTgc*KqBv7)#0&CyZ>u4`^9sBtBdysrjJCppw
z8A}OjVFm3MtE<+kAsC@c9Ee~;&h$l%u%lUGHh~|W{qY(a6LnOgI#1f?B!<l_)jTe<
z)#>9XlTy=vl*;-@=f<pRwXlIR!7*DKGv80N&hYxUO8nY{>Rd7W9nK(ke%AL)sGo~6
zlgf^Z4pf^i?(2QFc)An(^Hl+B)l1s;A=*c`n0aX55k8g{LC{UjfgNQVa+v)fqBzG=
zs{4Mf7R?y}ncaqO04SgHdthEIekb;FSkt@s6$ViN5>gpN-_lOQgI5I30(XOG)B?v!
zd_h*jhZVF$BTu*}9H$U<bdD!E?4QRzMzgM8#D~8XO&E$gd2LHs%KS2v{G6%nLK(1K
z!$J$&iU(c_PZxSJ8QTpkMHhI-E2ap?#S|MEne`KqlU}*TI%e@cpmGlRYH~=8{@*lO
zI_t+@LsCP3kN|nq);j>nI70<Rfj+q)@EpWl23W|^6&W`#v;B3=U;Xpf*!ZtLFD|90
za8#Gy?ygc}eXTik*GTU+!DYY}E%3JlDnZbHf+f@6n8f`(z~?{m`t7=ZRxHPtZD*_E
z?~5IQmxY#`2-X#eQ0Gv3Y;})zlb=lk?Ay$bH|(zQ@2uZw=qZ?hmUMU*);fY7iYRm;
zX%6QpHt8M+Buto3xNNWZ_TFOWd!=PjM)@i(Y{L{u0$t8?<T?f*iU%#uTT_v7+fXMM
z4YZy6gM<85LzpMLw(?8GKw(tRv=>zqrLhm&!FP8iW)6?`&2;$k{JK%gq;?$=s9Eck
zVc(`g=4>8RhtDpX1OJ@q<`3N*UuEG*CH-n^d!LzR#t4Sh;{=}WumN2>z6Zz)Y<;Wr
z6r%~ej)uL)QT;}S1YdJ9*qr*-2zgaIP5dQ2=xfe{hI9i+jVgq&(XgZPl<<Dj?bbz{
z95#DtFBXihsGY=F7W+B?*~%;@6Gj<xkJ%Xx6}^Zls+DA4hQ1oV2;q;y(=TLFADhdh
zP2lJt`_7}r(}Q>_#9^X=3V9$Y;sz(XGT$5D@CxNQDOkh3;~H0_efEm${L1dB&l1a$
zTxmKNHJ;6E(>x+uC5|^tZ)Bix&RnZxTDZR3o6L>x{uf>GD8m`z#k(x<6g*uAAe}hg
zL%}<>TOI~<q2aHZ6+}P!c~%J~YBx2GUjL%Hl1qJyhVRL~WURGmRj*QIY`N1M1qL3c
z=uz}_3KA6vwn14MZ)!wacFJZPqWK<Gk`lRm)A#&3rP;RHpm(ttepYvYfZa%$X!QQP
z#%aPoosL;S&UzKuB^DxoFHxv;;v&Y0e8z-fABwSqv=;N9?0hV&aDs%sN`AyVA0oG5
zVHJs?SY!h1d(<n$Vpm38jvyRc2>Qf($?f9`4iyXdA2R!7jN*c+=Ht>cU`K<{rs_2V
zz||0U9gO~lq(Y(u0W<yJ>7t9WZ3JWrisC<*7?@qgAM9{BA1HqNMbLm;o;B;4tJ&B+
zJD+W;D=KpYgL3nOGG7LiojFWRtipO_e;k7UFd%?L(eE`b&3(EtoML98iq8}cQ9PJo
ztfnv9cG|rCZD8*y0q#H7n0lnKbN3WsHLR@DJRY371m&7{Hn-gwOxT0g91XH>A+7;M
z59GUOzfW-&v-vG~zsfF(n<}J8i%*Y;PcX~NoLK{pCwZS9Twkf%8>&0;0Pil_`=MZf
zum)3b{LvP*$I?+bx3N>dKmTaV)s3%S;r&aV+~i32PiyQY9A#!e*)?Syb5{}RvcgeR
zxajk0Q1PQ)x|&;>y)`e4pTA&^qm3g`1xp9AVbd1}BUFfLTVw`1B`kk-EWoKOr*%uF
zKz{xP^FX;zofm<{OCs$0;nwU8XU$b|&-_)`d}}(nh=RFs`lBv5edyAxQ-3?2qv720
z(A?kBT%^0mhmDbjznSI1+ph>~wezNFHyziT`~B%-fdEY#oc>PB{fp>0&JU^;bU_{T
z$Bs?>L(gyZNW){nbW%tvY+p}s$p<SlE3CH?5NCA`<N~Gff@<_N1<YKxZ}{T!ILRe%
ziO=7l?j!vYgrV!6dR!5-z!LcK#SLZoI`k&CeYfwcRhzTS+lIFL!a{C#jw&zWp7PRC
zmROUeG9H4*tH9^(7evPvO_sRi7W@*>*58lZ|3h4J0A&q)gn>qSaS{4zXiiA$a%q+H
zmWhnFpzhTA!dC(`)IfGj?iIzU4-!|m^*URpZSH**D-9&Y*nL@oMED06E#Bb`Gd|d@
znD75mm-RKtLW|vP$65GIG&xVYp*zD#arbRn6BF>n(S?IRVH}W;{Db`IkD|1RLVZnY
zfMfUjk@|mh=m;Pp{x5J<4NeuUa(~#9>hT#{i&{Z_ecE%AWQl@QG6IzRbCXKr-!lXW
zxPHpcABt~6Pj-PMY~0XA7%*~FIe|_I-hpBVT!!v4&;NDBU;Xjd`1sG94fM_R(MGK5
z?^Ewe+#!7TYIK1xWt$)u_M1^h2$w6)Na@~hr~i<v^S@@WA@R4l#xcs@i@l34Kl>BF
z;lBqqQi}Xb9D6SCHv*&(`T202mt2Xp^|djAl_~-Ca{1yPFGkjc8KRby;sg53ob)$G
zoT5=XR9<o)wH?Kbk6sVbN=m2?9tWlx`*3TuigtQQqhNm<&_V)VjNkx*HQ7$x1wMUA
zCjfQm|18%WOJNcq)TF&9{VnDDo53q1-8Mj7U8xX;z^<>-q81fUOO=96<-{{Y8J=s1
z^VO2dWl*3-RcL<~f(Ts{Bm5!dd(6D~OCNSfsO+s?msS}v7WRWD>v$Sn>Wiq2U<x!M
z^Ozx9B)CT`_1Nou$2Z<-7nx`dqfF+Ry8?VBnZDp9nG&vWMT%Ts^?MlIPXBN6uX~G~
ztLr0p>ucU=2sU-n%3JXEQW6CwHQrs0LaR9j>Ro)O14lHvCHWLPSh1nk8>>KNiC7(+
zt#kgZZyt@q#1CNK-@oi_*NCu?QwSZ+C<@qt9)R<}A?gRyXZT7ld1!}{kk+SB7nfiT
zUzxss{qPF7YsDCE!sZQo?*t>p&t4+JaYof-Nr@QNp=NZ|XrW&vjsulhVs(ed@kVBV
z#69jbBED(Qg!`5x=&G-6tIhWScG`f7)_5Hx$|qgcc#f_S%@tYatR#xZJdF^9L)@0`
zFIOc<Xy1p6UBC5w`DQshTBWuwy>l=;SbP|#QZi-b?5~Cou}eYi#Tak6iR>QT%oV>E
zz|tVz*Tx^88W*5iG8^%=(oC(mpw#<tA|1WI&W03l4SE7Cov%pzn1`Nua{=oIbO9sV
zC)G}RJ6VahHg)i}?Y{`}UJCt-phhnekl;g>7%?tjY8*FL%;ZCv#!NNLliT-T+f8T<
z4(CKPI2%whv?KT@wgDkbi|8?Zym3#Dl`Lp|)pF)*wN{B`k6U!|ux#J7Mql+c#!re0
z=Qd`}tSWI2H_ooMl&`ByR&_9iDe5p*Jq?h??Lb@!8pR*jqzqWxalji{cLcvQ81MHw
z+bCMWQW$^989c*HZrZn|RV;BYE-mPi!5E^NlPY()-1adI_ON5i^OxO{aYmnUQW!r%
zMQ&6rBk<*J@4z@MM>BD~6BlgJpn(*R#CNChDF>_vVr+m`DKBqxo@!@TQ7j+pZmHS@
zo6N|q{482gxue{nc>W=M8)2t>|1lP2jV5Bzke~t+<A0c^N5K{OgNm__MFiTFC4CY*
z8grfO0^VzeYDXc)+@2bRXw!@K{}1-wJF2PeTN@6?f(i(NbOoh~G(n0ACMwcJK<O<<
z>BNBa0D*YuML<A6KtMW3iAagG<j@5KL`tM3BuE!Xr~yL!_V14GecyY$-|x0@@3{BA
zpZ`F{9y{54&$ZTEbI&=S=Xq}#bbP`<(vfFhp2kk44}WJj%*6GX?n9){XXl2^4#RT@
z!>x9_b(o;D!vdEaPsF=q)my)iefL<%o&;VbOnJE_Ml$%9?AWRsYCeO{BUCkN8p6KU
zOfJ1mqxPM)Y_2z{V@g-NK7r`Nl$C`rBeh0DWRa?0)NRWaRS4%3%_G&C)UGKXK`=D(
zUwlzb<ujYIo)e7xTyi0~oM_r}4a1M@_D8MkF9ytqQ3mxW^<fe*`ScqU_oM>dr|BLg
z7ME|k4;M%m7Ax1Cvb8)DW`E&r(r43t`_^hiXuEm!sQMWaT(8f<C*?tX^9%j+o0s2~
z->ZCbH(bVr>B)>pLR(YdWji3V3>^)x{kplTPr_zYBD5ykGNalDnV|@~kqcpW8(73&
zT7&1m3XJ5>)-!Xr6iqB=3T44dmH_cUpF{ANawLnjH9q!Ik`}iy&(txH59o2@4Jjf@
z!EZwW`&S^vUqBR<nNqC(JglPL<qhjUi_BPYiUs|L|3pPDugyv&>86W8iR;pt_pj6>
z@<gvdW#8Yr!us0p<31j0K>#3tNvd{#;~q&o<QW0|71T2H6I{hw$>_N1Pmq<L5GpmA
zAp2iG@>kT(FPi|+reE{qKW6P5ll`JtmccE4ndPm5BU{2PG1lk1f9@L_{$Ez|J@3(0
zURzDwNmq}P7TIgy=-2>A*+?dwfjitw1QE(WdO4yB%DH4U7^f~(Wn-mn@9$ga`h3No
z^~ow{4*(3Qo?}gZG^+9Fql^}46BUNrFs@BJfd2sH+co7*@OBDwG&AS7q7nL*l>VyR
zofe2;a<|oIcX@GFIgiKniZKN(;m3>>>)XPUqvi#%rlaCbeg^=Y^_g;2Wbi_pFpFHq
z8_m4s9C=DHI=?4CMqn+aq@g*rYBHFZDOr~+rM{%n<OJq7DgMNyUyhU>s;aY=)xSqI
zPgG*NBP(&9b8T&`zp7*e{@B^Ua;~X1(_lR9fkS-Kz-IE(4=ddrI`m_nB=l*-tPN&_
z4y{rmItZK&^FL>PDp{+dO*|-t^G&Agzy=&tll@lkiQT!wH|lKU{=h+<VP#nQ2i)fI
zTyM`J+sSb^@v}|A$;&!vg0C;vgv=2y7bY+%>Gk_t{Q#xxxvkZY##xlNKmdZ8a<s<t
zVa9R9>F)NWf)sLGmIYm{7~!s(HZY@q&Wy9U9&>8xx1<{%4=}+~`SgO<i>mF}S07Lg
zcQVMrJc!LsMO(sGySlX!2h&-V^pl>P+Glx^{*1L#Xs)<DC1gR;>+v`+MjX)|QVLMc
zJsY=D-XcV_N?tek>@mh(-8&HHHx&MIvb|BYPmh)u|6TPrKjqwn9BztVQH5p|EmD<Z
znX<UElXgRMDy8UCn!ZmOfClCij5)7*<JAazhJCB&rmLi!vm{U?7V;Qa-`WtC#!iOl
z%C|~~I1-~9{lchM+f;LeM)1}c01}enF6-0K@vfh>OzD9f&GMwBs8)NqR2LwaZO#nj
zksh`OhJ&TxsE&~9RFyokpqilWw)6_S2n<!-FEjT3mTU!Chj3kd>hzbBnjax5YXnpx
zfjO$JQ9GLJVRO^_(c&~h??kt5a|13@Cr;+0%xp~65iR4eFg~f;wN7tdh%hf%9H3>W
zg+}=AJ5Dodl~3lHj8&|u$MMM&>B#X-I`NC6z!tM7u?oAB{=G*uv!@>5sX)hm2~_?9
zHZ2GRef!lCEU<)0yP-3JgwC#7n-BZe;vbE9Jc%1zkdbG8DZ=x3T0}u-(d=4^THHc_
z{Jno8m|#gJ1b|;0qhO2tsW;SY!7#P$>N?!e>bF$o7t6(x^)*_VzDXGw>WSH#{1@K8
zaK%GwC7*pqUVT(GSYaccy^}w2_X^5!I%^Pff?)W*t~72?x4x@G>m%RXt)uTlvx|Hi
zQE&C#Gp!9q_Z5Rtg;k7IA`F`6O<JvrqTv7B=Z)~`99tOP3cB{a@>|oeoPgO~qc;r}
zCR~4fWm>gm{|Rw`K2A$Nu$ZuNV{p|2dN<sWiEmvqi>rglY=G8VaRGOs%Dpl%IIEq$
z)-uESmDtnA8*T5yE;|n)9S^h+>?C|nCZpP?cmU|+Ta!_OO@5HcbCf|prD%#wu7Of5
zE%p8SZL&h>jks>;hNG_1m*D%uKVb-le#T9#m`#w?`Vb`%n9j@0TCMyK6xqR};*!1v
z;!2}K|1!#=M)Tq;;!jYA#J|EE{vihY$qxSn+R%TN>0fg(ko`}{3&>wl7ypnv{I}l@
z5OyMR1^^|;5&hHo{NIa>pau<^J@HX8_8@jIPupO7#4zLTbP%Ii4zgu!#PqWb{|WZw
zZ}HQA%8mXvv)ixR0kYQr04@3-Fv))n);}7oWAfMCxgCvKZ8boP#_zNO9xyR~eLnPm
zRZQxDYU3rX?}Gp;1!jOzj_Z~X(_g^P?~qQEn?Ui+@JPI%K^lQ2@9k3h<XT3miZ!&=
z^FnW#TI}6FJ)b8fGmE1Bg~p@;r(A#*f9+-hszc~FZZzZyLZkC(zk6m#Vv$?g)voUZ
zZ(XmQc>dd%qfG95J?G6qn~8jsZh&ordLAo|;O!8S7?G7HwqrJ@h;C@|>V3*`5WtMl
zk+hU}%F42SDJizQ>Bv*I?*1(Z{bC4E+<$W#Ab)t~34CAar)Sjb9ovAd!x`x9arTT4
zA9le=Ao2FSlbkI{!nTn1&|7{jIGQtT=3}95kwV<Z&P7+gSDSOF!tO{AhHZm#^|KZI
z2=%BihwOG&^RPi-h+GZ6_oTxnc<^+%&yGz1g5bFKY(T~PMTpf7gm#AxnD5qkpa2H~
z+TqDA07y{2ow&sW@Nk@lFeGXhkOQ{RA0YAmy@Ur|E&Ih-S}Cqobe_oNvKcZMSRX-d
z(e4o9t{zFKA#1SOiH|rB*EUr83U)T0!V~Vsc{~lh_O2r8(;>O++7#DJ?jJzf3SoYW
zmYUaH&_hmoON)h8FNXvd<rEbYm)Wk3%k_wbU)4BT{n0+<(tY!V-a|=%JqsFzTRZ{~
zXPiUP4TLEXv>?W_8_o=-ZeZ9)G(VL4nc3;1iTKgj2lFlaO*mWB;D&mW4b7j*+mY#>
zgm$};7UxMm6z&kU9j?799WO_af8XQw1Uqtp6Z^o`VRM30<6TFP*5!l@Kn-&jeBley
z;O&&F%6<nySx9PB=1EDKeh|W3FS+$%s8-B9bfQ$;d+pb_BOqbGDt|Fm;?K5p?c8WW
zuK>7&Z|1p#b-A_O<;*>>2Ls5m>sl{V2JJke-Y+YLrBR`a%e!M><@)E%U!DnCH0Z+h
zAIrtQz58n5w95L5jcRwL*DMK}O9hy!@7IvBTCBSQ@BO|cP&C>h&}0kFKA+iLGI;{f
z^ZO|B{I}!_F@0+TDtJ8cuXX|=roWg`F7~FZM_LszNO>nQ!|*MfK54+voaPz*UK_*`
zpfIm-)U*DE!=>Xu)$?DPtfo@eRV7Cj!c9c-8wHjA<?|m{)@TA@#a|Y+_CL7@(|^m8
ze$Lzd4}FaPvG)IU<Nqy}#jjxr@z{>nyxNm6ZpLSCksp6Y%;nAHQa%vVXC_UM!T;R4
zv((7+|675U8u@!PHUFKCGnxHu9Ju}l#bMeZc0L^80sa7E@b7~J9|3p=|B)iOQU5>I
z|9z(N@0AVtd*5sO-=w112LI>dzndQQ|BEgDKkFE`(+`lvi1GjQHj7`={nvE=*DR4=
z)BWs!h1dV(-T(b!;{Tdy{rArKHQj$r_rDjk{JrD9f|>u-Kl)$QGx{r-`4!CkJ&*p6
zkN@)Szr6e3^WuN+_%HAN%e!m-o0iS5*!@@R{%>dF`QHGTU%||;VCEkOEr0LxuVCii
z<dy%UvN*rI`+vQ6zs`ORktWZ2(qo&=ZT8ORt;5T^Cb-7>2yw>*_Dg9K_FlKlyd(I*
zhtmDExQQUrd?{Aw74jC<F(NA3Id)QSWZUD^)K?2T-3t=R3CqDZFN*)U7ZMU=R<f!F
zkizGZnGd6LE&xQQ3g&T0;lmvIL+Z+oIqiBrt)y$;M(DTDZoN@Kogw;7F<$)1Q_`JB
z)4gmzei-=j=|$ePSr>9PB+xL+VRlzl`mo34`oTK`_dV=Zfk)?ZFhG5IV)X~;kaZt@
zxL*rUUY|v^si6QUHMksrQddG~ok&4&169P9d8dXN6@k<s_rtpE_^IdiJ)*XVt3l;)
z`YawP11?XR2c@#@leKI=PU=Y+otQTaLvc03g2^mPSRoqmKGibrPk(cNA>p{rMogS|
zma$Gu-3O<J_jhYDirSQ`Q&o?&Il`{>=;cCoO|Z0v5Iq7`x9&cBCXwKIP*|{sYw&*$
zZl{(!S3*TKmPGmxIS?zR!|O=>Qn~aZ&7Erw2L?Syqc_qzBti6nTf3DWY(0{(6+#<*
z9d_cOVi_~iWv){85_8wYv~T7sK(&s*PTv;6sVG4O5zbv;@xzq0LRwrEt%mG77@Plr
zDwkRg^H+sD&K1)W=>OX3_CTZBl>4L&MO!qe9mlKg7(iJkmD6gv?L@yYgosG3u?nOy
z8L=G{?Ji47Fx--MZays*dK})X;|KyNs;%^nGzjgw9V+>ZuCJ)B)E~_CkHQ`3JEc48
zF7SojkZ_o4n%#NquuL5^n#(mkv#^-ZM%b|CD{Okf`ZC6f!{PVqZQsbvq4=E!+M@tc
z9aRi<M%_mJiqi&rv=}8~D$os=F-NMJyw6d}zL9JL3e4PFnLUa<*^E2Pq-#0?Vn$$b
ziYewfS{5KMkfLgp`MQWE_D#K&v;2LNLowB|$LT3Q5)X>4KDdAsq~><Be+|U9>Ca{Z
z#E~b)s?e)zZ!naV!LmIF{6%k^4?f3;7<zTb%S6>n)#T2`LqD1oJs;oz99+O__X(<w
z5f1<y{#Nd0^{y{A&foW_w+6XI&&9!LNzB!$w^!@-(c2$I0EM1#zaJn*WUx8o&<aJ3
zrFoFMqh~0pT^{FX8r@J1qyc5N6?~ooe;oockhlCW{9qpdyuCjj5j9c8scCZN%9F&I
zn}!OrhB4q9DWu)DX>&4fM?<rGW&cchR=-BuL{0;WbQ0}o$tOO=yF=ys*7Q|G*zx6+
zKKG5TS-l6biM0vD5*+Oyvnt(u#gVUfS&;yc2z){9$!F)L{s7$p40Rs&1C*C*3MfOG
z<<cKFKzV669g0v3DVW!mI#zC9M$4pDD7^xAv*x;Hn|m7cJh|gJU{kt_jzc<9Yr4!Q
za{t_`D_?bP7iQfyJlW$d#!#f%;v*fJM~~b%+B_-)-%a0071xyF>}m3soGW(SUx+~h
zNLALX0J0%x>+H$I1m$Dz#`0BV+@EyZy<?{z2lLN()(g;wh@Il&O0m-QYC;77tyzC;
zKyppL5IxxrUK6K107u~kl}mIERv)YbizC{#X%K*(Q%Bu)w$fbp3wW5VWCFNhVoPe9
zR+D|b(_A!v7c<h_{AUlqy?2vpPgX|9gMaIn?=Tcv!BLtDsZ&OF2*q{thFBJz<VASA
z&u4~SNA<BsQuopz$Q<@!<O5nA^*AI(nEy%h?b%mI7IGY$+qLEk7>3Nq)-c-(&mC9U
z+Boa$<dHG=#HJulownZvjK`#~xYRdWYllBT2EuTkmsYReG`eoiXX5*&Hv_BoLja1_
zpPZ4m6(**(QfS5OeM)y`>RNf>r1<n&V$tLz(M{pb!igUshtTo)i;4Fx4!Y4`vwkhJ
zFaRDA@;NYqn5b@9K%}Yt*?7O~k^P&M-Fg)csP3uS_b0ehF1*i89eLXAwNS!Ez1qbK
z7NVAP{s8e}Vzy<EbY*t8Rq~n)S3Mf5l9gy^ND_6Om_AzY)acx*(Otyj-!@?UhG}xP
zFLB3=(T6I4Oj~kdi?|RS7Z<@;cfkqgF4ir?ZN))WPvA&5M%}k%!2@asEdpHwf|U>W
z#Jaj+e3h9q&NtPYw`ap_Wvdc{0$^+%ArfO|*Nq)a(<UZVLC@nqj*(vOhccZ%f&P>m
zJ3-H3rtWq@IW)=H)*4h_VDby>>NPP44WaYPeirokD8;agMio!rgzrV4@`AmmrBm&@
z0O$y3-FZkC7+|#vH*eRU2?o!$t2U;``wFvZtFJ$a_x4M+n(W^>2!ECIt#6V~Ys$cF
z>t+5vl;X?F7@up|nRpa1G5r7)2=u=8>@7BKAD;Yq2}4Lie_1iuleD_oK^{vQj8wPq
z9DK#v3W^CWZf*t|;p*$4%*ExG>;a|iJ@~`mA*c+C?rfK}e|ltpuxKG|(vm9q{kkvN
znk6`p%?o;b+Rz;<=6fS16<&B^qfz9z(`CMEnp3ChkGoy&8IAy&`2b|e&K&TwxciOM
zy!XLhd`Os^I6+kIR-OEG$md-4O7*ME(SqcJB8d#&w+024b=cP_`Y3;{C&_9S$)9_?
zBagbr*cv`|l|{K2M(qM*tj^Kh9n*GfkTwKkI&xi?XSdP)KJP$na=)|93(o`WQ_s)%
z{t&@o#yMA1jmP(zV6>Z%DCOQxS|nVOd_1y##WA0t%j0Cd@LI_&1#CVaAI5#*{qxLu
ztQ2D90hPUtIN_|Dwhj=}et)8_3Tq-4#}%|Y%5D+dnCv}euE&xJ9=`6nWy4{@72!*)
zHhnPKojKhyirO5S^mQUEr?209UH<@h+@3=G9NgS`S{$<|mtwEdP&B@*Vka}|RMQ+G
zcOm6toh`L7@?n_!iI1Y8SIz#3L&K?uTtMEZx~EWoTi2pOVnrz);c)uN;vSOjI?}Ne
z@t8TO(*EjH$5`4X%lK^i+=ke9(22=-Kk#*0dD&N2-0|;76gk^sB`_-TTT{cRU>C35
z3Cif~k;EI;5q0`qjndWO$Ih~ikN3lkwb&+M0B4sTJ($=%Hr}@J+<H0XU<AsTUMbkX
zP2wmCL_ivD@GYR5UT(-xuvth@*DR*141f1%)}O7^*L~lDjk+H>Zw>QH989h6s_J`k
z=eWBn_$()rs0dEAS(Hk&raFg(oREATav3%uA0a^{V2a!XM$s{AaDku`1ram4>$l^-
z-5SL7rvRFy^pcW17~x`iSMGcCsx{0T?Phm{1}&P#&@b-qVEn~rKlO5Fi!`Rj4W9Eq
z6`NpN^A)4l*_Xyyx)Y1*oAMZ%u=02t9RK)b5g*jwBYPe_j~152_ieE-&H&H2&%X^l
z*mea7!!G^2KKf~p;L_#XlSuA>ij%x%ABJR#nz?|j^7@-Q2OrL@Nh;&bMlXI}xCBn<
zQd(HO$iHiT<(~{%B263c@r%x*sP$1`CGu7FN?+AT#fBLTZ5y0D@3pfpSY97$;ckyP
zW&Sa>wfAn}`Hw9tm7jlr@KZlPjk*2huc_Mf@MfUZ_z@4V-RZ|JZvE#8_9jt_i{uiE
zmv<>|mB#Ad-;cW>t}o;G?3(EF*vcIUCotyLdVqwA4UV2<Km#^ik--?qcVP%4M)?~G
z;~%4pIuX)?I^!G9Y}zuU470wLN%#S}{~;)VBtgEQrxULDn)BMtlNHN`<E{N$N+rUs
z%BL65!HFHXEuJqpin%<n7aIR`8R*N9n}-;67@7TRRvUqRX#3EwefC)VVJAA;U8FgU
z&5Y37$%^(AUE{3Ub}^~pu;^CY@P^tT!G{%eLtoM~q+RGN#UAkkWJAG(`m}g=WTv6f
zg1<TB42fIJIz>KIA87Eh$X_UTs!ujvXre(06-3qI#D(H{1$Ub=+2K89alyc8bn|}7
z_YlBky{QKtI<x_`Cz4oAOJ{^mm@`fR%lca^g-v*8)G$UMv3**QgbvH?0C0<OK2qvh
z@a-1u14ZKf8eIV%#ZT`BEI)nx$jy;(^~DMF85-|_3z4o!v*bifEF7435+taM>SD`8
zE<|)|$<n*FA^xNIje@<Woz0gcp0A0)jWe%9A)-SoyJeRq*F$1I`R|&fVdxDu=w11a
z1I#~vgS7Nn?bTpNr(vkSDwt=)x;^9q&>x>zo6Y+mw_RX&VpZ;OvP~4GIb1UPXju-w
z38lBMQ<s7>!mE2zVY9ETuu)WKciO(;b9b2v!8XSAV8yXEQ0(2`Bt`)xF`bEUkY~>5
zYwHvkej|(?fS*!zz*TKR5A4BX2f{I^McE?MOcj)&Q}EMGVwDaHuyXn&S*}@hxVo}%
ztU9EM25l7x-V)ft%3&K)?<R{xSmbC~%p1QLE>92fS6HzXsyy&pVjOFZYo)ZmA;JRL
zBPx7t%~m=I(usEuOukHtm3z22#A<k}NxNUa^yb~M4aC_3>G4&Dx8d41RMA#bO`{p3
zWpFTC*yLZBc{wN^h^t{ti27jxSO76R)3R0+oCkS^B3*8z7>s_4;VEMEzgqN3=#Gb0
z!{kOe=)S$6?u8=&BJ4EEO*%1}Seu@JG?<C&*Y679OC(L_l-ZL|t0r|3a*WSBnx0B=
zteohJ(+y09gc(81rfWsY#@swzSX6(InxW=aXP8DnPT|rOLvq!E<9=Eb+F$AVKR~zP
z(tyAapyRcniEIE|%(34>hdZotqRL$Cct?Css4J8rBByAOuA~*C7}<-i4mAzF^@$%}
zyyf6y=4a-w005*6f)E%8M#7%J!7Q#TLFConP@JyO3`N0Xh1MAUtnW}!jsmw*<hsr3
zUTk}EY!d@3&(MN;>OwvKp+CQ1_v`-PPXqeXve<JXGEQA3O~VXmUX6Yk1xq(@akVjZ
zWe}1gRlsH6FPqh?=k!?9m7aN>FCEKot2d6%*0&7rc+;pwAw#*Wt909<36-3XVdePg
zj1LJj8lGHmUtN1O2aR^=1b&NitmL|H{;>KVAW3K1qPF6Dz%`qq`X0XjXXhN|Qt74k
zf8mTbtdPI1Z*Vsmy3gVYrAep%03}O%#e#nQ#4lTXi~Th({?8ux|7%;YvoG`B$jSLa
z;LDtobC(@|b0?Z$Gst&QUxEL`X~n=HBZML2HU_eD1`7mn4K8m^2}w~UvS$vBRwi1y
z$Ca|>?~Zu5uc_Cg<oG70r|U~)zM4vcI6eP-3+ATn&aCsn#%?GVHjd?%5(|19d74Pm
z+Pb`t%)L)otr5giL+PDH$skF-`38>(7(*ADi2JSy<VLpcnMy`5cxbo|UcxE7x+}Sb
z*AhdSIS5HFaS3G2e5xw%Mme6N>%OKRk>GlTLwl3s{?lV(42c~xx(7p3*Gd8A-%s<J
z+{AC(%x@UvMBe$_ELpX9L$1L4+Nf`joivM@_4PG}6sPI6HZc!5B|=tVaA{Mdg@bvo
zFo+?6z<1cGQ+#vEDi`0;GPRn((nBbc+=%k|TTHAG$p%st9XI@Xn)55fG%u@2>&~-o
z3JFo>-}^wrbFO~zg`P!3k-3TO@&4xZL4bw}bSh^3i}aCG(E|4T4+g{d4Bm4TwR4;Z
za@(0c{1S+oQA_39>}kE}2WZnU>cM;YSlJPEPY-93yg(0Sn0&l}oUH<h3hZPwpggOl
zpgqNIL*4v!k3^dU)gq_83>pI-A>aAQ4Vi_AE$%10#Dc92`4PR-$g7lIV#f`HA*Wp;
z?Nx>{KU+u1r%%mY1$~bnU*^{1zI$roebXlWI{F8wuqm*zu68)CjPJwtp$!j?CV(p-
z>K~?PB2&ngz7Q50#DNBex>H_gIYqO-Ggdh2yZT^*tQM4kc7O0~uXpZyUWSN?bJXKo
zCzI~L(He&-NHeSig*RF~tGx2gSf2%;G&M`MRM&L8vQc2aBYVe0*E#&^nS=gL8=bsp
zqa!Z^?5{)K=gt7Y%JYqTA~z%n#OEgIGuxYMLD}3?|29A{K<E;Nrf|A!79}`c5=R#d
zDqGQJaK?Ru#muOHHAKbo-b^lF?@3&h6nYfmomh^vNVlK`QKp-3lda2x{We?E%$HQq
zvnz?GET^Zy+s+u6P7Z9cIX^w7{Q}#WQ>;-}AMUym3xH4<x5vC&;NdjAilLvyJzPa`
zArkLXP*KVTLFOM0BvOke`hxsDt8;rJmE>0IZrevZc*-Ps?o<oh^ZZ=%!?sCs>JdQr
ztc;&I*Nx_MU^~VTMM_b5s>y69i{>22A8CP2Rz>u8RtA8uYH7uAp9jCiHz&~3UadQ?
z9nV&6(uic{NX9w3M*WEVngp%H<8+o^J)pPVUEKXfeJC%d)l5Ztt_#m1)25o?V9+i7
zt;qG#hMsW{dK0#Vd->4AUbR+WWxrJU2dK!pe_YtfOe>MxFKstk*RR;NGKd|L<NU&q
zm&Ny_k`f$kUL0|{ZvUiAbiUi@7HtP2gb0Juf@ht+ARqls5Mqo{MvICD!;y|CbW6ps
zo2Rx9@54-j%z8}0adH3gP4`n`tu?dp*48xdGUVN;_KjCt*N`5C<14BYRXAVAZ%eax
z>f5K>^im8tGt7<!3|P;ry*lp11-JbHs{9?h9<;wqZb*R@0bw6%<<r5PSlSTXv0_E|
znEU#A$VD3J-e=y<peG)pRKHTCm8^^Kw3?I-5dyz4@^B3!-J)w6=6$ZxiI&wJWkeIE
z9GHKgR!khvtX__WBlTyqrX>%|9a=>)+58B~t~;0PPwDfkp80tG<x5$O_*7y#8JO=!
z^!F!?o(E*=R|0DAKR|u^f$|s^bS^^-k#0JIbA+-kr&&csBX5>2KS1~jER%hsOk$jI
zHM;6}O_q7ZrsI3J)S{(B&6+o@sz;E5rHGvsL#lI}81w{<*uEp6*|HwJ?*3RmsHQmU
z)jX%3K`(B?5p>_->xqq>$K^=}+6G&G^vAfJIa-Qv;3@sG9j+^;Vq53%qJ?xLN^YDw
z#I@A5F?M;!Dk8h?o8M0Tm;R?a7&-e$v-%T`7UN;Refr~;w#+Rv(S>+n@}~_%O&ftL
zy;C)jY{8IRxw0(q&`!r)bohDT>$I{P&eBIN{E=_o_D)mkMKfYBwCW8dC!55kH*7ni
zlpSP5(GIn*i@oVYmao<F#a*7W5BqegSJY)RwCe1=@$SsY{MV+|lC`hR8OMN^cXUQY
z!S|`u?L;78?fU^@UITxlU8h88(az6QlMGK!L3Q&q{5`#0VL>9gJ4$I97ehtfmh_rA
zO$d5}o?1|CJCUOYnr;H?fG#TsR+ScWI*bJ5FZc>oZ=^@2&E9r-0&WkS7mVp2DGcoq
zT4IW3b!XCHZAQ~2u6B~FM=;coTH<jWNjY+u!F7CeO~`a;AV(}I<$?320dA7YI}x5?
z>h#+^*6>`2tq*ix57<C{vt4(dLl(XSFUTKCd)S#5XQ9=CQJTnr3PgTTZp41ySGood
z_Gm5Dt21u}kY0^OKdIafR=1RHX<B(&V{rD)T+z6XE#{S(%!eoE%Gcurwoq)0vv!AN
z6?87&!*Zd#w={IO`3(n7gR`wNyD%ebez6GrfQ#*@`kX~y^%D=6#b#j0<^8=a3<{8h
z9;A5CaV0^)SE)y4s$!_Z-62d=ry}3d+od$Qji70;wB?ycsdI^akFGd8{_}K}-saS3
zVjN)RE%|r`Pgr}1co63E7=y4s-#ejsw`Idd=UsoQSciqw?@hWB*H>(y?TOqgl?PV{
zu>G+D?{cWdqtYpB0?o0osWEL_$le`JSSE<7=U)`4NaPu})S&n7L6j!w79*U7=UYi0
zyoLMLbX9zvhC~XLbF-6EGYEv)a0-x!*L=Awftqf;2z_;q?X4coX(kPy-VcDJc8FL)
z(H*%GX)I&CcjpIa61|)I=2V`^<!2y1kYB|sq-Bu};{AHN@wEHcs!ZO9MBT;|Jk&h-
z8qrVXb*&=nktECuCjLE-jvI+F*z^6faLKz=$ryRK7-b>KH?FT~a<5LXNjZms@D(ga
zUBRS<^b3MN{&`pGrRkC_FR)Me&a4bhh<2f8JkCxdnlyyCznHjjt)vU|2IRs<SY-UY
zbw#i7^D}*w`ZyDM{B_QD3yOuoYX!z?ncF1j*FqV``u6qL*Ak0e@?$?0<o?|p)q=V)
zAO$JU`UCV;6StL2ZoJ8U7(94TMxXeP2ZNEY;;4@`WWxjIL4Z8Vz=7^WGb}`U(HyI2
zf_5KTYQOkKrwPW?F5Jz&)A-@*!1FhA+NSDGb$|VqJ<?^Y*$}{=M>cl;0EJcL96%F$
z&S-PvCd70m94w1;F!Gty4hP>v$khur=Vm(gY^>5c+m>~^y?c}szjBch>d1-bXrY$M
zrM?i}Mg2E+J!~BpZlBOeDhlU|wzRpEm4P|^;fMs<E~FEDHi6{s!_JLpYQ?9}LWu2F
z*#lt#aKYvJr&`r_oi}BtPW~oGRJ<q%dV1n5;TUZiUeG>!-V`e`MFV8ZutJ{~M(fVY
zTMdg2H|swVvwRHht>1amwphiSA;7T=_G9(inP6Y#FMtBPiIe-x(dsOxmbg~NHW3LY
zM6H|n?qn3-X8rR1C8w6v^*cI864R@c0?RSOlwI8YU^QsonQ7~mbaWHLc*}3uI~`h`
z8~S8Toi~6aYY6Oe<p{%j!Muhe&1$~vuXo%nGx;Mr9*R9{y85kZ_0xSApx^W^@ejK6
z%%PypaI+@)V_iJ2%?Ik~Y;ryJk2)BvtgV9P&78J$;TFF`e3*ZKpmJ_^Hm7nE9CdHB
z`NF%!VZ<MhdeP{;9gfw_;@d})XPs!fqHq*Jz7NlbhYCH$Qjq(_c_g$T;v(mY9!JY)
zL`jzS`ROU64&&yYyopbf4)L*{qC388U2<z#^MH)@ZO$8v!+Z`Dq%GxgDYBcp4(SE@
zZ!aUFc8mdjrJ0AMTu#JbSBNNERV$f!Wc#L`l>0r(2Zwmx9*$0S;bn3R^K8E_Q0N1n
zTFgBO@cz4HYL4Z31N;b%+fCaoN&KSOFQqI(^IyC=5-AhB(#7BaXi$MfVXW_=b~gqq
z`B`Rcw&TI%5@u}Qe8;vjEz9$qmqL)Ycuy+#OO>TlH33<-jUro2Auep31H(oi@LPaV
zTt5D8Zih~0p;3*^*IQ@%Wghc8%&B(23`QSx?gf&$ljm(s8D`MJ8z}h{Ib(Mlv{^RR
zwQjEm(6Y93hETb7v?-{LmLRfF<y$waZ@l1ZL+~wGO}J#lgG(2me*7F66Wbt~+!^(*
z*e&<HQ#j#biXDX*N4K6aa9Yt0171bculqkqENwk?f)o%wwo#gT_L1XrQ4r#$bQ8^$
zdL1ivJ3y3X4y6pfsMZ(WX|q=MH`XaEe9;eA8oD+Unzt)2_|n0-&4+y=uB#D~;A63d
z<gU_vvTG;x9?q;A*|X5h!<#mATW+`Ea->p#>8`#&&4SC5?v(WLz7)9Np4&8_<0R=j
zzjmg%Wt%{Zz%yP{&{GplfHIU5@w5w$r}-xJjIwRWE_4W~1{6X+K_+;AfZTLf>FpQO
zoRjXozh{|X(Ua&#s`>%?T{yzrPMnreE_S2nHHt$9+xd2>KK)E5jD4N+q0A}KsRW1T
z*ZUN|<)$-pFJ1}xY$Z``UkLDE=!_vBOhm@<MzabyJe5hSf3VP(Y^+rMyxJ0^DbvwK
z_p}O;qm2S?;v`)PLkmwL!mz@P`ckr5Ngnf4HkBjQ+9wlKF5kcR(G$-<lXK~%+_o*A
zadaBrPn^$vBt0$H2%#U74(=sN`>R*2><Cikt1GSD0v8h1UD5ikq?JL|@pfN{Y*v+e
z|0_zICuT!?ubh*1P#C=6?5AMM8xE9Aq0BUh8AYYQu?&_x2q<T;^^<~*<skI>VDf|j
zPmWJH5qjtYiN^Divvqab*cR@EFuG|WB1}7yKsG#WW2NDVQdd3Qb2N#?wJ1I*`@F6q
zlTer96C(r9Tn?0*tbRLzZ*(Ag=6v8_=Y%OEk$dcq3KoeMRz$lKoQ}!T?JVw-igFU-
z4ySDIIDE#d%>@z->8DnHfIL$7m)9;WVGJ2##Ql44AUO~0`0WgK3OHq^v{{<$!HsaX
zrevpfWjT^%!Sf?TLsuoV^yJh@SpvV2S{kS`_1cB`mMqeG=tI(a1YGIvSRMKdAIsI{
zkc+sWzMk<{)y_%pZsc&^dHWpn#ypuWdYC7#40khN@7Oo1n};IMRzpgMe%3IT@BWih
zHnnB}x<@0%POz`5zkhC&cm=EJhRE6p%W8%2s++%oS+Pxp@D9--kIFjBCJP_vdU|_%
zE1mr9QdSWEdHV$AjaI^Y_ltU5?hjJ44uE|ADluVad3$UrzB_GY&xalWxv5_NW$b2g
zEE!;q0kfZT6R1k{&RZt%P>-j*=<_dt-8phYL*)lZ!f0oskE`I-H&8`*=ySd!cI<CL
zM(mVFo4xb9y=N-D^qm{21}QE#?x`5woO#z=AA8cjrN=~}pcma}F`BEU;iTSe9iY0)
zCNYW+3SnbJjlMKbZ#}r}lG&M%<V~6uwWR913CP%8xZ;}h=X<Mh{aiWSyI;B#I~|zL
z)^G_@JADGvJ0Yi+fV?!E{uBcbARROX1!Y}_7l^M#v}$5>MrPKM3gd5Gn~U`9S9$Fb
zl1nk42RK(cwlUkobUY!UK6azhc_(_PS)mwEDBX#ua;b5FJ4;kLd!BDj{%SeZBm%u2
z>#X5kBu4o7>7%HZ2~%bdbfQ3eBrz)e7R5hK{XRn^_}1b|v9(ocOK3k?SI%_)L;M3?
z#W2B-<_La<7{S^aWewdp3{n8-#Br5P1rv+t%aB^F#DM%oa{-k<(6D_%wd@oJ&AnF|
zu}%)s-dO7(Acxw8lWAHKHk`d2^`uu$hlNU6c_f!wc{W5;M`Kj<BU~~I3twe^+IU4#
zf?(+yTL)DJel8g<>Up#PWZ2i}u;<Tcj0u`VLGkuuivT;H`N2I~S|4dJ!WYP^GcV&q
zq(+o3ZTD;68ydsIE1RxfHazI7c#@QC5PiPz;%SpJvCG@SiJAAlxKL<N&9P=*a7mG7
z)#XJgcktjNxx40^f*E;yUV?5?uBW@&p1O`4POlGd;r1X6JC9Zl;SAW=O5@ENv3mt3
zv{thCVGLbrwhw*wJhRP+Y^^Ge%Ax`1$LMGExvWV@%fL@PzAkpqQx`ZJ8`cgvN21_k
z>L#aE$&TG!iA=6wJ}h22SZ~~CGDB&6bwVZKXl57RBY7pK$08~7hhwXM3Y5uN>Tw3C
zVCVd4XY;?MSP|)84h>66(Gtyvc_J8VK6o<=xnJKb^Ur1c4TC~6t@*Z&k3Hu*CaX*p
zB)X&ouJ+z}VSh~L+$Lirt)8JiI<(c4mCK|-g3CjPx0`fk{acbZ0v<;y8LEAfcp#!P
z(ouLV#S@fwd4P_?LpNak#hAMinec!RS^2{Th*nN_i?N*vvG#s^i3?b)8DD!osNJe6
zG;PqAbU^z(bN{y2yqcl(_f7q?IX=(2njA=fuwl4wc6VC<Q4Cw<PSFhU?2^?;W3np`
zR-myiY!#nwDEzAYXq-ZQ(!EEoKT3YhJEtaP#^iz7M$vqCt4JX@JgRYGYHL}#qs}lM
zFR)WKbD|cp-VE`EJiq@3nN9d%D3|wfrOB+^i@mim1Cs$qI=*pe@r7!7LwUg$-RdDi
zXmwS|c%&~Wo|D+_XBDo(I7!Ke$!zuj{TE8z2P)JT@W?A~zB98u-_*|sDiD2HRBPVe
zFCOm=*VE;R!GndRXnvYBwV~Pk4NY#)${SUQ2*sN=*J0}tl;c+`){-v_mwHSO^eBlJ
z7)PKKbPSqcZD8q;Q9Ct!U348g(-Kx@SY1^mB9CaLT#1eC(i%a>H>4SRU%lc8o$e@q
zCh@@WxQs$M*pVA=IJj+a)LhA?QGT>&Ki7T8?qNRNYgl&nJ*MeBzGV3EhzNzwki_a6
zURV|%<2oD4f5c;R@a5n+JM&jM%pS_sB;mN2*(?Z;%@UXoFLeI5wRJ;F37+bBz4-~t
z$7yO$1`iiMxFT!8kLgp_Or^Ettc%cd4A{DDkNHZ8VGiyaW8v$1Df)Tyv4cB^b`zQ*
ztbADjA5Ef$<ncx+>;66}Q=9jQt&&kC`^xGt+j6H-ru<;XndKcFow?7ARdXpTSJE~Z
zg;eue2pmDgIY1uq&Kt9*B&rSo)!Vyj^vW#S^z%%tc93ri){R)MkH6jHML2}?{s5(6
zGjozOJ%!JJFHh1TBT8^NHfBqz8LoV9(6wOEl_@(v!lYkXeWR08#N#<~;H2UuBi)K%
z@7d|JEV*2GG0m_OEHdJ*)G_8>J1KbSO~Y`-*HlY6(Z)2+3QGrL&b1e>h$lQPH>H8(
zi!DkUo$|f&FeEoiXKP=&lWuTeig3RDY3zn%w7NmDHfhC!+r};MW~~H)c<)v3vGxq7
z_=F)Ag?N*XxA#onO+ErD`!2Fho1M95l*W-KtK8}*c&cH=u17RO3A*xW)|aYS)UvXZ
z5;Jl?y=wbINL=!hxLd{OFmUDL>GP{AA?@|)jC#Wj*FA}*g2_^}NDXbZh!(o8feC`1
zp}|an4u)5>z)kHO!snjzpR3arHJ`Q?X`G3c`D68)q~~+@E~?ETZuQX{L{_W$K6*z!
zu36UJlLQ;&Txxu|llB<aWbN}%&017nWUo5oA<K=+Dz^oPGjBlmC;n#S2rW4SO&|_d
z!6^l;a?fIfYq4=9wB!Ztly-(#&hY&ar9){WJY&#Zrqa1Nfw^7|IrE)ws;i_gzFuK3
zw9HJJ%syeS_t~RL4`JA`U;GEWqWz5lS)WhVAQpPF*P4&A%c5{n<`1jeXKtB3d#ok3
z(NsiaFz=yT{IB8~i-72_osqca{B8&e2^of-Yko-S+J_6b@89i|Qx`<9Xjq&4!RYpO
zW`(mHOEyk@DEX;avL{)_;iST67f5a7iwU_(iIxIb<MF)0af|7?sv?gQSw;z~6Vk31
zEap(#HKJOahO(p4r^_A+2dek&kD`GKn#&n`>SA#e$R40+zJ+@?WRn|)6|qLtn2q`b
z)V65c!uxz6<t*rrs;69T2e$glMxT4co>V+eMR?tg|KrGW%8qdDT2fHCO$URMYLF8K
zL+g?n=my^KCyjx2l|A*QWN!tnbhj8Dk!K%RP1cykiRoy%B7MTHs-uM~WgOMc&|_rg
z^k@cX=B{!V5<5$q!GcsKj+OT`_f9J{#~XF@2mR-&Cyzkx@~q9xZQwn;G%n8~N_OI!
z>PG3jj(IewQA4f~)sjMJa`cgn=z7&Qp#Pd{=S4(pz=KE)kyq`BSW~^tEjq9tt=o1D
z#}6-|+IH!=QwKsM%>dpgluOtw`?19k(a+d(h)L5p>xaq_g=Ra(J+dF}+N<e`e`)8O
z+lffKo77u<*HN$>e6W*F842{uB^%O8^J)xu?MBv=Qfo0s&#UMZG$s7L;GZj)cK1xr
zT)j=&qj#UT8iOI4sA0QFg;I?8**X+{dpBoNGe9o)K-cA}<;t*IBhb3puvHFI>!=xp
zlembt{Y5ohC^B9+#oov($!C3^Fq>8iCMF6<q!iO);ob(LNJ#0TE=+tNG2q;T=^IlG
z3HaU(u_Q6Y-b#UIdsBtyn+4$4fQ}SAU0Ak<=&^i&%+;=5QZi3KUZ<6ih53<Bsd7<e
zi&otcJ}bhfm+;jt`(G?G%v{|)&dhOI>npx<h*e0uGW%%HBqdjk%G`-Qy9AaW%flms
z<pFIzj_wM5c-|uT!cej>o103gQT;*Pkt0|2m5{k|b9z_*_?F#d8C=?p+nZe9I2rWT
zunxzKY8r&m6U!FSYCtn>zquVPf@*jV*SPt;*7ot(Sev{oZ%_Qzq-BN|NH#+}>HRoF
z<pzj3g62=Qt0Ws(cQTvNGN{mdAL^)b<s}-<q(mlhMlLG{q3`VFbe8-9!4u^lGG2wf
zp859BZd1R}a2>Pv5-$^w8x*{R`@Bz%rN}4bayJ{0pgbDDOVsG3T;j|DAJ8lQF=X&u
z<b0PU9ewGloN*c-T5`|tbvxexUsw{=np~8VMg6jHYm9M?J)9wI$oh5J>@r5+=7*LX
z>#W+OFyf>h^yFEaiy-Hl%O3f`lTa@r#c$nxB6dP`TynP-XF}A!x}+!F4J7H!M=Eg6
zS?lL%J?`@9x|k8P&^+PuZIuFhxER5pw{6nY%0*NcM7mGWV&eORJFHNemb+<LMQ57T
z_bzfBTJeX8*S%73^_YJQWJ8Lc<~F7%2yZ36WGsZ4!k^&K%Y)yt7TfyE_Zb2qaCwzc
zjCRA2ojDoEd1+9J%ip#i##b)yYek)Vu+q4beU{C!P$pXJp~{|_dB>CYS0b<LmH+m&
z-y7&d#ksJcParhNvbgZto8IQ5bhUgN8=I9t2Nkqyp^czfCcBd1<=PV0gge#Ra@^@p
z`ODltKD8Sv%voqOwlgn9BEK}>goWS({mM(n7gg#v5n&cJ3Bon5kveG)9N9!ZevKc~
zIwQJjCam`vm;}=gJA^qxq+Ics$Mfx)W3q4`T89hvo?-FCr@mBBy+i7xjP73bXC^DJ
zrVl%~nJ<U;X60PU<Qj8gG|!rx{a*mfpZ|M<xqg6Nzy}Y*J#br9R+lKb^e}hScTpf(
zDe@Y+#j!jQZe<u#=~1Ff5QIzYJ6l@^NWOgAQhiZ0-g5fgHPd4;Grz&?JHV2|NNG~M
ze#ecvcFQFJ-@<hK!`l1zjJhJGV?h=#-ekXVer~TkFI=lq5|B^7NzKVCnI3$Eg3Yw}
z1qGE`3S57bWFL__^fI&pP*3(cfj$b5!5b9WAs&*qOA^%`%Uv=*fzL^kWChsV=KGZe
z{lARnnAFFQdSp}y_@;c)J?a*=miTG_yWNOtJs_sTvhJMhJ$YD+pYCHgPnzM+5Ue}Z
zWwBC5H&jmaCcOyQDFgPKSYgg)=X_dsN2c@6SUjnZR;jd2339{AjVLWinAJt6huy=}
zWcXg)2@`2N-~Xq<+bjLUw21>74As8_#g)iK3uRenrwzB8@!SZ;qMv5^7)=$Yr7L9p
z-YRB!tmCy#$<Z~V3XALp53nh5TYnp0J!r7m3H3_>zsl++2!&oRhMOI6bWiv2E|stv
zz<>Y5Gh_IaDVFDsj<l%Mp74=xPdv0m3<PuqE0R92R+L-_`y93e&7LSK<e*uT<-?U)
zye+bY2FXej3CWq;Q(sQbHez5<1=h}&ufASk7VH?fzpfk?n3I8aJ435!hq57+$-(wU
zKwsDMAE5Vjhw?UImYLMbtrbK~sV4WZ&U$s`*fhd^;LX?4h+?4sKSRE2cTb$>kW1-p
zsv33xwI{8Mob=lU|Cx_P9KKbDRnfFNz{ia}OI!)84ZI~L*wJQVZ&H(HgK6CNGTFms
zMfKk!Ux%xij^C1f!QI^66>?>)kfuCiVBUc{qm;#$;cI@ThrIf*eQdu_<ab$to3?Or
zvTlAH(*vE@sdd+<tr{jJz9`5FgPE8w5dW=UYX1l50o5tBXj<P@7N3ne)hd>z;@Q&Q
ztF2Ug>6?o^D_M~gxzTh>TkKhdRtr*;?7>l)n;_Gy&x0o4Wa-ExhSPZ{)_}lkwcS8s
zW%ODqN!wyP9L|^G@H_kUfzeG1Rb%@Sx6FZZ(&3%it`K=k-@%y7^(eX7s19D^Hl-1`
zZW`g*Xj^xfhmiO!@!M@5-&_*dxY;<3Zq$Qh;TJZ|Y^s8}aa#%gG>H~?{leGL^a@&m
z{%4Es0Kfe2lLi6jpI;uf9XxySUgg;ntjVi{XW?eoefj!8A-BiC7bKEqf(3Q(bloyd
zm<L(@%u+pOVj-bQ{kpxFU;66VGX%Gx@@e(9tI?koHh<G|m@+!EAb^fT%QrKVp!_hf
z)QBun15Sbpe{~If9wO4nR9ydo#r2$*lYFg?h}K1df><bb(jP9ZHYttsz?=R=JvLYt
zD7t4Y7^RzgWWs|dZlaEC*<C+6R8abR#Syk%lg(xJ3kL$eNN=iqH!qJr;ac#+uE2q~
z*#kjoRT?U9V!7s4W0{KQjrW%2;Ufb2HA-G5ISpb9fCWeZq)VGT$m`LBzxB&zb@1RB
z!qk$;iYGo_l>_kiBL!c6`7$r+&ZjN%QBLfm?-Q;Dzv{<6#x3^^SIp9hRpohqTO#0<
z{jH6)U}>^pv=wjo_B6BRh`X~uRfBZrBWIZN%}>csCR@JVji2kv8um%7<9*u48oRhM
z51lZ)J7PYo8mYX_7O#imwV+9NU<6Dnu?1c)dVJZxemVvJa{F9Hz-iN$F<1VmsaB!v
zM<vL9r;E=5LIlWLc=d<H-IR3;9A17+c`|NKso(0_tV)HLzrFL*$tqopp^SXm>COph
zae2S>VAN@#2ZBDa2@ARJ+9ti~N&{-G7-|2S4ebdN^C-Fr>I~3h`fAOQ626~MP~OQ_
zkv#1t5D#$S-ou>9>bW=PnB>>7dF4{&+#eefXUv+>?WnWzxqm{ql&_MlVe8CR-No8R
zu`JZ=vBoiPT?PHf8!glMGM;A@`2DW7G|D8X*E=eE@7mxu0F@|zg1y0Jmm<UdyWU5<
zx@8f<eN(e;qzc;)t_eEKeVk*B$%dqgH7UM{*t_^*?j6|1S?mZE5*2a@w)%!{Rr?uf
z<VabKFNs<UnIsMJ+0roId4>stJVl<z4nK+5uE4nzOl7M1`lQ%isO5540xg4nz^k?o
zk(svH0lq{T!9vck8TdBE)kiVJc_Ve;z7i6a9GmHu*-6=>wx=Srd)Lx4=~f46jS#y=
zcEi9!E|gC|I-Q%g4UVJkMr^gf4p^*52z2&SL>0`@K%~Ld9+64IlMgvXo5LK|?r5p<
zusq+odC%))fwobr#c{!xWeX!hw|A5`ebZx<vxLWa$<1YzZV#4)grR4@y6zY5ChG>B
z^Q)QecJJ~c*z3eb9}QpIyDgRji=_HA&V46xY;RH1W1pb663-%>?++K$j+G=#ObC|x
z6gCJuPtWN#+vuPItGE;9PIbnJG!*p-{|%C(g!}l#*tKxi1_=89w4%$6?>KUA@WwXZ
z$FXk_g=t;~9pDR4c2}%o&AYm$5u~l><20&4VY`=o%jle<&EzD1#=FJI-Da25W70P3
zp&{aK6Y0;CEqo1m3|waf4TZalm#yXjv64#NCrXjSBCnr}iRo0a%KD1Axu0Z>lVaL=
z)@HxHIu6051KowTJ0a#VB}cP1QC=QDKybS8&f$4!QdVMgW?E_<t)I?IZW`pIohEjh
zsiZ9LFT}3fhvN0&5>kQ!Z*+?zr>0&W_nm(6ExkQm=<whQEybLw8iv&H4=P^@9C+DS
zTUKb}dZTl!n$?ed38N*yCQE|75m!tSSqG`~f6HTngg*R^rfL2y+ZndenBGXvh0-8%
zG_+daTFP6EO4l^Z&XVUj-ZL@A?yH6MFMm@!(kW7&_XFhk^32o&H-Y}h6!F*on6Vu%
zyZ)UZlN?j69NXs%o8NOY>I>?JQX=d>3(%|$#*n&XkBG*|2LD+ow<anNDKy-y?E?4h
zdu3Y3;d}hxUaL*OC*(`k_wMGx!PYF8DU?g}*7Tj(89PiBvA=k-o%}(-`qlRJir>vr
z3>&|mD~v7i$rlr>$)3B;saR)b*AJ2O=<MD?MfK@+C)MstFc!;Dt3AOC^PTmRv~qoF
zccjlChf6vX{zX~Zr?K|K1_7d|!f!S)bLHdrTgsKdMk<xk31JW6qpM}L2IkbfT?t%o
zXR~6+%I)u4mpftW6F!aa@qEAM$$eGvv9j$<{6W@y)h0kqy;e;|M;lvD(xUE;)P{e~
z*oFxT^DTo#KO}5^iq_f`jJuZHbxVC47QB2Pk+|y+f6iPjbP3hkgIYze+xf-DJk;F<
zCL;qCwt*W01Mc-jdL-(*2ytK6(^e>ofn7KpVZdz`Vj<teNx26nFpOjyb8W!*=mwyk
z3&}FO^B>Ny`BO<%GND<)m%M(u?~#2K6yZhF=kCN7lwgIfizYc3^&*m0r0ZtY6|b^B
z|0dso-s3^)0tJ{opb<^Cav8L@IV;&Fd<F(Uq1L0fv$cpFSjogztFD;BaT-b|?eiIL
z4(EsbIsTbih)1WGK);i8uM3z7*D#7^gSaF+63?`wMQL6u?e+D6b}xN~Jm7vwagpgh
z!~BvWe1ATwK0X}XsvZ=7wj#4;(&O;*4-o3AbOu4ux*T^XxiP0bfIWGbpk&DYl@{6l
zAg)cy`R$_j;&H>1G{s@fCKm}NJIe(&X<a$cL5@v8k3tH@!@!dQ?XnZ6_%sT~g<PCm
zQFE^?tF#-uZLKqFE&ZX2<wiy}!4>(Yf;7r}>@lBOOL8pM*;n4Mbvi`sQW4DOZ=<L~
zsXt8JI=lmWRR?!nukHItH$2u~;fJ+W5A6KpeZ`zGdGGeAu&@uWNuFWWB#-aayzOe0
zdTW3fiGV3R(<IXtC?Q55*&U0fipda^=SA0h8_4ohsGR5~7yq8$!78`(^pGE>54!27
zcmMu+6G~+bp!)xaWu_m4oY)@ZTHdf-gn=8rNiQ+DKPSW$ch7kou8%kGVeZdsU~Stk
za{O-pN}cNfGELJbiy&g5oEky)9^O?`R$q%VAEzEK@n-9Y6T^BlZxAFv;Wp*Zv`o@=
z^e&}0uN_!_+hX6u4UEw<HcuBFoN@JD0&`j{P>QQ7T(Cl6?J3hVTq_mEq0cIwArZ-&
z!O9t#v`1LJ&6HXnswSKOVfU!`0dme_KS{IV7h3K*xDz^F;q&ze$nV{<N<{I2`laf(
zZwu5eUw&h%e(78g{On#f!<D{{8XyJrWv>G}6GMl3({KlcN1B!(s@6NeN|CdjjjNTZ
z&v0LAOWs$9PK^2GTr<E-2<dWaC*Qu}%JA`e^QKB$`!GINYnTpn&q3QJ*ci=9{{K&V
z*B%Y^yY6+Bgj{kBMRJQtxfG&8nlz+D3?@>O2BU<*FhyuwL+&xTCHGurWH1wElv_lu
z<2o+IxD8=2rt|&nz5m(!>~;3pYn`>uS?m1%n7`h4&Ajt`-}m`G@8|hEpC`uf5bL5V
z(YtEPdve579;()(NpiYx)1LLA?(`SlU%ev9wbg6=tGu<gzqz28czuAW!>yjq9Yz`2
zvWMidntyXqOPK>HeLh67)MtB?2Si2#*#tBx#FvMuIDUwH7!5CAxYCbwPmsEc-Pn2M
ztMK5tR_4;F#e1*5i0n)J${LhHaqTzvV*I)_zl}*GgI0&7fNlW*qu+ikoy^|z3}E2>
zum-|;ES@aF%5lUq1+Z&S1cE!=yeF-(oA-{Sb<DX5zjt@dI#thVhM$Pqx93O)N^u=i
zyPQlhW{QwbgsnG^{&4t+FNbJSRrt&Sa*!w(JtHt(cYW7^8XK;t0=4<qR<sqw)P`7c
z24@=CGO!tbbLVAM%`Y6!DKFNoJAIlkflCuDi)`393;$9fOsGA<h_^kA$LK11$@|!|
zR@x<*qDjzmJF#d3p0Xxr`N3IByP*c*l29V0%;P6hgJ5;X2N`W<eE9g@QZB7o&D%Qd
zX@L$LU*zh&+L<3f!jO+++G!^+pS2S=5$$Ul@kg&znOKUjT~lFMaGf+iR%8o(KZ?*+
z0e`LEkZNa>*jS(^$fIT}CcB^QHBQFD_^Ud3P!S3}Re^?%5lxo8Rm#Ycs_pX%A5Pz~
zAkfcD-*&i%(croBOoes_m#>t~+0a?TlLb~e@Tc*A^ki2Ugi_rx@OslG?XCSe6Btob
z_@tV)!!<-S)ViC~j|kk!I`^^GRxKw<sPMkA3(v8^4|~Xpx{iD-Zn2OxpoU_ut)wNu
zE0@NQW9AWdl`Y~`Q#=%L$IHcqj0djkw2s54Th<~}+<T^19(Atn`11L3fUXPeo6fL-
zpiu~@De}Tbi1!ZE^jl9;BZDf#6KnNs(0H+L;piFr5a~v&xaRX#1HYF}7k`DZ#0pW!
z8$BQ(R}&z6XxKCxiVE8Cc&d~iMjzVm6L!@pbUcRVb({79>qvoErDv-BKl6!|oN55A
z6&Ur!^o>oCCR^NtcBf;WkwjUKyeE$XKy(F}{>bgq*`0_Pp5I)(+?&3eQb9~wt^Oo8
zft5VI#JYpyN8g%&;eq4&o6E!1*A6+vF4JjQ&3?Es|79e4u=&d4?>1nQ^E!H~-tJ@@
z@|2a!cu&mFl`<jp@qVBZP7V^#z){05j}|=6Y08>|1m>pyTs#(<`!E-4%oKlTq&e{5
zjJoVTsogahM(dV%O9VWJ6<XCH*WW~R#3GL-T}Hfz_mtYrv@SeYJ@Vv)v&)sJrB4m!
zUQ16ceS;7T@(-Y*QqO0$;<&2%Z2~S7;3DRhePmuRTlf3UcF1R?tx{#w!<L$LD_mRL
zi)JNl2Gy#5bA7Pa>d1;20>8V59n9Fc<{R?-gLBoI#fwaJrc7l1Fv36=U&G8yC@L+_
zo%0zw9BxthnD?U7HYn32UNHvj8}?D%?(*%feU=Q(n|AIy80*b7*)<k)rHi=$8t0?<
zC&iV_Wy(X0D3+{zj!=gAqO@U4*y5%97Ux$+f{9TY#qKgX`r%4X%;zIXDy(~x=M9t{
zSe<~Gx~gpIzb#Znre8od^*p-x^5n-v7t@&aF^{0GOM7)jj`kN?ni|&vZbFUQXDNsn
zq(7EPUKS9n+{w@a@t3_QW$R4}!b=}66z)`PTXy%&g7)WmR*SjCor-&j>$rID1o7=e
zyg5EPm(#*+wmh1eCo3uuofJK>Am6F$x!+>G{!*EKI}d8r*Nw&&{|e0u6fikgEt<B7
z8^Nq%P!59tm*NMeeRF7R{d^Z-MXVPprhBY$e`BCKe$#QIflzJNm65lad2UTYeBouw
zmz|clUB_MQm0?g8a$JCA1EBBuK;{8p2=sYNL)%x-^|(jGcnnA?#>X$+_wgI}o^z$d
zi1dKW&b*S99`-9kKf+cm8#rOw=+$>YErx@m(mhpV)-`r#`Oi_AZuHpvJd)G)U2?Vk
zbNc7>+6Tkz6kyS3{@vAOKP;L-507|^byST_1w%Tt)XCY^Q+_`vxG-WTwhjysJVa1X
ziuI@@Q#+QgHx2UMHU9=avv0EP%;yuTnrg3>FV8hz3eG?NoO+KT%Z<J*Q;sZU1$-_C
z0*%Ql;l_R5d{7fJa`k87tIV%)O{aJd&GkS+!mb+}g72>Ph={T3SEaF+ILIHR6FWi5
z99JOWGvDBZm>3g$XlIRANXjTc35|VPXJ~x@nhCyi*!84c8r^K@I_+7P^<!pIj|!{g
z&nq24J)@zD%sHLC-n`O~s(KsafLd@~Fx72T^<&?hhL^`r?6Jrxnw+Yr=k9*8<!!}w
zS|A2Us%7TPVEVIRuWi=6evF0VD$qfDwuMcKW<`s*Bgqr0+I9lr{gCo2x=z4F`cviU
z=qJ}}vG=|oO(5pg?j#2FE^jnBiUANn4su;5c+KJfz)dnV@TJLB$1X_ek0rz7GE()b
z11Dkv*CV9dp<CVh<Rsj(sl+v*n=uSFjZM?{y0McPH1dvB(Z7v>udmnSAoMs?oc#Rm
zFMg<11LWCxSU@xheudWg3f0qOpjgA)_Zh(1`S^mQg&56A#J1VwOu-aWA<RQ+fR`>~
zx6vbatgXu(AMD2I>#?;kCrb#ppxL3Avk2pufXq8@8hn<Utsd^5^8<YbJ+H=y<yR-Q
zQBME7u6Y4=@uTo;^E`MT+cM{j^TTyrO^pwVp7NfbzBrlE1?JoG*w>JUPnNweCZv^y
zyn9F=nUB96W2RbAdM`Dx#r8<137@d~!GgUR!vQ8ocukOuvQ2V(Dcprz(!-{cCR>62
z(+3GGLv{R9HMWAgU%{r*H<n?e=R@uHabeB39c^8l?c*J-EuZ6UW=4LD@dq7ZRa3Qf
zN5Wrf;k()5Xs-f<ypg*go&1Jc1orYUp+TzW7h6Q$T&j7^Jhq>doL#t-YF7B0OA`>m
zP+MHt99u@L>4T`S3Q{V)lZza_##^s<>ti@ohon*wc+b>~rjSPWn_7Y6eJS5we0+ND
z>4|5bvi1o6#8mR4kWCz!5b$@jR@NL5hBbXj+ZJf_oZUhIjnBbuU24T=iUdZk4a9?U
zJqJ!yJPR3BDj-Dr2Z>fsk&6;fPSFC9EFiyvIlJA7TZ<ihfanZ26hN!V)MU@ap)4$P
zpAtyQ%iE9q&9e_S@9nnbO$kp61I0x}iGYtajKtyIaXxFzl6Dc6DF!F2gOxOq!mbDU
z(#|v-<a6|Xedcn-f=P%+*RzSR^Vgk8ud)jWw7?}yMh|=N+7t6L7Vb7{tNO5i>uI^5
zkn4~S-DpKRL>H_N%AT{!CD@2=IPxq<6R`W<^u=QrrdOKUP82KLejs$s>UpSDoAA;v
zOJASu-&`3>jG%WugFe;c@vvaiRJ0)rMu%|IPgYMFim*yQQ*<-)^Xyz<@Sdk$`7~&*
z(C}4>Bl)Rcm#U72d3`x?_`vOUrfGwrAPUzMB)1&eA5C2f#`k{&YAq}vhzxU8zu_tZ
z(_wQiYyZrhw#I3(hi&fr>w_ae)yeF0YxP^RELG5_6+OdHN1f%*6af)LkT^3ueEt{g
zu;b9Ey(_UgTpgq?zfy9SppwjkIlP&~_2gSTuNSd^<vU?PS=vLuBPPf)EOTi&pa%iD
zmmYq5Ik2~ba&m44JfxkdbuMu|^}t@aDgML-k(&#%$RylUc=w1VTi9&bjrj$*?34Gh
zBJG&!HB9a%jx;Nz3}KV$OT$ABiXrs8v2qbZvS-^}Ob$Sg*T5glJ?d{DKle2}=&c%1
zR3;3j3%gIyKzxlob_a4-)ZjHJ2U>s!Qhh>R|4>RKM)Vs)^Oi+2a;~MUg<NN0)x*JC
zteP#F-An)P2PDGu&S&K1*O*3E;65}~iD?9nf2orPQaS0I&dLSYtSn<cd#7me@rJ>M
zzJISsEpS0ifEB5kTjGhBG|U^DkvUW>mJD~dxcvvd3F&6TWD)^__azYn`5qv9W#4W+
zqLv$AOKBy{Ty-{fXLf8nsRe6}y~I?0t){DDq-RND%*3+LDv*?IidF#l7mUyvzmw>W
zrn%~~`YwDk6(SPB8L2m4%+xnVxQ}5NZ9Q$y$G2;r{Gx4bqC4N{yKhVUB&jeHIkiHl
zD>l8|z23@oM&mRG=QT<3c<FVo_U*ePM-hqf=aW5$MO=RL>VuZ)tBnSynFi%fg+u9v
zhabHfvmXTC^)(T(7PKq2B)0C(Sk6svjjO6~|9aH-=`FzPw{Tv96{-<{B;3u@*%3-y
z|KzhL_3zU}rU2`3N(xAcIY}+;s3t{l1z^4~0Z?>6cl9^dm7#u9M=z$j?F9aNX8-Nv
zOGFso9Lc(fFUR3RI^aAOQZB!(3gwr2I5l}>RW>Z>1S+O~0!dY3OBA-y5j&?{K_`py
zP@_u}|C`hk+v-ESHCX|Xt}k|tUhV%XD4^EmG7-97SF%Z)inAvwj?e&9)R1AeLp4D=
z<*{!tjmZKA@{e?7v6{58Lr{Y_iMr){>{8v$Eqe>E>YJH*+mUxRj^K~8t6UE$@V=@I
z^U^}fxcZypCkG8=&{z7=mnQhMn;QF<fP5H0L@aTCm~ZzArtf>x?6}5GhCrH8MW~JS
zF=b2Zjn@GjNynwO&i-=MQAinx3b1S*+nj}u>jhwB$56ycOY(QnOchiLib#r`-$I|a
zj=O7Rf*jSA*Sq=^Y{FIe48QU@nsgT7UxZ><FB9rTi^dx!qqX}OH*pRhdy3Vnn!k<5
zrGm0ztO?bK*i)(54*Bhhp8WQZELI4`;9M74?7mLq04u<X6}B$g?<mOBTKkcjAFVQD
zO5pUwi=U{mIDZs-%V+e6i@;%(w{OxAfrcyhWxFxn+quIfM~aA*7~iwoJ7aBM$L@WD
ziRVWNjdlY4fMe@?OHlz6)0x%uj5C4$5Py23Exd}P;kv$|8$oO}^jZ(rxpt`JVA!bX
zS+pH1q3EIRM)=!wO5Er1)H=m|%?YI_GPRl{l(zn>F4)uy>2iHzw}!_U{HX=$YIFq+
zo@#5B1JEGX=6}&T^k@$zU9W-O;hxq!T_l?)8+^&GlcFnP>nw6PZT|b~3+KxAleyRV
z`pOGkvjc7mMz-DZQ$7^ea`XyF>y7To@9Gj4_u8be$aH5^v+T>Qm<*;|zXkBZIQ_v4
zTD#-FYSz8DfGHpMsdg|QSggcO-&5(3pvk|7rFYB>yDMJ-w0<_JFkmrk@q#kfBkI4o
zro9*~_^Wq&)!Zc14{ZpxsOUnkc0CI<;(o1ce6P;Q_PLvOmk&X$ziF$CG6M9JQ^yA6
zAsppCM6vcXCw$gwCTzq*6e?*Y^{o#3qJ~w1o29~UOp9xrceXI2s~J(q0{5AOxNKAc
zrJ4|r@@gVdr$?sQL!l#S!>~^N9B<@|s79RoUrRG;j`-htTv@$VmsFpay|iPyGITJA
z91_8vBFn1o_?n={whdV!P5ooi`uFous-GqEqv~V@!ZmAO+A%0Py1m6VATR%l6pxef
z5Xy@vZRrm-NGYG#Q_LUK)MC=9PeUNAx*C3fjfCe756qGD-aJx$wDsC&MUEA84pKH~
zr?@lx)%tQwyr}w>2g<FU;j0)NXl?3xnRqL#4cQC?bo4;q$yjWr49izKcD=XvQw_<-
z%#6>Vna)|6dOys$jVJ1tiaeoOh3<t*6sFlUgSZ(!x(Xvn6_wQDIZ&=0C{Cqioe~3<
z8MRW8lsMv)@T65F_~W?~dz#n*l6wtKEoaQ+)1#piU(1n5e<Ct!Ty)YR#4o&~sz&GF
z^309*t)*{q#|gTxF0Dx6UcQ<w7%vHEmaGmd`Gt_hh8j+s$4980wdkXBs5O>8)4BJD
zkM6u7trR;+J>IySzoci#7&Tj4`Gv4-v32mlynw=z*=G%sm`l_pKDPGvsuDe0cRL3o
zs~A`JQ@;ZFx2@8QUo<N`NqW`VT6jM;i3{CW{1YRrc_E;2EBWP(N5lsS`8<+y5MPt}
z?kN*v{-tRLBdLmghA3t3y?b6=1h7sJ<u}(}R3yE>u|BA2@=7W|(H0;jmR-zL>DF>k
z8;`d<iA#U1_TB3uc+QG%47ArP740>NKuV!6bl35WgRr773GXst2Etvdx39#xJnywi
zwT@?pb_SwCX-J7(oMKz~R5$pv>PJ`+`vL>jJyahyU%Z7$T;Hx+yFi3RkMhM%gpw)8
z3n4n8>iStndPz91Q(wEUe~{z#j_RmuZP0cbmI~g_2OcL!Wr{Hu@dmN)ItuL`EjO&^
z<TKI~785*|7Z2XQx;TN&y+~2Mzjyb=6F_0_s5Y5t5sEfs?wR1=^Jitmdy#dQvQTAB
zOj}1t;ee_1*UE><DiU-8e9-5eP1Iic!bPV(Qo?dsBT&SPX>1@canWeVJH{3Xmv50b
z+x0U|b(YX&q8|r7OKkD-TsnT*co%;mm%OW@mStsCCU8mFy5_T5g;Q00NGrrlmk=Eg
zwSe!xo=q}h;dUx@um&m&PgE0C(t9*=62w27=Ay0J&a@k9dXnaBzsp0aH39YU72U+M
z*jLH=lo`I`!A8S|%&BE+e}oo8Ae;z`ahqRrc))-X<Cn~P9(2onypPyCan?lGq3+TB
z)m_o|?@E}Cq-?hstOIoKO|6YH(^4n5PjaTl*cxlSk9C)ONJ@P0i`{jW#Se8qhc8W-
zrFRIjQx`lADW6dmYVK+;5&3eBcs1Nt{t0uHhU%L|i=B?84wbVE^xcvSms5bEh4vo4
z!vow8Bx9B|8`hVGcgi}3593DfdX%bigt?a&VkuHfT!`pm))}PCY!{G3DxgEZ$U7Kr
zOYU}9yOCi2YP0R&&RdU7i?M^$#^}7Fp_VP}3B-D}&L}RiucPcFXVLuvOtq)Ok2crI
z`!h&}rIfU;<3F19o*|>&2Mx+Eq(d{V?Ub~kIoA;)056qGMQ+z&@s=N5jL9(`RJs-O
z7z-Kj9wBL6o|8{05ChkA<I_u=4=jX5qwF*=kCD6%yFh8iA*fvhq{0DNdsqLGq?&g_
zbGq)+B!K)usichzwN_9y-OiN8<c8fma4aO4&lTyz!6~$+!J?~rAP0OP&_Z%&8kl2J
zsFeNI3Ftb;ZAg#Z<UzqXjcNd%lYHdcPYh!bGh#8*`xTYLx$GzbKsb?@Sp|UCF1xjj
z+qkuf=W?_B_ms8&m4@rCCiJ!6Tyby;Ze0NPo69(Dv{)7kTRFu_;4}cy&d&ts_kZ;j
z{};x0wX^<Cbs(n^v%*&c*e`m6lh|g$b%8vE>IS;xO8WsJnmZz=6m#pF9cOE3Mfx(s
zye9a(oPe90USg~0-m6i%&Gj0Vn0t>5f1RDfZOrum9FrT|X7&qZ&*#;Gj?EcpfIay0
z1K&YiZyL9gu14^{6D6*U&?4;(z|N<|xSp69mg8jsr7erUDQ$6@zheMREW^KQV&#-H
z<p2y<;Iyad+Nq}f!{D}*<sY(+R)3RqeD*&zrIl$D-cNOut?y31@m;g5>SS+ycPXU4
z{nP=KpOaa8`46aEG1|?;ZTJtrA3N`E3aDV-`l|{iH1?hj01`ER(dR2dNOpWB7o*B3
zx)qoB`pQc1A5)t6fmq{+8OjMOc)db4=``5ql!f>L<6r#6l$fN58S;N+vA6vHuV(O|
zfi3V)t4M2Hme~t|a4+8y*w{|I(J9aN4=Nr?EeRD0hTgP#w*nPPk*;vVz<domn#)_n
z=Il(%Z>|dufUVTY(e$gpblQ@afCC-P_4NCnV}IJ^@At%m|MU`whFU=2*XznNrdCR`
zP|}GN6Smjlc{<Nsn2TN42JSdU1_gf;VCH&SWuvtd{r)d9iI~5rB%1y?g+=|-E`M%@
zKM%>Do`IwBzvm7W{{hLY$(cpu3`E{fgW^WPsu1mw#cn-4;i_F5DZh3{aQz<q5Aln)
ASO5S3

literal 0
HcmV?d00001

diff --git a/docs/_static/bonsai-predictposes.jpg b/docs/_static/bonsai-predictposes.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..2e4f04a22020342a8f45bde98ab45eb6de607a1f
GIT binary patch
literal 79334
zcmeFZ2UJsQ*De|aMSAZdDov$HS1D1ECL0lyUZT>85J5_)Q52*jARr*1bfiRjCqk$q
zz1NV?q$iXRAmwJCbMF6td!Kv0^WSsE9rup=e~YY)u?8#ensdJMo#mPHU6d)xB7psY
zj=m0nhK2?JrTzgZvj8ms?U^%wey9f>^>g<8*|T(XXBp|~&z)yxWM*b!WMX1r<z#1J
z<zQuEV&`S&;Ns@t;bCUu<LBk(=j7(${&Nr-TIz4;&N7@m%fQXT#KQf5`=I;)aGXCA
zL5HEGxdJ%DK|{+yL+JpB001;+snq^S@P9jK&QNJQM}MAyk%@Xk4Ljfr4K3{%I@&*}
zrrsS){Xc+?<1FW8`Mc-1jGxh8@#I!`lkoMt=)H<I9+P3LnBsG<PzFX`K7Ii~aS6$*
z*RCrmtEk?*bzAGcwvMiz{)5L)OwG(KEUoMv9G#q9T-{!I`}q1n{R7^<3k!cA5gC=3
z^f5Ul^;24UPHtX)L19txx5}#On%cVhhQ@YeM`u@ePw&r>(XsJ~$*JiXG-hdeWp(ZM
z`UY-q|KJdRL^wYAlP?+o?SEzK-#GglUmR4v&d|}(($W9P7tI+z>PgE%clNUUInKMr
z^v^uGt|+`Y&wVf9YegG_sG<p$=egG~Bd?eeS{(N$YyX9_{}^MT|4W?x8)N^@*9?G}
zmWDccv>X5+;FK04*<r-~+fIH=`h^eMl?gdHhmb$qW4iM8o@C`s-l71QAf4F<JQP5$
z3xW`pi<=OB0OQ+?%(I=D+Mw^<D{QF&e@Z>_T=NEtIo9`}yBO%R-gmYN5aK%n#><Pp
zL6gjf9;Vv+2RjT;yL!!hSJ`h40Y5l00pwkN-UJ<7SW0R+VVwaDbT?4|?Xx&%(plJ^
zl^pp>Ji$6dngaL&CmU!}03GH@1Q^CIocPc~JDHe3%3`Zgoh_57nx`k?jc#%MiYBK^
zm?sP>N0tyBMm*GoFu_K5G7HDNTO@)Xo>vwUigwI2QE^P5iwd`h2B>yD*wggTme?l2
zY(MEPxXL0XH{xGW0AY2Q!4pObV7VGZ)E2?k5-ll!yC7mpGzAa=&sIXkr8{Ovs%z&H
zLZ3M)9dD=Haxe7|i@Ta%6K;FN`K|JP)ACD!pUiSWT()KnI5lL5@C2|5vb!9$a!+Yc
zIB%rw#RW9;{uO>S%j_@5{>Nq&7k*A^IFor%(ZJtQ6hPuI495-oM-eoCZIyt!`n=7{
zG!QJrzaG`>TsBw>{?fhrrpm+j7zRu`y1_7St09{z)w2aeyA!I&p@?G{J_>+Cmneha
zg$<cg0OC0m07fc4=d~K!Uj=IXw~1Z){e>0zPl%~Mxk~%}pUbFRGYasZjF5RU8==Vz
zXM{)K^HDR+@sfR{8HxGX{i8dBcPB?$R3?7ew2LY9<fZW2*oW#C)<7q>C;$pTe)rUm
z0+>nOZiEp1;irQXz*j+lGv0~<81#qXP}tm48Q5Agg0M?4zl`JRPJbU7WQiqDyKgTK
z9g?gJublYv{Q96bG1BMp^X?B8F*Vjj{{+t{Tp!y^Qb%KUC|MMQEL>+Ou|ioFzT?RG
zj0^<(L-8t{4^_`}u7>M~Gsec$cd2p&&ISY>j7~mLvxcpM;rZh&MtO+JNpfHx1rSG_
zX_09#1)#A)0d&LK>vy?`URdYQ<~vFnV6zmKd&-gJ%WsVuvL-t8JOvu-<5Z%?<WE~)
z73f}RfJcF<HPg_r?N~|V2+ReVtvjxiA#-nG>0u+-+M%hbDkl8%_0Cm(4`j}9niyC3
z`9+SSyqr1@R4l?+#>kA^2Rh!BBdd|QVUBR(Q7Q!>OaU-rccVqB4-1k-!b_W3UXk+s
zOQFfl>L(T0o!oC0v(+zq#X2<y*dP9Ur2Fn8AZSE?Pn;=tF?$BA1qlnGn{2j!IWcn-
zp=41L7%jv0pt*%z=ZSxk(#yOuX2$M|4tZC9TSxI$c>PBFqyRq5c>2O7%0gUor-1(h
zX*4iVC6q-s`T`D-Ro6Z<V>v%%X>hw!^lReL&r^vm)5{}9x7GF19zN8YnOgT2;dQl>
zXq<4bo`u&K<St}t9&t35`khhw@Z~w+g>-RZ-fMk<o=a*PuyWLD>)~9+?`+uS0qo^z
z38?beL2tLA_QTYH<ad10#zOJYB?_QC`&7#GffnP{oY()W=gsVB%`0Q-is(LYnaa}E
zXv<7Te+B5e@xYBAabL`n{%VL@M3xZ5C>7WtG5bXJ0Ri@i5W55xVIUNCiFjal@tV=F
z;PWGSqmue-Q444cqio1%wI2GWrm;{2e{R`(XD|@~+|x|^=avI5mWqiF--i1H8bq(l
ztPNu<YiPx*bEMpR;=YY+Pd7($G4iLZ0Twam2yJM+AYjp^`R$kq-rL>@y3o1_Ro>Ab
z4o@pmL-i)^@P9}0v@Ndzh8|2zYy|TC)ec_$8x{VtX*eyURc%HwNH<uf`nHQ%Qyh69
zP(R1)^j+t5sSoqqVMvZQN&22R1Eg7i2ENcIP&vOQ=0m4fs@daCbwjCl(Kjj?Zv6T|
zTSuP-y-DBNUG`Vm{oRR${ssz4s~cjXGbMN$M>PAtX+0Adt1hXZIr>$lD#U$Tcex%J
zOhR4vzW6+SLkSUK#adIWYwsGct5k58F5X@G<sA1ZM!@o(K@8WP{36$o;L`&QgMVVQ
zo3oKXE_Jmf7X2-yriPUUi;$5lpO08E><-gv``(q(-b788KDJM<ddhR6Z)^-t{l)$K
zop=%LOaF2r|8lGU3N-#bD*R6v@%;ZQ2GnfLiMmR=CK3(XdsG8Db)*1t5}S39=<2;7
zO&loPhbb@q)MM?5r$({NO1F<vY2gRWr<KfcX><;8;B{aGh=r(+VdHn>I|=zTsgigc
zC}#P{nWIRZw`W;!4TgS5=qA5TgrTM=0Kzw0;!hK|iQl`&6hPX3*(4IkPFUI&?UXBU
zedaCH!bqEIB-RD0K93xjU5a;7&+pKJE)!S3OtcCSs=87H6DD94FXSF0Q)hUh){hi=
zFATmcVjf9}{KJh|KVnQ%m8*Wyd<2Ls46>&H?&t$|w_rq5k-uI73BcIhW5l}D)m8)R
z0@V9sx%H~mSOaujOaVm5RVDYHpk$QEO^4R>9Z2<z<Wx@WhX^`r2cRg=#pk|aS;>%&
zY-V*W1wxhXh{LGZ5Ybdn*WWK#Df*Q?IWA?uK1Qcc#YFiMUp$y=%zQZeuaEv4C2ACa
z-95YQ#PvEfS$Cihi?E5ud=9x~>5UBJDo;r%OJq&tOLW*f3eNpSl~&W6D)kh&S@|HD
zdp8F5n?aI9il+dga{qdnKZ2S(oE(p$eji<M<e8>A;!IpTiAIb9KvYlwxF@i`T{ACT
z{hti^g;gPn8m|9!8i9Yk1pI51|Ek`9m&+#5nF2VDhs`3$V{rFT*it`i0Y*AMnH(g7
zjf10gCiqONRNPgQzAKh+#UB4QDuSE-+f76COt@3@P8m)S9a5MF_1&8_#N9W!3(G~>
z@mZAZUP^Y1{QO0l&&&NL9Xg&q-;<j1R+)BaDiQAb<hk+IE?mGS$M_1ieEzoJeqwzI
zeC8c&@K&1_{}RjGWXtgV;O>6mo$P)I_z`fUM|@W|dE2S=;Y<kqYvMhE)RM;YO^jLW
zm)<Xk@DS1Ku@5N#lZG@o?;Lm0s*gy<_H{32W2vjIm#3CVXk)@A-gAuj>}!zW@XTuG
zR*)%1n9n%Op&^am%1Sn?Dle|2!Qq*!?0g<+xli_?W@t`WeyhQ9x@%Br)T-+|a1*{9
z4;UiT5HgS>P~hF#mXT#F<So&NRGj0V_)``$YSpzNQh9spdOvp^W6^A@%j@~@KU$sN
zw9*G@5e#w(UPi1`EtfBIL$<4E@{K<mhm>Ecd$XPH6t_A*v7nRu%JY&-n*DqY>)~o^
zVos7xZ|5OuP*<dPx8-n7Znh7bs&JZZ*y#RsZ=mXwZ?Fi~el15ALkDBDfSM{98j7)2
z&D&X94MH4WM*sX))Kc%I_+hv$W_xLxe2&0pK?uaIc0a_4AGQ`^q9Y5(e4=L+RvF-u
z$@w3<o6k)#xj(woUbJWq)3#1>ucKl+Dii=*4}V|7JtTVtvN?X_DD!ydE(Kr<#G8Ff
z2Ae*uXt#K+K2S)S_y%-|CI?884zBCyYwO}85NkOw)V=s>Dhi~*MLyxC5$6$^xk3SC
zMN$C4jGDuI!~>(#-;d}%K7swWA6`AZ5859F?tZ<)X>3f&kvsmKeKeH2%>f&{PXV-s
z0sfN>d(j_<B5Ml)Psm0jKJlQi#q7|Nb76t761S|2LlAW>l?wl&&TnNOAumg0kiC4n
zmbBrfju$S<zt)3)Z&jHY3`S=#1zLzV)+dtgY0){(S8(So(j@(m#KHwBfOo(SkpsXu
z40{elmWa(n`t*@}C;-Lo8hF@ATxpidH9s(OV{M83xAm>rG}G<e+L<a@dKU2pdVynW
z4QRO$#{7Mdw#fTK6qnnuYf5_m;lz7K$TP=R;+@T4)|{Bqy=u=YUe^rZr5OON@&SZ+
znM_TAznm+f0A4Flt-2DdT?E5+0&qhRXyqsXu6rX3W<;42Db}?xXf#?Afjc+Z;g`NO
zaa*&b$l7AoYG&Q<RUh}g33NF}R&a%$LDL7{>mLE^`m4p$M1k?ukwd?9CN-9a_)UXn
ztKgIH(x43Mn#{LGuLea!#w^Cp_}e$FoY0fmDFAlZ5UnW%@IwGbNC%?XaP%aAGKfe+
zMJpl;`TD-sp-#x9(EK%|u-6(!sj(;+cq)82rJP9txO}jA=Dk+!f<KR6^PONpd@!P=
zy@_HAIk%z7F`>l2^G?+&B-BAAbszfnaf82p?9x5y%no+p^1CxO_Loilf)Mxq#<K2z
zCzBMa6LdUcHSB|J!uHk8(QsrYtUcs<;)bzfu8V;SHdD@ET(-QdTtYVQkCgpO2Fj0w
zI65qX2x8lGCz({t=dU9D@5i5Q2^+7v5;gg^*^eh^5veQs4vRSj0H8)v7+0DE?f<}V
z|GMk{ROWw4;$M>ZyMg<cPW)%r3FA2^$hsy}zhKzs>a$pb=0wW}-@cvG7lQ*B0Vm~A
zSHG~rFwPxnh%f}JoDT`8|K1uQ7#17b65Efij?$O)_)=@K7qNWxn-jC1&P}`Iv|OUo
zq!LI-goUV!B}bzrQ*|3BWeyGHK*9GqtgY1zJlA@a%PQU#W(qkJn=l-|;V)`cgR&VD
zQnBIt$Qj4xNmV|<fh6fjcVBmLcI_goKE=GR!wD%@=iivIBp-*oFl7@V56oz!w#P@{
zTX{wgW$ZiCQQ?q-XzNTr+orLdj9+ist@Aq%_V{yM856%IGXS#EiVeGD4MFCGq(&-&
zngr|@?2=E4v96h!{JDA1@LMtHDOR{P*}k}c3BuZ0fYO&%NsTYDj_?qRjnYeK2vQS0
zQ%WPo@wVRz7vB-jK-7r(X~Yf9)yB@C^p~eIL65?RY<fvO#4e%Pgde@DtuR6{HlsVm
z?q=m=vnmC^@JFL*_q%PPnr!Ogv65Kgf?i&7&quce1%Fm?Qwgqm{x0LL`9>=?WIQ+V
z%0x9&ACP{5a570|&tPMW&?$D>Ur(H^E?u>BKE2&bm-B>8%rjMD-YD@d;1=B125#$4
z<;Z%sa`JS+AcO9}5D$u%Tk+fM8#xJ(Y)8e`X0ta7<rC6RPety0Z^CZFBb-BRpXz%$
zUTA&hl*a7B%PxBJQ!Rh#Teq!Vr<&70f>cJHVck14Wrq*Ja=aU5x$CJj9)<48%}J#m
zYYGfEo^V?X%y$P>lxw;jZ)%e=h{1eF14D%+WC{vgh1Csv0W!$K)3U5gWELCwyn1SB
z0Jx}7bm8^89?ci$x{IJm_i#p0<;2UFKv0+o!E3d<D`CTL#>;Mf&7`r8JdiasZ^ND1
zJ6Cp{CC6XrXAH;wb9tZA&kbti3loU>Bdc9quN8-t^+LEf1#nOXWa%P+@E>uRN1@F^
zOR+FCIl6!#*HJZfSGX$`;|hmww@zQzrqzFeXD=37zB4QST=9bt&sG^uwB2Co?&3Zs
zlN(yyxfV;e!A@%Kj-4gOE;!A-`fSg$L{0bxx7u={s%XS(p-vUS&hE&AKp|I&(HamZ
z%1uQjHRA1B)cB5AjgHPhSL}`KwkTvO8WLk=5Py!1@m3t@B0Rb~jyws*dwM9JZJv4j
z>r{8<zSh<ke>=M&&0vJglhSWq-mPTwD5A#N8X>13^&<=(iNYGJ9i8@PckI%;*(o$)
zV+*5|)ZjcbUlW(^y*CNl{rZST8hdUbm$^{P?_!O!0d6har$6B6{%DYP-iF|K%5IJ8
zMBT7~=eZG{x~@{omwNP<o+j9xMN<IbtqLO{Y8L(Zgg~>AAgPP+;tjvSQk(UF@Ki1v
zv5q-cD&TSXyoY$AwzDo1=lirf&+7bcVRaWF9c)a5Xf!z)V;S@;5WHB>;4>zB$ik{2
z=;R{)PPx9WR(0<<)WlMjIq;!Hb4*F&Z8-xX1t1jQ(H^py+zOgJgeb1Vn1dc-{bFm}
zjUG^|Y)Tri$0M_jq?FWRtyjvq9#WCU{FN6%HJ{$1GdBb&kp~g1LC+RZ5fivR^t`ne
z)-LU9Z{c0~$a1BoV;M1p4*R-M*)qgP`q1@TkeUzOFT}2@RUFhjNz_+5q5!Cija4!)
zFQw;dPSpRjdUZqD=$JvDjG+LmhKFIx7qWMO<Z!Us6KaX>WA={RVVN)=`ED<Hp91*p
z^jgkXn;T$kIdc@Ll+8FLv<-ezIU@aWTUGUlf$r&v$Mdr<@f1LHYI{@0U}a&42pfZ!
zxm=M>ardc17|DNIzq(Gf@Lri|z)N?3i9!9GlJ%3=f-+~8mDpdZwMnwScVw11b(Y30
zE^r)_aQ`|(wYmShC!5BkMR^&z@^<tw8~K5a`P$$+3INQ;vw^A{Bi~HLhq;Exc6%z!
zKzXDOEc-X}r{k__b_#!KBw=XwrD8nhKQof(C;5CU(D~~b$Mk=WaQRf8$V^owF9NUt
z%6AGObBa{h*bGwHt_;qAWh(Z?%LOh8;f@-*pz9B=Ef}^&E!0i+dHim5$SJL9JT0XF
ztS{6ipW3Wb0NBsAP&sofETYV(pG~WJf@oNzlTuvZGHh_>=eAg-jr2PS3!W!VYC?=t
z%HTFo%~4phV=+|anmQ0;ORJB|{Ub|oQW*^PsTvo&VeNMOsyTu0ktH3jyPowHrZjsx
zvHcW-=!EdPH6l^2+3ZSc!Hk=_adRCLI9;y$Tav7Pyyn36reH&V3ZM-&Wj2a7iU^W=
zIg|a)heZ(?vkehPyq1<Zj2;-u$h@1bBO@ZBuFq8=AT#kzVRhocBn5EG^~h^>e1W9s
zxe}YR3n6R6w&x~~t!6*J-xfIabmJ3YlvN&lH{%!a^?P(jZKDEVacWvWzdv$ZQg6ai
z(CqR4rKFzOrd)63byZN!<lCTgD6U;)cZ1G=DrjKG2N#CCK}QEW>uGcOnf1Q&I?2{a
zX|sOxs{nKQ5?C9sSz8k6NN~<2T1G!4K+d%)_t~m!BJZu9+z{UnYqAY4roEITeLhB2
zJ@KXnFdFzQ5XLRnAr(*fW3_U;SqrlX5^z-=T3@69sE;V0LS9bgV+x(7ced+r`sn?D
z<hKj2M5Gdlbw&uQ?5;WA^;xo=_24lz^`+G#tA|WwiDT)Jddl1MXn2HonG=w6+9fpr
z7G1p@dH<CvD*S!}hhcrtx8VmHW#lRsxWq8bHQ3q)6!!`W$G^)!*EXQx9B#gvBG}o^
zuL9~3sW-mXB~Iv<TWCH$^5X1#$)G;jLewLr1v%ni=(4r<L|dG`vhdT*lz{QD<%<`h
z7P^pII|9$F9W(lUJpB0Toz_FYum@yMFD+`I+Gv+%W3@US5u#2z7ewoE+<VrpT|cji
zeY)y0%si6#rT9|H0IQnAujeR@`D)+OV)C~@Vx3e*YI1(C`QjQP6?wAtGK+>(E>}im
z!*(0sAa4)n?J<*)2U;aIP!O{ALs`e8%a`VU)^Xkckh@KHEoi0owWrm4{4xRNgU&v`
z5nvnMQnGPTa?Ha!Lxxex#U1{<sSYrASH2<T>T~a(v1PIr-=`bvG_QsXre{ZK-f#hO
z2MJ@Qh~i7ass<VN({>N4iY@1-oA}NTMSi{hbNRa??D_d}IYUzPj=Cy#P!C&>hS`=g
zV=9^74A=^}fz2$hDBLtxrIqe~VW;xe9`N{QvS*L`@wbnkXOqi(4=8}TYJ<Mikh<A>
zqC7%6N0lh@{QU90`^fB(nxPT?+mI5Bg`l~&+fJ6L#)(AB<&9Qn)7!ZggG&vSoR8A<
z^YoIrpT&gRetN2PVZyHX)Q!NWh4m;da|;M*DcrEKLx_aSso#3#&Ir0&r2orP(I4S_
zzHgr)X2DpbJ4<R1hMo_&LjW(9oh2Hx7MBss%^SyG>?pTzs%?vv4WMohtq)3^ZRPNV
zXEgBhvh@238f_&6e<3PQ1gRAe-nR!h<(9Y&gXT%^(K{Ke2^y6zrSrdwC?&8P4UDGU
zy=89v!4R|Cntcj^KlGpg-Zzi8ps3Kt9vHUW(s%;5A%Jj-WL}R{m^YaRC;85~vJq@4
zP1JYzuKo!Sc?G(lR}k|S*|o*WraWUFP40(PMTHFdl4WK~XHkfV8o6!W(SuCYQ^g0P
zWSjHbGU~lSUtaA8t=+Wy1a~PM4v=e|A&cpppqhvWM5>w2d-Pvi!=LWq|HacxgxXTG
ztcznKy)6eGJT_Z-M%X3ahf$HFsoT_OMq3#G%89yqrfdWn=E#!gmhPV4B-hV_|9r%P
zkQ7rGyhy{p_;wVo%ttgs9S9ErvARGCK*9u?VUley@;rUP=-RI)19$ou`5lcCBoTOO
z34PTuvi^L~bo^`GxR_okq=t<JWjQ@Z1!7B{G^9^oc}|6SXoKdf?u}2=8yskKlcgXh
ztiMGFqS@4FmD(o}HA{ZmNKZb$My2}a3bhC#Jq}yjq(&iy1Zw9PgA)~i1A>lOfbFku
z#}mpBztz6L7UN*#KXU&vkkmjKSzczfA~1h;gKG|CX_p^1_Hum(!;nSu^6Qjt&6BR5
zeo1xGd}<LqB}LG^UH$J3t*v9|%F5c<f~?i0zPOvYj||l0`Aq?w`inHW25|#;Y74^u
z_;H;+@R^MJLq4%ae`G(J@NgbwWSXwtWk%V}TJJ_W+U_G?;MnFHEBaotK9p)04GQ)t
zU7x7K-~R09Yi=x!wYgW+EB1oM%!K|@=y@gY`ndn6q>@TS&Z?;BVK$r?o{fH@42tCX
zU1_1~Uvz8&vMI~2l%|={e^BL4Lk)Ep6*-|a`eHP9{4)`MnS1Ii@YIp$z1fHygVf&4
z<fTUQZPhPetMS@2R}6!dK|}h?$ey8sODRNLsOkcrJA0_pvGq#qRA%AOq!BgbjAFxN
zSZ8F)JwAP+a`bBIC(3!c*2Tdk{||+jZSQ9ZZC!Uo(^bc5<|%+qY65e2o|XcjqxMMd
zIf~${VMBg*sJ{6E1wc%*K@i2$;Y1Q_S&2$5@E`Q#5&hq2{0RHoP@Zx)L5zl*jQF8&
z`nw_JUoWdKL1HaU(;m)U;QQmXm#S<))IphwbN%SH<A)u;NXt0+#5)r4H>kibrk08V
z@4TS^LT^#4ObUNd!LpxMY1#gwhJV-Rxt!NG{y#E?LfPf9aXUk+y$SsK?E9)3Fm(J5
zzmS#gt$~Mre(RmyR&vC*NhInj9O`$RqX1akTo8M#Mn}|Ds-;8|&@6>7A{D%#0I0P#
zW+cz&mncG&@ngKuHHK+HR^~mLfgvdErHaLu48rl<a`AR-q^ux5$;f8MB2xGdm7eh@
z)roaSrn+Jz1@PYJl3$k;a7MK&y{<!;>xtpxd;W=C{7>`}Ew00RxpIrHTeBVum>4V`
z^7Ka(U-zENZ8-R>5sF;G;qoJsb6XH;cygm?8m_y8jf1E^DK<_tESU%M?S9sOdCU>|
zs)jE|?({mKJfL-k6iVd7R&^|};4YA_R}T9aad!I}mmTKE5=)F^7@J)sX{8<d;&cE@
zF~k-eNPAdli9VtlekRvVP|?lkT$*tCdh-u!8qqF;*p9Hxgt!LynfzN`-(nZmwm+!{
z3p{sQyf*6jBynOQzVBd1ADg=LKAOz98Ce7cKS?}q5MDo;|9qnEj&Gv+A}*o+Svd3x
z!c6lV(Pg;xHk5U;!%9zOZ>_hb!nbc`03iMD=t8Jaz3A5-k=Pgc7A0r+FR%who#c`k
zP)nQHY(XyBNbR&zbuH&BpWZAQGFBH+)sexROuubwaeFUjT;;%d5NBqj%N3TT@Y&r&
z5Y@N|GaNk@$*(rn4(vvwwYeYB{BNA}S(~NiX4q+AZQ^@spRcUcibxo>SMt!%uN}y$
zY%8_&21fzJ`vjJ$Xqf+ij!1BHD@D}3>vVNXy>VXDIVrG#X$8HXeuA*Eo#_rwEr2YG
z9AEQ%M3lse5U*3yh!B2k{4AoH0;upNya<_IKDP7iWI{Nt&VKgpJsB1jTR!c^Qhj?!
z<VQ07ML01=g|MAdP&zEE`h?tJ`_zR6qw#_C4wQCZ^PxIQjpWI`yH^WY@>Kx;B#V=W
zvo7P#6{1`C+BAi6zNyasOV^r}12Iw-)t?(``Rsyx`+Q!`rPR*1pl4U!1CBpgjn<Lk
zyk&qGn}sXYb=db)l9NK9<lP`4uumDo4CvNUym3O@RLtl2xiof8UhVee%@-knvb+FB
z%~}aVxh|{t=m#0<3}KVAz~NSE_8T&NbUc`Pq_cB=TR`IgK{i07cJ5UkibH~R2T<PI
z`v^%4Nt&RGe>dE0UFynP@jHDrJ0Dzj-Q_mv;D$`f_u<^0cHzY8;!jHnlLK)#Hr|y(
zOIalG<=-a{5KOk~FgBtqOf^b+$A{(0&zmpIz!^~9x!!9szMRSwfF8g<CO^hKzUAVn
z9SbSjM~-y?s?%+ZbFiv=7sS6S`QE#Fl$lSSi!KEh{B7}9-_0n2OmhcgUzz%-mPL`*
z2<%&`<F2AsnRaJkXeR981oir22xwM)w~~BBdbWorrS1@?QDgQW@xI5wBd7slNiD)t
z2n5<BpIh+TTZnsqY6dm87^MJeJd|PswTB9@q`}BiAKM2L;!)(E^Y-BINKUhlN+E7R
zl{y(!9~YV@9kGL$S3!E=d%D>27IggO=P8P%;#HkK+))qonI}dKx!;B|zMpIMFiyx)
zL&B<By2HKW0ai26VoW18HC!%8jb}L^fH%32_lFIgucK}Ftfn<R3nEdzA5OL3G)C<*
zuR2beQv(DC%b9%K)!G|5hP|hnsW4#6scN>$`dp7+U%kT>X69FQ4SodJ;zLxBJK=Cz
zUo1Hbfkk0K$39;rw~g~D05?(`s!da_Im*^)A6c9FeP-rZH9fiN55XCYNt)JYbTQ&S
z7OzE=BD824oN*WpF!HuXVGdNJGyRY>-#`DkQ8~Hg+n}RzX_wHu`gH4eHTNp70sNN{
z;;OIpN(zW)$y6`<4n~%2b;cXT9J2Ajy0g?#HNjOR5t4IX-F7zY^D)fZW9bu4ZVRqb
zQE(T0u_0$eUbl2pu(Z)RDoAstAnB#tPV5g(E2d9x`xuT*OqKy7L2v!|H-Dzu0*%HI
ziFV~HeZP)<n@(FcLl{MF0kuch5!E5F)Nom6P#;U(gCMuttOpsE8Cf_#YXp{x*xuC~
zaY(y&Y@+w-d+T;=mFHEVbOQaUfh$>_V4DxxRIL-tUrk2W*MPVQQCCF`gczEWlXE^6
zgHYTOMzUh=QRhe*Ln`s@py?c7m={#z>%{iX73{(7t!oh1{IWb*Wucy(JNM0{I6GIi
z7-{G@jy$7&Vqsfv@}X+&wPeBS30@8AW(lMM6w)!FGQq?Mz7<pI7e$+W1?|_HdVExZ
zt~3>Jv(m9?Y|UytYz49Pmo!OS-P@L>HsMkLYA~d*&*U~oo~xh^Jh;q75dRK;n`JQa
zSygM?Tz|0rnYU8yqG@9ayD{7!a%rzD`*h!O#+?Kj2@^t(Y-+|1K+LlQdxJ3<Z|feJ
zesn_vqPgx5|A)t}%}p<h2D?_~IO^WKM-NU-^l4#_qLtcI-+E8`#Iw#reHTuVn)Edz
zYxgr%1a=%O1=c<Q8lJe=163ZYJvfB1Bb;pIFpmjul!LV$tH-^C6?J)c9TH1TrsfS_
zUHEk&Il*$}(XhheH+KE`PR%Df2zWG(5CkvJl(^kK2qRFVifyu0z~%<3S|SqN&(~od
zu$oZf(wGu1-QN}MkB3y(H>LL0ICy7%nHBr`LGi(vWcL+7zc%dj?##L-PvzL5$T>o`
z;W(rQwq;<%W+(nx)lXGI<)gn>O;3%teC~}8M%CNbSt)>&@9I3KiglqxMNEc)KoA!-
zW5Bd=!P?ap4Qzpb$m{6K8-Z8VQ<*L~nX}wHHrgRpr6@N8Pi5x_5Og^l9c^gd{;RQ|
zbS85k3HZ*$i1mehDa^@GzFhHwzjo@%XD(Tb#Jv2AI|#Xc5VjTJwIgK-Y9dw>Vd#^r
zWMHR(ArP%nxU_9>f1)~CL)EBZ%;O;A)HI4L{>G+PAf|N$)#RRcOQ+{@)Ccez1gX?}
zURJ}hoJg}|{VLyec+(FoUh_%b6|(xAKmp{x%x+AQ7deRkNM3tubt3YEXBg&eXpMfp
z`IK5LUvCVH922dcW~pf0_qtTO&z8y@;xQHtV~lEQ8f~`3@{u{?35GCCARa={R~O)W
z2U{A1P2+=0NTQ^tsO*~H0nN3AwJ@ny4`#iIFl#v-lBu?KQL$#@iJjD*S4eSyhQ{Ni
z?)TO8K?YK@Zd|%w+)g}VUk-H(aP|kt{2_PI#*u#Z%G|^A$fp;W37@m8*CG$3w(lV=
zD#vz}9;v(fyn?$vvJC###iD@i281~#n5N(Cp%0?&i{SL5xh6ar33Mt0nU5=nRp3f*
zhrBIyv(l^xwwH)svr+DUQ^Uv!xpl8&-I-r%C5735zPrbTkc&5JnyB9F1IzIt9VUx0
z(Sh%?^%dDfEmQj)@5pk2+xsr^C-R>=Hh+1&`XQN`go&HM2t24j9>cC-Q`KhcgDS+5
zeGLw@ExZV$heDpT2)09$j7L~L`psO7RN}KLm3gJ~sB_g?Dm1UTw~Idb%V*jPz-qHt
zZ~-(5!v6cuhK?EyP9e+HVIRGxi96682YGA$kU^!2-qoLF>@<59o`sHsW>ni7TZ^%G
zm24S2WR>*}HJG)<Yl<_IevXw{7Ra-Hi;CGEI8)pUaQe0oGDvm4mv9%qVs?3TVVHgs
z8BHN~1Kkla(!G!rjxH|~RP#%5a^h<6laulgxjJ=i9S*M9s|=7mIUH(N#e*|o1Zwsn
z+?cK^LN$fC1Xzdp4E{|y6l^-w+Kn%xHQ~($Kl77Ia_JTGWAiYxq`A}e^qXnj$2R4!
zY~4(b*zwlfBB)VEx@L*bH<pW|$fF_avX%K>txN2Uzb5uB$0#00tNh^D`%%o3%+>~B
zU(mvoJ+{W#r4~aGcGWebS?lVW=TcI9cSXyZEDh94*Dt3(lW<~gBR{@KQ+^#dh0xDH
zH@9AFt;Zhrj}YFIISthAV@KH7X8yRm7xZpq!TF-rWv?D_X4}5-$1&nnh}2a$0d!!y
zYY<Fb!l`?!+hqtg53FvybkHksF~3ZC_3XARu2fn0Ht|x{$jeBH?yCdu<$Hy~b)IU)
zz4~wq!A$fU4l*us@$4?xNU;J%2tFA)EUA8HwGpo5&-RhiEWuCrLZV!HWW%GS?}|Zu
z1T9x=Qv1=2mDU%r)FX(Wr${&vcrNxi)AcX+>tZT?n2IPR>MbYj+q=O(6R(lt4nSd=
zvLiJtnsVn0GIRY=)BBbLl@S*Ax{@oTLSyOGdvD_VFL)|Gg*XYl6%GN#))0Ls#}6g#
z%onI2j}bvU(;ojV*Z{2>qYgqt=m?<}-V^}aJUCr5AI3t3)T4A*g@VlQ7igNh6v+Sa
zNM%g=Jnt6FRZ8QQqiVWHv@(mwzD3n=(w(@4&a9a|cNvHof0{Vc%18QFy;APJkWV@a
z4b(2O%KLRzsWH1XM(i#)8$vX4vQ;E~3E3)SD{Bp}el55|EZwT}V~S4>?fBCC$+-UP
z+z`-N?No%_0-RnL<qoeoAEXV&2y|wBoWSV~t<xLsC0qG3o*T?a?2m7Hd-N?;X6sar
zBn2X-+P2)zwe;HTBW(r`uMj<4pt|^!ks!;`-?ftM3lhWVab>1V9=XG-GAZYMq<qo}
zDua`%QjH;UyBt=W=4b9fOW4Un^0gI(7PbVo&Xaqa=b#4$Wz=mc++(Xz0+M@Xik@qC
zwlA0{L>Hk&WEtqA;*ALOc%C78{0;@M1sB-Rq;>#agv_)~ljD!bw%IREiEDkQxqcT^
zLlO2}SVV^i12k*X3cI~M`QD=-(?<=)bl#`M*lr>)2;?U7O7tPM*Uem-gT~>sIkrKq
z*LM_G_j@O`ppLu2TvZAIQM+Zm`#|QT$vTgkwj<p1e0FJPp}C#-#@2*l)v$wkyEWbL
znfYzoopR?;E;mtsaaW0iGw%}r06a%qwD|I&stXdn5@5Eilj`yqLpIZ4r_rS*EUANc
zAGFm{&u(oXlAxZT7qjxd_wuuqdZ~JVzX(f6PvyFuP||}jb-Owf>)+{i@dXQ(t}y37
zol&?`Q+55*Q9ZGjBRc#+Uz@99zhawC-ftb1K)<oLopxT7t{GD|`ex&Dx?8<{xy=1i
z%_v-Z?q#>f5$AW=Roj7)EF1POvZJL9j^gfIFfYWK`-$(?HS5e-cc3^d1-}Vv-U<@N
z*{Zwd5K;nRESpv#b?D=Tk$q$x1>k5seB+JeeKY+T+EuZNWQZ!Hf1jtVJC4Xk%2RGa
zh1-d=Odg2vL;E_@Yd@=N1v@*s=fm+nG4I{<-blp$aW?U;AaCMIt>%-ZcnCK8H4x2*
zGmrDht|JwaIfL&B?TYqJ;~}yGOzs(V!pD26&P=gW-2$3mG-|pVM1bWILgeV7n8^p@
zGU{qzblHQw)BCE%3wJ#nTneI;>tm1#X$43{z4{>T7mU5NNnqbploN_<=(t|zjgpgT
zo?L^^!69mPYtCPYUc=3nOYTz08P<$%V0kw|ad?DImHPz)dzQ@Z&zJoa^eP==#N3}a
z9gi*1{H*KD7FL2VxZOT5Sx+w`XQIbr59jUrFRRR0cyopvD3yr1N2ciPC16V#)}kQD
z>`)aE<{({`o#bHA_1Xpg(I7>Lv(nR^Vpnlwq}81eCm1lH+ADAEGb<BDcPluDu!*@?
zgMog<Mh5$T3DzfSEpFO6;U%J!qkTT_-?Z@4iDK-%&{yxlsBm|}rg>u)eB2Zw+?aqG
z4tYQU#1i9{Fl^mm*n2OxI=&DVoY`0O_C@jRCpezu(lb9t6y8k#NO4H<d2*?uF7DAS
z5hxEQD#)|qZMV6m(0A-$r1I}Ar<%I4!`SP+;?<-1^`<U~j~uveP2rmOa}w+=c|I5Q
zUIgQ+@EX)m6j=;STOKFg!ppdgyQ6~x+mWWN!x3+uiA?X;{1mTr<h_@3#jonakMFY!
z@OBYSmj%sn<#-n8w2~p4O=ETIgwo^)CkCVcCF-Wd?7+y`FvlqV1Z3wmbd-WT-5)ZA
zs)w5^;O(;Q`uvqy+$f~Gj|$>qv+<!#V>Rp5Y@F+jzheC{{Sd9u<~tJ;xGBZiIh$&g
zx8O)7CMc`O&#MpE9pn;hB|F&S1XZb9xi<TXVGtA*?s1)YvRi0j)g}#EPaT=LkMmB>
zS76ID;9)HO?jg;jgnEsjHo#-4RpXAMU*uj~7S;a}_PaCi@LE_d1yFkIORC!4&!*-h
zb0QqAXQ(M93{z%qMV>#UpDn94fQfE}I1O%*^mnV8ehKWW9z%c3#-dc`b5xGF_KbGS
zVZr3xw9;e$f~Bf)*dmM=mW^J1R7)!IUMezlE`o}7rYl-(Of;&%Wwy+8WUgbD+hv+p
z<JLQ(C1<X*z3rd6;GZD{wjkgjXi|ujt!`WE^~#~V@hiir9iuaMQgk--@5>j(7Vv2c
znmyr+P_L@L@x0LI36AdvXq&o`!jOvG1unzd4KTuzc2tBz1LbteHihsZe}|y(axGmb
z#a#EbFr3I%z=9ZVWOMR1)jX=hIodG%%2K;bOql_4vN+k09%a6U<EKJa;!h%dzZ&m9
z-ML))wZAs?pc+E0%vcXYCLB^lJp&G?F(b?>Gdlw48Y^|Twc#2X$hFaUwzi7(jcTOy
zr5crMK}XkK*dynpJH#r_y3%RWn9)S?m!<p%v5>YPFglxSju5(pKBB&};PS9+vn#l?
zHPs(2@Xk{bA7fspSpHp11>q#_nDA^ya;{B5bK8&V0xJf>3W)|89cC$H_EWl(CGcL$
z?pt?{ny+dGYwDgG3i7iYNeQP(SI*q)sW-`&QWh|9M$EJArv@rJ`=-BoJiltZu(ObL
z*czaV+bR1k!W|;x3Sw3+@vqUZc^}2bwsEqYVilya%Z!~aWx4j808?TB#7JGw97N<K
zK+uz2LUIuS1BFz-@pRMf!dPjUd(GL}2~4<ZLv_8vc@I&JJ_()}lQZy2ZK-hW(!rQ)
zIvCpBMhP(snzn9ME+qs)rh|5=k-r|_!09F#-ZOy0%#V8xEWE)^p0Z(-*P;;BLDX8k
z3-4`_Y4zBqkbaqtEDut*TI+CZ{gU(Jtn)JqM~9_|Q&0=B+-%u~TGD8XU;jlm7$)da
zfiO2Y!HtDm!}bkWvpVsrd1ygzwXlQ`U)KH>U&`=Z*`tQ*)L6;%$Uv}a&E-XRw(Wa-
z=&*9M@Rc=Wqh=K*WlYJU`Oec((Kik?m8qIGiIMp+j!`{lE;~!8E=BcWji~!U{Lpd{
zk(g^i`YhYAH3Lna;OFVCQKo{Kzf?J;6OVt1^g%zJ;qWS_5iqe18$&!7mUWkGH@HCN
zo=}eELk>Z`0ug)!O&KSKXYZ=ShT8Wphdvqss?F<NQm7reUqJe7T2qODx4>-uL-0x4
zCd(d%e!d5*)GCbc@x?)mQH<RM>~8y1P>4j536?a*TQ%MC#We7%7cRdMp5F7#Hb2Hm
zN85J#`4>*bX9V<%R>8V4A2tR;lJ2-i&E4D?XvhhZPjgelRfjWtmgx)k<-=j}PE%(z
zEqJa}@3vm31~G0Eg>vmIT{qpGSw4E%tn=<$=|ylWc};(bQ{;|ftoA&HGT-A~jK30(
zmm6^+K;+)T%hferdADAsrVfb%!|F<Jk@%$ZLD`p;>e9MD^w?A~Xz_WWKQ0Ms+PPYT
zO9E^|sMQbIW8Fn_LJc*ebg%e52@<Kcwmk>V&D-SMOxReWmKNSQ-9+Dh>vm3F;Ri5q
zd7puESK;`Xf&udd6YSdd<vehDM}RgeoT!Z^*$~3im1W|JS?Zx4>22`0irQ)=x+}7(
zxhfVhXEfc`s<WrP34!gm9Rx2HQ>z{<ghGQ!1o~2C^N_mKx$OLI`LMDXiHfK@Blm}z
zqYMTd`W*BCrWe-2$&7?i8>%0dQ)XlDcE-f2cKHV2Gs<mTWR*5+1(j3=O5E3^lrB4_
z$k%$c<qBm+JbmKc4^LfVn%zxBgJ-t$15bd6g*x&qf!LxW5}&U5#eMsrOvqwFIWX41
zWc2BXyV<;C(iYbz52s&AV$a2c>!la7<uz?SGn2INABS_}k&(Fsa7WhFF`p?PhNq+6
z8G)Wka`w`mww?0nF2lkX?={zs7>8Df(|THUZ-Jf5TpEq=ak*chsU5dHN=i25=Lqn2
zoSRipNs&$3N}K^uS<~^dN>ActAtV0hHM3AYd`K-;C4Vux$$Im=IWpb1*a4sJ6n!q+
zU2UO{t`yhFIu`b*?RCA=<u}e$hXU?fJK00D;}OShYURPasW6}UR^Pe}71)fuR>NDZ
zR3up#Cl_-rKL?2s;JQ~Cr!W0NTufKGSgh#RFS9#W-sxYH*JNLGe%wJtE0sV@3*cyp
z)Dn<umWW>7qz~94hxg<Wv?F)GJi+mKg+fpJ&AD&i2i47ms4-GqM+nU*Q9X;5@P;=U
z21fURm^b`V1qJ7WzX|#JhC|tZ(J2<EzYtJgM#gl_`sd@g@M^=&?n~RCj*7x}D@isg
z+r|ADqjNK0)gGKq0;DR-jDf~w*sD@o!9yrC=7V=o^NhRS@%mof00rQc>M&nBg7QKv
z!KAF$wid48CbH4OTJ3m0&pAWMh#KHliB6rVqa`&XKK-1+Z~Nh9Pj9cCeBw>v$N2PN
z_mDnF0Tm6qcx8MCerSDODfXsi{x8Jz=rNxOJlj)H;k{IiRvQ1!%Qs(mBe+9^?4UE<
zNE4Uzmex0=&{t35pNy_GepmHvnh<z;O<GFqj~@F>T<!f!VYe=^2Q8VBQhhYE(3^zk
z5DtPL$RO7I@wjYvw7TfUd{?nlKe?L;m$|xz`#F3~Jtam2D%095Y(rrA2+*d<q)aX>
zn`{Up)uI%?;DFl<UkR#@#^m5%aQ7>axxbIPb!~578je1lmiam><eoi<>PV<)|0X*=
zUoT`FGMuhI3~{Q*PDZ>SYJ*E9UaLz<`}P<02Vl!OJiqGCoDt;R_e%>s%hHhQ%Jata
zRJt_9ccK^5RBXl~ikx|rjz<lq8*Me=oA+&rZ4LSJRs2E$M3iqykQu>VGHj;?A!NDW
zR=II>QL<{X)M><9a%>=Fy|C=<@kB+Bcw<;D=Q#_{OZz(XJyzv5kuBa$rwzG~w!Yxf
z7QuN<{ODu?Y<oUlkKl-VK8tZ!5UmwFSA%ZRMIIFwqNc!}?Ui|DbgyCy1RQj26XCx~
z@80UtVdxzutIcj)ygsZH{Ac8W9TPklnf-lIP~}SZXrU)Ty;D6ut^AVx1(O?m9NO&(
zZ9mkh=~!<bAx?+z9s}aYs7>k!*s!%S^vi2a+}9FLu0ec~UbdG$Zzsu^m?SZ7(y`P3
zDukXCpV6Mpp)N|@>hFxKtFseoN<sTb{gV`}SS~<))~n>C8oq{@pIC14dN61j7zE#|
zA(K;7LT(cFZLo-V_p%x}PB57U%RkBs{nPxhYTVZka(s!M*B5VTpY2G9m{Qc`w-*gL
zaPPFhns;P#Y%C}c%<W%cH26T@Ck}+2@EfPduXcULE)rGNu1n9EE>A0_Xg~j&8E$bs
zeTS&MxXX=*TA2(|#2k!jj=Boxccqq2^%hw;srXucjg<_``M7(1S>B8>&88gu<~`nf
zQeDc2x<J~pWJkUG&2<c)&g+-`_5!u}OXR=e=||>y+WvTU#$sTc`G?fg<(Ra>o!N}R
zy<`M0f<;Z$S&E7tyOUwDhYxy?gc=-^X$9fP58a>+UV=0KP9Y0bj$N+98ijFLx=JRl
zZi~6b<K34o%`!i`JH;v_aH~AJV?mDko(^t;vp#k(O8GRK7g}O0XqQn=NYaiQee`}c
z(XudQc>Ag8-j{IquU`Nd_3-&5uYEMA14{G_0_8*3+9w)27b3yA8`u2NYH8amE|T@h
zj$wQzIuRn(?562PjjTJK564;o*%9O`!!<@x`vge$#EfZdXR{!XDEeHAt=izae8XO0
z*3ueBS4B)x?7&^2?>ulG`gzpI_Q^(Rqjg{I$Cjh3c(cP5gHc`IBShK{=;%^Pu4+3l
zETFt_$r$@H*q{S6Wp{17t>Uh;I`3rkEB?UxyfB)ET7bNrj(qI+bPx@0<tQr17DKw-
z*R_U$y_I&~${!_K^t$-X^*HhI=74YJ)@AfDFgbE79M4)y<|+ZFA%@ua=(1h)TSZpC
zHF;}rVww`Q*#4AH3Dz?MO2?(EdQl4R^woQoe;6evoE)ctJj>*osc}>yk*fu<m?WDJ
z*9+SoA-aQGjJT5UMI+Q=QAjsrF`I8eTeiv{J!>%K?jDrU|6sa+#r@*uYt)TU9nQwm
zJ8bFqFTWm*#>b589vJcl%ON!{j~|lER!>HNOf&i4HrRWeuSM<lyUOy;n@Q3b_|IB2
zzL2RmaYfL;s1+@=$e&fMG*Z%%hE^vEw-&Fu4Wd@IbtwSLz3o$H)S4z}UZ{THEk1BO
zNW%wQO0e(rZS;$h@iBDUlj_&f1DEL5j+M8L%2a=s>3A_~`Fv*c`u;^dBu|kwRne^E
z#;N_O$_pD-yGaPNe@K+vgwYDMm54rP-X%+<ap=Gn(lJw8KdBE!lQgf2wwl$zWVa0P
zk<X)m5`T#DemK){c0M7?8V%~+oHs{;NhJkfK~!Xo9Q)?PgrIP<%J=zBSLZ-6Xp(nA
zM~>Kyug{k&9NWBJS(zcKsNFN+?&`N-Xb3mbeu+i=P%_(Wyn3+K8bk%SAIi$n)#Cb@
z08URFX0J|vr)_pt&ni9@wAvJ8jZ@Yzx`We=IH0~PYRqJ}GlVzVcy!J7-r}HwzVm9~
zuTm~>UWCqQ7k?QO;;?bHujO!|W4#XoY#%{kXN~H&4p7k}Py{9wIUCHmW6`WpvTn7N
zI-&R@`uhCzoRwC^0E1cNg7;->*IU3FjNmD7SF_xyM&Du*S}2pvY((jFmv!B$EXH`2
z5FiES%fefT?`zjIbrcIK@$a?GwY9pcp6TC;&dl|sTS`24eZDktpE!?TV#?u-x>^LF
zHZ%R(Sqc=U0D9svBB%5=u-^q`Kan^TF@j8u7{B{;lcU@K2x=MpoMaqu85B=$q*~iD
z&;=qJ!4m_!K;$C0L*USi2QY@&S=X750Up^o`Yf}b1SG<5$=Yq)cVvyQesSHJU+5Lc
z-EBq}KaqXmA|$EiXlJ&mf`3B;EJS7(ISHXY9Y{*%II<4Ork0+5`5pNiMbwl9b_7Vi
z@evL|nYJ{A?slh#g;#i%K%;hkPV^auZ{VIA-<Hm{kGL6Tr8jhey|nGlHD@DSZG}G;
z&s6rf1CJ%}LU@DI_pOI~W}7-IN65b{;P|``?b`y4j0I?hsY;&xb0g;4Mua=C^-~Jq
z5(R*#Zb7GZ(HNnQ>8U7Y&TT0TU*auTA1=gC;y7<+<cR2x;DNRITchD;HhoiCf!SmZ
zHg?cPfbd#}6xgRJ$t3-^-92@6Yx{>={px|kdQQ&tZXX`fd^2_v)QxF9OBfJ}7bGN^
z6W)5Fg&#mTpg|kvvMF^@?<(~Hn4Y}XT?vIB#qa$n{c<6gs*&f}NI=-K2>ECHKN4tQ
zQGaYrEh^Lz`SxZeWffLFJETkyg<xff5)r{5k&$L?42yO5<|0JkckO~u%F@FKn+4Z>
zFdr9fo@Km(8oahh(&L7$!?Hulx*5^Zo*Hn6pL8Xm**FMIrr7wmN)TpiY6ugd?C-tW
z(xQ_fJ})LClC6+qo&0*oY@dJ21NHPDMU?vwlgFb$svg}ll-z5Xw7+^<VZ04#gAYs`
zXmSuTmuz7MLqw5MXl_T61-J~|Bv6Q*6C@41DQa04&r2Lo(g`O!-_!zM;koziCg&?Z
z%u$CC?>d6L+4{MUp&WAZXDtF7O-&pLS@?Ey7DjoO>din)^GBNg3ASF{uRj@whEVx<
zM^Da1?6FbNLS5n|-~ox+gbq8sKKyfM-3PKuuG`YgmJ$(w?sq2rhFvVBBTFPbYOEhu
zB3#1aGcalyLZ_w<u@0uA0?d399aTO1A4<R3oI7_p-UN4l03{M6%#B_c(V#W0uZ@Iq
z+C<85FNvLLSA9vNA%I{D{Ksco3ZmMYY`so(TT)HZnObDmwpI4;a)z%6TvX}3rDJv}
z{^?!1Gm59zHQ<Dj5TRqljbKnBQJLtgd1uHcP~+TayP)6O=094mj11)3pC>qD0;T+Y
zG$eqx{9cH@7-}i{m6qM#oqY73#bV}tbWJwXv_(-?JF?yHq7?4#7(^M!G3UbD_p<RV
zDO0Q^^qK(8x#!&5=GSR%(O6R<>wnavzs@4t5@hcX@n!cUgpmrXqgR6rN(uN?V0i0U
z&FeVkryFcV>Vofy5`1<dBhHZ$XXc&9S_g%HB}RPm4}2uu=I~)h$f{wNm2e^qen_wj
z3dYes!s6fDz$uxlJ|ZLrUiAnWgmJ4B%lzJRcv;-uGb?<$x5b8nVP8`KqetYsRC`jw
z^oSrz?NIO2!AUoM-h|~_*;*n6yILQMBkv7Z_=xdy@r#x?T)Z*a8L~|_3?kiq(F7*|
zy(Y1mc0)DpR=b7-EkBw5!<9B%EGa*19&7LEmv^j+xh5tP{&Yx8t|i?1e0Ng%>07F=
z`e2w!dTB(WGiUR``i>lRq}-l6d6zgh)h4DgT&>j?(PJL;+F7mVQG?{05AqFF3)8ct
zl#%nQkiFr@zVe@$oub#_hm|C`jcE-C$pf_sxgS}S6ZL=jvad@GWWlk2-xndfMgTw7
zrV`)<Yo-8B1ir9+LM(m-RXV{>sXL-at5IwqZP&7ORoIKVxRr<~`<hhn#X0#ky8nZ{
z_Y7-lP1}Z}fQW!dS3sgv6#)xH5eVu=L<9t-3sC_nAw;Bw8bkqUA|N0jM5Rd=A%vPp
z2NCJLgoNG`YJd>m<(}vL-tC^UXZGx8j+y!V5f5FidaZTe_j#4`yjHisEf99v;nJO?
zA_6zk$OIp_U8flLJdLIhd<0u$tQd3SOO|<GQ`Du=T~+7hxAUQFz63&H%zB!quLRYt
zJ<l)TUeCfC{SMG$7UyTPJWLKQdo$8W=TBdAjJ~0o5op1C|JKOy!A&n`7yMdfgxpR7
z>Yn-KG5#RO3JsE0Hk!k>@MCk@rY#$mi*xz20e<2+<kHSlA#^Il8r(N2>%S2vEK0dN
zXzzPsMJ%wjQT$C&sJMIP5o>@4<elZzh4FPrFMU^^mHp2A#Ji6M@qP@Qxw7l)89>RR
zAby)p7_(4O>dNe2wy$XIF4#1*d}eUGLO)EhUBh~S*4vwp(&os+BV_w?kVXd4B*h?U
zh3SpuF+*&yuNf>*`8sGiN*MPHH30G;%=4$*o<$d7gBI_ECrc|0EYA{ye!A>P*^L@~
z_?iZ*Mq%VH%$p^amq*Aw;q=HEY??lu3|qynAKgiE#Boqnr=S$3{C?Wayc3Jq=icjX
zH8H`Jha*1M6!Qq*>gl?d-8&WX!8F;;&eTLf`=o?OzVi4h%d|vYZC5WDTu|pNpkyqA
zbo8Gn7yDLf#*ChWEDMlZcb))POH)QFEe~b_d@;R&sCi(0+R}6~mj?0ZssCI5c!S39
zo1C=E-=dEZJ3QThrrCwBph$2nWPN|hS5UjQ8vpd$5D<oA3zN|aKn%m8EfByq`dvT&
zUyZ4pax^hl`<JXEePN_>i0O*%L2ug<@>0Zo^0UZMFgLaJ7Pj+Fw+#5FIPH*QM-xm^
z!cLc0_{B;yzB+b<&ACPRa$%tlo`_fQz}KPr6p2%{W~FO~kLix?=--ZopT0Fv&b^Vq
z@Nw(tMGFHkjeF1OW3rT^3ChdaZ~UCrwlC5R)+K8Q!gd6%3MqKrOyrPRooF*-Ph3d0
zN?NVH+|7%}^qX<#IolTQ>l>Z;yjL<c&x4qG&FF=z3Jr4L{!4q`FEQ?j3y>Cthyzsl
zZ7@xhde(z4!^TDdo3G)Yket;lTH7~X>&K*P>0fhh*n)5yK2<w-)f<HZayXLuU6@1_
zg;s+|_!Fr^{-nvPyV%9DKOMAk0}q2`){nvpByGi3d2?=AvNova6H}C5>4@M?WG9R0
z=&!7n(GC_FMcbo_?#GDvvY}Y02q@K0+2pQs=1j_EGcD(JTehR)DC})B_#{bLR=W35
z$Sh9CTeFB7ZkhSPt}O6z){J8K-OTBM_dyGWHyKTMZ`X)Lp0XW(rq(37Tj=55Bev0~
z-Xy$=ijJM6AT)n%>#z@x-qnl$#Hjf(V*as_^V=iF=m{Xqfe-f9+t!>bi%lf8a#KaM
zg?T!q>*&EYmSRQ^bBDL(95!O3^TziFu@pTzBK<1YtCGi6cWa=|+3Q1umoGb{ZL_?V
zP`H=`A0q?_NXLdsVA5txC7;R|>Gj-Ibjr(=uM49LeVMrC(W;nuYV*n=<G0~{9DB%!
z8ZfwqSRSgq@5|UVN*2<0nNZ7ojm$Lp*gbRT<+b<aJ&v+rMk{jV@sfU5wF(}zrxY4Z
z8}PWmdRL-%9JW)7dC+G<7N(|9X6rsO+vvS5KvAA%83h}>NIyh7)l(rw^vJ0Y75Iqb
zDeDkTK+3O3*PP^17vy^;owI~}yD0Xktdf<8engW62o#nkNIWsQG}Bu@(WL833dv5+
zF6?Z7W9O!}Li(i|CtLW)!Mcoi2}0!cBlXvOOw=W!g9^R8MsO>-tMufQNjJA;nn()v
z6CG!y0mM`p=m%hy7pWYnGpr;;!1A=OP|t#VXp8J6|9(=%S!2g%7d`AwYP49paWU;o
z#a=4~O|{`9y}|7?K}(`@-b_MoZ9c0xGQ{U*-B*x9$Cm=9cc0gF*(+B11}_&Ik6Ass
zq`s%CUvPd?D<bV(_CvS~kXuA78_kTDy^fxwp`a_Yt+S)D+}@`<m*h$oA;kS%^A%!c
zfjM*m@_ISBXQNi&x?$9&C}Q1e&xIsa=QCPj`n=(W=gl8zf?xpmR5oNy_JAEfPJ&E3
zby4w2QJ;&K8J=<0LLu%-7pK+sb+_Id=2v>$cMWi_$2(0rgjW!T58$2K^{ktkI(5yt
zp}u8!X>5g5=6UJ0)$WW9JM0(+Cssg3W!j+x<qlsPs~Lz3uDCN08`~S_D@L~d1XbhO
zn^ICzL}5x@f-}8WTNXSoto=nIG&M_Vq3l@pB#<sFN5yTTt~Fir4=r0jw9>wUGTeZL
zQS<Kam`Z>tq}|*n`$`$+Lo9$r<HAoq|Ja_ra7tVF%DUc0l2<pHdNm*5uo`!xmKFfq
z0?uWMm+mmJf)~^W^jCt$6M(q%lMiPb-&cqh?DWwQW;aum5~>b%<?i#;2HX}F-mjn?
zR)PGT9iUcwY!y|lj`|9^RG<A7gn~n;J798*<PO6UWSRdfi2r6EHAJvE6?;k-Av$G~
zz|W@E1eOwhz%td5XKY`fT@opzeO&IvQwUK64=55_0t^*=F`DKZ@#`D>B=vFzD0Td;
z-Bfh-vF)pa<=ailiCV_V9;pLn_9tR3{aDU{R<q@H1^wpDJL;(&M3EzbI7wSU)>jZ=
zP?rwTpotj)-13w5_>Fa-iaDZZg6GGXe%R^9SFrPAwEUP0sDJS*Fu{p;$!lBIl;u8A
zJ=7Ho%htSeCCW_gP5eH8r}JYHWwXb2$MEQ>@xA%kU9y7A_Mp|6fDOOrci{H-EH(K!
zXPSKGTOb5`9s&qJvb8bI{u9dOe@XHDT#F@L%@e)j*$i+36DYb*4RO+lgVg(QUYDAb
zs_a8+C78pU`4bhL^k&Rmf3=!VA`Zn$<Jsd`MK`xC3V_h8iY~p<8dzhrkr}nJywu0)
zl!!n$P-owu*WoB{$%k?B%DHXaWRtNI_pq7KDU+)pQB+d$h{W!4E#{0XBULL<-=Kg9
zd~rZ7Y6;%!Y(K*8*lLJLFu$p9n>*h<-Weil)_mlN|Dn3<^P7Q7vG<(&B3`VU9<CXN
z+Hp03Uo1#Pp%uYp(c7OgR(6-jcSVf>6dAqFyp&`ZNCv|KDeO7;$gl8hHXa=|;q*Sw
zGrUxOi2CpX($U2V69WL%ZLzQ;H@8cb9Co3OTI4J=U2sE1PU#_%6J|1CJ4v)J7wUnw
z0>RS0VguA?&Dve*?Q)^hMRxF`A4)xBj>NOx)`@(&f7NJtud5G!dZFus0U?~QtnUj1
zs{6*AzJk)F4GSy05PbO*`)-Lg*IVr_K4md$-FeZ<Em_0!23YJf^-FA5_e>pPEjgwl
zP>8aVWu{%HNY=^v$0f6rbdI&H4*LRCMde{dvGxv&_Y^tr?rjS6da`SSLvGN!{HOTM
zEzHh`5KO57KL5fntu-rv>B_P`UI7=eMdP?ljsFS~Bh;bRS;_&=Ue?rCTu&nAj1X}=
zTt{iN@ok~6q{32=>dABUF4;Kd9VgE=%9W-0coWMg^QhX1d*Mm%KV+UWU`*aMx^GHO
zx%CEl4Oeg|DT-VtlEqOtTp$%A?w<ALgj=byO;mD-dL@Haac70W>!!U4$44fwj%$K$
z+#38`<=2-w{w&p%xl^vEBp$BMx&OVP{5}Oc_x5N%eyXvX-VC9uYd~>esp9VIMz}oF
zGv+(Xx_6bYWgU6ByH(<$=7Tdj9<Mp{ewhyWH~IU2^pM}VZ1BEaZacI^A1YBVzguIu
zd1rFp9%f0gkTC$4DpH*X4NCQ<vt2;66TNllP056Y4)D(5!@k}~4N`q6Nn5fe!kwZ>
z?1|2s<1=)Zdef3tTJ}uqY1B_ok9NhxN4$LP{kr%fLqZR)jqYgl73Yk7XrY^`K+ylm
z7_oPUDGx7|mE1~_teInZ(l#5l`TCCB^P%%#?T;j!E)_=7LLa0lk&E%3k?I~<Gltks
zG0@W<&AUc-UPoP;cy>m!!P!EEJhh%sN?~q52vfOhIcJLbOl6a|d_JF;5esJZC0{jZ
zv@D$5R3n-u)^^909<NL!yd8=OO5^?V8mr})Z@K!I&d=-*I1K<5n#uzBliV6K84f03
z4g=;fG8Z7liI>4~(0fgTYT;6K0?CE$DM+YJL3hslyd?b0<aA?n#Usl|eTU5_B7S50
z_G{{&%ZF7^DWh-*6fW%J5L$#YKRsb~Aqb;Y*s-UC^JT1#@5}=Y&t`LeBf2L2U9Yw`
zeiYCNP;Kc|6{GUk3P`+38-h8DSFTmWcE}liVBR-->-ja#SHh#+W0asO8}tHD^Yu{^
zkjkXS8B>==&6#VXr;WB|t566z!wvRzpb~Aw1k#;a_~i7Jmgz6-?nN#oTYX#3zx9wf
z%eQ{2uYe-ns!{wF8=k&f3n+MP2O8Tvu0pLILeUt3%LeDaoPN!b0U<s>%F-VLrt*vW
zlWyoJ8M?Cva4mL;0|lk9Nu^^9Gp*Fy!mPj|0xTq>VtAT1m|UU^_)Qo8@7-@c3jzJO
zs~@)d@m>ADiRQ-`{5G%s*Ub!F>rh^0(RU*k?GV|@aomjsSJ2ik9PgiF5Pxu9Wp3ir
z02xYqX3DXH8U$9My4M_;K#oPtG-$`PbLKgqF=OiUyOT&4zxNk<%v;Gl>7f8jAO8yS
z03ttU)%byw=<mws{IxvLGCCBptScXp3fWRpqJ~!cGEpCmF72BhV2TR^s*96jtl7h4
zA%Tu;;X+qAwK(5_%=Vr2+vBxw+nv-(i*B);;h;ZzhC{q&gY2f-&=2G@BE0}*vu&Go
zog~EFb+l~riub!1#f7H9q|$nwvt7NPAqVi5vImfOJ&HnZ#_YiTm^^z6k%VPd>SX+F
zyH53qX10JETn*c=GMNI2CB4(=k3RgmJrJxa;V#T;VpCAHeP0t+^}f?{r&k=Gl1Q>4
zBh2KT5$rK4$|n<vP)o{PL-Ux-#G?DRyk74y9~?JH<D8nZ6?uM?jfKgW<-^039T_}(
z+?U~N<yyKk{rM~O$jxn!ZE;v^HYrhCWKN^F6M@MUS+ACiH8?#uaNN*n_Zf!BFv}w=
zQBtWC$tn(k5$eEK&@D1;OZLe`dG}a_kF!Qs{_6yob=B!1<2zN#8P9K*Fh67Rjn@ER
z0(*;qrf_hYrl8-aW;;$s;U_{aK$x?dpSZLInn6>U!2o<Q%Cf8Mr2zn@IKQDZDJH=T
zXjU=V`UT8JJ55=t!5>1FLD_CoANp|ACiJdX#F-W(<>k5qy|BVA3j{<LcUa$c-&6Sk
zmq>2O_xHYQ<La7f=V?8}hs6qbx9%io79+)N)32Y|?N)ZGadhlRWD5-SKbH@p4*`I2
zpP9N0A*P7NisDB2Am)5f+~c!SvvAoOrNi_Xgu63Q%3Q|Il1n9KIJaDTm#DZ<y*Vq#
zx=GL*`3Y4P>KM79!9t0g5#T^OE}i~{gkaAIo|mZTl<wgW@BjEH-ux+=vvFHZ3Ue@U
zjAy~!ddHS9jvhKJR3^KEootnTTbZa%4IhPaduviM9<3snv@+*ZxfP&}$fE6^{B?_7
z-It6~AD#f!Ump!mm090}bwU8)hm9SOODl~b0suH^?OG{(Ee%<!+HWiGw~~4v$t>n9
z>iQw1Y2seKW!Cv&lAUnw>j}h(Wy^iagdzuzQ`x7V{8aZMsETo`m3Guu7(hJx0Z+>h
zaQ7&3j(Pa}Io)X8cPfYDm*Ko?p-9=9tkfjt>mR4+%7Cp%upp`tiG!VH+Md#rqbjVM
zI&E}Xx*nZrAL}M{n08&KH(WGk{DvBPu-^WIPOto=FQ>Kcd5e2q99^{mJ>Uvv&Ippe
zl3|%{X8++@&#6s`;}gkZ1`w|YF!#n2d-jbct4-s-f4r2ms%e0Akt@4-c6>l;PL_I!
z2Jp7j4xa^7#=qT<Er7~pYY3~mIw9o!5W0jPU<cfXUu}%|v!fXPO-}Z~HGl84p1tYr
z^w5mAO`eUj#RY!N>vmr_gu?uMU*BN<>|)ddB?nNw=zdBz^j!^qLb!XOLHKEgbkwL2
zilz<nPE_{+fd6+J!QUU##@tl&<3vAf^y54Ff8*wVafo0rqxemsBJdW#CImuPq0sC9
zcg6Xi1y47qT8X!)y=8xBM%_t=!(v*-HBt=EFEaSEgG{~&0d{>0!e0M|r>6hEI@*6Q
zaI~BC`)ZXBi>N@19CVXea=R5s?<IgLe@8a+TN%+``!T=x)-gBABgY4*yJ>IlZCP6l
z`bF@?o1iA#vaJLZ;8uMFIerCg8<w^*QL~MxU?R{@gDOY8P@615InuLsVVGyoZly9X
zE6)Al@J|+-qDc6qO}bh9r{|XUcpu%K^JadZyuIz*)RkAXgS+DUrQq=9iqg`RnT2PV
z0sE0t<jKdRP+9g_6!${W5-0SI=?uSnnYWSqp_z&5_8^CChD&&PAf8w=GnFxwai1=z
znm~swc5RlAp&rwmGQD*v*a6Yw3hz!f1gr^ycaOk&W3L?vH#W8B9Ij%2oc1vW5WNz;
zS#th9W7a%!ka)OdDikM*EJ0e#Zq&un<;X(=s(G<z<eD8b)6%+P8<-uef0{e4voD2N
zOcJb%x<f#<HDm(0%yKCG?5%mCK)~=?Zhn|do@0)cu_N(GfB9g!NLfzE(NCu}3_+vM
zKdu77%E)X$NZ}m5zQ<>g-yI(VEYNu8iHIizCYM2ISb};5ngdn1AeUrQx6vxy5Hunv
zo}hi{i@DOBN%@eQ<5@5JwcyL|_u7)npnH}1oieD2#J&&Y!2GwgpE@T?AP1Dy-P5!Y
z!s&rMu09(%{ue0X%@@Tk3hPQ6y=4-?y`fj(fUhOuCXftv!_)4U=9<+D3$=OcSraRD
zNj}9%B&m5;bCJaj-H^pG73C7sQw!)jFjWiND{(P+$V^_QOqJy43CV8k1|ZqTI=!W$
zo2bk&ij>2%)0A|97Cs!92p-$_NKxIw4wo9@=9=vMu!-=^2V#%(!i&CyJMekzKJm1V
zU&r|oQ8vg|mfRZPEk*J|?VYnRC)+%#IO_wu({qM3coJKJH*4l$biQEPD&jE9mKoWY
zj{)k-5%_7#uCJh_;ry?lLn+&Pw&{UDw0{Rcuat!Drs9EiEDBxGJ0_bHmoyj^C*NCV
zwWBhu>WvTe$z4{>CFj>EdB-hC1iUk!RSHZZUYOg=%@vCj-mK72nV8Mq)%rsGWDWJk
z@pHVnM8acdJNl6}Wpy|)v;|U&#RD~Jy!7|Q4_RY6GU%QCm5q(X_fIK36Fu2jGwFbS
zhyR=RmL%N_m)}k_^)xYS%kwbVaO-WkSPu~6cXauvkTG)@aifO7K$90{p|DSd78%3k
zE~GeHDhnX`iGw_GH8++MhM(658iYX-YO{E6zIw2Nl$oaX0Qhe2vmD=3is7;Akb^)e
zv-Ws6_rj-bRFT`}rCpXUwKEb+a47favOjz2nW-Y$rf_NHyDkH-PNjX0z2@+_iM#N`
z)u!>qeV?B(aQf@{P1~)UzqKCKV}vK6hZpB&X}QrM6BDMgRCd$?!G8t+xlFd|U`kJs
zwmpimNg_AX`wE2}m+$JFAs3+#ey)omH27lj%cVPaGBzB(<m`b;J_8y|x!3ygQ|)_9
zWgok;bGMKs1FvLQTi)wAYr)f%XVsP4<riP-^^h2Mq4nkK<&V9<OYKu6te&1lZ7lYF
z1%<WmJoW)(5`_V0!*>BiZ8?TBu%4hw<j&*{$iEsIN}K~^GgnX_S2Y?E8vG`4WmL3c
zRqL)<Vb%_riVJUQc=VC?d`_?~N$HhFZJo;n&(D9pb1w&7Zl`z$oOHVhT(g6yg>qED
znR3I2w)j%AsP^YpeMCRTjqqXE`7*eNb=h+^&ghnZlA656C0BdxvO<q2aK(ASrLBW&
zknIiFS5Up}uf|q_8|)L|ab+b$R26Fh_19YMos5W~X^|_C_s43><b3zy;)p8f2#5T9
zW%*P)eNDJ?hC=iKXFxC3w|zH3?x7}+?dXT|16S?iNQ!|n`1d%WA*xX|))ZW$bzc6W
z(`#yIRWI&{u~=m1d{mNzU_g?{M|AUT-_dKY>%e3mjbGUd{GBzs=3t(_EYe;JuKl!J
z?UCC+pQT|yna9_{eGG4-Eo#_CKQP^s&{8*1*XR{2!ao4B>riKC6}`;gKj4udcD3Jt
zb6n=#yMt1@Q#!H*zaNN@?}tKG^rZXQkPiig>XzS6mm2Hr8#`=3rf(Y<+GFzKf^vq6
z-))lnet)&!-a$S*AJurX2E9fH>R+`~zJKH%9n9MiiL}=Y`Li;nYkv2i4IcAtIp0{T
z@?h_jDC0&I=&%oO_}?^I+n$ymd;DTZ{;<dQU)T?O{1Nt0J<PX!JjJbd?NE#{w_t+<
zub;~Pqs+{GkJb2W|8=F{f3Ujvt_cVqYCIVRzKU9Pab5B8IZCmiYfcud0%da=8c@YH
z=0e{u_?1)N8f<pwUH%9m2coK9Q40W&1)c|B_>&-G;18@Iy}fWjcU_xqwB+s+H=k_S
znuv%id}Sn^PH*!>eMFyK*wX%zxsx4^wu!uH6Z2|H6?cc04L01#Q{{(^94ny6WRR14
z^(ef-dpeEEF__pz!eQmoc-B0Xr};J<QIiMD>@oXfZbaFC1<7yamiH61y}hrv#t|i^
zjdD`X6)h~IQc6d>uTtYk<$ToqyqRQ~xSDmK3kVSl1Mk<gV~m_7=8ToFyLHrel8ho>
z>xz#ZO*uc*&!99{SN0{EBGusO81oJiC{9s~kgJp!Gg*Ect9@Qj4I4MkC8aT0Ui&`N
z)tdd-rY>kjpm*UBxfO6BI^}nuL@TBA^0cp@hQLfO1YEv5<NQunPFi%bul@=6&GpYz
z_dMe{KF^36@vf6=v4&j>i*JjgE|9X`rMp^-T;k=tGCwNv%+KgL(o0_FE66Brl{A8L
zTQ$#l2x#Y+XeSsyolThQw#$g6`{-oNk(i&!<mV_KhL4%0MsYlakVBhD?@l@l%r19s
zi)hue<+AwcM3K%k9k^Yq3av<=s-cY5Q57k-y{{@b5XjE~%aLQz=ml|K0H|6KCVV&C
z`+b`x95#zLz0TJ;3D_*ekmuAp<-!$psrO<9+#)2C(g#fI;6jG_aBIT86ORPpk@dG2
zg6<A=!_C%9sOI^|w+*#r=xQ%>jaox7L2Zl0o*uJuZ<pmsE18mYk${@gEKMKth5HK=
zVAVY7!Sc%0p0rvMEldHYVXylLG&L4yX6Cp;(x}Dr_F?OM+^<tya9}LeoxR+c`cUht
zScDO~&Eb!*;#dVVS4H@457|TPhrUcU^3qHwf(@~%GL(8ObI9V!RiJ@Ep<2glGNDhJ
z?d(M_;a6A|x%9Kgj|`U|)zHs<6?gjtW$W*JXTuE(xub%uSRnqtUjQHoTy}S#6fgIA
zi!AogZ;IJWQhrxl&;*q}9iFmZNF|~Enl6V7IsvhyxhXX*UTXp<O2tv=cqtF}3$5%l
zLyBP<#Zxwr%YI3<^vxXqR&4$Q={H=rZT-cgnw$#l<Gjj|*msOOu?`0w!d_8jDXk=M
zxV&!AR(eVr?0_BanAaOG;aul8b(SDw>*vz`mmeU^$B^BJ#IL|{ckcus-+&i-N2gtb
z*KBDl&cfMF`^;LrbN41+eO@T$KpisBD&42E>wd#c<1OfdD;a;80_Yh97cB+w4WOQ`
zfz&4pU~t&siN<q{QPyfpulA>|@`7g<;mwG<QUe5){FRQKq=6xQ(#f=uzNY}3<e(~M
z#6<V?bOJ3oERCtM<R$ZBN^D_{y+BW-0ybMFh%ab=?!L_%ds`~_olbI@zOgwozF9sI
z0SmaWC%JI@GBVtyIrs9_-g{-8dEQ`R;<nGVIc+m!Zd#W-OKwJ&u!z=)oV!-b{ryJG
zi;CP0FapM?8jzDk>9k?;AFm7Gg3`z4P2oE+sp~x@f*LX0M{j)`n_vj(Lkz+`hHdO8
z>`zK2wYY4)%sCX>?GAs0v#=B-?Df^FEE(@<4O-E{uiKe!Y{Js$Fvs4?1y~%XwY1&S
zh&3^w?5Fq3pv4ZyGu5NWxsMO1gz6q)1++P?W0Kqzn=IDlaB^UBXjUMue-fK;?p_Fg
zt?gkaf%hh*=LfoHAxuEmC2eOCIM8+2ZLGGHSZOb%Wv;)z_h?D%AqRNgrN;LfLRQ(=
zXZO7w-<5n(a;lPly_jl3%-;`iitNBFaF<-x1y=W+IBOV;J^xi+-iOU=kKBWPy3^Nq
zBI6w|)@Iusd0Dp{SB4?WcVoI&L*<9sDvoXJfA=m|bXVM+r&%mbyVRzczD*t>cEkz5
z^MoPFT9+hu?tOwI5a7A3(zRjU<sh1}U=5JYIz9`favlbg$p{Xhd?O<hHTsh)WOfm7
z57d_Vx9Vb0Z9nN|Pv^gY5JyLVOvxKveV~-$=`<j?_!x<z2dHR$1??71*)eZHO$%S7
z7g$4AIk{)Iq$W71NJU_&n7-MCCyOZk+0jPM&@JZYs2mfc9mb!wa3PzQ>B8(l9A=AW
zO_sLzS;Y58cMtW_AB}_O_zZw*2u{`fx40X>x`FSHUUQcRk_sNozx)2{d-3DPiLF5R
z47DSkjI;axYI&hV!h*I9CA)q=2aw6}^tv$U7P~6H){*0o&HCncRsr)y{IsW`{D<oe
z9NOL4>hE&<F+xi+wYP3%8tUj;`3S08`efHaR#}0RfwkmUkZAjQ{sF-HveyAEsTXKk
zgvFu7Z@$Ao#bw|%R05fXJ^H9+6|g<fzGL@lEA5abP|moz3vf{W_Gx~0Kz^L$JJbHK
z$)A62|39$FWv6Ud@qE_S{a5C=Sq;?Lf(8qwQn_7tb!YUT>Bi13K;?I74G2q%bge6;
z7rMGVZ%XgKq`I<L9)DZX&A^u5jk#ni<Y%D>rgy0BT-Eq*{LZ$&7K#3??Tq}88q<Wo
z*;?Zd=<&Au7w6zOPAvphwH4znU`k0z$xob{;rK<-!6wd6*?m3mE9fQjP6cW~X;#^q
z+^b1(iy*gdpQ!KEij{(^wweO<h+##}rs5yVq8VbV&8;IO@1%R*0xAwsPf`6Sa`-gX
zw4skiDDp|VxV$kx_$qnZY&i`}#y7B<nO_?LGwSlVZ;1);yjDX7igD(I#*wrWUtq6a
zr7|e{)<usUt2t3~jC7Qx2CX56jUI*;!o4FJ3cU}TO!j3cD^KazpP+<#9h>i%?wt$)
z^HzX~t(?fHjoR4)s$%tn86}|Z+or>1HQ}=S=L;#}A_Re`OBc@agI$s}>zp>gd*Ln~
zYl$tE7J&L{-(?#)rGS_?w0vKASd$|cUnEGJfp2KcvHHxAO=VCO`rtuLpcJ6Spu4EQ
zXj?@IuCREB@r?8broc;<@m40F))S+LKc9XaVo#MF;~9u8P)H$<+Hd)03LR)o9afvM
z4+XJ0iY}*VayRQ09oVcoKOAi_EL5Mg*XtbVja5t~RYhfFYsK4aCTj&7VI~_Z;@e@Y
zIlQ!}4YL{4fd(PGP)P_*8dpS}DRw|^uk50nkD^+KnDd+IbkgFi$tWYFzq@Pqpm|a#
z*{eV#mnutCwm8r$Tjb-e6S!qdyG+C%U&`@sgf9vkT3J1~#;2b}NuIs<<*cEsvMlBC
zEqnz~e@OJHoQ!w3fmtg&>*>(2Hyn2R*gWz1gXpo=6JtvUuIxWioH&XPGgU}&b-ha-
zxgdwW@xDtvw-O!4rfX|l8<KrwQ)jHC;N-BDRn&v6<rv9p#$|3r@THoRP|uN#4s+@I
z;o^_%EjrF!d({IQMK3qmCV0YoHjLI9dKKrA<vUzhp`28S8H9WvoJ(4?x<Fk%@BTo}
z#Qa?@5%NC7(Xs<qw6!FwHkEBMkcOQ)7=#kxdRM+HOHr<zw(CQYF7aTuO99>uVs8UZ
zf)0N)ZSBDAqnWhR?%I&9dUJU8_F$Xdi9J+*85Vlv`ECXqM~ytEkuoh5GtDL-Cx#A|
zZ%M1n?S4`0Y*rAm$TDR&uIY3Hf$L}57m4v`Y(d;S6qt7Kq{4}WVmy!HMAQsy5xW^F
zB<<Pxj&jg67H&3-uKA2M8`N4@K1z<LxgZ;;G=hbMx*T8g5g4ofV9z-)+q0)*SzOWS
z;?d<HZ1kD%3FFhm7gLNO)#X$hI1$fG#R&$(?_&Ti-qmQW-tkBAreB<$f?s6Cf&3J+
z{4XE9%~g`?ffMvLG?Bk}ms~9^((@IhMQ?(c2^!wI8kTwdPK(rqMib^1(eqUeYNru3
zbROyyaM}Ut!@;VJhbAVo4`z#_bU_~QiQv$zH^UJp3r^nhKcjiD>GO^!x}M@P=%~0s
zD$-V6kSKO*A6~MOeb=S8uEnA!mS)o=nH`R`Q%LM~HHBwBcHLHjM5YW4j?Bfzx({@x
zKkJ0ygnc0H1FTF5AZx_=JsnrYIiwnt)vd`IHJ(orY=i2t4sip`NM8kdZPoUj<;^Pe
zj>Z!6mj?jM{JcIN5k!|Ud-){^@aL}~7QD<I4YAKAuOyRpPbd~)_q!!6#OSLOv{>2!
z&FIhAq5FM}{H74|3$7Ix`KX%154<a~eOF%WbdXPZC~uvcbktn2Jc3`C^(kA#@+>zt
z5UT(`W)O1$uiYz6cU!y%v`>NcD(j`q+}t$9CS|qC@pTx_aG9G_TRyKVn;bp0e5*0%
z3@<swDgS7Rzm~~(puzEz4tiZ{j6$5yizZYs>9gCZ!C5e?G^QRCYnOAsYeD$(ed~LR
zS=lF2Ln~rgPc-h5e5QlEvXkg5S;*zhpXuDn`Hqg6COX^ib4G?4D3p1%FVeP6H|7ea
z#?B<<oi@HL{nOUSnfWr)!I2&>#KhzaBOpFdBTv%>K(#E8-NNzHfKrk`7!?ihCoO=a
zfwwHRH*Kc?_)yxeL&)d8g5oDomXpv`b_i`ZblfA2cLqgnL2&~ZTkBhjDc~ze076Wl
z19-A6E)&S`vC`#G4%YZpZVuXh$RxOLGj4|gxELsIQ`k=36G}S^1(>^ATtG5_h^mjE
z-p2<6?@1iUUQ`2#9`N8H#113iATUI12Cb`F130u30RG~1jRZXlpdOj6mja<bsnn$d
zz4TJ(hw@3N9Tp%mPR5UEtk;_XDE$<$_Za=|PDDJaO%4b#u%IIUu19YF<rBaB!yedT
zdmDg(O>RYBLHVp`Kp#c$`)9~E-HG$%N0t-Ob<HZ!THN6-mr9k~j7QeC0+%|@SRK37
z%pG9)ahe@@e%<C#Zmdy;PKCHc(KStc*;M88iI=nIllH|N(>bkcNb1<0_vr;v2h}E)
zUqQ<95E`#Jic$tNCIdoa$pUElzC1j27zF{hNdU2VfP!irK_dh2k8I!yWMlxp>dA)&
zI?^6RJy8ZE8C<@B{J(uFkVCsMfgB4&+gjOy*UMxCr8!1WwN9d@0)danKRchU_%ptv
ziN``#YJd-|$^ght38v7wn!%ey;6oYB0v!49pF_bvZs&(}{wuy6gYXan^u5@b!5V+1
zu!?y4rFXv?L>bPcmv7>yFh#gmjiEx%TQbb+_cA`eyjPRaYJEx$G#C*qER4Y>lMp<1
z&#M>}<G7k7-U#G4nKX)2L<T!B`CsBPRK8NXr;3Cf4hzpH=W1CG=64z&YkWSo$~)D|
z=POi(O&&o64LG#vvQJbLXr>Q+XyP6BH*4$@ROU@q6E;x3`jM|`vdjlkTfMbDEcv0d
z$iR6b)y>_J^}~zi3)PR**gPo&<r{l)|8t3v${F94C+DZCo&=^~<sQ}Nz?wBZH@O8`
zNA&m7?{jRse+3O->ShUtaFTXt4w^fKf7_@kqXu`V8)IhOCebdFQdZ{wW%256RsxoX
zt5NhI#75M47dBx5VgucL30WWj@@XC3r*`6L;zV;@1InIqpze2{y{c>pg)c3qDqkj^
z4I7!d@vwz?Y2@?~m+)?+GN<3w$p8wpc@3J@Ty|rnY&%zquAoTM+P;JYR}H*aSNf9T
z0h3RtA9&I>W8|)oaDUB6&%>2Xm%U)%6QR;|(^&BmyED)O6J;N<&s(0}g9#+uqU<k7
zCCscVu6w#g`^dQgHx)Z{+>V5eQW-J-)R_6I@k-rh;j^>rK~xrsAgL?)Y-f%ut63%3
z<b%T2ZLF|N*XoOoLPt^LctO$AI?voAHqvrQVp2s}e#6Y6xIM_YTOGb);i^6oTEkwo
z0t%(+b<R#F^&;=Z3Z}&ONSt1RDej7a?s_qH-t=h4&B6S%o}~54?ZJztEm8#C2HjR@
zNd`G?JswZtA@J*lot!V4L+!N~8Jg)Z(Ga^or4{J)l;P@!(-L>ZQ{NOkTG3bz>-0T9
z?+}R17&TLYkz9Ksh;Y*pX<wh4*E(0X<NfQ>hjJb^Mg{rc^dl>;6+ACFUUZ!K0{sxG
z*4=1;eW)xp?!NBRHCUzRUaMI3@(#x}4s7m+Xo1O#CvO}d0u8KyJPw6g?7Z~8nPrzS
zQDL;DRJj%r4o{Jld%dSeM&C@oFZ69zq{P|B;pJPMrqb>cH2uz=Ey>BG+|su<Vzd-(
zR{%}&{KQ2}(X#!Nd%kaH@<9aioIp@O8r;UDIHCSir*4;Yp37QE^449Z=)i;{rX!jh
zThld}w;rYygp80!GAXPzTjK>tA`Oy>^vSWRfO~0R;Y!j<S3ga%q$E!oMw%RYc|A>S
zx$lh3#XLwop>QN~jN^d)mS2~HLsHRUuEFaiRvzKFv<!<@ccBj&0qlGmPw$+%rx+3U
z*whpI$XCSMpxs;ZEwy9i3K<uWX=~(p>OFa)a~5nH>&miTDxNUH7tw59byxf268y^p
zCaxkx6Al+!oP<G`U&wn^uUAxhtH@$ud$=DT)$NTo6ExLb;2f2<)a+S1=jxU1r2^K8
zHZFzjoLFv8sM~eFEI=aIlqAHl_B`n3^Ic+L6fids_|@f@ChT|hv8-Aiwohomj3n6z
z8L5K1Qg87`KP8o&%Fz;Q<K;S|Y$cZCFX!|5br(gk3i=enwVGs0jiYlWN!|>tO2`Os
zdAFL_G#NN)R{K=`^L*QtyK3;U+7se~rO(!n7%V*ngJ>5?xxQOl9V^<;j;|8fO|n`}
zK6*SekUgTKBdRRtu-g;Mcx*Bu1TL{!wq1&z;whK+-C{iaHg#hGDEmges`T(hq_xqN
zraf)^=P`#-wNiP|+xGmX4<#NGlP+Bdp7Phww2aUd_GvopaJ0{gjc@ZJuP(85fioQ$
zUDc91cjk;r4?b(g6P_NIZm#R>#Gcz#AmtO!J2@m8em%%`pYG9zOm}8bj5M3kq<b~B
zTbymKaJN=N3J<<NA$+x`xX{Gjy?R17`Tf=Q7H|xo6E{QL?N$4UJcu>nU{l7lX=O_{
zfUbjAcg!rx6LmY4wX$saum|jo?u|`U6x}={=Owy%O$qy%L;X&^57b~oMfEGlB5v+(
zQB22-<ch@xjiFzDX@LKu#o-Tai;k|ri~MH(q$!;s?M|7cjFPK|_`KB_b%-fv*r#uu
zAujK!ukpUPt?sQtZpn*q_Q;QyUNhU3ZPsFG>5L4%=SEtzWOI}r-3(f??t9EUIXV_Y
zbt3cYlXbA8f=kpmv&$)VTjQPqm6dA??TU&U3M&)hWp?5(7<dxLBfAO@SMexxPtjP&
zx+J6O3$pB5#CSls#y*rN1!kMIG%PK>3Kz$CSX_9i@RKQkN3X`YNRK-Xpet^_QjyI^
z(Y0tPc>_DmI$CQ0vcxUOEG)0BEqIObKk-(~(!7hxQqtU~F5XhB;M@wc1xnzvVWCgk
zfnK-chYlg9GJs~<pCQ~R>q<nOg(AggEu$afBUul&fnSvMHp`v4F>uDL0dY)0L4R+)
zUdJi7Cb8>hZk67)oGC{dj_f7fUI@Pw7NrAfxY<OrlX^(VU)zCZs=#+p(74avRJQK^
zx-isuV1pK*eEus4y7U#4TG-QK2_r1?kf(M~czRST0Ez#FLUa26#U-_cWUzYYTEs#t
zw$F5A)Avei(WMXY*7#ReLN*M`sokp5NPex7ZpVn4g-cV_3EJ}sZYiW@pA%g(>xB)H
zamTFAu*O~`oahjf7Ff7vw(n}?cNMx7g!+WTJ7LL^UCO+vK%C=SLm6CYRHs%4-|R-s
z7N+X?rXu%!iGrVA!9C12^z)|JssgV;1#}TqDf<2ax+rRYB0T}{M7w8JY`3lhRFYIH
zThqJXwW)8UA0T!09c96L^BvL>NWlbb<`a;Fk4hWC-s-OJW*$<_%sO%;6Qo#g1s)aA
z@J!Qh(RVG4y5%8yCS}Ya(I0%CG8O1*wiJ5GUdiaumuQj}O0d5I`yQ!ytD;-HK;XsQ
zMdqh$Z<(-O4~+*;lr;#-0UfnX{U=|D(g=Z>zWlwI<SmRW@gX4nNh_u&MQ@=$E+Zig
z1Mr<iG_l$E9?<L0C6jNmp4U<uQ&T0%Ub?Au<yvwo=W`C;018L#2$(DzT*SfAPm?RB
z_!j1x4Vzzty&1G?ny7QU%XJphMLsr(B<~ViY2Y2T5GNzT;L=6eiQ&gk)lwWGG$YkF
z*4ap<m0Bl$^o{!;O5sW)#uG97Jj%sy%;N4i+<h#25r9&U%kqOUh$nm_;wWCtjQ$x5
zk;wcUz<CW4FEVjl5a=1Ex4Ly)@;_pD&gx}x<BKzamw45kfno)yH9j`mu|BHE{+f!c
zY*kUE?i*#vlo?-<cFW2xWstJ$!_S{whD}H6rquQ`aKlfwkFFhj=%Y4c_N&tP-Y33o
zBkTp8F0D+{aH5^E?rXE$KBb6Yxp1@k_XAolKf~moJOqMBqfNyg#+#SUypA}_{MlP-
zbns;b1G&Xl5-Dj)f%wbK<#SzjeG{?Yd{~TIehzj2(}2y=#DUzh&yl-@eiDnAWMEg{
z%v<jv^d?T16_;^!hJ*rB2Q#^Ct4t8eIE4JLRHHZJX$f<Sy)nye3WS^A4Kn77IpG!(
zbxaTK+0{7M2nqvS>F=3kP%-KlF)=?B0|Z4lkPPOX6I~U&Z`9JY&P#Mzbg*9>)GLVE
zG*y!kX#Erc`)sPld#jib3wI|D32X!^!(pQchFMQ=P57maYRVuLwY_gli)*MME1S^D
z8`>ZgDf#^F2;DSrx&OnZ*bG=voDqVS(7U>W=+v`ofz4!)qJ4|u>+SBU?|+MrWWOGr
z+BdCao}QR)Vm342`bvulla#8(bLz@#rk~V7C;UJlHqeP-jz%`bpHZ29b?#!@)IExS
z#_ReO|G-S{#(&XH^3l0SjZ2~WZM<^kDtuZ!8U~oKJ<vj(2gq?NhPB;gTi@rC{WyNx
zhOhq~L;u*x_J;|6n4tLIzyx+=s>Y?_g!6COU8rf{k5hwT2d(b_lCTdh=Gy;9%Uk_+
z6lV?Xx6VyKIeI9?^q{egt^GZscvTzo>0y<Eo+1q58Jf-Iu;J46F8T$eyG~cmgsqIj
z;@a?u<8LB$4jn#p(U9@L#Y2p@poRIg&){`n#l_eY!XV>Qs)ta0mH9xZXjtLtoV@S`
zjT`m4r@==!++EI2%vYkzD*U6hwU+)~mA$IE*K)Hi$z9OJ*jt<`>s;_|B(CIzztY5M
z+yZt!;%H-iLn-Rxer;syQ`IBL*oIp0L7;1xlSitZDPLxWZGu5KE&}c*0yx!shF)it
zf9}v}I^>~Tbe1pD{1y}7*~)ugZj+>Z3G(qFv3i9I&sBGrX_Fe|+`3|9*DfO-gzvmE
z9jb57XntYB?*(2Ic(5^`l`l)zhO0Pq-XW3l0P-;2$Z{4jFV)^e5l&~`amIY(4;X3s
z!3M1JQqqytOM0D)DO`iY1A^!8#qVvja(xoW=+4YLn#@NH8pBZTKh7plad3@m-QGUA
z)Em>%cVUMs4Brc``(G5P7eDg`qdo+M9k0^vcdi}qzC%qU$p&F9j)pck&qUkeS#y-{
zik-4^(~(_js+19zWf|noEz3UX#R1BeY$<t`otTc6%8+exp?SjQ(>@csQOmc{=usb!
zUl9wOElyJ0gS;7>teJ!9l4jJ0UX%THe7Rm31DG0{gJGHnPdd*4;%6af?g~pE;!>E!
z(N1L_Nx#I81Gk-x?Vk7dYx52^Wu$#-ec&kKeqy=tUf7lPuv+Je3D`TTu2Z@Na&AnF
zej2W_X6IYKEZ=wWXxo>hU6+6e+eu}B!|1Z<zL_+0p&`EVyuq3q%@vChcZZ*}6x3AD
z$6$$c*wmZ(u}hVrqO}R};I%K`*f$+OG2Jh@ZWf#fl4Osov|=@V4<&aB$ZRp%j_fPy
zEy^y02^B)C7f#3~AV`t52O8Ma>$z@D<(VT}!Fk?R)Cf`p+ibE#0p%p_0R0mcnnCTF
z;VW}3DroU0QAQN)#@r8_dUPu*#v6XLEpXVdJ2wjWgb@h*@|idMijHNz%Eezn9qp^q
zy&-afU_LXd=@jXf0$sA?M%(m<;XdhkzrZ_&+``Ly4VCwY_(;9|6(t8}Ku*SxWLat6
z6sDbwU{}FV1Q&9>{?qpL<dM(yZP*~yu6H@_Px@txUClrZhdam0v=k5_!NO;|#Hf;_
z(EO(#ba2YuBhr_vuY9V5=UQ&_cI7g%9pwkri0P#Uj6XC+fi7HqyCAr@und>LtD+yB
zp<82R6`P^UHXYFK2spo{GdZ;D@A;ebm!Bct-{Qyod+q>WIa_RSrY&f0D>Sz9&Y_gz
zPvAnGho?Di?>^sHr9#$#{wlK2w(D0z<$F#+PYchkzs=J5vBwXKeNPqqVUK^*9)FdC
zW7>6t=>(woDSerqYq#3?t#;4vFR6I`J==b(<q2!o9~@Kkm<m^hIFZn-gc-L89(V31
z1D#w4;E~4Te!6O2MlVp+eiRNG-44C6u1i<6SVGX>@0_Rud7x|ZcT<BtLI%L)owD@a
ziLapMp~!{7a{9|xQ|>e7$~~qr^_j3tE#ks{K-m2JQEY`*&)G{+yk}0F$9@Vt77rSo
zzrD)a(XJ-sx)f-pkLi)lqk)Op*>D^Br3WpltGG=*2;nh2SXJF%vvB#{VNi|_D^WFo
z=*8mTf(Z0w={T8gYpMl1nBuQ^s?qvMjM)tqO%~VI;d@uBCMzXE_!3cu;K+ssVn%2q
z3z38UHMTNZFUWzqyqHRu)WJ^=T^n&oeC6Hn*#Z5s+t!;K0k@QYp}L2hV|de8CPupK
z9A{<1(xAg9>G$`9tzL%FQoNj+@+L!eFQMk}Q<*Cf^}78HwP=1!Y4RNXKx2_t=;^!k
z_>{Ds><qO}`nIdvOVaEmk(ytq6IJ(&7(kvj!8k!XoPv@a1zfYlObu8a5|r#N9+odS
zSQXY}?<UNA-6j9h+km&&&l)Tw!fsRQrh`!}oso=wTK5F!`nvn2xlgi=Dw5WGv!FyM
zO!x$l(+HaE7Q=fHSg3B4HP+J^Bl2}s&d&CVB$l)Am4@vm)yt79SfPZucdr%t#*C`#
z5KJWMg&HY0tdyq=BcXzymE-w`z-<Nhr+7EPce-)sNEM+q&_Yx&P)8GDsCudgLS)A!
zRIkG?T?m)e&{ovPmmt#uz&FaTHkKY2>aeuZW3N3u6SAcmVpp^oOYDq$6aK6Ea*baV
zg*&ugWkX<)LG(wSHgvXBAhYJF+D!s)F#-BWXQF}jCUf^gK=s>3Pf^qSfauF7UAjif
zg2hpw#@-?xhz=fyYzqJEgZ(?6A`Qjy6p>fdk1t%U-9B7}Myx*x^1T<OjCwb=b05Xw
zIr9z(A^bbu_<zkC`0*hBg~RF(c&XlsN^{x=6LmT}L_S`Scq4l1_OjY3wpVVATRj+_
zUDtr>b=Dg{Yx$i4ufcR*|FWF@ANl)UneTOHn!jsYX1jascPfqlHFa{K3H(;gYBsM0
zVdw6a!5^y>_}3H!|Jm(nK=$-Mli0BT;h)T$|3Bd8-@Sj>xBL6gO$VR<la-CIUqQyc
zG53BqlKy+p_8+y+aex`W%l_LGlK!&~0Q#Lw`CUxERsuC~|DMJJvH<e@zgI5nFCEA5
z>VFKKJqR$rA=e6iS4)JyHc|a=ZRdISpR0QC1xV9=GqeGK;(Nep#=m>j`A1f$e|Fz}
zw}46IKUMs8O8QTh&RYKpn&arm`K$77zeTG1SN{KhZb-p7_D^Ou@ItKx%^3cn)74*^
z!v2VD|H-W)&Nu&T=M$AbgF>F)yZHPokNW|=|5-Twj~xF4djBgS`w!^-N5Sg9=R7~4
z_a7Wv_$|o(1A6~L^WdL_)ql_NKcM%&5_W&P)(`0YM?mVo_Z&Z<_rDaf|A5~A2z35u
z_x%CAe?aeF)5iY7znTAIquwi14ND=hC}wG5H;2C*{?IJIg$m12mWT7DJt)rPzHNF%
zZvZNCf8dbEF4R->y=HJotg0ZjHG`&1jU@M0C-WrUYF~5>hnRF=#JWz#OE24KCYWit
zJ;`Dp42cn~w4!$tm@70WDH?4q`291T_K^OsAkNj?n|j+STm8*UFWhV;V#HQBA0Kmy
zkj9Qrc<>G*Q`W7>d2vLv;H<Au`w|8!WS=Zl(Y}C@b~Y>TT3lDw;bKf)_h7qfUb-vB
zqbX<T)$vbm<vOOQ!4Sgc+0EH-k6UR6C1Gjqg9Lyy-eff10)sq7?_N8+XZv;$MX{zi
zm#g|Jsl4kA{=FB>-Fx0dj2**zXOxSjst(>i;w{tVEh-tZAW07B4M@2Xxz=dWJs-nE
zeECIHZP%=Gop*e;Z!?6;!IzExdFpEC;`~~E&`-@dNH6vB&W@EYF(AbQDvePsli@Xr
zOJ@W>JDgJP2J`qf;SVTZB%^Dw38<ho%Ul9=y|peb`5@&=X?l}HDQ}608tKzs_z{m_
z!($fe5~s8n%-Ln^q~;e;QgzVQD<es~<%0)@>(9Vky}VNw##gV=eU!G+@C&kJLM@bA
z*^Y={wF7ftd}M2Ihl2o>t0!4w=EXtR+&y9y{OwZw4W7u1YE5iCDzcHEGi?aLaUwNc
zY`t|esjM{vf!VvYVQ>#Gn}en7k6yARvSmj5Re3ay+VCzz>Nxzqg7{QrkTqlYr_#Q>
zw~_jUTHM~^wxZArm1f#@`7)Lz!C$_DB;2lk>Au6NaO&}7)5p0xRMD}eAmBh{O_CHk
zfDg@I6v2apTy5{vt>|*N7kWKm;MiC~kJH2PgNI)<nu<k7Zm>)iMgaY-a<x?_z*>3v
zdNh6PS{kIgiH3$NaXiBpGgl)XV^IfI%^-Vce2-F~k8|L%>7DVp(r?PSTZHZ(8{e(5
zBFYuydiufXxdZ}^nyL7R<^)L6Ql_dQ`|M1i&t$$N-OZw+3<V<nr|X5!O@^2WK4b_B
z|75)Td_LDs2I@XEMEE3yxz37=uUV@spdg+UsYvOXi|T&}>G<TO^)fY<VM>eji9y&C
z>j*|T(Su=*W3O+E>fSb26bJR$m~OW`l-qWF;$$8`b7sG6q`)@KcX{`)?ulBR(=R}y
z`)xB)&u3(7&kRefuW6bP@-6bbOjQs7=d>!nwA7`i#jb<qD?wF}4dZwcZ9u+2&4e~9
z2@IJej`n>ah+bHE<`g2UtW9>;rabZGF7Ni{fRW<OEw%T<LgjF7cVqN+U1$=0e!9gQ
z4%<FP(}zCc1MjsW%a_}`;S}j3h~N+QhB&Qi8Rey-fE9Maz&`qtV-=k)l|||O2zmOR
zrq!FL%tBrt(-DXZBle`joo}?Sp;$+b9BFzdceY`D&@6hVk7q%2;^FzXN+r<T{FU+E
zPeZLfKEv2Tba|^|qQ{&p-s&r8VPO%o6>3Idt_HJF69#J%>eEJ(q-qt-7zMOf*B=Zw
zR}d9yYcDveoI0OQ&8|f=SFq%m&Zvq5k3(EAoZf>i=yp80VDpl@h(~kpeMZX1Ij9J7
zY~6?&n=C>dPSHw5qU%P4r~$d`*N)4nKra^+#`4`U6vX1O3A23IhEHwG2U#(<dX^f%
z{e%X|T8m@oDq*G`SxmUDddKQYj%sd-y{SSmStNs|?SR=E^P-+V=#eqb9TYT?Sw0??
zJBu^IHqH1T`Z9Q1#F9h4f)b22GrU|D*I=`V;G0NpBD7`)Rj)J|kk*iDTdZ#}ba4LO
z$xorrvradku6{HWX(7to@J`;`i->>XLU>%ZkGfGWz)TYhPq!}%GmJIlOMNAEdnabZ
z<iYvyqhij<du9iAFv?&OnA5jL^=t>@1E3&$a5}b^|AF1HpV)eqcnVXmUGc8CLOuB9
z^lHu$zwHX`UZ?NjSw%2Q#d(S|Q6_`fXJrs_!MdktFrM>}Rq$d)2t+KM=jo@$y^_yC
zvUb`QGx8gzo+GVpadZZ_>_h&_y()>i<Oe%~+gk&5x;(yZhRJ)W_voS&@JIYSyW?P;
z!pOnW4tbtu=fuvQ({qQpj&nM3D4$YqV)U-m#lU=zP%jhF+-68;<E{$sG-Tl1oyVHE
z<S&Ay?yn^JUIxJ%(wXI#buim!X%^kM{V)L`0rgeS>&Yse@`(LZja$9<D$CkkO2j^w
zFIvUCm<r@P`2k+POYBwESVTXRWQq(Oa|@S=B%%0`&LqAf`D3YVYd-rsz8IT4GIkuH
zkHFrU62MYwLU&}kZ|cZw=g-9^C&9C0%VrZ+-Fi`T37)IC5GA`5-QGFWA!^}hFF=O`
zO6g&$OytV>_RWGpn317jl#-2LjMa&&jk70Yxsim?IDgYxWGIxp-ygBiv)E?u9;-mG
zeX<&>GOze%zDA^ipVUy@8p?;0p)lbd5xue(_N-L*Y~5DE28gslF!ClHU};jfBSxOB
z-G2|;=~WdTiiIpWy4qKoDoq(1-TyeB2C1MZh7=)!04Z(TSR*$Ml|sRp;>Hi!N|zqw
zWG^?pI$!@lU2Jh-vUeyQNNr(_=!}w>g?30~6)-m-ASGKHBuJesFqLw&)+Fgb*r!dG
z9(&)-mUxB}>+}x3gNe0<+(*UiYZ5Grkr9HAt_=1<h>Yak9o7LO`Yv^V2;Dn{ujP15
zm!!IHgQrpZF`<G!XX`DLsS)fwk;K8{>b0hqT3jTT86-a$iiL-snYc2xmfs=Yc!Q;B
z(Z1t$_$H827JrOQdX<3JH``d~AZ)wD(Cs8Gx#<UV*->qkX{b>i=(irk{zw=-U_lB{
z4tx8p{occns4xiiqOv?7Q??!h=r`!0C~~N37GnMue^EYdcPOBgko)Nv1IKTk2LzBX
zo^aPdS7q&>R3`qPyGS2N76A0nya37UkH_%TgdspZ0FW#E?z!6m+O}jMe%hh~Pu?d>
zm%tu-y<bDC!@&i0(7vtD#H_EnhtJWnIM`Bu#JcedNNAMfQ+Xvn{bCaxkPMl=Ujk?s
zM2_78n)A4#>EUI0HGnOjKSWgkT4;w}|1VVVK7FhBf7pBPpr*U1T@)2XL_p~s1q7u^
z?;!F50%9oAOH^tIAs{VCAOg}m2q*}sRHaD`B_v7}klv(*gbo4;H9&|r@11k!e&^or
zn>lCZ{Q1rB4>KhEva{FPYd`B*&qLKi`I3X@@|Jmj=S;Dq{nzK2z)LT0_%=U1+_pvE
zK@k8EJN2#58?XuwYqcV*+Zb*fWun-+V4SMa6ghZ3Da)+kWw1a)c?DJV7u=zJbOUiN
z(o`L*y;#1$I7JREgR_soC7;5g9d*s~J2~$@{x;il|C(=ypbW#TFZV4!+dfYw<>`9#
zk;98M|5oK=i2@h!OiVagq7tO0g~U&c96}t)ZE952x6Uyg&8=m|Z{7RRsVi@v>#W37
z`U?`Ge<Oyv%c(Ta1PnzTn9fRen8J8MX8?<cuGv;wg?ZK|M2-a9qly)-eVNS51Ak&K
z+>&{>czy3_iH{EZ&Jm^|s7sWj^PKn!z&Be}8nA@;!$EQt%>^$Pi7xe?-G@p#o`00t
zRfcNsyl4@olL;VOBQta36HfL^S0=kz@~Wld#MVL#8}^nx&dMD|$?CWyp<T<#fP^+N
z<NCyYztGTXhYs4iYIUuT*&iH=mevp!Qoh>EE=+q;y{+;(!3xSZz(D}MvhVBx&!w-W
z>d^Y=REPhi@eoW9kR-JcdX9>DkSRTp^(j{tj6d4Kj20$`AP2^;@8lceL2(~Gh`rrN
z^lfdMmsvXQoNe<ZGvLnh+&fO)qrwE%kB?Y95<Y!t$+5pXH}BY3POYSs`fhsba6tx3
z+1~;c98ei2Dt*9oU|7j$g87OQ==6)e_6N+*Gi&Q+0x;&#a`X_u_Ls(O(*z`SI=ukP
zf|V$ZB=+9de<vtKqh7$pKk(8ns2du=)BOqZ{5dhC1Bsk6uefh-Pg*yWtd+fB#^$i{
zE}iC5x(Pgqc&VFbHw7&W*CsN@eH`09+}2&h#=4;#MRz31b;Yu8$*|QJ3dt!cGgfsl
zCK8>_&zLrVH>QqYBr`vmcnvuUJEaTxyhJQ|hvTg8cVP7ldcnd+r9YTU<?QoU`o@zM
z#GXY8WjeIweY3}hSAS{t?q^fxNdicBJTOKr<tS)c!w(#AUKlJ|;JQqzGB~pwM{M0n
zU?uO#_&a>tk@!jtb)^nUm?DtO=SX$Gnjs&Kpzpy7@RQAFYTz)5aZJOZiq_pNE7^e~
z*r(3>Uu?#_JPX6SeKY2ACI~Xyw#3qIC&3zGC%_LPiN3F9*%Q+sQNV2LVKx(&sQ*jj
z_~zzC0`ohLvw`o4U#^rS9Q6jz?szhW$)DJV9a<d27p%|5a>iCJZC5k*FJd`>R97dl
zxWT&MBEXTOJ{CX}+3$5#dbGFt^xZqGI9B2rAygd7)PVzV+08k6Xx+}WZ$vhQ7?zBo
z^n88N9>*}Vvb}q~`RdDZQ1u`%T1_xfi=XUFIwy{O-wv(-;4p%lN&-TJA{Vv^g_o;_
zcDIpQ@W=43wbY_l<$uBk5Zq!7BnhfIU%MiHz^~)j={_l~($KX8hIrV3o(S;xnG6uW
zCnNZ$u~XuT`&;G4`yM6Z<3MY+nf;8UHjCA<0maG~vk+C4gH=z8nK@?Q<*9Ay#;FW7
zFG6C-WKRr9rIk8hSdVDueg2n5o2Qz>B>k6$v>oOTOQ0&v_cF)84RB}LoK7IM?p&gb
z5#g8GWO2!c#U{Ob(DfC3o>a$-d|@|6!RejIO9|sHtYDZDjKPG|QAvH^v6hK#iU6zw
zFuYwMqoAp1Z59!1=5nUnb1~It@e;?cXKSjw4xuOd5|tPM0L50Y^F7)I*;Y-W4)X1>
zkeP>pBwa7<wIxfc3}F{lwTr)ya$}bHgfxL5_)_n4Jz!h`N7&e38ZJc)3*571cn7M$
zx)|!}zby|4_Gbv??b8=TjW3Uj#ji^)lB?<UXp)L!h0p}TWT9UL%+CmeLK+*lmOm=Z
zSx~jIjGMkB7h`VuS|sUvs&A@^N|ndGQ|&fK+P3MJrxYH!YVz$y!F?iiXD;G*|I#G?
zd#?P)^=oL>yZsLFKc_!~|IceCSJ*C4kMRHLb^pJ0_cnXL_22LFU+>HRUZ4NoKmVh=
z`0x1o5Bd5(i3?jTNy2>Ndz_=&-psCL^6Ek#=mlaMT-cDMbZfCU^5?E&=5x|6<sZZi
zsu`i$(q9@qMH=Ae_`8?-;b0=C4T0$sdsX-+o2^q!EXqP4jaDtI<cNb<@aKnW+R7s_
z@2ur#w|k%etOQ;n+l&V=v%m^E^5sXoG^8rlIjm;#Ivdr>fuLQG>^sQ_{G(^C;*gc=
z)jUfy5eC8;q+-J0E<UjNqg@yKK4>N=sSq>nHWr}jzZEOholM*{{G6_8uzM%Ym)HAt
z3UAQU0yb@T`(Ujbq=J745>9ro81wwT9Y+Jx@f9?>?M$)rWvtWY!QiiJg`XvJ3(d(@
zqu-1E<Wl_<8Ke-}jtS}3@s8QpUsPr6u<Hl2h`0r&R>B|3t;nt%zM1PR6)c-+o2CsF
zFXhF*Dtu~;zB5eo#L$($E4U{T;kkR6gzeT+YWPMJe3SA<L*lwr0A3dFAsF%~4Nc_v
zqK5>^KYw2=h_W(4xuF&Sky?t>`Fs4?xxo{n<7#PzrvRTWWJzNC#*k)P*5{1ARVg)*
zaCT?VeLL%5n(m32Pb~m;xCHS9X}>f7UgI!R0|>Gd!hF+year8nRyl+tg}*R3WZujr
z`Gt#)#4RbBWmmh(I|Ef%${^-M9wBvHzm5}|01_stW;<&k0t3E8=N}wl>~%!No<x*3
zUG%OYM@wf;5rezxwO<b(1D&5c`~6r?4B;vFCD!5QOFjiW*G_@UG9DILI5^n#R@hg3
z>tU69yi>gQH8HUQaP4ViA`kIg8K)c{K*h8cgj)z}N0x1Coj{V2=!jSoGFYMCbyTE&
z|IY7i?#wlY+cxhBu5@D8LH;S>r{l~xgC@du;~6ks4HLJ^oqtiU2s0y~*L#9?(}Fb8
z^jG(%Dt)+nHi^zg_K`8!@`o!MBfTP34V2Q8y<+N}SLy&$vx(u>xbO+){)8x6yp+PY
zLE#@yo|SUivPQ;_t}lMjlUFfV6l?+L+Mi}q)tz`BIa-@eNLKha%7%yf`J{b^cUz{c
zz2IDRGgV#wcBvBHQ3)3&hGkQA!L-IYoJ7=?dDh~JYg!3a#l^vdnFVZpFJI&F_7O;G
zRfku}FzwpiYs#_$;Y-kHv@p@&&lC<r@3emEIN8Rqu05mXoyTYT`|Iq&Oz+E@!W$8a
zkCxk$uR2`#$u?>pAL7`vod1_5$Pr@-&hk`7Q}b0VLUVA-m^4q=a>zAIctSjmI;0N>
zc4AbO3djZV9Q|`k?IHWnyZ*G)W?_wqNz#|aCbEl$(1OgcbyrN<$Gj&O;9-u~O=ajA
z>R@94Hk5#7ghik3;`1g8K6M>cLc6v2CcH!qpBv;%4sVozlsPU3a!KE>jPMDJY;sR0
z4s~e9p+$%qYW_}w6P5nCP(ItO7yNhpc|$LPZ4On|l%-nLA|1KHgPDKwJRtqNB^T3n
zSFD(dIhRo!hW^rA*r<$?(CO9Q0+JX`Yj$#udjL$oN#z|aS5nGAVM61I{C;OFrwsf?
z%Gr;hLOYIo-L0|e70W4L$g;BPLy5u`(t8{O|Me|vE%6bl1K*cN2Bz6&Q4q>-;!k35
zRO7CdfqC;!aUDLM6KrQe9U>Oxx;P-yQP+d`Xy8i}w3|F+I2SnL^8Wm-i6=HWzy}~*
z_r2(P6LX?8J$*S`o|G$)L71e>KAG^B7`m)m=3((0#(v<RzQcK=h;uRxTW(+El}a8_
zP5x(}`)R6HfMAT{0ofcgcW{FBs2MksuMeCeysbPgDsU~`7hbQl#|yrxFFPn6r^<28
zRMpdrS#46-IM3q6V}q1;UygPbdcSnhA?s>bw-3cTT^XY=k~nSN!a{J3nU%i{0-k?v
zo8$PVi=y+u;}yL)duIxF-T9_C3CKk>28UpQ%eJA$s!h<DrLd%3kqKy}w72ArOjTt`
z5`Dzo!u|`&H+pJRAN*3P3to6nP{s#E)@xs$`ym(~Ssh+MRzwuRtkxcaAXP%aj-5T`
zNlW*e-d-G*_3G(snr*-YXlsuZkh9R9{VnlNlo6AS<OXEc+C;t>={;4IjdqjAH6|8i
z{n<nLG}r4}!-OIiq`1{9s?+EX!-v{L>0sf>f5Dir8>3`L5Y8CO?vWLC7iD@hbEmuT
zn$uvZ2;1w%BS8s&{*+?)Eb-D}L81bJVV%eQ4OzPi3-%$|^*jg}+SZgX9h1(wcm_H2
zfI27kUU&U{(pW>!r%o<=7&%PUHXBM@CSe$lW8j)!M3iSHbIUw~EQU%PNr5ZR9+~a9
zL@b@lF%&&&>^XXll<n*~fqQ-uaOFJ&uY?NDS5gcEg#R=L1J_Clrl4s%`7)#=oOaB&
z9_4sApb+MH?`DyqV$wqiJy$C|w=4SG7m5;Df43y#zr7&3x6j*mAy4OC4ZV?}%E4_E
zS1Yo&G=$8n{fK$@INcp?dNXD@O4Bnt5ycQv4fiK-#zyC|FJu(1pAMj=%(;3aLZa(#
zUba%N?dR(dSD}f}#}(W_o_`=e!(JuR<8mQ9v&QHK5~zygO;+qQX5*alwY!?u$J!*{
zBf;_DJG;n-v=1C}LqmI1v3?;8K?VbRDEIcod`1Ihe=badwAG<?Gor^S<H;yo7tb%Z
z+OBSsV-q!Bq!zUt81_nO(qDgW?)9Y;J-_icBwylghag9TFlW@EiRjS`Z<|p&j`wNo
zLU3b-i0kJFP5h&GDxoXmUr}P9eJvpO6osGskQDxPv!w`qrFEFaM7=%B#}4bMWd3;D
zxuR`g1S;;X$8d8&%el!}P{|ZT5?dUKTP3}XtMaf*&i$wpgx{5aT0dbvU6NvkX>4xv
z5J7%$&^MBZ@JXzQsWG1ssqUv}w@;oOQqt1v;-Miu#02P0R!lRg6+ab@x)97?8b@}g
z=36GmpJf|Ys=w-hg(b4CKK!|u+Qr9go|GJ20V~HFyg@Mt6o|#aEk2W8X+8vq4>UMF
z<TZOqhS$93f@&mlYU;r_zPQRnAVB;t`gI8$r9|4EMH97d<y?!1kOcpb0bQYxgp<w`
z)<wG8MTpEW{Us&>wjIL-za8^%A1)du>i)gai8V)^x7DNT13{tWA=Wcpa@Oe1gDz$5
z9nNF-*K#k9`O<gyNthG%>SI*HORJ0&Z(x@mruF^l>=57;QM+@9jwEI}CaIx^Z)4TL
zjL*78RedSzfJn`{NDEm%bD-j1e!T(>&Q`>!=mUQ5e*JN>ZZT-4Z3kPsaU4{m=c{G!
zG0l75t>5Q9vxap)MS;Y>ly~U|nrBpjQP^@41%To-wL7NNii;QC`xXKfcrkuo>^v{|
zBJ~5G9c_-u{ajr9x{62;qz$_T7LX%FLZLG8hnjTV{T|m~FZchcuYG-|K5);#|G05)
z{+~<2k1D-HZmG7%&^8Xrnj0NYRPP<MOrH6R6_c|!&h10BjJNFWWf9*!OX8l<x4h;!
zX#6Zdg9&jh)a}mKaVkyX%D;_%KWa7*?xrgKKikqi=M8X-rv83w`7aIC)$l;xfr{z;
zBz{HwR->K{`t#P$sbKF<RqX}D`Ssz;n^XaX{}sOH|LWgp{H3v_qOMcXh(EWexT~SB
z9S`!&fo90l(L*VK{@5YEfl*rV*sk=pxoyRKHdsOW`@3E~mLeCcu)4oAx2vyHi(4J)
zL#N|Wvqx_TQd7w&J={NyH3V#u^VmmJQ!@#H2k#1DpX%?Wc{iTCEU%{%r%jV{L!e@4
z`$G`Gslzd|Dx~I>DI9aSB*cFU4805sT)>*Nsbwb0f=tEj<})&UZ_vMaTd?Gy+$KV`
zj-lgBWKOX7p~VMamp_Sda1@;ba&#*PpDC~LxhpPe@sZ-bk_<Xdp-~ip4dD)+6>Ku^
z1N1PQ+Bt(X4hL=QIk7dIDSTAmHCGjWKssL%+ab{-tHVYGgb5}|M=_PNwhOSYQ(dMc
zkIfwEn9yL)EveTS`>1bG%S|tY`na<6PolTa{1O@mg6-G(scmlq3b#ka5x5XKqKe3%
z$_=<z!0HwO9Y8@ab2+R4w-yCA_}U@~2`%@?nzG0QwiGWv{LLQ?`+_nd*yDDSDi3~Z
z=q+kr-x}UXO$gyS)D&1)fe6uD;L{oY<8YyU(MO`MGE58BV>j%-t<|CLDBt<2G~OV_
zOEKjp3}HIvRUGptDR3&HZ?Q-9f^gPt!FS4!-ZLjl<enFy0861_KdkMn3I(IKQMNJ>
z8W#Qwx4zZp8^lR<Rg**9o~F+Y@V$4ziuTwHZqOn$TDv@`$orlXw12JIf2307%%~>n
z@D;pjYPDc~HRbaznxAB0jzYGg2;E6R0Z^Ay7x|&m&RzDxq-yU9YavlSHo<%RdZ6`J
z0=c+6p)sS8gEL1l<GMs$h_my>e_>&9q1RvOK5|v3lf{UJXj+mt)pSJ-5lz-Cbwaco
zlzPZIk}TST^u<8_vDGfkS=|%MIYHO>X{>vB4OOfZ=LU#lJ*P1pt@8%Pgtk+sX>f;0
zfmI73u;f$4<BLF~UeQmYW8bCP@6HAy_uVhQg;d_*Rmy4&Q$m~t;m_2a1z*WRjmyau
ze`!v5;pu-~ofovH(1NhjSQkQ2%r>H~no6rc<w?_?NsSk5i6zkQJ({2Q>1i&rU&Ou^
z>=y2-T!b<r<2&<xrk<ohk9Os`uxo#?t$gmP)v`ojCnTf`TZ2hNS(a>_x@d{SJQ&BG
zSr{N&qE(8+U9X$pvs2w(y>RvSn!kAf7P-2Qp(oAuAeiB1OWVP_3XH}kKo-)=ep`c@
zJZ~9+&5p;7pj%LlSS4;t0ws{5Z^0YKh4EOd_rxZc0ZVavtxorjF*qp2cwei07%un`
z<^*{H;nW@Tk83c_LSfR|M^$)s0?$VKo-)f01$t#{IUE#Rr0|;Hd~v0uP@FYkZN_Bn
zNY|sWdu&Mf3Q){#U>U(7zw2b^d}a!Mg-oovJyy>MFde%^)FlNU%mV2)N1Ni+t84pV
z0gAdF7w>}}yZFREZSu~;b^*9y_#wP|1d4yLnXLURpq1a~XxDVAqO^67W4iL;iLai5
zeXla@4aX2Za;qK7YNFjVB_VeYxzfV;bk2UliPBPfO5FzXP!rN6LWh10haI9L<?8S0
zgr@4Qw#Fzh9|9D6KR&daOW?&7NcBVz-_4vcG|6l+?+Ml#i6R7s2VJRoncec%V#7xb
z`{WH<)-wwhwbZL_y<A+1OH)8W6d+*n{C4`;gJH7HdEr-3(3yg!fn!#%q_;_HuY1Xi
z+%KZ+h4(Xnd2hYj$1bq7&Y^@b1ybeWT}AzIsrn@8t?vt}$nk@yZ>?_}`&~95Yr!x3
zk{I|T#Ep41?28}Tpe_BprvM{<i0$$rh=O>Y@;!ZK<5u|27_cDgbTc#rFB1u5CO^kt
zRoo8wzQ8+q*dH3f4=@{*+L3nt{W90pUs5SwpP9eBrdp7sz64|+z;7G>KubV)R_#p~
zOeVdm6hc7HZSYOt*hH4#bnSOPA=DAQs_#EA6D6`vU_u>Uu(Q8B5_c7@^LYUlX-O1}
z+Yz2@2$fq636YTCCDKPdJ9`E)3{Dk_Gd1Aq7W9xR`n)7&5{zk{rXUYCT?!P~u<2*`
zWeoGsr$bAsDzO|X6EDLV_LE8;($IC)(BdaPd0(;>_|nn#B!w@C1E3*i?A^D%)ztng
zsr9a{Un_`i5GA$*<tJ+r4N@A`P{WE+ig06Gzm%&$cR&*=q?}Q1x+omAck8#Sb;rr<
z3HM`fafW<vt%3x@FxS2HYE0W-n)I11BANMHS#_=?!jlndMa=9}3k?6V$>Sn$7s0HR
z3tl|8v%YB%aAXT*Dk+1|=ri6+OulK9%{W#^!V<r3;fCTh#Bc*z+@re=@p~K9p9w3J
z*5xkf%Eix<{l{<Wo)oA)`^vT^Qubv*Hhqg?P0ssE!#MV^Zq;WKGf+K%cyx44VkLz)
zY(hYZkV4(b_$!2lOjN#b{?>yOz|`%g(b5=mRyn(c76-O~nl#g&o2G_b{wKL&JZKS8
z8_~j`iS(sxwtw+NK_<qx<IV&s2ZEhxaBT7C47!#q9r*9Kuur0u(^oEd@%veX%t-Y<
zVB}-`%RuUQtMPuziTsAR#2S{H{E`@oLfnAs;Yj`}D5*s*=WMgFLt~KB+0%g!*6Fcs
zwwY2jY(Kx+sUn-8Evi)(OH^0pIGzJ7^~jE#ofKT2a0f+{aFh$VJ?346F4&d>)u!*=
z8P_r8&9ZucQDbR-V$;B}TnO1F)n|}R3wK2(hbVL>g_|=HtCh+}1yc<`PLs=7QC~c*
zI4Gh}rvB1tlCpIgs%a;zUvcdy{Y!(T)&E%b`7uHKaU#l<$ojn{x`y_YJtg7~8CvSG
zRov?=zbVJ-7`Q<9VeaGytAAR}O)mP`NQ@Xw{F+bWD46I=m9xn{$cx!&EFPI%t138=
zpyXgC>a7O~>a!Axwy=C?PJ6I6BLWASVH9-q{T!MLC~&45BPz{qd`?*D<z{SMU369~
zz?~s@;e4Y9^U579N*Rmg2p-aOWqmdC=19U%e+G^*m(;TON!KKY0nfc)Gb87V=9exS
zvNFF~KolrTVSs7oNtx%sUQz6FsDU1wg5^cMKtmJYZP!UWGVCeW_{ILzAv|TzU}FQr
z1%tn}!-B4Aq#H3GzE|ucOOj%Ua+`~!hk`dekomVJ;0E}R*uzTT<L3N@zP@hq!{;_y
zzj@4^T&{k8db3+LW`^M=s*zF;*h@!5pVF7I2aQm7gVFhmsihifwC%6v5?i|$)S&dR
zkdEPX9<g{udd2$j21fIlkXbHF?M~JxP)H<;BV!=zAJ<<Pp`F1!38)Pb+ZhVzDj>vr
z6t=fWw-%AYX6+~4#&0)a8uD$6aukq_5_ej~-Db?)d}eiT?ez|bjywN4I%nUZ>SwPV
zmuJ1&Xrw&Zu-|pBHv?*uxrp@$7PNdn{%56Aw<Z)f8D7R<4e6FqS~CsJ{if`!GN{@O
z{vMEWaR5Tt#bXi|Nub3BupkR5dIg1m2G?~?-V-R)I*Vof1AZ@LI!iY*b=yE#-E+>t
zwO>ZN-J#HTWN)j&nsQkHN}|J#TSt*HmhxLa&ysb(_7E*I+%D_7-JM($wl}?(*T=|H
z8BB6Z+vyNt4_@nAi)$-+ynqR~AW<A3@UI01kVP~zXbF15>7RU_w~mheB@6$=E`aBI
z<8JUb?pUG+qsF20O-2j5dTzVxZ!u!0>|ClpjMN%OjVF{a!Uvx5C((j4(5`PJV3+mw
z=Q&_Q%o`|57h2{(QnnfKf_QH)F*VjUEMD|Ei(}@!-3i!B`ko{>V}$KGJ^s@VV%#bJ
zB>m-!XN!hi^1OIZUCu(NKJA`Hb;A{jxSI}jX+jl|p23l>D!mG*(X&}#4kCZxCa-a}
zBR>@zDA+_J<kx2;miT4xm#bd~3<!3Lb+l;Zx#RdHdiu|^ak<RjW*;|NHzvG|<@`N*
zwYEnrOSCOy5!;QuL0amXoY5~W+T>KSwzKf0vlL*Os5cH#&0G?6b8mC3%eG~Z_yjzZ
z$~=jtZLaK}0RFqbgs`}_Dnk{SkXE3=C?bcFq1hh4cUJ>S1)(|MrbMMYTrpuD6Qdzj
zHUmO7w$xi+**h<akccaacM1JQo9sVRc^&O6eDNm}Zm^fSL$8g&>h(TRJC5|633O3U
zQXy`XaqW#IfHKfyKad&AzhID-9(nHeMBI7zW$`37DgA5*<CvF$qIMQ5bIJb}#WlMJ
z)&r-2?xEP6=(ZR8$mnOo4N#6Pts8am5)3gd5ysD7J!?%H_--w@9`1DG>ysdQPIb-%
z0&_P;&;vnB5uiue7H;zVTHr=bNQ*!jzyd<!RYBSGYdWBNFjo4=0O6*z4!pu_q(-m7
z`|VPEE;*ADdZ6Gx<E?3Z(nJAO_dozOcdOq@V?a_bca0Ld?8Kts+ThB~d{#n!A@72L
zuwkCDdHNM(R)F{b+w<vE2(N8ktm>>IACMjO&w}izhB7{ZnQ_Z*Fng)#z-oNEep)Ej
zKr}&W69DDQcr^U$Qm`JfNgT+94W}NjVctpA%~atog@@KARaAo8iH!1cD7PM*Dh4_e
z(%k}$JOHqhM*DTpV~W+Zo&r+Hnqi201lYaKILIU4@J{L@5!=11dV4>A?9&Z6Vx!=G
zL@@xV5+}n1YusvYiGGil;`mGRl<_m@m}B40_>x^X(}}c*tu?3Eakxbw-Oo0x#Olf+
zS+$x<d&F#fOkY73Y=9}?)5k}B5=%1NPNXwa(pnIRGuD6h6PT#fht-ch9mIVa>wtNK
zxH8Xl7CyPqcIR82Txwj|zn_;_(7=QNyW{c%_GVPzh9QV5TcD>MAUZ(G?Kh*`y1jgt
ztLMOdmh2=n=xU-1OJ~&!wJzf}j0m+m@Rug@u%VAhN@mLmO3|Y<Qe|MDm^o=+xD}_h
z-Kts053+T#9}%x)6nh34J}>|6Fe6!O7l&$o%u4%S=&gS3Wv~0TG=uu!6rx!7P@Ev6
zf}=i=HHE<wc^r0q0uCrn675^PS}9hWET*)7lF^*UJM{Pw+U$hOi<?$?r?CLA2i*fu
z_&+;yDzhOokn`00pF*gx9()!Lm|^Hb^G;QuWPF$Vt3@LPS_6WLND?*^+bi+|azDJ%
zX4LD|{}E9bm1TNjNhgjfh}BYpKN=%b)p2t=XE#ipNY?G6T{xK=HuINC=dhdU_C|Nk
zPhb2qf6Hy}ZR79oOa1CtiI--M+xt_8CSfxA)#3xdk@Fqxq71U~<I~gllb9ugZN?>;
zH#OiXhOe*?TLhy`kLFQsG0DOj4Q{4okc<Go<Nq{@^uEIZ)xK&r{ZoZAU?_pAdrJzh
z@W^K$CCis=s-+}(9f*y~NV>ix_I^z0(7>zS7!YTB>hzRN?{PJjRS<!r!c#I+E0bNe
zl?Nl3bZbP=(&*Z#g7vSLVdHA&l?BoMA^jdW;+<_uvuo%VkQVoNL+&$${!d;*?3v{X
z-;JD<PKtzNQ;k`g*A)tqoD_)@$cTCqVWUPuN_@cY2IJg>YBrClh`Y~f@>L$kx$Ht^
z<xfT)#qSi~Ob)u)-*A|WnWRR3S<wf_;#D;G^gwej{$vWrro10UWisEaJQTb+O|pKP
z?LhoTjug#t|HpHA;JnO?wO_>7#Gpj9zSi-pHHskVML$lvv-*n77LaKhEx0+g8mfi`
z61+?KPNA`)9M!i|Os$etXCBsQeBBdH-AGZ`xs5kp!F&#Dal&TjqGQdEq07tYHSk{=
zI|<LBuVvNQ2hbD!KPY#KR9g=vRnv>iPEyI|o;=`xPPqaXBcx<Nngghan$t<nTfNW1
zcQMnZFQ>(^ZG7^fJ2AEfS$>fSX0qTOU$&&hE}|W~8p^B(w^-CpcojsJVHCsb#5hI4
zT`T5DFE_y{%$~B;>*cS0fiyWM4WW6Y;YDY!7%?-r|C0P;l09^3mzk{JIE-eHTu^Bv
zI$<~ph_V?*NDg@_A!<;#>Fhzlc1Q2`vv28v5N?Z{*t0&FmZ^TrQt)&MWmmROqq5`(
z6uN15P7H&9mqW+&f>XK$xnYGrdKWe2J7kI*zu@_O{XyP`Wg<CuO{eF6xze+C+eSVp
z^<Lz1ec3#RqKgbuhCj%)B6KbYQ3*lb5J6L-e(~bjJB3z;*d3k_J(LiUf9+Z9>X4wj
z_%hv4PS`d-8MBDUIV0y~=EP?>%A8Qa0y%_?BBIcvnuZ|Lz02j9MA+?Ao3rC3S>S0?
z1aeD3lXX&*EXIl{pZ6tlEpu)v+zujUUT%W0OAbj>|CW%~&K8xE0ze0M;WAaQOz0nl
z5MJ)6?GMmmZ2b%SNO_T8rsN;bQ^0%|fA<E~)z{Y<o;au}Pb+Uxqq(o=-@*|=j{pt$
z;oZLARI_!)OGn_3MctCMuX;|#74|HnzV%8M%7x~vJn2e&cf>gQ#r(V90;QMw+z`R*
zm_HXle`!8<X<55rjl-156};*K)p~4CWB;>JCJE*q#GWuIhVo-d@bq5!%Wq^VMMg+F
z7V(FlTGM{cc`Y+SRW=%$2IoMdv%6Zds>|#)>;ljqqrv^tSzOLnCVyWX?7Yx>DPy4m
z2!BkVw#jbebFu8}f%#k_7b*c|?OC=`^~@@bLv4G9TV280-PIRJ+1-s20l1JhBi`*x
zU6ZHvS*N<wl`Ea#(hAq-%CSe37w;{{jV`7a1V5SNYbb4(Vd+>@A28b2C(f_Mdc&>Y
zIt-%7&j!`DkEFDINjpEnCp?+!yH<aFr`5%MZfNf^e{gj%<_x%x2&ybMF2+)rPh>E=
zhcawoqfT_g4G_*pjYQF?fHLrm;!Z=c!yNdAZs`DH%a^%hnwqCRzjQyV%w8WLb7?9L
zbq&#Ov~)Ell)?Ht8twl45jK%Cv7=JuS1O7vxg%wNub;~|yj`NT{<R=V(k3dKlc1mN
z2XK~c6RQMdf<LJ85AT!u5D@ei)Z<t@-QYr6r%h(%51?Z2C6J^V;s*KZ)M2O~{xJa8
z8j_pj*NNsvMgB^k%gt3+9Q$V_4%(8hH`MoAjIo}-L+x(K33Tc&jW-g<s~(kExQ6%_
zJg?Op#p2h(y-|YN3Ac!)F(rfwCg*I26wB}m=VuSZ$6|h86}iqjKFQR4rbpHs5hge+
zb*Vimbj5E!-mtSgOfG~<p=x92<MuUvv0tVQHtB(<RcU$dr*OkHiGu~zuqDe84GpMp
z8!<Gi9CRlgwC^c9E1kwzpRII}*m-k~KEuteXNNp@DRb!F;?NuHje~@;!vWonnT2AX
zFp<*Od`+V*L!P&dfX@j^Ic8oX_E!vTcAl<zvS@TfOgq+N-vY7y0VK>5pgoI`_%YR#
z_MSb_cyevHk{H>ZBNnZpSq>}g3gZ~oua7%2WNK!$bA0qb9a5TVu$HWU)%4M=MXQuJ
zt-CPgA2V+_sC?c8e<`NZ#m?Ie;|+HqE2%$iUtFvDp0}!cap{YGw-0zNB?fby3?U$O
z37yCJG}+x21B>c1409WS$Bjm(uAKl0BN6$>vBSn|w*upeyuM?EuRli$))mBip?L9a
zLXKKkr}11Id%+v824*FhgQSOwsC^;b+t<sRw|~0NF!veueE><^Gp>cM5fOV5z};Z5
z(Y=B0QxK_Xp4?nOxr9DC+)W1ZE9^!};EW9g4KnPAN@XpsQH>21qVMVE+=HFW)9>re
zl<yU~uEfC>7z0kVJE7NLk)5xyW9o-k*ITk<6@KKNiw~K`+m}Cipyg%Mm61;Oi_UI|
zavLV&K)l@3Ze5QA+5a<D*fw6YZ>n&UQEi9iFOBTU(0fT|hKzl8F50ONaAlG+{)~fU
zhNB|9^o`L4%&SdV`<<&h)40h?c7O|KYDGy(?!j3gjrFA;x)D8BK09q@Tje5iYRTy1
zP$jvxai^(xgnB=qH`sd!K|m&<)Rz+ORck0$gTijrN9(X5c^sF6WDDO~U;CB#K%JkL
zuI}Eto_q&dgD{lHL0H!k8V#s%d(f;Fa9}N<quPC*7<fLpwH_+bPi7@#J4PO9){^{0
zv5cfFAU!(uPgRAKcM;M>(oXI>ZdT-r>T>)(_0~ZJS<lCFDp$)(z*Da>QzEN1w4Fgq
z<$2c}MK^8#33<kG3Tdftoc-ibzsq}B<SHdl!s6Ifr78iSDY%)9J>GB)sM_6`2unK0
z1`RJTlZqFIV$B!2a6nEEC(-oIaQ_^wV#aBt*e!ZmufPLM-{C14e98=K`zu5^R0Dvk
z1nUtqi?hBs7tjBtarGST<!_yH?C^&73NNsI%nch$-0tE2Nr<IfCqE{I;~{ns{i)8J
z$I4C@C*jJMhvI<`lhxJBvl{l&X~!<9L>@zbCH}eSZ~k&VCYH>M7mNd5UmJBHX@ha<
zO+|#m!s3aQH+4?-Gn+RCG<5sifwf&cKYxm(%6t2+-0NN~tnfag-it1RDA$hmv3l^a
z3E=6d_Uhr@tfS^}J6e8NBcL$JzQb-c7tb%1S7^U2zNh|<LFPKz#idGTQ@+d3hwMf0
zE+|^BgT-_^-K37nsYtEFvrXaBxf5sKg#DAU@;<Ha(ai|PS>8*LmSKD_?Wcs@WVD3S
zylsM`y8UG3uK$tsLs6ZQ8C~IRi%cz-Crpd<v-xj$2-?h)LCSq335e%cI_@ydb6dh)
zu8xwj+>x42{)q9GvyWMLz@2r`{x-9q?2NCb%yI$6XcqX&cnxWQdd}^r48d?@QVi!$
z{k8JP!bD-(cJ-50^;9R+^`N5`;d;A@$(cmCJCS8~vI4DyglJS56Q7OX3gXbr#FChx
zN>WHyjzC+X?atYg!V=H>bM_q#K}sh#<E!Yz#KH$)$V4;;tajFlcxNq!EL-X{i;2?E
zE}fZi=mARtG>Z>A>qLBhwri`Y6j?<Tb-n<|tbX?jKK$3QoM&E#sN1P^m#4V6suYYH
zs_R|a+U8iWKdK!U&9c9j`{W|6A$wt*ahBL1O%1nYx#z?ImBRr(IY#1tg$*f}<bP3z
zSG$q(6H3ROxJhk`KpxwLc~VF>nt!U&>(iOZ_^Dg@<#rtoo8aW1%i=47t15!RuXv^i
zQwjAl+#B1WxN|UI$}97iiC@XQfiIWO1#w$??jLjlc(frA8Ur^Z_TxFbS~bNwj<x8f
zVq)M*CdCsI3G;mfPi1QXhsv8x;dYOfW<Dh*)PM4=&#VTLE-hkMKxlDdP`rjwzI6PS
zWklZq0w>)%vao6yIjBzU8n8#v?8isTsefDidf6e(+KyUpQh)r}+NPj@F0oSzouzUo
zRS4Y8c<2T!*NFg)ffj`%Tq;A?(8$fsIw(9B0MWW|Ej$yD{>R+NTI54+0V`S#W{ne!
z_(%lC2B_Q`5o@V$h+pv*vP_*bN-Hc^|C)K>SE3H?=5KfQW%Nao44zRN9~fB<=D98g
zs6`X=e&m3jW%P5NzLB&0b|LOpYMk2jG7+UwANLGgN`lv48fWYk(qoU>c~ZGv%)|6h
zA&GYy-AX}OW5z6xMz1D#MMbQ%XM{H~f%NBUpl8Sf2zG}$2O{rQI5?FH^s6On<AxrP
zT!~v9+BX{*w(kUJiuQB{o+cN9IARr5id*v;?7u)LqcuKff_?Mtsq#rZ+*Pf{)b{P+
zz{`ETF_dmGa7)P^JyA;ao4eD=`^XB*Q{(#M7GE+w`yz#hq>eY1#i>&Zw2}o~qJ%Ax
zf4`!U;%{YrudKpPULv)#;%M_Jhu5>4Q7Yd_!UTgJG$&E4<6OS84u;+t`;p?aWZH}#
z9y7?yp?^N_{mnkK1$f`c_|A_MI!Pye%Q<{TZ77CUl&Zu&unoxykm8hf3<#tSxfEqD
zW}H+WBWy?X*oFR|_Ni*ue?6j7rHNREO=$n&cmGRM=w1G81hLFlu&Ygp{y#5dD}bQ(
zt^VJMlVT#GqwP|Uyp4nBWCriN24-53fw#G$YP;jL$kTi}Lw1MDbDzG=QbDxmiUlW(
zFn3BjGZhZVW*;4<OCmow&r`s#+Q8%u$ArO}VvmWprxaCrpSC8SE-3}wf6uY)FgUiG
zC8?Tv|0xb_Dp3Xu&!-2__2SZs4A1lKOdUlILaP|}9V4!`DCfJ+Xb@l1F1QhBc)v-Y
zKBsK${iP9MwkmLic45w&^Iu_U$joAq*bYFkgW<E|4h!&mMfJvUt>0I>hGr8np#9gT
zp8@hdW=M4y6)}kpBQSV;+mQO0U7G?JSNNs5-C-!g*=4f!{&ogD-N?+gI8e@AEYMJF
zc)jqHEOi3f_$U)lf;3?M0%s?9SGrf&JAfFGmMsydt2R)vc-N)~E?l%nu>mt}9G_X4
z$9n+xprPnj@@0ZU7`G#nI{!m1j>85)q;4fpw8;DYHHB`{V`<~J>I~d3EmpK;*w>$$
zoreLJ(aXGc^GwzzBp@zH3`eHWH;#_O-0%w_wtrS~3?FYOEy<?qI}6ukza0E!o51qq
z30SUU5g2iB&T?jGLT$vTZ?#ICRX2JDS+fjietd2JXtY;C_1=+(cEalPzPp&_Hrvk0
zhZfA3C25oHk}pB#hqV)R(CST=iS*K$?l!orp=HzB_}MWy2`S}r=S=<aRk{J$?7qkk
zuff9jm>+j3cc~W&ZzRIkmnr4*#%U%BGd)wekoQWoDMXi7PpVEh>LUFm_>%hTj*GsV
z_2x7ER5pP`n9c-xt-w?80{mu~M~i5-!25tA`}J~h8{<!#$XBkN68ElH?r#?-Td~NM
ztGKEEzT)+5z_j4VjpXZ#JLA8xZqZXR-BZ@Zc7GF@mcf7GhB63vrw}J0`=aU*?L8WH
zbIphGS8}b8x~u?#Q&KT09`b5tsBgmUTcmMFisB&TQ?S6@qpaDz`cb|mu^^E}=M40D
zKAExX&M-vVv@w0jJlC3&Lv8-B;^^4IeJ~{W#e1hY!*oZXyH$pg^rCG7@d7QFEqltV
z(~OmKj|vQ<rh@9vFn-d@+3)EjIVn*tvaD25mK<o)M!4hFbD^3+yuBaCiy)THjEo#A
z0L;1Aru`3k)PjoI51#JyQMR^o@+La#03NqEMipv}28}{<nDHl#@af*65IqKi0L+Oe
zdadlJ){*JT67`F%LOrKeg_1drJx4J!Q)b|R@v5vJJ5#*gc0Y}e0^i+W&T*E>q*JAt
z#3YCDdI;Y9SZm_Bur*kd+7AjT1<rwv1ApmH?(Ik|CqKQe@~q+#6BD@fb8@`qjezF0
z8p;rbCV4=LImS6_LLnjVx~^vDXz8@HU}$^lpiZ#Om0-XwUwCtpZw}|5a@;P-1pfRh
zEQ0<}Q^pKc8^X)HU49g89}>draS`~y<>vfNk?S3UEXy|xdxhF2<2jP4y~AVj?lBrS
zik$-N0bi<Uv%wh&z(kbZIkFriHw4>aap++@|HXmq=l=C#b4HbeH?{4}PZA%K_+%>q
z(H<sJLsTrLych1mtCz_Hh(_<hT(V#sccP#A`;5pjz<iv=bsoH0o^sWbklETjzu>Wx
zLN#EHqF;=}#~$vNIP#7+8!ug$@E&-Z{fNf?y_5!4$grF3WPlr92AQUr2=zM(Akf7r
zo-~B&>NjQK!M_7krIsdSgzCUQ^Z2Drz8c-$kJNvk;kJ5-HuaHKhMu{%l;Rk*jV053
zt^bsQ!Y;Thz$}c$H2@{B=n$a1%@|}yN=h2uBj8~1F0s6xztlHc`jH4^iCa~1=o-jL
z1U2H_AOWl&Ds-ZPE;g8VTgyW$h8*O37tn@9TwW=;DnDt#X2!^Dvi4fp;7v3gNeblx
zH}(I5sUH}|P=nON;rz|~h47bN<e|Fv-E_OCwiBA20gM05&1_}WfT`5(4ekGi02Z8J
z+VT+<@XG|P2HZTeUWP$d-;l1f|Av3w!`1f9S(uOsQfW^Fi<%1}|D{n?h4kdOHJ@pd
zSb;Soe^y3ZzadAdr+Xd5+F)GTA1|dsB@m^~_Ht(lTzQdN%i7DAbD(y&mV(A{sR6?M
z`57D`&QHc()S@A3Si|`Xu1SzvARex)`+;{YB;S3fqvNN!l#oG<dW6!aP(Q`w8qK!D
zgO;SpOyxMmdkW2KSr(FsM^ZkJcM5N6bOq+}GdG`F%wtb?Pd2r(HYi5@J9e)Gg7E;V
z62<p?DWb^C&X+=Ctxp<9NDIkVQR7N`Ltg8X`zZQbam4`f`yzLNo^J-RS@n`=CXzZ}
zoa0D}MJew|5h7F}y%6DHROb|}|Dr;~-&}|G`g0$8_URGh;ubJz_xO+>d$@v}3$DsO
zQzg+EjvXf4J;?3Pg;<+p>*pylzBt<Lt4V#nUHNNFwlOvlp87{noVZt|_Lt@cOh%um
zAI_BUW%|7IaWoh-%B(oX3>7z_pR3?qfu$b)=w{LpYFz)?N}-aBf7wIMAr#{s18!GB
z(2tq0kH{L2wEE2e;j@jROy0AHmN^G<^6bxrKc`>4<;?K<Q5lLFS?xfk07FH7mTn@H
z2m8mjzsxf`2<wKN<Nb-`WNC_!j5?pQw_b75GE2tC8=GLNH+OgGcA@71DkdR|&<K>t
zk_b`1un^VK1QW^bZ)=L5*;s@OYj)K>y4?gDyrxjr6<Bq3=DzaxzciPTi7D|+naW4+
zTR2w*w;2k)<$k;`ZBp?5$WP__`EfSp7XNhBPCQ2z60mgu^vMDEr6WlIoW0pIq0z<!
zez&DgYFA-{B`<sIWWj$9qmCX^?t58?pn2w;9dq=Y6aoghLcq@}Y<{?3s85na8Gf1F
zMJ^0VEl>Ms{AO$2jye{id-nBXnp3%+giv$dE@JbTJR4TAKZ|&+b+KWrQc`F*(9X^V
z)Z!wW&`dFgXd$!)HBR1ZIm844q(%~$B+Q|2DC|u{hOKmveS`jlg`NEQXS3YSdL^**
z{2A=^+u5WxVk_0B<twD`c<XfQr1+SD!b?D|lzwWL`GH{s*-2-B3PDQWj6ba(t4Sv#
zyTQ3Isa2bf9uYU4wcC3yemXZr-MZgftw54L)7%ocNy>J`4}bxq^_HUt!3VkOQp%lb
z$qlQ)qYr=e9mhvcvfZG`8j051=h|6xU-!Ig^2HAe+{aLh8_L;knel=ZViLQYEDDX(
zy4ggmgEg8cIaetz+r_;F&c%j}QDsC#79M2uE!<8kl&TYf3<g%+n$WsTlCc`&=sK7!
z*Gn;LnRwAM)?nPK-sJ7=0+JR`f9Gmg$RgjkVtq}g`#w$8QM{wn*eUh@Da~R}v1IJR
zkUBBD-QGSWT;MJ#1ZNxqk=E4x2R-IBi8qj~sOu8i%^X~N;q-8D&5y<AL&gAri82^Q
zKT^+V3@CWqJkD!5nrnd0Hl3)JaXAyN5<x~77Q08VemL=EbGq>Tmh;$pVL7y#EV%!i
z+<2yp9Hf3Y!EWGX@|~rg+Ki{QZSnZQx#=FiVrq5ZzwPb&P9%NCX68{G&55_I-=MtZ
zA@p_r<;m&_?;QpGxQa?C$y?EbNBEgJ9kAM*y|6A@0&i@wX@9e!!#6(pOY3YFtxvYv
zD6h7*%Fm{ifK&bC?70*rBJ$7nA|N*Vv|+Nj)RPtArMJ9k)tIw#UR+GL8T%uDfqV3x
z%`Ho&bsO3zVV2!v@H>P;Fm~2}dNGMBI4XKIRB#lBSlUwu)D*P_EyXeoG|kQ>%}oLI
z=x3o@r?ukSyBm8dQWn$`{177R1vF`?w>jjkw34V)f7)8t8oj!zpN?B<j((aWE3EEq
z=~Kk#Omoe6H(cu)nlPC3F-aQV-ER<-)gG#B-cpM~22tk;g06qj+EAAD!^p81?+gTu
zn{Pn3Ox_#@h5V(-U%Q7xM1`>uppo7vv0{%bBFTHB%4x}@Ny_&+)*~>uBhJ{{TKdGI
z=a;^0mnC8C2bKY5Z&8n_ZN%i@`;_8`&J3(k=;Q(0z%*LFzAfhdy_Ji3Mb*0T+a8~`
zLJk>^HLCT_-)mgPoilCyQy<eQ^=WQnA7l<vc9NW^@`|b`bB2x5OXn!?8q$1?z_ja`
zI8(EuE`ZnkxZ6T^xvjael{kp>y^xgp$$~9A@eRZ|F<C{J_L5Vc9`nN7zn=g&86#5;
zmnC#JFLIGtLyCH2qLz@+lbV{2!v=hJ9^ZDfk$Zdh&HY;uAJ-HMaRu}nPE2(z^`lZ@
zy->ek-!db#b8k!V3djv!BjO$89R7Wf=2HCB=UhnR!k@S4MiMSATls9GD8t~-^PW~1
zP=S6MWw-_*JJnmzmo7yzzXZ>Cbbne<$v5@tPbA3#XAk&8EnAJJj~J<$N(uXKqT2sA
zQ6c|BqN@CyKKAc7eHP#9IrM+W%-ILA=rB5x%n!;P%KT{kmyc-u%*v%~)8?GXpM8dJ
zXJ!qBb-jRMzEP~Iw+<BzOwRM*7m7%^fT>D<@Ei0PR0K1IhnN$$CEWtjT)-i2Z#%9&
zm)Nlxv5f1beQi)Q^6(hq{bm}x>Z2t{VqGj?AiY<dS2u2o-s8+M_7EAbDz2*yet8EB
zHhBBP`xke(tWY{P2IbefxJ$E+zA#p)aDI&;K~rnAk5Ox3q}n})>r+DD_b<Y=(I<lr
zjK3~Mi+L`6qPktxQZVHyn0Nn`SAvxwwXRz>ATpmvR%sttSgri$V`o`V&UL5DkzKEP
z;uPD|J5C>~y#?Q)ln3$GU$vW@yDO)9jb7}MX#Gx@2+U}xLn%zSiTEUeE~Yy0JliV6
zGzua%NAH-P^?RCTiS7f>J|G^=*+!Hu2$9Y!+G@%AR7^^i*aDGd%k1nnt;`uUg;^g+
z6QTDy6O;HlrKh<PYhS|`pl|-tuvNm<31d#(HP6dE+->5IjS28q?C9d!9aor}OLO9H
zg;|T^Zn@X>Rn8(VDvgk1vJ~Z2Fx%+Z>Fzng?py-T9E)ZnWJlwOOwIUEUj1Rw=@rkG
z=ud1MwNP%8+SbeJF~W5D<L8_`3tU8*-+yUpH4KV2F<!Pavwy5dOx9v;&0pl&?A)vq
zZ`9g&#v~8ahPmTn618{?2t&Yx_ETg+hJlHP)4JT|EB~C#imCmIEuWMG`-e{@(`s*<
zMg%^UyO8mV%sWiBaQ3A#?>RsM*OW&#PInXQTL(Bz(2a?|djq}-3wvi4>c=s*mcQHU
zxb;<A#eAgIA3&I$JoT&}H*^wr3MBKMD`^#s|BK5Tjw!SmoprUa=jVNuK4WN2GGdG0
z_F;PQ@^|X9$V{=a!p;IV4c^7=hYXB}t`?4%P@fk!W;X5KD((Grk*v9F{p`s;U%2mh
zp|gtOUP%T)9gm1S5WMWr#Tjp9Nl+|^ZeSE_53*x5hWN1>?bJ3F^L7toaxQpqH9B)!
z^9CNAmj2nrUxmvo@ylWLDI+<+H;wr$V`8Yb2s}ten6vfcetcKH6ZSY-Wg>YjJmN}{
z2sQ(u(2P-+LbNScd&>H)pmV^Rcnm{p_0Db^I)0prvs>#C^$T#7x}7bWF|mj2Sl7aN
zU>E<&=>9LCIy&Dyq&TzghPfa=%nMdPxmkzjzY9y?#nfuea?TKuf$+RpB64+OEQbh@
z`4J{c80wT^H3%l)vE4a{c$tH|2ViR|Q3m=<YI`v*9p@ryPCzj#I6XDLIC{O_T~dWa
z4fj()eg_aIQs6#j*p2vw_>Xm?^$Z`+Kfbv&{!!E01*ZEJ7v*_Y`0kns2|s^ozG?;y
z!zi9Oo~V2L+4;75th!y51J5OoIv$<~J{_vCvD*xs@tT?(@#Hk?Rp3ZBDwnM3e)?$r
z%}S<m$s;b_DNQ}Uoa}Cy3m#jaa{tosl3sR#KiL~9t2LkyGc>6eko_qNEPYpg9SSt`
z!OchD#$ckw_XVk8l&W&vw*MKUg>iF3Lu&;6b8tz6$0}5HnCtEA{m|kM-i~GDrFLmF
zJzQsm!a-7hGD^NvE|$pJuB67lWrqbRzxrp0RXsHKX(MWIT^PjuCM$gTxp1;ll~G`3
zb1SvQ30!X2@bk8W93r>Qsf|viPy+-}UWc&VT?^O4Gc&DcS#cVB);WN0mS0h7U#!@~
z(aL!(;FOh|%|suF&<DuGFVK&Wg&}zAdvSSiXmPa6E`x*AKHQw9)+~WfY=qtExHRx6
zXLFJ=9%77NJL;NDxICFn<$YZ2aoSHzeS4<iB&by4leiK+oBEZLH~hlu@AEdN){$f#
z!cZ55bE0&Z)1etTNw&RH5CsX~y<6=Pu}1>1uY4U+@NN7V&2L;T?vO8^dC#lrEX<H1
z)v)3o)@WX9SS{0&yH`uSQDn^kH%#cd)MVo8)4qD-v_;pHaQJgJa^BNWPtEjgqj&g1
zD*(H-TBjjM)csL#Q?xCi9Zg+!6R9+Y0D+V*x`D$ik1p!)^0zTN_$w}H(|A$YU}y5A
zE`>Z+{WG`9ucY@C3c1@z9&mm3l#Wn_2ZP8RwAJ?xY2P1TwXeb%mQJ)ZJkSzEj`#Qw
zh-goT@{F052!#GJYSh8DMGqp7o3&@uC8{y?Ldkh$&?dg^iS_3ON}xCP6>?~K)i_h#
zn~I|%$%1LvEg(1C(07RnH{~-CT&v?U+TGYD&=Zvm5?@5JF6TMdztO#wi9$Ycg*~nL
z?i^vDi-X34@hJiT-1x{&{ULza28=gee9;7KF!ss+Czbt|cY*twWbljoO3NqMu9O%#
zSbygXYT-7?JYS+Y@d6IZ26H=N)HbV+?Uv^|W^w4y@T!B}YPU6#BYz6Nk`jG@iN!F$
zL!G~RqFJW2?l>4P<_YVgClz=r)wJcP0#S{5dEPsV&aH#`*Dw5EJLm5=W#!xVz@cvM
zuil?_%WNxq3p`%Q{Y(CzseSwWyW#~*&&yRDcl<EDX^-iO8$a?saZBcxt`~`k@HUwg
z<y<_q%hs+yg!9moeV)Mwmi;ODvE@E*oy-mIbroKhkL+q&F1>E?!`IdOtr&A<_)R~z
zRXEPFc<go{<9cOEkZ`?R=IR>Vjk|tq{Ky{vG5^g&^&_p<^>yBD`Rf%O&wtm(VBP#&
z;X}vXeU$Y&Ei;pg>pRPH(<{I=Y)e1NAMx)nyLEi#x*DYq|C%5E*wQQRcRAR7R*<^l
zyNi!2);;VMaL$(C?EjkI6w)=fQs{MR{jPWad|$lX`sH@i_mK6M?mvCK>VDj9sqM3_
z$a9s>=lG%jkp1v`j@kJ_QSsSa8}IW99S!2&%=Gxpwc?*Wj~*J7^$D+hwA;S5UP$jx
zrF_SmueFjUvh{~|O}z9~+q-&mdG51HL%;Pr&n!+ab~j0X=91va>R!+DAstvUch60J
z_^$EMjj!{6xc?E(UiaeLu6NTPNt(Rw-t9K+%*%6c(@lb&9F|sI#`LaXl0m@yMSo-;
z$nrn@)@?R>!A;XG#$MlrPn#Z@ZZ2+Yz_@f)W#p1+P7JJ1(sDE;4~Q1mY%Gd;@%Daw
z+26hQT1vWJU;ntjZvXWfw;%r*KJ3kXzmES%KSzg7Qdjm)zuno}?pYQ-wP`<dX!X0;
zB@fKpjAvc{D1Vfn?T6K6?_JmbL@#_%wre%(<!7LI<x3wXa0nW25EA0B*Q@FRZtk{c
zE}hQ>4E(Nt7e72*yZn!`oZyu$fBkuxYo~py-}r!cO}Y4|cSa{Ce&1A=SNKo<kLkz#
zhyL;(`YwKtarUvvXCG+GbKS|D5Fg2YS-o{>rkJzIiQasTsmwEGx&1x#r^m3V?p4{f
z>p%b0&({JL47*-d7e+6?{V}tS>;2Qq#V^)N)VO_Ee)z|lB_HiOx772k{W8yY;nl4i
zYh91bztevvapz7Mo_X0bKg1u;=gq#)16=<N92&@~3U6WW%t_j_9k?LZV)8VHqk{L-
zXSh8)!Mx%@_$=U1@cN!Q{`(bq4}RqNck=Axxco<3JM!*rmz;IFz*)YHJ1ait>d$DL
zX@2r|>zUkRz)e=}Kg2&;AFPx6;PS!oQO8{y?(~PV|4JVUw-WZuF68z&2d<hE3N?a&
zJ5G-otuuRDw54+Gx?MFp|1(&-$4$TVH_a?NUjO*C{|u$~?>=9CWtyGc@ARX&(Fc=`
zXKc8b^R<3^<;~AFKi?&N`j-1R|K{f1h3EbxB&)Am&+#Mv!{hiLiI1hd+b+KUYHM`=
zU==&?ESCM-cg*l*J$hkw@%D*(Tqe7{d1P{r@kZ-DukTewU45PR>(Ac)u+e;X?OOT2
zs_`Fp1iFecgC<1==Y+_*X*<`Z|8tI$1fC_<3>?Q$JzBDbp7*qnWfzeRcVpm`>L;O(
z=6_ql|1)^%zFq*l(o1XE?8P@SHk{@IPP8P7Z1yZxDZ8kX_36aZClgwkb<P}}xoeWA
z%`D)pjU~X%hM*0HAS~MYSe~&W{gK+VUHM&mtbrXJzetwV>1!g6mU9$v@CdmoE33~b
zU(^--j&0rSWw$bZ)o$%=U1Mfub`f;Ym78*(N?>bf07C!+1LJ|!B7vd%)|dfP!~}d7
zA;<vp67p8Y&?Wby4<~i19`SfMU+Wy>?_*sWO!vP{UH<2<^M8hV_8%AjE(Nw#>;wKY
t+&}owB+!9<RBkj}(9(pA+h)|Ar@Nvw2LIh~=qm<+3I8wfMiTk|HvtpCChq_M

literal 0
HcmV?d00001

diff --git a/docs/_static/bonsai-workflow.jpg b/docs/_static/bonsai-workflow.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..0481c3dcfdc1105ea5676c4cc55f907580168188
GIT binary patch
literal 33481
zcmeFY2UwHMx;7f5cah#jK#(p~YNU$@2r9j(^e!b3YEYy%0RaI8=`{rDy%%ZHLocC9
zC)5BToP7H~d$09<XYaN4+1Eb%I@f>wZ!$?{-pR~7GxN?o&;8tUGkvoRxcyX3Lk)m~
zg989#Ux1rAfHDC0)~!F@*b5K)#wWtZ$HT)XAtWRqA|oLqBPAgvC8xM^o1B7*f|T?&
z-EAsr8d_RfGD>;|IvR#MG_*8-6oP|`{SO{KF+M&q4LK<}&HwA`rVT(vgky0l3>W7f
z;1(4QE)~vA7l0iAz`@71_Kybt-xtm;Y#Rv(iHJ!^u^Fmw18(8q;@-l;{iD^`+yU78
z06Z%EI~)%c38;175Z-g65q%$@MZ~G}qn%c7^pH#Jt@{UJ5;}SYMken2JiHJ1#3dx9
z9!bk6tEj4}Khb#l;^ixS14AQYODk&|TRVFP4^NPnH`vEF_+v<DSa?KaLgMG7<diR|
zY1uisdHDr}Ma31BRn;}Mb@dG$on75My+8Z<$Hpfn;ZxHyvxt?|we^k7-&@<rqvMm)
zvvbtN<sWk40C4}8SpO#3U*w{~%5@754;PQ{54mt|d1C`E6&^mvLxMYsx`c1ssPBos
zC!$e`&-&3$%qgaKNc+}(l!T5;9KntJL$rU9>^~>ihyRjf|0dYq<eCML;o@Ki50?r6
z1YC||=v!KE01CdY*O!SmfT*0m#9<_;)D@^ly8=@){Nx+64qUM6%IH9tvuPr?@?CsB
zTD^u?IRO{MA5EbR{*i;qKAY-%;m9T&rG7Zs^!llF*em<xC*chC+{p||EMG10JU74U
zE}SusE!x)H06Na6ZUCFZS~q}JJqXeh^5-~cjkAY|6zq+;g9F{_Rh-#?I0TddEB1a#
z`P9!vBfrAU^N@EsXb-C{O_OIALYoT%BKZeo^T3~ds5LkEc{zIby_Tt_U<?l$U4(%<
z%JZ!VznvSv@n-%FV5ta*Y5@Kvey9`1PAASDKC7M1U6WIw3N@Wk_fd$VJ@GJ|<L7;3
z&EQ1dnUGTmErGCr-H-JqJW+Crc|(x=BELDq15117mmvF>lFupiZxx&GR8n$IL;h0!
zU+X(c2j1Q?PU;ddcJT6kG2~PmuTJ)RkBjVGMIqs5o*WAK?d!Yp?0*@K|Htq=$`1NJ
znf|A{Kq2GMAY4dJ6~V8ix5cTkeuvIbk@TpMw?{)BaA2l>=G1z;06dtv0j!jKL>b%w
zxaPEP0A8-wOs89(IGNY!!sv6It1}3y7+7h21IXCA0R#@;0BFs<R}SLn7D*=+IE^f}
zR5TFsB;<i)X1p?=^&%#971=p`YWCP=>#gOHubOy}`#QA3osZrgqM~$>ubVL=Raan-
z+<<rn)3uM+X9ICU6cK~(>#lN*_pV(r+Sd)i<LcLTnQN1cC-PEPk`@>HndiGV0OU)E
z&A<&nejIb1iQ%ghheWX5K@vtt04ePQc}GFQx9z_!ZRz`d6kRa(h+FI`HIlyT-v%l(
zA75Px1r4+PmLYgwY_GsEe#*?MbQEBzHF+<CU$UAmX1!&oQy?YsLO=xYh~xdODOZcB
z6UK!bz~5CCw>tybY=A6P-T=A{-h+?Nh7U3Ckpu&s=8O;uR5jC*W@CL_Sl18r2eFq(
z!Idj<-Cp60S7tU&6*$a2in85)FZ`bzCZK+TRhx9PVUcJU@QP^r_I2Yx&mKl?@5=3$
zTeR+(PlF@Mg(n|l5vdhO&IFmMxB(1%V%Hnn8$e{v-*v+){obBMwuxV90;hYTyW9G(
z0LwLVwHts#tybek6!~#a<G4UQuJ7?rC^`SDiYd<lEp@G?+6nQix@14UMk9U0{Fu;J
zqZ6O{74dGrQ_RPIz-f2?VI`rgvK1MKXK6MeujK|1H+};!ow}%2eR3<3_xpcmpbS&E
z|NA_H1Le8`71jycqd%X0HCf$c^5GzU#roC2Z2}FJ1bj=>L1)9a#YBW`i7amMK-N-c
zfWqerwHWsjviLovwfiUbbgJ|1?!ocxu$MQ0kILq(o?7ZR035Cx00|I5H3j)iT6hCk
z=E_8<Erl&h71zR6t!@CEn`ehQ;bvAy`ZH<*1zs}<hALQ&%|6_i)Sqr?-DmCw0ErW`
zd3;ehq-nWq7Pw<r4!6*7mXw9jeRI%~*bHy}%sw;hV$}abZ|^=s_)OS=Oy-xVX6)X;
zS#|>$q1U|uv}pi4;!rA}rOR8}Kol_#O0OpQs1}3WtbS(6yF!3HHvn{^f*Z<pIg?5~
z8~tds49$h1b0w{>YntGytM>dB)AU1Bi6eeG#a{X0yWn*UXJj=3LIh6Ksy)<BmkPD2
zbtO#QNPii-F<(}!)%3$T$S1u$b<F@GZ3&1i2)0my{DZc<NOW($ya8BDDefO!LyRvk
zGAIv)5sR?HP2I_q=VOA1D8oou?DjhQe4wCZ;|37%4!`eVchhHHN%YveCfjL83k(&j
zHzued4s%NbL-Uua(!DXQu$=%^_zhs{^xDZv*Z2l-7{1ebf<FN{yrXlN^QS)t{mibu
z0sLvC)Y1234z6)P>}b41G%xPGNbH9DQQhnY@Lee91`w1031R>H^1hQr^8dpI1&Q)P
zA%8y>*zqBXZM*?IHv8A90{Gvh6_lQTS-L@+KyoP050u{<{6SHZ!4k>e>q@0!objIm
z!eMIPDTz@#hhgT(;cnruhWt2}gx^ClGpH|BKhvEMb)lC8v-7P7P=9F2FCk_*tB<HR
zhjCpB%qZmQkgeVa27f^&O>9qgce$7JrC;dp==Qrmj|v}WY`c3XGu1fy^nt1o#P0@h
zf9dl0jj$g?xJ4Fdo)IAwHV7dPq(^mhw@{;8u+NNERzc@X8Fv#>v9SgW3W6ujWmfV%
z8MpCHwMiU~m7{qi@Xj%=^*YgCcU3qHNZ>E#&5Qyy$s?Grv@|T!hy`2bhu$Jh%+zwP
zvlrDvmG9*%<8!2w56^sl(b$-D3i+1SiqX38-xhVl&P!;3PO_ZE;nE6am!Q&)Y5k%1
z8Ta~GBI1e7c*^sd_}3W&rU@H+RPbk_TwF4>%t=AGfzM465#G^EnsEeiB=s8p;1(e^
znJY8cOn2P`-7&u0a8*FAib6kO()u3m=3IAxw5ncGnqyAuDr2CV%D|#2Y~V1CXU+e`
zit9|{%4Y-2-Nuo7AEZ6gXR0hnZvc5JsiUU^4URM6zo`4A%2y;AU3Y&TQXn!b#O2xM
zIva#C^NiH)RsN`rg%eEx=)|8F1K5H{npK|#Odr8^E}dJA6cEC`rU#d;XuOnc&8*ab
z8vuI&rXF_e5c#Lo`!PO^?tg2zvz@^grWmJLsma{@LboX<jg+5TYqyZ5kuE--hJ4`s
zi&Z(Ugm-5KXAM#!cU+O?Y;0~2eoj}bQDMmaoLC<aKk|Wy>$W0Z<VKUVrLE6(B#2yC
zFJXowAgv{it7H3@5<LkW)dsEG*Z8D<K<6jN&u{_%7k&r{{p20pG3aR?$k{;KFDi8x
z&+z)R!pHS4HOpSx+hSrumb}ethm_Ov#kLy&tiE+;|Hp2XL1bBqH~R-0&6tO$@-nN2
zi^)g7y+!6`yVotj<8b|C!=`o&v<Cgl#16oegE$~Ib$D(xz4&`bD7J;~IAQAQzKN`^
zlN0lXi9IR9is&689_EOjjNN($UMFa03?xV%wu7=jx3ae`GLsN@+k#!qYLRNeZ%T}F
z<LAsfDtjscEc|3akGX7-F!{rpcx(HWxta-o*&>Y<kD`Sq5}vhg>@(@!N$Jrcb|0;&
z`{uhu*m)RfFbpVpv;=550x@>KJS~6%e5&z`taL0GC&q>=0C5~b{uA)Nt($kJHDzsS
z>-B!@@l)jDw-WGx@!_EEyn5jlcK56*#B(nxbY1wQA?q#VP9Q~5)k2Ace_IWM6MSRs
zC`<J{6P5c(YXt)AKch~PUj7=uKP;!2yEAJM>94%P%B)qNaSfua2^Y}1nzpG36<uCU
z(nyee`{<VBNnzfh|E$zm%-1yQg@Z7`(<or|T(8R!bK}eMSh<gY)Xz_ER}?z!$+1wW
zfSz8bEbd+KTpPL{ouVZ#DrZZio=rhd%{28ZLb!}P`x1y#+eijSxq`|lIVogTzfPsF
zQn>C)@#id35<pyuZva13b0Wy&KGDQgT6(9P#bx%*eYIWRq*d1$r`Me1o)kZz*p>+l
zaa}QV8Cw(-ed$Jc<`e$&6aG@|;G2#zHSNdk?X2EMXdlSXb;rr^HDmxOAGl*)TZJUZ
z%$0fK;`$~HJo9*(j+5cFwnyfo(4$>n{UmTmM$`&(#kl<$eFNxZ6dXR+=SHgbHo-A`
z_Ev9p@-~H^m1hB1Q@J_uN<}~_S$|VCd3Jm0>ETI(NfWddPj0+C4b-^@!Vk1eUb8as
zhh5wNQdZ=Jk84*_(A0B|UniPC1zRP)&!nWtT|JYZBqPHdqe7oAzO<D4MAY{P_iN_o
z)z2uYe91%~XeUJyEi`mm(!4u&I5#Q`Cm7e6qi0!iLzRU29n0QIX!;v(ip0Eadq+1P
zk8E%<#_!4mv>+bEInVJL7#)(zj+>B!`<29*YcBw}4BN)Uz42^4t%OWne1pQe1ZA%U
zCHVcI6#J9QM$W?VOpose#EU!>E8zzYeCowxS*flL@Lf`E82kQ>%XY%0rH?B57A?9q
zQ7}Q|A>&rYlQP^D{2Amyi10Xus-zJrG`8Jiw|LZ{E52s>Al<sMhpVGE{;U7o7l69y
z5PSQP(yZBJO_G}>$Ze&uZ==CPsiKtGrAvD*<s<Wy3!+wdd}CeE`v7CHfmyy1yu@f>
z!b%)xyVzJqPRaGxx%~W?eT1wGgM?nu{2(F5eZ=~+G)B0)AvBLmv55m})P9eio_fIQ
zxqt5ak99VE%M>#4U&+$eB;vv@;VFnE_QA`uJSgi^IO2Qd5p+OXD_f0}Xjy-$o(P8s
z4}Uqq;XK%+^Xz=qcTPAO|E%pjhzqV&C;QHuxc4)msmrok?1W(8I%=LR!v6-K+&Cto
z8Q}4(rnUO*i1z!bcb$~a)!$MmvW82z)E$I4tahFsd=Si=JXot?ZOZYuPqh4+Izc|h
z>!nYaCpS&i^IN)_s(?TJAmo<1Ophzg9Go&(FE}?iD${d4m0~vnrk2$NG5<?S&VbH7
zY+XD1Pf_KDp1jl#g=>{8IL|^in4>TBavObjU%OpA=fD_4KoCB;D$F+K<u)(Fk|405
zHHqzC5<xULWXWE~l3RCUhM!g7KEn)E-2gHTAqUe~rMXfSp#P@?V5Ku~6V85h2J9#u
zeN5Z>w`kn%oX*8*80I?eutfYhzU$wjSR-W|6YvIWvRI<|h7=`En;UnN!8w7AW^LlT
ziyh6i%-8AmSOv2ji)#G9Dv&m_kl|9``7R{HDEm|Y(o9X}@ebwn6+1%x%((!4*@-nZ
zdr_~_cyHDIU0nC?%b>N11VH~j;m?Hz@&8>N?0+j?-|jofe^{O4&mTp$uZwfIzbumg
zg8xn=-amCv0&}5D<=PO-A3Z;fE;jaaVcz`*UHIfbuJC_YzW>jwQ#HyCdej4NGI1Cp
zV`7VP!?P-N?<*Oi7R5bER&;1%vAp~6atzX;x+;80?<tIC!(wVbLD&gs`Ij^-$o7{!
zrbAO*2ZV;G7IqR-RHbq)@>!d>F-yZ|e&6Nj59M@=?$XP;@b<CwCDfB9ELfqmA94c-
z?rXiEi2G|ojoo)L*S{wI&*v(hvq0tlI@;f6C;M$LF~>is*6bE${rvmt?iWG;+52Av
zz9heY|Cd{uF4Vlrm<^umoV}}$uf5<4lDnj`k3XfYRb~F&*7jr4-ynac-QF4TjeyyX
z$}(1B>V8Kd0|1hB`Euzn9z1{vS-|S5go%(NR;SU&41Y~y|Nnm;MiW~T>a)BUA!E_8
zQUN1essx%(#DEN9+~54=WyV+gLzpehtBDaDb46tBYLx{A76f~dTk7q55>X#@%vzig
z66PhJf|P%MBIw->pOvdH?#74b6bhKT&In}L8&Y`f-+iOV89mkQYY_%=4bP;Zem1e@
zsP#gx(W^e;rF2{ga9|gw?d(?lj<&9bdJ)2Ax_IJ2&o}%7t=a3=<x2XljE8WvndcN!
zPk~3FwCsoe&9^szhE|Yz`K?N)10nINj~Q?u!2;!^1bW`i->Ly-?7Q)ZyFm&CP%+Dd
zN+rVebpukQ$(eKQAv84812H$3UR&Q3Ij}o5m@3FI`%#31aOnNdgQ>`sB&MiE+dln%
zx2lyLE%QZh%fLIBv9#DkKZ4Yi#?T4z86wrz#MJNS9%#p%%rAZ?u)OLx8hc;zJm}=<
zin1%{o+?>WMWvLsIzzAY)_V&Uv@S|os$+pvHRJ5UrJ8i}lO_x2xh_7*W4xo@jh)`H
zW)lS4_drg6^2rna)+Mtojsya^Z&}{m!#~Jc70twLxrJ@T7W;p^+Jb`x^v@bvuJ_K_
zl}1flJ6Y|~ip%nBNBT<-!uq}Tjfj<Wrd%LQ+x+F1ZP}5$fImX%4dAvTT!DVH?9@yf
z;q2u6yiXuz+LuyK2jI;`G6GA1IK?qNt+d0?z3=aOv3GtKLe=-A7ZLX8oGod~dRi*B
zt0%PY;I~hG?8DdZ#$b-swIz>$5jTLUf(SWRZ;6&9H_71aO=9xKy)?xg&R8GJ??;;~
zr!bBL^)xka8*j^G8)>e~w3FeP%{CdSE**AOSyL~`c;n`dxzZ5wP6FLZ&aZEx-gz8d
zJDl%vA8sn5JXQvn1MsRu7EFA`jH>dLiOC+{{lP=X+V|V>)KuP-jq-NIrAQStpz5M|
zKk!v}UUbi!#W%*@F?%ho+|{wbb_h+A?lSY2wb4_1M2L=L!b32NFT*f<c>9UO8Kk{6
z+8ELuL%XJ8B5yu{<89seXjon6CGd+JTNj$x4Awr(IIwIH-k^{htG|$M;+MBOQctpE
z`mS?9)nnsAOTd{?o2pv1H~^hG`q^gk-K!kp{*%QRuVeM!S~R*ApV(m`|7JnzdxuvS
zHc5YXFfLscnV!>dUj}i|)Bv9rcJ~!RXG9Ys-TgTG6=X3Z&(RWC^hhB#ei0n1GT(g>
zJF279*+})%LR;r`pbAndrh19dth%SRc)>xGx=vt>H3G98Fv-?Nl2_(v5O@zR4AR_^
zn;Ym$vp2cUmwt4@7ChGSbrfRTio1xW4Kqqpmx*q$olAdprHmheZtnK7b(sV{z_Shh
zDRBwqGn5NWlG2m<`mHPpi{YyN4t%9wCZ<sGVy1s2vN^q?cKcc<Mc|;!bLV4SRb4XN
zQ`+c_dvvRkH0?nvA;dvYN4=$XtjXftY-|b#Raq8~n8{{i=BkA$x^v<Lmhl_rP5&V)
zCOE`yX}I{rHsoW^RJ)dTG<1FRGUsHUF=2%(na79|L{=8Y@qX>LUt|EdU^iz~+F#>v
z>trrP)YR@c;G5y8NcjT$`myk+s-Ib)U{vP-vAHLAlhKtnH5H2)kwNU3qAHC62b@M%
z74VP3yUg3Uc1^e|25P)J`KC$MkM4%2$D3P}In2t3bvyHD)S}+jGMU(?eU5XbjZI4L
zjrHS>T#>YY|MFFWVs+%h<CHa7?P2T5s-GYz-?>ixyz@T7&&TU4YgwK)rDd==#0Nne
z*K4dA>}Z^ul4xIB*&8@I<MSz7(*bt-2B7wf<!AJ}$Ag-mz&%xK*$2GysrS#kq~r*_
zoKMDWo%08s40M}g`NN-S;^{^;elMm6zot=(3TizXefWu`;Fp5Pcq@V#<%(oB9nI{7
z+)<#fsV>}3+XyLAbSv1oZO+~2_yy6Re5}b*6#gnZrf0e>h1s|TIwEe^B%jx7NmlWz
z_~^;lprW_8m$zjV&cH)C0Kd=`BJ-n!&K>m?=3x@M6koJcY+C81r?6+~muZze?^|^Y
z(yO$QV-6PZEllR&Q{SFpEOblIYJnmQK|e)huFTmh2!O-1kyaTjVTcpSBO|bbrmT0I
z{(?Pgo0Qs<S*NE*;zS(apLV~zS>8)Uw@84a)GZ=7B);tOjLTUfhB&o97=B+Oqj$`&
zIt(({6yT{_TMyM9lZ^I@tv=l8$lyZKrD82HhS6#*Z!CZoN3M6S6X{r@?QvCPq{$<P
z&VA`y+X*DxS-flJ_u?==tZ*la#Z9QdyFgj+r^=@%k=)jTAs_=>IT1W}zMTC&shZ4S
zt?RQ|>j3f5D4m`0qp7K)<6X$@CXmk?qLkhg8CNt{CY39ipZs82pyv0P*YT3iTfoSt
z@qsgN)`CcUpmm2!pEHoj{}~bz24cqnt3R2N&|%jo>MIe1HzOWZfnn=@>Z&F-^zP8Y
zB$c?0-RE6xbC!yA!j0Q(w+?^75LeeESDk5E`(Y6NVefoX!*1^Ct~j1A2VAONjbh){
zK54k<NW9QPnJhOuc1SMLOvrJ3Ujmy?+znQ#ALmHu7tHi<BFm#^Eq_#G%<mhh5$J&N
zo`P;`N!JTjqvtGwHTxG_^MP$~bdDL4DBFVVK+X@3BwQ7?7toFAMuQQwUGY8p@QOa7
zPuUMZ>Kqon(U!O)d}S$lMs~i=HJSM{b6b}IrH*8DH9rJ&RU$=-!@uJLJOOW^ulg1i
z7<*N)kCq1ZL3@t#PXa#kL8H`F!G*yK)9Uku^y_hSNzr!-k?m11kKxq)$-8xQ-id+}
zz8@tIM3>dVEdzp*_@1wFGRS+R9e$(lF?Njx+7Ai8EWafy&1E+_{kr~$Q4y_k3@ciT
zv5twKkP+$RC;LF@fO1^v?VMW*Srkp+{rF>AgCkSBL0@{7;_<@u`Xj<t<#_`MpGUdy
z-eFtX;*k*P9;wI#*R}nY<9*-dz=swlTeIq6n++MVV9u#s*0rP$#l?e+5j^F9%<r&x
zEX?kV8pnjjyto0}N=G~Ln<GQ(S3)8{0mjFAh8ho{gM$c-?Z@|jBy&qx#Ywh6wDY8u
zZhPB~uvf43x#gntpM0~5S^JhWu_nULMkL>DaQ*q^G}ybxH+qIHNJZ4wKZ0NC_kcOR
zy(APS7#iksPG7PS?`XwPPJtq{+xtp(&A<H{!@iwCgpzkg#N3-KtCm+&)!bU;baK8g
z*CT%xbjy~NK#S%;4<;?`pBFdC_E4=)xCj0Nvb+=~bRPvN7`lcQ8kIshOGU)rJYMVs
zic<d)n`-u2I*gu&jI@_Pmfm4~gAL8!x2Wux1Y^GuJ~Zz5fxGkm1B7~b)nKRT>IEr=
zDn0I!LBjQFpB*Emj^>nNKM`?QI-dukU23t0fG|qU1^D~*Wa~j41hv5a=W-XagG5Wk
zHA-Fd^(RkBs{z5ZoodgKOGK1A7Mz$VxO+QNY1W&3U8g`rxAEOR5&LFyK)Si~)iZy=
z%oW0qS*_!5#2lG4?C)@2jjKFf4DxotVOJu&-;9o<B@>bnE!@q>grKS0hD9p7Ut2Rd
z>^Yu8+7w=(ES3d5RtO`VcdH~5W}k#7n4EI!PkNjWP+)@Gbyn2N?JO#O)ff6dt+${y
z*YL|SNlLB8<q)-aa8=aDt-~(q>-9?!Mls?KS`j9m8*t9qk!mp2Q@G%)`P{?!Bl;^y
zqUZ)dn%E0Gk#tx>(cA!5Epp%f-k`PRQF4nmuK9(j1xj`34{RwZEZ+H?*s{HtjA>c@
zEj(+o3~D|L&EPsbO;7~ocl?;2HOU|4ZXhFgF8iJR{O1k8%Ah~ZIkzD+%h3^V-)G%0
zhjgOyHUW^@Oas-snvBOzyWItu8i&OA((kq&)@*j03&wSwk5p{QSY=pgciwV*-~WD_
z{(${mc736rQx3i`bVx(p1(SBM>p68M-6F~UGB_&vy4mw+{_;w1eCc_FuCR;K(;Ce=
zTD&XPesJl*p9a!p)Z$YnjKSj1;|WAC*4+TIg7t3zEeNdSphi`|8%T$k8y4UpfZhNC
zu&_hVZ>;OdW)%}kZi2ZAy8*<&_h@jiwj|L|?==}y?e#h34S=~65@~FD<j;+i`0AnM
z{b=0wo`#5({{tug?-aqT{MGq=Ohe3^g`RJ!lR^~<s~w*fR1`6NOM|#283)3@7M*Ja
zY;Lgn*`!>_gSWAdXz(_b>Ad<zT^f*5JUZX-@_<i!<eErNEcT+U%w*-fGacKaz(<}R
zHP|w_@>jJQSkT|s$Zh~W6gX%~H>0J)dN~iQ05(_w-s@l_uH=oTdh`km@Ofu;%36hw
zw*wzOr4a!+==Jm%7~2;N4-aIW0y~yw#?=p7KO6{v=D^kDFZ6>4niC5nw25NH=J0Uq
zK(8s^SUf=Mpu7VP2f9;a<TtFs1CM4m8K}E-`j^bRyzWl4{-s)_t1bl|R8+v5H$DLY
zWecftjfOTJ{4P#pLcX@RM`Z#Ol&m2znd!)4-Q=sSjYh8%S-0sy3cioLh^<Hcvd%CV
z<X9o(I&zO&3)r2a!hTP~Is3(&djx4<Z1qXCn*||;ou^ahmf;p&iMgb06G)m|45yKB
z1VH1*TLy2H7<x$6z3QMJI<$49e@!`7(Ii*?h0oKr!5EEI1fOR|^7qSQKmV5z@gIES
z#vnuU6k-spuFq>YT$}}LGwyZX73~=ml|91uX{lU=k#Dh6m5Q5WM(U}k#$41lB)%HY
zi5;`cuU2#wFjrf8cwUcR-hcNKPN=`HZKQb0RFhbVeSAYV_g*f%e{WPX>$y&?#AeE@
zWIa#ZiDUiiaRcM1XDyMl^A43ht}aKjj9)S<PW3h4DZDwK3!Vh!FK4!tJ*#i&n?@OQ
zeHwB-W8TwetX65NuYQ@ds<Hku1Hs8n&r)IX6a1Vqd5!ErZwz<`;q?u4r3zNomN}1c
zKH2kl2Crt`k=0>#lss%2)wEv!{A7Rf?W^N;98QutC(HM7?+ydwJ+gk=JA+5l^D|af
zbhxALiVhy_kv-;NMt&t*{a$>T@_tuBht<wpAlW2j!qZQ7fEgVBS*g!`L_DP3G?`%G
z;RixNtRAR4=Ae^z1K_!w9+ek(=izTX^#)1UB`>RU7*f5FvOM)+WXW`=&*fZ9!`1nb
z0C%*2+db_b4JM^FHObuQIfJmV8-P<`H7j9v#q1UvJ5mUSxHth122vtl`sQc4OOOs%
z^ptqLiOwx-Pt5_h^$e$z$8WRULs5SRrw#PfoJ$#hwCy$y$uIH{&|%v#Nw?dB6Mhu8
z4t(KAz4u8bnecAWoMAj>NXB~NauQy<azG7E)SHY=Jd4^>+#T)MnsNrY6n-*MXZo<O
z{6^YBZGfJ9N)}^+VOfG9oE4Ae7BE;A*oW6UOI7<t2q4)`qGKT4*h#aW+SJMj41Mjc
zv+JlRSZt;al0a20&~UR-dnk5A7lUO>B9XN{Ly#9pgN_>j^)?-R*8BbPh(so}!>8YY
z!U-y0XPR(*<x5S3?-8}<#4PArNPmoAj3zi2)LfpYc2KxCags@c3hOkksV!12*7D06
zB6yo1`(#f^z(bi!hcQ;<X@6X(Z0Hz0Ba`s%!16VZ&e(3g<F(%6Sf~E@KHb?9WT8+V
zm;0i|BNg4|GzGcl$gD5;JtGtmrxuw!!+VZ_X-4#55B|NW%ac)IlNMA%!5$S1C8Yr`
zK!$w~JPA-@C$m5mXjWCgW5yOb?ZUDCyp2WRJ-3bSPuVX!@37V^Z=f>IJ|p8o4GfMI
zSnbDaA`szdlVWEi2Ptj3hNoQN@QfH2#;n{qm(Rt<LIhtwFatOP(`Xn2HL|*PvvDKE
zwwY*$Lzf#Td1}P$;nan9x^ZIh^m?7F`b6UUV%)o|SX*zSDMp&GIx{hFw(OJvyIOJ_
z4hGUR99cy>RqIr}S?y=h#3b_Y@OPk_SYqMhSbmaAEoKg)zP<7dN-KYPec0S$dUoN<
z8_44Sa?;;umE5v3?I7&#;q__RH`~(LfOI{nY)9FYZ`6a=6K|!HpmXjI;?>k<nA1+u
z)=%QXd|jeeH8ZK(<;R!i?dj6^)j5m8m0h{UH~}=QYrig}du}aLcuc%VY8gH)%M6L0
zyu1i}irq9YSoQMd=YM3v{c%sX<*E|W07(K#3|Iurv@QVq8`1#$JY9>lV`%H#f&A?=
zdmnoqS(6XYna{B=uOi_v(dz3QIp#vgDJHGY<_{)>1W$>@>s<u=nM+NLpk3}?axJR4
zi^52@`z-=TgA?DUS<p8;P<=|7Y8FVC4o*K)N#CR72|KMD6bLkwpt%b}lj96uAMx{R
zUB}C*a<|c6!c(0IeEeKRZLmA!(}H7I_G{!>gjBebC#QTDQ}xlJhO?uGwVWWIaUY9E
zoluZ7xqzR^qj&9}DvpiQHtGw&w(8B3)qXnxVwx2pWkRmP+=f}95{$OqZ*LXia;3G8
z6d)zCRx~-=`@Po!Q*1d7mzI0OFify-w#?=ZPL1Z6O?>1$@9&(g;UbB}DdOeKm<(zp
zrhHaA)vajOAw#hta=>&{xGJ^m4IV)<USlYy=mTrA(S6cpDXM0Vg(;=Dd)1}B6Gbm_
z*ljh8?q(M`Fu}j&YFW}=Uqe6Fjz^iO)TLTFMnqPIui~K(s#$BIzw)#7l#Mw3ZJ-QW
zM3Bi*%j}f%JArp6?v0C2OO|Nf*YJKz%c=6*Sc5hERSb@;4axTeeEvDNgCCm^>Q<Xp
z3V)@B6y0jCk!l}`Eg``XaraDvI8L7V+W{Wt(x59QRJF|;*HXKR&Y8RfD@zp7@G&e0
zO^)>F1X2mZ1UyquNdvUUB+{IvjK_q<OWLXL(CNB*)98t-2J?P;px56ijYrYnj8^cD
z?PymLYuEyLv~{h0VR<>fNvcpYa?#ih`7xCAZP1&IXY$wgVt1#~mpw6<lduiWX+Xfv
z?*J*Ss?pNBKd_8~v0qq1wMgMBRn{O{he$;CD)Q0BH=o4?nSfp|YMjuvlIJ0?DcB;7
zt`QqA(mZLonaja<Xli;lW~ljl%7FLv?lIRxAsXsqjeA0F@4tQpX!%?_`4Tk$8%m`6
zm;u9g7@N52%w{HRToHTrOWUgz-=vbC0y}vxhPy3(e7ynK`N&DbRNh!7)lQ6j$XI-f
zn|ZZn;0C&Su4wC6wmp1!&_mcE!<Oa!rVyNE@=Wp_snINR1R7Rzo!b>1+v~&~S;V39
z^z2KXMveyOYh)c+X~Es#k;qdKpJe`9FeNXY0-gF@pD**3q|r$~Zl5Qo%xpC@?E1&$
z2lfPVCohvLn~p(W4+}eP3@)7y@KM5>R}l~_ae%^K&_bbKvU$?Tt0fEc0kwbIQ}JZ&
z{C+b<-4ZTZs|xj=d(PfhJ6B*&LQwr_T~podQa2r2x3x!ipPEK6NGlbCl^fsiQB0Sv
zGRK(r=bkJMXvtwtEJOOr!BTtnon~dHQwEn}71=-h%~o02+kBZFF87wEJ=Zf9r-#<%
zGMmz@Lb*F`1=@3|{t^*O3?haWBXz8Jvm*ijEUbgBjXYUN^(Gd@_>4?6+((PrUq_k%
zu2$-0cVa++ckO;?=G&2$)TH>s55H;uZhXvLMJ?jyTn-ol@U{Jz&wE0$e8MJOnf@G0
zj}vsxu`lBp-@g|r!6-CD4Ir$)-xO?gh?c4C(j*Uvi_)7EfMwSiuEAG7X^F$jGW6$t
z+*TO9@6noGKpxCCg7UvFwe0?(v+S|4P3FR@mZ~&w%{@yTcIMLHh*$3B@0v4w|LN%+
zV%!gpXqQLOKFHAKRrWiEb}1PYOc_x{!&aFTr*wV+-p^|zwKDBB7CoPI6_dpWyZlsq
zY|Y3W;MX>*n5d<tv?c7kpWEC}XK0mdXW(dRT0!^mleAGEM+(mx&u6WcH%Qs#spvpn
zu&v=?a;q!C;<nL(=`+s*y2a!$y|K3mN~2Z!11>*pgKihy8)93D?6v>pX2G;~1Hgy0
zFX^biUFvk*p2hO{<}8wa4H%;~5fB20v_Y90K=|zorY0FVI(=i+`4p-J4XavBC;uND
zb@AgKMKO;rlI7^xB?UM4ShFA<>R9%ovt9=~u7j^aWs!wLTC}@yp9B1bao4Mljby6E
z_OHfN<{2L^f1vPRx>db_%J5eaF17lBRrLF(x*jDM9M3Kal5Oqg5}jRyzG!5etNj{$
zG{pf8>`hS+pY()J9qhfWE}xLoDw>_2Hhn52DCAmAR90%M%1b-TE>Ow8t{l?l_A}3!
z87ox{?7F_up=PSyE3WbA4`P(DJR6n|xn)LaRD-08vo1D@3aM%eu?;sf(R3DqGcso<
zy&74*hCf$#Gbo{M=j_r%$P4-#U4Qj|QqH&3lJLvf>hn8u`E3S*p1oYw`W_SJ*ac}L
zjl3VxE@O4Ns<hnyS>6$iYs>D^e&)KFmiyF_m+QICQCWTv$;I56&!IxK*Wx|+3Mynp
zQ}CgjGwjFVJl*A+@^=k+ZG2|Cv#Vrk+V=^$mfyd(P#M)0PFlKL0FF2>vSip__rWGm
z-d!1d$P#lihYZ$+rimsqYeyuVyx^Lkhd`iyi9<Y>LwnzJ<gL5<SH)Q-A<}6I^~I=8
zm5H2@Q$gXL461SN(bHxf&BODUY8~czy4s1o{G_Ix^p_8T@hmZCs_y%5neX*EUGJzq
z>g#U50SLC7%uO{k&h=l-PU@MDT?@@!NcjmZz-|C|7jr9Nd$h~Kv|x3!l(vS?fpSQK
zq-=}rhYsfC$%|)Z)VD=1VjgLzR>dV0x0w8#_E&3XXP2JCTJ|HJt>F!lsmLt<SBNV=
zSJEY5IM$ljrq0h8-%(H<%Ntra-A#|?P{vDo!NTxfu<=B{eu)jt*1N(J))QAa;D%q`
zT-mbZ&=1=%G3|C^7<j;}+3_9umHGoEhg!c|ik9$Z36@s5x89N+_4{h+2+P=OW<-WX
z2+2tlgO}Q;(kx;?W_}2pL?rGr8Ay1UoRJU6nYOC|G+h@WJnL=~{j?^CF21ABP2h3R
z5UZCi+B^eD2bS%iTqO-fnnGwnPe803AVSB&6S&z^6T{8NGlopS$|?tx9<5|tN*wHW
zVP!Q8OBTJvk})|@7Y_5yi!U2YZU9&{Zw+&L9terM?btxs*%~}`xfxl2eJEVJ0$H3j
zvHF`ClK<u58ufOgV3+-M7Rl4mc&g9Px`$k0e({`HoV)Bn4K=l8{+x#k=i#o-r-J%+
zj2TB$za9>=f9c>M7Z+*V5DD{%z`b5)He@>Z4K;-YNI^=xj;nW@5>#{+Gl8eqpny3i
z?2;Qqa%2&gNKSEu*o=fF>mN>qfzF!YT(MH0OqAdTBHRPTAN=dqk|k|8iR$CSiJqJm
z^G>OSjTTHwoR>f>B9O2i%D8fHls?iACDjv1fpq>DNQA`i4y{J`5wkxZS&M~BQDIrz
zy&x`=g`Q)kTo*jIAA~q@^j8{=iT-kNMv-$XEqTjxLzEpFQ2Xbo#4eey$By0=0om$T
zG;Zc|7iBOe0;85mCW09KMc#AbKumW25Wniq=-JY#_l1LT+SvA)>t>(@MirL1bo9}8
zDN49S#J-h6yBZo0*G=zbF0G~Ll27iK-%WB|(Z@}mr0h*wH%$_gGWgwDk3aK}FxGVs
z^Ju|v8r6(;g_+Lp;P#KBO~6x<+eHHL>P29$4Ip37qtrXJ8FXsiN7@UOBCsM)SlHRD
zcU*NG{cw@Rwr|pPmHu;8@Y^HL9Z&9PeFUc-ZvEo!`b=cuLN%@WD+7K_#Ml;8#=Oty
zLY4pAI{?%pWi0RP`mNU4r)DXyk(~i$utFHN7f8K53y(vx_n32_ve-Wg9muJEpWP8N
z8?yK1RDPHSH{6WVOnabNtQM~t-$aym)Mf!3^*250Tcsp3^4A*2@YsnAEbnOud!}FQ
z>^S<ROpQrX*)+8@dpQ@eDLpL|0eV5}GDo)gR)vfs^QyGZwsKq+pPtobIX=m2O<w|`
zs?q`**{Q&UEA-JWEuzM=+iV=8&?bjP-ny!ZD#tkHZI8uo;-*h%3B!_ypa`AH@Kl$i
zpt}Mw5AP@()Y@fN#(vVp)L3}U`J6YppQfikP{2(Tcf}vzvj`T4z}<)oYXwN?Fn}RF
zbmsSwkZAEGaxBeqify%nA3vu0+d^(y^e)b)*#KC7D$!egw|dh(jwiHzXj7DQ@eX|}
zBj{SV(*b0QWJ|0#TUATA)NE!HI60_kCwq!)9u!`e8Nkzf@I`GDJ@iolYP;$h$$mQr
z-lV)7kMP5uoM;GH%(Ttf(KooWup{Uw*2y1GCi_uSBH_uJ%6-PeEGwt|qp70A)z%OJ
ziGcvhmucrZaH99ZmP`TN-TCQq#icdp1|vOZ<xjo8lN^2U?om_D->5O$Z4MkT_njSb
zDI2dIvX&Z_Rwy$#nzGQ8&onVJ?;U8M3^HDv9SiX2lwvw>|9x7w)^b=De+WNUAD?XL
zlPN9A))Y7wQ$GuRpLvcW*H#hY47<0E@<3!pr*tHCE)NqsVwmFh&|vWro0>$e<XPDm
zwb4CJgMe&;1a4P;ZnxUIF}ZQYi0fPo`vN%9jAH3!OWg!#Lu2_hJj#DWexdO5utR0I
zT+KHXQIQbCMqaYw_Rn8|p*MhP+Yo=}Wnmo0*^ruoYzK3Z>V%a6H8Z};ZHt8_`-k|>
z?GLlkM6LOXs#1PU+7JVMs^NnExlo9F%ihn(cjJ{8TSEn)#@4i!ut>BT%JFcu8$!^e
z>9`F1aM_b;i{BIC?Pm-8DCG2}{s=_0=*If-N6EY?ck8znyJgFJj}*DSOe+kHrxA?r
zX_cyxTHg7RZTF0u8ke*@=mjqO13>H-+I3OPxH-Y3c!0T<lv%?#SFlFoq%%m%qSH6h
za2Sv3z%QI}HNw^M`Kz<JpPE_yUvE#wAv4di#uV7LYz9=mZnxamQr4)%VtKm_{y@iu
z+L#2B&ldL?qs8bAb3X%z7-y9M6R?Wysl}74*5!SEI~W#|DLA?Iwezc{-Mpmijl&10
zckdveK)uEk-2@FpZJHL}K0kjw2bUIW^J<sirswR8VeVT=AUnV12%j&{yxHHGRsX&_
z)1!8(+`h(u-dmK&8M<n39pAuL*vIo5r%dSByfVF)NxN78ds0AXO1r6=%Z}bpOrWMz
z)XRnNoColLaHG$Z*BKVRbfn|2a|3wY2)m}4+|0OJkA(L2$W!Z9wTz9y1vy_uGDnDO
zti>*H)s<UE>Int;FsPA~^FE}><QVRQK6=n7_12n)Ek)i@b5gK6imS283v84c2%eQB
zD}9xe6vU_e0Pi-@-KkV?)LG=19MBuuFDJXgs@ds_x93`l%Kk+JtUNh?Ox9RCSoxH5
z_}7T)T_uJ)B;~9+@3l^cqEo0$O?uzXe!CJUcNuzc19;Z35Vi>Ddq{Qk70>P0;fp?h
z`W2|fG9(PGyga^qa0eXK6>Ktb;Q7M2;MW;bvDv}^h{|s*27#qnT)5csDaM5B<9kG+
zTA^N;((0WG>+AH3>;7SjW??~{tep7!F<de5C-VgQFFq;$Ry-*%z=HCOP0VgZsGJq)
zNjc^c$3WmjcI^!y1X4C_pJ=9@7w^ba-0fSa*OoV(rop5ZHK<Rx!OjaqGB2m2qD)a)
zum!N*&T}LO)=LPilrVJWriK>AX~rIDs#~Pq{$6LqCMsIk{);*?u9>TG4rw2-eR&DO
zYmmBzv~L|NWFReK1?1R^z;FAHm?Nx_1O?Yg4kLU?8iB*&SGfaZLenf(R%<YZKBn`j
zabqp3czdk($W1DV>y!8!=M+n+)cfap27L5Hj1@aat8VPf;A4bMCSt*6M?D3BDUj;a
z6po-vSydhJ<d1-cq=WrS$JAY2J~j~`!sB(HR*Dso83vDlE!r+aTB;Exag^W}dLvE5
zSl7pG2PVmcOQ+<_*FP{_84s)_^OTKM&efEtCZ#{ImNel?nbw{_d_yxY#7-+OUat0^
zVDvGLndp?f%hSuTN`I{tdD(Uh>!ie<`RZ<(eu|^u-a=0gM2KOev1SrKmR6$;@9R7M
z)f+%4kQhW67I+605Rd~V)zz{&NAy4mTAf`t?a!soBa>}rXE)Z~tfVNos|_lvN#V7b
zQV95qG$V14&9Pz2j$m=E7l@5oyoSJ}!vo#JU1|qKXxb<by-Ua0y3QBN^tWHb?j~qY
zN%YSfsPde91X89zB-Z_&n~pY)R<#&dY@&SUPDo}~WCy?nA+_Cnglg@KdZLsHlidCq
zE0nmiX7=XyBiX|mCKmac#|3gw&V?pmmkrucTBmH4819ESc~#z;;t9mhWRt=c*?G@h
zKQ7{!0aw+#Vbt5c_wr$&!kM{QM`~kQduVH=z{hODH<;1^Mlzb^s@E)^8AiK3s%i|-
z@Qj8XEMyq${=wC&b>Pzc8>KZCi5KnoC#T9tr2f?!sCWEV|LcG0AOjnx<8o%C1tkX@
zEx!ywsluvQpJEQ4S4^G`?1a0Vcr~PMBq<EAHs;OtawW*v*oo#@IyoORe?ao=ro)!N
zm0eFORaXS*yHeg+SU51HM1!E>-FG7y@{h$!=wjgq-ib}DhXL2Qw}WaYJzqqv+e6PT
z><gBw7fzjCm~Itd^;<uz^TAjsf{oJgkh4qj5T%QMmQnB&6;gyMTP_{*hkNrppJ76O
z?Z;p9`-}iXzAW2{At1ZQD-a5-6N;j$*>_b#6GDn|G&r0;+N#wXJSTn!|G+pQ$4iq`
z>;6+?_G!>BQT0*Jm+mFvXApFp@cKo%pOMEg))BPahjF(ZklAb#j%ay=g5;xsCOX7h
zH7%@%@T{DzhS<zvGp+kMZju>iB7Ndo4w*J_R?)j9T$^4aVzu`-qg!sFRhq^nhDm=r
zBO?5zpy1$VNSe#!tN;(%^-4vF^dXmW?nbWnEqzI!f7UHk6)+EvztnF!a`*6SMxSn=
z$9%o^Jn(55dPr62LF?HT3AIC+3Si3R%xHfxq`SN`=a|1(7U|sF021qkqjC^6-cfL`
zdp;}PcV;3b)F(DX6T6G7<$8*`q=@9Bi`U)DGBa)b8yD%{Kh?sZ*Ol0V{29gNLQ)~}
z^R@2>*~sMsInwXFt1<WCKuSB!{DA^<P4WBv)3W^)v*FVZ%+~OI6#0*01{xEpm-N|+
ze9u~V$M1+IATr?obKQJJoWE*b+=r(s5-}wB*WR@v&}-)gOji7;8F#j<sj7{H&p!MZ
zN<$PW4`3k~Na<x`3Dzv1SVIz|u%$!ZC3~gB_%VcTcn3t#xHeMkwiQ0t;Cf9q2?f5a
zZ;GdLW;<&)l1ixq6TPSrtJ;l^b>D5eB}RJuAq4Xf-bC%UDohP-M3ATa#85O$$r%(_
z>wfg6w@cXiNM`FW;O7@Bmi~^<C&4Qq{X)><*OcREQvd9DgGJ!bg?Zm)roA$BYf7dW
zo<x3K+K`v4JM8v0@j(W|ngFNwV{g76F>PG0cbz;Xyjq?ewsvdA=pc0BFyeN1+Ff}f
zt7D|J%O}d-VqJBS64i-=tDNqg;wcq~a$wWLv{7Vl7ds2;4m&eSf1u~wcwsSBD%5<c
ztF}f`ZGU>3DP?A}HtAmIBZlyC-xZN|o3+I15_P)VY<Stg#$G``*3s!z8s6e;a0+b&
zZqH_F&g{@Jg)E#xp?l3klhIM+^CJCaXZt_PO6~=Jej8hg)6Vh!x2hG^C_c<DS}AEI
zn*nF4H6gjb+tiKjof2gn-IiQy;;1cTPLXEsl2`R_&ILDivyoqC%e_L>OE`WL?wDHS
z9d`y%GMlM8a(6Fmp2^20+C0A-N&W7pver8OK&e}Z4{c)fp&1-@Skgt0;jEpgrli}J
z@wA|VLH9sJ&=mg>2fAL286rH9^RCS=AFjK)A;czXXX~!}kr0@Py=~>QnHSE5&^gTm
zy&C|(*F}C@vmGqCNpTex&PM)m)?Pt2v$xh@kt(yh?_qfbIj-@L!{W<(yzyGigYE-N
z7wK!v?iE&dc){3HHArAw%RQ6d4S@x<T1fLoiY!FQ7w8*TRJ=)bhxzNIi=^3~QeWKL
zK@2tVAsmIZ+*kxBw;rMaeA0I!qNmL^UVz5C0laKB-Vd<<&p00bL+8UkLmU4)?st_&
z71f|BwW5Z4c9`>zng(6El2>Kr+P}K&tk+0c`uo-N!3=%#^9V9mVnj?AdYxI4D+`Tm
ziCAYHtrV;JrSAMH271lc;f#CD0HpXn%iRCwT(aO$K5g7xY}U;}Pv1upu@NiUBkqms
z!I#Psq)l+v`$BgM!!-u7zaz`4*MFu#J)}}wr$Ksudg&eDNOFkOcI+jvOYCSf)}=`i
z%UJ5dD#xenY$%64kVcs}jK%I;)B@$+(>X-2DEm$3fl`QJn^A^iVBYkzN7fow+w$1u
zHc$rh8bfyjC_6tHKd3Ff3?1&2QOL}yiE_-Hp0gWFP@DCa{F%INny_t&>%vfZ=i#r_
zjOvtZwD6&8Xp`*q`&_W5(pZ&nmEh8uS#mm=WQ%)4!@xRU$RaCMu;tz2jPC~YE0l4;
zN)}A<*`)D^*4y?>bgCPp9%lzX0<ZR2(-r2<1DzKwbQ|M<C~ssz4;y2}lyLPh<qg0d
zB?@HSmf^>`j8^4O%w3?|RcvNi5$``pf^#;ktpR({#nBT6V){F-*ELI*dX@=QH0Zta
zd@P~hH>+~HIol-kRMT&d@L5yvtN8m9_v$KR;72M%g-0Li;y<f1;8#3W?N!vNJPZvk
zuu%80&$frx##eXdbxFqOMov+1)o}#<6m2L0o`>VDMINRP&lYqp64%sJ)(7w?w@g%D
zdyPGivslTox1sAnW+y9yM5WDk`_|jSl|uRIUP1#n&%P%1UP^tW>^z{HMrE2H0TWh7
zE)sdQ<+ZUry7+^c4zUlSBdx)3(I1}>ova@cChxaGf&78S9}}b80+tXU3?yqigBInA
z5Wa==JW!IOOjHjKC#;TLAr)09=avaHddx|;>Bz_FRKii~&^7GDYizbtY23*~v&NC1
zpr$HT%-&_rTbT}UG_w1wc$qb-_g2$DJC<qZR)!g2$`3BFr@NF2<2+1GFD6%ToL+CN
zYE1L{{BF@y)i9AtIU@CO1f#pv>y`oSA?)#VD6_@p5%p-@W#9wgq(4>-%Kc{qv_J0u
zyCk$*MccE@$sq+p0flC&wzfnIT;r0H#<k;#7Bv3vP<bFEbN8J&lV`okm7jX;$*lH1
z2M0em6`UIh&~_`*ePJNi)XB<1E@T7asgU>XTB@#zDVUd&xu57(g8%4B<tQLSY(Iyr
zD?kaZpMLea#1y~vu<r@+`d-<&2o^T{%!uv=`$NdlCie(+NFB9HMHxLW!hxtg8wU*A
zq(AW9{i1wx`!aO#;$tM;B>}EP$Ih#HRNux-r2pQGul)tPCyIW>HVkubwyHsX<06MK
zdF)MvT}#Q<J~7+rq`_wF`%;Gt+6h4ksHBFoO{mVoJg=-K#nX(g*GKZ?0p4y%W?iII
zhaZs6j^dMP+W_}o*+?gmn}Gij!8?yBS9rs3_*t6#aAE&3u&h8kyJ~7B?nQQyT|=(S
zMxt$@N#^{5N(HQ(^U-BwqZv=j-#mZjCRa-SFCn4k6c*(TF^;-6AI7~E3~@J8SkKlh
z7eYMW93FSSieYjzVnIc#h*aen=Zs=E4K(Pmwo`*?_Hz9IBqDZ3D#E6i5fX7%c*P?&
zXew?Ya5m9r=nO*qJ*+m}Uz0tMbEPF<qzMQGADh$pIyYUG6}cO$F;@)wR*F4rJLfvj
zL)r#|GI6)Vw5J5P#t$UJsOrfWwcqvkKY8=Xhwuk%Tx5pAOl_h>!DserZeL^XJTrAg
zY`3R+d(`w(bCto>oXK@BoQ?Dv+VpUR`O98)lo481>?^UAn>gT-9zb=<ANRt1m27mr
zBcb!&08Jz#ml8?2+y|m<{^L6>;nztmLZ({{6Xp~Xo1*qpek~niEMKg5%Gvj8<sg0x
z+m8K1b-v%2v|Fh2p{@n-+v{@5deB$O^iwKlQ-hbYk-V@V1@zx`h{S^XK@WDAhT_6b
zW=_eUVzB~VwJo0155;lEvrWw17B!L)1*UQ`ba8H_f1Y(y(P)2Ey$6!N`-RDzwtCpE
zmEGPwwF~RJ%Yy-$UMCEv(i?l2JE(@8j+hn0D%r5nS7iCiw7=RXJ8a#|uY!7{VIzmq
z6ii9HlY?v(Gv)*1jkcR#cCV=QON*>`p6u^=@62${pA64w87~XvAoM;_4<^^9r$Uwm
zg4hQQYSBPc5n?#P%ULMOuf?us%V~p|Cu+O7w~otiYBRtd<<Vsv_VKjD%s{i=Pjk+e
zBH;nLowfb9;x_+i&%4DtrR<YaT+JpHvP*|@yX<$~ZfOO)Iqn_hIjU}@J}HY_Z(&4j
zQ-&{+j#vMDWtm_-C&%Zp{<KNvTI}c9y@oV&OP;iDdVz#ZmF+gowptbK6WM}d65i}e
zaqLIXKMD{3=y?7C;TwQ$^Zlbvga50&^9*aMPuD$)ihzhT=|m}_bg5FKC|y8$2}J~H
zp$JHaL`8~J0f83~kS;ZJB!nPMs(^Gt2>~fm6Ql$Z@T_;vHM3{tynANPK6~%;<$T}@
zS!;cP)t={m?)&%umQT3Q;)0{ISqzQ=y50`aW()n*rX-vkN>s_KXJ9!=-k5Ha`Q{yj
zQC@r_JzuYq5lPk=c9U?q@MBi6?Z{@m@D|tz_3io5f{y%lc?RXV;^(5*j@<K9qZcJv
z75(i>TXHrh3x6Dkf3FS5@8W29DcdObhO<452}64o9i04}jpgMHq1r3`oVv@)##J6#
z2MV{`q0a$}+K~mMKBa+YeaMqY2h%;-X<EpkKfSQdfq3oit8BSyL4+@{Ov+T4jx0UQ
z7=-k525(H?klppviE)t=Ofq+W#(vMVFQT&1g2(_ZbvflEfn$|SiUPv!L0&fyq~+cY
zebh{2lcN^LTu-`cCKTzJdX{ape~OuAZ~Yb;JGqZ-HBff=e&`PmZfYmOaoh<JWG9zJ
z9bKqp(mFRIH%)-JSn3bY>)~^YsV)NQ>4m1Lw!djodMbuJwnyeg`@<(II);wFL3RNJ
z*YHn~6hOuON1tco{%jontI!eEb(xCE*nLsQ>3!h}#fP=Yn=svU<0RADWl~aCG})d}
zu>Hv1|G63o17QLL50XZtjlKL`xGJP!j?|z}F03jGO9XaAfeeD?tB;otykOm*L#0sK
z&U*18q0iV(C%=T9TkaAUjD^q;vy`RZ-pX=qa(`+WIVMt><+11!xz@9ky}*=E=ws?-
z)wmsU;BFWvr!^jr>bL2YffMbwkw1sxES0kx^VVb=Cw_RH%h-y2SaCA>j_*>ux=Ga1
zMAHt@b?7_*s1XHR#b?q07*{1|5wUf>kQk;rdL6%C*kOgic7!;fV}&T0JWM`6(KEla
zdxwWCKOQMOHp=?j<7HKyEHE#2%f1NX!a@IfbJIPqu0C38;*Q<jsC?>S8t_w#_!68x
zyYo**cevk-zW#hSZ{rU(a#<L$ivUbUio!l1rV6UY+`hk?)yVpm)%MH)d7#A5-D<f9
zdL`=)NUo>X3EaaH{U5hUcRb`M3i-e>wJ=JV*S0dyZu{<z!8@^7%p?D}*2@r>lRw_S
zOUo1gi$d=eL|4`Eb<ZnJk?)C`^PP6@>|>Ykm7}MO8}MVQbu3W~@L|3w1=H;L52McH
zTlO?Rl{gxte`J-`cI-ikudHkys+VZRi2(xJ2N=+TzxvTi%mDx0cfe8nA2~*{TuJS6
zu6LDaQ7)|nzf9Nr#A3>3hL^Siuqlprjkb)NTkM&=^`lEuEsau5_C#_%9#K=aEIuRL
z?xR-WjOLG*g}UUDE)7s!T-PZIosxoMS@y{CH@hVg<czo0LxvdCz5AE@zm*;HSm4bT
zUy*MT!dz>nIX5h}FcZ$h;m-<F8gD7L-D}(U>IRG&!hH!#Uoh6_xB^{<IF){=$COKf
zqg%DxGpfeN)Pf_CZCTLIr>qm;1|<vb<^8YClrB!^k*Ir-D)vbo{5)Z7+@ZHOkxtzS
zuaV#wak7?Efvd#&;YUW7)N%O213KdQPqP|Up*{xzy}?|NPQwmuAA_1SU=-Nldjw()
zOK4~pQMP($PbK_mA9yzF=Op6b%xDHEh{n?`cg#ZUar0{hYgPI*)ETr&a0xi7ZRx}J
zJ!@@iPV?&4m(jXVC7*opM}2LP>%u>P34Tr;RflKPJWKO8&xP?62hQC0me-x1f_%G&
zX7<m;@BiHC5U18zvIRLGKM)$}7VWDW$lq=+7MsW31jh^4Qbt!8(cZGp?OO_-vyj1(
z2Os)fjZ-qVf0-khY+Bo_-*^Pd<wz~cu+rh{@VVoNxr({)F2S3bi)Ij*VIr}oFOM5f
zX@Syu5$YFBb~t$lxB~xXv-9752M(k7&m034Uo`!6&`LwrD6xF=;hoDU8@Xq|n~z24
z@@P++eC3y#5Q|*#H)QcLm&-Hf_sbhdtw;+^D!iaf3$?n8-L1`x)_FS{$&_1PIO~9P
zUdH~OdJ4%zt(Yttb5DUkp$sw=ry}2hnC_vjwu*21JUII}am){|RW~_WA6YdPHr0D>
zh4Y1x&x8ARXsg)7A$CNr^v)aG3vMC6L*wH64AT;yC*QS1cHCaM+@L)1SU95l5l^vH
z%yPW#LStgD4;Voz_I|AXqM$4}0-oS?g+@s=#f9}7%RnAAcg5xmk#i9o1>+o-DZBXE
zZ{j0tu!b_T_slZWn$*z4g`mXM8b0;S_(E<ynyNOc;mIka)0P2#M}2f_)6K135u*5<
zq3e=g6btin%2h-TOg12<-A0^nZ-OkA=C|XY+S2&xwL02~yk~WrG}rxcC}`+R+?R<?
zy*n9V67b8JfneK#QP8V4gN}A{*FGcLtIbCU{hkBQsYb(8YfzW$`8N^N3t47WE-d4%
znv=VTf{k{#gFf}wJmTd;(@$~vp_U8M_ada2AD1>RcKxEbWY%o9YB83t)Ls@bNjB6v
z6<li)ak_*~%03{^K=Huq9(8}_AHDQ{-}k2%|G!mT|J*V3*T?3E9s*WB;I%BY4o&^D
z8&R6V>^0esNF^9k$xon<G`LW1c5-9jZQd%^1;YUP_@;%EAW}UG)2obHd9OQ`0Z5M&
zliZ`PB`*FM%HLDd4r(pt)Sux8!?|G}#w|}+#-)|ztYS4vTn>=0q2I(lSllJEaTFH^
z2zDTqKWBH`8x<*{;?)Vhvt&mMuLPGiuMH)oUGC4ED=E(1&1tutCYcf{zO!PwP#Bw6
z1&Y!i3K}U%AL*}wc6j8MWcF2R7ee{0M(R#vgS}I)N~+^$+PfZ|W1}(4wem{P*T&`{
zOGb?+B1aaUF;__}M^tQg{`kCL_)X0<==T}ypZYJr@`sPlTFSXR4>Soj?uC5|DY(Q)
zq4i)1rC${LWv<bm^o+k3Q#nSxw>K1HG<!&7TJ&$DXCO<VQYuBWe^F>s9R_6~nb^4-
z6Bbmu6I7u?!h3r!KEIL_&O#EsaKh6W&;;7G+JV`0fl-_z_>CnACC8#IVbE#RyxrPE
ze5g<J#b~8ttY-5#WN4_BE<oR$5RuoQAeN<)?~U!C`F?ZXaV>xLt?(B8OYpgCM6;q^
z{~*E~F=FfNS4r=xVp~!AqPzTu*MiK7&0QKYOg=T%)JF`S>4~0~hz3*KHih;M{9K<R
z2H(_3<qmE&n&_A0xI;~{K)dX1wwwiO_B53be|<$9kq%y+WGxd?2D16^BiANq0A!H`
zBeT`gi*@}ogh?fk{p9yKBkh)No&~jIqBm-qV{y1ovH6HGjuWX;pbRYBAY80hmK`;C
z6C^N8dlM8!X=|P4#bU)Rqh1%d*s$F?aKzOe>0&RWF!r=pu+`~QV<j$FybU|uZ%!EB
zZ#k{^lugDhU!N{COqxyD&da_Fo>%hu0^Mkl`;6xJ=5g*A9$8mL-PV(Y>AN`v6Uu9Z
z*0Ogol~7%hmWQm*xAG-Wwd0FrlpY=>tu-9IHJ_UKk!pJ2_qMR50tIC@P?z{vD)Bo|
zDLYyim+3y2<DI141N`~-|DW7teSmW`v^-$6o1+C?NDy3f&ze;c8LL}+{#CfmKr2t=
zimeW_fec}U{}I~qroWv0xdsb;{nD}|A~04angyv8+a2s|(Kl=;#3NP$RZNHKt^2)}
zm%Ijtq&&8zVi%m4mkKILu(DHX!`gzyMkJ|1e4;yP7B*~p42qN9Xm0U0NN8(lm={Si
zRY)ngpf9D5ZU}$ygn8CLZlL<*Ph`!`EO{a=7VKQ1KG{(_g$Xb-z(>Z53j!LrG=PDu
zsEhjH@6S>S3_0?wv=LnHVwO_8ntm@-sZ2|c5t}JR#OV3mezQpj<ftbfLByzyB_p3@
ze22HGM`WPsbtmhJso9_=rH~Iwki_YtV3S#X?<2fjpkQvU&6A~m2Z-Waup|aS0`0QI
zc$=KW6?=vHnIE!xckJTw-0<r2i>dvv9aD^rUrj4np)}5#q_RDq_N^W>C$~3!cohSV
zQ(^h4QeQ$`G9`xLqTGry4Qt2hnjOdN0*33{pQK)`4$O>Sb}lfpe<!q2pQ@+W0dJ&w
z8KRGCWmuocfE6oYm@|TaX4rQ=vs4cVCp29hW`=V1E)J!?wRgFC_k!Q!N<WkaUhWH)
zu};}un=siKC06cHSC*wPKhu89_v6`4x-9#45~M7lE(PSLQ~;OSHWkngT<wCtXuZOz
z53nnDl#J{b!Hll9__!ZeQ_&ziDz(bI3tHa&%xPD5Z0nkVaPzCr=VRZ0T$`wk4ptP`
zwOwjBw%EzNt#yoDK7iBOx_h7o%Q^w)!i_SgMKF~!68=J_ncI@o^~l+BBDp*?W=r<1
zj+0EMk2+zV?Kg1R+rHgP_Z8l931uIE6Kp_9$}C;kg>jHBRv2U0p#izkZ{@q@)=vTh
ziT^h~(Vt?*4oh!aPTrI6RX|NugvLXhif4(7Seu&hxJ?eGuHvJs6F+APOD!{t67NCE
zBIlj#S5&;}&K1Ku<Tk+g-&|OAmW0j8!Ue>d)vksN1?TQ9QRQ~4H~Fb+vse08tQD`D
zzx{ZmcY1LM0s)F;dhcHpB>2#>-%;T6d2GTC*hW8XAm{E$$tKyNV9m@D)Um_n+K61v
z9jluYx?Qa1<MpJBhlJ`6NbKf7mGJQU`q{rzziZQC+s-30it@rWbNgte6jn7GKI#eQ
zwlFWtzqM>(VWFu^lW0SSfPF#wmEb^;rWb5!p&6}Vd+<%Q%r7>~x{mGZXZiY1Qi|j9
z-brhljMz{rRE4}kkK-plPhjan=n9A)qnl{mST*?YbE3$#*#Qewa{!Bdod50im+$Qb
zGDUntkTb0~f6w;GpRKBb1X45H7Vo|cU4Oo3pmz6s;N%w&^M<8;fLRBO3zeWf5n@u(
zkjYaf`YC7gwg769>0x3u&4}CUQjgi(Ig6BW?Hskr`-^5BKbN+o^kp*&GOo51*ofs9
ztXHchhh{7rjIPFgN|Mv!j~f-zRy^<ErcgZkU1svz$-(F`nWsCggP=|S8pca_8ec(h
z?O0T92+N&=a-_xt`xJ}6DYrRip&36va+3iU5bwpyC+w~`?Qn*d_>$y8w8gn8JVy)T
zyQ7rA+{?{?=|oJr`5`3YngTWA(QLH=s>6L{;N@E-=fUU+B?u5=8(p-w)m;Kb1-TE@
ztO2U1yJud2dwGQ&WGYoQD^;hdAdvav(pQsfXX?1BZKL6}kx?k;5ED>jh^HNH2F?=T
zDo*G61%NXuyzXa)DwNG=iapt_5d(Q6dkS8tk0$BVb7Hn&ihw28c6DSVtbT7CJ33d7
zgi9TTo;=K{(~mUqm4?t+>X1MT7@`wN`AL+%a##C(e#tMFUa&osqV{LJXdF^$<}L-;
zHS2DqFX?)$(28FoMuX|qyn>z4LP^3-qn1$hr&*x>tl}3D@pmr~)Ttj*raqeW3&W#a
zZR^ixH!%n0x!5q__pX{5+L&v(yM&f(B`F4eaC#(g8>0KXl8|a}5H;!LaNQl~GHS?4
zWK)&f6Z`HL_S>52f(waHrRSiJ;+-uS=uY;G+?(F;dA+`(ks5t<{ry-9n(nz@Yf2LG
zr%P|3mfs{>!(Lj-FACYT!jb)lYbZI=eM~^i!%Q<CDnemSBvv=gbkx^>WMom)aE{{f
zaJMIat@?I(dEyY4-U7K)yNbt0ZY!oBL>q5YE%?eQS$pz%7UW*^GSA_|vB+tJ%pLcc
z!&HB{N3VKUubVzi^yYN#tM`3$Is_9&5q&;}fG1#l71*#}6jepQ74BBY|7!C=Pj#2*
z_aTv(XN0_$QA;mDK0AgNX>+_tQUUp$^StBxLoCD{9W=V7@tY_wfYV+wpPA_@cD41O
z;G65~?9o!y!TG$s$mzP&zhzHugdFI$gv=#D-F4!7`{AOEX4t~YN$bm_nkOIX;_PiA
z=9Ot7ksXZ+F>41fk#y0BOy8g+Y4vjkw?9!0Xz^Xfm%H_LWc5JCpsV0Zr(cf6!(n_H
z2ulc2oFvO_qwC0?*(QrXys=SF>vuPpt7S_ZWV<VHktJj{Zgs715Pa4XPNdy$-7iZc
z)B%aE$P-cGF)!doumeg#s*Yk4g_VS*BWY-B3`b;F6iUAW;;jD^95th+wKR4^=Yf4Q
zN#gEH`^%Mn=S?q3CEBzt4PlaYaC57%Ge-E+l*^I0EY*Is&?mnrtgpH#@6TSXQMI<b
zU@kf~n<HuZ_MLHG!Rsxa3Ou$02@2oKG9A4))>^l<GT}=*lYWWN_XL@}qE#GolfpVl
zv%ik~nSxp{W2J~_gadKS1Yn@t9xIF1J{Hjj!6%ZtH<BNYl`ln{laMT-=GS!%PrCby
zLc;f%@w;4QUZe$ltFf`~#BHt;D#&j9zPSmqGGhj~sG?UXad6C$k}p3Dgs+bbM;R@Z
zh>v`A-veKz{mNv2Pqgjb`YkuBGhAXcq3>JMU&a*=R4U@Sx2XIV=A;b9l|1drY}c<I
z&P$^LDtt>kOy?Dqwb>>H@65OX#O1sP+f4n&pD%L1ILJi25PVeInEDg2Od^!@p+Vwq
zWCM|dC)?j%>oRD5xY|Cz2VAjqO>@mYjD4{vfQl;xjCW|)E`S)ILD{)Q{U;SJJ}S*0
z)j*gWGi3M*Jukn5rlaPBa=51W%jCwhBksQ_d=Z)26Gw<q;|DfxbXrbLlO}CtYApkQ
zYh_1^hS=kC+M7=_1EkdOgN;1-`2zKGjZxwI1&LO+U6-nAm6MP<5j5B)2Wqjs#4K4e
z(4wyGgK-s+5J)}2WT*=oXWDl#!Yw)4#;S57Soo>C(1^Y$U4*DfY~>bdxL|vB%E+d=
zGK6+ij4b7k!(Jup5=^UAnnbfIa{|k3E!3o$EPFh1A?Egt@l@}m)m4X><2TqZS(s4{
zqxcp$!Bp#~{9_9!R7(rt{Fq!?jrDK8C?+(ZbX7tO%oK*_pMX&iHZ}^a;K_|(l~QuE
zN}R>9t<GE;yLUxBSpmiYCUWxV6p$^6IfcZaQFTBsxUWM1$Al>2%IGw@yj?k_NA)p<
zZ~E`MRbd8`dVk{=6y<pOR``k@W*&4>nMZKy{3_m&;gOJk|LI4W+{Q>V0+%1h++0f+
za<I`al~I*(W$ykJOU1b9=toy-+?7`o(C0grtl<XZ<stTqwp9A1Q#Ue8yFb+Dq<vJ)
zJfI?})pfBysJR&s>mKnvtE{0Qr?x2lMqyDdLH`Y;JtK6v@T4_tC41goLB67ajRy5r
zW<*HI;8UES@ocOA`Iq%#uJ1=pbx$jS@wp||r(sn(Z!FO68@Ly;8a#jOYaP$p(b_N>
z0T5-POflD*@pVDI+x0hILeIf&MeD-1bfi>y+{ODY+_8M%CXt>4RUe3EZwM8%amgjq
zM|Pn3ScBlPAqG&~a%-|-Li@AXHl8f5T`A4)3hiPK9X0%SAFha0bIRg03db}S%&@0t
zPds<4^IH`guz(s^i1z<SEx~_B(`}%}Oz|)NoY|cZuY?+8+S0BalTVp=1c+}Cm|3Xl
zvzdM$#!e3>230Y>q-(ouFZzMk>&=i$xU_y4V5IwU9NfLoN{0=|F^d3V+oEGm6J%Z_
z$GX`_Tj`NK2=lO^`Sz{AbD9vZgfpzyPD(1s^7x!Lb+mJ|=6G~m=~`NCW?6+SNk0PW
zvLcS^A=IQeiU0Jt_WtKYt*`hi_~reCOC7j=KpYNuLh2Zv<@{8anZ+}&qq!<SP{yG8
za&YD7#n3<~RKe?+fqsE;==12-)an*w+ASZSz~25*3zt=b$G)`!NOB#^rgAeNx&2gh
zBDo8JV<$^3Ff8g<xk$LA_4)>Le<@^V?2mc9JLk@a<uZI~Zs%ZKfz;*fL|?<VM#1EW
z#kvIULyRXGg1e)JIyk6cHki{WEXchZ>3>{;x!#mXQwlq^I<cC+fS<w4LgXU%8XB5L
zC$+Mfo?h*BdnW;L<a{>#lJ6z^k@~?>vdMQM6vtjK9H**Jvxb0saiff^WwWSLAH3|C
zDU+Q|<SeZ@3p4x>%K%X;G|<}mMKNQ2@<YHQCw-u9wEk3E_@^Pvf9%K;=N}cKgP-gm
z#DT5GWNJ&v<^=AnuKKx_YL~?aW6Qf~{lH^R6y>40>Zgj+!{p;^<8!Vel=uoZlB|b{
zdKX(RFK*NRqTsR(@SA*4+e`vyvQDY=Dq2gcs$Pi+PfyX7x?&O*>JG&pwl^pg9H%r!
zJP!_455{iY^5m_0<CwJ5?~EQEWW8Z4IxUyn3}txFC={QLhkp(dE3R0#^nVqiX!mjB
z4S!I@TxGH+7xGn)5AWdmi$1m_O>b{~nIv7K?;Cw-h54Rc8_U!21x>3%+%ldMKfTFW
zDAMLLP;G0BlFmoK9S__=F`g*n<G~!3tm#p*ty7`sT%Aw(3&~j*_k+MZ$MAy^wG{2!
zZe>Nv4Ncf{zh7S+)(Y<6b{@8GC}BHkbVvj8Mh?q+L#-Fh)AlCkt0;O`Da_O$T1DHj
zPu`pS2IP#+|D&7{+q-{4Q2MvI@Be8{{=-=AUuz`)UB7q!lVQ@oJZL(j#_}D+2~$+V
z@0<!KXm|S~G&6r|y^SeZGxxWjhBHmEzkA0rJrvbx)IRDf)XX2W@$cb9%~prrDA`vK
zkv}SB;`CM@dHbJF83ZIuU)XxW<MWQw^Hwg0)XL?%5WPrYeldf{;pnB~2_)b0(@w9I
zLl>8gUlbH4+<9nYq7F&hRY7-?!zp^3^U<io^{c0Pj$v~{1VkmwZf69YI-z7xxF(z0
zXg2wVu4U{Ev*d2yz??jnJ=bc{d0(TCrNa4Z_<(iuDIcRB$n`kOn|XDL?SC<p3WGa2
z_7lSQpe<%bpCAxUV@Vx8Kc29$<j>r_XG?E;yWTc)OV?79*P7;2{g81yKX;&`WYB6|
zxxfK<q~Tk)7WFyjzIql~>3&#iYQOJ6a(o*E49noi%ktgI&>n}xmBg&+tf}SncrBG8
z^bB45c<DXu7Cke}Bx&Eagm5SOR=%Ex^sS#_hmCXjVEvnjLF{?5AwlDXhkdffcyR=a
z5=OIY!E4Q43C0I#)4;+T7fx@x`~oQ7W{C`L`@xUS&!uiIDRV17I(>`}W?FW%;0{5n
zjF5-7kwN+6{XTu{uUoR3oTm5poN~XZvHXnu_BXD`XJ7`jN{Ad$y&cR>T)l^%{q6yA
zS3SM6Y8U(cf!@#V58HrGPK{&&!|>cDmJn!7iMn`2Qx2ZAg5K`|3EifvVo44Wjdgsv
z=dM`mE%y1Pn9R;I;c}+SJ|A--0zdxPT8y)Cl(bq`*vL#5^!6U;!-{DQxg96l9M~o;
z`dlPw23KzbzM-}Osd8CcF)h-_W~<R&SqG@W|8pywE5p<eTb2-Unc29SJC0+F7yOjZ
z9))|uQZ}o5lGSU}`_?{7*1Ug3uSy5I4+Wor@jx4j=9k)zAKrvV4^F$khkwq;z%3CA
z;HVr%xZTWDeZoc8w0UWw8Lg8M-lIg`$Av-O#oZQkoh$o-VwGLb`j=-1!qikIGW?A|
zsqfVvw#X$W=9;W0x)qB4d2l&;8O9o-i$DCLPvMM~#8>RDD_liTWY61mT-J2(B8!!?
zTLeocTAPG4?f*$6{<B#8?|fbm_^S}~KYB3z(Y`+nLQ2$QSs~85y-;RW%lHq~fvpZT
z;f0UoXFk0QKNl}$%kl%pJ&xuLd88|4lvp97tG$I;(R#F^0uU0#tEecaXpV{hmzb(b
z*KdZ)&Fditp6Xei4y8I<0v?C1tG@q94&`ytCs3@}#Iau@?r6I>e*WS<i7HUwacODo
zg(q5@6oP9vvu=8~&~`ojO-5Dsp19PCcr=*m1`(3$Tsw~yoyZN@rM7<!Tfdg8Ns(<%
zo(niOD+}Qpqwn;5iA0C{lQvrTCGU18*NG9E4z8!I34t)X4xdIsg!%Pq)OYpz*1g$S
z63+1~mYwzqxhW|wDXY81+~s9gR(X(;x%agEag(rjw|!5CQ9P^cf%NmZ)`OW-r+V`3
ziJVRz%5OUY?z_K*6niYceaC;|w@RqYL)HFke5xIlvAMje*1w|<0gT(fBM(vfMgKnT
z`)m5px4$C~eW1e}`5*KU3d&>?aL!!iYART7DgxKLSQwj6g4YjQ^(p6M^V#x>vUOJ1
zGZ%y-Vsn?!S3nbJjT1!qa@b;c-8N@mj$Ty0A8z&vcqr)DgS(g}{dFMn20)NgQZszM
zd*r?E0>1zLBs+vMKjZ;{Fk0Vj@4`|xD~jxSG%N9xQ&Rp5TO$=%d;$R*vAz5p6saOo
z-8igspKiP{OCVF*LE>%aSz5+tUbEQ|P2%CFXD3yz1r*Kq-**u$h<YQ?YG~qa1`es(
zS)<}V<qy6PP7>M3i0mv&A;^4BL$WkkWImE)64=c_zE3hH{B$aI`_4g%HRa7_<=8Ok
zv=jnU?N1VpK=k1qV_lC^ILw24`<G6UFp$(L%@*u(z@Y6}vf8(Oe(#1Y$(cEE$M7E0
z<E03mbzv;vgY<kkafsmvNbZ=>IzN#~)9vHpk@+BXO@jUI_6s<c?<CPy$fY4^og=)W
zoFaWS<w>l#;lN!svaX+9^>ZmV{(=U7UB5eoJK9HjVz+~fGD3$&ErqfRi_y7^c{a0C
zDQ;;ELevL}?+5^p>2Kg<WaE>M|Al8CkC+Kw5*YfNFzw$B6CN#whT4gArDh3*u5CN{
zhV$D#pSL|dV*@0~{{~ovNNQau=b)R#R@;;lw3OFfGAhEyJCi6Z@_(XQbDSwT3-#vT
zo!RYfqd(Jy-@#QT6kR;9>1CX=l_w2(UAuIoOlVJd%5o?cLsBO494?eA42;N+Jsnk<
z#UL<jBKphQ!nKNiB}bZkN<8-v$#tYPUmaR#zds{n_Cc-sQob<7J#t)d_J06)g_yd%
zhf|N^c)mg6&(Ik<CN<|IEC}h<yN(MzI;beZ0_!{+DCs3VuweKdpx0)~elCcE1jTu7
z{>UfVw)v+jY@sY8O3Lmn&-ac!C_bmbzV+(NtU?u472@5gAXyPje7e}5#q(400XWdz
zCzg6qaM=^_MDYqVpDNRiFkY;hIiVO>xpDkc;>*)Jv2<)^AM1jTR0aTr`dlaBX?Mp?
z43<TOx?dzNU;WKu9L-gEb6&p(S*$aXpKq_n1}EF*n;eCaSRrU2ws!ss5cth~8mWoD
zuMyC5d1S}yjIsUt{RqmF1lJQ4MDI_UXZbN@dYreU_CwGG8+Nyhv$utJfu<AnbNC;%
zp1<$=?`S`#`7G=md@sjJ^mSYd!=O2c)%b%$U>%f#=<6bWjG9iPl_~LlVQ=N&fqG|l
zh7VZ|l^uwg?MU0$y7cFBHeba>`B$OOOqwMV`*OU@$BVaE)4w~KpN}s}@=Kb|0gTA)
zGlb$3I*#Jg^K!%q)0QPVpDtR)_2c@Lf~q8YNP2pE^4JUAV8okN2QBpQZU7t-+upq#
zW}pfTAU_Xt0oZunnmX(>Q>UTT8F+<D>abH<!_OO@r0=D0g@xzqn?o@rX)C%U`5UJ}
zFsC?6-7;vuUG7Xv*ZNU`<@5y+O7t0wq7x-p`xpt1JT88vLJtYB*Q|$D=o=&&9tZTy
zaB3yoP6+)#k=}hT*;>2<9>xPAx}tsGy9bukYz33vvvxnr&r2v)DVPkle(rJyDmqls
zx7%wcZd!|)&yLFdV*=q?9x@^CFDH{Qgt5oD0SIFy0lL#n>P+(y9nD1X`3&9*dz8)3
zE*@Wzm)U%7X3EN>B;Y(*jFk5>cYd<g7p|Lr)cS$A(P_ifm^z~|;raYiLFlfA`>j|h
zd-gl)P@_{;!Fc$Z+juMf7J#oqR}!;GG5~m(Wegct)(~FS<FDoA@iPD2F)EN<Wg%=M
zgV2wdn(Qi*q7AiIOXc}0lFSGCtPanDW!{(TBN(ifzI!^cvENX<;=3C=N-L(V+ovo7
zMDy1IbN@w-Uvtj+({5Q)|J$>G&Y<wBq_?Kph(nGv0{SiC@tH$J?%tg2sCw0_FJI;r
zb?z|pKWfi4VfLdiOH5x4a5hZ2bauj0Hl^;wV&iGLn2@i%EzewjsV>%bxE>(nuCA{A
zr7E*Z>q10eY`=&{z;596YMY+V`L#pX9q!V0wLGg43z&_IVBgL|Vs26r8W*-}D*&&P
z%uc&F$M#5w5AGH!)Aa3v&CIR2OT)>BFLVp+iGJB%6LMQ|-T|wp5nll&!Py)#^amJ|
zd~gS^v1L$zQK>zMqwNX;tU7OOqoq*-%SD$L3Vu=aYaTjmDN_q*C+UV5EYeSV9+VO_
zo=r4ZnuZf%)Dn1wHxxUoTc%WgBnmiiwxkV!w>U_sW~C0&&2Qab$jDok$ASAQOt*Uo
zJQF5Wj5Vp1at&x}`UgN*vv(=ym{Z1+OamuWQZC7Ly8sIznN6VVE3I8`TVA+YW%U)U
z<zRPODG~hs((#$(&*e~1`0sJ%*aC>P_(E4sgj`nKM<{ZDj#>Gh`^)j6YfG+;25sE1
zS0_2}g-kDJ6|pM9&X-n34FafZ!fBv(#4ffnu^4^t(#Oy2nxC(wm<$P+zZjp69(pFt
z7sADqqtWlsYLq~mXur`n)mwkLEMx05qMOX5u^`)sKg409RCtMlU&~Q0gbP1>sokGi
z$rL6<kdz+q7-wyF`gUC76M}A*0-aHzpuD`+T<xBDm#WZhQ%#_yPI~p1KJJf~ox7y{
zPdu=HxzqmOxSQBN6SkHSf98%*u&kK2)&Y)dIv{CR6Vs&--i6M+ktW)e5at^bBl9WV
zh&kq<pk4xJZK1P+k??*dWs_;E2pdwNny8%Vw4AlQg3@|gBp6qJpu+hI`%v~WtMRAv
zIxU%2c2T{ZpTEDqB5|wR!eYyF9G-4yTCmcHv<*r=p?3=F#_-QyuO5y)yy809A1F{?
z@%0_Pt^~^OHMoF0Qzu=$<%a9HpxXb{6`HY<h0EeOG#%fm*!wEhclA>f3YGFCf^F<+
z)?5_Lhf7J&#a70EvM&{ZzNG<v>pu+%zA1=K^S#Em^#j#MmKCKB63eIrpWM9o60%(y
z?!mm`e7sV&M8{;9aDS|>l_NmoVVcIozOs17yey#bKeNOdHDs9ntaN|WdHTRxZknaP
zDatKIphnHdTNpz^$C#DvYQc3(O!ziuHXJ7_vdU4|{cqq(6ceGg9R{A@)EX6&QPef+
zku$J>W(XW4%M4t5G4eU>S*N3#NN$F5n}5-wTxBrwe+-P$v?D-=z^)#Of@W3CCI_*{
z+DFa3Aqm5A_bo^3-jqLSZLP0KJ@$e+r!k4Rd)+qm;(DBub&%iiEjiqNY8*pU>`dd~
z0RAxx8(+o0AS+w&wOYiNMcm*{GaeBXvLSlWT4B>lHEl~?N?ibQmsS(!@G14$UB~$5
zo2TOXpco}?92FtFBcKu-4a5)P-y*~6$AX|@g`(v?AKc36K+_nRkXtmF8vS2F`sV!`
zPZNJp^qzFx!6;4sfS<?%OJjs?9-l^&e^KyUn*=ItI28hzw&??<9X5)52a{S-N^D*<
zEy48&j{8BPMoFs7Dfzs7w;<1+Xk9+!)lk)#0-y19_p~Bv;F@=?%+IV_mJdrQFC*UD
zXM=CQ*5h!t^B=Ig{F9N2)F)ddQx=E#zI7<us)G;)^9pY3Cm)~IPu3z8k8kFEI9nZJ
zw<OB?Sq0QILB97<t;nUy-8A)K(srkiden!yB*iOb9A^gySnl!DU=RYtBZ=W1G*3Y3
z(Nj;l0DoKP+!M{E!1HyYPkqi>AwLTC$7|U;K9Pt?XM6Gb^>Ii9@)+EIe8QxVKpA3%
zAL`WE=;yu>s){wLi*wA>ov4fXDujQ~8U7A3*To#i@VJkk#<-?x@E&CiA`T`I5*pTN
z*%*UuLrdZPW8C=g8c|{*G6>o!&Etqje!uEZ2_#d+65jc3x!=b5p7F&+-MJLdiqWyk
z#C;DMFvYU$i(i;`hB4|wwIe&TuaB)}LRyly7{OEd(FfSCz5&CMUk^SGOS?7@G(Ktj
z2=F-NZ4gh6#dk%Gyz8^K5Y1`u_lbx5EdL{jm12IJuh^Rw8KQe|kRrmom5;4g>jrX$
z0I;fr!5Jz;<q6CI2w?D=B%|!OvFWG2GC!=d+<bZvn8cPwLw!_`$a?4VpdQ5yv?d<W
zNl%X<3y#-UfEK~AsQUy}@I~wSfS7yr6OPRpBYWx*_A=$wLp*8HhH1ueyfJ(M3_8m%
z(C|EB({l0`9<~}{W-(cB*e<ONlVtNl?r?Hq;!qq>;LZCrFf?d$#!%^X%A?1yEr^$b
zx?CBD?)J#&`24)7-q3d|Uu|opz2L&e;CkQTlgoK%UOauLINMx^Lw)_2(*1l$z&Ogl
z%caD5l1FZAwMD`C(_0~WdeKZsdyW%><eV_YT?>vU{GnKn39rKs#aMFk3B1s8`G?$P
zEFRv!yYW2wX5P!+H8CC|FOq%;xPDB$q7XTmM9X<3J=5}t2r)SwXv~&>C`m1j{(9j1
z;4O>GrA*dysT7Va#$3DdEW0@`VC=-v#mJYiYXQC}yF3>d<%E*iqOkA`8(CQ-BeyYB
z59jm%C6{ej8~6EkgL^KLQOX<;@WfEMg>xZKChFXwW+va1)Cyc+53;YJYN|%-(nac4
zTOMC$ym(>4dy<w3b-f^whWo3&j_{7d>8Z(~jpa@u30j^@>iwdq)*CvAl_gm~J&r?T
ztR^E)d4accQxZ4t;^JXHvZXr7xcdElu7ji~r?~Go8O-HB_j2O9<j(Vlgnl63O<l)~
zU~LFnsX2J}`cF92q>}v7iN104aHiJmeeo!PXzHtSpRPQsWi0P!QYeq%^qHq3z@wgn
zsIc&M<V1bkT}_p1b3eA2dV^(VG)gc>Oh=wMb=oC-@5I=Dt0v9F8xm)2H5(84_e13X
zxB;(hhUr{0jBy5C{o_Z|;Y4+X%RH!B4q13ond!(M&*v2=*~N3jZ74I_d8xMMBX>WA
zyv*vU06?!~0f1~93Zhstf0{o=ig#VKfT&|S_PE~h&5bjw4fRP9Ax#WXbD!<rn|-xq
zh2zYLF$+#_T5<HmtVIy(+JRX%sAX(fMHD?ID&ker$)Ap}RCjjRZLIBBJs+#Vg*JAe
zQc1*dGj#)XNERVIgkcGMc@4vSToLQ2t8p*K^i#55oc)_8S)O_NNHfl*<isR%u1Cos
z^T3FF#%u)}igYzLwroCaI00ejJoh5fOE6GN_sZPjx%<I({)|JOJ|4KGRM|5r<B{hz
zuf0-OsW9udQE0^Lx3}DXgg1*(xen#I_>C-wqosvkD2Ltr^evn4O;D>q?Ui}z!Gglw
zOKu?0jVsib$)fa&7aFO$b}qF>p{&Kq7RzR~Zpf46>=*mLikppQ=mexLQhv~4sBfvi
zH1hE9F5g9rYPltgy^Ff9{I$K7w5%DO?Z%cG>m3K5K0O9hv)my+cD3kmZ;TWjO@sho
zNtBad=W4`ZC&ngc2~O0k+^VOOAa)p|$ISFB3TR4j1vLQ-YM<VW;4X90Yzb3?{GCXA
z#1;fO)TbbUm7eSCNwgDsd=JiHz2WC6IS^fAJvCQ!a%1K0rYYtYjh&+*_<2NL&W@-D
zw(xYrzZTtFq4N<~jBD=~g(BPrGQAe+4k)F+BaZ%PW<kmLPb$5$|6u<=#zXj^mGH_5
zo%-}~AcNFJTDxGUQ1p0;?Lb3Q^y2MQL)o<OdrNl&&+%xesc}k^=oDRlFFNo2ovKSE
zTo?jw7)1fzOmrxKb2S=6Y~|Iw67FD4E{5caJZYrnc*{L_=p;5^qQ3R9RH5`{ex7w$
nStzjnB#|fMk0joIW!wLsUH$Ksnf~si{H5XhFWgS`Yx;iyo?w)7

literal 0
HcmV?d00001

diff --git a/docs/_static/css/tabs.css b/docs/_static/css/tabs.css
new file mode 100644
index 000000000..95765dff6
--- /dev/null
+++ b/docs/_static/css/tabs.css
@@ -0,0 +1,91 @@
+.sphinx-tabs {
+  margin-bottom: 1rem;
+}
+
+[role="tablist"] {
+  border-bottom: 1px solid #a0b3bf;
+}
+
+.sphinx-tabs-tab {
+  position: relative;
+  font-family: Lato,'Helvetica Neue',Arial,Helvetica,sans-serif;
+  color: var(--pst-color-link);
+  line-height: 24px;
+  margin: 3px;
+  font-size: 16px;
+  font-weight: 400;
+  background-color: rgb(241 244 249);
+  border-radius: 5px 5px 0 0;
+  border: 0;
+  padding: 1rem 1.5rem;
+  margin-bottom: 0;
+}
+
+.sphinx-tabs-tab[aria-selected="true"] {
+  font-weight: 700;
+  border: 1px solid #a0b3bf;
+  border-bottom: 1px solid rgb(241 244 249);
+  margin: -1px;
+  background-color: rgb(242 247 255);
+}
+
+.admonition .sphinx-tabs-tab[aria-selected="true"]:last-child {
+  margin-bottom: -1px;
+}
+
+.sphinx-tabs-tab:focus {
+  z-index: 1;
+  outline-offset: 1px;
+}
+
+.sphinx-tabs-panel {
+  position: relative;
+  padding: 1rem;
+  border: 1px solid #a0b3bf;
+  margin: 0px -1px -1px -1px;
+  border-radius: 0 0 5px 5px;
+  border-top: 0;
+  background: rgb(242 247 255);
+}
+
+.sphinx-tabs-panel.code-tab {
+  padding: 0.4rem;
+}
+
+.sphinx-tab img {
+    margin-bottom: 24px;
+}
+
+/* Dark theme preference styling */
+
+html[data-theme="dark"] .sphinx-tabs-panel {
+  color: white;
+  background-color: rgb(50, 50, 50);
+}
+
+html[data-theme="dark"] .sphinx-tabs-tab {
+  color: var(--pst-color-link);
+  background-color: rgba(255, 255, 255, 0.05);
+}
+
+html[data-theme="dark"] .sphinx-tabs-tab[aria-selected="true"] {
+  border-bottom: 2px solid rgb(50, 50, 50);
+  background-color: rgb(50, 50, 50);
+}
+
+/* Light theme preference styling */
+
+html[data-theme="light"] .sphinx-tabs-panel {
+  color: black;
+  background-color: white;
+}
+
+html[data-theme="light"] .sphinx-tabs-tab {
+  color: var(--pst-color-link);
+  background-color: rgba(0, 0, 0, 0.05);
+}
+
+html[data-theme="light"] .sphinx-tabs-tab[aria-selected="true"] {
+  border-bottom: 2px solid white;
+  background-color: white;
+}
\ No newline at end of file
diff --git a/genindex.html b/genindex.html
index c5b920caf..973a17649 100644
--- a/genindex.html
+++ b/genindex.html
@@ -3,12 +3,12 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <title>Index &#8212; SLEAP (v1.3.3)</title>
+    <title>Index &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -31,8 +31,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -41,22 +41,19 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'genindex';</script>
     <link rel="canonical" href="/develop/genindex.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
     <link rel="index" title="Index" href="#" />
     <link rel="search" title="Search" href="search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -162,6 +159,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -398,6 +396,8 @@ <h2 id="A">A</h2>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.add_video">add_video() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.nn.architectures.hrnet.html#sleap.nn.architectures.hrnet.adjust_prefix">adjust_prefix() (in module sleap.nn.architectures.hrnet)</a>
+</li>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter">AlbumentationsAugmenter (class in sleap.nn.data.augmentation)</a>
 </li>
       <li><a href="api/sleap.info.align.html#sleap.info.align.align_instance_points">align_instance_points() (in module sleap.info.align)</a>
 </li>
@@ -461,9 +461,11 @@ <h2 id="A">A</h2>
       </ul></li>
       <li><a href="api/sleap.nn.data.instance_centroids.html#sleap.nn.data.instance_centroids.InstanceCentroidFinder.anchor_part_names">anchor_part_names (sleap.nn.data.instance_centroids.InstanceCentroidFinder attribute)</a>
 </li>
-      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.append">append() (sleap.io.dataset.Labels method)</a>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.append">append() (sleap.instance.InstancesList method)</a>
 
       <ul>
+        <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.append">(sleap.io.dataset.Labels method)</a>
+</li>
         <li><a href="api/sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline.append">(sleap.nn.data.pipelines.Pipeline method)</a>
 </li>
       </ul></li>
@@ -476,10 +478,6 @@ <h2 id="A">A</h2>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Node.as_node">as_node() (sleap.skeleton.Node class method)</a>
 </li>
       <li><a href="api/sleap.nn.paf_grouping.html#sleap.nn.paf_grouping.assign_connections_to_instances">assign_connections_to_instances() (in module sleap.nn.paf_grouping)</a>
-</li>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideo">AsyncVideo (class in sleap.io.asyncvideo)</a>
-</li>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideoServer">AsyncVideoServer (class in sleap.io.asyncvideo)</a>
 </li>
       <li><a href="api/sleap.util.html#sleap.util.attr_to_dtype">attr_to_dtype() (in module sleap.util)</a>
 </li>
@@ -487,7 +485,7 @@ <h2 id="A">A</h2>
 </li>
       <li><a href="api/sleap.nn.config.optimization.html#sleap.nn.config.optimization.AugmentationConfig">AugmentationConfig (class in sleap.nn.config.optimization)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.augmenter">augmenter (sleap.nn.data.augmentation.ImgaugAugmenter attribute)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter">augmenter (sleap.nn.data.augmentation.AlbumentationsAugmenter attribute)</a>
 </li>
       <li><a href="api/sleap.io.video.html#sleap.io.video.available_video_exts">available_video_exts() (in module sleap.io.video)</a>
 </li>
@@ -681,6 +679,8 @@ <h2 id="C">C</h2>
         <li><a href="api/sleap.io.format.text.html#sleap.io.format.text.TextAdaptor.can_read_file">(sleap.io.format.text.TextAdaptor method)</a>
 </li>
       </ul></li>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriter.can_use_ffmpeg">can_use_ffmpeg() (sleap.io.videowriter.VideoWriter static method)</a>
+</li>
       <li><a href="api/sleap.io.format.adaptor.html#sleap.io.format.adaptor.Adaptor.can_write_filename">can_write_filename() (sleap.io.format.adaptor.Adaptor method)</a>
 
       <ul>
@@ -825,8 +825,6 @@ <h2 id="C">C</h2>
 </li>
   </ul></td>
   <td style="width: 33%; vertical-align: top;"><ul>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideo.chunks">chunks (sleap.io.asyncvideo.AsyncVideo property)</a>
-</li>
       <li><a href="api/sleap.nn.data.identity.html#sleap.nn.data.identity.ClassMapGenerator.class_map_threshold">class_map_threshold (sleap.nn.data.identity.ClassMapGenerator attribute)</a>
 </li>
       <li><a href="api/sleap.nn.config.model.html#sleap.nn.config.model.MultiClassBottomUpConfig.class_maps">class_maps (sleap.nn.config.model.MultiClassBottomUpConfig attribute)</a>
@@ -872,16 +870,16 @@ <h2 id="C">C</h2>
       <li><a href="api/sleap.nn.tracking.html#sleap.nn.tracking.Tracker.cleaner">cleaner (sleap.nn.tracking.Tracker attribute)</a>
 </li>
       <li><a href="api/sleap.nn.training.html#sleap.nn.training.Trainer.cleanup">cleanup() (sleap.nn.training.Trainer method)</a>
+</li>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.clear">clear() (sleap.instance.InstancesList method)</a>
 </li>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Skeleton.clear_edges">clear_edges() (sleap.skeleton.Skeleton method)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.clear_suggestions">clear_suggestions() (sleap.io.dataset.Labels method)</a>
 </li>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideo.close">close() (sleap.io.asyncvideo.AsyncVideo method)</a>
+      <li><a href="api/sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle.close">close() (sleap.io.format.filehandle.FileHandle method)</a>
 
       <ul>
-        <li><a href="api/sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle.close">(sleap.io.format.filehandle.FileHandle method)</a>
-</li>
         <li><a href="api/sleap.io.video.html#sleap.io.video.HDF5Video.close">(sleap.io.video.HDF5Video method)</a>
 </li>
         <li><a href="api/sleap.io.video.html#sleap.io.video.ImgStoreVideo.close">(sleap.io.video.ImgStoreVideo method)</a>
@@ -1011,13 +1009,17 @@ <h2 id="C">C</h2>
 </li>
       <li><a href="api/sleap.nn.architectures.unet.html#sleap.nn.architectures.unet.UNet.convs_per_block">convs_per_block (sleap.nn.architectures.unet.UNet attribute)</a>
 </li>
-      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.copy">copy() (sleap.io.dataset.Labels method)</a>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.copy">copy() (sleap.instance.InstancesList method)</a>
+
+      <ul>
+        <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.copy">(sleap.io.dataset.Labels method)</a>
 </li>
+      </ul></li>
       <li><a href="api/sleap.nn.tracker.components.html#sleap.nn.tracker.components.FrameMatches.cost_matrix">cost_matrix (sleap.nn.tracker.components.FrameMatches attribute)</a>
 </li>
       <li><a href="api/sleap.nn.training.html#sleap.nn.training.create_trainer_using_cli">create_trainer_using_cli() (in module sleap.nn.training)</a>
 </li>
-      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.crf">crf (sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.crf">crf (sleap.io.videowriter.VideoWriterImageio attribute)</a>
 </li>
       <li><a href="api/sleap.nn.data.instance_cropping.html#sleap.nn.data.instance_cropping.crop_bboxes">crop_bboxes() (in module sleap.nn.data.instance_cropping)</a>
 
@@ -1087,7 +1089,9 @@ <h2 id="D">D</h2>
 </li>
       <li><a href="api/sleap.nn.training.html#sleap.nn.training.DataReaders">DataReaders (class in sleap.nn.training)</a>
 </li>
-      <li><a href="api/sleap.util.html#sleap.util.decode_preview_image">decode_preview_image() (in module sleap.util)</a>
+      <li><a href="api/sleap.skeleton.html#sleap.skeleton.SkeletonDecoder.decode">decode() (sleap.skeleton.SkeletonDecoder class method)</a>
+</li>
+      <li><a href="api/sleap.skeleton.html#sleap.skeleton.SkeletonDecoder.decode_preview_image">decode_preview_image() (sleap.skeleton.SkeletonDecoder static method)</a>
 </li>
       <li><a href="api/sleap.nn.config.model.html#sleap.nn.config.model.PretrainedEncoderConfig.decoder_batchnorm">decoder_batchnorm (sleap.nn.config.model.PretrainedEncoderConfig attribute)</a>
 </li>
@@ -1154,8 +1158,6 @@ <h2 id="D">D</h2>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.delete_suggestions">delete_suggestions() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Skeleton.delete_symmetry">delete_symmetry() (sleap.skeleton.Skeleton method)</a>
-</li>
-      <li><a href="api/sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig.delete_viz_images">delete_viz_images (sleap.nn.config.outputs.OutputsConfig attribute)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.describe">describe() (sleap.io.dataset.Labels method)</a>
 
@@ -1344,6 +1346,8 @@ <h2 id="E">E</h2>
       <li><a href="api/sleap.nn.system.html#sleap.nn.system.enable_preallocation">enable_preallocation() (in module sleap.nn.system)</a>
 </li>
       <li><a href="api/sleap.io.video.html#sleap.io.video.HDF5Video.enable_source_video">enable_source_video (sleap.io.video.HDF5Video property)</a>
+</li>
+      <li><a href="api/sleap.skeleton.html#sleap.skeleton.SkeletonEncoder.encode">encode() (sleap.skeleton.SkeletonEncoder class method)</a>
 </li>
       <li><a href="api/sleap.nn.architectures.pretrained_encoders.html#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.encoder">encoder (sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder attribute)</a>
 
@@ -1436,6 +1440,8 @@ <h2 id="E">E</h2>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.export">export() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.nn.inference.html#sleap.nn.inference.export_cli">export_cli() (in module sleap.nn.inference)</a>
+</li>
+      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.export_csv">export_csv() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.nn.inference.html#sleap.nn.inference.export_model">export_model() (in module sleap.nn.inference)</a>
 
@@ -1454,6 +1460,8 @@ <h2 id="E">E</h2>
 </li>
       </ul></li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.export_nwb">export_nwb() (sleap.io.dataset.Labels method)</a>
+</li>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.extend">extend() (sleap.instance.InstancesList method)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.extend_from">extend_from() (sleap.io.dataset.Labels method)</a>
 </li>
@@ -1469,6 +1477,8 @@ <h2 id="E">E</h2>
 <h2 id="F">F</h2>
 <table style="width: 100%" class="indextable genindextable"><tr>
   <td style="width: 33%; vertical-align: top;"><ul>
+      <li><a href="api/sleap.nn.tracker.components.html#sleap.nn.tracker.components.factory_object_keypoint_similarity">factory_object_keypoint_similarity() (in module sleap.nn.tracker.components)</a>
+</li>
       <li><a href="api/sleap.nn.architectures.resnet.html#sleap.nn.architectures.resnet.ResNet101.features_output_stride">features_output_stride (sleap.nn.architectures.resnet.ResNet101 attribute)</a>
 
       <ul>
@@ -1483,7 +1493,7 @@ <h2 id="F">F</h2>
 </li>
       <li><a href="api/sleap.io.format.filehandle.html#sleap.io.format.filehandle.FileHandle">FileHandle (class in sleap.io.format.filehandle)</a>
 </li>
-      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.filename">filename (sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.filename">filename (sleap.io.videowriter.VideoWriterImageio attribute)</a>
 
       <ul>
         <li><a href="api/sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig.filename">(sleap.nn.config.training_job.TrainingJobConfig attribute)</a>
@@ -1666,7 +1676,7 @@ <h2 id="F">F</h2>
       <li><a href="api/sleap.io.video.html#sleap.io.video.MediaVideo.fps">fps (sleap.io.video.MediaVideo property)</a>
 
       <ul>
-        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.fps">(sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.fps">(sleap.io.videowriter.VideoWriterImageio attribute)</a>
 </li>
       </ul></li>
       <li><a href="api/sleap.instance.html#sleap.instance.Instance.frame_idx">frame_idx (sleap.instance.Instance property)</a>
@@ -1714,7 +1724,7 @@ <h2 id="F">F</h2>
 </li>
         <li><a href="api/sleap.nn.architectures.upsampling.html#sleap.nn.architectures.upsampling.UpsamplingStack.from_config">(sleap.nn.architectures.upsampling.UpsamplingStack class method)</a>
 </li>
-        <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.from_config">(sleap.nn.data.augmentation.ImgaugAugmenter class method)</a>
+        <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config">(sleap.nn.data.augmentation.AlbumentationsAugmenter class method)</a>
 </li>
         <li><a href="api/sleap.nn.data.instance_centroids.html#sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config">(sleap.nn.data.instance_centroids.InstanceCentroidFinder class method)</a>
 </li>
@@ -1828,8 +1838,6 @@ <h2 id="F">F</h2>
       <li><a href="api/sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader.from_user_instances">from_user_instances() (sleap.nn.data.providers.LabelsReader class method)</a>
 </li>
       <li><a href="api/sleap.nn.data.providers.html#sleap.nn.data.providers.LabelsReader.from_user_labeled_frames">from_user_labeled_frames() (sleap.nn.data.providers.LabelsReader class method)</a>
-</li>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideo.from_video">from_video() (sleap.io.asyncvideo.AsyncVideo class method)</a>
 </li>
       <li><a href="api/sleap.nn.architectures.resnet.html#sleap.nn.architectures.resnet.ResNet101.frozen">frozen (sleap.nn.architectures.resnet.ResNet101 attribute)</a>
 
@@ -1941,10 +1949,10 @@ <h2 id="G">G</h2>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.get_next_suggestion">get_next_suggestion() (sleap.io.dataset.Labels method)</a>
 </li>
-  </ul></td>
-  <td style="width: 33%; vertical-align: top;"><ul>
       <li><a href="api/sleap.info.write_tracking_h5.html#sleap.info.write_tracking_h5.get_nodes_as_np_strings">get_nodes_as_np_strings() (in module sleap.info.write_tracking_h5)</a>
 </li>
+  </ul></td>
+  <td style="width: 33%; vertical-align: top;"><ul>
       <li><a href="api/sleap.info.write_tracking_h5.html#sleap.info.write_tracking_h5.get_occupancy_and_points_matrices">get_occupancy_and_points_matrices() (in module sleap.info.write_tracking_h5)</a>
 </li>
       <li><a href="api/sleap.util.html#sleap.util.get_package_file">get_package_file() (in module sleap.util)</a>
@@ -1991,6 +1999,8 @@ <h2 id="G">G</h2>
         <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.LabelsDataCache.get_track_occupancy">(sleap.io.dataset.LabelsDataCache method)</a>
 </li>
       </ul></li>
+      <li><a href="api/sleap.info.summary.html#sleap.info.summary.StatisticSeries.get_tracking_score_series">get_tracking_score_series() (sleap.info.summary.StatisticSeries method)</a>
+</li>
       <li><a href="api/sleap.info.write_tracking_h5.html#sleap.info.write_tracking_h5.get_tracks_as_np_strings">get_tracks_as_np_strings() (in module sleap.info.write_tracking_h5)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.get_unlabeled_suggestion_inds">get_unlabeled_suggestion_inds() (sleap.io.dataset.Labels method)</a>
@@ -2136,7 +2146,7 @@ <h2 id="H">H</h2>
 </li>
         <li><a href="api/sleap.io.video.html#sleap.io.video.SingleImageVideo.height">(sleap.io.video.SingleImageVideo property)</a>
 </li>
-        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.height">(sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.height">(sleap.io.videowriter.VideoWriterImageio attribute)</a>
 </li>
       </ul></li>
       <li><a href="api/sleap.nn.architectures.hrnet.html#sleap.nn.architectures.hrnet.HigherHRNet">HigherHRNet (class in sleap.nn.architectures.hrnet)</a>
@@ -2161,7 +2171,7 @@ <h2 id="I">I</h2>
   <td style="width: 33%; vertical-align: top;"><ul>
       <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame.image">image (sleap.instance.LabeledFrame property)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.image_key">image_key (sleap.nn.data.augmentation.ImgaugAugmenter attribute)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key">image_key (sleap.nn.data.augmentation.AlbumentationsAugmenter attribute)</a>
 
       <ul>
         <li><a href="api/sleap.nn.data.instance_cropping.html#sleap.nn.data.instance_cropping.InstanceCropper.image_key">(sleap.nn.data.instance_cropping.InstanceCropper attribute)</a>
@@ -2184,8 +2194,6 @@ <h2 id="I">I</h2>
       <li><a href="api/sleap.nn.tracking.html#sleap.nn.tracking.FlowCandidateMaker.img_scale">img_scale (sleap.nn.tracking.FlowCandidateMaker attribute)</a>
 </li>
       <li><a href="api/sleap.io.visuals.html#sleap.io.visuals.img_to_cv">img_to_cv() (in module sleap.io.visuals)</a>
-</li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter">ImgaugAugmenter (class in sleap.nn.data.augmentation)</a>
 </li>
       <li><a href="api/sleap.nn.viz.html#sleap.nn.viz.imgfig">imgfig() (in module sleap.nn.viz)</a>
 </li>
@@ -2237,7 +2245,7 @@ <h2 id="I">I</h2>
 </li>
       <li><a href="api/sleap.nn.data.general.html#sleap.nn.data.general.LambdaMap.input_key_names">input_key_names (sleap.nn.data.general.LambdaMap attribute)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.input_keys">input_keys (sleap.nn.data.augmentation.ImgaugAugmenter property)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys">input_keys (sleap.nn.data.augmentation.AlbumentationsAugmenter property)</a>
 
       <ul>
         <li><a href="api/sleap.nn.data.confidence_maps.html#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.input_keys">(sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator property)</a>
@@ -2327,9 +2335,11 @@ <h2 id="I">I</h2>
       </ul></li>
   </ul></td>
   <td style="width: 33%; vertical-align: top;"><ul>
-      <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame.insert">insert() (sleap.instance.LabeledFrame method)</a>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.insert">insert() (sleap.instance.InstancesList method)</a>
 
       <ul>
+        <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame.insert">(sleap.instance.LabeledFrame method)</a>
+</li>
         <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.insert">(sleap.io.dataset.Labels method)</a>
 </li>
       </ul></li>
@@ -2363,7 +2373,7 @@ <h2 id="I">I</h2>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.instances">instances() (sleap.io.dataset.Labels method)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.instances_key">instances_key (sleap.nn.data.augmentation.ImgaugAugmenter attribute)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key">instances_key (sleap.nn.data.augmentation.AlbumentationsAugmenter attribute)</a>
 
       <ul>
         <li><a href="api/sleap.nn.data.instance_centroids.html#sleap.nn.data.instance_centroids.InstanceCentroidFinder.instances_key">(sleap.nn.data.instance_centroids.InstanceCentroidFinder attribute)</a>
@@ -2372,6 +2382,8 @@ <h2 id="I">I</h2>
 </li>
       </ul></li>
       <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame.instances_to_show">instances_to_show (sleap.instance.LabeledFrame property)</a>
+</li>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList">InstancesList (class in sleap.instance)</a>
 </li>
       <li><a href="api/sleap.nn.inference.html#sleap.nn.inference.BottomUpInferenceLayer.integral_patch_size">integral_patch_size (sleap.nn.inference.BottomUpInferenceLayer attribute)</a>
 
@@ -2501,6 +2513,8 @@ <h2 id="K">K</h2>
         <li><a href="api/sleap.nn.data.resizing.html#sleap.nn.data.resizing.SizeMatcher.keep_full_image">(sleap.nn.data.resizing.SizeMatcher attribute)</a>
 </li>
       </ul></li>
+      <li><a href="api/sleap.nn.config.outputs.html#sleap.nn.config.outputs.OutputsConfig.keep_viz_images">keep_viz_images (sleap.nn.config.outputs.OutputsConfig attribute)</a>
+</li>
       <li><a href="api/sleap.nn.inference.html#sleap.nn.inference.BottomUpInferenceLayer.keras_model">keras_model (sleap.nn.inference.BottomUpInferenceLayer attribute)</a>
 
       <ul>
@@ -2551,6 +2565,8 @@ <h2 id="K">K</h2>
 <h2 id="L">L</h2>
 <table style="width: 100%" class="indextable genindextable"><tr>
   <td style="width: 33%; vertical-align: top;"><ul>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.labeled_frame">labeled_frame (sleap.instance.InstancesList property)</a>
+</li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.labeled_frames">labeled_frames (sleap.io.dataset.Labels attribute)</a>
 </li>
       <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame">LabeledFrame (class in sleap.instance)</a>
@@ -2612,8 +2628,6 @@ <h2 id="L">L</h2>
       <li><a href="api/sleap.info.metrics.html#sleap.info.metrics.list_points_array">list_points_array() (in module sleap.info.metrics)</a>
 </li>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Skeleton.load_all_hdf5">load_all_hdf5() (sleap.skeleton.Skeleton class method)</a>
-</li>
-      <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideo.load_by_chunk">load_by_chunk() (sleap.io.asyncvideo.AsyncVideo method)</a>
 </li>
       <li><a href="api/sleap.nn.config.training_job.html#sleap.nn.config.training_job.load_config">load_config() (in module sleap.nn.config.training_job)</a>
 </li>
@@ -2984,11 +2998,11 @@ <h2 id="M">M</h2>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.merge_container_dicts">merge_container_dicts() (sleap.io.dataset.Labels static method)</a>
 </li>
       <li><a href="api/sleap.instance.html#sleap.instance.LabeledFrame.merge_frames">merge_frames() (sleap.instance.LabeledFrame static method)</a>
-</li>
-      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.merge_matching_frames">merge_matching_frames() (sleap.io.dataset.Labels method)</a>
 </li>
   </ul></td>
   <td style="width: 33%; vertical-align: top;"><ul>
+      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.merge_matching_frames">merge_matching_frames() (sleap.io.dataset.Labels method)</a>
+</li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.merge_nodes">merge_nodes() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.info.feature_suggestions.html#sleap.info.feature_suggestions.ItemStack.meta">meta (sleap.info.feature_suggestions.ItemStack attribute)</a>
@@ -3076,8 +3090,6 @@ <h2 id="M">M</h2>
         <li><a href="api/sleap.info.write_tracking_h5.html#module-sleap.info.write_tracking_h5">sleap.info.write_tracking_h5</a>
 </li>
         <li><a href="api/sleap.instance.html#module-sleap.instance">sleap.instance</a>
-</li>
-        <li><a href="api/sleap.io.asyncvideo.html#module-sleap.io.asyncvideo">sleap.io.asyncvideo</a>
 </li>
         <li><a href="api/sleap.io.convert.html#module-sleap.io.convert">sleap.io.convert</a>
 </li>
@@ -3345,6 +3357,8 @@ <h2 id="N">N</h2>
       <li><a href="api/sleap.instance.html#sleap.instance.Instance.nodes_points">nodes_points (sleap.instance.Instance property)</a>
 </li>
       <li><a href="api/sleap.nn.data.instance_cropping.html#sleap.nn.data.instance_cropping.normalize_bboxes">normalize_bboxes() (in module sleap.nn.data.instance_cropping)</a>
+</li>
+      <li><a href="api/sleap.nn.tracker.components.html#sleap.nn.tracker.components.normalized_instance_similarity">normalized_instance_similarity() (in module sleap.nn.tracker.components)</a>
 </li>
       <li><a href="api/sleap.nn.data.normalization.html#sleap.nn.data.normalization.Normalizer">Normalizer (class in sleap.nn.data.normalization)</a>
 </li>
@@ -3512,7 +3526,7 @@ <h2 id="O">O</h2>
 </li>
       <li><a href="api/sleap.nn.data.general.html#sleap.nn.data.general.LambdaMap.output_key_names">output_key_names (sleap.nn.data.general.LambdaMap attribute)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.output_keys">output_keys (sleap.nn.data.augmentation.ImgaugAugmenter property)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys">output_keys (sleap.nn.data.augmentation.AlbumentationsAugmenter property)</a>
 
       <ul>
         <li><a href="api/sleap.nn.data.confidence_maps.html#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_keys">(sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator property)</a>
@@ -3875,6 +3889,8 @@ <h2 id="P">P</h2>
 </li>
       </ul></li>
       <li><a href="api/sleap.nn.architectures.unet.html#sleap.nn.architectures.unet.PoolingBlock">PoolingBlock (class in sleap.nn.architectures.unet)</a>
+</li>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.pop">pop() (sleap.instance.InstancesList method)</a>
 </li>
       <li><a href="api/sleap.nn.inference.html#sleap.nn.inference.InferenceModel.predict">predict() (sleap.nn.inference.InferenceModel method)</a>
 
@@ -3922,7 +3938,7 @@ <h2 id="P">P</h2>
 </li>
       <li><a href="api/sleap.nn.config.data.html#sleap.nn.config.data.PreprocessingConfig">PreprocessingConfig (class in sleap.nn.config.data)</a>
 </li>
-      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.preset">preset (sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.preset">preset (sleap.io.videowriter.VideoWriterImageio attribute)</a>
 </li>
       <li><a href="api/sleap.nn.architectures.pretrained_encoders.html#sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.pretrained">pretrained (sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder attribute)</a>
 
@@ -4068,8 +4084,12 @@ <h2 id="R">R</h2>
 </li>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Skeleton.relabel_nodes">relabel_nodes() (sleap.skeleton.Skeleton method)</a>
 </li>
-      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove">remove() (sleap.io.dataset.Labels method)</a>
+      <li><a href="api/sleap.instance.html#sleap.instance.InstancesList.remove">remove() (sleap.instance.InstancesList method)</a>
+
+      <ul>
+        <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove">(sleap.io.dataset.Labels method)</a>
 </li>
+      </ul></li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove_all_tracks">remove_all_tracks() (sleap.io.dataset.Labels method)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove_empty_frames">remove_empty_frames() (sleap.io.dataset.Labels method)</a>
@@ -4096,10 +4116,10 @@ <h2 id="R">R</h2>
         <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.LabelsDataCache.remove_instance">(sleap.io.dataset.LabelsDataCache method)</a>
 </li>
       </ul></li>
-      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove_predictions">remove_predictions() (sleap.io.dataset.Labels method)</a>
-</li>
   </ul></td>
   <td style="width: 33%; vertical-align: top;"><ul>
+      <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove_predictions">remove_predictions() (sleap.io.dataset.Labels method)</a>
+</li>
       <li><a href="api/sleap.nn.tracker.kalman.html#sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix">remove_second_bests_from_cost_matrix() (in module sleap.nn.tracker.kalman)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.remove_suggestion">remove_suggestion() (sleap.io.dataset.Labels method)</a>
@@ -4207,8 +4227,6 @@ <h2 id="R">R</h2>
       <li><a href="api/sleap.info.feature_suggestions.html#sleap.info.feature_suggestions.ParallelFeaturePipeline.run">run() (sleap.info.feature_suggestions.ParallelFeaturePipeline class method)</a>
 
       <ul>
-        <li><a href="api/sleap.io.asyncvideo.html#sleap.io.asyncvideo.AsyncVideoServer.run">(sleap.io.asyncvideo.AsyncVideoServer method)</a>
-</li>
         <li><a href="api/sleap.io.visuals.html#sleap.io.visuals.VideoMarkerThread.run">(sleap.io.visuals.VideoMarkerThread method)</a>
 </li>
         <li><a href="api/sleap.nn.data.pipelines.html#sleap.nn.data.pipelines.Pipeline.run">(sleap.nn.data.pipelines.Pipeline method)</a>
@@ -4441,6 +4459,10 @@ <h2 id="S">S</h2>
       <li><a href="api/sleap.skeleton.html#sleap.skeleton.Skeleton">Skeleton (class in sleap.skeleton)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.skeleton">skeleton (sleap.io.dataset.Labels property)</a>
+</li>
+      <li><a href="api/sleap.skeleton.html#sleap.skeleton.SkeletonDecoder">SkeletonDecoder (class in sleap.skeleton)</a>
+</li>
+      <li><a href="api/sleap.skeleton.html#sleap.skeleton.SkeletonEncoder">SkeletonEncoder (class in sleap.skeleton)</a>
 </li>
       <li><a href="api/sleap.io.dataset.html#sleap.io.dataset.Labels.skeletons">skeletons (sleap.io.dataset.Labels attribute)</a>
 
@@ -4526,13 +4548,6 @@ <h2 id="S">S</h2>
 
       <ul>
         <li><a href="api/sleap.instance.html#module-sleap.instance">module</a>
-</li>
-      </ul></li>
-      <li>
-    sleap.io.asyncvideo
-
-      <ul>
-        <li><a href="api/sleap.io.asyncvideo.html#module-sleap.io.asyncvideo">module</a>
 </li>
       </ul></li>
       <li>
@@ -5314,7 +5329,7 @@ <h2 id="T">T</h2>
 </li>
       <li><a href="api/sleap.nn.config.training_job.html#sleap.nn.config.training_job.TrainingJobConfig">TrainingJobConfig (class in sleap.nn.config.training_job)</a>
 </li>
-      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.ImgaugAugmenter.transform_dataset">transform_dataset() (sleap.nn.data.augmentation.ImgaugAugmenter method)</a>
+      <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset">transform_dataset() (sleap.nn.data.augmentation.AlbumentationsAugmenter method)</a>
 
       <ul>
         <li><a href="api/sleap.nn.data.augmentation.html#sleap.nn.data.augmentation.RandomCropper.transform_dataset">(sleap.nn.data.augmentation.RandomCropper method)</a>
@@ -5594,9 +5609,9 @@ <h2 id="V">V</h2>
       </ul></li>
       <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriter">VideoWriter (class in sleap.io.videowriter)</a>
 </li>
-      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterOpenCV">VideoWriterOpenCV (class in sleap.io.videowriter)</a>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio">VideoWriterImageio (class in sleap.io.videowriter)</a>
 </li>
-      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo">VideoWriterSkvideo (class in sleap.io.videowriter)</a>
+      <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterOpenCV">VideoWriterOpenCV (class in sleap.io.videowriter)</a>
 </li>
       <li><a href="api/sleap.nn.training.html#sleap.nn.training.Trainer.visualization_callbacks">visualization_callbacks (sleap.nn.training.Trainer attribute)</a>
 </li>
@@ -5627,7 +5642,7 @@ <h2 id="W">W</h2>
 </li>
         <li><a href="api/sleap.io.video.html#sleap.io.video.SingleImageVideo.width">(sleap.io.video.SingleImageVideo property)</a>
 </li>
-        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterSkvideo.width">(sleap.io.videowriter.VideoWriterSkvideo attribute)</a>
+        <li><a href="api/sleap.io.videowriter.html#sleap.io.videowriter.VideoWriterImageio.width">(sleap.io.videowriter.VideoWriterImageio attribute)</a>
 </li>
       </ul></li>
       <li><a href="api/sleap.nn.data.confidence_maps.html#sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.with_offsets">with_offsets (sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator attribute)</a>
@@ -5734,7 +5749,7 @@ <h2 id="Z">Z</h2>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/bonsai.html b/guides/bonsai.html
new file mode 100644
index 000000000..9421f06c5
--- /dev/null
+++ b/guides/bonsai.html
@@ -0,0 +1,512 @@
+
+
+<!DOCTYPE html>
+
+
+<html lang="en" >
+
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
+
+    <title>Using Bonsai with SLEAP &#8212; SLEAP (v1.4.1)</title>
+  
+  
+  
+  <script data-cfasync="false">
+    document.documentElement.dataset.mode = localStorage.getItem("mode") || "";
+    document.documentElement.dataset.theme = localStorage.getItem("theme") || "light";
+  </script>
+  
+  <!-- Loaded before other Sphinx assets -->
+  <link href="../_static/styles/theme.css?digest=e353d410970836974a52" rel="stylesheet" />
+<link href="../_static/styles/bootstrap.css?digest=e353d410970836974a52" rel="stylesheet" />
+<link href="../_static/styles/pydata-sphinx-theme.css?digest=e353d410970836974a52" rel="stylesheet" />
+
+  
+  <link href="../_static/vendor/fontawesome/6.1.2/css/all.min.css?digest=e353d410970836974a52" rel="stylesheet" />
+  <link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.1.2/webfonts/fa-solid-900.woff2" />
+<link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.1.2/webfonts/fa-brands-400.woff2" />
+<link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.1.2/webfonts/fa-regular-400.woff2" />
+
+    <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
+    <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
+    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
+  
+  <!-- Pre-loaded scripts that we'll load fully later -->
+  <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
+<link rel="preload" as="script" href="../_static/scripts/pydata-sphinx-theme.js?digest=e353d410970836974a52" />
+
+    <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
+    <script src="../_static/jquery.js"></script>
+    <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
+    <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
+    <script src="../_static/clipboard.min.js"></script>
+    <script src="../_static/copybutton.js"></script>
+    <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
+    <script>DOCUMENTATION_OPTIONS.pagename = 'guides/bonsai';</script>
+    <link rel="canonical" href="/develop/guides/bonsai.html" />
+    <link rel="shortcut icon" href="../_static/favicon.ico"/>
+    <link rel="index" title="Index" href="../genindex.html" />
+    <link rel="search" title="Search" href="../search.html" />
+    <link rel="next" title="Notebooks" href="../notebooks/index.html" />
+    <link rel="prev" title="Running SLEAP remotely" href="remote.html" />
+  <meta name="viewport" content="width=device-width, initial-scale=1"/>
+  <meta name="docsearch:language" content="en"/>
+  </head>
+  
+  
+  <body data-bs-spy="scroll" data-bs-target=".bd-toc-nav" data-offset="180" data-bs-root-margin="0px 0px -60%" data-default-mode="">
+
+  
+  
+  <a class="skip-link" href="#main-content">Skip to main content</a>
+  
+  <input type="checkbox"
+          class="sidebar-toggle"
+          name="__primary"
+          id="__primary"/>
+  <label class="overlay overlay-primary" for="__primary"></label>
+  
+  <input type="checkbox"
+          class="sidebar-toggle"
+          name="__secondary"
+          id="__secondary"/>
+  <label class="overlay overlay-secondary" for="__secondary"></label>
+  
+  <div class="search-button__wrapper">
+    <div class="search-button__overlay"></div>
+    <div class="search-button__search-container">
+<form class="bd-search d-flex align-items-center"
+      action="../search.html"
+      method="get">
+  <i class="fa-solid fa-magnifying-glass"></i>
+  <input type="search"
+         class="form-control"
+         name="q"
+         id="search-input"
+         placeholder="Search..."
+         aria-label="Search..."
+         autocomplete="off"
+         autocorrect="off"
+         autocapitalize="off"
+         spellcheck="false"/>
+  <span class="search-button__kbd-shortcut"><kbd class="kbd-shortcut__modifier">Ctrl</kbd>+<kbd>K</kbd></span>
+</form></div>
+  </div>
+  
+    <nav class="bd-header navbar navbar-expand-lg bd-navbar">
+    </nav>
+  
+  <div class="bd-container">
+    <div class="bd-container__inner bd-page-width">
+      
+      <div class="bd-sidebar-primary bd-sidebar">
+        
+
+  
+  <div class="sidebar-header-items sidebar-primary__section">
+    
+    
+    
+    
+  </div>
+  
+    <div class="sidebar-primary-items__start sidebar-primary__section">
+        <div class="sidebar-primary-item">
+  
+
+<a class="navbar-brand logo" href="../index.html">
+  
+  
+  
+  
+    
+    
+      
+    
+    
+    <img src="../_static/logo.png" class="logo__image only-light" alt="Logo image"/>
+    <script>document.write(`<img src="../_static/logo.png" class="logo__image only-dark" alt="Logo image"/>`);</script>
+  
+  
+</a></div>
+        <div class="sidebar-primary-item"><nav class="bd-links" id="bd-docs-nav" aria-label="Main">
+    <div class="bd-toc-item navbar-nav active">
+        <ul class="current nav bd-sidenav">
+<li class="toctree-l1"><a class="reference internal" href="../index.html">Home</a></li>
+<li class="toctree-l1"><a class="reference internal" href="../overview.html">Overview</a></li>
+<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
+<li class="toctree-l1 has-children"><a class="reference internal" href="../tutorials/tutorial.html">Tutorial</a><input class="toctree-checkbox" id="toctree-checkbox-1" name="toctree-checkbox-1" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-1"><i class="fa-solid fa-chevron-down"></i></label><ul>
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/new-project.html">Creating a project</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/initial-labeling.html">Initial Labeling</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/initial-training.html">Training and Inference</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/assisted-labeling.html">Prediction-assisted labeling</a></li>
+
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/proofreading.html">Tracking instances across frames</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../tutorials/analysis.html">Export Data For Analysis</a></li>
+</ul>
+</li>
+<li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Guides</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-2" name="toctree-checkbox-2" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-2"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
+<li class="toctree-l2"><a class="reference internal" href="gui.html">GUI</a></li>
+<li class="toctree-l2"><a class="reference internal" href="cli.html">Command line interfaces</a></li>
+<li class="toctree-l2"><a class="reference internal" href="troubleshooting-workflows.html">Troubleshooting workflows</a></li>
+<li class="toctree-l2"><a class="reference internal" href="skeletons.html">Skeleton design</a></li>
+<li class="toctree-l2"><a class="reference internal" href="choosing-models.html">Configuring models</a></li>
+<li class="toctree-l2"><a class="reference internal" href="merging.html">Importing predictions for labeling</a></li>
+<li class="toctree-l2"><a class="reference internal" href="proofreading.html">Tracking and proofreading</a></li>
+<li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
+<li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
+<li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2 current active"><a class="current reference internal" href="#">Using Bonsai with SLEAP</a></li>
+</ul>
+</li>
+<li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Training_and_inference_on_an_example_dataset.html">Training and inference on an example dataset</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Training_and_inference_using_Google_Drive.html">Training and inference on your own data using Google Drive</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Model_evaluation.html">Model evaluation</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Analysis_examples.html">Analysis examples</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Data_structures.html">Data structures</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Post_inference_tracking.html">Post-inference tracking</a></li>
+
+
+
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Interactive_and_resumable_training.html">Interactive and resumable training</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../notebooks/Interactive_and_realtime_inference.html">Interactive and realtime inference</a></li>
+</ul>
+</li>
+<li class="toctree-l1 has-children"><a class="reference internal" href="../api.html">Developer API</a><input class="toctree-checkbox" id="toctree-checkbox-4" name="toctree-checkbox-4" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-4"><i class="fa-solid fa-chevron-down"></i></label><ul class="simple">
+</ul>
+</li>
+<li class="toctree-l1"><a class="reference internal" href="../datasets.html">Datasets</a></li>
+<li class="toctree-l1"><a class="reference external" href="https://github.com/talmolab/sleap">GitHub</a></li>
+<li class="toctree-l1"><a class="reference external" href="https://github.com/talmolab/sleap/releases">Releases</a></li>
+<li class="toctree-l1"><a class="reference internal" href="../help.html">Help</a></li>
+</ul>
+
+    </div>
+</nav></div>
+    </div>
+  
+  
+  <div class="sidebar-primary-items__end sidebar-primary__section">
+  </div>
+  
+  <div id="rtd-footer-container"></div>
+
+
+      </div>
+      
+      <main id="main-content" class="bd-main">
+        
+        
+
+<div class="sbt-scroll-pixel-helper"></div>
+
+          <div class="bd-content">
+            <div class="bd-article-container">
+              
+              <div class="bd-header-article">
+<div class="header-article-items header-article__inner">
+  
+    <div class="header-article-items__start">
+      
+        <div class="header-article-item"><label class="sidebar-toggle primary-toggle btn btn-sm" for="__primary" title="Toggle primary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+  <span class="fa-solid fa-bars"></span>
+</label></div>
+      
+    </div>
+  
+  
+    <div class="header-article-items__end">
+      
+        <div class="header-article-item">
+
+<div class="article-header-buttons">
+
+
+
+
+
+<div class="dropdown dropdown-source-buttons">
+  <button class="btn dropdown-toggle" type="button" data-bs-toggle="dropdown" aria-expanded="false" aria-label="Source repositories">
+    <i class="fab fa-github"></i>
+  </button>
+  <ul class="dropdown-menu">
+      
+      
+      
+      <li><a href="https://github.com/talmolab/sleap" target="_blank"
+   class="btn btn-sm btn-source-repository-button dropdown-item"
+   title="Source repository"
+   data-bs-placement="left" data-bs-toggle="tooltip"
+>
+  
+
+<span class="btn__icon-container">
+  <i class="fab fa-github"></i>
+  </span>
+<span class="btn__text-container">Repository</span>
+</a>
+</li>
+      
+      
+      
+      
+      <li><a href="https://github.com/talmolab/sleap/issues/new?title=Issue%20on%20page%20%2Fguides/bonsai.html&body=Your%20issue%20content%20here." target="_blank"
+   class="btn btn-sm btn-source-issues-button dropdown-item"
+   title="Open an issue"
+   data-bs-placement="left" data-bs-toggle="tooltip"
+>
+  
+
+<span class="btn__icon-container">
+  <i class="fas fa-lightbulb"></i>
+  </span>
+<span class="btn__text-container">Open issue</span>
+</a>
+</li>
+      
+  </ul>
+</div>
+
+
+<script>
+document.write(`
+  <button class="theme-switch-button btn btn-sm btn-outline-primary navbar-btn rounded-circle" title="light/dark" aria-label="light/dark" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="theme-switch" data-mode="light"><i class="fa-solid fa-sun"></i></span>
+    <span class="theme-switch" data-mode="dark"><i class="fa-solid fa-moon"></i></span>
+    <span class="theme-switch" data-mode="auto"><i class="fa-solid fa-circle-half-stroke"></i></span>
+  </button>
+`);
+</script>
+
+<script>
+document.write(`
+  <button class="btn btn-sm navbar-btn search-button search-button__button" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <i class="fa-solid fa-magnifying-glass"></i>
+  </button>
+`);
+</script>
+<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
+    <span class="fa-solid fa-list"></span>
+</label>
+</div></div>
+      
+    </div>
+  
+</div>
+</div>
+              
+              
+
+<div id="jb-print-docs-body" class="onlyprint">
+    <h1>Using Bonsai with SLEAP</h1>
+    <!-- Table of contents -->
+    <div id="print-main-content">
+        <div id="jb-print-toc">
+            
+            <div>
+                <h2> Contents </h2>
+            </div>
+            <nav aria-label="Page">
+                <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#exporting-a-sleap-trained-model">Exporting a SLEAP trained model</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installing-bonsai-and-necessary-packages">Installing Bonsai and necessary packages</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#using-bonsai-sleap-modules">Using Bonsai SLEAP modules</a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#top-down-model">Top-down model</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#top-down-id-model">Top-Down-ID model</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#single-instance-model">Single instance model</a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#connecting-the-modules">Connecting the modules</a></li>
+</ul>
+            </nav>
+        </div>
+    </div>
+</div>
+
+              
+                
+<div id="searchbox"></div>
+                <article class="bd-article" role="main">
+                  
+  <section class="tex2jax_ignore mathjax_ignore" id="using-bonsai-with-sleap">
+<span id="bonsai"></span><h1>Using Bonsai with SLEAP<a class="headerlink" href="#using-bonsai-with-sleap" title="Permalink to this heading">#</a></h1>
+<p>Bonsai is a visual language for reactive programming and currently supports SLEAP models.</p>
+<div class="admonition note">
+<p class="admonition-title">Note</p>
+<p>Currently Bonsai supports only single instance, top-down and top-down-id SLEAP models.</p>
+</div>
+<section id="exporting-a-sleap-trained-model">
+<h2>Exporting a SLEAP trained model<a class="headerlink" href="#exporting-a-sleap-trained-model" title="Permalink to this heading">#</a></h2>
+<p>Before we can import a trained model into Bonsai, we need to use the <code class="code docutils literal notranslate"><span class="pre">sleap-export</span></code> command to convert the model to a format supported by Bonsai. For example, to export a top-down-id model, the command is as follows:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>sleap-export<span class="w"> </span>-m<span class="w"> </span>centroid/model/folder/path<span class="w"> </span>-m<span class="w"> </span>top_down_id/model/folder/path<span class="w"> </span>-e<span class="w"> </span>exported/model/path
+</pre></div>
+</div>
+<p>Please refer to the <a class="reference internal" href="cli.html#sleap-export"><span class="std std-ref">sleap-export</span></a> docs for more details on using the command.</p>
+<p>This will generate the necessary <code class="docutils literal notranslate"><span class="pre">.pb</span></code> file and other information files required by Bonsai. In this example, these files were saved to the specified <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder.</p>
+<p>The <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder will have a structure like the following:</p>
+<div class="highlight-plaintext notranslate"><div class="highlight"><pre><span></span>exported/model/path
+├── centroid_config.json
+├── confmap_config.json
+├── frozen_graph.pb
+└── info.json
+</pre></div>
+</div>
+</section>
+<section id="installing-bonsai-and-necessary-packages">
+<h2>Installing Bonsai and necessary packages<a class="headerlink" href="#installing-bonsai-and-necessary-packages" title="Permalink to this heading">#</a></h2>
+<ol class="arabic simple">
+<li><p>Install Bonsai. See the <a class="reference external" href="https://bonsai-rx.org/docs/articles/installation.html">Bonsai installation instructions</a>.</p></li>
+<li><p>Download and add the necessary packages for Bonsai to run with SLEAP. See the official <a class="reference external" href="https://github.com/bonsai-rx/sleap?tab=readme-ov-file#bonsai---sleap">Bonsai SLEAP documentation</a> for more information.</p></li>
+</ol>
+</section>
+<section id="using-bonsai-sleap-modules">
+<h2>Using Bonsai SLEAP modules<a class="headerlink" href="#using-bonsai-sleap-modules" title="Permalink to this heading">#</a></h2>
+<p>Once you have Bonsai installed with the required packages, you should be able to open the Bonsai application. The workflow must have a source module <code class="docutils literal notranslate"><span class="pre">FileCapture</span></code> which can be found in the toolbox search in the workflow editor. Provide the path to the video that was used to train the SLEAP model in the <code class="docutils literal notranslate"><span class="pre">FileName</span></code> field of the module.</p>
+<p><img alt="Bonsai FileCapture module" src="../_images/bonsai-filecapture.jpg" /></p>
+<section id="top-down-model">
+<h3>Top-down model<a class="headerlink" href="#top-down-model" title="Permalink to this heading">#</a></h3>
+<p>The top-down model requires both the <code class="docutils literal notranslate"><span class="pre">PredictCentroids</span></code> and the <code class="docutils literal notranslate"><span class="pre">PredictPoses</span></code> modules.</p>
+<p>The <code class="docutils literal notranslate"><span class="pre">PredictCentroids</span></code> module will predict the centroids of detections. There are two fields inside the <code class="docutils literal notranslate"><span class="pre">PredictCentroids</span></code> module: the <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field and the <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field. The <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field expects the path to the training config JSON file for the centroid model. The <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field expects the path to the <code class="docutils literal notranslate"><span class="pre">frozen_graph.pb</span></code> file in the <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder.</p>
+<p><img alt="Bonsai PredictCentroids module" src="../_images/bonsai-predictcentroids.jpg" /></p>
+<p>The <code class="docutils literal notranslate"><span class="pre">PredictPoses</span></code> module will predict the instances of detections. Similar to the <code class="docutils literal notranslate"><span class="pre">PredictCentroid</span></code> module, there are two fields inside the <code class="docutils literal notranslate"><span class="pre">PredictPoses</span></code> module: the <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field and the <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field. The <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field expects the path to the training config JSON file for the centered instance model. The <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field expects the path to the <code class="docutils literal notranslate"><span class="pre">frozen_graph.pb</span></code> file in the <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder.</p>
+<p><img alt="Bonsai PredictPoses module" src="../_images/bonsai-predictposes.jpg" /></p>
+</section>
+<section id="top-down-id-model">
+<h3>Top-Down-ID model<a class="headerlink" href="#top-down-id-model" title="Permalink to this heading">#</a></h3>
+<p>The <code class="docutils literal notranslate"><span class="pre">PredictPoseIdentities</span></code> module will predict the instances with identities. This module has two fields: the <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field and the <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field. The <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field expects the path to the training config JSON file for the top-down-id model. The <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field expects the path to the <code class="docutils literal notranslate"><span class="pre">frozen_graph.pb</span></code> file in the <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder.</p>
+<p><img alt="Bonsai PredictPoseIdentities module" src="../_images/bonsai-predictposeidentities.jpg" /></p>
+</section>
+<section id="single-instance-model">
+<h3>Single instance model<a class="headerlink" href="#single-instance-model" title="Permalink to this heading">#</a></h3>
+<p>The <code class="docutils literal notranslate"><span class="pre">PredictSinglePose</span></code> module will predict the poses for single instance models. This module also has two fields: the <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field and the <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field. The <code class="docutils literal notranslate"><span class="pre">TrainingConfig</span></code> field expects the path to the training config JSON file for the single instance model. The <code class="docutils literal notranslate"><span class="pre">ModelFileName</span></code> field expects the path to the <code class="docutils literal notranslate"><span class="pre">frozen_graph.pb</span></code> file in the <code class="docutils literal notranslate"><span class="pre">exported/model/path</span></code> folder.</p>
+</section>
+</section>
+<section id="connecting-the-modules">
+<h2>Connecting the modules<a class="headerlink" href="#connecting-the-modules" title="Permalink to this heading">#</a></h2>
+<p>Right-click on the <code class="docutils literal notranslate"><span class="pre">FileCapture</span></code> module and select <strong>Create Connection</strong>. Now click on the required SLEAP module to complete the connection.</p>
+<p><img alt="Bonsai module connection " src="../_images/bonsai-connection.jpg" /></p>
+<p>Once it is done, the workflow in Bonsai will look something like the following:</p>
+<p><img alt="Bonsai.SLEAP workflow" src="../_images/bonsai-workflow.jpg" /></p>
+<p>Now you can click the green start button to run the workflow and you can add more modules to analyze and visualize the results in Bonsai.</p>
+<p>For more documentation on various modules and workflows, please refer to the <a class="reference external" href="https://bonsai-rx.org/docs/articles/editor.html">official Bonsai docs</a>.</p>
+</section>
+</section>
+
+
+                </article>
+              
+
+              
+              
+                <footer class="bd-footer-article">
+                  
+<div class="footer-article-items footer-article__inner">
+  
+    <div class="footer-article-item"><!-- Previous / next buttons -->
+<div class="prev-next-area">
+    <a class="left-prev"
+       href="remote.html"
+       title="previous page">
+      <i class="fa-solid fa-angle-left"></i>
+      <div class="prev-next-info">
+        <p class="prev-next-subtitle">previous</p>
+        <p class="prev-next-title">Running SLEAP remotely</p>
+      </div>
+    </a>
+    <a class="right-next"
+       href="../notebooks/index.html"
+       title="next page">
+      <div class="prev-next-info">
+        <p class="prev-next-subtitle">next</p>
+        <p class="prev-next-title">Notebooks</p>
+      </div>
+      <i class="fa-solid fa-angle-right"></i>
+    </a>
+</div></div>
+  
+</div>
+
+                </footer>
+              
+            </div>
+            
+            
+              
+                <div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
+
+  <div class="sidebar-secondary-item">
+  <div class="page-toc tocsection onthispage">
+    <i class="fa-solid fa-list"></i> Contents
+  </div>
+  <nav class="bd-toc-nav page-toc">
+    <ul class="visible nav section-nav flex-column">
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#exporting-a-sleap-trained-model">Exporting a SLEAP trained model</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installing-bonsai-and-necessary-packages">Installing Bonsai and necessary packages</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#using-bonsai-sleap-modules">Using Bonsai SLEAP modules</a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#top-down-model">Top-down model</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#top-down-id-model">Top-Down-ID model</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#single-instance-model">Single instance model</a></li>
+</ul>
+</li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#connecting-the-modules">Connecting the modules</a></li>
+</ul>
+  </nav></div>
+
+</div></div>
+              
+            
+          </div>
+          <footer class="bd-footer-content">
+            
+<div class="bd-footer-content__inner container">
+  
+  <div class="footer-item">
+    
+<p class="component-author">
+By SLEAP Developers
+</p>
+
+  </div>
+  
+  <div class="footer-item">
+    
+  <p class="copyright">
+    
+      © Copyright 2019–2024, Talmo Lab.
+      <br/>
+    
+  </p>
+
+  </div>
+  
+  <div class="footer-item">
+    
+  </div>
+  
+  <div class="footer-item">
+    
+  </div>
+  
+</div>
+          </footer>
+        
+
+      </main>
+    </div>
+  </div>
+  
+  <!-- Scripts loaded after <body> so the DOM is not blocked -->
+  <script src="../_static/scripts/bootstrap.js?digest=e353d410970836974a52"></script>
+<script src="../_static/scripts/pydata-sphinx-theme.js?digest=e353d410970836974a52"></script>
+
+  <footer class="bd-footer">
+  </footer>
+  </body>
+</html>
\ No newline at end of file
diff --git a/guides/choosing-models.html b/guides/choosing-models.html
index bb9dd98fe..6e232461d 100644
--- a/guides/choosing-models.html
+++ b/guides/choosing-models.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Configuring models &#8212; SLEAP (v1.3.3)</title>
+    <title>Configuring models &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/choosing-models';</script>
     <link rel="canonical" href="/develop/guides/choosing-models.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Importing predictions for labeling" href="merging.html" />
     <link rel="prev" title="Skeleton design" href="skeletons.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -331,18 +329,18 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="configuring-models">
-<span id="choosing-models"></span><h1>Configuring models<a class="headerlink" href="#configuring-models" title="Permalink to this headline">#</a></h1>
+<span id="choosing-models"></span><h1>Configuring models<a class="headerlink" href="#configuring-models" title="Permalink to this heading">#</a></h1>
 <section id="model-type">
-<h2>Model type<a class="headerlink" href="#model-type" title="Permalink to this headline">#</a></h2>
+<h2>Model type<a class="headerlink" href="#model-type" title="Permalink to this heading">#</a></h2>
 <p>Training and inference will run in different modes depending on the output types of the models you supply. SLEAP currently support two distinct modes for multi-animal inference.</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>The “<strong>bottom-up</strong>” approach uses a single model which outputs <strong>confidence maps</strong> and <strong>part affinity fields</strong> for all instances in a given frame. The confidence maps are used to predict node locations and the part affinity fields are used to group nodes into distinct animal instances.</p></li>
 <li><p>The “<strong>top-down</strong>” approach starts by using a <strong>centroid</strong> model to predict the location of each animal in a given frame, and then a <strong>instance centered confidence map</strong> model is used to predict the locations of all the nodes for each animal separately.</p></li>
 </ol>
 <p>Each approach has its advantages, and you may wish to try out both to see which gives you better results.</p>
 </section>
 <section id="hyperparameters">
-<h2>Hyperparameters<a class="headerlink" href="#hyperparameters" title="Permalink to this headline">#</a></h2>
+<h2>Hyperparameters<a class="headerlink" href="#hyperparameters" title="Permalink to this heading">#</a></h2>
 <p>You can configure the hyperparameters for training your model in the “Model Configuration” tabs of the Training Dialog:</p>
 <p><img alt="Model configuration" src="../_images/model-config.jpg" /></p>
 <p>Hover over each field to get an explanation. Here are some hyperparameters which are especially worth your attention:</p>
@@ -430,7 +428,7 @@ <h2>Hyperparameters<a class="headerlink" href="#hyperparameters" title="Permalin
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/cli.html b/guides/cli.html
index 997871def..bb886e10a 100644
--- a/guides/cli.html
+++ b/guides/cli.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Command line interfaces &#8212; SLEAP (v1.3.3)</title>
+    <title>Command line interfaces &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/cli';</script>
     <link rel="canonical" href="/develop/guides/cli.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Troubleshooting workflows" href="troubleshooting-workflows.html" />
     <link rel="prev" title="GUI" href="gui.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -358,12 +356,12 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="command-line-interfaces">
-<span id="cli"></span><h1>Command line interfaces<a class="headerlink" href="#command-line-interfaces" title="Permalink to this headline">#</a></h1>
+<span id="cli"></span><h1>Command line interfaces<a class="headerlink" href="#command-line-interfaces" title="Permalink to this heading">#</a></h1>
 <p>SLEAP provides several types of functionality accessible through a command prompt.</p>
 <section id="gui">
-<h2>GUI<a class="headerlink" href="#gui" title="Permalink to this headline">#</a></h2>
+<h2>GUI<a class="headerlink" href="#gui" title="Permalink to this heading">#</a></h2>
 <section id="sleap-label">
-<span id="id1"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-label</span></code><a class="headerlink" href="#sleap-label" title="Permalink to this headline">#</a></h3>
+<span id="id1"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-label</span></code><a class="headerlink" href="#sleap-label" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-label</span></code> runs the GUI application for labeling and viewing <code class="code docutils literal notranslate"><span class="pre">.slp</span></code> files.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-label [-h] [--nonnative] [--profiling] [--reset] [labels_path]
 
@@ -381,13 +379,13 @@ <h2>GUI<a class="headerlink" href="#gui" title="Permalink to this headline">#</a
 </section>
 </section>
 <section id="training">
-<h2>Training<a class="headerlink" href="#training" title="Permalink to this headline">#</a></h2>
+<h2>Training<a class="headerlink" href="#training" title="Permalink to this heading">#</a></h2>
 <section id="sleap-train">
-<span id="id2"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-train</span></code><a class="headerlink" href="#sleap-train" title="Permalink to this headline">#</a></h3>
+<span id="id2"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-train</span></code><a class="headerlink" href="#sleap-train" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-train</span></code> is the command-line interface for training. Use this for training on a remote machine/cluster/colab notebook instead of through the GUI.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-train [-h] [--video-paths VIDEO_PATHS] [--val_labels VAL_LABELS]
-                   [--test_labels TEST_LABELS] [--tensorboard] [--save_viz]
-                   [--zmq] [--run_name RUN_NAME] [--prefix PREFIX]
+                   [--test_labels TEST_LABELS] [--tensorboard] [--save_viz] 
+                   [--keep_viz] [--zmq] [--run_name RUN_NAME] [--prefix PREFIX]
                    [--suffix SUFFIX]
                    training_job_path [labels_path]
 
@@ -418,6 +416,8 @@ <h2>Training<a class="headerlink" href="#training" title="Permalink to this head
   --save_viz            Enable saving of prediction visualizations to the run
                         folder if not already specified in the training job
                         config.
+  --keep_viz            Keep prediction visualization images in the run
+                        folder after training if --save_viz is enabled.
   --zmq                 Enable ZMQ logging (for GUI) if not already specified
                         in the training job config.
   --run_name RUN_NAME   Run name to use when saving file, overrides other run
@@ -434,7 +434,7 @@ <h2>Training<a class="headerlink" href="#training" title="Permalink to this head
 </div>
 </section>
 <section id="sleap-export">
-<span id="id3"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-export</span></code><a class="headerlink" href="#sleap-export" title="Permalink to this headline">#</a></h3>
+<span id="id3"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-export</span></code><a class="headerlink" href="#sleap-export" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-export</span></code> is a command-line interface for exporting trained models as a TensorFlow graph for use in other applications. See <a class="reference external" href="https://www.tensorflow.org/guide/saved_model">this guide</a> for details on how TensorFlow saves models and the <a class="reference internal" href="../api/sleap.nn.inference.html#sleap.nn.inference.InferenceModel.export_model" title="sleap.nn.inference.InferenceModel.export_model"><span class="xref myst py py-meth"><code class="docutils literal notranslate"><span class="pre">sleap.nn.inference.InferenceModel.export_model</span></code></span></a> documentation.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-export [-h] [-m MODELS] [-e [EXPORT_PATH]]
 
@@ -446,9 +446,9 @@ <h2>Training<a class="headerlink" href="#training" title="Permalink to this head
   -e [EXPORT_PATH], --export_path [EXPORT_PATH]
                         Path to output directory where the frozen model will be exported to.
                         Defaults to a folder named &#39;exported_model&#39;.
-  -u, --unrag UNRAG
-                        Convert ragged tensors into regular tensors with NaN padding.
-                        Defaults to True.
+  -r, --ragged RAGGED
+                        Keep tensors ragged if present. If ommited, convert
+                        ragged tensors into regular tensors with NaN padding.
   -n, --max_instances MAX_INSTANCES
                         Limit maximum number of instances in multi-instance models.
                         Not available for ID models. Defaults to None.
@@ -457,9 +457,9 @@ <h2>Training<a class="headerlink" href="#training" title="Permalink to this head
 </section>
 </section>
 <section id="inference-and-tracking">
-<h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" title="Permalink to this headline">#</a></h2>
+<h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" title="Permalink to this heading">#</a></h2>
 <section id="sleap-track">
-<span id="id4"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-track</span></code><a class="headerlink" href="#sleap-track" title="Permalink to this headline">#</a></h3>
+<span id="id4"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-track</span></code><a class="headerlink" href="#sleap-track" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-track</span></code> is the command-line interface for running inference using models which have already been trained. Use this for running inference on a remote machine such as an HPC cluster or Colab notebook.</p>
 <p>If you specify how many identities there should be in a frame (i.e., the number of animals) with the <code class="code docutils literal notranslate"><span class="pre">--tracking.clean_instance_count</span></code> argument, then we will use a heuristic method to connect “breaks” in the track identities where we lose one identity and spawn another. This can be used as part of the inference pipeline (if models are specified), as part of the tracking-only pipeline (if the predictions file is specified and no models are specified), or by itself on predictions with pre-tracked identities (if you specify <code class="code docutils literal notranslate"><span class="pre">--tracking.tracker</span> <span class="pre">none</span></code>). See <a class="reference internal" href="proofreading.html#proofreading"><span class="std std-ref">Tracking and proofreading</span></a> for more details on tracking.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-track [-h] [-m MODELS] [--frames FRAMES] [--only-labeled-frames] [--only-suggested-frames] [-o OUTPUT] [--no-empty-frames]
@@ -480,7 +480,10 @@ <h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" t
                    [data_path]
 
 positional arguments:
-  data_path             Path to data to predict on. This can be a labels (.slp) file or any supported video format.
+  data_path             Path to data to predict on. This can be one of the following: A .slp file containing labeled data; A folder containing multiple
+                        video files in supported formats; An individual video file in a supported format; A CSV file with a column of video file paths. 
+                        If more than one column is provided in the CSV file, the first will be used for the input data paths and the next column will be
+                        used as the output paths; A text file with a path to a video file on each line
 
 optional arguments:
   -h, --help            show this help message and exit
@@ -495,7 +498,7 @@ <h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" t
                         Only run inference on unlabeled suggested frames when running on labels dataset. This is useful for generating predictions for
                         initialization during labeling.
   -o OUTPUT, --output OUTPUT
-                        The output filename to use for the predicted data. If not provided, defaults to &#39;[data_path].predictions.slp&#39;.
+                        The output filename or directory path to use for the predicted data. If not provided, defaults to &#39;[data_path].predictions.slp&#39;.
   --no-empty-frames     Clear any empty frames that did not have any detected instances before saving to output.
   --verbosity {none,rich,json}
                         Verbosity of inference progress reporting. &#39;none&#39; does not output anything during inference, &#39;rich&#39; displays an updating
@@ -546,7 +549,7 @@ <h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" t
   --tracking.clean_iou_threshold TRACKING.CLEAN_IOU_THRESHOLD
                         IOU to use when culling instances *after* tracking. (default: 0)
   --tracking.similarity TRACKING.SIMILARITY
-                        Options: instance, centroid, iou (default: instance)
+                        Options: instance, normalized_instance, object_keypoint, centroid, iou (default: instance)
   --tracking.match TRACKING.MATCH
                         Options: hungarian, greedy (default: greedy)
   --tracking.robust TRACKING.ROBUST
@@ -573,7 +576,7 @@ <h2>Inference and Tracking<a class="headerlink" href="#inference-and-tracking" t
 </pre></div>
 </div>
 <section id="examples">
-<h4>Examples:<a class="headerlink" href="#examples" title="Permalink to this headline">#</a></h4>
+<h4>Examples:<a class="headerlink" href="#examples" title="Permalink to this heading">#</a></h4>
 <p><strong>1. Simple inference without tracking:</strong></p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>sleap-track -m &quot;models/my_model&quot; -o &quot;output_predictions.slp&quot; &quot;input_video.mp4&quot;
 </pre></div>
@@ -611,9 +614,9 @@ <h4>Examples:<a class="headerlink" href="#examples" title="Permalink to this hea
 </section>
 </section>
 <section id="dataset-files">
-<h2>Dataset files<a class="headerlink" href="#dataset-files" title="Permalink to this headline">#</a></h2>
+<h2>Dataset files<a class="headerlink" href="#dataset-files" title="Permalink to this heading">#</a></h2>
 <section id="sleap-convert">
-<span id="id5"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-convert</span></code><a class="headerlink" href="#sleap-convert" title="Permalink to this headline">#</a></h3>
+<span id="id5"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-convert</span></code><a class="headerlink" href="#sleap-convert" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-convert</span></code> allows you to convert between various dataset file formats. Amongst other things, it can be used to export data from a SLEAP dataset into an HDF5 file that can be easily used for analysis (e.g., read from MATLAB). See <a class="reference internal" href="../api/sleap.io.convert.html#module-sleap.io.convert" title="sleap.io.convert"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sleap.io.convert</span></code></a> for more information.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-convert [-h] [-o OUTPUT] [--format FORMAT] [--video VIDEO]
                      input_path
@@ -648,7 +651,8 @@ <h2>Dataset files<a class="headerlink" href="#dataset-files" title="Permalink to
                               analysis file for the latter video is given a default name.
   --format FORMAT       Output format. Default (&#39;slp&#39;) is SLEAP dataset;
                         &#39;analysis&#39; results in analysis.h5 file; &#39;analysis.nix&#39; results
-                        in an analysis nix file; &#39;h5&#39; or &#39;json&#39; results in SLEAP dataset
+                        in an analysis nix file; &#39;analysis.csv&#39; results
+                        in an analysis csv file; &#39;h5&#39; or &#39;json&#39; results in SLEAP dataset
                         with specified file format.
   --video VIDEO         Path to video (if needed for conversion).
 </pre></div>
@@ -660,7 +664,7 @@ <h2>Dataset files<a class="headerlink" href="#dataset-files" title="Permalink to
 <p>See <span class="xref myst">Analysis examples</span> for how to work with these outputs.</p>
 </section>
 <section id="sleap-inspect">
-<span id="id6"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-inspect</span></code><a class="headerlink" href="#sleap-inspect" title="Permalink to this headline">#</a></h3>
+<span id="id6"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-inspect</span></code><a class="headerlink" href="#sleap-inspect" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-inspect</span></code> gives you various information about a SLEAP dataset file such as a list of videos and a count of the frames with labels. If you’re inspecting a predictions dataset (i.e., the output from running <code class="code docutils literal notranslate"><span class="pre">sleap-track</span></code> or inference in the GUI) it will also include details about how those predictions were created (i.e., the models, the version of SLEAP, and any inference parameters).</p>
 <p>You can also specify a model folder to get a quick summary of the configuration and metrics (if available).</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-inspect [-h] [--verbose] data_path
@@ -676,9 +680,9 @@ <h2>Dataset files<a class="headerlink" href="#dataset-files" title="Permalink to
 </section>
 </section>
 <section id="rendering">
-<h2>Rendering<a class="headerlink" href="#rendering" title="Permalink to this headline">#</a></h2>
+<h2>Rendering<a class="headerlink" href="#rendering" title="Permalink to this heading">#</a></h2>
 <section id="sleap-render">
-<span id="id7"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-render</span></code><a class="headerlink" href="#sleap-render" title="Permalink to this headline">#</a></h3>
+<span id="id7"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-render</span></code><a class="headerlink" href="#sleap-render" title="Permalink to this heading">#</a></h3>
 <p><code class="code docutils literal notranslate"><span class="pre">sleap-render</span></code> allows you to render videos directly from the CLI. It is used to render video clips with Instances.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-render [-h] [-o OUTPUT] [-f FPS] [--scale SCALE] [--crop CROP] [--frames FRAMES] [--video-index VIDEO_INDEX] data_path
 positional arguments:
@@ -706,14 +710,17 @@ <h2>Rendering<a class="headerlink" href="#rendering" title="Permalink to this he
   --distinctly_color DISTINCTLY_COLOR
                         Specify how to color instances. Options include: &quot;instances&quot;,
                         &quot;edges&quot;, and &quot;nodes&quot; (default: &quot;instances&quot;)
+  --background BACKGROUND
+                        Specify the type of background to be used to save the videos.
+                        Options: original, black, white and grey. (default: &quot;original&quot;)
 </pre></div>
 </div>
 </section>
 </section>
 <section id="debugging">
-<h2>Debugging<a class="headerlink" href="#debugging" title="Permalink to this headline">#</a></h2>
+<h2>Debugging<a class="headerlink" href="#debugging" title="Permalink to this heading">#</a></h2>
 <section id="sleap-diagnostic">
-<span id="id8"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-diagnostic</span></code><a class="headerlink" href="#sleap-diagnostic" title="Permalink to this headline">#</a></h3>
+<span id="id8"></span><h3><code class="docutils literal notranslate"><span class="pre">sleap-diagnostic</span></code><a class="headerlink" href="#sleap-diagnostic" title="Permalink to this heading">#</a></h3>
 <p>There’s also a script to output diagnostic information which may help us if you need to contact us about problems installing or running SLEAP. If you were able to install the SLEAP Python package, you can run this script with <code class="code docutils literal notranslate"><span class="pre">sleap-diagnostic</span></code>. Otherwise, you can download <a class="reference external" href="https://raw.githubusercontent.com/talmolab/sleap/main/sleap/diagnostic.py">diagnostic.py</a> and run <code class="code docutils literal notranslate"><span class="pre">python</span> <span class="pre">diagnostic.py</span></code>.</p>
 <div class="highlight-none notranslate"><div class="highlight"><pre><span></span>usage: sleap-diagnostic [-h] [-o OUTPUT] [--gui-check]
 
@@ -832,7 +839,7 @@ <h2>Debugging<a class="headerlink" href="#debugging" title="Permalink to this he
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/colab.html b/guides/colab.html
index 5199b9ee2..362ce0388 100644
--- a/guides/colab.html
+++ b/guides/colab.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Run training and inference on Colab &#8212; SLEAP (v1.3.3)</title>
+    <title>Run training and inference on Colab &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/colab';</script>
     <link rel="canonical" href="/develop/guides/colab.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Creating a custom training profile" href="custom-training.html" />
     <link rel="prev" title="Tracking and proofreading" href="proofreading.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2 current active"><a class="current reference internal" href="#">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -331,18 +329,18 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="run-training-and-inference-on-colab">
-<span id="colab"></span><h1>Run training and inference on Colab<a class="headerlink" href="#run-training-and-inference-on-colab" title="Permalink to this headline">#</a></h1>
+<span id="colab"></span><h1>Run training and inference on Colab<a class="headerlink" href="#run-training-and-inference-on-colab" title="Permalink to this heading">#</a></h1>
 <p><em>Case: You already have a project with labeled training data and you’d like to run training or inference in a Colab notebook.</em></p>
 <p>Take a look at our <a class="reference external" href="https://sleap.ai/notebooks">example notebooks</a> which you can run on Colab! These will walk you through the entire process of running training and inference on Colab.</p>
 <section id="about-colab">
-<h2>About Colab<a class="headerlink" href="#about-colab" title="Permalink to this headline">#</a></h2>
+<h2>About Colab<a class="headerlink" href="#about-colab" title="Permalink to this heading">#</a></h2>
 <p><a class="reference external" href="https://colab.research.google.com/">Google Colaboratory</a> is a hosted service from Google that lets you run Python code (like SLEAP) on Google’s servers with access to GPUs for training and inference. It’s free!</p>
 <p>When you start running code in a notebook, Google connects your notebook to a virtual server instance. Think of this as a new computer with Python, TensorFlow and a few other things pre-installed. You can run things on this server instance for at most 12 hours (it will timeout sooner if idle but this depends on how much demand there is for Google’s servers).</p>
 <p>Once the server instance times out, you lose all the data that’s stored on the server instance. This means that you’ll need to install SLEAP from the notebook at the beginning of each new session, you’ll need to move your data to the server instance, and you’ll need to copy any results off the server instance when you’re done.</p>
 <p>You can also pay for <a class="reference external" href="https://colab.research.google.com/signup">Colab Pro</a> (currently available only in the US) which gives you priority access to better hardware and longer runtimes (they say 24 hours and more lenient idle timeouts).</p>
 </section>
 <section id="moving-data-to-from-colab">
-<h2>Moving Data to/from Colab<a class="headerlink" href="#moving-data-to-from-colab" title="Permalink to this headline">#</a></h2>
+<h2>Moving Data to/from Colab<a class="headerlink" href="#moving-data-to-from-colab" title="Permalink to this heading">#</a></h2>
 <p>The easiest way to work with your own data on Colab is to use <a class="reference external" href="https://www.google.com/drive/">Google Drive</a>. This is another service provided by Google. It’s also free! If your university has <a class="reference external" href="https://edu.google.com/products/gsuite-for-education/">G Suite for Education</a> you’ll have unlimited storage space.</p>
 <p>Google Drive is similar to Dropbox, Microsoft OneDrive, or Apple iCloud. You can install software and you’ll you get a special folder which is automatically synced: anything you copy into it will be available from other computers (using your account), and anything copied into it from elsewhere will automatically be downloaded into the folder on your computer.</p>
 <p>For training on your own data, it’s probably easiest to export a training job package from the <code class="xref py py-obj docutils literal notranslate"><span class="pre">sleap-label</span></code> GUI. The training job package will contain model files as well as all of your labeled data and all of the images for labeled frames within a single file, so that you don’t have to copy the full training videos and you don’t have to worry about paths to the videos.</p>
@@ -423,7 +421,7 @@ <h2>Moving Data to/from Colab<a class="headerlink" href="#moving-data-to-from-co
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/custom-training.html b/guides/custom-training.html
index 6a172c4ac..68a89d8be 100644
--- a/guides/custom-training.html
+++ b/guides/custom-training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Creating a custom training profile &#8212; SLEAP (v1.3.3)</title>
+    <title>Creating a custom training profile &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/custom-training';</script>
     <link rel="canonical" href="/develop/guides/custom-training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Running SLEAP remotely" href="remote.html" />
     <link rel="prev" title="Run training and inference on Colab" href="colab.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2 current active"><a class="current reference internal" href="#">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Creating a custom training profile</h1>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="creating-a-custom-training-profile">
-<span id="custom-training"></span><h1>Creating a custom training profile<a class="headerlink" href="#creating-a-custom-training-profile" title="Permalink to this headline">#</a></h1>
+<span id="custom-training"></span><h1>Creating a custom training profile<a class="headerlink" href="#creating-a-custom-training-profile" title="Permalink to this heading">#</a></h1>
 <p><em>Case: You’ve created a project with training data on one computer, and you want to use a different computer for training models with custom hyperparameters.</em></p>
 <p>Hyperparameters include the model architecture, learning rate, and data augmentation settings. While model <strong>parameters</strong> are learned from your data during training, <strong>hyperparameters</strong> are not learned from your data—they have to be set before training since they control the training process.</p>
 <p>This guide will explain how to create a custom training profile but doesn’t cover how to decide what the hyperpameters should be. For more information about the hyperparameters, see our guide to <a class="reference internal" href="choosing-models.html#choosing-models"><span class="std std-ref">Configuring models</span></a>.</p>
@@ -409,7 +407,7 @@ <h1>Creating a custom training profile</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/gui.html b/guides/gui.html
index 31352164a..0d02479b3 100644
--- a/guides/gui.html
+++ b/guides/gui.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>GUI &#8212; SLEAP (v1.3.3)</title>
+    <title>GUI &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/gui';</script>
     <link rel="canonical" href="/develop/guides/gui.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Command line interfaces" href="cli.html" />
     <link rel="prev" title="Guides" href="index.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -342,14 +340,14 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="gui">
-<span id="id1"></span><h1>GUI<a class="headerlink" href="#gui" title="Permalink to this headline">#</a></h1>
+<span id="id1"></span><h1>GUI<a class="headerlink" href="#gui" title="Permalink to this heading">#</a></h1>
 <p>The SLEAP labeling interface is accessible via the <code class="code docutils literal notranslate"><span class="pre">sleap_label</span></code> command (see <a class="reference internal" href="cli.html#cli"><span class="std std-ref">Command line interfaces</span></a>).</p>
 <section id="menus">
-<h2>Menus<a class="headerlink" href="#menus" title="Permalink to this headline">#</a></h2>
+<h2>Menus<a class="headerlink" href="#menus" title="Permalink to this heading">#</a></h2>
 <p>Note that many of the menu command have keyboard shortcuts which can be configured from “<strong>Keyboard Reference</strong>” in the “<strong>Help</strong>” menu.</p>
 </section>
 <section id="file">
-<h2>File<a class="headerlink" href="#file" title="Permalink to this headline">#</a></h2>
+<h2>File<a class="headerlink" href="#file" title="Permalink to this heading">#</a></h2>
 <p>“<strong>New…</strong>”, “<strong>Open…</strong>”, “Save, and “<strong>Save As…</strong>” have their usual behavior.</p>
 <p>“<strong>Import…</strong>” allows you to import the data external formats into a new SLEAP dataset. This includes <a class="reference external" href="http://cocodataset.org/#format-data">COCO</a> keypoint detection (<code class="code docutils literal notranslate"><span class="pre">.json</span></code> files), <a class="reference external" href="http://deeplabcut.org">DeepLabCut</a> (<code class="code docutils literal notranslate"><span class="pre">.csv</span></code>), <a class="reference external" href="http://deepposekit.org">DeepPoseKit</a> (<code class="code docutils literal notranslate"><span class="pre">.h5</span></code>), and <a class="reference external" href="https://github.com/talmo/leap">LEAP</a> (<code class="code docutils literal notranslate"><span class="pre">.mat</span></code>).</p>
 <p>“<strong>Merge Data From…</strong>” allows you to copy labels and/or predictions from another SLEAP dataset into the currently open project. This is useful because you can only train on data from a single SLEAP dataset, so you may need to import data before training. For instance, suppose you’ve trained a model and use it to get predictions on another video. You then open the predictions file in the GUI and realize that you’d like to correct some of the predictions and add those corrections to your training data. You can do this by importing data from the predictions file into the SLEAP project which has your training data.</p>
@@ -357,7 +355,7 @@ <h2>File<a class="headerlink" href="#file" title="Permalink to this headline">#<
 <p>“<strong>Replace Videos…</strong>” allows you to <em>swap</em> the videos currently in your project with other videos. This is useful if you want to have your project access copies of the videos at a different path, e.g., if you copy the videos between a network drive and a local drive and want to change which is used by your project. This can also be used if you want to replace you videos with copies that have been re-encoded, cropped, or edited in some other way that doesn’t affect the frame numbers (since your annotations will be placed directly on the corresponding frames of the new video).</p>
 </section>
 <section id="go">
-<h2>Go<a class="headerlink" href="#go" title="Permalink to this headline">#</a></h2>
+<h2>Go<a class="headerlink" href="#go" title="Permalink to this heading">#</a></h2>
 <p>“<strong>Next Labeled Frame</strong>” will take you to the next frame of the video which has any labeled data—either labels added by the user or predictions.</p>
 <p>“<strong>Next User Labeled Frame</strong>” will take you to the next frame which has labels added by the user, skipping any frames which are unlabeled or have only predictions.</p>
 <p>“<strong>Next Suggestion</strong>” will take you to the next frame in the list of suggested frames. If you are currently on a suggested frame, it will take you to the next frame in the list, which may not be a later frame in the video (or even a frame in the same video). If you are not currently on a suggested frame, it will take you to the nearest subsequent suggested frame.</p>
@@ -367,7 +365,7 @@ <h2>Go<a class="headerlink" href="#go" title="Permalink to this headline">#</a><
 <p>“<strong>Select to Frame…</strong>” allows you to select the clip from the current frame to a specified frame. If you want to select from frames X to Y, you can use <strong>Go to</strong> to go to X then <strong>Select to</strong> to select from X to Y.</p>
 </section>
 <section id="view">
-<span id="id2"></span><h2>View<a class="headerlink" href="#view" title="Permalink to this headline">#</a></h2>
+<span id="id2"></span><h2>View<a class="headerlink" href="#view" title="Permalink to this heading">#</a></h2>
 <p>“<strong>Color Predicted Instances</strong>” allows you to toggle whether <em>predicted</em> instances are all shown in yellow, or whether to apply distinct colors to (e.g.) track identities. You should turn this on when proofreading predictions for a video.</p>
 <p>“<strong>Color Palette</strong>” allows you to choose the palette which will be used for coloring instances. Most of the palettes cycle colors, so that if there are five colors in the palette, the sixth item to color will be the same as the first. A few things to know:</p>
 <ul class="simple">
@@ -381,13 +379,13 @@ <h2>Go<a class="headerlink" href="#go" title="Permalink to this headline">#</a><
 <p>“<strong>Show Node Names</strong>” allows you to toggle the visibility of the node names. This is useful if you have lots of nearby instances or very dense skeletons and the node names make it hard to see where the nodes are located.</p>
 <p>“<strong>Show Edges</strong>” allows you to toggle the visibility of the edges which connect the nodes. This can be useful when you have lots of edges which make it hard to see the features of animals in your video.</p>
 <p>“<strong>Edge Style</strong>” controls whether edges are drawn as thin lines or as wedges which indicate the <a class="reference internal" href="proofreading.html#orientation"><span class="std std-ref">Orientation</span></a> of the instance (as well as the direction of the part affinity field which would be used to predict the connection between nodes when using a “bottom-up” approach).</p>
-<p>“<strong>Trail Length</strong>” allows you to show a trail of where each instance was located in prior frames (the length of the trail is the number of prior frames). This can be useful when proofreading predictions since it can help you detect swaps in the identities of animals across frames.</p>
+<p>“<strong>Trail Length</strong>” allows you to show a trail of where each instance was located in prior frames (the length of the trail is the number of prior frames). This can be useful when proofreading predictions since it can help you detect swaps in the identities of animals across frames. By default, you can only select trail lengths of up to 250 frames. You can use a custom trail length by modifying the default length in the <code class="docutils literal notranslate"><span class="pre">preferences.yaml</span></code> file. However, using trail lengths longer than about 500 frames can result in significant lag.</p>
 <p>“<strong>Fit Instances to View</strong>” allows you to toggle whether the view is auto-zoomed to the instances in each frame. This can be useful when proofreading predictions.</p>
 <p>“<strong>Seekbar Header</strong>” allows you to plot relevant information above the seekbar. Note that the seekbar is not updated when you modify instances; it only updates when you select a new seekbar header.</p>
 <p>“<strong>Videos</strong>”, “<strong>Skeleton</strong>”, “<strong>Instances</strong>”, and “<strong>Labeling Suggestions</strong>” allow you to toggle which information windows are shown (by default these are docked to the right side of the main GUI window).</p>
 </section>
 <section id="labels">
-<h2>Labels<a class="headerlink" href="#labels" title="Permalink to this headline">#</a></h2>
+<h2>Labels<a class="headerlink" href="#labels" title="Permalink to this heading">#</a></h2>
 <p>“<strong>Add Instance</strong>” will add an instance to the current frame. You can also add an instance by right-clicking within the frame. For predicted instances, you can also “convert” the predicted instance to a regular, editable instance by double-clicking on the predicted instance (the predictions are still there, but they won’t be shown unless you delete the editable instance you just created).</p>
 <p>“<strong>Instance Placement Method</strong>” allows you to pick how we determine where to place the instance and its nodes on the video frame. “Best” (the default) will first check for predicted instances in the current frame and create a new editable instance from one of the predicted instances (if you add multiple instances, it will do this for each predicted instance in turn). Otherwise, it will copy the location from another instance in a prior or the current frame, or will use a force-directed graph algorithm to place the nodes. You can also choose the “average” method which creates an “average” instance from the instances you’ve already labeled.</p>
 <p>“<strong>Delete Instance</strong>” will delete the currently selected instance (the selected instance will have an outline drawn around it and will be highlighted in the “Instances” window).</p>
@@ -400,7 +398,7 @@ <h2>Labels<a class="headerlink" href="#labels" title="Permalink to this headline
 <p>“<strong>Clear Selection</strong>” allows you to deselect the selected instance.</p>
 </section>
 <section id="predict">
-<h2>Predict<a class="headerlink" href="#predict" title="Permalink to this headline">#</a></h2>
+<h2>Predict<a class="headerlink" href="#predict" title="Permalink to this heading">#</a></h2>
 <p>“<strong>Run Training…</strong>” allows you to train a set of models from the data in your open project, and then optionally predict on some frames in the project.</p>
 <p>“<strong>Run Inference…</strong>” allows you to generate predictions using a pre-trained set of models. Any trained models in the <code class="docutils literal notranslate"><span class="pre">models</span></code> directory next to your current project will be listed, and you also have the option to select models saved elsewhere.</p>
 <p>“<strong>Evaluate Metrics for Trained Models…</strong>” provides you with information to evaluate all of your trained models. For example, you can see the recall and precision of predictions on frames with ground truth validation data (i.e., data withheld when training).</p>
@@ -414,14 +412,14 @@ <h2>Predict<a class="headerlink" href="#predict" title="Permalink to this headli
 <p>“<strong>Export Video with Visual Annotations…</strong>” allows you to export a video clip with any instances drawn on the frame (much as you can see in the GUI). If you select a clip in the seekbar, just those frames will be included in the new video; otherwise the whole (current) video will be used.</p>
 </section>
 <section id="help">
-<h2>Help<a class="headerlink" href="#help" title="Permalink to this headline">#</a></h2>
+<h2>Help<a class="headerlink" href="#help" title="Permalink to this heading">#</a></h2>
 <p>“<strong>Keyboard Shortcuts</strong>” allows you to view and change keyboard shortcuts for common menu and GUI actions.</p>
 </section>
 <section id="application-gui">
-<h2>Application GUI<a class="headerlink" href="#application-gui" title="Permalink to this headline">#</a></h2>
+<h2>Application GUI<a class="headerlink" href="#application-gui" title="Permalink to this heading">#</a></h2>
 </section>
 <section id="mouse">
-<h2>Mouse<a class="headerlink" href="#mouse" title="Permalink to this headline">#</a></h2>
+<h2>Mouse<a class="headerlink" href="#mouse" title="Permalink to this heading">#</a></h2>
 <p><strong>Mouse wheel</strong> on image: zoom in / out</p>
 <p><strong>Left-click (hold) + drag</strong> on image: drag image</p>
 <p><strong>Right-click</strong> (or control + click) on node: Toggle visibility</p>
@@ -437,7 +435,7 @@ <h2>Mouse<a class="headerlink" href="#mouse" title="Permalink to this headline">
 <p><strong>Click</strong> elsewhere on image: Clear selection</p>
 </section>
 <section id="navigation-keys">
-<h2>Navigation Keys<a class="headerlink" href="#navigation-keys" title="Permalink to this headline">#</a></h2>
+<h2>Navigation Keys<a class="headerlink" href="#navigation-keys" title="Permalink to this heading">#</a></h2>
 <p><strong>Right arrow</strong>: Move one frame forward</p>
 <p><strong>Left arrow</strong>: Move one frame back</p>
 <p><strong>Control + Right arrow</strong>: Move a <em>medium</em> step forward (4 frames by default)</p>
@@ -453,7 +451,7 @@ <h2>Navigation Keys<a class="headerlink" href="#navigation-keys" title="Permalin
 </div>
 </section>
 <section id="selection-keys">
-<span id="id3"></span><h2>Selection Keys<a class="headerlink" href="#selection-keys" title="Permalink to this headline">#</a></h2>
+<span id="id3"></span><h2>Selection Keys<a class="headerlink" href="#selection-keys" title="Permalink to this heading">#</a></h2>
 <p><em>Number</em> (e.g., <strong>2</strong>): Select the instance corresponding to that number</p>
 <p><strong>Control (hold)</strong> while instance is selected: Show color-coded list of numbered
 tracks (<strong>Show tracks legend</strong> key)</p>
@@ -461,14 +459,14 @@ <h2>Navigation Keys<a class="headerlink" href="#navigation-keys" title="Permalin
 <p><strong>Escape</strong>: Deselect all instances</p>
 </section>
 <section id="seekbar">
-<h2>Seekbar<a class="headerlink" href="#seekbar" title="Permalink to this headline">#</a></h2>
+<h2>Seekbar<a class="headerlink" href="#seekbar" title="Permalink to this heading">#</a></h2>
 <p><strong>Shift + drag</strong>: Select a range of frames</p>
 <p><strong>Shift + left-click</strong>: Clear frame selection</p>
 <p><strong>Alt + left-click (hold) + drag</strong>: Zoom into a range of frames</p>
 <p><strong>Alt + left-click</strong>: Zoom out so that all frames are visible in seekbar</p>
 </section>
 <section id="labeling-suggestions">
-<span id="suggestion-methods"></span><h2>Labeling Suggestions<a class="headerlink" href="#labeling-suggestions" title="Permalink to this headline">#</a></h2>
+<span id="suggestion-methods"></span><h2>Labeling Suggestions<a class="headerlink" href="#labeling-suggestions" title="Permalink to this heading">#</a></h2>
 <p>There are various methods to generate a list “suggested” frames for labeling or proofreading.</p>
 <p>The <strong>sample</strong> method is a quick way to get some number of frames for every video in your project. You can tell it how many samples (frames) to take from each video, and whether they should be evenly spaced throughout a video (the “stride” sampling method) or randomly distributed.</p>
 <p>The <strong>image feature</strong> method uses various algorithms to give you visually distinctive frames, since you will be able to train more robust models if the frames you’ve labeled are more representative of the visual variations in your videos. Generating suggestions based on image features can be slow.</p>
@@ -561,7 +559,7 @@ <h2>Seekbar<a class="headerlink" href="#seekbar" title="Permalink to this headli
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/index.html b/guides/index.html
index fc04dec46..f2a43bd04 100644
--- a/guides/index.html
+++ b/guides/index.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Guides &#8212; SLEAP (v1.3.3)</title>
+    <title>Guides &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/index';</script>
     <link rel="canonical" href="/develop/guides/index.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="GUI" href="gui.html" />
     <link rel="prev" title="Export Data For Analysis" href="../tutorials/analysis.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,6 +318,7 @@ <h2> Contents </h2>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#general">General</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#getting-better-results">Getting better results</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#running-remotely">Running remotely</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap-with-bonsai">SLEAP with Bonsai</a></li>
 </ul>
             </nav>
         </div>
@@ -332,14 +331,14 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="guides">
-<span id="id1"></span><h1>Guides<a class="headerlink" href="#guides" title="Permalink to this headline">#</a></h1>
+<span id="id1"></span><h1>Guides<a class="headerlink" href="#guides" title="Permalink to this heading">#</a></h1>
 <section id="general">
-<h2>General<a class="headerlink" href="#general" title="Permalink to this headline">#</a></h2>
+<h2>General<a class="headerlink" href="#general" title="Permalink to this heading">#</a></h2>
 <p><a class="reference internal" href="gui.html#gui"><span class="std std-ref">GUI</span></a> is a comprehensive list of the functionality in the SLEAP GUI for labeling, proofreading and inspection.</p>
 <p><a class="reference internal" href="cli.html#cli"><span class="std std-ref">Command line interfaces</span></a> describes the command line interfaces provided by SLEAP.</p>
 </section>
 <section id="getting-better-results">
-<h2>Getting better results<a class="headerlink" href="#getting-better-results" title="Permalink to this headline">#</a></h2>
+<h2>Getting better results<a class="headerlink" href="#getting-better-results" title="Permalink to this heading">#</a></h2>
 <p><a class="reference internal" href="troubleshooting-workflows.html"><span class="doc std std-doc">Troubleshooting workflows</span></a> provide a good place to start with suggestions based on how much you’ve labeled so far.</p>
 <p><a class="reference internal" href="skeletons.html#skeletons"><span class="std std-ref">Skeleton design</span></a> answers some questions about designing the skeleton for your animals.</p>
 <p><a class="reference internal" href="choosing-models.html#choosing-models"><span class="std std-ref">Configuring models</span></a> describes the types of models in SLEAP and how to configure them.</p>
@@ -347,11 +346,15 @@ <h2>Getting better results<a class="headerlink" href="#getting-better-results" t
 <p><a class="reference internal" href="proofreading.html#proofreading"><span class="std std-ref">Tracking and proofreading</span></a> provides tips and tools you can use to speed up proofreading when you’re happy enough with the frame-by-frame predictions but you need to correct the identities tracked across frames.</p>
 </section>
 <section id="running-remotely">
-<h2>Running remotely<a class="headerlink" href="#running-remotely" title="Permalink to this headline">#</a></h2>
+<h2>Running remotely<a class="headerlink" href="#running-remotely" title="Permalink to this heading">#</a></h2>
 <p><a class="reference internal" href="colab.html#colab"><span class="std std-ref">Run training and inference on Colab</span></a> when you have a project with labeled training data and you’d like to run training or inference in a <strong>Colab</strong> notebook.</p>
 <p><a class="reference internal" href="remote.html#remote-train"><span class="std std-ref">Remote training</span></a> when you have a project with training data and you want to train on a different machine using a <strong>command-line interface</strong>.</p>
 <p><a class="reference internal" href="custom-training.html#custom-training"><span class="std std-ref">Creating a custom training profile</span></a> for creating <strong>custom training profiles</strong> (i.e., non-default model hyperparameters) from the GUI.</p>
 <p><a class="reference internal" href="remote.html#remote-inference"><span class="std std-ref">Remote inference</span></a> when you trained models and you want to run inference on a different machine using a <strong>command-line interface</strong>.</p>
+</section>
+<section id="sleap-with-bonsai">
+<h2>SLEAP with Bonsai<a class="headerlink" href="#sleap-with-bonsai" title="Permalink to this heading">#</a></h2>
+<p><a class="reference internal" href="bonsai.html#bonsai"><span class="std std-ref">Using Bonsai with SLEAP</span></a> when you want to analyze the trained SLEAP model to visualize the poses, centroids and identities for further visual analysis.</p>
 <div class="toctree-wrapper compound">
 </div>
 </section>
@@ -408,6 +411,7 @@ <h2>Running remotely<a class="headerlink" href="#running-remotely" title="Permal
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#general">General</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#getting-better-results">Getting better results</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#running-remotely">Running remotely</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#sleap-with-bonsai">SLEAP with Bonsai</a></li>
 </ul>
   </nav></div>
 
@@ -431,7 +435,7 @@ <h2>Running remotely<a class="headerlink" href="#running-remotely" title="Permal
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/merging.html b/guides/merging.html
index 3a05c2ada..940dc603c 100644
--- a/guides/merging.html
+++ b/guides/merging.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Importing predictions for labeling &#8212; SLEAP (v1.3.3)</title>
+    <title>Importing predictions for labeling &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/merging';</script>
     <link rel="canonical" href="/develop/guides/merging.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Tracking and proofreading" href="proofreading.html" />
     <link rel="prev" title="Configuring models" href="choosing-models.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Importing predictions for labeling</h1>
                 <article class="bd-article" role="main">
                   
   <section id="importing-predictions-for-labeling">
-<span id="merging"></span><h1>Importing predictions for labeling<a class="headerlink" href="#importing-predictions-for-labeling" title="Permalink to this headline">#</a></h1>
+<span id="merging"></span><h1>Importing predictions for labeling<a class="headerlink" href="#importing-predictions-for-labeling" title="Permalink to this heading">#</a></h1>
 <p><em>Case: You have predictions that aren’t in the same project as your original training data and you want to correct some of the predictions and use these corrections to train a better model.</em></p>
 <p>All of your training data must be in a single SLEAP project file (or labels package), so if you have data in multiple files, you’ll need to merge them before you can train on the entire set of data.</p>
 <p>When you run inference from the GUI, the predictions will be added to the same project as your training data (they’ll also be saved in a separate file). When you run inference from the command-line, they’ll only be in a separate file.</p>
@@ -393,7 +391,7 @@ <h1>Importing predictions for labeling</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/proofreading.html b/guides/proofreading.html
index dd60aa518..de6552910 100644
--- a/guides/proofreading.html
+++ b/guides/proofreading.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Tracking and proofreading &#8212; SLEAP (v1.3.3)</title>
+    <title>Tracking and proofreading &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/proofreading';</script>
     <link rel="canonical" href="/develop/guides/proofreading.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Run training and inference on Colab" href="colab.html" />
     <link rel="prev" title="Importing predictions for labeling" href="merging.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -335,14 +333,14 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="tracking-and-proofreading">
-<span id="proofreading"></span><h1>Tracking and proofreading<a class="headerlink" href="#tracking-and-proofreading" title="Permalink to this headline">#</a></h1>
+<span id="proofreading"></span><h1>Tracking and proofreading<a class="headerlink" href="#tracking-and-proofreading" title="Permalink to this heading">#</a></h1>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p><em>Case: You’re happy enough with the frame-by-frame predictions but you need to correct the identities tracked across frames.</em></p>
 <p>The basics of <a class="reference internal" href="../tutorials/proofreading.html#track-proofreading"><span class="std std-ref">Track proofreading</span></a> are covered in the <a class="reference internal" href="../tutorials/tutorial.html#tutorial"><span class="std std-ref">Tutorial</span></a>. You should go read that if you haven’t already. Here we’ll go into more details.</p>
 </div>
 <section id="tracking-methods">
-<span id="tracking-method-details"></span><h2>Tracking methods<a class="headerlink" href="#tracking-methods" title="Permalink to this headline">#</a></h2>
+<span id="tracking-method-details"></span><h2>Tracking methods<a class="headerlink" href="#tracking-methods" title="Permalink to this heading">#</a></h2>
 <p>The process of predicting instances frame-by-frame and the process of putting these together into <strong>tracks</strong> (i.e., identities across frames) are distinct, although it’s common to run them together during the inference pipeline. Obviously you can only track identities after you’ve predicted instances, but once you have predictions, it’s easy to then run tracking by itself to try out different methods and parameters.</p>
 <p>If you’re getting poor results, you may want to try out different methods and parameters. Changing the <strong>track window</strong> and the <strong>similarity</strong> method—both explained below—can make a big difference.</p>
 <p>The tracker will go through the predictions frame by frame and try to match instances on frame N to candidates which already have track identities assigned. These candidates are generated from a certain number of frames immediately prior to frame N (what we refer to as the “tracking window”).</p>
@@ -352,6 +350,8 @@ <h2> Contents </h2>
 <li><p>“<strong>centroid</strong>” measures similarity by the distance between the instance centroids</p></li>
 <li><p>“<strong>iou</strong>” measures similarity by the intersection/overlap of the instance bounding boxes</p></li>
 <li><p>“<strong>instance</strong>” measures similarity by looking at the distances between corresponding nodes in the instances, normalized by the number of valid nodes in the candidate instance.</p></li>
+<li><p>“<strong>normalized_instance</strong>” measures similarity by looking at the distances between corresponding nodes in the instances, normalized by the number of valid nodes in the candidate instance and the keypoints normalized by the image size.</p></li>
+<li><p>“<strong>object_keypoint</strong>” measures similarity by measuring the distance between each keypoints from a reference instance and a query instance, takes the exp(-d**2), sum for all the keypoints and divide by the number of visible keypoints in the reference instance.</p></li>
 </ul>
 <p>Once SLEAP has measured the similarity between all the candidates and the instances in frame N, you need to choose a way to pair them up. You can do this either by picking the best match, and the picking the best remaining match for each remaining instance in turn—this is “<strong>greedy</strong>” matching—or you can find the way of matching identities which minimizes the total cost (or: maximizes the total similarity)—this is “<strong>Hungarian</strong>” matching.</p>
 <p>Finally, you have an optional second-pass method which “cleans” the resulting
@@ -361,7 +361,7 @@ <h2> Contents </h2>
 frame by frame and removes (or culls) instances over this target number. To cull
 to a target number of instances per frame, navigate to the Inference Pipeline
 via Predict &gt;&gt; Inference, then:</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>Specify the <strong>Tracker (cross-frame identity) Method</strong></p></li>
 <li><p>Uncheck the <strong>No target</strong> checkbox</p></li>
 <li><p>Specify the <strong>Target Number of Instances Per Frame</strong></p></li>
@@ -371,11 +371,11 @@ <h2> Contents </h2>
 <p>Once you have the desired number of instances in every frame, SLEAP connects identities with a simple heuristic: if exactly one track identity was dropped from frame N and exactly one new track identity was added in frame N+1, it matches up the dropped and the new tracks.</p>
 </section>
 <section id="more-training-data">
-<h2>More training data?<a class="headerlink" href="#more-training-data" title="Permalink to this headline">#</a></h2>
+<h2>More training data?<a class="headerlink" href="#more-training-data" title="Permalink to this heading">#</a></h2>
 <p>Often your models will fail to predict <em>all</em> of the instances on <em>all</em> of the frames. Even if you’re happy enough with the result since you can interpolate missing data, it’s possible that the missing instances will cause problems when we try to determine track identities across frames, so if your tracking results are poor, you may wish to <a class="reference internal" href="merging.html#merging"><span class="std std-ref">Importing predictions for labeling</span></a>.</p>
 </section>
 <section id="the-track-cleaning-script">
-<h2>The “track cleaning” script<a class="headerlink" href="#the-track-cleaning-script" title="Permalink to this headline">#</a></h2>
+<h2>The “track cleaning” script<a class="headerlink" href="#the-track-cleaning-script" title="Permalink to this heading">#</a></h2>
 <p>There’s an experimental command-line utility which tries to match up lost identities. You need to give it a predictions file which already has track assignments, and specify how many instances there should be. It looks for frames where there’s exactly one lost identity and exactly one newly spawned identity, and it joins these into a single track. Suppose you have a predictions file at <code class="code docutils literal notranslate"><span class="pre">path/to/predictions.h5</span></code> and you want to end up with three distinct tracks. You can run</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">python</span> <span class="o">-</span><span class="n">m</span> <span class="n">sleap</span><span class="o">.</span><span class="n">info</span><span class="o">.</span><span class="n">trackcleaner</span> <span class="n">path</span><span class="o">/</span><span class="n">to</span><span class="o">/</span><span class="n">predictions</span><span class="o">.</span><span class="n">h5</span> <span class="o">-</span><span class="n">c</span> <span class="mi">3</span>
 </pre></div>
@@ -384,7 +384,7 @@ <h2>The “track cleaning” script<a class="headerlink" href="#the-track-cleani
 <p>The main worry is that this script will connect identities which should be distinct, so that in place of <strong>lost</strong> identities you’ll now have more <strong>mistaken</strong> identities, which can be harder to locate when proofreading. Tools and techniques for finding <strong>mistaken</strong> identities during proofreading are explained below.</p>
 </section>
 <section id="color-palettes">
-<h2>Color palettes<a class="headerlink" href="#color-palettes" title="Permalink to this headline">#</a></h2>
+<h2>Color palettes<a class="headerlink" href="#color-palettes" title="Permalink to this heading">#</a></h2>
 <p>When you’re proofreading track identities, the first step should always be to enable “<strong>Color Predicted Instances</strong>” in the View menu. Choosing the right color palette can also make a difference. If there are a small number of instances you’re tracking, the “five+” palette will make it easier to see instances which were assigned to later tracks, both in on the video frame:</p>
 <img alt="../_images/orange-leg.jpg" src="../_images/orange-leg.jpg" />
 <p>and on the seekbar:</p>
@@ -393,13 +393,13 @@ <h2>Color palettes<a class="headerlink" href="#color-palettes" title="Permalink
 <p>Sometimes the background in the video will make it hard to see certain colors in a palette. It’s possible to edit palettes, as explained in the <a class="reference internal" href="gui.html#view"><span class="std std-ref">View</span></a> menu section of the <a class="reference internal" href="gui.html#gui"><span class="std std-ref">GUI</span></a>.</p>
 </section>
 <section id="id1">
-<h2>Proofreading<a class="headerlink" href="#id1" title="Permalink to this headline">#</a></h2>
+<h2>Proofreading<a class="headerlink" href="#id1" title="Permalink to this heading">#</a></h2>
 <p>As discussed in the <a class="reference internal" href="../tutorials/proofreading.html#track-proofreading"><span class="std std-ref">Track proofreading</span></a> section of the <a class="reference internal" href="../tutorials/tutorial.html#tutorial"><span class="std std-ref">Tutorial</span></a>, there are two main types of mistakes made by the tracking code: lost identities and mistaken
 identities.</p>
 <p><strong>Lost Identities:</strong> The code may fail to identity an instance in one
 frame with any instances from previous frames.</p>
 <p>Here’s a strategy that works well for fixing <strong>lost</strong> identities:</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>Turn on colors for predicted instances and use a good color palette (as explained above).</p></li>
 <li><p>Change the “<strong>Trail Length</strong>” to a number greater than zero. These trails show where instances in each track were in prior frames.</p></li>
 <li><p>Use the keyboard shortcut for the “<strong>Next Track Spawn Frame</strong>” command in the <strong>“Go”</strong> menu to jump to frames where a new track identity is spawned.</p></li>
@@ -424,7 +424,7 @@ <h2>Proofreading<a class="headerlink" href="#id1" title="Permalink to this headl
 switching the tracks in one frame will also be applied in all subsequent frames.</p>
 </section>
 <section id="orientation">
-<span id="id2"></span><h2>Orientation<a class="headerlink" href="#orientation" title="Permalink to this headline">#</a></h2>
+<span id="id2"></span><h2>Orientation<a class="headerlink" href="#orientation" title="Permalink to this heading">#</a></h2>
 <p>In some cases it may be difficult to see the orientation of the predicted instances. You can make it easier to see the orientation by changing the style of the edges drawn between nodes from thin lines (as shown above) to <strong>wedges</strong>, as shown here:</p>
 <img alt="../_images/wedges.jpg" src="../_images/wedges.jpg" />
 <p>The wedges point from each <strong>source</strong> node to its <strong>destination</strong> node(s) in your skeleton. You can set the edge style using the “<strong>Edge Style</strong>” submenu in the “View” menu.</p>
@@ -508,7 +508,7 @@ <h2>Proofreading<a class="headerlink" href="#id1" title="Permalink to this headl
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/remote.html b/guides/remote.html
index 92e26c88c..e5766de87 100644
--- a/guides/remote.html
+++ b/guides/remote.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Running SLEAP remotely &#8212; SLEAP (v1.3.3)</title>
+    <title>Running SLEAP remotely &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,24 +42,21 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/remote';</script>
     <link rel="canonical" href="/develop/guides/remote.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
-    <link rel="next" title="Notebooks" href="../notebooks/index.html" />
+    <link rel="next" title="Using Bonsai with SLEAP" href="bonsai.html" />
     <link rel="prev" title="Creating a custom training profile" href="custom-training.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2 current active"><a class="current reference internal" href="#">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -331,12 +329,12 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="running-sleap-remotely">
-<span id="remote"></span><h1>Running SLEAP remotely<a class="headerlink" href="#running-sleap-remotely" title="Permalink to this headline">#</a></h1>
+<span id="remote"></span><h1>Running SLEAP remotely<a class="headerlink" href="#running-sleap-remotely" title="Permalink to this heading">#</a></h1>
 <section id="remote-training">
-<span id="remote-train"></span><h2>Remote training<a class="headerlink" href="#remote-training" title="Permalink to this headline">#</a></h2>
+<span id="remote-train"></span><h2>Remote training<a class="headerlink" href="#remote-training" title="Permalink to this heading">#</a></h2>
 <p><em>Case: You already have a project with training data and you want to train on a different machine using a command-line interface.</em></p>
 <p>You need three things to run training:</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>You need to install SLEAP on the remote machine where you’ll run training.</p></li>
 <li><p>Labels and images to use for training.</p></li>
 <li><p>A training profile which defines the training parameters (e.g., learning rate, image augmentation methods).</p></li>
@@ -363,10 +361,10 @@ <h2> Contents </h2>
 </div>
 </section>
 <section id="remote-inference">
-<span id="id1"></span><h2>Remote inference<a class="headerlink" href="#remote-inference" title="Permalink to this headline">#</a></h2>
+<span id="id1"></span><h2>Remote inference<a class="headerlink" href="#remote-inference" title="Permalink to this heading">#</a></h2>
 <p><em>Case: You already have models and you want to run inference on a different machine using a command-line interface.</em></p>
 <p>Here’s what you need to run inference:</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>You need to install SLEAP on the remote machine where you’ll run training.</p></li>
 <li><p>You need a compatible set of trained model files.</p></li>
 <li><p>You need a video for which you want predictions.</p></li>
@@ -426,11 +424,11 @@ <h2> Contents </h2>
       </div>
     </a>
     <a class="right-next"
-       href="../notebooks/index.html"
+       href="bonsai.html"
        title="next page">
       <div class="prev-next-info">
         <p class="prev-next-subtitle">next</p>
-        <p class="prev-next-title">Notebooks</p>
+        <p class="prev-next-title">Using Bonsai with SLEAP</p>
       </div>
       <i class="fa-solid fa-angle-right"></i>
     </a>
@@ -477,7 +475,7 @@ <h2> Contents </h2>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/skeletons.html b/guides/skeletons.html
index c2db96e90..7478c261a 100644
--- a/guides/skeletons.html
+++ b/guides/skeletons.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Skeleton design &#8212; SLEAP (v1.3.3)</title>
+    <title>Skeleton design &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/skeletons';</script>
     <link rel="canonical" href="/develop/guides/skeletons.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Configuring models" href="choosing-models.html" />
     <link rel="prev" title="Troubleshooting workflows" href="troubleshooting-workflows.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -330,7 +328,7 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="skeleton-design">
-<span id="skeletons"></span><h1>Skeleton design<a class="headerlink" href="#skeleton-design" title="Permalink to this headline">#</a></h1>
+<span id="skeletons"></span><h1>Skeleton design<a class="headerlink" href="#skeleton-design" title="Permalink to this heading">#</a></h1>
 <p>In SLEAP, skeletons are defined as a set of <em>nodes</em> (landmarks types) and <em>edges</em>
 (connections between landmarks).</p>
 <p><strong>Nodes</strong> are essentially just a list of body part names. The actual naming doesn’t
@@ -341,11 +339,11 @@ <h2> Contents </h2>
 serve primarily for visualization.</p>
 <p>In bottom-up models, the edges are important and must connect all nodes to one another.</p>
 <section id="tips">
-<h2>Tips<a class="headerlink" href="#tips" title="Permalink to this headline">#</a></h2>
+<h2>Tips<a class="headerlink" href="#tips" title="Permalink to this heading">#</a></h2>
 <p><strong>How do I choose the best skeleton for my data?</strong></p>
 <p>In general, here are the rules of thumb you want to follow for designing an ideal
 skeleton:</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>Choose <em>nodes</em> that will be easy to locate in new images. It’s important to be as
 consistent as possible about the relative placement of body parts.</p></li>
 <li><p>When choosing <em>edges</em>, try to form a shallow tree. This is because it’s preferable to
@@ -433,7 +431,7 @@ <h2>Tips<a class="headerlink" href="#tips" title="Permalink to this headline">#<
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/training.html b/guides/training.html
index ec4250f47..43146b9d1 100644
--- a/guides/training.html
+++ b/guides/training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Training with GUI &#8212; SLEAP (v1.3.3)</title>
+    <title>Training with GUI &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/training';</script>
     <link rel="canonical" href="/develop/guides/training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -318,7 +316,7 @@ <h1>Training with GUI</h1>
                 <article class="bd-article" role="main">
                   
   <section id="training-with-gui">
-<h1>Training with GUI<a class="headerlink" href="#training-with-gui" title="Permalink to this headline">#</a></h1>
+<h1>Training with GUI<a class="headerlink" href="#training-with-gui" title="Permalink to this heading">#</a></h1>
 <p>To run training, select “<strong>Run Training…</strong>” from the “Predict”
 menu. This trains new models using the frames that you’ve already
 labeled in your project, and then uses these models to predict instances
@@ -407,7 +405,7 @@ <h1>Training with GUI<a class="headerlink" href="#training-with-gui" title="Perm
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/guides/troubleshooting-workflows.html b/guides/troubleshooting-workflows.html
index 8aab75b85..be826a9ac 100644
--- a/guides/troubleshooting-workflows.html
+++ b/guides/troubleshooting-workflows.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Troubleshooting workflows &#8212; SLEAP (v1.3.3)</title>
+    <title>Troubleshooting workflows &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'guides/troubleshooting-workflows';</script>
     <link rel="canonical" href="/develop/guides/troubleshooting-workflows.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Skeleton design" href="skeletons.html" />
     <link rel="prev" title="Command line interfaces" href="cli.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Troubleshooting workflows</h1>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="troubleshooting-workflows">
-<h1>Troubleshooting workflows<a class="headerlink" href="#troubleshooting-workflows" title="Permalink to this headline">#</a></h1>
+<h1>Troubleshooting workflows<a class="headerlink" href="#troubleshooting-workflows" title="Permalink to this heading">#</a></h1>
 <p>SLEAP can work with any type of data, but sometimes it may be helpful to tweak certain
 configurations or try out different parameters to improve performance.</p>
 <p>For the majority of cases, we recommend following these steps based on how far along you
@@ -400,7 +398,7 @@ <h1>Troubleshooting workflows<a class="headerlink" href="#troubleshooting-workfl
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/help.html b/help.html
index 7b8fd7d51..aa687e1ef 100644
--- a/help.html
+++ b/help.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Help &#8212; SLEAP (v1.3.3)</title>
+    <title>Help &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'help';</script>
     <link rel="canonical" href="/develop/help.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -58,7 +55,7 @@
     <link rel="search" title="Search" href="search.html" />
     <link rel="prev" title="Datasets" href="datasets.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -164,6 +161,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -352,35 +350,35 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="help">
-<h1>Help<a class="headerlink" href="#help" title="Permalink to this headline">#</a></h1>
+<h1>Help<a class="headerlink" href="#help" title="Permalink to this heading">#</a></h1>
 <p>Stuck? Can’t get SLEAP to run? Crashing? Try the recommended tips below.</p>
 <section id="installation">
-<h2>Installation<a class="headerlink" href="#installation" title="Permalink to this headline">#</a></h2>
+<h2>Installation<a class="headerlink" href="#installation" title="Permalink to this heading">#</a></h2>
 <section id="i-cant-get-sleap-to-install">
-<h3>I can’t get SLEAP to install!<a class="headerlink" href="#i-cant-get-sleap-to-install" title="Permalink to this headline">#</a></h3>
+<h3>I can’t get SLEAP to install!<a class="headerlink" href="#i-cant-get-sleap-to-install" title="Permalink to this heading">#</a></h3>
 <p>Have you tried all of the steps in the <a class="reference internal" href="installation.html"><span class="doc std std-doc">installation instructions</span></a>?</p>
 <p>If so, please feel free to <a class="reference external" href="https://github.com/talmolab/sleap/discussions">start a discussion</a> or <a class="reference external" href="https://github.com/talmolab/sleap/issues">open an issue</a> and tell us how you’re trying to install it, what error messages you’re getting and which operating system you’re on.</p>
 </section>
 <section id="can-i-install-it-on-a-computer-without-a-gpu">
-<h3>Can I install it on a computer without a GPU?<a class="headerlink" href="#can-i-install-it-on-a-computer-without-a-gpu" title="Permalink to this headline">#</a></h3>
+<h3>Can I install it on a computer without a GPU?<a class="headerlink" href="#can-i-install-it-on-a-computer-without-a-gpu" title="Permalink to this heading">#</a></h3>
 <p>Yes! You can install SLEAP as you normally would using the <code class="docutils literal notranslate"><span class="pre">conda</span></code> or <code class="docutils literal notranslate"><span class="pre">pip</span></code>-based methods and the GPU support will be ignored.</p>
 </section>
 <section id="what-if-i-already-have-cuda-set-up-on-my-system">
-<h3>What if I already have CUDA set up on my system?<a class="headerlink" href="#what-if-i-already-have-cuda-set-up-on-my-system" title="Permalink to this headline">#</a></h3>
-<p>You can use the system CUDA installation by simply using the <a class="reference internal" href="installation.html#pip-package"><span class="std std-doc">pip package</span></a> installation method.</p>
+<h3>What if I already have CUDA set up on my system?<a class="headerlink" href="#what-if-i-already-have-cuda-set-up-on-my-system" title="Permalink to this heading">#</a></h3>
+<p>You can use the system CUDA installation by simply using the <span class="xref myst"></span> installation method.</p>
 <p>Note that you will need to use a version compatible with <strong>TensorFlow 2.6+</strong> (<strong>CUDA Toolkit v11.3</strong> and <strong>cuDNN v8.2</strong>).</p>
 </section>
 </section>
 <section id="usage">
-<h2>Usage<a class="headerlink" href="#usage" title="Permalink to this headline">#</a></h2>
+<h2>Usage<a class="headerlink" href="#usage" title="Permalink to this heading">#</a></h2>
 <section id="how-do-i-use-sleap">
-<h3>How do I use SLEAP?<a class="headerlink" href="#how-do-i-use-sleap" title="Permalink to this headline">#</a></h3>
+<h3>How do I use SLEAP?<a class="headerlink" href="#how-do-i-use-sleap" title="Permalink to this heading">#</a></h3>
 <p>If you’re new to pose tracking in general, check out <a class="reference external" href="https://cbmm.mit.edu/video/decoding-animal-behavior-through-pose-tracking">this talk</a> or our review in <em><a class="reference external" href="https://rdcu.be/caH3H">Nature Neuroscience</a></em>.</p>
 <p>If you’re just new to SLEAP, we suggest starting with the <a class="reference internal" href="overview.html#high-level-overview"><span class="std std-ref">Overview</span></a> and then following the <a class="reference internal" href="tutorials/tutorial.html#tutorial"><span class="std std-ref">Tutorial</span></a>.</p>
 <p>Once you get the hang of it, check out the <a class="reference internal" href="guides/index.html#guides"><span class="std std-ref">Guides</span></a> for more detailed info.</p>
 </section>
 <section id="does-my-data-need-to-be-in-a-particular-format">
-<span id="reencoding"></span><h3>Does my data need to be in a particular format?<a class="headerlink" href="#does-my-data-need-to-be-in-a-particular-format" title="Permalink to this headline">#</a></h3>
+<span id="reencoding"></span><h3>Does my data need to be in a particular format?<a class="headerlink" href="#does-my-data-need-to-be-in-a-particular-format" title="Permalink to this heading">#</a></h3>
 <p>SLEAP supports a large number of formats, including all video formats and imported data from DeepLabCut and others.</p>
 <p>Many types of video acquisition software, however, do not save videos in a format suitable for computer vision-based processing. A very common issue is that videos are not <strong>reliably seekable</strong>, meaning that you may not get the same data when trying to read a particular frame index. This is because many video formats are optimized for realtime and sequential playback and save space by reconstructing the image using data in adjacent frames. The consequence is that you may not get the exact same image depending on what the last frame you read was. Check out <a class="reference external" href="http://blog.loopbio.com/video-io-1-introduction.html">this blog post</a> for more details.</p>
 <p>If you think you may be affected by this issue, or just want to be safe, re-encode your videos using the following command:</p>
@@ -399,35 +397,35 @@ <h3>How do I use SLEAP?<a class="headerlink" href="#how-do-i-use-sleap" title="P
 <p>If you don’t have <code class="docutils literal notranslate"><span class="pre">ffmpeg</span></code> on your system, you can install it using <code class="docutils literal notranslate"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">ffmpeg</span></code> or by downloading it from the <a class="reference external" href="https://ffmpeg.org/download.html">official website</a>.</p>
 </section>
 <section id="i-get-strange-results-where-the-poses-appear-to-be-correct-but-shifted-relative-to-the-image">
-<h3>I get strange results where the poses appear to be correct but shifted relative to the image.<a class="headerlink" href="#i-get-strange-results-where-the-poses-appear-to-be-correct-but-shifted-relative-to-the-image" title="Permalink to this headline">#</a></h3>
+<h3>I get strange results where the poses appear to be correct but shifted relative to the image.<a class="headerlink" href="#i-get-strange-results-where-the-poses-appear-to-be-correct-but-shifted-relative-to-the-image" title="Permalink to this heading">#</a></h3>
 <p>This is most likely an issue with the video compression format. <a class="reference internal" href="#reencoding"><span class="std std-ref">See above</span></a>.</p>
 </section>
 <section id="how-do-i-get-predictions-out">
-<h3>How do I get predictions out?<a class="headerlink" href="#how-do-i-get-predictions-out" title="Permalink to this headline">#</a></h3>
+<h3>How do I get predictions out?<a class="headerlink" href="#how-do-i-get-predictions-out" title="Permalink to this heading">#</a></h3>
 <p>See <a class="reference internal" href="tutorials/analysis.html#export-analysis"><span class="std std-ref">Export Data For Analysis</span></a> and <a class="reference internal" href="guides/cli.html#sleap-convert"><span class="std std-ref">sleap-convert</span></a>.</p>
 </section>
 <section id="what-do-i-do-with-the-output-of-sleap">
-<h3>What do I do with the output of SLEAP?<a class="headerlink" href="#what-do-i-do-with-the-output-of-sleap" title="Permalink to this headline">#</a></h3>
+<h3>What do I do with the output of SLEAP?<a class="headerlink" href="#what-do-i-do-with-the-output-of-sleap" title="Permalink to this heading">#</a></h3>
 <p>Check out the <a class="reference internal" href="notebooks/Analysis_examples.html"><span class="doc std std-doc">Analysis examples</span></a> notebooks.</p>
 </section>
 </section>
 <section id="getting-more-help">
-<h2>Getting more help<a class="headerlink" href="#getting-more-help" title="Permalink to this headline">#</a></h2>
+<h2>Getting more help<a class="headerlink" href="#getting-more-help" title="Permalink to this heading">#</a></h2>
 <section id="ive-found-a-bug-or-have-another-problem">
-<h3>I’ve found a bug or have another problem!<a class="headerlink" href="#ive-found-a-bug-or-have-another-problem" title="Permalink to this headline">#</a></h3>
+<h3>I’ve found a bug or have another problem!<a class="headerlink" href="#ive-found-a-bug-or-have-another-problem" title="Permalink to this heading">#</a></h3>
 <p>Feel free to <a class="reference external" href="https://github.com/talmolab/sleap/discussions">start a discussion</a> to get help from the developers and community.
 Or <a class="reference external" href="https://github.com/talmolab/sleap/issues">open an issue</a> and we’ll get back to you as soon as we can!</p>
 </section>
 <section id="can-i-just-talk-to-someone">
-<h3>Can I just talk to someone?<a class="headerlink" href="#can-i-just-talk-to-someone" title="Permalink to this headline">#</a></h3>
+<h3>Can I just talk to someone?<a class="headerlink" href="#can-i-just-talk-to-someone" title="Permalink to this heading">#</a></h3>
 <p>SLEAP is a complex machine learning system intended for general use, so it’s possible that we failed to consider the specifics of the situation in which you may be interested in using it with.</p>
 <p>Feel free to reach out to us at <code class="docutils literal notranslate"><span class="pre">talmo&#64;salk.edu</span></code> if you have a question that isn’t covered here.</p>
 </section>
 </section>
 <section id="improving-sleap">
-<h2>Improving SLEAP<a class="headerlink" href="#improving-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Improving SLEAP<a class="headerlink" href="#improving-sleap" title="Permalink to this heading">#</a></h2>
 <section id="how-can-i-help-improve-sleap">
-<h3>How can I help improve SLEAP?<a class="headerlink" href="#how-can-i-help-improve-sleap" title="Permalink to this headline">#</a></h3>
+<h3>How can I help improve SLEAP?<a class="headerlink" href="#how-can-i-help-improve-sleap" title="Permalink to this heading">#</a></h3>
 <ul>
 <li><p>Tell your friends about SLEAP! We also love to hear stories about what worked or didn’t work, or your experience if you came from other software tools (<code class="docutils literal notranslate"><span class="pre">talmo&#64;salk.edu</span></code>).</p></li>
 <li><p><a class="reference external" href="https://www.nature.com/articles/s41592-022-01426-1">Cite our paper</a>! Here’s a BibTeX citation for your reference manager:</p>
@@ -469,7 +467,7 @@ <h3>How can I help improve SLEAP?<a class="headerlink" href="#how-can-i-help-imp
 </ul>
 </section>
 <section id="what-is-usage-data">
-<span id="usage-data"></span><h3>What is usage data?<a class="headerlink" href="#what-is-usage-data" title="Permalink to this headline">#</a></h3>
+<span id="usage-data"></span><h3>What is usage data?<a class="headerlink" href="#what-is-usage-data" title="Permalink to this heading">#</a></h3>
 <p>To help us improve SLEAP, you may allow us to collect basic and <strong>anonymous</strong> usage data. If enabled from the <strong>Help</strong> menu, the SLEAP GUI will transmit information such as which version of Python and operating system you are running SLEAP on.</p>
 <p>This helps us understand on which types of computers SLEAP is being used so we can ensure that our software is maximally accessible to the broadest userbase possible, for example, by telling us whether it’s safe to update our dependencies without breaking SLEAP for most users. Collecting usage data also helps us get a sense for how often SLEAP is being used so that we can report its impact to external grant funding agencies.</p>
 <p>You can opt out at any time from the menu (this preference will be stored). If you want to prevent these data from being shared with us, you can launch the GUI with <code class="docutils literal notranslate"><span class="pre">sleap-label</span> <span class="pre">--no-usage-data</span></code>. Usage data is only shared when the GUI is used, not the API or CLIs. You can check out the <a class="reference external" href="https://github.com/talmolab/sleap/blob/main/sleap/gui/web.py">source code</a> to see exactly what data is collected.</p>
@@ -563,7 +561,7 @@ <h3>How can I help improve SLEAP?<a class="headerlink" href="#how-can-i-help-imp
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/index.html b/index.html
index 97e9b4974..bc78d05c0 100644
--- a/index.html
+++ b/index.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Social LEAP Estimates Animal Poses (SLEAP) &#8212; SLEAP (v1.3.3)</title>
+    <title>Social LEAP Estimates Animal Poses (SLEAP) &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'index';</script>
     <link rel="canonical" href="/develop/index.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -58,7 +55,7 @@
     <link rel="search" title="Search" href="search.html" />
     <link rel="next" title="Overview" href="overview.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -164,6 +161,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -339,11 +337,11 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="social-leap-estimates-animal-poses-sleap">
-<h1>Social LEAP Estimates Animal Poses (SLEAP)<a class="headerlink" href="#social-leap-estimates-animal-poses-sleap" title="Permalink to this headline">#</a></h1>
+<h1>Social LEAP Estimates Animal Poses (SLEAP)<a class="headerlink" href="#social-leap-estimates-animal-poses-sleap" title="Permalink to this heading">#</a></h1>
 <a class="reference internal image-reference" href="https://sleap.ai/docs/_static/sleap_movie.gif"><img alt="https://sleap.ai/docs/_static/sleap_movie.gif" src="https://sleap.ai/docs/_static/sleap_movie.gif" style="width: 600px;" /></a>
 <p><strong>SLEAP</strong> is an open source deep-learning based framework for multi-animal pose tracking <a class="reference external" href="https://www.nature.com/articles/s41592-022-01426-1">(Pereira et al., Nature Methods, 2022)</a>. It can be used to track any type or number of animals and includes an advanced labeling/training GUI for active learning and proofreading.</p>
 <section id="features">
-<h2>Features<a class="headerlink" href="#features" title="Permalink to this headline">#</a></h2>
+<h2>Features<a class="headerlink" href="#features" title="Permalink to this heading">#</a></h2>
 <ul class="simple">
 <li><p>Easy, one-line installation with support for all OSes</p></li>
 <li><p>Purpose-built GUI and human-in-the-loop workflow for rapidly labeling large datasets</p></li>
@@ -356,34 +354,34 @@ <h2>Features<a class="headerlink" href="#features" title="Permalink to this head
 </ul>
 </section>
 <section id="get-some-sleap">
-<h2>Get some SLEAP<a class="headerlink" href="#get-some-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Get some SLEAP<a class="headerlink" href="#get-some-sleap" title="Permalink to this heading">#</a></h2>
 <p>SLEAP is installed as a Python package. We strongly recommend using <a class="reference external" href="https://https://docs.conda.io/en/latest/miniconda.html">Miniconda</a> to install SLEAP in its own environment.</p>
 <p>You can find the latest version of SLEAP in the <a class="reference external" href="https://github.com/talmolab/sleap/releases">Releases</a> page.</p>
 <section id="quick-install">
-<h3>Quick install<a class="headerlink" href="#quick-install" title="Permalink to this headline">#</a></h3>
+<h3>Quick install<a class="headerlink" href="#quick-install" title="Permalink to this heading">#</a></h3>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">conda</span></code> <strong>(Windows/Linux/GPU)</strong>:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>-y<span class="w"> </span>-n<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>nvidia<span class="w"> </span>-c<span class="w"> </span>sleap/label/dev<span class="w"> </span>-c<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>anaconda<span class="w"> </span>sleap
 </pre></div>
 </div>
 <p><code class="xref py py-obj docutils literal notranslate"><span class="pre">pip</span></code> <strong>(any OS except Apple silicon)</strong>:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip install sleap<span class="o">[</span>pypi<span class="o">]</span>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>sleap<span class="o">[</span>pypi<span class="o">]</span>
 </pre></div>
 </div>
 <p>See the docs for <a class="reference external" href="https://sleap.ai/installation.html">full installation instructions</a>.</p>
 </section>
 </section>
 <section id="learn-to-sleap">
-<h2>Learn to SLEAP<a class="headerlink" href="#learn-to-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Learn to SLEAP<a class="headerlink" href="#learn-to-sleap" title="Permalink to this heading">#</a></h2>
 <ul class="simple">
 <li><p><strong>Learn step-by-step</strong>: <a class="reference external" href="https://sleap.ai/tutorials/tutorial.html">Tutorial</a></p></li>
 <li><p><strong>Learn more advanced usage</strong>: <a class="reference external" href="https://sleap.ai/guides/">Guides</a> and <a class="reference external" href="https://sleap.ai/notebooks/">Notebooks</a></p></li>
-<li><p><strong>Learn by watching</strong>: <a class="reference external" href="https://cbmm.mit.edu/video/decoding-animal-behavior-through-pose-tracking">MIT CBMM Tutorial</a></p></li>
+<li><p><strong>Learn by watching</strong>: <a class="reference external" href="https://www.youtube.com/watch?v=BfW-HgeDfMI">ABL:AOC 2023 Workshop</a> and <a class="reference external" href="https://cbmm.mit.edu/video/decoding-animal-behavior-through-pose-tracking">MIT CBMM Tutorial</a></p></li>
 <li><p><strong>Learn by reading</strong>: <a class="reference external" href="https://www.nature.com/articles/s41592-022-01426-1">Paper (Pereira et al., Nature Methods, 2022)</a> and <a class="reference external" href="https://rdcu.be/caH3H">Review on behavioral quantification (Pereira et al., Nature Neuroscience, 2020)</a></p></li>
 <li><p><strong>Learn from others</strong>: <a class="reference external" href="https://github.com/talmolab/sleap/discussions">Discussions on Github</a></p></li>
 </ul>
 </section>
 <section id="references">
-<h2>References<a class="headerlink" href="#references" title="Permalink to this headline">#</a></h2>
+<h2>References<a class="headerlink" href="#references" title="Permalink to this heading">#</a></h2>
 <p>SLEAP is the successor to the single-animal pose estimation software <a class="reference external" href="https://github.com/talmo/leap">LEAP</a> (<a class="reference external" href="https://www.nature.com/articles/s41592-018-0234-5">Pereira et al., Nature Methods, 2019</a>).</p>
 <p>If you use SLEAP in your research, please cite:</p>
 <blockquote>
@@ -424,7 +422,7 @@ <h2>References<a class="headerlink" href="#references" title="Permalink to this
 </div>
 </section>
 <section id="contact">
-<h2>Contact<a class="headerlink" href="#contact" title="Permalink to this headline">#</a></h2>
+<h2>Contact<a class="headerlink" href="#contact" title="Permalink to this heading">#</a></h2>
 <p>Follow <a class="reference external" href="https://twitter.com/talmop">&#64;talmop</a> on Twitter for news and updates!</p>
 <p><strong>Technical issue with the software?</strong></p>
 <ol class="arabic simple">
@@ -436,7 +434,7 @@ <h2>Contact<a class="headerlink" href="#contact" title="Permalink to this headli
 Reach out to <code class="xref py py-obj docutils literal notranslate"><span class="pre">talmo&#64;salk.edu</span></code>.</p>
 </section>
 <section id="contributors">
-<span id="id2"></span><h2>Contributors<a class="headerlink" href="#contributors" title="Permalink to this headline">#</a></h2>
+<span id="id2"></span><h2>Contributors<a class="headerlink" href="#contributors" title="Permalink to this heading">#</a></h2>
 <ul class="simple">
 <li><p><strong>Talmo Pereira</strong>, Salk Institute for Biological Studies</p></li>
 <li><p><strong>Liezl Maree</strong>, Salk Institute for Biological Studies</p></li>
@@ -456,11 +454,11 @@ <h2>Contact<a class="headerlink" href="#contact" title="Permalink to this headli
 </ul>
 </section>
 <section id="license">
-<h2>License<a class="headerlink" href="#license" title="Permalink to this headline">#</a></h2>
+<h2>License<a class="headerlink" href="#license" title="Permalink to this heading">#</a></h2>
 <p>SLEAP is released under a <a class="reference external" href="https://raw.githubusercontent.com/talmolab/sleap/main/LICENSE">Clear BSD License</a> and is intended for research/academic use only. For commercial use, please contact: Laurie Tzodikov (Assistant Director, Office of Technology Licensing), Princeton University, 609-258-7256.</p>
 </section>
 <section id="sleap-documentation">
-<h2>SLEAP Documentation<a class="headerlink" href="#sleap-documentation" title="Permalink to this headline">#</a></h2>
+<h2>SLEAP Documentation<a class="headerlink" href="#sleap-documentation" title="Permalink to this heading">#</a></h2>
 <div class="toctree-wrapper compound" id="sleap">
 <ul class="current">
 <li class="toctree-l1 current"><a class="current reference internal" href="#">Home</a></li>
@@ -552,7 +550,7 @@ <h2>SLEAP Documentation<a class="headerlink" href="#sleap-documentation" title="
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/installation.html b/installation.html
index 302beb752..4a242f453 100644
--- a/installation.html
+++ b/installation.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Installation &#8212; SLEAP (v1.3.3)</title>
+    <title>Installation &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,13 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
+    <script src="_static/tabs.js"></script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'installation';</script>
     <link rel="canonical" href="/develop/installation.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -59,7 +57,7 @@
     <link rel="next" title="Tutorial" href="tutorials/tutorial.html" />
     <link rel="prev" title="Overview" href="overview.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +163,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -317,14 +316,12 @@ <h2> Contents </h2>
             </div>
             <nav aria-label="Page">
                 <ul class="visible nav section-nav flex-column">
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#macs-pre-m1-pre-installation">Macs Pre-M1 (Pre-Installation)</a></li>
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installing-mambaforge">Installing Mambaforge</a></li>
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installation-methods">Installation methods</a><ul class="nav section-nav flex-column">
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#conda-package"><code class="docutils literal notranslate"><span class="pre">conda</span></code> package</a></li>
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#conda-from-source"><code class="docutils literal notranslate"><span class="pre">conda</span></code> from source</a></li>
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#pip-package"><code class="docutils literal notranslate"><span class="pre">pip</span></code> package</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#package-manager">Package Manager</a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#miniforge-recommended">Miniforge (recommended)</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#miniconda">Miniconda</a></li>
 </ul>
 </li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installation-methods">Installation methods</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#testing-that-things-are-working">Testing that things are working</a><ul class="nav section-nav flex-column">
 <li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#gui-support">GUI support</a></li>
 <li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#importing">Importing</a></li>
@@ -345,137 +342,123 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="installation">
-<h1>Installation<a class="headerlink" href="#installation" title="Permalink to this headline">#</a></h1>
-<p>SLEAP can be installed as a Python package on Windows, Linux, Mac OS X, and Mac OS Apple Silicon.</p>
-<p>SLEAP requires many complex dependencies, so we <strong>strongly</strong> recommend using <a class="reference external" href="https://mamba.readthedocs.io/en/latest/installation.html">Mambaforge</a> to install it in its own isolated environment. See <a class="reference internal" href="#mambaforge"><span class="std std-ref">Installing Mambaforge</span></a> below for more instructions.</p>
-<p>The newest version of SLEAP can always be found in the <a class="reference external" href="https://github.com/talmolab/sleap/releases">Releases page</a>.</p>
-<div class="contents local topic" id="contents">
+<h1>Installation<a class="headerlink" href="#installation" title="Permalink to this heading">#</a></h1>
+<p>SLEAP can be installed as a Python package on Windows, Linux, and Mac OS. For quick install using conda, see below:</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-0-V2luZG93cyBhbmQgTGludXg=" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-0-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tab" tabindex="0">Windows and Linux</button><button aria-controls="panel-0-TWFjIE9T" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-0-TWFjIE9T" name="TWFjIE9T" role="tab" tabindex="-1">Mac OS</button></div><div aria-labelledby="tab-0-V2luZG93cyBhbmQgTGludXg=" class="sphinx-tabs-panel group-tab" id="panel-0-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>-y<span class="w"> </span>-n<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>nvidia<span class="w"> </span>-c<span class="w"> </span>sleap/label/dev<span class="w"> </span>-c<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>anaconda<span class="w"> </span><span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.4.1
+</pre></div>
+</div>
+</div><div aria-labelledby="tab-0-TWFjIE9T" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-0-TWFjIE9T" name="TWFjIE9T" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>-y<span class="w"> </span>-n<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>anaconda<span class="w"> </span>-c<span class="w"> </span>sleap<span class="w"> </span><span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.4.1
+</pre></div>
+</div>
+</div></div>
+<p>. For more in-depth installation instructions, see the <a class="reference internal" href="#installation-methods"><span class="std std-ref">installation methods</span></a>. The newest version of SLEAP can always be found in the <a class="reference external" href="https://github.com/talmolab/sleap/releases">Releases page</a>.</p>
+<nav class="contents local" id="contents">
 <p class="topic-title">Contents</p>
 <ul class="simple">
-<li><p><a class="reference internal" href="#macs-pre-m1-pre-installation" id="id1">Macs Pre-M1 (Pre-Installation)</a></p></li>
-<li><p><a class="reference internal" href="#installing-mambaforge" id="id2">Installing Mambaforge</a></p></li>
-<li><p><a class="reference internal" href="#installation-methods" id="id3">Installation methods</a></p>
+<li><p><a class="reference internal" href="#package-manager" id="id2">Package Manager</a></p>
 <ul>
-<li><p><a class="reference internal" href="#conda-package" id="id4"><code class="docutils literal notranslate"><span class="pre">conda</span></code> package</a></p></li>
-<li><p><a class="reference internal" href="#conda-from-source" id="id5"><code class="docutils literal notranslate"><span class="pre">conda</span></code> from source</a></p></li>
-<li><p><a class="reference internal" href="#pip-package" id="id6"><code class="docutils literal notranslate"><span class="pre">pip</span></code> package</a></p></li>
+<li><p><a class="reference internal" href="#miniforge-recommended" id="id3">Miniforge (recommended)</a></p></li>
+<li><p><a class="reference internal" href="#miniconda" id="id4">Miniconda</a></p></li>
 </ul>
 </li>
-<li><p><a class="reference internal" href="#testing-that-things-are-working" id="id7">Testing that things are working</a></p>
+<li><p><a class="reference internal" href="#installation-methods" id="id5">Installation methods</a></p></li>
+<li><p><a class="reference internal" href="#testing-that-things-are-working" id="id6">Testing that things are working</a></p>
 <ul>
-<li><p><a class="reference internal" href="#gui-support" id="id8">GUI support</a></p></li>
-<li><p><a class="reference internal" href="#importing" id="id9">Importing</a></p></li>
-<li><p><a class="reference internal" href="#gpu-support" id="id10">GPU support</a></p></li>
+<li><p><a class="reference internal" href="#gui-support" id="id7">GUI support</a></p></li>
+<li><p><a class="reference internal" href="#importing" id="id8">Importing</a></p></li>
+<li><p><a class="reference internal" href="#gpu-support" id="id9">GPU support</a></p></li>
 </ul>
 </li>
-<li><p><a class="reference internal" href="#upgrading-and-uninstalling" id="id11">Upgrading and uninstalling</a></p></li>
-<li><p><a class="reference internal" href="#getting-help" id="id12">Getting help</a></p></li>
+<li><p><a class="reference internal" href="#upgrading-and-uninstalling" id="id10">Upgrading and uninstalling</a></p></li>
+<li><p><a class="reference internal" href="#getting-help" id="id11">Getting help</a></p></li>
 </ul>
-</div>
+</nav>
 <div class="admonition hint">
 <p class="admonition-title">Hint</p>
 <p>Installation requires entering commands in a terminal. To open one:</p>
-<p><strong>Windows:</strong> Open the <em>Start menu</em> and search for the <em>Miniforge Prompt</em> (if using Mambaforge) or the <em>Command Prompt</em> if not.</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-1-1-0" aria-selected="true" class="sphinx-tabs-tab" id="tab-1-1-0" name="1-0" role="tab" tabindex="0">Windows</button><button aria-controls="panel-1-1-1" aria-selected="false" class="sphinx-tabs-tab" id="tab-1-1-1" name="1-1" role="tab" tabindex="-1">Linux</button><button aria-controls="panel-1-TWFjIE9T" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-1-TWFjIE9T" name="TWFjIE9T" role="tab" tabindex="-1">Mac OS</button></div><div aria-labelledby="tab-1-1-0" class="sphinx-tabs-panel" id="panel-1-1-0" name="1-0" role="tabpanel" tabindex="0"><p>Open the <em>Start menu</em> and search for the <em>Anaconda Prompt</em> (if using Miniconda) or the <em>Command Prompt</em> if not.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p>On Windows, our personal preference is to use alternative terminal apps like <a class="reference external" href="https://cmder.net">Cmder</a> or <a class="reference external" href="https://aka.ms/terminal">Windows Terminal</a>.</p>
 </div>
-<p><strong>Linux:</strong> Launch a new terminal by pressing <kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>.</p>
-<p><strong>Mac:</strong> Launch a new terminal by pressing <kbd>Cmd</kbd> + <kbd>Space</kbd> and searching for <em>Terminal</em>.</p>
-</div>
-<section id="macs-pre-m1-pre-installation">
-<h2><a class="toc-backref" href="#id1">Macs Pre-M1 (Pre-Installation)</a><a class="headerlink" href="#macs-pre-m1-pre-installation" title="Permalink to this headline">#</a></h2>
-<p>SLEAP can be installed on Macs by following these instructions:</p>
-<ol>
-<li><p>Make sure you’re on <strong>macOS Monterey</strong> or later, i.e., version 12+.</p></li>
-<li><p>If you don’t have it yet, <a class="reference external" href="https://brew.sh/">install <strong>homebrew</strong></a>, a convenient package manager for Macs (skip this if you can run <code class="docutils literal notranslate"><span class="pre">brew</span></code> from the terminal):</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>/bin/bash -c <span class="s2">&quot;</span><span class="k">$(</span>curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh<span class="k">)</span><span class="s2">&quot;</span>
-</pre></div>
+</div><div aria-labelledby="tab-1-1-1" class="sphinx-tabs-panel" hidden="true" id="panel-1-1-1" name="1-1" role="tabpanel" tabindex="0"><p>Launch a new terminal by pressing <kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>.</p>
+</div><div aria-labelledby="tab-1-TWFjIE9T" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-1-TWFjIE9T" name="TWFjIE9T" role="tabpanel" tabindex="0"><p>Launch a new terminal by pressing <kbd>Cmd</kbd> + <kbd>Space</kbd> and searching for <em>Terminal</em>.</p>
+</div></div>
 </div>
-<p>This might take a little while since it’ll also install Xcode (which we’ll need later). Once it’s finished, your terminal should give you two extra commands to run listed under <strong>Next Steps</strong>.</p>
+<section id="package-manager">
+<h2><a class="toc-backref" href="#id2" role="doc-backlink">Package Manager</a><a class="headerlink" href="#package-manager" title="Permalink to this heading">#</a></h2>
+<p>SLEAP requires many complex dependencies, so we <strong>strongly</strong> recommend using a package manager such as <a class="reference external" href="https://github.com/conda-forge/miniforge">Miniforge</a> or <a class="reference external" href="https://docs.anaconda.com/free/miniconda/">Miniconda</a> to install SLEAP in its own isolated environment.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
-<p>We recommend running the commands given in your terminal which will be similar to (but may differ slightly) from the commands below:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span><span class="nb">echo</span> <span class="s1">&#39;eval &quot;$(/opt/homebrew/bin/brew shellenv)&quot;&#39;</span> &gt;&gt; ~/.zprofile
+<p>If you already have Anaconda on your computer (and it is an <a class="reference external" href="https://conda.org/blog/2023-11-06-conda-23-10-0-release/">older installation</a>), then make sure to <a class="reference external" href="https://www.anaconda.com/blog/a-faster-conda-for-a-growing-community">set the solver to <code class="docutils literal notranslate"><span class="pre">libmamba</span></code></a> in the <code class="docutils literal notranslate"><span class="pre">base</span></code> environment.</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>update<span class="w"> </span>-n<span class="w"> </span>base<span class="w"> </span>conda
+conda<span class="w"> </span>install<span class="w"> </span>-n<span class="w"> </span>base<span class="w"> </span>conda-libmamba-solver
+conda<span class="w"> </span>config<span class="w"> </span>--set<span class="w"> </span>solver<span class="w"> </span>libmamba
 </pre></div>
 </div>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span><span class="nb">eval</span> <span class="s2">&quot;</span><span class="k">$(</span>/opt/homebrew/bin/brew shellenv<span class="k">)</span><span class="s2">&quot;</span>
-</pre></div>
+<div class="admonition warning">
+<p class="admonition-title">Warning</p>
+<p>Any subsequent <code class="docutils literal notranslate"><span class="pre">conda</span></code> commands in the docs will need to be replaced with <code class="docutils literal notranslate"><span class="pre">mamba</span></code> if you have <a class="reference external" href="https://mamba.readthedocs.io/en/latest/">Mamba</a> installed instead of Anaconda or Miniconda.</p>
 </div>
 </div>
-<p>Then, close and re-open the terminal for it to take effect.</p>
-</li>
-<li><p>Install wget, a CLI downloading utility (also makes sure your homebrew setup worked):</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>brew install wget
+<p>If you don’t have a <code class="docutils literal notranslate"><span class="pre">conda</span></code> package manager installation, here are some quick install options:</p>
+<section id="miniforge-recommended">
+<h3><a class="toc-backref" href="#id3" role="doc-backlink">Miniforge (recommended)</a><a class="headerlink" href="#miniforge-recommended" title="Permalink to this heading">#</a></h3>
+<p>Miniforge is a minimal installer for conda that includes the <code class="docutils literal notranslate"><span class="pre">conda</span></code> package manager and is maintained by the <a class="reference external" href="https://conda-forge.org">conda-forge</a> community. The only difference between Miniforge and Miniconda is that Miniforge uses the <code class="docutils literal notranslate"><span class="pre">conda-forge</span></code> channel by default, which provides a much wider selection of community-maintained packages.</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-2-V2luZG93cw==" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-2-V2luZG93cw==" name="V2luZG93cw==" role="tab" tabindex="0">Windows</button><button aria-controls="panel-2-TGludXg=" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-2-TGludXg=" name="TGludXg=" role="tab" tabindex="-1">Linux</button><button aria-controls="panel-2-TWFjIChBcHBsZSBTaWxpY29uKQ==" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-2-TWFjIChBcHBsZSBTaWxpY29uKQ==" name="TWFjIChBcHBsZSBTaWxpY29uKQ==" role="tab" tabindex="-1">Mac (Apple Silicon)</button><button aria-controls="panel-2-TWFjIChJbnRlbCk=" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-2-TWFjIChJbnRlbCk=" name="TWFjIChJbnRlbCk=" role="tab" tabindex="-1">Mac (Intel)</button></div><div aria-labelledby="tab-2-V2luZG93cw==" class="sphinx-tabs-panel group-tab" id="panel-2-V2luZG93cw==" name="V2luZG93cw==" role="tabpanel" tabindex="0"><p>Open a new PowerShell terminal (does not need to be admin) and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>Invoke-WebRequest<span class="w"> </span>-Uri<span class="w"> </span><span class="s2">&quot;https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Windows-x86_64.exe&quot;</span><span class="w"> </span>-OutFile<span class="w"> </span><span class="s2">&quot;</span><span class="nv">$env</span><span class="s2">:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe&quot;</span><span class="p">;</span><span class="w"> </span>Start-Process<span class="w"> </span>-FilePath<span class="w"> </span><span class="s2">&quot;</span><span class="nv">$env</span><span class="s2">:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe&quot;</span><span class="w"> </span>-ArgumentList<span class="w"> </span><span class="s2">&quot;/InstallationType=JustMe /RegisterPython=1 /S&quot;</span><span class="w"> </span>-Wait<span class="p">;</span><span class="w"> </span>Remove-Item<span class="w"> </span>-Path<span class="w"> </span><span class="s2">&quot;</span><span class="nv">$env</span><span class="s2">:UserProfile/Downloads/Miniforge3-Windows-x86_64.exe&quot;</span>
 </pre></div>
 </div>
-</li>
-</ol>
-</section>
-<section id="installing-mambaforge">
-<span id="mambaforge"></span><h2><a class="toc-backref" href="#id2">Installing Mambaforge</a><a class="headerlink" href="#installing-mambaforge" title="Permalink to this headline">#</a></h2>
-<p><strong>Anaconda</strong> is a Python environment manager that makes it easy to install SLEAP and its necessary dependencies without affecting other Python software on your computer.</p>
-<p><a class="reference external" href="https://mamba.readthedocs.io/en/latest/installation.html"><strong>Mambaforge</strong></a> is a lightweight installer of Anaconda with speedy package resolution that we recommend.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>If you already have Anaconda on your computer, then you can <a class="reference external" href="https://www.anaconda.com/blog/a-faster-conda-for-a-growing-community">set the solver to <code class="docutils literal notranslate"><span class="pre">libmamba</span></code></a> in the <code class="docutils literal notranslate"><span class="pre">base</span></code> environment (and skip the Mambaforge installation):</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda update -n base conda
-conda install -n base conda-libmamba-solver
-conda config --set solver libmamba
+</div><div aria-labelledby="tab-2-TGludXg=" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-2-TGludXg=" name="TGludXg=" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>-fsSL<span class="w"> </span>--compressed<span class="w"> </span>https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh<span class="w"> </span>-o<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-Linux-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>chmod<span class="w"> </span>+x<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-Linux-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-Linux-x86_64.sh&quot;</span><span class="w"> </span>-b<span class="w"> </span>-p<span class="w"> </span>~/miniforge3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-Linux-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniforge3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
 </pre></div>
 </div>
-<div class="admonition warning">
-<p class="admonition-title">Warning</p>
-<p>Any subsequent <code class="docutils literal notranslate"><span class="pre">mamba</span></code> commands in the docs will need to be replaced with <code class="docutils literal notranslate"><span class="pre">conda</span></code> if you choose to use your existing Anaconda installation.</p>
+</div><div aria-labelledby="tab-2-TWFjIChBcHBsZSBTaWxpY29uKQ==" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-2-TWFjIChBcHBsZSBTaWxpY29uKQ==" name="TWFjIChBcHBsZSBTaWxpY29uKQ==" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>-fsSL<span class="w"> </span>--compressed<span class="w"> </span>https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh<span class="w"> </span>-o<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>chmod<span class="w"> </span>+x<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-arm64.sh&quot;</span><span class="w"> </span>-b<span class="w"> </span>-p<span class="w"> </span>~/miniforge3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniforge3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
+</pre></div>
 </div>
+</div><div aria-labelledby="tab-2-TWFjIChJbnRlbCk=" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-2-TWFjIChJbnRlbCk=" name="TWFjIChJbnRlbCk=" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>-fsSL<span class="w"> </span>--compressed<span class="w"> </span>https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-x86_64.sh<span class="w"> </span>-o<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>chmod<span class="w"> </span>+x<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-x86_64.sh&quot;</span><span class="w"> </span>-b<span class="w"> </span>-p<span class="w"> </span>~/miniforge3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniforge3-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniforge3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
+</pre></div>
 </div>
-<p>Otherwise, to install Mamba:</p>
-<p><strong>On Windows</strong>, just click through the installation steps.</p>
-<ol class="simple">
-<li><p>Go to: <a class="github reference external" href="https://github.com/conda-forge/miniforge#mambaforge">conda-forge/miniforge</a></p></li>
-<li><p>Download the latest version for your OS.</p></li>
-<li><p>Follow the installer instructions.</p></li>
-</ol>
-<p>We recommend using the following settings:</p>
-<ul class="simple">
-<li><p>Install for: All Users (requires admin privileges)</p></li>
-<li><p>Destination folder: <code class="docutils literal notranslate"><span class="pre">C:\mambaforge</span></code></p></li>
-<li><p>Advanced Options: Add MambaForge to the system PATH environment variable</p></li>
-<li><p>Advanced Options: Register MambaForge as the system Python 3.X
-These will make sure that MambaForge is easily accessible from most places on your computer.</p></li>
-</ul>
-<p><strong>On Linux</strong>, it might be easier to do this straight from the terminal (<kbd>Ctrl</kbd> + <kbd>Alt</kbd> + <kbd>T</kbd>) with this one-liner:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>wget -nc https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-Linux-x86_64.sh <span class="o">&amp;&amp;</span> bash Mambaforge-Linux-x86_64.sh -b <span class="o">&amp;&amp;</span> ~/mambaforge/bin/conda init bash
+</div></div>
+</section>
+<section id="miniconda">
+<h3><a class="toc-backref" href="#id4" role="doc-backlink">Miniconda</a><a class="headerlink" href="#miniconda" title="Permalink to this heading">#</a></h3>
+<p>This is a minimal installer for conda that includes the <code class="docutils literal notranslate"><span class="pre">conda</span></code> package manager and is maintained by the <a class="reference external" href="https://www.anaconda.com">Anaconda</a> company.</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-3-V2luZG93cw==" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-3-V2luZG93cw==" name="V2luZG93cw==" role="tab" tabindex="0">Windows</button><button aria-controls="panel-3-TGludXg=" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-3-TGludXg=" name="TGludXg=" role="tab" tabindex="-1">Linux</button><button aria-controls="panel-3-TWFjIChBcHBsZSBTaWxpY29uKQ==" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-3-TWFjIChBcHBsZSBTaWxpY29uKQ==" name="TWFjIChBcHBsZSBTaWxpY29uKQ==" role="tab" tabindex="-1">Mac (Apple Silicon)</button><button aria-controls="panel-3-TWFjIChJbnRlbCk=" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-3-TWFjIChJbnRlbCk=" name="TWFjIChJbnRlbCk=" role="tab" tabindex="-1">Mac (Intel)</button></div><div aria-labelledby="tab-3-V2luZG93cw==" class="sphinx-tabs-panel group-tab" id="panel-3-V2luZG93cw==" name="V2luZG93cw==" role="tabpanel" tabindex="0"><p>Open a new PowerShell terminal (does not need to be admin) and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>https://repo.anaconda.com/miniconda/Miniconda3-latest-Windows-x86_64.exe<span class="w"> </span>-o<span class="w"> </span>miniconda.exe<span class="p">;</span><span class="w"> </span>Start-Process<span class="w"> </span>-FilePath<span class="w"> </span><span class="s2">&quot;.\miniconda.exe&quot;</span><span class="w"> </span>-ArgumentList<span class="w"> </span><span class="s2">&quot;/S&quot;</span><span class="w"> </span>-Wait<span class="p">;</span><span class="w"> </span>del<span class="w"> </span>miniconda.exe
 </pre></div>
 </div>
-<p>Restart the terminal after running this command.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>For other Linux architectures (arm64 and POWER8/9), replace the <code class="docutils literal notranslate"><span class="pre">.sh</span></code> filenames above with the correct installer name for your architecture. See the Download column in <a class="reference external" href="https://github.com/conda-forge/miniforge#mambaforge">this table</a> for the correct filename.</p>
+</div><div aria-labelledby="tab-3-TGludXg=" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-3-TGludXg=" name="TGludXg=" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mkdir<span class="w"> </span>-p<span class="w"> </span>~/miniconda3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>wget<span class="w"> </span>https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh<span class="w"> </span>-O<span class="w"> </span>~/miniconda3/miniconda.sh<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>bash<span class="w"> </span>~/miniconda3/miniconda.sh<span class="w"> </span>-b<span class="w"> </span>-u<span class="w"> </span>-p<span class="w"> </span>~/miniconda3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span>~/miniconda3/miniconda.sh<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniconda3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
+</pre></div>
 </div>
-<p><strong>On Macs (pre-M1)</strong>, you can run the installer using this terminal command:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>wget -nc https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-MacOSX-x86_64.sh <span class="o">&amp;&amp;</span> bash Mambaforge-MacOSX-x86_64.sh -b <span class="o">&amp;&amp;</span> ~/mambaforge/bin/conda init zsh
+</div><div aria-labelledby="tab-3-TWFjIChBcHBsZSBTaWxpY29uKQ==" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-3-TWFjIChBcHBsZSBTaWxpY29uKQ==" name="TWFjIChBcHBsZSBTaWxpY29uKQ==" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>-fsSL<span class="w"> </span>--compressed<span class="w"> </span>https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-arm64.sh<span class="w"> </span>-o<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>chmod<span class="w"> </span>+x<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-arm64.sh&quot;</span><span class="w"> </span>-b<span class="w"> </span>-u<span class="w"> </span>-p<span class="w"> </span>~/miniconda3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-arm64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniconda3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
 </pre></div>
 </div>
-<p><strong>On Macs (Apple Silicon)</strong>, use this terminal command:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl -fsSL --compressed https://github.com/conda-forge/miniforge/releases/latest/download/Mambaforge-MacOSX-arm64.sh -o Mambaforge3-MacOSX-arm64.sh <span class="o">&amp;&amp;</span> chmod +x Mambaforge3-MacOSX-arm64.sh <span class="o">&amp;&amp;</span> ./Mambaforge3-MacOSX-arm64.sh -b -p ~/mambaforge3 <span class="o">&amp;&amp;</span> rm Mambaforge3-MacOSX-arm64.sh <span class="o">&amp;&amp;</span> ~/mambaforge3/bin/conda init <span class="s2">&quot;</span><span class="k">$(</span>basename <span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span> <span class="o">&amp;&amp;</span> <span class="nb">source</span> <span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename <span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
+</div><div aria-labelledby="tab-3-TWFjIChJbnRlbCk=" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-3-TWFjIChJbnRlbCk=" name="TWFjIChJbnRlbCk=" role="tabpanel" tabindex="0"><p>Open a new terminal and enter:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>curl<span class="w"> </span>-fsSL<span class="w"> </span>--compressed<span class="w"> </span>https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh<span class="w"> </span>-o<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>chmod<span class="w"> </span>+x<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh&quot;</span><span class="w"> </span>-b<span class="w"> </span>-u<span class="w"> </span>-p<span class="w"> </span>~/miniconda3<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>rm<span class="w"> </span><span class="s2">&quot;~/Downloads/Miniconda3-latest-MacOSX-x86_64.sh&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span>~/miniconda3/bin/conda<span class="w"> </span>init<span class="w"> </span><span class="s2">&quot;</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">&quot;</span><span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">source</span><span class="w"> </span><span class="s2">&quot;</span><span class="nv">$HOME</span><span class="s2">/.</span><span class="k">$(</span>basename<span class="w"> </span><span class="s2">&quot;</span><span class="si">${</span><span class="nv">SHELL</span><span class="si">}</span><span class="s2">&quot;</span><span class="k">)</span><span class="s2">rc&quot;</span>
 </pre></div>
 </div>
+</div></div>
+<p>See the <a class="reference external" href="https://docs.anaconda.com/free/miniconda/">Miniconda website</a> for up-to-date installation instructions if the above instructions don’t work for your system.</p>
+</section>
 </section>
 <section id="installation-methods">
-<h2><a class="toc-backref" href="#id3">Installation methods</a><a class="headerlink" href="#installation-methods" title="Permalink to this headline">#</a></h2>
-<p>SLEAP can be installed three different ways: via <a class="reference internal" href="#condapackage"><span class="std std-ref">conda package</span></a>, <a class="reference internal" href="#condasource"><span class="std std-ref">conda from source</span></a>, or <a class="reference internal" href="#pippackage"><span class="std std-ref">pip package</span></a>. Select one of the methods below to install SLEAP. We recommend <a class="reference internal" href="#condapackage"><span class="std std-ref">conda package</span></a>.</p>
-<section id="conda-package">
-<span id="condapackage"></span><h3><a class="toc-backref" href="#id4"><code class="docutils literal notranslate"><span class="pre">conda</span></code> package</a><a class="headerlink" href="#conda-package" title="Permalink to this headline">#</a></h3>
-<p><strong>Windows</strong> and <strong>Linux</strong></p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda <span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.3.3
+<span id="id1"></span><h2><a class="toc-backref" href="#id5" role="doc-backlink">Installation methods</a><a class="headerlink" href="#installation-methods" title="Permalink to this heading">#</a></h2>
+<p>SLEAP can be installed three different ways: via <span class="xref std std-ref">conda package</span>, <span class="xref std std-ref">conda from source</span>, or <span class="xref std std-ref">pip package</span>. Select one of the methods below to install SLEAP. We recommend <span class="xref std std-ref">conda package</span>.</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-4-4-0" aria-selected="true" class="sphinx-tabs-tab" id="tab-4-4-0" name="4-0" role="tab" tabindex="0">conda package</button><button aria-controls="panel-4-4-1" aria-selected="false" class="sphinx-tabs-tab" id="tab-4-4-1" name="4-1" role="tab" tabindex="-1">conda from source</button><button aria-controls="panel-4-4-2" aria-selected="false" class="sphinx-tabs-tab" id="tab-4-4-2" name="4-2" role="tab" tabindex="-1">pip package</button></div><div aria-labelledby="tab-4-4-0" class="sphinx-tabs-panel" id="panel-4-4-0" name="4-0" role="tabpanel" tabindex="0"><p><strong>This is the recommended installation method</strong>.</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-5-V2luZG93cyBhbmQgTGludXg=" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-5-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tab" tabindex="0">Windows and Linux</button><button aria-controls="panel-5-TWFjIE9T" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-5-TWFjIE9T" name="TWFjIE9T" role="tab" tabindex="-1">Mac OS</button></div><div aria-labelledby="tab-5-V2luZG93cyBhbmQgTGludXg=" class="sphinx-tabs-panel group-tab" id="panel-5-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>-y<span class="w"> </span>-n<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>nvidia<span class="w"> </span>-c<span class="w"> </span>sleap/label/dev<span class="w"> </span>-c<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>anaconda<span class="w"> </span><span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.4.1
 </pre></div>
 </div>
-<p><strong>Mac OS X</strong> and <strong>Apple Silicon</strong></p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba create -y -n sleap -c conda-forge -c anaconda -c sleap <span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.3.3
-</pre></div>
-</div>
-<p><strong>This is the recommended installation method</strong>.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <ul class="simple">
@@ -484,35 +467,41 @@ <h2><a class="toc-backref" href="#id3">Installation methods</a><a class="headerl
 <li><p>This will also work in CPU mode if you don’t have a GPU on your machine.</p></li>
 </ul>
 </div>
-</section>
-<section id="conda-from-source">
-<span id="condasource"></span><h3><a class="toc-backref" href="#id5"><code class="docutils literal notranslate"><span class="pre">conda</span></code> from source</a><a class="headerlink" href="#conda-from-source" title="Permalink to this headline">#</a></h3>
-<ol>
+</div><div aria-labelledby="tab-5-TWFjIE9T" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-5-TWFjIE9T" name="TWFjIE9T" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>-y<span class="w"> </span>-n<span class="w"> </span>sleap<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>anaconda<span class="w"> </span>-c<span class="w"> </span>sleap<span class="w"> </span><span class="nv">sleap</span><span class="o">=</span><span class="m">1</span>.4.1
+</pre></div>
+</div>
+<div class="admonition note">
+<p class="admonition-title">Note</p>
+<p>This will also work in CPU mode if you don’t have a GPU on your machine.</p>
+</div>
+</div></div>
+</div><div aria-labelledby="tab-4-4-1" class="sphinx-tabs-panel" hidden="true" id="panel-4-4-1" name="4-1" role="tabpanel" tabindex="0"><p>This is the <strong>recommended method for development</strong>.</p>
+<ol class="arabic">
 <li><p>First, ensure git is installed:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>git --version
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>git<span class="w"> </span>--version
 </pre></div>
 </div>
-<p>If ‘git’ is not recognized, then <a class="reference external" href="https://git-scm.com/book/en/v2/Getting-Started-Installing-Git">install git</a>.</p>
+<p>If <code class="docutils literal notranslate"><span class="pre">git</span></code> is not recognized, then <a class="reference external" href="https://git-scm.com/book/en/v2/Getting-Started-Installing-Git">install git</a>.</p>
 </li>
 <li><p>Then, clone the repository:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>git clone https://github.com/talmolab/sleap <span class="o">&amp;&amp;</span> <span class="nb">cd</span> sleap
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>git<span class="w"> </span>clone<span class="w"> </span>https://github.com/talmolab/sleap<span class="w"> </span><span class="o">&amp;&amp;</span><span class="w"> </span><span class="nb">cd</span><span class="w"> </span>sleap
 </pre></div>
 </div>
 </li>
-<li><p>Finally, install from the environment file (differs based on OS and GPU):</p>
-<p><strong>Windows</strong> and <strong>Linux</strong></p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba env create -f environment.yml -n sleap
+<li><p>Finally, install SLEAP from the environment file:</p>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-6-V2luZG93cyBhbmQgTGludXg=" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-6-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tab" tabindex="0">Windows and Linux</button><button aria-controls="panel-6-TWFjIE9T" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-6-TWFjIE9T" name="TWFjIE9T" role="tab" tabindex="-1">Mac OS</button></div><div aria-labelledby="tab-6-V2luZG93cyBhbmQgTGludXg=" class="sphinx-tabs-panel group-tab" id="panel-6-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tabpanel" tabindex="0"><div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-7-TlZJRElBIEdQVQ==" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-7-TlZJRElBIEdQVQ==" name="TlZJRElBIEdQVQ==" role="tab" tabindex="0">NVIDIA GPU</button><button aria-controls="panel-7-Q1BVIG9yIG90aGVyIEdQVQ==" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-7-Q1BVIG9yIG90aGVyIEdQVQ==" name="Q1BVIG9yIG90aGVyIEdQVQ==" role="tab" tabindex="-1">CPU or other GPU</button></div><div aria-labelledby="tab-7-TlZJRElBIEdQVQ==" class="sphinx-tabs-panel group-tab" id="panel-7-TlZJRElBIEdQVQ==" name="TlZJRElBIEdQVQ==" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>env<span class="w"> </span>create<span class="w"> </span>-f<span class="w"> </span>environment.yml<span class="w"> </span>-n<span class="w"> </span>sleap
 </pre></div>
 </div>
-<p>If you do not have a NVIDIA GPU, then you should use the no CUDA environment file:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba env create -f environment_no_cuda.yml -n sleap
+</div><div aria-labelledby="tab-7-Q1BVIG9yIG90aGVyIEdQVQ==" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-7-Q1BVIG9yIG90aGVyIEdQVQ==" name="Q1BVIG9yIG90aGVyIEdQVQ==" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>env<span class="w"> </span>create<span class="w"> </span>-f<span class="w"> </span>environment_no_cuda.yml<span class="w"> </span>-n<span class="w"> </span>sleap
 </pre></div>
 </div>
-<p><strong>Mac OS X</strong> and <strong>Apple Silicon</strong></p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba env create -f environment_mac.yml -n sleap
+</div></div>
+</div><div aria-labelledby="tab-6-TWFjIE9T" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-6-TWFjIE9T" name="TWFjIE9T" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>env<span class="w"> </span>create<span class="w"> </span>-f<span class="w"> </span>environment_mac.yml<span class="w"> </span>-n<span class="w"> </span>sleap
 </pre></div>
 </div>
-<p>This is the <strong>recommended method for development</strong>.</p>
+</div></div>
 </li>
 </ol>
 <div class="admonition note">
@@ -522,23 +511,34 @@ <h2><a class="toc-backref" href="#id3">Installation methods</a><a class="headerl
 <li><p>Change the <code class="docutils literal notranslate"><span class="pre">-n</span> <span class="pre">sleap</span></code> in the command to create an environment with a different name (e.g., <code class="docutils literal notranslate"><span class="pre">-n</span> <span class="pre">sleap_develop</span></code>).</p></li>
 </ul>
 </div>
-</section>
-<section id="pip-package">
-<span id="pippackage"></span><h3><a class="toc-backref" href="#id6"><code class="docutils literal notranslate"><span class="pre">pip</span></code> package</a><a class="headerlink" href="#pip-package" title="Permalink to this headline">#</a></h3>
-<p>Although you do not need Mambaforge installed to perform a <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code>, we recommend <a class="reference internal" href="#mambaforge"><span class="std std-ref">installing Mambaforge</span></a> to create a new environment where we can isolate the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code>. Alternatively, you can use a venv if you have an existing python installation. If you are working on <strong>Google Colab</strong>, skip to step 3 to perform the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code> without using a conda environment.</p>
-<ol>
+</div><div aria-labelledby="tab-4-4-2" class="sphinx-tabs-panel" hidden="true" id="panel-4-4-2" name="4-2" role="tabpanel" tabindex="0"><p>This is the <strong>recommended method for Google Colab only</strong>.</p>
+<div class="admonition warning">
+<p class="admonition-title">Warning</p>
+<p>This will uninstall existing libraries and potentially install conflicting ones.</p>
+<p>We strongly recommend that you <strong>only use this method if you know what you’re doing</strong>!</p>
+</div>
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-8-V2luZG93cyBhbmQgTGludXg=" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-8-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tab" tabindex="0">Windows and Linux</button><button aria-controls="panel-8-TWFjIE9T" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-8-TWFjIE9T" name="TWFjIE9T" role="tab" tabindex="-1">Mac OS</button></div><div aria-labelledby="tab-8-V2luZG93cyBhbmQgTGludXg=" class="sphinx-tabs-panel group-tab" id="panel-8-V2luZG93cyBhbmQgTGludXg=" name="V2luZG93cyBhbmQgTGludXg=" role="tabpanel" tabindex="0"><div class="admonition note">
+<p class="admonition-title">Note</p>
+<ul class="simple">
+<li><p>Requires Python 3.7</p></li>
+<li><p>To enable GPU support, make sure that you have <strong>CUDA Toolkit v11.3</strong> and <strong>cuDNN v8.2</strong> installed.</p></li>
+</ul>
+</div>
+<p>Although you do not need Miniconda installed to perform a <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code>, we recommend <a class="reference external" href="https://docs.anaconda.com/free/miniconda/">installing Miniconda</a> to create a new environment where we can isolate the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code>. Alternatively, you can use a venv if you have an existing Python 3.7 installation. If you are working on <strong>Google Colab</strong>, skip to step 3 to perform the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code> without using a conda environment.</p>
+<ol class="arabic">
 <li><p>Otherwise, create a new conda environment where we will <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span> <span class="pre">sleap</span></code>:</p>
-<p>either without GPU support:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba create --name sleap pip <span class="nv">python</span><span class="o">=</span><span class="m">3</span>.7.12
+<div class="sphinx-tabs docutils container">
+<div aria-label="Tabbed content" class="closeable" role="tablist"><button aria-controls="panel-9-TlZJRElBIEdQVQ==" aria-selected="true" class="sphinx-tabs-tab group-tab" id="tab-9-TlZJRElBIEdQVQ==" name="TlZJRElBIEdQVQ==" role="tab" tabindex="0">NVIDIA GPU</button><button aria-controls="panel-9-Q1BVIG9yIG90aGVyIEdQVQ==" aria-selected="false" class="sphinx-tabs-tab group-tab" id="tab-9-Q1BVIG9yIG90aGVyIEdQVQ==" name="Q1BVIG9yIG90aGVyIEdQVQ==" role="tab" tabindex="-1">CPU or other GPU</button></div><div aria-labelledby="tab-9-TlZJRElBIEdQVQ==" class="sphinx-tabs-panel group-tab" id="panel-9-TlZJRElBIEdQVQ==" name="TlZJRElBIEdQVQ==" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>--name<span class="w"> </span>sleap<span class="w"> </span>pip<span class="w"> </span><span class="nv">python</span><span class="o">=</span><span class="m">3</span>.7.12<span class="w"> </span><span class="nv">cudatoolkit</span><span class="o">=</span><span class="m">11</span>.3<span class="w"> </span><span class="nv">cudnn</span><span class="o">=</span><span class="m">8</span>.2<span class="w"> </span>-c<span class="w"> </span>conda-forge<span class="w"> </span>-c<span class="w"> </span>nvidia
 </pre></div>
 </div>
-<p>or with GPU support:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba create --name sleap pip <span class="nv">python</span><span class="o">=</span><span class="m">3</span>.7.12 <span class="nv">cudatoolkit</span><span class="o">=</span><span class="m">11</span>.3 <span class="nv">cudnn</span><span class="o">=</span><span class="m">8</span>.2
+</div><div aria-labelledby="tab-9-Q1BVIG9yIG90aGVyIEdQVQ==" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-9-Q1BVIG9yIG90aGVyIEdQVQ==" name="Q1BVIG9yIG90aGVyIEdQVQ==" role="tabpanel" tabindex="0"><div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>create<span class="w"> </span>--name<span class="w"> </span>sleap<span class="w"> </span>pip<span class="w"> </span><span class="nv">python</span><span class="o">=</span><span class="m">3</span>.7.12
 </pre></div>
 </div>
+</div></div>
 </li>
 <li><p>Then activate the environment to isolate the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code> from other environments on your computer:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba activate sleap
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>activate<span class="w"> </span>sleap
 </pre></div>
 </div>
 <div class="admonition warning">
@@ -547,52 +547,44 @@ <h2><a class="toc-backref" href="#id3">Installation methods</a><a class="headerl
 </div>
 </li>
 <li><p>Finally, we can perform the <code class="docutils literal notranslate"><span class="pre">pip</span> <span class="pre">install</span></code>:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip install sleap<span class="o">[</span>pypi<span class="o">]==</span><span class="m">1</span>.3.3
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>sleap<span class="o">[</span>pypi<span class="o">]==</span><span class="m">1</span>.4.1
 </pre></div>
 </div>
-<p>This works on <strong>any OS except Apple silicon</strong> and on <strong>Google Colab</strong>.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p>The pypi distributed package of SLEAP ships with the following extras:</p>
 <ul class="simple">
-<li><p><strong>pypi</strong>: For installation without an mamba environment file. All dependencies come from PyPI.</p></li>
+<li><p><strong>pypi</strong>: For installation without an conda environment file. All dependencies come from PyPI.</p></li>
 <li><p><strong>jupyter</strong>: This installs all <em>pypi</em> and jupyter lab dependencies.</p></li>
 <li><p><strong>dev</strong>: This installs all <em>jupyter</em> dependencies and developement tools for testing and building docs.</p></li>
-<li><p><strong>conda_jupyter</strong>: For installation using a mamba environment file included in the source code. Most dependencies are listed as conda packages in the environment file and only a few come from PyPI to allow jupyter lab support.</p></li>
-<li><p><strong>conda_dev</strong>: For installation using <a class="reference external" href="https://github.com/search?q=repo%3Atalmolab%2Fsleap+path%3Aenvironment*.yml&amp;type=code">a mamba environment</a> with a few PyPI dependencies for development tools.</p></li>
+<li><p><strong>conda_jupyter</strong>: For installation using a conda environment file included in the source code. Most dependencies are listed as conda packages in the environment file and only a few come from PyPI to allow jupyter lab support.</p></li>
+<li><p><strong>conda_dev</strong>: For installation using <a class="reference external" href="https://github.com/search?q=repo%3Atalmolab%2Fsleap+path%3Aenvironment*.yml&amp;type=code">a conda environment</a> with a few PyPI dependencies for development tools.</p></li>
 </ul>
 </div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<ul class="simple">
-<li><p>Requires Python 3.7</p></li>
-<li><p>To enable GPU support, make sure that you have <strong>CUDA Toolkit v11.3</strong> and <strong>cuDNN v8.2</strong> installed.</p></li>
-</ul>
-</div>
-<div class="admonition warning">
-<p class="admonition-title">Warning</p>
-<p>This will uninstall existing libraries and potentially install conflicting ones.</p>
-<p>We strongly recommend that you <strong>only use this method if you know what you’re doing</strong>!</p>
-</div>
 </li>
 </ol>
-</section>
+</div><div aria-labelledby="tab-8-TWFjIE9T" class="sphinx-tabs-panel group-tab" hidden="true" id="panel-8-TWFjIE9T" name="TWFjIE9T" role="tabpanel" tabindex="0"><p>Not supported.</p>
+</div><div class="highlight-default notranslate"><div class="highlight"><pre><span></span>
+</pre></div>
+</div>
+</div>
+</div></div>
 </section>
 <section id="testing-that-things-are-working">
-<h2><a class="toc-backref" href="#id7">Testing that things are working</a><a class="headerlink" href="#testing-that-things-are-working" title="Permalink to this headline">#</a></h2>
-<p>If you installed using <code class="docutils literal notranslate"><span class="pre">mamba</span></code>, first activate the <code class="docutils literal notranslate"><span class="pre">sleap</span></code> environment by opening a terminal and typing:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba activate sleap
+<h2><a class="toc-backref" href="#id6" role="doc-backlink">Testing that things are working</a><a class="headerlink" href="#testing-that-things-are-working" title="Permalink to this heading">#</a></h2>
+<p>If you installed using <code class="docutils literal notranslate"><span class="pre">conda</span></code>, first activate the <code class="docutils literal notranslate"><span class="pre">sleap</span></code> environment by opening a terminal and typing:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>activate<span class="w"> </span>sleap
 </pre></div>
 </div>
 <div class="admonition hint">
 <p class="admonition-title">Hint</p>
-<p>Not sure what <code class="docutils literal notranslate"><span class="pre">mamba</span></code> environments you already installed? You can get a list of the environments on your system with:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">mamba</span> <span class="n">env</span> <span class="nb">list</span>
+<p>Not sure what <code class="docutils literal notranslate"><span class="pre">conda</span></code> environments you already installed? You can get a list of the environments on your system with:</p>
+<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">conda</span> <span class="n">env</span> <span class="nb">list</span>
 </pre></div>
 </div>
 </div>
 <section id="gui-support">
-<h3><a class="toc-backref" href="#id8">GUI support</a><a class="headerlink" href="#gui-support" title="Permalink to this headline">#</a></h3>
+<h3><a class="toc-backref" href="#id7" role="doc-backlink">GUI support</a><a class="headerlink" href="#gui-support" title="Permalink to this heading">#</a></h3>
 <p>To check that the GUI is working, simply type:</p>
 <div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>sleap-label
 </pre></div>
@@ -600,9 +592,9 @@ <h3><a class="toc-backref" href="#id8">GUI support</a><a class="headerlink" href
 <p>You should see the SLEAP labeling interface pop up within a few moments.</p>
 </section>
 <section id="importing">
-<h3><a class="toc-backref" href="#id9">Importing</a><a class="headerlink" href="#importing" title="Permalink to this headline">#</a></h3>
+<h3><a class="toc-backref" href="#id8" role="doc-backlink">Importing</a><a class="headerlink" href="#importing" title="Permalink to this heading">#</a></h3>
 <p>To check if SLEAP is installed correctly in non-interactive environments, such as remote servers, confirm that you can import it with:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python -c <span class="s2">&quot;import sleap; sleap.versions()&quot;</span>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python<span class="w"> </span>-c<span class="w"> </span><span class="s2">&quot;import sleap; sleap.versions()&quot;</span>
 </pre></div>
 </div>
 <p><small><strong>Output:</strong></small></p>
@@ -616,10 +608,10 @@ <h3><a class="toc-backref" href="#id9">Importing</a><a class="headerlink" href="
 </div>
 </section>
 <section id="gpu-support">
-<h3><a class="toc-backref" href="#id10">GPU support</a><a class="headerlink" href="#gpu-support" title="Permalink to this headline">#</a></h3>
-<p>Assuming you installed using either of the <code class="docutils literal notranslate"><span class="pre">mamba</span></code>-based methods on Windows or Linux, SLEAP should automatically have GPU support enabled.</p>
+<h3><a class="toc-backref" href="#id9" role="doc-backlink">GPU support</a><a class="headerlink" href="#gpu-support" title="Permalink to this heading">#</a></h3>
+<p>Assuming you installed using either of the <code class="docutils literal notranslate"><span class="pre">conda</span></code>-based methods on Windows or Linux, SLEAP should automatically have GPU support enabled.</p>
 <p>To check, verify that SLEAP can detect the GPUs on your system:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python -c <span class="s2">&quot;import sleap; sleap.system_summary()&quot;</span>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python<span class="w"> </span>-c<span class="w"> </span><span class="s2">&quot;import sleap; sleap.system_summary()&quot;</span>
 </pre></div>
 </div>
 <p><small><strong>Output:</strong></small></p>
@@ -636,7 +628,7 @@ <h3><a class="toc-backref" href="#id10">GPU support</a><a class="headerlink" hre
 </pre></div>
 </div>
 <p>SLEAP uses TensorFlow for GPU acceleration. To directly check if TensorFlow is detecting your GPUs:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python -c <span class="s2">&quot;import tensorflow as tf; print(tf.config.list_physical_devices(&#39;GPU&#39;))&quot;</span>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>python<span class="w"> </span>-c<span class="w"> </span><span class="s2">&quot;import tensorflow as tf; print(tf.config.list_physical_devices(&#39;GPU&#39;))&quot;</span>
 </pre></div>
 </div>
 <p><small><strong>Output:</strong></small></p>
@@ -657,7 +649,7 @@ <h3><a class="toc-backref" href="#id10">GPU support</a><a class="headerlink" hre
 <p class="admonition-title">Warning</p>
 <p>TensorFlow 2.7+ is currently failing to detect CUDA Toolkit and CuDNN on some systems (see <a class="reference external" href="https://github.com/tensorflow/tensorflow/issues/52988">Issue thread</a>).</p>
 <p>If you run into issues, either try downgrading the TensorFlow 2.6:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip install <span class="nv">tensorflow</span><span class="o">==</span><span class="m">2</span>.6.3
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span><span class="nv">tensorflow</span><span class="o">==</span><span class="m">2</span>.6.3
 </pre></div>
 </div>
 <p>or follow the note below.</p>
@@ -665,27 +657,27 @@ <h3><a class="toc-backref" href="#id10">GPU support</a><a class="headerlink" hre
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p>If you are on Linux, have a NVIDIA GPU, but cannot detect your GPU:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>W tensorflow/stream_executor/platform/default/dso_loader.cc:64 Could not load dynamic
-library <span class="s1">&#39;libcudart.so.11.0&#39;</span><span class="p">;</span> dlerror: libcudart.so.11.0: cannot open shared object
-file: No such file or directory
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>W<span class="w"> </span>tensorflow/stream_executor/platform/default/dso_loader.cc:64<span class="w"> </span>Could<span class="w"> </span>not<span class="w"> </span>load<span class="w"> </span>dynamic
+library<span class="w"> </span><span class="s1">&#39;libcudart.so.11.0&#39;</span><span class="p">;</span><span class="w"> </span>dlerror:<span class="w"> </span>libcudart.so.11.0:<span class="w"> </span>cannot<span class="w"> </span>open<span class="w"> </span>shared<span class="w"> </span>object
+file:<span class="w"> </span>No<span class="w"> </span>such<span class="w"> </span>file<span class="w"> </span>or<span class="w"> </span>directory
 </pre></div>
 </div>
 <p>then activate the environment:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba activate sleap
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>activate<span class="w"> </span>sleap
 </pre></div>
 </div>
 <p>and run the commands:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mkdir -p <span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d
-<span class="nb">echo</span> <span class="s1">&#39;#!/bin/sh&#39;</span> &gt;&gt; <span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
-<span class="nb">echo</span> <span class="s1">&#39;export SLEAP_OLD_LD_LIBRARY_PATH=$LD_LIBRARY_PATH&#39;</span> &gt;&gt; <span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
-<span class="nb">echo</span> <span class="s1">&#39;export LD_LIBRARY_PATH=$CONDA_PREFIX/lib:$LD_LIBRARY_PATH&#39;</span> &gt;&gt; <span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
-<span class="nb">source</span> <span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mkdir<span class="w"> </span>-p<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d
+<span class="nb">echo</span><span class="w"> </span><span class="s1">&#39;#!/bin/sh&#39;</span><span class="w"> </span>&gt;&gt;<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
+<span class="nb">echo</span><span class="w"> </span><span class="s1">&#39;export SLEAP_OLD_LD_LIBRARY_PATH=$LD_LIBRARY_PATH&#39;</span><span class="w"> </span>&gt;&gt;<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
+<span class="nb">echo</span><span class="w"> </span><span class="s1">&#39;export LD_LIBRARY_PATH=$CONDA_PREFIX/lib:$LD_LIBRARY_PATH&#39;</span><span class="w"> </span>&gt;&gt;<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
+<span class="nb">source</span><span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/activate.d/sleap_activate.sh
 </pre></div>
 </div>
 <p>This will set the environment variable <code class="docutils literal notranslate"><span class="pre">LD_LIBRARY_PATH</span></code> each time the environment is activated. The environment variable will remain set in the current terminal even if we deactivate the environment. Although not strictly necessary, if you would also like the environment variable to be reset to the original value when deactivating the environment, we can run the following commands:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mkdir -p <span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d
-<span class="nb">echo</span> <span class="s1">&#39;#!/bin/sh&#39;</span> &gt;&gt; <span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d/sleap_deactivate.sh
-<span class="nb">echo</span> <span class="s1">&#39;export LD_LIBRARY_PATH=$SLEAP_OLD_LD_LIBRARY_PATH&#39;</span> &gt;&gt; <span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d/sleap_deactivate.sh
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mkdir<span class="w"> </span>-p<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d
+<span class="nb">echo</span><span class="w"> </span><span class="s1">&#39;#!/bin/sh&#39;</span><span class="w"> </span>&gt;&gt;<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d/sleap_deactivate.sh
+<span class="nb">echo</span><span class="w"> </span><span class="s1">&#39;export LD_LIBRARY_PATH=$SLEAP_OLD_LD_LIBRARY_PATH&#39;</span><span class="w"> </span>&gt;&gt;<span class="w"> </span><span class="nv">$CONDA_PREFIX</span>/etc/conda/deactivate.d/sleap_deactivate.sh
 </pre></div>
 </div>
 <p>These commands only need to be run once and will subsequently run automatically upon [de]activating your <code class="docutils literal notranslate"><span class="pre">sleap</span></code> environment.</p>
@@ -693,30 +685,30 @@ <h3><a class="toc-backref" href="#id10">GPU support</a><a class="headerlink" hre
 </section>
 </section>
 <section id="upgrading-and-uninstalling">
-<h2><a class="toc-backref" href="#id11">Upgrading and uninstalling</a><a class="headerlink" href="#upgrading-and-uninstalling" title="Permalink to this headline">#</a></h2>
+<h2><a class="toc-backref" href="#id10" role="doc-backlink">Upgrading and uninstalling</a><a class="headerlink" href="#upgrading-and-uninstalling" title="Permalink to this heading">#</a></h2>
 <p>We <strong>strongly recommend</strong> installing SLEAP in a fresh environment when updating. This is because dependency versions might change, and depending on the state of your previous environment, directly updating might break compatibility with some of them.</p>
 <p>To uninstall an existing environment named <code class="docutils literal notranslate"><span class="pre">sleap</span></code>:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba env remove -n sleap
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>env<span class="w"> </span>remove<span class="w"> </span>-n<span class="w"> </span>sleap
 </pre></div>
 </div>
 <div class="admonition hint">
 <p class="admonition-title">Hint</p>
-<p>Not sure what <code class="docutils literal notranslate"><span class="pre">mamba</span></code> environments you already installed? You can get a list of the environments on your system with:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba env list
+<p>Not sure what <code class="docutils literal notranslate"><span class="pre">conda</span></code> environments you already installed? You can get a list of the environments on your system with:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>env<span class="w"> </span>list
 </pre></div>
 </div>
 </div>
 <p>Once the environment has been removed, you are free to install SLEAP using any of the installation methods above into an environment of the same name.</p>
 </section>
 <section id="getting-help">
-<h2><a class="toc-backref" href="#id12">Getting help</a><a class="headerlink" href="#getting-help" title="Permalink to this headline">#</a></h2>
+<h2><a class="toc-backref" href="#id11" role="doc-backlink">Getting help</a><a class="headerlink" href="#getting-help" title="Permalink to this heading">#</a></h2>
 <p>If you run into any problems, check out the <a class="reference external" href="https://github.com/talmolab/sleap/discussions">Github Discussions</a> and <a class="reference external" href="https://github.com/talmolab/sleap/issues">GitHub Issues</a> to see if others have had the same problem.</p>
 <p>If you get any errors or the GUI fails to launch, try running the diagnostics to see what SLEAP is able to detect on your system:</p>
 <div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>sleap-diagnostic
 </pre></div>
 </div>
-<p>If you were not able to get SLEAP installed, activate the mamba environment it is in and generate a list of the package versions installed:</p>
-<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>mamba list
+<p>If you were not able to get SLEAP installed, activate the conda environment it is in and generate a list of the package versions installed:</p>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>conda<span class="w"> </span>list
 </pre></div>
 </div>
 <p>Then, <a class="reference external" href="https://github.com/talmolab/sleap/issues">open a new Issue</a> providing the versions from either command above, as well as any errors you saw in the console during the installation. Or <a class="reference external" href="https://github.com/talmolab/sleap/discussions">start a discussion</a> to get help from the community.</p>
@@ -771,14 +763,12 @@ <h2><a class="toc-backref" href="#id12">Getting help</a><a class="headerlink" hr
   </div>
   <nav class="bd-toc-nav page-toc">
     <ul class="visible nav section-nav flex-column">
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#macs-pre-m1-pre-installation">Macs Pre-M1 (Pre-Installation)</a></li>
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installing-mambaforge">Installing Mambaforge</a></li>
-<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installation-methods">Installation methods</a><ul class="nav section-nav flex-column">
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#conda-package"><code class="docutils literal notranslate"><span class="pre">conda</span></code> package</a></li>
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#conda-from-source"><code class="docutils literal notranslate"><span class="pre">conda</span></code> from source</a></li>
-<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#pip-package"><code class="docutils literal notranslate"><span class="pre">pip</span></code> package</a></li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#package-manager">Package Manager</a><ul class="nav section-nav flex-column">
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#miniforge-recommended">Miniforge (recommended)</a></li>
+<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#miniconda">Miniconda</a></li>
 </ul>
 </li>
+<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#installation-methods">Installation methods</a></li>
 <li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#testing-that-things-are-working">Testing that things are working</a><ul class="nav section-nav flex-column">
 <li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#gui-support">GUI support</a></li>
 <li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#importing">Importing</a></li>
@@ -810,7 +800,7 @@ <h2><a class="toc-backref" href="#id12">Getting help</a><a class="headerlink" hr
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Analysis_examples.html b/notebooks/Analysis_examples.html
index b4f876962..e237e362a 100644
--- a/notebooks/Analysis_examples.html
+++ b/notebooks/Analysis_examples.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Analysis examples &#8212; SLEAP (v1.3.3)</title>
+    <title>Analysis examples &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Analysis_examples';</script>
     <link rel="canonical" href="/develop/notebooks/Analysis_examples.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Data structures" href="Data_structures.html" />
     <link rel="prev" title="Model evaluation" href="Model_evaluation.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -341,17 +339,17 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Analysis_examples.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="analysis-examples">
-<h1>Analysis examples<a class="headerlink" href="#analysis-examples" title="Permalink to this headline">#</a></h1>
+<h1>Analysis examples<a class="headerlink" href="#analysis-examples" title="Permalink to this heading">#</a></h1>
 <p>In this notebook we’ll show examples of how you might use the predictions exported from SLEAP. We’ll work with an <a class="reference external" href="https://sleap.ai/tutorials/analysis.html">Analysis HDF5</a> file (rather than the <code class="docutils literal notranslate"><span class="pre">.slp</span></code> predictions file). This HDF5 file can be exported from SLEAP.</p>
 <p>We advise building your post-SLEAP analysis pipeline around these HDF5 files rather than trying to work directly with the <code class="docutils literal notranslate"><span class="pre">.slp</span></code> files used by SLEAP.</p>
 <p><strong>Note</strong>: You can work with these HDF5 directly in Python (as we’ll do here) or MATLAB without having SLEAP itself installed.</p>
 <section id="example-analysis-data">
-<h2>Example analysis data<a class="headerlink" href="#example-analysis-data" title="Permalink to this headline">#</a></h2>
+<h2>Example analysis data<a class="headerlink" href="#example-analysis-data" title="Permalink to this heading">#</a></h2>
 <p>Let’s start by download a sample HDF5. These predictions were created with models trained on our <a class="reference external" href="https://github.com/talmolab/sleap-datasets">sample Drosophila melanogaster courtship dataset</a>. Using these models we can inference on a video clip with 3000 frames. The video clip, resulting predictions, and exported HDF5 are all available <a class="reference external" href="https://github.com/talmolab/sleap/tree/main/docs/notebooks/analysis_example">here</a>.</p>
 <p>We’ll just download the <code class="docutils literal notranslate"><span class="pre">predictions.analysis.h5</span></code> file:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget -O predictions.analysis.h5 https://github.com/talmolab/sleap/raw/main/docs/notebooks/analysis_example/predictions.analysis.h5
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget<span class="w"> </span>-O<span class="w"> </span>predictions.analysis.h5<span class="w"> </span>https://github.com/talmolab/sleap/raw/main/docs/notebooks/analysis_example/predictions.analysis.h5
 </pre></div>
 </div>
 </div>
@@ -385,7 +383,7 @@ <h2>Example analysis data<a class="headerlink" href="#example-analysis-data" tit
 </div>
 </section>
 <section id="loading-the-data">
-<h2>Loading the data<a class="headerlink" href="#loading-the-data" title="Permalink to this headline">#</a></h2>
+<h2>Loading the data<a class="headerlink" href="#loading-the-data" title="Permalink to this heading">#</a></h2>
 <p>We use the <a class="reference external" href="https://www.h5py.org">h5py</a> package to load data from the HDF5. This is already installed in Colab. If your running analysis code on your local machine and have SLEAP installed, then <code class="docutils literal notranslate"><span class="pre">h5py</span></code> and other packages we use are already installed in your SLEAP conda environment. Otherwise, you may need to use <code class="docutils literal notranslate"><span class="pre">conda</span></code> or <code class="docutils literal notranslate"><span class="pre">pip</span></code> to install <code class="docutils literal notranslate"><span class="pre">h5py</span></code> as well as <code class="docutils literal notranslate"><span class="pre">numpy</span></code>, <code class="docutils literal notranslate"><span class="pre">scipy</span></code>, <code class="docutils literal notranslate"><span class="pre">matplotlib</span></code>, <code class="docutils literal notranslate"><span class="pre">seaborn</span></code>, and any other packages you want use in your analysis code.</p>
 <p>Let’s load the file and take a peek.</p>
 <div class="cell docutils container">
@@ -475,13 +473,13 @@ <h2>Loading the data<a class="headerlink" href="#loading-the-data" title="Permal
 <p>Now that we’ve loaded the data, let’s see some different things we can do with it…</p>
 </section>
 <section id="fill-missing-values">
-<h2>Fill missing values<a class="headerlink" href="#fill-missing-values" title="Permalink to this headline">#</a></h2>
+<h2>Fill missing values<a class="headerlink" href="#fill-missing-values" title="Permalink to this heading">#</a></h2>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">scipy.interpolate</span> <span class="kn">import</span> <span class="n">interp1d</span>
 
 <span class="k">def</span> <span class="nf">fill_missing</span><span class="p">(</span><span class="n">Y</span><span class="p">,</span> <span class="n">kind</span><span class="o">=</span><span class="s2">&quot;linear&quot;</span><span class="p">):</span>
-    <span class="sd">&quot;&quot;&quot;Fills missing values independently along each dimension after the first.&quot;&quot;&quot;</span>
+<span class="w">    </span><span class="sd">&quot;&quot;&quot;Fills missing values independently along each dimension after the first.&quot;&quot;&quot;</span>
 
     <span class="c1"># Store initial shape.</span>
     <span class="n">initial_shape</span> <span class="o">=</span> <span class="n">Y</span><span class="o">.</span><span class="n">shape</span>
@@ -525,7 +523,7 @@ <h2>Fill missing values<a class="headerlink" href="#fill-missing-values" title="
 </div>
 </section>
 <section id="visualize-thorax-movement-across-video">
-<h2>Visualize thorax movement across video<a class="headerlink" href="#visualize-thorax-movement-across-video" title="Permalink to this headline">#</a></h2>
+<h2>Visualize thorax movement across video<a class="headerlink" href="#visualize-thorax-movement-across-video" title="Permalink to this heading">#</a></h2>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="n">HEAD_INDEX</span> <span class="o">=</span> <span class="mi">0</span>
@@ -587,20 +585,20 @@ <h2>Visualize thorax movement across video<a class="headerlink" href="#visualize
 <div class="output text_plain highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>Text(0.5, 1.0, &#39;Thorax tracks&#39;)
 </pre></div>
 </div>
-<img alt="../_images/Analysis_examples_19_1.png" src="../_images/Analysis_examples_19_1.png" />
-<img alt="../_images/Analysis_examples_19_2.png" src="../_images/Analysis_examples_19_2.png" />
+<img alt="../_images/dc4bfaee117747c1eeba0b62c375704e720b95ba9694ecd512a2cc28d501fbab.png" src="../_images/dc4bfaee117747c1eeba0b62c375704e720b95ba9694ecd512a2cc28d501fbab.png" />
+<img alt="../_images/35c10296694a5eb3635e87b56ab9545d9a293bed86bd144682549fe750d7a34d.png" src="../_images/35c10296694a5eb3635e87b56ab9545d9a293bed86bd144682549fe750d7a34d.png" />
 </div>
 </div>
 </section>
 <section id="more-advanced-visualizations">
-<h2>More advanced visualizations<a class="headerlink" href="#more-advanced-visualizations" title="Permalink to this headline">#</a></h2>
+<h2>More advanced visualizations<a class="headerlink" href="#more-advanced-visualizations" title="Permalink to this heading">#</a></h2>
 <p>For some additional analysis, we’ll first smooth and differentiate the data with a Savitzky-Golay filter to extract velocities of each joint.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">scipy.signal</span> <span class="kn">import</span> <span class="n">savgol_filter</span>
 
 <span class="k">def</span> <span class="nf">smooth_diff</span><span class="p">(</span><span class="n">node_loc</span><span class="p">,</span> <span class="n">win</span><span class="o">=</span><span class="mi">25</span><span class="p">,</span> <span class="n">poly</span><span class="o">=</span><span class="mi">3</span><span class="p">):</span>
-    <span class="sd">&quot;&quot;&quot;</span>
+<span class="w">    </span><span class="sd">&quot;&quot;&quot;</span>
 <span class="sd">    node_loc is a [frames, 2] array</span>
 <span class="sd">    </span>
 <span class="sd">    win defines the window to smooth over</span>
@@ -631,7 +629,7 @@ <h2>More advanced visualizations<a class="headerlink" href="#more-advanced-visua
 </div>
 </div>
 <section id="visualizing-thorax-x-y-dynamics-and-velocity-for-fly-0">
-<h3>Visualizing thorax x-y dynamics and velocity for fly 0<a class="headerlink" href="#visualizing-thorax-x-y-dynamics-and-velocity-for-fly-0" title="Permalink to this headline">#</a></h3>
+<h3>Visualizing thorax x-y dynamics and velocity for fly 0<a class="headerlink" href="#visualizing-thorax-x-y-dynamics-and-velocity-for-fly-0" title="Permalink to this heading">#</a></h3>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="n">fig</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">figure</span><span class="p">(</span><span class="n">figsize</span><span class="o">=</span><span class="p">(</span><span class="mi">15</span><span class="p">,</span><span class="mi">7</span><span class="p">))</span>
@@ -653,12 +651,12 @@ <h3>Visualizing thorax x-y dynamics and velocity for fly 0<a class="headerlink"
 <div class="output text_plain highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>Text(0.5, 1.0, &#39;Velocity&#39;)
 </pre></div>
 </div>
-<img alt="../_images/Analysis_examples_25_1.png" src="../_images/Analysis_examples_25_1.png" />
+<img alt="../_images/6a87f28526a92715315c4766f197b24f30f433e554ae3059a3d226b457135a00.png" src="../_images/6a87f28526a92715315c4766f197b24f30f433e554ae3059a3d226b457135a00.png" />
 </div>
 </div>
 </section>
 <section id="visualize-thorax-colored-by-magnitude-of-fly-speed">
-<h3>Visualize thorax colored by magnitude of fly speed<a class="headerlink" href="#visualize-thorax-colored-by-magnitude-of-fly-speed" title="Permalink to this headline">#</a></h3>
+<h3>Visualize thorax colored by magnitude of fly speed<a class="headerlink" href="#visualize-thorax-colored-by-magnitude-of-fly-speed" title="Permalink to this heading">#</a></h3>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="n">fig</span> <span class="o">=</span> <span class="n">plt</span><span class="o">.</span><span class="n">figure</span><span class="p">(</span><span class="n">figsize</span><span class="o">=</span><span class="p">(</span><span class="mi">15</span><span class="p">,</span><span class="mi">6</span><span class="p">))</span>
@@ -688,18 +686,18 @@ <h3>Visualize thorax colored by magnitude of fly speed<a class="headerlink" href
 <div class="output text_plain highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>Text(0.5, 1.0, &#39;Thorax tracks colored by magnitude of fly speed&#39;)
 </pre></div>
 </div>
-<img alt="../_images/Analysis_examples_27_1.png" src="../_images/Analysis_examples_27_1.png" />
+<img alt="../_images/4c35c39e8dcf293ab2362ab3534303852fcba727e88b0345cabc9999244acf84.png" src="../_images/4c35c39e8dcf293ab2362ab3534303852fcba727e88b0345cabc9999244acf84.png" />
 </div>
 </div>
 </section>
 <section id="find-covariance-in-thorax-velocities-between-fly-0-and-fly-1">
-<h3>Find covariance in thorax velocities between fly-0 and fly-1<a class="headerlink" href="#find-covariance-in-thorax-velocities-between-fly-0-and-fly-1" title="Permalink to this headline">#</a></h3>
+<h3>Find covariance in thorax velocities between fly-0 and fly-1<a class="headerlink" href="#find-covariance-in-thorax-velocities-between-fly-0-and-fly-1" title="Permalink to this heading">#</a></h3>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">pandas</span> <span class="k">as</span> <span class="nn">pd</span>
 
 <span class="k">def</span> <span class="nf">corr_roll</span><span class="p">(</span><span class="n">datax</span><span class="p">,</span> <span class="n">datay</span><span class="p">,</span> <span class="n">win</span><span class="p">):</span>
-    <span class="sd">&quot;&quot;&quot;</span>
+<span class="w">    </span><span class="sd">&quot;&quot;&quot;</span>
 <span class="sd">    datax, datay are the two timeseries to find correlations between</span>
 <span class="sd">    </span>
 <span class="sd">    win sets the number of frames over which the covariance is computed</span>
@@ -735,15 +733,15 @@ <h3>Find covariance in thorax velocities between fly-0 and fly-1<a class="header
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Analysis_examples_30_0.png" src="../_images/Analysis_examples_30_0.png" />
+<img alt="../_images/b21d966bde1fb74635e96843f91a4ec1072e73e79320eae51395ff6e3a9fe551.png" src="../_images/b21d966bde1fb74635e96843f91a4ec1072e73e79320eae51395ff6e3a9fe551.png" />
 </div>
 </div>
 </section>
 </section>
 <section id="clustering">
-<h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this headline">#</a></h2>
+<h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this heading">#</a></h2>
 <p>For an example of clustering the data, we’ll</p>
-<ol class="simple">
+<ol class="arabic simple">
 <li><p>extract joint velocities for each joint,</p></li>
 <li><p>run simple k-means on the velocities from each frame.</p></li>
 </ol>
@@ -783,7 +781,7 @@ <h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Analysis_examples_34_0.png" src="../_images/Analysis_examples_34_0.png" />
+<img alt="../_images/42e0a303da85179624b2247d2d06868f157496f5fefdd2f4e8ad741179cb6e69.png" src="../_images/42e0a303da85179624b2247d2d06868f157496f5fefdd2f4e8ad741179cb6e69.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -795,7 +793,7 @@ <h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Analysis_examples_35_0.png" src="../_images/Analysis_examples_35_0.png" />
+<img alt="../_images/6dcf42bdd95e0ded30ed1fb43ad73bdc855130d58eaa29f55050031f80d0a148.png" src="../_images/6dcf42bdd95e0ded30ed1fb43ad73bdc855130d58eaa29f55050031f80d0a148.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -838,7 +836,7 @@ <h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Analysis_examples_38_0.png" src="../_images/Analysis_examples_38_0.png" />
+<img alt="../_images/b44ac03a6b1270ea102001c424a678d09aa630125b1a04b4b4429de3f8b4f1fc.png" src="../_images/b44ac03a6b1270ea102001c424a678d09aa630125b1a04b4b4429de3f8b4f1fc.png" />
 </div>
 </div>
 </section>
@@ -926,7 +924,7 @@ <h2>Clustering<a class="headerlink" href="#clustering" title="Permalink to this
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Data_structures.html b/notebooks/Data_structures.html
index 98f5ac08c..bd464b6be 100644
--- a/notebooks/Data_structures.html
+++ b/notebooks/Data_structures.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Data structures &#8212; SLEAP (v1.3.3)</title>
+    <title>Data structures &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Data_structures';</script>
     <link rel="canonical" href="/develop/notebooks/Data_structures.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Post-inference tracking" href="Post_inference_tracking.html" />
     <link rel="prev" title="Analysis examples" href="Analysis_examples.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -332,7 +330,7 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Data_structures.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="data-structures">
-<h1>Data structures<a class="headerlink" href="#data-structures" title="Permalink to this headline">#</a></h1>
+<h1>Data structures<a class="headerlink" href="#data-structures" title="Permalink to this heading">#</a></h1>
 <p>In this notebook, we will explore some of the major data structures used in SLEAP and how they can be manipulated when generating predictions from trained models.</p>
 <p>A quick overview of the data structures before we start:</p>
 <ul class="simple">
@@ -343,14 +341,14 @@ <h1>Data structures<a class="headerlink" href="#data-structures" title="Permalin
 <li><p><code class="docutils literal notranslate"><span class="pre">Labels</span></code> → Contains a set of <code class="docutils literal notranslate"><span class="pre">LabeledFrame</span></code>s and the associated metadata for the videos and other information related to the project or predictions.</p></li>
 </ul>
 <section id="setup-sleap-and-data">
-<h2>1. Setup SLEAP and data<a class="headerlink" href="#setup-sleap-and-data" title="Permalink to this headline">#</a></h2>
+<h2>1. Setup SLEAP and data<a class="headerlink" href="#setup-sleap-and-data" title="Permalink to this heading">#</a></h2>
 <p>We’ll start by installing SLEAP and downloading some data and models to play around with.</p>
 <p>If you get a dependency error in subsequent cells, just click <strong>Runtime</strong> → <strong>Restart runtime</strong> to reload the packages.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># This should take care of all the dependencies on colab:</span>
-<span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 
 <span class="c1"># But to do it locally, we&#39;d recommend the conda package (available on Windows + Linux):</span>
 <span class="c1"># conda create -n sleap -c sleap -c conda-forge -c nvidia sleap</span>
@@ -361,17 +359,17 @@ <h2>1. Setup SLEAP and data<a class="headerlink" href="#setup-sleap-and-data" ti
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># Test video:</span>
-<span class="o">!</span>wget https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.mp4
+<span class="o">!</span>wget<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.mp4
 
 <span class="c1"># Test video labels (from predictions/not necessary for inference benchmarking):</span>
-<span class="o">!</span>wget https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.slp
+<span class="o">!</span>wget<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.slp
 
 <span class="c1"># Bottom-up model:</span>
 <span class="c1"># !wget https://storage.googleapis.com/sleap-data/reference/flies13/bu.210506_230852.multi_instance.n%3D1800.zip</span>
 
 <span class="c1"># Top-down model (two-stage):</span>
-<span class="o">!</span>wget https://storage.googleapis.com/sleap-data/reference/flies13/centroid.fast.210504_182918.centroid.n%3D1800.zip
-<span class="o">!</span>wget https://storage.googleapis.com/sleap-data/reference/flies13/td_fast.210505_012601.centered_instance.n%3D1800.zip
+<span class="o">!</span>wget<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/centroid.fast.210504_182918.centroid.n%3D1800.zip
+<span class="o">!</span>wget<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/td_fast.210505_012601.centered_instance.n%3D1800.zip
 </pre></div>
 </div>
 </div>
@@ -425,7 +423,7 @@ <h2>1. Setup SLEAP and data<a class="headerlink" href="#setup-sleap-and-data" ti
 </div>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>ls -lah
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>ls<span class="w"> </span>-lah
 </pre></div>
 </div>
 </div>
@@ -508,7 +506,7 @@ <h2>1. Setup SLEAP and data<a class="headerlink" href="#setup-sleap-and-data" ti
 </div>
 </section>
 <section id="data-structures-and-inference">
-<h2>2. Data structures and inference<a class="headerlink" href="#data-structures-and-inference" title="Permalink to this headline">#</a></h2>
+<h2>2. Data structures and inference<a class="headerlink" href="#data-structures-and-inference" title="Permalink to this heading">#</a></h2>
 <p>SLEAP can read videos in a variety of different formats through the <code class="docutils literal notranslate"><span class="pre">sleap.load_video</span></code> high level API. Once loaded, the <code class="docutils literal notranslate"><span class="pre">sleap.Video</span></code> object allows you to access individual frames as if the it were a standard numpy array.</p>
 <p><strong>Note:</strong> The actual frames are not loaded until you access them so we don’t blow up our memory when using long videos.</p>
 <div class="cell docutils container">
@@ -568,9 +566,7 @@ <h2>2. Data structures and inference<a class="headerlink" href="#data-structures
 </div>
 </div>
 <div class="cell_output docutils container">
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "cf38d776e9fc48ada47705ce018c64af", "version_major": 2, "version_minor": 0}
-</script><div class="output stderr highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>2023-08-31 12:04:01.923466: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_FLOAT } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_FLOAT shape { dim { size: -45 } dim { size: -46 } dim { size: -47 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -15 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: -48 } dim { size: -49 } dim { size: 1 } } }
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "cf38d776e9fc48ada47705ce018c64af", "version_major": 2, "version_minor": 0}</script><div class="output stderr highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>2023-08-31 12:04:01.923466: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_FLOAT } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_FLOAT shape { dim { size: -45 } dim { size: -46 } dim { size: -47 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -15 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: -48 } dim { size: -49 } dim { size: 1 } } }
 2023-08-31 12:04:01.923717: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_UINT8 } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_UINT8 shape { dim { size: 4 } dim { size: 1024 } dim { size: 1024 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -15 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -15 } dim { size: -56 } dim { size: -57 } dim { size: 1 } } }
 2023-08-31 12:04:01.926044: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_FLOAT } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_FLOAT shape { dim { size: -90 } dim { size: -91 } dim { size: -92 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -20 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -20 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -20 } dim { size: -94 } dim { size: -95 } dim { size: 1 } } }
 </pre></div>
@@ -630,7 +626,7 @@ <h2>2. Data structures and inference<a class="headerlink" href="#data-structures
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Data_structures_20_0.png" src="../_images/Data_structures_20_0.png" />
+<img alt="../_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png" src="../_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png" />
 </div>
 </div>
 <p>The labeled frame is itself a container for instances:</p>
@@ -722,7 +718,7 @@ <h2>2. Data structures and inference<a class="headerlink" href="#data-structures
 <p>Let’s delete the source data:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>rm <span class="s2">&quot;190719_090330_wt_18159206_rig1.2@15000-17560.mp4&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>rm<span class="w"> </span><span class="s2">&quot;190719_090330_wt_18159206_rig1.2@15000-17560.mp4&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -748,7 +744,7 @@ <h2>2. Data structures and inference<a class="headerlink" href="#data-structures
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Data_structures_34_0.png" src="../_images/Data_structures_34_0.png" />
+<img alt="../_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png" src="../_images/7e506e2f6407ee4d65215f596800b1ab55b0a5571d0d2bb8b55ff041fe2e88f0.png" />
 </div>
 </div>
 </section>
@@ -827,7 +823,7 @@ <h2>2. Data structures and inference<a class="headerlink" href="#data-structures
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Interactive_and_realtime_inference.html b/notebooks/Interactive_and_realtime_inference.html
index 08ab557dd..11def9424 100644
--- a/notebooks/Interactive_and_realtime_inference.html
+++ b/notebooks/Interactive_and_realtime_inference.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Interactive and realtime inference &#8212; SLEAP (v1.3.3)</title>
+    <title>Interactive and realtime inference &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Interactive_and_realtime_inference';</script>
     <link rel="canonical" href="/develop/notebooks/Interactive_and_realtime_inference.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Developer API" href="../api.html" />
     <link rel="prev" title="Interactive and resumable training" href="Interactive_and_resumable_training.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -334,18 +332,18 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Interactive_and_realtime_inference.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="interactive-and-realtime-inference">
-<h1>Interactive and realtime inference<a class="headerlink" href="#interactive-and-realtime-inference" title="Permalink to this headline">#</a></h1>
+<h1>Interactive and realtime inference<a class="headerlink" href="#interactive-and-realtime-inference" title="Permalink to this heading">#</a></h1>
 <p>For most workflows, using the <a class="reference external" href="https://sleap.ai/guides/cli.html#sleap-track"><code class="docutils literal notranslate"><span class="pre">sleap-track</span></code> CLI</a> is probably the most convenient option, but if you’re developing a custom application you can take advantage of SLEAP’s inference API to use your trained models in your own custom scripts.</p>
 <p>In this notebook we will explore how to predict poses from raw images in pure Python, and do some basic benchmarking on a simulated realtime predictor that could be used to enable closed-loop experiments.</p>
 <section id="setup-sleap">
-<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this headline">#</a></h2>
+<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this heading">#</a></h2>
 <p>Run this cell first to install SLEAP. If you get a dependency error in subsequent cells, just click <strong>Runtime</strong> → <strong>Restart runtime</strong> to reload the packages.</p>
 <p>Don’t forget to set <strong>Runtime</strong> → <strong>Change runtime type</strong> → <strong>GPU</strong> as the accelerator.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># This should take care of all the dependencies on colab:</span>
-<span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 
 
 <span class="c1"># But to do it locally, we&#39;d recommend the conda package (available on Windows + Linux):</span>
@@ -393,14 +391,14 @@ <h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to
 </div>
 </section>
 <section id="setup-data">
-<h2>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to this headline">#</a></h2>
+<h2>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to this heading">#</a></h2>
 <p>Before we start, let’s download a raw video and a set of trained top-down ID models that we’ll use to build our application around.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>curl -L --output video.mp4 https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.mp4
-<span class="o">!</span>curl -L --output centroid_model.zip https://storage.googleapis.com/sleap-data/reference/flies13/centroid.fast.210504_182918.centroid.n%3D1800.zip
-<span class="o">!</span>curl -L --output centered_instance_id_model.zip https://storage.googleapis.com/sleap-data/reference/flies13/td_id.fast.v2.210519_111253.multi_class_topdown.n%3D1800.zip
-<span class="o">!</span>ls -lah
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>curl<span class="w"> </span>-L<span class="w"> </span>--output<span class="w"> </span>video.mp4<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/190719_090330_wt_18159206_rig1.2%4015000-17560.mp4
+<span class="o">!</span>curl<span class="w"> </span>-L<span class="w"> </span>--output<span class="w"> </span>centroid_model.zip<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/centroid.fast.210504_182918.centroid.n%3D1800.zip
+<span class="o">!</span>curl<span class="w"> </span>-L<span class="w"> </span>--output<span class="w"> </span>centered_instance_id_model.zip<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/td_id.fast.v2.210519_111253.multi_class_topdown.n%3D1800.zip
+<span class="o">!</span>ls<span class="w"> </span>-lah
 </pre></div>
 </div>
 </div>
@@ -454,7 +452,7 @@ <h2>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to th
 <p><strong>Note:</strong> These zip files just have the contents of standard SLEAP model folders that are generated during training.</p>
 </section>
 <section id="interactive-inference">
-<h2>3. Interactive inference<a class="headerlink" href="#interactive-inference" title="Permalink to this headline">#</a></h2>
+<h2>3. Interactive inference<a class="headerlink" href="#interactive-inference" title="Permalink to this heading">#</a></h2>
 <p>SLEAP provides a high-level API for performing inference in the form of <code class="docutils literal notranslate"><span class="pre">Predictor</span></code> classes specific to each approach/model type.</p>
 <p>To create one from a set of trained models, we can use the high-level <code class="docutils literal notranslate"><span class="pre">sleap.load_model()</span></code> function:</p>
 <div class="cell docutils container">
@@ -506,9 +504,7 @@ <h2>3. Interactive inference<a class="headerlink" href="#interactive-inference"
 </div>
 </div>
 <div class="cell_output docutils container">
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "0cc2e3a471764285a58d023906ba1f7a", "version_major": 2, "version_minor": 0}
-</script><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>imgs.shape: (100, 1024, 1024, 1)
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "0cc2e3a471764285a58d023906ba1f7a", "version_major": 2, "version_minor": 0}</script><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>imgs.shape: (100, 1024, 1024, 1)
 </pre></div>
 </div>
 <div class="output stderr highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>2023-09-01 13:57:13.455046: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8201
@@ -534,7 +530,7 @@ <h2>3. Interactive inference<a class="headerlink" href="#interactive-inference"
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Interactive_and_realtime_inference_16_0.png" src="../_images/Interactive_and_realtime_inference_16_0.png" />
+<img alt="../_images/ca67ed7d78a0a607f47eea0f22d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png" src="../_images/ca67ed7d78a0a607f47eea0f22d937fc65e07bfc3c875c09ac3e2290aea8f4e7.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -1270,7 +1266,7 @@ <h2>3. Interactive inference<a class="headerlink" href="#interactive-inference"
 </div>
 </section>
 <section id="realtime-performance">
-<h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" title="Permalink to this headline">#</a></h2>
+<h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" title="Permalink to this heading">#</a></h2>
 <p>Now that we know how to do inference with different types of outputs, let’s try to use that to build a simulated “realtime” application with timing.</p>
 <p>First, we’ll create a class that simulates a camera grabber API that provides a sequence of pre-loaded frames.</p>
 <div class="cell docutils container">
@@ -1280,7 +1276,7 @@ <h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" ti
 
 
 <span class="k">class</span> <span class="nc">SimulatedCamera</span><span class="p">:</span>
-    <span class="sd">&quot;&quot;&quot;Simulated camera class that serves frames from memory continuously.</span>
+<span class="w">    </span><span class="sd">&quot;&quot;&quot;Simulated camera class that serves frames from memory continuously.</span>
 
 <span class="sd">    Attributes:</span>
 <span class="sd">        frames: Numpy array with pre-loaded frames.</span>
@@ -1355,7 +1351,7 @@ <h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" ti
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Interactive_and_realtime_inference_27_0.png" src="../_images/Interactive_and_realtime_inference_27_0.png" />
+<img alt="../_images/507b37f5c35704f694d1e81f64e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png" src="../_images/507b37f5c35704f694d1e81f64e1dc7a4c8f1ea2f81e7ea81aa3d61b9cc5c3a9.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -1368,7 +1364,7 @@ <h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" ti
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Interactive_and_realtime_inference_28_0.png" src="../_images/Interactive_and_realtime_inference_28_0.png" />
+<img alt="../_images/7d06f5baca09a6bf3099c8ae1060ddf3eba8874e355ff047933d9a606da9318e.png" src="../_images/7d06f5baca09a6bf3099c8ae1060ddf3eba8874e355ff047933d9a606da9318e.png" />
 </div>
 </div>
 </section>
@@ -1449,7 +1445,7 @@ <h2>4. Realtime performance<a class="headerlink" href="#realtime-performance" ti
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Interactive_and_resumable_training.html b/notebooks/Interactive_and_resumable_training.html
index 9d05e9ee4..eb4a496a0 100644
--- a/notebooks/Interactive_and_resumable_training.html
+++ b/notebooks/Interactive_and_resumable_training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Interactive and resumable training &#8212; SLEAP (v1.3.3)</title>
+    <title>Interactive and resumable training &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Interactive_and_resumable_training';</script>
     <link rel="canonical" href="/develop/notebooks/Interactive_and_resumable_training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Interactive and realtime inference" href="Interactive_and_realtime_inference.html" />
     <link rel="prev" title="Post-inference tracking" href="Post_inference_tracking.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -335,19 +333,19 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Interactive_and_resumable_training.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="interactive-and-resumable-training">
-<h1>Interactive and resumable training<a class="headerlink" href="#interactive-and-resumable-training" title="Permalink to this headline">#</a></h1>
+<h1>Interactive and resumable training<a class="headerlink" href="#interactive-and-resumable-training" title="Permalink to this heading">#</a></h1>
 <p>Most of the time, you will be training models through the GUI or using the <a class="reference external" href="https://sleap.ai/guides/cli.html#sleap-train"><code class="docutils literal notranslate"><span class="pre">sleap-train</span></code> CLI</a>.</p>
 <p>If you’d like to customize the training process, however, you can use SLEAP’s low-level training functionality interactively. This allows you to define scripts that train models according to your own workflow, for example, to <strong>resume training</strong> on an already trained model. Another possible application would be to train a model using <strong>transfer learning</strong>, where a pretrained model can be used to initialize the weights of the new model.</p>
 <p>In this notebook we will explore how to set up a training job and train a model for multiple rounds without the GUI or CLI.</p>
 <section id="setup-sleap">
-<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this headline">#</a></h2>
+<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this heading">#</a></h2>
 <p>Run this cell first to install SLEAP. If you get a dependency error in subsequent cells, just click <strong>Runtime</strong> → <strong>Restart runtime</strong> to reload the packages.</p>
 <p>Don’t forget to set <strong>Runtime</strong> → <strong>Change runtime type</strong> → <strong>GPU</strong> as the accelerator.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># This should take care of all the dependencies on colab:</span>
-<span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 
 
 <span class="c1"># But to do it locally, we&#39;d recommend the conda package (available on Windows + Linux):</span>
@@ -389,14 +387,14 @@ <h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to
 </div>
 </section>
 <section id="setup-training-data">
-<h2>2. Setup training data<a class="headerlink" href="#setup-training-data" title="Permalink to this headline">#</a></h2>
+<h2>2. Setup training data<a class="headerlink" href="#setup-training-data" title="Permalink to this heading">#</a></h2>
 <p>Here we will download an existing training dataset package. This is an <code class="docutils literal notranslate"><span class="pre">.slp</span></code> file that contains both the labeled poses, as well as the image data for labeled frames.</p>
 <p>If running on Google Colab, you’ll want to replace this with mounting your Google Drive folder containing your own data, or if running locally, simply change the path to your labels below in <code class="docutils literal notranslate"><span class="pre">TRAINING_SLP_FILE</span></code>.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># !curl -L --output labels.pkg.slp https://www.dropbox.com/s/b990gxjt3d3j3jh/210205.sleap_wt_gold.13pt.pkg.slp?dl=1</span>
-<span class="o">!</span>curl -L --output labels.pkg.slp https://storage.googleapis.com/sleap-data/datasets/wt_gold.13pt/tracking_split2/train.pkg.slp
-<span class="o">!</span>ls -lah
+<span class="o">!</span>curl<span class="w"> </span>-L<span class="w"> </span>--output<span class="w"> </span>labels.pkg.slp<span class="w"> </span>https://storage.googleapis.com/sleap-data/datasets/wt_gold.13pt/tracking_split2/train.pkg.slp
+<span class="o">!</span>ls<span class="w"> </span>-lah
 </pre></div>
 </div>
 </div>
@@ -431,7 +429,7 @@ <h2>2. Setup training data<a class="headerlink" href="#setup-training-data" titl
 </div>
 </section>
 <section id="setup-training-job">
-<h2>3. Setup training job<a class="headerlink" href="#setup-training-job" title="Permalink to this headline">#</a></h2>
+<h2>3. Setup training job<a class="headerlink" href="#setup-training-job" title="Permalink to this heading">#</a></h2>
 <p>A SLEAP <code class="docutils literal notranslate"><span class="pre">TrainingJobConfig</span></code> is a structure that contains all of the hyperparameters needed to train a SLEAP model. This is typically saved out to <code class="docutils literal notranslate"><span class="pre">initial_config.json</span></code> and <code class="docutils literal notranslate"><span class="pre">training_config.json</span></code> in the model folder so that training runs can be reproduced if needed, as well as to store metadata necessary for inference.</p>
 <p>Normally, these are generated interactively by the GUI, or manually by editing an existing JSON file in a text editor. Here, we will define a configuration interactively entirely in Python.</p>
 <div class="cell docutils container">
@@ -468,12 +466,12 @@ <h2>3. Setup training job<a class="headerlink" href="#setup-training-job" title=
 </div>
 </div>
 <p>Existing configs can also be loaded from a <code class="docutils literal notranslate"><span class="pre">.json</span></code> file with:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">cfg</span> <span class="o">=</span> <span class="n">sleap</span><span class="o">.</span><span class="n">load_config</span><span class="p">(</span><span class="s2">&quot;training_config.json&quot;</span><span class="p">)</span>
+<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="linenos">1</span><span class="n">cfg</span> <span class="o">=</span> <span class="n">sleap</span><span class="o">.</span><span class="n">load_config</span><span class="p">(</span><span class="s2">&quot;training_config.json&quot;</span><span class="p">)</span>
 </pre></div>
 </div>
 </section>
 <section id="training">
-<h2>4. Training<a class="headerlink" href="#training" title="Permalink to this headline">#</a></h2>
+<h2>4. Training<a class="headerlink" href="#training" title="Permalink to this heading">#</a></h2>
 <p>Next we will create a SLEAP <code class="docutils literal notranslate"><span class="pre">Trainer</span></code> from the configuration we just specified. This handles all the nitty gritty mechanics necessary to setup training in the backend.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
@@ -569,9 +567,7 @@ <h2>4. Training<a class="headerlink" href="#training" title="Permalink to this h
 INFO:sleap.nn.training:Saving evaluation metrics to model folder...
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "9864dea73605449cb08b26c938812cfb", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "9864dea73605449cb08b26c938812cfb", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.train.slp
@@ -579,9 +575,7 @@ <h2>4. Training<a class="headerlink" href="#training" title="Permalink to this h
 INFO:sleap.nn.evals:OKS mAP: 0.508754
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "243984a359bc41e9975653fa6206ac27", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "243984a359bc41e9975653fa6206ac27", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.val.slp
@@ -593,7 +587,7 @@ <h2>4. Training<a class="headerlink" href="#training" title="Permalink to this h
 </div>
 </section>
 <section id="continuing-training">
-<h2>5. Continuing training<a class="headerlink" href="#continuing-training" title="Permalink to this headline">#</a></h2>
+<h2>5. Continuing training<a class="headerlink" href="#continuing-training" title="Permalink to this heading">#</a></h2>
 <p>If we still have the trainer in memory, we can continue training by simply calling <code class="docutils literal notranslate"><span class="pre">trainer.train()</span></code> again with a potentially different number of epochs:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
@@ -617,9 +611,7 @@ <h2>5. Continuing training<a class="headerlink" href="#continuing-training" titl
 INFO:sleap.nn.training:Saving evaluation metrics to model folder...
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "f1bb0ee48431420d9cb6d99c4db4680d", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "f1bb0ee48431420d9cb6d99c4db4680d", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.train.slp
@@ -627,9 +619,7 @@ <h2>5. Continuing training<a class="headerlink" href="#continuing-training" titl
 INFO:sleap.nn.evals:OKS mAP: 0.559100
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "db5de880cd154476a097178972c8f0a3", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "db5de880cd154476a097178972c8f0a3", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.val.slp
@@ -712,9 +702,7 @@ <h2>5. Continuing training<a class="headerlink" href="#continuing-training" titl
 INFO:sleap.nn.training:Saving evaluation metrics to model folder...
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "b94057057f6442c990c6fc548910a685", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "b94057057f6442c990c6fc548910a685", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.train.slp
@@ -722,9 +710,7 @@ <h2>5. Continuing training<a class="headerlink" href="#continuing-training" titl
 INFO:sleap.nn.evals:OKS mAP: 0.585451
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "8f2e64c8d4d6457986ee8b43b47e2876", "version_major": 2, "version_minor": 0}
-</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "8f2e64c8d4d6457986ee8b43b47e2876", "version_major": 2, "version_minor": 0}</script><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"></pre>
 </div><div class="output text_html"><pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">
 </pre>
 </div><div class="output stream highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.val.slp
@@ -815,7 +801,7 @@ <h2>5. Continuing training<a class="headerlink" href="#continuing-training" titl
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Model_evaluation.html b/notebooks/Model_evaluation.html
index 104094f4d..b7d4acfcc 100644
--- a/notebooks/Model_evaluation.html
+++ b/notebooks/Model_evaluation.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Model evaluation &#8212; SLEAP (v1.3.3)</title>
+    <title>Model evaluation &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Model_evaluation';</script>
     <link rel="canonical" href="/develop/notebooks/Model_evaluation.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Analysis examples" href="Analysis_examples.html" />
     <link rel="prev" title="Training and inference on your own data using Google Drive" href="Training_and_inference_using_Google_Drive.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -321,16 +319,16 @@ <h1>Model evaluation</h1>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Model_evaluation.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="model-evaluation">
-<h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permalink to this headline">#</a></h1>
+<h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permalink to this heading">#</a></h1>
 <p>After you’ve trained several models, you may want to generate some accuracy metrics to compare between them. This notebook demonstrates how you’ll do that given a trained model.</p>
 <p>Let’s start by installing SLEAP and downloading the trained model.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
-<span class="o">!</span>apt -qq install tree
-<span class="o">!</span>wget -q https://storage.googleapis.com/sleap-data/reference/flies13/td_fast.210505_012601.centered_instance.n%3D1800.zip
-<span class="o">!</span>unzip -qq -o -d <span class="s2">&quot;td_fast.210505_012601.centered_instance.n=1800&quot;</span> <span class="s2">&quot;td_fast.210505_012601.centered_instance.n=1800.zip&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<span class="o">!</span>apt<span class="w"> </span>-qq<span class="w"> </span>install<span class="w"> </span>tree
+<span class="o">!</span>wget<span class="w"> </span>-q<span class="w"> </span>https://storage.googleapis.com/sleap-data/reference/flies13/td_fast.210505_012601.centered_instance.n%3D1800.zip
+<span class="o">!</span>unzip<span class="w"> </span>-qq<span class="w"> </span>-o<span class="w"> </span>-d<span class="w"> </span><span class="s2">&quot;td_fast.210505_012601.centered_instance.n=1800&quot;</span><span class="w"> </span><span class="s2">&quot;td_fast.210505_012601.centered_instance.n=1800.zip&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -349,7 +347,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 <p><strong>Note:</strong> A test split will also be evaluated if it was provided during training and saved to <code class="docutils literal notranslate"><span class="pre">metrics.test.npz</span></code>.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree td_fast.210505_012601.centered_instance.n<span class="o">=</span><span class="m">1800</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree<span class="w"> </span>td_fast.210505_012601.centered_instance.n<span class="o">=</span><span class="m">1800</span>
 </pre></div>
 </div>
 </div>
@@ -534,7 +532,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Model_evaluation_14_0.png" src="../_images/Model_evaluation_14_0.png" />
+<img alt="../_images/d8517178d972bc0703903cfab70470e9c921794a84a0af190ce345f8321a14e9.png" src="../_images/d8517178d972bc0703903cfab70470e9c921794a84a0af190ce345f8321a14e9.png" />
 </div>
 </div>
 <p>This metric is intuitive, but it does not incorporate other sources of error like those stemming from poor instance detection and grouping, or missing points.</p>
@@ -549,7 +547,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Model_evaluation_16_0.png" src="../_images/Model_evaluation_16_0.png" />
+<img alt="../_images/d2b61f464140b9d880c782c1347c1961ff1a43b6b62847e515dadb3047025028.png" src="../_images/d2b61f464140b9d880c782c1347c1961ff1a43b6b62847e515dadb3047025028.png" />
 </div>
 </div>
 <p>Since these range from 0 to 1, it seems like we’re doing pretty well!</p>
@@ -567,7 +565,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Model_evaluation_18_0.png" src="../_images/Model_evaluation_18_0.png" />
+<img alt="../_images/b63ef4ea81a463798c2aa2540688b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png" src="../_images/b63ef4ea81a463798c2aa2540688b427d4c48b56a8bfa3f2b8e6edf7a74b37ef.png" />
 </div>
 </div>
 <p>An easy way to summarize this analysis is to take the average over all of these thresholds to compute the <strong>mean Average Precision (mAP)</strong> and <strong>mean Average Recall (mAR)</strong> which are widely used in the pose estimation literature.</p>
@@ -591,7 +589,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 <p>First, let’s download a new SLEAP labels package file (<code class="docutils literal notranslate"><span class="pre">.pkg.slp</span></code>). This is important since it denotes that the labels contain the images as well which we need for predicting.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget -q https://storage.googleapis.com/sleap-data/datasets/wt_gold.13pt/tracking_split2/test.pkg.slp
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget<span class="w"> </span>-q<span class="w"> </span>https://storage.googleapis.com/sleap-data/datasets/wt_gold.13pt/tracking_split2/test.pkg.slp
 </pre></div>
 </div>
 </div>
@@ -621,9 +619,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
 To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
 </pre></div>
 </div>
-<script type="application/vnd.jupyter.widget-view+json">
-{"model_id": "061ef3f7278a47bbbe199d38ccd6be37", "version_major": 2, "version_minor": 0}
-</script><div class="output stderr highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>2023-09-01 14:14:07.317060: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_UINT8 } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_UINT8 shape { dim { size: 4 } dim { size: 1024 } dim { size: 1024 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -2 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -2 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -2 } dim { size: -27 } dim { size: -28 } dim { size: 1 } } }
+<script type="application/vnd.jupyter.widget-view+json">{"model_id": "061ef3f7278a47bbbe199d38ccd6be37", "version_major": 2, "version_minor": 0}</script><div class="output stderr highlight-myst-ansi notranslate"><div class="highlight"><pre><span></span>2023-09-01 14:14:07.317060: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_UINT8 } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_UINT8 shape { dim { size: 4 } dim { size: 1024 } dim { size: 1024 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -2 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -2 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -2 } dim { size: -27 } dim { size: -28 } dim { size: 1 } } }
 2023-09-01 14:14:07.320224: W tensorflow/core/grappler/costs/op_level_cost_estimator.cc:690] Error in PredictCost() for the op: op: &quot;CropAndResize&quot; attr { key: &quot;T&quot; value { type: DT_FLOAT } } attr { key: &quot;extrapolation_value&quot; value { f: 0 } } attr { key: &quot;method&quot; value { s: &quot;bilinear&quot; } } inputs { dtype: DT_FLOAT shape { dim { size: -42 } dim { size: -43 } dim { size: -44 } dim { size: 1 } } } inputs { dtype: DT_FLOAT shape { dim { size: -10 } dim { size: 4 } } } inputs { dtype: DT_INT32 shape { dim { size: -10 } } } inputs { dtype: DT_INT32 shape { dim { size: 2 } } } device { type: &quot;CPU&quot; vendor: &quot;GenuineIntel&quot; model: &quot;103&quot; frequency: 3600 num_cores: 16 environment { key: &quot;cpu_instruction_set&quot; value: &quot;AVX SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2&quot; } environment { key: &quot;eigen&quot; value: &quot;3.4.90&quot; } l1_cache_size: 49152 l2_cache_size: 524288 l3_cache_size: 16777216 memory_size: 268435456 } outputs { dtype: DT_FLOAT shape { dim { size: -10 } dim { size: -48 } dim { size: -49 } dim { size: 1 } } }
 </pre></div>
 </div>
@@ -715,7 +711,7 @@ <h1>Model evaluation<a class="headerlink" href="#model-evaluation" title="Permal
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Post_inference_tracking.html b/notebooks/Post_inference_tracking.html
index 8360e8c8d..8cbd89ddd 100644
--- a/notebooks/Post_inference_tracking.html
+++ b/notebooks/Post_inference_tracking.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Post-inference tracking &#8212; SLEAP (v1.3.3)</title>
+    <title>Post-inference tracking &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Post_inference_tracking';</script>
     <link rel="canonical" href="/develop/notebooks/Post_inference_tracking.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Interactive and resumable training" href="Interactive_and_resumable_training.html" />
     <link rel="prev" title="Data structures" href="Data_structures.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -342,21 +340,21 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Post_inference_tracking.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="post-inference-tracking">
-<h1>Post-inference tracking<a class="headerlink" href="#post-inference-tracking" title="Permalink to this headline">#</a></h1>
+<h1>Post-inference tracking<a class="headerlink" href="#post-inference-tracking" title="Permalink to this heading">#</a></h1>
 <p>In tracking, we associate the poses (that were detected within individual frames) across time such that they belong to the same individual.</p>
 <p>Since we typically do this separately, it is sometimes desirable to tweak the tracking parameters to optimize accuracy without having to re-run inference (e.g., <code class="docutils literal notranslate"><span class="pre">sleap-track</span></code>).</p>
 <p>In this notebook, we will explore how to re-run the tracking given an existing predictions SLP file.</p>
 <p><strong>Note:</strong> Tracking does not run on the GPU, so this notebook can be run locally on your computer without the hassle of uploading your data if desired.</p>
 <section id="setup-sleap">
-<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this headline">#</a></h2>
+<h2>1. Setup SLEAP<a class="headerlink" href="#setup-sleap" title="Permalink to this heading">#</a></h2>
 <p>Run this cell first to install SLEAP. If you get a dependency error in subsequent cells, just click <strong>Runtime</strong> → <strong>Restart runtime</strong> to reload the packages.</p>
 <section id="install">
-<h3>Install<a class="headerlink" href="#install" title="Permalink to this headline">#</a></h3>
+<h3>Install<a class="headerlink" href="#install" title="Permalink to this heading">#</a></h3>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># This should take care of all the dependencies on colab:</span>
-<span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 
 <span class="c1"># But to do it locally, we&#39;d recommend the conda package (available on Windows + Linux):</span>
 <span class="c1"># conda create -n sleap -c sleap -c conda-forge -c nvidia sleap</span>
@@ -366,7 +364,7 @@ <h3>Install<a class="headerlink" href="#install" title="Permalink to this headli
 </div>
 </section>
 <section id="test">
-<h3>Test<a class="headerlink" href="#test" title="Permalink to this headline">#</a></h3>
+<h3>Test<a class="headerlink" href="#test" title="Permalink to this heading">#</a></h3>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1">#@title SLEAP and system versions: { display-mode: &quot;form&quot; }</span>
@@ -408,13 +406,13 @@ <h3>Test<a class="headerlink" href="#test" title="Permalink to this headline">#<
 </section>
 </section>
 <section class="tex2jax_ignore mathjax_ignore" id="setup-data">
-<h1>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to this headline">#</a></h1>
+<h1>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to this heading">#</a></h1>
 <p>Here we’re downloading an existing <code class="docutils literal notranslate"><span class="pre">.slp</span></code> file with predictions and the corresponding <code class="docutils literal notranslate"><span class="pre">.mp4</span></code> video.</p>
 <p>You should replace this with Google Drive mounting if running this on Google Colab, or simply skip it altogether and just set the paths below if running locally.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget -q -O fly_clip.mp4 https://github.com/talmolab/sleap-tutorial-uo/blob/main/data/fly_clip.mp4?raw<span class="o">=</span>true
-<span class="o">!</span>wget -q -O predictions.slp https://github.com/talmolab/sleap-tutorial-uo/blob/main/data/predictions.slp?raw<span class="o">=</span>true
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>wget<span class="w"> </span>-q<span class="w"> </span>-O<span class="w"> </span>fly_clip.mp4<span class="w"> </span>https://github.com/talmolab/sleap-tutorial-uo/blob/main/data/fly_clip.mp4?raw<span class="o">=</span>true
+<span class="o">!</span>wget<span class="w"> </span>-q<span class="w"> </span>-O<span class="w"> </span>predictions.slp<span class="w"> </span>https://github.com/talmolab/sleap-tutorial-uo/blob/main/data/predictions.slp?raw<span class="o">=</span>true
 </pre></div>
 </div>
 </div>
@@ -428,7 +426,7 @@ <h1>2. Setup data<a class="headerlink" href="#setup-data" title="Permalink to th
 </div>
 </section>
 <section class="tex2jax_ignore mathjax_ignore" id="track">
-<h1>3. Track<a class="headerlink" href="#track" title="Permalink to this headline">#</a></h1>
+<h1>3. Track<a class="headerlink" href="#track" title="Permalink to this heading">#</a></h1>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># Load predictions</span>
@@ -502,7 +500,7 @@ <h1>3. Track<a class="headerlink" href="#track" title="Permalink to this headlin
 </div>
 </section>
 <section class="tex2jax_ignore mathjax_ignore" id="inspect-and-save">
-<h1>4. Inspect and save<a class="headerlink" href="#inspect-and-save" title="Permalink to this headline">#</a></h1>
+<h1>4. Inspect and save<a class="headerlink" href="#inspect-and-save" title="Permalink to this heading">#</a></h1>
 <p>Let’s see the results and save out the tracked predictions.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
@@ -511,7 +509,7 @@ <h1>4. Inspect and save<a class="headerlink" href="#inspect-and-save" title="Per
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Post_inference_tracking_17_0.png" src="../_images/Post_inference_tracking_17_0.png" />
+<img alt="../_images/3f78e9f7b93df6d2f3452d162ca631a9de6f98afaf55b8f4f59ee8351db6feeb.png" src="../_images/3f78e9f7b93df6d2f3452d162ca631a9de6f98afaf55b8f4f59ee8351db6feeb.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -521,7 +519,7 @@ <h1>4. Inspect and save<a class="headerlink" href="#inspect-and-save" title="Per
 </div>
 </div>
 <div class="cell_output docutils container">
-<img alt="../_images/Post_inference_tracking_18_0.png" src="../_images/Post_inference_tracking_18_0.png" />
+<img alt="../_images/1b0b6e7baba606deccb7311077642a87dad38e9d08625d98d4cf0ffd2b414281.png" src="../_images/1b0b6e7baba606deccb7311077642a87dad38e9d08625d98d4cf0ffd2b414281.png" />
 </div>
 </div>
 <div class="cell docutils container">
@@ -616,7 +614,7 @@ <h1>4. Inspect and save<a class="headerlink" href="#inspect-and-save" title="Per
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Training_and_inference_on_an_example_dataset.html b/notebooks/Training_and_inference_on_an_example_dataset.html
index b1092b312..237445199 100644
--- a/notebooks/Training_and_inference_on_an_example_dataset.html
+++ b/notebooks/Training_and_inference_on_an_example_dataset.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Training and inference on an example dataset &#8212; SLEAP (v1.3.3)</title>
+    <title>Training and inference on an example dataset &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Training_and_inference_on_an_example_dataset';</script>
     <link rel="canonical" href="/develop/notebooks/Training_and_inference_on_an_example_dataset.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Training and inference on your own data using Google Drive" href="Training_and_inference_using_Google_Drive.html" />
     <link rel="prev" title="Notebooks" href="index.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -334,15 +332,15 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Training_and_inference_on_an_example_dataset.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="training-and-inference-on-an-example-dataset">
-<h1>Training and inference on an example dataset<a class="headerlink" href="#training-and-inference-on-an-example-dataset" title="Permalink to this headline">#</a></h1>
+<h1>Training and inference on an example dataset<a class="headerlink" href="#training-and-inference-on-an-example-dataset" title="Permalink to this heading">#</a></h1>
 <p>In this notebook we’ll install SLEAP, download a sample dataset, run training and inference on that dataset using the SLEAP command-line interface, and then download the predictions.</p>
 <section id="install-sleap">
-<h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to this heading">#</a></h2>
 <p>Note: Before installing SLEAP check <a class="reference external" href="https://github.com/talmolab/sleap/releases">SLEAP releases</a> page for the latest version.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -356,16 +354,16 @@ <h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to
 </div>
 </section>
 <section id="download-sample-training-data-into-colab">
-<h2>Download sample training data into Colab<a class="headerlink" href="#download-sample-training-data-into-colab" title="Permalink to this headline">#</a></h2>
+<h2>Download sample training data into Colab<a class="headerlink" href="#download-sample-training-data-into-colab" title="Permalink to this heading">#</a></h2>
 <p>Let’s download a sample dataset from the SLEAP <a class="reference external" href="https://github.com/talmolab/sleap-datasets">sample datasets repository</a> into Colab.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>apt-get install tree
-<span class="o">!</span>wget -O dataset.zip https://github.com/talmolab/sleap-datasets/releases/download/dm-courtship-v1/drosophila-melanogaster-courtship.zip
-<span class="o">!</span>mkdir dataset
-<span class="o">!</span>unzip dataset.zip -d dataset
-<span class="o">!</span>rm dataset.zip
-<span class="o">!</span>tree dataset
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>apt-get<span class="w"> </span>install<span class="w"> </span>tree
+<span class="o">!</span>wget<span class="w"> </span>-O<span class="w"> </span>dataset.zip<span class="w"> </span>https://github.com/talmolab/sleap-datasets/releases/download/dm-courtship-v1/drosophila-melanogaster-courtship.zip
+<span class="o">!</span>mkdir<span class="w"> </span>dataset
+<span class="o">!</span>unzip<span class="w"> </span>dataset.zip<span class="w"> </span>-d<span class="w"> </span>dataset
+<span class="o">!</span>rm<span class="w"> </span>dataset.zip
+<span class="o">!</span>tree<span class="w"> </span>dataset
 </pre></div>
 </div>
 </div>
@@ -415,7 +413,7 @@ <h2>Download sample training data into Colab<a class="headerlink" href="#downloa
 </div>
 </section>
 <section id="train-models">
-<h2>Train models<a class="headerlink" href="#train-models" title="Permalink to this headline">#</a></h2>
+<h2>Train models<a class="headerlink" href="#train-models" title="Permalink to this heading">#</a></h2>
 <p>For the top-down pipeline, we’ll need train two models: a centroid model and a centered-instance model.</p>
 <p>Using the command-line interface, we’ll first train a model for centroids using the default <strong>training profile</strong>. The training profile determines the model architecture, the learning rate, and other parameters.</p>
 <p>When you start training, you’ll first see the training parameters and then the training and validation loss for each training epoch.</p>
@@ -423,7 +421,7 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
 <p>If you don’t stop training, it will run for 200 epochs or until validation loss fails to improve for some number of epochs (controlled by the <code class="docutils literal notranslate"><span class="pre">early_stopping</span></code> fields in the training profile).</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train baseline.centroid.json <span class="s2">&quot;dataset/drosophila-melanogaster-courtship/courtship_labels.slp&quot;</span> --run_name <span class="s2">&quot;courtship.centroid&quot;</span> --video-paths <span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train<span class="w"> </span>baseline.centroid.json<span class="w"> </span><span class="s2">&quot;dataset/drosophila-melanogaster-courtship/courtship_labels.slp&quot;</span><span class="w"> </span>--run_name<span class="w"> </span><span class="s2">&quot;courtship.centroid&quot;</span><span class="w"> </span>--video-paths<span class="w"> </span><span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -593,7 +591,7 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
         &quot;runs_folder&quot;: &quot;models&quot;,
         &quot;tags&quot;: [],
         &quot;save_visualizations&quot;: true,
-        &quot;delete_viz_images&quot;: true,
+        &quot;keep_viz_images&quot;: true,
         &quot;zip_outputs&quot;: false,
         &quot;log_to_csv&quot;: true,
         &quot;checkpointing&quot;: {
@@ -799,7 +797,7 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
 <p>Let’s now train a centered-instance model.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train baseline_medium_rf.topdown.json <span class="s2">&quot;dataset/drosophila-melanogaster-courtship/courtship_labels.slp&quot;</span> --run_name <span class="s2">&quot;courtship.topdown_confmaps&quot;</span> --video-paths <span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train<span class="w"> </span>baseline_medium_rf.topdown.json<span class="w"> </span><span class="s2">&quot;dataset/drosophila-melanogaster-courtship/courtship_labels.slp&quot;</span><span class="w"> </span>--run_name<span class="w"> </span><span class="s2">&quot;courtship.topdown_confmaps&quot;</span><span class="w"> </span>--video-paths<span class="w"> </span><span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -970,7 +968,7 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
         &quot;runs_folder&quot;: &quot;models&quot;,
         &quot;tags&quot;: [],
         &quot;save_visualizations&quot;: true,
-        &quot;delete_viz_images&quot;: true,
+        &quot;keep_viz_images&quot;: true,
         &quot;zip_outputs&quot;: false,
         &quot;log_to_csv&quot;: true,
         &quot;checkpointing&quot;: {
@@ -1154,7 +1152,7 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
 <p>The models (along with the profiles and ground truth data used to train and validate the model) are saved in the <code class="docutils literal notranslate"><span class="pre">models/</span></code> directory:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree models/
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree<span class="w"> </span>models/
 </pre></div>
 </div>
 </div>
@@ -1190,11 +1188,11 @@ <h2>Train models<a class="headerlink" href="#train-models" title="Permalink to t
 </div>
 </section>
 <section id="inference">
-<h2>Inference<a class="headerlink" href="#inference" title="Permalink to this headline">#</a></h2>
+<h2>Inference<a class="headerlink" href="#inference" title="Permalink to this heading">#</a></h2>
 <p>Let’s run inference with our trained models for centroids and centered instances.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-track <span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span> --frames <span class="m">0</span>-100 -m <span class="s2">&quot;models/courtship.centroid&quot;</span> -m <span class="s2">&quot;models/courtship.topdown_confmaps&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-track<span class="w"> </span><span class="s2">&quot;dataset/drosophila-melanogaster-courtship/20190128_113421.mp4&quot;</span><span class="w"> </span>--frames<span class="w"> </span><span class="m">0</span>-100<span class="w"> </span>-m<span class="w"> </span><span class="s2">&quot;models/courtship.centroid&quot;</span><span class="w"> </span>-m<span class="w"> </span><span class="s2">&quot;models/courtship.topdown_confmaps&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -1305,7 +1303,7 @@ <h2>Inference<a class="headerlink" href="#inference" title="Permalink to this he
 <p>When inference is finished, predictions are saved in a file. Since we didn’t specify a path, it will be saved as <code class="docutils literal notranslate"><span class="pre">&lt;video</span> <span class="pre">filename&gt;.predictions.slp</span></code> in the same directory as the video:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree dataset/drosophila-melanogaster-courtship
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>tree<span class="w"> </span>dataset/drosophila-melanogaster-courtship
 </pre></div>
 </div>
 </div>
@@ -1324,7 +1322,7 @@ <h2>Inference<a class="headerlink" href="#inference" title="Permalink to this he
 <p>You can inspect your predictions file using <code class="docutils literal notranslate"><span class="pre">sleap-inspect</span></code>:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-inspect dataset/drosophila-melanogaster-courtship/20190128_113421.mp4.predictions.slp
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-inspect<span class="w"> </span>dataset/drosophila-melanogaster-courtship/20190128_113421.mp4.predictions.slp
 </pre></div>
 </div>
 </div>
@@ -1360,7 +1358,7 @@ <h2>Inference<a class="headerlink" href="#inference" title="Permalink to this he
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="c1"># Zip up the models directory</span>
-<span class="o">!</span>zip -r trained_models.zip models/
+<span class="o">!</span>zip<span class="w"> </span>-r<span class="w"> </span>trained_models.zip<span class="w"> </span>models/
 
 <span class="c1"># Download.</span>
 <span class="kn">from</span> <span class="nn">google.colab</span> <span class="kn">import</span> <span class="n">files</span>
@@ -1484,7 +1482,7 @@ <h2>Inference<a class="headerlink" href="#inference" title="Permalink to this he
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/Training_and_inference_using_Google_Drive.html b/notebooks/Training_and_inference_using_Google_Drive.html
index 0b51b9640..fa37c643b 100644
--- a/notebooks/Training_and_inference_using_Google_Drive.html
+++ b/notebooks/Training_and_inference_using_Google_Drive.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Training and inference on your own data using Google Drive &#8212; SLEAP (v1.3.3)</title>
+    <title>Training and inference on your own data using Google Drive &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/Training_and_inference_using_Google_Drive';</script>
     <link rel="canonical" href="/develop/notebooks/Training_and_inference_using_Google_Drive.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Model evaluation" href="Model_evaluation.html" />
     <link rel="prev" title="Training and inference on an example dataset" href="Training_and_inference_on_an_example_dataset.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="reference internal" href="index.html">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul class="current">
@@ -348,15 +346,15 @@ <h2> Contents </h2>
                   
   <p><a href="https://colab.research.google.com/github/talmolab/sleap/blob/main/docs/notebooks/Training_and_inference_using_Google_Drive.ipynb" target="_parent"><img alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg" /></a></p>
 <section class="tex2jax_ignore mathjax_ignore" id="training-and-inference-on-your-own-data-using-google-drive">
-<h1>Training and inference on your own data using Google Drive<a class="headerlink" href="#training-and-inference-on-your-own-data-using-google-drive" title="Permalink to this headline">#</a></h1>
+<h1>Training and inference on your own data using Google Drive<a class="headerlink" href="#training-and-inference-on-your-own-data-using-google-drive" title="Permalink to this heading">#</a></h1>
 <p>In this notebook we’ll install SLEAP, import training data into Colab using <a class="reference external" href="https://www.google.com/drive">Google Drive</a>, and run training and inference.</p>
 <section id="install-sleap">
-<h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to this heading">#</a></h2>
 <p>Note: Before installing SLEAP check <a class="reference external" href="https://github.com/talmolab/sleap/releases">SLEAP releases</a> page for the latest version.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip uninstall -qqq -y opencv-python opencv-contrib-python
-<span class="o">!</span>pip install -qqq <span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>pip<span class="w"> </span>uninstall<span class="w"> </span>-qqq<span class="w"> </span>-y<span class="w"> </span>opencv-python<span class="w"> </span>opencv-contrib-python
+<span class="o">!</span>pip<span class="w"> </span>install<span class="w"> </span>-qqq<span class="w"> </span><span class="s2">&quot;sleap[pypi]&gt;=1.3.3&quot;</span>
 </pre></div>
 </div>
 </div>
@@ -370,20 +368,20 @@ <h2>Install SLEAP<a class="headerlink" href="#install-sleap" title="Permalink to
 </div>
 </section>
 <section id="import-training-data-into-colab-with-google-drive">
-<h2>Import training data into Colab with Google Drive<a class="headerlink" href="#import-training-data-into-colab-with-google-drive" title="Permalink to this headline">#</a></h2>
+<h2>Import training data into Colab with Google Drive<a class="headerlink" href="#import-training-data-into-colab-with-google-drive" title="Permalink to this heading">#</a></h2>
 <p>We’ll first prepare and export the training data from SLEAP, then upload it to Google Drive, and then mount Google Drive  into this Colab notebook.</p>
 <section id="create-and-export-the-training-job-package">
-<h3>Create and export the training job package<a class="headerlink" href="#create-and-export-the-training-job-package" title="Permalink to this headline">#</a></h3>
+<h3>Create and export the training job package<a class="headerlink" href="#create-and-export-the-training-job-package" title="Permalink to this heading">#</a></h3>
 <p>A self-contained <strong>training job package</strong> contains a .slp file with labeled data and images which will be used for training, as well as .json training configuration file(s).</p>
 <p>A training job package can be exported in the SLEAP GUI fron the “Run Training…” dialog under the “Predict” menu.</p>
 </section>
 <section id="upload-training-job-package-to-google-drive">
-<h3>Upload training job package to Google Drive<a class="headerlink" href="#upload-training-job-package-to-google-drive" title="Permalink to this headline">#</a></h3>
+<h3>Upload training job package to Google Drive<a class="headerlink" href="#upload-training-job-package-to-google-drive" title="Permalink to this heading">#</a></h3>
 <p>To be consistent with the examples in this notebook, name the SLEAP project <code class="docutils literal notranslate"><span class="pre">colab</span></code> and create a directory called <code class="docutils literal notranslate"><span class="pre">sleap</span></code> in the root of your Google Drive. Then upload the exported training job package <code class="docutils literal notranslate"><span class="pre">colab.slp.training_job.zip</span></code> into <code class="docutils literal notranslate"><span class="pre">sleap</span></code> directory.</p>
 <p>If you place your training pckage somewhere else, or name it differently, adjust the paths/filenames/parameters below accordingly.</p>
 </section>
 <section id="mount-your-google-drive">
-<h3>Mount your Google Drive<a class="headerlink" href="#mount-your-google-drive" title="Permalink to this headline">#</a></h3>
+<h3>Mount your Google Drive<a class="headerlink" href="#mount-your-google-drive" title="Permalink to this heading">#</a></h3>
 <p>Mounting your Google Drive will allow you to accessed the uploaded training job package in this notebook. When prompted to log into your Google account, give Colab access and the copy the authorization code into a field below (+ hit ‘return’).</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
@@ -403,7 +401,7 @@ <h3>Mount your Google Drive<a class="headerlink" href="#mount-your-google-drive"
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">os</span>
 <span class="n">os</span><span class="o">.</span><span class="n">chdir</span><span class="p">(</span><span class="s2">&quot;/content/drive/My Drive/sleap&quot;</span><span class="p">)</span>
-<span class="o">!</span>unzip colab.slp.training_job.zip
+<span class="o">!</span>unzip<span class="w"> </span>colab.slp.training_job.zip
 <span class="o">!</span>ls
 </pre></div>
 </div>
@@ -412,10 +410,10 @@ <h3>Mount your Google Drive<a class="headerlink" href="#mount-your-google-drive"
 </section>
 </section>
 <section id="train-a-model">
-<h2>Train a model<a class="headerlink" href="#train-a-model" title="Permalink to this headline">#</a></h2>
+<h2>Train a model<a class="headerlink" href="#train-a-model" title="Permalink to this heading">#</a></h2>
 <p>Let’s train a model with the training profile (.json file) and the project data (.slp file) you have exported from SLEAP.</p>
 <section id="note-on-training-profiles">
-<h3>Note on training profiles<a class="headerlink" href="#note-on-training-profiles" title="Permalink to this headline">#</a></h3>
+<h3>Note on training profiles<a class="headerlink" href="#note-on-training-profiles" title="Permalink to this heading">#</a></h3>
 <p>Depending on the pipeline you chose in the training dialog, the config filename(s) will be:</p>
 <ul class="simple">
 <li><p>for a <strong>bottom-up</strong> pipeline approach: <code class="docutils literal notranslate"><span class="pre">multi_instance.json</span></code> (this is the pipeline we assume here),</p></li>
@@ -424,13 +422,13 @@ <h3>Note on training profiles<a class="headerlink" href="#note-on-training-profi
 </ul>
 </section>
 <section id="note-on-training-process">
-<h3>Note on training process<a class="headerlink" href="#note-on-training-process" title="Permalink to this headline">#</a></h3>
+<h3>Note on training process<a class="headerlink" href="#note-on-training-process" title="Permalink to this heading">#</a></h3>
 <p>When you start training, you’ll first see the training parameters and then the training and validation loss for each training epoch.</p>
 <p>As soon as you’re satisfied with the validation loss you see for an epoch during training, you’re welcome to stop training by clicking the stop button. The version of the model with the lowest validation loss is saved during training, and that’s what will be used for inference.</p>
 <p>If you don’t stop training, it will run for 200 epochs or until validation loss fails to improve for some number of epochs (controlled by the early_stopping fields in the training profile).</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train multi_instance.json colab.pkg.slp
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-train<span class="w"> </span>multi_instance.json<span class="w"> </span>colab.pkg.slp
 </pre></div>
 </div>
 </div>
@@ -443,7 +441,7 @@ <h3>Note on training process<a class="headerlink" href="#note-on-training-proces
 </section>
 </section>
 <section id="run-inference-to-predict-instances">
-<h2>Run inference to predict instances<a class="headerlink" href="#run-inference-to-predict-instances" title="Permalink to this headline">#</a></h2>
+<h2>Run inference to predict instances<a class="headerlink" href="#run-inference-to-predict-instances" title="Permalink to this heading">#</a></h2>
 <p>Once training finishes, you’ll see a new directory (or two new directories for top-down training pipeline) containing all the model files SLEAP needs to use for inference.</p>
 <p>Here we’ll use the created model files to run inference in two modes:</p>
 <ul class="simple">
@@ -452,7 +450,7 @@ <h2>Run inference to predict instances<a class="headerlink" href="#run-inference
 </ul>
 <p>You can also download the trained models for running inference from the SLEAP GUI on your computer (or anywhere else).</p>
 <section id="predicting-instances-in-suggested-frames">
-<h3>Predicting instances in suggested frames<a class="headerlink" href="#predicting-instances-in-suggested-frames" title="Permalink to this headline">#</a></h3>
+<h3>Predicting instances in suggested frames<a class="headerlink" href="#predicting-instances-in-suggested-frames" title="Permalink to this heading">#</a></h3>
 <p>This mode of predicting instances is useful for accelerating the manual labeling work; it allows you to get early predictions on suggested frames and merge them back into the project for faster labeling.</p>
 <p>Here we assume you’ve trained a bottom-up model and that the model files were written in directory named <code class="docutils literal notranslate"><span class="pre">colab_demo.bottomup</span></code>; later in this notebook we’ll also show how to run inference with the pair of top-down models instead.</p>
 <div class="cell docutils container">
@@ -469,7 +467,7 @@ <h3>Predicting instances in suggested frames<a class="headerlink" href="#predict
 <p>Now, you can download the generated <code class="docutils literal notranslate"><span class="pre">colab.predicted_suggestions.slp</span></code> file and merge it into your labeling project (<strong>File -&gt; Merge into Project…</strong> from the GUI) to get new predictions for your suggested frames.</p>
 </section>
 <section id="predicting-and-tracking-instances-in-uploaded-video">
-<h3>Predicting and tracking instances in uploaded video<a class="headerlink" href="#predicting-and-tracking-instances-in-uploaded-video" title="Permalink to this headline">#</a></h3>
+<h3>Predicting and tracking instances in uploaded video<a class="headerlink" href="#predicting-and-tracking-instances-in-uploaded-video" title="Permalink to this heading">#</a></h3>
 <p>Let’s first upload the video we want to run inference on and name it <code class="docutils literal notranslate"><span class="pre">colab_demo.mp4</span></code>. (If your video is not named <code class="docutils literal notranslate"><span class="pre">colab_demo.mp4</span></code>, adjust the names below accordingly.)</p>
 <p>For this demo we’ll just get predictions for the first 200 frames (or you can adjust the –frames parameter below or remove it to run on the whole video).</p>
 <div class="cell docutils container">
@@ -486,7 +484,7 @@ <h3>Predicting and tracking instances in uploaded video<a class="headerlink" hre
 <p>Let’s inspect the predictions file:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
-<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-inspect colab_demo.mp4.predictions.slp
+<div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="o">!</span>sleap-inspect<span class="w"> </span>colab_demo.mp4.predictions.slp
 </pre></div>
 </div>
 </div>
@@ -494,7 +492,7 @@ <h3>Predicting and tracking instances in uploaded video<a class="headerlink" hre
 <p>You can copy this file from your Google Drive to a local drive and open it in the SLEAP GUI app (or open it directly if you have your Google Drive mounted on your local machine). If the video is in the same directory as the predictions file, SLEAP will automatically find it; otherwise, you’ll be prompted to locate the video (since the path to the video on your local machine will be different than the path to the video on Colab).</p>
 </section>
 <section id="inference-with-top-down-models">
-<h3>Inference with top-down models<a class="headerlink" href="#inference-with-top-down-models" title="Permalink to this headline">#</a></h3>
+<h3>Inference with top-down models<a class="headerlink" href="#inference-with-top-down-models" title="Permalink to this heading">#</a></h3>
 <p>If you trained the pair of models needed for top-down inference, you can call <code class="docutils literal notranslate"><span class="pre">sleap-track</span></code> with <code class="docutils literal notranslate"><span class="pre">-m</span> <span class="pre">path/to/model</span></code> for each model, like so:</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
@@ -600,7 +598,7 @@ <h3>Inference with top-down models<a class="headerlink" href="#inference-with-to
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/analysis_example/README.html b/notebooks/analysis_example/README.html
index ce5164c35..734c7fcaa 100644
--- a/notebooks/analysis_example/README.html
+++ b/notebooks/analysis_example/README.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>&lt;no title&gt; &#8212; SLEAP (v1.3.3)</title>
+    <title>&lt;no title&gt; &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../../_static/pygments.css" />
     <link rel="stylesheet" href="../../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,22 +42,19 @@
     <script data-url_root="../../" id="documentation_options" src="../../_static/documentation_options.js"></script>
     <script src="../../_static/jquery.js"></script>
     <script src="../../_static/underscore.js"></script>
+    <script src="../../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../../_static/doctools.js"></script>
+    <script src="../../_static/sphinx_highlight.js"></script>
     <script src="../../_static/clipboard.min.js"></script>
     <script src="../../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../../_static/togglebutton.js"></script>
     <script src="../../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/analysis_example/README';</script>
     <link rel="canonical" href="/develop/notebooks/analysis_example/README.html" />
     <link rel="shortcut icon" href="../../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../../genindex.html" />
     <link rel="search" title="Search" href="../../search.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -163,6 +160,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -385,7 +383,7 @@ <h2> Contents </h2>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/notebooks/index.html b/notebooks/index.html
index 115900190..ec1e476bd 100644
--- a/notebooks/index.html
+++ b/notebooks/index.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Notebooks &#8212; SLEAP (v1.3.3)</title>
+    <title>Notebooks &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,24 +42,21 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'notebooks/index';</script>
     <link rel="canonical" href="/develop/notebooks/index.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
     <link rel="index" title="Index" href="../genindex.html" />
     <link rel="search" title="Search" href="../search.html" />
     <link rel="next" title="Training and inference on an example dataset" href="Training_and_inference_on_an_example_dataset.html" />
-    <link rel="prev" title="Running SLEAP remotely" href="../guides/remote.html" />
+    <link rel="prev" title="Using Bonsai with SLEAP" href="../guides/bonsai.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 current active has-children"><a class="current reference internal" href="#">Notebooks</a><input checked="" class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -343,45 +341,45 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="notebooks">
-<span id="id1"></span><h1>Notebooks<a class="headerlink" href="#notebooks" title="Permalink to this headline">#</a></h1>
+<span id="id1"></span><h1>Notebooks<a class="headerlink" href="#notebooks" title="Permalink to this heading">#</a></h1>
 <p>Here are Jupyter notebooks you can run to try SLEAP on <a class="reference external" href="https://colab.research.google.com">Google Colaboratory</a> (Colab). Colab is great for running training and inference on your data if you don’t have access to a local machine with a supported GPU.</p>
 <section id="basic-usage">
-<h2>Basic usage<a class="headerlink" href="#basic-usage" title="Permalink to this headline">#</a></h2>
+<h2>Basic usage<a class="headerlink" href="#basic-usage" title="Permalink to this heading">#</a></h2>
 <section id="id2">
-<h3><a class="reference external" href="./Training_and_inference_on_an_example_dataset.html">Training and inference on an example dataset</a><a class="headerlink" href="#id2" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Training_and_inference_on_an_example_dataset.html">Training and inference on an example dataset</a><a class="headerlink" href="#id2" title="Permalink to this heading">#</a></h3>
 <p>In this notebook we’ll show you how to install SLEAP on Colab, download a dataset from the <a class="reference external" href="https://github.com/talmolab/sleap-datasets">repository of sample datasets</a>, run training and inference on that dataset using the SLEAP command-line interface, and then download the predictions.
 This notebook can be a good place to start since you’ll be able to see how training and inference work without any of your own data and without having to edit anything in the notebook to get it to run correctly.</p>
 </section>
 <section id="id3">
-<h3><a class="reference external" href="./Training_and_inference_using_Google_Drive.html">Training and inference on your own data using Google Drive</a><a class="headerlink" href="#id3" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Training_and_inference_using_Google_Drive.html">Training and inference on your own data using Google Drive</a><a class="headerlink" href="#id3" title="Permalink to this heading">#</a></h3>
 <p>Once you’re ready to run training and inference on your own SLEAP dataset, this notebook walks you through the process of using <a class="reference external" href="https://www.google.com/drive">Google Drive</a> to copy data to and from Colab (as well as running training and inference on your dataset).</p>
 </section>
 <section id="id4">
-<h3><a class="reference external" href="./Analysis_examples.html">Analysis examples</a><a class="headerlink" href="#id4" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Analysis_examples.html">Analysis examples</a><a class="headerlink" href="#id4" title="Permalink to this heading">#</a></h3>
 <p>Once you’ve used SLEAP to successfully estimate animal pose and track animals in your videos, you’ll want to use the resulting data. This notebook walks you through some analysis examples which illustrate how to read and interpret the data in the analysis HDF5 files which you can export from SLEAP.</p>
 </section>
 </section>
 <section id="advanced-sleaping">
-<h2>Advanced SLEAPing<a class="headerlink" href="#advanced-sleaping" title="Permalink to this headline">#</a></h2>
+<h2>Advanced SLEAPing<a class="headerlink" href="#advanced-sleaping" title="Permalink to this heading">#</a></h2>
 <section id="id5">
-<h3><a class="reference external" href="./Data_structures.html">Data structures</a><a class="headerlink" href="#id5" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Data_structures.html">Data structures</a><a class="headerlink" href="#id5" title="Permalink to this heading">#</a></h3>
 <p>SLEAP uses a set of core data structures that contain labels, predictions and other metadata. In this notebook we show you how to use them to develop custom analysis scripts and applications.</p>
 <p>We demonstrate how to work with these data structures by interactively generating predictions from a trained model.</p>
 </section>
 <section id="id6">
-<h3><a class="reference external" href="./Post_inference_tracking.html">Post-inference tracking</a><a class="headerlink" href="#id6" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Post_inference_tracking.html">Post-inference tracking</a><a class="headerlink" href="#id6" title="Permalink to this heading">#</a></h3>
 <p>In this notebook, we show how to use the tracking functionality within SLEAP to re-track existing predictions. This is useful when experimenting with new ID tracking parameters without having to re-run pose estimation.</p>
 </section>
 <section id="id7">
-<h3><a class="reference external" href="./Interactive_and_resumable_training.html">Interactive and resumable training</a><a class="headerlink" href="#id7" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Interactive_and_resumable_training.html">Interactive and resumable training</a><a class="headerlink" href="#id7" title="Permalink to this heading">#</a></h3>
 <p>Training in SLEAP can be done via the GUI, CLI or interactively in Python. Here we show how to use SLEAP’s Python API to enable customizable training workflows, including resumable training for initialization from existing models.</p>
 </section>
 <section id="id8">
-<h3><a class="reference external" href="./Interactive_and_realtime_inference.html">Interactive and realtime inference</a><a class="headerlink" href="#id8" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Interactive_and_realtime_inference.html">Interactive and realtime inference</a><a class="headerlink" href="#id8" title="Permalink to this heading">#</a></h3>
 <p>Once you have trained models, you can run inference via the GUI, CLI or interactively in Python. Here we show how to load trained models, use them to predict on new frames, and implement a basic version of a realtime SLEAP tracker for closed-loop applications.</p>
 </section>
 <section id="id9">
-<h3><a class="reference external" href="./Model_evaluation.html">Model evaluation</a><a class="headerlink" href="#id9" title="Permalink to this headline">#</a></h3>
+<h3><a class="reference external" href="./Model_evaluation.html">Model evaluation</a><a class="headerlink" href="#id9" title="Permalink to this heading">#</a></h3>
 <p>After you’ve trained several models, you may want to compute some metrics for benchmarking and comparisons. This notebook walks through some of the types of metrics that SLEAP can compute for you, as well as how to recompute them.</p>
 <div class="toctree-wrapper compound">
 </div>
@@ -402,12 +400,12 @@ <h3><a class="reference external" href="./Model_evaluation.html">Model evaluatio
     <div class="footer-article-item"><!-- Previous / next buttons -->
 <div class="prev-next-area">
     <a class="left-prev"
-       href="../guides/remote.html"
+       href="../guides/bonsai.html"
        title="previous page">
       <i class="fa-solid fa-angle-left"></i>
       <div class="prev-next-info">
         <p class="prev-next-subtitle">previous</p>
-        <p class="prev-next-title">Running SLEAP remotely</p>
+        <p class="prev-next-title">Using Bonsai with SLEAP</p>
       </div>
     </a>
     <a class="right-next"
@@ -474,7 +472,7 @@ <h3><a class="reference external" href="./Model_evaluation.html">Model evaluatio
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/objects.inv b/objects.inv
index b465c43e5083a67708a27c069ef8a086b88b6df9..635ee932d010617d202b8f65be1a7b20823183fc 100644
GIT binary patch
literal 20650
zcmV)EK)}BvAX9K?X>NERX>N99Zgg*Qc_4OWa&u{KZXhxWBOp+6Z)#;@bUGkYOhrLZ
z3L_v^WpZ<AZ*DpuF)lPNF$yCfRA^-&a%F8{X>Md?av*PJAarPHb0B7EY-J#6b0A}H
zZE$jBb8}^6Aa!$TZf78RY-wUH3V7PheaW)qMwZ<^zanJCMQagJMY1ThwaL_$%w&?;
z#U!(HU`xG$Ap#x|;6nf?3~%77e?w-{&-ItcN5BF22)~Y(3npIx7w7nJ{rdIkdRw5h
zKIG-OI;2H@DR=eLKi#UVFVOBUyR^>#+|C3J*X~yQiHYpL+U;K~1n9p>UbbCYo={R(
zdD*oD_uQAKF0abn*9K$<NLa9>DwA8)b_xD{#MutQzMxJ*hR8ZWyrHyA(lWzI%#w4}
zzyXi>PRc67m6hgAIP4q|Af+9;)kWF~bxP8vNuSojo>AKM4NBVnazQQmx;5#&Lfq5}
zRN`-%^oD+Hs=oe+I>L1-(zacuSxe>1fpcFZIW9Gtv%0Bj)O1f#3F{hR7453zLLSiq
zMcuW^Vv?fAj*wEuF}v}3X*<VsTjrHzGE2LZ@TFbX<j1~4(J436AjXDEcGEmgt8o)b
zt}9#7wJElpD5{da#f2EjtP$vsU3$8Ujej^cd3%=}(@RpF6D}Ifo{Y4ozG+a|B|`n<
zHMfy7atj-YxPh)7O4PL1ypCU-XY3|m&iF*64;Xsc<lIz|dxI%6HCv(@H>6AQ?6KX3
zvq>M4{FYv}XoReD#jUU1q*+|6%l3w5@Q<Qr?~aUEHptN}y~D3)_qJ8r!|2;s&WXv4
z&Q_l`FeGwKMbnP=X@jjYD&DYn@ADcJd5I!<4;nUNzc$VwZ46s;HhsA@W!GaJYRNcr
zK70>MM3Ej*VVnqH5~i^~PD7KQj38XZ0p)N?i&M{LJ<_~+$Xk@0s<tyvdra_kz}2lz
zaghv@bVc6U=964-35>@KjSJ*0?M_$lgsM>Kd=)8VAb6CG1X=Dhb?JglQi&S?pIxNL
z;h^wXtb|WhUxGoFRSAOH_P1NwJee9P7cn(bA%eTM4wqYg`iPpGRj{U?nSi>XO(v9+
z_EZr&Ht^^y-dZA7%(Qf)vtV7Ti}ZwUXaln71=r-{-L_fl8thZ$<pm^n6zqh5W2e*9
zFl0GX>zK%$bDmm@`CEd|P#+pRh8Z({{mzusg|I2TOJP#PD8n8@lH#rbj}5UwDKC~B
zXgHn7h+*H1tU8_g8XJ`-x~{~#Y;0Yz(L@nxnHFr3$y=J2agZBytL{;PZgux0GDK&8
z;uar!a%%6Bb8cIv4C5Ge1B8inUfPz*UTfPXfBja1QFX5}isAFa5qEAGN$IhAG)%O*
z!kw^*Z>l^4fWl>6B)7b6aq6^I+!>5a%9E)&h=dW{F8f>k1duEY#f;4`JO~0Dx_D{4
z5d_FQ_ck0S&uF>;%D+tR^EM|7>f^G1C?x4>pl|?_1)^kN$(buwARIQ5WqYogn^a&y
z8NJUlRD~z3{h1H@0_-Zq;9CGf^Tn`;8`uCrRpDC2kl1P#3>JQj>kI~qPE4UJp=KXQ
zBe43{FsWk>(^{;bS>E1=tgX;e8G_Dz73h(bRdPek1te($E=5d^sCz($RY)J?bG}~Z
zCO}v0ch?Qd@>3^GP6e3C1v+&EBVqTry?_JL^+hPGGDD9b6%}S1=1{N&HtcRBGw|2y
z0SK2h8`>yakc^iRRlSU+=w%!=5Av)gNOe(lU_fIP!9Yf<@X45e0v?=b4&9err9ePq
zsVBzGOEP?h2JTC&_nFWKv7`q94Ah+7i_zQ)|AT&e$hxN*nHKY%Q1Z4+SGyGqBFGp}
zh?inWw6-uP0Gkd78n;VebT3GT&S_t)Hc;<{0STOLTHr`oZNU=9KqnJyaIj%YqQ(T6
z`B@^1Q93yuK^8)S2kc7<h6{igWJD$);nrjdH`Y&Z&4gQsCDH*qn?ZW>sWI#zcKJ4R
zW{{evVI}uOYRxdSt^up?D6_m5Lj%QZ`XHd2Pag%$3nVk32?E&$%J(U&)4&DvX%GTG
z4_j^&)iT2|036aRv)VyVxN)}-5MU=X5CChK`3uD&+n4RWkq@GA3eX$KWsjE!D4kS<
z#`Kx@@xs;jzaB7TvrvbC^PHjfggYIGwuT<-N<Y=%4MY}5;lRp=BM2Bh13|i27&L52
zAWF<MujmProb#rIN>hUS0x&F4J5CX)As-jXVPNx?Y-^J&2?VThGR*_Xu$kEpB`LA?
z!d3}AZwzYOhJ7mIK0`1>NUb5LeJPfH#xlS?dInVv)X@MgUow|JEK~tt@jVkPtuS7U
zaVFmj$25=c2gC;sKq_gS!}P(zb<h`GK5@Z=>tM2`X*#y#iwUD5MU`g5{;0qLq+>(Y
z)K1tXKf(+Ym_v$(aUCA%p(Y_?>J|)_4u?sz1M6iN?yxl_5UoU)w9D^Nf)(;wH9!hv
zoE7uH@mi_y!maL6DcWFoe$7kw-az!c?6?yc29qtyfT^M1Qjb3{RM#iN0VWS{8igay
z)^uP%L48P$2{l7AU(t)9MSQ&sx-BhN%V2P!Cal^N#2OJ8sLq>V(al8keEc;#e-{BZ
zzg^m{YCuv9K4fa|$Sf0Ui@|-E00Zi>+q%dhJmy)6g(s}5s7nTrG>r6n3t|Z|QY{^r
z4n?D5-c~xo1_%0()=*$T^{E<*!+3SBdl-BnCqt}Z+8_cp46bFrvo8`*wLj$DH5s!^
z;t3W2Lxk7(*Q@k&MZv3M0}}8u*%TMZc^X04v@$z}Bu=xm@B~2a^9W0(&9H0Hp=NAZ
zz$vE`6d7nbXbKHJRXK%c6CDvrU<X<Y+!!qg)=eI9Av?<vwXg9T1dqtnxuqdagKLjA
z?>eQ%naop>KU>p=2Rr0s&d(o)oc&2VY%S9od-D426H__;E{d7l%Mn-EU#MUm@aQH1
z^w=5u@U_^dv`iXgok<OWXD3t%ereRGwH$oQAt|C5C6raD<-s&K190*QKB<C8(6spB
zvDsrXfl5{stJ>;puO+&sWmcdl(53OP)_UAHwFz92AIbAykH4c+_lfUh46|BAeNEYf
zn}cbUgzffK7Jf;7d}2Rb4eo+$4U3l<hGFzH&Tyo@CL4y?(~tv^dYX4YRyQ*NNa$!m
zU>H|-5HR#CI4pk7Pl`2SF4s8Y7ds77JhJCZh6t0iT^c-kaTr2rNuP{n=rL0oW^St-
zyE+#zY+KqANHcP(PL<VLAVO9zfeNu<?JNAJqi--myJPxz0;}--`u)`up1u4k9=|?)
z9ffmWx*pz5<yYamTgFxJj%H{THbkNG-QvdP-`?9>KbCp*@i*62FA(ieUX9%Y*0YU!
zBKtON4|vZO?St&zkbO}d+ph<#N3-=nb#1L4kN{cF5UPvxc<MWATOCEMO;!@+v(fkg
z{_9oWy=w0rTk@Q5HIYv{ha>f~0<gTEHUf^^&w>E*I@=duPPZlqN9t~y;5gThx8Ty?
zXg%Qww@{}4o!*LVjG@G|p*56QH#LWI;%)7r%)Y5XoDgqi5f`MJm_(U%Gn*)rZex&S
zM$pQvs}|kmZcZPUb@KXz?3j`F_{rb$&cn`4_|_8nnmHI!e=7&ayTZu9ko(&>AYLyM
z2gvDe;b2Jp3>*xnhkb(~1gyxst;wpN-MDc<t7RjNidBLYa}@PJJ-oqx{xlGn4CEZ&
z8TfE2s@ATx>+yxjvuI2J0ezX6BEkc(<n3_GBqoPAR1l3LIst|)BdSy9)}6QzTu_Oc
z{4|{7FtuMv*4lt&*&o^8|APN^bafU?{uDk=4g{T_djrMj>DWNf`nfb9jLyys7@M0L
z13~BRz(BA$x-Jk@n^hQKT@8mQO=ge?S?wPcVh-JJUprd85!e;c$H=Wf_p@m$2|P{M
z3Kl<$wT8giK&=IL)A1GP?&fC&xTCdMfgG!_A<CJb$8&P2Bw4FTk!4xBJv^!fkvD^n
ztpQ;3vpqnBp0)^p)z3CT5IWl`C^EM;1;9pw36pm;bio#k?$#54%+Z<xFm0tCkeY#i
zk`3)x*lu)3S-a7dWLXA&#!ueS2d<`fPPdZB%K$@hdYWNKT3=%f#qDX5!D!tLGbp8F
zBZT7gFqu%4u7(qe53UkLuymt2BSKa?MukMebZ3YcfLf$MX@a22>@lgu%)2RzLeQGc
zLGYK?|BEc5IGX4wz6Y?6O<qspXN1FeJgx3}GCwmLz~gLJ13BCb>3R}(%ekJz(NwM{
zh+Gk1b&x;WjZ*$-w?|m!m;BMy3<*R#l$YW0fb}#rp2)t&#RJ~cocJKS8xUVqM|Jgp
z^)M11sIF$g0}{MI0m_p8PSP?hp4z;%8i1*|)fUXeX(WP|{FkGFP%*Bg@-ZNgw0;H#
zn%UC;LDKsfC{Si+0|w6O)+ixq-3=fltD}L0q};qhVe3-tVKG`*B4o9zROl1_$I*z6
zpe|@WcG45s&lIiz_q2AN41Na97u?xqtpRpZWlv;x%i{^`Xk0u|V-#*sxfbY>@`#K|
zo50@d%>jQ;qvc6!21!*IC%t4?E#7!>0t7ON3D|deVQ?`A5q*9~5z)e1lFoDfn4GXK
zUX1f|Gkh&@nQ<@@Uie7fgrk!4anve_J9O4Uj&UoMWk_G^%a6T&-QR9czX>2^zail$
z0DgM?<~xw@52T<=%SAEnSjZF?*BQpVG1?VoxHUF3Tj+K~S>_x&+8MS?k60bss&7tG
zIMh&Bdk|_@q#^WfcwLaYtA=8v2UbB+ct!c;+Cd<ZYxu?p#t?{>J9OQF8q}KZq;rLV
z)-xF|N$=CVAnFW({E%nN;;#dGd%Jujw?~B7#|Z|2jhx|y7`Q1^pplz`g_@2jTCn+;
z!iBc=DVlu62m~IK{g|Q#+K?%j`Fw~EQjkTOLIt+<ci2NtLmK)B0MN=Mu)uaHA)RZO
z6ooc!2`ac{OIV@JT7n8}EfP|Yf=f7o+WiZ;fiFb6M*slrmcRnFTS5xeN(m}hD<!N@
zt(2hhRvvK%!v!k2gcPLY5>B9!f5atIptod95TfdH7yyM|KnAM)0xwkg7l6S{V1XFg
z5Eg)e`n$jj(&7bLpe~0fFxOggNz>ZwaIMN8`#KS`W&t>4XS;nK-x<n$S)elv3wJfl
zJ^~=#NQVu8W4SQb1t+)x&tO_8oTN?9s0hrR7<45Q)e1}_ME?pqhGS@08<J21&u3b$
zX-K!jK)MQ`0b&sQ-JFg<_PZC&K*N+1pMnz0rMfwNGa6lR8Ro`TVbFD`#<DWNge>a=
zT!<FczYVDRK1Bv~L8I%meT&=S`BtNgr3SpwiK$4AeO_d!0mi0XbzZ?12M0ZznzQkm
z#2|3FBaOsRxADM7KQIn{J@|2P@CU~Z7XH^7m9KtVt&VM<n$}~vfXP}g0h*$BX>*5~
zPgk+Vu1S#u>4tS^dgHvI4G{Mj3AXpAd$VP=!H}dddlG`t=6@jOLWUvHE33u5uMYi&
z4R7o`JTb6;(em6Vl!(98XpngOecGaT?C<w!)8(hUPT5tPOC6dZE}=f46Uy&(9hT$A
zY#}+3v$oW|Rak8xSlzeR<d}vg?jtI*m1V>}J+%0LN!m2&Q%DLJjNx(`4N6%a(&lu<
zDtqdB+^7zx>UPsc9?(c@lu{@Cw!{sD4%IsJ2BqYkx&V!It9TCWDK#u@#vsI<RA%rM
zB2adyY}r0WSV~l$Dy(WrhUouYb!7ILtFp{Ld(Hm;eNnNUDGutQWL`<hVs#${yP}1~
zrSI!Dz10P|6V_X$LKE^sxp~pcl+RG~6Z*QUif1Zq%Ke;m{(5Zd+toTJJ@CiLnHjcL
zDg`(59VtC+$8U{)mQ;w>UrMQn*Uwuj%tqi894qZi)Fx?=d&g%=0+)HtB(xp6JsbaH
ze(P_~-7KoUtNZS`Q$ZVL`)vKjf#|^(W2tCjHa1fqbL_RDS5<kxrTa8vJ*1~Dzfb98
z#CEDMPsEX47+hA(?Tbi6aztmcg!N^V4f&U!zj(%Eobv*YC);OC!5wN!R3v0}?-^6E
z{fJ|p0<!3M!oQxgE;b(csOUu_{Su1eS;k@WgpkF)W1enb7N3U+i0wPF8TTP4vtD(=
z+vPKt#IqTTc{)hvQ9f5wF(u|XyI5nop3x(5OxG9Sq!?5BTn>XJQ4KpH`sF3<Io%s$
z!oHM#<<Q>cb&_V;myr-$I3+rL5$&#P(y}FS$1+-FUs`&;Boj4H+81Q!^-22zP5nHv
zUy4bcr|V0ys`IpcF@|-^6~t@RHy1iOa9kuFK5$(g9zS?pJ(O#PCNs))Tq|ZwV0irp
z9cy@hlIt;2nGPIl*Vh6`anZ7Zj9rN%=gK9sFjA(eNnV~Bl(uNI%x=>{BtCK^0h``)
ze%YjHte%YsWA$Fy38h1~Wb3M2vtR?l=-EAnarX5#_U9wcTkH;U;Ml0y%Y0<ePhGA}
z>ThJ;rYj!#_{#Fu1?;&R5KgwrQ(Z&5ZfYBVX`@C0HGK}jy-i~g6C}zl_#q5<oY^^#
zX^61%c!mo9lwYo>`RT(isB_iAk;Di6@XDv&S8^P4ax5@yl5yy{1#8uH1uA&}yU1i@
z$g*l;WHsS*xZ>y&tt89Kyvx%f$)w>Y+1gcak<&8ah07Yx7L=@Kbv8^vEQ#k)1Ru6E
z`@P4`(72t%(GGf0jIC_6=I|^noym&BK%Tmqa{AQKlGCT;TBldP{NiZDM|elvwNoW4
zh*nLHtmI)!Q8*uCjYlI~IS1oU<li@}=17{fLve}xCs+JQ)(nVOv23Ytn2!0~RS}_*
zu&M8w6nk5l*!}H1^{)8T)zqmz{fuqz`{DTyrO){hZBii8FdirIZjh1FwZ^53n<VK3
z!52|PE6h&Z1g$!55<5xj&OO9U)yiWKaT7I4elZ$8>GR^15&q9(27^hv_6A(~18VUo
z=j*?G?P&C65)bslaH?dXVvZHMnz*U{K;;HRtJHKY*=)O0y8-27M?&i5f32eY?wgn@
z`Q10qq>_Yq6P3h7+o>cb+gc?t<;E)c-CrVV<ad909)%>lTj(R^*-RZV*T&k288=qO
z`<OlU{nzWpIV-uIMe*;yUNcZx&lJCjVV+eZoz-0HhA?aCu+Lm1=2(7ET<a$N1Ie~2
z#?6fI{pMaOVI-E9H|Y)NHnaKc#4)q^Zp$iu_%t*h{)4PSVf8yjq?(wh<XlGOVSgvo
zcEloLIv20NI&yoS99AEcr1zL%=so5asO)>TG0115PVbuN3HXYgDl3URH&z9(tlClK
z&iC1U+ln73H&)mP<BrDnG?pjT`ZXa>qLoXyzAUTOZ#_v?uBCdi_^;cnBk)~C^=0uN
zB6u>aI}YVZa?7(!-kK4Ih3?0Wt-h3GlR{gwq1}^W&EmN~#k!?zH=fC6K_@3-_`=zh
z7{3@dsD2dn;?jT3C|7Z<G6f*SEJoYv1J00%5QV;7=GRUM!LYB4(g9D-jvjj+ijI(s
z72SHYM0j`oASqc9@&JZ-eMo^Q{LiS7;V0drrsH|^66TmaOj<%tuNzZvN{iwsj#8?<
zjX^~3UW_8|(ylJ5uE>uc(|g0tw;8A#+My~FR9B~Kg33(aYp%m4KltpKwqREx+v6Ip
z1E!aEoU6jNSIrTf%;~)%w@lO(c5r_!f4E0w7g~c#&;t_CQPNB=DH0b4#$f=}L@fnf
zSV)fOl9#bb0P^v^sV>;yv>$NEG}vzc`@4S))E)&QNimBW15zwThB#$1>Q2lYfN7@&
zA5b-o3;;R!L~jK7zDk>Hw4kQ@=pc{}MRmz?2)k?|fvji~4_xPHV(HQg1Ow3<?CrcF
z^W{}hKo}w=Ca9rMGqk8nie@D?omn|#X_p$DyHUvC-bq5Q@jtJIIU9A6&M~!MZ_Bnz
z%M*IlRCUcV9~1a+LS*7D!Sz<Bjm7^6=E6Vig9SL6BtspW%EFn;GTUimcjBMP?Qktv
z)yQyIzsXpIUh<q=v0KtcNb{k+Lj~$=+a=b5<B?UD9G{ZB%%Tu<sO{o{)9e4@z;mM(
zFa3fbULNk#BF|_*LUIkVeXcm1CRSld6rjcRo^5eK^zv#jJE1^5VBeu{P;zO~r(r8u
zoD2x6=5mZkGVBr}`)OkGl#|_ekyvoILoP^ln4k*d;kTcrWtMQtE^RKTi%7``65H*;
zP&}4feqCvXU0O!pk4jmcKPL4(-)834$|Yo1$Z70MTlUuwh`<Liu*dUdG0=NR>{rwk
zM%mz2;+MCyZht~)23Wgxngf1(tr};K9E!H6+BSJ`ZqQJ))%i>Yb!_?VW}O8ECU;)@
z1tunyOShW0Te7zK<(7gHHdUU*Xd*M*Xl-K76RQ!-dE#pQf-Rbw&zXQVjHuGP2u65`
zd!V3@%~8P-MY|L{k?LE)6rsL<L#M84qEz_;I!>)GX`)s9l8W3Ec0S<&?UQZ80vuUi
z=3R8II7&?jT0lLM@c8U1%95OS$uUw*S%?{q0)Fd@E>CEwBqYXLb?fWsL|s*red-Z;
zqE-kQz7{9~hZqS$s_PWfX;VD_zFXM&5TaZ92yQB76p;+ljKv|adl<Z1gmWK`pMQAy
z9#BU|IB@k`Lfg02t;CO@kP><vU6crb)lfqAI6mOR!1njRX`zI7kk7wECl-UZ<TR$(
zm=uNMO1djJWYVWnukCy4B7CkP%UuJ`T=#^`a751tm5RjlsjA6!K2b_+g3Sx;o;lXB
za}ue@<jF0_;+o=alK4ZNo+A=On7g)cL>cAUIO9#ES(&4Bxi;1a6UR16*g2l}I0G-w
zXA(>P=}nFb@aba{!2x9#z>+gk#b^yi!Z-?Pqy$!tc^FJctx#hR$L`(E8bLBbk7XSB
zU<WulQkRIq5Im7|Uc=@X+Q^5BMfw3$bQv(lpmTyBr}!rnj+e!(#=pmH!f^|mKR7f0
zaBch}om!0S=GjQ1^lVnrc#fW(I*OKCX(PCLR!-Os{WGj-8{xnw0+-OR{p9DUHd#h@
zjl;4MKTcba2!JOVC&=(q-Jj)-D#|?Oc=8#tCCxy2L@nsTnX`DD>JQ=pw+<QtU>h-K
zak!B&2;VvONDzxxjLsmYwE+s;NUOTbZ}}goIPhTHLo5qhw-Qf;U-lQ`-OwLyEq=+h
zg5+>)h(A{%w~fZooC)a_@6+OWxbt#bZWg()?Y3v6<$OtRZcZLFH!kOmxz7bQKG&+r
z|G<gZvWhQh-?n*LCS{eoZF7jho7_-ZJ{x)4Wf{8Pjvkx;j6_gts`^<-J&wb`b2}fA
z0i`DGYeUkib9A@>`N;7=Vl`7TbbbzQY&AFLc0L#fNJq+ebIN=}<4C>JVpIJcD7P(K
ziFV5Z?#uj)hm~9RxD(p#{n#0OR9E(gv?-oGc0REwU|0?{QKz)0`NN~MF5NDN19y!@
zkmH74AnF$XOolW4B$5S7OQa<Hls4Jl&=cKv_D)1!iBriss>2n(AV-){+CGu_mAu+G
zvut@IW*4YsCdPy2|Bk<HH%|d>U(cI@sw~K))QvuiME)~M8v_3UH>QuLD<U&9(KKC8
zMoYWkubt--VUr+?+o`Gw*FzU^QV`p;J0u~pg*T+2!9PwYzJ<3XfrWKR$rjN_Dfm_T
z5v|QFNe-#xI!Qs+H_73FMMZO+hz8_D?Cug86K^|eL|mDq<-W{@yd<0Lc6^hbGH#cp
zU?*Gh^0e5b2T{;ZXtBS<jLCNVq`!DZ9p@vh=S%{3mD+26J{Oy+9)FLv>m+E<Xw+?4
zyqK}5(ku}HBe!8?wTY4={vWwRY12L<wnEd*BdvY)l*!o_PPyC|#@T(yDpPo3AC4qS
z;1Ep8S~ek6=pHps9`z8yb<Rs%|J&qCbY9!A5uZ9rmHo>853F+&)Wb>6P`uP1?hvYz
z`}~iDM&WBCG8NT@O!1uB`-mh>U)nG8g`r4XZZv>lYH_y^ok?63pu3SQn4xvLr6Gug
z-UT6bI5BDN^R`b5-u=a7>(Wb9cp8iRNMeykWceenH>j9DZhr?wm*TVoTl)p_<h~jb
z+_k>O%?g|HJ~!y^Jt|L6k;$-CA@`ANlCg*_b4m6<qV^}abAu)N_jmsatZ@S(+nYtB
z>ysPD8rR)_K}0qFADK76BL;*8BN5j9vB=wN<ZKC+zRrk?8HuIskL?MABI-2_B?-bd
ztS<gu9jya1S=c%{n+S{R9R8&`x+(LL%Ae!lI&_C2L!^>w*MR6CoWzj2I=L`}<c`wc
z+sbswO$@pu%TIjfAUY|X_1P|u&ZANgwoSEVzGr?LQg`1HgpUpZwo?}>o_HKv1!2V<
zPpmA`<~FX#o%@nqDiK1^cX?r-XQWcrevC;~C92L2>60)gU^Ftlya=gJ-9l8K#HHxV
z;@A(rTE1b-aIu$$Y#oabviK&joKDwf!&1^2GP3fwVK&*F6_~R%P0c`avXjE?#(CwZ
zOt;M}-L^{iQj%|4cDDeeHhrLb=IMP^;0<zueQ-UE0LEoMeE>nVQ4PH*@@O^0v6;~l
z7jZ?+#=-IU69g4*sK@k}7kT#-ovVGi;WBITll^`u2)gLHM~|wF2AAD(sB<<zDEbV9
zf_A+b**7;tV>I|RGii%!Gnbx((4)`arS*@fL`}-Ktm#{<k~x&}kQT*Qq;@TglZbPH
zSy7p=ASut53=bD2P6BxTlkVD}_F5I$bCraT0~0j`4j~;PCc++0CL+17#g3DThrGK+
zZ<j0x>vlHLz&G4S@GQCA%o3W?5Z?@%*tfHR##F>NfDZJLt)D3caZR6zc_X`LO2Q^a
zZyC&NBkSivyqO_#N=$r<<dlr<Op{x3;@c>nbVN2*PDzPtwVaZ$k@<2;%O-a0mfq1I
zTtmcp<A^+jgkQ3>JykL<NC;l89!X`LbC4u0^9{1(_jDt?r3+a=J7KyW2n(Ue$h>TW
zZfRa-F6%KXX*{}?l?*;+dL@Og%~?s)mrZ&BuZf@2+c8Vu(2#_n$|?i%kcY{+jGF$4
z>NKH=s8bW3$cpuzABqC5_!0KtdJHL$CSJ8mBoQlJVsSdC7RV7_t;nVfc&%cn2lPk+
zce+fMucnBqP@E&KK5>%BRs8|gD247g1gOh_>>&JjwLDNntmJ{Cp?}0HrpP+<ksVou
z)gb{IwFeqV;wjQV5mAu_4l=F6jt`?0k1z5LHTJ!iL5!R)9zkT?{ahX+4D$xmh>A6G
zMAWR2Btkuq%eY*&3=DFX!TTt;ytK-3MtDZGl4F?sJzdkXM1?~S2t@3mZa|Da$_Gp7
zI$sLJ>NqtD#m2XWZKpcE4cvn^%h@a|jq8V>T4w+~urJug&|hP?Iw)v5kvTC^pvY{B
zRNMjz3Jn8#p}J96tTUIg*k&R1H}v!x-RCFtuDZ8%x)WqqNFNipj=;y_tz+<Qita2f
zvxL5AaY}r}M4dNyymJ0PPvwCXG>3~*(gHkAQ491)WzA+=-=#H3Q4eSbVma+yuQ%*a
zfeshD1u}qD;)p$JNVMcfP{I)$S_hk|7Xc1U9V8dC^^ii0{*lOYq{eQ13h#_dCkKMT
z$ue!ncl#TPK!{n2U=XAC$-D34nZo<v`s`K=VIfb5ujRA#C4kSv2n4YBwAFAL*Wq3O
zlj9ICfWg-S`BD9bNZ{Tmjb;WNPs8R!@sw=N6s>O91|MMp>=u0u@IB280CsssjfjD5
zvgkt2X3___AB*ak7VlAd_h#6UV$jYhz6tnmxc=IAB#6TM29Yok6&&`@)hSAv-&Lo(
zXa&NH0cEe~!B(-y?H8=tBo@+^OkQ>ZlGn>@KoWaf5j#1bJe#;G#-=&{z}iL^9}N2<
z`4cw{Rr3ZA9`K`U)g+ErVMj~GiPkm=&o3qDd6MED?H~rbS96X*?-h?AMiTrC3p~Ha
z!axjqfAbcQhBM~m^ZLp0SUo1UJflDDZcYd-SyC7tvj*QrVQj_=5}fQgmv=VjLVl=A
zm(UX$x)nU*;M%vjhBNyFz<}m;42A(nyQWhJCiMt{1IfCk2LfPS+3^H3)2Q6%o<tcS
z<-IBQ4UbX=dzOsHGB!=bx(+v(c6~<|Ox?3hH&IKO8Y<GKBA>3!n}+t_lgGnGzF3rO
zQiM9}DBB>5r5O8dz}1)W*aWgmaJq@C#T}W-x=unU7~G)`sBB5->~m4MY^w#~x-{vN
z{nAJ{zUFd_&E)P?gz<Pd5&%RcVi$o#Cv$gH!iao2$S@|~K5+#bR=0bcdZONBx7hJw
zU!z>XGPe(r@FzAQ5{JuZcs)9~<%~5?T=jPmiMTm~MrZfD4>`D?cqCX`2m5kL&u27P
znfQF&c}udYIb8wy*s<`WK^GFJ-|l=Y<Sh=!Ql7equBXv8bZ$KD9XwB+*N9yFXIlpt
z<SgYg%6(sTU3L3hv(PTYH+#?DnVq#6(>jWAlSG?bQR=Wk0*F?`DA<yeNV?oY3rI<o
z{)l_+q?Rhx(0c>Y4ne!OrzRJgZOSTw+-#{%a-98`Y8A~KY}gt%YlEA%yv^I&maS~N
zw)Jc+>OhWcz~@V>-o4_@3cLa92o95JW3uycv%E!X*jmDv4ctr`v7wuZ<1~0JXS9Z|
zXN+tBz+2*w^0cl|nZZ@c*sc(}g)9OzooYN5)#<K{g+3iN!4l9W;Wj6>2Ht0eZw-Ep
z&|QRN<8<EV8EV=n4Tsz#8zIrd9A~t2-={@RZVQZHKbJubqjdUanrQv6g*Ij{>_Q!}
zFLt4i(<57$qxH)+-pIYPg}$Y)rG@QSKJKTgeqBA35hvud42Q|Vf(^44aSlMw8Ovtr
z8Dp9&3u{D6WnqqEr1YfG?2~~qvT4H18D@^wxohxawh$%XCdO6C9JyucIpelVJ#VyD
zsichACe?&-TBJO>F)hOhIomSaAop{6L`C}n*&}eyuw6k#08VmqTJCgh)mkDt@|G{D
zHmOz!M_(4?8OoBjzg$pD?{<tz*rH~7@mEwPk-KvvPC-SW%iDx9vdBbaqh><HDu-Ez
zm5el75%te!VrDhPBR^@7q-8d4{*O{#UNb=}IdCt;r{SW&Pa95BmBmxM`j4!>^jZU2
zUhw(&jU*ko6Ob^Y_5m`_&;`^fYZx$XP^O5}%BKu*8@bFCsg27Nv{^)}af(j#DlP*+
z=`;v;a`qE=Ld@d94Px)<k;M;)?wzx=-YK>-sL%G|wG=*zxR%3rOR+Q02l5A4;d3;P
zE%E-)!x=BPt?=FiQL!Uvqrmn)0q4ySMbQO#on~bFtZVope|&a6I!?Vho@iNz8!64v
zOv*<L|IQ^^_!>kJW1@DM6j?<O!FEmHW7#j6BszWsJ0|1Dvt=?1@m5($xgLQJq;Wj)
zzKSJ^ah3_97*LrYhOv|JqZvFFe<RH@GAE4*q)yCul);%`(Ko`t(m&AH#DY-8q`YN}
zAjB+VghA|Y=;>W*kH#^Db;0woVJpx{@K+JV&aBHTz<t{K3V7d!;!M%jMcyUjWTsK!
zR-q=%FcTC3oIXwNS*-UU{JzTz+lmO4v9uE=BX%4n9V%OwC_ECh4wD}^Eb_JkfqSgU
zj1abBPA4QC5HD%pws~46b#{ipqqAcdY+;>)L^^=dt1Lw@VK_im&#UneR$+to(bsy#
zW3Nxu4cj1;kV`8PY#f^WWGpcW)Pe>8!y{7T-VN7gb%IOZ81D)jnsB)44j@hS=?*S~
z_j#Kib8PdT=on}{Ao;gXI~0yU&(Id@aj1Jn7X-1BmNx-~MD7|4FX|h;)EFc|3~I8x
zx(=v8b&;M(de{KP7lmw6Fjy4&T5nO<>sLG&&3rrIPkaM}()*K;Js}USF)9Nn<cqCz
zURnseeUnB2y3<5R0De)mZSr6rc?VFq$3zbxag4euG341_JUQ$k*X<dQ&+i1bsW}v2
zTUtQFw4o&|MB7=Q0=1b1E?iq#Ktr~X1$eOD@~Z)Y6jsJKq_|(ZAA)zSRHz9efPzl(
z$S)sb^0gH)Sl1_T5K2A)1k&yaCYV}JP(k#W-KY?((h>}uR?DD3iamk!)$MOwKta@7
zMugIE8Ivmc;8(*HDPZG~Vd->;*n1380A+tm&ZneEpKNCa*8%rs-htvmnUH{n4%WDV
z8Z@LP!>AcOS^_v^)Zzh@o&h`5gz2FHY?&S$#`x&*p_-tcVa@36TOWc@m*H1>4&(zm
z=OwyfA3H#|RZswGt%eQNXeBCuGAnUc=y9I?H%D+NC02s3(ciy1VpG+1yo+x_@ZFMN
zd8c-OBCEjwR9Occs>~W(a<{Zcz5RD*V1Ra_L58vT8gQtVqJa-J#p>hk0eY+l1khy@
zc&I)bumM!hfDWU61_CG*G*GND+5hOpuwK#aH`UrNsU0L}i^$JUNsZDwacItRwWHzY
zMI)}mqP+_=yZx8NbfD;K0qqw=I@}F8iikjXZ^e_J$uL@Lc1+jh*+37cv?%mVHw^lg
zxHqscCJ`OOcEBcxg>C&4dJ>_)&Q<e}He`EY8;gQ%gtKp94<zZ(rAZ62!hE_8x0}Qv
zyCYYfj@t&}kcjj7QT3}%&s#MZLe-wChFO(p9I*=ww?)ShOAt>pjFe(>bbV8bz#eUM
zBBBAfjn>yKn4)CoooMoqogxK29O+~18x*4BjgolK&UXyE6UA-N!EcfWISxLoMR8$_
zWs~*Xw`bst<3efq;GHQwRRp(6OBTn$GSJ0zwY0QxoGvY09QR907tOEGGDh(*3~bET
z7);!jr0@=of)dg<(80&<`54I{3hs+YarzpR<zz+$yu=3N!!fRX8i^dHJHtunEEGNW
zgED3YBlIjI-KYfdOs>qfO`oD*jBE!b#&&uWoo?1WoN0?1+n~de$fT%<5oyMdTR<ex
zRxYw_li<Z@6x_5aIH1*aq@ah#YE-60_vEswvz8~KZQjBnZ)JzKvc9oxu4HM`$Am1h
zCj}}m-8E^*Ho7P|p-~>>gIMqkd6^I+%VufgaF8iL%E6*N#Nmk5Gezx>X1x%I+dUka
zz*-;J1IbVzqmU^e>`c%UTfXfc5*T*cWd<F?3{NOvr&A`-(57kv4m&S1g$AAgoq)T?
zA6o+sZSp4QpndcV8G3?736KAV#{RLX`ZD{}^z1w{kY$xn!Hw($JAzFFod{Edf|kh~
zdblW&cb0+0@&+3yv3KgdSaS#{kp&(|VJJ{xyW%AQY`3YxkI{E7d0;!xB|dCNr@#m9
z?UoGSz0#5bv};nqL-$V$4!DlQ1Px*=Ko|KVEa}f6d4#ERQKb;tp+=v~F2FLt8p^1v
zsf%q`u)gjbzCm?U;ju*vrZEsc%gmubk9E~_V-drplN11%D~N-dRAoYEmqX#-Rrlx@
z)PY(JB_@dRP$L7|5F-hsDNz%_S`;-0tYJ}eQ1nFHWfX-1+JwznAPkNYGfHEKwJ(74
zKrNHa;b2$E7U&UHeHI+>i#|&V&^3|;e2m5#ak78O05{hUrRQVfAf1evVzYW90sC4w
zWLRC3SI6#Lu8H)81#+}TK<v4a`oek6t1P?#KBOIbRTcfM4AiI*2q&@R5spK$JUu)J
z>I0gfL3%$k0qp#kkqg#evrxsDH8WDedWuFW*ts+_74Wp0kqp{PHIqSiwMI73nLP^=
z^dz5#4#MX*LB}+HjG(12$0*fM<A5nCl$)eMg&h^>NW^v7U>oRcL~#sh99L7v5n)87
zqk<n;>6zkpt+993Y1u&ic^ymSo+rv!?^9L3t{zHI<FCX7vHBWh;5BhI%g~z)NAqB4
z#vR^dWej2sUJnde2}*c+Zi>es0diwabg+XQ0~72J$4C^z3L6Mvhc^Zy*ujm7sP6I`
zE<qZ@hF`@MkqvktWuOA@UkqHJ-LjEKI&dikFN)6yH;!76#pq3=p0%`c#OYzRe2`AO
zjVpe?4mD8J1H+CL4LqYFK#dn&$fEU55y!y}TpMI7YD0=TAU3eUkBN<Bn>b{#N5&>X
z(7~~R2z2$!z!Tq3ia8Y5p7;fs!nBb9c7VX$Ufj-ZgDx(9#}>aki{6Pv?Yg$;ur}zX
zHtL)<>XIUNL~*;Hcn6;s1hIOxn0;A{9_+Wg?b8CPbD5#RD#;uiXrIm;woT8dc;dJH
zcEPFUhWAwD#-PtWSk|j+-Oy-cJaJ1-7}N!ki{po>PwS1Z?x+lJF)bJd%&y#pDa9M^
z^&n90c*lnSy~nE8t;Hm<?2+NtT=rfc9h|@a^{2nR`?+fMizhOU3BWR=<_zO>1AsHy
zmBeZ6ssyF&lSBLtXIy%dfPG=Z8ISaXG_7~&$su}%J8oV=;*zT|g0w{dh#M-b>l)60
z);-A1NZ41v<Gl6CHvsi@Ms6U--i@hyvB3@3y<hKcNx=(WXEC&3<A6YEE$C~#v0$&w
z^YRje?pr(tJ>4BEi?8GY<YXmWfQGxrgOcOA9O!~Ib`fDY!zLxa^*5<A)r|~vXe^}n
zxE!cf9a&1RK@+JTvJ1I#q(s9dfJ%*8BD+E!G8!5TyZ$!&62L(tDMbTsPH{5u>8VCw
zB~HL-S(jLaq9m0T1UUNpmv24}w-p4J#DL}<l5QOfjY)=!qm6~#;ew9^9s}2@H&_SF
zuJoBqQ9#1q=nT{;Ewh~X3K8&qL)fk4!G5O4lJI&G*BEx|7|JA;ik8%f#X)v(?AJll
z8&@3m=&VI|y2)%MxNXvC9W2$nL<BknZ?0|CbSO(A7<k+y-@b<F?I_d%zE5Vj;7$*@
z?d?K$E^*(gB@+S@i>@0@`Ot`A8gT)*6{Kze6h4_G8}l&hGN1@1iVXrf?LNoY6EVQs
z^i~%L3Z0I_Sa;JOTToPPoN!AASHrnO0bAk=bl4(SSl0-!utm~Ubw$kf=sB!KC5Dg8
z)tg7#l9XYzElbgOfX8F?Xi44zSr18<6fQ+|B#WrFi!OtJ6S0aKj{#T7<r)wEA;*3<
z-^LBfQdFlX#V_H;DSep$zVK!AXk{<9<ukG@Z^o%C2o7oMzy*(VT0m&p0bSA_gQg`}
z%qtOC?Y6h7!Udk+E+M7eKD8aX=}jPeuQz`D{SmbtuAfUzYY5%v*e4T<1iF;T&?-#q
z+@;5YcHn72>fToAHbCqng5nV;4kIO(x^IJV?$f-W$%8;U^-Y7yj$nmCVTk00ZdLOX
zf<*3j&(k9R199yL$|Bi?gOfGtZUE4jvMdCt?Qgd<tc>~=B`2J%sw{$W&Uw=YS3dy)
zfWTwm^-1BIa_ES!DiNQ0QQNAetln-ZWZyK{V7(*mD$=jH>NtGB@pD`v*u8J^N8`vF
zgpAGCicy5i>?#QDHEpRTb5%TDE7B+2#^Tj2kb$sEJaJ36FBw(lxU+68={qru0@cc4
zm#>&eF}2avBy1uj;UO=mL~?9wuJmfU9(T!SQb%qB{ve*H4Cj>WoqI~E<L{*W#I6!t
zkT>irG`V@cYc=blz49?KG!b^TQT7y-;@X!NTu|8ej>$5t#pG2V6+dppm$x=K<Gwly
zcGcXNla<8G&0SjD($elBFgd$-z~zV^%*@u1c&UT2rCcU6v9S3*8aSAR4Wz#ZbU>yC
zk^PWKi?C_V5;Asc6TYlFaZJMaj8khu!cqrB;`_8kpT@n}`VLS%!W~|{DXNE8!Vh|v
z!owZT_)*ec^K<9yX8Ds4WuzQ)pCdjCQxcRmMUI-J%WnvrH@{sRwtB7vCu`ac>$fY>
zt8+qj02Jt+AJvIK@Q{}o(J|X8nw1Q7(~~8b=-Sn$_Zk}L95J)aiVXz+UHS-Rj7F$~
za8D-YKCH+k^s%RnfZqb!R1Ge*NDVqeYNd{fG-HROrmi#)`tmp0)q*3D8L)fom}S-x
zxfd#(_m<96t*@pjE`1yZ_%SDosDL`l0D;WH6s!eLCPW%3#w+}|1OyrfIj5?(oZa-t
zu6oB~iO$dvS*Jmk#k|}twVgfm!X2vz1n-=e^ayVoXmySK9Y{(C+ieT0<pqiz#}Ym8
zK_HL|DZrcR*yCXrT|yJB01RBy^@CsoSmOy4l$$0%`!fG~&-)>=(39X_JqqpVu!}5r
zZN@RO7Ah}oqhl&z933+WaqBTXhw_Pr3ox1{3D+Hv&E}6w`-bvchtGl+-l3;B@cgJN
zDE<(h1is{z&{66@%@m<F)I@;wGF;HggXV$NiU}C9PK-QoHDaUztdG}emwrHLhMG30
z_$3^-qbI9Ci}YxWui^{Z(fu$3C{RJluy`z-W_AzH51v1ak@6{YjMUR{EZx)-A^i(*
zs2ofKXoFcTJuIZnGO)l|ECUIky;?*a1FH%nGGuL-iQp>4%mZ5|csx*ocj*RP;QG+g
zfYyeV1+p$6%AF^RL@zmt=~aM4vFVP!4(q=MFnd5cA2{4=q`b3e!9Wr2k9uRI3m-Hv
zziQNOx~Vb&E=s^g!x3k(Bp@K|fhl~@Xr%c3kCdN7liCdoJTnl{fFbLu#tBR>Kq@cd
zrzOWeFChFl0ZgWcM5HbdK-?2SSvL9r5Q_NoP7(Q5A>ZeJz@tlkd`#>(X{SN;;Ellv
z_f8$+u|}JuCmS_wvGrVV>9(o-hR3YwtqujVg@kj#Xj9TJ{Xt_I)b;}^PLQN^CZ{Vp
z-O*(M`xZk-$T8$xI2La1O@0CbBkLf<$ek-;H^VXYy&M=eSW*=|W;kkx%C>5bRoY}Q
zfD>B*1Nq18@4?ybySy-*;GnPdiyrK?IQ(zfZ7X3ZK-h8iv}w|(^&o7A=C~ZV4y8f3
zd=uu|4%a}ggX-9+>$FMQ{+6&{3?dY=FCy;3^c96)>Z^&o6p>z&_6`-OBi(I-?sIHA
z@*8*JbO{=;7ohp=!iN+7-_}LdO>Wd#34zlP@^xuHXW<ADk_i%PCggM5*$u__LJu@L
zC0=W^YItpqsy0PYeEv(7g?xIzgT<0VD12x)3_98#Lh-DIk#?Z3#xTe6bi=SkUbCWn
zlr&>Sy@<~bN#5dq^i~&qUL-I%htYK9*&_sCnz3j@YpZMm6rzR;pcVwFeYz1(^C=jN
zOy@OYjE^9!jQk(Llsm@)DSD0tR%W;oPfcE%H~6^2O`>8cXKy^Q27?CyMq^54=rOp|
z1|0)PRhyjqq5uG+K`%;TR*(Fu!Js@cI0&7kNPJfnjM<=lp&@phC(`e&c4xaM5>?kE
zEobrJf`I9n8mRUzgAFqoVTesyUjfvY;!H9c2uPOXK<xK5(0y9~h3f?NSyDXi$78?~
z;^TtA(wccefoSwr2NLAunw`H3&MifQLjq~;?B0qr+yc=<4=2*`0zd`%_5+Tv3xP^N
z!%sqW;1mq310PypV5%0g6QK@NgXlgX7<(gwwE$?OE|bZKJOq!P<^!`46b%ju#76MW
zEI&cH!~zh(f<bwp6UPD~iMYLkczGBHqC#JzQ<dp9v-qOmq1OqZkx$W<)1@KUKa>#p
z<tPVKuE;N$X4cI8kyn~|Kl+N8V$x$|9iRA9PCJe$v(ITng;U@8e@VwX=e!_cOrMgr
zKc5Y85%efD7EO`2D^kD%5ufu5O~*O_zHwqD0%>na*53@X=H_pSgSFyY>Q2#x+Kk!`
z{Glnpnii){n-(Q~lbbJ4(|%8T5voI21x3H)2SJC-`FTa+dlmwbgoPW<MfxE;u=tM!
z>+^5MXuMO9SHHggA^G)9^6J-LUVnb|Y1ekyKV{WvxBq%Kdquz5y{by)GF1)!`yLbG
z^u8jNqB@h8S%2#0pod%bC*f(4)DRBp>gZS|(l5uqRluh)LHiR=(4LsCeRbZSSb`iF
zonY2x9+l^+IkEdn?MacPe;EH(0atyquM0YCw&O~|zxK`UJ^fRMipQ#TRkrytFY=Cf
zFmCvZ&j{0p*+U&CS@O!#CUYel<q6A2@k<3yO^gF9llw^i#W25o`tbJq&!67@^5e8!
zjL&C{f@}=ws@=z*e|Y(R@;%d3qf*X5|Fl%Txt`78r9@b<C0RN3M$~;*?Wqm%fge9|
zE2S)NPd$y#+uCSk(bwbO3&^}}dt^o9Qg7-178o`5xv_Tty(e*$Y*16-_rzWQz<(`(
z+wKWFLzh0!>`9!+ek4D3KhhsBuz8nN6#rzJ`I^M3Ax{17<@>jT)IcHRa96O9<lRM%
zEzlt?@=G}*SiEP9^QUvDGfKO@Ay)ugF1SF3dR7A~q!%;_nK>yxDx~LO0t;yNc#bA^
zTIu1q!~%~$p96}xvU)%+u>j=H=YaHFvIq~LN0r4E(DKVU{zH>@XgGnNT)#7rEMyl<
z3yBFG-%Ft(U~i|8?3oNwk^TPncxbDE8QxEThNn{~ve&;wdMN&W2F0IFq0S^?UUx{d
zv?e}<4sj|r1D?KA;}`Wc7020aB$BL#Ns>^~ovKr1qaiYCc=ARK+umDI+j|XYd#}XI
z5UOz_Cj&<d^eCxj7~gA&>Z(O|xf2r?)l+c+HCNu&bO>N0;&K`m{#Hpr#u``bZxzT!
zGD@mx#`j7hy2UI%(fJ??lO(F9k_41oblhhlAu?)o@<s{EkCM7%K@wGyNCIju{F99T
zu$+}^6xDEz0!l8rS=&NFWYp;7jS{xZ9uuP7HfSNIh8J(u6#1hSJ%7{y=Z{KE`a4O>
zw0LUs)&@V7Rx?dyl%yScw3-tlqb4A4rmzLNq^Bn>B6~dpV-Ke|WD<t$0@gtd5;L^n
z`4sRTd%5~Pg}$8u(Ps(}>TEP-C?=x<WN+rMWEXa8fTZsiaP+AJbwh2NUNm~oU&<i-
z(TmBZ%(OXOaS~6&Bc{8kZ#t`^5tbP=A|^&QZQLXopL}<to_%FZ!%Wtcm}}KH7djR0
zmW)B(n1DgbNZph=7f~&1rX&(3x?xmqXQNL|WaO!dt8NfGMU*AHE?mo3@ue{(3VBn)
z@PF=w&V-HhOvJ?4yYuxnw%*27)2rdc?2ZHD@PV0k5Sy}(r$BUcgTxGfcs>K9OB{3{
zq=y`3=Ah#nB`iCUYr!4wW#1`ZDUs!KOEv;oR?Q<zsOjk0qt?pA#55%Ar4oNQdii^G
zw2{cNY9?7iNoQ~aW<qL$xif&@Q0++j3z-!+*UCb2#<6&>AfoBW5Q_(2w`Q)Hu!?Rb
zqG04x1xaONlLS>v(rX0^-STM9(IBQ`7+xyy+2NqVjLwBr*!-0WnP2&6ftD|Z+^OFw
zz)96~<hlYQxR6qTi+3vU@{Ag!oz7Z_DdELSCBBHS$^(G4Hln$_nsF{=q;9I(UONkM
zGcgxZ@QO<(91OukRK+(DP;jx3G4d6HM!to(igO{P;1!3=Ewwp_sd$E$D*UD*qnpNt
zFT|Ai;-vx~=PdujsuDs<g)ZJHz=uGaM)1jT)t&mC0^HntjUp;oMbAYdy6EBP+Zhsl
zHU+t&)X=W!+ZhObHU+VoIgN!=8#lXM)0s4lHRA{4b!hDM42nIR;+VZcl{={L1LbRm
zwOEV0h<D0YFSPe|>XGM{d+wC4UMTL`(_%_j@$F_0Us!JD(PC;Y@!1&!UKnnivGPb4
z)9f@aUnryXnvs{KEFQm59)LHYsZ5UA+L#TE%El<n3pvfLgAk{6ck+`L%PoCsfXRlx
z-SpXu>E0qWz_9CRH~xM>&aPvX%l33yd$Dh!BO7gUZ2$he5EQG{@7cZYxZB(dnVpb%
zajeQVovnDY#q&S2R|4fT35K%!o{5tij``k%)3vI|t35RetrBS_BOqVV(M#VDVD{z(
z3AUm{DDv^E8`fH!1rkw)Tv^<l(-XF7r%FudFFvls{5}5sLtpL&wIn#BUB(sriip0D
zHag8RMEB}T=^w;iocgw_Zu^PGFA@Q35R~sSl^JkI*cqCf9g`Rnh&=|f`167fQzqIk
zp7C7_{MR5#HclMTe8a@#Ble#BGx=&lrulM^QSzRP9JOD6YXT4={QtAhpT7OZ{4o)r
z|4u%<pyrqt#)JzI{{PA6H*(Y8^<?^-i0rWl7hQf{ovoi^5ia@yJ=Rsz?I%kjs<tgJ
zf?$pLfqq8Ie)pzob}z*O2p$l!<0Mn{<4>8Uyd{6Wz$&TKQT#u_^5f^XFI1auxa6cQ
z@I5vY(;+MC(xzxF%2br?uOjzm(v4DAv}b#xTdU2|eD-R1yu*)o{PB{NRSo$%Wt+41
ze4cebI!q>d+xI_+_qf^9jPLk+W($`Iq9eV6SnWM~h}9<B<GheHj^=64cQ@M$#u@1C
zASkMd7|cO%Cj^&irJ=I~I?KIlwHZF0XbdYeZS_rHh<2&A<=N<^?=0G?>ISE(uj&L<
zU7$vswF4pS-P<a@_2Rxh)u^dcvV*@vjR*1k!=I+V4V3h+Jw58+%qD$9<G=36pD#uy
zLRBx^wa_{ik{qoBDzzt4`%%j5wdg)_qvT!4o*uw36HP^V-@K#0=Ah`<UoKitWdIz+
zUW>nGK)Q}(fFJ^6S0X+e+%9T-Cc_RblIEMYg9k^-@{<^eO!=@th8ioyDE2}8H3MQM
zN!}}`PjB|!ZX&LP%JTG5R&AG`Y-s<un}4AMWIc}!?Ax6ztpvgyMU75Yr2n>?h%2Gk
z=N6!^capFYhE<OZ=fCe{X(bT5vd4n+*PSG+gUL%9%r`qpcm`u0F9?i5>XR54%R?KL
zQ5%LlP6HnG_(X-P7{R!VUo@W+d&)n?UefICo`38<lybmwPphzN(*h9gv24-EGje>V
zu8aIc=f(F{?|Lr$a{R4;))vEnM<w1&lTWR4W&RGo|FN2XV@o_~smnxOQBj+}qY^*i
ze`a5b*@*po!@Cwq`Nhv-hUBBP%U!|bg3Zz9-``N#TRvwZ@5sONkE_15R3j64NB*6C
zT&DN=h3N~8C2fS63;#0w-u(^#o<Y}ag2)Uhg!pXI45CHo?wB@a2qE;5{eJ;bpp#?%
zNbmJg$g8qMP7K>1zS;eTrC!LRSw(}F2F}wlE_*>I9mYjm!x&kJ{m9pBmoOc*T}S+X
zL{EwJMc3L<m0$hH9`D%W9e=#!bQ;~$=C00r)0fn|l8e+kR6N<B7V;n1gPnM=U|?M*
z9f%p`DkX~L2W+HB$82o;CP;3mxv2X&{)SJO+{8$|9;Q;;-MMPUX;l6h?n0_Fx#B>L
zD!*pX!dayqv;10x$c7(9)!-3&K`K^Va|@RFN3Y1wo%pF^x>XGt`_ExVlATTQ;XC|j
z>}biOf7a1fO+MMOV&@|*`PcC)9q^)VT#&!U_Rd}@WO|AEh<kb#i1te~2Aw01{`*f?
zwENHh_`l3s6)*quKmOm2z1Jp)s}&t5t;o3kdvApupje26);9&>vw@gD5PZyVC>k4k
zG5Ek#^Oz-=+ZrxO2Z`jb{dB2zo>&Q(YRCtcf^n5r2W#~8Wz@MmXXs|mzeK>zsu=bd
zjPZm={h7J6#FCb3j%#?wihT%Zir<lAjtePb-}#R{lToI53cL8kg{I5v)&-l&eZv0;
z6vJR>Itmg)GuF{a?s3;t{jnf-E?2axeyEzeb5T9CN=S{Y;F!GL4X>?StSOojGA|f&
zQVd}FL=Gbao@w%%WsD9A=ugterR6@~^I7VB-tEcC<(@}7Z)n$E^&Nh6srzZt2S7iJ
zS82I>yW_&Uydwi+JU-;YJ1&eL{rLX#KhKi+J03;u@ppIh5O;M$Rf)`rQhvdbKwj<0
zD{}UmzZwa%3bp&&6YWy>B|;eiI1+P(^HHbRMLG3F+U(iU6<s1cQC?&5-R+5X+B+gn
zl6EBK4wuI$zGLEJVufQ$y@BP>a7@oqP&lvT2^AbbO*rG=a*(bb_N>UslYM?8`>&8{
z9Y8`cee+ZGushL;qGbFfvy=Zkq0K6cfOEE&JjQ+1cP3ixH-t8;Fe`pJaY6A*1uY&s
z+VeP*94N@}bC#c+6i7!?#^<pr(Xez3rcd6IBK*()_&*)Z4qy4IMC<!65-t548`<)L
z#K30%a7Dzs#10cVxz_H}x<*)8nF^n(rom&v{Smk9_L`r`ZU_=FuFG*1ViEEg3nQgX
zcSKHBe;iy4$#xW+u0&Gs9Ruf$jDzovL~|JUPJj~;@)-p`jKxNka_H5yIh~2U6~Fni
z)&R$<Md6Q2ShCDp>BoqI%i>)1*qKamL0$HEG;&Sv$uxe2v(OM16ZaT5ErH|bZ?Rh8
zf}x_=I$_Zr9_SF!9T(kk(NWF%)}RmU<!5??OnmTgO*__n@IQ2zNvfEgCLpm$jK%Ku
z@xNgy)*SIa)SVy|&7??plUW0$p!petrUh}LpIGUf`Yhd(#h|_>y&pFD(uOx(+B{%r
zD)Tc=C6{L<{p$1UmpiiR)YqgR#a#!JpQpFo)6!LpEayp@NHW%T$Et=6C9nDAx-Up@
zjvcng&2sWbJmvOp5Z@2tzw$p`Eariv380VFDNl<XHCaEj9d3DKIQ_E3e-yj-wE6x6
zt(*V)f?xcn8%m5@JA~kTQ_!%eH9*-mfX(GbB>H@q9(^&V^iA3)c{M>!=gd?L1vowz
zJ5VGWA-`4+5*Gk8QxcgN5>5mqA7tq^EFmYd!ULhs#{#wP_ibM<*Vq)Gk|_2acl_@q
zWS=`jN}_m>p@qzcNh=HBC^*b>Ey1+CfC5k38<<dCYn}|ao$CIERg{E<N^OXv;5+e`
z5)3=I)F&Wo`D9%&a{*p{^iDk8{e$%Qzb+{G-kSX_E%E;v8`)g^J-;LJ9s9k2Cl{!*
zol)4rc31o-_G_@0Lk4RON@M%{6MMfS?|1C|4!_^=_lxB8^`)V+KXk^aFUKuPwv-sY
z{48EkLnXg@F{~e*^9GN`Xk<!XyE-g4k<OXo>?qC%PKPw~S-TpX4uTlQ3wgeJK3tb)
ztciWJ&spE9OtDw6e%gx{%vir;aZ1^)!4f?XZ|nEPu7EK|l89o1D}O_QCfljKk;JUe
z3+&mUV^!VF%6ND@-KhABcszdhXR&=UL678^@AOcGL5vlUc)SyjO~{A5ekzZD5*s@A
z6Glba)trj&*QDZ9^w=|7Zi=zeOCP=({&XPZKZpIU3g^$tt;YC)J>3mEY3U(ody39^
zQS5JdOLvlyyQICVLWS@KKilzVJNC>eE!cvcs{0hHKiTAzm)NvcP5QVeVdqXXE+kp#
zdqAk%pNpscU#QI^e|_zp=m4rW_~YI9F_{WtAMeJGcl_fW{g@hH`msYXkTAJo)9woW
zSbY97dqB;!qgGJEoo4qWd)a<+g`}@~7@_F8n0UFHzH~^Y_yCE@qYHhQ=7^1Wdcj>w
z&*Ec}qHk+7_gSooUqSGj_(bRg|73T3+6|wOLF}0<P>@g7Wb}pyZsb7vK0P&6+wRBM
zHvySbgDx^8#CA+<$D-)CXM!mwWqjN}?X#ycz2&EtHZb0Rp2+MgYj-O%<w^3}_?gHi
z{%j{cqwQCGM!g79dw-EKO}3gQZP)ZHVurc{oe7NxS9aq=Jti;4ulMv8Rz9&v@?$a*
zD#|}80lq%$KGGY;*_0&7@yuGN)AJ3RNgtn6NX@w4k}xyUH-~v)P51LzlRASBNt&VB
z-jaRH>50ry(tD$@%T|!vuM_17SfB62?7<*M^DpF-1>pdyP%B-@I3`h6u$f-1N!m9&
zX(m1QOBqG_MkH*e=FaK*mThHcqIQjUO`;m~+ry;dgX;RPd+};tqlT_r0i;`r`Pakl
zgV0au`I=PA<w6X@z!#<DH4K4`ukCL%j=ySE-RC=<CE#-7RtYcAAWNW3;#y0fNt&<(
zhHY7ElQ@MOY{O*J)|w=LJ5RC*sRWl|g81zs!6ewyiuC?fGcE3)$^i%4fKLJX_#HVt
z(&c4OeLA4TsUW{Q&?C+4aZnRl8WJuDmVO5YFY7>0izOtINWWOX&`S+7WH=k+Qv>WU
zz9h%gW^9VmQ1UImHYDJE)pnzqPgLO@timjp<fYUR3;`ePpL>{9QwrQ{sW}lYu9QO>
zjQ*231>dTU9yyY;FTXvo&5+KNe-$UQXiFj2_^e`Ifc%xXWbA|(X<y`-bDESJNZb>`
zc9O9FJ9E|(4c44FbWDQ|E3}o6g_U)>K|UnQE)tdKb*LpBKdegpK+GE?ictOQ5N6vT
z7e~#~Qp`<@pa?6IDn_%_LrEt4_AQ$bo_H{1Xj1JJF+l?I=Fa$<3+|ZU)QchqeJg(O
zE~o&ke!qVXWk-MM&#y`&+kt7c`fJiD!*WBg=nGb)m)4Z>bbUy2C#dvuapFEav0~|S
zG)82rYbjQS_Hz4%%Zw{4tT~{;8Im8X>VgyXy2<a6m)Tl=ZnIfWRp0DzV`dGPe!pYi
z@96hyk}MYA$WkVAUU?K4ehg$7)QCVFo>iyGr4J!pW@TP8KTJ**PvoJMi2V^z0Bqxl
z)`+t0aTogCJ{#&fyd3;_*<!lZit`=+FlsS2D7HPY!rmrUz;z}`TLfm3F`xq5e8^ur
z+bl!o*7pYbyaaRw+LYeNC6jy7{9~w~4*eay+L9U_Aax>N64NXngeT#9fHI<mC)+HD
z0hLeJlNZ2QQG`SR%?ra<WroF6S)UuREr<7`!+AFm&$lIO@_kLbNV;xpSl{%_=Njhs
z!#A{37=}B;H)Iol_-3&p8c;TqMDfG^I8~}mRS*SYrjtrxlJP-MY}8tx^`<Z7VPz!_
zgN!Ky%CrL!i5wygWh=3jY|MN>Mwv^Ts)DJl&eIf=`PnHCrxut}rN2uJIO#X33&A3X
zMd&R867VdZEwckk&ypxwgc9;${7hw)&MeV!O&O%K?G}Ij@Wad3?|x7ZKQzK@u<<|2
zYS-mmf&S-<*%n1Ln6wph+`H(1O8%?K8TrXE2~IgVBrzU{SaE#(E5k8y`Nv-D_0;VE
zHA0Vuif_c>##h?GYT4TQ{;9jN>GN@iYlG-DI*ZJXhf==sJ`*XOQ8yjO{z`Z9+WT1g
zSIa}`@|kk??}wF=7LglqrXR9r{Adh&|DFs9c_bIIyL#2$=a1rhUZbRRSxB%UMoeOl
z5i^S?GYc0pJ;iQ06#muXh`C|_GudU#Z=QDx6E}*xhJqWvaEbLcN;?H~8^s-hwvECr
zq1i@hhxls~VOx|nDm;XJG6tKFBHSt?M$x1tdb4LTPb`YZGGfNm2F`+8>@(D*dC~e?
zS8@1d==jLZNeiX+IMMZM&C9@In8jZM0RR1B8M9dKVl)wr_3?}i;mtL;$|&gO80e<2
z$M9J4w`+s38=*qd*dWqZ3m@^EKeF-bc=E$9omRLHLV)e;rULRwlNWj`udX*`U$8v7
zs8?k(tR_^)^Z=vs<cS)0el$gO#6>MV4SDE$mPYaEN@nlFH>z({$8ogeRgKsvMm3vZ
zAfA8sEBlQ(**t@m5W9yg0JoyIHt@|p`^bOtDMhvQS8ItO-j~jKU7A=m^uz#<=GT6t
V`YYuLYJ0rAuUNFa{}0(-?5sd>u7v;q

literal 20518
zcmV)9K*hf!AX9K?X>NERX>N99Zgg*Qc_4OWa&u{KZXhxWBOp+6Z)#;@bUGkYOhrLZ
z3L_v^WpZ<AZ*DpuF)lMMGYTUhRA^-&a%F8{X>Md?av*PJAarPHb0B7EY-J#6b0A}H
zZE$jBb8}^6Aa!$TZf78RY-wUH3V7Phean*MHkRc+zrx2#i=Kshm88;4Tie*Qr81RL
zb}6Lnj7;@RZ)kK9?oI`rWbqLZu2uiWn9clLf63rMkRXXCaly!il;PyXIUsOxaq+rn
zauipGtT>g2IM2?-uDboxwM@Gl?f$ZhtL)FsOyF>7ulb*t$o{L{{?$T&{y)l!rj3gP
zMOB#<Z9{NRU6Hg|S?s>nAX`Aff<<K!UCXA8@b4r3Y%lCHY6WD7PbY|16c<rkq&SOd
zbSi5&;PJboqD*mR#aZnSJ4Xb_X^XB^9=BYbqPVW(+gjKYircP6QPZ8zs3EBvo!(2t
zb*(@q{;rO%=*PP3s(aKDt|X6}W|?Otl`9AS`aH^TsnIX1>as$0dkacfRS2tSTSjN{
zh!!YnFJ&5%6g{_ulroN=8=jZ8b4<78x3Z5+<2ELIaog6}v1?Iq%FQ&0vEh>4^c$zu
zxCv#iD_hYuF}9s3sFJ?Ng&4@J5$O9iP98$zAC7g_JVeL%9F?c27Y$}lM%tvSYgDum
zSO0L$ZR8Bx!bYOrK$lMis+&tzg)hz%b`vmXcp}mV483f0s>{H=!IYVrEnbZq(nVSN
z+-$>H$4^mqjn7*&Le{zB*4J#(EH34FdqXq$N5Mbufs9x-_@Zn4fM3z>ZL7A2(YLXj
z6O$R7tx7g9Bw|ek(~b{ujjb}u-?4WevI^x{fdY9C3N~WDHvT}`7`EoDyJBm~w!=Eq
zka6aG`0kmBJU*gaI}yMnbYp*<hdN8NAianM${~sKq+_!laaKQN4T_SoY4y_{6MPkL
zd97kxB*P?Ko;9ZVBpX};<1j;G19^?x<N}^i87iHxB83bDkJ6DK&8(&_ZLmo$aRcaQ
z7b&toC_EM`;iT*eFvy}TKv0|RdX4Lwu95a4x<<-HaM#x2a?6r?RA;P$75&Tv)DCSj
zp&T_yN$l9bqq2BwiC8hy(hbgnb*;+d1YOYvWWfur&d59XPit%JHzoKt_62qND$41U
zb|QDm+K4_dSfwTS4E3qTLkk9#73b*koi3{jVO@Tg!lbxSiaYWs#ugh7*>F2A&W0~g
za9WWOd%GD~nIv6>+xm?z_w+p}HhS2Wp@6hZ3%2;kYn&BfkZW`;A5ny^ReR$fL?>m!
z79V?(G>_3KGp!!_ag4eF!o>O%WlLqRm2I<s{a%2P^?D+TCi&rr`<#^I^jJP?=09!W
z5^Q<vGD`uVa9QWkHESB2JEax31|xIjWCjf)V??)$?poadvW20Tu^GH}FTkPa6@~#`
zfXtz8{Bg3B<_n<wi|8?HGP2G+uEzU9lCA~{2QXV8N(Pplxng<3VIx^Ir?S2Z1?Kgk
zk6DV!^%?t_OfsO9E~~9Ht0Dv)>*we>Vz%ds>N8sViBepO7#&gjgfy!zF36{>NI`%u
z*rBOvlx9gQ48|Lnx}oT_hYf_?;h_UKFkMf9!YWeq3{p|}Db5QN6l{SFySd1W<E4B8
z!ez~dHp&L%!;65bUIbJ0B8-}Qby*XnDlc0wps|WzAcIx-WH2=W_s(<=-4(q`fq;fm
zPmG%tWLyFb+!a{wQ?3tu2?GKcs5yP$Zt@EMgQh*C?X5z(#Y-!ctSREvZUuwzG6odl
zHTN~w76t`i<1DWMkpM>5zEgCHyL`2Q`oIl{=Qw_WBYbKDmN^DGnYn<2^-~f#Ccuz-
ziOh#TWIH&S*7hC=Eh!i-0HT)>nSl6PlPTO-Ki)OtZy}aQ2kdNm>CLBFvt!aGt?!fz
zHBZCJRYGb_Gi9j&D-UsVE0SSEEt{znWmMCZ6PefK&)+ztq<of$W~f7Q>52zUn^^~%
zmyOcUtwN-FKb9vHY+lA`bi75h6d;lBQT@o4zJMx<0qQoc7*Lb4c;tgssXF!oYF*?4
z`n8Z5&;)^O#O!wp%k$6+sPn)DejYZxNXvJEqX9U?X=?159pT3PIzWIew}}8)h9+c>
zMK%zdy>%BvdJYF7YwSRe#V#ucP_m3%Lu0zkL^|QBJwgW<vR$V|z~7vrCc$nHL|a48
zRjHnQa|Vivry~d$Jvc!6D;PA~_#r-0H#zMH6P>cUfyz^a%{LepsOhDM)b@@`*f6kJ
zL-v9RjUNJ5J0t4=q}c)Nh7uK6dts}D9?}K1N5d|XadRFRBBa(3)UL?cjG0gdxD!sH
z%7OX@z~xI8gid@(5CE&+(ZKQw<9KM7gPd@5i$QKc{pb@&C5;o|E?Bq@y1dOMUbA-{
zaK#?2m}0ub6%$59=rED*Uh*tJI>Er2T7tdkBaCm)d`LX<YVqI<)e)JPX~1yloC;}n
zV7&~(9X5`5q7~>Ix7j0#utHwS8c2b(Q>zX*PRm_RxaA`%cpLQpUbeyo@WK6385tvv
z`*dM2*~%W68oI3)x&uRXeIguS_5f!-dc@hnAPgv|57}cvO{&b7g+pjjzvv3xmX=Gd
zFt|{2Oic`8jR*`>=gqL_jx>6x{F0u&3jmv4&rMs_AUOsfGBpomVu-ax?>>x!0d?6;
zm1hth^Q^@DGuD>WB?HJBM*6)0v4j|@5>6TWqR}yLBb@t!1AU4sC@`S<l#Rt<ygJ*B
zFs_i3A=WT$5C9tn*RUJe6$z-?pR)E6jh{?{@Z|tQgtz$Dn>e{3?***^33z2|iVNgC
zji78=nH>`1XH6P-0wDK!geB8v*tKX;J+>_1oKp&l3^W}yg$AFhoWiq-j({Yv1FZ#a
zh!zCvCXcv~o#lwyRrn2pM`Y^U(72%9wMUzG6;tC(=Bdb^jc$*E8FDh`=MF<o=EN<w
zmT`qWdG(HmshoP_!%Xh=h%4<cWUv-^bX^}kx7t2@Ep`$YQH_i<sXp-R{3yXMv>LUR
zqu;kmiYP=0r6p?m2oya7aP|p4se(YzwD{q%*<&(+N>&uh%IIvbCA!2#nxi1lTpyuv
z!_+2lMc$L=zaD=_N&Bha4(n&N^6C<^2{#MVDhQkHt0?@M{P@Iv*c#jg*%}rnGYrG%
zXq@3lT}?I&v!fvgB6T$HfUI_A0+7(sg1|7Y?jT_3SwvO*ygzPO^SNB@j9=(H2tmAF
zG8;U$&~|xn=*3|Og(ZD5nxSWvX>7KwzS!2efMMIxmOz@3q)bYqw?Kr9UIG<j!`e6a
zPfOomgto_Y@dQ@kyY>64DI9zGRXlEe`Z@~hzH~jjoyxDmx3`R|;4RJ2Dr|^CZywFf
zzj-vbek}9m{x{oJ&k@Z~PL16G*0GH{BD*$i2YAO8?SgFIkX=zN+phzxL$h^2wQa2q
zkN}^aB2?w^G3i=kTOCD>O;!-?XQS~0{MVbVebYQzw&XeAY9g0*4oB){1z>p{Z3G;-
zn*{;nwYD$7oOVqRj?~^Z!Evr1Z^7lk(t5%XZlO&7JHF=I7(;<+Lu)9qZfXwa#M|0K
zk$qExI3eE3A}&ZbF^MAUW;RhI-Nqn^jG&cSl?{5x?3}({*2?SSvg3z*z)${`wGMV}
z!nc;l)y%<=x?4Fo-W5g;hTPr80r5JSI6zK&3kO5$X5e5r9qbznAz(#jO+{Ay%*L%3
zG+H*&s8}T!F-K7c)WbXc=T8HH&Opxbt$|HZg1qs$$unq-2LXK<>Eg8muw>0}%p@cS
zKU5HmBRBzuEdr`jYuBCF5S&qg>MR+~ap>BwAZu*EqU?|C?|;F6Te><6CU*)KCkKMg
z&AowQbaZSWXx&^I5Jqcf28_+lje(%EcVHmcEL|4}s?914u&#zflsYp=gpBr&3NeT7
zx34X&-Uw`q=wjqnpu5?$l?09^Yz2#(#acsPZJ^eI+v)fUbbIr&0^HKttUwM`*bwF1
zpU1OusRUW0NfBjPz8yTO1(7p@i>(1*bF)1_gpRfdfYr@5K@eKoDkw6$HU+>&BlD7X
zG<3lhjP}+OfXvdG0x)f*9*~-Wf07ODSlDcIM_IGc6=c~5{EVO6(FeAscTTsG$H@Rg
zaXOk|NLp8848`qelEG;04KpaEWg~>*bTFAvl(vQwiVv<5d3<xNIU_<wJ4S_g##CpB
z7l2BnL2108iu5_E_{_WR6S<)AGY7$6-u^GLh+=7?r}z%QE;e~RiJKAj<8idQ>&e{A
ztOt*^UG?O!Go<TD>@DYd4og$Ho*-~VfYm|vY&J^Sv)LYDnP0MJTQkHF%}`E;#{t&S
z)Hou$8W#t6M|0wWY;QnZQ7zTg0oK7tIH1~^1qVp*0tF~b`a6n?IKMSnV>AF$aicAm
ziPJ~~ulrw?213TTlFG$^K+?Jy7-(ik0|ZI$W}rZstqmABt6ignq_sDIkgS#l5|VQB
z3WcppzK2C?VTq8@u2P{-_<xp0bOg0QbFq_-$Zn=^1-PTNb7XKcXs+PaHfs&Aohmyb
z+glz-U`ylTh#I1By~;I5=eUo^D76Xfz1ke~?`gC=Va*^abM2&;2&=>!FHV3!CNTlK
zHp?|G<{+Z}-cf|N@P_2`ls!iY*2S}Ser|@Z1TNAZjD#0Hk~jXS<a`{pO8gF;wUCBe
zO$`^Cve$3B>-DC;%T#rbC42tU+jrlAq(6{?E({mN*kd7GTwH4y{l;ipnBnSS-_JtV
zBT7^2*wNOoMSR5S*pywJ2;oqDVa-9PZIOo1yZ&`S?yee&kseqD$>9ZM=SvHLM6Tf*
zBN#&<8ZKhC2dYt{yOYiq23pTVyeNK*vz(|i1oBgsGK;?s=>7G4Pi~L!v5ylB02?{O
z^D%H!C{H6d1@ko>Q#5b$F@^JO>r*uGiV+ArDEl!*^|T>VF#Y)u7bGu>G==hP>vz~g
zj(r;X2msK^1u)NcDIlFHm>l^wZUM@>WeZro&02u+Y%Kzkmx2p8p4$Bjxq;6|yGH;3
z?H0g1wOc^))k*=%TPp=DU#%3NvPK+ndc%1txq#%Q<N}VTlHcQ!$<cc<Ch$@9IShcp
zFCabDeu3w!{0l(uCa^&CZ3qiMPyJosd1>(i%~O|s6u8$~a!J$L&*`-)d+w@;&zgDQ
zke%)3d3<Xq{bhmHFf82FF#8CAcqJV+0FL3pTpJwk20Vjlpm3rlKB3(6>-eB6ny8j%
z9wGWy*fAVK!`cw}8u)&u<r4dJI}D^P0U96%G2hK;31q%|Q4cgsIrURee7RILr>{q&
z3ogUl*eVRV4pkdg2AGgxeV`YjMfGn3s=7}RL2b|uq%lXwF3VG7>TF!Gf2~mQ<~P{<
z-akB=JJyw;honAfFGutO7rW0u3{?{jeDwap;8%n976yNC=T^fvU94L%?FDr0nF-L$
z-Np3-sy|)$D!4+K1xP!rL)~d-_Dq1-rbpN!Cy#pDXM!PFVfMraqsjh2%*pdZqL)hZ
zn^`R^y9sYJ?~WMQ18DlMktk9BR-r-s?e}qm?%3ZS;=0X}tcuxXm`fd+ATFUkq6B4+
zsxHMcF<a`5#7|r5-YTrd6Rhf*OLUBVGk1@QbY&T_q=yFo6UB8M-+XewU^LgqC{T*x
z5ZB2Ct8CJCxKSOF@_JRq22e;Vlu{-AzQ7HH4yIc44#ni1ya2Uyt9TAg5^EM8LlEK)
zCDrun0VrEkG;F6KEF~(E602I2BKohbj?9wTD$6|CTlVkwdCB&USg4DXc_k%_rhVk>
zk`@-0zN?z}TIJ-{R%ey+&BzbM=0z`3zCh7W=&Q2KU#PSx_e<9K+p(>0SL;3Lp?@5m
zm|<&#Qm`}Mf%4O|{k`rlk_-O&ODPrp`ejRn*$AA2W2u}g+9VHR@AyJl=w)6q3r&lz
zFUJ3zUAyZ`H;b}stFC?NT+l|@yjZ_+AbRk{SSp&Cjm^}3hP^iQrYs(}bf0Fdhd61o
z$Cyq=Y^MtIOdRQzzC~GIzlcmkM|2{KOkYOXkpK55BwjEZr!2?g$>s%f@PO(9<q?_P
zd%;|6KjN6@fGi6p_}5F;#l{0489i^LUqVs5$T)1Ck*>jtnZ?_e#phuHV*8G4#(l`h
ztXCEF?ec|7;>C=`JRhX<C|;_m=rZ$?U92`=FX)k2=IaY^QnWdJDTl$3shS<){PME)
zlJ1Q*V_!<Ya%dj1DvHze%g6{WoB}0ZM7wM2xM)aZu#{HWmzJL|$wbYw_66B_b=JN>
zQ$Nq_mts=q`TCNq>O5~>jA7k!1@TgL^_h+iEEkD~1h&h=V}iHkQ?X`fGNW9_wPMBu
zhSz=2v4;02u^t1J>A<meea({;7cDEu(3LoHu3R$nBW0QzWkph>xIvqJ>^3b#;3J0y
zu<1Qz=S`Z%>e&cCR_B!+UpjP+wyw%G3pOB(j@_dlXIE`wf9~<Oh3+5+j*XhV$VUeK
z)aBZw{zm3)y5f<qUs>L|fE`x@{K-~%s%vQ1O>F}(ZPZAhrq3a`w`nYVf<(Fn--iK@
zGh6F14H32;&rso?vhxMiKYjcKwYFL~lDMEB-ni8JN{(Yj4(`QuG!8wtU@hA=M@1iE
zE-)DxvaFgISxq<{u2}j+E6K7VYqL0yQepT>wsuupd}$H&h07Yx7L=@KwKhy%ERo|;
z1Q)hA{k_A^(72t%(hfRMjIC_6=I|^HoyiKrK%Ux~a{AQLlGCT;;-xpg{9<XuM|exz
zwNoW4h*nLHtmI)!QTRK?8V^Reat_9w$h~h^%@Ngci^3AQPp-I=tQiomV%bvPFdg%|
z%RE3OVN>7MG4{4nzWdvH>Rpo5*3`+8e#W-<!|?ov!sYykGAj^y7>|=UH^{*GTI14%
z&603};EO1t6=o-FhE^Rn37w^N=N`i5YUMGAu$dYqzZeal^m*aR2><6WgTbV2dxKv3
zBWmy{=j*?GZE5sH5(o6daH?dXVwM%Unz%0iK*a_`tJHKY*=)O0vjOE~M?&o6f32eY
z?wgP*`Q0}!q>_Yq6P3h7+o>cb+gc?t<;E)c-CqJ~<ad908HFUgTj(R^*-RZV*T&k2
z88=qO$B;et<JarPIV-tdMDZWLUNcZx&lJ9iVV+eZoz-0HhA?aCu+N-F`nY>gT<a$N
zJ;}By#?6fI<K|u}VI-EfZ_*pkZD#Y?iDPE-?Uq&C@M&m1{0CWu!s?d<q?(wh=u`yd
zVSh)|c7!5gIv1|MT5`Xg82BEPq;r^I=p5!3sO$%}G00`3PG_6x3HXYgDl3T`H&%JD
ztlCj!&-dAW+lm`#Z>+Ep#vP68X)H&o^=m?oL@SqYU0GJG-#U`4TuXIiabLGtN8q}O
z>dN9iL~vwScO1%*<eH_axHTgT3*C<$T74<WCWW?UL%Sozn#FT>igiobc07~Kf>ut%
zkiy!P7*mWJRPTAc*z{jB%2iyeOaTZni_y0FfHPz!c%g6i@oT4qVAxki>3}C^M~@v3
zMF+^nif%nvCY-x|kesXtc>qJaKBPbp{uk6p|C8=P^YJoz33JRICNDmx*R?rF;ygd{
zqm*)QV-Qih7o*5s+*Wzn=GpN+e$?!In}OP)9m*m?Rhe8ORHW)&a~(GM(a)YK3w9;4
zIj-S4V0wASsmx7#)hyA;oZcI9%S2UT2lv<FheuSjzBMQXJs=JpB+v9hA%3-A7zR*H
zR8r7|h3JURSrM89;5$Cl<ry2C<|8hd8r$uEzx!8D?U5jo9J8n~AjP6(2va7b?u7gT
zFzr;}1FEK$0pJTh(HlX&FXK8LEvW9sI0)oJUY^qo!Y=DbAS+tO1J^m~Slaju!9er|
zdpmE)e0iDY5Qa#K32G?h3=L|dyk3b-XI2hr+{PN`ZWPkEcY@Gc{J%HDoQ=Fl=a@>c
z_eImjMS|YcWmU0{4+(ro5Sh4(aJ^M=ZSX(5x$sZ(U;&OMN>R(Evhc^Hsp+(_J@L=v
zcDM$tYD74!-(;*puVIcZ*ez)S<oVD%pd7WP?GkIj@yM!;j<+Z)(jWvaYTB^i^!mRr
z@V!wBmw!$W?i(KCJWFXnLUK{Dd9K)>CRAYw6rjcRfo*X?^m=J9J0U?mVBevuQFN~3
z+pv`^%m;9)`g{yXGVBr}`)NY*6qDU|fmm?2LoP_Qn4t3G>2E)ci!|yf+qgcXHXtP<
zh)g&ALh%@G<h7+4c4=vSKPqK({+QJF{WdeZRxTjhLQZ36nm&IG0S|l-0((4P76QG4
z#C}6vVU!MTC4PC0tL7&Zr+~F<ra9opm$J6@$f0QSvT33x>jn))Tb@p2P{)?vuEtqV
zU~=oVUtnTVxpb>}vn6Yjov$${VO?fvh$b?_4b~>+JfRxFoF}Z-FW7>q`J4$@!-y)~
z1u()(+ye!LY>pC+AlfD22~^(_rU3Q*8%o-;4pQX{=rFauqzP8>ODb|x*eSsS+M8*^
z0vuUaWNmP+FiK4b8bCc0@cP+Rltx+KC5Om0Wg%ub2>7+j+bp7`5|J2h<+ZDV6SZYU
z_NfQt30fg!_?n{t9DF1QsjgE@t4;L)_;z9EeTc5{Gq|akQ3NtbGZu@$?tbug5zbvW
ze*WS02S6Pi;lR~%32ok5w-P^qLJH_%bWtDxRzm^V;rM_H1KZyNr-cIEK|KEsommXt
zkkgoaV^R=~3+b-lkV&6Py|(YEi}1OEEO!MsbKN5{!x8*Ls8k@PPh~}}^9fR76YRdg
z9+_hu`bz>8nK-!xSzJ@xO)`I|;!{9^0CU#{jv%946KA-IG$V76F4x2wVB*+B2|LI0
z0e`^j(}~2AzrD*)4nBRXBRHV+3|Mknst~O~OBhBWwUofBF%N?YsTFeU;n=<1StCeB
z=&_6?AM5}}MQRf<7=kB|&MVj~LmRnJu}D9FiY@|%7<5kX!xaC7!t%11)%cINO;~PW
za|dVUAFhqx)2YS4Zk~xGNY7>@4d>{Yse@>_kv4#<XXJ$K&_BbPwh<0|B5(-}+fRND
zYLi8D+c+#M@x!zQfdF`-ae@p#)%{uQsDjL6jwhcXTha`)kEjJ*ICB<<Q~h2%;MPGw
z0Bj@XEEYF1df{8g9`RyviqYxCv^GG29cfv$*){tk<_8|Mdx%A0<5uE{@ayhOyc_z%
zuEj68R*)QyHSy;PWVg{6nlmE3;$xh@40l#+%grJeHtqI|w4BfJ)y~O-=Emi`G50CQ
z#^+Mj*&jF)TUPNo?wTfxi>N3wyKN33c%v(dix(qr+B8Lv+tFk5pOOe_by>X#sl#y?
zcy8+>GN9C?eQiiuww4YTARjp%NUUZ|hR!d+jjiU!+|~!<0O?2>Z%&zQXdJ0`ns2JV
z1Ld}bE6{FPz+I7@@UU|09(O{!y&pTH59-SP5ZC$bzIBOB0mE{rh&sg`{XRTOYvb*{
zaNw@62(sMJ3q;-EpV4rppG2}?c?pz-pW-_G8@kbbXLmgMN|;JkQ5`P$1v$cu;^s!?
zSF&>Bk7dgnA-g~&Gcg|Y`ycwZ&E`44?dxT8P!>6vl)BPqfyjSGaZTVq;>L8JTo9R=
z38v|4GFsYs|Jr&k5jG3lxFuzo+a9_IlY`Kv-69KtExaZNHU5#H@D|>b1s2vNCR;=U
z<=|K0N3b@xBss)_>m&qO-{cDqEGnAqL^L2LVt1F&n0VV!BjU<LP4CNW$V)!6*^Y10
zQ-<xbB<y5MUKZz@^dJ)Y2`%=Qm@(OopY#_msN?-e>m{?$yGqTaJDu`PRj+@Kw(BHl
z&}h_dS-hArFXJ@g0VB6zWwnW-BmRfnp|okA5n7??=8?v{ddlR?3#Yx@7{=Lt$SP7e
zVjqqqO5hMo%2+l&ROk`aH-~!g;W}jnuK#Vmi+5hzuo0g+NR|D{{sq>#3F;vcKPX)8
z4=ITXilWEtkBG+MYXWlR<(W+KB+X+$lDaF**ZIOwL^gLCz%VtqV+hW~FAGrJN#@Pa
zSl!d$#X@g`5W1X@G>=)+#d+TWhGc8wb5M91jr>TWkq2b?Bda#3m_M$62SpbmwFBGx
z1@h#^8WP^Mx`h1{HtBt;(ce2%B)7n1*s72lNjCYgfGu-L_C(_LN4Rr?CHnWfe+AaK
z0g>&_qVe^~ony7_{y#6G3jaf94)B-(VZlIzRd>v@<`Vd`I7?Th#K{cA(sair!JvqG
zO+!h7unp^rzn4ej&`cCIj?X5-{6dF+DUWu_yrgpHIM@!}VaO1vq{=oVItV8*q_$2j
z3?aFt^!KLJU2_wHE=secpF0RnO6PsH`$p?=DG1x9+A`lYzYVFqZwbOjhXC8D3l)zX
z4z7Z*;*KZOC(`CNuF0Lcf?X@&L(sKZZk}nRQpSFaNtFdEPZs$TFehL%HodqEsm|R(
zRGr18==<v#|L(VnX|D8Akhr1qmj+)1hRf=LY#1s!14h>NHux-d8wKWSbW<)+T<q*`
zv(dNp=S8>qIJ)ta?uI0?4Z9rxQcFJ4UGMb9De#6h!9LiYCjjFzpC5pr+Ng%!<yo*A
z>am$I;#X+}{fvX<0VW74+(3`<G0U^|7M!cOU2&P!Sz^B734$)T?$N_$qk(1j6snvB
z5Q;7Zp`cxAN_M^t@f0<F$xPbdn#`pqLGbAFySVxh6{wE;O=s#BD`objJj8iE7O7ba
z!(`%=V^&l|EX>J^Wy8T036p{T`=j<!qvlfP=}VP_i{lbB2NppY0%pP-5GEkGtHlnJ
zi>Iu;1aFrtjOunaQNuUfDDWa*yO|}_<srNo)Uj`81GTvbZvZXm16w~`4#Ju~9rH$Z
zPnU&FjNURF*+$mShIlhWWR;ol7Rf3b+nFZ2?1Z;bF8K&-tgNyU)@oU0VI%Wpm6uKI
z*foBjVYG&r@!Ao&4@rMb()Lu*ydWWXd-aG)<D7#aX_#*iCBLUz*$rLD0@@1G_DGix
zMM~ynYjlmXBDGmxSxMv2wX9@tG1Dt4Ty4%uny#qhGk8t>9AA%V{Eh|{1XWfMxDQ$w
zor<9O52#KPnt(br;R&o*ANpfWz!g8j9&8U1dD4Wdc7Y^dr3)-pN6kDr!mAb8Q~|G5
z4E2DXN$5?R>GIVSK^3a!2&+#$N#Lsfh$<9AcN`ql=18>{ez;m5C<0dUz){mb!WC0s
z9lB?SPGNP3g9hz^29j`!G*ARoq=AD>tFQyWD8}Q9tVOkX@1z%F&lip$u<m{?jsb>w
z18P9U8aV=L)<_bd9>{fCHd_V;Im6(6lv`dqWjP}}qgu%^O#UX9xF}F=(E|bzJE$8F
zql@yvQrga!0<l_7jY6^YTf??f9oGi#K$~W4mX${4!%vMffDYJaY-8xJG2k2&G@Zzt
z7%5O>Hbu(s@dSm20X<XQ$PL!HmonI9F7-Ebdy5{k1l^U7rcQT)YzygP0@o3^SiE%%
zu1(RN#b%b!6)jAOub8NF1`k)xAL*$&u!8R4!jv=z4^z|}Jy2P*+19(b0x9YN?LaK2
zx$AX?9V*b`YBo;>uu2@UM-2&<{Qychf<x<IU3NS~p{|4M#Y{aU7o&ef;vA{A8=t~k
z<I>3iZ*a0q+xENtHATS13`H=A(Hr6I_wh{OU2t7?EBdgIC&brs+4|za=U@apSX|nw
zKaK5h&x6Twi08rJYJuFSenTYmZjeSZgAS)*bE0rcHfIV}H*A9sHvwjgz6SV_ehdJ1
zaY8kZVy(02T+U?D2e}{f@)+kIQ2g+2*pZ^q&MCeQ_;0xWnmZB%;X{o`n1>P$`{yzV
zlIOcJc?eb@eKDZ)4L#V(_qhFnRhz^z+LFo1PC)WHnGHx{XDebR?<dbDu8Ov4&J!5h
z=pw<eFS37Pr=fCQ55fa}bSdk|@-pjS*+^(@lK}fdhF&Hs_E8K%usb#95cE#*2SQ}Q
z&9K1pJ1h)@pm#TK9(g!n&VF7$Iv&gC=$fVUr`gT%p(RTS!(-Oqn;?w!c*TL0J@4hM
z%{iAJ>e43kgo17b&p5btZLa>zE&(v0c`bus0Mf4M6ue0tg4{r|uIYgQSXXvD-pmgO
zRne&<Q3gnPZ_0heqm;p(MdPuIO*65s!}X?J-_dzfcWl#5&{C#`iu9?-rE7Dhp*{HI
z@vxCE6eXJ!p$<DrH~7SYkNq~_>WXk|9N8wg+(g#kj!b23Cm|#Z_Rt4ZG$i!(DKBld
z)q-%H>-c8A_7RS+J|9Cf*?Sd!JPwWo08xS1RUW~~>>U+9B9{)*kIA)9T)~Fb?Gfjm
zs5jXyc08NcC|9t|?L#2^iA@N^VKW+Dk529>W6cv+{X;;a-kd?Bvzyw7FF2!cBv@Mq
z`;x?`6B?{cc)s?uC0SV~7eGFCEIg{wnFQK5J0A;qg9EaZr!J!HX><jh9Z&NB&r@X;
zB3JjB)&T}NL-~wy-<NG$UO!hXv<vagKJ@SO&f1J=9mTjwqKPgjw%8y6M9X6oY)MKa
zZSIi;q@+rJ#JzS@36-kpy#Z;5pxfJ1lMBf<W#vI`wp1rEPJT$W3T6&AY>k_>!A)D<
z=Iw3ER<>Q+da)LDAV)Ufzl*Qlz3H13cmq}u946DoWar~%-xjQ4D+xn3a6M_jhOQ?L
z)8Lhy!5Y4rF|YvuZ;3<B<ElbM3Rfv(yF%<1u>jC?s_~eY$wLzgeL8G{C7?~hZcc0s
zyvq#V8vF{OhXC2e`FzY$R5w8y4!H+5K&FQ|&S3d|jPs1#_7}i@E`u5d>Gbt9!TMhV
zZOC5OhB{zhY(pQWM>a4A>z7TufqQ2IeM4Uh3)`W5+)riowtOlAPRJ=44wHig8-7~A
zIRG_hD4V5b3~8<mtN|^RfjNwkQj-R=Pa4X=rU^5rm^oPIuD}o3LZp0~7*{EC;FhW8
z4BIZ%yun(flrm(SloN(&k>cn^w+zSSOv`YC+|Th5<;_QAj=(v?wgu$@ILXavnboyb
zYl-N{TfAgir&>N7U6GS#D2<x#d`1na>n12+gX-zUUqPAp*PUv93Mv3y)<l$%MJ6H}
zH6tQcTFg4EWTc-JQ2$&eW>!-?_a_aaxJbv%|3S*jX(nhT2kwRXX}BQp(}ok3MScrc
z|AEz)UTZ+h3;sKPBUuOT1O&{WeSpX_bOBY&8U{=o6e+^A@+m{uMlNy%YU3gWZ5F|5
zoT3xG3d;acIyJ(bocRPE7c+QpgV+apWbq@Sd*=+TcZzKd>ax9fErp9BuH|suQf$rh
zk^BKx_&u7#mUwsQ;f$BvR(NNEpx6Plkz+fbfOBRDqUapFic_+E);4^RJ3c!f9j0C_
zPqeJV4V33#CgmQ(-`PY9UxUbFOjI_LBC7}j*sckDDElRn1ji3x$3*;awoGIp-YP38
z+avIvG?pjcSFr>!&LTk&11b`PFm@t-FoVb9Z^UUz=A<!!(1~e}GFTHV`bHR7`Ue`D
zm=`LalsAkKxR_y#Fo^vP-R@#@G>$2(4W5e)TY*l3zw#({dR<-t?$Xv*z`HgSYl^1I
zvo;zhGqno03N@;SnV<mR^l9|SV!eCecWsuNRz#?bp`9=pzT+@zQPJ2$;Sr!!nEipn
zJZoAIxaW$@2w^K`bwbht@f>$elf^|;r6&kHIy-j87S=jQqy;Fw%8~~Yh67~vyc!Q-
z71n4UeXUkJ_BtuA*ao49Tv`!f<4|Xbw!{Qb0~!EKACVgOZn!qf1TKGLyen*I{Nc(6
zfIL;n1KbBbW=(d?u+6*CG0=KI^8YSwkv{@GLz}P1q3#)75X4SeR(li@xognBsIT-=
zV~`9nsLAr`I-nX=d7O~^umOrI3fZKfu_*Ml+M=-6Z~9;~^X-H`^&23h-k*T%2zjuL
zQRzV;UTmfE(tO~}n>0MottLV|@bj{1q9^mn+k?VBCb|cSWz<!HAxr<_$YBn-ZvFuA
z{ElavnnMA$r3Exh8(P9bw4DViP@7rc!nKtJG-MlDfP3q0e>H%Y!ipG&824+BL-4MZ
z3N=9lP|ztJ`Q>6vzBVFy>-q!^Ldhq9K-xXQ1XJq?Du_O_8x_1&T7ZGmY7rDju_utO
zy8Vp}D2RHCh)^0XVp1g^{A#!&1#CPbES(PFdyhd1py;mADT(s<W;!#t4!A3_78DoC
zgm^S`u*MD4pdmFGM$PEa62Kv&5)YvCG}yi-ObrcS%hcd7#z&10)dbZHYesM1x)Auf
z48PKIARo~wE6^4D*dDqqg91=%IjpZnOHl!oS&F+tkF)f@S%O0;u@rob{{Gbxo2stm
zU3?RQ@0JA1JGDI&Sq=uE$|_)AWme#lyQLlK?Y~<C1GEzbGK|GnfJ3zu1-!2*mLGTb
z&|@_qfG+F6ef3#`4WN1&bQty15J0J*hGLD${zoT<^@?u3sn&c+?I1y0M0UDG6^bAD
zp*h3Vj+&bnwYV0G_BPP$_Fop$fugSkG+z*DaW~{B!UN&G=TCkn!)T@1(Os8k0zD*g
zo~xU#AM`b{Z(x2*JUWKyfQ=Um+xo}#BtU_k%K9m;$@an~6b0J|XWqi@Nz$To9p_|)
zIk^nCn}i{IAXlA^+Xljr@bmdW^{a|cTQwL$)g)!jtV%GB&;^FuqUDIii>Dq&N+CJg
zt}b|Bk0v+~(E!{=tIHNlK|be>H+jfTk%AtM^fC4ga^CR<$vkN1JBHl};x?$@H~9uJ
z4nC|!v0)5llhxeU7vK!zLMi#+ohda{0Jlm>7RJHS(1mohl(b=-E+t(U_e)6^%&$-~
z2JtX7Y|PgfOx%{F@D7fI64EzN!H4eo7|9?C?u!U<`f8MBWJU$N#0KQUF|2(Wi7ci&
z{YmI76g~HYQf3AN^eiLYs086muE@4cpCVxlYzGC#W_lByYS!JKX^R@$pu-Z#gs6xC
zX~vLSKqSysF0gGA;Q442+_XtJpw)GxpohmQRK$6Evsu+y%M;KxZ()(Qvcp?h-_SN!
zu(a`WL>Aej92MvG64hiIU63!KQ69vDSnv#45fLNHW@*E4kSRdQ!J<8c;Rw|;1?`V!
zy%31oJsg?9TJPBdiBKS;kSQSSOwbftyzL$m7<Sra1|7l-PbgrgQzp>RrfLEXJ1;YZ
z2A%+&fZN9(TLTVl@+Rn@ee?_&dV)p@kN<|o{;@8*BK=f%>^w7&Wff4tjqC(FfK3FQ
z2$O??mdPA?xG1sjEIo_m9X3#W@6-pr=HO8x3p|j*kf6eL#Y+O%Zc~OIqVHVtz;>WZ
zeAtdof)Cu=Eg8Ujr6mPu*Cd07?w=MMa2<&b8pK$DF7gLh(w{;05vERg8AE7?9DOpo
z0LuVtD1)x1F1BI8`nq%Y8dY_P#}*Bk#z6QiGlv2_S7qIfMGTiuQUGMGAPj0$77?9Y
z_JzMIAJH$U1+^MdOc3KCM+UYbS`tW8A}4~iC~^*1!y@OP=n=omC<q0#37fM(7#t~P
zkj4;dUqH_TwM;gLgIy(Cpa)p>S#ZEF`Yb6x*GLxdAsTDI$^In+++07Do{yP>bTVd&
z&FYN=>}%kVVRcPh9ec33Cejra$k7@BvFA$a3x9J~rv3%+F>cYDGViX1r$&uHIEzIe
z;n?Ssr-uhYeLx*FNbjd7fSn)Ha>4p*2C5LVW?D*EPf<$+JC~-X0-jdWl0kc^dNSy)
zR?7xDvu9v}p5#-|LHPVS=#a*b5j6DW7^NC&955w?a+4IOu%iMM3BN8IYy+K%D2yQu
z<7%on0*t6sRPX~UHB<PmHT2FpB^#(euVM+@^8^{|eJZQB<x>G_{H2&6R$qY(ye2MZ
z8G4i9XdVpBxWk*Qj6tlytAQaaK?+aLP4zKIfZSLe9qb@S!vs6T(GrEQ!Wu%@;f;m}
zc5tI3s@m*|OOVE}>967n$Ob%+(oli-FB&e;ZduDC9JmyM7lmhp8%Hh3LiDCV&srKe
z!t}68K1e6t#1+0@hZ-oVfnmpr8lF)RpvH?fWWjo;fa72bt_?mbXhRA*Al9(KkBPNp
zn>b{lN5(or(7~~W2z2#J!xP?53ON*4p7;fs!ZeWpc7VX$Uf9lVgDx(7#}>Xj3*LzZ
z?Yg$;ur}zXHtL)<>XHI?L}9z1a0j0k1fhDhkbPN*9_+WQ>Eax!bD5#RD#;uiXrIm;
zwuw(DzxB8Mw%)n!4R4ZX+Mv%aSjMYs?a*jsJbp`#AJiF<i{po>PvecR_NWYRF)io^
z%&y#pDa9-9^&n6l`i>3%{}HQRyB3qgvPYU*bJ=@!bnyQDuRs0m?&q>mFP?}vIsn6r
znl+5o4FJ|?7ZRtjEh7{+H;eck*0}U00rSF!H6G~)X<i@D%_4e+J#JP&;*zT|ytIV_
z@Ea<O>l)U8#y!Z^NZ41v<Gl6CHvn~2N^T&>-i@w%F~JSjy<Z-#QQjB6%3>(N#sPuU
zTF}>OW5HhQ=j8<mRa!g-O&*S=!B?^Yva%93K*QbRUfHo-4z$4<y9mE8!zLxWc2}V@
zRgDZ(Xe^}nxEv@~9Z^cHK@+JT(lfboq(H+ZfKrWGBHKb9QW_czyZ$Eo;=w^9DFp+s
zlQ19f=~f}I5+`7^tP89{L9$8<0v!Fr>v#9VZ3W&XF`zkzq+7>AW0K+GXhLDPxZp#9
zhrqS!4c0-kD}5$Y6p-*&Is=u&MVb*`AppK>2)mU$*w1uW5?)VY8^dlLLlK2i(SjPW
zFvu2;{W@rR<BG)|owevzH<_&jH%%I?gQc1m@j!>*&9y1(7Nt=D1CN{J`?oN?9fdl;
z_sI+w-02~=y`AaKrQWw{$b`VeqN_$zE;M|YMqGg23Q{)!3ZG1pjo&cqGN1@1iVXrf
z?LNoY6CuEx_*&%%3Z0I_Sa;JM8&K5VIN_EKu7-1m0=C2#=&(gDv91ANVT+_KtCE=Q
z(Q{afN(>*FtJjaV1u4yFTa=>l0FTG=*^s>jvKo>uDV+22NET5|7hQS*M|>4E90RVB
z^CcYoQ-=NSej7I^OI{{HieJDDQ~DwSeBq1e!OEU*%V%V1R*zFz5FEnRfioWIG=R{w
z13JeY28|1{m{%aM+HJ07i3>cto_$KYxiu}is!bq!uQq=D`y*;vTtDZG))0EkuusMp
z2~;@~p=FrZxr>iE?ZD%l)V-<FO@P=(1jQpx97c-HRo8gqJjPj0vj>5ebajo2mSFip
zVTkC8u4R4mK_d6NXK|kWfqL!m`b4q|2WM;4T@Ro!XQ>ZT(_OEzUm0}`iW2-;Ws!U1
zoU*#{u6_ap0D;HA>$B2t%Aq5^vOxXRi`-T%Wz}{|BKxk!2J4Qvt4O`(s^V|~$Io$z
zVE4Yxp0y)y5HdDj%SRD5v#TJqm$;#t+^b@8DM_Dj9gA1BKnB7t@PsYhu3%KD<<7db
zq<4H61*(<9E?+*8qHCk8N!Ub6)Q7yF649|XxzelYI@~3nNFCV?_`P_7GAvt8KU#$*
z!fr20Z$T+8U2%?U;su9f8CFHIvI~meWLNC2)tP?EaV<@8ZK4zIHNmhS;`|yH=AM(u
znS0G%j`)YHIS&OR7+zD(MKT==n+Kx-by+w)`g=eJWU3L_Cm1ydo2)b<W11%FmpdmN
zwjZB%%85%D>VQamA2;aJxMx}2F~~=L!>f0B`SgbS2X`?%+~I^DMa?BUwbt&3I|)%n
z$}#VY!?Q3YL2;dDsE*p~iokjCFyCSe<VtX|)@!k@xe~oRMPy|>N00pxmH-4#S&<SQ
zGo3D3$xziDS(*s0T}}L`pn?7(X0}<ef#Bc8&rrr_ggS6%VPfw6id;Y+d+`AH4X{mB
z;8KfJqZ6c7s;Edawm5=mO9P=Vf1_P3I1-r|dW`cdHI7uBQ0aWIaL#CbHBE8pV?V(A
zj4Un!>MR2UG7D3%20WV(d88O`@Z$myXdH4RWoJ0k=#E`>mdEw1p&>q<1|R10RoB>b
zCeI0XtR4`&Q&!NUt&OME)%JHFIUQ`bDcqD3D0UnR^wbXmfm}!qUYExX54-5n5^n`y
z;EJvv1RKB_PoSXOG!EJo+26aqA0i7N5&qR7-<}S;$YR%~9V08DvcfbvrV`rGF_Rc>
zrzhtvKG9?JTGJ%p+5@s#?s3UmQ-164Sx~<RbPEID9{~l$AHsIPm%J1@NFB(T0@Q|_
z2(Vs;i%a`pa$vQh1BR><Ee~9cXlVfJ<89o=A5om5y74N00mtr0#wyS}K5FBRxPrEH
zpU40TRFEPp9t+2**~4>#?;nQ9`4l=t?&&y|ZgvTf{{=YI7fb_7gIO&#ETqlSu)tX?
z4GEyVT14Xls|qbLWNqk);3`DV16wC}JWzml=^9+%`cTq<)`pS=vMwOXz0Vej>T%@L
zs{ol|(;Z#q*MARS=0I&OaJbiq`_7^P14Xz$>a-CcT+qP$s!_Y?X2S@$C;=Nyk2s65
z00C(Z*x-UjBdBM8#Qpg!q20j1GXoI}7_zP^oWb}Er1CtjSaj^N9Kw&|z+`%eN3a3`
z#61y|Wup%Op@=_k6{&6&@?-V~Ji6e=hr|w(cN%OC-WVjfcWO}|gR_o0vVqYQbIt~r
zZkx)kc+47KD_<~Es5BdlCMNyTAM~t$Wj?rJ1xZ?Glw45qK$iu~8~!XI$53qkSh%^@
zSpoz`)<KAoJC)pS`eUklIWTOntV()Jf7BKgO<5nyxK3dJC$<6x@{jA^y+6Bavs`ns
zfxcERda&30P`qL9tAHf|VaL4Vx{hz_LD&w>aXD}uN{w*&M$ESzu18!4)w0vbaUC_?
zHDW;uc!*$EMBIg`D+<3<R}*>3BcMjj1IkfLy4xB(X4rOQ*WQU!WoW=&fM(ZoKb&w+
zTjgatxoKu41kQs`>cV{fz!JjeBS;JupX8=94U+U+4-`7ZUn{g~c&(2%Hbs#n|D{ZQ
zk{<A2vFITrKC~SM9c>RGYF5JtJJ45S=;K@3VVEM48By*9%~(+{JozEY8r+XwtGu5V
z@yyO)G+kNx3;`IYEDF!sDw_cLsNn*r0YPeRSK?{jyurwHUOmRX@WM*TzX+z>dn}Nm
z@3Fwj3|C@OXO(_~k3HNZT9kD5#t~~Uco1MT)>4X|y-RJ-F@Rh((W%RG05BR1q9A7V
z*k8ft^^FV;LT4#bzpDzyY|y??6I0C*>Gwvtv)vJis%sP%v$$zq!1PQFRC^b}hM5dM
z#5%4n0P0I|CLihvNS5S4?DxgfeOmze>jd^$Vm$80W55LU;{wmpnt4H<X!N!N68z(m
zp1$+`T8ajT1k&8ueGaj|1)_%@64LPkKzaH00}i(fo=QN&GD3CW6b!5bA6lVjt`@Ts
zz7AB4=sqDBdn1B10BEExqsfTe2alfS1G5nn4Gsy!M)aLomO#1091y{RL3yAP#{wdY
zxV?jTc^GF!g_0>w#k>=FrI>S~uZS5S)+%lLNhj%)*N8Gpo}z@9JG;40H42FdkaVqk
zO6swnvYZ5{yhTlSI%!V!t5Il_kvwZw<e(2Rd&<r<AL{`6rT>)(L}Q|~yK3fr^=XNV
zup%vW2<TE{O0^#+)P>Gc;#6_nps1@e{nc2?^wbVg9m0ytEvp*@9rR_V6`Ai?a61z4
zYdG2GhVaB9`{k_Hx@zOBPC?%M`u2zD*LTsIUw?V~`OT+Y)29EFmdS4a^=|fxrrEtI
z3)bnEHU9e%6XN{7Asu*mA}`Y}Y3HDaYx*bdE0Ej}4%%6(SSHf1$G>I3$(W&i(r0K-
zEXuw-?Gu(E3q~uLwI7d)Q&}hMCQ5TwMCl*Kzh%H>SMRHw4*blxg7B|hz577_RH5S0
zR#g^FcFgjuC4I0R{vsJ+`Z#;2;v|bUX<VnaWTQM`x#z!R@YKXuz%sde@_!8T%cqa;
zzyJK{{VzXG+r{{N)+or<hPK?@|NO)250msvQ?*Jt1O3xd`Q~~yhZizo#1>@5+-p(y
zZMmm5#07p#WLHXQ)+8N`yxN#(MA5h7-wVjBX*y&?>!se)e-;=O_8PHv|GguTfJ{(R
z;Sa>&{Mi3m05|OoyHw{cOU+rF$ljA5yC3O~SJ=FZDvEzH&3sEDtq`|#_xi*8L297j
za=0s4NcP=DhAq$`&a-ndBUrp=jPs{+s1u6Yt|r0V&u3g9Lp{rZCDIETiOig|KkTCB
zVFC+i_IQrQ_X?@udWi*I|9lR}BZjI0dx-^L|9lQePuB7S?rKy~YymC4oZ~;$S&N1<
z<k96j4aq`w!L*Q=(Dl0k6a?(;6p}p?L9i}1w<3AOZKYHoi6Rp?@x>Gqo4H3k;8Ozj
z?`LTJ(<v0$0pB1sRR4Yk)jyp=VTS=bfQLAZE8?f95T{}@;OR>_eqLQ-e#qNIBFHM3
z1PLi!QYNK|hRDd_$s0Lr^JqkE9u=U?qZBhmsKSPW3`PymqokZ+e6J*`$_71TR!qI9
znyMF&b7c*^wcSY6%PCm;w^9l+ez{;DE=MMkQBqDbzLyfwjaOMhXGsi9f~cHI5Rh`w
zv7~{7$jH&j8!2pmz|!Uu1W`GOARy<$KgpO9%Ne;wQ3cm1AmyT4m<=REMvhM2NMVcg
zIU?F^g648^c>Y#Skv$vHvu6cx_AJGuzoWQ_^IMZOCitneoM|c}C2i5O(VP$&IRSYy
zh0W19PHsv>_Id`!9!_z{#17jVtAZLNW@y9nDWKjPkU0VpAAmlh4ot0wVlpy7_GS)C
z_9izPNcw&ON1qB%SJX7|S)up+OA(}h^lGwIF|Lyf&SFBmY`VAjs<Ju?VUa;0qGLn_
z84K4Dn<=g>sb{IQd6>!SGIJ@r`b@Xr*=1vp*Jfal(o)xj%0-mR>M4nYj&2y~o7w17
z9T|D5<Em=J&eo(+Ul+FJD@m!%i9}wPG5nu>p)+AEJrmI}cJ_RoiLEnnRrDe;K4WCT
zI3&>X4q{XG@oa>OZjhMa56@?SblHRsgw&9u%p7!lBZXz>R1LV}E#5onD=D&gn#V*S
z%F1~}2{|1-k<%ENn3#fuy_Dh)2NHiTk0ugPR?Z|!Na-|gz)VO^Fn0#<8>$^~cP2CK
z`dV2?&NvqDB}6nIDPmF5t47Z?6PD4<L?n#;R6$gl*aSftlki%?LN@}Mzi1GXF$^yy
z`0U6`u1D{MWZ3;H8FGI?paELEMsO#8Cjlo_(~=7UwBTGy2F~Bfz>5>Ak#dS`Atr_A
zFQxcA>S!Mpt1%JH<>icXDJ^wfR_5ASi0g@aAqj6m#xl9dJTFCMd=mi)7Yn2!UO1=a
zTZqdz7eW$Ve)QT<n}e8)XLu>YuS+tzsZID?Op4E6O7QWQWq%k|f=kKJ`8x^t5Pnh%
zJ~_m>lfRRI>wB+JL<TGAxk#K7H5`3AL!!^7AQu#C+BJPU1EJ5RAVxE%uyAtYX0~f8
zlcum{{c&*>8hbs1Vh^V{daqFC4odt$@lskP*5c;co%Gc!<t>kD<oQ*QJL#)elAGC-
zn8L+vyV=86hWl!in2M`yb{c_KntND`Ji>J?JH^Xa(ipaS<Yg)I$FHQv!gXjelOwk_
zdP5_#F%t7a&N!<e_!-un_~g}cv!NVdvK4VRefDa)Ur7!y>}1-Fzj4AUNBbtrvt+*2
zh<5#XK6ce~2NXF1j&(HMQ@O?FiI(j4$*{rv^Fokgd|=m#W7IvD*>Ray$)zkww2(bD
z0<$H4wv_+pnGDB$Lq{uJP4dLvydoh-qzHL7rn+J+#W^5xRmg?5^(juU4NFQsfxk#x
z^7(q4{9{+_1~nu&qaDWO`G$zTB3n{o9&5;0#8i6Mi?g&%+0}`4QgYg8$CA7f_GvZq
z^eK84if#~lm2^#8UiXv6wE^r86Ywa5smy>w;=a(5*fEK*Y_M}e)|p=QBce$|<Inm{
z82zt7l<fOBqWON0Nh0>1{BM$KLZ(PL$Vhold4`&=ztsV7k^bLV^4quH=o1qG`u|D7
zGpdhSuFW_X>HnQ1zmkjZE{QHV5!qvrUUZqfJQ<T?kzVu_dala4-A`6uWL803Z^0V#
z15HNDe)q1dcdz+c3m$v2^Bhz4>z^`Bc~Aa)g;i3jqxc`r^5f_CuVj0MddW#!=(h;T
zZH=h48fzr8MGA4YX^aayn>|TdC+^FwLUk3B?bt1<9f;=%f13W*P}0BlboaY8n~;Wv
zquh}{UyasWRxiDcS6X;Mj#dJd+7qe$D5dW{REg{;`v$eAyXo~rQ_<d1?dY#LC^~lM
zv(j!4fP>gu{?`mh)vN{xA}}^Ao^0s<_M~7Y{+<#^k>>r-m6Edj#9hEC9~QtvUL(aQ
z_L2WJ1ENmRUfiUl&TsnNZXzy)O0)P}luer@CbWOt%~MDLnTa)leY+E-r9jwss8C`=
z`ft05xD<*dHvoOT6NIHOta?m1|9vM)OM%!8A_knl?gU{KOjej+zS#-FGZ?-1#4!e`
zPkclo`j|4KHu#=ao~PVfli^DIIU8?Ik(}7mKJh-@&))9)#P*J*1(tnYxm}wU0CyHd
zi$<Q2vlw`YlqJ;B-Wz?5x$x`pw**>Q3<DmOcsI>Hwa%qJ9e)2~IZtECJZY)ROkPl4
z>C;h(pYZ=?Dfy^wKi|EqM3ScXnU6H@g)PexCL8P?ZJz#$iq4RniQJL@&l8tjW2i<Z
za!39@OI*Z{*_r7J4Q;E1nG63i{NDWy|DHitY@n@&<U;+Zd<M}V^l*%8Jp>oJXa6oB
za+FwpAL+gBxxDO?$eBkI#5cR&u+%Hjx0f_{VHh#>_}L3O1RGb+G$RHs_G7<TwuEV^
z?K<Lr5k2zL<!xg|m8SZUJ>Id$yZ-T#(`u|ko4fqm>#m^YmE7yxqWorpTF8H74|e<k
zXSkyJEboH)H~q-&N}LyFz4>fWvpeAd;xJRxC&SjgJdt4Na#Zm(gXVrD?K8#KGDJ2q
z$jf?P#W^W+d7&;?=84{rpF93j#dIxeG<I#nE)6rAB;g%?H1>hy(Lbwb@z8v-jBn;6
zjqKMkl?r%KH#W#$V^d_Vl=gy{`?jZNE@+oQqg7e*sMCMCpxuA|$Ny!HB!Bsz|MCBJ
z?7cEWy;{*B%!&_Jr*~G^0g4@HZ(MW0lMTfE1J1|Vw4|}I7lV6BHIE)!TbO}fl8)cV
zU;F9u)jYETFx8NcEC=Jtl?qnthKs0k`JSPj`Q-vZZ}IquL}Mo>Jo04v(&E$nvgxPc
z9V_-B^dWvnj*~5<NRQnA*fSYvo~N*jBsMf{Ry8)*RPGc0A4k!QQ>FtHK5E)MW$i`u
zT-&<rjybuVwWPi1Q&~Tp^75&XLMmiA$K>^Hcx~ijO;MMTdBNzjq5)H9a&X@Jk0!s_
zhf#l>{v>@|T<o)bzoxLy+C5pE+4pgg>pdyovTO09bJb0=J^;F5yorn5d%lRV%UUvK
z#Y4AVc-IT#M?Ze}{Liy&{*DKDdz|ip9+oYys4S2^Q_>VH3*^;~ydr08`&T1jTB2ru
zz0odpUm%ncfFm(i_&cf?yC_MQ$Mv3V=T~LI5#=ow-(7FC)7}wrlC>i-cRR8$#Km_^
ze9SEO8L2n09JP)4SqgH$l02b;1E>yX9QO^<<<p)OIeD_ru4GRNk{yE<P;_a2DxY=<
zttd*?zhrjupC`0wi4pLZ?FEl<Uv{mIR+)y-rX^;@FDEW2ekq~F!#eXf&Ljs4GBTWf
zPeux)r8(=Tm`dKTR1CT#?@1B<=YRa4mVOTFW3Wo~um6icOOs<GTbvOebL^ikh<KOS
zVIrq_ntfbV2rDa7;iRl<JY3r!am#Kl*@-OplGsvhhN}>ZkYp@IjWXW>Ia&Sn;C!^S
zqu@I%wWHuW2HrQa9(;Esn!~_%9Gr-dWEA``78_N{p;uRa=}hcB|J@($AK-+QT_u(%
z^IrHdqTsSPl^u2_V_Z<@Jsyo*;zu&WUgA%viHnJQjH{BsGWmO~R=8lOD7H>mbcY8z
zM0D4S?t0Nt&ALXT5A5Y<daQ;gc)G+bYd-jYq-&s6%yw9lm=D@ww@dtQSc)}A{6FeW
zkcwtfB#gwYfl|=yghAt+IMFv&I!Twtd$LZ`RiyXBCSTYDq{^EE49#VB!ny3_SxLY7
z{Py*ZEC_WK=|^$b!Q|)pZT7TuB_qpumL`&nwb`+%VMEDFcE0R$5(Hs~?Qy-Fm*G!)
z`!|U12k~F~KVB`SXoR_$`!dPmd`C^z4^4|(9vM!*F7O}u?gMSU|3K^JzrO0P2vZFu
z#x-azIN!Z6tTzo%radlmxe<vzA7(CJ%_&{8_DP)R7xOtY6+;2eHit}Y#N;(wfW`_(
z_KocAwwf#xEVrJ>Ni4U)OogXgV&=fyn^}4)OWcVpcY3J5vp}u7ebZIT6*mc}Ad219
zUH|tIvdgR?1yMY%Q9@?JJedJ-6db13mSD<`M}nv9fOM$x83{f5kVa$~rX&<>)VqzZ
z#~Ay_W^I!vdOtkf-3?C{gx5G}i2Jy|eog;zDMW4BJN%N&%J1*K-u(mlSN#6}y2wYB
zRkn>w6vcll8y2}6nua7IT|#@oe_}@rXM~I{EmY2w{1bb>Bky<Y{SLq1_3szi>8f*0
zr>p1`Syzm;X3B{s<!AniT2}GZt6?qbl+}3nNJAF7b_B!H74ar0&W_@Y;B*W|pEaw&
z>8Ooiyt3E&uM`?CPSMY2D8xy60$`q|zm}<FA6IqT^B1gtx?_Pf*^aIPJrM`<_r)HC
zu}lOJ$>ziU4F#I)s&z)<6J4*cGmDO8`7o=c;qi16*)RO@nC{PfYg>e#$!X8&-g=D~
zD<J-O#~<sE4_S39j(_5t?e-H!N#5n0lJskGk>nkAMweT3jPydnH^ZM6g#FK9FQUZX
zHFC=_eqc{`!!Ad97S^1jQ<mrZYu3<xZRDa|=c<q)yu;6S{j(i=W|bFg1C#PG#_CVD
zTxA6|&Sf1x?@2_ngoYd>%Z^V775h_u+y8}HP4d^*&YAY0dWRG5#>8ayizVKTiFbYC
z9ZgJ4I8AI(3?u?8>z&XV{a7UbnLVI}-cl>5>5tP#@_Ffga=)IddT62OQXqf1o4&Nj
zCQpEP<<YfGOmoCWJU-*jsbirRNzpeInXh!q{0f5KcoMD?eUjbrwi}X=ar=p^$&e&#
zKJ<#ms^nbqK2GYgY4&3X6pxRSgDyUZi|v@$js;V({|Kg>l<|3g+o!i8zGm3nBn_-@
zKsPes%-Y?`k8&jWZA`{L6DQm8WVHSAWYmiwwf7g{qshkCsA=nt1x}E6pfjQIP{D3|
zsK)Gz@#{UkqOzY$CEuen5;Dp^Nddk-?C$CH%WM{weDTa$$n*0JTRa$_Q%J>BhaqEn
zq;C%M!kX?bG-h=M?~^r6wY?|1s^f%A%hF4Au*;T{tAQiwK}eT$e8Oaqqu<Zvqy=FC
zDp4y{$yjDlRIr&|t;yOqeb#h(?v^u(^bOD0OwFzH^)1_2&qU1{Z<|Fq=(mSS#e3ED
zU-$ggzCtx!tOLlm6!Wi#-AAsU!t*t`7R!YghJi0g$txH<8(-61X^42)Mgo^~DofDI
zjk`mfK)o!1G>dC3fg)?%5@>d!t<B;Ta<C1PeQIm6{Ovr;JhBd4jtSzoiwu*PJ1g@0
zSH((zdoBkYYy&=d=;L?fEK8de9rfveGAD!l?m&;5vd3OcXlY1#L9qN=FgRHUdUh=$
z0VTS{nvGg&m?6WTF+SD64&zI5Xl}+PDGe##GPNcHAIhd3&3q&~<!=;bUrAmH9l>z!
z-<;gRtm<-LXG`^&uyLg<@}Tvf^f~xeb`Z*voTdEs!1jMyQ~s5o6QeDK+z_*heF5@U
z;*zoRYNUOUXV!U=?wPUA2-`fv{<Y?;CK{|b0cn{B6_#%+Axkssbb}-$3pWB4?{%mp
z9TS!Xe!%AqB1y#dbqKSapNoTUalz-N`GNJ7S>>bI@~I$`efx$@2v0m1GBhdol28#I
z-)7JFwinzn!KoKTjvbf%F+i^ZF#7%OUnn~IOMiY<9@(Z-t<_(XPZ72^1dG04MSf{b
zNzeNGWOsr}lk@Zc{+Z>Ar=u|<yL$`1GPD=lH{8eAvcj5!jx4A|KbGYgXX<U8Jt8Nw
zwM=fZS+}yQcepXLhD+1$So$4Jzb4Ck@r^8HGUv6A(9xfK83r}-T?1>s(uI&Nvof!_
zKiW<fPsE|sur*<Q0kDlPN+Zg41fFTSeLB>&e>u4Gvc+_z74Hu@z^KL8pxBPc5__9i
z0oR!%WfAB}#_$eozh(d0+Gc4!Zhdc{{uYm}K$+7kxxa8vntuotRH5I|TNA0#0a7QD
zl9*=kSbl;UC`D4jlWi8nfQo1R$qV35kpyS~{T7BV`xqACW&PLiZ8?2EIvnxiVO<-t
zCf`-Wi=^wun)OY`?_9(DVMs$eg<-feq#>ICc$&qEs7KjMGS!D3z`2rb%YrBnJ)KYr
zlMf#R`Bty>pI&!`IIJxB`Ij+gKp$;EMB<nTL)nsVB^y89<3pKC9Ph^<r2gMD#bkbV
z%z7=~oJ##&YQRaqNnHpE5Tdv6$iRzuw#*JlJxii!0ZPb)@iUc`I<rK_6=e|5wp;x9
z;}5Uj-u)mSekg?5VB>!j<*v=z9R1H%vn`5pFkvg^xOdS#Fu+%numF>jAwA{fT*!C^
zq{=Gm2$|xZ<rOFGztWtj7k}*e9#PfKQ7!an;Q9?bz0nqSz8bc~ez>(4Htjy{jI9y9
zg=O)v<N2MhoR5zR$N6<9*S}I7Qg=SV|JCyRzIgoJKK-!5(jao7$}}N+)}JC`?>`WY
z>EqFmech{eUElNceT|aRWr59xNH>u=(#<R?&MfH7^uWL2{QOso)A*8s&t!LUe;K-6
zu)9&*HlW>@!X{$fC~Xy%ZWOl&JvR#51ehD8EuzRxgiW#HsPGWHN*k9$h*~O*dquMr
zsm-3wJh6Z$%itnY8#s#tvQJSPXL;jpUHKWEp~EB>?<|zs;qcq9RbK`M!_5C00Q%pr
zmT_psu1OK4Rvn{O6YN@nD~*Y*kBF^HJqGy_)2<E8t%dT*V~t23&wa~N_RPk;<B1Ws
zd>Y|Casjp@oC=61R9>mAyu99|eZlgSr&^Wuu!^WA)3ct^6DxAu`H>dc(HOb(G;pKq
z85~K{mCVkEc4TQ~$9R<FWrf)2M>Y$hA)crEmHoz?Y<@yXh}}babV@c&so|R?yYGMY
tQ;~A(FV_-HXf&1cx-_wJ=!pRy&9C|N`B&1z=H{q>U$K~f{~!Bl6w-5qU8w*7

diff --git a/overview.html b/overview.html
index 8bc758484..34457c079 100644
--- a/overview.html
+++ b/overview.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Overview &#8212; SLEAP (v1.3.3)</title>
+    <title>Overview &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'overview';</script>
     <link rel="canonical" href="/develop/overview.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Installation" href="installation.html" />
     <link rel="prev" title="Social LEAP Estimates Animal Poses (SLEAP)" href="index.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Overview</h1>
                 <article class="bd-article" role="main">
                   
   <section id="overview">
-<span id="high-level-overview"></span><h1>Overview<a class="headerlink" href="#overview" title="Permalink to this headline">#</a></h1>
+<span id="high-level-overview"></span><h1>Overview<a class="headerlink" href="#overview" title="Permalink to this heading">#</a></h1>
 <img alt="_images/workflow.png" src="_images/workflow.png" />
 <p>In a nutshell, SLEAP allows you to train and use deep learning-based models that can automatically track the movements of any body parts of any number or type of animal from recorded videos. This enables precise and quantitative analysis of behavioral data.</p>
 <p>Below is a high-level overview of a typical end-to-end workflow in SLEAP. You can also step through this workflow yourself in the <a class="reference internal" href="tutorials/tutorial.html#tutorial"><span class="std std-ref">Tutorial</span></a> section.</p>
@@ -452,7 +450,7 @@ <h1>Overview</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/py-modindex.html b/py-modindex.html
index a822c4135..caefe310e 100644
--- a/py-modindex.html
+++ b/py-modindex.html
@@ -3,12 +3,12 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <title>Python Module Index &#8212; SLEAP (v1.3.3)</title>
+    <title>Python Module Index &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -31,8 +31,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -41,15 +41,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'py-modindex';</script>
     <link rel="canonical" href="/develop/py-modindex.html" />
     <link rel="shortcut icon" href="_static/favicon.ico"/>
@@ -57,7 +54,7 @@
     <link rel="search" title="Search" href="search.html" />
 
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
 
 
   </head>
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -376,11 +374,6 @@ <h1>Python Module Index</h1>
        <td>&#160;&#160;&#160;
        <a href="api/sleap.instance.html#module-sleap.instance"><code class="xref">sleap.instance</code></a></td><td>
        <em></em></td></tr>
-     <tr class="cg-1">
-       <td></td>
-       <td>&#160;&#160;&#160;
-       <a href="api/sleap.io.asyncvideo.html#module-sleap.io.asyncvideo"><code class="xref">sleap.io.asyncvideo</code></a></td><td>
-       <em></em></td></tr>
      <tr class="cg-1">
        <td></td>
        <td>&#160;&#160;&#160;
@@ -798,7 +791,7 @@ <h1>Python Module Index</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/search.html b/search.html
index ae6e5fdb3..829af95d4 100644
--- a/search.html
+++ b/search.html
@@ -3,11 +3,11 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><title>Search - SLEAP (v1.3.3)</title>
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><title>Search - SLEAP (v1.4.1)</title>
   
   
   
@@ -30,8 +30,8 @@
     <link rel="stylesheet" type="text/css" href="_static/pygments.css" />
     <link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -40,15 +40,12 @@
     <script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
     <script src="_static/jquery.js"></script>
     <script src="_static/underscore.js"></script>
+    <script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="_static/doctools.js"></script>
+    <script src="_static/sphinx_highlight.js"></script>
     <script src="_static/clipboard.min.js"></script>
     <script src="_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="_static/togglebutton.js"></script>
     <script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'search';</script>
   <script src="_static/searchtools.js"></script>
   <script src="_static/language_data.js"></script>
@@ -58,7 +55,7 @@
     <link rel="index" title="Index" href="genindex.html" />
     <link rel="search" title="Search" href="#" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -164,6 +161,7 @@
 <li class="toctree-l2"><a class="reference internal" href="guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -375,7 +373,7 @@ <h1>Search</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/searchindex.js b/searchindex.js
index f2078a9a3..7820c5cdc 100644
--- a/searchindex.js
+++ b/searchindex.js
@@ -1 +1 @@
-Search.setIndex({docnames:["CODE_OF_CONDUCT","CONTRIBUTING","api","api/sleap.info.align","api/sleap.info.feature_suggestions","api/sleap.info.labels","api/sleap.info.metrics","api/sleap.info.summary","api/sleap.info.trackcleaner","api/sleap.info.write_tracking_h5","api/sleap.instance","api/sleap.io.asyncvideo","api/sleap.io.convert","api/sleap.io.dataset","api/sleap.io.format.adaptor","api/sleap.io.format.alphatracker","api/sleap.io.format.coco","api/sleap.io.format.csv","api/sleap.io.format.deeplabcut","api/sleap.io.format.deepposekit","api/sleap.io.format.dispatch","api/sleap.io.format.filehandle","api/sleap.io.format.genericjson","api/sleap.io.format.hdf5","api/sleap.io.format.labels_json","api/sleap.io.format.leap_matlab","api/sleap.io.format.main","api/sleap.io.format.ndx_pose","api/sleap.io.format.nix","api/sleap.io.format.sleap_analysis","api/sleap.io.format.text","api/sleap.io.legacy","api/sleap.io.pathutils","api/sleap.io.video","api/sleap.io.videowriter","api/sleap.io.visuals","api/sleap.message","api/sleap.nn.architectures.common","api/sleap.nn.architectures.encoder_decoder","api/sleap.nn.architectures.hourglass","api/sleap.nn.architectures.hrnet","api/sleap.nn.architectures.leap","api/sleap.nn.architectures.pretrained_encoders","api/sleap.nn.architectures.resnet","api/sleap.nn.architectures.unet","api/sleap.nn.architectures.upsampling","api/sleap.nn.callbacks","api/sleap.nn.config.data","api/sleap.nn.config.model","api/sleap.nn.config.optimization","api/sleap.nn.config.outputs","api/sleap.nn.config.training_job","api/sleap.nn.config.utils","api/sleap.nn.data.augmentation","api/sleap.nn.data.confidence_maps","api/sleap.nn.data.dataset_ops","api/sleap.nn.data.edge_maps","api/sleap.nn.data.general","api/sleap.nn.data.grouping","api/sleap.nn.data.identity","api/sleap.nn.data.inference","api/sleap.nn.data.instance_centroids","api/sleap.nn.data.instance_cropping","api/sleap.nn.data.normalization","api/sleap.nn.data.offset_regression","api/sleap.nn.data.pipelines","api/sleap.nn.data.providers","api/sleap.nn.data.resizing","api/sleap.nn.data.training","api/sleap.nn.data.utils","api/sleap.nn.evals","api/sleap.nn.heads","api/sleap.nn.identity","api/sleap.nn.inference","api/sleap.nn.losses","api/sleap.nn.model","api/sleap.nn.paf_grouping","api/sleap.nn.peak_finding","api/sleap.nn.system","api/sleap.nn.tracker.components","api/sleap.nn.tracker.kalman","api/sleap.nn.tracking","api/sleap.nn.training","api/sleap.nn.utils","api/sleap.nn.viz","api/sleap.skeleton","api/sleap.util","datasets","guides/choosing-models","guides/cli","guides/colab","guides/custom-training","guides/gui","guides/index","guides/merging","guides/proofreading","guides/remote","guides/skeletons","guides/training","guides/troubleshooting-workflows","help","index","installation","notebooks/Analysis_examples","notebooks/Data_structures","notebooks/Interactive_and_realtime_inference","notebooks/Interactive_and_resumable_training","notebooks/Model_evaluation","notebooks/Post_inference_tracking","notebooks/Training_and_inference_on_an_example_dataset","notebooks/Training_and_inference_using_Google_Drive","notebooks/analysis_example/README","notebooks/index","overview","tutorials/analysis","tutorials/assisted-labeling","tutorials/initial-labeling","tutorials/initial-training","tutorials/new-project","tutorials/proofreading","tutorials/tutorial"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":5,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":3,"sphinx.domains.rst":2,"sphinx.domains.std":2,sphinx:56},filenames:["CODE_OF_CONDUCT.md","CONTRIBUTING.md","api.rst","api/sleap.info.align.rst","api/sleap.info.feature_suggestions.rst","api/sleap.info.labels.rst","api/sleap.info.metrics.rst","api/sleap.info.summary.rst","api/sleap.info.trackcleaner.rst","api/sleap.info.write_tracking_h5.rst","api/sleap.instance.rst","api/sleap.io.asyncvideo.rst","api/sleap.io.convert.rst","api/sleap.io.dataset.rst","api/sleap.io.format.adaptor.rst","api/sleap.io.format.alphatracker.rst","api/sleap.io.format.coco.rst","api/sleap.io.format.csv.rst","api/sleap.io.format.deeplabcut.rst","api/sleap.io.format.deepposekit.rst","api/sleap.io.format.dispatch.rst","api/sleap.io.format.filehandle.rst","api/sleap.io.format.genericjson.rst","api/sleap.io.format.hdf5.rst","api/sleap.io.format.labels_json.rst","api/sleap.io.format.leap_matlab.rst","api/sleap.io.format.main.rst","api/sleap.io.format.ndx_pose.rst","api/sleap.io.format.nix.rst","api/sleap.io.format.sleap_analysis.rst","api/sleap.io.format.text.rst","api/sleap.io.legacy.rst","api/sleap.io.pathutils.rst","api/sleap.io.video.rst","api/sleap.io.videowriter.rst","api/sleap.io.visuals.rst","api/sleap.message.rst","api/sleap.nn.architectures.common.rst","api/sleap.nn.architectures.encoder_decoder.rst","api/sleap.nn.architectures.hourglass.rst","api/sleap.nn.architectures.hrnet.rst","api/sleap.nn.architectures.leap.rst","api/sleap.nn.architectures.pretrained_encoders.rst","api/sleap.nn.architectures.resnet.rst","api/sleap.nn.architectures.unet.rst","api/sleap.nn.architectures.upsampling.rst","api/sleap.nn.callbacks.rst","api/sleap.nn.config.data.rst","api/sleap.nn.config.model.rst","api/sleap.nn.config.optimization.rst","api/sleap.nn.config.outputs.rst","api/sleap.nn.config.training_job.rst","api/sleap.nn.config.utils.rst","api/sleap.nn.data.augmentation.rst","api/sleap.nn.data.confidence_maps.rst","api/sleap.nn.data.dataset_ops.rst","api/sleap.nn.data.edge_maps.rst","api/sleap.nn.data.general.rst","api/sleap.nn.data.grouping.rst","api/sleap.nn.data.identity.rst","api/sleap.nn.data.inference.rst","api/sleap.nn.data.instance_centroids.rst","api/sleap.nn.data.instance_cropping.rst","api/sleap.nn.data.normalization.rst","api/sleap.nn.data.offset_regression.rst","api/sleap.nn.data.pipelines.rst","api/sleap.nn.data.providers.rst","api/sleap.nn.data.resizing.rst","api/sleap.nn.data.training.rst","api/sleap.nn.data.utils.rst","api/sleap.nn.evals.rst","api/sleap.nn.heads.rst","api/sleap.nn.identity.rst","api/sleap.nn.inference.rst","api/sleap.nn.losses.rst","api/sleap.nn.model.rst","api/sleap.nn.paf_grouping.rst","api/sleap.nn.peak_finding.rst","api/sleap.nn.system.rst","api/sleap.nn.tracker.components.rst","api/sleap.nn.tracker.kalman.rst","api/sleap.nn.tracking.rst","api/sleap.nn.training.rst","api/sleap.nn.utils.rst","api/sleap.nn.viz.rst","api/sleap.skeleton.rst","api/sleap.util.rst","datasets.md","guides/choosing-models.md","guides/cli.md","guides/colab.rst","guides/custom-training.md","guides/gui.md","guides/index.md","guides/merging.rst","guides/proofreading.md","guides/remote.md","guides/skeletons.md","guides/training.rst","guides/troubleshooting-workflows.md","help.md","index.rst","installation.md","notebooks/Analysis_examples.ipynb","notebooks/Data_structures.ipynb","notebooks/Interactive_and_realtime_inference.ipynb","notebooks/Interactive_and_resumable_training.ipynb","notebooks/Model_evaluation.ipynb","notebooks/Post_inference_tracking.ipynb","notebooks/Training_and_inference_on_an_example_dataset.ipynb","notebooks/Training_and_inference_using_Google_Drive.ipynb","notebooks/analysis_example/README.md","notebooks/index.rst","overview.rst","tutorials/analysis.rst","tutorials/assisted-labeling.md","tutorials/initial-labeling.rst","tutorials/initial-training.rst","tutorials/new-project.md","tutorials/proofreading.rst","tutorials/tutorial.rst"],objects:{"sleap.info":[[3,0,0,"-","align"],[4,0,0,"-","feature_suggestions"],[5,0,0,"-","labels"],[6,0,0,"-","metrics"],[7,0,0,"-","summary"],[8,0,0,"-","trackcleaner"],[9,0,0,"-","write_tracking_h5"]],"sleap.info.align":[[3,1,1,"","align_instance_points"],[3,1,1,"","align_instances"],[3,1,1,"","align_instances_on_most_stable"],[3,1,1,"","get_instances_points"],[3,1,1,"","get_mean_and_std_for_points"],[3,1,1,"","get_most_stable_node_pair"],[3,1,1,"","get_stable_node_pairs"],[3,1,1,"","get_template_points_array"]],"sleap.info.feature_suggestions":[[4,2,1,"","FrameGroupSet"],[4,2,1,"","FrameItem"],[4,2,1,"","ItemStack"],[4,2,1,"","ParallelFeaturePipeline"]],"sleap.info.feature_suggestions.FrameGroupSet":[[4,3,1,"","all_items"],[4,4,1,"","append_to_group"],[4,4,1,"","extend_group_items"],[4,4,1,"","get_item_group"],[4,5,1,"","group_data"],[4,3,1,"","groups"],[4,5,1,"","groupset_data"],[4,5,1,"","item_group"],[4,5,1,"","method"],[4,4,1,"","sample"]],"sleap.info.feature_suggestions.ItemStack":[[4,4,1,"","brisk_bag_of_features"],[4,3,1,"","current_groupset"],[4,5,1,"","data"],[4,4,1,"","extend_ownership"],[4,4,1,"","flatten"],[4,4,1,"","get_all_items_from_group"],[4,4,1,"","get_item_data"],[4,4,1,"","get_item_data_idxs"],[4,4,1,"","get_raw_images"],[4,5,1,"","group_sets"],[4,4,1,"","hog_bag_of_features"],[4,5,1,"","items"],[4,4,1,"","kmeans"],[4,4,1,"","make_sample_group"],[4,5,1,"","meta"],[4,5,1,"","ownership"],[4,4,1,"","pca"],[4,4,1,"","sample_groups"]],"sleap.info.feature_suggestions.ParallelFeaturePipeline":[[4,4,1,"","get"],[4,4,1,"","make"],[4,4,1,"","run"],[4,4,1,"","tuples_to_suggestions"]],"sleap.info.metrics":[[6,1,1,"","calculate_pairwise_cost"],[6,1,1,"","compare_instance_lists"],[6,1,1,"","list_points_array"],[6,1,1,"","match_instance_lists"],[6,1,1,"","match_instance_lists_nodewise"],[6,1,1,"","matched_instance_distances"],[6,1,1,"","nodeless_point_dist"],[6,1,1,"","point_dist"],[6,1,1,"","point_match_count"],[6,1,1,"","point_nonmatch_count"]],"sleap.info.summary":[[7,2,1,"","StatisticSeries"]],"sleap.info.summary.StatisticSeries":[[7,4,1,"","get_instance_score_series"],[7,4,1,"","get_point_count_series"],[7,4,1,"","get_point_displacement_series"],[7,4,1,"","get_point_score_series"],[7,4,1,"","get_primary_point_displacement_series"]],"sleap.info.trackcleaner":[[8,1,1,"","fit_tracks"]],"sleap.info.write_tracking_h5":[[9,1,1,"","get_edges_as_np_strings"],[9,1,1,"","get_nodes_as_np_strings"],[9,1,1,"","get_occupancy_and_points_matrices"],[9,1,1,"","get_tracks_as_np_strings"],[9,1,1,"","main"],[9,1,1,"","remove_empty_tracks_from_matrices"],[9,1,1,"","write_csv_file"],[9,1,1,"","write_occupancy_file"]],"sleap.instance":[[10,2,1,"","Instance"],[10,2,1,"","LabeledFrame"],[10,2,1,"","Point"],[10,2,1,"","PointArray"],[10,2,1,"","PredictedInstance"],[10,2,1,"","PredictedPoint"],[10,2,1,"","PredictedPointArray"],[10,2,1,"","Track"],[10,1,1,"","make_instance_cattr"]],"sleap.instance.Instance":[[10,3,1,"","bounding_box"],[10,3,1,"","centroid"],[10,4,1,"","fill_missing"],[10,3,1,"","frame_idx"],[10,4,1,"","from_numpy"],[10,4,1,"","from_pointsarray"],[10,4,1,"","get_points_array"],[10,4,1,"","matches"],[10,3,1,"","midpoint"],[10,3,1,"","n_visible_points"],[10,3,1,"","nodes"],[10,3,1,"","nodes_points"],[10,4,1,"","numpy"],[10,3,1,"","points"],[10,3,1,"","points_array"],[10,4,1,"","transform_points"],[10,3,1,"","video"]],"sleap.instance.LabeledFrame":[[10,4,1,"","complex_frame_merge"],[10,4,1,"","complex_merge_between"],[10,4,1,"","find"],[10,3,1,"","has_predicted_instances"],[10,3,1,"","has_tracked_instances"],[10,3,1,"","has_user_instances"],[10,3,1,"","image"],[10,4,1,"","index"],[10,4,1,"","insert"],[10,3,1,"","instances"],[10,3,1,"","instances_to_show"],[10,4,1,"","merge_frames"],[10,3,1,"","n_predicted_instances"],[10,3,1,"","n_tracked_instances"],[10,3,1,"","n_user_instances"],[10,4,1,"","numpy"],[10,4,1,"","plot"],[10,4,1,"","plot_predicted"],[10,3,1,"","predicted_instances"],[10,4,1,"","remove_empty_instances"],[10,4,1,"","remove_untracked"],[10,3,1,"","tracked_instances"],[10,3,1,"","training_instances"],[10,3,1,"","unused_predictions"],[10,3,1,"","user_instances"]],"sleap.instance.Point":[[10,5,1,"","dtype"],[10,4,1,"","isnan"],[10,4,1,"","numpy"]],"sleap.instance.PointArray":[[10,4,1,"","from_array"],[10,4,1,"","make_default"]],"sleap.instance.PredictedInstance":[[10,4,1,"","from_arrays"],[10,4,1,"","from_instance"],[10,4,1,"","from_numpy"],[10,4,1,"","from_pointsarray"],[10,3,1,"","points_and_scores_array"],[10,3,1,"","scores"]],"sleap.instance.PredictedPoint":[[10,5,1,"","dtype"],[10,4,1,"","from_point"]],"sleap.instance.PredictedPointArray":[[10,4,1,"","to_array"]],"sleap.instance.Track":[[10,4,1,"","matches"]],"sleap.io":[[11,0,0,"-","asyncvideo"],[12,0,0,"-","convert"],[13,0,0,"-","dataset"],[31,0,0,"-","legacy"],[32,0,0,"-","pathutils"],[33,0,0,"-","video"],[34,0,0,"-","videowriter"],[35,0,0,"-","visuals"]],"sleap.io.asyncvideo":[[11,2,1,"","AsyncVideo"],[11,2,1,"","AsyncVideoServer"]],"sleap.io.asyncvideo.AsyncVideo":[[11,3,1,"","chunks"],[11,4,1,"","close"],[11,4,1,"","from_video"],[11,4,1,"","load_by_chunk"]],"sleap.io.asyncvideo.AsyncVideoServer":[[11,4,1,"","run"]],"sleap.io.convert":[[12,1,1,"","main"]],"sleap.io.dataset":[[13,2,1,"","Labels"],[13,2,1,"","LabelsDataCache"],[13,1,1,"","find_path_using_paths"],[13,1,1,"","load_file"]],"sleap.io.dataset.Labels":[[13,4,1,"","add_instance"],[13,4,1,"","add_suggestion"],[13,4,1,"","add_track"],[13,4,1,"","add_video"],[13,3,1,"","all_instances"],[13,4,1,"","append"],[13,4,1,"","append_suggestions"],[13,4,1,"","clear_suggestions"],[13,4,1,"","complex_merge_between"],[13,4,1,"","copy"],[13,4,1,"","delete_suggestions"],[13,4,1,"","describe"],[13,4,1,"","export"],[13,4,1,"","export_nwb"],[13,4,1,"","extend_from"],[13,4,1,"","extract"],[13,4,1,"","find"],[13,4,1,"","find_first"],[13,4,1,"","find_last"],[13,4,1,"","find_suggestion"],[13,4,1,"","find_track_occupancy"],[13,4,1,"","finish_complex_merge"],[13,4,1,"","frames"],[13,4,1,"","get"],[13,4,1,"","get_next_suggestion"],[13,4,1,"","get_suggestions"],[13,4,1,"","get_track_count"],[13,4,1,"","get_track_occupancy"],[13,4,1,"","get_unlabeled_suggestion_inds"],[13,4,1,"","get_video_suggestions"],[13,4,1,"","has_frame"],[13,3,1,"","has_missing_videos"],[13,4,1,"","index"],[13,4,1,"","insert"],[13,4,1,"","instance_count"],[13,4,1,"","instances"],[13,3,1,"","is_multi_instance"],[13,5,1,"","labeled_frames"],[13,3,1,"","labels"],[13,4,1,"","load_file"],[13,4,1,"","make_video_callback"],[13,4,1,"","merge_container_dicts"],[13,4,1,"","merge_matching_frames"],[13,4,1,"","merge_nodes"],[13,5,1,"","negative_anchors"],[13,4,1,"","numpy"],[13,3,1,"","predicted_instances"],[13,5,1,"","provenance"],[13,4,1,"","remove"],[13,4,1,"","remove_all_tracks"],[13,4,1,"","remove_empty_frames"],[13,4,1,"","remove_empty_instances"],[13,4,1,"","remove_frame"],[13,4,1,"","remove_frames"],[13,4,1,"","remove_instance"],[13,4,1,"","remove_predictions"],[13,4,1,"","remove_suggestion"],[13,4,1,"","remove_track"],[13,4,1,"","remove_untracked_instances"],[13,4,1,"","remove_unused_tracks"],[13,4,1,"","remove_user_instances"],[13,4,1,"","remove_video"],[13,4,1,"","save"],[13,4,1,"","save_file"],[13,4,1,"","save_frame_data_hdf5"],[13,4,1,"","save_frame_data_imgstore"],[13,4,1,"","set_suggestions"],[13,3,1,"","skeleton"],[13,5,1,"","skeletons"],[13,4,1,"","split"],[13,5,1,"","suggestions"],[13,4,1,"","to_dict"],[13,4,1,"","to_json"],[13,4,1,"","to_pipeline"],[13,4,1,"","track_set_instance"],[13,4,1,"","track_swap"],[13,5,1,"","tracks"],[13,3,1,"","unlabeled_suggestions"],[13,3,1,"","user_instances"],[13,3,1,"","user_labeled_frame_inds"],[13,3,1,"","user_labeled_frames"],[13,3,1,"","video"],[13,5,1,"","videos"],[13,4,1,"","with_user_labels_only"]],"sleap.io.dataset.LabelsDataCache":[[13,4,1,"","add_instance"],[13,4,1,"","add_track"],[13,4,1,"","find_fancy_frame_idxs"],[13,4,1,"","find_frames"],[13,4,1,"","get_filtered_frame_idxs"],[13,4,1,"","get_frame_count"],[13,4,1,"","get_track_occupancy"],[13,4,1,"","get_video_track_occupancy"],[13,4,1,"","remove_frame"],[13,4,1,"","remove_instance"],[13,4,1,"","remove_video"],[13,4,1,"","track_swap"],[13,4,1,"","update"],[13,4,1,"","update_counts_for_frame"]],"sleap.io.format":[[14,0,0,"-","adaptor"],[15,0,0,"-","alphatracker"],[16,0,0,"-","coco"],[17,0,0,"-","csv"],[18,0,0,"-","deeplabcut"],[19,0,0,"-","deepposekit"],[20,0,0,"-","dispatch"],[21,0,0,"-","filehandle"],[22,0,0,"-","genericjson"],[23,0,0,"-","hdf5"],[24,0,0,"-","labels_json"],[25,0,0,"-","leap_matlab"],[26,0,0,"-","main"],[27,0,0,"-","ndx_pose"],[28,0,0,"-","nix"],[29,0,0,"-","sleap_analysis"],[30,0,0,"-","text"]],"sleap.io.format.adaptor":[[14,2,1,"","Adaptor"],[14,2,1,"","SleapObjectType"]],"sleap.io.format.adaptor.Adaptor":[[14,3,1,"","all_exts"],[14,4,1,"","can_read_file"],[14,4,1,"","can_write_filename"],[14,3,1,"","default_ext"],[14,4,1,"","does_match_ext"],[14,4,1,"","does_read"],[14,4,1,"","does_write"],[14,3,1,"","formatted_ext_options"],[14,3,1,"","handles"],[14,3,1,"","name"],[14,4,1,"","read"],[14,4,1,"","write"]],"sleap.io.format.alphatracker":[[15,2,1,"","AlphaTrackerAdaptor"]],"sleap.io.format.alphatracker.AlphaTrackerAdaptor":[[15,3,1,"","all_exts"],[15,4,1,"","can_read_file"],[15,4,1,"","can_write_filename"],[15,3,1,"","default_ext"],[15,4,1,"","does_match_ext"],[15,4,1,"","does_read"],[15,4,1,"","does_write"],[15,3,1,"","formatted_ext_options"],[15,4,1,"","get_alpha_tracker_frame_dict"],[15,4,1,"","get_alpha_tracker_instance_dict"],[15,4,1,"","get_alpha_tracker_point_dict"],[15,3,1,"","handles"],[15,4,1,"","make_video_for_image_list"],[15,3,1,"","name"],[15,4,1,"","read"],[15,4,1,"","write"]],"sleap.io.format.coco":[[16,2,1,"","LabelsCocoAdaptor"]],"sleap.io.format.coco.LabelsCocoAdaptor":[[16,3,1,"","all_exts"],[16,4,1,"","can_read_file"],[16,4,1,"","can_write_filename"],[16,3,1,"","default_ext"],[16,4,1,"","does_read"],[16,4,1,"","does_write"],[16,3,1,"","handles"],[16,3,1,"","name"],[16,4,1,"","read"]],"sleap.io.format.csv":[[17,2,1,"","CSVAdaptor"]],"sleap.io.format.csv.CSVAdaptor":[[17,3,1,"","all_exts"],[17,4,1,"","can_read_file"],[17,4,1,"","can_write_filename"],[17,3,1,"","default_ext"],[17,4,1,"","does_read"],[17,4,1,"","does_write"],[17,3,1,"","handles"],[17,3,1,"","name"],[17,4,1,"","write"]],"sleap.io.format.deeplabcut":[[18,2,1,"","LabelsDeepLabCutCsvAdaptor"],[18,2,1,"","LabelsDeepLabCutYamlAdaptor"]],"sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor":[[18,3,1,"","all_exts"],[18,4,1,"","can_read_file"],[18,4,1,"","can_write_filename"],[18,3,1,"","default_ext"],[18,4,1,"","does_read"],[18,4,1,"","does_write"],[18,3,1,"","handles"],[18,4,1,"","make_video_for_image_list"],[18,3,1,"","name"],[18,4,1,"","read"]],"sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor":[[18,3,1,"","all_exts"],[18,4,1,"","can_read_file"],[18,4,1,"","can_write_filename"],[18,3,1,"","default_ext"],[18,4,1,"","does_read"],[18,4,1,"","does_write"],[18,3,1,"","handles"],[18,3,1,"","name"],[18,4,1,"","read"]],"sleap.io.format.deepposekit":[[19,2,1,"","LabelsDeepPoseKitAdaptor"]],"sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor":[[19,3,1,"","all_exts"],[19,4,1,"","can_read_file"],[19,4,1,"","can_write_filename"],[19,3,1,"","default_ext"],[19,4,1,"","does_read"],[19,4,1,"","does_write"],[19,3,1,"","handles"],[19,3,1,"","name"],[19,4,1,"","read"]],"sleap.io.format.filehandle":[[21,2,1,"","FileHandle"]],"sleap.io.format.filehandle.FileHandle":[[21,4,1,"","close"],[21,3,1,"","file"],[21,3,1,"","format_id"],[21,3,1,"","is_hdf5"],[21,3,1,"","is_json"],[21,3,1,"","json"],[21,4,1,"","open"],[21,3,1,"","text"]],"sleap.io.format.genericjson":[[22,2,1,"","GenericJsonAdaptor"]],"sleap.io.format.genericjson.GenericJsonAdaptor":[[22,3,1,"","all_exts"],[22,4,1,"","can_read_file"],[22,4,1,"","can_write_filename"],[22,3,1,"","default_ext"],[22,4,1,"","does_read"],[22,4,1,"","does_write"],[22,3,1,"","handles"],[22,3,1,"","name"],[22,4,1,"","read"],[22,4,1,"","write"]],"sleap.io.format.hdf5":[[23,2,1,"","LabelsV1Adaptor"]],"sleap.io.format.hdf5.LabelsV1Adaptor":[[23,3,1,"","all_exts"],[23,4,1,"","can_read_file"],[23,4,1,"","can_write_filename"],[23,3,1,"","default_ext"],[23,4,1,"","does_read"],[23,4,1,"","does_write"],[23,3,1,"","handles"],[23,3,1,"","name"],[23,4,1,"","read"],[23,4,1,"","write"]],"sleap.io.format.labels_json":[[24,2,1,"","LabelsJsonAdaptor"]],"sleap.io.format.labels_json.LabelsJsonAdaptor":[[24,3,1,"","all_exts"],[24,4,1,"","can_read_file"],[24,4,1,"","can_write_filename"],[24,3,1,"","default_ext"],[24,4,1,"","does_read"],[24,4,1,"","does_write"],[24,4,1,"","from_json_data"],[24,3,1,"","handles"],[24,3,1,"","name"],[24,4,1,"","read"],[24,4,1,"","write"]],"sleap.io.format.leap_matlab":[[25,2,1,"","LabelsLeapMatlabAdaptor"]],"sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor":[[25,3,1,"","all_exts"],[25,4,1,"","can_read_file"],[25,4,1,"","can_write_filename"],[25,3,1,"","default_ext"],[25,4,1,"","does_read"],[25,4,1,"","does_write"],[25,3,1,"","handles"],[25,3,1,"","name"],[25,4,1,"","read"]],"sleap.io.format.main":[[26,1,1,"","read"],[26,1,1,"","write"]],"sleap.io.format.ndx_pose":[[27,2,1,"","NDXPoseAdaptor"]],"sleap.io.format.ndx_pose.NDXPoseAdaptor":[[27,3,1,"","all_exts"],[27,4,1,"","can_read_file"],[27,4,1,"","can_write_filename"],[27,3,1,"","default_ext"],[27,4,1,"","does_read"],[27,4,1,"","does_write"],[27,3,1,"","handles"],[27,3,1,"","name"],[27,4,1,"","read"],[27,4,1,"","write"]],"sleap.io.format.nix":[[28,2,1,"","NixAdaptor"]],"sleap.io.format.nix.NixAdaptor":[[28,3,1,"","all_exts"],[28,4,1,"","can_read_file"],[28,4,1,"","can_write_filename"],[28,3,1,"","default_ext"],[28,4,1,"","does_read"],[28,4,1,"","does_write"],[28,3,1,"","handles"],[28,3,1,"","name"],[28,4,1,"","read"],[28,4,1,"","write"]],"sleap.io.format.sleap_analysis":[[29,2,1,"","SleapAnalysisAdaptor"]],"sleap.io.format.sleap_analysis.SleapAnalysisAdaptor":[[29,3,1,"","all_exts"],[29,4,1,"","can_read_file"],[29,4,1,"","can_write_filename"],[29,3,1,"","default_ext"],[29,4,1,"","does_read"],[29,4,1,"","does_write"],[29,3,1,"","handles"],[29,3,1,"","name"],[29,4,1,"","read"],[29,4,1,"","write"]],"sleap.io.format.text":[[30,2,1,"","TextAdaptor"]],"sleap.io.format.text.TextAdaptor":[[30,3,1,"","all_exts"],[30,4,1,"","can_read_file"],[30,4,1,"","can_write_filename"],[30,3,1,"","default_ext"],[30,4,1,"","does_read"],[30,4,1,"","does_write"],[30,3,1,"","handles"],[30,3,1,"","name"],[30,4,1,"","read"],[30,4,1,"","write"]],"sleap.io.legacy":[[31,1,1,"","load_labels_json_old"],[31,1,1,"","load_predicted_labels_json_old"]],"sleap.io.pathutils":[[32,1,1,"","filenames_prefix_change"],[32,1,1,"","find_changed_subpath"],[32,1,1,"","list_file_missing"]],"sleap.io.video":[[33,2,1,"","DummyVideo"],[33,2,1,"","HDF5Video"],[33,2,1,"","ImgStoreVideo"],[33,2,1,"","MediaVideo"],[33,2,1,"","NumpyVideo"],[33,2,1,"","SingleImageVideo"],[33,2,1,"","Video"],[33,1,1,"","available_video_exts"],[33,1,1,"","load_video"]],"sleap.io.video.HDF5Video":[[33,3,1,"","channels"],[33,4,1,"","check"],[33,4,1,"","close"],[33,3,1,"","dtype"],[33,3,1,"","embedded_frame_inds"],[33,3,1,"","enable_source_video"],[33,3,1,"","frames"],[33,4,1,"","get_frame"],[33,3,1,"","has_embedded_images"],[33,3,1,"","height"],[33,3,1,"","last_frame_idx"],[33,4,1,"","matches"],[33,4,1,"","reset"],[33,3,1,"","source_video"],[33,3,1,"","source_video_available"],[33,3,1,"","width"]],"sleap.io.video.ImgStoreVideo":[[33,3,1,"","channels"],[33,4,1,"","close"],[33,3,1,"","dtype"],[33,3,1,"","frames"],[33,4,1,"","get_frame"],[33,3,1,"","height"],[33,3,1,"","imgstore"],[33,3,1,"","last_frame_idx"],[33,4,1,"","matches"],[33,4,1,"","open"],[33,4,1,"","reset"],[33,3,1,"","width"]],"sleap.io.video.MediaVideo":[[33,3,1,"","channels"],[33,3,1,"","dtype"],[33,3,1,"","fps"],[33,3,1,"","frames"],[33,4,1,"","get_frame"],[33,3,1,"","height"],[33,4,1,"","matches"],[33,4,1,"","reset"],[33,3,1,"","width"]],"sleap.io.video.NumpyVideo":[[33,3,1,"","channels"],[33,3,1,"","dtype"],[33,3,1,"","frames"],[33,4,1,"","get_frame"],[33,3,1,"","height"],[33,3,1,"","is_missing"],[33,4,1,"","matches"],[33,4,1,"","reset"],[33,3,1,"","width"]],"sleap.io.video.SingleImageVideo":[[33,3,1,"","channels"],[33,3,1,"","dtype"],[33,3,1,"","frames"],[33,4,1,"","get_frame"],[33,3,1,"","height"],[33,4,1,"","matches"],[33,4,1,"","reset"],[33,3,1,"","width"]],"sleap.io.video.Video":[[33,4,1,"","cattr"],[33,4,1,"","fixup_path"],[33,4,1,"","from_filename"],[33,4,1,"","from_hdf5"],[33,4,1,"","from_image_filenames"],[33,4,1,"","from_media"],[33,4,1,"","from_numpy"],[33,4,1,"","get_frame"],[33,4,1,"","get_frames"],[33,4,1,"","get_frames_safely"],[33,4,1,"","imgstore_from_filenames"],[33,3,1,"","is_missing"],[33,3,1,"","last_frame_idx"],[33,3,1,"","num_frames"],[33,3,1,"","shape"],[33,4,1,"","to_hdf5"],[33,4,1,"","to_imgstore"],[33,4,1,"","to_pipeline"]],"sleap.io.videowriter":[[34,2,1,"","VideoWriter"],[34,2,1,"","VideoWriterOpenCV"],[34,2,1,"","VideoWriterSkvideo"]],"sleap.io.videowriter.VideoWriter":[[34,4,1,"","safe_builder"]],"sleap.io.videowriter.VideoWriterSkvideo":[[34,5,1,"","crf"],[34,5,1,"","filename"],[34,5,1,"","fps"],[34,5,1,"","height"],[34,5,1,"","preset"],[34,5,1,"","width"]],"sleap.io.visuals":[[35,2,1,"","VideoMarkerThread"],[35,1,1,"","img_to_cv"],[35,1,1,"","reader"],[35,1,1,"","resize_image"],[35,1,1,"","save_labeled_video"],[35,1,1,"","writer"]],"sleap.io.visuals.VideoMarkerThread":[[35,4,1,"","run"]],"sleap.message":[[36,2,1,"","BaseMessageParticipant"],[36,2,1,"","Receiver"],[36,2,1,"","Sender"]],"sleap.message.Receiver":[[36,4,1,"","check_message"],[36,4,1,"","check_messages"],[36,4,1,"","push_back_message"]],"sleap.message.Sender":[[36,4,1,"","send_array"],[36,4,1,"","send_dict"]],"sleap.nn":[[46,0,0,"-","callbacks"],[70,0,0,"-","evals"],[71,0,0,"-","heads"],[72,0,0,"-","identity"],[73,0,0,"-","inference"],[74,0,0,"-","losses"],[75,0,0,"-","model"],[76,0,0,"-","paf_grouping"],[77,0,0,"-","peak_finding"],[78,0,0,"-","system"],[81,0,0,"-","tracking"],[82,0,0,"-","training"],[83,0,0,"-","utils"],[84,0,0,"-","viz"]],"sleap.nn.architectures":[[37,0,0,"-","common"],[38,0,0,"-","encoder_decoder"],[39,0,0,"-","hourglass"],[40,0,0,"-","hrnet"],[41,0,0,"-","leap"],[42,0,0,"-","pretrained_encoders"],[43,0,0,"-","resnet"],[44,0,0,"-","unet"],[45,0,0,"-","upsampling"]],"sleap.nn.architectures.common":[[37,2,1,"","IntermediateFeature"]],"sleap.nn.architectures.common.IntermediateFeature":[[37,3,1,"","scale"],[37,5,1,"","stride"],[37,5,1,"","tensor"]],"sleap.nn.architectures.encoder_decoder":[[38,2,1,"","DecoderBlock"],[38,2,1,"","EncoderBlock"],[38,2,1,"","EncoderDecoder"],[38,2,1,"","SimpleConvBlock"],[38,2,1,"","SimpleUpsamplingBlock"]],"sleap.nn.architectures.encoder_decoder.DecoderBlock":[[38,4,1,"","make_block"],[38,5,1,"","upsampling_stride"]],"sleap.nn.architectures.encoder_decoder.EncoderBlock":[[38,4,1,"","make_block"],[38,5,1,"","pool"],[38,5,1,"","pooling_stride"]],"sleap.nn.architectures.encoder_decoder.EncoderDecoder":[[38,3,1,"","decoder_features_stride"],[38,3,1,"","decoder_stack"],[38,3,1,"","encoder_features_stride"],[38,3,1,"","encoder_stack"],[38,4,1,"","make_backbone"],[38,4,1,"","make_decoder"],[38,4,1,"","make_encoder"],[38,4,1,"","make_stem"],[38,3,1,"","maximum_stride"],[38,3,1,"","output_stride"],[38,5,1,"","stacks"],[38,3,1,"","stem_features_stride"],[38,3,1,"","stem_stack"]],"sleap.nn.architectures.encoder_decoder.SimpleConvBlock":[[38,5,1,"","activation"],[38,5,1,"","batch_norm"],[38,5,1,"","batch_norm_before_activation"],[38,5,1,"","block_prefix"],[38,5,1,"","filters"],[38,5,1,"","kernel_size"],[38,4,1,"","make_block"],[38,5,1,"","num_convs"],[38,5,1,"","pool"],[38,5,1,"","pool_before_convs"],[38,5,1,"","pooling_stride"],[38,5,1,"","use_bias"]],"sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock":[[38,5,1,"","interp_method"],[38,4,1,"","make_block"],[38,5,1,"","refine_convs"],[38,5,1,"","refine_convs_activation"],[38,5,1,"","refine_convs_batch_norm"],[38,5,1,"","refine_convs_batch_norm_before_activation"],[38,5,1,"","refine_convs_filters"],[38,5,1,"","refine_convs_first_filters"],[38,5,1,"","refine_convs_kernel_size"],[38,5,1,"","refine_convs_use_bias"],[38,5,1,"","skip_add"],[38,5,1,"","skip_connection"],[38,5,1,"","transposed_conv"],[38,5,1,"","transposed_conv_activation"],[38,5,1,"","transposed_conv_batch_norm"],[38,5,1,"","transposed_conv_batch_norm_before_activation"],[38,5,1,"","transposed_conv_filters"],[38,5,1,"","transposed_conv_kernel_size"],[38,5,1,"","transposed_conv_use_bias"],[38,5,1,"","upsampling_stride"]],"sleap.nn.architectures.hourglass":[[39,2,1,"","DownsamplingBlock"],[39,2,1,"","Hourglass"],[39,2,1,"","StemBlock"],[39,2,1,"","UpsamplingBlock"],[39,1,1,"","conv"]],"sleap.nn.architectures.hourglass.DownsamplingBlock":[[39,5,1,"","filters"],[39,4,1,"","make_block"]],"sleap.nn.architectures.hourglass.Hourglass":[[39,3,1,"","decoder_stack"],[39,5,1,"","down_blocks"],[39,3,1,"","encoder_stack"],[39,5,1,"","filter_increase"],[39,5,1,"","filters"],[39,4,1,"","from_config"],[39,5,1,"","interp_method"],[39,5,1,"","stacks"],[39,5,1,"","stem_filters"],[39,3,1,"","stem_stack"],[39,5,1,"","stem_stride"],[39,5,1,"","up_blocks"]],"sleap.nn.architectures.hourglass.StemBlock":[[39,5,1,"","filters"],[39,5,1,"","first_conv_stride"],[39,4,1,"","make_block"],[39,5,1,"","output_filters"],[39,5,1,"","pool"],[39,5,1,"","pooling_stride"]],"sleap.nn.architectures.hourglass.UpsamplingBlock":[[39,5,1,"","filters"],[39,5,1,"","interp_method"],[39,4,1,"","make_block"]],"sleap.nn.architectures.hrnet":[[40,2,1,"","HigherHRNet"],[40,1,1,"","adjust_prefix"],[40,1,1,"","bottleneck_block"],[40,1,1,"","simple_block"]],"sleap.nn.architectures.hrnet.HigherHRNet":[[40,5,1,"","C"],[40,5,1,"","bilinear_upsampling"],[40,5,1,"","bottleneck"],[40,5,1,"","deconv_filters"],[40,3,1,"","down_blocks"],[40,5,1,"","initial_downsampling_steps"],[40,5,1,"","n_deconv_modules"],[40,4,1,"","output"],[40,3,1,"","output_scale"]],"sleap.nn.architectures.leap":[[41,2,1,"","LeapCNN"]],"sleap.nn.architectures.leap.LeapCNN":[[41,3,1,"","decoder_stack"],[41,5,1,"","down_blocks"],[41,5,1,"","down_convs_per_block"],[41,3,1,"","encoder_stack"],[41,5,1,"","filters"],[41,5,1,"","filters_rate"],[41,4,1,"","from_config"],[41,5,1,"","up_blocks"],[41,5,1,"","up_convs_per_block"],[41,5,1,"","up_interpolate"]],"sleap.nn.architectures.pretrained_encoders":[[42,2,1,"","UnetPretrainedEncoder"]],"sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder":[[42,5,1,"","decoder_filters"],[42,3,1,"","down_blocks"],[42,5,1,"","encoder"],[42,4,1,"","from_config"],[42,4,1,"","make_backbone"],[42,3,1,"","maximum_stride"],[42,3,1,"","output_stride"],[42,5,1,"","pretrained"],[42,3,1,"","up_blocks"]],"sleap.nn.architectures.resnet":[[43,2,1,"","ResNet101"],[43,2,1,"","ResNet152"],[43,2,1,"","ResNet50"],[43,2,1,"","ResNetv1"],[43,1,1,"","block_v1"],[43,1,1,"","imagenet_preproc_v1"],[43,1,1,"","make_backbone_fn"],[43,1,1,"","make_resnet_model"],[43,1,1,"","stack_v1"],[43,1,1,"","tile_channels"]],"sleap.nn.architectures.resnet.ResNet101":[[43,5,1,"","features_output_stride"],[43,5,1,"","frozen"],[43,5,1,"","pretrained"],[43,5,1,"","skip_connections"],[43,5,1,"","upsampling_stack"]],"sleap.nn.architectures.resnet.ResNet152":[[43,5,1,"","features_output_stride"],[43,5,1,"","frozen"],[43,5,1,"","pretrained"],[43,5,1,"","skip_connections"],[43,5,1,"","upsampling_stack"]],"sleap.nn.architectures.resnet.ResNet50":[[43,5,1,"","features_output_stride"],[43,5,1,"","frozen"],[43,5,1,"","pretrained"],[43,5,1,"","skip_connections"],[43,5,1,"","upsampling_stack"]],"sleap.nn.architectures.resnet.ResNetv1":[[43,3,1,"","down_blocks"],[43,5,1,"","features_output_stride"],[43,4,1,"","from_config"],[43,5,1,"","frozen"],[43,4,1,"","make_backbone"],[43,3,1,"","maximum_stride"],[43,5,1,"","model_name"],[43,3,1,"","output_scale"],[43,3,1,"","output_stride"],[43,5,1,"","pretrained"],[43,5,1,"","skip_connections"],[43,5,1,"","stack_configs"],[43,5,1,"","upsampling_stack"]],"sleap.nn.architectures.unet":[[44,2,1,"","PoolingBlock"],[44,2,1,"","UNet"]],"sleap.nn.architectures.unet.PoolingBlock":[[44,4,1,"","make_block"],[44,5,1,"","pool"],[44,5,1,"","pooling_stride"]],"sleap.nn.architectures.unet.UNet":[[44,5,1,"","block_contraction"],[44,5,1,"","convs_per_block"],[44,3,1,"","decoder_stack"],[44,5,1,"","down_blocks"],[44,3,1,"","encoder_stack"],[44,5,1,"","filters"],[44,5,1,"","filters_rate"],[44,4,1,"","from_config"],[44,5,1,"","kernel_size"],[44,5,1,"","middle_block"],[44,5,1,"","stem_blocks"],[44,5,1,"","stem_kernel_size"],[44,3,1,"","stem_stack"],[44,5,1,"","up_blocks"],[44,5,1,"","up_interpolate"]],"sleap.nn.architectures.upsampling":[[45,2,1,"","UpsamplingStack"]],"sleap.nn.architectures.upsampling.UpsamplingStack":[[45,4,1,"","from_config"],[45,5,1,"","make_skip_connection"],[45,4,1,"","make_stack"],[45,5,1,"","output_stride"],[45,5,1,"","refine_convs"],[45,5,1,"","refine_convs_batchnorm"],[45,5,1,"","refine_convs_filters"],[45,5,1,"","refine_convs_filters_rate"],[45,5,1,"","skip_add"],[45,5,1,"","transposed_conv"],[45,5,1,"","transposed_conv_batchnorm"],[45,5,1,"","transposed_conv_filters"],[45,5,1,"","transposed_conv_filters_rate"],[45,5,1,"","transposed_conv_kernel_size"],[45,5,1,"","upsampling_stride"]],"sleap.nn.callbacks":[[46,2,1,"","MatplotlibSaver"],[46,2,1,"","ModelCheckpointOnEvent"],[46,2,1,"","ProgressReporterZMQ"],[46,2,1,"","TensorBoardMatplotlibWriter"],[46,2,1,"","TrainingControllerZMQ"]],"sleap.nn.callbacks.MatplotlibSaver":[[46,4,1,"","on_epoch_end"],[46,5,1,"","plot_fn"],[46,5,1,"","prefix"],[46,5,1,"","save_folder"]],"sleap.nn.callbacks.ModelCheckpointOnEvent":[[46,5,1,"","event"],[46,5,1,"","filepath"],[46,4,1,"","on_epoch_end"],[46,4,1,"","on_train_begin"],[46,4,1,"","on_train_end"]],"sleap.nn.callbacks.ProgressReporterZMQ":[[46,4,1,"","on_batch_begin"],[46,4,1,"","on_batch_end"],[46,4,1,"","on_epoch_begin"],[46,4,1,"","on_epoch_end"],[46,4,1,"","on_train_begin"],[46,4,1,"","on_train_end"]],"sleap.nn.callbacks.TensorBoardMatplotlibWriter":[[46,5,1,"","logdir"],[46,4,1,"","on_epoch_end"],[46,5,1,"","plot_fn"],[46,5,1,"","tag"]],"sleap.nn.callbacks.TrainingControllerZMQ":[[46,4,1,"","on_batch_end"],[46,4,1,"","set_lr"]],"sleap.nn.config":[[47,0,0,"-","data"],[48,0,0,"-","model"],[49,0,0,"-","optimization"],[50,0,0,"-","outputs"],[51,0,0,"-","training_job"],[52,0,0,"-","utils"]],"sleap.nn.config.data":[[47,2,1,"","DataConfig"],[47,2,1,"","InstanceCroppingConfig"],[47,2,1,"","LabelsConfig"],[47,2,1,"","PreprocessingConfig"]],"sleap.nn.config.data.InstanceCroppingConfig":[[47,5,1,"","center_on_part"],[47,5,1,"","crop_size"],[47,5,1,"","crop_size_detection_padding"]],"sleap.nn.config.data.LabelsConfig":[[47,5,1,"","search_path_hints"],[47,5,1,"","skeletons"],[47,5,1,"","split_by_inds"],[47,5,1,"","test_inds"],[47,5,1,"","test_labels"],[47,5,1,"","training_inds"],[47,5,1,"","training_labels"],[47,5,1,"","validation_fraction"],[47,5,1,"","validation_inds"],[47,5,1,"","validation_labels"]],"sleap.nn.config.data.PreprocessingConfig":[[47,5,1,"","ensure_grayscale"],[47,5,1,"","ensure_rgb"],[47,5,1,"","imagenet_mode"],[47,5,1,"","input_scale"],[47,5,1,"","pad_to_stride"],[47,5,1,"","resize_and_pad_to_target"],[47,5,1,"","target_height"],[47,5,1,"","target_width"]],"sleap.nn.config.model":[[48,2,1,"","BackboneConfig"],[48,2,1,"","CenteredInstanceConfmapsHeadConfig"],[48,2,1,"","CentroidsHeadConfig"],[48,2,1,"","ClassMapsHeadConfig"],[48,2,1,"","ClassVectorsHeadConfig"],[48,2,1,"","HeadsConfig"],[48,2,1,"","HourglassConfig"],[48,2,1,"","LEAPConfig"],[48,2,1,"","ModelConfig"],[48,2,1,"","MultiClassBottomUpConfig"],[48,2,1,"","MultiClassTopDownConfig"],[48,2,1,"","MultiInstanceConfig"],[48,2,1,"","MultiInstanceConfmapsHeadConfig"],[48,2,1,"","PartAffinityFieldsHeadConfig"],[48,2,1,"","PretrainedEncoderConfig"],[48,2,1,"","ResNetConfig"],[48,2,1,"","SingleInstanceConfmapsHeadConfig"],[48,2,1,"","UNetConfig"],[48,2,1,"","UpsamplingConfig"]],"sleap.nn.config.model.BackboneConfig":[[48,5,1,"","hourglass"],[48,5,1,"","leap"],[48,5,1,"","resnet"],[48,5,1,"","unet"]],"sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig":[[48,5,1,"","anchor_part"],[48,5,1,"","loss_weight"],[48,5,1,"","offset_refinement"],[48,5,1,"","output_stride"],[48,5,1,"","part_names"],[48,5,1,"","sigma"]],"sleap.nn.config.model.CentroidsHeadConfig":[[48,5,1,"","anchor_part"],[48,5,1,"","loss_weight"],[48,5,1,"","offset_refinement"],[48,5,1,"","output_stride"],[48,5,1,"","sigma"]],"sleap.nn.config.model.ClassMapsHeadConfig":[[48,5,1,"","classes"],[48,5,1,"","loss_weight"],[48,5,1,"","output_stride"],[48,5,1,"","sigma"]],"sleap.nn.config.model.ClassVectorsHeadConfig":[[48,5,1,"","classes"],[48,5,1,"","loss_weight"],[48,5,1,"","num_fc_layers"],[48,5,1,"","num_fc_units"],[48,5,1,"","output_stride"]],"sleap.nn.config.model.HeadsConfig":[[48,5,1,"","centered_instance"],[48,5,1,"","centroid"],[48,5,1,"","multi_class_bottomup"],[48,5,1,"","multi_class_topdown"],[48,5,1,"","multi_instance"],[48,5,1,"","single_instance"]],"sleap.nn.config.model.HourglassConfig":[[48,5,1,"","filters"],[48,5,1,"","filters_increase"],[48,5,1,"","max_stride"],[48,5,1,"","output_stride"],[48,5,1,"","stacks"],[48,5,1,"","stem_stride"]],"sleap.nn.config.model.LEAPConfig":[[48,5,1,"","filters"],[48,5,1,"","filters_rate"],[48,5,1,"","max_stride"],[48,5,1,"","output_stride"],[48,5,1,"","stacks"],[48,5,1,"","up_interpolate"]],"sleap.nn.config.model.ModelConfig":[[48,5,1,"","backbone"],[48,5,1,"","base_checkpoint"],[48,5,1,"","heads"]],"sleap.nn.config.model.MultiClassBottomUpConfig":[[48,5,1,"","class_maps"],[48,5,1,"","confmaps"]],"sleap.nn.config.model.MultiClassTopDownConfig":[[48,5,1,"","class_vectors"],[48,5,1,"","confmaps"]],"sleap.nn.config.model.MultiInstanceConfig":[[48,5,1,"","confmaps"],[48,5,1,"","pafs"]],"sleap.nn.config.model.MultiInstanceConfmapsHeadConfig":[[48,5,1,"","loss_weight"],[48,5,1,"","offset_refinement"],[48,5,1,"","output_stride"],[48,5,1,"","part_names"],[48,5,1,"","sigma"]],"sleap.nn.config.model.PartAffinityFieldsHeadConfig":[[48,5,1,"","edges"],[48,5,1,"","loss_weight"],[48,5,1,"","output_stride"],[48,5,1,"","sigma"]],"sleap.nn.config.model.PretrainedEncoderConfig":[[48,5,1,"","decoder_batchnorm"],[48,5,1,"","decoder_filters"],[48,5,1,"","decoder_filters_rate"],[48,5,1,"","encoder"],[48,5,1,"","output_stride"],[48,5,1,"","pretrained"]],"sleap.nn.config.model.ResNetConfig":[[48,5,1,"","max_stride"],[48,5,1,"","output_stride"],[48,5,1,"","upsampling"],[48,5,1,"","version"],[48,5,1,"","weights"]],"sleap.nn.config.model.SingleInstanceConfmapsHeadConfig":[[48,5,1,"","loss_weight"],[48,5,1,"","offset_refinement"],[48,5,1,"","output_stride"],[48,5,1,"","part_names"],[48,5,1,"","sigma"]],"sleap.nn.config.model.UNetConfig":[[48,5,1,"","filters"],[48,5,1,"","filters_rate"],[48,5,1,"","max_stride"],[48,5,1,"","middle_block"],[48,5,1,"","output_stride"],[48,5,1,"","stacks"],[48,5,1,"","stem_stride"],[48,5,1,"","up_interpolate"]],"sleap.nn.config.model.UpsamplingConfig":[[48,5,1,"","block_stride"],[48,5,1,"","conv_batchnorm"],[48,5,1,"","filters"],[48,5,1,"","filters_rate"],[48,5,1,"","method"],[48,5,1,"","refine_convs"],[48,5,1,"","skip_connections"],[48,5,1,"","transposed_conv_kernel_size"]],"sleap.nn.config.optimization":[[49,2,1,"","AugmentationConfig"],[49,2,1,"","EarlyStoppingConfig"],[49,2,1,"","HardKeypointMiningConfig"],[49,2,1,"","LearningRateScheduleConfig"],[49,2,1,"","OptimizationConfig"]],"sleap.nn.config.optimization.AugmentationConfig":[[49,5,1,"","brightness"],[49,5,1,"","brightness_max_val"],[49,5,1,"","brightness_min_val"],[49,5,1,"","contrast"],[49,5,1,"","contrast_max_gamma"],[49,5,1,"","contrast_min_gamma"],[49,5,1,"","flip_horizontal"],[49,5,1,"","gaussian_noise"],[49,5,1,"","gaussian_noise_mean"],[49,5,1,"","gaussian_noise_stddev"],[49,5,1,"","random_crop"],[49,5,1,"","random_crop_height"],[49,5,1,"","random_crop_width"],[49,5,1,"","random_flip"],[49,5,1,"","rotate"],[49,5,1,"","rotation_max_angle"],[49,5,1,"","rotation_min_angle"],[49,5,1,"","scale"],[49,5,1,"","scale_max"],[49,5,1,"","scale_min"],[49,5,1,"","translate"],[49,5,1,"","translate_max"],[49,5,1,"","translate_min"],[49,5,1,"","uniform_noise"],[49,5,1,"","uniform_noise_max_val"],[49,5,1,"","uniform_noise_min_val"]],"sleap.nn.config.optimization.EarlyStoppingConfig":[[49,5,1,"","plateau_min_delta"],[49,5,1,"","plateau_patience"],[49,5,1,"","stop_training_on_plateau"]],"sleap.nn.config.optimization.HardKeypointMiningConfig":[[49,5,1,"","hard_to_easy_ratio"],[49,5,1,"","loss_scale"],[49,5,1,"","max_hard_keypoints"],[49,5,1,"","min_hard_keypoints"],[49,5,1,"","online_mining"]],"sleap.nn.config.optimization.LearningRateScheduleConfig":[[49,5,1,"","min_learning_rate"],[49,5,1,"","plateau_cooldown"],[49,5,1,"","plateau_min_delta"],[49,5,1,"","plateau_patience"],[49,5,1,"","reduce_on_plateau"],[49,5,1,"","reduction_factor"]],"sleap.nn.config.optimization.OptimizationConfig":[[49,5,1,"","augmentation_config"],[49,5,1,"","batch_size"],[49,5,1,"","batches_per_epoch"],[49,5,1,"","early_stopping"],[49,5,1,"","epochs"],[49,5,1,"","hard_keypoint_mining"],[49,5,1,"","initial_learning_rate"],[49,5,1,"","learning_rate_schedule"],[49,5,1,"","min_batches_per_epoch"],[49,5,1,"","min_val_batches_per_epoch"],[49,5,1,"","online_shuffling"],[49,5,1,"","optimizer"],[49,5,1,"","prefetch"],[49,5,1,"","preload_data"],[49,5,1,"","shuffle_buffer_size"],[49,5,1,"","val_batches_per_epoch"]],"sleap.nn.config.outputs":[[50,2,1,"","CheckpointingConfig"],[50,2,1,"","OutputsConfig"],[50,2,1,"","TensorBoardConfig"],[50,2,1,"","ZMQConfig"]],"sleap.nn.config.outputs.CheckpointingConfig":[[50,5,1,"","best_model"],[50,5,1,"","every_epoch"],[50,5,1,"","final_model"],[50,5,1,"","initial_model"],[50,5,1,"","latest_model"]],"sleap.nn.config.outputs.OutputsConfig":[[50,5,1,"","checkpointing"],[50,5,1,"","delete_viz_images"],[50,5,1,"","log_to_csv"],[50,5,1,"","run_name"],[50,5,1,"","run_name_prefix"],[50,5,1,"","run_name_suffix"],[50,3,1,"","run_path"],[50,5,1,"","runs_folder"],[50,5,1,"","save_outputs"],[50,5,1,"","save_visualizations"],[50,5,1,"","tags"],[50,5,1,"","tensorboard"],[50,5,1,"","zip_outputs"],[50,5,1,"","zmq"]],"sleap.nn.config.outputs.TensorBoardConfig":[[50,5,1,"","architecture_graph"],[50,5,1,"","loss_frequency"],[50,5,1,"","profile_graph"],[50,5,1,"","visualizations"],[50,5,1,"","write_logs"]],"sleap.nn.config.outputs.ZMQConfig":[[50,5,1,"","controller_address"],[50,5,1,"","controller_polling_timeout"],[50,5,1,"","publish_address"],[50,5,1,"","publish_updates"],[50,5,1,"","subscribe_to_controller"]],"sleap.nn.config.training_job":[[51,2,1,"","TrainingJobConfig"],[51,1,1,"","load_config"]],"sleap.nn.config.training_job.TrainingJobConfig":[[51,5,1,"","data"],[51,5,1,"","description"],[51,5,1,"","filename"],[51,4,1,"","from_json"],[51,4,1,"","from_json_dicts"],[51,4,1,"","load_json"],[51,5,1,"","model"],[51,5,1,"","name"],[51,5,1,"","optimization"],[51,5,1,"","outputs"],[51,4,1,"","save_json"],[51,5,1,"","sleap_version"],[51,4,1,"","to_json"]],"sleap.nn.config.utils":[[52,1,1,"","oneof"]],"sleap.nn.data":[[53,0,0,"-","augmentation"],[54,0,0,"-","confidence_maps"],[55,0,0,"-","dataset_ops"],[56,0,0,"-","edge_maps"],[57,0,0,"-","general"],[58,0,0,"-","grouping"],[59,0,0,"-","identity"],[60,0,0,"-","inference"],[61,0,0,"-","instance_centroids"],[62,0,0,"-","instance_cropping"],[63,0,0,"-","normalization"],[64,0,0,"-","offset_regression"],[65,0,0,"-","pipelines"],[66,0,0,"-","providers"],[67,0,0,"-","resizing"],[68,0,0,"-","training"],[69,0,0,"-","utils"]],"sleap.nn.data.augmentation":[[53,2,1,"","ImgaugAugmenter"],[53,2,1,"","RandomCropper"],[53,2,1,"","RandomFlipper"],[53,1,1,"","flip_instances_lr"],[53,1,1,"","flip_instances_ud"]],"sleap.nn.data.augmentation.ImgaugAugmenter":[[53,5,1,"","augmenter"],[53,4,1,"","from_config"],[53,5,1,"","image_key"],[53,3,1,"","input_keys"],[53,5,1,"","instances_key"],[53,3,1,"","output_keys"],[53,4,1,"","transform_dataset"]],"sleap.nn.data.augmentation.RandomCropper":[[53,5,1,"","crop_height"],[53,5,1,"","crop_width"],[53,4,1,"","transform_dataset"]],"sleap.nn.data.augmentation.RandomFlipper":[[53,4,1,"","from_skeleton"],[53,5,1,"","horizontal"],[53,5,1,"","probability"],[53,5,1,"","symmetric_inds"],[53,4,1,"","transform_dataset"]],"sleap.nn.data.confidence_maps":[[54,2,1,"","InstanceConfidenceMapGenerator"],[54,2,1,"","MultiConfidenceMapGenerator"],[54,2,1,"","SingleInstanceConfidenceMapGenerator"],[54,1,1,"","make_confmaps"],[54,1,1,"","make_multi_confmaps"],[54,1,1,"","make_multi_confmaps_with_offsets"]],"sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator":[[54,5,1,"","all_instances"],[54,5,1,"","flatten_offsets"],[54,3,1,"","input_keys"],[54,5,1,"","offsets_threshold"],[54,3,1,"","output_keys"],[54,5,1,"","output_stride"],[54,5,1,"","sigma"],[54,4,1,"","transform_dataset"],[54,5,1,"","with_offsets"]],"sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator":[[54,5,1,"","centroids"],[54,5,1,"","flatten_offsets"],[54,3,1,"","input_keys"],[54,5,1,"","offsets_threshold"],[54,3,1,"","output_keys"],[54,5,1,"","output_stride"],[54,5,1,"","sigma"],[54,4,1,"","transform_dataset"],[54,5,1,"","with_offsets"]],"sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator":[[54,5,1,"","flatten_offsets"],[54,3,1,"","input_keys"],[54,5,1,"","offsets_threshold"],[54,3,1,"","output_keys"],[54,5,1,"","output_stride"],[54,5,1,"","sigma"],[54,4,1,"","transform_dataset"],[54,5,1,"","with_offsets"]],"sleap.nn.data.dataset_ops":[[55,2,1,"","Batcher"],[55,2,1,"","LambdaFilter"],[55,2,1,"","Prefetcher"],[55,2,1,"","Preloader"],[55,2,1,"","Repeater"],[55,2,1,"","Shuffler"],[55,2,1,"","Unbatcher"]],"sleap.nn.data.dataset_ops.Batcher":[[55,5,1,"","batch_size"],[55,5,1,"","drop_remainder"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,4,1,"","transform_dataset"],[55,5,1,"","unrag"]],"sleap.nn.data.dataset_ops.LambdaFilter":[[55,5,1,"","filter_fn"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.dataset_ops.Prefetcher":[[55,5,1,"","buffer_size"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,5,1,"","prefetch"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.dataset_ops.Preloader":[[55,5,1,"","examples"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.dataset_ops.Repeater":[[55,5,1,"","epochs"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,5,1,"","repeat"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.dataset_ops.Shuffler":[[55,5,1,"","buffer_size"],[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,5,1,"","reshuffle_each_iteration"],[55,5,1,"","shuffle"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.dataset_ops.Unbatcher":[[55,3,1,"","input_keys"],[55,3,1,"","output_keys"],[55,4,1,"","transform_dataset"]],"sleap.nn.data.edge_maps":[[56,2,1,"","PartAffinityFieldsGenerator"],[56,1,1,"","distance_to_edge"],[56,1,1,"","get_edge_points"],[56,1,1,"","make_edge_maps"],[56,1,1,"","make_multi_pafs"],[56,1,1,"","make_pafs"]],"sleap.nn.data.edge_maps.PartAffinityFieldsGenerator":[[56,5,1,"","flatten_channels"],[56,3,1,"","input_keys"],[56,3,1,"","output_keys"],[56,5,1,"","output_stride"],[56,5,1,"","sigma"],[56,5,1,"","skeletons"],[56,4,1,"","transform_dataset"]],"sleap.nn.data.general":[[57,2,1,"","KeyDeviceMover"],[57,2,1,"","KeyFilter"],[57,2,1,"","KeyRenamer"],[57,2,1,"","LambdaMap"]],"sleap.nn.data.general.KeyDeviceMover":[[57,3,1,"","input_keys"],[57,3,1,"","output_keys"],[57,4,1,"","transform_dataset"]],"sleap.nn.data.general.KeyFilter":[[57,3,1,"","input_keys"],[57,3,1,"","output_keys"],[57,4,1,"","transform_dataset"]],"sleap.nn.data.general.KeyRenamer":[[57,3,1,"","input_keys"],[57,3,1,"","output_keys"],[57,4,1,"","transform_dataset"]],"sleap.nn.data.general.LambdaMap":[[57,5,1,"","func"],[57,5,1,"","input_key_names"],[57,3,1,"","input_keys"],[57,5,1,"","output_key_names"],[57,3,1,"","output_keys"],[57,4,1,"","transform_dataset"]],"sleap.nn.data.grouping":[[58,1,1,"","group_examples"],[58,1,1,"","group_examples_iter"]],"sleap.nn.data.identity":[[59,2,1,"","ClassMapGenerator"],[59,2,1,"","ClassVectorGenerator"],[59,1,1,"","make_class_maps"],[59,1,1,"","make_class_vectors"]],"sleap.nn.data.identity.ClassMapGenerator":[[59,5,1,"","centroids"],[59,5,1,"","class_map_threshold"],[59,3,1,"","input_keys"],[59,3,1,"","output_keys"],[59,5,1,"","output_stride"],[59,5,1,"","sigma"],[59,4,1,"","transform_dataset"]],"sleap.nn.data.identity.ClassVectorGenerator":[[59,3,1,"","input_keys"],[59,3,1,"","output_keys"],[59,4,1,"","transform_dataset"]],"sleap.nn.data.inference":[[60,2,1,"","GlobalPeakFinder"],[60,2,1,"","KerasModelPredictor"],[60,2,1,"","LocalPeakFinder"],[60,2,1,"","MockGlobalPeakFinder"],[60,2,1,"","PredictedCenterInstanceNormalizer"]],"sleap.nn.data.inference.PredictedCenterInstanceNormalizer":[[60,3,1,"","input_keys"],[60,3,1,"","output_keys"],[60,4,1,"","transform_dataset"]],"sleap.nn.data.instance_centroids":[[61,2,1,"","InstanceCentroidFinder"],[61,1,1,"","find_points_bbox_midpoint"],[61,1,1,"","get_instance_anchors"]],"sleap.nn.data.instance_centroids.InstanceCentroidFinder":[[61,5,1,"","anchor_part_names"],[61,5,1,"","center_on_anchor_part"],[61,4,1,"","from_config"],[61,3,1,"","input_keys"],[61,5,1,"","instances_key"],[61,3,1,"","output_keys"],[61,5,1,"","skeletons"],[61,4,1,"","transform_dataset"]],"sleap.nn.data.instance_cropping":[[62,2,1,"","InstanceCropper"],[62,1,1,"","crop_bboxes"],[62,1,1,"","find_instance_crop_size"],[62,1,1,"","make_centered_bboxes"],[62,1,1,"","normalize_bboxes"],[62,1,1,"","unnormalize_bboxes"]],"sleap.nn.data.instance_cropping.InstanceCropper":[[62,5,1,"","centroids_key"],[62,5,1,"","crop_height"],[62,5,1,"","crop_width"],[62,4,1,"","from_config"],[62,5,1,"","image_key"],[62,3,1,"","input_keys"],[62,5,1,"","instances_key"],[62,5,1,"","keep_full_image"],[62,5,1,"","mock_centroid_confidence"],[62,3,1,"","output_keys"],[62,4,1,"","transform_dataset"],[62,5,1,"","unbatch"]],"sleap.nn.data.normalization":[[63,2,1,"","Normalizer"],[63,1,1,"","convert_rgb_to_bgr"],[63,1,1,"","ensure_float"],[63,1,1,"","ensure_grayscale"],[63,1,1,"","ensure_int"],[63,1,1,"","ensure_min_image_rank"],[63,1,1,"","ensure_rgb"],[63,1,1,"","scale_image_range"],[63,1,1,"","scale_to_imagenet_caffe_mode"],[63,1,1,"","scale_to_imagenet_tf_mode"],[63,1,1,"","scale_to_imagenet_torch_mode"]],"sleap.nn.data.normalization.Normalizer":[[63,5,1,"","ensure_float"],[63,5,1,"","ensure_grayscale"],[63,5,1,"","ensure_rgb"],[63,4,1,"","from_config"],[63,5,1,"","image_key"],[63,5,1,"","imagenet_mode"],[63,3,1,"","input_keys"],[63,3,1,"","output_keys"],[63,4,1,"","transform_dataset"]],"sleap.nn.data.offset_regression":[[64,1,1,"","make_offsets"],[64,1,1,"","mask_offsets"]],"sleap.nn.data.pipelines":[[65,2,1,"","BottomUpMultiClassPipeline"],[65,2,1,"","BottomUpPipeline"],[65,2,1,"","CentroidConfmapsPipeline"],[65,2,1,"","Pipeline"],[65,2,1,"","SingleInstanceConfmapsPipeline"],[65,2,1,"","TopDownMultiClassPipeline"],[65,2,1,"","TopdownConfmapsPipeline"]],"sleap.nn.data.pipelines.BottomUpMultiClassPipeline":[[65,5,1,"","class_maps_head"],[65,5,1,"","confmaps_head"],[65,5,1,"","data_config"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"]],"sleap.nn.data.pipelines.BottomUpPipeline":[[65,5,1,"","confmaps_head"],[65,5,1,"","data_config"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"],[65,5,1,"","pafs_head"]],"sleap.nn.data.pipelines.CentroidConfmapsPipeline":[[65,5,1,"","centroid_confmap_head"],[65,5,1,"","data_config"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"]],"sleap.nn.data.pipelines.Pipeline":[[65,4,1,"","append"],[65,4,1,"","describe"],[65,4,1,"","from_blocks"],[65,4,1,"","from_pipelines"],[65,4,1,"","make_dataset"],[65,3,1,"","output_keys"],[65,4,1,"","peek"],[65,5,1,"","providers"],[65,4,1,"","run"],[65,5,1,"","transformers"],[65,4,1,"","validate_pipeline"]],"sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline":[[65,5,1,"","data_config"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"],[65,5,1,"","single_instance_confmap_head"]],"sleap.nn.data.pipelines.TopDownMultiClassPipeline":[[65,5,1,"","class_vectors_head"],[65,5,1,"","confmaps_head"],[65,5,1,"","data_config"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"]],"sleap.nn.data.pipelines.TopdownConfmapsPipeline":[[65,5,1,"","data_config"],[65,5,1,"","instance_confmap_head"],[65,4,1,"","make_base_pipeline"],[65,4,1,"","make_training_pipeline"],[65,4,1,"","make_viz_pipeline"],[65,5,1,"","offsets_head"],[65,5,1,"","optimization_config"]],"sleap.nn.data.providers":[[66,2,1,"","LabelsReader"],[66,2,1,"","VideoReader"]],"sleap.nn.data.providers.LabelsReader":[[66,5,1,"","example_indices"],[66,4,1,"","from_filename"],[66,4,1,"","from_unlabeled_suggestions"],[66,4,1,"","from_user_instances"],[66,4,1,"","from_user_labeled_frames"],[66,3,1,"","is_from_multi_size_videos"],[66,5,1,"","labels"],[66,4,1,"","make_dataset"],[66,3,1,"","max_height_and_width"],[66,3,1,"","output_keys"],[66,3,1,"","tracks"],[66,5,1,"","user_instances_only"],[66,3,1,"","videos"],[66,5,1,"","with_track_only"]],"sleap.nn.data.providers.VideoReader":[[66,5,1,"","example_indices"],[66,4,1,"","from_filepath"],[66,4,1,"","make_dataset"],[66,3,1,"","max_height_and_width"],[66,3,1,"","output_keys"],[66,5,1,"","video"],[66,5,1,"","video_ind"],[66,3,1,"","videos"]],"sleap.nn.data.resizing":[[67,2,1,"","PointsRescaler"],[67,2,1,"","Resizer"],[67,2,1,"","SizeMatcher"],[67,1,1,"","find_padding_for_stride"],[67,1,1,"","pad_to_stride"],[67,1,1,"","resize_image"]],"sleap.nn.data.resizing.PointsRescaler":[[67,3,1,"","input_keys"],[67,3,1,"","output_keys"],[67,4,1,"","transform_dataset"]],"sleap.nn.data.resizing.Resizer":[[67,4,1,"","from_config"],[67,5,1,"","full_image_key"],[67,5,1,"","image_key"],[67,3,1,"","input_keys"],[67,5,1,"","keep_full_image"],[67,3,1,"","output_keys"],[67,5,1,"","pad_to_stride"],[67,5,1,"","points_key"],[67,5,1,"","scale"],[67,5,1,"","scale_key"],[67,4,1,"","transform_dataset"]],"sleap.nn.data.resizing.SizeMatcher":[[67,5,1,"","center_pad"],[67,4,1,"","from_config"],[67,5,1,"","full_image_key"],[67,5,1,"","image_key"],[67,3,1,"","input_keys"],[67,5,1,"","keep_full_image"],[67,5,1,"","max_image_height"],[67,5,1,"","max_image_width"],[67,3,1,"","output_keys"],[67,5,1,"","points_key"],[67,5,1,"","scale_key"],[67,4,1,"","transform_dataset"]],"sleap.nn.data.training":[[68,2,1,"","KeyMapper"],[68,1,1,"","split_labels"],[68,1,1,"","split_labels_reader"],[68,1,1,"","split_labels_train_val"]],"sleap.nn.data.training.KeyMapper":[[68,3,1,"","input_keys"],[68,5,1,"","key_maps"],[68,3,1,"","output_keys"],[68,4,1,"","transform_dataset"]],"sleap.nn.data.utils":[[69,1,1,"","describe_tensors"],[69,1,1,"","ensure_list"],[69,1,1,"","expand_to_rank"],[69,1,1,"","gaussian_pdf"],[69,1,1,"","make_grid_vectors"],[69,1,1,"","unrag_example"],[69,1,1,"","unrag_tensor"]],"sleap.nn.evals":[[70,1,1,"","compute_dist_metrics"],[70,1,1,"","compute_dists"],[70,1,1,"","compute_generalized_voc_metrics"],[70,1,1,"","compute_instance_area"],[70,1,1,"","compute_oks"],[70,1,1,"","compute_pck_metrics"],[70,1,1,"","compute_visibility_conf"],[70,1,1,"","evaluate"],[70,1,1,"","evaluate_model"],[70,1,1,"","find_frame_pairs"],[70,1,1,"","load_metrics"],[70,1,1,"","match_frame_pairs"],[70,1,1,"","match_instances"],[70,1,1,"","replace_path"]],"sleap.nn.heads":[[71,2,1,"","CenteredInstanceConfmapsHead"],[71,2,1,"","CentroidConfmapsHead"],[71,2,1,"","ClassMapsHead"],[71,2,1,"","ClassVectorsHead"],[71,2,1,"","Head"],[71,2,1,"","MultiInstanceConfmapsHead"],[71,2,1,"","OffsetRefinementHead"],[71,2,1,"","PartAffinityFieldsHead"],[71,2,1,"","SingleInstanceConfmapsHead"]],"sleap.nn.heads.CenteredInstanceConfmapsHead":[[71,5,1,"","anchor_part"],[71,3,1,"","channels"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"id0","part_names"],[71,5,1,"","sigma"]],"sleap.nn.heads.CentroidConfmapsHead":[[71,5,1,"","anchor_part"],[71,3,1,"","channels"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"","sigma"]],"sleap.nn.heads.ClassMapsHead":[[71,3,1,"","activation"],[71,3,1,"","channels"],[71,5,1,"id1","classes"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"","sigma"]],"sleap.nn.heads.ClassVectorsHead":[[71,3,1,"","activation"],[71,3,1,"","channels"],[71,5,1,"id2","classes"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,3,1,"","loss_function"],[71,5,1,"","loss_weight"],[71,4,1,"","make_head"],[71,5,1,"","num_fc_layers"],[71,5,1,"","num_fc_units"],[71,5,1,"","output_stride"]],"sleap.nn.heads.Head":[[71,3,1,"","activation"],[71,3,1,"","channels"],[71,3,1,"","loss_function"],[71,4,1,"","make_head"]],"sleap.nn.heads.MultiInstanceConfmapsHead":[[71,3,1,"","channels"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"id3","part_names"],[71,5,1,"","sigma"]],"sleap.nn.heads.OffsetRefinementHead":[[71,3,1,"","channels"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"id6","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"id4","part_names"],[71,5,1,"id5","sigma_threshold"]],"sleap.nn.heads.PartAffinityFieldsHead":[[71,3,1,"","channels"],[71,5,1,"","config"],[71,5,1,"id7","edges"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"","sigma"]],"sleap.nn.heads.SingleInstanceConfmapsHead":[[71,3,1,"","channels"],[71,5,1,"","config"],[71,4,1,"","from_config"],[71,5,1,"","loss_weight"],[71,5,1,"","output_stride"],[71,5,1,"id8","part_names"],[71,5,1,"","sigma"]],"sleap.nn.identity":[[72,1,1,"","classify_peaks_from_maps"],[72,1,1,"","classify_peaks_from_vectors"],[72,1,1,"","group_class_peaks"]],"sleap.nn.inference":[[73,2,1,"","BottomUpInferenceLayer"],[73,2,1,"","BottomUpInferenceModel"],[73,2,1,"","BottomUpMultiClassInferenceLayer"],[73,2,1,"","BottomUpMultiClassInferenceModel"],[73,2,1,"","BottomUpMultiClassPredictor"],[73,2,1,"","BottomUpPredictor"],[73,2,1,"","CentroidCrop"],[73,2,1,"","CentroidCropGroundTruth"],[73,2,1,"","CentroidInferenceModel"],[73,2,1,"","FindInstancePeaks"],[73,2,1,"","FindInstancePeaksGroundTruth"],[73,2,1,"","InferenceLayer"],[73,2,1,"","InferenceModel"],[73,2,1,"","MoveNetInferenceLayer"],[73,2,1,"","MoveNetInferenceModel"],[73,2,1,"","MoveNetPredictor"],[73,2,1,"","Predictor"],[73,2,1,"","RateColumn"],[73,2,1,"","SingleInstanceInferenceLayer"],[73,2,1,"","SingleInstanceInferenceModel"],[73,2,1,"","SingleInstancePredictor"],[73,2,1,"","TopDownInferenceModel"],[73,2,1,"","TopDownMultiClassFindPeaks"],[73,2,1,"","TopDownMultiClassInferenceModel"],[73,2,1,"","TopDownMultiClassPredictor"],[73,2,1,"","TopDownPredictor"],[73,2,1,"","VisualPredictor"],[73,1,1,"","export_cli"],[73,1,1,"","export_model"],[73,1,1,"","find_head"],[73,1,1,"","get_keras_model_path"],[73,1,1,"","get_model_output_stride"],[73,1,1,"","load_model"],[73,1,1,"","main"],[73,1,1,"","make_model_movenet"]],"sleap.nn.inference.BottomUpInferenceLayer":[[73,4,1,"","call"],[73,5,1,"","cm_output_stride"],[73,5,1,"","confmaps_ind"],[73,4,1,"","find_peaks"],[73,4,1,"","forward_pass"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","offsets_ind"],[73,5,1,"","paf_output_stride"],[73,5,1,"","paf_scorer"],[73,5,1,"","pafs_ind"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_confmaps"],[73,5,1,"","return_paf_graph"],[73,5,1,"","return_pafs"]],"sleap.nn.inference.BottomUpInferenceModel":[[73,5,1,"","bottomup_layer"],[73,4,1,"","call"]],"sleap.nn.inference.BottomUpMultiClassInferenceLayer":[[73,4,1,"","call"],[73,5,1,"","class_maps_ind"],[73,5,1,"","class_maps_output_stride"],[73,5,1,"","cm_output_stride"],[73,5,1,"","confmaps_ind"],[73,4,1,"","find_peaks"],[73,4,1,"","forward_pass"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","offsets_ind"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_class_maps"],[73,5,1,"","return_confmaps"]],"sleap.nn.inference.BottomUpMultiClassInferenceModel":[[73,4,1,"","call"],[73,5,1,"","inference_layer"]],"sleap.nn.inference.BottomUpMultiClassPredictor":[[73,5,1,"","batch_size"],[73,5,1,"","config"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,5,1,"","integral_patch_size"],[73,5,1,"","integral_refinement"],[73,3,1,"","is_grayscale"],[73,5,1,"","model"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"],[73,5,1,"","tracks"]],"sleap.nn.inference.BottomUpPredictor":[[73,5,1,"","batch_size"],[73,5,1,"","bottomup_config"],[73,5,1,"","bottomup_model"],[73,5,1,"","dist_penalty_weight"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,5,1,"","integral_patch_size"],[73,5,1,"","integral_refinement"],[73,3,1,"","is_grayscale"],[73,5,1,"","max_edge_length_ratio"],[73,5,1,"","max_instances"],[73,5,1,"","min_line_scores"],[73,5,1,"","paf_line_points"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"],[73,5,1,"","tracker"]],"sleap.nn.inference.CentroidCrop":[[73,4,1,"","call"],[73,5,1,"","confmaps_ind"],[73,5,1,"","crop_size"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","max_instances"],[73,5,1,"","offsets_ind"],[73,5,1,"","output_stride"],[73,5,1,"","pad_to_stride"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_confmaps"],[73,5,1,"","return_crops"]],"sleap.nn.inference.CentroidCropGroundTruth":[[73,4,1,"","call"],[73,5,1,"","crop_size"]],"sleap.nn.inference.CentroidInferenceModel":[[73,4,1,"","call"],[73,5,1,"","centroid_crop"]],"sleap.nn.inference.FindInstancePeaks":[[73,4,1,"","call"],[73,5,1,"","confmaps_ind"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","offsets_ind"],[73,5,1,"","output_stride"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_confmaps"]],"sleap.nn.inference.FindInstancePeaksGroundTruth":[[73,4,1,"","call"]],"sleap.nn.inference.InferenceLayer":[[73,4,1,"","call"],[73,5,1,"","ensure_float"],[73,5,1,"","ensure_grayscale"],[73,5,1,"","input_scale"],[73,5,1,"","keras_model"],[73,5,1,"","pad_to_stride"],[73,4,1,"","preprocess"]],"sleap.nn.inference.InferenceModel":[[73,4,1,"","export_model"],[73,4,1,"","predict"],[73,4,1,"","predict_on_batch"]],"sleap.nn.inference.MoveNetInferenceLayer":[[73,4,1,"","call"],[73,5,1,"","ensure_float"],[73,5,1,"","ensure_grayscale"],[73,5,1,"","input_scale"],[73,5,1,"","keras_model"],[73,5,1,"","model_name"],[73,5,1,"","pad_to_stride"]],"sleap.nn.inference.MoveNetInferenceModel":[[73,4,1,"","call"],[73,5,1,"","inference_layer"]],"sleap.nn.inference.MoveNetPredictor":[[73,5,1,"","batch_size"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,3,1,"","is_grayscale"],[73,5,1,"","model_name"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"]],"sleap.nn.inference.Predictor":[[73,4,1,"","export_model"],[73,4,1,"","from_model_paths"],[73,3,1,"","is_grayscale"],[73,4,1,"","make_pipeline"],[73,4,1,"","predict"],[73,3,1,"","report_period"]],"sleap.nn.inference.RateColumn":[[73,4,1,"","render"]],"sleap.nn.inference.SingleInstanceInferenceLayer":[[73,4,1,"","call"],[73,5,1,"","confmaps_ind"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","offsets_ind"],[73,5,1,"","output_stride"],[73,5,1,"","pad_to_stride"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_confmaps"]],"sleap.nn.inference.SingleInstanceInferenceModel":[[73,4,1,"","call"],[73,5,1,"","single_instance_layer"]],"sleap.nn.inference.SingleInstancePredictor":[[73,5,1,"","batch_size"],[73,5,1,"","confmap_config"],[73,5,1,"","confmap_model"],[73,4,1,"","export_model"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,5,1,"","integral_patch_size"],[73,5,1,"","integral_refinement"],[73,3,1,"","is_grayscale"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"]],"sleap.nn.inference.TopDownInferenceModel":[[73,4,1,"","call"],[73,5,1,"","centroid_crop"],[73,5,1,"","instance_peaks"]],"sleap.nn.inference.TopDownMultiClassFindPeaks":[[73,4,1,"","call"],[73,5,1,"","class_vectors_ind"],[73,5,1,"","confmaps_ind"],[73,5,1,"","input_scale"],[73,5,1,"","integral_patch_size"],[73,5,1,"","keras_model"],[73,5,1,"","offsets_ind"],[73,5,1,"","optimal_grouping"],[73,5,1,"","output_stride"],[73,5,1,"","peak_threshold"],[73,5,1,"","refinement"],[73,5,1,"","return_class_vectors"],[73,5,1,"","return_confmaps"]],"sleap.nn.inference.TopDownMultiClassInferenceModel":[[73,4,1,"","call"],[73,5,1,"","centroid_crop"],[73,4,1,"","export_model"],[73,5,1,"","instance_peaks"]],"sleap.nn.inference.TopDownMultiClassPredictor":[[73,5,1,"","batch_size"],[73,5,1,"","centroid_config"],[73,5,1,"","centroid_model"],[73,5,1,"","confmap_config"],[73,5,1,"","confmap_model"],[73,4,1,"","export_model"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,5,1,"","integral_patch_size"],[73,5,1,"","integral_refinement"],[73,3,1,"","is_grayscale"],[73,4,1,"","make_pipeline"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"],[73,5,1,"","tracker"],[73,5,1,"","tracks"]],"sleap.nn.inference.TopDownPredictor":[[73,5,1,"","batch_size"],[73,5,1,"","centroid_config"],[73,5,1,"","centroid_model"],[73,5,1,"","confmap_config"],[73,5,1,"","confmap_model"],[73,4,1,"","export_model"],[73,4,1,"","from_trained_models"],[73,5,1,"","inference_model"],[73,5,1,"","integral_patch_size"],[73,5,1,"","integral_refinement"],[73,3,1,"","is_grayscale"],[73,4,1,"","make_pipeline"],[73,5,1,"","max_instances"],[73,5,1,"","peak_threshold"],[73,5,1,"","pipeline"],[73,5,1,"","tracker"]],"sleap.nn.inference.VisualPredictor":[[73,4,1,"","make_pipeline"],[73,4,1,"","predict"],[73,4,1,"","safely_generate"]],"sleap.nn.losses":[[74,2,1,"","OHKMLoss"],[74,2,1,"","PartLoss"],[74,1,1,"","compute_ohkm_loss"]],"sleap.nn.losses.OHKMLoss":[[74,4,1,"","call"],[74,4,1,"","from_config"],[74,5,1,"","hard_to_easy_ratio"],[74,5,1,"","loss_scale"],[74,5,1,"","max_hard_keypoints"],[74,5,1,"","min_hard_keypoints"]],"sleap.nn.losses.PartLoss":[[74,5,1,"","channel_ind"],[74,5,1,"","name"],[74,4,1,"","result"],[74,4,1,"","update_state"]],"sleap.nn.model":[[75,2,1,"","Model"]],"sleap.nn.model.Model":[[75,5,1,"","backbone"],[75,4,1,"","from_config"],[75,5,1,"","heads"],[75,5,1,"","keras_model"],[75,4,1,"","make_model"],[75,3,1,"","maximum_stride"]],"sleap.nn.paf_grouping":[[76,2,1,"","EdgeConnection"],[76,2,1,"","EdgeType"],[76,2,1,"","PAFScorer"],[76,2,1,"","PeakID"],[76,1,1,"","assign_connections_to_instances"],[76,1,1,"","compute_distance_penalty"],[76,1,1,"","get_connection_candidates"],[76,1,1,"","get_paf_lines"],[76,1,1,"","group_instances_batch"],[76,1,1,"","group_instances_sample"],[76,1,1,"","make_line_subs"],[76,1,1,"","make_predicted_instances"],[76,1,1,"","match_candidates_batch"],[76,1,1,"","match_candidates_sample"],[76,1,1,"","score_paf_lines"],[76,1,1,"","score_paf_lines_batch"],[76,1,1,"","toposort_edges"]],"sleap.nn.paf_grouping.EdgeConnection":[[76,5,1,"","dst_peak_ind"],[76,5,1,"","score"],[76,5,1,"","src_peak_ind"]],"sleap.nn.paf_grouping.EdgeType":[[76,5,1,"","dst_node_ind"],[76,5,1,"","src_node_ind"]],"sleap.nn.paf_grouping.PAFScorer":[[76,5,1,"","dist_penalty_weight"],[76,5,1,"","edge_inds"],[76,5,1,"","edge_types"],[76,5,1,"","edges"],[76,4,1,"","from_config"],[76,4,1,"","group_instances"],[76,4,1,"","match_candidates"],[76,5,1,"","max_edge_length_ratio"],[76,5,1,"","min_instance_peaks"],[76,5,1,"","min_line_scores"],[76,5,1,"","n_edges"],[76,5,1,"","n_nodes"],[76,5,1,"","n_points"],[76,5,1,"","pafs_stride"],[76,5,1,"","part_names"],[76,4,1,"","predict"],[76,4,1,"","score_paf_lines"],[76,5,1,"","sorted_edge_inds"]],"sleap.nn.paf_grouping.PeakID":[[76,5,1,"","node_ind"],[76,5,1,"","peak_ind"]],"sleap.nn.peak_finding":[[77,1,1,"","crop_bboxes"],[77,1,1,"","find_global_peaks"],[77,1,1,"","find_global_peaks_integral"],[77,1,1,"","find_global_peaks_rough"],[77,1,1,"","find_global_peaks_with_offsets"],[77,1,1,"","find_local_peaks"],[77,1,1,"","find_local_peaks_integral"],[77,1,1,"","find_local_peaks_rough"],[77,1,1,"","find_local_peaks_with_offsets"],[77,1,1,"","find_offsets_local_direction"],[77,1,1,"","integral_regression"],[77,1,1,"","make_gaussian_kernel"],[77,1,1,"","smooth_imgs"]],"sleap.nn.system":[[78,1,1,"","best_logical_device_name"],[78,1,1,"","disable_preallocation"],[78,1,1,"","enable_preallocation"],[78,1,1,"","get_all_gpus"],[78,1,1,"","get_available_gpus"],[78,1,1,"","get_current_gpu"],[78,1,1,"","get_gpu_memory"],[78,1,1,"","initialize_devices"],[78,1,1,"","is_gpu_system"],[78,1,1,"","is_initialized"],[78,1,1,"","summary"],[78,1,1,"","use_cpu_only"],[78,1,1,"","use_first_gpu"],[78,1,1,"","use_gpu"],[78,1,1,"","use_last_gpu"]],"sleap.nn.tracker":[[79,0,0,"-","components"],[80,0,0,"-","kalman"]],"sleap.nn.tracker.components":[[79,2,1,"","FrameMatches"],[79,2,1,"","Match"],[79,1,1,"","centroid_distance"],[79,1,1,"","connect_single_track_breaks"],[79,1,1,"","cull_frame_instances"],[79,1,1,"","cull_instances"],[79,1,1,"","first_choice_matching"],[79,1,1,"","greedy_matching"],[79,1,1,"","hungarian_matching"],[79,1,1,"","instance_iou"],[79,1,1,"","instance_similarity"],[79,1,1,"","nms_fast"]],"sleap.nn.tracker.components.FrameMatches":[[79,5,1,"","cost_matrix"],[79,4,1,"","from_candidate_instances"],[79,3,1,"","has_only_first_choice_matches"],[79,5,1,"","matches"],[79,5,1,"","unmatched_instances"]],"sleap.nn.tracker.kalman":[[80,1,1,"","get_track_instance_matches"],[80,1,1,"","match_dict_from_match_function"],[80,1,1,"","remove_second_bests_from_cost_matrix"]],"sleap.nn.tracking":[[81,2,1,"","BaseTracker"],[81,2,1,"","FlowCandidateMaker"],[81,2,1,"","FlowMaxTracker"],[81,2,1,"","FlowMaxTracksCandidateMaker"],[81,2,1,"","FlowTracker"],[81,2,1,"","KalmanTracker"],[81,2,1,"","SimpleCandidateMaker"],[81,2,1,"","SimpleMaxTracker"],[81,2,1,"","SimpleMaxTracksCandidateMaker"],[81,2,1,"","SimpleTracker"],[81,2,1,"","TrackCleaner"],[81,2,1,"","Tracker"],[81,1,1,"","run_tracker"]],"sleap.nn.tracking.FlowCandidateMaker":[[81,4,1,"","flow_shift_instances"],[81,4,1,"","get_shifted_instances"],[81,4,1,"","get_shifted_instances_from_earlier_time"],[81,5,1,"","img_scale"],[81,5,1,"","min_points"],[81,5,1,"","of_max_levels"],[81,5,1,"","of_window_size"],[81,4,1,"","prune_shifted_instances"],[81,5,1,"","save_shifted_instances"],[81,5,1,"","track_window"]],"sleap.nn.tracking.FlowMaxTracker":[[81,4,1,"","matching_function"],[81,4,1,"","similarity_function"]],"sleap.nn.tracking.FlowMaxTracksCandidateMaker":[[81,4,1,"","get_ref_instances"],[81,5,1,"","max_tracks"]],"sleap.nn.tracking.KalmanTracker":[[81,3,1,"","is_valid"],[81,4,1,"","make_tracker"],[81,4,1,"","track"]],"sleap.nn.tracking.TrackCleaner":[[81,5,1,"","instance_count"],[81,5,1,"","iou_threshold"]],"sleap.nn.tracking.Tracker":[[81,5,1,"","candidate_maker"],[81,5,1,"","cleaner"],[81,4,1,"","final_pass"],[81,5,1,"","matching_function"],[81,5,1,"","max_tracking"],[81,5,1,"","min_new_track_points"],[81,5,1,"","robust_best_instance"],[81,5,1,"","similarity_function"],[81,4,1,"","track"],[81,5,1,"","track_window"],[81,3,1,"","unique_tracks_in_queue"]],"sleap.nn.training":[[82,2,1,"","BottomUpModelTrainer"],[82,2,1,"","BottomUpMultiClassModelTrainer"],[82,2,1,"","CentroidConfmapsModelTrainer"],[82,2,1,"","DataReaders"],[82,2,1,"","SingleInstanceModelTrainer"],[82,2,1,"","TopDownMultiClassModelTrainer"],[82,2,1,"","TopdownConfmapsModelTrainer"],[82,2,1,"","Trainer"],[82,1,1,"","create_trainer_using_cli"],[82,1,1,"","get_timestamp"],[82,1,1,"","main"],[82,1,1,"","sanitize_scope_name"],[82,1,1,"","setup_checkpointing"],[82,1,1,"","setup_losses"],[82,1,1,"","setup_metrics"],[82,1,1,"","setup_new_run_folder"],[82,1,1,"","setup_optimization_callbacks"],[82,1,1,"","setup_optimizer"],[82,1,1,"","setup_output_callbacks"],[82,1,1,"","setup_tensorboard"],[82,1,1,"","setup_visualization"],[82,1,1,"","setup_zmq_callbacks"]],"sleap.nn.training.BottomUpModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.BottomUpMultiClassModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.CentroidConfmapsModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.DataReaders":[[82,4,1,"","from_config"],[82,4,1,"","from_labels"],[82,3,1,"","test_labels"],[82,5,1,"","test_labels_reader"],[82,3,1,"","training_labels"],[82,5,1,"","training_labels_reader"],[82,3,1,"","validation_labels"],[82,5,1,"","validation_labels_reader"]],"sleap.nn.training.SingleInstanceModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.TopDownMultiClassModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.TopdownConfmapsModelTrainer":[[82,3,1,"","has_offsets"],[82,3,1,"","input_keys"],[82,3,1,"","output_keys"]],"sleap.nn.training.Trainer":[[82,3,1,"","callbacks"],[82,4,1,"","cleanup"],[82,5,1,"","config"],[82,5,1,"","data_readers"],[82,4,1,"","evaluate"],[82,4,1,"","from_config"],[82,5,1,"","initial_config"],[82,3,1,"","input_keys"],[82,3,1,"","keras_model"],[82,5,1,"","model"],[82,5,1,"","optimization_callbacks"],[82,5,1,"","output_callbacks"],[82,3,1,"","output_keys"],[82,4,1,"","package"],[82,5,1,"","pipeline_builder"],[82,5,1,"","run_path"],[82,4,1,"","setup"],[82,4,1,"","train"],[82,5,1,"","training_pipeline"],[82,5,1,"","training_viz_pipeline"],[82,5,1,"","validation_pipeline"],[82,5,1,"","validation_viz_pipeline"],[82,5,1,"","visualization_callbacks"]],"sleap.nn.utils":[[83,1,1,"","compute_iou"],[83,1,1,"","group_array"],[83,1,1,"","match_points"],[83,1,1,"","reset_input_layer"],[83,1,1,"","tf_linear_sum_assignment"]],"sleap.nn.viz":[[84,1,1,"","generate_skeleton_preview_image"],[84,1,1,"","imgfig"],[84,1,1,"","plot_confmaps"],[84,1,1,"","plot_img"],[84,1,1,"","plot_instance"],[84,1,1,"","plot_instances"],[84,1,1,"","plot_pafs"],[84,1,1,"","plot_peaks"]],"sleap.skeleton":[[85,2,1,"","EdgeType"],[85,2,1,"","Node"],[85,2,1,"","Skeleton"]],"sleap.skeleton.Node":[[85,4,1,"","as_node"],[85,4,1,"","from_names"],[85,4,1,"","matches"],[85,5,1,"","name"],[85,5,1,"","weight"]],"sleap.skeleton.Skeleton":[[85,5,1,"","_is_template"],[85,5,1,"","_skeleton_idx"],[85,4,1,"","add_edge"],[85,4,1,"","add_node"],[85,4,1,"","add_nodes"],[85,4,1,"","add_symmetry"],[85,4,1,"","clear_edges"],[85,4,1,"","delete_edge"],[85,4,1,"","delete_node"],[85,4,1,"","delete_symmetry"],[85,5,1,"","description"],[85,3,1,"","edge_inds"],[85,3,1,"","edge_names"],[85,4,1,"","edge_to_index"],[85,3,1,"","edges"],[85,3,1,"","edges_full"],[85,4,1,"","find_neighbors"],[85,4,1,"","find_node"],[85,4,1,"","find_unique_nodes"],[85,4,1,"","from_dict"],[85,4,1,"","from_json"],[85,4,1,"","from_names_and_edge_inds"],[85,4,1,"","get_symmetry"],[85,4,1,"","get_symmetry_name"],[85,3,1,"","graph"],[85,3,1,"","graph_symmetry"],[85,4,1,"","has_edge"],[85,4,1,"","has_node"],[85,4,1,"","has_nodes"],[85,3,1,"","is_arborescence"],[85,3,1,"","is_template"],[85,4,1,"","load_all_hdf5"],[85,4,1,"","load_hdf5"],[85,4,1,"","load_json"],[85,4,1,"","load_mat"],[85,4,1,"","make_cattr"],[85,4,1,"","matches"],[85,3,1,"","name"],[85,3,1,"","node_names"],[85,4,1,"","node_to_index"],[85,3,1,"","nodes"],[85,5,1,"","preview_image"],[85,4,1,"","relabel_node"],[85,4,1,"","relabel_nodes"],[85,4,1,"","rename_skeleton"],[85,4,1,"","save_all_hdf5"],[85,4,1,"","save_hdf5"],[85,4,1,"","save_json"],[85,3,1,"","symmetric_inds"],[85,3,1,"","symmetries"],[85,3,1,"","symmetries_full"],[85,3,1,"","symmetry_names"],[85,4,1,"","to_dict"],[85,4,1,"","to_json"]],"sleap.util":[[86,1,1,"","attr_to_dtype"],[86,1,1,"","decode_preview_image"],[86,1,1,"","dict_cut"],[86,1,1,"","find_files_by_suffix"],[86,1,1,"","frame_list"],[86,1,1,"","get_config_file"],[86,1,1,"","get_package_file"],[86,1,1,"","json_dumps"],[86,1,1,"","json_loads"],[86,1,1,"","make_scoped_dictionary"],[86,1,1,"","parse_uri_path"],[86,1,1,"","save_dict_to_hdf5"],[86,1,1,"","uniquify"],[86,1,1,"","usable_cpu_count"],[86,1,1,"","weak_filename_match"]],sleap:[[10,0,0,"-","instance"],[36,0,0,"-","message"],[85,0,0,"-","skeleton"],[86,0,0,"-","util"]]},objnames:{"0":["py","module","Python module"],"1":["py","function","Python function"],"2":["py","class","Python class"],"3":["py","property","Python property"],"4":["py","method","Python method"],"5":["py","attribute","Python attribute"]},objtypes:{"0":"py:module","1":"py:function","2":"py:class","3":"py:property","4":"py:method","5":"py:attribute"},terms:{"0":[3,4,10,13,15,27,33,35,36,37,38,43,44,45,46,47,48,49,50,53,54,55,56,57,59,60,62,63,64,66,67,68,69,70,71,72,73,74,76,77,79,80,81,83,84,85,86,89,92,95,102,104,105,106,107,108,109,110],"00":[104,105,106,109],"0000":[105,106,109],"0000e":[106,109],"0001":[49,109],"0001038":105,"00012":105,"00022":109,"00025":109,"00031":109,"00032":109,"00037":109,"00038":109,"00045":109,"00046":109,"0010":[106,109],"0011":[106,109],"0012":[106,109],"0013":[106,109],"0014":[106,109],"0015":[106,109],"0016":[106,109],"0017":[106,109],"0018":[106,109],"0018e":109,"0019":[106,109],"001_second":89,"0020":[106,109],"0021":[106,109],"0022":[106,109],"0023":[106,109],"0024":[106,109],"0025":[106,109],"0026":[106,109],"0027":[106,109],"0028":[106,109],"0029":[106,109],"0030":[106,109],"00305":105,"0031":[106,109],"0032":[106,109],"0032e":109,"0033":[106,109],"0034":[106,109],"0035":[106,109],"0036":[106,109],"0037":[106,109],"0038":[106,109],"0039":[106,109],"003_my_video":[13,27],"0040":[106,109],"0041":[106,109],"0042":109,"0043":109,"0044":109,"0045":109,"0046":109,"0046191":105,"0047":109,"0048":109,"0049":109,"0050":109,"0051":109,"0052":109,"0053":109,"0054":109,"0054291":105,"00543":105,"0055e":109,"0056":109,"0057":109,"0058":109,"0060":109,"0061":109,"0062":109,"0063":109,"0063e":109,"0064":109,"0064e":106,"0065":109,"0066":109,"0067":109,"0067e":109,"0068":109,"0069":109,"0070":109,"0071":109,"0072":109,"0073":109,"0074e":109,"0075":109,"0076":109,"0078":109,"0079":109,"0080":109,"0081":109,"0082":109,"0083":109,"0085":109,"0085e":106,"0086":109,"0087":109,"0088":109,"0089":109,"009":105,"0091":109,"00915":43,"0092":109,"0093":109,"0097":109,"00977":105,"0098":105,"0099":109,"00995":105,"00m":[107,109],"01":[70,103,104,105,106,107,108,109],"0100e":109,"0102":109,"0105":109,"0106":109,"0108":109,"0111":109,"01126":105,"01157":105,"0116":109,"0118455":105,"0118e":109,"0123":109,"0125":109,"0126":109,"0127":109,"0128":109,"01282":105,"0128e":109,"0131":109,"01324":105,"0134e":106,"0152":105,"0159e":109,"01648":105,"0166608700891056":107,"018":38,"0190e":109,"0192e":109,"0193176269531":104,"01931763":104,"0196e":109,"0197e":106,"0198":105,"0198e":109,"02":[70,79,105,106],"02057":105,"0210e":109,"0213e":109,"0215":105,"0233e":109,"0234":38,"02444":105,"025":70,"0259":105,"0259e":109,"0263e":109,"0264e":109,"0266e":109,"0270e":109,"02728":105,"0275":105,"0277e":109,"02960205":104,"02960205078125":104,"03":[13,27,70,103,104,105,109],"0301":105,"0303e":106,"0311976":105,"03162":105,"03174":105,"0318":105,"0318e":109,"0323e":109,"03244":105,"0365":105,"03793":105,"0385e":106,"038687":109,"0386e":109,"03918":105,"0394e":109,"03998":105,"04":[70,103,104,105,106,107,109],"0420e":109,"0426e":109,"04318":105,"04324":105,"04352":105,"04425":105,"04437362":105,"0444e":109,"0446e":109,"0451e":109,"0462":105,"0466e":109,"04684":105,"0474e":109,"04776":105,"04788":105,"0479":105,"04814":105,"0488e":109,"04938014":105,"04953":105,"04962":105,"0497":105,"04d":50,"05":[70,103,105,109],"0507e":109,"0515":105,"0518e":109,"05255":105,"0531e":[106,109],"05379":105,"05388":107,"0541e":106,"0545800141197903":107,"05463":105,"05493":105,"055738":106,"05611448":105,"0561e":109,"056597":106,"057389":106,"0575e":109,"0576e":109,"058046":106,"058048":104,"0581e":106,"0584":105,"05853":105,"0585e":109,"06":[49,70,106,109],"060007":104,"060013":104,"0606e":109,"0607e":109,"0612e":109,"06203796":105,"06238":105,"0630e":106,"0631e":109,"06328589":105,"0634":105,"0634e":109,"0639e":109,"06403984":105,"06436":105,"0646795":105,"06516132":105,"0652":105,"0657e":109,"06647":105,"0668":105,"066840":109,"0668e":109,"06784":105,"0690e":106,"06937":38,"07":[70,107,109],"07055778":105,"0711e":109,"0713":105,"072":70,"0720e":109,"0727e":109,"07603623":105,"0771e":109,"0772e":106,"0778e":109,"07797408":105,"08":[49,70,104,106,109],"08127":105,"08318384":105,"0835e":109,"08496":105,"08594":105,"08648682":105,"08688":105,"08721923828125":104,"08721924":104,"0888e":109,"089801":109,"08987":105,"08994":105,"08m":104,"09":[70,105,106,107,108,109],"09042":105,"0904e":109,"0905e":109,"090652":109,"09125":105,"09143":105,"091464":109,"09204":105,"09212":40,"092164":109,"09326":38,"09349":105,"09372":105,"09384":105,"0939e":106,"0943e":106,"0944e":109,"097":105,"09735":105,"0978":105,"0986e":109,"098811":109,"09885":105,"0989e":109,"0k":[104,105,106],"0m":[105,106,107,109,110],"0m0":109,"0m2023":109,"0m31m48":109,"0m31m49":109,"0m31m51":109,"0m31m52":109,"0mb":109,"0mcould":107,"0mfp":109,"0mm":109,"0munabl":107,"1":[0,3,4,9,10,12,13,27,33,35,36,38,39,40,41,42,43,44,45,46,47,48,49,50,51,54,55,56,59,60,61,62,63,64,65,66,67,68,69,70,71,73,74,76,77,79,80,81,83,84,85,87,89,95,98,102,107,109,110,119],"10":[36,41,46,49,50,70,73,76,81,84,102,103,104,105,106,107,108,109,117],"100":[49,81,87,92,103,104,105,106,108,109],"1000":[87,89,105],"10000":33,"1004e":109,"1007e":109,"101":[103,106,109],"1018e":109,"1019":105,"1024":[87,103,104,105,107],"103":[43,63,104,107],"103255":109,"10357":40,"103658":105,"103982":109,"104":104,"10415649":104,"10415649414062":104,"104377":105,"104556":106,"105059":105,"10544":105,"10553":105,"106":109,"106019":105,"1060e":109,"107":70,"1078e":109,"107m":109,"108":109,"10843":105,"1089e":106,"109":109,"1097e":109,"10m":101,"11":[70,87,102,103,104,105,106,107,108,109],"110":109,"1100":33,"11035":105,"11078":105,"11148":105,"111973079":109,"1119e":109,"112":104,"11206":105,"1129e":109,"113":[103,109],"114":109,"1149e":109,"11572":105,"1158e":109,"116":[43,63],"1164e":109,"11816":105,"1183e":109,"1192":105,"11ea":109,"12":[70,90,102,103,104,105,106,107,108,109],"120":105,"1201e":109,"1204e":109,"120k":[105,106],"121":103,"1218e":109,"122":103,"1221e":109,"1223e":109,"123":[43,63,96],"1233e":109,"12372":105,"1238e":109,"12497":105,"12512":105,"127":[36,43,46,50,63,109],"12731166":105,"1278e":109,"128":[39,42,43,48,49,84,103,109],"1280":87,"12836":105,"12885":105,"12k":[104,105,106],"13":[70,87,103,105,106,107,109],"13065":105,"13086":105,"13202":105,"1323e":109,"13245":105,"13287":105,"132k":106,"133":[103,109],"13306":105,"13345":105,"1338":105,"1339":105,"134":109,"13475017":105,"13495":105,"1350":108,"1375e":106,"1395":105,"13pt":[106,107],"14":[70,87,104,105,106,107,108,109],"140":103,"1400e":109,"14189687":105,"142":104,"1421":105,"14230077":105,"14267":105,"14295959":104,"14295959472656":104,"14362":105,"144":[104,109],"1440":106,"145283":105,"1452e":109,"14550781":105,"14614868":104,"1461486816406":104,"1474":87,"1484228079076315":107,"1486e":109,"1497e":106,"149k":104,"14m":109,"15":[35,48,70,71,88,98,100,101,103,104,105,106,107,108,109,110],"150":[87,107],"1500":87,"15000":[104,105],"15082283":105,"1508674":105,"1508e":109,"150k":104,"151":[103,105,106,107,109],"1523e":109,"1525":[105,106,109],"15356":105,"1536":87,"1550e":109,"1555e":109,"1558":105,"1559":105,"156":109,"1560e":109,"15668":105,"15704":105,"157392":109,"1578e":109,"157k":[105,106],"158019":109,"1581400":104,"15857":105,"1588":87,"158864":109,"15892":105,"158k":105,"15924":105,"159656":109,"15m":109,"16":[43,45,47,48,70,79,104,105,106,107,108,109],"160":106,"1600e":109,"1603":38,"1604":87,"1606":43,"1611":38,"16116264":105,"1619e":109,"1631":105,"16318":105,"16357":105,"16394":105,"16397":105,"164":105,"16411652":105,"16653292":105,"1667":105,"16724":105,"16742":105,"16777216":[104,107],"16904":105,"1692e":106,"1696e":109,"1699e":109,"16df8d00":109,"16m":109,"17":[70,73,106,108,109],"1707":107,"17175":105,"172":104,"17218":105,"17297":105,"17343":105,"1743e":106,"1751e":106,"1753e":109,"17560":[104,105],"1757":105,"17578":105,"176":104,"1763e":109,"176m":105,"1774":[105,106,109],"17917":105,"17926":105,"179k":106,"18":[34,40,70,87,106,109],"180":[49,88],"1800":[104,105,107],"1802e":109,"18051867":105,"1807e":106,"1815e":109,"1817":105,"1826e":106,"1827e":109,"18298":105,"1831665":105,"185":109,"1850":[107,108],"1859":105,"18735":105,"188":104,"18854":105,"189":104,"18994":105,"18m":109,"19":[70,100,101,109],"1902":40,"1903e":109,"19041":102,"19046":105,"1905e":109,"190719_090330_wt_18":104,"190719_090330_wt_18159206_rig1":[104,105],"1907e":106,"1908":40,"191205_162402":96,"191205_163413":96,"1913e":109,"192":[87,109],"1925e":109,"193":104,"1934":104,"1936":105,"19464":105,"19696":105,"19714":105,"19766915":105,"19778":105,"197896466395166":107,"197k":[104,105],"1985e":109,"199":109,"1998e":109,"19m":[106,109],"1e":[49,106,109],"1f":105,"1g":105,"1k":[104,105,106],"1m":[104,105,109],"1s":[104,106],"1x1":[38,45,48],"2":[3,6,9,10,12,13,34,37,38,39,41,43,44,45,48,49,53,54,55,56,59,60,61,62,63,64,66,67,70,71,72,73,74,76,77,80,81,83,84,87,89,92,100,102,103,107,109,119],"20":[4,76,81,87,103,104,105,106,107,109,116,117],"200":[15,49,103,104,109,110],"2000":[33,87,89],"20065":105,"2012e":[106,109],"2014":87,"2015":[44,79],"2016":77,"2017":[70,76,117],"2018":[42,95],"2019":[41,87,101],"20190128_113421":109,"2020":101,"2021":[103,104,105],"2022":[87,100,101],"2022e":109,"2023":[104,105,106,107,108,109],"20230901t203033z":109,"20264":105,"203":105,"20336914":105,"2035e":109,"2043":105,"2048":87,"20544":105,"20596":105,"2068":105,"20728":105,"20752":105,"20844":105,"20847":105,"2087e":106,"208933":107,"2092e":106,"209734":107,"209771":107,"209801":107,"209829":107,"209859":107,"209886":107,"2098e":106,"2099":105,"209912":107,"209939":107,"209945":107,"20filenam":109,"20m":106,"21":[1,34,70,81,87,89,102,104,105,106,107,108,109],"210205":106,"21050":104,"210504_182918":[104,105],"210505_0126":104,"210505_012601":[104,105,107],"210506_230852":104,"210519_111253":105,"21084":109,"211":103,"2110e":109,"21112":105,"21129":105,"21145":109,"21151":109,"21155":105,"21188":105,"212":[103,105],"21204":105,"21255928":105,"21261":106,"2130e":106,"2156e":109,"2159e":109,"216":105,"2162e":[106,109],"21640316":105,"2166":105,"2169e":109,"217":104,"2172e":109,"2174e":109,"21753":105,"2177e":109,"2181147":105,"219":105,"21m":106,"22":[70,109],"220":105,"22018598":105,"2208":105,"22192":105,"222":105,"2224":105,"2224e":109,"223":105,"223182":104,"223923":104,"223968":104,"223999":104,"224":[43,63,105],"224028":104,"224057":104,"224084":104,"224111":104,"224140":104,"224144":104,"2247":105,"2248e":109,"225":[63,105],"2254e":106,"2262":105,"22705":105,"2276387":105,"22766":105,"2278e":106,"22808443":105,"2288e":106,"229":[63,105],"22980":109,"23":[43,70,87,100,105,106,109],"23050":109,"23056":109,"23196":105,"232":105,"2339e":106,"234":[104,105],"23578":105,"2364e":109,"2365e":109,"23718":105,"2374e":109,"2377":105,"237m":109,"2389":105,"239":105,"2397":105,"2397e":109,"239m":104,"24":[70,90,103,106,109],"240":[104,105],"2402e":106,"2403":105,"241":105,"2417e":109,"242":[104,105],"2425":105,"2428e":109,"243":105,"2437":105,"2438e":109,"244":105,"24440002":104,"24440002441406":104,"2447e":109,"2449e":109,"245":[104,105],"245745":107,"246":[104,105],"2467":105,"24688721":105,"2471e":[106,109],"2476":105,"24774":105,"24854":105,"2485e":109,"2486e":109,"24900974":105,"24979575":105,"24988":105,"24997":105,"24999984211172e":109,"249999968422344e":109,"24m":106,"25":[70,73,76,77,87,89,103,105,108,109],"250":104,"2500e":109,"2502e":109,"250591":108,"250602":108,"251":105,"2511e":109,"2513e":109,"252":105,"25208":105,"25244":105,"2535e":109,"25458":105,"2546":105,"254912":109,"255":[33,43,47,49,63,92,109],"2551e":109,"255468":109,"25574":105,"256":[39,40,42,48,49,53,105,109],"2560":[104,105],"256291":109,"2563e":109,"257158":109,"2579e":109,"258":101,"25829":105,"25912":105,"25969282":105,"25983":105,"25hfinish":109,"25hinfo":109,"26":[70,92,95,104,109],"2600708":105,"2600e":109,"2604e":109,"26096":105,"2615":105,"26282":105,"2632598876953":104,"26325989":104,"2633":105,"263375180":109,"2635e":106,"26373":105,"2637e":106,"2648e":109,"2650e":106,"26575":105,"26588":105,"26732486":105,"26816":105,"268435456":[104,107],"2695e":109,"26993":105,"27":[70,104,107,109],"271":[104,105],"27118":105,"2713":105,"272":[104,105],"273":105,"27322":105,"274":105,"2743e":109,"275":105,"2752e":109,"27567":105,"2759":105,"2764":105,"27682":105,"27792":105,"2783e":109,"2793e":109,"27942":105,"28":[70,107,109],"280029296875":104,"2800293":104,"2823e":109,"28253":105,"2828e":109,"283":105,"2838e":109,"284":105,"285":[104,105],"286":105,"28660768":105,"287":105,"2873":105,"2879":105,"288":105,"2889e":106,"2893e":109,"28958":105,"29":[70,104,107,108,109],"29004":105,"29108":105,"292":105,"2922e":109,"2925e":106,"2940e":109,"29464462":105,"2947e":109,"2948":87,"2950":87,"29553":105,"29578":105,"296230286359787":104,"29968":105,"29m":104,"2d":[9,54,56,59,76,77],"2e65b":109,"2f":107,"2f20230901":109,"2faws4_request":109,"2foctet":109,"2fs3":109,"2fu":109,"2kpredict":109,"2m":105,"2s":[105,106,109],"3":[1,3,10,13,27,33,38,39,41,42,43,44,48,49,51,54,56,61,62,63,66,70,76,77,80,81,83,86,87,89,95,100,102,103,104,107,109,110,119],"30":[34,87,100,104,105,109],"300":109,"3000":103,"3003e":109,"3006e":109,"30078":105,"30173":105,"30188":105,"30194":105,"302":[103,109],"3027e":109,"3028e":109,"30325472":105,"30384466":105,"304":105,"30457":105,"305":[104,105],"3055e":109,"306":105,"30673":105,"307":[104,105],"30713":105,"30775963":104,"3077e":109,"308":[104,105],"30814":105,"3086853":105,"309":105,"3093e":106,"3099":105,"30m":[40,104,105],"31":[70,104,105,106,109],"310":105,"311":[105,109],"3113e":109,"312":105,"3123e":106,"3126e":109,"313":105,"31342":105,"3135e":109,"31372":105,"3139":105,"314":105,"315":105,"31561279":105,"3157":105,"3157e":109,"316":105,"31602":105,"3161e":109,"31635":105,"317":[104,105],"317060":107,"318":105,"3180e":106,"319":[104,105],"31976":105,"31982":105,"31m":[105,106,109,110],"31m126":109,"31m14":109,"31m184":109,"31m270":109,"31m33":109,"31m39":109,"31m51":109,"31m7":109,"31me":107,"31merror":[105,106,109,110],"32":[34,38,42,43,47,48,70,87,105,106,109],"320":105,"32009887695312":104,"32009888":104,"3201e":106,"32022":105,"320224":107,"3208e":109,"321":[104,105],"3215e":106,"3218e":109,"322":[104,105],"3221e":109,"32227":105,"32256":105,"323":105,"32303":105,"32346":105,"324":105,"324520":109,"32465":105,"325":[104,105],"32507":105,"32578":105,"326":[105,109],"3265e":109,"3269":105,"327":[104,105],"3279e":109,"328":105,"32881752":105,"329":[104,105],"329181":109,"3299":105,"329961":109,"32m":[106,109],"32m20190128_113421":109,"32mcourtship_label":109,"33":[70,109],"330":105,"331":[104,105],"33124":105,"332":105,"333":105,"334":105,"33406225":105,"3346e":109,"33477783":105,"335":105,"33521563":105,"33566353":105,"3358e":109,"336":105,"3368e":109,"337":105,"337645":109,"338":105,"339":105,"33m":105,"34":[44,70,106,109],"340":105,"341":105,"3421e":109,"3431e":109,"34476":105,"345":105,"3458":105,"3465":105,"347":105,"3472e":109,"348":105,"3483e":106,"3487e":106,"3493":105,"3497e":109,"3499":105,"34m__macosx":109,"34mcourtship":109,"34mdataset":109,"34mdrosophila":109,"34mmodel":109,"34mtd_fast":107,"35":[70,87,105,109],"35016":105,"351":105,"35138":105,"3515e":109,"3519e":109,"353":105,"35394":105,"354":[104,105],"35414153":105,"35440978":105,"3545e":109,"35481":105,"355":105,"35507":105,"356":105,"3565":105,"357":105,"3573e":109,"35754":105,"358":105,"3584":105,"35846":105,"358483":105,"35867":105,"3589e":109,"359":105,"35m":[104,109],"35m100":109,"35mexampl":109,"35mnone":109,"36":[70,109],"360":[105,106],"3600":[104,107],"3604":105,"36052382":105,"361":105,"36105":105,"3611e":109,"3617e":[106,109],"362":105,"36204475":105,"36255":105,"36267":105,"363":[104,105],"3632":105,"36377":105,"3638e":109,"364":105,"3648e":109,"36536":105,"3658e":106,"366":[105,106,109],"36603":105,"3664e":106,"36795":105,"368":104,"369":105,"36913":105,"36966":105,"3697e":109,"36m":109,"36m0":109,"36m1":109,"36m4":109,"36m7":109,"37":[70,105,109],"370":105,"3719":105,"372":105,"3721e":109,"3737e":106,"374":105,"375":105,"3752e":106,"376":104,"37636":105,"377":105,"3770e":109,"3774":105,"3775":105,"3776e":106,"378":105,"379":105,"3791":105,"3792":105,"38":[70,104,109],"38037":105,"38046":105,"3809e":109,"381":105,"38232":105,"38303":105,"384":[33,105],"3846e":109,"3847992":105,"385":[104,105],"3852":105,"3858":105,"3859e":109,"386":105,"3863e":109,"3869":105,"387":105,"388":105,"3881e":109,"3886e":109,"389":105,"389239":108,"3896":105,"38m":109,"39":[70,104,105,106,109],"390":105,"390139":108,"390188":108,"390230":108,"390267":108,"390306":108,"390345":108,"390383":108,"390421":108,"390425":108,"3906e":109,"3908":105,"391":105,"3911e":109,"39127233624458313":104,"39133":105,"3913e":109,"393":[104,105,109],"3931e":109,"39355":105,"394":105,"3947":105,"39497375":105,"39508":105,"3959e":109,"396":105,"397":105,"39786857":105,"398":[104,105],"39813":105,"3986":105,"398k":104,"399":105,"39m":104,"3b":109,"3d":9,"3d1800":[104,105,107],"3ddrosophila":109,"3m":105,"3mb":104,"3s":[106,109],"3x3":[45,48,62,77],"4":[0,3,13,33,35,39,40,42,43,44,45,48,49,54,56,62,70,73,77,80,83,84,87,89,92,98,100,101,103,104,107,109,110],"40":[4,87,106,109],"400":[49,105],"4000":87,"40005":105,"40057":105,"4005e":106,"401":105,"40115":105,"4015000":[104,105],"4018e":109,"40190646":105,"4019e":106,"402":105,"40244":105,"4029e":109,"4033e":109,"40347":105,"404":105,"4047e":109,"405":105,"40552":105,"4058":105,"406":[63,104,105],"4065":105,"4066":105,"407":105,"40717":105,"408":105,"4080e":109,"4087e":109,"409":105,"4097":105,"41":[70,109],"410":105,"4100e":109,"4107e":109,"4109717":105,"411":105,"4111":105,"4112e":109,"41144":105,"412":105,"412k":105,"413":[104,105],"414":105,"41422":105,"4143e":109,"4145e":106,"41486":105,"4151e":106,"4152e":109,"41534":105,"416":105,"417":105,"4179e":106,"418":105,"41803":105,"419":105,"4191e":106,"4193":105,"42":[70,105,106,107,109],"420":[104,105],"42017":105,"421":[104,105],"4215e":109,"42197":105,"422":105,"422k":105,"423":105,"42343":105,"42352":105,"42374":105,"42407":105,"4241":105,"425":[87,104,105],"42508":105,"4254e":109,"4258":105,"426":105,"427":105,"42723":105,"428":105,"42874":105,"4288e":106,"42896":105,"429":105,"4292e":109,"42938":105,"4297":105,"43":[70,105,106,107,109],"430":[104,105],"43085":105,"431":105,"43195":105,"432":105,"43228":105,"43257":105,"4325e":106,"4326e":109,"433":[104,105],"43307":105,"4330e":109,"43337214":105,"434":105,"4342":105,"4343e":109,"435":[104,105],"43555":105,"436":[104,105],"437":105,"43756":105,"4376e":106,"438":[104,105],"4382e":109,"43842":105,"43861":105,"43884":105,"439":105,"4398e":106,"44":[70,105,106,107,109],"440":[104,105],"441":[104,105],"4413e":106,"4415e":109,"442":105,"4425e":109,"4428":105,"443":[103,104,105,109],"4437e":109,"444":105,"4440e":[106,109],"4441e":106,"44467":105,"445":[104,105],"445179":104,"445232":104,"445236":104,"4453":105,"446":105,"44635":105,"4464":105,"44681":105,"4471e":109,"44776785":105,"448":105,"44873":105,"449":105,"44904":105,"4493e":109,"45":[70,87,104,109],"450":[104,105],"4500e":109,"4505e":109,"45065":105,"4506e":109,"4507":105,"451":105,"45117":105,"4526":105,"4529e":109,"4532":105,"45367":105,"45417705":105,"4541e":109,"4546e":106,"455046":105,"455402":109,"45566":105,"456":[63,96,105],"456138":109,"4567e":109,"456803":109,"457":105,"45706":105,"457464":109,"45755":105,"458":105,"45853":105,"459":[104,105],"4593":105,"4597148":105,"46":[70,104,109],"460":105,"461":105,"46105212":105,"46127":105,"462":105,"46237":105,"46246":105,"46274":105,"4627e":106,"46356":105,"4637":105,"4640e":109,"46423":105,"4643e":109,"46442":105,"465":105,"46533":105,"4653e":109,"4657e":106,"466":105,"467":105,"4673":105,"46782":105,"46788":105,"4680e":106,"4682":105,"469":105,"46994":105,"47":[70,104],"4706e":109,"471":105,"4712e":106,"4716e":109,"472":105,"4725":105,"4725e":109,"4726e":109,"4728e":109,"473":105,"47393468022346497":104,"474":[104,105],"475":[104,105],"4751":105,"47512543":105,"47549438":105,"4755":105,"4756":105,"47607":105,"4763e":109,"4767e":109,"4768":105,"4770e":106,"477220":106,"47867":105,"4787e":109,"479":105,"47983":105,"48":[70,104,107],"480":104,"481":105,"4811":105,"481k":[105,106],"4821e":109,"4829e":106,"48315":105,"4831e":109,"48407":105,"48416":105,"4847e":106,"485":63,"48532":105,"48575":105,"4858":105,"4860e":106,"48629727959632874":104,"48642":105,"4865":105,"486k":104,"48721":105,"4886e":109,"4893e":106,"4898e":109,"48993":105,"49":[70,104,105,107,109],"4912":105,"49136":105,"49152":[104,107],"4916":105,"4916717":105,"4919e":106,"4927e":109,"49347":105,"49423":105,"49429995":105,"49533":105,"4963e":109,"49683":105,"49707":105,"49756864":105,"49778372":105,"49817":105,"498908":104,"49898362":105,"499999936844688e":109,"4d":[9,33,40],"4m":106,"4s":[104,109],"5":[1,4,6,13,27,37,38,42,43,48,49,53,60,62,63,70,71,73,74,76,77,81,83,86,87,89,102,103,104,105,107,108,109,110],"50":[95,103,104,107],"5000e":109,"5004723":105,"5007e":109,"5009e":109,"501":106,"50101817":105,"5017098":105,"5036":105,"5036e":109,"50397":105,"5043":105,"5052e":109,"50543":105,"5054e":106,"5058":105,"5058e":109,"50616":105,"5061e":109,"50806":105,"508754":106,"50894":105,"5089e":109,"50916994":105,"50th":[70,107],"51":70,"5118":105,"512":[105,109],"51236":105,"51400757":105,"51434":105,"5147":105,"5150e":109,"51591":105,"51675177":105,"5170047":105,"5174e":106,"5177e":109,"5178":105,"51843":105,"51874":105,"51895":105,"5190e":[106,109],"51m":104,"52":[70,104,105],"52002":105,"52005":105,"52008056640625":104,"52008057":104,"5206e":109,"5207e":109,"5212e":109,"52148":105,"5228e":[106,109],"5232e":106,"523632":105,"524288":[104,107],"52533":105,"52646":105,"52814":105,"5282e":106,"529680":106,"53":[70,104,105],"53275":105,"53324103":105,"53396606":105,"53412354":105,"5342":105,"5347014":105,"535":105,"5352e":106,"5362e":109,"5364e":109,"5365":105,"5365e":109,"53745":105,"538":103,"5381e":109,"5387e":109,"5388288":105,"5396e":106,"539k":103,"53m":109,"54":[70,104,109],"54178":105,"5420e":109,"54233":105,"5428e":109,"5451e":109,"5454":105,"54587":105,"546117":109,"5463986":105,"5468":105,"546866":109,"5469":105,"5473":105,"547533":109,"5475e":109,"5481076":105,"548184":109,"5494046":105,"54964":105,"5496521":105,"54977":105,"55":[70,109],"55066":105,"5512e":106,"55141157":105,"551501":103,"5518e":109,"5519e":109,"5523e":106,"5540e":106,"55441":105,"5555e":106,"5559":105,"55792":105,"5580e":109,"5588e":106,"5589e":109,"559":105,"5590857":105,"559100":106,"5591e":109,"55927":105,"5596e":109,"55972":105,"55m":109,"56":[70,104,105],"5609e":109,"56104":105,"5611e":106,"561211":105,"5621643066406":104,"56216431":104,"5624":105,"56253":105,"5631e":106,"5642e":109,"56442":105,"5647e":109,"5649e":106,"565":105,"5654e":109,"5656e":109,"56747":105,"56854":105,"56874":105,"56883":105,"56894":105,"5694e":109,"56982":105,"56m":109,"57":[70,104,105,109],"57007":105,"570176":105,"57062":105,"57127994":105,"57135009765625":104,"5713501":104,"57254":105,"57278":105,"5733776":105,"5735":105,"57388":105,"5739e":109,"5740245":105,"5744":105,"574921":106,"57523":105,"5752e":109,"5753":105,"5754e":109,"575856":105,"5764e":109,"57666":105,"5772e":109,"57848":105,"5797":105,"5798e":109,"57m":109,"58":[70,104,106],"5806e":109,"5809e":109,"5814e":109,"58252":105,"5831e":109,"58366936":105,"58368":105,"58414":105,"58466":105,"58485055":83,"585451":106,"5855e":109,"5856":105,"586040":106,"5861632227897644":104,"5869e":109,"5874e":109,"5877e":109,"588":105,"58817774":105,"5883e":106,"5890152":105,"5894165":104,"58941650390625":104,"58961":105,"58k":103,"58m":109,"59":[70,87,104],"59024":105,"59042":105,"5914":105,"5916":105,"5931e":106,"59384":105,"59448":105,"5945e":109,"5949e":109,"5957e":109,"5965e":106,"5981e":109,"59887695":104,"598876953125":104,"5989e":109,"59937":105,"59984":105,"5998e":109,"5k":104,"5m":[44,104,105],"5s":109,"5x":48,"6":[43,70,84,89,100,102,103,104,105,106,107,108,109,110],"60":101,"600":101,"60114":105,"6015e":106,"60196":105,"60254":105,"6026":105,"6030e":109,"60315":105,"6031595":105,"6039":105,"6045e":109,"6051e":109,"6053466796875":104,"60534668":104,"60617":105,"6061791":105,"6070e":106,"6076e":109,"6084e":109,"609":101,"6091":105,"6096e":106,"61":[70,109],"6100e":109,"61078":105,"61108":105,"6120859":105,"61237":105,"6125536":105,"6145e":106,"61487":105,"6160e":109,"6171893":105,"6172e":109,"61740112":105,"6176859":105,"6177351":105,"6194322":105,"6195e":109,"619m":106,"62":70,"620m":[105,106],"6215":105,"6223k":105,"6224e":109,"62262":105,"62274":105,"6229":105,"622m":106,"6231525":105,"624":105,"6250e":109,"62582":105,"62598":105,"6272e":109,"6274e":109,"6281e":109,"62988":105,"63":[70,105,109],"63016":105,"6322e":109,"63318":105,"6341":105,"6343123":105,"63477":105,"6352e":109,"63538":105,"6368e":109,"6372537":104,"63763":105,"6378e":106,"63806784":105,"6394e":109,"6399e":109,"64":[11,38,40,41,43,44,45,48,55,70,71,102,103,104,107,108],"64005":105,"6401e":109,"64023":105,"6407e":109,"6409":105,"641":105,"6411":105,"6412e":109,"64131":105,"64159":105,"6417e":109,"6432e":109,"64355469":105,"64372706":105,"64392":105,"6439e":106,"6440e":109,"6441":105,"6446":105,"64484":105,"64542":105,"6464e":109,"6468e":109,"64728":105,"6477e":109,"6479e":109,"6483e":109,"64882857":105,"64886":105,"64966":105,"65":70,"65015":105,"6513e":106,"65207":105,"6521e":109,"6535e":109,"65408":105,"6547e":109,"6551e":106,"6553":105,"65547":105,"6554e":109,"6555e":109,"65576":105,"65726":105,"6575725":105,"6584e":109,"6591e":109,"65952":105,"65994":105,"66":70,"6601":105,"66077":105,"6607e":109,"6610e":109,"66180676":105,"661k":105,"6627722":105,"66278":105,"66385":105,"66406":105,"6642":105,"6647e":109,"6663e":109,"6666e":109,"6669e":109,"6671":105,"66855":105,"66882":105,"67":[70,109],"6708e":109,"67142":105,"6716e":109,"6717e":109,"67194":105,"67218":105,"6727e":109,"6739e":109,"6743e":106,"6751":105,"6759e":109,"6765e":109,"6766":105,"6766811":105,"67697":105,"67744":105,"67786":105,"6780e":109,"6782e":109,"67886":105,"67926":105,"6795893":105,"68":[43,63,70,104,105],"68062663":105,"68115":105,"68118":105,"6811e":109,"6819e":109,"68225":105,"6822e":109,"68262":105,"68353":105,"6837e":109,"684":109,"68488":105,"6862":105,"68756":105,"6879568":105,"68842":105,"6884535":105,"68896":105,"69":[70,104],"690":[104,107],"6906e":106,"691378":109,"6915e":109,"6921e":109,"6930795":105,"69318":105,"6934e":109,"6936e":109,"6939e":109,"6943":105,"6945e":[106,109],"6946":105,"6961e":106,"69723856":105,"6987e":109,"69904":105,"69998":105,"6b0":1,"6c03a2f89e1c":109,"6j":87,"6m":[104,105],"7":[39,70,86,87,89,102,103,104,105,106,107,108,109],"70065":105,"7012927":105,"7020806":105,"7034e":109,"7035529":105,"7037247":105,"7037e":106,"7041355":105,"7051566":105,"7055e":109,"7064677":105,"7064e":109,"70679":105,"70728":105,"70898":105,"71":[70,104],"7110e":106,"7111":105,"7111e":109,"71122557":105,"71146":105,"7118e":106,"71207":105,"71269846":105,"7130e":106,"713k":[104,105,106],"7147":105,"71594":105,"71674":105,"71683":105,"71751094":105,"71759033":104,"71759033203125":104,"7180e":106,"71852":105,"71884066":105,"7197965":105,"7199e":106,"72":[70,84,104,109],"7203e":109,"72043":105,"72051835":105,"7208e":109,"7216":105,"72241":105,"72244436":105,"72284":105,"7234e":109,"72357":105,"7239725":105,"7244":105,"7247e":109,"72516":105,"725241":109,"7256":101,"72562":105,"7257e":106,"72635":105,"7274184":105,"7278":105,"72796327":105,"72803":105,"72806":105,"7281":105,"72864616":105,"7287e":109,"7288e":109,"7290425":105,"7290e":109,"73":[70,109],"73026":105,"7307":105,"7312":105,"731425":105,"7323e":106,"73245":105,"7324e":109,"7328e":109,"7331":105,"7343e":109,"73447":105,"7345e":106,"73501164":105,"73508":105,"7351e":106,"7352e":109,"7357e":109,"735933":105,"7361e":109,"736867":105,"73703":105,"7370e":106,"7373e":109,"7375747":105,"73767":105,"7378067":105,"7379e":109,"7396e":106,"7399e":109,"74":[70,105,109],"7404":105,"74078":105,"7414":105,"74188":105,"743163":105,"7434e":106,"74356":105,"74454":105,"74457":105,"74493":105,"7453e":106,"7456":105,"74594":105,"7461e":109,"74677":105,"7468e":[106,109],"7469e":109,"74753":105,"7489":105,"7489e":109,"7494e":109,"7495e":109,"75":70,"7511e":109,"7513e":[106,109],"75169766":105,"75269":105,"7544447":105,"755154":109,"7552497":105,"7553143":105,"7555":105,"7555e":109,"75566101":105,"75584614":105,"756024":109,"75647":105,"7566e":109,"757213":109,"7574282":105,"758315":109,"7586e":109,"75934":105,"75th":[70,107],"76":70,"7613073":105,"7613e":109,"7617":105,"7624657":105,"76257":105,"7625e":109,"7628807":105,"76299024":105,"7635e":109,"7636e":109,"7640e":106,"7656e":109,"7662":105,"76627":105,"76758":105,"7679e":109,"7682":105,"7685":105,"7697851":105,"77":[70,104],"7702941894531":104,"77029419":104,"77054":105,"7713":105,"77133":105,"7715297":105,"7721256017684937":104,"77222":105,"77227783":105,"7722e":109,"77295":105,"77327":105,"77335":105,"77365":105,"7738e":109,"7741e":[106,109],"77439654":105,"7750215":105,"775633":106,"775644779205322":109,"7761":105,"77646":105,"776555":106,"77664":105,"77673894":105,"77695":105,"77728":105,"777493":106,"77814513":105,"778196":106,"779":[43,63],"78":[70,104,105,106,107,108,109],"78006":105,"7803e":109,"780803382396698":104,"7812":105,"781426":105,"78146":105,"78155517578125":104,"78155518":104,"78204":105,"7820e":109,"7827e":106,"7838e":109,"7842e":109,"78436":105,"7844e":109,"78534484":105,"7854e":106,"78555703":105,"7855e":109,"78564":105,"7869e":109,"78735":105,"7880e":109,"7884":105,"7885e":109,"7888e":109,"7891e":109,"78940654":105,"7895":105,"78m":105,"79":[70,109],"7902795":105,"7903e":109,"79062":105,"7908e":109,"79095":105,"7911e":109,"79128":105,"79132":105,"7932":105,"79385":105,"7938e":109,"7941327":105,"79416":105,"7941e":109,"7946e":106,"794787":105,"7948513":105,"79486":105,"7954e":109,"79565":105,"7970e":109,"7977e":109,"797836431061851":107,"7978e":109,"79828":105,"7989652":105,"798996":105,"799":106,"7991047":105,"7991e":109,"79m":109,"7m":105,"7mb":104,"7s":[106,109],"8":[13,43,48,49,55,70,81,87,89,102,103,104,105,106,107,108,109],"80":[87,104],"80017":105,"80063":105,"80096":105,"8017e":109,"8021e":109,"8027e":109,"8031e":109,"80396":105,"804":87,"8041":105,"8042e":109,"80545":105,"8055":105,"806004":105,"8066e":109,"807011":105,"8070e":109,"8071e":109,"8074e":109,"8076":105,"8076e":106,"80786":105,"807970":105,"8081e":109,"8082221":105,"8082359":105,"8082982":105,"80847":105,"8088e":109,"808962":105,"8091e":109,"80968696":105,"81":[70,104,105],"8100913":105,"8101312":105,"8105e":109,"81119704":105,"8112e":109,"81192":105,"8121e":109,"8127704":105,"81348":105,"81354":105,"81366":105,"8137":105,"81372":105,"8141":105,"81424":105,"814560":109,"81458":105,"8147":105,"8150621709973537":107,"8156":105,"816599":105,"81688696":105,"81714":105,"8173e":109,"81768798828125":104,"81768799":104,"81785023":105,"8180281":105,"81816167":105,"81857":105,"8189e":[106,109],"81934":105,"81m":104,"82":[70,103,109],"8201":[105,106,109],"82016":105,"8213e":109,"82147":105,"8217e":106,"8224e":109,"8229":105,"82294524":105,"8234e":109,"8240589":105,"8240e":109,"8241":105,"82428":105,"8243384":105,"82465":105,"8248e":109,"82546026":105,"8261":105,"8262e":109,"8265212":105,"82654":105,"82689875":105,"8269":105,"827290":109,"8279644":105,"82819":105,"8283e":109,"8284e":109,"8288708":105,"82891417":105,"82898":105,"8293e":109,"8294e":109,"8295e":109,"82977":105,"82m":[104,105],"83":[70,104,105],"83106405":105,"83140534":105,"8317e":109,"831845":109,"8321578":105,"832633":109,"8339e":109,"83435":105,"8343e":109,"8346e":109,"83553046":105,"8356":105,"83655":105,"83666444":105,"83694":105,"83694774":105,"8370e":106,"8374":105,"83746":105,"8376e":109,"83797425":105,"8379e":109,"8384e":106,"8386477":105,"8388201":105,"83882904":105,"8390e":109,"83948356":105,"8396003":105,"84":[70,104,105],"84057474":105,"84071857":105,"8409559":105,"8414e":109,"8417":105,"8417e":106,"84183":105,"8423e":[106,109],"842469":109,"8434063":105,"84363":105,"8436978":105,"8439":105,"8445456":105,"8446":105,"8454e":106,"8461558":105,"84665924":105,"8467e":106,"84747":105,"84778":105,"8496e":109,"84995":105,"85":70,"850639":105,"8507803":105,"8509916":105,"8509e":109,"8510e":106,"8512343":105,"8518939":105,"85189575":105,"8525362":105,"85304886":105,"85324234":105,"85343812":104,"85375595":105,"85390973":105,"8547e":109,"8551e":109,"85558087":105,"8557e":106,"8559173":105,"85625255":105,"85666":105,"85703963":105,"8572e":109,"8575352":105,"85782766":105,"85786":105,"85938":105,"86":[70,105],"86021763":105,"8602e":109,"8608399":105,"8608e":106,"8609162":105,"8611444":105,"8616991":105,"8620718":105,"86227":105,"86304706":105,"86414":105,"8643":105,"864786":105,"86514":105,"8651793":105,"8653e":109,"8655":105,"86606":105,"86616135":105,"86683005":105,"8676515":105,"8684657":105,"8689e":109,"8693e":109,"86963":105,"8699e":106,"87":[70,105],"8703e":109,"87045":105,"87049603":105,"870526":109,"87054":105,"87125":105,"87129986":105,"87139845":105,"8714e":106,"87155":105,"87157995":105,"87177":105,"8718e":109,"8726":105,"87274":105,"87278444":105,"87323":105,"8733":105,"8735":105,"8743328":105,"8750135":105,"8751741":105,"87518":105,"87591183":105,"8761287":105,"87616193":105,"8762751":105,"8765895":105,"8768972":105,"877":109,"8772":105,"8775":105,"87766":105,"8776e":109,"8778":105,"87787":105,"8778759":105,"8782499999999999":107,"8785":105,"8788193":105,"8793":105,"88":[70,104,105,109],"880064":105,"88055":105,"88107586":105,"8814398":105,"8816869":105,"8817":105,"8818e":109,"8820076":105,"8822638":105,"8824023":105,"8825e":109,"8828186":105,"88284":105,"88294595":105,"8831e":109,"883279":105,"8834407":105,"88361436":105,"883681":105,"88368326":105,"8839093":105,"88438815":105,"8845e":109,"88504":105,"8853136":105,"8855596780776978":104,"886":105,"8863e":109,"88647026":105,"886791":105,"88775337":105,"8885":107,"88854957":105,"8892915":105,"8896506":105,"88998085":105,"89":[70,104,105],"89014":105,"8901e":109,"89034307":105,"89056396":105,"89063996":105,"8906e":106,"89081717":105,"89086473":105,"8910252":105,"89124495":105,"8912647":105,"89133":105,"8916":105,"8923e":109,"89285":105,"8932258":105,"89369905":105,"89383173":105,"8941e":109,"89442694":105,"89460224":105,"89484":105,"8954e":109,"89554":105,"8958":105,"89645":105,"89681005":105,"8968e":109,"8971772":105,"89723104":105,"8972577":105,"8973187":105,"8976605":105,"8979453":105,"89824754":105,"8984147543126978":107,"89858156":105,"8985e":109,"89860517":105,"89868015":105,"89886":105,"899":105,"89908":105,"8990e":109,"899237":109,"89928836":105,"8997":105,"8997571":105,"8997e":109,"89996":105,"89m":105,"8m":41,"8s":109,"9":[49,70,87,102,103,104,105,106,109,119],"90":[104,105,107],"9000":[46,50,109],"9001":[36,46,50,109],"90029544":105,"9007e":[106,109],"9010":11,"9010142093245521":107,"9011465":105,"90115":105,"9016738":105,"9018014":105,"90228":105,"90283":105,"9031525":105,"9033977":105,"9033e":109,"90368915":105,"9041e":109,"9042e":109,"9046e":109,"9048104":105,"90489644":105,"9048e":109,"90536344":105,"90552235":105,"9058e":109,"9067393":105,"90770215":105,"9081669":105,"9090e":109,"9092e":109,"9093149":105,"90th":[70,107],"91":[70,104,105],"9105379":105,"91081876":105,"9112e":109,"911333":105,"9113763":105,"9116395":105,"91168":105,"9121e":109,"9128e":106,"9131e":106,"9136e":109,"9138":105,"9141492":105,"91429555":105,"9143e":109,"91440874":105,"9149e":109,"9152e":109,"9153e":109,"91552734":104,"91552734375":104,"9155e":106,"9156e":109,"9157e":109,"91586":105,"91632":105,"9163557":105,"9167":105,"9169966":105,"9171e":106,"91743165":105,"9176925420761108":104,"918121":105,"9181e":109,"9189269":105,"91900206":105,"9191e":109,"91922":105,"9194614":105,"9197768":105,"91993105":105,"91998947":105,"91mfals":109,"92":[70,104,105],"9200019240379333":104,"9202296":105,"9207018":105,"9208387":105,"92098916":105,"92133":105,"9214447":105,"92154":105,"92188":105,"92344654":105,"923466":104,"9236755":105,"923717":104,"9238624":105,"9239":105,"9241573":105,"92460406":105,"925":[105,106],"9252076":105,"9255538":105,"9256e":109,"9257e":109,"9260271":105,"926044":104,"92618316":105,"9266":105,"9266098":105,"92688":105,"92693084":105,"9271338":105,"92730576":105,"92737":105,"9277581":105,"92798":105,"9286875":105,"9289e":106,"9293e":106,"9294e":106,"9295e":106,"92982674":105,"93":70,"9305":105,"9313457":105,"9313e":109,"9316707":105,"93173":105,"93176":105,"9318":105,"9318335":105,"931835":105,"93198806":105,"9319e":106,"93260366":105,"9328e":109,"9330e":106,"9332":105,"93347":105,"9335175":105,"93381244":105,"9344":105,"93462956":105,"93469065":105,"9352003":105,"9353e":109,"9354":105,"93542594":105,"9355017":105,"9358043670654297":104,"936":[107,108],"93606567":104,"9360656738281":104,"93695":105,"93707":105,"9372e":109,"9375356":105,"93761605":105,"93762034":105,"9381855":105,"9381e":109,"93841267":105,"93872":105,"939":[43,63,105,106,109],"9391487":105,"93915":105,"9391e":109,"9396e":109,"93973":105,"94":[70,104,105],"940083":109,"9400e":106,"940585":109,"94125":105,"94138":105,"9413e":106,"94214964":105,"9421855":105,"9422818":105,"9423e":109,"94244254":105,"9426196":105,"94275":105,"94284":105,"9439e":109,"9446315":105,"94498163":105,"9450149":105,"94503784":104,"9450378417969":104,"9452e":109,"94554764":105,"9460e":106,"94632":105,"94662":105,"9466e":106,"9470e":109,"94733":105,"94746464":105,"94770956":105,"9486e":106,"94920886":105,"9492276":105,"94934":105,"9497375488281":104,"94973755":104,"9499294":105,"94f1":109,"94k":[105,106],"95":[70,79,81,104,105,107],"9501":105,"9501e":106,"9504":105,"9505e":109,"95087":105,"9508924":105,"95218":105,"9522048":105,"9527973":105,"953":109,"95354944":105,"9537":105,"9539e":109,"95540506":105,"95544":105,"9554e":106,"95602":105,"95603913":105,"95605":105,"9561e":106,"9564e":109,"95667875":105,"95673":105,"9570074":105,"9572475":105,"9573141":105,"9573e":106,"9577e":109,"9581e":109,"9582136":105,"95822":105,"9582e":109,"9585e":106,"95904541":105,"9591119":105,"95961505":105,"9598888":105,"95th":[70,107],"96":[70,105],"9603e":109,"96054":105,"96090955":105,"9611e":106,"96142364":105,"9617e":109,"96207":105,"9620e":109,"9628e":109,"9629":105,"9633726":105,"96359974":105,"9639e":109,"96436":105,"96454924":105,"9653242":105,"96551245":105,"9657e":109,"9660537":105,"9668305":105,"96696764":105,"9671":105,"9674e":109,"96838":105,"9694":105,"9696":105,"9697e":109,"9698766":105,"9699e":109,"97":[70,87,105],"97024536":105,"9704e":109,"9705e":109,"9706679":105,"97103214":105,"9711":105,"97113":105,"9716":105,"9723e":109,"9726e":106,"97273135":105,"97276276":105,"97341204":105,"9735":105,"97403":105,"974096":105,"9743e":109,"97470093":105,"97501004":105,"97522974":105,"97578657":105,"9765e":109,"9771716":105,"9773e":109,"97743154":105,"97747803":105,"9776":105,"97779256":105,"97818244":105,"97897":105,"9789e":109,"97906494":105,"9790770411491394":104,"9790e":109,"9794":105,"97974":105,"9798533":105,"9799493":105,"98":[70,104],"98030627":105,"98035":105,"9804845":105,"9807":105,"981":104,"981288":105,"9814e":109,"9819":105,"9821":105,"982109":107,"982120":107,"9826e":106,"9830136":105,"9832e":109,"9833999":105,"98342896":105,"98374254":105,"9841e":109,"98425347":105,"9845507":105,"98462":105,"9847":105,"9849":105,"98565674":105,"9866e":109,"98671097":105,"9867e":106,"9871e":[106,109],"98755246":105,"9876475":105,"98822":105,"9882784":105,"98843":105,"9886":105,"98887885":105,"989133":104,"98916876":105,"98956":105,"98962784":105,"9896e":109,"9898":105,"98d1":109,"99":[70,104,105],"99023":105,"9910501":105,"9914025":105,"9914e":[106,109],"9915406":105,"99155337":105,"9916019":105,"9916197":105,"99162734":105,"991653":105,"99172956":105,"9917e":109,"9920751":105,"99216":105,"9929e":106,"9932013":105,"9932473":105,"9933716":105,"993386":105,"9933976":105,"99394447":105,"9940315":105,"9943127":105,"9944642":105,"9945123":105,"9946308":105,"99465877":105,"9947187":105,"99476504":105,"9948527":105,"99486005":105,"9949":105,"99496":105,"99503":105,"99526805":105,"9953135":105,"99564207":105,"99573624":105,"9959078":105,"99593395":105,"99603134":105,"99658":105,"9967":105,"9968e":109,"9973e":109,"99805":105,"9981":105,"9987816":105,"9988":105,"999999873689376e":109,"99m":109,"99mb":104,"99th":[70,107],"9m":106,"9s":[104,106,109],"\u03bb":102,"abstract":[13,34,71,73,81,82],"boolean":[13,27,73,84],"break":[8,79,81,89,100,102],"byte":[84,85,86],"case":[1,8,13,26,33,43,62,73,88,89,90,91,94,95,96,99,103,106,116,119],"catch":[13,73],"class":[4,7,10,11,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,59,60,61,62,63,65,66,67,68,71,72,73,74,75,76,79,81,82,83,85,86,105],"default":[3,7,9,10,12,13,14,15,16,17,18,19,22,23,24,25,26,27,28,29,30,33,34,39,41,42,43,44,48,50,51,53,61,62,70,71,73,78,85,86,89,91,92,93,95,96,98,102,104,106,107,108,109,117,119],"do":[0,10,13,38,42,44,45,48,50,61,73,74,81,92,94,95,97,102,103,104,105,106,107,108,119],"export":[12,13,28,47,73,90,91,92,96,100,102,103,112,113,119,120],"final":[1,38,40,43,45,48,55,65,95,102,104,113,116],"float":[3,6,7,10,13,15,33,35,37,41,43,44,45,47,48,49,50,53,54,56,59,60,62,63,64,67,68,69,70,71,73,74,76,77,79,80,81,82,83,84,85,104],"function":[3,6,7,9,13,14,15,24,26,33,35,38,43,46,48,51,55,57,62,65,67,69,70,71,72,73,74,76,77,78,79,81,82,83,86,89,92,93,95,98,105,106,107,112],"import":[1,9,13,15,18,66,70,76,88,92,93,95,97,100,103,104,105,106,107,108,109,113,114,118],"int":[3,4,6,7,8,10,11,13,15,23,24,33,34,35,36,37,38,39,40,41,42,43,44,45,47,48,49,50,53,54,55,56,59,60,62,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,81,83,84,85,86,105],"long":[1,50,55,72,90,104,105],"new":[1,4,10,13,14,24,26,32,33,38,40,50,55,65,67,68,73,79,81,82,85,89,90,91,92,94,95,96,97,98,100,101,102,106,107,110,112,113,115,116,118,119],"null":109,"public":0,"return":[3,4,6,7,9,10,11,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,31,32,33,35,37,38,39,40,41,42,43,44,45,46,50,51,52,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,103,104,105,106,107,108,109,110,115],"short":[49,50,76,86],"static":[10,13,33,34,81,85],"switch":[49,74,92,95],"true":[4,9,10,13,24,25,27,31,33,35,36,38,39,40,41,42,43,44,45,47,48,49,50,51,52,53,54,55,56,57,59,60,61,62,63,65,66,67,69,70,71,73,75,77,78,80,81,84,85,86,89,102,103,104,105,106,107,108,109],"try":[1,13,26,33,80,81,86,88,92,95,96,97,98,99,100,102,103,105,112,116,117,120],"var":[107,109],"while":[13,21,26,39,48,49,55,91,92,95,98,100,102,105,115,116,117],A:[10,12,13,15,24,27,31,33,36,38,39,40,42,43,45,46,47,48,50,51,53,54,55,56,57,59,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,81,82,83,84,85,86,89,92,96,100,101,104,106,107,110],AND:42,AS:42,And:[104,109],As:[1,74,88,92,94,95,106,109,110,116],At:[38,49,98],BE:42,BUT:[42,104],Being:0,But:[91,95,96,104,105,106,108],By:[9,33,98,105,119],FOR:42,For:[3,6,10,13,24,26,28,33,38,42,50,62,70,73,76,77,79,81,85,87,88,89,90,91,92,96,98,99,100,101,102,103,104,105,107,109,110,113,115,116,117,118,119,120],IN:42,IS:42,If:[1,6,7,9,10,12,13,15,17,18,24,25,26,27,29,32,33,38,39,40,41,42,43,44,45,47,48,49,50,51,52,53,54,55,56,59,61,62,63,65,66,67,68,69,70,71,73,75,76,77,78,79,80,81,82,84,85,86,88,89,90,91,92,94,95,96,97,98,99,100,101,102,103,104,105,106,108,109,110,111,116,117,118,119],In:[0,8,39,44,51,73,76,77,88,92,95,96,97,99,103,104,105,106,108,109,110,112,113,117,119],It:[10,13,33,36,43,46,47,70,73,76,80,82,89,90,91,92,95,97,101,104,105,107,116],NO:42,NOT:[13,42],No:[45,47,48,49,63,73,95,102,104,107,108],Not:[13,38,89,102],OF:42,OR:42,On:[88,92,102,118],One:[68,73,79,85,95],Or:[33,100,102,117],THE:42,THe:[13,27],TO:42,That:33,The:[0,1,4,7,8,9,10,11,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,31,32,33,35,37,38,39,40,42,43,44,45,47,48,49,50,51,52,53,54,55,56,59,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,85,86,88,89,90,91,92,94,96,97,98,100,102,103,104,105,106,107,109,110,111,113,114,115,116,117,119,120],Their:68,Then:[80,91,94,97,102,105,110,120],There:[13,89,91,92,95,98,103,116,119],These:[1,13,29,36,38,39,41,42,44,45,47,48,50,51,54,55,56,61,66,68,69,70,72,73,76,87,88,90,92,95,102,103,104,105,106,107,113,119],To:[13,14,26,27,29,50,54,56,62,66,70,73,77,85,89,92,94,95,96,98,100,102,104,105,106,107,109,110,114,115,117,118],WITH:42,Will:72,With:97,_1:50,_20190128_113421:109,_:[12,13,27,38,70,89,103,109],__call__:73,__getitem__:13,__init__:[52,105],__macosx:109,_courtship_label:109,_cvpr_:76,_european:77,_exampl:109,_hl:[33,85,86],_is_templ:85,_skeleton_idx:85,a5000:[105,106,109],ab:[43,107],abdo_index:103,abdo_loc:103,abdomen:[103,104,105,106,109],abid:1,abil:1,abl:[41,44,48,89,90,92,102,105,106,109,110,112,117,118],abort:13,about:[1,4,5,13,16,48,62,73,74,89,91,93,94,96,97,99,100,105,106,107,108,114,115,116,117,119],abov:[3,12,42,49,50,73,76,87,91,92,95,100,102,104,107,108],absolut:[31,33,37,47,49,50,62,66,76],abus:0,academ:101,acceler:[101,102,105,106,109,110,113],accept:[0,33,73,85],access:[10,11,13,27,33,49,55,76,89,90,92,96,100,102,104,110,112],accord:[43,63,68,106],accordingli:[3,13,49,110],account:[0,53,54,56,59,62,70,90,110,116],accumul:[74,76],accur:[88,101,113,115,117],accuraci:[47,48,70,81,106,107,108,113,115,117],achiev:[38,43,48,88,113,117],acquir:[107,109],acquisit:[100,105],across:[10,13,33,39,47,49,50,55,56,66,68,70,73,74,76,83,85,92,93,95,96,108,113,115,120],acryl:87,act:[0,1,36],action:[0,42,46,92],activ:[35,38,39,42,43,45,48,71,98,101,102,113,115,118],actor_id:109,actual:[13,33,51,68,75,81,97,104,105,106,107,108],ad:[1,10,13,38,39,44,49,62,63,67,69,73,74,90,92,94,95,97,107,109,115,116],adam:[49,109],adapt:[0,43],adaptor:[15,16,17,18,19,21,22,23,24,25,26,27,28,29,30],add:[1,3,4,10,13,14,26,40,44,48,49,61,62,68,74,85,92,94,97,102,115,116,118],add_edg:85,add_fram:34,add_inst:13,add_nod:85,add_subplot:103,add_suggest:13,add_symmetri:85,add_track:13,add_video:13,addelementwis:49,addit:[1,13,26,38,44,45,48,49,51,62,73,79,96,99,103,107,113,117],addition:[28,42,48,53,107],additivegaussiannois:49,address:[0,36,46,50],adher:1,adjac:[95,100],adjust:[31,43,46,47,49,53,54,60,64,67,73,76,110],adjust_matlab_index:31,adjust_prefix:40,admin:102,admonit:13,adopt:99,adult:87,advanc:[0,101,102],advantag:[42,48,73,88,105],advic:91,advis:103,affect:[50,92,100,102],affin:[10,48,49,56,65,71,73,76,88,92,96,97,117],after:[3,13,38,39,41,43,45,47,48,49,50,55,65,67,69,71,73,77,81,89,91,92,94,95,96,97,98,99,102,103,104,105,107,112,113,115,116,117,118],ag:0,again:[13,49,106],against:[13,81,86,89,95],agenc:100,aggreg:51,agnost:104,agreement:76,aim:1,akiaiwnjyax4csveh53a:109,al:[41,44,87,95,101,117],alejandro:77,alert:115,algorithm:[72,76,77,79,80,83,92,109,113,115],alia:[10,13,46,82],alias:[38,45,48],align:[0,10,54,67,73,77],align_inst:3,align_instance_point:3,align_instances_on_most_st:3,aligned_points_arrai:3,all:[0,1,3,4,9,10,11,13,14,15,16,17,18,19,22,23,24,25,26,27,28,29,30,32,33,38,39,42,45,47,48,49,50,51,54,56,59,61,62,65,66,67,68,70,73,74,75,76,78,79,80,81,82,85,87,88,90,92,94,95,96,97,99,100,101,102,103,104,105,106,107,108,110,113,115,116,117,118,119],all_ext:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],all_fram:[9,13],all_inst:[13,54,62],all_instance_confidence_map:54,all_item:4,all_label:[13,23,24],all_labels_adaptor:26,all_peaks_in_kei:60,all_points_arrai:3,alloc:[10,13,78,104,109],allow:[10,13,21,24,26,36,48,49,50,65,73,75,85,89,92,96,100,102,104,106,110,113,119],almost:[33,119],alon:48,along:[3,15,24,28,48,55,56,61,73,76,77,83,99,103,109],alphabet:[35,89,92,95],alphatrackeradaptor:15,alreadi:[1,8,10,13,21,31,33,47,50,63,68,69,73,78,80,81,85,89,90,91,92,95,96,97,98,102,103,106,119],also:[9,13,24,26,33,42,47,48,51,54,59,61,62,63,64,66,67,69,70,72,73,76,77,78,79,81,83,89,90,91,92,94,95,96,98,100,102,103,104,105,106,107,110,113,115,116,117,119],alt:[88,92,102,116,118],altern:[10,78,89,96,102,118],although:[9,10,26,48,94,95,96,102,115,117],altogeth:[49,74,104,108],alwai:[55,66,73,95,102,116],amen:76,among:113,amongst:[89,96],amount:[98,113],amsgrad:49,amz:109,an:[0,1,3,4,6,9,10,12,13,14,15,21,24,26,27,29,33,36,37,38,39,41,42,43,44,45,46,47,48,49,50,52,53,55,57,61,62,63,65,67,68,69,70,71,73,74,75,76,77,78,79,80,81,83,84,85,86,87,88,89,91,92,94,95,96,97,98,100,101,102,103,104,105,106,107,108,110,114,115,116,117,119],anaconda:[101,102,118],analog:10,analysi:[9,12,13,17,23,26,28,29,89,100,107,111,113,119,120],analysis_exampl:[103,104,105,106],analysis_video_0:89,anatom:70,anchor:[13,48,61,71,73,77,88,117],anchor_ind:61,anchor_part:[48,71,106,109],anchor_part_nam:61,angl:[3,49],ani:[0,1,4,6,9,10,13,21,22,26,28,30,33,35,36,38,42,43,46,47,48,49,50,51,55,56,60,61,62,63,65,66,67,68,69,70,73,74,75,76,80,81,82,85,86,89,90,92,94,95,96,97,99,100,101,102,103,107,108,112,113,116,119],anim:[10,15,41,48,59,85,87,88,89,91,92,93,95,99,100,103,107,109,110,112,113,115,116,117],annegret:[87,100,101],annot:[13,15,18,28,35,92],anonym:100,anoth:[3,10,13,33,50,74,80,85,89,90,92,95,96,97,98,106,107,117,119],answer:93,anyth:[13,50,89,90,102,112],anywher:110,ap:107,apart:76,api:[8,26,28,73,76,82,85,100,101,104,105,107,112],app:[101,102,110,118],appear:[0,10,87,89,92,94,98,116,119],append:[13,23,24,27,38,46,50,63,65,70,89,105,108],append_suggest:13,append_to_group:4,appendag:95,appl:[90,101,102],appli:[0,4,7,10,32,38,39,44,45,47,48,49,53,55,62,63,67,73,74,76,77,81,89,92,95,97,102,108,113,119],applic:[43,47,48,63,66,73,77,89,103,104,105,106,109,112],appoint:0,approach:[28,48,70,73,76,83,88,92,105,110,113,117],appropri:[0,14,15,16,18,19,22,23,24,25,26,27,28,29,30,33,38,47,65,70,73,80,82,92,97,98,104,105,106,107,109,116],approxim:[63,88,113],apt:[107,109],ar:[0,1,3,4,6,7,8,9,10,11,12,13,17,24,26,27,28,29,32,33,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,70,72,73,74,76,77,78,79,80,81,82,85,86,87,88,89,91,92,94,95,96,97,98,99,100,103,104,105,106,107,108,109,112,114,115,116,117,119],arang:[83,103],arbitrari:[13,33,50,57],arboresc:85,architectur:[47,48,50,51,62,75,91,96,98,101,102,104,109],architecture_graph:[50,109],archiv:[24,109],are_too_close_funct:80,area:[1,70,83,92,99],aren:[13,80,81,93,94,95,99],arena:87,arg:[12,13,16,18,19,22,23,24,25,26,29,30,33,35,48,73,74,81,82,107,109],argument:[9,12,13,24,33,35,43,45,46,61,62,66,67,73,77,80,89,96,109],ari:[100,101],aris:42,arithmet:49,arlo:101,arm64:102,arm:85,around:[13,15,35,43,47,48,49,59,64,71,73,75,77,86,92,98,103,104,105],arrai:[4,6,9,10,13,33,53,66,69,70,73,76,77,81,83,85,103,104,105],arrow:92,art:[42,101],articl:[38,100,101],arxiv:[38,40,43,107],as_format:26,as_nod:85,ascend:80,ask:[1,92,101,116,118],aspect:[47,66,103],assembl:[76,116],assess:115,asset:[33,109],assign:[1,10,13,48,66,68,72,73,76,80,81,83,92,95,119],assign_connections_to_inst:76,assigne:1,assist:[101,113,117,120],associ:[10,13,27,39,42,48,54,61,66,70,71,73,76,81,104,108,113],associt:73,assum:[9,13,33,38,61,63,73,84,85,86,92,102,107,110,117],assumpt:48,async:11,asynchron:50,asyncvideoserv:11,atrou:43,attach:109,attack:0,attempt:[13,25,33,36,50,81],attent:[0,88],attr:[33,52,85,104,107],attr_to_dtyp:86,attribut:[10,13,33,38,42,43,45,47,48,49,50,51,52,54,55,56,59,61,62,63,65,66,67,68,71,73,82,85,105],attrs_cl:52,aug:[104,105,106],augment:[49,88,91,96],augmentation_config:[49,106,109],augmentationconfig:[49,53],author:[28,42,100,101,110],auto:[33,50,89,92,103,109],automat:[10,12,13,25,26,42,47,49,50,55,73,76,90,98,102,110,113],autotun:55,avail:[0,1,9,33,34,38,49,62,68,70,73,78,88,89,90,96,102,103,104,105,106,107,108,109],available_video_ext:33,avc1:24,averag:[48,49,70,76,77,88,92,105,106,107],avg:[70,107],avi:[24,33,34,96,100,118],avoid:[28,54,81,116],avx2:[104,105,106,107,109],avx512_vnni:104,avx512f:[104,105,106,107,109],avx:[104,107],awai:13,await:[103,104,109],aws4:109,ax1:103,ax2:103,ax:[56,73,103],axi:[3,43,54,55,56,61,62,63,64,69,73,76,77,83,103,105],b990gxjt3d3j3jh:106,b9b0638744af3144affdc46668c749128bd6c4f23ca2a1313821c7bbcd54ccdd:109,b:[1,6,62,86,92,102],back:[4,10,33,36,55,62,76,80,81,89,92,100,110],backbon:[38,39,40,42,43,45,48,75,98,106,109,117],backbone_fn:43,backboneconfig:48,backend:[13,33,104,105,106],background:[11,87,95],backlight:87,backward:[46,85,92],bag:4,balanc:49,ban:0,bar:[73,89,116],bare:33,barekalmantrack:81,base:[4,7,10,13,14,21,33,34,36,38,39,40,41,43,44,45,46,47,48,50,53,63,65,70,71,73,76,77,79,81,82,84,89,92,93,94,95,98,99,100,101,102,113],base_checkpoint:[48,89,109],base_fram:10,base_label:[10,13],base_nod:13,base_port:11,base_run_nam:82,baselin:[50,91,109],baseline_medium_rf:[96,109],baseline_model:106,basemessageparticip:36,basenam:102,basetrack:81,bash:102,basic:[13,33,39,40,61,73,86,95,99,100,105],batch:[13,33,38,39,45,46,48,49,50,55,56,72,73,74,76,77,88,98,101,105],batch_norm:[38,48],batch_norm_before_activ:38,batch_siz:[13,33,49,55,73,74,89,105,109],batcher:55,batches_per_epoch:[49,109],bbox1:83,bbox2:83,bbox:[62,77,84],bear:44,becaus:[10,24,81,85,90,92,96,97,100,102,103,106],becom:62,bed:87,been:[1,8,10,13,36,75,78,80,81,89,92,95,102,105,115,116],beeswax:87,befor:[1,8,9,13,32,33,35,38,44,45,48,49,50,55,62,67,69,73,75,77,80,81,82,85,89,91,92,94,96,104,105,106,109,110,115,116],begin:[9,46,49,50,90],behavior:[0,1,33,52,87,92,101,113],being:[15,34,49,56,73,74,85,100,101],believ:1,belong:[4,10,13,62,76,77,108],below:[38,44,47,49,54,59,62,64,70,71,74,76,77,87,92,95,100,102,106,108,110,113],benchmark:[47,70,87,104,105,112],bendalab:28,benefici:[48,49,88],benefit:48,berman:87,best:[0,1,3,6,78,79,80,92,95,96,97,116,117],best_logical_device_nam:78,best_model:[50,73,89,96,104,106,107,109],better:[8,24,73,76,80,88,90,94,107,115,117],between:[3,6,7,9,10,12,13,35,36,43,47,48,56,61,67,68,70,73,76,79,81,83,85,89,92,95,97,107,116,119],beyond:[48,49,73,81,92],bgr:[33,43,47,63,104],bia:38,bias:43,biasadd:[106,109],bibtex:[100,101],big:95,bilinear:[38,39,40,41,44,45,48,104,107],bilinear_upsampl:40,bin:[102,105,109],binari:[59,104,105,106,107,109],binrang:107,biolog:101,bipartit:[79,81],bit:[99,108,117],black:[1,115],blob:108,block:[38,39,40,41,42,43,44,48,65],block_contract:[44,106,109],block_prefix:38,block_strid:48,block_v1:43,blog:[83,100],blow:104,blue:[33,88,98,115,116],bmp:24,bn:38,bodi:[0,3,48,49,53,61,70,71,73,74,85,95,97,113,116,117,118],bombu:87,bone:70,bookworm:[104,105,106,107,108,109],bool:[3,4,9,10,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,31,32,33,35,36,38,39,40,41,42,43,44,45,47,48,49,50,51,52,53,54,55,56,57,59,60,61,62,63,65,66,67,69,70,71,73,75,78,79,80,81,82,84,85,86],both:[0,1,13,15,24,39,47,48,50,54,56,62,67,68,69,70,73,80,88,91,94,95,96,106,117],bottleneck:[40,43,44,49,55],bottleneck_block:40,bottom:[47,48,67,73,88,89,91,92,97,101,104,110,117],bottomup:[73,96,110],bottomup_config:73,bottomup_lay:73,bottomup_model:73,bottomupinferencelay:73,bottomupinferencemodel:73,bottomupmodeltrain:82,bottomupmulticlassinferencelay:73,bottomupmulticlassinferencemodel:73,bottomupmulticlassmodeltrain:82,bottomupmulticlasspipelin:65,bottomupmulticlasspredictor:73,bottomuppipelin:65,bottomuppredictor:73,bound:[10,15,47,48,49,53,61,62,69,70,71,73,77,79,83,95],boundari:[55,71],bounding_box:10,bounds_error:103,box:[10,15,33,47,48,53,61,62,70,71,73,77,79,83,88,95],box_height:62,box_width:62,brain:101,branch:[1,38,48],brew:102,briefli:70,bright:[49,88,109],brightness_max_v:[49,109],brightness_min_v:[49,109],brisk:4,brisk_bag_of_featur:4,brisk_threshold:4,broadcast:[50,56,69],broadest:100,broadfield:49,browser:[50,109],bsd:101,bu:[104,105,106,109],buf:10,buffer:[49,55],buffer_s:55,bug:1,build:[9,13,33,34,37,38,40,42,52,61,62,63,65,67,73,75,101,102,103,105,106,109,113],builder:[65,82,106,109],built:[75,96,101,113,116],bumblebe:87,bundl:96,button:[1,91,98,109,110,116,117,118],byteord:10,bytes_:9,c57bl:87,c:[10,40,42,62,95,100,101,102,103,104,105,106,108],cach:[13,33,55,79],caff:[43,47,63],cage:87,calcul:[3,6,7,61,62,67,69,70,74,79,83,107],calculate_pairwise_cost:6,call:[4,13,14,15,16,17,18,19,22,23,24,25,26,27,28,29,30,32,33,43,46,55,70,73,74,77,79,81,86,91,96,105,106,110,113],callabl:[6,13,23,24,32,35,43,46,55,57,79,80,81,82],callback:[13,33,82],callback_funct:13,came:100,camera:[87,88,105],can:[1,3,4,7,10,11,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,31,33,35,38,39,42,43,45,47,48,49,50,51,53,54,55,56,59,62,64,65,66,69,70,73,74,76,77,78,80,81,85,87,88,89,90,91,92,93,94,95,96,97,98,99,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,119,120],can_read_fil:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],can_write_filenam:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],cancel:13,candid:[48,70,73,76,79,81,89,95,116],candidate_inst:79,candidate_mak:81,cannot:[13,25,26,29,33,43,48,78,85,86,97,102,104,105,106,107,108,109,110],canon:[43,44,45],cao:[76,117],cap:89,capabl:[65,105,106,109],capac:[41,43,44,48],captur:[73,105],care:[4,44,104,105,106,107,108],castro:[87,100,101],catalin:[87,100,101],categor:74,categori:1,cattr:[10,33,85],caus:95,cbmm:101,cc:[102,104,105,106,107,108,109],cd:102,cell:[104,105,106,108],center:[10,13,47,48,49,54,60,61,62,64,65,67,69,71,73,77,82,87,88,91,96,98,103,104,109,117],center_inst:[54,62],center_instance_ind:62,center_on_anchor_part:61,center_on_part:[47,61,106,109],center_pad:67,centered_inst:[48,89,91,104,105,106,107,109,110],centered_instance_id_model:105,centered_pair_smal:33,centered_patch:77,centeredinstanceconfmapshead:[65,71,73,106,109],centeredinstanceconfmapsheadconfig:[48,71,106],centroid:[10,28,47,48,54,59,60,61,62,63,65,67,71,73,77,79,82,88,89,91,95,96,98,104,105,109,110,117],centroid_confid:[60,62],centroid_confidence_kei:60,centroid_confidence_map:[54,60],centroid_config:73,centroid_confmap:73,centroid_confmap_head:65,centroid_crop:73,centroid_dist:79,centroid_kei:60,centroid_model:[73,105],centroid_model_path:73,centroid_v:[73,105],centroidconfmapshead:[65,71,73,109],centroidconfmapsmodeltrain:82,centroidconfmapspipelin:65,centroidcrop:73,centroidcropgroundtruth:73,centroidinferencemodel:73,centroids_kei:62,centroidsheadconfig:[48,71],certain:[81,85,95,98,99],certainli:33,cfg:[70,76,106],chain:[65,73],chanc:36,chang:[1,13,32,39,45,46,47,48,49,62,63,66,68,73,78,85,92,95,97,102,105,106,116,119],changed_on_load:13,channel:[10,33,35,38,39,40,42,43,45,47,48,54,59,62,63,66,67,71,72,73,74,75,76,77,84,96,104,118],channel_ind:74,channels_:33,channels_first:[33,43],channels_last:[33,43,96,104,109],channelwis:[38,74,77],charg:42,chdir:110,check:[1,10,13,15,29,32,33,36,50,52,65,73,78,81,85,86,89,92,95,100,101,102,104,109,110],check_messag:36,checkbox:95,checkerboard:[38,45,48],checkout:1,checkpoint:[46,48,50,82,89,104,109],checkpointingconfig:[50,82],child:10,children:97,chmod:102,choic:[79,80,117],choos:[88,92,94,95,96,97,102,114,116,117,119],chose:110,chosen:[95,110],chrome:109,chunk:[11,28,35],circular:87,circumst:0,citat:100,cite:[100,101],cl:86,claim:42,clarifi:0,class_ind:[59,72],class_map:[48,59,72,73],class_map_threshold:59,class_maps_head:65,class_maps_ind:73,class_maps_output_strid:73,class_prob:72,class_vector:[48,73],class_vectors_head:65,class_vectors_ind:73,classif:[48,71,72,73],classifi:[48,72,73,76,107],classification_model:42,classify_peaks_from_map:72,classify_peaks_from_vector:72,classmapgener:59,classmapshead:[65,71,73],classmapsheadconfig:[48,71],classmethod:[4,10,11,13,16,17,18,19,23,24,25,28,29,33,39,41,42,43,44,45,51,53,61,62,63,65,66,67,71,73,74,75,76,79,81,82,85],classvectorgener:59,classvectorshead:[65,71,73],classvectorsheadconfig:[48,71],clean:[79,81,89],clean_instance_count:[89,108,109],clean_iou_threshold:[89,108,109],cleaner:[8,81],cleanli:[10,13],cleanup:82,clear:[13,80,81,87,89,92,101],clear_edg:85,clear_suggest:13,cli:[8,12,73,82,89,100,102,105,106,108,112],click:[88,91,92,94,97,98,102,104,105,106,108,109,110,115,116,117,118,119],client:50,clip:[87,89,92,99,103,107,113,115],clipboard:92,clone:102,close:[1,11,21,33,34,48,80,102,105,112,116],closest:[6,7,73,83],cloth:70,cloud:117,cluster:[4,73,89],cm:[73,77,87],cm_output_strid:73,cmap:[84,103],cmd:[102,118],cmder:[102,118],cnn:[38,41,45],coco:[12,26,70,92],cocoapi:70,cocodataset:[16,70],cocoev:70,code:[15,26,41,70,74,85,90,92,95,100,102,103,110,119],codec:24,codecov:1,coeffici:[73,76,89],coher:48,col:[76,83],colab:[89,91,93,102,103,104,105,106,108,112,117],colab_demo:110,colaboratori:[90,112],collabor:101,collect:[13,33,50,74,100],color:[33,35,84,87,89,92,115,119],color_by_nod:84,color_by_track:84,color_manag:35,colormanag:35,column:[9,10,12,67,73,77,79,80,102,114],column_item:80,com:[38,79,83,102,103,104,105,106,107,108,109],comapar:33,combin:[13,28,48,55,56],come:[33,38,48,53,66,96,102,107],comma:89,command:[5,12,24,26,50,91,92,93,94,95,96,100,102,109,112,116,117,118,119],comment:[0,1],commerci:101,commit:[0,1],common:[21,26,33,34,38,39,40,42,43,45,57,65,76,85,92,95,100,105],common_runtim:[104,105,106,107,108,109],commonli:[47,63,77],commun:[0,1,2,11,100,101,102],compar:[6,13,33,85,86,89,107],compare_instance_list:6,comparison:[50,83,107,112],compat:[46,62,73,85,96,100,102],compens:[43,44,49],compet:80,compil:[104,105,106,107,109],complaint:0,complet:[10,13,15,50,76,85,104,113],complex:[41,44,48,100,102],complex_frame_merg:10,complex_merge_between:[10,13],complic:97,compon:[33,45,76,80,81],compos:[13,65],composit:[49,86],comprehens:93,compress:[24,34,50,100,102],comput:[10,13,33,38,45,47,48,49,50,54,56,61,62,63,67,69,70,73,74,76,77,79,81,82,83,90,91,101,102,103,104,105,106,107,108,109,110,112],compute_dist:70,compute_dist_metr:70,compute_distance_penalti:76,compute_generalized_voc_metr:70,compute_i:83,compute_instance_area:70,compute_ohkm_loss:74,compute_ok:70,compute_pck_metr:70,compute_visibility_conf:70,concaten:[38,45,48,55,65],concret:104,conda:[1,100,101,103,104,105,106,108,109,110,118],conda_dev:102,conda_jupyt:102,conda_prefix:102,condit:42,confer:77,confid:[10,48,54,56,59,62,64,65,71,72,73,76,77,82,84,88,89,92,96,104],confidence_map:[59,64],confidenti:0,config:[39,41,42,43,44,45,53,61,62,63,65,67,70,71,73,74,75,76,78,82,86,89,96,102,106,110],configur:[38,39,40,41,42,43,44,45,47,48,49,50,51,53,61,62,63,65,67,71,73,75,81,82,89,91,92,93,96,98,99,105,106,107,110,113,117],confirm:[32,36,92,102],confirm_callback:32,conflict:[10,13,38,39,50,94,102],confmap:[48,54,59,64,73,84],confmap_config:73,confmap_model:73,confmap_model_path:73,confmapconfig:71,confmaps_head:65,confmaps_in_kei:60,confmaps_ind:73,confmaps_kei:60,confmaps_out_kei:60,confmaps_strid:60,confus:70,conncet:76,connect:[8,36,37,38,39,41,42,43,44,45,48,71,73,75,76,79,85,89,90,92,95,96,97,103,104,109,113,118],connect_single_track_break:79,connected:76,consecut:[43,48,76],consequ:[100,113],consid:[0,49,70,73,74,76,81,85,89,100],consider:[47,48,49,50,55,80,86,88],consist:[45,48,97,110,116],consol:[73,89,102],constant:48,constitu:85,constraint:[63,67],constrast:49,construct:[0,10,13,31,38,43,75,105],constructor:[35,66,73],consum:50,consumpt:55,contact:[0,89],contain:[4,10,13,15,18,24,28,29,33,38,42,43,47,48,50,51,53,54,55,56,57,59,60,61,62,63,66,67,68,70,72,73,76,77,80,82,83,85,86,87,89,90,91,94,95,96,104,105,106,107,110,112],content:[13,105,109,110],context:[9,13,36,78,98],continu:[55,87,105,115,116,117,118],contract:42,contrast:[48,49,87,88,109],contrast_max_gamma:[49,109],contrast_min_gamma:[49,109],contrib:[104,105,106,107,108,109,110],contribut:[0,59,88,100],control:[8,34,45,48,50,76,84,88,91,92,98,109,110],controller_address:[50,109],controller_polling_timeout:[50,109],conv:[38,39,43,44,45,48],conv_batchnorm:48,conv_block:[38,44],conv_shortcut:43,conveni:[13,39,51,56,62,65,66,67,76,77,78,84,85,102,104,105],convent:89,converg:49,convers:[1,33,89],convert:[1,4,10,13,21,26,33,43,47,55,62,63,66,67,69,73,76,85,86,92,94,100,104,105,114,115],convert_image_dtyp:63,convert_rang:33,convert_rgb_to_bgr:63,convolut:[38,39,40,41,42,43,44,45,48],convolv:77,convs_per_block:[44,106,109],cooldown:81,coor:15,coordin:[9,10,13,15,21,48,49,53,54,56,60,62,64,66,70,72,73,76,77,83,103,104],copi:[10,13,15,36,42,48,49,66,73,82,83,85,86,90,92,94,96,103,110,112,116],copyright:42,core:[73,104,105,106,107,108,109,112],corner:[15,83],corr:103,corr_rol:103,correct:[0,13,26,70,71,76,92,93,94,95,97,102,107,113,115,117,119],correctli:[49,50,102,105,106,112,117],correl:103,correspond:[3,4,6,7,9,10,13,32,36,38,54,56,62,63,64,70,72,73,76,77,81,83,85,88,92,95,104,108],cost:[6,39,41,44,47,48,49,55,62,73,79,80,81,83,88,95,104,107],cost_funct:6,cost_matrix:[79,80,83],could:[0,13,14,33,70,91,102,104,105,107,108,109],count:[13,33,45,48,49,73,76,79,80,81,89,95,103,105],counter:50,courtship:[87,103,109,111],courtship_label:109,cov_vel:103,covari:[54,77],cover:[1,48,91,95,100],coverag:1,covert:[13,27],cpu:[57,73,78,86,89,102,104,105,106,107,109],cpu_feature_guard:[104,105,106,107,109],cpu_instruction_set:[104,107],cpuset:86,crash:100,creat:[0,1,4,9,10,11,13,14,15,18,24,26,27,33,38,39,40,41,42,43,44,45,46,47,50,51,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,71,73,75,76,78,79,81,82,84,85,86,89,92,93,94,96,97,101,102,103,104,105,106,108,109,113,120],create_trainer_using_cli:82,creation:[38,43,51],credenti:109,credit:87,crf:[34,100],criteria:[13,47,55,67,92,107],critic:[0,48,73,104,105,106,107,109],crop:[13,35,47,48,49,53,60,61,62,66,67,73,77,88,89,92,98,104],crop_and_res:[62,77],crop_bbox:[53,62,77],crop_height:[53,62,77],crop_offset:73,crop_output:73,crop_sample_ind:72,crop_siz:[47,62,73,77,109],crop_size_detection_pad:[47,109],crop_size_xi:35,crop_width:[53,62,77],cropandres:[104,107],cropper:73,cross:[95,96],crossentropi:74,csv:[9,12,18,50,92,96,107,109],csvadaptor:17,ctrl:[102,118],cuda:[78,102,104,105,106,107,108,109],cuda_bla:[105,106,109],cuda_dnn:[105,106,109],cuda_gpu_executor:[104,105,106,107,108,109],cudart:[104,107,108],cudart_stub:[104,107,108],cudatoolkit:102,cudnn:[100,102,105,106,109],cull:[8,79,80,81,89,95],cull_frame_inst:79,cull_funct:81,cull_inst:79,curl:[102,105,106],current:[4,13,15,24,27,33,38,43,45,46,48,50,54,56,75,78,80,81,82,85,86,88,90,92,95,98,101,102,105,106,110,116,118],current_groupset:4,current_strid:[38,39,43,45],curv:107,custom:[26,73,74,79,92,93,96,101,104,105,106,112],customiz:[101,112],cv2:[104,107,108],cycl:[92,95,117],d0:74,d:[4,62,73,83,85,86,87,90,92,93,94,95,100,101,102,104,105,106,107,108,109],d_0:56,d_n:56,dai:87,damag:42,dan:[87,100,101],dark:115,dashboard:50,data:[4,5,6,7,9,10,11,13,15,16,17,18,21,24,25,26,27,28,29,32,33,35,36,43,46,48,49,50,51,70,73,81,82,85,86,87,88,89,91,92,93,94,96,97,98,99,107,113,119,120],data_config:65,data_dict:9,data_path:[31,89,109],data_provid:[65,73],data_read:82,data_structur:[104,105,106],dataconfig:[47,51,65],datai:103,dataread:82,dataset:[6,7,9,12,15,16,17,18,19,20,23,24,26,27,28,29,31,33,35,46,47,49,53,54,55,56,57,59,60,61,62,63,65,66,67,68,70,73,82,85,91,92,96,101,103,104,105,106,107,111,113,114,117,118,120],dataset_h5:33,dataset_op:[53,54,56,57,59,60,61,62,63,65,66,67,68,73],datasetv2:[53,54,55,56,57,59,60,61,62,63,65,66,67,68,73],datax:103,date:[82,96,109],datetim:[13,27],david:[100,101],de:102,deactiv:102,deal:42,debian:[104,105,106,107,108,109],dec:38,decid:[91,92,94],decod:[38,39,41,42,44,45,48,51,85,86,103],decode_preview_imag:86,decoder_batchnorm:48,decoder_features_strid:38,decoder_filt:[42,48],decoder_filters_r:48,decoder_stack:[38,39,41,44],decoderblock:[38,39],deconv:40,deconv_filt:40,deconv_modul:40,deconvolut:40,decor:52,decoupl:51,decreas:[40,41,44,48,49,55,81,88],deem:0,deep:[13,15,41,85,86,100,101,104,105,106,107,109,113],deeplabcut:[12,92,100],deeplabv2:43,deepposekit:[12,92],def:[103,105],default_ext:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],default_suffix:13,defin:[0,10,15,28,38,39,43,44,45,48,53,54,56,59,65,69,71,73,75,76,85,96,97,103,104,105,106,113],definit:[28,39,43,71],deflat:109,degre:49,delai:50,delet:[13,50,81,82,85,92,94,104,106,109],delete_edg:85,delete_nod:85,delete_suggest:13,delete_symmetri:85,delete_viz_imag:[50,109],delimit:40,deliv:101,delta:77,demand:90,demo:110,demonstr:[39,107,112],deng:77,deni:[107,109],denot:[9,10,13,42,62,70,72,73,76,83,107],dens:[38,55,92],densenet121:[42,48],densenet169:[42,48],densenet201:[42,48],densenet:[38,47,63],densiti:[48,77],dep:[105,106,109,110],depend:[34,45,50,54,59,69,88,90,92,96,100,102,104,105,106,107,108,110,113],deprec:[8,24],depth:86,dequ:81,deriv:103,derogatori:0,descend:[49,70],describ:[12,13,38,39,43,44,50,65,73,75,76,77,82,93,97,114,120],describe_tensor:[69,73],descript:[13,27,48,51,65,69,70,85,87,104,106,109],descriptor:4,deselect:92,deseri:[4,13,14,15,16,18,19,22,23,24,25,27,28,29,30,33,51,85],design:[51,93],desir:[4,13,39,41,43,44,45,48,50,53,61,62,63,67,81,91,95,108,113,117],dest_nam:85,destin:[56,71,76,85,95,97,102],destination_nod:[48,71],detail:[0,7,13,16,38,39,41,44,70,73,76,88,89,91,95,100,113,115],detect:[10,12,13,16,26,33,47,48,49,50,73,76,81,84,88,89,92,97,102,104,107,108],detect_video:13,detector:73,determin:[0,6,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,33,35,39,41,44,45,47,48,50,66,67,70,71,73,79,80,81,92,95,98,109,116],determinist:48,deutsch:[100,101],dev:[3,102],dev_requir:1,develop:[1,51,100,101,102,105,112],deviat:[3,47,49,54,56,59,63,70,77],devic:[57,78,102,104,105,106,107,108,109],device_ind:78,device_nam:[57,60],device_typ:102,diagnosi:70,diagnost:102,diagon:54,diagram:62,dialog:[14,15,35,88,89,91,98,110,117,119],dic:86,dict:[4,7,9,10,13,15,22,24,31,36,46,51,55,57,65,68,69,70,73,76,79,80,81,83,85,86,107],dict_a:13,dict_b:13,dict_cut:86,dictionari:[4,7,9,10,13,15,21,24,26,36,43,50,51,55,57,58,65,66,68,69,70,73,79,81,83,85,86,107],did:89,didn:[36,100,103,109,116],diez:[87,100,101],differ:[0,12,13,15,26,33,38,42,43,44,45,47,49,61,64,66,70,73,76,81,83,85,88,91,92,93,95,96,99,102,103,104,105,106,107,110,116,119],differenti:103,difficult:[48,49,95],difficulti:[70,107],dilat:[43,48],dilation_r:43,dim:[103,104,107],dimens:[47,48,54,56,63,66,67,69,70,71,73,76,83,103],direct:[13,45,48,56,71,76,77,92],directli:[33,45,47,48,55,73,76,82,89,92,102,103,105,107,110,118],director:101,directori:[13,15,26,33,46,50,51,73,86,89,90,92,96,102,104,106,107,108,109,110],directoryimgstor:33,direntri:86,disabl:[0,73,78,109],disable_gpu_prealloc:[73,78],disable_prealloc:[78,104,105],disambigu:48,discard:[73,76],discover:78,discuss:[2,95,100,101,102],dish:87,disk:[13,33,46,50,51],dispatch:[14,15,16,17,18,19,21,22,23,24,25,27,28,29,30],displac:[7,76,81],displai:[50,73,89,103,108,115],disposit:109,dist:[70,107],dist_arrai:6,dist_penalty_weight:[73,76,89,109],distanc:[3,6,7,48,56,70,73,76,79,83,89,95,107],distance_to_edg:56,distinct:[26,48,80,81,88,90,92,95,96,103,111,116,117],distinctly_color:[35,89],distinguish:[21,87],distribut:[42,48,49,54,56,59,69,77,92,102,107],dists_dict:70,divid:[54,63,64,73,76,77],divis:[38,43,47,62,67,73],dl:106,dlc:45,dlerror:[102,104,107,108],dload:[105,106],dlopen:[104,107,108],dm:109,dn:74,doc:[7,28,101,102,103],dock:92,docstr:[1,39,41,44],document:[1,42,43,51,76,89,108],doe:[13,26,42,44,46,48,50,53,54,56,59,61,62,65,68,85,86,89,104,107,108,115],does_match_ext:[14,15],does_read:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],does_writ:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],doesn:[1,3,10,13,33,55,86,91,92,96,97,98,115],don:[10,12,29,33,74,86,89,90,91,99,100,102,104,105,106,109,110,112,117],done:[35,55,61,62,68,73,81,90,112,117,119],dot:[76,84],doubl:[38,45,48,92,94,97,115,116,118,119],down:[1,41,44,48,49,50,72,73,76,88,89,91,95,96,100,101,104,105,109,116,117,118],down_block:[39,40,41,42,43,44,106,109],down_convs_per_block:41,downgrad:102,download:[87,89,90,100,102,103,104,105,106,107,108,110,111,112],downsampl:[39,40,42,43,44,48,67,88],downsampling_lay:40,downsamplingblock:39,downsid:95,downstream:[55,66],dpi:[84,105,107],dpkg:[107,109],draft:1,drag:[88,92,98,116,118],draw:[35,89],drawn:[4,88,92,95],drive:[90,92,96,103,106,108],driver:102,drop:[10,36,55,65,95,117,118],drop_old:57,drop_remaind:55,dropbox:[90,106],dropdown:119,drosophila:[87,103,109,111],drwxrwxr:[104,105,106],ds:73,ds_index:66,ds_input:[55,61,62,63,67,68],ds_store:109,dset_nam:103,dso_load:[102,104,107,108],dst_node:[76,85],dst_node_ind:[76,85],dst_peak_ind:76,dt:105,dt_float:[104,107],dt_int32:[104,107],dt_uint8:[104,107],dtype:[10,33,43,53,54,56,59,61,62,63,64,66,67,69,72,76,77,86,104,105,106,107,109],due:[76,104],dummi:[33,73],dummyvideo:33,duplic:[24,85,86],dure:[13,33,38,46,47,48,49,50,51,55,65,71,73,74,76,89,91,95,98,102,104,105,107,109,110,115,117],dx_i:77,dy_i:77,dynam:[20,73,74,102,104,107,108],e:[0,1,3,6,10,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,33,37,38,43,45,48,49,50,53,54,55,56,59,61,62,66,67,68,69,70,72,73,74,77,78,79,80,81,85,86,88,89,91,92,93,94,95,96,101,102,104,108,109,110,113,118],each:[3,4,6,7,10,13,24,27,28,32,33,36,38,39,40,41,42,43,44,45,46,48,49,50,53,54,55,56,57,59,61,62,64,65,66,68,70,71,72,73,74,76,77,78,79,81,83,85,88,89,90,91,92,95,96,97,98,102,103,104,105,108,109,110,113,117],eager:[74,78],earli:[36,47,49,50,106,109,110],earlier:[7,81],early_stop:[49,109,110],earlystoppingconfig:[49,106,109],easi:[1,4,48,49,70,74,75,78,91,95,96,97,101,102,104,107],easier:[8,48,51,70,74,88,95,102,113],easiest:[70,90,96,114],easili:[51,89,102],east:109,echo:102,eclos:87,edg:[1,3,9,35,48,49,56,71,76,84,85,89,92,95,97,104,118],edge_destin:56,edge_ind:[9,12,56,73,76,85],edge_inds_sampl:76,edge_is_wedg:[35,89],edge_nam:[9,12,85],edge_peak_ind:[73,76],edge_peak_inds_sampl:76,edge_sourc:56,edge_to_index:85,edge_typ:76,edgeconnect:76,edges_ful:85,edgetyp:[76,85],edit:[0,51,91,92,94,95,96,102,106,112,115,119],editor:[91,96,106],edna:[100,101],edu:[0,1,99,100,101],educ:90,effect:[33,37,38,39,40,41,43,44,45,47,48,49,54,55,56,59,61,63,73,98,102,119],effici:[49,54,78],efficientnetb0:[42,48],efficientnetb1:[42,48],efficientnetb2:[42,48],efficientnetb3:[42,48],efficientnetb4:[42,48],efficientnetb5:[42,48],efficientnetb6:[42,48],efficientnetb7:[42,48],eg:1,eigen:[104,107],either:[1,9,10,18,33,36,50,54,59,62,67,68,70,73,77,85,89,92,94,95,96,102,116],elaps:[35,81,89],electron:0,element:[10,13,43,45,48,53,54,55,56,57,59,60,61,62,63,66,67,68,73,76,77,86],elementwis:54,eleni:[87,100,101],elimin:55,els:[13,76,110],elsewher:[90,92],email:99,embed:[13,33,39,87,104],embed_all_label:[13,104],embed_suggest:13,embedded_frame_ind:33,empathi:0,empti:[9,13,33,40,66,89,116,118],en:76,enabl:[4,33,42,48,49,50,54,55,73,78,89,95,100,102,104,105,106,107,109,112,113],enable_gpu_prealloc:78,enable_prealloc:78,enable_source_video:33,enc:38,encapsul:[73,79],enclos:87,encod:[33,38,39,41,42,44,45,48,51,77,84,85,86,89,92,95,100],encoder_decod:[39,41,44],encoder_features_strid:38,encoder_stack:[38,39,41,44],encoderblock:[38,39],encoderdecod:38,encourag:[1,48,49],end:[9,13,38,39,44,46,49,50,51,55,63,65,86,92,95,100,103,113],endpoint:50,enforc:[73,76,89],engag:87,engin:[43,60,65,73,75,82,83],enhanc:1,enough:[81,88,93,95,99,117],ensur:[4,13,24,44,47,50,52,62,67,68,73,76,78,80,87,96,100,102,119],ensure_float:[63,73],ensure_grayscal:[47,63,73,109],ensure_int:63,ensure_list:69,ensure_min_image_rank:63,ensure_rgb:[47,63,109],entail:77,enter:[102,117,118],entir:[4,9,13,27,33,38,49,54,66,77,89,90,92,94,96,105,106,107,115,116,120],entireti:92,entiti:[10,28],entrypoint:[12,73],enumer:103,env:[1,86,102,104,107,108,109],environ:[0,78,87,101,102,103,104,107,118],environment_mac:102,environment_no_cuda:102,epoch:[46,49,50,55,88,98,106,109,110,117],equal:[33,39,40,41,44,45,49,55,62,67,84,89],equival:[26,37,38,55,62,69,77],error:[1,33,52,70,76,78,98,100,102,104,105,106,107,108,109,113,115],escap:92,especi:[15,18,48,49,50,55,62,66,88,92],essenti:97,establish:85,estim:[41,48,70,76,77,81,85,107,109,112,113,117],et:[41,44,87,95,101,117],eta:109,etc:[1,4,13,33,45,86,102,104,117],ethnic:0,ethogram:103,euclidean:[70,76,83],eval:[102,106,107,109],evalu:[48,49,62,65,70,73,76,77,82,92,106,109],evaluate_model:70,even:[13,38,47,48,49,54,56,61,62,68,74,79,92,95,102,116],evenli:[73,92,116],event:[0,42,46],ever:92,everi:[1,3,6,13,27,38,39,43,45,48,49,50,55,70,89,92,95],every_epoch:[50,109],everyon:[0,1],everyth:[10,81,92,117],ex:[13,27,73],exact:[50,69,100,107],exactli:[43,62,70,77,81,88,89,95,100,103],exampl:[0,13,21,22,26,30,33,38,45,46,47,49,50,53,54,55,57,58,61,62,65,66,67,68,69,73,74,76,77,82,83,85,87,88,90,91,92,96,100,104,106,110,111,113,117],example_gt:73,example_ind:66,example_indic:[66,68],except:[10,13,26,28,38,55,73,74,97,101,102],exclud:[24,48,76,79,81,86],exclude_nod:86,exclude_non:86,exclus:13,execut:[49,74,76,82],exist:[1,13,26,27,32,33,38,39,46,50,53,85,86,92,96,97,102,106,108,112],exit:89,expand:[43,47,63,69],expand_dim:105,expand_to_rank:[63,69],expans:[43,55],expansion_r:40,expect:[0,1,9,12,13,50,53,54,55,56,57,59,60,61,62,63,67,68,70,71,73,76,77,89,103,114],experi:[0,1,48,87,100,105,108,112,113],experiment:[55,95],expir:109,explain:[91,95,96],explan:[88,114],explicit:[0,62,67,78],explicitli:[47,49,50,51],explor:[104,105,106,108],export_cli:73,export_model:[73,89],export_nwb:13,export_path:89,exportanalysisfil:26,exported_model:[73,89],express:[0,42],ext:26,extend:[4,87],extend_from:13,extend_group_item:4,extend_ownership:4,extens:[12,13,14,15,16,17,18,19,22,23,24,25,26,27,28,29,30,33],extern:[13,50,70,92,100,102],extra:[4,47,48,102],extract:[13,38,39,42,73,86,103],extran:73,extrapolation_valu:[104,107],ey:70,eyel:[103,104,105,106,109],eyer:[103,104,105,106,109],f8:10,f:[89,95,102,103,104,105,107,114],face:[0,15,51,99],facecolor:[105,107],facilit:[33,76],fact:95,factor:[34,40,41,43,44,45,47,48,49,54,62,64,66,67,70,73,74,81,84,107],fail:[1,49,50,55,70,89,95,100,102,109,110,119],fair:0,fairli:[3,87,91],faith:0,fake:33,falkner:[87,100,101],fall:1,fals:[3,9,10,13,16,21,23,24,27,33,35,36,38,40,41,42,43,44,45,47,48,49,50,52,54,55,56,59,61,62,63,65,66,67,69,70,73,75,79,81,82,84,85,86,89,92,102,103,104,105,106,107,109],false_neg:70,familiar:94,far:[13,28,76,93,95,99,107],fashion:48,fast:[41,43,81,101,104,105],faster:[13,41,44,73,79,89,110,115],fc:38,featur:[4,37,38,39,42,43,45,47,48,70,71,79,88,92,100,113,116],features_output_strid:43,featuresuggestionpipelin:4,feautur:1,feedback:115,feel:[1,99,100,116],femal:[87,105,116],fetch:[11,65],few:[48,49,74,88,90,92,98,101,102,104,113,119],fewer:[38,45,48,55,68,76,81,97],ffmpeg:[33,34,100],field:[10,39,41,43,44,48,50,52,56,65,67,71,73,75,76,87,88,92,96,97,98,109,110,117],fifth:15,fig:103,figsiz:[103,105,107],figur:[10,46,82,84,103,105,107],file:[1,5,9,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,42,43,47,48,50,51,66,70,73,82,85,86,87,90,91,94,95,96,100,102,103,104,105,106,107,108,109,110,111,112,113,114,118],file_h5:33,filehandl:[14,15,16,17,18,19,22,23,24,25,27,28,29,30],filenam:[8,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,32,33,34,35,46,51,66,70,85,86,89,90,91,96,102,103,104,105,109,110],filename_a:86,filename_b:86,filenames_prefix_chang:32,filenotfounderror:[33,86],filepath:[46,47],filetyp:13,fill:[55,73,105,116],fill_miss:[10,103],fill_valu:103,filter:[13,38,39,40,41,42,43,44,45,48,55,57,66,73,76,80,81,89,103,106,108,109],filter_connection_candid:76,filter_fn:55,filter_increas:[39,48],filters_increas:48,filters_r:[41,44,48,106,109],final_model:[50,96,109],final_pass:81,find:[3,6,10,13,15,18,25,32,33,45,48,49,57,60,61,70,73,76,77,79,80,85,89,90,92,95,101,105,110,117,119],find_changed_subpath:32,find_fancy_frame_idx:13,find_files_by_suffix:86,find_first:13,find_fram:13,find_frame_pair:70,find_global_peak:77,find_global_peaks_integr:77,find_global_peaks_rough:77,find_global_peaks_with_offset:77,find_head:73,find_instance_crop_s:62,find_last:13,find_local_peak:77,find_local_peaks_integr:77,find_local_peaks_rough:77,find_local_peaks_with_offset:77,find_neighbor:85,find_nod:85,find_offsets_local_direct:77,find_padding_for_strid:67,find_path_using_path:13,find_peak:73,find_points_bbox_midpoint:61,find_suggest:13,find_track_occup:13,find_unique_nod:85,finder:76,findinstancepeak:73,findinstancepeaksgroundtruth:73,fine:[39,43,49,74,88,116],finer:[49,76,81],finish:[13,50,89,94,96,98,102,106,109,110,115],finish_complex_merg:13,finish_timestamp:109,first:[6,7,9,12,13,15,17,29,33,35,38,39,41,43,44,45,49,53,54,55,56,62,63,65,66,69,73,78,79,80,81,85,86,89,91,92,94,95,97,102,103,104,105,106,107,108,109,110,113,114,115,117,118,119,120],first_choice_match:79,first_conv_strid:39,first_gpu:109,first_inference_tim:105,fit:[3,42,92,103],fit_predict:103,fit_track:8,five:[35,89,92,95],fix:[1,3,31,33,43,46,47,48,50,63,67,69,86,95,120],fix_rel_path:31,fixup_path:33,flag:[36,84,89,104,105,106,107,109],flat_dict:86,flatnonzero:103,flatten:[4,54,56,69,71,103,107],flatten_channel:56,flatten_offset:54,flexibl:[38,101,105],fli:[87,103],flies13:[104,105,107],flip:[47,49,53,63],flip_horizont:[49,109],flip_instances_lr:53,flip_instances_ud:53,float32:[10,43,53,54,56,59,61,62,63,64,66,69,72,73,76,77,105,106,109],float64:[77,86,105],floor:87,flow:[73,81,89,95,96,108],flow_shift_inst:81,flowcandidatemak:81,flowmaxtrack:[81,89],flowmaxtrackscandidatemak:81,flowtrack:81,fly:87,fly_clip:[105,108],fly_id:103,fly_node_loc:103,fly_node_veloc:103,fma:[104,105,106,107,109],fn:[13,70,107],focu:[48,49,74],focus:0,folder:[1,13,33,46,48,50,51,70,73,82,89,90,91,102,104,105,106,107,109,118],follow:[0,1,13,28,33,38,42,43,48,55,62,70,85,87,97,99,100,101,102,103,104,105,106,107,108,109],font_scal:103,foo:[26,86],footag:113,for_object:26,forc:[24,92,105,106,109,110],forelegl4:[103,104,105,106,109],forelegr4:[103,104,105,106,109],forg:[101,102,104,105,106,108],forget:[105,106],fork:1,form:[10,13,38,39,43,44,45,48,50,53,55,56,57,65,68,70,71,73,76,85,97,105,108],format:[9,10,12,13,31,33,43,51,62,63,65,73,85,86,89,91,92,95,96,104,105],format_id:21,formatted_ext_opt:[14,15],former:73,forum:100,forward:[73,92,103],forward_pass:73,foster:0,found:[3,13,25,26,33,38,55,70,73,76,81,85,86,95,98,102,103,105,106,109,110,117],fourth:[15,92],fp:[33,34,35,70,87,89,101,107,109],fraction:[13,47,68,73,76,89,106,109],fragment:80,frame:[4,6,7,9,10,11,12,13,15,17,18,24,29,33,34,35,47,48,56,58,62,66,68,70,73,76,79,80,81,84,87,88,89,90,91,92,93,94,95,96,97,98,99,100,103,104,105,106,108,109,112,113,115,117,120],frame_count:[103,105],frame_data_format:[23,24],frame_gt:70,frame_idx:[4,10,11,13,70,104,105],frame_ind:[58,66],frame_index:[13,33],frame_indic:[13,33],frame_list:86,frame_numb:33,frame_pair:70,frame_pr:70,frame_predict:105,frame_rang:[6,13],frame_str:86,framegroupset:4,frameitem:4,framematch:[79,81],framer:34,frames_per_chunk:11,frames_record:105,framework:[37,38,39,42,43,44,45,53,54,55,56,57,59,61,62,63,64,65,67,69,71,72,73,74,76,77,82,83,101],free:[0,1,42,90,99,100,102,109,116],freeli:87,freez:[50,73],frequenc:[50,104,107],frequent:87,fresh:[36,102],friend:100,from:[0,1,2,3,4,6,7,9,10,11,13,14,15,17,18,21,24,26,27,28,29,31,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,53,54,55,59,61,62,63,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,91,92,93,94,95,96,98,100,101,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119],from_arrai:10,from_block:65,from_candidate_inst:79,from_config:[39,41,42,43,44,45,53,61,62,63,67,71,74,75,76,82,106],from_cost_matrix:79,from_dict:85,from_filenam:[33,66],from_filepath:66,from_frame_idx:13,from_hdf5:33,from_image_filenam:33,from_inst:10,from_json:[51,85],from_json_data:24,from_json_dict:51,from_label:82,from_media:33,from_model_path:73,from_nam:85,from_names_and_edge_ind:85,from_numpi:[10,33],from_pipelin:65,from_point:10,from_pointsarrai:10,from_predict:10,from_skeleton:53,from_trained_model:73,from_unlabeled_suggest:66,from_user_inst:66,from_user_labeled_fram:66,from_video:11,fron:110,front:13,frontend:[107,109],frozen:[43,48,73,89],frozen_graph:73,fruit:87,fssl:102,full:[10,11,13,15,18,26,29,35,43,47,48,55,62,65,66,67,69,73,76,78,86,87,90,96,98,101,103],full_imag:67,full_image_height:62,full_image_kei:67,full_image_width:62,full_video:[15,18],fulli:[38,42,43,44,48,71],func:57,functon:43,fund:[100,101],furnish:42,further:[0,49,88,113],fuse:[38,45,48],futur:[46,96,115],g:[1,6,10,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,33,37,45,48,50,53,55,62,66,67,70,72,73,74,78,79,85,86,89,90,92,95,96,101,102,103,104,108,113],gain:49,gamma:49,gammacontrast:49,gap:81,gather:61,gather_nd:76,gaussian:[48,54,56,59,69,77],gaussian_nois:[49,109],gaussian_noise_mean:[49,109],gaussian_noise_stddev:[49,109],gaussian_pdf:69,gender:0,gener:[1,4,9,11,13,22,27,28,30,35,38,39,41,43,44,45,47,48,49,50,51,53,54,55,56,59,62,64,65,66,68,69,70,73,74,76,77,79,81,82,83,84,88,89,92,94,95,96,97,100,101,102,104,105,106,107,108,109,110,112,113,115,116],generaliz:38,generate_skeleton_preview_imag:84,genericjsonadaptor:[22,26],genuineintel:[104,107],geometr:49,geometri:48,get:[1,2,3,4,7,9,10,13,17,29,33,35,36,38,42,45,62,65,76,78,79,85,86,88,89,90,92,94,95,96,98,99,103,104,105,106,108,109,110,112,115,116,117,119],get_all_gpu:78,get_all_items_from_group:4,get_alpha_tracker_frame_dict:15,get_alpha_tracker_instance_dict:15,get_alpha_tracker_point_dict:15,get_available_gpu:78,get_candid:81,get_config:74,get_config_fil:86,get_connection_candid:76,get_current_gpu:78,get_default:86,get_edge_point:56,get_edges_as_np_str:9,get_filtered_frame_idx:13,get_fram:33,get_frame_count:13,get_frames_saf:33,get_gpu_memori:78,get_instance_anchor:61,get_instance_score_seri:7,get_instances_point:3,get_item_data:4,get_item_data_idx:4,get_item_group:4,get_keras_model_path:73,get_mean_and_std_for_point:3,get_model_output_strid:73,get_most_stable_node_pair:3,get_next_suggest:13,get_nodes_as_np_str:9,get_occupancy_and_points_matric:9,get_package_fil:86,get_paf_lin:76,get_point_count_seri:7,get_point_displacement_seri:7,get_point_score_seri:7,get_points_arrai:10,get_primary_point_displacement_seri:7,get_raw_imag:4,get_ref_inst:81,get_shifted_inst:81,get_shifted_instances_from_earlier_tim:81,get_stable_node_pair:3,get_suggest:13,get_symmetri:85,get_symmetry_nam:85,get_template_points_arrai:3,get_timestamp:82,get_track_count:13,get_track_instance_match:80,get_track_occup:13,get_tracks_as_np_str:9,get_unlabeled_suggestion_ind:13,get_video_suggest:13,get_video_track_occup:13,git:[1,102],github:[1,99,101,102,103,108,109],githubusercont:[102,103,109],give:[7,73,85,88,89,90,92,95,96,98,102,110,115,116],given:[1,3,6,9,10,13,24,26,27,32,33,48,62,70,73,75,76,80,81,84,85,88,89,102,107,108],global:[48,49,60,73,77],global_pool:[48,71],globalpeakfind:60,go:[1,11,34,55,80,81,95,96,102,105,116,117],goal:[85,115,117],goe:[1,95,119],golai:103,good:[0,1,22,26,30,36,49,70,79,80,81,88,91,93,95,99,112,117],googl:[1,90,102,103,106,108,109],googleapi:[104,105,106,107],gordon:87,got:[79,117],gotten:95,govern:1,gpl:24,gpu:[49,55,73,78,88,89,90,98,101,104,105,106,107,108,109,112,117],gpu_devic:[104,105,106,107,108,109],grab:105,grab_fram:105,grabber:105,grace:[87,100,101],gracefulli:0,gracious:42,gradient:[49,73,77,88],grant:[42,100],graph:[48,56,71,73,74,76,85,89,92,97,105,117],graph_symmetri:85,graphic:[50,115],grappler:[104,107],grayscal:[33,42,43,47,63,66,73,104,118],grayscale_to_rgb:63,great:[33,106,107,112],greater:[38,45,47,48,69,73,76,89,95],greedi:[73,76,79,80,81,89,95,108],greedili:70,greedy_match:[79,81],green:[33,115],grei:[115,119],grew:28,grid:[54,56,59,64,66,69,73,77,105],grid_height:[54,56,59,64,69],grid_width:[54,56,59,64,69],gridlin:21,gritti:106,ground:[6,60,70,73,74,84,89,92,107,109,113],group:[4,10,48,49,72,73,74,76,81,83,85,86,88,97,100,101,105,107,116,117],group_arrai:83,group_class_peak:72,group_data:4,group_exampl:58,group_examples_it:58,group_inst:76,group_instances_batch:76,group_instances_sampl:76,group_set:4,groupset:4,groupset_data:4,grow:1,growth:[78,102,105,106,109],gt:107,guarante:[13,15,18],guess:94,gui:[9,10,13,25,26,35,50,90,91,93,94,95,96,100,101,106,110,111,112,113,115,118],gui_progress:35,guid:[13,27,73,76,89,91,96,100,101,104,107,108,117,119],guidelin:[1,100],gvoc:70,h264:[24,100],h5:[12,13,26,33,48,50,73,89,92,95,96,103,104,106,107,109,111,114,118],h5file:86,h5py:[33,85,86,103,114],h5read:114,h:[35,84,87,89,100,101],ha:[1,3,9,10,12,13,26,27,33,38,39,40,43,44,47,48,50,52,55,61,65,68,69,73,75,76,78,81,85,88,89,90,92,95,96,98,102,109,115,116,117,119],had:[55,81,90,102,104,105,106,107,108,109],hand:[15,85,91],handi:[66,92],handl:[14,15,16,17,18,19,22,23,24,25,27,28,29,30,33,46,73,76,81,85,105,106],handshak:36,hang:100,happen:[26,78,104],happi:[93,95],harass:0,hard:[49,70,74,88,92,95],hard_keypoint_min:[49,109],hard_to_easy_ratio:[49,74,109],harder:95,hardest:70,hardkeypointminingconfig:[49,74],hardwar:90,harm:0,has_edg:85,has_embedded_imag:33,has_fram:13,has_missing_video:13,has_nod:85,has_offset:82,has_only_first_choice_match:79,has_predicted_inst:10,has_tracked_inst:10,has_user_inst:10,hash:85,hashabl:86,hassl:108,have:[0,1,3,4,8,9,10,13,24,26,29,32,33,36,38,39,42,43,44,47,48,49,50,53,54,55,56,57,59,60,61,62,63,66,67,68,69,70,72,73,76,79,80,81,85,86,88,89,90,91,92,93,94,95,96,97,98,99,102,103,104,105,106,107,108,110,112,115,116,117,118,119],haven:[95,98,102,118],hdf5:[9,12,13,19,21,24,26,28,29,33,85,86,89,96,103,104,111,112,114],hdf5video:[13,33],head:[2,3,13,27,38,40,43,45,48,65,73,75,76,102,103,104,105,106,109],head_index:103,head_loc:103,header:92,header_data:36,headless:73,headsconfig:48,hear:100,heart:1,heatmap:77,heavili:48,height:[10,15,33,34,35,38,39,43,44,45,47,48,49,53,54,56,62,63,66,67,69,72,73,75,76,77,84,89,96],height_:33,held:47,help:[1,2,38,45,48,49,66,73,74,89,95,99,101,107,108,113],helper:[33,86],here:[1,21,33,70,73,77,80,86,87,88,92,95,96,97,100,103,106,107,108,110,111,112,113,114],herebi:42,heurist:[49,89,95,115],hexagon:87,hidden:78,hide:[78,92,104],high:[13,28,33,49,50,51,55,65,73,76,81,82,104,105,113],higher:[34,38,39,40,41,44,45,47,48,49,55,59,69,73,75,88,95,107],higher_hrnet_model:40,higherhrnet:40,highest:[40,49,72,89],highli:100,highlight:92,hindlegl4:[103,104,105,106,109],hindlegr4:[103,104,105,106,109],hint:[1,105,106,109,110],hip:70,hist:105,histori:73,histplot:107,hit:[1,95,110,117,118],hmac:109,hoc:48,hog:4,hog_bag_of_featur:4,hold:[10,21,33,75,76,92,95,116],holder:42,holist:107,home:[87,92,101,102,104,107,108,109],homebrew:102,hook:[10,85],horizont:[10,53],host:[90,109],hostnam:50,hotkei:95,hour:90,hourglass:[38,45,48,77,109],hourglassconfig:[39,48],hover:88,how:[8,11,35,39,47,48,50,67,70,81,86,87,89,90,91,92,93,94,95,96,97,98,99,103,104,105,106,107,108,110,112,113,115,116,117],howev:[48,100,106,113],hpc:89,hrnet32:40,http:[0,16,38,40,43,73,79,83,102,103,104,105,106,107,108,109],human:[14,15,16,17,18,19,22,23,24,25,27,28,29,30,51,77,101,113,117],hungarian:[72,79,83,89,95],hungarian_match:[79,81],hyperpamet:91,hyperparamet:[47,49,91,93,98,99,106],hyphen:89,i:[3,8,13,24,33,38,43,48,49,54,56,59,61,65,66,68,69,73,77,79,80,81,85,88,89,91,92,93,94,95,96,97,102,103,104,105,106,107,108,109,110,118],iccv:70,icloud:90,icon:109,id:[4,21,66,76,89,105,106,109,112],idea:[1,80,91,100,115,117],ideal:[10,55,97],idempot:74,ident:[0,6,38,43,44,48,55,66,68,69,71,73,80,81,84,85,87,89,92,93,95,96,103,115,119],identifi:[10,13,27,48,76,88,92,116],idl:90,idx:[4,6,13,33,105],idx_to_nod:85,idx_train:68,idx_val:68,ignor:[6,9,10,38,47,50,61,70,76,77,85,100,104,107,108],ignore_file_not_found:86,ii:8,illumin:49,illustr:112,imag:[4,10,13,15,18,24,33,35,39,40,42,43,46,47,48,49,50,53,54,56,59,62,63,66,67,69,73,76,77,81,82,84,85,86,87,88,89,90,91,92,95,96,97,98,104,105,106,107,108,110,113,116,117],image_dir:[15,18],image_height:[62,67,69],image_kei:[53,62,63,67],image_width:[62,67,69],imagenet:[42,43,47,48,63],imagenet_mod:[47,63,109],imagenet_preproc_v1:43,imageri:0,img:[33,34,35,73,77,81,84,104,105,108],img_b64:86,img_dir:16,img_height:53,img_scal:[81,89,108,109],img_to_cv:35,img_width:53,imgaug:[49,53],imgaugaugment:53,imgfig:84,imgstor:[13,24,33],imgstore_from_filenam:33,imgstorevideo:[13,33],immedi:[54,95],immut:85,impact:[49,50,73,74,100],impatien:87,implement:[14,26,33,38,39,40,41,42,43,44,46,51,63,67,70,72,73,76,80,82,85,112],impli:[42,55],importanalysisfil:26,importantli:[54,56],improv:[1,39,48,49,50,55,62,73,74,81,98,99,106,109,110],imshow:103,in_q:35,in_transact:13,inappropri:0,inceptionresnetv2:[42,48],inceptionv3:[42,48],inch:84,incid:0,includ:[0,8,9,10,13,23,24,29,33,35,38,42,43,70,73,78,86,88,89,91,92,94,96,100,101,102,104,105,107,112,113,115],include_top:43,inclus:0,incom:[45,48,53,54,55,56,57,59,60,61,62,63,67,68],incomplet:82,incorpor:[81,107],incorrectli:[89,92],increas:[39,41,44,47,48,49,50,54,56,59,62,73,81,84,88,116],increment:[39,81],incur:41,ind:[13,76],indefinit:55,independ:[49,62,103],index:[4,7,9,10,12,13,27,31,33,35,46,47,56,59,61,62,66,72,73,74,76,77,78,81,83,85,86,89,96,100,104,105,106,109],index_by_origin:33,indic:[6,11,13,33,46,47,53,54,56,59,61,62,66,68,70,72,73,76,77,79,83,85,89,92,95,115],indirectli:45,individu:[0,10,15,18,33,42,49,54,55,62,64,65,66,74,81,95,104,108,117],inds1:83,inds2:83,induc:69,inf:[80,83],infer:[10,13,33,47,48,50,51,55,58,62,65,66,67,72,75,76,77,83,88,91,92,93,94,95,97,98,101,103,106,107,113,115,116,119,120],inference_lay:73,inference_model:[73,105],inference_tim:105,inferencelay:73,inferencemodel:[73,89],infin:76,infinit:55,inflat:109,info:[42,73,95,100,104,106,109],inform:[0,13,15,28,39,41,42,44,48,51,61,73,79,89,91,92,94,96,99,100,104,105,106,114,116,117,119],inherit:[14,38],init:[81,102,105,106,109],init_don:81,init_frame_count:81,init_set:81,init_track:81,initi:[32,36,38,39,40,42,43,44,48,49,50,55,56,66,73,76,78,80,81,82,89,99,101,103,105,106,109,112,113,117,118,120],initial_config:[51,82,96,106,107,109],initial_downsampling_step:40,initial_learning_r:[49,109],initial_model:[50,109],initial_shap:103,initialize_devic:78,innov:101,input:[12,33,37,38,39,40,41,42,43,44,45,47,48,53,54,55,56,57,59,61,62,63,64,65,66,67,68,69,71,73,74,75,76,77,81,82,83,86,88,89,98,100,104,105,106,107,109],input_d:[53,54,56,57,59,60,62,67],input_format:[33,89,96,104,109],input_ind:73,input_kei:[53,54,55,56,57,59,60,61,62,63,67,68,82],input_key_nam:57,input_path:89,input_predict:89,input_sc:[47,62,109],input_scal:[47,73],input_shap:[43,75],input_tensor:43,input_video:89,inquiri:101,insert:[10,13],insid:[73,86,88,89,92,96,114],inspect:[15,18,65,92,93,104,105,109,110,113],inspir:52,inst_a:6,inst_b:6,instal:[1,24,28,89,90,96,98,103,104,105,106,107,112,114,118],instanc:[0,3,6,7,8,9,13,15,24,26,28,31,33,35,39,41,42,43,44,45,47,48,49,51,53,54,56,59,60,61,62,63,65,66,67,68,70,71,72,73,74,75,76,79,80,81,82,84,85,87,88,89,90,91,92,94,95,96,97,98,103,104,105,107,108,109,113,115,116,117,120],instance_assign:76,instance_confidence_map:[54,60],instance_confmap:73,instance_confmap_head:65,instance_count:[8,13,79,81,103],instance_crop:[47,106,109],instance_gt:70,instance_i:[79,81],instance_idx:103,instance_imag:[54,60,62],instance_iou_threshold:81,instance_node_veloc:103,instance_peak:[73,105],instance_peak_v:[73,105],instance_pr:70,instance_scor:[9,10,12,73,105],instance_similar:[79,81],instancecentroidfind:[61,73],instanceconfidencemapgener:54,instancecropp:62,instancecroppingconfig:[47,61,62],instances_a:6,instances_b:6,instances_kei:[53,61,62],instances_list:79,instances_to_show:10,instancetyp:[79,80,81],instanti:[24,33,38,39,42,43,44,51,65,71,74,75,81,82,85],instead:[13,38,40,41,44,45,48,49,53,65,69,82,89,107,109,110],institut:101,instruct:[1,96,100,101,102,104,105,106,107,109,118],insult:0,int32:[53,56,59,61,62,66,72,76,77,83,105],int64:[66,86],integ:[9,13,38,42,43,45,46,47,48,49,50,59,62,63,67,69,72,73,76,77,89],integr:[39,60,73,76,77,101],integral_patch_s:[60,73,77],integral_refin:73,integral_regress:77,intend:[13,51,82,100,101],inter:110,interact:[11,13,50,73,87,102],interactive_and_realtime_infer:[104,105,106],interactive_and_resumable_train:[104,105,106],interest:[0,100],interfac:[13,33,73,85,92,93,96,102,104,105,109,112,113],interfer:102,intermedi:[37,38,39,42,43,45,48,73,92],intermediate_activ:42,intermediate_feat:43,intermediateactiv:42,intermediatefeatur:[37,38,39,42,43,45],intern:[13,81],interp1d:103,interp:103,interp_method:[38,39],interpol:[38,39,41,44,48,76,95,103],interpret:[59,76,112],intersect:[79,81,83,95],intimid:91,introduc:[40,44,48],intuit:107,invalid:[43,80],invalid_v:80,invert:[67,85],investig:0,invis:10,invisible_as_nan:10,invok:[33,35,74,110],involv:113,io:[4,6,7,9,10,62,66,68,70,73,82,89,114],iou:[79,81,89,95],iou_threshold:[79,81],ip:50,ip_address:50,ipynb:[104,105,106],is_arboresc:85,is_first_choic:79,is_from_multi_size_video:66,is_gpu_system:78,is_grayscal:73,is_hdf5:21,is_initi:78,is_json:21,is_miss:33,is_multi_inst:13,is_open:21,is_templ:85,is_valid:81,isinst:[13,27],isn:[1,33,88,91,92,96,100],isnan:[10,103],isol:102,isomorph:85,issu:[0,67,99,100,101,102,104],item:[1,4,10,13,66,86,92,105],item_group:4,item_list:4,itemstack:4,iter:[4,11,13,33,55,58,66,68,85,86,113,117],its:[0,4,13,26,33,49,55,61,62,63,67,73,74,76,78,80,88,92,95,96,100,101,102,113,116,117,118,119],itself:[3,10,49,75,86,89,94,95,103,104],j:101,jan:28,jia:77,job:[50,51,73,82,89,90,91,96,105,107,109,113],john:[1,87,100,101],join:[92,95,107],joint:[73,88,103,118],joshua:[87,100,101],journal:[100,101],jpg:[13,24,109],json:[12,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,31,50,51,73,85,86,89,91,92,96,104,106,107,109,110],json_data:51,json_data_dict:51,json_dump:86,json_load:86,json_str:[85,86],jump:[92,95],jun:[104,105,106],junyu:[87,100,101],jupyt:[102,112],just:[4,10,13,24,33,50,62,73,80,85,86,88,92,94,95,96,97,98,102,103,104,105,106,108,110,111,115,116,117,119],k:[4,74,103],kaiyu:77,kalman:[81,89],kalman_track:81,kalmaninitset:81,kalmantrack:81,kanad:81,kb:[103,104],kde:107,kde_kw:107,keep:[21,36,47,48,49,50,55,66,67,81,86,89,92,97,115],keep_confmap:60,keep_empty_fram:13,keep_full_imag:[62,67],keep_kei:57,kei:[4,7,9,10,13,15,27,43,46,50,53,54,55,56,57,58,59,60,61,62,63,65,66,67,68,69,70,73,79,80,82,83,85,86,95,103,104,105,107,116,119],kept:[55,62],kera:[38,40,42,43,46,47,49,60,63,65,69,73,75,82,83,104,106,107,109],keras_model:[60,65,73,75,82,83,106],kerasmodelpredictor:60,kerastensor:[106,109],kernel:[38,39,43,44,45,48,77],kernel_s:[38,39,43,44,77,106,109],key_by_column:80,key_id:109,key_map:68,keyboard:[92,94,95],keydevicemov:57,keyerror:13,keyfilt:57,keymapp:[65,68],keypoint:[12,16,26,49,70,74,88,92,107],keyrenam:57,keyword:[1,35,43,66],kf_init_frame_count:[89,109],kf_node_indic:[89,109],kind:[24,42,99,103],kislin:[87,100,101],km:103,kmean:[4,103],knew:95,know:[10,92,102,105,119],knowledg:62,known:32,kocher:[87,100,101],kp:103,kwarg:[13,16,18,19,22,23,24,25,26,29,30,33,35,48,66,73,74,84],l1_cache_s:[104,107],l2_cache_s:[104,107],l3_cache_s:[104,107],l:[100,101,105,106],lab:[101,102],label:[4,6,7,9,10,13,15,16,17,18,19,23,24,26,27,29,31,33,35,43,46,47,48,49,62,66,68,70,73,81,82,83,85,87,88,90,93,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,117,118,119,120],labeled_fram:[10,13,104,105,108],labeledfram:[10,13,31,33,66,70,79,81,104],labels_a:13,labels_b:13,labels_config:82,labels_gt:[6,70,107,109],labels_path:[9,89,109],labels_pr:[6,70,106,107,109],labels_read:68,labels_test:13,labels_train:[13,68],labels_v:[13,68],labels_val_test:13,labels_with_imag:[104,105],labels_with_predict:89,labelscocoadaptor:16,labelsconfig:[47,82],labelsdatacach:13,labelsdeeplabcutcsvadaptor:18,labelsdeeplabcutyamladaptor:18,labelsdeepposekitadaptor:19,labelsjsonadaptor:24,labelsleapmatlabadaptor:25,labelsread:[13,65,66,68,70,73,82],labelsv1adaptor:[21,23,26],labori:113,lack:29,lah:[104,105,106],lambdafilt:55,lambdamap:57,landmark:[48,76,77,85,92,97,104,107],languag:0,larg:[28,48,49,50,54,62,88,92,95,100,101],larger:[33,48,49,62,73,76,88,89,98,100],largest:[47,48,49,55,62],last:[9,13,33,36,38,42,43,54,56,62,63,64,69,73,76,77,78,81,89,92,96,100,118],last_frame_idx:33,last_gpu:109,last_init_t:81,last_match:81,last_t:81,latenc:[55,101,105],later:[49,92,95,102,110],latest:[50,101,102,109,110],latest_model:[50,109],latter:[55,73,89,99],launch:[100,102,118],lauri:101,layer:[37,38,39,40,41,42,43,44,45,47,48,67,68,71,73,83,106,109],ld_library_path:[102,104,107,108],lead:[49,55,78,88,103,104],leadership:0,leap:[12,25,31,38,45,48,85,92,109,117],leapcnn:[38,41],leapconfig:[41,48],learn:[41,43,44,46,47,48,49,50,72,77,88,91,96,98,100,106,109,113,115,116],learnabl:[38,45,48],learning_rate_schedul:[49,109],learningratescheduleconfig:[49,106,109],least:[49,68,96,104,105,106,107,108,109],leav:[55,69],left:[15,47,49,53,55,67,68,70,73,85,88,92,94,105,106,107,118],legend:[92,95,103,107],len:[13,33,105],length:[3,10,42,55,62,67,68,69,70,73,76,77,83,89,92,95,103,104,105,109],lenient:90,less:[34,48,68,100],let:[8,90,91,92,94,96,98,103,104,105,107,108,109,110,116,117,119],level:[0,10,13,15,28,33,44,45,51,62,65,70,73,75,76,81,82,89,104,105,106,113],lewi:101,lf:[13,108],li:[87,100,101],liabil:42,liabl:42,lib64:[104,107,108],lib:[86,102,104,107,108,109],libcubla:[104,107,108],libcublaslt:[104,107,108],libcudart:[102,104,107,108],libcudnn:[104,107,108],libcufft:[104,107,108],libcurand:[104,107,108],libcusolv:[104,107,108],libcuspars:[104,107,108],libmamba:102,libnvinf:104,libnvinfer_plugin:104,librari:[28,42,53,102,104,105,106,107,108,109],libx264:[34,100],licens:[24,42],liezl:101,lighter:115,lightli:87,lightn:73,lightweight:[33,102],like:[3,13,33,36,38,42,45,47,50,70,74,80,81,88,90,92,93,94,95,96,98,100,102,103,104,105,106,107,108,109,110,114,116,118],likewis:[88,109],limit:[13,33,36,42,48,62,77,81,89,92],linalg:103,line:[1,3,5,9,12,35,48,73,76,91,92,93,94,95,96,98,101,109,112,115],line_scor:[73,76],line_scores_sampl:76,linear:[13,38,45,48,103],linear_sum_assign:[76,83],linearli:[39,48],liner:102,link:[81,104],lint:1,linux:[96,101,102,104,105,106,107,108,109,118],list:[3,4,6,9,10,11,12,13,14,15,16,17,18,19,22,23,24,25,27,28,29,30,31,32,33,35,36,38,39,40,41,42,43,44,45,47,48,50,53,54,55,56,57,58,59,60,61,62,63,65,66,67,68,69,70,71,73,75,76,78,79,80,81,82,83,84,85,86,89,92,93,95,97,102,103,104,113,116],list_file_miss:32,list_physical_devic:102,list_points_arrai:6,listen:50,literatur:107,littl:[97,99,102],ll:[3,7,10,13,14,15,18,24,26,32,36,70,81,85,86,88,90,91,92,94,95,96,97,98,100,102,103,104,105,106,107,108,109,110,112,114,115,116,118,119],load:[10,11,13,21,24,31,33,42,43,48,49,51,55,66,70,73,85,102,104,105,106,107,108,109,112],load_all_hdf5:85,load_by_chunk:11,load_config:[51,106,107],load_fil:[13,70,104,107,108],load_hdf5:85,load_json:[51,85],load_labels_json_old:31,load_mat:85,load_metr:[70,107],load_model:[73,104,105,107],load_predicted_labels_json_old:31,load_training_config:51,load_video:[33,104,105],load_weight:106,loader:13,loc:[103,107],local:[11,48,49,59,60,64,70,72,73,77,91,92,96,103,104,105,106,107,108,110,112,117],localhost:[105,106,109],localpeakfind:60,locat:[3,7,9,10,13,15,29,33,35,51,64,70,73,77,88,90,92,95,97,103,107,109,110,116,117],locations_matrix:9,lock:[107,109],log:[45,46,50,73,89,105,106,109,110],log_dir:46,log_to_csv:[50,109],logdir:46,logic:[51,72,73,78,79,105],logicaldevic:78,longer:[49,62,73,76,89,90,94,108],look:[1,26,33,38,56,61,70,81,86,88,89,90,91,92,95,107,111,116,117],lookup:13,loop:[13,48,49,50,55,76,82,101,105,106,109,112,113,117],loopbio:33,looser:80,lose:[39,89,90,116],loss:[40,41,48,49,50,71,82,88,98,106,109,110,117],loss_frequ:[50,109],loss_funct:71,loss_scal:[49,74,109],loss_weight:[48,71,106,109],lossi:[13,34,100],lossless:[13,33,100],lost:[79,88,89,95,119],lot:[50,92],love:100,low:[28,38,45,49,73,87,92,105,106],lower:[15,34,43,48,49,70,76,80,86,88,95,100,107],lowest:[42,49,74,81,92,109,110],lr:[46,50,106,109],ls:[104,105,106,110],luca:81,lw:84,m:[83,86,89,95,96,100,101,108,109,110],mac:[92,116,118],machin:[73,89,91,93,96,98,100,102,103,104,107,108,110,112,113,117],maco:102,macosx:102,made:[1,42,95,101,119],magnitud:49,mai:[0,1,12,13,21,35,38,39,41,43,44,46,47,48,49,50,51,53,55,61,62,66,68,70,73,74,76,88,89,92,95,96,99,100,102,103,104,105,107,112,113,114,116,117,119],mail:0,main:[1,9,12,13,39,48,73,75,79,82,85,92,95,103,108,115,118,119],maintain:[0,13,43,49,54,55,101],major:[9,12,99,104,114],make:[0,1,4,9,10,13,33,48,51,54,56,59,64,66,68,69,71,73,74,78,80,83,85,88,91,92,94,95,96,97,102,104,105,106,107,108,113,116,117],make_arch:24,make_backbon:[38,42,43],make_backbone_fn:43,make_base_pipelin:65,make_block:[38,39,44],make_cattr:85,make_centered_bbox:[62,77],make_class_map:59,make_class_vector:59,make_confmap:[54,59,64],make_dataset:[65,66],make_decod:38,make_default:10,make_edge_map:56,make_encod:38,make_gaussian_kernel:77,make_grid_vector:[54,56,64,69],make_head:71,make_instance_cattr:10,make_label:73,make_line_sub:76,make_model:75,make_model_movenet:73,make_multi_confmap:54,make_multi_confmaps_with_offset:54,make_multi_paf:56,make_offset:64,make_paf:56,make_pipelin:73,make_predicted_inst:76,make_resnet_model:43,make_sample_group:4,make_scoped_dictionari:86,make_skip_connect:45,make_stack:45,make_stem:38,make_track:81,make_tracker_by_nam:108,make_training_pipelin:65,make_video_callback:13,make_video_for_image_list:[15,18],make_viz_pipelin:65,mala:[87,100,101],male:[87,105,116],mamba:102,mambaforge3:102,manag:[100,102],mani:[10,11,13,38,48,73,81,86,89,92,95,98,100,102,113,115,117],manipul:[10,104],manner:10,manual:[15,18,49,50,92,106,110,113,115],map:[4,6,28,38,43,45,48,50,54,56,57,59,62,64,65,66,68,70,71,72,73,76,77,81,82,83,84,85,88,89,92,96,104,106,107,109,113],mar:[70,107],mare:101,margin:62,marielisa:[87,100,101],mark:[10,49,92,97,115,116,117],marker:[35,89],marker_s:[35,89],markers:103,mask:[13,56,59,64,73,103],mask_offset:64,masking_and_pad:73,mat:[12,85,92],match:[4,10,12,13,24,33,35,38,45,47,48,58,61,66,70,72,73,76,79,80,81,83,85,86,89,92,95,96,108,109,114,119],match_candid:76,match_candidates_batch:76,match_candidates_sampl:76,match_dict_from_match_funct:80,match_dst_peak_ind:76,match_dst_peak_inds_sampl:76,match_edge_ind:76,match_edge_inds_sampl:76,match_frame_pair:70,match_funct:80,match_inst:70,match_instance_list:6,match_instance_lists_nodewis:6,match_line_scor:76,match_line_scores_sampl:76,match_lists_funct:6,match_point:83,match_scor:[70,107],match_score_threshold:[70,107],match_src_peak_ind:76,match_src_peak_inds_sampl:76,match_threshold:70,match_to:[13,23,24],matched_instance_dist:6,matchedframeinst:81,matching_funct:[79,81],matlab:[9,12,25,31,85,89,103,113,120],matplotlib:[46,82,84,103,105,107],matplotlibsav:46,matric:[3,6,9,103,113],matrix:[3,6,9,10,33,54,70,79,80,81,83,103,105,106,109],matsliah:[100,101],matter:[53,96,97],max:[7,38,39,44,47,54,56,67,73,76,79,81,88,89,98,106,109],max_edge_length:76,max_edge_length_ratio:[73,76,89,109],max_hard_keypoint:[49,74,109],max_height_and_width:[66,67],max_i:10,max_image_height:67,max_image_width:67,max_inst:[73,89,109],max_level:81,max_siz:69,max_strid:[48,67,109],max_track:[81,89],max_val:63,max_x:10,maxim:[95,100,117],maxima:77,maximum:[38,41,42,43,44,49,54,59,63,66,67,69,73,74,75,76,77,79,81,89,106],maximum_strid:[38,42,43,62,75],maxpool:39,mayb:13,mb:[103,104,105,106,109],mckenzi:[87,100,101],mean:[3,4,7,12,13,33,34,43,47,49,54,63,70,74,79,80,89,90,92,95,100,102,103,105,107,114,117],meant:[73,76],measur:[70,76,95],mechan:106,media:[0,33],median:[10,70],mediavideo:[33,104,105],medium:92,meet:[47,55,65,67,92],melanogast:[87,103,109,111],member:[0,1],memori:[28,41,44,47,48,49,54,55,62,73,78,88,89,98,102,104,105,106,109],memory_s:[104,107],mention:[104,107,108],menu:[91,94,95,98,100,102,109,110,114,116,117,118,119],merchant:42,merg:[1,10,13,24,42,79,81,89,92,94,96,110,119],merge_container_dict:13,merge_fram:10,merge_matching_fram:13,merge_nod:13,merion:87,meshgrid:69,mess:80,messag:[50,73,89,100,115],message_queu:36,meta:[4,53,73,85],metadata:[4,10,13,15,18,21,28,29,37,50,73,104,106,107,112],method:[4,7,11,13,14,15,16,17,18,19,22,23,24,25,27,28,29,30,33,35,36,38,39,45,46,48,52,55,62,70,73,75,76,77,78,79,81,82,83,85,87,89,92,96,100,101,104,107,108,115,116,118,119],metric:[46,47,50,70,74,82,89,92,106,107,109,112],mib:[78,109],mice:87,micromamba:[104,107,108,109],microsecond:50,microsoft:90,middl:44,middle_block:[44,48,106,109],midlegl4:[103,104,105,106,109],midlegr4:[103,104,105,106,109],midpixel:21,midpoint:[10,48,61],might:[13,79,91,94,95,102,103,105,106,108,109,110],mikhail:[87,100,101],millisecond:105,mimic:60,min:[3,7,101,106,107,109],min_batches_per_epoch:[49,109],min_crop_s:62,min_dist:3,min_edge_scor:76,min_hard_keypoint:[49,74,109],min_instance_peak:76,min_learning_r:[49,106,109],min_line_scor:[73,76],min_match_point:[89,108,109],min_new_track_point:[81,89,108,109],min_point:81,min_shifted_point:81,min_stable_dist:3,min_val:63,min_val_batches_per_epoch:[49,109],mine:[49,74,88],mini:[38,49,55,74],minibatch:[49,88],miniconda:[86,101,118],miniforg:102,minim:[49,50,55,95],minima:49,minimum:[33,49,54,59,63,64,70,71,73,74,76,77,80,81,89],minut:[87,117],mirror:53,miscellan:[69,86],misidentifi:[95,119],mismatch:67,miss:[10,13,32,33,47,54,70,72,76,81,92,95,104,107,108],missing_path:13,mistak:[95,119],mistaken:[95,119],mit:[42,101],mitelut:[87,100,101],mix:[38,45,47,48],mjpeg:[24,33],mkdir:[102,109],mkv:24,mm:87,mobilenet:[42,48],mobilenetv1:[47,63],mobilenetv2:[42,47,48,63],mock_centroid_confid:62,mockglobalpeakfind:60,mode:[46,47,63,73,74,88,96,102,108,110,117],model:[28,37,38,39,40,41,42,43,44,45,46,47,49,50,51,56,59,60,62,63,65,66,67,70,71,72,73,74,76,82,83,87,89,90,91,92,93,94,95,96,97,98,99,103,104,105,106,113,115,116,117,119],model_evalu:[104,105,106],model_info:107,model_input_kei:60,model_nam:[43,73],model_output_kei:60,model_path:[70,73,107,109],modelcheckpointonev:46,modelconfig:[48,51,75],modifi:[10,13,39,40,42,53,66,78,79,80,83,92],modul:[4,6,7,10,31,34,35,36,38,39,41,42,43,44,45,51,65,66,73,75,76,77,78,80,83,85,86,107],mok:[70,107],moment:102,monitor:[50,73,74,87,98,115],monochrom:43,monterei:102,month:87,more:[4,8,10,13,28,38,39,41,42,43,44,45,48,49,52,59,65,68,69,73,76,78,80,81,82,85,88,89,90,91,92,97,99,101,102,107,108,113,114,115,117,118,119,120],most:[3,4,8,13,34,36,38,40,45,49,72,73,74,88,90,92,96,100,102,105,106,107,113,115,119],mostli:[8,13,50,55,78,115],motif:33,mount:[96,106,108],mous:[95,116],move:[13,47,50,57,76,87,92,94,95,97,115,116,119],movement:[81,92,113],movenet:73,movenetinferencelay:73,movenetinferencemodel:73,movenetpredictor:73,movi:[34,87],mp4:[13,24,27,33,34,87,89,96,100,104,105,108,109,110,118],mpck:[70,107],mpck_part:107,mpl:[103,107],ms:[84,105],mse:74,mu:87,much:[39,40,47,90,92,93,105],multi:[10,38,42,43,45,48,54,55,62,67,70,71,73,76,77,78,82,88,89,96,100,101,117],multi_class_bottomup:[48,109],multi_class_topdown:[48,105,109],multi_inst:[48,76,91,104,109,110],multiclassbottomupconfig:48,multiclasstopdownconfig:48,multiconfidencemapgener:54,multiinstanceconfig:[48,76],multiinstanceconfmapshead:[65,71,73],multiinstanceconfmapsheadconfig:[48,71],multipl:[1,10,13,15,18,20,21,26,33,36,38,39,47,48,50,54,56,62,65,67,68,72,76,78,79,80,85,89,91,92,94,105,106,109,116,117],multipli:[39,54,56,62],munkr:76,murthi:[87,100,101],musculu:87,must:[13,15,24,33,38,43,47,48,49,50,54,56,61,67,69,70,71,73,80,81,82,83,85,86,94,97,100,104,105,106,107,108,109],must_be_set:52,musuculu:87,mutual:13,my:[97,110],my_model:89,my_video:[13,27],n:[13,65,80,83,86,89,95,101,102,103,104,105,106,107,108,109],n_bbox:[62,77],n_candid:76,n_centroid:[62,73],n_channel:72,n_class:[59,72,73],n_cluster:[4,103],n_compon:4,n_connect:76,n_deconv_modul:40,n_down_block:67,n_dst:83,n_ed:70,n_edg:[56,76],n_frame:13,n_gt:70,n_instanc:[53,54,56,59,61,62,66,73,76],n_instances_gt:73,n_line_point:76,n_node:[10,13,53,54,56,61,62,66,70,73,76],n_output_channel:40,n_peak:[72,76,77],n_point:[61,64,76],n_positive_pair:70,n_pr:70,n_predicted_inst:10,n_sampl:[72,76],n_src:83,n_symmetri:53,n_total:13,n_track:[13,59],n_tracked_inst:10,n_user_inst:10,n_valid:[73,105],n_visible_point:10,name:[1,7,9,10,13,14,15,16,17,18,19,22,23,24,25,27,28,29,30,33,34,37,38,39,42,43,46,47,48,49,50,51,53,57,61,62,63,67,68,70,71,73,74,76,78,82,85,86,87,89,92,96,97,100,102,103,104,105,106,107,109,110,116,118,119],name_list:85,name_prefix:40,nan:[10,13,54,55,61,69,70,72,73,76,77,80,81,83,89,103,104,105],narrow:1,nasnet:[47,63],nat:101,nathaniel:[100,101],nation:0,nativ:[46,69,89],natur:[38,87,100,101],navig:95,nc:102,ndarrai:[3,4,6,9,10,11,13,33,35,36,53,66,69,70,73,76,79,80,81,83,84,85,86,105],ndx:27,ndx_pose:13,ndxposeadaptor:27,nearbi:[48,80,92],nearest:[38,39,76,92,103],nearli:100,necessari:[0,13,73,82,91,96,100,102,104,106],necessarili:[1,76],need:[1,12,13,21,26,33,35,42,44,47,48,50,61,63,66,67,70,74,76,80,81,85,86,87,88,89,90,92,93,94,95,96,97,102,103,104,105,106,107,109,110,113,114,115,116,118,119],needn:[21,32],neg:[13,70,79,104,105,106,107,108,109],negative_anchor:13,neglig:73,neighbor:[39,76,85],neighborhood:[48,59,64,77],neither:94,nest:[13,68],net:[38,98],network:[38,39,41,42,43,44,45,47,48,49,50,61,73,75,76,77,92,96,98,101,104,105,106,107,109,117],neural:[41,42,76,98,101,104,105,106,107,109],neurosci:[100,101],never:[47,55,94],new_centroid_confidence_kei:60,new_centroid_kei:60,new_fold:106,new_fram:[10,13],new_img:81,new_key_nam:57,new_label:13,new_nam:85,new_path:[32,70],new_peak_confidences_kei:60,new_peaks_kei:60,new_prefix:32,new_shap:83,new_skeleton:85,new_track:13,newaxi:103,newel:77,newer:107,newest:102,newli:[24,31,95,117],next:[4,13,92,95,102,105,106,107,108,113,116,117,119],nice:26,nicer:1,nih:101,nitti:106,nix:[78,89],nixadaptor:28,nixio:28,nixtrack:28,nm91:87,nms_fast:79,nn:[10,89,106,107,108,109],no_empty_fram:109,node1:85,node2:85,node:[3,6,7,9,10,12,13,15,27,28,29,35,47,48,53,54,56,61,62,70,71,72,73,76,81,85,87,88,89,92,94,95,96,97,98,103,104,105,106,107,108,109,115,116,118],node_a:3,node_b:3,node_count:103,node_ind:76,node_indic:81,node_loc:103,node_loc_vel:103,node_nam:[3,9,12,13,85,103],node_to_idx:85,node_to_index:85,node_type_ind:76,node_vel:103,node_veloc:103,nodeless_point_dist:6,nodes_point:10,nois:49,non:[10,13,26,33,49,50,67,73,77,79,81,89,92,93,97,102,103,105],none:[4,6,7,9,10,11,12,13,15,17,18,21,23,24,26,27,28,29,31,32,33,35,36,38,39,40,43,45,46,47,48,49,50,52,53,56,60,61,62,63,65,66,67,69,70,71,73,74,75,76,77,78,79,81,82,83,84,85,86,89,98,102,103,104,106,108,109],nonetyp:33,noninfring:42,nonn:89,nonzero:48,norm:[45,48,103],normal:[10,38,39,47,48,49,54,62,69,70,73,76,77,95,100,106],normalize_bbox:62,normalized_bbox:62,normand:[100,101],not_set:46,note:[3,4,9,12,23,24,29,33,45,49,55,74,85,86,92,94,98,100,102,103,104,105,107,108,109,111,114,116,119],notebook:[73,89,90,93,100,101,103,104,105,106,107,108,109,110,111,113],noth:[4,10,13,33,36,43,47,48,49,50,51,57,65,73,79,81],notic:42,notimplementederror:[15,26],now:[24,63,80,91,95,103,105,106,109,110,115,116],np:[9,13,33,73,76,77,83,86,103,105,107],npy:[24,33],npz:[70,106,107,109],ns104899:101,nstate:103,num:87,num_color_channel:73,num_conv:38,num_cor:[104,107],num_fc_lay:[48,71],num_fc_unit:[48,71],num_fram:33,numa:[104,105,106,107,108,109],number:[4,6,7,10,13,27,33,34,38,39,40,41,42,43,44,45,47,48,49,50,52,54,55,59,62,63,65,67,70,71,72,73,74,76,79,80,81,86,87,88,89,92,95,98,100,101,103,106,109,110,113,119],numer:[7,13,15,81,86,104],numpi:[3,4,6,9,10,11,13,27,28,33,35,36,53,66,69,70,73,76,77,79,80,81,83,84,85,86,102,103,104,105,106,107,108,109],numpyvideo:33,nutshel:113,nvidia:[101,102,104,105,106,108,109],nwb:[13,27],nwbfile:[13,27],o:[13,33,65,66,89,96,102,103,107,108,109,110],obj:85,object:[4,6,7,10,11,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,31,33,35,51,68,70,79,81,84,85,86,102,104,105,107,108,109],objectss:10,oblig:0,obtain:[42,70,73],obvious:95,occlud:[48,70,92,116],occup:[9,13,113],occupancy_matrix:[9,114],occupi:[13,48],occur:[10,13,50,55,95,113,115,116,119],occurr:13,octet:[103,104,109],odd:77,of_max_level:[81,89,108,109],of_window_s:[81,89,108,109],off:[1,46,70,73,90,95,99,104,106,107],offens:0,offic:101,offici:[0,100,102],offlin:0,offset:[10,48,54,59,64,65,67,71,73,77,82],offset_i:67,offset_refin:[48,109],offset_x:67,offsetrefinementhead:[65,71,73],offsets_head:65,offsets_ind:73,offsets_threshold:54,often:[50,70,74,77,95,96,100],ohkm:[49,74],ohkmloss:74,ok:[70,103,104,106,107,109],oks_scal:70,oks_stddev:70,oks_voc:[70,107],old:[24,31,32,85,87,94],old_key_nam:57,old_nam:85,old_path:32,old_prefix:32,old_skeleton:85,old_track:13,older:81,on_batch_begin:46,on_batch_end:46,on_epoch_begin:46,on_epoch_end:46,on_train_batch_begin:46,on_train_batch_end:46,on_train_begin:46,on_train_end:46,onc:[1,48,55,78,90,91,94,95,96,99,100,102,104,105,106,109,110,112,113,115,117,118,119],one:[4,10,13,33,43,48,50,51,52,53,59,65,67,68,69,70,73,75,81,82,85,89,91,92,95,97,98,101,102,104,105,106,107,108,109,117,118,119,120],oneapi:[104,105,106,107,109],onednn:[104,105,106,107,109],onedr:90,oneof:52,ones:[13,44,62,68,73,76,82,102],onli:[1,9,10,13,33,35,39,43,44,46,47,48,49,50,52,62,65,66,68,73,76,78,79,80,81,85,89,90,91,92,94,95,96,98,100,101,102,105,106,107,108,109,110,113],onlin:[0,49,74,88],online_min:[49,109],online_shuffl:[49,109],only_labeled_fram:109,only_suggested_fram:109,onto:[62,68,105],op:[37,38,39,42,43,44,45,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,71,72,73,74,76,77,78,82,83,104,107],op_level_cost_estim:[104,107],open:[0,1,13,21,33,85,86,87,89,91,92,94,99,100,101,102,104,107,108,109,110,111,120],open_in_gui:109,opencv:[33,34,35,96,104,105,106,107,108,109,110],oper:[4,13,38,39,41,44,47,54,55,56,62,63,65,66,67,68,69,70,73,74,76,77,100,104,105,106,107,109],oppos:51,opt:[100,102],optic:[81,89,95,96,108],optim:[47,48,50,51,53,55,65,71,74,76,82,83,88,100,104,105,106,107,108,109],optimal_group:73,optimization_callback:82,optimization_config:65,optimizationconfig:[49,51,65,82],optimizer_v2:82,optimizerv2:82,option:[1,4,6,8,9,10,11,12,13,14,15,17,18,21,23,24,26,27,28,29,31,32,33,35,36,38,39,42,43,44,45,46,47,48,49,50,51,53,56,60,61,62,63,65,66,67,68,69,70,71,73,75,76,77,78,79,81,82,83,84,85,86,88,89,92,95,96,102,105,108,113,116,119,120],orang:[92,95,116],order:[4,10,13,33,38,43,47,48,49,53,55,61,63,66,67,68,73,74,76,77,80,81,86,96,97,103,104,116,117],org:[0,16,28,38,40,43,73,104,107,108],organ:50,orient:[0,33,88,92],origin:[1,9,10,13,33,38,39,41,44,54,56,66,67,86,88,93,94,96,98,102,115],os:[86,101,102,104,105,106,107,108,109,110],osf:87,other:[0,3,6,10,13,24,26,29,33,42,44,49,50,51,65,66,67,68,75,79,80,81,85,86,89,90,91,92,96,98,99,100,101,102,103,104,105,106,107,109,111,112,116,117],otherwis:[0,1,9,10,13,15,18,33,38,42,43,55,63,65,69,73,85,89,92,96,102,103,110,116],our:[1,2,80,90,91,96,100,101,102,103,104,105,107,109,117,118,120],out:[3,10,13,28,42,47,49,50,55,65,73,76,88,90,92,95,98,99,101,102,104,105,106,107,108],out_q:35,outgo:[53,54,55,56,57,59,60,61,62,63,67,68],outlin:92,output:[9,10,12,13,17,27,29,33,35,37,38,39,40,41,42,43,44,45,48,51,53,54,56,57,59,61,62,65,66,67,68,69,71,73,74,75,76,82,85,88,89,92,96,102,104,105,106,107,109,110],output_callback:82,output_dir:13,output_filenam:33,output_filt:39,output_height:73,output_ind:73,output_kei:[53,54,55,56,57,59,60,61,62,63,65,66,67,68,82],output_key_nam:57,output_path:[9,13,109],output_predict:89,output_scal:[40,43,84],output_strid:[38,42,43,45,48,54,56,59,69,71,73,106,109],output_width:73,outputsconfig:[50,51,82],outsid:[10,85],over:[3,4,6,8,10,13,49,50,51,55,66,68,69,73,76,77,79,81,83,84,88,89,90,92,95,103,105,106,107],overal:[6,10,48],overfit:[49,88],overlai:35,overlap:[13,79,81,89,92,95,99],overload:73,overrid:[33,35,46,89],overridden:[11,73],overview:[100,101,104,120],overwrit:[13,27,50,82],own:[4,26,55,90,91,92,101,102,105,106,113,116,120],ownership:4,p15:87,p50:[70,107],p75:[70,107],p90:[70,107],p95:[70,107],p99:[70,107],p:102,packag:[47,82,86,89,90,91,94,96,100,101,103,104,105,106,107,108,109,114,118],pad:[47,55,62,66,67,69,73,89],pad_bottom:67,pad_right:67,pad_to_strid:[47,67,73,109],pade:73,paf:[10,48,49,56,65,73,74,76,82,84,89],paf_group:73,paf_line_point:73,paf_lines_sampl:76,paf_output_strid:73,paf_scor:73,pafs_head:65,pafs_ind:73,pafs_sampl:76,pafs_strid:76,pafscor:[73,76],page:[1,2,101,102,109,110],pai:90,pair:[3,6,48,53,70,71,73,76,79,83,84,85,87,89,95,107,110,118,119],pairedreceiv:36,pairedsend:36,pairwis:[6,56,70,81],palett:[35,89,92],pan:88,panda:[103,107],panel:[95,109,116,118,119],papadoyanni:[87,100,101],paper:[39,41,70,87,100,101],parallel:[4,48,49,50,55],parallelfeaturepipelin:4,param:[13,25,40,66],paramet:[4,6,7,9,10,11,12,13,15,17,24,26,27,29,31,32,33,35,38,39,40,41,42,43,44,45,49,51,52,53,54,55,56,59,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,88,89,91,95,96,98,99,100,106,108,109,110,112,113],parametr:51,parent:[50,97],pars:[31,51,73,86],parse_uri_path:86,parsed_json:31,part:[3,32,47,48,49,53,56,61,65,70,71,73,74,76,85,88,89,92,96,97,113,116,117,118],part_affinity_field:[56,73],part_nam:[48,71,76,82,106,109],partaffinityfieldsgener:56,partaffinityfieldshead:[65,71,73],partaffinityfieldsheadconfig:[48,71],partial:[85,87,113],particip:0,particular:[24,42,68,92],particularli:[44,66,76,78],partit:76,partloss:74,pascal:70,pass:[1,10,12,21,33,35,38,46,60,73,76,95],patch:[62,73,77],path:[9,12,13,17,27,31,32,33,34,35,39,43,46,47,48,50,51,66,70,73,82,86,89,90,91,92,95,96,100,102,103,106,107,108,109,110,114],path_a:13,path_b:13,path_to_video:[13,27],pathlib:[13,27],pattern:[38,45,46,50],pavel:42,paw:87,pb:73,pca:4,pci:[105,106,109],pck:[70,107],pck_voc:[70,107],pckage:110,pd:[103,107],pdf:[38,40,54,69,77,105],peak:[48,54,55,59,60,64,72,73,76,77,84,89,105],peak_channel_ind:[72,73,76,77],peak_channel_inds_kei:60,peak_channel_inds_sampl:76,peak_class_prob:72,peak_confidences_kei:60,peak_ind:[72,76],peak_point:[72,77],peak_sample_ind:[72,77],peak_sample_inds_kei:60,peak_scor:76,peak_scores_sampl:76,peak_threshold:[60,73,89,109],peak_val:[72,73,76,77],peak_vals_kei:60,peakid:76,peaks_kei:60,peaks_out_kei:60,peaks_sampl:76,peek:[65,103],penal:[73,76,89],penalti:[73,76,89],pep8:1,per:[4,7,11,33,35,41,44,48,49,73,79,80,81,84,88,89,95],per_group:4,percentag:[70,89,107],percentil:[70,107],pereira2022sleap:[100,101],pereira:[41,87,100,101],perf_count:105,perfect:[10,49,70],perfectli:[10,13,15,18,54,103],perform:[38,39,40,45,46,47,48,49,50,51,55,60,62,70,73,74,76,78,79,81,83,88,89,99,102,104,106,107,109,113],perhap:90,period:81,perman:0,permiss:[0,42,107,109],permit:42,perona:70,persist:79,person:[0,42,76,85,102,118],petri:87,pgm:24,physic:[0,78,101],physical_devic:[102,105,106,109],physicaldevic:[78,102],pi:77,pick:[79,91,92,95,106,116],pil:86,pip:[28,100,101,103,104,105,106,107,108,109,110],pipelin:[4,13,33,49,55,57,62,66,70,73,76,81,82,89,91,95,98,103,106,109,110,117],pipeline_build:82,pipelinebuild:82,pix_fmt:100,pixel:[10,33,35,47,48,49,53,54,56,59,62,67,70,73,76,77,89,98],pkg:[87,91,96,104,105,106,107,110],pkzip:24,place:[1,13,55,66,70,79,86,90,92,93,95,96,102,110,112,113,119],placement:[78,92,97,116],plai:104,plan:90,plateau:49,plateau_cooldown:[49,106,109],plateau_min_delta:[49,106,109],plateau_pati:[49,106,109],platform:[102,104,105,106,107,108,109],playback:[34,100],pleas:[1,87,100,101,104,107,108],plot:[10,84,92,103,104,105,107,108],plot_confmap:84,plot_fn:46,plot_img:[10,84],plot_inst:[10,84],plot_instance_node_veloc:103,plot_paf:84,plot_peak:84,plot_predict:10,plt:[103,105,107],plu:[4,10,95],png:[13,23,24,33,46,50],point:[1,3,6,7,9,10,13,15,35,48,53,54,56,59,61,62,64,65,66,67,70,72,73,76,77,81,83,86,89,91,95,98,104,105,107,116,117],point_confid:10,point_dist:6,point_match_count:6,point_nonmatch_count:6,point_scor:[9,12],point_val:72,pointarrai:10,points1:83,points2:83,points_and_scores_arrai:10,points_arrai:10,points_gt:70,points_kei:67,points_pr:70,pointsrescal:67,poli:103,polici:[0,78],polit:0,poll:50,poll_timeout:46,poller:50,polynomi:103,pool:[4,38,39,41,44,48,67,70],pool_before_conv:38,pooling_strid:[38,39,44],poolingblock:44,poor:[48,95,107],poorli:115,pop:[92,102,116],popul:79,popular:42,port:[11,50,104],port_numb:50,portabl:[33,82],portion:[42,45],pose:[13,27,41,48,70,76,77,81,85,89,99,105,106,107,108,109,112,113,117],poseestim:[13,27],poseestimationseri:[13,27],posit:[0,13,28,43,70,88,89,92,95,104,107,116,117],positive_pair:70,posix:86,possibl:[13,42,73,76,80,81,82,88,91,95,96,97,100,101,106],post:[0,48,81,83,87,100,103],post_connect_single_break:[81,89,108,109],post_inference_track:[104,105,106],postpon:13,postprocess:73,potenti:[18,50,55,61,73,102,106,113],power8:102,ppm:24,pr:1,practic:49,pre:[43,48,78,81,89,90,92,105,108,109],pre_cull_funct:81,pre_cull_iou_threshold:[89,108,109],pre_cull_to_target:[89,108,109],pre_track:81,preact:43,prealloc:[73,78],preced:89,preceed:89,precis:[41,48,70,92,107,113],predecessor:85,predict:[6,8,10,13,31,47,48,50,62,65,70,71,73,74,76,79,81,88,89,91,93,95,96,98,101,103,104,105,106,107,108,109,111,112,113,116,117,119,120],predict_on_batch:[73,105],predictcost:[104,107],predicted_center_instance_confid:60,predicted_center_instance_point:60,predicted_centroid:60,predicted_centroid_channel_ind:60,predicted_centroid_confid:60,predicted_centroid_sample_ind:60,predicted_inst:[10,13,60,67,76],predicted_instance_confid:60,predicted_instance_confidence_map:60,predicted_instance_scor:76,predicted_peak_scor:76,predicted_suggest:110,predictedcenterinstancenorm:60,predictedinst:[6,10,13,27,70,73,104,105],predictedpoint:[10,104],predictedpointarrai:10,predictions_fil:108,predictor:[73,104,105,107,109],preexist:[13,27],prefer:[33,48,51,80,89,97,99,100,102,118],prefetch:[13,33,49,55,73,109],prefix:[38,39,40,44,46,70,86,89,109],preform:92,preload:[49,55],preload_data:[49,109],prepar:[35,110],prepend:[38,39,46,50,69,89],preprocess:[13,42,43,47,49,63,73,75,109],preprocessingconfig:[47,63,67],preprocesss:106,present:[10,13,33,38,48,51,54,56,61,62,65,73,77,81,85,92],preserv:[13,47,86],preset:[34,100],press:[88,102,118],pretrain:[42,43,47,48,50,63,101,106],pretrained_encod:109,pretrainedencoderconfig:[42,48],pretti:[76,105,107],prev:116,prevent:[13,38,45,48,49,50,55,73,100,104,105],preview:[84,85,86,88,98],preview_imag:85,previou:[21,50,54,56,59,81,94,95,96,102,106,119],previous:[73,85],primari:39,primarili:[13,43,67,97],primary_nod:7,primit:[13,69,73],princeton:101,principl:26,print:[5,13,65,69,73,78,102,103,104,105,106,107,114],prior:[13,54,55,56,71,81,92,95,106,116],prioriti:[1,90],privat:0,privileg:[26,102],pro:90,probabl:[13,48,53,56,59,72,73,77,86,88,90,103,105,107],problem:[1,76,89,95,99,102,115],procedur:[10,49,70,104,107],process:[4,11,13,27,33,36,48,49,55,57,62,65,66,67,69,81,86,88,90,91,95,96,98,100,106,112,113,115,117],processingmodul:[13,27],produc:[6,33,48,54,55,56,63,65,66,71,73,81,88],product:[76,109],profession:0,profil:[50,51,89,93,96,109],profile_graph:[50,109],program:50,programmat:73,progress:[13,24,28,35,41,44,73,89,115],progress_callback:[13,23,24],progress_queu:35,progress_report:73,progressreporterzmq:46,project:[0,10,13,27,89,90,91,92,93,94,96,98,104,110,113,116,117,120],promot:[49,55],prompt:[25,89,91,92,102,110,118],prone:1,proof:115,proofread:[89,92,93,96,101,113,116,120],propag:[92,95],properli:[66,104,107,108],properti:[4,10,11,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,33,37,38,39,40,41,42,43,44,50,53,54,55,56,57,59,60,61,62,63,65,66,67,68,71,73,75,79,81,82,85],proport:[54,56],propos:1,protobuff:52,protocol:50,proven:[9,13,68,109],provid:[13,15,26,33,38,39,41,42,43,44,45,47,48,49,51,53,61,62,65,67,68,70,71,73,75,76,77,78,81,82,85,87,89,90,92,93,102,105,107,113,115],prune:81,prune_shifted_inst:81,pseudo:13,pt:104,pts_gt:84,pts_pr:84,pub:[36,50],publish:[0,36,42,50,100,101],publish_address:[50,109],publish_upd:[50,109],pull:[47,65,76,109],pup:87,pure:105,purepath:[13,27],purpos:[10,38,42,45,48,51,57,73,101,113],push:1,push_back_messag:36,put:[42,76,86,95],px:[87,107],py:[11,13,26,86,89],py_funct:76,py_util:104,pyimagesearch:79,pynwb:[13,27],pypi:[101,102,104,105,106,107,108,109,110],pyplot:[103,105,107],pyramid:[81,89],python3:[86,104,107,108],python:[12,13,24,28,37,38,39,42,43,44,45,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,71,72,73,74,76,77,78,79,82,83,86,89,90,95,100,101,102,103,104,105,106,107,108,109,110,112,113,120],pytorch:40,q:[107,108],qq:107,qqq:[104,105,106,107,108,109,110],qt:[14,15,35,89],qualiti:[34,100,113],quantif:101,quantil:[79,81,89],quantit:113,quarter:[73,77],qubvel:42,queri:[75,78,85],query_inst:[79,81],question:[1,93,100],queue:[35,81],quick:[87,89,92,104],quickli:[65,92,95,115],quiver:84,r01:101,r:[92,103,104,105,106,109,113,114],race:0,rag:[55,69,73,76,89],ragged_tensor:[69,73,76],raggedtensor:[55,69,73,76],rais:[13,15,33,43,50,52,61,62,65,67,68,73,78,85,86],raise_error:[13,33],raise_warn:33,ran:98,random:[43,48,49,50,53,68,98,116,117],random_crop:[49,109],random_crop_height:[49,109],random_crop_width:[49,109],random_flip:[49,109],randomcropp:53,randomflipp:53,randomli:[13,42,49,92,106],rang:[6,10,13,27,33,34,42,43,49,62,63,69,70,73,76,86,88,89,92,96,98,103,107],rangelist:13,rank:[13,33,42,54,55,56,61,63,69,73,77],rapidli:101,rate:[34,43,46,47,49,50,73,91,96,106,109],ratecolumn:73,rather:[50,54,55,59,61,67,90,96,103],ratio:[47,49,66,67,73,74,83],ravindranath:[87,100,101],raw:[4,21,47,49,51,53,54,56,66,73,102,103,105,108],raw_image_s:66,rc:102,rcparam:103,re:[1,8,12,13,26,81,89,90,91,92,93,94,95,96,97,99,100,102,103,104,105,106,107,108,109,110,112,114,115,117,119],re_init_aft:81,re_init_cooldown:81,reach:[36,55,99,100,101],read:[12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,33,35,51,66,68,70,89,91,95,96,100,101,104,105,106,107,108,109,112,114,115,120],readabl:33,reader:[13,33,35,66,82],readi:[1,73,106,112,115,117,120],real:[33,73,76],realiz:92,realli:[55,86],realtim:[76,100,101],reappli:73,reason:[0,49,76],rebuild:[13,104,105,106,107,109],rec:105,recal:[70,92,107,117],recall_threshold:[70,107],recarrai:10,receiv:[35,36,42,50,105,106,117],recent:[4,119],recept:[39,41,43,44,48,88,98],reciproc:[37,48,54,56,59,69],recogn:102,recommend:[1,50,55,73,99,100,101,102,104,105,106,107,108,117,118],recomput:112,reconstruct:100,record:[10,33,87,105,106,109,110,113],recording_dur:105,recov:[41,44,48,62,105,106,109,110],rectangl:92,rectangular:92,recurs:86,red:[33,95,115],reduc:[39,44,48,49,50,54,61,74,84,109],reduce_on_plateau:[49,106,109],reducelronplateau:109,reduct:[7,49,54,56,67],reduction_factor:[49,106,109],redund:[13,80,81],ref:40,ref_img:81,ref_inst:[79,81],ref_t:81,refer:[10,13,21,33,47,70,73,76,77,81,83,92,95,100,104,105,107],referenc:[13,70],refin:[38,39,45,48,54,62,65,71,73,77,82,99,113],refine_conv:[38,45,48],refine_convs_:38,refine_convs_activ:38,refine_convs_batch_norm:38,refine_convs_batch_norm_before_activ:38,refine_convs_batchnorm:45,refine_convs_filt:[38,45],refine_convs_filters_r:45,refine_convs_first_filt:38,refine_convs_kernel_s:38,refine_convs_use_bia:38,reflect:[49,53,70],refrain:102,regard:0,regardless:[0,13,49,50,73,86,88],regim:49,region:[48,53,88,92],regist:[10,14,15,16,17,18,19,22,23,24,25,27,28,29,30,85,102,104,107,108],regress:[64,73,77],regular:[10,81,89,92,115],regularli:115,reinstal:[105,106,109,110],reject:[0,73,76],rel:[3,10,37,38,40,42,43,45,48,49,50,54,56,59,61,62,66,73,76,77,88,95,97,98,117],relabel:85,relabel_nod:85,relat:[26,46,47,48,49,50,51,65,68,82,85,104],relationship:[10,48,85],releas:[1,101,102,109,110],relev:[1,14,15,73,92,119],reli:[43,48,51,73,81],reliabl:[48,81,100],relianc:48,religion:0,reload:[33,104,105,106,108],relu:[38,39,45,48],remain:[13,47,48,69,70,73,80,95,102],remaind:55,remap:65,remot:[13,89,91,101,102,117],remov:[0,9,10,13,66,70,72,79,80,81,85,86,89,94,95,97,102,108,110],remove_all_track:13,remove_empty_fram:13,remove_empty_inst:[10,13],remove_empty_tracks_from_matric:9,remove_fram:13,remove_inst:13,remove_predict:13,remove_redund:10,remove_second_bests_from_cost_matrix:80,remove_suggest:13,remove_track:13,remove_untrack:10,remove_untracked_inst:13,remove_unused_track:13,remove_user_inst:13,remove_video:13,renam:[57,85,118],rename_skeleton:85,render:[46,50,73],repeat:[1,38,39,48,55,67,113,115,118],repeatedli:[13,105],repercuss:0,replac:[10,13,24,32,33,39,54,59,64,68,69,70,71,77,80,83,85,91,92,100,102,106,108,115],replace_path:70,replic:[43,62],replica:[105,106,109],repo_id:109,report:[0,1,13,73,87,89,100],report_period:73,report_r:73,repositori:[42,87,102,109,112],repres:[0,4,10,13,15,33,35,42,48,59,66,70,76,77,81,85,86,92,97,99,104],represent:[0,13,41,43,44,48,51,73,76,77,85,86],reproduc:106,request:[11,33,55,103,104,109],requir:[10,13,24,26,33,38,47,48,49,50,51,62,63,65,67,73,75,76,88,89,96,102,104,107,108],rescal:67,research:101,reset:[33,55,83,89,102,117],reset_input_lay:83,reshap:[77,103],reshuffle_each_iter:55,residu:[38,39,40,43],resiz:[35,47,62,66,73],resize_and_pad_to_target:[47,67,109],resize_imag:[35,67],resize_input_lay:73,resnet101:[42,43,48],resnet152:[42,43,48],resnet18:[42,48],resnet34:[42,48],resnet50:[42,43,48],resnet:[48,109,117],resnetconfig:[43,48],resnetv1:[43,47,48,63],resnetv2:[43,47,63],resnext101:[42,48],resnext50:[42,48],resnext:43,resolut:[37,38,39,40,42,45,47,48,49,84,87,102],resolv:[76,94,103,104,109],resolved_fram:13,resourc:[48,49,74],respect:[0,28,35,38,43,44,48,49,50,55,56,69,74,77,87,116],respons:[38,50,89,103,104,109],rest:[38,85,105],restart:[102,104,105,106,108],restor:103,restrict:[42,64,86],result:[0,3,4,10,13,15,18,21,28,33,38,39,43,44,45,46,48,49,58,62,68,73,74,81,82,85,86,88,89,90,92,95,99,103,104,105,106,108,112],resum:89,retain:[48,79],retrack:[89,105,108],retriev:[10,76,107],return_class_map:73,return_class_vector:73,return_confid:13,return_confmap:73,return_crop:73,return_descript:[65,69],return_dict:85,return_new:13,return_paf:73,return_paf_graph:73,reus:[13,38],revers:[13,43,63],review:[0,1,13,81,95,100,101,120],reweight:74,rgb:[33,42,43,47,63,73],rgb_to_grayscal:63,rich:[73,89,109],richer:[41,44],right:[0,1,42,47,49,53,67,85,92,95,97,103,113,116,118,119],rm:[102,104,109],roadmap:1,robust:[79,81,88,89,92,109],robust_best_inst:[79,81],roll:103,ronch:70,ronneberg:44,root:[76,107,109,110],root_dir:86,rotat:[3,49,88,92,103,106,109,116],rotate_on_node_a:3,rotation_max_angl:[49,109],rotation_min_angl:[49,109],rough:[73,77],roughli:[26,76,87],round:[68,104,106],routin:73,row:[4,9,10,53,59,62,64,67,76,79,80,83],row_count:4,row_item:80,rst:[104,105,106],rtx:[105,106,109],rubric:38,rule:[88,97],run:[4,8,11,13,35,46,50,51,65,66,73,76,78,81,82,83,88,89,91,92,94,95,98,100,102,103,105,106,108,109,112,117,119,120],run_fold:50,run_nam:[50,89,96,106,109],run_name_prefix:[50,109],run_name_suffix:[50,109],run_path:[50,82],run_track:81,runs_fold:[50,109],runtim:[41,44,51,78,90,104,105,106,108,109],rw:[104,105,106],s0:109,s1:103,s2:103,s41592:38,s:[0,1,10,13,15,17,18,21,24,26,29,31,33,35,36,38,40,42,43,45,47,50,54,55,56,59,61,63,65,66,67,69,70,73,75,76,80,81,83,85,86,87,88,89,90,91,92,94,95,96,97,98,100,101,102,103,104,105,106,107,108,109,110,112,113,116,117,119],safari:109,safe:[13,100],safe_build:34,safely_gener:73,sai:90,salk:[0,1,99,100,101],same:[7,10,13,15,28,38,39,40,41,42,43,44,45,48,49,50,51,53,54,55,56,59,62,63,64,65,66,67,68,69,70,71,72,73,74,76,77,79,80,84,85,86,88,90,91,92,93,94,95,96,100,102,104,108,109,110],sampl:[4,13,43,46,47,49,54,55,56,59,64,68,69,72,73,76,77,87,88,92,98,99,103,111,112,113,116,120],sample_group:4,sample_ind:77,sample_method:4,sample_weight:74,samples_per_group:4,samples_per_video:4,samuel:[87,100,101],sane:[87,100,101],sanit:82,sanitize_scope_nam:82,sarah:[87,100,101],satisfactori:113,satisfi:[109,110,119],save:[9,12,13,24,26,31,33,34,35,46,47,48,49,50,51,66,70,73,81,82,85,86,87,89,91,92,94,96,100,103,104,106,107,109,110,117,120],save_all_hdf5:85,save_dict_to_hdf5:86,save_fil:[13,26],save_fold:46,save_frame_data:[23,24],save_frame_data_hdf5:13,save_frame_data_imgstor:13,save_hdf5:85,save_json:[51,85],save_labeled_video:35,save_output:[50,109],save_path:73,save_shifted_inst:[81,89,109],save_trac:73,save_tracked_inst:81,save_visu:[50,109],save_viz:[89,109],savedmodel:73,savgol_filt:103,savitzki:103,saw:102,scalar:[9,10,13,47,48,55,62,66,67,69,70,73,76,77,83,84],scale:[4,10,35,37,38,39,40,41,43,44,45,47,48,49,53,54,56,59,62,63,64,66,67,69,70,72,73,74,76,77,81,84,88,89,98,104,105,108,109],scale_i:67,scale_image_rang:63,scale_kei:67,scale_max:[49,109],scale_min:[49,109],scale_to_imagenet_caffe_mod:63,scale_to_imagenet_tf_mod:63,scale_to_imagenet_torch_mod:63,scale_x:67,scatter:103,schedul:[46,49,50,106,109],scientif:28,scikit:34,scipi:[76,79,103],scope:[82,86],score:[7,9,10,13,28,48,70,73,76,79,81,89,92,104,105,107],score_paf_lin:76,score_paf_lines_batch:76,scorer:76,scratch:[94,106,109,115],screen:1,screenshot:99,script:[89,91,105,106,112],scroll:116,seaborn:[103,107],search:[13,45,47,73,86,101,102,118],search_path:13,search_path_hint:[47,109],sec:109,second:[6,15,33,35,50,73,80,85,89,95,98,115],secondary_kei:13,section:[32,95,113],see:[1,7,10,13,15,16,20,21,28,33,38,39,41,42,43,44,45,46,48,49,54,55,59,62,63,64,67,69,70,72,73,74,76,77,78,79,83,87,88,89,91,92,94,95,96,98,99,100,101,102,103,104,105,106,107,108,109,110,112,113,114,115,116,117,118,119],seek:[13,100],seek_direct:13,seekabl:100,seekbar:[95,115,116,117,119],seem:107,seen:15,segment:[48,76],segmentation_model:42,select:[8,13,33,45,47,68,79,81,87,89,91,94,95,96,98,99,102,109,113,117,118,119,120],self:[1,81,82,87,105,110],sell:42,semant:48,send:[11,35,36],send_arrai:36,send_dict:36,sender:36,senet154:[42,48],sens:100,sent:[36,103,104,109],sep:[105,106],separ:[0,1,13,24,27,50,73,88,89,94,96,103,104,105,108,110,117],seq:86,sequenc:[13,38,43,45,48,65,66,68,71,92,105,110],sequenti:[35,50,53,100],seresnet101:[42,48],seresnet152:[42,48],seresnet18:[42,48],seresnet34:[42,48],seresnet50:[42,48],seresnext101:[42,48],seresnext50:[42,48],seri:[7,103],serial:[4,13,33,50,51,73,85,104],serialiaz:33,serializ:51,serv:[38,39,45,48,51,62,74,77,82,97,105],server:[11,90,102],servic:90,serving_default:73,session1:89,session:[78,90,105,117],session_descript:[13,27],session_start_tim:[13,27],set:[0,1,3,4,6,9,10,13,24,27,33,38,39,41,43,44,45,47,48,49,50,52,53,54,55,56,59,61,62,63,64,66,67,69,70,71,73,76,77,80,81,82,83,86,87,89,91,92,94,95,96,97,98,102,103,104,105,106,107,108,109,110,112,113,116,117,119],set_lr:[46,50],set_suggest:13,set_titl:103,set_xlabel:103,set_xlim:103,set_xtick:103,set_ylabel:103,set_ylim:103,set_ytick:103,set_yticklabel:103,setup:[82,87,102,107],setup_checkpoint:82,setup_loss:82,setup_metr:82,setup_new_run_fold:82,setup_optim:82,setup_optimization_callback:82,setup_output_callback:82,setup_tensorboard:82,setup_visu:82,setup_zmq_callback:82,sever:[49,89,107,112,113],sexual:0,sh:[91,102],sha256:109,shaevitz:[87,100,101],shall:42,shallow:[10,13,45,97],shape:[3,9,10,12,13,33,35,43,53,54,55,56,59,61,62,63,64,66,67,69,70,71,72,73,74,75,76,77,79,83,95,103,104,105,106,107,109,114],share:[4,78,86,96,100,102,104,107,108],sharex:103,shave:87,sheikh:76,shell:102,shellenv:102,sheridan:101,shiboken2:[105,106,109,110],shift:[81,89,92,95],shifted_inst:81,shiftedinst:81,shih:76,ship:102,shortcut:[38,43,86,92,94,95,104],shortnam:86,shot:49,should:[1,10,11,13,24,26,33,39,43,45,46,47,48,49,50,51,53,55,61,68,70,72,73,74,76,77,80,81,82,83,84,85,86,88,89,91,92,94,95,97,98,102,103,104,105,106,108,113,115,117],shouldn:26,show:[0,10,35,73,89,92,95,103,109,110,112,116,117,119],show_edg:[35,89],shown:[92,94,95,116],shruthi:[87,100,101],shuffl:[49,55],shuffle_buffer_s:[49,109],shuffler:55,shutil:24,sid:[104,105,106,107,108,109],side:[62,88,92,109,118,119],sigler:101,sigma:[48,54,56,59,69,71,77,106,109],sigma_threshold:71,sign:1,signal:[13,103],signatur:[43,73,80,109],signedhead:109,significantli:[48,79],silicon:[101,102],similar:[38,39,45,48,49,55,70,79,81,89,90,92,95,96,102,107,108,109,113],similarity_funct:[79,81],simon:76,simpl:[4,22,26,30,36,38,39,40,41,50,51,73,81,86,89,95,96,103,110],simple_block:40,simplecandidatemak:81,simpleconvblock:[38,41,44],simplemaxtrack:[81,89],simplemaxtrackscandidatemak:81,simpler:[39,45],simpletrack:81,simpleupsamplingblock:[38,41,44],simpli:[39,49,74,100,102,105,106,107,108,118],simplifi:[39,73],simul:[49,73,105],simulatedcamera:105,sinc:[10,21,24,26,33,36,47,48,49,54,55,70,73,76,79,80,81,86,88,89,91,92,95,97,98,102,107,108,109,110,112,113,115,116,117],singl:[3,4,7,10,13,15,18,24,33,35,36,43,47,48,49,50,51,52,54,56,61,62,63,65,66,67,71,73,76,78,79,81,82,84,85,87,88,89,90,91,92,94,95,101,104,105,110,117],single_inst:[48,91,109,110],single_instance_confmap_head:65,single_instance_lay:73,single_path_to_search:13,singleimagevideo:33,singleinstanceconfidencemapgener:54,singleinstanceconfmapshead:[65,71,73],singleinstanceconfmapsheadconfig:[48,71],singleinstanceconfmapspipelin:65,singleinstanceinferencelay:73,singleinstanceinferencemodel:73,singleinstancemodeltrain:82,singleinstancepredictor:73,singleton:[13,63,69],site:[86,104,107,108],situat:100,sixth:92,size:[0,10,35,38,39,42,43,44,45,47,48,49,50,55,62,63,66,67,69,70,73,76,77,81,84,87,88,89,98,104,107],sizematch:67,skeleton:[3,10,13,15,27,28,29,47,48,49,53,56,61,66,73,75,76,80,84,86,87,88,92,93,95,103,104,105,108,109,113,116,120],skeleton_edg:76,skeleton_ind:[56,61,66],skeleton_path:19,skew:[49,74],skip:[13,37,38,39,41,42,43,44,45,48,92,102,104,107,108],skip_add:[38,45],skip_connect:[38,43,48],skip_in:39,skip_label:13,skip_sourc:[38,39,45],skip_source_featur:38,sklearn:103,sleap:[2,87,88,90,91,92,93,94,95,97,98,99,102,103,107,111,113,114,115,116,117,120],sleap_activ:102,sleap_ci:86,sleap_deactiv:102,sleap_develop:102,sleap_jupyt:[104,107,108],sleap_label:92,sleap_old_ld_library_path:102,sleap_vers:[51,109],sleap_video_:[13,27],sleap_wt_gold:106,sleapanalysisadaptor:[26,29],sleapobjecttyp:[14,15,27],slice:[13,61,83,103],slightli:[102,104],slow:[13,49,50,88,92],slower:[73,79,88,95,105],slowli:117,slp:[9,12,13,23,33,70,87,89,91,96,103,104,105,106,107,108,109,110,111],small:[48,50,74,80,92,95],smaller:[39,48,54,56,59,64,67,69,70,76,98,116],smallest:[3,32,47,67],smith:[87,100,101],smooth:[77,103],smooth_diff:103,smooth_img:77,sn:[103,107],so:[3,9,10,12,13,21,28,31,33,35,39,42,44,47,48,50,54,62,63,64,70,74,76,78,79,81,87,88,90,92,93,94,95,96,100,102,103,104,105,106,107,108,109,110,114,115,117],social:0,socket:[36,50],softwar:[1,42,90,100,101,102],solar:[35,89],solv:1,solver:102,some:[1,3,7,13,21,29,31,44,55,65,73,76,78,85,87,88,92,93,94,95,96,100,102,103,104,105,106,107,108,109,110,112,115,116,117],some_issu:1,someth:[1,3,26,92],sometim:[77,81,95,99,108],somewher:110,soon:[100,109,110],sooner:90,sort:[3,6,49,70,80],sorted_edge_ind:76,souce_object:15,sound:91,sourc:[1,3,4,6,7,8,9,10,11,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,95,97,100,101,104,107],source_nam:85,source_nod:[48,71],source_object:[9,14,15,17,22,23,24,26,28,29,30],source_path:[17,28,29],source_points_arrai:3,source_video:33,source_video_avail:33,sp0:102,space:[0,10,48,50,69,76,90,92,100,102,116,118],span:[62,69,72,77],spars:74,spatial:[39,41,48],spatial_vec_length:76,spawn:[10,81,89,92,95,119],spawned_on:[10,105],spawned_track:81,speak:1,special:[42,50,73,90],specif:[0,10,13,14,15,26,27,45,48,50,51,63,68,73,74,79,82,85,92,98,100,105],specifi:[4,7,9,10,12,13,17,26,27,29,33,35,38,39,41,42,43,44,45,46,47,48,49,50,51,53,61,62,63,64,66,67,68,69,71,73,75,76,77,80,81,82,83,84,85,86,89,92,95,96,100,106,107,109],speed:[48,49,50,73,89,93,105,106,115,116],speedi:102,spent:[105,106],split:[13,47,50,62,66,68,70,82,86,106,107,109],split_by_ind:[47,109],split_fract:68,split_label:68,split_labels_read:68,split_labels_train_v:68,split_nam:70,spread:[48,54,56,59,70,71],spuriou:[73,76],squar:[73,77,84],square_bb:84,squash:1,squeez:99,src_node:[76,85],src_node_ind:[76,85],src_peak_ind:76,sse2:[104,107],sse3:[104,107],sse4:[104,107],sse:[104,107],ssse3:[104,107],stabl:[3,88,95,117],stack:[4,38,39,41,42,43,44,45,48,49,55,77,106,109],stack_config:43,stack_fn:43,stack_v1:43,stackoverflow:83,stage:[1,62,73,99,104,116,117],standalon:[76,104],standard:[3,35,38,43,45,47,48,49,54,55,56,59,63,70,77,80,87,89,96,104,105],start:[11,13,42,43,44,45,46,49,50,70,76,80,81,86,88,90,91,92,93,95,99,100,102,103,104,105,106,107,109,110,112,116,120],start_timestamp:109,startup:78,stat:107,state:[42,74,78,89,101,102],statist:[1,7,13,50,74],statisticseri:7,statu:1,std:[3,105],stddev:70,stem:[13,27,38,39,40,44,48,107],stem_block:[44,106,109],stem_features_strid:38,stem_filt:[39,40,43,48],stem_kernel_s:[44,106,109],stem_stack:[38,39,44],stem_strid:[39,48,109],stem_stride1:43,stem_stride2:43,stemblock:39,step:[1,3,13,38,39,40,41,42,43,44,45,48,49,54,56,59,63,67,69,73,76,80,81,88,92,95,99,100,101,102,105,106,109,113,117],still:[33,48,92,95,97,99,106],sto:85,stop:[47,49,50,98,106,109,110,117],stop_training_on_plateau:[49,106,109],storag:[13,47,49,50,90,104,105,106,107],store:[4,9,10,12,13,21,24,27,33,37,47,49,50,53,55,61,62,66,67,68,73,79,85,90,96,100,103,104,106,109,114,118],stori:100,str:[4,7,8,9,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,38,39,42,43,44,46,47,48,49,50,51,53,54,55,56,57,59,60,61,62,63,65,66,67,68,69,70,71,73,76,77,78,82,84,85,86,107],straight:102,straightforward:[3,97],strain:87,strategi:[95,101],stream:[61,63,67,103,104,109],stream_executor:[102,104,105,106,107,108,109],strictli:[73,76,89,102],stride1:43,stride:[4,10,37,38,39,40,41,42,43,44,45,47,48,54,56,59,62,64,67,71,73,75,76,84,88,92,98,106,109],string:[13,14,15,26,33,38,39,43,46,47,48,50,51,55,61,63,65,67,68,69,70,71,73,76,82,84,85,86],string_:9,stringent:107,strong:48,strongli:[101,102],structur:[10,13,38,39,49,51,68,73,85,87,105,106],stuck:100,studi:101,style:[41,42,43,92,95,107],sub:[10,11,36,50],subclass:[35,46,73,82],subdictionari:86,subdir:86,subdirectori:[50,82,86,96],subgraph:85,subject:42,sublay:39,sublicens:42,submenu:[95,119],submit:1,subpixel:[48,77],subplot:103,subsampl:[13,69],subscribe_to_control:[50,109],subscript:[76,85],subsequ:[39,44,62,92,95,102,104,105,106,108,116,119],subset:[10,13,33,66,68,83,86,89],substanti:[42,62],substitut:92,subtest:1,subtract:[43,47,63],success:[104,105,106,107,108,109,117],successfulli:[33,81,112,115],successor:[85,101],suffic:45,suffici:[43,48,88],suffix:[13,50,86,89,109],sugar:36,suggest:[4,13,23,24,66,81,89,93,95,98,99,100,115,116],suggestionfram:[4,13],suit:90,suitabl:100,sum:[7,56],summar:107,summari:[46,50,78,89,107],superfast:[34,100],supervis:[39,66],supplement:13,suppli:[88,96],support:[11,13,14,15,16,17,18,19,20,22,23,24,25,26,27,28,29,30,33,36,76,85,88,89,100,101,112,113,118],suppos:[92,94,95,96,116],suppres:77,suppress:79,sure:[1,10,13,47,80,83,91,96,97,102,104,105,106,107,108],surround:73,swap:[13,49,53,81,92,95,119],swiss:87,symmetr:[39,49,53,77,85],symmetri:[85,104],symmetric_ind:[53,85],symmetries_ful:85,symmetry_nam:85,sync:[10,90],sync_to_numpy_or_python_typ:69,sysf:[104,105,106,107,108,109],system:[13,24,33,47,50,73,89,101,102,105,106,108,109,116],system_summari:[102,104,105,106,108],t0:105,t:[1,3,10,12,13,21,26,29,32,33,36,45,48,55,74,79,80,81,86,88,89,90,91,92,93,94,95,96,97,98,99,101,102,103,104,105,106,107,109,110,112,113,115,116,117,118],tab10:103,tab:[50,88,91,98,109,119],tabl:[73,92,102,109,116],table_column:73,tabri:[100,101],tackl:1,tag:[46,50,109],tail:104,take:[0,4,6,26,38,42,43,45,49,54,55,56,59,73,77,81,85,86,90,91,92,95,98,102,103,104,105,106,107,108,111,116,117,119],taken:35,talmo:[0,1,31,87,99,100,101],talmolab:[102,103,104,105,106,107,108,109],talmop:101,target:[3,35,38,43,47,67,69,79,81,89,95],target_count:79,target_height:[47,67,109],target_instance_count:[81,89,108,109],target_points_arrai:3,target_rank:69,target_width:[47,67,109],task:[48,57,73,105,106,109],tcp:[36,46,50,109],td_fast:[104,105,107],td_id:105,team:[0,1],technic:[55,62,77,101],techniqu:95,technolog:101,tell:[13,88,92,99,100,117],templat:[1,26,85,96],temporari:[0,116],temporarili:0,tensor:[13,33,37,38,39,40,42,43,44,45,48,53,54,55,56,57,59,61,62,63,64,65,66,67,68,69,71,72,73,74,76,77,78,82,83,89],tensorboard:[46,50,82,89,109],tensorboardconfig:[50,82],tensorboardmatplotlibwrit:46,tensorfloat:[105,106,109],tensorflow:[37,38,39,42,43,44,45,53,54,55,56,57,59,60,61,62,63,64,65,66,67,68,69,71,72,73,74,76,77,78,82,83,89,90,100,102,104,105,106,107,108,109,117],tensorrt:104,tensorspec:[106,109],term:[38,48,71],termin:[49,102,118],teset:82,test:[1,13,33,47,49,50,70,73,82,87,89,104,105,106,107,109],test_ind:[47,109],test_label:[47,82,89,109],test_labels_read:82,text:[21,46,48,51,71,73,85,91,96,103,106],textadaptor:30,textual:74,tf2tensorrt:104,tf:[13,33,38,40,42,43,46,47,49,53,54,55,56,59,60,61,62,63,64,65,66,67,68,69,71,72,73,75,76,77,78,83,102,104,106,107,109],tf_enable_onednn_opt:104,tf_linear_sum_assign:[76,83],th:[62,77,89],than:[38,39,45,47,48,49,50,52,54,55,56,59,61,67,68,69,73,76,79,80,81,85,88,89,92,95,96,103,110,115],thei:[0,1,10,12,13,24,25,38,45,47,48,49,50,51,56,66,67,69,73,74,76,78,81,86,90,91,92,94,95,97,102,103,104,105,106,107,108,114],them:[35,48,49,54,55,56,67,76,82,85,90,92,93,94,95,97,99,102,104,105,106,107,109,110,112,113,117],themselv:1,therefor:[42,73],thi:[0,1,3,4,7,8,9,10,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,32,33,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,59,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,88,89,90,91,92,94,95,96,97,98,100,101,102,103,104,105,106,107,108,109,110,112,113,114,115,116,117,119,120],thicker:115,thin:[92,95,115],thing:[1,74,85,86,89,90,92,96,103,116],think:[90,98,100],third:15,thorax:[3,104,105,106,109],thorax_index:103,thorax_loc:103,those:[10,13,33,47,48,73,85,89,92,94,96,102,104,107,115,117],though:55,thought:69,thread:[11,35,102],threaten:0,three:[10,13,15,76,95,96,102,117],thresh:[6,80,107],threshold:[3,6,59,64,70,71,76,77,79,80,81,89,92,95,107],through:[11,13,49,51,54,55,73,76,89,90,92,95,101,102,104,106,107,112,113,117,120],throughout:92,throughput:[13,33],thrown:26,thu:3,thumb:[88,97],thumbnail:84,thumbnail_s:84,thunder:73,thx_vel_fly0:103,thx_vel_fly1:103,tick:103,tif:24,tigher:49,tight:84,tight_layout:103,tile:[42,43],tile_channel:43,time:[1,10,31,35,36,38,39,48,49,50,51,55,73,81,82,89,90,96,98,100,102,105,106,107,108,115,116],timeout:[36,50,90],times_to_check:36,timeseri:103,timestamp:50,timestep:81,tip:[93,100,119],titl:[10,100,101,103,108],tn:[70,107],to_arrai:10,to_dict:[13,85],to_hdf5:33,to_imgstor:33,to_json:[13,51,85],to_pipelin:[13,33],todo:[1,15,43,50],toget:[17,29],togeth:[13,42,48,55,65,73,80,87,95,96,104],toggl:[92,94,116],told:115,toma:76,too:[40,49,55,80,95,113],took:105,tool:[81,93,95,100,102,115,119],toolkit:[100,102],top:[13,33,38,43,48,67,72,73,74,88,89,91,96,98,101,104,105,109,117],topdown:[47,48,62,73,106,109,117],topdown_confmap:109,topdownconfmapsmodeltrain:82,topdownconfmapspipelin:65,topdowninferencemodel:73,topdownmulticlassfindpeak:73,topdownmulticlassinferencemodel:73,topdownmulticlassmodeltrain:82,topdownmulticlasspipelin:65,topdownmulticlasspredictor:73,topdownpredictor:[73,109],topic:46,topolog:76,toposort_edg:76,torch:[47,63],tort:42,total:[6,7,13,35,49,62,68,73,76,95,104,105,106,109],total_elaps:109,toward:0,tp:[70,107],trace:73,track001:[13,27],track:[7,8,9,10,12,13,27,28,29,36,47,48,59,62,66,67,68,73,75,79,80,84,87,92,93,96,97,100,101,103,104,105,109,113,114,115,116,120],track_idx:[13,27],track_ind:[59,62,66],track_matching_queu:81,track_matching_queue_dict:81,track_nam:[9,12,13,103],track_occup:[9,12,13,29,103,114],track_set_inst:13,track_swap:13,track_window:[81,89,108,109],trackclean:[81,95],tracked_inst:[10,81],tracked_label:108,tracked_lf:108,tracker:[73,81,89,92,95,96,108,109,110,112,119],tracker_max_inst:73,tracker_window:73,tracking_scor:[9,10,12,104,105],tracking_split2:[106,107],tracks_matrix:114,tradit:33,trail:[92,95,103],train:[10,13,24,33,43,46,47,48,49,50,51,55,56,60,65,66,70,73,74,75,83,87,88,92,93,94,97,99,101,103,104,105,107,111,113,115,116,119,120],train_end:46,train_start:46,trainabl:[43,48,75,88,105],trained_model:[105,109],trainer:[46,82,106,109],training_and_inference_on_an_example_dataset:[104,105,106],training_and_inference_using_google_dr:[104,105,106],training_config:[73,89,96,104,106,107,109],training_ind:[47,109],training_inst:10,training_job:[70,73,82,110],training_job_path:[89,109],training_label:[47,82,106,109],training_labels_read:82,training_log:[50,96,107,109],training_pipelin:82,training_profil:[96,109],training_slp_fil:106,training_viz_pipelin:82,trainingcontrollerzmq:46,trainingjobconfig:[51,70,73,82,104,106],transfer:[33,43,106],transform:[3,4,10,48,53,54,55,56,57,59,60,61,62,63,65,67,68,73],transform_dataset:[53,54,55,56,57,59,60,61,62,63,67,68],transform_point:10,transformation_matrix:10,translat:[49,109],translate_max:[49,109],translate_min:[49,109],transmit:100,transpos:[9,12,38,40,41,44,45,48,92,95,103,114,119],transposed_conv:[38,45,48],transposed_conv_:38,transposed_conv_activ:38,transposed_conv_batch_norm:38,transposed_conv_batch_norm_before_activ:38,transposed_conv_batchnorm:45,transposed_conv_filt:[38,45],transposed_conv_filters_r:45,transposed_conv_kernel_s:[38,45,48],transposed_conv_use_bia:38,treat:33,tree:[97,103,107,109],tri:[90,95,100],tricki:[49,96],trigger:[46,78],troll:0,troubl:[1,117],troubleshoot:93,trt:104,truncat:62,truth:[6,60,70,73,74,84,89,92,107,109,113],tunabl:48,tune:[43,47,49,74,99],tupl:[3,4,6,9,10,11,13,32,33,35,37,38,42,43,45,48,54,56,65,66,67,68,69,70,71,72,73,75,76,77,79,81,83,84,85],tuples_to_suggest:4,turn:[49,88,92,95,104,117],turner:[100,101],tutori:[94,95,100,101,108,113,117],tweak:[99,108],twice:21,twitter:101,two:[3,6,10,13,38,51,54,70,73,76,79,80,83,85,86,88,89,92,94,95,96,102,103,104,109,110,115,117,118,119],type:[1,4,6,9,10,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,33,37,38,39,40,41,42,43,44,45,47,48,49,50,51,53,54,55,56,57,59,61,62,63,65,66,67,68,69,70,71,73,74,75,76,79,81,82,85,86,89,91,93,95,96,97,99,100,101,102,104,105,106,107,109,112,113,117,118,119],type_spec:[106,109],typeerror:26,typic:[26,33,38,42,45,47,48,49,51,53,65,67,70,76,78,79,88,92,96,101,106,108,113],tzodikov:101,u:[38,89,98],uint8:[63,73,104,105],unabl:109,unaccept:[0,1],unaug:88,unavail:78,unbatch:[55,62],uncach:33,uncheck:95,unclear:80,uncorrect:94,under:[13,27,62,80,86,101,102,110,116],underl:13,underli:[10,13,21,33,51,55,73,76,82],underneath:48,understand:[70,92,100],undirect:56,undo:116,unet:[38,42,45,48,106,109,117],unetconfig:[44,48,106],unetpretrainedencod:42,unguiculatu:87,unifi:13,uniform:[49,67],uniform_nois:[49,109],uniform_noise_max_v:[49,109],uniform_noise_min_v:[49,109],uniformli:[49,55],uninstal:[104,105,106,107,108,109,110],union:[6,10,13,23,24,26,29,33,38,61,65,66,70,71,73,76,79,81,82,83,84,85],uniqu:[10,13,27,48,76,81,83,85,86,104],unique_sampl:4,unique_tracks_in_queu:81,uniquifi:86,unit:[48,49,53,54,56,59,71,76],univers:[90,101],unlabel:[13,66,76,89,92,113],unlabeled_suggest:13,unless:[13,26,33,86,92,115],unlik:38,unlimit:90,unmatch:[70,83],unmatched_inst:79,unmodifi:55,unnorm:[54,69,77],unnormalize_bbox:62,unoccupi:9,unpack:110,unrag:[55,89],unrag_exampl:69,unrag_output:73,unrag_tensor:69,unstructur:[10,70,85],until:[36,98,104,109,110,113,117,118],untrack:[13,27,79],untracked_inst:[79,81],unus:10,unused_predict:10,unwelcom:0,unzip:[91,96,107,109,110],uo:108,up:[9,33,36,38,44,45,48,49,50,56,61,66,68,73,79,80,82,86,87,88,89,91,92,93,95,97,101,102,103,104,106,107,108,109,110,115,116,117],up_block:[39,41,42,44,106,109],up_convs_per_block:41,up_interpol:[41,44,48,106,109],updat:[13,33,49,50,53,67,73,74,75,79,82,88,89,92,100,101,102,106,107],update_cach:13,update_config:[67,75,82],update_counts_for_fram:13,update_st:74,upload:[105,106,108],upon:[39,102],upper:15,upsampl:[38,39,40,41,42,43,44,48,67],upsampling2d:38,upsampling_stack:43,upsampling_strid:[38,39,45],upsamplingblock:39,upsamplingconfig:[45,48],upsamplingstack:[43,45],upshot:80,uri:86,us:[0,1,3,4,6,8,9,10,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,54,55,56,59,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,101,102,103,104,105,106,107,108,109,111,113,115,116,117,118,119,120],usabl:[86,117],usable_cpu_count:86,usag:[1,20,26,34,49,62,73,89,101],use_bia:[38,43],use_cach:13,use_cocoev:70,use_cpu_onli:[78,104],use_first_gpu:78,use_gpu:78,use_gui:13,use_last_gpu:78,use_missing_gui:16,user:[1,10,13,25,32,47,50,51,62,66,86,89,92,100,102,109,113],user_inst:[10,13,66],user_instances_onli:[13,66],user_label:13,user_labeled_fram:13,user_labeled_frame_ind:13,user_labeled_onli:13,user_labels_onli:70,userbas:100,usr:86,usual:[3,26,33,80,88,92,94,95,96,106],util:[5,12,32,37,54,56,59,63,64,68,70,72,73,76,78,84,95,102,104],uva:[87,100,101],v005:89,v11:[100,102],v1:109,v2:[43,105],v8:[100,102],v:100,val:[13,58,70,89,106,107,109],val_:46,val_abdomen:[106,109],val_batches_per_epoch:[49,109],val_ey:[106,109],val_eyel:[106,109],val_forelegl4:[106,109],val_forelegr4:[106,109],val_head:[106,109],val_hindlegl4:[106,109],val_hindlegr4:[106,109],val_label:[89,109],val_loss:[106,109],val_midlegl4:[106,109],val_midlegr4:[106,109],val_thorax:[106,109],val_wingl:[106,109],val_wingr:[106,109],valid:[10,13,33,39,42,43,46,47,48,49,50,52,63,68,73,80,82,86,87,89,92,95,98,106,107,109,110,117],validate_pipelin:65,validation_fract:[47,68,106,109],validation_ind:[47,109],validation_label:[47,82,109],validation_labels_read:82,validation_pipelin:82,validation_viz_pipelin:82,valu:[4,7,9,10,13,14,33,34,39,43,47,48,49,50,54,55,59,62,63,64,67,68,69,70,71,72,73,74,75,76,77,79,80,81,83,85,86,89,92,102,104,105,106,107,108,109],valueerror:[13,43,50,61,62,65,67,68,78,85,86],vari:[49,88,105],variabl:[55,67,69,73,74,85,102,104],variant:[38,40,43,48,65],variat:92,varieti:[42,96,104,105,115],variou:[7,12,13,33,89,92],ve:[13,81,91,92,93,94,95,96,97,98,103,107,110,112,115,118],vector:[3,4,48,54,56,59,64,65,69,72,73,76,77,83],veloc:[81,92,95],vendor:[104,107],venv:102,verbos:[73,89,109],veri:[22,30,49,74,92,95,100,101],verifi:[10,102,113],version:[0,1,13,48,51,65,69,86,89,94,96,100,101,102,104,105,106,107,108,109,110,112],vertic:[10,53],vgg16:[42,48],vgg19:[42,48],vi:[70,107],via:[0,1,11,13,38,45,48,55,73,76,77,89,92,95,101,102,103,105,106,109,110,112,118],vid:58,video:[4,7,9,10,11,12,13,15,17,18,24,25,27,28,29,31,34,35,47,66,68,70,73,87,88,89,90,92,94,95,96,97,98,100,104,105,108,109,111,112,113,115,116,120],video_fn:[13,27],video_idx:[4,13,27,35],video_ind:[9,58,66],video_index:89,video_list:70,video_path:[9,19,70,89,109],video_search:[13,23,24,33],video_search_path:82,videocaptur:33,videomarkerthread:35,videoread:[33,66],videos_as_dict:4,videowriteropencv:34,videowriterskvideo:34,view:[10,50,85,88,89,95,98,109],viewpoint:0,virgin:87,virtual:[26,90],visibl:[10,13,48,66,70,74,87,92,94,97,104,107,116],vision:100,vision_:77,visual:[46,49,50,62,65,73,82,84,89,92,95,97,98,105,106,107,109,115,117],visualization_callback:82,visualize_predict:46,visualpredictor:73,viz:[10,46,50,96,106,109],viz_fn:82,vmax:103,vmin:103,voc:[70,107],vocab_s:4,volum:[100,101],vs:[53,65,70,107],w:[35,84,100,101,102,104,105,107,108,113],wa:[10,13,15,18,32,33,38,39,46,50,51,54,62,63,66,73,78,80,83,85,89,92,94,95,100,101,104,105,106,107,109,110],wai:[13,39,42,65,74,76,78,90,92,95,96,97,98,102,105,107,114,115],wait:[50,81],walk:[90,112,120],wang:[87,100,101],want:[3,10,11,12,13,26,32,33,35,36,45,70,79,81,85,88,91,92,93,94,95,96,97,100,103,105,106,107,108,110,111,112,114,115,116,117,119],warn:[13,33,50,73,104],warrant:1,warranti:42,wasn:[80,107],watch:101,we:[0,3,6,7,10,11,12,13,15,18,21,24,26,29,31,32,33,35,36,79,80,81,85,86,87,89,92,94,95,98,99,100,101,102,103,104,105,106,107,108,109,110,112,115,116,117],weak_filename_match:86,websit:100,webster:87,wedg:[35,89,92,95],wei:76,weigh:[48,59,70],weight:[42,43,48,49,50,56,71,73,76,77,85,89,96,104,106,107],weird:99,welcom:[0,109,110],well:[3,13,33,39,44,45,49,50,51,54,56,59,65,70,73,74,90,92,95,98,99,102,103,104,106,107,110,112,115],were:[10,13,33,47,55,59,63,68,73,78,79,81,87,89,92,95,102,103,104,108,110,111,113,115],weren:79,wget:[102,103,104,107,108,109],what:[0,35,46,80,84,88,91,95,96,99,102,103,104,105,107,109,110,119],whatev:[33,94,96],wheel:[92,116],when:[0,1,10,11,13,24,28,33,35,38,43,45,47,48,49,50,53,54,55,56,61,62,63,66,69,70,73,74,76,78,79,81,85,88,89,90,92,93,94,95,96,97,98,99,100,102,103,104,106,108,109,110,112,115,116,117,118,119],whenev:80,where:[3,10,13,27,33,39,43,45,49,50,51,53,54,56,57,59,61,62,64,67,70,73,74,76,77,79,80,83,85,86,88,89,91,92,95,96,98,102,106,113,115,116],wherev:91,whether:[4,10,13,14,15,16,17,18,19,21,22,23,24,25,27,28,29,30,31,32,33,35,42,43,45,48,50,54,59,70,73,79,81,82,84,85,86,89,92,95,96,100,118],which:[0,3,4,5,6,7,8,9,10,11,13,14,15,16,17,18,19,21,22,23,24,25,26,27,28,29,30,32,33,35,38,39,40,45,46,48,49,50,53,54,55,58,59,61,62,64,66,67,70,71,73,74,76,77,79,80,81,82,84,85,86,87,88,89,90,91,92,94,95,96,97,98,100,102,103,104,105,107,110,112,115,116,117,119],whichev:[49,94],white:87,who:[0,42],whole:[76,88,92,105,110],whom:42,whose:[10,40,66,85],why:92,wide:[42,70,98,107],widget:89,width:[10,15,33,34,35,38,39,43,44,45,47,48,49,53,54,56,62,63,66,67,69,72,73,75,76,77,84,89,96],width_:33,wiki:0,win:103,window:[13,81,89,92,94,95,96,98,101,102,103,104,105,106,108,117,118,119],window_s:81,wingl:[103,104,105,106,109],wingr:[103,104,105,106,109],wise:[45,48,55,59,76,79],wish:[88,95],with_centroid:107,with_imag:[13,104],with_offset:54,with_track_onli:[13,66,82],with_user_labels_onli:13,withheld:92,within:[0,9,10,13,27,28,33,38,39,45,47,48,50,55,62,66,68,70,72,73,76,78,85,86,90,92,96,102,104,108,112],without:[0,9,10,13,14,15,16,17,18,19,22,23,24,25,27,28,29,30,33,36,41,42,49,50,55,63,68,70,75,76,78,79,85,89,94,96,101,102,103,106,108,112],won:[45,48,80,81,88,92,95,98],work:[1,3,12,13,26,28,32,33,42,50,55,73,74,78,86,87,88,89,90,95,96,98,99,100,101,103,110,112,113,114,116,117],worker:87,workflow:[93,101,105,106,112,113,116,120],worri:[74,90,95],worst:[74,88],worth:88,would:[1,26,55,62,67,68,73,77,79,80,88,92,94,100,102,104,106,107,108,110,116],wouldn:95,wrap:[8,15,18,65,73,76,78,105],wrapper:[4,13,33,34,55,67,70,75,76,79,85,86],write:[1,9,12,13,14,15,16,17,18,19,20,22,23,24,25,26,27,28,29,30,33,34,35,46,50,73,86],write_csv_fil:9,write_log:[50,109],write_occupancy_fil:9,writer:[34,35],written:[9,14,15,16,17,18,19,22,23,24,25,27,28,29,30,35,50,110],wrong:[33,95],wt_gold:[106,107],www:[38,73,79,104,106,107,108],x1:[10,53,62,77,83],x2:[10,53,62,77,83],x86_64:[102,104,105,106,107,108,109],x:[3,10,13,15,38,39,43,45,49,53,54,56,62,64,69,73,76,77,83,87,92,102,104,105,106,109],x_hat:77,x_in:[38,39,40,42,43,44,71],x_main:42,x_middl:39,x_out:[39,43],x_scale:66,x_up:39,xcode:102,xferd:[105,106],xiao:95,xla:104,xlabel:[103,105,107],xlim:103,xq:103,xtick:103,xv:[54,56,64,69,77],xy:77,xy_max:61,xy_mid:61,xy_min:61,y1:[10,53,62,77,83],y2:[10,53,62,77,83],y:[3,10,13,15,49,53,54,56,62,64,69,73,76,77,92,100,101,102,104,105,106,107,108,109,110],y_gt:74,y_hat:77,y_pr:74,y_pred:74,y_scale:66,y_true:74,yakubovskii:42,yaml:[12,18,86,92],yan:[87,100,101],yang:77,yaser:76,ye:[85,97,100],year:[100,101],yellow:[92,115,119],yet:[3,36,81,86,98,102,115,116,118],yield:[4,11,13,55,58,62,73],ylabel:[103,105,107],ylim:103,yml:102,you:[3,8,12,13,14,24,26,33,35,36,50,70,74,80,81,85,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120],your:[1,24,26,70,88,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,113,115,116,117,118,119,120],yourself:113,ytick:103,yuv420p:100,yv:[54,56,64,69,77],z:[87,100,101],zero:[9,10,33,54,59,64,67,71,89,95,103,104,105,106,107,108,109],zeromq:[50,82],zeros_lik:103,zhe:76,zip:[12,24,50,82,91,96,104,105,107,109,110],zip_output:[50,109],zmq:[36,50,89,109],zmq_config:82,zmqconfig:[50,82],zoom:[88,92,98,116],zprofil:102,zsh:102},titles:["Contributor Covenant Code of Conduct","Contributing to SLEAP","Developer API","sleap.info.align","sleap.info.feature_suggestions","sleap.info.labels","sleap.info.metrics","sleap.info.summary","sleap.info.trackcleaner","sleap.info.write_tracking_h5","sleap.instance","sleap.io.asyncvideo","sleap.io.convert","sleap.io.dataset","sleap.io.format.adaptor","sleap.io.format.alphatracker","sleap.io.format.coco","sleap.io.format.csv","sleap.io.format.deeplabcut","sleap.io.format.deepposekit","sleap.io.format.dispatch","sleap.io.format.filehandle","sleap.io.format.genericjson","sleap.io.format.hdf5","sleap.io.format.labels_json","sleap.io.format.leap_matlab","sleap.io.format.main","sleap.io.format.ndx_pose","sleap.io.format.nix","sleap.io.format.sleap_analysis","sleap.io.format.text","sleap.io.legacy","sleap.io.pathutils","sleap.io.video","sleap.io.videowriter","sleap.io.visuals","sleap.message","sleap.nn.architectures.common","sleap.nn.architectures.encoder_decoder","sleap.nn.architectures.hourglass","sleap.nn.architectures.hrnet","sleap.nn.architectures.leap","sleap.nn.architectures.pretrained_encoders","sleap.nn.architectures.resnet","sleap.nn.architectures.unet","sleap.nn.architectures.upsampling","sleap.nn.callbacks","sleap.nn.config.data","sleap.nn.config.model","sleap.nn.config.optimization","sleap.nn.config.outputs","sleap.nn.config.training_job","sleap.nn.config.utils","sleap.nn.data.augmentation","sleap.nn.data.confidence_maps","sleap.nn.data.dataset_ops","sleap.nn.data.edge_maps","sleap.nn.data.general","sleap.nn.data.grouping","sleap.nn.data.identity","sleap.nn.data.inference","sleap.nn.data.instance_centroids","sleap.nn.data.instance_cropping","sleap.nn.data.normalization","sleap.nn.data.offset_regression","sleap.nn.data.pipelines","sleap.nn.data.providers","sleap.nn.data.resizing","sleap.nn.data.training","sleap.nn.data.utils","sleap.nn.evals","sleap.nn.heads","sleap.nn.identity","sleap.nn.inference","sleap.nn.losses","sleap.nn.model","sleap.nn.paf_grouping","sleap.nn.peak_finding","sleap.nn.system","sleap.nn.tracker.components","sleap.nn.tracker.kalman","sleap.nn.tracking","sleap.nn.training","sleap.nn.utils","sleap.nn.viz","sleap.skeleton","sleap.util","Datasets","Configuring models","Command line interfaces","Run training and inference on Colab","Creating a custom training profile","GUI","Guides","Importing predictions for labeling","Tracking and proofreading","Running SLEAP remotely","Skeleton design","Training with GUI","Troubleshooting workflows","Help","Social LEAP Estimates Animal Poses (SLEAP)","Installation","Analysis examples","Data structures","Interactive and realtime inference","Interactive and resumable training","Model evaluation","Post-inference tracking","Training and inference on an example dataset","Training and inference on your own data using Google Drive","&lt;no title&gt;","Notebooks","Overview","Export Data For Analysis","Prediction-assisted labeling","Initial Labeling","Training and Inference","Creating a project","Tracking instances across frames","Tutorial"],titleterms:{"0":103,"1":[103,104,105,106,108],"2":[104,105,106,108],"3":[105,106,108],"4":[105,106,108],"5":106,"do":100,"export":[89,110,114],"import":[94,102,110],For:114,Of:1,The:95,about:90,across:[103,119],adaptor:14,advanc:[103,112],align:3,alphatrack:15,alreadi:100,an:[109,112],analysi:[103,112,114],anim:101,anoth:100,api:2,appear:100,applic:92,ar:102,architectur:[37,38,39,40,41,42,43,44,45],assist:115,asyncvideo:11,attribut:0,augment:53,basic:112,bee:87,better:93,between:103,bug:100,callback:46,can:100,clean:95,cluster:103,coco:16,code:[0,1],colab:[90,109,110],color:[95,103],command:89,common:37,compon:79,comput:100,conda:102,conduct:[0,1],confidence_map:54,config:[47,48,49,50,51,52],configur:88,contact:101,content:[1,102],continu:106,contribut:1,contributor:[0,101],convert:[12,89],correct:100,covari:103,coven:0,creat:[91,110,118],csv:17,cuda:100,custom:91,data:[47,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,90,95,100,103,104,105,106,108,109,110,112,114],dataset:[13,87,89,109,112],dataset_op:55,debug:89,deeplabcut:18,deepposekit:19,design:97,develop:2,diagnost:89,discuss:1,dispatch:20,document:101,doe:100,down:110,download:109,drive:[110,112],dynam:103,edge_map:56,encoder_decod:38,enforc:0,estim:101,eval:70,evalu:[107,112],exampl:[89,103,109,112],featur:101,feature_suggest:4,file:[89,92],filehandl:21,fill:103,find:103,first:116,fix:115,flies13:87,fly32:87,fly:103,format:[14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,100],found:100,frame:[110,116,119],from:[90,102],gener:[57,93],genericjson:22,gerbil:87,get:[93,100,101,102],go:92,googl:[110,112],gpu:[100,102],group:58,gui:[89,92,98,102],guid:[1,93],have:100,hdf5:23,head:71,help:[92,100,102],hourglass:39,how:100,hrnet:40,hyperparamet:88,i:100,ident:[59,72],imag:100,improv:100,infer:[60,73,89,90,96,104,105,108,109,110,112,117],info:[3,4,5,6,7,8,9],initi:116,inspect:[89,108],instal:[100,101,102,108,109,110],instanc:[10,110,119],instance_centroid:61,instance_crop:62,interact:[105,106,112],interfac:89,io:[11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35],issu:1,job:[106,110],just:100,kalman:80,kei:92,label:[5,89,92,94,115,116],labels_json:24,leap:[41,101],leap_matlab:25,learn:101,legaci:31,licens:101,line:89,load:103,loss:74,m1:102,mac:102,magnitud:103,main:26,mambaforg:102,matlab:114,menu:92,messag:36,method:[95,102],metric:6,mice_hc:87,mice_of:87,miss:103,model:[48,75,88,107,109,110,112],more:[95,100,103,116],mount:110,mous:92,move:90,movement:103,my:100,navig:92,ndx_pose:27,need:100,nix:28,nn:[37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84],normal:63,note:[10,13,46,50,53,54,56,59,61,62,63,65,66,70,72,73,76,77,78,81,83,110],notebook:112,offset_regress:64,open:118,optim:49,option:117,orient:95,our:0,out:100,output:[50,100],overview:113,own:[110,112],packag:[102,110],paf_group:76,palett:95,particular:100,pathutil:32,peak_find:77,perform:105,pip:102,pipelin:65,pledg:0,pose:[100,101],post:[108,112],pre:102,predict:[92,94,100,110,115],pretrained_encod:42,problem:100,process:110,profil:[91,110],project:118,proofread:[95,119],provid:66,pull:1,python:114,quick:101,realtim:[105,112],refer:101,rel:100,remot:[93,96],render:89,request:1,resiz:67,resnet:43,respons:0,result:[93,100],resum:[106,112],review:115,run:[90,93,96,110],sampl:109,save:[108,116],scope:0,script:95,seekbar:92,select:[92,116],set:100,setup:[104,105,106,108],shift:100,skeleton:[85,97,118],sleap:[1,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,89,96,100,101,104,105,106,108,109,110,112,118],sleap_analysi:29,social:101,some:101,someon:100,sourc:102,speed:103,standard:0,start:[117,118],strang:100,structur:[104,112],style:1,suggest:[92,110],summari:7,support:102,system:[78,100],t:100,tabl:1,talk:100,test:[102,108],text:30,thank:1,thing:102,thorax:103,tip:97,top:110,track:[81,89,95,108,110,112,119],trackclean:8,tracker:[79,80],train:[68,82,89,90,91,95,96,98,106,109,110,112,117],training_job:51,troubleshoot:99,tutori:120,type:88,unet:44,uninstal:102,up:100,upgrad:102,upload:110,upsampl:45,us:[100,110,112],usag:[100,112],util:[52,69,83,86],valu:103,ve:100,veloc:103,video:[33,103,110,118],videowrit:34,view:92,visual:[35,103],viz:84,what:100,where:100,without:100,work:102,workflow:99,write_tracking_h5:9,x:103,y:103,you:1,your:[110,112]}})
\ No newline at end of file
+Search.setIndex({"docnames": ["CODE_OF_CONDUCT", "CONTRIBUTING", "api", "api/sleap.info.align", "api/sleap.info.feature_suggestions", "api/sleap.info.labels", "api/sleap.info.metrics", "api/sleap.info.summary", "api/sleap.info.trackcleaner", "api/sleap.info.write_tracking_h5", "api/sleap.instance", "api/sleap.io.convert", "api/sleap.io.dataset", "api/sleap.io.format.adaptor", "api/sleap.io.format.alphatracker", "api/sleap.io.format.coco", "api/sleap.io.format.csv", "api/sleap.io.format.deeplabcut", "api/sleap.io.format.deepposekit", "api/sleap.io.format.dispatch", "api/sleap.io.format.filehandle", "api/sleap.io.format.genericjson", "api/sleap.io.format.hdf5", "api/sleap.io.format.labels_json", "api/sleap.io.format.leap_matlab", "api/sleap.io.format.main", "api/sleap.io.format.ndx_pose", "api/sleap.io.format.nix", "api/sleap.io.format.sleap_analysis", "api/sleap.io.format.text", "api/sleap.io.legacy", "api/sleap.io.pathutils", "api/sleap.io.video", "api/sleap.io.videowriter", "api/sleap.io.visuals", "api/sleap.message", "api/sleap.nn.architectures.common", "api/sleap.nn.architectures.encoder_decoder", "api/sleap.nn.architectures.hourglass", "api/sleap.nn.architectures.hrnet", "api/sleap.nn.architectures.leap", "api/sleap.nn.architectures.pretrained_encoders", "api/sleap.nn.architectures.resnet", "api/sleap.nn.architectures.unet", "api/sleap.nn.architectures.upsampling", "api/sleap.nn.callbacks", "api/sleap.nn.config.data", "api/sleap.nn.config.model", "api/sleap.nn.config.optimization", "api/sleap.nn.config.outputs", "api/sleap.nn.config.training_job", "api/sleap.nn.config.utils", "api/sleap.nn.data.augmentation", "api/sleap.nn.data.confidence_maps", "api/sleap.nn.data.dataset_ops", "api/sleap.nn.data.edge_maps", "api/sleap.nn.data.general", "api/sleap.nn.data.grouping", "api/sleap.nn.data.identity", "api/sleap.nn.data.inference", "api/sleap.nn.data.instance_centroids", "api/sleap.nn.data.instance_cropping", "api/sleap.nn.data.normalization", "api/sleap.nn.data.offset_regression", "api/sleap.nn.data.pipelines", "api/sleap.nn.data.providers", "api/sleap.nn.data.resizing", "api/sleap.nn.data.training", "api/sleap.nn.data.utils", "api/sleap.nn.evals", "api/sleap.nn.heads", "api/sleap.nn.identity", "api/sleap.nn.inference", "api/sleap.nn.losses", "api/sleap.nn.model", "api/sleap.nn.paf_grouping", "api/sleap.nn.peak_finding", "api/sleap.nn.system", "api/sleap.nn.tracker.components", "api/sleap.nn.tracker.kalman", "api/sleap.nn.tracking", "api/sleap.nn.training", "api/sleap.nn.utils", "api/sleap.nn.viz", "api/sleap.skeleton", "api/sleap.util", "datasets", "guides/bonsai", "guides/choosing-models", "guides/cli", "guides/colab", "guides/custom-training", "guides/gui", "guides/index", "guides/merging", "guides/proofreading", "guides/remote", "guides/skeletons", "guides/training", "guides/troubleshooting-workflows", "help", "index", "installation", "notebooks/Analysis_examples", "notebooks/Data_structures", "notebooks/Interactive_and_realtime_inference", "notebooks/Interactive_and_resumable_training", "notebooks/Model_evaluation", "notebooks/Post_inference_tracking", "notebooks/Training_and_inference_on_an_example_dataset", "notebooks/Training_and_inference_using_Google_Drive", "notebooks/analysis_example/README", "notebooks/index", "overview", "tutorials/analysis", "tutorials/assisted-labeling", "tutorials/initial-labeling", "tutorials/initial-training", "tutorials/new-project", "tutorials/proofreading", "tutorials/tutorial"], "filenames": ["CODE_OF_CONDUCT.md", "CONTRIBUTING.md", "api.rst", "api/sleap.info.align.rst", "api/sleap.info.feature_suggestions.rst", "api/sleap.info.labels.rst", "api/sleap.info.metrics.rst", "api/sleap.info.summary.rst", "api/sleap.info.trackcleaner.rst", "api/sleap.info.write_tracking_h5.rst", "api/sleap.instance.rst", "api/sleap.io.convert.rst", "api/sleap.io.dataset.rst", "api/sleap.io.format.adaptor.rst", "api/sleap.io.format.alphatracker.rst", "api/sleap.io.format.coco.rst", "api/sleap.io.format.csv.rst", "api/sleap.io.format.deeplabcut.rst", "api/sleap.io.format.deepposekit.rst", "api/sleap.io.format.dispatch.rst", "api/sleap.io.format.filehandle.rst", "api/sleap.io.format.genericjson.rst", "api/sleap.io.format.hdf5.rst", "api/sleap.io.format.labels_json.rst", "api/sleap.io.format.leap_matlab.rst", "api/sleap.io.format.main.rst", "api/sleap.io.format.ndx_pose.rst", "api/sleap.io.format.nix.rst", "api/sleap.io.format.sleap_analysis.rst", "api/sleap.io.format.text.rst", "api/sleap.io.legacy.rst", "api/sleap.io.pathutils.rst", "api/sleap.io.video.rst", "api/sleap.io.videowriter.rst", "api/sleap.io.visuals.rst", "api/sleap.message.rst", "api/sleap.nn.architectures.common.rst", "api/sleap.nn.architectures.encoder_decoder.rst", "api/sleap.nn.architectures.hourglass.rst", "api/sleap.nn.architectures.hrnet.rst", "api/sleap.nn.architectures.leap.rst", "api/sleap.nn.architectures.pretrained_encoders.rst", "api/sleap.nn.architectures.resnet.rst", "api/sleap.nn.architectures.unet.rst", "api/sleap.nn.architectures.upsampling.rst", "api/sleap.nn.callbacks.rst", "api/sleap.nn.config.data.rst", "api/sleap.nn.config.model.rst", "api/sleap.nn.config.optimization.rst", "api/sleap.nn.config.outputs.rst", "api/sleap.nn.config.training_job.rst", "api/sleap.nn.config.utils.rst", "api/sleap.nn.data.augmentation.rst", "api/sleap.nn.data.confidence_maps.rst", "api/sleap.nn.data.dataset_ops.rst", "api/sleap.nn.data.edge_maps.rst", "api/sleap.nn.data.general.rst", "api/sleap.nn.data.grouping.rst", "api/sleap.nn.data.identity.rst", "api/sleap.nn.data.inference.rst", "api/sleap.nn.data.instance_centroids.rst", "api/sleap.nn.data.instance_cropping.rst", "api/sleap.nn.data.normalization.rst", "api/sleap.nn.data.offset_regression.rst", "api/sleap.nn.data.pipelines.rst", "api/sleap.nn.data.providers.rst", "api/sleap.nn.data.resizing.rst", "api/sleap.nn.data.training.rst", "api/sleap.nn.data.utils.rst", "api/sleap.nn.evals.rst", "api/sleap.nn.heads.rst", "api/sleap.nn.identity.rst", "api/sleap.nn.inference.rst", "api/sleap.nn.losses.rst", "api/sleap.nn.model.rst", "api/sleap.nn.paf_grouping.rst", "api/sleap.nn.peak_finding.rst", "api/sleap.nn.system.rst", "api/sleap.nn.tracker.components.rst", "api/sleap.nn.tracker.kalman.rst", "api/sleap.nn.tracking.rst", "api/sleap.nn.training.rst", "api/sleap.nn.utils.rst", "api/sleap.nn.viz.rst", "api/sleap.skeleton.rst", "api/sleap.util.rst", "datasets.md", "guides/bonsai.md", "guides/choosing-models.md", "guides/cli.md", "guides/colab.rst", "guides/custom-training.md", "guides/gui.md", "guides/index.md", "guides/merging.rst", "guides/proofreading.md", "guides/remote.md", "guides/skeletons.md", "guides/training.rst", "guides/troubleshooting-workflows.md", "help.md", "index.rst", "installation.md", "notebooks/Analysis_examples.ipynb", "notebooks/Data_structures.ipynb", "notebooks/Interactive_and_realtime_inference.ipynb", "notebooks/Interactive_and_resumable_training.ipynb", "notebooks/Model_evaluation.ipynb", "notebooks/Post_inference_tracking.ipynb", "notebooks/Training_and_inference_on_an_example_dataset.ipynb", "notebooks/Training_and_inference_using_Google_Drive.ipynb", "notebooks/analysis_example/README.md", "notebooks/index.rst", "overview.rst", "tutorials/analysis.rst", "tutorials/assisted-labeling.md", "tutorials/initial-labeling.rst", "tutorials/initial-training.rst", "tutorials/new-project.md", "tutorials/proofreading.rst", "tutorials/tutorial.rst"], "titles": ["Contributor Covenant Code of Conduct", "Contributing to SLEAP", "Developer API", "sleap.info.align", "sleap.info.feature_suggestions", "sleap.info.labels", "sleap.info.metrics", "sleap.info.summary", "sleap.info.trackcleaner", "sleap.info.write_tracking_h5", "sleap.instance", "sleap.io.convert", "sleap.io.dataset", "sleap.io.format.adaptor", "sleap.io.format.alphatracker", "sleap.io.format.coco", "sleap.io.format.csv", "sleap.io.format.deeplabcut", "sleap.io.format.deepposekit", "sleap.io.format.dispatch", "sleap.io.format.filehandle", "sleap.io.format.genericjson", "sleap.io.format.hdf5", "sleap.io.format.labels_json", "sleap.io.format.leap_matlab", "sleap.io.format.main", "sleap.io.format.ndx_pose", "sleap.io.format.nix", "sleap.io.format.sleap_analysis", "sleap.io.format.text", "sleap.io.legacy", "sleap.io.pathutils", "sleap.io.video", "sleap.io.videowriter", "sleap.io.visuals", "sleap.message", "sleap.nn.architectures.common", "sleap.nn.architectures.encoder_decoder", "sleap.nn.architectures.hourglass", "sleap.nn.architectures.hrnet", "sleap.nn.architectures.leap", "sleap.nn.architectures.pretrained_encoders", "sleap.nn.architectures.resnet", "sleap.nn.architectures.unet", "sleap.nn.architectures.upsampling", "sleap.nn.callbacks", "sleap.nn.config.data", "sleap.nn.config.model", "sleap.nn.config.optimization", "sleap.nn.config.outputs", "sleap.nn.config.training_job", "sleap.nn.config.utils", "sleap.nn.data.augmentation", "sleap.nn.data.confidence_maps", "sleap.nn.data.dataset_ops", "sleap.nn.data.edge_maps", "sleap.nn.data.general", "sleap.nn.data.grouping", "sleap.nn.data.identity", "sleap.nn.data.inference", "sleap.nn.data.instance_centroids", "sleap.nn.data.instance_cropping", "sleap.nn.data.normalization", "sleap.nn.data.offset_regression", "sleap.nn.data.pipelines", "sleap.nn.data.providers", "sleap.nn.data.resizing", "sleap.nn.data.training", "sleap.nn.data.utils", "sleap.nn.evals", "sleap.nn.heads", "sleap.nn.identity", "sleap.nn.inference", "sleap.nn.losses", "sleap.nn.model", "sleap.nn.paf_grouping", "sleap.nn.peak_finding", "sleap.nn.system", "sleap.nn.tracker.components", "sleap.nn.tracker.kalman", "sleap.nn.tracking", "sleap.nn.training", "sleap.nn.utils", "sleap.nn.viz", "sleap.skeleton", "sleap.util", "Datasets", "Using Bonsai with SLEAP", "Configuring models", "Command line interfaces", "Run training and inference on Colab", "Creating a custom training profile", "GUI", "Guides", "Importing predictions for labeling", "Tracking and proofreading", "Running SLEAP remotely", "Skeleton design", "Training with GUI", "Troubleshooting workflows", "Help", "Social LEAP Estimates Animal Poses (SLEAP)", "Installation", "Analysis examples", "Data structures", "Interactive and realtime inference", "Interactive and resumable training", "Model evaluation", "Post-inference tracking", "Training and inference on an example dataset", "Training and inference on your own data using Google Drive", "&lt;no title&gt;", "Notebooks", "Overview", "Export Data For Analysis", "Prediction-assisted labeling", "Initial Labeling", "Training and Inference", "Creating a project", "Tracking instances across frames", "Tutorial"], "terms": {"In": [0, 8, 38, 43, 50, 72, 75, 76, 87, 88, 92, 95, 96, 97, 99, 103, 104, 105, 106, 108, 109, 110, 112, 113, 117, 119], "interest": [0, 100], "foster": 0, "an": [0, 1, 3, 4, 6, 9, 10, 11, 12, 13, 14, 20, 23, 25, 26, 28, 32, 35, 36, 37, 38, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 51, 52, 54, 56, 60, 61, 62, 64, 66, 67, 68, 69, 70, 72, 73, 74, 75, 76, 77, 78, 79, 80, 82, 83, 84, 85, 86, 88, 89, 91, 92, 94, 95, 96, 97, 98, 100, 101, 102, 103, 104, 105, 106, 107, 108, 110, 114, 115, 116, 117, 119], "open": [0, 1, 12, 20, 32, 84, 85, 86, 87, 89, 91, 92, 94, 99, 100, 101, 102, 104, 107, 108, 109, 110, 111, 120], "welcom": [0, 109, 110], "environ": [0, 77, 86, 101, 102, 103, 104, 107, 118], "we": [0, 3, 6, 7, 10, 11, 12, 14, 17, 20, 23, 25, 28, 30, 31, 32, 34, 35, 78, 79, 80, 84, 85, 86, 87, 89, 92, 94, 95, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 115, 116, 117], "maintain": [0, 12, 42, 48, 53, 54, 101, 102], "make": [0, 1, 4, 9, 10, 12, 32, 47, 50, 53, 55, 58, 63, 65, 67, 68, 70, 72, 73, 77, 79, 82, 84, 88, 91, 92, 94, 95, 96, 97, 102, 104, 105, 106, 107, 108, 113, 116, 117], "particip": 0, "project": [0, 10, 12, 26, 89, 90, 91, 92, 93, 94, 96, 98, 104, 110, 113, 116, 117, 120], "commun": [0, 1, 2, 100, 101, 102], "harass": 0, "free": [0, 1, 41, 90, 99, 100, 102, 109, 116], "experi": [0, 1, 47, 86, 100, 105, 108, 112, 113], "everyon": [0, 1], "regardless": [0, 12, 48, 49, 72, 85, 88], "ag": 0, "bodi": [0, 3, 47, 48, 52, 60, 69, 70, 72, 73, 84, 95, 97, 113, 116, 117, 118], "size": [0, 10, 34, 37, 38, 41, 42, 43, 44, 46, 47, 48, 49, 54, 61, 62, 65, 66, 68, 69, 72, 75, 76, 80, 83, 86, 88, 89, 95, 98, 104, 107], "disabl": [0, 72, 77, 109], "ethnic": 0, "gender": 0, "ident": [0, 6, 37, 42, 43, 47, 54, 65, 67, 68, 70, 72, 79, 80, 83, 84, 86, 87, 89, 92, 93, 95, 96, 103, 115, 119], "express": [0, 41], "level": [0, 10, 12, 14, 27, 32, 43, 44, 50, 61, 64, 69, 72, 74, 75, 80, 81, 89, 104, 105, 106, 113], "nation": 0, "person": [0, 41, 75, 84, 102, 118], "appear": [0, 10, 86, 89, 92, 94, 98, 116, 119], "race": 0, "religion": 0, "sexual": 0, "orient": [0, 32, 88, 92], "exampl": [0, 12, 20, 21, 25, 29, 32, 37, 44, 45, 46, 48, 49, 52, 53, 54, 56, 57, 60, 61, 64, 65, 66, 67, 68, 72, 73, 75, 76, 81, 82, 84, 86, 87, 88, 90, 91, 92, 96, 100, 104, 106, 110, 111, 113, 117], "behavior": [0, 1, 32, 51, 86, 92, 101, 113], "contribut": [0, 58, 88, 100], "creat": [0, 1, 4, 9, 10, 12, 13, 14, 17, 23, 25, 26, 32, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 49, 50, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 70, 72, 74, 75, 77, 78, 80, 81, 83, 84, 85, 87, 89, 92, 93, 94, 96, 97, 101, 102, 103, 104, 105, 106, 108, 109, 113, 120], "posit": [0, 12, 27, 42, 69, 88, 89, 92, 95, 104, 107, 116, 117], "includ": [0, 8, 9, 10, 12, 22, 23, 28, 32, 34, 37, 41, 42, 69, 72, 77, 85, 88, 89, 91, 92, 94, 96, 100, 101, 102, 104, 105, 107, 112, 113, 115], "us": [0, 1, 3, 4, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 53, 54, 55, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 101, 102, 103, 104, 105, 106, 107, 108, 109, 111, 113, 115, 116, 117, 118, 119, 120], "inclus": 0, "languag": [0, 87], "Being": 0, "respect": [0, 27, 34, 37, 42, 43, 47, 48, 49, 54, 55, 68, 73, 76, 86, 116], "differ": [0, 11, 12, 14, 25, 32, 37, 41, 42, 43, 44, 46, 48, 60, 63, 65, 69, 72, 75, 80, 82, 84, 88, 91, 92, 93, 95, 96, 99, 102, 103, 104, 105, 106, 107, 110, 116, 119], "viewpoint": 0, "gracefulli": 0, "accept": [0, 32, 72, 84], "construct": [0, 10, 12, 30, 37, 42, 74, 105], "critic": [0, 47, 72, 104, 105, 106, 107, 109], "focus": 0, "what": [0, 34, 45, 79, 83, 88, 91, 95, 96, 99, 102, 103, 104, 105, 107, 109, 110, 119], "i": [0, 1, 3, 4, 7, 8, 9, 10, 11, 12, 14, 16, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119], "best": [0, 1, 3, 6, 77, 78, 79, 92, 95, 96, 97, 116, 117], "show": [0, 10, 34, 72, 89, 92, 95, 103, 109, 110, 112, 116, 117, 119], "empathi": 0, "toward": 0, "other": [0, 3, 6, 10, 12, 23, 25, 28, 32, 41, 43, 48, 49, 50, 64, 65, 66, 67, 74, 78, 79, 80, 84, 85, 87, 89, 90, 91, 92, 96, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 109, 111, 112, 116, 117], "member": [0, 1], "unaccept": [0, 1], "The": [0, 1, 4, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 34, 36, 37, 38, 39, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 84, 85, 87, 88, 89, 90, 91, 92, 94, 96, 97, 98, 100, 102, 103, 104, 105, 106, 107, 109, 110, 111, 113, 114, 115, 116, 117, 119, 120], "imageri": 0, "unwelcom": 0, "attent": [0, 88], "advanc": [0, 101], "troll": 0, "insult": 0, "derogatori": 0, "comment": [0, 1], "polit": 0, "attack": 0, "public": 0, "privat": 0, "publish": [0, 35, 41, 49, 100, 101], "inform": [0, 12, 14, 27, 38, 40, 41, 43, 47, 50, 60, 72, 78, 87, 89, 91, 92, 94, 96, 99, 100, 104, 105, 106, 114, 116, 117, 119], "physic": [0, 77, 101], "electron": 0, "address": [0, 35, 45, 49], "without": [0, 9, 10, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 32, 35, 40, 41, 48, 49, 54, 62, 67, 69, 74, 75, 77, 78, 84, 89, 94, 96, 101, 102, 103, 106, 108, 112], "explicit": [0, 61, 66, 77], "permiss": [0, 41, 107, 109], "which": [0, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 34, 37, 38, 39, 44, 45, 47, 48, 49, 52, 53, 54, 57, 58, 60, 61, 63, 65, 66, 69, 70, 72, 73, 75, 76, 78, 79, 80, 81, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 100, 102, 103, 104, 105, 107, 110, 112, 115, 116, 117, 119], "could": [0, 12, 13, 32, 69, 91, 102, 104, 105, 107, 108, 109], "reason": [0, 48, 75], "consid": [0, 48, 69, 72, 73, 75, 80, 84, 89, 100], "inappropri": 0, "profession": 0, "set": [0, 1, 3, 4, 6, 9, 10, 12, 23, 26, 32, 37, 38, 40, 42, 43, 44, 46, 47, 48, 49, 51, 52, 53, 54, 55, 58, 60, 61, 62, 63, 65, 66, 68, 69, 70, 72, 75, 76, 79, 80, 81, 82, 85, 86, 89, 91, 92, 94, 95, 96, 97, 98, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 113, 116, 117, 119], "ar": [0, 1, 3, 4, 6, 7, 8, 9, 10, 11, 12, 16, 23, 25, 26, 27, 28, 31, 32, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 71, 72, 73, 75, 76, 77, 78, 79, 80, 81, 84, 85, 86, 87, 88, 89, 91, 92, 94, 95, 96, 97, 98, 99, 100, 103, 104, 105, 106, 107, 108, 109, 112, 114, 115, 116, 117, 119], "clarifi": 0, "expect": [0, 1, 9, 11, 12, 49, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 66, 67, 69, 70, 72, 75, 76, 87, 89, 103, 114], "take": [0, 4, 6, 25, 37, 41, 42, 44, 48, 53, 54, 55, 58, 72, 76, 80, 84, 85, 90, 91, 92, 95, 98, 103, 104, 105, 106, 107, 108, 111, 116, 117, 119], "appropri": [0, 13, 14, 15, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 37, 46, 64, 69, 72, 79, 81, 92, 97, 98, 104, 105, 106, 107, 109, 116], "fair": 0, "correct": [0, 12, 25, 69, 70, 75, 92, 93, 94, 95, 97, 107, 113, 115, 117, 119], "action": [0, 41, 45, 92], "ani": [0, 1, 4, 6, 9, 10, 12, 20, 21, 25, 27, 29, 32, 34, 35, 37, 41, 42, 45, 46, 47, 48, 49, 50, 54, 55, 59, 60, 61, 62, 64, 65, 66, 67, 68, 69, 72, 73, 74, 75, 79, 80, 81, 84, 85, 89, 90, 92, 94, 95, 96, 97, 99, 100, 101, 102, 103, 107, 108, 112, 113, 116, 119], "instanc": [0, 3, 6, 7, 8, 9, 12, 14, 23, 25, 27, 30, 32, 34, 38, 40, 41, 42, 43, 44, 46, 47, 48, 50, 52, 53, 55, 58, 59, 60, 61, 62, 64, 65, 66, 67, 69, 70, 71, 72, 73, 74, 75, 78, 79, 80, 81, 83, 84, 86, 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 103, 104, 105, 107, 108, 109, 113, 115, 116, 117, 120], "have": [0, 1, 3, 4, 8, 9, 10, 12, 23, 25, 28, 31, 32, 35, 37, 38, 41, 42, 43, 46, 47, 48, 49, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 65, 66, 67, 68, 69, 71, 72, 75, 78, 79, 80, 84, 85, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 102, 103, 104, 105, 106, 107, 108, 110, 112, 115, 116, 117, 118, 119], "right": [0, 1, 41, 46, 48, 52, 66, 84, 87, 92, 95, 97, 103, 113, 116, 118, 119], "remov": [0, 9, 10, 12, 65, 69, 71, 78, 79, 80, 84, 85, 89, 94, 95, 97, 102, 108, 110], "edit": [0, 50, 91, 92, 94, 95, 96, 102, 106, 112, 115, 119], "reject": [0, 72, 75], "commit": [0, 1], "wiki": 0, "issu": [0, 66, 99, 100, 101, 102, 104], "align": [0, 10, 53, 66, 72, 76], "thi": [0, 1, 3, 4, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 113, 114, 115, 116, 117, 119, 120], "ban": 0, "temporarili": 0, "perman": 0, "thei": [0, 1, 10, 11, 12, 23, 24, 37, 44, 46, 47, 48, 49, 50, 55, 65, 66, 68, 72, 73, 75, 77, 80, 84, 85, 90, 91, 92, 94, 95, 97, 102, 103, 104, 105, 106, 107, 108, 114], "deem": 0, "threaten": 0, "offens": 0, "harm": 0, "appli": [0, 4, 7, 10, 31, 37, 38, 43, 44, 46, 47, 48, 52, 54, 61, 62, 66, 72, 73, 75, 76, 80, 89, 92, 95, 97, 102, 108, 113, 119], "both": [0, 1, 12, 14, 23, 38, 46, 47, 49, 53, 55, 61, 66, 67, 68, 69, 72, 78, 79, 87, 88, 91, 94, 95, 96, 106, 117], "within": [0, 9, 10, 12, 26, 27, 32, 37, 38, 44, 46, 47, 49, 54, 61, 65, 67, 69, 71, 72, 75, 77, 84, 85, 90, 92, 96, 102, 104, 108, 112], "space": [0, 10, 47, 49, 68, 75, 90, 92, 100, 102, 116, 118], "when": [0, 1, 10, 12, 23, 27, 32, 34, 37, 42, 44, 46, 47, 48, 49, 52, 53, 54, 55, 60, 61, 62, 65, 68, 69, 72, 73, 75, 77, 78, 80, 84, 88, 89, 90, 92, 93, 94, 95, 96, 97, 98, 99, 100, 102, 103, 104, 106, 108, 109, 110, 112, 115, 116, 117, 118, 119], "individu": [0, 10, 14, 17, 32, 41, 48, 53, 54, 61, 63, 64, 65, 73, 80, 89, 95, 104, 108, 117], "repres": [0, 4, 10, 12, 14, 32, 34, 41, 47, 58, 65, 69, 75, 76, 80, 84, 85, 92, 97, 99, 104], "its": [0, 4, 12, 25, 32, 48, 54, 60, 61, 62, 66, 72, 73, 75, 77, 79, 84, 88, 92, 95, 96, 100, 101, 102, 113, 116, 117, 118, 119], "offici": [0, 87, 100, 102], "e": [0, 1, 3, 6, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 36, 37, 42, 44, 47, 48, 49, 52, 53, 54, 55, 58, 60, 61, 65, 66, 67, 68, 69, 71, 72, 73, 76, 77, 78, 79, 80, 84, 85, 87, 88, 89, 91, 92, 93, 94, 95, 96, 101, 102, 104, 108, 109, 110, 113, 118], "mail": 0, "post": [0, 47, 80, 82, 86, 100, 103], "via": [0, 1, 12, 37, 44, 47, 54, 72, 75, 76, 89, 92, 95, 101, 102, 103, 105, 106, 109, 110, 112, 118], "social": 0, "media": [0, 32], "account": [0, 52, 53, 55, 58, 61, 69, 90, 110, 116], "act": [0, 1, 35], "appoint": 0, "onlin": [0, 48, 73, 88], "offlin": 0, "event": [0, 41, 45], "represent": [0, 12, 40, 42, 43, 47, 50, 72, 75, 76, 84], "mai": [0, 1, 11, 12, 20, 34, 37, 38, 40, 42, 43, 45, 46, 47, 48, 49, 50, 52, 54, 60, 61, 65, 67, 69, 72, 73, 75, 88, 89, 92, 95, 96, 99, 100, 103, 104, 105, 107, 112, 113, 114, 116, 117, 119], "further": [0, 48, 88, 93, 113], "defin": [0, 10, 14, 27, 37, 38, 42, 43, 44, 47, 52, 53, 55, 58, 64, 68, 70, 72, 74, 75, 78, 84, 96, 97, 103, 104, 105, 106, 113], "abus": 0, "otherwis": [0, 1, 9, 10, 12, 14, 17, 32, 37, 41, 42, 54, 62, 64, 68, 72, 84, 89, 92, 96, 102, 103, 110, 116], "report": [0, 1, 12, 72, 86, 89, 100], "contact": [0, 89], "team": [0, 1], "talmo": [0, 1, 30, 86, 99, 100, 101], "salk": [0, 1, 99, 100, 101], "edu": [0, 1, 99, 100, 101], "all": [0, 1, 3, 4, 9, 10, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 37, 38, 41, 44, 46, 47, 48, 49, 50, 53, 55, 58, 60, 61, 64, 65, 66, 67, 69, 72, 73, 74, 75, 77, 78, 79, 80, 81, 84, 86, 88, 90, 92, 94, 95, 96, 97, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 110, 113, 115, 116, 117, 118, 119], "complaint": 0, "review": [0, 1, 12, 80, 95, 100, 101, 120], "investig": 0, "result": [0, 3, 4, 10, 12, 14, 17, 20, 27, 32, 37, 38, 42, 43, 44, 45, 47, 48, 57, 61, 67, 72, 73, 80, 81, 84, 85, 87, 88, 89, 90, 92, 95, 99, 103, 104, 105, 106, 108, 112], "necessari": [0, 12, 72, 81, 91, 96, 100, 102, 104, 106], "circumst": 0, "oblig": 0, "confidenti": 0, "regard": 0, "incid": 0, "detail": [0, 7, 12, 15, 37, 38, 40, 43, 69, 72, 75, 87, 88, 89, 91, 95, 100, 113, 115], "specif": [0, 10, 12, 13, 14, 25, 26, 44, 47, 49, 50, 62, 67, 72, 73, 78, 81, 84, 92, 98, 100, 105], "polici": [0, 77], "separ": [0, 1, 12, 23, 26, 49, 72, 88, 89, 94, 96, 103, 104, 105, 108, 110, 117], "who": [0, 41], "do": [0, 10, 12, 37, 41, 43, 44, 47, 49, 60, 72, 73, 78, 80, 92, 94, 95, 97, 102, 103, 104, 105, 106, 107, 108, 119], "follow": [0, 1, 12, 27, 32, 37, 41, 42, 47, 54, 61, 69, 84, 86, 87, 89, 97, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109], "good": [0, 1, 21, 25, 29, 35, 48, 69, 78, 79, 80, 88, 91, 93, 95, 99, 112, 117], "faith": 0, "face": [0, 14, 50, 99], "temporari": [0, 116], "repercuss": 0, "determin": [0, 6, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 32, 34, 38, 40, 43, 44, 46, 47, 49, 65, 66, 69, 70, 72, 78, 79, 80, 92, 95, 98, 109, 116], "": [0, 1, 10, 12, 14, 16, 17, 20, 23, 25, 28, 30, 32, 34, 35, 37, 39, 41, 42, 44, 46, 49, 53, 54, 55, 58, 60, 62, 64, 65, 66, 68, 69, 72, 74, 75, 78, 79, 80, 82, 84, 85, 86, 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 113, 116, 117, 119], "leadership": 0, "adapt": [0, 42], "from": [0, 1, 2, 3, 4, 6, 7, 9, 10, 12, 13, 14, 16, 17, 20, 23, 25, 26, 27, 28, 30, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 52, 53, 54, 58, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 88, 89, 91, 92, 93, 94, 95, 96, 98, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119], "version": [0, 1, 12, 47, 50, 64, 68, 85, 89, 94, 96, 100, 101, 102, 104, 105, 106, 107, 108, 109, 110, 112], "1": [0, 3, 4, 9, 10, 11, 12, 26, 32, 34, 35, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 53, 54, 55, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 72, 73, 75, 76, 78, 79, 80, 82, 83, 84, 86, 89, 95, 98, 102, 107, 109, 110, 119], "4": [0, 3, 12, 32, 34, 38, 39, 41, 42, 43, 44, 47, 48, 50, 53, 55, 61, 69, 72, 76, 79, 82, 83, 86, 89, 92, 98, 100, 101, 102, 103, 104, 107, 109, 110], "avail": [0, 1, 9, 32, 33, 37, 48, 61, 67, 69, 72, 77, 88, 89, 90, 96, 102, 103, 104, 105, 106, 107, 108, 109], "http": [0, 15, 37, 39, 42, 72, 78, 82, 102, 103, 104, 105, 106, 107, 108, 109], "org": [0, 15, 27, 37, 39, 42, 72, 104, 107, 108], "As": [1, 73, 88, 92, 94, 95, 106, 109, 110, 116], "our": [1, 2, 79, 84, 90, 91, 96, 100, 101, 102, 103, 104, 105, 107, 109, 117, 118, 120], "grow": 1, "import": [1, 9, 12, 14, 17, 65, 69, 75, 87, 88, 92, 93, 95, 97, 100, 103, 104, 105, 106, 107, 108, 109, 113, 114, 118], "adher": 1, "guidelin": [1, 100], "These": [1, 12, 28, 35, 37, 38, 40, 41, 43, 44, 46, 47, 49, 50, 53, 54, 55, 60, 65, 67, 68, 69, 71, 72, 75, 86, 88, 90, 92, 95, 102, 103, 104, 105, 106, 107, 113, 119], "chang": [1, 12, 31, 38, 44, 45, 46, 47, 48, 61, 62, 65, 67, 72, 77, 84, 92, 95, 97, 102, 105, 106, 116, 119], "need": [1, 11, 12, 20, 25, 32, 34, 41, 43, 46, 47, 49, 60, 62, 65, 66, 69, 73, 75, 79, 80, 84, 85, 86, 87, 88, 89, 90, 92, 93, 94, 95, 96, 97, 102, 103, 104, 105, 106, 107, 109, 110, 113, 114, 115, 116, 118, 119], "pleas": [1, 86, 87, 100, 101, 104, 107, 108], "feel": [1, 99, 100, 116], "propos": 1, "sourc": [1, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 87, 95, 97, 100, 101, 102, 104, 107], "govern": 1, "abid": 1, "github": [1, 99, 101, 102, 103, 108, 109], "ha": [1, 3, 9, 10, 11, 12, 25, 26, 32, 37, 38, 39, 42, 43, 46, 47, 49, 51, 54, 60, 64, 67, 68, 72, 74, 75, 77, 78, 80, 84, 87, 88, 89, 90, 92, 95, 96, 98, 102, 109, 115, 116, 117, 119], "made": [1, 41, 95, 101, 119], "easi": [1, 4, 47, 48, 69, 73, 74, 77, 91, 95, 96, 97, 101, 104, 107], "gener": [1, 4, 9, 12, 21, 26, 27, 29, 34, 37, 38, 40, 42, 43, 44, 46, 47, 48, 49, 50, 52, 53, 54, 55, 58, 61, 63, 64, 65, 67, 68, 69, 72, 73, 75, 76, 78, 80, 81, 82, 83, 87, 88, 89, 92, 94, 95, 96, 97, 100, 101, 102, 104, 105, 106, 107, 108, 109, 110, 112, 113, 115, 116], "speak": 1, "someth": [1, 3, 25, 87, 92], "warrant": 1, "g": [1, 6, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 36, 44, 47, 49, 52, 54, 61, 65, 66, 69, 71, 72, 73, 77, 78, 84, 85, 89, 90, 92, 95, 96, 101, 102, 103, 104, 108, 113], "bug": 1, "fix": [1, 3, 30, 32, 42, 45, 46, 47, 49, 62, 66, 68, 85, 95, 120], "todo": [1, 14, 42, 49], "should": [1, 10, 12, 23, 25, 32, 38, 42, 44, 45, 46, 47, 48, 49, 50, 52, 54, 60, 67, 69, 71, 72, 73, 75, 76, 78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89, 91, 92, 94, 95, 97, 98, 102, 103, 104, 105, 106, 108, 113, 115, 117], "submit": 1, "convers": [1, 32, 89], "question": [1, 93, 100], "about": [1, 4, 5, 12, 15, 47, 61, 72, 73, 89, 91, 92, 93, 94, 96, 97, 99, 100, 105, 106, 107, 108, 114, 115, 116, 117, 119], "usag": [1, 19, 25, 33, 48, 61, 72, 89, 101], "improv": [1, 38, 47, 48, 49, 54, 61, 72, 73, 80, 98, 99, 106, 109, 110], "idea": [1, 79, 91, 100, 115, 117], "go": [1, 33, 54, 79, 80, 95, 96, 105, 116, 117], "stage": [1, 61, 72, 99, 104, 116, 117], "area": [1, 69, 82, 92, 99], "befor": [1, 8, 9, 10, 12, 31, 32, 34, 37, 43, 44, 47, 48, 49, 54, 61, 66, 68, 72, 74, 76, 79, 80, 81, 84, 87, 89, 91, 92, 94, 96, 104, 105, 106, 109, 110, 115, 116], "ad": [1, 10, 12, 37, 38, 43, 48, 61, 62, 66, 68, 72, 73, 90, 92, 94, 95, 97, 107, 109, 115, 116], "item": [1, 4, 10, 12, 65, 85, 92, 102, 105], "roadmap": 1, "onc": [1, 47, 54, 77, 87, 90, 91, 94, 95, 96, 99, 100, 102, 104, 105, 106, 109, 110, 112, 113, 115, 117, 118, 119], "can": [1, 3, 4, 7, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 34, 37, 38, 41, 42, 44, 46, 47, 48, 49, 50, 52, 53, 54, 55, 58, 61, 63, 64, 65, 68, 69, 72, 73, 75, 76, 77, 79, 80, 84, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 119, 120], "given": [1, 3, 6, 9, 10, 12, 23, 25, 26, 31, 32, 47, 61, 69, 72, 74, 75, 79, 80, 83, 84, 88, 89, 107, 108], "prioriti": [1, 90], "statu": 1, "assigne": 1, "etc": [1, 4, 12, 32, 44, 85, 102, 104, 117], "check": [1, 10, 12, 14, 28, 31, 32, 33, 35, 49, 51, 64, 72, 77, 80, 84, 85, 89, 92, 95, 100, 101, 102, 104, 109, 110], "close": [1, 20, 32, 33, 47, 79, 105, 112, 116], "help": [1, 2, 37, 44, 47, 48, 65, 72, 73, 89, 95, 99, 101, 107, 108, 113], "page": [1, 2, 101, 102, 109, 110], "sure": [1, 10, 12, 46, 79, 82, 91, 96, 97, 102, 104, 105, 106, 107, 108], "doesn": [1, 3, 10, 12, 32, 54, 85, 91, 92, 96, 97, 98, 115], "t": [1, 3, 10, 11, 12, 20, 25, 28, 31, 32, 35, 44, 47, 54, 73, 78, 79, 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 101, 102, 103, 104, 105, 106, 107, 109, 110, 112, 113, 115, 116, 117, 118], "alreadi": [1, 8, 10, 12, 20, 30, 32, 46, 49, 62, 67, 68, 72, 77, 79, 80, 84, 89, 90, 91, 92, 95, 96, 97, 98, 102, 103, 106, 119], "exist": [1, 12, 25, 26, 31, 32, 37, 38, 45, 49, 52, 84, 85, 92, 96, 97, 102, 106, 108, 112], "been": [1, 8, 10, 12, 35, 74, 77, 79, 80, 84, 89, 92, 95, 102, 105, 115, 116], "solv": 1, "new": [1, 4, 10, 12, 13, 23, 25, 31, 32, 37, 39, 49, 54, 64, 66, 67, 72, 78, 80, 81, 84, 89, 90, 91, 92, 94, 95, 96, 97, 98, 100, 101, 102, 106, 107, 110, 112, 113, 115, 116, 118, 119], "templat": [1, 25, 84, 96], "place": [1, 12, 54, 65, 69, 78, 85, 90, 92, 93, 95, 96, 110, 112, 113, 119], "ask": [1, 92, 101, 116, 118], "3": [1, 3, 10, 12, 26, 32, 37, 38, 40, 41, 42, 43, 47, 48, 53, 55, 60, 61, 62, 65, 69, 75, 76, 79, 80, 82, 85, 86, 89, 95, 100, 102, 103, 104, 107, 109, 110, 119], "categori": 1, "troubl": [1, 117], "softwar": [1, 41, 90, 100, 101], "user": [1, 10, 12, 24, 31, 46, 49, 50, 61, 65, 85, 89, 92, 100, 109, 113], "enhanc": 1, "thing": [1, 73, 84, 85, 89, 90, 92, 96, 103, 116], "would": [1, 25, 54, 61, 66, 67, 72, 76, 78, 79, 88, 92, 94, 100, 102, 104, 106, 107, 108, 110, 116], "nicer": 1, "necessarili": [1, 75], "problem": [1, 75, 89, 95, 99, 102, 115], "If": [1, 6, 7, 9, 10, 11, 12, 14, 16, 17, 23, 24, 25, 26, 28, 31, 32, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 58, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 72, 74, 75, 76, 77, 78, 79, 80, 81, 83, 84, 85, 88, 89, 90, 91, 92, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 108, 109, 110, 111, 116, 117, 118, 119], "fall": 1, "goe": [1, 95, 119], "here": [1, 20, 32, 69, 72, 76, 79, 85, 86, 88, 92, 95, 96, 97, 100, 102, 103, 106, 107, 108, 110, 111, 112, 113, 114], "long": [1, 49, 54, 71, 90, 104, 105], "isn": [1, 32, 88, 91, 92, 96, 100], "instal": [1, 23, 27, 89, 90, 96, 98, 103, 104, 105, 106, 107, 112, 114, 118], "develop": [1, 50, 100, 101, 102, 105, 112], "branch": [1, 37, 47], "instruct": [1, 87, 96, 100, 101, 102, 104, 105, 106, 107, 109, 118], "conda": [1, 100, 101, 102, 103, 104, 105, 106, 108, 109, 110, 118], "env": [1, 102, 104, 107, 108, 109], "fork": 1, "either": [1, 9, 10, 17, 32, 34, 35, 49, 53, 58, 61, 66, 67, 69, 72, 76, 84, 89, 92, 94, 95, 96, 102, 116], "work": [1, 3, 11, 12, 25, 27, 31, 32, 41, 49, 54, 72, 73, 77, 85, 86, 88, 89, 90, 95, 96, 98, 99, 100, 101, 103, 110, 112, 113, 114, 116, 117], "recommend": [1, 49, 54, 72, 99, 100, 101, 104, 105, 106, 107, 108, 117, 118], "tackl": 1, "multipl": [1, 10, 12, 14, 17, 19, 20, 25, 32, 35, 37, 38, 46, 47, 49, 53, 55, 61, 64, 66, 67, 71, 75, 77, 78, 79, 84, 89, 91, 92, 94, 105, 106, 109, 116, 117], "time": [1, 10, 30, 34, 35, 37, 38, 47, 48, 49, 50, 54, 72, 80, 81, 84, 89, 90, 96, 98, 100, 102, 105, 106, 107, 108, 115, 116], "your": [1, 23, 25, 69, 88, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 113, 115, 116, 117, 118, 119, 120], "name": [1, 7, 9, 10, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 32, 33, 36, 37, 38, 41, 42, 45, 46, 47, 48, 49, 50, 52, 56, 60, 61, 62, 66, 67, 69, 70, 72, 73, 75, 77, 81, 84, 85, 86, 89, 92, 96, 97, 100, 102, 103, 104, 105, 106, 107, 109, 110, 116, 118, 119], "relev": [1, 13, 14, 72, 92, 119], "keyword": [1, 34, 42, 65], "eg": 1, "git": [1, 102], "checkout": 1, "b": [1, 6, 61, 84, 85, 92, 102], "john": [1, 86, 100, 101], "some_issu": 1, "some": [1, 3, 7, 12, 20, 28, 30, 43, 54, 64, 72, 75, 77, 84, 86, 88, 92, 93, 94, 95, 96, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 115, 116, 117], "addit": [1, 12, 25, 37, 43, 44, 47, 48, 50, 61, 72, 78, 96, 99, 103, 107, 113, 117], "write": [1, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 33, 34, 45, 49, 72, 85], "test": [1, 12, 32, 46, 48, 49, 69, 72, 81, 86, 89, 104, 105, 106, 107, 109], "draft": 1, "pr": 1, "get": [1, 2, 3, 4, 7, 9, 10, 12, 16, 28, 32, 34, 35, 37, 41, 44, 61, 64, 75, 77, 78, 84, 85, 88, 89, 90, 92, 94, 95, 96, 98, 99, 103, 104, 105, 106, 108, 109, 110, 112, 115, 116, 117, 119], "coverag": 1, "statist": [1, 7, 12, 49, 73], "codecov": 1, "narrow": 1, "down": [1, 40, 43, 47, 48, 49, 71, 72, 75, 88, 89, 91, 95, 96, 100, 101, 104, 105, 109, 116, 117, 118], "error": [1, 32, 51, 69, 75, 77, 78, 98, 100, 102, 104, 105, 106, 107, 108, 109, 113, 115], "prone": 1, "line": [1, 3, 5, 9, 11, 34, 47, 72, 75, 91, 92, 93, 94, 95, 96, 98, 101, 109, 112, 115], "subtest": 1, "folder": [1, 12, 32, 45, 47, 49, 50, 69, 72, 81, 87, 89, 90, 91, 104, 105, 106, 107, 109, 118], "aim": 1, "hit": [1, 95, 110, 117, 118], "everi": [1, 3, 6, 12, 26, 37, 38, 42, 44, 47, 48, 49, 54, 69, 78, 89, 92, 95], "point": [1, 3, 6, 7, 9, 10, 12, 14, 34, 47, 52, 53, 55, 58, 60, 61, 63, 64, 65, 66, 69, 71, 72, 75, 76, 78, 80, 82, 85, 89, 91, 95, 98, 104, 105, 107, 116, 117], "cover": [1, 47, 91, 95, 100], "edg": [1, 3, 9, 34, 47, 48, 55, 70, 75, 83, 84, 89, 92, 95, 97, 104, 118], "case": [1, 8, 12, 25, 32, 42, 61, 72, 88, 89, 90, 91, 94, 95, 96, 99, 103, 106, 116, 119], "abil": 1, "try": [1, 12, 25, 32, 79, 80, 85, 88, 92, 95, 96, 97, 98, 99, 100, 102, 103, 105, 112, 116, 117, 120], "add": [1, 3, 4, 10, 12, 13, 25, 39, 43, 47, 48, 60, 61, 67, 73, 78, 84, 87, 92, 94, 97, 115, 116, 118], "file": [1, 5, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 41, 42, 46, 47, 49, 50, 65, 69, 72, 81, 84, 85, 86, 87, 90, 91, 94, 95, 96, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 118], "push": 1, "origin": [1, 9, 10, 12, 32, 34, 37, 38, 40, 43, 53, 55, 65, 66, 85, 88, 89, 93, 94, 96, 98, 102, 115], "pass": [1, 10, 11, 20, 32, 34, 37, 45, 59, 72, 75, 95], "fail": [1, 48, 49, 54, 69, 89, 95, 100, 102, 109, 110, 119], "repeat": [1, 37, 38, 47, 54, 66, 84, 113, 115, 118], "step": [1, 3, 12, 37, 38, 39, 40, 41, 42, 43, 44, 47, 48, 53, 55, 58, 62, 66, 68, 72, 75, 79, 80, 88, 92, 95, 99, 100, 101, 102, 105, 106, 109, 113, 117], "5": [1, 4, 6, 12, 26, 36, 37, 41, 42, 47, 48, 52, 59, 61, 62, 69, 70, 72, 73, 75, 76, 80, 82, 85, 86, 89, 102, 103, 104, 105, 107, 108, 109, 110], "look": [1, 25, 32, 37, 55, 60, 69, 80, 85, 87, 88, 89, 90, 91, 92, 95, 107, 111, 116, 117], "convert": [1, 4, 10, 12, 20, 25, 32, 42, 46, 54, 61, 62, 65, 66, 68, 72, 75, 84, 85, 87, 92, 94, 100, 104, 105, 114, 115], "final": [1, 37, 39, 42, 44, 47, 54, 64, 95, 102, 104, 113, 116], "readi": [1, 72, 106, 112, 115, 117, 120], "button": [1, 87, 91, 98, 109, 110, 116, 117, 118], "onli": [1, 9, 10, 12, 32, 34, 38, 42, 43, 45, 46, 47, 48, 49, 51, 61, 64, 65, 67, 72, 75, 77, 78, 79, 80, 84, 87, 89, 90, 91, 92, 94, 95, 96, 98, 100, 101, 102, 105, 106, 107, 108, 109, 110, 113], "believ": 1, "merg": [1, 10, 12, 23, 41, 78, 80, 89, 92, 94, 96, 110, 119], "option": [1, 4, 6, 8, 9, 10, 11, 12, 13, 14, 16, 17, 20, 22, 23, 25, 26, 27, 28, 30, 31, 32, 34, 35, 37, 38, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 52, 55, 59, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 72, 74, 75, 76, 77, 78, 80, 81, 82, 83, 84, 85, 88, 89, 92, 95, 96, 102, 105, 108, 113, 116, 119, 120], "assign": [1, 10, 12, 47, 65, 67, 71, 72, 75, 79, 80, 82, 92, 95, 119], "screen": 1, "self": [1, 80, 81, 86, 105, 110], "themselv": 1, "re": [1, 8, 11, 12, 25, 80, 89, 90, 91, 92, 93, 94, 95, 96, 97, 99, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 114, 115, 117, 119], "sign": 1, "off": [1, 45, 69, 72, 90, 95, 99, 104, 106, 107], "squash": 1, "feautur": 1, "main": [1, 9, 11, 12, 38, 47, 72, 74, 78, 81, 84, 92, 95, 103, 108, 115, 118, 119], "releas": [1, 101, 102, 109, 110], "lint": 1, "black": [1, 34, 89, 115], "21": [1, 33, 69, 80, 86, 89, 102, 104, 105, 106, 107, 108, 109], "6b0": 1, "see": [1, 7, 10, 12, 14, 15, 19, 20, 27, 32, 37, 38, 40, 41, 42, 43, 44, 45, 47, 48, 53, 54, 58, 61, 62, 63, 66, 68, 69, 71, 72, 73, 75, 76, 77, 78, 82, 84, 86, 87, 88, 89, 91, 92, 94, 95, 96, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 113, 114, 115, 116, 117, 118, 119], "dev_requir": 1, "pep8": 1, "type": [1, 4, 6, 9, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 52, 53, 54, 55, 56, 58, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 72, 73, 74, 75, 78, 80, 81, 84, 85, 89, 91, 93, 95, 96, 97, 99, 100, 101, 102, 104, 105, 106, 107, 109, 112, 113, 117, 118, 119], "hint": [1, 105, 106, 109, 110], "encourag": [1, 47, 48], "document": [1, 41, 42, 50, 75, 87, 89, 108], "googl": [1, 90, 102, 103, 106, 108, 109], "docstr": [1, 38, 40, 43], "heart": 1, "head": [2, 3, 12, 26, 37, 39, 42, 44, 47, 64, 72, 74, 75, 103, 104, 105, 106, 109], "discuss": [2, 95, 100, 101, 102], "sleap": [2, 86, 88, 90, 91, 92, 94, 95, 97, 98, 99, 102, 103, 107, 111, 113, 114, 115, 116, 117, 120], "function": [3, 6, 7, 9, 12, 13, 14, 23, 25, 32, 34, 37, 42, 45, 47, 50, 54, 56, 61, 64, 66, 68, 69, 70, 71, 72, 73, 75, 76, 77, 78, 80, 81, 82, 84, 85, 89, 92, 93, 95, 98, 105, 106, 107, 112], "usual": [3, 25, 32, 79, 88, 92, 94, 95, 96, 106], "you": [3, 8, 11, 12, 13, 23, 25, 32, 34, 35, 49, 69, 73, 79, 80, 84, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120], "ll": [3, 7, 10, 12, 13, 14, 17, 23, 25, 31, 35, 69, 80, 84, 85, 88, 90, 91, 92, 94, 95, 96, 97, 98, 100, 103, 104, 105, 106, 107, 108, 109, 110, 112, 114, 115, 116, 118, 119], "want": [3, 10, 11, 12, 25, 31, 32, 34, 35, 44, 69, 78, 80, 84, 88, 91, 92, 93, 94, 95, 96, 97, 100, 103, 105, 106, 107, 108, 110, 111, 112, 114, 115, 116, 117, 119], "find": [3, 6, 10, 12, 14, 17, 24, 31, 32, 44, 47, 48, 56, 59, 60, 69, 72, 75, 76, 78, 79, 84, 89, 90, 92, 95, 101, 105, 110, 117, 119], "out": [3, 10, 12, 27, 41, 46, 48, 49, 54, 64, 72, 75, 88, 90, 92, 95, 98, 99, 101, 102, 104, 105, 106, 107, 108], "skeleton": [3, 10, 12, 14, 26, 27, 28, 46, 47, 48, 52, 55, 60, 65, 72, 74, 75, 78, 79, 83, 86, 88, 92, 93, 95, 103, 104, 105, 108, 109, 113, 116, 120], "calcul": [3, 6, 7, 60, 61, 66, 68, 69, 73, 78, 82, 107], "mean": [3, 4, 7, 11, 12, 32, 33, 42, 46, 48, 53, 62, 69, 73, 78, 79, 89, 90, 92, 95, 100, 102, 103, 105, 107, 114, 117], "std": [3, 105], "node": [3, 6, 7, 9, 10, 11, 12, 14, 26, 27, 28, 34, 46, 47, 52, 53, 55, 60, 61, 69, 70, 71, 72, 75, 80, 84, 86, 88, 89, 92, 94, 95, 96, 97, 98, 103, 104, 105, 106, 107, 108, 109, 115, 116, 118], "locat": [3, 7, 9, 10, 12, 14, 28, 32, 34, 50, 63, 69, 72, 76, 88, 90, 92, 95, 97, 103, 107, 109, 110, 116, 117], "For": [3, 6, 10, 12, 23, 25, 27, 32, 37, 41, 49, 61, 69, 72, 75, 76, 78, 80, 84, 86, 87, 88, 89, 90, 91, 92, 96, 98, 99, 100, 101, 102, 103, 104, 105, 107, 109, 110, 113, 115, 116, 117, 118, 119, 120], "most": [3, 4, 8, 12, 33, 35, 37, 39, 44, 48, 71, 72, 73, 88, 90, 92, 96, 100, 102, 105, 106, 107, 113, 115, 119], "stabl": [3, 88, 95, 117], "smallest": [3, 31, 46, 66], "length": [3, 10, 41, 54, 61, 66, 67, 68, 69, 72, 75, 76, 78, 82, 89, 92, 95, 103, 104, 105, 109], "abov": [3, 11, 41, 48, 49, 72, 75, 86, 91, 92, 95, 100, 102, 104, 107, 108], "threshold": [3, 6, 58, 63, 69, 70, 75, 76, 78, 79, 80, 89, 92, 95, 107], "like": [3, 12, 32, 35, 37, 41, 44, 46, 49, 69, 73, 79, 80, 87, 88, 90, 92, 93, 94, 95, 96, 98, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110, 114, 116, 118], "thorax": [3, 104, 105, 106, 109], "between": [3, 6, 7, 9, 10, 11, 12, 34, 35, 42, 46, 47, 55, 60, 66, 67, 69, 72, 75, 78, 80, 82, 84, 89, 92, 95, 97, 102, 107, 116, 119], "two": [3, 6, 10, 12, 37, 50, 53, 69, 72, 75, 78, 79, 82, 84, 85, 87, 88, 89, 92, 94, 95, 96, 103, 104, 109, 110, 115, 117, 118, 119], "part": [3, 31, 46, 47, 48, 52, 55, 60, 64, 69, 70, 72, 73, 75, 84, 88, 89, 92, 96, 97, 113, 116, 117, 118], "rel": [3, 10, 36, 37, 39, 41, 42, 44, 47, 48, 49, 53, 55, 58, 60, 61, 65, 72, 75, 76, 88, 95, 97, 98, 117], "each": [3, 4, 6, 7, 10, 12, 23, 26, 27, 31, 32, 35, 37, 38, 39, 40, 41, 42, 43, 44, 45, 47, 48, 49, 52, 53, 54, 55, 56, 58, 60, 61, 63, 64, 65, 67, 69, 70, 71, 72, 73, 75, 76, 77, 78, 80, 82, 84, 88, 89, 90, 91, 92, 95, 96, 97, 98, 102, 103, 104, 105, 108, 109, 110, 113, 117], "thu": 3, "well": [3, 12, 32, 38, 43, 44, 48, 49, 50, 53, 55, 58, 64, 69, 72, 73, 90, 92, 95, 98, 99, 102, 103, 104, 106, 107, 110, 112, 115], "axi": [3, 42, 53, 54, 55, 60, 61, 62, 63, 68, 72, 75, 76, 82, 103, 105], "2": [3, 6, 9, 10, 11, 12, 33, 36, 37, 38, 40, 42, 43, 44, 47, 48, 52, 53, 54, 55, 58, 59, 60, 61, 62, 63, 65, 66, 69, 70, 71, 72, 73, 75, 76, 79, 80, 82, 83, 84, 86, 89, 92, 95, 100, 102, 103, 107, 109, 119], "fairli": [3, 86, 91], "straightforward": [3, 97], "angl": [3, 48], "found": [3, 12, 24, 25, 32, 37, 54, 69, 72, 75, 80, 84, 85, 87, 95, 98, 102, 103, 105, 106, 109, 110, 117], "rotat": [3, 48, 88, 92, 103, 106, 109, 116], "accordingli": [3, 12, 48, 110], "standard": [3, 34, 37, 42, 44, 46, 47, 48, 53, 54, 55, 58, 62, 69, 76, 78, 79, 86, 89, 96, 104, 105], "deviat": [3, 46, 48, 53, 55, 58, 62, 69, 76], "matrix": [3, 6, 9, 10, 32, 53, 69, 78, 79, 80, 82, 103, 105, 106, 109], "note": [3, 4, 9, 11, 22, 23, 28, 32, 44, 48, 54, 73, 84, 85, 92, 94, 98, 100, 102, 103, 104, 105, 107, 108, 109, 111, 114, 116, 119], "vector": [3, 4, 47, 53, 55, 58, 63, 64, 68, 71, 72, 75, 76, 82], "matric": [3, 6, 9, 103, 113], "shape": [3, 9, 10, 11, 12, 32, 34, 42, 52, 53, 54, 55, 58, 60, 61, 62, 63, 65, 66, 68, 69, 70, 71, 72, 73, 74, 75, 76, 78, 82, 95, 103, 104, 105, 106, 107, 109, 114], "where": [3, 10, 12, 26, 32, 38, 42, 44, 48, 49, 50, 52, 53, 55, 56, 58, 60, 61, 63, 66, 69, 72, 73, 75, 76, 78, 79, 82, 84, 85, 88, 89, 91, 92, 95, 96, 98, 102, 106, 113, 115, 116], "correspond": [3, 4, 6, 7, 9, 10, 12, 31, 35, 37, 53, 55, 61, 62, 63, 69, 71, 72, 75, 76, 80, 82, 84, 88, 92, 95, 104, 108], "x": [3, 10, 12, 14, 37, 38, 42, 44, 48, 52, 53, 55, 61, 63, 68, 72, 75, 76, 82, 86, 92, 102, 104, 105, 106, 109], "y": [3, 10, 12, 14, 48, 52, 53, 55, 61, 63, 68, 72, 75, 76, 92, 100, 101, 102, 104, 105, 106, 107, 108, 109, 110], "after": [3, 12, 37, 38, 40, 42, 44, 46, 47, 48, 49, 54, 64, 66, 68, 70, 72, 76, 80, 89, 91, 92, 94, 95, 96, 97, 98, 99, 103, 104, 105, 107, 112, 113, 115, 116, 117, 118], "itself": [3, 10, 48, 74, 85, 89, 94, 95, 103, 104], "anoth": [3, 10, 12, 32, 49, 73, 79, 84, 89, 90, 92, 95, 96, 97, 98, 106, 107, 117, 119], "align_instance_point": 3, "so": [3, 9, 10, 11, 12, 20, 27, 30, 32, 34, 38, 41, 43, 46, 47, 49, 53, 61, 62, 63, 69, 73, 75, 77, 78, 80, 84, 86, 88, 90, 92, 93, 94, 95, 96, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110, 114, 115, 117], "default": [3, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 33, 38, 40, 41, 42, 43, 47, 49, 50, 52, 60, 61, 69, 70, 72, 77, 78, 84, 85, 89, 91, 92, 93, 95, 96, 98, 102, 104, 106, 107, 108, 109, 117, 119], "yet": [3, 35, 80, 85, 98, 102, 115, 116, 118], "source_points_arrai": 3, "target_points_arrai": 3, "transform": [3, 4, 10, 47, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 66, 67, 72], "fit": [3, 41, 92, 103], "target": [3, 34, 37, 42, 46, 66, 68, 78, 80, 89, 95], "align_inst": 3, "all_points_arrai": 3, "ndarrai": [3, 4, 6, 9, 10, 12, 32, 34, 35, 52, 65, 68, 69, 72, 75, 78, 79, 80, 82, 83, 84, 85, 105], "node_a": 3, "int": [3, 4, 6, 7, 8, 10, 12, 14, 22, 23, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 52, 53, 54, 55, 58, 59, 61, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80, 82, 83, 84, 85, 105], "node_b": 3, "rotate_on_node_a": 3, "bool": [3, 4, 9, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 68, 69, 70, 72, 74, 77, 78, 79, 80, 81, 83, 84, 85], "fals": [3, 9, 10, 12, 15, 20, 22, 23, 26, 32, 34, 35, 37, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 51, 53, 54, 55, 58, 60, 61, 62, 64, 65, 66, 68, 69, 72, 74, 78, 80, 81, 83, 84, 85, 89, 92, 102, 103, 104, 105, 106, 107, 109], "align_instances_on_most_st": 3, "min_stable_dist": 3, "float": [3, 6, 7, 10, 12, 14, 32, 34, 36, 40, 42, 43, 44, 46, 47, 48, 49, 52, 53, 55, 58, 59, 61, 62, 63, 66, 67, 68, 69, 70, 72, 73, 75, 76, 78, 79, 80, 81, 82, 83, 84, 104], "0": [3, 4, 10, 12, 14, 26, 32, 34, 35, 36, 37, 42, 43, 44, 45, 46, 47, 48, 49, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 65, 66, 67, 68, 69, 70, 71, 72, 73, 75, 76, 78, 79, 80, 82, 83, 84, 85, 89, 92, 95, 102, 104, 105, 106, 107, 108, 109, 110], "pair": [3, 6, 47, 52, 69, 70, 72, 75, 78, 82, 83, 84, 86, 89, 95, 107, 110, 118, 119], "along": [3, 14, 23, 27, 47, 54, 55, 60, 72, 75, 76, 82, 99, 103, 109], "get_instances_point": 3, "list": [3, 4, 6, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 35, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 49, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 72, 74, 75, 77, 78, 79, 80, 81, 82, 83, 84, 85, 89, 92, 93, 95, 97, 102, 103, 104, 113, 116], "return": [3, 4, 6, 7, 9, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 103, 104, 105, 106, 107, 108, 109, 110, 115], "singl": [3, 4, 7, 10, 12, 14, 17, 23, 32, 34, 35, 42, 46, 47, 48, 49, 50, 51, 53, 55, 60, 61, 62, 64, 65, 66, 70, 72, 75, 77, 78, 80, 81, 83, 84, 86, 88, 89, 90, 91, 92, 94, 95, 101, 104, 105, 110, 117], "get_mean_and_std_for_point": 3, "aligned_points_arrai": 3, "tupl": [3, 4, 6, 9, 10, 12, 31, 32, 34, 36, 37, 41, 42, 44, 47, 53, 55, 64, 65, 66, 67, 68, 69, 70, 71, 72, 74, 75, 76, 78, 80, 82, 83, 84], "get_most_stable_node_pair": 3, "min_dist": 3, "distanc": [3, 6, 7, 47, 55, 69, 72, 75, 78, 82, 89, 95, 107], "over": [3, 4, 6, 8, 10, 12, 48, 49, 50, 54, 65, 67, 68, 72, 75, 76, 78, 80, 82, 83, 88, 89, 90, 92, 95, 103, 105, 106, 107], "min": [3, 7, 101, 106, 107, 109], "get_stable_node_pair": 3, "node_nam": [3, 9, 11, 12, 84, 103], "sort": [3, 6, 48, 69, 79], "dev": [3, 101, 102], "get_template_points_arrai": 3, "modul": [4, 6, 7, 10, 30, 33, 34, 35, 37, 38, 40, 41, 42, 43, 44, 50, 64, 65, 72, 74, 75, 76, 77, 79, 82, 84, 107], "frame": [4, 6, 7, 9, 10, 11, 12, 14, 16, 17, 23, 28, 32, 33, 34, 46, 47, 55, 57, 61, 65, 67, 69, 72, 75, 78, 79, 80, 83, 86, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 103, 104, 105, 106, 108, 109, 112, 113, 115, 117, 120], "featur": [4, 36, 37, 38, 41, 42, 44, 46, 47, 69, 70, 78, 88, 92, 100, 113, 116], "pca": 4, "kmean": [4, 103], "class": [4, 7, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 70, 71, 72, 73, 74, 75, 78, 80, 81, 82, 84, 85, 105], "framegroupset": 4, "method": [4, 7, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 32, 34, 35, 37, 38, 44, 45, 47, 51, 54, 61, 69, 72, 74, 75, 76, 77, 78, 80, 81, 82, 84, 86, 89, 92, 96, 100, 101, 104, 107, 108, 115, 116, 118, 119], "str": [4, 7, 8, 9, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 37, 38, 41, 42, 43, 45, 46, 47, 48, 49, 50, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 72, 75, 76, 77, 78, 81, 83, 84, 85, 107], "item_group": 4, "dict": [4, 7, 9, 10, 12, 14, 21, 23, 30, 35, 45, 50, 54, 56, 64, 67, 68, 69, 72, 75, 78, 79, 80, 82, 84, 85, 107], "frameitem": 4, "_noth": [4, 10, 12, 32, 35, 42, 46, 47, 48, 49, 50, 56, 64, 72, 78, 80], "noth": [4, 10, 12, 32, 35, 42, 46, 47, 48, 49, 50, 56, 64, 72, 78, 80], "group_data": 4, "groupset_data": 4, "group": [4, 10, 47, 48, 71, 72, 73, 75, 80, 82, 84, 85, 88, 97, 100, 101, 105, 107, 116, 117], "object": [4, 6, 7, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 34, 50, 67, 69, 78, 80, 83, 84, 85, 102, 104, 105, 107, 108, 109], "one": [4, 10, 12, 32, 42, 47, 49, 50, 51, 52, 58, 64, 66, 67, 68, 69, 72, 74, 80, 81, 84, 89, 91, 92, 95, 97, 98, 101, 102, 104, 105, 106, 107, 108, 109, 117, 118, 119, 120], "label": [4, 6, 7, 9, 10, 12, 14, 15, 16, 17, 18, 22, 23, 25, 26, 28, 30, 32, 34, 42, 45, 46, 47, 48, 61, 65, 67, 69, 72, 80, 81, 82, 84, 86, 88, 90, 93, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 117, 118, 119, 120], "dictionari": [4, 7, 9, 10, 12, 14, 20, 23, 25, 35, 42, 49, 50, 54, 56, 57, 64, 65, 67, 68, 69, 72, 78, 80, 82, 84, 85, 107], "map": [4, 6, 27, 37, 42, 44, 47, 49, 53, 55, 56, 58, 61, 63, 64, 65, 67, 69, 70, 71, 72, 75, 76, 80, 81, 82, 83, 84, 88, 89, 92, 96, 104, 106, 107, 109, 113], "extra": [4, 46, 47, 102], "data": [4, 5, 6, 7, 9, 10, 12, 14, 15, 16, 17, 20, 23, 24, 25, 26, 27, 28, 31, 32, 34, 35, 42, 45, 47, 48, 49, 50, 69, 72, 80, 81, 84, 85, 86, 88, 89, 91, 92, 93, 94, 96, 97, 98, 99, 107, 113, 119, 120], "kei": [4, 7, 9, 10, 12, 14, 26, 42, 45, 49, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 64, 65, 66, 67, 68, 69, 72, 78, 79, 81, 82, 84, 85, 95, 103, 104, 105, 107, 116, 119], "id": [4, 20, 65, 75, 84, 89, 105, 106, 109, 112], "valu": [4, 7, 9, 10, 12, 13, 32, 33, 38, 42, 46, 47, 48, 49, 53, 54, 58, 61, 62, 63, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 78, 79, 80, 82, 84, 85, 89, 92, 102, 104, 105, 106, 107, 108, 109], "entir": [4, 9, 12, 26, 32, 37, 48, 53, 65, 76, 89, 90, 92, 94, 96, 105, 106, 107, 115, 116, 120], "properti": [4, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 32, 36, 37, 38, 39, 40, 41, 42, 43, 49, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 70, 72, 74, 78, 80, 81, 84], "all_item": 4, "append_to_group": 4, "extend_group_item": 4, "item_list": 4, "get_item_group": 4, "contain": [4, 10, 12, 14, 17, 23, 27, 28, 32, 37, 41, 42, 46, 47, 49, 50, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 65, 66, 67, 69, 71, 72, 75, 76, 79, 81, 82, 84, 85, 86, 89, 90, 91, 94, 95, 96, 104, 105, 106, 107, 110, 112], "iter": [4, 12, 32, 54, 57, 65, 67, 84, 85, 113, 117], "yield": [4, 12, 54, 57, 61, 72], "sampl": [4, 12, 42, 45, 46, 48, 53, 54, 55, 58, 63, 67, 68, 71, 72, 75, 76, 86, 88, 92, 98, 99, 103, 111, 112, 113, 116, 120], "per_group": 4, "unique_sampl": 4, "true": [4, 9, 10, 12, 23, 24, 26, 30, 32, 34, 35, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 68, 69, 70, 72, 74, 76, 77, 78, 79, 80, 83, 84, 85, 89, 102, 103, 104, 105, 106, 107, 108, 109], "current": [4, 12, 14, 23, 26, 32, 37, 42, 44, 45, 47, 49, 53, 55, 74, 77, 79, 80, 81, 84, 85, 87, 88, 90, 92, 95, 98, 101, 102, 105, 106, 110, 116, 118], "order": [4, 10, 12, 32, 37, 42, 46, 47, 48, 52, 54, 60, 62, 65, 66, 67, 72, 73, 75, 76, 79, 80, 84, 85, 96, 97, 103, 104, 116, 117], "match": [4, 10, 11, 12, 23, 32, 34, 37, 44, 46, 47, 57, 60, 65, 69, 71, 72, 75, 78, 79, 80, 82, 84, 85, 89, 92, 95, 96, 108, 109, 114, 119], "drawn": [4, 88, 92, 95], "paramet": [4, 6, 7, 9, 10, 11, 12, 14, 16, 23, 25, 26, 28, 30, 31, 32, 34, 37, 38, 39, 40, 41, 42, 43, 44, 48, 50, 51, 52, 53, 54, 55, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 88, 89, 91, 95, 96, 98, 99, 100, 106, 108, 109, 110, 112, 113], "number": [4, 6, 7, 10, 12, 26, 32, 33, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 51, 53, 54, 58, 61, 62, 64, 66, 69, 70, 71, 72, 73, 75, 78, 79, 80, 85, 86, 88, 89, 92, 95, 98, 100, 101, 103, 106, 109, 110, 113, 119], "whether": [4, 10, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 31, 32, 34, 41, 42, 44, 47, 49, 53, 58, 69, 72, 78, 80, 81, 83, 84, 85, 89, 92, 95, 96, 100, 118], "ensur": [4, 12, 23, 43, 46, 49, 51, 61, 66, 67, 72, 75, 77, 79, 86, 96, 100, 102, 119], "share": [4, 77, 96, 100, 102, 104, 107, 108], "video": [4, 7, 9, 10, 11, 12, 14, 16, 17, 23, 24, 26, 27, 28, 30, 33, 34, 46, 65, 67, 69, 72, 86, 87, 88, 89, 90, 92, 94, 95, 96, 97, 98, 100, 104, 105, 108, 109, 111, 112, 113, 115, 116, 120], "frame_idx": [4, 10, 12, 69, 104, 105], "just": [4, 10, 12, 23, 32, 49, 61, 72, 79, 84, 85, 88, 92, 94, 95, 96, 97, 98, 103, 104, 105, 106, 108, 110, 111, 115, 116, 117, 119], "simpl": [4, 21, 25, 29, 35, 37, 38, 39, 40, 49, 50, 72, 80, 85, 89, 95, 96, 103, 110], "wrapper": [4, 12, 32, 33, 54, 66, 69, 74, 75, 78, 84, 85], "plu": [4, 10, 95], "imag": [4, 10, 12, 14, 17, 23, 32, 34, 38, 39, 41, 42, 45, 46, 47, 48, 49, 52, 53, 55, 58, 61, 62, 65, 66, 68, 72, 75, 76, 80, 81, 83, 84, 86, 88, 89, 90, 91, 92, 95, 96, 97, 98, 104, 105, 106, 107, 108, 110, 113, 116, 117], "itemstack": 4, "none": [4, 6, 7, 9, 10, 11, 12, 14, 16, 17, 20, 22, 23, 25, 26, 27, 28, 30, 31, 32, 34, 35, 37, 38, 39, 42, 44, 45, 46, 47, 48, 49, 51, 52, 55, 59, 60, 61, 62, 64, 65, 66, 68, 69, 70, 72, 73, 74, 75, 76, 77, 78, 80, 81, 82, 83, 84, 85, 89, 98, 102, 103, 104, 106, 108, 109], "ownership": 4, "meta": [4, 72, 84], "group_set": 4, "own": [4, 25, 54, 84, 90, 91, 92, 101, 102, 105, 106, 113, 116, 120], "more": [4, 8, 10, 12, 27, 37, 38, 40, 41, 42, 43, 44, 47, 48, 51, 58, 64, 67, 68, 72, 75, 77, 79, 80, 81, 84, 87, 88, 89, 90, 91, 92, 97, 99, 101, 102, 107, 108, 113, 114, 115, 117, 118, 119, 120], "row": [4, 9, 10, 52, 58, 61, 63, 66, 75, 78, 79, 82], "numpi": [4, 9, 10, 12, 26, 27, 32, 35, 52, 65, 68, 69, 72, 76, 78, 80, 82, 85, 102, 103, 104, 105, 106, 107, 108, 109], "specifi": [4, 7, 9, 10, 11, 12, 16, 25, 26, 28, 32, 34, 37, 38, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 52, 60, 61, 62, 63, 65, 66, 67, 68, 70, 72, 74, 75, 76, 79, 80, 81, 82, 83, 84, 85, 87, 89, 92, 95, 96, 100, 106, 107, 109], "store": [4, 9, 10, 11, 12, 20, 23, 26, 32, 36, 46, 48, 49, 52, 54, 60, 61, 65, 66, 67, 72, 78, 84, 90, 96, 100, 103, 104, 106, 109, 114, 118], "metadata": [4, 10, 12, 14, 17, 20, 27, 28, 36, 49, 72, 104, 106, 107, 112], "oper": [4, 12, 37, 38, 40, 43, 46, 53, 54, 55, 61, 62, 64, 65, 66, 67, 68, 69, 72, 73, 75, 76, 100, 104, 105, 106, 107, 109], "stack": [4, 37, 38, 40, 41, 42, 43, 44, 47, 48, 54, 76, 106, 109], "groupset": 4, "brisk_bag_of_featur": 4, "brisk_threshold": 4, "40": [4, 86, 106, 109], "vocab_s": 4, "20": [4, 75, 80, 86, 103, 104, 105, 106, 107, 109, 116, 117], "bag": 4, "base": [4, 7, 10, 12, 13, 20, 32, 33, 35, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 49, 52, 62, 64, 69, 70, 72, 75, 76, 78, 80, 81, 83, 89, 92, 93, 94, 95, 98, 99, 100, 101, 102, 113], "brisk": 4, "current_groupset": 4, "recent": [4, 119], "extend_ownership": 4, "row_count": 4, "extend": [4, 10, 86], "next": [4, 12, 89, 92, 95, 102, 105, 106, 107, 108, 113, 116, 117, 119], "flatten": [4, 53, 55, 68, 70, 103, 107], "d": [4, 61, 72, 82, 84, 85, 86, 90, 92, 93, 94, 95, 100, 101, 102, 104, 105, 106, 107, 108, 109], "arrai": [4, 6, 9, 10, 12, 32, 52, 65, 68, 69, 72, 75, 76, 80, 82, 84, 103, 104, 105], "get_all_items_from_group": 4, "get_item_data": 4, "belong": [4, 10, 12, 61, 75, 76, 108], "get_item_data_idx": 4, "index": [4, 7, 9, 10, 11, 12, 26, 30, 32, 34, 45, 46, 55, 58, 60, 61, 65, 71, 72, 73, 75, 76, 77, 80, 82, 84, 85, 89, 96, 100, 104, 105, 106, 109], "get_raw_imag": 4, "scale": [4, 10, 34, 36, 37, 38, 39, 40, 42, 43, 44, 46, 47, 48, 52, 53, 55, 58, 61, 62, 63, 65, 66, 68, 69, 71, 72, 73, 75, 76, 80, 83, 88, 89, 98, 104, 105, 108, 109], "raw": [4, 20, 46, 48, 50, 52, 53, 55, 65, 72, 103, 105, 108], "hog_bag_of_featur": 4, "hog": 4, "descriptor": 4, "n_cluster": [4, 103], "k": [4, 73, 103], "cluster": [4, 72, 89], "make_sample_group": 4, "samples_per_video": 4, "sample_method": 4, "stride": [4, 10, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 53, 55, 58, 61, 63, 66, 70, 72, 74, 75, 83, 88, 92, 98, 106, 109], "n_compon": 4, "sample_group": 4, "samples_per_group": 4, "parallelfeaturepipelin": 4, "pipelin": [4, 12, 32, 48, 54, 56, 61, 65, 69, 72, 75, 80, 81, 89, 91, 95, 98, 103, 106, 109, 110, 117], "featuresuggestionpipelin": 4, "videos_as_dict": 4, "enabl": [4, 32, 41, 47, 48, 49, 53, 54, 72, 77, 89, 95, 100, 102, 104, 105, 106, 107, 109, 112, 113], "per": [4, 7, 32, 34, 40, 43, 47, 48, 72, 78, 79, 80, 83, 88, 89, 95], "parallel": [4, 47, 48, 49, 54], "suggest": [4, 12, 22, 23, 65, 80, 89, 93, 95, 98, 99, 100, 115, 116], "desir": [4, 12, 38, 40, 42, 43, 44, 47, 49, 52, 60, 61, 62, 66, 80, 91, 95, 108, 113, 117], "call": [4, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 42, 45, 54, 69, 72, 73, 76, 78, 80, 85, 91, 96, 105, 106, 110, 113], "run": [4, 8, 12, 34, 45, 49, 50, 64, 65, 72, 75, 77, 80, 81, 82, 87, 88, 89, 91, 92, 94, 95, 98, 100, 102, 103, 105, 106, 108, 109, 112, 117, 119, 120], "process": [4, 12, 26, 32, 35, 47, 48, 54, 56, 61, 64, 65, 66, 68, 80, 85, 88, 90, 91, 95, 96, 98, 100, 102, 106, 112, 113, 115, 117], "care": [4, 43, 104, 105, 106, 107, 108], "serial": [4, 12, 32, 49, 50, 72, 84, 104], "pool": [4, 37, 38, 40, 43, 47, 66, 69], "deseri": [4, 12, 13, 14, 15, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 32, 50, 84], "back": [4, 10, 32, 35, 54, 61, 75, 79, 80, 89, 92, 100, 110], "suggestionfram": [4, 12], "video_idx": [4, 12, 26, 34], "idx": [4, 6, 12, 32, 105], "classmethod": [4, 10, 12, 15, 16, 17, 18, 22, 23, 24, 27, 28, 32, 38, 40, 41, 42, 43, 44, 50, 52, 60, 61, 62, 64, 65, 66, 70, 72, 73, 74, 75, 78, 80, 81, 84], "tuples_to_suggest": 4, "command": [5, 11, 23, 25, 49, 87, 91, 92, 93, 94, 95, 96, 100, 102, 109, 112, 116, 117, 118, 119], "util": [5, 11, 31, 36, 53, 55, 58, 62, 63, 67, 69, 71, 72, 75, 77, 83, 95, 104], "print": [5, 12, 64, 68, 72, 77, 102, 103, 104, 105, 106, 107, 114], "produc": [6, 32, 47, 53, 54, 55, 62, 64, 65, 70, 72, 80, 88], "predict": [6, 8, 10, 12, 30, 46, 47, 49, 61, 64, 69, 70, 72, 73, 75, 78, 80, 87, 88, 89, 91, 93, 95, 96, 98, 101, 103, 104, 105, 106, 107, 108, 109, 111, 112, 113, 116, 117, 119, 120], "dataset": [6, 9, 11, 14, 15, 17, 18, 19, 22, 23, 25, 27, 28, 30, 32, 34, 45, 46, 48, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 69, 72, 81, 84, 91, 92, 96, 101, 103, 104, 105, 106, 107, 111, 113, 114, 117, 118, 120], "calculate_pairwise_cost": 6, "instances_a": 6, "union": [6, 10, 12, 22, 23, 25, 28, 32, 37, 60, 64, 65, 69, 70, 72, 75, 78, 80, 81, 82, 83, 84], "predictedinst": [6, 10, 12, 26, 69, 72, 104, 105], "instances_b": 6, "cost_funct": 6, "callabl": [6, 12, 22, 23, 31, 34, 42, 45, 54, 56, 78, 79, 80, 81], "pairwis": [6, 55, 69, 80], "cost": [6, 38, 40, 43, 46, 47, 48, 54, 61, 72, 78, 79, 80, 82, 88, 95, 104, 107], "compare_instance_list": 6, "list_points_arrai": 6, "match_instance_list": 6, "overal": [6, 10, 47], "total": [6, 7, 12, 34, 48, 61, 67, 72, 75, 95, 104, 105, 106, 109], "match_instance_lists_nodewis": 6, "thresh": [6, 79, 107], "first": [6, 7, 9, 11, 12, 14, 16, 28, 32, 34, 37, 38, 40, 42, 43, 44, 48, 52, 53, 54, 55, 61, 62, 64, 65, 68, 72, 77, 78, 79, 80, 84, 85, 89, 91, 92, 94, 95, 97, 102, 103, 104, 105, 106, 107, 108, 109, 110, 113, 114, 115, 117, 118, 119, 120], "closest": [6, 7, 72, 82], "second": [6, 14, 32, 34, 49, 72, 79, 84, 89, 95, 98, 115], "matched_instance_dist": 6, "labels_gt": [6, 69, 107, 109], "io": [6, 65, 89, 114], "labels_pr": [6, 69, 106, 107, 109], "match_lists_funct": 6, "frame_rang": [6, 12], "rang": [6, 10, 12, 26, 32, 33, 41, 42, 48, 61, 62, 68, 69, 72, 75, 85, 88, 89, 92, 96, 98, 103, 107], "ground": [6, 59, 69, 72, 73, 83, 89, 92, 107, 109, 113], "truth": [6, 59, 69, 72, 73, 83, 89, 92, 107, 109, 113], "compar": [6, 12, 32, 84, 85, 89, 107], "indic": [6, 12, 32, 45, 46, 52, 53, 55, 58, 60, 61, 65, 67, 69, 71, 72, 75, 76, 78, 82, 84, 89, 92, 95, 115], "nodeless_point_dist": 6, "inst_a": 6, "inst_b": 6, "ignor": [6, 9, 10, 37, 46, 49, 60, 69, 75, 76, 84, 100, 104, 107, 108], "point_dist": 6, "point_match_count": 6, "dist_arrai": 6, "point_nonmatch_count": 6, "seri": [7, 103], "give": [7, 72, 84, 88, 89, 90, 92, 95, 96, 98, 110, 115, 116], "statisticseri": 7, "variou": [7, 11, 12, 32, 87, 89, 92], "numer": [7, 12, 14, 80, 85, 104], "get_instance_score_seri": 7, "reduct": [7, 48, 53, 55, 66], "sum": [7, 55, 95], "score": [7, 9, 10, 12, 27, 47, 69, 72, 75, 78, 80, 89, 92, 104, 105, 107], "doc": [7, 27, 87, 101, 102, 103], "get_point_count_seri": 7, "get_point_displacement_seri": 7, "displac": [7, 75, 80], "same": [7, 10, 12, 14, 27, 37, 38, 39, 40, 41, 42, 43, 44, 47, 48, 49, 50, 52, 53, 54, 55, 58, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 75, 76, 78, 79, 83, 84, 85, 88, 90, 91, 92, 93, 94, 95, 96, 100, 102, 104, 108, 109, 110], "track": [7, 8, 9, 10, 11, 12, 26, 27, 28, 35, 46, 47, 58, 61, 65, 66, 67, 72, 74, 78, 79, 83, 86, 92, 93, 96, 97, 100, 101, 103, 104, 105, 109, 113, 114, 115, 116, 120], "earlier": [7, 80], "max": [7, 37, 38, 43, 46, 53, 55, 66, 72, 75, 78, 80, 88, 89, 98, 106, 109], "get_point_score_seri": 7, "get_primary_point_displacement_seri": 7, "primary_nod": 7, "get_tracking_score_seri": 7, "cli": [8, 11, 72, 81, 89, 100, 105, 106, 108, 112], "mostli": [8, 12, 49, 54, 77, 115], "deprec": [8, 23], "cleaner": [8, 80], "cull": [8, 78, 79, 80, 89, 95], "ii": 8, "connect": [8, 35, 36, 37, 38, 40, 41, 42, 43, 44, 47, 70, 72, 74, 75, 78, 84, 89, 90, 92, 95, 96, 97, 103, 104, 109, 113, 118], "break": [8, 78, 80, 89, 100, 102], "let": [8, 90, 91, 92, 94, 96, 98, 103, 104, 105, 107, 108, 109, 110, 116, 117, 119], "control": [8, 33, 44, 47, 49, 75, 83, 88, 91, 92, 98, 109, 110], "how": [8, 34, 38, 46, 47, 49, 66, 69, 78, 80, 85, 86, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 103, 104, 105, 106, 107, 108, 110, 112, 113, 115, 116, 117], "select": [8, 12, 32, 44, 46, 67, 78, 80, 86, 87, 89, 91, 94, 95, 96, 98, 99, 102, 109, 113, 117, 118, 119, 120], "better": [8, 23, 72, 75, 79, 88, 90, 94, 107, 115, 117], "fit_track": 8, "filenam": [8, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 32, 33, 34, 45, 50, 65, 69, 84, 85, 87, 89, 90, 91, 96, 103, 104, 105, 109, 110], "instance_count": [8, 12, 78, 80, 103], "wrap": [8, 14, 17, 64, 72, 75, 77, 105], "easier": [8, 47, 50, 69, 73, 88, 95, 113], "api": [8, 25, 27, 72, 75, 81, 84, 100, 101, 104, 105, 107, 112], "hdf5": [9, 11, 12, 18, 20, 23, 25, 27, 28, 32, 84, 85, 89, 96, 103, 104, 111, 112, 114], "csv": [9, 11, 12, 17, 49, 89, 92, 96, 107, 109], "occup": [9, 12, 113], "empti": [9, 12, 32, 39, 65, 78, 89, 116, 118], "By": [9, 32, 92, 98, 105, 119], "also": [9, 12, 23, 25, 32, 41, 46, 47, 50, 53, 58, 60, 61, 62, 63, 65, 66, 68, 69, 71, 72, 75, 76, 77, 78, 80, 82, 87, 89, 90, 91, 92, 94, 95, 96, 98, 100, 102, 103, 104, 105, 106, 107, 110, 113, 115, 116, 117, 119], "begin": [9, 45, 48, 49, 90], "end": [9, 12, 37, 38, 43, 45, 48, 49, 50, 54, 62, 64, 85, 92, 95, 100, 103, 113], "although": [9, 10, 25, 47, 94, 95, 96, 102, 115, 117], "argument": [9, 11, 12, 23, 32, 34, 42, 44, 45, 60, 61, 65, 66, 72, 76, 79, 89, 96, 109], "track_occup": [9, 11, 12, 28, 103, 114], "track_nam": [9, 11, 12, 103], "edge_nam": [9, 11, 84], "edge_ind": [9, 11, 55, 72, 75, 84], "point_scor": [9, 11], "instance_scor": [9, 10, 11, 72, 105], "tracking_scor": [9, 10, 11, 104, 105], "labels_path": [9, 89, 109], "path": [9, 11, 12, 16, 26, 30, 31, 32, 33, 34, 38, 42, 45, 46, 47, 49, 50, 65, 69, 72, 81, 85, 87, 89, 90, 91, 92, 95, 96, 100, 102, 103, 106, 107, 108, 109, 110, 114], "slp": [9, 11, 12, 22, 32, 69, 86, 89, 91, 96, 103, 104, 105, 106, 107, 108, 109, 110, 111], "gui": [9, 10, 12, 24, 25, 34, 49, 90, 91, 93, 94, 95, 96, 100, 101, 106, 110, 111, 112, 113, 115, 118], "context": [9, 12, 35, 98], "video_path": [9, 18, 69, 89, 109], "video_ind": [9, 57, 65], "scalar": [9, 10, 12, 46, 47, 54, 61, 65, 66, 68, 69, 72, 75, 76, 78, 82, 83], "integ": [9, 12, 37, 41, 42, 44, 45, 46, 47, 48, 49, 58, 61, 62, 66, 68, 71, 72, 75, 76, 89], "proven": [9, 12, 67, 109], "denot": [9, 10, 12, 41, 61, 69, 71, 72, 75, 82, 107], "column": [9, 10, 11, 66, 72, 76, 78, 79, 89, 114], "major": [9, 11, 99, 104, 114], "matlab": [9, 11, 24, 30, 84, 89, 103, 113, 120], "get_edges_as_np_str": 9, "bytes_": 9, "np": [9, 12, 32, 72, 75, 76, 82, 85, 103, 105, 107], "string_": 9, "get_nodes_as_np_str": 9, "get_occupancy_and_points_matric": 9, "all_fram": [9, 12], "build": [9, 12, 32, 33, 36, 37, 39, 41, 51, 60, 61, 62, 64, 66, 72, 74, 101, 102, 103, 105, 106, 109, 113], "assum": [9, 12, 32, 37, 60, 62, 72, 83, 84, 85, 92, 102, 107, 110, 117], "zero": [9, 10, 32, 53, 58, 63, 66, 70, 89, 95, 103, 104, 105, 106, 107, 108, 109], "up": [9, 32, 35, 37, 43, 44, 47, 48, 49, 55, 60, 65, 67, 72, 78, 79, 81, 85, 86, 88, 89, 91, 92, 93, 95, 97, 101, 102, 103, 104, 106, 107, 108, 109, 110, 115, 116, 117], "output": [9, 10, 11, 12, 16, 26, 28, 32, 34, 36, 37, 38, 39, 40, 41, 42, 43, 44, 47, 50, 52, 53, 55, 56, 58, 60, 61, 64, 65, 66, 67, 68, 70, 72, 73, 74, 75, 81, 84, 88, 89, 92, 96, 102, 104, 105, 106, 107, 109, 110], "last": [9, 12, 32, 35, 37, 41, 42, 53, 55, 61, 62, 63, 68, 72, 75, 76, 77, 80, 89, 92, 96, 100, 118], "source_object": [9, 13, 14, 16, 21, 22, 23, 25, 27, 28, 29], "get_tracks_as_np_str": 9, "output_path": [9, 12, 109], "coordin": [9, 10, 12, 14, 20, 47, 48, 52, 53, 55, 59, 61, 63, 65, 69, 71, 72, 75, 76, 82, 103, 104], "written": [9, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 34, 49, 110], "save": [9, 11, 12, 23, 25, 30, 32, 33, 34, 45, 46, 47, 48, 49, 50, 65, 69, 72, 80, 81, 84, 85, 86, 87, 89, 91, 92, 94, 96, 100, 103, 104, 106, 107, 109, 110, 117, 120], "analysi": [9, 11, 12, 16, 22, 25, 27, 28, 89, 93, 100, 107, 111, 113, 119, 120], "remove_empty_tracks_from_matric": 9, "occupancy_matrix": [9, 114], "locations_matrix": 9, "unoccupi": 9, "2d": [9, 53, 55, 58, 75, 76], "4d": [9, 32, 39], "3d": 9, "write_csv_fil": 9, "data_dict": 9, "write_occupancy_fil": 9, "transpos": [9, 11, 37, 39, 40, 43, 44, 47, 92, 95, 103, 114, 119], "format": [9, 10, 11, 12, 30, 32, 42, 50, 61, 62, 64, 72, 84, 85, 87, 89, 91, 92, 95, 96, 104, 105], "structur": [10, 12, 37, 38, 48, 50, 67, 72, 84, 86, 87, 105, 106], "relationship": [10, 47, 84], "A": [10, 11, 12, 14, 23, 26, 30, 32, 35, 37, 38, 39, 41, 42, 44, 45, 46, 47, 49, 50, 52, 53, 54, 55, 56, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 80, 81, 82, 83, 84, 85, 89, 92, 96, 100, 101, 104, 106, 107, 110], "labeledfram": [10, 12, 30, 32, 65, 69, 78, 80, 104], "pointarrai": 10, "predictedpointarrai": 10, "associ": [10, 12, 26, 38, 41, 47, 53, 60, 65, 69, 70, 72, 75, 80, 104, 108, 113], "predictedpoint": [10, 78, 104], "objectss": 10, "ideal": [10, 54, 97], "mani": [10, 12, 37, 47, 72, 80, 85, 89, 92, 95, 98, 100, 102, 113, 115, 117], "sync": [10, 90], "manipul": [10, 104], "from_predict": 10, "altern": [10, 77, 89, 96, 102, 118], "whose": [10, 39, 65, 84], "multi": [10, 37, 41, 42, 44, 47, 53, 54, 61, 66, 69, 70, 72, 75, 76, 77, 81, 88, 89, 96, 100, 101, 117], "allow": [10, 12, 20, 23, 25, 35, 47, 48, 49, 64, 72, 74, 84, 89, 92, 96, 100, 102, 104, 106, 110, 113, 119], "anim": [10, 14, 40, 47, 58, 84, 86, 88, 89, 91, 92, 93, 95, 99, 100, 103, 107, 109, 110, 112, 113, 115, 116, 117], "across": [10, 12, 32, 38, 46, 48, 49, 54, 55, 65, 67, 69, 72, 73, 75, 82, 84, 92, 93, 95, 96, 108, 113, 115, 120], "wa": [10, 12, 14, 17, 31, 32, 37, 38, 45, 49, 50, 53, 61, 62, 65, 72, 77, 79, 82, 84, 87, 89, 92, 94, 95, 100, 101, 104, 105, 106, 107, 109, 110], "copi": [10, 12, 14, 35, 41, 47, 48, 65, 72, 81, 82, 84, 85, 90, 92, 94, 96, 103, 110, 112, 116], "refer": [10, 12, 20, 32, 46, 69, 72, 75, 76, 78, 80, 82, 87, 92, 95, 100, 104, 105, 107], "field": [10, 38, 40, 42, 43, 47, 49, 51, 55, 64, 66, 70, 72, 74, 75, 86, 87, 88, 92, 96, 97, 98, 109, 110, 117], "bounding_box": 10, "bound": [10, 14, 46, 47, 48, 52, 60, 61, 68, 69, 70, 72, 76, 78, 82, 95], "box": [10, 14, 32, 46, 47, 52, 60, 61, 69, 70, 72, 76, 78, 82, 88, 95], "y1": [10, 52, 61, 76, 82], "x1": [10, 52, 61, 76, 82], "y2": [10, 52, 61, 76, 82], "x2": [10, 52, 61, 76, 82], "centroid": [10, 27, 46, 47, 53, 58, 59, 60, 61, 62, 64, 66, 70, 72, 76, 78, 81, 87, 88, 89, 91, 93, 95, 96, 98, 104, 105, 109, 110, 117], "comput": [10, 12, 32, 37, 44, 46, 47, 48, 49, 53, 55, 60, 61, 62, 66, 68, 69, 72, 73, 75, 76, 78, 80, 81, 82, 90, 91, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112], "median": [10, 69], "visibl": [10, 12, 47, 65, 69, 73, 78, 86, 92, 94, 95, 97, 104, 107, 116], "fill_miss": [10, 103], "max_x": 10, "max_i": 10, "miss": [10, 12, 31, 32, 46, 53, 69, 71, 75, 80, 92, 95, 104, 107, 108], "modifi": [10, 12, 38, 39, 41, 52, 65, 77, 78, 79, 82, 92], "outsid": [10, 84], "valid": [10, 12, 32, 38, 41, 42, 45, 46, 47, 48, 49, 51, 62, 67, 72, 79, 81, 85, 86, 89, 92, 95, 98, 106, 107, 109, 110, 117], "from_numpi": [10, 32], "n_node": [10, 12, 52, 53, 55, 60, 61, 65, 69, 72, 75], "dtype": [10, 32, 42, 52, 53, 55, 58, 60, 61, 62, 63, 65, 66, 68, 71, 75, 76, 85, 104, 105, 106, 107, 109], "float32": [10, 42, 52, 53, 55, 58, 60, 61, 62, 63, 65, 68, 71, 72, 75, 76, 105, 106, 109], "nan": [10, 12, 53, 54, 60, 68, 69, 71, 72, 75, 76, 79, 80, 82, 89, 103, 104, 105], "alia": [10, 12, 45, 81], "from_pointsarrai": 10, "get_points_arrai": 10, "invisible_as_nan": 10, "full": [10, 12, 14, 17, 25, 28, 34, 42, 46, 47, 54, 61, 64, 65, 66, 68, 72, 75, 77, 84, 85, 86, 90, 96, 98, 101, 103], "recarrai": 10, "form": [10, 12, 37, 38, 42, 43, 44, 47, 49, 52, 54, 55, 56, 64, 67, 69, 70, 72, 75, 84, 97, 105, 108], "view": [10, 49, 84, 88, 89, 95, 98, 109], "underli": [10, 12, 20, 32, 50, 54, 72, 75, 81], "invis": 10, "mark": [10, 48, 92, 97, 115, 116, 117], "sinc": [10, 20, 23, 25, 32, 35, 46, 47, 48, 53, 54, 69, 72, 75, 78, 79, 80, 85, 88, 89, 91, 92, 95, 97, 98, 107, 108, 109, 110, 112, 113, 115, 116, 117], "don": [10, 11, 28, 32, 73, 85, 89, 90, 91, 99, 100, 102, 104, 105, 106, 109, 110, 112, 117], "present": [10, 12, 32, 37, 47, 50, 53, 55, 60, 61, 64, 72, 76, 80, 84, 89, 92], "access": [10, 12, 26, 32, 48, 54, 75, 89, 90, 92, 96, 100, 104, 110, 112], "midpoint": [10, 47, 60], "center": [10, 12, 46, 47, 48, 53, 59, 60, 61, 63, 64, 66, 68, 70, 72, 76, 81, 86, 87, 88, 91, 96, 98, 103, 104, 109, 117], "n_visible_point": 10, "nodes_point": 10, "points_arrai": 10, "replac": [10, 12, 23, 31, 32, 38, 53, 58, 63, 67, 68, 69, 70, 76, 79, 82, 84, 91, 92, 100, 102, 106, 108, 115], "were": [10, 12, 32, 46, 54, 58, 62, 67, 72, 77, 78, 80, 86, 87, 89, 92, 95, 102, 103, 104, 108, 110, 111, 113, 115], "transform_point": 10, "transformation_matrix": 10, "affin": [10, 47, 48, 55, 64, 70, 72, 75, 88, 92, 96, 97, 117], "instanceslist": 10, "arg": [10, 11, 12, 15, 17, 18, 21, 22, 23, 24, 25, 28, 29, 32, 34, 47, 72, 73, 80, 81, 107, 109], "labeled_fram": [10, 12, 104, 105, 108], "attribut": [10, 12, 32, 37, 41, 42, 44, 46, 47, 48, 49, 50, 51, 53, 54, 55, 58, 60, 61, 62, 64, 65, 66, 67, 70, 72, 81, 84, 105], "append": [10, 12, 22, 23, 26, 37, 45, 49, 62, 64, 69, 89, 105, 108], "clear": [10, 12, 79, 80, 86, 89, 92, 101], "shallow": [10, 12, 44, 97], "normal": [10, 37, 38, 46, 47, 48, 53, 61, 68, 69, 72, 75, 76, 78, 80, 95, 100, 106], "insert": [10, 12], "pop": [10, 92, 102, 116], "hold": [10, 20, 32, 74, 75, 92, 95, 116], "complex_frame_merg": 10, "base_fram": 10, "new_fram": [10, 12], "conflict": [10, 12, 37, 38, 49, 94, 102], "occur": [10, 12, 49, 54, 95, 113, 115, 116, 119], "perfectli": [10, 12, 14, 17, 53, 103], "those": [10, 12, 32, 46, 47, 72, 84, 89, 92, 94, 96, 102, 104, 107, 115, 117], "three": [10, 12, 14, 75, 95, 96, 102, 117], "complex_merge_between": [10, 12], "base_label": [10, 12], "everyth": [10, 80, 92, 117], "cleanli": [10, 12], "retriev": [10, 75, 107], "non": [10, 12, 25, 32, 48, 49, 66, 72, 76, 78, 80, 89, 92, 93, 97, 102, 103, 105], "has_predicted_inst": 10, "has_tracked_inst": 10, "has_user_inst": 10, "height": [10, 14, 32, 33, 34, 37, 38, 42, 43, 44, 46, 47, 48, 52, 53, 55, 61, 62, 65, 66, 68, 71, 72, 74, 75, 76, 80, 83, 89, 96], "width": [10, 14, 32, 33, 34, 37, 38, 42, 43, 44, 46, 47, 48, 52, 53, 55, 61, 62, 65, 66, 68, 71, 72, 74, 75, 76, 80, 83, 89, 96], "channel": [10, 32, 34, 37, 38, 39, 41, 42, 44, 46, 47, 53, 58, 61, 62, 65, 66, 70, 71, 72, 73, 74, 75, 76, 83, 96, 102, 104, 118], "instances_to_show": 10, "regular": [10, 80, 89, 92, 115], "static": [10, 12, 32, 33, 80, 84], "merge_fram": 10, "remove_redund": 10, "know": [10, 92, 102, 105, 119], "drop": [10, 35, 54, 64, 95, 117, 118], "perfect": [10, 48, 69], "n_predicted_inst": 10, "n_tracked_inst": 10, "n_user_inst": 10, "plot": [10, 83, 92, 103, 104, 105, 107, 108], "load": [10, 12, 20, 23, 30, 32, 41, 42, 47, 48, 50, 54, 65, 69, 72, 84, 102, 104, 105, 106, 107, 108, 109, 112], "figur": [10, 45, 81, 83, 103, 105, 107], "nn": [10, 89, 106, 107, 108, 109], "viz": [10, 45, 49, 96, 106, 109], "plot_img": [10, 83], "plot_inst": [10, 83], "plot_predict": 10, "predicted_inst": [10, 12, 59, 66, 75], "remove_empty_inst": [10, 12], "remove_untrack": 10, "tracked_inst": [10, 80], "training_inst": 10, "train": [10, 12, 23, 32, 42, 45, 46, 47, 48, 49, 50, 54, 55, 64, 65, 69, 72, 73, 74, 82, 86, 88, 92, 93, 94, 97, 99, 101, 103, 104, 105, 107, 111, 113, 115, 116, 119, 120], "unused_predict": 10, "unus": 10, "user_inst": [10, 12, 65], "complet": [10, 12, 14, 49, 75, 84, 87, 104, 113], "horizont": [10, 52], "pixel": [10, 32, 34, 46, 47, 48, 52, 53, 55, 58, 61, 66, 69, 72, 75, 76, 78, 89, 98], "vertic": [10, 52], "verifi": [10, 102, 113], "f8": 10, "isnan": [10, 103], "buf": 10, "offset": [10, 47, 53, 58, 63, 64, 66, 70, 72, 76, 81], "titl": [10, 100, 101, 103, 108], "byteord": 10, "c": [10, 39, 41, 61, 95, 100, 101, 102, 103, 104, 105, 106, 108], "sub": [10, 35, 49], "record": [10, 32, 86, 105, 106, 109, 110, 113], "from_arrai": 10, "child": 10, "make_default": 10, "alloc": [10, 12, 77, 104, 109], "element": [10, 12, 42, 44, 47, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 65, 66, 67, 72, 75, 76, 85], "infer": [10, 12, 32, 46, 47, 49, 50, 54, 57, 61, 64, 65, 66, 71, 74, 75, 76, 82, 88, 91, 92, 93, 94, 95, 97, 98, 101, 103, 106, 107, 113, 115, 116, 119, 120], "procedur": [10, 48, 69, 104, 107], "point_confid": 10, "confid": [10, 47, 53, 55, 58, 61, 63, 64, 70, 71, 72, 75, 76, 81, 83, 88, 89, 92, 96, 104], "paf": [10, 47, 48, 55, 64, 72, 73, 75, 81, 83, 89], "from_inst": 10, "manner": 10, "except": [10, 12, 25, 27, 37, 54, 72, 73, 97, 101], "points_and_scores_arrai": 10, "It": [10, 12, 32, 35, 42, 45, 46, 69, 72, 75, 79, 81, 89, 90, 91, 92, 95, 97, 101, 104, 105, 107, 116], "from_point": 10, "analog": 10, "to_arrai": 10, "spawned_on": [10, 105], "uniqu": [10, 12, 26, 47, 75, 80, 82, 84, 85, 104], "entiti": [10, 27], "spawn": [10, 80, 89, 92, 95, 119], "identifi": [10, 12, 26, 47, 75, 88, 92, 116], "purpos": [10, 37, 41, 44, 47, 50, 56, 72, 101, 113], "make_instance_cattr": 10, "cattr": [10, 32, 84], "requir": [10, 12, 23, 25, 32, 37, 46, 47, 48, 49, 50, 61, 62, 64, 66, 72, 74, 75, 87, 88, 89, 96, 102, 104, 107, 108], "becaus": [10, 23, 80, 84, 90, 92, 96, 97, 100, 102, 103, 106], "automat": [10, 11, 12, 24, 25, 41, 46, 48, 49, 54, 72, 75, 90, 98, 102, 110, 113], "detect": [10, 11, 12, 15, 25, 32, 46, 47, 48, 49, 72, 75, 80, 83, 87, 88, 89, 92, 97, 102, 104, 107, 108], "subset": [10, 12, 32, 65, 67, 82, 85, 89], "hook": [10, 84], "regist": [10, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 84, 104, 107, 108], "unstructur": [10, 69, 84], "read": [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 34, 50, 65, 67, 69, 89, 91, 95, 96, 100, 101, 104, 105, 106, 107, 108, 109, 112, 114, 115, 120], "h5": [11, 12, 25, 32, 47, 49, 72, 89, 92, 95, 96, 103, 104, 106, 107, 109, 111, 114, 118], "json": [11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 49, 50, 72, 84, 85, 87, 89, 91, 92, 96, 104, 106, 107, 109, 110], "zip": [11, 23, 49, 81, 91, 96, 104, 105, 107, 109, 110], "leap": [11, 24, 30, 37, 44, 47, 84, 92, 109, 117], "mat": [11, 84, 92], "deeplabcut": [11, 92, 100], "yaml": [11, 17, 85, 92], "deepposekit": [11, 92], "coco": [11, 25, 69, 92], "keypoint": [11, 15, 25, 48, 69, 73, 78, 80, 88, 92, 95, 107], "extens": [11, 12, 13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32], "input": [11, 32, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 52, 53, 54, 55, 56, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 70, 72, 73, 74, 75, 76, 80, 81, 82, 84, 85, 88, 89, 98, 100, 104, 105, 106, 107, 109], "export": [11, 12, 27, 46, 72, 90, 91, 92, 96, 100, 102, 103, 112, 113, 119, 120], "_": [11, 12, 26, 37, 69, 89, 103, 109], "python": [11, 12, 23, 27, 36, 54, 56, 68, 72, 78, 84, 85, 89, 90, 95, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 112, 113, 120], "describ": [11, 12, 37, 38, 42, 43, 49, 64, 72, 74, 75, 76, 81, 93, 97, 114, 120], "entrypoint": [11, 72], "collect": [12, 32, 49, 73, 100], "togeth": [12, 41, 47, 54, 64, 72, 79, 86, 95, 96, 104], "annot": [12, 14, 17, 27, 34, 92], "conveni": [12, 38, 50, 55, 61, 64, 65, 66, 75, 76, 77, 83, 84, 104, 105], "wai": [12, 38, 41, 64, 73, 75, 77, 90, 92, 95, 96, 97, 98, 102, 105, 107, 114, 115], "high": [12, 27, 32, 48, 49, 50, 54, 64, 72, 75, 80, 81, 104, 105, 113], "loader": 12, "load_fil": [12, 69, 104, 107, 108], "provid": [12, 14, 25, 32, 37, 38, 40, 41, 42, 43, 44, 46, 47, 48, 50, 52, 60, 61, 64, 66, 67, 69, 70, 72, 74, 75, 76, 77, 80, 81, 84, 86, 87, 89, 90, 92, 93, 102, 105, 107, 113, 115], "To": [12, 13, 25, 26, 28, 49, 53, 55, 61, 65, 69, 72, 76, 84, 89, 92, 94, 95, 96, 98, 100, 102, 104, 105, 106, 107, 109, 110, 114, 115, 117, 118], "disk": [12, 32, 45, 49, 50], "ve": [12, 80, 91, 92, 93, 94, 95, 96, 97, 98, 103, 107, 110, 112, 115, 118], "move": [12, 46, 49, 56, 75, 86, 92, 94, 95, 97, 115, 116, 119], "directori": [12, 14, 25, 32, 45, 49, 50, 72, 85, 89, 90, 92, 96, 102, 104, 106, 107, 108, 109, 110], "aren": [12, 79, 80, 93, 94, 95, 99], "tell": [12, 88, 92, 99, 100, 117], "search": [12, 44, 46, 72, 85, 87, 101, 102, 118], "There": [12, 87, 89, 91, 92, 95, 98, 103, 116, 119], "single_path_to_search": 12, "path_a": 12, "path_b": 12, "callback_funct": 12, "video_search": [12, 22, 23, 32], "make_video_callback": 12, "callback": [12, 32, 81], "window": [12, 80, 89, 92, 94, 95, 96, 98, 101, 102, 103, 104, 105, 106, 108, 117, 118, 119], "save_fil": [12, 25], "support": [12, 13, 14, 15, 16, 17, 18, 19, 21, 22, 23, 24, 25, 26, 27, 28, 29, 32, 35, 75, 84, 87, 88, 89, 100, 101, 112, 113, 118], "negative_anchor": 12, "front": 12, "interact": [12, 49, 72, 86, 102], "actual": [12, 32, 50, 67, 74, 80, 97, 104, 105, 106, 107, 108], "storag": [12, 46, 48, 49, 90, 104, 105, 106, 107], "backend": [12, 32, 104, 105, 106], "abstract": [12, 33, 70, 72, 80, 81], "awai": 12, "interfac": [12, 32, 72, 84, 92, 93, 96, 102, 104, 105, 109, 112, 113], "again": [12, 48, 106], "referenc": [12, 69, 84], "around": [12, 14, 34, 42, 46, 47, 48, 58, 63, 70, 72, 74, 76, 85, 92, 98, 103, 104, 105], "crop": [12, 34, 46, 47, 48, 52, 59, 60, 61, 65, 66, 72, 76, 88, 89, 92, 98, 104], "neg": [12, 69, 78, 104, 105, 106, 107, 108, 109], "add_inst": 12, "updat": [12, 32, 48, 49, 52, 66, 72, 73, 74, 78, 81, 88, 89, 92, 100, 101, 102, 106, 107], "add_suggest": 12, "add_track": 12, "add_video": 12, "all_inst": [12, 53, 61], "append_suggest": 12, "clear_suggest": 12, "delet": [12, 80, 81, 84, 92, 94, 104, 106, 109], "new_label": 12, "unifi": 12, "anyth": [12, 49, 89, 90, 102, 112], "redund": [12, 79, 80], "mayb": 12, "deep": [12, 14, 40, 84, 85, 100, 101, 104, 105, 106, 107, 109, 113], "admonit": 12, "slow": [12, 48, 49, 88, 92], "delete_suggest": 12, "basic": [12, 32, 38, 39, 60, 72, 85, 95, 99, 100, 105], "content": [12, 105, 109, 110], "mask": [12, 55, 58, 63, 72, 103], "export_csv": 12, "export_nwb": 12, "overwrit": [12, 26, 49, 81], "session_descript": [12, 26], "pose": [12, 26, 40, 47, 69, 75, 76, 80, 84, 87, 89, 93, 99, 105, 106, 107, 108, 109, 112, 113, 117], "session_start_tim": [12, 26], "datetim": [12, 26], "nwb": [12, 26], "pynwb": [12, 26], "nwbfile": [12, 26], "processingmodul": [12, 26], "sleap_video_": [12, 26], "03": [12, 26, 69, 103, 104, 105, 109], "video_fn": [12, 26], "stem": [12, 26, 37, 38, 39, 43, 47, 107], "isinst": [12, 26], "pathlib": [12, 26], "purepath": [12, 26], "ex": [12, 26, 72, 102], "path_to_video": [12, 26], "my_video": [12, 26], "mp4": [12, 23, 26, 32, 33, 86, 89, 96, 100, 104, 105, 108, 109, 110, 118], "003_my_video": [12, 26], "ndx_pose": 12, "poseestim": [12, 26], "under": [12, 26, 61, 79, 85, 101, 110, 116], "track_idx": [12, 26], "untrack": [12, 26, 78], "track001": [12, 26], "poseestimationseri": [12, 26], "covert": [12, 26], "boolean": [12, 26, 72, 83], "descript": [12, 26, 47, 50, 64, 68, 69, 84, 86, 104, 106, 109], "preexist": [12, 26], "guid": [12, 26, 72, 75, 89, 91, 96, 100, 101, 104, 107, 108, 117, 119], "THe": [12, 26], "extend_from": 12, "extract": [12, 37, 38, 41, 72, 85, 103], "ind": [12, 75], "param": [12, 24, 32, 39, 65], "slice": [12, 60, 82, 103], "__getitem__": 12, "preserv": [12, 46, 85], "even": [12, 37, 46, 47, 48, 53, 55, 60, 61, 67, 73, 78, 92, 95, 102, 116], "return_new": 12, "singleton": [12, 62, 68, 78], "criteria": [12, 46, 54, 66, 92, 107], "unless": [12, 25, 32, 85, 92, 115], "frame_index": [12, 32], "find_first": 12, "use_cach": 12, "occurr": 12, "instead": [12, 37, 39, 40, 43, 44, 47, 48, 52, 64, 68, 81, 89, 102, 107, 109, 110], "cach": [12, 32, 54, 78], "els": [12, 75, 110], "loop": [12, 47, 48, 49, 54, 75, 81, 101, 105, 106, 109, 112, 113, 117], "through": [12, 48, 50, 53, 54, 72, 75, 89, 90, 92, 95, 101, 104, 106, 107, 112, 113, 117, 120], "find_last": 12, "find_suggest": 12, "find_track_occup": 12, "pseudo": 12, "finish_complex_merg": 12, "resolved_fram": 12, "finish": [12, 49, 89, 94, 96, 98, 106, 109, 110, 115], "from_frame_idx": 12, "revers": [12, 42, 62], "start": [12, 41, 42, 43, 44, 45, 48, 49, 69, 75, 79, 80, 84, 85, 87, 88, 90, 91, 92, 93, 95, 99, 100, 102, 103, 104, 105, 106, 107, 109, 110, 112, 116, 120], "secondary_kei": 12, "raise_error": [12, 32], "safe": [12, 100], "rais": [12, 14, 32, 42, 49, 51, 60, 61, 64, 66, 67, 72, 77, 84, 85], "against": [12, 80, 85, 89, 95], "linear": [12, 37, 44, 47, 103], "supplement": 12, "keyerror": 12, "catch": [12, 72], "get_next_suggest": 12, "seek_direct": 12, "seek": [12, 100], "get_suggest": 12, "get_track_count": 12, "occupi": [12, 47], "get_track_occup": 12, "get_unlabeled_suggestion_ind": 12, "unlabel": [12, 65, 75, 89, 92, 113], "doe": [12, 25, 41, 43, 45, 47, 49, 52, 53, 55, 58, 60, 61, 64, 67, 84, 85, 89, 102, 104, 107, 108, 115], "remove_empty_fram": 12, "get_video_suggest": 12, "user_label": 12, "has_fram": 12, "lf": [12, 108], "must": [12, 14, 23, 32, 37, 42, 46, 47, 48, 49, 53, 55, 60, 66, 68, 69, 70, 72, 79, 80, 81, 82, 84, 85, 87, 94, 97, 100, 104, 105, 106, 107, 108, 109], "Not": [12, 37, 89, 102], "lookup": 12, "faster": [12, 40, 43, 72, 78, 89, 110, 115], "ones": [12, 43, 61, 67, 72, 75, 81, 102], "match_to": [12, 22, 23], "has_missing_video": 12, "filter": [12, 37, 38, 39, 40, 41, 42, 43, 44, 47, 54, 56, 65, 72, 75, 79, 80, 89, 103, 106, 108, 109], "is_multi_inst": 12, "kwarg": [12, 15, 17, 18, 21, 22, 23, 24, 25, 28, 29, 32, 34, 47, 65, 72, 73, 83], "search_path": 12, "use_gui": 12, "while": [12, 20, 25, 38, 47, 48, 54, 91, 92, 95, 98, 100, 102, 105, 115, 116, 117], "system": [12, 23, 32, 46, 49, 72, 89, 101, 102, 105, 106, 108, 109, 116], "signal": [12, 103], "abort": 12, "changed_on_load": 12, "extern": [12, 49, 69, 92, 100, 102], "merge_container_dict": 12, "dict_a": 12, "dict_b": 12, "merge_matching_fram": 12, "combin": [12, 27, 47, 54, 55], "merge_nod": 12, "base_nod": 12, "remain": [12, 46, 47, 68, 69, 72, 79, 95, 102], "might": [12, 78, 91, 94, 95, 102, 103, 105, 106, 108, 109, 110], "return_confid": 12, "arbitrari": [12, 32, 49, 56], "n_frame": 12, "n_track": [12, 58], "NOT": [12, 41], "intend": [12, 50, 81, 100, 101], "primarili": [12, 42, 66, 97], "remove_all_track": 12, "keep_empty_fram": 12, "remove_fram": 12, "update_cach": 12, "intern": [12, 80], "postpon": 12, "sequenc": [12, 37, 42, 44, 47, 64, 65, 67, 70, 92, 105, 110], "remove_inst": 12, "in_transact": 12, "remove_predict": 12, "prior": [12, 53, 54, 55, 70, 80, 92, 95, 106, 116], "prevent": [12, 37, 44, 47, 48, 49, 54, 72, 100, 104, 105], "overlap": [12, 78, 80, 89, 92, 95, 99], "remove_suggest": 12, "remove_track": 12, "remove_untracked_inst": 12, "remove_unused_track": 12, "remove_user_inst": 12, "remove_video": 12, "with_imag": [12, 104], "embed_all_label": [12, 104], "embed_suggest": 12, "embed": [12, 32, 38, 86, 104], "remot": [12, 89, 91, 101, 102, 117], "default_suffix": 12, "attempt": [12, 24, 32, 35, 49, 80], "suffix": [12, 49, 85, 89, 109], "valueerror": [12, 42, 49, 60, 61, 64, 66, 67, 77, 84, 85], "cannot": [12, 24, 25, 28, 32, 42, 47, 77, 84, 85, 97, 102, 104, 105, 106, 107, 108, 109, 110], "filetyp": 12, "save_frame_data_hdf5": 12, "png": [12, 22, 23, 32, 45, 49], "all_label": [12, 22, 23], "progress_callback": [12, 22, 23], "hdf5video": [12, 32], "progress": [12, 23, 27, 34, 40, 43, 72, 89, 115], "fn": [12, 69, 107], "n": [12, 64, 79, 82, 85, 89, 95, 101, 102, 103, 104, 105, 106, 107, 108, 109], "n_total": 12, "far": [12, 27, 75, 93, 95, 99, 107], "cancel": 12, "save_frame_data_imgstor": 12, "output_dir": 12, "imgstorevideo": [12, 32], "imgstor": [12, 23, 32], "lossless": [12, 32, 100], "jpg": [12, 23, 109], "lossi": [12, 33, 100], "probabl": [12, 47, 52, 55, 58, 71, 72, 76, 85, 88, 90, 103, 105, 107], "set_suggest": 12, "split": [12, 46, 49, 61, 65, 67, 69, 81, 85, 106, 107, 109], "randomli": [12, 41, 48, 92, 106], "fraction": [12, 46, 67, 72, 75, 89, 106, 109], "labels_a": 12, "labels_b": 12, "subsampl": [12, 68], "len": [12, 32, 105], "guarante": [12, 14, 17], "mutual": 12, "exclus": 12, "repeatedli": [12, 105], "py": [12, 25, 84, 89], "8": [12, 42, 47, 48, 54, 69, 80, 86, 89, 102, 103, 104, 105, 106, 107, 108, 109], "val": [12, 57, 69, 89, 106, 107, 109], "labels_train": [12, 67], "labels_val_test": 12, "labels_v": [12, 67], "labels_test": 12, "to_dict": [12, 84], "skip_label": 12, "underl": 12, "nest": [12, 67], "compos": [12, 52, 64], "primit": [12, 68, 72], "skip": [12, 36, 37, 38, 40, 41, 42, 43, 44, 47, 92, 102, 104, 107, 108], "anchor": [12, 47, 60, 70, 72, 76, 88, 117], "top": [12, 32, 37, 42, 47, 66, 71, 72, 73, 84, 88, 89, 91, 96, 98, 101, 104, 105, 109, 117], "to_json": [12, 50, 84], "string": [12, 13, 14, 25, 32, 37, 38, 42, 45, 46, 47, 49, 50, 54, 60, 62, 64, 66, 67, 68, 69, 70, 72, 75, 81, 83, 84, 85], "to_pipelin": [12, 32], "batch_siz": [12, 32, 48, 54, 72, 73, 89, 105, 109], "prefetch": [12, 32, 48, 54, 72, 109], "frame_indic": [12, 32], "user_labeled_onli": 12, "batch": [12, 32, 37, 38, 44, 45, 47, 48, 49, 54, 55, 71, 72, 73, 75, 76, 88, 98, 101, 105], "rank": [12, 32, 41, 53, 54, 55, 60, 62, 68, 72, 76], "tensor": [12, 32, 36, 37, 38, 39, 41, 42, 43, 44, 47, 52, 53, 54, 55, 56, 58, 60, 61, 62, 63, 64, 65, 66, 67, 68, 70, 71, 72, 73, 75, 76, 77, 81, 82, 89], "limit": [12, 32, 35, 41, 47, 61, 76, 80, 89, 92], "reader": [12, 32, 34, 65, 81], "tf": [12, 32, 37, 39, 41, 42, 45, 46, 48, 52, 53, 54, 55, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 70, 71, 72, 74, 75, 76, 77, 82, 102, 104, 106, 107, 109], "throughput": [12, 32], "o": [12, 32, 64, 65, 85, 89, 96, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110], "dure": [12, 32, 37, 45, 46, 47, 48, 49, 50, 54, 64, 70, 72, 73, 75, 89, 91, 95, 98, 102, 104, 105, 107, 109, 110, 115, 117], "labelsread": [12, 64, 65, 67, 69, 72, 81], "track_set_inst": 12, "new_track": 12, "track_swap": 12, "old_track": 12, "swap": [12, 48, 52, 80, 92, 95, 119], "unlabeled_suggest": 12, "user_labeled_frame_ind": 12, "user_labeled_fram": 12, "with_user_labels_onli": 12, "user_instances_onli": [12, 65], "with_track_onli": [12, 65, 81], "preprocess": [12, 41, 42, 46, 48, 62, 72, 74, 109], "warn": [12, 32, 49, 72, 104], "labelsdatacach": 12, "find_fancy_frame_idx": 12, "find_fram": 12, "get_filtered_frame_idx": 12, "get_frame_count": 12, "possibli": [12, 81, 91], "count": [12, 32, 44, 47, 48, 72, 75, 78, 79, 80, 89, 95, 103, 105], "rangelist": 12, "get_video_track_occup": 12, "rebuild": [12, 104, 105, 106, 107, 109], "update_counts_for_fram": 12, "find_path_using_path": 12, "missing_path": 12, "detect_video": 12, "direct": [12, 44, 47, 55, 70, 75, 76, 92], "reus": [12, 37], "handl": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 32, 45, 72, 75, 80, 84, 105, 106], "inherit": [13, 37], "implement": [13, 25, 32, 37, 38, 39, 40, 41, 42, 43, 45, 50, 62, 66, 69, 71, 72, 75, 79, 81, 84, 112], "all_ext": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "can_read_fil": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "filehandl": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "can_write_filenam": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "default_ext": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "does_match_ext": [13, 14], "does_read": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "does_writ": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29], "formatted_ext_opt": [13, 14], "qt": [13, 14, 34, 89], "dialog": [13, 14, 34, 88, 89, 91, 98, 110, 117, 119], "sleapobjecttyp": [13, 14, 26], "dispatch": [13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 26, 27, 28, 29], "human": [13, 14, 15, 16, 17, 18, 21, 22, 23, 24, 26, 27, 28, 29, 50, 76, 101, 113, 117], "adaptor": [14, 15, 16, 17, 18, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29], "manual": [14, 17, 48, 49, 92, 106, 110, 113, 115], "inspect": [14, 17, 64, 92, 93, 104, 105, 109, 110, 113], "especi": [14, 17, 47, 48, 49, 54, 61, 65, 88, 92], "alphatrackeradaptor": 14, "upper": 14, "third": 14, "fourth": [14, 92], "fifth": 14, "get_alpha_tracker_frame_dict": 14, "get_alpha_tracker_instance_dict": 14, "200": [14, 48, 103, 104, 109, 110], "coor": 14, "lower": [14, 33, 42, 47, 48, 69, 75, 79, 85, 88, 95, 100, 107], "left": [14, 46, 48, 52, 54, 66, 67, 69, 72, 84, 88, 92, 94, 105, 106, 107, 118], "hand": [14, 84, 91], "corner": [14, 82], "get_alpha_tracker_point_dict": 14, "make_video_for_image_list": [14, 17], "image_dir": [14, 17], "full_video": [14, 17], "being": [14, 33, 48, 55, 72, 73, 84, 100, 101], "souce_object": 14, "notimplementederror": [14, 25], "code": [14, 25, 40, 69, 73, 84, 90, 92, 95, 100, 102, 103, 110, 119], "seen": [14, 84], "cocodataset": [15, 69], "labelscocoadaptor": 15, "img_dir": 15, "use_missing_gui": 15, "csvadaptor": 16, "source_path": [16, 27, 28], "toget": [16, 28], "potenti": [17, 49, 54, 60, 72, 102, 106, 113], "labelsdeeplabcutcsvadaptor": 17, "labelsdeeplabcutyamladaptor": 17, "labelsdeepposekitadaptor": 18, "skeleton_path": 18, "dynam": [19, 72, 73, 102, 104, 107, 108], "u": [20, 35, 37, 89, 90, 98, 99, 100, 102], "keep": [20, 35, 46, 47, 48, 49, 54, 65, 66, 80, 85, 89, 92, 97, 115], "previou": [20, 49, 53, 55, 58, 80, 94, 95, 96, 102, 106, 119], "is_hdf5": 20, "is_json": 20, "is_open": 20, "text": [20, 45, 47, 50, 70, 72, 84, 89, 91, 96, 103, 106], "needn": [20, 31], "twice": 20, "format_id": 20, "distinguish": [20, 86], "common": [20, 25, 32, 33, 39, 41, 44, 56, 64, 75, 84, 92, 95, 100, 105], "labelsv1adaptor": [20, 22, 25], "gridlin": 20, "midpixel": 20, "veri": [21, 29, 48, 73, 92, 95, 100, 101], "genericjsonadaptor": [21, 25], "save_frame_data": [22, 23], "frame_data_format": [22, 23], "old": [23, 30, 31, 84, 86, 94], "kind": [23, 41, 99, 103], "now": [23, 62, 79, 87, 91, 95, 103, 105, 106, 109, 110, 115, 116], "labelsjsonadaptor": 23, "from_json_data": 23, "particular": [23, 41, 67, 92], "newli": [23, 30, 95, 117], "instanti": [23, 32, 37, 38, 41, 42, 43, 50, 64, 70, 73, 74, 80, 81, 84], "duplic": [23, 84, 85], "compress": [23, 33, 49, 100, 102], "shutil": 23, "make_arch": 23, "pkzip": 23, "forc": [23, 92, 105, 106, 109, 110], "archiv": [23, 109], "pgm": 23, "bmp": 23, "ppm": 23, "tif": 23, "npy": [23, 32], "mjpeg": [23, 32], "avi": [23, 32, 33, 96, 100, 118], "h264": [23, 100], "mkv": 23, "avc1": 23, "codec": 23, "exclud": [23, 47, 75, 78, 80, 85], "gpl": 23, "licens": [23, 41], "prompt": [24, 89, 91, 92, 102, 110, 118], "labelsleapmatlabadaptor": 24, "roughli": [25, 75, 86], "equival": [25, 36, 37, 54, 61, 68, 76], "principl": 25, "as_format": 25, "nice": 25, "virtual": [25, 90], "sleapanalysisadaptor": [25, 28], "typic": [25, 32, 37, 41, 44, 46, 47, 48, 50, 52, 64, 66, 69, 75, 77, 78, 88, 92, 96, 101, 106, 108, 113], "all_labels_adaptor": 25, "distinct": [25, 47, 79, 80, 88, 90, 92, 95, 96, 103, 111, 116, 117], "ext": 25, "foo": [25, 85], "exportanalysisfil": 25, "importanalysisfil": 25, "custom": [25, 72, 73, 78, 92, 93, 96, 101, 104, 105, 106, 112], "for_object": 25, "typeerror": 25, "shouldn": 25, "happen": [25, 77, 104], "relat": [25, 45, 46, 47, 48, 49, 50, 64, 67, 81, 84, 104], "thrown": 25, "privileg": 25, "ndx": 26, "ndxposeadaptor": 26, "nixadaptor": 27, "model": [27, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 48, 49, 50, 55, 58, 59, 61, 62, 64, 65, 66, 69, 70, 71, 72, 73, 75, 81, 82, 86, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 103, 104, 105, 106, 113, 115, 116, 117, 119], "scientif": 27, "librari": [27, 41, 52, 102, 104, 105, 106, 107, 108, 109], "definit": [27, 38, 42, 70], "nixio": 27, "low": [27, 37, 44, 48, 72, 86, 92, 105, 106], "pip": [27, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110], "nixtrack": 27, "bendalab": 27, "addition": [27, 41, 47, 52, 107], "chunk": [27, 34], "approach": [27, 47, 69, 72, 75, 82, 88, 92, 105, 110, 113, 117], "avoid": [27, 53, 80, 116], "memori": [27, 40, 43, 46, 47, 48, 53, 54, 61, 72, 77, 88, 89, 98, 102, 104, 105, 106, 109], "larg": [27, 47, 48, 49, 53, 61, 88, 92, 95, 100, 101], "author": [27, 41, 100, 101, 110], "jan": 27, "grew": 27, "lack": 28, "textadaptor": 29, "load_labels_json_old": 30, "data_path": [30, 89, 109], "parsed_json": 30, "adjust_matlab_index": 30, "fix_rel_path": 30, "pars": [30, 50, 72, 85], "adjust": [30, 42, 45, 46, 48, 52, 53, 59, 63, 66, 72, 75, 110], "absolut": [30, 32, 36, 46, 48, 49, 61, 65, 75], "load_predicted_labels_json_old": 30, "filenames_prefix_chang": 31, "old_prefix": 31, "new_prefix": 31, "confirm_callback": 31, "initi": [31, 35, 37, 38, 39, 41, 42, 43, 47, 48, 49, 54, 55, 65, 72, 75, 77, 79, 80, 81, 89, 99, 101, 103, 105, 106, 109, 112, 113, 117, 118, 120], "known": 31, "confirm": [31, 35, 92, 102], "find_changed_subpath": 31, "old_path": 31, "new_path": [31, 69], "section": [31, 95, 113], "list_file_miss": 31, "dummyvideo": 32, "2000": [32, 86, 89], "10000": 32, "dummi": [32, 72], "fake": 32, "real": [32, 72, 75], "input_format": [32, 89, 96, 104, 109], "channels_last": [32, 42, 96, 104, 109], "convert_rang": 32, "file_h5": 32, "dataset_h5": 32, "equal": [32, 38, 39, 40, 43, 44, 48, 54, 61, 66, 78, 83, 89], "channels_first": [32, 42], "255": [32, 42, 46, 48, 62, 92, 109], "attr": [32, 51, 84, 104, 107], "embedded_frame_ind": 32, "enable_source_video": 32, "get_fram": 32, "has_embedded_imag": 32, "last_frame_idx": 32, "overrid": [32, 34, 45, 89], "reset": [32, 54, 82, 89, 102, 117], "reload": [32, 104, 105, 106, 108], "source_video": 32, "source_video_avail": 32, "uncach": 32, "index_by_origin": 32, "loopbio": 32, "lightweight": 32, "great": [32, 106, 107, 112], "larger": [32, 47, 48, 61, 72, 75, 88, 89, 98, 100], "directli": [32, 44, 46, 47, 54, 72, 75, 81, 89, 92, 102, 103, 105, 107, 110, 118], "frame_numb": 32, "That": 32, "comapar": 32, "mediavideo": [32, 104, 105], "grayscal": [32, 41, 42, 46, 62, 65, 72, 104, 118], "bgr": [32, 42, 46, 62, 104], "tradit": 32, "readabl": 32, "ffmpeg": [32, 33, 100], "bare": 32, "minimum": [32, 48, 53, 58, 62, 63, 69, 70, 72, 73, 75, 76, 79, 80, 89], "opencv": [32, 33, 34, 96, 104, 105, 106, 107, 108, 109, 110], "videocaptur": 32, "auto": [32, 49, 89, 92, 103, 109], "color": [32, 34, 83, 86, 89, 92, 115, 119], "blue": [32, 88, 98, 115, 116], "green": [32, 87, 115], "red": [32, 95, 115], "fp": [32, 33, 34, 69, 86, 89, 101, 107, 109], "numpyvideo": 32, "is_miss": 32, "come": [32, 37, 47, 52, 65, 96, 102, 107], "singleimagevideo": 32, "height_": 32, "width_": 32, "channels_": 32, "compon": [32, 44, 75, 80], "from_hdf5": 32, "from_media": 32, "Or": [32, 100, 102, 117], "from_filenam": [32, 65], "rgb": [32, 41, 42, 46, 62, 72], "serialiaz": 32, "fixup_path": 32, "raise_warn": 32, "portabl": [32, 81], "whatev": [32, 94, 96], "almost": [32, 119], "certainli": 32, "wrong": [32, 95], "transfer": [32, 42, 106], "prefer": [32, 47, 50, 79, 89, 92, 97, 99, 100, 102, 118], "asset": [32, 109], "filenotfounderror": [32, 85], "still": [32, 47, 92, 95, 97, 99, 106], "motif": 32, "helper": [32, 85], "invok": [32, 34, 73, 102, 110], "from_image_filenam": 32, "request": [32, 54, 103, 104, 109], "get_frames_saf": 32, "successfulli": [32, 80, 112, 115], "imgstore_from_filenam": 32, "output_filenam": 32, "num_fram": 32, "to_hdf5": 32, "encod": [32, 37, 38, 40, 41, 43, 44, 47, 50, 76, 83, 84, 85, 89, 92, 95, 100], "to_imgstor": 32, "facilit": [32, 75], "directoryimgstor": 32, "videoread": [32, 65], "available_video_ext": 32, "load_video": [32, 104, 105], "nonetyp": 32, "treat": 32, "effect": [32, 36, 37, 38, 39, 40, 42, 43, 44, 46, 47, 48, 53, 54, 55, 58, 60, 62, 72, 98, 119], "centered_pair_smal": 32, "1100": 32, "384": [32, 105], "img": [32, 33, 34, 72, 76, 80, 83, 104, 105, 108], "writer": [33, 34], "safe_build": 33, "add_fram": 33, "can_use_ffmpeg": 33, "depend": [33, 44, 49, 53, 58, 68, 88, 90, 92, 96, 100, 102, 104, 105, 106, 107, 108, 110, 113], "videowriterimageio": 33, "crf": [33, 100], "preset": [33, 100], "superfast": [33, 100], "imageio": 33, "movi": [33, 86], "playback": [33, 100], "framer": 33, "rate": [33, 42, 45, 46, 48, 49, 72, 91, 96, 106, 109], "factor": [33, 39, 40, 42, 43, 44, 46, 47, 48, 53, 61, 63, 65, 66, 69, 72, 73, 80, 83, 107], "32": [33, 37, 41, 42, 46, 47, 69, 86, 105, 106, 109], "18": [33, 39, 69, 86, 106, 109], "30": [33, 86, 100, 104, 105, 109], "less": [33, 47, 67, 100], "higher": [33, 37, 38, 39, 40, 43, 44, 46, 47, 48, 54, 58, 68, 72, 74, 88, 95, 107], "qualiti": [33, 100, 113], "libx264": [33, 100], "videowriteropencv": 33, "overlai": 34, "videomarkerthread": 34, "in_q": 34, "queue": [34, 80], "out_q": 34, "show_edg": [34, 89], "edge_is_wedg": [34, 89], "marker_s": [34, 89], "crop_size_xi": 34, "color_manag": 34, "colormanag": 34, "palett": [34, 89, 92], "distinctly_color": [34, 89], "draw": [34, 89], "send": [34, 35], "h": [34, 83, 86, 89, 100, 101], "w": [34, 83, 100, 101, 102, 104, 105, 107, 108, 113], "thread": [34, 102], "activ": [34, 37, 38, 41, 42, 44, 47, 70, 98, 101, 102, 113, 115, 118], "subclass": [34, 45, 72, 81], "constructor": [34, 65, 72], "sequenti": [34, 49, 100], "taken": 34, "img_to_cv": 34, "prepar": [34, 110], "background": [34, 86, 89, 95], "them": [34, 47, 48, 53, 54, 55, 66, 75, 81, 84, 90, 92, 93, 94, 95, 97, 99, 102, 104, 105, 106, 107, 109, 110, 112, 113, 117], "white": [34, 86, 89], "grei": [34, 89, 115, 119], "resize_imag": [34, 66], "resiz": [34, 46, 61, 65, 72], "save_labeled_video": 34, "15": [34, 47, 69, 70, 88, 98, 100, 101, 103, 104, 105, 106, 107, 108, 109, 110], "gui_progress": 34, "wedg": [34, 89, 92, 95], "marker": [34, 89], "alphabet": [34, 89, 92, 95], "five": [34, 89, 92, 95], "solar": [34, 89], "progress_queu": 34, "receiv": [34, 35, 41, 49, 105, 106, 117], "elaps": [34, 80, 89], "done": [34, 54, 60, 61, 67, 72, 80, 87, 90, 112, 117, 119], "zmq": [35, 49, 89, 109], "pub": [35, 49], "socket": [35, 49], "pairedsend": 35, "pairedreceiv": 35, "handshak": 35, "chanc": 35, "earli": [35, 46, 48, 49, 106, 109, 110], "basemessageparticip": 35, "tcp": [35, 45, 49, 109], "127": [35, 42, 45, 49, 62, 109], "9001": [35, 45, 49, 109], "sender": 35, "message_queu": 35, "check_messag": 35, "timeout": [35, 49, 90], "10": [35, 40, 45, 48, 49, 69, 72, 75, 80, 83, 102, 103, 104, 105, 106, 107, 108, 109, 117], "fresh": [35, 102], "times_to_check": 35, "sent": [35, 103, 104, 109], "until": [35, 98, 104, 109, 110, 113, 117, 118], "reach": [35, 54, 99, 100, 101], "push_back_messag": 35, "didn": [35, 100, 103, 109, 116], "send_arrai": 35, "header_data": 35, "flag": [35, 83, 89, 104, 105, 106, 107, 109], "send_dict": 35, "intermediatefeatur": [36, 37, 38, 41, 42, 44], "intermedi": [36, 37, 38, 41, 42, 44, 47, 72, 92], "resolut": [36, 37, 38, 39, 41, 44, 46, 47, 48, 83, 86], "layer": [36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 66, 67, 70, 72, 82, 106, 109], "tensorflow": [36, 42, 54, 56, 61, 72, 76, 77, 81, 82, 89, 90, 100, 102, 104, 105, 106, 107, 108, 109, 117], "framework": [36, 54, 56, 101], "op": [36, 54, 56, 72, 73, 77, 104, 107], "reciproc": [36, 47, 53, 55, 58, 68], "decod": [37, 38, 40, 41, 43, 44, 47, 50, 84, 85, 103], "fulli": [37, 41, 42, 43, 47, 70], "convolut": [37, 38, 39, 40, 41, 42, 43, 44, 47], "backbon": [37, 38, 39, 41, 42, 44, 47, 74, 98, 106, 109, 117], "block": [37, 38, 39, 40, 41, 42, 43, 47, 64], "network": [37, 38, 40, 41, 42, 43, 44, 46, 47, 48, 49, 60, 72, 74, 75, 76, 92, 96, 98, 101, 104, 105, 106, 107, 109, 117], "respons": [37, 49, 89, 103, 104, 109], "upsampl": [37, 38, 39, 40, 41, 42, 43, 47, 66], "achiev": [37, 42, 47, 88, 113, 117], "pattern": [37, 44, 45, 49], "generaliz": 37, "rubric": 37, "cnn": [37, 40, 44], "www": [37, 72, 78, 104, 106, 107, 108], "natur": [37, 86, 100, 101], "com": [37, 78, 82, 102, 103, 104, 105, 106, 107, 108, 109], "articl": [37, 100, 101], "s41592": 37, "018": 37, "0234": 37, "net": [37, 98], "residu": [37, 38, 39, 42], "hourglass": [37, 44, 47, 76, 109], "arxiv": [37, 39, 42, 107], "pdf": [37, 39, 53, 68, 76, 105], "1603": 37, "06937": 37, "dens": [37, 54, 92], "fc": 37, "densenet": [37, 46, 62], "1611": 37, "09326": 37, "variant": [37, 39, 42, 47, 64], "encoderdecod": 37, "decoderblock": [37, 38], "upsampling_strid": [37, 38, 44], "doubl": [37, 44, 47, 92, 94, 97, 115, 116, 118, 119], "fewer": [37, 44, 47, 54, 67, 75, 80, 97], "make_block": [37, 38, 43], "current_strid": [37, 38, 42, 44], "skip_sourc": [37, 38, 44], "prefix": [37, 38, 39, 43, 45, 69, 85, 89, 109], "x_in": [37, 38, 39, 41, 42, 43, 70], "configur": [37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 52, 60, 61, 62, 64, 66, 70, 72, 74, 80, 81, 89, 91, 92, 93, 96, 98, 99, 105, 106, 107, 110, 113, 117], "encoderblock": [37, 38], "pooling_strid": [37, 38, 43], "encoder_stack": [37, 38, 40, 43], "decoder_stack": [37, 38, 40, 43], "greater": [37, 44, 46, 47, 68, 72, 75, 89, 95], "than": [37, 38, 44, 46, 47, 48, 49, 51, 53, 54, 55, 58, 60, 66, 67, 68, 72, 75, 78, 79, 80, 84, 88, 89, 92, 95, 96, 103, 110, 115], "decoder_features_strid": 37, "encoder_features_strid": 37, "make_backbon": [37, 41, 42], "make_decod": 37, "skip_source_featur": 37, "dec": 37, "At": [37, 48, 98], "make_encod": 37, "enc": 37, "perform": [37, 38, 39, 44, 45, 46, 47, 48, 49, 50, 54, 59, 61, 69, 72, 73, 75, 77, 78, 80, 82, 88, 89, 99, 102, 104, 106, 107, 109, 113], "make_stem": 37, "unlik": 37, "rest": [37, 84, 105], "maximum_strid": [37, 41, 42, 61, 74], "maximum": [37, 40, 41, 42, 43, 48, 53, 58, 62, 65, 66, 68, 72, 73, 74, 75, 76, 78, 80, 89, 106], "divis": [37, 42, 46, 61, 66, 72], "output_strid": [37, 41, 42, 44, 47, 53, 55, 58, 68, 70, 72, 106, 109], "stem_features_strid": 37, "stem_stack": [37, 38, 43], "simpleconvblock": [37, 40, 43], "pool_before_conv": 37, "num_conv": 37, "kernel_s": [37, 38, 42, 43, 76, 106, 109], "use_bia": [37, 42], "batch_norm": [37, 47], "batch_norm_before_activ": 37, "relu": [37, 38, 44, 47], "block_prefix": 37, "flexibl": [37, 101, 105], "below": [37, 43, 46, 48, 53, 58, 61, 63, 69, 70, 73, 75, 76, 86, 92, 95, 100, 102, 106, 108, 110, 113], "kernel": [37, 38, 42, 43, 44, 47, 76], "bia": 37, "term": [37, 47, 70], "conv": [37, 38, 42, 43, 44, 47], "bn": 37, "mini": [37, 48, 54, 73], "creation": [37, 42, 50], "leapcnn": [37, 40], "unet": [37, 41, 44, 47, 106, 109, 117], "conv_block": [37, 43], "simpleupsamplingblock": [37, 40, 43], "transposed_conv": [37, 44, 47], "transposed_conv_filt": [37, 44], "64": [37, 39, 40, 42, 43, 44, 47, 54, 69, 70, 102, 103, 104, 107, 108], "transposed_conv_kernel_s": [37, 44, 47], "transposed_conv_use_bia": 37, "transposed_conv_batch_norm": 37, "transposed_conv_batch_norm_before_activ": 37, "transposed_conv_activ": 37, "interp_method": [37, 38], "bilinear": [37, 38, 39, 40, 43, 44, 47, 104, 107], "skip_connect": [37, 42, 47], "skip_add": [37, 44], "refine_conv": [37, 44, 47], "refine_convs_first_filt": 37, "refine_convs_filt": [37, 44], "refine_convs_use_bia": 37, "refine_convs_kernel_s": 37, "refine_convs_batch_norm": 37, "refine_convs_batch_norm_before_activ": 37, "refine_convs_activ": 37, "refin": [37, 38, 44, 47, 53, 61, 64, 70, 72, 76, 81, 99, 113], "learnabl": [37, 44, 47], "interpol": [37, 38, 40, 43, 47, 75, 95, 103], "transposed_conv_": 37, "nearest": [37, 38, 75, 92, 103], "kera": [37, 39, 41, 42, 45, 46, 48, 59, 62, 64, 68, 72, 74, 81, 82, 104, 106, 107, 109], "upsampling2d": 37, "1x1": [37, 44, 47], "similar": [37, 38, 44, 47, 48, 54, 69, 78, 80, 87, 89, 90, 92, 95, 96, 107, 108, 109, 113], "shortcut": [37, 42, 85, 92, 94, 95, 104], "concaten": [37, 44, 47, 54, 64], "channelwis": [37, 73, 76], "serv": [37, 38, 44, 47, 50, 61, 73, 76, 81, 97, 105], "mix": [37, 44, 46, 47], "fuse": [37, 44, 47], "alias": [37, 44, 47], "checkerboard": [37, 44, 47], "refine_convs_": 37, "prepend": [37, 38, 45, 49, 68, 89], "downsamplingblock": 38, "256": [38, 39, 41, 47, 48, 52, 105, 109], "downsampl": [38, 39, 41, 42, 43, 47, 66, 88], "simplifi": [38, 72], "paper": [38, 40, 69, 86, 100, 101], "simpler": [38, 44], "demonstr": [38, 107, 112], "simpli": [38, 48, 73, 100, 102, 105, 106, 107, 108, 118], "maxpool": 38, "down_block": [38, 39, 40, 41, 42, 43, 106, 109], "up_block": [38, 40, 41, 43, 106, 109], "stem_filt": [38, 39, 42, 47], "128": [38, 41, 42, 47, 48, 83, 103, 109], "stem_strid": [38, 47, 109], "filter_increas": [38, 47], "upon": [38, 102], "primari": 38, "x_out": [38, 42], "symmetr": [38, 48, 52, 76, 84], "reduc": [38, 43, 47, 48, 49, 53, 60, 73, 83, 84, 109], "fine": [38, 42, 48, 73, 88, 116], "increas": [38, 40, 43, 46, 47, 48, 49, 53, 55, 58, 61, 72, 80, 83, 88, 116], "spatial": [38, 40, 47], "recept": [38, 40, 42, 43, 47, 88, 98], "lose": [38, 89, 90, 116], "subsequ": [38, 43, 61, 92, 95, 102, 104, 105, 106, 108, 116, 119], "linearli": [38, 47], "increment": [38, 80], "multipli": [38, 53, 55, 61], "neighbor": [38, 75, 84], "supervis": [38, 65], "from_config": [38, 40, 41, 42, 43, 44, 52, 60, 61, 62, 66, 70, 73, 74, 75, 81, 106], "config": [38, 40, 41, 42, 43, 44, 52, 60, 61, 62, 64, 66, 70, 72, 73, 74, 75, 77, 81, 85, 87, 89, 96, 102, 106, 110], "hourglassconfig": [38, 47], "stemblock": 38, "output_filt": 38, "7": [38, 69, 86, 89, 102, 103, 104, 105, 106, 107, 108, 109], "much": [38, 39, 46, 90, 92, 93, 102, 105], "first_conv_strid": 38, "upsamplingblock": 38, "integr": [38, 59, 72, 75, 76, 101], "x_up": 38, "skip_in": 38, "x_middl": 38, "smaller": [38, 47, 53, 55, 58, 63, 66, 68, 69, 75, 98, 116], "sublay": 38, "pytorch": 39, "higherhrnet": 39, "deconvolut": 39, "ref": [39, 78], "1902": 39, "09212": 39, "1908": 39, "10357": 39, "initial_downsampling_step": 39, "n_deconv_modul": 39, "bottleneck": [39, 42, 43, 48, 54], "deconv_filt": 39, "bilinear_upsampl": 39, "hrnet32": 39, "30m": [39, 104, 105], "highest": [39, 48, 71, 89], "decreas": [39, 40, 43, 47, 48, 54, 80, 88], "introduc": [39, 43, 47], "too": [39, 48, 54, 79, 95, 113], "loss": [39, 40, 47, 48, 49, 70, 81, 88, 98, 106, 109, 110, 117], "deconv": 39, "n_output_channel": 39, "deconv_modul": 39, "higher_hrnet_model": 39, "output_scal": [39, 42, 83], "adjust_prefix": 39, "name_prefix": 39, "delimit": 39, "bottleneck_block": 39, "expansion_r": 39, "simple_block": 39, "downsampling_lay": 39, "filters_r": [40, 43, 47, 106, 109], "down_convs_per_block": 40, "up_interpol": [40, 43, 47, 106, 109], "up_convs_per_block": 40, "fast": [40, 42, 80, 101, 104, 105], "estim": [40, 47, 69, 75, 76, 80, 84, 107, 109, 112, 113, 117], "neural": [40, 41, 75, 98, 101, 104, 105, 106, 107, 109], "2019": [40, 86, 101], "style": [40, 41, 42, 92, 95, 107], "pereira": [40, 86, 100, 101], "et": [40, 43, 86, 95, 101, 117], "al": [40, 43, 86, 95, 101, 117], "8m": 40, "capac": [40, 42, 43, 47], "runtim": [40, 43, 50, 77, 90, 104, 105, 106, 108, 109], "incur": 40, "precis": [40, 47, 69, 92, 107, 113], "abl": [40, 43, 47, 87, 89, 90, 92, 101, 102, 105, 106, 109, 110, 112, 117, 118], "learn": [40, 42, 43, 45, 46, 47, 48, 49, 71, 76, 88, 91, 96, 98, 100, 106, 109, 113, 115, 116], "richer": [40, 43], "complex": [40, 43, 47, 100, 102], "recov": [40, 43, 47, 61, 105, 106, 109, 110], "leapconfig": [40, 47], "pretrain": [41, 42, 46, 47, 49, 62, 101, 106], "wide": [41, 69, 98, 107], "state": [41, 73, 77, 84, 89, 101, 102], "art": [41, 101], "advantag": [41, 47, 72, 88, 105], "imagenet": [41, 42, 46, 47, 62], "weight": [41, 42, 47, 48, 49, 55, 70, 72, 75, 76, 78, 84, 89, 96, 104, 106, 107], "possibl": [41, 72, 75, 79, 80, 88, 95, 96, 97, 100, 101, 106], "pavel": 41, "yakubovskii": 41, "gracious": 41, "put": [41, 75, 85, 95], "info": [41, 72, 87, 95, 100, 104, 106, 109], "repositori": [41, 86, 102, 109, 112], "qubvel": 41, "segmentation_model": 41, "classification_model": 41, "mit": [41, 101], "copyright": 41, "2018": [41, 95], "herebi": 41, "grant": [41, 100], "charg": 41, "obtain": [41, 69, 72], "deal": 41, "restrict": [41, 63, 85], "distribut": [41, 47, 48, 53, 55, 58, 68, 76, 92, 102, 107], "sublicens": 41, "sell": 41, "permit": 41, "whom": 41, "furnish": 41, "subject": 41, "condit": 41, "notic": 41, "shall": 41, "substanti": [41, 61], "portion": [41, 44], "THE": 41, "AS": 41, "warranti": 41, "OF": 41, "OR": 41, "impli": [41, 54], "BUT": [41, 104], "TO": 41, "merchant": 41, "FOR": 41, "AND": 41, "noninfring": 41, "IN": 41, "NO": 41, "holder": 41, "BE": 41, "liabl": 41, "claim": 41, "damag": 41, "liabil": 41, "contract": 41, "tort": 41, "aris": 41, "WITH": 41, "unetpretrainedencod": 41, "efficientnetb0": [41, 47], "decoder_filt": [41, 47], "varieti": [41, 96, 104, 105, 115], "popular": 41, "special": [41, 49, 72, 90], "tile": [41, 42], "vgg16": [41, 47], "vgg19": [41, 47], "resnet18": [41, 47], "resnet34": [41, 47], "resnet50": [41, 42, 47], "resnet101": [41, 42, 47], "resnet152": [41, 42, 47], "resnext50": [41, 47], "resnext101": [41, 47], "inceptionv3": [41, 47], "inceptionresnetv2": [41, 47], "densenet121": [41, 47], "densenet169": [41, 47], "densenet201": [41, 47], "seresnet18": [41, 47], "seresnet34": [41, 47], "seresnet50": [41, 47], "seresnet101": [41, 47], "seresnet152": [41, 47], "seresnext50": [41, 47], "seresnext101": [41, 47], "senet154": [41, 47], "mobilenet": [41, 47], "mobilenetv2": [41, 46, 47, 62], "efficientnetb1": [41, 47], "efficientnetb2": [41, 47], "efficientnetb3": [41, 47], "efficientnetb4": [41, 47], "efficientnetb5": [41, 47], "efficientnetb6": [41, 47], "efficientnetb7": [41, 47], "lowest": [41, 48, 73, 80, 92, 109, 110], "therefor": [41, 72], "pretrainedencoderconfig": [41, 47], "x_main": 41, "intermediate_activ": 41, "intermediateactiv": 41, "atrou": 42, "deeplabv2": 42, "ab": [42, 107], "1606": 42, "00915": 42, "applic": [42, 46, 47, 62, 65, 72, 76, 87, 89, 103, 104, 105, 106, 109, 112], "upsampling_stack": 42, "upsamplingstack": [42, 44], "features_output_strid": 42, "16": [42, 44, 46, 47, 69, 78, 104, 105, 106, 107, 108, 109], "frozen": [42, 47, 72, 89], "model_nam": [42, 72], "stack_config": 42, "23": [42, 69, 86, 100, 105, 106, 109], "canon": [42, 43, 44], "compens": [42, 43, 48], "dilat": [42, 47], "expans": [42, 54], "random": [42, 47, 48, 49, 52, 67, 98, 116, 117], "trainabl": [42, 47, 74, 88, 105], "tune": [42, 46, 48, 73, 99], "reli": [42, 47, 50, 72, 80], "suffici": [42, 47, 88], "resnetv1": [42, 46, 47, 62], "v2": [42, 105], "6": [42, 69, 83, 89, 100, 102, 103, 104, 105, 106, 107, 108, 109, 110], "stack_fn": 42, "consecut": [42, 47, 75], "expand": [42, 46, 62, 68], "stride1": 42, "resnetconfig": [42, 47], "block_v1": 42, "dilation_r": 42, "conv_shortcut": 42, "imagenet_preproc_v1": 42, "accord": [42, 62, 67, 106], "caff": [42, 46, 62], "subtract": [42, 46, 62], "103": [42, 62, 104, 107], "939": [42, 62, 105, 106, 109], "116": [42, 62], "779": [42, 62], "123": [42, 62, 96], "68": [42, 62, 69, 104, 105], "make_backbone_fn": 42, "signatur": [42, 72, 79, 109], "intermediate_feat": 42, "backbone_fn": 42, "make_resnet_model": 42, "preact": 42, "input_tensor": 42, "input_shap": [42, 74], "stem_stride1": 42, "stem_stride2": 42, "resnetv2": [42, 46, 62], "resnext": 42, "pre": [42, 47, 77, 80, 89, 90, 92, 105, 108, 109], "bias": 42, "include_top": 42, "224": [42, 62, 105], "exactli": [42, 61, 69, 76, 80, 88, 89, 95, 100, 103], "invalid": [42, 79], "stack_v1": 42, "tile_channel": 42, "functon": 42, "replic": [42, 61], "monochrom": 42, "poolingblock": 43, "stem_kernel_s": [43, 106, 109], "convs_per_block": [43, 106, 109], "stem_block": [43, 106, 109], "middle_block": [43, 47, 106, 109], "block_contract": [43, 106, 109], "ronneberg": 43, "2015": [43, 78], "34": [43, 69, 106, 109], "5m": [43, 104, 105], "middl": 43, "bear": 43, "particularli": [43, 65, 75, 77], "unetconfig": [43, 47, 106], "consist": [44, 47, 97, 110, 116], "3x3": [44, 47, 61, 76], "suffic": 44, "dlc": 44, "transposed_conv_filters_r": 44, "transposed_conv_batchnorm": 44, "make_skip_connect": 44, "refine_convs_filters_r": 44, "refine_convs_batchnorm": 44, "make_stack": 44, "No": [44, 46, 47, 48, 62, 72, 95, 102, 104, 107, 108], "won": [44, 47, 79, 80, 88, 92, 95, 98], "norm": [44, 47, 103], "incom": [44, 47, 52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 66, 67], "wise": [44, 47, 54, 58, 75, 78], "upsamplingconfig": [44, 47], "log": [44, 45, 49, 72, 89, 105, 106, 109, 110], "indirectli": 44, "matplotlibsav": 45, "save_fold": 45, "plot_fn": 45, "render": [45, 49, 72], "matplotlib": [45, 83, 103, 105, 107], "visual": [45, 48, 49, 61, 64, 72, 81, 83, 87, 89, 92, 93, 95, 97, 98, 105, 106, 107, 109, 115, 117], "epoch": [45, 48, 49, 54, 88, 98, 106, 109, 110, 117], "trainer": [45, 81, 106, 109], "visualize_predict": 45, "on_epoch_end": 45, "modelcheckpointonev": 45, "filepath": [45, 46, 102], "train_end": 45, "checkpoint": [45, 47, 49, 81, 89, 104, 109], "trigger": [45, 77], "train_start": 45, "on_train_begin": 45, "on_train_end": 45, "progressreporterzmq": 45, "not_set": 45, "on_batch_begin": 45, "backward": [45, 84, 92], "compat": [45, 61, 72, 84, 96, 100, 102], "on_train_batch_begin": 45, "on_batch_end": 45, "on_train_batch_end": 45, "on_epoch_begin": 45, "mode": [45, 46, 62, 72, 73, 88, 96, 102, 108, 110, 117], "futur": [45, 96, 115], "metric": [45, 46, 49, 69, 73, 81, 89, 92, 106, 107, 109, 112], "val_": 45, "tensorboardmatplotlibwrit": 45, "log_dir": 45, "tag": [45, 49, 109], "summari": [45, 49, 77, 89, 107], "logdir": 45, "tensorboard": [45, 49, 81, 89, 109], "trainingcontrollerzmq": 45, "9000": [45, 49, 109], "topic": 45, "poll_timeout": 45, "set_lr": [45, 49], "lr": [45, 49, 106, 109], "nativ": [45, 68, 89], "schedul": [45, 48, 49, 106, 109], "dataconfig": [46, 50, 64], "labelsconfig": [46, 81], "preprocessingconfig": [46, 62, 66], "instance_crop": [46, 106, 109], "instancecroppingconfig": [46, 60, 61], "topdown": [46, 47, 61, 72, 106, 109, 117], "center_on_part": [46, 60, 106, 109], "crop_siz": [46, 61, 72, 76, 109], "crop_size_detection_pad": [46, 109], "largest": [46, 47, 48, 54, 61], "pad": [46, 54, 61, 65, 66, 68, 72, 89], "training_label": [46, 81, 106, 109], "validation_label": [46, 81, 109], "validation_fract": [46, 67, 106, 109], "test_label": [46, 81, 89, 109], "split_by_ind": [46, 109], "training_ind": [46, 109], "validation_ind": [46, 109], "test_ind": [46, 109], "search_path_hint": [46, 109], "hyperparamet": [46, 48, 91, 93, 98, 99, 106], "stop": [46, 48, 49, 98, 106, 109, 110, 117], "held": 46, "never": [46, 54, 94], "benchmark": [46, 69, 86, 104, 105, 112], "optim": [46, 47, 49, 50, 54, 64, 70, 73, 75, 81, 82, 88, 100, 104, 105, 106, 107, 108, 109], "explicitli": [46, 48, 49, 50], "packag": [46, 81, 85, 89, 90, 91, 94, 96, 101, 103, 104, 105, 106, 107, 108, 109, 114], "pull": [46, 64, 75, 109], "ensure_rgb": [46, 62, 109], "ensure_grayscal": [46, 62, 72, 109], "imagenet_mod": [46, 62, 109], "input_sc": [46, 61, 109], "pad_to_strid": [46, 66, 72, 109], "resize_and_pad_to_target": [46, 66, 109], "target_height": [46, 66, 109], "target_width": [46, 66, 109], "commonli": [46, 62, 76], "mobilenetv1": [46, 62], "nasnet": [46, 62], "flip": [46, 48, 52, 62], "torch": [46, 62], "input_scal": [46, 72], "consider": [46, 47, 48, 49, 54, 79, 85, 88], "accuraci": [46, 47, 69, 80, 106, 107, 108, 113, 115, 117], "bottom": [46, 47, 66, 72, 84, 88, 89, 91, 92, 97, 101, 104, 110, 117], "meet": [46, 54, 64, 66, 92], "architectur": [46, 47, 49, 50, 61, 74, 91, 96, 98, 101, 104, 109], "dimens": [46, 47, 53, 55, 62, 65, 66, 68, 69, 70, 72, 75, 82, 103], "aspect": [46, 65, 103], "ratio": [46, 48, 65, 66, 72, 73, 82], "backboneconfig": 47, "resnet": [47, 109, 117], "centeredinstanceconfmapsheadconfig": [47, 70, 106], "anchor_part": [47, 70, 106, 109], "part_nam": [47, 70, 75, 81, 106, 109], "sigma": [47, 53, 55, 58, 68, 70, 76, 106, 109], "loss_weight": [47, 70, 106, 109], "offset_refin": [47, 109], "assumpt": 47, "reliabl": [47, 80, 100], "geometri": 47, "occlud": [47, 69, 92, 116], "strong": 47, "relianc": 47, "heavili": 47, "evalu": [47, 48, 61, 64, 69, 72, 75, 76, 81, 92, 106, 109], "poor": [47, 95, 107], "few": [47, 48, 73, 88, 90, 92, 98, 101, 102, 104, 113, 119], "small": [47, 49, 73, 79, 92, 95], "region": [47, 52, 88, 92], "significantli": [47, 78], "benefit": 47, "spread": [47, 53, 55, 58, 69, 70], "gaussian": [47, 53, 55, 58, 68, 76], "difficult": [47, 48, 95], "densiti": [47, 76], "peak": [47, 53, 54, 58, 59, 63, 71, 72, 75, 76, 83, 89, 105], "unit": [47, 48, 52, 53, 55, 58, 70, 75], "5x": 47, "speed": [47, 48, 49, 72, 89, 93, 105, 106, 115, 116], "weigh": [47, 58, 69, 78], "focu": [47, 48, 73], "subpixel": [47, 76], "local": [47, 48, 58, 59, 63, 69, 71, 72, 76, 91, 92, 96, 103, 104, 105, 106, 107, 108, 110, 112, 117], "hoc": 47, "determinist": 47, "resourc": [47, 48, 73], "centroidsheadconfig": [47, 70], "classmapsheadconfig": [47, 70], "classifi": [47, 71, 72, 75, 107], "semant": 47, "segment": [47, 75], "howev": [47, 92, 100, 106, 113], "neighborhood": [47, 58, 63, 76], "landmark": [47, 75, 76, 84, 92, 97, 104, 107], "nonzero": 47, "classvectorsheadconfig": [47, 70], "num_fc_lay": [47, 70], "num_fc_unit": [47, 70], "global_pool": [47, 70], "classif": [47, 70, 71, 72], "task": [47, 56, 72, 105, 106, 109], "headsconfig": 47, "single_inst": [47, 91, 109, 110], "singleinstanceconfmapsheadconfig": [47, 70], "centered_inst": [47, 89, 91, 104, 105, 106, 107, 109, 110], "multi_inst": [47, 75, 91, 104, 109, 110], "multiinstanceconfig": [47, 75], "multi_class_bottomup": [47, 109], "multiclassbottomupconfig": 47, "multi_class_topdown": [47, 105, 109], "multiclasstopdownconfig": 47, "max_strid": [47, 66, 109], "retain": [47, 78], "filters_increas": 47, "constant": 47, "modelconfig": [47, 50, 74], "base_checkpoint": [47, 89, 109], "confmap": [47, 53, 58, 63, 72, 83], "multiinstanceconfmapsheadconfig": [47, 70], "class_map": [47, 58, 71, 72], "fashion": 47, "beyond": [47, 48, 72, 80, 92], "class_vector": [47, 72], "partaffinityfieldsheadconfig": [47, 70], "alon": 47, "contrast": [47, 48, 86, 88, 109], "benefici": [47, 48, 88], "graph": [47, 55, 70, 72, 73, 75, 84, 89, 92, 97, 105, 117], "averag": [47, 48, 69, 75, 76, 88, 92, 105, 106, 107], "underneath": 47, "candid": [47, 69, 72, 75, 78, 80, 89, 95, 116], "coher": 47, "source_nod": [47, 70], "destination_nod": [47, 70], "disambigu": 47, "nearbi": [47, 79, 92], "decoder_filters_r": 47, "decoder_batchnorm": 47, "tunabl": 47, "global": [47, 48, 59, 72, 76], "block_strid": 47, "conv_batchnorm": 47, "augmentationconfig": [48, 52], "rotation_min_angl": [48, 109], "180": [48, 88], "rotation_max_angl": [48, 109], "translat": [48, 109], "translate_min": [48, 109], "translate_max": [48, 109], "scale_min": [48, 109], "9": [48, 69, 86, 103, 104, 105, 106, 109, 119], "scale_max": [48, 109], "uniform_nois": [48, 109], "uniform_noise_min_v": [48, 109], "uniform_noise_max_v": [48, 109], "gaussian_nois": [48, 109], "gaussian_noise_mean": [48, 109], "gaussian_noise_stddev": [48, 109], "contrast_min_gamma": [48, 109], "contrast_max_gamma": [48, 109], "bright": [48, 88, 109], "brightness_min_v": [48, 109], "brightness_max_v": [48, 109], "random_crop": [48, 109], "random_crop_height": [48, 109], "random_crop_width": [48, 109], "random_flip": [48, 109], "flip_horizont": [48, 109], "augment": [48, 88, 91, 96], "imgaug": 48, "geometr": 48, "degre": 48, "independ": [48, 61, 103], "uniformli": [48, 54], "nois": 48, "simul": [48, 72, 105], "shot": 48, "arithmet": 48, "addelementwis": 48, "uniform": [48, 66], "tigher": 48, "magnitud": 48, "additivegaussiannois": 48, "gamma": 48, "constrast": 48, "gammacontrast": 48, "broadfield": 48, "illumin": 48, "effici": [48, 53, 77], "reflect": [48, 52, 69], "correctli": [48, 49, 102, 105, 106, 112, 117], "earlystoppingconfig": [48, 106, 109], "stop_training_on_plateau": [48, 106, 109], "plateau_min_delta": [48, 106, 109], "1e": [48, 106, 109], "06": [48, 69, 106, 109], "plateau_pati": [48, 106, 109], "termin": [48, 102, 118], "plateau": 48, "minim": [48, 49, 54, 95, 102], "gain": 48, "overfit": [48, 88], "regim": 48, "least": [48, 67, 96, 104, 105, 106, 107, 108, 109], "hardkeypointminingconfig": [48, 73], "online_min": [48, 109], "hard_to_easy_ratio": [48, 73, 109], "min_hard_keypoint": [48, 73, 109], "max_hard_keypoint": [48, 73, 109], "loss_scal": [48, 73, 109], "hard": [48, 69, 73, 88, 92, 95], "mine": [48, 73, 88], "ohkm": [48, 73], "tricki": [48, 96], "switch": [48, 73, 92, 95], "skew": [48, 73], "impact": [48, 49, 72, 73, 100], "altogeth": [48, 73, 104, 108], "learningratescheduleconfig": [48, 106, 109], "reduce_on_plateau": [48, 106, 109], "reduction_factor": [48, 106, 109], "plateau_cooldown": [48, 106, 109], "min_learning_r": [48, 106, 109], "08": [48, 69, 104, 106, 109], "later": [48, 92, 95, 110], "finer": [48, 75, 80], "balanc": 48, "practic": 48, "optimizationconfig": [48, 50, 64, 81], "preload_data": [48, 109], "augmentation_config": [48, 106, 109], "online_shuffl": [48, 109], "shuffle_buffer_s": [48, 109], "batches_per_epoch": [48, 109], "min_batches_per_epoch": [48, 109], "val_batches_per_epoch": [48, 109], "min_val_batches_per_epoch": [48, 109], "100": [48, 80, 86, 92, 103, 104, 105, 106, 108, 109], "adam": [48, 109], "initial_learning_r": [48, 109], "0001": [48, 109], "learning_rate_schedul": [48, 109], "hard_keypoint_min": [48, 109], "early_stop": [48, 109, 110], "preload": [48, 54], "shuffl": [48, 54], "buffer": [48, 54], "vari": [48, 88, 105], "composit": [48, 52, 85], "turn": [48, 88, 92, 95, 104, 117], "promot": [48, 54], "minibatch": [48, 88], "gradient": [48, 72, 76, 88], "gpu": [48, 54, 72, 77, 88, 89, 90, 98, 101, 104, 105, 106, 107, 108, 109, 112, 117], "lead": [48, 54, 77, 88, 103, 104], "whichev": [48, 94], "sever": [48, 89, 107, 112, 113], "execut": [48, 73, 75, 81], "longer": [48, 61, 72, 75, 89, 90, 92, 94, 108], "heurist": [48, 89, 95, 115], "400": [48, 105], "short": [48, 49, 75, 85], "amsgrad": 48, "minima": 48, "descend": [48, 69], "converg": 48, "checkpointingconfig": [49, 81], "initial_model": [49, 109], "best_model": [49, 72, 89, 96, 104, 106, 107, 109], "every_epoch": [49, 109], "latest_model": [49, 109], "final_model": [49, 96, 109], "comparison": [49, 82, 107, 112], "baselin": [49, 91, 109], "run_fold": 49, "04d": 49, "latest": [49, 101, 102, 109, 110], "outputsconfig": [49, 50, 81], "save_output": [49, 109], "run_nam": [49, 89, 96, 106, 109], "run_name_prefix": [49, 109], "run_name_suffix": [49, 109], "runs_fold": [49, 109], "save_visu": [49, 109], "keep_viz_imag": [49, 109], "zip_output": [49, 109], "log_to_csv": [49, 109], "tensorboardconfig": [49, 81], "zmqconfig": [49, 81], "job": [49, 50, 72, 81, 89, 90, 91, 96, 105, 107, 109, 113], "timestamp": 49, "rather": [49, 53, 54, 58, 60, 66, 90, 96, 103], "exact": [49, 68, 100, 107], "_1": 49, "counter": 49, "delai": 49, "client": 49, "subdirectori": [49, 81, 85, 96], "organ": 49, "uncheck": [49, 95], "training_log": [49, 96, 107, 109], "zeromq": [49, 81], "monitor": [49, 72, 73, 86, 98, 115], "run_path": [49, 81], "write_log": [49, 109], "loss_frequ": [49, 109], "architecture_graph": [49, 109], "profile_graph": [49, 109], "parent": [49, 97], "displai": [49, 72, 89, 103, 108, 115], "dashboard": 49, "often": [49, 69, 73, 76, 95, 96, 100], "frequenc": [49, 104, 107], "affect": [49, 92, 100], "graphic": [49, 115], "consum": 49, "lot": [49, 92], "freez": [49, 72], "browser": [49, 109], "tab": [49, 88, 91, 98, 109, 119], "profil": [49, 50, 89, 93, 96, 109], "subscribe_to_control": [49, 109], "controller_address": [49, 109], "controller_polling_timeout": [49, 109], "publish_upd": [49, 109], "publish_address": [49, 109], "listen": 49, "broadcast": [49, 55, 68], "messag": [49, 72, 89, 100, 115], "endpoint": 49, "protocol": 49, "asynchron": 49, "program": [49, 87], "poll": 49, "ip": 49, "hostnam": 49, "port": [49, 104], "ip_address": 49, "port_numb": 49, "microsecond": 49, "poller": 49, "wait": [49, 80, 102], "serializ": 50, "parametr": 50, "easili": [50, 89], "decoupl": 50, "design": [50, 93], "logic": [50, 71, 72, 77, 78, 105], "aggreg": 50, "oppos": 50, "trainingjobconfig": [50, 69, 72, 81, 104, 106], "sleap_vers": [50, 109], "from_json": [50, 84], "json_data": 50, "from_json_dict": 50, "json_data_dict": 50, "load_json": [50, 84], "load_training_config": 50, "initial_config": [50, 81, 96, 106, 107, 109], "save_json": [50, 84], "load_config": [50, 106, 107], "oneof": 51, "attrs_cl": 51, "must_be_set": 51, "decor": 51, "inspir": 51, "protobuff": 51, "__init__": [51, 105], "albumentationsaugment": 52, "image_kei": [52, 61, 62, 66], "instances_kei": [52, 60, 61], "albument": 52, "core": [52, 72, 104, 105, 106, 107, 108, 109, 112], "input_kei": [52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 66, 67, 81], "output_kei": [52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 81], "outgo": [52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 66, 67], "transform_dataset": [52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 66, 67], "input_d": [52, 53, 55, 56, 58, 59, 61, 66], "datasetv2": [52, 53, 54, 55, 56, 58, 59, 60, 61, 62, 64, 65, 66, 67, 72], "randomcropp": 52, "crop_height": [52, 61, 76], "crop_width": [52, 61, 76], "crop_bbox": [52, 61, 76], "randomflipp": 52, "symmetric_ind": [52, 84], "n_symmetri": 52, "mirror": 52, "v": [52, 64, 69, 100, 107], "matter": [52, 96, 97], "from_skeleton": 52, "flip_instances_lr": 52, "img_width": 52, "n_instanc": [52, 53, 55, 58, 60, 61, 65, 72, 75], "int32": [52, 55, 58, 60, 61, 65, 71, 75, 76, 82, 105], "flip_instances_ud": 52, "img_height": 52, "instanceconfidencemapgener": 53, "with_offset": 53, "offsets_threshold": 53, "flatten_offset": 53, "grid": [53, 55, 58, 63, 65, 68, 72, 76, 105], "instance_imag": [53, 59, 61], "center_inst": [53, 61], "instance_confidence_map": [53, 59], "all_instance_confidence_map": 53, "importantli": [53, 55], "proport": [53, 55], "multiconfidencemapgener": 53, "centroid_confidence_map": [53, 59], "singleinstanceconfidencemapgener": 53, "make_confmap": [53, 58, 63], "xv": [53, 55, 63, 68, 76], "yv": [53, 55, 63, 68, 76], "grid_width": [53, 55, 58, 63, 68], "make_grid_vector": [53, 55, 63, 68], "grid_height": [53, 55, 58, 63, 68], "unnorm": [53, 68, 76], "diagon": 53, "covari": [53, 76], "make_multi_confmap": 53, "elementwis": 53, "immedi": [53, 95], "make_multi_confmaps_with_offset": 53, "divid": [53, 62, 63, 72, 75, 76, 95], "batcher": 54, "drop_remaind": 54, "unrag": 54, "variabl": [54, 66, 68, 72, 73, 84, 102, 104], "rag": [54, 68, 72, 89], "remaind": 54, "raggedtensor": [54, 68, 72, 75], "leav": [54, 68], "downstream": [54, 65], "continu": [54, 86, 105, 115, 116, 117, 118], "ds_input": [54, 60, 61, 62, 66, 67], "had": [54, 80, 90, 102, 104, 105, 106, 107, 108, 109], "lambdafilt": 54, "filter_fn": 54, "elimin": 54, "kept": [54, 61], "buffer_s": 54, "unmodifi": 54, "experiment": [54, 95], "autotun": 54, "latenc": [54, 101, 105], "consumpt": 54, "infinit": 54, "indefinit": 54, "boundari": [54, 70], "shuffler": 54, "reshuffle_each_iter": 54, "though": 54, "latter": [54, 72, 89, 99], "technic": [54, 61, 76, 101], "alwai": [54, 65, 72, 84, 95, 102, 116], "fill": [54, 72, 105, 116], "realli": [54, 85], "unbatch": [54, 61], "partaffinityfieldsgener": 55, "flatten_channel": 55, "n_edg": [55, 75], "ax": [55, 72, 103], "skeleton_ind": [55, 60, 65], "part_affinity_field": [55, 72], "distance_to_edg": 55, "edge_sourc": 55, "edge_destin": 55, "undirect": 55, "d_0": 55, "d_n": 55, "get_edge_point": 55, "destin": [55, 70, 75, 84, 95, 97], "make_edge_map": 55, "make_multi_paf": 55, "make_paf": 55, "keydevicemov": 56, "device_nam": [56, 59], "cpu": [56, 72, 77, 85, 89, 102, 104, 105, 106, 107, 109], "devic": [56, 77, 102, 104, 105, 106, 107, 108, 109], "keyfilt": 56, "keep_kei": 56, "keyrenam": 56, "old_key_nam": 56, "new_key_nam": 56, "drop_old": 56, "renam": [56, 84, 118], "lambdamap": 56, "func": 56, "input_key_nam": 56, "output_key_nam": 56, "group_exampl": 57, "frame_ind": [57, 65], "group_examples_it": 57, "vid": 57, "classmapgener": 58, "class_map_threshold": 58, "track_ind": [58, 61, 65], "classvectorgener": 58, "make_class_map": 58, "class_ind": [58, 71], "n_class": [58, 71, 72], "confidence_map": [58, 63], "binari": [58, 104, 105, 106, 107, 109], "make_class_vector": 58, "interpret": [58, 75, 112], "globalpeakfind": 59, "confmaps_kei": 59, "predicted_instance_confidence_map": 59, "confmaps_strid": 59, "peak_threshold": [59, 72, 89, 109], "peaks_kei": 59, "predicted_center_instance_point": 59, "peak_vals_kei": 59, "predicted_center_instance_confid": 59, "keep_confmap": 59, "integral_patch_s": [59, 72, 76], "kerasmodelpredictor": 59, "keras_model": [59, 64, 72, 74, 81, 82, 106], "model_input_kei": 59, "model_output_kei": 59, "localpeakfind": 59, "predicted_centroid": 59, "predicted_centroid_confid": 59, "peak_sample_inds_kei": 59, "predicted_centroid_sample_ind": 59, "peak_channel_inds_kei": 59, "predicted_centroid_channel_ind": 59, "mockglobalpeakfind": 59, "all_peaks_in_kei": 59, "peaks_out_kei": 59, "confmaps_in_kei": 59, "confmaps_out_kei": 59, "mimic": 59, "predictedcenterinstancenorm": 59, "centroid_kei": 59, "centroid_confidence_kei": 59, "centroid_confid": [59, 61], "peak_confidences_kei": 59, "new_centroid_kei": 59, "new_centroid_confidence_kei": 59, "new_peaks_kei": 59, "new_peak_confidences_kei": 59, "predicted_instance_confid": 59, "instancecentroidfind": [60, 72], "center_on_anchor_part": 60, "anchor_part_nam": 60, "stream": [60, 62, 66, 103, 104, 109], "find_points_bbox_midpoint": 60, "n_point": [60, 63, 75], "xy_mid": 60, "xy_min": 60, "xy_max": 60, "get_instance_anchor": 60, "anchor_ind": 60, "gather": 60, "instancecropp": 61, "keep_full_imag": [61, 66], "mock_centroid_confid": 61, "centroids_kei": 61, "knowledg": 61, "bbox": [61, 76, 83], "center_instance_ind": 61, "th": [61, 76, 89], "full_image_height": 61, "full_image_width": 61, "crop_and_res": [61, 76], "becom": 61, "n_bbox": [61, 76], "make_centered_bbox": [61, 76], "span": [61, 68, 71, 76], "patch": [61, 72, 76], "find_instance_crop_s": 61, "min_crop_s": 61, "margin": 61, "truncat": 61, "side": [61, 88, 92, 109, 118, 119], "box_height": 61, "box_width": 61, "n_centroid": [61, 72], "onto": [61, 67, 105], "diagram": 61, "normalize_bbox": 61, "image_height": [61, 66, 68], "image_width": [61, 66, 68], "unnormalize_bbox": 61, "normalized_bbox": 61, "ensure_float": [62, 72], "constraint": [62, 66], "convert_rgb_to_bgr": 62, "convert_image_dtyp": 62, "rgb_to_grayscal": 62, "ensure_int": 62, "uint8": [62, 72, 104, 105], "ensure_min_image_rank": 62, "expand_to_rank": [62, 68], "grayscale_to_rgb": 62, "scale_image_rang": 62, "min_val": 62, "max_val": 62, "scale_to_imagenet_caffe_mod": 62, "approxim": [62, 88, 113], "scale_to_imagenet_tf_mod": 62, "scale_to_imagenet_torch_mod": 62, "485": 62, "456": [62, 96, 105], "406": [62, 104, 105], "229": [62, 105], "225": [62, 105], "regress": [63, 72, 76], "make_offset": 63, "mask_offset": 63, "capabl": [64, 105, 106, 109], "bottomupmulticlasspipelin": 64, "data_config": 64, "optimization_config": 64, "confmaps_head": 64, "multiinstanceconfmapshead": [64, 70, 72], "class_maps_head": 64, "classmapshead": [64, 70, 72], "offsets_head": 64, "offsetrefinementhead": [64, 70, 72], "builder": [64, 81, 106, 109], "make_base_pipelin": 64, "data_provid": [64, 72], "make_training_pipelin": 64, "remap": 64, "keymapp": [64, 67], "make_viz_pipelin": 64, "fetch": 64, "bottomuppipelin": 64, "pafs_head": 64, "partaffinityfieldshead": [64, 70, 72], "centroidconfmapspipelin": 64, "centroid_confmap_head": 64, "centroidconfmapshead": [64, 70, 72, 109], "return_descript": [64, 68], "from_block": 64, "chain": [64, 72], "from_pipelin": 64, "make_dataset": [64, 65], "peek": [64, 103], "quickli": [64, 92, 95, 115], "validate_pipelin": 64, "singleinstanceconfmapspipelin": 64, "single_instance_confmap_head": 64, "singleinstanceconfmapshead": [64, 70, 72], "topdownmulticlasspipelin": 64, "instance_confmap_head": 64, "centeredinstanceconfmapshead": [64, 70, 72, 106, 109], "class_vectors_head": 64, "classvectorshead": [64, 70, 72], "topdownconfmapspipelin": 64, "example_indic": [65, 67], "handi": [65, 92], "from_unlabeled_suggest": 65, "from_user_inst": 65, "from_user_labeled_fram": 65, "is_from_multi_size_video": 65, "ds_index": 65, "raw_image_s": 65, "example_ind": 65, "int64": [65, 85], "x_scale": 65, "y_scale": 65, "properli": [65, 104, 107, 108], "max_height_and_width": [65, 66], "from_filepath": 65, "pointsrescal": 66, "points_kei": 66, "scale_kei": 66, "invert": [66, 84], "full_image_kei": 66, "full_imag": 66, "sizematch": 66, "max_image_height": 66, "max_image_width": 66, "center_pad": 66, "offset_x": 66, "offset_i": 66, "update_config": [66, 74, 81], "find_padding_for_strid": 66, "pad_bottom": 66, "pad_right": 66, "mismatch": 66, "n_down_block": 66, "rescal": 66, "scale_x": 66, "scale_i": 66, "key_map": 67, "split_label": 67, "split_fract": 67, "One": [67, 72, 78, 84, 95], "round": [67, 104, 106], "split_labels_read": 67, "labels_read": 67, "split_labels_train_v": 67, "idx_train": 67, "idx_val": 67, "Their": 67, "miscellan": [68, 85], "describe_tensor": [68, 72], "ensure_list": 68, "target_rank": 68, "gaussian_pdf": 68, "meshgrid": 68, "thought": 68, "induc": 68, "unrag_exampl": 68, "sync_to_numpy_or_python_typ": 68, "unrag_tensor": 68, "max_siz": 68, "measur": [69, 75, 95], "understand": [69, 92, 100], "oks_voc": [69, 107], "vi": [69, 107], "recal": [69, 92, 107, 117], "dist": [69, 107], "p95": [69, 107], "compute_dist_metr": 69, "dists_dict": 69, "euclidean": [69, 75, 82], "percentil": [69, 107], "n_positive_pair": 69, "compute_dist": 69, "positive_pair": 69, "instance_gt": 69, "instance_pr": 69, "compute_generalized_voc_metr": 69, "false_neg": 69, "match_scor": [69, 107], "match_score_threshold": [69, 107], "55": [69, 109], "65": 69, "75": 69, "85": 69, "95": [69, 78, 80, 104, 105, 107], "recall_threshold": [69, 107], "01": [69, 103, 104, 105, 106, 107, 108, 109], "02": [69, 78, 105, 106], "04": [69, 103, 104, 105, 106, 107, 109], "05": [69, 103, 105, 109], "07": [69, 107, 109], "09": [69, 105, 106, 107, 108, 109], "11": [69, 86, 102, 103, 104, 105, 106, 107, 108, 109], "12": [69, 90, 102, 103, 104, 105, 106, 107, 108, 109], "13": [69, 86, 103, 105, 106, 107, 109], "14": [69, 86, 104, 105, 106, 107, 108, 109], "17": [69, 72, 106, 108, 109], "19": [69, 100, 101, 109], "22": [69, 109], "24": [69, 90, 103, 106, 109], "25": [69, 72, 75, 76, 86, 89, 103, 105, 108, 109], "26": [69, 92, 95, 104, 109], "27": [69, 104, 107, 109], "28": [69, 107, 109], "29": [69, 104, 107, 108, 109], "31": [69, 104, 105, 106, 109], "33": [69, 109], "35": [69, 86, 105, 109], "36": [69, 109], "37": [69, 105, 109], "38": [69, 104, 109], "39": [69, 104, 105, 106, 109], "41": [69, 109], "42": [69, 105, 106, 107, 109], "43": [69, 105, 106, 107, 109], "44": [69, 105, 106, 107, 109], "45": [69, 86, 104, 109], "46": [69, 104, 109], "47": [69, 104], "48": [69, 104, 107], "49": [69, 104, 105, 107, 109], "51": 69, "52": [69, 104, 105], "53": [69, 104, 105], "54": [69, 104, 109], "56": [69, 104, 105], "57": [69, 104, 105, 109], "58": [69, 104, 106], "59": [69, 86, 104], "61": [69, 109], "62": 69, "63": [69, 105, 109], "66": 69, "67": [69, 109], "69": [69, 104], "71": [69, 104], "72": [69, 83, 104, 109], "73": [69, 109], "74": [69, 105, 109], "76": 69, "77": [69, 104], "78": [69, 104, 105, 106, 107, 108, 109], "79": [69, 109], "81": [69, 104, 105], "82": [69, 103, 109], "83": [69, 104, 105], "84": [69, 104, 105], "86": [69, 105], "87": [69, 105], "88": [69, 104, 105, 109], "89": [69, 104, 105], "91": [69, 104, 105], "92": [69, 104, 105], "93": 69, "94": [69, 104, 105], "96": [69, 105], "97": [69, 86, 105], "98": [69, 104], "99": [69, 104, 105], "gvoc": 69, "voc": [69, 107], "unmatch": [69, 82], "ok": [69, 103, 104, 106, 107, 109], "compute_instance_area": 69, "compute_ok": 69, "points_gt": 69, "points_pr": 69, "stddev": 69, "025": 69, "use_cocoev": 69, "n_gt": 69, "n_ed": 69, "n_pr": 69, "cocoev": 69, "cocoapi": 69, "difficulti": [69, 107], "072": 69, "easiest": [69, 90, 96, 114], "ey": 69, "hardest": 69, "hip": 69, "107": 69, "bone": 69, "anatom": 69, "cloth": 69, "ronch": 69, "perona": 69, "diagnosi": 69, "iccv": 69, "2017": [69, 75, 117], "compute_pck_metr": 69, "pck": [69, 107], "compute_visibility_conf": 69, "confus": 69, "oks_stddev": 69, "oks_scal": 69, "match_threshold": 69, "user_labels_onli": 69, "evaluate_model": 69, "cfg": [69, 75, 106], "split_nam": 69, "find_frame_pair": 69, "frame_gt": 69, "frame_pr": 69, "load_metr": [69, 107], "model_path": [69, 72, 107, 109], "npz": [69, 106, 107, 109], "tp": [69, 107], "tn": [69, 107], "avg": [69, 107], "p50": [69, 107], "50th": [69, 107], "p75": [69, 107], "75th": [69, 107], "p90": [69, 107], "90th": [69, 107], "95th": [69, 107], "p99": [69, 107], "99th": [69, 107], "mpck": [69, 107], "percentag": [69, 89, 107], "mok": [69, 107], "mar": [69, 107], "pck_voc": [69, 107], "match_frame_pair": 69, "frame_pair": 69, "match_inst": 69, "pascal": 69, "briefli": 69, "greedili": 69, "replace_path": 69, "video_list": 69, "loss_funct": 70, "make_head": 70, "sigma_threshold": 70, "confmapconfig": 70, "classify_peaks_from_map": 71, "peak_point": [71, 76], "peak_val": [71, 72, 75, 76], "peak_sample_ind": [71, 76], "peak_channel_ind": [71, 72, 75, 76], "n_channel": 71, "n_sampl": [71, 75], "n_peak": [71, 75, 76], "point_val": 71, "class_prob": 71, "group_class_peak": 71, "classify_peaks_from_vector": 71, "peak_class_prob": 71, "crop_sample_ind": 71, "peak_ind": [71, 75], "Will": 71, "hungarian": [71, 78, 82, 89, 95], "algorithm": [71, 75, 76, 78, 79, 82, 92, 109, 113, 115], "inferencemodel": [72, 89], "surround": 72, "routin": 72, "predictor": [72, 104, 105, 107, 109], "postprocess": 72, "load_model": [72, 104, 105, 107], "bottomupinferencelay": 72, "encapsul": [72, 78], "forward": [72, 92, 103], "paf_scor": 72, "paf_group": 72, "pafscor": [72, 75], "cm_output_strid": 72, "paf_output_strid": 72, "quarter": [72, 76], "rough": [72, 76], "return_confmap": 72, "slower": [72, 78, 88, 95, 105], "return_paf": 72, "return_paf_graph": 72, "confmaps_ind": 72, "pafs_ind": 72, "offsets_ind": 72, "instance_peak": [72, 105], "instance_peak_v": [72, 105], "edge_peak_ind": [72, 75], "line_scor": [72, 75], "find_peak": 72, "cm": [72, 76, 86], "forward_pass": 72, "bottomupinferencemodel": 72, "bottomup_lay": 72, "bottomupmulticlassinferencelay": 72, "class_maps_output_strid": 72, "return_class_map": 72, "class_maps_ind": 72, "inference_lay": 72, "bottomupmulticlassinferencemodel": 72, "bottomupmulticlasspredictor": 72, "inference_model": [72, 105], "integral_refin": 72, "verbos": [72, 89, 109], "rich": [72, 89, 109], "report_r": 72, "from_trained_model": 72, "training_job": [72, 81, 110], "resize_input_lay": 72, "insid": [72, 85, 87, 88, 89, 92, 96, 114], "training_config": [72, 89, 96, 104, 106, 107, 109], "num_color_channel": 72, "bottomuppredictor": 72, "is_grayscal": 72, "bottomup_config": 72, "bottomup_model": 72, "max_edge_length_ratio": [72, 75, 89, 109], "dist_penalty_weight": [72, 75, 89, 109], "paf_line_point": 72, "min_line_scor": [72, 75], "max_inst": [72, 89, 109], "bottomup": [72, 96, 110], "tracker": [72, 80, 89, 92, 95, 96, 108, 109, 110, 112, 119], "penal": [72, 75, 89], "coeffici": [72, 75, 89], "penalti": [72, 75, 89], "enforc": [72, 75, 89], "strictli": [72, 75, 89, 102], "spuriou": [72, 75], "discard": [72, 75], "extran": 72, "centroidcrop": 72, "pade": 72, "return_crop": 72, "findinstancepeak": 72, "centroid_v": [72, 105], "centroid_confmap": 72, "output_height": 72, "output_width": 72, "crop_offset": 72, "former": 72, "centroidcropgroundtruth": 72, "squar": [72, 76, 83], "example_gt": 72, "centroidinferencemodel": 72, "centroid_crop": 72, "instance_confmap": 72, "findinstancepeaksgroundtruth": 72, "crop_output": 72, "n_instances_gt": 72, "greedi": [72, 75, 78, 79, 80, 89, 95, 108], "inferencelay": 72, "overload": 72, "previous": [72, 84], "export_model": [72, 89], "save_path": 72, "serving_default": 72, "save_trac": 72, "unrag_output": 72, "savedmodel": 72, "trace": 72, "frozen_graph": [72, 87], "pb": [72, 87], "n_valid": [72, 105], "predict_on_batch": [72, 105], "movenetinferencelay": 72, "movenet": 72, "lightn": 72, "thunder": 72, "movenetinferencemodel": 72, "joint": [72, 88, 103, 118], "reappli": 72, "meant": [72, 75], "overridden": 72, "__call__": 72, "masking_and_pad": 72, "movenetpredictor": 72, "from_model_path": 72, "singleinstancepredictor": 72, "topdownpredictor": [72, 109], "topdownmulticlasspredictor": 72, "make_pipelin": 72, "make_label": 72, "report_period": 72, "ratecolumn": 72, "table_column": 72, "singleinstanceinferencelay": 72, "singleinstanceinferencemodel": 72, "detector": 72, "single_instance_lay": 72, "confmap_config": [72, 87], "confmap_model": 72, "topdowninferencemodel": 72, "cropper": 72, "topdownmulticlassfindpeak": 72, "return_class_vector": 72, "associt": 72, "class_vectors_ind": 72, "optimal_group": 72, "topdownmulticlassinferencemodel": 72, "centroid_config": [72, 87], "centroid_model": [72, 105], "centroid_model_path": 72, "confmap_model_path": 72, "visualpredictor": 72, "safely_gener": 72, "export_cli": 72, "exported_model": [72, 89], "find_head": 72, "get_keras_model_path": 72, "get_model_output_strid": 72, "input_ind": 72, "output_ind": 72, "evenli": [72, 92, 116], "tracker_window": 72, "tracker_max_inst": 72, "disable_gpu_prealloc": [72, 77], "progress_report": 72, "flow": [72, 80, 89, 95, 96, 108], "histori": 72, "prealloc": [72, 77], "neglig": 72, "bar": [72, 89, 116], "consol": [72, 89, 102], "notebook": [72, 89, 90, 93, 100, 101, 103, 104, 105, 106, 107, 108, 109, 110, 111, 113], "captur": [72, 105], "programmat": 72, "headless": 72, "machin": [72, 89, 91, 93, 96, 98, 100, 102, 103, 104, 107, 108, 110, 112, 113, 117], "make_model_movenet": 72, "ohkmloss": 73, "reweight": 73, "mse": 73, "worst": [73, 88], "y_gt": 73, "y_pr": 73, "sample_weight": 73, "y_true": 73, "d0": 73, "dn": 73, "spars": 73, "categor": 73, "crossentropi": 73, "y_pred": 73, "get_config": 73, "partloss": 73, "channel_ind": 73, "idempot": 73, "update_st": 73, "accumul": [73, 75], "textual": 73, "worri": [73, 90, 95], "eager": 73, "compute_ohkm_loss": 73, "queri": [74, 77, 78, 84, 95], "engin": 74, "built": [74, 96, 101, 113, 116], "make_model": 74, "resolv": [75, 94, 103, 104, 109], "agreement": 75, "connected": 75, "amen": 75, "zhe": 75, "cao": [75, 117], "toma": 75, "simon": 75, "shih": 75, "en": 75, "wei": 75, "yaser": 75, "sheikh": 75, "realtim": [75, 100, 101], "_cvpr_": 75, "edgeconnect": 75, "src_peak_ind": 75, "dst_peak_ind": 75, "edgetyp": [75, 84], "src_node_ind": [75, 84], "dst_node_ind": [75, 84], "pafs_strid": 75, "min_instance_peak": 75, "src_node": [75, 84], "dst_node": [75, 84], "subscript": [75, 84], "edge_typ": 75, "sorted_edge_ind": 75, "topolog": 75, "assembli": 75, "assign_connections_to_inst": 75, "finder": 75, "scorer": 75, "min_edge_scor": 75, "group_inst": 75, "match_edge_ind": 75, "match_src_peak_ind": 75, "match_dst_peak_ind": 75, "match_line_scor": 75, "n_connect": 75, "match_candid": 75, "whole": [75, 88, 92, 105, 110], "predicted_peak_scor": 75, "predicted_instance_scor": 75, "standalon": [75, 104], "group_instances_batch": 75, "n_candid": 75, "score_paf_lin": 75, "match_candidates_batch": 75, "score_paf_lines_batch": 75, "peakid": 75, "node_ind": 75, "partit": 75, "filter_connection_candid": 75, "node_type_ind": 75, "instance_assign": 75, "compute_distance_penalti": 75, "spatial_vec_length": 75, "max_edge_length": 75, "peaks_sampl": 75, "apart": 75, "infin": 75, "get_connection_candid": 75, "peak_channel_inds_sampl": 75, "skeleton_edg": 75, "get_paf_lin": 75, "pafs_sampl": 75, "n_line_point": 75, "pretti": [75, 105, 107], "gather_nd": 75, "col": [75, 82], "make_line_sub": 75, "group_instances_sampl": 75, "peak_scores_sampl": 75, "match_edge_inds_sampl": 75, "match_src_peak_inds_sampl": 75, "match_dst_peak_inds_sampl": 75, "match_line_scores_sampl": 75, "match_candidates_sampl": 75, "py_funct": 75, "make_predicted_inst": 75, "peak_scor": 75, "munkr": 75, "scipi": [75, 78, 103], "linear_sum_assign": [75, 82], "tf_linear_sum_assign": [75, 82], "edge_inds_sampl": 75, "edge_peak_inds_sampl": 75, "line_scores_sampl": 75, "paf_lines_sampl": 75, "conncet": 75, "dot": [75, 83], "product": [75, 109], "due": [75, 104], "toposort_edg": 75, "assembl": [75, 116], "root": [75, 107, 109, 110], "sometim": [76, 80, 95, 99, 108], "heatmap": 76, "entail": 76, "maxima": 76, "sample_ind": 76, "find_global_peak": 76, "xy": 76, "find_global_peaks_integr": 76, "find_global_peaks_rough": 76, "find_global_peaks_with_offset": 76, "find_local_peak": 76, "find_local_peaks_integr": 76, "find_local_peaks_rough": 76, "suppres": 76, "find_local_peaks_with_offset": 76, "find_offsets_local_direct": 76, "centered_patch": 76, "delta": 76, "dx_i": 76, "dy_i": 76, "reshap": [76, 103], "float64": [76, 85, 105], "alejandro": 76, "newel": 76, "kaiyu": 76, "yang": 76, "jia": 76, "deng": 76, "_european": 76, "confer": 76, "vision_": 76, "2016": 76, "integral_regress": 76, "x_hat": 76, "y_hat": 76, "make_gaussian_kernel": 76, "odd": 76, "pi": 76, "smooth_img": 76, "smooth": [76, 103], "convolv": 76, "best_logical_device_nam": 77, "placement": [77, 92, 97, 116], "disable_prealloc": [77, 104, 105], "growth": [77, 102, 105, 106, 109], "cuda": [77, 102, 104, 105, 106, 107, 108, 109], "startup": 77, "session": [77, 90, 105, 117], "nix": [77, 89], "enable_gpu_prealloc": 77, "enable_prealloc": 77, "get_all_gpu": 77, "physicaldevic": [77, 102], "unavail": 77, "get_available_gpu": 77, "get_current_gpu": 77, "get_gpu_memori": 77, "mib": [77, 109], "initialize_devic": 77, "is_gpu_system": 77, "discover": 77, "is_initi": 77, "hidden": 77, "logicaldevic": 77, "use_cpu_onli": [77, 104], "hide": [77, 92, 104], "use_first_gpu": 77, "use_gpu": 77, "device_ind": 77, "use_last_gpu": 77, "popul": 78, "pick": [78, 91, 92, 95, 106, 116], "clean": [78, 80, 89], "framematch": [78, 80], "cost_matrix": [78, 79, 82], "unmatched_inst": 78, "instancetyp": [78, 79, 80], "choic": [78, 79, 117], "got": [78, 117], "weren": 78, "from_candidate_inst": 78, "untracked_inst": [78, 80], "candidate_inst": 78, "similarity_funct": [78, 80], "matching_funct": [78, 80], "robust_best_inst": [78, 80], "from_cost_matrix": 78, "robust": [78, 80, 88, 89, 92, 109], "quantil": [78, 80, 89], "has_only_first_choice_match": 78, "is_first_choic": 78, "centroid_dist": 78, "ref_inst": [78, 80], "query_inst": [78, 80], "persist": 78, "connect_single_track_break": 78, "lost": [78, 88, 89, 95, 119], "cull_frame_inst": 78, "instances_list": 78, "iou_threshold": [78, 80], "intersect": [78, 80, 82, 95], "iou": [78, 80, 89, 95], "cull_inst": 78, "factory_object_keypoint_similar": 78, "keypoint_error": 78, "score_weight": 78, "normalization_keypoint": 78, "factori": 78, "first_choice_match": 78, "greedy_match": [78, 80], "bipartit": [78, 80], "hungarian_match": [78, 80], "instance_i": [78, 80], "instance_similar": [78, 80], "nms_fast": 78, "target_count": 78, "pyimagesearch": 78, "suppress": 78, "normalized_instance_similar": 78, "img_hw": [78, 80], "Then": [79, 91, 94, 97, 102, 105, 110, 120], "fragment": 79, "mess": 79, "get_track_instance_match": 79, "are_too_close_funct": 79, "ascend": 79, "whenev": 79, "wasn": [79, 107], "upshot": 79, "compet": 79, "looser": 79, "match_dict_from_match_funct": 79, "row_item": 79, "column_item": 79, "match_funct": 79, "key_by_column": 79, "remove_second_bests_from_cost_matrix": 79, "invalid_v": 79, "unclear": 79, "inf": [79, 82], "tool": [80, 93, 95, 100, 102, 115, 119], "link": [80, 104], "basetrack": 80, "flowcandidatemak": 80, "min_point": 80, "img_scal": [80, 89, 108, 109], "of_window_s": [80, 89, 108, 109], "of_max_level": [80, 89, 108, 109], "save_shifted_inst": [80, 89, 109], "track_window": [80, 89, 108, 109], "shifted_inst": 80, "shiftedinst": 80, "optic": [80, 89, 95, 96, 108], "shift": [80, 89, 92, 95], "movement": [80, 92, 113], "pyramid": [80, 89], "flow_shift_inst": 80, "ref_img": 80, "new_img": 80, "min_shifted_point": 80, "window_s": 80, "max_level": 80, "luca": 80, "kanad": 80, "get_shifted_inst": 80, "ref_t": 80, "get_shifted_instances_from_earlier_tim": 80, "prune_shifted_inst": 80, "prune": 80, "older": [80, 102], "flowmaxtrack": [80, 89], "max_track": [80, 89], "candidate_mak": 80, "target_instance_count": [80, 89, 108, 109], "pre_cull_funct": 80, "post_connect_single_break": [80, 89, 108, 109], "min_new_track_point": [80, 89, 108, 109], "track_matching_queu": 80, "dequ": 80, "matchedframeinst": 80, "track_matching_queue_dict": 80, "spawned_track": 80, "save_tracked_inst": 80, "last_match": 80, "flowmaxtrackscandidatemak": 80, "get_ref_inst": 80, "flowtrack": 80, "kalmantrack": 80, "init_track": 80, "init_set": 80, "kalmaninitset": 80, "kalman_track": 80, "barekalmantrack": 80, "cull_funct": 80, "init_frame_count": 80, "re_init_cooldown": 80, "re_init_aft": 80, "init_don": 80, "pre_track": 80, "last_t": 80, "last_init_t": 80, "kalman": [80, 89], "certain": [80, 84, 95, 98, 99], "init": [80, 102, 105, 106, 109], "enough": [80, 88, 93, 95, 99, 117], "gap": 80, "cooldown": 80, "period": 80, "is_valid": 80, "make_track": 80, "node_indic": 80, "instance_iou_threshold": 80, "simplecandidatemak": 80, "simplemaxtrack": [80, 89], "simplemaxtrackscandidatemak": 80, "simpletrack": 80, "trackclean": [80, 95], "veloc": [80, 92, 95], "get_candid": 80, "incorpor": [80, 107], "final_pass": 80, "timestep": 80, "unique_tracks_in_queu": 80, "run_track": 80, "bottomupmodeltrain": 81, "data_read": 81, "dataread": 81, "has_offset": 81, "bottomupmulticlassmodeltrain": 81, "centroidconfmapsmodeltrain": 81, "training_labels_read": 81, "validation_labels_read": 81, "test_labels_read": 81, "labels_config": 81, "video_search_path": 81, "incomplet": 81, "from_label": 81, "singleinstancemodeltrain": 81, "topdownmulticlassmodeltrain": 81, "topdownconfmapsmodeltrain": 81, "pipeline_build": 81, "pipelinebuild": 81, "training_pipelin": 81, "validation_pipelin": 81, "training_viz_pipelin": 81, "validation_viz_pipelin": 81, "optimization_callback": 81, "output_callback": 81, "visualization_callback": 81, "cleanup": 81, "teset": 81, "setup": [81, 86, 107], "create_trainer_using_cli": 81, "get_timestamp": 81, "date": [81, 96, 102, 109], "sanitize_scope_nam": 81, "sanit": 81, "scope": [81, 85], "setup_checkpoint": 81, "setup_loss": 81, "setup_metr": 81, "setup_new_run_fold": 81, "base_run_nam": 81, "setup_optimization_callback": 81, "setup_optim": 81, "optimizerv2": 81, "setup_output_callback": 81, "setup_tensorboard": 81, "setup_visu": 81, "viz_fn": 81, "setup_zmq_callback": 81, "zmq_config": 81, "compute_i": 82, "bbox1": 82, "bbox2": 82, "group_arrai": 82, "blog": [82, 100], "arang": [82, 103], "match_point": 82, "points1": 82, "points2": 82, "m": [82, 83, 85, 87, 89, 95, 96, 100, 101, 105, 108, 109, 110], "inds1": 82, "inds2": 82, "reset_input_lay": 82, "new_shap": 82, "stackoverflow": 82, "58485055": 82, "n_src": 82, "n_dst": 82, "generate_skeleton_preview_imag": 83, "square_bb": 83, "thumbnail_s": 83, "byte": [83, 84, 85], "preview": [83, 84, 88, 98], "thumbnail": 83, "imgfig": 83, "dpi": [83, 105, 107], "tight": 83, "inch": 83, "plot_confmap": 83, "cmap": [83, 103], "color_by_nod": 83, "lw": 83, "color_by_track": 83, "plot_paf": 83, "quiver": 83, "plot_peak": 83, "pts_gt": 83, "pts_pr": 83, "goal": [84, 115, 117], "symmetri": [84, 104], "arm": 84, "as_nod": 84, "from_nam": 84, "name_list": 84, "constitu": 84, "_skeleton_idx": 84, "preview_imag": 84, "_is_templ": 84, "add_edg": 84, "add_nod": 84, "add_symmetri": 84, "node1": 84, "node2": 84, "clear_edg": 84, "delete_edg": 84, "delete_nod": 84, "delete_symmetri": 84, "establish": 84, "edge_to_index": 84, "edges_ful": 84, "find_neighbor": 84, "predecessor": 84, "successor": [84, 101], "find_nod": 84, "find_unique_nod": 84, "from_dict": 84, "node_to_idx": 84, "sto": 84, "json_str": [84, 85], "idx_to_nod": 84, "from_names_and_edge_ind": 84, "get_symmetri": 84, "get_symmetry_nam": 84, "subgraph": 84, "graph_symmetri": 84, "has_edg": 84, "source_nam": 84, "dest_nam": 84, "ye": [84, 97, 100], "has_nod": 84, "is_arboresc": 84, "arboresc": 84, "is_templ": 84, "load_all_hdf5": 84, "return_dict": 84, "h5py": [84, 103, 114], "load_hdf5": 84, "load_mat": 84, "make_cattr": 84, "isomorph": 84, "node_to_index": 84, "relabel_nod": 84, "old_nam": 84, "new_nam": 84, "relabel": 84, "partial": [84, 86, 113], "rename_skeleton": 84, "immut": 84, "hash": 84, "new_skeleton": 84, "old_skeleton": 84, "save_all_hdf5": 84, "save_hdf5": 84, "symmetries_ful": 84, "symmetry_nam": 84, "obj": 84, "skeletondecod": 84, "jsonpickl": 84, "thorax1": 84, "b64": 84, "avzc": 84, "ivbo": 84, "reconstruct": [84, 100], "decode_preview_imag": 84, "img_b64": 84, "return_byt": 84, "pil": 84, "skeletonencod": 84, "neck": 84, "enum": 84, "attr_to_dtyp": 85, "cl": 85, "dict_cut": 85, "subdictionari": 85, "find_files_by_suffix": 85, "root_dir": 85, "depth": [85, 102], "direntri": 85, "subdir": 85, "frame_list": 85, "frame_str": 85, "get_config_fil": 85, "shortnam": 85, "ignore_file_not_found": 85, "get_default": 85, "get_package_fil": 85, "json_dump": 85, "json_load": 85, "make_scoped_dictionari": 85, "flat_dict": 85, "exclude_non": 85, "exclude_nod": 85, "parse_uri_path": 85, "uri": [85, 102], "posix": 85, "save_dict_to_hdf5": 85, "h5file": 85, "dic": 85, "recurs": 85, "uniquifi": 85, "seq": 85, "hashabl": 85, "usable_cpu_count": 85, "usabl": [85, 117], "cpuset": 85, "weak_filename_match": 85, "filename_a": 85, "filename_b": 85, "pkg": [86, 91, 96, 104, 105, 106, 107, 110], "2022": [86, 100, 101], "osf": 86, "quick": [86, 89, 92, 102, 104], "clip": [86, 89, 92, 99, 103, 107, 113, 115], "mice": 86, "home": [86, 92, 101, 102, 104, 107, 108, 109], "cage": 86, "fruit": 86, "fly": 86, "drosophila": [86, 103, 109, 111], "melanogast": [86, 103, 109, 111], "freeli": 86, "enclos": 86, "circular": 86, "arena": 86, "backlight": 86, "camera": [86, 88, 105], "192": [86, 109], "px": [86, 107], "mm": 86, "num": 86, "1500": 86, "download": [86, 87, 89, 90, 100, 102, 103, 104, 105, 106, 107, 108, 110, 111, 112], "credit": 86, "berman": 86, "2014": 86, "gordon": 86, "joshua": [86, 100, 101], "shaevitz": [86, 100, 101], "virgin": 86, "male": [86, 105, 116], "femal": [86, 105, 116], "fli": [86, 103], "nm91": 86, "strain": 86, "dai": 86, "eclos": 86, "frequent": 86, "engag": 86, "courtship": [86, 103, 109, 111], "1024": [86, 103, 104, 105, 107], "150": [86, 107], "4000": 86, "junyu": [86, 100, 101], "li": [86, 100, 101], "shruthi": [86, 100, 101], "ravindranath": [86, 100, 101], "mala": [86, 100, 101], "murthi": [86, 100, 101], "c57bl": 86, "6j": 86, "mu": 86, "musculu": 86, "acryl": 86, "floor": 86, "paw": 86, "1280": 86, "80": [86, 104], "1000": [86, 89, 105], "2950": 86, "uva": [86, 100, 101], "mikhail": [86, 100, 101], "kislin": [86, 100, 101], "samuel": [86, 100, 101], "wang": [86, 100, 101], "swiss": 86, "webster": 86, "musuculu": 86, "lightli": 86, "bed": 86, "1474": 86, "2948": 86, "eleni": [86, 100, 101], "papadoyanni": [86, 100, 101], "annegret": [86, 100, 101], "falkner": [86, 100, 101], "worker": 86, "bumblebe": 86, "bombu": 86, "impatien": 86, "petri": 86, "dish": 86, "hexagon": 86, "beeswax": 86, "minut": [86, 117], "1536": 86, "2048": 86, "804": 86, "1604": 86, "grace": [86, 100, 101], "mckenzi": [86, 100, 101], "smith": [86, 100, 101], "z": [86, 100, 101], "yan": [86, 100, 101], "sarah": [86, 100, 101], "kocher": [86, 100, 101], "pup": 86, "adult": 86, "merion": 86, "unguiculatu": 86, "p15": 86, "month": 86, "shave": 86, "425": [86, 104, 105], "1588": 86, "catalin": [86, 100, 101], "mitelut": [86, 100, 101], "marielisa": [86, 100, 101], "diez": [86, 100, 101], "castro": [86, 100, 101], "dan": [86, 100, 101], "sane": [86, 100, 101], "reactiv": 87, "top_down_id": 87, "workflow": [87, 93, 101, 105, 106, 112, 113, 116, 120], "filecaptur": 87, "toolbox": 87, "editor": [87, 91, 96, 106], "predictcentroid": 87, "predictpos": 87, "modelfilenam": 87, "trainingconfig": 87, "predictposeident": 87, "predictsinglepos": 87, "click": [87, 88, 91, 92, 94, 97, 98, 104, 105, 106, 108, 109, 110, 115, 116, 117, 118, 119], "analyz": [87, 93], "suppli": [88, 96], "wish": [88, 95], "hover": 88, "explan": [88, 114], "worth": 88, "On": [88, 92, 102, 118], "zoom": [88, 92, 98, 116], "pan": 88, "drag": [88, 92, 98, 116, 118], "press": [88, 102, 118], "alt": [88, 92, 102, 116, 118], "rule": [88, 97], "thumb": [88, 97], "accur": [88, 101, 113, 115, 117], "likewis": [88, 109], "unaug": 88, "choos": [88, 92, 94, 95, 96, 97, 114, 116, 117, 119], "nonn": 89, "exit": 89, "incorrectli": [89, 92], "colab": [89, 91, 93, 102, 103, 104, 105, 106, 108, 112, 117], "val_label": [89, 109], "save_viz": [89, 109], "keep_viz": 89, "training_job_path": [89, 109], "resum": 89, "export_path": 89, "preced": 89, "r": [89, 92, 103, 104, 105, 106, 109, 113, 114], "ommit": 89, "hpc": 89, "clean_instance_count": [89, 108, 109], "proofread": [89, 92, 93, 96, 101, 113, 116, 120], "pre_cull_to_target": [89, 108, 109], "pre_cull_iou_threshold": [89, 108, 109], "clean_iou_threshold": [89, 108, 109], "min_match_point": [89, 108, 109], "kf_node_indic": [89, 109], "kf_init_frame_count": [89, 109], "comma": 89, "hyphen": 89, "did": 89, "cap": 89, "normalized_inst": [89, 95], "object_keypoint": [89, 95], "my_model": 89, "output_predict": 89, "input_video": 89, "labels_with_predict": 89, "v005": 89, "retrack": [89, 105, 108], "input_predict": 89, "amongst": [89, 96], "input_path": 89, "convent": 89, "preceed": 89, "analysis_video_0": 89, "001_second": 89, "session1": 89, "f": [89, 95, 102, 103, 104, 105, 107, 114], "video_index": 89, "script": [89, 91, 105, 106, 112], "widget": 89, "walk": [90, 112, 120], "colaboratori": [90, 112], "host": [90, 109], "servic": 90, "server": [90, 102], "think": [90, 98, 100], "hour": 90, "sooner": 90, "idl": 90, "demand": 90, "pai": 90, "pro": 90, "hardwar": 90, "sai": 90, "lenient": 90, "drive": [90, 92, 96, 103, 106, 108], "univers": [90, 101], "suit": 90, "educ": 90, "unlimit": 90, "dropbox": [90, 106], "microsoft": 90, "onedr": 90, "appl": [90, 101, 102], "icloud": 90, "elsewher": [90, 92], "perhap": 90, "plan": 90, "tri": [90, 95, 100], "explain": [91, 95, 96], "decid": [91, 92, 94], "hyperpamet": 91, "But": [91, 95, 96, 104, 105, 106, 108], "sound": 91, "intimid": 91, "menu": [91, 94, 95, 98, 100, 102, 109, 110, 114, 116, 117, 118, 119], "advic": 91, "wherev": 91, "sh": [91, 102], "unzip": [91, 96, 107, 109, 110], "sleap_label": 92, "keyboard": [92, 94, 95], "suppos": [92, 94, 95, 96, 116], "realiz": 92, "join": [92, 95, 107], "clipboard": 92, "toggl": [92, 94, 116], "shown": [92, 94, 95, 116], "yellow": [92, 115, 119], "cycl": [92, 95, 117], "sixth": 92, "orang": [92, 95, 116], "thin": [92, 95, 115], "trail": [92, 95, 103], "250": [92, 104], "500": 92, "signific": 92, "lag": 92, "header": 92, "dock": 92, "outlin": 92, "highlight": 92, "propag": [92, 95], "entireti": 92, "deselect": 92, "withheld": 92, "why": 92, "preform": 92, "ever": 92, "rectangular": 92, "rectangl": 92, "tabl": [92, 109, 116], "wheel": [92, 116], "mac": [92, 102, 116, 118], "substitut": 92, "arrow": 92, "medium": 92, "legend": [92, 95, 103, 107], "escap": 92, "throughout": 92, "variat": 92, "jump": [92, 95], "comprehens": 93, "troubleshoot": 93, "answer": 93, "tip": [93, 100, 119], "happi": [93, 95], "scratch": [94, 106, 109, 115], "uncorrect": 94, "guess": 94, "tutori": [94, 95, 100, 101, 108, 113, 117], "familiar": 94, "neither": 94, "haven": [95, 98, 102, 118], "obvious": 95, "big": 95, "chosen": [95, 110], "xiao": 95, "exp": 95, "maxim": [95, 100, 117], "navig": 95, "cross": [95, 96], "checkbox": 95, "caus": 95, "mistaken": [95, 119], "harder": 95, "techniqu": 95, "seekbar": [95, 115, 116, 117, 119], "mistak": [95, 119], "strategi": [95, 101], "mous": [95, 116], "misidentifi": [95, 119], "knew": 95, "wouldn": 95, "gotten": 95, "50": [95, 103, 104, 107], "hotkei": 95, "downsid": 95, "panel": [95, 109, 116, 118, 119], "appendag": 95, "fact": 95, "adjac": [95, 100], "submenu": [95, 119], "bundl": 96, "linux": [96, 101, 102, 104, 105, 106, 107, 108, 109, 118], "mount": [96, 106, 108], "training_profil": [96, 109], "baseline_medium_rf": [96, 109], "191205_162402": 96, "191205_163413": 96, "essenti": 97, "With": 97, "my": [97, 110], "tree": [97, 103, 107, 109], "children": 97, "littl": [97, 99], "complic": 97, "ran": 98, "amount": [98, 113], "tweak": [99, 108], "adopt": 99, "weird": 99, "squeez": 99, "email": 99, "bit": [99, 108, 117], "screenshot": 99, "stuck": 100, "crash": 100, "toolkit": [100, 102], "v11": [100, 102], "cudnn": [100, 102, 105, 106, 109], "v8": [100, 102], "neurosci": [100, 101], "overview": [100, 101, 104, 120], "hang": 100, "acquisit": [100, 105], "suitabl": 100, "vision": 100, "seekabl": 100, "consequ": [100, 113], "pix_fmt": 100, "yuv420p": 100, "nearli": 100, "highli": 100, "websit": [100, 102], "soon": [100, 109, 110], "situat": 100, "friend": 100, "love": 100, "hear": 100, "stori": 100, "came": 100, "cite": [100, 101], "bibtex": [100, 101], "citat": 100, "manag": 100, "pereira2022sleap": [100, 101], "tabri": [100, 101], "nathaniel": [100, 101], "matsliah": [100, 101], "ari": [100, 101], "turner": [100, 101], "david": [100, 101], "normand": [100, 101], "edna": [100, 101], "deutsch": [100, 101], "l": [100, 101, 104, 105, 106, 110], "journal": [100, 101], "volum": [100, 101], "year": [100, 101], "forum": 100, "anonym": 100, "transmit": 100, "broadest": 100, "userbas": 100, "sens": 100, "fund": [100, 101], "agenc": 100, "opt": 100, "launch": [100, 102, 118], "os": 101, "rapidli": 101, "customiz": [101, 112], "deliv": 101, "60": 101, "600": 101, "10m": 101, "app": [101, 102, 110, 118], "strongli": [101, 102], "miniconda": [101, 118], "forg": [101, 102, 104, 105, 106, 108], "nvidia": [101, 102, 104, 105, 106, 108, 109], "anaconda": [101, 102, 118], "silicon": [101, 102], "pypi": [101, 102, 104, 105, 106, 107, 108, 109, 110], "watch": 101, "aoc": 101, "2023": [101, 104, 105, 106, 107, 108, 109], "workshop": 101, "cbmm": 101, "quantif": 101, "2020": 101, "research": 101, "j": 101, "talmop": 101, "twitter": 101, "inquiri": 101, "institut": 101, "biolog": 101, "studi": 101, "liezl": 101, "mare": 101, "arlo": 101, "sheridan": 101, "princeton": 101, "nat": 101, "lewi": 101, "sigler": 101, "lab": [101, 102], "collabor": 101, "nih": 101, "brain": 101, "r01": 101, "ns104899": 101, "innov": 101, "acceler": [101, 102, 105, 106, 109, 110, 113], "bsd": 101, "academ": 101, "commerci": 101, "lauri": 101, "tzodikov": 101, "assist": [101, 113, 117, 120], "director": 101, "offic": 101, "technologi": 101, "609": 101, "258": 101, "7256": 101, "newest": 102, "enter": [102, 117, 118], "cmder": [102, 118], "ctrl": [102, 118], "cmd": [102, 118], "isol": 102, "solver": 102, "libmamba": 102, "mamba": 102, "wider": 102, "intel": 102, "powershel": 102, "admin": 102, "webrequest": 102, "miniforge3": 102, "x86_64": [102, 104, 105, 106, 107, 108, 109], "outfil": 102, "userprofil": 102, "argumentlist": 102, "installationtyp": 102, "justm": 102, "registerpython": 102, "curl": [102, 105, 106], "fssl": 102, "chmod": 102, "p": 102, "rm": [102, 104, 109], "bin": [102, 105, 109], "basenam": 102, "shell": 102, "rc": 102, "macosx": 102, "arm64": 102, "compani": 102, "repo": 102, "miniconda3": 102, "del": 102, "mkdir": [102, 109], "wget": [102, 103, 104, 107, 108, 109], "bash": 102, "driver": 102, "recogn": 102, "clone": 102, "talmolab": [102, 103, 104, 105, 106, 107, 108, 109], "cd": 102, "yml": 102, "environment_no_cuda": 102, "environment_mac": 102, "sleap_develop": 102, "venv": 102, "cudatoolkit": 102, "refrain": 102, "ship": 102, "jupyt": [102, 112], "conda_jupyt": 102, "conda_dev": 102, "moment": 102, "\u03bb": 102, "19041": 102, "sp0": 102, "system_summari": [102, 104, 105, 106, 108], "physical_devic": [102, 105, 106, 109], "list_physical_devic": 102, "device_typ": 102, "interfer": 102, "downgrad": 102, "stream_executor": [102, 104, 105, 106, 107, 108, 109], "platform": [102, 104, 105, 106, 107, 108, 109], "dso_load": [102, 104, 107, 108], "cc": [102, 104, 105, 106, 107, 108, 109], "libcudart": [102, 104, 107, 108], "dlerror": [102, 104, 107, 108], "conda_prefix": 102, "echo": 102, "sleap_activ": 102, "sleap_old_ld_library_path": 102, "ld_library_path": [102, 104, 107, 108], "lib": [102, 104, 107, 108, 109], "deactiv": 102, "sleap_deactiv": 102, "de": 102, "diagnost": 102, "saw": 102, "advis": 103, "3000": 103, "analysis_exampl": [103, 104, 105, 106], "2021": [103, 104, 105], "140": 103, "113": [103, 109], "443": [103, 104, 105, 109], "await": [103, 104, 109], "302": [103, 109], "githubusercont": [103, 109], "151": [103, 105, 106, 107, 109], "101": [103, 106, 109], "133": [103, 109], "551501": 103, "539k": 103, "octet": [103, 104, 109], "538": 103, "58k": 103, "kb": [103, 104], "mb": [103, 104, 105, 106, 109], "seaborn": [103, 107], "dset_nam": 103, "enumer": 103, "abdomen": [103, 104, 105, 106, 109], "wingl": [103, 104, 105, 106, 109], "wingr": [103, 104, 105, 106, 109], "forelegl4": [103, 104, 105, 106, 109], "forelegr4": [103, 104, 105, 106, 109], "midlegl4": [103, 104, 105, 106, 109], "midlegr4": [103, 104, 105, 106, 109], "hindlegl4": [103, 104, 105, 106, 109], "hindlegr4": [103, 104, 105, 106, 109], "eyel": [103, 104, 105, 106, 109], "eyer": [103, 104, 105, 106, 109], "frame_count": [103, 105], "node_count": 103, "interp1d": 103, "def": [103, 105], "initial_shap": 103, "dim": [103, 104, 107], "flatnonzero": 103, "fill_valu": 103, "bounds_error": 103, "xq": 103, "interp": 103, "restor": 103, "head_index": 103, "thorax_index": 103, "abdo_index": 103, "head_loc": 103, "thorax_loc": 103, "abdo_loc": 103, "sn": [103, 107], "mpl": [103, 107], "pyplot": [103, 105, 107], "plt": [103, 105, 107], "tick": 103, "font_scal": 103, "rcparam": 103, "figsiz": [103, 105, 107], "loc": [103, 107], "xlim": 103, "xtick": 103, "ylim": 103, "ytick": 103, "differenti": 103, "savitzki": 103, "golai": 103, "savgol_filt": 103, "smooth_diff": 103, "node_loc": 103, "win": 103, "poli": 103, "polynomi": 103, "node_loc_vel": 103, "zeros_lik": 103, "deriv": 103, "node_vel": 103, "linalg": 103, "thx_vel_fly0": 103, "thx_vel_fly1": 103, "fig": 103, "ax1": 103, "add_subplot": 103, "211": 103, "set_xtick": 103, "set_titl": 103, "ax2": 103, "212": [103, 105], "sharex": 103, "imshow": 103, "newaxi": 103, "vmin": 103, "vmax": 103, "set_ytick": 103, "121": 103, "set_xlim": 103, "set_ylim": 103, "kp": 103, "122": 103, "scatter": 103, "panda": [103, 107], "pd": [103, 107], "corr_rol": 103, "datax": 103, "datai": 103, "timeseri": 103, "correl": 103, "s1": 103, "s2": 103, "roll": 103, "corr": 103, "cov_vel": 103, "subplot": 103, "markers": 103, "tight_layout": 103, "instance_node_veloc": 103, "instance_idx": 103, "fly_node_loc": 103, "fly_node_veloc": 103, "plot_instance_node_veloc": 103, "node_veloc": 103, "xlabel": [103, 105, 107], "ylabel": [103, 105, 107], "fly_id": 103, "sklearn": 103, "nstate": 103, "km": 103, "fit_predict": 103, "set_xlabel": 103, "set_ylabel": 103, "set_yticklabel": 103, "tab10": 103, "ethogram": 103, "explor": [104, 105, 106, 108], "tail": 104, "plai": 104, "cell": [104, 105, 106, 108], "restart": [104, 105, 106, 108], "uninstal": [104, 105, 106, 107, 108, 109, 110], "qqq": [104, 105, 106, 107, 108, 109, 110], "contrib": [104, 105, 106, 107, 108, 109, 110], "googleapi": [104, 105, 106, 107], "flies13": [104, 105, 107], "190719_090330_wt_18159206_rig1": [104, 105], "4015000": [104, 105], "17560": [104, 105], "bu": [104, 105, 106, 109], "210506_230852": 104, "3d1800": [104, 105, 107], "210504_182918": [104, 105], "td_fast": [104, 105, 107], "210505_012601": [104, 105, 107], "142": 104, "176": 104, "144": [104, 109], "172": 104, "217": 104, "85343812": 104, "81m": 104, "15000": [104, 105], "190719_090330_wt_18": 104, "39m": 104, "7mb": 104, "188": 104, "240": [104, 105], "1581400": 104, "51m": 104, "99mb": 104, "189": 104, "6372537": 104, "1m": [104, 105, 109], "1800": [104, 105, 107], "21050": 104, "08m": 104, "112": 104, "30775963": 104, "29m": 104, "210505_0126": 104, "35m": [104, 109], "3mb": 104, "lah": [104, 105, 106], "239m": 104, "drwxrwxr": [104, 105, 106], "0k": [104, 105, 106], "aug": [104, 105, 106], "rw": [104, 105, 106], "82m": [104, 105], "6m": [104, 105], "jun": [104, 105, 106], "00": [104, 105, 106, 109], "713k": [104, 105, 106], "ipynb": [104, 105, 106], "486k": 104, "data_structur": [104, 105, 106], "1k": [104, 105, 106], "rst": [104, 105, 106], "197k": [104, 105], "interactive_and_realtime_infer": [104, 105, 106], "398k": 104, "interactive_and_resumable_train": [104, 105, 106], "149k": 104, "model_evalu": [104, 105, 106], "150k": 104, "post_inference_track": [104, 105, 106], "5k": 104, "training_and_inference_on_an_example_dataset": [104, 105, 106], "12k": [104, 105, 106], "training_and_inference_using_google_dr": [104, 105, 106], "989133": 104, "cpu_feature_guard": [104, 105, 106, 107, 109], "193": 104, "oneapi": [104, 105, 106, 107, 109], "onednn": [104, 105, 106, 107, 109], "avx2": [104, 105, 106, 107, 109], "avx512f": [104, 105, 106, 107, 109], "avx512_vnni": 104, "fma": [104, 105, 106, 107, 109], "compil": [104, 105, 106, 107, 109], "058048": 104, "104": 104, "slightli": 104, "tf_enable_onednn_opt": 104, "060007": 104, "xla": 104, "micromamba": [104, 107, 108, 109], "sleap_jupyt": [104, 107, 108], "python3": [104, 107, 108], "site": [104, 107, 108], "cv2": [104, 107, 108], "lib64": [104, 107, 108], "060013": 104, "cudart_stub": [104, 107, 108], "cudart": [104, 107, 108], "445179": 104, "libnvinf": 104, "445232": 104, "libnvinfer_plugin": 104, "445236": 104, "tf2tensorrt": 104, "py_util": 104, "trt": 104, "dlopen": [104, 107, 108], "tensorrt": 104, "mention": [104, 107, 108], "debian": [104, 105, 106, 107, 108, 109], "bookworm": [104, 105, 106, 107, 108, 109], "sid": [104, 105, 106, 107, 108, 109], "223182": 104, "cuda_gpu_executor": [104, 105, 106, 107, 108, 109], "981": 104, "success": [104, 105, 106, 107, 108, 109, 117], "numa": [104, 105, 106, 107, 108, 109], "sysf": [104, 105, 106, 107, 108, 109], "223923": 104, "223968": 104, "libcubla": [104, 107, 108], "223999": 104, "libcublaslt": [104, 107, 108], "224028": 104, "libcufft": [104, 107, 108], "224057": 104, "libcurand": [104, 107, 108], "224084": 104, "libcusolv": [104, 107, 108], "224111": 104, "libcuspars": [104, 107, 108], "224140": 104, "libcudnn": [104, 107, 108], "224144": 104, "common_runtim": [104, 105, 106, 107, 108, 109], "gpu_devic": [104, 105, 106, 107, 108, 109], "1934": 104, "blow": 104, "agnost": 104, "And": [104, 109], "2560": [104, 105], "498908": 104, "923466": 104, "grappler": [104, 107], "op_level_cost_estim": [104, 107], "690": [104, 107], "predictcost": [104, 107], "cropandres": [104, 107], "dt_float": [104, 107], "extrapolation_valu": [104, 107], "dt_int32": [104, 107], "vendor": [104, 107], "genuineintel": [104, 107], "3600": [104, 107], "num_cor": [104, 107], "cpu_instruction_set": [104, 107], "avx": [104, 107], "sse": [104, 107], "sse2": [104, 107], "sse3": [104, 107], "ssse3": [104, 107], "sse4": [104, 107], "eigen": [104, 107], "90": [104, 105, 107], "l1_cache_s": [104, 107], "49152": [104, 107], "l2_cache_s": [104, 107], "524288": [104, 107], "l3_cache_s": [104, 107], "16777216": [104, 107], "memory_s": [104, 107], "268435456": [104, 107], "923717": 104, "dt_uint8": [104, 107], "926044": 104, "234": [104, 105], "430": [104, 105], "271": [104, 105], "436": [104, 105], "308": [104, 105], "438": [104, 105], "321": [104, 105], "440": [104, 105], "322": [104, 105], "246": [104, 105], "450": [104, 105], "242": [104, 105], "413": [104, 105], "285": [104, 105], "459": [104, 105], "272": [104, 105], "317": [104, 105], "441": [104, 105], "245": [104, 105], "420": [104, 105], "319": [104, 105], "435": [104, 105], "354": [104, 105], "368": 104, "433": [104, 105], "393": [104, 105, 109], "480": 104, "398": [104, 105], "307": [104, 105], "445": [104, 105], "305": [104, 105], "421": [104, 105], "325": [104, 105], "475": [104, 105], "331": [104, 105], "385": [104, 105], "363": [104, 105], "474": [104, 105], "376": 104, "329": [104, 105], "327": [104, 105], "24440002441406": 104, "52008056640625": 104, "9790770411491394": 104, "58941650390625": 104, "1461486816406": 104, "9358043670654297": 104, "02960205078125": 104, "57135009765625": 104, "5861632227897644": 104, "81768798828125": 104, "08721923828125": 104, "39127233624458313": 104, "0193176269531": 104, "7702941894531": 104, "48629727959632874": 104, "14295959472656": 104, "5621643066406": 104, "9176925420761108": 104, "2632598876953": 104, "9497375488281": 104, "780803382396698": 104, "78155517578125": 104, "91552734375": 104, "47393468022346497": 104, "280029296875": 104, "71759033203125": 104, "7721256017684937": 104, "598876953125": 104, "6053466796875": 104, "296230286359787": 104, "10415649414062": 104, "9450378417969": 104, "8855596780776978": 104, "32009887695312": 104, "9360656738281": 104, "9200019240379333": 104, "concret": 104, "pt": 104, "24440002": 104, "52008057": 104, "5894165": 104, "14614868": 104, "02960205": 104, "5713501": 104, "81768799": 104, "08721924": 104, "01931763": 104, "77029419": 104, "14295959": 104, "56216431": 104, "26325989": 104, "94973755": 104, "78155518": 104, "91552734": 104, "2800293": 104, "71759033": 104, "59887695": 104, "60534668": 104, "10415649": 104, "94503784": 104, "32009888": 104, "93606567": 104, "labels_with_imag": [104, 105], "pure": 105, "forget": [105, 106], "31merror": [105, 106, 109, 110], "0m": [105, 106, 107, 109, 110], "31m": [105, 106, 109, 110], "shiboken2": [105, 106, 109, 110], "reinstal": [105, 106, 109, 110], "dep": [105, 106, 109, 110], "731425": 105, "735933": 105, "736867": 105, "centered_instance_id_model": 105, "td_id": 105, "210519_111253": 105, "xferd": [105, 106], "dload": [105, 106], "upload": [105, 106, 108], "spent": [105, 106], "3m": 105, "7m": 105, "6223k": 105, "2m": 105, "1g": 105, "sep": [105, 106], "33m": 105, "481k": [105, 106], "661k": 105, "fly_clip": [105, 108], "120k": [105, 106], "620m": [105, 106], "158k": 105, "157k": [105, 106], "412k": 105, "422k": 105, "78m": 105, "89m": 105, "trained_model": [105, 109], "94k": [105, 106], "806004": 105, "807011": 105, "807970": 105, "808962": 105, "103658": 105, "104377": 105, "105059": 105, "106019": 105, "1525": [105, 106, 109], "localhost": [105, 106, 109], "replica": [105, 106, 109], "21129": 105, "rtx": [105, 106, 109], "a5000": [105, 106, 109], "pci": [105, 106, 109], "0000": [105, 106, 109], "455046": 105, "cuda_dnn": [105, 106, 109], "366": [105, 106, 109], "8201": [105, 106, 109], "358483": 105, "cuda_bla": [105, 106, 109], "1774": [105, 106, 109], "tensorfloat": [105, 106, 109], "427": 105, "252": 105, "288": 105, "439": 105, "304": 105, "306": 105, "216": 105, "410": 105, "244": 105, "471": 105, "408": 105, "275": 105, "292": 105, "412": 105, "220": 105, "223": 105, "417": 105, "313": 105, "432": 105, "348": 105, "378": 105, "397": 105, "428": 105, "394": 105, "446": 105, "422": 105, "341": 105, "332": 105, "386": 105, "458": 105, "387": 105, "323": 105, "442": 105, "320": 105, "rec": 105, "51400757": 105, "426": 105, "97024536": 105, "251": 105, "97747803": 105, "08648682": 105, "64355469": 105, "3086853": 105, "53396606": 105, "33477783": 105, "20336914": 105, "77227783": 105, "24688721": 105, "47549438": 105, "14550781": 105, "409": 105, "98342896": 105, "39497375": 105, "31561279": 105, "61740112": 105, "89056396": 105, "274": 105, "97470093": 105, "1831665": 105, "2600708": 105, "411": 105, "95904541": 105, "219": 105, "98565674": 105, "437": 105, "97906494": 105, "75566101": 105, "5496521": 105, "176m": 105, "2224": 105, "62598": 105, "5043": 105, "13202": 105, "309": 105, "87125": 105, "64966": 105, "324": 105, "12512": 105, "3908": 105, "3458": 105, "9504": 105, "42352": 105, "67786": 105, "37636": 105, "81458": 105, "5624": 105, "460": 105, "22766": 105, "273": 105, "45117": 105, "51895": 105, "241": 105, "9716": 105, "32303": 105, "46788": 105, "90228": 105, "80017": 105, "48407": 105, "351": 105, "93695": 105, "434": 105, "0301": 105, "369": 105, "43228": 105, "431": 105, "78564": 105, "89014": 105, "481": 105, "0584": 105, "4241": 105, "429": 105, "79565": 105, "42896": 105, "419": 105, "3896": 105, "67926": 105, "0098": 105, "97974": 105, "30814": 105, "66406": 105, "473": 105, "9354": 105, "377": 105, "3852": 105, "13065": 105, "328": 105, "40244": 105, "51434": 105, "1667": 105, "423": 105, "94733": 105, "36913": 105, "38037": 105, "65576": 105, "0479": 105, "311": [105, 109], "6751": 105, "00995": 105, "48315": 105, "1421": 105, "20544": 105, "06784": 105, "43257": 105, "61487": 105, "3986": 105, "8269": 105, "565": 105, "00977": 105, "78204": 105, "46442": 105, "11816": 105, "0634": 105, "55441": 105, "7281": 105, "03793": 105, "2389": 105, "353": 105, "87274": 105, "77695": 105, "370": 105, "67218": 105, "9711": 105, "91922": 105, "09735": 105, "399": 105, "77133": 105, "25983": 105, "48993": 105, "82016": 105, "90115": 105, "94632": 105, "0652": 105, "65408": 105, "70728": 105, "68225": 105, "30194": 105, "72806": 105, "15356": 105, "48532": 105, "624": 105, "5559": 105, "06238": 105, "8775": 105, "9898": 105, "312": 105, "13086": 105, "16318": 105, "77222": 105, "65994": 105, "40115": 105, "7244": 105, "44681": 105, "51874": 105, "45566": 105, "7617": 105, "8958": 105, "56442": 105, "66855": 105, "2377": 105, "26588": 105, "80545": 105, "77664": 105, "7662": 105, "46994": 105, "2546": 105, "89484": 105, "93176": 105, "372": 105, "25574": 105, "46127": 105, "40717": 105, "479": 105, "5797": 105, "30173": 105, "96054": 105, "9821": 105, "449": 105, "3157": 105, "8817": 105, "52148": 105, "98843": 105, "9167": 105, "17917": 105, "04684": 105, "0318": 105, "50616": 105, "391": 105, "05493": 105, "396": 105, "85666": 105, "16904": 105, "04953": 105, "89996": 105, "52533": 105, "65547": 105, "6946": 105, "38303": 105, "68842": 105, "04352": 105, "86963": 105, "80847": 105, "3792": 105, "84747": 105, "93973": 105, "71852": 105, "451": 105, "2873": 105, "57388": 105, "58414": 105, "286": 105, "164": 105, "461": 105, "83655": 105, "8726": 105, "21753": 105, "43861": 105, "46246": 105, "25829": 105, "48416": 105, "7713": 105, "55927": 105, "356": 105, "25912": 105, "81424": 105, "98462": 105, "9266": 105, "402": 105, "0365": 105, "465": 105, "8439": 105, "7685": 105, "310": 105, "4258": 105, "7895": 105, "16397": 105, "86514": 105, "16724": 105, "9967": 105, "362": 105, "87766": 105, "12836": 105, "390": 105, "43555": 105, "69998": 105, "330": 105, "20596": 105, "4066": 105, "0497": 105, "68896": 105, "51591": 105, "5735": 105, "3791": 105, "4755": 105, "74457": 105, "20264": 105, "24997": 105, "69904": 105, "1339": 105, "8241": 105, "75269": 105, "22192": 105, "58466": 105, "53275": 105, "0668": 105, "6229": 105, "87787": 105, "2068": 105, "3858": 105, "31342": 105, "15892": 105, "27942": 105, "91632": 105, "5178": 105, "588": 105, "3604": 105, "374": 105, "51236": 105, "42508": 105, "405": 105, "0515": 105, "2759": 105, "401": 105, "2467": 105, "2713": 105, "314": 105, "74677": 105, "78735": 105, "76758": 105, "29553": 105, "20752": 105, "6215": 105, "86606": 105, "8446": 105, "472": 105, "95822": 105, "388": 105, "92188": 105, "203": 105, "54233": 105, "43842": 105, "1192": 105, "79416": 105, "54964": 105, "56854": 105, "30457": 105, "13345": 105, "08594": 105, "02444": 105, "13245": 105, "11148": 105, "71674": 105, "81714": 105, "79828": 105, "9629": 105, "6766": 105, "53745": 105, "09372": 105, "14362": 105, "87155": 105, "23718": 105, "4111": 105, "2425": 105, "13495": 105, "83694": 105, "7404": 105, "43884": 105, "4725": 105, "68488": 105, "375": 105, "05853": 105, "87177": 105, "404": 105, "3775": 105, "92688": 105, "39508": 105, "9776": 105, "77365": 105, "6409": 105, "17343": 105, "26575": 105, "44904": 105, "56747": 105, "05463": 105, "54587": 105, "72284": 105, "13287": 105, "25458": 105, "28958": 105, "9332": 105, "74493": 105, "15704": 105, "3947": 105, "1558": 105, "1859": 105, "46423": 105, "5753": 105, "42407": 105, "18854": 105, "80786": 105, "41486": 105, "72241": 105, "9671": 105, "64005": 105, "65726": 105, "9537": 105, "01648": 105, "73447": 105, "31635": 105, "89285": 105, "9849": 105, "21112": 105, "0713": 105, "77327": 105, "9886": 105, "03162": 105, "8643": 105, "5856": 105, "9501": 105, "2166": 105, "29108": 105, "09204": 105, "2633": 105, "97113": 105, "12497": 105, "74753": 105, "16742": 105, "3774": 105, "73508": 105, "5342": 105, "97403": 105, "56894": 105, "469": 105, "3632": 105, "81372": 105, "89886": 105, "641": 105, "67194": 105, "04425": 105, "418": 105, "99023": 105, "232": 105, "79128": 105, "2476": 105, "7884": 105, "45706": 105, "26096": 105, "58252": 105, "67697": 105, "28253": 105, "35138": 105, "4916": 105, "49533": 105, "1817": 105, "4297": 105, "56104": 105, "01126": 105, "4526": 105, "72516": 105, "3869": 105, "284": 105, "4912": 105, "79095": 105, "58961": 105, "1936": 105, "4464": 105, "00543": 105, "69318": 105, "96207": 105, "358": 105, "88284": 105, "98035": 105, "379": 105, "26816": 105, "0259": 105, "7312": 105, "5473": 105, "13306": 105, "02057": 105, "64542": 105, "59024": 105, "46237": 105, "79062": 105, "63318": 105, "9981": 105, "9735": 105, "466": 105, "93915": 105, "87518": 105, "89645": 105, "4858": 105, "98822": 105, "72357": 105, "30713": 105, "9138": 105, "26993": 105, "89908": 105, "6341": 105, "36536": 105, "9696": 105, "63763": 105, "87323": 105, "4065": 105, "7932": 105, "48575": 105, "27322": 105, "48721": 105, "6446": 105, "74454": 105, "08987": 105, "75647": 105, "338": 105, "82465": 105, "17297": 105, "64159": 105, "22705": 105, "54178": 105, "04788": 105, "2764": 105, "7331": 105, "359": 105, "43756": 105, "0462": 105, "8793": 105, "56253": 105, "407": 105, "32346": 105, "448": 105, "95087": 105, "74594": 105, "27792": 105, "315": 105, "46356": 105, "38046": 105, "48642": 105, "8147": 105, "63016": 105, "81934": 105, "73767": 105, "03244": 105, "19778": 105, "1395": 105, "32227": 105, "00305": 105, "94138": 105, "76627": 105, "20728": 105, "03998": 105, "59984": 105, "94275": 105, "68756": 105, "925": [105, 106], "74356": 105, "45367": 105, "3493": 105, "94934": 105, "39355": 105, "09885": 105, "09349": 105, "45755": 105, "8041": 105, "40057": 105, "55066": 105, "6091": 105, "43307": 105, "35394": 105, "9949": 105, "04814": 105, "10544": 105, "36105": 105, "95673": 105, "50397": 105, "86414": 105, "65952": 105, "01282": 105, "381": 105, "80063": 105, "2879": 105, "9239": 105, "02728": 105, "27682": 105, "24774": 105, "42343": 105, "444": 105, "4193": 105, "38232": 105, "35754": 105, "6553": 105, "45853": 105, "8156": 105, "7812": 105, "62988": 105, "58368": 105, "52002": 105, "27118": 105, "3299": 105, "78006": 105, "43085": 105, "03174": 105, "25208": 105, "7414": 105, "57523": 105, "99503": 105, "347": 105, "97897": 105, "64392": 105, "52814": 105, "16394": 105, "4637": 105, "9819": 105, "81857": 105, "35481": 105, "535": 105, "59384": 105, "70065": 105, "50806": 105, "95544": 105, "72635": 105, "96436": 105, "3465": 105, "5469": 105, "239": 105, "56883": 105, "8733": 105, "04318": 105, "60315": 105, "36966": 105, "4342": 105, "360": [105, 106], "08127": 105, "41803": 105, "283": 105, "4751": 105, "8785": 105, "59448": 105, "36377": 105, "55792": 105, "316": 105, "73703": 105, "6411": 105, "78436": 105, "899": 105, "92154": 105, "19464": 105, "91168": 105, "32022": 105, "73245": 105, "457": 105, "32578": 105, "57062": 105, "18298": 105, "334": 105, "3139": 105, "40005": 105, "89133": 105, "64728": 105, "70679": 105, "36255": 105, "08994": 105, "64023": 105, "14267": 105, "50543": 105, "68262": 105, "5147": 105, "82147": 105, "87054": 105, "32256": 105, "4768": 105, "57848": 105, "34476": 105, "7278": 105, "26282": 105, "8772": 105, "77335": 105, "55972": 105, "77054": 105, "95602": 105, "7761": 105, "31602": 105, "2043": 105, "92737": 105, "62262": 105, "1631": 105, "82898": 105, "76257": 105, "35846": 105, "0152": 105, "79132": 105, "68118": 105, "318": 105, "88504": 105, "35867": 105, "8374": 105, "72562": 105, "326": [105, 109], "25244": 105, "88055": 105, "2403": 105, "27567": 105, "44467": 105, "21188": 105, "84778": 105, "12372": 105, "335": 105, "4058": 105, "62274": 105, "9344": 105, "79385": 105, "12885": 105, "08496": 105, "77728": 105, "3099": 105, "81348": 105, "31982": 105, "3584": 105, "60254": 105, "39813": 105, "222": 105, "91586": 105, "43195": 105, "6026": 105, "74078": 105, "7489": 105, "09384": 105, "8778": 105, "4865": 105, "287": 105, "56982": 105, "68353": 105, "8655": 105, "80096": 105, "19046": 105, "23196": 105, "46782": 105, "99805": 105, "60196": 105, "36603": 105, "8261": 105, "19696": 105, "17218": 105, "47867": 105, "81366": 105, "09143": 105, "99658": 105, "15668": 105, "84363": 105, "00012": 105, "57254": 105, "2615": 105, "333": 105, "60617": 105, "467": 105, "9318": 105, "6039": 105, "32465": 105, "44635": 105, "16357": 105, "92133": 105, "11078": 105, "336": 105, "72043": 105, "8229": 105, "6642": 105, "31372": 105, "83435": 105, "26373": 105, "11572": 105, "61078": 105, "17926": 105, "66278": 105, "83746": 105, "15924": 105, "40552": 105, "6441": 105, "87045": 105, "41144": 105, "64131": 105, "59937": 105, "06647": 105, "35507": 105, "84183": 105, "3719": 105, "41534": 105, "61237": 105, "42723": 105, "2208": 105, "4097": 105, "95218": 105, "3499": 105, "52005": 105, "361": 105, "36267": 105, "389": 105, "47607": 105, "60114": 105, "6601": 105, "50894": 105, "51843": 105, "6943": 105, "90283": 105, "82428": 105, "70898": 105, "5036": 105, "345": 105, "66882": 105, "1757": 105, "79486": 105, "46274": 105, "46533": 105, "10553": 105, "24854": 105, "009": 105, "337": 105, "8076": 105, "06436": 105, "29004": 105, "49707": 105, "41422": 105, "4673": 105, "8141": 105, "7682": 105, "01324": 105, "98956": 105, "19714": 105, "71683": 105, "71207": 105, "39133": 105, "85786": 105, "4593": 105, "5914": 105, "65207": 105, "67142": 105, "77646": 105, "7307": 105, "5118": 105, "71594": 105, "21155": 105, "99216": 105, "57278": 105, "243": 105, "82819": 105, "339": 105, "47983": 105, "74188": 105, "93173": 105, "2397": 105, "49423": 105, "05255": 105, "68115": 105, "6671": 105, "5454": 105, "09042": 105, "30078": 105, "80396": 105, "82977": 105, "7456": 105, "64886": 105, "89554": 105, "84995": 105, "1559": 105, "78146": 105, "11206": 105, "5744": 105, "94125": 105, "9305": 105, "99496": 105, "5468": 105, "42017": 105, "05379": 105, "01157": 105, "71146": 105, "17175": 105, "08688": 105, "66077": 105, "65015": 105, "097": 105, "3269": 105, "45065": 105, "67744": 105, "92798": 105, "56874": 105, "94662": 105, "72803": 105, "462": 105, "40347": 105, "67886": 105, "66385": 105, "6862": 105, "8137": 105, "559": 105, "416": 105, "42374": 105, "62582": 105, "11035": 105, "73026": 105, "8417": 105, "30188": 105, "77295": 105, "364": 105, "57666": 105, "20844": 105, "32507": 105, "96838": 105, "414": 105, "35016": 105, "2262": 105, "04324": 105, "17578": 105, "63538": 105, "82654": 105, "36795": 105, "21204": 105, "9988": 105, "7216": 105, "20065": 105, "24988": 105, "20847": 105, "9794": 105, "04962": 105, "340": 105, "75934": 105, "0198": 105, "4428": 105, "33124": 105, "9914025": 105, "9798533": 105, "7552497": 105, "45417705": 105, "49756864": 105, "8265212": 105, "89824754": 105, "7941327": 105, "81785023": 105, "05611448": 105, "06403984": 105, "88647026": 105, "96359974": 105, "9033977": 105, "25969282": 105, "6343123": 105, "8396003": 105, "7613073": 105, "04938014": 105, "84057474": 105, "8820076": 105, "8816869": 105, "8243384": 105, "33521563": 105, "8434063": 105, "8127704": 105, "9598888": 105, "97341204": 105, "6766811": 105, "35414153": 105, "49778372": 105, "883279": 105, "9271338": 105, "7989652": 105, "7574282": 105, "04437362": 105, "06203796": 105, "8609162": 105, "89723104": 105, "8814398": 105, "43337214": 105, "6627722": 105, "8388201": 105, "71751094": 105, "08318384": 105, "7553143": 105, "8750135": 105, "8972577": 105, "85390973": 105, "87049603": 105, "84071857": 105, "8853136": 105, "9277581": 105, "9876475": 105, "71884066": 105, "36052382": 105, "53324103": 105, "89681005": 105, "92098916": 105, "8180281": 105, "6177351": 105, "0311976": 105, "07055778": 105, "83666444": 105, "8608399": 105, "8386477": 105, "58817774": 105, "72051835": 105, "7902795": 105, "7041355": 105, "2181147": 105, "76299024": 105, "8507803": 105, "8824023": 105, "8892915": 105, "8559173": 105, "83882904": 105, "9163557": 105, "9318335": 105, "0054291": 105, "7037247": 105, "44776785": 105, "55141157": 105, "8751741": 105, "8788193": 105, "7378067": 105, "6061791": 105, "06516132": 105, "145283": 105, "81688696": 105, "88854957": 105, "85625255": 105, "86021763": 105, "82891417": 105, "5004723": 105, "8896506": 105, "15082283": 105, "57127994": 105, "86683005": 105, "94244254": 105, "8910252": 105, "9375356": 105, "92730576": 105, "8518939": 105, "9335175": 105, "98755246": 105, "66180676": 105, "5590857": 105, "5017098": 105, "89124495": 105, "8839093": 105, "77439654": 105, "5733776": 105, "0646795": 105, "12731166": 105, "816599": 105, "90029544": 105, "9238624": 105, "8279644": 105, "7274184": 105, "8509916": 105, "9116395": 105, "21640316": 105, "4109717": 105, "92344654": 105, "8912647": 105, "8676515": 105, "91081876": 105, "9236755": 105, "9313457": 105, "9660537": 105, "97779256": 105, "6795893": 105, "5347014": 105, "49429995": 105, "89868015": 105, "88998085": 105, "82294524": 105, "49898362": 105, "14230077": 105, "13475017": 105, "8461558": 105, "89860517": 105, "8971772": 105, "85703963": 105, "743163": 105, "87278444": 105, "90552235": 105, "19766915": 105, "33566353": 105, "89383173": 105, "87157995": 105, "83140534": 105, "92693084": 105, "9499294": 105, "85782766": 105, "9214447": 105, "9804845": 105, "6575725": 105, "46105212": 105, "5740245": 105, "88368326": 105, "89460224": 105, "81119704": 105, "50101817": 105, "24979575": 105, "16411652": 105, "83694774": 105, "9241573": 105, "8916": 105, "87129986": 105, "7239725": 105, "8828186": 105, "7020806": 105, "16116264": 105, "36204475": 105, "8973187": 105, "8997571": 105, "51675177": 105, "89034307": 105, "98887885": 105, "88438815": 105, "8979453": 105, "97743154": 105, "5481076": 105, "523632": 105, "570176": 105, "8288708": 105, "9113763": 105, "9194614": 105, "575856": 105, "07603623": 105, "21255928": 105, "9018014": 105, "9266098": 105, "91993105": 105, "8616991": 105, "781426": 105, "7750215": 105, "85324234": 105, "14189687": 105, "5463986": 105, "8761287": 105, "93542594": 105, "50916994": 105, "87139845": 105, "8620718": 105, "9169966": 105, "90489644": 105, "9633726": 105, "6176859": 105, "6120859": 105, "53412354": 105, "8082982": 105, "9141492": 105, "8100913": 105, "7064677": 105, "07797408": 105, "28660768": 105, "9255538": 105, "9081669": 105, "9197768": 105, "89081717": 105, "7697851": 105, "850639": 105, "8240589": 105, "2276387": 105, "7375747": 105, "9573141": 105, "95667875": 105, "7197965": 105, "8762751": 105, "8575352": 105, "8765895": 105, "9522048": 105, "96551245": 105, "72864616": 105, "5890152": 105, "561211": 105, "7051566": 105, "9421855": 105, "39786857": 105, "7715297": 105, "6171893": 105, "06328589": 105, "0118455": 105, "886791": 105, "9031525": 105, "9011465": 105, "7290425": 105, "84665924": 105, "85558087": 105, "35440978": 105, "8101312": 105, "931835": 105, "91998947": 105, "9771716": 105, "88361436": 105, "8611444": 105, "88294595": 105, "93872": 105, "97103214": 105, "63806784": 105, "89063996": 105, "68062663": 105, "9067393": 105, "89928836": 105, "40190646": 105, "75169766": 105, "5388288": 105, "30325472": 105, "86616135": 105, "864786": 105, "9355017": 105, "93469065": 105, "73501164": 105, "89369905": 105, "794787": 105, "29464462": 105, "91743165": 105, "88107586": 105, "89442694": 105, "97276276": 105, "9208387": 105, "8436978": 105, "9492276": 105, "91440874": 105, "97273135": 105, "64372706": 105, "85304886": 105, "6125536": 105, "89858156": 105, "89086473": 105, "33406225": 105, "7624657": 105, "64882857": 105, "18051867": 105, "93381244": 105, "90368915": 105, "9286875": 105, "93761605": 105, "7948513": 105, "81816167": 105, "7628807": 105, "30384466": 105, "83553046": 105, "83106405": 105, "9189269": 105, "93762034": 105, "94770956": 105, "8512343": 105, "9446315": 105, "9450149": 105, "9582136": 105, "78555703": 105, "7544447": 105, "58366936": 105, "85938": 105, "94498163": 105, "6194322": 105, "7035529": 105, "22808443": 105, "24900974": 105, "981288": 105, "92618316": 105, "93841267": 105, "9422818": 105, "80968696": 105, "8445456": 105, "7991047": 105, "4916717": 105, "77814513": 105, "6231525": 105, "93198806": 105, "9570074": 105, "95540506": 105, "9207018": 105, "8778759": 105, "9381855": 105, "94920886": 105, "77673894": 105, "87591183": 105, "3847992": 105, "88775337": 105, "92982674": 105, "8082221": 105, "6930795": 105, "16653292": 105, "26732486": 105, "9830136": 105, "93462956": 105, "9093149": 105, "96090955": 105, "8409559": 105, "83797425": 105, "8743328": 105, "82546026": 105, "32881752": 105, "5494046": 105, "9653242": 105, "9882784": 105, "85375595": 105, "95603913": 105, "9316707": 105, "9048104": 105, "92460406": 105, "75584614": 105, "8082359": 105, "47512543": 105, "8684657": 105, "9260271": 105, "8822638": 105, "71269846": 105, "1508674": 105, "22018598": 105, "9016738": 105, "90536344": 105, "918121": 105, "96696764": 105, "78534484": 105, "883681": 105, "798996": 105, "69723856": 105, "5170047": 105, "8321578": 105, "9426196": 105, "9527973": 105, "91900206": 105, "9706679": 105, "90770215": 105, "9391487": 105, "9352003": 105, "85189575": 105, "72796327": 105, "6884535": 105, "8768972": 105, "9508924": 105, "6879568": 105, "71122557": 105, "7012927": 105, "6031595": 105, "87616193": 105, "91429555": 105, "8932258": 105, "97501004": 105, "78940654": 105, "8651793": 105, "72244436": 105, "82689875": 105, "4597148": 105, "93260366": 105, "9202296": 105, "94214964": 105, "8834407": 105, "98030627": 105, "8976605": 105, "9953135": 105, "99476504": 105, "99593395": 105, "99526805": 105, "9959078": 105, "9945123": 105, "99573624": 105, "993386": 105, "99603134": 105, "99172956": 105, "99564207": 105, "9916197": 105, "9947187": 105, "9915406": 105, "9940315": 105, "98916876": 105, "99394447": 105, "98962784": 105, "9944642": 105, "9910501": 105, "99155337": 105, "9933716": 105, "9916019": 105, "9933976": 105, "9932473": 105, "9932013": 105, "9920751": 105, "9946308": 105, "991653": 105, "99465877": 105, "99162734": 105, "99486005": 105, "8735": 105, "4811": 105, "355": 105, "93707": 105, "63477": 105, "0215": 105, "42197": 105, "2099": 105, "4682": 105, "23578": 105, "31976": 105, "61108": 105, "4756": 105, "57007": 105, "15857": 105, "7147": 105, "9847": 105, "93347": 105, "7111": 105, "8055": 105, "86227": 105, "03918": 105, "64484": 105, "49347": 105, "357": 105, "0275": 105, "29968": 105, "49817": 105, "54977": 105, "29578": 105, "42874": 105, "1338": 105, "81354": 105, "7555": 105, "4507": 105, "2437": 105, "95605": 105, "9807": 105, "4453": 105, "04776": 105, "2247": 105, "3565": 105, "81192": 105, "30673": 105, "8997": 105, "5365": 105, "4532": 105, "9694": 105, "10843": 105, "52646": 105, "42938": 105, "09125": 105, "59042": 105, "5916": 105, "44873": 105, "94284": 105, "18994": 105, "5058": 105, "18735": 105, "0978": 105, "8356": 105, "49683": 105, "1019": 105, "49136": 105, "94554764": 105, "83948356": 105, "9591119": 105, "8525362": 105, "95961505": 105, "86304706": 105, "9252076": 105, "97578657": 105, "974096": 105, "9668305": 105, "9845507": 105, "9572475": 105, "9105379": 105, "97522974": 105, "880064": 105, "9943127": 105, "911333": 105, "0001038": 105, "9698766": 105, "9948527": 105, "96454924": 105, "9799493": 105, "96142364": 105, "0046191": 105, "95354944": 105, "9987816": 105, "94746464": 105, "98374254": 105, "97818244": 105, "98671097": 105, "9833999": 105, "98425347": 105, "grabber": 105, "perf_count": 105, "simulatedcamera": 105, "grab": 105, "grab_fram": 105, "recording_dur": 105, "512": [105, 109], "inference_tim": 105, "frames_record": 105, "took": 105, "t0": 105, "frame_predict": 105, "expand_dim": 105, "dt": 105, "millisecond": 105, "first_inference_tim": 105, "1f": 105, "886": 105, "120": 105, "facecolor": [105, 107], "hist": 105, "training_slp_fil": 106, "b990gxjt3d3j3jh": 106, "210205": 106, "sleap_wt_gold": 106, "13pt": [106, 107], "dl": 106, "wt_gold": [106, 107], "tracking_split2": [106, 107], "619m": 106, "9m": 106, "4m": 106, "622m": 106, "179k": 106, "132k": 106, "reproduc": 106, "preprocesss": 106, "baseline_model": 106, "nitti": 106, "gritti": 106, "mechan": 106, "1440": 106, "160": 106, "775633": 106, "776555": 106, "777493": 106, "778196": 106, "055738": 106, "056597": 106, "057389": 106, "058046": 106, "21261": 106, "799": 106, "501": 106, "kerastensor": [106, 109], "type_spec": [106, 109], "tensorspec": [106, 109], "biasadd": [106, 109], "586040": 106, "104556": 106, "0037": [106, 109], "0030": [106, 109], "0036": [106, 109], "0040": [106, 109], "0038": [106, 109], "0041": [106, 109], "0039": [106, 109], "0035": [106, 109], "val_loss": [106, 109], "0033": [106, 109], "val_head": [106, 109], "0020": [106, 109], "val_thorax": [106, 109], "0029": [106, 109], "val_abdomen": [106, 109], "val_wingl": [106, 109], "val_wingr": [106, 109], "0034": [106, 109], "val_forelegl4": [106, 109], "val_forelegr4": [106, 109], "val_midlegl4": [106, 109], "val_midlegr4": [106, 109], "val_hindlegl4": [106, 109], "val_hindlegr4": [106, 109], "val_eyel": [106, 109], "val_ey": [106, 109], "0027": [106, 109], "0000e": [106, 109], "32m": [106, 109], "0028": [106, 109], "0013": [106, 109], "0018": [106, 109], "0026": [106, 109], "0032": [106, 109], "0015": [106, 109], "0025": [106, 109], "7323e": 106, "0011": [106, 109], "0024": [106, 109], "0012": [106, 109], "21m": 106, "0022": [106, 109], "0630e": 106, "7199e": 106, "0021": [106, 109], "7345e": 106, "4145e": 106, "6439e": 106, "9914e": [106, 109], "0019": [106, 109], "0031": [106, 109], "9466e": 106, "5174e": 106, "20m": 106, "7854e": 106, "6945e": [106, 109], "0017": [106, 109], "0023": [106, 109], "4546e": 106, "9585e": 106, "7640e": 106, "3180e": 106, "2650e": 106, "0064e": 106, "8714e": 106, "0531e": [106, 109], "7827e": 106, "2254e": 106, "5523e": 106, "4019e": 106, "0016": [106, 109], "9871e": [106, 109], "8608e": 106, "3215e": 106, "5232e": 106, "0014": [106, 109], "9826e": 106, "6906e": 106, "3776e": 106, "7946e": 106, "6378e": 106, "5611e": 106, "19m": [106, 109], "8557e": 106, "1089e": 106, "6096e": 106, "3123e": 106, "2092e": 106, "4376e": 106, "4288e": 106, "0581e": 106, "3752e": 106, "7513e": [106, 109], "1807e": 106, "9554e": 106, "6743e": 106, "5883e": 106, "2925e": 106, "5965e": 106, "2635e": 106, "4829e": 106, "0010": [106, 109], "0197e": 106, "8384e": 106, "8699e": 106, "5631e": 106, "1692e": 106, "8370e": 106, "8454e": 106, "9611e": 106, "2278e": 106, "4893e": 106, "5555e": 106, "7396e": 106, "4770e": 106, "9330e": 106, "9460e": 106, "5190e": [106, 109], "9289e": 106, "5512e": 106, "3737e": 106, "eval": [106, 107, 109], "508754": 106, "477220": 106, "3201e": 106, "7118e": 106, "0303e": 106, "9319e": 106, "0134e": 106, "4919e": 106, "2012e": [106, 109], "4680e": 106, "9131e": 106, "4191e": 106, "9155e": 106, "9295e": 106, "3658e": 106, "0085e": 106, "8906e": 106, "6015e": 106, "9128e": 106, "5054e": 106, "5352e": 106, "3093e": 106, "0690e": 106, "9501e": 106, "1907e": 106, "3487e": 106, "6145e": 106, "4151e": 106, "2130e": 106, "9293e": 106, "5396e": 106, "4440e": [106, 109], "7180e": 106, "9867e": 106, "1743e": 106, "2288e": 106, "7110e": 106, "1497e": 106, "9294e": 106, "8076e": 106, "7130e": 106, "4712e": 106, "2889e": 106, "5931e": 106, "9400e": 106, "9486e": 106, "6961e": 106, "559100": 106, "529680": 106, "new_fold": 106, "load_weight": 106, "3664e": 106, "7037e": 106, "8467e": 106, "9929e": 106, "0385e": 106, "5228e": [106, 109], "8510e": 106, "0772e": 106, "9413e": 106, "7351e": 106, "0943e": 106, "7453e": 106, "4413e": 106, "3617e": [106, 109], "4860e": 106, "4441e": 106, "4847e": 106, "4179e": 106, "24m": 106, "0541e": 106, "4627e": 106, "6070e": 106, "4325e": 106, "7257e": 106, "7434e": 106, "9573e": 106, "3483e": 106, "0939e": 106, "8417e": 106, "2339e": 106, "9561e": 106, "2637e": 106, "6513e": 106, "1751e": 106, "5540e": 106, "5588e": 106, "8189e": [106, 109], "2402e": 106, "7741e": [106, 109], "2087e": 106, "4398e": 106, "1826e": 106, "4005e": 106, "5282e": 106, "6551e": 106, "9726e": 106, "8423e": [106, 109], "7468e": [106, 109], "4657e": 106, "5649e": 106, "2162e": [106, 109], "9171e": 106, "9007e": [106, 109], "2471e": [106, 109], "1375e": 106, "8217e": 106, "7370e": 106, "2098e": 106, "585451": 106, "574921": 106, "apt": [107, 109], "qq": 107, "q": [107, 108], "31me": 107, "0mcould": 107, "lock": [107, 109], "var": [107, 109], "dpkg": [107, 109], "frontend": [107, 109], "deni": [107, 109], "0munabl": 107, "acquir": [107, 109], "34mtd_fast": 107, "00m": [107, 109], "with_centroid": 107, "model_info": 107, "982109": 107, "982120": 107, "mpck_part": 107, "ap": 107, "9010142093245521": 107, "0545800141197903": 107, "0166608700891056": 107, "histplot": 107, "binrang": 107, "kde": 107, "kde_kw": 107, "stat": 107, "intuit": 107, "holist": 107, "1707": 107, "05388": 107, "seem": 107, "summar": 107, "curv": 107, "stringent": 107, "2f": 107, "literatur": 107, "8150621709973537": 107, "8885": 107, "newer": 107, "gt": 107, "208933": 107, "936": [107, 108], "209734": 107, "209771": 107, "209801": 107, "209829": 107, "209859": 107, "209886": 107, "209912": 107, "209939": 107, "209945": 107, "1850": [107, 108], "245745": 107, "317060": 107, "320224": 107, "8984147543126978": 107, "197896466395166": 107, "1484228079076315": 107, "797836431061851": 107, "8782499999999999": 107, "hassl": 108, "250591": 108, "250602": 108, "389239": 108, "390139": 108, "390188": 108, "390230": 108, "390267": 108, "390306": 108, "390345": 108, "390383": 108, "390421": 108, "390425": 108, "uo": 108, "blob": 108, "predictions_fil": 108, "1350": 108, "make_tracker_by_nam": 108, "tracked_lf": 108, "tracked_label": 108, "dm": 109, "v1": 109, "unabl": 109, "2e65b": 109, "263375180": 109, "16df8d00": 109, "94f1": 109, "11ea": 109, "98d1": 109, "6c03a2f89e1c": 109, "amz": 109, "aws4": 109, "hmac": 109, "sha256": 109, "credenti": 109, "akiaiwnjyax4csveh53a": 109, "2f20230901": 109, "2fu": 109, "east": 109, "2fs3": 109, "2faws4_request": 109, "20230901t203033z": 109, "expir": 109, "300": 109, "b9b0638744af3144affdc46668c749128bd6c4f23ca2a1313821c7bbcd54ccdd": 109, "signedhead": 109, "actor_id": 109, "key_id": 109, "repo_id": 109, "disposit": 109, "attach": 109, "3b": 109, "20filenam": 109, "3ddrosophila": 109, "2foctet": 109, "185": 109, "199": 109, "108": 109, "109": 109, "110": 109, "111973079": 109, "107m": 109, "106": 109, "79m": 109, "0mb": 109, "inflat": 109, "ds_store": 109, "__macosx": 109, "20190128_113421": 109, "_20190128_113421": 109, "courtship_label": 109, "_courtship_label": 109, "_exampl": 109, "34mdataset": 109, "34mdrosophila": 109, "32m20190128_113421": 109, "32mcourtship_label": 109, "35mexampl": 109, "34m__macosx": 109, "satisfi": [109, 110, 119], "null": 109, "first_gpu": 109, "last_gpu": 109, "pretrained_encod": 109, "827290": 109, "831845": 109, "832633": 109, "22980": 109, "134": 109, "755154": 109, "756024": 109, "757213": 109, "758315": 109, "089801": 109, "090652": 109, "091464": 109, "092164": 109, "21084": 109, "953": 109, "814560": 109, "940585": 109, "5945e": 109, "99m": 109, "2513e": 109, "5694e": 109, "57m": 109, "6987e": 109, "8224e": 109, "1486e": 109, "0657e": 109, "58m": 109, "2174e": 109, "3859e": 109, "55m": 109, "9181e": 109, "0259e": 109, "9353e": 109, "9832e": 109, "8997e": 109, "4787e": 109, "5596e": 109, "5150e": 109, "9256e": 109, "8968e": 109, "8572e": 109, "5451e": 109, "2156e": 109, "8602e": 109, "53m": 109, "7656e": 109, "1905e": 109, "6440e": 109, "6607e": 109, "4415e": 109, "1699e": 109, "3589e": 109, "5362e": 109, "56m": 109, "0888e": 109, "1600e": 109, "0426e": 109, "6782e": 109, "9092e": 109, "8284e": 109, "0018e": 109, "9439e": 109, "7977e": 109, "8703e": 109, "00022": 109, "reducelronplateau": 109, "999999873689376e": 109, "5981e": 109, "6030e": 109, "6479e": 109, "8081e": 109, "2579e": 109, "7954e": 109, "9628e": 109, "6399e": 109, "6915e": 109, "9973e": 109, "4726e": 109, "5831e": 109, "2110e": 109, "7290e": 109, "3421e": 109, "1827e": 109, "3472e": 109, "4653e": 109, "00031": 109, "499999936844688e": 109, "1221e": 109, "5739e": 109, "2486e": 109, "5000e": 109, "5589e": 109, "3135e": 109, "4215e": 109, "8923e": 109, "4033e": 109, "8776e": 109, "3358e": 109, "5874e": 109, "2922e": 109, "6051e": 109, "00038": 109, "249999968422344e": 109, "1278e": 109, "4898e": 109, "0474e": 109, "8901e": 109, "2500e": 109, "0612e": 109, "7469e": 109, "8414e": 109, "8496e": 109, "0196e": 109, "5206e": 109, "8551e": 109, "6483e": 109, "9705e": 109, "4643e": 109, "00045": 109, "24999984211172e": 109, "9136e": 109, "8379e": 109, "7911e": 109, "0055e": 109, "00046": 109, "2kpredict": 109, "114": 109, "156": 109, "35m100": 109, "eta": 109, "36m0": 109, "31m33": 109, "0m31m51": 109, "0m31m52": 109, "0mfp": 109, "25hinfo": 109, "725241": 109, "31m7": 109, "0m0": 109, "31m184": 109, "0mm": 109, "870526": 109, "topdown_confmap": 109, "324520": 109, "329181": 109, "329961": 109, "23056": 109, "254912": 109, "255468": 109, "256291": 109, "257158": 109, "546117": 109, "546866": 109, "547533": 109, "548184": 109, "21151": 109, "684": 109, "877": 109, "940083": 109, "337645": 109, "0108": 109, "0073": 109, "0067": 109, "0111": 109, "0125": 109, "0126": 109, "0127": 109, "0128": 109, "0131": 109, "0082": 109, "0083": 109, "0087": 109, "0089": 109, "0105": 109, "0106": 109, "0091": 109, "0123": 109, "0116": 109, "0045": 109, "38m": 109, "0064": 109, "0057": 109, "0061": 109, "0075": 109, "0078": 109, "0092": 109, "0099": 109, "0102": 109, "0049": 109, "0056": 109, "0079": 109, "0086": 109, "0093": 109, "0081": 109, "0048": 109, "9048e": 109, "0051": 109, "0063": 109, "0066": 109, "0076": 109, "0080": 109, "0058": 109, "0044": 109, "0070": 109, "0072": 109, "0088": 109, "0085": 109, "0097": 109, "16m": 109, "6417e": 109, "0060": 109, "6555e": 109, "0042": 109, "0065": 109, "0069": 109, "0071": 109, "0062": 109, "15m": 109, "1233e": 109, "0052": 109, "0054": 109, "3809e": 109, "0053": 109, "0068": 109, "5957e": 109, "3519e": 109, "0047": 109, "7287e": 109, "6837e": 109, "9723e": 109, "0046": 109, "7941e": 109, "5772e": 109, "3055e": 109, "9191e": 109, "1078e": 109, "4425e": 109, "8283e": 109, "7111e": 109, "2022e": 109, "1555e": 109, "1508e": 109, "8990e": 109, "4963e": 109, "9041e": 109, "7679e": 109, "6666e": 109, "9156e": 109, "0043": 109, "6281e": 109, "0986e": 109, "9581e": 109, "8376e": 109, "2502e": 109, "1183e": 109, "7891e": 109, "4567e": 109, "6811e": 109, "7635e": 109, "3648e": 109, "2940e": 109, "6554e": 109, "8863e": 109, "2604e": 109, "0488e": 109, "1238e": 109, "5798e": 109, "1007e": 109, "5654e": 109, "1802e": 109, "9789e": 109, "0050": 109, "14m": 109, "8818e": 109, "1018e": 109, "8027e": 109, "8017e": 109, "4529e": 109, "6272e": 109, "3265e": 109, "5806e": 109, "6352e": 109, "4029e": 109, "8339e": 109, "6739e": 109, "0118e": 109, "4831e": 109, "1158e": 109, "2793e": 109, "1400e": 109, "4292e": 109, "1119e": 109, "8295e": 109, "3028e": 109, "3006e": 109, "9391e": 109, "8825e": 109, "0989e": 109, "3611e": 109, "4493e": 109, "1164e": 109, "4152e": 109, "4500e": 109, "3208e": 109, "3573e": 109, "0631e": 109, "5007e": 109, "3431e": 109, "9773e": 109, "1004e": 109, "7803e": 109, "3279e": 109, "9841e": 109, "5381e": 109, "0128e": 109, "3161e": 109, "0100e": 109, "8088e": 109, "9153e": 109, "7586e": 109, "9866e": 109, "6100e": 109, "0266e": 109, "9697e": 109, "9149e": 109, "7208e": 109, "8070e": 109, "1903e": 109, "8509e": 109, "7838e": 109, "7844e": 109, "9896e": 109, "9112e": 109, "7495e": 109, "5591e": 109, "5877e": 109, "9042e": 109, "6195e": 109, "9452e": 109, "6819e": 109, "5989e": 109, "4763e": 109, "1925e": 109, "0394e": 109, "0835e": 109, "1323e": 109, "8346e": 109, "5475e": 109, "2159e": 109, "3027e": 109, "9814e": 109, "5814e": 109, "9765e": 109, "9968e": 109, "7869e": 109, "9121e": 109, "4343e": 109, "3077e": 109, "3346e": 109, "4254e": 109, "7991e": 109, "7328e": 109, "2397e": 109, "2417e": 109, "9313e": 109, "8547e": 109, "9704e": 109, "3515e": 109, "8294e": 109, "5212e": 109, "2828e": 109, "2438e": 109, "7034e": 109, "7978e": 109, "7903e": 109, "6477e": 109, "9423e": 109, "4541e": 109, "4100e": 109, "6076e": 109, "4080e": 109, "8343e": 109, "6759e": 109, "7489e": 109, "1060e": 109, "7288e": 109, "2695e": 109, "6401e": 109, "9033e": 109, "8954e": 109, "5764e": 109, "8718e": 109, "6939e": 109, "7738e": 109, "4725e": 109, "0905e": 109, "2447e": 109, "2224e": 109, "0585e": 109, "2169e": 109, "7379e": 109, "5061e": 109, "3931e": 109, "1129e": 109, "2449e": 109, "9372e": 109, "8213e": 109, "2947e": 109, "4640e": 109, "8293e": 109, "6663e": 109, "6084e": 109, "0277e": 109, "0561e": 109, "0192e": 109, "8845e": 109, "3221e": 109, "7722e": 109, "9143e": 109, "3545e": 109, "1985e": 109, "5058e": 109, "7636e": 109, "6468e": 109, "9639e": 109, "6412e": 109, "0466e": 109, "7357e": 109, "0067e": 109, "4886e": 109, "4382e": 109, "9157e": 109, "2551e": 109, "0634e": 109, "0727e": 109, "3863e": 109, "0904e": 109, "5949e": 109, "1201e": 109, "2893e": 109, "8121e": 109, "6669e": 109, "6464e": 109, "2743e": 109, "00025": 109, "0444e": 109, "2563e": 109, "7247e": 109, "6934e": 109, "5754e": 109, "4728e": 109, "8390e": 109, "3959e": 109, "3003e": 109, "6432e": 109, "0270e": 109, "5518e": 109, "5609e": 109, "5365e": 109, "5420e": 109, "5809e": 109, "6708e": 109, "0263e": 109, "3368e": 109, "1149e": 109, "5177e": 109, "1763e": 109, "9396e": 109, "8234e": 109, "8248e": 109, "7555e": 109, "8653e": 109, "7842e": 109, "8074e": 109, "3157e": 109, "2374e": 109, "1815e": 109, "0063e": 109, "2248e": 109, "8071e": 109, "0711e": 109, "4018e": 109, "0210e": 109, "8693e": 109, "7399e": 109, "1753e": 109, "7613e": 109, "2838e": 109, "4471e": 109, "4706e": 109, "1696e": 109, "1452e": 109, "7855e": 109, "7885e": 109, "0074e": 109, "4821e": 109, "8941e": 109, "0606e": 109, "9539e": 109, "4716e": 109, "8173e": 109, "6716e": 109, "2783e": 109, "2172e": 109, "2648e": 109, "3846e": 109, "2823e": 109, "1204e": 109, "0944e": 109, "9505e": 109, "8105e": 109, "7888e": 109, "8985e": 109, "1223e": 109, "0778e": 109, "8262e": 109, "2364e": 109, "3126e": 109, "7625e": 109, "7494e": 109, "5998e": 109, "0159e": 109, "9470e": 109, "0446e": 109, "0576e": 109, "1560e": 109, "1218e": 109, "0720e": 109, "2535e": 109, "8031e": 109, "5428e": 109, "6936e": 109, "7461e": 109, "9617e": 109, "2428e": 109, "7234e": 109, "6780e": 109, "5580e": 109, "7324e": 109, "8042e": 109, "9090e": 109, "0420e": 109, "0318e": 109, "0518e": 109, "0386e": 109, "9058e": 109, "9603e": 109, "7566e": 109, "9699e": 109, "3721e": 109, "7064e": 109, "7373e": 109, "2365e": 109, "6765e": 109, "5656e": 109, "4505e": 109, "7352e": 109, "6274e": 109, "0323e": 109, "9917e": 109, "0639e": 109, "0032e": 109, "7938e": 109, "6727e": 109, "7820e": 109, "4437e": 109, "0607e": 109, "6172e": 109, "00032": 109, "6547e": 109, "6407e": 109, "1578e": 109, "6160e": 109, "5752e": 109, "4326e": 109, "5855e": 109, "8317e": 109, "7880e": 109, "8021e": 109, "9743e": 109, "8831e": 109, "9381e": 109, "9257e": 109, "7361e": 109, "0451e": 109, "8240e": 109, "9328e": 109, "3886e": 109, "18m": 109, "4767e": 109, "0575e": 109, "1097e": 109, "4927e": 109, "4112e": 109, "3113e": 109, "1913e": 109, "1998e": 109, "6045e": 109, "6535e": 109, "8091e": 109, "7343e": 109, "5387e": 109, "2035e": 109, "9046e": 109, "6921e": 109, "4087e": 109, "5647e": 109, "4506e": 109, "9790e": 109, "0771e": 109, "5052e": 109, "4143e": 109, "2485e": 109, "1619e": 109, "6584e": 109, "6250e": 109, "6521e": 109, "6717e": 109, "1550e": 109, "8112e": 109, "8689e": 109, "7203e": 109, "3770e": 109, "0190e": 109, "0213e": 109, "5036e": 109, "3911e": 109, "9674e": 109, "0668e": 109, "4330e": 109, "3906e": 109, "2752e": 109, "9657e": 109, "9577e": 109, "5642e": 109, "6610e": 109, "6394e": 109, "1523e": 109, "9620e": 109, "7511e": 109, "5009e": 109, "7908e": 109, "3697e": 109, "5207e": 109, "0507e": 109, "3913e": 109, "3497e": 109, "2511e": 109, "9152e": 109, "0264e": 109, "5519e": 109, "6368e": 109, "5869e": 109, "0233e": 109, "7055e": 109, "8066e": 109, "6591e": 109, "9582e": 109, "2600e": 109, "6224e": 109, "3638e": 109, "6822e": 109, "4107e": 109, "4047e": 109, "2177e": 109, "9564e": 109, "7970e": 109, "5364e": 109, "5089e": 109, "6647e": 109, "6322e": 109, "0198e": 109, "3218e": 109, "3881e": 109, "00037": 109, "31m39": 109, "0m31m48": 109, "0m31m49": 109, "899237": 109, "31m14": 109, "31m270": 109, "691378": 109, "34mmodel": 109, "34mcourtship": 109, "066840": 109, "only_labeled_fram": 109, "91mfals": 109, "only_suggested_fram": 109, "35mnone": 109, "no_empty_fram": 109, "36m1": 109, "36m4": 109, "open_in_gui": 109, "098811": 109, "103255": 109, "103982": 109, "23050": 109, "157392": 109, "158019": 109, "158864": 109, "159656": 109, "455402": 109, "456138": 109, "456803": 109, "457464": 109, "21145": 109, "237m": 109, "36m": 109, "0m2023": 109, "038687": 109, "31m51": 109, "31m126": 109, "25hfinish": 109, "842469": 109, "775644779205322": 109, "sec": 109, "s0": 109, "total_elaps": 109, "36m7": 109, "start_timestamp": 109, "finish_timestamp": 109, "chrome": 109, "deflat": 109, "safari": 109, "icon": 109, "fron": 110, "pckage": 110, "somewher": 110, "unpack": 110, "inter": 110, "chdir": 110, "chose": 110, "anywher": 110, "colab_demo": 110, "predicted_suggest": 110, "demo": 110, "illustr": 112, "recomput": 112, "nutshel": 113, "quantit": 113, "yourself": 113, "footag": 113, "labori": 113, "involv": 113, "among": 113, "satisfactori": 113, "h5read": 114, "tracks_matrix": 114, "feedback": 115, "poorli": 115, "alert": 115, "told": 115, "thicker": 115, "dark": 115, "lighter": 115, "regularli": 115, "proof": 115, "assess": 115, "prev": 116, "scroll": 116, "undo": 116, "cloud": 117, "slowli": 117, "dropdown": 119}, "objects": {"sleap.info": [[3, 0, 0, "-", "align"], [4, 0, 0, "-", "feature_suggestions"], [5, 0, 0, "-", "labels"], [6, 0, 0, "-", "metrics"], [7, 0, 0, "-", "summary"], [8, 0, 0, "-", "trackcleaner"], [9, 0, 0, "-", "write_tracking_h5"]], "sleap.info.align": [[3, 1, 1, "", "align_instance_points"], [3, 1, 1, "", "align_instances"], [3, 1, 1, "", "align_instances_on_most_stable"], [3, 1, 1, "", "get_instances_points"], [3, 1, 1, "", "get_mean_and_std_for_points"], [3, 1, 1, "", "get_most_stable_node_pair"], [3, 1, 1, "", "get_stable_node_pairs"], [3, 1, 1, "", "get_template_points_array"]], "sleap.info.feature_suggestions": [[4, 2, 1, "", "FrameGroupSet"], [4, 2, 1, "", "FrameItem"], [4, 2, 1, "", "ItemStack"], [4, 2, 1, "", "ParallelFeaturePipeline"]], "sleap.info.feature_suggestions.FrameGroupSet": [[4, 3, 1, "", "all_items"], [4, 4, 1, "", "append_to_group"], [4, 4, 1, "", "extend_group_items"], [4, 4, 1, "", "get_item_group"], [4, 5, 1, "", "group_data"], [4, 3, 1, "", "groups"], [4, 5, 1, "", "groupset_data"], [4, 5, 1, "", "item_group"], [4, 5, 1, "", "method"], [4, 4, 1, "", "sample"]], "sleap.info.feature_suggestions.ItemStack": [[4, 4, 1, "", "brisk_bag_of_features"], [4, 3, 1, "", "current_groupset"], [4, 5, 1, "", "data"], [4, 4, 1, "", "extend_ownership"], [4, 4, 1, "", "flatten"], [4, 4, 1, "", "get_all_items_from_group"], [4, 4, 1, "", "get_item_data"], [4, 4, 1, "", "get_item_data_idxs"], [4, 4, 1, "", "get_raw_images"], [4, 5, 1, "", "group_sets"], [4, 4, 1, "", "hog_bag_of_features"], [4, 5, 1, "", "items"], [4, 4, 1, "", "kmeans"], [4, 4, 1, "", "make_sample_group"], [4, 5, 1, "", "meta"], [4, 5, 1, "", "ownership"], [4, 4, 1, "", "pca"], [4, 4, 1, "", "sample_groups"]], "sleap.info.feature_suggestions.ParallelFeaturePipeline": [[4, 4, 1, "", "get"], [4, 4, 1, "", "make"], [4, 4, 1, "", "run"], [4, 4, 1, "", "tuples_to_suggestions"]], "sleap.info.metrics": [[6, 1, 1, "", "calculate_pairwise_cost"], [6, 1, 1, "", "compare_instance_lists"], [6, 1, 1, "", "list_points_array"], [6, 1, 1, "", "match_instance_lists"], [6, 1, 1, "", "match_instance_lists_nodewise"], [6, 1, 1, "", "matched_instance_distances"], [6, 1, 1, "", "nodeless_point_dist"], [6, 1, 1, "", "point_dist"], [6, 1, 1, "", "point_match_count"], [6, 1, 1, "", "point_nonmatch_count"]], "sleap.info.summary": [[7, 2, 1, "", "StatisticSeries"]], "sleap.info.summary.StatisticSeries": [[7, 4, 1, "", "get_instance_score_series"], [7, 4, 1, "", "get_point_count_series"], [7, 4, 1, "", "get_point_displacement_series"], [7, 4, 1, "", "get_point_score_series"], [7, 4, 1, "", "get_primary_point_displacement_series"], [7, 4, 1, "", "get_tracking_score_series"]], "sleap.info.trackcleaner": [[8, 1, 1, "", "fit_tracks"]], "sleap.info.write_tracking_h5": [[9, 1, 1, "", "get_edges_as_np_strings"], [9, 1, 1, "", "get_nodes_as_np_strings"], [9, 1, 1, "", "get_occupancy_and_points_matrices"], [9, 1, 1, "", "get_tracks_as_np_strings"], [9, 1, 1, "", "main"], [9, 1, 1, "", "remove_empty_tracks_from_matrices"], [9, 1, 1, "", "write_csv_file"], [9, 1, 1, "", "write_occupancy_file"]], "sleap": [[10, 0, 0, "-", "instance"], [35, 0, 0, "-", "message"], [84, 0, 0, "-", "skeleton"], [85, 0, 0, "-", "util"]], "sleap.instance": [[10, 2, 1, "", "Instance"], [10, 2, 1, "", "InstancesList"], [10, 2, 1, "", "LabeledFrame"], [10, 2, 1, "", "Point"], [10, 2, 1, "", "PointArray"], [10, 2, 1, "", "PredictedInstance"], [10, 2, 1, "", "PredictedPoint"], [10, 2, 1, "", "PredictedPointArray"], [10, 2, 1, "", "Track"], [10, 1, 1, "", "make_instance_cattr"]], "sleap.instance.Instance": [[10, 3, 1, "", "bounding_box"], [10, 3, 1, "", "centroid"], [10, 4, 1, "", "fill_missing"], [10, 3, 1, "", "frame_idx"], [10, 4, 1, "", "from_numpy"], [10, 4, 1, "", "from_pointsarray"], [10, 4, 1, "", "get_points_array"], [10, 4, 1, "", "matches"], [10, 3, 1, "", "midpoint"], [10, 3, 1, "", "n_visible_points"], [10, 3, 1, "", "nodes"], [10, 3, 1, "", "nodes_points"], [10, 4, 1, "", "numpy"], [10, 3, 1, "", "points"], [10, 3, 1, "", "points_array"], [10, 4, 1, "", "transform_points"], [10, 3, 1, "", "video"]], "sleap.instance.InstancesList": [[10, 4, 1, "", "append"], [10, 4, 1, "", "clear"], [10, 4, 1, "", "copy"], [10, 4, 1, "", "extend"], [10, 4, 1, "", "insert"], [10, 3, 1, "", "labeled_frame"], [10, 4, 1, "", "pop"], [10, 4, 1, "", "remove"]], "sleap.instance.LabeledFrame": [[10, 4, 1, "", "complex_frame_merge"], [10, 4, 1, "", "complex_merge_between"], [10, 4, 1, "", "find"], [10, 3, 1, "", "has_predicted_instances"], [10, 3, 1, "", "has_tracked_instances"], [10, 3, 1, "", "has_user_instances"], [10, 3, 1, "", "image"], [10, 4, 1, "", "index"], [10, 4, 1, "", "insert"], [10, 3, 1, "", "instances"], [10, 3, 1, "", "instances_to_show"], [10, 4, 1, "", "merge_frames"], [10, 3, 1, "", "n_predicted_instances"], [10, 3, 1, "", "n_tracked_instances"], [10, 3, 1, "", "n_user_instances"], [10, 4, 1, "", "numpy"], [10, 4, 1, "", "plot"], [10, 4, 1, "", "plot_predicted"], [10, 3, 1, "", "predicted_instances"], [10, 4, 1, "", "remove_empty_instances"], [10, 4, 1, "", "remove_untracked"], [10, 3, 1, "", "tracked_instances"], [10, 3, 1, "", "training_instances"], [10, 3, 1, "", "unused_predictions"], [10, 3, 1, "", "user_instances"]], "sleap.instance.Point": [[10, 5, 1, "", "dtype"], [10, 4, 1, "", "isnan"], [10, 4, 1, "", "numpy"]], "sleap.instance.PointArray": [[10, 4, 1, "", "from_array"], [10, 4, 1, "", "make_default"]], "sleap.instance.PredictedInstance": [[10, 4, 1, "", "from_arrays"], [10, 4, 1, "", "from_instance"], [10, 4, 1, "", "from_numpy"], [10, 4, 1, "", "from_pointsarray"], [10, 3, 1, "", "points_and_scores_array"], [10, 3, 1, "", "scores"]], "sleap.instance.PredictedPoint": [[10, 5, 1, "", "dtype"], [10, 4, 1, "", "from_point"]], "sleap.instance.PredictedPointArray": [[10, 4, 1, "", "to_array"]], "sleap.instance.Track": [[10, 4, 1, "", "matches"]], "sleap.io": [[11, 0, 0, "-", "convert"], [12, 0, 0, "-", "dataset"], [30, 0, 0, "-", "legacy"], [31, 0, 0, "-", "pathutils"], [32, 0, 0, "-", "video"], [33, 0, 0, "-", "videowriter"], [34, 0, 0, "-", "visuals"]], "sleap.io.convert": [[11, 1, 1, "", "main"]], "sleap.io.dataset": [[12, 2, 1, "", "Labels"], [12, 2, 1, "", "LabelsDataCache"], [12, 1, 1, "", "find_path_using_paths"], [12, 1, 1, "", "load_file"]], "sleap.io.dataset.Labels": [[12, 4, 1, "", "add_instance"], [12, 4, 1, "", "add_suggestion"], [12, 4, 1, "", "add_track"], [12, 4, 1, "", "add_video"], [12, 3, 1, "", "all_instances"], [12, 4, 1, "", "append"], [12, 4, 1, "", "append_suggestions"], [12, 4, 1, "", "clear_suggestions"], [12, 4, 1, "", "complex_merge_between"], [12, 4, 1, "", "copy"], [12, 4, 1, "", "delete_suggestions"], [12, 4, 1, "", "describe"], [12, 4, 1, "", "export"], [12, 4, 1, "", "export_csv"], [12, 4, 1, "", "export_nwb"], [12, 4, 1, "", "extend_from"], [12, 4, 1, "", "extract"], [12, 4, 1, "", "find"], [12, 4, 1, "", "find_first"], [12, 4, 1, "", "find_last"], [12, 4, 1, "", "find_suggestion"], [12, 4, 1, "", "find_track_occupancy"], [12, 4, 1, "", "finish_complex_merge"], [12, 4, 1, "", "frames"], [12, 4, 1, "", "get"], [12, 4, 1, "", "get_next_suggestion"], [12, 4, 1, "", "get_suggestions"], [12, 4, 1, "", "get_track_count"], [12, 4, 1, "", "get_track_occupancy"], [12, 4, 1, "", "get_unlabeled_suggestion_inds"], [12, 4, 1, "", "get_video_suggestions"], [12, 4, 1, "", "has_frame"], [12, 3, 1, "", "has_missing_videos"], [12, 4, 1, "", "index"], [12, 4, 1, "", "insert"], [12, 4, 1, "", "instance_count"], [12, 4, 1, "", "instances"], [12, 3, 1, "", "is_multi_instance"], [12, 5, 1, "", "labeled_frames"], [12, 3, 1, "", "labels"], [12, 4, 1, "", "load_file"], [12, 4, 1, "", "make_video_callback"], [12, 4, 1, "", "merge_container_dicts"], [12, 4, 1, "", "merge_matching_frames"], [12, 4, 1, "", "merge_nodes"], [12, 5, 1, "", "negative_anchors"], [12, 4, 1, "", "numpy"], [12, 3, 1, "", "predicted_instances"], [12, 5, 1, "", "provenance"], [12, 4, 1, "", "remove"], [12, 4, 1, "", "remove_all_tracks"], [12, 4, 1, "", "remove_empty_frames"], [12, 4, 1, "", "remove_empty_instances"], [12, 4, 1, "", "remove_frame"], [12, 4, 1, "", "remove_frames"], [12, 4, 1, "", "remove_instance"], [12, 4, 1, "", "remove_predictions"], [12, 4, 1, "", "remove_suggestion"], [12, 4, 1, "", "remove_track"], [12, 4, 1, "", "remove_untracked_instances"], [12, 4, 1, "", "remove_unused_tracks"], [12, 4, 1, "", "remove_user_instances"], [12, 4, 1, "", "remove_video"], [12, 4, 1, "", "save"], [12, 4, 1, "", "save_file"], [12, 4, 1, "", "save_frame_data_hdf5"], [12, 4, 1, "", "save_frame_data_imgstore"], [12, 4, 1, "", "set_suggestions"], [12, 3, 1, "", "skeleton"], [12, 5, 1, "", "skeletons"], [12, 4, 1, "", "split"], [12, 5, 1, "", "suggestions"], [12, 4, 1, "", "to_dict"], [12, 4, 1, "", "to_json"], [12, 4, 1, "", "to_pipeline"], [12, 4, 1, "", "track_set_instance"], [12, 4, 1, "", "track_swap"], [12, 5, 1, "", "tracks"], [12, 3, 1, "", "unlabeled_suggestions"], [12, 3, 1, "", "user_instances"], [12, 3, 1, "", "user_labeled_frame_inds"], [12, 3, 1, "", "user_labeled_frames"], [12, 3, 1, "", "video"], [12, 5, 1, "", "videos"], [12, 4, 1, "", "with_user_labels_only"]], "sleap.io.dataset.LabelsDataCache": [[12, 4, 1, "", "add_instance"], [12, 4, 1, "", "add_track"], [12, 4, 1, "", "find_fancy_frame_idxs"], [12, 4, 1, "", "find_frames"], [12, 4, 1, "", "get_filtered_frame_idxs"], [12, 4, 1, "", "get_frame_count"], [12, 4, 1, "", "get_track_occupancy"], [12, 4, 1, "", "get_video_track_occupancy"], [12, 4, 1, "", "remove_frame"], [12, 4, 1, "", "remove_instance"], [12, 4, 1, "", "remove_video"], [12, 4, 1, "", "track_swap"], [12, 4, 1, "", "update"], [12, 4, 1, "", "update_counts_for_frame"]], "sleap.io.format": [[13, 0, 0, "-", "adaptor"], [14, 0, 0, "-", "alphatracker"], [15, 0, 0, "-", "coco"], [16, 0, 0, "-", "csv"], [17, 0, 0, "-", "deeplabcut"], [18, 0, 0, "-", "deepposekit"], [19, 0, 0, "-", "dispatch"], [20, 0, 0, "-", "filehandle"], [21, 0, 0, "-", "genericjson"], [22, 0, 0, "-", "hdf5"], [23, 0, 0, "-", "labels_json"], [24, 0, 0, "-", "leap_matlab"], [25, 0, 0, "-", "main"], [26, 0, 0, "-", "ndx_pose"], [27, 0, 0, "-", "nix"], [28, 0, 0, "-", "sleap_analysis"], [29, 0, 0, "-", "text"]], "sleap.io.format.adaptor": [[13, 2, 1, "", "Adaptor"], [13, 2, 1, "", "SleapObjectType"]], "sleap.io.format.adaptor.Adaptor": [[13, 3, 1, "", "all_exts"], [13, 4, 1, "", "can_read_file"], [13, 4, 1, "", "can_write_filename"], [13, 3, 1, "", "default_ext"], [13, 4, 1, "", "does_match_ext"], [13, 4, 1, "", "does_read"], [13, 4, 1, "", "does_write"], [13, 3, 1, "", "formatted_ext_options"], [13, 3, 1, "", "handles"], [13, 3, 1, "", "name"], [13, 4, 1, "", "read"], [13, 4, 1, "", "write"]], "sleap.io.format.alphatracker": [[14, 2, 1, "", "AlphaTrackerAdaptor"]], "sleap.io.format.alphatracker.AlphaTrackerAdaptor": [[14, 3, 1, "", "all_exts"], [14, 4, 1, "", "can_read_file"], [14, 4, 1, "", "can_write_filename"], [14, 3, 1, "", "default_ext"], [14, 4, 1, "", "does_match_ext"], [14, 4, 1, "", "does_read"], [14, 4, 1, "", "does_write"], [14, 3, 1, "", "formatted_ext_options"], [14, 4, 1, "", "get_alpha_tracker_frame_dict"], [14, 4, 1, "", "get_alpha_tracker_instance_dict"], [14, 4, 1, "", "get_alpha_tracker_point_dict"], [14, 3, 1, "", "handles"], [14, 4, 1, "", "make_video_for_image_list"], [14, 3, 1, "", "name"], [14, 4, 1, "", "read"], [14, 4, 1, "", "write"]], "sleap.io.format.coco": [[15, 2, 1, "", "LabelsCocoAdaptor"]], "sleap.io.format.coco.LabelsCocoAdaptor": [[15, 3, 1, "", "all_exts"], [15, 4, 1, "", "can_read_file"], [15, 4, 1, "", "can_write_filename"], [15, 3, 1, "", "default_ext"], [15, 4, 1, "", "does_read"], [15, 4, 1, "", "does_write"], [15, 3, 1, "", "handles"], [15, 3, 1, "", "name"], [15, 4, 1, "", "read"]], "sleap.io.format.csv": [[16, 2, 1, "", "CSVAdaptor"]], "sleap.io.format.csv.CSVAdaptor": [[16, 3, 1, "", "all_exts"], [16, 4, 1, "", "can_read_file"], [16, 4, 1, "", "can_write_filename"], [16, 3, 1, "", "default_ext"], [16, 4, 1, "", "does_read"], [16, 4, 1, "", "does_write"], [16, 3, 1, "", "handles"], [16, 3, 1, "", "name"], [16, 4, 1, "", "write"]], "sleap.io.format.deeplabcut": [[17, 2, 1, "", "LabelsDeepLabCutCsvAdaptor"], [17, 2, 1, "", "LabelsDeepLabCutYamlAdaptor"]], "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor": [[17, 3, 1, "", "all_exts"], [17, 4, 1, "", "can_read_file"], [17, 4, 1, "", "can_write_filename"], [17, 3, 1, "", "default_ext"], [17, 4, 1, "", "does_read"], [17, 4, 1, "", "does_write"], [17, 3, 1, "", "handles"], [17, 4, 1, "", "make_video_for_image_list"], [17, 3, 1, "", "name"], [17, 4, 1, "", "read"]], "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor": [[17, 3, 1, "", "all_exts"], [17, 4, 1, "", "can_read_file"], [17, 4, 1, "", "can_write_filename"], [17, 3, 1, "", "default_ext"], [17, 4, 1, "", "does_read"], [17, 4, 1, "", "does_write"], [17, 3, 1, "", "handles"], [17, 3, 1, "", "name"], [17, 4, 1, "", "read"]], "sleap.io.format.deepposekit": [[18, 2, 1, "", "LabelsDeepPoseKitAdaptor"]], "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor": [[18, 3, 1, "", "all_exts"], [18, 4, 1, "", "can_read_file"], [18, 4, 1, "", "can_write_filename"], [18, 3, 1, "", "default_ext"], [18, 4, 1, "", "does_read"], [18, 4, 1, "", "does_write"], [18, 3, 1, "", "handles"], [18, 3, 1, "", "name"], [18, 4, 1, "", "read"]], "sleap.io.format.filehandle": [[20, 2, 1, "", "FileHandle"]], "sleap.io.format.filehandle.FileHandle": [[20, 4, 1, "", "close"], [20, 3, 1, "", "file"], [20, 3, 1, "", "format_id"], [20, 3, 1, "", "is_hdf5"], [20, 3, 1, "", "is_json"], [20, 3, 1, "", "json"], [20, 4, 1, "", "open"], [20, 3, 1, "", "text"]], "sleap.io.format.genericjson": [[21, 2, 1, "", "GenericJsonAdaptor"]], "sleap.io.format.genericjson.GenericJsonAdaptor": [[21, 3, 1, "", "all_exts"], [21, 4, 1, "", "can_read_file"], [21, 4, 1, "", "can_write_filename"], [21, 3, 1, "", "default_ext"], [21, 4, 1, "", "does_read"], [21, 4, 1, "", "does_write"], [21, 3, 1, "", "handles"], [21, 3, 1, "", "name"], [21, 4, 1, "", "read"], [21, 4, 1, "", "write"]], "sleap.io.format.hdf5": [[22, 2, 1, "", "LabelsV1Adaptor"]], "sleap.io.format.hdf5.LabelsV1Adaptor": [[22, 3, 1, "", "all_exts"], [22, 4, 1, "", "can_read_file"], [22, 4, 1, "", "can_write_filename"], [22, 3, 1, "", "default_ext"], [22, 4, 1, "", "does_read"], [22, 4, 1, "", "does_write"], [22, 3, 1, "", "handles"], [22, 3, 1, "", "name"], [22, 4, 1, "", "read"], [22, 4, 1, "", "write"]], "sleap.io.format.labels_json": [[23, 2, 1, "", "LabelsJsonAdaptor"]], "sleap.io.format.labels_json.LabelsJsonAdaptor": [[23, 3, 1, "", "all_exts"], [23, 4, 1, "", "can_read_file"], [23, 4, 1, "", "can_write_filename"], [23, 3, 1, "", "default_ext"], [23, 4, 1, "", "does_read"], [23, 4, 1, "", "does_write"], [23, 4, 1, "", "from_json_data"], [23, 3, 1, "", "handles"], [23, 3, 1, "", "name"], [23, 4, 1, "", "read"], [23, 4, 1, "", "write"]], "sleap.io.format.leap_matlab": [[24, 2, 1, "", "LabelsLeapMatlabAdaptor"]], "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor": [[24, 3, 1, "", "all_exts"], [24, 4, 1, "", "can_read_file"], [24, 4, 1, "", "can_write_filename"], [24, 3, 1, "", "default_ext"], [24, 4, 1, "", "does_read"], [24, 4, 1, "", "does_write"], [24, 3, 1, "", "handles"], [24, 3, 1, "", "name"], [24, 4, 1, "", "read"]], "sleap.io.format.main": [[25, 1, 1, "", "read"], [25, 1, 1, "", "write"]], "sleap.io.format.ndx_pose": [[26, 2, 1, "", "NDXPoseAdaptor"]], "sleap.io.format.ndx_pose.NDXPoseAdaptor": [[26, 3, 1, "", "all_exts"], [26, 4, 1, "", "can_read_file"], [26, 4, 1, "", "can_write_filename"], [26, 3, 1, "", "default_ext"], [26, 4, 1, "", "does_read"], [26, 4, 1, "", "does_write"], [26, 3, 1, "", "handles"], [26, 3, 1, "", "name"], [26, 4, 1, "", "read"], [26, 4, 1, "", "write"]], "sleap.io.format.nix": [[27, 2, 1, "", "NixAdaptor"]], "sleap.io.format.nix.NixAdaptor": [[27, 3, 1, "", "all_exts"], [27, 4, 1, "", "can_read_file"], [27, 4, 1, "", "can_write_filename"], [27, 3, 1, "", "default_ext"], [27, 4, 1, "", "does_read"], [27, 4, 1, "", "does_write"], [27, 3, 1, "", "handles"], [27, 3, 1, "", "name"], [27, 4, 1, "", "read"], [27, 4, 1, "", "write"]], "sleap.io.format.sleap_analysis": [[28, 2, 1, "", "SleapAnalysisAdaptor"]], "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor": [[28, 3, 1, "", "all_exts"], [28, 4, 1, "", "can_read_file"], [28, 4, 1, "", "can_write_filename"], [28, 3, 1, "", "default_ext"], [28, 4, 1, "", "does_read"], [28, 4, 1, "", "does_write"], [28, 3, 1, "", "handles"], [28, 3, 1, "", "name"], [28, 4, 1, "", "read"], [28, 4, 1, "", "write"]], "sleap.io.format.text": [[29, 2, 1, "", "TextAdaptor"]], "sleap.io.format.text.TextAdaptor": [[29, 3, 1, "", "all_exts"], [29, 4, 1, "", "can_read_file"], [29, 4, 1, "", "can_write_filename"], [29, 3, 1, "", "default_ext"], [29, 4, 1, "", "does_read"], [29, 4, 1, "", "does_write"], [29, 3, 1, "", "handles"], [29, 3, 1, "", "name"], [29, 4, 1, "", "read"], [29, 4, 1, "", "write"]], "sleap.io.legacy": [[30, 1, 1, "", "load_labels_json_old"], [30, 1, 1, "", "load_predicted_labels_json_old"]], "sleap.io.pathutils": [[31, 1, 1, "", "filenames_prefix_change"], [31, 1, 1, "", "find_changed_subpath"], [31, 1, 1, "", "list_file_missing"]], "sleap.io.video": [[32, 2, 1, "", "DummyVideo"], [32, 2, 1, "", "HDF5Video"], [32, 2, 1, "", "ImgStoreVideo"], [32, 2, 1, "", "MediaVideo"], [32, 2, 1, "", "NumpyVideo"], [32, 2, 1, "", "SingleImageVideo"], [32, 2, 1, "", "Video"], [32, 1, 1, "", "available_video_exts"], [32, 1, 1, "", "load_video"]], "sleap.io.video.HDF5Video": [[32, 3, 1, "", "channels"], [32, 4, 1, "", "check"], [32, 4, 1, "", "close"], [32, 3, 1, "", "dtype"], [32, 3, 1, "", "embedded_frame_inds"], [32, 3, 1, "", "enable_source_video"], [32, 3, 1, "", "frames"], [32, 4, 1, "", "get_frame"], [32, 3, 1, "", "has_embedded_images"], [32, 3, 1, "", "height"], [32, 3, 1, "", "last_frame_idx"], [32, 4, 1, "", "matches"], [32, 4, 1, "", "reset"], [32, 3, 1, "", "source_video"], [32, 3, 1, "", "source_video_available"], [32, 3, 1, "", "width"]], "sleap.io.video.ImgStoreVideo": [[32, 3, 1, "", "channels"], [32, 4, 1, "", "close"], [32, 3, 1, "", "dtype"], [32, 3, 1, "", "frames"], [32, 4, 1, "", "get_frame"], [32, 3, 1, "", "height"], [32, 3, 1, "", "imgstore"], [32, 3, 1, "", "last_frame_idx"], [32, 4, 1, "", "matches"], [32, 4, 1, "", "open"], [32, 4, 1, "", "reset"], [32, 3, 1, "", "width"]], "sleap.io.video.MediaVideo": [[32, 3, 1, "", "channels"], [32, 3, 1, "", "dtype"], [32, 3, 1, "", "fps"], [32, 3, 1, "", "frames"], [32, 4, 1, "", "get_frame"], [32, 3, 1, "", "height"], [32, 4, 1, "", "matches"], [32, 4, 1, "", "reset"], [32, 3, 1, "", "width"]], "sleap.io.video.NumpyVideo": [[32, 3, 1, "", "channels"], [32, 3, 1, "", "dtype"], [32, 3, 1, "", "frames"], [32, 4, 1, "", "get_frame"], [32, 3, 1, "", "height"], [32, 3, 1, "", "is_missing"], [32, 4, 1, "", "matches"], [32, 4, 1, "", "reset"], [32, 3, 1, "", "width"]], "sleap.io.video.SingleImageVideo": [[32, 3, 1, "", "channels"], [32, 3, 1, "", "dtype"], [32, 3, 1, "", "frames"], [32, 4, 1, "", "get_frame"], [32, 3, 1, "", "height"], [32, 4, 1, "", "matches"], [32, 4, 1, "", "reset"], [32, 3, 1, "", "width"]], "sleap.io.video.Video": [[32, 4, 1, "", "cattr"], [32, 4, 1, "", "fixup_path"], [32, 4, 1, "", "from_filename"], [32, 4, 1, "", "from_hdf5"], [32, 4, 1, "", "from_image_filenames"], [32, 4, 1, "", "from_media"], [32, 4, 1, "", "from_numpy"], [32, 4, 1, "", "get_frame"], [32, 4, 1, "", "get_frames"], [32, 4, 1, "", "get_frames_safely"], [32, 4, 1, "", "imgstore_from_filenames"], [32, 3, 1, "", "is_missing"], [32, 3, 1, "", "last_frame_idx"], [32, 3, 1, "", "num_frames"], [32, 3, 1, "", "shape"], [32, 4, 1, "", "to_hdf5"], [32, 4, 1, "", "to_imgstore"], [32, 4, 1, "", "to_pipeline"]], "sleap.io.videowriter": [[33, 2, 1, "", "VideoWriter"], [33, 2, 1, "", "VideoWriterImageio"], [33, 2, 1, "", "VideoWriterOpenCV"]], "sleap.io.videowriter.VideoWriter": [[33, 4, 1, "", "can_use_ffmpeg"], [33, 4, 1, "", "safe_builder"]], "sleap.io.videowriter.VideoWriterImageio": [[33, 5, 1, "", "crf"], [33, 5, 1, "", "filename"], [33, 5, 1, "", "fps"], [33, 5, 1, "", "height"], [33, 5, 1, "", "preset"], [33, 5, 1, "", "width"]], "sleap.io.visuals": [[34, 2, 1, "", "VideoMarkerThread"], [34, 1, 1, "", "img_to_cv"], [34, 1, 1, "", "reader"], [34, 1, 1, "", "resize_image"], [34, 1, 1, "", "save_labeled_video"], [34, 1, 1, "", "writer"]], "sleap.io.visuals.VideoMarkerThread": [[34, 4, 1, "", "run"]], "sleap.message": [[35, 2, 1, "", "BaseMessageParticipant"], [35, 2, 1, "", "Receiver"], [35, 2, 1, "", "Sender"]], "sleap.message.Receiver": [[35, 4, 1, "", "check_message"], [35, 4, 1, "", "check_messages"], [35, 4, 1, "", "push_back_message"]], "sleap.message.Sender": [[35, 4, 1, "", "send_array"], [35, 4, 1, "", "send_dict"]], "sleap.nn.architectures": [[36, 0, 0, "-", "common"], [37, 0, 0, "-", "encoder_decoder"], [38, 0, 0, "-", "hourglass"], [39, 0, 0, "-", "hrnet"], [40, 0, 0, "-", "leap"], [41, 0, 0, "-", "pretrained_encoders"], [42, 0, 0, "-", "resnet"], [43, 0, 0, "-", "unet"], [44, 0, 0, "-", "upsampling"]], "sleap.nn.architectures.common": [[36, 2, 1, "", "IntermediateFeature"]], "sleap.nn.architectures.common.IntermediateFeature": [[36, 3, 1, "", "scale"], [36, 5, 1, "", "stride"], [36, 5, 1, "", "tensor"]], "sleap.nn.architectures.encoder_decoder": [[37, 2, 1, "", "DecoderBlock"], [37, 2, 1, "", "EncoderBlock"], [37, 2, 1, "", "EncoderDecoder"], [37, 2, 1, "", "SimpleConvBlock"], [37, 2, 1, "", "SimpleUpsamplingBlock"]], "sleap.nn.architectures.encoder_decoder.DecoderBlock": [[37, 4, 1, "", "make_block"], [37, 5, 1, "", "upsampling_stride"]], "sleap.nn.architectures.encoder_decoder.EncoderBlock": [[37, 4, 1, "", "make_block"], [37, 5, 1, "", "pool"], [37, 5, 1, "", "pooling_stride"]], "sleap.nn.architectures.encoder_decoder.EncoderDecoder": [[37, 3, 1, "", "decoder_features_stride"], [37, 3, 1, "", "decoder_stack"], [37, 3, 1, "", "encoder_features_stride"], [37, 3, 1, "", "encoder_stack"], [37, 4, 1, "", "make_backbone"], [37, 4, 1, "", "make_decoder"], [37, 4, 1, "", "make_encoder"], [37, 4, 1, "", "make_stem"], [37, 3, 1, "", "maximum_stride"], [37, 3, 1, "", "output_stride"], [37, 5, 1, "", "stacks"], [37, 3, 1, "", "stem_features_stride"], [37, 3, 1, "", "stem_stack"]], "sleap.nn.architectures.encoder_decoder.SimpleConvBlock": [[37, 5, 1, "", "activation"], [37, 5, 1, "", "batch_norm"], [37, 5, 1, "", "batch_norm_before_activation"], [37, 5, 1, "", "block_prefix"], [37, 5, 1, "", "filters"], [37, 5, 1, "", "kernel_size"], [37, 4, 1, "", "make_block"], [37, 5, 1, "", "num_convs"], [37, 5, 1, "", "pool"], [37, 5, 1, "", "pool_before_convs"], [37, 5, 1, "", "pooling_stride"], [37, 5, 1, "", "use_bias"]], "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock": [[37, 5, 1, "", "interp_method"], [37, 4, 1, "", "make_block"], [37, 5, 1, "", "refine_convs"], [37, 5, 1, "", "refine_convs_activation"], [37, 5, 1, "", "refine_convs_batch_norm"], [37, 5, 1, "", "refine_convs_batch_norm_before_activation"], [37, 5, 1, "", "refine_convs_filters"], [37, 5, 1, "", "refine_convs_first_filters"], [37, 5, 1, "", "refine_convs_kernel_size"], [37, 5, 1, "", "refine_convs_use_bias"], [37, 5, 1, "", "skip_add"], [37, 5, 1, "", "skip_connection"], [37, 5, 1, "", "transposed_conv"], [37, 5, 1, "", "transposed_conv_activation"], [37, 5, 1, "", "transposed_conv_batch_norm"], [37, 5, 1, "", "transposed_conv_batch_norm_before_activation"], [37, 5, 1, "", "transposed_conv_filters"], [37, 5, 1, "", "transposed_conv_kernel_size"], [37, 5, 1, "", "transposed_conv_use_bias"], [37, 5, 1, "", "upsampling_stride"]], "sleap.nn.architectures.hourglass": [[38, 2, 1, "", "DownsamplingBlock"], [38, 2, 1, "", "Hourglass"], [38, 2, 1, "", "StemBlock"], [38, 2, 1, "", "UpsamplingBlock"], [38, 1, 1, "", "conv"]], "sleap.nn.architectures.hourglass.DownsamplingBlock": [[38, 5, 1, "", "filters"], [38, 4, 1, "", "make_block"]], "sleap.nn.architectures.hourglass.Hourglass": [[38, 3, 1, "", "decoder_stack"], [38, 5, 1, "", "down_blocks"], [38, 3, 1, "", "encoder_stack"], [38, 5, 1, "", "filter_increase"], [38, 5, 1, "", "filters"], [38, 4, 1, "", "from_config"], [38, 5, 1, "", "interp_method"], [38, 5, 1, "", "stacks"], [38, 5, 1, "", "stem_filters"], [38, 3, 1, "", "stem_stack"], [38, 5, 1, "", "stem_stride"], [38, 5, 1, "", "up_blocks"]], "sleap.nn.architectures.hourglass.StemBlock": [[38, 5, 1, "", "filters"], [38, 5, 1, "", "first_conv_stride"], [38, 4, 1, "", "make_block"], [38, 5, 1, "", "output_filters"], [38, 5, 1, "", "pool"], [38, 5, 1, "", "pooling_stride"]], "sleap.nn.architectures.hourglass.UpsamplingBlock": [[38, 5, 1, "", "filters"], [38, 5, 1, "", "interp_method"], [38, 4, 1, "", "make_block"]], "sleap.nn.architectures.hrnet": [[39, 2, 1, "", "HigherHRNet"], [39, 1, 1, "", "adjust_prefix"], [39, 1, 1, "", "bottleneck_block"], [39, 1, 1, "", "simple_block"]], "sleap.nn.architectures.hrnet.HigherHRNet": [[39, 5, 1, "", "C"], [39, 5, 1, "", "bilinear_upsampling"], [39, 5, 1, "", "bottleneck"], [39, 5, 1, "", "deconv_filters"], [39, 3, 1, "", "down_blocks"], [39, 5, 1, "", "initial_downsampling_steps"], [39, 5, 1, "", "n_deconv_modules"], [39, 4, 1, "", "output"], [39, 3, 1, "", "output_scale"]], "sleap.nn.architectures.leap": [[40, 2, 1, "", "LeapCNN"]], "sleap.nn.architectures.leap.LeapCNN": [[40, 3, 1, "", "decoder_stack"], [40, 5, 1, "", "down_blocks"], [40, 5, 1, "", "down_convs_per_block"], [40, 3, 1, "", "encoder_stack"], [40, 5, 1, "", "filters"], [40, 5, 1, "", "filters_rate"], [40, 4, 1, "", "from_config"], [40, 5, 1, "", "up_blocks"], [40, 5, 1, "", "up_convs_per_block"], [40, 5, 1, "", "up_interpolate"]], "sleap.nn.architectures.pretrained_encoders": [[41, 2, 1, "", "UnetPretrainedEncoder"]], "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder": [[41, 5, 1, "", "decoder_filters"], [41, 3, 1, "", "down_blocks"], [41, 5, 1, "", "encoder"], [41, 4, 1, "", "from_config"], [41, 4, 1, "", "make_backbone"], [41, 3, 1, "", "maximum_stride"], [41, 3, 1, "", "output_stride"], [41, 5, 1, "", "pretrained"], [41, 3, 1, "", "up_blocks"]], "sleap.nn.architectures.resnet": [[42, 2, 1, "", "ResNet101"], [42, 2, 1, "", "ResNet152"], [42, 2, 1, "", "ResNet50"], [42, 2, 1, "", "ResNetv1"], [42, 1, 1, "", "block_v1"], [42, 1, 1, "", "imagenet_preproc_v1"], [42, 1, 1, "", "make_backbone_fn"], [42, 1, 1, "", "make_resnet_model"], [42, 1, 1, "", "stack_v1"], [42, 1, 1, "", "tile_channels"]], "sleap.nn.architectures.resnet.ResNet101": [[42, 5, 1, "", "features_output_stride"], [42, 5, 1, "", "frozen"], [42, 5, 1, "", "pretrained"], [42, 5, 1, "", "skip_connections"], [42, 5, 1, "", "upsampling_stack"]], "sleap.nn.architectures.resnet.ResNet152": [[42, 5, 1, "", "features_output_stride"], [42, 5, 1, "", "frozen"], [42, 5, 1, "", "pretrained"], [42, 5, 1, "", "skip_connections"], [42, 5, 1, "", "upsampling_stack"]], "sleap.nn.architectures.resnet.ResNet50": [[42, 5, 1, "", "features_output_stride"], [42, 5, 1, "", "frozen"], [42, 5, 1, "", "pretrained"], [42, 5, 1, "", "skip_connections"], [42, 5, 1, "", "upsampling_stack"]], "sleap.nn.architectures.resnet.ResNetv1": [[42, 3, 1, "", "down_blocks"], [42, 5, 1, "", "features_output_stride"], [42, 4, 1, "", "from_config"], [42, 5, 1, "", "frozen"], [42, 4, 1, "", "make_backbone"], [42, 3, 1, "", "maximum_stride"], [42, 5, 1, "", "model_name"], [42, 3, 1, "", "output_scale"], [42, 3, 1, "", "output_stride"], [42, 5, 1, "", "pretrained"], [42, 5, 1, "", "skip_connections"], [42, 5, 1, "", "stack_configs"], [42, 5, 1, "", "upsampling_stack"]], "sleap.nn.architectures.unet": [[43, 2, 1, "", "PoolingBlock"], [43, 2, 1, "", "UNet"]], "sleap.nn.architectures.unet.PoolingBlock": [[43, 4, 1, "", "make_block"], [43, 5, 1, "", "pool"], [43, 5, 1, "", "pooling_stride"]], "sleap.nn.architectures.unet.UNet": [[43, 5, 1, "", "block_contraction"], [43, 5, 1, "", "convs_per_block"], [43, 3, 1, "", "decoder_stack"], [43, 5, 1, "", "down_blocks"], [43, 3, 1, "", "encoder_stack"], [43, 5, 1, "", "filters"], [43, 5, 1, "", "filters_rate"], [43, 4, 1, "", "from_config"], [43, 5, 1, "", "kernel_size"], [43, 5, 1, "", "middle_block"], [43, 5, 1, "", "stem_blocks"], [43, 5, 1, "", "stem_kernel_size"], [43, 3, 1, "", "stem_stack"], [43, 5, 1, "", "up_blocks"], [43, 5, 1, "", "up_interpolate"]], "sleap.nn.architectures.upsampling": [[44, 2, 1, "", "UpsamplingStack"]], "sleap.nn.architectures.upsampling.UpsamplingStack": [[44, 4, 1, "", "from_config"], [44, 5, 1, "", "make_skip_connection"], [44, 4, 1, "", "make_stack"], [44, 5, 1, "", "output_stride"], [44, 5, 1, "", "refine_convs"], [44, 5, 1, "", "refine_convs_batchnorm"], [44, 5, 1, "", "refine_convs_filters"], [44, 5, 1, "", "refine_convs_filters_rate"], [44, 5, 1, "", "skip_add"], [44, 5, 1, "", "transposed_conv"], [44, 5, 1, "", "transposed_conv_batchnorm"], [44, 5, 1, "", "transposed_conv_filters"], [44, 5, 1, "", "transposed_conv_filters_rate"], [44, 5, 1, "", "transposed_conv_kernel_size"], [44, 5, 1, "", "upsampling_stride"]], "sleap.nn": [[45, 0, 0, "-", "callbacks"], [69, 0, 0, "-", "evals"], [70, 0, 0, "-", "heads"], [71, 0, 0, "-", "identity"], [72, 0, 0, "-", "inference"], [73, 0, 0, "-", "losses"], [74, 0, 0, "-", "model"], [75, 0, 0, "-", "paf_grouping"], [76, 0, 0, "-", "peak_finding"], [77, 0, 0, "-", "system"], [80, 0, 0, "-", "tracking"], [81, 0, 0, "-", "training"], [82, 0, 0, "-", "utils"], [83, 0, 0, "-", "viz"]], "sleap.nn.callbacks": [[45, 2, 1, "", "MatplotlibSaver"], [45, 2, 1, "", "ModelCheckpointOnEvent"], [45, 2, 1, "", "ProgressReporterZMQ"], [45, 2, 1, "", "TensorBoardMatplotlibWriter"], [45, 2, 1, "", "TrainingControllerZMQ"]], "sleap.nn.callbacks.MatplotlibSaver": [[45, 4, 1, "", "on_epoch_end"], [45, 5, 1, "", "plot_fn"], [45, 5, 1, "", "prefix"], [45, 5, 1, "", "save_folder"]], "sleap.nn.callbacks.ModelCheckpointOnEvent": [[45, 5, 1, "", "event"], [45, 5, 1, "", "filepath"], [45, 4, 1, "", "on_epoch_end"], [45, 4, 1, "", "on_train_begin"], [45, 4, 1, "", "on_train_end"]], "sleap.nn.callbacks.ProgressReporterZMQ": [[45, 4, 1, "", "on_batch_begin"], [45, 4, 1, "", "on_batch_end"], [45, 4, 1, "", "on_epoch_begin"], [45, 4, 1, "", "on_epoch_end"], [45, 4, 1, "", "on_train_begin"], [45, 4, 1, "", "on_train_end"]], "sleap.nn.callbacks.TensorBoardMatplotlibWriter": [[45, 5, 1, "", "logdir"], [45, 4, 1, "", "on_epoch_end"], [45, 5, 1, "", "plot_fn"], [45, 5, 1, "", "tag"]], "sleap.nn.callbacks.TrainingControllerZMQ": [[45, 4, 1, "", "on_batch_end"], [45, 4, 1, "", "set_lr"]], "sleap.nn.config": [[46, 0, 0, "-", "data"], [47, 0, 0, "-", "model"], [48, 0, 0, "-", "optimization"], [49, 0, 0, "-", "outputs"], [50, 0, 0, "-", "training_job"], [51, 0, 0, "-", "utils"]], "sleap.nn.config.data": [[46, 2, 1, "", "DataConfig"], [46, 2, 1, "", "InstanceCroppingConfig"], [46, 2, 1, "", "LabelsConfig"], [46, 2, 1, "", "PreprocessingConfig"]], "sleap.nn.config.data.InstanceCroppingConfig": [[46, 5, 1, "", "center_on_part"], [46, 5, 1, "", "crop_size"], [46, 5, 1, "", "crop_size_detection_padding"]], "sleap.nn.config.data.LabelsConfig": [[46, 5, 1, "", "search_path_hints"], [46, 5, 1, "", "skeletons"], [46, 5, 1, "", "split_by_inds"], [46, 5, 1, "", "test_inds"], [46, 5, 1, "", "test_labels"], [46, 5, 1, "", "training_inds"], [46, 5, 1, "", "training_labels"], [46, 5, 1, "", "validation_fraction"], [46, 5, 1, "", "validation_inds"], [46, 5, 1, "", "validation_labels"]], "sleap.nn.config.data.PreprocessingConfig": [[46, 5, 1, "", "ensure_grayscale"], [46, 5, 1, "", "ensure_rgb"], [46, 5, 1, "", "imagenet_mode"], [46, 5, 1, "", "input_scale"], [46, 5, 1, "", "pad_to_stride"], [46, 5, 1, "", "resize_and_pad_to_target"], [46, 5, 1, "", "target_height"], [46, 5, 1, "", "target_width"]], "sleap.nn.config.model": [[47, 2, 1, "", "BackboneConfig"], [47, 2, 1, "", "CenteredInstanceConfmapsHeadConfig"], [47, 2, 1, "", "CentroidsHeadConfig"], [47, 2, 1, "", "ClassMapsHeadConfig"], [47, 2, 1, "", "ClassVectorsHeadConfig"], [47, 2, 1, "", "HeadsConfig"], [47, 2, 1, "", "HourglassConfig"], [47, 2, 1, "", "LEAPConfig"], [47, 2, 1, "", "ModelConfig"], [47, 2, 1, "", "MultiClassBottomUpConfig"], [47, 2, 1, "", "MultiClassTopDownConfig"], [47, 2, 1, "", "MultiInstanceConfig"], [47, 2, 1, "", "MultiInstanceConfmapsHeadConfig"], [47, 2, 1, "", "PartAffinityFieldsHeadConfig"], [47, 2, 1, "", "PretrainedEncoderConfig"], [47, 2, 1, "", "ResNetConfig"], [47, 2, 1, "", "SingleInstanceConfmapsHeadConfig"], [47, 2, 1, "", "UNetConfig"], [47, 2, 1, "", "UpsamplingConfig"]], "sleap.nn.config.model.BackboneConfig": [[47, 5, 1, "", "hourglass"], [47, 5, 1, "", "leap"], [47, 5, 1, "", "resnet"], [47, 5, 1, "", "unet"]], "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig": [[47, 5, 1, "", "anchor_part"], [47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "offset_refinement"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "part_names"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.CentroidsHeadConfig": [[47, 5, 1, "", "anchor_part"], [47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "offset_refinement"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.ClassMapsHeadConfig": [[47, 5, 1, "", "classes"], [47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.ClassVectorsHeadConfig": [[47, 5, 1, "", "classes"], [47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "num_fc_layers"], [47, 5, 1, "", "num_fc_units"], [47, 5, 1, "", "output_stride"]], "sleap.nn.config.model.HeadsConfig": [[47, 5, 1, "", "centered_instance"], [47, 5, 1, "", "centroid"], [47, 5, 1, "", "multi_class_bottomup"], [47, 5, 1, "", "multi_class_topdown"], [47, 5, 1, "", "multi_instance"], [47, 5, 1, "", "single_instance"]], "sleap.nn.config.model.HourglassConfig": [[47, 5, 1, "", "filters"], [47, 5, 1, "", "filters_increase"], [47, 5, 1, "", "max_stride"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "stacks"], [47, 5, 1, "", "stem_stride"]], "sleap.nn.config.model.LEAPConfig": [[47, 5, 1, "", "filters"], [47, 5, 1, "", "filters_rate"], [47, 5, 1, "", "max_stride"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "stacks"], [47, 5, 1, "", "up_interpolate"]], "sleap.nn.config.model.ModelConfig": [[47, 5, 1, "", "backbone"], [47, 5, 1, "", "base_checkpoint"], [47, 5, 1, "", "heads"]], "sleap.nn.config.model.MultiClassBottomUpConfig": [[47, 5, 1, "", "class_maps"], [47, 5, 1, "", "confmaps"]], "sleap.nn.config.model.MultiClassTopDownConfig": [[47, 5, 1, "", "class_vectors"], [47, 5, 1, "", "confmaps"]], "sleap.nn.config.model.MultiInstanceConfig": [[47, 5, 1, "", "confmaps"], [47, 5, 1, "", "pafs"]], "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig": [[47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "offset_refinement"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "part_names"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.PartAffinityFieldsHeadConfig": [[47, 5, 1, "", "edges"], [47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.PretrainedEncoderConfig": [[47, 5, 1, "", "decoder_batchnorm"], [47, 5, 1, "", "decoder_filters"], [47, 5, 1, "", "decoder_filters_rate"], [47, 5, 1, "", "encoder"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "pretrained"]], "sleap.nn.config.model.ResNetConfig": [[47, 5, 1, "", "max_stride"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "upsampling"], [47, 5, 1, "", "version"], [47, 5, 1, "", "weights"]], "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig": [[47, 5, 1, "", "loss_weight"], [47, 5, 1, "", "offset_refinement"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "part_names"], [47, 5, 1, "", "sigma"]], "sleap.nn.config.model.UNetConfig": [[47, 5, 1, "", "filters"], [47, 5, 1, "", "filters_rate"], [47, 5, 1, "", "max_stride"], [47, 5, 1, "", "middle_block"], [47, 5, 1, "", "output_stride"], [47, 5, 1, "", "stacks"], [47, 5, 1, "", "stem_stride"], [47, 5, 1, "", "up_interpolate"]], "sleap.nn.config.model.UpsamplingConfig": [[47, 5, 1, "", "block_stride"], [47, 5, 1, "", "conv_batchnorm"], [47, 5, 1, "", "filters"], [47, 5, 1, "", "filters_rate"], [47, 5, 1, "", "method"], [47, 5, 1, "", "refine_convs"], [47, 5, 1, "", "skip_connections"], [47, 5, 1, "", "transposed_conv_kernel_size"]], "sleap.nn.config.optimization": [[48, 2, 1, "", "AugmentationConfig"], [48, 2, 1, "", "EarlyStoppingConfig"], [48, 2, 1, "", "HardKeypointMiningConfig"], [48, 2, 1, "", "LearningRateScheduleConfig"], [48, 2, 1, "", "OptimizationConfig"]], "sleap.nn.config.optimization.AugmentationConfig": [[48, 5, 1, "", "brightness"], [48, 5, 1, "", "brightness_max_val"], [48, 5, 1, "", "brightness_min_val"], [48, 5, 1, "", "contrast"], [48, 5, 1, "", "contrast_max_gamma"], [48, 5, 1, "", "contrast_min_gamma"], [48, 5, 1, "", "flip_horizontal"], [48, 5, 1, "", "gaussian_noise"], [48, 5, 1, "", "gaussian_noise_mean"], [48, 5, 1, "", "gaussian_noise_stddev"], [48, 5, 1, "", "random_crop"], [48, 5, 1, "", "random_crop_height"], [48, 5, 1, "", "random_crop_width"], [48, 5, 1, "", "random_flip"], [48, 5, 1, "", "rotate"], [48, 5, 1, "", "rotation_max_angle"], [48, 5, 1, "", "rotation_min_angle"], [48, 5, 1, "", "scale"], [48, 5, 1, "", "scale_max"], [48, 5, 1, "", "scale_min"], [48, 5, 1, "", "translate"], [48, 5, 1, "", "translate_max"], [48, 5, 1, "", "translate_min"], [48, 5, 1, "", "uniform_noise"], [48, 5, 1, "", "uniform_noise_max_val"], [48, 5, 1, "", "uniform_noise_min_val"]], "sleap.nn.config.optimization.EarlyStoppingConfig": [[48, 5, 1, "", "plateau_min_delta"], [48, 5, 1, "", "plateau_patience"], [48, 5, 1, "", "stop_training_on_plateau"]], "sleap.nn.config.optimization.HardKeypointMiningConfig": [[48, 5, 1, "", "hard_to_easy_ratio"], [48, 5, 1, "", "loss_scale"], [48, 5, 1, "", "max_hard_keypoints"], [48, 5, 1, "", "min_hard_keypoints"], [48, 5, 1, "", "online_mining"]], "sleap.nn.config.optimization.LearningRateScheduleConfig": [[48, 5, 1, "", "min_learning_rate"], [48, 5, 1, "", "plateau_cooldown"], [48, 5, 1, "", "plateau_min_delta"], [48, 5, 1, "", "plateau_patience"], [48, 5, 1, "", "reduce_on_plateau"], [48, 5, 1, "", "reduction_factor"]], "sleap.nn.config.optimization.OptimizationConfig": [[48, 5, 1, "", "augmentation_config"], [48, 5, 1, "", "batch_size"], [48, 5, 1, "", "batches_per_epoch"], [48, 5, 1, "", "early_stopping"], [48, 5, 1, "", "epochs"], [48, 5, 1, "", "hard_keypoint_mining"], [48, 5, 1, "", "initial_learning_rate"], [48, 5, 1, "", "learning_rate_schedule"], [48, 5, 1, "", "min_batches_per_epoch"], [48, 5, 1, "", "min_val_batches_per_epoch"], [48, 5, 1, "", "online_shuffling"], [48, 5, 1, "", "optimizer"], [48, 5, 1, "", "prefetch"], [48, 5, 1, "", "preload_data"], [48, 5, 1, "", "shuffle_buffer_size"], [48, 5, 1, "", "val_batches_per_epoch"]], "sleap.nn.config.outputs": [[49, 2, 1, "", "CheckpointingConfig"], [49, 2, 1, "", "OutputsConfig"], [49, 2, 1, "", "TensorBoardConfig"], [49, 2, 1, "", "ZMQConfig"]], "sleap.nn.config.outputs.CheckpointingConfig": [[49, 5, 1, "", "best_model"], [49, 5, 1, "", "every_epoch"], [49, 5, 1, "", "final_model"], [49, 5, 1, "", "initial_model"], [49, 5, 1, "", "latest_model"]], "sleap.nn.config.outputs.OutputsConfig": [[49, 5, 1, "", "checkpointing"], [49, 5, 1, "", "keep_viz_images"], [49, 5, 1, "", "log_to_csv"], [49, 5, 1, "", "run_name"], [49, 5, 1, "", "run_name_prefix"], [49, 5, 1, "", "run_name_suffix"], [49, 3, 1, "", "run_path"], [49, 5, 1, "", "runs_folder"], [49, 5, 1, "", "save_outputs"], [49, 5, 1, "", "save_visualizations"], [49, 5, 1, "", "tags"], [49, 5, 1, "", "tensorboard"], [49, 5, 1, "", "zip_outputs"], [49, 5, 1, "", "zmq"]], "sleap.nn.config.outputs.TensorBoardConfig": [[49, 5, 1, "", "architecture_graph"], [49, 5, 1, "", "loss_frequency"], [49, 5, 1, "", "profile_graph"], [49, 5, 1, "", "visualizations"], [49, 5, 1, "", "write_logs"]], "sleap.nn.config.outputs.ZMQConfig": [[49, 5, 1, "", "controller_address"], [49, 5, 1, "", "controller_polling_timeout"], [49, 5, 1, "", "publish_address"], [49, 5, 1, "", "publish_updates"], [49, 5, 1, "", "subscribe_to_controller"]], "sleap.nn.config.training_job": [[50, 2, 1, "", "TrainingJobConfig"], [50, 1, 1, "", "load_config"]], "sleap.nn.config.training_job.TrainingJobConfig": [[50, 5, 1, "", "data"], [50, 5, 1, "", "description"], [50, 5, 1, "", "filename"], [50, 4, 1, "", "from_json"], [50, 4, 1, "", "from_json_dicts"], [50, 4, 1, "", "load_json"], [50, 5, 1, "", "model"], [50, 5, 1, "", "name"], [50, 5, 1, "", "optimization"], [50, 5, 1, "", "outputs"], [50, 4, 1, "", "save_json"], [50, 5, 1, "", "sleap_version"], [50, 4, 1, "", "to_json"]], "sleap.nn.config.utils": [[51, 1, 1, "", "oneof"]], "sleap.nn.data": [[52, 0, 0, "-", "augmentation"], [53, 0, 0, "-", "confidence_maps"], [54, 0, 0, "-", "dataset_ops"], [55, 0, 0, "-", "edge_maps"], [56, 0, 0, "-", "general"], [57, 0, 0, "-", "grouping"], [58, 0, 0, "-", "identity"], [59, 0, 0, "-", "inference"], [60, 0, 0, "-", "instance_centroids"], [61, 0, 0, "-", "instance_cropping"], [62, 0, 0, "-", "normalization"], [63, 0, 0, "-", "offset_regression"], [64, 0, 0, "-", "pipelines"], [65, 0, 0, "-", "providers"], [66, 0, 0, "-", "resizing"], [67, 0, 0, "-", "training"], [68, 0, 0, "-", "utils"]], "sleap.nn.data.augmentation": [[52, 2, 1, "", "AlbumentationsAugmenter"], [52, 2, 1, "", "RandomCropper"], [52, 2, 1, "", "RandomFlipper"], [52, 1, 1, "", "flip_instances_lr"], [52, 1, 1, "", "flip_instances_ud"]], "sleap.nn.data.augmentation.AlbumentationsAugmenter": [[52, 5, 1, "", "augmenter"], [52, 4, 1, "", "from_config"], [52, 5, 1, "", "image_key"], [52, 3, 1, "", "input_keys"], [52, 5, 1, "", "instances_key"], [52, 3, 1, "", "output_keys"], [52, 4, 1, "", "transform_dataset"]], "sleap.nn.data.augmentation.RandomCropper": [[52, 5, 1, "", "crop_height"], [52, 5, 1, "", "crop_width"], [52, 4, 1, "", "transform_dataset"]], "sleap.nn.data.augmentation.RandomFlipper": [[52, 4, 1, "", "from_skeleton"], [52, 5, 1, "", "horizontal"], [52, 5, 1, "", "probability"], [52, 5, 1, "", "symmetric_inds"], [52, 4, 1, "", "transform_dataset"]], "sleap.nn.data.confidence_maps": [[53, 2, 1, "", "InstanceConfidenceMapGenerator"], [53, 2, 1, "", "MultiConfidenceMapGenerator"], [53, 2, 1, "", "SingleInstanceConfidenceMapGenerator"], [53, 1, 1, "", "make_confmaps"], [53, 1, 1, "", "make_multi_confmaps"], [53, 1, 1, "", "make_multi_confmaps_with_offsets"]], "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator": [[53, 5, 1, "", "all_instances"], [53, 5, 1, "", "flatten_offsets"], [53, 3, 1, "", "input_keys"], [53, 5, 1, "", "offsets_threshold"], [53, 3, 1, "", "output_keys"], [53, 5, 1, "", "output_stride"], [53, 5, 1, "", "sigma"], [53, 4, 1, "", "transform_dataset"], [53, 5, 1, "", "with_offsets"]], "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator": [[53, 5, 1, "", "centroids"], [53, 5, 1, "", "flatten_offsets"], [53, 3, 1, "", "input_keys"], [53, 5, 1, "", "offsets_threshold"], [53, 3, 1, "", "output_keys"], [53, 5, 1, "", "output_stride"], [53, 5, 1, "", "sigma"], [53, 4, 1, "", "transform_dataset"], [53, 5, 1, "", "with_offsets"]], "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator": [[53, 5, 1, "", "flatten_offsets"], [53, 3, 1, "", "input_keys"], [53, 5, 1, "", "offsets_threshold"], [53, 3, 1, "", "output_keys"], [53, 5, 1, "", "output_stride"], [53, 5, 1, "", "sigma"], [53, 4, 1, "", "transform_dataset"], [53, 5, 1, "", "with_offsets"]], "sleap.nn.data.dataset_ops": [[54, 2, 1, "", "Batcher"], [54, 2, 1, "", "LambdaFilter"], [54, 2, 1, "", "Prefetcher"], [54, 2, 1, "", "Preloader"], [54, 2, 1, "", "Repeater"], [54, 2, 1, "", "Shuffler"], [54, 2, 1, "", "Unbatcher"]], "sleap.nn.data.dataset_ops.Batcher": [[54, 5, 1, "", "batch_size"], [54, 5, 1, "", "drop_remainder"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 4, 1, "", "transform_dataset"], [54, 5, 1, "", "unrag"]], "sleap.nn.data.dataset_ops.LambdaFilter": [[54, 5, 1, "", "filter_fn"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.dataset_ops.Prefetcher": [[54, 5, 1, "", "buffer_size"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 5, 1, "", "prefetch"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.dataset_ops.Preloader": [[54, 5, 1, "", "examples"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.dataset_ops.Repeater": [[54, 5, 1, "", "epochs"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 5, 1, "", "repeat"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.dataset_ops.Shuffler": [[54, 5, 1, "", "buffer_size"], [54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 5, 1, "", "reshuffle_each_iteration"], [54, 5, 1, "", "shuffle"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.dataset_ops.Unbatcher": [[54, 3, 1, "", "input_keys"], [54, 3, 1, "", "output_keys"], [54, 4, 1, "", "transform_dataset"]], "sleap.nn.data.edge_maps": [[55, 2, 1, "", "PartAffinityFieldsGenerator"], [55, 1, 1, "", "distance_to_edge"], [55, 1, 1, "", "get_edge_points"], [55, 1, 1, "", "make_edge_maps"], [55, 1, 1, "", "make_multi_pafs"], [55, 1, 1, "", "make_pafs"]], "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator": [[55, 5, 1, "", "flatten_channels"], [55, 3, 1, "", "input_keys"], [55, 3, 1, "", "output_keys"], [55, 5, 1, "", "output_stride"], [55, 5, 1, "", "sigma"], [55, 5, 1, "", "skeletons"], [55, 4, 1, "", "transform_dataset"]], "sleap.nn.data.general": [[56, 2, 1, "", "KeyDeviceMover"], [56, 2, 1, "", "KeyFilter"], [56, 2, 1, "", "KeyRenamer"], [56, 2, 1, "", "LambdaMap"]], "sleap.nn.data.general.KeyDeviceMover": [[56, 3, 1, "", "input_keys"], [56, 3, 1, "", "output_keys"], [56, 4, 1, "", "transform_dataset"]], "sleap.nn.data.general.KeyFilter": [[56, 3, 1, "", "input_keys"], [56, 3, 1, "", "output_keys"], [56, 4, 1, "", "transform_dataset"]], "sleap.nn.data.general.KeyRenamer": [[56, 3, 1, "", "input_keys"], [56, 3, 1, "", "output_keys"], [56, 4, 1, "", "transform_dataset"]], "sleap.nn.data.general.LambdaMap": [[56, 5, 1, "", "func"], [56, 5, 1, "", "input_key_names"], [56, 3, 1, "", "input_keys"], [56, 5, 1, "", "output_key_names"], [56, 3, 1, "", "output_keys"], [56, 4, 1, "", "transform_dataset"]], "sleap.nn.data.grouping": [[57, 1, 1, "", "group_examples"], [57, 1, 1, "", "group_examples_iter"]], "sleap.nn.data.identity": [[58, 2, 1, "", "ClassMapGenerator"], [58, 2, 1, "", "ClassVectorGenerator"], [58, 1, 1, "", "make_class_maps"], [58, 1, 1, "", "make_class_vectors"]], "sleap.nn.data.identity.ClassMapGenerator": [[58, 5, 1, "", "centroids"], [58, 5, 1, "", "class_map_threshold"], [58, 3, 1, "", "input_keys"], [58, 3, 1, "", "output_keys"], [58, 5, 1, "", "output_stride"], [58, 5, 1, "", "sigma"], [58, 4, 1, "", "transform_dataset"]], "sleap.nn.data.identity.ClassVectorGenerator": [[58, 3, 1, "", "input_keys"], [58, 3, 1, "", "output_keys"], [58, 4, 1, "", "transform_dataset"]], "sleap.nn.data.inference": [[59, 2, 1, "", "GlobalPeakFinder"], [59, 2, 1, "", "KerasModelPredictor"], [59, 2, 1, "", "LocalPeakFinder"], [59, 2, 1, "", "MockGlobalPeakFinder"], [59, 2, 1, "", "PredictedCenterInstanceNormalizer"]], "sleap.nn.data.inference.PredictedCenterInstanceNormalizer": [[59, 3, 1, "", "input_keys"], [59, 3, 1, "", "output_keys"], [59, 4, 1, "", "transform_dataset"]], "sleap.nn.data.instance_centroids": [[60, 2, 1, "", "InstanceCentroidFinder"], [60, 1, 1, "", "find_points_bbox_midpoint"], [60, 1, 1, "", "get_instance_anchors"]], "sleap.nn.data.instance_centroids.InstanceCentroidFinder": [[60, 5, 1, "", "anchor_part_names"], [60, 5, 1, "", "center_on_anchor_part"], [60, 4, 1, "", "from_config"], [60, 3, 1, "", "input_keys"], [60, 5, 1, "", "instances_key"], [60, 3, 1, "", "output_keys"], [60, 5, 1, "", "skeletons"], [60, 4, 1, "", "transform_dataset"]], "sleap.nn.data.instance_cropping": [[61, 2, 1, "", "InstanceCropper"], [61, 1, 1, "", "crop_bboxes"], [61, 1, 1, "", "find_instance_crop_size"], [61, 1, 1, "", "make_centered_bboxes"], [61, 1, 1, "", "normalize_bboxes"], [61, 1, 1, "", "unnormalize_bboxes"]], "sleap.nn.data.instance_cropping.InstanceCropper": [[61, 5, 1, "", "centroids_key"], [61, 5, 1, "", "crop_height"], [61, 5, 1, "", "crop_width"], [61, 4, 1, "", "from_config"], [61, 5, 1, "", "image_key"], [61, 3, 1, "", "input_keys"], [61, 5, 1, "", "instances_key"], [61, 5, 1, "", "keep_full_image"], [61, 5, 1, "", "mock_centroid_confidence"], [61, 3, 1, "", "output_keys"], [61, 4, 1, "", "transform_dataset"], [61, 5, 1, "", "unbatch"]], "sleap.nn.data.normalization": [[62, 2, 1, "", "Normalizer"], [62, 1, 1, "", "convert_rgb_to_bgr"], [62, 1, 1, "", "ensure_float"], [62, 1, 1, "", "ensure_grayscale"], [62, 1, 1, "", "ensure_int"], [62, 1, 1, "", "ensure_min_image_rank"], [62, 1, 1, "", "ensure_rgb"], [62, 1, 1, "", "scale_image_range"], [62, 1, 1, "", "scale_to_imagenet_caffe_mode"], [62, 1, 1, "", "scale_to_imagenet_tf_mode"], [62, 1, 1, "", "scale_to_imagenet_torch_mode"]], "sleap.nn.data.normalization.Normalizer": [[62, 5, 1, "", "ensure_float"], [62, 5, 1, "", "ensure_grayscale"], [62, 5, 1, "", "ensure_rgb"], [62, 4, 1, "", "from_config"], [62, 5, 1, "", "image_key"], [62, 5, 1, "", "imagenet_mode"], [62, 3, 1, "", "input_keys"], [62, 3, 1, "", "output_keys"], [62, 4, 1, "", "transform_dataset"]], "sleap.nn.data.offset_regression": [[63, 1, 1, "", "make_offsets"], [63, 1, 1, "", "mask_offsets"]], "sleap.nn.data.pipelines": [[64, 2, 1, "", "BottomUpMultiClassPipeline"], [64, 2, 1, "", "BottomUpPipeline"], [64, 2, 1, "", "CentroidConfmapsPipeline"], [64, 2, 1, "", "Pipeline"], [64, 2, 1, "", "SingleInstanceConfmapsPipeline"], [64, 2, 1, "", "TopDownMultiClassPipeline"], [64, 2, 1, "", "TopdownConfmapsPipeline"]], "sleap.nn.data.pipelines.BottomUpMultiClassPipeline": [[64, 5, 1, "", "class_maps_head"], [64, 5, 1, "", "confmaps_head"], [64, 5, 1, "", "data_config"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"]], "sleap.nn.data.pipelines.BottomUpPipeline": [[64, 5, 1, "", "confmaps_head"], [64, 5, 1, "", "data_config"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"], [64, 5, 1, "", "pafs_head"]], "sleap.nn.data.pipelines.CentroidConfmapsPipeline": [[64, 5, 1, "", "centroid_confmap_head"], [64, 5, 1, "", "data_config"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"]], "sleap.nn.data.pipelines.Pipeline": [[64, 4, 1, "", "append"], [64, 4, 1, "", "describe"], [64, 4, 1, "", "from_blocks"], [64, 4, 1, "", "from_pipelines"], [64, 4, 1, "", "make_dataset"], [64, 3, 1, "", "output_keys"], [64, 4, 1, "", "peek"], [64, 5, 1, "", "providers"], [64, 4, 1, "", "run"], [64, 5, 1, "", "transformers"], [64, 4, 1, "", "validate_pipeline"]], "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline": [[64, 5, 1, "", "data_config"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"], [64, 5, 1, "", "single_instance_confmap_head"]], "sleap.nn.data.pipelines.TopDownMultiClassPipeline": [[64, 5, 1, "", "class_vectors_head"], [64, 5, 1, "", "confmaps_head"], [64, 5, 1, "", "data_config"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"]], "sleap.nn.data.pipelines.TopdownConfmapsPipeline": [[64, 5, 1, "", "data_config"], [64, 5, 1, "", "instance_confmap_head"], [64, 4, 1, "", "make_base_pipeline"], [64, 4, 1, "", "make_training_pipeline"], [64, 4, 1, "", "make_viz_pipeline"], [64, 5, 1, "", "offsets_head"], [64, 5, 1, "", "optimization_config"]], "sleap.nn.data.providers": [[65, 2, 1, "", "LabelsReader"], [65, 2, 1, "", "VideoReader"]], "sleap.nn.data.providers.LabelsReader": [[65, 5, 1, "", "example_indices"], [65, 4, 1, "", "from_filename"], [65, 4, 1, "", "from_unlabeled_suggestions"], [65, 4, 1, "", "from_user_instances"], [65, 4, 1, "", "from_user_labeled_frames"], [65, 3, 1, "", "is_from_multi_size_videos"], [65, 5, 1, "", "labels"], [65, 4, 1, "", "make_dataset"], [65, 3, 1, "", "max_height_and_width"], [65, 3, 1, "", "output_keys"], [65, 3, 1, "", "tracks"], [65, 5, 1, "", "user_instances_only"], [65, 3, 1, "", "videos"], [65, 5, 1, "", "with_track_only"]], "sleap.nn.data.providers.VideoReader": [[65, 5, 1, "", "example_indices"], [65, 4, 1, "", "from_filepath"], [65, 4, 1, "", "make_dataset"], [65, 3, 1, "", "max_height_and_width"], [65, 3, 1, "", "output_keys"], [65, 5, 1, "", "video"], [65, 5, 1, "", "video_ind"], [65, 3, 1, "", "videos"]], "sleap.nn.data.resizing": [[66, 2, 1, "", "PointsRescaler"], [66, 2, 1, "", "Resizer"], [66, 2, 1, "", "SizeMatcher"], [66, 1, 1, "", "find_padding_for_stride"], [66, 1, 1, "", "pad_to_stride"], [66, 1, 1, "", "resize_image"]], "sleap.nn.data.resizing.PointsRescaler": [[66, 3, 1, "", "input_keys"], [66, 3, 1, "", "output_keys"], [66, 4, 1, "", "transform_dataset"]], "sleap.nn.data.resizing.Resizer": [[66, 4, 1, "", "from_config"], [66, 5, 1, "", "full_image_key"], [66, 5, 1, "", "image_key"], [66, 3, 1, "", "input_keys"], [66, 5, 1, "", "keep_full_image"], [66, 3, 1, "", "output_keys"], [66, 5, 1, "", "pad_to_stride"], [66, 5, 1, "", "points_key"], [66, 5, 1, "", "scale"], [66, 5, 1, "", "scale_key"], [66, 4, 1, "", "transform_dataset"]], "sleap.nn.data.resizing.SizeMatcher": [[66, 5, 1, "", "center_pad"], [66, 4, 1, "", "from_config"], [66, 5, 1, "", "full_image_key"], [66, 5, 1, "", "image_key"], [66, 3, 1, "", "input_keys"], [66, 5, 1, "", "keep_full_image"], [66, 5, 1, "", "max_image_height"], [66, 5, 1, "", "max_image_width"], [66, 3, 1, "", "output_keys"], [66, 5, 1, "", "points_key"], [66, 5, 1, "", "scale_key"], [66, 4, 1, "", "transform_dataset"]], "sleap.nn.data.training": [[67, 2, 1, "", "KeyMapper"], [67, 1, 1, "", "split_labels"], [67, 1, 1, "", "split_labels_reader"], [67, 1, 1, "", "split_labels_train_val"]], "sleap.nn.data.training.KeyMapper": [[67, 3, 1, "", "input_keys"], [67, 5, 1, "", "key_maps"], [67, 3, 1, "", "output_keys"], [67, 4, 1, "", "transform_dataset"]], "sleap.nn.data.utils": [[68, 1, 1, "", "describe_tensors"], [68, 1, 1, "", "ensure_list"], [68, 1, 1, "", "expand_to_rank"], [68, 1, 1, "", "gaussian_pdf"], [68, 1, 1, "", "make_grid_vectors"], [68, 1, 1, "", "unrag_example"], [68, 1, 1, "", "unrag_tensor"]], "sleap.nn.evals": [[69, 1, 1, "", "compute_dist_metrics"], [69, 1, 1, "", "compute_dists"], [69, 1, 1, "", "compute_generalized_voc_metrics"], [69, 1, 1, "", "compute_instance_area"], [69, 1, 1, "", "compute_oks"], [69, 1, 1, "", "compute_pck_metrics"], [69, 1, 1, "", "compute_visibility_conf"], [69, 1, 1, "", "evaluate"], [69, 1, 1, "", "evaluate_model"], [69, 1, 1, "", "find_frame_pairs"], [69, 1, 1, "", "load_metrics"], [69, 1, 1, "", "match_frame_pairs"], [69, 1, 1, "", "match_instances"], [69, 1, 1, "", "replace_path"]], "sleap.nn.heads": [[70, 2, 1, "", "CenteredInstanceConfmapsHead"], [70, 2, 1, "", "CentroidConfmapsHead"], [70, 2, 1, "", "ClassMapsHead"], [70, 2, 1, "", "ClassVectorsHead"], [70, 2, 1, "", "Head"], [70, 2, 1, "", "MultiInstanceConfmapsHead"], [70, 2, 1, "", "OffsetRefinementHead"], [70, 2, 1, "", "PartAffinityFieldsHead"], [70, 2, 1, "", "SingleInstanceConfmapsHead"]], "sleap.nn.heads.CenteredInstanceConfmapsHead": [[70, 5, 1, "", "anchor_part"], [70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "id0", "part_names"], [70, 5, 1, "", "sigma"]], "sleap.nn.heads.CentroidConfmapsHead": [[70, 5, 1, "", "anchor_part"], [70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "", "sigma"]], "sleap.nn.heads.ClassMapsHead": [[70, 3, 1, "", "activation"], [70, 3, 1, "", "channels"], [70, 5, 1, "id1", "classes"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "", "sigma"]], "sleap.nn.heads.ClassVectorsHead": [[70, 3, 1, "", "activation"], [70, 3, 1, "", "channels"], [70, 5, 1, "id2", "classes"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 3, 1, "", "loss_function"], [70, 5, 1, "", "loss_weight"], [70, 4, 1, "", "make_head"], [70, 5, 1, "", "num_fc_layers"], [70, 5, 1, "", "num_fc_units"], [70, 5, 1, "", "output_stride"]], "sleap.nn.heads.Head": [[70, 3, 1, "", "activation"], [70, 3, 1, "", "channels"], [70, 3, 1, "", "loss_function"], [70, 4, 1, "", "make_head"]], "sleap.nn.heads.MultiInstanceConfmapsHead": [[70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "id3", "part_names"], [70, 5, 1, "", "sigma"]], "sleap.nn.heads.OffsetRefinementHead": [[70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "id6", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "id4", "part_names"], [70, 5, 1, "id5", "sigma_threshold"]], "sleap.nn.heads.PartAffinityFieldsHead": [[70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 5, 1, "id7", "edges"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "", "sigma"]], "sleap.nn.heads.SingleInstanceConfmapsHead": [[70, 3, 1, "", "channels"], [70, 5, 1, "", "config"], [70, 4, 1, "", "from_config"], [70, 5, 1, "", "loss_weight"], [70, 5, 1, "", "output_stride"], [70, 5, 1, "id8", "part_names"], [70, 5, 1, "", "sigma"]], "sleap.nn.identity": [[71, 1, 1, "", "classify_peaks_from_maps"], [71, 1, 1, "", "classify_peaks_from_vectors"], [71, 1, 1, "", "group_class_peaks"]], "sleap.nn.inference": [[72, 2, 1, "", "BottomUpInferenceLayer"], [72, 2, 1, "", "BottomUpInferenceModel"], [72, 2, 1, "", "BottomUpMultiClassInferenceLayer"], [72, 2, 1, "", "BottomUpMultiClassInferenceModel"], [72, 2, 1, "", "BottomUpMultiClassPredictor"], [72, 2, 1, "", "BottomUpPredictor"], [72, 2, 1, "", "CentroidCrop"], [72, 2, 1, "", "CentroidCropGroundTruth"], [72, 2, 1, "", "CentroidInferenceModel"], [72, 2, 1, "", "FindInstancePeaks"], [72, 2, 1, "", "FindInstancePeaksGroundTruth"], [72, 2, 1, "", "InferenceLayer"], [72, 2, 1, "", "InferenceModel"], [72, 2, 1, "", "MoveNetInferenceLayer"], [72, 2, 1, "", "MoveNetInferenceModel"], [72, 2, 1, "", "MoveNetPredictor"], [72, 2, 1, "", "Predictor"], [72, 2, 1, "", "RateColumn"], [72, 2, 1, "", "SingleInstanceInferenceLayer"], [72, 2, 1, "", "SingleInstanceInferenceModel"], [72, 2, 1, "", "SingleInstancePredictor"], [72, 2, 1, "", "TopDownInferenceModel"], [72, 2, 1, "", "TopDownMultiClassFindPeaks"], [72, 2, 1, "", "TopDownMultiClassInferenceModel"], [72, 2, 1, "", "TopDownMultiClassPredictor"], [72, 2, 1, "", "TopDownPredictor"], [72, 2, 1, "", "VisualPredictor"], [72, 1, 1, "", "export_cli"], [72, 1, 1, "", "export_model"], [72, 1, 1, "", "find_head"], [72, 1, 1, "", "get_keras_model_path"], [72, 1, 1, "", "get_model_output_stride"], [72, 1, 1, "", "load_model"], [72, 1, 1, "", "main"], [72, 1, 1, "", "make_model_movenet"]], "sleap.nn.inference.BottomUpInferenceLayer": [[72, 4, 1, "", "call"], [72, 5, 1, "", "cm_output_stride"], [72, 5, 1, "", "confmaps_ind"], [72, 4, 1, "", "find_peaks"], [72, 4, 1, "", "forward_pass"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "paf_output_stride"], [72, 5, 1, "", "paf_scorer"], [72, 5, 1, "", "pafs_ind"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_confmaps"], [72, 5, 1, "", "return_paf_graph"], [72, 5, 1, "", "return_pafs"]], "sleap.nn.inference.BottomUpInferenceModel": [[72, 5, 1, "", "bottomup_layer"], [72, 4, 1, "", "call"]], "sleap.nn.inference.BottomUpMultiClassInferenceLayer": [[72, 4, 1, "", "call"], [72, 5, 1, "", "class_maps_ind"], [72, 5, 1, "", "class_maps_output_stride"], [72, 5, 1, "", "cm_output_stride"], [72, 5, 1, "", "confmaps_ind"], [72, 4, 1, "", "find_peaks"], [72, 4, 1, "", "forward_pass"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_class_maps"], [72, 5, 1, "", "return_confmaps"]], "sleap.nn.inference.BottomUpMultiClassInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "inference_layer"]], "sleap.nn.inference.BottomUpMultiClassPredictor": [[72, 5, 1, "", "batch_size"], [72, 5, 1, "", "config"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "integral_refinement"], [72, 3, 1, "", "is_grayscale"], [72, 5, 1, "", "model"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"], [72, 5, 1, "", "tracks"]], "sleap.nn.inference.BottomUpPredictor": [[72, 5, 1, "", "batch_size"], [72, 5, 1, "", "bottomup_config"], [72, 5, 1, "", "bottomup_model"], [72, 5, 1, "", "dist_penalty_weight"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "integral_refinement"], [72, 3, 1, "", "is_grayscale"], [72, 5, 1, "", "max_edge_length_ratio"], [72, 5, 1, "", "max_instances"], [72, 5, 1, "", "min_line_scores"], [72, 5, 1, "", "paf_line_points"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"], [72, 5, 1, "", "tracker"]], "sleap.nn.inference.CentroidCrop": [[72, 4, 1, "", "call"], [72, 5, 1, "", "confmaps_ind"], [72, 5, 1, "", "crop_size"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "max_instances"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "output_stride"], [72, 5, 1, "", "pad_to_stride"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_confmaps"], [72, 5, 1, "", "return_crops"]], "sleap.nn.inference.CentroidCropGroundTruth": [[72, 4, 1, "", "call"], [72, 5, 1, "", "crop_size"]], "sleap.nn.inference.CentroidInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "centroid_crop"]], "sleap.nn.inference.FindInstancePeaks": [[72, 4, 1, "", "call"], [72, 5, 1, "", "confmaps_ind"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "output_stride"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_confmaps"]], "sleap.nn.inference.FindInstancePeaksGroundTruth": [[72, 4, 1, "", "call"]], "sleap.nn.inference.InferenceLayer": [[72, 4, 1, "", "call"], [72, 5, 1, "", "ensure_float"], [72, 5, 1, "", "ensure_grayscale"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "pad_to_stride"], [72, 4, 1, "", "preprocess"]], "sleap.nn.inference.InferenceModel": [[72, 4, 1, "", "export_model"], [72, 4, 1, "", "predict"], [72, 4, 1, "", "predict_on_batch"]], "sleap.nn.inference.MoveNetInferenceLayer": [[72, 4, 1, "", "call"], [72, 5, 1, "", "ensure_float"], [72, 5, 1, "", "ensure_grayscale"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "model_name"], [72, 5, 1, "", "pad_to_stride"]], "sleap.nn.inference.MoveNetInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "inference_layer"]], "sleap.nn.inference.MoveNetPredictor": [[72, 5, 1, "", "batch_size"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 3, 1, "", "is_grayscale"], [72, 5, 1, "", "model_name"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"]], "sleap.nn.inference.Predictor": [[72, 4, 1, "", "export_model"], [72, 4, 1, "", "from_model_paths"], [72, 3, 1, "", "is_grayscale"], [72, 4, 1, "", "make_pipeline"], [72, 4, 1, "", "predict"], [72, 3, 1, "", "report_period"]], "sleap.nn.inference.RateColumn": [[72, 4, 1, "", "render"]], "sleap.nn.inference.SingleInstanceInferenceLayer": [[72, 4, 1, "", "call"], [72, 5, 1, "", "confmaps_ind"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "output_stride"], [72, 5, 1, "", "pad_to_stride"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_confmaps"]], "sleap.nn.inference.SingleInstanceInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "single_instance_layer"]], "sleap.nn.inference.SingleInstancePredictor": [[72, 5, 1, "", "batch_size"], [72, 5, 1, "", "confmap_config"], [72, 5, 1, "", "confmap_model"], [72, 4, 1, "", "export_model"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "integral_refinement"], [72, 3, 1, "", "is_grayscale"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"]], "sleap.nn.inference.TopDownInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "centroid_crop"], [72, 5, 1, "", "instance_peaks"]], "sleap.nn.inference.TopDownMultiClassFindPeaks": [[72, 4, 1, "", "call"], [72, 5, 1, "", "class_vectors_ind"], [72, 5, 1, "", "confmaps_ind"], [72, 5, 1, "", "input_scale"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "keras_model"], [72, 5, 1, "", "offsets_ind"], [72, 5, 1, "", "optimal_grouping"], [72, 5, 1, "", "output_stride"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "refinement"], [72, 5, 1, "", "return_class_vectors"], [72, 5, 1, "", "return_confmaps"]], "sleap.nn.inference.TopDownMultiClassInferenceModel": [[72, 4, 1, "", "call"], [72, 5, 1, "", "centroid_crop"], [72, 4, 1, "", "export_model"], [72, 5, 1, "", "instance_peaks"]], "sleap.nn.inference.TopDownMultiClassPredictor": [[72, 5, 1, "", "batch_size"], [72, 5, 1, "", "centroid_config"], [72, 5, 1, "", "centroid_model"], [72, 5, 1, "", "confmap_config"], [72, 5, 1, "", "confmap_model"], [72, 4, 1, "", "export_model"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "integral_refinement"], [72, 3, 1, "", "is_grayscale"], [72, 4, 1, "", "make_pipeline"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"], [72, 5, 1, "", "tracker"], [72, 5, 1, "", "tracks"]], "sleap.nn.inference.TopDownPredictor": [[72, 5, 1, "", "batch_size"], [72, 5, 1, "", "centroid_config"], [72, 5, 1, "", "centroid_model"], [72, 5, 1, "", "confmap_config"], [72, 5, 1, "", "confmap_model"], [72, 4, 1, "", "export_model"], [72, 4, 1, "", "from_trained_models"], [72, 5, 1, "", "inference_model"], [72, 5, 1, "", "integral_patch_size"], [72, 5, 1, "", "integral_refinement"], [72, 3, 1, "", "is_grayscale"], [72, 4, 1, "", "make_pipeline"], [72, 5, 1, "", "max_instances"], [72, 5, 1, "", "peak_threshold"], [72, 5, 1, "", "pipeline"], [72, 5, 1, "", "tracker"]], "sleap.nn.inference.VisualPredictor": [[72, 4, 1, "", "make_pipeline"], [72, 4, 1, "", "predict"], [72, 4, 1, "", "safely_generate"]], "sleap.nn.losses": [[73, 2, 1, "", "OHKMLoss"], [73, 2, 1, "", "PartLoss"], [73, 1, 1, "", "compute_ohkm_loss"]], "sleap.nn.losses.OHKMLoss": [[73, 4, 1, "", "call"], [73, 4, 1, "", "from_config"], [73, 5, 1, "", "hard_to_easy_ratio"], [73, 5, 1, "", "loss_scale"], [73, 5, 1, "", "max_hard_keypoints"], [73, 5, 1, "", "min_hard_keypoints"]], "sleap.nn.losses.PartLoss": [[73, 5, 1, "", "channel_ind"], [73, 5, 1, "", "name"], [73, 4, 1, "", "result"], [73, 4, 1, "", "update_state"]], "sleap.nn.model": [[74, 2, 1, "", "Model"]], "sleap.nn.model.Model": [[74, 5, 1, "", "backbone"], [74, 4, 1, "", "from_config"], [74, 5, 1, "", "heads"], [74, 5, 1, "", "keras_model"], [74, 4, 1, "", "make_model"], [74, 3, 1, "", "maximum_stride"]], "sleap.nn.paf_grouping": [[75, 2, 1, "", "EdgeConnection"], [75, 2, 1, "", "EdgeType"], [75, 2, 1, "", "PAFScorer"], [75, 2, 1, "", "PeakID"], [75, 1, 1, "", "assign_connections_to_instances"], [75, 1, 1, "", "compute_distance_penalty"], [75, 1, 1, "", "get_connection_candidates"], [75, 1, 1, "", "get_paf_lines"], [75, 1, 1, "", "group_instances_batch"], [75, 1, 1, "", "group_instances_sample"], [75, 1, 1, "", "make_line_subs"], [75, 1, 1, "", "make_predicted_instances"], [75, 1, 1, "", "match_candidates_batch"], [75, 1, 1, "", "match_candidates_sample"], [75, 1, 1, "", "score_paf_lines"], [75, 1, 1, "", "score_paf_lines_batch"], [75, 1, 1, "", "toposort_edges"]], "sleap.nn.paf_grouping.EdgeConnection": [[75, 5, 1, "", "dst_peak_ind"], [75, 5, 1, "", "score"], [75, 5, 1, "", "src_peak_ind"]], "sleap.nn.paf_grouping.EdgeType": [[75, 5, 1, "", "dst_node_ind"], [75, 5, 1, "", "src_node_ind"]], "sleap.nn.paf_grouping.PAFScorer": [[75, 5, 1, "", "dist_penalty_weight"], [75, 5, 1, "", "edge_inds"], [75, 5, 1, "", "edge_types"], [75, 5, 1, "", "edges"], [75, 4, 1, "", "from_config"], [75, 4, 1, "", "group_instances"], [75, 4, 1, "", "match_candidates"], [75, 5, 1, "", "max_edge_length_ratio"], [75, 5, 1, "", "min_instance_peaks"], [75, 5, 1, "", "min_line_scores"], [75, 5, 1, "", "n_edges"], [75, 5, 1, "", "n_nodes"], [75, 5, 1, "", "n_points"], [75, 5, 1, "", "pafs_stride"], [75, 5, 1, "", "part_names"], [75, 4, 1, "", "predict"], [75, 4, 1, "", "score_paf_lines"], [75, 5, 1, "", "sorted_edge_inds"]], "sleap.nn.paf_grouping.PeakID": [[75, 5, 1, "", "node_ind"], [75, 5, 1, "", "peak_ind"]], "sleap.nn.peak_finding": [[76, 1, 1, "", "crop_bboxes"], [76, 1, 1, "", "find_global_peaks"], [76, 1, 1, "", "find_global_peaks_integral"], [76, 1, 1, "", "find_global_peaks_rough"], [76, 1, 1, "", "find_global_peaks_with_offsets"], [76, 1, 1, "", "find_local_peaks"], [76, 1, 1, "", "find_local_peaks_integral"], [76, 1, 1, "", "find_local_peaks_rough"], [76, 1, 1, "", "find_local_peaks_with_offsets"], [76, 1, 1, "", "find_offsets_local_direction"], [76, 1, 1, "", "integral_regression"], [76, 1, 1, "", "make_gaussian_kernel"], [76, 1, 1, "", "smooth_imgs"]], "sleap.nn.system": [[77, 1, 1, "", "best_logical_device_name"], [77, 1, 1, "", "disable_preallocation"], [77, 1, 1, "", "enable_preallocation"], [77, 1, 1, "", "get_all_gpus"], [77, 1, 1, "", "get_available_gpus"], [77, 1, 1, "", "get_current_gpu"], [77, 1, 1, "", "get_gpu_memory"], [77, 1, 1, "", "initialize_devices"], [77, 1, 1, "", "is_gpu_system"], [77, 1, 1, "", "is_initialized"], [77, 1, 1, "", "summary"], [77, 1, 1, "", "use_cpu_only"], [77, 1, 1, "", "use_first_gpu"], [77, 1, 1, "", "use_gpu"], [77, 1, 1, "", "use_last_gpu"]], "sleap.nn.tracker": [[78, 0, 0, "-", "components"], [79, 0, 0, "-", "kalman"]], "sleap.nn.tracker.components": [[78, 2, 1, "", "FrameMatches"], [78, 2, 1, "", "Match"], [78, 1, 1, "", "centroid_distance"], [78, 1, 1, "", "connect_single_track_breaks"], [78, 1, 1, "", "cull_frame_instances"], [78, 1, 1, "", "cull_instances"], [78, 1, 1, "", "factory_object_keypoint_similarity"], [78, 1, 1, "", "first_choice_matching"], [78, 1, 1, "", "greedy_matching"], [78, 1, 1, "", "hungarian_matching"], [78, 1, 1, "", "instance_iou"], [78, 1, 1, "", "instance_similarity"], [78, 1, 1, "", "nms_fast"], [78, 1, 1, "", "normalized_instance_similarity"]], "sleap.nn.tracker.components.FrameMatches": [[78, 5, 1, "", "cost_matrix"], [78, 4, 1, "", "from_candidate_instances"], [78, 3, 1, "", "has_only_first_choice_matches"], [78, 5, 1, "", "matches"], [78, 5, 1, "", "unmatched_instances"]], "sleap.nn.tracker.kalman": [[79, 1, 1, "", "get_track_instance_matches"], [79, 1, 1, "", "match_dict_from_match_function"], [79, 1, 1, "", "remove_second_bests_from_cost_matrix"]], "sleap.nn.tracking": [[80, 2, 1, "", "BaseTracker"], [80, 2, 1, "", "FlowCandidateMaker"], [80, 2, 1, "", "FlowMaxTracker"], [80, 2, 1, "", "FlowMaxTracksCandidateMaker"], [80, 2, 1, "", "FlowTracker"], [80, 2, 1, "", "KalmanTracker"], [80, 2, 1, "", "SimpleCandidateMaker"], [80, 2, 1, "", "SimpleMaxTracker"], [80, 2, 1, "", "SimpleMaxTracksCandidateMaker"], [80, 2, 1, "", "SimpleTracker"], [80, 2, 1, "", "TrackCleaner"], [80, 2, 1, "", "Tracker"], [80, 1, 1, "", "run_tracker"]], "sleap.nn.tracking.FlowCandidateMaker": [[80, 4, 1, "", "flow_shift_instances"], [80, 4, 1, "", "get_shifted_instances"], [80, 4, 1, "", "get_shifted_instances_from_earlier_time"], [80, 5, 1, "", "img_scale"], [80, 5, 1, "", "min_points"], [80, 5, 1, "", "of_max_levels"], [80, 5, 1, "", "of_window_size"], [80, 4, 1, "", "prune_shifted_instances"], [80, 5, 1, "", "save_shifted_instances"], [80, 5, 1, "", "track_window"]], "sleap.nn.tracking.FlowMaxTracker": [[80, 4, 1, "", "matching_function"], [80, 4, 1, "", "similarity_function"]], "sleap.nn.tracking.FlowMaxTracksCandidateMaker": [[80, 4, 1, "", "get_ref_instances"], [80, 5, 1, "", "max_tracks"]], "sleap.nn.tracking.KalmanTracker": [[80, 3, 1, "", "is_valid"], [80, 4, 1, "", "make_tracker"], [80, 4, 1, "", "track"]], "sleap.nn.tracking.TrackCleaner": [[80, 5, 1, "", "instance_count"], [80, 5, 1, "", "iou_threshold"]], "sleap.nn.tracking.Tracker": [[80, 5, 1, "", "candidate_maker"], [80, 5, 1, "", "cleaner"], [80, 4, 1, "", "final_pass"], [80, 5, 1, "", "matching_function"], [80, 5, 1, "", "max_tracking"], [80, 5, 1, "", "min_new_track_points"], [80, 5, 1, "", "robust_best_instance"], [80, 5, 1, "", "similarity_function"], [80, 4, 1, "", "track"], [80, 5, 1, "", "track_window"], [80, 3, 1, "", "unique_tracks_in_queue"]], "sleap.nn.training": [[81, 2, 1, "", "BottomUpModelTrainer"], [81, 2, 1, "", "BottomUpMultiClassModelTrainer"], [81, 2, 1, "", "CentroidConfmapsModelTrainer"], [81, 2, 1, "", "DataReaders"], [81, 2, 1, "", "SingleInstanceModelTrainer"], [81, 2, 1, "", "TopDownMultiClassModelTrainer"], [81, 2, 1, "", "TopdownConfmapsModelTrainer"], [81, 2, 1, "", "Trainer"], [81, 1, 1, "", "create_trainer_using_cli"], [81, 1, 1, "", "get_timestamp"], [81, 1, 1, "", "main"], [81, 1, 1, "", "sanitize_scope_name"], [81, 1, 1, "", "setup_checkpointing"], [81, 1, 1, "", "setup_losses"], [81, 1, 1, "", "setup_metrics"], [81, 1, 1, "", "setup_new_run_folder"], [81, 1, 1, "", "setup_optimization_callbacks"], [81, 1, 1, "", "setup_optimizer"], [81, 1, 1, "", "setup_output_callbacks"], [81, 1, 1, "", "setup_tensorboard"], [81, 1, 1, "", "setup_visualization"], [81, 1, 1, "", "setup_zmq_callbacks"]], "sleap.nn.training.BottomUpModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.BottomUpMultiClassModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.CentroidConfmapsModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.DataReaders": [[81, 4, 1, "", "from_config"], [81, 4, 1, "", "from_labels"], [81, 3, 1, "", "test_labels"], [81, 5, 1, "", "test_labels_reader"], [81, 3, 1, "", "training_labels"], [81, 5, 1, "", "training_labels_reader"], [81, 3, 1, "", "validation_labels"], [81, 5, 1, "", "validation_labels_reader"]], "sleap.nn.training.SingleInstanceModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.TopDownMultiClassModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.TopdownConfmapsModelTrainer": [[81, 3, 1, "", "has_offsets"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "output_keys"]], "sleap.nn.training.Trainer": [[81, 3, 1, "", "callbacks"], [81, 4, 1, "", "cleanup"], [81, 5, 1, "", "config"], [81, 5, 1, "", "data_readers"], [81, 4, 1, "", "evaluate"], [81, 4, 1, "", "from_config"], [81, 5, 1, "", "initial_config"], [81, 3, 1, "", "input_keys"], [81, 3, 1, "", "keras_model"], [81, 5, 1, "", "model"], [81, 5, 1, "", "optimization_callbacks"], [81, 5, 1, "", "output_callbacks"], [81, 3, 1, "", "output_keys"], [81, 4, 1, "", "package"], [81, 5, 1, "", "pipeline_builder"], [81, 5, 1, "", "run_path"], [81, 4, 1, "", "setup"], [81, 4, 1, "", "train"], [81, 5, 1, "", "training_pipeline"], [81, 5, 1, "", "training_viz_pipeline"], [81, 5, 1, "", "validation_pipeline"], [81, 5, 1, "", "validation_viz_pipeline"], [81, 5, 1, "", "visualization_callbacks"]], "sleap.nn.utils": [[82, 1, 1, "", "compute_iou"], [82, 1, 1, "", "group_array"], [82, 1, 1, "", "match_points"], [82, 1, 1, "", "reset_input_layer"], [82, 1, 1, "", "tf_linear_sum_assignment"]], "sleap.nn.viz": [[83, 1, 1, "", "generate_skeleton_preview_image"], [83, 1, 1, "", "imgfig"], [83, 1, 1, "", "plot_confmaps"], [83, 1, 1, "", "plot_img"], [83, 1, 1, "", "plot_instance"], [83, 1, 1, "", "plot_instances"], [83, 1, 1, "", "plot_pafs"], [83, 1, 1, "", "plot_peaks"]], "sleap.skeleton": [[84, 2, 1, "", "EdgeType"], [84, 2, 1, "", "Node"], [84, 2, 1, "", "Skeleton"], [84, 2, 1, "", "SkeletonDecoder"], [84, 2, 1, "", "SkeletonEncoder"]], "sleap.skeleton.Node": [[84, 4, 1, "", "as_node"], [84, 4, 1, "", "from_names"], [84, 4, 1, "", "matches"], [84, 5, 1, "", "name"], [84, 5, 1, "", "weight"]], "sleap.skeleton.Skeleton": [[84, 5, 1, "", "_is_template"], [84, 5, 1, "", "_skeleton_idx"], [84, 4, 1, "", "add_edge"], [84, 4, 1, "", "add_node"], [84, 4, 1, "", "add_nodes"], [84, 4, 1, "", "add_symmetry"], [84, 4, 1, "", "clear_edges"], [84, 4, 1, "", "delete_edge"], [84, 4, 1, "", "delete_node"], [84, 4, 1, "", "delete_symmetry"], [84, 5, 1, "", "description"], [84, 3, 1, "", "edge_inds"], [84, 3, 1, "", "edge_names"], [84, 4, 1, "", "edge_to_index"], [84, 3, 1, "", "edges"], [84, 3, 1, "", "edges_full"], [84, 4, 1, "", "find_neighbors"], [84, 4, 1, "", "find_node"], [84, 4, 1, "", "find_unique_nodes"], [84, 4, 1, "", "from_dict"], [84, 4, 1, "", "from_json"], [84, 4, 1, "", "from_names_and_edge_inds"], [84, 4, 1, "", "get_symmetry"], [84, 4, 1, "", "get_symmetry_name"], [84, 3, 1, "", "graph"], [84, 3, 1, "", "graph_symmetry"], [84, 4, 1, "", "has_edge"], [84, 4, 1, "", "has_node"], [84, 4, 1, "", "has_nodes"], [84, 3, 1, "", "is_arborescence"], [84, 3, 1, "", "is_template"], [84, 4, 1, "", "load_all_hdf5"], [84, 4, 1, "", "load_hdf5"], [84, 4, 1, "", "load_json"], [84, 4, 1, "", "load_mat"], [84, 4, 1, "", "make_cattr"], [84, 4, 1, "", "matches"], [84, 3, 1, "", "name"], [84, 3, 1, "", "node_names"], [84, 4, 1, "", "node_to_index"], [84, 3, 1, "", "nodes"], [84, 5, 1, "", "preview_image"], [84, 4, 1, "", "relabel_node"], [84, 4, 1, "", "relabel_nodes"], [84, 4, 1, "", "rename_skeleton"], [84, 4, 1, "", "save_all_hdf5"], [84, 4, 1, "", "save_hdf5"], [84, 4, 1, "", "save_json"], [84, 3, 1, "", "symmetric_inds"], [84, 3, 1, "", "symmetries"], [84, 3, 1, "", "symmetries_full"], [84, 3, 1, "", "symmetry_names"], [84, 4, 1, "", "to_dict"], [84, 4, 1, "", "to_json"]], "sleap.skeleton.SkeletonDecoder": [[84, 4, 1, "", "decode"], [84, 4, 1, "", "decode_preview_image"]], "sleap.skeleton.SkeletonEncoder": [[84, 4, 1, "", "encode"]], "sleap.util": [[85, 1, 1, "", "attr_to_dtype"], [85, 1, 1, "", "dict_cut"], [85, 1, 1, "", "find_files_by_suffix"], [85, 1, 1, "", "frame_list"], [85, 1, 1, "", "get_config_file"], [85, 1, 1, "", "get_package_file"], [85, 1, 1, "", "json_dumps"], [85, 1, 1, "", "json_loads"], [85, 1, 1, "", "make_scoped_dictionary"], [85, 1, 1, "", "parse_uri_path"], [85, 1, 1, "", "save_dict_to_hdf5"], [85, 1, 1, "", "uniquify"], [85, 1, 1, "", "usable_cpu_count"], [85, 1, 1, "", "weak_filename_match"]]}, "objtypes": {"0": "py:module", "1": "py:function", "2": "py:class", "3": "py:property", "4": "py:method", "5": "py:attribute"}, "objnames": {"0": ["py", "module", "Python module"], "1": ["py", "function", "Python function"], "2": ["py", "class", "Python class"], "3": ["py", "property", "Python property"], "4": ["py", "method", "Python method"], "5": ["py", "attribute", "Python attribute"]}, "titleterms": {"contributor": [0, 101], "coven": 0, "code": [0, 1], "conduct": [0, 1], "our": 0, "pledg": 0, "standard": 0, "respons": 0, "scope": 0, "enforc": 0, "attribut": 0, "contribut": 1, "sleap": [1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 87, 89, 93, 96, 100, 101, 104, 105, 106, 108, 109, 110, 112, 118], "tabl": 1, "Of": 1, "content": [1, 102], "issu": 1, "discuss": 1, "pull": 1, "request": 1, "style": 1, "guid": [1, 93], "thank": 1, "you": 1, "develop": 2, "api": 2, "info": [3, 4, 5, 6, 7, 8, 9], "align": 3, "feature_suggest": 4, "label": [5, 89, 92, 94, 115, 116], "metric": 6, "summari": 7, "trackclean": 8, "write_tracking_h5": 9, "instanc": [10, 87, 110, 119], "note": [10, 12, 45, 49, 52, 53, 55, 58, 60, 61, 62, 64, 65, 69, 71, 72, 75, 76, 77, 80, 82, 110], "io": [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34], "convert": [11, 89], "dataset": [12, 86, 89, 109, 112], "format": [13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 100], "adaptor": 13, "alphatrack": 14, "coco": 15, "csv": 16, "deeplabcut": 17, "deepposekit": 18, "dispatch": 19, "filehandl": 20, "genericjson": 21, "hdf5": 22, "labels_json": 23, "leap_matlab": 24, "main": 25, "ndx_pose": 26, "nix": 27, "sleap_analysi": 28, "text": 29, "legaci": 30, "pathutil": 31, "video": [32, 103, 110, 118], "videowrit": 33, "visual": [34, 103], "messag": 35, "nn": [36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83], "architectur": [36, 37, 38, 39, 40, 41, 42, 43, 44], "common": 36, "encoder_decod": 37, "hourglass": 38, "hrnet": 39, "leap": [40, 101], "pretrained_encod": 41, "resnet": 42, "unet": 43, "upsampl": 44, "callback": 45, "config": [46, 47, 48, 49, 50, 51], "data": [46, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 90, 95, 100, 103, 104, 105, 106, 108, 109, 110, 112, 114], "model": [47, 74, 87, 88, 107, 109, 110, 112], "optim": 48, "output": [49, 100], "training_job": 50, "util": [51, 68, 82, 85], "augment": 52, "confidence_map": 53, "dataset_op": 54, "edge_map": 55, "gener": [56, 93], "group": 57, "ident": [58, 71], "infer": [59, 72, 89, 90, 96, 104, 105, 108, 109, 110, 112, 117], "instance_centroid": 60, "instance_crop": 61, "normal": 62, "offset_regress": 63, "pipelin": 64, "provid": 65, "resiz": 66, "train": [67, 81, 87, 89, 90, 91, 95, 96, 98, 106, 109, 110, 112, 117], "eval": 69, "head": 70, "loss": 73, "paf_group": 75, "peak_find": 76, "system": [77, 100], "tracker": [78, 79], "compon": 78, "kalman": 79, "track": [80, 89, 95, 108, 110, 112, 119], "viz": 83, "skeleton": [84, 97, 118], "fly32": 86, "flies13": 86, "mice_of": 86, "mice_hc": 86, "bee": 86, "gerbil": 86, "us": [87, 100, 110, 112], "bonsai": [87, 93], "export": [87, 89, 110, 114], "instal": [87, 100, 101, 102, 108, 109, 110], "necessari": 87, "packag": [87, 102, 110], "modul": 87, "top": [87, 110], "down": [87, 110], "id": 87, "singl": 87, "connect": 87, "configur": 88, "type": 88, "hyperparamet": 88, "command": 89, "line": 89, "interfac": 89, "gui": [89, 92, 98, 102], "exampl": [89, 103, 109, 112], "file": [89, 92], "inspect": [89, 108], "render": 89, "debug": 89, "diagnost": 89, "run": [90, 93, 96, 110], "colab": [90, 109, 110], "about": 90, "move": 90, "from": 90, "creat": [91, 110, 118], "custom": 91, "profil": [91, 110], "menu": 92, "go": 92, "view": 92, "predict": [92, 94, 100, 110, 115], "help": [92, 100, 102], "applic": 92, "mous": 92, "navig": 92, "kei": 92, "select": [92, 116], "seekbar": 92, "suggest": [92, 110], "get": [93, 100, 101, 102], "better": 93, "result": [93, 100], "remot": [93, 96], "import": [94, 102, 110], "proofread": [95, 119], "method": [95, 102], "more": [95, 100, 103, 116], "The": 95, "clean": 95, "script": 95, "color": [95, 103], "palett": 95, "orient": 95, "design": 97, "tip": 97, "troubleshoot": 99, "workflow": 99, "i": 100, "can": 100, "t": 100, "comput": 100, "without": 100, "gpu": [100, 102], "what": 100, "alreadi": 100, "have": 100, "cuda": 100, "set": 100, "up": 100, "my": 100, "usag": [100, 112], "how": 100, "do": 100, "doe": 100, "need": 100, "particular": 100, "strang": 100, "where": 100, "pose": [100, 101], "appear": 100, "correct": 100, "shift": 100, "rel": 100, "imag": 100, "out": 100, "ve": 100, "found": 100, "bug": 100, "anoth": 100, "problem": 100, "just": 100, "talk": 100, "someon": 100, "improv": 100, "social": 101, "estim": 101, "anim": 101, "featur": 101, "some": 101, "quick": 101, "learn": 101, "refer": 101, "contact": 101, "licens": 101, "document": 101, "manag": 102, "miniforg": 102, "recommend": 102, "miniconda": 102, "test": [102, 108], "thing": 102, "ar": 102, "work": 102, "support": 102, "upgrad": 102, "uninstal": 102, "analysi": [103, 112, 114], "load": 103, "fill": 103, "miss": 103, "valu": 103, "thorax": 103, "movement": 103, "across": [103, 119], "advanc": [103, 112], "x": 103, "y": 103, "dynam": 103, "veloc": 103, "fly": 103, "0": 103, "magnitud": 103, "speed": 103, "find": 103, "covari": 103, "between": 103, "1": [103, 104, 105, 106, 108], "cluster": 103, "structur": [104, 112], "setup": [104, 105, 106, 108], "2": [104, 105, 106, 108], "interact": [105, 106, 112], "realtim": [105, 112], "3": [105, 106, 108], "4": [105, 106, 108], "perform": 105, "resum": [106, 112], "job": [106, 110], "5": 106, "continu": 106, "evalu": [107, 112], "post": [108, 112], "save": [108, 116], "an": [109, 112], "download": 109, "sampl": 109, "your": [110, 112], "own": [110, 112], "googl": [110, 112], "drive": [110, 112], "upload": 110, "mount": 110, "process": 110, "frame": [110, 116, 119], "notebook": 112, "basic": 112, "overview": 113, "For": 114, "matlab": 114, "python": 114, "assist": 115, "review": 115, "fix": 115, "initi": 116, "first": 116, "option": 117, "start": [117, 118], "project": 118, "open": 118, "tutori": 120}, "envversion": {"sphinx.domains.c": 2, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 8, "sphinx.domains.index": 1, "sphinx.domains.javascript": 2, "sphinx.domains.math": 2, "sphinx.domains.python": 3, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx": 57}, "alltitles": {"Contributor Covenant Code of Conduct": [[0, "contributor-covenant-code-of-conduct"]], "Our Pledge": [[0, "our-pledge"]], "Our Standards": [[0, "our-standards"]], "Our Responsibilities": [[0, "our-responsibilities"]], "Scope": [[0, "scope"]], "Enforcement": [[0, "enforcement"]], "Attribution": [[0, "attribution"]], "Contributing to SLEAP": [[1, "contributing-to-sleap"]], "Table Of Contents": [[1, "table-of-contents"]], "Code of Conduct": [[1, "code-of-conduct"]], "Contributing": [[1, "contributing"]], "Issues": [[1, "issues"]], "Discussions": [[1, "discussions"]], "Pull Requests": [[1, "pull-requests"]], "Style Guides": [[1, "style-guides"]], "Thank you for contributing to SLEAP!": [[1, "thank-you-for-contributing-to-sleap"]], "Developer API": [[2, "developer-api"]], "sleap.info.align": [[3, "module-sleap.info.align"]], "sleap.info.feature_suggestions": [[4, "module-sleap.info.feature_suggestions"]], "sleap.info.labels": [[5, "module-sleap.info.labels"]], "sleap.info.metrics": [[6, "module-sleap.info.metrics"]], "sleap.info.summary": [[7, "module-sleap.info.summary"]], "sleap.info.trackcleaner": [[8, "module-sleap.info.trackcleaner"]], "sleap.info.write_tracking_h5": [[9, "module-sleap.info.write_tracking_h5"]], "sleap.instance": [[10, "module-sleap.instance"]], "Notes": [[10, null], [10, null], [10, null], [10, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [12, null], [45, null], [49, null], [52, null], [53, null], [53, null], [53, null], [53, null], [53, null], [55, null], [58, null], [60, null], [61, null], [61, null], [62, null], [62, null], [62, null], [64, null], [64, null], [64, null], [64, null], [64, null], [64, null], [65, null], [69, null], [69, null], [71, null], [72, null], [72, null], [72, null], [72, null], [72, null], [72, null], [72, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [75, null], [76, null], [76, null], [77, null], [77, null], [80, null], [82, null]], "sleap.io.convert": [[11, "module-sleap.io.convert"]], "sleap.io.dataset": [[12, "module-sleap.io.dataset"]], "sleap.io.format.adaptor": [[13, "module-sleap.io.format.adaptor"]], "sleap.io.format.alphatracker": [[14, "module-sleap.io.format.alphatracker"]], "sleap.io.format.coco": [[15, "module-sleap.io.format.coco"]], "sleap.io.format.csv": [[16, "module-sleap.io.format.csv"]], "sleap.io.format.deeplabcut": [[17, "module-sleap.io.format.deeplabcut"]], "sleap.io.format.deepposekit": [[18, "module-sleap.io.format.deepposekit"]], "sleap.io.format.dispatch": [[19, "module-sleap.io.format.dispatch"]], "sleap.io.format.filehandle": [[20, "module-sleap.io.format.filehandle"]], "sleap.io.format.genericjson": [[21, "module-sleap.io.format.genericjson"]], "sleap.io.format.hdf5": [[22, "module-sleap.io.format.hdf5"]], "sleap.io.format.labels_json": [[23, "module-sleap.io.format.labels_json"]], "sleap.io.format.leap_matlab": [[24, "module-sleap.io.format.leap_matlab"]], "sleap.io.format.main": [[25, "module-sleap.io.format.main"]], "sleap.io.format.ndx_pose": [[26, "module-sleap.io.format.ndx_pose"]], "sleap.io.format.nix": [[27, "module-sleap.io.format.nix"]], "sleap.io.format.sleap_analysis": [[28, "module-sleap.io.format.sleap_analysis"]], "sleap.io.format.text": [[29, "module-sleap.io.format.text"]], "sleap.io.legacy": [[30, "module-sleap.io.legacy"]], "sleap.io.pathutils": [[31, "module-sleap.io.pathutils"]], "sleap.io.video": [[32, "module-sleap.io.video"]], "sleap.io.videowriter": [[33, "module-sleap.io.videowriter"]], "sleap.io.visuals": [[34, "module-sleap.io.visuals"]], "sleap.message": [[35, "module-sleap.message"]], "sleap.nn.architectures.common": [[36, "module-sleap.nn.architectures.common"]], "sleap.nn.architectures.encoder_decoder": [[37, "module-sleap.nn.architectures.encoder_decoder"]], "sleap.nn.architectures.hourglass": [[38, "module-sleap.nn.architectures.hourglass"]], "sleap.nn.architectures.hrnet": [[39, "module-sleap.nn.architectures.hrnet"]], "sleap.nn.architectures.leap": [[40, "module-sleap.nn.architectures.leap"]], "sleap.nn.architectures.pretrained_encoders": [[41, "module-sleap.nn.architectures.pretrained_encoders"]], "sleap.nn.architectures.resnet": [[42, "module-sleap.nn.architectures.resnet"]], "sleap.nn.architectures.unet": [[43, "module-sleap.nn.architectures.unet"]], "sleap.nn.architectures.upsampling": [[44, "module-sleap.nn.architectures.upsampling"]], "sleap.nn.callbacks": [[45, "module-sleap.nn.callbacks"]], "sleap.nn.config.data": [[46, "module-sleap.nn.config.data"]], "sleap.nn.config.model": [[47, "module-sleap.nn.config.model"]], "sleap.nn.config.optimization": [[48, "module-sleap.nn.config.optimization"]], "sleap.nn.config.outputs": [[49, "module-sleap.nn.config.outputs"]], "sleap.nn.config.training_job": [[50, "module-sleap.nn.config.training_job"]], "sleap.nn.config.utils": [[51, "module-sleap.nn.config.utils"]], "sleap.nn.data.augmentation": [[52, "module-sleap.nn.data.augmentation"]], "sleap.nn.data.confidence_maps": [[53, "module-sleap.nn.data.confidence_maps"]], "sleap.nn.data.dataset_ops": [[54, "module-sleap.nn.data.dataset_ops"]], "sleap.nn.data.edge_maps": [[55, "module-sleap.nn.data.edge_maps"]], "sleap.nn.data.general": [[56, "module-sleap.nn.data.general"]], "sleap.nn.data.grouping": [[57, "module-sleap.nn.data.grouping"]], "sleap.nn.data.identity": [[58, "module-sleap.nn.data.identity"]], "sleap.nn.data.inference": [[59, "module-sleap.nn.data.inference"]], "sleap.nn.data.instance_centroids": [[60, "module-sleap.nn.data.instance_centroids"]], "sleap.nn.data.instance_cropping": [[61, "module-sleap.nn.data.instance_cropping"]], "sleap.nn.data.normalization": [[62, "module-sleap.nn.data.normalization"]], "sleap.nn.data.offset_regression": [[63, "module-sleap.nn.data.offset_regression"]], "sleap.nn.data.pipelines": [[64, "module-sleap.nn.data.pipelines"]], "sleap.nn.data.providers": [[65, "module-sleap.nn.data.providers"]], "sleap.nn.data.resizing": [[66, "module-sleap.nn.data.resizing"]], "sleap.nn.data.training": [[67, "module-sleap.nn.data.training"]], "sleap.nn.data.utils": [[68, "module-sleap.nn.data.utils"]], "sleap.nn.evals": [[69, "module-sleap.nn.evals"]], "sleap.nn.heads": [[70, "module-sleap.nn.heads"]], "sleap.nn.identity": [[71, "module-sleap.nn.identity"]], "sleap.nn.inference": [[72, "module-sleap.nn.inference"]], "sleap.nn.losses": [[73, "module-sleap.nn.losses"]], "sleap.nn.model": [[74, "module-sleap.nn.model"]], "sleap.nn.paf_grouping": [[75, "module-sleap.nn.paf_grouping"]], "sleap.nn.peak_finding": [[76, "module-sleap.nn.peak_finding"]], "sleap.nn.system": [[77, "module-sleap.nn.system"]], "sleap.nn.tracker.components": [[78, "module-sleap.nn.tracker.components"]], "sleap.nn.tracker.kalman": [[79, "module-sleap.nn.tracker.kalman"]], "sleap.nn.tracking": [[80, "module-sleap.nn.tracking"]], "sleap.nn.training": [[81, "module-sleap.nn.training"]], "sleap.nn.utils": [[82, "module-sleap.nn.utils"]], "sleap.nn.viz": [[83, "module-sleap.nn.viz"]], "sleap.skeleton": [[84, "module-sleap.skeleton"]], "sleap.util": [[85, "module-sleap.util"]], "Datasets": [[86, "datasets"]], "fly32": [[86, "fly32"]], "flies13": [[86, "flies13"]], "mice_of": [[86, "mice-of"]], "mice_hc": [[86, "mice-hc"]], "bees": [[86, "bees"]], "gerbils": [[86, "gerbils"]], "Using Bonsai with SLEAP": [[87, "using-bonsai-with-sleap"]], "Exporting a SLEAP trained model": [[87, "exporting-a-sleap-trained-model"]], "Installing Bonsai and necessary packages": [[87, "installing-bonsai-and-necessary-packages"]], "Using Bonsai SLEAP modules": [[87, "using-bonsai-sleap-modules"]], "Top-down model": [[87, "top-down-model"]], "Top-Down-ID model": [[87, "top-down-id-model"]], "Single instance model": [[87, "single-instance-model"]], "Connecting the modules": [[87, "connecting-the-modules"]], "Configuring models": [[88, "configuring-models"]], "Model type": [[88, "model-type"]], "Hyperparameters": [[88, "hyperparameters"]], "Command line interfaces": [[89, "command-line-interfaces"]], "GUI": [[89, "gui"], [92, "gui"]], "sleap-label": [[89, "sleap-label"]], "Training": [[89, "training"]], "sleap-train": [[89, "sleap-train"]], "sleap-export": [[89, "sleap-export"]], "Inference and Tracking": [[89, "inference-and-tracking"]], "sleap-track": [[89, "sleap-track"]], "Examples:": [[89, "examples"]], "Dataset files": [[89, "dataset-files"]], "sleap-convert": [[89, "sleap-convert"]], "sleap-inspect": [[89, "sleap-inspect"]], "Rendering": [[89, "rendering"]], "sleap-render": [[89, "sleap-render"]], "Debugging": [[89, "debugging"]], "sleap-diagnostic": [[89, "sleap-diagnostic"]], "Run training and inference on Colab": [[90, "run-training-and-inference-on-colab"]], "About Colab": [[90, "about-colab"]], "Moving Data to/from Colab": [[90, "moving-data-to-from-colab"]], "Creating a custom training profile": [[91, "creating-a-custom-training-profile"]], "Menus": [[92, "menus"]], "File": [[92, "file"]], "Go": [[92, "go"]], "View": [[92, "view"]], "Labels": [[92, "labels"]], "Predict": [[92, "predict"]], "Help": [[92, "help"], [100, "help"]], "Application GUI": [[92, "application-gui"]], "Mouse": [[92, "mouse"]], "Navigation Keys": [[92, "navigation-keys"]], "Selection Keys": [[92, "selection-keys"]], "Seekbar": [[92, "seekbar"]], "Labeling Suggestions": [[92, "labeling-suggestions"]], "Guides": [[93, "guides"]], "General": [[93, "general"]], "Getting better results": [[93, "getting-better-results"]], "Running remotely": [[93, "running-remotely"]], "SLEAP with Bonsai": [[93, "sleap-with-bonsai"]], "Importing predictions for labeling": [[94, "importing-predictions-for-labeling"]], "Tracking and proofreading": [[95, "tracking-and-proofreading"]], "Tracking methods": [[95, "tracking-methods"]], "More training data?": [[95, "more-training-data"]], "The \u201ctrack cleaning\u201d script": [[95, "the-track-cleaning-script"]], "Color palettes": [[95, "color-palettes"]], "Proofreading": [[95, "id1"]], "Orientation": [[95, "orientation"]], "Running SLEAP remotely": [[96, "running-sleap-remotely"]], "Remote training": [[96, "remote-training"]], "Remote inference": [[96, "remote-inference"]], "Skeleton design": [[97, "skeleton-design"]], "Tips": [[97, "tips"]], "Training with GUI": [[98, "training-with-gui"]], "Troubleshooting workflows": [[99, "troubleshooting-workflows"]], "Installation": [[100, "installation"], [102, "installation"]], "I can\u2019t get SLEAP to install!": [[100, "i-cant-get-sleap-to-install"]], "Can I install it on a computer without a GPU?": [[100, "can-i-install-it-on-a-computer-without-a-gpu"]], "What if I already have CUDA set up on my system?": [[100, "what-if-i-already-have-cuda-set-up-on-my-system"]], "Usage": [[100, "usage"]], "How do I use SLEAP?": [[100, "how-do-i-use-sleap"]], "Does my data need to be in a particular format?": [[100, "does-my-data-need-to-be-in-a-particular-format"]], "I get strange results where the poses appear to be correct but shifted relative to the image.": [[100, "i-get-strange-results-where-the-poses-appear-to-be-correct-but-shifted-relative-to-the-image"]], "How do I get predictions out?": [[100, "how-do-i-get-predictions-out"]], "What do I do with the output of SLEAP?": [[100, "what-do-i-do-with-the-output-of-sleap"]], "Getting more help": [[100, "getting-more-help"]], "I\u2019ve found a bug or have another problem!": [[100, "ive-found-a-bug-or-have-another-problem"]], "Can I just talk to someone?": [[100, "can-i-just-talk-to-someone"]], "Improving SLEAP": [[100, "improving-sleap"]], "How can I help improve SLEAP?": [[100, "how-can-i-help-improve-sleap"]], "What is usage data?": [[100, "what-is-usage-data"]], "Social LEAP Estimates Animal Poses (SLEAP)": [[101, "social-leap-estimates-animal-poses-sleap"]], "Features": [[101, "features"]], "Get some SLEAP": [[101, "get-some-sleap"]], "Quick install": [[101, "quick-install"]], "Learn to SLEAP": [[101, "learn-to-sleap"]], "References": [[101, "references"]], "Contact": [[101, "contact"]], "Contributors": [[101, "contributors"]], "License": [[101, "license"]], "SLEAP Documentation": [[101, "sleap-documentation"]], "Contents": [[102, "contents"]], "Package Manager": [[102, "package-manager"]], "Miniforge (recommended)": [[102, "miniforge-recommended"]], "Miniconda": [[102, "miniconda"]], "Installation methods": [[102, "installation-methods"]], "Testing that things are working": [[102, "testing-that-things-are-working"]], "GUI support": [[102, "gui-support"]], "Importing": [[102, "importing"]], "GPU support": [[102, "gpu-support"]], "Upgrading and uninstalling": [[102, "upgrading-and-uninstalling"]], "Getting help": [[102, "getting-help"]], "Analysis examples": [[103, "analysis-examples"], [112, "id4"]], "Example analysis data": [[103, "example-analysis-data"]], "Loading the data": [[103, "loading-the-data"]], "Fill missing values": [[103, "fill-missing-values"]], "Visualize thorax movement across video": [[103, "visualize-thorax-movement-across-video"]], "More advanced visualizations": [[103, "more-advanced-visualizations"]], "Visualizing thorax x-y dynamics and velocity for fly 0": [[103, "visualizing-thorax-x-y-dynamics-and-velocity-for-fly-0"]], "Visualize thorax colored by magnitude of fly speed": [[103, "visualize-thorax-colored-by-magnitude-of-fly-speed"]], "Find covariance in thorax velocities between fly-0 and fly-1": [[103, "find-covariance-in-thorax-velocities-between-fly-0-and-fly-1"]], "Clustering": [[103, "clustering"]], "Data structures": [[104, "data-structures"], [112, "id5"]], "1. Setup SLEAP and data": [[104, "setup-sleap-and-data"]], "2. Data structures and inference": [[104, "data-structures-and-inference"]], "Interactive and realtime inference": [[105, "interactive-and-realtime-inference"], [112, "id8"]], "1. Setup SLEAP": [[105, "setup-sleap"], [106, "setup-sleap"], [108, "setup-sleap"]], "2. Setup data": [[105, "setup-data"], [108, "setup-data"]], "3. Interactive inference": [[105, "interactive-inference"]], "4. Realtime performance": [[105, "realtime-performance"]], "Interactive and resumable training": [[106, "interactive-and-resumable-training"], [112, "id7"]], "2. Setup training data": [[106, "setup-training-data"]], "3. Setup training job": [[106, "setup-training-job"]], "4. Training": [[106, "training"]], "5. Continuing training": [[106, "continuing-training"]], "Model evaluation": [[107, "model-evaluation"], [112, "id9"]], "Post-inference tracking": [[108, "post-inference-tracking"], [112, "id6"]], "Install": [[108, "install"]], "Test": [[108, "test"]], "3. Track": [[108, "track"]], "4. Inspect and save": [[108, "inspect-and-save"]], "Training and inference on an example dataset": [[109, "training-and-inference-on-an-example-dataset"], [112, "id2"]], "Install SLEAP": [[109, "install-sleap"], [110, "install-sleap"]], "Download sample training data into Colab": [[109, "download-sample-training-data-into-colab"]], "Train models": [[109, "train-models"]], "Inference": [[109, "inference"], [117, "inference"]], "Training and inference on your own data using Google Drive": [[110, "training-and-inference-on-your-own-data-using-google-drive"], [112, "id3"]], "Import training data into Colab with Google Drive": [[110, "import-training-data-into-colab-with-google-drive"]], "Create and export the training job package": [[110, "create-and-export-the-training-job-package"]], "Upload training job package to Google Drive": [[110, "upload-training-job-package-to-google-drive"]], "Mount your Google Drive": [[110, "mount-your-google-drive"]], "Train a model": [[110, "train-a-model"]], "Note on training profiles": [[110, "note-on-training-profiles"]], "Note on training process": [[110, "note-on-training-process"]], "Run inference to predict instances": [[110, "run-inference-to-predict-instances"]], "Predicting instances in suggested frames": [[110, "predicting-instances-in-suggested-frames"]], "Predicting and tracking instances in uploaded video": [[110, "predicting-and-tracking-instances-in-uploaded-video"]], "Inference with top-down models": [[110, "inference-with-top-down-models"]], "Notebooks": [[112, "notebooks"]], "Basic usage": [[112, "basic-usage"]], "Advanced SLEAPing": [[112, "advanced-sleaping"]], "Overview": [[113, "overview"]], "Export Data For Analysis": [[114, "export-data-for-analysis"]], "MATLAB": [[114, "matlab"]], "Python": [[114, "python"]], "Prediction-assisted labeling": [[115, "prediction-assisted-labeling"]], "Reviewing and fixing predictions": [[115, "reviewing-and-fixing-predictions"]], "Initial Labeling": [[116, "initial-labeling"]], "Selecting frames to label": [[116, "selecting-frames-to-label"]], "Labeling the first frame": [[116, "labeling-the-first-frame"]], "Saving": [[116, "saving"]], "Labeling more frames": [[116, "labeling-more-frames"]], "Training and Inference": [[117, "training-and-inference"]], "Training Options": [[117, "training-options"]], "Start Training": [[117, "start-training"]], "Creating a project": [[118, "creating-a-project"]], "Starting SLEAP": [[118, "starting-sleap"]], "Opening a video": [[118, "opening-a-video"]], "Creating a Skeleton": [[118, "creating-a-skeleton"]], "Tracking instances across frames": [[119, "tracking-instances-across-frames"]], "Track proofreading": [[119, "track-proofreading"]], "Tutorial": [[120, "tutorial"]]}, "indexentries": {"align_instance_points() (in module sleap.info.align)": [[3, "sleap.info.align.align_instance_points"]], "align_instances() (in module sleap.info.align)": [[3, "sleap.info.align.align_instances"]], "align_instances_on_most_stable() (in module sleap.info.align)": [[3, "sleap.info.align.align_instances_on_most_stable"]], "get_instances_points() (in module sleap.info.align)": [[3, "sleap.info.align.get_instances_points"]], "get_mean_and_std_for_points() (in module sleap.info.align)": [[3, "sleap.info.align.get_mean_and_std_for_points"]], "get_most_stable_node_pair() (in module sleap.info.align)": [[3, "sleap.info.align.get_most_stable_node_pair"]], "get_stable_node_pairs() (in module sleap.info.align)": [[3, "sleap.info.align.get_stable_node_pairs"]], "get_template_points_array() (in module sleap.info.align)": [[3, "sleap.info.align.get_template_points_array"]], "module": [[3, "module-sleap.info.align"], [4, "module-sleap.info.feature_suggestions"], [5, "module-sleap.info.labels"], [6, "module-sleap.info.metrics"], [7, "module-sleap.info.summary"], [8, "module-sleap.info.trackcleaner"], [9, "module-sleap.info.write_tracking_h5"], [10, "module-sleap.instance"], [11, "module-sleap.io.convert"], [12, "module-sleap.io.dataset"], [13, "module-sleap.io.format.adaptor"], [14, "module-sleap.io.format.alphatracker"], [15, "module-sleap.io.format.coco"], [16, "module-sleap.io.format.csv"], [17, "module-sleap.io.format.deeplabcut"], [18, "module-sleap.io.format.deepposekit"], [19, "module-sleap.io.format.dispatch"], [20, "module-sleap.io.format.filehandle"], [21, "module-sleap.io.format.genericjson"], [22, "module-sleap.io.format.hdf5"], [23, "module-sleap.io.format.labels_json"], [24, "module-sleap.io.format.leap_matlab"], [25, "module-sleap.io.format.main"], [26, "module-sleap.io.format.ndx_pose"], [27, "module-sleap.io.format.nix"], [28, "module-sleap.io.format.sleap_analysis"], [29, "module-sleap.io.format.text"], [30, "module-sleap.io.legacy"], [31, "module-sleap.io.pathutils"], [32, "module-sleap.io.video"], [33, "module-sleap.io.videowriter"], [34, "module-sleap.io.visuals"], [35, "module-sleap.message"], [36, "module-sleap.nn.architectures.common"], [37, "module-sleap.nn.architectures.encoder_decoder"], [38, "module-sleap.nn.architectures.hourglass"], [39, "module-sleap.nn.architectures.hrnet"], [40, "module-sleap.nn.architectures.leap"], [41, "module-sleap.nn.architectures.pretrained_encoders"], [42, "module-sleap.nn.architectures.resnet"], [43, "module-sleap.nn.architectures.unet"], [44, "module-sleap.nn.architectures.upsampling"], [45, "module-sleap.nn.callbacks"], [46, "module-sleap.nn.config.data"], [47, "module-sleap.nn.config.model"], [48, "module-sleap.nn.config.optimization"], [49, "module-sleap.nn.config.outputs"], [50, "module-sleap.nn.config.training_job"], [51, "module-sleap.nn.config.utils"], [52, "module-sleap.nn.data.augmentation"], [53, "module-sleap.nn.data.confidence_maps"], [54, "module-sleap.nn.data.dataset_ops"], [55, "module-sleap.nn.data.edge_maps"], [56, "module-sleap.nn.data.general"], [57, "module-sleap.nn.data.grouping"], [58, "module-sleap.nn.data.identity"], [59, "module-sleap.nn.data.inference"], [60, "module-sleap.nn.data.instance_centroids"], [61, "module-sleap.nn.data.instance_cropping"], [62, "module-sleap.nn.data.normalization"], [63, "module-sleap.nn.data.offset_regression"], [64, "module-sleap.nn.data.pipelines"], [65, "module-sleap.nn.data.providers"], [66, "module-sleap.nn.data.resizing"], [67, "module-sleap.nn.data.training"], [68, "module-sleap.nn.data.utils"], [69, "module-sleap.nn.evals"], [70, "module-sleap.nn.heads"], [71, "module-sleap.nn.identity"], [72, "module-sleap.nn.inference"], [73, "module-sleap.nn.losses"], [74, "module-sleap.nn.model"], [75, "module-sleap.nn.paf_grouping"], [76, "module-sleap.nn.peak_finding"], [77, "module-sleap.nn.system"], [78, "module-sleap.nn.tracker.components"], [79, "module-sleap.nn.tracker.kalman"], [80, "module-sleap.nn.tracking"], [81, "module-sleap.nn.training"], [82, "module-sleap.nn.utils"], [83, "module-sleap.nn.viz"], [84, "module-sleap.skeleton"], [85, "module-sleap.util"]], "sleap.info.align": [[3, "module-sleap.info.align"]], "framegroupset (class in sleap.info.feature_suggestions)": [[4, "sleap.info.feature_suggestions.FrameGroupSet"]], "frameitem (class in sleap.info.feature_suggestions)": [[4, "sleap.info.feature_suggestions.FrameItem"]], "itemstack (class in sleap.info.feature_suggestions)": [[4, "sleap.info.feature_suggestions.ItemStack"]], "parallelfeaturepipeline (class in sleap.info.feature_suggestions)": [[4, "sleap.info.feature_suggestions.ParallelFeaturePipeline"]], "all_items (sleap.info.feature_suggestions.framegroupset property)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.all_items"]], "append_to_group() (sleap.info.feature_suggestions.framegroupset method)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.append_to_group"]], "brisk_bag_of_features() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.brisk_bag_of_features"]], "current_groupset (sleap.info.feature_suggestions.itemstack property)": [[4, "sleap.info.feature_suggestions.ItemStack.current_groupset"]], "data (sleap.info.feature_suggestions.itemstack attribute)": [[4, "sleap.info.feature_suggestions.ItemStack.data"]], "extend_group_items() (sleap.info.feature_suggestions.framegroupset method)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.extend_group_items"]], "extend_ownership() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.extend_ownership"]], "flatten() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.flatten"]], "get() (sleap.info.feature_suggestions.parallelfeaturepipeline method)": [[4, "sleap.info.feature_suggestions.ParallelFeaturePipeline.get"]], "get_all_items_from_group() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.get_all_items_from_group"]], "get_item_data() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.get_item_data"]], "get_item_data_idxs() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.get_item_data_idxs"]], "get_item_group() (sleap.info.feature_suggestions.framegroupset method)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.get_item_group"]], "get_raw_images() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.get_raw_images"]], "group_data (sleap.info.feature_suggestions.framegroupset attribute)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.group_data"]], "group_sets (sleap.info.feature_suggestions.itemstack attribute)": [[4, "sleap.info.feature_suggestions.ItemStack.group_sets"]], "groups (sleap.info.feature_suggestions.framegroupset property)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.groups"]], "groupset_data (sleap.info.feature_suggestions.framegroupset attribute)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.groupset_data"]], "hog_bag_of_features() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.hog_bag_of_features"]], "item_group (sleap.info.feature_suggestions.framegroupset attribute)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.item_group"]], "items (sleap.info.feature_suggestions.itemstack attribute)": [[4, "sleap.info.feature_suggestions.ItemStack.items"]], "kmeans() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.kmeans"]], "make() (sleap.info.feature_suggestions.parallelfeaturepipeline class method)": [[4, "sleap.info.feature_suggestions.ParallelFeaturePipeline.make"]], "make_sample_group() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.make_sample_group"]], "meta (sleap.info.feature_suggestions.itemstack attribute)": [[4, "sleap.info.feature_suggestions.ItemStack.meta"]], "method (sleap.info.feature_suggestions.framegroupset attribute)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.method"]], "ownership (sleap.info.feature_suggestions.itemstack attribute)": [[4, "sleap.info.feature_suggestions.ItemStack.ownership"]], "pca() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.pca"]], "run() (sleap.info.feature_suggestions.parallelfeaturepipeline class method)": [[4, "sleap.info.feature_suggestions.ParallelFeaturePipeline.run"]], "sample() (sleap.info.feature_suggestions.framegroupset method)": [[4, "sleap.info.feature_suggestions.FrameGroupSet.sample"]], "sample_groups() (sleap.info.feature_suggestions.itemstack method)": [[4, "sleap.info.feature_suggestions.ItemStack.sample_groups"]], "sleap.info.feature_suggestions": [[4, "module-sleap.info.feature_suggestions"]], "tuples_to_suggestions() (sleap.info.feature_suggestions.parallelfeaturepipeline class method)": [[4, "sleap.info.feature_suggestions.ParallelFeaturePipeline.tuples_to_suggestions"]], "sleap.info.labels": [[5, "module-sleap.info.labels"]], "calculate_pairwise_cost() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.calculate_pairwise_cost"]], "compare_instance_lists() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.compare_instance_lists"]], "list_points_array() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.list_points_array"]], "match_instance_lists() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.match_instance_lists"]], "match_instance_lists_nodewise() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.match_instance_lists_nodewise"]], "matched_instance_distances() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.matched_instance_distances"]], "nodeless_point_dist() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.nodeless_point_dist"]], "point_dist() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.point_dist"]], "point_match_count() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.point_match_count"]], "point_nonmatch_count() (in module sleap.info.metrics)": [[6, "sleap.info.metrics.point_nonmatch_count"]], "sleap.info.metrics": [[6, "module-sleap.info.metrics"]], "statisticseries (class in sleap.info.summary)": [[7, "sleap.info.summary.StatisticSeries"]], "get_instance_score_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_instance_score_series"]], "get_point_count_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_point_count_series"]], "get_point_displacement_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_point_displacement_series"]], "get_point_score_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_point_score_series"]], "get_primary_point_displacement_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_primary_point_displacement_series"]], "get_tracking_score_series() (sleap.info.summary.statisticseries method)": [[7, "sleap.info.summary.StatisticSeries.get_tracking_score_series"]], "sleap.info.summary": [[7, "module-sleap.info.summary"]], "fit_tracks() (in module sleap.info.trackcleaner)": [[8, "sleap.info.trackcleaner.fit_tracks"]], "sleap.info.trackcleaner": [[8, "module-sleap.info.trackcleaner"]], "get_edges_as_np_strings() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.get_edges_as_np_strings"]], "get_nodes_as_np_strings() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.get_nodes_as_np_strings"]], "get_occupancy_and_points_matrices() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.get_occupancy_and_points_matrices"]], "get_tracks_as_np_strings() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.get_tracks_as_np_strings"]], "main() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.main"]], "remove_empty_tracks_from_matrices() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.remove_empty_tracks_from_matrices"]], "sleap.info.write_tracking_h5": [[9, "module-sleap.info.write_tracking_h5"]], "write_csv_file() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.write_csv_file"]], "write_occupancy_file() (in module sleap.info.write_tracking_h5)": [[9, "sleap.info.write_tracking_h5.write_occupancy_file"]], "instance (class in sleap.instance)": [[10, "sleap.instance.Instance"]], "instanceslist (class in sleap.instance)": [[10, "sleap.instance.InstancesList"]], "labeledframe (class in sleap.instance)": [[10, "sleap.instance.LabeledFrame"]], "point (class in sleap.instance)": [[10, "sleap.instance.Point"]], "pointarray (class in sleap.instance)": [[10, "sleap.instance.PointArray"]], "predictedinstance (class in sleap.instance)": [[10, "sleap.instance.PredictedInstance"]], "predictedpoint (class in sleap.instance)": [[10, "sleap.instance.PredictedPoint"]], "predictedpointarray (class in sleap.instance)": [[10, "sleap.instance.PredictedPointArray"]], "track (class in sleap.instance)": [[10, "sleap.instance.Track"]], "append() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.append"]], "bounding_box (sleap.instance.instance property)": [[10, "sleap.instance.Instance.bounding_box"]], "centroid (sleap.instance.instance property)": [[10, "sleap.instance.Instance.centroid"]], "clear() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.clear"]], "complex_frame_merge() (sleap.instance.labeledframe class method)": [[10, "sleap.instance.LabeledFrame.complex_frame_merge"]], "complex_merge_between() (sleap.instance.labeledframe class method)": [[10, "sleap.instance.LabeledFrame.complex_merge_between"]], "copy() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.copy"]], "dtype (sleap.instance.point attribute)": [[10, "sleap.instance.Point.dtype"]], "dtype (sleap.instance.predictedpoint attribute)": [[10, "sleap.instance.PredictedPoint.dtype"]], "extend() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.extend"]], "fill_missing() (sleap.instance.instance method)": [[10, "sleap.instance.Instance.fill_missing"]], "find() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.find"]], "frame_idx (sleap.instance.instance property)": [[10, "sleap.instance.Instance.frame_idx"]], "from_array() (sleap.instance.pointarray class method)": [[10, "sleap.instance.PointArray.from_array"]], "from_arrays() (sleap.instance.predictedinstance class method)": [[10, "sleap.instance.PredictedInstance.from_arrays"]], "from_instance() (sleap.instance.predictedinstance class method)": [[10, "sleap.instance.PredictedInstance.from_instance"]], "from_numpy() (sleap.instance.instance class method)": [[10, "sleap.instance.Instance.from_numpy"]], "from_numpy() (sleap.instance.predictedinstance class method)": [[10, "sleap.instance.PredictedInstance.from_numpy"]], "from_point() (sleap.instance.predictedpoint class method)": [[10, "sleap.instance.PredictedPoint.from_point"]], "from_pointsarray() (sleap.instance.instance class method)": [[10, "sleap.instance.Instance.from_pointsarray"]], "from_pointsarray() (sleap.instance.predictedinstance class method)": [[10, "sleap.instance.PredictedInstance.from_pointsarray"]], "get_points_array() (sleap.instance.instance method)": [[10, "sleap.instance.Instance.get_points_array"]], "has_predicted_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.has_predicted_instances"]], "has_tracked_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.has_tracked_instances"]], "has_user_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.has_user_instances"]], "image (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.image"]], "index() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.index"]], "insert() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.insert"]], "insert() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.insert"]], "instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.instances"]], "instances_to_show (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.instances_to_show"]], "isnan() (sleap.instance.point method)": [[10, "sleap.instance.Point.isnan"]], "labeled_frame (sleap.instance.instanceslist property)": [[10, "sleap.instance.InstancesList.labeled_frame"]], "make_default() (sleap.instance.pointarray class method)": [[10, "sleap.instance.PointArray.make_default"]], "make_instance_cattr() (in module sleap.instance)": [[10, "sleap.instance.make_instance_cattr"]], "matches() (sleap.instance.instance method)": [[10, "sleap.instance.Instance.matches"]], "matches() (sleap.instance.track method)": [[10, "sleap.instance.Track.matches"]], "merge_frames() (sleap.instance.labeledframe static method)": [[10, "sleap.instance.LabeledFrame.merge_frames"]], "midpoint (sleap.instance.instance property)": [[10, "sleap.instance.Instance.midpoint"]], "n_predicted_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.n_predicted_instances"]], "n_tracked_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.n_tracked_instances"]], "n_user_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.n_user_instances"]], "n_visible_points (sleap.instance.instance property)": [[10, "sleap.instance.Instance.n_visible_points"]], "nodes (sleap.instance.instance property)": [[10, "sleap.instance.Instance.nodes"]], "nodes_points (sleap.instance.instance property)": [[10, "sleap.instance.Instance.nodes_points"]], "numpy() (sleap.instance.instance method)": [[10, "sleap.instance.Instance.numpy"]], "numpy() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.numpy"]], "numpy() (sleap.instance.point method)": [[10, "sleap.instance.Point.numpy"]], "plot() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.plot"]], "plot_predicted() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.plot_predicted"]], "points (sleap.instance.instance property)": [[10, "sleap.instance.Instance.points"]], "points_and_scores_array (sleap.instance.predictedinstance property)": [[10, "sleap.instance.PredictedInstance.points_and_scores_array"]], "points_array (sleap.instance.instance property)": [[10, "sleap.instance.Instance.points_array"]], "pop() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.pop"]], "predicted_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.predicted_instances"]], "remove() (sleap.instance.instanceslist method)": [[10, "sleap.instance.InstancesList.remove"]], "remove_empty_instances() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.remove_empty_instances"]], "remove_untracked() (sleap.instance.labeledframe method)": [[10, "sleap.instance.LabeledFrame.remove_untracked"]], "scores (sleap.instance.predictedinstance property)": [[10, "sleap.instance.PredictedInstance.scores"]], "sleap.instance": [[10, "module-sleap.instance"]], "to_array() (sleap.instance.predictedpointarray class method)": [[10, "sleap.instance.PredictedPointArray.to_array"]], "tracked_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.tracked_instances"]], "training_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.training_instances"]], "transform_points() (sleap.instance.instance method)": [[10, "sleap.instance.Instance.transform_points"]], "unused_predictions (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.unused_predictions"]], "user_instances (sleap.instance.labeledframe property)": [[10, "sleap.instance.LabeledFrame.user_instances"]], "video (sleap.instance.instance property)": [[10, "sleap.instance.Instance.video"]], "main() (in module sleap.io.convert)": [[11, "sleap.io.convert.main"]], "sleap.io.convert": [[11, "module-sleap.io.convert"]], "labels (class in sleap.io.dataset)": [[12, "sleap.io.dataset.Labels"]], "labelsdatacache (class in sleap.io.dataset)": [[12, "sleap.io.dataset.LabelsDataCache"]], "add_instance() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.add_instance"]], "add_instance() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.add_instance"]], "add_suggestion() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.add_suggestion"]], "add_track() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.add_track"]], "add_track() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.add_track"]], "add_video() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.add_video"]], "all_instances (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.all_instances"]], "append() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.append"]], "append_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.append_suggestions"]], "clear_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.clear_suggestions"]], "complex_merge_between() (sleap.io.dataset.labels class method)": [[12, "sleap.io.dataset.Labels.complex_merge_between"]], "copy() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.copy"]], "delete_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.delete_suggestions"]], "describe() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.describe"]], "export() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.export"]], "export_csv() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.export_csv"]], "export_nwb() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.export_nwb"]], "extend_from() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.extend_from"]], "extract() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.extract"]], "find() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.find"]], "find_fancy_frame_idxs() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.find_fancy_frame_idxs"]], "find_first() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.find_first"]], "find_frames() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.find_frames"]], "find_last() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.find_last"]], "find_path_using_paths() (in module sleap.io.dataset)": [[12, "sleap.io.dataset.find_path_using_paths"]], "find_suggestion() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.find_suggestion"]], "find_track_occupancy() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.find_track_occupancy"]], "finish_complex_merge() (sleap.io.dataset.labels static method)": [[12, "sleap.io.dataset.Labels.finish_complex_merge"]], "frames() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.frames"]], "get() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get"]], "get_filtered_frame_idxs() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.get_filtered_frame_idxs"]], "get_frame_count() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.get_frame_count"]], "get_next_suggestion() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_next_suggestion"]], "get_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_suggestions"]], "get_track_count() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_track_count"]], "get_track_occupancy() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_track_occupancy"]], "get_track_occupancy() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.get_track_occupancy"]], "get_unlabeled_suggestion_inds() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_unlabeled_suggestion_inds"]], "get_video_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.get_video_suggestions"]], "get_video_track_occupancy() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.get_video_track_occupancy"]], "has_frame() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.has_frame"]], "has_missing_videos (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.has_missing_videos"]], "index() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.index"]], "insert() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.insert"]], "instance_count() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.instance_count"]], "instances() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.instances"]], "is_multi_instance (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.is_multi_instance"]], "labeled_frames (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.labeled_frames"]], "labels (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.labels"]], "load_file() (in module sleap.io.dataset)": [[12, "sleap.io.dataset.load_file"]], "load_file() (sleap.io.dataset.labels class method)": [[12, "sleap.io.dataset.Labels.load_file"]], "make_video_callback() (sleap.io.dataset.labels class method)": [[12, "sleap.io.dataset.Labels.make_video_callback"]], "merge_container_dicts() (sleap.io.dataset.labels static method)": [[12, "sleap.io.dataset.Labels.merge_container_dicts"]], "merge_matching_frames() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.merge_matching_frames"]], "merge_nodes() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.merge_nodes"]], "negative_anchors (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.negative_anchors"]], "numpy() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.numpy"]], "predicted_instances (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.predicted_instances"]], "provenance (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.provenance"]], "remove() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove"]], "remove_all_tracks() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_all_tracks"]], "remove_empty_frames() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_empty_frames"]], "remove_empty_instances() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_empty_instances"]], "remove_frame() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_frame"]], "remove_frame() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.remove_frame"]], "remove_frames() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_frames"]], "remove_instance() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_instance"]], "remove_instance() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.remove_instance"]], "remove_predictions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_predictions"]], "remove_suggestion() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_suggestion"]], "remove_track() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_track"]], "remove_untracked_instances() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_untracked_instances"]], "remove_unused_tracks() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_unused_tracks"]], "remove_user_instances() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_user_instances"]], "remove_video() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.remove_video"]], "remove_video() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.remove_video"]], "save() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.save"]], "save_file() (sleap.io.dataset.labels class method)": [[12, "sleap.io.dataset.Labels.save_file"]], "save_frame_data_hdf5() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.save_frame_data_hdf5"]], "save_frame_data_imgstore() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.save_frame_data_imgstore"]], "set_suggestions() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.set_suggestions"]], "skeleton (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.skeleton"]], "skeletons (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.skeletons"]], "sleap.io.dataset": [[12, "module-sleap.io.dataset"]], "split() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.split"]], "suggestions (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.suggestions"]], "to_dict() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.to_dict"]], "to_json() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.to_json"]], "to_pipeline() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.to_pipeline"]], "track_set_instance() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.track_set_instance"]], "track_swap() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.track_swap"]], "track_swap() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.track_swap"]], "tracks (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.tracks"]], "unlabeled_suggestions (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.unlabeled_suggestions"]], "update() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.update"]], "update_counts_for_frame() (sleap.io.dataset.labelsdatacache method)": [[12, "sleap.io.dataset.LabelsDataCache.update_counts_for_frame"]], "user_instances (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.user_instances"]], "user_labeled_frame_inds (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.user_labeled_frame_inds"]], "user_labeled_frames (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.user_labeled_frames"]], "video (sleap.io.dataset.labels property)": [[12, "sleap.io.dataset.Labels.video"]], "videos (sleap.io.dataset.labels attribute)": [[12, "sleap.io.dataset.Labels.videos"]], "with_user_labels_only() (sleap.io.dataset.labels method)": [[12, "sleap.io.dataset.Labels.with_user_labels_only"]], "adaptor (class in sleap.io.format.adaptor)": [[13, "sleap.io.format.adaptor.Adaptor"]], "sleapobjecttype (class in sleap.io.format.adaptor)": [[13, "sleap.io.format.adaptor.SleapObjectType"]], "all_exts (sleap.io.format.adaptor.adaptor property)": [[13, "sleap.io.format.adaptor.Adaptor.all_exts"]], "can_read_file() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.can_read_file"]], "can_write_filename() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.can_write_filename"]], "default_ext (sleap.io.format.adaptor.adaptor property)": [[13, "sleap.io.format.adaptor.Adaptor.default_ext"]], "does_match_ext() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.does_match_ext"]], "does_read() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.does_read"]], "does_write() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.does_write"]], "formatted_ext_options (sleap.io.format.adaptor.adaptor property)": [[13, "sleap.io.format.adaptor.Adaptor.formatted_ext_options"]], "handles (sleap.io.format.adaptor.adaptor property)": [[13, "sleap.io.format.adaptor.Adaptor.handles"]], "name (sleap.io.format.adaptor.adaptor property)": [[13, "sleap.io.format.adaptor.Adaptor.name"]], "read() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.read"]], "sleap.io.format.adaptor": [[13, "module-sleap.io.format.adaptor"]], "write() (sleap.io.format.adaptor.adaptor method)": [[13, "sleap.io.format.adaptor.Adaptor.write"]], "alphatrackeradaptor (class in sleap.io.format.alphatracker)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor"]], "all_exts (sleap.io.format.alphatracker.alphatrackeradaptor property)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.all_exts"]], "can_read_file() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.can_write_filename"]], "default_ext (sleap.io.format.alphatracker.alphatrackeradaptor property)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.default_ext"]], "does_match_ext() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_match_ext"]], "does_read() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_read"]], "does_write() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.does_write"]], "formatted_ext_options (sleap.io.format.alphatracker.alphatrackeradaptor property)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.formatted_ext_options"]], "get_alpha_tracker_frame_dict() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_frame_dict"]], "get_alpha_tracker_instance_dict() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_instance_dict"]], "get_alpha_tracker_point_dict() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.get_alpha_tracker_point_dict"]], "handles (sleap.io.format.alphatracker.alphatrackeradaptor property)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.handles"]], "make_video_for_image_list() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.make_video_for_image_list"]], "name (sleap.io.format.alphatracker.alphatrackeradaptor property)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.name"]], "read() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.read"]], "sleap.io.format.alphatracker": [[14, "module-sleap.io.format.alphatracker"]], "write() (sleap.io.format.alphatracker.alphatrackeradaptor method)": [[14, "sleap.io.format.alphatracker.AlphaTrackerAdaptor.write"]], "labelscocoadaptor (class in sleap.io.format.coco)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor"]], "all_exts (sleap.io.format.coco.labelscocoadaptor property)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.all_exts"]], "can_read_file() (sleap.io.format.coco.labelscocoadaptor method)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.coco.labelscocoadaptor method)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.can_write_filename"]], "default_ext (sleap.io.format.coco.labelscocoadaptor property)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.default_ext"]], "does_read() (sleap.io.format.coco.labelscocoadaptor method)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.does_read"]], "does_write() (sleap.io.format.coco.labelscocoadaptor method)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.does_write"]], "handles (sleap.io.format.coco.labelscocoadaptor property)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.handles"]], "name (sleap.io.format.coco.labelscocoadaptor property)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.name"]], "read() (sleap.io.format.coco.labelscocoadaptor class method)": [[15, "sleap.io.format.coco.LabelsCocoAdaptor.read"]], "sleap.io.format.coco": [[15, "module-sleap.io.format.coco"]], "csvadaptor (class in sleap.io.format.csv)": [[16, "sleap.io.format.csv.CSVAdaptor"]], "all_exts (sleap.io.format.csv.csvadaptor property)": [[16, "sleap.io.format.csv.CSVAdaptor.all_exts"]], "can_read_file() (sleap.io.format.csv.csvadaptor method)": [[16, "sleap.io.format.csv.CSVAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.csv.csvadaptor method)": [[16, "sleap.io.format.csv.CSVAdaptor.can_write_filename"]], "default_ext (sleap.io.format.csv.csvadaptor property)": [[16, "sleap.io.format.csv.CSVAdaptor.default_ext"]], "does_read() (sleap.io.format.csv.csvadaptor method)": [[16, "sleap.io.format.csv.CSVAdaptor.does_read"]], "does_write() (sleap.io.format.csv.csvadaptor method)": [[16, "sleap.io.format.csv.CSVAdaptor.does_write"]], "handles (sleap.io.format.csv.csvadaptor property)": [[16, "sleap.io.format.csv.CSVAdaptor.handles"]], "name (sleap.io.format.csv.csvadaptor property)": [[16, "sleap.io.format.csv.CSVAdaptor.name"]], "sleap.io.format.csv": [[16, "module-sleap.io.format.csv"]], "write() (sleap.io.format.csv.csvadaptor class method)": [[16, "sleap.io.format.csv.CSVAdaptor.write"]], "labelsdeeplabcutcsvadaptor (class in sleap.io.format.deeplabcut)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor"]], "labelsdeeplabcutyamladaptor (class in sleap.io.format.deeplabcut)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor"]], "all_exts (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.all_exts"]], "all_exts (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.all_exts"]], "can_read_file() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_read_file"]], "can_read_file() (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.can_write_filename"]], "can_write_filename() (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.can_write_filename"]], "default_ext (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.default_ext"]], "default_ext (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.default_ext"]], "does_read() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_read"]], "does_read() (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_read"]], "does_write() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.does_write"]], "does_write() (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.does_write"]], "handles (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.handles"]], "handles (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.handles"]], "make_video_for_image_list() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor class method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.make_video_for_image_list"]], "name (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.name"]], "name (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor property)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.name"]], "read() (sleap.io.format.deeplabcut.labelsdeeplabcutcsvadaptor class method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutCsvAdaptor.read"]], "read() (sleap.io.format.deeplabcut.labelsdeeplabcutyamladaptor class method)": [[17, "sleap.io.format.deeplabcut.LabelsDeepLabCutYamlAdaptor.read"]], "sleap.io.format.deeplabcut": [[17, "module-sleap.io.format.deeplabcut"]], "labelsdeepposekitadaptor (class in sleap.io.format.deepposekit)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor"]], "all_exts (sleap.io.format.deepposekit.labelsdeepposekitadaptor property)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.all_exts"]], "can_read_file() (sleap.io.format.deepposekit.labelsdeepposekitadaptor method)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.deepposekit.labelsdeepposekitadaptor method)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.can_write_filename"]], "default_ext (sleap.io.format.deepposekit.labelsdeepposekitadaptor property)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.default_ext"]], "does_read() (sleap.io.format.deepposekit.labelsdeepposekitadaptor method)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_read"]], "does_write() (sleap.io.format.deepposekit.labelsdeepposekitadaptor method)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.does_write"]], "handles (sleap.io.format.deepposekit.labelsdeepposekitadaptor property)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.handles"]], "name (sleap.io.format.deepposekit.labelsdeepposekitadaptor property)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.name"]], "read() (sleap.io.format.deepposekit.labelsdeepposekitadaptor class method)": [[18, "sleap.io.format.deepposekit.LabelsDeepPoseKitAdaptor.read"]], "sleap.io.format.deepposekit": [[18, "module-sleap.io.format.deepposekit"]], "sleap.io.format.dispatch": [[19, "module-sleap.io.format.dispatch"]], "filehandle (class in sleap.io.format.filehandle)": [[20, "sleap.io.format.filehandle.FileHandle"]], "close() (sleap.io.format.filehandle.filehandle method)": [[20, "sleap.io.format.filehandle.FileHandle.close"]], "file (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.file"]], "format_id (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.format_id"]], "is_hdf5 (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.is_hdf5"]], "is_json (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.is_json"]], "json (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.json"]], "open() (sleap.io.format.filehandle.filehandle method)": [[20, "sleap.io.format.filehandle.FileHandle.open"]], "sleap.io.format.filehandle": [[20, "module-sleap.io.format.filehandle"]], "text (sleap.io.format.filehandle.filehandle property)": [[20, "sleap.io.format.filehandle.FileHandle.text"]], "genericjsonadaptor (class in sleap.io.format.genericjson)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor"]], "all_exts (sleap.io.format.genericjson.genericjsonadaptor property)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.all_exts"]], "can_read_file() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.can_write_filename"]], "default_ext (sleap.io.format.genericjson.genericjsonadaptor property)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.default_ext"]], "does_read() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.does_read"]], "does_write() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.does_write"]], "handles (sleap.io.format.genericjson.genericjsonadaptor property)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.handles"]], "name (sleap.io.format.genericjson.genericjsonadaptor property)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.name"]], "read() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.read"]], "sleap.io.format.genericjson": [[21, "module-sleap.io.format.genericjson"]], "write() (sleap.io.format.genericjson.genericjsonadaptor method)": [[21, "sleap.io.format.genericjson.GenericJsonAdaptor.write"]], "labelsv1adaptor (class in sleap.io.format.hdf5)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor"]], "all_exts (sleap.io.format.hdf5.labelsv1adaptor property)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.all_exts"]], "can_read_file() (sleap.io.format.hdf5.labelsv1adaptor method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.can_read_file"]], "can_write_filename() (sleap.io.format.hdf5.labelsv1adaptor method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.can_write_filename"]], "default_ext (sleap.io.format.hdf5.labelsv1adaptor property)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.default_ext"]], "does_read() (sleap.io.format.hdf5.labelsv1adaptor method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.does_read"]], "does_write() (sleap.io.format.hdf5.labelsv1adaptor method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.does_write"]], "handles (sleap.io.format.hdf5.labelsv1adaptor property)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.handles"]], "name (sleap.io.format.hdf5.labelsv1adaptor property)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.name"]], "read() (sleap.io.format.hdf5.labelsv1adaptor class method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.read"]], "sleap.io.format.hdf5": [[22, "module-sleap.io.format.hdf5"]], "write() (sleap.io.format.hdf5.labelsv1adaptor class method)": [[22, "sleap.io.format.hdf5.LabelsV1Adaptor.write"]], "labelsjsonadaptor (class in sleap.io.format.labels_json)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor"]], "all_exts (sleap.io.format.labels_json.labelsjsonadaptor property)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.all_exts"]], "can_read_file() (sleap.io.format.labels_json.labelsjsonadaptor method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.labels_json.labelsjsonadaptor method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.can_write_filename"]], "default_ext (sleap.io.format.labels_json.labelsjsonadaptor property)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.default_ext"]], "does_read() (sleap.io.format.labels_json.labelsjsonadaptor method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.does_read"]], "does_write() (sleap.io.format.labels_json.labelsjsonadaptor method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.does_write"]], "from_json_data() (sleap.io.format.labels_json.labelsjsonadaptor class method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.from_json_data"]], "handles (sleap.io.format.labels_json.labelsjsonadaptor property)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.handles"]], "name (sleap.io.format.labels_json.labelsjsonadaptor property)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.name"]], "read() (sleap.io.format.labels_json.labelsjsonadaptor class method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.read"]], "sleap.io.format.labels_json": [[23, "module-sleap.io.format.labels_json"]], "write() (sleap.io.format.labels_json.labelsjsonadaptor class method)": [[23, "sleap.io.format.labels_json.LabelsJsonAdaptor.write"]], "labelsleapmatlabadaptor (class in sleap.io.format.leap_matlab)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor"]], "all_exts (sleap.io.format.leap_matlab.labelsleapmatlabadaptor property)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.all_exts"]], "can_read_file() (sleap.io.format.leap_matlab.labelsleapmatlabadaptor method)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.leap_matlab.labelsleapmatlabadaptor method)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.can_write_filename"]], "default_ext (sleap.io.format.leap_matlab.labelsleapmatlabadaptor property)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.default_ext"]], "does_read() (sleap.io.format.leap_matlab.labelsleapmatlabadaptor method)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_read"]], "does_write() (sleap.io.format.leap_matlab.labelsleapmatlabadaptor method)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.does_write"]], "handles (sleap.io.format.leap_matlab.labelsleapmatlabadaptor property)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.handles"]], "name (sleap.io.format.leap_matlab.labelsleapmatlabadaptor property)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.name"]], "read() (sleap.io.format.leap_matlab.labelsleapmatlabadaptor class method)": [[24, "sleap.io.format.leap_matlab.LabelsLeapMatlabAdaptor.read"]], "sleap.io.format.leap_matlab": [[24, "module-sleap.io.format.leap_matlab"]], "read() (in module sleap.io.format.main)": [[25, "sleap.io.format.main.read"]], "sleap.io.format.main": [[25, "module-sleap.io.format.main"]], "write() (in module sleap.io.format.main)": [[25, "sleap.io.format.main.write"]], "ndxposeadaptor (class in sleap.io.format.ndx_pose)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor"]], "all_exts (sleap.io.format.ndx_pose.ndxposeadaptor property)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.all_exts"]], "can_read_file() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.can_write_filename"]], "default_ext (sleap.io.format.ndx_pose.ndxposeadaptor property)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.default_ext"]], "does_read() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.does_read"]], "does_write() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.does_write"]], "handles (sleap.io.format.ndx_pose.ndxposeadaptor property)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.handles"]], "name (sleap.io.format.ndx_pose.ndxposeadaptor property)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.name"]], "read() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.read"]], "sleap.io.format.ndx_pose": [[26, "module-sleap.io.format.ndx_pose"]], "write() (sleap.io.format.ndx_pose.ndxposeadaptor method)": [[26, "sleap.io.format.ndx_pose.NDXPoseAdaptor.write"]], "nixadaptor (class in sleap.io.format.nix)": [[27, "sleap.io.format.nix.NixAdaptor"]], "all_exts (sleap.io.format.nix.nixadaptor property)": [[27, "sleap.io.format.nix.NixAdaptor.all_exts"]], "can_read_file() (sleap.io.format.nix.nixadaptor class method)": [[27, "sleap.io.format.nix.NixAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.nix.nixadaptor method)": [[27, "sleap.io.format.nix.NixAdaptor.can_write_filename"]], "default_ext (sleap.io.format.nix.nixadaptor property)": [[27, "sleap.io.format.nix.NixAdaptor.default_ext"]], "does_read() (sleap.io.format.nix.nixadaptor class method)": [[27, "sleap.io.format.nix.NixAdaptor.does_read"]], "does_write() (sleap.io.format.nix.nixadaptor class method)": [[27, "sleap.io.format.nix.NixAdaptor.does_write"]], "handles (sleap.io.format.nix.nixadaptor property)": [[27, "sleap.io.format.nix.NixAdaptor.handles"]], "name (sleap.io.format.nix.nixadaptor property)": [[27, "sleap.io.format.nix.NixAdaptor.name"]], "read() (sleap.io.format.nix.nixadaptor class method)": [[27, "sleap.io.format.nix.NixAdaptor.read"]], "sleap.io.format.nix": [[27, "module-sleap.io.format.nix"]], "write() (sleap.io.format.nix.nixadaptor class method)": [[27, "sleap.io.format.nix.NixAdaptor.write"]], "sleapanalysisadaptor (class in sleap.io.format.sleap_analysis)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor"]], "all_exts (sleap.io.format.sleap_analysis.sleapanalysisadaptor property)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.all_exts"]], "can_read_file() (sleap.io.format.sleap_analysis.sleapanalysisadaptor method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.sleap_analysis.sleapanalysisadaptor method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.can_write_filename"]], "default_ext (sleap.io.format.sleap_analysis.sleapanalysisadaptor property)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.default_ext"]], "does_read() (sleap.io.format.sleap_analysis.sleapanalysisadaptor method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_read"]], "does_write() (sleap.io.format.sleap_analysis.sleapanalysisadaptor method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.does_write"]], "handles (sleap.io.format.sleap_analysis.sleapanalysisadaptor property)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.handles"]], "name (sleap.io.format.sleap_analysis.sleapanalysisadaptor property)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.name"]], "read() (sleap.io.format.sleap_analysis.sleapanalysisadaptor class method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.read"]], "sleap.io.format.sleap_analysis": [[28, "module-sleap.io.format.sleap_analysis"]], "write() (sleap.io.format.sleap_analysis.sleapanalysisadaptor class method)": [[28, "sleap.io.format.sleap_analysis.SleapAnalysisAdaptor.write"]], "textadaptor (class in sleap.io.format.text)": [[29, "sleap.io.format.text.TextAdaptor"]], "all_exts (sleap.io.format.text.textadaptor property)": [[29, "sleap.io.format.text.TextAdaptor.all_exts"]], "can_read_file() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.can_read_file"]], "can_write_filename() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.can_write_filename"]], "default_ext (sleap.io.format.text.textadaptor property)": [[29, "sleap.io.format.text.TextAdaptor.default_ext"]], "does_read() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.does_read"]], "does_write() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.does_write"]], "handles (sleap.io.format.text.textadaptor property)": [[29, "sleap.io.format.text.TextAdaptor.handles"]], "name (sleap.io.format.text.textadaptor property)": [[29, "sleap.io.format.text.TextAdaptor.name"]], "read() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.read"]], "sleap.io.format.text": [[29, "module-sleap.io.format.text"]], "write() (sleap.io.format.text.textadaptor method)": [[29, "sleap.io.format.text.TextAdaptor.write"]], "load_labels_json_old() (in module sleap.io.legacy)": [[30, "sleap.io.legacy.load_labels_json_old"]], "load_predicted_labels_json_old() (in module sleap.io.legacy)": [[30, "sleap.io.legacy.load_predicted_labels_json_old"]], "sleap.io.legacy": [[30, "module-sleap.io.legacy"]], "filenames_prefix_change() (in module sleap.io.pathutils)": [[31, "sleap.io.pathutils.filenames_prefix_change"]], "find_changed_subpath() (in module sleap.io.pathutils)": [[31, "sleap.io.pathutils.find_changed_subpath"]], "list_file_missing() (in module sleap.io.pathutils)": [[31, "sleap.io.pathutils.list_file_missing"]], "sleap.io.pathutils": [[31, "module-sleap.io.pathutils"]], "dummyvideo (class in sleap.io.video)": [[32, "sleap.io.video.DummyVideo"]], "hdf5video (class in sleap.io.video)": [[32, "sleap.io.video.HDF5Video"]], "imgstorevideo (class in sleap.io.video)": [[32, "sleap.io.video.ImgStoreVideo"]], "mediavideo (class in sleap.io.video)": [[32, "sleap.io.video.MediaVideo"]], "numpyvideo (class in sleap.io.video)": [[32, "sleap.io.video.NumpyVideo"]], "singleimagevideo (class in sleap.io.video)": [[32, "sleap.io.video.SingleImageVideo"]], "video (class in sleap.io.video)": [[32, "sleap.io.video.Video"]], "available_video_exts() (in module sleap.io.video)": [[32, "sleap.io.video.available_video_exts"]], "cattr() (sleap.io.video.video static method)": [[32, "sleap.io.video.Video.cattr"]], "channels (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.channels"]], "channels (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.channels"]], "channels (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.channels"]], "channels (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.channels"]], "channels (sleap.io.video.singleimagevideo property)": [[32, "sleap.io.video.SingleImageVideo.channels"]], "check() (sleap.io.video.hdf5video method)": [[32, "sleap.io.video.HDF5Video.check"]], "close() (sleap.io.video.hdf5video method)": [[32, "sleap.io.video.HDF5Video.close"]], "close() (sleap.io.video.imgstorevideo method)": [[32, "sleap.io.video.ImgStoreVideo.close"]], "dtype (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.dtype"]], "dtype (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.dtype"]], "dtype (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.dtype"]], "dtype (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.dtype"]], "dtype (sleap.io.video.singleimagevideo property)": [[32, "sleap.io.video.SingleImageVideo.dtype"]], "embedded_frame_inds (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.embedded_frame_inds"]], "enable_source_video (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.enable_source_video"]], "fixup_path() (sleap.io.video.video static method)": [[32, "sleap.io.video.Video.fixup_path"]], "fps (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.fps"]], "frames (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.frames"]], "frames (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.frames"]], "frames (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.frames"]], "frames (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.frames"]], "frames (sleap.io.video.singleimagevideo property)": [[32, "sleap.io.video.SingleImageVideo.frames"]], "from_filename() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.from_filename"]], "from_hdf5() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.from_hdf5"]], "from_image_filenames() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.from_image_filenames"]], "from_media() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.from_media"]], "from_numpy() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.from_numpy"]], "get_frame() (sleap.io.video.hdf5video method)": [[32, "sleap.io.video.HDF5Video.get_frame"]], "get_frame() (sleap.io.video.imgstorevideo method)": [[32, "sleap.io.video.ImgStoreVideo.get_frame"]], "get_frame() (sleap.io.video.mediavideo method)": [[32, "sleap.io.video.MediaVideo.get_frame"]], "get_frame() (sleap.io.video.numpyvideo method)": [[32, "sleap.io.video.NumpyVideo.get_frame"]], "get_frame() (sleap.io.video.singleimagevideo method)": [[32, "sleap.io.video.SingleImageVideo.get_frame"]], "get_frame() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.get_frame"]], "get_frames() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.get_frames"]], "get_frames_safely() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.get_frames_safely"]], "has_embedded_images (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.has_embedded_images"]], "height (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.height"]], "height (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.height"]], "height (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.height"]], "height (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.height"]], "height (sleap.io.video.singleimagevideo property)": [[32, "sleap.io.video.SingleImageVideo.height"]], "imgstore (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.imgstore"]], "imgstore_from_filenames() (sleap.io.video.video class method)": [[32, "sleap.io.video.Video.imgstore_from_filenames"]], "is_missing (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.is_missing"]], "is_missing (sleap.io.video.video property)": [[32, "sleap.io.video.Video.is_missing"]], "last_frame_idx (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.last_frame_idx"]], "last_frame_idx (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.last_frame_idx"]], "last_frame_idx (sleap.io.video.video property)": [[32, "sleap.io.video.Video.last_frame_idx"]], "load_video() (in module sleap.io.video)": [[32, "sleap.io.video.load_video"]], "matches() (sleap.io.video.hdf5video method)": [[32, "sleap.io.video.HDF5Video.matches"]], "matches() (sleap.io.video.imgstorevideo method)": [[32, "sleap.io.video.ImgStoreVideo.matches"]], "matches() (sleap.io.video.mediavideo method)": [[32, "sleap.io.video.MediaVideo.matches"]], "matches() (sleap.io.video.numpyvideo method)": [[32, "sleap.io.video.NumpyVideo.matches"]], "matches() (sleap.io.video.singleimagevideo method)": [[32, "sleap.io.video.SingleImageVideo.matches"]], "num_frames (sleap.io.video.video property)": [[32, "sleap.io.video.Video.num_frames"]], "open() (sleap.io.video.imgstorevideo method)": [[32, "sleap.io.video.ImgStoreVideo.open"]], "reset() (sleap.io.video.hdf5video method)": [[32, "sleap.io.video.HDF5Video.reset"]], "reset() (sleap.io.video.imgstorevideo method)": [[32, "sleap.io.video.ImgStoreVideo.reset"]], "reset() (sleap.io.video.mediavideo method)": [[32, "sleap.io.video.MediaVideo.reset"]], "reset() (sleap.io.video.numpyvideo method)": [[32, "sleap.io.video.NumpyVideo.reset"]], "reset() (sleap.io.video.singleimagevideo method)": [[32, "sleap.io.video.SingleImageVideo.reset"]], "shape (sleap.io.video.video property)": [[32, "sleap.io.video.Video.shape"]], "sleap.io.video": [[32, "module-sleap.io.video"]], "source_video (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.source_video"]], "source_video_available (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.source_video_available"]], "to_hdf5() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.to_hdf5"]], "to_imgstore() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.to_imgstore"]], "to_pipeline() (sleap.io.video.video method)": [[32, "sleap.io.video.Video.to_pipeline"]], "width (sleap.io.video.hdf5video property)": [[32, "sleap.io.video.HDF5Video.width"]], "width (sleap.io.video.imgstorevideo property)": [[32, "sleap.io.video.ImgStoreVideo.width"]], "width (sleap.io.video.mediavideo property)": [[32, "sleap.io.video.MediaVideo.width"]], "width (sleap.io.video.numpyvideo property)": [[32, "sleap.io.video.NumpyVideo.width"]], "width (sleap.io.video.singleimagevideo property)": [[32, "sleap.io.video.SingleImageVideo.width"]], "videowriter (class in sleap.io.videowriter)": [[33, "sleap.io.videowriter.VideoWriter"]], "videowriterimageio (class in sleap.io.videowriter)": [[33, "sleap.io.videowriter.VideoWriterImageio"]], "videowriteropencv (class in sleap.io.videowriter)": [[33, "sleap.io.videowriter.VideoWriterOpenCV"]], "can_use_ffmpeg() (sleap.io.videowriter.videowriter static method)": [[33, "sleap.io.videowriter.VideoWriter.can_use_ffmpeg"]], "crf (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.crf"]], "filename (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.filename"]], "fps (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.fps"]], "height (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.height"]], "preset (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.preset"]], "safe_builder() (sleap.io.videowriter.videowriter static method)": [[33, "sleap.io.videowriter.VideoWriter.safe_builder"]], "sleap.io.videowriter": [[33, "module-sleap.io.videowriter"]], "width (sleap.io.videowriter.videowriterimageio attribute)": [[33, "sleap.io.videowriter.VideoWriterImageio.width"]], "videomarkerthread (class in sleap.io.visuals)": [[34, "sleap.io.visuals.VideoMarkerThread"]], "img_to_cv() (in module sleap.io.visuals)": [[34, "sleap.io.visuals.img_to_cv"]], "reader() (in module sleap.io.visuals)": [[34, "sleap.io.visuals.reader"]], "resize_image() (in module sleap.io.visuals)": [[34, "sleap.io.visuals.resize_image"]], "run() (sleap.io.visuals.videomarkerthread method)": [[34, "sleap.io.visuals.VideoMarkerThread.run"]], "save_labeled_video() (in module sleap.io.visuals)": [[34, "sleap.io.visuals.save_labeled_video"]], "sleap.io.visuals": [[34, "module-sleap.io.visuals"]], "writer() (in module sleap.io.visuals)": [[34, "sleap.io.visuals.writer"]], "basemessageparticipant (class in sleap.message)": [[35, "sleap.message.BaseMessageParticipant"]], "receiver (class in sleap.message)": [[35, "sleap.message.Receiver"]], "sender (class in sleap.message)": [[35, "sleap.message.Sender"]], "check_message() (sleap.message.receiver method)": [[35, "sleap.message.Receiver.check_message"]], "check_messages() (sleap.message.receiver method)": [[35, "sleap.message.Receiver.check_messages"]], "push_back_message() (sleap.message.receiver method)": [[35, "sleap.message.Receiver.push_back_message"]], "send_array() (sleap.message.sender method)": [[35, "sleap.message.Sender.send_array"]], "send_dict() (sleap.message.sender method)": [[35, "sleap.message.Sender.send_dict"]], "sleap.message": [[35, "module-sleap.message"]], "intermediatefeature (class in sleap.nn.architectures.common)": [[36, "sleap.nn.architectures.common.IntermediateFeature"]], "scale (sleap.nn.architectures.common.intermediatefeature property)": [[36, "sleap.nn.architectures.common.IntermediateFeature.scale"]], "sleap.nn.architectures.common": [[36, "module-sleap.nn.architectures.common"]], "stride (sleap.nn.architectures.common.intermediatefeature attribute)": [[36, "sleap.nn.architectures.common.IntermediateFeature.stride"]], "tensor (sleap.nn.architectures.common.intermediatefeature attribute)": [[36, "sleap.nn.architectures.common.IntermediateFeature.tensor"]], "decoderblock (class in sleap.nn.architectures.encoder_decoder)": [[37, "sleap.nn.architectures.encoder_decoder.DecoderBlock"]], "encoderblock (class in sleap.nn.architectures.encoder_decoder)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderBlock"]], "encoderdecoder (class in sleap.nn.architectures.encoder_decoder)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder"]], "simpleconvblock (class in sleap.nn.architectures.encoder_decoder)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock"]], "simpleupsamplingblock (class in sleap.nn.architectures.encoder_decoder)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock"]], "activation (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.activation"]], "batch_norm (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm"]], "batch_norm_before_activation (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.batch_norm_before_activation"]], "block_prefix (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.block_prefix"]], "decoder_features_stride (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_features_stride"]], "decoder_stack (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.decoder_stack"]], "encoder_features_stride (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_features_stride"]], "encoder_stack (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.encoder_stack"]], "filters (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.filters"]], "interp_method (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.interp_method"]], "kernel_size (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.kernel_size"]], "make_backbone() (sleap.nn.architectures.encoder_decoder.encoderdecoder method)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_backbone"]], "make_block() (sleap.nn.architectures.encoder_decoder.decoderblock method)": [[37, "sleap.nn.architectures.encoder_decoder.DecoderBlock.make_block"]], "make_block() (sleap.nn.architectures.encoder_decoder.encoderblock method)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderBlock.make_block"]], "make_block() (sleap.nn.architectures.encoder_decoder.simpleconvblock method)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.make_block"]], "make_block() (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock method)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.make_block"]], "make_decoder() (sleap.nn.architectures.encoder_decoder.encoderdecoder method)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_decoder"]], "make_encoder() (sleap.nn.architectures.encoder_decoder.encoderdecoder method)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_encoder"]], "make_stem() (sleap.nn.architectures.encoder_decoder.encoderdecoder method)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.make_stem"]], "maximum_stride (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.maximum_stride"]], "num_convs (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.num_convs"]], "output_stride (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.output_stride"]], "pool (sleap.nn.architectures.encoder_decoder.encoderblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderBlock.pool"]], "pool (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool"]], "pool_before_convs (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pool_before_convs"]], "pooling_stride (sleap.nn.architectures.encoder_decoder.encoderblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderBlock.pooling_stride"]], "pooling_stride (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.pooling_stride"]], "refine_convs (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs"]], "refine_convs_activation (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_activation"]], "refine_convs_batch_norm (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm"]], "refine_convs_batch_norm_before_activation (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_batch_norm_before_activation"]], "refine_convs_filters (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_filters"]], "refine_convs_first_filters (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_first_filters"]], "refine_convs_kernel_size (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_kernel_size"]], "refine_convs_use_bias (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.refine_convs_use_bias"]], "skip_add (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_add"]], "skip_connection (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.skip_connection"]], "sleap.nn.architectures.encoder_decoder": [[37, "module-sleap.nn.architectures.encoder_decoder"]], "stacks (sleap.nn.architectures.encoder_decoder.encoderdecoder attribute)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.stacks"]], "stem_features_stride (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_features_stride"]], "stem_stack (sleap.nn.architectures.encoder_decoder.encoderdecoder property)": [[37, "sleap.nn.architectures.encoder_decoder.EncoderDecoder.stem_stack"]], "transposed_conv (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv"]], "transposed_conv_activation (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_activation"]], "transposed_conv_batch_norm (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm"]], "transposed_conv_batch_norm_before_activation (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_batch_norm_before_activation"]], "transposed_conv_filters (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_filters"]], "transposed_conv_kernel_size (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_kernel_size"]], "transposed_conv_use_bias (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.transposed_conv_use_bias"]], "upsampling_stride (sleap.nn.architectures.encoder_decoder.decoderblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.DecoderBlock.upsampling_stride"]], "upsampling_stride (sleap.nn.architectures.encoder_decoder.simpleupsamplingblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleUpsamplingBlock.upsampling_stride"]], "use_bias (sleap.nn.architectures.encoder_decoder.simpleconvblock attribute)": [[37, "sleap.nn.architectures.encoder_decoder.SimpleConvBlock.use_bias"]], "downsamplingblock (class in sleap.nn.architectures.hourglass)": [[38, "sleap.nn.architectures.hourglass.DownsamplingBlock"]], "hourglass (class in sleap.nn.architectures.hourglass)": [[38, "sleap.nn.architectures.hourglass.Hourglass"]], "stemblock (class in sleap.nn.architectures.hourglass)": [[38, "sleap.nn.architectures.hourglass.StemBlock"]], "upsamplingblock (class in sleap.nn.architectures.hourglass)": [[38, "sleap.nn.architectures.hourglass.UpsamplingBlock"]], "conv() (in module sleap.nn.architectures.hourglass)": [[38, "sleap.nn.architectures.hourglass.conv"]], "decoder_stack (sleap.nn.architectures.hourglass.hourglass property)": [[38, "sleap.nn.architectures.hourglass.Hourglass.decoder_stack"]], "down_blocks (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.down_blocks"]], "encoder_stack (sleap.nn.architectures.hourglass.hourglass property)": [[38, "sleap.nn.architectures.hourglass.Hourglass.encoder_stack"]], "filter_increase (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.filter_increase"]], "filters (sleap.nn.architectures.hourglass.downsamplingblock attribute)": [[38, "sleap.nn.architectures.hourglass.DownsamplingBlock.filters"]], "filters (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.filters"]], "filters (sleap.nn.architectures.hourglass.stemblock attribute)": [[38, "sleap.nn.architectures.hourglass.StemBlock.filters"]], "filters (sleap.nn.architectures.hourglass.upsamplingblock attribute)": [[38, "sleap.nn.architectures.hourglass.UpsamplingBlock.filters"]], "first_conv_stride (sleap.nn.architectures.hourglass.stemblock attribute)": [[38, "sleap.nn.architectures.hourglass.StemBlock.first_conv_stride"]], "from_config() (sleap.nn.architectures.hourglass.hourglass class method)": [[38, "sleap.nn.architectures.hourglass.Hourglass.from_config"]], "interp_method (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.interp_method"]], "interp_method (sleap.nn.architectures.hourglass.upsamplingblock attribute)": [[38, "sleap.nn.architectures.hourglass.UpsamplingBlock.interp_method"]], "make_block() (sleap.nn.architectures.hourglass.downsamplingblock method)": [[38, "sleap.nn.architectures.hourglass.DownsamplingBlock.make_block"]], "make_block() (sleap.nn.architectures.hourglass.stemblock method)": [[38, "sleap.nn.architectures.hourglass.StemBlock.make_block"]], "make_block() (sleap.nn.architectures.hourglass.upsamplingblock method)": [[38, "sleap.nn.architectures.hourglass.UpsamplingBlock.make_block"]], "output_filters (sleap.nn.architectures.hourglass.stemblock attribute)": [[38, "sleap.nn.architectures.hourglass.StemBlock.output_filters"]], "pool (sleap.nn.architectures.hourglass.stemblock attribute)": [[38, "sleap.nn.architectures.hourglass.StemBlock.pool"]], "pooling_stride (sleap.nn.architectures.hourglass.stemblock attribute)": [[38, "sleap.nn.architectures.hourglass.StemBlock.pooling_stride"]], "sleap.nn.architectures.hourglass": [[38, "module-sleap.nn.architectures.hourglass"]], "stacks (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.stacks"]], "stem_filters (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.stem_filters"]], "stem_stack (sleap.nn.architectures.hourglass.hourglass property)": [[38, "sleap.nn.architectures.hourglass.Hourglass.stem_stack"]], "stem_stride (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.stem_stride"]], "up_blocks (sleap.nn.architectures.hourglass.hourglass attribute)": [[38, "sleap.nn.architectures.hourglass.Hourglass.up_blocks"]], "c (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.C"]], "higherhrnet (class in sleap.nn.architectures.hrnet)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet"]], "adjust_prefix() (in module sleap.nn.architectures.hrnet)": [[39, "sleap.nn.architectures.hrnet.adjust_prefix"]], "bilinear_upsampling (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.bilinear_upsampling"]], "bottleneck (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.bottleneck"]], "bottleneck_block() (in module sleap.nn.architectures.hrnet)": [[39, "sleap.nn.architectures.hrnet.bottleneck_block"]], "deconv_filters (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.deconv_filters"]], "down_blocks (sleap.nn.architectures.hrnet.higherhrnet property)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.down_blocks"]], "initial_downsampling_steps (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.initial_downsampling_steps"]], "n_deconv_modules (sleap.nn.architectures.hrnet.higherhrnet attribute)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.n_deconv_modules"]], "output() (sleap.nn.architectures.hrnet.higherhrnet method)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.output"]], "output_scale (sleap.nn.architectures.hrnet.higherhrnet property)": [[39, "sleap.nn.architectures.hrnet.HigherHRNet.output_scale"]], "simple_block() (in module sleap.nn.architectures.hrnet)": [[39, "sleap.nn.architectures.hrnet.simple_block"]], "sleap.nn.architectures.hrnet": [[39, "module-sleap.nn.architectures.hrnet"]], "leapcnn (class in sleap.nn.architectures.leap)": [[40, "sleap.nn.architectures.leap.LeapCNN"]], "decoder_stack (sleap.nn.architectures.leap.leapcnn property)": [[40, "sleap.nn.architectures.leap.LeapCNN.decoder_stack"]], "down_blocks (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.down_blocks"]], "down_convs_per_block (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.down_convs_per_block"]], "encoder_stack (sleap.nn.architectures.leap.leapcnn property)": [[40, "sleap.nn.architectures.leap.LeapCNN.encoder_stack"]], "filters (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.filters"]], "filters_rate (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.filters_rate"]], "from_config() (sleap.nn.architectures.leap.leapcnn class method)": [[40, "sleap.nn.architectures.leap.LeapCNN.from_config"]], "sleap.nn.architectures.leap": [[40, "module-sleap.nn.architectures.leap"]], "up_blocks (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.up_blocks"]], "up_convs_per_block (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.up_convs_per_block"]], "up_interpolate (sleap.nn.architectures.leap.leapcnn attribute)": [[40, "sleap.nn.architectures.leap.LeapCNN.up_interpolate"]], "unetpretrainedencoder (class in sleap.nn.architectures.pretrained_encoders)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder"]], "decoder_filters (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder attribute)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.decoder_filters"]], "down_blocks (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder property)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.down_blocks"]], "encoder (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder attribute)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.encoder"]], "from_config() (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder class method)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.from_config"]], "make_backbone() (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder method)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.make_backbone"]], "maximum_stride (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder property)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.maximum_stride"]], "output_stride (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder property)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.output_stride"]], "pretrained (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder attribute)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.pretrained"]], "sleap.nn.architectures.pretrained_encoders": [[41, "module-sleap.nn.architectures.pretrained_encoders"]], "up_blocks (sleap.nn.architectures.pretrained_encoders.unetpretrainedencoder property)": [[41, "sleap.nn.architectures.pretrained_encoders.UnetPretrainedEncoder.up_blocks"]], "resnet101 (class in sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.ResNet101"]], "resnet152 (class in sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.ResNet152"]], "resnet50 (class in sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.ResNet50"]], "resnetv1 (class in sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.ResNetv1"]], "block_v1() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.block_v1"]], "down_blocks (sleap.nn.architectures.resnet.resnetv1 property)": [[42, "sleap.nn.architectures.resnet.ResNetv1.down_blocks"]], "features_output_stride (sleap.nn.architectures.resnet.resnet101 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet101.features_output_stride"]], "features_output_stride (sleap.nn.architectures.resnet.resnet152 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet152.features_output_stride"]], "features_output_stride (sleap.nn.architectures.resnet.resnet50 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet50.features_output_stride"]], "features_output_stride (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.features_output_stride"]], "from_config() (sleap.nn.architectures.resnet.resnetv1 class method)": [[42, "sleap.nn.architectures.resnet.ResNetv1.from_config"]], "frozen (sleap.nn.architectures.resnet.resnet101 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet101.frozen"]], "frozen (sleap.nn.architectures.resnet.resnet152 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet152.frozen"]], "frozen (sleap.nn.architectures.resnet.resnet50 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet50.frozen"]], "frozen (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.frozen"]], "imagenet_preproc_v1() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.imagenet_preproc_v1"]], "make_backbone() (sleap.nn.architectures.resnet.resnetv1 method)": [[42, "sleap.nn.architectures.resnet.ResNetv1.make_backbone"]], "make_backbone_fn() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.make_backbone_fn"]], "make_resnet_model() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.make_resnet_model"]], "maximum_stride (sleap.nn.architectures.resnet.resnetv1 property)": [[42, "sleap.nn.architectures.resnet.ResNetv1.maximum_stride"]], "model_name (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.model_name"]], "output_scale (sleap.nn.architectures.resnet.resnetv1 property)": [[42, "sleap.nn.architectures.resnet.ResNetv1.output_scale"]], "output_stride (sleap.nn.architectures.resnet.resnetv1 property)": [[42, "sleap.nn.architectures.resnet.ResNetv1.output_stride"]], "pretrained (sleap.nn.architectures.resnet.resnet101 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet101.pretrained"]], "pretrained (sleap.nn.architectures.resnet.resnet152 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet152.pretrained"]], "pretrained (sleap.nn.architectures.resnet.resnet50 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet50.pretrained"]], "pretrained (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.pretrained"]], "skip_connections (sleap.nn.architectures.resnet.resnet101 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet101.skip_connections"]], "skip_connections (sleap.nn.architectures.resnet.resnet152 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet152.skip_connections"]], "skip_connections (sleap.nn.architectures.resnet.resnet50 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet50.skip_connections"]], "skip_connections (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.skip_connections"]], "sleap.nn.architectures.resnet": [[42, "module-sleap.nn.architectures.resnet"]], "stack_configs (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.stack_configs"]], "stack_v1() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.stack_v1"]], "tile_channels() (in module sleap.nn.architectures.resnet)": [[42, "sleap.nn.architectures.resnet.tile_channels"]], "upsampling_stack (sleap.nn.architectures.resnet.resnet101 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet101.upsampling_stack"]], "upsampling_stack (sleap.nn.architectures.resnet.resnet152 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet152.upsampling_stack"]], "upsampling_stack (sleap.nn.architectures.resnet.resnet50 attribute)": [[42, "sleap.nn.architectures.resnet.ResNet50.upsampling_stack"]], "upsampling_stack (sleap.nn.architectures.resnet.resnetv1 attribute)": [[42, "sleap.nn.architectures.resnet.ResNetv1.upsampling_stack"]], "poolingblock (class in sleap.nn.architectures.unet)": [[43, "sleap.nn.architectures.unet.PoolingBlock"]], "unet (class in sleap.nn.architectures.unet)": [[43, "sleap.nn.architectures.unet.UNet"]], "block_contraction (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.block_contraction"]], "convs_per_block (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.convs_per_block"]], "decoder_stack (sleap.nn.architectures.unet.unet property)": [[43, "sleap.nn.architectures.unet.UNet.decoder_stack"]], "down_blocks (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.down_blocks"]], "encoder_stack (sleap.nn.architectures.unet.unet property)": [[43, "sleap.nn.architectures.unet.UNet.encoder_stack"]], "filters (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.filters"]], "filters_rate (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.filters_rate"]], "from_config() (sleap.nn.architectures.unet.unet class method)": [[43, "sleap.nn.architectures.unet.UNet.from_config"]], "kernel_size (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.kernel_size"]], "make_block() (sleap.nn.architectures.unet.poolingblock method)": [[43, "sleap.nn.architectures.unet.PoolingBlock.make_block"]], "middle_block (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.middle_block"]], "pool (sleap.nn.architectures.unet.poolingblock attribute)": [[43, "sleap.nn.architectures.unet.PoolingBlock.pool"]], "pooling_stride (sleap.nn.architectures.unet.poolingblock attribute)": [[43, "sleap.nn.architectures.unet.PoolingBlock.pooling_stride"]], "sleap.nn.architectures.unet": [[43, "module-sleap.nn.architectures.unet"]], "stem_blocks (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.stem_blocks"]], "stem_kernel_size (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.stem_kernel_size"]], "stem_stack (sleap.nn.architectures.unet.unet property)": [[43, "sleap.nn.architectures.unet.UNet.stem_stack"]], "up_blocks (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.up_blocks"]], "up_interpolate (sleap.nn.architectures.unet.unet attribute)": [[43, "sleap.nn.architectures.unet.UNet.up_interpolate"]], "upsamplingstack (class in sleap.nn.architectures.upsampling)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack"]], "from_config() (sleap.nn.architectures.upsampling.upsamplingstack class method)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.from_config"]], "make_skip_connection (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.make_skip_connection"]], "make_stack() (sleap.nn.architectures.upsampling.upsamplingstack method)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.make_stack"]], "output_stride (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.output_stride"]], "refine_convs (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs"]], "refine_convs_batchnorm (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_batchnorm"]], "refine_convs_filters (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters"]], "refine_convs_filters_rate (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.refine_convs_filters_rate"]], "skip_add (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.skip_add"]], "sleap.nn.architectures.upsampling": [[44, "module-sleap.nn.architectures.upsampling"]], "transposed_conv (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv"]], "transposed_conv_batchnorm (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_batchnorm"]], "transposed_conv_filters (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters"]], "transposed_conv_filters_rate (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_filters_rate"]], "transposed_conv_kernel_size (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.transposed_conv_kernel_size"]], "upsampling_stride (sleap.nn.architectures.upsampling.upsamplingstack attribute)": [[44, "sleap.nn.architectures.upsampling.UpsamplingStack.upsampling_stride"]], "matplotlibsaver (class in sleap.nn.callbacks)": [[45, "sleap.nn.callbacks.MatplotlibSaver"]], "modelcheckpointonevent (class in sleap.nn.callbacks)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent"]], "progressreporterzmq (class in sleap.nn.callbacks)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ"]], "tensorboardmatplotlibwriter (class in sleap.nn.callbacks)": [[45, "sleap.nn.callbacks.TensorBoardMatplotlibWriter"]], "trainingcontrollerzmq (class in sleap.nn.callbacks)": [[45, "sleap.nn.callbacks.TrainingControllerZMQ"]], "event (sleap.nn.callbacks.modelcheckpointonevent attribute)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent.event"]], "filepath (sleap.nn.callbacks.modelcheckpointonevent attribute)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent.filepath"]], "logdir (sleap.nn.callbacks.tensorboardmatplotlibwriter attribute)": [[45, "sleap.nn.callbacks.TensorBoardMatplotlibWriter.logdir"]], "on_batch_begin() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_batch_begin"]], "on_batch_end() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_batch_end"]], "on_batch_end() (sleap.nn.callbacks.trainingcontrollerzmq method)": [[45, "sleap.nn.callbacks.TrainingControllerZMQ.on_batch_end"]], "on_epoch_begin() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_begin"]], "on_epoch_end() (sleap.nn.callbacks.matplotlibsaver method)": [[45, "sleap.nn.callbacks.MatplotlibSaver.on_epoch_end"]], "on_epoch_end() (sleap.nn.callbacks.modelcheckpointonevent method)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent.on_epoch_end"]], "on_epoch_end() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_epoch_end"]], "on_epoch_end() (sleap.nn.callbacks.tensorboardmatplotlibwriter method)": [[45, "sleap.nn.callbacks.TensorBoardMatplotlibWriter.on_epoch_end"]], "on_train_begin() (sleap.nn.callbacks.modelcheckpointonevent method)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_begin"]], "on_train_begin() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_train_begin"]], "on_train_end() (sleap.nn.callbacks.modelcheckpointonevent method)": [[45, "sleap.nn.callbacks.ModelCheckpointOnEvent.on_train_end"]], "on_train_end() (sleap.nn.callbacks.progressreporterzmq method)": [[45, "sleap.nn.callbacks.ProgressReporterZMQ.on_train_end"]], "plot_fn (sleap.nn.callbacks.matplotlibsaver attribute)": [[45, "sleap.nn.callbacks.MatplotlibSaver.plot_fn"]], "plot_fn (sleap.nn.callbacks.tensorboardmatplotlibwriter attribute)": [[45, "sleap.nn.callbacks.TensorBoardMatplotlibWriter.plot_fn"]], "prefix (sleap.nn.callbacks.matplotlibsaver attribute)": [[45, "sleap.nn.callbacks.MatplotlibSaver.prefix"]], "save_folder (sleap.nn.callbacks.matplotlibsaver attribute)": [[45, "sleap.nn.callbacks.MatplotlibSaver.save_folder"]], "set_lr() (sleap.nn.callbacks.trainingcontrollerzmq method)": [[45, "sleap.nn.callbacks.TrainingControllerZMQ.set_lr"]], "sleap.nn.callbacks": [[45, "module-sleap.nn.callbacks"]], "tag (sleap.nn.callbacks.tensorboardmatplotlibwriter attribute)": [[45, "sleap.nn.callbacks.TensorBoardMatplotlibWriter.tag"]], "dataconfig (class in sleap.nn.config.data)": [[46, "sleap.nn.config.data.DataConfig"]], "instancecroppingconfig (class in sleap.nn.config.data)": [[46, "sleap.nn.config.data.InstanceCroppingConfig"]], "labelsconfig (class in sleap.nn.config.data)": [[46, "sleap.nn.config.data.LabelsConfig"]], "preprocessingconfig (class in sleap.nn.config.data)": [[46, "sleap.nn.config.data.PreprocessingConfig"]], "center_on_part (sleap.nn.config.data.instancecroppingconfig attribute)": [[46, "sleap.nn.config.data.InstanceCroppingConfig.center_on_part"]], "crop_size (sleap.nn.config.data.instancecroppingconfig attribute)": [[46, "sleap.nn.config.data.InstanceCroppingConfig.crop_size"]], "crop_size_detection_padding (sleap.nn.config.data.instancecroppingconfig attribute)": [[46, "sleap.nn.config.data.InstanceCroppingConfig.crop_size_detection_padding"]], "ensure_grayscale (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.ensure_grayscale"]], "ensure_rgb (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.ensure_rgb"]], "imagenet_mode (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.imagenet_mode"]], "input_scale (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.input_scale"]], "pad_to_stride (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.pad_to_stride"]], "resize_and_pad_to_target (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.resize_and_pad_to_target"]], "search_path_hints (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.search_path_hints"]], "skeletons (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.skeletons"]], "sleap.nn.config.data": [[46, "module-sleap.nn.config.data"]], "split_by_inds (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.split_by_inds"]], "target_height (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.target_height"]], "target_width (sleap.nn.config.data.preprocessingconfig attribute)": [[46, "sleap.nn.config.data.PreprocessingConfig.target_width"]], "test_inds (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.test_inds"]], "test_labels (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.test_labels"]], "training_inds (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.training_inds"]], "training_labels (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.training_labels"]], "validation_fraction (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.validation_fraction"]], "validation_inds (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.validation_inds"]], "validation_labels (sleap.nn.config.data.labelsconfig attribute)": [[46, "sleap.nn.config.data.LabelsConfig.validation_labels"]], "backboneconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.BackboneConfig"]], "centeredinstanceconfmapsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig"]], "centroidsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.CentroidsHeadConfig"]], "classmapsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.ClassMapsHeadConfig"]], "classvectorsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig"]], "headsconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.HeadsConfig"]], "hourglassconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.HourglassConfig"]], "leapconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.LEAPConfig"]], "modelconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.ModelConfig"]], "multiclassbottomupconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.MultiClassBottomUpConfig"]], "multiclasstopdownconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.MultiClassTopDownConfig"]], "multiinstanceconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.MultiInstanceConfig"]], "multiinstanceconfmapsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig"]], "partaffinityfieldsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.PartAffinityFieldsHeadConfig"]], "pretrainedencoderconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig"]], "resnetconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.ResNetConfig"]], "singleinstanceconfmapsheadconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig"]], "unetconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.UNetConfig"]], "upsamplingconfig (class in sleap.nn.config.model)": [[47, "sleap.nn.config.model.UpsamplingConfig"]], "anchor_part (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.anchor_part"]], "anchor_part (sleap.nn.config.model.centroidsheadconfig attribute)": [[47, "sleap.nn.config.model.CentroidsHeadConfig.anchor_part"]], "backbone (sleap.nn.config.model.modelconfig attribute)": [[47, "sleap.nn.config.model.ModelConfig.backbone"]], "base_checkpoint (sleap.nn.config.model.modelconfig attribute)": [[47, "sleap.nn.config.model.ModelConfig.base_checkpoint"]], "block_stride (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.block_stride"]], "centered_instance (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.centered_instance"]], "centroid (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.centroid"]], "class_maps (sleap.nn.config.model.multiclassbottomupconfig attribute)": [[47, "sleap.nn.config.model.MultiClassBottomUpConfig.class_maps"]], "class_vectors (sleap.nn.config.model.multiclasstopdownconfig attribute)": [[47, "sleap.nn.config.model.MultiClassTopDownConfig.class_vectors"]], "classes (sleap.nn.config.model.classmapsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassMapsHeadConfig.classes"]], "classes (sleap.nn.config.model.classvectorsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig.classes"]], "confmaps (sleap.nn.config.model.multiclassbottomupconfig attribute)": [[47, "sleap.nn.config.model.MultiClassBottomUpConfig.confmaps"]], "confmaps (sleap.nn.config.model.multiclasstopdownconfig attribute)": [[47, "sleap.nn.config.model.MultiClassTopDownConfig.confmaps"]], "confmaps (sleap.nn.config.model.multiinstanceconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfig.confmaps"]], "conv_batchnorm (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.conv_batchnorm"]], "decoder_batchnorm (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.decoder_batchnorm"]], "decoder_filters (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters"]], "decoder_filters_rate (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.decoder_filters_rate"]], "edges (sleap.nn.config.model.partaffinityfieldsheadconfig attribute)": [[47, "sleap.nn.config.model.PartAffinityFieldsHeadConfig.edges"]], "encoder (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.encoder"]], "filters (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.filters"]], "filters (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.filters"]], "filters (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.filters"]], "filters (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.filters"]], "filters_increase (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.filters_increase"]], "filters_rate (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.filters_rate"]], "filters_rate (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.filters_rate"]], "filters_rate (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.filters_rate"]], "heads (sleap.nn.config.model.modelconfig attribute)": [[47, "sleap.nn.config.model.ModelConfig.heads"]], "hourglass (sleap.nn.config.model.backboneconfig attribute)": [[47, "sleap.nn.config.model.BackboneConfig.hourglass"]], "leap (sleap.nn.config.model.backboneconfig attribute)": [[47, "sleap.nn.config.model.BackboneConfig.leap"]], "loss_weight (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.centroidsheadconfig attribute)": [[47, "sleap.nn.config.model.CentroidsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.classmapsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassMapsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.classvectorsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.multiinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.partaffinityfieldsheadconfig attribute)": [[47, "sleap.nn.config.model.PartAffinityFieldsHeadConfig.loss_weight"]], "loss_weight (sleap.nn.config.model.singleinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.loss_weight"]], "max_stride (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.max_stride"]], "max_stride (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.max_stride"]], "max_stride (sleap.nn.config.model.resnetconfig attribute)": [[47, "sleap.nn.config.model.ResNetConfig.max_stride"]], "max_stride (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.max_stride"]], "method (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.method"]], "middle_block (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.middle_block"]], "multi_class_bottomup (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.multi_class_bottomup"]], "multi_class_topdown (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.multi_class_topdown"]], "multi_instance (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.multi_instance"]], "num_fc_layers (sleap.nn.config.model.classvectorsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_layers"]], "num_fc_units (sleap.nn.config.model.classvectorsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig.num_fc_units"]], "offset_refinement (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.offset_refinement"]], "offset_refinement (sleap.nn.config.model.centroidsheadconfig attribute)": [[47, "sleap.nn.config.model.CentroidsHeadConfig.offset_refinement"]], "offset_refinement (sleap.nn.config.model.multiinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.offset_refinement"]], "offset_refinement (sleap.nn.config.model.singleinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.offset_refinement"]], "output_stride (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.centroidsheadconfig attribute)": [[47, "sleap.nn.config.model.CentroidsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.classmapsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassMapsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.classvectorsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassVectorsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.output_stride"]], "output_stride (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.output_stride"]], "output_stride (sleap.nn.config.model.multiinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.partaffinityfieldsheadconfig attribute)": [[47, "sleap.nn.config.model.PartAffinityFieldsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.output_stride"]], "output_stride (sleap.nn.config.model.resnetconfig attribute)": [[47, "sleap.nn.config.model.ResNetConfig.output_stride"]], "output_stride (sleap.nn.config.model.singleinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.output_stride"]], "output_stride (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.output_stride"]], "pafs (sleap.nn.config.model.multiinstanceconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfig.pafs"]], "part_names (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.part_names"]], "part_names (sleap.nn.config.model.multiinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.part_names"]], "part_names (sleap.nn.config.model.singleinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.part_names"]], "pretrained (sleap.nn.config.model.pretrainedencoderconfig attribute)": [[47, "sleap.nn.config.model.PretrainedEncoderConfig.pretrained"]], "refine_convs (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.refine_convs"]], "resnet (sleap.nn.config.model.backboneconfig attribute)": [[47, "sleap.nn.config.model.BackboneConfig.resnet"]], "sigma (sleap.nn.config.model.centeredinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.CenteredInstanceConfmapsHeadConfig.sigma"]], "sigma (sleap.nn.config.model.centroidsheadconfig attribute)": [[47, "sleap.nn.config.model.CentroidsHeadConfig.sigma"]], "sigma (sleap.nn.config.model.classmapsheadconfig attribute)": [[47, "sleap.nn.config.model.ClassMapsHeadConfig.sigma"]], "sigma (sleap.nn.config.model.multiinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.MultiInstanceConfmapsHeadConfig.sigma"]], "sigma (sleap.nn.config.model.partaffinityfieldsheadconfig attribute)": [[47, "sleap.nn.config.model.PartAffinityFieldsHeadConfig.sigma"]], "sigma (sleap.nn.config.model.singleinstanceconfmapsheadconfig attribute)": [[47, "sleap.nn.config.model.SingleInstanceConfmapsHeadConfig.sigma"]], "single_instance (sleap.nn.config.model.headsconfig attribute)": [[47, "sleap.nn.config.model.HeadsConfig.single_instance"]], "skip_connections (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.skip_connections"]], "sleap.nn.config.model": [[47, "module-sleap.nn.config.model"]], "stacks (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.stacks"]], "stacks (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.stacks"]], "stacks (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.stacks"]], "stem_stride (sleap.nn.config.model.hourglassconfig attribute)": [[47, "sleap.nn.config.model.HourglassConfig.stem_stride"]], "stem_stride (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.stem_stride"]], "transposed_conv_kernel_size (sleap.nn.config.model.upsamplingconfig attribute)": [[47, "sleap.nn.config.model.UpsamplingConfig.transposed_conv_kernel_size"]], "unet (sleap.nn.config.model.backboneconfig attribute)": [[47, "sleap.nn.config.model.BackboneConfig.unet"]], "up_interpolate (sleap.nn.config.model.leapconfig attribute)": [[47, "sleap.nn.config.model.LEAPConfig.up_interpolate"]], "up_interpolate (sleap.nn.config.model.unetconfig attribute)": [[47, "sleap.nn.config.model.UNetConfig.up_interpolate"]], "upsampling (sleap.nn.config.model.resnetconfig attribute)": [[47, "sleap.nn.config.model.ResNetConfig.upsampling"]], "version (sleap.nn.config.model.resnetconfig attribute)": [[47, "sleap.nn.config.model.ResNetConfig.version"]], "weights (sleap.nn.config.model.resnetconfig attribute)": [[47, "sleap.nn.config.model.ResNetConfig.weights"]], "augmentationconfig (class in sleap.nn.config.optimization)": [[48, "sleap.nn.config.optimization.AugmentationConfig"]], "earlystoppingconfig (class in sleap.nn.config.optimization)": [[48, "sleap.nn.config.optimization.EarlyStoppingConfig"]], "hardkeypointminingconfig (class in sleap.nn.config.optimization)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig"]], "learningratescheduleconfig (class in sleap.nn.config.optimization)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig"]], "optimizationconfig (class in sleap.nn.config.optimization)": [[48, "sleap.nn.config.optimization.OptimizationConfig"]], "augmentation_config (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.augmentation_config"]], "batch_size (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.batch_size"]], "batches_per_epoch (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.batches_per_epoch"]], "brightness (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.brightness"]], "brightness_max_val (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.brightness_max_val"]], "brightness_min_val (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.brightness_min_val"]], "contrast (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.contrast"]], "contrast_max_gamma (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.contrast_max_gamma"]], "contrast_min_gamma (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.contrast_min_gamma"]], "early_stopping (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.early_stopping"]], "epochs (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.epochs"]], "flip_horizontal (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.flip_horizontal"]], "gaussian_noise (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.gaussian_noise"]], "gaussian_noise_mean (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_mean"]], "gaussian_noise_stddev (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.gaussian_noise_stddev"]], "hard_keypoint_mining (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.hard_keypoint_mining"]], "hard_to_easy_ratio (sleap.nn.config.optimization.hardkeypointminingconfig attribute)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig.hard_to_easy_ratio"]], "initial_learning_rate (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.initial_learning_rate"]], "learning_rate_schedule (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.learning_rate_schedule"]], "loss_scale (sleap.nn.config.optimization.hardkeypointminingconfig attribute)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig.loss_scale"]], "max_hard_keypoints (sleap.nn.config.optimization.hardkeypointminingconfig attribute)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig.max_hard_keypoints"]], "min_batches_per_epoch (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.min_batches_per_epoch"]], "min_hard_keypoints (sleap.nn.config.optimization.hardkeypointminingconfig attribute)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig.min_hard_keypoints"]], "min_learning_rate (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.min_learning_rate"]], "min_val_batches_per_epoch (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.min_val_batches_per_epoch"]], "online_mining (sleap.nn.config.optimization.hardkeypointminingconfig attribute)": [[48, "sleap.nn.config.optimization.HardKeypointMiningConfig.online_mining"]], "online_shuffling (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.online_shuffling"]], "optimizer (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.optimizer"]], "plateau_cooldown (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_cooldown"]], "plateau_min_delta (sleap.nn.config.optimization.earlystoppingconfig attribute)": [[48, "sleap.nn.config.optimization.EarlyStoppingConfig.plateau_min_delta"]], "plateau_min_delta (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_min_delta"]], "plateau_patience (sleap.nn.config.optimization.earlystoppingconfig attribute)": [[48, "sleap.nn.config.optimization.EarlyStoppingConfig.plateau_patience"]], "plateau_patience (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.plateau_patience"]], "prefetch (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.prefetch"]], "preload_data (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.preload_data"]], "random_crop (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.random_crop"]], "random_crop_height (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.random_crop_height"]], "random_crop_width (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.random_crop_width"]], "random_flip (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.random_flip"]], "reduce_on_plateau (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.reduce_on_plateau"]], "reduction_factor (sleap.nn.config.optimization.learningratescheduleconfig attribute)": [[48, "sleap.nn.config.optimization.LearningRateScheduleConfig.reduction_factor"]], "rotate (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.rotate"]], "rotation_max_angle (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.rotation_max_angle"]], "rotation_min_angle (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.rotation_min_angle"]], "scale (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.scale"]], "scale_max (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.scale_max"]], "scale_min (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.scale_min"]], "shuffle_buffer_size (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.shuffle_buffer_size"]], "sleap.nn.config.optimization": [[48, "module-sleap.nn.config.optimization"]], "stop_training_on_plateau (sleap.nn.config.optimization.earlystoppingconfig attribute)": [[48, "sleap.nn.config.optimization.EarlyStoppingConfig.stop_training_on_plateau"]], "translate (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.translate"]], "translate_max (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.translate_max"]], "translate_min (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.translate_min"]], "uniform_noise (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.uniform_noise"]], "uniform_noise_max_val (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.uniform_noise_max_val"]], "uniform_noise_min_val (sleap.nn.config.optimization.augmentationconfig attribute)": [[48, "sleap.nn.config.optimization.AugmentationConfig.uniform_noise_min_val"]], "val_batches_per_epoch (sleap.nn.config.optimization.optimizationconfig attribute)": [[48, "sleap.nn.config.optimization.OptimizationConfig.val_batches_per_epoch"]], "checkpointingconfig (class in sleap.nn.config.outputs)": [[49, "sleap.nn.config.outputs.CheckpointingConfig"]], "outputsconfig (class in sleap.nn.config.outputs)": [[49, "sleap.nn.config.outputs.OutputsConfig"]], "tensorboardconfig (class in sleap.nn.config.outputs)": [[49, "sleap.nn.config.outputs.TensorBoardConfig"]], "zmqconfig (class in sleap.nn.config.outputs)": [[49, "sleap.nn.config.outputs.ZMQConfig"]], "architecture_graph (sleap.nn.config.outputs.tensorboardconfig attribute)": [[49, "sleap.nn.config.outputs.TensorBoardConfig.architecture_graph"]], "best_model (sleap.nn.config.outputs.checkpointingconfig attribute)": [[49, "sleap.nn.config.outputs.CheckpointingConfig.best_model"]], "checkpointing (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.checkpointing"]], "controller_address (sleap.nn.config.outputs.zmqconfig attribute)": [[49, "sleap.nn.config.outputs.ZMQConfig.controller_address"]], "controller_polling_timeout (sleap.nn.config.outputs.zmqconfig attribute)": [[49, "sleap.nn.config.outputs.ZMQConfig.controller_polling_timeout"]], "every_epoch (sleap.nn.config.outputs.checkpointingconfig attribute)": [[49, "sleap.nn.config.outputs.CheckpointingConfig.every_epoch"]], "final_model (sleap.nn.config.outputs.checkpointingconfig attribute)": [[49, "sleap.nn.config.outputs.CheckpointingConfig.final_model"]], "initial_model (sleap.nn.config.outputs.checkpointingconfig attribute)": [[49, "sleap.nn.config.outputs.CheckpointingConfig.initial_model"]], "keep_viz_images (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.keep_viz_images"]], "latest_model (sleap.nn.config.outputs.checkpointingconfig attribute)": [[49, "sleap.nn.config.outputs.CheckpointingConfig.latest_model"]], "log_to_csv (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.log_to_csv"]], "loss_frequency (sleap.nn.config.outputs.tensorboardconfig attribute)": [[49, "sleap.nn.config.outputs.TensorBoardConfig.loss_frequency"]], "profile_graph (sleap.nn.config.outputs.tensorboardconfig attribute)": [[49, "sleap.nn.config.outputs.TensorBoardConfig.profile_graph"]], "publish_address (sleap.nn.config.outputs.zmqconfig attribute)": [[49, "sleap.nn.config.outputs.ZMQConfig.publish_address"]], "publish_updates (sleap.nn.config.outputs.zmqconfig attribute)": [[49, "sleap.nn.config.outputs.ZMQConfig.publish_updates"]], "run_name (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.run_name"]], "run_name_prefix (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.run_name_prefix"]], "run_name_suffix (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.run_name_suffix"]], "run_path (sleap.nn.config.outputs.outputsconfig property)": [[49, "sleap.nn.config.outputs.OutputsConfig.run_path"]], "runs_folder (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.runs_folder"]], "save_outputs (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.save_outputs"]], "save_visualizations (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.save_visualizations"]], "sleap.nn.config.outputs": [[49, "module-sleap.nn.config.outputs"]], "subscribe_to_controller (sleap.nn.config.outputs.zmqconfig attribute)": [[49, "sleap.nn.config.outputs.ZMQConfig.subscribe_to_controller"]], "tags (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.tags"]], "tensorboard (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.tensorboard"]], "visualizations (sleap.nn.config.outputs.tensorboardconfig attribute)": [[49, "sleap.nn.config.outputs.TensorBoardConfig.visualizations"]], "write_logs (sleap.nn.config.outputs.tensorboardconfig attribute)": [[49, "sleap.nn.config.outputs.TensorBoardConfig.write_logs"]], "zip_outputs (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.zip_outputs"]], "zmq (sleap.nn.config.outputs.outputsconfig attribute)": [[49, "sleap.nn.config.outputs.OutputsConfig.zmq"]], "trainingjobconfig (class in sleap.nn.config.training_job)": [[50, "sleap.nn.config.training_job.TrainingJobConfig"]], "data (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.data"]], "description (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.description"]], "filename (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.filename"]], "from_json() (sleap.nn.config.training_job.trainingjobconfig class method)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.from_json"]], "from_json_dicts() (sleap.nn.config.training_job.trainingjobconfig class method)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.from_json_dicts"]], "load_config() (in module sleap.nn.config.training_job)": [[50, "sleap.nn.config.training_job.load_config"]], "load_json() (sleap.nn.config.training_job.trainingjobconfig class method)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.load_json"]], "model (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.model"]], "name (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.name"]], "optimization (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.optimization"]], "outputs (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.outputs"]], "save_json() (sleap.nn.config.training_job.trainingjobconfig method)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.save_json"]], "sleap.nn.config.training_job": [[50, "module-sleap.nn.config.training_job"]], "sleap_version (sleap.nn.config.training_job.trainingjobconfig attribute)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.sleap_version"]], "to_json() (sleap.nn.config.training_job.trainingjobconfig method)": [[50, "sleap.nn.config.training_job.TrainingJobConfig.to_json"]], "oneof() (in module sleap.nn.config.utils)": [[51, "sleap.nn.config.utils.oneof"]], "sleap.nn.config.utils": [[51, "module-sleap.nn.config.utils"]], "albumentationsaugmenter (class in sleap.nn.data.augmentation)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter"]], "randomcropper (class in sleap.nn.data.augmentation)": [[52, "sleap.nn.data.augmentation.RandomCropper"]], "randomflipper (class in sleap.nn.data.augmentation)": [[52, "sleap.nn.data.augmentation.RandomFlipper"]], "augmenter (sleap.nn.data.augmentation.albumentationsaugmenter attribute)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.augmenter"]], "crop_height (sleap.nn.data.augmentation.randomcropper attribute)": [[52, "sleap.nn.data.augmentation.RandomCropper.crop_height"]], "crop_width (sleap.nn.data.augmentation.randomcropper attribute)": [[52, "sleap.nn.data.augmentation.RandomCropper.crop_width"]], "flip_instances_lr() (in module sleap.nn.data.augmentation)": [[52, "sleap.nn.data.augmentation.flip_instances_lr"]], "flip_instances_ud() (in module sleap.nn.data.augmentation)": [[52, "sleap.nn.data.augmentation.flip_instances_ud"]], "from_config() (sleap.nn.data.augmentation.albumentationsaugmenter class method)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.from_config"]], "from_skeleton() (sleap.nn.data.augmentation.randomflipper class method)": [[52, "sleap.nn.data.augmentation.RandomFlipper.from_skeleton"]], "horizontal (sleap.nn.data.augmentation.randomflipper attribute)": [[52, "sleap.nn.data.augmentation.RandomFlipper.horizontal"]], "image_key (sleap.nn.data.augmentation.albumentationsaugmenter attribute)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.image_key"]], "input_keys (sleap.nn.data.augmentation.albumentationsaugmenter property)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.input_keys"]], "instances_key (sleap.nn.data.augmentation.albumentationsaugmenter attribute)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.instances_key"]], "output_keys (sleap.nn.data.augmentation.albumentationsaugmenter property)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.output_keys"]], "probability (sleap.nn.data.augmentation.randomflipper attribute)": [[52, "sleap.nn.data.augmentation.RandomFlipper.probability"]], "sleap.nn.data.augmentation": [[52, "module-sleap.nn.data.augmentation"]], "symmetric_inds (sleap.nn.data.augmentation.randomflipper attribute)": [[52, "sleap.nn.data.augmentation.RandomFlipper.symmetric_inds"]], "transform_dataset() (sleap.nn.data.augmentation.albumentationsaugmenter method)": [[52, "sleap.nn.data.augmentation.AlbumentationsAugmenter.transform_dataset"]], "transform_dataset() (sleap.nn.data.augmentation.randomcropper method)": [[52, "sleap.nn.data.augmentation.RandomCropper.transform_dataset"]], "transform_dataset() (sleap.nn.data.augmentation.randomflipper method)": [[52, "sleap.nn.data.augmentation.RandomFlipper.transform_dataset"]], "instanceconfidencemapgenerator (class in sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator"]], "multiconfidencemapgenerator (class in sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator"]], "singleinstanceconfidencemapgenerator (class in sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator"]], "all_instances (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.all_instances"]], "centroids (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.centroids"]], "flatten_offsets (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.flatten_offsets"]], "flatten_offsets (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.flatten_offsets"]], "flatten_offsets (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.flatten_offsets"]], "input_keys (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.input_keys"]], "input_keys (sleap.nn.data.confidence_maps.multiconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.input_keys"]], "input_keys (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.input_keys"]], "make_confmaps() (in module sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.make_confmaps"]], "make_multi_confmaps() (in module sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.make_multi_confmaps"]], "make_multi_confmaps_with_offsets() (in module sleap.nn.data.confidence_maps)": [[53, "sleap.nn.data.confidence_maps.make_multi_confmaps_with_offsets"]], "offsets_threshold (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.offsets_threshold"]], "offsets_threshold (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.offsets_threshold"]], "offsets_threshold (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.offsets_threshold"]], "output_keys (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_keys"]], "output_keys (sleap.nn.data.confidence_maps.multiconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_keys"]], "output_keys (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator property)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_keys"]], "output_stride (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.output_stride"]], "output_stride (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.output_stride"]], "output_stride (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.output_stride"]], "sigma (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.sigma"]], "sigma (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.sigma"]], "sigma (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.sigma"]], "sleap.nn.data.confidence_maps": [[53, "module-sleap.nn.data.confidence_maps"]], "transform_dataset() (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator method)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.transform_dataset"]], "transform_dataset() (sleap.nn.data.confidence_maps.multiconfidencemapgenerator method)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.transform_dataset"]], "transform_dataset() (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator method)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.transform_dataset"]], "with_offsets (sleap.nn.data.confidence_maps.instanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.InstanceConfidenceMapGenerator.with_offsets"]], "with_offsets (sleap.nn.data.confidence_maps.multiconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.MultiConfidenceMapGenerator.with_offsets"]], "with_offsets (sleap.nn.data.confidence_maps.singleinstanceconfidencemapgenerator attribute)": [[53, "sleap.nn.data.confidence_maps.SingleInstanceConfidenceMapGenerator.with_offsets"]], "batcher (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Batcher"]], "lambdafilter (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.LambdaFilter"]], "prefetcher (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Prefetcher"]], "preloader (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Preloader"]], "repeater (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Repeater"]], "shuffler (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Shuffler"]], "unbatcher (class in sleap.nn.data.dataset_ops)": [[54, "sleap.nn.data.dataset_ops.Unbatcher"]], "batch_size (sleap.nn.data.dataset_ops.batcher attribute)": [[54, "sleap.nn.data.dataset_ops.Batcher.batch_size"]], "buffer_size (sleap.nn.data.dataset_ops.prefetcher attribute)": [[54, "sleap.nn.data.dataset_ops.Prefetcher.buffer_size"]], "buffer_size (sleap.nn.data.dataset_ops.shuffler attribute)": [[54, "sleap.nn.data.dataset_ops.Shuffler.buffer_size"]], "drop_remainder (sleap.nn.data.dataset_ops.batcher attribute)": [[54, "sleap.nn.data.dataset_ops.Batcher.drop_remainder"]], "epochs (sleap.nn.data.dataset_ops.repeater attribute)": [[54, "sleap.nn.data.dataset_ops.Repeater.epochs"]], "examples (sleap.nn.data.dataset_ops.preloader attribute)": [[54, "sleap.nn.data.dataset_ops.Preloader.examples"]], "filter_fn (sleap.nn.data.dataset_ops.lambdafilter attribute)": [[54, "sleap.nn.data.dataset_ops.LambdaFilter.filter_fn"]], "input_keys (sleap.nn.data.dataset_ops.batcher property)": [[54, "sleap.nn.data.dataset_ops.Batcher.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.lambdafilter property)": [[54, "sleap.nn.data.dataset_ops.LambdaFilter.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.prefetcher property)": [[54, "sleap.nn.data.dataset_ops.Prefetcher.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.preloader property)": [[54, "sleap.nn.data.dataset_ops.Preloader.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.repeater property)": [[54, "sleap.nn.data.dataset_ops.Repeater.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.shuffler property)": [[54, "sleap.nn.data.dataset_ops.Shuffler.input_keys"]], "input_keys (sleap.nn.data.dataset_ops.unbatcher property)": [[54, "sleap.nn.data.dataset_ops.Unbatcher.input_keys"]], "output_keys (sleap.nn.data.dataset_ops.batcher property)": [[54, "sleap.nn.data.dataset_ops.Batcher.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.lambdafilter property)": [[54, "sleap.nn.data.dataset_ops.LambdaFilter.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.prefetcher property)": [[54, "sleap.nn.data.dataset_ops.Prefetcher.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.preloader property)": [[54, "sleap.nn.data.dataset_ops.Preloader.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.repeater property)": [[54, "sleap.nn.data.dataset_ops.Repeater.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.shuffler property)": [[54, "sleap.nn.data.dataset_ops.Shuffler.output_keys"]], "output_keys (sleap.nn.data.dataset_ops.unbatcher property)": [[54, "sleap.nn.data.dataset_ops.Unbatcher.output_keys"]], "prefetch (sleap.nn.data.dataset_ops.prefetcher attribute)": [[54, "sleap.nn.data.dataset_ops.Prefetcher.prefetch"]], "repeat (sleap.nn.data.dataset_ops.repeater attribute)": [[54, "sleap.nn.data.dataset_ops.Repeater.repeat"]], "reshuffle_each_iteration (sleap.nn.data.dataset_ops.shuffler attribute)": [[54, "sleap.nn.data.dataset_ops.Shuffler.reshuffle_each_iteration"]], "shuffle (sleap.nn.data.dataset_ops.shuffler attribute)": [[54, "sleap.nn.data.dataset_ops.Shuffler.shuffle"]], "sleap.nn.data.dataset_ops": [[54, "module-sleap.nn.data.dataset_ops"]], "transform_dataset() (sleap.nn.data.dataset_ops.batcher method)": [[54, "sleap.nn.data.dataset_ops.Batcher.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.lambdafilter method)": [[54, "sleap.nn.data.dataset_ops.LambdaFilter.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.prefetcher method)": [[54, "sleap.nn.data.dataset_ops.Prefetcher.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.preloader method)": [[54, "sleap.nn.data.dataset_ops.Preloader.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.repeater method)": [[54, "sleap.nn.data.dataset_ops.Repeater.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.shuffler method)": [[54, "sleap.nn.data.dataset_ops.Shuffler.transform_dataset"]], "transform_dataset() (sleap.nn.data.dataset_ops.unbatcher method)": [[54, "sleap.nn.data.dataset_ops.Unbatcher.transform_dataset"]], "unrag (sleap.nn.data.dataset_ops.batcher attribute)": [[54, "sleap.nn.data.dataset_ops.Batcher.unrag"]], "partaffinityfieldsgenerator (class in sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator"]], "distance_to_edge() (in module sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.distance_to_edge"]], "flatten_channels (sleap.nn.data.edge_maps.partaffinityfieldsgenerator attribute)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.flatten_channels"]], "get_edge_points() (in module sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.get_edge_points"]], "input_keys (sleap.nn.data.edge_maps.partaffinityfieldsgenerator property)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.input_keys"]], "make_edge_maps() (in module sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.make_edge_maps"]], "make_multi_pafs() (in module sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.make_multi_pafs"]], "make_pafs() (in module sleap.nn.data.edge_maps)": [[55, "sleap.nn.data.edge_maps.make_pafs"]], "output_keys (sleap.nn.data.edge_maps.partaffinityfieldsgenerator property)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_keys"]], "output_stride (sleap.nn.data.edge_maps.partaffinityfieldsgenerator attribute)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.output_stride"]], "sigma (sleap.nn.data.edge_maps.partaffinityfieldsgenerator attribute)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.sigma"]], "skeletons (sleap.nn.data.edge_maps.partaffinityfieldsgenerator attribute)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.skeletons"]], "sleap.nn.data.edge_maps": [[55, "module-sleap.nn.data.edge_maps"]], "transform_dataset() (sleap.nn.data.edge_maps.partaffinityfieldsgenerator method)": [[55, "sleap.nn.data.edge_maps.PartAffinityFieldsGenerator.transform_dataset"]], "keydevicemover (class in sleap.nn.data.general)": [[56, "sleap.nn.data.general.KeyDeviceMover"]], "keyfilter (class in sleap.nn.data.general)": [[56, "sleap.nn.data.general.KeyFilter"]], "keyrenamer (class in sleap.nn.data.general)": [[56, "sleap.nn.data.general.KeyRenamer"]], "lambdamap (class in sleap.nn.data.general)": [[56, "sleap.nn.data.general.LambdaMap"]], "func (sleap.nn.data.general.lambdamap attribute)": [[56, "sleap.nn.data.general.LambdaMap.func"]], "input_key_names (sleap.nn.data.general.lambdamap attribute)": [[56, "sleap.nn.data.general.LambdaMap.input_key_names"]], "input_keys (sleap.nn.data.general.keydevicemover property)": [[56, "sleap.nn.data.general.KeyDeviceMover.input_keys"]], "input_keys (sleap.nn.data.general.keyfilter property)": [[56, "sleap.nn.data.general.KeyFilter.input_keys"]], "input_keys (sleap.nn.data.general.keyrenamer property)": [[56, "sleap.nn.data.general.KeyRenamer.input_keys"]], "input_keys (sleap.nn.data.general.lambdamap property)": [[56, "sleap.nn.data.general.LambdaMap.input_keys"]], "output_key_names (sleap.nn.data.general.lambdamap attribute)": [[56, "sleap.nn.data.general.LambdaMap.output_key_names"]], "output_keys (sleap.nn.data.general.keydevicemover property)": [[56, "sleap.nn.data.general.KeyDeviceMover.output_keys"]], "output_keys (sleap.nn.data.general.keyfilter property)": [[56, "sleap.nn.data.general.KeyFilter.output_keys"]], "output_keys (sleap.nn.data.general.keyrenamer property)": [[56, "sleap.nn.data.general.KeyRenamer.output_keys"]], "output_keys (sleap.nn.data.general.lambdamap property)": [[56, "sleap.nn.data.general.LambdaMap.output_keys"]], "sleap.nn.data.general": [[56, "module-sleap.nn.data.general"]], "transform_dataset() (sleap.nn.data.general.keydevicemover method)": [[56, "sleap.nn.data.general.KeyDeviceMover.transform_dataset"]], "transform_dataset() (sleap.nn.data.general.keyfilter method)": [[56, "sleap.nn.data.general.KeyFilter.transform_dataset"]], "transform_dataset() (sleap.nn.data.general.keyrenamer method)": [[56, "sleap.nn.data.general.KeyRenamer.transform_dataset"]], "transform_dataset() (sleap.nn.data.general.lambdamap method)": [[56, "sleap.nn.data.general.LambdaMap.transform_dataset"]], "group_examples() (in module sleap.nn.data.grouping)": [[57, "sleap.nn.data.grouping.group_examples"]], "group_examples_iter() (in module sleap.nn.data.grouping)": [[57, "sleap.nn.data.grouping.group_examples_iter"]], "sleap.nn.data.grouping": [[57, "module-sleap.nn.data.grouping"]], "classmapgenerator (class in sleap.nn.data.identity)": [[58, "sleap.nn.data.identity.ClassMapGenerator"]], "classvectorgenerator (class in sleap.nn.data.identity)": [[58, "sleap.nn.data.identity.ClassVectorGenerator"]], "centroids (sleap.nn.data.identity.classmapgenerator attribute)": [[58, "sleap.nn.data.identity.ClassMapGenerator.centroids"]], "class_map_threshold (sleap.nn.data.identity.classmapgenerator attribute)": [[58, "sleap.nn.data.identity.ClassMapGenerator.class_map_threshold"]], "input_keys (sleap.nn.data.identity.classmapgenerator property)": [[58, "sleap.nn.data.identity.ClassMapGenerator.input_keys"]], "input_keys (sleap.nn.data.identity.classvectorgenerator property)": [[58, "sleap.nn.data.identity.ClassVectorGenerator.input_keys"]], "make_class_maps() (in module sleap.nn.data.identity)": [[58, "sleap.nn.data.identity.make_class_maps"]], "make_class_vectors() (in module sleap.nn.data.identity)": [[58, "sleap.nn.data.identity.make_class_vectors"]], "output_keys (sleap.nn.data.identity.classmapgenerator property)": [[58, "sleap.nn.data.identity.ClassMapGenerator.output_keys"]], "output_keys (sleap.nn.data.identity.classvectorgenerator property)": [[58, "sleap.nn.data.identity.ClassVectorGenerator.output_keys"]], "output_stride (sleap.nn.data.identity.classmapgenerator attribute)": [[58, "sleap.nn.data.identity.ClassMapGenerator.output_stride"]], "sigma (sleap.nn.data.identity.classmapgenerator attribute)": [[58, "sleap.nn.data.identity.ClassMapGenerator.sigma"]], "sleap.nn.data.identity": [[58, "module-sleap.nn.data.identity"]], "transform_dataset() (sleap.nn.data.identity.classmapgenerator method)": [[58, "sleap.nn.data.identity.ClassMapGenerator.transform_dataset"]], "transform_dataset() (sleap.nn.data.identity.classvectorgenerator method)": [[58, "sleap.nn.data.identity.ClassVectorGenerator.transform_dataset"]], "globalpeakfinder (class in sleap.nn.data.inference)": [[59, "sleap.nn.data.inference.GlobalPeakFinder"]], "kerasmodelpredictor (class in sleap.nn.data.inference)": [[59, "sleap.nn.data.inference.KerasModelPredictor"]], "localpeakfinder (class in sleap.nn.data.inference)": [[59, "sleap.nn.data.inference.LocalPeakFinder"]], "mockglobalpeakfinder (class in sleap.nn.data.inference)": [[59, "sleap.nn.data.inference.MockGlobalPeakFinder"]], "predictedcenterinstancenormalizer (class in sleap.nn.data.inference)": [[59, "sleap.nn.data.inference.PredictedCenterInstanceNormalizer"]], "input_keys (sleap.nn.data.inference.predictedcenterinstancenormalizer property)": [[59, "sleap.nn.data.inference.PredictedCenterInstanceNormalizer.input_keys"]], "output_keys (sleap.nn.data.inference.predictedcenterinstancenormalizer property)": [[59, "sleap.nn.data.inference.PredictedCenterInstanceNormalizer.output_keys"]], "sleap.nn.data.inference": [[59, "module-sleap.nn.data.inference"]], "transform_dataset() (sleap.nn.data.inference.predictedcenterinstancenormalizer method)": [[59, "sleap.nn.data.inference.PredictedCenterInstanceNormalizer.transform_dataset"]], "instancecentroidfinder (class in sleap.nn.data.instance_centroids)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder"]], "anchor_part_names (sleap.nn.data.instance_centroids.instancecentroidfinder attribute)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.anchor_part_names"]], "center_on_anchor_part (sleap.nn.data.instance_centroids.instancecentroidfinder attribute)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.center_on_anchor_part"]], "find_points_bbox_midpoint() (in module sleap.nn.data.instance_centroids)": [[60, "sleap.nn.data.instance_centroids.find_points_bbox_midpoint"]], "from_config() (sleap.nn.data.instance_centroids.instancecentroidfinder class method)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.from_config"]], "get_instance_anchors() (in module sleap.nn.data.instance_centroids)": [[60, "sleap.nn.data.instance_centroids.get_instance_anchors"]], "input_keys (sleap.nn.data.instance_centroids.instancecentroidfinder property)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.input_keys"]], "instances_key (sleap.nn.data.instance_centroids.instancecentroidfinder attribute)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.instances_key"]], "output_keys (sleap.nn.data.instance_centroids.instancecentroidfinder property)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.output_keys"]], "skeletons (sleap.nn.data.instance_centroids.instancecentroidfinder attribute)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.skeletons"]], "sleap.nn.data.instance_centroids": [[60, "module-sleap.nn.data.instance_centroids"]], "transform_dataset() (sleap.nn.data.instance_centroids.instancecentroidfinder method)": [[60, "sleap.nn.data.instance_centroids.InstanceCentroidFinder.transform_dataset"]], "instancecropper (class in sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper"]], "centroids_key (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.centroids_key"]], "crop_bboxes() (in module sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.crop_bboxes"]], "crop_height (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.crop_height"]], "crop_width (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.crop_width"]], "find_instance_crop_size() (in module sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.find_instance_crop_size"]], "from_config() (sleap.nn.data.instance_cropping.instancecropper class method)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.from_config"]], "image_key (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.image_key"]], "input_keys (sleap.nn.data.instance_cropping.instancecropper property)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.input_keys"]], "instances_key (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.instances_key"]], "keep_full_image (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.keep_full_image"]], "make_centered_bboxes() (in module sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.make_centered_bboxes"]], "mock_centroid_confidence (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.mock_centroid_confidence"]], "normalize_bboxes() (in module sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.normalize_bboxes"]], "output_keys (sleap.nn.data.instance_cropping.instancecropper property)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.output_keys"]], "sleap.nn.data.instance_cropping": [[61, "module-sleap.nn.data.instance_cropping"]], "transform_dataset() (sleap.nn.data.instance_cropping.instancecropper method)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.transform_dataset"]], "unbatch (sleap.nn.data.instance_cropping.instancecropper attribute)": [[61, "sleap.nn.data.instance_cropping.InstanceCropper.unbatch"]], "unnormalize_bboxes() (in module sleap.nn.data.instance_cropping)": [[61, "sleap.nn.data.instance_cropping.unnormalize_bboxes"]], "normalizer (class in sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.Normalizer"]], "convert_rgb_to_bgr() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.convert_rgb_to_bgr"]], "ensure_float (sleap.nn.data.normalization.normalizer attribute)": [[62, "sleap.nn.data.normalization.Normalizer.ensure_float"]], "ensure_float() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.ensure_float"]], "ensure_grayscale (sleap.nn.data.normalization.normalizer attribute)": [[62, "sleap.nn.data.normalization.Normalizer.ensure_grayscale"]], "ensure_grayscale() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.ensure_grayscale"]], "ensure_int() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.ensure_int"]], "ensure_min_image_rank() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.ensure_min_image_rank"]], "ensure_rgb (sleap.nn.data.normalization.normalizer attribute)": [[62, "sleap.nn.data.normalization.Normalizer.ensure_rgb"]], "ensure_rgb() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.ensure_rgb"]], "from_config() (sleap.nn.data.normalization.normalizer class method)": [[62, "sleap.nn.data.normalization.Normalizer.from_config"]], "image_key (sleap.nn.data.normalization.normalizer attribute)": [[62, "sleap.nn.data.normalization.Normalizer.image_key"]], "imagenet_mode (sleap.nn.data.normalization.normalizer attribute)": [[62, "sleap.nn.data.normalization.Normalizer.imagenet_mode"]], "input_keys (sleap.nn.data.normalization.normalizer property)": [[62, "sleap.nn.data.normalization.Normalizer.input_keys"]], "output_keys (sleap.nn.data.normalization.normalizer property)": [[62, "sleap.nn.data.normalization.Normalizer.output_keys"]], "scale_image_range() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.scale_image_range"]], "scale_to_imagenet_caffe_mode() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.scale_to_imagenet_caffe_mode"]], "scale_to_imagenet_tf_mode() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.scale_to_imagenet_tf_mode"]], "scale_to_imagenet_torch_mode() (in module sleap.nn.data.normalization)": [[62, "sleap.nn.data.normalization.scale_to_imagenet_torch_mode"]], "sleap.nn.data.normalization": [[62, "module-sleap.nn.data.normalization"]], "transform_dataset() (sleap.nn.data.normalization.normalizer method)": [[62, "sleap.nn.data.normalization.Normalizer.transform_dataset"]], "make_offsets() (in module sleap.nn.data.offset_regression)": [[63, "sleap.nn.data.offset_regression.make_offsets"]], "mask_offsets() (in module sleap.nn.data.offset_regression)": [[63, "sleap.nn.data.offset_regression.mask_offsets"]], "sleap.nn.data.offset_regression": [[63, "module-sleap.nn.data.offset_regression"]], "bottomupmulticlasspipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline"]], "bottomuppipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline"]], "centroidconfmapspipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline"]], "pipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.Pipeline"]], "singleinstanceconfmapspipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline"]], "topdownmulticlasspipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline"]], "topdownconfmapspipeline (class in sleap.nn.data.pipelines)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline"]], "append() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.append"]], "centroid_confmap_head (sleap.nn.data.pipelines.centroidconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.centroid_confmap_head"]], "class_maps_head (sleap.nn.data.pipelines.bottomupmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.class_maps_head"]], "class_vectors_head (sleap.nn.data.pipelines.topdownmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.class_vectors_head"]], "confmaps_head (sleap.nn.data.pipelines.bottomupmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.confmaps_head"]], "confmaps_head (sleap.nn.data.pipelines.bottomuppipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.confmaps_head"]], "confmaps_head (sleap.nn.data.pipelines.topdownmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.confmaps_head"]], "data_config (sleap.nn.data.pipelines.bottomupmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.data_config"]], "data_config (sleap.nn.data.pipelines.bottomuppipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.data_config"]], "data_config (sleap.nn.data.pipelines.centroidconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.data_config"]], "data_config (sleap.nn.data.pipelines.singleinstanceconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.data_config"]], "data_config (sleap.nn.data.pipelines.topdownmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.data_config"]], "data_config (sleap.nn.data.pipelines.topdownconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.data_config"]], "describe() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.describe"]], "from_blocks() (sleap.nn.data.pipelines.pipeline class method)": [[64, "sleap.nn.data.pipelines.Pipeline.from_blocks"]], "from_pipelines() (sleap.nn.data.pipelines.pipeline class method)": [[64, "sleap.nn.data.pipelines.Pipeline.from_pipelines"]], "instance_confmap_head (sleap.nn.data.pipelines.topdownconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.instance_confmap_head"]], "make_base_pipeline() (sleap.nn.data.pipelines.bottomupmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_base_pipeline"]], "make_base_pipeline() (sleap.nn.data.pipelines.bottomuppipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.make_base_pipeline"]], "make_base_pipeline() (sleap.nn.data.pipelines.centroidconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_base_pipeline"]], "make_base_pipeline() (sleap.nn.data.pipelines.singleinstanceconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_base_pipeline"]], "make_base_pipeline() (sleap.nn.data.pipelines.topdownmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_base_pipeline"]], "make_base_pipeline() (sleap.nn.data.pipelines.topdownconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_base_pipeline"]], "make_dataset() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.make_dataset"]], "make_training_pipeline() (sleap.nn.data.pipelines.bottomupmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_training_pipeline"]], "make_training_pipeline() (sleap.nn.data.pipelines.bottomuppipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.make_training_pipeline"]], "make_training_pipeline() (sleap.nn.data.pipelines.centroidconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_training_pipeline"]], "make_training_pipeline() (sleap.nn.data.pipelines.singleinstanceconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_training_pipeline"]], "make_training_pipeline() (sleap.nn.data.pipelines.topdownmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_training_pipeline"]], "make_training_pipeline() (sleap.nn.data.pipelines.topdownconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_training_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.bottomupmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.make_viz_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.bottomuppipeline method)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.make_viz_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.centroidconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.make_viz_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.singleinstanceconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.make_viz_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.topdownmulticlasspipeline method)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.make_viz_pipeline"]], "make_viz_pipeline() (sleap.nn.data.pipelines.topdownconfmapspipeline method)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.make_viz_pipeline"]], "offsets_head (sleap.nn.data.pipelines.bottomupmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.offsets_head"]], "offsets_head (sleap.nn.data.pipelines.bottomuppipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.offsets_head"]], "offsets_head (sleap.nn.data.pipelines.centroidconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.offsets_head"]], "offsets_head (sleap.nn.data.pipelines.singleinstanceconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.offsets_head"]], "offsets_head (sleap.nn.data.pipelines.topdownmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.offsets_head"]], "offsets_head (sleap.nn.data.pipelines.topdownconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.offsets_head"]], "optimization_config (sleap.nn.data.pipelines.bottomupmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpMultiClassPipeline.optimization_config"]], "optimization_config (sleap.nn.data.pipelines.bottomuppipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.optimization_config"]], "optimization_config (sleap.nn.data.pipelines.centroidconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.CentroidConfmapsPipeline.optimization_config"]], "optimization_config (sleap.nn.data.pipelines.singleinstanceconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.optimization_config"]], "optimization_config (sleap.nn.data.pipelines.topdownmulticlasspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopDownMultiClassPipeline.optimization_config"]], "optimization_config (sleap.nn.data.pipelines.topdownconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.TopdownConfmapsPipeline.optimization_config"]], "output_keys (sleap.nn.data.pipelines.pipeline property)": [[64, "sleap.nn.data.pipelines.Pipeline.output_keys"]], "pafs_head (sleap.nn.data.pipelines.bottomuppipeline attribute)": [[64, "sleap.nn.data.pipelines.BottomUpPipeline.pafs_head"]], "peek() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.peek"]], "providers (sleap.nn.data.pipelines.pipeline attribute)": [[64, "sleap.nn.data.pipelines.Pipeline.providers"]], "run() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.run"]], "single_instance_confmap_head (sleap.nn.data.pipelines.singleinstanceconfmapspipeline attribute)": [[64, "sleap.nn.data.pipelines.SingleInstanceConfmapsPipeline.single_instance_confmap_head"]], "sleap.nn.data.pipelines": [[64, "module-sleap.nn.data.pipelines"]], "transformers (sleap.nn.data.pipelines.pipeline attribute)": [[64, "sleap.nn.data.pipelines.Pipeline.transformers"]], "validate_pipeline() (sleap.nn.data.pipelines.pipeline method)": [[64, "sleap.nn.data.pipelines.Pipeline.validate_pipeline"]], "labelsreader (class in sleap.nn.data.providers)": [[65, "sleap.nn.data.providers.LabelsReader"]], "videoreader (class in sleap.nn.data.providers)": [[65, "sleap.nn.data.providers.VideoReader"]], "example_indices (sleap.nn.data.providers.labelsreader attribute)": [[65, "sleap.nn.data.providers.LabelsReader.example_indices"]], "example_indices (sleap.nn.data.providers.videoreader attribute)": [[65, "sleap.nn.data.providers.VideoReader.example_indices"]], "from_filename() (sleap.nn.data.providers.labelsreader class method)": [[65, "sleap.nn.data.providers.LabelsReader.from_filename"]], "from_filepath() (sleap.nn.data.providers.videoreader class method)": [[65, "sleap.nn.data.providers.VideoReader.from_filepath"]], "from_unlabeled_suggestions() (sleap.nn.data.providers.labelsreader class method)": [[65, "sleap.nn.data.providers.LabelsReader.from_unlabeled_suggestions"]], "from_user_instances() (sleap.nn.data.providers.labelsreader class method)": [[65, "sleap.nn.data.providers.LabelsReader.from_user_instances"]], "from_user_labeled_frames() (sleap.nn.data.providers.labelsreader class method)": [[65, "sleap.nn.data.providers.LabelsReader.from_user_labeled_frames"]], "is_from_multi_size_videos (sleap.nn.data.providers.labelsreader property)": [[65, "sleap.nn.data.providers.LabelsReader.is_from_multi_size_videos"]], "labels (sleap.nn.data.providers.labelsreader attribute)": [[65, "sleap.nn.data.providers.LabelsReader.labels"]], "make_dataset() (sleap.nn.data.providers.labelsreader method)": [[65, "sleap.nn.data.providers.LabelsReader.make_dataset"]], "make_dataset() (sleap.nn.data.providers.videoreader method)": [[65, "sleap.nn.data.providers.VideoReader.make_dataset"]], "max_height_and_width (sleap.nn.data.providers.labelsreader property)": [[65, "sleap.nn.data.providers.LabelsReader.max_height_and_width"]], "max_height_and_width (sleap.nn.data.providers.videoreader property)": [[65, "sleap.nn.data.providers.VideoReader.max_height_and_width"]], "output_keys (sleap.nn.data.providers.labelsreader property)": [[65, "sleap.nn.data.providers.LabelsReader.output_keys"]], "output_keys (sleap.nn.data.providers.videoreader property)": [[65, "sleap.nn.data.providers.VideoReader.output_keys"]], "sleap.nn.data.providers": [[65, "module-sleap.nn.data.providers"]], "tracks (sleap.nn.data.providers.labelsreader property)": [[65, "sleap.nn.data.providers.LabelsReader.tracks"]], "user_instances_only (sleap.nn.data.providers.labelsreader attribute)": [[65, "sleap.nn.data.providers.LabelsReader.user_instances_only"]], "video (sleap.nn.data.providers.videoreader attribute)": [[65, "sleap.nn.data.providers.VideoReader.video"]], "video_ind (sleap.nn.data.providers.videoreader attribute)": [[65, "sleap.nn.data.providers.VideoReader.video_ind"]], "videos (sleap.nn.data.providers.labelsreader property)": [[65, "sleap.nn.data.providers.LabelsReader.videos"]], "videos (sleap.nn.data.providers.videoreader property)": [[65, "sleap.nn.data.providers.VideoReader.videos"]], "with_track_only (sleap.nn.data.providers.labelsreader attribute)": [[65, "sleap.nn.data.providers.LabelsReader.with_track_only"]], "pointsrescaler (class in sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.PointsRescaler"]], "resizer (class in sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.Resizer"]], "sizematcher (class in sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.SizeMatcher"]], "center_pad (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.center_pad"]], "find_padding_for_stride() (in module sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.find_padding_for_stride"]], "from_config() (sleap.nn.data.resizing.resizer class method)": [[66, "sleap.nn.data.resizing.Resizer.from_config"]], "from_config() (sleap.nn.data.resizing.sizematcher class method)": [[66, "sleap.nn.data.resizing.SizeMatcher.from_config"]], "full_image_key (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.full_image_key"]], "full_image_key (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.full_image_key"]], "image_key (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.image_key"]], "image_key (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.image_key"]], "input_keys (sleap.nn.data.resizing.pointsrescaler property)": [[66, "sleap.nn.data.resizing.PointsRescaler.input_keys"]], "input_keys (sleap.nn.data.resizing.resizer property)": [[66, "sleap.nn.data.resizing.Resizer.input_keys"]], "input_keys (sleap.nn.data.resizing.sizematcher property)": [[66, "sleap.nn.data.resizing.SizeMatcher.input_keys"]], "keep_full_image (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.keep_full_image"]], "keep_full_image (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.keep_full_image"]], "max_image_height (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.max_image_height"]], "max_image_width (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.max_image_width"]], "output_keys (sleap.nn.data.resizing.pointsrescaler property)": [[66, "sleap.nn.data.resizing.PointsRescaler.output_keys"]], "output_keys (sleap.nn.data.resizing.resizer property)": [[66, "sleap.nn.data.resizing.Resizer.output_keys"]], "output_keys (sleap.nn.data.resizing.sizematcher property)": [[66, "sleap.nn.data.resizing.SizeMatcher.output_keys"]], "pad_to_stride (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.pad_to_stride"]], "pad_to_stride() (in module sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.pad_to_stride"]], "points_key (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.points_key"]], "points_key (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.points_key"]], "resize_image() (in module sleap.nn.data.resizing)": [[66, "sleap.nn.data.resizing.resize_image"]], "scale (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.scale"]], "scale_key (sleap.nn.data.resizing.resizer attribute)": [[66, "sleap.nn.data.resizing.Resizer.scale_key"]], "scale_key (sleap.nn.data.resizing.sizematcher attribute)": [[66, "sleap.nn.data.resizing.SizeMatcher.scale_key"]], "sleap.nn.data.resizing": [[66, "module-sleap.nn.data.resizing"]], "transform_dataset() (sleap.nn.data.resizing.pointsrescaler method)": [[66, "sleap.nn.data.resizing.PointsRescaler.transform_dataset"]], "transform_dataset() (sleap.nn.data.resizing.resizer method)": [[66, "sleap.nn.data.resizing.Resizer.transform_dataset"]], "transform_dataset() (sleap.nn.data.resizing.sizematcher method)": [[66, "sleap.nn.data.resizing.SizeMatcher.transform_dataset"]], "keymapper (class in sleap.nn.data.training)": [[67, "sleap.nn.data.training.KeyMapper"]], "input_keys (sleap.nn.data.training.keymapper property)": [[67, "sleap.nn.data.training.KeyMapper.input_keys"]], "key_maps (sleap.nn.data.training.keymapper attribute)": [[67, "sleap.nn.data.training.KeyMapper.key_maps"]], "output_keys (sleap.nn.data.training.keymapper property)": [[67, "sleap.nn.data.training.KeyMapper.output_keys"]], "sleap.nn.data.training": [[67, "module-sleap.nn.data.training"]], "split_labels() (in module sleap.nn.data.training)": [[67, "sleap.nn.data.training.split_labels"]], "split_labels_reader() (in module sleap.nn.data.training)": [[67, "sleap.nn.data.training.split_labels_reader"]], "split_labels_train_val() (in module sleap.nn.data.training)": [[67, "sleap.nn.data.training.split_labels_train_val"]], "transform_dataset() (sleap.nn.data.training.keymapper method)": [[67, "sleap.nn.data.training.KeyMapper.transform_dataset"]], "describe_tensors() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.describe_tensors"]], "ensure_list() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.ensure_list"]], "expand_to_rank() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.expand_to_rank"]], "gaussian_pdf() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.gaussian_pdf"]], "make_grid_vectors() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.make_grid_vectors"]], "sleap.nn.data.utils": [[68, "module-sleap.nn.data.utils"]], "unrag_example() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.unrag_example"]], "unrag_tensor() (in module sleap.nn.data.utils)": [[68, "sleap.nn.data.utils.unrag_tensor"]], "compute_dist_metrics() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_dist_metrics"]], "compute_dists() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_dists"]], "compute_generalized_voc_metrics() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_generalized_voc_metrics"]], "compute_instance_area() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_instance_area"]], "compute_oks() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_oks"]], "compute_pck_metrics() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_pck_metrics"]], "compute_visibility_conf() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.compute_visibility_conf"]], "evaluate() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.evaluate"]], "evaluate_model() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.evaluate_model"]], "find_frame_pairs() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.find_frame_pairs"]], "load_metrics() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.load_metrics"]], "match_frame_pairs() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.match_frame_pairs"]], "match_instances() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.match_instances"]], "replace_path() (in module sleap.nn.evals)": [[69, "sleap.nn.evals.replace_path"]], "sleap.nn.evals": [[69, "module-sleap.nn.evals"]], "centeredinstanceconfmapshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead"]], "centroidconfmapshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.CentroidConfmapsHead"]], "classmapshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.ClassMapsHead"]], "classvectorshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.ClassVectorsHead"]], "head (class in sleap.nn.heads)": [[70, "sleap.nn.heads.Head"]], "multiinstanceconfmapshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead"]], "offsetrefinementhead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.OffsetRefinementHead"]], "partaffinityfieldshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.PartAffinityFieldsHead"]], "singleinstanceconfmapshead (class in sleap.nn.heads)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead"]], "activation (sleap.nn.heads.classmapshead property)": [[70, "sleap.nn.heads.ClassMapsHead.activation"]], "activation (sleap.nn.heads.classvectorshead property)": [[70, "sleap.nn.heads.ClassVectorsHead.activation"]], "activation (sleap.nn.heads.head property)": [[70, "sleap.nn.heads.Head.activation"]], "anchor_part (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.anchor_part"]], "anchor_part (sleap.nn.heads.centroidconfmapshead attribute)": [[70, "sleap.nn.heads.CentroidConfmapsHead.anchor_part"]], "channels (sleap.nn.heads.centeredinstanceconfmapshead property)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.channels"]], "channels (sleap.nn.heads.centroidconfmapshead property)": [[70, "sleap.nn.heads.CentroidConfmapsHead.channels"]], "channels (sleap.nn.heads.classmapshead property)": [[70, "sleap.nn.heads.ClassMapsHead.channels"]], "channels (sleap.nn.heads.classvectorshead property)": [[70, "sleap.nn.heads.ClassVectorsHead.channels"]], "channels (sleap.nn.heads.head property)": [[70, "sleap.nn.heads.Head.channels"]], "channels (sleap.nn.heads.multiinstanceconfmapshead property)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.channels"]], "channels (sleap.nn.heads.offsetrefinementhead property)": [[70, "sleap.nn.heads.OffsetRefinementHead.channels"]], "channels (sleap.nn.heads.partaffinityfieldshead property)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.channels"]], "channels (sleap.nn.heads.singleinstanceconfmapshead property)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.channels"]], "classes (sleap.nn.heads.classmapshead attribute)": [[70, "id1"], [70, "sleap.nn.heads.ClassMapsHead.classes"]], "classes (sleap.nn.heads.classvectorshead attribute)": [[70, "id2"], [70, "sleap.nn.heads.ClassVectorsHead.classes"]], "config (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.config"]], "config (sleap.nn.heads.centroidconfmapshead attribute)": [[70, "sleap.nn.heads.CentroidConfmapsHead.config"]], "config (sleap.nn.heads.classmapshead attribute)": [[70, "sleap.nn.heads.ClassMapsHead.config"]], "config (sleap.nn.heads.classvectorshead attribute)": [[70, "sleap.nn.heads.ClassVectorsHead.config"]], "config (sleap.nn.heads.multiinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.config"]], "config (sleap.nn.heads.offsetrefinementhead attribute)": [[70, "sleap.nn.heads.OffsetRefinementHead.config"]], "config (sleap.nn.heads.partaffinityfieldshead attribute)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.config"]], "config (sleap.nn.heads.singleinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.config"]], "edges (sleap.nn.heads.partaffinityfieldshead attribute)": [[70, "id7"], [70, "sleap.nn.heads.PartAffinityFieldsHead.edges"]], "from_config() (sleap.nn.heads.centeredinstanceconfmapshead class method)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.from_config"]], "from_config() (sleap.nn.heads.centroidconfmapshead class method)": [[70, "sleap.nn.heads.CentroidConfmapsHead.from_config"]], "from_config() (sleap.nn.heads.classmapshead class method)": [[70, "sleap.nn.heads.ClassMapsHead.from_config"]], "from_config() (sleap.nn.heads.classvectorshead class method)": [[70, "sleap.nn.heads.ClassVectorsHead.from_config"]], "from_config() (sleap.nn.heads.multiinstanceconfmapshead class method)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.from_config"]], "from_config() (sleap.nn.heads.offsetrefinementhead class method)": [[70, "sleap.nn.heads.OffsetRefinementHead.from_config"]], "from_config() (sleap.nn.heads.partaffinityfieldshead class method)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.from_config"]], "from_config() (sleap.nn.heads.singleinstanceconfmapshead class method)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.from_config"]], "loss_function (sleap.nn.heads.classvectorshead property)": [[70, "sleap.nn.heads.ClassVectorsHead.loss_function"]], "loss_function (sleap.nn.heads.head property)": [[70, "sleap.nn.heads.Head.loss_function"]], "loss_weight (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.loss_weight"]], "loss_weight (sleap.nn.heads.centroidconfmapshead attribute)": [[70, "sleap.nn.heads.CentroidConfmapsHead.loss_weight"]], "loss_weight (sleap.nn.heads.classmapshead attribute)": [[70, "sleap.nn.heads.ClassMapsHead.loss_weight"]], "loss_weight (sleap.nn.heads.classvectorshead attribute)": [[70, "sleap.nn.heads.ClassVectorsHead.loss_weight"]], "loss_weight (sleap.nn.heads.multiinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.loss_weight"]], "loss_weight (sleap.nn.heads.offsetrefinementhead attribute)": [[70, "id6"], [70, "sleap.nn.heads.OffsetRefinementHead.loss_weight"]], "loss_weight (sleap.nn.heads.partaffinityfieldshead attribute)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.loss_weight"]], "loss_weight (sleap.nn.heads.singleinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.loss_weight"]], "make_head() (sleap.nn.heads.classvectorshead method)": [[70, "sleap.nn.heads.ClassVectorsHead.make_head"]], "make_head() (sleap.nn.heads.head method)": [[70, "sleap.nn.heads.Head.make_head"]], "num_fc_layers (sleap.nn.heads.classvectorshead attribute)": [[70, "sleap.nn.heads.ClassVectorsHead.num_fc_layers"]], "num_fc_units (sleap.nn.heads.classvectorshead attribute)": [[70, "sleap.nn.heads.ClassVectorsHead.num_fc_units"]], "output_stride (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.output_stride"]], "output_stride (sleap.nn.heads.centroidconfmapshead attribute)": [[70, "sleap.nn.heads.CentroidConfmapsHead.output_stride"]], "output_stride (sleap.nn.heads.classmapshead attribute)": [[70, "sleap.nn.heads.ClassMapsHead.output_stride"]], "output_stride (sleap.nn.heads.classvectorshead attribute)": [[70, "sleap.nn.heads.ClassVectorsHead.output_stride"]], "output_stride (sleap.nn.heads.multiinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.output_stride"]], "output_stride (sleap.nn.heads.offsetrefinementhead attribute)": [[70, "sleap.nn.heads.OffsetRefinementHead.output_stride"]], "output_stride (sleap.nn.heads.partaffinityfieldshead attribute)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.output_stride"]], "output_stride (sleap.nn.heads.singleinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.output_stride"]], "part_names (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "id0"], [70, "sleap.nn.heads.CenteredInstanceConfmapsHead.part_names"]], "part_names (sleap.nn.heads.multiinstanceconfmapshead attribute)": [[70, "id3"], [70, "sleap.nn.heads.MultiInstanceConfmapsHead.part_names"]], "part_names (sleap.nn.heads.offsetrefinementhead attribute)": [[70, "id4"], [70, "sleap.nn.heads.OffsetRefinementHead.part_names"]], "part_names (sleap.nn.heads.singleinstanceconfmapshead attribute)": [[70, "id8"], [70, "sleap.nn.heads.SingleInstanceConfmapsHead.part_names"]], "sigma (sleap.nn.heads.centeredinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.CenteredInstanceConfmapsHead.sigma"]], "sigma (sleap.nn.heads.centroidconfmapshead attribute)": [[70, "sleap.nn.heads.CentroidConfmapsHead.sigma"]], "sigma (sleap.nn.heads.classmapshead attribute)": [[70, "sleap.nn.heads.ClassMapsHead.sigma"]], "sigma (sleap.nn.heads.multiinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.MultiInstanceConfmapsHead.sigma"]], "sigma (sleap.nn.heads.partaffinityfieldshead attribute)": [[70, "sleap.nn.heads.PartAffinityFieldsHead.sigma"]], "sigma (sleap.nn.heads.singleinstanceconfmapshead attribute)": [[70, "sleap.nn.heads.SingleInstanceConfmapsHead.sigma"]], "sigma_threshold (sleap.nn.heads.offsetrefinementhead attribute)": [[70, "id5"], [70, "sleap.nn.heads.OffsetRefinementHead.sigma_threshold"]], "sleap.nn.heads": [[70, "module-sleap.nn.heads"]], "classify_peaks_from_maps() (in module sleap.nn.identity)": [[71, "sleap.nn.identity.classify_peaks_from_maps"]], "classify_peaks_from_vectors() (in module sleap.nn.identity)": [[71, "sleap.nn.identity.classify_peaks_from_vectors"]], "group_class_peaks() (in module sleap.nn.identity)": [[71, "sleap.nn.identity.group_class_peaks"]], "sleap.nn.identity": [[71, "module-sleap.nn.identity"]], "bottomupinferencelayer (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpInferenceLayer"]], "bottomupinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpInferenceModel"]], "bottomupmulticlassinferencelayer (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer"]], "bottomupmulticlassinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceModel"]], "bottomupmulticlasspredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor"]], "bottomuppredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.BottomUpPredictor"]], "centroidcrop (class in sleap.nn.inference)": [[72, "sleap.nn.inference.CentroidCrop"]], "centroidcropgroundtruth (class in sleap.nn.inference)": [[72, "sleap.nn.inference.CentroidCropGroundTruth"]], "centroidinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.CentroidInferenceModel"]], "findinstancepeaks (class in sleap.nn.inference)": [[72, "sleap.nn.inference.FindInstancePeaks"]], "findinstancepeaksgroundtruth (class in sleap.nn.inference)": [[72, "sleap.nn.inference.FindInstancePeaksGroundTruth"]], "inferencelayer (class in sleap.nn.inference)": [[72, "sleap.nn.inference.InferenceLayer"]], "inferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.InferenceModel"]], "movenetinferencelayer (class in sleap.nn.inference)": [[72, "sleap.nn.inference.MoveNetInferenceLayer"]], "movenetinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.MoveNetInferenceModel"]], "movenetpredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.MoveNetPredictor"]], "predictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.Predictor"]], "ratecolumn (class in sleap.nn.inference)": [[72, "sleap.nn.inference.RateColumn"]], "singleinstanceinferencelayer (class in sleap.nn.inference)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer"]], "singleinstanceinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.SingleInstanceInferenceModel"]], "singleinstancepredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.SingleInstancePredictor"]], "topdowninferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.TopDownInferenceModel"]], "topdownmulticlassfindpeaks (class in sleap.nn.inference)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks"]], "topdownmulticlassinferencemodel (class in sleap.nn.inference)": [[72, "sleap.nn.inference.TopDownMultiClassInferenceModel"]], "topdownmulticlasspredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor"]], "topdownpredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.TopDownPredictor"]], "visualpredictor (class in sleap.nn.inference)": [[72, "sleap.nn.inference.VisualPredictor"]], "batch_size (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.batch_size"]], "batch_size (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.batch_size"]], "batch_size (sleap.nn.inference.movenetpredictor attribute)": [[72, "sleap.nn.inference.MoveNetPredictor.batch_size"]], "batch_size (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.batch_size"]], "batch_size (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.batch_size"]], "batch_size (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.batch_size"]], "bottomup_config (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.bottomup_config"]], "bottomup_layer (sleap.nn.inference.bottomupinferencemodel attribute)": [[72, "sleap.nn.inference.BottomUpInferenceModel.bottomup_layer"]], "bottomup_model (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.bottomup_model"]], "call() (sleap.nn.inference.bottomupinferencelayer method)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.call"]], "call() (sleap.nn.inference.bottomupinferencemodel method)": [[72, "sleap.nn.inference.BottomUpInferenceModel.call"]], "call() (sleap.nn.inference.bottomupmulticlassinferencelayer method)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.call"]], "call() (sleap.nn.inference.bottomupmulticlassinferencemodel method)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceModel.call"]], "call() (sleap.nn.inference.centroidcrop method)": [[72, "sleap.nn.inference.CentroidCrop.call"]], "call() (sleap.nn.inference.centroidcropgroundtruth method)": [[72, "sleap.nn.inference.CentroidCropGroundTruth.call"]], "call() (sleap.nn.inference.centroidinferencemodel method)": [[72, "sleap.nn.inference.CentroidInferenceModel.call"]], "call() (sleap.nn.inference.findinstancepeaks method)": [[72, "sleap.nn.inference.FindInstancePeaks.call"]], "call() (sleap.nn.inference.findinstancepeaksgroundtruth method)": [[72, "sleap.nn.inference.FindInstancePeaksGroundTruth.call"]], "call() (sleap.nn.inference.inferencelayer method)": [[72, "sleap.nn.inference.InferenceLayer.call"]], "call() (sleap.nn.inference.movenetinferencelayer method)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.call"]], "call() (sleap.nn.inference.movenetinferencemodel method)": [[72, "sleap.nn.inference.MoveNetInferenceModel.call"]], "call() (sleap.nn.inference.singleinstanceinferencelayer method)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.call"]], "call() (sleap.nn.inference.singleinstanceinferencemodel method)": [[72, "sleap.nn.inference.SingleInstanceInferenceModel.call"]], "call() (sleap.nn.inference.topdowninferencemodel method)": [[72, "sleap.nn.inference.TopDownInferenceModel.call"]], "call() (sleap.nn.inference.topdownmulticlassfindpeaks method)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.call"]], "call() (sleap.nn.inference.topdownmulticlassinferencemodel method)": [[72, "sleap.nn.inference.TopDownMultiClassInferenceModel.call"]], "centroid_config (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.centroid_config"]], "centroid_config (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.centroid_config"]], "centroid_crop (sleap.nn.inference.centroidinferencemodel attribute)": [[72, "sleap.nn.inference.CentroidInferenceModel.centroid_crop"]], "centroid_crop (sleap.nn.inference.topdowninferencemodel attribute)": [[72, "sleap.nn.inference.TopDownInferenceModel.centroid_crop"]], "centroid_crop (sleap.nn.inference.topdownmulticlassinferencemodel attribute)": [[72, "sleap.nn.inference.TopDownMultiClassInferenceModel.centroid_crop"]], "centroid_model (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.centroid_model"]], "centroid_model (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.centroid_model"]], "class_maps_ind (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_ind"]], "class_maps_output_stride (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.class_maps_output_stride"]], "class_vectors_ind (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.class_vectors_ind"]], "cm_output_stride (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.cm_output_stride"]], "cm_output_stride (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.cm_output_stride"]], "config (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.config"]], "confmap_config (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.confmap_config"]], "confmap_config (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.confmap_config"]], "confmap_config (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.confmap_config"]], "confmap_model (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.confmap_model"]], "confmap_model (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.confmap_model"]], "confmap_model (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.confmap_model"]], "confmaps_ind (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.confmaps_ind"]], "confmaps_ind (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.confmaps_ind"]], "confmaps_ind (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.confmaps_ind"]], "confmaps_ind (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.confmaps_ind"]], "confmaps_ind (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.confmaps_ind"]], "confmaps_ind (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.confmaps_ind"]], "crop_size (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.crop_size"]], "crop_size (sleap.nn.inference.centroidcropgroundtruth attribute)": [[72, "sleap.nn.inference.CentroidCropGroundTruth.crop_size"]], "dist_penalty_weight (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.dist_penalty_weight"]], "ensure_float (sleap.nn.inference.inferencelayer attribute)": [[72, "sleap.nn.inference.InferenceLayer.ensure_float"]], "ensure_float (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.ensure_float"]], "ensure_grayscale (sleap.nn.inference.inferencelayer attribute)": [[72, "sleap.nn.inference.InferenceLayer.ensure_grayscale"]], "ensure_grayscale (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.ensure_grayscale"]], "export_cli() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.export_cli"]], "export_model() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.export_model"]], "export_model() (sleap.nn.inference.inferencemodel method)": [[72, "sleap.nn.inference.InferenceModel.export_model"]], "export_model() (sleap.nn.inference.predictor method)": [[72, "sleap.nn.inference.Predictor.export_model"]], "export_model() (sleap.nn.inference.singleinstancepredictor method)": [[72, "sleap.nn.inference.SingleInstancePredictor.export_model"]], "export_model() (sleap.nn.inference.topdownmulticlassinferencemodel method)": [[72, "sleap.nn.inference.TopDownMultiClassInferenceModel.export_model"]], "export_model() (sleap.nn.inference.topdownmulticlasspredictor method)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.export_model"]], "export_model() (sleap.nn.inference.topdownpredictor method)": [[72, "sleap.nn.inference.TopDownPredictor.export_model"]], "find_head() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.find_head"]], "find_peaks() (sleap.nn.inference.bottomupinferencelayer method)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.find_peaks"]], "find_peaks() (sleap.nn.inference.bottomupmulticlassinferencelayer method)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.find_peaks"]], "forward_pass() (sleap.nn.inference.bottomupinferencelayer method)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.forward_pass"]], "forward_pass() (sleap.nn.inference.bottomupmulticlassinferencelayer method)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.forward_pass"]], "from_model_paths() (sleap.nn.inference.predictor class method)": [[72, "sleap.nn.inference.Predictor.from_model_paths"]], "from_trained_models() (sleap.nn.inference.bottomupmulticlasspredictor class method)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.from_trained_models"]], "from_trained_models() (sleap.nn.inference.bottomuppredictor class method)": [[72, "sleap.nn.inference.BottomUpPredictor.from_trained_models"]], "from_trained_models() (sleap.nn.inference.movenetpredictor class method)": [[72, "sleap.nn.inference.MoveNetPredictor.from_trained_models"]], "from_trained_models() (sleap.nn.inference.singleinstancepredictor class method)": [[72, "sleap.nn.inference.SingleInstancePredictor.from_trained_models"]], "from_trained_models() (sleap.nn.inference.topdownmulticlasspredictor class method)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.from_trained_models"]], "from_trained_models() (sleap.nn.inference.topdownpredictor class method)": [[72, "sleap.nn.inference.TopDownPredictor.from_trained_models"]], "get_keras_model_path() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.get_keras_model_path"]], "get_model_output_stride() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.get_model_output_stride"]], "inference_layer (sleap.nn.inference.bottomupmulticlassinferencemodel attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceModel.inference_layer"]], "inference_layer (sleap.nn.inference.movenetinferencemodel attribute)": [[72, "sleap.nn.inference.MoveNetInferenceModel.inference_layer"]], "inference_model (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.inference_model"]], "inference_model (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.inference_model"]], "inference_model (sleap.nn.inference.movenetpredictor attribute)": [[72, "sleap.nn.inference.MoveNetPredictor.inference_model"]], "inference_model (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.inference_model"]], "inference_model (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.inference_model"]], "inference_model (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.inference_model"]], "input_scale (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.input_scale"]], "input_scale (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.input_scale"]], "input_scale (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.input_scale"]], "input_scale (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.input_scale"]], "input_scale (sleap.nn.inference.inferencelayer attribute)": [[72, "sleap.nn.inference.InferenceLayer.input_scale"]], "input_scale (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.input_scale"]], "input_scale (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.input_scale"]], "input_scale (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.input_scale"]], "instance_peaks (sleap.nn.inference.topdowninferencemodel attribute)": [[72, "sleap.nn.inference.TopDownInferenceModel.instance_peaks"]], "instance_peaks (sleap.nn.inference.topdownmulticlassinferencemodel attribute)": [[72, "sleap.nn.inference.TopDownMultiClassInferenceModel.instance_peaks"]], "integral_patch_size (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.integral_patch_size"]], "integral_patch_size (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.integral_patch_size"]], "integral_refinement (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.integral_refinement"]], "integral_refinement (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.integral_refinement"]], "integral_refinement (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.integral_refinement"]], "integral_refinement (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.integral_refinement"]], "integral_refinement (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.integral_refinement"]], "is_grayscale (sleap.nn.inference.bottomupmulticlasspredictor property)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.bottomuppredictor property)": [[72, "sleap.nn.inference.BottomUpPredictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.movenetpredictor property)": [[72, "sleap.nn.inference.MoveNetPredictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.predictor property)": [[72, "sleap.nn.inference.Predictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.singleinstancepredictor property)": [[72, "sleap.nn.inference.SingleInstancePredictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.topdownmulticlasspredictor property)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.is_grayscale"]], "is_grayscale (sleap.nn.inference.topdownpredictor property)": [[72, "sleap.nn.inference.TopDownPredictor.is_grayscale"]], "keras_model (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.keras_model"]], "keras_model (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.keras_model"]], "keras_model (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.keras_model"]], "keras_model (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.keras_model"]], "keras_model (sleap.nn.inference.inferencelayer attribute)": [[72, "sleap.nn.inference.InferenceLayer.keras_model"]], "keras_model (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.keras_model"]], "keras_model (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.keras_model"]], "keras_model (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.keras_model"]], "load_model() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.load_model"]], "main() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.main"]], "make_model_movenet() (in module sleap.nn.inference)": [[72, "sleap.nn.inference.make_model_movenet"]], "make_pipeline() (sleap.nn.inference.predictor method)": [[72, "sleap.nn.inference.Predictor.make_pipeline"]], "make_pipeline() (sleap.nn.inference.topdownmulticlasspredictor method)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.make_pipeline"]], "make_pipeline() (sleap.nn.inference.topdownpredictor method)": [[72, "sleap.nn.inference.TopDownPredictor.make_pipeline"]], "make_pipeline() (sleap.nn.inference.visualpredictor method)": [[72, "sleap.nn.inference.VisualPredictor.make_pipeline"]], "max_edge_length_ratio (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.max_edge_length_ratio"]], "max_instances (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.max_instances"]], "max_instances (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.max_instances"]], "max_instances (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.max_instances"]], "min_line_scores (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.min_line_scores"]], "model (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.model"]], "model_name (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.model_name"]], "model_name (sleap.nn.inference.movenetpredictor attribute)": [[72, "sleap.nn.inference.MoveNetPredictor.model_name"]], "offsets_ind (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.offsets_ind"]], "offsets_ind (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.offsets_ind"]], "offsets_ind (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.offsets_ind"]], "offsets_ind (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.offsets_ind"]], "offsets_ind (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.offsets_ind"]], "offsets_ind (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.offsets_ind"]], "optimal_grouping (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.optimal_grouping"]], "output_stride (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.output_stride"]], "output_stride (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.output_stride"]], "output_stride (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.output_stride"]], "output_stride (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.output_stride"]], "pad_to_stride (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.pad_to_stride"]], "pad_to_stride (sleap.nn.inference.inferencelayer attribute)": [[72, "sleap.nn.inference.InferenceLayer.pad_to_stride"]], "pad_to_stride (sleap.nn.inference.movenetinferencelayer attribute)": [[72, "sleap.nn.inference.MoveNetInferenceLayer.pad_to_stride"]], "pad_to_stride (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.pad_to_stride"]], "paf_line_points (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.paf_line_points"]], "paf_output_stride (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.paf_output_stride"]], "paf_scorer (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.paf_scorer"]], "pafs_ind (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.pafs_ind"]], "peak_threshold (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.peak_threshold"]], "peak_threshold (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.peak_threshold"]], "peak_threshold (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.peak_threshold"]], "peak_threshold (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.peak_threshold"]], "peak_threshold (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.peak_threshold"]], "peak_threshold (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.peak_threshold"]], "peak_threshold (sleap.nn.inference.movenetpredictor attribute)": [[72, "sleap.nn.inference.MoveNetPredictor.peak_threshold"]], "peak_threshold (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.peak_threshold"]], "peak_threshold (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.peak_threshold"]], "peak_threshold (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.peak_threshold"]], "peak_threshold (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.peak_threshold"]], "peak_threshold (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.peak_threshold"]], "pipeline (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.pipeline"]], "pipeline (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.pipeline"]], "pipeline (sleap.nn.inference.movenetpredictor attribute)": [[72, "sleap.nn.inference.MoveNetPredictor.pipeline"]], "pipeline (sleap.nn.inference.singleinstancepredictor attribute)": [[72, "sleap.nn.inference.SingleInstancePredictor.pipeline"]], "pipeline (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.pipeline"]], "pipeline (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.pipeline"]], "predict() (sleap.nn.inference.inferencemodel method)": [[72, "sleap.nn.inference.InferenceModel.predict"]], "predict() (sleap.nn.inference.predictor method)": [[72, "sleap.nn.inference.Predictor.predict"]], "predict() (sleap.nn.inference.visualpredictor method)": [[72, "sleap.nn.inference.VisualPredictor.predict"]], "predict_on_batch() (sleap.nn.inference.inferencemodel method)": [[72, "sleap.nn.inference.InferenceModel.predict_on_batch"]], "preprocess() (sleap.nn.inference.inferencelayer method)": [[72, "sleap.nn.inference.InferenceLayer.preprocess"]], "refinement (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.refinement"]], "refinement (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.refinement"]], "refinement (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.refinement"]], "refinement (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.refinement"]], "refinement (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.refinement"]], "refinement (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.refinement"]], "render() (sleap.nn.inference.ratecolumn method)": [[72, "sleap.nn.inference.RateColumn.render"]], "report_period (sleap.nn.inference.predictor property)": [[72, "sleap.nn.inference.Predictor.report_period"]], "return_class_maps (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_class_maps"]], "return_class_vectors (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.return_class_vectors"]], "return_confmaps (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.return_confmaps"]], "return_confmaps (sleap.nn.inference.bottomupmulticlassinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassInferenceLayer.return_confmaps"]], "return_confmaps (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.return_confmaps"]], "return_confmaps (sleap.nn.inference.findinstancepeaks attribute)": [[72, "sleap.nn.inference.FindInstancePeaks.return_confmaps"]], "return_confmaps (sleap.nn.inference.singleinstanceinferencelayer attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceLayer.return_confmaps"]], "return_confmaps (sleap.nn.inference.topdownmulticlassfindpeaks attribute)": [[72, "sleap.nn.inference.TopDownMultiClassFindPeaks.return_confmaps"]], "return_crops (sleap.nn.inference.centroidcrop attribute)": [[72, "sleap.nn.inference.CentroidCrop.return_crops"]], "return_paf_graph (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.return_paf_graph"]], "return_pafs (sleap.nn.inference.bottomupinferencelayer attribute)": [[72, "sleap.nn.inference.BottomUpInferenceLayer.return_pafs"]], "safely_generate() (sleap.nn.inference.visualpredictor method)": [[72, "sleap.nn.inference.VisualPredictor.safely_generate"]], "single_instance_layer (sleap.nn.inference.singleinstanceinferencemodel attribute)": [[72, "sleap.nn.inference.SingleInstanceInferenceModel.single_instance_layer"]], "sleap.nn.inference": [[72, "module-sleap.nn.inference"]], "tracker (sleap.nn.inference.bottomuppredictor attribute)": [[72, "sleap.nn.inference.BottomUpPredictor.tracker"]], "tracker (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.tracker"]], "tracker (sleap.nn.inference.topdownpredictor attribute)": [[72, "sleap.nn.inference.TopDownPredictor.tracker"]], "tracks (sleap.nn.inference.bottomupmulticlasspredictor attribute)": [[72, "sleap.nn.inference.BottomUpMultiClassPredictor.tracks"]], "tracks (sleap.nn.inference.topdownmulticlasspredictor attribute)": [[72, "sleap.nn.inference.TopDownMultiClassPredictor.tracks"]], "ohkmloss (class in sleap.nn.losses)": [[73, "sleap.nn.losses.OHKMLoss"]], "partloss (class in sleap.nn.losses)": [[73, "sleap.nn.losses.PartLoss"]], "call() (sleap.nn.losses.ohkmloss method)": [[73, "sleap.nn.losses.OHKMLoss.call"]], "channel_ind (sleap.nn.losses.partloss attribute)": [[73, "sleap.nn.losses.PartLoss.channel_ind"]], "compute_ohkm_loss() (in module sleap.nn.losses)": [[73, "sleap.nn.losses.compute_ohkm_loss"]], "from_config() (sleap.nn.losses.ohkmloss class method)": [[73, "sleap.nn.losses.OHKMLoss.from_config"]], "hard_to_easy_ratio (sleap.nn.losses.ohkmloss attribute)": [[73, "sleap.nn.losses.OHKMLoss.hard_to_easy_ratio"]], "loss_scale (sleap.nn.losses.ohkmloss attribute)": [[73, "sleap.nn.losses.OHKMLoss.loss_scale"]], "max_hard_keypoints (sleap.nn.losses.ohkmloss attribute)": [[73, "sleap.nn.losses.OHKMLoss.max_hard_keypoints"]], "min_hard_keypoints (sleap.nn.losses.ohkmloss attribute)": [[73, "sleap.nn.losses.OHKMLoss.min_hard_keypoints"]], "name (sleap.nn.losses.partloss attribute)": [[73, "sleap.nn.losses.PartLoss.name"]], "result() (sleap.nn.losses.partloss method)": [[73, "sleap.nn.losses.PartLoss.result"]], "sleap.nn.losses": [[73, "module-sleap.nn.losses"]], "update_state() (sleap.nn.losses.partloss method)": [[73, "sleap.nn.losses.PartLoss.update_state"]], "model (class in sleap.nn.model)": [[74, "sleap.nn.model.Model"]], "backbone (sleap.nn.model.model attribute)": [[74, "sleap.nn.model.Model.backbone"]], "from_config() (sleap.nn.model.model class method)": [[74, "sleap.nn.model.Model.from_config"]], "heads (sleap.nn.model.model attribute)": [[74, "sleap.nn.model.Model.heads"]], "keras_model (sleap.nn.model.model attribute)": [[74, "sleap.nn.model.Model.keras_model"]], "make_model() (sleap.nn.model.model method)": [[74, "sleap.nn.model.Model.make_model"]], "maximum_stride (sleap.nn.model.model property)": [[74, "sleap.nn.model.Model.maximum_stride"]], "sleap.nn.model": [[74, "module-sleap.nn.model"]], "edgeconnection (class in sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.EdgeConnection"]], "edgetype (class in sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.EdgeType"]], "pafscorer (class in sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.PAFScorer"]], "peakid (class in sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.PeakID"]], "assign_connections_to_instances() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.assign_connections_to_instances"]], "compute_distance_penalty() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.compute_distance_penalty"]], "dist_penalty_weight (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.dist_penalty_weight"]], "dst_node_ind (sleap.nn.paf_grouping.edgetype attribute)": [[75, "sleap.nn.paf_grouping.EdgeType.dst_node_ind"]], "dst_peak_ind (sleap.nn.paf_grouping.edgeconnection attribute)": [[75, "sleap.nn.paf_grouping.EdgeConnection.dst_peak_ind"]], "edge_inds (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.edge_inds"]], "edge_types (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.edge_types"]], "edges (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.edges"]], "from_config() (sleap.nn.paf_grouping.pafscorer class method)": [[75, "sleap.nn.paf_grouping.PAFScorer.from_config"]], "get_connection_candidates() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.get_connection_candidates"]], "get_paf_lines() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.get_paf_lines"]], "group_instances() (sleap.nn.paf_grouping.pafscorer method)": [[75, "sleap.nn.paf_grouping.PAFScorer.group_instances"]], "group_instances_batch() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.group_instances_batch"]], "group_instances_sample() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.group_instances_sample"]], "make_line_subs() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.make_line_subs"]], "make_predicted_instances() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.make_predicted_instances"]], "match_candidates() (sleap.nn.paf_grouping.pafscorer method)": [[75, "sleap.nn.paf_grouping.PAFScorer.match_candidates"]], "match_candidates_batch() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.match_candidates_batch"]], "match_candidates_sample() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.match_candidates_sample"]], "max_edge_length_ratio (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.max_edge_length_ratio"]], "min_instance_peaks (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.min_instance_peaks"]], "min_line_scores (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.min_line_scores"]], "n_edges (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.n_edges"]], "n_nodes (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.n_nodes"]], "n_points (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.n_points"]], "node_ind (sleap.nn.paf_grouping.peakid attribute)": [[75, "sleap.nn.paf_grouping.PeakID.node_ind"]], "pafs_stride (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.pafs_stride"]], "part_names (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.part_names"]], "peak_ind (sleap.nn.paf_grouping.peakid attribute)": [[75, "sleap.nn.paf_grouping.PeakID.peak_ind"]], "predict() (sleap.nn.paf_grouping.pafscorer method)": [[75, "sleap.nn.paf_grouping.PAFScorer.predict"]], "score (sleap.nn.paf_grouping.edgeconnection attribute)": [[75, "sleap.nn.paf_grouping.EdgeConnection.score"]], "score_paf_lines() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.score_paf_lines"]], "score_paf_lines() (sleap.nn.paf_grouping.pafscorer method)": [[75, "sleap.nn.paf_grouping.PAFScorer.score_paf_lines"]], "score_paf_lines_batch() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.score_paf_lines_batch"]], "sleap.nn.paf_grouping": [[75, "module-sleap.nn.paf_grouping"]], "sorted_edge_inds (sleap.nn.paf_grouping.pafscorer attribute)": [[75, "sleap.nn.paf_grouping.PAFScorer.sorted_edge_inds"]], "src_node_ind (sleap.nn.paf_grouping.edgetype attribute)": [[75, "sleap.nn.paf_grouping.EdgeType.src_node_ind"]], "src_peak_ind (sleap.nn.paf_grouping.edgeconnection attribute)": [[75, "sleap.nn.paf_grouping.EdgeConnection.src_peak_ind"]], "toposort_edges() (in module sleap.nn.paf_grouping)": [[75, "sleap.nn.paf_grouping.toposort_edges"]], "crop_bboxes() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.crop_bboxes"]], "find_global_peaks() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_global_peaks"]], "find_global_peaks_integral() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_global_peaks_integral"]], "find_global_peaks_rough() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_global_peaks_rough"]], "find_global_peaks_with_offsets() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_global_peaks_with_offsets"]], "find_local_peaks() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_local_peaks"]], "find_local_peaks_integral() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_local_peaks_integral"]], "find_local_peaks_rough() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_local_peaks_rough"]], "find_local_peaks_with_offsets() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_local_peaks_with_offsets"]], "find_offsets_local_direction() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.find_offsets_local_direction"]], "integral_regression() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.integral_regression"]], "make_gaussian_kernel() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.make_gaussian_kernel"]], "sleap.nn.peak_finding": [[76, "module-sleap.nn.peak_finding"]], "smooth_imgs() (in module sleap.nn.peak_finding)": [[76, "sleap.nn.peak_finding.smooth_imgs"]], "best_logical_device_name() (in module sleap.nn.system)": [[77, "sleap.nn.system.best_logical_device_name"]], "disable_preallocation() (in module sleap.nn.system)": [[77, "sleap.nn.system.disable_preallocation"]], "enable_preallocation() (in module sleap.nn.system)": [[77, "sleap.nn.system.enable_preallocation"]], "get_all_gpus() (in module sleap.nn.system)": [[77, "sleap.nn.system.get_all_gpus"]], "get_available_gpus() (in module sleap.nn.system)": [[77, "sleap.nn.system.get_available_gpus"]], "get_current_gpu() (in module sleap.nn.system)": [[77, "sleap.nn.system.get_current_gpu"]], "get_gpu_memory() (in module sleap.nn.system)": [[77, "sleap.nn.system.get_gpu_memory"]], "initialize_devices() (in module sleap.nn.system)": [[77, "sleap.nn.system.initialize_devices"]], "is_gpu_system() (in module sleap.nn.system)": [[77, "sleap.nn.system.is_gpu_system"]], "is_initialized() (in module sleap.nn.system)": [[77, "sleap.nn.system.is_initialized"]], "sleap.nn.system": [[77, "module-sleap.nn.system"]], "summary() (in module sleap.nn.system)": [[77, "sleap.nn.system.summary"]], "use_cpu_only() (in module sleap.nn.system)": [[77, "sleap.nn.system.use_cpu_only"]], "use_first_gpu() (in module sleap.nn.system)": [[77, "sleap.nn.system.use_first_gpu"]], "use_gpu() (in module sleap.nn.system)": [[77, "sleap.nn.system.use_gpu"]], "use_last_gpu() (in module sleap.nn.system)": [[77, "sleap.nn.system.use_last_gpu"]], "framematches (class in sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.FrameMatches"]], "match (class in sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.Match"]], "centroid_distance() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.centroid_distance"]], "connect_single_track_breaks() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.connect_single_track_breaks"]], "cost_matrix (sleap.nn.tracker.components.framematches attribute)": [[78, "sleap.nn.tracker.components.FrameMatches.cost_matrix"]], "cull_frame_instances() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.cull_frame_instances"]], "cull_instances() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.cull_instances"]], "factory_object_keypoint_similarity() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.factory_object_keypoint_similarity"]], "first_choice_matching() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.first_choice_matching"]], "from_candidate_instances() (sleap.nn.tracker.components.framematches class method)": [[78, "sleap.nn.tracker.components.FrameMatches.from_candidate_instances"]], "greedy_matching() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.greedy_matching"]], "has_only_first_choice_matches (sleap.nn.tracker.components.framematches property)": [[78, "sleap.nn.tracker.components.FrameMatches.has_only_first_choice_matches"]], "hungarian_matching() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.hungarian_matching"]], "instance_iou() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.instance_iou"]], "instance_similarity() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.instance_similarity"]], "matches (sleap.nn.tracker.components.framematches attribute)": [[78, "sleap.nn.tracker.components.FrameMatches.matches"]], "nms_fast() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.nms_fast"]], "normalized_instance_similarity() (in module sleap.nn.tracker.components)": [[78, "sleap.nn.tracker.components.normalized_instance_similarity"]], "sleap.nn.tracker.components": [[78, "module-sleap.nn.tracker.components"]], "unmatched_instances (sleap.nn.tracker.components.framematches attribute)": [[78, "sleap.nn.tracker.components.FrameMatches.unmatched_instances"]], "get_track_instance_matches() (in module sleap.nn.tracker.kalman)": [[79, "sleap.nn.tracker.kalman.get_track_instance_matches"]], "match_dict_from_match_function() (in module sleap.nn.tracker.kalman)": [[79, "sleap.nn.tracker.kalman.match_dict_from_match_function"]], "remove_second_bests_from_cost_matrix() (in module sleap.nn.tracker.kalman)": [[79, "sleap.nn.tracker.kalman.remove_second_bests_from_cost_matrix"]], "sleap.nn.tracker.kalman": [[79, "module-sleap.nn.tracker.kalman"]], "basetracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.BaseTracker"]], "flowcandidatemaker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.FlowCandidateMaker"]], "flowmaxtracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.FlowMaxTracker"]], "flowmaxtrackscandidatemaker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.FlowMaxTracksCandidateMaker"]], "flowtracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.FlowTracker"]], "kalmantracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.KalmanTracker"]], "simplecandidatemaker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.SimpleCandidateMaker"]], "simplemaxtracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.SimpleMaxTracker"]], "simplemaxtrackscandidatemaker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.SimpleMaxTracksCandidateMaker"]], "simpletracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.SimpleTracker"]], "trackcleaner (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.TrackCleaner"]], "tracker (class in sleap.nn.tracking)": [[80, "sleap.nn.tracking.Tracker"]], "candidate_maker (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.candidate_maker"]], "cleaner (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.cleaner"]], "final_pass() (sleap.nn.tracking.tracker method)": [[80, "sleap.nn.tracking.Tracker.final_pass"]], "flow_shift_instances() (sleap.nn.tracking.flowcandidatemaker static method)": [[80, "sleap.nn.tracking.FlowCandidateMaker.flow_shift_instances"]], "get_ref_instances() (sleap.nn.tracking.flowmaxtrackscandidatemaker static method)": [[80, "sleap.nn.tracking.FlowMaxTracksCandidateMaker.get_ref_instances"]], "get_shifted_instances() (sleap.nn.tracking.flowcandidatemaker method)": [[80, "sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances"]], "get_shifted_instances_from_earlier_time() (sleap.nn.tracking.flowcandidatemaker method)": [[80, "sleap.nn.tracking.FlowCandidateMaker.get_shifted_instances_from_earlier_time"]], "img_scale (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.img_scale"]], "instance_count (sleap.nn.tracking.trackcleaner attribute)": [[80, "sleap.nn.tracking.TrackCleaner.instance_count"]], "iou_threshold (sleap.nn.tracking.trackcleaner attribute)": [[80, "sleap.nn.tracking.TrackCleaner.iou_threshold"]], "is_valid (sleap.nn.tracking.kalmantracker property)": [[80, "sleap.nn.tracking.KalmanTracker.is_valid"]], "make_tracker() (sleap.nn.tracking.kalmantracker class method)": [[80, "sleap.nn.tracking.KalmanTracker.make_tracker"]], "matching_function (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.matching_function"]], "matching_function() (sleap.nn.tracking.flowmaxtracker method)": [[80, "sleap.nn.tracking.FlowMaxTracker.matching_function"]], "max_tracking (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.max_tracking"]], "max_tracks (sleap.nn.tracking.flowmaxtrackscandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowMaxTracksCandidateMaker.max_tracks"]], "min_new_track_points (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.min_new_track_points"]], "min_points (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.min_points"]], "of_max_levels (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.of_max_levels"]], "of_window_size (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.of_window_size"]], "prune_shifted_instances() (sleap.nn.tracking.flowcandidatemaker method)": [[80, "sleap.nn.tracking.FlowCandidateMaker.prune_shifted_instances"]], "robust_best_instance (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.robust_best_instance"]], "run_tracker() (in module sleap.nn.tracking)": [[80, "sleap.nn.tracking.run_tracker"]], "save_shifted_instances (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.save_shifted_instances"]], "similarity_function (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.similarity_function"]], "similarity_function() (sleap.nn.tracking.flowmaxtracker method)": [[80, "sleap.nn.tracking.FlowMaxTracker.similarity_function"]], "sleap.nn.tracking": [[80, "module-sleap.nn.tracking"]], "track() (sleap.nn.tracking.kalmantracker method)": [[80, "sleap.nn.tracking.KalmanTracker.track"]], "track() (sleap.nn.tracking.tracker method)": [[80, "sleap.nn.tracking.Tracker.track"]], "track_window (sleap.nn.tracking.flowcandidatemaker attribute)": [[80, "sleap.nn.tracking.FlowCandidateMaker.track_window"]], "track_window (sleap.nn.tracking.tracker attribute)": [[80, "sleap.nn.tracking.Tracker.track_window"]], "unique_tracks_in_queue (sleap.nn.tracking.tracker property)": [[80, "sleap.nn.tracking.Tracker.unique_tracks_in_queue"]], "bottomupmodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.BottomUpModelTrainer"]], "bottomupmulticlassmodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.BottomUpMultiClassModelTrainer"]], "centroidconfmapsmodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.CentroidConfmapsModelTrainer"]], "datareaders (class in sleap.nn.training)": [[81, "sleap.nn.training.DataReaders"]], "singleinstancemodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.SingleInstanceModelTrainer"]], "topdownmulticlassmodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.TopDownMultiClassModelTrainer"]], "topdownconfmapsmodeltrainer (class in sleap.nn.training)": [[81, "sleap.nn.training.TopdownConfmapsModelTrainer"]], "trainer (class in sleap.nn.training)": [[81, "sleap.nn.training.Trainer"]], "callbacks (sleap.nn.training.trainer property)": [[81, "sleap.nn.training.Trainer.callbacks"]], "cleanup() (sleap.nn.training.trainer method)": [[81, "sleap.nn.training.Trainer.cleanup"]], "config (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.config"]], "create_trainer_using_cli() (in module sleap.nn.training)": [[81, "sleap.nn.training.create_trainer_using_cli"]], "data_readers (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.data_readers"]], "evaluate() (sleap.nn.training.trainer method)": [[81, "sleap.nn.training.Trainer.evaluate"]], "from_config() (sleap.nn.training.datareaders class method)": [[81, "sleap.nn.training.DataReaders.from_config"]], "from_config() (sleap.nn.training.trainer class method)": [[81, "sleap.nn.training.Trainer.from_config"]], "from_labels() (sleap.nn.training.datareaders class method)": [[81, "sleap.nn.training.DataReaders.from_labels"]], "get_timestamp() (in module sleap.nn.training)": [[81, "sleap.nn.training.get_timestamp"]], "has_offsets (sleap.nn.training.bottomupmodeltrainer property)": [[81, "sleap.nn.training.BottomUpModelTrainer.has_offsets"]], "has_offsets (sleap.nn.training.bottomupmulticlassmodeltrainer property)": [[81, "sleap.nn.training.BottomUpMultiClassModelTrainer.has_offsets"]], "has_offsets (sleap.nn.training.centroidconfmapsmodeltrainer property)": [[81, "sleap.nn.training.CentroidConfmapsModelTrainer.has_offsets"]], "has_offsets (sleap.nn.training.singleinstancemodeltrainer property)": [[81, "sleap.nn.training.SingleInstanceModelTrainer.has_offsets"]], "has_offsets (sleap.nn.training.topdownmulticlassmodeltrainer property)": [[81, "sleap.nn.training.TopDownMultiClassModelTrainer.has_offsets"]], "has_offsets (sleap.nn.training.topdownconfmapsmodeltrainer property)": [[81, "sleap.nn.training.TopdownConfmapsModelTrainer.has_offsets"]], "initial_config (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.initial_config"]], "input_keys (sleap.nn.training.bottomupmodeltrainer property)": [[81, "sleap.nn.training.BottomUpModelTrainer.input_keys"]], "input_keys (sleap.nn.training.bottomupmulticlassmodeltrainer property)": [[81, "sleap.nn.training.BottomUpMultiClassModelTrainer.input_keys"]], "input_keys (sleap.nn.training.centroidconfmapsmodeltrainer property)": [[81, "sleap.nn.training.CentroidConfmapsModelTrainer.input_keys"]], "input_keys (sleap.nn.training.singleinstancemodeltrainer property)": [[81, "sleap.nn.training.SingleInstanceModelTrainer.input_keys"]], "input_keys (sleap.nn.training.topdownmulticlassmodeltrainer property)": [[81, "sleap.nn.training.TopDownMultiClassModelTrainer.input_keys"]], "input_keys (sleap.nn.training.topdownconfmapsmodeltrainer property)": [[81, "sleap.nn.training.TopdownConfmapsModelTrainer.input_keys"]], "input_keys (sleap.nn.training.trainer property)": [[81, "sleap.nn.training.Trainer.input_keys"]], "keras_model (sleap.nn.training.trainer property)": [[81, "sleap.nn.training.Trainer.keras_model"]], "main() (in module sleap.nn.training)": [[81, "sleap.nn.training.main"]], "model (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.model"]], "optimization_callbacks (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.optimization_callbacks"]], "output_callbacks (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.output_callbacks"]], "output_keys (sleap.nn.training.bottomupmodeltrainer property)": [[81, "sleap.nn.training.BottomUpModelTrainer.output_keys"]], "output_keys (sleap.nn.training.bottomupmulticlassmodeltrainer property)": [[81, "sleap.nn.training.BottomUpMultiClassModelTrainer.output_keys"]], "output_keys (sleap.nn.training.centroidconfmapsmodeltrainer property)": [[81, "sleap.nn.training.CentroidConfmapsModelTrainer.output_keys"]], "output_keys (sleap.nn.training.singleinstancemodeltrainer property)": [[81, "sleap.nn.training.SingleInstanceModelTrainer.output_keys"]], "output_keys (sleap.nn.training.topdownmulticlassmodeltrainer property)": [[81, "sleap.nn.training.TopDownMultiClassModelTrainer.output_keys"]], "output_keys (sleap.nn.training.topdownconfmapsmodeltrainer property)": [[81, "sleap.nn.training.TopdownConfmapsModelTrainer.output_keys"]], "output_keys (sleap.nn.training.trainer property)": [[81, "sleap.nn.training.Trainer.output_keys"]], "package() (sleap.nn.training.trainer method)": [[81, "sleap.nn.training.Trainer.package"]], "pipeline_builder (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.pipeline_builder"]], "run_path (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.run_path"]], "sanitize_scope_name() (in module sleap.nn.training)": [[81, "sleap.nn.training.sanitize_scope_name"]], "setup() (sleap.nn.training.trainer method)": [[81, "sleap.nn.training.Trainer.setup"]], "setup_checkpointing() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_checkpointing"]], "setup_losses() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_losses"]], "setup_metrics() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_metrics"]], "setup_new_run_folder() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_new_run_folder"]], "setup_optimization_callbacks() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_optimization_callbacks"]], "setup_optimizer() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_optimizer"]], "setup_output_callbacks() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_output_callbacks"]], "setup_tensorboard() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_tensorboard"]], "setup_visualization() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_visualization"]], "setup_zmq_callbacks() (in module sleap.nn.training)": [[81, "sleap.nn.training.setup_zmq_callbacks"]], "sleap.nn.training": [[81, "module-sleap.nn.training"]], "test_labels (sleap.nn.training.datareaders property)": [[81, "sleap.nn.training.DataReaders.test_labels"]], "test_labels_reader (sleap.nn.training.datareaders attribute)": [[81, "sleap.nn.training.DataReaders.test_labels_reader"]], "train() (sleap.nn.training.trainer method)": [[81, "sleap.nn.training.Trainer.train"]], "training_labels (sleap.nn.training.datareaders property)": [[81, "sleap.nn.training.DataReaders.training_labels"]], "training_labels_reader (sleap.nn.training.datareaders attribute)": [[81, "sleap.nn.training.DataReaders.training_labels_reader"]], "training_pipeline (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.training_pipeline"]], "training_viz_pipeline (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.training_viz_pipeline"]], "validation_labels (sleap.nn.training.datareaders property)": [[81, "sleap.nn.training.DataReaders.validation_labels"]], "validation_labels_reader (sleap.nn.training.datareaders attribute)": [[81, "sleap.nn.training.DataReaders.validation_labels_reader"]], "validation_pipeline (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.validation_pipeline"]], "validation_viz_pipeline (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.validation_viz_pipeline"]], "visualization_callbacks (sleap.nn.training.trainer attribute)": [[81, "sleap.nn.training.Trainer.visualization_callbacks"]], "compute_iou() (in module sleap.nn.utils)": [[82, "sleap.nn.utils.compute_iou"]], "group_array() (in module sleap.nn.utils)": [[82, "sleap.nn.utils.group_array"]], "match_points() (in module sleap.nn.utils)": [[82, "sleap.nn.utils.match_points"]], "reset_input_layer() (in module sleap.nn.utils)": [[82, "sleap.nn.utils.reset_input_layer"]], "sleap.nn.utils": [[82, "module-sleap.nn.utils"]], "tf_linear_sum_assignment() (in module sleap.nn.utils)": [[82, "sleap.nn.utils.tf_linear_sum_assignment"]], "generate_skeleton_preview_image() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.generate_skeleton_preview_image"]], "imgfig() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.imgfig"]], "plot_confmaps() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_confmaps"]], "plot_img() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_img"]], "plot_instance() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_instance"]], "plot_instances() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_instances"]], "plot_pafs() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_pafs"]], "plot_peaks() (in module sleap.nn.viz)": [[83, "sleap.nn.viz.plot_peaks"]], "sleap.nn.viz": [[83, "module-sleap.nn.viz"]], "edgetype (class in sleap.skeleton)": [[84, "sleap.skeleton.EdgeType"]], "node (class in sleap.skeleton)": [[84, "sleap.skeleton.Node"]], "skeleton (class in sleap.skeleton)": [[84, "sleap.skeleton.Skeleton"]], "skeletondecoder (class in sleap.skeleton)": [[84, "sleap.skeleton.SkeletonDecoder"]], "skeletonencoder (class in sleap.skeleton)": [[84, "sleap.skeleton.SkeletonEncoder"]], "_is_template (sleap.skeleton.skeleton attribute)": [[84, "sleap.skeleton.Skeleton._is_template"]], "_skeleton_idx (sleap.skeleton.skeleton attribute)": [[84, "sleap.skeleton.Skeleton._skeleton_idx"]], "add_edge() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.add_edge"]], "add_node() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.add_node"]], "add_nodes() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.add_nodes"]], "add_symmetry() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.add_symmetry"]], "as_node() (sleap.skeleton.node class method)": [[84, "sleap.skeleton.Node.as_node"]], "clear_edges() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.clear_edges"]], "decode() (sleap.skeleton.skeletondecoder class method)": [[84, "sleap.skeleton.SkeletonDecoder.decode"]], "decode_preview_image() (sleap.skeleton.skeletondecoder static method)": [[84, "sleap.skeleton.SkeletonDecoder.decode_preview_image"]], "delete_edge() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.delete_edge"]], "delete_node() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.delete_node"]], "delete_symmetry() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.delete_symmetry"]], "description (sleap.skeleton.skeleton attribute)": [[84, "sleap.skeleton.Skeleton.description"]], "edge_inds (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.edge_inds"]], "edge_names (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.edge_names"]], "edge_to_index() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.edge_to_index"]], "edges (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.edges"]], "edges_full (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.edges_full"]], "encode() (sleap.skeleton.skeletonencoder class method)": [[84, "sleap.skeleton.SkeletonEncoder.encode"]], "find_neighbors() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.find_neighbors"]], "find_node() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.find_node"]], "find_unique_nodes() (sleap.skeleton.skeleton static method)": [[84, "sleap.skeleton.Skeleton.find_unique_nodes"]], "from_dict() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.from_dict"]], "from_json() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.from_json"]], "from_names() (sleap.skeleton.node static method)": [[84, "sleap.skeleton.Node.from_names"]], "from_names_and_edge_inds() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.from_names_and_edge_inds"]], "get_symmetry() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.get_symmetry"]], "get_symmetry_name() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.get_symmetry_name"]], "graph (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.graph"]], "graph_symmetry (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.graph_symmetry"]], "has_edge() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.has_edge"]], "has_node() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.has_node"]], "has_nodes() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.has_nodes"]], "is_arborescence (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.is_arborescence"]], "is_template (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.is_template"]], "load_all_hdf5() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.load_all_hdf5"]], "load_hdf5() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.load_hdf5"]], "load_json() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.load_json"]], "load_mat() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.load_mat"]], "make_cattr() (sleap.skeleton.skeleton static method)": [[84, "sleap.skeleton.Skeleton.make_cattr"]], "matches() (sleap.skeleton.node method)": [[84, "sleap.skeleton.Node.matches"]], "matches() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.matches"]], "name (sleap.skeleton.node attribute)": [[84, "sleap.skeleton.Node.name"]], "name (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.name"]], "node_names (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.node_names"]], "node_to_index() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.node_to_index"]], "nodes (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.nodes"]], "preview_image (sleap.skeleton.skeleton attribute)": [[84, "sleap.skeleton.Skeleton.preview_image"]], "relabel_node() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.relabel_node"]], "relabel_nodes() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.relabel_nodes"]], "rename_skeleton() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.rename_skeleton"]], "save_all_hdf5() (sleap.skeleton.skeleton class method)": [[84, "sleap.skeleton.Skeleton.save_all_hdf5"]], "save_hdf5() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.save_hdf5"]], "save_json() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.save_json"]], "sleap.skeleton": [[84, "module-sleap.skeleton"]], "symmetric_inds (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.symmetric_inds"]], "symmetries (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.symmetries"]], "symmetries_full (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.symmetries_full"]], "symmetry_names (sleap.skeleton.skeleton property)": [[84, "sleap.skeleton.Skeleton.symmetry_names"]], "to_dict() (sleap.skeleton.skeleton static method)": [[84, "sleap.skeleton.Skeleton.to_dict"]], "to_json() (sleap.skeleton.skeleton method)": [[84, "sleap.skeleton.Skeleton.to_json"]], "weight (sleap.skeleton.node attribute)": [[84, "sleap.skeleton.Node.weight"]], "attr_to_dtype() (in module sleap.util)": [[85, "sleap.util.attr_to_dtype"]], "dict_cut() (in module sleap.util)": [[85, "sleap.util.dict_cut"]], "find_files_by_suffix() (in module sleap.util)": [[85, "sleap.util.find_files_by_suffix"]], "frame_list() (in module sleap.util)": [[85, "sleap.util.frame_list"]], "get_config_file() (in module sleap.util)": [[85, "sleap.util.get_config_file"]], "get_package_file() (in module sleap.util)": [[85, "sleap.util.get_package_file"]], "json_dumps() (in module sleap.util)": [[85, "sleap.util.json_dumps"]], "json_loads() (in module sleap.util)": [[85, "sleap.util.json_loads"]], "make_scoped_dictionary() (in module sleap.util)": [[85, "sleap.util.make_scoped_dictionary"]], "parse_uri_path() (in module sleap.util)": [[85, "sleap.util.parse_uri_path"]], "save_dict_to_hdf5() (in module sleap.util)": [[85, "sleap.util.save_dict_to_hdf5"]], "sleap.util": [[85, "module-sleap.util"]], "uniquify() (in module sleap.util)": [[85, "sleap.util.uniquify"]], "usable_cpu_count() (in module sleap.util)": [[85, "sleap.util.usable_cpu_count"]], "weak_filename_match() (in module sleap.util)": [[85, "sleap.util.weak_filename_match"]]}})
\ No newline at end of file
diff --git a/tutorials/analysis.html b/tutorials/analysis.html
index e7ef40783..b8152f9e7 100644
--- a/tutorials/analysis.html
+++ b/tutorials/analysis.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Export Data For Analysis &#8212; SLEAP (v1.3.3)</title>
+    <title>Export Data For Analysis &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/analysis';</script>
     <link rel="canonical" href="/develop/tutorials/analysis.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Guides" href="../guides/index.html" />
     <link rel="prev" title="Tracking instances across frames" href="proofreading.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -331,11 +329,11 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="export-data-for-analysis">
-<span id="export-analysis"></span><h1>Export Data For Analysis<a class="headerlink" href="#export-data-for-analysis" title="Permalink to this headline">#</a></h1>
+<span id="export-analysis"></span><h1>Export Data For Analysis<a class="headerlink" href="#export-data-for-analysis" title="Permalink to this heading">#</a></h1>
 <p>The easiest way to work with the data from SLEAP is to export an HDF5 file by choosing “<strong>Export Analysis HDF5…</strong>” in the “File” menu.</p>
 <p>See <a class="reference internal" href="../api/sleap.io.convert.html#module-sleap.io.convert" title="sleap.io.convert"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sleap.io.convert</span></code></a> for an explanation of the datasets inside the file.</p>
 <section id="matlab">
-<h2>MATLAB<a class="headerlink" href="#matlab" title="Permalink to this headline">#</a></h2>
+<h2>MATLAB<a class="headerlink" href="#matlab" title="Permalink to this heading">#</a></h2>
 <p>You can read this file in MATLAB like this:</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">occupancy_matrix</span> <span class="o">=</span> <span class="n">h5read</span><span class="p">(</span><span class="s1">&#39;path/to/analysis.h5&#39;</span><span class="p">,</span><span class="s1">&#39;/track_occupancy&#39;</span><span class="p">)</span>
 <span class="n">tracks_matrix</span> <span class="o">=</span> <span class="n">h5read</span><span class="p">(</span><span class="s1">&#39;path/to/analysis.h5&#39;</span><span class="p">,</span><span class="s1">&#39;/tracks&#39;</span><span class="p">)</span>
@@ -344,7 +342,7 @@ <h2>MATLAB<a class="headerlink" href="#matlab" title="Permalink to this headline
 <p>See <a class="reference external" href="https://www.mathworks.com/help/matlab/ref/h5read.html">here</a> for more information about working with HDF5 files in MATLAB.</p>
 </section>
 <section id="python">
-<h2>Python<a class="headerlink" href="#python" title="Permalink to this headline">#</a></h2>
+<h2>Python<a class="headerlink" href="#python" title="Permalink to this heading">#</a></h2>
 <p>To read the file in Python you’ll first need to install the <a class="reference external" href="http://docs.h5py.org/en/stable/">h5py package</a>. You can then read data from the file like this:</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">h5py</span>
 <span class="k">with</span> <span class="n">h5py</span><span class="o">.</span><span class="n">File</span><span class="p">(</span><span class="s1">&#39;path/to/analysis.h5&#39;</span><span class="p">,</span> <span class="s1">&#39;r&#39;</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
@@ -432,7 +430,7 @@ <h2>Python<a class="headerlink" href="#python" title="Permalink to this headline
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/assisted-labeling.html b/tutorials/assisted-labeling.html
index 43fbd5b91..6a2dfcef3 100644
--- a/tutorials/assisted-labeling.html
+++ b/tutorials/assisted-labeling.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Prediction-assisted labeling &#8212; SLEAP (v1.3.3)</title>
+    <title>Prediction-assisted labeling &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/assisted-labeling';</script>
     <link rel="canonical" href="/develop/tutorials/assisted-labeling.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Tracking instances across frames" href="proofreading.html" />
     <link rel="prev" title="Training and Inference" href="initial-training.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -332,7 +330,7 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="prediction-assisted-labeling">
-<span id="assisted-labeling"></span><h1>Prediction-assisted labeling<a class="headerlink" href="#prediction-assisted-labeling" title="Permalink to this headline">#</a></h1>
+<span id="assisted-labeling"></span><h1>Prediction-assisted labeling<a class="headerlink" href="#prediction-assisted-labeling" title="Permalink to this heading">#</a></h1>
 <p><em>Prediction-assisted labeling</em> has two main goals. First, it speeds up the labeling
 process as it is faster to correct a predicted instance which is mostly
 correct than it is to add a new instance from scratch. Second, it
@@ -346,7 +344,7 @@ <h2> Contents </h2>
 red on the seekbar.</p>
 </section>
 <section class="tex2jax_ignore mathjax_ignore" id="reviewing-and-fixing-predictions">
-<h1>Reviewing and fixing predictions<a class="headerlink" href="#reviewing-and-fixing-predictions" title="Permalink to this headline">#</a></h1>
+<h1>Reviewing and fixing predictions<a class="headerlink" href="#reviewing-and-fixing-predictions" title="Permalink to this heading">#</a></h1>
 <p>After you’ve successfully trained models and predicted some instances,
 you’ll get a message that inference has finished.
 Predictions will be marked with a thin black line on the seekbar, while frames
@@ -448,7 +446,7 @@ <h1>Reviewing and fixing predictions<a class="headerlink" href="#reviewing-and-f
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/initial-labeling.html b/tutorials/initial-labeling.html
index 8c9dbdd2d..bf7aa7adc 100644
--- a/tutorials/initial-labeling.html
+++ b/tutorials/initial-labeling.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Initial Labeling &#8212; SLEAP (v1.3.3)</title>
+    <title>Initial Labeling &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/initial-labeling';</script>
     <link rel="canonical" href="/develop/tutorials/initial-labeling.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Training and Inference" href="initial-training.html" />
     <link rel="prev" title="Creating a project" href="new-project.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -333,9 +331,9 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="initial-labeling">
-<span id="id1"></span><h1>Initial Labeling<a class="headerlink" href="#initial-labeling" title="Permalink to this headline">#</a></h1>
+<span id="id1"></span><h1>Initial Labeling<a class="headerlink" href="#initial-labeling" title="Permalink to this heading">#</a></h1>
 <section id="selecting-frames-to-label">
-<span id="selecting-frames"></span><h2>Selecting frames to label<a class="headerlink" href="#selecting-frames-to-label" title="Permalink to this headline">#</a></h2>
+<span id="selecting-frames"></span><h2>Selecting frames to label<a class="headerlink" href="#selecting-frames-to-label" title="Permalink to this heading">#</a></h2>
 <p>We start by assembling a candidate group of images to label. You can
 either pick your own frames or let the system suggest a set of frames
 using the “Labeling Suggestions” panel. SLEAP can give you random or
@@ -347,7 +345,7 @@ <h2> Contents </h2>
 <p>You can then move between suggested frames by double-clicking the frame information in the table or by clicking the “Prev” or “Next” buttons.</p>
 </section>
 <section id="labeling-the-first-frame">
-<span id="labeling-first-frame"></span><h2>Labeling the first frame<a class="headerlink" href="#labeling-the-first-frame" title="Permalink to this headline">#</a></h2>
+<span id="labeling-first-frame"></span><h2>Labeling the first frame<a class="headerlink" href="#labeling-the-first-frame" title="Permalink to this heading">#</a></h2>
 <p>Start by adding an <strong>instance</strong> of the skeleton to the current image by
 clicking the “<strong>New Instance</strong>” button in the Instances panel. The
 first instance will have its points located at random. Move the points
@@ -368,7 +366,7 @@ <h2> Contents </h2>
 identify things that look like the instances you didn’t label.</p>
 </section>
 <section id="saving">
-<h2>Saving<a class="headerlink" href="#saving" title="Permalink to this headline">#</a></h2>
+<h2>Saving<a class="headerlink" href="#saving" title="Permalink to this heading">#</a></h2>
 <p>Since this is a new project, you’ll need to select a location and name
 the first time you save. SLEAP will ask you to save before closing any
 project that has been changed to avoid losing any work. Note: There is
@@ -377,7 +375,7 @@ <h2>Saving<a class="headerlink" href="#saving" title="Permalink to this headline
 a copy of your project.</p>
 </section>
 <section id="labeling-more-frames">
-<span id="id2"></span><h2>Labeling more frames<a class="headerlink" href="#labeling-more-frames" title="Permalink to this headline">#</a></h2>
+<span id="id2"></span><h2>Labeling more frames<a class="headerlink" href="#labeling-more-frames" title="Permalink to this heading">#</a></h2>
 <p>After labeling the first frame and saving the project, it’s time to label
 more frames. Since you generated a list of suggested frames, you can go to the next frame in the labeling set by clicking “<strong>Next</strong>” under the list of suggested frames. You can also always pick a frame to label by using the seekbar under
 the video. Suggested and labeled frames will be shown as empty and filled blue bars, respectively.</p>
@@ -467,7 +465,7 @@ <h2>Saving<a class="headerlink" href="#saving" title="Permalink to this headline
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/initial-training.html b/tutorials/initial-training.html
index 6a65d078f..6ec38d4e6 100644
--- a/tutorials/initial-training.html
+++ b/tutorials/initial-training.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Training and Inference &#8212; SLEAP (v1.3.3)</title>
+    <title>Training and Inference &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/initial-training';</script>
     <link rel="canonical" href="/develop/tutorials/initial-training.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Prediction-assisted labeling" href="assisted-labeling.html" />
     <link rel="prev" title="Initial Labeling" href="initial-labeling.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -332,14 +330,14 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="training-and-inference">
-<span id="initial-training"></span><h1>Training and Inference<a class="headerlink" href="#training-and-inference" title="Permalink to this headline">#</a></h1>
+<span id="initial-training"></span><h1>Training and Inference<a class="headerlink" href="#training-and-inference" title="Permalink to this heading">#</a></h1>
 <p>After labeling about 10 frames and saving the project you can train your first model and start getting initial predictions.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p>This tutorial assumes you have a GPU in your local machine and that TensorFlow is able to use your GPU. If you don’t have a GPU or you’re having trouble getting it to work, you can run training and inference in the cloud. See our <a class="reference internal" href="../guides/colab.html#colab"><span class="std std-ref">Run training and inference on Colab</span></a> guide! Or take a look at our other <a class="reference internal" href="../guides/index.html#guides"><span class="std std-ref">Guides</span></a> about running SLEAP on remote machines.</p>
 </div>
 <section id="training-options">
-<span id="id1"></span><h2>Training Options<a class="headerlink" href="#training-options" title="Permalink to this headline">#</a></h2>
+<span id="id1"></span><h2>Training Options<a class="headerlink" href="#training-options" title="Permalink to this heading">#</a></h2>
 <p>When you’re ready to train you will have three choices for models: <em>single animal</em>, multi-animal <strong>top-down</strong>, or multi-animal <strong>bottom-up</strong>.</p>
 <p>In <strong>top-down</strong> mode for multiple animals, a network first finds each animal and then a separate network estimates the pose of each found animal:</p>
 <img alt="../_images/topdown_approach.jpg" src="../_images/topdown_approach.jpg" />
@@ -350,7 +348,7 @@ <h2> Contents </h2>
 <p>For more information about the types of models you can train, see <a class="reference internal" href="../guides/choosing-models.html#choosing-models"><span class="std std-ref">Configuring models</span></a>.</p>
 </section>
 <section id="start-training">
-<h2>Start Training<a class="headerlink" href="#start-training" title="Permalink to this headline">#</a></h2>
+<h2>Start Training<a class="headerlink" href="#start-training" title="Permalink to this heading">#</a></h2>
 <p>To run training on your local machine, select “<strong>Run Training…</strong>” from the “Predict”
 menu. For this tutorial, let’s use the default settings for training with the “top-down” pipeline and predict on 20 random frames.</p>
 <p><img alt="image6" src="../_images/training-dialog.jpg" /></p>
@@ -360,7 +358,7 @@ <h2>Start Training<a class="headerlink" href="#start-training" title="Permalink
 <p>Just for this tutorial, let’s stop each training session after about 10 epochs. This should take a minute or two for each model (assuming you have a usable GPU!), and should be good enough to get some initial predictions.</p>
 </section>
 <section id="inference">
-<span id="initial-inference"></span><h2>Inference<a class="headerlink" href="#inference" title="Permalink to this headline">#</a></h2>
+<span id="initial-inference"></span><h2>Inference<a class="headerlink" href="#inference" title="Permalink to this heading">#</a></h2>
 <p>After each model is trained, inference will run and if everything is successful, you should get a dialog telling you how many frames got predictions. Frames with labels will be marked in the seekbar, so try clicking on the newly marked frames or use the “<strong>Next Labeled Frame</strong>” command in the “Go” menu to step through frames with labels.</p>
 <p>Once you got the initial predictions from the trained models you enter into a human-in-the-loop training cycle, in which you receive predictions and use them to continue to label more examples. You will continue to iterate this process until you have achieved the desired pose labeling accuracy for your dataset. The goal at this stage is to get accurate pose predictions on individual frames. This means that the points for each body part are predicted in the correct locations and that the points are correctly grouped into distinct animal instances.</p>
 <p>Continue to <a class="reference internal" href="assisted-labeling.html#assisted-labeling"><span class="std std-ref">Prediction-assisted labeling</span></a>.</p>
@@ -441,7 +439,7 @@ <h2>Start Training<a class="headerlink" href="#start-training" title="Permalink
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/new-project.html b/tutorials/new-project.html
index 14b6d988c..d49ac24ac 100644
--- a/tutorials/new-project.html
+++ b/tutorials/new-project.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Creating a project &#8212; SLEAP (v1.3.3)</title>
+    <title>Creating a project &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/new-project';</script>
     <link rel="canonical" href="/develop/tutorials/new-project.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Initial Labeling" href="initial-labeling.html" />
     <link rel="prev" title="Tutorial" href="tutorial.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -332,11 +330,11 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section class="tex2jax_ignore mathjax_ignore" id="creating-a-project">
-<span id="new-project"></span><h1>Creating a project<a class="headerlink" href="#creating-a-project" title="Permalink to this headline">#</a></h1>
+<span id="new-project"></span><h1>Creating a project<a class="headerlink" href="#creating-a-project" title="Permalink to this heading">#</a></h1>
 <section id="starting-sleap">
-<h2>Starting SLEAP<a class="headerlink" href="#starting-sleap" title="Permalink to this headline">#</a></h2>
+<h2>Starting SLEAP<a class="headerlink" href="#starting-sleap" title="Permalink to this heading">#</a></h2>
 <p>If you haven’t installed SLEAP yet, see <a class="reference internal" href="../installation.html"><span class="doc std std-doc">Installation</span></a> for instructions.</p>
-<p>Once you have SLEAP installed, start by opening a terminal. If you installed via the recommended <a class="reference internal" href="../installation.html#conda-package"><span class="std std-doc">conda package</span></a> method, activate the environment with:</p>
+<p>Once you have SLEAP installed, start by opening a terminal. If you installed via the recommended <span class="xref myst"></span> method, activate the environment with:</p>
 <div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">conda</span> <span class="n">activate</span> <span class="n">sleap</span>
 </pre></div>
 </div>
@@ -358,7 +356,7 @@ <h2>Starting SLEAP<a class="headerlink" href="#starting-sleap" title="Permalink
 <p>When you first start SLEAP you’ll see a new, empty project.</p>
 </section>
 <section id="opening-a-video">
-<h2>Opening a video<a class="headerlink" href="#opening-a-video" title="Permalink to this headline">#</a></h2>
+<h2>Opening a video<a class="headerlink" href="#opening-a-video" title="Permalink to this heading">#</a></h2>
 <p>Add a video by clicking the “<strong>Add Video</strong>” button in the “<strong>Videos</strong>” panel
 on the right side of the main window, or by dragging-and-dropping your video file from its
 folder into the SLEAP GUI.</p>
@@ -371,7 +369,7 @@ <h2>Opening a video<a class="headerlink" href="#opening-a-video" title="Permalin
 <img alt="../_images/video-options.gif" src="../_images/video-options.gif" />
 </section>
 <section id="creating-a-skeleton">
-<span id="new-skeleton"></span><h2>Creating a Skeleton<a class="headerlink" href="#creating-a-skeleton" title="Permalink to this headline">#</a></h2>
+<span id="new-skeleton"></span><h2>Creating a Skeleton<a class="headerlink" href="#creating-a-skeleton" title="Permalink to this heading">#</a></h2>
 <p>Create a new <strong>skeleton</strong> using the “Skeleton” panel on the right side
 of the main window.</p>
 <p>Use the “<strong>New Node</strong>” button to add a node (i.e., joint or body part).
@@ -460,7 +458,7 @@ <h2>Opening a video<a class="headerlink" href="#opening-a-video" title="Permalin
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/proofreading.html b/tutorials/proofreading.html
index 1c412bbe5..32611f835 100644
--- a/tutorials/proofreading.html
+++ b/tutorials/proofreading.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Tracking instances across frames &#8212; SLEAP (v1.3.3)</title>
+    <title>Tracking instances across frames &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/proofreading';</script>
     <link rel="canonical" href="/develop/tutorials/proofreading.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Export Data For Analysis" href="analysis.html" />
     <link rel="prev" title="Prediction-assisted labeling" href="assisted-labeling.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -330,7 +328,7 @@ <h2> Contents </h2>
                 <article class="bd-article" role="main">
                   
   <section id="tracking-instances-across-frames">
-<span id="proofreading-tutorial"></span><h1>Tracking instances across frames<a class="headerlink" href="#tracking-instances-across-frames" title="Permalink to this headline">#</a></h1>
+<span id="proofreading-tutorial"></span><h1>Tracking instances across frames<a class="headerlink" href="#tracking-instances-across-frames" title="Permalink to this heading">#</a></h1>
 <p>When you’re satisfied with the predictions you’re getting, you can use your models to predict on more frames by selecting
 “<strong>Run Inference…</strong>” from the “Predict” menu. This will use the most
 recently trained set of models.</p>
@@ -344,7 +342,7 @@ <h2> Contents </h2>
 by using the dropdown menu on the tab for the relevant model type.</p>
 <p><img alt="model-selection" src="../_images/model-selection.jpg" /></p>
 <section id="track-proofreading">
-<span id="id1"></span><h2>Track proofreading<a class="headerlink" href="#track-proofreading" title="Permalink to this headline">#</a></h2>
+<span id="id1"></span><h2>Track proofreading<a class="headerlink" href="#track-proofreading" title="Permalink to this heading">#</a></h2>
 <p>Once you have predicted tracks, you’ll need to proofread these to ensure
 that the identities of instances across frames are correct. By default,
 predicted instances all appear in grey and yellow. Select “Color
@@ -455,7 +453,7 @@ <h2> Contents </h2>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>
diff --git a/tutorials/tutorial.html b/tutorials/tutorial.html
index 6e9ef0147..0998b6a68 100644
--- a/tutorials/tutorial.html
+++ b/tutorials/tutorial.html
@@ -3,13 +3,13 @@
 <!DOCTYPE html>
 
 
-<html >
+<html lang="en" >
 
   <head>
     <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.17.1: http://docutils.sourceforge.net/" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
 
-    <title>Tutorial &#8212; SLEAP (v1.3.3)</title>
+    <title>Tutorial &#8212; SLEAP (v1.4.1)</title>
   
   
   
@@ -32,8 +32,8 @@
     <link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
     <link rel="stylesheet" href="../_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
     <link rel="stylesheet" type="text/css" href="../_static/copybutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/togglebutton.css" />
-    <link rel="stylesheet" type="text/css" href="../_static/mystnb.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css" />
+    <link rel="stylesheet" type="text/css" href="../_static/css/tabs.css" />
   
   <!-- Pre-loaded scripts that we'll load fully later -->
   <link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
@@ -42,15 +42,12 @@
     <script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
     <script src="../_static/jquery.js"></script>
     <script src="../_static/underscore.js"></script>
+    <script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
     <script src="../_static/doctools.js"></script>
+    <script src="../_static/sphinx_highlight.js"></script>
     <script src="../_static/clipboard.min.js"></script>
     <script src="../_static/copybutton.js"></script>
-    <script>let toggleHintShow = 'Click to show';</script>
-    <script>let toggleHintHide = 'Click to hide';</script>
-    <script>let toggleOpenOnPrint = 'true';</script>
-    <script src="../_static/togglebutton.js"></script>
     <script src="../_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
-    <script>var togglebuttonSelector = '.toggle, .admonition.dropdown, .tag_hide_input div.cell_input, .tag_hide-input div.cell_input, .tag_hide_output div.cell_output, .tag_hide-output div.cell_output, .tag_hide_cell.cell, .tag_hide-cell.cell';</script>
     <script>DOCUMENTATION_OPTIONS.pagename = 'tutorials/tutorial';</script>
     <link rel="canonical" href="/develop/tutorials/tutorial.html" />
     <link rel="shortcut icon" href="../_static/favicon.ico"/>
@@ -59,7 +56,7 @@
     <link rel="next" title="Creating a project" href="new-project.html" />
     <link rel="prev" title="Installation" href="../installation.html" />
   <meta name="viewport" content="width=device-width, initial-scale=1"/>
-  <meta name="docsearch:language" content="None"/>
+  <meta name="docsearch:language" content="en"/>
   </head>
   
   
@@ -165,6 +162,7 @@
 <li class="toctree-l2"><a class="reference internal" href="../guides/colab.html">Run training and inference on Colab</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/custom-training.html">Creating a custom training profile</a></li>
 <li class="toctree-l2"><a class="reference internal" href="../guides/remote.html">Running SLEAP remotely</a></li>
+<li class="toctree-l2"><a class="reference internal" href="../guides/bonsai.html">Using Bonsai with SLEAP</a></li>
 </ul>
 </li>
 <li class="toctree-l1 has-children"><a class="reference internal" href="../notebooks/index.html">Notebooks</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
@@ -320,7 +318,7 @@ <h1>Tutorial</h1>
                 <article class="bd-article" role="main">
                   
   <section id="tutorial">
-<span id="id1"></span><h1>Tutorial<a class="headerlink" href="#tutorial" title="Permalink to this headline">#</a></h1>
+<span id="id1"></span><h1>Tutorial<a class="headerlink" href="#tutorial" title="Permalink to this heading">#</a></h1>
 <p>The tutorial will walk you through the entire SLEAP workflow as described in the <a class="reference internal" href="../overview.html#high-level-overview"><span class="std std-ref">Overview</span></a>. You can read through this tutorial and try running SLEAP on one of our <a class="reference external" href="https://github.com/talmolab/sleap-datasets">sample datasets</a>. Then you will be ready to start using SLEAP on your own data.</p>
 <div class="toctree-wrapper compound">
 <ul>
@@ -416,7 +414,7 @@ <h1>Tutorial</h1>
     
   <p class="copyright">
     
-      © Copyright 2019–2023, Talmo Lab.
+      © Copyright 2019–2024, Talmo Lab.
       <br/>
     
   </p>